diff options
Diffstat (limited to 'src')
868 files changed, 34921 insertions, 23050 deletions
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index 71efbb40d..61adbef28 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -32,7 +32,6 @@ if (MSVC) # /Zc:inline - Let codegen omit inline functions in object files # /Zc:throwingNew - Let codegen assume `operator new` (without std::nothrow) will never return null add_compile_options( - /W3 /MP /Zi /Zo @@ -43,6 +42,18 @@ if (MSVC) /Zc:externConstexpr /Zc:inline /Zc:throwingNew + + # Warnings + /W3 + /we4062 # enumerator 'identifier' in a switch of enum 'enumeration' is not handled + /we4101 # 'identifier': unreferenced local variable + /we4265 # 'class': class has virtual functions, but destructor is not virtual + /we4388 # signed/unsigned mismatch + /we4547 # 'operator' : operator before comma has no effect; expected operator with side-effect + /we4549 # 'operator1': operator before comma has no effect; did you intend 'operator2'? + /we4555 # Expression has no effect; expected expression with side-effect + /we4834 # Discarding return value of function with 'nodiscard' attribute + /we5038 # data member 'member1' will be initialized after data member 'member2' ) # /GS- - No stack buffer overflow checks @@ -56,9 +67,12 @@ else() -Werror=implicit-fallthrough -Werror=missing-declarations -Werror=reorder + -Werror=uninitialized + -Werror=unused-result -Wextra -Wmissing-declarations -Wno-attributes + -Wno-invalid-offsetof -Wno-unused-parameter ) diff --git a/src/audio_core/CMakeLists.txt b/src/audio_core/CMakeLists.txt index cb00ef60e..d1d177b51 100644 --- a/src/audio_core/CMakeLists.txt +++ b/src/audio_core/CMakeLists.txt @@ -44,6 +44,24 @@ add_library(audio_core STATIC create_target_directory_groups(audio_core) +if (NOT MSVC) + target_compile_options(audio_core PRIVATE + -Werror=conversion + -Werror=ignored-qualifiers + -Werror=implicit-fallthrough + -Werror=reorder + -Werror=sign-compare + -Werror=shadow + -Werror=unused-parameter + -Werror=unused-variable + + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable> + + -Wno-sign-conversion + ) +endif() + target_link_libraries(audio_core PUBLIC common core) target_link_libraries(audio_core PRIVATE SoundTouch) diff --git a/src/audio_core/algorithm/filter.cpp b/src/audio_core/algorithm/filter.cpp index f65bf64f7..01b8dff6b 100644 --- a/src/audio_core/algorithm/filter.cpp +++ b/src/audio_core/algorithm/filter.cpp @@ -31,8 +31,8 @@ Filter Filter::LowPass(double cutoff, double Q) { Filter::Filter() : Filter(1.0, 0.0, 0.0, 1.0, 0.0, 0.0) {} -Filter::Filter(double a0, double a1, double a2, double b0, double b1, double b2) - : a1(a1 / a0), a2(a2 / a0), b0(b0 / a0), b1(b1 / a0), b2(b2 / a0) {} +Filter::Filter(double a0_, double a1_, double a2_, double b0_, double b1_, double b2_) + : a1(a1_ / a0_), a2(a2_ / a0_), b0(b0_ / a0_), b1(b1_ / a0_), b2(b2_ / a0_) {} void Filter::Process(std::vector<s16>& signal) { const std::size_t num_frames = signal.size() / 2; @@ -55,7 +55,8 @@ void Filter::Process(std::vector<s16>& signal) { /// @param total_count The total number of biquads to be cascaded. /// @param index 0-index of the biquad to calculate the Q value for. static double CascadingBiquadQ(std::size_t total_count, std::size_t index) { - const double pole = M_PI * (2 * index + 1) / (4.0 * total_count); + const auto pole = + M_PI * static_cast<double>(2 * index + 1) / (4.0 * static_cast<double>(total_count)); return 1.0 / (2.0 * std::cos(pole)); } @@ -68,7 +69,7 @@ CascadingFilter CascadingFilter::LowPass(double cutoff, std::size_t cascade_size } CascadingFilter::CascadingFilter() = default; -CascadingFilter::CascadingFilter(std::vector<Filter> filters) : filters(std::move(filters)) {} +CascadingFilter::CascadingFilter(std::vector<Filter> filters_) : filters(std::move(filters_)) {} void CascadingFilter::Process(std::vector<s16>& signal) { for (auto& filter : filters) { diff --git a/src/audio_core/algorithm/filter.h b/src/audio_core/algorithm/filter.h index 3546d149b..a291fe79b 100644 --- a/src/audio_core/algorithm/filter.h +++ b/src/audio_core/algorithm/filter.h @@ -25,7 +25,7 @@ public: /// Passthrough filter. Filter(); - Filter(double a0, double a1, double a2, double b0, double b1, double b2); + Filter(double a0_, double a1_, double a2_, double b0_, double b1_, double b2_); void Process(std::vector<s16>& signal); @@ -51,7 +51,7 @@ public: /// Passthrough. CascadingFilter(); - explicit CascadingFilter(std::vector<Filter> filters); + explicit CascadingFilter(std::vector<Filter> filters_); void Process(std::vector<s16>& signal); diff --git a/src/audio_core/algorithm/interpolate.cpp b/src/audio_core/algorithm/interpolate.cpp index 689a54508..3b4144e21 100644 --- a/src/audio_core/algorithm/interpolate.cpp +++ b/src/audio_core/algorithm/interpolate.cpp @@ -146,7 +146,7 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, return {}; if (ratio <= 0) { - LOG_CRITICAL(Audio, "Nonsensical interpolation ratio {}", ratio); + LOG_ERROR(Audio, "Nonsensical interpolation ratio {}", ratio); return input; } @@ -164,7 +164,8 @@ std::vector<s16> Interpolate(InterpolationState& state, std::vector<s16> input, const std::size_t num_frames{input.size() / 2}; std::vector<s16> output; - output.reserve(static_cast<std::size_t>(input.size() / ratio + InterpolationState::taps)); + output.reserve(static_cast<std::size_t>(static_cast<double>(input.size()) / ratio + + InterpolationState::taps)); for (std::size_t frame{}; frame < num_frames; ++frame) { const std::size_t lut_index{(state.fraction >> 8) * InterpolationState::taps}; @@ -217,7 +218,7 @@ void Resample(s32* output, const s32* input, s32 pitch, s32& fraction, std::size const auto l2 = lut[lut_index + 2]; const auto l3 = lut[lut_index + 3]; - const auto s0 = static_cast<s32>(input[index]); + const auto s0 = static_cast<s32>(input[index + 0]); const auto s1 = static_cast<s32>(input[index + 1]); const auto s2 = static_cast<s32>(input[index + 2]); const auto s3 = static_cast<s32>(input[index + 3]); diff --git a/src/audio_core/audio_out.cpp b/src/audio_core/audio_out.cpp index 8619a3f03..fe3a898ad 100644 --- a/src/audio_core/audio_out.cpp +++ b/src/audio_core/audio_out.cpp @@ -43,6 +43,10 @@ std::vector<Buffer::Tag> AudioOut::GetTagsAndReleaseBuffers(StreamPtr stream, return stream->GetTagsAndReleaseBuffers(max_count); } +std::vector<Buffer::Tag> AudioOut::GetTagsAndReleaseBuffers(StreamPtr stream) { + return stream->GetTagsAndReleaseBuffers(); +} + void AudioOut::StartStream(StreamPtr stream) { stream->Play(); } diff --git a/src/audio_core/audio_out.h b/src/audio_core/audio_out.h index b07588287..6ce08cd0d 100644 --- a/src/audio_core/audio_out.h +++ b/src/audio_core/audio_out.h @@ -31,6 +31,9 @@ public: /// Returns a vector of recently released buffers specified by tag for the specified stream std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(StreamPtr stream, std::size_t max_count); + /// Returns a vector of all recently released buffers specified by tag for the specified stream + std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(StreamPtr stream); + /// Starts an audio stream for playback void StartStream(StreamPtr stream); diff --git a/src/audio_core/audio_renderer.cpp b/src/audio_core/audio_renderer.cpp index 56dc892b1..d2ce8c814 100644 --- a/src/audio_core/audio_renderer.cpp +++ b/src/audio_core/audio_renderer.cpp @@ -2,43 +2,90 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <limits> #include <vector> -#include "audio_core/algorithm/interpolate.h" + #include "audio_core/audio_out.h" #include "audio_core/audio_renderer.h" -#include "audio_core/codec.h" #include "audio_core/common.h" #include "audio_core/info_updater.h" #include "audio_core/voice_context.h" -#include "common/assert.h" #include "common/logging/log.h" -#include "core/core.h" -#include "core/hle/kernel/writable_event.h" #include "core/memory.h" #include "core/settings.h" +namespace { +[[nodiscard]] static constexpr s16 ClampToS16(s32 value) { + return static_cast<s16>(std::clamp(value, s32{std::numeric_limits<s16>::min()}, + s32{std::numeric_limits<s16>::max()})); +} + +[[nodiscard]] static constexpr s16 Mix2To1(s16 l_channel, s16 r_channel) { + // Mix 50% from left and 50% from right channel + constexpr float l_mix_amount = 50.0f / 100.0f; + constexpr float r_mix_amount = 50.0f / 100.0f; + return ClampToS16(static_cast<s32>((static_cast<float>(l_channel) * l_mix_amount) + + (static_cast<float>(r_channel) * r_mix_amount))); +} + +[[nodiscard]] static constexpr std::tuple<s16, s16> Mix6To2(s16 fl_channel, s16 fr_channel, + s16 fc_channel, + [[maybe_unused]] s16 lf_channel, + s16 bl_channel, s16 br_channel) { + // Front channels are mixed 36.94%, Center channels are mixed to be 26.12% & the back channels + // are mixed to be 36.94% + + constexpr float front_mix_amount = 36.94f / 100.0f; + constexpr float center_mix_amount = 26.12f / 100.0f; + constexpr float back_mix_amount = 36.94f / 100.0f; + + // Mix 50% from left and 50% from right channel + const auto left = front_mix_amount * static_cast<float>(fl_channel) + + center_mix_amount * static_cast<float>(fc_channel) + + back_mix_amount * static_cast<float>(bl_channel); + + const auto right = front_mix_amount * static_cast<float>(fr_channel) + + center_mix_amount * static_cast<float>(fc_channel) + + back_mix_amount * static_cast<float>(br_channel); + + return {ClampToS16(static_cast<s32>(left)), ClampToS16(static_cast<s32>(right))}; +} + +[[nodiscard]] static constexpr std::tuple<s16, s16> Mix6To2WithCoefficients( + s16 fl_channel, s16 fr_channel, s16 fc_channel, s16 lf_channel, s16 bl_channel, s16 br_channel, + const std::array<float_le, 4>& coeff) { + const auto left = + static_cast<float>(fl_channel) * coeff[0] + static_cast<float>(fc_channel) * coeff[1] + + static_cast<float>(lf_channel) * coeff[2] + static_cast<float>(bl_channel) * coeff[0]; + + const auto right = + static_cast<float>(fr_channel) * coeff[0] + static_cast<float>(fc_channel) * coeff[1] + + static_cast<float>(lf_channel) * coeff[2] + static_cast<float>(br_channel) * coeff[0]; + + return {ClampToS16(static_cast<s32>(left)), ClampToS16(static_cast<s32>(right))}; +} + +} // namespace + namespace AudioCore { AudioRenderer::AudioRenderer(Core::Timing::CoreTiming& core_timing, Core::Memory::Memory& memory_, AudioCommon::AudioRendererParameter params, - std::shared_ptr<Kernel::WritableEvent> buffer_event, + Stream::ReleaseCallback&& release_callback, std::size_t instance_number) - : worker_params{params}, buffer_event{buffer_event}, - memory_pool_info(params.effect_count + params.voice_count * 4), + : worker_params{params}, memory_pool_info(params.effect_count + params.voice_count * 4), voice_context(params.voice_count), effect_context(params.effect_count), mix_context(), sink_context(params.sink_count), splitter_context(), voices(params.voice_count), memory{memory_}, command_generator(worker_params, voice_context, mix_context, splitter_context, effect_context, - memory), - temp_mix_buffer(AudioCommon::TOTAL_TEMP_MIX_SIZE) { + memory) { behavior_info.SetUserRevision(params.revision); splitter_context.Initialize(behavior_info, params.splitter_count, params.num_splitter_send_channels); mix_context.Initialize(behavior_info, params.submix_count + 1, params.effect_count); audio_out = std::make_unique<AudioCore::AudioOut>(); - stream = - audio_out->OpenStream(core_timing, params.sample_rate, AudioCommon::STREAM_NUM_CHANNELS, - fmt::format("AudioRenderer-Instance{}", instance_number), - [=]() { buffer_event->Signal(); }); + stream = audio_out->OpenStream( + core_timing, params.sample_rate, AudioCommon::STREAM_NUM_CHANNELS, + fmt::format("AudioRenderer-Instance{}", instance_number), std::move(release_callback)); audio_out->StartStream(stream); QueueMixedBuffer(0); @@ -65,10 +112,6 @@ Stream::State AudioRenderer::GetStreamState() const { return stream->GetState(); } -static constexpr s16 ClampToS16(s32 value) { - return static_cast<s16>(std::clamp(value, -32768, 32767)); -} - ResultCode AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_params, std::vector<u8>& output_params) { @@ -107,8 +150,8 @@ ResultCode AudioRenderer::UpdateAudioRenderer(const std::vector<u8>& input_param } } - auto mix_result = info_updater.UpdateMixes(mix_context, worker_params.mix_buffer_count, - splitter_context, effect_context); + const auto mix_result = info_updater.UpdateMixes(mix_context, worker_params.mix_buffer_count, + splitter_context, effect_context); if (mix_result.IsError()) { LOG_ERROR(Audio, "Failed to update mix parameters"); @@ -197,20 +240,22 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) { for (std::size_t i = 0; i < BUFFER_SIZE; i++) { if (channel_count == 1) { const auto sample = ClampToS16(mix_buffers[0][i]); - buffer[i * stream_channel_count + 0] = sample; - if (stream_channel_count > 1) { - buffer[i * stream_channel_count + 1] = sample; + + // Place sample in all channels + for (u32 channel = 0; channel < stream_channel_count; channel++) { + buffer[i * stream_channel_count + channel] = sample; } + if (stream_channel_count == 6) { - buffer[i * stream_channel_count + 2] = sample; - buffer[i * stream_channel_count + 4] = sample; - buffer[i * stream_channel_count + 5] = sample; + // Output stream has a LF channel, mute it! + buffer[i * stream_channel_count + 3] = 0; } + } else if (channel_count == 2) { const auto l_sample = ClampToS16(mix_buffers[0][i]); const auto r_sample = ClampToS16(mix_buffers[1][i]); if (stream_channel_count == 1) { - buffer[i * stream_channel_count + 0] = l_sample; + buffer[i * stream_channel_count + 0] = Mix2To1(l_sample, r_sample); } else if (stream_channel_count == 2) { buffer[i * stream_channel_count + 0] = l_sample; buffer[i * stream_channel_count + 1] = r_sample; @@ -218,8 +263,8 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) { buffer[i * stream_channel_count + 0] = l_sample; buffer[i * stream_channel_count + 1] = r_sample; - buffer[i * stream_channel_count + 2] = - ClampToS16((static_cast<s32>(l_sample) + static_cast<s32>(r_sample)) / 2); + // Combine both left and right channels to the center channel + buffer[i * stream_channel_count + 2] = Mix2To1(l_sample, r_sample); buffer[i * stream_channel_count + 4] = l_sample; buffer[i * stream_channel_count + 5] = r_sample; @@ -234,17 +279,25 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) { const auto br_sample = ClampToS16(mix_buffers[5][i]); if (stream_channel_count == 1) { - buffer[i * stream_channel_count + 0] = fc_sample; + // Games seem to ignore the center channel half the time, we use the front left + // and right channel for mixing as that's where majority of the audio goes + buffer[i * stream_channel_count + 0] = Mix2To1(fl_sample, fr_sample); } else if (stream_channel_count == 2) { - buffer[i * stream_channel_count + 0] = - static_cast<s16>(0.3694f * static_cast<float>(fl_sample) + - 0.2612f * static_cast<float>(fc_sample) + - 0.3694f * static_cast<float>(bl_sample)); - buffer[i * stream_channel_count + 1] = - static_cast<s16>(0.3694f * static_cast<float>(fr_sample) + - 0.2612f * static_cast<float>(fc_sample) + - 0.3694f * static_cast<float>(br_sample)); + // Mix all channels into 2 channels + if (sink_context.HasDownMixingCoefficients()) { + const auto [left, right] = Mix6To2WithCoefficients( + fl_sample, fr_sample, fc_sample, lf_sample, bl_sample, br_sample, + sink_context.GetDownmixCoefficients()); + buffer[i * stream_channel_count + 0] = left; + buffer[i * stream_channel_count + 1] = right; + } else { + const auto [left, right] = Mix6To2(fl_sample, fr_sample, fc_sample, + lf_sample, bl_sample, br_sample); + buffer[i * stream_channel_count + 0] = left; + buffer[i * stream_channel_count + 1] = right; + } } else if (stream_channel_count == 6) { + // Pass through buffer[i * stream_channel_count + 0] = fl_sample; buffer[i * stream_channel_count + 1] = fr_sample; buffer[i * stream_channel_count + 2] = fc_sample; @@ -262,7 +315,7 @@ void AudioRenderer::QueueMixedBuffer(Buffer::Tag tag) { } void AudioRenderer::ReleaseAndQueueBuffers() { - const auto released_buffers{audio_out->GetTagsAndReleaseBuffers(stream, 2)}; + const auto released_buffers{audio_out->GetTagsAndReleaseBuffers(stream)}; for (const auto& tag : released_buffers) { QueueMixedBuffer(tag); } diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h index 2bca795ba..18567f618 100644 --- a/src/audio_core/audio_renderer.h +++ b/src/audio_core/audio_renderer.h @@ -21,53 +21,41 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "common/swap.h" -#include "core/hle/kernel/object.h" #include "core/hle/result.h" namespace Core::Timing { class CoreTiming; } -namespace Kernel { -class WritableEvent; -} - namespace Core::Memory { class Memory; } namespace AudioCore { -using DSPStateHolder = std::array<VoiceState*, 6>; +using DSPStateHolder = std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT>; class AudioOut; -struct RendererInfo { - u64_le elasped_frame_count{}; - INSERT_PADDING_WORDS(2); -}; -static_assert(sizeof(RendererInfo) == 0x10, "RendererInfo is an invalid size"); - class AudioRenderer { public: AudioRenderer(Core::Timing::CoreTiming& core_timing, Core::Memory::Memory& memory_, AudioCommon::AudioRendererParameter params, - std::shared_ptr<Kernel::WritableEvent> buffer_event, std::size_t instance_number); + Stream::ReleaseCallback&& release_callback, std::size_t instance_number); ~AudioRenderer(); - ResultCode UpdateAudioRenderer(const std::vector<u8>& input_params, - std::vector<u8>& output_params); + [[nodiscard]] ResultCode UpdateAudioRenderer(const std::vector<u8>& input_params, + std::vector<u8>& output_params); void QueueMixedBuffer(Buffer::Tag tag); void ReleaseAndQueueBuffers(); - u32 GetSampleRate() const; - u32 GetSampleCount() const; - u32 GetMixBufferCount() const; - Stream::State GetStreamState() const; + [[nodiscard]] u32 GetSampleRate() const; + [[nodiscard]] u32 GetSampleCount() const; + [[nodiscard]] u32 GetMixBufferCount() const; + [[nodiscard]] Stream::State GetStreamState() const; private: BehaviorInfo behavior_info{}; AudioCommon::AudioRendererParameter worker_params; - std::shared_ptr<Kernel::WritableEvent> buffer_event; std::vector<ServerMemoryPoolInfo> memory_pool_info; VoiceContext voice_context; EffectContext effect_context; @@ -80,7 +68,6 @@ private: Core::Memory::Memory& memory; CommandGenerator command_generator; std::size_t elapsed_frame_count{}; - std::vector<s32> temp_mix_buffer{}; }; } // namespace AudioCore diff --git a/src/audio_core/behavior_info.cpp b/src/audio_core/behavior_info.cpp index 5d62adb0b..3c2e3e6f1 100644 --- a/src/audio_core/behavior_info.cpp +++ b/src/audio_core/behavior_info.cpp @@ -57,15 +57,15 @@ bool BehaviorInfo::IsLongSizePreDelaySupported() const { return AudioCommon::IsRevisionSupported(3, user_revision); } -bool BehaviorInfo::IsAudioRenererProcessingTimeLimit80PercentSupported() const { +bool BehaviorInfo::IsAudioRendererProcessingTimeLimit80PercentSupported() const { return AudioCommon::IsRevisionSupported(5, user_revision); } -bool BehaviorInfo::IsAudioRenererProcessingTimeLimit75PercentSupported() const { +bool BehaviorInfo::IsAudioRendererProcessingTimeLimit75PercentSupported() const { return AudioCommon::IsRevisionSupported(4, user_revision); } -bool BehaviorInfo::IsAudioRenererProcessingTimeLimit70PercentSupported() const { +bool BehaviorInfo::IsAudioRendererProcessingTimeLimit70PercentSupported() const { return AudioCommon::IsRevisionSupported(1, user_revision); } diff --git a/src/audio_core/behavior_info.h b/src/audio_core/behavior_info.h index 50948e8df..5a96bf75e 100644 --- a/src/audio_core/behavior_info.h +++ b/src/audio_core/behavior_info.h @@ -43,22 +43,22 @@ public: void ClearError(); void UpdateFlags(u64_le dest_flags); void SetUserRevision(u32_le revision); - u32_le GetUserRevision() const; - u32_le GetProcessRevision() const; + [[nodiscard]] u32_le GetUserRevision() const; + [[nodiscard]] u32_le GetProcessRevision() const; - bool IsAdpcmLoopContextBugFixed() const; - bool IsSplitterSupported() const; - bool IsLongSizePreDelaySupported() const; - bool IsAudioRenererProcessingTimeLimit80PercentSupported() const; - bool IsAudioRenererProcessingTimeLimit75PercentSupported() const; - bool IsAudioRenererProcessingTimeLimit70PercentSupported() const; - bool IsElapsedFrameCountSupported() const; - bool IsMemoryPoolForceMappingEnabled() const; - bool IsFlushVoiceWaveBuffersSupported() const; - bool IsVoicePlayedSampleCountResetAtLoopPointSupported() const; - bool IsVoicePitchAndSrcSkippedSupported() const; - bool IsMixInParameterDirtyOnlyUpdateSupported() const; - bool IsSplitterBugFixed() const; + [[nodiscard]] bool IsAdpcmLoopContextBugFixed() const; + [[nodiscard]] bool IsSplitterSupported() const; + [[nodiscard]] bool IsLongSizePreDelaySupported() const; + [[nodiscard]] bool IsAudioRendererProcessingTimeLimit80PercentSupported() const; + [[nodiscard]] bool IsAudioRendererProcessingTimeLimit75PercentSupported() const; + [[nodiscard]] bool IsAudioRendererProcessingTimeLimit70PercentSupported() const; + [[nodiscard]] bool IsElapsedFrameCountSupported() const; + [[nodiscard]] bool IsMemoryPoolForceMappingEnabled() const; + [[nodiscard]] bool IsFlushVoiceWaveBuffersSupported() const; + [[nodiscard]] bool IsVoicePlayedSampleCountResetAtLoopPointSupported() const; + [[nodiscard]] bool IsVoicePitchAndSrcSkippedSupported() const; + [[nodiscard]] bool IsMixInParameterDirtyOnlyUpdateSupported() const; + [[nodiscard]] bool IsSplitterBugFixed() const; void CopyErrorInfo(OutParams& dst); private: diff --git a/src/audio_core/buffer.h b/src/audio_core/buffer.h index 5ee09e9aa..ccc46ef82 100644 --- a/src/audio_core/buffer.h +++ b/src/audio_core/buffer.h @@ -18,7 +18,7 @@ class Buffer { public: using Tag = u64; - Buffer(Tag tag, std::vector<s16>&& samples) : tag{tag}, samples{std::move(samples)} {} + Buffer(Tag tag_, std::vector<s16>&& samples_) : tag{tag_}, samples{std::move(samples_)} {} /// Returns the raw audio data for the buffer std::vector<s16>& GetSamples() { diff --git a/src/audio_core/codec.cpp b/src/audio_core/codec.cpp index c5a0d98ce..2fb91c13a 100644 --- a/src/audio_core/codec.cpp +++ b/src/audio_core/codec.cpp @@ -16,8 +16,9 @@ std::vector<s16> DecodeADPCM(const u8* const data, std::size_t size, const ADPCM constexpr std::size_t FRAME_LEN = 8; constexpr std::size_t SAMPLES_PER_FRAME = 14; - constexpr std::array<int, 16> SIGNED_NIBBLES = { - {0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1}}; + static constexpr std::array<int, 16> SIGNED_NIBBLES{ + 0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1, + }; const std::size_t sample_count = (size / FRAME_LEN) * SAMPLES_PER_FRAME; const std::size_t ret_size = diff --git a/src/audio_core/codec.h b/src/audio_core/codec.h index ef2ce01a8..9507abb1b 100644 --- a/src/audio_core/codec.h +++ b/src/audio_core/codec.h @@ -38,7 +38,7 @@ using ADPCM_Coeff = std::array<s16, 16>; * @param state ADPCM state, this is updated with new state * @return Decoded stereo signed PCM16 data, sample_count in length */ -std::vector<s16> DecodeADPCM(const u8* const data, std::size_t size, const ADPCM_Coeff& coeff, +std::vector<s16> DecodeADPCM(const u8* data, std::size_t size, const ADPCM_Coeff& coeff, ADPCMState& state); }; // namespace AudioCore::Codec diff --git a/src/audio_core/command_generator.cpp b/src/audio_core/command_generator.cpp index 8f7da49e6..a4a9a757d 100644 --- a/src/audio_core/command_generator.cpp +++ b/src/audio_core/command_generator.cpp @@ -67,12 +67,12 @@ s32 ApplyMixDepop(s32* output, s32 first_sample, s32 delta, s32 sample_count) { } // namespace -CommandGenerator::CommandGenerator(AudioCommon::AudioRendererParameter& worker_params, - VoiceContext& voice_context, MixContext& mix_context, - SplitterContext& splitter_context, EffectContext& effect_context, - Core::Memory::Memory& memory) - : worker_params(worker_params), voice_context(voice_context), mix_context(mix_context), - splitter_context(splitter_context), effect_context(effect_context), memory(memory), +CommandGenerator::CommandGenerator(AudioCommon::AudioRendererParameter& worker_params_, + VoiceContext& voice_context_, MixContext& mix_context_, + SplitterContext& splitter_context_, + EffectContext& effect_context_, Core::Memory::Memory& memory_) + : worker_params(worker_params_), voice_context(voice_context_), mix_context(mix_context_), + splitter_context(splitter_context_), effect_context(effect_context_), memory(memory_), mix_buffer((worker_params.mix_buffer_count + AudioCommon::MAX_CHANNEL_COUNT) * worker_params.sample_count), sample_buffer(MIX_BUFFER_SIZE), @@ -152,7 +152,7 @@ void CommandGenerator::GenerateVoiceCommand(ServerVoiceInfo& voice_info) { if (!destination_data->IsConfigured()) { continue; } - if (destination_data->GetMixId() >= mix_context.GetCount()) { + if (destination_data->GetMixId() >= static_cast<int>(mix_context.GetCount())) { continue; } @@ -255,7 +255,8 @@ void CommandGenerator::GenerateDataSourceCommand(ServerVoiceInfo& voice_info, Vo void CommandGenerator::GenerateBiquadFilterCommandForVoice(ServerVoiceInfo& voice_info, VoiceState& dsp_state, - s32 mix_buffer_count, s32 channel) { + [[maybe_unused]] s32 mix_buffer_count, + [[maybe_unused]] s32 channel) { for (std::size_t i = 0; i < AudioCommon::MAX_BIQUAD_FILTERS; i++) { const auto& in_params = voice_info.GetInParams(); auto& biquad_filter = in_params.biquad_filter[i]; @@ -278,9 +279,12 @@ void CommandGenerator::GenerateBiquadFilterCommandForVoice(ServerVoiceInfo& voic } } -void AudioCore::CommandGenerator::GenerateBiquadFilterCommand( - s32 mix_buffer, const BiquadFilterParameter& params, std::array<s64, 2>& state, - std::size_t input_offset, std::size_t output_offset, s32 sample_count, s32 node_id) { +void CommandGenerator::GenerateBiquadFilterCommand([[maybe_unused]] s32 mix_buffer_id, + const BiquadFilterParameter& params, + std::array<s64, 2>& state, + std::size_t input_offset, + std::size_t output_offset, s32 sample_count, + s32 node_id) { if (dumping_frame) { LOG_DEBUG(Audio, "(DSP_TRACE) GenerateBiquadFilterCommand node_id={}, " @@ -435,7 +439,7 @@ void CommandGenerator::GenerateAuxCommand(s32 mix_buffer_offset, EffectBase* inf GetMixBuffer(output_index), worker_params.sample_count, offset, write_count); memory.WriteBlock(aux->GetRecvInfo(), &recv_info, sizeof(AuxInfoDSP)); - if (samples_read != worker_params.sample_count && + if (samples_read != static_cast<int>(worker_params.sample_count) && samples_read <= params.sample_count) { std::memset(GetMixBuffer(output_index), 0, params.sample_count - samples_read); } @@ -611,7 +615,8 @@ void CommandGenerator::GenerateMixCommands(ServerMixInfo& mix_info) { const auto& dest_mix = mix_context.GetInfo(destination_data->GetMixId()); const auto& dest_in_params = dest_mix.GetInParams(); const auto mix_index = (base - 1) % in_params.buffer_count + in_params.buffer_offset; - for (std::size_t i = 0; i < dest_in_params.buffer_count; i++) { + for (std::size_t i = 0; i < static_cast<std::size_t>(dest_in_params.buffer_count); + i++) { const auto mixed_volume = in_params.volume * destination_data->GetMixVolume(i); if (mixed_volume != 0.0f) { GenerateMixCommand(dest_in_params.buffer_offset + i, mix_index, mixed_volume, @@ -704,7 +709,7 @@ s32 CommandGenerator::DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_s std::vector<s16> buffer(samples_processed * channel_count); memory.ReadBlock(buffer_pos, buffer.data(), buffer.size() * sizeof(s16)); - for (std::size_t i = 0; i < samples_processed; i++) { + for (std::size_t i = 0; i < static_cast<std::size_t>(samples_processed); i++) { sample_buffer[mix_offset + i] = buffer[i * channel_count + channel]; } } @@ -713,7 +718,8 @@ s32 CommandGenerator::DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_s } s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_state, - s32 sample_count, s32 channel, std::size_t mix_offset) { + s32 sample_count, [[maybe_unused]] s32 channel, + std::size_t mix_offset) { const auto& in_params = voice_info.GetInParams(); const auto& wave_buffer = in_params.wave_buffer[dsp_state.wave_buffer_index]; if (wave_buffer.buffer_address == 0) { @@ -726,8 +732,9 @@ s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_s return 0; } - constexpr std::array<int, 16> SIGNED_NIBBLES = { - {0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1}}; + static constexpr std::array<int, 16> SIGNED_NIBBLES{ + 0, 1, 2, 3, 4, 5, 6, 7, -8, -7, -6, -5, -4, -3, -2, -1, + }; constexpr std::size_t FRAME_LEN = 8; constexpr std::size_t NIBBLES_PER_SAMPLE = 16; @@ -789,9 +796,8 @@ s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_s position_in_frame += 2; // Decode entire frame - if (remaining_samples >= SAMPLES_PER_FRAME) { + if (remaining_samples >= static_cast<int>(SAMPLES_PER_FRAME)) { for (std::size_t i = 0; i < SAMPLES_PER_FRAME / 2; i++) { - // Sample 1 const s32 s0 = SIGNED_NIBBLES[buffer[buffer_offset] >> 4]; const s32 s1 = SIGNED_NIBBLES[buffer[buffer_offset++] & 0xf]; @@ -800,7 +806,7 @@ s32 CommandGenerator::DecodeAdpcm(ServerVoiceInfo& voice_info, VoiceState& dsp_s sample_buffer[cur_mix_offset++] = sample_1; sample_buffer[cur_mix_offset++] = sample_2; } - remaining_samples -= SAMPLES_PER_FRAME; + remaining_samples -= static_cast<int>(SAMPLES_PER_FRAME); position_in_frame += SAMPLES_PER_FRAME; continue; } @@ -866,7 +872,6 @@ void CommandGenerator::DecodeFromWaveBuffers(ServerVoiceInfo& voice_info, s32* o const auto resample_rate = static_cast<s32>( static_cast<float>(in_params.sample_rate) / static_cast<float>(target_sample_rate) * static_cast<float>(static_cast<s32>(in_params.pitch * 32768.0f))); - auto* output_base = output; if (dsp_state.fraction + sample_count * resample_rate > static_cast<s32>(SCALED_MIX_BUFFER_SIZE - 4ULL)) { return; diff --git a/src/audio_core/command_generator.h b/src/audio_core/command_generator.h index 967d24078..b937350b1 100644 --- a/src/audio_core/command_generator.h +++ b/src/audio_core/command_generator.h @@ -7,7 +7,6 @@ #include <array> #include "audio_core/common.h" #include "audio_core/voice_context.h" -#include "common/common_funcs.h" #include "common/common_types.h" namespace Core::Memory { @@ -26,10 +25,10 @@ using MixVolumeBuffer = std::array<float, AudioCommon::MAX_MIX_BUFFERS>; class CommandGenerator { public: - explicit CommandGenerator(AudioCommon::AudioRendererParameter& worker_params, - VoiceContext& voice_context, MixContext& mix_context, - SplitterContext& splitter_context, EffectContext& effect_context, - Core::Memory::Memory& memory); + explicit CommandGenerator(AudioCommon::AudioRendererParameter& worker_params_, + VoiceContext& voice_context_, MixContext& mix_context_, + SplitterContext& splitter_context_, EffectContext& effect_context_, + Core::Memory::Memory& memory_); ~CommandGenerator(); void ClearMixBuffers(); @@ -40,13 +39,13 @@ public: void PreCommand(); void PostCommand(); - s32* GetChannelMixBuffer(s32 channel); - const s32* GetChannelMixBuffer(s32 channel) const; - s32* GetMixBuffer(std::size_t index); - const s32* GetMixBuffer(std::size_t index) const; - std::size_t GetMixChannelBufferOffset(s32 channel) const; + [[nodiscard]] s32* GetChannelMixBuffer(s32 channel); + [[nodiscard]] const s32* GetChannelMixBuffer(s32 channel) const; + [[nodiscard]] s32* GetMixBuffer(std::size_t index); + [[nodiscard]] const s32* GetMixBuffer(std::size_t index) const; + [[nodiscard]] std::size_t GetMixChannelBufferOffset(s32 channel) const; - std::size_t GetTotalMixBufferCount() const; + [[nodiscard]] std::size_t GetTotalMixBufferCount() const; private: void GenerateDataSourceCommand(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 channel); @@ -74,7 +73,7 @@ private: void GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled); void GenerateBiquadFilterEffectCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled); void GenerateAuxCommand(s32 mix_buffer_offset, EffectBase* info, bool enabled); - ServerSplitterDestinationData* GetDestinationData(s32 splitter_id, s32 index); + [[nodiscard]] ServerSplitterDestinationData* GetDestinationData(s32 splitter_id, s32 index); s32 WriteAuxBuffer(AuxInfoDSP& dsp_info, VAddr send_buffer, u32 max_samples, const s32* data, u32 sample_count, u32 write_offset, u32 write_count); diff --git a/src/audio_core/common.h b/src/audio_core/common.h index 72ebce221..ec59a3ba9 100644 --- a/src/audio_core/common.h +++ b/src/audio_core/common.h @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #pragma once + #include "common/common_funcs.h" #include "common/common_types.h" #include "common/swap.h" @@ -21,7 +22,7 @@ constexpr std::size_t MAX_CHANNEL_COUNT = 6; constexpr std::size_t MAX_WAVE_BUFFERS = 4; constexpr std::size_t MAX_SAMPLE_HISTORY = 4; constexpr u32 STREAM_SAMPLE_RATE = 48000; -constexpr u32 STREAM_NUM_CHANNELS = 6; +constexpr u32 STREAM_NUM_CHANNELS = 2; constexpr s32 NO_SPLITTER = -1; constexpr s32 NO_MIX = 0x7fffffff; constexpr s32 NO_FINAL_MIX = std::numeric_limits<s32>::min(); diff --git a/src/audio_core/cubeb_sink.cpp b/src/audio_core/cubeb_sink.cpp index 83c06c0ed..043447eaa 100644 --- a/src/audio_core/cubeb_sink.cpp +++ b/src/audio_core/cubeb_sink.cpp @@ -21,15 +21,16 @@ namespace AudioCore { class CubebSinkStream final : public SinkStream { public: - CubebSinkStream(cubeb* ctx, u32 sample_rate, u32 num_channels_, cubeb_devid output_device, + CubebSinkStream(cubeb* ctx_, u32 sample_rate, u32 num_channels_, cubeb_devid output_device, const std::string& name) - : ctx{ctx}, num_channels{std::min(num_channels_, 6u)}, time_stretch{sample_rate, - num_channels} { + : ctx{ctx_}, num_channels{std::min(num_channels_, 6u)}, time_stretch{sample_rate, + num_channels} { cubeb_stream_params params{}; params.rate = sample_rate; params.channels = num_channels; params.format = CUBEB_SAMPLE_S16NE; + params.prefs = CUBEB_STREAM_PREF_PERSIST; switch (num_channels) { case 1: params.layout = CUBEB_LAYOUT_MONO; @@ -93,8 +94,10 @@ public: constexpr s32 clev{707}; // center mixing level coefficient constexpr s32 slev{707}; // surround mixing level coefficient - buf.push_back(left + (clev * center / 1000) + (slev * surround_left / 1000)); - buf.push_back(right + (clev * center / 1000) + (slev * surround_right / 1000)); + buf.push_back(static_cast<s16>(left + (clev * center / 1000) + + (slev * surround_left / 1000))); + buf.push_back(static_cast<s16>(right + (clev * center / 1000) + + (slev * surround_right / 1000))); } queue.Push(buf); return; @@ -190,10 +193,11 @@ SinkStream& CubebSink::AcquireSinkStream(u32 sample_rate, u32 num_channels, return *sink_streams.back(); } -long CubebSinkStream::DataCallback(cubeb_stream* stream, void* user_data, const void* input_buffer, - void* output_buffer, long num_frames) { - CubebSinkStream* impl = static_cast<CubebSinkStream*>(user_data); - u8* buffer = reinterpret_cast<u8*>(output_buffer); +long CubebSinkStream::DataCallback([[maybe_unused]] cubeb_stream* stream, void* user_data, + [[maybe_unused]] const void* input_buffer, void* output_buffer, + long num_frames) { + auto* impl = static_cast<CubebSinkStream*>(user_data); + auto* buffer = static_cast<u8*>(output_buffer); if (!impl) { return {}; @@ -234,7 +238,9 @@ long CubebSinkStream::DataCallback(cubeb_stream* stream, void* user_data, const return num_frames; } -void CubebSinkStream::StateCallback(cubeb_stream* stream, void* user_data, cubeb_state state) {} +void CubebSinkStream::StateCallback([[maybe_unused]] cubeb_stream* stream, + [[maybe_unused]] void* user_data, + [[maybe_unused]] cubeb_state state) {} std::vector<std::string> ListCubebSinkDevices() { std::vector<std::string> device_list; diff --git a/src/audio_core/effect_context.cpp b/src/audio_core/effect_context.cpp index adfec3df5..f770b9608 100644 --- a/src/audio_core/effect_context.cpp +++ b/src/audio_core/effect_context.cpp @@ -12,7 +12,7 @@ bool ValidChannelCountForEffect(s32 channel_count) { } } // namespace -EffectContext::EffectContext(std::size_t effect_count) : effect_count(effect_count) { +EffectContext::EffectContext(std::size_t effect_count_) : effect_count(effect_count_) { effects.reserve(effect_count); std::generate_n(std::back_inserter(effects), effect_count, [] { return std::make_unique<EffectStubbed>(); }); @@ -61,13 +61,13 @@ const EffectBase* EffectContext::GetInfo(std::size_t i) const { return effects.at(i).get(); } -EffectStubbed::EffectStubbed() : EffectBase::EffectBase(EffectType::Invalid) {} +EffectStubbed::EffectStubbed() : EffectBase(EffectType::Invalid) {} EffectStubbed::~EffectStubbed() = default; -void EffectStubbed::Update(EffectInfo::InParams& in_params) {} +void EffectStubbed::Update([[maybe_unused]] EffectInfo::InParams& in_params) {} void EffectStubbed::UpdateForCommandGeneration() {} -EffectBase::EffectBase(EffectType effect_type) : effect_type(effect_type) {} +EffectBase::EffectBase(EffectType effect_type_) : effect_type(effect_type_) {} EffectBase::~EffectBase() = default; UsageState EffectBase::GetUsage() const { @@ -90,32 +90,32 @@ s32 EffectBase::GetProcessingOrder() const { return processing_order; } -EffectI3dl2Reverb::EffectI3dl2Reverb() : EffectGeneric::EffectGeneric(EffectType::I3dl2Reverb) {} +EffectI3dl2Reverb::EffectI3dl2Reverb() : EffectGeneric(EffectType::I3dl2Reverb) {} EffectI3dl2Reverb::~EffectI3dl2Reverb() = default; void EffectI3dl2Reverb::Update(EffectInfo::InParams& in_params) { - auto& internal_params = GetParams(); + auto& params = GetParams(); const auto* reverb_params = reinterpret_cast<I3dl2ReverbParams*>(in_params.raw.data()); if (!ValidChannelCountForEffect(reverb_params->max_channels)) { UNREACHABLE_MSG("Invalid reverb max channel count {}", reverb_params->max_channels); return; } - const auto last_status = internal_params.status; + const auto last_status = params.status; mix_id = in_params.mix_id; processing_order = in_params.processing_order; - internal_params = *reverb_params; + params = *reverb_params; if (!ValidChannelCountForEffect(reverb_params->channel_count)) { - internal_params.channel_count = internal_params.max_channels; + params.channel_count = params.max_channels; } enabled = in_params.is_enabled; if (last_status != ParameterStatus::Updated) { - internal_params.status = last_status; + params.status = last_status; } if (in_params.is_new || skipped) { usage = UsageState::Initialized; - internal_params.status = ParameterStatus::Initialized; + params.status = ParameterStatus::Initialized; skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0; } } @@ -129,15 +129,15 @@ void EffectI3dl2Reverb::UpdateForCommandGeneration() { GetParams().status = ParameterStatus::Updated; } -EffectBiquadFilter::EffectBiquadFilter() : EffectGeneric::EffectGeneric(EffectType::BiquadFilter) {} +EffectBiquadFilter::EffectBiquadFilter() : EffectGeneric(EffectType::BiquadFilter) {} EffectBiquadFilter::~EffectBiquadFilter() = default; void EffectBiquadFilter::Update(EffectInfo::InParams& in_params) { - auto& internal_params = GetParams(); + auto& params = GetParams(); const auto* biquad_params = reinterpret_cast<BiquadFilterParams*>(in_params.raw.data()); mix_id = in_params.mix_id; processing_order = in_params.processing_order; - internal_params = *biquad_params; + params = *biquad_params; enabled = in_params.is_enabled; } @@ -150,7 +150,7 @@ void EffectBiquadFilter::UpdateForCommandGeneration() { GetParams().status = ParameterStatus::Updated; } -EffectAuxInfo::EffectAuxInfo() : EffectGeneric::EffectGeneric(EffectType::Aux) {} +EffectAuxInfo::EffectAuxInfo() : EffectGeneric(EffectType::Aux) {} EffectAuxInfo::~EffectAuxInfo() = default; void EffectAuxInfo::Update(EffectInfo::InParams& in_params) { @@ -184,48 +184,48 @@ void EffectAuxInfo::UpdateForCommandGeneration() { } } -const VAddr EffectAuxInfo::GetSendInfo() const { +VAddr EffectAuxInfo::GetSendInfo() const { return send_info; } -const VAddr EffectAuxInfo::GetSendBuffer() const { +VAddr EffectAuxInfo::GetSendBuffer() const { return send_buffer; } -const VAddr EffectAuxInfo::GetRecvInfo() const { +VAddr EffectAuxInfo::GetRecvInfo() const { return recv_info; } -const VAddr EffectAuxInfo::GetRecvBuffer() const { +VAddr EffectAuxInfo::GetRecvBuffer() const { return recv_buffer; } -EffectDelay::EffectDelay() : EffectGeneric::EffectGeneric(EffectType::Delay) {} +EffectDelay::EffectDelay() : EffectGeneric(EffectType::Delay) {} EffectDelay::~EffectDelay() = default; void EffectDelay::Update(EffectInfo::InParams& in_params) { const auto* delay_params = reinterpret_cast<DelayParams*>(in_params.raw.data()); - auto& internal_params = GetParams(); + auto& params = GetParams(); if (!ValidChannelCountForEffect(delay_params->max_channels)) { return; } - const auto last_status = internal_params.status; + const auto last_status = params.status; mix_id = in_params.mix_id; processing_order = in_params.processing_order; - internal_params = *delay_params; + params = *delay_params; if (!ValidChannelCountForEffect(delay_params->channels)) { - internal_params.channels = internal_params.max_channels; + params.channels = params.max_channels; } enabled = in_params.is_enabled; if (last_status != ParameterStatus::Updated) { - internal_params.status = last_status; + params.status = last_status; } if (in_params.is_new || skipped) { usage = UsageState::Initialized; - internal_params.status = ParameterStatus::Initialized; + params.status = ParameterStatus::Initialized; skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0; } } @@ -239,7 +239,7 @@ void EffectDelay::UpdateForCommandGeneration() { GetParams().status = ParameterStatus::Updated; } -EffectBufferMixer::EffectBufferMixer() : EffectGeneric::EffectGeneric(EffectType::BufferMixer) {} +EffectBufferMixer::EffectBufferMixer() : EffectGeneric(EffectType::BufferMixer) {} EffectBufferMixer::~EffectBufferMixer() = default; void EffectBufferMixer::Update(EffectInfo::InParams& in_params) { @@ -257,32 +257,32 @@ void EffectBufferMixer::UpdateForCommandGeneration() { } } -EffectReverb::EffectReverb() : EffectGeneric::EffectGeneric(EffectType::Reverb) {} +EffectReverb::EffectReverb() : EffectGeneric(EffectType::Reverb) {} EffectReverb::~EffectReverb() = default; void EffectReverb::Update(EffectInfo::InParams& in_params) { const auto* reverb_params = reinterpret_cast<ReverbParams*>(in_params.raw.data()); - auto& internal_params = GetParams(); + auto& params = GetParams(); if (!ValidChannelCountForEffect(reverb_params->max_channels)) { return; } - const auto last_status = internal_params.status; + const auto last_status = params.status; mix_id = in_params.mix_id; processing_order = in_params.processing_order; - internal_params = *reverb_params; + params = *reverb_params; if (!ValidChannelCountForEffect(reverb_params->channels)) { - internal_params.channels = internal_params.max_channels; + params.channels = params.max_channels; } enabled = in_params.is_enabled; if (last_status != ParameterStatus::Updated) { - internal_params.status = last_status; + params.status = last_status; } if (in_params.is_new || skipped) { usage = UsageState::Initialized; - internal_params.status = ParameterStatus::Initialized; + params.status = ParameterStatus::Initialized; skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0; } } diff --git a/src/audio_core/effect_context.h b/src/audio_core/effect_context.h index 2f2da72dd..c5e0b398c 100644 --- a/src/audio_core/effect_context.h +++ b/src/audio_core/effect_context.h @@ -166,13 +166,13 @@ public: std::array<u8, 0xa0> raw; }; }; - static_assert(sizeof(EffectInfo::InParams) == 0xc0, "InParams is an invalid size"); + static_assert(sizeof(InParams) == 0xc0, "InParams is an invalid size"); struct OutParams { UsageStatus status{}; INSERT_PADDING_BYTES(15); }; - static_assert(sizeof(EffectInfo::OutParams) == 0x10, "OutParams is an invalid size"); + static_assert(sizeof(OutParams) == 0x10, "OutParams is an invalid size"); }; struct AuxAddress { @@ -184,16 +184,16 @@ struct AuxAddress { class EffectBase { public: - EffectBase(EffectType effect_type); - ~EffectBase(); + explicit EffectBase(EffectType effect_type_); + virtual ~EffectBase(); virtual void Update(EffectInfo::InParams& in_params) = 0; virtual void UpdateForCommandGeneration() = 0; - UsageState GetUsage() const; - EffectType GetType() const; - bool IsEnabled() const; - s32 GetMixID() const; - s32 GetProcessingOrder() const; + [[nodiscard]] UsageState GetUsage() const; + [[nodiscard]] EffectType GetType() const; + [[nodiscard]] bool IsEnabled() const; + [[nodiscard]] s32 GetMixID() const; + [[nodiscard]] s32 GetProcessingOrder() const; protected: UsageState usage{UsageState::Invalid}; @@ -206,8 +206,7 @@ protected: template <typename T> class EffectGeneric : public EffectBase { public: - EffectGeneric(EffectType effect_type) : EffectBase::EffectBase(effect_type) {} - ~EffectGeneric() = default; + explicit EffectGeneric(EffectType effect_type_) : EffectBase(effect_type_) {} T& GetParams() { return internal_params; @@ -224,7 +223,7 @@ private: class EffectStubbed : public EffectBase { public: explicit EffectStubbed(); - ~EffectStubbed(); + ~EffectStubbed() override; void Update(EffectInfo::InParams& in_params) override; void UpdateForCommandGeneration() override; @@ -233,7 +232,7 @@ public: class EffectI3dl2Reverb : public EffectGeneric<I3dl2ReverbParams> { public: explicit EffectI3dl2Reverb(); - ~EffectI3dl2Reverb(); + ~EffectI3dl2Reverb() override; void Update(EffectInfo::InParams& in_params) override; void UpdateForCommandGeneration() override; @@ -245,7 +244,7 @@ private: class EffectBiquadFilter : public EffectGeneric<BiquadFilterParams> { public: explicit EffectBiquadFilter(); - ~EffectBiquadFilter(); + ~EffectBiquadFilter() override; void Update(EffectInfo::InParams& in_params) override; void UpdateForCommandGeneration() override; @@ -254,14 +253,14 @@ public: class EffectAuxInfo : public EffectGeneric<AuxInfo> { public: explicit EffectAuxInfo(); - ~EffectAuxInfo(); + ~EffectAuxInfo() override; void Update(EffectInfo::InParams& in_params) override; void UpdateForCommandGeneration() override; - const VAddr GetSendInfo() const; - const VAddr GetSendBuffer() const; - const VAddr GetRecvInfo() const; - const VAddr GetRecvBuffer() const; + [[nodiscard]] VAddr GetSendInfo() const; + [[nodiscard]] VAddr GetSendBuffer() const; + [[nodiscard]] VAddr GetRecvInfo() const; + [[nodiscard]] VAddr GetRecvBuffer() const; private: VAddr send_info{}; @@ -275,7 +274,7 @@ private: class EffectDelay : public EffectGeneric<DelayParams> { public: explicit EffectDelay(); - ~EffectDelay(); + ~EffectDelay() override; void Update(EffectInfo::InParams& in_params) override; void UpdateForCommandGeneration() override; @@ -287,7 +286,7 @@ private: class EffectBufferMixer : public EffectGeneric<BufferMixerParams> { public: explicit EffectBufferMixer(); - ~EffectBufferMixer(); + ~EffectBufferMixer() override; void Update(EffectInfo::InParams& in_params) override; void UpdateForCommandGeneration() override; @@ -296,7 +295,7 @@ public: class EffectReverb : public EffectGeneric<ReverbParams> { public: explicit EffectReverb(); - ~EffectReverb(); + ~EffectReverb() override; void Update(EffectInfo::InParams& in_params) override; void UpdateForCommandGeneration() override; @@ -307,13 +306,13 @@ private: class EffectContext { public: - explicit EffectContext(std::size_t effect_count); + explicit EffectContext(std::size_t effect_count_); ~EffectContext(); - std::size_t GetCount() const; - EffectBase* GetInfo(std::size_t i); - EffectBase* RetargetEffect(std::size_t i, EffectType effect); - const EffectBase* GetInfo(std::size_t i) const; + [[nodiscard]] std::size_t GetCount() const; + [[nodiscard]] EffectBase* GetInfo(std::size_t i); + [[nodiscard]] EffectBase* RetargetEffect(std::size_t i, EffectType effect); + [[nodiscard]] const EffectBase* GetInfo(std::size_t i) const; private: std::size_t effect_count{}; diff --git a/src/audio_core/info_updater.cpp b/src/audio_core/info_updater.cpp index f53ce21a5..d3ac90827 100644 --- a/src/audio_core/info_updater.cpp +++ b/src/audio_core/info_updater.cpp @@ -14,9 +14,9 @@ namespace AudioCore { -InfoUpdater::InfoUpdater(const std::vector<u8>& in_params, std::vector<u8>& out_params, - BehaviorInfo& behavior_info) - : in_params(in_params), out_params(out_params), behavior_info(behavior_info) { +InfoUpdater::InfoUpdater(const std::vector<u8>& in_params_, std::vector<u8>& out_params_, + BehaviorInfo& behavior_info_) + : in_params(in_params_), out_params(out_params_), behavior_info(behavior_info_) { ASSERT( AudioCommon::CanConsumeBuffer(in_params.size(), 0, sizeof(AudioCommon::UpdateDataHeader))); std::memcpy(&input_header, in_params.data(), sizeof(AudioCommon::UpdateDataHeader)); @@ -64,7 +64,6 @@ bool InfoUpdater::UpdateBehaviorInfo(BehaviorInfo& in_behavior_info) { } bool InfoUpdater::UpdateMemoryPools(std::vector<ServerMemoryPoolInfo>& memory_pool_info) { - const auto force_mapping = behavior_info.IsMemoryPoolForceMappingEnabled(); const auto memory_pool_count = memory_pool_info.size(); const auto total_memory_pool_in = sizeof(ServerMemoryPoolInfo::InParams) * memory_pool_count; const auto total_memory_pool_out = sizeof(ServerMemoryPoolInfo::OutParams) * memory_pool_count; @@ -136,8 +135,8 @@ bool InfoUpdater::UpdateVoiceChannelResources(VoiceContext& voice_context) { } bool InfoUpdater::UpdateVoices(VoiceContext& voice_context, - std::vector<ServerMemoryPoolInfo>& memory_pool_info, - VAddr audio_codec_dsp_addr) { + [[maybe_unused]] std::vector<ServerMemoryPoolInfo>& memory_pool_info, + [[maybe_unused]] VAddr audio_codec_dsp_addr) { const auto voice_count = voice_context.GetVoiceCount(); std::vector<VoiceInfo::InParams> voice_in(voice_count); std::vector<VoiceInfo::OutParams> voice_out(voice_count); @@ -166,28 +165,28 @@ bool InfoUpdater::UpdateVoices(VoiceContext& voice_context, // Update our voices for (std::size_t i = 0; i < voice_count; i++) { - auto& in_params = voice_in[i]; - const auto channel_count = static_cast<std::size_t>(in_params.channel_count); + auto& voice_in_params = voice_in[i]; + const auto channel_count = static_cast<std::size_t>(voice_in_params.channel_count); // Skip if it's not currently in use - if (!in_params.is_in_use) { + if (!voice_in_params.is_in_use) { continue; } // Voice states for each channel std::array<VoiceState*, AudioCommon::MAX_CHANNEL_COUNT> voice_states{}; - ASSERT(in_params.id < voice_count); + ASSERT(static_cast<std::size_t>(voice_in_params.id) < voice_count); // Grab our current voice info - auto& voice_info = voice_context.GetInfo(static_cast<std::size_t>(in_params.id)); + auto& voice_info = voice_context.GetInfo(static_cast<std::size_t>(voice_in_params.id)); ASSERT(channel_count <= AudioCommon::MAX_CHANNEL_COUNT); // Get all our channel voice states for (std::size_t channel = 0; channel < channel_count; channel++) { voice_states[channel] = - &voice_context.GetState(in_params.voice_channel_resource_ids[channel]); + &voice_context.GetState(voice_in_params.voice_channel_resource_ids[channel]); } - if (in_params.is_new) { + if (voice_in_params.is_new) { // Default our values for our voice voice_info.Initialize(); if (channel_count == 0 || channel_count > AudioCommon::MAX_CHANNEL_COUNT) { @@ -201,12 +200,12 @@ bool InfoUpdater::UpdateVoices(VoiceContext& voice_context, } // Update our voice - voice_info.UpdateParameters(in_params, behavior_info); + voice_info.UpdateParameters(voice_in_params, behavior_info); // TODO(ogniK): Handle mapping errors with behavior info based on in params response // Update our wave buffers - voice_info.UpdateWaveBuffers(in_params, voice_states, behavior_info); - voice_info.WriteOutStatus(voice_out[i], in_params, voice_states); + voice_info.UpdateWaveBuffers(voice_in_params, voice_states, behavior_info); + voice_info.WriteOutStatus(voice_out[i], voice_in_params, voice_states); } if (!AudioCommon::CanConsumeBuffer(out_params.size(), output_offset, voice_out_size)) { @@ -352,8 +351,8 @@ ResultCode InfoUpdater::UpdateMixes(MixContext& mix_context, std::size_t mix_buf for (std::size_t i = 0; i < mix_count; i++) { const auto& in = mix_in_params[i]; total_buffer_count += in.buffer_count; - if (in.dest_mix_id > mix_count && in.dest_mix_id != AudioCommon::NO_MIX && - in.mix_id != AudioCommon::FINAL_MIX) { + if (static_cast<std::size_t>(in.dest_mix_id) > mix_count && + in.dest_mix_id != AudioCommon::NO_MIX && in.mix_id != AudioCommon::FINAL_MIX) { LOG_ERROR( Audio, "Invalid mix destination, mix_id={:X}, dest_mix_id={:X}, mix_buffer_count={:X}", @@ -446,7 +445,7 @@ bool InfoUpdater::UpdatePerformanceBuffer() { return true; } -bool InfoUpdater::UpdateErrorInfo(BehaviorInfo& in_behavior_info) { +bool InfoUpdater::UpdateErrorInfo([[maybe_unused]] BehaviorInfo& in_behavior_info) { const auto total_beahvior_info_out = sizeof(BehaviorInfo::OutParams); if (!AudioCommon::CanConsumeBuffer(out_params.size(), output_offset, total_beahvior_info_out)) { diff --git a/src/audio_core/info_updater.h b/src/audio_core/info_updater.h index 06f9d770f..d315c91ed 100644 --- a/src/audio_core/info_updater.h +++ b/src/audio_core/info_updater.h @@ -21,8 +21,8 @@ class SplitterContext; class InfoUpdater { public: // TODO(ogniK): Pass process handle when we support it - InfoUpdater(const std::vector<u8>& in_params, std::vector<u8>& out_params, - BehaviorInfo& behavior_info); + InfoUpdater(const std::vector<u8>& in_params_, std::vector<u8>& out_params_, + BehaviorInfo& behavior_info_); ~InfoUpdater(); bool UpdateBehaviorInfo(BehaviorInfo& in_behavior_info); diff --git a/src/audio_core/memory_pool.cpp b/src/audio_core/memory_pool.cpp index 5a3453063..6b6908d26 100644 --- a/src/audio_core/memory_pool.cpp +++ b/src/audio_core/memory_pool.cpp @@ -10,11 +10,10 @@ namespace AudioCore { ServerMemoryPoolInfo::ServerMemoryPoolInfo() = default; ServerMemoryPoolInfo::~ServerMemoryPoolInfo() = default; -bool ServerMemoryPoolInfo::Update(const ServerMemoryPoolInfo::InParams& in_params, - ServerMemoryPoolInfo::OutParams& out_params) { + +bool ServerMemoryPoolInfo::Update(const InParams& in_params, OutParams& out_params) { // Our state does not need to be changed - if (in_params.state != ServerMemoryPoolInfo::State::RequestAttach && - in_params.state != ServerMemoryPoolInfo::State::RequestDetach) { + if (in_params.state != State::RequestAttach && in_params.state != State::RequestDetach) { return true; } @@ -32,11 +31,11 @@ bool ServerMemoryPoolInfo::Update(const ServerMemoryPoolInfo::InParams& in_param return false; } - if (in_params.state == ServerMemoryPoolInfo::State::RequestAttach) { + if (in_params.state == State::RequestAttach) { cpu_address = in_params.address; size = in_params.size; used = true; - out_params.state = ServerMemoryPoolInfo::State::Attached; + out_params.state = State::Attached; } else { // Unexpected address if (cpu_address != in_params.address) { @@ -54,7 +53,7 @@ bool ServerMemoryPoolInfo::Update(const ServerMemoryPoolInfo::InParams& in_param cpu_address = 0; size = 0; used = false; - out_params.state = ServerMemoryPoolInfo::State::Detached; + out_params.state = State::Detached; } return true; } diff --git a/src/audio_core/memory_pool.h b/src/audio_core/memory_pool.h index 8ac503f1c..3e9e777ae 100644 --- a/src/audio_core/memory_pool.h +++ b/src/audio_core/memory_pool.h @@ -28,19 +28,18 @@ public: struct InParams { u64_le address{}; u64_le size{}; - ServerMemoryPoolInfo::State state{}; + State state{}; INSERT_PADDING_WORDS(3); }; - static_assert(sizeof(ServerMemoryPoolInfo::InParams) == 0x20, "InParams are an invalid size"); + static_assert(sizeof(InParams) == 0x20, "InParams are an invalid size"); struct OutParams { - ServerMemoryPoolInfo::State state{}; + State state{}; INSERT_PADDING_WORDS(3); }; - static_assert(sizeof(ServerMemoryPoolInfo::OutParams) == 0x10, "OutParams are an invalid size"); + static_assert(sizeof(OutParams) == 0x10, "OutParams are an invalid size"); - bool Update(const ServerMemoryPoolInfo::InParams& in_params, - ServerMemoryPoolInfo::OutParams& out_params); + bool Update(const InParams& in_params, OutParams& out_params); private: // There's another entry here which is the DSP address, however since we're not talking to the diff --git a/src/audio_core/mix_context.cpp b/src/audio_core/mix_context.cpp index 042891490..4bca72eb0 100644 --- a/src/audio_core/mix_context.cpp +++ b/src/audio_core/mix_context.cpp @@ -53,7 +53,7 @@ void MixContext::UpdateDistancesFromFinalMix() { auto mix_id = in_params.mix_id; // Needs to be referenced out of scope s32 distance_to_final_mix{AudioCommon::FINAL_MIX}; - for (; distance_to_final_mix < info_count; distance_to_final_mix++) { + for (; distance_to_final_mix < static_cast<s32>(info_count); distance_to_final_mix++) { if (mix_id == AudioCommon::FINAL_MIX) { // If we're at the final mix, we're done break; @@ -77,7 +77,7 @@ void MixContext::UpdateDistancesFromFinalMix() { } // If we're out of range for our distance, mark it as no final mix - if (distance_to_final_mix >= info_count) { + if (distance_to_final_mix >= static_cast<s32>(info_count)) { distance_to_final_mix = AudioCommon::NO_FINAL_MIX; } diff --git a/src/audio_core/mix_context.h b/src/audio_core/mix_context.h index 6a588eeb4..68bc673c6 100644 --- a/src/audio_core/mix_context.h +++ b/src/audio_core/mix_context.h @@ -62,17 +62,17 @@ public: ServerMixInfo(); ~ServerMixInfo(); - const ServerMixInfo::InParams& GetInParams() const; - ServerMixInfo::InParams& GetInParams(); + [[nodiscard]] const ServerMixInfo::InParams& GetInParams() const; + [[nodiscard]] ServerMixInfo::InParams& GetInParams(); bool Update(EdgeMatrix& edge_matrix, const MixInfo::InParams& mix_in, BehaviorInfo& behavior_info, SplitterContext& splitter_context, EffectContext& effect_context); - bool HasAnyConnection() const; + [[nodiscard]] bool HasAnyConnection() const; void Cleanup(); void SetEffectCount(std::size_t count); void ResetEffectProcessingOrder(); - s32 GetEffectOrder(std::size_t i) const; + [[nodiscard]] s32 GetEffectOrder(std::size_t i) const; private: std::vector<s32> effect_processing_order; @@ -91,15 +91,15 @@ public: void SortInfo(); bool TsortInfo(SplitterContext& splitter_context); - std::size_t GetCount() const; - ServerMixInfo& GetInfo(std::size_t i); - const ServerMixInfo& GetInfo(std::size_t i) const; - ServerMixInfo& GetSortedInfo(std::size_t i); - const ServerMixInfo& GetSortedInfo(std::size_t i) const; - ServerMixInfo& GetFinalMixInfo(); - const ServerMixInfo& GetFinalMixInfo() const; - EdgeMatrix& GetEdgeMatrix(); - const EdgeMatrix& GetEdgeMatrix() const; + [[nodiscard]] std::size_t GetCount() const; + [[nodiscard]] ServerMixInfo& GetInfo(std::size_t i); + [[nodiscard]] const ServerMixInfo& GetInfo(std::size_t i) const; + [[nodiscard]] ServerMixInfo& GetSortedInfo(std::size_t i); + [[nodiscard]] const ServerMixInfo& GetSortedInfo(std::size_t i) const; + [[nodiscard]] ServerMixInfo& GetFinalMixInfo(); + [[nodiscard]] const ServerMixInfo& GetFinalMixInfo() const; + [[nodiscard]] EdgeMatrix& GetEdgeMatrix(); + [[nodiscard]] const EdgeMatrix& GetEdgeMatrix() const; private: void CalcMixBufferOffset(); diff --git a/src/audio_core/sink_context.cpp b/src/audio_core/sink_context.cpp index 0882b411a..a69543696 100644 --- a/src/audio_core/sink_context.cpp +++ b/src/audio_core/sink_context.cpp @@ -5,17 +5,23 @@ #include "audio_core/sink_context.h" namespace AudioCore { -SinkContext::SinkContext(std::size_t sink_count) : sink_count(sink_count) {} +SinkContext::SinkContext(std::size_t sink_count_) : sink_count{sink_count_} {} SinkContext::~SinkContext() = default; std::size_t SinkContext::GetCount() const { return sink_count; } -void SinkContext::UpdateMainSink(SinkInfo::InParams& in) { +void SinkContext::UpdateMainSink(const SinkInfo::InParams& in) { + ASSERT(in.type == SinkTypes::Device); + + has_downmix_coefs = in.device.down_matrix_enabled; + if (has_downmix_coefs) { + downmix_coefficients = in.device.down_matrix_coef; + } in_use = in.in_use; use_count = in.device.input_count; - std::memcpy(buffers.data(), in.device.input.data(), AudioCommon::MAX_CHANNEL_COUNT); + buffers = in.device.input; } bool SinkContext::InUse() const { @@ -28,4 +34,12 @@ std::vector<u8> SinkContext::OutputBuffers() const { return buffer_ret; } +bool SinkContext::HasDownMixingCoefficients() const { + return has_downmix_coefs; +} + +const DownmixCoefficients& SinkContext::GetDownmixCoefficients() const { + return downmix_coefficients; +} + } // namespace AudioCore diff --git a/src/audio_core/sink_context.h b/src/audio_core/sink_context.h index d7aa72ba7..05541becb 100644 --- a/src/audio_core/sink_context.h +++ b/src/audio_core/sink_context.h @@ -11,6 +11,8 @@ namespace AudioCore { +using DownmixCoefficients = std::array<float_le, 4>; + enum class SinkTypes : u8 { Invalid = 0, Device = 1, @@ -40,7 +42,7 @@ public: bool in_use; INSERT_UNION_PADDING_BYTES(5); }; - static_assert(sizeof(SinkInfo::CircularBufferIn) == 0x28, + static_assert(sizeof(CircularBufferIn) == 0x28, "SinkInfo::CircularBufferIn is in invalid size"); struct DeviceIn { @@ -50,9 +52,9 @@ public: std::array<u8, AudioCommon::MAX_CHANNEL_COUNT> input; INSERT_UNION_PADDING_BYTES(1); bool down_matrix_enabled; - std::array<float_le, 4> down_matrix_coef; + DownmixCoefficients down_matrix_coef; }; - static_assert(sizeof(SinkInfo::DeviceIn) == 0x11c, "SinkInfo::DeviceIn is an invalid size"); + static_assert(sizeof(DeviceIn) == 0x11c, "SinkInfo::DeviceIn is an invalid size"); struct InParams { SinkTypes type{}; @@ -62,28 +64,33 @@ public: INSERT_PADDING_WORDS(6); union { // std::array<u8, 0x120> raw{}; - SinkInfo::DeviceIn device; - SinkInfo::CircularBufferIn circular_buffer; + DeviceIn device; + CircularBufferIn circular_buffer; }; }; - static_assert(sizeof(SinkInfo::InParams) == 0x140, "SinkInfo::InParams are an invalid size!"); + static_assert(sizeof(InParams) == 0x140, "SinkInfo::InParams are an invalid size!"); }; class SinkContext { public: - explicit SinkContext(std::size_t sink_count); + explicit SinkContext(std::size_t sink_count_); ~SinkContext(); - std::size_t GetCount() const; + [[nodiscard]] std::size_t GetCount() const; + + void UpdateMainSink(const SinkInfo::InParams& in); + [[nodiscard]] bool InUse() const; + [[nodiscard]] std::vector<u8> OutputBuffers() const; - void UpdateMainSink(SinkInfo::InParams& in); - bool InUse() const; - std::vector<u8> OutputBuffers() const; + [[nodiscard]] bool HasDownMixingCoefficients() const; + [[nodiscard]] const DownmixCoefficients& GetDownmixCoefficients() const; private: bool in_use{false}; s32 use_count{}; std::array<u8, AudioCommon::MAX_CHANNEL_COUNT> buffers{}; std::size_t sink_count{}; + bool has_downmix_coefs{false}; + DownmixCoefficients downmix_coefficients{}; }; } // namespace AudioCore diff --git a/src/audio_core/splitter_context.cpp b/src/audio_core/splitter_context.cpp index 79bb2f516..f4bcd0391 100644 --- a/src/audio_core/splitter_context.cpp +++ b/src/audio_core/splitter_context.cpp @@ -10,7 +10,7 @@ namespace AudioCore { -ServerSplitterDestinationData::ServerSplitterDestinationData(s32 id) : id(id) {} +ServerSplitterDestinationData::ServerSplitterDestinationData(s32 id_) : id{id_} {} ServerSplitterDestinationData::~ServerSplitterDestinationData() = default; void ServerSplitterDestinationData::Update(SplitterInfo::InDestinationParams& header) { @@ -87,7 +87,7 @@ void ServerSplitterDestinationData::UpdateInternalState() { needs_update = false; } -ServerSplitterInfo::ServerSplitterInfo(s32 id) : id(id) {} +ServerSplitterInfo::ServerSplitterInfo(s32 id_) : id(id_) {} ServerSplitterInfo::~ServerSplitterInfo() = default; void ServerSplitterInfo::InitializeInfos() { @@ -121,7 +121,7 @@ const ServerSplitterDestinationData* ServerSplitterInfo::GetHead() const { } ServerSplitterDestinationData* ServerSplitterInfo::GetData(std::size_t depth) { - auto current_head = head; + auto* current_head = head; for (std::size_t i = 0; i < depth; i++) { if (current_head == nullptr) { return nullptr; @@ -132,7 +132,7 @@ ServerSplitterDestinationData* ServerSplitterInfo::GetData(std::size_t depth) { } const ServerSplitterDestinationData* ServerSplitterInfo::GetData(std::size_t depth) const { - auto current_head = head; + auto* current_head = head; for (std::size_t i = 0; i < depth; i++) { if (current_head == nullptr) { return nullptr; @@ -245,7 +245,7 @@ ServerSplitterDestinationData* SplitterContext::GetDestinationData(std::size_t i const ServerSplitterDestinationData* SplitterContext::GetDestinationData(std::size_t info, std::size_t data) const { ASSERT(info < info_count); - auto& cur_info = GetInfo(info); + const auto& cur_info = GetInfo(info); return cur_info.GetData(data); } @@ -267,11 +267,11 @@ std::size_t SplitterContext::GetDataCount() const { return data_count; } -void SplitterContext::Setup(std::size_t _info_count, std::size_t _data_count, +void SplitterContext::Setup(std::size_t info_count_, std::size_t data_count_, bool is_splitter_bug_fixed) { - info_count = _info_count; - data_count = _data_count; + info_count = info_count_; + data_count = data_count_; for (std::size_t i = 0; i < info_count; i++) { auto& splitter = infos.emplace_back(static_cast<s32>(i)); @@ -306,7 +306,7 @@ bool SplitterContext::UpdateInfo(const std::vector<u8>& input, std::size_t& inpu break; } - if (header.send_id < 0 || header.send_id > info_count) { + if (header.send_id < 0 || static_cast<std::size_t>(header.send_id) > info_count) { LOG_ERROR(Audio, "Bad splitter data id"); break; } @@ -348,7 +348,7 @@ bool SplitterContext::UpdateData(const std::vector<u8>& input, std::size_t& inpu break; } - if (header.splitter_id < 0 || header.splitter_id > data_count) { + if (header.splitter_id < 0 || static_cast<std::size_t>(header.splitter_id) > data_count) { LOG_ERROR(Audio, "Bad splitter data id"); break; } @@ -364,7 +364,7 @@ bool SplitterContext::RecomposeDestination(ServerSplitterInfo& info, // Clear our current destinations auto* current_head = info.GetHead(); while (current_head != nullptr) { - auto next_head = current_head->GetNextDestination(); + auto* next_head = current_head->GetNextDestination(); current_head->SetNextDestination(nullptr); current_head = next_head; } @@ -434,7 +434,7 @@ const std::vector<s32>& NodeStates::GetIndexList() const { } void NodeStates::PushTsortResult(s32 index) { - ASSERT(index < node_count); + ASSERT(index < static_cast<s32>(node_count)); index_list[index_pos++] = index; } @@ -471,8 +471,8 @@ bool NodeStates::DepthFirstSearch(EdgeMatrix& edge_matrix) { continue; } - const auto node_count = edge_matrix.GetNodeCount(); - for (s32 j = 0; j < static_cast<s32>(node_count); j++) { + const auto edge_node_count = edge_matrix.GetNodeCount(); + for (s32 j = 0; j < static_cast<s32>(edge_node_count); j++) { // Check if our node is connected to our edge matrix if (!edge_matrix.Connected(current_stack_index, j)) { continue; diff --git a/src/audio_core/splitter_context.h b/src/audio_core/splitter_context.h index ea6239fdb..b490627f5 100644 --- a/src/audio_core/splitter_context.h +++ b/src/audio_core/splitter_context.h @@ -63,7 +63,7 @@ public: NodeStates(); ~NodeStates(); - void Initialize(std::size_t _node_count); + void Initialize(std::size_t node_count_); bool Tsort(EdgeMatrix& edge_matrix); std::size_t GetIndexPos() const; const std::vector<s32>& GetIndexList() const; @@ -72,15 +72,15 @@ private: void PushTsortResult(s32 index); bool DepthFirstSearch(EdgeMatrix& edge_matrix); void ResetState(); - void UpdateState(NodeStates::State state, std::size_t i); - NodeStates::State GetState(std::size_t i); + void UpdateState(State state, std::size_t i); + State GetState(std::size_t i); std::size_t node_count{}; std::vector<bool> was_node_found{}; std::vector<bool> was_node_completed{}; std::size_t index_pos{}; std::vector<s32> index_list{}; - NodeStates::Stack index_stack{}; + Stack index_stack{}; }; enum class SplitterMagic : u32_le { @@ -97,8 +97,7 @@ public: s32_le data_count{}; INSERT_PADDING_WORDS(5); }; - static_assert(sizeof(SplitterInfo::InHeader) == 0x20, - "SplitterInfo::InHeader is an invalid size"); + static_assert(sizeof(InHeader) == 0x20, "SplitterInfo::InHeader is an invalid size"); struct InInfoPrams { SplitterMagic magic{}; @@ -107,8 +106,7 @@ public: s32_le length{}; s32_le resource_id_base{}; }; - static_assert(sizeof(SplitterInfo::InInfoPrams) == 0x14, - "SplitterInfo::InInfoPrams is an invalid size"); + static_assert(sizeof(InInfoPrams) == 0x14, "SplitterInfo::InInfoPrams is an invalid size"); struct InDestinationParams { SplitterMagic magic{}; @@ -118,13 +116,13 @@ public: bool in_use{}; INSERT_PADDING_BYTES(3); }; - static_assert(sizeof(SplitterInfo::InDestinationParams) == 0x70, + static_assert(sizeof(InDestinationParams) == 0x70, "SplitterInfo::InDestinationParams is an invalid size"); }; class ServerSplitterDestinationData { public: - explicit ServerSplitterDestinationData(s32 id); + explicit ServerSplitterDestinationData(s32 id_); ~ServerSplitterDestinationData(); void Update(SplitterInfo::InDestinationParams& header); @@ -153,7 +151,7 @@ private: class ServerSplitterInfo { public: - explicit ServerSplitterInfo(s32 id); + explicit ServerSplitterInfo(s32 id_); ~ServerSplitterInfo(); void InitializeInfos(); diff --git a/src/audio_core/stream.cpp b/src/audio_core/stream.cpp index cb33926bc..afe68c9ed 100644 --- a/src/audio_core/stream.cpp +++ b/src/audio_core/stream.cpp @@ -12,7 +12,6 @@ #include "common/assert.h" #include "common/logging/log.h" #include "core/core_timing.h" -#include "core/core_timing_util.h" #include "core/settings.h" namespace AudioCore { @@ -32,10 +31,10 @@ u32 Stream::GetNumChannels() const { return {}; } -Stream::Stream(Core::Timing::CoreTiming& core_timing, u32 sample_rate, Format format, - ReleaseCallback&& release_callback, SinkStream& sink_stream, std::string&& name_) - : sample_rate{sample_rate}, format{format}, release_callback{std::move(release_callback)}, - sink_stream{sink_stream}, core_timing{core_timing}, name{std::move(name_)} { +Stream::Stream(Core::Timing::CoreTiming& core_timing_, u32 sample_rate_, Format format_, + ReleaseCallback&& release_callback_, SinkStream& sink_stream_, std::string&& name_) + : sample_rate{sample_rate_}, format{format_}, release_callback{std::move(release_callback_)}, + sink_stream{sink_stream_}, core_timing{core_timing_}, name{std::move(name_)} { release_event = Core::Timing::CreateEvent(name, [this](std::uintptr_t, std::chrono::nanoseconds ns_late) { ReleaseActiveBuffer(ns_late); @@ -123,7 +122,7 @@ bool Stream::QueueBuffer(BufferPtr&& buffer) { return false; } -bool Stream::ContainsBuffer(Buffer::Tag tag) const { +bool Stream::ContainsBuffer([[maybe_unused]] Buffer::Tag tag) const { UNIMPLEMENTED(); return {}; } @@ -131,7 +130,25 @@ bool Stream::ContainsBuffer(Buffer::Tag tag) const { std::vector<Buffer::Tag> Stream::GetTagsAndReleaseBuffers(std::size_t max_count) { std::vector<Buffer::Tag> tags; for (std::size_t count = 0; count < max_count && !released_buffers.empty(); ++count) { - tags.push_back(released_buffers.front()->GetTag()); + if (released_buffers.front()) { + tags.push_back(released_buffers.front()->GetTag()); + } else { + ASSERT_MSG(false, "Invalid tag in released_buffers!"); + } + released_buffers.pop(); + } + return tags; +} + +std::vector<Buffer::Tag> Stream::GetTagsAndReleaseBuffers() { + std::vector<Buffer::Tag> tags; + tags.reserve(released_buffers.size()); + while (!released_buffers.empty()) { + if (released_buffers.front()) { + tags.push_back(released_buffers.front()->GetTag()); + } else { + ASSERT_MSG(false, "Invalid tag in released_buffers!"); + } released_buffers.pop(); } return tags; diff --git a/src/audio_core/stream.h b/src/audio_core/stream.h index 6437b8591..506ac536b 100644 --- a/src/audio_core/stream.h +++ b/src/audio_core/stream.h @@ -44,8 +44,8 @@ public: /// Callback function type, used to change guest state on a buffer being released using ReleaseCallback = std::function<void()>; - Stream(Core::Timing::CoreTiming& core_timing, u32 sample_rate, Format format, - ReleaseCallback&& release_callback, SinkStream& sink_stream, std::string&& name_); + Stream(Core::Timing::CoreTiming& core_timing_, u32 sample_rate_, Format format_, + ReleaseCallback&& release_callback_, SinkStream& sink_stream_, std::string&& name_); /// Plays the audio stream void Play(); @@ -57,37 +57,40 @@ public: bool QueueBuffer(BufferPtr&& buffer); /// Returns true if the audio stream contains a buffer with the specified tag - bool ContainsBuffer(Buffer::Tag tag) const; + [[nodiscard]] bool ContainsBuffer(Buffer::Tag tag) const; /// Returns a vector of recently released buffers specified by tag - std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(std::size_t max_count); + [[nodiscard]] std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(std::size_t max_count); + + /// Returns a vector of all recently released buffers specified by tag + [[nodiscard]] std::vector<Buffer::Tag> GetTagsAndReleaseBuffers(); void SetVolume(float volume); - float GetVolume() const { + [[nodiscard]] float GetVolume() const { return game_volume; } /// Returns true if the stream is currently playing - bool IsPlaying() const { + [[nodiscard]] bool IsPlaying() const { return state == State::Playing; } /// Returns the number of queued buffers - std::size_t GetQueueSize() const { + [[nodiscard]] std::size_t GetQueueSize() const { return queued_buffers.size(); } /// Gets the sample rate - u32 GetSampleRate() const { + [[nodiscard]] u32 GetSampleRate() const { return sample_rate; } /// Gets the number of channels - u32 GetNumChannels() const; + [[nodiscard]] u32 GetNumChannels() const; /// Get the state - State GetState() const; + [[nodiscard]] State GetState() const; private: /// Plays the next queued buffer in the audio stream, starting playback if necessary @@ -97,7 +100,7 @@ private: void ReleaseActiveBuffer(std::chrono::nanoseconds ns_late = {}); /// Gets the number of core cycles when the specified buffer will be released - std::chrono::nanoseconds GetBufferReleaseNS(const Buffer& buffer) const; + [[nodiscard]] std::chrono::nanoseconds GetBufferReleaseNS(const Buffer& buffer) const; u32 sample_rate; ///< Sample rate of the stream Format format; ///< Format of the stream diff --git a/src/audio_core/voice_context.cpp b/src/audio_core/voice_context.cpp index 1d8f69844..867b8fc6b 100644 --- a/src/audio_core/voice_context.cpp +++ b/src/audio_core/voice_context.cpp @@ -8,7 +8,7 @@ namespace AudioCore { -ServerVoiceChannelResource::ServerVoiceChannelResource(s32 id) : id(id) {} +ServerVoiceChannelResource::ServerVoiceChannelResource(s32 id_) : id(id_) {} ServerVoiceChannelResource::~ServerVoiceChannelResource() = default; bool ServerVoiceChannelResource::InUse() const { @@ -128,7 +128,10 @@ void ServerVoiceInfo::UpdateParameters(const VoiceInfo::InParams& voice_in, in_params.wave_buffer_count = voice_in.wave_buffer_count; in_params.wave_bufffer_head = voice_in.wave_buffer_head; if (behavior_info.IsFlushVoiceWaveBuffersSupported()) { - in_params.wave_buffer_flush_request_count += voice_in.wave_buffer_flush_request_count; + const auto in_request_count = in_params.wave_buffer_flush_request_count; + const auto voice_request_count = voice_in.wave_buffer_flush_request_count; + in_params.wave_buffer_flush_request_count = + static_cast<u8>(in_request_count + voice_request_count); } in_params.mix_id = voice_in.mix_id; if (behavior_info.IsSplitterSupported()) { @@ -206,7 +209,8 @@ void ServerVoiceInfo::UpdateWaveBuffers( void ServerVoiceInfo::UpdateWaveBuffer(ServerWaveBuffer& out_wavebuffer, const WaveBuffer& in_wave_buffer, SampleFormat sample_format, - bool is_buffer_valid, BehaviorInfo& behavior_info) { + bool is_buffer_valid, + [[maybe_unused]] BehaviorInfo& behavior_info) { if (!is_buffer_valid && out_wavebuffer.sent_to_dsp) { out_wavebuffer.buffer_address = 0; out_wavebuffer.buffer_size = 0; @@ -397,7 +401,7 @@ bool ServerVoiceInfo::HasValidWaveBuffer(const VoiceState* state) const { return std::find(valid_wb.begin(), valid_wb.end(), true) != valid_wb.end(); } -VoiceContext::VoiceContext(std::size_t voice_count) : voice_count(voice_count) { +VoiceContext::VoiceContext(std::size_t voice_count_) : voice_count{voice_count_} { for (std::size_t i = 0; i < voice_count; i++) { voice_channel_resources.emplace_back(static_cast<s32>(i)); sorted_voice_info.push_back(&voice_info.emplace_back()); @@ -488,11 +492,11 @@ s32 VoiceContext::DecodePcm16(s32* output_buffer, ServerWaveBuffer* wave_buffer, // Fast path if (channel_count == 1) { - for (std::size_t i = 0; i < samples_processed; i++) { + for (std::ptrdiff_t i = 0; i < samples_processed; i++) { output_buffer[i] = buffer_data[i]; } } else { - for (std::size_t i = 0; i < samples_processed; i++) { + for (std::ptrdiff_t i = 0; i < samples_processed; i++) { output_buffer[i] = buffer_data[i * channel_count + channel]; } } diff --git a/src/audio_core/voice_context.h b/src/audio_core/voice_context.h index 59d3d7dfb..863248761 100644 --- a/src/audio_core/voice_context.h +++ b/src/audio_core/voice_context.h @@ -118,12 +118,12 @@ public: bool in_use{}; INSERT_PADDING_BYTES(11); }; - static_assert(sizeof(VoiceChannelResource::InParams) == 0x70, "InParams is an invalid size"); + static_assert(sizeof(InParams) == 0x70, "InParams is an invalid size"); }; class ServerVoiceChannelResource { public: - explicit ServerVoiceChannelResource(s32 id); + explicit ServerVoiceChannelResource(s32 id_); ~ServerVoiceChannelResource(); bool InUse() const; @@ -174,7 +174,7 @@ public: BehaviorFlags behavior_flags{}; INSERT_PADDING_BYTES(16); }; - static_assert(sizeof(VoiceInfo::InParams) == 0x170, "InParams is an invalid size"); + static_assert(sizeof(InParams) == 0x170, "InParams is an invalid size"); struct OutParams { u64_le played_sample_count{}; @@ -182,7 +182,7 @@ public: u8 voice_dropped{}; INSERT_PADDING_BYTES(3); }; - static_assert(sizeof(VoiceInfo::OutParams) == 0x10, "OutParams is an invalid size"); + static_assert(sizeof(OutParams) == 0x10, "OutParams is an invalid size"); }; class ServerVoiceInfo { @@ -263,7 +263,7 @@ private: class VoiceContext { public: - VoiceContext(std::size_t voice_count); + explicit VoiceContext(std::size_t voice_count_); ~VoiceContext(); std::size_t GetVoiceCount() const; diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt index 5d54516eb..2c2bd2ee8 100644 --- a/src/common/CMakeLists.txt +++ b/src/common/CMakeLists.txt @@ -102,7 +102,9 @@ add_library(common STATIC atomic_ops.h detached_tasks.cpp detached_tasks.h + bit_cast.h bit_field.h + bit_set.h bit_util.h cityhash.cpp cityhash.h @@ -111,6 +113,7 @@ add_library(common STATIC common_paths.h common_types.h concepts.h + div_ceil.h dynamic_library.cpp dynamic_library.h fiber.cpp @@ -132,13 +135,10 @@ add_library(common STATIC math_util.h memory_detect.cpp memory_detect.h - memory_hook.cpp - memory_hook.h microprofile.cpp microprofile.h microprofileui.h misc.cpp - multi_level_queue.h page_table.cpp page_table.h param_package.cpp @@ -150,6 +150,8 @@ add_library(common STATIC scope_exit.h spin_lock.cpp spin_lock.h + stream.cpp + stream.h string_util.cpp string_util.h swap.h @@ -158,6 +160,8 @@ add_library(common STATIC thread.cpp thread.h thread_queue_list.h + thread_worker.cpp + thread_worker.h threadsafe_queue.h time_zone.cpp time_zone.h @@ -188,8 +192,28 @@ if(ARCHITECTURE_x86_64) ) endif() +if (MSVC) + target_compile_definitions(common PRIVATE + # The standard library doesn't provide any replacement for codecvt yet + # so we can disable this deprecation warning for the time being. + _SILENCE_CXX17_CODECVT_HEADER_DEPRECATION_WARNING + ) + target_compile_options(common PRIVATE + /W4 + /WX + ) +else() + target_compile_options(common PRIVATE + -Werror + ) +endif() + create_target_directory_groups(common) -find_package(Boost 1.71 COMPONENTS context headers REQUIRED) target_link_libraries(common PUBLIC ${Boost_LIBRARIES} fmt::fmt microprofile) -target_link_libraries(common PRIVATE lz4::lz4 zstd::zstd xbyak) +target_link_libraries(common PRIVATE lz4::lz4 xbyak) +if (MSVC) + target_link_libraries(common PRIVATE zstd::zstd) +else() + target_link_libraries(common PRIVATE zstd) +endif() diff --git a/src/common/bit_cast.h b/src/common/bit_cast.h new file mode 100644 index 000000000..a32a063d1 --- /dev/null +++ b/src/common/bit_cast.h @@ -0,0 +1,22 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <cstring> +#include <type_traits> + +namespace Common { + +template <typename To, typename From> +[[nodiscard]] std::enable_if_t<sizeof(To) == sizeof(From) && std::is_trivially_copyable_v<From> && + std::is_trivially_copyable_v<To>, + To> +BitCast(const From& src) noexcept { + To dst; + std::memcpy(&dst, &src, sizeof(To)); + return dst; +} + +} // namespace Common diff --git a/src/common/bit_set.h b/src/common/bit_set.h new file mode 100644 index 000000000..9235ad412 --- /dev/null +++ b/src/common/bit_set.h @@ -0,0 +1,99 @@ +/* + * Copyright (c) 2018-2020 Atmosphère-NX + * + * This program is free software; you can redistribute it and/or modify it + * under the terms and conditions of the GNU General Public License, + * version 2, as published by the Free Software Foundation. + * + * This program is distributed in the hope it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#pragma once + +#include <array> +#include <bit> + +#include "common/alignment.h" +#include "common/bit_util.h" +#include "common/common_types.h" + +namespace Common { + +namespace impl { + +template <typename Storage, size_t N> +class BitSet { + +public: + constexpr BitSet() = default; + + constexpr void SetBit(size_t i) { + this->words[i / FlagsPerWord] |= GetBitMask(i % FlagsPerWord); + } + + constexpr void ClearBit(size_t i) { + this->words[i / FlagsPerWord] &= ~GetBitMask(i % FlagsPerWord); + } + + constexpr size_t CountLeadingZero() const { + for (size_t i = 0; i < NumWords; i++) { + if (this->words[i]) { + return FlagsPerWord * i + CountLeadingZeroImpl(this->words[i]); + } + } + return FlagsPerWord * NumWords; + } + + constexpr size_t GetNextSet(size_t n) const { + for (size_t i = (n + 1) / FlagsPerWord; i < NumWords; i++) { + Storage word = this->words[i]; + if (!IsAligned(n + 1, FlagsPerWord)) { + word &= GetBitMask(n % FlagsPerWord) - 1; + } + if (word) { + return FlagsPerWord * i + CountLeadingZeroImpl(word); + } + } + return FlagsPerWord * NumWords; + } + +private: + static_assert(std::is_unsigned_v<Storage>); + static_assert(sizeof(Storage) <= sizeof(u64)); + + static constexpr size_t FlagsPerWord = BitSize<Storage>(); + static constexpr size_t NumWords = AlignUp(N, FlagsPerWord) / FlagsPerWord; + + static constexpr auto CountLeadingZeroImpl(Storage word) { + return std::countl_zero(static_cast<unsigned long long>(word)) - + (BitSize<unsigned long long>() - FlagsPerWord); + } + + static constexpr Storage GetBitMask(size_t bit) { + return Storage(1) << (FlagsPerWord - 1 - bit); + } + + std::array<Storage, NumWords> words{}; +}; + +} // namespace impl + +template <size_t N> +using BitSet8 = impl::BitSet<u8, N>; + +template <size_t N> +using BitSet16 = impl::BitSet<u16, N>; + +template <size_t N> +using BitSet32 = impl::BitSet<u32, N>; + +template <size_t N> +using BitSet64 = impl::BitSet<u64, N>; + +} // namespace Common diff --git a/src/common/concepts.h b/src/common/concepts.h index 5bef3ad67..aa08065a7 100644 --- a/src/common/concepts.h +++ b/src/common/concepts.h @@ -31,4 +31,8 @@ concept DerivedFrom = requires { std::is_convertible_v<const volatile Derived*, const volatile Base*>; }; +// TODO: Replace with std::convertible_to when libc++ implements it. +template <typename From, typename To> +concept ConvertibleTo = std::is_convertible_v<From, To>; + } // namespace Common diff --git a/src/common/div_ceil.h b/src/common/div_ceil.h new file mode 100644 index 000000000..95e1489a9 --- /dev/null +++ b/src/common/div_ceil.h @@ -0,0 +1,26 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <cstddef> +#include <type_traits> + +namespace Common { + +/// Ceiled integer division. +template <typename N, typename D> +requires std::is_integral_v<N>&& std::is_unsigned_v<D>[[nodiscard]] constexpr N DivCeil(N number, + D divisor) { + return static_cast<N>((static_cast<D>(number) + divisor - 1) / divisor); +} + +/// Ceiled integer division with logarithmic divisor in base 2 +template <typename N, typename D> +requires std::is_integral_v<N>&& std::is_unsigned_v<D>[[nodiscard]] constexpr N DivCeilLog2( + N value, D alignment_log2) { + return static_cast<N>((static_cast<D>(value) + (D(1) << alignment_log2) - 1) >> alignment_log2); +} + +} // namespace Common diff --git a/src/common/fiber.cpp b/src/common/fiber.cpp index 1c1d09ccb..3c1eefcb7 100644 --- a/src/common/fiber.cpp +++ b/src/common/fiber.cpp @@ -4,129 +4,51 @@ #include "common/assert.h" #include "common/fiber.h" -#if defined(_WIN32) || defined(WIN32) -#include <windows.h> -#else +#include "common/spin_lock.h" +#include "common/virtual_buffer.h" + #include <boost/context/detail/fcontext.hpp> -#endif namespace Common { -constexpr std::size_t default_stack_size = 256 * 1024; // 256kb - -#if defined(_WIN32) || defined(WIN32) +constexpr std::size_t default_stack_size = 256 * 1024; struct Fiber::FiberImpl { - LPVOID handle = nullptr; - LPVOID rewind_handle = nullptr; + FiberImpl() : stack{default_stack_size}, rewind_stack{default_stack_size} {} + + VirtualBuffer<u8> stack; + VirtualBuffer<u8> rewind_stack; + + SpinLock guard{}; + std::function<void(void*)> entry_point; + std::function<void(void*)> rewind_point; + void* rewind_parameter{}; + void* start_parameter{}; + std::shared_ptr<Fiber> previous_fiber; + bool is_thread_fiber{}; + bool released{}; + + u8* stack_limit{}; + u8* rewind_stack_limit{}; + boost::context::detail::fcontext_t context{}; + boost::context::detail::fcontext_t rewind_context{}; }; -void Fiber::Start() { - ASSERT(previous_fiber != nullptr); - previous_fiber->guard.unlock(); - previous_fiber.reset(); - entry_point(start_parameter); - UNREACHABLE(); -} - -void Fiber::OnRewind() { - ASSERT(impl->handle != nullptr); - DeleteFiber(impl->handle); - impl->handle = impl->rewind_handle; - impl->rewind_handle = nullptr; - rewind_point(rewind_parameter); - UNREACHABLE(); -} - -void Fiber::FiberStartFunc(void* fiber_parameter) { - auto fiber = static_cast<Fiber*>(fiber_parameter); - fiber->Start(); -} - -void Fiber::RewindStartFunc(void* fiber_parameter) { - auto fiber = static_cast<Fiber*>(fiber_parameter); - fiber->OnRewind(); -} - -Fiber::Fiber(std::function<void(void*)>&& entry_point_func, void* start_parameter) - : entry_point{std::move(entry_point_func)}, start_parameter{start_parameter} { - impl = std::make_unique<FiberImpl>(); - impl->handle = CreateFiber(default_stack_size, &FiberStartFunc, this); -} - -Fiber::Fiber() : impl{std::make_unique<FiberImpl>()} {} - -Fiber::~Fiber() { - if (released) { - return; - } - // Make sure the Fiber is not being used - const bool locked = guard.try_lock(); - ASSERT_MSG(locked, "Destroying a fiber that's still running"); - if (locked) { - guard.unlock(); - } - DeleteFiber(impl->handle); -} - -void Fiber::Exit() { - ASSERT_MSG(is_thread_fiber, "Exitting non main thread fiber"); - if (!is_thread_fiber) { - return; - } - ConvertFiberToThread(); - guard.unlock(); - released = true; -} - -void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* start_parameter) { - rewind_point = std::move(rewind_func); - rewind_parameter = start_parameter; -} - -void Fiber::Rewind() { - ASSERT(rewind_point); - ASSERT(impl->rewind_handle == nullptr); - impl->rewind_handle = CreateFiber(default_stack_size, &RewindStartFunc, this); - SwitchToFiber(impl->rewind_handle); -} - -void Fiber::YieldTo(std::shared_ptr<Fiber>& from, std::shared_ptr<Fiber>& to) { - ASSERT_MSG(from != nullptr, "Yielding fiber is null!"); - ASSERT_MSG(to != nullptr, "Next fiber is null!"); - to->guard.lock(); - to->previous_fiber = from; - SwitchToFiber(to->impl->handle); - ASSERT(from->previous_fiber != nullptr); - from->previous_fiber->guard.unlock(); - from->previous_fiber.reset(); +void Fiber::SetStartParameter(void* new_parameter) { + impl->start_parameter = new_parameter; } -std::shared_ptr<Fiber> Fiber::ThreadToFiber() { - std::shared_ptr<Fiber> fiber = std::shared_ptr<Fiber>{new Fiber()}; - fiber->guard.lock(); - fiber->impl->handle = ConvertThreadToFiber(nullptr); - fiber->is_thread_fiber = true; - return fiber; +void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param) { + impl->rewind_point = std::move(rewind_func); + impl->rewind_parameter = rewind_param; } -#else - -struct Fiber::FiberImpl { - alignas(64) std::array<u8, default_stack_size> stack; - alignas(64) std::array<u8, default_stack_size> rewind_stack; - u8* stack_limit; - u8* rewind_stack_limit; - boost::context::detail::fcontext_t context; - boost::context::detail::fcontext_t rewind_context; -}; - void Fiber::Start(boost::context::detail::transfer_t& transfer) { - ASSERT(previous_fiber != nullptr); - previous_fiber->impl->context = transfer.fctx; - previous_fiber->guard.unlock(); - previous_fiber.reset(); - entry_point(start_parameter); + ASSERT(impl->previous_fiber != nullptr); + impl->previous_fiber->impl->context = transfer.fctx; + impl->previous_fiber->impl->guard.unlock(); + impl->previous_fiber.reset(); + impl->entry_point(impl->start_parameter); UNREACHABLE(); } @@ -137,23 +59,24 @@ void Fiber::OnRewind([[maybe_unused]] boost::context::detail::transfer_t& transf u8* tmp = impl->stack_limit; impl->stack_limit = impl->rewind_stack_limit; impl->rewind_stack_limit = tmp; - rewind_point(rewind_parameter); + impl->rewind_point(impl->rewind_parameter); UNREACHABLE(); } void Fiber::FiberStartFunc(boost::context::detail::transfer_t transfer) { - auto fiber = static_cast<Fiber*>(transfer.data); + auto* fiber = static_cast<Fiber*>(transfer.data); fiber->Start(transfer); } void Fiber::RewindStartFunc(boost::context::detail::transfer_t transfer) { - auto fiber = static_cast<Fiber*>(transfer.data); + auto* fiber = static_cast<Fiber*>(transfer.data); fiber->OnRewind(transfer); } Fiber::Fiber(std::function<void(void*)>&& entry_point_func, void* start_parameter) - : entry_point{std::move(entry_point_func)}, start_parameter{start_parameter} { - impl = std::make_unique<FiberImpl>(); + : impl{std::make_unique<FiberImpl>()} { + impl->entry_point = std::move(entry_point_func); + impl->start_parameter = start_parameter; impl->stack_limit = impl->stack.data(); impl->rewind_stack_limit = impl->rewind_stack.data(); u8* stack_base = impl->stack_limit + default_stack_size; @@ -161,37 +84,31 @@ Fiber::Fiber(std::function<void(void*)>&& entry_point_func, void* start_paramete boost::context::detail::make_fcontext(stack_base, impl->stack.size(), FiberStartFunc); } -void Fiber::SetRewindPoint(std::function<void(void*)>&& rewind_func, void* start_parameter) { - rewind_point = std::move(rewind_func); - rewind_parameter = start_parameter; -} - Fiber::Fiber() : impl{std::make_unique<FiberImpl>()} {} Fiber::~Fiber() { - if (released) { + if (impl->released) { return; } // Make sure the Fiber is not being used - const bool locked = guard.try_lock(); + const bool locked = impl->guard.try_lock(); ASSERT_MSG(locked, "Destroying a fiber that's still running"); if (locked) { - guard.unlock(); + impl->guard.unlock(); } } void Fiber::Exit() { - - ASSERT_MSG(is_thread_fiber, "Exitting non main thread fiber"); - if (!is_thread_fiber) { + ASSERT_MSG(impl->is_thread_fiber, "Exitting non main thread fiber"); + if (!impl->is_thread_fiber) { return; } - guard.unlock(); - released = true; + impl->guard.unlock(); + impl->released = true; } void Fiber::Rewind() { - ASSERT(rewind_point); + ASSERT(impl->rewind_point); ASSERT(impl->rewind_context == nullptr); u8* stack_base = impl->rewind_stack_limit + default_stack_size; impl->rewind_context = @@ -199,24 +116,23 @@ void Fiber::Rewind() { boost::context::detail::jump_fcontext(impl->rewind_context, this); } -void Fiber::YieldTo(std::shared_ptr<Fiber>& from, std::shared_ptr<Fiber>& to) { +void Fiber::YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to) { ASSERT_MSG(from != nullptr, "Yielding fiber is null!"); ASSERT_MSG(to != nullptr, "Next fiber is null!"); - to->guard.lock(); - to->previous_fiber = from; + to->impl->guard.lock(); + to->impl->previous_fiber = from; auto transfer = boost::context::detail::jump_fcontext(to->impl->context, to.get()); - ASSERT(from->previous_fiber != nullptr); - from->previous_fiber->impl->context = transfer.fctx; - from->previous_fiber->guard.unlock(); - from->previous_fiber.reset(); + ASSERT(from->impl->previous_fiber != nullptr); + from->impl->previous_fiber->impl->context = transfer.fctx; + from->impl->previous_fiber->impl->guard.unlock(); + from->impl->previous_fiber.reset(); } std::shared_ptr<Fiber> Fiber::ThreadToFiber() { std::shared_ptr<Fiber> fiber = std::shared_ptr<Fiber>{new Fiber()}; - fiber->guard.lock(); - fiber->is_thread_fiber = true; + fiber->impl->guard.lock(); + fiber->impl->is_thread_fiber = true; return fiber; } -#endif } // namespace Common diff --git a/src/common/fiber.h b/src/common/fiber.h index 89dde5e36..f7f587f8c 100644 --- a/src/common/fiber.h +++ b/src/common/fiber.h @@ -7,14 +7,9 @@ #include <functional> #include <memory> -#include "common/common_types.h" -#include "common/spin_lock.h" - -#if !defined(_WIN32) && !defined(WIN32) namespace boost::context::detail { struct transfer_t; } -#endif namespace Common { @@ -46,10 +41,10 @@ public: /// Yields control from Fiber 'from' to Fiber 'to' /// Fiber 'from' must be the currently running fiber. - static void YieldTo(std::shared_ptr<Fiber>& from, std::shared_ptr<Fiber>& to); + static void YieldTo(std::shared_ptr<Fiber> from, std::shared_ptr<Fiber> to); [[nodiscard]] static std::shared_ptr<Fiber> ThreadToFiber(); - void SetRewindPoint(std::function<void(void*)>&& rewind_func, void* start_parameter); + void SetRewindPoint(std::function<void(void*)>&& rewind_func, void* rewind_param); void Rewind(); @@ -57,36 +52,18 @@ public: void Exit(); /// Changes the start parameter of the fiber. Has no effect if the fiber already started - void SetStartParameter(void* new_parameter) { - start_parameter = new_parameter; - } + void SetStartParameter(void* new_parameter); private: Fiber(); -#if defined(_WIN32) || defined(WIN32) - void OnRewind(); - void Start(); - static void FiberStartFunc(void* fiber_parameter); - static void RewindStartFunc(void* fiber_parameter); -#else void OnRewind(boost::context::detail::transfer_t& transfer); void Start(boost::context::detail::transfer_t& transfer); static void FiberStartFunc(boost::context::detail::transfer_t transfer); static void RewindStartFunc(boost::context::detail::transfer_t transfer); -#endif struct FiberImpl; - - SpinLock guard{}; - std::function<void(void*)> entry_point; - std::function<void(void*)> rewind_point; - void* rewind_parameter{}; - void* start_parameter{}; - std::shared_ptr<Fiber> previous_fiber; std::unique_ptr<FiberImpl> impl; - bool is_thread_fiber{}; - bool released{}; }; } // namespace Common diff --git a/src/common/file_util.cpp b/src/common/file_util.cpp index 16c3713e0..18fbfa25b 100644 --- a/src/common/file_util.cpp +++ b/src/common/file_util.cpp @@ -472,13 +472,14 @@ u64 ScanDirectoryTree(const std::string& directory, FSTEntry& parent_entry, } bool DeleteDirRecursively(const std::string& directory, unsigned int recursion) { - const auto callback = [recursion](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { - std::string new_path = directory + DIR_SEP_CHR + virtual_name; + const auto callback = [recursion](u64*, const std::string& directory, + const std::string& virtual_name) { + const std::string new_path = directory + DIR_SEP_CHR + virtual_name; if (IsDirectory(new_path)) { - if (recursion == 0) + if (recursion == 0) { return false; + } return DeleteDirRecursively(new_path, recursion - 1); } return Delete(new_path); @@ -492,7 +493,8 @@ bool DeleteDirRecursively(const std::string& directory, unsigned int recursion) return true; } -void CopyDir(const std::string& source_path, const std::string& dest_path) { +void CopyDir([[maybe_unused]] const std::string& source_path, + [[maybe_unused]] const std::string& dest_path) { #ifndef _WIN32 if (source_path == dest_path) { return; @@ -553,7 +555,7 @@ std::optional<std::string> GetCurrentDir() { std::string strDir = dir; #endif free(dir); - return std::move(strDir); + return strDir; } bool SetCurrentDir(const std::string& directory) { @@ -772,21 +774,23 @@ std::size_t ReadFileToString(bool text_file, const std::string& filename, std::s void SplitFilename83(const std::string& filename, std::array<char, 9>& short_name, std::array<char, 4>& extension) { - const std::string forbidden_characters = ".\"/\\[]:;=, "; + static constexpr std::string_view forbidden_characters = ".\"/\\[]:;=, "; // On a FAT32 partition, 8.3 names are stored as a 11 bytes array, filled with spaces. short_name = {{' ', ' ', ' ', ' ', ' ', ' ', ' ', ' ', '\0'}}; extension = {{' ', ' ', ' ', '\0'}}; - std::string::size_type point = filename.rfind('.'); - if (point == filename.size() - 1) + auto point = filename.rfind('.'); + if (point == filename.size() - 1) { point = filename.rfind('.', point); + } // Get short name. int j = 0; for (char letter : filename.substr(0, point)) { - if (forbidden_characters.find(letter, 0) != std::string::npos) + if (forbidden_characters.find(letter, 0) != std::string::npos) { continue; + } if (j == 8) { // TODO(Link Mauve): also do that for filenames containing a space. // TODO(Link Mauve): handle multiple files having the same short name. @@ -794,14 +798,15 @@ void SplitFilename83(const std::string& filename, std::array<char, 9>& short_nam short_name[7] = '1'; break; } - short_name[j++] = toupper(letter); + short_name[j++] = static_cast<char>(std::toupper(letter)); } // Get extension. if (point != std::string::npos) { j = 0; - for (char letter : filename.substr(point + 1, 3)) - extension[j++] = toupper(letter); + for (char letter : filename.substr(point + 1, 3)) { + extension[j++] = static_cast<char>(std::toupper(letter)); + } } } diff --git a/src/common/file_util.h b/src/common/file_util.h index 8b587320f..840cde2a6 100644 --- a/src/common/file_util.h +++ b/src/common/file_util.h @@ -232,7 +232,7 @@ public: void Swap(IOFile& other) noexcept; - [[nodiscard]] bool Open(const std::string& filename, const char openmode[], int flags = 0); + bool Open(const std::string& filename, const char openmode[], int flags = 0); bool Close(); template <typename T> diff --git a/src/common/hex_util.h b/src/common/hex_util.h index 120f1a5e6..a8d414fb8 100644 --- a/src/common/hex_util.h +++ b/src/common/hex_util.h @@ -16,14 +16,14 @@ namespace Common { [[nodiscard]] constexpr u8 ToHexNibble(char c) { if (c >= 65 && c <= 70) { - return c - 55; + return static_cast<u8>(c - 55); } if (c >= 97 && c <= 102) { - return c - 87; + return static_cast<u8>(c - 87); } - return c - 48; + return static_cast<u8>(c - 48); } [[nodiscard]] std::vector<u8> HexStringToVector(std::string_view str, bool little_endian); @@ -33,11 +33,11 @@ template <std::size_t Size, bool le = false> std::array<u8, Size> out{}; if constexpr (le) { for (std::size_t i = 2 * Size - 2; i <= 2 * Size; i -= 2) { - out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]); + out[i / 2] = static_cast<u8>((ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1])); } } else { for (std::size_t i = 0; i < 2 * Size; i += 2) { - out[i / 2] = (ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1]); + out[i / 2] = static_cast<u8>((ToHexNibble(str[i]) << 4) | ToHexNibble(str[i + 1])); } } return out; diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp index 62cfde397..631f64d05 100644 --- a/src/common/logging/backend.cpp +++ b/src/common/logging/backend.cpp @@ -23,6 +23,7 @@ #include "common/logging/text_formatter.h" #include "common/string_util.h" #include "common/threadsafe_queue.h" +#include "core/settings.h" namespace Log { @@ -152,10 +153,19 @@ FileBackend::FileBackend(const std::string& filename) void FileBackend::Write(const Entry& entry) { // prevent logs from going over the maximum size (in case its spamming and the user doesn't // know) - constexpr std::size_t MAX_BYTES_WRITTEN = 50 * 1024L * 1024L; - if (!file.IsOpen() || bytes_written > MAX_BYTES_WRITTEN) { + constexpr std::size_t MAX_BYTES_WRITTEN = 100 * 1024 * 1024; + constexpr std::size_t MAX_BYTES_WRITTEN_EXTENDED = 1024 * 1024 * 1024; + + if (!file.IsOpen()) { + return; + } + + if (Settings::values.extended_logging && bytes_written > MAX_BYTES_WRITTEN_EXTENDED) { + return; + } else if (!Settings::values.extended_logging && bytes_written > MAX_BYTES_WRITTEN) { return; } + bytes_written += file.WriteString(FormatLogMessage(entry).append(1, '\n')); if (entry.log_level >= Level::Error) { file.Flush(); @@ -222,6 +232,7 @@ void DebuggerBackend::Write(const Entry& entry) { SUB(Service, NPNS) \ SUB(Service, NS) \ SUB(Service, NVDRV) \ + SUB(Service, OLSC) \ SUB(Service, PCIE) \ SUB(Service, PCTL) \ SUB(Service, PCV) \ @@ -274,7 +285,6 @@ const char* GetLogClassName(Class log_class) { case Class::Count: break; } - UNREACHABLE(); return "Invalid"; } @@ -293,7 +303,6 @@ const char* GetLevelName(Level log_level) { break; } #undef LVL - UNREACHABLE(); return "Invalid"; } diff --git a/src/common/logging/log.h b/src/common/logging/log.h index 13a4f1e30..835894918 100644 --- a/src/common/logging/log.h +++ b/src/common/logging/log.h @@ -95,6 +95,7 @@ enum class Class : ClassType { Service_NPNS, ///< The NPNS service Service_NS, ///< The NS services Service_NVDRV, ///< The NVDRV (Nvidia driver) service + Service_OLSC, ///< The OLSC service Service_PCIE, ///< The PCIe service Service_PCTL, ///< The PCTL (Parental control) service Service_PCV, ///< The PCV service diff --git a/src/common/math_util.h b/src/common/math_util.h index b35ad8507..4c38d8040 100644 --- a/src/common/math_util.h +++ b/src/common/math_util.h @@ -20,14 +20,14 @@ struct Rectangle { constexpr Rectangle() = default; - constexpr Rectangle(T left, T top, T right, T bottom) - : left(left), top(top), right(right), bottom(bottom) {} + constexpr Rectangle(T left_, T top_, T right_, T bottom_) + : left(left_), top(top_), right(right_), bottom(bottom_) {} [[nodiscard]] T GetWidth() const { if constexpr (std::is_floating_point_v<T>) { return std::abs(right - left); } else { - return std::abs(static_cast<std::make_signed_t<T>>(right - left)); + return static_cast<T>(std::abs(static_cast<std::make_signed_t<T>>(right - left))); } } @@ -35,7 +35,7 @@ struct Rectangle { if constexpr (std::is_floating_point_v<T>) { return std::abs(bottom - top); } else { - return std::abs(static_cast<std::make_signed_t<T>>(bottom - top)); + return static_cast<T>(std::abs(static_cast<std::make_signed_t<T>>(bottom - top))); } } diff --git a/src/common/memory_hook.cpp b/src/common/memory_hook.cpp deleted file mode 100644 index 3986986d6..000000000 --- a/src/common/memory_hook.cpp +++ /dev/null @@ -1,11 +0,0 @@ -// Copyright 2018 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "common/memory_hook.h" - -namespace Common { - -MemoryHook::~MemoryHook() = default; - -} // namespace Common diff --git a/src/common/memory_hook.h b/src/common/memory_hook.h deleted file mode 100644 index adaa4c2c5..000000000 --- a/src/common/memory_hook.h +++ /dev/null @@ -1,47 +0,0 @@ -// Copyright 2016 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <memory> -#include <optional> - -#include "common/common_types.h" - -namespace Common { - -/** - * Memory hooks have two purposes: - * 1. To allow reads and writes to a region of memory to be intercepted. This is used to implement - * texture forwarding and memory breakpoints for debugging. - * 2. To allow for the implementation of MMIO devices. - * - * A hook may be mapped to multiple regions of memory. - * - * If a std::nullopt or false is returned from a function, the read/write request is passed through - * to the underlying memory region. - */ -class MemoryHook { -public: - virtual ~MemoryHook(); - - virtual std::optional<bool> IsValidAddress(VAddr addr) = 0; - - virtual std::optional<u8> Read8(VAddr addr) = 0; - virtual std::optional<u16> Read16(VAddr addr) = 0; - virtual std::optional<u32> Read32(VAddr addr) = 0; - virtual std::optional<u64> Read64(VAddr addr) = 0; - - virtual bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) = 0; - - virtual bool Write8(VAddr addr, u8 data) = 0; - virtual bool Write16(VAddr addr, u16 data) = 0; - virtual bool Write32(VAddr addr, u32 data) = 0; - virtual bool Write64(VAddr addr, u64 data) = 0; - - virtual bool WriteBlock(VAddr dest_addr, const void* src_buffer, std::size_t size) = 0; -}; - -using MemoryHookPointer = std::shared_ptr<MemoryHook>; -} // namespace Common diff --git a/src/common/misc.cpp b/src/common/misc.cpp index 68cb86cd1..1d5393597 100644 --- a/src/common/misc.cpp +++ b/src/common/misc.cpp @@ -16,16 +16,23 @@ // Call directly after the command or use the error num. // This function might change the error code. std::string GetLastErrorMsg() { - static const std::size_t buff_size = 255; + static constexpr std::size_t buff_size = 255; char err_str[buff_size]; #ifdef _WIN32 FormatMessageA(FORMAT_MESSAGE_FROM_SYSTEM, nullptr, GetLastError(), MAKELANGID(LANG_NEUTRAL, SUBLANG_DEFAULT), err_str, buff_size, nullptr); + return std::string(err_str, buff_size); +#elif defined(__GLIBC__) && (_GNU_SOURCE || (_POSIX_C_SOURCE < 200112L && _XOPEN_SOURCE < 600)) + // Thread safe (GNU-specific) + const char* str = strerror_r(errno, err_str, buff_size); + return std::string(str); #else // Thread safe (XSI-compliant) - strerror_r(errno, err_str, buff_size); + const int success = strerror_r(errno, err_str, buff_size); + if (success != 0) { + return {}; + } + return std::string(err_str); #endif - - return std::string(err_str, buff_size); } diff --git a/src/common/multi_level_queue.h b/src/common/multi_level_queue.h deleted file mode 100644 index 4b305bf40..000000000 --- a/src/common/multi_level_queue.h +++ /dev/null @@ -1,345 +0,0 @@ -// Copyright 2019 TuxSH -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <array> -#include <iterator> -#include <list> -#include <utility> - -#include "common/bit_util.h" -#include "common/common_types.h" - -namespace Common { - -/** - * A MultiLevelQueue is a type of priority queue which has the following characteristics: - * - iteratable through each of its elements. - * - back can be obtained. - * - O(1) add, lookup (both front and back) - * - discrete priorities and a max of 64 priorities (limited domain) - * This type of priority queue is normaly used for managing threads within an scheduler - */ -template <typename T, std::size_t Depth> -class MultiLevelQueue { -public: - using value_type = T; - using reference = value_type&; - using const_reference = const value_type&; - using pointer = value_type*; - using const_pointer = const value_type*; - - using difference_type = typename std::pointer_traits<pointer>::difference_type; - using size_type = std::size_t; - - template <bool is_constant> - class iterator_impl { - public: - using iterator_category = std::bidirectional_iterator_tag; - using value_type = T; - using pointer = std::conditional_t<is_constant, T*, const T*>; - using reference = std::conditional_t<is_constant, const T&, T&>; - using difference_type = typename std::pointer_traits<pointer>::difference_type; - - friend bool operator==(const iterator_impl& lhs, const iterator_impl& rhs) { - if (lhs.IsEnd() && rhs.IsEnd()) - return true; - return std::tie(lhs.current_priority, lhs.it) == std::tie(rhs.current_priority, rhs.it); - } - - friend bool operator!=(const iterator_impl& lhs, const iterator_impl& rhs) { - return !operator==(lhs, rhs); - } - - reference operator*() const { - return *it; - } - - pointer operator->() const { - return it.operator->(); - } - - iterator_impl& operator++() { - if (IsEnd()) { - return *this; - } - - ++it; - - if (it == GetEndItForPrio()) { - u64 prios = mlq.used_priorities; - prios &= ~((1ULL << (current_priority + 1)) - 1); - if (prios == 0) { - current_priority = static_cast<u32>(mlq.depth()); - } else { - current_priority = CountTrailingZeroes64(prios); - it = GetBeginItForPrio(); - } - } - return *this; - } - - iterator_impl& operator--() { - if (IsEnd()) { - if (mlq.used_priorities != 0) { - current_priority = 63 - CountLeadingZeroes64(mlq.used_priorities); - it = GetEndItForPrio(); - --it; - } - } else if (it == GetBeginItForPrio()) { - u64 prios = mlq.used_priorities; - prios &= (1ULL << current_priority) - 1; - if (prios != 0) { - current_priority = CountTrailingZeroes64(prios); - it = GetEndItForPrio(); - --it; - } - } else { - --it; - } - return *this; - } - - iterator_impl operator++(int) { - const iterator_impl v{*this}; - ++(*this); - return v; - } - - iterator_impl operator--(int) { - const iterator_impl v{*this}; - --(*this); - return v; - } - - // allow implicit const->non-const - iterator_impl(const iterator_impl<false>& other) - : mlq(other.mlq), it(other.it), current_priority(other.current_priority) {} - - iterator_impl(const iterator_impl<true>& other) - : mlq(other.mlq), it(other.it), current_priority(other.current_priority) {} - - iterator_impl& operator=(const iterator_impl<false>& other) { - mlq = other.mlq; - it = other.it; - current_priority = other.current_priority; - return *this; - } - - friend class iterator_impl<true>; - iterator_impl() = default; - - private: - friend class MultiLevelQueue; - using container_ref = - std::conditional_t<is_constant, const MultiLevelQueue&, MultiLevelQueue&>; - using list_iterator = std::conditional_t<is_constant, typename std::list<T>::const_iterator, - typename std::list<T>::iterator>; - - explicit iterator_impl(container_ref mlq, list_iterator it, u32 current_priority) - : mlq(mlq), it(it), current_priority(current_priority) {} - explicit iterator_impl(container_ref mlq, u32 current_priority) - : mlq(mlq), it(), current_priority(current_priority) {} - - bool IsEnd() const { - return current_priority == mlq.depth(); - } - - list_iterator GetBeginItForPrio() const { - return mlq.levels[current_priority].begin(); - } - - list_iterator GetEndItForPrio() const { - return mlq.levels[current_priority].end(); - } - - container_ref mlq; - list_iterator it; - u32 current_priority; - }; - - using iterator = iterator_impl<false>; - using const_iterator = iterator_impl<true>; - - void add(const T& element, u32 priority, bool send_back = true) { - if (send_back) - levels[priority].push_back(element); - else - levels[priority].push_front(element); - used_priorities |= 1ULL << priority; - } - - void remove(const T& element, u32 priority) { - auto it = ListIterateTo(levels[priority], element); - if (it == levels[priority].end()) - return; - levels[priority].erase(it); - if (levels[priority].empty()) { - used_priorities &= ~(1ULL << priority); - } - } - - void adjust(const T& element, u32 old_priority, u32 new_priority, bool adjust_front = false) { - remove(element, old_priority); - add(element, new_priority, !adjust_front); - } - void adjust(const_iterator it, u32 old_priority, u32 new_priority, bool adjust_front = false) { - adjust(*it, old_priority, new_priority, adjust_front); - } - - void transfer_to_front(const T& element, u32 priority, MultiLevelQueue& other) { - ListSplice(other.levels[priority], other.levels[priority].begin(), levels[priority], - ListIterateTo(levels[priority], element)); - - other.used_priorities |= 1ULL << priority; - - if (levels[priority].empty()) { - used_priorities &= ~(1ULL << priority); - } - } - - void transfer_to_front(const_iterator it, u32 priority, MultiLevelQueue& other) { - transfer_to_front(*it, priority, other); - } - - void transfer_to_back(const T& element, u32 priority, MultiLevelQueue& other) { - ListSplice(other.levels[priority], other.levels[priority].end(), levels[priority], - ListIterateTo(levels[priority], element)); - - other.used_priorities |= 1ULL << priority; - - if (levels[priority].empty()) { - used_priorities &= ~(1ULL << priority); - } - } - - void transfer_to_back(const_iterator it, u32 priority, MultiLevelQueue& other) { - transfer_to_back(*it, priority, other); - } - - void yield(u32 priority, std::size_t n = 1) { - ListShiftForward(levels[priority], n); - } - - [[nodiscard]] std::size_t depth() const { - return Depth; - } - - [[nodiscard]] std::size_t size(u32 priority) const { - return levels[priority].size(); - } - - [[nodiscard]] std::size_t size() const { - u64 priorities = used_priorities; - std::size_t size = 0; - while (priorities != 0) { - const u64 current_priority = CountTrailingZeroes64(priorities); - size += levels[current_priority].size(); - priorities &= ~(1ULL << current_priority); - } - return size; - } - - [[nodiscard]] bool empty() const { - return used_priorities == 0; - } - - [[nodiscard]] bool empty(u32 priority) const { - return (used_priorities & (1ULL << priority)) == 0; - } - - [[nodiscard]] u32 highest_priority_set(u32 max_priority = 0) const { - const u64 priorities = - max_priority == 0 ? used_priorities : (used_priorities & ~((1ULL << max_priority) - 1)); - return priorities == 0 ? Depth : static_cast<u32>(CountTrailingZeroes64(priorities)); - } - - [[nodiscard]] u32 lowest_priority_set(u32 min_priority = Depth - 1) const { - const u64 priorities = min_priority >= Depth - 1 - ? used_priorities - : (used_priorities & ((1ULL << (min_priority + 1)) - 1)); - return priorities == 0 ? Depth : 63 - CountLeadingZeroes64(priorities); - } - - [[nodiscard]] const_iterator cbegin(u32 max_prio = 0) const { - const u32 priority = highest_priority_set(max_prio); - return priority == Depth ? cend() - : const_iterator{*this, levels[priority].cbegin(), priority}; - } - [[nodiscard]] const_iterator begin(u32 max_prio = 0) const { - return cbegin(max_prio); - } - [[nodiscard]] iterator begin(u32 max_prio = 0) { - const u32 priority = highest_priority_set(max_prio); - return priority == Depth ? end() : iterator{*this, levels[priority].begin(), priority}; - } - - [[nodiscard]] const_iterator cend(u32 min_prio = Depth - 1) const { - return min_prio == Depth - 1 ? const_iterator{*this, Depth} : cbegin(min_prio + 1); - } - [[nodiscard]] const_iterator end(u32 min_prio = Depth - 1) const { - return cend(min_prio); - } - [[nodiscard]] iterator end(u32 min_prio = Depth - 1) { - return min_prio == Depth - 1 ? iterator{*this, Depth} : begin(min_prio + 1); - } - - [[nodiscard]] T& front(u32 max_priority = 0) { - const u32 priority = highest_priority_set(max_priority); - return levels[priority == Depth ? 0 : priority].front(); - } - [[nodiscard]] const T& front(u32 max_priority = 0) const { - const u32 priority = highest_priority_set(max_priority); - return levels[priority == Depth ? 0 : priority].front(); - } - - [[nodiscard]] T& back(u32 min_priority = Depth - 1) { - const u32 priority = lowest_priority_set(min_priority); // intended - return levels[priority == Depth ? 63 : priority].back(); - } - [[nodiscard]] const T& back(u32 min_priority = Depth - 1) const { - const u32 priority = lowest_priority_set(min_priority); // intended - return levels[priority == Depth ? 63 : priority].back(); - } - - void clear() { - used_priorities = 0; - for (std::size_t i = 0; i < Depth; i++) { - levels[i].clear(); - } - } - -private: - using const_list_iterator = typename std::list<T>::const_iterator; - - static void ListShiftForward(std::list<T>& list, const std::size_t shift = 1) { - if (shift >= list.size()) { - return; - } - - const auto begin_range = list.begin(); - const auto end_range = std::next(begin_range, shift); - list.splice(list.end(), list, begin_range, end_range); - } - - static void ListSplice(std::list<T>& in_list, const_list_iterator position, - std::list<T>& out_list, const_list_iterator element) { - in_list.splice(position, out_list, element); - } - - [[nodiscard]] static const_list_iterator ListIterateTo(const std::list<T>& list, - const T& element) { - auto it = list.cbegin(); - while (it != list.cend() && *it != element) { - ++it; - } - return it; - } - - std::array<std::list<T>, Depth> levels; - u64 used_priorities = 0; -}; - -} // namespace Common diff --git a/src/common/page_table.cpp b/src/common/page_table.cpp index e5d3090d5..8fd8620fd 100644 --- a/src/common/page_table.cpp +++ b/src/common/page_table.cpp @@ -8,18 +8,12 @@ namespace Common { PageTable::PageTable() = default; -PageTable::~PageTable() = default; +PageTable::~PageTable() noexcept = default; -void PageTable::Resize(std::size_t address_space_width_in_bits, std::size_t page_size_in_bits, - bool has_attribute) { - const std::size_t num_page_table_entries{1ULL - << (address_space_width_in_bits - page_size_in_bits)}; +void PageTable::Resize(size_t address_space_width_in_bits, size_t page_size_in_bits) { + const size_t num_page_table_entries{1ULL << (address_space_width_in_bits - page_size_in_bits)}; pointers.resize(num_page_table_entries); backing_addr.resize(num_page_table_entries); - - if (has_attribute) { - attributes.resize(num_page_table_entries); - } } } // namespace Common diff --git a/src/common/page_table.h b/src/common/page_table.h index cf5eed780..61c5552e0 100644 --- a/src/common/page_table.h +++ b/src/common/page_table.h @@ -4,12 +4,10 @@ #pragma once -#include <vector> - -#include <boost/icl/interval_map.hpp> +#include <atomic> +#include <tuple> #include "common/common_types.h" -#include "common/memory_hook.h" #include "common/virtual_buffer.h" namespace Common { @@ -22,27 +20,6 @@ enum class PageType : u8 { /// Page is mapped to regular memory, but also needs to check for rasterizer cache flushing and /// invalidation RasterizerCachedMemory, - /// Page is mapped to a I/O region. Writing and reading to this page is handled by functions. - Special, - /// Page is allocated for use. - Allocated, -}; - -struct SpecialRegion { - enum class Type { - DebugHook, - IODevice, - } type; - - MemoryHookPointer handler; - - [[nodiscard]] bool operator<(const SpecialRegion& other) const { - return std::tie(type, handler) < std::tie(other.type, other.handler); - } - - [[nodiscard]] bool operator==(const SpecialRegion& other) const { - return std::tie(type, handler) == std::tie(other.type, other.handler); - } }; /** @@ -50,27 +27,84 @@ struct SpecialRegion { * mimics the way a real CPU page table works. */ struct PageTable { + /// Number of bits reserved for attribute tagging. + /// This can be at most the guaranteed alignment of the pointers in the page table. + static constexpr int ATTRIBUTE_BITS = 2; + + /** + * Pair of host pointer and page type attribute. + * This uses the lower bits of a given pointer to store the attribute tag. + * Writing and reading the pointer attribute pair is guaranteed to be atomic for the same method + * call. In other words, they are guaranteed to be synchronized at all times. + */ + class PageInfo { + public: + /// Returns the page pointer + [[nodiscard]] u8* Pointer() const noexcept { + return ExtractPointer(raw.load(std::memory_order_relaxed)); + } + + /// Returns the page type attribute + [[nodiscard]] PageType Type() const noexcept { + return ExtractType(raw.load(std::memory_order_relaxed)); + } + + /// Returns the page pointer and attribute pair, extracted from the same atomic read + [[nodiscard]] std::pair<u8*, PageType> PointerType() const noexcept { + const uintptr_t non_atomic_raw = raw.load(std::memory_order_relaxed); + return {ExtractPointer(non_atomic_raw), ExtractType(non_atomic_raw)}; + } + + /// Returns the raw representation of the page information. + /// Use ExtractPointer and ExtractType to unpack the value. + [[nodiscard]] uintptr_t Raw() const noexcept { + return raw.load(std::memory_order_relaxed); + } + + /// Write a page pointer and type pair atomically + void Store(u8* pointer, PageType type) noexcept { + raw.store(reinterpret_cast<uintptr_t>(pointer) | static_cast<uintptr_t>(type)); + } + + /// Unpack a pointer from a page info raw representation + [[nodiscard]] static u8* ExtractPointer(uintptr_t raw) noexcept { + return reinterpret_cast<u8*>(raw & (~uintptr_t{0} << ATTRIBUTE_BITS)); + } + + /// Unpack a page type from a page info raw representation + [[nodiscard]] static PageType ExtractType(uintptr_t raw) noexcept { + return static_cast<PageType>(raw & ((uintptr_t{1} << ATTRIBUTE_BITS) - 1)); + } + + private: + std::atomic<uintptr_t> raw; + }; + PageTable(); - ~PageTable(); + ~PageTable() noexcept; + + PageTable(const PageTable&) = delete; + PageTable& operator=(const PageTable&) = delete; + + PageTable(PageTable&&) noexcept = default; + PageTable& operator=(PageTable&&) noexcept = default; /** - * Resizes the page table to be able to accomodate enough pages within + * Resizes the page table to be able to accommodate enough pages within * a given address space. * * @param address_space_width_in_bits The address size width in bits. + * @param page_size_in_bits The page size in bits. */ - void Resize(std::size_t address_space_width_in_bits, std::size_t page_size_in_bits, - bool has_attribute); + void Resize(size_t address_space_width_in_bits, size_t page_size_in_bits); /** * Vector of memory pointers backing each page. An entry can only be non-null if the - * corresponding entry in the `attributes` vector is of type `Memory`. + * corresponding attribute element is of type `Memory`. */ - VirtualBuffer<u8*> pointers; + VirtualBuffer<PageInfo> pointers; VirtualBuffer<u64> backing_addr; - - VirtualBuffer<PageType> attributes; }; } // namespace Common diff --git a/src/common/scope_exit.h b/src/common/scope_exit.h index 68ef5f197..fa46cb394 100644 --- a/src/common/scope_exit.h +++ b/src/common/scope_exit.h @@ -10,7 +10,7 @@ namespace detail { template <typename Func> struct ScopeExitHelper { - explicit ScopeExitHelper(Func&& func) : func(std::move(func)) {} + explicit ScopeExitHelper(Func&& func_) : func(std::move(func_)) {} ~ScopeExitHelper() { if (active) { func(); diff --git a/src/common/spin_lock.h b/src/common/spin_lock.h index 4f946a258..06ac2f5bb 100644 --- a/src/common/spin_lock.h +++ b/src/common/spin_lock.h @@ -15,6 +15,14 @@ namespace Common { */ class SpinLock { public: + SpinLock() = default; + + SpinLock(const SpinLock&) = delete; + SpinLock& operator=(const SpinLock&) = delete; + + SpinLock(SpinLock&&) = delete; + SpinLock& operator=(SpinLock&&) = delete; + void lock(); void unlock(); [[nodiscard]] bool try_lock(); diff --git a/src/common/stream.cpp b/src/common/stream.cpp new file mode 100644 index 000000000..bf0496c26 --- /dev/null +++ b/src/common/stream.cpp @@ -0,0 +1,47 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <stdexcept> +#include "common/common_types.h" +#include "common/stream.h" + +namespace Common { + +Stream::Stream() = default; +Stream::~Stream() = default; + +void Stream::Seek(s32 offset, SeekOrigin origin) { + if (origin == SeekOrigin::SetOrigin) { + if (offset < 0) { + position = 0; + } else if (position >= buffer.size()) { + position = buffer.size(); + } else { + position = offset; + } + } else if (origin == SeekOrigin::FromCurrentPos) { + Seek(static_cast<s32>(position) + offset, SeekOrigin::SetOrigin); + } else if (origin == SeekOrigin::FromEnd) { + Seek(static_cast<s32>(buffer.size()) - offset, SeekOrigin::SetOrigin); + } +} + +u8 Stream::ReadByte() { + if (position < buffer.size()) { + return buffer[position++]; + } else { + throw std::out_of_range("Attempting to read a byte not within the buffer range"); + } +} + +void Stream::WriteByte(u8 byte) { + if (position == buffer.size()) { + buffer.push_back(byte); + position++; + } else { + buffer.insert(buffer.begin() + position, byte); + } +} + +} // namespace Common diff --git a/src/common/stream.h b/src/common/stream.h new file mode 100644 index 000000000..0e40692de --- /dev/null +++ b/src/common/stream.h @@ -0,0 +1,56 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> +#include "common/common_types.h" + +namespace Common { + +enum class SeekOrigin { + SetOrigin, + FromCurrentPos, + FromEnd, +}; + +class Stream { +public: + /// Stream creates a bitstream and provides common functionality on the stream. + explicit Stream(); + ~Stream(); + + Stream(const Stream&) = delete; + Stream& operator=(const Stream&) = delete; + + Stream(Stream&&) = default; + Stream& operator=(Stream&&) = default; + + /// Reposition bitstream "cursor" to the specified offset from origin + void Seek(s32 offset, SeekOrigin origin); + + /// Reads next byte in the stream buffer and increments position + u8 ReadByte(); + + /// Writes byte at current position + void WriteByte(u8 byte); + + [[nodiscard]] std::size_t GetPosition() const { + return position; + } + + [[nodiscard]] std::vector<u8>& GetBuffer() { + return buffer; + } + + [[nodiscard]] const std::vector<u8>& GetBuffer() const { + return buffer; + } + +private: + std::vector<u8> buffer; + std::size_t position{0}; +}; + +} // namespace Common diff --git a/src/common/string_util.cpp b/src/common/string_util.cpp index 84883a1d3..4cba2aaa4 100644 --- a/src/common/string_util.cpp +++ b/src/common/string_util.cpp @@ -8,6 +8,7 @@ #include <cstdlib> #include <locale> #include <sstream> + #include "common/common_paths.h" #include "common/logging/log.h" #include "common/string_util.h" @@ -21,14 +22,14 @@ namespace Common { /// Make a string lowercase std::string ToLower(std::string str) { std::transform(str.begin(), str.end(), str.begin(), - [](unsigned char c) { return std::tolower(c); }); + [](unsigned char c) { return static_cast<char>(std::tolower(c)); }); return str; } /// Make a string uppercase std::string ToUpper(std::string str) { std::transform(str.begin(), str.end(), str.begin(), - [](unsigned char c) { return std::toupper(c); }); + [](unsigned char c) { return static_cast<char>(std::toupper(c)); }); return str; } diff --git a/src/common/swap.h b/src/common/swap.h index 7665942a2..a80e191dc 100644 --- a/src/common/swap.h +++ b/src/common/swap.h @@ -394,7 +394,7 @@ public: template <typename S, typename T2, typename F2> friend S operator%(const S& p, const swapped_t v); - // Arithmetics + assignements + // Arithmetics + assignments template <typename S, typename T2, typename F2> friend S operator+=(const S& p, const swapped_t v); @@ -451,7 +451,7 @@ S operator%(const S& i, const swap_struct_t<T, F> v) { return i % v.swap(); } -// Arithmetics + assignements +// Arithmetics + assignments template <typename S, typename T, typename F> S& operator+=(S& i, const swap_struct_t<T, F> v) { i += v.swap(); diff --git a/src/common/telemetry.h b/src/common/telemetry.h index a50c5d1de..49186e848 100644 --- a/src/common/telemetry.h +++ b/src/common/telemetry.h @@ -52,8 +52,8 @@ public: template <typename T> class Field : public FieldInterface { public: - Field(FieldType type, std::string name, T value) - : name(std::move(name)), type(type), value(std::move(value)) {} + Field(FieldType type_, std::string name_, T value_) + : name(std::move(name_)), type(type_), value(std::move(value_)) {} Field(const Field&) = default; Field& operator=(const Field&) = default; diff --git a/src/common/thread_worker.cpp b/src/common/thread_worker.cpp new file mode 100644 index 000000000..8f9bf447a --- /dev/null +++ b/src/common/thread_worker.cpp @@ -0,0 +1,58 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/thread.h" +#include "common/thread_worker.h" + +namespace Common { + +ThreadWorker::ThreadWorker(std::size_t num_workers, const std::string& name) { + for (std::size_t i = 0; i < num_workers; ++i) + threads.emplace_back([this, thread_name{std::string{name}}] { + Common::SetCurrentThreadName(thread_name.c_str()); + + // Wait for first request + { + std::unique_lock lock{queue_mutex}; + condition.wait(lock, [this] { return stop || !requests.empty(); }); + } + + while (true) { + std::function<void()> task; + + { + std::unique_lock lock{queue_mutex}; + condition.wait(lock, [this] { return stop || !requests.empty(); }); + if (stop || requests.empty()) { + return; + } + task = std::move(requests.front()); + requests.pop(); + } + + task(); + } + }); +} + +ThreadWorker::~ThreadWorker() { + { + std::unique_lock lock{queue_mutex}; + stop = true; + } + condition.notify_all(); + for (std::thread& thread : threads) { + thread.join(); + } +} + +void ThreadWorker::QueueWork(std::function<void()>&& work) { + { + std::unique_lock lock{queue_mutex}; + requests.emplace(work); + } + condition.notify_one(); +} + +} // namespace Common diff --git a/src/common/thread_worker.h b/src/common/thread_worker.h new file mode 100644 index 000000000..f1859971f --- /dev/null +++ b/src/common/thread_worker.h @@ -0,0 +1,30 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <atomic> +#include <functional> +#include <mutex> +#include <string> +#include <vector> +#include <queue> + +namespace Common { + +class ThreadWorker final { +public: + explicit ThreadWorker(std::size_t num_workers, const std::string& name); + ~ThreadWorker(); + void QueueWork(std::function<void()>&& work); + +private: + std::vector<std::thread> threads; + std::queue<std::function<void()>> requests; + std::mutex queue_mutex; + std::condition_variable condition; + std::atomic_bool stop{}; +}; + +} // namespace Common diff --git a/src/common/timer.cpp b/src/common/timer.cpp index 2dc15e434..d17dc2a50 100644 --- a/src/common/timer.cpp +++ b/src/common/timer.cpp @@ -142,20 +142,18 @@ std::string Timer::GetTimeFormatted() { // ---------------- double Timer::GetDoubleTime() { // Get continuous timestamp - u64 TmpSeconds = static_cast<u64>(Common::Timer::GetTimeSinceJan1970().count()); - double ms = static_cast<u64>(GetTimeMs().count()) % 1000; + auto tmp_seconds = static_cast<u64>(GetTimeSinceJan1970().count()); + const auto ms = static_cast<double>(static_cast<u64>(GetTimeMs().count()) % 1000); // Remove a few years. We only really want enough seconds to make // sure that we are detecting actual actions, perhaps 60 seconds is // enough really, but I leave a year of seconds anyway, in case the // user's clock is incorrect or something like that. - TmpSeconds = TmpSeconds - (38 * 365 * 24 * 60 * 60); + tmp_seconds = tmp_seconds - (38 * 365 * 24 * 60 * 60); // Make a smaller integer that fits in the double - u32 Seconds = static_cast<u32>(TmpSeconds); - double TmpTime = Seconds + ms; - - return TmpTime; + const auto seconds = static_cast<u32>(tmp_seconds); + return seconds + ms; } } // Namespace Common diff --git a/src/common/vector_math.h b/src/common/vector_math.h index 2a0fcf541..22dba3c2d 100644 --- a/src/common/vector_math.h +++ b/src/common/vector_math.h @@ -87,7 +87,13 @@ public: template <typename V> [[nodiscard]] constexpr Vec2<decltype(T{} * V{})> operator*(const V& f) const { - return {x * f, y * f}; + using TV = decltype(T{} * V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)), + }; } template <typename V> @@ -98,7 +104,13 @@ public: template <typename V> [[nodiscard]] constexpr Vec2<decltype(T{} / V{})> operator/(const V& f) const { - return {x / f, y / f}; + using TV = decltype(T{} / V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)), + }; } template <typename V> @@ -168,7 +180,10 @@ public: template <typename T, typename V> [[nodiscard]] constexpr Vec2<T> operator*(const V& f, const Vec2<T>& vec) { - return Vec2<T>(f * vec.x, f * vec.y); + using C = std::common_type_t<T, V>; + + return Vec2<T>(static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.x)), + static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.y))); } using Vec2f = Vec2<float>; @@ -237,7 +252,14 @@ public: template <typename V> [[nodiscard]] constexpr Vec3<decltype(T{} * V{})> operator*(const V& f) const { - return {x * f, y * f, z * f}; + using TV = decltype(T{} * V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) * static_cast<C>(f)), + }; } template <typename V> @@ -247,7 +269,14 @@ public: } template <typename V> [[nodiscard]] constexpr Vec3<decltype(T{} / V{})> operator/(const V& f) const { - return {x / f, y / f, z / f}; + using TV = decltype(T{} / V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) / static_cast<C>(f)), + }; } template <typename V> @@ -367,7 +396,11 @@ public: template <typename T, typename V> [[nodiscard]] constexpr Vec3<T> operator*(const V& f, const Vec3<T>& vec) { - return Vec3<T>(f * vec.x, f * vec.y, f * vec.z); + using C = std::common_type_t<T, V>; + + return Vec3<T>(static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.x)), + static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.y)), + static_cast<T>(static_cast<C>(f) * static_cast<C>(vec.z))); } template <> @@ -446,7 +479,15 @@ public: template <typename V> [[nodiscard]] constexpr Vec4<decltype(T{} * V{})> operator*(const V& f) const { - return {x * f, y * f, z * f, w * f}; + using TV = decltype(T{} * V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) * static_cast<C>(f)), + static_cast<TV>(static_cast<C>(w) * static_cast<C>(f)), + }; } template <typename V> @@ -457,7 +498,15 @@ public: template <typename V> [[nodiscard]] constexpr Vec4<decltype(T{} / V{})> operator/(const V& f) const { - return {x / f, y / f, z / f, w / f}; + using TV = decltype(T{} / V{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(x) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(y) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(z) / static_cast<C>(f)), + static_cast<TV>(static_cast<C>(w) / static_cast<C>(f)), + }; } template <typename V> @@ -582,7 +631,15 @@ public: template <typename T, typename V> [[nodiscard]] constexpr Vec4<decltype(V{} * T{})> operator*(const V& f, const Vec4<T>& vec) { - return {f * vec.x, f * vec.y, f * vec.z, f * vec.w}; + using TV = decltype(V{} * T{}); + using C = std::common_type_t<T, V>; + + return { + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.x)), + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.y)), + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.z)), + static_cast<TV>(static_cast<C>(f) * static_cast<C>(vec.w)), + }; } using Vec4f = Vec4<float>; diff --git a/src/common/virtual_buffer.cpp b/src/common/virtual_buffer.cpp index b009cb500..e3ca29258 100644 --- a/src/common/virtual_buffer.cpp +++ b/src/common/virtual_buffer.cpp @@ -13,7 +13,7 @@ namespace Common { -void* AllocateMemoryPages(std::size_t size) { +void* AllocateMemoryPages(std::size_t size) noexcept { #ifdef _WIN32 void* base{VirtualAlloc(nullptr, size, MEM_COMMIT, PAGE_READWRITE)}; #else @@ -29,7 +29,7 @@ void* AllocateMemoryPages(std::size_t size) { return base; } -void FreeMemoryPages(void* base, [[maybe_unused]] std::size_t size) { +void FreeMemoryPages(void* base, [[maybe_unused]] std::size_t size) noexcept { if (!base) { return; } diff --git a/src/common/virtual_buffer.h b/src/common/virtual_buffer.h index 125cb42f0..fb1a6f81f 100644 --- a/src/common/virtual_buffer.h +++ b/src/common/virtual_buffer.h @@ -4,29 +4,55 @@ #pragma once -#include "common/common_funcs.h" +#include <type_traits> +#include <utility> namespace Common { -void* AllocateMemoryPages(std::size_t size); -void FreeMemoryPages(void* base, std::size_t size); +void* AllocateMemoryPages(std::size_t size) noexcept; +void FreeMemoryPages(void* base, std::size_t size) noexcept; template <typename T> -class VirtualBuffer final : NonCopyable { +class VirtualBuffer final { public: + // TODO: Uncomment this and change Common::PageTable::PageInfo to be trivially constructible + // using std::atomic_ref once libc++ has support for it + // static_assert( + // std::is_trivially_constructible_v<T>, + // "T must be trivially constructible, as non-trivial constructors will not be executed " + // "with the current allocator"); + constexpr VirtualBuffer() = default; explicit VirtualBuffer(std::size_t count) : alloc_size{count * sizeof(T)} { base_ptr = reinterpret_cast<T*>(AllocateMemoryPages(alloc_size)); } - ~VirtualBuffer() { + ~VirtualBuffer() noexcept { FreeMemoryPages(base_ptr, alloc_size); } + VirtualBuffer(const VirtualBuffer&) = delete; + VirtualBuffer& operator=(const VirtualBuffer&) = delete; + + VirtualBuffer(VirtualBuffer&& other) noexcept + : alloc_size{std::exchange(other.alloc_size, 0)}, base_ptr{std::exchange(other.base_ptr), + nullptr} {} + + VirtualBuffer& operator=(VirtualBuffer&& other) noexcept { + alloc_size = std::exchange(other.alloc_size, 0); + base_ptr = std::exchange(other.base_ptr, nullptr); + return *this; + } + void resize(std::size_t count) { + const auto new_size = count * sizeof(T); + if (new_size == alloc_size) { + return; + } + FreeMemoryPages(base_ptr, alloc_size); - alloc_size = count * sizeof(T); + alloc_size = new_size; base_ptr = reinterpret_cast<T*>(AllocateMemoryPages(alloc_size)); } diff --git a/src/common/wall_clock.cpp b/src/common/wall_clock.cpp index 3afbdb898..a8c143f85 100644 --- a/src/common/wall_clock.cpp +++ b/src/common/wall_clock.cpp @@ -15,10 +15,10 @@ namespace Common { using base_timer = std::chrono::steady_clock; using base_time_point = std::chrono::time_point<base_timer>; -class StandardWallClock : public WallClock { +class StandardWallClock final : public WallClock { public: - StandardWallClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency) - : WallClock(emulated_cpu_frequency, emulated_clock_frequency, false) { + explicit StandardWallClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_) + : WallClock(emulated_cpu_frequency_, emulated_clock_frequency_, false) { start_time = base_timer::now(); } @@ -53,7 +53,7 @@ public: return Common::Divide128On32(temporary, 1000000000).first; } - void Pause(bool is_paused) override { + void Pause([[maybe_unused]] bool is_paused) override { // Do nothing in this clock type. } diff --git a/src/common/wall_clock.h b/src/common/wall_clock.h index 5db30083d..cef3e9499 100644 --- a/src/common/wall_clock.h +++ b/src/common/wall_clock.h @@ -13,6 +13,8 @@ namespace Common { class WallClock { public: + virtual ~WallClock() = default; + /// Returns current wall time in nanoseconds [[nodiscard]] virtual std::chrono::nanoseconds GetTimeNS() = 0; @@ -36,9 +38,9 @@ public: } protected: - WallClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency, bool is_native) - : emulated_cpu_frequency{emulated_cpu_frequency}, - emulated_clock_frequency{emulated_clock_frequency}, is_native{is_native} {} + explicit WallClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_, bool is_native_) + : emulated_cpu_frequency{emulated_cpu_frequency_}, + emulated_clock_frequency{emulated_clock_frequency_}, is_native{is_native_} {} u64 emulated_cpu_frequency; u64 emulated_clock_frequency; diff --git a/src/common/x64/native_clock.cpp b/src/common/x64/native_clock.cpp index 424b39b1f..eb8a7782f 100644 --- a/src/common/x64/native_clock.cpp +++ b/src/common/x64/native_clock.cpp @@ -43,10 +43,10 @@ u64 EstimateRDTSCFrequency() { } namespace X64 { -NativeClock::NativeClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency, - u64 rtsc_frequency) - : WallClock(emulated_cpu_frequency, emulated_clock_frequency, true), rtsc_frequency{ - rtsc_frequency} { +NativeClock::NativeClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_, + u64 rtsc_frequency_) + : WallClock(emulated_cpu_frequency_, emulated_clock_frequency_, true), rtsc_frequency{ + rtsc_frequency_} { _mm_mfence(); last_measure = __rdtsc(); accumulated_ticks = 0U; diff --git a/src/common/x64/native_clock.h b/src/common/x64/native_clock.h index 891a3bbfd..6d1e32ac8 100644 --- a/src/common/x64/native_clock.h +++ b/src/common/x64/native_clock.h @@ -12,9 +12,10 @@ namespace Common { namespace X64 { -class NativeClock : public WallClock { +class NativeClock final : public WallClock { public: - NativeClock(u64 emulated_cpu_frequency, u64 emulated_clock_frequency, u64 rtsc_frequency); + explicit NativeClock(u64 emulated_cpu_frequency_, u64 emulated_clock_frequency_, + u64 rtsc_frequency_); std::chrono::nanoseconds GetTimeNS() override; @@ -34,7 +35,7 @@ private: /// value used to reduce the native clocks accuracy as some apss rely on /// undefined behavior where the level of accuracy in the clock shouldn't /// be higher. - static constexpr u64 inaccuracy_mask = ~(0x400 - 1); + static constexpr u64 inaccuracy_mask = ~(UINT64_C(0x400) - 1); SpinLock rtsc_serialize{}; u64 last_measure{}; diff --git a/src/common/x64/xbyak_abi.h b/src/common/x64/xbyak_abi.h index 26e4bfda5..c2c9b6134 100644 --- a/src/common/x64/xbyak_abi.h +++ b/src/common/x64/xbyak_abi.h @@ -11,25 +11,25 @@ namespace Common::X64 { -constexpr std::size_t RegToIndex(const Xbyak::Reg& reg) { +constexpr size_t RegToIndex(const Xbyak::Reg& reg) { using Kind = Xbyak::Reg::Kind; ASSERT_MSG((reg.getKind() & (Kind::REG | Kind::XMM)) != 0, "RegSet only support GPRs and XMM registers."); ASSERT_MSG(reg.getIdx() < 16, "RegSet only supports XXM0-15."); - return reg.getIdx() + (reg.getKind() == Kind::REG ? 0 : 16); + return static_cast<size_t>(reg.getIdx()) + (reg.getKind() == Kind::REG ? 0 : 16); } -constexpr Xbyak::Reg64 IndexToReg64(std::size_t reg_index) { +constexpr Xbyak::Reg64 IndexToReg64(size_t reg_index) { ASSERT(reg_index < 16); return Xbyak::Reg64(static_cast<int>(reg_index)); } -constexpr Xbyak::Xmm IndexToXmm(std::size_t reg_index) { +constexpr Xbyak::Xmm IndexToXmm(size_t reg_index) { ASSERT(reg_index >= 16 && reg_index < 32); return Xbyak::Xmm(static_cast<int>(reg_index - 16)); } -constexpr Xbyak::Reg IndexToReg(std::size_t reg_index) { +constexpr Xbyak::Reg IndexToReg(size_t reg_index) { if (reg_index < 16) { return IndexToReg64(reg_index); } else { @@ -182,7 +182,7 @@ inline size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, std::b size_t rsp_alignment, size_t needed_frame_size = 0) { auto frame_info = ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size); - for (std::size_t i = 0; i < regs.size(); ++i) { + for (size_t i = 0; i < regs.size(); ++i) { if (regs[i] && ABI_ALL_GPRS[i]) { code.push(IndexToReg64(i)); } @@ -192,7 +192,7 @@ inline size_t ABI_PushRegistersAndAdjustStack(Xbyak::CodeGenerator& code, std::b code.sub(code.rsp, frame_info.subtraction); } - for (std::size_t i = 0; i < regs.size(); ++i) { + for (size_t i = 0; i < regs.size(); ++i) { if (regs[i] && ABI_ALL_XMMS[i]) { code.movaps(code.xword[code.rsp + frame_info.xmm_offset], IndexToXmm(i)); frame_info.xmm_offset += 0x10; @@ -206,7 +206,7 @@ inline void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, std::bits size_t rsp_alignment, size_t needed_frame_size = 0) { auto frame_info = ABI_CalculateFrameSize(regs, rsp_alignment, needed_frame_size); - for (std::size_t i = 0; i < regs.size(); ++i) { + for (size_t i = 0; i < regs.size(); ++i) { if (regs[i] && ABI_ALL_XMMS[i]) { code.movaps(IndexToXmm(i), code.xword[code.rsp + frame_info.xmm_offset]); frame_info.xmm_offset += 0x10; @@ -218,8 +218,8 @@ inline void ABI_PopRegistersAndAdjustStack(Xbyak::CodeGenerator& code, std::bits } // GPRs need to be popped in reverse order - for (std::size_t j = 0; j < regs.size(); ++j) { - const std::size_t i = regs.size() - j - 1; + for (size_t j = 0; j < regs.size(); ++j) { + const size_t i = regs.size() - j - 1; if (regs[i] && ABI_ALL_GPRS[i]) { code.pop(IndexToReg64(i)); } diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt index d0c405ec7..893df433a 100644 --- a/src/core/CMakeLists.txt +++ b/src/core/CMakeLists.txt @@ -1,9 +1,3 @@ -if (YUZU_ENABLE_BOXCAT) - set(BCAT_BOXCAT_ADDITIONAL_SOURCES hle/service/bcat/backend/boxcat.cpp hle/service/bcat/backend/boxcat.h) -else() - set(BCAT_BOXCAT_ADDITIONAL_SOURCES) -endif() - add_library(core STATIC arm/arm_interface.h arm/arm_interface.cpp @@ -19,8 +13,6 @@ add_library(core STATIC arm/dynarmic/arm_exclusive_monitor.h arm/exclusive_monitor.cpp arm/exclusive_monitor.h - arm/unicorn/arm_unicorn.cpp - arm/unicorn/arm_unicorn.h constants.cpp constants.h core.cpp @@ -49,6 +41,7 @@ add_library(core STATIC file_sys/bis_factory.h file_sys/card_image.cpp file_sys/card_image.h + file_sys/common_funcs.h file_sys/content_archive.cpp file_sys/content_archive.h file_sys/control_metadata.cpp @@ -142,9 +135,9 @@ add_library(core STATIC frontend/emu_window.h frontend/framebuffer_layout.cpp frontend/framebuffer_layout.h + frontend/input_interpreter.cpp + frontend/input_interpreter.h frontend/input.h - gdbstub/gdbstub.cpp - gdbstub/gdbstub.h hardware_interrupt_manager.cpp hardware_interrupt_manager.h hle/ipc.h @@ -158,10 +151,19 @@ add_library(core STATIC hle/kernel/code_set.cpp hle/kernel/code_set.h hle/kernel/errors.h + hle/kernel/global_scheduler_context.cpp + hle/kernel/global_scheduler_context.h hle/kernel/handle_table.cpp hle/kernel/handle_table.h hle/kernel/hle_ipc.cpp hle/kernel/hle_ipc.h + hle/kernel/k_affinity_mask.h + hle/kernel/k_priority_queue.h + hle/kernel/k_scheduler.cpp + hle/kernel/k_scheduler.h + hle/kernel/k_scheduler_lock.h + hle/kernel/k_scoped_lock.h + hle/kernel/k_scoped_scheduler_lock_and_sleep.h hle/kernel/kernel.cpp hle/kernel/kernel.h hle/kernel/memory/address_space_info.cpp @@ -196,12 +198,12 @@ add_library(core STATIC hle/kernel/readable_event.h hle/kernel/resource_limit.cpp hle/kernel/resource_limit.h - hle/kernel/scheduler.cpp - hle/kernel/scheduler.h hle/kernel/server_port.cpp hle/kernel/server_port.h hle/kernel/server_session.cpp hle/kernel/server_session.h + hle/kernel/service_thread.cpp + hle/kernel/service_thread.h hle/kernel/session.cpp hle/kernel/session.h hle/kernel/shared_memory.cpp @@ -303,7 +305,6 @@ add_library(core STATIC hle/service/audio/hwopus.h hle/service/bcat/backend/backend.cpp hle/service/bcat/backend/backend.h - ${BCAT_BOXCAT_ADDITIONAL_SOURCES} hle/service/bcat/bcat.cpp hle/service/bcat/bcat.h hle/service/bcat/module.cpp @@ -446,6 +447,8 @@ add_library(core STATIC hle/service/nvdrv/devices/nvhost_gpu.h hle/service/nvdrv/devices/nvhost_nvdec.cpp hle/service/nvdrv/devices/nvhost_nvdec.h + hle/service/nvdrv/devices/nvhost_nvdec_common.cpp + hle/service/nvdrv/devices/nvhost_nvdec_common.h hle/service/nvdrv/devices/nvhost_nvjpg.cpp hle/service/nvdrv/devices/nvhost_nvjpg.h hle/service/nvdrv/devices/nvhost_vic.cpp @@ -459,10 +462,14 @@ add_library(core STATIC hle/service/nvdrv/nvdrv.h hle/service/nvdrv/nvmemp.cpp hle/service/nvdrv/nvmemp.h + hle/service/nvdrv/syncpoint_manager.cpp + hle/service/nvdrv/syncpoint_manager.h hle/service/nvflinger/buffer_queue.cpp hle/service/nvflinger/buffer_queue.h hle/service/nvflinger/nvflinger.cpp hle/service/nvflinger/nvflinger.h + hle/service/olsc/olsc.cpp + hle/service/olsc/olsc.h hle/service/pcie/pcie.cpp hle/service/pcie/pcie.h hle/service/pctl/module.cpp @@ -495,7 +502,6 @@ add_library(core STATIC hle/service/sm/controller.h hle/service/sm/sm.cpp hle/service/sm/sm.h - hle/service/sockets/blocking_worker.h hle/service/sockets/bsd.cpp hle/service/sockets/bsd.h hle/service/sockets/ethc.cpp @@ -608,6 +614,13 @@ add_library(core STATIC tools/freezer.h ) +if (YUZU_ENABLE_BOXCAT) + target_sources(core PRIVATE + hle/service/bcat/backend/boxcat.cpp + hle/service/bcat/backend/boxcat.h + ) +endif() + if (MSVC) target_compile_options(core PRIVATE # 'expression' : signed/unsigned mismatch @@ -622,13 +635,29 @@ if (MSVC) /we4267 # 'context' : truncation from 'type1' to 'type2' /we4305 + # 'function' : not all control paths return a value + /we4715 + ) +else() + target_compile_options(core PRIVATE + -Werror=conversion + -Werror=ignored-qualifiers + -Werror=implicit-fallthrough + -Werror=reorder + -Werror=sign-compare + -Werror=unused-variable + + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable> + + -Wno-sign-conversion ) endif() create_target_directory_groups(core) target_link_libraries(core PUBLIC common PRIVATE audio_core video_core) -target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt::fmt nlohmann_json::nlohmann_json mbedtls opus unicorn zip) +target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt::fmt nlohmann_json::nlohmann_json mbedtls opus zip) if (YUZU_ENABLE_BOXCAT) target_compile_definitions(core PRIVATE -DYUZU_ENABLE_BOXCAT) diff --git a/src/core/arm/arm_interface.cpp b/src/core/arm/arm_interface.cpp index d2295ed90..0951e1976 100644 --- a/src/core/arm/arm_interface.cpp +++ b/src/core/arm/arm_interface.cpp @@ -147,10 +147,18 @@ std::vector<ARM_Interface::BacktraceEntry> ARM_Interface::GetBacktraceFromContex auto fp = ctx.cpu_registers[29]; auto lr = ctx.cpu_registers[30]; while (true) { - out.push_back({"", 0, lr, 0}); - if (!fp) { + out.push_back({ + .module = "", + .address = 0, + .original_address = lr, + .offset = 0, + .name = {}, + }); + + if (fp == 0) { break; } + lr = memory.Read64(fp + 8) - 4; fp = memory.Read64(fp); } diff --git a/src/core/arm/arm_interface.h b/src/core/arm/arm_interface.h index 1f24051e4..70098c526 100644 --- a/src/core/arm/arm_interface.h +++ b/src/core/arm/arm_interface.h @@ -64,15 +64,25 @@ public: /// Step CPU by one instruction virtual void Step() = 0; + /// Exits execution from a callback, the callback must rewind the stack + virtual void ExceptionalExit() = 0; + /// Clear all instruction cache virtual void ClearInstructionCache() = 0; - /// Notifies CPU emulation that the current page table has changed. - /// - /// @param new_page_table The new page table. - /// @param new_address_space_size_in_bits The new usable size of the address space in bits. - /// This can be either 32, 36, or 39 on official software. - /// + /** + * Clear instruction cache range + * @param addr Start address of the cache range to clear + * @param size Size of the cache range to clear, starting at addr + */ + virtual void InvalidateCacheRange(VAddr addr, std::size_t size) = 0; + + /** + * Notifies CPU emulation that the current page table has changed. + * @param new_page_table The new page table. + * @param new_address_space_size_in_bits The new usable size of the address space in bits. + * This can be either 32, 36, or 39 on official software. + */ virtual void PageTableChanged(Common::PageTable& new_page_table, std::size_t new_address_space_size_in_bits) = 0; diff --git a/src/core/arm/cpu_interrupt_handler.h b/src/core/arm/cpu_interrupt_handler.h index 71e582f79..c20c280f1 100644 --- a/src/core/arm/cpu_interrupt_handler.h +++ b/src/core/arm/cpu_interrupt_handler.h @@ -21,8 +21,8 @@ public: CPUInterruptHandler(const CPUInterruptHandler&) = delete; CPUInterruptHandler& operator=(const CPUInterruptHandler&) = delete; - CPUInterruptHandler(CPUInterruptHandler&&) = default; - CPUInterruptHandler& operator=(CPUInterruptHandler&&) = default; + CPUInterruptHandler(CPUInterruptHandler&&) = delete; + CPUInterruptHandler& operator=(CPUInterruptHandler&&) = delete; bool IsInterrupted() const { return is_interrupted; diff --git a/src/core/arm/dynarmic/arm_dynarmic_32.cpp b/src/core/arm/dynarmic/arm_dynarmic_32.cpp index b5f28a86e..6c4c8e9e4 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_32.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic_32.cpp @@ -7,6 +7,7 @@ #include <dynarmic/A32/a32.h> #include <dynarmic/A32/config.h> #include <dynarmic/A32/context.h> +#include "common/assert.h" #include "common/logging/log.h" #include "common/page_table.h" #include "core/arm/cpu_interrupt_handler.h" @@ -70,15 +71,8 @@ public: } void ExceptionRaised(u32 pc, Dynarmic::A32::Exception exception) override { - switch (exception) { - case Dynarmic::A32::Exception::UndefinedInstruction: - case Dynarmic::A32::Exception::UnpredictableInstruction: - break; - case Dynarmic::A32::Exception::Breakpoint: - break; - } LOG_CRITICAL(Core_ARM, "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X})", - static_cast<std::size_t>(exception), pc, MemoryReadCode(pc)); + exception, pc, MemoryReadCode(pc)); UNIMPLEMENTED(); } @@ -132,6 +126,7 @@ std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& config.page_table = reinterpret_cast<std::array<std::uint8_t*, NUM_PAGE_TABLE_ENTRIES>*>( page_table.pointers.data()); config.absolute_offset_page_table = true; + config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS; config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128; config.only_detect_misalignment_via_page_table_on_page_boundary = true; @@ -179,6 +174,9 @@ std::shared_ptr<Dynarmic::A32::Jit> ARM_Dynarmic_32::MakeJit(Common::PageTable& if (Settings::values.cpuopt_unsafe_reduce_fp_error) { config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_ReducedErrorFP; } + if (Settings::values.cpuopt_unsafe_inaccurate_nan) { + config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_InaccurateNaN; + } } return std::make_unique<Dynarmic::A32::Jit>(config); @@ -188,6 +186,10 @@ void ARM_Dynarmic_32::Run() { jit->Run(); } +void ARM_Dynarmic_32::ExceptionalExit() { + jit->ExceptionalExit(); +} + void ARM_Dynarmic_32::Step() { jit->Step(); } @@ -281,7 +283,17 @@ void ARM_Dynarmic_32::ClearInstructionCache() { jit->ClearCache(); } +void ARM_Dynarmic_32::InvalidateCacheRange(VAddr addr, std::size_t size) { + if (!jit) { + return; + } + jit->InvalidateCacheRange(static_cast<u32>(addr), size); +} + void ARM_Dynarmic_32::ClearExclusiveState() { + if (!jit) { + return; + } jit->ClearExclusiveState(); } diff --git a/src/core/arm/dynarmic/arm_dynarmic_32.h b/src/core/arm/dynarmic/arm_dynarmic_32.h index 2bab31b92..35e9ced48 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_32.h +++ b/src/core/arm/dynarmic/arm_dynarmic_32.h @@ -42,6 +42,7 @@ public: u32 GetPSTATE() const override; void SetPSTATE(u32 pstate) override; void Run() override; + void ExceptionalExit() override; void Step() override; VAddr GetTlsAddress() const override; void SetTlsAddress(VAddr address) override; @@ -58,6 +59,7 @@ public: void ClearExclusiveState() override; void ClearInstructionCache() override; + void InvalidateCacheRange(VAddr addr, std::size_t size) override; void PageTableChanged(Common::PageTable& new_page_table, std::size_t new_address_space_size_in_bits) override; diff --git a/src/core/arm/dynarmic/arm_dynarmic_64.cpp b/src/core/arm/dynarmic/arm_dynarmic_64.cpp index ce9968724..4c5ebca22 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_64.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic_64.cpp @@ -6,6 +6,7 @@ #include <memory> #include <dynarmic/A64/a64.h> #include <dynarmic/A64/config.h> +#include "common/assert.h" #include "common/logging/log.h" #include "common/page_table.h" #include "core/arm/cpu_interrupt_handler.h" @@ -13,11 +14,9 @@ #include "core/arm/dynarmic/arm_exclusive_monitor.h" #include "core/core.h" #include "core/core_timing.h" -#include "core/core_timing_util.h" -#include "core/gdbstub/gdbstub.h" #include "core/hardware_properties.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/svc.h" #include "core/memory.h" #include "core/settings.h" @@ -82,16 +81,9 @@ public: } void InterpreterFallback(u64 pc, std::size_t num_instructions) override { - LOG_INFO(Core_ARM, "Unicorn fallback @ 0x{:X} for {} instructions (instr = {:08X})", pc, - num_instructions, MemoryReadCode(pc)); - - ARM_Interface::ThreadContext64 ctx; - parent.SaveContext(ctx); - parent.inner_unicorn.LoadContext(ctx); - parent.inner_unicorn.ExecuteInstructions(num_instructions); - parent.inner_unicorn.SaveContext(ctx); - parent.LoadContext(ctx); - num_interpreted_instructions += num_instructions; + LOG_ERROR(Core_ARM, + "Unimplemented instruction @ 0x{:X} for {} instructions (instr = {:08X})", pc, + num_instructions, MemoryReadCode(pc)); } void ExceptionRaised(u64 pc, Dynarmic::A64::Exception exception) override { @@ -103,16 +95,6 @@ public: case Dynarmic::A64::Exception::Yield: return; case Dynarmic::A64::Exception::Breakpoint: - if (GDBStub::IsServerEnabled()) { - parent.jit->HaltExecution(); - parent.SetPC(pc); - Kernel::Thread* const thread = parent.system.CurrentScheduler().GetCurrentThread(); - parent.SaveContext(thread->GetContext64()); - GDBStub::Break(); - GDBStub::SendTrap(thread, 5); - return; - } - [[fallthrough]]; default: ASSERT_MSG(false, "ExceptionRaised(exception = {}, pc = {:08X}, code = {:08X})", static_cast<std::size_t>(exception), pc, MemoryReadCode(pc)); @@ -127,18 +109,17 @@ public: if (parent.uses_wall_clock) { return; } + // Divide the number of ticks by the amount of CPU cores. TODO(Subv): This yields only a // rough approximation of the amount of executed ticks in the system, it may be thrown off // if not all cores are doing a similar amount of work. Instead of doing this, we should // device a way so that timing is consistent across all cores without increasing the ticks 4 // times. - u64 amortized_ticks = - (ticks - num_interpreted_instructions) / Core::Hardware::NUM_CPU_CORES; + u64 amortized_ticks = ticks / Core::Hardware::NUM_CPU_CORES; // Always execute at least one tick. amortized_ticks = std::max<u64>(amortized_ticks, 1); parent.system.CoreTiming().AddTicks(amortized_ticks); - num_interpreted_instructions = 0; } u64 GetTicksRemaining() override { @@ -156,7 +137,6 @@ public: } ARM_Dynarmic_64& parent; - std::size_t num_interpreted_instructions = 0; u64 tpidrro_el0 = 0; u64 tpidr_el0 = 0; static constexpr u64 minimum_run_cycles = 1000U; @@ -172,6 +152,7 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& // Memory config.page_table = reinterpret_cast<void**>(page_table.pointers.data()); config.page_table_address_space_bits = address_space_bits; + config.page_table_pointer_mask_bits = Common::PageTable::ATTRIBUTE_BITS; config.silently_mirror_page_table = false; config.absolute_offset_page_table = true; config.detect_misaligned_access_via_page_table = 16 | 32 | 64 | 128; @@ -231,6 +212,9 @@ std::shared_ptr<Dynarmic::A64::Jit> ARM_Dynarmic_64::MakeJit(Common::PageTable& if (Settings::values.cpuopt_unsafe_reduce_fp_error) { config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_ReducedErrorFP; } + if (Settings::values.cpuopt_unsafe_inaccurate_nan) { + config.optimizations |= Dynarmic::OptimizationFlag::Unsafe_InaccurateNaN; + } } return std::make_shared<Dynarmic::A64::Jit>(config); @@ -240,6 +224,10 @@ void ARM_Dynarmic_64::Run() { jit->Run(); } +void ARM_Dynarmic_64::ExceptionalExit() { + jit->ExceptionalExit(); +} + void ARM_Dynarmic_64::Step() { cb->InterpreterFallback(jit->GetPC(), 1); } @@ -248,12 +236,8 @@ ARM_Dynarmic_64::ARM_Dynarmic_64(System& system, CPUInterrupts& interrupt_handle bool uses_wall_clock, ExclusiveMonitor& exclusive_monitor, std::size_t core_index) : ARM_Interface{system, interrupt_handlers, uses_wall_clock}, - cb(std::make_unique<DynarmicCallbacks64>(*this)), inner_unicorn{system, interrupt_handlers, - uses_wall_clock, - ARM_Unicorn::Arch::AArch64, - core_index}, - core_index{core_index}, exclusive_monitor{ - dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {} + cb(std::make_unique<DynarmicCallbacks64>(*this)), core_index{core_index}, + exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} {} ARM_Dynarmic_64::~ARM_Dynarmic_64() = default; @@ -342,7 +326,17 @@ void ARM_Dynarmic_64::ClearInstructionCache() { jit->ClearCache(); } +void ARM_Dynarmic_64::InvalidateCacheRange(VAddr addr, std::size_t size) { + if (!jit) { + return; + } + jit->InvalidateCacheRange(addr, size); +} + void ARM_Dynarmic_64::ClearExclusiveState() { + if (!jit) { + return; + } jit->ClearExclusiveState(); } diff --git a/src/core/arm/dynarmic/arm_dynarmic_64.h b/src/core/arm/dynarmic/arm_dynarmic_64.h index 403c55961..329b59a32 100644 --- a/src/core/arm/dynarmic/arm_dynarmic_64.h +++ b/src/core/arm/dynarmic/arm_dynarmic_64.h @@ -12,7 +12,6 @@ #include "common/hash.h" #include "core/arm/arm_interface.h" #include "core/arm/exclusive_monitor.h" -#include "core/arm/unicorn/arm_unicorn.h" namespace Core::Memory { class Memory; @@ -41,6 +40,7 @@ public: void SetPSTATE(u32 pstate) override; void Run() override; void Step() override; + void ExceptionalExit() override; VAddr GetTlsAddress() const override; void SetTlsAddress(VAddr address) override; void SetTPIDR_EL0(u64 value) override; @@ -56,6 +56,7 @@ public: void ClearExclusiveState() override; void ClearInstructionCache() override; + void InvalidateCacheRange(VAddr addr, std::size_t size) override; void PageTableChanged(Common::PageTable& new_page_table, std::size_t new_address_space_size_in_bits) override; @@ -71,7 +72,6 @@ private: std::unique_ptr<DynarmicCallbacks64> cb; JitCacheType jit_cache; std::shared_ptr<Dynarmic::A64::Jit> jit; - ARM_Unicorn inner_unicorn; std::size_t core_index; DynarmicExclusiveMonitor& exclusive_monitor; diff --git a/src/core/arm/unicorn/arm_unicorn.cpp b/src/core/arm/unicorn/arm_unicorn.cpp deleted file mode 100644 index 1df3f3ed1..000000000 --- a/src/core/arm/unicorn/arm_unicorn.cpp +++ /dev/null @@ -1,295 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <unicorn/arm64.h> -#include "common/assert.h" -#include "common/microprofile.h" -#include "core/arm/cpu_interrupt_handler.h" -#include "core/arm/unicorn/arm_unicorn.h" -#include "core/core.h" -#include "core/core_timing.h" -#include "core/hle/kernel/scheduler.h" -#include "core/hle/kernel/svc.h" -#include "core/memory.h" - -namespace Core { - -// Load Unicorn DLL once on Windows using RAII -#ifdef _MSC_VER -#include <unicorn_dynload.h> -struct LoadDll { -private: - LoadDll() { - ASSERT(uc_dyn_load(NULL, 0)); - } - ~LoadDll() { - ASSERT(uc_dyn_free()); - } - static LoadDll g_load_dll; -}; -LoadDll LoadDll::g_load_dll; -#endif - -#define CHECKED(expr) \ - do { \ - if (auto _cerr = (expr)) { \ - ASSERT_MSG(false, "Call " #expr " failed with error: {} ({})\n", _cerr, \ - uc_strerror(_cerr)); \ - } \ - } while (0) - -static void CodeHook(uc_engine* uc, uint64_t address, uint32_t size, void* user_data) { - GDBStub::BreakpointAddress bkpt = - GDBStub::GetNextBreakpointFromAddress(address, GDBStub::BreakpointType::Execute); - if (GDBStub::IsMemoryBreak() || - (bkpt.type != GDBStub::BreakpointType::None && address == bkpt.address)) { - auto core = static_cast<ARM_Unicorn*>(user_data); - core->RecordBreak(bkpt); - uc_emu_stop(uc); - } -} - -static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int size, u64 value, - void* user_data) { - auto* const system = static_cast<System*>(user_data); - - ARM_Interface::ThreadContext64 ctx{}; - system->CurrentArmInterface().SaveContext(ctx); - ASSERT_MSG(false, "Attempted to read from unmapped memory: 0x{:X}, pc=0x{:X}, lr=0x{:X}", addr, - ctx.pc, ctx.cpu_registers[30]); - - return false; -} - -ARM_Unicorn::ARM_Unicorn(System& system, CPUInterrupts& interrupt_handlers, bool uses_wall_clock, - Arch architecture, std::size_t core_index) - : ARM_Interface{system, interrupt_handlers, uses_wall_clock}, core_index{core_index} { - const auto arch = architecture == Arch::AArch32 ? UC_ARCH_ARM : UC_ARCH_ARM64; - CHECKED(uc_open(arch, UC_MODE_ARM, &uc)); - - auto fpv = 3 << 20; - CHECKED(uc_reg_write(uc, UC_ARM64_REG_CPACR_EL1, &fpv)); - - uc_hook hook{}; - CHECKED(uc_hook_add(uc, &hook, UC_HOOK_INTR, (void*)InterruptHook, this, 0, UINT64_MAX)); - CHECKED(uc_hook_add(uc, &hook, UC_HOOK_MEM_INVALID, (void*)UnmappedMemoryHook, &system, 0, - UINT64_MAX)); - if (GDBStub::IsServerEnabled()) { - CHECKED(uc_hook_add(uc, &hook, UC_HOOK_CODE, (void*)CodeHook, this, 0, UINT64_MAX)); - last_bkpt_hit = false; - } -} - -ARM_Unicorn::~ARM_Unicorn() { - CHECKED(uc_close(uc)); -} - -void ARM_Unicorn::SetPC(u64 pc) { - CHECKED(uc_reg_write(uc, UC_ARM64_REG_PC, &pc)); -} - -u64 ARM_Unicorn::GetPC() const { - u64 val{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_PC, &val)); - return val; -} - -u64 ARM_Unicorn::GetReg(int regn) const { - u64 val{}; - auto treg = UC_ARM64_REG_SP; - if (regn <= 28) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X0 + regn); - } else if (regn < 31) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X29 + regn - 29); - } - CHECKED(uc_reg_read(uc, treg, &val)); - return val; -} - -void ARM_Unicorn::SetReg(int regn, u64 val) { - auto treg = UC_ARM64_REG_SP; - if (regn <= 28) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X0 + regn); - } else if (regn < 31) { - treg = (uc_arm64_reg)(UC_ARM64_REG_X29 + regn - 29); - } - CHECKED(uc_reg_write(uc, treg, &val)); -} - -u128 ARM_Unicorn::GetVectorReg(int /*index*/) const { - UNIMPLEMENTED(); - static constexpr u128 res{}; - return res; -} - -void ARM_Unicorn::SetVectorReg(int /*index*/, u128 /*value*/) { - UNIMPLEMENTED(); -} - -u32 ARM_Unicorn::GetPSTATE() const { - u64 nzcv{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &nzcv)); - return static_cast<u32>(nzcv); -} - -void ARM_Unicorn::SetPSTATE(u32 pstate) { - u64 nzcv = pstate; - CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &nzcv)); -} - -VAddr ARM_Unicorn::GetTlsAddress() const { - u64 base{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_TPIDRRO_EL0, &base)); - return base; -} - -void ARM_Unicorn::SetTlsAddress(VAddr base) { - CHECKED(uc_reg_write(uc, UC_ARM64_REG_TPIDRRO_EL0, &base)); -} - -u64 ARM_Unicorn::GetTPIDR_EL0() const { - u64 value{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_TPIDR_EL0, &value)); - return value; -} - -void ARM_Unicorn::SetTPIDR_EL0(u64 value) { - CHECKED(uc_reg_write(uc, UC_ARM64_REG_TPIDR_EL0, &value)); -} - -void ARM_Unicorn::ChangeProcessorID(std::size_t new_core_id) { - core_index = new_core_id; -} - -void ARM_Unicorn::Run() { - if (GDBStub::IsServerEnabled()) { - ExecuteInstructions(std::max(4000000U, 0U)); - } else { - while (true) { - if (interrupt_handlers[core_index].IsInterrupted()) { - return; - } - ExecuteInstructions(10); - } - } -} - -void ARM_Unicorn::Step() { - ExecuteInstructions(1); -} - -MICROPROFILE_DEFINE(ARM_Jit_Unicorn, "ARM JIT", "Unicorn", MP_RGB(255, 64, 64)); - -void ARM_Unicorn::ExecuteInstructions(std::size_t num_instructions) { - MICROPROFILE_SCOPE(ARM_Jit_Unicorn); - - // Temporarily map the code page for Unicorn - u64 map_addr{GetPC() & ~Memory::PAGE_MASK}; - std::vector<u8> page_buffer(Memory::PAGE_SIZE); - system.Memory().ReadBlock(map_addr, page_buffer.data(), page_buffer.size()); - - CHECKED(uc_mem_map_ptr(uc, map_addr, page_buffer.size(), - UC_PROT_READ | UC_PROT_WRITE | UC_PROT_EXEC, page_buffer.data())); - CHECKED(uc_emu_start(uc, GetPC(), 1ULL << 63, 0, num_instructions)); - CHECKED(uc_mem_unmap(uc, map_addr, page_buffer.size())); - if (GDBStub::IsServerEnabled()) { - if (last_bkpt_hit && last_bkpt.type == GDBStub::BreakpointType::Execute) { - uc_reg_write(uc, UC_ARM64_REG_PC, &last_bkpt.address); - } - - Kernel::Thread* const thread = system.CurrentScheduler().GetCurrentThread(); - SaveContext(thread->GetContext64()); - if (last_bkpt_hit || GDBStub::IsMemoryBreak() || GDBStub::GetCpuStepFlag()) { - last_bkpt_hit = false; - GDBStub::Break(); - GDBStub::SendTrap(thread, 5); - } - } -} - -void ARM_Unicorn::SaveContext(ThreadContext64& ctx) { - int uregs[32]; - void* tregs[32]; - - CHECKED(uc_reg_read(uc, UC_ARM64_REG_SP, &ctx.sp)); - CHECKED(uc_reg_read(uc, UC_ARM64_REG_PC, &ctx.pc)); - CHECKED(uc_reg_read(uc, UC_ARM64_REG_NZCV, &ctx.pstate)); - - for (auto i = 0; i < 29; ++i) { - uregs[i] = UC_ARM64_REG_X0 + i; - tregs[i] = &ctx.cpu_registers[i]; - } - uregs[29] = UC_ARM64_REG_X29; - tregs[29] = (void*)&ctx.cpu_registers[29]; - uregs[30] = UC_ARM64_REG_X30; - tregs[30] = (void*)&ctx.cpu_registers[30]; - - CHECKED(uc_reg_read_batch(uc, uregs, tregs, 31)); - - for (int i = 0; i < 32; ++i) { - uregs[i] = UC_ARM64_REG_Q0 + i; - tregs[i] = &ctx.vector_registers[i]; - } - - CHECKED(uc_reg_read_batch(uc, uregs, tregs, 32)); -} - -void ARM_Unicorn::LoadContext(const ThreadContext64& ctx) { - int uregs[32]; - void* tregs[32]; - - CHECKED(uc_reg_write(uc, UC_ARM64_REG_SP, &ctx.sp)); - CHECKED(uc_reg_write(uc, UC_ARM64_REG_PC, &ctx.pc)); - CHECKED(uc_reg_write(uc, UC_ARM64_REG_NZCV, &ctx.pstate)); - - for (int i = 0; i < 29; ++i) { - uregs[i] = UC_ARM64_REG_X0 + i; - tregs[i] = (void*)&ctx.cpu_registers[i]; - } - uregs[29] = UC_ARM64_REG_X29; - tregs[29] = (void*)&ctx.cpu_registers[29]; - uregs[30] = UC_ARM64_REG_X30; - tregs[30] = (void*)&ctx.cpu_registers[30]; - - CHECKED(uc_reg_write_batch(uc, uregs, tregs, 31)); - - for (auto i = 0; i < 32; ++i) { - uregs[i] = UC_ARM64_REG_Q0 + i; - tregs[i] = (void*)&ctx.vector_registers[i]; - } - - CHECKED(uc_reg_write_batch(uc, uregs, tregs, 32)); -} - -void ARM_Unicorn::PrepareReschedule() { - CHECKED(uc_emu_stop(uc)); -} - -void ARM_Unicorn::ClearExclusiveState() {} - -void ARM_Unicorn::ClearInstructionCache() {} - -void ARM_Unicorn::RecordBreak(GDBStub::BreakpointAddress bkpt) { - last_bkpt = bkpt; - last_bkpt_hit = true; -} - -void ARM_Unicorn::InterruptHook(uc_engine* uc, u32 int_no, void* user_data) { - u32 esr{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_ESR, &esr)); - - const auto ec = esr >> 26; - const auto iss = esr & 0xFFFFFF; - - auto* const arm_instance = static_cast<ARM_Unicorn*>(user_data); - - switch (ec) { - case 0x15: // SVC - Kernel::Svc::Call(arm_instance->system, iss); - break; - } -} - -} // namespace Core diff --git a/src/core/arm/unicorn/arm_unicorn.h b/src/core/arm/unicorn/arm_unicorn.h deleted file mode 100644 index 810aff311..000000000 --- a/src/core/arm/unicorn/arm_unicorn.h +++ /dev/null @@ -1,63 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <unicorn/unicorn.h> -#include "common/common_types.h" -#include "core/arm/arm_interface.h" -#include "core/gdbstub/gdbstub.h" - -namespace Core { - -class System; - -class ARM_Unicorn final : public ARM_Interface { -public: - enum class Arch { - AArch32, // 32-bit ARM - AArch64, // 64-bit ARM - }; - - explicit ARM_Unicorn(System& system, CPUInterrupts& interrupt_handlers, bool uses_wall_clock, - Arch architecture, std::size_t core_index); - ~ARM_Unicorn() override; - - void SetPC(u64 pc) override; - u64 GetPC() const override; - u64 GetReg(int index) const override; - void SetReg(int index, u64 value) override; - u128 GetVectorReg(int index) const override; - void SetVectorReg(int index, u128 value) override; - u32 GetPSTATE() const override; - void SetPSTATE(u32 pstate) override; - VAddr GetTlsAddress() const override; - void SetTlsAddress(VAddr address) override; - void SetTPIDR_EL0(u64 value) override; - u64 GetTPIDR_EL0() const override; - void ChangeProcessorID(std::size_t new_core_id) override; - void PrepareReschedule() override; - void ClearExclusiveState() override; - void ExecuteInstructions(std::size_t num_instructions); - void Run() override; - void Step() override; - void ClearInstructionCache() override; - void PageTableChanged(Common::PageTable&, std::size_t) override {} - void RecordBreak(GDBStub::BreakpointAddress bkpt); - - void SaveContext(ThreadContext32& ctx) override {} - void SaveContext(ThreadContext64& ctx) override; - void LoadContext(const ThreadContext32& ctx) override {} - void LoadContext(const ThreadContext64& ctx) override; - -private: - static void InterruptHook(uc_engine* uc, u32 int_no, void* user_data); - - uc_engine* uc{}; - GDBStub::BreakpointAddress last_bkpt{}; - bool last_bkpt_hit = false; - std::size_t core_index; -}; - -} // namespace Core diff --git a/src/core/core.cpp b/src/core/core.cpp index 81e8cc338..1a2002dec 100644 --- a/src/core/core.cpp +++ b/src/core/core.cpp @@ -25,13 +25,12 @@ #include "core/file_sys/sdmc_factory.h" #include "core/file_sys/vfs_concat.h" #include "core/file_sys/vfs_real.h" -#include "core/gdbstub/gdbstub.h" #include "core/hardware_interrupt_manager.h" #include "core/hle/kernel/client_port.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/physical_core.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "core/hle/service/am/applets/applets.h" #include "core/hle/service/apm/controller.h" @@ -40,6 +39,7 @@ #include "core/hle/service/lm/manager.h" #include "core/hle/service/service.h" #include "core/hle/service/sm/sm.h" +#include "core/hle/service/time/time_manager.h" #include "core/loader/loader.h" #include "core/memory.h" #include "core/memory/cheat_engine.h" @@ -91,37 +91,47 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs, std::string dir_name; std::string filename; Common::SplitPath(path, &dir_name, &filename, nullptr); + if (filename == "00") { const auto dir = vfs->OpenDirectory(dir_name, FileSys::Mode::Read); std::vector<FileSys::VirtualFile> concat; - for (u8 i = 0; i < 0x10; ++i) { - auto next = dir->GetFile(fmt::format("{:02X}", i)); - if (next != nullptr) + + for (u32 i = 0; i < 0x10; ++i) { + const auto file_name = fmt::format("{:02X}", i); + auto next = dir->GetFile(file_name); + + if (next != nullptr) { concat.push_back(std::move(next)); - else { - next = dir->GetFile(fmt::format("{:02x}", i)); - if (next != nullptr) - concat.push_back(std::move(next)); - else + } else { + next = dir->GetFile(file_name); + + if (next == nullptr) { break; + } + + concat.push_back(std::move(next)); } } - if (concat.empty()) + if (concat.empty()) { return nullptr; + } - return FileSys::ConcatenatedVfsFile::MakeConcatenatedFile(concat, dir->GetName()); + return FileSys::ConcatenatedVfsFile::MakeConcatenatedFile(std::move(concat), + dir->GetName()); } - if (Common::FS::IsDirectory(path)) - return vfs->OpenFile(path + "/" + "main", FileSys::Mode::Read); + if (Common::FS::IsDirectory(path)) { + return vfs->OpenFile(path + "/main", FileSys::Mode::Read); + } return vfs->OpenFile(path, FileSys::Mode::Read); } + struct System::Impl { explicit Impl(System& system) : kernel{system}, fs_controller{system}, memory{system}, - cpu_manager{system}, reporter{system}, applet_manager{system} {} + cpu_manager{system}, reporter{system}, applet_manager{system}, time_manager{system} {} ResultStatus Run() { status = ResultStatus::Success; @@ -144,12 +154,12 @@ struct System::Impl { } ResultStatus Init(System& system, Frontend::EmuWindow& emu_window) { - LOG_DEBUG(HW_Memory, "initialized OK"); + LOG_DEBUG(Core, "initialized OK"); device_memory = std::make_unique<Core::DeviceMemory>(); is_multicore = Settings::values.use_multi_core.GetValue(); - is_async_gpu = is_multicore || Settings::values.use_asynchronous_gpu_emulation.GetValue(); + is_async_gpu = Settings::values.use_asynchronous_gpu_emulation.GetValue(); kernel.SetMulticore(is_multicore); cpu_manager.SetMulticore(is_multicore); @@ -178,17 +188,19 @@ struct System::Impl { arp_manager.ResetAll(); telemetry_session = std::make_unique<Core::TelemetrySession>(); - service_manager = std::make_shared<Service::SM::ServiceManager>(kernel); - Service::Init(service_manager, system); - GDBStub::DeferStart(); - - interrupt_manager = std::make_unique<Core::Hardware::InterruptManager>(system); gpu_core = VideoCore::CreateGPU(emu_window, system); if (!gpu_core) { return ResultStatus::ErrorVideoCore; } + service_manager = std::make_shared<Service::SM::ServiceManager>(kernel); + services = std::make_unique<Service::Services>(service_manager, system); + interrupt_manager = std::make_unique<Hardware::InterruptManager>(system); + + // Initialize time manager, which must happen after kernel is created + time_manager.Initialize(); + is_powered_on = true; exit_lock = false; @@ -202,9 +214,11 @@ struct System::Impl { return ResultStatus::Success; } - ResultStatus Load(System& system, Frontend::EmuWindow& emu_window, - const std::string& filepath) { - app_loader = Loader::GetLoader(GetGameFileFromPath(virtual_filesystem, filepath)); + ResultStatus Load(System& system, Frontend::EmuWindow& emu_window, const std::string& filepath, + std::size_t program_index) { + app_loader = Loader::GetLoader(system, GetGameFileFromPath(virtual_filesystem, filepath), + program_index); + if (!app_loader) { LOG_CRITICAL(Core, "Failed to obtain loader for {}!", filepath); return ResultStatus::ErrorGetLoader; @@ -218,12 +232,12 @@ struct System::Impl { return init_result; } - telemetry_session->AddInitialInfo(*app_loader); + telemetry_session->AddInitialInfo(*app_loader, fs_controller, *content_provider); auto main_process = Kernel::Process::Create(system, "main", Kernel::Process::ProcessType::Userland); const auto [load_result, load_parameters] = app_loader->Load(*main_process, system); if (load_result != Loader::ResultStatus::Success) { - LOG_CRITICAL(Core, "Failed to load ROM (Error {})!", static_cast<int>(load_result)); + LOG_CRITICAL(Core, "Failed to load ROM (Error {})!", load_result); Shutdown(); return static_cast<ResultStatus>(static_cast<u32>(ResultStatus::ErrorLoader) + @@ -231,6 +245,7 @@ struct System::Impl { } AddGlueRegistrationForProcess(*app_loader, *main_process); kernel.MakeCurrentProcess(main_process.get()); + kernel.InitializeCores(); // Initialize cheat engine if (cheat_engine) { @@ -252,8 +267,7 @@ struct System::Impl { u64 title_id{0}; if (app_loader->ReadProgramId(title_id) != Loader::ResultStatus::Success) { - LOG_ERROR(Core, "Failed to find title id for ROM (Error {})", - static_cast<u32>(load_result)); + LOG_ERROR(Core, "Failed to find title id for ROM (Error {})", load_result); } perf_stats = std::make_unique<PerfStats>(title_id); // Reset counters and set time origin to current frame @@ -289,19 +303,17 @@ struct System::Impl { } // Shutdown emulation session - GDBStub::Shutdown(); - Service::Shutdown(); + services.reset(); service_manager.reset(); cheat_engine.reset(); telemetry_session.reset(); - device_memory.reset(); // Close all CPU/threading state cpu_manager.Shutdown(); // Shutdown kernel and core timing - kernel.Shutdown(); core_timing.Shutdown(); + kernel.Shutdown(); // Close app loader app_loader.reset(); @@ -332,7 +344,7 @@ struct System::Impl { Service::Glue::ApplicationLaunchProperty launch{}; launch.title_id = process.GetTitleID(); - FileSys::PatchManager pm{launch.title_id}; + FileSys::PatchManager pm{launch.title_id, fs_controller, *content_provider}; launch.version = pm.GetGameVersion().value_or(0); // TODO(DarkLordZach): When FSController/Game Card Support is added, if @@ -387,10 +399,14 @@ struct System::Impl { /// Service State Service::Glue::ARPManager arp_manager; Service::LM::Manager lm_manager{reporter}; + Service::Time::TimeManager time_manager; /// Service manager std::shared_ptr<Service::SM::ServiceManager> service_manager; + /// Services + std::unique_ptr<Service::Services> services; + /// Telemetry session for this emulation session std::unique_ptr<Core::TelemetrySession> telemetry_session; @@ -406,6 +422,8 @@ struct System::Impl { bool is_multicore{}; bool is_async_gpu{}; + ExecuteProgramCallback execute_program_callback; + std::array<u64, Core::Hardware::NUM_CPU_CORES> dynarmic_ticks{}; std::array<MicroProfileToken, Core::Hardware::NUM_CPU_CORES> microprofile_dynarmic{}; }; @@ -437,8 +455,17 @@ void System::InvalidateCpuInstructionCaches() { impl->kernel.InvalidateAllInstructionCaches(); } -System::ResultStatus System::Load(Frontend::EmuWindow& emu_window, const std::string& filepath) { - return impl->Load(*this, emu_window, filepath); +void System::InvalidateCpuInstructionCacheRange(VAddr addr, std::size_t size) { + impl->kernel.InvalidateCpuInstructionCacheRange(addr, size); +} + +void System::Shutdown() { + impl->Shutdown(); +} + +System::ResultStatus System::Load(Frontend::EmuWindow& emu_window, const std::string& filepath, + std::size_t program_index) { + return impl->Load(*this, emu_window, filepath, program_index); } bool System::IsPoweredOn() const { @@ -466,11 +493,11 @@ const TelemetrySession& System::TelemetrySession() const { } ARM_Interface& System::CurrentArmInterface() { - return impl->kernel.CurrentScheduler().GetCurrentThread()->ArmInterface(); + return impl->kernel.CurrentPhysicalCore().ArmInterface(); } const ARM_Interface& System::CurrentArmInterface() const { - return impl->kernel.CurrentScheduler().GetCurrentThread()->ArmInterface(); + return impl->kernel.CurrentPhysicalCore().ArmInterface(); } std::size_t System::CurrentCoreIndex() const { @@ -479,14 +506,6 @@ std::size_t System::CurrentCoreIndex() const { return core; } -Kernel::Scheduler& System::CurrentScheduler() { - return impl->kernel.CurrentScheduler(); -} - -const Kernel::Scheduler& System::CurrentScheduler() const { - return impl->kernel.CurrentScheduler(); -} - Kernel::PhysicalCore& System::CurrentPhysicalCore() { return impl->kernel.CurrentPhysicalCore(); } @@ -495,22 +514,14 @@ const Kernel::PhysicalCore& System::CurrentPhysicalCore() const { return impl->kernel.CurrentPhysicalCore(); } -Kernel::Scheduler& System::Scheduler(std::size_t core_index) { - return impl->kernel.Scheduler(core_index); -} - -const Kernel::Scheduler& System::Scheduler(std::size_t core_index) const { - return impl->kernel.Scheduler(core_index); -} - /// Gets the global scheduler -Kernel::GlobalScheduler& System::GlobalScheduler() { - return impl->kernel.GlobalScheduler(); +Kernel::GlobalSchedulerContext& System::GlobalSchedulerContext() { + return impl->kernel.GlobalSchedulerContext(); } /// Gets the global scheduler -const Kernel::GlobalScheduler& System::GlobalScheduler() const { - return impl->kernel.GlobalScheduler(); +const Kernel::GlobalSchedulerContext& System::GlobalSchedulerContext() const { + return impl->kernel.GlobalSchedulerContext(); } Kernel::Process* System::CurrentProcess() { @@ -530,15 +541,11 @@ const Kernel::Process* System::CurrentProcess() const { } ARM_Interface& System::ArmInterface(std::size_t core_index) { - auto* thread = impl->kernel.Scheduler(core_index).GetCurrentThread(); - ASSERT(thread && !thread->IsHLEThread()); - return thread->ArmInterface(); + return impl->kernel.PhysicalCore(core_index).ArmInterface(); } const ARM_Interface& System::ArmInterface(std::size_t core_index) const { - auto* thread = impl->kernel.Scheduler(core_index).GetCurrentThread(); - ASSERT(thread && !thread->IsHLEThread()); - return thread->ArmInterface(); + return impl->kernel.PhysicalCore(core_index).ArmInterface(); } ExclusiveMonitor& System::Monitor() { @@ -625,7 +632,11 @@ const std::string& System::GetStatusDetails() const { return impl->status_details; } -Loader::AppLoader& System::GetAppLoader() const { +Loader::AppLoader& System::GetAppLoader() { + return *impl->app_loader; +} + +const Loader::AppLoader& System::GetAppLoader() const { return *impl->app_loader; } @@ -717,6 +728,14 @@ const Service::LM::Manager& System::GetLogManager() const { return impl->lm_manager; } +Service::Time::TimeManager& System::GetTimeManager() { + return impl->time_manager; +} + +const Service::Time::TimeManager& System::GetTimeManager() const { + return impl->time_manager; +} + void System::SetExitLock(bool locked) { impl->exit_lock = locked; } @@ -733,14 +752,6 @@ const System::CurrentBuildProcessID& System::GetCurrentProcessBuildID() const { return impl->build_id; } -System::ResultStatus System::Init(Frontend::EmuWindow& emu_window) { - return impl->Init(*this, emu_window); -} - -void System::Shutdown() { - impl->Shutdown(); -} - Service::SM::ServiceManager& System::ServiceManager() { return *impl->service_manager; } @@ -771,4 +782,16 @@ bool System::IsMulticore() const { return impl->is_multicore; } +void System::RegisterExecuteProgramCallback(ExecuteProgramCallback&& callback) { + impl->execute_program_callback = std::move(callback); +} + +void System::ExecuteProgram(std::size_t program_index) { + if (impl->execute_program_callback) { + impl->execute_program_callback(program_index); + } else { + LOG_CRITICAL(Core, "execute_program_callback must be initialized by the frontend"); + } +} + } // namespace Core diff --git a/src/core/core.h b/src/core/core.h index 83ded63a5..579a774e4 100644 --- a/src/core/core.h +++ b/src/core/core.h @@ -5,6 +5,7 @@ #pragma once #include <cstddef> +#include <functional> #include <memory> #include <string> #include <vector> @@ -25,11 +26,11 @@ class VfsFilesystem; } // namespace FileSys namespace Kernel { -class GlobalScheduler; +class GlobalSchedulerContext; class KernelCore; class PhysicalCore; class Process; -class Scheduler; +class KScheduler; } // namespace Kernel namespace Loader { @@ -69,6 +70,10 @@ namespace SM { class ServiceManager; } // namespace SM +namespace Time { +class TimeManager; +} // namespace Time + } // namespace Service namespace Tegra { @@ -120,7 +125,7 @@ public: * Gets the instance of the System singleton class. * @returns Reference to the instance of the System singleton class. */ - static System& GetInstance() { + [[deprecated("Use of the global system instance is deprecated")]] static System& GetInstance() { return s_instance; } @@ -140,19 +145,19 @@ public: * Run the OS and Application * This function will start emulation and run the relevant devices */ - ResultStatus Run(); + [[nodiscard]] ResultStatus Run(); /** * Pause the OS and Application * This function will pause emulation and stop the relevant devices */ - ResultStatus Pause(); + [[nodiscard]] ResultStatus Pause(); /** * Step the CPU one instruction * @return Result status, indicating whether or not the operation succeeded. */ - ResultStatus SingleStep(); + [[nodiscard]] ResultStatus SingleStep(); /** * Invalidate the CPU instruction caches @@ -161,6 +166,8 @@ public: */ void InvalidateCpuInstructionCaches(); + void InvalidateCpuInstructionCacheRange(VAddr addr, std::size_t size); + /// Shutdown the emulated system. void Shutdown(); @@ -169,22 +176,24 @@ public: * @param emu_window Reference to the host-system window used for video output and keyboard * input. * @param filepath String path to the executable application to load on the host file system. + * @param program_index Specifies the index within the container of the program to launch. * @returns ResultStatus code, indicating if the operation succeeded. */ - ResultStatus Load(Frontend::EmuWindow& emu_window, const std::string& filepath); + [[nodiscard]] ResultStatus Load(Frontend::EmuWindow& emu_window, const std::string& filepath, + std::size_t program_index = 0); /** * Indicates if the emulated system is powered on (all subsystems initialized and able to run an * application). * @returns True if the emulated system is powered on, otherwise false. */ - bool IsPoweredOn() const; + [[nodiscard]] bool IsPoweredOn() const; /// Gets a reference to the telemetry session for this emulation session. - Core::TelemetrySession& TelemetrySession(); + [[nodiscard]] Core::TelemetrySession& TelemetrySession(); /// Gets a reference to the telemetry session for this emulation session. - const Core::TelemetrySession& TelemetrySession() const; + [[nodiscard]] const Core::TelemetrySession& TelemetrySession() const; /// Prepare the core emulation for a reschedule void PrepareReschedule(); @@ -193,181 +202,166 @@ public: void PrepareReschedule(u32 core_index); /// Gets and resets core performance statistics - PerfStatsResults GetAndResetPerfStats(); + [[nodiscard]] PerfStatsResults GetAndResetPerfStats(); /// Gets an ARM interface to the CPU core that is currently running - ARM_Interface& CurrentArmInterface(); + [[nodiscard]] ARM_Interface& CurrentArmInterface(); /// Gets an ARM interface to the CPU core that is currently running - const ARM_Interface& CurrentArmInterface() const; + [[nodiscard]] const ARM_Interface& CurrentArmInterface() const; /// Gets the index of the currently running CPU core - std::size_t CurrentCoreIndex() const; - - /// Gets the scheduler for the CPU core that is currently running - Kernel::Scheduler& CurrentScheduler(); - - /// Gets the scheduler for the CPU core that is currently running - const Kernel::Scheduler& CurrentScheduler() const; + [[nodiscard]] std::size_t CurrentCoreIndex() const; /// Gets the physical core for the CPU core that is currently running - Kernel::PhysicalCore& CurrentPhysicalCore(); + [[nodiscard]] Kernel::PhysicalCore& CurrentPhysicalCore(); /// Gets the physical core for the CPU core that is currently running - const Kernel::PhysicalCore& CurrentPhysicalCore() const; + [[nodiscard]] const Kernel::PhysicalCore& CurrentPhysicalCore() const; /// Gets a reference to an ARM interface for the CPU core with the specified index - ARM_Interface& ArmInterface(std::size_t core_index); + [[nodiscard]] ARM_Interface& ArmInterface(std::size_t core_index); /// Gets a const reference to an ARM interface from the CPU core with the specified index - const ARM_Interface& ArmInterface(std::size_t core_index) const; + [[nodiscard]] const ARM_Interface& ArmInterface(std::size_t core_index) const; - CpuManager& GetCpuManager(); + /// Gets a reference to the underlying CPU manager. + [[nodiscard]] CpuManager& GetCpuManager(); - const CpuManager& GetCpuManager() const; + /// Gets a const reference to the underlying CPU manager + [[nodiscard]] const CpuManager& GetCpuManager() const; /// Gets a reference to the exclusive monitor - ExclusiveMonitor& Monitor(); + [[nodiscard]] ExclusiveMonitor& Monitor(); /// Gets a constant reference to the exclusive monitor - const ExclusiveMonitor& Monitor() const; + [[nodiscard]] const ExclusiveMonitor& Monitor() const; /// Gets a mutable reference to the system memory instance. - Core::Memory::Memory& Memory(); + [[nodiscard]] Core::Memory::Memory& Memory(); /// Gets a constant reference to the system memory instance. - const Core::Memory::Memory& Memory() const; + [[nodiscard]] const Core::Memory::Memory& Memory() const; /// Gets a mutable reference to the GPU interface - Tegra::GPU& GPU(); + [[nodiscard]] Tegra::GPU& GPU(); /// Gets an immutable reference to the GPU interface. - const Tegra::GPU& GPU() const; + [[nodiscard]] const Tegra::GPU& GPU() const; /// Gets a mutable reference to the renderer. - VideoCore::RendererBase& Renderer(); + [[nodiscard]] VideoCore::RendererBase& Renderer(); /// Gets an immutable reference to the renderer. - const VideoCore::RendererBase& Renderer() const; - - /// Gets the scheduler for the CPU core with the specified index - Kernel::Scheduler& Scheduler(std::size_t core_index); - - /// Gets the scheduler for the CPU core with the specified index - const Kernel::Scheduler& Scheduler(std::size_t core_index) const; + [[nodiscard]] const VideoCore::RendererBase& Renderer() const; /// Gets the global scheduler - Kernel::GlobalScheduler& GlobalScheduler(); + [[nodiscard]] Kernel::GlobalSchedulerContext& GlobalSchedulerContext(); /// Gets the global scheduler - const Kernel::GlobalScheduler& GlobalScheduler() const; + [[nodiscard]] const Kernel::GlobalSchedulerContext& GlobalSchedulerContext() const; /// Gets the manager for the guest device memory - Core::DeviceMemory& DeviceMemory(); + [[nodiscard]] Core::DeviceMemory& DeviceMemory(); /// Gets the manager for the guest device memory - const Core::DeviceMemory& DeviceMemory() const; + [[nodiscard]] const Core::DeviceMemory& DeviceMemory() const; /// Provides a pointer to the current process - Kernel::Process* CurrentProcess(); + [[nodiscard]] Kernel::Process* CurrentProcess(); /// Provides a constant pointer to the current process. - const Kernel::Process* CurrentProcess() const; + [[nodiscard]] const Kernel::Process* CurrentProcess() const; /// Provides a reference to the core timing instance. - Timing::CoreTiming& CoreTiming(); + [[nodiscard]] Timing::CoreTiming& CoreTiming(); /// Provides a constant reference to the core timing instance. - const Timing::CoreTiming& CoreTiming() const; + [[nodiscard]] const Timing::CoreTiming& CoreTiming() const; /// Provides a reference to the interrupt manager instance. - Core::Hardware::InterruptManager& InterruptManager(); + [[nodiscard]] Core::Hardware::InterruptManager& InterruptManager(); /// Provides a constant reference to the interrupt manager instance. - const Core::Hardware::InterruptManager& InterruptManager() const; + [[nodiscard]] const Core::Hardware::InterruptManager& InterruptManager() const; /// Provides a reference to the kernel instance. - Kernel::KernelCore& Kernel(); + [[nodiscard]] Kernel::KernelCore& Kernel(); /// Provides a constant reference to the kernel instance. - const Kernel::KernelCore& Kernel() const; + [[nodiscard]] const Kernel::KernelCore& Kernel() const; /// Provides a reference to the internal PerfStats instance. - Core::PerfStats& GetPerfStats(); + [[nodiscard]] Core::PerfStats& GetPerfStats(); /// Provides a constant reference to the internal PerfStats instance. - const Core::PerfStats& GetPerfStats() const; + [[nodiscard]] const Core::PerfStats& GetPerfStats() const; /// Provides a reference to the frame limiter; - Core::FrameLimiter& FrameLimiter(); + [[nodiscard]] Core::FrameLimiter& FrameLimiter(); /// Provides a constant referent to the frame limiter - const Core::FrameLimiter& FrameLimiter() const; + [[nodiscard]] const Core::FrameLimiter& FrameLimiter() const; /// Gets the name of the current game - Loader::ResultStatus GetGameName(std::string& out) const; + [[nodiscard]] Loader::ResultStatus GetGameName(std::string& out) const; void SetStatus(ResultStatus new_status, const char* details); - const std::string& GetStatusDetails() const; + [[nodiscard]] const std::string& GetStatusDetails() const; - Loader::AppLoader& GetAppLoader() const; + [[nodiscard]] Loader::AppLoader& GetAppLoader(); + [[nodiscard]] const Loader::AppLoader& GetAppLoader() const; - Service::SM::ServiceManager& ServiceManager(); - const Service::SM::ServiceManager& ServiceManager() const; + [[nodiscard]] Service::SM::ServiceManager& ServiceManager(); + [[nodiscard]] const Service::SM::ServiceManager& ServiceManager() const; void SetFilesystem(FileSys::VirtualFilesystem vfs); - FileSys::VirtualFilesystem GetFilesystem() const; + [[nodiscard]] FileSys::VirtualFilesystem GetFilesystem() const; void RegisterCheatList(const std::vector<Memory::CheatEntry>& list, const std::array<u8, 0x20>& build_id, VAddr main_region_begin, u64 main_region_size); void SetAppletFrontendSet(Service::AM::Applets::AppletFrontendSet&& set); - void SetDefaultAppletFrontendSet(); - Service::AM::Applets::AppletManager& GetAppletManager(); - - const Service::AM::Applets::AppletManager& GetAppletManager() const; + [[nodiscard]] Service::AM::Applets::AppletManager& GetAppletManager(); + [[nodiscard]] const Service::AM::Applets::AppletManager& GetAppletManager() const; void SetContentProvider(std::unique_ptr<FileSys::ContentProviderUnion> provider); - FileSys::ContentProvider& GetContentProvider(); - - const FileSys::ContentProvider& GetContentProvider() const; + [[nodiscard]] FileSys::ContentProvider& GetContentProvider(); + [[nodiscard]] const FileSys::ContentProvider& GetContentProvider() const; - Service::FileSystem::FileSystemController& GetFileSystemController(); - - const Service::FileSystem::FileSystemController& GetFileSystemController() const; + [[nodiscard]] Service::FileSystem::FileSystemController& GetFileSystemController(); + [[nodiscard]] const Service::FileSystem::FileSystemController& GetFileSystemController() const; void RegisterContentProvider(FileSys::ContentProviderUnionSlot slot, FileSys::ContentProvider* provider); void ClearContentProvider(FileSys::ContentProviderUnionSlot slot); - const Reporter& GetReporter() const; - - Service::Glue::ARPManager& GetARPManager(); - - const Service::Glue::ARPManager& GetARPManager() const; + [[nodiscard]] const Reporter& GetReporter() const; - Service::APM::Controller& GetAPMController(); + [[nodiscard]] Service::Glue::ARPManager& GetARPManager(); + [[nodiscard]] const Service::Glue::ARPManager& GetARPManager() const; - const Service::APM::Controller& GetAPMController() const; + [[nodiscard]] Service::APM::Controller& GetAPMController(); + [[nodiscard]] const Service::APM::Controller& GetAPMController() const; - Service::LM::Manager& GetLogManager(); + [[nodiscard]] Service::LM::Manager& GetLogManager(); + [[nodiscard]] const Service::LM::Manager& GetLogManager() const; - const Service::LM::Manager& GetLogManager() const; + [[nodiscard]] Service::Time::TimeManager& GetTimeManager(); + [[nodiscard]] const Service::Time::TimeManager& GetTimeManager() const; void SetExitLock(bool locked); - - bool GetExitLock() const; + [[nodiscard]] bool GetExitLock() const; void SetCurrentProcessBuildID(const CurrentBuildProcessID& id); - - const CurrentBuildProcessID& GetCurrentProcessBuildID() const; + [[nodiscard]] const CurrentBuildProcessID& GetCurrentProcessBuildID() const; /// Register a host thread as an emulated CPU Core. void RegisterCoreThread(std::size_t id); @@ -382,18 +376,27 @@ public: void ExitDynarmicProfile(); /// Tells if system is running on multicore. - bool IsMulticore() const; + [[nodiscard]] bool IsMulticore() const; -private: - System(); + /// Type used for the frontend to designate a callback for System to re-launch the application + /// using a specified program index. + using ExecuteProgramCallback = std::function<void(std::size_t)>; /** - * Initialize the emulated system. - * @param emu_window Reference to the host-system window used for video output and keyboard - * input. - * @return ResultStatus code, indicating if the operation succeeded. + * Registers a callback from the frontend for System to re-launch the application using a + * specified program index. + * @param callback Callback from the frontend to relaunch the application. */ - ResultStatus Init(Frontend::EmuWindow& emu_window); + void RegisterExecuteProgramCallback(ExecuteProgramCallback&& callback); + + /** + * Instructs the frontend to re-launch the application using the specified program_index. + * @param program_index Specifies the index within the application of the program to launch. + */ + void ExecuteProgram(std::size_t program_index); + +private: + System(); struct Impl; std::unique_ptr<Impl> impl; diff --git a/src/core/core_timing.h b/src/core/core_timing.h index b0b6036e4..77ff4c6fe 100644 --- a/src/core/core_timing.h +++ b/src/core/core_timing.h @@ -27,8 +27,8 @@ using TimedCallback = /// Contains the characteristics of a particular event. struct EventType { - EventType(TimedCallback&& callback, std::string&& name) - : callback{std::move(callback)}, name{std::move(name)} {} + explicit EventType(TimedCallback&& callback_, std::string&& name_) + : callback{std::move(callback_)}, name{std::move(name_)} {} /// The event's callback function. TimedCallback callback; @@ -67,8 +67,8 @@ public: void Shutdown(); /// Sets if emulation is multicore or single core, must be set before Initialize - void SetMulticore(bool is_multicore) { - this->is_multicore = is_multicore; + void SetMulticore(bool is_multicore_) { + is_multicore = is_multicore_; } /// Check if it's using host timing. diff --git a/src/core/cpu_manager.cpp b/src/core/cpu_manager.cpp index 688b99eba..373395047 100644 --- a/src/core/cpu_manager.cpp +++ b/src/core/cpu_manager.cpp @@ -4,15 +4,15 @@ #include "common/fiber.h" #include "common/microprofile.h" +#include "common/scope_exit.h" #include "common/thread.h" #include "core/arm/exclusive_monitor.h" #include "core/core.h" #include "core/core_timing.h" #include "core/cpu_manager.h" -#include "core/gdbstub/gdbstub.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/physical_core.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "video_core/gpu.h" @@ -109,28 +109,26 @@ void* CpuManager::GetStartFuncParamater() { void CpuManager::MultiCoreRunGuestThread() { auto& kernel = system.Kernel(); - { - auto& sched = kernel.CurrentScheduler(); - sched.OnThreadStart(); - } + kernel.CurrentScheduler()->OnThreadStart(); + auto* thread = kernel.CurrentScheduler()->GetCurrentThread(); + auto& host_context = thread->GetHostContext(); + host_context->SetRewindPoint(GuestRewindFunction, this); MultiCoreRunGuestLoop(); } void CpuManager::MultiCoreRunGuestLoop() { auto& kernel = system.Kernel(); - auto* thread = kernel.CurrentScheduler().GetCurrentThread(); + while (true) { auto* physical_core = &kernel.CurrentPhysicalCore(); - auto& arm_interface = thread->ArmInterface(); system.EnterDynarmicProfile(); while (!physical_core->IsInterrupted()) { - arm_interface.Run(); + physical_core->Run(); physical_core = &kernel.CurrentPhysicalCore(); } system.ExitDynarmicProfile(); - arm_interface.ClearExclusiveState(); - auto& scheduler = kernel.CurrentScheduler(); - scheduler.TryDoContextSwitch(); + physical_core->ArmInterface().ClearExclusiveState(); + kernel.CurrentScheduler()->RescheduleCurrentCore(); } } @@ -139,25 +137,21 @@ void CpuManager::MultiCoreRunIdleThread() { while (true) { auto& physical_core = kernel.CurrentPhysicalCore(); physical_core.Idle(); - auto& scheduler = kernel.CurrentScheduler(); - scheduler.TryDoContextSwitch(); + kernel.CurrentScheduler()->RescheduleCurrentCore(); } } void CpuManager::MultiCoreRunSuspendThread() { auto& kernel = system.Kernel(); - { - auto& sched = kernel.CurrentScheduler(); - sched.OnThreadStart(); - } + kernel.CurrentScheduler()->OnThreadStart(); while (true) { auto core = kernel.GetCurrentHostThreadID(); - auto& scheduler = kernel.CurrentScheduler(); + auto& scheduler = *kernel.CurrentScheduler(); Kernel::Thread* current_thread = scheduler.GetCurrentThread(); Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[core].host_context); ASSERT(scheduler.ContextSwitchPending()); ASSERT(core == kernel.GetCurrentHostThreadID()); - scheduler.TryDoContextSwitch(); + scheduler.RescheduleCurrentCore(); } } @@ -205,32 +199,31 @@ void CpuManager::MultiCorePause(bool paused) { void CpuManager::SingleCoreRunGuestThread() { auto& kernel = system.Kernel(); - { - auto& sched = kernel.CurrentScheduler(); - sched.OnThreadStart(); - } + kernel.CurrentScheduler()->OnThreadStart(); + auto* thread = kernel.CurrentScheduler()->GetCurrentThread(); + auto& host_context = thread->GetHostContext(); + host_context->SetRewindPoint(GuestRewindFunction, this); SingleCoreRunGuestLoop(); } void CpuManager::SingleCoreRunGuestLoop() { auto& kernel = system.Kernel(); - auto* thread = kernel.CurrentScheduler().GetCurrentThread(); + auto* thread = kernel.CurrentScheduler()->GetCurrentThread(); while (true) { auto* physical_core = &kernel.CurrentPhysicalCore(); - auto& arm_interface = thread->ArmInterface(); system.EnterDynarmicProfile(); if (!physical_core->IsInterrupted()) { - arm_interface.Run(); + physical_core->Run(); physical_core = &kernel.CurrentPhysicalCore(); } system.ExitDynarmicProfile(); thread->SetPhantomMode(true); system.CoreTiming().Advance(); thread->SetPhantomMode(false); - arm_interface.ClearExclusiveState(); + physical_core->ArmInterface().ClearExclusiveState(); PreemptSingleCore(); auto& scheduler = kernel.Scheduler(current_core); - scheduler.TryDoContextSwitch(); + scheduler.RescheduleCurrentCore(); } } @@ -242,51 +235,53 @@ void CpuManager::SingleCoreRunIdleThread() { system.CoreTiming().AddTicks(1000U); idle_count++; auto& scheduler = physical_core.Scheduler(); - scheduler.TryDoContextSwitch(); + scheduler.RescheduleCurrentCore(); } } void CpuManager::SingleCoreRunSuspendThread() { auto& kernel = system.Kernel(); - { - auto& sched = kernel.CurrentScheduler(); - sched.OnThreadStart(); - } + kernel.CurrentScheduler()->OnThreadStart(); while (true) { auto core = kernel.GetCurrentHostThreadID(); - auto& scheduler = kernel.CurrentScheduler(); + auto& scheduler = *kernel.CurrentScheduler(); Kernel::Thread* current_thread = scheduler.GetCurrentThread(); Common::Fiber::YieldTo(current_thread->GetHostContext(), core_data[0].host_context); ASSERT(scheduler.ContextSwitchPending()); ASSERT(core == kernel.GetCurrentHostThreadID()); - scheduler.TryDoContextSwitch(); + scheduler.RescheduleCurrentCore(); } } void CpuManager::PreemptSingleCore(bool from_running_enviroment) { - std::size_t old_core = current_core; - auto& scheduler = system.Kernel().Scheduler(old_core); - Kernel::Thread* current_thread = scheduler.GetCurrentThread(); - if (idle_count >= 4 || from_running_enviroment) { - if (!from_running_enviroment) { - system.CoreTiming().Idle(); - idle_count = 0; + { + auto& scheduler = system.Kernel().Scheduler(current_core); + Kernel::Thread* current_thread = scheduler.GetCurrentThread(); + if (idle_count >= 4 || from_running_enviroment) { + if (!from_running_enviroment) { + system.CoreTiming().Idle(); + idle_count = 0; + } + current_thread->SetPhantomMode(true); + system.CoreTiming().Advance(); + current_thread->SetPhantomMode(false); } - current_thread->SetPhantomMode(true); - system.CoreTiming().Advance(); - current_thread->SetPhantomMode(false); + current_core.store((current_core + 1) % Core::Hardware::NUM_CPU_CORES); + system.CoreTiming().ResetTicks(); + scheduler.Unload(scheduler.GetCurrentThread()); + + auto& next_scheduler = system.Kernel().Scheduler(current_core); + Common::Fiber::YieldTo(current_thread->GetHostContext(), next_scheduler.ControlContext()); } - current_core.store((current_core + 1) % Core::Hardware::NUM_CPU_CORES); - system.CoreTiming().ResetTicks(); - scheduler.Unload(); - auto& next_scheduler = system.Kernel().Scheduler(current_core); - Common::Fiber::YieldTo(current_thread->GetHostContext(), next_scheduler.ControlContext()); - /// May have changed scheduler - auto& current_scheduler = system.Kernel().Scheduler(current_core); - current_scheduler.Reload(); - auto* currrent_thread2 = current_scheduler.GetCurrentThread(); - if (!currrent_thread2->IsIdleThread()) { - idle_count = 0; + + // May have changed scheduler + { + auto& scheduler = system.Kernel().Scheduler(current_core); + scheduler.Reload(scheduler.GetCurrentThread()); + auto* currrent_thread2 = scheduler.GetCurrentThread(); + if (!currrent_thread2->IsIdleThread()) { + idle_count = 0; + } } } @@ -343,6 +338,16 @@ void CpuManager::RunThread(std::size_t core) { data.initialized = true; const bool sc_sync = !is_async_gpu && !is_multicore; bool sc_sync_first_use = sc_sync; + + // Cleanup + SCOPE_EXIT({ + data.host_context->Exit(); + data.enter_barrier.reset(); + data.exit_barrier.reset(); + data.initialized = false; + MicroProfileOnThreadExit(); + }); + /// Running while (running_mode) { data.is_running = false; @@ -351,8 +356,13 @@ void CpuManager::RunThread(std::size_t core) { system.GPU().ObtainContext(); sc_sync_first_use = false; } - auto& scheduler = system.Kernel().CurrentScheduler(); - Kernel::Thread* current_thread = scheduler.GetCurrentThread(); + + // Abort if emulation was killed before the session really starts + if (!system.IsPoweredOn()) { + return; + } + + auto current_thread = system.Kernel().CurrentScheduler()->GetCurrentThread(); data.is_running = true; Common::Fiber::YieldTo(data.host_context, current_thread->GetHostContext()); data.is_running = false; @@ -360,11 +370,6 @@ void CpuManager::RunThread(std::size_t core) { data.exit_barrier->Wait(); data.is_paused = false; } - /// Time to cleanup - data.host_context->Exit(); - data.enter_barrier.reset(); - data.exit_barrier.reset(); - data.initialized = false; } } // namespace Core diff --git a/src/core/crypto/key_manager.cpp b/src/core/crypto/key_manager.cpp index 65d246050..cebe2ce37 100644 --- a/src/core/crypto/key_manager.cpp +++ b/src/core/crypto/key_manager.cpp @@ -143,6 +143,7 @@ u64 GetSignatureTypeDataSize(SignatureType type) { return 0x3C; } UNREACHABLE(); + return 0; } u64 GetSignatureTypePaddingSize(SignatureType type) { @@ -157,6 +158,7 @@ u64 GetSignatureTypePaddingSize(SignatureType type) { return 0x40; } UNREACHABLE(); + return 0; } SignatureType Ticket::GetSignatureType() const { @@ -169,8 +171,7 @@ SignatureType Ticket::GetSignatureType() const { if (const auto* ticket = std::get_if<ECDSATicket>(&data)) { return ticket->sig_type; } - - UNREACHABLE(); + throw std::bad_variant_access{}; } TicketData& Ticket::GetData() { @@ -183,8 +184,7 @@ TicketData& Ticket::GetData() { if (auto* ticket = std::get_if<ECDSATicket>(&data)) { return ticket->data; } - - UNREACHABLE(); + throw std::bad_variant_access{}; } const TicketData& Ticket::GetData() const { @@ -197,8 +197,7 @@ const TicketData& Ticket::GetData() const { if (const auto* ticket = std::get_if<ECDSATicket>(&data)) { return ticket->data; } - - UNREACHABLE(); + throw std::bad_variant_access{}; } u64 Ticket::GetSize() const { @@ -411,7 +410,7 @@ Loader::ResultStatus DeriveSDKeys(std::array<Key256, 2>& sd_keys, KeyManager& ke // Combine sources and seed for (auto& source : sd_key_sources) { for (std::size_t i = 0; i < source.size(); ++i) { - source[i] ^= sd_seed[i & 0xF]; + source[i] = static_cast<u8>(source[i] ^ sd_seed[i & 0xF]); } } diff --git a/src/core/file_sys/card_image.cpp b/src/core/file_sys/card_image.cpp index 956da68f7..8dee5590b 100644 --- a/src/core/file_sys/card_image.cpp +++ b/src/core/file_sys/card_image.cpp @@ -29,7 +29,7 @@ constexpr std::array partition_names{ "logo", }; -XCI::XCI(VirtualFile file_) +XCI::XCI(VirtualFile file_, std::size_t program_index) : file(std::move(file_)), program_nca_status{Loader::ResultStatus::ErrorXCIMissingProgramNCA}, partitions(partition_names.size()), partitions_raw(partition_names.size()), keys{Core::Crypto::KeyManager::Instance()} { @@ -62,7 +62,8 @@ XCI::XCI(VirtualFile file_) } secure_partition = std::make_shared<NSP>( - main_hfs.GetFile(partition_names[static_cast<std::size_t>(XCIPartition::Secure)])); + main_hfs.GetFile(partition_names[static_cast<std::size_t>(XCIPartition::Secure)]), + program_index); ncas = secure_partition->GetNCAsCollapsed(); program = diff --git a/src/core/file_sys/card_image.h b/src/core/file_sys/card_image.h index 2d0a0f285..4960e90fe 100644 --- a/src/core/file_sys/card_image.h +++ b/src/core/file_sys/card_image.h @@ -78,7 +78,7 @@ enum class XCIPartition : u8 { Update, Normal, Secure, Logo }; class XCI : public ReadOnlyVfsDirectory { public: - explicit XCI(VirtualFile file); + explicit XCI(VirtualFile file, std::size_t program_index = 0); ~XCI() override; Loader::ResultStatus GetStatus() const; diff --git a/src/core/file_sys/common_funcs.h b/src/core/file_sys/common_funcs.h new file mode 100644 index 000000000..7ed97aa50 --- /dev/null +++ b/src/core/file_sys/common_funcs.h @@ -0,0 +1,56 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" + +namespace FileSys { + +constexpr u64 AOC_TITLE_ID_MASK = 0x7FF; +constexpr u64 AOC_TITLE_ID_OFFSET = 0x1000; +constexpr u64 BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000; + +/** + * Gets the base title ID from a given title ID. + * + * @param title_id The title ID. + * @returns The base title ID. + */ +[[nodiscard]] constexpr u64 GetBaseTitleID(u64 title_id) { + return title_id & BASE_TITLE_ID_MASK; +} + +/** + * Gets the base title ID with a program index offset from a given title ID. + * + * @param title_id The title ID. + * @param program_index The program index. + * @returns The base title ID with a program index offset. + */ +[[nodiscard]] constexpr u64 GetBaseTitleIDWithProgramIndex(u64 title_id, u64 program_index) { + return GetBaseTitleID(title_id) + program_index; +} + +/** + * Gets the AOC (Add-On Content) base title ID from a given title ID. + * + * @param title_id The title ID. + * @returns The AOC base title ID. + */ +[[nodiscard]] constexpr u64 GetAOCBaseTitleID(u64 title_id) { + return GetBaseTitleID(title_id) + AOC_TITLE_ID_OFFSET; +} + +/** + * Gets the AOC (Add-On Content) ID from a given AOC title ID. + * + * @param aoc_title_id The AOC title ID. + * @returns The AOC ID. + */ +[[nodiscard]] constexpr u64 GetAOCID(u64 aoc_title_id) { + return aoc_title_id & AOC_TITLE_ID_MASK; +} + +} // namespace FileSys diff --git a/src/core/file_sys/content_archive.cpp b/src/core/file_sys/content_archive.cpp index 76af47ff9..a6c0337fa 100644 --- a/src/core/file_sys/content_archive.cpp +++ b/src/core/file_sys/content_archive.cpp @@ -410,8 +410,9 @@ u8 NCA::GetCryptoRevision() const { std::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type) const { const auto master_key_id = GetCryptoRevision(); - if (!keys.HasKey(Core::Crypto::S128KeyType::KeyArea, master_key_id, header.key_index)) - return {}; + if (!keys.HasKey(Core::Crypto::S128KeyType::KeyArea, master_key_id, header.key_index)) { + return std::nullopt; + } std::vector<u8> key_area(header.key_area.begin(), header.key_area.end()); Core::Crypto::AESCipher<Core::Crypto::Key128> cipher( @@ -420,15 +421,17 @@ std::optional<Core::Crypto::Key128> NCA::GetKeyAreaKey(NCASectionCryptoType type cipher.Transcode(key_area.data(), key_area.size(), key_area.data(), Core::Crypto::Op::Decrypt); Core::Crypto::Key128 out; - if (type == NCASectionCryptoType::XTS) + if (type == NCASectionCryptoType::XTS) { std::copy(key_area.begin(), key_area.begin() + 0x10, out.begin()); - else if (type == NCASectionCryptoType::CTR || type == NCASectionCryptoType::BKTR) + } else if (type == NCASectionCryptoType::CTR || type == NCASectionCryptoType::BKTR) { std::copy(key_area.begin() + 0x20, key_area.begin() + 0x30, out.begin()); - else + } else { LOG_CRITICAL(Crypto, "Called GetKeyAreaKey on invalid NCASectionCryptoType type={:02X}", - static_cast<u8>(type)); + type); + } + u128 out_128{}; - memcpy(out_128.data(), out.data(), 16); + std::memcpy(out_128.data(), out.data(), sizeof(u128)); LOG_TRACE(Crypto, "called with crypto_rev={:02X}, kak_index={:02X}, key={:016X}{:016X}", master_key_id, header.key_index, out_128[1], out_128[0]); @@ -507,7 +510,7 @@ VirtualFile NCA::Decrypt(const NCASectionHeader& s_header, VirtualFile in, u64 s // TODO(DarkLordZach): Find a test case for XTS-encrypted NCAs default: LOG_ERROR(Crypto, "called with unhandled crypto type={:02X}", - static_cast<u8>(s_header.raw.header.crypto_type)); + s_header.raw.header.crypto_type); return nullptr; } } @@ -516,15 +519,17 @@ Loader::ResultStatus NCA::GetStatus() const { return status; } -std::vector<std::shared_ptr<VfsFile>> NCA::GetFiles() const { - if (status != Loader::ResultStatus::Success) +std::vector<VirtualFile> NCA::GetFiles() const { + if (status != Loader::ResultStatus::Success) { return {}; + } return files; } -std::vector<std::shared_ptr<VfsDirectory>> NCA::GetSubdirectories() const { - if (status != Loader::ResultStatus::Success) +std::vector<VirtualDir> NCA::GetSubdirectories() const { + if (status != Loader::ResultStatus::Success) { return {}; + } return dirs; } @@ -532,7 +537,7 @@ std::string NCA::GetName() const { return file->GetName(); } -std::shared_ptr<VfsDirectory> NCA::GetParentDirectory() const { +VirtualDir NCA::GetParentDirectory() const { return file->GetContainingDirectory(); } diff --git a/src/core/file_sys/content_archive.h b/src/core/file_sys/content_archive.h index 69292232a..e9eccdea3 100644 --- a/src/core/file_sys/content_archive.h +++ b/src/core/file_sys/content_archive.h @@ -82,7 +82,7 @@ struct NCAHeader { }; static_assert(sizeof(NCAHeader) == 0x400, "NCAHeader has incorrect size."); -inline bool IsDirectoryExeFS(const std::shared_ptr<VfsDirectory>& pfs) { +inline bool IsDirectoryExeFS(const VirtualDir& pfs) { // According to switchbrew, an exefs must only contain these two files: return pfs->GetFile("main") != nullptr && pfs->GetFile("main.npdm") != nullptr; } @@ -104,10 +104,10 @@ public: Loader::ResultStatus GetStatus() const; - std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; - std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; + std::vector<VirtualFile> GetFiles() const override; + std::vector<VirtualDir> GetSubdirectories() const override; std::string GetName() const override; - std::shared_ptr<VfsDirectory> GetParentDirectory() const override; + VirtualDir GetParentDirectory() const override; NCAContentType GetType() const; u64 GetTitleId() const; diff --git a/src/core/file_sys/fsmitm_romfsbuild.cpp b/src/core/file_sys/fsmitm_romfsbuild.cpp index 2aff2708a..c52fafb6f 100644 --- a/src/core/file_sys/fsmitm_romfsbuild.cpp +++ b/src/core/file_sys/fsmitm_romfsbuild.cpp @@ -266,8 +266,9 @@ std::multimap<u64, VirtualFile> RomFSBuildContext::Build() { cur_file->offset = file_partition_size; file_partition_size += cur_file->size; cur_file->entry_offset = entry_offset; - entry_offset += sizeof(RomFSFileEntry) + - Common::AlignUp(cur_file->path_len - cur_file->cur_path_ofs, 4); + entry_offset += + static_cast<u32>(sizeof(RomFSFileEntry) + + Common::AlignUp(cur_file->path_len - cur_file->cur_path_ofs, 4)); prev_file = cur_file; } // Assign deferred parent/sibling ownership. @@ -284,8 +285,9 @@ std::multimap<u64, VirtualFile> RomFSBuildContext::Build() { for (const auto& it : directories) { cur_dir = it.second; cur_dir->entry_offset = entry_offset; - entry_offset += sizeof(RomFSDirectoryEntry) + - Common::AlignUp(cur_dir->path_len - cur_dir->cur_path_ofs, 4); + entry_offset += + static_cast<u32>(sizeof(RomFSDirectoryEntry) + + Common::AlignUp(cur_dir->path_len - cur_dir->cur_path_ofs, 4)); } // Assign deferred parent/sibling ownership. for (auto it = directories.rbegin(); it->second != root; ++it) { diff --git a/src/core/file_sys/ips_layer.cpp b/src/core/file_sys/ips_layer.cpp index dd779310f..a6101f1c0 100644 --- a/src/core/file_sys/ips_layer.cpp +++ b/src/core/file_sys/ips_layer.cpp @@ -299,7 +299,7 @@ void IPSwitchCompiler::Parse() { patch_text->GetName(), offset, Common::HexToString(replace)); } - patch.records.insert_or_assign(offset, std::move(replace)); + patch.records.insert_or_assign(static_cast<u32>(offset), std::move(replace)); } patches.push_back(std::move(patch)); diff --git a/src/core/file_sys/nca_metadata.cpp b/src/core/file_sys/nca_metadata.cpp index 2d1476e3a..3596541b2 100644 --- a/src/core/file_sys/nca_metadata.cpp +++ b/src/core/file_sys/nca_metadata.cpp @@ -108,7 +108,7 @@ std::vector<u8> CNMT::Serialize() const { memcpy(out.data() + sizeof(CNMTHeader), &opt_header, sizeof(OptionalHeader)); } - auto offset = header.table_offset; + u64_le offset = header.table_offset; for (const auto& rec : content_records) { memcpy(out.data() + offset + sizeof(CNMTHeader), &rec, sizeof(ContentRecord)); diff --git a/src/core/file_sys/nca_patch.cpp b/src/core/file_sys/nca_patch.cpp index 5990a2fd5..a65ec6798 100644 --- a/src/core/file_sys/nca_patch.cpp +++ b/src/core/file_sys/nca_patch.cpp @@ -51,8 +51,8 @@ std::pair<std::size_t, std::size_t> SearchBucketEntry(u64 offset, const BlockTyp low = mid + 1; } } - UNREACHABLE_MSG("Offset could not be found in BKTR block."); + return {0, 0}; } } // Anonymous namespace @@ -191,7 +191,7 @@ bool BKTR::Resize(std::size_t new_size) { return false; } -std::shared_ptr<VfsDirectory> BKTR::GetContainingDirectory() const { +VirtualDir BKTR::GetContainingDirectory() const { return base_romfs->GetContainingDirectory(); } diff --git a/src/core/file_sys/nca_patch.h b/src/core/file_sys/nca_patch.h index 60c544f8e..503cf473e 100644 --- a/src/core/file_sys/nca_patch.h +++ b/src/core/file_sys/nca_patch.h @@ -106,7 +106,7 @@ public: bool Resize(std::size_t new_size) override; - std::shared_ptr<VfsDirectory> GetContainingDirectory() const override; + VirtualDir GetContainingDirectory() const override; bool IsWritable() const override; diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp index b9c09b456..7c3284df8 100644 --- a/src/core/file_sys/patch_manager.cpp +++ b/src/core/file_sys/patch_manager.cpp @@ -12,6 +12,7 @@ #include "common/logging/log.h" #include "common/string_util.h" #include "core/core.h" +#include "core/file_sys/common_funcs.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/ips_layer.h" @@ -29,8 +30,7 @@ namespace FileSys { namespace { -constexpr u64 SINGLE_BYTE_MODULUS = 0x100; -constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000; +constexpr u32 SINGLE_BYTE_MODULUS = 0x100; constexpr std::array<const char*, 14> EXEFS_FILE_NAMES{ "main", "main.npdm", "rtld", "sdk", "subsdk0", "subsdk1", "subsdk2", @@ -112,7 +112,10 @@ bool IsDirValidAndNonEmpty(const VirtualDir& dir) { } } // Anonymous namespace -PatchManager::PatchManager(u64 title_id) : title_id(title_id) {} +PatchManager::PatchManager(u64 title_id_, + const Service::FileSystem::FileSystemController& fs_controller_, + const ContentProvider& content_provider_) + : title_id{title_id_}, fs_controller{fs_controller_}, content_provider{content_provider_} {} PatchManager::~PatchManager() = default; @@ -128,34 +131,30 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const { if (Settings::values.dump_exefs) { LOG_INFO(Loader, "Dumping ExeFS for title_id={:016X}", title_id); - const auto dump_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationDumpRoot(title_id); + const auto dump_dir = fs_controller.GetModificationDumpRoot(title_id); if (dump_dir != nullptr) { const auto exefs_dir = GetOrCreateDirectoryRelative(dump_dir, "/exefs"); VfsRawCopyD(exefs, exefs_dir); } } - const auto& installed = Core::System::GetInstance().GetContentProvider(); - const auto& disabled = Settings::values.disabled_addons[title_id]; const auto update_disabled = std::find(disabled.cbegin(), disabled.cend(), "Update") != disabled.cend(); // Game Updates const auto update_tid = GetUpdateTitleID(title_id); - const auto update = installed.GetEntry(update_tid, ContentRecordType::Program); + const auto update = content_provider.GetEntry(update_tid, ContentRecordType::Program); if (!update_disabled && update != nullptr && update->GetExeFS() != nullptr && update->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { LOG_INFO(Loader, " ExeFS: Update ({}) applied successfully", - FormatTitleVersion(installed.GetEntryVersion(update_tid).value_or(0))); + FormatTitleVersion(content_provider.GetEntryVersion(update_tid).value_or(0))); exefs = update->GetExeFS(); } // LayeredExeFS - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir != nullptr && load_dir->GetSize() > 0) { auto patch_dirs = load_dir->GetSubdirectories(); std::sort( @@ -241,8 +240,7 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::st if (Settings::values.dump_nso) { LOG_INFO(Loader, "Dumping NSO for name={}, build_id={}, title_id={:016X}", name, build_id, title_id); - const auto dump_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationDumpRoot(title_id); + const auto dump_dir = fs_controller.GetModificationDumpRoot(title_id); if (dump_dir != nullptr) { const auto nso_dir = GetOrCreateDirectoryRelative(dump_dir, "/nso"); const auto file = nso_dir->CreateFile(fmt::format("{}-{}.nso", name, build_id)); @@ -254,8 +252,7 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::st LOG_INFO(Loader, "Patching NSO for name={}, build_id={}", name, build_id); - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir == nullptr) { LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id); return nso; @@ -298,8 +295,7 @@ bool PatchManager::HasNSOPatch(const BuildID& build_id_) const { LOG_INFO(Loader, "Querying NSO patch existence for build_id={}", build_id); - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir == nullptr) { LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id); return false; @@ -313,8 +309,8 @@ bool PatchManager::HasNSOPatch(const BuildID& build_id_) const { } std::vector<Core::Memory::CheatEntry> PatchManager::CreateCheatList( - const Core::System& system, const BuildID& build_id_) const { - const auto load_dir = system.GetFileSystemController().GetModificationLoadRoot(title_id); + const BuildID& build_id_) const { + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if (load_dir == nullptr) { LOG_ERROR(Loader, "Cannot load mods for invalid title_id={:016X}", title_id); return {}; @@ -347,9 +343,9 @@ std::vector<Core::Memory::CheatEntry> PatchManager::CreateCheatList( return out; } -static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type) { - const auto load_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); +static void ApplyLayeredFS(VirtualFile& romfs, u64 title_id, ContentRecordType type, + const Service::FileSystem::FileSystemController& fs_controller) { + const auto load_dir = fs_controller.GetModificationLoadRoot(title_id); if ((type != ContentRecordType::Program && type != ContentRecordType::Data) || load_dir == nullptr || load_dir->GetSize() <= 0) { return; @@ -411,19 +407,19 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content const auto log_string = fmt::format("Patching RomFS for title_id={:016X}, type={:02X}", title_id, static_cast<u8>(type)); - if (type == ContentRecordType::Program || type == ContentRecordType::Data) + if (type == ContentRecordType::Program || type == ContentRecordType::Data) { LOG_INFO(Loader, "{}", log_string); - else + } else { LOG_DEBUG(Loader, "{}", log_string); + } - if (romfs == nullptr) + if (romfs == nullptr) { return romfs; - - const auto& installed = Core::System::GetInstance().GetContentProvider(); + } // Game Updates const auto update_tid = GetUpdateTitleID(title_id); - const auto update = installed.GetEntryRaw(update_tid, type); + const auto update = content_provider.GetEntryRaw(update_tid, type); const auto& disabled = Settings::values.disabled_addons[title_id]; const auto update_disabled = @@ -434,7 +430,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content if (new_nca->GetStatus() == Loader::ResultStatus::Success && new_nca->GetRomFS() != nullptr) { LOG_INFO(Loader, " RomFS: Update ({}) applied successfully", - FormatTitleVersion(installed.GetEntryVersion(update_tid).value_or(0))); + FormatTitleVersion(content_provider.GetEntryVersion(update_tid).value_or(0))); romfs = new_nca->GetRomFS(); } } else if (!update_disabled && update_raw != nullptr) { @@ -447,7 +443,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content } // LayeredFS - ApplyLayeredFS(romfs, title_id, type); + ApplyLayeredFS(romfs, title_id, type, fs_controller); return romfs; } @@ -458,12 +454,11 @@ PatchManager::PatchVersionNames PatchManager::GetPatchVersionNames(VirtualFile u } std::map<std::string, std::string, std::less<>> out; - const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto& disabled = Settings::values.disabled_addons[title_id]; // Game Updates const auto update_tid = GetUpdateTitleID(title_id); - PatchManager update{update_tid}; + PatchManager update{update_tid, fs_controller, content_provider}; const auto metadata = update.GetControlMetadata(); const auto& nacp = metadata.first; @@ -474,8 +469,8 @@ PatchManager::PatchVersionNames PatchManager::GetPatchVersionNames(VirtualFile u if (nacp != nullptr) { out.insert_or_assign(update_label, nacp->GetVersionString()); } else { - if (installed.HasEntry(update_tid, ContentRecordType::Program)) { - const auto meta_ver = installed.GetEntryVersion(update_tid); + if (content_provider.HasEntry(update_tid, ContentRecordType::Program)) { + const auto meta_ver = content_provider.GetEntryVersion(update_tid); if (meta_ver.value_or(0) == 0) { out.insert_or_assign(update_label, ""); } else { @@ -487,8 +482,7 @@ PatchManager::PatchVersionNames PatchManager::GetPatchVersionNames(VirtualFile u } // General Mods (LayeredFS and IPS) - const auto mod_dir = - Core::System::GetInstance().GetFileSystemController().GetModificationLoadRoot(title_id); + const auto mod_dir = fs_controller.GetModificationLoadRoot(title_id); if (mod_dir != nullptr && mod_dir->GetSize() > 0) { for (const auto& mod : mod_dir->GetSubdirectories()) { std::string types; @@ -532,13 +526,15 @@ PatchManager::PatchVersionNames PatchManager::GetPatchVersionNames(VirtualFile u } // DLC - const auto dlc_entries = installed.ListEntriesFilter(TitleType::AOC, ContentRecordType::Data); + const auto dlc_entries = + content_provider.ListEntriesFilter(TitleType::AOC, ContentRecordType::Data); std::vector<ContentProviderEntry> dlc_match; dlc_match.reserve(dlc_entries.size()); std::copy_if(dlc_entries.begin(), dlc_entries.end(), std::back_inserter(dlc_match), - [this, &installed](const ContentProviderEntry& entry) { - return (entry.title_id & DLC_BASE_TITLE_ID_MASK) == title_id && - installed.GetEntry(entry)->GetStatus() == Loader::ResultStatus::Success; + [this](const ContentProviderEntry& entry) { + return GetBaseTitleID(entry.title_id) == title_id && + content_provider.GetEntry(entry)->GetStatus() == + Loader::ResultStatus::Success; }); if (!dlc_match.empty()) { // Ensure sorted so DLC IDs show in order. @@ -559,19 +555,16 @@ PatchManager::PatchVersionNames PatchManager::GetPatchVersionNames(VirtualFile u } std::optional<u32> PatchManager::GetGameVersion() const { - const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto update_tid = GetUpdateTitleID(title_id); - if (installed.HasEntry(update_tid, ContentRecordType::Program)) { - return installed.GetEntryVersion(update_tid); + if (content_provider.HasEntry(update_tid, ContentRecordType::Program)) { + return content_provider.GetEntryVersion(update_tid); } - return installed.GetEntryVersion(title_id); + return content_provider.GetEntryVersion(title_id); } PatchManager::Metadata PatchManager::GetControlMetadata() const { - const auto& installed = Core::System::GetInstance().GetContentProvider(); - - const auto base_control_nca = installed.GetEntry(title_id, ContentRecordType::Control); + const auto base_control_nca = content_provider.GetEntry(title_id, ContentRecordType::Control); if (base_control_nca == nullptr) { return {}; } diff --git a/src/core/file_sys/patch_manager.h b/src/core/file_sys/patch_manager.h index 1f28c6241..fb1853035 100644 --- a/src/core/file_sys/patch_manager.h +++ b/src/core/file_sys/patch_manager.h @@ -17,8 +17,13 @@ namespace Core { class System; } +namespace Service::FileSystem { +class FileSystemController; +} + namespace FileSys { +class ContentProvider; class NCA; class NACP; @@ -29,7 +34,9 @@ public: using Metadata = std::pair<std::unique_ptr<NACP>, VirtualFile>; using PatchVersionNames = std::map<std::string, std::string, std::less<>>; - explicit PatchManager(u64 title_id); + explicit PatchManager(u64 title_id_, + const Service::FileSystem::FileSystemController& fs_controller_, + const ContentProvider& content_provider_); ~PatchManager(); [[nodiscard]] u64 GetTitleID() const; @@ -50,7 +57,7 @@ public: // Creates a CheatList object with all [[nodiscard]] std::vector<Core::Memory::CheatEntry> CreateCheatList( - const Core::System& system, const BuildID& build_id) const; + const BuildID& build_id) const; // Currently tracked RomFS patches: // - Game Updates @@ -80,6 +87,8 @@ private: const std::string& build_id) const; u64 title_id; + const Service::FileSystem::FileSystemController& fs_controller; + const ContentProvider& content_provider; }; } // namespace FileSys diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp index da01002d5..431302f55 100644 --- a/src/core/file_sys/registered_cache.cpp +++ b/src/core/file_sys/registered_cache.cpp @@ -105,7 +105,8 @@ ContentRecordType GetCRTypeFromNCAType(NCAContentType type) { // TODO(DarkLordZach): Peek at NCA contents to differentiate Manual and Legal. return ContentRecordType::HtmlDocument; default: - UNREACHABLE_MSG("Invalid NCAContentType={:02X}", static_cast<u8>(type)); + UNREACHABLE_MSG("Invalid NCAContentType={:02X}", type); + return ContentRecordType{}; } } diff --git a/src/core/file_sys/registered_cache.h b/src/core/file_sys/registered_cache.h index 5b414b0f0..b08a1687a 100644 --- a/src/core/file_sys/registered_cache.h +++ b/src/core/file_sys/registered_cache.h @@ -67,18 +67,18 @@ public: virtual void Refresh() = 0; virtual bool HasEntry(u64 title_id, ContentRecordType type) const = 0; - virtual bool HasEntry(ContentProviderEntry entry) const; + bool HasEntry(ContentProviderEntry entry) const; virtual std::optional<u32> GetEntryVersion(u64 title_id) const = 0; virtual VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const = 0; - virtual VirtualFile GetEntryUnparsed(ContentProviderEntry entry) const; + VirtualFile GetEntryUnparsed(ContentProviderEntry entry) const; virtual VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const = 0; - virtual VirtualFile GetEntryRaw(ContentProviderEntry entry) const; + VirtualFile GetEntryRaw(ContentProviderEntry entry) const; virtual std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const = 0; - virtual std::unique_ptr<NCA> GetEntry(ContentProviderEntry entry) const; + std::unique_ptr<NCA> GetEntry(ContentProviderEntry entry) const; virtual std::vector<ContentProviderEntry> ListEntries() const; diff --git a/src/core/file_sys/romfs_factory.cpp b/src/core/file_sys/romfs_factory.cpp index e967a254e..f4e16e4be 100644 --- a/src/core/file_sys/romfs_factory.cpp +++ b/src/core/file_sys/romfs_factory.cpp @@ -7,6 +7,7 @@ #include "common/common_types.h" #include "common/logging/log.h" #include "core/file_sys/card_image.h" +#include "core/file_sys/common_funcs.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/nca_metadata.h" #include "core/file_sys/patch_manager.h" @@ -37,14 +38,37 @@ void RomFSFactory::SetPackedUpdate(VirtualFile update_raw) { } ResultVal<VirtualFile> RomFSFactory::OpenCurrentProcess(u64 current_process_title_id) const { - if (!updatable) + if (!updatable) { return MakeResult<VirtualFile>(file); + } - const PatchManager patch_manager(current_process_title_id); + const PatchManager patch_manager{current_process_title_id, filesystem_controller, + content_provider}; return MakeResult<VirtualFile>( patch_manager.PatchRomFS(file, ivfc_offset, ContentRecordType::Program, update_raw)); } +ResultVal<VirtualFile> RomFSFactory::OpenPatchedRomFS(u64 title_id, ContentRecordType type) const { + auto nca = content_provider.GetEntry(title_id, type); + + if (nca == nullptr) { + // TODO: Find the right error code to use here + return RESULT_UNKNOWN; + } + + const PatchManager patch_manager{title_id, filesystem_controller, content_provider}; + + return MakeResult<VirtualFile>( + patch_manager.PatchRomFS(nca->GetRomFS(), nca->GetBaseIVFCOffset(), type)); +} + +ResultVal<VirtualFile> RomFSFactory::OpenPatchedRomFSWithProgramIndex( + u64 title_id, u8 program_index, ContentRecordType type) const { + const auto res_title_id = GetBaseTitleIDWithProgramIndex(title_id, program_index); + + return OpenPatchedRomFS(res_title_id, type); +} + ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage, ContentRecordType type) const { const std::shared_ptr<NCA> res = GetEntry(title_id, storage, type); diff --git a/src/core/file_sys/romfs_factory.h b/src/core/file_sys/romfs_factory.h index ec704dfa8..96dd0d578 100644 --- a/src/core/file_sys/romfs_factory.h +++ b/src/core/file_sys/romfs_factory.h @@ -42,6 +42,10 @@ public: void SetPackedUpdate(VirtualFile update_raw); [[nodiscard]] ResultVal<VirtualFile> OpenCurrentProcess(u64 current_process_title_id) const; + [[nodiscard]] ResultVal<VirtualFile> OpenPatchedRomFS(u64 title_id, + ContentRecordType type) const; + [[nodiscard]] ResultVal<VirtualFile> OpenPatchedRomFSWithProgramIndex( + u64 title_id, u8 program_index, ContentRecordType type) const; [[nodiscard]] ResultVal<VirtualFile> Open(u64 title_id, StorageId storage, ContentRecordType type) const; diff --git a/src/core/file_sys/savedata_factory.cpp b/src/core/file_sys/savedata_factory.cpp index ba4efee3a..b7bfe0928 100644 --- a/src/core/file_sys/savedata_factory.cpp +++ b/src/core/file_sys/savedata_factory.cpp @@ -70,7 +70,8 @@ std::string SaveDataAttribute::DebugInfo() const { static_cast<u8>(rank), index); } -SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) { +SaveDataFactory::SaveDataFactory(Core::System& system_, VirtualDir save_directory_) + : dir{std::move(save_directory_)}, system{system_} { // Delete all temporary storages // On hardware, it is expected that temporary storage be empty at first use. dir->DeleteSubdirectoryRecursive("temp"); @@ -83,7 +84,7 @@ ResultVal<VirtualDir> SaveDataFactory::Create(SaveDataSpaceId space, PrintSaveDataAttributeWarnings(meta); const auto save_directory = - GetFullPath(space, meta.type, meta.title_id, meta.user_id, meta.save_id); + GetFullPath(system, space, meta.type, meta.title_id, meta.user_id, meta.save_id); auto out = dir->CreateDirectoryRelative(save_directory); @@ -100,7 +101,7 @@ ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, const SaveDataAttribute& meta) const { const auto save_directory = - GetFullPath(space, meta.type, meta.title_id, meta.user_id, meta.save_id); + GetFullPath(system, space, meta.type, meta.title_id, meta.user_id, meta.save_id); auto out = dir->GetDirectoryRelative(save_directory); @@ -135,13 +136,14 @@ std::string SaveDataFactory::GetSaveDataSpaceIdPath(SaveDataSpaceId space) { } } -std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType type, u64 title_id, - u128 user_id, u64 save_id) { +std::string SaveDataFactory::GetFullPath(Core::System& system, SaveDataSpaceId space, + SaveDataType type, u64 title_id, u128 user_id, + u64 save_id) { // According to switchbrew, if a save is of type SaveData and the title id field is 0, it should // be interpreted as the title id of the current process. if (type == SaveDataType::SaveData || type == SaveDataType::DeviceSaveData) { if (title_id == 0) { - title_id = Core::System::GetInstance().CurrentProcess()->GetTitleID(); + title_id = system.CurrentProcess()->GetTitleID(); } } @@ -167,7 +169,7 @@ std::string SaveDataFactory::GetFullPath(SaveDataSpaceId space, SaveDataType typ SaveDataSize SaveDataFactory::ReadSaveDataSize(SaveDataType type, u64 title_id, u128 user_id) const { - const auto path = GetFullPath(SaveDataSpaceId::NandUser, type, title_id, user_id, 0); + const auto path = GetFullPath(system, SaveDataSpaceId::NandUser, type, title_id, user_id, 0); const auto dir = GetOrCreateDirectoryRelative(this->dir, path); const auto size_file = dir->GetFile(SAVE_DATA_SIZE_FILENAME); @@ -182,7 +184,7 @@ SaveDataSize SaveDataFactory::ReadSaveDataSize(SaveDataType type, u64 title_id, void SaveDataFactory::WriteSaveDataSize(SaveDataType type, u64 title_id, u128 user_id, SaveDataSize new_value) const { - const auto path = GetFullPath(SaveDataSpaceId::NandUser, type, title_id, user_id, 0); + const auto path = GetFullPath(system, SaveDataSpaceId::NandUser, type, title_id, user_id, 0); const auto dir = GetOrCreateDirectoryRelative(this->dir, path); const auto size_file = dir->CreateFile(SAVE_DATA_SIZE_FILENAME); diff --git a/src/core/file_sys/savedata_factory.h b/src/core/file_sys/savedata_factory.h index 6625bbbd8..17f774baa 100644 --- a/src/core/file_sys/savedata_factory.h +++ b/src/core/file_sys/savedata_factory.h @@ -12,6 +12,10 @@ #include "core/file_sys/vfs.h" #include "core/hle/result.h" +namespace Core { +class System; +} + namespace FileSys { enum class SaveDataSpaceId : u8 { @@ -84,7 +88,7 @@ struct SaveDataSize { /// File system interface to the SaveData archive class SaveDataFactory { public: - explicit SaveDataFactory(VirtualDir dir); + explicit SaveDataFactory(Core::System& system_, VirtualDir save_directory_); ~SaveDataFactory(); ResultVal<VirtualDir> Create(SaveDataSpaceId space, const SaveDataAttribute& meta) const; @@ -93,8 +97,8 @@ public: VirtualDir GetSaveDataSpaceDirectory(SaveDataSpaceId space) const; static std::string GetSaveDataSpaceIdPath(SaveDataSpaceId space); - static std::string GetFullPath(SaveDataSpaceId space, SaveDataType type, u64 title_id, - u128 user_id, u64 save_id); + static std::string GetFullPath(Core::System& system, SaveDataSpaceId space, SaveDataType type, + u64 title_id, u128 user_id, u64 save_id); SaveDataSize ReadSaveDataSize(SaveDataType type, u64 title_id, u128 user_id) const; void WriteSaveDataSize(SaveDataType type, u64 title_id, u128 user_id, @@ -102,6 +106,7 @@ public: private: VirtualDir dir; + Core::System& system; }; } // namespace FileSys diff --git a/src/core/file_sys/submission_package.cpp b/src/core/file_sys/submission_package.cpp index aab957bf2..c05735ddd 100644 --- a/src/core/file_sys/submission_package.cpp +++ b/src/core/file_sys/submission_package.cpp @@ -19,41 +19,9 @@ #include "core/loader/loader.h" namespace FileSys { -namespace { -void SetTicketKeys(const std::vector<VirtualFile>& files) { - auto& keys = Core::Crypto::KeyManager::Instance(); - for (const auto& ticket_file : files) { - if (ticket_file == nullptr) { - continue; - } - - if (ticket_file->GetExtension() != "tik") { - continue; - } - - if (ticket_file->GetSize() < - Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET + sizeof(Core::Crypto::Key128)) { - continue; - } - - Core::Crypto::Key128 key{}; - ticket_file->Read(key.data(), key.size(), Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET); - - // We get the name without the extension in order to create the rights ID. - std::string name_only(ticket_file->GetName()); - name_only.erase(name_only.size() - 4); - - const auto rights_id_raw = Common::HexStringToArray<16>(name_only); - u128 rights_id; - std::memcpy(rights_id.data(), rights_id_raw.data(), sizeof(u128)); - keys.SetKey(Core::Crypto::S128KeyType::Titlekey, key, rights_id[1], rights_id[0]); - } -} -} // Anonymous namespace - -NSP::NSP(VirtualFile file_) - : file(std::move(file_)), status{Loader::ResultStatus::Success}, +NSP::NSP(VirtualFile file_, std::size_t program_index) + : file(std::move(file_)), program_index(program_index), status{Loader::ResultStatus::Success}, pfs(std::make_shared<PartitionFilesystem>(file)), keys{Core::Crypto::KeyManager::Instance()} { if (pfs->GetStatus() != Loader::ResultStatus::Success) { status = pfs->GetStatus(); @@ -178,7 +146,7 @@ std::shared_ptr<NCA> NSP::GetNCA(u64 title_id, ContentRecordType type, TitleType if (extracted) LOG_WARNING(Service_FS, "called on an NSP that is of type extracted."); - const auto title_id_iter = ncas.find(title_id); + const auto title_id_iter = ncas.find(title_id + program_index); if (title_id_iter == ncas.end()) return nullptr; @@ -232,6 +200,35 @@ VirtualDir NSP::GetParentDirectory() const { return file->GetContainingDirectory(); } +void NSP::SetTicketKeys(const std::vector<VirtualFile>& files) { + for (const auto& ticket_file : files) { + if (ticket_file == nullptr) { + continue; + } + + if (ticket_file->GetExtension() != "tik") { + continue; + } + + if (ticket_file->GetSize() < + Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET + sizeof(Core::Crypto::Key128)) { + continue; + } + + Core::Crypto::Key128 key{}; + ticket_file->Read(key.data(), key.size(), Core::Crypto::TICKET_FILE_TITLEKEY_OFFSET); + + // We get the name without the extension in order to create the rights ID. + std::string name_only(ticket_file->GetName()); + name_only.erase(name_only.size() - 4); + + const auto rights_id_raw = Common::HexStringToArray<16>(name_only); + u128 rights_id; + std::memcpy(rights_id.data(), rights_id_raw.data(), sizeof(u128)); + keys.SetKey(Core::Crypto::S128KeyType::Titlekey, key, rights_id[1], rights_id[0]); + } +} + void NSP::InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files) { exefs = pfs; @@ -286,12 +283,31 @@ void NSP::ReadNCAs(const std::vector<VirtualFile>& files) { } auto next_nca = std::make_shared<NCA>(std::move(next_file), nullptr, 0); + if (next_nca->GetType() == NCAContentType::Program) { program_status[next_nca->GetTitleId()] = next_nca->GetStatus(); } - if (next_nca->GetStatus() == Loader::ResultStatus::Success || - (next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS && - (next_nca->GetTitleId() & 0x800) != 0)) { + + if (next_nca->GetStatus() != Loader::ResultStatus::Success && + next_nca->GetStatus() != Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { + continue; + } + + // If the last 3 hexadecimal digits of the CNMT TitleID is 0x800 or is missing the + // BKTRBaseRomFS, this is an update NCA. Otherwise, this is a base NCA. + if ((cnmt.GetTitleID() & 0x800) != 0 || + next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS) { + // If the last 3 hexadecimal digits of the NCA's TitleID is between 0x1 and + // 0x7FF, this is a multi-program update NCA. Otherwise, this is a regular + // update NCA. + if ((next_nca->GetTitleId() & 0x7FF) != 0 && + (next_nca->GetTitleId() & 0x800) == 0) { + ncas[next_nca->GetTitleId()][{cnmt.GetType(), rec.type}] = + std::move(next_nca); + } else { + ncas[cnmt.GetTitleID()][{cnmt.GetType(), rec.type}] = std::move(next_nca); + } + } else { ncas[next_nca->GetTitleId()][{cnmt.GetType(), rec.type}] = std::move(next_nca); } } diff --git a/src/core/file_sys/submission_package.h b/src/core/file_sys/submission_package.h index 2db5e46b8..54581a6f3 100644 --- a/src/core/file_sys/submission_package.h +++ b/src/core/file_sys/submission_package.h @@ -27,7 +27,7 @@ enum class ContentRecordType : u8; class NSP : public ReadOnlyVfsDirectory { public: - explicit NSP(VirtualFile file); + explicit NSP(VirtualFile file, std::size_t program_index = 0); ~NSP() override; Loader::ResultStatus GetStatus() const; @@ -63,11 +63,14 @@ public: VirtualDir GetParentDirectory() const override; private: + void SetTicketKeys(const std::vector<VirtualFile>& files); void InitializeExeFSAndRomFS(const std::vector<VirtualFile>& files); void ReadNCAs(const std::vector<VirtualFile>& files); VirtualFile file; + const std::size_t program_index; + bool extracted = false; Loader::ResultStatus status; std::map<u64, Loader::ResultStatus> program_status; diff --git a/src/core/file_sys/system_archive/data/font_nintendo_extended.cpp b/src/core/file_sys/system_archive/data/font_nintendo_extended.cpp index 69d62ce8f..29ef110a6 100644 --- a/src/core/file_sys/system_archive/data/font_nintendo_extended.cpp +++ b/src/core/file_sys/system_archive/data/font_nintendo_extended.cpp @@ -6,191 +6,384 @@ namespace FileSys::SystemArchive::SharedFontData { -const std::array<unsigned char, 2932> FONT_NINTENDO_EXTENDED{{ - 0x00, 0x01, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x80, 0x00, 0x03, 0x00, 0x70, 0x44, 0x53, 0x49, 0x47, - 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x0b, 0x6c, 0x00, 0x00, 0x00, 0x08, 0x4f, 0x53, 0x2f, 0x32, - 0x33, 0x86, 0x1d, 0x9b, 0x00, 0x00, 0x01, 0x78, 0x00, 0x00, 0x00, 0x60, 0x63, 0x6d, 0x61, 0x70, - 0xc2, 0x06, 0x20, 0xde, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0xa0, 0x63, 0x76, 0x74, 0x20, - 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x04, 0x2c, 0x00, 0x00, 0x00, 0x06, 0x66, 0x70, 0x67, 0x6d, - 0x06, 0x59, 0x9c, 0x37, 0x00, 0x00, 0x02, 0xa0, 0x00, 0x00, 0x01, 0x73, 0x67, 0x61, 0x73, 0x70, - 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x0b, 0x64, 0x00, 0x00, 0x00, 0x08, 0x67, 0x6c, 0x79, 0x66, - 0x10, 0x31, 0x88, 0x00, 0x00, 0x00, 0x04, 0x34, 0x00, 0x00, 0x04, 0x64, 0x68, 0x65, 0x61, 0x64, - 0x15, 0x9d, 0xef, 0x91, 0x00, 0x00, 0x00, 0xfc, 0x00, 0x00, 0x00, 0x36, 0x68, 0x68, 0x65, 0x61, - 0x09, 0x60, 0x03, 0x71, 0x00, 0x00, 0x01, 0x34, 0x00, 0x00, 0x00, 0x24, 0x68, 0x6d, 0x74, 0x78, - 0x0d, 0x2e, 0x03, 0xa7, 0x00, 0x00, 0x01, 0xd8, 0x00, 0x00, 0x00, 0x26, 0x6c, 0x6f, 0x63, 0x61, - 0x05, 0xc0, 0x04, 0x6c, 0x00, 0x00, 0x08, 0x98, 0x00, 0x00, 0x00, 0x1e, 0x6d, 0x61, 0x78, 0x70, - 0x02, 0x1c, 0x00, 0x5f, 0x00, 0x00, 0x01, 0x58, 0x00, 0x00, 0x00, 0x20, 0x6e, 0x61, 0x6d, 0x65, - 0x7c, 0xe0, 0x84, 0x5c, 0x00, 0x00, 0x08, 0xb8, 0x00, 0x00, 0x02, 0x09, 0x70, 0x6f, 0x73, 0x74, - 0x47, 0x4e, 0x74, 0x19, 0x00, 0x00, 0x0a, 0xc4, 0x00, 0x00, 0x00, 0x9e, 0x70, 0x72, 0x65, 0x70, - 0x1c, 0xfc, 0x7d, 0x9c, 0x00, 0x00, 0x04, 0x14, 0x00, 0x00, 0x00, 0x16, 0x00, 0x01, 0x00, 0x00, - 0x00, 0x01, 0x00, 0x00, 0x7c, 0xc7, 0xb1, 0x63, 0x5f, 0x0f, 0x3c, 0xf5, 0x00, 0x1b, 0x03, 0xe8, - 0x00, 0x00, 0x00, 0x00, 0xd9, 0x44, 0x2f, 0x5d, 0x00, 0x00, 0x00, 0x00, 0xd9, 0x45, 0x7b, 0x69, - 0x00, 0x00, 0x00, 0x00, 0x03, 0xe6, 0x03, 0xe8, 0x00, 0x00, 0x00, 0x06, 0x00, 0x02, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x03, 0x84, 0xff, 0x83, 0x01, 0xf4, 0x03, 0xe8, - 0x00, 0x00, 0x00, 0x00, 0x03, 0xe6, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x5e, - 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x03, 0x74, 0x01, 0x90, 0x00, 0x05, - 0x00, 0x04, 0x00, 0xcd, 0x00, 0xcd, 0x00, 0x00, 0x01, 0x1f, 0x00, 0xcd, 0x00, 0xcd, 0x00, 0x00, - 0x03, 0xc3, 0x00, 0x66, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, +const std::array<unsigned char, 6024> FONT_NINTENDO_EXTENDED{{ + 0x00, 0x01, 0x00, 0x00, 0x00, 0x0E, 0x00, 0x80, 0x00, 0x03, 0x00, 0x60, 0x4F, 0x53, 0x2F, 0x32, + 0x34, 0x00, 0x1E, 0x26, 0x00, 0x00, 0x01, 0x68, 0x00, 0x00, 0x00, 0x60, 0x63, 0x6D, 0x61, 0x70, + 0xC1, 0xE7, 0xC8, 0xF3, 0x00, 0x00, 0x02, 0x0C, 0x00, 0x00, 0x01, 0x72, 0x63, 0x76, 0x74, 0x20, + 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x05, 0x0C, 0x00, 0x00, 0x00, 0x06, 0x66, 0x70, 0x67, 0x6D, + 0x06, 0x59, 0x9C, 0x37, 0x00, 0x00, 0x03, 0x80, 0x00, 0x00, 0x01, 0x73, 0x67, 0x61, 0x73, 0x70, + 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x17, 0x80, 0x00, 0x00, 0x00, 0x08, 0x67, 0x6C, 0x79, 0x66, + 0x50, 0x0B, 0xEA, 0xFA, 0x00, 0x00, 0x05, 0x50, 0x00, 0x00, 0x0F, 0x04, 0x68, 0x65, 0x61, 0x64, + 0x18, 0x65, 0x81, 0x09, 0x00, 0x00, 0x00, 0xEC, 0x00, 0x00, 0x00, 0x36, 0x68, 0x68, 0x65, 0x61, + 0x09, 0x88, 0x03, 0x86, 0x00, 0x00, 0x01, 0x24, 0x00, 0x00, 0x00, 0x24, 0x68, 0x6D, 0x74, 0x78, + 0x0A, 0xF0, 0x01, 0x94, 0x00, 0x00, 0x01, 0xC8, 0x00, 0x00, 0x00, 0x42, 0x6C, 0x6F, 0x63, 0x61, + 0x34, 0x80, 0x30, 0x6E, 0x00, 0x00, 0x05, 0x14, 0x00, 0x00, 0x00, 0x3A, 0x6D, 0x61, 0x78, 0x70, + 0x02, 0x2C, 0x00, 0x72, 0x00, 0x00, 0x01, 0x48, 0x00, 0x00, 0x00, 0x20, 0x6E, 0x61, 0x6D, 0x65, + 0xDB, 0xC5, 0x42, 0x4D, 0x00, 0x00, 0x14, 0x54, 0x00, 0x00, 0x01, 0xFE, 0x70, 0x6F, 0x73, 0x74, + 0xF4, 0xB4, 0xAC, 0xAB, 0x00, 0x00, 0x16, 0x54, 0x00, 0x00, 0x01, 0x2A, 0x70, 0x72, 0x65, 0x70, + 0x1C, 0xFC, 0x7D, 0x9C, 0x00, 0x00, 0x04, 0xF4, 0x00, 0x00, 0x00, 0x16, 0x00, 0x01, 0x00, 0x00, + 0x00, 0x01, 0x00, 0x00, 0xC9, 0x16, 0x5B, 0x71, 0x5F, 0x0F, 0x3C, 0xF5, 0x00, 0x0B, 0x04, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xD9, 0x44, 0x2F, 0x5D, 0x00, 0x00, 0x00, 0x00, 0xDC, 0x02, 0x0D, 0xA7, + 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x00, 0x00, 0x08, 0x00, 0x02, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x03, 0x9A, 0xFF, 0x80, 0x02, 0x00, 0x04, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x03, 0xEC, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x00, 0x01, 0x00, 0x00, 0x00, 0x1C, 0x00, 0x71, + 0x00, 0x05, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0A, 0x00, 0x00, + 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x03, 0xC4, 0x01, 0x90, 0x00, 0x05, + 0x00, 0x04, 0x00, 0xD2, 0x00, 0xD2, 0x00, 0x00, 0x01, 0x26, 0x00, 0xD2, 0x00, 0xD2, 0x00, 0x00, + 0x03, 0xDA, 0x00, 0x68, 0x02, 0x0C, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x20, 0x20, 0x20, 0x20, 0x00, 0xc0, 0x00, 0x00, 0xe0, 0xe9, 0x03, 0x84, 0xff, 0x83, - 0x01, 0xf4, 0x02, 0xee, 0x00, 0xfa, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x03, 0xe8, - 0x02, 0xbc, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x03, 0xe8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x00, 0xfa, 0x00, 0x00, 0x00, 0xfa, 0x00, 0x00, 0x03, 0xe8, 0x00, 0xeb, 0x01, 0x21, 0x00, 0xff, - 0x00, 0xff, 0x01, 0x3d, 0x01, 0x17, 0x00, 0x42, 0x00, 0x1c, 0x00, 0x3e, 0x00, 0x17, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x68, 0x00, 0x01, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x1c, 0x00, 0x03, 0x00, 0x01, 0x00, 0x00, 0x00, 0x68, 0x00, 0x06, 0x00, 0x4c, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x21, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x20, 0x20, 0x20, 0x20, 0x00, 0xC0, 0x00, 0x0D, 0xE0, 0xF0, 0x03, 0x9A, 0xFF, 0x80, + 0x02, 0x00, 0x03, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, + 0x02, 0xCD, 0x00, 0x00, 0x00, 0x20, 0x00, 0x01, 0x04, 0x00, 0x00, 0xA4, 0x00, 0x00, 0x00, 0x00, + 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, + 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, + 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, + 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, + 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x1C, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x6C, + 0x00, 0x03, 0x00, 0x01, 0x00, 0x00, 0x00, 0x1C, 0x00, 0x04, 0x00, 0x50, 0x00, 0x00, 0x00, 0x10, + 0x00, 0x10, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0D, 0x00, 0x20, 0xE0, 0xA9, 0xE0, 0xB4, + 0xE0, 0xE9, 0xE0, 0xF0, 0xFF, 0xFF, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0D, 0x00, 0x20, 0xE0, 0xA0, + 0xE0, 0xB3, 0xE0, 0xE0, 0xE0, 0xEF, 0xFF, 0xFF, 0x00, 0x01, 0xFF, 0xF5, 0xFF, 0xE3, 0x1F, 0x64, + 0x1F, 0x5B, 0x1F, 0x30, 0x1F, 0x2B, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x06, 0x00, 0x00, 0x01, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x04, 0x00, 0x38, 0x00, 0x00, 0x00, 0x0a, - 0x00, 0x08, 0x00, 0x02, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x20, 0xe0, 0xe9, 0xff, 0xff, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x20, 0xe0, 0xe0, 0xff, 0xff, 0x00, 0x01, 0xff, 0xf5, - 0xff, 0xe3, 0x1f, 0x24, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0xb8, 0x00, 0x00, 0x2c, 0x4b, 0xb8, 0x00, 0x09, 0x50, 0x58, 0xb1, 0x01, 0x01, 0x8e, 0x59, 0xb8, - 0x01, 0xff, 0x85, 0xb8, 0x00, 0x44, 0x1d, 0xb9, 0x00, 0x09, 0x00, 0x03, 0x5f, 0x5e, 0x2d, 0xb8, - 0x00, 0x01, 0x2c, 0x20, 0x20, 0x45, 0x69, 0x44, 0xb0, 0x01, 0x60, 0x2d, 0xb8, 0x00, 0x02, 0x2c, - 0xb8, 0x00, 0x01, 0x2a, 0x21, 0x2d, 0xb8, 0x00, 0x03, 0x2c, 0x20, 0x46, 0xb0, 0x03, 0x25, 0x46, - 0x52, 0x58, 0x23, 0x59, 0x20, 0x8a, 0x20, 0x8a, 0x49, 0x64, 0x8a, 0x20, 0x46, 0x20, 0x68, 0x61, - 0x64, 0xb0, 0x04, 0x25, 0x46, 0x20, 0x68, 0x61, 0x64, 0x52, 0x58, 0x23, 0x65, 0x8a, 0x59, 0x2f, - 0x20, 0xb0, 0x00, 0x53, 0x58, 0x69, 0x20, 0xb0, 0x00, 0x54, 0x58, 0x21, 0xb0, 0x40, 0x59, 0x1b, - 0x69, 0x20, 0xb0, 0x00, 0x54, 0x58, 0x21, 0xb0, 0x40, 0x65, 0x59, 0x59, 0x3a, 0x2d, 0xb8, 0x00, - 0x04, 0x2c, 0x20, 0x46, 0xb0, 0x04, 0x25, 0x46, 0x52, 0x58, 0x23, 0x8a, 0x59, 0x20, 0x46, 0x20, - 0x6a, 0x61, 0x64, 0xb0, 0x04, 0x25, 0x46, 0x20, 0x6a, 0x61, 0x64, 0x52, 0x58, 0x23, 0x8a, 0x59, - 0x2f, 0xfd, 0x2d, 0xb8, 0x00, 0x05, 0x2c, 0x4b, 0x20, 0xb0, 0x03, 0x26, 0x50, 0x58, 0x51, 0x58, - 0xb0, 0x80, 0x44, 0x1b, 0xb0, 0x40, 0x44, 0x59, 0x1b, 0x21, 0x21, 0x20, 0x45, 0xb0, 0xc0, 0x50, - 0x58, 0xb0, 0xc0, 0x44, 0x1b, 0x21, 0x59, 0x59, 0x2d, 0xb8, 0x00, 0x06, 0x2c, 0x20, 0x20, 0x45, - 0x69, 0x44, 0xb0, 0x01, 0x60, 0x20, 0x20, 0x45, 0x7d, 0x69, 0x18, 0x44, 0xb0, 0x01, 0x60, 0x2d, - 0xb8, 0x00, 0x07, 0x2c, 0xb8, 0x00, 0x06, 0x2a, 0x2d, 0xb8, 0x00, 0x08, 0x2c, 0x4b, 0x20, 0xb0, - 0x03, 0x26, 0x53, 0x58, 0xb0, 0x40, 0x1b, 0xb0, 0x00, 0x59, 0x8a, 0x8a, 0x20, 0xb0, 0x03, 0x26, - 0x53, 0x58, 0x23, 0x21, 0xb0, 0x80, 0x8a, 0x8a, 0x1b, 0x8a, 0x23, 0x59, 0x20, 0xb0, 0x03, 0x26, - 0x53, 0x58, 0x23, 0x21, 0xb8, 0x00, 0xc0, 0x8a, 0x8a, 0x1b, 0x8a, 0x23, 0x59, 0x20, 0xb0, 0x03, - 0x26, 0x53, 0x58, 0x23, 0x21, 0xb8, 0x01, 0x00, 0x8a, 0x8a, 0x1b, 0x8a, 0x23, 0x59, 0x20, 0xb0, - 0x03, 0x26, 0x53, 0x58, 0x23, 0x21, 0xb8, 0x01, 0x40, 0x8a, 0x8a, 0x1b, 0x8a, 0x23, 0x59, 0x20, - 0xb8, 0x00, 0x03, 0x26, 0x53, 0x58, 0xb0, 0x03, 0x25, 0x45, 0xb8, 0x01, 0x80, 0x50, 0x58, 0x23, - 0x21, 0xb8, 0x01, 0x80, 0x23, 0x21, 0x1b, 0xb0, 0x03, 0x25, 0x45, 0x23, 0x21, 0x23, 0x21, 0x59, - 0x1b, 0x21, 0x59, 0x44, 0x2d, 0xb8, 0x00, 0x09, 0x2c, 0x4b, 0x53, 0x58, 0x45, 0x44, 0x1b, 0x21, - 0x21, 0x59, 0x2d, 0x00, 0xb8, 0x00, 0x00, 0x2b, 0x00, 0xba, 0x00, 0x01, 0x00, 0x01, 0x00, 0x07, - 0x2b, 0xb8, 0x00, 0x00, 0x20, 0x45, 0x7d, 0x69, 0x18, 0x44, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x03, 0xe6, 0x03, 0xe8, 0x00, 0x06, - 0x00, 0x00, 0x35, 0x01, 0x33, 0x15, 0x01, 0x23, 0x35, 0x03, 0x52, 0x94, 0xfc, 0xa6, 0x8c, 0x90, - 0x03, 0x58, 0x86, 0xfc, 0xa0, 0x8e, 0x00, 0x00, 0x00, 0x02, 0x00, 0xeb, 0x00, 0xcc, 0x02, 0xfb, - 0x03, 0x1e, 0x00, 0x08, 0x00, 0x0f, 0x00, 0x00, 0x01, 0x33, 0x13, 0x23, 0x27, 0x23, 0x07, 0x23, - 0x13, 0x17, 0x07, 0x06, 0x15, 0x33, 0x27, 0x07, 0x01, 0xbc, 0x6d, 0xd2, 0x7c, 0x26, 0xcc, 0x26, - 0x7c, 0xd1, 0x35, 0x40, 0x02, 0x89, 0x45, 0x02, 0x03, 0x1e, 0xfd, 0xae, 0x77, 0x77, 0x02, 0x52, - 0x9b, 0xcc, 0x08, 0x04, 0xda, 0x02, 0x00, 0x00, 0x00, 0x03, 0x01, 0x21, 0x00, 0xcc, 0x02, 0xc5, - 0x03, 0x1e, 0x00, 0x15, 0x00, 0x1f, 0x00, 0x2b, 0x00, 0x00, 0x25, 0x11, 0x33, 0x32, 0x1e, 0x02, - 0x15, 0x14, 0x0e, 0x02, 0x07, 0x1e, 0x01, 0x15, 0x14, 0x0e, 0x02, 0x2b, 0x01, 0x13, 0x33, 0x32, - 0x36, 0x35, 0x34, 0x26, 0x2b, 0x01, 0x1d, 0x01, 0x33, 0x32, 0x3e, 0x02, 0x35, 0x34, 0x26, 0x2b, - 0x01, 0x15, 0x01, 0x21, 0xea, 0x25, 0x3f, 0x2e, 0x1a, 0x0e, 0x15, 0x1b, 0x0e, 0x2d, 0x2d, 0x1a, - 0x2e, 0x3f, 0x25, 0xf8, 0x76, 0x62, 0x20, 0x2a, 0x28, 0x22, 0x62, 0x76, 0x10, 0x18, 0x11, 0x09, - 0x22, 0x22, 0x74, 0xcc, 0x02, 0x52, 0x18, 0x2b, 0x3c, 0x24, 0x1d, 0x1f, 0x17, 0x17, 0x14, 0x0f, - 0x48, 0x2f, 0x24, 0x3f, 0x2e, 0x1a, 0x01, 0x5b, 0x29, 0x20, 0x20, 0x2b, 0x94, 0xf8, 0x0e, 0x16, - 0x1c, 0x0e, 0x1f, 0x31, 0x9e, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0xff, 0x00, 0xcc, 0x02, 0xe7, - 0x03, 0x1e, 0x00, 0x0c, 0x00, 0x00, 0x01, 0x33, 0x17, 0x37, 0x33, 0x03, 0x13, 0x23, 0x27, 0x07, - 0x23, 0x13, 0x03, 0x01, 0x04, 0x86, 0x69, 0x69, 0x86, 0xa3, 0xa8, 0x88, 0x6c, 0x6c, 0x88, 0xa8, - 0xa3, 0x03, 0x1e, 0xcb, 0xcb, 0xfe, 0xda, 0xfe, 0xd4, 0xcf, 0xcf, 0x01, 0x2c, 0x01, 0x26, 0x00, - 0x00, 0x01, 0x00, 0xff, 0x00, 0xcc, 0x02, 0xe7, 0x03, 0x1e, 0x00, 0x0f, 0x00, 0x00, 0x01, 0x03, - 0x33, 0x17, 0x32, 0x15, 0x1e, 0x01, 0x15, 0x1b, 0x01, 0x33, 0x03, 0x15, 0x23, 0x35, 0x01, 0xb8, - 0xb9, 0x7e, 0x01, 0x01, 0x01, 0x03, 0x70, 0x75, 0x7f, 0xb9, 0x76, 0x01, 0xa3, 0x01, 0x7b, 0x01, - 0x01, 0x01, 0x05, 0x02, 0xff, 0x00, 0x01, 0x0a, 0xfe, 0x85, 0xd7, 0xd7, 0x00, 0x01, 0x01, 0x3d, - 0x00, 0xcc, 0x02, 0xa9, 0x03, 0x1e, 0x00, 0x06, 0x00, 0x00, 0x25, 0x11, 0x33, 0x11, 0x33, 0x15, - 0x21, 0x01, 0x3d, 0x75, 0xf7, 0xfe, 0x94, 0xcc, 0x02, 0x52, 0xfe, 0x10, 0x62, 0x00, 0x00, 0x00, - 0x00, 0x02, 0x01, 0x17, 0x00, 0xbc, 0x02, 0xcf, 0x03, 0x0e, 0x00, 0x15, 0x00, 0x21, 0x00, 0x00, - 0x25, 0x11, 0x33, 0x32, 0x1e, 0x02, 0x1d, 0x01, 0x0e, 0x03, 0x1d, 0x01, 0x17, 0x15, 0x23, 0x27, - 0x23, 0x15, 0x23, 0x13, 0x33, 0x32, 0x3e, 0x02, 0x35, 0x34, 0x26, 0x2b, 0x01, 0x15, 0x01, 0x17, - 0xf4, 0x27, 0x40, 0x2e, 0x19, 0x01, 0x1f, 0x24, 0x1e, 0x78, 0x7d, 0x6a, 0x5c, 0x75, 0x76, 0x72, - 0x12, 0x19, 0x11, 0x08, 0x26, 0x26, 0x6a, 0xbc, 0x02, 0x52, 0x1d, 0x31, 0x42, 0x25, 0x16, 0x18, - 0x32, 0x2a, 0x1b, 0x02, 0x01, 0xef, 0x06, 0xd7, 0xd7, 0x01, 0x3f, 0x10, 0x1a, 0x1e, 0x0f, 0x23, - 0x36, 0xb0, 0x00, 0x00, 0x00, 0x02, 0x00, 0x42, 0x00, 0xbc, 0x03, 0xa4, 0x03, 0x0e, 0x00, 0x0a, - 0x00, 0x11, 0x00, 0x00, 0x13, 0x35, 0x21, 0x15, 0x01, 0x21, 0x15, 0x21, 0x35, 0x01, 0x21, 0x01, - 0x11, 0x33, 0x11, 0x33, 0x15, 0x21, 0x42, 0x01, 0xa7, 0xfe, 0xeb, 0x01, 0x1b, 0xfe, 0x53, 0x01, - 0x15, 0xfe, 0xeb, 0x01, 0xf7, 0x75, 0xf6, 0xfe, 0x95, 0x02, 0xac, 0x62, 0x45, 0xfe, 0x55, 0x62, - 0x47, 0x01, 0xa9, 0xfe, 0x10, 0x02, 0x52, 0xfe, 0x10, 0x62, 0x00, 0x00, 0x00, 0x03, 0x00, 0x1c, - 0x00, 0xbc, 0x03, 0xca, 0x03, 0x0e, 0x00, 0x0a, 0x00, 0x21, 0x00, 0x2f, 0x00, 0x00, 0x13, 0x35, - 0x21, 0x15, 0x01, 0x21, 0x15, 0x21, 0x35, 0x01, 0x21, 0x01, 0x11, 0x33, 0x32, 0x1e, 0x02, 0x15, - 0x14, 0x06, 0x07, 0x0e, 0x03, 0x15, 0x17, 0x15, 0x23, 0x27, 0x23, 0x15, 0x23, 0x13, 0x33, 0x32, - 0x3e, 0x02, 0x35, 0x34, 0x2e, 0x02, 0x2b, 0x01, 0x15, 0x1c, 0x01, 0xa7, 0xfe, 0xeb, 0x01, 0x1b, - 0xfe, 0x53, 0x01, 0x15, 0xfe, 0xeb, 0x01, 0xf7, 0xf3, 0x27, 0x41, 0x2d, 0x19, 0x1c, 0x20, 0x01, - 0x0d, 0x0e, 0x0a, 0x78, 0x7d, 0x69, 0x5c, 0x75, 0x76, 0x71, 0x11, 0x1a, 0x12, 0x09, 0x0a, 0x14, - 0x1d, 0x13, 0x69, 0x02, 0xac, 0x62, 0x45, 0xfe, 0x55, 0x62, 0x47, 0x01, 0xa9, 0xfe, 0x10, 0x02, - 0x52, 0x1d, 0x31, 0x42, 0x25, 0x2b, 0x44, 0x1d, 0x01, 0x08, 0x09, 0x07, 0x01, 0xf1, 0x06, 0xd7, - 0xd7, 0x01, 0x3f, 0x11, 0x19, 0x1f, 0x0e, 0x11, 0x20, 0x19, 0x0f, 0xb0, 0x00, 0x02, 0x00, 0x3e, - 0x00, 0xb3, 0x03, 0xa8, 0x03, 0x17, 0x00, 0x3a, 0x00, 0x41, 0x00, 0x00, 0x13, 0x34, 0x3e, 0x02, - 0x33, 0x32, 0x1e, 0x02, 0x15, 0x23, 0x27, 0x34, 0x27, 0x2e, 0x01, 0x23, 0x22, 0x0e, 0x02, 0x15, - 0x14, 0x16, 0x15, 0x1e, 0x05, 0x15, 0x14, 0x0e, 0x02, 0x23, 0x22, 0x2e, 0x02, 0x35, 0x33, 0x1e, - 0x01, 0x33, 0x32, 0x3e, 0x02, 0x35, 0x34, 0x2e, 0x04, 0x35, 0x01, 0x11, 0x33, 0x11, 0x33, 0x15, - 0x21, 0x50, 0x24, 0x3b, 0x4a, 0x27, 0x28, 0x4b, 0x39, 0x22, 0x73, 0x01, 0x01, 0x08, 0x2b, 0x29, - 0x10, 0x20, 0x19, 0x0f, 0x01, 0x0b, 0x35, 0x41, 0x46, 0x3b, 0x25, 0x23, 0x3a, 0x4b, 0x27, 0x2b, - 0x50, 0x3f, 0x26, 0x74, 0x05, 0x34, 0x33, 0x10, 0x20, 0x1a, 0x11, 0x2c, 0x42, 0x4d, 0x42, 0x2c, - 0x01, 0xef, 0x73, 0xf6, 0xfe, 0x97, 0x02, 0x70, 0x2a, 0x3f, 0x2a, 0x14, 0x18, 0x2e, 0x44, 0x2c, - 0x02, 0x03, 0x01, 0x27, 0x27, 0x07, 0x10, 0x1a, 0x12, 0x02, 0x0b, 0x02, 0x1f, 0x22, 0x19, 0x17, - 0x27, 0x3f, 0x34, 0x2c, 0x3e, 0x28, 0x13, 0x1a, 0x32, 0x48, 0x2e, 0x30, 0x30, 0x06, 0x0f, 0x1a, - 0x13, 0x21, 0x27, 0x1e, 0x1b, 0x29, 0x3e, 0x31, 0xfe, 0x4c, 0x02, 0x53, 0xfe, 0x10, 0x63, 0x00, - 0x00, 0x03, 0x00, 0x17, 0x00, 0xb3, 0x03, 0xce, 0x03, 0x17, 0x00, 0x38, 0x00, 0x4f, 0x00, 0x5d, - 0x00, 0x00, 0x13, 0x34, 0x3e, 0x02, 0x33, 0x32, 0x1e, 0x02, 0x15, 0x23, 0x27, 0x34, 0x23, 0x2e, - 0x01, 0x23, 0x22, 0x0e, 0x02, 0x15, 0x14, 0x1e, 0x04, 0x15, 0x14, 0x0e, 0x02, 0x23, 0x22, 0x2e, - 0x02, 0x35, 0x33, 0x1e, 0x01, 0x33, 0x32, 0x3e, 0x02, 0x35, 0x34, 0x26, 0x27, 0x2e, 0x03, 0x35, - 0x01, 0x11, 0x33, 0x32, 0x1e, 0x02, 0x15, 0x14, 0x06, 0x07, 0x30, 0x0e, 0x02, 0x31, 0x17, 0x15, - 0x23, 0x27, 0x23, 0x15, 0x23, 0x13, 0x33, 0x32, 0x3e, 0x02, 0x35, 0x34, 0x2e, 0x02, 0x2b, 0x01, - 0x15, 0x2a, 0x24, 0x3a, 0x4a, 0x26, 0x29, 0x4b, 0x39, 0x23, 0x73, 0x01, 0x01, 0x08, 0x2a, 0x2a, - 0x10, 0x1f, 0x1a, 0x10, 0x2c, 0x42, 0x4d, 0x42, 0x2c, 0x23, 0x39, 0x4b, 0x27, 0x2b, 0x51, 0x3f, - 0x27, 0x75, 0x05, 0x34, 0x33, 0x10, 0x20, 0x1a, 0x10, 0x1f, 0x1c, 0x25, 0x53, 0x47, 0x2e, 0x01, - 0xed, 0xf3, 0x27, 0x41, 0x2d, 0x19, 0x1c, 0x20, 0x0c, 0x0e, 0x0c, 0x78, 0x7d, 0x68, 0x5d, 0x75, - 0x76, 0x71, 0x11, 0x1a, 0x12, 0x09, 0x0a, 0x14, 0x1d, 0x13, 0x69, 0x02, 0x71, 0x2a, 0x3e, 0x2a, - 0x14, 0x18, 0x2e, 0x44, 0x2c, 0x02, 0x02, 0x27, 0x29, 0x07, 0x11, 0x1a, 0x12, 0x1d, 0x24, 0x1c, - 0x1d, 0x2b, 0x40, 0x32, 0x2c, 0x3f, 0x29, 0x13, 0x1a, 0x31, 0x49, 0x2e, 0x30, 0x30, 0x06, 0x0f, - 0x19, 0x13, 0x1e, 0x22, 0x0b, 0x0e, 0x20, 0x2f, 0x43, 0x30, 0xfe, 0x4b, 0x02, 0x52, 0x1d, 0x32, - 0x42, 0x25, 0x2c, 0x42, 0x1d, 0x08, 0x0a, 0x08, 0xf1, 0x06, 0xd7, 0xd7, 0x01, 0x3f, 0x11, 0x19, - 0x1f, 0x0e, 0x11, 0x20, 0x19, 0x0f, 0xb0, 0x00, 0x00, 0x00, 0x00, 0x12, 0x00, 0x12, 0x00, 0x12, - 0x00, 0x12, 0x00, 0x32, 0x00, 0x72, 0x00, 0x8e, 0x00, 0xac, 0x00, 0xbe, 0x00, 0xf0, 0x01, 0x14, - 0x01, 0x5c, 0x01, 0xb6, 0x02, 0x32, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0d, 0x00, 0xa2, 0x00, 0x01, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x10, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x02, 0x00, 0x07, 0x00, 0x10, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x2f, - 0x00, 0x17, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x12, 0x00, 0x46, 0x00, 0x01, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x00, 0x0d, 0x00, 0x58, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x06, 0x00, 0x12, 0x00, 0x65, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, 0x00, 0x01, 0x00, 0x20, - 0x00, 0x77, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, 0x00, 0x02, 0x00, 0x0e, 0x00, 0x97, 0x00, 0x03, - 0x00, 0x01, 0x04, 0x09, 0x00, 0x03, 0x00, 0x5e, 0x00, 0xa5, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, - 0x00, 0x04, 0x00, 0x24, 0x01, 0x03, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, 0x00, 0x05, 0x00, 0x1a, - 0x01, 0x27, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, 0x00, 0x06, 0x00, 0x24, 0x01, 0x41, 0x00, 0x03, - 0x00, 0x01, 0x04, 0x09, 0x00, 0x11, 0x00, 0x02, 0x01, 0x65, 0x59, 0x75, 0x7a, 0x75, 0x4f, 0x53, - 0x53, 0x45, 0x78, 0x74, 0x65, 0x6e, 0x73, 0x69, 0x6f, 0x6e, 0x52, 0x65, 0x67, 0x75, 0x6c, 0x61, - 0x72, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x20, 0x31, 0x2e, 0x30, 0x30, 0x30, 0x3b, 0x3b, - 0x59, 0x75, 0x7a, 0x75, 0x4f, 0x53, 0x53, 0x45, 0x78, 0x74, 0x65, 0x6e, 0x73, 0x69, 0x6f, 0x6e, - 0x2d, 0x52, 0x3b, 0x32, 0x30, 0x31, 0x39, 0x3b, 0x46, 0x4c, 0x56, 0x49, 0x2d, 0x36, 0x31, 0x34, - 0x59, 0x75, 0x7a, 0x75, 0x4f, 0x53, 0x53, 0x45, 0x78, 0x74, 0x65, 0x6e, 0x73, 0x69, 0x6f, 0x6e, - 0x20, 0x52, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x20, 0x31, 0x2e, 0x30, 0x30, 0x30, 0x59, - 0x75, 0x7a, 0x75, 0x4f, 0x53, 0x53, 0x45, 0x78, 0x74, 0x65, 0x6e, 0x73, 0x69, 0x6f, 0x6e, 0x2d, - 0x52, 0x00, 0x59, 0x00, 0x75, 0x00, 0x7a, 0x00, 0x75, 0x00, 0x4f, 0x00, 0x53, 0x00, 0x53, 0x00, - 0x45, 0x00, 0x78, 0x00, 0x74, 0x00, 0x65, 0x00, 0x6e, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6f, 0x00, - 0x6e, 0x00, 0x52, 0x00, 0x65, 0x00, 0x67, 0x00, 0x75, 0x00, 0x6c, 0x00, 0x61, 0x00, 0x72, 0x00, - 0x56, 0x00, 0x65, 0x00, 0x72, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6f, 0x00, 0x6e, 0x00, 0x20, 0x00, - 0x31, 0x00, 0x2e, 0x00, 0x30, 0x00, 0x30, 0x00, 0x30, 0x00, 0x3b, 0x00, 0x3b, 0x00, 0x59, 0x00, - 0x75, 0x00, 0x7a, 0x00, 0x75, 0x00, 0x4f, 0x00, 0x53, 0x00, 0x53, 0x00, 0x45, 0x00, 0x78, 0x00, - 0x74, 0x00, 0x65, 0x00, 0x6e, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6f, 0x00, 0x6e, 0x00, 0x2d, 0x00, - 0x52, 0x00, 0x3b, 0x00, 0x32, 0x00, 0x30, 0x00, 0x31, 0x00, 0x39, 0x00, 0x3b, 0x00, 0x46, 0x00, - 0x4c, 0x00, 0x56, 0x00, 0x49, 0x00, 0x2d, 0x00, 0x36, 0x00, 0x31, 0x00, 0x34, 0x00, 0x59, 0x00, - 0x75, 0x00, 0x7a, 0x00, 0x75, 0x00, 0x4f, 0x00, 0x53, 0x00, 0x53, 0x00, 0x45, 0x00, 0x78, 0x00, - 0x74, 0x00, 0x65, 0x00, 0x6e, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6f, 0x00, 0x6e, 0x00, 0x20, 0x00, - 0x52, 0x00, 0x56, 0x00, 0x65, 0x00, 0x72, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6f, 0x00, 0x6e, 0x00, - 0x20, 0x00, 0x31, 0x00, 0x2e, 0x00, 0x30, 0x00, 0x30, 0x00, 0x30, 0x00, 0x59, 0x00, 0x75, 0x00, - 0x7a, 0x00, 0x75, 0x00, 0x4f, 0x00, 0x53, 0x00, 0x53, 0x00, 0x45, 0x00, 0x78, 0x00, 0x74, 0x00, - 0x65, 0x00, 0x6e, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6f, 0x00, 0x6e, 0x00, 0x2d, 0x00, 0x52, 0x00, - 0x52, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x9c, 0x00, 0x32, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x01, 0x02, 0x01, 0x03, 0x00, 0x03, 0x01, 0x04, - 0x01, 0x05, 0x01, 0x06, 0x01, 0x07, 0x01, 0x08, 0x01, 0x09, 0x01, 0x0a, 0x01, 0x0b, 0x01, 0x0c, - 0x01, 0x0d, 0x07, 0x75, 0x6e, 0x69, 0x30, 0x30, 0x30, 0x30, 0x07, 0x75, 0x6e, 0x69, 0x30, 0x30, - 0x30, 0x44, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, 0x45, 0x30, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, - 0x45, 0x31, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, 0x45, 0x32, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, - 0x45, 0x33, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, 0x45, 0x34, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, - 0x45, 0x35, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, 0x45, 0x36, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, - 0x45, 0x37, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, 0x45, 0x38, 0x07, 0x75, 0x6e, 0x69, 0x45, 0x30, - 0x45, 0x39, 0x00, 0x00, 0x00, 0x01, 0x00, 0x01, 0xff, 0xff, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x01, - 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0xB8, 0x00, 0x00, 0x2C, 0x4B, 0xB8, 0x00, 0x09, 0x50, 0x58, 0xB1, 0x01, 0x01, 0x8E, 0x59, 0xB8, + 0x01, 0xFF, 0x85, 0xB8, 0x00, 0x44, 0x1D, 0xB9, 0x00, 0x09, 0x00, 0x03, 0x5F, 0x5E, 0x2D, 0xB8, + 0x00, 0x01, 0x2C, 0x20, 0x20, 0x45, 0x69, 0x44, 0xB0, 0x01, 0x60, 0x2D, 0xB8, 0x00, 0x02, 0x2C, + 0xB8, 0x00, 0x01, 0x2A, 0x21, 0x2D, 0xB8, 0x00, 0x03, 0x2C, 0x20, 0x46, 0xB0, 0x03, 0x25, 0x46, + 0x52, 0x58, 0x23, 0x59, 0x20, 0x8A, 0x20, 0x8A, 0x49, 0x64, 0x8A, 0x20, 0x46, 0x20, 0x68, 0x61, + 0x64, 0xB0, 0x04, 0x25, 0x46, 0x20, 0x68, 0x61, 0x64, 0x52, 0x58, 0x23, 0x65, 0x8A, 0x59, 0x2F, + 0x20, 0xB0, 0x00, 0x53, 0x58, 0x69, 0x20, 0xB0, 0x00, 0x54, 0x58, 0x21, 0xB0, 0x40, 0x59, 0x1B, + 0x69, 0x20, 0xB0, 0x00, 0x54, 0x58, 0x21, 0xB0, 0x40, 0x65, 0x59, 0x59, 0x3A, 0x2D, 0xB8, 0x00, + 0x04, 0x2C, 0x20, 0x46, 0xB0, 0x04, 0x25, 0x46, 0x52, 0x58, 0x23, 0x8A, 0x59, 0x20, 0x46, 0x20, + 0x6A, 0x61, 0x64, 0xB0, 0x04, 0x25, 0x46, 0x20, 0x6A, 0x61, 0x64, 0x52, 0x58, 0x23, 0x8A, 0x59, + 0x2F, 0xFD, 0x2D, 0xB8, 0x00, 0x05, 0x2C, 0x4B, 0x20, 0xB0, 0x03, 0x26, 0x50, 0x58, 0x51, 0x58, + 0xB0, 0x80, 0x44, 0x1B, 0xB0, 0x40, 0x44, 0x59, 0x1B, 0x21, 0x21, 0x20, 0x45, 0xB0, 0xC0, 0x50, + 0x58, 0xB0, 0xC0, 0x44, 0x1B, 0x21, 0x59, 0x59, 0x2D, 0xB8, 0x00, 0x06, 0x2C, 0x20, 0x20, 0x45, + 0x69, 0x44, 0xB0, 0x01, 0x60, 0x20, 0x20, 0x45, 0x7D, 0x69, 0x18, 0x44, 0xB0, 0x01, 0x60, 0x2D, + 0xB8, 0x00, 0x07, 0x2C, 0xB8, 0x00, 0x06, 0x2A, 0x2D, 0xB8, 0x00, 0x08, 0x2C, 0x4B, 0x20, 0xB0, + 0x03, 0x26, 0x53, 0x58, 0xB0, 0x40, 0x1B, 0xB0, 0x00, 0x59, 0x8A, 0x8A, 0x20, 0xB0, 0x03, 0x26, + 0x53, 0x58, 0x23, 0x21, 0xB0, 0x80, 0x8A, 0x8A, 0x1B, 0x8A, 0x23, 0x59, 0x20, 0xB0, 0x03, 0x26, + 0x53, 0x58, 0x23, 0x21, 0xB8, 0x00, 0xC0, 0x8A, 0x8A, 0x1B, 0x8A, 0x23, 0x59, 0x20, 0xB0, 0x03, + 0x26, 0x53, 0x58, 0x23, 0x21, 0xB8, 0x01, 0x00, 0x8A, 0x8A, 0x1B, 0x8A, 0x23, 0x59, 0x20, 0xB0, + 0x03, 0x26, 0x53, 0x58, 0x23, 0x21, 0xB8, 0x01, 0x40, 0x8A, 0x8A, 0x1B, 0x8A, 0x23, 0x59, 0x20, + 0xB8, 0x00, 0x03, 0x26, 0x53, 0x58, 0xB0, 0x03, 0x25, 0x45, 0xB8, 0x01, 0x80, 0x50, 0x58, 0x23, + 0x21, 0xB8, 0x01, 0x80, 0x23, 0x21, 0x1B, 0xB0, 0x03, 0x25, 0x45, 0x23, 0x21, 0x23, 0x21, 0x59, + 0x1B, 0x21, 0x59, 0x44, 0x2D, 0xB8, 0x00, 0x09, 0x2C, 0x4B, 0x53, 0x58, 0x45, 0x44, 0x1B, 0x21, + 0x21, 0x59, 0x2D, 0x00, 0xB8, 0x00, 0x00, 0x2B, 0x00, 0xBA, 0x00, 0x01, 0x00, 0x01, 0x00, 0x07, + 0x2B, 0xB8, 0x00, 0x00, 0x20, 0x45, 0x7D, 0x69, 0x18, 0x44, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x16, 0x00, 0x16, 0x00, 0x16, 0x00, 0x16, 0x00, 0x70, + 0x00, 0xDC, 0x01, 0x34, 0x01, 0x7C, 0x01, 0xA2, 0x01, 0xF4, 0x02, 0x3C, 0x02, 0xA8, 0x03, 0x4C, + 0x03, 0xE2, 0x04, 0x20, 0x04, 0x58, 0x04, 0x9A, 0x04, 0xEE, 0x05, 0x32, 0x05, 0x64, 0x05, 0x80, + 0x05, 0xC6, 0x05, 0xF6, 0x06, 0x54, 0x06, 0xB2, 0x07, 0x38, 0x07, 0x60, 0x07, 0x82, 0x00, 0x00, + 0x00, 0x02, 0x00, 0xA4, 0xFF, 0xFF, 0x03, 0x5C, 0x03, 0x09, 0x00, 0x03, 0x00, 0x07, 0x00, 0x00, + 0x13, 0x11, 0x21, 0x11, 0x25, 0x21, 0x11, 0x21, 0xCD, 0x02, 0x66, 0xFD, 0x71, 0x02, 0xB8, 0xFD, + 0x48, 0x02, 0xE0, 0xFD, 0x48, 0x02, 0xB8, 0x29, 0xFC, 0xF6, 0x00, 0x00, 0x00, 0x04, 0x00, 0x14, + 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0F, 0x00, 0x1F, 0x00, 0x2F, 0x00, 0x39, 0x00, 0x00, + 0x00, 0x22, 0x0E, 0x02, 0x14, 0x1E, 0x02, 0x32, 0x3E, 0x02, 0x34, 0x2E, 0x01, 0x24, 0x32, 0x1E, + 0x02, 0x14, 0x0E, 0x02, 0x22, 0x2E, 0x02, 0x34, 0x3E, 0x01, 0x13, 0x12, 0x37, 0x33, 0x13, 0x12, + 0x15, 0x16, 0x23, 0x2F, 0x01, 0x23, 0x07, 0x23, 0x22, 0x26, 0x25, 0x30, 0x27, 0x26, 0x2F, 0x01, + 0x06, 0x07, 0x06, 0x32, 0x02, 0x5A, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, + 0x46, 0x46, 0x77, 0xFE, 0x9E, 0xC8, 0xB7, 0x83, 0x4E, 0x4E, 0x83, 0xB7, 0xC8, 0xB7, 0x83, 0x4E, + 0x4E, 0x83, 0x23, 0x6C, 0x5E, 0x6D, 0x68, 0x68, 0x01, 0x39, 0x38, 0x2E, 0xD1, 0x2B, 0x37, 0x33, + 0x04, 0x01, 0x48, 0x1D, 0x1C, 0x0A, 0x05, 0x01, 0x45, 0x01, 0x89, 0x03, 0x3F, 0x46, 0x77, 0xA4, + 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x77, 0x4E, 0x83, 0xB7, 0xC8, 0xB7, + 0x83, 0x4E, 0x4E, 0x83, 0xB7, 0xC8, 0xB7, 0x83, 0xFD, 0x64, 0x01, 0x1A, 0xEB, 0xFE, 0xFE, 0xFE, + 0xFD, 0x03, 0x01, 0x01, 0x77, 0x78, 0x01, 0xCF, 0x4C, 0x4C, 0x1C, 0x0C, 0x02, 0xBE, 0x02, 0x00, + 0x00, 0x05, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0F, 0x00, 0x1B, 0x00, 0x2F, + 0x00, 0x3A, 0x00, 0x44, 0x00, 0x00, 0x12, 0x14, 0x1E, 0x02, 0x32, 0x3E, 0x02, 0x34, 0x2E, 0x02, + 0x22, 0x0E, 0x01, 0x02, 0x10, 0x3E, 0x01, 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x26, 0x01, + 0x16, 0x17, 0x14, 0x06, 0x07, 0x06, 0x2B, 0x01, 0x19, 0x01, 0x17, 0x32, 0x17, 0x16, 0x17, 0x16, + 0x07, 0x06, 0x0F, 0x01, 0x36, 0x37, 0x34, 0x2E, 0x01, 0x27, 0x23, 0x15, 0x33, 0x32, 0x27, 0x32, + 0x37, 0x36, 0x26, 0x27, 0x26, 0x2B, 0x01, 0x15, 0x45, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, + 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x77, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, + 0xF4, 0xE2, 0x01, 0xF7, 0x61, 0x01, 0x4E, 0x3E, 0x29, 0xAF, 0x4E, 0x81, 0x8B, 0x1D, 0x3C, 0x1F, + 0x19, 0x04, 0x06, 0x39, 0x57, 0x44, 0x01, 0x1B, 0x2D, 0x51, 0x46, 0x46, 0x47, 0x66, 0x70, 0x16, + 0x1F, 0x01, 0x2C, 0x08, 0x4B, 0x4C, 0x01, 0xDE, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, + 0xA4, 0x77, 0x46, 0x46, 0x77, 0xFE, 0x7C, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, + 0x84, 0x84, 0x01, 0x6D, 0x21, 0x5B, 0x40, 0x50, 0x05, 0x03, 0x01, 0x03, 0x01, 0x05, 0x01, 0x05, + 0x09, 0x30, 0x25, 0x29, 0x40, 0x21, 0xC2, 0x06, 0x3E, 0x1A, 0x21, 0x0B, 0x01, 0x8C, 0xE1, 0x0A, + 0x0E, 0x54, 0x0B, 0x02, 0x79, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, + 0x03, 0x70, 0x00, 0x0F, 0x00, 0x1B, 0x00, 0x38, 0x00, 0x00, 0x12, 0x14, 0x1E, 0x02, 0x32, 0x3E, + 0x02, 0x34, 0x2E, 0x02, 0x22, 0x0E, 0x01, 0x02, 0x10, 0x3E, 0x01, 0x20, 0x1E, 0x01, 0x10, 0x0E, + 0x01, 0x20, 0x26, 0x36, 0x34, 0x3F, 0x01, 0x27, 0x26, 0x27, 0x33, 0x17, 0x16, 0x33, 0x36, 0x3F, + 0x02, 0x32, 0x14, 0x06, 0x16, 0x12, 0x14, 0x2B, 0x01, 0x27, 0x26, 0x06, 0x0F, 0x01, 0x23, 0x45, + 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x77, 0x84, 0xE2, + 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x7B, 0x58, 0x58, 0x4D, 0x4F, 0x05, 0x7A, + 0x34, 0x34, 0x02, 0x01, 0x33, 0x32, 0x3C, 0x3C, 0xA1, 0x01, 0xB0, 0x3E, 0x3F, 0x39, 0x3B, 0x02, + 0x3A, 0x38, 0x3F, 0x01, 0xDE, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, + 0x46, 0x77, 0xFE, 0x7C, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x60, + 0x02, 0x87, 0x88, 0x79, 0x7A, 0x06, 0x54, 0x54, 0x01, 0x53, 0x53, 0x01, 0x01, 0xFB, 0x04, 0xFE, + 0xF8, 0x02, 0x5B, 0x5A, 0x03, 0x59, 0x59, 0x00, 0x00, 0x03, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, + 0x03, 0x70, 0x00, 0x0F, 0x00, 0x1B, 0x00, 0x2B, 0x00, 0x00, 0x00, 0x22, 0x0E, 0x02, 0x14, 0x1E, + 0x02, 0x32, 0x3E, 0x02, 0x34, 0x2E, 0x01, 0x24, 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x2E, + 0x01, 0x10, 0x36, 0x01, 0x35, 0x27, 0x26, 0x34, 0x3B, 0x01, 0x17, 0x16, 0x36, 0x3F, 0x01, 0x33, + 0x03, 0x15, 0x23, 0x02, 0x5A, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, + 0x46, 0x77, 0xFE, 0x7C, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x01, + 0x36, 0x5E, 0x5F, 0x3C, 0x3D, 0x3D, 0x3D, 0x03, 0x3B, 0x3B, 0x77, 0xBE, 0x68, 0x03, 0x3F, 0x46, + 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x77, 0x84, 0xE2, 0xFE, + 0xF4, 0xE2, 0x84, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0xFD, 0xF9, 0x6E, 0x96, 0x95, 0x01, 0x67, 0x67, + 0x03, 0x66, 0x65, 0xFE, 0xD3, 0xDA, 0x00, 0x00, 0x00, 0x03, 0x00, 0x14, 0xFF, 0xBD, 0x03, 0xEC, + 0x03, 0x4B, 0x00, 0x06, 0x00, 0x0C, 0x00, 0x12, 0x00, 0x00, 0x01, 0x21, 0x22, 0x15, 0x30, 0x11, + 0x21, 0x17, 0x21, 0x11, 0x10, 0x25, 0x21, 0x01, 0x11, 0x33, 0x11, 0x21, 0x15, 0x03, 0xBB, 0xFD, + 0x77, 0xED, 0x03, 0x76, 0x31, 0xFC, 0x28, 0x01, 0x1E, 0x02, 0xBA, 0xFD, 0x5C, 0x68, 0x01, 0x08, + 0x03, 0x1A, 0xEE, 0xFD, 0xC2, 0x31, 0x02, 0x6F, 0x01, 0x1E, 0x01, 0xFD, 0x36, 0x02, 0x07, 0xFE, + 0x50, 0x57, 0x00, 0x00, 0x00, 0x04, 0x00, 0x14, 0xFF, 0xBD, 0x03, 0xEC, 0x03, 0x4B, 0x00, 0x06, + 0x00, 0x0C, 0x00, 0x27, 0x00, 0x32, 0x00, 0x00, 0x05, 0x11, 0x34, 0x27, 0x30, 0x21, 0x11, 0x07, + 0x11, 0x21, 0x20, 0x19, 0x01, 0x25, 0x11, 0x33, 0x32, 0x17, 0x16, 0x17, 0x16, 0x17, 0x16, 0x07, + 0x06, 0x07, 0x06, 0x07, 0x1E, 0x02, 0x15, 0x07, 0x23, 0x27, 0x2E, 0x01, 0x2F, 0x01, 0x15, 0x13, + 0x36, 0x35, 0x34, 0x27, 0x26, 0x27, 0x23, 0x15, 0x33, 0x36, 0x03, 0xBB, 0xED, 0xFD, 0x77, 0x31, + 0x02, 0xBA, 0x01, 0x1E, 0xFD, 0x2A, 0x77, 0x76, 0x15, 0x49, 0x20, 0x35, 0x08, 0x04, 0x06, 0x13, + 0x66, 0x0C, 0x01, 0x1F, 0x2E, 0x65, 0x3D, 0x3D, 0x2A, 0x56, 0x28, 0x2E, 0x19, 0x99, 0x3C, 0x20, + 0x10, 0x56, 0x4F, 0x46, 0x47, 0x12, 0x02, 0x3E, 0xED, 0x01, 0xFC, 0xD4, 0x31, 0x03, 0x8E, 0xFE, + 0xE1, 0xFD, 0x91, 0xC4, 0x02, 0x07, 0x01, 0x04, 0x13, 0x21, 0x44, 0x1D, 0x19, 0x58, 0x15, 0x02, + 0x01, 0x13, 0x2D, 0xA2, 0x01, 0x01, 0x3D, 0x81, 0x1A, 0x01, 0x01, 0xDA, 0x01, 0x2D, 0x08, 0x3A, + 0x29, 0x0F, 0x08, 0x01, 0x85, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x14, 0xFF, 0xF5, 0x03, 0xEC, + 0x03, 0x13, 0x00, 0x09, 0x00, 0x11, 0x00, 0x26, 0x00, 0x32, 0x00, 0x00, 0x37, 0x21, 0x34, 0x10, + 0x35, 0x34, 0x27, 0x21, 0x04, 0x11, 0x23, 0x10, 0x25, 0x21, 0x16, 0x15, 0x11, 0x21, 0x37, 0x35, + 0x37, 0x36, 0x22, 0x2B, 0x01, 0x3D, 0x01, 0x3B, 0x01, 0x1D, 0x01, 0x0F, 0x01, 0x3B, 0x01, 0x1D, + 0x01, 0x2B, 0x01, 0x25, 0x35, 0x3B, 0x01, 0x1D, 0x01, 0x3B, 0x01, 0x1D, 0x01, 0x2B, 0x01, 0x45, + 0x03, 0x76, 0x45, 0xFE, 0x2D, 0xFE, 0xA2, 0x31, 0x01, 0x8F, 0x01, 0xD3, 0x76, 0xFC, 0x28, 0xA7, + 0x68, 0x68, 0x01, 0x5B, 0x5D, 0x90, 0x91, 0x6C, 0x6D, 0x71, 0x70, 0xA0, 0xA0, 0x01, 0x75, 0x27, + 0x28, 0x63, 0x63, 0x8B, 0x8A, 0x27, 0x69, 0x01, 0xA4, 0x69, 0x44, 0x01, 0x02, 0xFE, 0xA4, 0x01, + 0x8C, 0x03, 0x01, 0x75, 0xFD, 0x58, 0xBB, 0x24, 0x80, 0x80, 0x21, 0x21, 0x1F, 0x1E, 0x85, 0x86, + 0x20, 0x22, 0xC3, 0xC3, 0xA1, 0xA3, 0x20, 0x22, 0x00, 0x05, 0x00, 0x14, 0xFF, 0xF5, 0x03, 0xEC, + 0x03, 0x13, 0x00, 0x08, 0x00, 0x10, 0x00, 0x2B, 0x00, 0x37, 0x00, 0x44, 0x00, 0x00, 0x37, 0x21, + 0x11, 0x10, 0x25, 0x30, 0x21, 0x06, 0x15, 0x03, 0x11, 0x34, 0x37, 0x21, 0x04, 0x19, 0x01, 0x01, + 0x35, 0x17, 0x32, 0x17, 0x16, 0x17, 0x16, 0x07, 0x06, 0x07, 0x06, 0x17, 0x16, 0x17, 0x16, 0x17, + 0x16, 0x23, 0x2F, 0x01, 0x2E, 0x01, 0x2F, 0x01, 0x15, 0x23, 0x37, 0x32, 0x36, 0x37, 0x36, 0x35, + 0x26, 0x27, 0x26, 0x2B, 0x01, 0x15, 0x05, 0x35, 0x37, 0x36, 0x26, 0x2B, 0x01, 0x35, 0x21, 0x15, + 0x03, 0x17, 0x15, 0x45, 0x03, 0x76, 0xFE, 0xA2, 0xFE, 0x2D, 0x45, 0x31, 0x76, 0x01, 0xD3, 0x01, + 0x8F, 0xFE, 0x1E, 0x65, 0x6F, 0x15, 0x46, 0x10, 0x05, 0x04, 0x0D, 0x4F, 0x09, 0x09, 0x1F, 0x1D, + 0x3A, 0x06, 0x01, 0x30, 0x2F, 0x22, 0x37, 0x1E, 0x29, 0x14, 0x4E, 0x82, 0x34, 0x19, 0x0E, 0x13, + 0x0A, 0x22, 0x07, 0x38, 0x37, 0xFE, 0x3E, 0x68, 0x68, 0x01, 0x5C, 0x5C, 0x01, 0x20, 0xD8, 0xE1, + 0x27, 0x01, 0x5D, 0x01, 0x5B, 0x03, 0x01, 0x44, 0xFD, 0x58, 0x02, 0xA8, 0x75, 0x01, 0x03, 0xFE, + 0x74, 0xFE, 0x71, 0x01, 0x5C, 0xC5, 0x01, 0x04, 0x0C, 0x43, 0x15, 0x1D, 0x44, 0x10, 0x04, 0x06, + 0x14, 0x2B, 0x56, 0x10, 0x01, 0x01, 0x34, 0x52, 0x1C, 0x01, 0x01, 0xA5, 0xE3, 0x04, 0x06, 0x0A, + 0x20, 0x2C, 0x04, 0x01, 0x65, 0xE3, 0x47, 0x80, 0x80, 0x01, 0x42, 0x3D, 0xFE, 0xF5, 0x01, 0x41, + 0x00, 0x04, 0x00, 0x14, 0x00, 0x52, 0x03, 0xEC, 0x02, 0xB6, 0x00, 0x08, 0x00, 0x16, 0x00, 0x64, + 0x00, 0x70, 0x00, 0x00, 0x25, 0x11, 0x21, 0x22, 0x15, 0x30, 0x15, 0x14, 0x33, 0x11, 0x21, 0x32, + 0x15, 0x11, 0x14, 0x27, 0x21, 0x22, 0x26, 0x3D, 0x01, 0x34, 0x36, 0x13, 0x26, 0x27, 0x26, 0x27, + 0x26, 0x37, 0x33, 0x36, 0x37, 0x36, 0x33, 0x16, 0x17, 0x16, 0x17, 0x16, 0x37, 0x36, 0x37, 0x36, + 0x35, 0x34, 0x27, 0x26, 0x27, 0x26, 0x27, 0x26, 0x27, 0x26, 0x27, 0x26, 0x34, 0x37, 0x36, 0x37, + 0x36, 0x37, 0x36, 0x17, 0x16, 0x17, 0x16, 0x17, 0x16, 0x17, 0x16, 0x0F, 0x01, 0x22, 0x06, 0x23, + 0x27, 0x26, 0x27, 0x26, 0x23, 0x22, 0x07, 0x06, 0x07, 0x06, 0x17, 0x16, 0x17, 0x16, 0x17, 0x16, + 0x17, 0x16, 0x17, 0x16, 0x07, 0x06, 0x07, 0x06, 0x27, 0x37, 0x35, 0x3B, 0x01, 0x1D, 0x01, 0x3B, + 0x01, 0x1D, 0x01, 0x2B, 0x01, 0x03, 0xBB, 0xFD, 0x2A, 0xA0, 0xA0, 0x02, 0xEE, 0x19, 0x19, 0xFD, + 0x12, 0x57, 0x7A, 0x7A, 0xCA, 0x38, 0x1D, 0x16, 0x08, 0x03, 0x01, 0x02, 0x0F, 0x0C, 0x1E, 0x01, + 0x02, 0x04, 0x0C, 0x2B, 0x0F, 0x0E, 0x18, 0x0C, 0x09, 0x04, 0x15, 0x32, 0x23, 0x12, 0x1C, 0x0E, + 0x09, 0x03, 0x01, 0x01, 0x09, 0x21, 0x0F, 0x14, 0x2E, 0x2A, 0x13, 0x0F, 0x0C, 0x08, 0x0B, 0x05, + 0x02, 0x01, 0x02, 0x03, 0x36, 0x03, 0x02, 0x03, 0x08, 0x0D, 0x23, 0x16, 0x0E, 0x10, 0x01, 0x01, + 0x07, 0x0B, 0x32, 0x25, 0x13, 0x26, 0x0F, 0x09, 0x01, 0x01, 0x0F, 0x11, 0x24, 0x21, 0x2A, 0xE3, + 0x20, 0x20, 0x52, 0x50, 0x71, 0x71, 0x84, 0x02, 0x00, 0xAF, 0xA2, 0xAF, 0x02, 0x32, 0x19, 0xFD, + 0xCE, 0x19, 0x01, 0x84, 0x5C, 0xA2, 0x5C, 0x85, 0xFE, 0x29, 0x04, 0x1E, 0x18, 0x26, 0x0F, 0x01, + 0x02, 0x01, 0x03, 0x05, 0x0B, 0x29, 0x06, 0x02, 0x03, 0x04, 0x11, 0x0B, 0x0D, 0x0A, 0x06, 0x12, + 0x0D, 0x0A, 0x07, 0x0C, 0x18, 0x0D, 0x10, 0x06, 0x18, 0x05, 0x27, 0x14, 0x09, 0x03, 0x0A, 0x0D, + 0x06, 0x09, 0x09, 0x0D, 0x0F, 0x14, 0x0C, 0x06, 0x03, 0x02, 0x04, 0x10, 0x0A, 0x11, 0x08, 0x09, + 0x0E, 0x0C, 0x07, 0x0C, 0x0C, 0x0A, 0x07, 0x0F, 0x20, 0x11, 0x18, 0x1E, 0x1A, 0x1E, 0x0C, 0x0B, + 0x03, 0xAA, 0xA5, 0x89, 0x8A, 0x1C, 0x1B, 0x00, 0x00, 0x05, 0x00, 0x14, 0x00, 0x53, 0x03, 0xEC, + 0x02, 0xB6, 0x00, 0x08, 0x00, 0x16, 0x00, 0x2E, 0x00, 0x38, 0x00, 0x65, 0x00, 0x00, 0x01, 0x30, + 0x21, 0x11, 0x21, 0x32, 0x3D, 0x01, 0x34, 0x27, 0x32, 0x16, 0x1D, 0x01, 0x14, 0x06, 0x23, 0x21, + 0x26, 0x35, 0x11, 0x34, 0x33, 0x01, 0x11, 0x33, 0x32, 0x17, 0x16, 0x17, 0x16, 0x07, 0x06, 0x07, + 0x17, 0x1E, 0x01, 0x1F, 0x01, 0x23, 0x2A, 0x01, 0x2E, 0x01, 0x23, 0x27, 0x15, 0x37, 0x32, 0x37, + 0x36, 0x27, 0x2E, 0x01, 0x2B, 0x01, 0x15, 0x05, 0x26, 0x27, 0x37, 0x32, 0x3F, 0x01, 0x16, 0x17, + 0x1E, 0x01, 0x37, 0x36, 0x27, 0x2E, 0x04, 0x37, 0x3E, 0x01, 0x33, 0x32, 0x17, 0x16, 0x17, 0x14, + 0x06, 0x27, 0x26, 0x27, 0x26, 0x0E, 0x01, 0x1E, 0x02, 0x17, 0x16, 0x06, 0x07, 0x06, 0x07, 0x06, + 0x03, 0x1B, 0xFD, 0x2A, 0x02, 0xD6, 0xA0, 0xA0, 0x57, 0x7A, 0x7A, 0x57, 0xFD, 0x12, 0x19, 0x19, + 0x01, 0xD3, 0x47, 0x44, 0x11, 0x3E, 0x18, 0x21, 0x0B, 0x0C, 0x43, 0x04, 0x17, 0x1C, 0x1E, 0x16, + 0x26, 0x26, 0x03, 0x4D, 0x18, 0x1E, 0x11, 0x25, 0x3A, 0x0C, 0x22, 0x08, 0x03, 0x1B, 0x3E, 0x29, + 0xFE, 0xAC, 0x0D, 0x04, 0x02, 0x02, 0x1E, 0x1D, 0x03, 0x02, 0x0C, 0x4C, 0x13, 0x20, 0x07, 0x04, + 0x1B, 0x56, 0x2D, 0x1C, 0x01, 0x02, 0x44, 0x35, 0x49, 0x1F, 0x10, 0x03, 0x41, 0x01, 0x06, 0x0A, + 0x16, 0x3C, 0x18, 0x0C, 0x16, 0x5D, 0x15, 0x33, 0x03, 0x2B, 0x1E, 0x34, 0x59, 0x02, 0x84, 0xFE, + 0x00, 0xAF, 0xA2, 0xAF, 0x32, 0x85, 0x5C, 0xA2, 0x5C, 0x84, 0x01, 0x17, 0x02, 0x32, 0x19, 0xFE, + 0x2F, 0x01, 0x45, 0x01, 0x02, 0x19, 0x22, 0x32, 0x39, 0x0B, 0x08, 0x0F, 0x27, 0x2F, 0x24, 0x75, + 0x12, 0x01, 0x88, 0xBB, 0x04, 0x09, 0x2A, 0x0F, 0x0D, 0x53, 0x8A, 0x17, 0x1E, 0x04, 0x03, 0x03, + 0x0C, 0x04, 0x26, 0x0E, 0x0C, 0x14, 0x1A, 0x0E, 0x0E, 0x16, 0x16, 0x2C, 0x1A, 0x2D, 0x2D, 0x2A, + 0x16, 0x1D, 0x06, 0x04, 0x01, 0x1A, 0x09, 0x11, 0x09, 0x17, 0x18, 0x0D, 0x17, 0x0C, 0x1B, 0x71, + 0x1B, 0x12, 0x01, 0x03, 0x00, 0x03, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0F, + 0x00, 0x1B, 0x00, 0x27, 0x00, 0x00, 0x00, 0x22, 0x0E, 0x02, 0x14, 0x1E, 0x02, 0x32, 0x3E, 0x02, + 0x34, 0x2E, 0x01, 0x24, 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x2E, 0x01, 0x10, 0x36, 0x13, + 0x33, 0x35, 0x33, 0x15, 0x33, 0x15, 0x23, 0x15, 0x23, 0x35, 0x23, 0x02, 0x5A, 0xB4, 0xA4, 0x77, + 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xFE, 0x7C, 0x01, 0x0C, 0xE2, 0x84, + 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x7C, 0xC5, 0x4E, 0xC5, 0xC4, 0x50, 0xC4, 0x03, 0x3F, + 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x77, 0x84, 0xE2, + 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0xFE, 0xC0, 0xC4, 0xC5, 0x4E, 0xC5, 0xC5, + 0x00, 0x03, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0F, 0x00, 0x1B, 0x00, 0x1F, + 0x00, 0x00, 0x00, 0x22, 0x0E, 0x02, 0x14, 0x1E, 0x02, 0x32, 0x3E, 0x02, 0x34, 0x2E, 0x01, 0x24, + 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x2E, 0x01, 0x10, 0x36, 0x13, 0x35, 0x21, 0x15, 0x02, + 0x5A, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xFE, 0x7C, + 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x7C, 0x01, 0xD8, 0x03, 0x3F, + 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x46, 0x46, 0x77, 0xA4, 0xB4, 0xA4, 0x77, 0x77, 0x84, 0xE2, + 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0xFE, 0x71, 0x4E, 0x4E, 0x00, 0x00, 0x00, + 0x00, 0x03, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0B, 0x00, 0x1B, 0x00, 0x25, + 0x00, 0x00, 0x00, 0x20, 0x0E, 0x01, 0x10, 0x1E, 0x01, 0x20, 0x3E, 0x01, 0x10, 0x26, 0x01, 0x12, + 0x37, 0x33, 0x13, 0x12, 0x15, 0x16, 0x23, 0x2F, 0x01, 0x23, 0x07, 0x23, 0x22, 0x26, 0x25, 0x30, + 0x27, 0x26, 0x2F, 0x01, 0x06, 0x07, 0x06, 0x32, 0x02, 0x86, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0xE2, + 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xFD, 0xA0, 0x6C, 0x5E, 0x6D, 0x68, 0x68, 0x01, 0x39, 0x38, 0x2E, + 0xD1, 0x2B, 0x37, 0x33, 0x04, 0x01, 0x48, 0x1D, 0x1C, 0x0A, 0x05, 0x01, 0x45, 0x01, 0x89, 0x03, + 0x70, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0xFD, 0x9A, 0x01, 0x1A, + 0xEB, 0xFE, 0xFE, 0xFE, 0xFD, 0x03, 0x01, 0x01, 0x77, 0x78, 0x01, 0xCF, 0x4C, 0x4C, 0x1C, 0x0C, + 0x02, 0xBE, 0x02, 0x00, 0x00, 0x04, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0B, + 0x00, 0x20, 0x00, 0x2B, 0x00, 0x35, 0x00, 0x00, 0x36, 0x10, 0x3E, 0x01, 0x20, 0x1E, 0x01, 0x10, + 0x0E, 0x01, 0x20, 0x26, 0x01, 0x30, 0x37, 0x36, 0x37, 0x36, 0x27, 0x26, 0x27, 0x26, 0x23, 0x27, + 0x19, 0x01, 0x33, 0x32, 0x37, 0x3E, 0x01, 0x35, 0x26, 0x07, 0x06, 0x2B, 0x01, 0x35, 0x33, 0x1E, + 0x02, 0x15, 0x06, 0x27, 0x23, 0x35, 0x33, 0x16, 0x17, 0x16, 0x14, 0x07, 0x06, 0x14, 0x84, 0xE2, + 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x01, 0xF7, 0x0A, 0x3A, 0x05, 0x04, 0x19, + 0x20, 0x3B, 0x1D, 0x8B, 0x81, 0x4E, 0xAF, 0x29, 0x3E, 0x4E, 0x01, 0xAE, 0x0D, 0x47, 0x46, 0x46, + 0x52, 0x2C, 0x1B, 0x01, 0xB7, 0x27, 0x4C, 0x4C, 0x07, 0x2C, 0x1E, 0x16, 0xFE, 0x01, 0x0C, 0xE2, + 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x01, 0x6D, 0x06, 0x21, 0x40, 0x2A, 0x24, 0x30, + 0x09, 0x05, 0x01, 0xFE, 0xFB, 0xFE, 0xFD, 0x03, 0x05, 0x4F, 0x41, 0x5B, 0x9B, 0x01, 0x8C, 0x01, + 0x0B, 0x21, 0x1A, 0x3E, 0xDA, 0x79, 0x01, 0x01, 0x0B, 0x54, 0x0E, 0x0A, 0x00, 0x02, 0x00, 0x14, + 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0B, 0x00, 0x29, 0x00, 0x00, 0x36, 0x10, 0x3E, 0x01, + 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x26, 0x36, 0x14, 0x3B, 0x01, 0x37, 0x36, 0x37, 0x36, + 0x1F, 0x01, 0x33, 0x32, 0x34, 0x02, 0x26, 0x36, 0x34, 0x23, 0x0F, 0x01, 0x06, 0x07, 0x22, 0x2F, + 0x01, 0x23, 0x16, 0x1F, 0x01, 0x07, 0x14, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, + 0xF4, 0xE2, 0x7B, 0x3D, 0x3F, 0x38, 0x3A, 0x01, 0x02, 0x3A, 0x39, 0x3F, 0x3E, 0xB0, 0x01, 0xA1, + 0x3C, 0x3C, 0x32, 0x33, 0x01, 0x02, 0x34, 0x34, 0x7A, 0x05, 0x4F, 0x4D, 0x58, 0xFE, 0x01, 0x0C, + 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x62, 0x02, 0x59, 0x59, 0x02, 0x01, 0x5A, + 0x5B, 0x02, 0x01, 0x08, 0x04, 0xFB, 0x01, 0x01, 0x53, 0x53, 0x01, 0x54, 0x54, 0x06, 0x7A, 0x79, + 0x88, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0B, + 0x00, 0x1B, 0x00, 0x00, 0x00, 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x2E, 0x01, 0x10, 0x36, + 0x01, 0x15, 0x33, 0x35, 0x13, 0x23, 0x07, 0x0E, 0x01, 0x2F, 0x01, 0x23, 0x22, 0x16, 0x1F, 0x01, + 0x01, 0x7A, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x01, 0x36, 0x68, + 0xBE, 0x77, 0x3B, 0x3C, 0x02, 0x3D, 0x3D, 0x3D, 0x3D, 0x01, 0x5F, 0x5E, 0x03, 0x70, 0x84, 0xE2, + 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0xFD, 0xF9, 0x6D, 0xDA, 0x01, 0x2D, 0x65, + 0x66, 0x03, 0x67, 0x67, 0x01, 0x95, 0x96, 0x00, 0x00, 0x02, 0x00, 0x14, 0xFF, 0xBF, 0x03, 0xEC, + 0x03, 0x4A, 0x00, 0x05, 0x00, 0x0B, 0x00, 0x00, 0x05, 0x21, 0x11, 0x10, 0x05, 0x21, 0x01, 0x21, + 0x35, 0x21, 0x11, 0x23, 0x03, 0xEC, 0xFC, 0x28, 0x01, 0x14, 0x02, 0xC4, 0xFD, 0x5C, 0x01, 0x70, + 0xFE, 0xF8, 0x68, 0x41, 0x02, 0x77, 0x01, 0x14, 0x01, 0xFD, 0x38, 0x57, 0x01, 0xB0, 0x00, 0x00, + 0x00, 0x03, 0x00, 0x14, 0xFF, 0xBF, 0x03, 0xEC, 0x03, 0x49, 0x00, 0x05, 0x00, 0x20, 0x00, 0x2B, + 0x00, 0x00, 0x17, 0x11, 0x21, 0x20, 0x19, 0x01, 0x25, 0x33, 0x35, 0x17, 0x1E, 0x01, 0x1F, 0x01, + 0x33, 0x37, 0x2E, 0x02, 0x27, 0x34, 0x37, 0x36, 0x37, 0x36, 0x27, 0x26, 0x27, 0x26, 0x27, 0x26, + 0x2B, 0x01, 0x05, 0x06, 0x2B, 0x01, 0x35, 0x33, 0x16, 0x17, 0x16, 0x15, 0x14, 0x14, 0x02, 0xC4, + 0x01, 0x14, 0xFD, 0x2A, 0x69, 0x19, 0x2E, 0x28, 0x56, 0x2A, 0x3D, 0x3D, 0x01, 0x65, 0x2C, 0x20, + 0x0D, 0x66, 0x13, 0x06, 0x04, 0x09, 0x34, 0x20, 0x49, 0x15, 0x76, 0x77, 0x01, 0x02, 0x0C, 0x47, + 0x46, 0x4F, 0x56, 0x10, 0x20, 0x41, 0x03, 0x8A, 0xFE, 0xED, 0xFD, 0x89, 0xC2, 0xDA, 0x01, 0x01, + 0x1A, 0x81, 0x3D, 0x01, 0x01, 0xA3, 0x2C, 0x13, 0x01, 0x02, 0x13, 0x5A, 0x1A, 0x1C, 0x44, 0x21, + 0x13, 0x04, 0x01, 0xDA, 0x02, 0x85, 0x01, 0x08, 0x0F, 0x29, 0x3A, 0x00, 0x00, 0x03, 0x00, 0x14, + 0xFF, 0xFB, 0x03, 0xEC, 0x03, 0x0E, 0x00, 0x08, 0x00, 0x15, 0x00, 0x1B, 0x00, 0x00, 0x05, 0x21, + 0x11, 0x10, 0x21, 0x30, 0x21, 0x32, 0x15, 0x01, 0x21, 0x35, 0x23, 0x13, 0x35, 0x21, 0x15, 0x33, + 0x32, 0x22, 0x0F, 0x01, 0x05, 0x21, 0x35, 0x23, 0x11, 0x23, 0x03, 0xEC, 0xFC, 0x28, 0x01, 0x8A, + 0x01, 0xEC, 0x62, 0xFC, 0xCF, 0x01, 0x40, 0xE1, 0xD9, 0xFE, 0xDF, 0x5D, 0x5C, 0x01, 0x67, 0x68, + 0x01, 0x75, 0x01, 0x15, 0xC6, 0x4F, 0x05, 0x01, 0x89, 0x01, 0x8A, 0x63, 0xFD, 0xE1, 0x42, 0x01, + 0x0B, 0x3D, 0x42, 0x80, 0x80, 0x48, 0x42, 0x01, 0x44, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x14, + 0xFF, 0xFB, 0x03, 0xEC, 0x03, 0x0E, 0x00, 0x07, 0x00, 0x22, 0x00, 0x2F, 0x00, 0x3C, 0x00, 0x00, + 0x17, 0x11, 0x34, 0x37, 0x21, 0x20, 0x19, 0x01, 0x01, 0x15, 0x33, 0x35, 0x17, 0x1E, 0x01, 0x1F, + 0x02, 0x32, 0x35, 0x26, 0x27, 0x26, 0x27, 0x26, 0x37, 0x36, 0x37, 0x36, 0x27, 0x26, 0x27, 0x26, + 0x23, 0x27, 0x17, 0x30, 0x23, 0x35, 0x33, 0x32, 0x17, 0x16, 0x17, 0x14, 0x07, 0x0E, 0x01, 0x05, + 0x21, 0x35, 0x27, 0x13, 0x35, 0x21, 0x15, 0x33, 0x32, 0x14, 0x0F, 0x01, 0x14, 0x62, 0x01, 0xEC, + 0x01, 0x8A, 0xFE, 0x1E, 0x4E, 0x14, 0x29, 0x1E, 0x37, 0x22, 0x2F, 0x2F, 0x06, 0x3A, 0x1D, 0x1F, + 0x09, 0x09, 0x4E, 0x0E, 0x04, 0x05, 0x0F, 0x47, 0x15, 0x6F, 0x65, 0x82, 0x34, 0x37, 0x38, 0x07, + 0x23, 0x09, 0x13, 0x0D, 0x1A, 0xFD, 0xD6, 0x01, 0x40, 0xE1, 0xD8, 0xFE, 0xE0, 0x5C, 0x5C, 0x67, + 0x68, 0x05, 0x02, 0xB0, 0x62, 0x01, 0xFE, 0x76, 0xFE, 0x77, 0x01, 0x56, 0xC5, 0xA5, 0x01, 0x01, + 0x1C, 0x52, 0x34, 0x01, 0x01, 0x0E, 0x58, 0x2C, 0x13, 0x06, 0x04, 0x0F, 0x45, 0x1E, 0x14, 0x42, + 0x0D, 0x04, 0x01, 0xA7, 0x65, 0x01, 0x04, 0x2C, 0x21, 0x09, 0x07, 0x03, 0xE3, 0x41, 0x01, 0x01, + 0x0B, 0x3D, 0x42, 0x01, 0x80, 0x80, 0x00, 0x00, 0x00, 0x03, 0x00, 0x14, 0x00, 0x5D, 0x03, 0xEC, + 0x02, 0xAB, 0x00, 0x08, 0x00, 0x37, 0x00, 0x3D, 0x00, 0x00, 0x13, 0x30, 0x21, 0x11, 0x21, 0x22, + 0x3D, 0x01, 0x34, 0x05, 0x37, 0x34, 0x27, 0x26, 0x27, 0x26, 0x07, 0x06, 0x07, 0x0E, 0x01, 0x17, + 0x1E, 0x01, 0x17, 0x16, 0x14, 0x07, 0x06, 0x26, 0x27, 0x26, 0x27, 0x22, 0x06, 0x07, 0x22, 0x17, + 0x1E, 0x01, 0x17, 0x16, 0x37, 0x36, 0x27, 0x26, 0x27, 0x2E, 0x02, 0x37, 0x36, 0x33, 0x32, 0x1F, + 0x02, 0x33, 0x35, 0x23, 0x11, 0x23, 0xD6, 0x03, 0x16, 0xFC, 0xEA, 0xC2, 0x01, 0xC6, 0x02, 0x01, + 0x0C, 0x3A, 0x2B, 0x2D, 0x13, 0x10, 0x2B, 0x01, 0x33, 0x17, 0x55, 0x15, 0x04, 0x09, 0x14, 0x58, + 0x0C, 0x04, 0x02, 0x02, 0x26, 0x14, 0x01, 0x03, 0x08, 0x33, 0x38, 0x5F, 0x20, 0x10, 0x01, 0x03, + 0x3C, 0x12, 0x59, 0x11, 0x01, 0x02, 0x39, 0x2C, 0x09, 0x02, 0x9D, 0xE2, 0xA2, 0x40, 0x02, 0xAB, + 0xFD, 0xB2, 0xD2, 0xAA, 0xD2, 0xDC, 0x03, 0x07, 0x0B, 0x38, 0x10, 0x0C, 0x09, 0x04, 0x08, 0x19, + 0x6C, 0x17, 0x0B, 0x17, 0x11, 0x07, 0x17, 0x0A, 0x1A, 0x0A, 0x29, 0x0C, 0x04, 0x04, 0x02, 0x10, + 0x25, 0x37, 0x04, 0x06, 0x37, 0x1D, 0x1C, 0x3F, 0x19, 0x08, 0x16, 0x13, 0x0B, 0x1F, 0x2B, 0x04, + 0xE9, 0x37, 0x01, 0x13, 0x00, 0x04, 0x00, 0x14, 0x00, 0x5D, 0x03, 0xEC, 0x02, 0xAB, 0x00, 0x07, + 0x00, 0x1F, 0x00, 0x2A, 0x00, 0x58, 0x00, 0x00, 0x01, 0x32, 0x1D, 0x01, 0x14, 0x23, 0x21, 0x11, + 0x01, 0x33, 0x35, 0x17, 0x1E, 0x03, 0x3B, 0x01, 0x27, 0x2E, 0x01, 0x2F, 0x01, 0x36, 0x37, 0x36, + 0x27, 0x26, 0x27, 0x26, 0x2B, 0x01, 0x17, 0x30, 0x23, 0x35, 0x33, 0x32, 0x16, 0x17, 0x16, 0x07, + 0x06, 0x05, 0x16, 0x37, 0x36, 0x37, 0x3E, 0x01, 0x27, 0x2E, 0x03, 0x3E, 0x01, 0x17, 0x16, 0x17, + 0x30, 0x37, 0x36, 0x27, 0x26, 0x27, 0x26, 0x27, 0x22, 0x06, 0x07, 0x06, 0x1E, 0x03, 0x17, 0x16, + 0x07, 0x06, 0x26, 0x27, 0x26, 0x27, 0x07, 0x06, 0x23, 0x07, 0x16, 0x03, 0x2A, 0xC2, 0xC2, 0xFC, + 0xEA, 0x01, 0xEC, 0x41, 0x11, 0x1F, 0x17, 0x4D, 0x02, 0x27, 0x26, 0x16, 0x1E, 0x1C, 0x17, 0x04, + 0x43, 0x0C, 0x0B, 0x21, 0x18, 0x3E, 0x0F, 0x46, 0x47, 0x66, 0x25, 0x29, 0x3E, 0x1B, 0x03, 0x08, + 0x22, 0x0C, 0xFE, 0x4D, 0x22, 0x59, 0x34, 0x1E, 0x2B, 0x03, 0x33, 0x16, 0x5C, 0x16, 0x0C, 0x18, + 0x3C, 0x16, 0x0B, 0x05, 0x22, 0x21, 0x01, 0x03, 0x10, 0x1F, 0x49, 0x36, 0x43, 0x02, 0x01, 0x1C, + 0x2D, 0x56, 0x1B, 0x04, 0x07, 0x20, 0x13, 0x4B, 0x0D, 0x01, 0x04, 0x1D, 0x1E, 0x02, 0x02, 0x04, + 0x02, 0xAB, 0xD2, 0xAA, 0xD2, 0x02, 0x4E, 0xFE, 0x39, 0x89, 0x01, 0x01, 0x11, 0x75, 0x01, 0x25, + 0x2F, 0x27, 0x0F, 0x08, 0x0C, 0x38, 0x33, 0x21, 0x19, 0x02, 0x01, 0x8A, 0x53, 0x0D, 0x0F, 0x2A, + 0x09, 0x04, 0x8A, 0x3A, 0x03, 0x01, 0x12, 0x1B, 0x71, 0x1B, 0x0C, 0x17, 0x0D, 0x18, 0x17, 0x09, + 0x11, 0x09, 0x1A, 0x01, 0x01, 0x07, 0x1E, 0x15, 0x29, 0x01, 0x2D, 0x2D, 0x1A, 0x2C, 0x16, 0x16, + 0x0D, 0x0F, 0x1A, 0x14, 0x0C, 0x0D, 0x27, 0x04, 0x0C, 0x03, 0x03, 0x04, 0x1E, 0x00, 0x00, 0x00, + 0x00, 0x02, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0B, 0x00, 0x17, 0x00, 0x00, + 0x00, 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x2E, 0x01, 0x10, 0x36, 0x13, 0x15, 0x33, 0x15, + 0x33, 0x35, 0x33, 0x35, 0x23, 0x35, 0x23, 0x15, 0x01, 0x7A, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, + 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x7C, 0xC4, 0x50, 0xC4, 0xC5, 0x4E, 0x03, 0x70, 0x84, 0xE2, 0xFE, + 0xF4, 0xE2, 0x84, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0xFE, 0xC0, 0x4F, 0xC5, 0xC5, 0x4E, 0xC5, 0xC4, + 0x00, 0x02, 0x00, 0x14, 0xFF, 0x98, 0x03, 0xEC, 0x03, 0x70, 0x00, 0x0B, 0x00, 0x0F, 0x00, 0x00, + 0x00, 0x20, 0x1E, 0x01, 0x10, 0x0E, 0x01, 0x20, 0x2E, 0x01, 0x10, 0x36, 0x13, 0x21, 0x35, 0x21, + 0x01, 0x7A, 0x01, 0x0C, 0xE2, 0x84, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0x7C, 0x01, 0xD8, + 0xFE, 0x28, 0x03, 0x70, 0x84, 0xE2, 0xFE, 0xF4, 0xE2, 0x84, 0x84, 0xE2, 0x01, 0x0C, 0xE2, 0xFE, + 0x71, 0x4E, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0E, 0x00, 0xAE, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x15, 0x00, 0x2C, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x10, + 0x00, 0x64, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x07, 0x00, 0x85, 0x00, 0x01, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x10, 0x00, 0xAF, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x04, 0x00, 0x10, 0x00, 0xE2, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x00, 0x0D, + 0x01, 0x0F, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x06, 0x00, 0x10, 0x01, 0x3F, 0x00, 0x03, + 0x00, 0x01, 0x04, 0x09, 0x00, 0x00, 0x00, 0x2A, 0x00, 0x00, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, + 0x00, 0x01, 0x00, 0x20, 0x00, 0x42, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, 0x00, 0x02, 0x00, 0x0E, + 0x00, 0x75, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, 0x00, 0x03, 0x00, 0x20, 0x00, 0x8D, 0x00, 0x03, + 0x00, 0x01, 0x04, 0x09, 0x00, 0x04, 0x00, 0x20, 0x00, 0xC0, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, + 0x00, 0x05, 0x00, 0x1A, 0x00, 0xF3, 0x00, 0x03, 0x00, 0x01, 0x04, 0x09, 0x00, 0x06, 0x00, 0x20, + 0x01, 0x1D, 0x00, 0x59, 0x00, 0x75, 0x00, 0x7A, 0x00, 0x75, 0x00, 0x20, 0x00, 0x45, 0x00, 0x6D, + 0x00, 0x75, 0x00, 0x6C, 0x00, 0x61, 0x00, 0x74, 0x00, 0x6F, 0x00, 0x72, 0x00, 0x20, 0x00, 0x50, + 0x00, 0x72, 0x00, 0x6F, 0x00, 0x6A, 0x00, 0x65, 0x00, 0x63, 0x00, 0x74, 0x00, 0x00, 0x59, 0x75, + 0x7A, 0x75, 0x20, 0x45, 0x6D, 0x75, 0x6C, 0x61, 0x74, 0x6F, 0x72, 0x20, 0x50, 0x72, 0x6F, 0x6A, + 0x65, 0x63, 0x74, 0x00, 0x00, 0x59, 0x00, 0x75, 0x00, 0x7A, 0x00, 0x75, 0x00, 0x4F, 0x00, 0x53, + 0x00, 0x53, 0x00, 0x45, 0x00, 0x78, 0x00, 0x74, 0x00, 0x65, 0x00, 0x6E, 0x00, 0x73, 0x00, 0x69, + 0x00, 0x6F, 0x00, 0x6E, 0x00, 0x00, 0x59, 0x75, 0x7A, 0x75, 0x4F, 0x53, 0x53, 0x45, 0x78, 0x74, + 0x65, 0x6E, 0x73, 0x69, 0x6F, 0x6E, 0x00, 0x00, 0x52, 0x00, 0x65, 0x00, 0x67, 0x00, 0x75, 0x00, + 0x6C, 0x00, 0x61, 0x00, 0x72, 0x00, 0x00, 0x52, 0x65, 0x67, 0x75, 0x6C, 0x61, 0x72, 0x00, 0x00, + 0x59, 0x00, 0x75, 0x00, 0x7A, 0x00, 0x75, 0x00, 0x4F, 0x00, 0x53, 0x00, 0x53, 0x00, 0x45, 0x00, + 0x78, 0x00, 0x74, 0x00, 0x65, 0x00, 0x6E, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6F, 0x00, 0x6E, 0x00, + 0x00, 0x59, 0x75, 0x7A, 0x75, 0x4F, 0x53, 0x53, 0x45, 0x78, 0x74, 0x65, 0x6E, 0x73, 0x69, 0x6F, + 0x6E, 0x00, 0x00, 0x59, 0x00, 0x75, 0x00, 0x7A, 0x00, 0x75, 0x00, 0x4F, 0x00, 0x53, 0x00, 0x53, + 0x00, 0x45, 0x00, 0x78, 0x00, 0x74, 0x00, 0x65, 0x00, 0x6E, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6F, + 0x00, 0x6E, 0x00, 0x00, 0x59, 0x75, 0x7A, 0x75, 0x4F, 0x53, 0x53, 0x45, 0x78, 0x74, 0x65, 0x6E, + 0x73, 0x69, 0x6F, 0x6E, 0x00, 0x00, 0x56, 0x00, 0x65, 0x00, 0x72, 0x00, 0x73, 0x00, 0x69, 0x00, + 0x6F, 0x00, 0x6E, 0x00, 0x20, 0x00, 0x31, 0x00, 0x2E, 0x00, 0x30, 0x00, 0x30, 0x00, 0x30, 0x00, + 0x00, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6F, 0x6E, 0x20, 0x31, 0x2E, 0x30, 0x30, 0x30, 0x00, 0x00, + 0x59, 0x00, 0x75, 0x00, 0x7A, 0x00, 0x75, 0x00, 0x4F, 0x00, 0x53, 0x00, 0x53, 0x00, 0x45, 0x00, + 0x78, 0x00, 0x74, 0x00, 0x65, 0x00, 0x6E, 0x00, 0x73, 0x00, 0x69, 0x00, 0x6F, 0x00, 0x6E, 0x00, + 0x00, 0x59, 0x75, 0x7A, 0x75, 0x4F, 0x53, 0x53, 0x45, 0x78, 0x74, 0x65, 0x6E, 0x73, 0x69, 0x6F, + 0x6E, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xFF, 0xB5, 0x00, 0x32, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x1C, 0x00, 0x00, 0x01, 0x02, 0x01, 0x03, 0x00, 0x03, 0x01, 0x04, + 0x01, 0x05, 0x01, 0x06, 0x01, 0x07, 0x01, 0x08, 0x01, 0x09, 0x01, 0x0A, 0x01, 0x0B, 0x01, 0x0C, + 0x01, 0x0D, 0x01, 0x0E, 0x01, 0x0F, 0x01, 0x10, 0x01, 0x11, 0x01, 0x12, 0x01, 0x13, 0x01, 0x14, + 0x01, 0x15, 0x01, 0x16, 0x01, 0x17, 0x01, 0x18, 0x01, 0x19, 0x01, 0x1A, 0x01, 0x1B, 0x07, 0x75, + 0x6E, 0x69, 0x30, 0x30, 0x30, 0x30, 0x07, 0x75, 0x6E, 0x69, 0x30, 0x30, 0x30, 0x44, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x41, 0x30, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x41, 0x31, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x41, 0x32, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x41, 0x33, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x41, 0x34, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x41, 0x35, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x41, 0x36, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x41, 0x37, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x41, 0x38, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x41, 0x39, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x42, 0x33, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x42, 0x34, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x45, 0x30, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x45, 0x31, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x45, 0x32, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x45, 0x33, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x45, 0x34, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x45, 0x35, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x45, 0x36, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x45, 0x37, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x45, 0x38, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x45, 0x39, 0x07, 0x75, + 0x6E, 0x69, 0x45, 0x30, 0x45, 0x46, 0x07, 0x75, 0x6E, 0x69, 0x45, 0x30, 0x46, 0x30, 0x00, 0x00, + 0x00, 0x01, 0x00, 0x01, 0xFF, 0xFF, 0x00, 0x0F, }}; } // namespace FileSys::SystemArchive::SharedFontData diff --git a/src/core/file_sys/system_archive/data/font_nintendo_extended.h b/src/core/file_sys/system_archive/data/font_nintendo_extended.h index 2089f3db9..edb9df914 100644 --- a/src/core/file_sys/system_archive/data/font_nintendo_extended.h +++ b/src/core/file_sys/system_archive/data/font_nintendo_extended.h @@ -8,6 +8,6 @@ namespace FileSys::SystemArchive::SharedFontData { -extern const std::array<unsigned char, 2932> FONT_NINTENDO_EXTENDED; +extern const std::array<unsigned char, 6024> FONT_NINTENDO_EXTENDED; } // namespace FileSys::SystemArchive::SharedFontData diff --git a/src/core/file_sys/system_archive/system_version.cpp b/src/core/file_sys/system_archive/system_version.cpp index aa313de66..7bfbc9a67 100644 --- a/src/core/file_sys/system_archive/system_version.cpp +++ b/src/core/file_sys/system_archive/system_version.cpp @@ -12,17 +12,17 @@ namespace SystemVersionData { // This section should reflect the best system version to describe yuzu's HLE api. // TODO(DarkLordZach): Update when HLE gets better. -constexpr u8 VERSION_MAJOR = 10; +constexpr u8 VERSION_MAJOR = 11; constexpr u8 VERSION_MINOR = 0; -constexpr u8 VERSION_MICRO = 2; +constexpr u8 VERSION_MICRO = 0; -constexpr u8 REVISION_MAJOR = 1; +constexpr u8 REVISION_MAJOR = 5; constexpr u8 REVISION_MINOR = 0; constexpr char PLATFORM_STRING[] = "NX"; -constexpr char VERSION_HASH[] = "f90143fa8bbc061d4f68c35f95f04f8080c0ecdc"; -constexpr char DISPLAY_VERSION[] = "10.0.2"; -constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 10.0.2-1.0"; +constexpr char VERSION_HASH[] = "34197eba8810e2edd5e9dfcfbde7b340882e856d"; +constexpr char DISPLAY_VERSION[] = "11.0.0"; +constexpr char DISPLAY_TITLE[] = "NintendoSDK Firmware for NX 11.0.0-5.0"; } // namespace SystemVersionData diff --git a/src/core/file_sys/vfs.cpp b/src/core/file_sys/vfs.cpp index b2f026b6d..f497e9396 100644 --- a/src/core/file_sys/vfs.cpp +++ b/src/core/file_sys/vfs.cpp @@ -203,7 +203,7 @@ std::string VfsFile::GetFullPath() const { return GetContainingDirectory()->GetFullPath() + "/" + GetName(); } -std::shared_ptr<VfsFile> VfsDirectory::GetFileRelative(std::string_view path) const { +VirtualFile VfsDirectory::GetFileRelative(std::string_view path) const { auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); @@ -231,7 +231,7 @@ std::shared_ptr<VfsFile> VfsDirectory::GetFileRelative(std::string_view path) co return dir->GetFile(vec.back()); } -std::shared_ptr<VfsFile> VfsDirectory::GetFileAbsolute(std::string_view path) const { +VirtualFile VfsDirectory::GetFileAbsolute(std::string_view path) const { if (IsRoot()) { return GetFileRelative(path); } @@ -239,7 +239,7 @@ std::shared_ptr<VfsFile> VfsDirectory::GetFileAbsolute(std::string_view path) co return GetParentDirectory()->GetFileAbsolute(path); } -std::shared_ptr<VfsDirectory> VfsDirectory::GetDirectoryRelative(std::string_view path) const { +VirtualDir VfsDirectory::GetDirectoryRelative(std::string_view path) const { auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); @@ -261,7 +261,7 @@ std::shared_ptr<VfsDirectory> VfsDirectory::GetDirectoryRelative(std::string_vie return dir; } -std::shared_ptr<VfsDirectory> VfsDirectory::GetDirectoryAbsolute(std::string_view path) const { +VirtualDir VfsDirectory::GetDirectoryAbsolute(std::string_view path) const { if (IsRoot()) { return GetDirectoryRelative(path); } @@ -269,14 +269,14 @@ std::shared_ptr<VfsDirectory> VfsDirectory::GetDirectoryAbsolute(std::string_vie return GetParentDirectory()->GetDirectoryAbsolute(path); } -std::shared_ptr<VfsFile> VfsDirectory::GetFile(std::string_view name) const { +VirtualFile VfsDirectory::GetFile(std::string_view name) const { const auto& files = GetFiles(); const auto iter = std::find_if(files.begin(), files.end(), [&name](const auto& file1) { return name == file1->GetName(); }); return iter == files.end() ? nullptr : *iter; } -std::shared_ptr<VfsDirectory> VfsDirectory::GetSubdirectory(std::string_view name) const { +VirtualDir VfsDirectory::GetSubdirectory(std::string_view name) const { const auto& subs = GetSubdirectories(); const auto iter = std::find_if(subs.begin(), subs.end(), [&name](const auto& file1) { return name == file1->GetName(); }); @@ -301,7 +301,7 @@ std::size_t VfsDirectory::GetSize() const { return file_total + subdir_total; } -std::shared_ptr<VfsFile> VfsDirectory::CreateFileRelative(std::string_view path) { +VirtualFile VfsDirectory::CreateFileRelative(std::string_view path) { auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); @@ -324,7 +324,7 @@ std::shared_ptr<VfsFile> VfsDirectory::CreateFileRelative(std::string_view path) return dir->CreateFileRelative(Common::FS::GetPathWithoutTop(path)); } -std::shared_ptr<VfsFile> VfsDirectory::CreateFileAbsolute(std::string_view path) { +VirtualFile VfsDirectory::CreateFileAbsolute(std::string_view path) { if (IsRoot()) { return CreateFileRelative(path); } @@ -332,7 +332,7 @@ std::shared_ptr<VfsFile> VfsDirectory::CreateFileAbsolute(std::string_view path) return GetParentDirectory()->CreateFileAbsolute(path); } -std::shared_ptr<VfsDirectory> VfsDirectory::CreateDirectoryRelative(std::string_view path) { +VirtualDir VfsDirectory::CreateDirectoryRelative(std::string_view path) { auto vec = Common::FS::SplitPathComponents(path); vec.erase(std::remove_if(vec.begin(), vec.end(), [](const auto& str) { return str.empty(); }), vec.end()); @@ -355,7 +355,7 @@ std::shared_ptr<VfsDirectory> VfsDirectory::CreateDirectoryRelative(std::string_ return dir->CreateDirectoryRelative(Common::FS::GetPathWithoutTop(path)); } -std::shared_ptr<VfsDirectory> VfsDirectory::CreateDirectoryAbsolute(std::string_view path) { +VirtualDir VfsDirectory::CreateDirectoryAbsolute(std::string_view path) { if (IsRoot()) { return CreateDirectoryRelative(path); } @@ -446,27 +446,27 @@ bool ReadOnlyVfsDirectory::IsReadable() const { return true; } -std::shared_ptr<VfsDirectory> ReadOnlyVfsDirectory::CreateSubdirectory(std::string_view name) { +VirtualDir ReadOnlyVfsDirectory::CreateSubdirectory(std::string_view name) { return nullptr; } -std::shared_ptr<VfsFile> ReadOnlyVfsDirectory::CreateFile(std::string_view name) { +VirtualFile ReadOnlyVfsDirectory::CreateFile(std::string_view name) { return nullptr; } -std::shared_ptr<VfsFile> ReadOnlyVfsDirectory::CreateFileAbsolute(std::string_view path) { +VirtualFile ReadOnlyVfsDirectory::CreateFileAbsolute(std::string_view path) { return nullptr; } -std::shared_ptr<VfsFile> ReadOnlyVfsDirectory::CreateFileRelative(std::string_view path) { +VirtualFile ReadOnlyVfsDirectory::CreateFileRelative(std::string_view path) { return nullptr; } -std::shared_ptr<VfsDirectory> ReadOnlyVfsDirectory::CreateDirectoryAbsolute(std::string_view path) { +VirtualDir ReadOnlyVfsDirectory::CreateDirectoryAbsolute(std::string_view path) { return nullptr; } -std::shared_ptr<VfsDirectory> ReadOnlyVfsDirectory::CreateDirectoryRelative(std::string_view path) { +VirtualDir ReadOnlyVfsDirectory::CreateDirectoryRelative(std::string_view path) { return nullptr; } diff --git a/src/core/file_sys/vfs.h b/src/core/file_sys/vfs.h index 954094772..afd64e95c 100644 --- a/src/core/file_sys/vfs.h +++ b/src/core/file_sys/vfs.h @@ -91,7 +91,7 @@ public: // Resizes the file to new_size. Returns whether or not the operation was successful. virtual bool Resize(std::size_t new_size) = 0; // Gets a pointer to the directory containing this file, returning nullptr if there is none. - virtual std::shared_ptr<VfsDirectory> GetContainingDirectory() const = 0; + virtual VirtualDir GetContainingDirectory() const = 0; // Returns whether or not the file can be written to. virtual bool IsWritable() const = 0; @@ -183,27 +183,27 @@ public: // Retrives the file located at path as if the current directory was root. Returns nullptr if // not found. - virtual std::shared_ptr<VfsFile> GetFileRelative(std::string_view path) const; + virtual VirtualFile GetFileRelative(std::string_view path) const; // Calls GetFileRelative(path) on the root of the current directory. - virtual std::shared_ptr<VfsFile> GetFileAbsolute(std::string_view path) const; + virtual VirtualFile GetFileAbsolute(std::string_view path) const; // Retrives the directory located at path as if the current directory was root. Returns nullptr // if not found. - virtual std::shared_ptr<VfsDirectory> GetDirectoryRelative(std::string_view path) const; + virtual VirtualDir GetDirectoryRelative(std::string_view path) const; // Calls GetDirectoryRelative(path) on the root of the current directory. - virtual std::shared_ptr<VfsDirectory> GetDirectoryAbsolute(std::string_view path) const; + virtual VirtualDir GetDirectoryAbsolute(std::string_view path) const; // Returns a vector containing all of the files in this directory. - virtual std::vector<std::shared_ptr<VfsFile>> GetFiles() const = 0; + virtual std::vector<VirtualFile> GetFiles() const = 0; // Returns the file with filename matching name. Returns nullptr if directory dosen't have a // file with name. - virtual std::shared_ptr<VfsFile> GetFile(std::string_view name) const; + virtual VirtualFile GetFile(std::string_view name) const; // Returns a vector containing all of the subdirectories in this directory. - virtual std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const = 0; + virtual std::vector<VirtualDir> GetSubdirectories() const = 0; // Returns the directory with name matching name. Returns nullptr if directory dosen't have a // directory with name. - virtual std::shared_ptr<VfsDirectory> GetSubdirectory(std::string_view name) const; + virtual VirtualDir GetSubdirectory(std::string_view name) const; // Returns whether or not the directory can be written to. virtual bool IsWritable() const = 0; @@ -219,31 +219,31 @@ public: virtual std::size_t GetSize() const; // Returns the parent directory of this directory. Returns nullptr if this directory is root or // has no parent. - virtual std::shared_ptr<VfsDirectory> GetParentDirectory() const = 0; + virtual VirtualDir GetParentDirectory() const = 0; // Creates a new subdirectory with name name. Returns a pointer to the new directory or nullptr // if the operation failed. - virtual std::shared_ptr<VfsDirectory> CreateSubdirectory(std::string_view name) = 0; + virtual VirtualDir CreateSubdirectory(std::string_view name) = 0; // Creates a new file with name name. Returns a pointer to the new file or nullptr if the // operation failed. - virtual std::shared_ptr<VfsFile> CreateFile(std::string_view name) = 0; + virtual VirtualFile CreateFile(std::string_view name) = 0; // Creates a new file at the path relative to this directory. Also creates directories if // they do not exist and is supported by this implementation. Returns nullptr on any failure. - virtual std::shared_ptr<VfsFile> CreateFileRelative(std::string_view path); + virtual VirtualFile CreateFileRelative(std::string_view path); // Creates a new file at the path relative to root of this directory. Also creates directories // if they do not exist and is supported by this implementation. Returns nullptr on any failure. - virtual std::shared_ptr<VfsFile> CreateFileAbsolute(std::string_view path); + virtual VirtualFile CreateFileAbsolute(std::string_view path); // Creates a new directory at the path relative to this directory. Also creates directories if // they do not exist and is supported by this implementation. Returns nullptr on any failure. - virtual std::shared_ptr<VfsDirectory> CreateDirectoryRelative(std::string_view path); + virtual VirtualDir CreateDirectoryRelative(std::string_view path); // Creates a new directory at the path relative to root of this directory. Also creates // directories if they do not exist and is supported by this implementation. Returns nullptr on // any failure. - virtual std::shared_ptr<VfsDirectory> CreateDirectoryAbsolute(std::string_view path); + virtual VirtualDir CreateDirectoryAbsolute(std::string_view path); // Deletes the subdirectory with the given name and returns true on success. virtual bool DeleteSubdirectory(std::string_view name) = 0; @@ -280,12 +280,12 @@ class ReadOnlyVfsDirectory : public VfsDirectory { public: bool IsWritable() const override; bool IsReadable() const override; - std::shared_ptr<VfsDirectory> CreateSubdirectory(std::string_view name) override; - std::shared_ptr<VfsFile> CreateFile(std::string_view name) override; - std::shared_ptr<VfsFile> CreateFileAbsolute(std::string_view path) override; - std::shared_ptr<VfsFile> CreateFileRelative(std::string_view path) override; - std::shared_ptr<VfsDirectory> CreateDirectoryAbsolute(std::string_view path) override; - std::shared_ptr<VfsDirectory> CreateDirectoryRelative(std::string_view path) override; + VirtualDir CreateSubdirectory(std::string_view name) override; + VirtualFile CreateFile(std::string_view name) override; + VirtualFile CreateFileAbsolute(std::string_view path) override; + VirtualFile CreateFileRelative(std::string_view path) override; + VirtualDir CreateDirectoryAbsolute(std::string_view path) override; + VirtualDir CreateDirectoryRelative(std::string_view path) override; bool DeleteSubdirectory(std::string_view name) override; bool DeleteSubdirectoryRecursive(std::string_view name) override; bool CleanSubdirectoryRecursive(std::string_view name) override; diff --git a/src/core/file_sys/vfs_concat.cpp b/src/core/file_sys/vfs_concat.cpp index e0ff70174..3c5a7d87a 100644 --- a/src/core/file_sys/vfs_concat.cpp +++ b/src/core/file_sys/vfs_concat.cpp @@ -46,7 +46,7 @@ VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(std::vector<VirtualFile> f if (files.size() == 1) return files[0]; - return std::shared_ptr<VfsFile>(new ConcatenatedVfsFile(std::move(files), std::move(name))); + return VirtualFile(new ConcatenatedVfsFile(std::move(files), std::move(name))); } VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(u8 filler_byte, @@ -71,20 +71,23 @@ VirtualFile ConcatenatedVfsFile::MakeConcatenatedFile(u8 filler_byte, if (files.begin()->first != 0) files.emplace(0, std::make_shared<StaticVfsFile>(filler_byte, files.begin()->first)); - return std::shared_ptr<VfsFile>(new ConcatenatedVfsFile(std::move(files), std::move(name))); + return VirtualFile(new ConcatenatedVfsFile(std::move(files), std::move(name))); } std::string ConcatenatedVfsFile::GetName() const { - if (files.empty()) + if (files.empty()) { return ""; - if (!name.empty()) + } + if (!name.empty()) { return name; + } return files.begin()->second->GetName(); } std::size_t ConcatenatedVfsFile::GetSize() const { - if (files.empty()) + if (files.empty()) { return 0; + } return files.rbegin()->first + files.rbegin()->second->GetSize(); } @@ -92,9 +95,10 @@ bool ConcatenatedVfsFile::Resize(std::size_t new_size) { return false; } -std::shared_ptr<VfsDirectory> ConcatenatedVfsFile::GetContainingDirectory() const { - if (files.empty()) +VirtualDir ConcatenatedVfsFile::GetContainingDirectory() const { + if (files.empty()) { return nullptr; + } return files.begin()->second->GetContainingDirectory(); } diff --git a/src/core/file_sys/vfs_concat.h b/src/core/file_sys/vfs_concat.h index 7a26343c0..287c72555 100644 --- a/src/core/file_sys/vfs_concat.h +++ b/src/core/file_sys/vfs_concat.h @@ -31,7 +31,7 @@ public: std::string GetName() const override; std::size_t GetSize() const override; bool Resize(std::size_t new_size) override; - std::shared_ptr<VfsDirectory> GetContainingDirectory() const override; + VirtualDir GetContainingDirectory() const override; bool IsWritable() const override; bool IsReadable() const override; std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override; diff --git a/src/core/file_sys/vfs_layered.cpp b/src/core/file_sys/vfs_layered.cpp index 338e398da..434b03cec 100644 --- a/src/core/file_sys/vfs_layered.cpp +++ b/src/core/file_sys/vfs_layered.cpp @@ -20,10 +20,10 @@ VirtualDir LayeredVfsDirectory::MakeLayeredDirectory(std::vector<VirtualDir> dir if (dirs.size() == 1) return dirs[0]; - return std::shared_ptr<VfsDirectory>(new LayeredVfsDirectory(std::move(dirs), std::move(name))); + return VirtualDir(new LayeredVfsDirectory(std::move(dirs), std::move(name))); } -std::shared_ptr<VfsFile> LayeredVfsDirectory::GetFileRelative(std::string_view path) const { +VirtualFile LayeredVfsDirectory::GetFileRelative(std::string_view path) const { for (const auto& layer : dirs) { const auto file = layer->GetFileRelative(path); if (file != nullptr) @@ -33,23 +33,23 @@ std::shared_ptr<VfsFile> LayeredVfsDirectory::GetFileRelative(std::string_view p return nullptr; } -std::shared_ptr<VfsDirectory> LayeredVfsDirectory::GetDirectoryRelative( - std::string_view path) const { +VirtualDir LayeredVfsDirectory::GetDirectoryRelative(std::string_view path) const { std::vector<VirtualDir> out; for (const auto& layer : dirs) { auto dir = layer->GetDirectoryRelative(path); - if (dir != nullptr) + if (dir != nullptr) { out.push_back(std::move(dir)); + } } return MakeLayeredDirectory(std::move(out)); } -std::shared_ptr<VfsFile> LayeredVfsDirectory::GetFile(std::string_view name) const { +VirtualFile LayeredVfsDirectory::GetFile(std::string_view name) const { return GetFileRelative(name); } -std::shared_ptr<VfsDirectory> LayeredVfsDirectory::GetSubdirectory(std::string_view name) const { +VirtualDir LayeredVfsDirectory::GetSubdirectory(std::string_view name) const { return GetDirectoryRelative(name); } @@ -57,7 +57,7 @@ std::string LayeredVfsDirectory::GetFullPath() const { return dirs[0]->GetFullPath(); } -std::vector<std::shared_ptr<VfsFile>> LayeredVfsDirectory::GetFiles() const { +std::vector<VirtualFile> LayeredVfsDirectory::GetFiles() const { std::vector<VirtualFile> out; for (const auto& layer : dirs) { for (const auto& file : layer->GetFiles()) { @@ -72,7 +72,7 @@ std::vector<std::shared_ptr<VfsFile>> LayeredVfsDirectory::GetFiles() const { return out; } -std::vector<std::shared_ptr<VfsDirectory>> LayeredVfsDirectory::GetSubdirectories() const { +std::vector<VirtualDir> LayeredVfsDirectory::GetSubdirectories() const { std::vector<std::string> names; for (const auto& layer : dirs) { for (const auto& sd : layer->GetSubdirectories()) { @@ -101,15 +101,15 @@ std::string LayeredVfsDirectory::GetName() const { return name.empty() ? dirs[0]->GetName() : name; } -std::shared_ptr<VfsDirectory> LayeredVfsDirectory::GetParentDirectory() const { +VirtualDir LayeredVfsDirectory::GetParentDirectory() const { return dirs[0]->GetParentDirectory(); } -std::shared_ptr<VfsDirectory> LayeredVfsDirectory::CreateSubdirectory(std::string_view name) { +VirtualDir LayeredVfsDirectory::CreateSubdirectory(std::string_view name) { return nullptr; } -std::shared_ptr<VfsFile> LayeredVfsDirectory::CreateFile(std::string_view name) { +VirtualFile LayeredVfsDirectory::CreateFile(std::string_view name) { return nullptr; } diff --git a/src/core/file_sys/vfs_layered.h b/src/core/file_sys/vfs_layered.h index 8a25c3428..6d7513ac6 100644 --- a/src/core/file_sys/vfs_layered.h +++ b/src/core/file_sys/vfs_layered.h @@ -21,20 +21,20 @@ public: /// Wrapper function to allow for more efficient handling of dirs.size() == 0, 1 cases. static VirtualDir MakeLayeredDirectory(std::vector<VirtualDir> dirs, std::string name = ""); - std::shared_ptr<VfsFile> GetFileRelative(std::string_view path) const override; - std::shared_ptr<VfsDirectory> GetDirectoryRelative(std::string_view path) const override; - std::shared_ptr<VfsFile> GetFile(std::string_view name) const override; - std::shared_ptr<VfsDirectory> GetSubdirectory(std::string_view name) const override; + VirtualFile GetFileRelative(std::string_view path) const override; + VirtualDir GetDirectoryRelative(std::string_view path) const override; + VirtualFile GetFile(std::string_view name) const override; + VirtualDir GetSubdirectory(std::string_view name) const override; std::string GetFullPath() const override; - std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; - std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; + std::vector<VirtualFile> GetFiles() const override; + std::vector<VirtualDir> GetSubdirectories() const override; bool IsWritable() const override; bool IsReadable() const override; std::string GetName() const override; - std::shared_ptr<VfsDirectory> GetParentDirectory() const override; - std::shared_ptr<VfsDirectory> CreateSubdirectory(std::string_view name) override; - std::shared_ptr<VfsFile> CreateFile(std::string_view name) override; + VirtualDir GetParentDirectory() const override; + VirtualDir CreateSubdirectory(std::string_view name) override; + VirtualFile CreateFile(std::string_view name) override; bool DeleteSubdirectory(std::string_view name) override; bool DeleteFile(std::string_view name) override; bool Rename(std::string_view name) override; diff --git a/src/core/file_sys/vfs_offset.cpp b/src/core/file_sys/vfs_offset.cpp index 7714d3de5..056737b54 100644 --- a/src/core/file_sys/vfs_offset.cpp +++ b/src/core/file_sys/vfs_offset.cpp @@ -9,7 +9,7 @@ namespace FileSys { -OffsetVfsFile::OffsetVfsFile(std::shared_ptr<VfsFile> file_, std::size_t size_, std::size_t offset_, +OffsetVfsFile::OffsetVfsFile(VirtualFile file_, std::size_t size_, std::size_t offset_, std::string name_, VirtualDir parent_) : file(file_), offset(offset_), size(size_), name(std::move(name_)), parent(parent_ == nullptr ? file->GetContainingDirectory() : std::move(parent_)) {} @@ -37,7 +37,7 @@ bool OffsetVfsFile::Resize(std::size_t new_size) { return true; } -std::shared_ptr<VfsDirectory> OffsetVfsFile::GetContainingDirectory() const { +VirtualDir OffsetVfsFile::GetContainingDirectory() const { return parent; } diff --git a/src/core/file_sys/vfs_offset.h b/src/core/file_sys/vfs_offset.h index f7b7a3256..b2ccc5c7b 100644 --- a/src/core/file_sys/vfs_offset.h +++ b/src/core/file_sys/vfs_offset.h @@ -17,14 +17,14 @@ namespace FileSys { // the size of this wrapper. class OffsetVfsFile : public VfsFile { public: - OffsetVfsFile(std::shared_ptr<VfsFile> file, std::size_t size, std::size_t offset = 0, + OffsetVfsFile(VirtualFile file, std::size_t size, std::size_t offset = 0, std::string new_name = "", VirtualDir new_parent = nullptr); ~OffsetVfsFile() override; std::string GetName() const override; std::size_t GetSize() const override; bool Resize(std::size_t new_size) override; - std::shared_ptr<VfsDirectory> GetContainingDirectory() const override; + VirtualDir GetContainingDirectory() const override; bool IsWritable() const override; bool IsReadable() const override; std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override; @@ -42,7 +42,7 @@ public: private: std::size_t TrimToFit(std::size_t r_size, std::size_t r_offset) const; - std::shared_ptr<VfsFile> file; + VirtualFile file; std::size_t offset; std::size_t size; std::string name; diff --git a/src/core/file_sys/vfs_real.cpp b/src/core/file_sys/vfs_real.cpp index 488687ba9..a287eebe3 100644 --- a/src/core/file_sys/vfs_real.cpp +++ b/src/core/file_sys/vfs_real.cpp @@ -263,7 +263,7 @@ bool RealVfsFile::Resize(std::size_t new_size) { return backing->Resize(new_size); } -std::shared_ptr<VfsDirectory> RealVfsFile::GetContainingDirectory() const { +VirtualDir RealVfsFile::GetContainingDirectory() const { return base.OpenDirectory(parent_path, perms); } @@ -352,7 +352,7 @@ RealVfsDirectory::RealVfsDirectory(RealVfsFilesystem& base_, const std::string& RealVfsDirectory::~RealVfsDirectory() = default; -std::shared_ptr<VfsFile> RealVfsDirectory::GetFileRelative(std::string_view path) const { +VirtualFile RealVfsDirectory::GetFileRelative(std::string_view path) const { const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); if (!FS::Exists(full_path) || FS::IsDirectory(full_path)) { return nullptr; @@ -360,7 +360,7 @@ std::shared_ptr<VfsFile> RealVfsDirectory::GetFileRelative(std::string_view path return base.OpenFile(full_path, perms); } -std::shared_ptr<VfsDirectory> RealVfsDirectory::GetDirectoryRelative(std::string_view path) const { +VirtualDir RealVfsDirectory::GetDirectoryRelative(std::string_view path) const { const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); if (!FS::Exists(full_path) || !FS::IsDirectory(full_path)) { return nullptr; @@ -368,20 +368,20 @@ std::shared_ptr<VfsDirectory> RealVfsDirectory::GetDirectoryRelative(std::string return base.OpenDirectory(full_path, perms); } -std::shared_ptr<VfsFile> RealVfsDirectory::GetFile(std::string_view name) const { +VirtualFile RealVfsDirectory::GetFile(std::string_view name) const { return GetFileRelative(name); } -std::shared_ptr<VfsDirectory> RealVfsDirectory::GetSubdirectory(std::string_view name) const { +VirtualDir RealVfsDirectory::GetSubdirectory(std::string_view name) const { return GetDirectoryRelative(name); } -std::shared_ptr<VfsFile> RealVfsDirectory::CreateFileRelative(std::string_view path) { +VirtualFile RealVfsDirectory::CreateFileRelative(std::string_view path) { const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); return base.CreateFile(full_path, perms); } -std::shared_ptr<VfsDirectory> RealVfsDirectory::CreateDirectoryRelative(std::string_view path) { +VirtualDir RealVfsDirectory::CreateDirectoryRelative(std::string_view path) { const auto full_path = FS::SanitizePath(this->path + DIR_SEP + std::string(path)); return base.CreateDirectory(full_path, perms); } @@ -391,11 +391,11 @@ bool RealVfsDirectory::DeleteSubdirectoryRecursive(std::string_view name) { return base.DeleteDirectory(full_path); } -std::vector<std::shared_ptr<VfsFile>> RealVfsDirectory::GetFiles() const { +std::vector<VirtualFile> RealVfsDirectory::GetFiles() const { return IterateEntries<RealVfsFile, VfsFile>(); } -std::vector<std::shared_ptr<VfsDirectory>> RealVfsDirectory::GetSubdirectories() const { +std::vector<VirtualDir> RealVfsDirectory::GetSubdirectories() const { return IterateEntries<RealVfsDirectory, VfsDirectory>(); } @@ -411,7 +411,7 @@ std::string RealVfsDirectory::GetName() const { return path_components.back(); } -std::shared_ptr<VfsDirectory> RealVfsDirectory::GetParentDirectory() const { +VirtualDir RealVfsDirectory::GetParentDirectory() const { if (path_components.size() <= 1) { return nullptr; } @@ -419,12 +419,12 @@ std::shared_ptr<VfsDirectory> RealVfsDirectory::GetParentDirectory() const { return base.OpenDirectory(parent_path, perms); } -std::shared_ptr<VfsDirectory> RealVfsDirectory::CreateSubdirectory(std::string_view name) { +VirtualDir RealVfsDirectory::CreateSubdirectory(std::string_view name) { const std::string subdir_path = (path + DIR_SEP).append(name); return base.CreateDirectory(subdir_path, perms); } -std::shared_ptr<VfsFile> RealVfsDirectory::CreateFile(std::string_view name) { +VirtualFile RealVfsDirectory::CreateFile(std::string_view name) { const std::string file_path = (path + DIR_SEP).append(name); return base.CreateFile(file_path, perms); } diff --git a/src/core/file_sys/vfs_real.h b/src/core/file_sys/vfs_real.h index 0b537b22c..23e99865e 100644 --- a/src/core/file_sys/vfs_real.h +++ b/src/core/file_sys/vfs_real.h @@ -50,7 +50,7 @@ public: std::string GetName() const override; std::size_t GetSize() const override; bool Resize(std::size_t new_size) override; - std::shared_ptr<VfsDirectory> GetContainingDirectory() const override; + VirtualDir GetContainingDirectory() const override; bool IsWritable() const override; bool IsReadable() const override; std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override; @@ -79,21 +79,21 @@ class RealVfsDirectory : public VfsDirectory { public: ~RealVfsDirectory() override; - std::shared_ptr<VfsFile> GetFileRelative(std::string_view path) const override; - std::shared_ptr<VfsDirectory> GetDirectoryRelative(std::string_view path) const override; - std::shared_ptr<VfsFile> GetFile(std::string_view name) const override; - std::shared_ptr<VfsDirectory> GetSubdirectory(std::string_view name) const override; - std::shared_ptr<VfsFile> CreateFileRelative(std::string_view path) override; - std::shared_ptr<VfsDirectory> CreateDirectoryRelative(std::string_view path) override; + VirtualFile GetFileRelative(std::string_view path) const override; + VirtualDir GetDirectoryRelative(std::string_view path) const override; + VirtualFile GetFile(std::string_view name) const override; + VirtualDir GetSubdirectory(std::string_view name) const override; + VirtualFile CreateFileRelative(std::string_view path) override; + VirtualDir CreateDirectoryRelative(std::string_view path) override; bool DeleteSubdirectoryRecursive(std::string_view name) override; - std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; - std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; + std::vector<VirtualFile> GetFiles() const override; + std::vector<VirtualDir> GetSubdirectories() const override; bool IsWritable() const override; bool IsReadable() const override; std::string GetName() const override; - std::shared_ptr<VfsDirectory> GetParentDirectory() const override; - std::shared_ptr<VfsDirectory> CreateSubdirectory(std::string_view name) override; - std::shared_ptr<VfsFile> CreateFile(std::string_view name) override; + VirtualDir GetParentDirectory() const override; + VirtualDir CreateSubdirectory(std::string_view name) override; + VirtualFile CreateFile(std::string_view name) override; bool DeleteSubdirectory(std::string_view name) override; bool DeleteFile(std::string_view name) override; bool Rename(std::string_view name) override; diff --git a/src/core/file_sys/vfs_static.h b/src/core/file_sys/vfs_static.h index 8b27c30fa..c840b24b9 100644 --- a/src/core/file_sys/vfs_static.h +++ b/src/core/file_sys/vfs_static.h @@ -31,7 +31,7 @@ public: return true; } - std::shared_ptr<VfsDirectory> GetContainingDirectory() const override { + VirtualDir GetContainingDirectory() const override { return parent; } diff --git a/src/core/file_sys/vfs_vector.cpp b/src/core/file_sys/vfs_vector.cpp index 75fc04302..c1ec1e645 100644 --- a/src/core/file_sys/vfs_vector.cpp +++ b/src/core/file_sys/vfs_vector.cpp @@ -25,7 +25,7 @@ bool VectorVfsFile::Resize(size_t new_size) { return true; } -std::shared_ptr<VfsDirectory> VectorVfsFile::GetContainingDirectory() const { +VirtualDir VectorVfsFile::GetContainingDirectory() const { return parent; } @@ -68,11 +68,11 @@ VectorVfsDirectory::VectorVfsDirectory(std::vector<VirtualFile> files_, VectorVfsDirectory::~VectorVfsDirectory() = default; -std::vector<std::shared_ptr<VfsFile>> VectorVfsDirectory::GetFiles() const { +std::vector<VirtualFile> VectorVfsDirectory::GetFiles() const { return files; } -std::vector<std::shared_ptr<VfsDirectory>> VectorVfsDirectory::GetSubdirectories() const { +std::vector<VirtualDir> VectorVfsDirectory::GetSubdirectories() const { return dirs; } @@ -88,7 +88,7 @@ std::string VectorVfsDirectory::GetName() const { return name; } -std::shared_ptr<VfsDirectory> VectorVfsDirectory::GetParentDirectory() const { +VirtualDir VectorVfsDirectory::GetParentDirectory() const { return parent; } @@ -116,11 +116,11 @@ bool VectorVfsDirectory::Rename(std::string_view name_) { return true; } -std::shared_ptr<VfsDirectory> VectorVfsDirectory::CreateSubdirectory(std::string_view name) { +VirtualDir VectorVfsDirectory::CreateSubdirectory(std::string_view name) { return nullptr; } -std::shared_ptr<VfsFile> VectorVfsDirectory::CreateFile(std::string_view name) { +VirtualFile VectorVfsDirectory::CreateFile(std::string_view name) { return nullptr; } diff --git a/src/core/file_sys/vfs_vector.h b/src/core/file_sys/vfs_vector.h index 95d3da2f2..2aff9ca34 100644 --- a/src/core/file_sys/vfs_vector.h +++ b/src/core/file_sys/vfs_vector.h @@ -17,9 +17,9 @@ namespace FileSys { template <std::size_t size> class ArrayVfsFile : public VfsFile { public: - explicit ArrayVfsFile(const std::array<u8, size>& data, std::string name = "", - VirtualDir parent = nullptr) - : data(data), name(std::move(name)), parent(std::move(parent)) {} + explicit ArrayVfsFile(const std::array<u8, size>& data_, std::string name_ = "", + VirtualDir parent_ = nullptr) + : data(data_), name(std::move(name_)), parent(std::move(parent_)) {} std::string GetName() const override { return name; @@ -33,7 +33,7 @@ public: return false; } - std::shared_ptr<VfsDirectory> GetContainingDirectory() const override { + VirtualDir GetContainingDirectory() const override { return parent; } @@ -51,12 +51,12 @@ public: return read; } - std::size_t Write(const u8* data, std::size_t length, std::size_t offset) override { + std::size_t Write(const u8* data_, std::size_t length, std::size_t offset) override { return 0; } - bool Rename(std::string_view name) override { - this->name = name; + bool Rename(std::string_view new_name) override { + name = new_name; return true; } @@ -82,7 +82,7 @@ public: std::string GetName() const override; std::size_t GetSize() const override; bool Resize(std::size_t new_size) override; - std::shared_ptr<VfsDirectory> GetContainingDirectory() const override; + VirtualDir GetContainingDirectory() const override; bool IsWritable() const override; bool IsReadable() const override; std::size_t Read(u8* data, std::size_t length, std::size_t offset) const override; @@ -106,17 +106,17 @@ public: VirtualDir parent = nullptr); ~VectorVfsDirectory() override; - std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; - std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; + std::vector<VirtualFile> GetFiles() const override; + std::vector<VirtualDir> GetSubdirectories() const override; bool IsWritable() const override; bool IsReadable() const override; std::string GetName() const override; - std::shared_ptr<VfsDirectory> GetParentDirectory() const override; + VirtualDir GetParentDirectory() const override; bool DeleteSubdirectory(std::string_view name) override; bool DeleteFile(std::string_view name) override; bool Rename(std::string_view name) override; - std::shared_ptr<VfsDirectory> CreateSubdirectory(std::string_view name) override; - std::shared_ptr<VfsFile> CreateFile(std::string_view name) override; + VirtualDir CreateSubdirectory(std::string_view name) override; + VirtualFile CreateFile(std::string_view name) override; virtual void AddFile(VirtualFile file); virtual void AddDirectory(VirtualDir dir); diff --git a/src/core/file_sys/xts_archive.cpp b/src/core/file_sys/xts_archive.cpp index 24c58e7ae..814fd5680 100644 --- a/src/core/file_sys/xts_archive.cpp +++ b/src/core/file_sys/xts_archive.cpp @@ -152,11 +152,11 @@ NAXContentType NAX::GetContentType() const { return type; } -std::vector<std::shared_ptr<VfsFile>> NAX::GetFiles() const { +std::vector<VirtualFile> NAX::GetFiles() const { return {dec_file}; } -std::vector<std::shared_ptr<VfsDirectory>> NAX::GetSubdirectories() const { +std::vector<VirtualDir> NAX::GetSubdirectories() const { return {}; } @@ -164,7 +164,7 @@ std::string NAX::GetName() const { return file->GetName(); } -std::shared_ptr<VfsDirectory> NAX::GetParentDirectory() const { +VirtualDir NAX::GetParentDirectory() const { return file->GetContainingDirectory(); } diff --git a/src/core/file_sys/xts_archive.h b/src/core/file_sys/xts_archive.h index c472e226e..63a032b68 100644 --- a/src/core/file_sys/xts_archive.h +++ b/src/core/file_sys/xts_archive.h @@ -47,13 +47,13 @@ public: NAXContentType GetContentType() const; - std::vector<std::shared_ptr<VfsFile>> GetFiles() const override; + std::vector<VirtualFile> GetFiles() const override; - std::vector<std::shared_ptr<VfsDirectory>> GetSubdirectories() const override; + std::vector<VirtualDir> GetSubdirectories() const override; std::string GetName() const override; - std::shared_ptr<VfsDirectory> GetParentDirectory() const override; + VirtualDir GetParentDirectory() const override; private: Loader::ResultStatus Parse(std::string_view path); diff --git a/src/core/frontend/applets/controller.cpp b/src/core/frontend/applets/controller.cpp index 4505da758..03bbedf8b 100644 --- a/src/core/frontend/applets/controller.cpp +++ b/src/core/frontend/applets/controller.cpp @@ -4,7 +4,6 @@ #include "common/assert.h" #include "common/logging/log.h" -#include "core/core.h" #include "core/frontend/applets/controller.h" #include "core/hle/service/hid/controllers/npad.h" #include "core/hle/service/hid/hid.h" @@ -14,32 +13,33 @@ namespace Core::Frontend { ControllerApplet::~ControllerApplet() = default; +DefaultControllerApplet::DefaultControllerApplet(Service::SM::ServiceManager& service_manager_) + : service_manager{service_manager_} {} + DefaultControllerApplet::~DefaultControllerApplet() = default; void DefaultControllerApplet::ReconfigureControllers(std::function<void()> callback, - ControllerParameters parameters) const { + const ControllerParameters& parameters) const { LOG_INFO(Service_HID, "called, deducing the best configuration based on the given parameters!"); auto& npad = - Core::System::GetInstance() - .ServiceManager() - .GetService<Service::HID::Hid>("hid") + service_manager.GetService<Service::HID::Hid>("hid") ->GetAppletResource() ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad); - auto& players = Settings::values.players; + auto& players = Settings::values.players.GetValue(); const std::size_t min_supported_players = parameters.enable_single_mode ? 1 : parameters.min_players; // Disconnect Handheld first. - npad.DisconnectNPadAtIndex(8); + npad.DisconnectNpadAtIndex(8); // Deduce the best configuration based on the input parameters. for (std::size_t index = 0; index < players.size() - 2; ++index) { // First, disconnect all controllers regardless of the value of keep_controllers_connected. // This makes it easy to connect the desired controllers. - npad.DisconnectNPadAtIndex(index); + npad.DisconnectNpadAtIndex(index); // Only connect the minimum number of required players. if (index >= min_supported_players) { @@ -66,7 +66,7 @@ void DefaultControllerApplet::ReconfigureControllers(std::function<void()> callb npad.MapSettingsTypeToNPad(Settings::ControllerType::RightJoycon), index); } } else if (index == 0 && parameters.enable_single_mode && parameters.allow_handheld && - !Settings::values.use_docked_mode) { + !Settings::values.use_docked_mode.GetValue()) { // We should *never* reach here under any normal circumstances. npad.AddNewControllerAt(npad.MapSettingsTypeToNPad(Settings::ControllerType::Handheld), index); diff --git a/src/core/frontend/applets/controller.h b/src/core/frontend/applets/controller.h index a227f15cd..dff71d8d9 100644 --- a/src/core/frontend/applets/controller.h +++ b/src/core/frontend/applets/controller.h @@ -8,6 +8,10 @@ #include "common/common_types.h" +namespace Service::SM { +class ServiceManager; +} + namespace Core::Frontend { using BorderColor = std::array<u8, 4>; @@ -34,15 +38,19 @@ public: virtual ~ControllerApplet(); virtual void ReconfigureControllers(std::function<void()> callback, - ControllerParameters parameters) const = 0; + const ControllerParameters& parameters) const = 0; }; class DefaultControllerApplet final : public ControllerApplet { public: + explicit DefaultControllerApplet(Service::SM::ServiceManager& service_manager_); ~DefaultControllerApplet() override; void ReconfigureControllers(std::function<void()> callback, - ControllerParameters parameters) const override; + const ControllerParameters& parameters) const override; + +private: + Service::SM::ServiceManager& service_manager; }; } // namespace Core::Frontend diff --git a/src/core/frontend/applets/error.cpp b/src/core/frontend/applets/error.cpp index 4002a9211..dceb20ff8 100644 --- a/src/core/frontend/applets/error.cpp +++ b/src/core/frontend/applets/error.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/logging/log.h" #include "core/frontend/applets/error.h" namespace Core::Frontend { @@ -10,7 +11,7 @@ ErrorApplet::~ErrorApplet() = default; void DefaultErrorApplet::ShowError(ResultCode error, std::function<void()> finished) const { LOG_CRITICAL(Service_Fatal, "Application requested error display: {:04}-{:04} (raw={:08X})", - static_cast<u32>(error.module.Value()), error.description.Value(), error.raw); + error.module.Value(), error.description.Value(), error.raw); } void DefaultErrorApplet::ShowErrorWithTimestamp(ResultCode error, std::chrono::seconds time, @@ -18,7 +19,7 @@ void DefaultErrorApplet::ShowErrorWithTimestamp(ResultCode error, std::chrono::s LOG_CRITICAL( Service_Fatal, "Application requested error display: {:04X}-{:04X} (raw={:08X}) with timestamp={:016X}", - static_cast<u32>(error.module.Value()), error.description.Value(), error.raw, time.count()); + error.module.Value(), error.description.Value(), error.raw, time.count()); } void DefaultErrorApplet::ShowCustomErrorText(ResultCode error, std::string main_text, @@ -26,7 +27,7 @@ void DefaultErrorApplet::ShowCustomErrorText(ResultCode error, std::string main_ std::function<void()> finished) const { LOG_CRITICAL(Service_Fatal, "Application requested custom error with error_code={:04X}-{:04X} (raw={:08X})", - static_cast<u32>(error.module.Value()), error.description.Value(), error.raw); + error.module.Value(), error.description.Value(), error.raw); LOG_CRITICAL(Service_Fatal, " Main Text: {}", main_text); LOG_CRITICAL(Service_Fatal, " Detail Text: {}", detail_text); } diff --git a/src/core/frontend/applets/general_frontend.cpp b/src/core/frontend/applets/general_frontend.cpp index c30b36de7..7483ffb76 100644 --- a/src/core/frontend/applets/general_frontend.cpp +++ b/src/core/frontend/applets/general_frontend.cpp @@ -53,72 +53,4 @@ void DefaultPhotoViewerApplet::ShowAllPhotos(std::function<void()> finished) con finished(); } -ECommerceApplet::~ECommerceApplet() = default; - -DefaultECommerceApplet::~DefaultECommerceApplet() = default; - -void DefaultECommerceApplet::ShowApplicationInformation( - std::function<void()> finished, u64 title_id, std::optional<u128> user_id, - std::optional<bool> full_display, std::optional<std::string> extra_parameter) { - const auto value = user_id.value_or(u128{}); - LOG_INFO(Service_AM, - "Application requested frontend show application information for EShop, " - "title_id={:016X}, user_id={:016X}{:016X}, full_display={}, extra_parameter={}", - title_id, value[1], value[0], - full_display.has_value() ? fmt::format("{}", *full_display) : "null", - extra_parameter.value_or("null")); - finished(); -} - -void DefaultECommerceApplet::ShowAddOnContentList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id, - std::optional<bool> full_display) { - const auto value = user_id.value_or(u128{}); - LOG_INFO(Service_AM, - "Application requested frontend show add on content list for EShop, " - "title_id={:016X}, user_id={:016X}{:016X}, full_display={}", - title_id, value[1], value[0], - full_display.has_value() ? fmt::format("{}", *full_display) : "null"); - finished(); -} - -void DefaultECommerceApplet::ShowSubscriptionList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id) { - const auto value = user_id.value_or(u128{}); - LOG_INFO(Service_AM, - "Application requested frontend show subscription list for EShop, title_id={:016X}, " - "user_id={:016X}{:016X}", - title_id, value[1], value[0]); - finished(); -} - -void DefaultECommerceApplet::ShowConsumableItemList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id) { - const auto value = user_id.value_or(u128{}); - LOG_INFO( - Service_AM, - "Application requested frontend show consumable item list for EShop, title_id={:016X}, " - "user_id={:016X}{:016X}", - title_id, value[1], value[0]); - finished(); -} - -void DefaultECommerceApplet::ShowShopHome(std::function<void()> finished, u128 user_id, - bool full_display) { - LOG_INFO(Service_AM, - "Application requested frontend show home menu for EShop, user_id={:016X}{:016X}, " - "full_display={}", - user_id[1], user_id[0], full_display); - finished(); -} - -void DefaultECommerceApplet::ShowSettings(std::function<void()> finished, u128 user_id, - bool full_display) { - LOG_INFO(Service_AM, - "Application requested frontend show settings menu for EShop, user_id={:016X}{:016X}, " - "full_display={}", - user_id[1], user_id[0], full_display); - finished(); -} - } // namespace Core::Frontend diff --git a/src/core/frontend/applets/general_frontend.h b/src/core/frontend/applets/general_frontend.h index 4b63f828e..b713b14ee 100644 --- a/src/core/frontend/applets/general_frontend.h +++ b/src/core/frontend/applets/general_frontend.h @@ -58,55 +58,4 @@ public: void ShowAllPhotos(std::function<void()> finished) const override; }; -class ECommerceApplet { -public: - virtual ~ECommerceApplet(); - - // Shows a page with application icons, description, name, and price. - virtual void ShowApplicationInformation(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id = {}, - std::optional<bool> full_display = {}, - std::optional<std::string> extra_parameter = {}) = 0; - - // Shows a page with all of the add on content available for a game, with name, description, and - // price. - virtual void ShowAddOnContentList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id = {}, - std::optional<bool> full_display = {}) = 0; - - // Shows a page with all of the subscriptions (recurring payments) for a game, with name, - // description, price, and renewal period. - virtual void ShowSubscriptionList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id = {}) = 0; - - // Shows a page with a list of any additional game related purchasable items (DLC, - // subscriptions, etc) for a particular game, with name, description, type, and price. - virtual void ShowConsumableItemList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id = {}) = 0; - - // Shows the home page of the shop. - virtual void ShowShopHome(std::function<void()> finished, u128 user_id, bool full_display) = 0; - - // Shows the user settings page of the shop. - virtual void ShowSettings(std::function<void()> finished, u128 user_id, bool full_display) = 0; -}; - -class DefaultECommerceApplet : public ECommerceApplet { -public: - ~DefaultECommerceApplet() override; - - void ShowApplicationInformation(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id, std::optional<bool> full_display, - std::optional<std::string> extra_parameter) override; - void ShowAddOnContentList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id, - std::optional<bool> full_display) override; - void ShowSubscriptionList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id) override; - void ShowConsumableItemList(std::function<void()> finished, u64 title_id, - std::optional<u128> user_id) override; - void ShowShopHome(std::function<void()> finished, u128 user_id, bool full_display) override; - void ShowSettings(std::function<void()> finished, u128 user_id, bool full_display) override; -}; - } // namespace Core::Frontend diff --git a/src/core/frontend/applets/web_browser.cpp b/src/core/frontend/applets/web_browser.cpp index 528295ffc..50db6a654 100644 --- a/src/core/frontend/applets/web_browser.cpp +++ b/src/core/frontend/applets/web_browser.cpp @@ -11,14 +11,22 @@ WebBrowserApplet::~WebBrowserApplet() = default; DefaultWebBrowserApplet::~DefaultWebBrowserApplet() = default; -void DefaultWebBrowserApplet::OpenPageLocal(std::string_view filename, - std::function<void()> unpack_romfs_callback, - std::function<void()> finished_callback) { - LOG_INFO(Service_AM, - "(STUBBED) called - No suitable web browser implementation found to open website page " - "at '{}'!", - filename); - finished_callback(); +void DefaultWebBrowserApplet::OpenLocalWebPage( + std::string_view local_url, std::function<void()> extract_romfs_callback, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> callback) const { + LOG_WARNING(Service_AM, "(STUBBED) called, backend requested to open local web page at {}", + local_url); + + callback(Service::AM::Applets::WebExitReason::WindowClosed, "http://localhost/"); +} + +void DefaultWebBrowserApplet::OpenExternalWebPage( + std::string_view external_url, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> callback) const { + LOG_WARNING(Service_AM, "(STUBBED) called, backend requested to open external web page at {}", + external_url); + + callback(Service::AM::Applets::WebExitReason::WindowClosed, "http://localhost/"); } } // namespace Core::Frontend diff --git a/src/core/frontend/applets/web_browser.h b/src/core/frontend/applets/web_browser.h index 110e33bc4..1c5ef19a9 100644 --- a/src/core/frontend/applets/web_browser.h +++ b/src/core/frontend/applets/web_browser.h @@ -7,22 +7,34 @@ #include <functional> #include <string_view> +#include "core/hle/service/am/applets/web_types.h" + namespace Core::Frontend { class WebBrowserApplet { public: virtual ~WebBrowserApplet(); - virtual void OpenPageLocal(std::string_view url, std::function<void()> unpack_romfs_callback, - std::function<void()> finished_callback) = 0; + virtual void OpenLocalWebPage( + std::string_view local_url, std::function<void()> extract_romfs_callback, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> callback) const = 0; + + virtual void OpenExternalWebPage( + std::string_view external_url, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> callback) const = 0; }; class DefaultWebBrowserApplet final : public WebBrowserApplet { public: ~DefaultWebBrowserApplet() override; - void OpenPageLocal(std::string_view url, std::function<void()> unpack_romfs_callback, - std::function<void()> finished_callback) override; + void OpenLocalWebPage(std::string_view local_url, std::function<void()> extract_romfs_callback, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> + callback) const override; + + void OpenExternalWebPage(std::string_view external_url, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> + callback) const override; }; } // namespace Core::Frontend diff --git a/src/core/frontend/emu_window.cpp b/src/core/frontend/emu_window.cpp index 9a081fbd4..8c1193894 100644 --- a/src/core/frontend/emu_window.cpp +++ b/src/core/frontend/emu_window.cpp @@ -84,10 +84,12 @@ void EmuWindow::TouchPressed(unsigned framebuffer_x, unsigned framebuffer_y) { return; std::lock_guard guard{touch_state->mutex}; - touch_state->touch_x = static_cast<float>(framebuffer_x - framebuffer_layout.screen.left) / - (framebuffer_layout.screen.right - framebuffer_layout.screen.left); - touch_state->touch_y = static_cast<float>(framebuffer_y - framebuffer_layout.screen.top) / - (framebuffer_layout.screen.bottom - framebuffer_layout.screen.top); + touch_state->touch_x = + static_cast<float>(framebuffer_x - framebuffer_layout.screen.left) / + static_cast<float>(framebuffer_layout.screen.right - framebuffer_layout.screen.left); + touch_state->touch_y = + static_cast<float>(framebuffer_y - framebuffer_layout.screen.top) / + static_cast<float>(framebuffer_layout.screen.bottom - framebuffer_layout.screen.top); touch_state->touch_pressed = true; } diff --git a/src/core/frontend/emu_window.h b/src/core/frontend/emu_window.h index 3e8780243..276d2b906 100644 --- a/src/core/frontend/emu_window.h +++ b/src/core/frontend/emu_window.h @@ -102,8 +102,8 @@ public: float render_surface_scale = 1.0f; }; - /// Polls window events - virtual void PollEvents() = 0; + /// Called from GPU thread when a frame is displayed. + virtual void OnFrameDisplayed() {} /** * Returns a GraphicsContext that the frontend provides to be used for rendering. diff --git a/src/core/frontend/framebuffer_layout.cpp b/src/core/frontend/framebuffer_layout.cpp index c1fbc235b..b9a270a55 100644 --- a/src/core/frontend/framebuffer_layout.cpp +++ b/src/core/frontend/framebuffer_layout.cpp @@ -14,8 +14,8 @@ namespace Layout { template <class T> static Common::Rectangle<T> MaxRectangle(Common::Rectangle<T> window_area, float screen_aspect_ratio) { - float scale = std::min(static_cast<float>(window_area.GetWidth()), - window_area.GetHeight() / screen_aspect_ratio); + const float scale = std::min(static_cast<float>(window_area.GetWidth()), + static_cast<float>(window_area.GetHeight()) / screen_aspect_ratio); return Common::Rectangle<T>{0, 0, static_cast<T>(std::round(scale)), static_cast<T>(std::round(scale * screen_aspect_ratio))}; } @@ -27,7 +27,7 @@ FramebufferLayout DefaultFrameLayout(u32 width, u32 height) { // so just calculate them both even if the other isn't showing. FramebufferLayout res{width, height, false, {}}; - const float window_aspect_ratio = static_cast<float>(height) / width; + const float window_aspect_ratio = static_cast<float>(height) / static_cast<float>(width); const float emulation_aspect_ratio = EmulationAspectRatio( static_cast<AspectRatio>(Settings::values.aspect_ratio.GetValue()), window_aspect_ratio); @@ -47,7 +47,7 @@ FramebufferLayout DefaultFrameLayout(u32 width, u32 height) { FramebufferLayout FrameLayoutFromResolutionScale(u32 res_scale) { u32 width, height; - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { width = ScreenDocked::Width * res_scale; height = ScreenDocked::Height * res_scale; } else { diff --git a/src/core/frontend/input.h b/src/core/frontend/input.h index 9da0d2829..de51a754e 100644 --- a/src/core/frontend/input.h +++ b/src/core/frontend/input.h @@ -30,7 +30,12 @@ public: virtual StatusType GetStatus() const { return {}; } - virtual bool GetAnalogDirectionStatus(AnalogDirection direction) const { + virtual bool GetAnalogDirectionStatus([[maybe_unused]] AnalogDirection direction) const { + return {}; + } + virtual bool SetRumblePlay([[maybe_unused]] f32 amp_low, [[maybe_unused]] f32 freq_low, + [[maybe_unused]] f32 amp_high, + [[maybe_unused]] f32 freq_high) const { return {}; } }; @@ -119,6 +124,13 @@ using ButtonDevice = InputDevice<bool>; using AnalogDevice = InputDevice<std::tuple<float, float>>; /** + * A vibration device is an input device that returns an unsigned byte as status. + * It represents whether the vibration device supports vibration or not. + * If the status returns 1, it supports vibration. Otherwise, it does not support vibration. + */ +using VibrationDevice = InputDevice<u8>; + +/** * A motion status is an object that returns a tuple of accelerometer state vector, * gyroscope state vector, rotation state vector and orientation state matrix. * @@ -151,10 +163,15 @@ using MotionStatus = std::tuple<Common::Vec3<float>, Common::Vec3<float>, Common using MotionDevice = InputDevice<MotionStatus>; /** - * A touch device is an input device that returns a tuple of two floats and a bool. The floats are + * A touch status is an object that returns a tuple of two floats and a bool. The floats are * x and y coordinates in the range 0.0 - 1.0, and the bool indicates whether it is pressed. */ -using TouchDevice = InputDevice<std::tuple<float, float, bool>>; +using TouchStatus = std::tuple<float, float, bool>; + +/** + * A touch device is an input device that returns a touch status object + */ +using TouchDevice = InputDevice<TouchStatus>; /** * A mouse device is an input device that returns a tuple of two floats and four ints. diff --git a/src/core/frontend/input_interpreter.cpp b/src/core/frontend/input_interpreter.cpp new file mode 100644 index 000000000..66ae506cd --- /dev/null +++ b/src/core/frontend/input_interpreter.cpp @@ -0,0 +1,45 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/core.h" +#include "core/frontend/input_interpreter.h" +#include "core/hle/service/hid/controllers/npad.h" +#include "core/hle/service/hid/hid.h" +#include "core/hle/service/sm/sm.h" + +InputInterpreter::InputInterpreter(Core::System& system) + : npad{system.ServiceManager() + .GetService<Service::HID::Hid>("hid") + ->GetAppletResource() + ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad)} {} + +InputInterpreter::~InputInterpreter() = default; + +void InputInterpreter::PollInput() { + const u32 button_state = npad.GetAndResetPressState(); + + previous_index = current_index; + current_index = (current_index + 1) % button_states.size(); + + button_states[current_index] = button_state; +} + +bool InputInterpreter::IsButtonPressedOnce(HIDButton button) const { + const bool current_press = + (button_states[current_index] & (1U << static_cast<u8>(button))) != 0; + const bool previous_press = + (button_states[previous_index] & (1U << static_cast<u8>(button))) != 0; + + return current_press && !previous_press; +} + +bool InputInterpreter::IsButtonHeld(HIDButton button) const { + u32 held_buttons{button_states[0]}; + + for (std::size_t i = 1; i < button_states.size(); ++i) { + held_buttons &= button_states[i]; + } + + return (held_buttons & (1U << static_cast<u8>(button))) != 0; +} diff --git a/src/core/frontend/input_interpreter.h b/src/core/frontend/input_interpreter.h new file mode 100644 index 000000000..fea9aebe6 --- /dev/null +++ b/src/core/frontend/input_interpreter.h @@ -0,0 +1,120 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> + +#include "common/common_types.h" + +namespace Core { +class System; +} + +namespace Service::HID { +class Controller_NPad; +} + +enum class HIDButton : u8 { + A, + B, + X, + Y, + LStick, + RStick, + L, + R, + ZL, + ZR, + Plus, + Minus, + + DLeft, + DUp, + DRight, + DDown, + + LStickLeft, + LStickUp, + LStickRight, + LStickDown, + + RStickLeft, + RStickUp, + RStickRight, + RStickDown, + + LeftSL, + LeftSR, + + RightSL, + RightSR, +}; + +/** + * The InputInterpreter class interfaces with HID to retrieve button press states. + * Input is intended to be polled every 50ms so that a button is considered to be + * held down after 400ms has elapsed since the initial button press and subsequent + * repeated presses occur every 50ms. + */ +class InputInterpreter { +public: + explicit InputInterpreter(Core::System& system); + virtual ~InputInterpreter(); + + /// Gets a button state from HID and inserts it into the array of button states. + void PollInput(); + + /** + * The specified button is considered to be pressed once + * if it is currently pressed and not pressed previously. + * + * @param button The button to check. + * + * @returns True when the button is pressed once. + */ + [[nodiscard]] bool IsButtonPressedOnce(HIDButton button) const; + + /** + * Checks whether any of the buttons in the parameter list is pressed once. + * + * @tparam HIDButton The buttons to check. + * + * @returns True when at least one of the buttons is pressed once. + */ + template <HIDButton... T> + [[nodiscard]] bool IsAnyButtonPressedOnce() { + return (IsButtonPressedOnce(T) || ...); + } + + /** + * The specified button is considered to be held down if it is pressed in all 9 button states. + * + * @param button The button to check. + * + * @returns True when the button is held down. + */ + [[nodiscard]] bool IsButtonHeld(HIDButton button) const; + + /** + * Checks whether any of the buttons in the parameter list is held down. + * + * @tparam HIDButton The buttons to check. + * + * @returns True when at least one of the buttons is held down. + */ + template <HIDButton... T> + [[nodiscard]] bool IsAnyButtonHeld() { + return (IsButtonHeld(T) || ...); + } + +private: + Service::HID::Controller_NPad& npad; + + /// Stores 9 consecutive button states polled from HID. + std::array<u32, 9> button_states{}; + + std::size_t previous_index{}; + std::size_t current_index{}; +}; diff --git a/src/core/gdbstub/gdbstub.cpp b/src/core/gdbstub/gdbstub.cpp deleted file mode 100644 index 79f22a403..000000000 --- a/src/core/gdbstub/gdbstub.cpp +++ /dev/null @@ -1,1397 +0,0 @@ -// Copyright 2013 Dolphin Emulator Project -// Licensed under GPLv2+ -// Refer to the license.txt file included. - -// Originally written by Sven Peter <sven@fail0verflow.com> for anergistic. - -#include <algorithm> -#include <atomic> -#include <climits> -#include <csignal> -#include <cstdarg> -#include <cstdio> -#include <cstring> -#include <map> -#include <numeric> -#include <fcntl.h> - -#ifdef _WIN32 -#include <winsock2.h> -// winsock2.h needs to be included first to prevent winsock.h being included by other includes -#include <io.h> -#include <iphlpapi.h> -#include <ws2tcpip.h> -#define SHUT_RDWR 2 -#else -#include <netinet/in.h> -#include <sys/select.h> -#include <sys/socket.h> -#include <sys/un.h> -#include <unistd.h> -#endif - -#include "common/logging/log.h" -#include "common/string_util.h" -#include "common/swap.h" -#include "core/arm/arm_interface.h" -#include "core/core.h" -#include "core/gdbstub/gdbstub.h" -#include "core/hle/kernel/memory/page_table.h" -#include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" -#include "core/loader/loader.h" -#include "core/memory.h" - -namespace GDBStub { -namespace { -constexpr int GDB_BUFFER_SIZE = 10000; - -constexpr char GDB_STUB_START = '$'; -constexpr char GDB_STUB_END = '#'; -constexpr char GDB_STUB_ACK = '+'; -constexpr char GDB_STUB_NACK = '-'; - -#ifndef SIGTRAP -constexpr u32 SIGTRAP = 5; -#endif - -#ifndef SIGTERM -constexpr u32 SIGTERM = 15; -#endif - -#ifndef MSG_WAITALL -constexpr u32 MSG_WAITALL = 8; -#endif - -constexpr u32 LR_REGISTER = 30; -constexpr u32 SP_REGISTER = 31; -constexpr u32 PC_REGISTER = 32; -constexpr u32 PSTATE_REGISTER = 33; -constexpr u32 UC_ARM64_REG_Q0 = 34; -constexpr u32 FPCR_REGISTER = 66; - -// For sample XML files see the GDB source /gdb/features -// GDB also wants the l character at the start -// This XML defines what the registers are for this specific ARM device -constexpr char target_xml[] = - R"(l<?xml version="1.0"?> -<!DOCTYPE target SYSTEM "gdb-target.dtd"> -<target version="1.0"> - <feature name="org.gnu.gdb.aarch64.core"> - <reg name="x0" bitsize="64"/> - <reg name="x1" bitsize="64"/> - <reg name="x2" bitsize="64"/> - <reg name="x3" bitsize="64"/> - <reg name="x4" bitsize="64"/> - <reg name="x5" bitsize="64"/> - <reg name="x6" bitsize="64"/> - <reg name="x7" bitsize="64"/> - <reg name="x8" bitsize="64"/> - <reg name="x9" bitsize="64"/> - <reg name="x10" bitsize="64"/> - <reg name="x11" bitsize="64"/> - <reg name="x12" bitsize="64"/> - <reg name="x13" bitsize="64"/> - <reg name="x14" bitsize="64"/> - <reg name="x15" bitsize="64"/> - <reg name="x16" bitsize="64"/> - <reg name="x17" bitsize="64"/> - <reg name="x18" bitsize="64"/> - <reg name="x19" bitsize="64"/> - <reg name="x20" bitsize="64"/> - <reg name="x21" bitsize="64"/> - <reg name="x22" bitsize="64"/> - <reg name="x23" bitsize="64"/> - <reg name="x24" bitsize="64"/> - <reg name="x25" bitsize="64"/> - <reg name="x26" bitsize="64"/> - <reg name="x27" bitsize="64"/> - <reg name="x28" bitsize="64"/> - <reg name="x29" bitsize="64"/> - <reg name="x30" bitsize="64"/> - <reg name="sp" bitsize="64" type="data_ptr"/> - - <reg name="pc" bitsize="64" type="code_ptr"/> - - <flags id="pstate_flags" size="4"> - <field name="SP" start="0" end="0"/> - <field name="" start="1" end="1"/> - <field name="EL" start="2" end="3"/> - <field name="nRW" start="4" end="4"/> - <field name="" start="5" end="5"/> - <field name="F" start="6" end="6"/> - <field name="I" start="7" end="7"/> - <field name="A" start="8" end="8"/> - <field name="D" start="9" end="9"/> - - <field name="IL" start="20" end="20"/> - <field name="SS" start="21" end="21"/> - - <field name="V" start="28" end="28"/> - <field name="C" start="29" end="29"/> - <field name="Z" start="30" end="30"/> - <field name="N" start="31" end="31"/> - </flags> - <reg name="pstate" bitsize="32" type="pstate_flags"/> - </feature> - <feature name="org.gnu.gdb.aarch64.fpu"> - </feature> -</target> -)"; - -int gdbserver_socket = -1; -bool defer_start = false; - -u8 command_buffer[GDB_BUFFER_SIZE]; -u32 command_length; - -u32 latest_signal = 0; -bool memory_break = false; - -Kernel::Thread* current_thread = nullptr; -u32 current_core = 0; - -// Binding to a port within the reserved ports range (0-1023) requires root permissions, -// so default to a port outside of that range. -u16 gdbstub_port = 24689; - -bool halt_loop = true; -bool step_loop = false; -bool send_trap = false; - -// If set to false, the server will never be started and no -// gdbstub-related functions will be executed. -std::atomic<bool> server_enabled(false); - -#ifdef _WIN32 -WSADATA InitData; -#endif - -struct Breakpoint { - bool active; - VAddr addr; - u64 len; - std::array<u8, 4> inst; -}; - -using BreakpointMap = std::map<VAddr, Breakpoint>; -BreakpointMap breakpoints_execute; -BreakpointMap breakpoints_read; -BreakpointMap breakpoints_write; - -struct Module { - std::string name; - VAddr beg; - VAddr end; -}; - -std::vector<Module> modules; -} // Anonymous namespace - -void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext) { - Module module; - if (add_elf_ext) { - Common::SplitPath(name, nullptr, &module.name, nullptr); - module.name += ".elf"; - } else { - module.name = std::move(name); - } - module.beg = beg; - module.end = end; - modules.push_back(std::move(module)); -} - -static Kernel::Thread* FindThreadById(s64 id) { - const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList(); - for (auto& thread : threads) { - if (thread->GetThreadID() == static_cast<u64>(id)) { - current_core = thread->GetProcessorID(); - return thread.get(); - } - } - return nullptr; -} - -static u64 RegRead(std::size_t id, Kernel::Thread* thread = nullptr) { - if (!thread) { - return 0; - } - - const auto& thread_context = thread->GetContext64(); - - if (id < SP_REGISTER) { - return thread_context.cpu_registers[id]; - } else if (id == SP_REGISTER) { - return thread_context.sp; - } else if (id == PC_REGISTER) { - return thread_context.pc; - } else if (id == PSTATE_REGISTER) { - return thread_context.pstate; - } else if (id > PSTATE_REGISTER && id < FPCR_REGISTER) { - return thread_context.vector_registers[id - UC_ARM64_REG_Q0][0]; - } else { - return 0; - } -} - -static void RegWrite(std::size_t id, u64 val, Kernel::Thread* thread = nullptr) { - if (!thread) { - return; - } - - auto& thread_context = thread->GetContext64(); - - if (id < SP_REGISTER) { - thread_context.cpu_registers[id] = val; - } else if (id == SP_REGISTER) { - thread_context.sp = val; - } else if (id == PC_REGISTER) { - thread_context.pc = val; - } else if (id == PSTATE_REGISTER) { - thread_context.pstate = static_cast<u32>(val); - } else if (id > PSTATE_REGISTER && id < FPCR_REGISTER) { - thread_context.vector_registers[id - (PSTATE_REGISTER + 1)][0] = val; - } -} - -static u128 FpuRead(std::size_t id, Kernel::Thread* thread = nullptr) { - if (!thread) { - return u128{0}; - } - - auto& thread_context = thread->GetContext64(); - - if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) { - return thread_context.vector_registers[id - UC_ARM64_REG_Q0]; - } else if (id == FPCR_REGISTER) { - return u128{thread_context.fpcr, 0}; - } else { - return u128{0}; - } -} - -static void FpuWrite(std::size_t id, u128 val, Kernel::Thread* thread = nullptr) { - if (!thread) { - return; - } - - auto& thread_context = thread->GetContext64(); - - if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) { - thread_context.vector_registers[id - UC_ARM64_REG_Q0] = val; - } else if (id == FPCR_REGISTER) { - thread_context.fpcr = static_cast<u32>(val[0]); - } -} - -/** - * Turns hex string character into the equivalent byte. - * - * @param hex Input hex character to be turned into byte. - */ -static u8 HexCharToValue(u8 hex) { - if (hex >= '0' && hex <= '9') { - return hex - '0'; - } else if (hex >= 'a' && hex <= 'f') { - return hex - 'a' + 0xA; - } else if (hex >= 'A' && hex <= 'F') { - return hex - 'A' + 0xA; - } - - LOG_ERROR(Debug_GDBStub, "Invalid nibble: {} ({:02X})", hex, hex); - return 0; -} - -/** - * Turn nibble of byte into hex string character. - * - * @param n Nibble to be turned into hex character. - */ -static u8 NibbleToHex(u8 n) { - n &= 0xF; - if (n < 0xA) { - return '0' + n; - } else { - return 'a' + n - 0xA; - } -} - -/** - * Converts input hex string characters into an array of equivalent of u8 bytes. - * - * @param src Pointer to array of output hex string characters. - * @param len Length of src array. - */ -static u32 HexToInt(const u8* src, std::size_t len) { - u32 output = 0; - while (len-- > 0) { - output = (output << 4) | HexCharToValue(src[0]); - src++; - } - return output; -} - -/** - * Converts input hex string characters into an array of equivalent of u8 bytes. - * - * @param src Pointer to array of output hex string characters. - * @param len Length of src array. - */ -static u64 HexToLong(const u8* src, std::size_t len) { - u64 output = 0; - while (len-- > 0) { - output = (output << 4) | HexCharToValue(src[0]); - src++; - } - return output; -} - -/** - * Converts input array of u8 bytes into their equivalent hex string characters. - * - * @param dest Pointer to buffer to store output hex string characters. - * @param src Pointer to array of u8 bytes. - * @param len Length of src array. - */ -static void MemToGdbHex(u8* dest, const u8* src, std::size_t len) { - while (len-- > 0) { - u8 tmp = *src++; - *dest++ = NibbleToHex(tmp >> 4); - *dest++ = NibbleToHex(tmp); - } -} - -/** - * Converts input gdb-formatted hex string characters into an array of equivalent of u8 bytes. - * - * @param dest Pointer to buffer to store u8 bytes. - * @param src Pointer to array of output hex string characters. - * @param len Length of src array. - */ -static void GdbHexToMem(u8* dest, const u8* src, std::size_t len) { - while (len-- > 0) { - *dest++ = (HexCharToValue(src[0]) << 4) | HexCharToValue(src[1]); - src += 2; - } -} - -/** - * Convert a u32 into a gdb-formatted hex string. - * - * @param dest Pointer to buffer to store output hex string characters. - * @param v Value to convert. - */ -static void IntToGdbHex(u8* dest, u32 v) { - for (int i = 0; i < 8; i += 2) { - dest[i + 1] = NibbleToHex(static_cast<u8>(v >> (4 * i))); - dest[i] = NibbleToHex(static_cast<u8>(v >> (4 * (i + 1)))); - } -} - -/** - * Convert a u64 into a gdb-formatted hex string. - * - * @param dest Pointer to buffer to store output hex string characters. - * @param v Value to convert. - */ -static void LongToGdbHex(u8* dest, u64 v) { - for (int i = 0; i < 16; i += 2) { - dest[i + 1] = NibbleToHex(static_cast<u8>(v >> (4 * i))); - dest[i] = NibbleToHex(static_cast<u8>(v >> (4 * (i + 1)))); - } -} - -/** - * Convert a gdb-formatted hex string into a u32. - * - * @param src Pointer to hex string. - */ -static u32 GdbHexToInt(const u8* src) { - u32 output = 0; - - for (int i = 0; i < 8; i += 2) { - output = (output << 4) | HexCharToValue(src[7 - i - 1]); - output = (output << 4) | HexCharToValue(src[7 - i]); - } - - return output; -} - -/** - * Convert a gdb-formatted hex string into a u64. - * - * @param src Pointer to hex string. - */ -static u64 GdbHexToLong(const u8* src) { - u64 output = 0; - - for (int i = 0; i < 16; i += 2) { - output = (output << 4) | HexCharToValue(src[15 - i - 1]); - output = (output << 4) | HexCharToValue(src[15 - i]); - } - - return output; -} - -/** - * Convert a gdb-formatted hex string into a u128. - * - * @param src Pointer to hex string. - */ -static u128 GdbHexToU128(const u8* src) { - u128 output; - - for (int i = 0; i < 16; i += 2) { - output[0] = (output[0] << 4) | HexCharToValue(src[15 - i - 1]); - output[0] = (output[0] << 4) | HexCharToValue(src[15 - i]); - } - - for (int i = 0; i < 16; i += 2) { - output[1] = (output[1] << 4) | HexCharToValue(src[16 + 15 - i - 1]); - output[1] = (output[1] << 4) | HexCharToValue(src[16 + 15 - i]); - } - - return output; -} - -/// Read a byte from the gdb client. -static u8 ReadByte() { - u8 c; - std::size_t received_size = recv(gdbserver_socket, reinterpret_cast<char*>(&c), 1, MSG_WAITALL); - if (received_size != 1) { - LOG_ERROR(Debug_GDBStub, "recv failed: {}", received_size); - Shutdown(); - } - - return c; -} - -/// Calculate the checksum of the current command buffer. -static u8 CalculateChecksum(const u8* buffer, std::size_t length) { - return static_cast<u8>(std::accumulate(buffer, buffer + length, u8{0}, - [](u8 lhs, u8 rhs) { return u8(lhs + rhs); })); -} - -/** - * Get the map of breakpoints for a given breakpoint type. - * - * @param type Type of breakpoint map. - */ -static BreakpointMap& GetBreakpointMap(BreakpointType type) { - switch (type) { - case BreakpointType::Execute: - return breakpoints_execute; - case BreakpointType::Read: - return breakpoints_read; - case BreakpointType::Write: - return breakpoints_write; - default: - return breakpoints_read; - } -} - -/** - * Remove the breakpoint from the given address of the specified type. - * - * @param type Type of breakpoint. - * @param addr Address of breakpoint. - */ -static void RemoveBreakpoint(BreakpointType type, VAddr addr) { - BreakpointMap& p = GetBreakpointMap(type); - - const auto bp = p.find(addr); - if (bp == p.end()) { - return; - } - - LOG_DEBUG(Debug_GDBStub, "gdb: removed a breakpoint: {:016X} bytes at {:016X} of type {}", - bp->second.len, bp->second.addr, static_cast<int>(type)); - - if (type == BreakpointType::Execute) { - auto& system = Core::System::GetInstance(); - system.Memory().WriteBlock(bp->second.addr, bp->second.inst.data(), bp->second.inst.size()); - system.InvalidateCpuInstructionCaches(); - } - p.erase(addr); -} - -BreakpointAddress GetNextBreakpointFromAddress(VAddr addr, BreakpointType type) { - const BreakpointMap& p = GetBreakpointMap(type); - const auto next_breakpoint = p.lower_bound(addr); - BreakpointAddress breakpoint; - - if (next_breakpoint != p.end()) { - breakpoint.address = next_breakpoint->first; - breakpoint.type = type; - } else { - breakpoint.address = 0; - breakpoint.type = BreakpointType::None; - } - - return breakpoint; -} - -bool CheckBreakpoint(VAddr addr, BreakpointType type) { - if (!IsConnected()) { - return false; - } - - const BreakpointMap& p = GetBreakpointMap(type); - const auto bp = p.find(addr); - - if (bp == p.end()) { - return false; - } - - u64 len = bp->second.len; - - // IDA Pro defaults to 4-byte breakpoints for all non-hardware breakpoints - // no matter if it's a 4-byte or 2-byte instruction. When you execute a - // Thumb instruction with a 4-byte breakpoint set, it will set a breakpoint on - // two instructions instead of the single instruction you placed the breakpoint - // on. So, as a way to make sure that execution breakpoints are only breaking - // on the instruction that was specified, set the length of an execution - // breakpoint to 1. This should be fine since the CPU should never begin executing - // an instruction anywhere except the beginning of the instruction. - if (type == BreakpointType::Execute) { - len = 1; - } - - if (bp->second.active && (addr >= bp->second.addr && addr < bp->second.addr + len)) { - LOG_DEBUG(Debug_GDBStub, - "Found breakpoint type {} @ {:016X}, range: {:016X}" - " - {:016X} ({:X} bytes)", - static_cast<int>(type), addr, bp->second.addr, bp->second.addr + len, len); - return true; - } - - return false; -} - -/** - * Send packet to gdb client. - * - * @param packet Packet to be sent to client. - */ -static void SendPacket(const char packet) { - std::size_t sent_size = send(gdbserver_socket, &packet, 1, 0); - if (sent_size != 1) { - LOG_ERROR(Debug_GDBStub, "send failed"); - } -} - -/** - * Send reply to gdb client. - * - * @param reply Reply to be sent to client. - */ -static void SendReply(const char* reply) { - if (!IsConnected()) { - return; - } - - LOG_DEBUG(Debug_GDBStub, "Reply: {}", reply); - - memset(command_buffer, 0, sizeof(command_buffer)); - - command_length = static_cast<u32>(strlen(reply)); - if (command_length + 4 > sizeof(command_buffer)) { - LOG_ERROR(Debug_GDBStub, "command_buffer overflow in SendReply"); - return; - } - - memcpy(command_buffer + 1, reply, command_length); - - u8 checksum = CalculateChecksum(command_buffer, command_length + 1); - command_buffer[0] = GDB_STUB_START; - command_buffer[command_length + 1] = GDB_STUB_END; - command_buffer[command_length + 2] = NibbleToHex(checksum >> 4); - command_buffer[command_length + 3] = NibbleToHex(checksum); - - u8* ptr = command_buffer; - u32 left = command_length + 4; - while (left > 0) { - int sent_size = send(gdbserver_socket, reinterpret_cast<char*>(ptr), left, 0); - if (sent_size < 0) { - LOG_ERROR(Debug_GDBStub, "gdb: send failed"); - return Shutdown(); - } - - left -= sent_size; - ptr += sent_size; - } -} - -/// Handle query command from gdb client. -static void HandleQuery() { - LOG_DEBUG(Debug_GDBStub, "gdb: query '{}'", command_buffer + 1); - - const char* query = reinterpret_cast<const char*>(command_buffer + 1); - - if (strcmp(query, "TStatus") == 0) { - SendReply("T0"); - } else if (strncmp(query, "Supported", strlen("Supported")) == 0) { - // PacketSize needs to be large enough for target xml - std::string buffer = "PacketSize=2000;qXfer:features:read+;qXfer:threads:read+"; - if (!modules.empty()) { - buffer += ";qXfer:libraries:read+"; - } - SendReply(buffer.c_str()); - } else if (strncmp(query, "Xfer:features:read:target.xml:", - strlen("Xfer:features:read:target.xml:")) == 0) { - SendReply(target_xml); - } else if (strncmp(query, "Offsets", strlen("Offsets")) == 0) { - const VAddr base_address = - Core::System::GetInstance().CurrentProcess()->PageTable().GetCodeRegionStart(); - std::string buffer = fmt::format("TextSeg={:0x}", base_address); - SendReply(buffer.c_str()); - } else if (strncmp(query, "fThreadInfo", strlen("fThreadInfo")) == 0) { - std::string val = "m"; - const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList(); - for (const auto& thread : threads) { - val += fmt::format("{:x},", thread->GetThreadID()); - } - val.pop_back(); - SendReply(val.c_str()); - } else if (strncmp(query, "sThreadInfo", strlen("sThreadInfo")) == 0) { - SendReply("l"); - } else if (strncmp(query, "Xfer:threads:read", strlen("Xfer:threads:read")) == 0) { - std::string buffer; - buffer += "l<?xml version=\"1.0\"?>"; - buffer += "<threads>"; - const auto& threads = Core::System::GetInstance().GlobalScheduler().GetThreadList(); - for (const auto& thread : threads) { - buffer += - fmt::format(R"*(<thread id="{:x}" core="{:d}" name="Thread {:x}"></thread>)*", - thread->GetThreadID(), thread->GetProcessorID(), thread->GetThreadID()); - } - buffer += "</threads>"; - SendReply(buffer.c_str()); - } else if (strncmp(query, "Xfer:libraries:read", strlen("Xfer:libraries:read")) == 0) { - std::string buffer; - buffer += "l<?xml version=\"1.0\"?>"; - buffer += "<library-list>"; - for (const auto& module : modules) { - buffer += - fmt::format(R"*("<library name = "{}"><segment address = "0x{:x}"/></library>)*", - module.name, module.beg); - } - buffer += "</library-list>"; - SendReply(buffer.c_str()); - } else { - SendReply(""); - } -} - -/// Handle set thread command from gdb client. -static void HandleSetThread() { - int thread_id = -1; - if (command_buffer[2] != '-') { - thread_id = static_cast<int>(HexToInt(command_buffer + 2, command_length - 2)); - } - if (thread_id >= 1) { - current_thread = FindThreadById(thread_id); - } - if (!current_thread) { - thread_id = 1; - current_thread = FindThreadById(thread_id); - } - if (current_thread) { - SendReply("OK"); - return; - } - SendReply("E01"); -} - -/// Handle thread alive command from gdb client. -static void HandleThreadAlive() { - int thread_id = static_cast<int>(HexToInt(command_buffer + 1, command_length - 1)); - if (thread_id == 0) { - thread_id = 1; - } - if (FindThreadById(thread_id)) { - SendReply("OK"); - return; - } - SendReply("E01"); -} - -/** - * Send signal packet to client. - * - * @param signal Signal to be sent to client. - */ -static void SendSignal(Kernel::Thread* thread, u32 signal, bool full = true) { - if (gdbserver_socket == -1) { - return; - } - - latest_signal = signal; - - if (!thread) { - full = false; - } - - std::string buffer; - if (full) { - buffer = fmt::format("T{:02x}{:02x}:{:016x};{:02x}:{:016x};{:02x}:{:016x}", latest_signal, - PC_REGISTER, Common::swap64(RegRead(PC_REGISTER, thread)), SP_REGISTER, - Common::swap64(RegRead(SP_REGISTER, thread)), LR_REGISTER, - Common::swap64(RegRead(LR_REGISTER, thread))); - } else { - buffer = fmt::format("T{:02x}", latest_signal); - } - - if (thread) { - buffer += fmt::format(";thread:{:x};", thread->GetThreadID()); - } - - SendReply(buffer.c_str()); -} - -/// Read command from gdb client. -static void ReadCommand() { - command_length = 0; - memset(command_buffer, 0, sizeof(command_buffer)); - - u8 c = ReadByte(); - if (c == '+') { - // ignore ack - return; - } else if (c == 0x03) { - LOG_INFO(Debug_GDBStub, "gdb: found break command"); - halt_loop = true; - SendSignal(current_thread, SIGTRAP); - return; - } else if (c != GDB_STUB_START) { - LOG_DEBUG(Debug_GDBStub, "gdb: read invalid byte {:02X}", c); - return; - } - - while ((c = ReadByte()) != GDB_STUB_END) { - if (command_length >= sizeof(command_buffer)) { - LOG_ERROR(Debug_GDBStub, "gdb: command_buffer overflow"); - SendPacket(GDB_STUB_NACK); - return; - } - command_buffer[command_length++] = c; - } - - u8 checksum_received = HexCharToValue(ReadByte()) << 4; - checksum_received |= HexCharToValue(ReadByte()); - - u8 checksum_calculated = CalculateChecksum(command_buffer, command_length); - - if (checksum_received != checksum_calculated) { - LOG_ERROR(Debug_GDBStub, - "gdb: invalid checksum: calculated {:02X} and read {:02X} for ${}# (length: {})", - checksum_calculated, checksum_received, command_buffer, command_length); - - command_length = 0; - - SendPacket(GDB_STUB_NACK); - return; - } - - SendPacket(GDB_STUB_ACK); -} - -/// Check if there is data to be read from the gdb client. -static bool IsDataAvailable() { - if (!IsConnected()) { - return false; - } - - fd_set fd_socket; - - FD_ZERO(&fd_socket); - FD_SET(static_cast<u32>(gdbserver_socket), &fd_socket); - - struct timeval t; - t.tv_sec = 0; - t.tv_usec = 0; - - if (select(gdbserver_socket + 1, &fd_socket, nullptr, nullptr, &t) < 0) { - LOG_ERROR(Debug_GDBStub, "select failed"); - return false; - } - - return FD_ISSET(gdbserver_socket, &fd_socket) != 0; -} - -/// Send requested register to gdb client. -static void ReadRegister() { - static u8 reply[64]; - memset(reply, 0, sizeof(reply)); - - u32 id = HexCharToValue(command_buffer[1]); - if (command_buffer[2] != '\0') { - id <<= 4; - id |= HexCharToValue(command_buffer[2]); - } - - if (id <= SP_REGISTER) { - LongToGdbHex(reply, RegRead(id, current_thread)); - } else if (id == PC_REGISTER) { - LongToGdbHex(reply, RegRead(id, current_thread)); - } else if (id == PSTATE_REGISTER) { - IntToGdbHex(reply, static_cast<u32>(RegRead(id, current_thread))); - } else if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) { - u128 r = FpuRead(id, current_thread); - LongToGdbHex(reply, r[0]); - LongToGdbHex(reply + 16, r[1]); - } else if (id == FPCR_REGISTER) { - u128 r = FpuRead(id, current_thread); - IntToGdbHex(reply, static_cast<u32>(r[0])); - } else if (id == FPCR_REGISTER + 1) { - u128 r = FpuRead(id, current_thread); - IntToGdbHex(reply, static_cast<u32>(r[0] >> 32)); - } - - SendReply(reinterpret_cast<char*>(reply)); -} - -/// Send all registers to the gdb client. -static void ReadRegisters() { - static u8 buffer[GDB_BUFFER_SIZE - 4]; - memset(buffer, 0, sizeof(buffer)); - - u8* bufptr = buffer; - - for (u32 reg = 0; reg <= SP_REGISTER; reg++) { - LongToGdbHex(bufptr + reg * 16, RegRead(reg, current_thread)); - } - - bufptr += 32 * 16; - - LongToGdbHex(bufptr, RegRead(PC_REGISTER, current_thread)); - - bufptr += 16; - - IntToGdbHex(bufptr, static_cast<u32>(RegRead(PSTATE_REGISTER, current_thread))); - - bufptr += 8; - - u128 r; - - for (u32 reg = UC_ARM64_REG_Q0; reg < FPCR_REGISTER; reg++) { - r = FpuRead(reg, current_thread); - LongToGdbHex(bufptr + reg * 32, r[0]); - LongToGdbHex(bufptr + reg * 32 + 16, r[1]); - } - - bufptr += 32 * 32; - - r = FpuRead(FPCR_REGISTER, current_thread); - IntToGdbHex(bufptr, static_cast<u32>(r[0])); - - bufptr += 8; - - SendReply(reinterpret_cast<char*>(buffer)); -} - -/// Modify data of register specified by gdb client. -static void WriteRegister() { - const u8* buffer_ptr = command_buffer + 3; - - u32 id = HexCharToValue(command_buffer[1]); - if (command_buffer[2] != '=') { - ++buffer_ptr; - id <<= 4; - id |= HexCharToValue(command_buffer[2]); - } - - if (id <= SP_REGISTER) { - RegWrite(id, GdbHexToLong(buffer_ptr), current_thread); - } else if (id == PC_REGISTER) { - RegWrite(id, GdbHexToLong(buffer_ptr), current_thread); - } else if (id == PSTATE_REGISTER) { - RegWrite(id, GdbHexToInt(buffer_ptr), current_thread); - } else if (id >= UC_ARM64_REG_Q0 && id < FPCR_REGISTER) { - FpuWrite(id, GdbHexToU128(buffer_ptr), current_thread); - } else if (id == FPCR_REGISTER) { - } else if (id == FPCR_REGISTER + 1) { - } - - // Update ARM context, skipping scheduler - no running threads at this point - Core::System::GetInstance() - .ArmInterface(current_core) - .LoadContext(current_thread->GetContext64()); - - SendReply("OK"); -} - -/// Modify all registers with data received from the client. -static void WriteRegisters() { - const u8* buffer_ptr = command_buffer + 1; - - if (command_buffer[0] != 'G') - return SendReply("E01"); - - for (u32 i = 0, reg = 0; reg <= FPCR_REGISTER; i++, reg++) { - if (reg <= SP_REGISTER) { - RegWrite(reg, GdbHexToLong(buffer_ptr + i * 16), current_thread); - } else if (reg == PC_REGISTER) { - RegWrite(PC_REGISTER, GdbHexToLong(buffer_ptr + i * 16), current_thread); - } else if (reg == PSTATE_REGISTER) { - RegWrite(PSTATE_REGISTER, GdbHexToInt(buffer_ptr + i * 16), current_thread); - } else if (reg >= UC_ARM64_REG_Q0 && reg < FPCR_REGISTER) { - RegWrite(reg, GdbHexToLong(buffer_ptr + i * 16), current_thread); - } else if (reg == FPCR_REGISTER) { - RegWrite(FPCR_REGISTER, GdbHexToLong(buffer_ptr + i * 16), current_thread); - } else if (reg == FPCR_REGISTER + 1) { - RegWrite(FPCR_REGISTER, GdbHexToLong(buffer_ptr + i * 16), current_thread); - } - } - - // Update ARM context, skipping scheduler - no running threads at this point - Core::System::GetInstance() - .ArmInterface(current_core) - .LoadContext(current_thread->GetContext64()); - - SendReply("OK"); -} - -/// Read location in memory specified by gdb client. -static void ReadMemory() { - static u8 reply[GDB_BUFFER_SIZE - 4]; - - auto start_offset = command_buffer + 1; - const auto addr_pos = std::find(start_offset, command_buffer + command_length, ','); - const VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset)); - - start_offset = addr_pos + 1; - const u64 len = - HexToLong(start_offset, static_cast<u64>((command_buffer + command_length) - start_offset)); - - LOG_DEBUG(Debug_GDBStub, "gdb: addr: {:016X} len: {:016X}", addr, len); - - if (len * 2 > sizeof(reply)) { - SendReply("E01"); - } - - auto& memory = Core::System::GetInstance().Memory(); - if (!memory.IsValidVirtualAddress(addr)) { - return SendReply("E00"); - } - - std::vector<u8> data(len); - memory.ReadBlock(addr, data.data(), len); - - MemToGdbHex(reply, data.data(), len); - reply[len * 2] = '\0'; - SendReply(reinterpret_cast<char*>(reply)); -} - -/// Modify location in memory with data received from the gdb client. -static void WriteMemory() { - auto start_offset = command_buffer + 1; - const auto addr_pos = std::find(start_offset, command_buffer + command_length, ','); - const VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset)); - - start_offset = addr_pos + 1; - const auto len_pos = std::find(start_offset, command_buffer + command_length, ':'); - const u64 len = HexToLong(start_offset, static_cast<u64>(len_pos - start_offset)); - - auto& system = Core::System::GetInstance(); - auto& memory = system.Memory(); - if (!memory.IsValidVirtualAddress(addr)) { - return SendReply("E00"); - } - - std::vector<u8> data(len); - GdbHexToMem(data.data(), len_pos + 1, len); - memory.WriteBlock(addr, data.data(), len); - system.InvalidateCpuInstructionCaches(); - SendReply("OK"); -} - -void Break(bool is_memory_break) { - send_trap = true; - - memory_break = is_memory_break; -} - -/// Tell the CPU that it should perform a single step. -static void Step() { - if (command_length > 1) { - RegWrite(PC_REGISTER, GdbHexToLong(command_buffer + 1), current_thread); - // Update ARM context, skipping scheduler - no running threads at this point - Core::System::GetInstance() - .ArmInterface(current_core) - .LoadContext(current_thread->GetContext64()); - } - step_loop = true; - halt_loop = true; - send_trap = true; - Core::System::GetInstance().InvalidateCpuInstructionCaches(); -} - -/// Tell the CPU if we hit a memory breakpoint. -bool IsMemoryBreak() { - if (!IsConnected()) { - return false; - } - - return memory_break; -} - -/// Tell the CPU to continue executing. -static void Continue() { - memory_break = false; - step_loop = false; - halt_loop = false; - Core::System::GetInstance().InvalidateCpuInstructionCaches(); -} - -/** - * Commit breakpoint to list of breakpoints. - * - * @param type Type of breakpoint. - * @param addr Address of breakpoint. - * @param len Length of breakpoint. - */ -static bool CommitBreakpoint(BreakpointType type, VAddr addr, u64 len) { - BreakpointMap& p = GetBreakpointMap(type); - - Breakpoint breakpoint; - breakpoint.active = true; - breakpoint.addr = addr; - breakpoint.len = len; - - auto& system = Core::System::GetInstance(); - auto& memory = system.Memory(); - memory.ReadBlock(addr, breakpoint.inst.data(), breakpoint.inst.size()); - - static constexpr std::array<u8, 4> btrap{0x00, 0x7d, 0x20, 0xd4}; - if (type == BreakpointType::Execute) { - memory.WriteBlock(addr, btrap.data(), btrap.size()); - system.InvalidateCpuInstructionCaches(); - } - p.insert({addr, breakpoint}); - - LOG_DEBUG(Debug_GDBStub, "gdb: added {} breakpoint: {:016X} bytes at {:016X}", - static_cast<int>(type), breakpoint.len, breakpoint.addr); - - return true; -} - -/// Handle add breakpoint command from gdb client. -static void AddBreakpoint() { - BreakpointType type; - - u8 type_id = HexCharToValue(command_buffer[1]); - switch (type_id) { - case 0: - case 1: - type = BreakpointType::Execute; - break; - case 2: - type = BreakpointType::Write; - break; - case 3: - type = BreakpointType::Read; - break; - case 4: - type = BreakpointType::Access; - break; - default: - return SendReply("E01"); - } - - auto start_offset = command_buffer + 3; - auto addr_pos = std::find(start_offset, command_buffer + command_length, ','); - VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset)); - - start_offset = addr_pos + 1; - u64 len = - HexToLong(start_offset, static_cast<u64>((command_buffer + command_length) - start_offset)); - - if (type == BreakpointType::Access) { - // Access is made up of Read and Write types, so add both breakpoints - type = BreakpointType::Read; - - if (!CommitBreakpoint(type, addr, len)) { - return SendReply("E02"); - } - - type = BreakpointType::Write; - } - - if (!CommitBreakpoint(type, addr, len)) { - return SendReply("E02"); - } - - SendReply("OK"); -} - -/// Handle remove breakpoint command from gdb client. -static void RemoveBreakpoint() { - BreakpointType type; - - u8 type_id = HexCharToValue(command_buffer[1]); - switch (type_id) { - case 0: - case 1: - type = BreakpointType::Execute; - break; - case 2: - type = BreakpointType::Write; - break; - case 3: - type = BreakpointType::Read; - break; - case 4: - type = BreakpointType::Access; - break; - default: - return SendReply("E01"); - } - - auto start_offset = command_buffer + 3; - auto addr_pos = std::find(start_offset, command_buffer + command_length, ','); - VAddr addr = HexToLong(start_offset, static_cast<u64>(addr_pos - start_offset)); - - if (type == BreakpointType::Access) { - // Access is made up of Read and Write types, so add both breakpoints - type = BreakpointType::Read; - RemoveBreakpoint(type, addr); - - type = BreakpointType::Write; - } - - RemoveBreakpoint(type, addr); - SendReply("OK"); -} - -void HandlePacket() { - if (!IsConnected()) { - if (defer_start) { - ToggleServer(true); - } - return; - } - - if (!IsDataAvailable()) { - return; - } - - ReadCommand(); - if (command_length == 0) { - return; - } - - LOG_DEBUG(Debug_GDBStub, "Packet: {}", command_buffer); - - switch (command_buffer[0]) { - case 'q': - HandleQuery(); - break; - case 'H': - HandleSetThread(); - break; - case '?': - SendSignal(current_thread, latest_signal); - break; - case 'k': - Shutdown(); - LOG_INFO(Debug_GDBStub, "killed by gdb"); - return; - case 'g': - ReadRegisters(); - break; - case 'G': - WriteRegisters(); - break; - case 'p': - ReadRegister(); - break; - case 'P': - WriteRegister(); - break; - case 'm': - ReadMemory(); - break; - case 'M': - WriteMemory(); - break; - case 's': - Step(); - return; - case 'C': - case 'c': - Continue(); - return; - case 'z': - RemoveBreakpoint(); - break; - case 'Z': - AddBreakpoint(); - break; - case 'T': - HandleThreadAlive(); - break; - default: - SendReply(""); - break; - } -} - -void SetServerPort(u16 port) { - gdbstub_port = port; -} - -void ToggleServer(bool status) { - if (status) { - server_enabled = status; - - // Start server - if (!IsConnected() && Core::System::GetInstance().IsPoweredOn()) { - Init(); - } - } else { - // Stop server - if (IsConnected()) { - Shutdown(); - } - - server_enabled = status; - } -} - -void DeferStart() { - defer_start = true; -} - -static void Init(u16 port) { - if (!server_enabled) { - // Set the halt loop to false in case the user enabled the gdbstub mid-execution. - // This way the CPU can still execute normally. - halt_loop = false; - step_loop = false; - return; - } - - // Setup initial gdbstub status - halt_loop = true; - step_loop = false; - - breakpoints_execute.clear(); - breakpoints_read.clear(); - breakpoints_write.clear(); - - modules.clear(); - - // Start gdb server - LOG_INFO(Debug_GDBStub, "Starting GDB server on port {}...", port); - - sockaddr_in saddr_server = {}; - saddr_server.sin_family = AF_INET; - saddr_server.sin_port = htons(port); - saddr_server.sin_addr.s_addr = INADDR_ANY; - -#ifdef _WIN32 - WSAStartup(MAKEWORD(2, 2), &InitData); -#endif - - int tmpsock = static_cast<int>(socket(PF_INET, SOCK_STREAM, 0)); - if (tmpsock == -1) { - LOG_ERROR(Debug_GDBStub, "Failed to create gdb socket"); - } - - // Set socket to SO_REUSEADDR so it can always bind on the same port - int reuse_enabled = 1; - if (setsockopt(tmpsock, SOL_SOCKET, SO_REUSEADDR, (const char*)&reuse_enabled, - sizeof(reuse_enabled)) < 0) { - LOG_ERROR(Debug_GDBStub, "Failed to set gdb socket option"); - } - - const sockaddr* server_addr = reinterpret_cast<const sockaddr*>(&saddr_server); - socklen_t server_addrlen = sizeof(saddr_server); - if (bind(tmpsock, server_addr, server_addrlen) < 0) { - LOG_ERROR(Debug_GDBStub, "Failed to bind gdb socket"); - } - - if (listen(tmpsock, 1) < 0) { - LOG_ERROR(Debug_GDBStub, "Failed to listen to gdb socket"); - } - - // Wait for gdb to connect - LOG_INFO(Debug_GDBStub, "Waiting for gdb to connect..."); - sockaddr_in saddr_client; - sockaddr* client_addr = reinterpret_cast<sockaddr*>(&saddr_client); - socklen_t client_addrlen = sizeof(saddr_client); - gdbserver_socket = static_cast<int>(accept(tmpsock, client_addr, &client_addrlen)); - if (gdbserver_socket < 0) { - // In the case that we couldn't start the server for whatever reason, just start CPU - // execution like normal. - halt_loop = false; - step_loop = false; - - LOG_ERROR(Debug_GDBStub, "Failed to accept gdb client"); - } else { - LOG_INFO(Debug_GDBStub, "Client connected."); - saddr_client.sin_addr.s_addr = ntohl(saddr_client.sin_addr.s_addr); - } - - // Clean up temporary socket if it's still alive at this point. - if (tmpsock != -1) { - shutdown(tmpsock, SHUT_RDWR); - } -} - -void Init() { - Init(gdbstub_port); -} - -void Shutdown() { - if (!server_enabled) { - return; - } - defer_start = false; - - LOG_INFO(Debug_GDBStub, "Stopping GDB ..."); - if (gdbserver_socket != -1) { - shutdown(gdbserver_socket, SHUT_RDWR); - gdbserver_socket = -1; - } - -#ifdef _WIN32 - WSACleanup(); -#endif - - LOG_INFO(Debug_GDBStub, "GDB stopped."); -} - -bool IsServerEnabled() { - return server_enabled; -} - -bool IsConnected() { - return IsServerEnabled() && gdbserver_socket != -1; -} - -bool GetCpuHaltFlag() { - return halt_loop; -} - -bool GetCpuStepFlag() { - return step_loop; -} - -void SetCpuStepFlag(bool is_step) { - step_loop = is_step; -} - -void SendTrap(Kernel::Thread* thread, int trap) { - if (!send_trap) { - return; - } - - current_thread = thread; - SendSignal(thread, trap); - - halt_loop = true; - send_trap = false; -} -}; // namespace GDBStub diff --git a/src/core/gdbstub/gdbstub.h b/src/core/gdbstub/gdbstub.h deleted file mode 100644 index 8fe3c320b..000000000 --- a/src/core/gdbstub/gdbstub.h +++ /dev/null @@ -1,114 +0,0 @@ -// Copyright 2013 Dolphin Emulator Project -// Licensed under GPLv2+ -// Refer to the license.txt file included. - -// Originally written by Sven Peter <sven@fail0verflow.com> for anergistic. - -#pragma once - -#include <string> -#include "common/common_types.h" -#include "core/hle/kernel/thread.h" - -namespace GDBStub { - -/// Breakpoint Method -enum class BreakpointType { - None, ///< None - Execute, ///< Execution Breakpoint - Read, ///< Read Breakpoint - Write, ///< Write Breakpoint - Access ///< Access (R/W) Breakpoint -}; - -struct BreakpointAddress { - VAddr address; - BreakpointType type; -}; - -/** - * Set the port the gdbstub should use to listen for connections. - * - * @param port Port to listen for connection - */ -void SetServerPort(u16 port); - -/** - * Starts or stops the server if possible. - * - * @param status Set the server to enabled or disabled. - */ -void ToggleServer(bool status); - -/// Start the gdbstub server. -void Init(); - -/** - * Defer initialization of the gdbstub to the first packet processing functions. - * This avoids a case where the gdbstub thread is frozen after initialization - * and fails to respond in time to packets. - */ -void DeferStart(); - -/// Stop gdbstub server. -void Shutdown(); - -/// Checks if the gdbstub server is enabled. -bool IsServerEnabled(); - -/// Returns true if there is an active socket connection. -bool IsConnected(); - -/// Register module. -void RegisterModule(std::string name, VAddr beg, VAddr end, bool add_elf_ext = true); - -/** - * Signal to the gdbstub server that it should halt CPU execution. - * - * @param is_memory_break If true, the break resulted from a memory breakpoint. - */ -void Break(bool is_memory_break = false); - -/// Determine if there was a memory breakpoint. -bool IsMemoryBreak(); - -/// Read and handle packet from gdb client. -void HandlePacket(); - -/** - * Get the nearest breakpoint of the specified type at the given address. - * - * @param addr Address to search from. - * @param type Type of breakpoint. - */ -BreakpointAddress GetNextBreakpointFromAddress(VAddr addr, GDBStub::BreakpointType type); - -/** - * Check if a breakpoint of the specified type exists at the given address. - * - * @param addr Address of breakpoint. - * @param type Type of breakpoint. - */ -bool CheckBreakpoint(VAddr addr, GDBStub::BreakpointType type); - -/// If set to true, the CPU will halt at the beginning of the next CPU loop. -bool GetCpuHaltFlag(); - -/// If set to true and the CPU is halted, the CPU will step one instruction. -bool GetCpuStepFlag(); - -/** - * When set to true, the CPU will step one instruction when the CPU is halted next. - * - * @param is_step - */ -void SetCpuStepFlag(bool is_step); - -/** - * Send trap signal from thread back to the gdbstub server. - * - * @param thread Sending thread. - * @param trap Trap no. - */ -void SendTrap(Kernel::Thread* thread, int trap); -} // namespace GDBStub diff --git a/src/core/hle/ipc_helpers.h b/src/core/hle/ipc_helpers.h index 1b503331f..56cc911d1 100644 --- a/src/core/hle/ipc_helpers.h +++ b/src/core/hle/ipc_helpers.h @@ -12,7 +12,6 @@ #include <utility> #include "common/assert.h" #include "common/common_types.h" -#include "core/core.h" #include "core/hle/ipc.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/client_session.h" @@ -38,10 +37,11 @@ public: explicit RequestHelperBase(Kernel::HLERequestContext& context) : context(&context), cmdbuf(context.CommandBuffer()) {} - void Skip(unsigned size_in_words, bool set_to_null) { - if (set_to_null) + void Skip(u32 size_in_words, bool set_to_null) { + if (set_to_null) { memset(cmdbuf + index, 0, size_in_words * sizeof(u32)); - index += size_in_words; + } + index += static_cast<ptrdiff_t>(size_in_words); } /** @@ -49,15 +49,15 @@ public: */ void AlignWithPadding() { if (index & 3) { - Skip(4 - (index & 3), true); + Skip(static_cast<u32>(4 - (index & 3)), true); } } - unsigned GetCurrentOffset() const { - return static_cast<unsigned>(index); + u32 GetCurrentOffset() const { + return static_cast<u32>(index); } - void SetCurrentOffset(unsigned offset) { + void SetCurrentOffset(u32 offset) { index = static_cast<ptrdiff_t>(offset); } }; @@ -72,14 +72,12 @@ public: AlwaysMoveHandles = 1, }; - explicit ResponseBuilder(u32* command_buffer) : RequestHelperBase(command_buffer) {} - explicit ResponseBuilder(Kernel::HLERequestContext& context, u32 normal_params_size, u32 num_handles_to_copy = 0, u32 num_objects_to_move = 0, Flags flags = Flags::None) - : RequestHelperBase(context), normal_params_size(normal_params_size), - num_handles_to_copy(num_handles_to_copy), num_objects_to_move(num_objects_to_move) { + num_handles_to_copy(num_handles_to_copy), + num_objects_to_move(num_objects_to_move), kernel{context.kernel} { memset(cmdbuf, 0, sizeof(u32) * IPC::COMMAND_BUFFER_LENGTH); @@ -89,7 +87,7 @@ public: // The entire size of the raw data section in u32 units, including the 16 bytes of mandatory // padding. - u32 raw_data_size = sizeof(IPC::DataPayloadHeader) / 4 + 4 + normal_params_size; + u64 raw_data_size = sizeof(IPC::DataPayloadHeader) / 4 + 4 + normal_params_size; u32 num_handles_to_move{}; u32 num_domain_objects{}; @@ -105,7 +103,7 @@ public: raw_data_size += sizeof(DomainMessageHeader) / 4 + num_domain_objects; } - header.data_size.Assign(raw_data_size); + header.data_size.Assign(static_cast<u32>(raw_data_size)); if (num_handles_to_copy || num_handles_to_move) { header.enable_handle_descriptor.Assign(1); } @@ -139,7 +137,6 @@ public: if (context->Session()->IsDomain()) { context->AddDomainObject(std::move(iface)); } else { - auto& kernel = Core::System::GetInstance().Kernel(); auto [client, server] = Kernel::Session::Create(kernel, iface->GetServiceName()); context->AddMoveObject(std::move(client)); iface->ClientConnected(std::move(server)); @@ -169,8 +166,23 @@ public: ValidateHeader(); } + void PushImpl(s8 value); + void PushImpl(s16 value); + void PushImpl(s32 value); + void PushImpl(s64 value); + void PushImpl(u8 value); + void PushImpl(u16 value); + void PushImpl(u32 value); + void PushImpl(u64 value); + void PushImpl(float value); + void PushImpl(double value); + void PushImpl(bool value); + void PushImpl(ResultCode value); + template <typename T> - void Push(T value); + void Push(T value) { + return PushImpl(value); + } template <typename First, typename... Other> void Push(const First& first_value, const Other&... other_values); @@ -213,17 +225,16 @@ private: u32 num_handles_to_copy{}; u32 num_objects_to_move{}; ///< Domain objects or move handles, context dependent std::ptrdiff_t datapayload_index{}; + Kernel::KernelCore& kernel; }; /// Push /// -template <> -inline void ResponseBuilder::Push(s32 value) { +inline void ResponseBuilder::PushImpl(s32 value) { cmdbuf[index++] = static_cast<u32>(value); } -template <> -inline void ResponseBuilder::Push(u32 value) { +inline void ResponseBuilder::PushImpl(u32 value) { cmdbuf[index++] = value; } @@ -235,62 +246,52 @@ void ResponseBuilder::PushRaw(const T& value) { index += (sizeof(T) + 3) / 4; // round up to word length } -template <> -inline void ResponseBuilder::Push(ResultCode value) { +inline void ResponseBuilder::PushImpl(ResultCode value) { // Result codes are actually 64-bit in the IPC buffer, but only the high part is discarded. Push(value.raw); Push<u32>(0); } -template <> -inline void ResponseBuilder::Push(s8 value) { +inline void ResponseBuilder::PushImpl(s8 value) { PushRaw(value); } -template <> -inline void ResponseBuilder::Push(s16 value) { +inline void ResponseBuilder::PushImpl(s16 value) { PushRaw(value); } -template <> -inline void ResponseBuilder::Push(s64 value) { - Push(static_cast<u32>(value)); - Push(static_cast<u32>(value >> 32)); +inline void ResponseBuilder::PushImpl(s64 value) { + PushImpl(static_cast<u32>(value)); + PushImpl(static_cast<u32>(value >> 32)); } -template <> -inline void ResponseBuilder::Push(u8 value) { +inline void ResponseBuilder::PushImpl(u8 value) { PushRaw(value); } -template <> -inline void ResponseBuilder::Push(u16 value) { +inline void ResponseBuilder::PushImpl(u16 value) { PushRaw(value); } -template <> -inline void ResponseBuilder::Push(u64 value) { - Push(static_cast<u32>(value)); - Push(static_cast<u32>(value >> 32)); +inline void ResponseBuilder::PushImpl(u64 value) { + PushImpl(static_cast<u32>(value)); + PushImpl(static_cast<u32>(value >> 32)); } -template <> -inline void ResponseBuilder::Push(float value) { +inline void ResponseBuilder::PushImpl(float value) { u32 integral; std::memcpy(&integral, &value, sizeof(u32)); - Push(integral); + PushImpl(integral); } -template <> -inline void ResponseBuilder::Push(double value) { +inline void ResponseBuilder::PushImpl(double value) { u64 integral; std::memcpy(&integral, &value, sizeof(u64)); - Push(integral); + PushImpl(integral); } -template <> -inline void ResponseBuilder::Push(bool value) { - Push(static_cast<u8>(value)); +inline void ResponseBuilder::PushImpl(bool value) { + PushImpl(static_cast<u8>(value)); } template <typename First, typename... Other> diff --git a/src/core/hle/kernel/address_arbiter.cpp b/src/core/hle/kernel/address_arbiter.cpp index b882eaa0f..20ffa7d47 100644 --- a/src/core/hle/kernel/address_arbiter.cpp +++ b/src/core/hle/kernel/address_arbiter.cpp @@ -12,8 +12,9 @@ #include "core/hle/kernel/address_arbiter.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h" #include "core/hle/kernel/kernel.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "core/hle/kernel/time_manager.h" #include "core/hle/result.h" @@ -58,7 +59,7 @@ ResultCode AddressArbiter::SignalToAddress(VAddr address, SignalType type, s32 v } ResultCode AddressArbiter::SignalToAddressOnly(VAddr address, s32 num_to_wake) { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(system.Kernel()); const std::vector<std::shared_ptr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address); WakeThreads(waiting_threads, num_to_wake); @@ -67,7 +68,7 @@ ResultCode AddressArbiter::SignalToAddressOnly(VAddr address, s32 num_to_wake) { ResultCode AddressArbiter::IncrementAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake) { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(system.Kernel()); auto& memory = system.Memory(); // Ensure that we can write to the address. @@ -92,7 +93,7 @@ ResultCode AddressArbiter::IncrementAndSignalToAddressIfEqual(VAddr address, s32 ResultCode AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake) { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(system.Kernel()); auto& memory = system.Memory(); // Ensure that we can write to the address. @@ -153,11 +154,11 @@ ResultCode AddressArbiter::WaitForAddressIfLessThan(VAddr address, s32 value, s6 bool should_decrement) { auto& memory = system.Memory(); auto& kernel = system.Kernel(); - Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); + Thread* current_thread = kernel.CurrentScheduler()->GetCurrentThread(); Handle event_handle = InvalidHandle; { - SchedulerLockAndSleep lock(kernel, event_handle, current_thread, timeout); + KScopedSchedulerLockAndSleep lock(kernel, event_handle, current_thread, timeout); if (current_thread->IsPendingTermination()) { lock.CancelSleep(); @@ -210,7 +211,7 @@ ResultCode AddressArbiter::WaitForAddressIfLessThan(VAddr address, s32 value, s6 } { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); if (current_thread->IsWaitingForArbitration()) { RemoveThread(SharedFrom(current_thread)); current_thread->WaitForArbitration(false); @@ -223,11 +224,11 @@ ResultCode AddressArbiter::WaitForAddressIfLessThan(VAddr address, s32 value, s6 ResultCode AddressArbiter::WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout) { auto& memory = system.Memory(); auto& kernel = system.Kernel(); - Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); + Thread* current_thread = kernel.CurrentScheduler()->GetCurrentThread(); Handle event_handle = InvalidHandle; { - SchedulerLockAndSleep lock(kernel, event_handle, current_thread, timeout); + KScopedSchedulerLockAndSleep lock(kernel, event_handle, current_thread, timeout); if (current_thread->IsPendingTermination()) { lock.CancelSleep(); @@ -265,7 +266,7 @@ ResultCode AddressArbiter::WaitForAddressIfEqual(VAddr address, s32 value, s64 t } { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); if (current_thread->IsWaitingForArbitration()) { RemoveThread(SharedFrom(current_thread)); current_thread->WaitForArbitration(false); @@ -275,12 +276,6 @@ ResultCode AddressArbiter::WaitForAddressIfEqual(VAddr address, s32 value, s64 t return current_thread->GetSignalingResult(); } -void AddressArbiter::HandleWakeupThread(std::shared_ptr<Thread> thread) { - ASSERT(thread->GetStatus() == ThreadStatus::WaitArb); - RemoveThread(thread); - thread->SetArbiterWaitAddress(0); -} - void AddressArbiter::InsertThread(std::shared_ptr<Thread> thread) { const VAddr arb_addr = thread->GetArbiterWaitAddress(); std::list<std::shared_ptr<Thread>>& thread_list = arb_threads[arb_addr]; diff --git a/src/core/hle/kernel/address_arbiter.h b/src/core/hle/kernel/address_arbiter.h index 0b05d533c..b91edc67d 100644 --- a/src/core/hle/kernel/address_arbiter.h +++ b/src/core/hle/kernel/address_arbiter.h @@ -50,9 +50,6 @@ public: /// Waits on an address with a particular arbitration type. ResultCode WaitForAddress(VAddr address, ArbitrationType type, s32 value, s64 timeout_ns); - /// Removes a thread from the container and resets its address arbiter adress to 0 - void HandleWakeupThread(std::shared_ptr<Thread> thread); - private: /// Signals an address being waited on. ResultCode SignalToAddressOnly(VAddr address, s32 num_to_wake); diff --git a/src/core/hle/kernel/global_scheduler_context.cpp b/src/core/hle/kernel/global_scheduler_context.cpp new file mode 100644 index 000000000..a133e8ed0 --- /dev/null +++ b/src/core/hle/kernel/global_scheduler_context.cpp @@ -0,0 +1,52 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <mutex> + +#include "common/assert.h" +#include "core/core.h" +#include "core/hle/kernel/global_scheduler_context.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/kernel.h" + +namespace Kernel { + +GlobalSchedulerContext::GlobalSchedulerContext(KernelCore& kernel) + : kernel{kernel}, scheduler_lock{kernel} {} + +GlobalSchedulerContext::~GlobalSchedulerContext() = default; + +void GlobalSchedulerContext::AddThread(std::shared_ptr<Thread> thread) { + std::scoped_lock lock{global_list_guard}; + thread_list.push_back(std::move(thread)); +} + +void GlobalSchedulerContext::RemoveThread(std::shared_ptr<Thread> thread) { + std::scoped_lock lock{global_list_guard}; + thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread), + thread_list.end()); +} + +void GlobalSchedulerContext::PreemptThreads() { + // The priority levels at which the global scheduler preempts threads every 10 ms. They are + // ordered from Core 0 to Core 3. + static constexpr std::array<u32, Core::Hardware::NUM_CPU_CORES> preemption_priorities{ + 59, + 59, + 59, + 63, + }; + + ASSERT(IsLocked()); + for (u32 core_id = 0; core_id < Core::Hardware::NUM_CPU_CORES; core_id++) { + const u32 priority = preemption_priorities[core_id]; + kernel.Scheduler(core_id).RotateScheduledQueue(core_id, priority); + } +} + +bool GlobalSchedulerContext::IsLocked() const { + return scheduler_lock.IsLockedByCurrentThread(); +} + +} // namespace Kernel diff --git a/src/core/hle/kernel/global_scheduler_context.h b/src/core/hle/kernel/global_scheduler_context.h new file mode 100644 index 000000000..5c7b89290 --- /dev/null +++ b/src/core/hle/kernel/global_scheduler_context.h @@ -0,0 +1,81 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <atomic> +#include <vector> + +#include "common/common_types.h" +#include "common/spin_lock.h" +#include "core/hardware_properties.h" +#include "core/hle/kernel/k_priority_queue.h" +#include "core/hle/kernel/k_scheduler_lock.h" +#include "core/hle/kernel/thread.h" + +namespace Kernel { + +class KernelCore; +class SchedulerLock; + +using KSchedulerPriorityQueue = + KPriorityQueue<Thread, Core::Hardware::NUM_CPU_CORES, THREADPRIO_LOWEST, THREADPRIO_HIGHEST>; +constexpr s32 HighestCoreMigrationAllowedPriority = 2; + +class GlobalSchedulerContext final { + friend class KScheduler; + +public: + using LockType = KAbstractSchedulerLock<KScheduler>; + + explicit GlobalSchedulerContext(KernelCore& kernel); + ~GlobalSchedulerContext(); + + /// Adds a new thread to the scheduler + void AddThread(std::shared_ptr<Thread> thread); + + /// Removes a thread from the scheduler + void RemoveThread(std::shared_ptr<Thread> thread); + + /// Returns a list of all threads managed by the scheduler + [[nodiscard]] const std::vector<std::shared_ptr<Thread>>& GetThreadList() const { + return thread_list; + } + + /** + * Rotates the scheduling queues of threads at a preemption priority and then does + * some core rebalancing. Preemption priorities can be found in the array + * 'preemption_priorities'. + * + * @note This operation happens every 10ms. + */ + void PreemptThreads(); + + /// Returns true if the global scheduler lock is acquired + bool IsLocked() const; + + [[nodiscard]] LockType& SchedulerLock() { + return scheduler_lock; + } + + [[nodiscard]] const LockType& SchedulerLock() const { + return scheduler_lock; + } + +private: + friend class KScopedSchedulerLock; + friend class KScopedSchedulerLockAndSleep; + + KernelCore& kernel; + + std::atomic_bool scheduler_update_needed{}; + KSchedulerPriorityQueue priority_queue; + LockType scheduler_lock; + + /// Lists all thread ids that aren't deleted/etc. + std::vector<std::shared_ptr<Thread>> thread_list; + Common::SpinLock global_list_guard{}; +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/handle_table.cpp b/src/core/hle/kernel/handle_table.cpp index fb30b6f8b..40988b0fd 100644 --- a/src/core/hle/kernel/handle_table.cpp +++ b/src/core/hle/kernel/handle_table.cpp @@ -8,9 +8,9 @@ #include "core/core.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" namespace Kernel { @@ -105,7 +105,7 @@ bool HandleTable::IsValid(Handle handle) const { std::shared_ptr<Object> HandleTable::GetGeneric(Handle handle) const { if (handle == CurrentThread) { - return SharedFrom(kernel.CurrentScheduler().GetCurrentThread()); + return SharedFrom(kernel.CurrentScheduler()->GetCurrentThread()); } else if (handle == CurrentProcess) { return SharedFrom(kernel.CurrentProcess()); } @@ -118,7 +118,7 @@ std::shared_ptr<Object> HandleTable::GetGeneric(Handle handle) const { void HandleTable::Clear() { for (u16 i = 0; i < table_size; ++i) { - generations[i] = i + 1; + generations[i] = static_cast<u16>(i + 1); objects[i] = nullptr; } next_free_slot = 0; diff --git a/src/core/hle/kernel/hle_ipc.cpp b/src/core/hle/kernel/hle_ipc.cpp index 81f85643b..83decf6cf 100644 --- a/src/core/hle/kernel/hle_ipc.cpp +++ b/src/core/hle/kernel/hle_ipc.cpp @@ -17,11 +17,12 @@ #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/hle_ipc.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/readable_event.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/server_session.h" #include "core/hle/kernel/thread.h" #include "core/hle/kernel/time_manager.h" @@ -45,44 +46,6 @@ void SessionRequestHandler::ClientDisconnected( boost::range::remove_erase(connected_sessions, server_session); } -std::shared_ptr<WritableEvent> HLERequestContext::SleepClientThread( - const std::string& reason, u64 timeout, WakeupCallback&& callback, - std::shared_ptr<WritableEvent> writable_event) { - // Put the client thread to sleep until the wait event is signaled or the timeout expires. - - if (!writable_event) { - // Create event if not provided - const auto pair = WritableEvent::CreateEventPair(kernel, "HLE Pause Event: " + reason); - writable_event = pair.writable; - } - - { - Handle event_handle = InvalidHandle; - SchedulerLockAndSleep lock(kernel, event_handle, thread.get(), timeout); - thread->SetHLECallback( - [context = *this, callback](std::shared_ptr<Thread> thread) mutable -> bool { - ThreadWakeupReason reason = thread->GetSignalingResult() == RESULT_TIMEOUT - ? ThreadWakeupReason::Timeout - : ThreadWakeupReason::Signal; - callback(thread, context, reason); - context.WriteToOutgoingCommandBuffer(*thread); - return true; - }); - const auto readable_event{writable_event->GetReadableEvent()}; - writable_event->Clear(); - thread->SetHLESyncObject(readable_event.get()); - thread->SetStatus(ThreadStatus::WaitHLEEvent); - thread->SetSynchronizationResults(nullptr, RESULT_TIMEOUT); - readable_event->AddWaitingThread(thread); - lock.Release(); - thread->SetHLETimeEvent(event_handle); - } - - is_thread_waiting = true; - - return writable_event; -} - HLERequestContext::HLERequestContext(KernelCore& kernel, Core::Memory::Memory& memory, std::shared_ptr<ServerSession> server_session, std::shared_ptr<Thread> thread) diff --git a/src/core/hle/kernel/hle_ipc.h b/src/core/hle/kernel/hle_ipc.h index f3277b766..b112e1ebd 100644 --- a/src/core/hle/kernel/hle_ipc.h +++ b/src/core/hle/kernel/hle_ipc.h @@ -24,6 +24,10 @@ namespace Core::Memory { class Memory; } +namespace IPC { +class ResponseBuilder; +} + namespace Service { class ServiceFrameworkBase; } @@ -125,23 +129,6 @@ public: using WakeupCallback = std::function<void( std::shared_ptr<Thread> thread, HLERequestContext& context, ThreadWakeupReason reason)>; - /** - * Puts the specified guest thread to sleep until the returned event is signaled or until the - * specified timeout expires. - * @param reason Reason for pausing the thread, to be used for debugging purposes. - * @param timeout Timeout in nanoseconds after which the thread will be awoken and the callback - * invoked with a Timeout reason. - * @param callback Callback to be invoked when the thread is resumed. This callback must write - * the entire command response once again, regardless of the state of it before this function - * was called. - * @param writable_event Event to use to wake up the thread. If unspecified, an event will be - * created. - * @returns Event that when signaled will resume the thread and call the callback function. - */ - std::shared_ptr<WritableEvent> SleepClientThread( - const std::string& reason, u64 timeout, WakeupCallback&& callback, - std::shared_ptr<WritableEvent> writable_event = nullptr); - /// Populates this context with data from the requesting process/thread. ResultCode PopulateFromIncomingCommandBuffer(const HandleTable& handle_table, u32_le* src_cmdbuf); @@ -287,6 +274,8 @@ public: } private: + friend class IPC::ResponseBuilder; + void ParseCommandBuffer(const HandleTable& handle_table, u32_le* src_cmdbuf, bool incoming); std::array<u32, IPC::COMMAND_BUFFER_LENGTH> cmd_buf; diff --git a/src/core/hle/kernel/k_affinity_mask.h b/src/core/hle/kernel/k_affinity_mask.h new file mode 100644 index 000000000..dd73781cd --- /dev/null +++ b/src/core/hle/kernel/k_affinity_mask.h @@ -0,0 +1,58 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include "common/assert.h" +#include "common/common_types.h" +#include "core/hardware_properties.h" + +namespace Kernel { + +class KAffinityMask { +public: + constexpr KAffinityMask() = default; + + [[nodiscard]] constexpr u64 GetAffinityMask() const { + return this->mask; + } + + constexpr void SetAffinityMask(u64 new_mask) { + ASSERT((new_mask & ~AllowedAffinityMask) == 0); + this->mask = new_mask; + } + + [[nodiscard]] constexpr bool GetAffinity(s32 core) const { + return this->mask & GetCoreBit(core); + } + + constexpr void SetAffinity(s32 core, bool set) { + ASSERT(0 <= core && core < static_cast<s32>(Core::Hardware::NUM_CPU_CORES)); + + if (set) { + this->mask |= GetCoreBit(core); + } else { + this->mask &= ~GetCoreBit(core); + } + } + + constexpr void SetAll() { + this->mask = AllowedAffinityMask; + } + +private: + [[nodiscard]] static constexpr u64 GetCoreBit(s32 core) { + ASSERT(0 <= core && core < static_cast<s32>(Core::Hardware::NUM_CPU_CORES)); + return (1ULL << core); + } + + static constexpr u64 AllowedAffinityMask = (1ULL << Core::Hardware::NUM_CPU_CORES) - 1; + + u64 mask{}; +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_priority_queue.h b/src/core/hle/kernel/k_priority_queue.h new file mode 100644 index 000000000..99fb8fe93 --- /dev/null +++ b/src/core/hle/kernel/k_priority_queue.h @@ -0,0 +1,451 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include <array> +#include <concepts> + +#include "common/assert.h" +#include "common/bit_set.h" +#include "common/bit_util.h" +#include "common/common_types.h" +#include "common/concepts.h" + +namespace Kernel { + +class Thread; + +template <typename T> +concept KPriorityQueueAffinityMask = !std::is_reference_v<T> && requires(T & t) { + { t.GetAffinityMask() } + ->Common::ConvertibleTo<u64>; + {t.SetAffinityMask(std::declval<u64>())}; + + { t.GetAffinity(std::declval<int32_t>()) } + ->std::same_as<bool>; + {t.SetAffinity(std::declval<int32_t>(), std::declval<bool>())}; + {t.SetAll()}; +}; + +template <typename T> +concept KPriorityQueueMember = !std::is_reference_v<T> && requires(T & t) { + {typename T::QueueEntry()}; + {(typename T::QueueEntry()).Initialize()}; + {(typename T::QueueEntry()).SetPrev(std::addressof(t))}; + {(typename T::QueueEntry()).SetNext(std::addressof(t))}; + { (typename T::QueueEntry()).GetNext() } + ->std::same_as<T*>; + { (typename T::QueueEntry()).GetPrev() } + ->std::same_as<T*>; + { t.GetPriorityQueueEntry(std::declval<s32>()) } + ->std::same_as<typename T::QueueEntry&>; + + {t.GetAffinityMask()}; + { typename std::remove_cvref<decltype(t.GetAffinityMask())>::type() } + ->KPriorityQueueAffinityMask; + + { t.GetActiveCore() } + ->Common::ConvertibleTo<s32>; + { t.GetPriority() } + ->Common::ConvertibleTo<s32>; +}; + +template <typename Member, size_t _NumCores, int LowestPriority, int HighestPriority> +requires KPriorityQueueMember<Member> class KPriorityQueue { +public: + using AffinityMaskType = typename std::remove_cv_t< + typename std::remove_reference<decltype(std::declval<Member>().GetAffinityMask())>::type>; + + static_assert(LowestPriority >= 0); + static_assert(HighestPriority >= 0); + static_assert(LowestPriority >= HighestPriority); + static constexpr size_t NumPriority = LowestPriority - HighestPriority + 1; + static constexpr size_t NumCores = _NumCores; + + static constexpr bool IsValidCore(s32 core) { + return 0 <= core && core < static_cast<s32>(NumCores); + } + + static constexpr bool IsValidPriority(s32 priority) { + return HighestPriority <= priority && priority <= LowestPriority + 1; + } + +private: + using Entry = typename Member::QueueEntry; + +public: + class KPerCoreQueue { + private: + std::array<Entry, NumCores> root{}; + + public: + constexpr KPerCoreQueue() { + for (auto& per_core_root : root) { + per_core_root.Initialize(); + } + } + + constexpr bool PushBack(s32 core, Member* member) { + // Get the entry associated with the member. + Entry& member_entry = member->GetPriorityQueueEntry(core); + + // Get the entry associated with the end of the queue. + Member* tail = this->root[core].GetPrev(); + Entry& tail_entry = + (tail != nullptr) ? tail->GetPriorityQueueEntry(core) : this->root[core]; + + // Link the entries. + member_entry.SetPrev(tail); + member_entry.SetNext(nullptr); + tail_entry.SetNext(member); + this->root[core].SetPrev(member); + + return tail == nullptr; + } + + constexpr bool PushFront(s32 core, Member* member) { + // Get the entry associated with the member. + Entry& member_entry = member->GetPriorityQueueEntry(core); + + // Get the entry associated with the front of the queue. + Member* head = this->root[core].GetNext(); + Entry& head_entry = + (head != nullptr) ? head->GetPriorityQueueEntry(core) : this->root[core]; + + // Link the entries. + member_entry.SetPrev(nullptr); + member_entry.SetNext(head); + head_entry.SetPrev(member); + this->root[core].SetNext(member); + + return (head == nullptr); + } + + constexpr bool Remove(s32 core, Member* member) { + // Get the entry associated with the member. + Entry& member_entry = member->GetPriorityQueueEntry(core); + + // Get the entries associated with next and prev. + Member* prev = member_entry.GetPrev(); + Member* next = member_entry.GetNext(); + Entry& prev_entry = + (prev != nullptr) ? prev->GetPriorityQueueEntry(core) : this->root[core]; + Entry& next_entry = + (next != nullptr) ? next->GetPriorityQueueEntry(core) : this->root[core]; + + // Unlink. + prev_entry.SetNext(next); + next_entry.SetPrev(prev); + + return (this->GetFront(core) == nullptr); + } + + constexpr Member* GetFront(s32 core) const { + return this->root[core].GetNext(); + } + }; + + class KPriorityQueueImpl { + public: + constexpr KPriorityQueueImpl() = default; + + constexpr void PushBack(s32 priority, s32 core, Member* member) { + ASSERT(IsValidCore(core)); + ASSERT(IsValidPriority(priority)); + + if (priority > LowestPriority) { + return; + } + + if (this->queues[priority].PushBack(core, member)) { + this->available_priorities[core].SetBit(priority); + } + } + + constexpr void PushFront(s32 priority, s32 core, Member* member) { + ASSERT(IsValidCore(core)); + ASSERT(IsValidPriority(priority)); + + if (priority > LowestPriority) { + return; + } + + if (this->queues[priority].PushFront(core, member)) { + this->available_priorities[core].SetBit(priority); + } + } + + constexpr void Remove(s32 priority, s32 core, Member* member) { + ASSERT(IsValidCore(core)); + ASSERT(IsValidPriority(priority)); + + if (priority > LowestPriority) { + return; + } + + if (this->queues[priority].Remove(core, member)) { + this->available_priorities[core].ClearBit(priority); + } + } + + constexpr Member* GetFront(s32 core) const { + ASSERT(IsValidCore(core)); + + const s32 priority = + static_cast<s32>(this->available_priorities[core].CountLeadingZero()); + if (priority <= LowestPriority) { + return this->queues[priority].GetFront(core); + } else { + return nullptr; + } + } + + constexpr Member* GetFront(s32 priority, s32 core) const { + ASSERT(IsValidCore(core)); + ASSERT(IsValidPriority(priority)); + + if (priority <= LowestPriority) { + return this->queues[priority].GetFront(core); + } else { + return nullptr; + } + } + + constexpr Member* GetNext(s32 core, const Member* member) const { + ASSERT(IsValidCore(core)); + + Member* next = member->GetPriorityQueueEntry(core).GetNext(); + if (next == nullptr) { + const s32 priority = static_cast<s32>( + this->available_priorities[core].GetNextSet(member->GetPriority())); + if (priority <= LowestPriority) { + next = this->queues[priority].GetFront(core); + } + } + return next; + } + + constexpr void MoveToFront(s32 priority, s32 core, Member* member) { + ASSERT(IsValidCore(core)); + ASSERT(IsValidPriority(priority)); + + if (priority <= LowestPriority) { + this->queues[priority].Remove(core, member); + this->queues[priority].PushFront(core, member); + } + } + + constexpr Member* MoveToBack(s32 priority, s32 core, Member* member) { + ASSERT(IsValidCore(core)); + ASSERT(IsValidPriority(priority)); + + if (priority <= LowestPriority) { + this->queues[priority].Remove(core, member); + this->queues[priority].PushBack(core, member); + return this->queues[priority].GetFront(core); + } else { + return nullptr; + } + } + + private: + std::array<KPerCoreQueue, NumPriority> queues{}; + std::array<Common::BitSet64<NumPriority>, NumCores> available_priorities{}; + }; + +private: + KPriorityQueueImpl scheduled_queue; + KPriorityQueueImpl suggested_queue; + +private: + constexpr void ClearAffinityBit(u64& affinity, s32 core) { + affinity &= ~(u64(1) << core); + } + + constexpr s32 GetNextCore(u64& affinity) { + const s32 core = Common::CountTrailingZeroes64(affinity); + ClearAffinityBit(affinity, core); + return core; + } + + constexpr void PushBack(s32 priority, Member* member) { + ASSERT(IsValidPriority(priority)); + + // Push onto the scheduled queue for its core, if we can. + u64 affinity = member->GetAffinityMask().GetAffinityMask(); + if (const s32 core = member->GetActiveCore(); core >= 0) { + this->scheduled_queue.PushBack(priority, core, member); + ClearAffinityBit(affinity, core); + } + + // And suggest the thread for all other cores. + while (affinity) { + this->suggested_queue.PushBack(priority, GetNextCore(affinity), member); + } + } + + constexpr void PushFront(s32 priority, Member* member) { + ASSERT(IsValidPriority(priority)); + + // Push onto the scheduled queue for its core, if we can. + u64 affinity = member->GetAffinityMask().GetAffinityMask(); + if (const s32 core = member->GetActiveCore(); core >= 0) { + this->scheduled_queue.PushFront(priority, core, member); + ClearAffinityBit(affinity, core); + } + + // And suggest the thread for all other cores. + // Note: Nintendo pushes onto the back of the suggested queue, not the front. + while (affinity) { + this->suggested_queue.PushBack(priority, GetNextCore(affinity), member); + } + } + + constexpr void Remove(s32 priority, Member* member) { + ASSERT(IsValidPriority(priority)); + + // Remove from the scheduled queue for its core. + u64 affinity = member->GetAffinityMask().GetAffinityMask(); + if (const s32 core = member->GetActiveCore(); core >= 0) { + this->scheduled_queue.Remove(priority, core, member); + ClearAffinityBit(affinity, core); + } + + // Remove from the suggested queue for all other cores. + while (affinity) { + this->suggested_queue.Remove(priority, GetNextCore(affinity), member); + } + } + +public: + constexpr KPriorityQueue() = default; + + // Getters. + constexpr Member* GetScheduledFront(s32 core) const { + return this->scheduled_queue.GetFront(core); + } + + constexpr Member* GetScheduledFront(s32 core, s32 priority) const { + return this->scheduled_queue.GetFront(priority, core); + } + + constexpr Member* GetSuggestedFront(s32 core) const { + return this->suggested_queue.GetFront(core); + } + + constexpr Member* GetSuggestedFront(s32 core, s32 priority) const { + return this->suggested_queue.GetFront(priority, core); + } + + constexpr Member* GetScheduledNext(s32 core, const Member* member) const { + return this->scheduled_queue.GetNext(core, member); + } + + constexpr Member* GetSuggestedNext(s32 core, const Member* member) const { + return this->suggested_queue.GetNext(core, member); + } + + constexpr Member* GetSamePriorityNext(s32 core, const Member* member) const { + return member->GetPriorityQueueEntry(core).GetNext(); + } + + // Mutators. + constexpr void PushBack(Member* member) { + this->PushBack(member->GetPriority(), member); + } + + constexpr void Remove(Member* member) { + this->Remove(member->GetPriority(), member); + } + + constexpr void MoveToScheduledFront(Member* member) { + this->scheduled_queue.MoveToFront(member->GetPriority(), member->GetActiveCore(), member); + } + + constexpr Thread* MoveToScheduledBack(Member* member) { + return this->scheduled_queue.MoveToBack(member->GetPriority(), member->GetActiveCore(), + member); + } + + // First class fancy operations. + constexpr void ChangePriority(s32 prev_priority, bool is_running, Member* member) { + ASSERT(IsValidPriority(prev_priority)); + + // Remove the member from the queues. + const s32 new_priority = member->GetPriority(); + this->Remove(prev_priority, member); + + // And enqueue. If the member is running, we want to keep it running. + if (is_running) { + this->PushFront(new_priority, member); + } else { + this->PushBack(new_priority, member); + } + } + + constexpr void ChangeAffinityMask(s32 prev_core, const AffinityMaskType& prev_affinity, + Member* member) { + // Get the new information. + const s32 priority = member->GetPriority(); + const AffinityMaskType& new_affinity = member->GetAffinityMask(); + const s32 new_core = member->GetActiveCore(); + + // Remove the member from all queues it was in before. + for (s32 core = 0; core < static_cast<s32>(NumCores); core++) { + if (prev_affinity.GetAffinity(core)) { + if (core == prev_core) { + this->scheduled_queue.Remove(priority, core, member); + } else { + this->suggested_queue.Remove(priority, core, member); + } + } + } + + // And add the member to all queues it should be in now. + for (s32 core = 0; core < static_cast<s32>(NumCores); core++) { + if (new_affinity.GetAffinity(core)) { + if (core == new_core) { + this->scheduled_queue.PushBack(priority, core, member); + } else { + this->suggested_queue.PushBack(priority, core, member); + } + } + } + } + + constexpr void ChangeCore(s32 prev_core, Member* member, bool to_front = false) { + // Get the new information. + const s32 new_core = member->GetActiveCore(); + const s32 priority = member->GetPriority(); + + // We don't need to do anything if the core is the same. + if (prev_core != new_core) { + // Remove from the scheduled queue for the previous core. + if (prev_core >= 0) { + this->scheduled_queue.Remove(priority, prev_core, member); + } + + // Remove from the suggested queue and add to the scheduled queue for the new core. + if (new_core >= 0) { + this->suggested_queue.Remove(priority, new_core, member); + if (to_front) { + this->scheduled_queue.PushFront(priority, new_core, member); + } else { + this->scheduled_queue.PushBack(priority, new_core, member); + } + } + + // Add to the suggested queue for the previous core. + if (prev_core >= 0) { + this->suggested_queue.PushBack(priority, prev_core, member); + } + } + } +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_scheduler.cpp b/src/core/hle/kernel/k_scheduler.cpp new file mode 100644 index 000000000..c5fd82a6b --- /dev/null +++ b/src/core/hle/kernel/k_scheduler.cpp @@ -0,0 +1,784 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#include "common/assert.h" +#include "common/bit_util.h" +#include "common/fiber.h" +#include "common/logging/log.h" +#include "core/arm/arm_interface.h" +#include "core/core.h" +#include "core/core_timing.h" +#include "core/cpu_manager.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h" +#include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/physical_core.h" +#include "core/hle/kernel/process.h" +#include "core/hle/kernel/thread.h" +#include "core/hle/kernel/time_manager.h" + +namespace Kernel { + +static void IncrementScheduledCount(Kernel::Thread* thread) { + if (auto process = thread->GetOwnerProcess(); process) { + process->IncrementScheduledCount(); + } +} + +void KScheduler::RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedule, + Core::EmuThreadHandle global_thread) { + u32 current_core = global_thread.host_handle; + bool must_context_switch = global_thread.guest_handle != InvalidHandle && + (current_core < Core::Hardware::NUM_CPU_CORES); + + while (cores_pending_reschedule != 0) { + u32 core = Common::CountTrailingZeroes64(cores_pending_reschedule); + ASSERT(core < Core::Hardware::NUM_CPU_CORES); + if (!must_context_switch || core != current_core) { + auto& phys_core = kernel.PhysicalCore(core); + phys_core.Interrupt(); + } else { + must_context_switch = true; + } + cores_pending_reschedule &= ~(1ULL << core); + } + if (must_context_switch) { + auto core_scheduler = kernel.CurrentScheduler(); + kernel.ExitSVCProfile(); + core_scheduler->RescheduleCurrentCore(); + kernel.EnterSVCProfile(); + } +} + +u64 KScheduler::UpdateHighestPriorityThread(Thread* highest_thread) { + std::scoped_lock lock{guard}; + if (Thread* prev_highest_thread = this->state.highest_priority_thread; + prev_highest_thread != highest_thread) { + if (prev_highest_thread != nullptr) { + IncrementScheduledCount(prev_highest_thread); + prev_highest_thread->SetLastScheduledTick(system.CoreTiming().GetCPUTicks()); + } + if (this->state.should_count_idle) { + if (highest_thread != nullptr) { + // if (Process* process = highest_thread->GetOwnerProcess(); process != nullptr) { + // process->SetRunningThread(this->core_id, highest_thread, + // this->state.idle_count); + //} + } else { + this->state.idle_count++; + } + } + + this->state.highest_priority_thread = highest_thread; + this->state.needs_scheduling = true; + return (1ULL << this->core_id); + } else { + return 0; + } +} + +u64 KScheduler::UpdateHighestPriorityThreadsImpl(KernelCore& kernel) { + ASSERT(kernel.GlobalSchedulerContext().IsLocked()); + + // Clear that we need to update. + ClearSchedulerUpdateNeeded(kernel); + + u64 cores_needing_scheduling = 0, idle_cores = 0; + Thread* top_threads[Core::Hardware::NUM_CPU_CORES]; + auto& priority_queue = GetPriorityQueue(kernel); + + /// We want to go over all cores, finding the highest priority thread and determining if + /// scheduling is needed for that core. + for (size_t core_id = 0; core_id < Core::Hardware::NUM_CPU_CORES; core_id++) { + Thread* top_thread = priority_queue.GetScheduledFront(static_cast<s32>(core_id)); + if (top_thread != nullptr) { + // If the thread has no waiters, we need to check if the process has a thread pinned. + // TODO(bunnei): Implement thread pinning + } else { + idle_cores |= (1ULL << core_id); + } + + top_threads[core_id] = top_thread; + cores_needing_scheduling |= + kernel.Scheduler(core_id).UpdateHighestPriorityThread(top_threads[core_id]); + } + + // Idle cores are bad. We're going to try to migrate threads to each idle core in turn. + while (idle_cores != 0) { + u32 core_id = Common::CountTrailingZeroes64(idle_cores); + if (Thread* suggested = priority_queue.GetSuggestedFront(core_id); suggested != nullptr) { + s32 migration_candidates[Core::Hardware::NUM_CPU_CORES]; + size_t num_candidates = 0; + + // While we have a suggested thread, try to migrate it! + while (suggested != nullptr) { + // Check if the suggested thread is the top thread on its core. + const s32 suggested_core = suggested->GetActiveCore(); + if (Thread* top_thread = + (suggested_core >= 0) ? top_threads[suggested_core] : nullptr; + top_thread != suggested) { + // Make sure we're not dealing with threads too high priority for migration. + if (top_thread != nullptr && + top_thread->GetPriority() < HighestCoreMigrationAllowedPriority) { + break; + } + + // The suggested thread isn't bound to its core, so we can migrate it! + suggested->SetActiveCore(core_id); + priority_queue.ChangeCore(suggested_core, suggested); + + top_threads[core_id] = suggested; + cores_needing_scheduling |= + kernel.Scheduler(core_id).UpdateHighestPriorityThread(top_threads[core_id]); + break; + } + + // Note this core as a candidate for migration. + ASSERT(num_candidates < Core::Hardware::NUM_CPU_CORES); + migration_candidates[num_candidates++] = suggested_core; + suggested = priority_queue.GetSuggestedNext(core_id, suggested); + } + + // If suggested is nullptr, we failed to migrate a specific thread. So let's try all our + // candidate cores' top threads. + if (suggested == nullptr) { + for (size_t i = 0; i < num_candidates; i++) { + // Check if there's some other thread that can run on the candidate core. + const s32 candidate_core = migration_candidates[i]; + suggested = top_threads[candidate_core]; + if (Thread* next_on_candidate_core = + priority_queue.GetScheduledNext(candidate_core, suggested); + next_on_candidate_core != nullptr) { + // The candidate core can run some other thread! We'll migrate its current + // top thread to us. + top_threads[candidate_core] = next_on_candidate_core; + cores_needing_scheduling |= + kernel.Scheduler(candidate_core) + .UpdateHighestPriorityThread(top_threads[candidate_core]); + + // Perform the migration. + suggested->SetActiveCore(core_id); + priority_queue.ChangeCore(candidate_core, suggested); + + top_threads[core_id] = suggested; + cores_needing_scheduling |= + kernel.Scheduler(core_id).UpdateHighestPriorityThread( + top_threads[core_id]); + break; + } + } + } + } + + idle_cores &= ~(1ULL << core_id); + } + + return cores_needing_scheduling; +} + +void KScheduler::OnThreadStateChanged(KernelCore& kernel, Thread* thread, u32 old_state) { + ASSERT(kernel.GlobalSchedulerContext().IsLocked()); + + // Check if the state has changed, because if it hasn't there's nothing to do. + const auto cur_state = thread->scheduling_state; + if (cur_state == old_state) { + return; + } + + // Update the priority queues. + if (old_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + // If we were previously runnable, then we're not runnable now, and we should remove. + GetPriorityQueue(kernel).Remove(thread); + IncrementScheduledCount(thread); + SetSchedulerUpdateNeeded(kernel); + } else if (cur_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + // If we're now runnable, then we weren't previously, and we should add. + GetPriorityQueue(kernel).PushBack(thread); + IncrementScheduledCount(thread); + SetSchedulerUpdateNeeded(kernel); + } +} + +void KScheduler::OnThreadPriorityChanged(KernelCore& kernel, Thread* thread, Thread* current_thread, + u32 old_priority) { + + ASSERT(kernel.GlobalSchedulerContext().IsLocked()); + + // If the thread is runnable, we want to change its priority in the queue. + if (thread->scheduling_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + GetPriorityQueue(kernel).ChangePriority( + old_priority, thread == kernel.CurrentScheduler()->GetCurrentThread(), thread); + IncrementScheduledCount(thread); + SetSchedulerUpdateNeeded(kernel); + } +} + +void KScheduler::OnThreadAffinityMaskChanged(KernelCore& kernel, Thread* thread, + const KAffinityMask& old_affinity, s32 old_core) { + ASSERT(kernel.GlobalSchedulerContext().IsLocked()); + + // If the thread is runnable, we want to change its affinity in the queue. + if (thread->scheduling_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + GetPriorityQueue(kernel).ChangeAffinityMask(old_core, old_affinity, thread); + IncrementScheduledCount(thread); + SetSchedulerUpdateNeeded(kernel); + } +} + +void KScheduler::RotateScheduledQueue(s32 core_id, s32 priority) { + ASSERT(system.GlobalSchedulerContext().IsLocked()); + + // Get a reference to the priority queue. + auto& kernel = system.Kernel(); + auto& priority_queue = GetPriorityQueue(kernel); + + // Rotate the front of the queue to the end. + Thread* top_thread = priority_queue.GetScheduledFront(core_id, priority); + Thread* next_thread = nullptr; + if (top_thread != nullptr) { + next_thread = priority_queue.MoveToScheduledBack(top_thread); + if (next_thread != top_thread) { + IncrementScheduledCount(top_thread); + IncrementScheduledCount(next_thread); + } + } + + // While we have a suggested thread, try to migrate it! + { + Thread* suggested = priority_queue.GetSuggestedFront(core_id, priority); + while (suggested != nullptr) { + // Check if the suggested thread is the top thread on its core. + const s32 suggested_core = suggested->GetActiveCore(); + if (Thread* top_on_suggested_core = + (suggested_core >= 0) ? priority_queue.GetScheduledFront(suggested_core) + : nullptr; + top_on_suggested_core != suggested) { + // If the next thread is a new thread that has been waiting longer than our + // suggestion, we prefer it to our suggestion. + if (top_thread != next_thread && next_thread != nullptr && + next_thread->GetLastScheduledTick() < suggested->GetLastScheduledTick()) { + suggested = nullptr; + break; + } + + // If we're allowed to do a migration, do one. + // NOTE: Unlike migrations in UpdateHighestPriorityThread, this moves the suggestion + // to the front of the queue. + if (top_on_suggested_core == nullptr || + top_on_suggested_core->GetPriority() >= HighestCoreMigrationAllowedPriority) { + suggested->SetActiveCore(core_id); + priority_queue.ChangeCore(suggested_core, suggested, true); + IncrementScheduledCount(suggested); + break; + } + } + + // Get the next suggestion. + suggested = priority_queue.GetSamePriorityNext(core_id, suggested); + } + } + + // Now that we might have migrated a thread with the same priority, check if we can do better. + + { + Thread* best_thread = priority_queue.GetScheduledFront(core_id); + if (best_thread == GetCurrentThread()) { + best_thread = priority_queue.GetScheduledNext(core_id, best_thread); + } + + // If the best thread we can choose has a priority the same or worse than ours, try to + // migrate a higher priority thread. + if (best_thread != nullptr && best_thread->GetPriority() >= static_cast<u32>(priority)) { + Thread* suggested = priority_queue.GetSuggestedFront(core_id); + while (suggested != nullptr) { + // If the suggestion's priority is the same as ours, don't bother. + if (suggested->GetPriority() >= best_thread->GetPriority()) { + break; + } + + // Check if the suggested thread is the top thread on its core. + const s32 suggested_core = suggested->GetActiveCore(); + if (Thread* top_on_suggested_core = + (suggested_core >= 0) ? priority_queue.GetScheduledFront(suggested_core) + : nullptr; + top_on_suggested_core != suggested) { + // If we're allowed to do a migration, do one. + // NOTE: Unlike migrations in UpdateHighestPriorityThread, this moves the + // suggestion to the front of the queue. + if (top_on_suggested_core == nullptr || + top_on_suggested_core->GetPriority() >= + HighestCoreMigrationAllowedPriority) { + suggested->SetActiveCore(core_id); + priority_queue.ChangeCore(suggested_core, suggested, true); + IncrementScheduledCount(suggested); + break; + } + } + + // Get the next suggestion. + suggested = priority_queue.GetSuggestedNext(core_id, suggested); + } + } + } + + // After a rotation, we need a scheduler update. + SetSchedulerUpdateNeeded(kernel); +} + +bool KScheduler::CanSchedule(KernelCore& kernel) { + return kernel.CurrentScheduler()->GetCurrentThread()->GetDisableDispatchCount() <= 1; +} + +bool KScheduler::IsSchedulerUpdateNeeded(const KernelCore& kernel) { + return kernel.GlobalSchedulerContext().scheduler_update_needed.load(std::memory_order_acquire); +} + +void KScheduler::SetSchedulerUpdateNeeded(KernelCore& kernel) { + kernel.GlobalSchedulerContext().scheduler_update_needed.store(true, std::memory_order_release); +} + +void KScheduler::ClearSchedulerUpdateNeeded(KernelCore& kernel) { + kernel.GlobalSchedulerContext().scheduler_update_needed.store(false, std::memory_order_release); +} + +void KScheduler::DisableScheduling(KernelCore& kernel) { + if (auto* scheduler = kernel.CurrentScheduler(); scheduler) { + ASSERT(scheduler->GetCurrentThread()->GetDisableDispatchCount() >= 0); + scheduler->GetCurrentThread()->DisableDispatch(); + } +} + +void KScheduler::EnableScheduling(KernelCore& kernel, u64 cores_needing_scheduling, + Core::EmuThreadHandle global_thread) { + if (auto* scheduler = kernel.CurrentScheduler(); scheduler) { + scheduler->GetCurrentThread()->EnableDispatch(); + } + RescheduleCores(kernel, cores_needing_scheduling, global_thread); +} + +u64 KScheduler::UpdateHighestPriorityThreads(KernelCore& kernel) { + if (IsSchedulerUpdateNeeded(kernel)) { + return UpdateHighestPriorityThreadsImpl(kernel); + } else { + return 0; + } +} + +KSchedulerPriorityQueue& KScheduler::GetPriorityQueue(KernelCore& kernel) { + return kernel.GlobalSchedulerContext().priority_queue; +} + +void KScheduler::YieldWithoutCoreMigration() { + auto& kernel = system.Kernel(); + + // Validate preconditions. + ASSERT(CanSchedule(kernel)); + ASSERT(kernel.CurrentProcess() != nullptr); + + // Get the current thread and process. + Thread& cur_thread = *GetCurrentThread(); + Process& cur_process = *kernel.CurrentProcess(); + + // If the thread's yield count matches, there's nothing for us to do. + if (cur_thread.GetYieldScheduleCount() == cur_process.GetScheduledCount()) { + return; + } + + // Get a reference to the priority queue. + auto& priority_queue = GetPriorityQueue(kernel); + + // Perform the yield. + { + KScopedSchedulerLock lock(kernel); + + const auto cur_state = cur_thread.scheduling_state; + if (cur_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + // Put the current thread at the back of the queue. + Thread* next_thread = priority_queue.MoveToScheduledBack(std::addressof(cur_thread)); + IncrementScheduledCount(std::addressof(cur_thread)); + + // If the next thread is different, we have an update to perform. + if (next_thread != std::addressof(cur_thread)) { + SetSchedulerUpdateNeeded(kernel); + } else { + // Otherwise, set the thread's yield count so that we won't waste work until the + // process is scheduled again. + cur_thread.SetYieldScheduleCount(cur_process.GetScheduledCount()); + } + } + } +} + +void KScheduler::YieldWithCoreMigration() { + auto& kernel = system.Kernel(); + + // Validate preconditions. + ASSERT(CanSchedule(kernel)); + ASSERT(kernel.CurrentProcess() != nullptr); + + // Get the current thread and process. + Thread& cur_thread = *GetCurrentThread(); + Process& cur_process = *kernel.CurrentProcess(); + + // If the thread's yield count matches, there's nothing for us to do. + if (cur_thread.GetYieldScheduleCount() == cur_process.GetScheduledCount()) { + return; + } + + // Get a reference to the priority queue. + auto& priority_queue = GetPriorityQueue(kernel); + + // Perform the yield. + { + KScopedSchedulerLock lock(kernel); + + const auto cur_state = cur_thread.scheduling_state; + if (cur_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + // Get the current active core. + const s32 core_id = cur_thread.GetActiveCore(); + + // Put the current thread at the back of the queue. + Thread* next_thread = priority_queue.MoveToScheduledBack(std::addressof(cur_thread)); + IncrementScheduledCount(std::addressof(cur_thread)); + + // While we have a suggested thread, try to migrate it! + bool recheck = false; + Thread* suggested = priority_queue.GetSuggestedFront(core_id); + while (suggested != nullptr) { + // Check if the suggested thread is the thread running on its core. + const s32 suggested_core = suggested->GetActiveCore(); + + if (Thread* running_on_suggested_core = + (suggested_core >= 0) + ? kernel.Scheduler(suggested_core).state.highest_priority_thread + : nullptr; + running_on_suggested_core != suggested) { + // If the current thread's priority is higher than our suggestion's we prefer + // the next thread to the suggestion. We also prefer the next thread when the + // current thread's priority is equal to the suggestions, but the next thread + // has been waiting longer. + if ((suggested->GetPriority() > cur_thread.GetPriority()) || + (suggested->GetPriority() == cur_thread.GetPriority() && + next_thread != std::addressof(cur_thread) && + next_thread->GetLastScheduledTick() < suggested->GetLastScheduledTick())) { + suggested = nullptr; + break; + } + + // If we're allowed to do a migration, do one. + // NOTE: Unlike migrations in UpdateHighestPriorityThread, this moves the + // suggestion to the front of the queue. + if (running_on_suggested_core == nullptr || + running_on_suggested_core->GetPriority() >= + HighestCoreMigrationAllowedPriority) { + suggested->SetActiveCore(core_id); + priority_queue.ChangeCore(suggested_core, suggested, true); + IncrementScheduledCount(suggested); + break; + } else { + // We couldn't perform a migration, but we should check again on a future + // yield. + recheck = true; + } + } + + // Get the next suggestion. + suggested = priority_queue.GetSuggestedNext(core_id, suggested); + } + + // If we still have a suggestion or the next thread is different, we have an update to + // perform. + if (suggested != nullptr || next_thread != std::addressof(cur_thread)) { + SetSchedulerUpdateNeeded(kernel); + } else if (!recheck) { + // Otherwise if we don't need to re-check, set the thread's yield count so that we + // won't waste work until the process is scheduled again. + cur_thread.SetYieldScheduleCount(cur_process.GetScheduledCount()); + } + } + } +} + +void KScheduler::YieldToAnyThread() { + auto& kernel = system.Kernel(); + + // Validate preconditions. + ASSERT(CanSchedule(kernel)); + ASSERT(kernel.CurrentProcess() != nullptr); + + // Get the current thread and process. + Thread& cur_thread = *GetCurrentThread(); + Process& cur_process = *kernel.CurrentProcess(); + + // If the thread's yield count matches, there's nothing for us to do. + if (cur_thread.GetYieldScheduleCount() == cur_process.GetScheduledCount()) { + return; + } + + // Get a reference to the priority queue. + auto& priority_queue = GetPriorityQueue(kernel); + + // Perform the yield. + { + KScopedSchedulerLock lock(kernel); + + const auto cur_state = cur_thread.scheduling_state; + if (cur_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { + // Get the current active core. + const s32 core_id = cur_thread.GetActiveCore(); + + // Migrate the current thread to core -1. + cur_thread.SetActiveCore(-1); + priority_queue.ChangeCore(core_id, std::addressof(cur_thread)); + IncrementScheduledCount(std::addressof(cur_thread)); + + // If there's nothing scheduled, we can try to perform a migration. + if (priority_queue.GetScheduledFront(core_id) == nullptr) { + // While we have a suggested thread, try to migrate it! + Thread* suggested = priority_queue.GetSuggestedFront(core_id); + while (suggested != nullptr) { + // Check if the suggested thread is the top thread on its core. + const s32 suggested_core = suggested->GetActiveCore(); + if (Thread* top_on_suggested_core = + (suggested_core >= 0) ? priority_queue.GetScheduledFront(suggested_core) + : nullptr; + top_on_suggested_core != suggested) { + // If we're allowed to do a migration, do one. + if (top_on_suggested_core == nullptr || + top_on_suggested_core->GetPriority() >= + HighestCoreMigrationAllowedPriority) { + suggested->SetActiveCore(core_id); + priority_queue.ChangeCore(suggested_core, suggested); + IncrementScheduledCount(suggested); + } + + // Regardless of whether we migrated, we had a candidate, so we're done. + break; + } + + // Get the next suggestion. + suggested = priority_queue.GetSuggestedNext(core_id, suggested); + } + + // If the suggestion is different from the current thread, we need to perform an + // update. + if (suggested != std::addressof(cur_thread)) { + SetSchedulerUpdateNeeded(kernel); + } else { + // Otherwise, set the thread's yield count so that we won't waste work until the + // process is scheduled again. + cur_thread.SetYieldScheduleCount(cur_process.GetScheduledCount()); + } + } else { + // Otherwise, we have an update to perform. + SetSchedulerUpdateNeeded(kernel); + } + } + } +} + +KScheduler::KScheduler(Core::System& system, std::size_t core_id) + : system(system), core_id(core_id) { + switch_fiber = std::make_shared<Common::Fiber>(OnSwitch, this); + this->state.needs_scheduling = true; + this->state.interrupt_task_thread_runnable = false; + this->state.should_count_idle = false; + this->state.idle_count = 0; + this->state.idle_thread_stack = nullptr; + this->state.highest_priority_thread = nullptr; +} + +KScheduler::~KScheduler() = default; + +Thread* KScheduler::GetCurrentThread() const { + if (current_thread) { + return current_thread; + } + return idle_thread; +} + +u64 KScheduler::GetLastContextSwitchTicks() const { + return last_context_switch_time; +} + +void KScheduler::RescheduleCurrentCore() { + ASSERT(GetCurrentThread()->GetDisableDispatchCount() == 1); + + auto& phys_core = system.Kernel().PhysicalCore(core_id); + if (phys_core.IsInterrupted()) { + phys_core.ClearInterrupt(); + } + guard.lock(); + if (this->state.needs_scheduling) { + Schedule(); + } else { + guard.unlock(); + } +} + +void KScheduler::OnThreadStart() { + SwitchContextStep2(); +} + +void KScheduler::Unload(Thread* thread) { + if (thread) { + thread->SetIsRunning(false); + if (thread->IsContinuousOnSVC() && !thread->IsHLEThread()) { + system.ArmInterface(core_id).ExceptionalExit(); + thread->SetContinuousOnSVC(false); + } + if (!thread->IsHLEThread() && !thread->HasExited()) { + Core::ARM_Interface& cpu_core = system.ArmInterface(core_id); + cpu_core.SaveContext(thread->GetContext32()); + cpu_core.SaveContext(thread->GetContext64()); + // Save the TPIDR_EL0 system register in case it was modified. + thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0()); + cpu_core.ClearExclusiveState(); + } + thread->context_guard.unlock(); + } +} + +void KScheduler::Reload(Thread* thread) { + if (thread) { + ASSERT_MSG(thread->GetSchedulingStatus() == ThreadSchedStatus::Runnable, + "Thread must be runnable."); + + // Cancel any outstanding wakeup events for this thread + thread->SetIsRunning(true); + thread->SetWasRunning(false); + + auto* const thread_owner_process = thread->GetOwnerProcess(); + if (thread_owner_process != nullptr) { + system.Kernel().MakeCurrentProcess(thread_owner_process); + } + if (!thread->IsHLEThread()) { + Core::ARM_Interface& cpu_core = system.ArmInterface(core_id); + cpu_core.LoadContext(thread->GetContext32()); + cpu_core.LoadContext(thread->GetContext64()); + cpu_core.SetTlsAddress(thread->GetTLSAddress()); + cpu_core.SetTPIDR_EL0(thread->GetTPIDR_EL0()); + cpu_core.ClearExclusiveState(); + } + } +} + +void KScheduler::SwitchContextStep2() { + // Load context of new thread + Reload(current_thread); + + RescheduleCurrentCore(); +} + +void KScheduler::ScheduleImpl() { + Thread* previous_thread = current_thread; + current_thread = state.highest_priority_thread; + + this->state.needs_scheduling = false; + + if (current_thread == previous_thread) { + guard.unlock(); + return; + } + + Process* const previous_process = system.Kernel().CurrentProcess(); + + UpdateLastContextSwitchTime(previous_thread, previous_process); + + // Save context for previous thread + Unload(previous_thread); + + std::shared_ptr<Common::Fiber>* old_context; + if (previous_thread != nullptr) { + old_context = &previous_thread->GetHostContext(); + } else { + old_context = &idle_thread->GetHostContext(); + } + guard.unlock(); + + Common::Fiber::YieldTo(*old_context, switch_fiber); + /// When a thread wakes up, the scheduler may have changed to other in another core. + auto& next_scheduler = *system.Kernel().CurrentScheduler(); + next_scheduler.SwitchContextStep2(); +} + +void KScheduler::OnSwitch(void* this_scheduler) { + KScheduler* sched = static_cast<KScheduler*>(this_scheduler); + sched->SwitchToCurrent(); +} + +void KScheduler::SwitchToCurrent() { + while (true) { + { + std::scoped_lock lock{guard}; + current_thread = state.highest_priority_thread; + this->state.needs_scheduling = false; + } + const auto is_switch_pending = [this] { + std::scoped_lock lock{guard}; + return state.needs_scheduling.load(std::memory_order_relaxed); + }; + do { + if (current_thread != nullptr && !current_thread->IsHLEThread()) { + current_thread->context_guard.lock(); + if (!current_thread->IsRunnable()) { + current_thread->context_guard.unlock(); + break; + } + if (static_cast<u32>(current_thread->GetProcessorID()) != core_id) { + current_thread->context_guard.unlock(); + break; + } + } + std::shared_ptr<Common::Fiber>* next_context; + if (current_thread != nullptr) { + next_context = ¤t_thread->GetHostContext(); + } else { + next_context = &idle_thread->GetHostContext(); + } + Common::Fiber::YieldTo(switch_fiber, *next_context); + } while (!is_switch_pending()); + } +} + +void KScheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) { + const u64 prev_switch_ticks = last_context_switch_time; + const u64 most_recent_switch_ticks = system.CoreTiming().GetCPUTicks(); + const u64 update_ticks = most_recent_switch_ticks - prev_switch_ticks; + + if (thread != nullptr) { + thread->UpdateCPUTimeTicks(update_ticks); + } + + if (process != nullptr) { + process->UpdateCPUTimeTicks(update_ticks); + } + + last_context_switch_time = most_recent_switch_ticks; +} + +void KScheduler::Initialize() { + std::string name = "Idle Thread Id:" + std::to_string(core_id); + std::function<void(void*)> init_func = Core::CpuManager::GetIdleThreadStartFunc(); + void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater(); + ThreadType type = static_cast<ThreadType>(THREADTYPE_KERNEL | THREADTYPE_HLE | THREADTYPE_IDLE); + auto thread_res = Thread::Create(system, type, name, 0, 64, 0, static_cast<u32>(core_id), 0, + nullptr, std::move(init_func), init_func_parameter); + idle_thread = thread_res.Unwrap().get(); + + { + KScopedSchedulerLock lock{system.Kernel()}; + idle_thread->SetStatus(ThreadStatus::Ready); + } +} + +KScopedSchedulerLock::KScopedSchedulerLock(KernelCore& kernel) + : KScopedLock(kernel.GlobalSchedulerContext().SchedulerLock()) {} + +KScopedSchedulerLock::~KScopedSchedulerLock() = default; + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_scheduler.h b/src/core/hle/kernel/k_scheduler.h new file mode 100644 index 000000000..e84abc84c --- /dev/null +++ b/src/core/hle/kernel/k_scheduler.h @@ -0,0 +1,201 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include <atomic> + +#include "common/common_types.h" +#include "common/spin_lock.h" +#include "core/hle/kernel/global_scheduler_context.h" +#include "core/hle/kernel/k_priority_queue.h" +#include "core/hle/kernel/k_scheduler_lock.h" +#include "core/hle/kernel/k_scoped_lock.h" + +namespace Common { +class Fiber; +} + +namespace Core { +class System; +} + +namespace Kernel { + +class KernelCore; +class Process; +class SchedulerLock; +class Thread; + +class KScheduler final { +public: + explicit KScheduler(Core::System& system, std::size_t core_id); + ~KScheduler(); + + /// Reschedules to the next available thread (call after current thread is suspended) + void RescheduleCurrentCore(); + + /// Reschedules cores pending reschedule, to be called on EnableScheduling. + static void RescheduleCores(KernelCore& kernel, u64 cores_pending_reschedule, + Core::EmuThreadHandle global_thread); + + /// The next two are for SingleCore Only. + /// Unload current thread before preempting core. + void Unload(Thread* thread); + + /// Reload current thread after core preemption. + void Reload(Thread* thread); + + /// Gets the current running thread + [[nodiscard]] Thread* GetCurrentThread() const; + + /// Gets the timestamp for the last context switch in ticks. + [[nodiscard]] u64 GetLastContextSwitchTicks() const; + + [[nodiscard]] bool ContextSwitchPending() const { + return state.needs_scheduling.load(std::memory_order_relaxed); + } + + void Initialize(); + + void OnThreadStart(); + + [[nodiscard]] std::shared_ptr<Common::Fiber>& ControlContext() { + return switch_fiber; + } + + [[nodiscard]] const std::shared_ptr<Common::Fiber>& ControlContext() const { + return switch_fiber; + } + + [[nodiscard]] u64 UpdateHighestPriorityThread(Thread* highest_thread); + + /** + * Takes a thread and moves it to the back of the it's priority list. + * + * @note This operation can be redundant and no scheduling is changed if marked as so. + */ + void YieldWithoutCoreMigration(); + + /** + * Takes a thread and moves it to the back of the it's priority list. + * Afterwards, tries to pick a suggested thread from the suggested queue that has worse time or + * a better priority than the next thread in the core. + * + * @note This operation can be redundant and no scheduling is changed if marked as so. + */ + void YieldWithCoreMigration(); + + /** + * Takes a thread and moves it out of the scheduling queue. + * and into the suggested queue. If no thread can be scheduled afterwards in that core, + * a suggested thread is obtained instead. + * + * @note This operation can be redundant and no scheduling is changed if marked as so. + */ + void YieldToAnyThread(); + + /// Notify the scheduler a thread's status has changed. + static void OnThreadStateChanged(KernelCore& kernel, Thread* thread, u32 old_state); + + /// Notify the scheduler a thread's priority has changed. + static void OnThreadPriorityChanged(KernelCore& kernel, Thread* thread, Thread* current_thread, + u32 old_priority); + + /// Notify the scheduler a thread's core and/or affinity mask has changed. + static void OnThreadAffinityMaskChanged(KernelCore& kernel, Thread* thread, + const KAffinityMask& old_affinity, s32 old_core); + + static bool CanSchedule(KernelCore& kernel); + static bool IsSchedulerUpdateNeeded(const KernelCore& kernel); + static void SetSchedulerUpdateNeeded(KernelCore& kernel); + static void ClearSchedulerUpdateNeeded(KernelCore& kernel); + static void DisableScheduling(KernelCore& kernel); + static void EnableScheduling(KernelCore& kernel, u64 cores_needing_scheduling, + Core::EmuThreadHandle global_thread); + [[nodiscard]] static u64 UpdateHighestPriorityThreads(KernelCore& kernel); + +private: + friend class GlobalSchedulerContext; + + /** + * Takes care of selecting the new scheduled threads in three steps: + * + * 1. First a thread is selected from the top of the priority queue. If no thread + * is obtained then we move to step two, else we are done. + * + * 2. Second we try to get a suggested thread that's not assigned to any core or + * that is not the top thread in that core. + * + * 3. Third is no suggested thread is found, we do a second pass and pick a running + * thread in another core and swap it with its current thread. + * + * returns the cores needing scheduling. + */ + [[nodiscard]] static u64 UpdateHighestPriorityThreadsImpl(KernelCore& kernel); + + [[nodiscard]] static KSchedulerPriorityQueue& GetPriorityQueue(KernelCore& kernel); + + void RotateScheduledQueue(s32 core_id, s32 priority); + + void Schedule() { + ASSERT(GetCurrentThread()->GetDisableDispatchCount() == 1); + this->ScheduleImpl(); + } + + /// Switches the CPU's active thread context to that of the specified thread + void ScheduleImpl(); + + /// When a thread wakes up, it must run this through it's new scheduler + void SwitchContextStep2(); + + /** + * Called on every context switch to update the internal timestamp + * This also updates the running time ticks for the given thread and + * process using the following difference: + * + * ticks += most_recent_ticks - last_context_switch_ticks + * + * The internal tick timestamp for the scheduler is simply the + * most recent tick count retrieved. No special arithmetic is + * applied to it. + */ + void UpdateLastContextSwitchTime(Thread* thread, Process* process); + + static void OnSwitch(void* this_scheduler); + void SwitchToCurrent(); + + Thread* current_thread{}; + Thread* idle_thread{}; + + std::shared_ptr<Common::Fiber> switch_fiber{}; + + struct SchedulingState { + std::atomic<bool> needs_scheduling; + bool interrupt_task_thread_runnable{}; + bool should_count_idle{}; + u64 idle_count{}; + Thread* highest_priority_thread{}; + void* idle_thread_stack{}; + }; + + SchedulingState state; + + Core::System& system; + u64 last_context_switch_time{}; + const std::size_t core_id; + + Common::SpinLock guard{}; +}; + +class KScopedSchedulerLock : KScopedLock<GlobalSchedulerContext::LockType> { +public: + explicit KScopedSchedulerLock(KernelCore& kernel); + ~KScopedSchedulerLock(); +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_scheduler_lock.h b/src/core/hle/kernel/k_scheduler_lock.h new file mode 100644 index 000000000..2f1c1f691 --- /dev/null +++ b/src/core/hle/kernel/k_scheduler_lock.h @@ -0,0 +1,75 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include "common/assert.h" +#include "common/spin_lock.h" +#include "core/hardware_properties.h" +#include "core/hle/kernel/kernel.h" + +namespace Kernel { + +class KernelCore; + +template <typename SchedulerType> +class KAbstractSchedulerLock { +public: + explicit KAbstractSchedulerLock(KernelCore& kernel) : kernel{kernel} {} + + bool IsLockedByCurrentThread() const { + return this->owner_thread == kernel.GetCurrentEmuThreadID(); + } + + void Lock() { + if (this->IsLockedByCurrentThread()) { + // If we already own the lock, we can just increment the count. + ASSERT(this->lock_count > 0); + this->lock_count++; + } else { + // Otherwise, we want to disable scheduling and acquire the spinlock. + SchedulerType::DisableScheduling(kernel); + this->spin_lock.lock(); + + // For debug, ensure that our state is valid. + ASSERT(this->lock_count == 0); + ASSERT(this->owner_thread == Core::EmuThreadHandle::InvalidHandle()); + + // Increment count, take ownership. + this->lock_count = 1; + this->owner_thread = kernel.GetCurrentEmuThreadID(); + } + } + + void Unlock() { + ASSERT(this->IsLockedByCurrentThread()); + ASSERT(this->lock_count > 0); + + // Release an instance of the lock. + if ((--this->lock_count) == 0) { + // We're no longer going to hold the lock. Take note of what cores need scheduling. + const u64 cores_needing_scheduling = + SchedulerType::UpdateHighestPriorityThreads(kernel); + Core::EmuThreadHandle leaving_thread = owner_thread; + + // Note that we no longer hold the lock, and unlock the spinlock. + this->owner_thread = Core::EmuThreadHandle::InvalidHandle(); + this->spin_lock.unlock(); + + // Enable scheduling, and perform a rescheduling operation. + SchedulerType::EnableScheduling(kernel, cores_needing_scheduling, leaving_thread); + } + } + +private: + KernelCore& kernel; + Common::SpinLock spin_lock{}; + s32 lock_count{}; + Core::EmuThreadHandle owner_thread{Core::EmuThreadHandle::InvalidHandle()}; +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_scoped_lock.h b/src/core/hle/kernel/k_scoped_lock.h new file mode 100644 index 000000000..d7cc557b2 --- /dev/null +++ b/src/core/hle/kernel/k_scoped_lock.h @@ -0,0 +1,41 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include "common/common_types.h" + +namespace Kernel { + +template <typename T> +concept KLockable = !std::is_reference_v<T> && requires(T & t) { + { t.Lock() } + ->std::same_as<void>; + { t.Unlock() } + ->std::same_as<void>; +}; + +template <typename T> +requires KLockable<T> class KScopedLock { +public: + explicit KScopedLock(T* l) : lock_ptr(l) { + this->lock_ptr->Lock(); + } + explicit KScopedLock(T& l) : KScopedLock(std::addressof(l)) { /* ... */ + } + ~KScopedLock() { + this->lock_ptr->Unlock(); + } + + KScopedLock(const KScopedLock&) = delete; + KScopedLock(KScopedLock&&) = delete; + +private: + T* lock_ptr; +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h b/src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h new file mode 100644 index 000000000..2bb3817fa --- /dev/null +++ b/src/core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h @@ -0,0 +1,50 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This file references various implementation details from Atmosphere, an open-source firmware for +// the Nintendo Switch. Copyright 2018-2020 Atmosphere-NX. + +#pragma once + +#include "common/common_types.h" +#include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/thread.h" +#include "core/hle/kernel/time_manager.h" + +namespace Kernel { + +class KScopedSchedulerLockAndSleep { +public: + explicit KScopedSchedulerLockAndSleep(KernelCore& kernel, Handle& event_handle, Thread* t, + s64 timeout) + : kernel(kernel), event_handle(event_handle), thread(t), timeout_tick(timeout) { + event_handle = InvalidHandle; + + // Lock the scheduler. + kernel.GlobalSchedulerContext().scheduler_lock.Lock(); + } + + ~KScopedSchedulerLockAndSleep() { + // Register the sleep. + if (this->timeout_tick > 0) { + kernel.TimeManager().ScheduleTimeEvent(event_handle, this->thread, this->timeout_tick); + } + + // Unlock the scheduler. + kernel.GlobalSchedulerContext().scheduler_lock.Unlock(); + } + + void CancelSleep() { + this->timeout_tick = 0; + } + +private: + KernelCore& kernel; + Handle& event_handle; + Thread* thread{}; + s64 timeout_tick{}; +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp index f2b0fe2fd..e8ece8164 100644 --- a/src/core/hle/kernel/kernel.cpp +++ b/src/core/hle/kernel/kernel.cpp @@ -7,15 +7,15 @@ #include <bitset> #include <functional> #include <memory> -#include <mutex> #include <thread> -#include <unordered_map> +#include <unordered_set> #include <utility> #include "common/assert.h" #include "common/logging/log.h" #include "common/microprofile.h" #include "common/thread.h" +#include "common/thread_worker.h" #include "core/arm/arm_interface.h" #include "core/arm/cpu_interrupt_handler.h" #include "core/arm/exclusive_monitor.h" @@ -28,6 +28,7 @@ #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/memory/memory_layout.h" #include "core/hle/kernel/memory/memory_manager.h" @@ -35,7 +36,7 @@ #include "core/hle/kernel/physical_core.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/resource_limit.h" -#include "core/hle/kernel/scheduler.h" +#include "core/hle/kernel/service_thread.h" #include "core/hle/kernel/shared_memory.h" #include "core/hle/kernel/synchronization.h" #include "core/hle/kernel/thread.h" @@ -50,17 +51,20 @@ namespace Kernel { struct KernelCore::Impl { explicit Impl(Core::System& system, KernelCore& kernel) - : global_scheduler{kernel}, synchronization{system}, time_manager{system}, - global_handle_table{kernel}, system{system} {} + : synchronization{system}, time_manager{system}, global_handle_table{kernel}, system{ + system} {} void SetMulticore(bool is_multicore) { this->is_multicore = is_multicore; } void Initialize(KernelCore& kernel) { - Shutdown(); RegisterHostThread(); + global_scheduler_context = std::make_unique<Kernel::GlobalSchedulerContext>(kernel); + service_thread_manager = + std::make_unique<Common::ThreadWorker>(1, "yuzu:ServiceThreadManager"); + InitializePhysicalCores(); InitializeSystemResourceLimit(kernel); InitializeMemoryLayout(); @@ -69,7 +73,19 @@ struct KernelCore::Impl { InitializeSuspendThreads(); } + void InitializeCores() { + for (auto& core : cores) { + core.Initialize(current_process->Is64BitProcess()); + } + } + void Shutdown() { + process_list.clear(); + + // Ensures all service threads gracefully shutdown + service_thread_manager.reset(); + service_threads.clear(); + next_object_id = 0; next_kernel_process_id = Process::InitialKIPIDMin; next_user_process_id = Process::ProcessIDMin; @@ -81,41 +97,30 @@ struct KernelCore::Impl { } } - for (std::size_t i = 0; i < cores.size(); i++) { - cores[i].Shutdown(); - schedulers[i].reset(); - } cores.clear(); - registered_core_threads.reset(); - - process_list.clear(); current_process = nullptr; system_resource_limit = nullptr; global_handle_table.Clear(); - preemption_event = nullptr; - global_scheduler.Shutdown(); + preemption_event = nullptr; named_ports.clear(); - for (auto& core : cores) { - core.Shutdown(); - } - cores.clear(); - exclusive_monitor.reset(); - host_thread_ids.clear(); + + // Next host thead ID to use, 0-3 IDs represent core threads, >3 represent others + next_host_thread_id = Core::Hardware::NUM_CPU_CORES; } void InitializePhysicalCores() { exclusive_monitor = Core::MakeExclusiveMonitor(system.Memory(), Core::Hardware::NUM_CPU_CORES); for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { - schedulers[i] = std::make_unique<Kernel::Scheduler>(system, i); - cores.emplace_back(system, i, *schedulers[i], interrupts[i]); + schedulers[i] = std::make_unique<Kernel::KScheduler>(system, i); + cores.emplace_back(i, system, *schedulers[i], interrupts); } } @@ -147,8 +152,8 @@ struct KernelCore::Impl { preemption_event = Core::Timing::CreateEvent( "PreemptionCallback", [this, &kernel](std::uintptr_t, std::chrono::nanoseconds) { { - SchedulerLock lock(kernel); - global_scheduler.PreemptThreads(); + KScopedSchedulerLock lock(kernel); + global_scheduler_context->PreemptThreads(); } const auto time_interval = std::chrono::nanoseconds{ Core::Timing::msToCycles(std::chrono::milliseconds(10))}; @@ -177,63 +182,62 @@ struct KernelCore::Impl { void MakeCurrentProcess(Process* process) { current_process = process; - if (process == nullptr) { return; } - u32 core_id = GetCurrentHostThreadID(); + const u32 core_id = GetCurrentHostThreadID(); if (core_id < Core::Hardware::NUM_CPU_CORES) { system.Memory().SetCurrentPageTable(*process, core_id); } } + /// Creates a new host thread ID, should only be called by GetHostThreadId + u32 AllocateHostThreadId(std::optional<std::size_t> core_id) { + if (core_id) { + // The first for slots are reserved for CPU core threads + ASSERT(*core_id < Core::Hardware::NUM_CPU_CORES); + return static_cast<u32>(*core_id); + } else { + return next_host_thread_id++; + } + } + + /// Gets the host thread ID for the caller, allocating a new one if this is the first time + u32 GetHostThreadId(std::optional<std::size_t> core_id = std::nullopt) { + const thread_local auto host_thread_id{AllocateHostThreadId(core_id)}; + return host_thread_id; + } + + /// Registers a CPU core thread by allocating a host thread ID for it void RegisterCoreThread(std::size_t core_id) { - std::unique_lock lock{register_thread_mutex}; + ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); + const auto this_id = GetHostThreadId(core_id); if (!is_multicore) { - single_core_thread_id = std::this_thread::get_id(); + single_core_thread_id = this_id; } - const std::thread::id this_id = std::this_thread::get_id(); - const auto it = host_thread_ids.find(this_id); - ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); - ASSERT(it == host_thread_ids.end()); - ASSERT(!registered_core_threads[core_id]); - host_thread_ids[this_id] = static_cast<u32>(core_id); - registered_core_threads.set(core_id); } + /// Registers a new host thread by allocating a host thread ID for it void RegisterHostThread() { - std::unique_lock lock{register_thread_mutex}; - const std::thread::id this_id = std::this_thread::get_id(); - const auto it = host_thread_ids.find(this_id); - if (it != host_thread_ids.end()) { - return; - } - host_thread_ids[this_id] = registered_thread_ids++; + [[maybe_unused]] const auto this_id = GetHostThreadId(); } - u32 GetCurrentHostThreadID() const { - const std::thread::id this_id = std::this_thread::get_id(); - if (!is_multicore) { - if (single_core_thread_id == this_id) { - return static_cast<u32>(system.GetCpuManager().CurrentCore()); - } - } - std::unique_lock lock{register_thread_mutex}; - const auto it = host_thread_ids.find(this_id); - if (it == host_thread_ids.end()) { - return Core::INVALID_HOST_THREAD_ID; + [[nodiscard]] u32 GetCurrentHostThreadID() { + const auto this_id = GetHostThreadId(); + if (!is_multicore && single_core_thread_id == this_id) { + return static_cast<u32>(system.GetCpuManager().CurrentCore()); } - return it->second; + return this_id; } - Core::EmuThreadHandle GetCurrentEmuThreadID() const { + [[nodiscard]] Core::EmuThreadHandle GetCurrentEmuThreadID() { Core::EmuThreadHandle result = Core::EmuThreadHandle::InvalidHandle(); result.host_handle = GetCurrentHostThreadID(); if (result.host_handle >= Core::Hardware::NUM_CPU_CORES) { return result; } - const Kernel::Scheduler& sched = cores[result.host_handle].Scheduler(); + const Kernel::KScheduler& sched = cores[result.host_handle].Scheduler(); const Kernel::Thread* current = sched.GetCurrentThread(); if (current != nullptr && !current->IsPhantomMode()) { result.guest_handle = current->GetGlobalHandle(); @@ -302,7 +306,7 @@ struct KernelCore::Impl { // Lists all processes that exist in the current session. std::vector<std::shared_ptr<Process>> process_list; Process* current_process = nullptr; - Kernel::GlobalScheduler global_scheduler; + std::unique_ptr<Kernel::GlobalSchedulerContext> global_scheduler_context; Kernel::Synchronization synchronization; Kernel::TimeManager time_manager; @@ -321,11 +325,8 @@ struct KernelCore::Impl { std::unique_ptr<Core::ExclusiveMonitor> exclusive_monitor; std::vector<Kernel::PhysicalCore> cores; - // 0-3 IDs represent core threads, >3 represent others - std::unordered_map<std::thread::id, u32> host_thread_ids; - u32 registered_thread_ids{Core::Hardware::NUM_CPU_CORES}; - std::bitset<Core::Hardware::NUM_CPU_CORES> registered_core_threads; - mutable std::mutex register_thread_mutex; + // Next host thead ID to use, 0-3 IDs represent core threads, >3 represent others + std::atomic<u32> next_host_thread_id{Core::Hardware::NUM_CPU_CORES}; // Kernel memory management std::unique_ptr<Memory::MemoryManager> memory_manager; @@ -337,12 +338,19 @@ struct KernelCore::Impl { std::shared_ptr<Kernel::SharedMemory> irs_shared_mem; std::shared_ptr<Kernel::SharedMemory> time_shared_mem; + // Threads used for services + std::unordered_set<std::shared_ptr<Kernel::ServiceThread>> service_threads; + + // Service threads are managed by a worker thread, so that a calling service thread can queue up + // the release of itself + std::unique_ptr<Common::ThreadWorker> service_thread_manager; + std::array<std::shared_ptr<Thread>, Core::Hardware::NUM_CPU_CORES> suspend_threads{}; std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES> interrupts{}; - std::array<std::unique_ptr<Kernel::Scheduler>, Core::Hardware::NUM_CPU_CORES> schedulers{}; + std::array<std::unique_ptr<Kernel::KScheduler>, Core::Hardware::NUM_CPU_CORES> schedulers{}; bool is_multicore{}; - std::thread::id single_core_thread_id{}; + u32 single_core_thread_id{}; std::array<u64, Core::Hardware::NUM_CPU_CORES> svc_ticks{}; @@ -363,6 +371,10 @@ void KernelCore::Initialize() { impl->Initialize(*this); } +void KernelCore::InitializeCores() { + impl->InitializeCores(); +} + void KernelCore::Shutdown() { impl->Shutdown(); } @@ -395,19 +407,19 @@ const std::vector<std::shared_ptr<Process>>& KernelCore::GetProcessList() const return impl->process_list; } -Kernel::GlobalScheduler& KernelCore::GlobalScheduler() { - return impl->global_scheduler; +Kernel::GlobalSchedulerContext& KernelCore::GlobalSchedulerContext() { + return *impl->global_scheduler_context; } -const Kernel::GlobalScheduler& KernelCore::GlobalScheduler() const { - return impl->global_scheduler; +const Kernel::GlobalSchedulerContext& KernelCore::GlobalSchedulerContext() const { + return *impl->global_scheduler_context; } -Kernel::Scheduler& KernelCore::Scheduler(std::size_t id) { +Kernel::KScheduler& KernelCore::Scheduler(std::size_t id) { return *impl->schedulers[id]; } -const Kernel::Scheduler& KernelCore::Scheduler(std::size_t id) const { +const Kernel::KScheduler& KernelCore::Scheduler(std::size_t id) const { return *impl->schedulers[id]; } @@ -431,16 +443,13 @@ const Kernel::PhysicalCore& KernelCore::CurrentPhysicalCore() const { return impl->cores[core_id]; } -Kernel::Scheduler& KernelCore::CurrentScheduler() { +Kernel::KScheduler* KernelCore::CurrentScheduler() { u32 core_id = impl->GetCurrentHostThreadID(); - ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); - return *impl->schedulers[core_id]; -} - -const Kernel::Scheduler& KernelCore::CurrentScheduler() const { - u32 core_id = impl->GetCurrentHostThreadID(); - ASSERT(core_id < Core::Hardware::NUM_CPU_CORES); - return *impl->schedulers[core_id]; + if (core_id >= Core::Hardware::NUM_CPU_CORES) { + // This is expected when called from not a guest thread + return {}; + } + return impl->schedulers[core_id].get(); } std::array<Core::CPUInterruptHandler, Core::Hardware::NUM_CPU_CORES>& KernelCore::Interrupts() { @@ -477,12 +486,17 @@ const Core::ExclusiveMonitor& KernelCore::GetExclusiveMonitor() const { } void KernelCore::InvalidateAllInstructionCaches() { - auto& threads = GlobalScheduler().GetThreadList(); - for (auto& thread : threads) { - if (!thread->IsHLEThread()) { - auto& arm_interface = thread->ArmInterface(); - arm_interface.ClearInstructionCache(); + for (auto& physical_core : impl->cores) { + physical_core.ArmInterface().ClearInstructionCache(); + } +} + +void KernelCore::InvalidateCpuInstructionCacheRange(VAddr addr, std::size_t size) { + for (auto& physical_core : impl->cores) { + if (!physical_core.IsInitialized()) { + continue; } + physical_core.ArmInterface().InvalidateCacheRange(addr, size); } } @@ -598,7 +612,7 @@ const Kernel::SharedMemory& KernelCore::GetTimeSharedMem() const { void KernelCore::Suspend(bool in_suspention) { const bool should_suspend = exception_exited || in_suspention; { - SchedulerLock lock(*this); + KScopedSchedulerLock lock(*this); ThreadStatus status = should_suspend ? ThreadStatus::Ready : ThreadStatus::WaitSleep; for (std::size_t i = 0; i < Core::Hardware::NUM_CPU_CORES; i++) { impl->suspend_threads[i]->SetStatus(status); @@ -625,4 +639,19 @@ void KernelCore::ExitSVCProfile() { MicroProfileLeave(MICROPROFILE_TOKEN(Kernel_SVC), impl->svc_ticks[core]); } +std::weak_ptr<Kernel::ServiceThread> KernelCore::CreateServiceThread(const std::string& name) { + auto service_thread = std::make_shared<Kernel::ServiceThread>(*this, 1, name); + impl->service_thread_manager->QueueWork( + [this, service_thread] { impl->service_threads.emplace(service_thread); }); + return service_thread; +} + +void KernelCore::ReleaseServiceThread(std::weak_ptr<Kernel::ServiceThread> service_thread) { + impl->service_thread_manager->QueueWork([this, service_thread] { + if (auto strong_ptr = service_thread.lock()) { + impl->service_threads.erase(strong_ptr); + } + }); +} + } // namespace Kernel diff --git a/src/core/hle/kernel/kernel.h b/src/core/hle/kernel/kernel.h index 16285c3f0..e3169f5a7 100644 --- a/src/core/hle/kernel/kernel.h +++ b/src/core/hle/kernel/kernel.h @@ -35,13 +35,14 @@ class SlabHeap; class AddressArbiter; class ClientPort; -class GlobalScheduler; +class GlobalSchedulerContext; class HandleTable; class PhysicalCore; class Process; class ResourceLimit; -class Scheduler; +class KScheduler; class SharedMemory; +class ServiceThread; class Synchronization; class Thread; class TimeManager; @@ -74,6 +75,9 @@ public: /// Resets the kernel to a clean slate for use. void Initialize(); + /// Initializes the CPU cores. + void InitializeCores(); + /// Clears all resources in use by the kernel instance. void Shutdown(); @@ -99,16 +103,16 @@ public: const std::vector<std::shared_ptr<Process>>& GetProcessList() const; /// Gets the sole instance of the global scheduler - Kernel::GlobalScheduler& GlobalScheduler(); + Kernel::GlobalSchedulerContext& GlobalSchedulerContext(); /// Gets the sole instance of the global scheduler - const Kernel::GlobalScheduler& GlobalScheduler() const; + const Kernel::GlobalSchedulerContext& GlobalSchedulerContext() const; /// Gets the sole instance of the Scheduler assoviated with cpu core 'id' - Kernel::Scheduler& Scheduler(std::size_t id); + Kernel::KScheduler& Scheduler(std::size_t id); /// Gets the sole instance of the Scheduler assoviated with cpu core 'id' - const Kernel::Scheduler& Scheduler(std::size_t id) const; + const Kernel::KScheduler& Scheduler(std::size_t id) const; /// Gets the an instance of the respective physical CPU core. Kernel::PhysicalCore& PhysicalCore(std::size_t id); @@ -117,10 +121,7 @@ public: const Kernel::PhysicalCore& PhysicalCore(std::size_t id) const; /// Gets the sole instance of the Scheduler at the current running core. - Kernel::Scheduler& CurrentScheduler(); - - /// Gets the sole instance of the Scheduler at the current running core. - const Kernel::Scheduler& CurrentScheduler() const; + Kernel::KScheduler* CurrentScheduler(); /// Gets the an instance of the current physical CPU core. Kernel::PhysicalCore& CurrentPhysicalCore(); @@ -153,6 +154,8 @@ public: void InvalidateAllInstructionCaches(); + void InvalidateCpuInstructionCacheRange(VAddr addr, std::size_t size); + /// Adds a port to the named port table void AddNamedPort(std::string name, std::shared_ptr<ClientPort> port); @@ -225,6 +228,22 @@ public: void ExitSVCProfile(); + /** + * Creates an HLE service thread, which are used to execute service routines asynchronously. + * While these are allocated per ServerSession, these need to be owned and managed outside of + * ServerSession to avoid a circular dependency. + * @param name String name for the ServerSession creating this thread, used for debug purposes. + * @returns The a weak pointer newly created service thread. + */ + std::weak_ptr<Kernel::ServiceThread> CreateServiceThread(const std::string& name); + + /** + * Releases a HLE service thread, instructing KernelCore to free it. This should be called when + * the ServerSession associated with the thread is destroyed. + * @param service_thread Service thread to release. + */ + void ReleaseServiceThread(std::weak_ptr<Kernel::ServiceThread> service_thread); + private: friend class Object; friend class Process; diff --git a/src/core/hle/kernel/memory/address_space_info.cpp b/src/core/hle/kernel/memory/address_space_info.cpp index e4288cab4..6cf43ba24 100644 --- a/src/core/hle/kernel/memory/address_space_info.cpp +++ b/src/core/hle/kernel/memory/address_space_info.cpp @@ -96,6 +96,7 @@ u64 AddressSpaceInfo::GetAddressSpaceStart(std::size_t width, Type type) { return AddressSpaceInfos[AddressSpaceIndices39Bit[index]].address; } UNREACHABLE(); + return 0; } std::size_t AddressSpaceInfo::GetAddressSpaceSize(std::size_t width, Type type) { @@ -112,6 +113,7 @@ std::size_t AddressSpaceInfo::GetAddressSpaceSize(std::size_t width, Type type) return AddressSpaceInfos[AddressSpaceIndices39Bit[index]].size; } UNREACHABLE(); + return 0; } } // namespace Kernel::Memory diff --git a/src/core/hle/kernel/memory/memory_block.h b/src/core/hle/kernel/memory/memory_block.h index 9d7839d08..83acece1e 100644 --- a/src/core/hle/kernel/memory/memory_block.h +++ b/src/core/hle/kernel/memory/memory_block.h @@ -73,12 +73,12 @@ enum class MemoryState : u32 { ThreadLocal = static_cast<u32>(Svc::MemoryState::ThreadLocal) | FlagMapped | FlagReferenceCounted, - Transfered = static_cast<u32>(Svc::MemoryState::Transfered) | FlagsMisc | - FlagCanAlignedDeviceMap | FlagCanChangeAttribute | FlagCanUseIpc | - FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + Transferred = static_cast<u32>(Svc::MemoryState::Transferred) | FlagsMisc | + FlagCanAlignedDeviceMap | FlagCanChangeAttribute | FlagCanUseIpc | + FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, - SharedTransfered = static_cast<u32>(Svc::MemoryState::SharedTransfered) | FlagsMisc | - FlagCanAlignedDeviceMap | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, + SharedTransferred = static_cast<u32>(Svc::MemoryState::SharedTransferred) | FlagsMisc | + FlagCanAlignedDeviceMap | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, SharedCode = static_cast<u32>(Svc::MemoryState::SharedCode) | FlagMapped | FlagReferenceCounted | FlagCanUseNonSecureIpc | FlagCanUseNonDeviceIpc, @@ -111,8 +111,8 @@ static_assert(static_cast<u32>(MemoryState::AliasCodeData) == 0x03FFBD09); static_assert(static_cast<u32>(MemoryState::Ipc) == 0x005C3C0A); static_assert(static_cast<u32>(MemoryState::Stack) == 0x005C3C0B); static_assert(static_cast<u32>(MemoryState::ThreadLocal) == 0x0040200C); -static_assert(static_cast<u32>(MemoryState::Transfered) == 0x015C3C0D); -static_assert(static_cast<u32>(MemoryState::SharedTransfered) == 0x005C380E); +static_assert(static_cast<u32>(MemoryState::Transferred) == 0x015C3C0D); +static_assert(static_cast<u32>(MemoryState::SharedTransferred) == 0x005C380E); static_assert(static_cast<u32>(MemoryState::SharedCode) == 0x0040380F); static_assert(static_cast<u32>(MemoryState::Inaccessible) == 0x00000010); static_assert(static_cast<u32>(MemoryState::NonSecureIpc) == 0x005C3811); @@ -222,9 +222,9 @@ public: public: constexpr MemoryBlock() = default; - constexpr MemoryBlock(VAddr addr, std::size_t num_pages, MemoryState state, - MemoryPermission perm, MemoryAttribute attribute) - : addr{addr}, num_pages(num_pages), state{state}, perm{perm}, attribute{attribute} {} + constexpr MemoryBlock(VAddr addr_, std::size_t num_pages_, MemoryState state_, + MemoryPermission perm_, MemoryAttribute attribute_) + : addr{addr_}, num_pages(num_pages_), state{state_}, perm{perm_}, attribute{attribute_} {} constexpr VAddr GetAddress() const { return addr; diff --git a/src/core/hle/kernel/memory/memory_block_manager.h b/src/core/hle/kernel/memory/memory_block_manager.h index 6e1d41075..f57d1bbcc 100644 --- a/src/core/hle/kernel/memory/memory_block_manager.h +++ b/src/core/hle/kernel/memory/memory_block_manager.h @@ -57,8 +57,8 @@ public: private: void MergeAdjacent(iterator it, iterator& next_it); - const VAddr start_addr; - const VAddr end_addr; + [[maybe_unused]] const VAddr start_addr; + [[maybe_unused]] const VAddr end_addr; MemoryBlockTree memory_block_tree; }; diff --git a/src/core/hle/kernel/memory/page_table.cpp b/src/core/hle/kernel/memory/page_table.cpp index a3fadb533..080886554 100644 --- a/src/core/hle/kernel/memory/page_table.cpp +++ b/src/core/hle/kernel/memory/page_table.cpp @@ -265,7 +265,7 @@ ResultCode PageTable::InitializeForProcess(FileSys::ProgramAddressSpaceType as_t physical_memory_usage = 0; memory_pool = pool; - page_table_impl.Resize(address_space_width, PageBits, true); + page_table_impl.Resize(address_space_width, PageBits); return InitializeMemoryLayout(start, end); } @@ -670,6 +670,11 @@ ResultCode PageTable::SetCodeMemoryPermission(VAddr addr, std::size_t size, Memo return RESULT_SUCCESS; } + if ((prev_perm & MemoryPermission::Execute) != (perm & MemoryPermission::Execute)) { + // Memory execution state is changing, invalidate CPU cache range + system.InvalidateCpuInstructionCacheRange(addr, size); + } + const std::size_t num_pages{size / PageSize}; const OperationType operation{(perm & MemoryPermission::Execute) != MemoryPermission::None ? OperationType::ChangePermissionsAndRefresh @@ -1002,8 +1007,8 @@ constexpr VAddr PageTable::GetRegionAddress(MemoryState state) const { case MemoryState::Shared: case MemoryState::AliasCode: case MemoryState::AliasCodeData: - case MemoryState::Transfered: - case MemoryState::SharedTransfered: + case MemoryState::Transferred: + case MemoryState::SharedTransferred: case MemoryState::SharedCode: case MemoryState::GeneratedCode: case MemoryState::CodeOut: @@ -1037,8 +1042,8 @@ constexpr std::size_t PageTable::GetRegionSize(MemoryState state) const { case MemoryState::Shared: case MemoryState::AliasCode: case MemoryState::AliasCodeData: - case MemoryState::Transfered: - case MemoryState::SharedTransfered: + case MemoryState::Transferred: + case MemoryState::SharedTransferred: case MemoryState::SharedCode: case MemoryState::GeneratedCode: case MemoryState::CodeOut: @@ -1075,8 +1080,8 @@ constexpr bool PageTable::CanContain(VAddr addr, std::size_t size, MemoryState s case MemoryState::AliasCodeData: case MemoryState::Stack: case MemoryState::ThreadLocal: - case MemoryState::Transfered: - case MemoryState::SharedTransfered: + case MemoryState::Transferred: + case MemoryState::SharedTransferred: case MemoryState::SharedCode: case MemoryState::GeneratedCode: case MemoryState::CodeOut: diff --git a/src/core/hle/kernel/mutex.cpp b/src/core/hle/kernel/mutex.cpp index 8f6c944d1..4f8075e0e 100644 --- a/src/core/hle/kernel/mutex.cpp +++ b/src/core/hle/kernel/mutex.cpp @@ -11,11 +11,11 @@ #include "core/core.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/mutex.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "core/hle/result.h" #include "core/memory.h" @@ -73,9 +73,9 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle, auto& kernel = system.Kernel(); std::shared_ptr<Thread> current_thread = - SharedFrom(kernel.CurrentScheduler().GetCurrentThread()); + SharedFrom(kernel.CurrentScheduler()->GetCurrentThread()); { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); // The mutex address must be 4-byte aligned if ((address % sizeof(u32)) != 0) { return ERR_INVALID_ADDRESS; @@ -114,7 +114,7 @@ ResultCode Mutex::TryAcquire(VAddr address, Handle holding_thread_handle, } { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); auto* owner = current_thread->GetLockOwner(); if (owner != nullptr) { owner->RemoveMutexWaiter(current_thread); @@ -153,10 +153,10 @@ std::pair<ResultCode, std::shared_ptr<Thread>> Mutex::Unlock(std::shared_ptr<Thr ResultCode Mutex::Release(VAddr address) { auto& kernel = system.Kernel(); - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); std::shared_ptr<Thread> current_thread = - SharedFrom(kernel.CurrentScheduler().GetCurrentThread()); + SharedFrom(kernel.CurrentScheduler()->GetCurrentThread()); auto [result, new_owner] = Unlock(current_thread, address); diff --git a/src/core/hle/kernel/physical_core.cpp b/src/core/hle/kernel/physical_core.cpp index c6bbdb080..7fea45f96 100644 --- a/src/core/hle/kernel/physical_core.cpp +++ b/src/core/hle/kernel/physical_core.cpp @@ -2,54 +2,60 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "common/assert.h" -#include "common/logging/log.h" #include "common/spin_lock.h" -#include "core/arm/arm_interface.h" -#ifdef ARCHITECTURE_x86_64 +#include "core/arm/cpu_interrupt_handler.h" #include "core/arm/dynarmic/arm_dynarmic_32.h" #include "core/arm/dynarmic/arm_dynarmic_64.h" -#endif -#include "core/arm/cpu_interrupt_handler.h" -#include "core/arm/exclusive_monitor.h" -#include "core/arm/unicorn/arm_unicorn.h" #include "core/core.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/physical_core.h" -#include "core/hle/kernel/scheduler.h" -#include "core/hle/kernel/thread.h" namespace Kernel { -PhysicalCore::PhysicalCore(Core::System& system, std::size_t id, Kernel::Scheduler& scheduler, - Core::CPUInterruptHandler& interrupt_handler) - : interrupt_handler{interrupt_handler}, core_index{id}, scheduler{scheduler} { - - guard = std::make_unique<Common::SpinLock>(); -} +PhysicalCore::PhysicalCore(std::size_t core_index, Core::System& system, + Kernel::KScheduler& scheduler, Core::CPUInterrupts& interrupts) + : core_index{core_index}, system{system}, scheduler{scheduler}, + interrupts{interrupts}, guard{std::make_unique<Common::SpinLock>()} {} PhysicalCore::~PhysicalCore() = default; -void PhysicalCore::Idle() { - interrupt_handler.AwaitInterrupt(); +void PhysicalCore::Initialize([[maybe_unused]] bool is_64_bit) { +#ifdef ARCHITECTURE_x86_64 + auto& kernel = system.Kernel(); + if (is_64_bit) { + arm_interface = std::make_unique<Core::ARM_Dynarmic_64>( + system, interrupts, kernel.IsMulticore(), kernel.GetExclusiveMonitor(), core_index); + } else { + arm_interface = std::make_unique<Core::ARM_Dynarmic_32>( + system, interrupts, kernel.IsMulticore(), kernel.GetExclusiveMonitor(), core_index); + } +#else +#error Platform not supported yet. +#endif } -void PhysicalCore::Shutdown() { - scheduler.Shutdown(); +void PhysicalCore::Run() { + arm_interface->Run(); +} + +void PhysicalCore::Idle() { + interrupts[core_index].AwaitInterrupt(); } bool PhysicalCore::IsInterrupted() const { - return interrupt_handler.IsInterrupted(); + return interrupts[core_index].IsInterrupted(); } void PhysicalCore::Interrupt() { guard->lock(); - interrupt_handler.SetInterrupt(true); + interrupts[core_index].SetInterrupt(true); guard->unlock(); } void PhysicalCore::ClearInterrupt() { guard->lock(); - interrupt_handler.SetInterrupt(false); + interrupts[core_index].SetInterrupt(false); guard->unlock(); } diff --git a/src/core/hle/kernel/physical_core.h b/src/core/hle/kernel/physical_core.h index d7a7a951c..f2b0911aa 100644 --- a/src/core/hle/kernel/physical_core.h +++ b/src/core/hle/kernel/physical_core.h @@ -4,19 +4,21 @@ #pragma once +#include <array> #include <cstddef> #include <memory> +#include "core/arm/arm_interface.h" + namespace Common { class SpinLock; } namespace Kernel { -class Scheduler; +class KScheduler; } // namespace Kernel namespace Core { -class ARM_Interface; class CPUInterruptHandler; class ExclusiveMonitor; class System; @@ -26,17 +28,24 @@ namespace Kernel { class PhysicalCore { public: - PhysicalCore(Core::System& system, std::size_t id, Kernel::Scheduler& scheduler, - Core::CPUInterruptHandler& interrupt_handler); + PhysicalCore(std::size_t core_index, Core::System& system, Kernel::KScheduler& scheduler, + Core::CPUInterrupts& interrupts); ~PhysicalCore(); PhysicalCore(const PhysicalCore&) = delete; PhysicalCore& operator=(const PhysicalCore&) = delete; PhysicalCore(PhysicalCore&&) = default; - PhysicalCore& operator=(PhysicalCore&&) = default; + PhysicalCore& operator=(PhysicalCore&&) = delete; + + /// Initialize the core for the specified parameters. + void Initialize(bool is_64_bit); + + /// Execute current jit state + void Run(); void Idle(); + /// Interrupt this physical core. void Interrupt(); @@ -46,8 +55,17 @@ public: /// Check if this core is interrupted bool IsInterrupted() const; - // Shutdown this physical core. - void Shutdown(); + bool IsInitialized() const { + return arm_interface != nullptr; + } + + Core::ARM_Interface& ArmInterface() { + return *arm_interface; + } + + const Core::ARM_Interface& ArmInterface() const { + return *arm_interface; + } bool IsMainCore() const { return core_index == 0; @@ -61,19 +79,21 @@ public: return core_index; } - Kernel::Scheduler& Scheduler() { + Kernel::KScheduler& Scheduler() { return scheduler; } - const Kernel::Scheduler& Scheduler() const { + const Kernel::KScheduler& Scheduler() const { return scheduler; } private: - Core::CPUInterruptHandler& interrupt_handler; - std::size_t core_index; - Kernel::Scheduler& scheduler; + const std::size_t core_index; + Core::System& system; + Kernel::KScheduler& scheduler; + Core::CPUInterrupts& interrupts; std::unique_ptr<Common::SpinLock> guard; + std::unique_ptr<Core::ARM_Interface> arm_interface; }; } // namespace Kernel diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp index ff9d9248b..b905b486a 100644 --- a/src/core/hle/kernel/process.cpp +++ b/src/core/hle/kernel/process.cpp @@ -4,6 +4,7 @@ #include <algorithm> #include <bitset> +#include <ctime> #include <memory> #include <random> #include "common/alignment.h" @@ -14,13 +15,13 @@ #include "core/file_sys/program_metadata.h" #include "core/hle/kernel/code_set.h" #include "core/hle/kernel/errors.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/memory/memory_block_manager.h" #include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/memory/slab_heap.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/resource_limit.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "core/hle/lock.h" #include "core/memory.h" @@ -53,7 +54,7 @@ void SetupMainThread(Core::System& system, Process& owner_process, u32 priority, auto& kernel = system.Kernel(); // Threads by default are dormant, wake up the main thread so it runs when the scheduler fires { - SchedulerLock lock{kernel}; + KScopedSchedulerLock lock{kernel}; thread->SetStatus(ThreadStatus::Ready); } } @@ -123,7 +124,7 @@ std::shared_ptr<Process> Process::Create(Core::System& system, std::string name, : kernel.CreateNewUserProcessID(); process->capabilities.InitializeForMetadatalessProcess(); - std::mt19937 rng(Settings::values.rng_seed.GetValue().value_or(0)); + std::mt19937 rng(Settings::values.rng_seed.GetValue().value_or(std::time(nullptr))); std::uniform_int_distribution<u64> distribution; std::generate(process->random_entropy.begin(), process->random_entropy.end(), [&] { return distribution(rng); }); @@ -212,7 +213,7 @@ void Process::UnregisterThread(const Thread* thread) { } ResultCode Process::ClearSignalState() { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(system.Kernel()); if (status == ProcessStatus::Exited) { LOG_ERROR(Kernel, "called on a terminated process instance."); return ERR_INVALID_STATE; @@ -313,7 +314,7 @@ void Process::PrepareForTermination() { if (thread->GetOwnerProcess() != this) continue; - if (thread.get() == system.CurrentScheduler().GetCurrentThread()) + if (thread.get() == kernel.CurrentScheduler()->GetCurrentThread()) continue; // TODO(Subv): When are the other running/ready threads terminated? @@ -324,7 +325,7 @@ void Process::PrepareForTermination() { } }; - stop_threads(system.GlobalScheduler().GetThreadList()); + stop_threads(system.GlobalSchedulerContext().GetThreadList()); FreeTLSRegion(tls_region_address); tls_region_address = 0; @@ -346,7 +347,7 @@ static auto FindTLSPageWithAvailableSlots(std::vector<TLSPage>& tls_pages) { } VAddr Process::CreateTLSRegion() { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(system.Kernel()); if (auto tls_page_iter{FindTLSPageWithAvailableSlots(tls_pages)}; tls_page_iter != tls_pages.cend()) { return *tls_page_iter->ReserveSlot(); @@ -377,7 +378,7 @@ VAddr Process::CreateTLSRegion() { } void Process::FreeTLSRegion(VAddr tls_address) { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(system.Kernel()); const VAddr aligned_address = Common::AlignDown(tls_address, Core::Memory::PAGE_SIZE); auto iter = std::find_if(tls_pages.begin(), tls_pages.end(), [aligned_address](const auto& page) { diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h index f45cb5674..e412e58aa 100644 --- a/src/core/hle/kernel/process.h +++ b/src/core/hle/kernel/process.h @@ -216,6 +216,16 @@ public: total_process_running_time_ticks += ticks; } + /// Gets the process schedule count, used for thread yelding + s64 GetScheduledCount() const { + return schedule_count; + } + + /// Increments the process schedule count, used for thread yielding. + void IncrementScheduledCount() { + ++schedule_count; + } + /// Gets 8 bytes of random data for svcGetInfo RandomEntropy u64 GetRandomEntropy(std::size_t index) const { return random_entropy.at(index); @@ -397,6 +407,9 @@ private: /// Name of this process std::string name; + /// Schedule count of this process + s64 schedule_count{}; + /// System context Core::System& system; }; diff --git a/src/core/hle/kernel/process_capability.cpp b/src/core/hle/kernel/process_capability.cpp index 63880f13d..0f128c586 100644 --- a/src/core/hle/kernel/process_capability.cpp +++ b/src/core/hle/kernel/process_capability.cpp @@ -199,7 +199,7 @@ ResultCode ProcessCapabilities::ParseSingleFlagCapability(u32& set_flags, u32& s break; } - LOG_ERROR(Kernel, "Invalid capability type! type={}", static_cast<u32>(type)); + LOG_ERROR(Kernel, "Invalid capability type! type={}", type); return ERR_INVALID_CAPABILITY_DESCRIPTOR; } diff --git a/src/core/hle/kernel/readable_event.cpp b/src/core/hle/kernel/readable_event.cpp index 6e286419e..cea262ce0 100644 --- a/src/core/hle/kernel/readable_event.cpp +++ b/src/core/hle/kernel/readable_event.cpp @@ -6,10 +6,10 @@ #include "common/assert.h" #include "common/logging/log.h" #include "core/hle/kernel/errors.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/readable_event.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" namespace Kernel { @@ -39,7 +39,7 @@ void ReadableEvent::Clear() { } ResultCode ReadableEvent::Reset() { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); if (!is_signaled) { LOG_TRACE(Kernel, "Handle is not signaled! object_id={}, object_type={}, object_name={}", GetObjectId(), GetTypeName(), GetName()); diff --git a/src/core/hle/kernel/resource_limit.cpp b/src/core/hle/kernel/resource_limit.cpp index 212e442f4..7bf50339d 100644 --- a/src/core/hle/kernel/resource_limit.cpp +++ b/src/core/hle/kernel/resource_limit.cpp @@ -65,8 +65,8 @@ ResultCode ResourceLimit::SetLimitValue(ResourceType resource, s64 value) { limit[index] = value; return RESULT_SUCCESS; } else { - LOG_ERROR(Kernel, "Limit value is too large! resource={}, value={}, index={}", - static_cast<u32>(resource), value, index); + LOG_ERROR(Kernel, "Limit value is too large! resource={}, value={}, index={}", resource, + value, index); return ERR_INVALID_STATE; } } diff --git a/src/core/hle/kernel/scheduler.cpp b/src/core/hle/kernel/scheduler.cpp deleted file mode 100644 index 5cbd3b912..000000000 --- a/src/core/hle/kernel/scheduler.cpp +++ /dev/null @@ -1,849 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. -// -// SelectThreads, Yield functions originally by TuxSH. -// licensed under GPLv2 or later under exception provided by the author. - -#include <algorithm> -#include <mutex> -#include <set> -#include <unordered_set> -#include <utility> - -#include "common/assert.h" -#include "common/bit_util.h" -#include "common/fiber.h" -#include "common/logging/log.h" -#include "core/arm/arm_interface.h" -#include "core/core.h" -#include "core/core_timing.h" -#include "core/cpu_manager.h" -#include "core/hle/kernel/kernel.h" -#include "core/hle/kernel/physical_core.h" -#include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" -#include "core/hle/kernel/time_manager.h" - -namespace Kernel { - -GlobalScheduler::GlobalScheduler(KernelCore& kernel) : kernel{kernel} {} - -GlobalScheduler::~GlobalScheduler() = default; - -void GlobalScheduler::AddThread(std::shared_ptr<Thread> thread) { - std::scoped_lock lock{global_list_guard}; - thread_list.push_back(std::move(thread)); -} - -void GlobalScheduler::RemoveThread(std::shared_ptr<Thread> thread) { - std::scoped_lock lock{global_list_guard}; - thread_list.erase(std::remove(thread_list.begin(), thread_list.end(), thread), - thread_list.end()); -} - -u32 GlobalScheduler::SelectThreads() { - ASSERT(is_locked); - const auto update_thread = [](Thread* thread, Scheduler& sched) { - std::scoped_lock lock{sched.guard}; - if (thread != sched.selected_thread_set.get()) { - if (thread == nullptr) { - ++sched.idle_selection_count; - } - sched.selected_thread_set = SharedFrom(thread); - } - const bool reschedule_pending = - sched.is_context_switch_pending || (sched.selected_thread_set != sched.current_thread); - sched.is_context_switch_pending = reschedule_pending; - std::atomic_thread_fence(std::memory_order_seq_cst); - return reschedule_pending; - }; - if (!is_reselection_pending.load()) { - return 0; - } - std::array<Thread*, Core::Hardware::NUM_CPU_CORES> top_threads{}; - - u32 idle_cores{}; - - // Step 1: Get top thread in schedule queue. - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - Thread* top_thread = - scheduled_queue[core].empty() ? nullptr : scheduled_queue[core].front(); - if (top_thread != nullptr) { - // TODO(Blinkhawk): Implement Thread Pinning - } else { - idle_cores |= (1ul << core); - } - top_threads[core] = top_thread; - } - - while (idle_cores != 0) { - u32 core_id = Common::CountTrailingZeroes32(idle_cores); - - if (!suggested_queue[core_id].empty()) { - std::array<s32, Core::Hardware::NUM_CPU_CORES> migration_candidates{}; - std::size_t num_candidates = 0; - auto iter = suggested_queue[core_id].begin(); - Thread* suggested = nullptr; - // Step 2: Try selecting a suggested thread. - while (iter != suggested_queue[core_id].end()) { - suggested = *iter; - iter++; - s32 suggested_core_id = suggested->GetProcessorID(); - Thread* top_thread = - suggested_core_id >= 0 ? top_threads[suggested_core_id] : nullptr; - if (top_thread != suggested) { - if (top_thread != nullptr && - top_thread->GetPriority() < THREADPRIO_MAX_CORE_MIGRATION) { - suggested = nullptr; - break; - // There's a too high thread to do core migration, cancel - } - TransferToCore(suggested->GetPriority(), static_cast<s32>(core_id), suggested); - break; - } - suggested = nullptr; - migration_candidates[num_candidates++] = suggested_core_id; - } - // Step 3: Select a suggested thread from another core - if (suggested == nullptr) { - for (std::size_t i = 0; i < num_candidates; i++) { - s32 candidate_core = migration_candidates[i]; - suggested = top_threads[candidate_core]; - auto it = scheduled_queue[candidate_core].begin(); - it++; - Thread* next = it != scheduled_queue[candidate_core].end() ? *it : nullptr; - if (next != nullptr) { - TransferToCore(suggested->GetPriority(), static_cast<s32>(core_id), - suggested); - top_threads[candidate_core] = next; - break; - } else { - suggested = nullptr; - } - } - } - top_threads[core_id] = suggested; - } - - idle_cores &= ~(1ul << core_id); - } - u32 cores_needing_context_switch{}; - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - Scheduler& sched = kernel.Scheduler(core); - ASSERT(top_threads[core] == nullptr || - static_cast<u32>(top_threads[core]->GetProcessorID()) == core); - if (update_thread(top_threads[core], sched)) { - cores_needing_context_switch |= (1ul << core); - } - } - return cores_needing_context_switch; -} - -bool GlobalScheduler::YieldThread(Thread* yielding_thread) { - ASSERT(is_locked); - // Note: caller should use critical section, etc. - if (!yielding_thread->IsRunnable()) { - // Normally this case shouldn't happen except for SetThreadActivity. - is_reselection_pending.store(true, std::memory_order_release); - return false; - } - const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID()); - const u32 priority = yielding_thread->GetPriority(); - - // Yield the thread - Reschedule(priority, core_id, yielding_thread); - const Thread* const winner = scheduled_queue[core_id].front(); - if (kernel.GetCurrentHostThreadID() != core_id) { - is_reselection_pending.store(true, std::memory_order_release); - } - - return AskForReselectionOrMarkRedundant(yielding_thread, winner); -} - -bool GlobalScheduler::YieldThreadAndBalanceLoad(Thread* yielding_thread) { - ASSERT(is_locked); - // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section, - // etc. - if (!yielding_thread->IsRunnable()) { - // Normally this case shouldn't happen except for SetThreadActivity. - is_reselection_pending.store(true, std::memory_order_release); - return false; - } - const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID()); - const u32 priority = yielding_thread->GetPriority(); - - // Yield the thread - Reschedule(priority, core_id, yielding_thread); - - std::array<Thread*, Core::Hardware::NUM_CPU_CORES> current_threads; - for (std::size_t i = 0; i < current_threads.size(); i++) { - current_threads[i] = scheduled_queue[i].empty() ? nullptr : scheduled_queue[i].front(); - } - - Thread* next_thread = scheduled_queue[core_id].front(priority); - Thread* winner = nullptr; - for (auto& thread : suggested_queue[core_id]) { - const s32 source_core = thread->GetProcessorID(); - if (source_core >= 0) { - if (current_threads[source_core] != nullptr) { - if (thread == current_threads[source_core] || - current_threads[source_core]->GetPriority() < min_regular_priority) { - continue; - } - } - } - if (next_thread->GetLastRunningTicks() >= thread->GetLastRunningTicks() || - next_thread->GetPriority() < thread->GetPriority()) { - if (thread->GetPriority() <= priority) { - winner = thread; - break; - } - } - } - - if (winner != nullptr) { - if (winner != yielding_thread) { - TransferToCore(winner->GetPriority(), s32(core_id), winner); - } - } else { - winner = next_thread; - } - - if (kernel.GetCurrentHostThreadID() != core_id) { - is_reselection_pending.store(true, std::memory_order_release); - } - - return AskForReselectionOrMarkRedundant(yielding_thread, winner); -} - -bool GlobalScheduler::YieldThreadAndWaitForLoadBalancing(Thread* yielding_thread) { - ASSERT(is_locked); - // Note: caller should check if !thread.IsSchedulerOperationRedundant and use critical section, - // etc. - if (!yielding_thread->IsRunnable()) { - // Normally this case shouldn't happen except for SetThreadActivity. - is_reselection_pending.store(true, std::memory_order_release); - return false; - } - Thread* winner = nullptr; - const u32 core_id = static_cast<u32>(yielding_thread->GetProcessorID()); - - // Remove the thread from its scheduled mlq, put it on the corresponding "suggested" one instead - TransferToCore(yielding_thread->GetPriority(), -1, yielding_thread); - - // If the core is idle, perform load balancing, excluding the threads that have just used this - // function... - if (scheduled_queue[core_id].empty()) { - // Here, "current_threads" is calculated after the ""yield"", unlike yield -1 - std::array<Thread*, Core::Hardware::NUM_CPU_CORES> current_threads; - for (std::size_t i = 0; i < current_threads.size(); i++) { - current_threads[i] = scheduled_queue[i].empty() ? nullptr : scheduled_queue[i].front(); - } - for (auto& thread : suggested_queue[core_id]) { - const s32 source_core = thread->GetProcessorID(); - if (source_core < 0 || thread == current_threads[source_core]) { - continue; - } - if (current_threads[source_core] == nullptr || - current_threads[source_core]->GetPriority() >= min_regular_priority) { - winner = thread; - } - break; - } - if (winner != nullptr) { - if (winner != yielding_thread) { - TransferToCore(winner->GetPriority(), static_cast<s32>(core_id), winner); - } - } else { - winner = yielding_thread; - } - } else { - winner = scheduled_queue[core_id].front(); - } - - if (kernel.GetCurrentHostThreadID() != core_id) { - is_reselection_pending.store(true, std::memory_order_release); - } - - return AskForReselectionOrMarkRedundant(yielding_thread, winner); -} - -void GlobalScheduler::PreemptThreads() { - ASSERT(is_locked); - for (std::size_t core_id = 0; core_id < Core::Hardware::NUM_CPU_CORES; core_id++) { - const u32 priority = preemption_priorities[core_id]; - - if (scheduled_queue[core_id].size(priority) > 0) { - if (scheduled_queue[core_id].size(priority) > 1) { - scheduled_queue[core_id].front(priority)->IncrementYieldCount(); - } - scheduled_queue[core_id].yield(priority); - if (scheduled_queue[core_id].size(priority) > 1) { - scheduled_queue[core_id].front(priority)->IncrementYieldCount(); - } - } - - Thread* current_thread = - scheduled_queue[core_id].empty() ? nullptr : scheduled_queue[core_id].front(); - Thread* winner = nullptr; - for (auto& thread : suggested_queue[core_id]) { - const s32 source_core = thread->GetProcessorID(); - if (thread->GetPriority() != priority) { - continue; - } - if (source_core >= 0) { - Thread* next_thread = scheduled_queue[source_core].empty() - ? nullptr - : scheduled_queue[source_core].front(); - if (next_thread != nullptr && next_thread->GetPriority() < 2) { - break; - } - if (next_thread == thread) { - continue; - } - } - if (current_thread != nullptr && - current_thread->GetLastRunningTicks() >= thread->GetLastRunningTicks()) { - winner = thread; - break; - } - } - - if (winner != nullptr) { - TransferToCore(winner->GetPriority(), s32(core_id), winner); - current_thread = - winner->GetPriority() <= current_thread->GetPriority() ? winner : current_thread; - } - - if (current_thread != nullptr && current_thread->GetPriority() > priority) { - for (auto& thread : suggested_queue[core_id]) { - const s32 source_core = thread->GetProcessorID(); - if (thread->GetPriority() < priority) { - continue; - } - if (source_core >= 0) { - Thread* next_thread = scheduled_queue[source_core].empty() - ? nullptr - : scheduled_queue[source_core].front(); - if (next_thread != nullptr && next_thread->GetPriority() < 2) { - break; - } - if (next_thread == thread) { - continue; - } - } - if (current_thread != nullptr && - current_thread->GetLastRunningTicks() >= thread->GetLastRunningTicks()) { - winner = thread; - break; - } - } - - if (winner != nullptr) { - TransferToCore(winner->GetPriority(), s32(core_id), winner); - current_thread = winner; - } - } - - is_reselection_pending.store(true, std::memory_order_release); - } -} - -void GlobalScheduler::EnableInterruptAndSchedule(u32 cores_pending_reschedule, - Core::EmuThreadHandle global_thread) { - u32 current_core = global_thread.host_handle; - bool must_context_switch = global_thread.guest_handle != InvalidHandle && - (current_core < Core::Hardware::NUM_CPU_CORES); - while (cores_pending_reschedule != 0) { - u32 core = Common::CountTrailingZeroes32(cores_pending_reschedule); - ASSERT(core < Core::Hardware::NUM_CPU_CORES); - if (!must_context_switch || core != current_core) { - auto& phys_core = kernel.PhysicalCore(core); - phys_core.Interrupt(); - } else { - must_context_switch = true; - } - cores_pending_reschedule &= ~(1ul << core); - } - if (must_context_switch) { - auto& core_scheduler = kernel.CurrentScheduler(); - kernel.ExitSVCProfile(); - core_scheduler.TryDoContextSwitch(); - kernel.EnterSVCProfile(); - } -} - -void GlobalScheduler::Suggest(u32 priority, std::size_t core, Thread* thread) { - ASSERT(is_locked); - suggested_queue[core].add(thread, priority); -} - -void GlobalScheduler::Unsuggest(u32 priority, std::size_t core, Thread* thread) { - ASSERT(is_locked); - suggested_queue[core].remove(thread, priority); -} - -void GlobalScheduler::Schedule(u32 priority, std::size_t core, Thread* thread) { - ASSERT(is_locked); - ASSERT_MSG(thread->GetProcessorID() == s32(core), "Thread must be assigned to this core."); - scheduled_queue[core].add(thread, priority); -} - -void GlobalScheduler::SchedulePrepend(u32 priority, std::size_t core, Thread* thread) { - ASSERT(is_locked); - ASSERT_MSG(thread->GetProcessorID() == s32(core), "Thread must be assigned to this core."); - scheduled_queue[core].add(thread, priority, false); -} - -void GlobalScheduler::Reschedule(u32 priority, std::size_t core, Thread* thread) { - ASSERT(is_locked); - scheduled_queue[core].remove(thread, priority); - scheduled_queue[core].add(thread, priority); -} - -void GlobalScheduler::Unschedule(u32 priority, std::size_t core, Thread* thread) { - ASSERT(is_locked); - scheduled_queue[core].remove(thread, priority); -} - -void GlobalScheduler::TransferToCore(u32 priority, s32 destination_core, Thread* thread) { - ASSERT(is_locked); - const bool schedulable = thread->GetPriority() < THREADPRIO_COUNT; - const s32 source_core = thread->GetProcessorID(); - if (source_core == destination_core || !schedulable) { - return; - } - thread->SetProcessorID(destination_core); - if (source_core >= 0) { - Unschedule(priority, static_cast<u32>(source_core), thread); - } - if (destination_core >= 0) { - Unsuggest(priority, static_cast<u32>(destination_core), thread); - Schedule(priority, static_cast<u32>(destination_core), thread); - } - if (source_core >= 0) { - Suggest(priority, static_cast<u32>(source_core), thread); - } -} - -bool GlobalScheduler::AskForReselectionOrMarkRedundant(Thread* current_thread, - const Thread* winner) { - if (current_thread == winner) { - current_thread->IncrementYieldCount(); - return true; - } else { - is_reselection_pending.store(true, std::memory_order_release); - return false; - } -} - -void GlobalScheduler::AdjustSchedulingOnStatus(Thread* thread, u32 old_flags) { - if (old_flags == thread->scheduling_state) { - return; - } - ASSERT(is_locked); - - if (old_flags == static_cast<u32>(ThreadSchedStatus::Runnable)) { - // In this case the thread was running, now it's pausing/exitting - if (thread->processor_id >= 0) { - Unschedule(thread->current_priority, static_cast<u32>(thread->processor_id), thread); - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(thread->processor_id) && - ((thread->affinity_mask >> core) & 1) != 0) { - Unsuggest(thread->current_priority, core, thread); - } - } - } else if (thread->scheduling_state == static_cast<u32>(ThreadSchedStatus::Runnable)) { - // The thread is now set to running from being stopped - if (thread->processor_id >= 0) { - Schedule(thread->current_priority, static_cast<u32>(thread->processor_id), thread); - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(thread->processor_id) && - ((thread->affinity_mask >> core) & 1) != 0) { - Suggest(thread->current_priority, core, thread); - } - } - } - - SetReselectionPending(); -} - -void GlobalScheduler::AdjustSchedulingOnPriority(Thread* thread, u32 old_priority) { - if (thread->scheduling_state != static_cast<u32>(ThreadSchedStatus::Runnable)) { - return; - } - ASSERT(is_locked); - if (thread->processor_id >= 0) { - Unschedule(old_priority, static_cast<u32>(thread->processor_id), thread); - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(thread->processor_id) && - ((thread->affinity_mask >> core) & 1) != 0) { - Unsuggest(old_priority, core, thread); - } - } - - if (thread->processor_id >= 0) { - if (thread == kernel.CurrentScheduler().GetCurrentThread()) { - SchedulePrepend(thread->current_priority, static_cast<u32>(thread->processor_id), - thread); - } else { - Schedule(thread->current_priority, static_cast<u32>(thread->processor_id), thread); - } - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (core != static_cast<u32>(thread->processor_id) && - ((thread->affinity_mask >> core) & 1) != 0) { - Suggest(thread->current_priority, core, thread); - } - } - thread->IncrementYieldCount(); - SetReselectionPending(); -} - -void GlobalScheduler::AdjustSchedulingOnAffinity(Thread* thread, u64 old_affinity_mask, - s32 old_core) { - if (thread->scheduling_state != static_cast<u32>(ThreadSchedStatus::Runnable) || - thread->current_priority >= THREADPRIO_COUNT) { - return; - } - ASSERT(is_locked); - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (((old_affinity_mask >> core) & 1) != 0) { - if (core == static_cast<u32>(old_core)) { - Unschedule(thread->current_priority, core, thread); - } else { - Unsuggest(thread->current_priority, core, thread); - } - } - } - - for (u32 core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - if (((thread->affinity_mask >> core) & 1) != 0) { - if (core == static_cast<u32>(thread->processor_id)) { - Schedule(thread->current_priority, core, thread); - } else { - Suggest(thread->current_priority, core, thread); - } - } - } - - thread->IncrementYieldCount(); - SetReselectionPending(); -} - -void GlobalScheduler::Shutdown() { - for (std::size_t core = 0; core < Core::Hardware::NUM_CPU_CORES; core++) { - scheduled_queue[core].clear(); - suggested_queue[core].clear(); - } - thread_list.clear(); -} - -void GlobalScheduler::Lock() { - Core::EmuThreadHandle current_thread = kernel.GetCurrentEmuThreadID(); - ASSERT(!current_thread.IsInvalid()); - if (current_thread == current_owner) { - ++scope_lock; - } else { - inner_lock.lock(); - is_locked = true; - current_owner = current_thread; - ASSERT(current_owner != Core::EmuThreadHandle::InvalidHandle()); - scope_lock = 1; - } -} - -void GlobalScheduler::Unlock() { - if (--scope_lock != 0) { - ASSERT(scope_lock > 0); - return; - } - u32 cores_pending_reschedule = SelectThreads(); - Core::EmuThreadHandle leaving_thread = current_owner; - current_owner = Core::EmuThreadHandle::InvalidHandle(); - scope_lock = 1; - is_locked = false; - inner_lock.unlock(); - EnableInterruptAndSchedule(cores_pending_reschedule, leaving_thread); -} - -Scheduler::Scheduler(Core::System& system, std::size_t core_id) : system(system), core_id(core_id) { - switch_fiber = std::make_shared<Common::Fiber>(std::function<void(void*)>(OnSwitch), this); -} - -Scheduler::~Scheduler() = default; - -bool Scheduler::HaveReadyThreads() const { - return system.GlobalScheduler().HaveReadyThreads(core_id); -} - -Thread* Scheduler::GetCurrentThread() const { - if (current_thread) { - return current_thread.get(); - } - return idle_thread.get(); -} - -Thread* Scheduler::GetSelectedThread() const { - return selected_thread.get(); -} - -u64 Scheduler::GetLastContextSwitchTicks() const { - return last_context_switch_time; -} - -void Scheduler::TryDoContextSwitch() { - auto& phys_core = system.Kernel().CurrentPhysicalCore(); - if (phys_core.IsInterrupted()) { - phys_core.ClearInterrupt(); - } - guard.lock(); - if (is_context_switch_pending) { - SwitchContext(); - } else { - guard.unlock(); - } -} - -void Scheduler::OnThreadStart() { - SwitchContextStep2(); -} - -void Scheduler::Unload() { - Thread* thread = current_thread.get(); - if (thread) { - thread->SetContinuousOnSVC(false); - thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); - thread->SetIsRunning(false); - if (!thread->IsHLEThread() && !thread->HasExited()) { - Core::ARM_Interface& cpu_core = thread->ArmInterface(); - cpu_core.SaveContext(thread->GetContext32()); - cpu_core.SaveContext(thread->GetContext64()); - // Save the TPIDR_EL0 system register in case it was modified. - thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0()); - cpu_core.ClearExclusiveState(); - } - thread->context_guard.unlock(); - } -} - -void Scheduler::Reload() { - Thread* thread = current_thread.get(); - if (thread) { - ASSERT_MSG(thread->GetSchedulingStatus() == ThreadSchedStatus::Runnable, - "Thread must be runnable."); - - // Cancel any outstanding wakeup events for this thread - thread->SetIsRunning(true); - thread->SetWasRunning(false); - thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); - - auto* const thread_owner_process = thread->GetOwnerProcess(); - if (thread_owner_process != nullptr) { - system.Kernel().MakeCurrentProcess(thread_owner_process); - } - if (!thread->IsHLEThread()) { - Core::ARM_Interface& cpu_core = thread->ArmInterface(); - cpu_core.LoadContext(thread->GetContext32()); - cpu_core.LoadContext(thread->GetContext64()); - cpu_core.SetTlsAddress(thread->GetTLSAddress()); - cpu_core.SetTPIDR_EL0(thread->GetTPIDR_EL0()); - cpu_core.ChangeProcessorID(this->core_id); - cpu_core.ClearExclusiveState(); - } - } -} - -void Scheduler::SwitchContextStep2() { - // Load context of new thread - if (selected_thread) { - ASSERT_MSG(selected_thread->GetSchedulingStatus() == ThreadSchedStatus::Runnable, - "Thread must be runnable."); - - // Cancel any outstanding wakeup events for this thread - selected_thread->SetIsRunning(true); - selected_thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); - selected_thread->SetWasRunning(false); - - auto* const thread_owner_process = current_thread->GetOwnerProcess(); - if (thread_owner_process != nullptr) { - system.Kernel().MakeCurrentProcess(thread_owner_process); - } - if (!selected_thread->IsHLEThread()) { - Core::ARM_Interface& cpu_core = selected_thread->ArmInterface(); - cpu_core.LoadContext(selected_thread->GetContext32()); - cpu_core.LoadContext(selected_thread->GetContext64()); - cpu_core.SetTlsAddress(selected_thread->GetTLSAddress()); - cpu_core.SetTPIDR_EL0(selected_thread->GetTPIDR_EL0()); - cpu_core.ChangeProcessorID(this->core_id); - cpu_core.ClearExclusiveState(); - } - } - - TryDoContextSwitch(); -} - -void Scheduler::SwitchContext() { - current_thread_prev = current_thread; - selected_thread = selected_thread_set; - Thread* previous_thread = current_thread_prev.get(); - Thread* new_thread = selected_thread.get(); - current_thread = selected_thread; - - is_context_switch_pending = false; - - if (new_thread == previous_thread) { - guard.unlock(); - return; - } - - Process* const previous_process = system.Kernel().CurrentProcess(); - - UpdateLastContextSwitchTime(previous_thread, previous_process); - - // Save context for previous thread - if (previous_thread) { - if (new_thread != nullptr && new_thread->IsSuspendThread()) { - previous_thread->SetWasRunning(true); - } - previous_thread->SetContinuousOnSVC(false); - previous_thread->last_running_ticks = system.CoreTiming().GetCPUTicks(); - previous_thread->SetIsRunning(false); - if (!previous_thread->IsHLEThread() && !previous_thread->HasExited()) { - Core::ARM_Interface& cpu_core = previous_thread->ArmInterface(); - cpu_core.SaveContext(previous_thread->GetContext32()); - cpu_core.SaveContext(previous_thread->GetContext64()); - // Save the TPIDR_EL0 system register in case it was modified. - previous_thread->SetTPIDR_EL0(cpu_core.GetTPIDR_EL0()); - cpu_core.ClearExclusiveState(); - } - previous_thread->context_guard.unlock(); - } - - std::shared_ptr<Common::Fiber>* old_context; - if (previous_thread != nullptr) { - old_context = &previous_thread->GetHostContext(); - } else { - old_context = &idle_thread->GetHostContext(); - } - guard.unlock(); - - Common::Fiber::YieldTo(*old_context, switch_fiber); - /// When a thread wakes up, the scheduler may have changed to other in another core. - auto& next_scheduler = system.Kernel().CurrentScheduler(); - next_scheduler.SwitchContextStep2(); -} - -void Scheduler::OnSwitch(void* this_scheduler) { - Scheduler* sched = static_cast<Scheduler*>(this_scheduler); - sched->SwitchToCurrent(); -} - -void Scheduler::SwitchToCurrent() { - while (true) { - { - std::scoped_lock lock{guard}; - selected_thread = selected_thread_set; - current_thread = selected_thread; - is_context_switch_pending = false; - } - const auto is_switch_pending = [this] { - std::scoped_lock lock{guard}; - return is_context_switch_pending; - }; - do { - if (current_thread != nullptr && !current_thread->IsHLEThread()) { - current_thread->context_guard.lock(); - if (!current_thread->IsRunnable()) { - current_thread->context_guard.unlock(); - break; - } - if (current_thread->GetProcessorID() != core_id) { - current_thread->context_guard.unlock(); - break; - } - } - std::shared_ptr<Common::Fiber>* next_context; - if (current_thread != nullptr) { - next_context = ¤t_thread->GetHostContext(); - } else { - next_context = &idle_thread->GetHostContext(); - } - Common::Fiber::YieldTo(switch_fiber, *next_context); - } while (!is_switch_pending()); - } -} - -void Scheduler::UpdateLastContextSwitchTime(Thread* thread, Process* process) { - const u64 prev_switch_ticks = last_context_switch_time; - const u64 most_recent_switch_ticks = system.CoreTiming().GetCPUTicks(); - const u64 update_ticks = most_recent_switch_ticks - prev_switch_ticks; - - if (thread != nullptr) { - thread->UpdateCPUTimeTicks(update_ticks); - } - - if (process != nullptr) { - process->UpdateCPUTimeTicks(update_ticks); - } - - last_context_switch_time = most_recent_switch_ticks; -} - -void Scheduler::Initialize() { - std::string name = "Idle Thread Id:" + std::to_string(core_id); - std::function<void(void*)> init_func = Core::CpuManager::GetIdleThreadStartFunc(); - void* init_func_parameter = system.GetCpuManager().GetStartFuncParamater(); - ThreadType type = static_cast<ThreadType>(THREADTYPE_KERNEL | THREADTYPE_HLE | THREADTYPE_IDLE); - auto thread_res = Thread::Create(system, type, name, 0, 64, 0, static_cast<u32>(core_id), 0, - nullptr, std::move(init_func), init_func_parameter); - idle_thread = std::move(thread_res).Unwrap(); -} - -void Scheduler::Shutdown() { - current_thread = nullptr; - selected_thread = nullptr; -} - -SchedulerLock::SchedulerLock(KernelCore& kernel) : kernel{kernel} { - kernel.GlobalScheduler().Lock(); -} - -SchedulerLock::~SchedulerLock() { - kernel.GlobalScheduler().Unlock(); -} - -SchedulerLockAndSleep::SchedulerLockAndSleep(KernelCore& kernel, Handle& event_handle, - Thread* time_task, s64 nanoseconds) - : SchedulerLock{kernel}, event_handle{event_handle}, time_task{time_task}, nanoseconds{ - nanoseconds} { - event_handle = InvalidHandle; -} - -SchedulerLockAndSleep::~SchedulerLockAndSleep() { - if (sleep_cancelled) { - return; - } - auto& time_manager = kernel.TimeManager(); - time_manager.ScheduleTimeEvent(event_handle, time_task, nanoseconds); -} - -void SchedulerLockAndSleep::Release() { - if (sleep_cancelled) { - return; - } - auto& time_manager = kernel.TimeManager(); - time_manager.ScheduleTimeEvent(event_handle, time_task, nanoseconds); - sleep_cancelled = true; -} - -} // namespace Kernel diff --git a/src/core/hle/kernel/scheduler.h b/src/core/hle/kernel/scheduler.h deleted file mode 100644 index b6f04dcea..000000000 --- a/src/core/hle/kernel/scheduler.h +++ /dev/null @@ -1,318 +0,0 @@ -// Copyright 2018 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <atomic> -#include <memory> -#include <mutex> -#include <vector> - -#include "common/common_types.h" -#include "common/multi_level_queue.h" -#include "common/spin_lock.h" -#include "core/hardware_properties.h" -#include "core/hle/kernel/thread.h" - -namespace Common { -class Fiber; -} - -namespace Core { -class ARM_Interface; -class System; -} // namespace Core - -namespace Kernel { - -class KernelCore; -class Process; -class SchedulerLock; - -class GlobalScheduler final { -public: - explicit GlobalScheduler(KernelCore& kernel); - ~GlobalScheduler(); - - /// Adds a new thread to the scheduler - void AddThread(std::shared_ptr<Thread> thread); - - /// Removes a thread from the scheduler - void RemoveThread(std::shared_ptr<Thread> thread); - - /// Returns a list of all threads managed by the scheduler - const std::vector<std::shared_ptr<Thread>>& GetThreadList() const { - return thread_list; - } - - /// Notify the scheduler a thread's status has changed. - void AdjustSchedulingOnStatus(Thread* thread, u32 old_flags); - - /// Notify the scheduler a thread's priority has changed. - void AdjustSchedulingOnPriority(Thread* thread, u32 old_priority); - - /// Notify the scheduler a thread's core and/or affinity mask has changed. - void AdjustSchedulingOnAffinity(Thread* thread, u64 old_affinity_mask, s32 old_core); - - /** - * Takes care of selecting the new scheduled threads in three steps: - * - * 1. First a thread is selected from the top of the priority queue. If no thread - * is obtained then we move to step two, else we are done. - * - * 2. Second we try to get a suggested thread that's not assigned to any core or - * that is not the top thread in that core. - * - * 3. Third is no suggested thread is found, we do a second pass and pick a running - * thread in another core and swap it with its current thread. - * - * returns the cores needing scheduling. - */ - u32 SelectThreads(); - - bool HaveReadyThreads(std::size_t core_id) const { - return !scheduled_queue[core_id].empty(); - } - - /** - * Takes a thread and moves it to the back of the it's priority list. - * - * @note This operation can be redundant and no scheduling is changed if marked as so. - */ - bool YieldThread(Thread* thread); - - /** - * Takes a thread and moves it to the back of the it's priority list. - * Afterwards, tries to pick a suggested thread from the suggested queue that has worse time or - * a better priority than the next thread in the core. - * - * @note This operation can be redundant and no scheduling is changed if marked as so. - */ - bool YieldThreadAndBalanceLoad(Thread* thread); - - /** - * Takes a thread and moves it out of the scheduling queue. - * and into the suggested queue. If no thread can be scheduled afterwards in that core, - * a suggested thread is obtained instead. - * - * @note This operation can be redundant and no scheduling is changed if marked as so. - */ - bool YieldThreadAndWaitForLoadBalancing(Thread* thread); - - /** - * Rotates the scheduling queues of threads at a preemption priority and then does - * some core rebalancing. Preemption priorities can be found in the array - * 'preemption_priorities'. - * - * @note This operation happens every 10ms. - */ - void PreemptThreads(); - - u32 CpuCoresCount() const { - return Core::Hardware::NUM_CPU_CORES; - } - - void SetReselectionPending() { - is_reselection_pending.store(true, std::memory_order_release); - } - - bool IsReselectionPending() const { - return is_reselection_pending.load(std::memory_order_acquire); - } - - void Shutdown(); - -private: - friend class SchedulerLock; - - /// Lock the scheduler to the current thread. - void Lock(); - - /// Unlocks the scheduler, reselects threads, interrupts cores for rescheduling - /// and reschedules current core if needed. - void Unlock(); - - void EnableInterruptAndSchedule(u32 cores_pending_reschedule, - Core::EmuThreadHandle global_thread); - - /** - * Add a thread to the suggested queue of a cpu core. Suggested threads may be - * picked if no thread is scheduled to run on the core. - */ - void Suggest(u32 priority, std::size_t core, Thread* thread); - - /** - * Remove a thread to the suggested queue of a cpu core. Suggested threads may be - * picked if no thread is scheduled to run on the core. - */ - void Unsuggest(u32 priority, std::size_t core, Thread* thread); - - /** - * Add a thread to the scheduling queue of a cpu core. The thread is added at the - * back the queue in its priority level. - */ - void Schedule(u32 priority, std::size_t core, Thread* thread); - - /** - * Add a thread to the scheduling queue of a cpu core. The thread is added at the - * front the queue in its priority level. - */ - void SchedulePrepend(u32 priority, std::size_t core, Thread* thread); - - /// Reschedule an already scheduled thread based on a new priority - void Reschedule(u32 priority, std::size_t core, Thread* thread); - - /// Unschedules a thread. - void Unschedule(u32 priority, std::size_t core, Thread* thread); - - /** - * Transfers a thread into an specific core. If the destination_core is -1 - * it will be unscheduled from its source code and added into its suggested - * queue. - */ - void TransferToCore(u32 priority, s32 destination_core, Thread* thread); - - bool AskForReselectionOrMarkRedundant(Thread* current_thread, const Thread* winner); - - static constexpr u32 min_regular_priority = 2; - std::array<Common::MultiLevelQueue<Thread*, THREADPRIO_COUNT>, Core::Hardware::NUM_CPU_CORES> - scheduled_queue; - std::array<Common::MultiLevelQueue<Thread*, THREADPRIO_COUNT>, Core::Hardware::NUM_CPU_CORES> - suggested_queue; - std::atomic<bool> is_reselection_pending{false}; - - // The priority levels at which the global scheduler preempts threads every 10 ms. They are - // ordered from Core 0 to Core 3. - std::array<u32, Core::Hardware::NUM_CPU_CORES> preemption_priorities = {59, 59, 59, 62}; - - /// Scheduler lock mechanisms. - bool is_locked{}; - std::mutex inner_lock; - std::atomic<s64> scope_lock{}; - Core::EmuThreadHandle current_owner{Core::EmuThreadHandle::InvalidHandle()}; - - Common::SpinLock global_list_guard{}; - - /// Lists all thread ids that aren't deleted/etc. - std::vector<std::shared_ptr<Thread>> thread_list; - KernelCore& kernel; -}; - -class Scheduler final { -public: - explicit Scheduler(Core::System& system, std::size_t core_id); - ~Scheduler(); - - /// Returns whether there are any threads that are ready to run. - bool HaveReadyThreads() const; - - /// Reschedules to the next available thread (call after current thread is suspended) - void TryDoContextSwitch(); - - /// The next two are for SingleCore Only. - /// Unload current thread before preempting core. - void Unload(); - /// Reload current thread after core preemption. - void Reload(); - - /// Gets the current running thread - Thread* GetCurrentThread() const; - - /// Gets the currently selected thread from the top of the multilevel queue - Thread* GetSelectedThread() const; - - /// Gets the timestamp for the last context switch in ticks. - u64 GetLastContextSwitchTicks() const; - - bool ContextSwitchPending() const { - return is_context_switch_pending; - } - - void Initialize(); - - /// Shutdowns the scheduler. - void Shutdown(); - - void OnThreadStart(); - - std::shared_ptr<Common::Fiber>& ControlContext() { - return switch_fiber; - } - - const std::shared_ptr<Common::Fiber>& ControlContext() const { - return switch_fiber; - } - -private: - friend class GlobalScheduler; - - /// Switches the CPU's active thread context to that of the specified thread - void SwitchContext(); - - /// When a thread wakes up, it must run this through it's new scheduler - void SwitchContextStep2(); - - /** - * Called on every context switch to update the internal timestamp - * This also updates the running time ticks for the given thread and - * process using the following difference: - * - * ticks += most_recent_ticks - last_context_switch_ticks - * - * The internal tick timestamp for the scheduler is simply the - * most recent tick count retrieved. No special arithmetic is - * applied to it. - */ - void UpdateLastContextSwitchTime(Thread* thread, Process* process); - - static void OnSwitch(void* this_scheduler); - void SwitchToCurrent(); - - std::shared_ptr<Thread> current_thread = nullptr; - std::shared_ptr<Thread> selected_thread = nullptr; - std::shared_ptr<Thread> current_thread_prev = nullptr; - std::shared_ptr<Thread> selected_thread_set = nullptr; - std::shared_ptr<Thread> idle_thread = nullptr; - - std::shared_ptr<Common::Fiber> switch_fiber = nullptr; - - Core::System& system; - u64 last_context_switch_time = 0; - u64 idle_selection_count = 0; - const std::size_t core_id; - - Common::SpinLock guard{}; - - bool is_context_switch_pending = false; -}; - -class SchedulerLock { -public: - [[nodiscard]] explicit SchedulerLock(KernelCore& kernel); - ~SchedulerLock(); - -protected: - KernelCore& kernel; -}; - -class SchedulerLockAndSleep : public SchedulerLock { -public: - explicit SchedulerLockAndSleep(KernelCore& kernel, Handle& event_handle, Thread* time_task, - s64 nanoseconds); - ~SchedulerLockAndSleep(); - - void CancelSleep() { - sleep_cancelled = true; - } - - void Release(); - -private: - Handle& event_handle; - Thread* time_task; - s64 nanoseconds; - bool sleep_cancelled{}; -}; - -} // namespace Kernel diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp index 8c19f2534..b40fe3916 100644 --- a/src/core/hle/kernel/server_session.cpp +++ b/src/core/hle/kernel/server_session.cpp @@ -14,9 +14,9 @@ #include "core/hle/kernel/client_session.h" #include "core/hle/kernel/handle_table.h" #include "core/hle/kernel/hle_ipc.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/server_session.h" #include "core/hle/kernel/session.h" #include "core/hle/kernel/thread.h" @@ -25,19 +25,19 @@ namespace Kernel { ServerSession::ServerSession(KernelCore& kernel) : SynchronizationObject{kernel} {} -ServerSession::~ServerSession() = default; + +ServerSession::~ServerSession() { + kernel.ReleaseServiceThread(service_thread); +} ResultVal<std::shared_ptr<ServerSession>> ServerSession::Create(KernelCore& kernel, std::shared_ptr<Session> parent, std::string name) { std::shared_ptr<ServerSession> session{std::make_shared<ServerSession>(kernel)}; - session->request_event = - Core::Timing::CreateEvent(name, [session](std::uintptr_t, std::chrono::nanoseconds) { - session->CompleteSyncRequest(); - }); session->name = std::move(name); session->parent = std::move(parent); + session->service_thread = kernel.CreateServiceThread(session->name); return MakeResult(std::move(session)); } @@ -130,8 +130,7 @@ ResultCode ServerSession::HandleDomainSyncRequest(Kernel::HLERequestContext& con } } - LOG_CRITICAL(IPC, "Unknown domain command={}", - static_cast<int>(domain_message_header.command.Value())); + LOG_CRITICAL(IPC, "Unknown domain command={}", domain_message_header.command.Value()); ASSERT(false); return RESULT_SUCCESS; } @@ -143,16 +142,16 @@ ResultCode ServerSession::QueueSyncRequest(std::shared_ptr<Thread> thread, std::make_shared<HLERequestContext>(kernel, memory, SharedFrom(this), std::move(thread)); context->PopulateFromIncomingCommandBuffer(kernel.CurrentProcess()->GetHandleTable(), cmd_buf); - request_queue.Push(std::move(context)); + + if (auto strong_ptr = service_thread.lock()) { + strong_ptr->QueueSyncRequest(*this, std::move(context)); + return RESULT_SUCCESS; + } return RESULT_SUCCESS; } -ResultCode ServerSession::CompleteSyncRequest() { - ASSERT(!request_queue.Empty()); - - auto& context = *request_queue.Front(); - +ResultCode ServerSession::CompleteSyncRequest(HLERequestContext& context) { ResultCode result = RESULT_SUCCESS; // If the session has been converted to a domain, handle the domain request if (IsDomain() && context.HasDomainMessageHeader()) { @@ -171,25 +170,20 @@ ResultCode ServerSession::CompleteSyncRequest() { // Some service requests require the thread to block { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); if (!context.IsThreadWaiting()) { context.GetThread().ResumeFromWait(); context.GetThread().SetSynchronizationResults(nullptr, result); } } - request_queue.Pop(); - return result; } ResultCode ServerSession::HandleSyncRequest(std::shared_ptr<Thread> thread, Core::Memory::Memory& memory, Core::Timing::CoreTiming& core_timing) { - const ResultCode result = QueueSyncRequest(std::move(thread), memory); - const auto delay = std::chrono::nanoseconds{kernel.IsMulticore() ? 0 : 20000}; - core_timing.ScheduleEvent(delay, request_event, {}); - return result; + return QueueSyncRequest(std::move(thread), memory); } } // namespace Kernel diff --git a/src/core/hle/kernel/server_session.h b/src/core/hle/kernel/server_session.h index d23e9ec68..e8d1d99ea 100644 --- a/src/core/hle/kernel/server_session.h +++ b/src/core/hle/kernel/server_session.h @@ -10,6 +10,7 @@ #include <vector> #include "common/threadsafe_queue.h" +#include "core/hle/kernel/service_thread.h" #include "core/hle/kernel/synchronization_object.h" #include "core/hle/result.h" @@ -43,6 +44,8 @@ class Thread; * TLS buffer and control is transferred back to it. */ class ServerSession final : public SynchronizationObject { + friend class ServiceThread; + public: explicit ServerSession(KernelCore& kernel); ~ServerSession() override; @@ -132,7 +135,7 @@ private: ResultCode QueueSyncRequest(std::shared_ptr<Thread> thread, Core::Memory::Memory& memory); /// Completes a sync request from the emulated application. - ResultCode CompleteSyncRequest(); + ResultCode CompleteSyncRequest(HLERequestContext& context); /// Handles a SyncRequest to a domain, forwarding the request to the proper object or closing an /// object handle. @@ -163,11 +166,8 @@ private: /// The name of this session (optional) std::string name; - /// Core timing event used to schedule the service request at some point in the future - std::shared_ptr<Core::Timing::EventType> request_event; - - /// Queue of scheduled service requests - Common::MPSCQueue<std::shared_ptr<Kernel::HLERequestContext>> request_queue; + /// Thread to dispatch service requests + std::weak_ptr<ServiceThread> service_thread; }; } // namespace Kernel diff --git a/src/core/hle/kernel/service_thread.cpp b/src/core/hle/kernel/service_thread.cpp new file mode 100644 index 000000000..ee46f3e21 --- /dev/null +++ b/src/core/hle/kernel/service_thread.cpp @@ -0,0 +1,110 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <condition_variable> +#include <functional> +#include <mutex> +#include <thread> +#include <vector> +#include <queue> + +#include "common/assert.h" +#include "common/scope_exit.h" +#include "common/thread.h" +#include "core/core.h" +#include "core/hle/kernel/kernel.h" +#include "core/hle/kernel/server_session.h" +#include "core/hle/kernel/service_thread.h" +#include "core/hle/lock.h" +#include "video_core/renderer_base.h" + +namespace Kernel { + +class ServiceThread::Impl final { +public: + explicit Impl(KernelCore& kernel, std::size_t num_threads, const std::string& name); + ~Impl(); + + void QueueSyncRequest(ServerSession& session, std::shared_ptr<HLERequestContext>&& context); + +private: + std::vector<std::thread> threads; + std::queue<std::function<void()>> requests; + std::mutex queue_mutex; + std::condition_variable condition; + const std::string service_name; + bool stop{}; +}; + +ServiceThread::Impl::Impl(KernelCore& kernel, std::size_t num_threads, const std::string& name) + : service_name{name} { + for (std::size_t i = 0; i < num_threads; ++i) + threads.emplace_back([this, &kernel] { + Common::SetCurrentThreadName(std::string{"yuzu:HleService:" + service_name}.c_str()); + + // Wait for first request before trying to acquire a render context + { + std::unique_lock lock{queue_mutex}; + condition.wait(lock, [this] { return stop || !requests.empty(); }); + } + + kernel.RegisterHostThread(); + + while (true) { + std::function<void()> task; + + { + std::unique_lock lock{queue_mutex}; + condition.wait(lock, [this] { return stop || !requests.empty(); }); + if (stop || requests.empty()) { + return; + } + task = std::move(requests.front()); + requests.pop(); + } + + task(); + } + }); +} + +void ServiceThread::Impl::QueueSyncRequest(ServerSession& session, + std::shared_ptr<HLERequestContext>&& context) { + { + std::unique_lock lock{queue_mutex}; + + // ServerSession owns the service thread, so we cannot caption a strong pointer here in the + // event that the ServerSession is terminated. + std::weak_ptr<ServerSession> weak_ptr{SharedFrom(&session)}; + requests.emplace([weak_ptr, context{std::move(context)}]() { + if (auto strong_ptr = weak_ptr.lock()) { + strong_ptr->CompleteSyncRequest(*context); + } + }); + } + condition.notify_one(); +} + +ServiceThread::Impl::~Impl() { + { + std::unique_lock lock{queue_mutex}; + stop = true; + } + condition.notify_all(); + for (std::thread& thread : threads) { + thread.join(); + } +} + +ServiceThread::ServiceThread(KernelCore& kernel, std::size_t num_threads, const std::string& name) + : impl{std::make_unique<Impl>(kernel, num_threads, name)} {} + +ServiceThread::~ServiceThread() = default; + +void ServiceThread::QueueSyncRequest(ServerSession& session, + std::shared_ptr<HLERequestContext>&& context) { + impl->QueueSyncRequest(session, std::move(context)); +} + +} // namespace Kernel diff --git a/src/core/hle/kernel/service_thread.h b/src/core/hle/kernel/service_thread.h new file mode 100644 index 000000000..025ab8fb5 --- /dev/null +++ b/src/core/hle/kernel/service_thread.h @@ -0,0 +1,28 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <string> + +namespace Kernel { + +class HLERequestContext; +class KernelCore; +class ServerSession; + +class ServiceThread final { +public: + explicit ServiceThread(KernelCore& kernel, std::size_t num_threads, const std::string& name); + ~ServiceThread(); + + void QueueSyncRequest(ServerSession& session, std::shared_ptr<HLERequestContext>&& context); + +private: + class Impl; + std::unique_ptr<Impl> impl; +}; + +} // namespace Kernel diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index bafd1ced7..de3ed25da 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -24,6 +24,8 @@ #include "core/hle/kernel/client_session.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/memory/memory_block.h" #include "core/hle/kernel/memory/page_table.h" @@ -32,7 +34,6 @@ #include "core/hle/kernel/process.h" #include "core/hle/kernel/readable_event.h" #include "core/hle/kernel/resource_limit.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/shared_memory.h" #include "core/hle/kernel/svc.h" #include "core/hle/kernel/svc_types.h" @@ -234,8 +235,7 @@ static ResultCode SetMemoryAttribute(Core::System& system, VAddr address, u64 si static ResultCode SetMemoryAttribute32(Core::System& system, u32 address, u32 size, u32 mask, u32 attribute) { - return SetMemoryAttribute(system, static_cast<VAddr>(address), static_cast<std::size_t>(size), - mask, attribute); + return SetMemoryAttribute(system, address, size, mask, attribute); } /// Maps a memory range into a different range. @@ -255,8 +255,7 @@ static ResultCode MapMemory(Core::System& system, VAddr dst_addr, VAddr src_addr } static ResultCode MapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) { - return MapMemory(system, static_cast<VAddr>(dst_addr), static_cast<VAddr>(src_addr), - static_cast<std::size_t>(size)); + return MapMemory(system, dst_addr, src_addr, size); } /// Unmaps a region that was previously mapped with svcMapMemory @@ -276,8 +275,7 @@ static ResultCode UnmapMemory(Core::System& system, VAddr dst_addr, VAddr src_ad } static ResultCode UnmapMemory32(Core::System& system, u32 dst_addr, u32 src_addr, u32 size) { - return UnmapMemory(system, static_cast<VAddr>(dst_addr), static_cast<VAddr>(src_addr), - static_cast<std::size_t>(size)); + return UnmapMemory(system, dst_addr, src_addr, size); } /// Connect to an OS service given the port name, returns the handle to the port to out @@ -332,7 +330,8 @@ static ResultCode ConnectToNamedPort32(Core::System& system, Handle* out_handle, /// Makes a blocking IPC call to an OS service. static ResultCode SendSyncRequest(Core::System& system, Handle handle) { - const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); + auto& kernel = system.Kernel(); + const auto& handle_table = kernel.CurrentProcess()->GetHandleTable(); std::shared_ptr<ClientSession> session = handle_table.Get<ClientSession>(handle); if (!session) { LOG_ERROR(Kernel_SVC, "called with invalid handle=0x{:08X}", handle); @@ -341,9 +340,9 @@ static ResultCode SendSyncRequest(Core::System& system, Handle handle) { LOG_TRACE(Kernel_SVC, "called handle=0x{:08X}({})", handle, session->GetName()); - auto thread = system.CurrentScheduler().GetCurrentThread(); + auto thread = kernel.CurrentScheduler()->GetCurrentThread(); { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(kernel); thread->InvalidateHLECallback(); thread->SetStatus(ThreadStatus::WaitIPC); session->SendSyncRequest(SharedFrom(thread), system.Memory(), system.CoreTiming()); @@ -352,12 +351,12 @@ static ResultCode SendSyncRequest(Core::System& system, Handle handle) { if (thread->HasHLECallback()) { Handle event_handle = thread->GetHLETimeEvent(); if (event_handle != InvalidHandle) { - auto& time_manager = system.Kernel().TimeManager(); + auto& time_manager = kernel.TimeManager(); time_manager.UnscheduleTimeEvent(event_handle); } { - SchedulerLock lock(system.Kernel()); + KScopedSchedulerLock lock(kernel); auto* sync_object = thread->GetHLESyncObject(); sync_object->RemoveWaitingThread(SharedFrom(thread)); } @@ -531,8 +530,7 @@ static ResultCode ArbitrateLock(Core::System& system, Handle holding_thread_hand static ResultCode ArbitrateLock32(Core::System& system, Handle holding_thread_handle, u32 mutex_addr, Handle requesting_thread_handle) { - return ArbitrateLock(system, holding_thread_handle, static_cast<VAddr>(mutex_addr), - requesting_thread_handle); + return ArbitrateLock(system, holding_thread_handle, mutex_addr, requesting_thread_handle); } /// Unlock a mutex @@ -555,7 +553,7 @@ static ResultCode ArbitrateUnlock(Core::System& system, VAddr mutex_addr) { } static ResultCode ArbitrateUnlock32(Core::System& system, u32 mutex_addr) { - return ArbitrateUnlock(system, static_cast<VAddr>(mutex_addr)); + return ArbitrateUnlock(system, mutex_addr); } enum class BreakType : u32 { @@ -658,7 +656,6 @@ static void Break(Core::System& system, u32 reason, u64 info1, u64 info2) { info2, has_dumped_buffer ? std::make_optional(debug_buffer) : std::nullopt); if (!break_reason.signal_debugger) { - SchedulerLock lock(system.Kernel()); LOG_CRITICAL( Debug_Emulated, "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}", @@ -666,22 +663,18 @@ static void Break(Core::System& system, u32 reason, u64 info1, u64 info2) { handle_debug_buffer(info1, info2); - auto* const current_thread = system.CurrentScheduler().GetCurrentThread(); + auto* const current_thread = system.Kernel().CurrentScheduler()->GetCurrentThread(); const auto thread_processor_id = current_thread->GetProcessorID(); system.ArmInterface(static_cast<std::size_t>(thread_processor_id)).LogBacktrace(); - - // Kill the current thread - system.Kernel().ExceptionalExit(); - current_thread->Stop(); } } static void Break32(Core::System& system, u32 reason, u32 info1, u32 info2) { - Break(system, reason, static_cast<u64>(info1), static_cast<u64>(info2)); + Break(system, reason, info1, info2); } /// Used to output a message on a debug hardware unit - does nothing on a retail unit -static void OutputDebugString([[maybe_unused]] Core::System& system, VAddr address, u64 len) { +static void OutputDebugString(Core::System& system, VAddr address, u64 len) { if (len == 0) { return; } @@ -922,7 +915,7 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha } const auto& core_timing = system.CoreTiming(); - const auto& scheduler = system.CurrentScheduler(); + const auto& scheduler = *system.Kernel().CurrentScheduler(); const auto* const current_thread = scheduler.GetCurrentThread(); const bool same_thread = current_thread == thread.get(); @@ -948,7 +941,7 @@ static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 ha static ResultCode GetInfo32(Core::System& system, u32* result_low, u32* result_high, u32 sub_id_low, u32 info_id, u32 handle, u32 sub_id_high) { - const u64 sub_id{static_cast<u64>(sub_id_low | (static_cast<u64>(sub_id_high) << 32))}; + const u64 sub_id{u64{sub_id_low} | (u64{sub_id_high} << 32)}; u64 res_value{}; const ResultCode result{GetInfo(system, &res_value, info_id, handle, sub_id)}; @@ -1009,7 +1002,7 @@ static ResultCode MapPhysicalMemory(Core::System& system, VAddr addr, u64 size) } static ResultCode MapPhysicalMemory32(Core::System& system, u32 addr, u32 size) { - return MapPhysicalMemory(system, static_cast<VAddr>(addr), static_cast<std::size_t>(size)); + return MapPhysicalMemory(system, addr, size); } /// Unmaps memory previously mapped via MapPhysicalMemory @@ -1063,7 +1056,7 @@ static ResultCode UnmapPhysicalMemory(Core::System& system, VAddr addr, u64 size } static ResultCode UnmapPhysicalMemory32(Core::System& system, u32 addr, u32 size) { - return UnmapPhysicalMemory(system, static_cast<VAddr>(addr), static_cast<std::size_t>(size)); + return UnmapPhysicalMemory(system, addr, size); } /// Sets the thread activity @@ -1090,7 +1083,7 @@ static ResultCode SetThreadActivity(Core::System& system, Handle handle, u32 act return ERR_INVALID_HANDLE; } - if (thread.get() == system.CurrentScheduler().GetCurrentThread()) { + if (thread.get() == system.Kernel().CurrentScheduler()->GetCurrentThread()) { LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread"); return ERR_BUSY; } @@ -1123,7 +1116,7 @@ static ResultCode GetThreadContext(Core::System& system, VAddr thread_context, H return ERR_INVALID_HANDLE; } - if (thread.get() == system.CurrentScheduler().GetCurrentThread()) { + if (thread.get() == system.Kernel().CurrentScheduler()->GetCurrentThread()) { LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread"); return ERR_BUSY; } @@ -1144,7 +1137,7 @@ static ResultCode GetThreadContext(Core::System& system, VAddr thread_context, H } static ResultCode GetThreadContext32(Core::System& system, u32 thread_context, Handle handle) { - return GetThreadContext(system, static_cast<VAddr>(thread_context), handle); + return GetThreadContext(system, thread_context, handle); } /// Gets the priority for the specified thread @@ -1281,8 +1274,7 @@ static ResultCode MapSharedMemory(Core::System& system, Handle shared_memory_han static ResultCode MapSharedMemory32(Core::System& system, Handle shared_memory_handle, u32 addr, u32 size, u32 permissions) { - return MapSharedMemory(system, shared_memory_handle, static_cast<VAddr>(addr), - static_cast<std::size_t>(size), permissions); + return MapSharedMemory(system, shared_memory_handle, addr, size, permissions); } static ResultCode QueryProcessMemory(Core::System& system, VAddr memory_info_address, @@ -1480,7 +1472,7 @@ static void ExitProcess(Core::System& system) { current_process->PrepareForTermination(); // Kill the current thread - system.CurrentScheduler().GetCurrentThread()->Stop(); + system.Kernel().CurrentScheduler()->GetCurrentThread()->Stop(); } static void ExitProcess32(Core::System& system) { @@ -1552,8 +1544,7 @@ static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr e static ResultCode CreateThread32(Core::System& system, Handle* out_handle, u32 priority, u32 entry_point, u32 arg, u32 stack_top, s32 processor_id) { - return CreateThread(system, out_handle, static_cast<VAddr>(entry_point), static_cast<u64>(arg), - static_cast<VAddr>(stack_top), priority, processor_id); + return CreateThread(system, out_handle, entry_point, arg, stack_top, priority, processor_id); } /// Starts the thread for the provided handle @@ -1581,8 +1572,8 @@ static ResultCode StartThread32(Core::System& system, Handle thread_handle) { static void ExitThread(Core::System& system) { LOG_DEBUG(Kernel_SVC, "called, pc=0x{:08X}", system.CurrentArmInterface().GetPC()); - auto* const current_thread = system.CurrentScheduler().GetCurrentThread(); - system.GlobalScheduler().RemoveThread(SharedFrom(current_thread)); + auto* const current_thread = system.Kernel().CurrentScheduler()->GetCurrentThread(); + system.GlobalSchedulerContext().RemoveThread(SharedFrom(current_thread)); current_thread->Stop(); } @@ -1592,53 +1583,39 @@ static void ExitThread32(Core::System& system) { /// Sleep the current thread static void SleepThread(Core::System& system, s64 nanoseconds) { - LOG_DEBUG(Kernel_SVC, "called nanoseconds={}", nanoseconds); + LOG_TRACE(Kernel_SVC, "called nanoseconds={}", nanoseconds); enum class SleepType : s64 { - YieldWithoutLoadBalancing = 0, - YieldWithLoadBalancing = -1, + YieldWithoutCoreMigration = 0, + YieldWithCoreMigration = -1, YieldAndWaitForLoadBalancing = -2, }; - auto& scheduler = system.CurrentScheduler(); - auto* const current_thread = scheduler.GetCurrentThread(); - bool is_redundant = false; - + auto& scheduler = *system.Kernel().CurrentScheduler(); if (nanoseconds <= 0) { switch (static_cast<SleepType>(nanoseconds)) { - case SleepType::YieldWithoutLoadBalancing: { - auto pair = current_thread->YieldSimple(); - is_redundant = pair.second; + case SleepType::YieldWithoutCoreMigration: { + scheduler.YieldWithoutCoreMigration(); break; } - case SleepType::YieldWithLoadBalancing: { - auto pair = current_thread->YieldAndBalanceLoad(); - is_redundant = pair.second; + case SleepType::YieldWithCoreMigration: { + scheduler.YieldWithCoreMigration(); break; } case SleepType::YieldAndWaitForLoadBalancing: { - auto pair = current_thread->YieldAndWaitForLoadBalancing(); - is_redundant = pair.second; + scheduler.YieldToAnyThread(); break; } default: UNREACHABLE_MSG("Unimplemented sleep yield type '{:016X}'!", nanoseconds); } } else { - current_thread->Sleep(nanoseconds); - } - - if (is_redundant && !system.Kernel().IsMulticore()) { - system.Kernel().ExitSVCProfile(); - system.CoreTiming().AddTicks(1000U); - system.GetCpuManager().PreemptSingleCore(); - system.Kernel().EnterSVCProfile(); + scheduler.GetCurrentThread()->Sleep(nanoseconds); } } static void SleepThread32(Core::System& system, u32 nanoseconds_low, u32 nanoseconds_high) { - const s64 nanoseconds = static_cast<s64>(static_cast<u64>(nanoseconds_low) | - (static_cast<u64>(nanoseconds_high) << 32)); + const auto nanoseconds = static_cast<s64>(u64{nanoseconds_low} | (u64{nanoseconds_high} << 32)); SleepThread(system, nanoseconds); } @@ -1668,10 +1645,10 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add ASSERT(condition_variable_addr == Common::AlignDown(condition_variable_addr, 4)); auto& kernel = system.Kernel(); Handle event_handle; - Thread* current_thread = system.CurrentScheduler().GetCurrentThread(); - auto* const current_process = system.Kernel().CurrentProcess(); + Thread* current_thread = kernel.CurrentScheduler()->GetCurrentThread(); + auto* const current_process = kernel.CurrentProcess(); { - SchedulerLockAndSleep lock(kernel, event_handle, current_thread, nano_seconds); + KScopedSchedulerLockAndSleep lock(kernel, event_handle, current_thread, nano_seconds); const auto& handle_table = current_process->GetHandleTable(); std::shared_ptr<Thread> thread = handle_table.Get<Thread>(thread_handle); ASSERT(thread); @@ -1707,7 +1684,7 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add } { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); auto* owner = current_thread->GetLockOwner(); if (owner != nullptr) { @@ -1724,10 +1701,8 @@ static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_add static ResultCode WaitProcessWideKeyAtomic32(Core::System& system, u32 mutex_addr, u32 condition_variable_addr, Handle thread_handle, u32 nanoseconds_low, u32 nanoseconds_high) { - const s64 nanoseconds = - static_cast<s64>(nanoseconds_low | (static_cast<u64>(nanoseconds_high) << 32)); - return WaitProcessWideKeyAtomic(system, static_cast<VAddr>(mutex_addr), - static_cast<VAddr>(condition_variable_addr), thread_handle, + const auto nanoseconds = static_cast<s64>(nanoseconds_low | (u64{nanoseconds_high} << 32)); + return WaitProcessWideKeyAtomic(system, mutex_addr, condition_variable_addr, thread_handle, nanoseconds); } @@ -1740,7 +1715,7 @@ static void SignalProcessWideKey(Core::System& system, VAddr condition_variable_ // Retrieve a list of all threads that are waiting for this condition variable. auto& kernel = system.Kernel(); - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); auto* const current_process = kernel.CurrentProcess(); std::vector<std::shared_ptr<Thread>> waiting_threads = current_process->GetConditionVariableThreads(condition_variable_addr); @@ -1833,8 +1808,8 @@ static ResultCode WaitForAddress(Core::System& system, VAddr address, u32 type, static ResultCode WaitForAddress32(Core::System& system, u32 address, u32 type, s32 value, u32 timeout_low, u32 timeout_high) { - s64 timeout = static_cast<s64>(timeout_low | (static_cast<u64>(timeout_high) << 32)); - return WaitForAddress(system, static_cast<VAddr>(address), type, value, timeout); + const auto timeout = static_cast<s64>(timeout_low | (u64{timeout_high} << 32)); + return WaitForAddress(system, address, type, value, timeout); } // Signals to an address (via Address Arbiter) @@ -1862,7 +1837,7 @@ static ResultCode SignalToAddress(Core::System& system, VAddr address, u32 type, static ResultCode SignalToAddress32(Core::System& system, u32 address, u32 type, s32 value, s32 num_to_wake) { - return SignalToAddress(system, static_cast<VAddr>(address), type, value, num_to_wake); + return SignalToAddress(system, address, type, value, num_to_wake); } static void KernelDebug([[maybe_unused]] Core::System& system, @@ -1893,7 +1868,7 @@ static u64 GetSystemTick(Core::System& system) { } static void GetSystemTick32(Core::System& system, u32* time_low, u32* time_high) { - u64 time = GetSystemTick(system); + const auto time = GetSystemTick(system); *time_low = static_cast<u32>(time); *time_high = static_cast<u32>(time >> 32); } @@ -1984,8 +1959,7 @@ static ResultCode CreateTransferMemory(Core::System& system, Handle* handle, VAd static ResultCode CreateTransferMemory32(Core::System& system, Handle* handle, u32 addr, u32 size, u32 permissions) { - return CreateTransferMemory(system, handle, static_cast<VAddr>(addr), - static_cast<std::size_t>(size), permissions); + return CreateTransferMemory(system, handle, addr, size, permissions); } static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, u32* core, @@ -2003,7 +1977,7 @@ static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, } *core = thread->GetIdealCore(); - *mask = thread->GetAffinityMask(); + *mask = thread->GetAffinityMask().GetAffinityMask(); return RESULT_SUCCESS; } @@ -2075,8 +2049,7 @@ static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, static ResultCode SetThreadCoreMask32(Core::System& system, Handle thread_handle, u32 core, u32 affinity_mask_low, u32 affinity_mask_high) { - const u64 affinity_mask = - static_cast<u64>(affinity_mask_low) | (static_cast<u64>(affinity_mask_high) << 32); + const auto affinity_mask = u64{affinity_mask_low} | (u64{affinity_mask_high} << 32); return SetThreadCoreMask(system, thread_handle, core, affinity_mask); } @@ -2341,9 +2314,10 @@ static ResultCode GetThreadList(Core::System& system, u32* out_num_threads, VAdd return RESULT_SUCCESS; } -static ResultCode FlushProcessDataCache32(Core::System& system, Handle handle, u32 address, - u32 size) { - // Note(Blinkhawk): For emulation purposes of the data cache this is mostly a nope +static ResultCode FlushProcessDataCache32([[maybe_unused]] Core::System& system, + [[maybe_unused]] Handle handle, + [[maybe_unused]] u32 address, [[maybe_unused]] u32 size) { + // Note(Blinkhawk): For emulation purposes of the data cache this is mostly a no-op, // as all emulation is done in the same cache level in host architecture, thus data cache // does not need flushing. LOG_DEBUG(Kernel_SVC, "called"); @@ -2639,6 +2613,9 @@ void Call(Core::System& system, u32 immediate) { auto& kernel = system.Kernel(); kernel.EnterSVCProfile(); + auto* thread = kernel.CurrentScheduler()->GetCurrentThread(); + thread->SetContinuousOnSVC(true); + const FunctionDef* info = system.CurrentProcess()->Is64BitProcess() ? GetSVCInfo64(immediate) : GetSVCInfo32(immediate); if (info) { @@ -2652,6 +2629,12 @@ void Call(Core::System& system, u32 immediate) { } kernel.ExitSVCProfile(); + + if (!thread->IsContinuousOnSVC()) { + auto* host_context = thread->GetHostContext().get(); + host_context->Rewind(); + } + system.EnterDynarmicProfile(); } diff --git a/src/core/hle/kernel/svc_types.h b/src/core/hle/kernel/svc_types.h index 986724beb..11e1d8e2d 100644 --- a/src/core/hle/kernel/svc_types.h +++ b/src/core/hle/kernel/svc_types.h @@ -23,8 +23,8 @@ enum class MemoryState : u32 { Ipc = 0x0A, Stack = 0x0B, ThreadLocal = 0x0C, - Transfered = 0x0D, - SharedTransfered = 0x0E, + Transferred = 0x0D, + SharedTransferred = 0x0E, SharedCode = 0x0F, Inaccessible = 0x10, NonSecureIpc = 0x11, diff --git a/src/core/hle/kernel/synchronization.cpp b/src/core/hle/kernel/synchronization.cpp index 8b875d853..d3f520ea2 100644 --- a/src/core/hle/kernel/synchronization.cpp +++ b/src/core/hle/kernel/synchronization.cpp @@ -5,8 +5,9 @@ #include "core/core.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h" #include "core/hle/kernel/kernel.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/synchronization.h" #include "core/hle/kernel/synchronization_object.h" #include "core/hle/kernel/thread.h" @@ -18,7 +19,7 @@ Synchronization::Synchronization(Core::System& system) : system{system} {} void Synchronization::SignalObject(SynchronizationObject& obj) const { auto& kernel = system.Kernel(); - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); if (obj.IsSignaled()) { for (auto thread : obj.GetWaitingThreads()) { if (thread->GetSchedulingStatus() == ThreadSchedStatus::Paused) { @@ -37,10 +38,10 @@ void Synchronization::SignalObject(SynchronizationObject& obj) const { std::pair<ResultCode, Handle> Synchronization::WaitFor( std::vector<std::shared_ptr<SynchronizationObject>>& sync_objects, s64 nano_seconds) { auto& kernel = system.Kernel(); - auto* const thread = system.CurrentScheduler().GetCurrentThread(); + auto* const thread = kernel.CurrentScheduler()->GetCurrentThread(); Handle event_handle = InvalidHandle; { - SchedulerLockAndSleep lock(kernel, event_handle, thread, nano_seconds); + KScopedSchedulerLockAndSleep lock(kernel, event_handle, thread, nano_seconds); const auto itr = std::find_if(sync_objects.begin(), sync_objects.end(), [thread](const std::shared_ptr<SynchronizationObject>& object) { @@ -89,7 +90,7 @@ std::pair<ResultCode, Handle> Synchronization::WaitFor( } { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); ResultCode signaling_result = thread->GetSignalingResult(); SynchronizationObject* signaling_object = thread->GetSignalingObject(); thread->SetSynchronizationObjects(nullptr); diff --git a/src/core/hle/kernel/synchronization_object.h b/src/core/hle/kernel/synchronization_object.h index f89b24204..7408ed51f 100644 --- a/src/core/hle/kernel/synchronization_object.h +++ b/src/core/hle/kernel/synchronization_object.h @@ -4,6 +4,7 @@ #pragma once +#include <atomic> #include <memory> #include <vector> @@ -56,7 +57,7 @@ public: void ClearWaitingThreads(); protected: - bool is_signaled{}; // Tells if this sync object is signalled; + std::atomic_bool is_signaled{}; // Tells if this sync object is signaled private: /// Threads waiting for this object to become available diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index d132aba34..a4f9e0d97 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -12,17 +12,16 @@ #include "common/fiber.h" #include "common/logging/log.h" #include "common/thread_queue_list.h" -#include "core/arm/arm_interface.h" -#include "core/arm/unicorn/arm_unicorn.h" #include "core/core.h" #include "core/cpu_manager.h" #include "core/hardware_properties.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/k_scoped_scheduler_lock_and_sleep.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/process.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "core/hle/kernel/time_manager.h" #include "core/hle/result.h" @@ -52,7 +51,7 @@ Thread::~Thread() = default; void Thread::Stop() { { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); SetStatus(ThreadStatus::Dead); Signal(); kernel.GlobalHandleTable().Close(global_handle); @@ -63,14 +62,13 @@ void Thread::Stop() { // Mark the TLS slot in the thread's page as free. owner_process->FreeTLSRegion(tls_address); } - arm_interface.reset(); has_exited = true; } global_handle = 0; } void Thread::ResumeFromWait() { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); switch (status) { case ThreadStatus::Paused: case ThreadStatus::WaitSynch: @@ -91,10 +89,6 @@ void Thread::ResumeFromWait() { // before actually resuming. We can ignore subsequent wakeups if the thread status has // already been set to ThreadStatus::Ready. return; - - case ThreadStatus::Running: - DEBUG_ASSERT_MSG(false, "Thread with object id {} has already resumed.", GetObjectId()); - return; case ThreadStatus::Dead: // This should never happen, as threads must complete before being stopped. DEBUG_ASSERT_MSG(false, "Thread with object id {} cannot be resumed because it's DEAD.", @@ -106,19 +100,18 @@ void Thread::ResumeFromWait() { } void Thread::OnWakeUp() { - SchedulerLock lock(kernel); - + KScopedSchedulerLock lock(kernel); SetStatus(ThreadStatus::Ready); } ResultCode Thread::Start() { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); SetStatus(ThreadStatus::Ready); return RESULT_SUCCESS; } void Thread::CancelWait() { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); if (GetSchedulingStatus() != ThreadSchedStatus::Paused || !is_waiting_on_sync) { is_sync_cancelled = true; return; @@ -193,12 +186,14 @@ ResultVal<std::shared_ptr<Thread>> Thread::Create(Core::System& system, ThreadTy thread->status = ThreadStatus::Dormant; thread->entry_point = entry_point; thread->stack_top = stack_top; + thread->disable_count = 1; thread->tpidr_el0 = 0; thread->nominal_priority = thread->current_priority = priority; - thread->last_running_ticks = 0; + thread->schedule_count = -1; + thread->last_scheduled_tick = 0; thread->processor_id = processor_id; thread->ideal_core = processor_id; - thread->affinity_mask = 1ULL << processor_id; + thread->affinity_mask.SetAffinity(processor_id, true); thread->wait_objects = nullptr; thread->mutex_wait_address = 0; thread->condvar_wait_address = 0; @@ -208,7 +203,7 @@ ResultVal<std::shared_ptr<Thread>> Thread::Create(Core::System& system, ThreadTy thread->owner_process = owner_process; thread->type = type_flags; if ((type_flags & THREADTYPE_IDLE) == 0) { - auto& scheduler = kernel.GlobalScheduler(); + auto& scheduler = kernel.GlobalSchedulerContext(); scheduler.AddThread(thread); } if (owner_process) { @@ -217,33 +212,10 @@ ResultVal<std::shared_ptr<Thread>> Thread::Create(Core::System& system, ThreadTy } else { thread->tls_address = 0; } + // TODO(peachum): move to ScheduleThread() when scheduler is added so selected core is used // to initialize the context - thread->arm_interface.reset(); if ((type_flags & THREADTYPE_HLE) == 0) { -#ifdef ARCHITECTURE_x86_64 - if (owner_process && !owner_process->Is64BitProcess()) { - thread->arm_interface = std::make_unique<Core::ARM_Dynarmic_32>( - system, kernel.Interrupts(), kernel.IsMulticore(), kernel.GetExclusiveMonitor(), - processor_id); - } else { - thread->arm_interface = std::make_unique<Core::ARM_Dynarmic_64>( - system, kernel.Interrupts(), kernel.IsMulticore(), kernel.GetExclusiveMonitor(), - processor_id); - } - -#else - if (owner_process && !owner_process->Is64BitProcess()) { - thread->arm_interface = std::make_shared<Core::ARM_Unicorn>( - system, kernel.Interrupts(), kernel.IsMulticore(), ARM_Unicorn::Arch::AArch32, - processor_id); - } else { - thread->arm_interface = std::make_shared<Core::ARM_Unicorn>( - system, kernel.Interrupts(), kernel.IsMulticore(), ARM_Unicorn::Arch::AArch64, - processor_id); - } - LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available"); -#endif ResetThreadContext32(thread->context_32, static_cast<u32>(stack_top), static_cast<u32>(entry_point), static_cast<u32>(arg)); ResetThreadContext64(thread->context_64, stack_top, entry_point, arg); @@ -255,7 +227,7 @@ ResultVal<std::shared_ptr<Thread>> Thread::Create(Core::System& system, ThreadTy } void Thread::SetPriority(u32 priority) { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); ASSERT_MSG(priority <= THREADPRIO_LOWEST && priority >= THREADPRIO_HIGHEST, "Invalid priority value."); nominal_priority = priority; @@ -279,14 +251,6 @@ VAddr Thread::GetCommandBufferAddress() const { return GetTLSAddress() + command_header_offset; } -Core::ARM_Interface& Thread::ArmInterface() { - return *arm_interface; -} - -const Core::ARM_Interface& Thread::ArmInterface() const { - return *arm_interface; -} - void Thread::SetStatus(ThreadStatus new_status) { if (new_status == status) { return; @@ -294,7 +258,6 @@ void Thread::SetStatus(ThreadStatus new_status) { switch (new_status) { case ThreadStatus::Ready: - case ThreadStatus::Running: SetSchedulingStatus(ThreadSchedStatus::Runnable); break; case ThreadStatus::Dormant: @@ -401,7 +364,7 @@ bool Thread::InvokeHLECallback(std::shared_ptr<Thread> thread) { } ResultCode Thread::SetActivity(ThreadActivity value) { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); auto sched_status = GetSchedulingStatus(); @@ -430,7 +393,7 @@ ResultCode Thread::SetActivity(ThreadActivity value) { ResultCode Thread::Sleep(s64 nanoseconds) { Handle event_handle{}; { - SchedulerLockAndSleep lock(kernel, event_handle, this, nanoseconds); + KScopedSchedulerLockAndSleep lock(kernel, event_handle, this, nanoseconds); SetStatus(ThreadStatus::WaitSleep); } @@ -441,39 +404,12 @@ ResultCode Thread::Sleep(s64 nanoseconds) { return RESULT_SUCCESS; } -std::pair<ResultCode, bool> Thread::YieldSimple() { - bool is_redundant = false; - { - SchedulerLock lock(kernel); - is_redundant = kernel.GlobalScheduler().YieldThread(this); - } - return {RESULT_SUCCESS, is_redundant}; -} - -std::pair<ResultCode, bool> Thread::YieldAndBalanceLoad() { - bool is_redundant = false; - { - SchedulerLock lock(kernel); - is_redundant = kernel.GlobalScheduler().YieldThreadAndBalanceLoad(this); - } - return {RESULT_SUCCESS, is_redundant}; -} - -std::pair<ResultCode, bool> Thread::YieldAndWaitForLoadBalancing() { - bool is_redundant = false; - { - SchedulerLock lock(kernel); - is_redundant = kernel.GlobalScheduler().YieldThreadAndWaitForLoadBalancing(this); - } - return {RESULT_SUCCESS, is_redundant}; -} - void Thread::AddSchedulingFlag(ThreadSchedFlags flag) { const u32 old_state = scheduling_state; pausing_state |= static_cast<u32>(flag); const u32 base_scheduling = static_cast<u32>(GetSchedulingStatus()); scheduling_state = base_scheduling | pausing_state; - kernel.GlobalScheduler().AdjustSchedulingOnStatus(this, old_state); + KScheduler::OnThreadStateChanged(kernel, this, old_state); } void Thread::RemoveSchedulingFlag(ThreadSchedFlags flag) { @@ -481,23 +417,24 @@ void Thread::RemoveSchedulingFlag(ThreadSchedFlags flag) { pausing_state &= ~static_cast<u32>(flag); const u32 base_scheduling = static_cast<u32>(GetSchedulingStatus()); scheduling_state = base_scheduling | pausing_state; - kernel.GlobalScheduler().AdjustSchedulingOnStatus(this, old_state); + KScheduler::OnThreadStateChanged(kernel, this, old_state); } void Thread::SetSchedulingStatus(ThreadSchedStatus new_status) { const u32 old_state = scheduling_state; scheduling_state = (scheduling_state & static_cast<u32>(ThreadSchedMasks::HighMask)) | static_cast<u32>(new_status); - kernel.GlobalScheduler().AdjustSchedulingOnStatus(this, old_state); + KScheduler::OnThreadStateChanged(kernel, this, old_state); } void Thread::SetCurrentPriority(u32 new_priority) { const u32 old_priority = std::exchange(current_priority, new_priority); - kernel.GlobalScheduler().AdjustSchedulingOnPriority(this, old_priority); + KScheduler::OnThreadPriorityChanged(kernel, this, kernel.CurrentScheduler()->GetCurrentThread(), + old_priority); } ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) { - SchedulerLock lock(kernel); + KScopedSchedulerLock lock(kernel); const auto HighestSetCore = [](u64 mask, u32 max_cores) { for (s32 core = static_cast<s32>(max_cores - 1); core >= 0; core--) { if (((mask >> core) & 1) != 0) { @@ -518,20 +455,21 @@ ResultCode Thread::SetCoreAndAffinityMask(s32 new_core, u64 new_affinity_mask) { } if (use_override) { ideal_core_override = new_core; - affinity_mask_override = new_affinity_mask; } else { - const u64 old_affinity_mask = std::exchange(affinity_mask, new_affinity_mask); + const auto old_affinity_mask = affinity_mask; + affinity_mask.SetAffinityMask(new_affinity_mask); ideal_core = new_core; - if (old_affinity_mask != new_affinity_mask) { + if (old_affinity_mask.GetAffinityMask() != new_affinity_mask) { const s32 old_core = processor_id; - if (processor_id >= 0 && ((affinity_mask >> processor_id) & 1) == 0) { + if (processor_id >= 0 && !affinity_mask.GetAffinity(processor_id)) { if (static_cast<s32>(ideal_core) < 0) { - processor_id = HighestSetCore(affinity_mask, Core::Hardware::NUM_CPU_CORES); + processor_id = HighestSetCore(affinity_mask.GetAffinityMask(), + Core::Hardware::NUM_CPU_CORES); } else { processor_id = ideal_core; } } - kernel.GlobalScheduler().AdjustSchedulingOnAffinity(this, old_affinity_mask, old_core); + KScheduler::OnThreadAffinityMaskChanged(kernel, this, old_affinity_mask, old_core); } } return RESULT_SUCCESS; diff --git a/src/core/hle/kernel/thread.h b/src/core/hle/kernel/thread.h index 8daf79fac..11ef29888 100644 --- a/src/core/hle/kernel/thread.h +++ b/src/core/hle/kernel/thread.h @@ -4,6 +4,7 @@ #pragma once +#include <array> #include <functional> #include <string> #include <utility> @@ -12,6 +13,7 @@ #include "common/common_types.h" #include "common/spin_lock.h" #include "core/arm/arm_interface.h" +#include "core/hle/kernel/k_affinity_mask.h" #include "core/hle/kernel/object.h" #include "core/hle/kernel/synchronization_object.h" #include "core/hle/result.h" @@ -27,10 +29,10 @@ class System; namespace Kernel { -class GlobalScheduler; +class GlobalSchedulerContext; class KernelCore; class Process; -class Scheduler; +class KScheduler; enum ThreadPriority : u32 { THREADPRIO_HIGHEST = 0, ///< Highest thread priority @@ -72,7 +74,6 @@ enum ThreadProcessorId : s32 { }; enum class ThreadStatus { - Running, ///< Currently running Ready, ///< Ready to run Paused, ///< Paused by SetThreadActivity or debug WaitHLEEvent, ///< Waiting for hle event to finish @@ -248,10 +249,6 @@ public: void SetSynchronizationResults(SynchronizationObject* object, ResultCode result); - Core::ARM_Interface& ArmInterface(); - - const Core::ARM_Interface& ArmInterface() const; - SynchronizationObject* GetSignalingObject() const { return signaling_object; } @@ -350,8 +347,12 @@ public: void SetStatus(ThreadStatus new_status); - u64 GetLastRunningTicks() const { - return last_running_ticks; + s64 GetLastScheduledTick() const { + return this->last_scheduled_tick; + } + + void SetLastScheduledTick(s64 tick) { + this->last_scheduled_tick = tick; } u64 GetTotalCPUTimeTicks() const { @@ -366,10 +367,18 @@ public: return processor_id; } + s32 GetActiveCore() const { + return GetProcessorID(); + } + void SetProcessorID(s32 new_core) { processor_id = new_core; } + void SetActiveCore(s32 new_core) { + processor_id = new_core; + } + Process* GetOwnerProcess() { return owner_process; } @@ -474,7 +483,7 @@ public: return ideal_core; } - u64 GetAffinityMask() const { + const KAffinityMask& GetAffinityMask() const { return affinity_mask; } @@ -483,21 +492,12 @@ public: /// Sleeps this thread for the given amount of nanoseconds. ResultCode Sleep(s64 nanoseconds); - /// Yields this thread without rebalancing loads. - std::pair<ResultCode, bool> YieldSimple(); - - /// Yields this thread and does a load rebalancing. - std::pair<ResultCode, bool> YieldAndBalanceLoad(); - - /// Yields this thread and if the core is left idle, loads are rebalanced - std::pair<ResultCode, bool> YieldAndWaitForLoadBalancing(); - - void IncrementYieldCount() { - yield_count++; + s64 GetYieldScheduleCount() const { + return this->schedule_count; } - u64 GetYieldCount() const { - return yield_count; + void SetYieldScheduleCount(s64 count) { + this->schedule_count = count; } ThreadSchedStatus GetSchedulingStatus() const { @@ -573,9 +573,59 @@ public: return has_exited; } + class QueueEntry { + public: + constexpr QueueEntry() = default; + + constexpr void Initialize() { + this->prev = nullptr; + this->next = nullptr; + } + + constexpr Thread* GetPrev() const { + return this->prev; + } + constexpr Thread* GetNext() const { + return this->next; + } + constexpr void SetPrev(Thread* thread) { + this->prev = thread; + } + constexpr void SetNext(Thread* thread) { + this->next = thread; + } + + private: + Thread* prev{}; + Thread* next{}; + }; + + QueueEntry& GetPriorityQueueEntry(s32 core) { + return this->per_core_priority_queue_entry[core]; + } + + const QueueEntry& GetPriorityQueueEntry(s32 core) const { + return this->per_core_priority_queue_entry[core]; + } + + s32 GetDisableDispatchCount() const { + return disable_count; + } + + void DisableDispatch() { + ASSERT(GetDisableDispatchCount() >= 0); + disable_count++; + } + + void EnableDispatch() { + ASSERT(GetDisableDispatchCount() > 0); + disable_count--; + } + private: - friend class GlobalScheduler; - friend class Scheduler; + friend class GlobalSchedulerContext; + friend class KScheduler; + friend class Process; void SetSchedulingStatus(ThreadSchedStatus new_status); void AddSchedulingFlag(ThreadSchedFlags flag); @@ -586,15 +636,16 @@ private: Common::SpinLock context_guard{}; ThreadContext32 context_32{}; ThreadContext64 context_64{}; - std::unique_ptr<Core::ARM_Interface> arm_interface{}; std::shared_ptr<Common::Fiber> host_context{}; - u64 thread_id = 0; - ThreadStatus status = ThreadStatus::Dormant; + u32 scheduling_state = 0; + + u64 thread_id = 0; VAddr entry_point = 0; VAddr stack_top = 0; + std::atomic_int disable_count = 0; ThreadType type; @@ -608,9 +659,8 @@ private: u32 current_priority = 0; u64 total_cpu_time_ticks = 0; ///< Total CPU running ticks. - u64 last_running_ticks = 0; ///< CPU tick when thread was last running - u64 yield_count = 0; ///< Number of redundant yields carried by this thread. - ///< a redundant yield is one where no scheduling is changed + s64 schedule_count{}; + s64 last_scheduled_tick{}; s32 processor_id = 0; @@ -652,16 +702,16 @@ private: Handle hle_time_event; SynchronizationObject* hle_object; - Scheduler* scheduler = nullptr; + KScheduler* scheduler = nullptr; + + std::array<QueueEntry, Core::Hardware::NUM_CPU_CORES> per_core_priority_queue_entry{}; u32 ideal_core{0xFFFFFFFF}; - u64 affinity_mask{0x1}; + KAffinityMask affinity_mask{}; s32 ideal_core_override = -1; - u64 affinity_mask_override = 0x1; u32 affinity_override_count = 0; - u32 scheduling_state = 0; u32 pausing_state = 0; bool is_running = false; bool is_waiting_on_sync = false; diff --git a/src/core/hle/kernel/time_manager.cpp b/src/core/hle/kernel/time_manager.cpp index 95f2446c9..79628e2b4 100644 --- a/src/core/hle/kernel/time_manager.cpp +++ b/src/core/hle/kernel/time_manager.cpp @@ -7,8 +7,8 @@ #include "core/core_timing.h" #include "core/core_timing_util.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/kernel.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/thread.h" #include "core/hle/kernel/time_manager.h" @@ -18,17 +18,27 @@ TimeManager::TimeManager(Core::System& system_) : system{system_} { time_manager_event_type = Core::Timing::CreateEvent( "Kernel::TimeManagerCallback", [this](std::uintptr_t thread_handle, std::chrono::nanoseconds) { - const SchedulerLock lock(system.Kernel()); + const KScopedSchedulerLock lock(system.Kernel()); const auto proper_handle = static_cast<Handle>(thread_handle); - if (cancelled_events[proper_handle]) { - return; + + std::shared_ptr<Thread> thread; + { + std::lock_guard lock{mutex}; + if (cancelled_events[proper_handle]) { + return; + } + thread = system.Kernel().RetrieveThreadFromGlobalHandleTable(proper_handle); + } + + if (thread) { + // Thread can be null if process has exited + thread->OnWakeUp(); } - auto thread = this->system.Kernel().RetrieveThreadFromGlobalHandleTable(proper_handle); - thread->OnWakeUp(); }); } void TimeManager::ScheduleTimeEvent(Handle& event_handle, Thread* timetask, s64 nanoseconds) { + std::lock_guard lock{mutex}; event_handle = timetask->GetGlobalHandle(); if (nanoseconds > 0) { ASSERT(timetask); @@ -43,6 +53,7 @@ void TimeManager::ScheduleTimeEvent(Handle& event_handle, Thread* timetask, s64 } void TimeManager::UnscheduleTimeEvent(Handle event_handle) { + std::lock_guard lock{mutex}; if (event_handle == InvalidHandle) { return; } @@ -51,7 +62,8 @@ void TimeManager::UnscheduleTimeEvent(Handle event_handle) { } void TimeManager::CancelTimeEvent(Thread* time_task) { - Handle event_handle = time_task->GetGlobalHandle(); + std::lock_guard lock{mutex}; + const Handle event_handle = time_task->GetGlobalHandle(); UnscheduleTimeEvent(event_handle); } diff --git a/src/core/hle/kernel/time_manager.h b/src/core/hle/kernel/time_manager.h index 307a18765..f39df39a0 100644 --- a/src/core/hle/kernel/time_manager.h +++ b/src/core/hle/kernel/time_manager.h @@ -5,6 +5,7 @@ #pragma once #include <memory> +#include <mutex> #include <unordered_map> #include "core/hle/kernel/object.h" @@ -42,6 +43,7 @@ private: Core::System& system; std::shared_ptr<Core::Timing::EventType> time_manager_event_type; std::unordered_map<Handle, bool> cancelled_events; + std::mutex mutex; }; } // namespace Kernel diff --git a/src/core/hle/result.h b/src/core/hle/result.h index b6bdbd988..8feda7ad7 100644 --- a/src/core/hle/result.h +++ b/src/core/hle/result.h @@ -119,7 +119,7 @@ union ResultCode { BitField<0, 9, ErrorModule> module; BitField<9, 13, u32> description; - constexpr explicit ResultCode(u32 raw) : raw(raw) {} + constexpr explicit ResultCode(u32 raw_) : raw(raw_) {} constexpr ResultCode(ErrorModule module_, u32 description_) : raw(module.FormatValue(module_) | description.FormatValue(description_)) {} diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp index 6b1613510..6981f8ee7 100644 --- a/src/core/hle/service/acc/acc.cpp +++ b/src/core/hle/service/acc/acc.cpp @@ -11,6 +11,7 @@ #include "common/string_util.h" #include "common/swap.h" #include "core/constants.h" +#include "core/core.h" #include "core/core_timing.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" @@ -46,8 +47,8 @@ static constexpr u32 SanitizeJPEGSize(std::size_t size) { class IManagerForSystemService final : public ServiceFramework<IManagerForSystemService> { public: - explicit IManagerForSystemService(Common::UUID user_id) - : ServiceFramework("IManagerForSystemService") { + explicit IManagerForSystemService(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IManagerForSystemService"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CheckAvailability"}, @@ -82,8 +83,8 @@ public: // 3.0.0+ class IFloatingRegistrationRequest final : public ServiceFramework<IFloatingRegistrationRequest> { public: - explicit IFloatingRegistrationRequest(Common::UUID user_id) - : ServiceFramework("IFloatingRegistrationRequest") { + explicit IFloatingRegistrationRequest(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IFloatingRegistrationRequest"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetSessionId"}, @@ -107,7 +108,8 @@ public: class IAdministrator final : public ServiceFramework<IAdministrator> { public: - explicit IAdministrator(Common::UUID user_id) : ServiceFramework("IAdministrator") { + explicit IAdministrator(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IAdministrator"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CheckAvailability"}, @@ -164,8 +166,8 @@ public: class IAuthorizationRequest final : public ServiceFramework<IAuthorizationRequest> { public: - explicit IAuthorizationRequest(Common::UUID user_id) - : ServiceFramework("IAuthorizationRequest") { + explicit IAuthorizationRequest(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IAuthorizationRequest"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetSessionId"}, @@ -183,7 +185,8 @@ public: class IOAuthProcedure final : public ServiceFramework<IOAuthProcedure> { public: - explicit IOAuthProcedure(Common::UUID user_id) : ServiceFramework("IOAuthProcedure") { + explicit IOAuthProcedure(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IOAuthProcedure"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "PrepareAsync"}, @@ -201,8 +204,8 @@ public: // 3.0.0+ class IOAuthProcedureForExternalNsa final : public ServiceFramework<IOAuthProcedureForExternalNsa> { public: - explicit IOAuthProcedureForExternalNsa(Common::UUID user_id) - : ServiceFramework("IOAuthProcedureForExternalNsa") { + explicit IOAuthProcedureForExternalNsa(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IOAuthProcedureForExternalNsa"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "PrepareAsync"}, @@ -224,8 +227,8 @@ public: class IOAuthProcedureForNintendoAccountLinkage final : public ServiceFramework<IOAuthProcedureForNintendoAccountLinkage> { public: - explicit IOAuthProcedureForNintendoAccountLinkage(Common::UUID user_id) - : ServiceFramework("IOAuthProcedureForNintendoAccountLinkage") { + explicit IOAuthProcedureForNintendoAccountLinkage(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IOAuthProcedureForNintendoAccountLinkage"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "PrepareAsync"}, @@ -245,7 +248,8 @@ public: class INotifier final : public ServiceFramework<INotifier> { public: - explicit INotifier(Common::UUID user_id) : ServiceFramework("INotifier") { + explicit INotifier(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "INotifier"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetSystemEvent"}, @@ -258,9 +262,9 @@ public: class IProfileCommon : public ServiceFramework<IProfileCommon> { public: - explicit IProfileCommon(const char* name, bool editor_commands, Common::UUID user_id, - ProfileManager& profile_manager) - : ServiceFramework(name), profile_manager(profile_manager), user_id(user_id) { + explicit IProfileCommon(Core::System& system_, const char* name, bool editor_commands, + Common::UUID user_id_, ProfileManager& profile_manager_) + : ServiceFramework{system_, name}, profile_manager{profile_manager_}, user_id{user_id_} { static const FunctionInfo functions[] = { {0, &IProfileCommon::Get, "Get"}, {1, &IProfileCommon::GetBase, "GetBase"}, @@ -426,19 +430,21 @@ protected: class IProfile final : public IProfileCommon { public: - IProfile(Common::UUID user_id, ProfileManager& profile_manager) - : IProfileCommon("IProfile", false, user_id, profile_manager) {} + explicit IProfile(Core::System& system_, Common::UUID user_id_, + ProfileManager& profile_manager_) + : IProfileCommon{system_, "IProfile", false, user_id_, profile_manager_} {} }; class IProfileEditor final : public IProfileCommon { public: - IProfileEditor(Common::UUID user_id, ProfileManager& profile_manager) - : IProfileCommon("IProfileEditor", true, user_id, profile_manager) {} + explicit IProfileEditor(Core::System& system_, Common::UUID user_id_, + ProfileManager& profile_manager_) + : IProfileCommon{system_, "IProfileEditor", true, user_id_, profile_manager_} {} }; class IAsyncContext final : public ServiceFramework<IAsyncContext> { public: - explicit IAsyncContext(Common::UUID user_id) : ServiceFramework("IAsyncContext") { + explicit IAsyncContext(Core::System& system_) : ServiceFramework{system_, "IAsyncContext"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetSystemEvent"}, @@ -454,7 +460,8 @@ public: class ISessionObject final : public ServiceFramework<ISessionObject> { public: - explicit ISessionObject(Common::UUID user_id) : ServiceFramework("ISessionObject") { + explicit ISessionObject(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "ISessionObject"} { // clang-format off static const FunctionInfo functions[] = { {999, nullptr, "Dummy"}, @@ -467,7 +474,8 @@ public: class IGuestLoginRequest final : public ServiceFramework<IGuestLoginRequest> { public: - explicit IGuestLoginRequest(Common::UUID) : ServiceFramework("IGuestLoginRequest") { + explicit IGuestLoginRequest(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IGuestLoginRequest"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetSessionId"}, @@ -486,8 +494,8 @@ public: class IManagerForApplication final : public ServiceFramework<IManagerForApplication> { public: - explicit IManagerForApplication(Common::UUID user_id) - : ServiceFramework("IManagerForApplication"), user_id(user_id) { + explicit IManagerForApplication(Core::System& system_, Common::UUID user_id_) + : ServiceFramework{system_, "IManagerForApplication"}, user_id{user_id_} { // clang-format off static const FunctionInfo functions[] = { {0, &IManagerForApplication::CheckAvailability, "CheckAvailability"}, @@ -496,7 +504,7 @@ public: {3, nullptr, "LoadIdTokenCache"}, {130, nullptr, "GetNintendoAccountUserResourceCacheForApplication"}, {150, nullptr, "CreateAuthorizationRequest"}, - {160, nullptr, "StoreOpenContext"}, + {160, &IManagerForApplication::StoreOpenContext, "StoreOpenContext"}, {170, nullptr, "LoadNetworkServiceLicenseKindAsync"}, }; // clang-format on @@ -520,6 +528,12 @@ private: rb.PushRaw<u64>(user_id.GetNintendoID()); } + void StoreOpenContext(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_ACC, "(STUBBED) called"); + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + Common::UUID user_id; }; @@ -527,8 +541,8 @@ private: class IAsyncNetworkServiceLicenseKindContext final : public ServiceFramework<IAsyncNetworkServiceLicenseKindContext> { public: - explicit IAsyncNetworkServiceLicenseKindContext(Common::UUID user_id) - : ServiceFramework("IAsyncNetworkServiceLicenseKindContext") { + explicit IAsyncNetworkServiceLicenseKindContext(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IAsyncNetworkServiceLicenseKindContext"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetSystemEvent"}, @@ -547,8 +561,8 @@ public: class IOAuthProcedureForUserRegistration final : public ServiceFramework<IOAuthProcedureForUserRegistration> { public: - explicit IOAuthProcedureForUserRegistration(Common::UUID user_id) - : ServiceFramework("IOAuthProcedureForUserRegistration") { + explicit IOAuthProcedureForUserRegistration(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IOAuthProcedureForUserRegistration"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "PrepareAsync"}, @@ -571,7 +585,7 @@ public: class DAUTH_O final : public ServiceFramework<DAUTH_O> { public: - explicit DAUTH_O(Common::UUID) : ServiceFramework("dauth:o") { + explicit DAUTH_O(Core::System& system_, Common::UUID) : ServiceFramework{system_, "dauth:o"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "EnsureAuthenticationTokenCacheAsync"}, // [5.0.0-5.1.0] GeneratePostData @@ -590,7 +604,8 @@ public: // 6.0.0+ class IAsyncResult final : public ServiceFramework<IAsyncResult> { public: - explicit IAsyncResult(Common::UUID user_id) : ServiceFramework("IAsyncResult") { + explicit IAsyncResult(Core::System& system_, Common::UUID) + : ServiceFramework{system_, "IAsyncResult"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetResult"}, @@ -649,7 +664,7 @@ void Module::Interface::GetProfile(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IProfile>(user_id, *profile_manager); + rb.PushIpcInterface<IProfile>(system, user_id, *profile_manager); } void Module::Interface::IsUserRegistrationRequestPermitted(Kernel::HLERequestContext& ctx) { @@ -724,7 +739,7 @@ void Module::Interface::GetBaasAccountManagerForApplication(Kernel::HLERequestCo LOG_DEBUG(Service_ACC, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IManagerForApplication>(profile_manager->GetLastOpenedUser()); + rb.PushIpcInterface<IManagerForApplication>(system, profile_manager->GetLastOpenedUser()); } void Module::Interface::IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx) { @@ -735,8 +750,10 @@ void Module::Interface::IsUserAccountSwitchLocked(Kernel::HLERequestContext& ctx bool is_locked = false; if (res != Loader::ResultStatus::Success) { - FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID()}; - auto nacp_unique = pm.GetControlMetadata().first; + const FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID(), + system.GetFileSystemController(), + system.GetContentProvider()}; + const auto nacp_unique = pm.GetControlMetadata().first; if (nacp_unique != nullptr) { is_locked = nacp_unique->GetUserAccountSwitchLock(); @@ -760,7 +777,7 @@ void Module::Interface::GetProfileEditor(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IProfileEditor>(user_id, *profile_manager); + rb.PushIpcInterface<IProfileEditor>(system, user_id, *profile_manager); } void Module::Interface::ListQualifiedUsers(Kernel::HLERequestContext& ctx) { @@ -782,7 +799,7 @@ void Module::Interface::LoadOpenContext(Kernel::HLERequestContext& ctx) { // TODO: Find the differences between this and GetBaasAccountManagerForApplication IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IManagerForApplication>(profile_manager->GetLastOpenedUser()); + rb.PushIpcInterface<IManagerForApplication>(system, profile_manager->GetLastOpenedUser()); } void Module::Interface::ListOpenContextStoredUsers(Kernel::HLERequestContext& ctx) { @@ -818,11 +835,11 @@ void Module::Interface::TrySelectUserWithoutInteraction(Kernel::HLERequestContex rb.PushRaw<u128>(profile_manager->GetUser(0)->uuid); } -Module::Interface::Interface(std::shared_ptr<Module> module, - std::shared_ptr<ProfileManager> profile_manager, Core::System& system, - const char* name) - : ServiceFramework(name), module(std::move(module)), - profile_manager(std::move(profile_manager)), system(system) {} +Module::Interface::Interface(std::shared_ptr<Module> module_, + std::shared_ptr<ProfileManager> profile_manager_, + Core::System& system_, const char* name) + : ServiceFramework{system_, name}, module{std::move(module_)}, profile_manager{std::move( + profile_manager_)} {} Module::Interface::~Interface() = default; diff --git a/src/core/hle/service/acc/acc.h b/src/core/hle/service/acc/acc.h index c611efd89..ab8edc049 100644 --- a/src/core/hle/service/acc/acc.h +++ b/src/core/hle/service/acc/acc.h @@ -15,8 +15,8 @@ class Module final { public: class Interface : public ServiceFramework<Interface> { public: - explicit Interface(std::shared_ptr<Module> module, - std::shared_ptr<ProfileManager> profile_manager, Core::System& system, + explicit Interface(std::shared_ptr<Module> module_, + std::shared_ptr<ProfileManager> profile_manager_, Core::System& system_, const char* name); ~Interface() override; @@ -60,7 +60,6 @@ public: protected: std::shared_ptr<Module> module; std::shared_ptr<ProfileManager> profile_manager; - Core::System& system; }; }; diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index d7a81f64a..c9808060a 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -64,7 +64,7 @@ struct LaunchParameterAccountPreselectedUser { static_assert(sizeof(LaunchParameterAccountPreselectedUser) == 0x88); IWindowController::IWindowController(Core::System& system_) - : ServiceFramework("IWindowController"), system{system_} { + : ServiceFramework{system_, "IWindowController"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CreateWindow"}, @@ -99,7 +99,8 @@ void IWindowController::AcquireForegroundRights(Kernel::HLERequestContext& ctx) rb.Push(RESULT_SUCCESS); } -IAudioController::IAudioController() : ServiceFramework("IAudioController") { +IAudioController::IAudioController(Core::System& system_) + : ServiceFramework{system_, "IAudioController"} { // clang-format off static const FunctionInfo functions[] = { {0, &IAudioController::SetExpectedMasterVolume, "SetExpectedMasterVolume"}, @@ -180,7 +181,8 @@ void IAudioController::SetTransparentAudioRate(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); } -IDisplayController::IDisplayController() : ServiceFramework("IDisplayController") { +IDisplayController::IDisplayController(Core::System& system_) + : ServiceFramework{system_, "IDisplayController"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetLastForegroundCaptureImage"}, @@ -219,7 +221,8 @@ IDisplayController::IDisplayController() : ServiceFramework("IDisplayController" IDisplayController::~IDisplayController() = default; -IDebugFunctions::IDebugFunctions() : ServiceFramework{"IDebugFunctions"} { +IDebugFunctions::IDebugFunctions(Core::System& system_) + : ServiceFramework{system_, "IDebugFunctions"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "NotifyMessageToHomeMenuForDebug"}, @@ -246,9 +249,8 @@ IDebugFunctions::IDebugFunctions() : ServiceFramework{"IDebugFunctions"} { IDebugFunctions::~IDebugFunctions() = default; -ISelfController::ISelfController(Core::System& system, - std::shared_ptr<NVFlinger::NVFlinger> nvflinger) - : ServiceFramework("ISelfController"), system(system), nvflinger(std::move(nvflinger)) { +ISelfController::ISelfController(Core::System& system_, NVFlinger::NVFlinger& nvflinger_) + : ServiceFramework{system_, "ISelfController"}, nvflinger{nvflinger_} { // clang-format off static const FunctionInfo functions[] = { {0, &ISelfController::Exit, "Exit"}, @@ -458,8 +460,8 @@ void ISelfController::CreateManagedDisplayLayer(Kernel::HLERequestContext& ctx) // TODO(Subv): Find out how AM determines the display to use, for now just // create the layer in the Default display. - const auto display_id = nvflinger->OpenDisplay("Default"); - const auto layer_id = nvflinger->CreateLayer(*display_id); + const auto display_id = nvflinger.OpenDisplay("Default"); + const auto layer_id = nvflinger.CreateLayer(*display_id); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); @@ -476,8 +478,8 @@ void ISelfController::CreateManagedDisplaySeparableLayer(Kernel::HLERequestConte // Outputting 1 layer id instead of the expected 2 has not been observed to cause any adverse // side effects. // TODO: Support multiple layers - const auto display_id = nvflinger->OpenDisplay("Default"); - const auto layer_id = nvflinger->CreateLayer(*display_id); + const auto display_id = nvflinger.OpenDisplay("Default"); + const auto layer_id = nvflinger.CreateLayer(*display_id); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); @@ -558,14 +560,14 @@ void ISelfController::GetAccumulatedSuspendedTickChangedEvent(Kernel::HLERequest AppletMessageQueue::AppletMessageQueue(Kernel::KernelCore& kernel) { on_new_message = - Kernel::WritableEvent::CreateEventPair(kernel, "AMMessageQueue:OnMessageRecieved"); + Kernel::WritableEvent::CreateEventPair(kernel, "AMMessageQueue:OnMessageReceived"); on_operation_mode_changed = Kernel::WritableEvent::CreateEventPair(kernel, "AMMessageQueue:OperationModeChanged"); } AppletMessageQueue::~AppletMessageQueue() = default; -const std::shared_ptr<Kernel::ReadableEvent>& AppletMessageQueue::GetMesssageRecieveEvent() const { +const std::shared_ptr<Kernel::ReadableEvent>& AppletMessageQueue::GetMessageReceiveEvent() const { return on_new_message.readable; } @@ -606,9 +608,9 @@ void AppletMessageQueue::RequestExit() { PushMessage(AppletMessage::ExitRequested); } -ICommonStateGetter::ICommonStateGetter(Core::System& system, - std::shared_ptr<AppletMessageQueue> msg_queue) - : ServiceFramework("ICommonStateGetter"), system(system), msg_queue(std::move(msg_queue)) { +ICommonStateGetter::ICommonStateGetter(Core::System& system_, + std::shared_ptr<AppletMessageQueue> msg_queue_) + : ServiceFramework{system_, "ICommonStateGetter"}, msg_queue{std::move(msg_queue_)} { // clang-format off static const FunctionInfo functions[] = { {0, &ICommonStateGetter::GetEventHandle, "GetEventHandle"}, @@ -673,7 +675,7 @@ void ICommonStateGetter::GetEventHandle(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); - rb.PushCopyObjects(msg_queue->GetMesssageRecieveEvent()); + rb.PushCopyObjects(msg_queue->GetMessageReceiveEvent()); } void ICommonStateGetter::ReceiveMessage(Kernel::HLERequestContext& ctx) { @@ -751,7 +753,7 @@ void ICommonStateGetter::GetDefaultDisplayResolution(Kernel::HLERequestContext& IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth) * static_cast<u32>(Settings::values.resolution_factor.GetValue())); rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight) * @@ -796,8 +798,9 @@ private: std::vector<u8> buffer; }; -IStorage::IStorage(std::vector<u8>&& buffer) - : ServiceFramework("IStorage"), impl{std::make_shared<StorageDataImpl>(std::move(buffer))} { +IStorage::IStorage(Core::System& system_, std::vector<u8>&& buffer) + : ServiceFramework{system_, "IStorage"}, impl{std::make_shared<StorageDataImpl>( + std::move(buffer))} { Register(); } @@ -820,11 +823,11 @@ void IStorage::Open(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IStorageAccessor>(*this); + rb.PushIpcInterface<IStorageAccessor>(system, *this); } void ICommonStateGetter::GetOperationMode(Kernel::HLERequestContext& ctx) { - const bool use_docked_mode{Settings::values.use_docked_mode}; + const bool use_docked_mode{Settings::values.use_docked_mode.GetValue()}; LOG_DEBUG(Service_AM, "called, use_docked_mode={}", use_docked_mode); IPC::ResponseBuilder rb{ctx, 3}; @@ -842,8 +845,8 @@ void ICommonStateGetter::GetPerformanceMode(Kernel::HLERequestContext& ctx) { class ILibraryAppletAccessor final : public ServiceFramework<ILibraryAppletAccessor> { public: - explicit ILibraryAppletAccessor(std::shared_ptr<Applets::Applet> applet) - : ServiceFramework("ILibraryAppletAccessor"), applet(std::move(applet)) { + explicit ILibraryAppletAccessor(Core::System& system_, std::shared_ptr<Applets::Applet> applet_) + : ServiceFramework{system_, "ILibraryAppletAccessor"}, applet{std::move(applet_)} { // clang-format off static const FunctionInfo functions[] = { {0, &ILibraryAppletAccessor::GetAppletStateChangedEvent, "GetAppletStateChangedEvent"}, @@ -998,8 +1001,8 @@ private: std::shared_ptr<Applets::Applet> applet; }; -IStorageAccessor::IStorageAccessor(IStorage& storage) - : ServiceFramework("IStorageAccessor"), backing(storage) { +IStorageAccessor::IStorageAccessor(Core::System& system_, IStorage& backing_) + : ServiceFramework{system_, "IStorageAccessor"}, backing{backing_} { // clang-format off static const FunctionInfo functions[] = { {0, &IStorageAccessor::GetSize, "GetSize"}, @@ -1070,7 +1073,7 @@ void IStorageAccessor::Read(Kernel::HLERequestContext& ctx) { } ILibraryAppletCreator::ILibraryAppletCreator(Core::System& system_) - : ServiceFramework("ILibraryAppletCreator"), system{system_} { + : ServiceFramework{system_, "ILibraryAppletCreator"} { static const FunctionInfo functions[] = { {0, &ILibraryAppletCreator::CreateLibraryApplet, "CreateLibraryApplet"}, {1, nullptr, "TerminateAllLibraryApplets"}, @@ -1089,14 +1092,14 @@ void ILibraryAppletCreator::CreateLibraryApplet(Kernel::HLERequestContext& ctx) const auto applet_id = rp.PopRaw<Applets::AppletId>(); const auto applet_mode = rp.PopRaw<u32>(); - LOG_DEBUG(Service_AM, "called with applet_id={:08X}, applet_mode={:08X}", - static_cast<u32>(applet_id), applet_mode); + LOG_DEBUG(Service_AM, "called with applet_id={:08X}, applet_mode={:08X}", applet_id, + applet_mode); const auto& applet_manager{system.GetAppletManager()}; const auto applet = applet_manager.GetApplet(applet_id); if (applet == nullptr) { - LOG_ERROR(Service_AM, "Applet doesn't exist! applet_id={}", static_cast<u32>(applet_id)); + LOG_ERROR(Service_AM, "Applet doesn't exist! applet_id={}", applet_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_UNKNOWN); @@ -1106,7 +1109,7 @@ void ILibraryAppletCreator::CreateLibraryApplet(Kernel::HLERequestContext& ctx) IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<AM::ILibraryAppletAccessor>(applet); + rb.PushIpcInterface<ILibraryAppletAccessor>(system, applet); } void ILibraryAppletCreator::CreateStorage(Kernel::HLERequestContext& ctx) { @@ -1118,7 +1121,7 @@ void ILibraryAppletCreator::CreateStorage(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<AM::IStorage>(std::move(buffer)); + rb.PushIpcInterface<IStorage>(system, std::move(buffer)); } void ILibraryAppletCreator::CreateTransferMemoryStorage(Kernel::HLERequestContext& ctx) { @@ -1145,11 +1148,11 @@ void ILibraryAppletCreator::CreateTransferMemoryStorage(Kernel::HLERequestContex IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IStorage>(std::move(memory)); + rb.PushIpcInterface<IStorage>(system, std::move(memory)); } IApplicationFunctions::IApplicationFunctions(Core::System& system_) - : ServiceFramework("IApplicationFunctions"), system{system_} { + : ServiceFramework{system_, "IApplicationFunctions"} { // clang-format off static const FunctionInfo functions[] = { {1, &IApplicationFunctions::PopLaunchParameter, "PopLaunchParameter"}, @@ -1189,9 +1192,9 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {102, &IApplicationFunctions::SetApplicationCopyrightVisibility, "SetApplicationCopyrightVisibility"}, {110, &IApplicationFunctions::QueryApplicationPlayStatistics, "QueryApplicationPlayStatistics"}, {111, &IApplicationFunctions::QueryApplicationPlayStatisticsByUid, "QueryApplicationPlayStatisticsByUid"}, - {120, nullptr, "ExecuteProgram"}, - {121, nullptr, "ClearUserChannel"}, - {122, nullptr, "UnpopToUserChannel"}, + {120, &IApplicationFunctions::ExecuteProgram, "ExecuteProgram"}, + {121, &IApplicationFunctions::ClearUserChannel, "ClearUserChannel"}, + {122, &IApplicationFunctions::UnpopToUserChannel, "UnpopToUserChannel"}, {123, &IApplicationFunctions::GetPreviousProgramIndex, "GetPreviousProgramIndex"}, {124, nullptr, "EnableApplicationAllThreadDumpOnCrash"}, {130, &IApplicationFunctions::GetGpuErrorDetectedSystemEvent, "GetGpuErrorDetectedSystemEvent"}, @@ -1201,6 +1204,8 @@ IApplicationFunctions::IApplicationFunctions(Core::System& system_) {151, nullptr, "TryPopFromNotificationStorageChannel"}, {160, nullptr, "GetHealthWarningDisappearedSystemEvent"}, {170, nullptr, "SetHdcpAuthenticationActivated"}, + {180, nullptr, "GetLaunchRequiredVersion"}, + {181, nullptr, "UpgradeLaunchRequiredVersion"}, {500, nullptr, "StartContinuousRecordingFlushForDebug"}, {1000, nullptr, "CreateMovieMaker"}, {1001, nullptr, "PrepareForJit"}, @@ -1285,7 +1290,7 @@ void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto kind = rp.PopEnum<LaunchParameterKind>(); - LOG_DEBUG(Service_AM, "called, kind={:08X}", static_cast<u8>(kind)); + LOG_DEBUG(Service_AM, "called, kind={:08X}", kind); if (kind == LaunchParameterKind::ApplicationSpecific && !launch_popped_application_specific) { const auto backend = BCAT::CreateBackendFromSettings(system, [this](u64 tid) { @@ -1299,7 +1304,7 @@ void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) { if (data.has_value()) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IStorage>(std::move(*data)); + rb.PushIpcInterface<IStorage>(system, std::move(*data)); launch_popped_application_specific = true; return; } @@ -1322,7 +1327,7 @@ void IApplicationFunctions::PopLaunchParameter(Kernel::HLERequestContext& ctx) { std::vector<u8> buffer(sizeof(LaunchParameterAccountPreselectedUser)); std::memcpy(buffer.data(), ¶ms, buffer.size()); - rb.PushIpcInterface<IStorage>(std::move(buffer)); + rb.PushIpcInterface<IStorage>(system, std::move(buffer)); launch_popped_account_preselect = true; return; } @@ -1379,13 +1384,16 @@ void IApplicationFunctions::GetDisplayVersion(Kernel::HLERequestContext& ctx) { const auto res = [this] { const auto title_id = system.CurrentProcess()->GetTitleID(); - FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; auto res = pm.GetControlMetadata(); if (res.first != nullptr) { return res; } - FileSys::PatchManager pm_update{FileSys::GetUpdateTitleID(title_id)}; + const FileSys::PatchManager pm_update{FileSys::GetUpdateTitleID(title_id), + system.GetFileSystemController(), + system.GetContentProvider()}; return pm_update.GetControlMetadata(); }(); @@ -1413,13 +1421,16 @@ void IApplicationFunctions::GetDesiredLanguage(Kernel::HLERequestContext& ctx) { const auto res = [this] { const auto title_id = system.CurrentProcess()->GetTitleID(); - FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; auto res = pm.GetControlMetadata(); if (res.first != nullptr) { return res; } - FileSys::PatchManager pm_update{FileSys::GetUpdateTitleID(title_id)}; + const FileSys::PatchManager pm_update{FileSys::GetUpdateTitleID(title_id), + system.GetFileSystemController(), + system.GetContentProvider()}; return pm_update.GetControlMetadata(); }(); @@ -1526,8 +1537,8 @@ void IApplicationFunctions::GetSaveDataSize(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto [type, user_id] = rp.PopRaw<Parameters>(); - LOG_DEBUG(Service_AM, "called with type={:02X}, user_id={:016X}{:016X}", static_cast<u8>(type), - user_id[1], user_id[0]); + LOG_DEBUG(Service_AM, "called with type={:02X}, user_id={:016X}{:016X}", type, user_id[1], + user_id[0]); const auto size = system.GetFileSystemController().ReadSaveDataSize( type, system.CurrentProcess()->GetTitleID(), user_id); @@ -1554,6 +1565,34 @@ void IApplicationFunctions::QueryApplicationPlayStatisticsByUid(Kernel::HLEReque rb.Push<u32>(0); } +void IApplicationFunctions::ExecuteProgram(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + IPC::RequestParser rp{ctx}; + [[maybe_unused]] const auto unk_1 = rp.Pop<u32>(); + [[maybe_unused]] const auto unk_2 = rp.Pop<u32>(); + const auto program_index = rp.Pop<u64>(); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + + system.ExecuteProgram(program_index); +} + +void IApplicationFunctions::ClearUserChannel(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void IApplicationFunctions::UnpopToUserChannel(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + void IApplicationFunctions::GetPreviousProgramIndex(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_AM, "(STUBBED) called"); @@ -1578,22 +1617,22 @@ void IApplicationFunctions::GetFriendInvitationStorageChannelEvent(Kernel::HLERe rb.PushCopyObjects(friend_invitation_storage_channel_event.readable); } -void InstallInterfaces(SM::ServiceManager& service_manager, - std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system) { +void InstallInterfaces(SM::ServiceManager& service_manager, NVFlinger::NVFlinger& nvflinger, + Core::System& system) { auto message_queue = std::make_shared<AppletMessageQueue>(system.Kernel()); // Needed on game boot message_queue->PushMessage(AppletMessageQueue::AppletMessage::FocusStateChanged); std::make_shared<AppletAE>(nvflinger, message_queue, system)->InstallAsService(service_manager); std::make_shared<AppletOE>(nvflinger, message_queue, system)->InstallAsService(service_manager); - std::make_shared<IdleSys>()->InstallAsService(service_manager); - std::make_shared<OMM>()->InstallAsService(service_manager); - std::make_shared<SPSM>()->InstallAsService(service_manager); - std::make_shared<TCAP>()->InstallAsService(service_manager); + std::make_shared<IdleSys>(system)->InstallAsService(service_manager); + std::make_shared<OMM>(system)->InstallAsService(service_manager); + std::make_shared<SPSM>(system)->InstallAsService(service_manager); + std::make_shared<TCAP>(system)->InstallAsService(service_manager); } -IHomeMenuFunctions::IHomeMenuFunctions(Kernel::KernelCore& kernel) - : ServiceFramework("IHomeMenuFunctions"), kernel(kernel) { +IHomeMenuFunctions::IHomeMenuFunctions(Core::System& system_) + : ServiceFramework{system_, "IHomeMenuFunctions"} { // clang-format off static const FunctionInfo functions[] = { {10, &IHomeMenuFunctions::RequestToGetForeground, "RequestToGetForeground"}, @@ -1612,7 +1651,7 @@ IHomeMenuFunctions::IHomeMenuFunctions(Kernel::KernelCore& kernel) RegisterHandlers(functions); pop_from_general_channel_event = Kernel::WritableEvent::CreateEventPair( - kernel, "IHomeMenuFunctions:PopFromGeneralChannelEvent"); + system.Kernel(), "IHomeMenuFunctions:PopFromGeneralChannelEvent"); } IHomeMenuFunctions::~IHomeMenuFunctions() = default; @@ -1632,7 +1671,8 @@ void IHomeMenuFunctions::GetPopFromGeneralChannelEvent(Kernel::HLERequestContext rb.PushCopyObjects(pop_from_general_channel_event.readable); } -IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStateController") { +IGlobalStateController::IGlobalStateController(Core::System& system_) + : ServiceFramework{system_, "IGlobalStateController"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestToEnterSleep"}, @@ -1655,7 +1695,8 @@ IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStat IGlobalStateController::~IGlobalStateController() = default; -IApplicationCreator::IApplicationCreator() : ServiceFramework("IApplicationCreator") { +IApplicationCreator::IApplicationCreator(Core::System& system_) + : ServiceFramework{system_, "IApplicationCreator"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CreateApplication"}, @@ -1670,8 +1711,8 @@ IApplicationCreator::IApplicationCreator() : ServiceFramework("IApplicationCreat IApplicationCreator::~IApplicationCreator() = default; -IProcessWindingController::IProcessWindingController() - : ServiceFramework("IProcessWindingController") { +IProcessWindingController::IProcessWindingController(Core::System& system_) + : ServiceFramework{system_, "IProcessWindingController"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetLaunchReason"}, diff --git a/src/core/hle/service/am/am.h b/src/core/hle/service/am/am.h index bcc06affe..f51aca1af 100644 --- a/src/core/hle/service/am/am.h +++ b/src/core/hle/service/am/am.h @@ -55,7 +55,7 @@ public: explicit AppletMessageQueue(Kernel::KernelCore& kernel); ~AppletMessageQueue(); - const std::shared_ptr<Kernel::ReadableEvent>& GetMesssageRecieveEvent() const; + const std::shared_ptr<Kernel::ReadableEvent>& GetMessageReceiveEvent() const; const std::shared_ptr<Kernel::ReadableEvent>& GetOperationModeChangedEvent() const; void PushMessage(AppletMessage msg); AppletMessage PopMessage(); @@ -77,13 +77,11 @@ public: private: void GetAppletResourceUserId(Kernel::HLERequestContext& ctx); void AcquireForegroundRights(Kernel::HLERequestContext& ctx); - - Core::System& system; }; class IAudioController final : public ServiceFramework<IAudioController> { public: - IAudioController(); + explicit IAudioController(Core::System& system_); ~IAudioController() override; private: @@ -109,20 +107,19 @@ private: class IDisplayController final : public ServiceFramework<IDisplayController> { public: - IDisplayController(); + explicit IDisplayController(Core::System& system_); ~IDisplayController() override; }; class IDebugFunctions final : public ServiceFramework<IDebugFunctions> { public: - IDebugFunctions(); + explicit IDebugFunctions(Core::System& system_); ~IDebugFunctions() override; }; class ISelfController final : public ServiceFramework<ISelfController> { public: - explicit ISelfController(Core::System& system_, - std::shared_ptr<NVFlinger::NVFlinger> nvflinger_); + explicit ISelfController(Core::System& system_, NVFlinger::NVFlinger& nvflinger_); ~ISelfController() override; private: @@ -155,8 +152,7 @@ private: Disable = 2, }; - Core::System& system; - std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + NVFlinger::NVFlinger& nvflinger; Kernel::EventPair launchable_event; Kernel::EventPair accumulated_suspended_tick_changed_event; @@ -168,8 +164,8 @@ private: class ICommonStateGetter final : public ServiceFramework<ICommonStateGetter> { public: - explicit ICommonStateGetter(Core::System& system, - std::shared_ptr<AppletMessageQueue> msg_queue); + explicit ICommonStateGetter(Core::System& system_, + std::shared_ptr<AppletMessageQueue> msg_queue_); ~ICommonStateGetter() override; private: @@ -197,7 +193,6 @@ private: void GetDefaultDisplayResolution(Kernel::HLERequestContext& ctx); void SetCpuBoostMode(Kernel::HLERequestContext& ctx); - Core::System& system; std::shared_ptr<AppletMessageQueue> msg_queue; bool vr_mode_state{}; }; @@ -212,7 +207,7 @@ public: class IStorage final : public ServiceFramework<IStorage> { public: - explicit IStorage(std::vector<u8>&& buffer); + explicit IStorage(Core::System& system_, std::vector<u8>&& buffer); ~IStorage() override; std::vector<u8>& GetData() { @@ -236,7 +231,7 @@ private: class IStorageAccessor final : public ServiceFramework<IStorageAccessor> { public: - explicit IStorageAccessor(IStorage& backing); + explicit IStorageAccessor(Core::System& system_, IStorage& backing_); ~IStorageAccessor() override; private: @@ -256,8 +251,6 @@ private: void CreateLibraryApplet(Kernel::HLERequestContext& ctx); void CreateStorage(Kernel::HLERequestContext& ctx); void CreateTransferMemoryStorage(Kernel::HLERequestContext& ctx); - - Core::System& system; }; class IApplicationFunctions final : public ServiceFramework<IApplicationFunctions> { @@ -288,6 +281,9 @@ private: void SetApplicationCopyrightVisibility(Kernel::HLERequestContext& ctx); void QueryApplicationPlayStatistics(Kernel::HLERequestContext& ctx); void QueryApplicationPlayStatisticsByUid(Kernel::HLERequestContext& ctx); + void ExecuteProgram(Kernel::HLERequestContext& ctx); + void ClearUserChannel(Kernel::HLERequestContext& ctx); + void UnpopToUserChannel(Kernel::HLERequestContext& ctx); void GetPreviousProgramIndex(Kernel::HLERequestContext& ctx); void GetGpuErrorDetectedSystemEvent(Kernel::HLERequestContext& ctx); void GetFriendInvitationStorageChannelEvent(Kernel::HLERequestContext& ctx); @@ -297,12 +293,11 @@ private: s32 previous_program_index{-1}; Kernel::EventPair gpu_error_detected_event; Kernel::EventPair friend_invitation_storage_channel_event; - Core::System& system; }; class IHomeMenuFunctions final : public ServiceFramework<IHomeMenuFunctions> { public: - explicit IHomeMenuFunctions(Kernel::KernelCore& kernel); + explicit IHomeMenuFunctions(Core::System& system_); ~IHomeMenuFunctions() override; private: @@ -310,29 +305,28 @@ private: void GetPopFromGeneralChannelEvent(Kernel::HLERequestContext& ctx); Kernel::EventPair pop_from_general_channel_event; - Kernel::KernelCore& kernel; }; class IGlobalStateController final : public ServiceFramework<IGlobalStateController> { public: - IGlobalStateController(); + explicit IGlobalStateController(Core::System& system_); ~IGlobalStateController() override; }; class IApplicationCreator final : public ServiceFramework<IApplicationCreator> { public: - IApplicationCreator(); + explicit IApplicationCreator(Core::System& system_); ~IApplicationCreator() override; }; class IProcessWindingController final : public ServiceFramework<IProcessWindingController> { public: - IProcessWindingController(); + explicit IProcessWindingController(Core::System& system_); ~IProcessWindingController() override; }; /// Registers all AM services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager, - std::shared_ptr<NVFlinger::NVFlinger> nvflinger, Core::System& system); +void InstallInterfaces(SM::ServiceManager& service_manager, NVFlinger::NVFlinger& nvflinger, + Core::System& system); } // namespace Service::AM diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp index 9df286d17..5421e0da0 100644 --- a/src/core/hle/service/am/applet_ae.cpp +++ b/src/core/hle/service/am/applet_ae.cpp @@ -3,8 +3,8 @@ // Refer to the license.txt file included. #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" -#include "core/hle/kernel/process.h" #include "core/hle/service/am/am.h" #include "core/hle/service/am/applet_ae.h" #include "core/hle/service/nvflinger/nvflinger.h" @@ -13,11 +13,11 @@ namespace Service::AM { class ILibraryAppletProxy final : public ServiceFramework<ILibraryAppletProxy> { public: - explicit ILibraryAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, - std::shared_ptr<AppletMessageQueue> msg_queue, - Core::System& system) - : ServiceFramework("ILibraryAppletProxy"), nvflinger(std::move(nvflinger)), - msg_queue(std::move(msg_queue)), system(system) { + explicit ILibraryAppletProxy(NVFlinger::NVFlinger& nvflinger_, + std::shared_ptr<AppletMessageQueue> msg_queue_, + Core::System& system_) + : ServiceFramework{system_, "ILibraryAppletProxy"}, nvflinger{nvflinger_}, + msg_queue{std::move(msg_queue_)} { // clang-format off static const FunctionInfo functions[] = { {0, &ILibraryAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"}, @@ -66,7 +66,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IAudioController>(); + rb.PushIpcInterface<IAudioController>(system); } void GetDisplayController(Kernel::HLERequestContext& ctx) { @@ -74,7 +74,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDisplayController>(); + rb.PushIpcInterface<IDisplayController>(system); } void GetProcessWindingController(Kernel::HLERequestContext& ctx) { @@ -82,7 +82,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IProcessWindingController>(); + rb.PushIpcInterface<IProcessWindingController>(system); } void GetDebugFunctions(Kernel::HLERequestContext& ctx) { @@ -90,7 +90,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDebugFunctions>(); + rb.PushIpcInterface<IDebugFunctions>(system); } void GetLibraryAppletCreator(Kernel::HLERequestContext& ctx) { @@ -109,17 +109,17 @@ private: rb.PushIpcInterface<IApplicationFunctions>(system); } - std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + NVFlinger::NVFlinger& nvflinger; std::shared_ptr<AppletMessageQueue> msg_queue; - Core::System& system; }; class ISystemAppletProxy final : public ServiceFramework<ISystemAppletProxy> { public: - explicit ISystemAppletProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, - std::shared_ptr<AppletMessageQueue> msg_queue, Core::System& system) - : ServiceFramework("ISystemAppletProxy"), nvflinger(std::move(nvflinger)), - msg_queue(std::move(msg_queue)), system(system) { + explicit ISystemAppletProxy(NVFlinger::NVFlinger& nvflinger_, + std::shared_ptr<AppletMessageQueue> msg_queue_, + Core::System& system_) + : ServiceFramework{system_, "ISystemAppletProxy"}, nvflinger{nvflinger_}, + msg_queue{std::move(msg_queue_)} { // clang-format off static const FunctionInfo functions[] = { {0, &ISystemAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"}, @@ -170,7 +170,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IAudioController>(); + rb.PushIpcInterface<IAudioController>(system); } void GetDisplayController(Kernel::HLERequestContext& ctx) { @@ -178,7 +178,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDisplayController>(); + rb.PushIpcInterface<IDisplayController>(system); } void GetDebugFunctions(Kernel::HLERequestContext& ctx) { @@ -186,7 +186,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDebugFunctions>(); + rb.PushIpcInterface<IDebugFunctions>(system); } void GetLibraryAppletCreator(Kernel::HLERequestContext& ctx) { @@ -202,7 +202,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IHomeMenuFunctions>(system.Kernel()); + rb.PushIpcInterface<IHomeMenuFunctions>(system); } void GetGlobalStateController(Kernel::HLERequestContext& ctx) { @@ -210,7 +210,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IGlobalStateController>(); + rb.PushIpcInterface<IGlobalStateController>(system); } void GetApplicationCreator(Kernel::HLERequestContext& ctx) { @@ -218,11 +218,11 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IApplicationCreator>(); + rb.PushIpcInterface<IApplicationCreator>(system); } - std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + + NVFlinger::NVFlinger& nvflinger; std::shared_ptr<AppletMessageQueue> msg_queue; - Core::System& system; }; void AppletAE::OpenSystemAppletProxy(Kernel::HLERequestContext& ctx) { @@ -249,10 +249,10 @@ void AppletAE::OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface<ILibraryAppletProxy>(nvflinger, msg_queue, system); } -AppletAE::AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, - std::shared_ptr<AppletMessageQueue> msg_queue, Core::System& system) - : ServiceFramework("appletAE"), nvflinger(std::move(nvflinger)), - msg_queue(std::move(msg_queue)), system(system) { +AppletAE::AppletAE(NVFlinger::NVFlinger& nvflinger_, std::shared_ptr<AppletMessageQueue> msg_queue_, + Core::System& system_) + : ServiceFramework{system_, "appletAE"}, nvflinger{nvflinger_}, msg_queue{ + std::move(msg_queue_)} { // clang-format off static const FunctionInfo functions[] = { {100, &AppletAE::OpenSystemAppletProxy, "OpenSystemAppletProxy"}, diff --git a/src/core/hle/service/am/applet_ae.h b/src/core/hle/service/am/applet_ae.h index 2e3e45915..adb207349 100644 --- a/src/core/hle/service/am/applet_ae.h +++ b/src/core/hle/service/am/applet_ae.h @@ -23,8 +23,8 @@ class AppletMessageQueue; class AppletAE final : public ServiceFramework<AppletAE> { public: - explicit AppletAE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, - std::shared_ptr<AppletMessageQueue> msg_queue, Core::System& system); + explicit AppletAE(NVFlinger::NVFlinger& nvflinger_, + std::shared_ptr<AppletMessageQueue> msg_queue_, Core::System& system_); ~AppletAE() override; const std::shared_ptr<AppletMessageQueue>& GetMessageQueue() const; @@ -34,9 +34,8 @@ private: void OpenLibraryAppletProxy(Kernel::HLERequestContext& ctx); void OpenLibraryAppletProxyOld(Kernel::HLERequestContext& ctx); - std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + NVFlinger::NVFlinger& nvflinger; std::shared_ptr<AppletMessageQueue> msg_queue; - Core::System& system; }; } // namespace AM diff --git a/src/core/hle/service/am/applet_oe.cpp b/src/core/hle/service/am/applet_oe.cpp index a2ffaa440..f9eba8f52 100644 --- a/src/core/hle/service/am/applet_oe.cpp +++ b/src/core/hle/service/am/applet_oe.cpp @@ -12,10 +12,11 @@ namespace Service::AM { class IApplicationProxy final : public ServiceFramework<IApplicationProxy> { public: - explicit IApplicationProxy(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, - std::shared_ptr<AppletMessageQueue> msg_queue, Core::System& system) - : ServiceFramework("IApplicationProxy"), nvflinger(std::move(nvflinger)), - msg_queue(std::move(msg_queue)), system(system) { + explicit IApplicationProxy(NVFlinger::NVFlinger& nvflinger_, + std::shared_ptr<AppletMessageQueue> msg_queue_, + Core::System& system_) + : ServiceFramework{system_, "IApplicationProxy"}, nvflinger{nvflinger_}, + msg_queue{std::move(msg_queue_)} { // clang-format off static const FunctionInfo functions[] = { {0, &IApplicationProxy::GetCommonStateGetter, "GetCommonStateGetter"}, @@ -39,7 +40,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IAudioController>(); + rb.PushIpcInterface<IAudioController>(system); } void GetDisplayController(Kernel::HLERequestContext& ctx) { @@ -47,7 +48,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDisplayController>(); + rb.PushIpcInterface<IDisplayController>(system); } void GetDebugFunctions(Kernel::HLERequestContext& ctx) { @@ -55,7 +56,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDebugFunctions>(); + rb.PushIpcInterface<IDebugFunctions>(system); } void GetWindowController(Kernel::HLERequestContext& ctx) { @@ -98,9 +99,8 @@ private: rb.PushIpcInterface<IApplicationFunctions>(system); } - std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + NVFlinger::NVFlinger& nvflinger; std::shared_ptr<AppletMessageQueue> msg_queue; - Core::System& system; }; void AppletOE::OpenApplicationProxy(Kernel::HLERequestContext& ctx) { @@ -111,10 +111,10 @@ void AppletOE::OpenApplicationProxy(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface<IApplicationProxy>(nvflinger, msg_queue, system); } -AppletOE::AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, - std::shared_ptr<AppletMessageQueue> msg_queue, Core::System& system) - : ServiceFramework("appletOE"), nvflinger(std::move(nvflinger)), - msg_queue(std::move(msg_queue)), system(system) { +AppletOE::AppletOE(NVFlinger::NVFlinger& nvflinger_, std::shared_ptr<AppletMessageQueue> msg_queue_, + Core::System& system_) + : ServiceFramework{system_, "appletOE"}, nvflinger{nvflinger_}, msg_queue{ + std::move(msg_queue_)} { static const FunctionInfo functions[] = { {0, &AppletOE::OpenApplicationProxy, "OpenApplicationProxy"}, }; diff --git a/src/core/hle/service/am/applet_oe.h b/src/core/hle/service/am/applet_oe.h index 758da792d..6c1aa255a 100644 --- a/src/core/hle/service/am/applet_oe.h +++ b/src/core/hle/service/am/applet_oe.h @@ -23,8 +23,8 @@ class AppletMessageQueue; class AppletOE final : public ServiceFramework<AppletOE> { public: - explicit AppletOE(std::shared_ptr<NVFlinger::NVFlinger> nvflinger, - std::shared_ptr<AppletMessageQueue> msg_queue, Core::System& system); + explicit AppletOE(NVFlinger::NVFlinger& nvflinger_, + std::shared_ptr<AppletMessageQueue> msg_queue_, Core::System& system_); ~AppletOE() override; const std::shared_ptr<AppletMessageQueue>& GetMessageQueue() const; @@ -32,9 +32,8 @@ public: private: void OpenApplicationProxy(Kernel::HLERequestContext& ctx); - std::shared_ptr<NVFlinger::NVFlinger> nvflinger; + NVFlinger::NVFlinger& nvflinger; std::shared_ptr<AppletMessageQueue> msg_queue; - Core::System& system; }; } // namespace AM diff --git a/src/core/hle/service/am/applets/applets.cpp b/src/core/hle/service/am/applets/applets.cpp index 4e0800f9a..08676c3fc 100644 --- a/src/core/hle/service/am/applets/applets.cpp +++ b/src/core/hle/service/am/applets/applets.cpp @@ -142,14 +142,14 @@ void Applet::Initialize() { AppletFrontendSet::AppletFrontendSet() = default; -AppletFrontendSet::AppletFrontendSet(ControllerApplet controller, ECommerceApplet e_commerce, - ErrorApplet error, ParentalControlsApplet parental_controls, - PhotoViewer photo_viewer, ProfileSelect profile_select, - SoftwareKeyboard software_keyboard, WebBrowser web_browser) - : controller{std::move(controller)}, e_commerce{std::move(e_commerce)}, error{std::move(error)}, - parental_controls{std::move(parental_controls)}, photo_viewer{std::move(photo_viewer)}, - profile_select{std::move(profile_select)}, software_keyboard{std::move(software_keyboard)}, - web_browser{std::move(web_browser)} {} +AppletFrontendSet::AppletFrontendSet(ControllerApplet controller_applet, ErrorApplet error_applet, + ParentalControlsApplet parental_controls_applet, + PhotoViewer photo_viewer_, ProfileSelect profile_select_, + SoftwareKeyboard software_keyboard_, WebBrowser web_browser_) + : controller{std::move(controller_applet)}, error{std::move(error_applet)}, + parental_controls{std::move(parental_controls_applet)}, + photo_viewer{std::move(photo_viewer_)}, profile_select{std::move(profile_select_)}, + software_keyboard{std::move(software_keyboard_)}, web_browser{std::move(web_browser_)} {} AppletFrontendSet::~AppletFrontendSet() = default; @@ -170,10 +170,6 @@ void AppletManager::SetAppletFrontendSet(AppletFrontendSet set) { frontend.controller = std::move(set.controller); } - if (set.e_commerce != nullptr) { - frontend.e_commerce = std::move(set.e_commerce); - } - if (set.error != nullptr) { frontend.error = std::move(set.error); } @@ -206,11 +202,8 @@ void AppletManager::SetDefaultAppletFrontendSet() { void AppletManager::SetDefaultAppletsIfMissing() { if (frontend.controller == nullptr) { - frontend.controller = std::make_unique<Core::Frontend::DefaultControllerApplet>(); - } - - if (frontend.e_commerce == nullptr) { - frontend.e_commerce = std::make_unique<Core::Frontend::DefaultECommerceApplet>(); + frontend.controller = + std::make_unique<Core::Frontend::DefaultControllerApplet>(system.ServiceManager()); } if (frontend.error == nullptr) { @@ -256,13 +249,14 @@ std::shared_ptr<Applet> AppletManager::GetApplet(AppletId id) const { return std::make_shared<ProfileSelect>(system, *frontend.profile_select); case AppletId::SoftwareKeyboard: return std::make_shared<SoftwareKeyboard>(system, *frontend.software_keyboard); + case AppletId::Web: + case AppletId::Shop: + case AppletId::OfflineWeb: + case AppletId::LoginShare: + case AppletId::WebAuth: + return std::make_shared<WebBrowser>(system, *frontend.web_browser); case AppletId::PhotoViewer: return std::make_shared<PhotoViewer>(system, *frontend.photo_viewer); - case AppletId::LibAppletShop: - return std::make_shared<WebBrowser>(system, *frontend.web_browser, - frontend.e_commerce.get()); - case AppletId::LibAppletOff: - return std::make_shared<WebBrowser>(system, *frontend.web_browser); default: UNIMPLEMENTED_MSG( "No backend implementation exists for applet_id={:02X}! Falling back to stub applet.", diff --git a/src/core/hle/service/am/applets/applets.h b/src/core/hle/service/am/applets/applets.h index a1f4cf897..4fd792c05 100644 --- a/src/core/hle/service/am/applets/applets.h +++ b/src/core/hle/service/am/applets/applets.h @@ -50,13 +50,13 @@ enum class AppletId : u32 { ProfileSelect = 0x10, SoftwareKeyboard = 0x11, MiiEdit = 0x12, - LibAppletWeb = 0x13, - LibAppletShop = 0x14, + Web = 0x13, + Shop = 0x14, PhotoViewer = 0x15, Settings = 0x16, - LibAppletOff = 0x17, - LibAppletWhitelisted = 0x18, - LibAppletAuth = 0x19, + OfflineWeb = 0x17, + LoginShare = 0x18, + WebAuth = 0x19, MyPage = 0x1A, }; @@ -157,7 +157,6 @@ protected: struct AppletFrontendSet { using ControllerApplet = std::unique_ptr<Core::Frontend::ControllerApplet>; - using ECommerceApplet = std::unique_ptr<Core::Frontend::ECommerceApplet>; using ErrorApplet = std::unique_ptr<Core::Frontend::ErrorApplet>; using ParentalControlsApplet = std::unique_ptr<Core::Frontend::ParentalControlsApplet>; using PhotoViewer = std::unique_ptr<Core::Frontend::PhotoViewerApplet>; @@ -166,10 +165,10 @@ struct AppletFrontendSet { using WebBrowser = std::unique_ptr<Core::Frontend::WebBrowserApplet>; AppletFrontendSet(); - AppletFrontendSet(ControllerApplet controller, ECommerceApplet e_commerce, ErrorApplet error, - ParentalControlsApplet parental_controls, PhotoViewer photo_viewer, - ProfileSelect profile_select, SoftwareKeyboard software_keyboard, - WebBrowser web_browser); + AppletFrontendSet(ControllerApplet controller_applet, ErrorApplet error_applet, + ParentalControlsApplet parental_controls_applet, PhotoViewer photo_viewer_, + ProfileSelect profile_select_, SoftwareKeyboard software_keyboard_, + WebBrowser web_browser_); ~AppletFrontendSet(); AppletFrontendSet(const AppletFrontendSet&) = delete; @@ -179,7 +178,6 @@ struct AppletFrontendSet { AppletFrontendSet& operator=(AppletFrontendSet&&) noexcept; ControllerApplet controller; - ECommerceApplet e_commerce; ErrorApplet error; ParentalControlsApplet parental_controls; PhotoViewer photo_viewer; diff --git a/src/core/hle/service/am/applets/controller.cpp b/src/core/hle/service/am/applets/controller.cpp index 2151da783..7edfca64e 100644 --- a/src/core/hle/service/am/applets/controller.cpp +++ b/src/core/hle/service/am/applets/controller.cpp @@ -25,18 +25,18 @@ namespace Service::AM::Applets { static Core::Frontend::ControllerParameters ConvertToFrontendParameters( ControllerSupportArgPrivate private_arg, ControllerSupportArgHeader header, bool enable_text, std::vector<IdentificationColor> identification_colors, std::vector<ExplainText> text) { - HID::Controller_NPad::NPadType npad_style_set; + HID::Controller_NPad::NpadStyleSet npad_style_set; npad_style_set.raw = private_arg.style_set; return { - .min_players = std::max(s8(1), header.player_count_min), + .min_players = std::max(s8{1}, header.player_count_min), .max_players = header.player_count_max, .keep_controllers_connected = header.enable_take_over_connection, .enable_single_mode = header.enable_single_mode, .enable_border_color = header.enable_identification_color, - .border_colors = identification_colors, + .border_colors = std::move(identification_colors), .enable_explain_text = enable_text, - .explain_text = text, + .explain_text = std::move(text), .allow_pro_controller = npad_style_set.pro_controller == 1, .allow_handheld = npad_style_set.handheld == 1, .allow_dual_joycons = npad_style_set.joycon_dual == 1, @@ -46,7 +46,7 @@ static Core::Frontend::ControllerParameters ConvertToFrontendParameters( } Controller::Controller(Core::System& system_, const Core::Frontend::ControllerApplet& frontend_) - : Applet{system_.Kernel()}, frontend(frontend_) {} + : Applet{system_.Kernel()}, frontend{frontend_}, system{system_} {} Controller::~Controller() = default; @@ -62,7 +62,7 @@ void Controller::Initialize() { common_args.play_startup_sound, common_args.size, common_args.system_tick, common_args.theme_color); - library_applet_version = LibraryAppletVersion{common_args.library_version}; + controller_applet_version = ControllerAppletVersion{common_args.library_version}; const auto private_arg_storage = broker.PopNormalDataToApplet(); ASSERT(private_arg_storage != nullptr); @@ -70,39 +70,78 @@ void Controller::Initialize() { const auto& private_arg = private_arg_storage->GetData(); ASSERT(private_arg.size() == sizeof(ControllerSupportArgPrivate)); - std::memcpy(&controller_private_arg, private_arg.data(), sizeof(ControllerSupportArgPrivate)); + std::memcpy(&controller_private_arg, private_arg.data(), private_arg.size()); ASSERT_MSG(controller_private_arg.arg_private_size == sizeof(ControllerSupportArgPrivate), "Unknown ControllerSupportArgPrivate revision={} with size={}", - library_applet_version, controller_private_arg.arg_private_size); + controller_applet_version, controller_private_arg.arg_private_size); + + // Some games such as Cave Story+ set invalid values for the ControllerSupportMode. + // Defer to arg_size to set the ControllerSupportMode. + if (controller_private_arg.mode >= ControllerSupportMode::MaxControllerSupportMode) { + switch (controller_private_arg.arg_size) { + case sizeof(ControllerSupportArgOld): + case sizeof(ControllerSupportArgNew): + controller_private_arg.mode = ControllerSupportMode::ShowControllerSupport; + break; + case sizeof(ControllerUpdateFirmwareArg): + controller_private_arg.mode = ControllerSupportMode::ShowControllerFirmwareUpdate; + break; + default: + UNIMPLEMENTED_MSG("Unknown ControllerPrivateArg mode={} with arg_size={}", + controller_private_arg.mode, controller_private_arg.arg_size); + controller_private_arg.mode = ControllerSupportMode::ShowControllerSupport; + break; + } + } + + // Some games such as Cave Story+ set invalid values for the ControllerSupportCaller. + // This is always 0 (Application) except with ShowControllerFirmwareUpdateForSystem. + if (controller_private_arg.caller >= ControllerSupportCaller::MaxControllerSupportCaller) { + if (controller_private_arg.flag_1 && + controller_private_arg.mode == ControllerSupportMode::ShowControllerFirmwareUpdate) { + controller_private_arg.caller = ControllerSupportCaller::System; + } else { + controller_private_arg.caller = ControllerSupportCaller::Application; + } + } switch (controller_private_arg.mode) { - case ControllerSupportMode::ShowControllerSupport: { + case ControllerSupportMode::ShowControllerSupport: + case ControllerSupportMode::ShowControllerStrapGuide: { const auto user_arg_storage = broker.PopNormalDataToApplet(); ASSERT(user_arg_storage != nullptr); const auto& user_arg = user_arg_storage->GetData(); - switch (library_applet_version) { - case LibraryAppletVersion::Version3: - case LibraryAppletVersion::Version4: - case LibraryAppletVersion::Version5: + switch (controller_applet_version) { + case ControllerAppletVersion::Version3: + case ControllerAppletVersion::Version4: + case ControllerAppletVersion::Version5: ASSERT(user_arg.size() == sizeof(ControllerSupportArgOld)); - std::memcpy(&controller_user_arg_old, user_arg.data(), sizeof(ControllerSupportArgOld)); + std::memcpy(&controller_user_arg_old, user_arg.data(), user_arg.size()); break; - case LibraryAppletVersion::Version7: + case ControllerAppletVersion::Version7: ASSERT(user_arg.size() == sizeof(ControllerSupportArgNew)); - std::memcpy(&controller_user_arg_new, user_arg.data(), sizeof(ControllerSupportArgNew)); + std::memcpy(&controller_user_arg_new, user_arg.data(), user_arg.size()); break; default: UNIMPLEMENTED_MSG("Unknown ControllerSupportArg revision={} with size={}", - library_applet_version, controller_private_arg.arg_size); + controller_applet_version, controller_private_arg.arg_size); ASSERT(user_arg.size() >= sizeof(ControllerSupportArgNew)); std::memcpy(&controller_user_arg_new, user_arg.data(), sizeof(ControllerSupportArgNew)); break; } break; } - case ControllerSupportMode::ShowControllerStrapGuide: - case ControllerSupportMode::ShowControllerFirmwareUpdate: + case ControllerSupportMode::ShowControllerFirmwareUpdate: { + const auto update_arg_storage = broker.PopNormalDataToApplet(); + ASSERT(update_arg_storage != nullptr); + + const auto& update_arg = update_arg_storage->GetData(); + ASSERT(update_arg.size() == sizeof(ControllerUpdateFirmwareArg)); + + std::memcpy(&controller_update_arg, update_arg.data(), update_arg.size()); + break; + } default: { UNIMPLEMENTED_MSG("Unimplemented ControllerSupportMode={}", controller_private_arg.mode); break; @@ -126,10 +165,10 @@ void Controller::Execute() { switch (controller_private_arg.mode) { case ControllerSupportMode::ShowControllerSupport: { const auto parameters = [this] { - switch (library_applet_version) { - case LibraryAppletVersion::Version3: - case LibraryAppletVersion::Version4: - case LibraryAppletVersion::Version5: + switch (controller_applet_version) { + case ControllerAppletVersion::Version3: + case ControllerAppletVersion::Version4: + case ControllerAppletVersion::Version5: return ConvertToFrontendParameters( controller_private_arg, controller_user_arg_old.header, controller_user_arg_old.enable_explain_text, @@ -138,7 +177,7 @@ void Controller::Execute() { controller_user_arg_old.identification_colors.end()), std::vector<ExplainText>(controller_user_arg_old.explain_text.begin(), controller_user_arg_old.explain_text.end())); - case LibraryAppletVersion::Version7: + case ControllerAppletVersion::Version7: default: return ConvertToFrontendParameters( controller_private_arg, controller_user_arg_new.header, @@ -170,6 +209,9 @@ void Controller::Execute() { } case ControllerSupportMode::ShowControllerStrapGuide: case ControllerSupportMode::ShowControllerFirmwareUpdate: + UNIMPLEMENTED_MSG("ControllerSupportMode={} is not implemented", + controller_private_arg.mode); + [[fallthrough]]; default: { ConfigurationComplete(); break; @@ -180,20 +222,19 @@ void Controller::Execute() { void Controller::ConfigurationComplete() { ControllerSupportResultInfo result_info{}; - const auto& players = Settings::values.players; + const auto& players = Settings::values.players.GetValue(); // If enable_single_mode is enabled, player_count is 1 regardless of any other parameters. // Otherwise, only count connected players from P1-P8. result_info.player_count = - is_single_mode ? 1 - : static_cast<s8>(std::count_if( - players.begin(), players.end() - 2, - [](Settings::PlayerInput player) { return player.connected; })); + is_single_mode + ? 1 + : static_cast<s8>(std::count_if(players.begin(), players.end() - 2, + [](const auto& player) { return player.connected; })); - result_info.selected_id = HID::Controller_NPad::IndexToNPad( - std::distance(players.begin(), - std::find_if(players.begin(), players.end(), - [](Settings::PlayerInput player) { return player.connected; }))); + result_info.selected_id = HID::Controller_NPad::IndexToNPad(std::distance( + players.begin(), std::find_if(players.begin(), players.end(), + [](const auto& player) { return player.connected; }))); result_info.result = 0; @@ -203,7 +244,7 @@ void Controller::ConfigurationComplete() { complete = true; out_data = std::vector<u8>(sizeof(ControllerSupportResultInfo)); std::memcpy(out_data.data(), &result_info, out_data.size()); - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(out_data))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::move(out_data))); broker.SignalStateChanged(); } diff --git a/src/core/hle/service/am/applets/controller.h b/src/core/hle/service/am/applets/controller.h index f7bb3fba9..d4c9da7b1 100644 --- a/src/core/hle/service/am/applets/controller.h +++ b/src/core/hle/service/am/applets/controller.h @@ -21,7 +21,7 @@ namespace Service::AM::Applets { using IdentificationColor = std::array<u8, 4>; using ExplainText = std::array<char, 0x81>; -enum class LibraryAppletVersion : u32_le { +enum class ControllerAppletVersion : u32_le { Version3 = 0x3, // 1.0.0 - 2.3.0 Version4 = 0x4, // 3.0.0 - 5.1.0 Version5 = 0x5, // 6.0.0 - 7.0.1 @@ -29,14 +29,18 @@ enum class LibraryAppletVersion : u32_le { }; enum class ControllerSupportMode : u8 { - ShowControllerSupport = 0, - ShowControllerStrapGuide = 1, - ShowControllerFirmwareUpdate = 2, + ShowControllerSupport, + ShowControllerStrapGuide, + ShowControllerFirmwareUpdate, + + MaxControllerSupportMode, }; enum class ControllerSupportCaller : u8 { - Application = 0, - System = 1, + Application, + System, + + MaxControllerSupportCaller, }; struct ControllerSupportArgPrivate { @@ -84,6 +88,13 @@ struct ControllerSupportArgNew { static_assert(sizeof(ControllerSupportArgNew) == 0x430, "ControllerSupportArgNew has incorrect size."); +struct ControllerUpdateFirmwareArg { + bool enable_force_update{}; + INSERT_PADDING_BYTES(3); +}; +static_assert(sizeof(ControllerUpdateFirmwareArg) == 0x4, + "ControllerUpdateFirmwareArg has incorrect size."); + struct ControllerSupportResultInfo { s8 player_count{}; INSERT_PADDING_BYTES(3); @@ -109,11 +120,13 @@ public: private: const Core::Frontend::ControllerApplet& frontend; + Core::System& system; - LibraryAppletVersion library_applet_version; + ControllerAppletVersion controller_applet_version; ControllerSupportArgPrivate controller_private_arg; ControllerSupportArgOld controller_user_arg_old; ControllerSupportArgNew controller_user_arg_new; + ControllerUpdateFirmwareArg controller_update_arg; bool complete{false}; ResultCode status{RESULT_SUCCESS}; bool is_single_mode{false}; diff --git a/src/core/hle/service/am/applets/error.cpp b/src/core/hle/service/am/applets/error.cpp index f12fd7f89..d85505082 100644 --- a/src/core/hle/service/am/applets/error.cpp +++ b/src/core/hle/service/am/applets/error.cpp @@ -87,7 +87,7 @@ ResultCode Decode64BitError(u64 error) { } // Anonymous namespace Error::Error(Core::System& system_, const Core::Frontend::ErrorApplet& frontend_) - : Applet{system_.Kernel()}, frontend(frontend_), system{system_} {} + : Applet{system_.Kernel()}, frontend{frontend_}, system{system_} {} Error::~Error() = default; @@ -125,7 +125,7 @@ void Error::Initialize() { error_code = Decode64BitError(args->error_record.error_code_64); break; default: - UNIMPLEMENTED_MSG("Unimplemented LibAppletError mode={:02X}!", static_cast<u8>(mode)); + UNIMPLEMENTED_MSG("Unimplemented LibAppletError mode={:02X}!", mode); } } @@ -179,14 +179,14 @@ void Error::Execute() { error_code, std::chrono::seconds{args->error_record.posix_time}, callback); break; default: - UNIMPLEMENTED_MSG("Unimplemented LibAppletError mode={:02X}!", static_cast<u8>(mode)); + UNIMPLEMENTED_MSG("Unimplemented LibAppletError mode={:02X}!", mode); DisplayCompleted(); } } void Error::DisplayCompleted() { complete = true; - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::vector<u8>{})); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::vector<u8>{})); broker.SignalStateChanged(); } diff --git a/src/core/hle/service/am/applets/general_backend.cpp b/src/core/hle/service/am/applets/general_backend.cpp index 104501ac5..4d1df5cbe 100644 --- a/src/core/hle/service/am/applets/general_backend.cpp +++ b/src/core/hle/service/am/applets/general_backend.cpp @@ -38,7 +38,7 @@ static void LogCurrentStorage(AppletDataBroker& broker, std::string_view prefix) } Auth::Auth(Core::System& system_, Core::Frontend::ParentalControlsApplet& frontend_) - : Applet{system_.Kernel()}, frontend(frontend_) {} + : Applet{system_.Kernel()}, frontend{frontend_}, system{system_} {} Auth::~Auth() = default; @@ -90,7 +90,7 @@ void Auth::Execute() { const auto unimplemented_log = [this] { UNIMPLEMENTED_MSG("Unimplemented Auth applet type for type={:08X}, arg0={:02X}, " "arg1={:02X}, arg2={:02X}", - static_cast<u32>(type), arg0, arg1, arg2); + type, arg0, arg1, arg2); }; switch (type) { @@ -135,8 +135,8 @@ void Auth::Execute() { } } -void Auth::AuthFinished(bool successful) { - this->successful = successful; +void Auth::AuthFinished(bool is_successful) { + successful = is_successful; struct Return { ResultCode result_code; @@ -148,12 +148,12 @@ void Auth::AuthFinished(bool successful) { std::vector<u8> out(sizeof(Return)); std::memcpy(out.data(), &return_, sizeof(Return)); - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(out))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::move(out))); broker.SignalStateChanged(); } PhotoViewer::PhotoViewer(Core::System& system_, const Core::Frontend::PhotoViewerApplet& frontend_) - : Applet{system_.Kernel()}, frontend(frontend_), system{system_} {} + : Applet{system_.Kernel()}, frontend{frontend_}, system{system_} {} PhotoViewer::~PhotoViewer() = default; @@ -193,17 +193,17 @@ void PhotoViewer::Execute() { frontend.ShowAllPhotos(callback); break; default: - UNIMPLEMENTED_MSG("Unimplemented PhotoViewer applet mode={:02X}!", static_cast<u8>(mode)); + UNIMPLEMENTED_MSG("Unimplemented PhotoViewer applet mode={:02X}!", mode); } } void PhotoViewer::ViewFinished() { - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::vector<u8>{})); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::vector<u8>{})); broker.SignalStateChanged(); } StubApplet::StubApplet(Core::System& system_, AppletId id_) - : Applet{system_.Kernel()}, id(id_), system{system_} {} + : Applet{system_.Kernel()}, id{id_}, system{system_} {} StubApplet::~StubApplet() = default; @@ -234,8 +234,9 @@ void StubApplet::ExecuteInteractive() { LOG_WARNING(Service_AM, "called (STUBBED)"); LogCurrentStorage(broker, "ExecuteInteractive"); - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::vector<u8>(0x1000))); - broker.PushInteractiveDataFromApplet(std::make_shared<IStorage>(std::vector<u8>(0x1000))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::vector<u8>(0x1000))); + broker.PushInteractiveDataFromApplet( + std::make_shared<IStorage>(system, std::vector<u8>(0x1000))); broker.SignalStateChanged(); } @@ -243,8 +244,9 @@ void StubApplet::Execute() { LOG_WARNING(Service_AM, "called (STUBBED)"); LogCurrentStorage(broker, "Execute"); - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::vector<u8>(0x1000))); - broker.PushInteractiveDataFromApplet(std::make_shared<IStorage>(std::vector<u8>(0x1000))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::vector<u8>(0x1000))); + broker.PushInteractiveDataFromApplet( + std::make_shared<IStorage>(system, std::vector<u8>(0x1000))); broker.SignalStateChanged(); } diff --git a/src/core/hle/service/am/applets/general_backend.h b/src/core/hle/service/am/applets/general_backend.h index cfa2df369..ba76ae3d3 100644 --- a/src/core/hle/service/am/applets/general_backend.h +++ b/src/core/hle/service/am/applets/general_backend.h @@ -29,10 +29,11 @@ public: void ExecuteInteractive() override; void Execute() override; - void AuthFinished(bool successful = true); + void AuthFinished(bool is_successful = true); private: Core::Frontend::ParentalControlsApplet& frontend; + Core::System& system; bool complete = false; bool successful = false; diff --git a/src/core/hle/service/am/applets/profile_select.cpp b/src/core/hle/service/am/applets/profile_select.cpp index 70cc23552..77fba16c7 100644 --- a/src/core/hle/service/am/applets/profile_select.cpp +++ b/src/core/hle/service/am/applets/profile_select.cpp @@ -17,7 +17,7 @@ constexpr ResultCode ERR_USER_CANCELLED_SELECTION{ErrorModule::Account, 1}; ProfileSelect::ProfileSelect(Core::System& system_, const Core::Frontend::ProfileSelectApplet& frontend_) - : Applet{system_.Kernel()}, frontend(frontend_) {} + : Applet{system_.Kernel()}, frontend{frontend_}, system{system_} {} ProfileSelect::~ProfileSelect() = default; @@ -50,7 +50,7 @@ void ProfileSelect::ExecuteInteractive() { void ProfileSelect::Execute() { if (complete) { - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(final_data))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::move(final_data))); return; } @@ -71,7 +71,7 @@ void ProfileSelect::SelectionComplete(std::optional<Common::UUID> uuid) { final_data = std::vector<u8>(sizeof(UserSelectionOutput)); std::memcpy(final_data.data(), &output, final_data.size()); - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(final_data))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::move(final_data))); broker.SignalStateChanged(); } diff --git a/src/core/hle/service/am/applets/profile_select.h b/src/core/hle/service/am/applets/profile_select.h index 16364ead7..648d33a24 100644 --- a/src/core/hle/service/am/applets/profile_select.h +++ b/src/core/hle/service/am/applets/profile_select.h @@ -53,6 +53,7 @@ private: bool complete = false; ResultCode status = RESULT_SUCCESS; std::vector<u8> final_data; + Core::System& system; }; } // namespace Service::AM::Applets diff --git a/src/core/hle/service/am/applets/software_keyboard.cpp b/src/core/hle/service/am/applets/software_keyboard.cpp index bdeb0737a..3022438b1 100644 --- a/src/core/hle/service/am/applets/software_keyboard.cpp +++ b/src/core/hle/service/am/applets/software_keyboard.cpp @@ -53,7 +53,7 @@ static Core::Frontend::SoftwareKeyboardParameters ConvertToFrontendParameters( SoftwareKeyboard::SoftwareKeyboard(Core::System& system_, const Core::Frontend::SoftwareKeyboardApplet& frontend_) - : Applet{system_.Kernel()}, frontend(frontend_) {} + : Applet{system_.Kernel()}, frontend{frontend_}, system{system_} {} SoftwareKeyboard::~SoftwareKeyboard() = default; @@ -122,7 +122,7 @@ void SoftwareKeyboard::ExecuteInteractive() { switch (request) { case Request::Calc: { - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::vector<u8>{1})); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::vector<u8>{1})); broker.SignalStateChanged(); break; } @@ -135,7 +135,7 @@ void SoftwareKeyboard::ExecuteInteractive() { void SoftwareKeyboard::Execute() { if (complete) { - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(final_data))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::move(final_data))); broker.SignalStateChanged(); return; } @@ -179,15 +179,17 @@ void SoftwareKeyboard::WriteText(std::optional<std::u16string> text) { final_data = output_main; if (complete) { - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(output_main))); + broker.PushNormalDataFromApplet( + std::make_shared<IStorage>(system, std::move(output_main))); broker.SignalStateChanged(); } else { - broker.PushInteractiveDataFromApplet(std::make_shared<IStorage>(std::move(output_sub))); + broker.PushInteractiveDataFromApplet( + std::make_shared<IStorage>(system, std::move(output_sub))); } } else { output_main[0] = 1; complete = true; - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(output_main))); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::move(output_main))); broker.SignalStateChanged(); } } diff --git a/src/core/hle/service/am/applets/software_keyboard.h b/src/core/hle/service/am/applets/software_keyboard.h index 5a3824b5a..1d260fef8 100644 --- a/src/core/hle/service/am/applets/software_keyboard.h +++ b/src/core/hle/service/am/applets/software_keyboard.h @@ -80,6 +80,7 @@ private: bool complete = false; bool is_inline = false; std::vector<u8> final_data; + Core::System& system; }; } // namespace Service::AM::Applets diff --git a/src/core/hle/service/am/applets/web_browser.cpp b/src/core/hle/service/am/applets/web_browser.cpp index efe595c4f..2ab420789 100644 --- a/src/core/hle/service/am/applets/web_browser.cpp +++ b/src/core/hle/service/am/applets/web_browser.cpp @@ -1,558 +1,478 @@ -// Copyright 2018 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <array> -#include <cstring> -#include <vector> - #include "common/assert.h" -#include "common/common_funcs.h" #include "common/common_paths.h" #include "common/file_util.h" -#include "common/hex_util.h" #include "common/logging/log.h" #include "common/string_util.h" #include "core/core.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/mode.h" #include "core/file_sys/nca_metadata.h" +#include "core/file_sys/patch_manager.h" #include "core/file_sys/registered_cache.h" #include "core/file_sys/romfs.h" #include "core/file_sys/system_archive/system_archive.h" -#include "core/file_sys/vfs_types.h" -#include "core/frontend/applets/general_frontend.h" +#include "core/file_sys/vfs_vector.h" #include "core/frontend/applets/web_browser.h" #include "core/hle/kernel/process.h" +#include "core/hle/result.h" +#include "core/hle/service/am/am.h" #include "core/hle/service/am/applets/web_browser.h" #include "core/hle/service/filesystem/filesystem.h" -#include "core/loader/loader.h" +#include "core/hle/service/ns/pl_u.h" namespace Service::AM::Applets { -enum class WebArgTLVType : u16 { - InitialURL = 0x1, - ShopArgumentsURL = 0x2, ///< TODO(DarkLordZach): This is not the official name. - CallbackURL = 0x3, - CallbackableURL = 0x4, - ApplicationID = 0x5, - DocumentPath = 0x6, - DocumentKind = 0x7, - SystemDataID = 0x8, - ShareStartPage = 0x9, - Whitelist = 0xA, - News = 0xB, - UserID = 0xE, - AlbumEntry0 = 0xF, - ScreenShotEnabled = 0x10, - EcClientCertEnabled = 0x11, - Unk12 = 0x12, - PlayReportEnabled = 0x13, - Unk14 = 0x14, - Unk15 = 0x15, - BootDisplayKind = 0x17, - BackgroundKind = 0x18, - FooterEnabled = 0x19, - PointerEnabled = 0x1A, - LeftStickMode = 0x1B, - KeyRepeatFrame1 = 0x1C, - KeyRepeatFrame2 = 0x1D, - BootAsMediaPlayerInv = 0x1E, - DisplayUrlKind = 0x1F, - BootAsMediaPlayer = 0x21, - ShopJumpEnabled = 0x22, - MediaAutoPlayEnabled = 0x23, - LobbyParameter = 0x24, - ApplicationAlbumEntry = 0x26, - JsExtensionEnabled = 0x27, - AdditionalCommentText = 0x28, - TouchEnabledOnContents = 0x29, - UserAgentAdditionalString = 0x2A, - AdditionalMediaData0 = 0x2B, - MediaPlayerAutoCloseEnabled = 0x2C, - PageCacheEnabled = 0x2D, - WebAudioEnabled = 0x2E, - Unk2F = 0x2F, - YouTubeVideoWhitelist = 0x31, - FooterFixedKind = 0x32, - PageFadeEnabled = 0x33, - MediaCreatorApplicationRatingAge = 0x34, - BootLoadingIconEnabled = 0x35, - PageScrollIndicationEnabled = 0x36, - MediaPlayerSpeedControlEnabled = 0x37, - AlbumEntry1 = 0x38, - AlbumEntry2 = 0x39, - AlbumEntry3 = 0x3A, - AdditionalMediaData1 = 0x3B, - AdditionalMediaData2 = 0x3C, - AdditionalMediaData3 = 0x3D, - BootFooterButton = 0x3E, - OverrideWebAudioVolume = 0x3F, - OverrideMediaAudioVolume = 0x40, - BootMode = 0x41, - WebSessionEnabled = 0x42, -}; - -enum class ShimKind : u32 { - Shop = 1, - Login = 2, - Offline = 3, - Share = 4, - Web = 5, - Wifi = 6, - Lobby = 7, -}; - -enum class ShopWebTarget { - ApplicationInfo, - AddOnContentList, - SubscriptionList, - ConsumableItemList, - Home, - Settings, -}; - namespace { -constexpr std::size_t SHIM_KIND_COUNT = 0x8; - -struct WebArgHeader { - u16 count; - INSERT_PADDING_BYTES(2); - ShimKind kind; -}; -static_assert(sizeof(WebArgHeader) == 0x8, "WebArgHeader has incorrect size."); - -struct WebArgTLV { - WebArgTLVType type; - u16 size; - u32 offset; -}; -static_assert(sizeof(WebArgTLV) == 0x8, "WebArgTLV has incorrect size."); - -struct WebCommonReturnValue { - u32 result_code; - INSERT_PADDING_BYTES(0x4); - std::array<char, 0x1000> last_url; - u64 last_url_size; -}; -static_assert(sizeof(WebCommonReturnValue) == 0x1010, "WebCommonReturnValue has incorrect size."); - -struct WebWifiPageArg { - INSERT_PADDING_BYTES(4); - std::array<char, 0x100> connection_test_url; - std::array<char, 0x400> initial_url; - std::array<u8, 0x10> nifm_network_uuid; - u32 nifm_requirement; -}; -static_assert(sizeof(WebWifiPageArg) == 0x518, "WebWifiPageArg has incorrect size."); - -struct WebWifiReturnValue { - INSERT_PADDING_BYTES(4); - u32 result; -}; -static_assert(sizeof(WebWifiReturnValue) == 0x8, "WebWifiReturnValue has incorrect size."); - -enum class OfflineWebSource : u32 { - OfflineHtmlPage = 0x1, - ApplicationLegalInformation = 0x2, - SystemDataPage = 0x3, -}; - -std::map<WebArgTLVType, std::vector<u8>> GetWebArguments(const std::vector<u8>& arg) { - if (arg.size() < sizeof(WebArgHeader)) - return {}; - - WebArgHeader header{}; - std::memcpy(&header, arg.data(), sizeof(WebArgHeader)); - - std::map<WebArgTLVType, std::vector<u8>> out; - u64 offset = sizeof(WebArgHeader); - for (std::size_t i = 0; i < header.count; ++i) { - if (arg.size() < (offset + sizeof(WebArgTLV))) - return out; +template <typename T> +void ParseRawValue(T& value, const std::vector<u8>& data) { + static_assert(std::is_trivially_copyable_v<T>, + "It's undefined behavior to use memcpy with non-trivially copyable objects"); + std::memcpy(&value, data.data(), data.size()); +} - WebArgTLV tlv{}; - std::memcpy(&tlv, arg.data() + offset, sizeof(WebArgTLV)); - offset += sizeof(WebArgTLV); +template <typename T> +T ParseRawValue(const std::vector<u8>& data) { + T value; + ParseRawValue(value, data); + return value; +} - offset += tlv.offset; - if (arg.size() < (offset + tlv.size)) - return out; +std::string ParseStringValue(const std::vector<u8>& data) { + return Common::StringFromFixedZeroTerminatedBuffer(reinterpret_cast<const char*>(data.data()), + data.size()); +} - std::vector<u8> data(tlv.size); - std::memcpy(data.data(), arg.data() + offset, tlv.size); - offset += tlv.size; +std::string GetMainURL(const std::string& url) { + const auto index = url.find('?'); - out.insert_or_assign(tlv.type, data); + if (index == std::string::npos) { + return url; } - return out; + return url.substr(0, index); } -FileSys::VirtualFile GetApplicationRomFS(const Core::System& system, u64 title_id, - FileSys::ContentRecordType type) { - const auto& installed{system.GetContentProvider()}; - const auto res = installed.GetEntry(title_id, type); +WebArgInputTLVMap ReadWebArgs(const std::vector<u8>& web_arg, WebArgHeader& web_arg_header) { + std::memcpy(&web_arg_header, web_arg.data(), sizeof(WebArgHeader)); - if (res != nullptr) { - return res->GetRomFS(); + if (web_arg.size() == sizeof(WebArgHeader)) { + return {}; } - if (type == FileSys::ContentRecordType::Data) { - return FileSys::SystemArchive::SynthesizeSystemArchive(title_id); + WebArgInputTLVMap input_tlv_map; + + u64 current_offset = sizeof(WebArgHeader); + + for (std::size_t i = 0; i < web_arg_header.total_tlv_entries; ++i) { + if (web_arg.size() < current_offset + sizeof(WebArgInputTLV)) { + return input_tlv_map; + } + + WebArgInputTLV input_tlv; + std::memcpy(&input_tlv, web_arg.data() + current_offset, sizeof(WebArgInputTLV)); + + current_offset += sizeof(WebArgInputTLV); + + if (web_arg.size() < current_offset + input_tlv.arg_data_size) { + return input_tlv_map; + } + + std::vector<u8> data(input_tlv.arg_data_size); + std::memcpy(data.data(), web_arg.data() + current_offset, input_tlv.arg_data_size); + + current_offset += input_tlv.arg_data_size; + + input_tlv_map.insert_or_assign(input_tlv.input_tlv_type, std::move(data)); } - return nullptr; + return input_tlv_map; } -} // Anonymous namespace +FileSys::VirtualFile GetOfflineRomFS(Core::System& system, u64 title_id, + FileSys::ContentRecordType nca_type) { + if (nca_type == FileSys::ContentRecordType::Data) { + const auto nca = + system.GetFileSystemController().GetSystemNANDContents()->GetEntry(title_id, nca_type); + + if (nca == nullptr) { + LOG_ERROR(Service_AM, + "NCA of type={} with title_id={:016X} is not found in the System NAND!", + nca_type, title_id); + return FileSys::SystemArchive::SynthesizeSystemArchive(title_id); + } -WebBrowser::WebBrowser(Core::System& system_, Core::Frontend::WebBrowserApplet& frontend_, - Core::Frontend::ECommerceApplet* frontend_e_commerce_) - : Applet{system_.Kernel()}, frontend(frontend_), - frontend_e_commerce(frontend_e_commerce_), system{system_} {} + return nca->GetRomFS(); + } else { + const auto nca = system.GetContentProvider().GetEntry(title_id, nca_type); -WebBrowser::~WebBrowser() = default; + if (nca == nullptr) { + LOG_ERROR(Service_AM, + "NCA of type={} with title_id={:016X} is not found in the ContentProvider!", + nca_type, title_id); + return nullptr; + } -void WebBrowser::Initialize() { - Applet::Initialize(); + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; - complete = false; - temporary_dir.clear(); - filename.clear(); - status = RESULT_SUCCESS; + return pm.PatchRomFS(nca->GetRomFS(), nca->GetBaseIVFCOffset(), nca_type); + } +} - const auto web_arg_storage = broker.PopNormalDataToApplet(); - ASSERT(web_arg_storage != nullptr); - const auto& web_arg = web_arg_storage->GetData(); +void ExtractSharedFonts(Core::System& system) { + static constexpr std::array<const char*, 7> DECRYPTED_SHARED_FONTS{ + "FontStandard.ttf", + "FontChineseSimplified.ttf", + "FontExtendedChineseSimplified.ttf", + "FontChineseTraditional.ttf", + "FontKorean.ttf", + "FontNintendoExtended.ttf", + "FontNintendoExtended2.ttf", + }; - ASSERT(web_arg.size() >= 0x8); - std::memcpy(&kind, web_arg.data() + 0x4, sizeof(ShimKind)); + for (std::size_t i = 0; i < NS::SHARED_FONTS.size(); ++i) { + const auto fonts_dir = Common::FS::SanitizePath( + fmt::format("{}/fonts", Common::FS::GetUserPath(Common::FS::UserPath::CacheDir)), + Common::FS::DirectorySeparator::PlatformDefault); - args = GetWebArguments(web_arg); + const auto font_file_path = + Common::FS::SanitizePath(fmt::format("{}/{}", fonts_dir, DECRYPTED_SHARED_FONTS[i]), + Common::FS::DirectorySeparator::PlatformDefault); - InitializeInternal(); -} + if (Common::FS::Exists(font_file_path)) { + continue; + } -bool WebBrowser::TransactionComplete() const { - return complete; -} + const auto font = NS::SHARED_FONTS[i]; + const auto font_title_id = static_cast<u64>(font.first); -ResultCode WebBrowser::GetStatus() const { - return status; -} + const auto nca = system.GetFileSystemController().GetSystemNANDContents()->GetEntry( + font_title_id, FileSys::ContentRecordType::Data); -void WebBrowser::ExecuteInteractive() { - UNIMPLEMENTED_MSG("Unexpected interactive data recieved!"); -} + FileSys::VirtualFile romfs; -void WebBrowser::Execute() { - if (complete) { - return; - } + if (!nca) { + romfs = FileSys::SystemArchive::SynthesizeSystemArchive(font_title_id); + } else { + romfs = nca->GetRomFS(); + } - if (status != RESULT_SUCCESS) { - complete = true; + if (!romfs) { + LOG_ERROR(Service_AM, "SharedFont RomFS with title_id={:016X} cannot be extracted!", + font_title_id); + continue; + } - // This is a workaround in order not to softlock yuzu when an error happens during the - // webapplet init. In order to avoid an svcBreak, the status is set to RESULT_SUCCESS - Finalize(); - status = RESULT_SUCCESS; + const auto extracted_romfs = FileSys::ExtractRomFS(romfs); - return; - } + if (!extracted_romfs) { + LOG_ERROR(Service_AM, "SharedFont RomFS with title_id={:016X} failed to extract!", + font_title_id); + continue; + } - ExecuteInternal(); -} + const auto font_file = extracted_romfs->GetFile(font.second); -void WebBrowser::UnpackRomFS() { - if (unpacked) - return; + if (!font_file) { + LOG_ERROR(Service_AM, "SharedFont RomFS with title_id={:016X} has no font file \"{}\"!", + font_title_id, font.second); + continue; + } - ASSERT(offline_romfs != nullptr); - const auto dir = - FileSys::ExtractRomFS(offline_romfs, FileSys::RomFSExtractionType::SingleDiscard); - const auto& vfs{system.GetFilesystem()}; - const auto temp_dir = vfs->CreateDirectory(temporary_dir, FileSys::Mode::ReadWrite); - FileSys::VfsRawCopyD(dir, temp_dir); + std::vector<u32> font_data_u32(font_file->GetSize() / sizeof(u32)); + font_file->ReadBytes<u32>(font_data_u32.data(), font_file->GetSize()); - unpacked = true; -} + std::transform(font_data_u32.begin(), font_data_u32.end(), font_data_u32.begin(), + Common::swap32); -void WebBrowser::Finalize() { - complete = true; + std::vector<u8> decrypted_data(font_file->GetSize() - 8); - WebCommonReturnValue out{}; - out.result_code = 0; - out.last_url_size = 0; + NS::DecryptSharedFontToTTF(font_data_u32, decrypted_data); - std::vector<u8> data(sizeof(WebCommonReturnValue)); - std::memcpy(data.data(), &out, sizeof(WebCommonReturnValue)); + FileSys::VirtualFile decrypted_font = std::make_shared<FileSys::VectorVfsFile>( + std::move(decrypted_data), DECRYPTED_SHARED_FONTS[i]); - broker.PushNormalDataFromApplet(std::make_shared<IStorage>(std::move(data))); - broker.SignalStateChanged(); + const auto temp_dir = + system.GetFilesystem()->CreateDirectory(fonts_dir, FileSys::Mode::ReadWrite); + + const auto out_file = temp_dir->CreateFile(DECRYPTED_SHARED_FONTS[i]); - if (!temporary_dir.empty() && Common::FS::IsDirectory(temporary_dir)) { - Common::FS::DeleteDirRecursively(temporary_dir); + FileSys::VfsRawCopy(decrypted_font, out_file); } } -void WebBrowser::InitializeInternal() { - using WebAppletInitializer = void (WebBrowser::*)(); +} // namespace - constexpr std::array<WebAppletInitializer, SHIM_KIND_COUNT> functions{ - nullptr, &WebBrowser::InitializeShop, - nullptr, &WebBrowser::InitializeOffline, - nullptr, nullptr, - nullptr, nullptr, - }; +WebBrowser::WebBrowser(Core::System& system_, const Core::Frontend::WebBrowserApplet& frontend_) + : Applet{system_.Kernel()}, frontend(frontend_), system{system_} {} - const auto index = static_cast<u32>(kind); +WebBrowser::~WebBrowser() = default; - if (index > functions.size() || functions[index] == nullptr) { - LOG_ERROR(Service_AM, "Invalid shim_kind={:08X}", index); - return; - } +void WebBrowser::Initialize() { + Applet::Initialize(); - const auto function = functions[index]; - (this->*function)(); -} + LOG_INFO(Service_AM, "Initializing Web Browser Applet."); -void WebBrowser::ExecuteInternal() { - using WebAppletExecutor = void (WebBrowser::*)(); + LOG_DEBUG(Service_AM, + "Initializing Applet with common_args: arg_version={}, lib_version={}, " + "play_startup_sound={}, size={}, system_tick={}, theme_color={}", + common_args.arguments_version, common_args.library_version, + common_args.play_startup_sound, common_args.size, common_args.system_tick, + common_args.theme_color); - constexpr std::array<WebAppletExecutor, SHIM_KIND_COUNT> functions{ - nullptr, &WebBrowser::ExecuteShop, - nullptr, &WebBrowser::ExecuteOffline, - nullptr, nullptr, - nullptr, nullptr, - }; + web_applet_version = WebAppletVersion{common_args.library_version}; - const auto index = static_cast<u32>(kind); + const auto web_arg_storage = broker.PopNormalDataToApplet(); + ASSERT(web_arg_storage != nullptr); - if (index > functions.size() || functions[index] == nullptr) { - LOG_ERROR(Service_AM, "Invalid shim_kind={:08X}", index); - return; - } + const auto& web_arg = web_arg_storage->GetData(); + ASSERT_OR_EXECUTE(web_arg.size() >= sizeof(WebArgHeader), { return; }); - const auto function = functions[index]; - (this->*function)(); -} + web_arg_input_tlv_map = ReadWebArgs(web_arg, web_arg_header); -void WebBrowser::InitializeShop() { - if (frontend_e_commerce == nullptr) { - LOG_ERROR(Service_AM, "Missing ECommerce Applet frontend!"); - status = RESULT_UNKNOWN; - return; - } + LOG_DEBUG(Service_AM, "WebArgHeader: total_tlv_entries={}, shim_kind={}", + web_arg_header.total_tlv_entries, web_arg_header.shim_kind); - const auto user_id_data = args.find(WebArgTLVType::UserID); + ExtractSharedFonts(system); - user_id = std::nullopt; - if (user_id_data != args.end()) { - user_id = u128{}; - std::memcpy(user_id->data(), user_id_data->second.data(), sizeof(u128)); + switch (web_arg_header.shim_kind) { + case ShimKind::Shop: + InitializeShop(); + break; + case ShimKind::Login: + InitializeLogin(); + break; + case ShimKind::Offline: + InitializeOffline(); + break; + case ShimKind::Share: + InitializeShare(); + break; + case ShimKind::Web: + InitializeWeb(); + break; + case ShimKind::Wifi: + InitializeWifi(); + break; + case ShimKind::Lobby: + InitializeLobby(); + break; + default: + UNREACHABLE_MSG("Invalid ShimKind={}", web_arg_header.shim_kind); + break; } +} - const auto url = args.find(WebArgTLVType::ShopArgumentsURL); +bool WebBrowser::TransactionComplete() const { + return complete; +} - if (url == args.end()) { - LOG_ERROR(Service_AM, "Missing EShop Arguments URL for initialization!"); - status = RESULT_UNKNOWN; - return; - } +ResultCode WebBrowser::GetStatus() const { + return status; +} - std::vector<std::string> split_query; - Common::SplitString(Common::StringFromFixedZeroTerminatedBuffer( - reinterpret_cast<const char*>(url->second.data()), url->second.size()), - '?', split_query); - - // 2 -> Main URL '?' Query Parameters - // Less is missing info, More is malformed - if (split_query.size() != 2) { - LOG_ERROR(Service_AM, "EShop Arguments has more than one question mark, malformed"); - status = RESULT_UNKNOWN; - return; - } +void WebBrowser::ExecuteInteractive() { + UNIMPLEMENTED_MSG("WebSession is not implemented"); +} - std::vector<std::string> queries; - Common::SplitString(split_query[1], '&', queries); +void WebBrowser::Execute() { + switch (web_arg_header.shim_kind) { + case ShimKind::Shop: + ExecuteShop(); + break; + case ShimKind::Login: + ExecuteLogin(); + break; + case ShimKind::Offline: + ExecuteOffline(); + break; + case ShimKind::Share: + ExecuteShare(); + break; + case ShimKind::Web: + ExecuteWeb(); + break; + case ShimKind::Wifi: + ExecuteWifi(); + break; + case ShimKind::Lobby: + ExecuteLobby(); + break; + default: + UNREACHABLE_MSG("Invalid ShimKind={}", web_arg_header.shim_kind); + WebBrowserExit(WebExitReason::EndButtonPressed); + break; + } +} - const auto split_single_query = - [](const std::string& in) -> std::pair<std::string, std::string> { - const auto index = in.find('='); - if (index == std::string::npos || index == in.size() - 1) { - return {in, ""}; - } +void WebBrowser::ExtractOfflineRomFS() { + LOG_DEBUG(Service_AM, "Extracting RomFS to {}", offline_cache_dir); - return {in.substr(0, index), in.substr(index + 1)}; - }; + const auto extracted_romfs_dir = + FileSys::ExtractRomFS(offline_romfs, FileSys::RomFSExtractionType::SingleDiscard); - std::transform(queries.begin(), queries.end(), - std::inserter(shop_query, std::next(shop_query.begin())), split_single_query); + const auto temp_dir = + system.GetFilesystem()->CreateDirectory(offline_cache_dir, FileSys::Mode::ReadWrite); - const auto scene = shop_query.find("scene"); + FileSys::VfsRawCopyD(extracted_romfs_dir, temp_dir); +} - if (scene == shop_query.end()) { - LOG_ERROR(Service_AM, "No scene parameter was passed via shop query!"); - status = RESULT_UNKNOWN; - return; +void WebBrowser::WebBrowserExit(WebExitReason exit_reason, std::string last_url) { + if ((web_arg_header.shim_kind == ShimKind::Share && + web_applet_version >= WebAppletVersion::Version196608) || + (web_arg_header.shim_kind == ShimKind::Web && + web_applet_version >= WebAppletVersion::Version524288)) { + // TODO: Push Output TLVs instead of a WebCommonReturnValue } - const std::map<std::string, ShopWebTarget, std::less<>> target_map{ - {"product_detail", ShopWebTarget::ApplicationInfo}, - {"aocs", ShopWebTarget::AddOnContentList}, - {"subscriptions", ShopWebTarget::SubscriptionList}, - {"consumption", ShopWebTarget::ConsumableItemList}, - {"settings", ShopWebTarget::Settings}, - {"top", ShopWebTarget::Home}, - }; + WebCommonReturnValue web_common_return_value; - const auto target = target_map.find(scene->second); - if (target == target_map.end()) { - LOG_ERROR(Service_AM, "Scene for shop query is invalid! (scene={})", scene->second); - status = RESULT_UNKNOWN; - return; - } + web_common_return_value.exit_reason = exit_reason; + std::memcpy(&web_common_return_value.last_url, last_url.data(), last_url.size()); + web_common_return_value.last_url_size = last_url.size(); - shop_web_target = target->second; + LOG_DEBUG(Service_AM, "WebCommonReturnValue: exit_reason={}, last_url={}, last_url_size={}", + exit_reason, last_url, last_url.size()); - const auto title_id_data = shop_query.find("dst_app_id"); - if (title_id_data != shop_query.end()) { - title_id = std::stoull(title_id_data->second, nullptr, 0x10); - } + complete = true; + std::vector<u8> out_data(sizeof(WebCommonReturnValue)); + std::memcpy(out_data.data(), &web_common_return_value, out_data.size()); + broker.PushNormalDataFromApplet(std::make_shared<IStorage>(system, std::move(out_data))); + broker.SignalStateChanged(); +} - const auto mode_data = shop_query.find("mode"); - if (mode_data != shop_query.end()) { - shop_full_display = mode_data->second == "full"; - } +bool WebBrowser::InputTLVExistsInMap(WebArgInputTLVType input_tlv_type) const { + return web_arg_input_tlv_map.find(input_tlv_type) != web_arg_input_tlv_map.end(); } -void WebBrowser::InitializeOffline() { - if (args.find(WebArgTLVType::DocumentPath) == args.end() || - args.find(WebArgTLVType::DocumentKind) == args.end() || - args.find(WebArgTLVType::ApplicationID) == args.end()) { - status = RESULT_UNKNOWN; - LOG_ERROR(Service_AM, "Missing necessary parameters for initialization!"); +std::optional<std::vector<u8>> WebBrowser::GetInputTLVData(WebArgInputTLVType input_tlv_type) { + const auto map_it = web_arg_input_tlv_map.find(input_tlv_type); + + if (map_it == web_arg_input_tlv_map.end()) { + return std::nullopt; } - const auto url_data = args[WebArgTLVType::DocumentPath]; - filename = Common::StringFromFixedZeroTerminatedBuffer( - reinterpret_cast<const char*>(url_data.data()), url_data.size()); + return map_it->second; +} - OfflineWebSource source; - ASSERT(args[WebArgTLVType::DocumentKind].size() >= 4); - std::memcpy(&source, args[WebArgTLVType::DocumentKind].data(), sizeof(OfflineWebSource)); +void WebBrowser::InitializeShop() {} - constexpr std::array<const char*, 3> WEB_SOURCE_NAMES{ - "manual", - "legal", - "system", - }; +void WebBrowser::InitializeLogin() {} + +void WebBrowser::InitializeOffline() { + const auto document_path = + ParseStringValue(GetInputTLVData(WebArgInputTLVType::DocumentPath).value()); + + const auto document_kind = + ParseRawValue<DocumentKind>(GetInputTLVData(WebArgInputTLVType::DocumentKind).value()); + + std::string additional_paths; - temporary_dir = - Common::FS::SanitizePath(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir) + - "web_applet_" + WEB_SOURCE_NAMES[static_cast<u32>(source) - 1], - Common::FS::DirectorySeparator::PlatformDefault); - Common::FS::DeleteDirRecursively(temporary_dir); - - u64 title_id = 0; // 0 corresponds to current process - ASSERT(args[WebArgTLVType::ApplicationID].size() >= 0x8); - std::memcpy(&title_id, args[WebArgTLVType::ApplicationID].data(), sizeof(u64)); - FileSys::ContentRecordType type = FileSys::ContentRecordType::Data; - - switch (source) { - case OfflineWebSource::OfflineHtmlPage: - // While there is an AppID TLV field, in official SW this is always ignored. - title_id = 0; - type = FileSys::ContentRecordType::HtmlDocument; + switch (document_kind) { + case DocumentKind::OfflineHtmlPage: + default: + title_id = system.CurrentProcess()->GetTitleID(); + nca_type = FileSys::ContentRecordType::HtmlDocument; + additional_paths = "html-document"; break; - case OfflineWebSource::ApplicationLegalInformation: - type = FileSys::ContentRecordType::LegalInformation; + case DocumentKind::ApplicationLegalInformation: + title_id = ParseRawValue<u64>(GetInputTLVData(WebArgInputTLVType::ApplicationID).value()); + nca_type = FileSys::ContentRecordType::LegalInformation; break; - case OfflineWebSource::SystemDataPage: - type = FileSys::ContentRecordType::Data; + case DocumentKind::SystemDataPage: + title_id = ParseRawValue<u64>(GetInputTLVData(WebArgInputTLVType::SystemDataID).value()); + nca_type = FileSys::ContentRecordType::Data; break; } - if (title_id == 0) { - title_id = system.CurrentProcess()->GetTitleID(); - } + static constexpr std::array<const char*, 3> RESOURCE_TYPES{ + "manual", + "legal_information", + "system_data", + }; - offline_romfs = GetApplicationRomFS(system, title_id, type); - if (offline_romfs == nullptr) { - status = RESULT_UNKNOWN; - LOG_ERROR(Service_AM, "Failed to find offline data for request!"); - } + offline_cache_dir = Common::FS::SanitizePath( + fmt::format("{}/offline_web_applet_{}/{:016X}", + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir), + RESOURCE_TYPES[static_cast<u32>(document_kind) - 1], title_id), + Common::FS::DirectorySeparator::PlatformDefault); - std::string path_additional_directory; - if (source == OfflineWebSource::OfflineHtmlPage) { - path_additional_directory = std::string(DIR_SEP).append("html-document"); - } + offline_document = Common::FS::SanitizePath( + fmt::format("{}/{}/{}", offline_cache_dir, additional_paths, document_path), + Common::FS::DirectorySeparator::PlatformDefault); +} + +void WebBrowser::InitializeShare() {} - filename = - Common::FS::SanitizePath(temporary_dir + path_additional_directory + DIR_SEP + filename, - Common::FS::DirectorySeparator::PlatformDefault); +void WebBrowser::InitializeWeb() { + external_url = ParseStringValue(GetInputTLVData(WebArgInputTLVType::InitialURL).value()); } +void WebBrowser::InitializeWifi() {} + +void WebBrowser::InitializeLobby() {} + void WebBrowser::ExecuteShop() { - const auto callback = [this]() { Finalize(); }; + LOG_WARNING(Service_AM, "(STUBBED) called, Shop Applet is not implemented"); + WebBrowserExit(WebExitReason::EndButtonPressed); +} - const auto check_optional_parameter = [this](const auto& p) { - if (!p.has_value()) { - LOG_ERROR(Service_AM, "Missing one or more necessary parameters for execution!"); - status = RESULT_UNKNOWN; - return false; - } +void WebBrowser::ExecuteLogin() { + LOG_WARNING(Service_AM, "(STUBBED) called, Login Applet is not implemented"); + WebBrowserExit(WebExitReason::EndButtonPressed); +} - return true; - }; +void WebBrowser::ExecuteOffline() { + const auto main_url = Common::FS::SanitizePath(GetMainURL(offline_document), + Common::FS::DirectorySeparator::PlatformDefault); - switch (shop_web_target) { - case ShopWebTarget::ApplicationInfo: - if (!check_optional_parameter(title_id)) - return; - frontend_e_commerce->ShowApplicationInformation(callback, *title_id, user_id, - shop_full_display, shop_extra_parameter); - break; - case ShopWebTarget::AddOnContentList: - if (!check_optional_parameter(title_id)) - return; - frontend_e_commerce->ShowAddOnContentList(callback, *title_id, user_id, shop_full_display); - break; - case ShopWebTarget::ConsumableItemList: - if (!check_optional_parameter(title_id)) - return; - frontend_e_commerce->ShowConsumableItemList(callback, *title_id, user_id); - break; - case ShopWebTarget::Home: - if (!check_optional_parameter(user_id)) - return; - if (!check_optional_parameter(shop_full_display)) - return; - frontend_e_commerce->ShowShopHome(callback, *user_id, *shop_full_display); - break; - case ShopWebTarget::Settings: - if (!check_optional_parameter(user_id)) - return; - if (!check_optional_parameter(shop_full_display)) - return; - frontend_e_commerce->ShowSettings(callback, *user_id, *shop_full_display); - break; - case ShopWebTarget::SubscriptionList: - if (!check_optional_parameter(title_id)) + if (!Common::FS::Exists(main_url)) { + offline_romfs = GetOfflineRomFS(system, title_id, nca_type); + + if (offline_romfs == nullptr) { + LOG_ERROR(Service_AM, + "RomFS with title_id={:016X} and nca_type={} cannot be extracted!", title_id, + nca_type); + WebBrowserExit(WebExitReason::WindowClosed); return; - frontend_e_commerce->ShowSubscriptionList(callback, *title_id, user_id); - break; - default: - UNREACHABLE(); + } } + + LOG_INFO(Service_AM, "Opening offline document at {}", offline_document); + + frontend.OpenLocalWebPage( + offline_document, [this] { ExtractOfflineRomFS(); }, + [this](WebExitReason exit_reason, std::string last_url) { + WebBrowserExit(exit_reason, last_url); + }); } -void WebBrowser::ExecuteOffline() { - frontend.OpenPageLocal( - filename, [this] { UnpackRomFS(); }, [this] { Finalize(); }); +void WebBrowser::ExecuteShare() { + LOG_WARNING(Service_AM, "(STUBBED) called, Share Applet is not implemented"); + WebBrowserExit(WebExitReason::EndButtonPressed); +} + +void WebBrowser::ExecuteWeb() { + LOG_INFO(Service_AM, "Opening external URL at {}", external_url); + + frontend.OpenExternalWebPage(external_url, + [this](WebExitReason exit_reason, std::string last_url) { + WebBrowserExit(exit_reason, last_url); + }); } +void WebBrowser::ExecuteWifi() { + LOG_WARNING(Service_AM, "(STUBBED) called, Wifi Applet is not implemented"); + WebBrowserExit(WebExitReason::EndButtonPressed); +} + +void WebBrowser::ExecuteLobby() { + LOG_WARNING(Service_AM, "(STUBBED) called, Lobby Applet is not implemented"); + WebBrowserExit(WebExitReason::EndButtonPressed); +} } // namespace Service::AM::Applets diff --git a/src/core/hle/service/am/applets/web_browser.h b/src/core/hle/service/am/applets/web_browser.h index 8d4027411..04c274754 100644 --- a/src/core/hle/service/am/applets/web_browser.h +++ b/src/core/hle/service/am/applets/web_browser.h @@ -1,28 +1,31 @@ -// Copyright 2018 yuzu emulator team +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. #pragma once -#include <map> +#include <optional> + +#include "common/common_funcs.h" +#include "common/common_types.h" #include "core/file_sys/vfs_types.h" -#include "core/hle/service/am/am.h" +#include "core/hle/result.h" #include "core/hle/service/am/applets/applets.h" +#include "core/hle/service/am/applets/web_types.h" namespace Core { class System; } -namespace Service::AM::Applets { +namespace FileSys { +enum class ContentRecordType : u8; +} -enum class ShimKind : u32; -enum class ShopWebTarget; -enum class WebArgTLVType : u16; +namespace Service::AM::Applets { class WebBrowser final : public Applet { public: - WebBrowser(Core::System& system_, Core::Frontend::WebBrowserApplet& frontend_, - Core::Frontend::ECommerceApplet* frontend_e_commerce_ = nullptr); + WebBrowser(Core::System& system_, const Core::Frontend::WebBrowserApplet& frontend_); ~WebBrowser() override; @@ -33,49 +36,50 @@ public: void ExecuteInteractive() override; void Execute() override; - // Callback to be fired when the frontend needs the manual RomFS unpacked to temporary - // directory. This is a blocking call and may take a while as some manuals can be up to 100MB in - // size. Attempting to access files at filename before invocation is likely to not work. - void UnpackRomFS(); + void ExtractOfflineRomFS(); - // Callback to be fired when the frontend is finished browsing. This will delete the temporary - // manual RomFS extracted files, so ensure this is only called at actual finalization. - void Finalize(); + void WebBrowserExit(WebExitReason exit_reason, std::string last_url = ""); private: - void InitializeInternal(); - void ExecuteInternal(); + bool InputTLVExistsInMap(WebArgInputTLVType input_tlv_type) const; - // Specific initializers for the types of web applets + std::optional<std::vector<u8>> GetInputTLVData(WebArgInputTLVType input_tlv_type); + + // Initializers for the various types of browser applets void InitializeShop(); + void InitializeLogin(); void InitializeOffline(); + void InitializeShare(); + void InitializeWeb(); + void InitializeWifi(); + void InitializeLobby(); - // Specific executors for the types of web applets + // Executors for the various types of browser applets void ExecuteShop(); + void ExecuteLogin(); void ExecuteOffline(); + void ExecuteShare(); + void ExecuteWeb(); + void ExecuteWifi(); + void ExecuteLobby(); - Core::Frontend::WebBrowserApplet& frontend; - - // Extra frontends for specialized functions - Core::Frontend::ECommerceApplet* frontend_e_commerce; + const Core::Frontend::WebBrowserApplet& frontend; - bool complete = false; - bool unpacked = false; - ResultCode status = RESULT_SUCCESS; + bool complete{false}; + ResultCode status{RESULT_SUCCESS}; - ShimKind kind; - std::map<WebArgTLVType, std::vector<u8>> args; + WebAppletVersion web_applet_version; + WebExitReason web_exit_reason; + WebArgHeader web_arg_header; + WebArgInputTLVMap web_arg_input_tlv_map; + u64 title_id; + FileSys::ContentRecordType nca_type; + std::string offline_cache_dir; + std::string offline_document; FileSys::VirtualFile offline_romfs; - std::string temporary_dir; - std::string filename; - - ShopWebTarget shop_web_target; - std::map<std::string, std::string, std::less<>> shop_query; - std::optional<u64> title_id = 0; - std::optional<u128> user_id; - std::optional<bool> shop_full_display; - std::string shop_extra_parameter; + + std::string external_url; Core::System& system; }; diff --git a/src/core/hle/service/am/applets/web_types.h b/src/core/hle/service/am/applets/web_types.h new file mode 100644 index 000000000..419c2bf79 --- /dev/null +++ b/src/core/hle/service/am/applets/web_types.h @@ -0,0 +1,178 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <unordered_map> +#include <vector> + +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "common/swap.h" + +namespace Service::AM::Applets { + +enum class WebAppletVersion : u32_le { + Version0 = 0x0, // Only used by WifiWebAuthApplet + Version131072 = 0x20000, // 1.0.0 - 2.3.0 + Version196608 = 0x30000, // 3.0.0 - 4.1.0 + Version327680 = 0x50000, // 5.0.0 - 5.1.0 + Version393216 = 0x60000, // 6.0.0 - 7.0.1 + Version524288 = 0x80000, // 8.0.0+ +}; + +enum class ShimKind : u32 { + Shop = 1, + Login = 2, + Offline = 3, + Share = 4, + Web = 5, + Wifi = 6, + Lobby = 7, +}; + +enum class WebExitReason : u32 { + EndButtonPressed = 0, + BackButtonPressed = 1, + ExitRequested = 2, + CallbackURL = 3, + WindowClosed = 4, + ErrorDialog = 7, +}; + +enum class WebArgInputTLVType : u16 { + InitialURL = 0x1, + CallbackURL = 0x3, + CallbackableURL = 0x4, + ApplicationID = 0x5, + DocumentPath = 0x6, + DocumentKind = 0x7, + SystemDataID = 0x8, + ShareStartPage = 0x9, + Whitelist = 0xA, + News = 0xB, + UserID = 0xE, + AlbumEntry0 = 0xF, + ScreenShotEnabled = 0x10, + EcClientCertEnabled = 0x11, + PlayReportEnabled = 0x13, + BootDisplayKind = 0x17, + BackgroundKind = 0x18, + FooterEnabled = 0x19, + PointerEnabled = 0x1A, + LeftStickMode = 0x1B, + KeyRepeatFrame1 = 0x1C, + KeyRepeatFrame2 = 0x1D, + BootAsMediaPlayerInverted = 0x1E, + DisplayURLKind = 0x1F, + BootAsMediaPlayer = 0x21, + ShopJumpEnabled = 0x22, + MediaAutoPlayEnabled = 0x23, + LobbyParameter = 0x24, + ApplicationAlbumEntry = 0x26, + JsExtensionEnabled = 0x27, + AdditionalCommentText = 0x28, + TouchEnabledOnContents = 0x29, + UserAgentAdditionalString = 0x2A, + AdditionalMediaData0 = 0x2B, + MediaPlayerAutoCloseEnabled = 0x2C, + PageCacheEnabled = 0x2D, + WebAudioEnabled = 0x2E, + YouTubeVideoWhitelist = 0x31, + FooterFixedKind = 0x32, + PageFadeEnabled = 0x33, + MediaCreatorApplicationRatingAge = 0x34, + BootLoadingIconEnabled = 0x35, + PageScrollIndicatorEnabled = 0x36, + MediaPlayerSpeedControlEnabled = 0x37, + AlbumEntry1 = 0x38, + AlbumEntry2 = 0x39, + AlbumEntry3 = 0x3A, + AdditionalMediaData1 = 0x3B, + AdditionalMediaData2 = 0x3C, + AdditionalMediaData3 = 0x3D, + BootFooterButton = 0x3E, + OverrideWebAudioVolume = 0x3F, + OverrideMediaAudioVolume = 0x40, + BootMode = 0x41, + WebSessionEnabled = 0x42, + MediaPlayerOfflineEnabled = 0x43, +}; + +enum class WebArgOutputTLVType : u16 { + ShareExitReason = 0x1, + LastURL = 0x2, + LastURLSize = 0x3, + SharePostResult = 0x4, + PostServiceName = 0x5, + PostServiceNameSize = 0x6, + PostID = 0x7, + PostIDSize = 0x8, + MediaPlayerAutoClosedByCompletion = 0x9, +}; + +enum class DocumentKind : u32 { + OfflineHtmlPage = 1, + ApplicationLegalInformation = 2, + SystemDataPage = 3, +}; + +enum class ShareStartPage : u32 { + Default, + Settings, +}; + +enum class BootDisplayKind : u32 { + Default, + White, + Black, +}; + +enum class BackgroundKind : u32 { + Default, +}; + +enum class LeftStickMode : u32 { + Pointer, + Cursor, +}; + +enum class WebSessionBootMode : u32 { + AllForeground, + AllForegroundInitiallyHidden, +}; + +struct WebArgHeader { + u16 total_tlv_entries{}; + INSERT_PADDING_BYTES(2); + ShimKind shim_kind{}; +}; +static_assert(sizeof(WebArgHeader) == 0x8, "WebArgHeader has incorrect size."); + +struct WebArgInputTLV { + WebArgInputTLVType input_tlv_type{}; + u16 arg_data_size{}; + INSERT_PADDING_WORDS(1); +}; +static_assert(sizeof(WebArgInputTLV) == 0x8, "WebArgInputTLV has incorrect size."); + +struct WebArgOutputTLV { + WebArgOutputTLVType output_tlv_type{}; + u16 arg_data_size{}; + INSERT_PADDING_WORDS(1); +}; +static_assert(sizeof(WebArgOutputTLV) == 0x8, "WebArgOutputTLV has incorrect size."); + +struct WebCommonReturnValue { + WebExitReason exit_reason{}; + INSERT_PADDING_WORDS(1); + std::array<char, 0x1000> last_url{}; + u64 last_url_size{}; +}; +static_assert(sizeof(WebCommonReturnValue) == 0x1010, "WebCommonReturnValue has incorrect size."); + +using WebArgInputTLVMap = std::unordered_map<WebArgInputTLVType, std::vector<u8>>; + +} // namespace Service::AM::Applets diff --git a/src/core/hle/service/am/idle.cpp b/src/core/hle/service/am/idle.cpp index d256d57c8..6196773d5 100644 --- a/src/core/hle/service/am/idle.cpp +++ b/src/core/hle/service/am/idle.cpp @@ -6,7 +6,7 @@ namespace Service::AM { -IdleSys::IdleSys() : ServiceFramework{"idle:sys"} { +IdleSys::IdleSys(Core::System& system_) : ServiceFramework{system_, "idle:sys"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetAutoPowerDownEvent"}, diff --git a/src/core/hle/service/am/idle.h b/src/core/hle/service/am/idle.h index c44e856b1..e290c30b1 100644 --- a/src/core/hle/service/am/idle.h +++ b/src/core/hle/service/am/idle.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::AM { class IdleSys final : public ServiceFramework<IdleSys> { public: - explicit IdleSys(); + explicit IdleSys(Core::System& system_); ~IdleSys() override; }; diff --git a/src/core/hle/service/am/omm.cpp b/src/core/hle/service/am/omm.cpp index 37389ccda..55de67e1d 100644 --- a/src/core/hle/service/am/omm.cpp +++ b/src/core/hle/service/am/omm.cpp @@ -6,7 +6,7 @@ namespace Service::AM { -OMM::OMM() : ServiceFramework{"omm"} { +OMM::OMM(Core::System& system_) : ServiceFramework{system_, "omm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetOperationMode"}, diff --git a/src/core/hle/service/am/omm.h b/src/core/hle/service/am/omm.h index 59dc91b72..3766150fe 100644 --- a/src/core/hle/service/am/omm.h +++ b/src/core/hle/service/am/omm.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::AM { class OMM final : public ServiceFramework<OMM> { public: - explicit OMM(); + explicit OMM(Core::System& system_); ~OMM() override; }; diff --git a/src/core/hle/service/am/spsm.cpp b/src/core/hle/service/am/spsm.cpp index f27729ce7..95218d9ee 100644 --- a/src/core/hle/service/am/spsm.cpp +++ b/src/core/hle/service/am/spsm.cpp @@ -6,7 +6,7 @@ namespace Service::AM { -SPSM::SPSM() : ServiceFramework{"spsm"} { +SPSM::SPSM(Core::System& system_) : ServiceFramework{system_, "spsm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetState"}, diff --git a/src/core/hle/service/am/spsm.h b/src/core/hle/service/am/spsm.h index 3a0b979fa..04bbf9e68 100644 --- a/src/core/hle/service/am/spsm.h +++ b/src/core/hle/service/am/spsm.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::AM { class SPSM final : public ServiceFramework<SPSM> { public: - explicit SPSM(); + explicit SPSM(Core::System& system_); ~SPSM() override; }; diff --git a/src/core/hle/service/am/tcap.cpp b/src/core/hle/service/am/tcap.cpp index a75cbdda8..4d0971c03 100644 --- a/src/core/hle/service/am/tcap.cpp +++ b/src/core/hle/service/am/tcap.cpp @@ -6,7 +6,7 @@ namespace Service::AM { -TCAP::TCAP() : ServiceFramework{"tcap"} { +TCAP::TCAP(Core::System& system_) : ServiceFramework{system_, "tcap"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetContinuousHighSkinTemperatureEvent"}, diff --git a/src/core/hle/service/am/tcap.h b/src/core/hle/service/am/tcap.h index 2021b55d1..e9578f16e 100644 --- a/src/core/hle/service/am/tcap.h +++ b/src/core/hle/service/am/tcap.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::AM { class TCAP final : public ServiceFramework<TCAP> { public: - explicit TCAP(); + explicit TCAP(Core::System& system_); ~TCAP() override; }; diff --git a/src/core/hle/service/aoc/aoc_u.cpp b/src/core/hle/service/aoc/aoc_u.cpp index 8e79f707b..23e28565b 100644 --- a/src/core/hle/service/aoc/aoc_u.cpp +++ b/src/core/hle/service/aoc/aoc_u.cpp @@ -6,6 +6,8 @@ #include <numeric> #include <vector> #include "common/logging/log.h" +#include "core/core.h" +#include "core/file_sys/common_funcs.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/nca_metadata.h" @@ -22,11 +24,8 @@ namespace Service::AOC { -constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000; -constexpr u64 DLC_BASE_TO_AOC_ID = 0x1000; - static bool CheckAOCTitleIDMatchesBase(u64 title_id, u64 base) { - return (title_id & DLC_BASE_TITLE_ID_MASK) == base; + return FileSys::GetBaseTitleID(title_id) == base; } static std::vector<u64> AccumulateAOCTitleIDs(Core::System& system) { @@ -47,8 +46,64 @@ static std::vector<u64> AccumulateAOCTitleIDs(Core::System& system) { return add_on_content; } -AOC_U::AOC_U(Core::System& system) - : ServiceFramework("aoc:u"), add_on_content(AccumulateAOCTitleIDs(system)), system(system) { +class IPurchaseEventManager final : public ServiceFramework<IPurchaseEventManager> { +public: + explicit IPurchaseEventManager(Core::System& system_) + : ServiceFramework{system_, "IPurchaseEventManager"} { + // clang-format off + static const FunctionInfo functions[] = { + {0, &IPurchaseEventManager::SetDefaultDeliveryTarget, "SetDefaultDeliveryTarget"}, + {1, &IPurchaseEventManager::SetDeliveryTarget, "SetDeliveryTarget"}, + {2, &IPurchaseEventManager::GetPurchasedEventReadableHandle, "GetPurchasedEventReadableHandle"}, + {3, nullptr, "PopPurchasedProductInfo"}, + {4, nullptr, "PopPurchasedProductInfoWithUid"}, + }; + // clang-format on + + RegisterHandlers(functions); + + purchased_event = Kernel::WritableEvent::CreateEventPair( + system.Kernel(), "IPurchaseEventManager:PurchasedEvent"); + } + +private: + void SetDefaultDeliveryTarget(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const auto unknown_1 = rp.Pop<u64>(); + [[maybe_unused]] const auto unknown_2 = ctx.ReadBuffer(); + + LOG_WARNING(Service_AOC, "(STUBBED) called, unknown_1={}", unknown_1); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void SetDeliveryTarget(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const auto unknown_1 = rp.Pop<u64>(); + [[maybe_unused]] const auto unknown_2 = ctx.ReadBuffer(); + + LOG_WARNING(Service_AOC, "(STUBBED) called, unknown_1={}", unknown_1); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void GetPurchasedEventReadableHandle(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AOC, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushCopyObjects(purchased_event.readable); + } + + Kernel::EventPair purchased_event; +}; + +AOC_U::AOC_U(Core::System& system_) + : ServiceFramework{system_, "aoc:u"}, add_on_content{AccumulateAOCTitleIDs(system)} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CountAddOnContentByApplicationId"}, @@ -61,8 +116,8 @@ AOC_U::AOC_U(Core::System& system) {7, &AOC_U::PrepareAddOnContent, "PrepareAddOnContent"}, {8, &AOC_U::GetAddOnContentListChangedEvent, "GetAddOnContentListChangedEvent"}, {9, nullptr, "GetAddOnContentLostErrorCode"}, - {100, nullptr, "CreateEcPurchasedEventManager"}, - {101, nullptr, "CreatePermanentEcPurchasedEventManager"}, + {100, &AOC_U::CreateEcPurchasedEventManager, "CreateEcPurchasedEventManager"}, + {101, &AOC_U::CreatePermanentEcPurchasedEventManager, "CreatePermanentEcPurchasedEventManager"}, }; // clang-format on @@ -122,11 +177,11 @@ void AOC_U::ListAddOnContent(Kernel::HLERequestContext& ctx) { const auto& disabled = Settings::values.disabled_addons[current]; if (std::find(disabled.begin(), disabled.end(), "DLC") == disabled.end()) { for (u64 content_id : add_on_content) { - if ((content_id & DLC_BASE_TITLE_ID_MASK) != current) { + if (FileSys::GetBaseTitleID(content_id) != current) { continue; } - out.push_back(static_cast<u32>(content_id & 0x7FF)); + out.push_back(static_cast<u32>(FileSys::GetAOCID(content_id))); } } @@ -163,11 +218,12 @@ void AOC_U::GetAddOnContentBaseId(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); const auto title_id = system.CurrentProcess()->GetTitleID(); - FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto res = pm.GetControlMetadata(); if (res.first == nullptr) { - rb.Push(title_id + DLC_BASE_TO_AOC_ID); + rb.Push(FileSys::GetAOCBaseTitleID(title_id)); return; } @@ -199,6 +255,22 @@ void AOC_U::GetAddOnContentListChangedEvent(Kernel::HLERequestContext& ctx) { rb.PushCopyObjects(aoc_change_event.readable); } +void AOC_U::CreateEcPurchasedEventManager(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AOC, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IPurchaseEventManager>(system); +} + +void AOC_U::CreatePermanentEcPurchasedEventManager(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_AOC, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IPurchaseEventManager>(system); +} + void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { std::make_shared<AOC_U>(system)->InstallAsService(service_manager); } diff --git a/src/core/hle/service/aoc/aoc_u.h b/src/core/hle/service/aoc/aoc_u.h index 848b2f416..26ee51be0 100644 --- a/src/core/hle/service/aoc/aoc_u.h +++ b/src/core/hle/service/aoc/aoc_u.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class WritableEvent; } @@ -23,10 +27,11 @@ private: void GetAddOnContentBaseId(Kernel::HLERequestContext& ctx); void PrepareAddOnContent(Kernel::HLERequestContext& ctx); void GetAddOnContentListChangedEvent(Kernel::HLERequestContext& ctx); + void CreateEcPurchasedEventManager(Kernel::HLERequestContext& ctx); + void CreatePermanentEcPurchasedEventManager(Kernel::HLERequestContext& ctx); std::vector<u64> add_on_content; Kernel::EventPair aoc_change_event; - Core::System& system; }; /// Registers all AOC services with the specified service manager. diff --git a/src/core/hle/service/apm/apm.cpp b/src/core/hle/service/apm/apm.cpp index 85bbf5988..97d6619dd 100644 --- a/src/core/hle/service/apm/apm.cpp +++ b/src/core/hle/service/apm/apm.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/service/apm/apm.h" #include "core/hle/service/apm/interface.h" @@ -13,13 +14,14 @@ Module::~Module() = default; void InstallInterfaces(Core::System& system) { auto module_ = std::make_shared<Module>(); - std::make_shared<APM>(module_, system.GetAPMController(), "apm") + std::make_shared<APM>(system, module_, system.GetAPMController(), "apm") ->InstallAsService(system.ServiceManager()); - std::make_shared<APM>(module_, system.GetAPMController(), "apm:p") + std::make_shared<APM>(system, module_, system.GetAPMController(), "apm:p") ->InstallAsService(system.ServiceManager()); - std::make_shared<APM>(module_, system.GetAPMController(), "apm:am") + std::make_shared<APM>(system, module_, system.GetAPMController(), "apm:am") + ->InstallAsService(system.ServiceManager()); + std::make_shared<APM_Sys>(system, system.GetAPMController()) ->InstallAsService(system.ServiceManager()); - std::make_shared<APM_Sys>(system.GetAPMController())->InstallAsService(system.ServiceManager()); } } // namespace Service::APM diff --git a/src/core/hle/service/apm/apm.h b/src/core/hle/service/apm/apm.h index cf4c2bb11..691fe6c16 100644 --- a/src/core/hle/service/apm/apm.h +++ b/src/core/hle/service/apm/apm.h @@ -4,7 +4,9 @@ #pragma once -#include "core/hle/service/service.h" +namespace Core { +class System; +} namespace Service::APM { diff --git a/src/core/hle/service/apm/controller.cpp b/src/core/hle/service/apm/controller.cpp index 25a886238..03636642b 100644 --- a/src/core/hle/service/apm/controller.cpp +++ b/src/core/hle/service/apm/controller.cpp @@ -48,8 +48,7 @@ void Controller::SetPerformanceConfiguration(PerformanceMode mode, [config](const auto& entry) { return entry.first == config; }); if (iter == config_to_speed.cend()) { - LOG_ERROR(Service_APM, "Invalid performance configuration value provided: {}", - static_cast<u32>(config)); + LOG_ERROR(Service_APM, "Invalid performance configuration value provided: {}", config); return; } @@ -69,7 +68,8 @@ void Controller::SetFromCpuBoostMode(CpuBoostMode mode) { } PerformanceMode Controller::GetCurrentPerformanceMode() const { - return Settings::values.use_docked_mode ? PerformanceMode::Docked : PerformanceMode::Handheld; + return Settings::values.use_docked_mode.GetValue() ? PerformanceMode::Docked + : PerformanceMode::Handheld; } PerformanceConfiguration Controller::GetCurrentPerformanceConfiguration(PerformanceMode mode) { diff --git a/src/core/hle/service/apm/interface.cpp b/src/core/hle/service/apm/interface.cpp index 06f0f8edd..0bff97a37 100644 --- a/src/core/hle/service/apm/interface.cpp +++ b/src/core/hle/service/apm/interface.cpp @@ -12,7 +12,8 @@ namespace Service::APM { class ISession final : public ServiceFramework<ISession> { public: - ISession(Controller& controller) : ServiceFramework("ISession"), controller(controller) { + explicit ISession(Core::System& system_, Controller& controller_) + : ServiceFramework{system_, "ISession"}, controller{controller_} { static const FunctionInfo functions[] = { {0, &ISession::SetPerformanceConfiguration, "SetPerformanceConfiguration"}, {1, &ISession::GetPerformanceConfiguration, "GetPerformanceConfiguration"}, @@ -27,8 +28,7 @@ private: const auto mode = rp.PopEnum<PerformanceMode>(); const auto config = rp.PopEnum<PerformanceConfiguration>(); - LOG_DEBUG(Service_APM, "called mode={} config={}", static_cast<u32>(mode), - static_cast<u32>(config)); + LOG_DEBUG(Service_APM, "called mode={} config={}", mode, config); controller.SetPerformanceConfiguration(mode, config); @@ -40,7 +40,7 @@ private: IPC::RequestParser rp{ctx}; const auto mode = rp.PopEnum<PerformanceMode>(); - LOG_DEBUG(Service_APM, "called mode={}", static_cast<u32>(mode)); + LOG_DEBUG(Service_APM, "called mode={}", mode); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); @@ -50,12 +50,13 @@ private: Controller& controller; }; -APM::APM(std::shared_ptr<Module> apm, Controller& controller, const char* name) - : ServiceFramework(name), apm(std::move(apm)), controller(controller) { +APM::APM(Core::System& system_, std::shared_ptr<Module> apm_, Controller& controller_, + const char* name) + : ServiceFramework{system_, name}, apm(std::move(apm_)), controller{controller_} { static const FunctionInfo functions[] = { {0, &APM::OpenSession, "OpenSession"}, {1, &APM::GetPerformanceMode, "GetPerformanceMode"}, - {6, nullptr, "IsCpuOverclockEnabled"}, + {6, &APM::IsCpuOverclockEnabled, "IsCpuOverclockEnabled"}, }; RegisterHandlers(functions); } @@ -67,7 +68,7 @@ void APM::OpenSession(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISession>(controller); + rb.PushIpcInterface<ISession>(system, controller); } void APM::GetPerformanceMode(Kernel::HLERequestContext& ctx) { @@ -77,7 +78,16 @@ void APM::GetPerformanceMode(Kernel::HLERequestContext& ctx) { rb.PushEnum(controller.GetCurrentPerformanceMode()); } -APM_Sys::APM_Sys(Controller& controller) : ServiceFramework{"apm:sys"}, controller(controller) { +void APM::IsCpuOverclockEnabled(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_APM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(false); +} + +APM_Sys::APM_Sys(Core::System& system_, Controller& controller_) + : ServiceFramework{system_, "apm:sys"}, controller{controller_} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestPerformanceMode"}, @@ -101,14 +111,14 @@ void APM_Sys::GetPerformanceEvent(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISession>(controller); + rb.PushIpcInterface<ISession>(system, controller); } void APM_Sys::SetCpuBoostMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto mode = rp.PopEnum<CpuBoostMode>(); - LOG_DEBUG(Service_APM, "called, mode={:08X}", static_cast<u32>(mode)); + LOG_DEBUG(Service_APM, "called, mode={:08X}", mode); controller.SetFromCpuBoostMode(mode); diff --git a/src/core/hle/service/apm/interface.h b/src/core/hle/service/apm/interface.h index de1b89437..063ad5308 100644 --- a/src/core/hle/service/apm/interface.h +++ b/src/core/hle/service/apm/interface.h @@ -13,12 +13,14 @@ class Module; class APM final : public ServiceFramework<APM> { public: - explicit APM(std::shared_ptr<Module> apm, Controller& controller, const char* name); + explicit APM(Core::System& system_, std::shared_ptr<Module> apm_, Controller& controller_, + const char* name); ~APM() override; private: void OpenSession(Kernel::HLERequestContext& ctx); void GetPerformanceMode(Kernel::HLERequestContext& ctx); + void IsCpuOverclockEnabled(Kernel::HLERequestContext& ctx); std::shared_ptr<Module> apm; Controller& controller; @@ -26,7 +28,7 @@ private: class APM_Sys final : public ServiceFramework<APM_Sys> { public: - explicit APM_Sys(Controller& controller); + explicit APM_Sys(Core::System& system_, Controller& controller); ~APM_Sys() override; void SetCpuBoostMode(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/audio/audctl.cpp b/src/core/hle/service/audio/audctl.cpp index 6ddb547fb..84890be72 100644 --- a/src/core/hle/service/audio/audctl.cpp +++ b/src/core/hle/service/audio/audctl.cpp @@ -8,7 +8,7 @@ namespace Service::Audio { -AudCtl::AudCtl() : ServiceFramework{"audctl"} { +AudCtl::AudCtl(Core::System& system_) : ServiceFramework{system_, "audctl"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetTargetVolume"}, diff --git a/src/core/hle/service/audio/audctl.h b/src/core/hle/service/audio/audctl.h index c7fafc02e..15f6c77a0 100644 --- a/src/core/hle/service/audio/audctl.h +++ b/src/core/hle/service/audio/audctl.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Audio { class AudCtl final : public ServiceFramework<AudCtl> { public: - explicit AudCtl(); + explicit AudCtl(Core::System& system_); ~AudCtl() override; private: diff --git a/src/core/hle/service/audio/auddbg.cpp b/src/core/hle/service/audio/auddbg.cpp index 8fff3e4b4..6264e4bda 100644 --- a/src/core/hle/service/audio/auddbg.cpp +++ b/src/core/hle/service/audio/auddbg.cpp @@ -6,7 +6,7 @@ namespace Service::Audio { -AudDbg::AudDbg(const char* name) : ServiceFramework{name} { +AudDbg::AudDbg(Core::System& system_, const char* name) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestSuspendForDebug"}, diff --git a/src/core/hle/service/audio/auddbg.h b/src/core/hle/service/audio/auddbg.h index 6689f4759..d1653eedd 100644 --- a/src/core/hle/service/audio/auddbg.h +++ b/src/core/hle/service/audio/auddbg.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Audio { class AudDbg final : public ServiceFramework<AudDbg> { public: - explicit AudDbg(const char* name); + explicit AudDbg(Core::System& system_, const char* name); ~AudDbg() override; }; diff --git a/src/core/hle/service/audio/audin_a.cpp b/src/core/hle/service/audio/audin_a.cpp index ddd12f35e..79c3aa920 100644 --- a/src/core/hle/service/audio/audin_a.cpp +++ b/src/core/hle/service/audio/audin_a.cpp @@ -6,7 +6,7 @@ namespace Service::Audio { -AudInA::AudInA() : ServiceFramework{"audin:a"} { +AudInA::AudInA(Core::System& system_) : ServiceFramework{system_, "audin:a"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestSuspendAudioIns"}, diff --git a/src/core/hle/service/audio/audin_a.h b/src/core/hle/service/audio/audin_a.h index e7623bc29..15120a4b6 100644 --- a/src/core/hle/service/audio/audin_a.h +++ b/src/core/hle/service/audio/audin_a.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Audio { class AudInA final : public ServiceFramework<AudInA> { public: - explicit AudInA(); + explicit AudInA(Core::System& system_); ~AudInA() override; }; diff --git a/src/core/hle/service/audio/audin_u.cpp b/src/core/hle/service/audio/audin_u.cpp index 3e2299426..26a6deddf 100644 --- a/src/core/hle/service/audio/audin_u.cpp +++ b/src/core/hle/service/audio/audin_u.cpp @@ -11,7 +11,7 @@ namespace Service::Audio { class IAudioIn final : public ServiceFramework<IAudioIn> { public: - IAudioIn() : ServiceFramework("IAudioIn") { + explicit IAudioIn(Core::System& system_) : ServiceFramework{system_, "IAudioIn"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetAudioInState"}, @@ -36,7 +36,7 @@ public: } }; -AudInU::AudInU() : ServiceFramework("audin:u") { +AudInU::AudInU(Core::System& system_) : ServiceFramework{system_, "audin:u"} { // clang-format off static const FunctionInfo functions[] = { {0, &AudInU::ListAudioIns, "ListAudioIns"}, @@ -96,7 +96,7 @@ void AudInU::OpenInOutImpl(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 6, 0, 1}; rb.Push(RESULT_SUCCESS); rb.PushRaw<AudInOutParams>(params); - rb.PushIpcInterface<IAudioIn>(); + rb.PushIpcInterface<IAudioIn>(system); } void AudInU::OpenAudioIn(Kernel::HLERequestContext& ctx) { diff --git a/src/core/hle/service/audio/audin_u.h b/src/core/hle/service/audio/audin_u.h index a599f4a64..0d75ae5ac 100644 --- a/src/core/hle/service/audio/audin_u.h +++ b/src/core/hle/service/audio/audin_u.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -14,7 +18,7 @@ namespace Service::Audio { class AudInU final : public ServiceFramework<AudInU> { public: - explicit AudInU(); + explicit AudInU(Core::System& system_); ~AudInU() override; private: diff --git a/src/core/hle/service/audio/audio.cpp b/src/core/hle/service/audio/audio.cpp index 1781bec83..b3f24f9bb 100644 --- a/src/core/hle/service/audio/audio.cpp +++ b/src/core/hle/service/audio/audio.cpp @@ -20,22 +20,22 @@ namespace Service::Audio { void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { - std::make_shared<AudCtl>()->InstallAsService(service_manager); - std::make_shared<AudOutA>()->InstallAsService(service_manager); + std::make_shared<AudCtl>(system)->InstallAsService(service_manager); + std::make_shared<AudOutA>(system)->InstallAsService(service_manager); std::make_shared<AudOutU>(system)->InstallAsService(service_manager); - std::make_shared<AudInA>()->InstallAsService(service_manager); - std::make_shared<AudInU>()->InstallAsService(service_manager); - std::make_shared<AudRecA>()->InstallAsService(service_manager); - std::make_shared<AudRecU>()->InstallAsService(service_manager); - std::make_shared<AudRenA>()->InstallAsService(service_manager); + std::make_shared<AudInA>(system)->InstallAsService(service_manager); + std::make_shared<AudInU>(system)->InstallAsService(service_manager); + std::make_shared<AudRecA>(system)->InstallAsService(service_manager); + std::make_shared<AudRecU>(system)->InstallAsService(service_manager); + std::make_shared<AudRenA>(system)->InstallAsService(service_manager); std::make_shared<AudRenU>(system)->InstallAsService(service_manager); - std::make_shared<CodecCtl>()->InstallAsService(service_manager); - std::make_shared<HwOpus>()->InstallAsService(service_manager); + std::make_shared<CodecCtl>(system)->InstallAsService(service_manager); + std::make_shared<HwOpus>(system)->InstallAsService(service_manager); - std::make_shared<AudDbg>("audin:d")->InstallAsService(service_manager); - std::make_shared<AudDbg>("audout:d")->InstallAsService(service_manager); - std::make_shared<AudDbg>("audrec:d")->InstallAsService(service_manager); - std::make_shared<AudDbg>("audren:d")->InstallAsService(service_manager); + std::make_shared<AudDbg>(system, "audin:d")->InstallAsService(service_manager); + std::make_shared<AudDbg>(system, "audout:d")->InstallAsService(service_manager); + std::make_shared<AudDbg>(system, "audrec:d")->InstallAsService(service_manager); + std::make_shared<AudDbg>(system, "audren:d")->InstallAsService(service_manager); } } // namespace Service::Audio diff --git a/src/core/hle/service/audio/audout_a.cpp b/src/core/hle/service/audio/audout_a.cpp index 85febbca3..19825fd5d 100644 --- a/src/core/hle/service/audio/audout_a.cpp +++ b/src/core/hle/service/audio/audout_a.cpp @@ -6,7 +6,7 @@ namespace Service::Audio { -AudOutA::AudOutA() : ServiceFramework{"audout:a"} { +AudOutA::AudOutA(Core::System& system_) : ServiceFramework{system_, "audout:a"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestSuspendAudioOuts"}, diff --git a/src/core/hle/service/audio/audout_a.h b/src/core/hle/service/audio/audout_a.h index d65b66e8e..2043dfb77 100644 --- a/src/core/hle/service/audio/audout_a.h +++ b/src/core/hle/service/audio/audout_a.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Audio { class AudOutA final : public ServiceFramework<AudOutA> { public: - explicit AudOutA(); + explicit AudOutA(Core::System& system_); ~AudOutA() override; }; diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp index 9b4910e53..0cd797109 100644 --- a/src/core/hle/service/audio/audout_u.cpp +++ b/src/core/hle/service/audio/audout_u.cpp @@ -40,11 +40,11 @@ enum class AudioState : u32 { class IAudioOut final : public ServiceFramework<IAudioOut> { public: - IAudioOut(Core::System& system, AudoutParams audio_params, AudioCore::AudioOut& audio_core, - std::string&& device_name, std::string&& unique_name) - : ServiceFramework("IAudioOut"), audio_core(audio_core), - device_name(std::move(device_name)), - audio_params(audio_params), main_memory{system.Memory()} { + IAudioOut(Core::System& system_, AudoutParams audio_params_, AudioCore::AudioOut& audio_core_, + std::string&& device_name_, std::string&& unique_name) + : ServiceFramework{system_, "IAudioOut"}, audio_core{audio_core_}, + device_name{std::move(device_name_)}, audio_params{audio_params_}, main_memory{ + system.Memory()} { // clang-format off static const FunctionInfo functions[] = { {0, &IAudioOut::GetAudioOutState, "GetAudioOutState"}, @@ -70,8 +70,10 @@ public: Kernel::WritableEvent::CreateEventPair(system.Kernel(), "IAudioOutBufferReleased"); stream = audio_core.OpenStream(system.CoreTiming(), audio_params.sample_rate, - audio_params.channel_count, std::move(unique_name), - [this] { buffer_event.writable->Signal(); }); + audio_params.channel_count, std::move(unique_name), [this] { + const auto guard = LockService(); + buffer_event.writable->Signal(); + }); } private: @@ -213,7 +215,7 @@ private: Core::Memory::Memory& main_memory; }; -AudOutU::AudOutU(Core::System& system_) : ServiceFramework("audout:u"), system{system_} { +AudOutU::AudOutU(Core::System& system_) : ServiceFramework{system_, "audout:u"} { // clang-format off static const FunctionInfo functions[] = { {0, &AudOutU::ListAudioOutsImpl, "ListAudioOuts"}, diff --git a/src/core/hle/service/audio/audout_u.h b/src/core/hle/service/audio/audout_u.h index c9f532ccd..f7ae2f2bf 100644 --- a/src/core/hle/service/audio/audout_u.h +++ b/src/core/hle/service/audio/audout_u.h @@ -34,8 +34,6 @@ private: std::vector<std::shared_ptr<IAudioOut>> audio_out_interfaces; std::unique_ptr<AudioCore::AudioOut> audio_core; - - Core::System& system; }; } // namespace Service::Audio diff --git a/src/core/hle/service/audio/audrec_a.cpp b/src/core/hle/service/audio/audrec_a.cpp index ce1bfb48d..c5ab7cad4 100644 --- a/src/core/hle/service/audio/audrec_a.cpp +++ b/src/core/hle/service/audio/audrec_a.cpp @@ -6,7 +6,7 @@ namespace Service::Audio { -AudRecA::AudRecA() : ServiceFramework{"audrec:a"} { +AudRecA::AudRecA(Core::System& system_) : ServiceFramework{system_, "audrec:a"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestSuspendFinalOutputRecorders"}, diff --git a/src/core/hle/service/audio/audrec_a.h b/src/core/hle/service/audio/audrec_a.h index 384d24c69..2cce90b1d 100644 --- a/src/core/hle/service/audio/audrec_a.h +++ b/src/core/hle/service/audio/audrec_a.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Audio { class AudRecA final : public ServiceFramework<AudRecA> { public: - explicit AudRecA(); + explicit AudRecA(Core::System& system_); ~AudRecA() override; }; diff --git a/src/core/hle/service/audio/audrec_u.cpp b/src/core/hle/service/audio/audrec_u.cpp index 1a5aed9ed..eb5c63c62 100644 --- a/src/core/hle/service/audio/audrec_u.cpp +++ b/src/core/hle/service/audio/audrec_u.cpp @@ -8,7 +8,8 @@ namespace Service::Audio { class IFinalOutputRecorder final : public ServiceFramework<IFinalOutputRecorder> { public: - IFinalOutputRecorder() : ServiceFramework("IFinalOutputRecorder") { + explicit IFinalOutputRecorder(Core::System& system_) + : ServiceFramework{system_, "IFinalOutputRecorder"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetFinalOutputRecorderState"}, @@ -29,7 +30,7 @@ public: } }; -AudRecU::AudRecU() : ServiceFramework("audrec:u") { +AudRecU::AudRecU(Core::System& system_) : ServiceFramework{system_, "audrec:u"} { static const FunctionInfo functions[] = { {0, nullptr, "OpenFinalOutputRecorder"}, }; diff --git a/src/core/hle/service/audio/audrec_u.h b/src/core/hle/service/audio/audrec_u.h index ca3d638e8..f79d49e5c 100644 --- a/src/core/hle/service/audio/audrec_u.h +++ b/src/core/hle/service/audio/audrec_u.h @@ -6,15 +6,15 @@ #include "core/hle/service/service.h" -namespace Kernel { -class HLERequestContext; +namespace Core { +class System; } namespace Service::Audio { class AudRecU final : public ServiceFramework<AudRecU> { public: - explicit AudRecU(); + explicit AudRecU(Core::System& system_); ~AudRecU() override; }; diff --git a/src/core/hle/service/audio/audren_a.cpp b/src/core/hle/service/audio/audren_a.cpp index edb66d985..5e9f866f0 100644 --- a/src/core/hle/service/audio/audren_a.cpp +++ b/src/core/hle/service/audio/audren_a.cpp @@ -6,7 +6,7 @@ namespace Service::Audio { -AudRenA::AudRenA() : ServiceFramework{"audren:a"} { +AudRenA::AudRenA(Core::System& system_) : ServiceFramework{system_, "audren:a"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestSuspendAudioRenderers"}, diff --git a/src/core/hle/service/audio/audren_a.h b/src/core/hle/service/audio/audren_a.h index 81fef0ffe..5d0a626ad 100644 --- a/src/core/hle/service/audio/audren_a.h +++ b/src/core/hle/service/audio/audren_a.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Audio { class AudRenA final : public ServiceFramework<AudRenA> { public: - explicit AudRenA(); + explicit AudRenA(Core::System& system_); ~AudRenA() override; }; diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp index a2d3ded7b..c5c22d053 100644 --- a/src/core/hle/service/audio/audren_u.cpp +++ b/src/core/hle/service/audio/audren_u.cpp @@ -28,7 +28,7 @@ class IAudioRenderer final : public ServiceFramework<IAudioRenderer> { public: explicit IAudioRenderer(Core::System& system, AudioCommon::AudioRendererParameter audren_params, const std::size_t instance_number) - : ServiceFramework("IAudioRenderer") { + : ServiceFramework{system, "IAudioRenderer"} { // clang-format off static const FunctionInfo functions[] = { {0, &IAudioRenderer::GetSampleRate, "GetSampleRate"}, @@ -49,16 +49,16 @@ public: system_event = Kernel::WritableEvent::CreateEventPair(system.Kernel(), "IAudioRenderer:SystemEvent"); - renderer = std::make_unique<AudioCore::AudioRenderer>(system.CoreTiming(), system.Memory(), - audren_params, system_event.writable, - instance_number); + renderer = std::make_unique<AudioCore::AudioRenderer>( + system.CoreTiming(), system.Memory(), audren_params, + [this]() { + const auto guard = LockService(); + system_event.writable->Signal(); + }, + instance_number); } private: - void UpdateAudioCallback() { - system_event.writable->Signal(); - } - void GetSampleRate(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Audio, "called"); @@ -167,8 +167,8 @@ private: class IAudioDevice final : public ServiceFramework<IAudioDevice> { public: - explicit IAudioDevice(Core::System& system, u32_le revision_num) - : ServiceFramework("IAudioDevice"), revision{revision_num} { + explicit IAudioDevice(Core::System& system_, u32_le revision_num) + : ServiceFramework{system_, "IAudioDevice"}, revision{revision_num} { static const FunctionInfo functions[] = { {0, &IAudioDevice::ListAudioDeviceName, "ListAudioDeviceName"}, {1, &IAudioDevice::SetAudioDeviceOutputVolume, "SetAudioDeviceOutputVolume"}, @@ -325,7 +325,7 @@ private: }; // namespace Audio -AudRenU::AudRenU(Core::System& system_) : ServiceFramework("audren:u"), system{system_} { +AudRenU::AudRenU(Core::System& system_) : ServiceFramework{system_, "audren:u"} { // clang-format off static const FunctionInfo functions[] = { {0, &AudRenU::OpenAudioRenderer, "OpenAudioRenderer"}, diff --git a/src/core/hle/service/audio/audren_u.h b/src/core/hle/service/audio/audren_u.h index 4e0ccc792..d693dc406 100644 --- a/src/core/hle/service/audio/audren_u.h +++ b/src/core/hle/service/audio/audren_u.h @@ -31,7 +31,6 @@ private: void OpenAudioRendererImpl(Kernel::HLERequestContext& ctx); std::size_t audren_instance_count = 0; - Core::System& system; }; // Describes a particular audio feature that may be supported in a particular revision. diff --git a/src/core/hle/service/audio/codecctl.cpp b/src/core/hle/service/audio/codecctl.cpp index c6864146d..94afec1b6 100644 --- a/src/core/hle/service/audio/codecctl.cpp +++ b/src/core/hle/service/audio/codecctl.cpp @@ -2,14 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "common/logging/log.h" -#include "core/hle/ipc_helpers.h" -#include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/audio/codecctl.h" namespace Service::Audio { -CodecCtl::CodecCtl() : ServiceFramework("codecctl") { +CodecCtl::CodecCtl(Core::System& system_) : ServiceFramework{system_, "codecctl"} { static const FunctionInfo functions[] = { {0, nullptr, "InitializeCodecController"}, {1, nullptr, "FinalizeCodecController"}, diff --git a/src/core/hle/service/audio/codecctl.h b/src/core/hle/service/audio/codecctl.h index 2fe75b6e2..58e53259e 100644 --- a/src/core/hle/service/audio/codecctl.h +++ b/src/core/hle/service/audio/codecctl.h @@ -6,15 +6,15 @@ #include "core/hle/service/service.h" -namespace Kernel { -class HLERequestContext; +namespace Core { +class System; } namespace Service::Audio { class CodecCtl final : public ServiceFramework<CodecCtl> { public: - explicit CodecCtl(); + explicit CodecCtl(Core::System& system_); ~CodecCtl() override; }; diff --git a/src/core/hle/service/audio/hwopus.cpp b/src/core/hle/service/audio/hwopus.cpp index f1d81602c..ea3414fd2 100644 --- a/src/core/hle/service/audio/hwopus.cpp +++ b/src/core/hle/service/audio/hwopus.cpp @@ -160,8 +160,9 @@ private: class IHardwareOpusDecoderManager final : public ServiceFramework<IHardwareOpusDecoderManager> { public: - explicit IHardwareOpusDecoderManager(OpusDecoderState decoder_state) - : ServiceFramework("IHardwareOpusDecoderManager"), decoder_state{std::move(decoder_state)} { + explicit IHardwareOpusDecoderManager(Core::System& system_, OpusDecoderState decoder_state) + : ServiceFramework{system_, "IHardwareOpusDecoderManager"}, decoder_state{ + std::move(decoder_state)} { // clang-format off static const FunctionInfo functions[] = { {0, &IHardwareOpusDecoderManager::DecodeInterleavedOld, "DecodeInterleavedOld"}, @@ -287,10 +288,10 @@ void HwOpus::OpenOpusDecoder(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); rb.PushIpcInterface<IHardwareOpusDecoderManager>( - OpusDecoderState{std::move(decoder), sample_rate, channel_count}); + system, OpusDecoderState{std::move(decoder), sample_rate, channel_count}); } -HwOpus::HwOpus() : ServiceFramework("hwopus") { +HwOpus::HwOpus(Core::System& system_) : ServiceFramework{system_, "hwopus"} { static const FunctionInfo functions[] = { {0, &HwOpus::OpenOpusDecoder, "OpenOpusDecoder"}, {1, &HwOpus::GetWorkBufferSize, "GetWorkBufferSize"}, diff --git a/src/core/hle/service/audio/hwopus.h b/src/core/hle/service/audio/hwopus.h index 602ede8ba..4f921f18e 100644 --- a/src/core/hle/service/audio/hwopus.h +++ b/src/core/hle/service/audio/hwopus.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Audio { class HwOpus final : public ServiceFramework<HwOpus> { public: - explicit HwOpus(); + explicit HwOpus(Core::System& system_); ~HwOpus() override; private: diff --git a/src/core/hle/service/bcat/backend/backend.cpp b/src/core/hle/service/bcat/backend/backend.cpp index def3410cc..174388445 100644 --- a/src/core/hle/service/bcat/backend/backend.cpp +++ b/src/core/hle/service/bcat/backend/backend.cpp @@ -84,7 +84,7 @@ void ProgressServiceBackend::FinishDownload(ResultCode result) { void ProgressServiceBackend::SignalUpdate() const { if (need_hle_lock) { - std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock); + std::lock_guard lock(HLE::g_hle_lock); event.writable->Signal(); } else { event.writable->Signal(); diff --git a/src/core/hle/service/bcat/backend/boxcat.cpp b/src/core/hle/service/bcat/backend/boxcat.cpp index ca021a99f..e43f3f47f 100644 --- a/src/core/hle/service/bcat/backend/boxcat.cpp +++ b/src/core/hle/service/bcat/backend/boxcat.cpp @@ -196,7 +196,9 @@ private: const std::string& content_type_name) { if (client == nullptr) { client = std::make_unique<httplib::SSLClient>(BOXCAT_HOSTNAME, PORT); - client->set_timeout_sec(timeout_seconds); + client->set_connection_timeout(timeout_seconds); + client->set_read_timeout(timeout_seconds); + client->set_write_timeout(timeout_seconds); } httplib::Headers headers{ @@ -255,7 +257,7 @@ private: return out; } - std::unique_ptr<httplib::Client> client; + std::unique_ptr<httplib::SSLClient> client; std::string path; u64 title_id; u64 build_id; @@ -443,13 +445,25 @@ std::optional<std::vector<u8>> Boxcat::GetLaunchParameter(TitleIDVersion title) Boxcat::StatusResult Boxcat::GetStatus(std::optional<std::string>& global, std::map<std::string, EventStatus>& games) { httplib::SSLClient client{BOXCAT_HOSTNAME, static_cast<int>(PORT)}; - client.set_timeout_sec(static_cast<int>(TIMEOUT_SECONDS)); + client.set_connection_timeout(static_cast<int>(TIMEOUT_SECONDS)); + client.set_read_timeout(static_cast<int>(TIMEOUT_SECONDS)); + client.set_write_timeout(static_cast<int>(TIMEOUT_SECONDS)); httplib::Headers headers{ {std::string("Game-Assets-API-Version"), std::string(BOXCAT_API_VERSION)}, {std::string("Boxcat-Client-Type"), std::string(BOXCAT_CLIENT_TYPE)}, }; + if (!client.is_valid()) { + LOG_ERROR(Service_BCAT, "Client is invalid, going offline!"); + return StatusResult::Offline; + } + + if (!client.is_socket_open()) { + LOG_ERROR(Service_BCAT, "Failed to open socket, going offline!"); + return StatusResult::Offline; + } + const auto response = client.Get(BOXCAT_PATHNAME_EVENTS, headers); if (response == nullptr) return StatusResult::Offline; @@ -469,7 +483,7 @@ Boxcat::StatusResult Boxcat::GetStatus(std::optional<std::string>& global, global = json["global"].get<std::string>(); if (json["games"].is_array()) { - for (const auto object : json["games"]) { + for (const auto& object : json["games"]) { if (object.is_object() && object.find("name") != object.end()) { EventStatus detail{}; if (object["header"].is_string()) { diff --git a/src/core/hle/service/bcat/module.cpp b/src/core/hle/service/bcat/module.cpp index db0e06ca1..b8696a395 100644 --- a/src/core/hle/service/bcat/module.cpp +++ b/src/core/hle/service/bcat/module.cpp @@ -8,6 +8,7 @@ #include "common/hex_util.h" #include "common/logging/log.h" #include "common/string_util.h" +#include "core/core.h" #include "core/file_sys/vfs.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/process.h" @@ -87,9 +88,11 @@ struct DeliveryCacheDirectoryEntry { class IDeliveryCacheProgressService final : public ServiceFramework<IDeliveryCacheProgressService> { public: - IDeliveryCacheProgressService(std::shared_ptr<Kernel::ReadableEvent> event, - const DeliveryCacheProgressImpl& impl) - : ServiceFramework{"IDeliveryCacheProgressService"}, event(std::move(event)), impl(impl) { + explicit IDeliveryCacheProgressService(Core::System& system_, + std::shared_ptr<Kernel::ReadableEvent> event_, + const DeliveryCacheProgressImpl& impl_) + : ServiceFramework{system_, "IDeliveryCacheProgressService"}, event{std::move(event_)}, + impl{impl_} { // clang-format off static const FunctionInfo functions[] = { {0, &IDeliveryCacheProgressService::GetEvent, "GetEvent"}, @@ -125,7 +128,7 @@ private: class IBcatService final : public ServiceFramework<IBcatService> { public: explicit IBcatService(Core::System& system_, Backend& backend_) - : ServiceFramework("IBcatService"), system{system_}, backend{backend_}, + : ServiceFramework{system_, "IBcatService"}, backend{backend_}, progress{{ ProgressServiceBackend{system_.Kernel(), "Normal"}, ProgressServiceBackend{system_.Kernel(), "Directory"}, @@ -170,7 +173,7 @@ private: std::shared_ptr<IDeliveryCacheProgressService> CreateProgressService(SyncType type) { auto& backend{progress.at(static_cast<std::size_t>(type))}; - return std::make_shared<IDeliveryCacheProgressService>(backend.GetEvent(), + return std::make_shared<IDeliveryCacheProgressService>(system, backend.GetEvent(), backend.GetImpl()); } @@ -260,7 +263,6 @@ private: rb.Push(RESULT_SUCCESS); } - Core::System& system; Backend& backend; std::array<ProgressServiceBackend, static_cast<std::size_t>(SyncType::Count)> progress; @@ -276,8 +278,8 @@ void Module::Interface::CreateBcatService(Kernel::HLERequestContext& ctx) { class IDeliveryCacheFileService final : public ServiceFramework<IDeliveryCacheFileService> { public: - IDeliveryCacheFileService(FileSys::VirtualDir root_) - : ServiceFramework{"IDeliveryCacheFileService"}, root(std::move(root_)) { + explicit IDeliveryCacheFileService(Core::System& system_, FileSys::VirtualDir root_) + : ServiceFramework{system_, "IDeliveryCacheFileService"}, root(std::move(root_)) { // clang-format off static const FunctionInfo functions[] = { {0, &IDeliveryCacheFileService::Open, "Open"}, @@ -393,8 +395,8 @@ private: class IDeliveryCacheDirectoryService final : public ServiceFramework<IDeliveryCacheDirectoryService> { public: - IDeliveryCacheDirectoryService(FileSys::VirtualDir root_) - : ServiceFramework{"IDeliveryCacheDirectoryService"}, root(std::move(root_)) { + explicit IDeliveryCacheDirectoryService(Core::System& system_, FileSys::VirtualDir root_) + : ServiceFramework{system_, "IDeliveryCacheDirectoryService"}, root(std::move(root_)) { // clang-format off static const FunctionInfo functions[] = { {0, &IDeliveryCacheDirectoryService::Open, "Open"}, @@ -491,8 +493,8 @@ private: class IDeliveryCacheStorageService final : public ServiceFramework<IDeliveryCacheStorageService> { public: - IDeliveryCacheStorageService(FileSys::VirtualDir root_) - : ServiceFramework{"IDeliveryCacheStorageService"}, root(std::move(root_)) { + explicit IDeliveryCacheStorageService(Core::System& system_, FileSys::VirtualDir root_) + : ServiceFramework{system_, "IDeliveryCacheStorageService"}, root(std::move(root_)) { // clang-format off static const FunctionInfo functions[] = { {0, &IDeliveryCacheStorageService::CreateFileService, "CreateFileService"}, @@ -517,7 +519,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDeliveryCacheFileService>(root); + rb.PushIpcInterface<IDeliveryCacheFileService>(system, root); } void CreateDirectoryService(Kernel::HLERequestContext& ctx) { @@ -525,7 +527,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDeliveryCacheDirectoryService>(root); + rb.PushIpcInterface<IDeliveryCacheDirectoryService>(system, root); } void EnumerateDeliveryCacheDirectory(Kernel::HLERequestContext& ctx) { @@ -550,10 +552,10 @@ private: void Module::Interface::CreateDeliveryCacheStorageService(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_BCAT, "called"); + const auto title_id = system.CurrentProcess()->GetTitleID(); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDeliveryCacheStorageService>( - fsc.GetBCATDirectory(system.CurrentProcess()->GetTitleID())); + rb.PushIpcInterface<IDeliveryCacheStorageService>(system, fsc.GetBCATDirectory(title_id)); } void Module::Interface::CreateDeliveryCacheStorageServiceWithApplicationId( @@ -565,7 +567,7 @@ void Module::Interface::CreateDeliveryCacheStorageServiceWithApplicationId( IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDeliveryCacheStorageService>(fsc.GetBCATDirectory(title_id)); + rb.PushIpcInterface<IDeliveryCacheStorageService>(system, fsc.GetBCATDirectory(title_id)); } std::unique_ptr<Backend> CreateBackendFromSettings([[maybe_unused]] Core::System& system, @@ -581,10 +583,9 @@ std::unique_ptr<Backend> CreateBackendFromSettings([[maybe_unused]] Core::System Module::Interface::Interface(Core::System& system_, std::shared_ptr<Module> module_, FileSystem::FileSystemController& fsc_, const char* name) - : ServiceFramework(name), fsc{fsc_}, module{std::move(module_)}, + : ServiceFramework{system_, name}, fsc{fsc_}, module{std::move(module_)}, backend{CreateBackendFromSettings(system_, - [&fsc_](u64 tid) { return fsc_.GetBCATDirectory(tid); })}, - system{system_} {} + [&fsc_](u64 tid) { return fsc_.GetBCATDirectory(tid); })} {} Module::Interface::~Interface() = default; diff --git a/src/core/hle/service/bcat/module.h b/src/core/hle/service/bcat/module.h index e4ba23ba0..738731c06 100644 --- a/src/core/hle/service/bcat/module.h +++ b/src/core/hle/service/bcat/module.h @@ -37,9 +37,6 @@ public: std::shared_ptr<Module> module; std::unique_ptr<Backend> backend; - - private: - Core::System& system; }; }; diff --git a/src/core/hle/service/bpc/bpc.cpp b/src/core/hle/service/bpc/bpc.cpp index fac6b2f9c..e4630320e 100644 --- a/src/core/hle/service/bpc/bpc.cpp +++ b/src/core/hle/service/bpc/bpc.cpp @@ -12,7 +12,7 @@ namespace Service::BPC { class BPC final : public ServiceFramework<BPC> { public: - explicit BPC() : ServiceFramework{"bpc"} { + explicit BPC(Core::System& system_) : ServiceFramework{system_, "bpc"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "ShutdownSystem"}, @@ -40,7 +40,7 @@ public: class BPC_R final : public ServiceFramework<BPC_R> { public: - explicit BPC_R() : ServiceFramework{"bpc:r"} { + explicit BPC_R(Core::System& system_) : ServiceFramework{system_, "bpc:r"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetRtcTime"}, @@ -55,9 +55,9 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<BPC>()->InstallAsService(sm); - std::make_shared<BPC_R>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<BPC>(system)->InstallAsService(sm); + std::make_shared<BPC_R>(system)->InstallAsService(sm); } } // namespace Service::BPC diff --git a/src/core/hle/service/bpc/bpc.h b/src/core/hle/service/bpc/bpc.h index eaa37be8d..6ec25aa9b 100644 --- a/src/core/hle/service/bpc/bpc.h +++ b/src/core/hle/service/bpc/bpc.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::BPC { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::BPC diff --git a/src/core/hle/service/btdrv/btdrv.cpp b/src/core/hle/service/btdrv/btdrv.cpp index f311afa2f..2de86f1f1 100644 --- a/src/core/hle/service/btdrv/btdrv.cpp +++ b/src/core/hle/service/btdrv/btdrv.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/kernel/kernel.h" @@ -16,7 +17,7 @@ namespace Service::BtDrv { class Bt final : public ServiceFramework<Bt> { public: - explicit Bt(Core::System& system) : ServiceFramework{"bt"} { + explicit Bt(Core::System& system_) : ServiceFramework{system_, "bt"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "LeClientReadCharacteristic"}, @@ -51,7 +52,7 @@ private: class BtDrv final : public ServiceFramework<BtDrv> { public: - explicit BtDrv() : ServiceFramework{"btdrv"} { + explicit BtDrv(Core::System& system_) : ServiceFramework{system_, "btdrv"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "InitializeBluetoothDriver"}, @@ -165,7 +166,7 @@ public: }; void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { - std::make_shared<BtDrv>()->InstallAsService(sm); + std::make_shared<BtDrv>(system)->InstallAsService(sm); std::make_shared<Bt>(system)->InstallAsService(sm); } diff --git a/src/core/hle/service/btm/btm.cpp b/src/core/hle/service/btm/btm.cpp index 0d251c6d0..38b55300e 100644 --- a/src/core/hle/service/btm/btm.cpp +++ b/src/core/hle/service/btm/btm.cpp @@ -5,6 +5,7 @@ #include <memory> #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/kernel/kernel.h" @@ -17,7 +18,7 @@ namespace Service::BTM { class IBtmUserCore final : public ServiceFramework<IBtmUserCore> { public: - explicit IBtmUserCore(Core::System& system) : ServiceFramework{"IBtmUserCore"} { + explicit IBtmUserCore(Core::System& system_) : ServiceFramework{system_, "IBtmUserCore"} { // clang-format off static const FunctionInfo functions[] = { {0, &IBtmUserCore::AcquireBleScanEvent, "AcquireBleScanEvent"}, @@ -106,7 +107,7 @@ private: class BTM_USR final : public ServiceFramework<BTM_USR> { public: - explicit BTM_USR(Core::System& system) : ServiceFramework{"btm:u"}, system(system) { + explicit BTM_USR(Core::System& system_) : ServiceFramework{system_, "btm:u"} { // clang-format off static const FunctionInfo functions[] = { {0, &BTM_USR::GetCore, "GetCore"}, @@ -123,13 +124,11 @@ private: rb.Push(RESULT_SUCCESS); rb.PushIpcInterface<IBtmUserCore>(system); } - - Core::System& system; }; class BTM final : public ServiceFramework<BTM> { public: - explicit BTM() : ServiceFramework{"btm"} { + explicit BTM(Core::System& system_) : ServiceFramework{system_, "btm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetState"}, @@ -206,7 +205,7 @@ public: class BTM_DBG final : public ServiceFramework<BTM_DBG> { public: - explicit BTM_DBG() : ServiceFramework{"btm:dbg"} { + explicit BTM_DBG(Core::System& system_) : ServiceFramework{system_, "btm:dbg"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "AcquireDiscoveryEvent"}, @@ -231,7 +230,7 @@ public: class IBtmSystemCore final : public ServiceFramework<IBtmSystemCore> { public: - explicit IBtmSystemCore() : ServiceFramework{"IBtmSystemCore"} { + explicit IBtmSystemCore(Core::System& system_) : ServiceFramework{system_, "IBtmSystemCore"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "StartGamepadPairing"}, @@ -253,7 +252,7 @@ public: class BTM_SYS final : public ServiceFramework<BTM_SYS> { public: - explicit BTM_SYS() : ServiceFramework{"btm:sys"} { + explicit BTM_SYS(Core::System& system_) : ServiceFramework{system_, "btm:sys"} { // clang-format off static const FunctionInfo functions[] = { {0, &BTM_SYS::GetCore, "GetCore"}, @@ -269,14 +268,14 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IBtmSystemCore>(); + rb.PushIpcInterface<IBtmSystemCore>(system); } }; void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { - std::make_shared<BTM>()->InstallAsService(sm); - std::make_shared<BTM_DBG>()->InstallAsService(sm); - std::make_shared<BTM_SYS>()->InstallAsService(sm); + std::make_shared<BTM>(system)->InstallAsService(sm); + std::make_shared<BTM_DBG>(system)->InstallAsService(sm); + std::make_shared<BTM_SYS>(system)->InstallAsService(sm); std::make_shared<BTM_USR>(system)->InstallAsService(sm); } diff --git a/src/core/hle/service/caps/caps.cpp b/src/core/hle/service/caps/caps.cpp index ba5749b84..5b7fe8e9b 100644 --- a/src/core/hle/service/caps/caps.cpp +++ b/src/core/hle/service/caps/caps.cpp @@ -13,13 +13,13 @@ namespace Service::Capture { -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<CAPS_A>()->InstallAsService(sm); - std::make_shared<CAPS_C>()->InstallAsService(sm); - std::make_shared<CAPS_U>()->InstallAsService(sm); - std::make_shared<CAPS_SC>()->InstallAsService(sm); - std::make_shared<CAPS_SS>()->InstallAsService(sm); - std::make_shared<CAPS_SU>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<CAPS_A>(system)->InstallAsService(sm); + std::make_shared<CAPS_C>(system)->InstallAsService(sm); + std::make_shared<CAPS_U>(system)->InstallAsService(sm); + std::make_shared<CAPS_SC>(system)->InstallAsService(sm); + std::make_shared<CAPS_SS>(system)->InstallAsService(sm); + std::make_shared<CAPS_SU>(system)->InstallAsService(sm); } } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps.h b/src/core/hle/service/caps/caps.h index b8c67b6e2..3c4290c88 100644 --- a/src/core/hle/service/caps/caps.h +++ b/src/core/hle/service/caps/caps.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } @@ -87,6 +91,6 @@ static_assert(sizeof(ApplicationAlbumFileEntry) == 0x30, "ApplicationAlbumFileEntry has incorrect size."); /// Registers all Capture services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_a.cpp b/src/core/hle/service/caps/caps_a.cpp index a0a3b2ae3..1fe4f0e14 100644 --- a/src/core/hle/service/caps/caps_a.cpp +++ b/src/core/hle/service/caps/caps_a.cpp @@ -8,7 +8,8 @@ namespace Service::Capture { class IAlbumAccessorSession final : public ServiceFramework<IAlbumAccessorSession> { public: - explicit IAlbumAccessorSession() : ServiceFramework{"IAlbumAccessorSession"} { + explicit IAlbumAccessorSession(Core::System& system_) + : ServiceFramework{system_, "IAlbumAccessorSession"} { // clang-format off static const FunctionInfo functions[] = { {2001, nullptr, "OpenAlbumMovieReadStream"}, @@ -26,7 +27,7 @@ public: } }; -CAPS_A::CAPS_A() : ServiceFramework("caps:a") { +CAPS_A::CAPS_A(Core::System& system_) : ServiceFramework{system_, "caps:a"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetAlbumFileCount"}, diff --git a/src/core/hle/service/caps/caps_a.h b/src/core/hle/service/caps/caps_a.h index cb93aad5b..389cc6dbe 100644 --- a/src/core/hle/service/caps/caps_a.h +++ b/src/core/hle/service/caps/caps_a.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -14,7 +18,7 @@ namespace Service::Capture { class CAPS_A final : public ServiceFramework<CAPS_A> { public: - explicit CAPS_A(); + explicit CAPS_A(Core::System& system_); ~CAPS_A() override; }; diff --git a/src/core/hle/service/caps/caps_c.cpp b/src/core/hle/service/caps/caps_c.cpp index ab17a187e..45c1c9d30 100644 --- a/src/core/hle/service/caps/caps_c.cpp +++ b/src/core/hle/service/caps/caps_c.cpp @@ -2,13 +2,16 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/logging/log.h" +#include "core/hle/ipc_helpers.h" #include "core/hle/service/caps/caps_c.h" namespace Service::Capture { class IAlbumControlSession final : public ServiceFramework<IAlbumControlSession> { public: - explicit IAlbumControlSession() : ServiceFramework{"IAlbumControlSession"} { + explicit IAlbumControlSession(Core::System& system_) + : ServiceFramework{system_, "IAlbumControlSession"} { // clang-format off static const FunctionInfo functions[] = { {2001, nullptr, "OpenAlbumMovieReadStream"}, @@ -42,12 +45,12 @@ public: } }; -CAPS_C::CAPS_C() : ServiceFramework("caps:c") { +CAPS_C::CAPS_C(Core::System& system_) : ServiceFramework{system_, "caps:c"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "CaptureRawImage"}, {2, nullptr, "CaptureRawImageWithTimeout"}, - {33, nullptr, "Unknown33"}, + {33, &CAPS_C::SetShimLibraryVersion, "SetShimLibraryVersion"}, {1001, nullptr, "RequestTakingScreenShot"}, {1002, nullptr, "RequestTakingScreenShotWithTimeout"}, {1011, nullptr, "NotifyTakingScreenShotRefused"}, @@ -72,4 +75,16 @@ CAPS_C::CAPS_C() : ServiceFramework("caps:c") { CAPS_C::~CAPS_C() = default; +void CAPS_C::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto library_version{rp.Pop<u64>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}", + library_version, applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_c.h b/src/core/hle/service/caps/caps_c.h index a9d028689..c6d1dfdce 100644 --- a/src/core/hle/service/caps/caps_c.h +++ b/src/core/hle/service/caps/caps_c.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -14,8 +18,11 @@ namespace Service::Capture { class CAPS_C final : public ServiceFramework<CAPS_C> { public: - explicit CAPS_C(); + explicit CAPS_C(Core::System& system_); ~CAPS_C() override; + +private: + void SetShimLibraryVersion(Kernel::HLERequestContext& ctx); }; } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_sc.cpp b/src/core/hle/service/caps/caps_sc.cpp index 822ee96c8..d91e18e80 100644 --- a/src/core/hle/service/caps/caps_sc.cpp +++ b/src/core/hle/service/caps/caps_sc.cpp @@ -6,7 +6,7 @@ namespace Service::Capture { -CAPS_SC::CAPS_SC() : ServiceFramework("caps:sc") { +CAPS_SC::CAPS_SC(Core::System& system_) : ServiceFramework{system_, "caps:sc"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "CaptureRawImage"}, diff --git a/src/core/hle/service/caps/caps_sc.h b/src/core/hle/service/caps/caps_sc.h index ac3e929ca..e79a33ee5 100644 --- a/src/core/hle/service/caps/caps_sc.h +++ b/src/core/hle/service/caps/caps_sc.h @@ -6,15 +6,15 @@ #include "core/hle/service/service.h" -namespace Kernel { -class HLERequestContext; +namespace Core { +class System; } namespace Service::Capture { class CAPS_SC final : public ServiceFramework<CAPS_SC> { public: - explicit CAPS_SC(); + explicit CAPS_SC(Core::System& system_); ~CAPS_SC() override; }; diff --git a/src/core/hle/service/caps/caps_ss.cpp b/src/core/hle/service/caps/caps_ss.cpp index 24dc716e7..2b5314691 100644 --- a/src/core/hle/service/caps/caps_ss.cpp +++ b/src/core/hle/service/caps/caps_ss.cpp @@ -6,7 +6,7 @@ namespace Service::Capture { -CAPS_SS::CAPS_SS() : ServiceFramework("caps:ss") { +CAPS_SS::CAPS_SS(Core::System& system_) : ServiceFramework{system_, "caps:ss"} { // clang-format off static const FunctionInfo functions[] = { {201, nullptr, "SaveScreenShot"}, diff --git a/src/core/hle/service/caps/caps_ss.h b/src/core/hle/service/caps/caps_ss.h index 450686e4f..1816f7885 100644 --- a/src/core/hle/service/caps/caps_ss.h +++ b/src/core/hle/service/caps/caps_ss.h @@ -6,15 +6,15 @@ #include "core/hle/service/service.h" -namespace Kernel { -class HLERequestContext; +namespace Core { +class System; } namespace Service::Capture { class CAPS_SS final : public ServiceFramework<CAPS_SS> { public: - explicit CAPS_SS(); + explicit CAPS_SS(Core::System& system_); ~CAPS_SS() override; }; diff --git a/src/core/hle/service/caps/caps_su.cpp b/src/core/hle/service/caps/caps_su.cpp index fffb2ecf9..eae39eb7b 100644 --- a/src/core/hle/service/caps/caps_su.cpp +++ b/src/core/hle/service/caps/caps_su.cpp @@ -8,7 +8,7 @@ namespace Service::Capture { -CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") { +CAPS_SU::CAPS_SU(Core::System& system_) : ServiceFramework{system_, "caps:su"} { // clang-format off static const FunctionInfo functions[] = { {32, &CAPS_SU::SetShimLibraryVersion, "SetShimLibraryVersion"}, @@ -25,7 +25,12 @@ CAPS_SU::CAPS_SU() : ServiceFramework("caps:su") { CAPS_SU::~CAPS_SU() = default; void CAPS_SU::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_Capture, "(STUBBED) called"); + IPC::RequestParser rp{ctx}; + const auto library_version{rp.Pop<u64>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}", + library_version, applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); diff --git a/src/core/hle/service/caps/caps_su.h b/src/core/hle/service/caps/caps_su.h index 62c9603a9..b366fdb13 100644 --- a/src/core/hle/service/caps/caps_su.h +++ b/src/core/hle/service/caps/caps_su.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -14,7 +18,7 @@ namespace Service::Capture { class CAPS_SU final : public ServiceFramework<CAPS_SU> { public: - explicit CAPS_SU(); + explicit CAPS_SU(Core::System& system_); ~CAPS_SU() override; private: diff --git a/src/core/hle/service/caps/caps_u.cpp b/src/core/hle/service/caps/caps_u.cpp index f36d8de2d..842316a2e 100644 --- a/src/core/hle/service/caps/caps_u.cpp +++ b/src/core/hle/service/caps/caps_u.cpp @@ -12,8 +12,8 @@ namespace Service::Capture { class IAlbumAccessorApplicationSession final : public ServiceFramework<IAlbumAccessorApplicationSession> { public: - explicit IAlbumAccessorApplicationSession() - : ServiceFramework{"IAlbumAccessorApplicationSession"} { + explicit IAlbumAccessorApplicationSession(Core::System& system_) + : ServiceFramework{system_, "IAlbumAccessorApplicationSession"} { // clang-format off static const FunctionInfo functions[] = { {2001, nullptr, "OpenAlbumMovieReadStream"}, @@ -28,11 +28,10 @@ public: } }; -CAPS_U::CAPS_U() : ServiceFramework("caps:u") { +CAPS_U::CAPS_U(Core::System& system_) : ServiceFramework{system_, "caps:u"} { // clang-format off static const FunctionInfo functions[] = { - {31, nullptr, "GetShimLibraryVersion"}, - {32, nullptr, "SetShimLibraryVersion"}, + {32, &CAPS_U::SetShimLibraryVersion, "SetShimLibraryVersion"}, {102, &CAPS_U::GetAlbumContentsFileListForApplication, "GetAlbumContentsFileListForApplication"}, {103, nullptr, "DeleteAlbumContentsFileForApplication"}, {104, nullptr, "GetAlbumContentsFileSizeForApplication"}, @@ -42,7 +41,7 @@ CAPS_U::CAPS_U() : ServiceFramework("caps:u") { {130, nullptr, "PrecheckToCreateContentsForApplication"}, {140, nullptr, "GetAlbumFileList1AafeAruidDeprecated"}, {141, nullptr, "GetAlbumFileList2AafeUidAruidDeprecated"}, - {142, nullptr, "GetAlbumFileList3AaeAruid"}, + {142, &CAPS_U::GetAlbumFileList3AaeAruid, "GetAlbumFileList3AaeAruid"}, {143, nullptr, "GetAlbumFileList4AaeUidAruid"}, {60002, nullptr, "OpenAccessorSessionForApplication"}, }; @@ -53,6 +52,18 @@ CAPS_U::CAPS_U() : ServiceFramework("caps:u") { CAPS_U::~CAPS_U() = default; +void CAPS_U::SetShimLibraryVersion(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto library_version{rp.Pop<u64>()}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + LOG_WARNING(Service_Capture, "(STUBBED) called. library_version={}, applet_resource_user_id={}", + library_version, applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + void CAPS_U::GetAlbumContentsFileListForApplication(Kernel::HLERequestContext& ctx) { // Takes a type-0x6 output buffer containing an array of ApplicationAlbumFileEntry, a PID, an // u8 ContentType, two s64s, and an u64 AppletResourceUserId. Returns an output u64 for total @@ -66,17 +77,24 @@ void CAPS_U::GetAlbumContentsFileListForApplication(Kernel::HLERequestContext& c // TODO: Update this when we implement the album. // Currently we do not have a method of accessing album entries, set this to 0 for now. - constexpr s32 total_entries{0}; + constexpr u32 total_entries_1{}; + constexpr u32 total_entries_2{}; - LOG_WARNING(Service_Capture, - "(STUBBED) called. pid={}, content_type={}, start_posix_time={}, " - "end_posix_time={}, applet_resource_user_id={}, total_entries={}", - pid, content_type, start_posix_time, end_posix_time, applet_resource_user_id, - total_entries); + LOG_WARNING( + Service_Capture, + "(STUBBED) called. pid={}, content_type={}, start_posix_time={}, " + "end_posix_time={}, applet_resource_user_id={}, total_entries_1={}, total_entries_2={}", + pid, content_type, start_posix_time, end_posix_time, applet_resource_user_id, + total_entries_1, total_entries_2); - IPC::ResponseBuilder rb{ctx, 3}; + IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push(total_entries); + rb.Push(total_entries_1); + rb.Push(total_entries_2); +} + +void CAPS_U::GetAlbumFileList3AaeAruid(Kernel::HLERequestContext& ctx) { + GetAlbumContentsFileListForApplication(ctx); } } // namespace Service::Capture diff --git a/src/core/hle/service/caps/caps_u.h b/src/core/hle/service/caps/caps_u.h index 689364de4..e7e0d8775 100644 --- a/src/core/hle/service/caps/caps_u.h +++ b/src/core/hle/service/caps/caps_u.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -14,11 +18,13 @@ namespace Service::Capture { class CAPS_U final : public ServiceFramework<CAPS_U> { public: - explicit CAPS_U(); + explicit CAPS_U(Core::System& system_); ~CAPS_U() override; private: + void SetShimLibraryVersion(Kernel::HLERequestContext& ctx); void GetAlbumContentsFileListForApplication(Kernel::HLERequestContext& ctx); + void GetAlbumFileList3AaeAruid(Kernel::HLERequestContext& ctx); }; } // namespace Service::Capture diff --git a/src/core/hle/service/erpt/erpt.cpp b/src/core/hle/service/erpt/erpt.cpp index 4ec8c3093..4924c61c3 100644 --- a/src/core/hle/service/erpt/erpt.cpp +++ b/src/core/hle/service/erpt/erpt.cpp @@ -12,7 +12,7 @@ namespace Service::ERPT { class ErrorReportContext final : public ServiceFramework<ErrorReportContext> { public: - explicit ErrorReportContext() : ServiceFramework{"erpt:c"} { + explicit ErrorReportContext(Core::System& system_) : ServiceFramework{system_, "erpt:c"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SubmitContext"}, @@ -35,7 +35,7 @@ public: class ErrorReportSession final : public ServiceFramework<ErrorReportSession> { public: - explicit ErrorReportSession() : ServiceFramework{"erpt:r"} { + explicit ErrorReportSession(Core::System& system_) : ServiceFramework{system_, "erpt:r"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "OpenReport"}, @@ -48,9 +48,9 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<ErrorReportContext>()->InstallAsService(sm); - std::make_shared<ErrorReportSession>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<ErrorReportContext>(system)->InstallAsService(sm); + std::make_shared<ErrorReportSession>(system)->InstallAsService(sm); } } // namespace Service::ERPT diff --git a/src/core/hle/service/erpt/erpt.h b/src/core/hle/service/erpt/erpt.h index de439ab6d..8cd5c081f 100644 --- a/src/core/hle/service/erpt/erpt.h +++ b/src/core/hle/service/erpt/erpt.h @@ -4,6 +4,10 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } @@ -11,6 +15,6 @@ class ServiceManager; namespace Service::ERPT { /// Registers all ERPT services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::ERPT diff --git a/src/core/hle/service/es/es.cpp b/src/core/hle/service/es/es.cpp index c2737a365..26d1e3306 100644 --- a/src/core/hle/service/es/es.cpp +++ b/src/core/hle/service/es/es.cpp @@ -14,7 +14,7 @@ constexpr ResultCode ERROR_INVALID_RIGHTS_ID{ErrorModule::ETicket, 3}; class ETicket final : public ServiceFramework<ETicket> { public: - explicit ETicket() : ServiceFramework{"es"} { + explicit ETicket(Core::System& system_) : ServiceFramework{system_, "es"} { // clang-format off static const FunctionInfo functions[] = { {1, &ETicket::ImportTicket, "ImportTicket"}, @@ -305,8 +305,8 @@ private: Core::Crypto::KeyManager& keys = Core::Crypto::KeyManager::Instance(); }; -void InstallInterfaces(SM::ServiceManager& service_manager) { - std::make_shared<ETicket>()->InstallAsService(service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { + std::make_shared<ETicket>(system)->InstallAsService(service_manager); } } // namespace Service::ES diff --git a/src/core/hle/service/es/es.h b/src/core/hle/service/es/es.h index afe70465b..2a7b27d12 100644 --- a/src/core/hle/service/es/es.h +++ b/src/core/hle/service/es/es.h @@ -4,6 +4,10 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } @@ -11,6 +15,6 @@ class ServiceManager; namespace Service::ES { /// Registers all ES services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); } // namespace Service::ES diff --git a/src/core/hle/service/eupld/eupld.cpp b/src/core/hle/service/eupld/eupld.cpp index 0d6d244f4..2d650b1b7 100644 --- a/src/core/hle/service/eupld/eupld.cpp +++ b/src/core/hle/service/eupld/eupld.cpp @@ -12,7 +12,7 @@ namespace Service::EUPLD { class ErrorUploadContext final : public ServiceFramework<ErrorUploadContext> { public: - explicit ErrorUploadContext() : ServiceFramework{"eupld:c"} { + explicit ErrorUploadContext(Core::System& system_) : ServiceFramework{system_, "eupld:c"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SetUrl"}, @@ -29,7 +29,7 @@ public: class ErrorUploadRequest final : public ServiceFramework<ErrorUploadRequest> { public: - explicit ErrorUploadRequest() : ServiceFramework{"eupld:r"} { + explicit ErrorUploadRequest(Core::System& system_) : ServiceFramework{system_, "eupld:r"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -45,9 +45,9 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<ErrorUploadContext>()->InstallAsService(sm); - std::make_shared<ErrorUploadRequest>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<ErrorUploadContext>(system)->InstallAsService(sm); + std::make_shared<ErrorUploadRequest>(system)->InstallAsService(sm); } } // namespace Service::EUPLD diff --git a/src/core/hle/service/eupld/eupld.h b/src/core/hle/service/eupld/eupld.h index 6eef2c15f..539993a9d 100644 --- a/src/core/hle/service/eupld/eupld.h +++ b/src/core/hle/service/eupld/eupld.h @@ -4,6 +4,10 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } @@ -11,6 +15,6 @@ class ServiceManager; namespace Service::EUPLD { /// Registers all EUPLD services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::EUPLD diff --git a/src/core/hle/service/fatal/fatal.cpp b/src/core/hle/service/fatal/fatal.cpp index 2546d7595..13147472e 100644 --- a/src/core/hle/service/fatal/fatal.cpp +++ b/src/core/hle/service/fatal/fatal.cpp @@ -20,8 +20,9 @@ namespace Service::Fatal { -Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& system, const char* name) - : ServiceFramework(name), module(std::move(module)), system(system) {} +Module::Interface::Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name) + : ServiceFramework{system_, name}, module{std::move(module_)} {} Module::Interface::~Interface() = default; @@ -110,8 +111,9 @@ static void GenerateErrorReport(Core::System& system, ResultCode error_code, static void ThrowFatalError(Core::System& system, ResultCode error_code, FatalType fatal_type, const FatalInfo& info) { - LOG_ERROR(Service_Fatal, "Threw fatal error type {} with error code 0x{:X}", - static_cast<u32>(fatal_type), error_code.raw); + LOG_ERROR(Service_Fatal, "Threw fatal error type {} with error code 0x{:X}", fatal_type, + error_code.raw); + switch (fatal_type) { case FatalType::ErrorReportAndScreen: GenerateErrorReport(system, error_code, info); diff --git a/src/core/hle/service/fatal/fatal.h b/src/core/hle/service/fatal/fatal.h index bd9339dfc..2095bf89f 100644 --- a/src/core/hle/service/fatal/fatal.h +++ b/src/core/hle/service/fatal/fatal.h @@ -16,7 +16,8 @@ class Module final { public: class Interface : public ServiceFramework<Interface> { public: - explicit Interface(std::shared_ptr<Module> module, Core::System& system, const char* name); + explicit Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name); ~Interface() override; void ThrowFatal(Kernel::HLERequestContext& ctx); @@ -25,7 +26,6 @@ public: protected: std::shared_ptr<Module> module; - Core::System& system; }; }; diff --git a/src/core/hle/service/fatal/fatal_p.cpp b/src/core/hle/service/fatal/fatal_p.cpp index 066ccf6b0..8672b85dc 100644 --- a/src/core/hle/service/fatal/fatal_p.cpp +++ b/src/core/hle/service/fatal/fatal_p.cpp @@ -6,8 +6,8 @@ namespace Service::Fatal { -Fatal_P::Fatal_P(std::shared_ptr<Module> module, Core::System& system) - : Module::Interface(std::move(module), system, "fatal:p") {} +Fatal_P::Fatal_P(std::shared_ptr<Module> module_, Core::System& system_) + : Interface(std::move(module_), system_, "fatal:p") {} Fatal_P::~Fatal_P() = default; diff --git a/src/core/hle/service/fatal/fatal_p.h b/src/core/hle/service/fatal/fatal_p.h index c6d953cb5..ffa5b7b98 100644 --- a/src/core/hle/service/fatal/fatal_p.h +++ b/src/core/hle/service/fatal/fatal_p.h @@ -10,7 +10,7 @@ namespace Service::Fatal { class Fatal_P final : public Module::Interface { public: - explicit Fatal_P(std::shared_ptr<Module> module, Core::System& system); + explicit Fatal_P(std::shared_ptr<Module> module_, Core::System& system_); ~Fatal_P() override; }; diff --git a/src/core/hle/service/fatal/fatal_u.cpp b/src/core/hle/service/fatal/fatal_u.cpp index 8d72ed485..82993938a 100644 --- a/src/core/hle/service/fatal/fatal_u.cpp +++ b/src/core/hle/service/fatal/fatal_u.cpp @@ -6,8 +6,8 @@ namespace Service::Fatal { -Fatal_U::Fatal_U(std::shared_ptr<Module> module, Core::System& system) - : Module::Interface(std::move(module), system, "fatal:u") { +Fatal_U::Fatal_U(std::shared_ptr<Module> module_, Core::System& system_) + : Interface(std::move(module_), system_, "fatal:u") { static const FunctionInfo functions[] = { {0, &Fatal_U::ThrowFatal, "ThrowFatal"}, {1, &Fatal_U::ThrowFatalWithPolicy, "ThrowFatalWithPolicy"}, diff --git a/src/core/hle/service/fatal/fatal_u.h b/src/core/hle/service/fatal/fatal_u.h index 34c5c7f95..0b58c9112 100644 --- a/src/core/hle/service/fatal/fatal_u.h +++ b/src/core/hle/service/fatal/fatal_u.h @@ -10,7 +10,7 @@ namespace Service::Fatal { class Fatal_U final : public Module::Interface { public: - explicit Fatal_U(std::shared_ptr<Module> module, Core::System& system); + explicit Fatal_U(std::shared_ptr<Module> module_, Core::System& system_); ~Fatal_U() override; }; diff --git a/src/core/hle/service/fgm/fgm.cpp b/src/core/hle/service/fgm/fgm.cpp index e461274c1..9dc1bc52e 100644 --- a/src/core/hle/service/fgm/fgm.cpp +++ b/src/core/hle/service/fgm/fgm.cpp @@ -14,7 +14,7 @@ namespace Service::FGM { class IRequest final : public ServiceFramework<IRequest> { public: - explicit IRequest() : ServiceFramework{"IRequest"} { + explicit IRequest(Core::System& system_) : ServiceFramework{system_, "IRequest"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -30,7 +30,7 @@ public: class FGM final : public ServiceFramework<FGM> { public: - explicit FGM(const char* name) : ServiceFramework{name} { + explicit FGM(Core::System& system_, const char* name) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { {0, &FGM::Initialize, "Initialize"}, @@ -46,13 +46,13 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IRequest>(); + rb.PushIpcInterface<IRequest>(system); } }; class FGM_DBG final : public ServiceFramework<FGM_DBG> { public: - explicit FGM_DBG() : ServiceFramework{"fgm:dbg"} { + explicit FGM_DBG(Core::System& system_) : ServiceFramework{system_, "fgm:dbg"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -65,11 +65,11 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<FGM>("fgm")->InstallAsService(sm); - std::make_shared<FGM>("fgm:0")->InstallAsService(sm); - std::make_shared<FGM>("fgm:9")->InstallAsService(sm); - std::make_shared<FGM_DBG>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<FGM>(system, "fgm")->InstallAsService(sm); + std::make_shared<FGM>(system, "fgm:0")->InstallAsService(sm); + std::make_shared<FGM>(system, "fgm:9")->InstallAsService(sm); + std::make_shared<FGM_DBG>(system)->InstallAsService(sm); } } // namespace Service::FGM diff --git a/src/core/hle/service/fgm/fgm.h b/src/core/hle/service/fgm/fgm.h index e59691264..75978f2ed 100644 --- a/src/core/hle/service/fgm/fgm.h +++ b/src/core/hle/service/fgm/fgm.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::FGM { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::FGM diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp index 54a5fb84b..b15c737e1 100644 --- a/src/core/hle/service/filesystem/filesystem.cpp +++ b/src/core/hle/service/filesystem/filesystem.cpp @@ -79,7 +79,7 @@ ResultCode VfsDirectoryServiceWrapper::DeleteFile(const std::string& path_) cons } auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); - if (dir->GetFile(Common::FS::GetFilename(path)) == nullptr) { + if (dir == nullptr || dir->GetFile(Common::FS::GetFilename(path)) == nullptr) { return FileSys::ERROR_PATH_NOT_FOUND; } if (!dir->DeleteFile(Common::FS::GetFilename(path))) { @@ -93,8 +93,9 @@ ResultCode VfsDirectoryServiceWrapper::DeleteFile(const std::string& path_) cons ResultCode VfsDirectoryServiceWrapper::CreateDirectory(const std::string& path_) const { std::string path(Common::FS::SanitizePath(path_)); auto dir = GetDirectoryRelativeWrapped(backing, Common::FS::GetParentPath(path)); - if (dir == nullptr && Common::FS::GetFilename(Common::FS::GetParentPath(path)).empty()) + if (dir == nullptr || Common::FS::GetFilename(Common::FS::GetParentPath(path)).empty()) { dir = backing; + } auto new_dir = dir->CreateSubdirectory(Common::FS::GetFilename(path)); if (new_dir == nullptr) { // TODO(DarkLordZach): Find a better error code for this @@ -297,10 +298,35 @@ ResultVal<FileSys::VirtualFile> FileSystemController::OpenRomFSCurrentProcess() return romfs_factory->OpenCurrentProcess(system.CurrentProcess()->GetTitleID()); } +ResultVal<FileSys::VirtualFile> FileSystemController::OpenPatchedRomFS( + u64 title_id, FileSys::ContentRecordType type) const { + LOG_TRACE(Service_FS, "Opening patched RomFS for title_id={:016X}", title_id); + + if (romfs_factory == nullptr) { + // TODO: Find a better error code for this + return RESULT_UNKNOWN; + } + + return romfs_factory->OpenPatchedRomFS(title_id, type); +} + +ResultVal<FileSys::VirtualFile> FileSystemController::OpenPatchedRomFSWithProgramIndex( + u64 title_id, u8 program_index, FileSys::ContentRecordType type) const { + LOG_TRACE(Service_FS, "Opening patched RomFS for title_id={:016X}, program_index={}", title_id, + program_index); + + if (romfs_factory == nullptr) { + // TODO: Find a better error code for this + return RESULT_UNKNOWN; + } + + return romfs_factory->OpenPatchedRomFSWithProgramIndex(title_id, program_index, type); +} + ResultVal<FileSys::VirtualFile> FileSystemController::OpenRomFS( u64 title_id, FileSys::StorageId storage_id, FileSys::ContentRecordType type) const { LOG_TRACE(Service_FS, "Opening RomFS for title_id={:016X}, storage_id={:02X}, type={:02X}", - title_id, static_cast<u8>(storage_id), static_cast<u8>(type)); + title_id, storage_id, type); if (romfs_factory == nullptr) { // TODO(bunnei): Find a better error code for this @@ -312,8 +338,8 @@ ResultVal<FileSys::VirtualFile> FileSystemController::OpenRomFS( ResultVal<FileSys::VirtualDir> FileSystemController::CreateSaveData( FileSys::SaveDataSpaceId space, const FileSys::SaveDataAttribute& save_struct) const { - LOG_TRACE(Service_FS, "Creating Save Data for space_id={:01X}, save_struct={}", - static_cast<u8>(space), save_struct.DebugInfo()); + LOG_TRACE(Service_FS, "Creating Save Data for space_id={:01X}, save_struct={}", space, + save_struct.DebugInfo()); if (save_data_factory == nullptr) { return FileSys::ERROR_ENTITY_NOT_FOUND; @@ -324,8 +350,8 @@ ResultVal<FileSys::VirtualDir> FileSystemController::CreateSaveData( ResultVal<FileSys::VirtualDir> FileSystemController::OpenSaveData( FileSys::SaveDataSpaceId space, const FileSys::SaveDataAttribute& attribute) const { - LOG_TRACE(Service_FS, "Opening Save Data for space_id={:01X}, save_struct={}", - static_cast<u8>(space), attribute.DebugInfo()); + LOG_TRACE(Service_FS, "Opening Save Data for space_id={:01X}, save_struct={}", space, + attribute.DebugInfo()); if (save_data_factory == nullptr) { return FileSys::ERROR_ENTITY_NOT_FOUND; @@ -336,7 +362,7 @@ ResultVal<FileSys::VirtualDir> FileSystemController::OpenSaveData( ResultVal<FileSys::VirtualDir> FileSystemController::OpenSaveDataSpace( FileSys::SaveDataSpaceId space) const { - LOG_TRACE(Service_FS, "Opening Save Data Space for space_id={:01X}", static_cast<u8>(space)); + LOG_TRACE(Service_FS, "Opening Save Data Space for space_id={:01X}", space); if (save_data_factory == nullptr) { return FileSys::ERROR_ENTITY_NOT_FOUND; @@ -357,7 +383,7 @@ ResultVal<FileSys::VirtualDir> FileSystemController::OpenSDMC() const { ResultVal<FileSys::VirtualDir> FileSystemController::OpenBISPartition( FileSys::BisPartitionId id) const { - LOG_TRACE(Service_FS, "Opening BIS Partition with id={:08X}", static_cast<u32>(id)); + LOG_TRACE(Service_FS, "Opening BIS Partition with id={:08X}", id); if (bis_factory == nullptr) { return FileSys::ERROR_ENTITY_NOT_FOUND; @@ -373,7 +399,7 @@ ResultVal<FileSys::VirtualDir> FileSystemController::OpenBISPartition( ResultVal<FileSys::VirtualFile> FileSystemController::OpenBISPartitionStorage( FileSys::BisPartitionId id) const { - LOG_TRACE(Service_FS, "Opening BIS Partition Storage with id={:08X}", static_cast<u32>(id)); + LOG_TRACE(Service_FS, "Opening BIS Partition Storage with id={:08X}", id); if (bis_factory == nullptr) { return FileSys::ERROR_ENTITY_NOT_FOUND; @@ -454,7 +480,9 @@ FileSys::SaveDataSize FileSystemController::ReadSaveDataSize(FileSys::SaveDataTy const auto res = system.GetAppLoader().ReadControlData(nacp); if (res != Loader::ResultStatus::Success) { - FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID()}; + const FileSys::PatchManager pm{system.CurrentProcess()->GetTitleID(), + system.GetFileSystemController(), + system.GetContentProvider()}; const auto metadata = pm.GetControlMetadata(); const auto& nacp_unique = metadata.first; @@ -714,7 +742,8 @@ void FileSystemController::CreateFactories(FileSys::VfsFilesystem& vfs, bool ove } if (save_data_factory == nullptr) { - save_data_factory = std::make_unique<FileSys::SaveDataFactory>(std::move(nand_directory)); + save_data_factory = + std::make_unique<FileSys::SaveDataFactory>(system, std::move(nand_directory)); } if (sdmc_factory == nullptr) { @@ -725,10 +754,9 @@ void FileSystemController::CreateFactories(FileSys::VfsFilesystem& vfs, bool ove } void InstallInterfaces(Core::System& system) { - std::make_shared<FSP_LDR>()->InstallAsService(system.ServiceManager()); - std::make_shared<FSP_PR>()->InstallAsService(system.ServiceManager()); - std::make_shared<FSP_SRV>(system.GetFileSystemController(), system.GetReporter()) - ->InstallAsService(system.ServiceManager()); + std::make_shared<FSP_LDR>(system)->InstallAsService(system.ServiceManager()); + std::make_shared<FSP_PR>(system)->InstallAsService(system.ServiceManager()); + std::make_shared<FSP_SRV>(system)->InstallAsService(system.ServiceManager()); } } // namespace Service::FileSystem diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h index 6dbbf0b2b..7102d3f9a 100644 --- a/src/core/hle/service/filesystem/filesystem.h +++ b/src/core/hle/service/filesystem/filesystem.h @@ -66,6 +66,10 @@ public: void SetPackedUpdate(FileSys::VirtualFile update_raw); ResultVal<FileSys::VirtualFile> OpenRomFSCurrentProcess() const; + ResultVal<FileSys::VirtualFile> OpenPatchedRomFS(u64 title_id, + FileSys::ContentRecordType type) const; + ResultVal<FileSys::VirtualFile> OpenPatchedRomFSWithProgramIndex( + u64 title_id, u8 program_index, FileSys::ContentRecordType type) const; ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId storage_id, FileSys::ContentRecordType type) const; ResultVal<FileSys::VirtualDir> CreateSaveData( diff --git a/src/core/hle/service/filesystem/fsp_ldr.cpp b/src/core/hle/service/filesystem/fsp_ldr.cpp index fb487d5bc..1f6c17ba5 100644 --- a/src/core/hle/service/filesystem/fsp_ldr.cpp +++ b/src/core/hle/service/filesystem/fsp_ldr.cpp @@ -7,7 +7,7 @@ namespace Service::FileSystem { -FSP_LDR::FSP_LDR() : ServiceFramework{"fsp:ldr"} { +FSP_LDR::FSP_LDR(Core::System& system_) : ServiceFramework{system_, "fsp:ldr"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "OpenCodeFileSystem"}, diff --git a/src/core/hle/service/filesystem/fsp_ldr.h b/src/core/hle/service/filesystem/fsp_ldr.h index 8210b7729..d6432a0e1 100644 --- a/src/core/hle/service/filesystem/fsp_ldr.h +++ b/src/core/hle/service/filesystem/fsp_ldr.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::FileSystem { class FSP_LDR final : public ServiceFramework<FSP_LDR> { public: - explicit FSP_LDR(); + explicit FSP_LDR(Core::System& system_); ~FSP_LDR() override; }; diff --git a/src/core/hle/service/filesystem/fsp_pr.cpp b/src/core/hle/service/filesystem/fsp_pr.cpp index 378201610..00e4d1662 100644 --- a/src/core/hle/service/filesystem/fsp_pr.cpp +++ b/src/core/hle/service/filesystem/fsp_pr.cpp @@ -7,7 +7,7 @@ namespace Service::FileSystem { -FSP_PR::FSP_PR() : ServiceFramework{"fsp:pr"} { +FSP_PR::FSP_PR(Core::System& system_) : ServiceFramework{system_, "fsp:pr"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RegisterProgram"}, diff --git a/src/core/hle/service/filesystem/fsp_pr.h b/src/core/hle/service/filesystem/fsp_pr.h index 556ae5ce9..9e622518c 100644 --- a/src/core/hle/service/filesystem/fsp_pr.h +++ b/src/core/hle/service/filesystem/fsp_pr.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::FileSystem { class FSP_PR final : public ServiceFramework<FSP_PR> { public: - explicit FSP_PR(); + explicit FSP_PR(Core::System& system_); ~FSP_PR() override; }; diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp index 649128be4..9cc260515 100644 --- a/src/core/hle/service/filesystem/fsp_srv.cpp +++ b/src/core/hle/service/filesystem/fsp_srv.cpp @@ -14,6 +14,7 @@ #include "common/hex_util.h" #include "common/logging/log.h" #include "common/string_util.h" +#include "core/core.h" #include "core/file_sys/directory.h" #include "core/file_sys/errors.h" #include "core/file_sys/mode.h" @@ -56,8 +57,8 @@ enum class FileSystemType : u8 { class IStorage final : public ServiceFramework<IStorage> { public: - explicit IStorage(FileSys::VirtualFile backend_) - : ServiceFramework("IStorage"), backend(std::move(backend_)) { + explicit IStorage(Core::System& system_, FileSys::VirtualFile backend_) + : ServiceFramework{system_, "IStorage"}, backend(std::move(backend_)) { static const FunctionInfo functions[] = { {0, &IStorage::Read, "Read"}, {1, nullptr, "Write"}, @@ -114,8 +115,8 @@ private: class IFile final : public ServiceFramework<IFile> { public: - explicit IFile(FileSys::VirtualFile backend_) - : ServiceFramework("IFile"), backend(std::move(backend_)) { + explicit IFile(Core::System& system_, FileSys::VirtualFile backend_) + : ServiceFramework{system_, "IFile"}, backend(std::move(backend_)) { static const FunctionInfo functions[] = { {0, &IFile::Read, "Read"}, {1, &IFile::Write, "Write"}, {2, &IFile::Flush, "Flush"}, {3, &IFile::SetSize, "SetSize"}, @@ -246,8 +247,8 @@ static void BuildEntryIndex(std::vector<FileSys::Entry>& entries, const std::vec class IDirectory final : public ServiceFramework<IDirectory> { public: - explicit IDirectory(FileSys::VirtualDir backend_) - : ServiceFramework("IDirectory"), backend(std::move(backend_)) { + explicit IDirectory(Core::System& system_, FileSys::VirtualDir backend_) + : ServiceFramework{system_, "IDirectory"}, backend(std::move(backend_)) { static const FunctionInfo functions[] = { {0, &IDirectory::Read, "Read"}, {1, &IDirectory::GetEntryCount, "GetEntryCount"}, @@ -302,8 +303,9 @@ private: class IFileSystem final : public ServiceFramework<IFileSystem> { public: - explicit IFileSystem(FileSys::VirtualDir backend, SizeGetter size) - : ServiceFramework("IFileSystem"), backend(std::move(backend)), size(std::move(size)) { + explicit IFileSystem(Core::System& system_, FileSys::VirtualDir backend_, SizeGetter size_) + : ServiceFramework{system_, "IFileSystem"}, backend{std::move(backend_)}, size{std::move( + size_)} { static const FunctionInfo functions[] = { {0, &IFileSystem::CreateFile, "CreateFile"}, {1, &IFileSystem::DeleteFile, "DeleteFile"}, @@ -411,7 +413,7 @@ public: const auto mode = static_cast<FileSys::Mode>(rp.Pop<u32>()); - LOG_DEBUG(Service_FS, "called. file={}, mode={}", name, static_cast<u32>(mode)); + LOG_DEBUG(Service_FS, "called. file={}, mode={}", name, mode); auto result = backend.OpenFile(name, mode); if (result.Failed()) { @@ -420,7 +422,7 @@ public: return; } - auto file = std::make_shared<IFile>(result.Unwrap()); + auto file = std::make_shared<IFile>(system, result.Unwrap()); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); @@ -445,7 +447,7 @@ public: return; } - auto directory = std::make_shared<IDirectory>(result.Unwrap()); + auto directory = std::make_shared<IDirectory>(system, result.Unwrap()); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); @@ -500,8 +502,9 @@ private: class ISaveDataInfoReader final : public ServiceFramework<ISaveDataInfoReader> { public: - explicit ISaveDataInfoReader(FileSys::SaveDataSpaceId space, FileSystemController& fsc) - : ServiceFramework("ISaveDataInfoReader"), fsc(fsc) { + explicit ISaveDataInfoReader(Core::System& system_, FileSys::SaveDataSpaceId space, + FileSystemController& fsc_) + : ServiceFramework{system_, "ISaveDataInfoReader"}, fsc{fsc_} { static const FunctionInfo functions[] = { {0, &ISaveDataInfoReader::ReadSaveDataInfo, "ReadSaveDataInfo"}, }; @@ -550,8 +553,7 @@ private: const auto save_root = fsc.OpenSaveDataSpace(space); if (save_root.Failed() || *save_root == nullptr) { - LOG_ERROR(Service_FS, "The save root for the space_id={:02X} was invalid!", - static_cast<u8>(space)); + LOG_ERROR(Service_FS, "The save root for the space_id={:02X} was invalid!", space); return; } @@ -650,8 +652,9 @@ private: u64 next_entry_index = 0; }; -FSP_SRV::FSP_SRV(FileSystemController& fsc, const Core::Reporter& reporter) - : ServiceFramework("fsp-srv"), fsc(fsc), reporter(reporter) { +FSP_SRV::FSP_SRV(Core::System& system_) + : ServiceFramework{system_, "fsp-srv"}, fsc{system.GetFileSystemController()}, + content_provider{system.GetContentProvider()}, reporter{system.GetReporter()} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "OpenFileSystem"}, @@ -714,7 +717,7 @@ FSP_SRV::FSP_SRV(FileSystemController& fsc, const Core::Reporter& reporter) {202, &FSP_SRV::OpenDataStorageByDataId, "OpenDataStorageByDataId"}, {203, &FSP_SRV::OpenPatchDataStorageByCurrentProcess, "OpenPatchDataStorageByCurrentProcess"}, {204, nullptr, "OpenDataFileSystemByProgramIndex"}, - {205, nullptr, "OpenDataStorageByProgramIndex"}, + {205, &FSP_SRV::OpenDataStorageWithProgramIndex, "OpenDataStorageWithProgramIndex"}, {400, nullptr, "OpenDeviceOperator"}, {500, nullptr, "OpenSdCardDetectionEventNotifier"}, {501, nullptr, "OpenGameCardDetectionEventNotifier"}, @@ -791,8 +794,7 @@ void FSP_SRV::OpenFileSystemWithPatch(Kernel::HLERequestContext& ctx) { const auto type = rp.PopRaw<FileSystemType>(); const auto title_id = rp.PopRaw<u64>(); - LOG_WARNING(Service_FS, "(STUBBED) called with type={}, title_id={:016X}", - static_cast<u8>(type), title_id); + LOG_WARNING(Service_FS, "(STUBBED) called with type={}, title_id={:016X}", type, title_id); IPC::ResponseBuilder rb{ctx, 2, 0, 0}; rb.Push(RESULT_UNKNOWN); @@ -801,8 +803,9 @@ void FSP_SRV::OpenFileSystemWithPatch(Kernel::HLERequestContext& ctx) { void FSP_SRV::OpenSdCardFileSystem(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_FS, "called"); - auto filesystem = std::make_shared<IFileSystem>( - fsc.OpenSDMC().Unwrap(), SizeGetter::FromStorageId(fsc, FileSys::StorageId::SdCard)); + auto filesystem = + std::make_shared<IFileSystem>(system, fsc.OpenSDMC().Unwrap(), + SizeGetter::FromStorageId(fsc, FileSys::StorageId::SdCard)); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); @@ -862,8 +865,8 @@ void FSP_SRV::OpenSaveDataFileSystem(Kernel::HLERequestContext& ctx) { UNREACHABLE(); } - auto filesystem = - std::make_shared<IFileSystem>(std::move(dir.Unwrap()), SizeGetter::FromStorageId(fsc, id)); + auto filesystem = std::make_shared<IFileSystem>(system, std::move(dir.Unwrap()), + SizeGetter::FromStorageId(fsc, id)); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); @@ -878,11 +881,12 @@ void FSP_SRV::OpenReadOnlySaveDataFileSystem(Kernel::HLERequestContext& ctx) { void FSP_SRV::OpenSaveDataInfoReaderBySaveDataSpaceId(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto space = rp.PopRaw<FileSys::SaveDataSpaceId>(); - LOG_INFO(Service_FS, "called, space={}", static_cast<u8>(space)); + LOG_INFO(Service_FS, "called, space={}", space); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISaveDataInfoReader>(std::make_shared<ISaveDataInfoReader>(space, fsc)); + rb.PushIpcInterface<ISaveDataInfoReader>( + std::make_shared<ISaveDataInfoReader>(system, space, fsc)); } void FSP_SRV::WriteSaveDataFileSystemExtraDataBySaveDataAttribute(Kernel::HLERequestContext& ctx) { @@ -909,10 +913,10 @@ void FSP_SRV::ReadSaveDataFileSystemExtraDataWithMaskBySaveDataAttribute( "(STUBBED) called, flags={}, space_id={}, attribute.title_id={:016X}\n" "attribute.user_id={:016X}{:016X}, attribute.save_id={:016X}\n" "attribute.type={}, attribute.rank={}, attribute.index={}", - flags, static_cast<u32>(parameters.space_id), parameters.attribute.title_id, + flags, parameters.space_id, parameters.attribute.title_id, parameters.attribute.user_id[1], parameters.attribute.user_id[0], - parameters.attribute.save_id, static_cast<u32>(parameters.attribute.type), - static_cast<u32>(parameters.attribute.rank), parameters.attribute.index); + parameters.attribute.save_id, parameters.attribute.type, parameters.attribute.rank, + parameters.attribute.index); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); @@ -931,7 +935,7 @@ void FSP_SRV::OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) { return; } - auto storage = std::make_shared<IStorage>(std::move(romfs.Unwrap())); + auto storage = std::make_shared<IStorage>(system, std::move(romfs.Unwrap())); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); @@ -945,7 +949,7 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) { const auto title_id = rp.PopRaw<u64>(); LOG_DEBUG(Service_FS, "called with storage_id={:02X}, unknown={:08X}, title_id={:016X}", - static_cast<u8>(storage_id), unknown, title_id); + storage_id, unknown, title_id); auto data = fsc.OpenRomFS(title_id, storage_id, FileSys::ContentRecordType::Data); @@ -955,23 +959,23 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) { if (archive != nullptr) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface(std::make_shared<IStorage>(archive)); + rb.PushIpcInterface(std::make_shared<IStorage>(system, archive)); return; } // TODO(DarkLordZach): Find the right error code to use here LOG_ERROR(Service_FS, "could not open data storage with title_id={:016X}, storage_id={:02X}", title_id, - static_cast<u8>(storage_id)); + storage_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_UNKNOWN); return; } - FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, fsc, content_provider}; auto storage = std::make_shared<IStorage>( - pm.PatchRomFS(std::move(data.Unwrap()), 0, FileSys::ContentRecordType::Data)); + system, pm.PatchRomFS(std::move(data.Unwrap()), 0, FileSys::ContentRecordType::Data)); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); @@ -981,21 +985,46 @@ void FSP_SRV::OpenDataStorageByDataId(Kernel::HLERequestContext& ctx) { void FSP_SRV::OpenPatchDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto storage_id = rp.PopRaw<FileSys::StorageId>(); - auto title_id = rp.PopRaw<u64>(); + const auto storage_id = rp.PopRaw<FileSys::StorageId>(); + const auto title_id = rp.PopRaw<u64>(); - LOG_DEBUG(Service_FS, "called with storage_id={:02X}, title_id={:016X}", - static_cast<u8>(storage_id), title_id); + LOG_DEBUG(Service_FS, "called with storage_id={:02X}, title_id={:016X}", storage_id, title_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND); } +void FSP_SRV::OpenDataStorageWithProgramIndex(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const auto program_index = rp.PopRaw<u8>(); + + LOG_DEBUG(Service_FS, "called, program_index={}", program_index); + + auto romfs = fsc.OpenPatchedRomFSWithProgramIndex( + system.CurrentProcess()->GetTitleID(), program_index, FileSys::ContentRecordType::Program); + + if (romfs.Failed()) { + // TODO: Find the right error code to use here + LOG_ERROR(Service_FS, "could not open storage with program_index={}", program_index); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_UNKNOWN); + return; + } + + auto storage = std::make_shared<IStorage>(system, std::move(romfs.Unwrap())); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IStorage>(std::move(storage)); +} + void FSP_SRV::SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; log_mode = rp.PopEnum<LogMode>(); - LOG_DEBUG(Service_FS, "called, log_mode={:08X}", static_cast<u32>(log_mode)); + LOG_DEBUG(Service_FS, "called, log_mode={:08X}", log_mode); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -1033,7 +1062,8 @@ void FSP_SRV::GetAccessLogVersionInfo(Kernel::HLERequestContext& ctx) { class IMultiCommitManager final : public ServiceFramework<IMultiCommitManager> { public: - explicit IMultiCommitManager() : ServiceFramework("IMultiCommitManager") { + explicit IMultiCommitManager(Core::System& system_) + : ServiceFramework{system_, "IMultiCommitManager"} { static const FunctionInfo functions[] = { {1, &IMultiCommitManager::Add, "Add"}, {2, &IMultiCommitManager::Commit, "Commit"}, @@ -1064,7 +1094,7 @@ void FSP_SRV::OpenMultiCommitManager(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IMultiCommitManager>(std::make_shared<IMultiCommitManager>()); + rb.PushIpcInterface<IMultiCommitManager>(std::make_shared<IMultiCommitManager>(system)); } } // namespace Service::FileSystem diff --git a/src/core/hle/service/filesystem/fsp_srv.h b/src/core/hle/service/filesystem/fsp_srv.h index 4964e874e..8ed933279 100644 --- a/src/core/hle/service/filesystem/fsp_srv.h +++ b/src/core/hle/service/filesystem/fsp_srv.h @@ -12,8 +12,9 @@ class Reporter; } namespace FileSys { +class ContentProvider; class FileSystemBackend; -} +} // namespace FileSys namespace Service::FileSystem { @@ -32,7 +33,7 @@ enum class LogMode : u32 { class FSP_SRV final : public ServiceFramework<FSP_SRV> { public: - explicit FSP_SRV(FileSystemController& fsc, const Core::Reporter& reporter); + explicit FSP_SRV(Core::System& system_); ~FSP_SRV() override; private: @@ -48,6 +49,7 @@ private: void OpenDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx); void OpenDataStorageByDataId(Kernel::HLERequestContext& ctx); void OpenPatchDataStorageByCurrentProcess(Kernel::HLERequestContext& ctx); + void OpenDataStorageWithProgramIndex(Kernel::HLERequestContext& ctx); void SetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); void GetGlobalAccessLogMode(Kernel::HLERequestContext& ctx); void OutputAccessLogToSdCard(Kernel::HLERequestContext& ctx); @@ -55,6 +57,7 @@ private: void OpenMultiCommitManager(Kernel::HLERequestContext& ctx); FileSystemController& fsc; + const FileSys::ContentProvider& content_provider; FileSys::VirtualFile romfs; u64 current_process_id = 0; diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp index b7adaffc7..c5b053c31 100644 --- a/src/core/hle/service/friend/friend.cpp +++ b/src/core/hle/service/friend/friend.cpp @@ -5,6 +5,7 @@ #include <queue> #include "common/logging/log.h" #include "common/uuid.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/readable_event.h" #include "core/hle/kernel/writable_event.h" @@ -16,7 +17,7 @@ namespace Service::Friend { class IFriendService final : public ServiceFramework<IFriendService> { public: - IFriendService() : ServiceFramework("IFriendService") { + explicit IFriendService(Core::System& system_) : ServiceFramework{system_, "IFriendService"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetCompletionEvent"}, @@ -170,8 +171,8 @@ private: class INotificationService final : public ServiceFramework<INotificationService> { public: - INotificationService(Common::UUID uuid, Core::System& system) - : ServiceFramework("INotificationService"), uuid(uuid) { + explicit INotificationService(Common::UUID uuid_, Core::System& system_) + : ServiceFramework{system_, "INotificationService"}, uuid{uuid_} { // clang-format off static const FunctionInfo functions[] = { {0, &INotificationService::GetEvent, "GetEvent"}, @@ -228,8 +229,7 @@ private: break; default: // HOS seems not have an error case for an unknown notification - LOG_WARNING(Service_ACC, "Unknown notification {:08X}", - static_cast<u32>(notification.notification_type)); + LOG_WARNING(Service_ACC, "Unknown notification {:08X}", notification.notification_type); break; } @@ -266,7 +266,7 @@ private: void Module::Interface::CreateFriendService(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IFriendService>(); + rb.PushIpcInterface<IFriendService>(system); LOG_DEBUG(Service_ACC, "called"); } @@ -281,8 +281,9 @@ void Module::Interface::CreateNotificationService(Kernel::HLERequestContext& ctx rb.PushIpcInterface<INotificationService>(uuid, system); } -Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& system, const char* name) - : ServiceFramework(name), module(std::move(module)), system(system) {} +Module::Interface::Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name) + : ServiceFramework{system_, name}, module{std::move(module_)} {} Module::Interface::~Interface() = default; diff --git a/src/core/hle/service/friend/friend.h b/src/core/hle/service/friend/friend.h index 24f3fc969..8be3321db 100644 --- a/src/core/hle/service/friend/friend.h +++ b/src/core/hle/service/friend/friend.h @@ -16,7 +16,8 @@ class Module final { public: class Interface : public ServiceFramework<Interface> { public: - explicit Interface(std::shared_ptr<Module> module, Core::System& system, const char* name); + explicit Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name); ~Interface() override; void CreateFriendService(Kernel::HLERequestContext& ctx); @@ -24,7 +25,6 @@ public: protected: std::shared_ptr<Module> module; - Core::System& system; }; }; diff --git a/src/core/hle/service/friend/interface.cpp b/src/core/hle/service/friend/interface.cpp index 58155f652..7368ccec2 100644 --- a/src/core/hle/service/friend/interface.cpp +++ b/src/core/hle/service/friend/interface.cpp @@ -6,8 +6,8 @@ namespace Service::Friend { -Friend::Friend(std::shared_ptr<Module> module, Core::System& system, const char* name) - : Interface(std::move(module), system, name) { +Friend::Friend(std::shared_ptr<Module> module_, Core::System& system_, const char* name) + : Interface(std::move(module_), system_, name) { static const FunctionInfo functions[] = { {0, &Friend::CreateFriendService, "CreateFriendService"}, {1, &Friend::CreateNotificationService, "CreateNotificationService"}, diff --git a/src/core/hle/service/friend/interface.h b/src/core/hle/service/friend/interface.h index 465a35770..43d914b32 100644 --- a/src/core/hle/service/friend/interface.h +++ b/src/core/hle/service/friend/interface.h @@ -10,7 +10,7 @@ namespace Service::Friend { class Friend final : public Module::Interface { public: - explicit Friend(std::shared_ptr<Module> module, Core::System& system, const char* name); + explicit Friend(std::shared_ptr<Module> module_, Core::System& system_, const char* name); ~Friend() override; }; diff --git a/src/core/hle/service/glue/arp.cpp b/src/core/hle/service/glue/arp.cpp index b591ce31b..fc77e7286 100644 --- a/src/core/hle/service/glue/arp.cpp +++ b/src/core/hle/service/glue/arp.cpp @@ -5,6 +5,7 @@ #include <memory> #include "common/logging/log.h" +#include "core/core.h" #include "core/file_sys/control_metadata.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" @@ -32,8 +33,8 @@ std::optional<u64> GetTitleIDForProcessID(const Core::System& system, u64 proces } } // Anonymous namespace -ARP_R::ARP_R(const Core::System& system, const ARPManager& manager) - : ServiceFramework{"arp:r"}, system(system), manager(manager) { +ARP_R::ARP_R(Core::System& system_, const ARPManager& manager_) + : ServiceFramework{system_, "arp:r"}, manager{manager_} { // clang-format off static const FunctionInfo functions[] = { {0, &ARP_R::GetApplicationLaunchProperty, "GetApplicationLaunchProperty"}, @@ -151,8 +152,9 @@ class IRegistrar final : public ServiceFramework<IRegistrar> { public: explicit IRegistrar( + Core::System& system_, std::function<ResultCode(u64, ApplicationLaunchProperty, std::vector<u8>)> issuer) - : ServiceFramework{"IRegistrar"}, issue_process_id(std::move(issuer)) { + : ServiceFramework{system_, "IRegistrar"}, issue_process_id{std::move(issuer)} { // clang-format off static const FunctionInfo functions[] = { {0, &IRegistrar::Issue, "Issue"}, @@ -236,8 +238,8 @@ private: std::vector<u8> control; }; -ARP_W::ARP_W(const Core::System& system, ARPManager& manager) - : ServiceFramework{"arp:w"}, system(system), manager(manager) { +ARP_W::ARP_W(Core::System& system_, ARPManager& manager_) + : ServiceFramework{system_, "arp:w"}, manager{manager_} { // clang-format off static const FunctionInfo functions[] = { {0, &ARP_W::AcquireRegistrar, "AcquireRegistrar"}, @@ -254,7 +256,7 @@ void ARP_W::AcquireRegistrar(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_ARP, "called"); registrar = std::make_shared<IRegistrar>( - [this](u64 process_id, ApplicationLaunchProperty launch, std::vector<u8> control) { + system, [this](u64 process_id, ApplicationLaunchProperty launch, std::vector<u8> control) { const auto res = GetTitleIDForProcessID(system, process_id); if (!res.has_value()) { return ERR_NOT_REGISTERED; diff --git a/src/core/hle/service/glue/arp.h b/src/core/hle/service/glue/arp.h index d5f8a7e7a..34b412e26 100644 --- a/src/core/hle/service/glue/arp.h +++ b/src/core/hle/service/glue/arp.h @@ -13,7 +13,7 @@ class IRegistrar; class ARP_R final : public ServiceFramework<ARP_R> { public: - explicit ARP_R(const Core::System& system, const ARPManager& manager); + explicit ARP_R(Core::System& system_, const ARPManager& manager_); ~ARP_R() override; private: @@ -22,20 +22,18 @@ private: void GetApplicationControlProperty(Kernel::HLERequestContext& ctx); void GetApplicationControlPropertyWithApplicationId(Kernel::HLERequestContext& ctx); - const Core::System& system; const ARPManager& manager; }; class ARP_W final : public ServiceFramework<ARP_W> { public: - explicit ARP_W(const Core::System& system, ARPManager& manager); + explicit ARP_W(Core::System& system_, ARPManager& manager_); ~ARP_W() override; private: void AcquireRegistrar(Kernel::HLERequestContext& ctx); void DeleteProperties(Kernel::HLERequestContext& ctx); - const Core::System& system; ARPManager& manager; std::shared_ptr<IRegistrar> registrar; }; diff --git a/src/core/hle/service/glue/bgtc.cpp b/src/core/hle/service/glue/bgtc.cpp index cd89d088f..a478b68e1 100644 --- a/src/core/hle/service/glue/bgtc.cpp +++ b/src/core/hle/service/glue/bgtc.cpp @@ -6,7 +6,7 @@ namespace Service::Glue { -BGTC_T::BGTC_T() : ServiceFramework{"bgtc:t"} { +BGTC_T::BGTC_T(Core::System& system_) : ServiceFramework{system_, "bgtc:t"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "NotifyTaskStarting"}, @@ -31,7 +31,7 @@ BGTC_T::BGTC_T() : ServiceFramework{"bgtc:t"} { BGTC_T::~BGTC_T() = default; -BGTC_SC::BGTC_SC() : ServiceFramework{"bgtc:sc"} { +BGTC_SC::BGTC_SC(Core::System& system_) : ServiceFramework{system_, "bgtc:sc"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "GetState"}, diff --git a/src/core/hle/service/glue/bgtc.h b/src/core/hle/service/glue/bgtc.h index 81844f03e..906116ba6 100644 --- a/src/core/hle/service/glue/bgtc.h +++ b/src/core/hle/service/glue/bgtc.h @@ -6,17 +6,21 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Glue { class BGTC_T final : public ServiceFramework<BGTC_T> { public: - BGTC_T(); + explicit BGTC_T(Core::System& system_); ~BGTC_T() override; }; class BGTC_SC final : public ServiceFramework<BGTC_SC> { public: - BGTC_SC(); + explicit BGTC_SC(Core::System& system_); ~BGTC_SC() override; }; diff --git a/src/core/hle/service/glue/glue.cpp b/src/core/hle/service/glue/glue.cpp index c728e815c..4eafbe5fa 100644 --- a/src/core/hle/service/glue/glue.cpp +++ b/src/core/hle/service/glue/glue.cpp @@ -18,8 +18,8 @@ void InstallInterfaces(Core::System& system) { ->InstallAsService(system.ServiceManager()); // BackGround Task Controller - std::make_shared<BGTC_T>()->InstallAsService(system.ServiceManager()); - std::make_shared<BGTC_SC>()->InstallAsService(system.ServiceManager()); + std::make_shared<BGTC_T>(system)->InstallAsService(system.ServiceManager()); + std::make_shared<BGTC_SC>(system)->InstallAsService(system.ServiceManager()); } } // namespace Service::Glue diff --git a/src/core/hle/service/grc/grc.cpp b/src/core/hle/service/grc/grc.cpp index 401e0b208..a502ab47f 100644 --- a/src/core/hle/service/grc/grc.cpp +++ b/src/core/hle/service/grc/grc.cpp @@ -12,7 +12,7 @@ namespace Service::GRC { class GRC final : public ServiceFramework<GRC> { public: - explicit GRC() : ServiceFramework{"grc:c"} { + explicit GRC(Core::System& system) : ServiceFramework{system, "grc:c"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "OpenContinuousRecorder"}, @@ -27,8 +27,8 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<GRC>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<GRC>(system)->InstallAsService(sm); } } // namespace Service::GRC diff --git a/src/core/hle/service/grc/grc.h b/src/core/hle/service/grc/grc.h index e0d29e70d..9069fe756 100644 --- a/src/core/hle/service/grc/grc.h +++ b/src/core/hle/service/grc/grc.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::GRC { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::GRC diff --git a/src/core/hle/service/hid/controllers/controller_base.h b/src/core/hle/service/hid/controllers/controller_base.h index 8bc69c372..f47a9e61c 100644 --- a/src/core/hle/service/hid/controllers/controller_base.h +++ b/src/core/hle/service/hid/controllers/controller_base.h @@ -31,6 +31,10 @@ public: virtual void OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) = 0; + // When the controller is requesting a motion update for the shared memory + virtual void OnMotionUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, + std::size_t size) {} + // Called when input devices should be loaded virtual void OnLoadInputDevices() = 0; diff --git a/src/core/hle/service/hid/controllers/keyboard.cpp b/src/core/hle/service/hid/controllers/keyboard.cpp index 0b896d5ad..59b694cd4 100644 --- a/src/core/hle/service/hid/controllers/keyboard.cpp +++ b/src/core/hle/service/hid/controllers/keyboard.cpp @@ -42,8 +42,8 @@ void Controller_Keyboard::OnUpdate(const Core::Timing::CoreTiming& core_timing, cur_entry.modifier = 0; if (Settings::values.keyboard_enabled) { for (std::size_t i = 0; i < keyboard_keys.size(); ++i) { - cur_entry.key[i / KEYS_PER_BYTE] |= - (keyboard_keys[i]->GetStatus() << (i % KEYS_PER_BYTE)); + auto& entry = cur_entry.key[i / KEYS_PER_BYTE]; + entry = static_cast<u8>(entry | (keyboard_keys[i]->GetStatus() << (i % KEYS_PER_BYTE))); } for (std::size_t i = 0; i < keyboard_mods.size(); ++i) { diff --git a/src/core/hle/service/hid/controllers/npad.cpp b/src/core/hle/service/hid/controllers/npad.cpp index 620386cd1..d280e7caf 100644 --- a/src/core/hle/service/hid/controllers/npad.cpp +++ b/src/core/hle/service/hid/controllers/npad.cpp @@ -116,8 +116,36 @@ u32 Controller_NPad::IndexToNPad(std::size_t index) { } } +bool Controller_NPad::IsNpadIdValid(u32 npad_id) { + switch (npad_id) { + case 0: + case 1: + case 2: + case 3: + case 4: + case 5: + case 6: + case 7: + case NPAD_UNKNOWN: + case NPAD_HANDHELD: + return true; + default: + LOG_ERROR(Service_HID, "Invalid npad id {}", npad_id); + return false; + } +} + +bool Controller_NPad::IsDeviceHandleValid(const DeviceHandle& device_handle) { + return IsNpadIdValid(device_handle.npad_id) && + device_handle.npad_type < NpadType::MaxNpadType && + device_handle.device_index < DeviceIndex::MaxDeviceIndex; +} + Controller_NPad::Controller_NPad(Core::System& system) : ControllerBase(system), system(system) {} -Controller_NPad::~Controller_NPad() = default; + +Controller_NPad::~Controller_NPad() { + OnRelease(); +} void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) { const auto controller_type = connected_controllers[controller_idx].type; @@ -139,7 +167,7 @@ void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) { controller.properties.is_vertical.Assign(1); controller.properties.use_plus.Assign(1); controller.properties.use_minus.Assign(1); - controller.pad_assignment = NPadAssignments::Single; + controller.pad_assignment = NpadAssignments::Single; break; case NPadControllerType::Handheld: controller.joy_styles.handheld.Assign(1); @@ -147,7 +175,7 @@ void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) { controller.properties.is_vertical.Assign(1); controller.properties.use_plus.Assign(1); controller.properties.use_minus.Assign(1); - controller.pad_assignment = NPadAssignments::Dual; + controller.pad_assignment = NpadAssignments::Dual; break; case NPadControllerType::JoyDual: controller.joy_styles.joycon_dual.Assign(1); @@ -156,26 +184,26 @@ void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) { controller.properties.is_vertical.Assign(1); controller.properties.use_plus.Assign(1); controller.properties.use_minus.Assign(1); - controller.pad_assignment = NPadAssignments::Dual; + controller.pad_assignment = NpadAssignments::Dual; break; case NPadControllerType::JoyLeft: controller.joy_styles.joycon_left.Assign(1); controller.device_type.joycon_left.Assign(1); controller.properties.is_horizontal.Assign(1); controller.properties.use_minus.Assign(1); - controller.pad_assignment = NPadAssignments::Single; + controller.pad_assignment = NpadAssignments::Single; break; case NPadControllerType::JoyRight: controller.joy_styles.joycon_right.Assign(1); controller.device_type.joycon_right.Assign(1); controller.properties.is_horizontal.Assign(1); controller.properties.use_plus.Assign(1); - controller.pad_assignment = NPadAssignments::Single; + controller.pad_assignment = NpadAssignments::Single; break; case NPadControllerType::Pokeball: controller.joy_styles.pokeball.Assign(1); controller.device_type.pokeball.Assign(1); - controller.pad_assignment = NPadAssignments::Single; + controller.pad_assignment = NpadAssignments::Single; break; } @@ -184,11 +212,14 @@ void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) { controller.single_color.button_color = 0; controller.dual_color_error = ColorReadError::ReadOk; - controller.left_color.body_color = Settings::values.players[controller_idx].body_color_left; - controller.left_color.button_color = Settings::values.players[controller_idx].button_color_left; - controller.right_color.body_color = Settings::values.players[controller_idx].body_color_right; + controller.left_color.body_color = + Settings::values.players.GetValue()[controller_idx].body_color_left; + controller.left_color.button_color = + Settings::values.players.GetValue()[controller_idx].button_color_left; + controller.right_color.body_color = + Settings::values.players.GetValue()[controller_idx].body_color_right; controller.right_color.button_color = - Settings::values.players[controller_idx].button_color_right; + Settings::values.players.GetValue()[controller_idx].button_color_right; controller.battery_level[0] = BATTERY_FULL; controller.battery_level[1] = BATTERY_FULL; @@ -199,7 +230,7 @@ void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) { void Controller_NPad::OnInit() { auto& kernel = system.Kernel(); - for (std::size_t i = 0; i < styleset_changed_events.size(); i++) { + for (std::size_t i = 0; i < styleset_changed_events.size(); ++i) { styleset_changed_events[i] = Kernel::WritableEvent::CreateEventPair( kernel, fmt::format("npad:NpadStyleSetChanged_{}", i)); } @@ -208,6 +239,8 @@ void Controller_NPad::OnInit() { return; } + OnLoadInputDevices(); + if (style.raw == 0) { // We want to support all controllers style.handheld.Assign(1); @@ -218,12 +251,27 @@ void Controller_NPad::OnInit() { style.pokeball.Assign(1); } - std::transform(Settings::values.players.begin(), Settings::values.players.end(), - connected_controllers.begin(), [](const Settings::PlayerInput& player) { + std::transform(Settings::values.players.GetValue().begin(), + Settings::values.players.GetValue().end(), connected_controllers.begin(), + [](const Settings::PlayerInput& player) { return ControllerHolder{MapSettingsTypeToNPad(player.controller_type), player.connected}; }); + // Connect the Player 1 or Handheld controller if none are connected. + if (std::none_of(connected_controllers.begin(), connected_controllers.end(), + [](const ControllerHolder& controller) { return controller.is_connected; })) { + const auto controller = + MapSettingsTypeToNPad(Settings::values.players.GetValue()[0].controller_type); + if (controller == NPadControllerType::Handheld) { + Settings::values.players.GetValue()[HANDHELD_INDEX].connected = true; + connected_controllers[HANDHELD_INDEX] = {controller, true}; + } else { + Settings::values.players.GetValue()[0].connected = true; + connected_controllers[0] = {controller, true}; + } + } + // Account for handheld if (connected_controllers[HANDHELD_INDEX].is_connected) { connected_controllers[HANDHELD_INDEX].type = NPadControllerType::Handheld; @@ -242,7 +290,7 @@ void Controller_NPad::OnInit() { } void Controller_NPad::OnLoadInputDevices() { - const auto& players = Settings::values.players; + const auto& players = Settings::values.players.GetValue(); for (std::size_t i = 0; i < players.size(); ++i) { std::transform(players[i].buttons.begin() + Settings::NativeButton::BUTTON_HID_BEGIN, players[i].buttons.begin() + Settings::NativeButton::BUTTON_HID_END, @@ -250,17 +298,30 @@ void Controller_NPad::OnLoadInputDevices() { std::transform(players[i].analogs.begin() + Settings::NativeAnalog::STICK_HID_BEGIN, players[i].analogs.begin() + Settings::NativeAnalog::STICK_HID_END, sticks[i].begin(), Input::CreateDevice<Input::AnalogDevice>); + std::transform(players[i].vibrations.begin() + + Settings::NativeVibration::VIBRATION_HID_BEGIN, + players[i].vibrations.begin() + Settings::NativeVibration::VIBRATION_HID_END, + vibrations[i].begin(), Input::CreateDevice<Input::VibrationDevice>); std::transform(players[i].motions.begin() + Settings::NativeMotion::MOTION_HID_BEGIN, players[i].motions.begin() + Settings::NativeMotion::MOTION_HID_END, motions[i].begin(), Input::CreateDevice<Input::MotionDevice>); + for (std::size_t device_idx = 0; device_idx < vibrations[i].size(); ++device_idx) { + InitializeVibrationDeviceAtIndex(i, device_idx); + } } } -void Controller_NPad::OnRelease() {} +void Controller_NPad::OnRelease() { + for (std::size_t npad_idx = 0; npad_idx < vibrations.size(); ++npad_idx) { + for (std::size_t device_idx = 0; device_idx < vibrations[npad_idx].size(); ++device_idx) { + VibrateControllerAtIndex(npad_idx, device_idx, {}); + } + } +} void Controller_NPad::RequestPadStateUpdate(u32 npad_id) { const auto controller_idx = NPadIdToIndex(npad_id); - [[maybe_unused]] const auto controller_type = connected_controllers[controller_idx].type; + const auto controller_type = connected_controllers[controller_idx].type; if (!connected_controllers[controller_idx].is_connected) { return; } @@ -269,61 +330,69 @@ void Controller_NPad::RequestPadStateUpdate(u32 npad_id) { auto& rstick_entry = npad_pad_states[controller_idx].r_stick; const auto& button_state = buttons[controller_idx]; const auto& analog_state = sticks[controller_idx]; - const auto& motion_state = motions[controller_idx]; const auto [stick_l_x_f, stick_l_y_f] = analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetStatus(); const auto [stick_r_x_f, stick_r_y_f] = analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)]->GetStatus(); using namespace Settings::NativeButton; - pad_state.a.Assign(button_state[A - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.b.Assign(button_state[B - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.x.Assign(button_state[X - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.y.Assign(button_state[Y - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.l_stick.Assign(button_state[LStick - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.r_stick.Assign(button_state[RStick - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.l.Assign(button_state[L - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.r.Assign(button_state[R - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.zl.Assign(button_state[ZL - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.zr.Assign(button_state[ZR - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.plus.Assign(button_state[Plus - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.minus.Assign(button_state[Minus - BUTTON_HID_BEGIN]->GetStatus()); - - pad_state.d_left.Assign(button_state[DLeft - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.d_up.Assign(button_state[DUp - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.d_right.Assign(button_state[DRight - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.d_down.Assign(button_state[DDown - BUTTON_HID_BEGIN]->GetStatus()); - - pad_state.l_stick_right.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::RIGHT)); - pad_state.l_stick_left.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::LEFT)); - pad_state.l_stick_up.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::UP)); - pad_state.l_stick_down.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)]->GetAnalogDirectionStatus( - Input::AnalogDirection::DOWN)); - - pad_state.r_stick_right.Assign( - analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::RIGHT)); - pad_state.r_stick_left.Assign(analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::LEFT)); - pad_state.r_stick_up.Assign(analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::UP)); - pad_state.r_stick_down.Assign(analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] - ->GetAnalogDirectionStatus(Input::AnalogDirection::DOWN)); - - pad_state.left_sl.Assign(button_state[SL - BUTTON_HID_BEGIN]->GetStatus()); - pad_state.left_sr.Assign(button_state[SR - BUTTON_HID_BEGIN]->GetStatus()); - - lstick_entry.x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX); - lstick_entry.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX); - rstick_entry.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX); - rstick_entry.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX); + if (controller_type != NPadControllerType::JoyLeft) { + pad_state.a.Assign(button_state[A - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.b.Assign(button_state[B - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.x.Assign(button_state[X - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.y.Assign(button_state[Y - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.r_stick.Assign(button_state[RStick - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.r.Assign(button_state[R - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.zr.Assign(button_state[ZR - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.plus.Assign(button_state[Plus - BUTTON_HID_BEGIN]->GetStatus()); + + pad_state.r_stick_right.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::RIGHT)); + pad_state.r_stick_left.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::LEFT)); + pad_state.r_stick_up.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::UP)); + pad_state.r_stick_down.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Right)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::DOWN)); + rstick_entry.x = static_cast<s32>(stick_r_x_f * HID_JOYSTICK_MAX); + rstick_entry.y = static_cast<s32>(stick_r_y_f * HID_JOYSTICK_MAX); + } + + if (controller_type != NPadControllerType::JoyRight) { + pad_state.d_left.Assign(button_state[DLeft - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.d_up.Assign(button_state[DUp - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.d_right.Assign(button_state[DRight - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.d_down.Assign(button_state[DDown - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.l_stick.Assign(button_state[LStick - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.l.Assign(button_state[L - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.zl.Assign(button_state[ZL - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.minus.Assign(button_state[Minus - BUTTON_HID_BEGIN]->GetStatus()); + + pad_state.l_stick_right.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::RIGHT)); + pad_state.l_stick_left.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::LEFT)); + pad_state.l_stick_up.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::UP)); + pad_state.l_stick_down.Assign( + analog_state[static_cast<std::size_t>(JoystickId::Joystick_Left)] + ->GetAnalogDirectionStatus(Input::AnalogDirection::DOWN)); + lstick_entry.x = static_cast<s32>(stick_l_x_f * HID_JOYSTICK_MAX); + lstick_entry.y = static_cast<s32>(stick_l_y_f * HID_JOYSTICK_MAX); + } + + if (controller_type == NPadControllerType::JoyLeft || + controller_type == NPadControllerType::JoyRight) { + pad_state.left_sl.Assign(button_state[SL - BUTTON_HID_BEGIN]->GetStatus()); + pad_state.left_sr.Assign(button_state[SR - BUTTON_HID_BEGIN]->GetStatus()); + } } void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, @@ -331,7 +400,7 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* if (!IsControllerActivated()) { return; } - for (std::size_t i = 0; i < shared_memory_entries.size(); i++) { + for (std::size_t i = 0; i < shared_memory_entries.size(); ++i) { auto& npad = shared_memory_entries[i]; const std::array<NPadGeneric*, 7> controller_npads{&npad.main_controller_states, &npad.handheld_states, @@ -365,6 +434,123 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* } const u32 npad_index = static_cast<u32>(i); + RequestPadStateUpdate(npad_index); + auto& pad_state = npad_pad_states[npad_index]; + + auto& main_controller = + npad.main_controller_states.npad[npad.main_controller_states.common.last_entry_index]; + auto& handheld_entry = + npad.handheld_states.npad[npad.handheld_states.common.last_entry_index]; + auto& dual_entry = npad.dual_states.npad[npad.dual_states.common.last_entry_index]; + auto& left_entry = npad.left_joy_states.npad[npad.left_joy_states.common.last_entry_index]; + auto& right_entry = + npad.right_joy_states.npad[npad.right_joy_states.common.last_entry_index]; + auto& pokeball_entry = + npad.pokeball_states.npad[npad.pokeball_states.common.last_entry_index]; + auto& libnx_entry = npad.libnx.npad[npad.libnx.common.last_entry_index]; + + libnx_entry.connection_status.raw = 0; + libnx_entry.connection_status.IsConnected.Assign(1); + + switch (controller_type) { + case NPadControllerType::None: + UNREACHABLE(); + break; + case NPadControllerType::ProController: + main_controller.connection_status.raw = 0; + main_controller.connection_status.IsConnected.Assign(1); + main_controller.connection_status.IsWired.Assign(1); + main_controller.pad.pad_states.raw = pad_state.pad_states.raw; + main_controller.pad.l_stick = pad_state.l_stick; + main_controller.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsWired.Assign(1); + break; + case NPadControllerType::Handheld: + handheld_entry.connection_status.raw = 0; + handheld_entry.connection_status.IsConnected.Assign(1); + handheld_entry.connection_status.IsWired.Assign(1); + handheld_entry.connection_status.IsLeftJoyConnected.Assign(1); + handheld_entry.connection_status.IsRightJoyConnected.Assign(1); + handheld_entry.connection_status.IsLeftJoyWired.Assign(1); + handheld_entry.connection_status.IsRightJoyWired.Assign(1); + handheld_entry.pad.pad_states.raw = pad_state.pad_states.raw; + handheld_entry.pad.l_stick = pad_state.l_stick; + handheld_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsWired.Assign(1); + libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); + libnx_entry.connection_status.IsRightJoyConnected.Assign(1); + libnx_entry.connection_status.IsLeftJoyWired.Assign(1); + libnx_entry.connection_status.IsRightJoyWired.Assign(1); + break; + case NPadControllerType::JoyDual: + dual_entry.connection_status.raw = 0; + dual_entry.connection_status.IsConnected.Assign(1); + dual_entry.connection_status.IsLeftJoyConnected.Assign(1); + dual_entry.connection_status.IsRightJoyConnected.Assign(1); + dual_entry.pad.pad_states.raw = pad_state.pad_states.raw; + dual_entry.pad.l_stick = pad_state.l_stick; + dual_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); + libnx_entry.connection_status.IsRightJoyConnected.Assign(1); + break; + case NPadControllerType::JoyLeft: + left_entry.connection_status.raw = 0; + left_entry.connection_status.IsConnected.Assign(1); + left_entry.connection_status.IsLeftJoyConnected.Assign(1); + left_entry.pad.pad_states.raw = pad_state.pad_states.raw; + left_entry.pad.l_stick = pad_state.l_stick; + left_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); + break; + case NPadControllerType::JoyRight: + right_entry.connection_status.raw = 0; + right_entry.connection_status.IsConnected.Assign(1); + right_entry.connection_status.IsRightJoyConnected.Assign(1); + right_entry.pad.pad_states.raw = pad_state.pad_states.raw; + right_entry.pad.l_stick = pad_state.l_stick; + right_entry.pad.r_stick = pad_state.r_stick; + + libnx_entry.connection_status.IsRightJoyConnected.Assign(1); + break; + case NPadControllerType::Pokeball: + pokeball_entry.connection_status.raw = 0; + pokeball_entry.connection_status.IsConnected.Assign(1); + pokeball_entry.pad.pad_states.raw = pad_state.pad_states.raw; + pokeball_entry.pad.l_stick = pad_state.l_stick; + pokeball_entry.pad.r_stick = pad_state.r_stick; + break; + } + + // LibNX exclusively uses this section, so we always update it since LibNX doesn't activate + // any controllers. + libnx_entry.pad.pad_states.raw = pad_state.pad_states.raw; + libnx_entry.pad.l_stick = pad_state.l_stick; + libnx_entry.pad.r_stick = pad_state.r_stick; + + press_state |= static_cast<u32>(pad_state.pad_states.raw); + } + std::memcpy(data + NPAD_OFFSET, shared_memory_entries.data(), + shared_memory_entries.size() * sizeof(NPadEntry)); +} + +void Controller_NPad::OnMotionUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, + std::size_t data_len) { + if (!IsControllerActivated()) { + return; + } + for (std::size_t i = 0; i < shared_memory_entries.size(); ++i) { + auto& npad = shared_memory_entries[i]; + + const auto& controller_type = connected_controllers[i].type; + + if (controller_type == NPadControllerType::None || !connected_controllers[i].is_connected) { + continue; + } + const std::array<SixAxisGeneric*, 6> controller_sixaxes{ &npad.sixaxis_full, &npad.sixaxis_handheld, &npad.sixaxis_dual_left, &npad.sixaxis_dual_right, &npad.sixaxis_left, &npad.sixaxis_right, @@ -390,7 +576,7 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* // Try to read sixaxis sensor states std::array<MotionDevice, 2> motion_devices; - if (sixaxis_sensors_enabled && Settings::values.motion_enabled) { + if (sixaxis_sensors_enabled && Settings::values.motion_enabled.GetValue()) { sixaxis_at_rest = true; for (std::size_t e = 0; e < motion_devices.size(); ++e) { const auto& device = motions[i][e]; @@ -403,23 +589,6 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* } } - RequestPadStateUpdate(npad_index); - auto& pad_state = npad_pad_states[npad_index]; - - auto& main_controller = - npad.main_controller_states.npad[npad.main_controller_states.common.last_entry_index]; - auto& handheld_entry = - npad.handheld_states.npad[npad.handheld_states.common.last_entry_index]; - auto& dual_entry = npad.dual_states.npad[npad.dual_states.common.last_entry_index]; - auto& left_entry = npad.left_joy_states.npad[npad.left_joy_states.common.last_entry_index]; - auto& right_entry = - npad.right_joy_states.npad[npad.right_joy_states.common.last_entry_index]; - auto& pokeball_entry = - npad.pokeball_states.npad[npad.pokeball_states.common.last_entry_index]; - auto& libnx_entry = npad.libnx.npad[npad.libnx.common.last_entry_index]; - - libnx_entry.connection_status.raw = 0; - libnx_entry.connection_status.IsConnected.Assign(1); auto& full_sixaxis_entry = npad.sixaxis_full.sixaxis[npad.sixaxis_full.common.last_entry_index]; auto& handheld_sixaxis_entry = @@ -438,15 +607,6 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* UNREACHABLE(); break; case NPadControllerType::ProController: - main_controller.connection_status.raw = 0; - main_controller.connection_status.IsConnected.Assign(1); - main_controller.connection_status.IsWired.Assign(1); - main_controller.pad.pad_states.raw = pad_state.pad_states.raw; - main_controller.pad.l_stick = pad_state.l_stick; - main_controller.pad.r_stick = pad_state.r_stick; - - libnx_entry.connection_status.IsWired.Assign(1); - if (sixaxis_sensors_enabled && motions[i][0]) { full_sixaxis_entry.accel = motion_devices[0].accel; full_sixaxis_entry.gyro = motion_devices[0].gyro; @@ -455,23 +615,6 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* } break; case NPadControllerType::Handheld: - handheld_entry.connection_status.raw = 0; - handheld_entry.connection_status.IsConnected.Assign(1); - handheld_entry.connection_status.IsWired.Assign(1); - handheld_entry.connection_status.IsLeftJoyConnected.Assign(1); - handheld_entry.connection_status.IsRightJoyConnected.Assign(1); - handheld_entry.connection_status.IsLeftJoyWired.Assign(1); - handheld_entry.connection_status.IsRightJoyWired.Assign(1); - handheld_entry.pad.pad_states.raw = pad_state.pad_states.raw; - handheld_entry.pad.l_stick = pad_state.l_stick; - handheld_entry.pad.r_stick = pad_state.r_stick; - - libnx_entry.connection_status.IsWired.Assign(1); - libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); - libnx_entry.connection_status.IsRightJoyConnected.Assign(1); - libnx_entry.connection_status.IsLeftJoyWired.Assign(1); - libnx_entry.connection_status.IsRightJoyWired.Assign(1); - if (sixaxis_sensors_enabled && motions[i][0]) { handheld_sixaxis_entry.accel = motion_devices[0].accel; handheld_sixaxis_entry.gyro = motion_devices[0].gyro; @@ -480,17 +623,6 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* } break; case NPadControllerType::JoyDual: - dual_entry.connection_status.raw = 0; - dual_entry.connection_status.IsConnected.Assign(1); - dual_entry.connection_status.IsLeftJoyConnected.Assign(1); - dual_entry.connection_status.IsRightJoyConnected.Assign(1); - dual_entry.pad.pad_states.raw = pad_state.pad_states.raw; - dual_entry.pad.l_stick = pad_state.l_stick; - dual_entry.pad.r_stick = pad_state.r_stick; - - libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); - libnx_entry.connection_status.IsRightJoyConnected.Assign(1); - if (sixaxis_sensors_enabled && motions[i][0]) { // Set motion for the left joycon dual_left_sixaxis_entry.accel = motion_devices[0].accel; @@ -507,15 +639,6 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* } break; case NPadControllerType::JoyLeft: - left_entry.connection_status.raw = 0; - left_entry.connection_status.IsConnected.Assign(1); - left_entry.connection_status.IsLeftJoyConnected.Assign(1); - left_entry.pad.pad_states.raw = pad_state.pad_states.raw; - left_entry.pad.l_stick = pad_state.l_stick; - left_entry.pad.r_stick = pad_state.r_stick; - - libnx_entry.connection_status.IsLeftJoyConnected.Assign(1); - if (sixaxis_sensors_enabled && motions[i][0]) { left_sixaxis_entry.accel = motion_devices[0].accel; left_sixaxis_entry.gyro = motion_devices[0].gyro; @@ -524,15 +647,6 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* } break; case NPadControllerType::JoyRight: - right_entry.connection_status.raw = 0; - right_entry.connection_status.IsConnected.Assign(1); - right_entry.connection_status.IsRightJoyConnected.Assign(1); - right_entry.pad.pad_states.raw = pad_state.pad_states.raw; - right_entry.pad.l_stick = pad_state.l_stick; - right_entry.pad.r_stick = pad_state.r_stick; - - libnx_entry.connection_status.IsRightJoyConnected.Assign(1); - if (sixaxis_sensors_enabled && motions[i][1]) { right_sixaxis_entry.accel = motion_devices[1].accel; right_sixaxis_entry.gyro = motion_devices[1].gyro; @@ -541,35 +655,22 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* } break; case NPadControllerType::Pokeball: - pokeball_entry.connection_status.raw = 0; - pokeball_entry.connection_status.IsConnected.Assign(1); - pokeball_entry.pad.pad_states.raw = pad_state.pad_states.raw; - pokeball_entry.pad.l_stick = pad_state.l_stick; - pokeball_entry.pad.r_stick = pad_state.r_stick; break; } - - // LibNX exclusively uses this section, so we always update it since LibNX doesn't activate - // any controllers. - libnx_entry.pad.pad_states.raw = pad_state.pad_states.raw; - libnx_entry.pad.l_stick = pad_state.l_stick; - libnx_entry.pad.r_stick = pad_state.r_stick; - - press_state |= static_cast<u32>(pad_state.pad_states.raw); } std::memcpy(data + NPAD_OFFSET, shared_memory_entries.data(), shared_memory_entries.size() * sizeof(NPadEntry)); } -void Controller_NPad::SetSupportedStyleSet(NPadType style_set) { +void Controller_NPad::SetSupportedStyleSet(NpadStyleSet style_set) { style.raw = style_set.raw; } -Controller_NPad::NPadType Controller_NPad::GetSupportedStyleSet() const { +Controller_NPad::NpadStyleSet Controller_NPad::GetSupportedStyleSet() const { return style; } -void Controller_NPad::SetSupportedNPadIdTypes(u8* data, std::size_t length) { +void Controller_NPad::SetSupportedNpadIdTypes(u8* data, std::size_t length) { ASSERT(length > 0 && (length % sizeof(u32)) == 0); supported_npad_id_types.clear(); supported_npad_id_types.resize(length / sizeof(u32)); @@ -581,7 +682,7 @@ void Controller_NPad::GetSupportedNpadIdTypes(u32* data, std::size_t max_length) std::memcpy(data, supported_npad_id_types.data(), supported_npad_id_types.size()); } -std::size_t Controller_NPad::GetSupportedNPadIdTypesSize() const { +std::size_t Controller_NPad::GetSupportedNpadIdTypesSize() const { return supported_npad_id_types.size(); } @@ -601,7 +702,15 @@ Controller_NPad::NpadHandheldActivationMode Controller_NPad::GetNpadHandheldActi return handheld_activation_mode; } -void Controller_NPad::SetNpadMode(u32 npad_id, NPadAssignments assignment_mode) { +void Controller_NPad::SetNpadCommunicationMode(NpadCommunicationMode communication_mode_) { + communication_mode = communication_mode_; +} + +Controller_NPad::NpadCommunicationMode Controller_NPad::GetNpadCommunicationMode() const { + return communication_mode; +} + +void Controller_NPad::SetNpadMode(u32 npad_id, NpadAssignments assignment_mode) { const std::size_t npad_index = NPadIdToIndex(npad_id); ASSERT(npad_index < shared_memory_entries.size()); if (shared_memory_entries[npad_index].pad_assignment != assignment_mode) { @@ -609,24 +718,156 @@ void Controller_NPad::SetNpadMode(u32 npad_id, NPadAssignments assignment_mode) } } -void Controller_NPad::VibrateController(const std::vector<u32>& controller_ids, - const std::vector<Vibration>& vibrations) { - LOG_DEBUG(Service_HID, "(STUBBED) called"); +bool Controller_NPad::VibrateControllerAtIndex(std::size_t npad_index, std::size_t device_index, + const VibrationValue& vibration_value) { + if (!connected_controllers[npad_index].is_connected || !vibrations[npad_index][device_index]) { + return false; + } - if (!Settings::values.vibration_enabled || !can_controllers_vibrate) { - return; + const auto& player = Settings::values.players.GetValue()[npad_index]; + + if (!player.vibration_enabled) { + if (latest_vibration_values[npad_index][device_index].amp_low != 0.0f || + latest_vibration_values[npad_index][device_index].amp_high != 0.0f) { + // Send an empty vibration to stop any vibrations. + vibrations[npad_index][device_index]->SetRumblePlay(0.0f, 160.0f, 0.0f, 320.0f); + // Then reset the vibration value to its default value. + latest_vibration_values[npad_index][device_index] = {}; + } + + return false; } - for (std::size_t i = 0; i < controller_ids.size(); i++) { - std::size_t controller_pos = NPadIdToIndex(static_cast<u32>(i)); - if (connected_controllers[controller_pos].is_connected) { - // TODO(ogniK): Vibrate the physical controller + + if (!Settings::values.enable_accurate_vibrations.GetValue()) { + using std::chrono::duration_cast; + using std::chrono::milliseconds; + using std::chrono::steady_clock; + + const auto now = steady_clock::now(); + + // Filter out non-zero vibrations that are within 10ms of each other. + if ((vibration_value.amp_low != 0.0f || vibration_value.amp_high != 0.0f) && + duration_cast<milliseconds>(now - last_vibration_timepoints[npad_index][device_index]) < + milliseconds(10)) { + return false; } + + last_vibration_timepoints[npad_index][device_index] = now; + } + + auto& vibration = vibrations[npad_index][device_index]; + const auto player_vibration_strength = static_cast<f32>(player.vibration_strength); + const auto amp_low = + std::min(vibration_value.amp_low * player_vibration_strength / 100.0f, 1.0f); + const auto amp_high = + std::min(vibration_value.amp_high * player_vibration_strength / 100.0f, 1.0f); + return vibration->SetRumblePlay(amp_low, vibration_value.freq_low, amp_high, + vibration_value.freq_high); +} + +void Controller_NPad::VibrateController(const DeviceHandle& vibration_device_handle, + const VibrationValue& vibration_value) { + if (!IsDeviceHandleValid(vibration_device_handle)) { + return; + } + + if (!Settings::values.vibration_enabled.GetValue() && !permit_vibration_session_enabled) { + return; + } + + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + + if (!vibration_devices_mounted[npad_index][device_index] || + !connected_controllers[npad_index].is_connected) { + return; + } + + if (vibration_device_handle.device_index == DeviceIndex::None) { + UNREACHABLE_MSG("DeviceIndex should never be None!"); + return; + } + + // Some games try to send mismatched parameters in the device handle, block these. + if ((connected_controllers[npad_index].type == NPadControllerType::JoyLeft && + (vibration_device_handle.npad_type == NpadType::JoyconRight || + vibration_device_handle.device_index == DeviceIndex::Right)) || + (connected_controllers[npad_index].type == NPadControllerType::JoyRight && + (vibration_device_handle.npad_type == NpadType::JoyconLeft || + vibration_device_handle.device_index == DeviceIndex::Left))) { + return; + } + + // Filter out vibrations with equivalent values to reduce unnecessary state changes. + if (vibration_value.amp_low == latest_vibration_values[npad_index][device_index].amp_low && + vibration_value.amp_high == latest_vibration_values[npad_index][device_index].amp_high) { + return; + } + + if (VibrateControllerAtIndex(npad_index, device_index, vibration_value)) { + latest_vibration_values[npad_index][device_index] = vibration_value; } - last_processed_vibration = vibrations.back(); } -Controller_NPad::Vibration Controller_NPad::GetLastVibration() const { - return last_processed_vibration; +void Controller_NPad::VibrateControllers(const std::vector<DeviceHandle>& vibration_device_handles, + const std::vector<VibrationValue>& vibration_values) { + if (!Settings::values.vibration_enabled.GetValue() && !permit_vibration_session_enabled) { + return; + } + + ASSERT_OR_EXECUTE_MSG( + vibration_device_handles.size() == vibration_values.size(), { return; }, + "The amount of device handles does not match with the amount of vibration values," + "this is undefined behavior!"); + + for (std::size_t i = 0; i < vibration_device_handles.size(); ++i) { + VibrateController(vibration_device_handles[i], vibration_values[i]); + } +} + +Controller_NPad::VibrationValue Controller_NPad::GetLastVibration( + const DeviceHandle& vibration_device_handle) const { + if (!IsDeviceHandleValid(vibration_device_handle)) { + return {}; + } + + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + return latest_vibration_values[npad_index][device_index]; +} + +void Controller_NPad::InitializeVibrationDevice(const DeviceHandle& vibration_device_handle) { + if (!IsDeviceHandleValid(vibration_device_handle)) { + return; + } + + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + InitializeVibrationDeviceAtIndex(npad_index, device_index); +} + +void Controller_NPad::InitializeVibrationDeviceAtIndex(std::size_t npad_index, + std::size_t device_index) { + if (vibrations[npad_index][device_index]) { + vibration_devices_mounted[npad_index][device_index] = + vibrations[npad_index][device_index]->GetStatus() == 1; + } else { + vibration_devices_mounted[npad_index][device_index] = false; + } +} + +void Controller_NPad::SetPermitVibrationSession(bool permit_vibration_session) { + permit_vibration_session_enabled = permit_vibration_session; +} + +bool Controller_NPad::IsVibrationDeviceMounted(const DeviceHandle& vibration_device_handle) const { + if (!IsDeviceHandleValid(vibration_device_handle)) { + return false; + } + + const auto npad_index = NPadIdToIndex(vibration_device_handle.npad_id); + const auto device_index = static_cast<std::size_t>(vibration_device_handle.device_index); + return vibration_devices_mounted[npad_index][device_index]; } std::shared_ptr<Kernel::ReadableEvent> Controller_NPad::GetStyleSetChangedEvent(u32 npad_id) const { @@ -645,31 +886,38 @@ void Controller_NPad::AddNewControllerAt(NPadControllerType controller, std::siz void Controller_NPad::UpdateControllerAt(NPadControllerType controller, std::size_t npad_index, bool connected) { if (!connected) { - DisconnectNPadAtIndex(npad_index); + DisconnectNpadAtIndex(npad_index); return; } if (controller == NPadControllerType::Handheld) { - Settings::values.players[HANDHELD_INDEX].controller_type = + Settings::values.players.GetValue()[HANDHELD_INDEX].controller_type = MapNPadToSettingsType(controller); - Settings::values.players[HANDHELD_INDEX].connected = true; + Settings::values.players.GetValue()[HANDHELD_INDEX].connected = true; connected_controllers[HANDHELD_INDEX] = {controller, true}; InitNewlyAddedController(HANDHELD_INDEX); return; } - Settings::values.players[npad_index].controller_type = MapNPadToSettingsType(controller); - Settings::values.players[npad_index].connected = true; + Settings::values.players.GetValue()[npad_index].controller_type = + MapNPadToSettingsType(controller); + Settings::values.players.GetValue()[npad_index].connected = true; connected_controllers[npad_index] = {controller, true}; InitNewlyAddedController(npad_index); } -void Controller_NPad::DisconnectNPad(u32 npad_id) { - DisconnectNPadAtIndex(NPadIdToIndex(npad_id)); +void Controller_NPad::DisconnectNpad(u32 npad_id) { + DisconnectNpadAtIndex(NPadIdToIndex(npad_id)); } -void Controller_NPad::DisconnectNPadAtIndex(std::size_t npad_index) { - Settings::values.players[npad_index].connected = false; +void Controller_NPad::DisconnectNpadAtIndex(std::size_t npad_index) { + for (std::size_t device_idx = 0; device_idx < vibrations[npad_index].size(); ++device_idx) { + // Send an empty vibration to stop any vibrations. + VibrateControllerAtIndex(npad_index, device_idx, {}); + vibration_devices_mounted[npad_index][device_idx] = false; + } + + Settings::values.players.GetValue()[npad_index].connected = false; connected_controllers[npad_index].is_connected = false; auto& controller = shared_memory_entries[npad_index]; @@ -707,7 +955,7 @@ void Controller_NPad::MergeSingleJoyAsDualJoy(u32 npad_id_1, u32 npad_id_2) { (connected_controllers[npad_index_2].type == NPadControllerType::JoyLeft && connected_controllers[npad_index_1].type == NPadControllerType::JoyRight)) { // Disconnect the joycon at the second id and connect the dual joycon at the first index. - DisconnectNPad(npad_id_2); + DisconnectNpad(npad_id_2); AddNewControllerAt(NPadControllerType::JoyDual, npad_index_1); } } @@ -770,12 +1018,13 @@ Controller_NPad::LedPattern Controller_NPad::GetLedPattern(u32 npad_id) { } } -void Controller_NPad::SetVibrationEnabled(bool can_vibrate) { - can_controllers_vibrate = can_vibrate; +bool Controller_NPad::IsUnintendedHomeButtonInputProtectionEnabled(u32 npad_id) const { + return unintended_home_button_input_protection[NPadIdToIndex(npad_id)]; } -bool Controller_NPad::IsVibrationEnabled() const { - return can_controllers_vibrate; +void Controller_NPad::SetUnintendedHomeButtonInputProtectionEnabled(bool is_protection_enabled, + u32 npad_id) { + unintended_home_button_input_protection[NPadIdToIndex(npad_id)] = is_protection_enabled; } void Controller_NPad::ClearAllConnectedControllers() { @@ -809,7 +1058,7 @@ void Controller_NPad::ClearAllControllers() { } u32 Controller_NPad::GetAndResetPressState() { - return std::exchange(press_state, 0); + return press_state.exchange(0); } bool Controller_NPad::IsControllerSupported(NPadControllerType controller) const { @@ -822,7 +1071,7 @@ bool Controller_NPad::IsControllerSupported(NPadControllerType controller) const return false; } // Handheld should not be supported in docked mode - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { return false; } diff --git a/src/core/hle/service/hid/controllers/npad.h b/src/core/hle/service/hid/controllers/npad.h index 654d97c3f..e2e826623 100644 --- a/src/core/hle/service/hid/controllers/npad.h +++ b/src/core/hle/service/hid/controllers/npad.h @@ -5,6 +5,7 @@ #pragma once #include <array> +#include <atomic> #include "common/bit_field.h" #include "common/common_types.h" #include "core/frontend/input.h" @@ -32,31 +33,39 @@ public: // When the controller is requesting an update for the shared memory void OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, std::size_t size) override; + // When the controller is requesting a motion update for the shared memory + void OnMotionUpdate(const Core::Timing::CoreTiming& core_timing, u8* data, + std::size_t size) override; + // Called when input devices should be loaded void OnLoadInputDevices() override; - struct NPadType { - union { - u32_le raw{}; - - BitField<0, 1, u32> pro_controller; - BitField<1, 1, u32> handheld; - BitField<2, 1, u32> joycon_dual; - BitField<3, 1, u32> joycon_left; - BitField<4, 1, u32> joycon_right; + enum class NPadControllerType { + None, + ProController, + Handheld, + JoyDual, + JoyLeft, + JoyRight, + Pokeball, + }; - BitField<6, 1, u32> pokeball; // TODO(ogniK): Confirm when possible - }; + enum class NpadType : u8 { + ProController = 3, + Handheld = 4, + JoyconDual = 5, + JoyconLeft = 6, + JoyconRight = 7, + Pokeball = 9, + MaxNpadType = 10, }; - static_assert(sizeof(NPadType) == 4, "NPadType is an invalid size"); - struct Vibration { - f32 amp_low; - f32 freq_low; - f32 amp_high; - f32 freq_high; + enum class DeviceIndex : u8 { + Left = 0, + Right = 1, + None = 2, + MaxDeviceIndex = 3, }; - static_assert(sizeof(Vibration) == 0x10, "Vibration is an invalid size"); enum class GyroscopeZeroDriftMode : u32 { Loose = 0, @@ -69,7 +78,7 @@ public: Horizontal = 1, }; - enum class NPadAssignments : u32_le { + enum class NpadAssignments : u32 { Dual = 0, Single = 1, }; @@ -80,15 +89,43 @@ public: None = 2, }; - enum class NPadControllerType { - None, - ProController, - Handheld, - JoyDual, - JoyLeft, - JoyRight, - Pokeball, + enum class NpadCommunicationMode : u64 { + Unknown0 = 0, + Unknown1 = 1, + Unknown2 = 2, + Unknown3 = 3, + }; + + struct DeviceHandle { + NpadType npad_type{}; + u8 npad_id{}; + DeviceIndex device_index{}; + INSERT_PADDING_BYTES(1); + }; + static_assert(sizeof(DeviceHandle) == 4, "DeviceHandle is an invalid size"); + + struct NpadStyleSet { + union { + u32_le raw{}; + + BitField<0, 1, u32> pro_controller; + BitField<1, 1, u32> handheld; + BitField<2, 1, u32> joycon_dual; + BitField<3, 1, u32> joycon_left; + BitField<4, 1, u32> joycon_right; + + BitField<6, 1, u32> pokeball; // TODO(ogniK): Confirm when possible + }; }; + static_assert(sizeof(NpadStyleSet) == 4, "NpadStyleSet is an invalid size"); + + struct VibrationValue { + f32 amp_low{0.0f}; + f32 freq_low{160.0f}; + f32 amp_high{0.0f}; + f32 freq_high{320.0f}; + }; + static_assert(sizeof(VibrationValue) == 0x10, "Vibration is an invalid size"); struct LedPattern { explicit LedPattern(u64 light1, u64 light2, u64 light3, u64 light4) { @@ -106,12 +143,12 @@ public: }; }; - void SetSupportedStyleSet(NPadType style_set); - NPadType GetSupportedStyleSet() const; + void SetSupportedStyleSet(NpadStyleSet style_set); + NpadStyleSet GetSupportedStyleSet() const; - void SetSupportedNPadIdTypes(u8* data, std::size_t length); + void SetSupportedNpadIdTypes(u8* data, std::size_t length); void GetSupportedNpadIdTypes(u32* data, std::size_t max_length); - std::size_t GetSupportedNPadIdTypesSize() const; + std::size_t GetSupportedNpadIdTypesSize() const; void SetHoldType(NpadHoldType joy_hold_type); NpadHoldType GetHoldType() const; @@ -119,12 +156,29 @@ public: void SetNpadHandheldActivationMode(NpadHandheldActivationMode activation_mode); NpadHandheldActivationMode GetNpadHandheldActivationMode() const; - void SetNpadMode(u32 npad_id, NPadAssignments assignment_mode); + void SetNpadCommunicationMode(NpadCommunicationMode communication_mode_); + NpadCommunicationMode GetNpadCommunicationMode() const; + + void SetNpadMode(u32 npad_id, NpadAssignments assignment_mode); + + bool VibrateControllerAtIndex(std::size_t npad_index, std::size_t device_index, + const VibrationValue& vibration_value); + + void VibrateController(const DeviceHandle& vibration_device_handle, + const VibrationValue& vibration_value); + + void VibrateControllers(const std::vector<DeviceHandle>& vibration_device_handles, + const std::vector<VibrationValue>& vibration_values); + + VibrationValue GetLastVibration(const DeviceHandle& vibration_device_handle) const; + + void InitializeVibrationDevice(const DeviceHandle& vibration_device_handle); + + void InitializeVibrationDeviceAtIndex(std::size_t npad_index, std::size_t device_index); - void VibrateController(const std::vector<u32>& controller_ids, - const std::vector<Vibration>& vibrations); + void SetPermitVibrationSession(bool permit_vibration_session); - Vibration GetLastVibration() const; + bool IsVibrationDeviceMounted(const DeviceHandle& vibration_device_handle) const; std::shared_ptr<Kernel::ReadableEvent> GetStyleSetChangedEvent(u32 npad_id) const; void SignalStyleSetChangedEvent(u32 npad_id) const; @@ -134,16 +188,16 @@ public: // Adds a new controller at an index with connection status. void UpdateControllerAt(NPadControllerType controller, std::size_t npad_index, bool connected); - void DisconnectNPad(u32 npad_id); - void DisconnectNPadAtIndex(std::size_t index); + void DisconnectNpad(u32 npad_id); + void DisconnectNpadAtIndex(std::size_t index); void SetGyroscopeZeroDriftMode(GyroscopeZeroDriftMode drift_mode); GyroscopeZeroDriftMode GetGyroscopeZeroDriftMode() const; bool IsSixAxisSensorAtRest() const; void SetSixAxisEnabled(bool six_axis_status); LedPattern GetLedPattern(u32 npad_id); - void SetVibrationEnabled(bool can_vibrate); - bool IsVibrationEnabled() const; + bool IsUnintendedHomeButtonInputProtectionEnabled(u32 npad_id) const; + void SetUnintendedHomeButtonInputProtectionEnabled(bool is_protection_enabled, u32 npad_id); void ClearAllConnectedControllers(); void DisconnectAllConnectedControllers(); void ConnectAllDisconnectedControllers(); @@ -162,6 +216,8 @@ public: static Settings::ControllerType MapNPadToSettingsType(Controller_NPad::NPadControllerType type); static std::size_t NPadIdToIndex(u32 npad_id); static u32 IndexToNPad(std::size_t index); + static bool IsNpadIdValid(u32 npad_id); + static bool IsDeviceHandleValid(const DeviceHandle& device_handle); private: struct CommonHeader { @@ -318,8 +374,8 @@ private: }; struct NPadEntry { - NPadType joy_styles; - NPadAssignments pad_assignment; + NpadStyleSet joy_styles; + NpadAssignments pad_assignment; ColorReadError single_color_error; ControllerColor single_color; @@ -360,9 +416,9 @@ private: bool IsControllerSupported(NPadControllerType controller) const; void RequestPadStateUpdate(u32 npad_id); - u32 press_state{}; + std::atomic<u32> press_state{}; - NPadType style{}; + NpadStyleSet style{}; std::array<NPadEntry, 10> shared_memory_entries{}; using ButtonArray = std::array< std::array<std::unique_ptr<Input::ButtonDevice>, Settings::NativeButton::NUM_BUTTONS_HID>, @@ -370,21 +426,30 @@ private: using StickArray = std::array< std::array<std::unique_ptr<Input::AnalogDevice>, Settings::NativeAnalog::NUM_STICKS_HID>, 10>; + using VibrationArray = std::array<std::array<std::unique_ptr<Input::VibrationDevice>, + Settings::NativeVibration::NUM_VIBRATIONS_HID>, + 10>; using MotionArray = std::array< - std::array<std::unique_ptr<Input::MotionDevice>, Settings::NativeMotion::NUM_MOTION_HID>, + std::array<std::unique_ptr<Input::MotionDevice>, Settings::NativeMotion::NUM_MOTIONS_HID>, 10>; ButtonArray buttons; StickArray sticks; + VibrationArray vibrations; MotionArray motions; std::vector<u32> supported_npad_id_types{}; NpadHoldType hold_type{NpadHoldType::Vertical}; NpadHandheldActivationMode handheld_activation_mode{NpadHandheldActivationMode::Dual}; + // NpadCommunicationMode is unknown, default value is 1 + NpadCommunicationMode communication_mode{NpadCommunicationMode::Unknown1}; // Each controller should have their own styleset changed event std::array<Kernel::EventPair, 10> styleset_changed_events; - Vibration last_processed_vibration{}; + std::array<std::array<std::chrono::steady_clock::time_point, 2>, 10> last_vibration_timepoints; + std::array<std::array<VibrationValue, 2>, 10> latest_vibration_values{}; + bool permit_vibration_session_enabled{false}; + std::array<std::array<bool, 2>, 10> vibration_devices_mounted{}; std::array<ControllerHolder, 10> connected_controllers{}; + std::array<bool, 10> unintended_home_button_input_protection{}; GyroscopeZeroDriftMode gyroscope_zero_drift_mode{GyroscopeZeroDriftMode::Standard}; - bool can_controllers_vibrate{true}; bool sixaxis_sensors_enabled{true}; bool sixaxis_at_rest{true}; std::array<ControllerPad, 10> npad_pad_states{}; diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp index 395e83b3f..8d95f74e6 100644 --- a/src/core/hle/service/hid/hid.cpp +++ b/src/core/hle/service/hid/hid.cpp @@ -40,11 +40,12 @@ namespace Service::HID { // Updating period for each HID device. // HID is polled every 15ms, this value was derived from // https://github.com/dekuNukem/Nintendo_Switch_Reverse_Engineering#joy-con-status-data-packet -constexpr auto pad_update_ns = std::chrono::nanoseconds{1000 * 1000}; // (1ms, 1000Hz) +constexpr auto pad_update_ns = std::chrono::nanoseconds{1000 * 1000}; // (1ms, 1000Hz) +constexpr auto motion_update_ns = std::chrono::nanoseconds{15 * 1000 * 1000}; // (15ms, 66.666Hz) constexpr std::size_t SHARED_MEMORY_SIZE = 0x40000; -IAppletResource::IAppletResource(Core::System& system) - : ServiceFramework("IAppletResource"), system(system) { +IAppletResource::IAppletResource(Core::System& system_) + : ServiceFramework{system_, "IAppletResource"} { static const FunctionInfo functions[] = { {0, &IAppletResource::GetSharedMemoryHandle, "GetSharedMemoryHandle"}, }; @@ -77,12 +78,18 @@ IAppletResource::IAppletResource(Core::System& system) pad_update_event = Core::Timing::CreateEvent( "HID::UpdatePadCallback", [this](std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + const auto guard = LockService(); UpdateControllers(user_data, ns_late); }); - - // TODO(shinyquagsire23): Other update callbacks? (accel, gyro?) + motion_update_event = Core::Timing::CreateEvent( + "HID::MotionPadCallback", + [this](std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + const auto guard = LockService(); + UpdateMotion(user_data, ns_late); + }); system.CoreTiming().ScheduleEvent(pad_update_ns, pad_update_event); + system.CoreTiming().ScheduleEvent(motion_update_ns, motion_update_event); ReloadInputDevices(); } @@ -122,22 +129,50 @@ void IAppletResource::UpdateControllers(std::uintptr_t user_data, core_timing.ScheduleEvent(pad_update_ns - ns_late, pad_update_event); } +void IAppletResource::UpdateMotion(std::uintptr_t user_data, std::chrono::nanoseconds ns_late) { + auto& core_timing = system.CoreTiming(); + + for (const auto& controller : controllers) { + controller->OnMotionUpdate(core_timing, shared_mem->GetPointer(), SHARED_MEMORY_SIZE); + } + + core_timing.ScheduleEvent(motion_update_ns - ns_late, motion_update_event); +} + class IActiveVibrationDeviceList final : public ServiceFramework<IActiveVibrationDeviceList> { public: - IActiveVibrationDeviceList() : ServiceFramework("IActiveVibrationDeviceList") { + explicit IActiveVibrationDeviceList(Core::System& system_, + std::shared_ptr<IAppletResource> applet_resource_) + : ServiceFramework{system_, "IActiveVibrationDeviceList"}, + applet_resource(applet_resource_) { + // clang-format off static const FunctionInfo functions[] = { - {0, &IActiveVibrationDeviceList::ActivateVibrationDevice, "ActivateVibrationDevice"}, + {0, &IActiveVibrationDeviceList::InitializeVibrationDevice, "InitializeVibrationDevice"}, }; + // clang-format on + RegisterHandlers(functions); } private: - void ActivateVibrationDevice(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_HID, "(STUBBED) called"); + void InitializeVibrationDevice(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto vibration_device_handle{rp.PopRaw<Controller_NPad::DeviceHandle>()}; + + if (applet_resource != nullptr) { + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .InitializeVibrationDevice(vibration_device_handle); + } + + LOG_DEBUG(Service_HID, "called, npad_type={}, npad_id={}, device_index={}", + vibration_device_handle.npad_type, vibration_device_handle.npad_id, + vibration_device_handle.device_index); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } + + std::shared_ptr<IAppletResource> applet_resource; }; std::shared_ptr<IAppletResource> Hid::GetAppletResource() { @@ -148,7 +183,7 @@ std::shared_ptr<IAppletResource> Hid::GetAppletResource() { return applet_resource; } -Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { +Hid::Hid(Core::System& system_) : ServiceFramework{system_, "hid"} { // clang-format off static const FunctionInfo functions[] = { {0, &Hid::CreateAppletResource, "CreateAppletResource"}, @@ -173,7 +208,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {66, &Hid::StartSixAxisSensor, "StartSixAxisSensor"}, {67, &Hid::StopSixAxisSensor, "StopSixAxisSensor"}, {68, nullptr, "IsSixAxisSensorFusionEnabled"}, - {69, nullptr, "EnableSixAxisSensorFusion"}, + {69, &Hid::EnableSixAxisSensorFusion, "EnableSixAxisSensorFusion"}, {70, nullptr, "SetSixAxisSensorFusionParameters"}, {71, nullptr, "GetSixAxisSensorFusionParameters"}, {72, nullptr, "ResetSixAxisSensorFusionParameters"}, @@ -209,8 +244,8 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {128, &Hid::SetNpadHandheldActivationMode, "SetNpadHandheldActivationMode"}, {129, &Hid::GetNpadHandheldActivationMode, "GetNpadHandheldActivationMode"}, {130, &Hid::SwapNpadAssignment, "SwapNpadAssignment"}, - {131, nullptr, "IsUnintendedHomeButtonInputProtectionEnabled"}, - {132, nullptr, "EnableUnintendedHomeButtonInputProtection"}, + {131, &Hid::IsUnintendedHomeButtonInputProtectionEnabled, "IsUnintendedHomeButtonInputProtectionEnabled"}, + {132, &Hid::EnableUnintendedHomeButtonInputProtection, "EnableUnintendedHomeButtonInputProtection"}, {133, nullptr, "SetNpadJoyAssignmentModeSingleWithDestination"}, {134, nullptr, "SetNpadAnalogStickUseCenterClamp"}, {135, nullptr, "SetNpadCaptureButtonAssignment"}, @@ -226,7 +261,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {208, nullptr, "GetActualVibrationGcErmCommand"}, {209, &Hid::BeginPermitVibrationSession, "BeginPermitVibrationSession"}, {210, &Hid::EndPermitVibrationSession, "EndPermitVibrationSession"}, - {211, nullptr, "IsVibrationDeviceMounted"}, + {211, &Hid::IsVibrationDeviceMounted, "IsVibrationDeviceMounted"}, {300, &Hid::ActivateConsoleSixAxisSensor, "ActivateConsoleSixAxisSensor"}, {301, &Hid::StartConsoleSixAxisSensor, "StartConsoleSixAxisSensor"}, {302, &Hid::StopConsoleSixAxisSensor, "StopConsoleSixAxisSensor"}, @@ -245,7 +280,7 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {404, nullptr, "HasLeftRightBattery"}, {405, nullptr, "GetNpadInterfaceType"}, {406, nullptr, "GetNpadLeftRightInterfaceType"}, - {407, nullptr, "GetNpadOfHighestBatteryLevelForJoyLeft"}, + {407, nullptr, "GetNpadOfHighestBatteryLevel"}, {408, nullptr, "GetNpadOfHighestBatteryLevelForJoyRight"}, {500, nullptr, "GetPalmaConnectionHandle"}, {501, nullptr, "InitializePalma"}, @@ -277,8 +312,8 @@ Hid::Hid(Core::System& system) : ServiceFramework("hid"), system(system) { {527, nullptr, "EnablePalmaBoostMode"}, {528, nullptr, "GetPalmaBluetoothAddress"}, {529, nullptr, "SetDisallowedPalmaConnection"}, - {1000, nullptr, "SetNpadCommunicationMode"}, - {1001, nullptr, "GetNpadCommunicationMode"}, + {1000, &Hid::SetNpadCommunicationMode, "SetNpadCommunicationMode"}, + {1001, &Hid::GetNpadCommunicationMode, "GetNpadCommunicationMode"}, {1002, nullptr, "SetTouchScreenConfiguration"}, {1003, nullptr, "IsFirmwareUpdateNeededForNotification"}, {2000, nullptr, "ActivateDigitizer"}, @@ -305,154 +340,195 @@ void Hid::CreateAppletResource(Kernel::HLERequestContext& ctx) { rb.PushIpcInterface<IAppletResource>(applet_resource); } -void Hid::ActivateXpad(Kernel::HLERequestContext& ctx) { +void Hid::ActivateDebugPad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto basic_xpad_id{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, basic_xpad_id={}, applet_resource_user_id={}", basic_xpad_id, - applet_resource_user_id); + applet_resource->ActivateController(HidController::DebugPad); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->ActivateController(HidController::XPad); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::GetXpadIDs(Kernel::HLERequestContext& ctx) { +void Hid::ActivateTouchScreen(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "(STUBBED) called, applet_resource_user_id={}", applet_resource_user_id); - - IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push(0); -} + applet_resource->ActivateController(HidController::Touchscreen); -void Hid::ActivateSixAxisSensor(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; - applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(true); - LOG_DEBUG(Service_HID, "called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::DeactivateSixAxisSensor(Kernel::HLERequestContext& ctx) { +void Hid::ActivateMouse(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(false); - LOG_DEBUG(Service_HID, "called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + applet_resource->ActivateController(HidController::Mouse); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateDebugPad(Kernel::HLERequestContext& ctx) { +void Hid::ActivateKeyboard(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->ActivateController(HidController::Keyboard); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->ActivateController(HidController::DebugPad); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateTouchScreen(Kernel::HLERequestContext& ctx) { +void Hid::SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + const auto flags{rp.Pop<u32>()}; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + LOG_WARNING(Service_HID, "(STUBBED) called. flags={}", flags); - applet_resource->ActivateController(HidController::Touchscreen); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateMouse(Kernel::HLERequestContext& ctx) { +void Hid::ActivateXpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 basic_xpad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->ActivateController(HidController::XPad); + + LOG_DEBUG(Service_HID, "called, basic_xpad_id={}, applet_resource_user_id={}", + parameters.basic_xpad_id, parameters.applet_resource_user_id); - applet_resource->ActivateController(HidController::Mouse); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateKeyboard(Kernel::HLERequestContext& ctx) { +void Hid::GetXpadIDs(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + LOG_DEBUG(Service_HID, "(STUBBED) called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->ActivateController(HidController::Keyboard); - IPC::ResponseBuilder rb{ctx, 2}; + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); + rb.Push(0); } -void Hid::SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx) { +void Hid::ActivateSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto flags{rp.Pop<u32>()}; - LOG_WARNING(Service_HID, "(STUBBED) called. flags={}", flags); + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(true); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateGesture(Kernel::HLERequestContext& ctx) { +void Hid::DeactivateSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto unknown{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", unknown, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(false); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); - applet_resource->ActivateController(HidController::Gesture); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::ActivateNpadWithRevision(Kernel::HLERequestContext& ctx) { - // Should have no effect with how our npad sets up the data +void Hid::StartSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto unknown{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", unknown, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(true); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); - applet_resource->ActivateController(HidController::NPad); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::StartSixAxisSensor(Kernel::HLERequestContext& ctx) { +void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetSixAxisEnabled(false); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } -void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) { +void Hid::EnableSixAxisSensorFusion(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + bool enable_sixaxis_sensor_fusion{}; + INSERT_PADDING_BYTES(3); + Controller_NPad::DeviceHandle sixaxis_handle{}; + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_WARNING(Service_HID, + "(STUBBED) called, enable_sixaxis_sensor_fusion={}, npad_type={}, npad_id={}, " + "device_index={}, applet_resource_user_id={}", + parameters.enable_sixaxis_sensor_fusion, parameters.sixaxis_handle.npad_type, + parameters.sixaxis_handle.npad_id, parameters.sixaxis_handle.device_index, + parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -460,14 +536,17 @@ void Hid::StopSixAxisSensor(Kernel::HLERequestContext& ctx) { void Hid::SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto drift_mode{rp.Pop<u32>()}; + const auto sixaxis_handle{rp.PopRaw<Controller_NPad::DeviceHandle>()}; + const auto drift_mode{rp.PopEnum<Controller_NPad::GyroscopeZeroDriftMode>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; applet_resource->GetController<Controller_NPad>(HidController::NPad) - .SetGyroscopeZeroDriftMode(Controller_NPad::GyroscopeZeroDriftMode{drift_mode}); + .SetGyroscopeZeroDriftMode(drift_mode); - LOG_DEBUG(Service_HID, "called, handle={}, drift_mode={}, applet_resource_user_id={}", handle, + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, drift_mode={}, " + "applet_resource_user_id={}", + sixaxis_handle.npad_type, sixaxis_handle.npad_id, sixaxis_handle.device_index, drift_mode, applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; @@ -476,29 +555,42 @@ void Hid::SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { void Hid::GetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; - LOG_DEBUG(Service_HID, "called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>( - static_cast<u32>(applet_resource->GetController<Controller_NPad>(HidController::NPad) - .GetGyroscopeZeroDriftMode())); + rb.PushEnum(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetGyroscopeZeroDriftMode()); } void Hid::ResetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; applet_resource->GetController<Controller_NPad>(HidController::NPad) .SetGyroscopeZeroDriftMode(Controller_NPad::GyroscopeZeroDriftMode::Standard); - LOG_DEBUG(Service_HID, "called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -506,11 +598,18 @@ void Hid::ResetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx) { void Hid::IsSixAxisSensorAtRest(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; - LOG_DEBUG(Service_HID, "called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); @@ -518,15 +617,34 @@ void Hid::IsSixAxisSensorAtRest(Kernel::HLERequestContext& ctx) { .IsSixAxisSensorAtRest()); } +void Hid::ActivateGesture(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + u32 unknown{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->ActivateController(HidController::Gesture); + + LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", parameters.unknown, + parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + void Hid::SetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto supported_styleset{rp.Pop<u32>()}; - LOG_DEBUG(Service_HID, "called, supported_styleset={}", supported_styleset); - applet_resource->GetController<Controller_NPad>(HidController::NPad) .SetSupportedStyleSet({supported_styleset}); + LOG_DEBUG(Service_HID, "called, supported_styleset={}", supported_styleset); + IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -537,21 +655,22 @@ void Hid::GetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(controller.GetSupportedStyleSet().raw); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetSupportedStyleSet() + .raw); } void Hid::SetSupportedNpadIdType(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetSupportedNpadIdTypes(ctx.ReadBuffer().data(), ctx.GetReadBufferSize()); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - applet_resource->GetController<Controller_NPad>(HidController::NPad) - .SetSupportedNPadIdTypes(ctx.ReadBuffer().data(), ctx.GetReadBufferSize()); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -560,48 +679,62 @@ void Hid::ActivateNpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->ActivateController(HidController::NPad); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - applet_resource->ActivateController(HidController::NPad); } void Hid::DeactivateNpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->DeactivateController(HidController::NPad); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - applet_resource->DeactivateController(HidController::NPad); } void Hid::AcquireNpadStyleSetUpdateEventHandle(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto unknown{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + u64 unknown{}; + }; - LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}, unknown={}", npad_id, - applet_resource_user_id, unknown); + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}, unknown={}", + parameters.npad_id, parameters.applet_resource_user_id, parameters.unknown); IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); rb.PushCopyObjects(applet_resource->GetController<Controller_NPad>(HidController::NPad) - .GetStyleSetChangedEvent(npad_id)); + .GetStyleSetChangedEvent(parameters.npad_id)); } void Hid::DisconnectNpad(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .DisconnectNpad(parameters.npad_id); - LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", npad_id, - applet_resource_user_id); + LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", parameters.npad_id, + parameters.applet_resource_user_id); - applet_resource->GetController<Controller_NPad>(HidController::NPad).DisconnectNPad(npad_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -614,22 +747,41 @@ void Hid::GetPlayerLedPattern(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.PushRaw<u64>(applet_resource->GetController<Controller_NPad>(HidController::NPad) - .GetLedPattern(npad_id) - .raw); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetLedPattern(npad_id) + .raw); +} + +void Hid::ActivateNpadWithRevision(Kernel::HLERequestContext& ctx) { + // Should have no effect with how our npad sets up the data + IPC::RequestParser rp{ctx}; + struct Parameters { + u32 unknown{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->ActivateController(HidController::NPad); + + LOG_DEBUG(Service_HID, "called, unknown={}, applet_resource_user_id={}", parameters.unknown, + parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); } void Hid::SetNpadJoyHoldType(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto hold_type{rp.Pop<u64>()}; + const auto hold_type{rp.PopEnum<Controller_NPad::NpadHoldType>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad).SetHoldType(hold_type); LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, hold_type={}", applet_resource_user_id, hold_type); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetHoldType(Controller_NPad::NpadHoldType{hold_type}); - IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -640,22 +792,26 @@ void Hid::GetNpadJoyHoldType(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - const auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<u64>(static_cast<u64>(controller.GetHoldType())); + rb.PushEnum(applet_resource->GetController<Controller_NPad>(HidController::NPad).GetHoldType()); } void Hid::SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", npad_id, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadMode(parameters.npad_id, Controller_NPad::NpadAssignments::Single); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Single); + LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", + parameters.npad_id, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -664,16 +820,22 @@ void Hid::SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx void Hid::SetNpadJoyAssignmentModeSingle(Kernel::HLERequestContext& ctx) { // TODO: Check the differences between this and SetNpadJoyAssignmentModeSingleByDefault IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto npad_joy_device_type{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + u64 npad_joy_device_type{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadMode(parameters.npad_id, Controller_NPad::NpadAssignments::Single); LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}, npad_joy_device_type={}", - npad_id, applet_resource_user_id, npad_joy_device_type); - - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Single); + parameters.npad_id, parameters.applet_resource_user_id, + parameters.npad_joy_device_type); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -681,14 +843,19 @@ void Hid::SetNpadJoyAssignmentModeSingle(Kernel::HLERequestContext& ctx) { void Hid::SetNpadJoyAssignmentModeDual(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; - LOG_DEBUG(Service_HID, "called, npad_id={}, applet_resource_user_id={}", npad_id, - applet_resource_user_id); + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadMode(parameters.npad_id, Controller_NPad::NpadAssignments::Dual); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.SetNpadMode(npad_id, Controller_NPad::NPadAssignments::Dual); + LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", + parameters.npad_id, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -700,12 +867,12 @@ void Hid::MergeSingleJoyAsDualJoy(Kernel::HLERequestContext& ctx) { const auto npad_id_2{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .MergeSingleJoyAsDualJoy(npad_id_1, npad_id_2); + LOG_DEBUG(Service_HID, "called, npad_id_1={}, npad_id_2={}, applet_resource_user_id={}", npad_id_1, npad_id_2, applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.MergeSingleJoyAsDualJoy(npad_id_1, npad_id_2); - IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } @@ -714,9 +881,9 @@ void Hid::StartLrAssignmentMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->GetController<Controller_NPad>(HidController::NPad).StartLRAssignmentMode(); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.StartLRAssignmentMode(); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -726,9 +893,9 @@ void Hid::StopLrAssignmentMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; + applet_resource->GetController<Controller_NPad>(HidController::NPad).StopLRAssignmentMode(); + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); - controller.StopLRAssignmentMode(); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -737,13 +904,13 @@ void Hid::StopLrAssignmentMode(Kernel::HLERequestContext& ctx) { void Hid::SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto applet_resource_user_id{rp.Pop<u64>()}; - const auto mode{rp.Pop<u64>()}; - - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, mode={}", applet_resource_user_id, - mode); + const auto activation_mode{rp.PopEnum<Controller_NPad::NpadHandheldActivationMode>()}; applet_resource->GetController<Controller_NPad>(HidController::NPad) - .SetNpadHandheldActivationMode(Controller_NPad::NpadHandheldActivationMode{mode}); + .SetNpadHandheldActivationMode(activation_mode); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, activation_mode={}", + applet_resource_user_id, activation_mode); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -757,23 +924,24 @@ void Hid::GetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<u64>( - static_cast<u64>(applet_resource->GetController<Controller_NPad>(HidController::NPad) - .GetNpadHandheldActivationMode())); + rb.PushEnum(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetNpadHandheldActivationMode()); } void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto npad_1{rp.Pop<u32>()}; - const auto npad_2{rp.Pop<u32>()}; + const auto npad_id_1{rp.Pop<u32>()}; + const auto npad_id_2{rp.Pop<u32>()}; const auto applet_resource_user_id{rp.Pop<u64>()}; - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}, npad_1={}, npad_2={}", - applet_resource_user_id, npad_1, npad_2); + const bool res = applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SwapNpadAssignment(npad_id_1, npad_id_2); + + LOG_DEBUG(Service_HID, "called, npad_id_1={}, npad_id_2={}, applet_resource_user_id={}", + npad_id_1, npad_id_2, applet_resource_user_id); - auto& controller = applet_resource->GetController<Controller_NPad>(HidController::NPad); IPC::ResponseBuilder rb{ctx, 2}; - if (controller.SwapNpadAssignment(npad_1, npad_2)) { + if (res) { rb.Push(RESULT_SUCCESS); } else { LOG_ERROR(Service_HID, "Npads are not connected!"); @@ -781,61 +949,99 @@ void Hid::SwapNpadAssignment(Kernel::HLERequestContext& ctx) { } } -void Hid::BeginPermitVibrationSession(Kernel::HLERequestContext& ctx) { +void Hid::IsUnintendedHomeButtonInputProtectionEnabled(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto applet_resource_user_id{rp.Pop<u64>()}; - - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + struct Parameters { + u32 npad_id{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - applet_resource->GetController<Controller_NPad>(HidController::NPad).SetVibrationEnabled(true); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); -} + const auto parameters{rp.PopRaw<Parameters>()}; -void Hid::EndPermitVibrationSession(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_HID, "called"); + LOG_WARNING(Service_HID, "(STUBBED) called, npad_id={}, applet_resource_user_id={}", + parameters.npad_id, parameters.applet_resource_user_id); - applet_resource->GetController<Controller_NPad>(HidController::NPad).SetVibrationEnabled(false); - IPC::ResponseBuilder rb{ctx, 2}; + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .IsUnintendedHomeButtonInputProtectionEnabled(parameters.npad_id)); } -void Hid::SendVibrationValue(Kernel::HLERequestContext& ctx) { +void Hid::EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto controller_id{rp.Pop<u32>()}; - const auto vibration_values{rp.PopRaw<Controller_NPad::Vibration>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + bool unintended_home_button_input_protection{}; + INSERT_PADDING_BYTES(3); + u32 npad_id{}; + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; - LOG_DEBUG(Service_HID, "called, controller_id={}, applet_resource_user_id={}", controller_id, - applet_resource_user_id); + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetUnintendedHomeButtonInputProtectionEnabled( + parameters.unintended_home_button_input_protection, parameters.npad_id); + + LOG_WARNING(Service_HID, + "(STUBBED) called, unintended_home_button_input_protection={}, npad_id={}," + "applet_resource_user_id={}", + parameters.unintended_home_button_input_protection, parameters.npad_id, + parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); - - applet_resource->GetController<Controller_NPad>(HidController::NPad) - .VibrateController({controller_id}, {vibration_values}); } -void Hid::SendVibrationValues(Kernel::HLERequestContext& ctx) { +void Hid::GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + const auto vibration_device_handle{rp.PopRaw<Controller_NPad::DeviceHandle>()}; + + VibrationDeviceInfo vibration_device_info; + + vibration_device_info.type = VibrationDeviceType::LinearResonantActuator; + + switch (vibration_device_handle.device_index) { + case Controller_NPad::DeviceIndex::Left: + vibration_device_info.position = VibrationDevicePosition::Left; + break; + case Controller_NPad::DeviceIndex::Right: + vibration_device_info.position = VibrationDevicePosition::Right; + break; + case Controller_NPad::DeviceIndex::None: + default: + UNREACHABLE_MSG("DeviceIndex should never be None!"); + vibration_device_info.position = VibrationDevicePosition::None; + break; + } - LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + LOG_DEBUG(Service_HID, "called, vibration_device_type={}, vibration_device_position={}", + vibration_device_info.type, vibration_device_info.position); - const auto controllers = ctx.ReadBuffer(0); - const auto vibrations = ctx.ReadBuffer(1); + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.PushRaw(vibration_device_info); +} - std::vector<u32> controller_list(controllers.size() / sizeof(u32)); - std::vector<Controller_NPad::Vibration> vibration_list(vibrations.size() / - sizeof(Controller_NPad::Vibration)); +void Hid::SendVibrationValue(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle vibration_device_handle{}; + Controller_NPad::VibrationValue vibration_value{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - std::memcpy(controller_list.data(), controllers.data(), controllers.size()); - std::memcpy(vibration_list.data(), vibrations.data(), vibrations.size()); - std::transform(controller_list.begin(), controller_list.end(), controller_list.begin(), - [](u32 controller_id) { return controller_id - 3; }); + const auto parameters{rp.PopRaw<Parameters>()}; applet_resource->GetController<Controller_NPad>(HidController::NPad) - .VibrateController(controller_list, vibration_list); + .VibrateController(parameters.vibration_device_handle, parameters.vibration_value); + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.vibration_device_handle.npad_type, + parameters.vibration_device_handle.npad_id, + parameters.vibration_device_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -843,25 +1049,24 @@ void Hid::SendVibrationValues(Kernel::HLERequestContext& ctx) { void Hid::GetActualVibrationValue(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto controller_id{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; - - LOG_DEBUG(Service_HID, "called, controller_id={}, applet_resource_user_id={}", controller_id, - applet_resource_user_id); + struct Parameters { + Controller_NPad::DeviceHandle vibration_device_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - IPC::ResponseBuilder rb{ctx, 6}; - rb.Push(RESULT_SUCCESS); - rb.PushRaw<Controller_NPad::Vibration>( - applet_resource->GetController<Controller_NPad>(HidController::NPad).GetLastVibration()); -} + const auto parameters{rp.PopRaw<Parameters>()}; -void Hid::GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_HID, "called"); + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.vibration_device_handle.npad_type, + parameters.vibration_device_handle.npad_id, + parameters.vibration_device_handle.device_index, parameters.applet_resource_user_id); - IPC::ResponseBuilder rb{ctx, 4}; + IPC::ResponseBuilder rb{ctx, 6}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(1); - rb.Push<u32>(0); + rb.PushRaw(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetLastVibration(parameters.vibration_device_handle)); } void Hid::CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx) { @@ -869,13 +1074,14 @@ void Hid::CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IActiveVibrationDeviceList>(); + rb.PushIpcInterface<IActiveVibrationDeviceList>(system, applet_resource); } void Hid::PermitVibration(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto can_vibrate{rp.Pop<bool>()}; - Settings::values.vibration_enabled = can_vibrate; + + Settings::values.vibration_enabled.SetValue(can_vibrate); LOG_DEBUG(Service_HID, "called, can_vibrate={}", can_vibrate); @@ -888,7 +1094,76 @@ void Hid::IsVibrationPermitted(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push(Settings::values.vibration_enabled); + rb.Push(Settings::values.vibration_enabled.GetValue()); +} + +void Hid::SendVibrationValues(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + const auto handles = ctx.ReadBuffer(0); + const auto vibrations = ctx.ReadBuffer(1); + + std::vector<Controller_NPad::DeviceHandle> vibration_device_handles( + handles.size() / sizeof(Controller_NPad::DeviceHandle)); + std::vector<Controller_NPad::VibrationValue> vibration_values( + vibrations.size() / sizeof(Controller_NPad::VibrationValue)); + + std::memcpy(vibration_device_handles.data(), handles.data(), handles.size()); + std::memcpy(vibration_values.data(), vibrations.data(), vibrations.size()); + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .VibrateControllers(vibration_device_handles, vibration_values); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::BeginPermitVibrationSession(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetPermitVibrationSession(true); + + LOG_DEBUG(Service_HID, "called, applet_resource_user_id={}", applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::EndPermitVibrationSession(Kernel::HLERequestContext& ctx) { + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetPermitVibrationSession(false); + + LOG_DEBUG(Service_HID, "called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::IsVibrationDeviceMounted(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + struct Parameters { + Controller_NPad::DeviceHandle vibration_device_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; + + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_DEBUG(Service_HID, + "called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.vibration_device_handle.npad_type, + parameters.vibration_device_handle.npad_id, + parameters.vibration_device_handle.device_index, parameters.applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .IsVibrationDeviceMounted(parameters.vibration_device_handle)); } void Hid::ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { @@ -904,11 +1179,19 @@ void Hid::ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { void Hid::StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_WARNING( + Service_HID, + "(STUBBED) called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -916,11 +1199,19 @@ void Hid::StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { void Hid::StopConsoleSixAxisSensor(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const auto handle{rp.Pop<u32>()}; - const auto applet_resource_user_id{rp.Pop<u64>()}; + struct Parameters { + Controller_NPad::DeviceHandle sixaxis_handle{}; + INSERT_PADDING_WORDS(1); + u64 applet_resource_user_id{}; + }; - LOG_WARNING(Service_HID, "(STUBBED) called, handle={}, applet_resource_user_id={}", handle, - applet_resource_user_id); + const auto parameters{rp.PopRaw<Parameters>()}; + + LOG_WARNING( + Service_HID, + "(STUBBED) called, npad_type={}, npad_id={}, device_index={}, applet_resource_user_id={}", + parameters.sixaxis_handle.npad_type, parameters.sixaxis_handle.npad_id, + parameters.sixaxis_handle.device_index, parameters.applet_resource_user_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -1011,9 +1302,37 @@ void Hid::SetPalmaBoostMode(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); } +void Hid::SetNpadCommunicationMode(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + const auto communication_mode{rp.PopEnum<Controller_NPad::NpadCommunicationMode>()}; + + applet_resource->GetController<Controller_NPad>(HidController::NPad) + .SetNpadCommunicationMode(communication_mode); + + LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}, communication_mode={}", + applet_resource_user_id, communication_mode); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + +void Hid::GetNpadCommunicationMode(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto applet_resource_user_id{rp.Pop<u64>()}; + + LOG_WARNING(Service_HID, "(STUBBED) called, applet_resource_user_id={}", + applet_resource_user_id); + + IPC::ResponseBuilder rb{ctx, 4}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(applet_resource->GetController<Controller_NPad>(HidController::NPad) + .GetNpadCommunicationMode()); +} + class HidDbg final : public ServiceFramework<HidDbg> { public: - explicit HidDbg() : ServiceFramework{"hid:dbg"} { + explicit HidDbg(Core::System& system_) : ServiceFramework{system_, "hid:dbg"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "DeactivateDebugPad"}, @@ -1140,7 +1459,7 @@ public: class HidSys final : public ServiceFramework<HidSys> { public: - explicit HidSys() : ServiceFramework{"hid:sys"} { + explicit HidSys(Core::System& system_) : ServiceFramework{system_, "hid:sys"} { // clang-format off static const FunctionInfo functions[] = { {31, nullptr, "SendKeyboardLockKeyEvent"}, @@ -1274,7 +1593,7 @@ public: class HidTmp final : public ServiceFramework<HidTmp> { public: - explicit HidTmp() : ServiceFramework{"hid:tmp"} { + explicit HidTmp(Core::System& system_) : ServiceFramework{system_, "hid:tmp"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetConsoleSixAxisSensorCalibrationValues"}, @@ -1287,7 +1606,7 @@ public: class HidBus final : public ServiceFramework<HidBus> { public: - explicit HidBus() : ServiceFramework{"hidbus"} { + explicit HidBus(Core::System& system_) : ServiceFramework{system_, "hidbus"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "GetBusHandle"}, @@ -1317,15 +1636,15 @@ void ReloadInputDevices() { void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { std::make_shared<Hid>(system)->InstallAsService(service_manager); - std::make_shared<HidBus>()->InstallAsService(service_manager); - std::make_shared<HidDbg>()->InstallAsService(service_manager); - std::make_shared<HidSys>()->InstallAsService(service_manager); - std::make_shared<HidTmp>()->InstallAsService(service_manager); + std::make_shared<HidBus>(system)->InstallAsService(service_manager); + std::make_shared<HidDbg>(system)->InstallAsService(service_manager); + std::make_shared<HidSys>(system)->InstallAsService(service_manager); + std::make_shared<HidTmp>(system)->InstallAsService(service_manager); std::make_shared<IRS>(system)->InstallAsService(service_manager); - std::make_shared<IRS_SYS>()->InstallAsService(service_manager); + std::make_shared<IRS_SYS>(system)->InstallAsService(service_manager); - std::make_shared<XCD_SYS>()->InstallAsService(service_manager); + std::make_shared<XCD_SYS>(system)->InstallAsService(service_manager); } } // namespace Service::HID diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h index e04aaf1e9..b87bfdde1 100644 --- a/src/core/hle/service/hid/hid.h +++ b/src/core/hle/service/hid/hid.h @@ -41,7 +41,7 @@ enum class HidController : std::size_t { class IAppletResource final : public ServiceFramework<IAppletResource> { public: - explicit IAppletResource(Core::System& system); + explicit IAppletResource(Core::System& system_); ~IAppletResource() override; void ActivateController(HidController controller); @@ -65,11 +65,12 @@ private: void GetSharedMemoryHandle(Kernel::HLERequestContext& ctx); void UpdateControllers(std::uintptr_t user_data, std::chrono::nanoseconds ns_late); + void UpdateMotion(std::uintptr_t user_data, std::chrono::nanoseconds ns_late); std::shared_ptr<Kernel::SharedMemory> shared_mem; std::shared_ptr<Core::Timing::EventType> pad_update_event; - Core::System& system; + std::shared_ptr<Core::Timing::EventType> motion_update_event; std::array<std::unique_ptr<ControllerBase>, static_cast<size_t>(HidController::MaxControllers)> controllers{}; @@ -77,30 +78,30 @@ private: class Hid final : public ServiceFramework<Hid> { public: - explicit Hid(Core::System& system); + explicit Hid(Core::System& system_); ~Hid() override; std::shared_ptr<IAppletResource> GetAppletResource(); private: void CreateAppletResource(Kernel::HLERequestContext& ctx); - void ActivateXpad(Kernel::HLERequestContext& ctx); - void GetXpadIDs(Kernel::HLERequestContext& ctx); - void ActivateSixAxisSensor(Kernel::HLERequestContext& ctx); - void DeactivateSixAxisSensor(Kernel::HLERequestContext& ctx); void ActivateDebugPad(Kernel::HLERequestContext& ctx); void ActivateTouchScreen(Kernel::HLERequestContext& ctx); void ActivateMouse(Kernel::HLERequestContext& ctx); void ActivateKeyboard(Kernel::HLERequestContext& ctx); void SendKeyboardLockKeyEvent(Kernel::HLERequestContext& ctx); - void ActivateGesture(Kernel::HLERequestContext& ctx); - void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx); + void ActivateXpad(Kernel::HLERequestContext& ctx); + void GetXpadIDs(Kernel::HLERequestContext& ctx); + void ActivateSixAxisSensor(Kernel::HLERequestContext& ctx); + void DeactivateSixAxisSensor(Kernel::HLERequestContext& ctx); void StartSixAxisSensor(Kernel::HLERequestContext& ctx); void StopSixAxisSensor(Kernel::HLERequestContext& ctx); + void EnableSixAxisSensorFusion(Kernel::HLERequestContext& ctx); void SetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx); void GetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx); void ResetGyroscopeZeroDriftMode(Kernel::HLERequestContext& ctx); void IsSixAxisSensorAtRest(Kernel::HLERequestContext& ctx); + void ActivateGesture(Kernel::HLERequestContext& ctx); void SetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx); void GetSupportedNpadStyleSet(Kernel::HLERequestContext& ctx); void SetSupportedNpadIdType(Kernel::HLERequestContext& ctx); @@ -109,6 +110,7 @@ private: void AcquireNpadStyleSetUpdateEventHandle(Kernel::HLERequestContext& ctx); void DisconnectNpad(Kernel::HLERequestContext& ctx); void GetPlayerLedPattern(Kernel::HLERequestContext& ctx); + void ActivateNpadWithRevision(Kernel::HLERequestContext& ctx); void SetNpadJoyHoldType(Kernel::HLERequestContext& ctx); void GetNpadJoyHoldType(Kernel::HLERequestContext& ctx); void SetNpadJoyAssignmentModeSingleByDefault(Kernel::HLERequestContext& ctx); @@ -120,15 +122,18 @@ private: void SetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx); void GetNpadHandheldActivationMode(Kernel::HLERequestContext& ctx); void SwapNpadAssignment(Kernel::HLERequestContext& ctx); - void BeginPermitVibrationSession(Kernel::HLERequestContext& ctx); - void EndPermitVibrationSession(Kernel::HLERequestContext& ctx); + void IsUnintendedHomeButtonInputProtectionEnabled(Kernel::HLERequestContext& ctx); + void EnableUnintendedHomeButtonInputProtection(Kernel::HLERequestContext& ctx); + void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx); void SendVibrationValue(Kernel::HLERequestContext& ctx); - void SendVibrationValues(Kernel::HLERequestContext& ctx); void GetActualVibrationValue(Kernel::HLERequestContext& ctx); - void GetVibrationDeviceInfo(Kernel::HLERequestContext& ctx); void CreateActiveVibrationDeviceList(Kernel::HLERequestContext& ctx); void PermitVibration(Kernel::HLERequestContext& ctx); void IsVibrationPermitted(Kernel::HLERequestContext& ctx); + void SendVibrationValues(Kernel::HLERequestContext& ctx); + void BeginPermitVibrationSession(Kernel::HLERequestContext& ctx); + void EndPermitVibrationSession(Kernel::HLERequestContext& ctx); + void IsVibrationDeviceMounted(Kernel::HLERequestContext& ctx); void ActivateConsoleSixAxisSensor(Kernel::HLERequestContext& ctx); void StartConsoleSixAxisSensor(Kernel::HLERequestContext& ctx); void StopConsoleSixAxisSensor(Kernel::HLERequestContext& ctx); @@ -140,9 +145,26 @@ private: void ResetSevenSixAxisSensorTimestamp(Kernel::HLERequestContext& ctx); void SetIsPalmaAllConnectable(Kernel::HLERequestContext& ctx); void SetPalmaBoostMode(Kernel::HLERequestContext& ctx); + void SetNpadCommunicationMode(Kernel::HLERequestContext& ctx); + void GetNpadCommunicationMode(Kernel::HLERequestContext& ctx); + + enum class VibrationDeviceType : u32 { + LinearResonantActuator = 1, + }; + + enum class VibrationDevicePosition : u32 { + None = 0, + Left = 1, + Right = 2, + }; + + struct VibrationDeviceInfo { + VibrationDeviceType type{}; + VibrationDevicePosition position{}; + }; + static_assert(sizeof(VibrationDeviceInfo) == 0x8, "VibrationDeviceInfo has incorrect size."); std::shared_ptr<IAppletResource> applet_resource; - Core::System& system; }; /// Reload input devices. Used when input configuration changed diff --git a/src/core/hle/service/hid/irs.cpp b/src/core/hle/service/hid/irs.cpp index e82fd031b..c8413099f 100644 --- a/src/core/hle/service/hid/irs.cpp +++ b/src/core/hle/service/hid/irs.cpp @@ -12,7 +12,7 @@ namespace Service::HID { -IRS::IRS(Core::System& system) : ServiceFramework{"irs"}, system(system) { +IRS::IRS(Core::System& system_) : ServiceFramework{system_, "irs"} { // clang-format off static const FunctionInfo functions[] = { {302, &IRS::ActivateIrsensor, "ActivateIrsensor"}, @@ -175,7 +175,7 @@ void IRS::ActivateIrsensorWithFunctionLevel(Kernel::HLERequestContext& ctx) { IRS::~IRS() = default; -IRS_SYS::IRS_SYS() : ServiceFramework{"irs:sys"} { +IRS_SYS::IRS_SYS(Core::System& system_) : ServiceFramework{system_, "irs:sys"} { // clang-format off static const FunctionInfo functions[] = { {500, nullptr, "SetAppletResourceUserId"}, diff --git a/src/core/hle/service/hid/irs.h b/src/core/hle/service/hid/irs.h index 8918ad6ca..be0c486ba 100644 --- a/src/core/hle/service/hid/irs.h +++ b/src/core/hle/service/hid/irs.h @@ -7,6 +7,10 @@ #include "core/hle/kernel/object.h" #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class SharedMemory; } @@ -15,7 +19,7 @@ namespace Service::HID { class IRS final : public ServiceFramework<IRS> { public: - explicit IRS(Core::System& system); + explicit IRS(Core::System& system_); ~IRS() override; private: @@ -37,14 +41,14 @@ private: void RunIrLedProcessor(Kernel::HLERequestContext& ctx); void StopImageProcessorAsync(Kernel::HLERequestContext& ctx); void ActivateIrsensorWithFunctionLevel(Kernel::HLERequestContext& ctx); + std::shared_ptr<Kernel::SharedMemory> shared_mem; const u32 device_handle{0xABCD}; - Core::System& system; }; class IRS_SYS final : public ServiceFramework<IRS_SYS> { public: - explicit IRS_SYS(); + explicit IRS_SYS(Core::System& system); ~IRS_SYS() override; }; diff --git a/src/core/hle/service/hid/xcd.cpp b/src/core/hle/service/hid/xcd.cpp index c8e9125f6..43a8840d0 100644 --- a/src/core/hle/service/hid/xcd.cpp +++ b/src/core/hle/service/hid/xcd.cpp @@ -6,7 +6,7 @@ namespace Service::HID { -XCD_SYS::XCD_SYS() : ServiceFramework{"xcd:sys"} { +XCD_SYS::XCD_SYS(Core::System& system_) : ServiceFramework{system_, "xcd:sys"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetDataFormat"}, diff --git a/src/core/hle/service/hid/xcd.h b/src/core/hle/service/hid/xcd.h index fd506d303..54932c228 100644 --- a/src/core/hle/service/hid/xcd.h +++ b/src/core/hle/service/hid/xcd.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::HID { class XCD_SYS final : public ServiceFramework<XCD_SYS> { public: - explicit XCD_SYS(); + explicit XCD_SYS(Core::System& system_); ~XCD_SYS() override; }; diff --git a/src/core/hle/service/lbl/lbl.cpp b/src/core/hle/service/lbl/lbl.cpp index 17350b403..6ad3a2877 100644 --- a/src/core/hle/service/lbl/lbl.cpp +++ b/src/core/hle/service/lbl/lbl.cpp @@ -15,7 +15,7 @@ namespace Service::LBL { class LBL final : public ServiceFramework<LBL> { public: - explicit LBL() : ServiceFramework{"lbl"} { + explicit LBL(Core::System& system_) : ServiceFramework{system_, "lbl"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SaveCurrentSetting"}, @@ -84,8 +84,8 @@ private: bool vr_mode_enabled = false; }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<LBL>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<LBL>(system)->InstallAsService(sm); } } // namespace Service::LBL diff --git a/src/core/hle/service/lbl/lbl.h b/src/core/hle/service/lbl/lbl.h index bf6f400f8..9c2021026 100644 --- a/src/core/hle/service/lbl/lbl.h +++ b/src/core/hle/service/lbl/lbl.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::LBL { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::LBL diff --git a/src/core/hle/service/ldn/ldn.cpp b/src/core/hle/service/ldn/ldn.cpp index 49972cd69..ee908f399 100644 --- a/src/core/hle/service/ldn/ldn.cpp +++ b/src/core/hle/service/ldn/ldn.cpp @@ -13,7 +13,7 @@ namespace Service::LDN { class IMonitorService final : public ServiceFramework<IMonitorService> { public: - explicit IMonitorService() : ServiceFramework{"IMonitorService"} { + explicit IMonitorService(Core::System& system_) : ServiceFramework{system_, "IMonitorService"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetStateForMonitor"}, @@ -33,7 +33,7 @@ public: class LDNM final : public ServiceFramework<LDNM> { public: - explicit LDNM() : ServiceFramework{"ldn:m"} { + explicit LDNM(Core::System& system_) : ServiceFramework{system_, "ldn:m"} { // clang-format off static const FunctionInfo functions[] = { {0, &LDNM::CreateMonitorService, "CreateMonitorService"} @@ -48,15 +48,15 @@ public: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IMonitorService>(); + rb.PushIpcInterface<IMonitorService>(system); } }; class ISystemLocalCommunicationService final : public ServiceFramework<ISystemLocalCommunicationService> { public: - explicit ISystemLocalCommunicationService() - : ServiceFramework{"ISystemLocalCommunicationService"} { + explicit ISystemLocalCommunicationService(Core::System& system_) + : ServiceFramework{system_, "ISystemLocalCommunicationService"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetState"}, @@ -99,7 +99,8 @@ public: class IUserLocalCommunicationService final : public ServiceFramework<IUserLocalCommunicationService> { public: - explicit IUserLocalCommunicationService() : ServiceFramework{"IUserLocalCommunicationService"} { + explicit IUserLocalCommunicationService(Core::System& system_) + : ServiceFramework{system_, "IUserLocalCommunicationService"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetState"}, @@ -148,7 +149,7 @@ public: class LDNS final : public ServiceFramework<LDNS> { public: - explicit LDNS() : ServiceFramework{"ldn:s"} { + explicit LDNS(Core::System& system_) : ServiceFramework{system_, "ldn:s"} { // clang-format off static const FunctionInfo functions[] = { {0, &LDNS::CreateSystemLocalCommunicationService, "CreateSystemLocalCommunicationService"}, @@ -163,13 +164,13 @@ public: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemLocalCommunicationService>(); + rb.PushIpcInterface<ISystemLocalCommunicationService>(system); } }; class LDNU final : public ServiceFramework<LDNU> { public: - explicit LDNU() : ServiceFramework{"ldn:u"} { + explicit LDNU(Core::System& system_) : ServiceFramework{system_, "ldn:u"} { // clang-format off static const FunctionInfo functions[] = { {0, &LDNU::CreateUserLocalCommunicationService, "CreateUserLocalCommunicationService"}, @@ -184,14 +185,14 @@ public: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IUserLocalCommunicationService>(); + rb.PushIpcInterface<IUserLocalCommunicationService>(system); } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<LDNM>()->InstallAsService(sm); - std::make_shared<LDNS>()->InstallAsService(sm); - std::make_shared<LDNU>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<LDNM>(system)->InstallAsService(sm); + std::make_shared<LDNS>(system)->InstallAsService(sm); + std::make_shared<LDNU>(system)->InstallAsService(sm); } } // namespace Service::LDN diff --git a/src/core/hle/service/ldn/ldn.h b/src/core/hle/service/ldn/ldn.h index 6b2a3c2b2..3ccd9738b 100644 --- a/src/core/hle/service/ldn/ldn.h +++ b/src/core/hle/service/ldn/ldn.h @@ -4,6 +4,10 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } @@ -11,6 +15,6 @@ class ServiceManager; namespace Service::LDN { /// Registers all LDN services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::LDN diff --git a/src/core/hle/service/ldr/ldr.cpp b/src/core/hle/service/ldr/ldr.cpp index d8cd10e31..9da786b4e 100644 --- a/src/core/hle/service/ldr/ldr.cpp +++ b/src/core/hle/service/ldr/ldr.cpp @@ -9,6 +9,7 @@ #include "common/alignment.h" #include "common/hex_util.h" #include "common/scope_exit.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/errors.h" #include "core/hle/kernel/memory/page_table.h" @@ -23,7 +24,7 @@ namespace Service::LDR { constexpr ResultCode ERROR_INSUFFICIENT_ADDRESS_SPACE{ErrorModule::RO, 2}; -constexpr ResultCode ERROR_INVALID_MEMORY_STATE{ErrorModule::Loader, 51}; +[[maybe_unused]] constexpr ResultCode ERROR_INVALID_MEMORY_STATE{ErrorModule::Loader, 51}; constexpr ResultCode ERROR_INVALID_NRO{ErrorModule::Loader, 52}; constexpr ResultCode ERROR_INVALID_NRR{ErrorModule::Loader, 53}; constexpr ResultCode ERROR_MISSING_NRR_HASH{ErrorModule::Loader, 54}; @@ -33,7 +34,7 @@ constexpr ResultCode ERROR_ALREADY_LOADED{ErrorModule::Loader, 57}; constexpr ResultCode ERROR_INVALID_ALIGNMENT{ErrorModule::Loader, 81}; constexpr ResultCode ERROR_INVALID_SIZE{ErrorModule::Loader, 82}; constexpr ResultCode ERROR_INVALID_NRO_ADDRESS{ErrorModule::Loader, 84}; -constexpr ResultCode ERROR_INVALID_NRR_ADDRESS{ErrorModule::Loader, 85}; +[[maybe_unused]] constexpr ResultCode ERROR_INVALID_NRR_ADDRESS{ErrorModule::Loader, 85}; constexpr ResultCode ERROR_NOT_INITIALIZED{ErrorModule::Loader, 87}; constexpr std::size_t MAXIMUM_LOADED_RO{0x40}; @@ -114,7 +115,7 @@ static_assert(sizeof(NROInfo) == 0x60, "NROInfo has invalid size."); class DebugMonitor final : public ServiceFramework<DebugMonitor> { public: - explicit DebugMonitor() : ServiceFramework{"ldr:dmnt"} { + explicit DebugMonitor(Core::System& system_) : ServiceFramework{system_, "ldr:dmnt"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "AddProcessToDebugLaunchQueue"}, @@ -129,7 +130,7 @@ public: class ProcessManager final : public ServiceFramework<ProcessManager> { public: - explicit ProcessManager() : ServiceFramework{"ldr:pm"} { + explicit ProcessManager(Core::System& system_) : ServiceFramework{system_, "ldr:pm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CreateProcess"}, @@ -146,7 +147,7 @@ public: class Shell final : public ServiceFramework<Shell> { public: - explicit Shell() : ServiceFramework{"ldr:shel"} { + explicit Shell(Core::System& system_) : ServiceFramework{system_, "ldr:shel"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "AddProcessToLaunchQueue"}, @@ -160,13 +161,13 @@ public: class RelocatableObject final : public ServiceFramework<RelocatableObject> { public: - explicit RelocatableObject(Core::System& system) : ServiceFramework{"ldr:ro"}, system(system) { + explicit RelocatableObject(Core::System& system_) : ServiceFramework{system_, "ldr:ro"} { // clang-format off static const FunctionInfo functions[] = { {0, &RelocatableObject::LoadNro, "LoadNro"}, {1, &RelocatableObject::UnloadNro, "UnloadNro"}, {2, &RelocatableObject::LoadNrr, "LoadNrr"}, - {3, nullptr, "UnloadNrr"}, + {3, &RelocatableObject::UnloadNrr, "UnloadNrr"}, {4, &RelocatableObject::Initialize, "Initialize"}, {10, nullptr, "LoadNrrEx"}, }; @@ -272,6 +273,20 @@ public: rb.Push(RESULT_SUCCESS); } + void UnloadNrr(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + const auto pid = rp.Pop<u64>(); + const auto nrr_address = rp.Pop<VAddr>(); + + LOG_DEBUG(Service_LDR, "called with pid={}, nrr_address={:016X}", pid, nrr_address); + + nrr.erase(nrr_address); + + IPC::ResponseBuilder rb{ctx, 2}; + + rb.Push(RESULT_SUCCESS); + } + bool ValidateRegionForMap(Kernel::Memory::PageTable& page_table, VAddr start, std::size_t size) const { constexpr std::size_t padding_size{4 * Kernel::Memory::PageSize}; @@ -512,9 +527,6 @@ public: header.segment_headers[RO_INDEX].memory_size, header.segment_headers[DATA_INDEX].memory_size, nro_address}); - // Invalidate JIT caches for the newly mapped process code - system.InvalidateCpuInstructionCaches(); - IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); rb.Push(*map_result); @@ -575,8 +587,6 @@ public: const auto result{UnmapNro(iter->second)}; - system.InvalidateCpuInstructionCaches(); - nro.erase(iter); IPC::ResponseBuilder rb{ctx, 2}; @@ -624,13 +634,12 @@ private: Common::Is4KBAligned(header.segment_headers[RO_INDEX].memory_size) && Common::Is4KBAligned(header.segment_headers[DATA_INDEX].memory_size); } - Core::System& system; }; void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { - std::make_shared<DebugMonitor>()->InstallAsService(sm); - std::make_shared<ProcessManager>()->InstallAsService(sm); - std::make_shared<Shell>()->InstallAsService(sm); + std::make_shared<DebugMonitor>(system)->InstallAsService(sm); + std::make_shared<ProcessManager>(system)->InstallAsService(sm); + std::make_shared<Shell>(system)->InstallAsService(sm); std::make_shared<RelocatableObject>(system)->InstallAsService(sm); } diff --git a/src/core/hle/service/ldr/ldr.h b/src/core/hle/service/ldr/ldr.h index 7ac8c0b65..104fc15c5 100644 --- a/src/core/hle/service/ldr/ldr.h +++ b/src/core/hle/service/ldr/ldr.h @@ -4,6 +4,10 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } diff --git a/src/core/hle/service/lm/lm.cpp b/src/core/hle/service/lm/lm.cpp index dec96b771..8e49b068c 100644 --- a/src/core/hle/service/lm/lm.cpp +++ b/src/core/hle/service/lm/lm.cpp @@ -7,6 +7,7 @@ #include "common/logging/log.h" #include "common/scope_exit.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/service/lm/lm.h" #include "core/hle/service/lm/manager.h" @@ -17,8 +18,9 @@ namespace Service::LM { class ILogger final : public ServiceFramework<ILogger> { public: - explicit ILogger(Manager& manager_, Core::Memory::Memory& memory_) - : ServiceFramework("ILogger"), manager{manager_}, memory{memory_} { + explicit ILogger(Core::System& system_) + : ServiceFramework{system_, "ILogger"}, manager{system_.GetLogManager()}, + memory{system_.Memory()} { static const FunctionInfo functions[] = { {0, &ILogger::Log, "Log"}, {1, &ILogger::SetDestination, "SetDestination"}, @@ -66,7 +68,7 @@ private: IPC::RequestParser rp{ctx}; const auto destination = rp.PopEnum<DestinationFlag>(); - LOG_DEBUG(Service_LM, "called, destination={:08X}", static_cast<u32>(destination)); + LOG_DEBUG(Service_LM, "called, destination={:08X}", destination); manager.SetDestination(destination); @@ -80,8 +82,7 @@ private: class LM final : public ServiceFramework<LM> { public: - explicit LM(Manager& manager_, Core::Memory::Memory& memory_) - : ServiceFramework{"lm"}, manager{manager_}, memory{memory_} { + explicit LM(Core::System& system_) : ServiceFramework{system_, "lm"} { // clang-format off static const FunctionInfo functions[] = { {0, &LM::OpenLogger, "OpenLogger"}, @@ -97,16 +98,12 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ILogger>(manager, memory); + rb.PushIpcInterface<ILogger>(system); } - - Manager& manager; - Core::Memory::Memory& memory; }; void InstallInterfaces(Core::System& system) { - std::make_shared<LM>(system.GetLogManager(), system.Memory()) - ->InstallAsService(system.ServiceManager()); + std::make_shared<LM>(system)->InstallAsService(system.ServiceManager()); } } // namespace Service::LM diff --git a/src/core/hle/service/mig/mig.cpp b/src/core/hle/service/mig/mig.cpp index 113a4665c..1599d941b 100644 --- a/src/core/hle/service/mig/mig.cpp +++ b/src/core/hle/service/mig/mig.cpp @@ -12,7 +12,7 @@ namespace Service::Migration { class MIG_USR final : public ServiceFramework<MIG_USR> { public: - explicit MIG_USR() : ServiceFramework{"mig:usr"} { + explicit MIG_USR(Core::System& system_) : ServiceFramework{system_, "mig:usr"} { // clang-format off static const FunctionInfo functions[] = { {10, nullptr, "TryGetLastMigrationInfo"}, @@ -33,8 +33,8 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<MIG_USR>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<MIG_USR>(system)->InstallAsService(sm); } } // namespace Service::Migration diff --git a/src/core/hle/service/mig/mig.h b/src/core/hle/service/mig/mig.h index 288c1c1b3..2b24cdf2c 100644 --- a/src/core/hle/service/mig/mig.h +++ b/src/core/hle/service/mig/mig.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::Migration { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::Migration diff --git a/src/core/hle/service/mii/manager.cpp b/src/core/hle/service/mii/manager.cpp index 4730070cb..d73b90015 100644 --- a/src/core/hle/service/mii/manager.cpp +++ b/src/core/hle/service/mii/manager.cpp @@ -131,7 +131,7 @@ template <typename T> T GetRandomValue(T min, T max) { std::random_device device; std::mt19937 gen(device()); - std::uniform_int_distribution<u64> distribution(0, static_cast<u64>(max)); + std::uniform_int_distribution<u64> distribution(static_cast<u64>(min), static_cast<u64>(max)); return static_cast<T>(distribution(gen)); } @@ -428,7 +428,7 @@ bool MiiManager::IsFullDatabase() const { } u32 MiiManager::GetCount(SourceFlag source_flag) const { - u32 count{}; + std::size_t count{}; if ((source_flag & SourceFlag::Database) != SourceFlag::None) { // TODO(bunnei): We don't implement the Mii database, but when we do, update this count += 0; @@ -436,7 +436,7 @@ u32 MiiManager::GetCount(SourceFlag source_flag) const { if ((source_flag & SourceFlag::Default) != SourceFlag::None) { count += DefaultMiiCount; } - return count; + return static_cast<u32>(count); } ResultVal<MiiInfo> MiiManager::UpdateLatest([[maybe_unused]] const MiiInfo& info, diff --git a/src/core/hle/service/mii/mii.cpp b/src/core/hle/service/mii/mii.cpp index b81bf6277..26be9e45b 100644 --- a/src/core/hle/service/mii/mii.cpp +++ b/src/core/hle/service/mii/mii.cpp @@ -18,7 +18,8 @@ constexpr ResultCode ERROR_INVALID_ARGUMENT{ErrorModule::Mii, 1}; class IDatabaseService final : public ServiceFramework<IDatabaseService> { public: - explicit IDatabaseService() : ServiceFramework{"IDatabaseService"} { + explicit IDatabaseService(Core::System& system_) + : ServiceFramework{system_, "IDatabaseService"} { // clang-format off static const FunctionInfo functions[] = { {0, &IDatabaseService::IsUpdated, "IsUpdated"}, @@ -47,6 +48,7 @@ public: {23, nullptr, "Convert"}, {24, nullptr, "ConvertCoreDataToCharInfo"}, {25, nullptr, "ConvertCharInfoToCoreData"}, + {26, nullptr, "Append"}, }; // clang-format on @@ -251,7 +253,8 @@ private: class MiiDBModule final : public ServiceFramework<MiiDBModule> { public: - explicit MiiDBModule(const char* name) : ServiceFramework{name} { + explicit MiiDBModule(Core::System& system_, const char* name) + : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { {0, &MiiDBModule::GetDatabaseService, "GetDatabaseService"}, @@ -265,7 +268,7 @@ private: void GetDatabaseService(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IDatabaseService>(); + rb.PushIpcInterface<IDatabaseService>(system); LOG_DEBUG(Service_Mii, "called"); } @@ -273,7 +276,7 @@ private: class MiiImg final : public ServiceFramework<MiiImg> { public: - explicit MiiImg() : ServiceFramework{"miiimg"} { + explicit MiiImg(Core::System& system_) : ServiceFramework{system_, "miiimg"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -297,11 +300,11 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<MiiDBModule>("mii:e")->InstallAsService(sm); - std::make_shared<MiiDBModule>("mii:u")->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<MiiDBModule>(system, "mii:e")->InstallAsService(sm); + std::make_shared<MiiDBModule>(system, "mii:u")->InstallAsService(sm); - std::make_shared<MiiImg>()->InstallAsService(sm); + std::make_shared<MiiImg>(system)->InstallAsService(sm); } } // namespace Service::Mii diff --git a/src/core/hle/service/mii/mii.h b/src/core/hle/service/mii/mii.h index 7ce9be50e..9d3238e72 100644 --- a/src/core/hle/service/mii/mii.h +++ b/src/core/hle/service/mii/mii.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::Mii { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::Mii diff --git a/src/core/hle/service/mm/mm_u.cpp b/src/core/hle/service/mm/mm_u.cpp index 25c24e537..b0cb07d24 100644 --- a/src/core/hle/service/mm/mm_u.cpp +++ b/src/core/hle/service/mm/mm_u.cpp @@ -6,12 +6,13 @@ #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/client_session.h" #include "core/hle/service/mm/mm_u.h" +#include "core/hle/service/sm/sm.h" namespace Service::MM { class MM_U final : public ServiceFramework<MM_U> { public: - explicit MM_U() : ServiceFramework{"mm:u"} { + explicit MM_U(Core::System& system_) : ServiceFramework{system_, "mm:u"} { // clang-format off static const FunctionInfo functions[] = { {0, &MM_U::InitializeOld, "InitializeOld"}, @@ -104,8 +105,8 @@ private: u32 id{1}; }; -void InstallInterfaces(SM::ServiceManager& service_manager) { - std::make_shared<MM_U>()->InstallAsService(service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { + std::make_shared<MM_U>(system)->InstallAsService(service_manager); } } // namespace Service::MM diff --git a/src/core/hle/service/mm/mm_u.h b/src/core/hle/service/mm/mm_u.h index 5439fa653..49b6a3355 100644 --- a/src/core/hle/service/mm/mm_u.h +++ b/src/core/hle/service/mm/mm_u.h @@ -4,11 +4,17 @@ #pragma once -#include "core/hle/service/service.h" +namespace Core { +class System; +} + +namespace Service::SM { +class ServiceManager; +} namespace Service::MM { /// Registers all MM services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); } // namespace Service::MM diff --git a/src/core/hle/service/ncm/ncm.cpp b/src/core/hle/service/ncm/ncm.cpp index e38dea1f4..2dcda16f6 100644 --- a/src/core/hle/service/ncm/ncm.cpp +++ b/src/core/hle/service/ncm/ncm.cpp @@ -14,8 +14,8 @@ namespace Service::NCM { class ILocationResolver final : public ServiceFramework<ILocationResolver> { public: - explicit ILocationResolver(FileSys::StorageId id) - : ServiceFramework{"ILocationResolver"}, storage(id) { + explicit ILocationResolver(Core::System& system_, FileSys::StorageId id) + : ServiceFramework{system_, "ILocationResolver"}, storage{id} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "ResolveProgramPath"}, @@ -45,12 +45,13 @@ public: } private: - FileSys::StorageId storage; + [[maybe_unused]] FileSys::StorageId storage; }; class IRegisteredLocationResolver final : public ServiceFramework<IRegisteredLocationResolver> { public: - explicit IRegisteredLocationResolver() : ServiceFramework{"IRegisteredLocationResolver"} { + explicit IRegisteredLocationResolver(Core::System& system_) + : ServiceFramework{system_, "IRegisteredLocationResolver"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "ResolveProgramPath"}, @@ -72,7 +73,8 @@ public: class IAddOnContentLocationResolver final : public ServiceFramework<IAddOnContentLocationResolver> { public: - explicit IAddOnContentLocationResolver() : ServiceFramework{"IAddOnContentLocationResolver"} { + explicit IAddOnContentLocationResolver(Core::System& system_) + : ServiceFramework{system_, "IAddOnContentLocationResolver"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "ResolveAddOnContentPath"}, @@ -89,7 +91,7 @@ public: class LR final : public ServiceFramework<LR> { public: - explicit LR() : ServiceFramework{"lr"} { + explicit LR(Core::System& system_) : ServiceFramework{system_, "lr"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "OpenLocationResolver"}, @@ -105,7 +107,7 @@ public: class NCM final : public ServiceFramework<NCM> { public: - explicit NCM() : ServiceFramework{"ncm"} { + explicit NCM(Core::System& system_) : ServiceFramework{system_, "ncm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CreateContentStorage"}, @@ -130,9 +132,9 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<LR>()->InstallAsService(sm); - std::make_shared<NCM>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<LR>(system)->InstallAsService(sm); + std::make_shared<NCM>(system)->InstallAsService(sm); } } // namespace Service::NCM diff --git a/src/core/hle/service/ncm/ncm.h b/src/core/hle/service/ncm/ncm.h index 7bc8518a6..ee01eddc0 100644 --- a/src/core/hle/service/ncm/ncm.h +++ b/src/core/hle/service/ncm/ncm.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::NCM { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::NCM diff --git a/src/core/hle/service/nfc/nfc.cpp b/src/core/hle/service/nfc/nfc.cpp index 780ea30fe..6ab35de47 100644 --- a/src/core/hle/service/nfc/nfc.cpp +++ b/src/core/hle/service/nfc/nfc.cpp @@ -16,7 +16,7 @@ namespace Service::NFC { class IAm final : public ServiceFramework<IAm> { public: - explicit IAm() : ServiceFramework{"NFC::IAm"} { + explicit IAm(Core::System& system_) : ServiceFramework{system_, "NFC::IAm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -31,7 +31,7 @@ public: class NFC_AM final : public ServiceFramework<NFC_AM> { public: - explicit NFC_AM() : ServiceFramework{"nfc:am"} { + explicit NFC_AM(Core::System& system_) : ServiceFramework{system_, "nfc:am"} { // clang-format off static const FunctionInfo functions[] = { {0, &NFC_AM::CreateAmInterface, "CreateAmInterface"}, @@ -47,13 +47,13 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IAm>(); + rb.PushIpcInterface<IAm>(system); } }; class MFIUser final : public ServiceFramework<MFIUser> { public: - explicit MFIUser() : ServiceFramework{"NFC::MFIUser"} { + explicit MFIUser(Core::System& system_) : ServiceFramework{system_, "NFC::MFIUser"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -79,7 +79,7 @@ public: class NFC_MF_U final : public ServiceFramework<NFC_MF_U> { public: - explicit NFC_MF_U() : ServiceFramework{"nfc:mf:u"} { + explicit NFC_MF_U(Core::System& system_) : ServiceFramework{system_, "nfc:mf:u"} { // clang-format off static const FunctionInfo functions[] = { {0, &NFC_MF_U::CreateUserInterface, "CreateUserInterface"}, @@ -95,13 +95,13 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<MFIUser>(); + rb.PushIpcInterface<MFIUser>(system); } }; class IUser final : public ServiceFramework<IUser> { public: - explicit IUser() : ServiceFramework{"NFC::IUser"} { + explicit IUser(Core::System& system_) : ServiceFramework{system_, "NFC::IUser"} { // clang-format off static const FunctionInfo functions[] = { {0, &IUser::InitializeOld, "InitializeOld"}, @@ -171,7 +171,7 @@ private: class NFC_U final : public ServiceFramework<NFC_U> { public: - explicit NFC_U() : ServiceFramework{"nfc:user"} { + explicit NFC_U(Core::System& system_) : ServiceFramework{system_, "nfc:user"} { // clang-format off static const FunctionInfo functions[] = { {0, &NFC_U::CreateUserInterface, "CreateUserInterface"}, @@ -187,13 +187,13 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IUser>(); + rb.PushIpcInterface<IUser>(system); } }; class ISystem final : public ServiceFramework<ISystem> { public: - explicit ISystem() : ServiceFramework{"ISystem"} { + explicit ISystem(Core::System& system_) : ServiceFramework{system_, "ISystem"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -230,7 +230,7 @@ public: class NFC_SYS final : public ServiceFramework<NFC_SYS> { public: - explicit NFC_SYS() : ServiceFramework{"nfc:sys"} { + explicit NFC_SYS(Core::System& system_) : ServiceFramework{system_, "nfc:sys"} { // clang-format off static const FunctionInfo functions[] = { {0, &NFC_SYS::CreateSystemInterface, "CreateSystemInterface"}, @@ -246,15 +246,15 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystem>(); + rb.PushIpcInterface<ISystem>(system); } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<NFC_AM>()->InstallAsService(sm); - std::make_shared<NFC_MF_U>()->InstallAsService(sm); - std::make_shared<NFC_U>()->InstallAsService(sm); - std::make_shared<NFC_SYS>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<NFC_AM>(system)->InstallAsService(sm); + std::make_shared<NFC_MF_U>(system)->InstallAsService(sm); + std::make_shared<NFC_U>(system)->InstallAsService(sm); + std::make_shared<NFC_SYS>(system)->InstallAsService(sm); } } // namespace Service::NFC diff --git a/src/core/hle/service/nfc/nfc.h b/src/core/hle/service/nfc/nfc.h index 4d2d815f9..5a94b076d 100644 --- a/src/core/hle/service/nfc/nfc.h +++ b/src/core/hle/service/nfc/nfc.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::NFC { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::NFC diff --git a/src/core/hle/service/nfp/nfp.cpp b/src/core/hle/service/nfp/nfp.cpp index a0469ffbd..5557da72e 100644 --- a/src/core/hle/service/nfp/nfp.cpp +++ b/src/core/hle/service/nfp/nfp.cpp @@ -21,8 +21,9 @@ namespace ErrCodes { constexpr ResultCode ERR_NO_APPLICATION_AREA(ErrorModule::NFP, 152); } // namespace ErrCodes -Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& system, const char* name) - : ServiceFramework(name), module(std::move(module)), system(system) { +Module::Interface::Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name) + : ServiceFramework{system_, name}, module{std::move(module_)} { auto& kernel = system.Kernel(); nfc_tag_load = Kernel::WritableEvent::CreateEventPair(kernel, "IUser:NFCTagDetected"); } @@ -31,8 +32,8 @@ Module::Interface::~Interface() = default; class IUser final : public ServiceFramework<IUser> { public: - IUser(Module::Interface& nfp_interface, Core::System& system) - : ServiceFramework("NFP::IUser"), nfp_interface(nfp_interface) { + explicit IUser(Module::Interface& nfp_interface_, Core::System& system_) + : ServiceFramework{system_, "NFP::IUser"}, nfp_interface{nfp_interface_} { static const FunctionInfo functions[] = { {0, &IUser::Initialize, "Initialize"}, {1, &IUser::Finalize, "Finalize"}, diff --git a/src/core/hle/service/nfp/nfp.h b/src/core/hle/service/nfp/nfp.h index 200013795..295de535b 100644 --- a/src/core/hle/service/nfp/nfp.h +++ b/src/core/hle/service/nfp/nfp.h @@ -16,7 +16,8 @@ class Module final { public: class Interface : public ServiceFramework<Interface> { public: - explicit Interface(std::shared_ptr<Module> module, Core::System& system, const char* name); + explicit Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name); ~Interface() override; struct ModelInfo { @@ -43,7 +44,6 @@ public: protected: std::shared_ptr<Module> module; - Core::System& system; }; }; diff --git a/src/core/hle/service/nfp/nfp_user.cpp b/src/core/hle/service/nfp/nfp_user.cpp index 298184f17..10b0ef944 100644 --- a/src/core/hle/service/nfp/nfp_user.cpp +++ b/src/core/hle/service/nfp/nfp_user.cpp @@ -6,8 +6,8 @@ namespace Service::NFP { -NFP_User::NFP_User(std::shared_ptr<Module> module, Core::System& system) - : Module::Interface(std::move(module), system, "nfp:user") { +NFP_User::NFP_User(std::shared_ptr<Module> module_, Core::System& system_) + : Interface(std::move(module_), system_, "nfp:user") { static const FunctionInfo functions[] = { {0, &NFP_User::CreateUserInterface, "CreateUserInterface"}, }; diff --git a/src/core/hle/service/nfp/nfp_user.h b/src/core/hle/service/nfp/nfp_user.h index 1686ebf20..7f3c124f6 100644 --- a/src/core/hle/service/nfp/nfp_user.h +++ b/src/core/hle/service/nfp/nfp_user.h @@ -10,7 +10,7 @@ namespace Service::NFP { class NFP_User final : public Module::Interface { public: - explicit NFP_User(std::shared_ptr<Module> module, Core::System& system); + explicit NFP_User(std::shared_ptr<Module> module_, Core::System& system_); ~NFP_User() override; }; diff --git a/src/core/hle/service/nifm/nifm.cpp b/src/core/hle/service/nifm/nifm.cpp index 2e9d95195..ef5176bea 100644 --- a/src/core/hle/service/nifm/nifm.cpp +++ b/src/core/hle/service/nifm/nifm.cpp @@ -23,7 +23,7 @@ enum class RequestState : u32 { class IScanRequest final : public ServiceFramework<IScanRequest> { public: - explicit IScanRequest() : ServiceFramework("IScanRequest") { + explicit IScanRequest(Core::System& system_) : ServiceFramework{system_, "IScanRequest"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Submit"}, @@ -40,7 +40,7 @@ public: class IRequest final : public ServiceFramework<IRequest> { public: - explicit IRequest(Core::System& system) : ServiceFramework("IRequest") { + explicit IRequest(Core::System& system_) : ServiceFramework{system_, "IRequest"} { static const FunctionInfo functions[] = { {0, &IRequest::GetRequestState, "GetRequestState"}, {1, &IRequest::GetResult, "GetResult"}, @@ -62,7 +62,7 @@ public: {18, nullptr, "SetRequirementByRevision"}, {19, nullptr, "GetRequirement"}, {20, nullptr, "GetRevision"}, - {21, nullptr, "GetAppletInfo"}, + {21, &IRequest::GetAppletInfo, "GetAppletInfo"}, {22, nullptr, "GetAdditionalInfo"}, {23, nullptr, "SetKeptInSleep"}, {24, nullptr, "RegisterSocketDescriptor"}, @@ -125,12 +125,22 @@ private: rb.Push(RESULT_SUCCESS); } + void GetAppletInfo(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NIFM, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 8}; + rb.Push(RESULT_SUCCESS); + rb.Push<u32>(0); + rb.Push<u32>(0); + rb.Push<u32>(0); + } + Kernel::EventPair event1, event2; }; class INetworkProfile final : public ServiceFramework<INetworkProfile> { public: - explicit INetworkProfile() : ServiceFramework("INetworkProfile") { + explicit INetworkProfile(Core::System& system_) : ServiceFramework{system_, "INetworkProfile"} { static const FunctionInfo functions[] = { {0, nullptr, "Update"}, {1, nullptr, "PersistOld"}, @@ -142,7 +152,7 @@ public: class IGeneralService final : public ServiceFramework<IGeneralService> { public: - IGeneralService(Core::System& system); + explicit IGeneralService(Core::System& system_); private: void GetClientId(Kernel::HLERequestContext& ctx) { @@ -159,7 +169,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IScanRequest>(); + rb.PushIpcInterface<IScanRequest>(system); } void CreateRequest(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NIFM, "called"); @@ -197,7 +207,7 @@ private: IPC::ResponseBuilder rb{ctx, 6, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<INetworkProfile>(); + rb.PushIpcInterface<INetworkProfile>(system); rb.PushRaw<u128>(uuid); } void IsWirelessCommunicationEnabled(Kernel::HLERequestContext& ctx) { @@ -229,11 +239,10 @@ private: rb.Push<u8>(1); } } - Core::System& system; }; -IGeneralService::IGeneralService(Core::System& system) - : ServiceFramework("IGeneralService"), system(system) { +IGeneralService::IGeneralService(Core::System& system_) + : ServiceFramework{system_, "IGeneralService"} { // clang-format off static const FunctionInfo functions[] = { {1, &IGeneralService::GetClientId, "GetClientId"}, @@ -286,8 +295,8 @@ IGeneralService::IGeneralService(Core::System& system) class NetworkInterface final : public ServiceFramework<NetworkInterface> { public: - explicit NetworkInterface(const char* name, Core::System& system) - : ServiceFramework{name}, system(system) { + explicit NetworkInterface(const char* name, Core::System& system_) + : ServiceFramework{system_, name} { static const FunctionInfo functions[] = { {4, &NetworkInterface::CreateGeneralServiceOld, "CreateGeneralServiceOld"}, {5, &NetworkInterface::CreateGeneralService, "CreateGeneralService"}, @@ -295,6 +304,7 @@ public: RegisterHandlers(functions); } +private: void CreateGeneralServiceOld(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NIFM, "called"); @@ -310,9 +320,6 @@ public: rb.Push(RESULT_SUCCESS); rb.PushIpcInterface<IGeneralService>(system); } - -private: - Core::System& system; }; void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { diff --git a/src/core/hle/service/nifm/nifm.h b/src/core/hle/service/nifm/nifm.h index 6857e18f9..c3dd4f386 100644 --- a/src/core/hle/service/nifm/nifm.h +++ b/src/core/hle/service/nifm/nifm.h @@ -4,14 +4,14 @@ #pragma once -namespace Service::SM { -class ServiceManager; -} - namespace Core { class System; } +namespace Service::SM { +class ServiceManager; +} + namespace Service::NIFM { /// Registers all NIFM services with the specified service manager. diff --git a/src/core/hle/service/nim/nim.cpp b/src/core/hle/service/nim/nim.cpp index 11aa74828..d16223064 100644 --- a/src/core/hle/service/nim/nim.cpp +++ b/src/core/hle/service/nim/nim.cpp @@ -17,7 +17,8 @@ namespace Service::NIM { class IShopServiceAsync final : public ServiceFramework<IShopServiceAsync> { public: - IShopServiceAsync() : ServiceFramework("IShopServiceAsync") { + explicit IShopServiceAsync(Core::System& system_) + : ServiceFramework{system_, "IShopServiceAsync"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Cancel"}, @@ -35,7 +36,8 @@ public: class IShopServiceAccessor final : public ServiceFramework<IShopServiceAccessor> { public: - IShopServiceAccessor() : ServiceFramework("IShopServiceAccessor") { + explicit IShopServiceAccessor(Core::System& system_) + : ServiceFramework{system_, "IShopServiceAccessor"} { // clang-format off static const FunctionInfo functions[] = { {0, &IShopServiceAccessor::CreateAsyncInterface, "CreateAsyncInterface"}, @@ -50,13 +52,14 @@ private: LOG_WARNING(Service_NIM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IShopServiceAsync>(); + rb.PushIpcInterface<IShopServiceAsync>(system); } }; class IShopServiceAccessServer final : public ServiceFramework<IShopServiceAccessServer> { public: - IShopServiceAccessServer() : ServiceFramework("IShopServiceAccessServer") { + explicit IShopServiceAccessServer(Core::System& system_) + : ServiceFramework{system_, "IShopServiceAccessServer"} { // clang-format off static const FunctionInfo functions[] = { {0, &IShopServiceAccessServer::CreateAccessorInterface, "CreateAccessorInterface"}, @@ -71,13 +74,13 @@ private: LOG_WARNING(Service_NIM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IShopServiceAccessor>(); + rb.PushIpcInterface<IShopServiceAccessor>(system); } }; class NIM final : public ServiceFramework<NIM> { public: - explicit NIM() : ServiceFramework{"nim"} { + explicit NIM(Core::System& system_) : ServiceFramework{system_, "nim"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CreateSystemUpdateTask"}, @@ -207,14 +210,14 @@ public: class NIM_ECA final : public ServiceFramework<NIM_ECA> { public: - explicit NIM_ECA() : ServiceFramework{"nim:eca"} { + explicit NIM_ECA(Core::System& system_) : ServiceFramework{system_, "nim:eca"} { // clang-format off static const FunctionInfo functions[] = { {0, &NIM_ECA::CreateServerInterface, "CreateServerInterface"}, {1, nullptr, "RefreshDebugAvailability"}, {2, nullptr, "ClearDebugResponse"}, {3, nullptr, "RegisterDebugResponse"}, - {4, nullptr, "IsLargeResourceAvailable"}, + {4, &NIM_ECA::IsLargeResourceAvailable, "IsLargeResourceAvailable"}, }; // clang-format on @@ -226,13 +229,25 @@ private: LOG_WARNING(Service_NIM, "(STUBBED) called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IShopServiceAccessServer>(); + rb.PushIpcInterface<IShopServiceAccessServer>(system); + } + + void IsLargeResourceAvailable(Kernel::HLERequestContext& ctx) { + IPC::RequestParser rp{ctx}; + + const auto unknown{rp.Pop<u64>()}; + + LOG_INFO(Service_NIM, "(STUBBED) called, unknown={}", unknown); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(false); } }; class NIM_SHP final : public ServiceFramework<NIM_SHP> { public: - explicit NIM_SHP() : ServiceFramework{"nim:shp"} { + explicit NIM_SHP(Core::System& system_) : ServiceFramework{system_, "nim:shp"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestDeviceAuthenticationToken"}, @@ -272,8 +287,8 @@ public: class IEnsureNetworkClockAvailabilityService final : public ServiceFramework<IEnsureNetworkClockAvailabilityService> { public: - explicit IEnsureNetworkClockAvailabilityService(Core::System& system) - : ServiceFramework("IEnsureNetworkClockAvailabilityService") { + explicit IEnsureNetworkClockAvailabilityService(Core::System& system_) + : ServiceFramework{system_, "IEnsureNetworkClockAvailabilityService"} { static const FunctionInfo functions[] = { {0, &IEnsureNetworkClockAvailabilityService::StartTask, "StartTask"}, {1, &IEnsureNetworkClockAvailabilityService::GetFinishNotificationEvent, @@ -345,7 +360,7 @@ private: class NTC final : public ServiceFramework<NTC> { public: - explicit NTC(Core::System& system) : ServiceFramework{"ntc"}, system(system) { + explicit NTC(Core::System& system_) : ServiceFramework{system_, "ntc"} { // clang-format off static const FunctionInfo functions[] = { {0, &NTC::OpenEnsureNetworkClockAvailabilityService, "OpenEnsureNetworkClockAvailabilityService"}, @@ -380,13 +395,12 @@ private: IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } - Core::System& system; }; void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { - std::make_shared<NIM>()->InstallAsService(sm); - std::make_shared<NIM_ECA>()->InstallAsService(sm); - std::make_shared<NIM_SHP>()->InstallAsService(sm); + std::make_shared<NIM>(system)->InstallAsService(sm); + std::make_shared<NIM_ECA>(system)->InstallAsService(sm); + std::make_shared<NIM_SHP>(system)->InstallAsService(sm); std::make_shared<NTC>(system)->InstallAsService(sm); } diff --git a/src/core/hle/service/nim/nim.h b/src/core/hle/service/nim/nim.h index dbe25dc01..571153fe6 100644 --- a/src/core/hle/service/nim/nim.h +++ b/src/core/hle/service/nim/nim.h @@ -4,14 +4,14 @@ #pragma once -namespace Service::SM { -class ServiceManager; -} - namespace Core { class System; } +namespace Service::SM { +class ServiceManager; +} + namespace Service::NIM { void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); diff --git a/src/core/hle/service/npns/npns.cpp b/src/core/hle/service/npns/npns.cpp index 8fa16fb08..f7a58f659 100644 --- a/src/core/hle/service/npns/npns.cpp +++ b/src/core/hle/service/npns/npns.cpp @@ -12,7 +12,7 @@ namespace Service::NPNS { class NPNS_S final : public ServiceFramework<NPNS_S> { public: - explicit NPNS_S() : ServiceFramework{"npns:s"} { + explicit NPNS_S(Core::System& system_) : ServiceFramework{system_, "npns:s"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "ListenAll"}, @@ -62,7 +62,7 @@ public: class NPNS_U final : public ServiceFramework<NPNS_U> { public: - explicit NPNS_U() : ServiceFramework{"npns:u"} { + explicit NPNS_U(Core::System& system_) : ServiceFramework{system_, "npns:u"} { // clang-format off static const FunctionInfo functions[] = { {1, nullptr, "ListenAll"}, @@ -91,9 +91,9 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<NPNS_S>()->InstallAsService(sm); - std::make_shared<NPNS_U>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<NPNS_S>(system)->InstallAsService(sm); + std::make_shared<NPNS_U>(system)->InstallAsService(sm); } } // namespace Service::NPNS diff --git a/src/core/hle/service/npns/npns.h b/src/core/hle/service/npns/npns.h index 861cd3e48..3b7596b6b 100644 --- a/src/core/hle/service/npns/npns.h +++ b/src/core/hle/service/npns/npns.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::NPNS { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::NPNS diff --git a/src/core/hle/service/ns/ns.cpp b/src/core/hle/service/ns/ns.cpp index 58ee1f712..6ccf8995c 100644 --- a/src/core/hle/service/ns/ns.cpp +++ b/src/core/hle/service/ns/ns.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include "common/logging/log.h" +#include "core/core.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" #include "core/file_sys/vfs.h" @@ -17,7 +18,8 @@ namespace Service::NS { -IAccountProxyInterface::IAccountProxyInterface() : ServiceFramework{"IAccountProxyInterface"} { +IAccountProxyInterface::IAccountProxyInterface(Core::System& system_) + : ServiceFramework{system_, "IAccountProxyInterface"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CreateUserAccount"}, @@ -29,8 +31,8 @@ IAccountProxyInterface::IAccountProxyInterface() : ServiceFramework{"IAccountPro IAccountProxyInterface::~IAccountProxyInterface() = default; -IApplicationManagerInterface::IApplicationManagerInterface() - : ServiceFramework{"IApplicationManagerInterface"} { +IApplicationManagerInterface::IApplicationManagerInterface(Core::System& system_) + : ServiceFramework{system_, "IApplicationManagerInterface"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "ListApplicationRecord"}, @@ -298,7 +300,8 @@ void IApplicationManagerInterface::GetApplicationControlData(Kernel::HLERequestC const auto size = ctx.GetWriteBufferSize(); - const FileSys::PatchManager pm{title_id}; + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto control = pm.GetControlMetadata(); std::vector<u8> out; @@ -426,8 +429,8 @@ ResultVal<u64> IApplicationManagerInterface::ConvertApplicationLanguageToLanguag return MakeResult(static_cast<u64>(*language_code)); } -IApplicationVersionInterface::IApplicationVersionInterface() - : ServiceFramework{"IApplicationVersionInterface"} { +IApplicationVersionInterface::IApplicationVersionInterface(Core::System& system_) + : ServiceFramework{system_, "IApplicationVersionInterface"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetLaunchRequiredVersion"}, @@ -447,8 +450,8 @@ IApplicationVersionInterface::IApplicationVersionInterface() IApplicationVersionInterface::~IApplicationVersionInterface() = default; -IContentManagementInterface::IContentManagementInterface() - : ServiceFramework{"IContentManagementInterface"} { +IContentManagementInterface::IContentManagementInterface(Core::System& system_) + : ServiceFramework{system_, "IContentManagementInterface"} { // clang-format off static const FunctionInfo functions[] = { {11, nullptr, "CalculateApplicationOccupiedSize"}, @@ -467,7 +470,8 @@ IContentManagementInterface::IContentManagementInterface() IContentManagementInterface::~IContentManagementInterface() = default; -IDocumentInterface::IDocumentInterface() : ServiceFramework{"IDocumentInterface"} { +IDocumentInterface::IDocumentInterface(Core::System& system_) + : ServiceFramework{system_, "IDocumentInterface"} { // clang-format off static const FunctionInfo functions[] = { {21, nullptr, "GetApplicationContentPath"}, @@ -481,7 +485,8 @@ IDocumentInterface::IDocumentInterface() : ServiceFramework{"IDocumentInterface" IDocumentInterface::~IDocumentInterface() = default; -IDownloadTaskInterface::IDownloadTaskInterface() : ServiceFramework{"IDownloadTaskInterface"} { +IDownloadTaskInterface::IDownloadTaskInterface(Core::System& system_) + : ServiceFramework{system_, "IDownloadTaskInterface"} { // clang-format off static const FunctionInfo functions[] = { {701, nullptr, "ClearTaskStatusList"}, @@ -501,7 +506,8 @@ IDownloadTaskInterface::IDownloadTaskInterface() : ServiceFramework{"IDownloadTa IDownloadTaskInterface::~IDownloadTaskInterface() = default; -IECommerceInterface::IECommerceInterface() : ServiceFramework{"IECommerceInterface"} { +IECommerceInterface::IECommerceInterface(Core::System& system_) + : ServiceFramework{system_, "IECommerceInterface"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RequestLinkDevice"}, @@ -519,8 +525,8 @@ IECommerceInterface::IECommerceInterface() : ServiceFramework{"IECommerceInterfa IECommerceInterface::~IECommerceInterface() = default; -IFactoryResetInterface::IFactoryResetInterface::IFactoryResetInterface() - : ServiceFramework{"IFactoryResetInterface"} { +IFactoryResetInterface::IFactoryResetInterface(Core::System& system_) + : ServiceFramework{system_, "IFactoryResetInterface"} { // clang-format off static const FunctionInfo functions[] = { {100, nullptr, "ResetToFactorySettings"}, @@ -538,14 +544,14 @@ IFactoryResetInterface::IFactoryResetInterface::IFactoryResetInterface() IFactoryResetInterface::~IFactoryResetInterface() = default; -NS::NS(const char* name) : ServiceFramework{name} { +NS::NS(const char* name, Core::System& system_) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { {7992, &NS::PushInterface<IECommerceInterface>, "GetECommerceInterface"}, {7993, &NS::PushInterface<IApplicationVersionInterface>, "GetApplicationVersionInterface"}, {7994, &NS::PushInterface<IFactoryResetInterface>, "GetFactoryResetInterface"}, {7995, &NS::PushInterface<IAccountProxyInterface>, "GetAccountProxyInterface"}, - {7996, &NS::PushInterface<IApplicationManagerInterface>, "GetApplicationManagerInterface"}, + {7996, &NS::PushIApplicationManagerInterface, "GetApplicationManagerInterface"}, {7997, &NS::PushInterface<IDownloadTaskInterface>, "GetDownloadTaskInterface"}, {7998, &NS::PushInterface<IContentManagementInterface>, "GetContentManagementInterface"}, {7999, &NS::PushInterface<IDocumentInterface>, "GetDocumentInterface"}, @@ -558,12 +564,12 @@ NS::NS(const char* name) : ServiceFramework{name} { NS::~NS() = default; std::shared_ptr<IApplicationManagerInterface> NS::GetApplicationManagerInterface() const { - return GetInterface<IApplicationManagerInterface>(); + return GetInterface<IApplicationManagerInterface>(system); } class NS_DEV final : public ServiceFramework<NS_DEV> { public: - explicit NS_DEV() : ServiceFramework{"ns:dev"} { + explicit NS_DEV(Core::System& system_) : ServiceFramework{system_, "ns:dev"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "LaunchProgram"}, @@ -590,7 +596,8 @@ public: class ISystemUpdateControl final : public ServiceFramework<ISystemUpdateControl> { public: - explicit ISystemUpdateControl() : ServiceFramework{"ISystemUpdateControl"} { + explicit ISystemUpdateControl(Core::System& system_) + : ServiceFramework{system_, "ISystemUpdateControl"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "HasDownloaded"}, @@ -625,7 +632,7 @@ public: class NS_SU final : public ServiceFramework<NS_SU> { public: - explicit NS_SU() : ServiceFramework{"ns:su"} { + explicit NS_SU(Core::System& system_) : ServiceFramework{system_, "ns:su"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetBackgroundNetworkUpdateState"}, @@ -657,16 +664,16 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemUpdateControl>(); + rb.PushIpcInterface<ISystemUpdateControl>(system); } }; class NS_VM final : public ServiceFramework<NS_VM> { public: - explicit NS_VM() : ServiceFramework{"ns:vm"} { + explicit NS_VM(Core::System& system_) : ServiceFramework{system_, "ns:vm"} { // clang-format off static const FunctionInfo functions[] = { - {1200, nullptr, "NeedsUpdateVulnerability"}, + {1200, &NS_VM::NeedsUpdateVulnerability, "NeedsUpdateVulnerability"}, {1201, nullptr, "UpdateSafeSystemVersionForDebug"}, {1202, nullptr, "GetSafeSystemVersion"}, }; @@ -674,19 +681,28 @@ public: RegisterHandlers(functions); } + +private: + void NeedsUpdateVulnerability(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_NS, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.Push(false); + } }; void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { - std::make_shared<NS>("ns:am2")->InstallAsService(service_manager); - std::make_shared<NS>("ns:ec")->InstallAsService(service_manager); - std::make_shared<NS>("ns:rid")->InstallAsService(service_manager); - std::make_shared<NS>("ns:rt")->InstallAsService(service_manager); - std::make_shared<NS>("ns:web")->InstallAsService(service_manager); + std::make_shared<NS>("ns:am2", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:ec", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:rid", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:rt", system)->InstallAsService(service_manager); + std::make_shared<NS>("ns:web", system)->InstallAsService(service_manager); - std::make_shared<NS_DEV>()->InstallAsService(service_manager); - std::make_shared<NS_SU>()->InstallAsService(service_manager); - std::make_shared<NS_VM>()->InstallAsService(service_manager); + std::make_shared<NS_DEV>(system)->InstallAsService(service_manager); + std::make_shared<NS_SU>(system)->InstallAsService(service_manager); + std::make_shared<NS_VM>(system)->InstallAsService(service_manager); std::make_shared<PL_U>(system)->InstallAsService(service_manager); } diff --git a/src/core/hle/service/ns/ns.h b/src/core/hle/service/ns/ns.h index c2554b878..991271f3e 100644 --- a/src/core/hle/service/ns/ns.h +++ b/src/core/hle/service/ns/ns.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service { namespace FileSystem { @@ -16,13 +20,13 @@ namespace NS { class IAccountProxyInterface final : public ServiceFramework<IAccountProxyInterface> { public: - explicit IAccountProxyInterface(); + explicit IAccountProxyInterface(Core::System& system_); ~IAccountProxyInterface() override; }; class IApplicationManagerInterface final : public ServiceFramework<IApplicationManagerInterface> { public: - explicit IApplicationManagerInterface(); + explicit IApplicationManagerInterface(Core::System& system_); ~IApplicationManagerInterface() override; ResultVal<u8> GetApplicationDesiredLanguage(u32 supported_languages); @@ -36,63 +40,71 @@ private: class IApplicationVersionInterface final : public ServiceFramework<IApplicationVersionInterface> { public: - explicit IApplicationVersionInterface(); + explicit IApplicationVersionInterface(Core::System& system_); ~IApplicationVersionInterface() override; }; class IContentManagementInterface final : public ServiceFramework<IContentManagementInterface> { public: - explicit IContentManagementInterface(); + explicit IContentManagementInterface(Core::System& system_); ~IContentManagementInterface() override; }; class IDocumentInterface final : public ServiceFramework<IDocumentInterface> { public: - explicit IDocumentInterface(); + explicit IDocumentInterface(Core::System& system_); ~IDocumentInterface() override; }; class IDownloadTaskInterface final : public ServiceFramework<IDownloadTaskInterface> { public: - explicit IDownloadTaskInterface(); + explicit IDownloadTaskInterface(Core::System& system_); ~IDownloadTaskInterface() override; }; class IECommerceInterface final : public ServiceFramework<IECommerceInterface> { public: - explicit IECommerceInterface(); + explicit IECommerceInterface(Core::System& system_); ~IECommerceInterface() override; }; class IFactoryResetInterface final : public ServiceFramework<IFactoryResetInterface> { public: - explicit IFactoryResetInterface(); + explicit IFactoryResetInterface(Core::System& system_); ~IFactoryResetInterface() override; }; class NS final : public ServiceFramework<NS> { public: - explicit NS(const char* name); + explicit NS(const char* name, Core::System& system_); ~NS() override; std::shared_ptr<IApplicationManagerInterface> GetApplicationManagerInterface() const; private: - template <typename T> + template <typename T, typename... Args> void PushInterface(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NS, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<T>(); + rb.PushIpcInterface<T>(system); + } + + void PushIApplicationManagerInterface(Kernel::HLERequestContext& ctx) { + LOG_DEBUG(Service_NS, "called"); + + IPC::ResponseBuilder rb{ctx, 2, 0, 1}; + rb.Push(RESULT_SUCCESS); + rb.PushIpcInterface<IApplicationManagerInterface>(system); } - template <typename T> - std::shared_ptr<T> GetInterface() const { + template <typename T, typename... Args> + std::shared_ptr<T> GetInterface(Args&&... args) const { static_assert(std::is_base_of_v<Kernel::SessionRequestHandler, T>, "Not a base of ServiceFrameworkBase"); - return std::make_shared<T>(); + return std::make_shared<T>(std::forward<Args>(args)...); } }; diff --git a/src/core/hle/service/ns/pl_u.cpp b/src/core/hle/service/ns/pl_u.cpp index 40838a225..71c7587db 100644 --- a/src/core/hle/service/ns/pl_u.cpp +++ b/src/core/hle/service/ns/pl_u.cpp @@ -27,42 +27,14 @@ namespace Service::NS { -enum class FontArchives : u64 { - Extension = 0x0100000000000810, - Standard = 0x0100000000000811, - Korean = 0x0100000000000812, - ChineseTraditional = 0x0100000000000813, - ChineseSimple = 0x0100000000000814, -}; - struct FontRegion { u32 offset; u32 size; }; -constexpr std::array<std::pair<FontArchives, const char*>, 7> SHARED_FONTS{ - std::make_pair(FontArchives::Standard, "nintendo_udsg-r_std_003.bfttf"), - std::make_pair(FontArchives::ChineseSimple, "nintendo_udsg-r_org_zh-cn_003.bfttf"), - std::make_pair(FontArchives::ChineseSimple, "nintendo_udsg-r_ext_zh-cn_003.bfttf"), - std::make_pair(FontArchives::ChineseTraditional, "nintendo_udjxh-db_zh-tw_003.bfttf"), - std::make_pair(FontArchives::Korean, "nintendo_udsg-r_ko_003.bfttf"), - std::make_pair(FontArchives::Extension, "nintendo_ext_003.bfttf"), - std::make_pair(FontArchives::Extension, "nintendo_ext2_003.bfttf"), -}; - -constexpr std::array<const char*, 7> SHARED_FONTS_TTF{ - "FontStandard.ttf", - "FontChineseSimplified.ttf", - "FontExtendedChineseSimplified.ttf", - "FontChineseTraditional.ttf", - "FontKorean.ttf", - "FontNintendoExtended.ttf", - "FontNintendoExtended2.ttf", -}; - // The below data is specific to shared font data dumped from Switch on f/w 2.2 // Virtual address and offsets/sizes likely will vary by dump -constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL}; +[[maybe_unused]] constexpr VAddr SHARED_FONT_MEM_VADDR{0x00000009d3016000ULL}; constexpr u32 EXPECTED_RESULT{0x7f9a0218}; // What we expect the decrypted bfttf first 4 bytes to be constexpr u32 EXPECTED_MAGIC{0x36f81a1e}; // What we expect the encrypted bfttf first 4 bytes to be constexpr u64 SHARED_FONT_MEM_SIZE{0x1100000}; @@ -90,6 +62,18 @@ static void DecryptSharedFont(const std::vector<u32>& input, Kernel::PhysicalMem offset += transformed_font.size() * sizeof(u32); } +void DecryptSharedFontToTTF(const std::vector<u32>& input, std::vector<u8>& output) { + ASSERT_MSG(input[0] == EXPECTED_MAGIC, "Failed to derive key, unexpected magic number"); + + const u32 KEY = input[0] ^ EXPECTED_RESULT; // Derive key using an inverse xor + std::vector<u32> transformed_font(input.size()); + // TODO(ogniK): Figure out a better way to do this + std::transform(input.begin(), input.end(), transformed_font.begin(), + [&KEY](u32 font_data) { return Common::swap32(font_data ^ KEY); }); + transformed_font[1] = Common::swap32(transformed_font[1]) ^ KEY; // "re-encrypt" the size + std::memcpy(output.data(), transformed_font.data() + 2, transformed_font.size() * sizeof(u32)); +} + void EncryptSharedFont(const std::vector<u32>& input, std::vector<u8>& output, std::size_t& offset) { ASSERT_MSG(offset + (input.size() * sizeof(u32)) < SHARED_FONT_MEM_SIZE, @@ -151,8 +135,8 @@ struct PL_U::Impl { std::vector<FontRegion> shared_font_regions; }; -PL_U::PL_U(Core::System& system) - : ServiceFramework("pl:u"), impl{std::make_unique<Impl>()}, system(system) { +PL_U::PL_U(Core::System& system_) + : ServiceFramework{system_, "pl:u"}, impl{std::make_unique<Impl>()} { // clang-format off static const FunctionInfo functions[] = { {0, &PL_U::RequestLoad, "RequestLoad"}, @@ -192,21 +176,18 @@ PL_U::PL_U(Core::System& system) } if (!romfs) { - LOG_ERROR(Service_NS, "Failed to find or synthesize {:016X}! Skipping", - static_cast<u64>(font.first)); + LOG_ERROR(Service_NS, "Failed to find or synthesize {:016X}! Skipping", font.first); continue; } const auto extracted_romfs = FileSys::ExtractRomFS(romfs); if (!extracted_romfs) { - LOG_ERROR(Service_NS, "Failed to extract RomFS for {:016X}! Skipping", - static_cast<u64>(font.first)); + LOG_ERROR(Service_NS, "Failed to extract RomFS for {:016X}! Skipping", font.first); continue; } const auto font_fp = extracted_romfs->GetFile(font.second); if (!font_fp) { - LOG_ERROR(Service_NS, "{:016X} has no file \"{}\"! Skipping", - static_cast<u64>(font.first), font.second); + LOG_ERROR(Service_NS, "{:016X} has no file \"{}\"! Skipping", font.first, font.second); continue; } std::vector<u32> font_data_u32(font_fp->GetSize() / sizeof(u32)); diff --git a/src/core/hle/service/ns/pl_u.h b/src/core/hle/service/ns/pl_u.h index 27161bd7a..f920c7f69 100644 --- a/src/core/hle/service/ns/pl_u.h +++ b/src/core/hle/service/ns/pl_u.h @@ -16,11 +16,30 @@ class FileSystemController; namespace NS { +enum class FontArchives : u64 { + Extension = 0x0100000000000810, + Standard = 0x0100000000000811, + Korean = 0x0100000000000812, + ChineseTraditional = 0x0100000000000813, + ChineseSimple = 0x0100000000000814, +}; + +constexpr std::array<std::pair<FontArchives, const char*>, 7> SHARED_FONTS{ + std::make_pair(FontArchives::Standard, "nintendo_udsg-r_std_003.bfttf"), + std::make_pair(FontArchives::ChineseSimple, "nintendo_udsg-r_org_zh-cn_003.bfttf"), + std::make_pair(FontArchives::ChineseSimple, "nintendo_udsg-r_ext_zh-cn_003.bfttf"), + std::make_pair(FontArchives::ChineseTraditional, "nintendo_udjxh-db_zh-tw_003.bfttf"), + std::make_pair(FontArchives::Korean, "nintendo_udsg-r_ko_003.bfttf"), + std::make_pair(FontArchives::Extension, "nintendo_ext_003.bfttf"), + std::make_pair(FontArchives::Extension, "nintendo_ext2_003.bfttf"), +}; + +void DecryptSharedFontToTTF(const std::vector<u32>& input, std::vector<u8>& output); void EncryptSharedFont(const std::vector<u32>& input, std::vector<u8>& output, std::size_t& offset); class PL_U final : public ServiceFramework<PL_U> { public: - explicit PL_U(Core::System& system); + explicit PL_U(Core::System& system_); ~PL_U() override; private: @@ -33,7 +52,6 @@ private: struct Impl; std::unique_ptr<Impl> impl; - Core::System& system; }; } // namespace NS diff --git a/src/core/hle/service/nvdrv/devices/nvdevice.h b/src/core/hle/service/nvdrv/devices/nvdevice.h index 0240d6643..5681599ba 100644 --- a/src/core/hle/service/nvdrv/devices/nvdevice.h +++ b/src/core/hle/service/nvdrv/devices/nvdevice.h @@ -24,25 +24,37 @@ public: explicit nvdevice(Core::System& system) : system{system} {} virtual ~nvdevice() = default; - union Ioctl { - u32_le raw; - BitField<0, 8, u32> cmd; - BitField<8, 8, u32> group; - BitField<16, 14, u32> length; - BitField<30, 1, u32> is_in; - BitField<31, 1, u32> is_out; - }; + /** + * Handles an ioctl1 request. + * @param command The ioctl command id. + * @param input A buffer containing the input data for the ioctl. + * @param output A buffer where the output data will be written to. + * @returns The result code of the ioctl. + */ + virtual NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) = 0; + + /** + * Handles an ioctl2 request. + * @param command The ioctl command id. + * @param input A buffer containing the input data for the ioctl. + * @param inline_input A buffer containing the input data for the ioctl which has been inlined. + * @param output A buffer where the output data will be written to. + * @returns The result code of the ioctl. + */ + virtual NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) = 0; /** - * Handles an ioctl request. + * Handles an ioctl3 request. * @param command The ioctl command id. * @param input A buffer containing the input data for the ioctl. * @param output A buffer where the output data will be written to. + * @param inline_output A buffer where the inlined output data will be written to. * @returns The result code of the ioctl. */ - virtual u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) = 0; + virtual NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) = 0; protected: Core::System& system; diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp index 3f7b8e670..ce615c758 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp @@ -18,11 +18,22 @@ nvdisp_disp0::nvdisp_disp0(Core::System& system, std::shared_ptr<nvmap> nvmap_de : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} nvdisp_disp0 ::~nvdisp_disp0() = default; -u32 nvdisp_disp0::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; +NvResult nvdisp_disp0::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvdisp_disp0::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvdisp_disp0::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height, diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h index 6fcdeee84..55a33b7e4 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h @@ -20,9 +20,11 @@ public: explicit nvdisp_disp0(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); ~nvdisp_disp0() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; /// Performs a screen flip, drawing the buffer pointed to by the handle. void flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height, u32 stride, diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp index 39bd2a45b..6b062e10e 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.cpp @@ -17,57 +17,77 @@ namespace Service::Nvidia::Devices { -namespace NvErrCodes { -constexpr u32 Success{}; -constexpr u32 OutOfMemory{static_cast<u32>(-12)}; -constexpr u32 InvalidInput{static_cast<u32>(-22)}; -} // namespace NvErrCodes - nvhost_as_gpu::nvhost_as_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev) : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} nvhost_as_gpu::~nvhost_as_gpu() = default; -u32 nvhost_as_gpu::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocInitalizeExCommand: - return InitalizeEx(input, output); - case IoctlCommand::IocAllocateSpaceCommand: - return AllocateSpace(input, output); - case IoctlCommand::IocMapBufferExCommand: - return MapBufferEx(input, output); - case IoctlCommand::IocBindChannelCommand: - return BindChannel(input, output); - case IoctlCommand::IocGetVaRegionsCommand: - return GetVARegions(input, output); - case IoctlCommand::IocUnmapBufferCommand: - return UnmapBuffer(input, output); +NvResult nvhost_as_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 'A': + switch (command.cmd) { + case 0x1: + return BindChannel(input, output); + case 0x2: + return AllocateSpace(input, output); + case 0x3: + return FreeSpace(input, output); + case 0x5: + return UnmapBuffer(input, output); + case 0x6: + return MapBufferEx(input, output); + case 0x8: + return GetVARegions(input, output); + case 0x9: + return InitalizeEx(input, output); + case 0x14: + return Remap(input, output); + default: + break; + } + break; default: break; } - if (static_cast<IoctlCommand>(command.cmd.Value()) == IoctlCommand::IocRemapCommand) { - return Remap(input, output); - } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - UNIMPLEMENTED_MSG("Unimplemented ioctl command"); - return 0; +NvResult nvhost_as_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_as_gpu::InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + switch (command.group) { + case 'A': + switch (command.cmd) { + case 0x8: + return GetVARegions(input, output, inline_output); + default: + break; + } + break; + default: + break; + } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_as_gpu::InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output) { IoctlInitalizeEx params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called, big_page_size=0x{:X}", params.big_page_size); - return 0; + return NvResult::Success; } -u32 nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output) { IoctlAllocSpace params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -81,22 +101,36 @@ u32 nvhost_as_gpu::AllocateSpace(const std::vector<u8>& input, std::vector<u8>& params.offset = system.GPU().MemoryManager().Allocate(size, params.align); } - auto result{NvErrCodes::Success}; + auto result = NvResult::Success; if (!params.offset) { LOG_CRITICAL(Service_NVDRV, "allocation failed for size {}", size); - result = NvErrCodes::OutOfMemory; + result = NvResult::InsufficientMemory; } std::memcpy(output.data(), ¶ms, output.size()); return result; } -u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::FreeSpace(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlFreeSpace params{}; + std::memcpy(¶ms, input.data(), input.size()); + + LOG_DEBUG(Service_NVDRV, "called, offset={:X}, pages={:X}, page_size={:X}", params.offset, + params.pages, params.page_size); + + system.GPU().MemoryManager().Unmap(params.offset, + static_cast<std::size_t>(params.pages) * params.page_size); + + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; +} + +NvResult nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) { const auto num_entries = input.size() / sizeof(IoctlRemapEntry); LOG_DEBUG(Service_NVDRV, "called, num_entries=0x{:X}", num_entries); - auto result{NvErrCodes::Success}; + auto result = NvResult::Success; std::vector<IoctlRemapEntry> entries(num_entries); std::memcpy(entries.data(), input.data(), input.size()); @@ -107,7 +141,7 @@ u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) const auto object{nvmap_dev->GetObject(entry.nvmap_handle)}; if (!object) { LOG_CRITICAL(Service_NVDRV, "invalid nvmap_handle={:X}", entry.nvmap_handle); - result = NvErrCodes::InvalidInput; + result = NvResult::InvalidState; break; } @@ -118,7 +152,7 @@ u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) if (!addr) { LOG_CRITICAL(Service_NVDRV, "map returned an invalid address!"); - result = NvErrCodes::InvalidInput; + result = NvResult::InvalidState; break; } } @@ -127,7 +161,7 @@ u32 nvhost_as_gpu::Remap(const std::vector<u8>& input, std::vector<u8>& output) return result; } -u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output) { IoctlMapBufferEx params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -141,7 +175,7 @@ u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& ou if (!object) { LOG_CRITICAL(Service_NVDRV, "invalid nvmap_handle={:X}", params.nvmap_handle); std::memcpy(output.data(), ¶ms, output.size()); - return NvErrCodes::InvalidInput; + return NvResult::InvalidState; } // The real nvservices doesn't make a distinction between handles and ids, and @@ -168,16 +202,16 @@ u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& ou params.mapping_size, params.offset); std::memcpy(output.data(), ¶ms, output.size()); - return NvErrCodes::InvalidInput; + return NvResult::InvalidState; } std::memcpy(output.data(), ¶ms, output.size()); - return NvErrCodes::Success; + return NvResult::Success; } else { LOG_CRITICAL(Service_NVDRV, "address not mapped offset={}", params.offset); std::memcpy(output.data(), ¶ms, output.size()); - return NvErrCodes::InvalidInput; + return NvResult::InvalidState; } } @@ -197,10 +231,10 @@ u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& ou params.offset = gpu.MemoryManager().Map(physical_address, params.offset, size); } - auto result{NvErrCodes::Success}; + auto result = NvResult::Success; if (!params.offset) { LOG_CRITICAL(Service_NVDRV, "failed to map size={}", size); - result = NvErrCodes::InvalidInput; + result = NvResult::InvalidState; } else { AddBufferMap(params.offset, size, physical_address, is_alloc); } @@ -209,7 +243,7 @@ u32 nvhost_as_gpu::MapBufferEx(const std::vector<u8>& input, std::vector<u8>& ou return result; } -u32 nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { IoctlUnmapBuffer params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -222,20 +256,42 @@ u32 nvhost_as_gpu::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& ou } std::memcpy(output.data(), ¶ms, output.size()); - return NvErrCodes::Success; + return NvResult::Success; } -u32 nvhost_as_gpu::BindChannel(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::BindChannel(const std::vector<u8>& input, std::vector<u8>& output) { IoctlBindChannel params{}; std::memcpy(¶ms, input.data(), input.size()); - - LOG_DEBUG(Service_NVDRV, "called, fd={:X}", params.fd); + LOG_WARNING(Service_NVDRV, "(STUBBED) called, fd={:X}", params.fd); channel = params.fd; - return 0; + return NvResult::Success; +} + +NvResult nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlGetVaRegions params{}; + std::memcpy(¶ms, input.data(), input.size()); + + LOG_WARNING(Service_NVDRV, "(STUBBED) called, buf_addr={:X}, buf_size={:X}", params.buf_addr, + params.buf_size); + + params.buf_size = 0x30; + params.regions[0].offset = 0x04000000; + params.regions[0].page_size = 0x1000; + params.regions[0].pages = 0x3fbfff; + + params.regions[1].offset = 0x04000000; + params.regions[1].page_size = 0x10000; + params.regions[1].pages = 0x1bffff; + + // TODO(ogniK): This probably can stay stubbed but should add support way way later + + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; } -u32 nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { IoctlGetVaRegions params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -254,7 +310,8 @@ u32 nvhost_as_gpu::GetVARegions(const std::vector<u8>& input, std::vector<u8>& o // TODO(ogniK): This probably can stay stubbed but should add support way way later std::memcpy(output.data(), ¶ms, output.size()); - return 0; + std::memcpy(inline_output.data(), ¶ms.regions, inline_output.size()); + return NvResult::Success; } std::optional<nvhost_as_gpu::BufferMap> nvhost_as_gpu::FindBufferMap(GPUVAddr gpu_addr) const { diff --git a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h index 9a0cdff0c..08035fa0e 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_as_gpu.h @@ -30,9 +30,11 @@ public: explicit nvhost_as_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); ~nvhost_as_gpu() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: class BufferMap final { @@ -74,31 +76,21 @@ private: bool is_allocated{}; }; - enum class IoctlCommand : u32_le { - IocInitalizeExCommand = 0x40284109, - IocAllocateSpaceCommand = 0xC0184102, - IocRemapCommand = 0x00000014, - IocMapBufferExCommand = 0xC0284106, - IocBindChannelCommand = 0x40044101, - IocGetVaRegionsCommand = 0xC0404108, - IocUnmapBufferCommand = 0xC0084105, - }; - struct IoctlInitalizeEx { - u32_le big_page_size; // depends on GPU's available_big_page_sizes; 0=default - s32_le as_fd; // ignored; passes 0 - u32_le flags; // passes 0 - u32_le reserved; // ignored; passes 0 - u64_le unk0; - u64_le unk1; - u64_le unk2; + u32_le big_page_size{}; // depends on GPU's available_big_page_sizes; 0=default + s32_le as_fd{}; // ignored; passes 0 + u32_le flags{}; // passes 0 + u32_le reserved{}; // ignored; passes 0 + u64_le unk0{}; + u64_le unk1{}; + u64_le unk2{}; }; static_assert(sizeof(IoctlInitalizeEx) == 40, "IoctlInitalizeEx is incorrect size"); struct IoctlAllocSpace { - u32_le pages; - u32_le page_size; - AddressSpaceFlags flags; + u32_le pages{}; + u32_le page_size{}; + AddressSpaceFlags flags{}; INSERT_PADDING_WORDS(1); union { u64_le offset; @@ -107,63 +99,74 @@ private: }; static_assert(sizeof(IoctlAllocSpace) == 24, "IoctlInitalizeEx is incorrect size"); + struct IoctlFreeSpace { + u64_le offset{}; + u32_le pages{}; + u32_le page_size{}; + }; + static_assert(sizeof(IoctlFreeSpace) == 16, "IoctlFreeSpace is incorrect size"); + struct IoctlRemapEntry { - u16_le flags; - u16_le kind; - u32_le nvmap_handle; - u32_le map_offset; - u32_le offset; - u32_le pages; + u16_le flags{}; + u16_le kind{}; + u32_le nvmap_handle{}; + u32_le map_offset{}; + u32_le offset{}; + u32_le pages{}; }; static_assert(sizeof(IoctlRemapEntry) == 20, "IoctlRemapEntry is incorrect size"); struct IoctlMapBufferEx { - AddressSpaceFlags flags; // bit0: fixed_offset, bit2: cacheable - u32_le kind; // -1 is default - u32_le nvmap_handle; - u32_le page_size; // 0 means don't care - s64_le buffer_offset; - u64_le mapping_size; - s64_le offset; + AddressSpaceFlags flags{}; // bit0: fixed_offset, bit2: cacheable + u32_le kind{}; // -1 is default + u32_le nvmap_handle{}; + u32_le page_size{}; // 0 means don't care + s64_le buffer_offset{}; + u64_le mapping_size{}; + s64_le offset{}; }; static_assert(sizeof(IoctlMapBufferEx) == 40, "IoctlMapBufferEx is incorrect size"); struct IoctlUnmapBuffer { - s64_le offset; + s64_le offset{}; }; static_assert(sizeof(IoctlUnmapBuffer) == 8, "IoctlUnmapBuffer is incorrect size"); struct IoctlBindChannel { - u32_le fd; + s32_le fd{}; }; static_assert(sizeof(IoctlBindChannel) == 4, "IoctlBindChannel is incorrect size"); struct IoctlVaRegion { - u64_le offset; - u32_le page_size; + u64_le offset{}; + u32_le page_size{}; INSERT_PADDING_WORDS(1); - u64_le pages; + u64_le pages{}; }; static_assert(sizeof(IoctlVaRegion) == 24, "IoctlVaRegion is incorrect size"); struct IoctlGetVaRegions { - u64_le buf_addr; // (contained output user ptr on linux, ignored) - u32_le buf_size; // forced to 2*sizeof(struct va_region) - u32_le reserved; - IoctlVaRegion regions[2]; + u64_le buf_addr{}; // (contained output user ptr on linux, ignored) + u32_le buf_size{}; // forced to 2*sizeof(struct va_region) + u32_le reserved{}; + IoctlVaRegion regions[2]{}; }; static_assert(sizeof(IoctlGetVaRegions) == 16 + sizeof(IoctlVaRegion) * 2, "IoctlGetVaRegions is incorrect size"); - u32 channel{}; + s32 channel{}; + + NvResult InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output); + NvResult AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output); + NvResult Remap(const std::vector<u8>& input, std::vector<u8>& output); + NvResult MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output); + NvResult UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output); + NvResult FreeSpace(const std::vector<u8>& input, std::vector<u8>& output); + NvResult BindChannel(const std::vector<u8>& input, std::vector<u8>& output); - u32 InitalizeEx(const std::vector<u8>& input, std::vector<u8>& output); - u32 AllocateSpace(const std::vector<u8>& input, std::vector<u8>& output); - u32 Remap(const std::vector<u8>& input, std::vector<u8>& output); - u32 MapBufferEx(const std::vector<u8>& input, std::vector<u8>& output); - u32 UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output); - u32 BindChannel(const std::vector<u8>& input, std::vector<u8>& output); - u32 GetVARegions(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetVARegions(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetVARegions(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output); std::optional<BufferMap> FindBufferMap(GPUVAddr gpu_addr) const; void AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, bool is_allocated); diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp index b27ee0502..fea3b7b9f 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.cpp @@ -15,45 +15,59 @@ namespace Service::Nvidia::Devices { -nvhost_ctrl::nvhost_ctrl(Core::System& system, EventInterface& events_interface) - : nvdevice(system), events_interface{events_interface} {} +nvhost_ctrl::nvhost_ctrl(Core::System& system, EventInterface& events_interface, + SyncpointManager& syncpoint_manager) + : nvdevice(system), events_interface{events_interface}, syncpoint_manager{syncpoint_manager} {} nvhost_ctrl::~nvhost_ctrl() = default; -u32 nvhost_ctrl::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocGetConfigCommand: - return NvOsGetConfigU32(input, output); - case IoctlCommand::IocCtrlEventWaitCommand: - return IocCtrlEventWait(input, output, false, ctrl); - case IoctlCommand::IocCtrlEventWaitAsyncCommand: - return IocCtrlEventWait(input, output, true, ctrl); - case IoctlCommand::IocCtrlEventRegisterCommand: - return IocCtrlEventRegister(input, output); - case IoctlCommand::IocCtrlEventUnregisterCommand: - return IocCtrlEventUnregister(input, output); - case IoctlCommand::IocCtrlEventSignalCommand: - return IocCtrlEventSignal(input, output); +NvResult nvhost_ctrl::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x1b: + return NvOsGetConfigU32(input, output); + case 0x1c: + return IocCtrlClearEventWait(input, output); + case 0x1d: + return IocCtrlEventWait(input, output, false); + case 0x1e: + return IocCtrlEventWait(input, output, true); + case 0x1f: + return IocCtrlEventRegister(input, output); + case 0x20: + return IocCtrlEventUnregister(input, output); + } + break; default: - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + break; } + + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_outpu) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output) { IocGetConfigParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_TRACE(Service_NVDRV, "called, setting={}!{}", params.domain_str.data(), params.param_str.data()); - return 0x30006; // Returns error on production mode + return NvResult::ConfigVarNotFound; // Returns error on production mode } -u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, - bool is_async, IoctlCtrl& ctrl) { +NvResult nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, + bool is_async) { IocCtrlEventWaitParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "syncpt_id={}, threshold={}, timeout={}, is_async={}", @@ -70,19 +84,33 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& return NvResult::BadParameter; } + if (syncpoint_manager.IsSyncpointExpired(params.syncpt_id, params.threshold)) { + params.value = syncpoint_manager.GetSyncpointMin(params.syncpt_id); + std::memcpy(output.data(), ¶ms, sizeof(params)); + return NvResult::Success; + } + + if (const auto new_value = syncpoint_manager.RefreshSyncpoint(params.syncpt_id); + syncpoint_manager.IsSyncpointExpired(params.syncpt_id, params.threshold)) { + params.value = new_value; + std::memcpy(output.data(), ¶ms, sizeof(params)); + return NvResult::Success; + } + auto event = events_interface.events[event_id]; auto& gpu = system.GPU(); + // This is mostly to take into account unimplemented features. As synced // gpu is always synced. if (!gpu.IsAsync()) { - event.writable->Signal(); + event.event.writable->Signal(); return NvResult::Success; } auto lock = gpu.LockSync(); - const u32 current_syncpoint_value = gpu.GetSyncpointValue(params.syncpt_id); + const u32 current_syncpoint_value = event.fence.value; const s32 diff = current_syncpoint_value - params.threshold; if (diff >= 0) { - event.writable->Signal(); + event.event.writable->Signal(); params.value = current_syncpoint_value; std::memcpy(output.data(), ¶ms, sizeof(params)); return NvResult::Success; @@ -109,14 +137,8 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& params.value = ((params.syncpt_id & 0xfff) << 16) | 0x10000000; } params.value |= event_id; - event.writable->Clear(); + event.event.writable->Clear(); gpu.RegisterSyncptInterrupt(params.syncpt_id, target_value); - if (!is_async && ctrl.fresh_call) { - ctrl.must_delay = true; - ctrl.timeout = params.timeout; - ctrl.event_id = event_id; - return NvResult::Timeout; - } std::memcpy(output.data(), ¶ms, sizeof(params)); return NvResult::Timeout; } @@ -124,7 +146,7 @@ u32 nvhost_ctrl::IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& return NvResult::BadParameter; } -u32 nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output) { IocCtrlEventRegisterParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); const u32 event_id = params.user_event_id & 0x00FF; @@ -139,7 +161,8 @@ u32 nvhost_ctrl::IocCtrlEventRegister(const std::vector<u8>& input, std::vector< return NvResult::Success; } -u32 nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input, + std::vector<u8>& output) { IocCtrlEventUnregisterParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); const u32 event_id = params.user_event_id & 0x00FF; @@ -154,24 +177,22 @@ u32 nvhost_ctrl::IocCtrlEventUnregister(const std::vector<u8>& input, std::vecto return NvResult::Success; } -u32 nvhost_ctrl::IocCtrlEventSignal(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl::IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output) { IocCtrlEventSignalParams params{}; std::memcpy(¶ms, input.data(), sizeof(params)); - // TODO(Blinkhawk): This is normally called when an NvEvents timeout on WaitSynchronization - // It is believed from RE to cancel the GPU Event. However, better research is required - u32 event_id = params.user_event_id & 0x00FF; - LOG_WARNING(Service_NVDRV, "(STUBBED) called, user_event_id: {:X}", event_id); + + u32 event_id = params.event_id & 0x00FF; + LOG_WARNING(Service_NVDRV, "cleared event wait on, event_id: {:X}", event_id); + if (event_id >= MaxNvEvents) { return NvResult::BadParameter; } if (events_interface.status[event_id] == EventState::Waiting) { - auto& gpu = system.GPU(); - if (gpu.CancelSyncptInterrupt(events_interface.assigned_syncpt[event_id], - events_interface.assigned_value[event_id])) { - events_interface.LiberateEvent(event_id); - events_interface.events[event_id].writable->Signal(); - } + events_interface.LiberateEvent(event_id); } + + syncpoint_manager.RefreshSyncpoint(events_interface.events[event_id].fence.id); + return NvResult::Success; } diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h index 9898623de..c5aa1362a 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl.h @@ -14,137 +14,120 @@ namespace Service::Nvidia::Devices { class nvhost_ctrl final : public nvdevice { public: - explicit nvhost_ctrl(Core::System& system, EventInterface& events_interface); + explicit nvhost_ctrl(Core::System& system, EventInterface& events_interface, + SyncpointManager& syncpoint_manager); ~nvhost_ctrl() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocSyncptReadCommand = 0xC0080014, - IocSyncptIncrCommand = 0x40040015, - IocSyncptWaitCommand = 0xC00C0016, - IocModuleMutexCommand = 0x40080017, - IocModuleRegRDWRCommand = 0xC0180018, - IocSyncptWaitexCommand = 0xC0100019, - IocSyncptReadMaxCommand = 0xC008001A, - IocGetConfigCommand = 0xC183001B, - IocCtrlEventSignalCommand = 0xC004001C, - IocCtrlEventWaitCommand = 0xC010001D, - IocCtrlEventWaitAsyncCommand = 0xC010001E, - IocCtrlEventRegisterCommand = 0xC004001F, - IocCtrlEventUnregisterCommand = 0xC0040020, - IocCtrlEventKillCommand = 0x40080021, - }; struct IocSyncptReadParams { - u32_le id; - u32_le value; + u32_le id{}; + u32_le value{}; }; static_assert(sizeof(IocSyncptReadParams) == 8, "IocSyncptReadParams is incorrect size"); struct IocSyncptIncrParams { - u32_le id; + u32_le id{}; }; static_assert(sizeof(IocSyncptIncrParams) == 4, "IocSyncptIncrParams is incorrect size"); struct IocSyncptWaitParams { - u32_le id; - u32_le thresh; - s32_le timeout; + u32_le id{}; + u32_le thresh{}; + s32_le timeout{}; }; static_assert(sizeof(IocSyncptWaitParams) == 12, "IocSyncptWaitParams is incorrect size"); struct IocModuleMutexParams { - u32_le id; - u32_le lock; // (0 = unlock and 1 = lock) + u32_le id{}; + u32_le lock{}; // (0 = unlock and 1 = lock) }; static_assert(sizeof(IocModuleMutexParams) == 8, "IocModuleMutexParams is incorrect size"); struct IocModuleRegRDWRParams { - u32_le id; - u32_le num_offsets; - u32_le block_size; - u32_le offsets; - u32_le values; - u32_le write; + u32_le id{}; + u32_le num_offsets{}; + u32_le block_size{}; + u32_le offsets{}; + u32_le values{}; + u32_le write{}; }; static_assert(sizeof(IocModuleRegRDWRParams) == 24, "IocModuleRegRDWRParams is incorrect size"); struct IocSyncptWaitexParams { - u32_le id; - u32_le thresh; - s32_le timeout; - u32_le value; + u32_le id{}; + u32_le thresh{}; + s32_le timeout{}; + u32_le value{}; }; static_assert(sizeof(IocSyncptWaitexParams) == 16, "IocSyncptWaitexParams is incorrect size"); struct IocSyncptReadMaxParams { - u32_le id; - u32_le value; + u32_le id{}; + u32_le value{}; }; static_assert(sizeof(IocSyncptReadMaxParams) == 8, "IocSyncptReadMaxParams is incorrect size"); struct IocGetConfigParams { - std::array<char, 0x41> domain_str; - std::array<char, 0x41> param_str; - std::array<char, 0x101> config_str; + std::array<char, 0x41> domain_str{}; + std::array<char, 0x41> param_str{}; + std::array<char, 0x101> config_str{}; }; static_assert(sizeof(IocGetConfigParams) == 387, "IocGetConfigParams is incorrect size"); struct IocCtrlEventSignalParams { - u32_le user_event_id; + u32_le event_id{}; }; static_assert(sizeof(IocCtrlEventSignalParams) == 4, "IocCtrlEventSignalParams is incorrect size"); struct IocCtrlEventWaitParams { - u32_le syncpt_id; - u32_le threshold; - s32_le timeout; - u32_le value; + u32_le syncpt_id{}; + u32_le threshold{}; + s32_le timeout{}; + u32_le value{}; }; static_assert(sizeof(IocCtrlEventWaitParams) == 16, "IocCtrlEventWaitParams is incorrect size"); struct IocCtrlEventWaitAsyncParams { - u32_le syncpt_id; - u32_le threshold; - u32_le timeout; - u32_le value; + u32_le syncpt_id{}; + u32_le threshold{}; + u32_le timeout{}; + u32_le value{}; }; static_assert(sizeof(IocCtrlEventWaitAsyncParams) == 16, "IocCtrlEventWaitAsyncParams is incorrect size"); struct IocCtrlEventRegisterParams { - u32_le user_event_id; + u32_le user_event_id{}; }; static_assert(sizeof(IocCtrlEventRegisterParams) == 4, "IocCtrlEventRegisterParams is incorrect size"); struct IocCtrlEventUnregisterParams { - u32_le user_event_id; + u32_le user_event_id{}; }; static_assert(sizeof(IocCtrlEventUnregisterParams) == 4, "IocCtrlEventUnregisterParams is incorrect size"); struct IocCtrlEventKill { - u64_le user_events; + u64_le user_events{}; }; static_assert(sizeof(IocCtrlEventKill) == 8, "IocCtrlEventKill is incorrect size"); - u32 NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output); - - u32 IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, bool is_async, - IoctlCtrl& ctrl); - - u32 IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output); - - u32 IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output); - - u32 IocCtrlEventSignal(const std::vector<u8>& input, std::vector<u8>& output); + NvResult NvOsGetConfigU32(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocCtrlEventWait(const std::vector<u8>& input, std::vector<u8>& output, bool is_async); + NvResult IocCtrlEventRegister(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocCtrlEventUnregister(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocCtrlClearEventWait(const std::vector<u8>& input, std::vector<u8>& output); EventInterface& events_interface; + SyncpointManager& syncpoint_manager; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp index fba89e7a6..0320d3ae2 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.cpp @@ -15,39 +15,66 @@ namespace Service::Nvidia::Devices { nvhost_ctrl_gpu::nvhost_ctrl_gpu(Core::System& system) : nvdevice(system) {} nvhost_ctrl_gpu::~nvhost_ctrl_gpu() = default; -u32 nvhost_ctrl_gpu::ioctl(Ioctl command, const std::vector<u8>& input, - const std::vector<u8>& input2, std::vector<u8>& output, - std::vector<u8>& output2, IoctlCtrl& ctrl, IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocGetCharacteristicsCommand: - return GetCharacteristics(input, output, output2, version); - case IoctlCommand::IocGetTPCMasksCommand: - return GetTPCMasks(input, output, output2, version); - case IoctlCommand::IocGetActiveSlotMaskCommand: - return GetActiveSlotMask(input, output); - case IoctlCommand::IocZcullGetCtxSizeCommand: - return ZCullGetCtxSize(input, output); - case IoctlCommand::IocZcullGetInfo: - return ZCullGetInfo(input, output); - case IoctlCommand::IocZbcSetTable: - return ZBCSetTable(input, output); - case IoctlCommand::IocZbcQueryTable: - return ZBCQueryTable(input, output); - case IoctlCommand::IocFlushL2: - return FlushL2(input, output); - case IoctlCommand::IocGetGpuTime: - return GetGpuTime(input, output); +NvResult nvhost_ctrl_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 'G': + switch (command.cmd) { + case 0x1: + return ZCullGetCtxSize(input, output); + case 0x2: + return ZCullGetInfo(input, output); + case 0x3: + return ZBCSetTable(input, output); + case 0x4: + return ZBCQueryTable(input, output); + case 0x5: + return GetCharacteristics(input, output); + case 0x6: + return GetTPCMasks(input, output); + case 0x7: + return FlushL2(input, output); + case 0x14: + return GetActiveSlotMask(input, output); + case 0x1c: + return GetGpuTime(input, output); + default: + break; + } + break; + } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_ctrl_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { + switch (command.group) { + case 'G': + switch (command.cmd) { + case 0x5: + return GetCharacteristics(input, output, inline_output); + case 0x6: + return GetTPCMasks(input, output, inline_output); + default: + break; + } + break; default: - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + break; } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& output2, IoctlVersion version) { +NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, + std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlCharacteristics params{}; std::memcpy(¶ms, input.data(), input.size()); @@ -88,36 +115,83 @@ u32 nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vecto params.gc.gr_compbit_store_base_hw = 0x0; params.gpu_characteristics_buf_size = 0xA0; params.gpu_characteristics_buf_addr = 0xdeadbeef; // Cannot be 0 (UNUSED) + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; +} - if (version == IoctlVersion::Version3) { - std::memcpy(output.data(), input.data(), output.size()); - std::memcpy(output2.data(), ¶ms.gc, output2.size()); - } else { - std::memcpy(output.data(), ¶ms, output.size()); - } - return 0; +NvResult nvhost_ctrl_gpu::GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + LOG_DEBUG(Service_NVDRV, "called"); + IoctlCharacteristics params{}; + std::memcpy(¶ms, input.data(), input.size()); + params.gc.arch = 0x120; + params.gc.impl = 0xb; + params.gc.rev = 0xa1; + params.gc.num_gpc = 0x1; + params.gc.l2_cache_size = 0x40000; + params.gc.on_board_video_memory_size = 0x0; + params.gc.num_tpc_per_gpc = 0x2; + params.gc.bus_type = 0x20; + params.gc.big_page_size = 0x20000; + params.gc.compression_page_size = 0x20000; + params.gc.pde_coverage_bit_count = 0x1B; + params.gc.available_big_page_sizes = 0x30000; + params.gc.gpc_mask = 0x1; + params.gc.sm_arch_sm_version = 0x503; + params.gc.sm_arch_spa_version = 0x503; + params.gc.sm_arch_warp_count = 0x80; + params.gc.gpu_va_bit_count = 0x28; + params.gc.reserved = 0x0; + params.gc.flags = 0x55; + params.gc.twod_class = 0x902D; + params.gc.threed_class = 0xB197; + params.gc.compute_class = 0xB1C0; + params.gc.gpfifo_class = 0xB06F; + params.gc.inline_to_memory_class = 0xA140; + params.gc.dma_copy_class = 0xB0B5; + params.gc.max_fbps_count = 0x1; + params.gc.fbp_en_mask = 0x0; + params.gc.max_ltc_per_fbp = 0x2; + params.gc.max_lts_per_ltc = 0x1; + params.gc.max_tex_per_tpc = 0x0; + params.gc.max_gpc_count = 0x1; + params.gc.rop_l2_en_mask_0 = 0x21D70; + params.gc.rop_l2_en_mask_1 = 0x0; + params.gc.chipname = 0x6230326D67; + params.gc.gr_compbit_store_base_hw = 0x0; + params.gpu_characteristics_buf_size = 0xA0; + params.gpu_characteristics_buf_addr = 0xdeadbeef; // Cannot be 0 (UNUSED) + + std::memcpy(output.data(), ¶ms, output.size()); + std::memcpy(inline_output.data(), ¶ms.gc, inline_output.size()); + return NvResult::Success; } -u32 nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& output2, IoctlVersion version) { +NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output) { IoctlGpuGetTpcMasksArgs params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); if (params.mask_buffer_size != 0) { params.tcp_mask = 3; } + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::Success; +} - if (version == IoctlVersion::Version3) { - std::memcpy(output.data(), input.data(), output.size()); - std::memcpy(output2.data(), ¶ms.tcp_mask, output2.size()); - } else { - std::memcpy(output.data(), ¶ms, output.size()); +NvResult nvhost_ctrl_gpu::GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + IoctlGpuGetTpcMasksArgs params{}; + std::memcpy(¶ms, input.data(), input.size()); + LOG_DEBUG(Service_NVDRV, "called, mask_buffer_size=0x{:X}", params.mask_buffer_size); + if (params.mask_buffer_size != 0) { + params.tcp_mask = 3; } - - return 0; + std::memcpy(output.data(), ¶ms, output.size()); + std::memcpy(inline_output.data(), ¶ms.tcp_mask, inline_output.size()); + return NvResult::Success; } -u32 nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlActiveSlotMask params{}; @@ -127,10 +201,10 @@ u32 nvhost_ctrl_gpu::GetActiveSlotMask(const std::vector<u8>& input, std::vector params.slot = 0x07; params.mask = 0x01; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlZcullGetCtxSize params{}; @@ -139,10 +213,10 @@ u32 nvhost_ctrl_gpu::ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u } params.size = 0x1; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlNvgpuGpuZcullGetInfoArgs params{}; @@ -162,47 +236,47 @@ u32 nvhost_ctrl_gpu::ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& params.subregion_height_align_pixels = 0x40; params.subregion_count = 0x10; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlZbcSetTable params{}; std::memcpy(¶ms, input.data(), input.size()); // TODO(ogniK): What does this even actually do? std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlZbcQueryTable params{}; std::memcpy(¶ms, input.data(), input.size()); // TODO : To implement properly std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::FlushL2(const std::vector<u8>& input, std::vector<u8>& output) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); IoctlFlushL2 params{}; std::memcpy(¶ms, input.data(), input.size()); // TODO : To implement properly std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_ctrl_gpu::GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_ctrl_gpu::GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlGetGpuTime params{}; std::memcpy(¶ms, input.data(), input.size()); params.gpu_time = static_cast<u64_le>(system.CoreTiming().GetGlobalTimeNs().count()); std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h index ef60f72ce..137b88238 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_ctrl_gpu.h @@ -16,32 +16,13 @@ public: explicit nvhost_ctrl_gpu(Core::System& system); ~nvhost_ctrl_gpu() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocGetCharacteristicsCommand = 0xC0B04705, - IocGetTPCMasksCommand = 0xC0184706, - IocGetActiveSlotMaskCommand = 0x80084714, - IocZcullGetCtxSizeCommand = 0x80044701, - IocZcullGetInfo = 0x80284702, - IocZbcSetTable = 0x402C4703, - IocZbcQueryTable = 0xC0344704, - IocFlushL2 = 0x40084707, - IocInvalICache = 0x4008470D, - IocSetMmudebugMode = 0x4008470E, - IocSetSmDebugMode = 0x4010470F, - IocWaitForPause = 0xC0084710, - IocGetTcpExceptionEnStatus = 0x80084711, - IocNumVsms = 0x80084712, - IocVsmsMapping = 0xC0044713, - IocGetErrorChannelUserData = 0xC008471B, - IocGetGpuTime = 0xC010471C, - IocGetCpuTimeCorrelationInfo = 0xC108471D, - }; - struct IoctlGpuCharacteristics { u32_le arch; // 0x120 (NVGPU_GPU_ARCH_GM200) u32_le impl; // 0xB (NVGPU_GPU_IMPL_GM20B) @@ -159,17 +140,21 @@ private: }; static_assert(sizeof(IoctlGetGpuTime) == 0x10, "IoctlGetGpuTime is incorrect size"); - u32 GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, - std::vector<u8>& output2, IoctlVersion version); - u32 GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output, std::vector<u8>& output2, - IoctlVersion version); - u32 GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output); - u32 FlushL2(const std::vector<u8>& input, std::vector<u8>& output); - u32 GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetCharacteristics(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output); + + NvResult GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetTPCMasks(const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output); + + NvResult GetActiveSlotMask(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZCullGetCtxSize(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZCullGetInfo(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZBCSetTable(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZBCQueryTable(const std::vector<u8>& input, std::vector<u8>& output); + NvResult FlushL2(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetGpuTime(const std::vector<u8>& input, std::vector<u8>& output); }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp index f1966ac0e..af8b3d9f1 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp @@ -7,117 +7,148 @@ #include "common/logging/log.h" #include "core/core.h" #include "core/hle/service/nvdrv/devices/nvhost_gpu.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" #include "core/memory.h" #include "video_core/gpu.h" #include "video_core/memory_manager.h" namespace Service::Nvidia::Devices { -nvhost_gpu::nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev) - : nvdevice(system), nvmap_dev(std::move(nvmap_dev)) {} +nvhost_gpu::nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager) + : nvdevice(system), nvmap_dev(std::move(nvmap_dev)), syncpoint_manager{syncpoint_manager} { + channel_fence.id = syncpoint_manager.AllocateSyncpoint(); + channel_fence.value = system.GPU().GetSyncpointValue(channel_fence.id); +} + nvhost_gpu::~nvhost_gpu() = default; -u32 nvhost_gpu::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); - case IoctlCommand::IocSetClientDataCommand: - return SetClientData(input, output); - case IoctlCommand::IocGetClientDataCommand: - return GetClientData(input, output); - case IoctlCommand::IocZCullBind: - return ZCullBind(input, output); - case IoctlCommand::IocSetErrorNotifierCommand: - return SetErrorNotifier(input, output); - case IoctlCommand::IocChannelSetPriorityCommand: - return SetChannelPriority(input, output); - case IoctlCommand::IocAllocGPFIFOEx2Command: - return AllocGPFIFOEx2(input, output); - case IoctlCommand::IocAllocObjCtxCommand: - return AllocateObjectContext(input, output); - case IoctlCommand::IocChannelGetWaitbaseCommand: - return GetWaitbase(input, output); - case IoctlCommand::IocChannelSetTimeoutCommand: - return ChannelSetTimeout(input, output); - case IoctlCommand::IocChannelSetTimeslice: - return ChannelSetTimeslice(input, output); - default: +NvResult nvhost_gpu::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x3: + return GetWaitbase(input, output); + default: + break; + } + break; + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input, output); + case 0x3: + return ChannelSetTimeout(input, output); + case 0x8: + return SubmitGPFIFOBase(input, output, false); + case 0x9: + return AllocateObjectContext(input, output); + case 0xb: + return ZCullBind(input, output); + case 0xc: + return SetErrorNotifier(input, output); + case 0xd: + return SetChannelPriority(input, output); + case 0x1a: + return AllocGPFIFOEx2(input, output); + case 0x1b: + return SubmitGPFIFOBase(input, output, true); + case 0x1d: + return ChannelSetTimeslice(input, output); + default: + break; + } + break; + case 'G': + switch (command.cmd) { + case 0x14: + return SetClientData(input, output); + case 0x15: + return GetClientData(input, output); + default: + break; + } break; } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +}; - if (command.group == NVGPU_IOCTL_MAGIC) { - if (command.cmd == NVGPU_IOCTL_CHANNEL_SUBMIT_GPFIFO) { - return SubmitGPFIFO(input, output); - } - if (command.cmd == NVGPU_IOCTL_CHANNEL_KICKOFF_PB) { - return KickoffPB(input, output, input2, version); +NvResult nvhost_gpu::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + switch (command.group) { + case 'H': + switch (command.cmd) { + case 0x1b: + return SubmitGPFIFOBase(input, inline_input, output); } + break; } + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; -}; +NvResult nvhost_gpu::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} -u32 nvhost_gpu::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); nvmap_fd = params.nvmap_fd; - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SetClientData(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetClientData(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlClientData params{}; std::memcpy(¶ms, input.data(), input.size()); user_data = params.data; - return 0; + return NvResult::Success; } -u32 nvhost_gpu::GetClientData(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::GetClientData(const std::vector<u8>& input, std::vector<u8>& output) { LOG_DEBUG(Service_NVDRV, "called"); IoctlClientData params{}; std::memcpy(¶ms, input.data(), input.size()); params.data = user_data; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::ZCullBind(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::ZCullBind(const std::vector<u8>& input, std::vector<u8>& output) { std::memcpy(&zcull_params, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, gpu_va={:X}, mode={:X}", zcull_params.gpu_va, zcull_params.mode); std::memcpy(output.data(), &zcull_params, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetErrorNotifier params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called, offset={:X}, size={:X}, mem={:X}", params.offset, params.size, params.mem); std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output) { std::memcpy(&channel_priority, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "(STUBBED) called, priority={:X}", channel_priority); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output) { IoctlAllocGpfifoEx2 params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, @@ -126,15 +157,15 @@ u32 nvhost_gpu::AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& ou params.num_entries, params.flags, params.unk0, params.unk1, params.unk2, params.unk3); - auto& gpu = system.GPU(); - params.fence_out.id = assigned_syncpoints; - params.fence_out.value = gpu.GetSyncpointValue(assigned_syncpoints); - assigned_syncpoints++; + channel_fence.value = system.GPU().GetSyncpointValue(channel_fence.id); + + params.fence_out = channel_fence; + std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output) { IoctlAllocObjCtx params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_WARNING(Service_NVDRV, "(STUBBED) called, class_num={:X}, flags={:X}", params.class_num, @@ -142,102 +173,149 @@ u32 nvhost_gpu::AllocateObjectContext(const std::vector<u8>& input, std::vector< params.obj_id = 0x0; std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output) { - if (input.size() < sizeof(IoctlSubmitGpfifo)) { - UNIMPLEMENTED(); +static std::vector<Tegra::CommandHeader> BuildWaitCommandList(Fence fence) { + return { + Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceValue, 1, + Tegra::SubmissionMode::Increasing), + {fence.value}, + Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceAction, 1, + Tegra::SubmissionMode::Increasing), + Tegra::GPU::FenceAction::Build(Tegra::GPU::FenceOperation::Acquire, fence.id), + }; +} + +static std::vector<Tegra::CommandHeader> BuildIncrementCommandList(Fence fence, u32 add_increment) { + std::vector<Tegra::CommandHeader> result{ + Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceValue, 1, + Tegra::SubmissionMode::Increasing), + {}}; + + for (u32 count = 0; count < add_increment; ++count) { + result.emplace_back(Tegra::BuildCommandHeader(Tegra::BufferMethods::FenceAction, 1, + Tegra::SubmissionMode::Increasing)); + result.emplace_back( + Tegra::GPU::FenceAction::Build(Tegra::GPU::FenceOperation::Increment, fence.id)); } - IoctlSubmitGpfifo params{}; - std::memcpy(¶ms, input.data(), sizeof(IoctlSubmitGpfifo)); + + return result; +} + +static std::vector<Tegra::CommandHeader> BuildIncrementWithWfiCommandList(Fence fence, + u32 add_increment) { + std::vector<Tegra::CommandHeader> result{ + Tegra::BuildCommandHeader(Tegra::BufferMethods::WaitForInterrupt, 1, + Tegra::SubmissionMode::Increasing), + {}}; + const std::vector<Tegra::CommandHeader> increment{ + BuildIncrementCommandList(fence, add_increment)}; + + result.insert(result.end(), increment.begin(), increment.end()); + + return result; +} + +NvResult nvhost_gpu::SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>& output, + Tegra::CommandList&& entries) { LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address, params.num_entries, params.flags.raw); - ASSERT_MSG(input.size() == sizeof(IoctlSubmitGpfifo) + - params.num_entries * sizeof(Tegra::CommandListHeader), - "Incorrect input size"); + auto& gpu = system.GPU(); - Tegra::CommandList entries(params.num_entries); - std::memcpy(entries.data(), &input[sizeof(IoctlSubmitGpfifo)], - params.num_entries * sizeof(Tegra::CommandListHeader)); + params.fence_out.id = channel_fence.id; - UNIMPLEMENTED_IF(params.flags.add_wait.Value() != 0); - UNIMPLEMENTED_IF(params.flags.add_increment.Value() != 0); + if (params.flags.add_wait.Value() && + !syncpoint_manager.IsSyncpointExpired(params.fence_out.id, params.fence_out.value)) { + gpu.PushGPUEntries(Tegra::CommandList{BuildWaitCommandList(params.fence_out)}); + } - auto& gpu = system.GPU(); - u32 current_syncpoint_value = gpu.GetSyncpointValue(params.fence_out.id); - if (params.flags.increment.Value()) { - params.fence_out.value += current_syncpoint_value; + if (params.flags.add_increment.Value() || params.flags.increment.Value()) { + const u32 increment_value = params.flags.increment.Value() ? params.fence_out.value : 0; + params.fence_out.value = syncpoint_manager.IncreaseSyncpoint( + params.fence_out.id, params.AddIncrementValue() + increment_value); } else { - params.fence_out.value = current_syncpoint_value; + params.fence_out.value = syncpoint_manager.GetSyncpointMax(params.fence_out.id); } + gpu.PushGPUEntries(std::move(entries)); + if (params.flags.add_increment.Value()) { + if (params.flags.suppress_wfi) { + gpu.PushGPUEntries(Tegra::CommandList{ + BuildIncrementCommandList(params.fence_out, params.AddIncrementValue())}); + } else { + gpu.PushGPUEntries(Tegra::CommandList{ + BuildIncrementWithWfiCommandList(params.fence_out, params.AddIncrementValue())}); + } + } + std::memcpy(output.data(), ¶ms, sizeof(IoctlSubmitGpfifo)); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::KickoffPB(const std::vector<u8>& input, std::vector<u8>& output, - const std::vector<u8>& input2, IoctlVersion version) { +NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input, std::vector<u8>& output, + bool kickoff) { if (input.size() < sizeof(IoctlSubmitGpfifo)) { UNIMPLEMENTED(); + return NvResult::InvalidSize; } IoctlSubmitGpfifo params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlSubmitGpfifo)); - LOG_TRACE(Service_NVDRV, "called, gpfifo={:X}, num_entries={:X}, flags={:X}", params.address, - params.num_entries, params.flags.raw); - Tegra::CommandList entries(params.num_entries); - if (version == IoctlVersion::Version2) { - std::memcpy(entries.data(), input2.data(), - params.num_entries * sizeof(Tegra::CommandListHeader)); - } else { - system.Memory().ReadBlock(params.address, entries.data(), - params.num_entries * sizeof(Tegra::CommandListHeader)); - } - UNIMPLEMENTED_IF(params.flags.add_wait.Value() != 0); - UNIMPLEMENTED_IF(params.flags.add_increment.Value() != 0); - auto& gpu = system.GPU(); - u32 current_syncpoint_value = gpu.GetSyncpointValue(params.fence_out.id); - if (params.flags.increment.Value()) { - params.fence_out.value += current_syncpoint_value; + if (kickoff) { + system.Memory().ReadBlock(params.address, entries.command_lists.data(), + params.num_entries * sizeof(Tegra::CommandListHeader)); } else { - params.fence_out.value = current_syncpoint_value; + std::memcpy(entries.command_lists.data(), &input[sizeof(IoctlSubmitGpfifo)], + params.num_entries * sizeof(Tegra::CommandListHeader)); } - gpu.PushGPUEntries(std::move(entries)); - std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return SubmitGPFIFOImpl(params, output, std::move(entries)); +} + +NvResult nvhost_gpu::SubmitGPFIFOBase(const std::vector<u8>& input, + const std::vector<u8>& input_inline, + std::vector<u8>& output) { + if (input.size() < sizeof(IoctlSubmitGpfifo)) { + UNIMPLEMENTED(); + return NvResult::InvalidSize; + } + IoctlSubmitGpfifo params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlSubmitGpfifo)); + Tegra::CommandList entries(params.num_entries); + std::memcpy(entries.command_lists.data(), input_inline.data(), input_inline.size()); + return SubmitGPFIFOImpl(params, output, std::move(entries)); } -u32 nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) { IoctlGetWaitbase params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlGetWaitbase)); LOG_INFO(Service_NVDRV, "called, unknown=0x{:X}", params.unknown); params.value = 0; // Seems to be hard coded at 0 std::memcpy(output.data(), ¶ms, output.size()); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output) { IoctlChannelSetTimeout params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlChannelSetTimeout)); LOG_INFO(Service_NVDRV, "called, timeout=0x{:X}", params.timeout); - return 0; + return NvResult::Success; } -u32 nvhost_gpu::ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_gpu::ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetTimeslice params{}; std::memcpy(¶ms, input.data(), sizeof(IoctlSetTimeslice)); LOG_INFO(Service_NVDRV, "called, timeslice=0x{:X}", params.timeslice); channel_timeslice = params.timeslice; - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h index 2ac74743f..e0298b4fe 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_gpu.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_gpu.h @@ -11,46 +11,28 @@ #include "common/swap.h" #include "core/hle/service/nvdrv/devices/nvdevice.h" #include "core/hle/service/nvdrv/nvdata.h" +#include "video_core/dma_pusher.h" + +namespace Service::Nvidia { +class SyncpointManager; +} namespace Service::Nvidia::Devices { class nvmap; -constexpr u32 NVGPU_IOCTL_MAGIC('H'); -constexpr u32 NVGPU_IOCTL_CHANNEL_SUBMIT_GPFIFO(0x8); -constexpr u32 NVGPU_IOCTL_CHANNEL_KICKOFF_PB(0x1b); - class nvhost_gpu final : public nvdevice { public: - explicit nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev); + explicit nvhost_gpu(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager); ~nvhost_gpu() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - IocAllocGPFIFOCommand = 0x40084805, - IocSetClientDataCommand = 0x40084714, - IocGetClientDataCommand = 0x80084715, - IocZCullBind = 0xc010480b, - IocSetErrorNotifierCommand = 0xC018480C, - IocChannelSetPriorityCommand = 0x4004480D, - IocEnableCommand = 0x0000480E, - IocDisableCommand = 0x0000480F, - IocPreemptCommand = 0x00004810, - IocForceResetCommand = 0x00004811, - IocEventIdControlCommand = 0x40084812, - IocGetErrorNotificationCommand = 0xC0104817, - IocAllocGPFIFOExCommand = 0x40204818, - IocAllocGPFIFOEx2Command = 0xC020481A, - IocAllocObjCtxCommand = 0xC0104809, - IocChannelGetWaitbaseCommand = 0xC0080003, - IocChannelSetTimeoutCommand = 0x40044803, - IocChannelSetTimeslice = 0xC004481D, - }; - enum class CtxObjects : u32_le { Ctx2D = 0x902D, Ctx3D = 0xB197, @@ -61,63 +43,63 @@ private: }; struct IoctlSetNvmapFD { - u32_le nvmap_fd; + s32_le nvmap_fd{}; }; static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); struct IoctlChannelSetTimeout { - u32_le timeout; + u32_le timeout{}; }; static_assert(sizeof(IoctlChannelSetTimeout) == 4, "IoctlChannelSetTimeout is incorrect size"); struct IoctlAllocGPFIFO { - u32_le num_entries; - u32_le flags; + u32_le num_entries{}; + u32_le flags{}; }; static_assert(sizeof(IoctlAllocGPFIFO) == 8, "IoctlAllocGPFIFO is incorrect size"); struct IoctlClientData { - u64_le data; + u64_le data{}; }; static_assert(sizeof(IoctlClientData) == 8, "IoctlClientData is incorrect size"); struct IoctlZCullBind { - u64_le gpu_va; - u32_le mode; // 0=global, 1=no_ctxsw, 2=separate_buffer, 3=part_of_regular_buf + u64_le gpu_va{}; + u32_le mode{}; // 0=global, 1=no_ctxsw, 2=separate_buffer, 3=part_of_regular_buf INSERT_PADDING_WORDS(1); }; static_assert(sizeof(IoctlZCullBind) == 16, "IoctlZCullBind is incorrect size"); struct IoctlSetErrorNotifier { - u64_le offset; - u64_le size; - u32_le mem; // nvmap object handle + u64_le offset{}; + u64_le size{}; + u32_le mem{}; // nvmap object handle INSERT_PADDING_WORDS(1); }; static_assert(sizeof(IoctlSetErrorNotifier) == 24, "IoctlSetErrorNotifier is incorrect size"); struct IoctlChannelSetPriority { - u32_le priority; + u32_le priority{}; }; static_assert(sizeof(IoctlChannelSetPriority) == 4, "IoctlChannelSetPriority is incorrect size"); struct IoctlSetTimeslice { - u32_le timeslice; + u32_le timeslice{}; }; static_assert(sizeof(IoctlSetTimeslice) == 4, "IoctlSetTimeslice is incorrect size"); struct IoctlEventIdControl { - u32_le cmd; // 0=disable, 1=enable, 2=clear - u32_le id; + u32_le cmd{}; // 0=disable, 1=enable, 2=clear + u32_le id{}; }; static_assert(sizeof(IoctlEventIdControl) == 8, "IoctlEventIdControl is incorrect size"); struct IoctlGetErrorNotification { - u64_le timestamp; - u32_le info32; - u16_le info16; - u16_le status; // always 0xFFFF + u64_le timestamp{}; + u32_le info32{}; + u16_le info16{}; + u16_le status{}; // always 0xFFFF }; static_assert(sizeof(IoctlGetErrorNotification) == 16, "IoctlGetErrorNotification is incorrect size"); @@ -125,80 +107,89 @@ private: static_assert(sizeof(Fence) == 8, "Fence is incorrect size"); struct IoctlAllocGpfifoEx { - u32_le num_entries; - u32_le flags; - u32_le unk0; - u32_le unk1; - u32_le unk2; - u32_le unk3; - u32_le unk4; - u32_le unk5; + u32_le num_entries{}; + u32_le flags{}; + u32_le unk0{}; + u32_le unk1{}; + u32_le unk2{}; + u32_le unk3{}; + u32_le unk4{}; + u32_le unk5{}; }; static_assert(sizeof(IoctlAllocGpfifoEx) == 32, "IoctlAllocGpfifoEx is incorrect size"); struct IoctlAllocGpfifoEx2 { - u32_le num_entries; // in - u32_le flags; // in - u32_le unk0; // in (1 works) - Fence fence_out; // out - u32_le unk1; // in - u32_le unk2; // in - u32_le unk3; // in + u32_le num_entries{}; // in + u32_le flags{}; // in + u32_le unk0{}; // in (1 works) + Fence fence_out{}; // out + u32_le unk1{}; // in + u32_le unk2{}; // in + u32_le unk3{}; // in }; static_assert(sizeof(IoctlAllocGpfifoEx2) == 32, "IoctlAllocGpfifoEx2 is incorrect size"); struct IoctlAllocObjCtx { - u32_le class_num; // 0x902D=2d, 0xB197=3d, 0xB1C0=compute, 0xA140=kepler, 0xB0B5=DMA, - // 0xB06F=channel_gpfifo - u32_le flags; - u64_le obj_id; // (ignored) used for FREE_OBJ_CTX ioctl, which is not supported + u32_le class_num{}; // 0x902D=2d, 0xB197=3d, 0xB1C0=compute, 0xA140=kepler, 0xB0B5=DMA, + // 0xB06F=channel_gpfifo + u32_le flags{}; + u64_le obj_id{}; // (ignored) used for FREE_OBJ_CTX ioctl, which is not supported }; static_assert(sizeof(IoctlAllocObjCtx) == 16, "IoctlAllocObjCtx is incorrect size"); struct IoctlSubmitGpfifo { - u64_le address; // pointer to gpfifo entry structs - u32_le num_entries; // number of fence objects being submitted + u64_le address{}; // pointer to gpfifo entry structs + u32_le num_entries{}; // number of fence objects being submitted union { u32_le raw; BitField<0, 1, u32_le> add_wait; // append a wait sync_point to the list BitField<1, 1, u32_le> add_increment; // append an increment to the list - BitField<2, 1, u32_le> new_hw_format; // Mostly ignored + BitField<2, 1, u32_le> new_hw_format; // mostly ignored + BitField<4, 1, u32_le> suppress_wfi; // suppress wait for interrupt BitField<8, 1, u32_le> increment; // increment the returned fence } flags; - Fence fence_out; // returned new fence object for others to wait on + Fence fence_out{}; // returned new fence object for others to wait on + + u32 AddIncrementValue() const { + return flags.add_increment.Value() << 1; + } }; static_assert(sizeof(IoctlSubmitGpfifo) == 16 + sizeof(Fence), "IoctlSubmitGpfifo is incorrect size"); struct IoctlGetWaitbase { - u32 unknown; // seems to be ignored? Nintendo added this - u32 value; + u32 unknown{}; // seems to be ignored? Nintendo added this + u32 value{}; }; static_assert(sizeof(IoctlGetWaitbase) == 8, "IoctlGetWaitbase is incorrect size"); - u32_le nvmap_fd{}; + s32_le nvmap_fd{}; u64_le user_data{}; IoctlZCullBind zcull_params{}; u32_le channel_priority{}; u32_le channel_timeslice{}; - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); - u32 SetClientData(const std::vector<u8>& input, std::vector<u8>& output); - u32 GetClientData(const std::vector<u8>& input, std::vector<u8>& output); - u32 ZCullBind(const std::vector<u8>& input, std::vector<u8>& output); - u32 SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output); - u32 SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output); - u32 AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output); - u32 AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output); - u32 SubmitGPFIFO(const std::vector<u8>& input, std::vector<u8>& output); - u32 KickoffPB(const std::vector<u8>& input, std::vector<u8>& output, - const std::vector<u8>& input2, IoctlVersion version); - u32 GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output); - u32 ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output); - u32 ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetClientData(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetClientData(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ZCullBind(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetErrorNotifier(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetChannelPriority(const std::vector<u8>& input, std::vector<u8>& output); + NvResult AllocGPFIFOEx2(const std::vector<u8>& input, std::vector<u8>& output); + NvResult AllocateObjectContext(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SubmitGPFIFOImpl(IoctlSubmitGpfifo& params, std::vector<u8>& output, + Tegra::CommandList&& entries); + NvResult SubmitGPFIFOBase(const std::vector<u8>& input, std::vector<u8>& output, + bool kickoff = false); + NvResult SubmitGPFIFOBase(const std::vector<u8>& input, const std::vector<u8>& input_inline, + std::vector<u8>& output); + NvResult GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ChannelSetTimeout(const std::vector<u8>& input, std::vector<u8>& output); + NvResult ChannelSetTimeslice(const std::vector<u8>& input, std::vector<u8>& output); std::shared_ptr<nvmap> nvmap_dev; - u32 assigned_syncpoints{}; + SyncpointManager& syncpoint_manager; + Fence channel_fence; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp index bdae8b887..36970f828 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp @@ -2,39 +2,72 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> - #include "common/assert.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/service/nvdrv/devices/nvhost_nvdec.h" +#include "video_core/memory_manager.h" +#include "video_core/renderer_base.h" namespace Service::Nvidia::Devices { -nvhost_nvdec::nvhost_nvdec(Core::System& system) : nvdevice(system) {} +nvhost_nvdec::nvhost_nvdec(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager) + : nvhost_nvdec_common(system, std::move(nvmap_dev), syncpoint_manager) {} nvhost_nvdec::~nvhost_nvdec() = default; -u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); +NvResult nvhost_nvdec::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x1: + return Submit(input, output); + case 0x2: + return GetSyncpoint(input, output); + case 0x3: + return GetWaitbase(input, output); + case 0x7: + return SetSubmitTimeout(input, output); + case 0x9: + return MapBuffer(input, output); + case 0xa: { + if (command.length == 0x1c) { + LOG_INFO(Service_NVDRV, "NVDEC video stream ended"); + Tegra::ChCommandHeaderList cmdlist(1); + cmdlist[0] = Tegra::ChCommandHeader{0xDEADB33F}; + system.GPU().PushCommandBuffer(cmdlist); + } + return UnmapBuffer(input, output); + } + default: + break; + } + break; + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input); + default: + break; + } + break; } - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_nvdec::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { - IoctlSetNvmapFD params{}; - std::memcpy(¶ms, input.data(), input.size()); - LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); +NvResult nvhost_nvdec::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - nvmap_fd = params.nvmap_fd; - return 0; +NvResult nvhost_nvdec::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h index cbdac8069..77ef53cdd 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.h @@ -4,35 +4,22 @@ #pragma once -#include <vector> -#include "common/common_types.h" -#include "common/swap.h" -#include "core/hle/service/nvdrv/devices/nvdevice.h" +#include <memory> +#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h" namespace Service::Nvidia::Devices { -class nvhost_nvdec final : public nvdevice { +class nvhost_nvdec final : public nvhost_nvdec_common { public: - explicit nvhost_nvdec(Core::System& system); + explicit nvhost_nvdec(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager); ~nvhost_nvdec() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; - -private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - }; - - struct IoctlSetNvmapFD { - u32_le nvmap_fd; - }; - static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); - - u32_le nvmap_fd{}; - - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp new file mode 100644 index 000000000..4898dc27a --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.cpp @@ -0,0 +1,244 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <cstring> + +#include "common/assert.h" +#include "common/common_types.h" +#include "common/logging/log.h" +#include "core/core.h" +#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h" +#include "core/hle/service/nvdrv/devices/nvmap.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" +#include "core/memory.h" +#include "video_core/memory_manager.h" +#include "video_core/renderer_base.h" + +namespace Service::Nvidia::Devices { + +namespace { +// Splice vectors will copy count amount of type T from the input vector into the dst vector. +template <typename T> +std::size_t SpliceVectors(const std::vector<u8>& input, std::vector<T>& dst, std::size_t count, + std::size_t offset) { + std::memcpy(dst.data(), input.data() + offset, count * sizeof(T)); + offset += count * sizeof(T); + return offset; +} + +// Write vectors will write data to the output buffer +template <typename T> +std::size_t WriteVectors(std::vector<u8>& dst, const std::vector<T>& src, std::size_t offset) { + std::memcpy(dst.data() + offset, src.data(), src.size() * sizeof(T)); + offset += src.size() * sizeof(T); + return offset; +} +} // Anonymous namespace + +nvhost_nvdec_common::nvhost_nvdec_common(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager) + : nvdevice(system), nvmap_dev(std::move(nvmap_dev)), syncpoint_manager(syncpoint_manager) {} +nvhost_nvdec_common::~nvhost_nvdec_common() = default; + +NvResult nvhost_nvdec_common::SetNVMAPfd(const std::vector<u8>& input) { + IoctlSetNvmapFD params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlSetNvmapFD)); + LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); + + nvmap_fd = params.nvmap_fd; + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::Submit(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlSubmit params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlSubmit)); + LOG_DEBUG(Service_NVDRV, "called NVDEC Submit, cmd_buffer_count={}", params.cmd_buffer_count); + + // Instantiate param buffers + std::size_t offset = sizeof(IoctlSubmit); + std::vector<CommandBuffer> command_buffers(params.cmd_buffer_count); + std::vector<Reloc> relocs(params.relocation_count); + std::vector<u32> reloc_shifts(params.relocation_count); + std::vector<SyncptIncr> syncpt_increments(params.syncpoint_count); + std::vector<SyncptIncr> wait_checks(params.syncpoint_count); + std::vector<Fence> fences(params.fence_count); + + // Splice input into their respective buffers + offset = SpliceVectors(input, command_buffers, params.cmd_buffer_count, offset); + offset = SpliceVectors(input, relocs, params.relocation_count, offset); + offset = SpliceVectors(input, reloc_shifts, params.relocation_count, offset); + offset = SpliceVectors(input, syncpt_increments, params.syncpoint_count, offset); + offset = SpliceVectors(input, wait_checks, params.syncpoint_count, offset); + offset = SpliceVectors(input, fences, params.fence_count, offset); + + auto& gpu = system.GPU(); + if (gpu.UseNvdec()) { + for (std::size_t i = 0; i < syncpt_increments.size(); i++) { + const SyncptIncr& syncpt_incr = syncpt_increments[i]; + fences[i].id = syncpt_incr.id; + fences[i].value = + syncpoint_manager.IncreaseSyncpoint(syncpt_incr.id, syncpt_incr.increments); + } + } + for (const auto& cmd_buffer : command_buffers) { + auto object = nvmap_dev->GetObject(cmd_buffer.memory_id); + ASSERT_OR_EXECUTE(object, return NvResult::InvalidState;); + const auto map = FindBufferMap(object->dma_map_addr); + if (!map) { + LOG_ERROR(Service_NVDRV, "Tried to submit an invalid offset 0x{:X} dma 0x{:X}", + object->addr, object->dma_map_addr); + return NvResult::Success; + } + Tegra::ChCommandHeaderList cmdlist(cmd_buffer.word_count); + gpu.MemoryManager().ReadBlock(map->StartAddr() + cmd_buffer.offset, cmdlist.data(), + cmdlist.size() * sizeof(u32)); + gpu.PushCommandBuffer(cmdlist); + } + if (gpu.UseNvdec()) { + + fences[0].value = syncpoint_manager.IncreaseSyncpoint(fences[0].id, 1); + + Tegra::ChCommandHeaderList cmdlist{{(4 << 28) | fences[0].id}}; + gpu.PushCommandBuffer(cmdlist); + } + std::memcpy(output.data(), ¶ms, sizeof(IoctlSubmit)); + // Some games expect command_buffers to be written back + offset = sizeof(IoctlSubmit); + offset = WriteVectors(output, command_buffers, offset); + offset = WriteVectors(output, relocs, offset); + offset = WriteVectors(output, reloc_shifts, offset); + offset = WriteVectors(output, syncpt_increments, offset); + offset = WriteVectors(output, wait_checks, offset); + offset = WriteVectors(output, fences, offset); + + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlGetSyncpoint params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlGetSyncpoint)); + LOG_DEBUG(Service_NVDRV, "called GetSyncpoint, id={}", params.param); + + if (device_syncpoints[params.param] == 0 && system.GPU().UseNvdec()) { + device_syncpoints[params.param] = syncpoint_manager.AllocateSyncpoint(); + } + params.value = device_syncpoints[params.param]; + std::memcpy(output.data(), ¶ms, sizeof(IoctlGetSyncpoint)); + + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlGetWaitbase params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlGetWaitbase)); + params.value = 0; // Seems to be hard coded at 0 + std::memcpy(output.data(), ¶ms, sizeof(IoctlGetWaitbase)); + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::MapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlMapBuffer params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlMapBuffer)); + std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries); + + SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer)); + + auto& gpu = system.GPU(); + + for (auto& cmf_buff : cmd_buffer_handles) { + auto object{nvmap_dev->GetObject(cmf_buff.map_handle)}; + if (!object) { + LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle); + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::InvalidState; + } + if (object->dma_map_addr == 0) { + // NVDEC and VIC memory is in the 32-bit address space + // MapAllocate32 will attempt to map a lower 32-bit value in the shared gpu memory space + const GPUVAddr low_addr = gpu.MemoryManager().MapAllocate32(object->addr, object->size); + object->dma_map_addr = static_cast<u32>(low_addr); + // Ensure that the dma_map_addr is indeed in the lower 32-bit address space. + ASSERT(object->dma_map_addr == low_addr); + } + if (!object->dma_map_addr) { + LOG_ERROR(Service_NVDRV, "failed to map size={}", object->size); + } else { + cmf_buff.map_address = object->dma_map_addr; + AddBufferMap(object->dma_map_addr, object->size, object->addr, + object->status == nvmap::Object::Status::Allocated); + } + } + std::memcpy(output.data(), ¶ms, sizeof(IoctlMapBuffer)); + std::memcpy(output.data() + sizeof(IoctlMapBuffer), cmd_buffer_handles.data(), + cmd_buffer_handles.size() * sizeof(MapBufferEntry)); + + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output) { + IoctlMapBuffer params{}; + std::memcpy(¶ms, input.data(), sizeof(IoctlMapBuffer)); + std::vector<MapBufferEntry> cmd_buffer_handles(params.num_entries); + SpliceVectors(input, cmd_buffer_handles, params.num_entries, sizeof(IoctlMapBuffer)); + + auto& gpu = system.GPU(); + + for (auto& cmf_buff : cmd_buffer_handles) { + const auto object{nvmap_dev->GetObject(cmf_buff.map_handle)}; + if (!object) { + LOG_ERROR(Service_NVDRV, "invalid cmd_buffer nvmap_handle={:X}", cmf_buff.map_handle); + std::memcpy(output.data(), ¶ms, output.size()); + return NvResult::InvalidState; + } + if (const auto size{RemoveBufferMap(object->dma_map_addr)}; size) { + gpu.MemoryManager().Unmap(object->dma_map_addr, *size); + } else { + // This occurs quite frequently, however does not seem to impact functionality + LOG_DEBUG(Service_NVDRV, "invalid offset=0x{:X} dma=0x{:X}", object->addr, + object->dma_map_addr); + } + object->dma_map_addr = 0; + } + std::memset(output.data(), 0, output.size()); + return NvResult::Success; +} + +NvResult nvhost_nvdec_common::SetSubmitTimeout(const std::vector<u8>& input, + std::vector<u8>& output) { + std::memcpy(&submit_timeout, input.data(), input.size()); + LOG_WARNING(Service_NVDRV, "(STUBBED) called"); + return NvResult::Success; +} + +std::optional<nvhost_nvdec_common::BufferMap> nvhost_nvdec_common::FindBufferMap( + GPUVAddr gpu_addr) const { + const auto it = std::find_if( + buffer_mappings.begin(), buffer_mappings.upper_bound(gpu_addr), [&](const auto& entry) { + return (gpu_addr >= entry.second.StartAddr() && gpu_addr < entry.second.EndAddr()); + }); + + ASSERT(it != buffer_mappings.end()); + return it->second; +} + +void nvhost_nvdec_common::AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, + bool is_allocated) { + buffer_mappings.insert_or_assign(gpu_addr, BufferMap{gpu_addr, size, cpu_addr, is_allocated}); +} + +std::optional<std::size_t> nvhost_nvdec_common::RemoveBufferMap(GPUVAddr gpu_addr) { + const auto iter{buffer_mappings.find(gpu_addr)}; + if (iter == buffer_mappings.end()) { + return std::nullopt; + } + std::size_t size = 0; + if (iter->second.IsAllocated()) { + size = iter->second.Size(); + } + buffer_mappings.erase(iter); + return size; +} + +} // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h new file mode 100644 index 000000000..4c9d4ba41 --- /dev/null +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec_common.h @@ -0,0 +1,170 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <map> +#include <vector> +#include "common/common_types.h" +#include "common/swap.h" +#include "core/hle/service/nvdrv/devices/nvdevice.h" + +namespace Service::Nvidia { +class SyncpointManager; + +namespace Devices { +class nvmap; + +class nvhost_nvdec_common : public nvdevice { +public: + explicit nvhost_nvdec_common(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager); + ~nvhost_nvdec_common() override; + +protected: + class BufferMap final { + public: + constexpr BufferMap() = default; + + constexpr BufferMap(GPUVAddr start_addr, std::size_t size) + : start_addr{start_addr}, end_addr{start_addr + size} {} + + constexpr BufferMap(GPUVAddr start_addr, std::size_t size, VAddr cpu_addr, + bool is_allocated) + : start_addr{start_addr}, end_addr{start_addr + size}, cpu_addr{cpu_addr}, + is_allocated{is_allocated} {} + + constexpr VAddr StartAddr() const { + return start_addr; + } + + constexpr VAddr EndAddr() const { + return end_addr; + } + + constexpr std::size_t Size() const { + return end_addr - start_addr; + } + + constexpr VAddr CpuAddr() const { + return cpu_addr; + } + + constexpr bool IsAllocated() const { + return is_allocated; + } + + private: + GPUVAddr start_addr{}; + GPUVAddr end_addr{}; + VAddr cpu_addr{}; + bool is_allocated{}; + }; + + struct IoctlSetNvmapFD { + s32_le nvmap_fd{}; + }; + static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); + + struct IoctlSubmitCommandBuffer { + u32_le id{}; + u32_le offset{}; + u32_le count{}; + }; + static_assert(sizeof(IoctlSubmitCommandBuffer) == 0xC, + "IoctlSubmitCommandBuffer is incorrect size"); + struct IoctlSubmit { + u32_le cmd_buffer_count{}; + u32_le relocation_count{}; + u32_le syncpoint_count{}; + u32_le fence_count{}; + }; + static_assert(sizeof(IoctlSubmit) == 0x10, "IoctlSubmit has incorrect size"); + + struct CommandBuffer { + s32 memory_id{}; + u32 offset{}; + s32 word_count{}; + }; + static_assert(sizeof(CommandBuffer) == 0xC, "CommandBuffer has incorrect size"); + + struct Reloc { + s32 cmdbuffer_memory{}; + s32 cmdbuffer_offset{}; + s32 target{}; + s32 target_offset{}; + }; + static_assert(sizeof(Reloc) == 0x10, "CommandBuffer has incorrect size"); + + struct SyncptIncr { + u32 id{}; + u32 increments{}; + }; + static_assert(sizeof(SyncptIncr) == 0x8, "CommandBuffer has incorrect size"); + + struct Fence { + u32 id{}; + u32 value{}; + }; + static_assert(sizeof(Fence) == 0x8, "CommandBuffer has incorrect size"); + + struct IoctlGetSyncpoint { + // Input + u32_le param{}; + // Output + u32_le value{}; + }; + static_assert(sizeof(IoctlGetSyncpoint) == 8, "IocGetIdParams has wrong size"); + + struct IoctlGetWaitbase { + u32_le unknown{}; // seems to be ignored? Nintendo added this + u32_le value{}; + }; + static_assert(sizeof(IoctlGetWaitbase) == 0x8, "IoctlGetWaitbase is incorrect size"); + + struct IoctlMapBuffer { + u32_le num_entries{}; + u32_le data_address{}; // Ignored by the driver. + u32_le attach_host_ch_das{}; + }; + static_assert(sizeof(IoctlMapBuffer) == 0x0C, "IoctlMapBuffer is incorrect size"); + + struct IocGetIdParams { + // Input + u32_le param{}; + // Output + u32_le value{}; + }; + static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size"); + + // Used for mapping and unmapping command buffers + struct MapBufferEntry { + u32_le map_handle{}; + u32_le map_address{}; + }; + static_assert(sizeof(IoctlMapBuffer) == 0x0C, "IoctlMapBuffer is incorrect size"); + + /// Ioctl command implementations + NvResult SetNVMAPfd(const std::vector<u8>& input); + NvResult Submit(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetSyncpoint(const std::vector<u8>& input, std::vector<u8>& output); + NvResult GetWaitbase(const std::vector<u8>& input, std::vector<u8>& output); + NvResult MapBuffer(const std::vector<u8>& input, std::vector<u8>& output); + NvResult UnmapBuffer(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetSubmitTimeout(const std::vector<u8>& input, std::vector<u8>& output); + + std::optional<BufferMap> FindBufferMap(GPUVAddr gpu_addr) const; + void AddBufferMap(GPUVAddr gpu_addr, std::size_t size, VAddr cpu_addr, bool is_allocated); + std::optional<std::size_t> RemoveBufferMap(GPUVAddr gpu_addr); + + s32_le nvmap_fd{}; + u32_le submit_timeout{}; + std::shared_ptr<nvmap> nvmap_dev; + SyncpointManager& syncpoint_manager; + std::array<u32, MaxSyncPoints> device_syncpoints{}; + // This is expected to be ordered, therefore we must use a map, not unordered_map + std::map<GPUVAddr, BufferMap> buffer_mappings; +}; +}; // namespace Devices +} // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp index 96e7b7dab..2d06955c0 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.cpp @@ -13,28 +13,44 @@ namespace Service::Nvidia::Devices { nvhost_nvjpg::nvhost_nvjpg(Core::System& system) : nvdevice(system) {} nvhost_nvjpg::~nvhost_nvjpg() = default; -u32 nvhost_nvjpg::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); +NvResult nvhost_nvjpg::Ioctl1(Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + switch (command.group) { + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input, output); + default: + break; + } + break; + default: + break; } - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_nvjpg::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvhost_nvjpg::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_nvjpg::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvhost_nvjpg::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { IoctlSetNvmapFD params{}; std::memcpy(¶ms, input.data(), input.size()); LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); nvmap_fd = params.nvmap_fd; - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h index 98dcac52f..43948d18d 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_nvjpg.h @@ -16,23 +16,21 @@ public: explicit nvhost_nvjpg(Core::System& system); ~nvhost_nvjpg() override; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - }; - struct IoctlSetNvmapFD { - u32_le nvmap_fd; + s32_le nvmap_fd{}; }; static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); - u32_le nvmap_fd{}; + s32_le nvmap_fd{}; - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + NvResult SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp index c695b8863..72499654c 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp @@ -2,39 +2,64 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cstring> - #include "common/assert.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/service/nvdrv/devices/nvhost_vic.h" +#include "video_core/memory_manager.h" +#include "video_core/renderer_base.h" namespace Service::Nvidia::Devices { +nvhost_vic::nvhost_vic(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager) + : nvhost_nvdec_common(system, std::move(nvmap_dev), syncpoint_manager) {} -nvhost_vic::nvhost_vic(Core::System& system) : nvdevice(system) {} nvhost_vic::~nvhost_vic() = default; -u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - LOG_DEBUG(Service_NVDRV, "called, command=0x{:08X}, input_size=0x{:X}, output_size=0x{:X}", - command.raw, input.size(), output.size()); - - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::IocSetNVMAPfdCommand: - return SetNVMAPfd(input, output); +NvResult nvhost_vic::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x0: + switch (command.cmd) { + case 0x1: + return Submit(input, output); + case 0x2: + return GetSyncpoint(input, output); + case 0x3: + return GetWaitbase(input, output); + case 0x9: + return MapBuffer(input, output); + case 0xa: + return UnmapBuffer(input, output); + default: + break; + } + break; + case 'H': + switch (command.cmd) { + case 0x1: + return SetNVMAPfd(input); + default: + break; + } + break; + default: + break; } - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } -u32 nvhost_vic::SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output) { - IoctlSetNvmapFD params{}; - std::memcpy(¶ms, input.data(), input.size()); - LOG_DEBUG(Service_NVDRV, "called, fd={}", params.nvmap_fd); +NvResult nvhost_vic::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} - nvmap_fd = params.nvmap_fd; - return 0; +NvResult nvhost_vic::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.h b/src/core/hle/service/nvdrv/devices/nvhost_vic.h index bec32bea1..f401c61fa 100644 --- a/src/core/hle/service/nvdrv/devices/nvhost_vic.h +++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.h @@ -4,35 +4,20 @@ #pragma once -#include <vector> -#include "common/common_types.h" -#include "common/swap.h" -#include "core/hle/service/nvdrv/devices/nvdevice.h" +#include "core/hle/service/nvdrv/devices/nvhost_nvdec_common.h" namespace Service::Nvidia::Devices { -class nvhost_vic final : public nvdevice { +class nvhost_vic final : public nvhost_nvdec_common { public: - explicit nvhost_vic(Core::System& system); - ~nvhost_vic() override; - - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; - -private: - enum class IoctlCommand : u32_le { - IocSetNVMAPfdCommand = 0x40044801, - }; - - struct IoctlSetNvmapFD { - u32_le nvmap_fd; - }; - static_assert(sizeof(IoctlSetNvmapFD) == 4, "IoctlSetNvmapFD is incorrect size"); - - u32_le nvmap_fd{}; - - u32 SetNVMAPfd(const std::vector<u8>& input, std::vector<u8>& output); + explicit nvhost_vic(Core::System& system, std::shared_ptr<nvmap> nvmap_dev, + SyncpointManager& syncpoint_manager); + ~nvhost_vic(); + + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; }; - } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvmap.cpp b/src/core/hle/service/nvdrv/devices/nvmap.cpp index 9436e16ad..4015a2740 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.cpp +++ b/src/core/hle/service/nvdrv/devices/nvmap.cpp @@ -11,13 +11,6 @@ namespace Service::Nvidia::Devices { -namespace NvErrCodes { -enum { - OperationNotPermitted = -1, - InvalidValue = -22, -}; -} - nvmap::nvmap(Core::System& system) : nvdevice(system) { // Handle 0 appears to be used when remapping, so we create a placeholder empty nvmap object to // represent this. @@ -26,6 +19,46 @@ nvmap::nvmap(Core::System& system) : nvdevice(system) { nvmap::~nvmap() = default; +NvResult nvmap::Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) { + switch (command.group) { + case 0x1: + switch (command.cmd) { + case 0x1: + return IocCreate(input, output); + case 0x3: + return IocFromId(input, output); + case 0x4: + return IocAlloc(input, output); + case 0x5: + return IocFree(input, output); + case 0x9: + return IocParam(input, output); + case 0xe: + return IocGetId(input, output); + default: + break; + } + break; + default: + break; + } + + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvmap::Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + +NvResult nvmap::Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) { + UNIMPLEMENTED_MSG("Unimplemented ioctl={:08X}", command.raw); + return NvResult::NotImplemented; +} + VAddr nvmap::GetObjectAddress(u32 handle) const { auto object = GetObject(handle); ASSERT(object); @@ -33,28 +66,6 @@ VAddr nvmap::GetObjectAddress(u32 handle) const { return object->addr; } -u32 nvmap::ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - switch (static_cast<IoctlCommand>(command.raw)) { - case IoctlCommand::Create: - return IocCreate(input, output); - case IoctlCommand::Alloc: - return IocAlloc(input, output); - case IoctlCommand::GetId: - return IocGetId(input, output); - case IoctlCommand::FromId: - return IocFromId(input, output); - case IoctlCommand::Param: - return IocParam(input, output); - case IoctlCommand::Free: - return IocFree(input, output); - } - - UNIMPLEMENTED_MSG("Unimplemented ioctl"); - return 0; -} - u32 nvmap::CreateObject(u32 size) { // Create a new nvmap object and obtain a handle to it. auto object = std::make_shared<Object>(); @@ -70,35 +81,35 @@ u32 nvmap::CreateObject(u32 size) { return handle; } -u32 nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocCreate(const std::vector<u8>& input, std::vector<u8>& output) { IocCreateParams params; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "size=0x{:08X}", params.size); if (!params.size) { LOG_ERROR(Service_NVDRV, "Size is 0"); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } params.handle = CreateObject(params.size); std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { IocAllocParams params; std::memcpy(¶ms, input.data(), sizeof(params)); LOG_DEBUG(Service_NVDRV, "called, addr={:X}", params.addr); if (!params.handle) { LOG_ERROR(Service_NVDRV, "Handle is 0"); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if ((params.align - 1) & params.align) { LOG_ERROR(Service_NVDRV, "Incorrect alignment used, alignment={:08X}", params.align); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } const u32 min_alignment = 0x1000; @@ -109,12 +120,12 @@ u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { auto object = GetObject(params.handle); if (!object) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if (object->status == Object::Status::Allocated) { LOG_ERROR(Service_NVDRV, "Object is already allocated, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::OperationNotPermitted); + return NvResult::InsufficientMemory; } object->flags = params.flags; @@ -124,10 +135,10 @@ u32 nvmap::IocAlloc(const std::vector<u8>& input, std::vector<u8>& output) { object->status = Object::Status::Allocated; std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) { IocGetIdParams params; std::memcpy(¶ms, input.data(), sizeof(params)); @@ -135,22 +146,22 @@ u32 nvmap::IocGetId(const std::vector<u8>& input, std::vector<u8>& output) { if (!params.handle) { LOG_ERROR(Service_NVDRV, "Handle is zero"); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } auto object = GetObject(params.handle); if (!object) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::OperationNotPermitted); + return NvResult::BadValue; } params.id = object->id; std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { IocFromIdParams params; std::memcpy(¶ms, input.data(), sizeof(params)); @@ -160,13 +171,13 @@ u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { [&](const auto& entry) { return entry.second->id == params.id; }); if (itr == handles.end()) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } auto& object = itr->second; if (object->status != Object::Status::Allocated) { LOG_ERROR(Service_NVDRV, "Object is not allocated, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } itr->second->refcount++; @@ -175,10 +186,10 @@ u32 nvmap::IocFromId(const std::vector<u8>& input, std::vector<u8>& output) { params.handle = itr->first; std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { enum class ParamTypes { Size = 1, Alignment = 2, Base = 3, Heap = 4, Kind = 5, Compr = 6 }; IocParamParams params; @@ -189,12 +200,12 @@ u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { auto object = GetObject(params.handle); if (!object) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if (object->status != Object::Status::Allocated) { LOG_ERROR(Service_NVDRV, "Object is not allocated, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::OperationNotPermitted); + return NvResult::BadValue; } switch (static_cast<ParamTypes>(params.param)) { @@ -216,10 +227,10 @@ u32 nvmap::IocParam(const std::vector<u8>& input, std::vector<u8>& output) { } std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } -u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { +NvResult nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { // TODO(Subv): These flags are unconfirmed. enum FreeFlags { Freed = 0, @@ -234,14 +245,14 @@ u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { auto itr = handles.find(params.handle); if (itr == handles.end()) { LOG_ERROR(Service_NVDRV, "Object does not exist, handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } if (!itr->second->refcount) { LOG_ERROR( Service_NVDRV, "There is no references to this object. The object is already freed. handle={:08X}", params.handle); - return static_cast<u32>(NvErrCodes::InvalidValue); + return NvResult::BadValue; } itr->second->refcount--; @@ -261,7 +272,7 @@ u32 nvmap::IocFree(const std::vector<u8>& input, std::vector<u8>& output) { handles.erase(params.handle); std::memcpy(output.data(), ¶ms, sizeof(params)); - return 0; + return NvResult::Success; } } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/devices/nvmap.h b/src/core/hle/service/nvdrv/devices/nvmap.h index 84624be00..4484bd79f 100644 --- a/src/core/hle/service/nvdrv/devices/nvmap.h +++ b/src/core/hle/service/nvdrv/devices/nvmap.h @@ -19,13 +19,15 @@ public: explicit nvmap(Core::System& system); ~nvmap() override; + NvResult Ioctl1(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; + NvResult Ioctl2(Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) override; + NvResult Ioctl3(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output, + std::vector<u8>& inline_output) override; + /// Returns the allocated address of an nvmap object given its handle. VAddr GetObjectAddress(u32 handle) const; - u32 ioctl(Ioctl command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) override; - /// Represents an nvmap object. struct Object { enum class Status { Created, Allocated }; @@ -37,6 +39,7 @@ public: VAddr addr; Status status; u32 refcount; + u32 dma_map_addr; }; std::shared_ptr<Object> GetObject(u32 handle) const { @@ -57,76 +60,68 @@ private: /// Mapping of currently allocated handles to the objects they represent. std::unordered_map<u32, std::shared_ptr<Object>> handles; - enum class IoctlCommand : u32 { - Create = 0xC0080101, - FromId = 0xC0080103, - Alloc = 0xC0200104, - Free = 0xC0180105, - Param = 0xC00C0109, - GetId = 0xC008010E, - }; struct IocCreateParams { // Input - u32_le size; + u32_le size{}; // Output - u32_le handle; + u32_le handle{}; }; static_assert(sizeof(IocCreateParams) == 8, "IocCreateParams has wrong size"); struct IocFromIdParams { // Input - u32_le id; + u32_le id{}; // Output - u32_le handle; + u32_le handle{}; }; static_assert(sizeof(IocFromIdParams) == 8, "IocFromIdParams has wrong size"); struct IocAllocParams { // Input - u32_le handle; - u32_le heap_mask; - u32_le flags; - u32_le align; - u8 kind; + u32_le handle{}; + u32_le heap_mask{}; + u32_le flags{}; + u32_le align{}; + u8 kind{}; INSERT_PADDING_BYTES(7); - u64_le addr; + u64_le addr{}; }; static_assert(sizeof(IocAllocParams) == 32, "IocAllocParams has wrong size"); struct IocFreeParams { - u32_le handle; + u32_le handle{}; INSERT_PADDING_BYTES(4); - u64_le address; - u32_le size; - u32_le flags; + u64_le address{}; + u32_le size{}; + u32_le flags{}; }; static_assert(sizeof(IocFreeParams) == 24, "IocFreeParams has wrong size"); struct IocParamParams { // Input - u32_le handle; - u32_le param; + u32_le handle{}; + u32_le param{}; // Output - u32_le result; + u32_le result{}; }; static_assert(sizeof(IocParamParams) == 12, "IocParamParams has wrong size"); struct IocGetIdParams { // Output - u32_le id; + u32_le id{}; // Input - u32_le handle; + u32_le handle{}; }; static_assert(sizeof(IocGetIdParams) == 8, "IocGetIdParams has wrong size"); u32 CreateObject(u32 size); - u32 IocCreate(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocAlloc(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocGetId(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocFromId(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocParam(const std::vector<u8>& input, std::vector<u8>& output); - u32 IocFree(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocCreate(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocAlloc(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocGetId(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocFromId(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocParam(const std::vector<u8>& input, std::vector<u8>& output); + NvResult IocFree(const std::vector<u8>& input, std::vector<u8>& output); }; } // namespace Service::Nvidia::Devices diff --git a/src/core/hle/service/nvdrv/interface.cpp b/src/core/hle/service/nvdrv/interface.cpp index 88fbfa9b0..cc23b001c 100644 --- a/src/core/hle/service/nvdrv/interface.cpp +++ b/src/core/hle/service/nvdrv/interface.cpp @@ -23,124 +23,167 @@ void NVDRV::SignalGPUInterruptSyncpt(const u32 syncpoint_id, const u32 value) { void NVDRV::Open(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NVDRV, "called"); + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + const auto& buffer = ctx.ReadBuffer(); - std::string device_name(buffer.begin(), buffer.end()); + const std::string device_name(buffer.begin(), buffer.end()); + DeviceFD fd = nvdrv->Open(device_name); - u32 fd = nvdrv->Open(device_name); IPC::ResponseBuilder rb{ctx, 4}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(fd); - rb.Push<u32>(0); + rb.Push<DeviceFD>(fd); + rb.PushEnum(fd != INVALID_NVDRV_FD ? NvResult::Success : NvResult::FileOperationFailed); +} + +void NVDRV::ServiceError(Kernel::HLERequestContext& ctx, NvResult result) { + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(result); } -void NVDRV::IoctlBase(Kernel::HLERequestContext& ctx, IoctlVersion version) { +void NVDRV::Ioctl1(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - u32 fd = rp.Pop<u32>(); - u32 command = rp.Pop<u32>(); - - /// Ioctl 3 has 2 outputs, first in the input params, second is the result - std::vector<u8> output(ctx.GetWriteBufferSize(0)); - std::vector<u8> output2; - if (version == IoctlVersion::Version3) { - output2.resize((ctx.GetWriteBufferSize(1))); + const auto fd = rp.Pop<DeviceFD>(); + const auto command = rp.PopRaw<Ioctl>(); + LOG_DEBUG(Service_NVDRV, "called fd={}, ioctl=0x{:08X}", fd, command.raw); + + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; } - /// Ioctl2 has 2 inputs. It's used to pass data directly instead of providing a pointer. - /// KickOfPB uses this - auto input = ctx.ReadBuffer(0); + // Check device + std::vector<u8> output_buffer(ctx.GetWriteBufferSize(0)); + const auto input_buffer = ctx.ReadBuffer(0); - std::vector<u8> input2; - if (version == IoctlVersion::Version2) { - input2 = ctx.ReadBuffer(1); + const auto nv_result = nvdrv->Ioctl1(fd, command, input_buffer, output_buffer); + if (command.is_out != 0) { + ctx.WriteBuffer(output_buffer); } - IoctlCtrl ctrl{}; - - u32 result = nvdrv->Ioctl(fd, command, input, input2, output, output2, ctrl, version); - - if (ctrl.must_delay) { - ctrl.fresh_call = false; - ctx.SleepClientThread( - "NVServices::DelayedResponse", ctrl.timeout, - [=, this](std::shared_ptr<Kernel::Thread> thread, Kernel::HLERequestContext& ctx_, - Kernel::ThreadWakeupReason reason) { - IoctlCtrl ctrl2{ctrl}; - std::vector<u8> tmp_output = output; - std::vector<u8> tmp_output2 = output2; - const u32 ioctl_result = nvdrv->Ioctl(fd, command, input, input2, tmp_output, - tmp_output2, ctrl2, version); - ctx_.WriteBuffer(tmp_output, 0); - if (version == IoctlVersion::Version3) { - ctx_.WriteBuffer(tmp_output2, 1); - } - IPC::ResponseBuilder rb{ctx_, 3}; - rb.Push(RESULT_SUCCESS); - rb.Push(ioctl_result); - }, - nvdrv->GetEventWriteable(ctrl.event_id)); - } else { - ctx.WriteBuffer(output); - if (version == IoctlVersion::Version3) { - ctx.WriteBuffer(output2, 1); - } - } IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push(result); -} - -void NVDRV::Ioctl(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_NVDRV, "called"); - IoctlBase(ctx, IoctlVersion::Version1); + rb.PushEnum(nv_result); } void NVDRV::Ioctl2(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_NVDRV, "called"); - IoctlBase(ctx, IoctlVersion::Version2); + IPC::RequestParser rp{ctx}; + const auto fd = rp.Pop<DeviceFD>(); + const auto command = rp.PopRaw<Ioctl>(); + LOG_DEBUG(Service_NVDRV, "called fd={}, ioctl=0x{:08X}", fd, command.raw); + + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + + const auto input_buffer = ctx.ReadBuffer(0); + const auto input_inlined_buffer = ctx.ReadBuffer(1); + std::vector<u8> output_buffer(ctx.GetWriteBufferSize(0)); + + const auto nv_result = + nvdrv->Ioctl2(fd, command, input_buffer, input_inlined_buffer, output_buffer); + if (command.is_out != 0) { + ctx.WriteBuffer(output_buffer); + } + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(nv_result); } void NVDRV::Ioctl3(Kernel::HLERequestContext& ctx) { - LOG_DEBUG(Service_NVDRV, "called"); - IoctlBase(ctx, IoctlVersion::Version3); + IPC::RequestParser rp{ctx}; + const auto fd = rp.Pop<DeviceFD>(); + const auto command = rp.PopRaw<Ioctl>(); + LOG_DEBUG(Service_NVDRV, "called fd={}, ioctl=0x{:08X}", fd, command.raw); + + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + + const auto input_buffer = ctx.ReadBuffer(0); + std::vector<u8> output_buffer(ctx.GetWriteBufferSize(0)); + std::vector<u8> output_buffer_inline(ctx.GetWriteBufferSize(1)); + + const auto nv_result = + nvdrv->Ioctl3(fd, command, input_buffer, output_buffer, output_buffer_inline); + if (command.is_out != 0) { + ctx.WriteBuffer(output_buffer, 0); + ctx.WriteBuffer(output_buffer_inline, 1); + } + + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(nv_result); } void NVDRV::Close(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_NVDRV, "called"); - IPC::RequestParser rp{ctx}; - u32 fd = rp.Pop<u32>(); + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } - auto result = nvdrv->Close(fd); + IPC::RequestParser rp{ctx}; + const auto fd = rp.Pop<DeviceFD>(); + const auto result = nvdrv->Close(fd); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(result); + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(result); } void NVDRV::Initialize(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); + is_initialized = true; + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); + rb.PushEnum(NvResult::Success); } void NVDRV::QueryEvent(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - u32 fd = rp.Pop<u32>(); - // TODO(Blinkhawk): Figure the meaning of the flag at bit 16 - u32 event_id = rp.Pop<u32>() & 0x000000FF; + const auto fd = rp.Pop<DeviceFD>(); + const auto event_id = rp.Pop<u32>() & 0x00FF; LOG_WARNING(Service_NVDRV, "(STUBBED) called, fd={:X}, event_id={:X}", fd, event_id); - IPC::ResponseBuilder rb{ctx, 3, 1}; - rb.Push(RESULT_SUCCESS); + if (!is_initialized) { + ServiceError(ctx, NvResult::NotInitialized); + LOG_ERROR(Service_NVDRV, "NvServices is not initalized!"); + return; + } + + const auto nv_result = nvdrv->VerifyFD(fd); + if (nv_result != NvResult::Success) { + LOG_ERROR(Service_NVDRV, "Invalid FD specified DeviceFD={}!", fd); + ServiceError(ctx, nv_result); + return; + } + if (event_id < MaxNvEvents) { + IPC::ResponseBuilder rb{ctx, 3, 1}; + rb.Push(RESULT_SUCCESS); auto event = nvdrv->GetEvent(event_id); event->Clear(); rb.PushCopyObjects(event); - rb.Push<u32>(NvResult::Success); + rb.PushEnum(NvResult::Success); } else { - rb.Push<u32>(0); - rb.Push<u32>(NvResult::BadParameter); + IPC::ResponseBuilder rb{ctx, 3}; + rb.Push(RESULT_SUCCESS); + rb.PushEnum(NvResult::BadParameter); } } @@ -151,7 +194,7 @@ void NVDRV::SetAruid(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); - rb.Push<u32>(0); + rb.PushEnum(NvResult::Success); } void NVDRV::SetGraphicsFirmwareMemoryMarginEnabled(Kernel::HLERequestContext& ctx) { @@ -164,8 +207,9 @@ void NVDRV::SetGraphicsFirmwareMemoryMarginEnabled(Kernel::HLERequestContext& ct void NVDRV::GetStatus(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_NVDRV, "(STUBBED) called"); - IPC::ResponseBuilder rb{ctx, 2}; + IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); + rb.PushEnum(NvResult::Success); } void NVDRV::DumpGraphicsMemoryInfo(Kernel::HLERequestContext& ctx) { @@ -177,11 +221,11 @@ void NVDRV::DumpGraphicsMemoryInfo(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); } -NVDRV::NVDRV(std::shared_ptr<Module> nvdrv, const char* name) - : ServiceFramework(name), nvdrv(std::move(nvdrv)) { +NVDRV::NVDRV(Core::System& system_, std::shared_ptr<Module> nvdrv_, const char* name) + : ServiceFramework{system_, name}, nvdrv{std::move(nvdrv_)} { static const FunctionInfo functions[] = { {0, &NVDRV::Open, "Open"}, - {1, &NVDRV::Ioctl, "Ioctl"}, + {1, &NVDRV::Ioctl1, "Ioctl"}, {2, &NVDRV::Close, "Close"}, {3, &NVDRV::Initialize, "Initialize"}, {4, &NVDRV::QueryEvent, "QueryEvent"}, diff --git a/src/core/hle/service/nvdrv/interface.h b/src/core/hle/service/nvdrv/interface.h index 72e17a728..5c777c59b 100644 --- a/src/core/hle/service/nvdrv/interface.h +++ b/src/core/hle/service/nvdrv/interface.h @@ -16,14 +16,14 @@ namespace Service::Nvidia { class NVDRV final : public ServiceFramework<NVDRV> { public: - NVDRV(std::shared_ptr<Module> nvdrv, const char* name); + explicit NVDRV(Core::System& system_, std::shared_ptr<Module> nvdrv_, const char* name); ~NVDRV() override; - void SignalGPUInterruptSyncpt(const u32 syncpoint_id, const u32 value); + void SignalGPUInterruptSyncpt(u32 syncpoint_id, u32 value); private: void Open(Kernel::HLERequestContext& ctx); - void Ioctl(Kernel::HLERequestContext& ctx); + void Ioctl1(Kernel::HLERequestContext& ctx); void Ioctl2(Kernel::HLERequestContext& ctx); void Ioctl3(Kernel::HLERequestContext& ctx); void Close(Kernel::HLERequestContext& ctx); @@ -33,11 +33,13 @@ private: void SetGraphicsFirmwareMemoryMarginEnabled(Kernel::HLERequestContext& ctx); void GetStatus(Kernel::HLERequestContext& ctx); void DumpGraphicsMemoryInfo(Kernel::HLERequestContext& ctx); - void IoctlBase(Kernel::HLERequestContext& ctx, IoctlVersion version); + + void ServiceError(Kernel::HLERequestContext& ctx, NvResult result); std::shared_ptr<Module> nvdrv; u64 pid{}; + bool is_initialized{}; }; } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/nvdata.h b/src/core/hle/service/nvdrv/nvdata.h index 529b03471..3294bc0e7 100644 --- a/src/core/hle/service/nvdrv/nvdata.h +++ b/src/core/hle/service/nvdrv/nvdata.h @@ -1,12 +1,16 @@ #pragma once #include <array> +#include "common/bit_field.h" #include "common/common_types.h" namespace Service::Nvidia { constexpr u32 MaxSyncPoints = 192; constexpr u32 MaxNvEvents = 64; +using DeviceFD = s32; + +constexpr DeviceFD INVALID_NVDRV_FD = -1; struct Fence { s32 id; @@ -20,11 +24,61 @@ struct MultiFence { std::array<Fence, 4> fences; }; -enum NvResult : u32 { - Success = 0, - BadParameter = 4, - Timeout = 5, - ResourceError = 15, +enum class NvResult : u32 { + Success = 0x0, + NotImplemented = 0x1, + NotSupported = 0x2, + NotInitialized = 0x3, + BadParameter = 0x4, + Timeout = 0x5, + InsufficientMemory = 0x6, + ReadOnlyAttribute = 0x7, + InvalidState = 0x8, + InvalidAddress = 0x9, + InvalidSize = 0xA, + BadValue = 0xB, + AlreadyAllocated = 0xD, + Busy = 0xE, + ResourceError = 0xF, + CountMismatch = 0x10, + OverFlow = 0x11, + InsufficientTransferMemory = 0x1000, + InsufficientVideoMemory = 0x10000, + BadSurfaceColorScheme = 0x10001, + InvalidSurface = 0x10002, + SurfaceNotSupported = 0x10003, + DispInitFailed = 0x20000, + DispAlreadyAttached = 0x20001, + DispTooManyDisplays = 0x20002, + DispNoDisplaysAttached = 0x20003, + DispModeNotSupported = 0x20004, + DispNotFound = 0x20005, + DispAttachDissallowed = 0x20006, + DispTypeNotSupported = 0x20007, + DispAuthenticationFailed = 0x20008, + DispNotAttached = 0x20009, + DispSamePwrState = 0x2000A, + DispEdidFailure = 0x2000B, + DispDsiReadAckError = 0x2000C, + DispDsiReadInvalidResp = 0x2000D, + FileWriteFailed = 0x30000, + FileReadFailed = 0x30001, + EndOfFile = 0x30002, + FileOperationFailed = 0x30003, + DirOperationFailed = 0x30004, + EndOfDirList = 0x30005, + ConfigVarNotFound = 0x30006, + InvalidConfigVar = 0x30007, + LibraryNotFound = 0x30008, + SymbolNotFound = 0x30009, + MemoryMapFailed = 0x3000A, + IoctlFailed = 0x3000F, + AccessDenied = 0x30010, + DeviceNotFound = 0x30011, + KernelDriverNotFound = 0x30012, + FileNotFound = 0x30013, + PathAlreadyExists = 0x30014, + ModuleNotPresent = 0xA000E, }; enum class EventState { @@ -34,21 +88,13 @@ enum class EventState { Busy = 3, }; -enum class IoctlVersion : u32 { - Version1, - Version2, - Version3, -}; - -struct IoctlCtrl { - // First call done to the servioce for services that call itself again after a call. - bool fresh_call{true}; - // Tells the Ioctl Wrapper that it must delay the IPC response and send the thread to sleep - bool must_delay{}; - // Timeout for the delay - s64 timeout{}; - // NV Event Id - s32 event_id{-1}; +union Ioctl { + u32_le raw; + BitField<0, 8, u32> cmd; + BitField<8, 8, u32> group; + BitField<16, 14, u32> length; + BitField<30, 1, u32> is_in; + BitField<31, 1, u32> is_out; }; } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/nvdrv.cpp b/src/core/hle/service/nvdrv/nvdrv.cpp index 197c77db0..620c18728 100644 --- a/src/core/hle/service/nvdrv/nvdrv.cpp +++ b/src/core/hle/service/nvdrv/nvdrv.cpp @@ -5,6 +5,7 @@ #include <utility> #include <fmt/format.h> +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/readable_event.h" #include "core/hle/kernel/writable_event.h" @@ -21,6 +22,7 @@ #include "core/hle/service/nvdrv/interface.h" #include "core/hle/service/nvdrv/nvdrv.h" #include "core/hle/service/nvdrv/nvmemp.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" #include "core/hle/service/nvflinger/nvflinger.h" namespace Service::Nvidia { @@ -28,66 +30,135 @@ namespace Service::Nvidia { void InstallInterfaces(SM::ServiceManager& service_manager, NVFlinger::NVFlinger& nvflinger, Core::System& system) { auto module_ = std::make_shared<Module>(system); - std::make_shared<NVDRV>(module_, "nvdrv")->InstallAsService(service_manager); - std::make_shared<NVDRV>(module_, "nvdrv:a")->InstallAsService(service_manager); - std::make_shared<NVDRV>(module_, "nvdrv:s")->InstallAsService(service_manager); - std::make_shared<NVDRV>(module_, "nvdrv:t")->InstallAsService(service_manager); - std::make_shared<NVMEMP>()->InstallAsService(service_manager); + std::make_shared<NVDRV>(system, module_, "nvdrv")->InstallAsService(service_manager); + std::make_shared<NVDRV>(system, module_, "nvdrv:a")->InstallAsService(service_manager); + std::make_shared<NVDRV>(system, module_, "nvdrv:s")->InstallAsService(service_manager); + std::make_shared<NVDRV>(system, module_, "nvdrv:t")->InstallAsService(service_manager); + std::make_shared<NVMEMP>(system)->InstallAsService(service_manager); nvflinger.SetNVDrvInstance(module_); } -Module::Module(Core::System& system) { +Module::Module(Core::System& system) : syncpoint_manager{system.GPU()} { auto& kernel = system.Kernel(); for (u32 i = 0; i < MaxNvEvents; i++) { std::string event_label = fmt::format("NVDRV::NvEvent_{}", i); - events_interface.events[i] = Kernel::WritableEvent::CreateEventPair(kernel, event_label); + events_interface.events[i] = {Kernel::WritableEvent::CreateEventPair(kernel, event_label)}; events_interface.status[i] = EventState::Free; events_interface.registered[i] = false; } auto nvmap_dev = std::make_shared<Devices::nvmap>(system); devices["/dev/nvhost-as-gpu"] = std::make_shared<Devices::nvhost_as_gpu>(system, nvmap_dev); - devices["/dev/nvhost-gpu"] = std::make_shared<Devices::nvhost_gpu>(system, nvmap_dev); + devices["/dev/nvhost-gpu"] = + std::make_shared<Devices::nvhost_gpu>(system, nvmap_dev, syncpoint_manager); devices["/dev/nvhost-ctrl-gpu"] = std::make_shared<Devices::nvhost_ctrl_gpu>(system); devices["/dev/nvmap"] = nvmap_dev; devices["/dev/nvdisp_disp0"] = std::make_shared<Devices::nvdisp_disp0>(system, nvmap_dev); - devices["/dev/nvhost-ctrl"] = std::make_shared<Devices::nvhost_ctrl>(system, events_interface); - devices["/dev/nvhost-nvdec"] = std::make_shared<Devices::nvhost_nvdec>(system); + devices["/dev/nvhost-ctrl"] = + std::make_shared<Devices::nvhost_ctrl>(system, events_interface, syncpoint_manager); + devices["/dev/nvhost-nvdec"] = + std::make_shared<Devices::nvhost_nvdec>(system, nvmap_dev, syncpoint_manager); devices["/dev/nvhost-nvjpg"] = std::make_shared<Devices::nvhost_nvjpg>(system); - devices["/dev/nvhost-vic"] = std::make_shared<Devices::nvhost_vic>(system); + devices["/dev/nvhost-vic"] = + std::make_shared<Devices::nvhost_vic>(system, nvmap_dev, syncpoint_manager); } Module::~Module() = default; -u32 Module::Open(const std::string& device_name) { - ASSERT_MSG(devices.find(device_name) != devices.end(), "Trying to open unknown device {}", - device_name); +NvResult Module::VerifyFD(DeviceFD fd) const { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + if (open_files.find(fd) == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return NvResult::Success; +} + +DeviceFD Module::Open(const std::string& device_name) { + if (devices.find(device_name) == devices.end()) { + LOG_ERROR(Service_NVDRV, "Trying to open unknown device {}", device_name); + return INVALID_NVDRV_FD; + } auto device = devices[device_name]; - const u32 fd = next_fd++; + const DeviceFD fd = next_fd++; open_files[fd] = std::move(device); return fd; } -u32 Module::Ioctl(u32 fd, u32 command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version) { - auto itr = open_files.find(fd); - ASSERT_MSG(itr != open_files.end(), "Tried to talk to an invalid device"); +NvResult Module::Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } - auto& device = itr->second; - return device->ioctl({command}, input, input2, output, output2, ctrl, version); + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return itr->second->Ioctl1(command, input, output); } -ResultCode Module::Close(u32 fd) { - auto itr = open_files.find(fd); - ASSERT_MSG(itr != open_files.end(), "Tried to talk to an invalid device"); +NvResult Module::Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return itr->second->Ioctl2(command, input, inline_input, output); +} + +NvResult Module::Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } + + return itr->second->Ioctl3(command, input, output, inline_output); +} + +NvResult Module::Close(DeviceFD fd) { + if (fd < 0) { + LOG_ERROR(Service_NVDRV, "Invalid DeviceFD={}!", fd); + return NvResult::InvalidState; + } + + const auto itr = open_files.find(fd); + + if (itr == open_files.end()) { + LOG_ERROR(Service_NVDRV, "Could not find DeviceFD={}!", fd); + return NvResult::NotImplemented; + } open_files.erase(itr); - // TODO(flerovium): return correct result code if operation failed. - return RESULT_SUCCESS; + return NvResult::Success; } void Module::SignalSyncpt(const u32 syncpoint_id, const u32 value) { @@ -95,17 +166,17 @@ void Module::SignalSyncpt(const u32 syncpoint_id, const u32 value) { if (events_interface.assigned_syncpt[i] == syncpoint_id && events_interface.assigned_value[i] == value) { events_interface.LiberateEvent(i); - events_interface.events[i].writable->Signal(); + events_interface.events[i].event.writable->Signal(); } } } std::shared_ptr<Kernel::ReadableEvent> Module::GetEvent(const u32 event_id) const { - return events_interface.events[event_id].readable; + return events_interface.events[event_id].event.readable; } std::shared_ptr<Kernel::WritableEvent> Module::GetEventWriteable(const u32 event_id) const { - return events_interface.events[event_id].writable; + return events_interface.events[event_id].event.writable; } } // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/nvdrv.h b/src/core/hle/service/nvdrv/nvdrv.h index 7706a5590..144e657e5 100644 --- a/src/core/hle/service/nvdrv/nvdrv.h +++ b/src/core/hle/service/nvdrv/nvdrv.h @@ -10,6 +10,7 @@ #include "common/common_types.h" #include "core/hle/kernel/writable_event.h" #include "core/hle/service/nvdrv/nvdata.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" #include "core/hle/service/service.h" namespace Core { @@ -22,15 +23,23 @@ class NVFlinger; namespace Service::Nvidia { +class SyncpointManager; + namespace Devices { class nvdevice; } +/// Represents an Nvidia event +struct NvEvent { + Kernel::EventPair event; + Fence fence{}; +}; + struct EventInterface { // Mask representing currently busy events u64 events_mask{}; // Each kernel event associated to an NV event - std::array<Kernel::EventPair, MaxNvEvents> events; + std::array<NvEvent, MaxNvEvents> events; // The status of the current NVEvent std::array<EventState, MaxNvEvents> status{}; // Tells if an NVEvent is registered or not @@ -91,7 +100,7 @@ struct EventInterface { class Module final { public: - Module(Core::System& system); + explicit Module(Core::System& system_); ~Module(); /// Returns a pointer to one of the available devices, identified by its name. @@ -103,14 +112,23 @@ public: return std::static_pointer_cast<T>(itr->second); } + NvResult VerifyFD(DeviceFD fd) const; + /// Opens a device node and returns a file descriptor to it. - u32 Open(const std::string& device_name); + DeviceFD Open(const std::string& device_name); + /// Sends an ioctl command to the specified file descriptor. - u32 Ioctl(u32 fd, u32 command, const std::vector<u8>& input, const std::vector<u8>& input2, - std::vector<u8>& output, std::vector<u8>& output2, IoctlCtrl& ctrl, - IoctlVersion version); + NvResult Ioctl1(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output); + + NvResult Ioctl2(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + const std::vector<u8>& inline_input, std::vector<u8>& output); + + NvResult Ioctl3(DeviceFD fd, Ioctl command, const std::vector<u8>& input, + std::vector<u8>& output, std::vector<u8>& inline_output); + /// Closes a device file descriptor and returns operation success. - ResultCode Close(u32 fd); + NvResult Close(DeviceFD fd); void SignalSyncpt(const u32 syncpoint_id, const u32 value); @@ -119,11 +137,14 @@ public: std::shared_ptr<Kernel::WritableEvent> GetEventWriteable(u32 event_id) const; private: + /// Manages syncpoints on the host + SyncpointManager syncpoint_manager; + /// Id to use for the next open file descriptor. - u32 next_fd = 1; + DeviceFD next_fd = 1; /// Mapping of file descriptors to the devices they reference. - std::unordered_map<u32, std::shared_ptr<Devices::nvdevice>> open_files; + std::unordered_map<DeviceFD, std::shared_ptr<Devices::nvdevice>> open_files; /// Mapping of device node names to their implementation. std::unordered_map<std::string, std::shared_ptr<Devices::nvdevice>> devices; diff --git a/src/core/hle/service/nvdrv/nvmemp.cpp b/src/core/hle/service/nvdrv/nvmemp.cpp index 73b37e805..331c02243 100644 --- a/src/core/hle/service/nvdrv/nvmemp.cpp +++ b/src/core/hle/service/nvdrv/nvmemp.cpp @@ -8,7 +8,7 @@ namespace Service::Nvidia { -NVMEMP::NVMEMP() : ServiceFramework("nvmemp") { +NVMEMP::NVMEMP(Core::System& system_) : ServiceFramework{system_, "nvmemp"} { static const FunctionInfo functions[] = { {0, &NVMEMP::Open, "Open"}, {1, &NVMEMP::GetAruid, "GetAruid"}, diff --git a/src/core/hle/service/nvdrv/nvmemp.h b/src/core/hle/service/nvdrv/nvmemp.h index c453ee4db..724c27ef9 100644 --- a/src/core/hle/service/nvdrv/nvmemp.h +++ b/src/core/hle/service/nvdrv/nvmemp.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Nvidia { class NVMEMP final : public ServiceFramework<NVMEMP> { public: - NVMEMP(); + explicit NVMEMP(Core::System& system_); ~NVMEMP() override; private: diff --git a/src/core/hle/service/nvdrv/syncpoint_manager.cpp b/src/core/hle/service/nvdrv/syncpoint_manager.cpp new file mode 100644 index 000000000..0151a03b7 --- /dev/null +++ b/src/core/hle/service/nvdrv/syncpoint_manager.cpp @@ -0,0 +1,39 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "core/hle/service/nvdrv/syncpoint_manager.h" +#include "video_core/gpu.h" + +namespace Service::Nvidia { + +SyncpointManager::SyncpointManager(Tegra::GPU& gpu) : gpu{gpu} {} + +SyncpointManager::~SyncpointManager() = default; + +u32 SyncpointManager::RefreshSyncpoint(u32 syncpoint_id) { + syncpoints[syncpoint_id].min = gpu.GetSyncpointValue(syncpoint_id); + return GetSyncpointMin(syncpoint_id); +} + +u32 SyncpointManager::AllocateSyncpoint() { + for (u32 syncpoint_id = 1; syncpoint_id < MaxSyncPoints; syncpoint_id++) { + if (!syncpoints[syncpoint_id].is_allocated) { + syncpoints[syncpoint_id].is_allocated = true; + return syncpoint_id; + } + } + UNREACHABLE_MSG("No more available syncpoints!"); + return {}; +} + +u32 SyncpointManager::IncreaseSyncpoint(u32 syncpoint_id, u32 value) { + for (u32 index = 0; index < value; ++index) { + syncpoints[syncpoint_id].max.fetch_add(1, std::memory_order_relaxed); + } + + return GetSyncpointMax(syncpoint_id); +} + +} // namespace Service::Nvidia diff --git a/src/core/hle/service/nvdrv/syncpoint_manager.h b/src/core/hle/service/nvdrv/syncpoint_manager.h new file mode 100644 index 000000000..d395c5d0b --- /dev/null +++ b/src/core/hle/service/nvdrv/syncpoint_manager.h @@ -0,0 +1,85 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <atomic> + +#include "common/common_types.h" +#include "core/hle/service/nvdrv/nvdata.h" + +namespace Tegra { +class GPU; +} + +namespace Service::Nvidia { + +class SyncpointManager final { +public: + explicit SyncpointManager(Tegra::GPU& gpu); + ~SyncpointManager(); + + /** + * Returns true if the specified syncpoint is expired for the given value. + * @param syncpoint_id Syncpoint ID to check. + * @param value Value to check against the specified syncpoint. + * @returns True if the specified syncpoint is expired for the given value, otherwise False. + */ + bool IsSyncpointExpired(u32 syncpoint_id, u32 value) const { + return (GetSyncpointMax(syncpoint_id) - value) >= (GetSyncpointMin(syncpoint_id) - value); + } + + /** + * Gets the lower bound for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to get the lower bound for. + * @returns The lower bound for the specified syncpoint. + */ + u32 GetSyncpointMin(u32 syncpoint_id) const { + return syncpoints.at(syncpoint_id).min.load(std::memory_order_relaxed); + } + + /** + * Gets the uper bound for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to get the upper bound for. + * @returns The upper bound for the specified syncpoint. + */ + u32 GetSyncpointMax(u32 syncpoint_id) const { + return syncpoints.at(syncpoint_id).max.load(std::memory_order_relaxed); + } + + /** + * Refreshes the minimum value for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to be refreshed. + * @returns The new syncpoint minimum value. + */ + u32 RefreshSyncpoint(u32 syncpoint_id); + + /** + * Allocates a new syncoint. + * @returns The syncpoint ID for the newly allocated syncpoint. + */ + u32 AllocateSyncpoint(); + + /** + * Increases the maximum value for the specified syncpoint. + * @param syncpoint_id Syncpoint ID to be increased. + * @param value Value to increase the specified syncpoint by. + * @returns The new syncpoint maximum value. + */ + u32 IncreaseSyncpoint(u32 syncpoint_id, u32 value); + +private: + struct Syncpoint { + std::atomic<u32> min; + std::atomic<u32> max; + std::atomic<bool> is_allocated; + }; + + std::array<Syncpoint, MaxSyncPoints> syncpoints{}; + + Tegra::GPU& gpu; +}; + +} // namespace Service::Nvidia diff --git a/src/core/hle/service/nvflinger/buffer_queue.cpp b/src/core/hle/service/nvflinger/buffer_queue.cpp index 637b310d7..5578181a4 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.cpp +++ b/src/core/hle/service/nvflinger/buffer_queue.cpp @@ -22,127 +22,169 @@ BufferQueue::BufferQueue(Kernel::KernelCore& kernel, u32 id, u64 layer_id) BufferQueue::~BufferQueue() = default; void BufferQueue::SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer) { + ASSERT(slot < buffer_slots); LOG_WARNING(Service, "Adding graphics buffer {}", slot); - free_buffers.push_back(slot); - queue.push_back({ + { + std::unique_lock lock{free_buffers_mutex}; + free_buffers.push_back(slot); + } + free_buffers_condition.notify_one(); + + buffers[slot] = { .slot = slot, .status = Buffer::Status::Free, .igbp_buffer = igbp_buffer, - }); + .transform = {}, + .crop_rect = {}, + .swap_interval = 0, + .multi_fence = {}, + }; buffer_wait_event.writable->Signal(); } std::optional<std::pair<u32, Service::Nvidia::MultiFence*>> BufferQueue::DequeueBuffer(u32 width, u32 height) { + // Wait for first request before trying to dequeue + { + std::unique_lock lock{free_buffers_mutex}; + free_buffers_condition.wait(lock, [this] { return !free_buffers.empty() || !is_connect; }); + } - if (free_buffers.empty()) { + if (!is_connect) { + // Buffer was disconnected while the thread was blocked, this is most likely due to + // emulation being stopped return std::nullopt; } + std::unique_lock lock{free_buffers_mutex}; + auto f_itr = free_buffers.begin(); - auto itr = queue.end(); + auto slot = buffers.size(); while (f_itr != free_buffers.end()) { - auto slot = *f_itr; - itr = std::find_if(queue.begin(), queue.end(), [&](const Buffer& buffer) { - // Only consider free buffers. Buffers become free once again after they've been - // Acquired and Released by the compositor, see the NVFlinger::Compose method. - if (buffer.status != Buffer::Status::Free) { - return false; - } - - if (buffer.slot != slot) { - return false; - } - - // Make sure that the parameters match. - return buffer.igbp_buffer.width == width && buffer.igbp_buffer.height == height; - }); - - if (itr != queue.end()) { + const Buffer& buffer = buffers[*f_itr]; + if (buffer.status == Buffer::Status::Free && buffer.igbp_buffer.width == width && + buffer.igbp_buffer.height == height) { + slot = *f_itr; free_buffers.erase(f_itr); break; } ++f_itr; } - - if (itr == queue.end()) { + if (slot == buffers.size()) { return std::nullopt; } - - itr->status = Buffer::Status::Dequeued; - return {{itr->slot, &itr->multi_fence}}; + buffers[slot].status = Buffer::Status::Dequeued; + return {{buffers[slot].slot, &buffers[slot].multi_fence}}; } const IGBPBuffer& BufferQueue::RequestBuffer(u32 slot) const { - auto itr = std::find_if(queue.begin(), queue.end(), - [&](const Buffer& buffer) { return buffer.slot == slot; }); - ASSERT(itr != queue.end()); - ASSERT(itr->status == Buffer::Status::Dequeued); - return itr->igbp_buffer; + ASSERT(slot < buffers.size()); + ASSERT(buffers[slot].status == Buffer::Status::Dequeued); + ASSERT(buffers[slot].slot == slot); + + return buffers[slot].igbp_buffer; } void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform, const Common::Rectangle<int>& crop_rect, u32 swap_interval, Service::Nvidia::MultiFence& multi_fence) { - auto itr = std::find_if(queue.begin(), queue.end(), - [&](const Buffer& buffer) { return buffer.slot == slot; }); - ASSERT(itr != queue.end()); - ASSERT(itr->status == Buffer::Status::Dequeued); - itr->status = Buffer::Status::Queued; - itr->transform = transform; - itr->crop_rect = crop_rect; - itr->swap_interval = swap_interval; - itr->multi_fence = multi_fence; + ASSERT(slot < buffers.size()); + ASSERT(buffers[slot].status == Buffer::Status::Dequeued); + ASSERT(buffers[slot].slot == slot); + + buffers[slot].status = Buffer::Status::Queued; + buffers[slot].transform = transform; + buffers[slot].crop_rect = crop_rect; + buffers[slot].swap_interval = swap_interval; + buffers[slot].multi_fence = multi_fence; + std::unique_lock lock{queue_sequence_mutex}; queue_sequence.push_back(slot); } +void BufferQueue::CancelBuffer(u32 slot, const Service::Nvidia::MultiFence& multi_fence) { + ASSERT(slot < buffers.size()); + ASSERT(buffers[slot].status != Buffer::Status::Free); + ASSERT(buffers[slot].slot == slot); + + buffers[slot].status = Buffer::Status::Free; + buffers[slot].multi_fence = multi_fence; + buffers[slot].swap_interval = 0; + + { + std::unique_lock lock{free_buffers_mutex}; + free_buffers.push_back(slot); + } + free_buffers_condition.notify_one(); + + buffer_wait_event.writable->Signal(); +} + std::optional<std::reference_wrapper<const BufferQueue::Buffer>> BufferQueue::AcquireBuffer() { - auto itr = queue.end(); + std::unique_lock lock{queue_sequence_mutex}; + std::size_t buffer_slot = buffers.size(); // Iterate to find a queued buffer matching the requested slot. - while (itr == queue.end() && !queue_sequence.empty()) { - const u32 slot = queue_sequence.front(); - itr = std::find_if(queue.begin(), queue.end(), [&slot](const Buffer& buffer) { - return buffer.status == Buffer::Status::Queued && buffer.slot == slot; - }); + while (buffer_slot == buffers.size() && !queue_sequence.empty()) { + const auto slot = static_cast<std::size_t>(queue_sequence.front()); + ASSERT(slot < buffers.size()); + if (buffers[slot].status == Buffer::Status::Queued) { + ASSERT(buffers[slot].slot == slot); + buffer_slot = slot; + } queue_sequence.pop_front(); } - if (itr == queue.end()) { + if (buffer_slot == buffers.size()) { return std::nullopt; } - itr->status = Buffer::Status::Acquired; - return *itr; + buffers[buffer_slot].status = Buffer::Status::Acquired; + return {{buffers[buffer_slot]}}; } void BufferQueue::ReleaseBuffer(u32 slot) { - auto itr = std::find_if(queue.begin(), queue.end(), - [&](const Buffer& buffer) { return buffer.slot == slot; }); - ASSERT(itr != queue.end()); - ASSERT(itr->status == Buffer::Status::Acquired); - itr->status = Buffer::Status::Free; - free_buffers.push_back(slot); + ASSERT(slot < buffers.size()); + ASSERT(buffers[slot].status == Buffer::Status::Acquired); + ASSERT(buffers[slot].slot == slot); + + buffers[slot].status = Buffer::Status::Free; + { + std::unique_lock lock{free_buffers_mutex}; + free_buffers.push_back(slot); + } + free_buffers_condition.notify_one(); buffer_wait_event.writable->Signal(); } -void BufferQueue::Disconnect() { - queue.clear(); +void BufferQueue::Connect() { + std::unique_lock lock{queue_sequence_mutex}; queue_sequence.clear(); - id = 1; - layer_id = 1; + is_connect = true; +} + +void BufferQueue::Disconnect() { + buffers.fill({}); + { + std::unique_lock lock{queue_sequence_mutex}; + queue_sequence.clear(); + } + buffer_wait_event.writable->Signal(); + is_connect = false; + free_buffers_condition.notify_one(); } u32 BufferQueue::Query(QueryType type) { - LOG_WARNING(Service, "(STUBBED) called type={}", static_cast<u32>(type)); + LOG_WARNING(Service, "(STUBBED) called type={}", type); switch (type) { case QueryType::NativeWindowFormat: return static_cast<u32>(PixelFormat::RGBA8888); + case QueryType::NativeWindowWidth: + case QueryType::NativeWindowHeight: + break; } - - UNIMPLEMENTED(); + UNIMPLEMENTED_MSG("Unimplemented query type={}", type); return 0; } diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h index 8a837e5aa..ad7469277 100644 --- a/src/core/hle/service/nvflinger/buffer_queue.h +++ b/src/core/hle/service/nvflinger/buffer_queue.h @@ -4,7 +4,9 @@ #pragma once +#include <condition_variable> #include <list> +#include <mutex> #include <optional> #include <vector> @@ -21,6 +23,7 @@ class KernelCore; namespace Service::NVFlinger { +constexpr u32 buffer_slots = 0x40; struct IGBPBuffer { u32_le magic; u32_le width; @@ -95,8 +98,10 @@ public: void QueueBuffer(u32 slot, BufferTransformFlags transform, const Common::Rectangle<int>& crop_rect, u32 swap_interval, Service::Nvidia::MultiFence& multi_fence); + void CancelBuffer(u32 slot, const Service::Nvidia::MultiFence& multi_fence); std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer(); void ReleaseBuffer(u32 slot); + void Connect(); void Disconnect(); u32 Query(QueryType type); @@ -104,18 +109,30 @@ public: return id; } + bool IsConnected() const { + return is_connect; + } + std::shared_ptr<Kernel::WritableEvent> GetWritableBufferWaitEvent() const; std::shared_ptr<Kernel::ReadableEvent> GetBufferWaitEvent() const; private: - u32 id; - u64 layer_id; + BufferQueue(const BufferQueue&) = delete; + + u32 id{}; + u64 layer_id{}; + std::atomic_bool is_connect{}; std::list<u32> free_buffers; - std::vector<Buffer> queue; + std::array<Buffer, buffer_slots> buffers; std::list<u32> queue_sequence; Kernel::EventPair buffer_wait_event; + + std::mutex free_buffers_mutex; + std::condition_variable free_buffers_condition; + + std::mutex queue_sequence_mutex; }; } // namespace Service::NVFlinger diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp index c64673dba..4b3581949 100644 --- a/src/core/hle/service/nvflinger/nvflinger.cpp +++ b/src/core/hle/service/nvflinger/nvflinger.cpp @@ -88,6 +88,10 @@ NVFlinger::NVFlinger(Core::System& system) : system(system) { } NVFlinger::~NVFlinger() { + for (auto& buffer_queue : buffer_queues) { + buffer_queue->Disconnect(); + } + if (system.IsMulticore()) { is_running = false; wait_event->Set(); @@ -104,6 +108,8 @@ void NVFlinger::SetNVDrvInstance(std::shared_ptr<Nvidia::Module> instance) { } std::optional<u64> NVFlinger::OpenDisplay(std::string_view name) { + const auto guard = Lock(); + LOG_DEBUG(Service, "Opening \"{}\" display", name); // TODO(Subv): Currently we only support the Default display. @@ -121,6 +127,7 @@ std::optional<u64> NVFlinger::OpenDisplay(std::string_view name) { } std::optional<u64> NVFlinger::CreateLayer(u64 display_id) { + const auto guard = Lock(); auto* const display = FindDisplay(display_id); if (display == nullptr) { @@ -129,18 +136,22 @@ std::optional<u64> NVFlinger::CreateLayer(u64 display_id) { const u64 layer_id = next_layer_id++; const u32 buffer_queue_id = next_buffer_queue_id++; - buffer_queues.emplace_back(system.Kernel(), buffer_queue_id, layer_id); - display->CreateLayer(layer_id, buffer_queues.back()); + buffer_queues.emplace_back( + std::make_unique<BufferQueue>(system.Kernel(), buffer_queue_id, layer_id)); + display->CreateLayer(layer_id, *buffer_queues.back()); return layer_id; } void NVFlinger::CloseLayer(u64 layer_id) { + const auto guard = Lock(); + for (auto& display : displays) { display.CloseLayer(layer_id); } } std::optional<u32> NVFlinger::FindBufferQueueId(u64 display_id, u64 layer_id) const { + const auto guard = Lock(); const auto* const layer = FindLayer(display_id, layer_id); if (layer == nullptr) { @@ -151,6 +162,7 @@ std::optional<u32> NVFlinger::FindBufferQueueId(u64 display_id, u64 layer_id) co } std::shared_ptr<Kernel::ReadableEvent> NVFlinger::FindVsyncEvent(u64 display_id) const { + const auto guard = Lock(); auto* const display = FindDisplay(display_id); if (display == nullptr) { @@ -160,20 +172,16 @@ std::shared_ptr<Kernel::ReadableEvent> NVFlinger::FindVsyncEvent(u64 display_id) return display->GetVSyncEvent(); } -BufferQueue& NVFlinger::FindBufferQueue(u32 id) { +BufferQueue* NVFlinger::FindBufferQueue(u32 id) { + const auto guard = Lock(); const auto itr = std::find_if(buffer_queues.begin(), buffer_queues.end(), - [id](const auto& queue) { return queue.GetId() == id; }); - - ASSERT(itr != buffer_queues.end()); - return *itr; -} + [id](const auto& queue) { return queue->GetId() == id; }); -const BufferQueue& NVFlinger::FindBufferQueue(u32 id) const { - const auto itr = std::find_if(buffer_queues.begin(), buffer_queues.end(), - [id](const auto& queue) { return queue.GetId() == id; }); + if (itr == buffer_queues.end()) { + return nullptr; + } - ASSERT(itr != buffer_queues.end()); - return *itr; + return itr->get(); } VI::Display* NVFlinger::FindDisplay(u64 display_id) { @@ -242,6 +250,10 @@ void NVFlinger::Compose() { const auto& igbp_buffer = buffer->get().igbp_buffer; + if (!system.IsPoweredOn()) { + return; // We are likely shutting down + } + auto& gpu = system.GPU(); const auto& multi_fence = buffer->get().multi_fence; guard->unlock(); diff --git a/src/core/hle/service/nvflinger/nvflinger.h b/src/core/hle/service/nvflinger/nvflinger.h index 1ebe949c0..c6765259f 100644 --- a/src/core/hle/service/nvflinger/nvflinger.h +++ b/src/core/hle/service/nvflinger/nvflinger.h @@ -75,10 +75,7 @@ public: [[nodiscard]] std::shared_ptr<Kernel::ReadableEvent> FindVsyncEvent(u64 display_id) const; /// Obtains a buffer queue identified by the ID. - [[nodiscard]] BufferQueue& FindBufferQueue(u32 id); - - /// Obtains a buffer queue identified by the ID. - [[nodiscard]] const BufferQueue& FindBufferQueue(u32 id) const; + [[nodiscard]] BufferQueue* FindBufferQueue(u32 id); /// Performs a composition request to the emulated nvidia GPU and triggers the vsync events when /// finished. @@ -86,11 +83,11 @@ public: [[nodiscard]] s64 GetNextTicks() const; +private: [[nodiscard]] std::unique_lock<std::mutex> Lock() const { return std::unique_lock{*guard}; } -private: /// Finds the display identified by the specified ID. [[nodiscard]] VI::Display* FindDisplay(u64 display_id); @@ -110,7 +107,7 @@ private: std::shared_ptr<Nvidia::Module> nvdrv; std::vector<VI::Display> displays; - std::vector<BufferQueue> buffer_queues; + std::vector<std::unique_ptr<BufferQueue>> buffer_queues; /// Id to use for the next layer that is created, this counter is shared among all displays. u64 next_layer_id = 1; diff --git a/src/core/hle/service/olsc/olsc.cpp b/src/core/hle/service/olsc/olsc.cpp new file mode 100644 index 000000000..4440135ed --- /dev/null +++ b/src/core/hle/service/olsc/olsc.cpp @@ -0,0 +1,69 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "core/hle/ipc_helpers.h" +#include "core/hle/kernel/hle_ipc.h" +#include "core/hle/service/olsc/olsc.h" +#include "core/hle/service/service.h" +#include "core/hle/service/sm/sm.h" + +namespace Service::OLSC { + +class OLSC final : public ServiceFramework<OLSC> { +public: + explicit OLSC(Core::System& system_) : ServiceFramework{system_, "olsc:u"} { + // clang-format off + static const FunctionInfo functions[] = { + {0, &OLSC::Initialize, "Initialize"}, + {10, nullptr, "VerifySaveDataBackupLicenseAsync"}, + {13, nullptr, "GetSaveDataBackupSetting"}, + {14, &OLSC::SetSaveDataBackupSettingEnabled, "SetSaveDataBackupSettingEnabled"}, + {15, nullptr, "SetCustomData"}, + {16, nullptr, "DeleteSaveDataBackupSetting"}, + {18, nullptr, "GetSaveDataBackupInfoCache"}, + {19, nullptr, "UpdateSaveDataBackupInfoCacheAsync"}, + {22, nullptr, "DeleteSaveDataBackupAsync"}, + {25, nullptr, "ListDownloadableSaveDataBackupInfoAsync"}, + {26, nullptr, "DownloadSaveDataBackupAsync"}, + {9010, nullptr, "VerifySaveDataBackupLicenseAsyncForDebug"}, + {9013, nullptr, "GetSaveDataBackupSettingForDebug"}, + {9014, nullptr, "SetSaveDataBackupSettingEnabledForDebug"}, + {9015, nullptr, "SetCustomDataForDebug"}, + {9016, nullptr, "DeleteSaveDataBackupSettingForDebug"}, + {9018, nullptr, "GetSaveDataBackupInfoCacheForDebug"}, + {9019, nullptr, "UpdateSaveDataBackupInfoCacheAsyncForDebug"}, + {9022, nullptr, "DeleteSaveDataBackupAsyncForDebug"}, + {9025, nullptr, "ListDownloadableSaveDataBackupInfoAsyncForDebug"}, + {9026, nullptr, "DownloadSaveDataBackupAsyncForDebug"}, + }; + // clang-format on + + RegisterHandlers(functions); + } + +private: + void Initialize(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_OLSC, "(STUBBED) called"); + + initialized = true; + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + void SetSaveDataBackupSettingEnabled(Kernel::HLERequestContext& ctx) { + LOG_WARNING(Service_OLSC, "(STUBBED) called"); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); + } + + bool initialized{}; +}; + +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { + std::make_shared<OLSC>(system)->InstallAsService(service_manager); +} + +} // namespace Service::OLSC diff --git a/src/core/hle/service/olsc/olsc.h b/src/core/hle/service/olsc/olsc.h new file mode 100644 index 000000000..24f24ca6b --- /dev/null +++ b/src/core/hle/service/olsc/olsc.h @@ -0,0 +1,20 @@ +// Copyright 2020 yuzu emulator team +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +namespace Core { +class System; +} + +namespace Service::SM { +class ServiceManager; +} + +namespace Service::OLSC { + +/// Registers all SSL services with the specified service manager. +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); + +} // namespace Service::OLSC diff --git a/src/core/hle/service/pcie/pcie.cpp b/src/core/hle/service/pcie/pcie.cpp index c568a0adc..f6686fc4d 100644 --- a/src/core/hle/service/pcie/pcie.cpp +++ b/src/core/hle/service/pcie/pcie.cpp @@ -12,7 +12,7 @@ namespace Service::PCIe { class ISession final : public ServiceFramework<ISession> { public: - explicit ISession() : ServiceFramework{"ISession"} { + explicit ISession(Core::System& system_) : ServiceFramework{system_, "ISession"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "QueryFunctions"}, @@ -48,7 +48,7 @@ public: class PCIe final : public ServiceFramework<PCIe> { public: - explicit PCIe() : ServiceFramework{"pcie"} { + explicit PCIe(Core::System& system_) : ServiceFramework{system_, "pcie"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "RegisterClassDriver"}, @@ -60,8 +60,8 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<PCIe>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<PCIe>(system)->InstallAsService(sm); } } // namespace Service::PCIe diff --git a/src/core/hle/service/pcie/pcie.h b/src/core/hle/service/pcie/pcie.h index 59c22ca45..e5709a72f 100644 --- a/src/core/hle/service/pcie/pcie.h +++ b/src/core/hle/service/pcie/pcie.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::PCIe { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::PCIe diff --git a/src/core/hle/service/pctl/module.cpp b/src/core/hle/service/pctl/module.cpp index caf14ed61..6ab1e4124 100644 --- a/src/core/hle/service/pctl/module.cpp +++ b/src/core/hle/service/pctl/module.cpp @@ -11,7 +11,8 @@ namespace Service::PCTL { class IParentalControlService final : public ServiceFramework<IParentalControlService> { public: - IParentalControlService() : ServiceFramework("IParentalControlService") { + explicit IParentalControlService(Core::System& system_) + : ServiceFramework{system_, "IParentalControlService"} { // clang-format off static const FunctionInfo functions[] = { {1, &IParentalControlService::Initialize, "Initialize"}, @@ -137,7 +138,7 @@ void Module::Interface::CreateService(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IParentalControlService>(); + rb.PushIpcInterface<IParentalControlService>(system); } void Module::Interface::CreateServiceWithoutInitialize(Kernel::HLERequestContext& ctx) { @@ -145,20 +146,20 @@ void Module::Interface::CreateServiceWithoutInitialize(Kernel::HLERequestContext IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IParentalControlService>(); + rb.PushIpcInterface<IParentalControlService>(system); } -Module::Interface::Interface(std::shared_ptr<Module> module, const char* name) - : ServiceFramework(name), module(std::move(module)) {} +Module::Interface::Interface(Core::System& system_, std::shared_ptr<Module> module_, const char* name) + : ServiceFramework{system_, name}, module{std::move(module_)} {} Module::Interface::~Interface() = default; -void InstallInterfaces(SM::ServiceManager& service_manager) { +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { auto module = std::make_shared<Module>(); - std::make_shared<PCTL>(module, "pctl")->InstallAsService(service_manager); - std::make_shared<PCTL>(module, "pctl:a")->InstallAsService(service_manager); - std::make_shared<PCTL>(module, "pctl:r")->InstallAsService(service_manager); - std::make_shared<PCTL>(module, "pctl:s")->InstallAsService(service_manager); + std::make_shared<PCTL>(system, module, "pctl")->InstallAsService(service_manager); + std::make_shared<PCTL>(system, module, "pctl:a")->InstallAsService(service_manager); + std::make_shared<PCTL>(system, module, "pctl:r")->InstallAsService(service_manager); + std::make_shared<PCTL>(system, module, "pctl:s")->InstallAsService(service_manager); } } // namespace Service::PCTL diff --git a/src/core/hle/service/pctl/module.h b/src/core/hle/service/pctl/module.h index 3e449110d..4c7e09a3b 100644 --- a/src/core/hle/service/pctl/module.h +++ b/src/core/hle/service/pctl/module.h @@ -6,13 +6,18 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::PCTL { class Module final { public: class Interface : public ServiceFramework<Interface> { public: - explicit Interface(std::shared_ptr<Module> module, const char* name); + explicit Interface(Core::System& system_, std::shared_ptr<Module> module_, + const char* name); ~Interface() override; void CreateService(Kernel::HLERequestContext& ctx); @@ -24,6 +29,6 @@ public: }; /// Registers all PCTL services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); } // namespace Service::PCTL diff --git a/src/core/hle/service/pctl/pctl.cpp b/src/core/hle/service/pctl/pctl.cpp index af9d1433a..16dd34f90 100644 --- a/src/core/hle/service/pctl/pctl.cpp +++ b/src/core/hle/service/pctl/pctl.cpp @@ -6,8 +6,8 @@ namespace Service::PCTL { -PCTL::PCTL(std::shared_ptr<Module> module, const char* name) - : Module::Interface(std::move(module), name) { +PCTL::PCTL(Core::System& system_, std::shared_ptr<Module> module_, const char* name) + : Interface{system_, std::move(module_), name} { static const FunctionInfo functions[] = { {0, &PCTL::CreateService, "CreateService"}, {1, &PCTL::CreateServiceWithoutInitialize, "CreateServiceWithoutInitialize"}, diff --git a/src/core/hle/service/pctl/pctl.h b/src/core/hle/service/pctl/pctl.h index c33ea80b6..275d23007 100644 --- a/src/core/hle/service/pctl/pctl.h +++ b/src/core/hle/service/pctl/pctl.h @@ -6,11 +6,15 @@ #include "core/hle/service/pctl/module.h" +namespace Core { +class System; +} + namespace Service::PCTL { class PCTL final : public Module::Interface { public: - explicit PCTL(std::shared_ptr<Module> module, const char* name); + explicit PCTL(Core::System& system_, std::shared_ptr<Module> module_, const char* name); ~PCTL() override; }; diff --git a/src/core/hle/service/pcv/pcv.cpp b/src/core/hle/service/pcv/pcv.cpp index 8bfc0276e..68b2c4178 100644 --- a/src/core/hle/service/pcv/pcv.cpp +++ b/src/core/hle/service/pcv/pcv.cpp @@ -12,7 +12,7 @@ namespace Service::PCV { class PCV final : public ServiceFramework<PCV> { public: - explicit PCV() : ServiceFramework{"pcv"} { + explicit PCV(Core::System& system_) : ServiceFramework{system_, "pcv"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SetPowerEnabled"}, @@ -54,7 +54,7 @@ public: class PCV_ARB final : public ServiceFramework<PCV_ARB> { public: - explicit PCV_ARB() : ServiceFramework{"pcv:arb"} { + explicit PCV_ARB(Core::System& system_) : ServiceFramework{system_, "pcv:arb"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "ReleaseControl"}, @@ -67,7 +67,7 @@ public: class PCV_IMM final : public ServiceFramework<PCV_IMM> { public: - explicit PCV_IMM() : ServiceFramework{"pcv:imm"} { + explicit PCV_IMM(Core::System& system_) : ServiceFramework{system_, "pcv:imm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SetClockRate"}, @@ -78,10 +78,10 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<PCV>()->InstallAsService(sm); - std::make_shared<PCV_ARB>()->InstallAsService(sm); - std::make_shared<PCV_IMM>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<PCV>(system)->InstallAsService(sm); + std::make_shared<PCV_ARB>(system)->InstallAsService(sm); + std::make_shared<PCV_IMM>(system)->InstallAsService(sm); } } // namespace Service::PCV diff --git a/src/core/hle/service/pcv/pcv.h b/src/core/hle/service/pcv/pcv.h index 219a893c3..c61a0b591 100644 --- a/src/core/hle/service/pcv/pcv.h +++ b/src/core/hle/service/pcv/pcv.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::PCV { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::PCV diff --git a/src/core/hle/service/pm/pm.cpp b/src/core/hle/service/pm/pm.cpp index f43122ad2..68736c40c 100644 --- a/src/core/hle/service/pm/pm.cpp +++ b/src/core/hle/service/pm/pm.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" @@ -43,7 +44,7 @@ void GetApplicationPidGeneric(Kernel::HLERequestContext& ctx, class BootMode final : public ServiceFramework<BootMode> { public: - explicit BootMode() : ServiceFramework{"pm:bm"} { + explicit BootMode(Core::System& system_) : ServiceFramework{system_, "pm:bm"} { static const FunctionInfo functions[] = { {0, &BootMode::GetBootMode, "GetBootMode"}, {1, &BootMode::SetMaintenanceBoot, "SetMaintenanceBoot"}, @@ -74,8 +75,8 @@ private: class DebugMonitor final : public ServiceFramework<DebugMonitor> { public: - explicit DebugMonitor(const Kernel::KernelCore& kernel) - : ServiceFramework{"pm:dmnt"}, kernel(kernel) { + explicit DebugMonitor(Core::System& system_) + : ServiceFramework{system_, "pm:dmnt"}, kernel{system_.Kernel()} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetJitDebugProcessIdList"}, @@ -124,8 +125,9 @@ private: class Info final : public ServiceFramework<Info> { public: - explicit Info(const std::vector<std::shared_ptr<Kernel::Process>>& process_list) - : ServiceFramework{"pm:info"}, process_list(process_list) { + explicit Info(Core::System& system_, + const std::vector<std::shared_ptr<Kernel::Process>>& process_list_) + : ServiceFramework{system_, "pm:info"}, process_list{process_list_} { static const FunctionInfo functions[] = { {0, &Info::GetTitleId, "GetTitleId"}, }; @@ -159,8 +161,8 @@ private: class Shell final : public ServiceFramework<Shell> { public: - explicit Shell(const Kernel::KernelCore& kernel) - : ServiceFramework{"pm:shell"}, kernel(kernel) { + explicit Shell(Core::System& system_) + : ServiceFramework{system_, "pm:shell"}, kernel{system_.Kernel()} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "LaunchProgram"}, @@ -189,11 +191,11 @@ private: }; void InstallInterfaces(Core::System& system) { - std::make_shared<BootMode>()->InstallAsService(system.ServiceManager()); - std::make_shared<DebugMonitor>(system.Kernel())->InstallAsService(system.ServiceManager()); - std::make_shared<Info>(system.Kernel().GetProcessList()) + std::make_shared<BootMode>(system)->InstallAsService(system.ServiceManager()); + std::make_shared<DebugMonitor>(system)->InstallAsService(system.ServiceManager()); + std::make_shared<Info>(system, system.Kernel().GetProcessList()) ->InstallAsService(system.ServiceManager()); - std::make_shared<Shell>(system.Kernel())->InstallAsService(system.ServiceManager()); + std::make_shared<Shell>(system)->InstallAsService(system.ServiceManager()); } } // namespace Service::PM diff --git a/src/core/hle/service/prepo/prepo.cpp b/src/core/hle/service/prepo/prepo.cpp index cde3312da..b417624c9 100644 --- a/src/core/hle/service/prepo/prepo.cpp +++ b/src/core/hle/service/prepo/prepo.cpp @@ -4,6 +4,7 @@ #include "common/hex_util.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/process.h" #include "core/hle/service/acc/profile_manager.h" @@ -15,8 +16,7 @@ namespace Service::PlayReport { class PlayReport final : public ServiceFramework<PlayReport> { public: - explicit PlayReport(const char* name, Core::System& system) - : ServiceFramework{name}, system(system) { + explicit PlayReport(const char* name, Core::System& system_) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { {10100, &PlayReport::SaveReport<Core::Reporter::PlayReportType::Old>, "SaveReportOld"}, @@ -65,7 +65,7 @@ private: } LOG_DEBUG(Service_PREPO, "called, type={:02X}, process_id={:016X}, data1_size={:016X}", - static_cast<u8>(Type), process_id, data[0].size()); + Type, process_id, data[0].size()); const auto& reporter{system.GetReporter()}; reporter.SavePlayReport(Type, system.CurrentProcess()->GetTitleID(), data, process_id); @@ -92,7 +92,7 @@ private: LOG_DEBUG( Service_PREPO, "called, type={:02X}, user_id={:016X}{:016X}, process_id={:016X}, data1_size={:016X}", - static_cast<u8>(Type), user_id[1], user_id[0], process_id, data[0].size()); + Type, user_id[1], user_id[0], process_id, data[0].size()); const auto& reporter{system.GetReporter()}; reporter.SavePlayReport(Type, system.CurrentProcess()->GetTitleID(), data, process_id, @@ -139,8 +139,6 @@ private: IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); } - - Core::System& system; }; void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { diff --git a/src/core/hle/service/prepo/prepo.h b/src/core/hle/service/prepo/prepo.h index a5682ee26..395b57ead 100644 --- a/src/core/hle/service/prepo/prepo.h +++ b/src/core/hle/service/prepo/prepo.h @@ -4,14 +4,14 @@ #pragma once -namespace Service::SM { -class ServiceManager; -} - namespace Core { class System; } +namespace Service::SM { +class ServiceManager; +} + namespace Service::PlayReport { void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); diff --git a/src/core/hle/service/psc/psc.cpp b/src/core/hle/service/psc/psc.cpp index 99e1c9042..5a52b2b05 100644 --- a/src/core/hle/service/psc/psc.cpp +++ b/src/core/hle/service/psc/psc.cpp @@ -14,7 +14,7 @@ namespace Service::PSC { class PSC_C final : public ServiceFramework<PSC_C> { public: - explicit PSC_C() : ServiceFramework{"psc:c"} { + explicit PSC_C(Core::System& system_) : ServiceFramework{system_, "psc:c"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -35,7 +35,7 @@ public: class IPmModule final : public ServiceFramework<IPmModule> { public: - explicit IPmModule() : ServiceFramework{"IPmModule"} { + explicit IPmModule(Core::System& system_) : ServiceFramework{system_, "IPmModule"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -52,7 +52,7 @@ public: class PSC_M final : public ServiceFramework<PSC_M> { public: - explicit PSC_M() : ServiceFramework{"psc:m"} { + explicit PSC_M(Core::System& system_) : ServiceFramework{system_, "psc:m"} { // clang-format off static const FunctionInfo functions[] = { {0, &PSC_M::GetPmModule, "GetPmModule"}, @@ -68,13 +68,13 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IPmModule>(); + rb.PushIpcInterface<IPmModule>(system); } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<PSC_C>()->InstallAsService(sm); - std::make_shared<PSC_M>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<PSC_C>(system)->InstallAsService(sm); + std::make_shared<PSC_M>(system)->InstallAsService(sm); } } // namespace Service::PSC diff --git a/src/core/hle/service/psc/psc.h b/src/core/hle/service/psc/psc.h index 5052eb02c..89344f32d 100644 --- a/src/core/hle/service/psc/psc.h +++ b/src/core/hle/service/psc/psc.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::PSC { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::PSC diff --git a/src/core/hle/service/ptm/psm.cpp b/src/core/hle/service/ptm/psm.cpp index 6d9e6bd09..b4b0dd241 100644 --- a/src/core/hle/service/ptm/psm.cpp +++ b/src/core/hle/service/ptm/psm.cpp @@ -14,7 +14,7 @@ namespace Service::PSM { class PSM final : public ServiceFramework<PSM> { public: - explicit PSM() : ServiceFramework{"psm"} { + explicit PSM(Core::System& system_) : ServiceFramework{system_, "psm"} { // clang-format off static const FunctionInfo functions[] = { {0, &PSM::GetBatteryChargePercentage, "GetBatteryChargePercentage"}, @@ -72,8 +72,8 @@ private: ChargerType charger_type{ChargerType::RegularCharger}; }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<PSM>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<PSM>(system)->InstallAsService(sm); } } // namespace Service::PSM diff --git a/src/core/hle/service/ptm/psm.h b/src/core/hle/service/ptm/psm.h index a286793ae..2930ce26a 100644 --- a/src/core/hle/service/ptm/psm.h +++ b/src/core/hle/service/ptm/psm.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::PSM { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::PSM diff --git a/src/core/hle/service/service.cpp b/src/core/hle/service/service.cpp index 76b3533ec..ff2a5b1db 100644 --- a/src/core/hle/service/service.cpp +++ b/src/core/hle/service/service.cpp @@ -51,6 +51,7 @@ #include "core/hle/service/ns/ns.h" #include "core/hle/service/nvdrv/nvdrv.h" #include "core/hle/service/nvflinger/nvflinger.h" +#include "core/hle/service/olsc/olsc.h" #include "core/hle/service/pcie/pcie.h" #include "core/hle/service/pctl/module.h" #include "core/hle/service/pcv/pcv.h" @@ -72,13 +73,36 @@ namespace Service { -ServiceFrameworkBase::ServiceFrameworkBase(const char* service_name, u32 max_sessions, - InvokerFn* handler_invoker) - : service_name(service_name), max_sessions(max_sessions), handler_invoker(handler_invoker) {} +/** + * Creates a function string for logging, complete with the name (or header code, depending + * on what's passed in) the port name, and all the cmd_buff arguments. + */ +[[maybe_unused]] static std::string MakeFunctionString(std::string_view name, + std::string_view port_name, + const u32* cmd_buff) { + // Number of params == bits 0-5 + bits 6-11 + int num_params = (cmd_buff[0] & 0x3F) + ((cmd_buff[0] >> 6) & 0x3F); -ServiceFrameworkBase::~ServiceFrameworkBase() = default; + std::string function_string = fmt::format("function '{}': port={}", name, port_name); + for (int i = 1; i <= num_params; ++i) { + function_string += fmt::format(", cmd_buff[{}]=0x{:X}", i, cmd_buff[i]); + } + return function_string; +} + +ServiceFrameworkBase::ServiceFrameworkBase(Core::System& system_, const char* service_name_, + u32 max_sessions_, InvokerFn* handler_invoker_) + : system{system_}, service_name{service_name_}, max_sessions{max_sessions_}, + handler_invoker{handler_invoker_} {} + +ServiceFrameworkBase::~ServiceFrameworkBase() { + // Wait for other threads to release access before destroying + const auto guard = LockService(); +} void ServiceFrameworkBase::InstallAsService(SM::ServiceManager& service_manager) { + const auto guard = LockService(); + ASSERT(!port_installed); auto port = service_manager.RegisterService(service_name, max_sessions).Unwrap(); @@ -87,6 +111,8 @@ void ServiceFrameworkBase::InstallAsService(SM::ServiceManager& service_manager) } void ServiceFrameworkBase::InstallAsNamedPort(Kernel::KernelCore& kernel) { + const auto guard = LockService(); + ASSERT(!port_installed); auto [server_port, client_port] = @@ -96,17 +122,6 @@ void ServiceFrameworkBase::InstallAsNamedPort(Kernel::KernelCore& kernel) { port_installed = true; } -std::shared_ptr<Kernel::ClientPort> ServiceFrameworkBase::CreatePort(Kernel::KernelCore& kernel) { - ASSERT(!port_installed); - - auto [server_port, client_port] = - Kernel::ServerPort::CreatePortPair(kernel, max_sessions, service_name); - auto port = MakeResult(std::move(server_port)).Unwrap(); - port->SetHleHandler(shared_from_this()); - port_installed = true; - return client_port; -} - void ServiceFrameworkBase::RegisterHandlersBase(const FunctionInfoBase* functions, std::size_t n) { handlers.reserve(handlers.size() + n); for (std::size_t i = 0; i < n; ++i) { @@ -128,8 +143,8 @@ void ServiceFrameworkBase::ReportUnimplementedFunction(Kernel::HLERequestContext } buf.push_back('}'); - Core::System::GetInstance().GetReporter().SaveUnimplementedFunctionReport( - ctx, ctx.GetCommand(), function_name, service_name); + system.GetReporter().SaveUnimplementedFunctionReport(ctx, ctx.GetCommand(), function_name, + service_name); UNIMPLEMENTED_MSG("Unknown / unimplemented {}", fmt::to_string(buf)); } @@ -145,6 +160,8 @@ void ServiceFrameworkBase::InvokeRequest(Kernel::HLERequestContext& ctx) { } ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& context) { + const auto guard = LockService(); + switch (context.GetCommandType()) { case IPC::CommandType::Close: { IPC::ResponseBuilder rb{context, 2}; @@ -153,7 +170,7 @@ ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& co } case IPC::CommandType::ControlWithContext: case IPC::CommandType::Control: { - Core::System::GetInstance().ServiceManager().InvokeControlRequest(context); + system.ServiceManager().InvokeControlRequest(context); break; } case IPC::CommandType::RequestWithContext: @@ -162,79 +179,82 @@ ResultCode ServiceFrameworkBase::HandleSyncRequest(Kernel::HLERequestContext& co break; } default: - UNIMPLEMENTED_MSG("command_type={}", static_cast<int>(context.GetCommandType())); + UNIMPLEMENTED_MSG("command_type={}", context.GetCommandType()); } - context.WriteToOutgoingCommandBuffer(context.GetThread()); + // If emulation was shutdown, we are closing service threads, do not write the response back to + // memory that may be shutting down as well. + if (system.IsPoweredOn()) { + context.WriteToOutgoingCommandBuffer(context.GetThread()); + } return RESULT_SUCCESS; } -/// Initialize ServiceManager -void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system) { +/// Initialize Services +Services::Services(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system) + : nv_flinger{std::make_unique<NVFlinger::NVFlinger>(system)} { + // NVFlinger needs to be accessed by several services like Vi and AppletOE so we instantiate it // here and pass it into the respective InstallInterfaces functions. - auto nv_flinger = std::make_shared<NVFlinger::NVFlinger>(system); + system.GetFileSystemController().CreateFactories(*system.GetFilesystem(), false); - SM::ServiceManager::InstallInterfaces(sm, system.Kernel()); + SM::ServiceManager::InstallInterfaces(sm, system); Account::InstallInterfaces(system); - AM::InstallInterfaces(*sm, nv_flinger, system); + AM::InstallInterfaces(*sm, *nv_flinger, system); AOC::InstallInterfaces(*sm, system); APM::InstallInterfaces(system); Audio::InstallInterfaces(*sm, system); BCAT::InstallInterfaces(system); - BPC::InstallInterfaces(*sm); + BPC::InstallInterfaces(*sm, system); BtDrv::InstallInterfaces(*sm, system); BTM::InstallInterfaces(*sm, system); - Capture::InstallInterfaces(*sm); - ERPT::InstallInterfaces(*sm); - ES::InstallInterfaces(*sm); - EUPLD::InstallInterfaces(*sm); + Capture::InstallInterfaces(*sm, system); + ERPT::InstallInterfaces(*sm, system); + ES::InstallInterfaces(*sm, system); + EUPLD::InstallInterfaces(*sm, system); Fatal::InstallInterfaces(*sm, system); - FGM::InstallInterfaces(*sm); + FGM::InstallInterfaces(*sm, system); FileSystem::InstallInterfaces(system); Friend::InstallInterfaces(*sm, system); Glue::InstallInterfaces(system); - GRC::InstallInterfaces(*sm); + GRC::InstallInterfaces(*sm, system); HID::InstallInterfaces(*sm, system); - LBL::InstallInterfaces(*sm); - LDN::InstallInterfaces(*sm); + LBL::InstallInterfaces(*sm, system); + LDN::InstallInterfaces(*sm, system); LDR::InstallInterfaces(*sm, system); LM::InstallInterfaces(system); - Migration::InstallInterfaces(*sm); - Mii::InstallInterfaces(*sm); - MM::InstallInterfaces(*sm); - NCM::InstallInterfaces(*sm); - NFC::InstallInterfaces(*sm); + Migration::InstallInterfaces(*sm, system); + Mii::InstallInterfaces(*sm, system); + MM::InstallInterfaces(*sm, system); + NCM::InstallInterfaces(*sm, system); + NFC::InstallInterfaces(*sm, system); NFP::InstallInterfaces(*sm, system); NIFM::InstallInterfaces(*sm, system); NIM::InstallInterfaces(*sm, system); - NPNS::InstallInterfaces(*sm); + NPNS::InstallInterfaces(*sm, system); NS::InstallInterfaces(*sm, system); Nvidia::InstallInterfaces(*sm, *nv_flinger, system); - PCIe::InstallInterfaces(*sm); - PCTL::InstallInterfaces(*sm); - PCV::InstallInterfaces(*sm); + OLSC::InstallInterfaces(*sm, system); + PCIe::InstallInterfaces(*sm, system); + PCTL::InstallInterfaces(*sm, system); + PCV::InstallInterfaces(*sm, system); PlayReport::InstallInterfaces(*sm, system); PM::InstallInterfaces(system); - PSC::InstallInterfaces(*sm); - PSM::InstallInterfaces(*sm); - Set::InstallInterfaces(*sm); + PSC::InstallInterfaces(*sm, system); + PSM::InstallInterfaces(*sm, system); + Set::InstallInterfaces(*sm, system); Sockets::InstallInterfaces(*sm, system); - SPL::InstallInterfaces(*sm); - SSL::InstallInterfaces(*sm); + SPL::InstallInterfaces(*sm, system); + SSL::InstallInterfaces(*sm, system); Time::InstallInterfaces(system); - USB::InstallInterfaces(*sm); - VI::InstallInterfaces(*sm, nv_flinger); - WLAN::InstallInterfaces(*sm); - - LOG_DEBUG(Service, "initialized OK"); + USB::InstallInterfaces(*sm, system); + VI::InstallInterfaces(*sm, system, *nv_flinger); + WLAN::InstallInterfaces(*sm, system); } -/// Shutdown ServiceManager -void Shutdown() { - LOG_DEBUG(Service, "shutdown OK"); -} +Services::~Services() = default; + } // namespace Service diff --git a/src/core/hle/service/service.h b/src/core/hle/service/service.h index a01ef3353..916445517 100644 --- a/src/core/hle/service/service.h +++ b/src/core/hle/service/service.h @@ -5,9 +5,11 @@ #pragma once #include <cstddef> +#include <mutex> #include <string> #include <boost/container/flat_map.hpp> #include "common/common_types.h" +#include "common/spin_lock.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/kernel/object.h" @@ -29,7 +31,11 @@ namespace Service { namespace FileSystem { class FileSystemController; -} // namespace FileSystem +} + +namespace NVFlinger { +class NVFlinger; +} namespace SM { class ServiceManager; @@ -64,11 +70,9 @@ public: void InstallAsService(SM::ServiceManager& service_manager); /// Creates a port pair and registers it on the kernel's global port registry. void InstallAsNamedPort(Kernel::KernelCore& kernel); - /// Creates and returns an unregistered port for the service. - std::shared_ptr<Kernel::ClientPort> CreatePort(Kernel::KernelCore& kernel); - + /// Invokes a service request routine. void InvokeRequest(Kernel::HLERequestContext& ctx); - + /// Handles a synchronization request for the service. ResultCode HandleSyncRequest(Kernel::HLERequestContext& context) override; protected: @@ -76,6 +80,14 @@ protected: template <typename Self> using HandlerFnP = void (Self::*)(Kernel::HLERequestContext&); + /// Used to gain exclusive access to the service members, e.g. from CoreTiming thread. + [[nodiscard]] std::scoped_lock<Common::SpinLock> LockService() { + return std::scoped_lock{lock_service}; + } + + /// System context that the service operates under. + Core::System& system; + private: template <typename T> friend class ServiceFramework; @@ -89,7 +101,8 @@ private: using InvokerFn = void(ServiceFrameworkBase* object, HandlerFnP<ServiceFrameworkBase> member, Kernel::HLERequestContext& ctx); - ServiceFrameworkBase(const char* service_name, u32 max_sessions, InvokerFn* handler_invoker); + explicit ServiceFrameworkBase(Core::System& system_, const char* service_name_, + u32 max_sessions_, InvokerFn* handler_invoker_); ~ServiceFrameworkBase() override; void RegisterHandlersBase(const FunctionInfoBase* functions, std::size_t n); @@ -107,6 +120,9 @@ private: /// Function used to safely up-cast pointers to the derived class before invoking a handler. InvokerFn* handler_invoker; boost::container::flat_map<u32, FunctionInfoBase> handlers; + + /// Used to gain exclusive access to the service members, e.g. from CoreTiming thread. + Common::SpinLock lock_service; }; /** @@ -147,11 +163,15 @@ protected: /** * Initializes the handler with no functions installed. - * @param max_sessions Maximum number of sessions that can be - * connected to this service at the same time. + * + * @param system_ The system context to construct this service under. + * @param service_name_ Name of the service. + * @param max_sessions_ Maximum number of sessions that can be + * connected to this service at the same time. */ - explicit ServiceFramework(const char* service_name, u32 max_sessions = DefaultMaxSessions) - : ServiceFrameworkBase(service_name, max_sessions, Invoker) {} + explicit ServiceFramework(Core::System& system_, const char* service_name_, + u32 max_sessions_ = DefaultMaxSessions) + : ServiceFrameworkBase(system_, service_name_, max_sessions_, Invoker) {} /// Registers handlers in the service. template <std::size_t N> @@ -181,10 +201,17 @@ private: } }; -/// Initialize ServiceManager -void Init(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system); +/** + * The purpose of this class is to own any objects that need to be shared across the other service + * implementations. Will be torn down when the global system instance is shutdown. + */ +class Services final { +public: + explicit Services(std::shared_ptr<SM::ServiceManager>& sm, Core::System& system); + ~Services(); -/// Shutdown ServiceManager -void Shutdown(); +private: + std::unique_ptr<NVFlinger::NVFlinger> nv_flinger; +}; } // namespace Service diff --git a/src/core/hle/service/set/set.cpp b/src/core/hle/service/set/set.cpp index e64777668..d953b4303 100644 --- a/src/core/hle/service/set/set.cpp +++ b/src/core/hle/service/set/set.cpp @@ -188,7 +188,7 @@ void SET::GetKeyCodeMap2(Kernel::HLERequestContext& ctx) { GetKeyCodeMapImpl(ctx); } -SET::SET() : ServiceFramework("set") { +SET::SET(Core::System& system_) : ServiceFramework{system_, "set"} { // clang-format off static const FunctionInfo functions[] = { {0, &SET::GetLanguageCode, "GetLanguageCode"}, @@ -202,6 +202,7 @@ SET::SET() : ServiceFramework("set") { {8, &SET::GetQuestFlag, "GetQuestFlag"}, {9, &SET::GetKeyCodeMap2, "GetKeyCodeMap2"}, {10, nullptr, "GetFirmwareVersionForDebug"}, + {11, nullptr, "GetDeviceNickName"}, }; // clang-format on diff --git a/src/core/hle/service/set/set.h b/src/core/hle/service/set/set.h index 8ac9c169d..d5bd7828d 100644 --- a/src/core/hle/service/set/set.h +++ b/src/core/hle/service/set/set.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Set { /// This is "nn::settings::LanguageCode", which is a NUL-terminated string stored in a u64. @@ -32,7 +36,7 @@ LanguageCode GetLanguageCodeFromIndex(std::size_t idx); class SET final : public ServiceFramework<SET> { public: - explicit SET(); + explicit SET(Core::System& system_); ~SET() override; private: diff --git a/src/core/hle/service/set/set_cal.cpp b/src/core/hle/service/set/set_cal.cpp index 3fbfecc9e..b2aa7bc0c 100644 --- a/src/core/hle/service/set/set_cal.cpp +++ b/src/core/hle/service/set/set_cal.cpp @@ -6,7 +6,7 @@ namespace Service::Set { -SET_CAL::SET_CAL() : ServiceFramework("set:cal") { +SET_CAL::SET_CAL(Core::System& system_) : ServiceFramework{system_, "set:cal"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetBluetoothBdAddress"}, diff --git a/src/core/hle/service/set/set_cal.h b/src/core/hle/service/set/set_cal.h index a0677e815..a29fc3ddd 100644 --- a/src/core/hle/service/set/set_cal.h +++ b/src/core/hle/service/set/set_cal.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Set { class SET_CAL final : public ServiceFramework<SET_CAL> { public: - explicit SET_CAL(); + explicit SET_CAL(Core::System& system_); ~SET_CAL() override; }; diff --git a/src/core/hle/service/set/set_fd.cpp b/src/core/hle/service/set/set_fd.cpp index 565882a31..f04dc5047 100644 --- a/src/core/hle/service/set/set_fd.cpp +++ b/src/core/hle/service/set/set_fd.cpp @@ -6,7 +6,7 @@ namespace Service::Set { -SET_FD::SET_FD() : ServiceFramework("set:fd") { +SET_FD::SET_FD(Core::System& system_) : ServiceFramework{system_, "set:fd"} { // clang-format off static const FunctionInfo functions[] = { {2, nullptr, "SetSettingsItemValue"}, diff --git a/src/core/hle/service/set/set_fd.h b/src/core/hle/service/set/set_fd.h index 216e65f1f..c28cb301e 100644 --- a/src/core/hle/service/set/set_fd.h +++ b/src/core/hle/service/set/set_fd.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Set { class SET_FD final : public ServiceFramework<SET_FD> { public: - explicit SET_FD(); + explicit SET_FD(Core::System& system_); ~SET_FD() override; }; diff --git a/src/core/hle/service/set/set_sys.cpp b/src/core/hle/service/set/set_sys.cpp index 8bd4c7e79..b58b2c8c5 100644 --- a/src/core/hle/service/set/set_sys.cpp +++ b/src/core/hle/service/set/set_sys.cpp @@ -34,9 +34,9 @@ void GetFirmwareVersionImpl(Kernel::HLERequestContext& ctx, GetFirmwareVersionTy // consistence (currently reports as 5.1.0-0.0) const auto archive = FileSys::SystemArchive::SystemVersion(); - const auto early_exit_failure = [&ctx](const std::string& desc, ResultCode code) { + const auto early_exit_failure = [&ctx](std::string_view desc, ResultCode code) { LOG_ERROR(Service_SET, "General failure while attempting to resolve firmware version ({}).", - desc.c_str()); + desc); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(code); }; @@ -103,7 +103,7 @@ void SET_SYS::SetColorSetId(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); } -SET_SYS::SET_SYS() : ServiceFramework("set:sys") { +SET_SYS::SET_SYS(Core::System& system_) : ServiceFramework{system_, "set:sys"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SetLanguageCode"}, @@ -300,6 +300,8 @@ SET_SYS::SET_SYS() : ServiceFramework("set:sys") { {198, nullptr, "SetButtonConfigRegisteredSettingsEmbedded"}, {199, nullptr, "GetButtonConfigRegisteredSettings"}, {200, nullptr, "SetButtonConfigRegisteredSettings"}, + {201, nullptr, "GetFieldTestingFlag"}, + {202, nullptr, "SetFieldTestingFlag"}, }; // clang-format on diff --git a/src/core/hle/service/set/set_sys.h b/src/core/hle/service/set/set_sys.h index 13ee2cf46..edb185a68 100644 --- a/src/core/hle/service/set/set_sys.h +++ b/src/core/hle/service/set/set_sys.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Set { class SET_SYS final : public ServiceFramework<SET_SYS> { public: - explicit SET_SYS(); + explicit SET_SYS(Core::System& system_); ~SET_SYS() override; private: diff --git a/src/core/hle/service/set/settings.cpp b/src/core/hle/service/set/settings.cpp index cf5541ca8..212ebc427 100644 --- a/src/core/hle/service/set/settings.cpp +++ b/src/core/hle/service/set/settings.cpp @@ -7,14 +7,15 @@ #include "core/hle/service/set/set_fd.h" #include "core/hle/service/set/set_sys.h" #include "core/hle/service/set/settings.h" +#include "core/hle/service/sm/sm.h" namespace Service::Set { -void InstallInterfaces(SM::ServiceManager& service_manager) { - std::make_shared<SET>()->InstallAsService(service_manager); - std::make_shared<SET_CAL>()->InstallAsService(service_manager); - std::make_shared<SET_FD>()->InstallAsService(service_manager); - std::make_shared<SET_SYS>()->InstallAsService(service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { + std::make_shared<SET>(system)->InstallAsService(service_manager); + std::make_shared<SET_CAL>(system)->InstallAsService(service_manager); + std::make_shared<SET_FD>(system)->InstallAsService(service_manager); + std::make_shared<SET_SYS>(system)->InstallAsService(service_manager); } } // namespace Service::Set diff --git a/src/core/hle/service/set/settings.h b/src/core/hle/service/set/settings.h index 6606ce776..7a6950dd0 100644 --- a/src/core/hle/service/set/settings.h +++ b/src/core/hle/service/set/settings.h @@ -4,11 +4,17 @@ #pragma once -#include "core/hle/service/service.h" +namespace Core { +class System; +} + +namespace Service::SM { +class ServiceManager; +} namespace Service::Set { /// Registers all Settings services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); } // namespace Service::Set diff --git a/src/core/hle/service/sm/controller.cpp b/src/core/hle/service/sm/controller.cpp index 972aaa6d9..916177efd 100644 --- a/src/core/hle/service/sm/controller.cpp +++ b/src/core/hle/service/sm/controller.cpp @@ -48,7 +48,7 @@ void Controller::QueryPointerBufferSize(Kernel::HLERequestContext& ctx) { } // https://switchbrew.org/wiki/IPC_Marshalling -Controller::Controller() : ServiceFramework("IpcController") { +Controller::Controller(Core::System& system_) : ServiceFramework{system_, "IpcController"} { static const FunctionInfo functions[] = { {0, &Controller::ConvertCurrentObjectToDomain, "ConvertCurrentObjectToDomain"}, {1, nullptr, "CopyFromCurrentDomain"}, diff --git a/src/core/hle/service/sm/controller.h b/src/core/hle/service/sm/controller.h index 180c6da50..7494f898d 100644 --- a/src/core/hle/service/sm/controller.h +++ b/src/core/hle/service/sm/controller.h @@ -6,11 +6,15 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::SM { class Controller final : public ServiceFramework<Controller> { public: - Controller(); + explicit Controller(Core::System& system_); ~Controller() override; private: diff --git a/src/core/hle/service/sm/sm.cpp b/src/core/hle/service/sm/sm.cpp index 9c1da361b..4da69f503 100644 --- a/src/core/hle/service/sm/sm.cpp +++ b/src/core/hle/service/sm/sm.cpp @@ -38,14 +38,13 @@ static ResultCode ValidateServiceName(const std::string& name) { return RESULT_SUCCESS; } -void ServiceManager::InstallInterfaces(std::shared_ptr<ServiceManager> self, - Kernel::KernelCore& kernel) { +void ServiceManager::InstallInterfaces(std::shared_ptr<ServiceManager> self, Core::System& system) { ASSERT(self->sm_interface.expired()); - auto sm = std::make_shared<SM>(self, kernel); - sm->InstallAsNamedPort(kernel); + auto sm = std::make_shared<SM>(self, system); + sm->InstallAsNamedPort(system.Kernel()); self->sm_interface = sm; - self->controller_interface = std::make_unique<Controller>(); + self->controller_interface = std::make_unique<Controller>(system); } ResultVal<std::shared_ptr<Kernel::ServerPort>> ServiceManager::RegisterService(std::string name, @@ -190,8 +189,9 @@ void SM::UnregisterService(Kernel::HLERequestContext& ctx) { rb.Push(service_manager->UnregisterService(name)); } -SM::SM(std::shared_ptr<ServiceManager> service_manager, Kernel::KernelCore& kernel) - : ServiceFramework{"sm:", 4}, service_manager{std::move(service_manager)}, kernel{kernel} { +SM::SM(std::shared_ptr<ServiceManager> service_manager_, Core::System& system_) + : ServiceFramework{system_, "sm:", 4}, + service_manager{std::move(service_manager_)}, kernel{system_.Kernel()} { static const FunctionInfo functions[] = { {0x00000000, &SM::Initialize, "Initialize"}, {0x00000001, &SM::GetService, "GetService"}, diff --git a/src/core/hle/service/sm/sm.h b/src/core/hle/service/sm/sm.h index 6790c86f0..3f46ae44f 100644 --- a/src/core/hle/service/sm/sm.h +++ b/src/core/hle/service/sm/sm.h @@ -16,6 +16,10 @@ #include "core/hle/result.h" #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class ClientPort; class ClientSession; @@ -31,7 +35,7 @@ class Controller; /// Interface to "sm:" service class SM final : public ServiceFramework<SM> { public: - explicit SM(std::shared_ptr<ServiceManager> service_manager, Kernel::KernelCore& kernel); + explicit SM(std::shared_ptr<ServiceManager> service_manager_, Core::System& system_); ~SM() override; private: @@ -46,7 +50,7 @@ private: class ServiceManager { public: - static void InstallInterfaces(std::shared_ptr<ServiceManager> self, Kernel::KernelCore& kernel); + static void InstallInterfaces(std::shared_ptr<ServiceManager> self, Core::System& system); explicit ServiceManager(Kernel::KernelCore& kernel_); ~ServiceManager(); diff --git a/src/core/hle/service/sockets/blocking_worker.h b/src/core/hle/service/sockets/blocking_worker.h deleted file mode 100644 index 2d53e52b6..000000000 --- a/src/core/hle/service/sockets/blocking_worker.h +++ /dev/null @@ -1,161 +0,0 @@ -// Copyright 2020 yuzu emulator team -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <atomic> -#include <memory> -#include <string> -#include <string_view> -#include <thread> -#include <variant> -#include <vector> - -#include <fmt/format.h> - -#include "common/assert.h" -#include "common/microprofile.h" -#include "common/thread.h" -#include "core/core.h" -#include "core/hle/kernel/hle_ipc.h" -#include "core/hle/kernel/kernel.h" -#include "core/hle/kernel/thread.h" -#include "core/hle/kernel/writable_event.h" - -namespace Service::Sockets { - -/** - * Worker abstraction to execute blocking calls on host without blocking the guest thread - * - * @tparam Service Service where the work is executed - * @tparam Types Types of work to execute - */ -template <class Service, class... Types> -class BlockingWorker { - using This = BlockingWorker<Service, Types...>; - using WorkVariant = std::variant<std::monostate, Types...>; - -public: - /// Create a new worker - static std::unique_ptr<This> Create(Core::System& system, Service* service, - std::string_view name) { - return std::unique_ptr<This>(new This(system, service, name)); - } - - ~BlockingWorker() { - while (!is_available.load(std::memory_order_relaxed)) { - // Busy wait until work is finished - std::this_thread::yield(); - } - // Monostate means to exit the thread - work = std::monostate{}; - work_event.Set(); - thread.join(); - } - - /** - * Try to capture the worker to send work after a success - * @returns True when the worker has been successfully captured - */ - bool TryCapture() { - bool expected = true; - return is_available.compare_exchange_weak(expected, false, std::memory_order_relaxed, - std::memory_order_relaxed); - } - - /** - * Send work to this worker abstraction - * @see TryCapture must be called before attempting to call this function - */ - template <class Work> - void SendWork(Work new_work) { - ASSERT_MSG(!is_available, "Trying to send work on a worker that's not captured"); - work = std::move(new_work); - work_event.Set(); - } - - /// Generate a callback for @see SleepClientThread - template <class Work> - auto Callback() { - return [this](std::shared_ptr<Kernel::Thread>, Kernel::HLERequestContext& ctx, - Kernel::ThreadWakeupReason reason) { - ASSERT(reason == Kernel::ThreadWakeupReason::Signal); - std::get<Work>(work).Response(ctx); - is_available.store(true); - }; - } - - /// Get kernel event that will be signalled by the worker when the host operation finishes - std::shared_ptr<Kernel::WritableEvent> KernelEvent() const { - return kernel_event; - } - -private: - explicit BlockingWorker(Core::System& system, Service* service, std::string_view name) { - auto pair = Kernel::WritableEvent::CreateEventPair(system.Kernel(), std::string(name)); - kernel_event = std::move(pair.writable); - thread = std::thread([this, &system, service, name] { Run(system, service, name); }); - } - - void Run(Core::System& system, Service* service, std::string_view name) { - system.RegisterHostThread(); - - const std::string thread_name = fmt::format("yuzu:{}", name); - MicroProfileOnThreadCreate(thread_name.c_str()); - Common::SetCurrentThreadName(thread_name.c_str()); - - bool keep_running = true; - while (keep_running) { - work_event.Wait(); - - const auto visit_fn = [service, &keep_running]<typename T>(T&& w) { - if constexpr (std::is_same_v<std::decay_t<T>, std::monostate>) { - keep_running = false; - } else { - w.Execute(service); - } - }; - std::visit(visit_fn, work); - - kernel_event->Signal(); - } - } - - std::thread thread; - WorkVariant work; - Common::Event work_event; - std::shared_ptr<Kernel::WritableEvent> kernel_event; - std::atomic_bool is_available{true}; -}; - -template <class Service, class... Types> -class BlockingWorkerPool { - using Worker = BlockingWorker<Service, Types...>; - -public: - explicit BlockingWorkerPool(Core::System& system_, Service* service_) - : system{system_}, service{service_} {} - - /// Returns a captured worker thread, creating new ones if necessary - Worker* CaptureWorker() { - for (auto& worker : workers) { - if (worker->TryCapture()) { - return worker.get(); - } - } - auto new_worker = Worker::Create(system, service, fmt::format("BSD:{}", workers.size())); - [[maybe_unused]] const bool success = new_worker->TryCapture(); - ASSERT(success); - - return workers.emplace_back(std::move(new_worker)).get(); - } - -private: - Core::System& system; - Service* const service; - - std::vector<std::unique_ptr<Worker>> workers; -}; - -} // namespace Service::Sockets diff --git a/src/core/hle/service/sockets/bsd.cpp b/src/core/hle/service/sockets/bsd.cpp index a74be9370..2b824059d 100644 --- a/src/core/hle/service/sockets/bsd.cpp +++ b/src/core/hle/service/sockets/bsd.cpp @@ -30,7 +30,7 @@ bool IsConnectionBased(Type type) { case Type::DGRAM: return false; default: - UNIMPLEMENTED_MSG("Unimplemented type={}", static_cast<int>(type)); + UNIMPLEMENTED_MSG("Unimplemented type={}", type); return false; } } @@ -178,13 +178,12 @@ void BSD::Poll(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. nfds={} timeout={}", nfds, timeout); - ExecuteWork(ctx, "BSD:Poll", timeout != 0, - PollWork{ - .nfds = nfds, - .timeout = timeout, - .read_buffer = ctx.ReadBuffer(), - .write_buffer = std::vector<u8>(ctx.GetWriteBufferSize()), - }); + ExecuteWork(ctx, PollWork{ + .nfds = nfds, + .timeout = timeout, + .read_buffer = ctx.ReadBuffer(), + .write_buffer = std::vector<u8>(ctx.GetWriteBufferSize()), + }); } void BSD::Accept(Kernel::HLERequestContext& ctx) { @@ -193,11 +192,10 @@ void BSD::Accept(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. fd={}", fd); - ExecuteWork(ctx, "BSD:Accept", IsBlockingSocket(fd), - AcceptWork{ - .fd = fd, - .write_buffer = std::vector<u8>(ctx.GetWriteBufferSize()), - }); + ExecuteWork(ctx, AcceptWork{ + .fd = fd, + .write_buffer = std::vector<u8>(ctx.GetWriteBufferSize()), + }); } void BSD::Bind(Kernel::HLERequestContext& ctx) { @@ -215,11 +213,10 @@ void BSD::Connect(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. fd={} addrlen={}", fd, ctx.GetReadBufferSize()); - ExecuteWork(ctx, "BSD:Connect", IsBlockingSocket(fd), - ConnectWork{ - .fd = fd, - .addr = ctx.ReadBuffer(), - }); + ExecuteWork(ctx, ConnectWork{ + .fd = fd, + .addr = ctx.ReadBuffer(), + }); } void BSD::GetPeerName(Kernel::HLERequestContext& ctx) { @@ -327,12 +324,11 @@ void BSD::Recv(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. fd={} flags=0x{:x} len={}", fd, flags, ctx.GetWriteBufferSize()); - ExecuteWork(ctx, "BSD:Recv", IsBlockingSocket(fd), - RecvWork{ - .fd = fd, - .flags = flags, - .message = std::vector<u8>(ctx.GetWriteBufferSize()), - }); + ExecuteWork(ctx, RecvWork{ + .fd = fd, + .flags = flags, + .message = std::vector<u8>(ctx.GetWriteBufferSize()), + }); } void BSD::RecvFrom(Kernel::HLERequestContext& ctx) { @@ -344,13 +340,12 @@ void BSD::RecvFrom(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. fd={} flags=0x{:x} len={} addrlen={}", fd, flags, ctx.GetWriteBufferSize(0), ctx.GetWriteBufferSize(1)); - ExecuteWork(ctx, "BSD:RecvFrom", IsBlockingSocket(fd), - RecvFromWork{ - .fd = fd, - .flags = flags, - .message = std::vector<u8>(ctx.GetWriteBufferSize(0)), - .addr = std::vector<u8>(ctx.GetWriteBufferSize(1)), - }); + ExecuteWork(ctx, RecvFromWork{ + .fd = fd, + .flags = flags, + .message = std::vector<u8>(ctx.GetWriteBufferSize(0)), + .addr = std::vector<u8>(ctx.GetWriteBufferSize(1)), + }); } void BSD::Send(Kernel::HLERequestContext& ctx) { @@ -361,12 +356,11 @@ void BSD::Send(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. fd={} flags=0x{:x} len={}", fd, flags, ctx.GetReadBufferSize()); - ExecuteWork(ctx, "BSD:Send", IsBlockingSocket(fd), - SendWork{ - .fd = fd, - .flags = flags, - .message = ctx.ReadBuffer(), - }); + ExecuteWork(ctx, SendWork{ + .fd = fd, + .flags = flags, + .message = ctx.ReadBuffer(), + }); } void BSD::SendTo(Kernel::HLERequestContext& ctx) { @@ -377,13 +371,12 @@ void BSD::SendTo(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. fd={} flags=0x{} len={} addrlen={}", fd, flags, ctx.GetReadBufferSize(0), ctx.GetReadBufferSize(1)); - ExecuteWork(ctx, "BSD:SendTo", IsBlockingSocket(fd), - SendToWork{ - .fd = fd, - .flags = flags, - .message = ctx.ReadBuffer(0), - .addr = ctx.ReadBuffer(1), - }); + ExecuteWork(ctx, SendToWork{ + .fd = fd, + .flags = flags, + .message = ctx.ReadBuffer(0), + .addr = ctx.ReadBuffer(1), + }); } void BSD::Write(Kernel::HLERequestContext& ctx) { @@ -392,12 +385,11 @@ void BSD::Write(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service, "called. fd={} len={}", fd, ctx.GetReadBufferSize()); - ExecuteWork(ctx, "BSD:Write", IsBlockingSocket(fd), - SendWork{ - .fd = fd, - .flags = 0, - .message = ctx.ReadBuffer(), - }); + ExecuteWork(ctx, SendWork{ + .fd = fd, + .flags = 0, + .message = ctx.ReadBuffer(), + }); } void BSD::Close(Kernel::HLERequestContext& ctx) { @@ -410,24 +402,9 @@ void BSD::Close(Kernel::HLERequestContext& ctx) { } template <typename Work> -void BSD::ExecuteWork(Kernel::HLERequestContext& ctx, std::string_view sleep_reason, - bool is_blocking, Work work) { - if (!is_blocking) { - work.Execute(this); - work.Response(ctx); - return; - } - - // Signal a dummy response to make IPC validation happy - // This will be overwritten by the SleepClientThread callback +void BSD::ExecuteWork(Kernel::HLERequestContext& ctx, Work work) { + work.Execute(this); work.Response(ctx); - - auto worker = worker_pool.CaptureWorker(); - - ctx.SleepClientThread(std::string(sleep_reason), std::numeric_limits<u64>::max(), - worker->Callback<Work>(), worker->KernelEvent()); - - worker->SendWork(std::move(work)); } std::pair<s32, Errno> BSD::SocketImpl(Domain domain, Type type, Protocol protocol) { @@ -489,18 +466,18 @@ std::pair<s32, Errno> BSD::PollImpl(std::vector<u8>& write_buffer, std::vector<u } for (PollFD& pollfd : fds) { - ASSERT(pollfd.revents == 0); + ASSERT(False(pollfd.revents)); if (pollfd.fd > static_cast<s32>(MAX_FD) || pollfd.fd < 0) { LOG_ERROR(Service, "File descriptor handle={} is invalid", pollfd.fd); - pollfd.revents = 0; + pollfd.revents = PollEvents{}; return {0, Errno::SUCCESS}; } const std::optional<FileDescriptor>& descriptor = file_descriptors[pollfd.fd]; if (!descriptor) { LOG_ERROR(Service, "File descriptor handle={} is not allocated", pollfd.fd); - pollfd.revents = POLL_NVAL; + pollfd.revents = PollEvents::Nval; return {0, Errno::SUCCESS}; } } @@ -510,7 +487,7 @@ std::pair<s32, Errno> BSD::PollImpl(std::vector<u8>& write_buffer, std::vector<u Network::PollFD result; result.socket = file_descriptors[pollfd.fd]->socket.get(); result.events = TranslatePollEventsToHost(pollfd.events); - result.revents = 0; + result.revents = Network::PollEvents{}; return result; }); @@ -636,7 +613,7 @@ std::pair<s32, Errno> BSD::FcntlImpl(s32 fd, FcntlCmd cmd, s32 arg) { return {0, Errno::SUCCESS}; } default: - UNIMPLEMENTED_MSG("Unimplemented cmd={}", static_cast<int>(cmd)); + UNIMPLEMENTED_MSG("Unimplemented cmd={}", cmd); return {-1, Errno::SUCCESS}; } } @@ -679,7 +656,7 @@ Errno BSD::SetSockOptImpl(s32 fd, u32 level, OptName optname, size_t optlen, con case OptName::RCVTIMEO: return Translate(socket->SetRcvTimeo(value)); default: - UNIMPLEMENTED_MSG("Unimplemented optname={}", static_cast<int>(optname)); + UNIMPLEMENTED_MSG("Unimplemented optname={}", optname); return Errno::SUCCESS; } } @@ -807,18 +784,6 @@ bool BSD::IsFileDescriptorValid(s32 fd) const noexcept { return true; } -bool BSD::IsBlockingSocket(s32 fd) const noexcept { - // Inform invalid sockets as non-blocking - // This way we avoid using a worker thread as it will fail without blocking host - if (fd > static_cast<s32>(MAX_FD) || fd < 0) { - return false; - } - if (!file_descriptors[fd]) { - return false; - } - return (file_descriptors[fd]->flags & FLAG_O_NONBLOCK) != 0; -} - void BSD::BuildErrnoResponse(Kernel::HLERequestContext& ctx, Errno bsd_errno) const noexcept { IPC::ResponseBuilder rb{ctx, 4}; @@ -827,8 +792,7 @@ void BSD::BuildErrnoResponse(Kernel::HLERequestContext& ctx, Errno bsd_errno) co rb.PushEnum(bsd_errno); } -BSD::BSD(Core::System& system, const char* name) - : ServiceFramework(name), worker_pool{system, this} { +BSD::BSD(Core::System& system_, const char* name) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { {0, &BSD::RegisterClient, "RegisterClient"}, @@ -873,7 +837,7 @@ BSD::BSD(Core::System& system, const char* name) BSD::~BSD() = default; -BSDCFG::BSDCFG() : ServiceFramework{"bsdcfg"} { +BSDCFG::BSDCFG(Core::System& system_) : ServiceFramework{system_, "bsdcfg"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SetIfUp"}, diff --git a/src/core/hle/service/sockets/bsd.h b/src/core/hle/service/sockets/bsd.h index 357531951..6da0bfeb2 100644 --- a/src/core/hle/service/sockets/bsd.h +++ b/src/core/hle/service/sockets/bsd.h @@ -11,7 +11,6 @@ #include "common/common_types.h" #include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/service.h" -#include "core/hle/service/sockets/blocking_worker.h" #include "core/hle/service/sockets/sockets.h" namespace Core { @@ -26,7 +25,7 @@ namespace Service::Sockets { class BSD final : public ServiceFramework<BSD> { public: - explicit BSD(Core::System& system, const char* name); + explicit BSD(Core::System& system_, const char* name); ~BSD() override; private: @@ -138,8 +137,7 @@ private: void Close(Kernel::HLERequestContext& ctx); template <typename Work> - void ExecuteWork(Kernel::HLERequestContext& ctx, std::string_view sleep_reason, - bool is_blocking, Work work); + void ExecuteWork(Kernel::HLERequestContext& ctx, Work work); std::pair<s32, Errno> SocketImpl(Domain domain, Type type, Protocol protocol); std::pair<s32, Errno> PollImpl(std::vector<u8>& write_buffer, std::vector<u8> read_buffer, @@ -163,20 +161,15 @@ private: s32 FindFreeFileDescriptorHandle() noexcept; bool IsFileDescriptorValid(s32 fd) const noexcept; - bool IsBlockingSocket(s32 fd) const noexcept; void BuildErrnoResponse(Kernel::HLERequestContext& ctx, Errno bsd_errno) const noexcept; std::array<std::optional<FileDescriptor>, MAX_FD> file_descriptors; - - BlockingWorkerPool<BSD, PollWork, AcceptWork, ConnectWork, RecvWork, RecvFromWork, SendWork, - SendToWork> - worker_pool; }; class BSDCFG final : public ServiceFramework<BSDCFG> { public: - explicit BSDCFG(); + explicit BSDCFG(Core::System& system_); ~BSDCFG() override; }; diff --git a/src/core/hle/service/sockets/ethc.cpp b/src/core/hle/service/sockets/ethc.cpp index abbeb4c50..05681ca2d 100644 --- a/src/core/hle/service/sockets/ethc.cpp +++ b/src/core/hle/service/sockets/ethc.cpp @@ -6,7 +6,7 @@ namespace Service::Sockets { -ETHC_C::ETHC_C() : ServiceFramework{"ethc:c"} { +ETHC_C::ETHC_C(Core::System& system_) : ServiceFramework{system_, "ethc:c"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Initialize"}, @@ -23,7 +23,7 @@ ETHC_C::ETHC_C() : ServiceFramework{"ethc:c"} { ETHC_C::~ETHC_C() = default; -ETHC_I::ETHC_I() : ServiceFramework{"ethc:i"} { +ETHC_I::ETHC_I(Core::System& system_) : ServiceFramework{system_, "ethc:i"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetReadableHandle"}, diff --git a/src/core/hle/service/sockets/ethc.h b/src/core/hle/service/sockets/ethc.h index da2c7f741..71884182e 100644 --- a/src/core/hle/service/sockets/ethc.h +++ b/src/core/hle/service/sockets/ethc.h @@ -6,17 +6,21 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Sockets { class ETHC_C final : public ServiceFramework<ETHC_C> { public: - explicit ETHC_C(); + explicit ETHC_C(Core::System& system_); ~ETHC_C() override; }; class ETHC_I final : public ServiceFramework<ETHC_I> { public: - explicit ETHC_I(); + explicit ETHC_I(Core::System& system_); ~ETHC_I() override; }; diff --git a/src/core/hle/service/sockets/nsd.cpp b/src/core/hle/service/sockets/nsd.cpp index 40d781124..51c3739bb 100644 --- a/src/core/hle/service/sockets/nsd.cpp +++ b/src/core/hle/service/sockets/nsd.cpp @@ -6,7 +6,7 @@ namespace Service::Sockets { -NSD::NSD(const char* name) : ServiceFramework(name) { +NSD::NSD(Core::System& system_, const char* name) : ServiceFramework{system_, name} { // clang-format off static const FunctionInfo functions[] = { {10, nullptr, "GetSettingName"}, diff --git a/src/core/hle/service/sockets/nsd.h b/src/core/hle/service/sockets/nsd.h index d842e3232..becf93125 100644 --- a/src/core/hle/service/sockets/nsd.h +++ b/src/core/hle/service/sockets/nsd.h @@ -4,14 +4,17 @@ #pragma once -#include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Sockets { class NSD final : public ServiceFramework<NSD> { public: - explicit NSD(const char* name); + explicit NSD(Core::System& system_, const char* name); ~NSD() override; }; diff --git a/src/core/hle/service/sockets/sfdnsres.cpp b/src/core/hle/service/sockets/sfdnsres.cpp index e3017451f..3a6329f56 100644 --- a/src/core/hle/service/sockets/sfdnsres.cpp +++ b/src/core/hle/service/sockets/sfdnsres.cpp @@ -7,25 +7,7 @@ namespace Service::Sockets { -void SFDNSRES::GetAddrInfoRequest(Kernel::HLERequestContext& ctx) { - struct Parameters { - u8 use_nsd_resolve; - u32 unknown; - u64 process_id; - }; - - IPC::RequestParser rp{ctx}; - const auto parameters = rp.PopRaw<Parameters>(); - - LOG_WARNING(Service, - "(STUBBED) called. use_nsd_resolve={}, unknown=0x{:08X}, process_id=0x{:016X}", - parameters.use_nsd_resolve, parameters.unknown, parameters.process_id); - - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); -} - -SFDNSRES::SFDNSRES() : ServiceFramework("sfdnsres") { +SFDNSRES::SFDNSRES(Core::System& system_) : ServiceFramework{system_, "sfdnsres"} { static const FunctionInfo functions[] = { {0, nullptr, "SetDnsAddressesPrivate"}, {1, nullptr, "GetDnsAddressPrivate"}, @@ -49,4 +31,22 @@ SFDNSRES::SFDNSRES() : ServiceFramework("sfdnsres") { SFDNSRES::~SFDNSRES() = default; +void SFDNSRES::GetAddrInfoRequest(Kernel::HLERequestContext& ctx) { + struct Parameters { + u8 use_nsd_resolve; + u32 unknown; + u64 process_id; + }; + + IPC::RequestParser rp{ctx}; + const auto parameters = rp.PopRaw<Parameters>(); + + LOG_WARNING(Service, + "(STUBBED) called. use_nsd_resolve={}, unknown=0x{:08X}, process_id=0x{:016X}", + parameters.use_nsd_resolve, parameters.unknown, parameters.process_id); + + IPC::ResponseBuilder rb{ctx, 2}; + rb.Push(RESULT_SUCCESS); +} + } // namespace Service::Sockets diff --git a/src/core/hle/service/sockets/sfdnsres.h b/src/core/hle/service/sockets/sfdnsres.h index acd3647bb..faa6b7d0d 100644 --- a/src/core/hle/service/sockets/sfdnsres.h +++ b/src/core/hle/service/sockets/sfdnsres.h @@ -7,11 +7,15 @@ #include "core/hle/kernel/hle_ipc.h" #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Sockets { class SFDNSRES final : public ServiceFramework<SFDNSRES> { public: - explicit SFDNSRES(); + explicit SFDNSRES(Core::System& system_); ~SFDNSRES() override; private: diff --git a/src/core/hle/service/sockets/sockets.cpp b/src/core/hle/service/sockets/sockets.cpp index 1d27f7906..96f73bce3 100644 --- a/src/core/hle/service/sockets/sockets.cpp +++ b/src/core/hle/service/sockets/sockets.cpp @@ -13,15 +13,15 @@ namespace Service::Sockets { void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { std::make_shared<BSD>(system, "bsd:s")->InstallAsService(service_manager); std::make_shared<BSD>(system, "bsd:u")->InstallAsService(service_manager); - std::make_shared<BSDCFG>()->InstallAsService(service_manager); + std::make_shared<BSDCFG>(system)->InstallAsService(service_manager); - std::make_shared<ETHC_C>()->InstallAsService(service_manager); - std::make_shared<ETHC_I>()->InstallAsService(service_manager); + std::make_shared<ETHC_C>(system)->InstallAsService(service_manager); + std::make_shared<ETHC_I>(system)->InstallAsService(service_manager); - std::make_shared<NSD>("nsd:a")->InstallAsService(service_manager); - std::make_shared<NSD>("nsd:u")->InstallAsService(service_manager); + std::make_shared<NSD>(system, "nsd:a")->InstallAsService(service_manager); + std::make_shared<NSD>(system, "nsd:u")->InstallAsService(service_manager); - std::make_shared<SFDNSRES>()->InstallAsService(service_manager); + std::make_shared<SFDNSRES>(system)->InstallAsService(service_manager); } } // namespace Service::Sockets diff --git a/src/core/hle/service/sockets/sockets.h b/src/core/hle/service/sockets/sockets.h index 89a410076..5a65ed2a9 100644 --- a/src/core/hle/service/sockets/sockets.h +++ b/src/core/hle/service/sockets/sockets.h @@ -69,10 +69,22 @@ struct SockAddrIn { std::array<u8, 8> zeroes; }; +enum class PollEvents : u16 { + // Using Pascal case because IN is a macro on Windows. + In = 1 << 0, + Pri = 1 << 1, + Out = 1 << 2, + Err = 1 << 3, + Hup = 1 << 4, + Nval = 1 << 5, +}; + +DECLARE_ENUM_FLAG_OPERATORS(PollEvents); + struct PollFD { s32 fd; - u16 events; - u16 revents; + PollEvents events; + PollEvents revents; }; struct Linger { @@ -80,13 +92,6 @@ struct Linger { u32 linger; }; -constexpr u16 POLL_IN = 0x01; -constexpr u16 POLL_PRI = 0x02; -constexpr u16 POLL_OUT = 0x04; -constexpr u16 POLL_ERR = 0x08; -constexpr u16 POLL_HUP = 0x10; -constexpr u16 POLL_NVAL = 0x20; - constexpr u32 FLAG_MSG_DONTWAIT = 0x80; constexpr u32 FLAG_O_NONBLOCK = 0x800; diff --git a/src/core/hle/service/sockets/sockets_translate.cpp b/src/core/hle/service/sockets/sockets_translate.cpp index 139743e1d..ca61d72ca 100644 --- a/src/core/hle/service/sockets/sockets_translate.cpp +++ b/src/core/hle/service/sockets/sockets_translate.cpp @@ -27,7 +27,7 @@ Errno Translate(Network::Errno value) { case Network::Errno::NOTCONN: return Errno::NOTCONN; default: - UNIMPLEMENTED_MSG("Unimplemented errno={}", static_cast<int>(value)); + UNIMPLEMENTED_MSG("Unimplemented errno={}", value); return Errno::SUCCESS; } } @@ -41,7 +41,7 @@ Network::Domain Translate(Domain domain) { case Domain::INET: return Network::Domain::INET; default: - UNIMPLEMENTED_MSG("Unimplemented domain={}", static_cast<int>(domain)); + UNIMPLEMENTED_MSG("Unimplemented domain={}", domain); return {}; } } @@ -51,7 +51,7 @@ Domain Translate(Network::Domain domain) { case Network::Domain::INET: return Domain::INET; default: - UNIMPLEMENTED_MSG("Unimplemented domain={}", static_cast<int>(domain)); + UNIMPLEMENTED_MSG("Unimplemented domain={}", domain); return {}; } } @@ -63,7 +63,8 @@ Network::Type Translate(Type type) { case Type::DGRAM: return Network::Type::DGRAM; default: - UNIMPLEMENTED_MSG("Unimplemented type={}", static_cast<int>(type)); + UNIMPLEMENTED_MSG("Unimplemented type={}", type); + return Network::Type{}; } } @@ -84,47 +85,47 @@ Network::Protocol Translate(Type type, Protocol protocol) { case Protocol::UDP: return Network::Protocol::UDP; default: - UNIMPLEMENTED_MSG("Unimplemented protocol={}", static_cast<int>(protocol)); + UNIMPLEMENTED_MSG("Unimplemented protocol={}", protocol); return Network::Protocol::TCP; } } -u16 TranslatePollEventsToHost(u16 flags) { - u16 result = 0; - const auto translate = [&result, &flags](u16 from, u16 to) { - if ((flags & from) != 0) { +Network::PollEvents TranslatePollEventsToHost(PollEvents flags) { + Network::PollEvents result{}; + const auto translate = [&result, &flags](PollEvents from, Network::PollEvents to) { + if (True(flags & from)) { flags &= ~from; result |= to; } }; - translate(POLL_IN, Network::POLL_IN); - translate(POLL_PRI, Network::POLL_PRI); - translate(POLL_OUT, Network::POLL_OUT); - translate(POLL_ERR, Network::POLL_ERR); - translate(POLL_HUP, Network::POLL_HUP); - translate(POLL_NVAL, Network::POLL_NVAL); - - UNIMPLEMENTED_IF_MSG(flags != 0, "Unimplemented flags={}", flags); + translate(PollEvents::In, Network::PollEvents::In); + translate(PollEvents::Pri, Network::PollEvents::Pri); + translate(PollEvents::Out, Network::PollEvents::Out); + translate(PollEvents::Err, Network::PollEvents::Err); + translate(PollEvents::Hup, Network::PollEvents::Hup); + translate(PollEvents::Nval, Network::PollEvents::Nval); + + UNIMPLEMENTED_IF_MSG((u16)flags != 0, "Unimplemented flags={}", (u16)flags); return result; } -u16 TranslatePollEventsToGuest(u16 flags) { - u16 result = 0; - const auto translate = [&result, &flags](u16 from, u16 to) { - if ((flags & from) != 0) { +PollEvents TranslatePollEventsToGuest(Network::PollEvents flags) { + PollEvents result{}; + const auto translate = [&result, &flags](Network::PollEvents from, PollEvents to) { + if (True(flags & from)) { flags &= ~from; result |= to; } }; - translate(Network::POLL_IN, POLL_IN); - translate(Network::POLL_PRI, POLL_PRI); - translate(Network::POLL_OUT, POLL_OUT); - translate(Network::POLL_ERR, POLL_ERR); - translate(Network::POLL_HUP, POLL_HUP); - translate(Network::POLL_NVAL, POLL_NVAL); + translate(Network::PollEvents::In, PollEvents::In); + translate(Network::PollEvents::Pri, PollEvents::Pri); + translate(Network::PollEvents::Out, PollEvents::Out); + translate(Network::PollEvents::Err, PollEvents::Err); + translate(Network::PollEvents::Hup, PollEvents::Hup); + translate(Network::PollEvents::Nval, PollEvents::Nval); - UNIMPLEMENTED_IF_MSG(flags != 0, "Unimplemented flags={}", flags); + UNIMPLEMENTED_IF_MSG((u16)flags != 0, "Unimplemented flags={}", (u16)flags); return result; } @@ -157,7 +158,7 @@ Network::ShutdownHow Translate(ShutdownHow how) { case ShutdownHow::RDWR: return Network::ShutdownHow::RDWR; default: - UNIMPLEMENTED_MSG("Unimplemented how={}", static_cast<int>(how)); + UNIMPLEMENTED_MSG("Unimplemented how={}", how); return {}; } } diff --git a/src/core/hle/service/sockets/sockets_translate.h b/src/core/hle/service/sockets/sockets_translate.h index 8ed041e31..057d1ff22 100644 --- a/src/core/hle/service/sockets/sockets_translate.h +++ b/src/core/hle/service/sockets/sockets_translate.h @@ -31,10 +31,10 @@ Network::Type Translate(Type type); Network::Protocol Translate(Type type, Protocol protocol); /// Translate abstract poll event flags to guest poll event flags -u16 TranslatePollEventsToHost(u16 flags); +Network::PollEvents TranslatePollEventsToHost(PollEvents flags); /// Translate guest poll event flags to abstract poll event flags -u16 TranslatePollEventsToGuest(u16 flags); +PollEvents TranslatePollEventsToGuest(Network::PollEvents flags); /// Translate guest socket address structure to abstract socket address structure Network::SockAddrIn Translate(SockAddrIn value); diff --git a/src/core/hle/service/spl/csrng.cpp b/src/core/hle/service/spl/csrng.cpp index 674928798..1beca417c 100644 --- a/src/core/hle/service/spl/csrng.cpp +++ b/src/core/hle/service/spl/csrng.cpp @@ -6,7 +6,8 @@ namespace Service::SPL { -CSRNG::CSRNG(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "csrng") { +CSRNG::CSRNG(Core::System& system_, std::shared_ptr<Module> module_) + : Interface(system_, std::move(module_), "csrng") { static const FunctionInfo functions[] = { {0, &CSRNG::GetRandomBytes, "GetRandomBytes"}, }; diff --git a/src/core/hle/service/spl/csrng.h b/src/core/hle/service/spl/csrng.h index 764d5ceb0..5c0bd2199 100644 --- a/src/core/hle/service/spl/csrng.h +++ b/src/core/hle/service/spl/csrng.h @@ -6,11 +6,15 @@ #include "core/hle/service/spl/module.h" +namespace Core { +class System; +} + namespace Service::SPL { class CSRNG final : public Module::Interface { public: - explicit CSRNG(std::shared_ptr<Module> module); + explicit CSRNG(Core::System& system_, std::shared_ptr<Module> module_); ~CSRNG() override; }; diff --git a/src/core/hle/service/spl/module.cpp b/src/core/hle/service/spl/module.cpp index 865ed3b91..dea6b0fe0 100644 --- a/src/core/hle/service/spl/module.cpp +++ b/src/core/hle/service/spl/module.cpp @@ -17,8 +17,9 @@ namespace Service::SPL { -Module::Interface::Interface(std::shared_ptr<Module> module, const char* name) - : ServiceFramework(name), module(std::move(module)), +Module::Interface::Interface(Core::System& system_, std::shared_ptr<Module> module_, + const char* name) + : ServiceFramework{system_, name}, module{std::move(module_)}, rng(Settings::values.rng_seed.GetValue().value_or(std::time(nullptr))) {} Module::Interface::~Interface() = default; @@ -38,10 +39,10 @@ void Module::Interface::GetRandomBytes(Kernel::HLERequestContext& ctx) { rb.Push(RESULT_SUCCESS); } -void InstallInterfaces(SM::ServiceManager& service_manager) { +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { auto module = std::make_shared<Module>(); - std::make_shared<CSRNG>(module)->InstallAsService(service_manager); - std::make_shared<SPL>(module)->InstallAsService(service_manager); + std::make_shared<CSRNG>(system, module)->InstallAsService(service_manager); + std::make_shared<SPL>(system, module)->InstallAsService(service_manager); } } // namespace Service::SPL diff --git a/src/core/hle/service/spl/module.h b/src/core/hle/service/spl/module.h index afa1f0295..71855c1bf 100644 --- a/src/core/hle/service/spl/module.h +++ b/src/core/hle/service/spl/module.h @@ -7,13 +7,18 @@ #include <random> #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::SPL { class Module final { public: class Interface : public ServiceFramework<Interface> { public: - explicit Interface(std::shared_ptr<Module> module, const char* name); + explicit Interface(Core::System& system_, std::shared_ptr<Module> module_, + const char* name); ~Interface() override; void GetRandomBytes(Kernel::HLERequestContext& ctx); @@ -27,6 +32,6 @@ public: }; /// Registers all SPL services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); } // namespace Service::SPL diff --git a/src/core/hle/service/spl/spl.cpp b/src/core/hle/service/spl/spl.cpp index 773551464..3fabc2c79 100644 --- a/src/core/hle/service/spl/spl.cpp +++ b/src/core/hle/service/spl/spl.cpp @@ -6,7 +6,8 @@ namespace Service::SPL { -SPL::SPL(std::shared_ptr<Module> module) : Module::Interface(std::move(module), "spl:") { +SPL::SPL(Core::System& system_, std::shared_ptr<Module> module_) + : Interface(system_, std::move(module_), "spl:") { static const FunctionInfo functions[] = { {0, nullptr, "GetConfig"}, {1, nullptr, "ModularExponentiate"}, diff --git a/src/core/hle/service/spl/spl.h b/src/core/hle/service/spl/spl.h index 3637d1623..d27d16b86 100644 --- a/src/core/hle/service/spl/spl.h +++ b/src/core/hle/service/spl/spl.h @@ -6,11 +6,15 @@ #include "core/hle/service/spl/module.h" +namespace Core { +class System; +} + namespace Service::SPL { class SPL final : public Module::Interface { public: - explicit SPL(std::shared_ptr<Module> module); + explicit SPL(Core::System& system_, std::shared_ptr<Module> module_); ~SPL() override; }; diff --git a/src/core/hle/service/ssl/ssl.cpp b/src/core/hle/service/ssl/ssl.cpp index 1ba8c19a0..dc2baca4a 100644 --- a/src/core/hle/service/ssl/ssl.cpp +++ b/src/core/hle/service/ssl/ssl.cpp @@ -12,7 +12,7 @@ namespace Service::SSL { class ISslConnection final : public ServiceFramework<ISslConnection> { public: - ISslConnection() : ServiceFramework("ISslConnection") { + explicit ISslConnection(Core::System& system_) : ServiceFramework{system_, "ISslConnection"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SetSocketDescriptor"}, @@ -52,7 +52,7 @@ public: class ISslContext final : public ServiceFramework<ISslContext> { public: - ISslContext() : ServiceFramework("ISslContext") { + explicit ISslContext(Core::System& system_) : ServiceFramework{system_, "ISslContext"} { static const FunctionInfo functions[] = { {0, &ISslContext::SetOption, "SetOption"}, {1, nullptr, "GetOption"}, @@ -92,13 +92,13 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISslConnection>(); + rb.PushIpcInterface<ISslConnection>(system); } }; class SSL final : public ServiceFramework<SSL> { public: - explicit SSL() : ServiceFramework{"ssl"} { + explicit SSL(Core::System& system_) : ServiceFramework{system_, "ssl"} { // clang-format off static const FunctionInfo functions[] = { {0, &SSL::CreateContext, "CreateContext"}, @@ -123,7 +123,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISslContext>(); + rb.PushIpcInterface<ISslContext>(system); } void SetInterfaceVersion(Kernel::HLERequestContext& ctx) { @@ -137,8 +137,8 @@ private: } }; -void InstallInterfaces(SM::ServiceManager& service_manager) { - std::make_shared<SSL>()->InstallAsService(service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system) { + std::make_shared<SSL>(system)->InstallAsService(service_manager); } } // namespace Service::SSL diff --git a/src/core/hle/service/ssl/ssl.h b/src/core/hle/service/ssl/ssl.h index 5cb04c3b9..a3aa4b4b5 100644 --- a/src/core/hle/service/ssl/ssl.h +++ b/src/core/hle/service/ssl/ssl.h @@ -4,6 +4,10 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } @@ -11,6 +15,6 @@ class ServiceManager; namespace Service::SSL { /// Registers all SSL services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system); } // namespace Service::SSL diff --git a/src/core/hle/service/time/interface.cpp b/src/core/hle/service/time/interface.cpp index ba8fd6152..a01d9e0ff 100644 --- a/src/core/hle/service/time/interface.cpp +++ b/src/core/hle/service/time/interface.cpp @@ -7,7 +7,7 @@ namespace Service::Time { Time::Time(std::shared_ptr<Module> module, Core::System& system, const char* name) - : Module::Interface(std::move(module), system, name) { + : Interface(std::move(module), system, name) { // clang-format off static const FunctionInfo functions[] = { {0, &Time::GetStandardUserSystemClock, "GetStandardUserSystemClock"}, diff --git a/src/core/hle/service/time/time.cpp b/src/core/hle/service/time/time.cpp index ee4fa4b48..abc753d5d 100644 --- a/src/core/hle/service/time/time.cpp +++ b/src/core/hle/service/time/time.cpp @@ -10,7 +10,8 @@ #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/client_session.h" -#include "core/hle/kernel/scheduler.h" +#include "core/hle/kernel/k_scheduler.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/service/time/interface.h" #include "core/hle/service/time/time.h" #include "core/hle/service/time/time_sharedmemory.h" @@ -20,8 +21,8 @@ namespace Service::Time { class ISystemClock final : public ServiceFramework<ISystemClock> { public: - explicit ISystemClock(Clock::SystemClockCore& clock_core, Core::System& system) - : ServiceFramework("ISystemClock"), clock_core{clock_core}, system{system} { + explicit ISystemClock(Clock::SystemClockCore& clock_core_, Core::System& system_) + : ServiceFramework{system_, "ISystemClock"}, clock_core{clock_core_} { // clang-format off static const FunctionInfo functions[] = { {0, &ISystemClock::GetCurrentTime, "GetCurrentTime"}, @@ -81,13 +82,12 @@ private: } Clock::SystemClockCore& clock_core; - Core::System& system; }; class ISteadyClock final : public ServiceFramework<ISteadyClock> { public: - explicit ISteadyClock(Clock::SteadyClockCore& clock_core, Core::System& system) - : ServiceFramework("ISteadyClock"), clock_core{clock_core}, system{system} { + explicit ISteadyClock(Clock::SteadyClockCore& clock_core_, Core::System& system_) + : ServiceFramework{system_, "ISteadyClock"}, clock_core{clock_core_} { static const FunctionInfo functions[] = { {0, &ISteadyClock::GetCurrentTimePoint, "GetCurrentTimePoint"}, {2, nullptr, "GetTestOffset"}, @@ -118,14 +118,13 @@ private: } Clock::SteadyClockCore& clock_core; - Core::System& system; }; ResultCode Module::Interface::GetClockSnapshotFromSystemClockContextInternal( Kernel::Thread* thread, Clock::SystemClockContext user_context, Clock::SystemClockContext network_context, u8 type, Clock::ClockSnapshot& clock_snapshot) { - auto& time_manager{module->GetTimeManager()}; + auto& time_manager{system.GetTimeManager()}; clock_snapshot.is_automatic_correction_enabled = time_manager.GetStandardUserSystemClockCore().IsAutomaticCorrectionEnabled(); @@ -182,7 +181,7 @@ void Module::Interface::GetStandardUserSystemClock(Kernel::HLERequestContext& ct LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemClock>(module->GetTimeManager().GetStandardUserSystemClockCore(), + rb.PushIpcInterface<ISystemClock>(system.GetTimeManager().GetStandardUserSystemClockCore(), system); } @@ -190,7 +189,7 @@ void Module::Interface::GetStandardNetworkSystemClock(Kernel::HLERequestContext& LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemClock>(module->GetTimeManager().GetStandardNetworkSystemClockCore(), + rb.PushIpcInterface<ISystemClock>(system.GetTimeManager().GetStandardNetworkSystemClockCore(), system); } @@ -198,29 +197,29 @@ void Module::Interface::GetStandardSteadyClock(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISteadyClock>(module->GetTimeManager().GetStandardSteadyClockCore(), - system); + rb.PushIpcInterface<ISteadyClock>(system.GetTimeManager().GetStandardSteadyClockCore(), system); } void Module::Interface::GetTimeZoneService(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ITimeZoneService>(module->GetTimeManager().GetTimeZoneContentManager()); + rb.PushIpcInterface<ITimeZoneService>(system, + system.GetTimeManager().GetTimeZoneContentManager()); } void Module::Interface::GetStandardLocalSystemClock(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemClock>(module->GetTimeManager().GetStandardLocalSystemClockCore(), + rb.PushIpcInterface<ISystemClock>(system.GetTimeManager().GetStandardLocalSystemClockCore(), system); } void Module::Interface::IsStandardNetworkSystemClockAccuracySufficient( Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); - auto& clock_core{module->GetTimeManager().GetStandardNetworkSystemClockCore()}; + auto& clock_core{system.GetTimeManager().GetStandardNetworkSystemClockCore()}; IPC::ResponseBuilder rb{ctx, 3}; rb.Push(RESULT_SUCCESS); rb.Push<u32>(clock_core.IsStandardNetworkSystemClockAccuracySufficient(system)); @@ -229,7 +228,7 @@ void Module::Interface::IsStandardNetworkSystemClockAccuracySufficient( void Module::Interface::CalculateMonotonicSystemClockBaseTimePoint(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Time, "called"); - auto& steady_clock_core{module->GetTimeManager().GetStandardSteadyClockCore()}; + auto& steady_clock_core{system.GetTimeManager().GetStandardSteadyClockCore()}; if (!steady_clock_core.IsInitialized()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_UNINITIALIZED_CLOCK); @@ -262,8 +261,8 @@ void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) { Clock::SystemClockContext user_context{}; if (const ResultCode result{ - module->GetTimeManager().GetStandardUserSystemClockCore().GetClockContext( - system, user_context)}; + system.GetTimeManager().GetStandardUserSystemClockCore().GetClockContext(system, + user_context)}; result.IsError()) { IPC::ResponseBuilder rb{ctx, 2}; rb.Push(result); @@ -271,7 +270,7 @@ void Module::Interface::GetClockSnapshot(Kernel::HLERequestContext& ctx) { } Clock::SystemClockContext network_context{}; if (const ResultCode result{ - module->GetTimeManager().GetStandardNetworkSystemClockCore().GetClockContext( + system.GetTimeManager().GetStandardNetworkSystemClockCore().GetClockContext( system, network_context)}; result.IsError()) { IPC::ResponseBuilder rb{ctx, 2}; @@ -372,16 +371,17 @@ void Module::Interface::GetSharedMemoryNativeHandle(Kernel::HLERequestContext& c LOG_DEBUG(Service_Time, "called"); IPC::ResponseBuilder rb{ctx, 2, 1}; rb.Push(RESULT_SUCCESS); - rb.PushCopyObjects(module->GetTimeManager().GetSharedMemory().GetSharedMemoryHolder()); + rb.PushCopyObjects(SharedFrom(&system.Kernel().GetTimeSharedMem())); } -Module::Interface::Interface(std::shared_ptr<Module> module, Core::System& system, const char* name) - : ServiceFramework(name), module{std::move(module)}, system{system} {} +Module::Interface::Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name) + : ServiceFramework{system_, name}, module{std::move(module_)} {} Module::Interface::~Interface() = default; void InstallInterfaces(Core::System& system) { - auto module{std::make_shared<Module>(system)}; + auto module{std::make_shared<Module>()}; std::make_shared<Time>(module, system, "time:a")->InstallAsService(system.ServiceManager()); std::make_shared<Time>(module, system, "time:s")->InstallAsService(system.ServiceManager()); std::make_shared<Time>(module, system, "time:u")->InstallAsService(system.ServiceManager()); diff --git a/src/core/hle/service/time/time.h b/src/core/hle/service/time/time.h index 41f3002e9..975a8ae5b 100644 --- a/src/core/hle/service/time/time.h +++ b/src/core/hle/service/time/time.h @@ -16,11 +16,12 @@ namespace Service::Time { class Module final { public: - Module(Core::System& system) : time_manager{system} {} + Module() = default; class Interface : public ServiceFramework<Interface> { public: - explicit Interface(std::shared_ptr<Module> module, Core::System& system, const char* name); + explicit Interface(std::shared_ptr<Module> module_, Core::System& system_, + const char* name); ~Interface() override; void GetStandardUserSystemClock(Kernel::HLERequestContext& ctx); @@ -44,15 +45,7 @@ public: protected: std::shared_ptr<Module> module; - Core::System& system; }; - - TimeManager& GetTimeManager() { - return time_manager; - } - -private: - TimeManager time_manager; }; /// Registers all Time services with the specified service manager. diff --git a/src/core/hle/service/time/time_manager.cpp b/src/core/hle/service/time/time_manager.cpp index b4dfe45e5..858623e2b 100644 --- a/src/core/hle/service/time/time_manager.cpp +++ b/src/core/hle/service/time/time_manager.cpp @@ -22,125 +22,282 @@ static std::chrono::seconds GetSecondsSinceEpoch() { Settings::values.custom_rtc_differential; } -static s64 GetExternalTimeZoneOffset() { - // With "auto" timezone setting, we use the external system's timezone offset - if (Settings::GetTimeZoneString() == "auto") { - return Common::TimeZone::GetCurrentOffsetSeconds().count(); - } - return 0; -} - static s64 GetExternalRtcValue() { - return GetSecondsSinceEpoch().count() + GetExternalTimeZoneOffset(); -} - -TimeManager::TimeManager(Core::System& system) - : shared_memory{system}, standard_local_system_clock_core{standard_steady_clock_core}, - standard_network_system_clock_core{standard_steady_clock_core}, - standard_user_system_clock_core{standard_local_system_clock_core, - standard_network_system_clock_core, system}, - ephemeral_network_system_clock_core{tick_based_steady_clock_core}, - local_system_clock_context_writer{ - std::make_shared<Clock::LocalSystemClockContextWriter>(shared_memory)}, - network_system_clock_context_writer{ - std::make_shared<Clock::NetworkSystemClockContextWriter>(shared_memory)}, - ephemeral_network_system_clock_context_writer{ - std::make_shared<Clock::EphemeralNetworkSystemClockContextWriter>()}, - time_zone_content_manager{*this, system} { - - const auto system_time{Clock::TimeSpanType::FromSeconds(GetExternalRtcValue())}; - SetupStandardSteadyClock(system, Common::UUID::Generate(), system_time, {}, {}); - SetupStandardLocalSystemClock(system, {}, system_time.ToSeconds()); - SetupStandardNetworkSystemClock({}, standard_network_clock_accuracy); - SetupStandardUserSystemClock(system, {}, Clock::SteadyClockTimePoint::GetRandom()); - SetupEphemeralNetworkSystemClock(); + return GetSecondsSinceEpoch().count() + TimeManager::GetExternalTimeZoneOffset(); } -TimeManager::~TimeManager() = default; +struct TimeManager::Impl final { + explicit Impl(Core::System& system) + : shared_memory{system}, standard_local_system_clock_core{standard_steady_clock_core}, + standard_network_system_clock_core{standard_steady_clock_core}, + standard_user_system_clock_core{standard_local_system_clock_core, + standard_network_system_clock_core, system}, + ephemeral_network_system_clock_core{tick_based_steady_clock_core}, + local_system_clock_context_writer{ + std::make_shared<Clock::LocalSystemClockContextWriter>(shared_memory)}, + network_system_clock_context_writer{ + std::make_shared<Clock::NetworkSystemClockContextWriter>(shared_memory)}, + ephemeral_network_system_clock_context_writer{ + std::make_shared<Clock::EphemeralNetworkSystemClockContextWriter>()}, + time_zone_content_manager{system} { -void TimeManager::SetupTimeZoneManager(std::string location_name, - Clock::SteadyClockTimePoint time_zone_updated_time_point, - std::size_t total_location_name_count, - u128 time_zone_rule_version, - FileSys::VirtualFile& vfs_file) { - if (time_zone_content_manager.GetTimeZoneManager().SetDeviceLocationNameWithTimeZoneRule( - location_name, vfs_file) != RESULT_SUCCESS) { - UNREACHABLE(); - return; - } - - time_zone_content_manager.GetTimeZoneManager().SetUpdatedTime(time_zone_updated_time_point); - time_zone_content_manager.GetTimeZoneManager().SetTotalLocationNameCount( - total_location_name_count); - time_zone_content_manager.GetTimeZoneManager().SetTimeZoneRuleVersion(time_zone_rule_version); - time_zone_content_manager.GetTimeZoneManager().MarkAsInitialized(); -} - -void TimeManager::SetupStandardSteadyClock(Core::System& system, Common::UUID clock_source_id, - Clock::TimeSpanType setup_value, - Clock::TimeSpanType internal_offset, - bool is_rtc_reset_detected) { - standard_steady_clock_core.SetClockSourceId(clock_source_id); - standard_steady_clock_core.SetSetupValue(setup_value); - standard_steady_clock_core.SetInternalOffset(internal_offset); - standard_steady_clock_core.MarkAsInitialized(); - - const auto current_time_point{standard_steady_clock_core.GetCurrentRawTimePoint(system)}; - shared_memory.SetupStandardSteadyClock(system, clock_source_id, current_time_point); -} - -void TimeManager::SetupStandardLocalSystemClock(Core::System& system, - Clock::SystemClockContext clock_context, - s64 posix_time) { - standard_local_system_clock_core.SetUpdateCallbackInstance(local_system_clock_context_writer); - - const auto current_time_point{ - standard_local_system_clock_core.GetSteadyClockCore().GetCurrentTimePoint(system)}; - if (current_time_point.clock_source_id == clock_context.steady_time_point.clock_source_id) { - standard_local_system_clock_core.SetSystemClockContext(clock_context); - } else { - if (standard_local_system_clock_core.SetCurrentTime(system, posix_time) != RESULT_SUCCESS) { + const auto system_time{Clock::TimeSpanType::FromSeconds(GetExternalRtcValue())}; + SetupStandardSteadyClock(system, Common::UUID::Generate(), system_time, {}, {}); + SetupStandardLocalSystemClock(system, {}, system_time.ToSeconds()); + SetupStandardNetworkSystemClock({}, standard_network_clock_accuracy); + SetupStandardUserSystemClock(system, {}, Clock::SteadyClockTimePoint::GetRandom()); + SetupEphemeralNetworkSystemClock(); + } + + ~Impl() = default; + + Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() { + return standard_steady_clock_core; + } + + const Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() const { + return standard_steady_clock_core; + } + + Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() { + return standard_local_system_clock_core; + } + + const Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() const { + return standard_local_system_clock_core; + } + + Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() { + return standard_network_system_clock_core; + } + + const Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() const { + return standard_network_system_clock_core; + } + + Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() { + return standard_user_system_clock_core; + } + + const Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() const { + return standard_user_system_clock_core; + } + + TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() { + return time_zone_content_manager; + } + + const TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() const { + return time_zone_content_manager; + } + + SharedMemory& GetSharedMemory() { + return shared_memory; + } + + const SharedMemory& GetSharedMemory() const { + return shared_memory; + } + + void SetupTimeZoneManager(std::string location_name, + Clock::SteadyClockTimePoint time_zone_updated_time_point, + std::size_t total_location_name_count, u128 time_zone_rule_version, + FileSys::VirtualFile& vfs_file) { + if (time_zone_content_manager.GetTimeZoneManager().SetDeviceLocationNameWithTimeZoneRule( + location_name, vfs_file) != RESULT_SUCCESS) { UNREACHABLE(); return; } + + time_zone_content_manager.GetTimeZoneManager().SetUpdatedTime(time_zone_updated_time_point); + time_zone_content_manager.GetTimeZoneManager().SetTotalLocationNameCount( + total_location_name_count); + time_zone_content_manager.GetTimeZoneManager().SetTimeZoneRuleVersion( + time_zone_rule_version); + time_zone_content_manager.GetTimeZoneManager().MarkAsInitialized(); } - standard_local_system_clock_core.MarkAsInitialized(); -} + static s64 GetExternalTimeZoneOffset() { + // With "auto" timezone setting, we use the external system's timezone offset + if (Settings::GetTimeZoneString() == "auto") { + return Common::TimeZone::GetCurrentOffsetSeconds().count(); + } + return 0; + } -void TimeManager::SetupStandardNetworkSystemClock(Clock::SystemClockContext clock_context, - Clock::TimeSpanType sufficient_accuracy) { - standard_network_system_clock_core.SetUpdateCallbackInstance( - network_system_clock_context_writer); + void SetupStandardSteadyClock(Core::System& system, Common::UUID clock_source_id, + Clock::TimeSpanType setup_value, + Clock::TimeSpanType internal_offset, bool is_rtc_reset_detected) { + standard_steady_clock_core.SetClockSourceId(clock_source_id); + standard_steady_clock_core.SetSetupValue(setup_value); + standard_steady_clock_core.SetInternalOffset(internal_offset); + standard_steady_clock_core.MarkAsInitialized(); - if (standard_network_system_clock_core.SetSystemClockContext(clock_context) != RESULT_SUCCESS) { - UNREACHABLE(); - return; + const auto current_time_point{standard_steady_clock_core.GetCurrentRawTimePoint(system)}; + shared_memory.SetupStandardSteadyClock(system, clock_source_id, current_time_point); } - standard_network_system_clock_core.SetStandardNetworkClockSufficientAccuracy( - sufficient_accuracy); - standard_network_system_clock_core.MarkAsInitialized(); -} + void SetupStandardLocalSystemClock(Core::System& system, + Clock::SystemClockContext clock_context, s64 posix_time) { + standard_local_system_clock_core.SetUpdateCallbackInstance( + local_system_clock_context_writer); + + const auto current_time_point{ + standard_local_system_clock_core.GetSteadyClockCore().GetCurrentTimePoint(system)}; + if (current_time_point.clock_source_id == clock_context.steady_time_point.clock_source_id) { + standard_local_system_clock_core.SetSystemClockContext(clock_context); + } else { + if (standard_local_system_clock_core.SetCurrentTime(system, posix_time) != + RESULT_SUCCESS) { + UNREACHABLE(); + return; + } + } + + standard_local_system_clock_core.MarkAsInitialized(); + } + + void SetupStandardNetworkSystemClock(Clock::SystemClockContext clock_context, + Clock::TimeSpanType sufficient_accuracy) { + standard_network_system_clock_core.SetUpdateCallbackInstance( + network_system_clock_context_writer); -void TimeManager::SetupStandardUserSystemClock( - Core::System& system, bool is_automatic_correction_enabled, - Clock::SteadyClockTimePoint steady_clock_time_point) { - if (standard_user_system_clock_core.SetAutomaticCorrectionEnabled( - system, is_automatic_correction_enabled) != RESULT_SUCCESS) { - UNREACHABLE(); - return; + if (standard_network_system_clock_core.SetSystemClockContext(clock_context) != + RESULT_SUCCESS) { + UNREACHABLE(); + return; + } + + standard_network_system_clock_core.SetStandardNetworkClockSufficientAccuracy( + sufficient_accuracy); + standard_network_system_clock_core.MarkAsInitialized(); } - standard_user_system_clock_core.SetAutomaticCorrectionUpdatedTime(steady_clock_time_point); - standard_user_system_clock_core.MarkAsInitialized(); - shared_memory.SetAutomaticCorrectionEnabled(is_automatic_correction_enabled); + void SetupStandardUserSystemClock(Core::System& system, bool is_automatic_correction_enabled, + Clock::SteadyClockTimePoint steady_clock_time_point) { + if (standard_user_system_clock_core.SetAutomaticCorrectionEnabled( + system, is_automatic_correction_enabled) != RESULT_SUCCESS) { + UNREACHABLE(); + return; + } + + standard_user_system_clock_core.SetAutomaticCorrectionUpdatedTime(steady_clock_time_point); + standard_user_system_clock_core.MarkAsInitialized(); + shared_memory.SetAutomaticCorrectionEnabled(is_automatic_correction_enabled); + } + + void SetupEphemeralNetworkSystemClock() { + ephemeral_network_system_clock_core.SetUpdateCallbackInstance( + ephemeral_network_system_clock_context_writer); + ephemeral_network_system_clock_core.MarkAsInitialized(); + } + + void UpdateLocalSystemClockTime(Core::System& system, s64 posix_time) { + const auto timespan{Service::Time::Clock::TimeSpanType::FromSeconds(posix_time)}; + if (GetStandardLocalSystemClockCore() + .SetCurrentTime(system, timespan.ToSeconds()) + .IsError()) { + UNREACHABLE(); + return; + } + } + + SharedMemory shared_memory; + + Clock::StandardSteadyClockCore standard_steady_clock_core; + Clock::TickBasedSteadyClockCore tick_based_steady_clock_core; + Clock::StandardLocalSystemClockCore standard_local_system_clock_core; + Clock::StandardNetworkSystemClockCore standard_network_system_clock_core; + Clock::StandardUserSystemClockCore standard_user_system_clock_core; + Clock::EphemeralNetworkSystemClockCore ephemeral_network_system_clock_core; + + std::shared_ptr<Clock::LocalSystemClockContextWriter> local_system_clock_context_writer; + std::shared_ptr<Clock::NetworkSystemClockContextWriter> network_system_clock_context_writer; + std::shared_ptr<Clock::EphemeralNetworkSystemClockContextWriter> + ephemeral_network_system_clock_context_writer; + + TimeZone::TimeZoneContentManager time_zone_content_manager; +}; + +TimeManager::TimeManager(Core::System& system) : system{system} {} + +TimeManager::~TimeManager() = default; + +void TimeManager::Initialize() { + impl = std::make_unique<Impl>(system); + + // Time zones can only be initialized after impl is valid + impl->time_zone_content_manager.Initialize(*this); +} + +Clock::StandardSteadyClockCore& TimeManager::GetStandardSteadyClockCore() { + return impl->standard_steady_clock_core; +} + +const Clock::StandardSteadyClockCore& TimeManager::GetStandardSteadyClockCore() const { + return impl->standard_steady_clock_core; +} + +Clock::StandardLocalSystemClockCore& TimeManager::GetStandardLocalSystemClockCore() { + return impl->standard_local_system_clock_core; +} + +const Clock::StandardLocalSystemClockCore& TimeManager::GetStandardLocalSystemClockCore() const { + return impl->standard_local_system_clock_core; +} + +Clock::StandardNetworkSystemClockCore& TimeManager::GetStandardNetworkSystemClockCore() { + return impl->standard_network_system_clock_core; } -void TimeManager::SetupEphemeralNetworkSystemClock() { - ephemeral_network_system_clock_core.SetUpdateCallbackInstance( - ephemeral_network_system_clock_context_writer); - ephemeral_network_system_clock_core.MarkAsInitialized(); +const Clock::StandardNetworkSystemClockCore& TimeManager::GetStandardNetworkSystemClockCore() + const { + return impl->standard_network_system_clock_core; +} + +Clock::StandardUserSystemClockCore& TimeManager::GetStandardUserSystemClockCore() { + return impl->standard_user_system_clock_core; +} + +const Clock::StandardUserSystemClockCore& TimeManager::GetStandardUserSystemClockCore() const { + return impl->standard_user_system_clock_core; +} + +TimeZone::TimeZoneContentManager& TimeManager::GetTimeZoneContentManager() { + return impl->time_zone_content_manager; +} + +const TimeZone::TimeZoneContentManager& TimeManager::GetTimeZoneContentManager() const { + return impl->time_zone_content_manager; +} + +SharedMemory& TimeManager::GetSharedMemory() { + return impl->shared_memory; +} + +const SharedMemory& TimeManager::GetSharedMemory() const { + return impl->shared_memory; +} + +void TimeManager::UpdateLocalSystemClockTime(s64 posix_time) { + impl->UpdateLocalSystemClockTime(system, posix_time); +} + +void TimeManager::SetupTimeZoneManager(std::string location_name, + Clock::SteadyClockTimePoint time_zone_updated_time_point, + std::size_t total_location_name_count, + u128 time_zone_rule_version, + FileSys::VirtualFile& vfs_file) { + impl->SetupTimeZoneManager(location_name, time_zone_updated_time_point, + total_location_name_count, time_zone_rule_version, vfs_file); +} + +/*static*/ s64 TimeManager::GetExternalTimeZoneOffset() { + // With "auto" timezone setting, we use the external system's timezone offset + if (Settings::GetTimeZoneString() == "auto") { + return Common::TimeZone::GetCurrentOffsetSeconds().count(); + } + return 0; } } // namespace Service::Time diff --git a/src/core/hle/service/time/time_manager.h b/src/core/hle/service/time/time_manager.h index 8e65f0d22..993c7c288 100644 --- a/src/core/hle/service/time/time_manager.h +++ b/src/core/hle/service/time/time_manager.h @@ -5,6 +5,7 @@ #pragma once #include "common/common_types.h" +#include "common/time_zone.h" #include "core/file_sys/vfs_types.h" #include "core/hle/service/time/clock_types.h" #include "core/hle/service/time/ephemeral_network_system_clock_core.h" @@ -32,86 +33,46 @@ public: explicit TimeManager(Core::System& system); ~TimeManager(); - Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() { - return standard_steady_clock_core; - } + void Initialize(); - const Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() const { - return standard_steady_clock_core; - } + Clock::StandardSteadyClockCore& GetStandardSteadyClockCore(); - Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() { - return standard_local_system_clock_core; - } + const Clock::StandardSteadyClockCore& GetStandardSteadyClockCore() const; - const Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() const { - return standard_local_system_clock_core; - } + Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore(); - Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() { - return standard_network_system_clock_core; - } + const Clock::StandardLocalSystemClockCore& GetStandardLocalSystemClockCore() const; - const Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() const { - return standard_network_system_clock_core; - } + Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore(); - Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() { - return standard_user_system_clock_core; - } + const Clock::StandardNetworkSystemClockCore& GetStandardNetworkSystemClockCore() const; - const Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() const { - return standard_user_system_clock_core; - } + Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore(); - TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() { - return time_zone_content_manager; - } + const Clock::StandardUserSystemClockCore& GetStandardUserSystemClockCore() const; - const TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() const { - return time_zone_content_manager; - } + TimeZone::TimeZoneContentManager& GetTimeZoneContentManager(); - SharedMemory& GetSharedMemory() { - return shared_memory; - } + const TimeZone::TimeZoneContentManager& GetTimeZoneContentManager() const; - const SharedMemory& GetSharedMemory() const { - return shared_memory; - } + void UpdateLocalSystemClockTime(s64 posix_time); + + SharedMemory& GetSharedMemory(); + + const SharedMemory& GetSharedMemory() const; void SetupTimeZoneManager(std::string location_name, Clock::SteadyClockTimePoint time_zone_updated_time_point, std::size_t total_location_name_count, u128 time_zone_rule_version, FileSys::VirtualFile& vfs_file); + static s64 GetExternalTimeZoneOffset(); + private: - void SetupStandardSteadyClock(Core::System& system, Common::UUID clock_source_id, - Clock::TimeSpanType setup_value, - Clock::TimeSpanType internal_offset, bool is_rtc_reset_detected); - void SetupStandardLocalSystemClock(Core::System& system, - Clock::SystemClockContext clock_context, s64 posix_time); - void SetupStandardNetworkSystemClock(Clock::SystemClockContext clock_context, - Clock::TimeSpanType sufficient_accuracy); - void SetupStandardUserSystemClock(Core::System& system, bool is_automatic_correction_enabled, - Clock::SteadyClockTimePoint steady_clock_time_point); - void SetupEphemeralNetworkSystemClock(); - - SharedMemory shared_memory; - - Clock::StandardSteadyClockCore standard_steady_clock_core; - Clock::TickBasedSteadyClockCore tick_based_steady_clock_core; - Clock::StandardLocalSystemClockCore standard_local_system_clock_core; - Clock::StandardNetworkSystemClockCore standard_network_system_clock_core; - Clock::StandardUserSystemClockCore standard_user_system_clock_core; - Clock::EphemeralNetworkSystemClockCore ephemeral_network_system_clock_core; - - std::shared_ptr<Clock::LocalSystemClockContextWriter> local_system_clock_context_writer; - std::shared_ptr<Clock::NetworkSystemClockContextWriter> network_system_clock_context_writer; - std::shared_ptr<Clock::EphemeralNetworkSystemClockContextWriter> - ephemeral_network_system_clock_context_writer; - - TimeZone::TimeZoneContentManager time_zone_content_manager; + Core::System& system; + + struct Impl; + std::unique_ptr<Impl> impl; }; } // namespace Service::Time diff --git a/src/core/hle/service/time/time_zone_content_manager.cpp b/src/core/hle/service/time/time_zone_content_manager.cpp index 320672add..4177d0a41 100644 --- a/src/core/hle/service/time/time_zone_content_manager.cpp +++ b/src/core/hle/service/time/time_zone_content_manager.cpp @@ -68,9 +68,10 @@ static std::vector<std::string> BuildLocationNameCache(Core::System& system) { return location_name_cache; } -TimeZoneContentManager::TimeZoneContentManager(TimeManager& time_manager, Core::System& system) - : system{system}, location_name_cache{BuildLocationNameCache(system)} { +TimeZoneContentManager::TimeZoneContentManager(Core::System& system) + : system{system}, location_name_cache{BuildLocationNameCache(system)} {} +void TimeZoneContentManager::Initialize(TimeManager& time_manager) { std::string location_name; const auto timezone_setting = Settings::GetTimeZoneString(); if (timezone_setting == "auto" || timezone_setting == "default") { diff --git a/src/core/hle/service/time/time_zone_content_manager.h b/src/core/hle/service/time/time_zone_content_manager.h index 4f302c3b9..52dd1a020 100644 --- a/src/core/hle/service/time/time_zone_content_manager.h +++ b/src/core/hle/service/time/time_zone_content_manager.h @@ -21,7 +21,9 @@ namespace Service::Time::TimeZone { class TimeZoneContentManager final { public: - TimeZoneContentManager(TimeManager& time_manager, Core::System& system); + explicit TimeZoneContentManager(Core::System& system); + + void Initialize(TimeManager& time_manager); TimeZoneManager& GetTimeZoneManager() { return time_zone_manager; diff --git a/src/core/hle/service/time/time_zone_manager.cpp b/src/core/hle/service/time/time_zone_manager.cpp index 69152d0ac..bdf0439f2 100644 --- a/src/core/hle/service/time/time_zone_manager.cpp +++ b/src/core/hle/service/time/time_zone_manager.cpp @@ -820,7 +820,10 @@ static ResultCode ToCalendarTimeImpl(const TimeZoneRule& rules, s64 time, Calend const ResultCode result{ ToCalendarTimeInternal(rules, time, calendar_time, calendar.additiona_info)}; calendar.time.year = static_cast<s16>(calendar_time.year); - calendar.time.month = calendar_time.month + 1; // Internal impl. uses 0-indexed month + + // Internal impl. uses 0-indexed month + calendar.time.month = static_cast<s8>(calendar_time.month + 1); + calendar.time.day = calendar_time.day; calendar.time.hour = calendar_time.hour; calendar.time.minute = calendar_time.minute; @@ -872,13 +875,15 @@ ResultCode TimeZoneManager::ToPosixTime(const TimeZoneRule& rules, const CalendarTime& calendar_time, s64& posix_time) const { posix_time = 0; - CalendarTimeInternal internal_time{}; - internal_time.year = calendar_time.year; - internal_time.month = calendar_time.month - 1; // Internal impl. uses 0-indexed month - internal_time.day = calendar_time.day; - internal_time.hour = calendar_time.hour; - internal_time.minute = calendar_time.minute; - internal_time.second = calendar_time.second; + CalendarTimeInternal internal_time{ + .year = calendar_time.year, + // Internal impl. uses 0-indexed month + .month = static_cast<s8>(calendar_time.month - 1), + .day = calendar_time.day, + .hour = calendar_time.hour, + .minute = calendar_time.minute, + .second = calendar_time.second, + }; s32 hour{internal_time.hour}; s32 minute{internal_time.minute}; diff --git a/src/core/hle/service/time/time_zone_service.cpp b/src/core/hle/service/time/time_zone_service.cpp index ff3a10b3e..25cecbc83 100644 --- a/src/core/hle/service/time/time_zone_service.cpp +++ b/src/core/hle/service/time/time_zone_service.cpp @@ -10,8 +10,9 @@ namespace Service::Time { -ITimeZoneService ::ITimeZoneService(TimeZone::TimeZoneContentManager& time_zone_content_manager) - : ServiceFramework("ITimeZoneService"), time_zone_content_manager{time_zone_content_manager} { +ITimeZoneService ::ITimeZoneService(Core::System& system_, + TimeZone::TimeZoneContentManager& time_zone_manager_) + : ServiceFramework{system_, "ITimeZoneService"}, time_zone_content_manager{time_zone_manager_} { static const FunctionInfo functions[] = { {0, &ITimeZoneService::GetDeviceLocationName, "GetDeviceLocationName"}, {1, nullptr, "SetDeviceLocationName"}, diff --git a/src/core/hle/service/time/time_zone_service.h b/src/core/hle/service/time/time_zone_service.h index cb495748b..2c9b97603 100644 --- a/src/core/hle/service/time/time_zone_service.h +++ b/src/core/hle/service/time/time_zone_service.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Service::Time { namespace TimeZone { @@ -14,7 +18,8 @@ class TimeZoneContentManager; class ITimeZoneService final : public ServiceFramework<ITimeZoneService> { public: - explicit ITimeZoneService(TimeZone::TimeZoneContentManager& time_zone_manager); + explicit ITimeZoneService(Core::System& system_, + TimeZone::TimeZoneContentManager& time_zone_manager_); private: void GetDeviceLocationName(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/usb/usb.cpp b/src/core/hle/service/usb/usb.cpp index d033f8603..579de83e4 100644 --- a/src/core/hle/service/usb/usb.cpp +++ b/src/core/hle/service/usb/usb.cpp @@ -15,7 +15,7 @@ namespace Service::USB { class IDsInterface final : public ServiceFramework<IDsInterface> { public: - explicit IDsInterface() : ServiceFramework{"IDsInterface"} { + explicit IDsInterface(Core::System& system_) : ServiceFramework{system_, "IDsInterface"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetDsEndpoint"}, @@ -40,7 +40,7 @@ public: class USB_DS final : public ServiceFramework<USB_DS> { public: - explicit USB_DS() : ServiceFramework{"usb:ds"} { + explicit USB_DS(Core::System& system_) : ServiceFramework{system_, "usb:ds"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "BindDevice"}, @@ -65,7 +65,8 @@ public: class IClientEpSession final : public ServiceFramework<IClientEpSession> { public: - explicit IClientEpSession() : ServiceFramework{"IClientEpSession"} { + explicit IClientEpSession(Core::System& system_) + : ServiceFramework{system_, "IClientEpSession"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Open"}, @@ -86,7 +87,8 @@ public: class IClientIfSession final : public ServiceFramework<IClientIfSession> { public: - explicit IClientIfSession() : ServiceFramework{"IClientIfSession"} { + explicit IClientIfSession(Core::System& system_) + : ServiceFramework{system_, "IClientIfSession"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Unknown0"}, @@ -108,7 +110,7 @@ public: class USB_HS final : public ServiceFramework<USB_HS> { public: - explicit USB_HS() : ServiceFramework{"usb:hs"} { + explicit USB_HS(Core::System& system_) : ServiceFramework{system_, "usb:hs"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "BindClientProcess"}, @@ -129,7 +131,7 @@ public: class IPdSession final : public ServiceFramework<IPdSession> { public: - explicit IPdSession() : ServiceFramework{"IPdSession"} { + explicit IPdSession(Core::System& system_) : ServiceFramework{system_, "IPdSession"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "BindNoticeEvent"}, @@ -148,7 +150,7 @@ public: class USB_PD final : public ServiceFramework<USB_PD> { public: - explicit USB_PD() : ServiceFramework{"usb:pd"} { + explicit USB_PD(Core::System& system_) : ServiceFramework{system_, "usb:pd"} { // clang-format off static const FunctionInfo functions[] = { {0, &USB_PD::GetPdSession, "GetPdSession"}, @@ -164,13 +166,14 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IPdSession>(); + rb.PushIpcInterface<IPdSession>(system); } }; class IPdCradleSession final : public ServiceFramework<IPdCradleSession> { public: - explicit IPdCradleSession() : ServiceFramework{"IPdCradleSession"} { + explicit IPdCradleSession(Core::System& system_) + : ServiceFramework{system_, "IPdCradleSession"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "VdmUserWrite"}, @@ -191,7 +194,7 @@ public: class USB_PD_C final : public ServiceFramework<USB_PD_C> { public: - explicit USB_PD_C() : ServiceFramework{"usb:pd:c"} { + explicit USB_PD_C(Core::System& system_) : ServiceFramework{system_, "usb:pd:c"} { // clang-format off static const FunctionInfo functions[] = { {0, &USB_PD_C::GetPdCradleSession, "GetPdCradleSession"}, @@ -205,7 +208,7 @@ private: void GetPdCradleSession(Kernel::HLERequestContext& ctx) { IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IPdCradleSession>(); + rb.PushIpcInterface<IPdCradleSession>(system); LOG_DEBUG(Service_USB, "called"); } @@ -213,7 +216,7 @@ private: class USB_PM final : public ServiceFramework<USB_PM> { public: - explicit USB_PM() : ServiceFramework{"usb:pm"} { + explicit USB_PM(Core::System& system_) : ServiceFramework{system_, "usb:pm"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Unknown0"}, @@ -229,12 +232,12 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<USB_DS>()->InstallAsService(sm); - std::make_shared<USB_HS>()->InstallAsService(sm); - std::make_shared<USB_PD>()->InstallAsService(sm); - std::make_shared<USB_PD_C>()->InstallAsService(sm); - std::make_shared<USB_PM>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<USB_DS>(system)->InstallAsService(sm); + std::make_shared<USB_HS>(system)->InstallAsService(sm); + std::make_shared<USB_PD>(system)->InstallAsService(sm); + std::make_shared<USB_PD_C>(system)->InstallAsService(sm); + std::make_shared<USB_PM>(system)->InstallAsService(sm); } } // namespace Service::USB diff --git a/src/core/hle/service/usb/usb.h b/src/core/hle/service/usb/usb.h index 970a11fe8..fc366df34 100644 --- a/src/core/hle/service/usb/usb.h +++ b/src/core/hle/service/usb/usb.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::USB { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::USB diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp index 480d34725..968cd16b6 100644 --- a/src/core/hle/service/vi/vi.cpp +++ b/src/core/hle/service/vi/vi.cpp @@ -159,7 +159,7 @@ public: header.data_size = static_cast<u32_le>(write_index - sizeof(Header)); header.data_offset = sizeof(Header); header.objects_size = 4; - header.objects_offset = sizeof(Header) + header.data_size; + header.objects_offset = static_cast<u32>(sizeof(Header) + header.data_size); std::memcpy(buffer.data(), &header, sizeof(Header)); return buffer; @@ -215,10 +215,9 @@ public: explicit IGBPConnectRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPConnectRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); } @@ -279,23 +278,28 @@ public: : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPSetPreallocatedBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); - buffer = Read<NVFlinger::IGBPBuffer>(); + if (data.contains_object != 0) { + buffer_container = Read<BufferContainer>(); + } } struct Data { u32_le slot; - INSERT_PADDING_WORDS(1); + u32_le contains_object; + }; + + struct BufferContainer { u32_le graphic_buffer_length; INSERT_PADDING_WORDS(1); + NVFlinger::IGBPBuffer buffer{}; }; - Data data; - NVFlinger::IGBPBuffer buffer; + Data data{}; + BufferContainer buffer_container{}; }; class IGBPSetPreallocatedBufferResponseParcel : public Parcel { @@ -306,15 +310,40 @@ protected: } }; +class IGBPCancelBufferRequestParcel : public Parcel { +public: + explicit IGBPCancelBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { + Deserialize(); + } + + void DeserializeData() override { + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); + data = Read<Data>(); + } + + struct Data { + u32_le slot; + Service::Nvidia::MultiFence multi_fence; + }; + + Data data; +}; + +class IGBPCancelBufferResponseParcel : public Parcel { +protected: + void SerializeData() override { + Write<u32>(0); // Success + } +}; + class IGBPDequeueBufferRequestParcel : public Parcel { public: explicit IGBPDequeueBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPDequeueBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); } @@ -333,7 +362,6 @@ class IGBPDequeueBufferResponseParcel : public Parcel { public: explicit IGBPDequeueBufferResponseParcel(u32 slot, Service::Nvidia::MultiFence& multi_fence) : slot(slot), multi_fence(multi_fence) {} - ~IGBPDequeueBufferResponseParcel() override = default; protected: void SerializeData() override { @@ -352,10 +380,9 @@ public: explicit IGBPRequestBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPRequestBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); slot = Read<u32_le>(); } @@ -384,10 +411,9 @@ public: explicit IGBPQueueBufferRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPQueueBufferRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); data = Read<Data>(); } @@ -447,10 +473,9 @@ public: explicit IGBPQueryRequestParcel(std::vector<u8> buffer) : Parcel(std::move(buffer)) { Deserialize(); } - ~IGBPQueryRequestParcel() override = default; void DeserializeData() override { - std::u16string token = ReadInterfaceToken(); + [[maybe_unused]] const std::u16string token = ReadInterfaceToken(); type = Read<u32_le>(); } @@ -473,8 +498,8 @@ private: class IHOSBinderDriver final : public ServiceFramework<IHOSBinderDriver> { public: - explicit IHOSBinderDriver(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) - : ServiceFramework("IHOSBinderDriver"), nv_flinger(std::move(nv_flinger)) { + explicit IHOSBinderDriver(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_) + : ServiceFramework{system_, "IHOSBinderDriver"}, nv_flinger(nv_flinger_) { static const FunctionInfo functions[] = { {0, &IHOSBinderDriver::TransactParcel, "TransactParcel"}, {1, &IHOSBinderDriver::AdjustRefcount, "AdjustRefcount"}, @@ -509,10 +534,9 @@ private: const u32 flags = rp.Pop<u32>(); LOG_DEBUG(Service_VI, "called. id=0x{:08X} transaction={:X}, flags=0x{:08X}", id, - static_cast<u32>(transaction), flags); + transaction, flags); - const auto guard = nv_flinger->Lock(); - auto& buffer_queue = nv_flinger->FindBufferQueue(id); + auto& buffer_queue = *nv_flinger.FindBufferQueue(id); switch (transaction) { case TransactionId::Connect: { @@ -522,13 +546,16 @@ private: Settings::values.resolution_factor.GetValue()), static_cast<u32>(static_cast<u32>(DisplayResolution::UndockedHeight) * Settings::values.resolution_factor.GetValue())}; + + buffer_queue.Connect(); + ctx.WriteBuffer(response.Serialize()); break; } case TransactionId::SetPreallocatedBuffer: { IGBPSetPreallocatedBufferRequestParcel request{ctx.ReadBuffer()}; - buffer_queue.SetPreallocatedBuffer(request.data.slot, request.buffer); + buffer_queue.SetPreallocatedBuffer(request.data.slot, request.buffer_container.buffer); IGBPSetPreallocatedBufferResponseParcel response{}; ctx.WriteBuffer(response.Serialize()); @@ -538,40 +565,25 @@ private: IGBPDequeueBufferRequestParcel request{ctx.ReadBuffer()}; const u32 width{request.data.width}; const u32 height{request.data.height}; - auto result = buffer_queue.DequeueBuffer(width, height); - - if (result) { - // Buffer is available - IGBPDequeueBufferResponseParcel response{result->first, *result->second}; - ctx.WriteBuffer(response.Serialize()); - } else { - // Wait the current thread until a buffer becomes available - ctx.SleepClientThread( - "IHOSBinderDriver::DequeueBuffer", UINT64_MAX, - [=, this](std::shared_ptr<Kernel::Thread> thread, - Kernel::HLERequestContext& ctx, Kernel::ThreadWakeupReason reason) { - // Repeat TransactParcel DequeueBuffer when a buffer is available - const auto guard = nv_flinger->Lock(); - auto& buffer_queue = nv_flinger->FindBufferQueue(id); - auto result = buffer_queue.DequeueBuffer(width, height); - ASSERT_MSG(result != std::nullopt, "Could not dequeue buffer."); - - IGBPDequeueBufferResponseParcel response{result->first, *result->second}; - ctx.WriteBuffer(response.Serialize()); - IPC::ResponseBuilder rb{ctx, 2}; - rb.Push(RESULT_SUCCESS); - }, - buffer_queue.GetWritableBufferWaitEvent()); - } + + do { + if (auto result = buffer_queue.DequeueBuffer(width, height); result) { + // Buffer is available + IGBPDequeueBufferResponseParcel response{result->first, *result->second}; + ctx.WriteBuffer(response.Serialize()); + break; + } + } while (buffer_queue.IsConnected()); + break; } case TransactionId::RequestBuffer: { IGBPRequestBufferRequestParcel request{ctx.ReadBuffer()}; auto& buffer = buffer_queue.RequestBuffer(request.slot); - IGBPRequestBufferResponseParcel response{buffer}; ctx.WriteBuffer(response.Serialize()); + break; } case TransactionId::QueueBuffer: { @@ -596,7 +608,12 @@ private: break; } case TransactionId::CancelBuffer: { - LOG_CRITICAL(Service_VI, "(STUBBED) called, transaction=CancelBuffer"); + IGBPCancelBufferRequestParcel request{ctx.ReadBuffer()}; + + buffer_queue.CancelBuffer(request.data.slot, request.data.multi_fence); + + IGBPCancelBufferResponseParcel response{}; + ctx.WriteBuffer(response.Serialize()); break; } case TransactionId::Disconnect: { @@ -652,7 +669,7 @@ private: LOG_WARNING(Service_VI, "(STUBBED) called id={}, unknown={:08X}", id, unknown); - const auto& buffer_queue = nv_flinger->FindBufferQueue(id); + const auto& buffer_queue = *nv_flinger.FindBufferQueue(id); // TODO(Subv): Find out what this actually is. IPC::ResponseBuilder rb{ctx, 2, 1}; @@ -660,12 +677,13 @@ private: rb.PushCopyObjects(buffer_queue.GetBufferWaitEvent()); } - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger; -}; // namespace VI + NVFlinger::NVFlinger& nv_flinger; +}; class ISystemDisplayService final : public ServiceFramework<ISystemDisplayService> { public: - explicit ISystemDisplayService() : ServiceFramework("ISystemDisplayService") { + explicit ISystemDisplayService(Core::System& system_) + : ServiceFramework{system_, "ISystemDisplayService"} { static const FunctionInfo functions[] = { {1200, nullptr, "GetZOrderCountMin"}, {1202, nullptr, "GetZOrderCountMax"}, @@ -747,7 +765,7 @@ private: IPC::ResponseBuilder rb{ctx, 6}; rb.Push(RESULT_SUCCESS); - if (Settings::values.use_docked_mode) { + if (Settings::values.use_docked_mode.GetValue()) { rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedWidth) * static_cast<u32>(Settings::values.resolution_factor.GetValue())); rb.Push(static_cast<u32>(Service::VI::DisplayResolution::DockedHeight) * @@ -766,8 +784,8 @@ private: class IManagerDisplayService final : public ServiceFramework<IManagerDisplayService> { public: - explicit IManagerDisplayService(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) - : ServiceFramework("IManagerDisplayService"), nv_flinger(std::move(nv_flinger)) { + explicit IManagerDisplayService(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_) + : ServiceFramework{system_, "IManagerDisplayService"}, nv_flinger{nv_flinger_} { // clang-format off static const FunctionInfo functions[] = { {200, nullptr, "AllocateProcessHeapBlock"}, @@ -869,7 +887,7 @@ private: "(STUBBED) called. unknown=0x{:08X}, display=0x{:016X}, aruid=0x{:016X}", unknown, display, aruid); - const auto layer_id = nv_flinger->CreateLayer(display); + const auto layer_id = nv_flinger.CreateLayer(display); if (!layer_id) { LOG_ERROR(Service_VI, "Layer not found! display=0x{:016X}", display); IPC::ResponseBuilder rb{ctx, 2}; @@ -906,12 +924,12 @@ private: rb.Push(RESULT_SUCCESS); } - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger; + NVFlinger::NVFlinger& nv_flinger; }; class IApplicationDisplayService final : public ServiceFramework<IApplicationDisplayService> { public: - explicit IApplicationDisplayService(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger); + explicit IApplicationDisplayService(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_); private: enum class ConvertedScaleMode : u64 { @@ -935,7 +953,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IHOSBinderDriver>(nv_flinger); + rb.PushIpcInterface<IHOSBinderDriver>(system, nv_flinger); } void GetSystemDisplayService(Kernel::HLERequestContext& ctx) { @@ -943,7 +961,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ISystemDisplayService>(); + rb.PushIpcInterface<ISystemDisplayService>(system); } void GetManagerDisplayService(Kernel::HLERequestContext& ctx) { @@ -951,7 +969,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IManagerDisplayService>(nv_flinger); + rb.PushIpcInterface<IManagerDisplayService>(system, nv_flinger); } void GetIndirectDisplayTransactionService(Kernel::HLERequestContext& ctx) { @@ -959,7 +977,7 @@ private: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IHOSBinderDriver>(nv_flinger); + rb.PushIpcInterface<IHOSBinderDriver>(system, nv_flinger); } void OpenDisplay(Kernel::HLERequestContext& ctx) { @@ -986,7 +1004,7 @@ private: ASSERT_MSG(name == "Default", "Non-default displays aren't supported yet"); - const auto display_id = nv_flinger->OpenDisplay(name); + const auto display_id = nv_flinger.OpenDisplay(name); if (!display_id) { LOG_ERROR(Service_VI, "Display not found! display_name={}", name); IPC::ResponseBuilder rb{ctx, 2}; @@ -1041,8 +1059,8 @@ private: const auto scaling_mode = rp.PopEnum<NintendoScaleMode>(); const u64 unknown = rp.Pop<u64>(); - LOG_DEBUG(Service_VI, "called. scaling_mode=0x{:08X}, unknown=0x{:016X}", - static_cast<u32>(scaling_mode), unknown); + LOG_DEBUG(Service_VI, "called. scaling_mode=0x{:08X}, unknown=0x{:016X}", scaling_mode, + unknown); IPC::ResponseBuilder rb{ctx, 2}; @@ -1086,7 +1104,7 @@ private: LOG_DEBUG(Service_VI, "called. layer_id=0x{:016X}, aruid=0x{:016X}", layer_id, aruid); - const auto display_id = nv_flinger->OpenDisplay(display_name); + const auto display_id = nv_flinger.OpenDisplay(display_name); if (!display_id) { LOG_ERROR(Service_VI, "Layer not found! layer_id={}", layer_id); IPC::ResponseBuilder rb{ctx, 2}; @@ -1094,7 +1112,7 @@ private: return; } - const auto buffer_queue_id = nv_flinger->FindBufferQueueId(*display_id, layer_id); + const auto buffer_queue_id = nv_flinger.FindBufferQueueId(*display_id, layer_id); if (!buffer_queue_id) { LOG_ERROR(Service_VI, "Buffer queue id not found! display_id={}", *display_id); IPC::ResponseBuilder rb{ctx, 2}; @@ -1114,7 +1132,7 @@ private: LOG_DEBUG(Service_VI, "called. layer_id=0x{:016X}", layer_id); - nv_flinger->CloseLayer(layer_id); + nv_flinger.CloseLayer(layer_id); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -1130,7 +1148,7 @@ private: // TODO(Subv): What's the difference between a Stray and a Managed layer? - const auto layer_id = nv_flinger->CreateLayer(display_id); + const auto layer_id = nv_flinger.CreateLayer(display_id); if (!layer_id) { LOG_ERROR(Service_VI, "Layer not found! layer_id={}", *layer_id); IPC::ResponseBuilder rb{ctx, 2}; @@ -1138,7 +1156,7 @@ private: return; } - const auto buffer_queue_id = nv_flinger->FindBufferQueueId(display_id, *layer_id); + const auto buffer_queue_id = nv_flinger.FindBufferQueueId(display_id, *layer_id); if (!buffer_queue_id) { LOG_ERROR(Service_VI, "Buffer queue id not found! display_id={}", display_id); IPC::ResponseBuilder rb{ctx, 2}; @@ -1169,7 +1187,7 @@ private: LOG_WARNING(Service_VI, "(STUBBED) called. display_id=0x{:016X}", display_id); - const auto vsync_event = nv_flinger->FindVsyncEvent(display_id); + const auto vsync_event = nv_flinger.FindVsyncEvent(display_id); if (!vsync_event) { LOG_ERROR(Service_VI, "Vsync event was not found for display_id={}", display_id); IPC::ResponseBuilder rb{ctx, 2}; @@ -1185,7 +1203,7 @@ private: void ConvertScalingMode(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; const auto mode = rp.PopEnum<NintendoScaleMode>(); - LOG_DEBUG(Service_VI, "called mode={}", static_cast<u32>(mode)); + LOG_DEBUG(Service_VI, "called mode={}", mode); const auto converted_mode = ConvertScalingModeImpl(mode); @@ -1205,8 +1223,8 @@ private: const auto height = rp.Pop<u64>(); LOG_DEBUG(Service_VI, "called width={}, height={}", width, height); - constexpr std::size_t base_size = 0x20000; - constexpr std::size_t alignment = 0x1000; + constexpr u64 base_size = 0x20000; + constexpr u64 alignment = 0x1000; const auto texture_size = width * height * 4; const auto out_size = (texture_size + base_size - 1) / base_size * base_size; @@ -1234,12 +1252,12 @@ private: } } - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger; + NVFlinger::NVFlinger& nv_flinger; }; -IApplicationDisplayService::IApplicationDisplayService( - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) - : ServiceFramework("IApplicationDisplayService"), nv_flinger(std::move(nv_flinger)) { +IApplicationDisplayService::IApplicationDisplayService(Core::System& system_, + NVFlinger::NVFlinger& nv_flinger_) + : ServiceFramework{system_, "IApplicationDisplayService"}, nv_flinger{nv_flinger_} { static const FunctionInfo functions[] = { {100, &IApplicationDisplayService::GetRelayService, "GetRelayService"}, {101, &IApplicationDisplayService::GetSystemDisplayService, "GetSystemDisplayService"}, @@ -1280,14 +1298,13 @@ static bool IsValidServiceAccess(Permission permission, Policy policy) { return false; } -void detail::GetDisplayServiceImpl(Kernel::HLERequestContext& ctx, - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger, - Permission permission) { +void detail::GetDisplayServiceImpl(Kernel::HLERequestContext& ctx, Core::System& system, + NVFlinger::NVFlinger& nv_flinger, Permission permission) { IPC::RequestParser rp{ctx}; const auto policy = rp.PopEnum<Policy>(); if (!IsValidServiceAccess(permission, policy)) { - LOG_ERROR(Service_VI, "Permission denied for policy {}", static_cast<u32>(policy)); + LOG_ERROR(Service_VI, "Permission denied for policy {}", policy); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERR_PERMISSION_DENIED); return; @@ -1295,14 +1312,14 @@ void detail::GetDisplayServiceImpl(Kernel::HLERequestContext& ctx, IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<IApplicationDisplayService>(std::move(nv_flinger)); + rb.PushIpcInterface<IApplicationDisplayService>(system, nv_flinger); } -void InstallInterfaces(SM::ServiceManager& service_manager, - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) { - std::make_shared<VI_M>(nv_flinger)->InstallAsService(service_manager); - std::make_shared<VI_S>(nv_flinger)->InstallAsService(service_manager); - std::make_shared<VI_U>(nv_flinger)->InstallAsService(service_manager); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system, + NVFlinger::NVFlinger& nv_flinger) { + std::make_shared<VI_M>(system, nv_flinger)->InstallAsService(service_manager); + std::make_shared<VI_S>(system, nv_flinger)->InstallAsService(service_manager); + std::make_shared<VI_U>(system, nv_flinger)->InstallAsService(service_manager); } } // namespace Service::VI diff --git a/src/core/hle/service/vi/vi.h b/src/core/hle/service/vi/vi.h index 6b66f8b81..eec531d54 100644 --- a/src/core/hle/service/vi/vi.h +++ b/src/core/hle/service/vi/vi.h @@ -7,6 +7,10 @@ #include <memory> #include "common/common_types.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -43,12 +47,12 @@ enum class Policy { }; namespace detail { -void GetDisplayServiceImpl(Kernel::HLERequestContext& ctx, - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger, Permission permission); +void GetDisplayServiceImpl(Kernel::HLERequestContext& ctx, Core::System& system, + NVFlinger::NVFlinger& nv_flinger, Permission permission); } // namespace detail /// Registers all VI services with the specified service manager. -void InstallInterfaces(SM::ServiceManager& service_manager, - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger); +void InstallInterfaces(SM::ServiceManager& service_manager, Core::System& system, + NVFlinger::NVFlinger& nv_flinger); } // namespace Service::VI diff --git a/src/core/hle/service/vi/vi_m.cpp b/src/core/hle/service/vi/vi_m.cpp index 06070087f..87db1c416 100644 --- a/src/core/hle/service/vi/vi_m.cpp +++ b/src/core/hle/service/vi/vi_m.cpp @@ -8,8 +8,8 @@ namespace Service::VI { -VI_M::VI_M(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) - : ServiceFramework{"vi:m"}, nv_flinger{std::move(nv_flinger)} { +VI_M::VI_M(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_) + : ServiceFramework{system_, "vi:m"}, nv_flinger{nv_flinger_} { static const FunctionInfo functions[] = { {2, &VI_M::GetDisplayService, "GetDisplayService"}, {3, nullptr, "GetDisplayServiceWithProxyNameExchange"}, @@ -22,7 +22,7 @@ VI_M::~VI_M() = default; void VI_M::GetDisplayService(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_VI, "called"); - detail::GetDisplayServiceImpl(ctx, nv_flinger, Permission::Manager); + detail::GetDisplayServiceImpl(ctx, system, nv_flinger, Permission::Manager); } } // namespace Service::VI diff --git a/src/core/hle/service/vi/vi_m.h b/src/core/hle/service/vi/vi_m.h index 290e06689..d79c41beb 100644 --- a/src/core/hle/service/vi/vi_m.h +++ b/src/core/hle/service/vi/vi_m.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -18,13 +22,13 @@ namespace Service::VI { class VI_M final : public ServiceFramework<VI_M> { public: - explicit VI_M(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger); + explicit VI_M(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_); ~VI_M() override; private: void GetDisplayService(Kernel::HLERequestContext& ctx); - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger; + NVFlinger::NVFlinger& nv_flinger; }; } // namespace Service::VI diff --git a/src/core/hle/service/vi/vi_s.cpp b/src/core/hle/service/vi/vi_s.cpp index 57c596cc4..5cd22f7df 100644 --- a/src/core/hle/service/vi/vi_s.cpp +++ b/src/core/hle/service/vi/vi_s.cpp @@ -8,8 +8,8 @@ namespace Service::VI { -VI_S::VI_S(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) - : ServiceFramework{"vi:s"}, nv_flinger{std::move(nv_flinger)} { +VI_S::VI_S(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_) + : ServiceFramework{system_, "vi:s"}, nv_flinger{nv_flinger_} { static const FunctionInfo functions[] = { {1, &VI_S::GetDisplayService, "GetDisplayService"}, {3, nullptr, "GetDisplayServiceWithProxyNameExchange"}, @@ -22,7 +22,7 @@ VI_S::~VI_S() = default; void VI_S::GetDisplayService(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_VI, "called"); - detail::GetDisplayServiceImpl(ctx, nv_flinger, Permission::System); + detail::GetDisplayServiceImpl(ctx, system, nv_flinger, Permission::System); } } // namespace Service::VI diff --git a/src/core/hle/service/vi/vi_s.h b/src/core/hle/service/vi/vi_s.h index 47804dc0b..5f1f8f290 100644 --- a/src/core/hle/service/vi/vi_s.h +++ b/src/core/hle/service/vi/vi_s.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -18,13 +22,13 @@ namespace Service::VI { class VI_S final : public ServiceFramework<VI_S> { public: - explicit VI_S(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger); + explicit VI_S(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_); ~VI_S() override; private: void GetDisplayService(Kernel::HLERequestContext& ctx); - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger; + NVFlinger::NVFlinger& nv_flinger; }; } // namespace Service::VI diff --git a/src/core/hle/service/vi/vi_u.cpp b/src/core/hle/service/vi/vi_u.cpp index 6b7329345..0079d51f0 100644 --- a/src/core/hle/service/vi/vi_u.cpp +++ b/src/core/hle/service/vi/vi_u.cpp @@ -8,8 +8,8 @@ namespace Service::VI { -VI_U::VI_U(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger) - : ServiceFramework{"vi:u"}, nv_flinger{std::move(nv_flinger)} { +VI_U::VI_U(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_) + : ServiceFramework{system_, "vi:u"}, nv_flinger{nv_flinger_} { static const FunctionInfo functions[] = { {0, &VI_U::GetDisplayService, "GetDisplayService"}, {1, nullptr, "GetDisplayServiceWithProxyNameExchange"}, @@ -22,7 +22,7 @@ VI_U::~VI_U() = default; void VI_U::GetDisplayService(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_VI, "called"); - detail::GetDisplayServiceImpl(ctx, nv_flinger, Permission::User); + detail::GetDisplayServiceImpl(ctx, system, nv_flinger, Permission::User); } } // namespace Service::VI diff --git a/src/core/hle/service/vi/vi_u.h b/src/core/hle/service/vi/vi_u.h index 19bdb73b0..8e3885c73 100644 --- a/src/core/hle/service/vi/vi_u.h +++ b/src/core/hle/service/vi/vi_u.h @@ -6,6 +6,10 @@ #include "core/hle/service/service.h" +namespace Core { +class System; +} + namespace Kernel { class HLERequestContext; } @@ -18,13 +22,13 @@ namespace Service::VI { class VI_U final : public ServiceFramework<VI_U> { public: - explicit VI_U(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger); + explicit VI_U(Core::System& system_, NVFlinger::NVFlinger& nv_flinger_); ~VI_U() override; private: void GetDisplayService(Kernel::HLERequestContext& ctx); - std::shared_ptr<NVFlinger::NVFlinger> nv_flinger; + NVFlinger::NVFlinger& nv_flinger; }; } // namespace Service::VI diff --git a/src/core/hle/service/wlan/wlan.cpp b/src/core/hle/service/wlan/wlan.cpp index 0260d7dcf..ddbf04069 100644 --- a/src/core/hle/service/wlan/wlan.cpp +++ b/src/core/hle/service/wlan/wlan.cpp @@ -12,7 +12,7 @@ namespace Service::WLAN { class WLANInfra final : public ServiceFramework<WLANInfra> { public: - explicit WLANInfra() : ServiceFramework{"wlan:inf"} { + explicit WLANInfra(Core::System& system_) : ServiceFramework{system_, "wlan:inf"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "OpenMode"}, @@ -55,7 +55,7 @@ public: class WLANLocal final : public ServiceFramework<WLANLocal> { public: - explicit WLANLocal() : ServiceFramework{"wlan:lcl"} { + explicit WLANLocal(Core::System& system_) : ServiceFramework{system_, "wlan:lcl"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Unknown0"}, @@ -120,7 +120,7 @@ public: class WLANLocalGetFrame final : public ServiceFramework<WLANLocalGetFrame> { public: - explicit WLANLocalGetFrame() : ServiceFramework{"wlan:lg"} { + explicit WLANLocalGetFrame(Core::System& system_) : ServiceFramework{system_, "wlan:lg"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Unknown"}, @@ -133,7 +133,7 @@ public: class WLANSocketGetFrame final : public ServiceFramework<WLANSocketGetFrame> { public: - explicit WLANSocketGetFrame() : ServiceFramework{"wlan:sg"} { + explicit WLANSocketGetFrame(Core::System& system_) : ServiceFramework{system_, "wlan:sg"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Unknown"}, @@ -146,7 +146,7 @@ public: class WLANSocketManager final : public ServiceFramework<WLANSocketManager> { public: - explicit WLANSocketManager() : ServiceFramework{"wlan:soc"} { + explicit WLANSocketManager(Core::System& system_) : ServiceFramework{system_, "wlan:soc"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Unknown0"}, @@ -169,12 +169,12 @@ public: } }; -void InstallInterfaces(SM::ServiceManager& sm) { - std::make_shared<WLANInfra>()->InstallAsService(sm); - std::make_shared<WLANLocal>()->InstallAsService(sm); - std::make_shared<WLANLocalGetFrame>()->InstallAsService(sm); - std::make_shared<WLANSocketGetFrame>()->InstallAsService(sm); - std::make_shared<WLANSocketManager>()->InstallAsService(sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system) { + std::make_shared<WLANInfra>(system)->InstallAsService(sm); + std::make_shared<WLANLocal>(system)->InstallAsService(sm); + std::make_shared<WLANLocalGetFrame>(system)->InstallAsService(sm); + std::make_shared<WLANSocketGetFrame>(system)->InstallAsService(sm); + std::make_shared<WLANSocketManager>(system)->InstallAsService(sm); } } // namespace Service::WLAN diff --git a/src/core/hle/service/wlan/wlan.h b/src/core/hle/service/wlan/wlan.h index 054ea928a..3899eedbb 100644 --- a/src/core/hle/service/wlan/wlan.h +++ b/src/core/hle/service/wlan/wlan.h @@ -4,12 +4,16 @@ #pragma once +namespace Core { +class System; +} + namespace Service::SM { class ServiceManager; } namespace Service::WLAN { -void InstallInterfaces(SM::ServiceManager& sm); +void InstallInterfaces(SM::ServiceManager& sm, Core::System& system); } // namespace Service::WLAN diff --git a/src/core/loader/deconstructed_rom_directory.cpp b/src/core/loader/deconstructed_rom_directory.cpp index 394a1bf26..79ebf11de 100644 --- a/src/core/loader/deconstructed_rom_directory.cpp +++ b/src/core/loader/deconstructed_rom_directory.cpp @@ -12,7 +12,6 @@ #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" #include "core/file_sys/romfs_factory.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" @@ -114,7 +113,8 @@ AppLoader_DeconstructedRomDirectory::LoadResult AppLoader_DeconstructedRomDirect } if (override_update) { - const FileSys::PatchManager patch_manager(metadata.GetTitleID()); + const FileSys::PatchManager patch_manager( + metadata.GetTitleID(), system.GetFileSystemController(), system.GetContentProvider()); dir = patch_manager.PatchExeFS(dir); } @@ -160,7 +160,8 @@ AppLoader_DeconstructedRomDirectory::LoadResult AppLoader_DeconstructedRomDirect modules.clear(); const VAddr base_address{process.PageTable().GetCodeRegionStart()}; VAddr next_load_addr{base_address}; - const FileSys::PatchManager pm{metadata.GetTitleID()}; + const FileSys::PatchManager pm{metadata.GetTitleID(), system.GetFileSystemController(), + system.GetContentProvider()}; for (const auto& module : static_modules) { const FileSys::VirtualFile module_file{dir->GetFile(module)}; if (!module_file) { @@ -178,8 +179,6 @@ AppLoader_DeconstructedRomDirectory::LoadResult AppLoader_DeconstructedRomDirect next_load_addr = *tentative_next_load_addr; modules.insert_or_assign(load_addr, module); LOG_DEBUG(Loader, "loaded module {} @ 0x{:X}", module, load_addr); - // Register module with GDBStub - GDBStub::RegisterModule(module, load_addr, next_load_addr - 1, false); } // Find the RomFS by searching for a ".romfs" file in this directory diff --git a/src/core/loader/deconstructed_rom_directory.h b/src/core/loader/deconstructed_rom_directory.h index 35d340317..3c968580f 100644 --- a/src/core/loader/deconstructed_rom_directory.h +++ b/src/core/loader/deconstructed_rom_directory.h @@ -32,7 +32,7 @@ public: /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/elf.h b/src/core/loader/elf.h index 3527933ad..2067932c7 100644 --- a/src/core/loader/elf.h +++ b/src/core/loader/elf.h @@ -21,7 +21,7 @@ public: /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/kip.cpp b/src/core/loader/kip.cpp index 5981bcd21..e162c4ff0 100644 --- a/src/core/loader/kip.cpp +++ b/src/core/loader/kip.cpp @@ -5,7 +5,6 @@ #include <cstring> #include "core/file_sys/kernel_executable.h" #include "core/file_sys/program_metadata.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/code_set.h" #include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" @@ -16,7 +15,7 @@ namespace Loader { namespace { constexpr u32 PageAlignSize(u32 size) { - return (size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK; + return static_cast<u32>((size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK); } } // Anonymous namespace @@ -91,8 +90,6 @@ AppLoader::LoadResult AppLoader_KIP::Load(Kernel::Process& process, program_image.resize(PageAlignSize(kip->GetBSSOffset()) + kip->GetBSSSize()); codeset.DataSegment().size += kip->GetBSSSize(); - GDBStub::RegisterModule(kip->GetName(), base_address, base_address + program_image.size()); - codeset.memory = std::move(program_image); process.LoadModule(std::move(codeset), base_address); diff --git a/src/core/loader/kip.h b/src/core/loader/kip.h index dee05a7b5..14a85e295 100644 --- a/src/core/loader/kip.h +++ b/src/core/loader/kip.h @@ -23,7 +23,7 @@ public: /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/loader.cpp b/src/core/loader/loader.cpp index 9bc3a8840..e4f5fd40c 100644 --- a/src/core/loader/loader.cpp +++ b/src/core/loader/loader.cpp @@ -10,6 +10,7 @@ #include "common/file_util.h" #include "common/logging/log.h" #include "common/string_util.h" +#include "core/core.h" #include "core/hle/kernel/process.h" #include "core/loader/deconstructed_rom_directory.h" #include "core/loader/elf.h" @@ -184,6 +185,10 @@ constexpr std::array<const char*, 66> RESULT_MESSAGES{ "The INI file contains more than the maximum allowable number of KIP files.", }; +std::string GetResultStatusString(ResultStatus status) { + return RESULT_MESSAGES.at(static_cast<std::size_t>(status)); +} + std::ostream& operator<<(std::ostream& os, ResultStatus status) { os << RESULT_MESSAGES.at(static_cast<std::size_t>(status)); return os; @@ -194,15 +199,15 @@ AppLoader::~AppLoader() = default; /** * Get a loader for a file with a specific type - * @param file The file to load - * @param type The type of the file - * @param file the file to retrieve the loader for - * @param type the file type + * @param system The system context to use. + * @param file The file to retrieve the loader for + * @param type The file type + * @param program_index Specifies the index within the container of the program to launch. * @return std::unique_ptr<AppLoader> a pointer to a loader object; nullptr for unsupported type */ -static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileType type) { +static std::unique_ptr<AppLoader> GetFileLoader(Core::System& system, FileSys::VirtualFile file, + FileType type, std::size_t program_index) { switch (type) { - // Standard ELF file format. case FileType::ELF: return std::make_unique<AppLoader_ELF>(std::move(file)); @@ -221,7 +226,8 @@ static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileT // NX XCI (nX Card Image) file format. case FileType::XCI: - return std::make_unique<AppLoader_XCI>(std::move(file)); + return std::make_unique<AppLoader_XCI>(std::move(file), system.GetFileSystemController(), + system.GetContentProvider(), program_index); // NX NAX (NintendoAesXts) file format. case FileType::NAX: @@ -229,7 +235,8 @@ static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileT // NX NSP (Nintendo Submission Package) file format case FileType::NSP: - return std::make_unique<AppLoader_NSP>(std::move(file)); + return std::make_unique<AppLoader_NSP>(std::move(file), system.GetFileSystemController(), + system.GetContentProvider(), program_index); // NX KIP (Kernel Internal Process) file format case FileType::KIP: @@ -244,20 +251,22 @@ static std::unique_ptr<AppLoader> GetFileLoader(FileSys::VirtualFile file, FileT } } -std::unique_ptr<AppLoader> GetLoader(FileSys::VirtualFile file) { +std::unique_ptr<AppLoader> GetLoader(Core::System& system, FileSys::VirtualFile file, + std::size_t program_index) { FileType type = IdentifyFile(file); - FileType filename_type = GuessFromFilename(file->GetName()); + const FileType filename_type = GuessFromFilename(file->GetName()); // Special case: 00 is either a NCA or NAX. if (type != filename_type && !(file->GetName() == "00" && type == FileType::NAX)) { LOG_WARNING(Loader, "File {} has a different type than its extension.", file->GetName()); - if (FileType::Unknown == type) + if (FileType::Unknown == type) { type = filename_type; + } } LOG_DEBUG(Loader, "Loading file {} as {}...", file->GetName(), GetFileTypeString(type)); - return GetFileLoader(std::move(file), type); + return GetFileLoader(system, std::move(file), type, program_index); } } // namespace Loader diff --git a/src/core/loader/loader.h b/src/core/loader/loader.h index ac60b097a..b2e5b13de 100644 --- a/src/core/loader/loader.h +++ b/src/core/loader/loader.h @@ -135,6 +135,7 @@ enum class ResultStatus : u16 { ErrorINITooManyKIPs, }; +std::string GetResultStatusString(ResultStatus status); std::ostream& operator<<(std::ostream& os, ResultStatus status); /// Interface for loading an application @@ -290,9 +291,14 @@ protected: /** * Identifies a bootable file and return a suitable loader - * @param file The bootable file - * @return the best loader for this file + * + * @param system The system context. + * @param file The bootable file. + * @param program_index Specifies the index within the container of the program to launch. + * + * @return the best loader for this file. */ -std::unique_ptr<AppLoader> GetLoader(FileSys::VirtualFile file); +std::unique_ptr<AppLoader> GetLoader(Core::System& system, FileSys::VirtualFile file, + std::size_t program_index = 0); } // namespace Loader diff --git a/src/core/loader/nax.h b/src/core/loader/nax.h index c2b7722b5..a5b5e2ae1 100644 --- a/src/core/loader/nax.h +++ b/src/core/loader/nax.h @@ -28,7 +28,7 @@ public: /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/nca.h b/src/core/loader/nca.h index 711070294..918792800 100644 --- a/src/core/loader/nca.h +++ b/src/core/loader/nca.h @@ -28,7 +28,7 @@ public: /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/nro.cpp b/src/core/loader/nro.cpp index 9fb5eddad..ccf8cc153 100644 --- a/src/core/loader/nro.cpp +++ b/src/core/loader/nro.cpp @@ -14,10 +14,10 @@ #include "core/file_sys/control_metadata.h" #include "core/file_sys/romfs_factory.h" #include "core/file_sys/vfs_offset.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/code_set.h" #include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" +#include "core/hle/kernel/thread.h" #include "core/hle/service/filesystem/filesystem.h" #include "core/loader/nro.h" #include "core/loader/nso.h" @@ -127,7 +127,7 @@ FileType AppLoader_NRO::IdentifyType(const FileSys::VirtualFile& file) { } static constexpr u32 PageAlignSize(u32 size) { - return (size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK; + return static_cast<u32>((size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK); } static bool LoadNroImpl(Kernel::Process& process, const std::vector<u8>& data, @@ -197,10 +197,6 @@ static bool LoadNroImpl(Kernel::Process& process, const std::vector<u8>& data, codeset.memory = std::move(program_image); process.LoadModule(std::move(codeset), process.PageTable().GetCodeRegionStart()); - // Register module with GDBStub - GDBStub::RegisterModule(name, process.PageTable().GetCodeRegionStart(), - process.PageTable().GetCodeRegionEnd()); - return true; } diff --git a/src/core/loader/nro.h b/src/core/loader/nro.h index a2aab2ecc..a82b66221 100644 --- a/src/core/loader/nro.h +++ b/src/core/loader/nro.h @@ -32,7 +32,7 @@ public: /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp index 1e70f6e11..95b6f339a 100644 --- a/src/core/loader/nso.cpp +++ b/src/core/loader/nso.cpp @@ -14,10 +14,10 @@ #include "common/swap.h" #include "core/core.h" #include "core/file_sys/patch_manager.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/code_set.h" #include "core/hle/kernel/memory/page_table.h" #include "core/hle/kernel/process.h" +#include "core/hle/kernel/thread.h" #include "core/loader/nso.h" #include "core/memory.h" #include "core/settings.h" @@ -47,7 +47,7 @@ std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data, } constexpr u32 PageAlignSize(u32 size) { - return (size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK; + return static_cast<u32>((size + Core::Memory::PAGE_MASK) & ~Core::Memory::PAGE_MASK); } } // Anonymous namespace @@ -149,7 +149,7 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, Core::S // Apply cheats if they exist and the program has a valid title ID if (pm) { system.SetCurrentProcessBuildID(nso_header.build_id); - const auto cheats = pm->CreateCheatList(system, nso_header.build_id); + const auto cheats = pm->CreateCheatList(nso_header.build_id); if (!cheats.empty()) { system.RegisterCheatList(cheats, nso_header.build_id, load_base, image_size); } @@ -159,9 +159,6 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, Core::S codeset.memory = std::move(program_image); process.LoadModule(std::move(codeset), load_base); - // Register module with GDBStub - GDBStub::RegisterModule(file.GetName(), load_base, load_base); - return load_base + image_size; } diff --git a/src/core/loader/nso.h b/src/core/loader/nso.h index 4bd47787d..3af461b5f 100644 --- a/src/core/loader/nso.h +++ b/src/core/loader/nso.h @@ -59,7 +59,7 @@ struct NSOHeader { static_assert(sizeof(NSOHeader) == 0x100, "NSOHeader has incorrect size."); static_assert(std::is_trivially_copyable_v<NSOHeader>, "NSOHeader must be trivially copyable."); -constexpr u64 NSO_ARGUMENT_DATA_ALLOCATION_SIZE = 0x9000; +constexpr u32 NSO_ARGUMENT_DATA_ALLOCATION_SIZE = 0x9000; struct NSOArgumentHeader { u32_le allocated_size; @@ -75,7 +75,7 @@ public: /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/nsp.cpp b/src/core/loader/nsp.cpp index 15e528fa8..928f64c8c 100644 --- a/src/core/loader/nsp.cpp +++ b/src/core/loader/nsp.cpp @@ -21,26 +21,34 @@ namespace Loader { -AppLoader_NSP::AppLoader_NSP(FileSys::VirtualFile file) - : AppLoader(file), nsp(std::make_unique<FileSys::NSP>(file)), +AppLoader_NSP::AppLoader_NSP(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider, + std::size_t program_index) + : AppLoader(file), nsp(std::make_unique<FileSys::NSP>(file, program_index)), title_id(nsp->GetProgramTitleID()) { - if (nsp->GetStatus() != ResultStatus::Success) + if (nsp->GetStatus() != ResultStatus::Success) { return; + } if (nsp->IsExtractedType()) { secondary_loader = std::make_unique<AppLoader_DeconstructedRomDirectory>(nsp->GetExeFS()); } else { const auto control_nca = nsp->GetNCA(nsp->GetProgramTitleID(), FileSys::ContentRecordType::Control); - if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) + if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) { return; + } - std::tie(nacp_file, icon_file) = - FileSys::PatchManager(nsp->GetProgramTitleID()).ParseControlNCA(*control_nca); + std::tie(nacp_file, icon_file) = [this, &content_provider, &control_nca, &fsc] { + const FileSys::PatchManager pm{nsp->GetProgramTitleID(), fsc, content_provider}; + return pm.ParseControlNCA(*control_nca); + }(); - if (title_id == 0) + if (title_id == 0) { return; + } secondary_loader = std::make_unique<AppLoader_NCA>( nsp->GetNCAFile(title_id, FileSys::ContentRecordType::Program)); diff --git a/src/core/loader/nsp.h b/src/core/loader/nsp.h index b27deb686..d48d87f2c 100644 --- a/src/core/loader/nsp.h +++ b/src/core/loader/nsp.h @@ -9,15 +9,16 @@ #include "core/file_sys/vfs.h" #include "core/loader/loader.h" -namespace Core { -class System; -} - namespace FileSys { +class ContentProvider; class NACP; class NSP; } // namespace FileSys +namespace Service::FileSystem { +class FileSystemController; +} + namespace Loader { class AppLoader_NCA; @@ -25,12 +26,15 @@ class AppLoader_NCA; /// Loads an XCI file class AppLoader_NSP final : public AppLoader { public: - explicit AppLoader_NSP(FileSys::VirtualFile file); + explicit AppLoader_NSP(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider, + std::size_t program_index); ~AppLoader_NSP() override; /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/loader/xci.cpp b/src/core/loader/xci.cpp index 25e83af0f..aaa250cea 100644 --- a/src/core/loader/xci.cpp +++ b/src/core/loader/xci.cpp @@ -20,18 +20,25 @@ namespace Loader { -AppLoader_XCI::AppLoader_XCI(FileSys::VirtualFile file) - : AppLoader(file), xci(std::make_unique<FileSys::XCI>(file)), +AppLoader_XCI::AppLoader_XCI(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider, + std::size_t program_index) + : AppLoader(file), xci(std::make_unique<FileSys::XCI>(file, program_index)), nca_loader(std::make_unique<AppLoader_NCA>(xci->GetProgramNCAFile())) { - if (xci->GetStatus() != ResultStatus::Success) + if (xci->GetStatus() != ResultStatus::Success) { return; + } const auto control_nca = xci->GetNCAByType(FileSys::NCAContentType::Control); - if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) + if (control_nca == nullptr || control_nca->GetStatus() != ResultStatus::Success) { return; + } - std::tie(nacp_file, icon_file) = - FileSys::PatchManager(xci->GetProgramTitleID()).ParseControlNCA(*control_nca); + std::tie(nacp_file, icon_file) = [this, &content_provider, &control_nca, &fsc] { + const FileSys::PatchManager pm{xci->GetProgramTitleID(), fsc, content_provider}; + return pm.ParseControlNCA(*control_nca); + }(); } AppLoader_XCI::~AppLoader_XCI() = default; diff --git a/src/core/loader/xci.h b/src/core/loader/xci.h index 04aea286f..9f0ceb5ef 100644 --- a/src/core/loader/xci.h +++ b/src/core/loader/xci.h @@ -9,15 +9,16 @@ #include "core/file_sys/vfs.h" #include "core/loader/loader.h" -namespace Core { -class System; -} - namespace FileSys { +class ContentProvider; class NACP; class XCI; } // namespace FileSys +namespace Service::FileSystem { +class FileSystemController; +} + namespace Loader { class AppLoader_NCA; @@ -25,12 +26,15 @@ class AppLoader_NCA; /// Loads an XCI file class AppLoader_XCI final : public AppLoader { public: - explicit AppLoader_XCI(FileSys::VirtualFile file); + explicit AppLoader_XCI(FileSys::VirtualFile file, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider, + std::size_t program_index); ~AppLoader_XCI() override; /** * Returns the type of the file - * @param file std::shared_ptr<VfsFile> open file + * @param file open file * @return FileType found, or FileType::Error if this loader doesn't know it */ static FileType IdentifyType(const FileSys::VirtualFile& file); diff --git a/src/core/memory.cpp b/src/core/memory.cpp index c3f4829d7..11609682a 100644 --- a/src/core/memory.cpp +++ b/src/core/memory.cpp @@ -44,44 +44,16 @@ struct Memory::Impl { MapPages(page_table, base / PAGE_SIZE, size / PAGE_SIZE, target, Common::PageType::Memory); } - void MapIoRegion(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer mmio_handler) { - UNIMPLEMENTED(); - } - void UnmapRegion(Common::PageTable& page_table, VAddr base, u64 size) { ASSERT_MSG((size & PAGE_MASK) == 0, "non-page aligned size: {:016X}", size); ASSERT_MSG((base & PAGE_MASK) == 0, "non-page aligned base: {:016X}", base); MapPages(page_table, base / PAGE_SIZE, size / PAGE_SIZE, 0, Common::PageType::Unmapped); } - void AddDebugHook(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer hook) { - UNIMPLEMENTED(); - } - - void RemoveDebugHook(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer hook) { - UNIMPLEMENTED(); - } - bool IsValidVirtualAddress(const Kernel::Process& process, const VAddr vaddr) const { const auto& page_table = process.PageTable().PageTableImpl(); - - const u8* const page_pointer = page_table.pointers[vaddr >> PAGE_BITS]; - if (page_pointer != nullptr) { - return true; - } - - if (page_table.attributes[vaddr >> PAGE_BITS] == Common::PageType::RasterizerCachedMemory) { - return true; - } - - if (page_table.attributes[vaddr >> PAGE_BITS] != Common::PageType::Special) { - return false; - } - - return false; + const auto [pointer, type] = page_table.pointers[vaddr >> PAGE_BITS].PointerType(); + return pointer != nullptr || type == Common::PageType::RasterizerCachedMemory; } bool IsValidVirtualAddress(VAddr vaddr) const { @@ -99,17 +71,15 @@ struct Memory::Impl { } u8* GetPointer(const VAddr vaddr) const { - u8* const page_pointer{current_page_table->pointers[vaddr >> PAGE_BITS]}; - if (page_pointer) { - return page_pointer + vaddr; + const uintptr_t raw_pointer = current_page_table->pointers[vaddr >> PAGE_BITS].Raw(); + if (u8* const pointer = Common::PageTable::PageInfo::ExtractPointer(raw_pointer)) { + return pointer + vaddr; } - - if (current_page_table->attributes[vaddr >> PAGE_BITS] == - Common::PageType::RasterizerCachedMemory) { + const auto type = Common::PageTable::PageInfo::ExtractType(raw_pointer); + if (type == Common::PageType::RasterizerCachedMemory) { return GetPointerFromRasterizerCachedMemory(vaddr); } - - return {}; + return nullptr; } u8 Read8(const VAddr addr) { @@ -120,9 +90,9 @@ struct Memory::Impl { if ((addr & 1) == 0) { return Read<u16_le>(addr); } else { - const u8 a{Read<u8>(addr)}; - const u8 b{Read<u8>(addr + sizeof(u8))}; - return (static_cast<u16>(b) << 8) | a; + const u32 a{Read<u8>(addr)}; + const u32 b{Read<u8>(addr + sizeof(u8))}; + return static_cast<u16>((b << 8) | a); } } @@ -130,9 +100,9 @@ struct Memory::Impl { if ((addr & 3) == 0) { return Read<u32_le>(addr); } else { - const u16 a{Read16(addr)}; - const u16 b{Read16(addr + sizeof(u16))}; - return (static_cast<u32>(b) << 16) | a; + const u32 a{Read16(addr)}; + const u32 b{Read16(addr + sizeof(u16))}; + return (b << 16) | a; } } @@ -221,7 +191,8 @@ struct Memory::Impl { std::min(static_cast<std::size_t>(PAGE_SIZE) - page_offset, remaining_size); const auto current_vaddr = static_cast<VAddr>((page_index << PAGE_BITS) + page_offset); - switch (page_table.attributes[page_index]) { + const auto [pointer, type] = page_table.pointers[page_index].PointerType(); + switch (type) { case Common::PageType::Unmapped: { LOG_ERROR(HW_Memory, "Unmapped ReadBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})", @@ -230,10 +201,8 @@ struct Memory::Impl { break; } case Common::PageType::Memory: { - DEBUG_ASSERT(page_table.pointers[page_index]); - - const u8* const src_ptr = - page_table.pointers[page_index] + page_offset + (page_index << PAGE_BITS); + DEBUG_ASSERT(pointer); + const u8* const src_ptr = pointer + page_offset + (page_index << PAGE_BITS); std::memcpy(dest_buffer, src_ptr, copy_amount); break; } @@ -267,7 +236,8 @@ struct Memory::Impl { std::min(static_cast<std::size_t>(PAGE_SIZE) - page_offset, remaining_size); const auto current_vaddr = static_cast<VAddr>((page_index << PAGE_BITS) + page_offset); - switch (page_table.attributes[page_index]) { + const auto [pointer, type] = page_table.pointers[page_index].PointerType(); + switch (type) { case Common::PageType::Unmapped: { LOG_ERROR(HW_Memory, "Unmapped ReadBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})", @@ -276,10 +246,8 @@ struct Memory::Impl { break; } case Common::PageType::Memory: { - DEBUG_ASSERT(page_table.pointers[page_index]); - - const u8* const src_ptr = - page_table.pointers[page_index] + page_offset + (page_index << PAGE_BITS); + DEBUG_ASSERT(pointer); + const u8* const src_ptr = pointer + page_offset + (page_index << PAGE_BITS); std::memcpy(dest_buffer, src_ptr, copy_amount); break; } @@ -319,7 +287,8 @@ struct Memory::Impl { std::min(static_cast<std::size_t>(PAGE_SIZE) - page_offset, remaining_size); const auto current_vaddr = static_cast<VAddr>((page_index << PAGE_BITS) + page_offset); - switch (page_table.attributes[page_index]) { + const auto [pointer, type] = page_table.pointers[page_index].PointerType(); + switch (type) { case Common::PageType::Unmapped: { LOG_ERROR(HW_Memory, "Unmapped WriteBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})", @@ -327,10 +296,8 @@ struct Memory::Impl { break; } case Common::PageType::Memory: { - DEBUG_ASSERT(page_table.pointers[page_index]); - - u8* const dest_ptr = - page_table.pointers[page_index] + page_offset + (page_index << PAGE_BITS); + DEBUG_ASSERT(pointer); + u8* const dest_ptr = pointer + page_offset + (page_index << PAGE_BITS); std::memcpy(dest_ptr, src_buffer, copy_amount); break; } @@ -363,7 +330,8 @@ struct Memory::Impl { std::min(static_cast<std::size_t>(PAGE_SIZE) - page_offset, remaining_size); const auto current_vaddr = static_cast<VAddr>((page_index << PAGE_BITS) + page_offset); - switch (page_table.attributes[page_index]) { + const auto [pointer, type] = page_table.pointers[page_index].PointerType(); + switch (type) { case Common::PageType::Unmapped: { LOG_ERROR(HW_Memory, "Unmapped WriteBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})", @@ -371,10 +339,8 @@ struct Memory::Impl { break; } case Common::PageType::Memory: { - DEBUG_ASSERT(page_table.pointers[page_index]); - - u8* const dest_ptr = - page_table.pointers[page_index] + page_offset + (page_index << PAGE_BITS); + DEBUG_ASSERT(pointer); + u8* const dest_ptr = pointer + page_offset + (page_index << PAGE_BITS); std::memcpy(dest_ptr, src_buffer, copy_amount); break; } @@ -413,7 +379,8 @@ struct Memory::Impl { std::min(static_cast<std::size_t>(PAGE_SIZE) - page_offset, remaining_size); const auto current_vaddr = static_cast<VAddr>((page_index << PAGE_BITS) + page_offset); - switch (page_table.attributes[page_index]) { + const auto [pointer, type] = page_table.pointers[page_index].PointerType(); + switch (type) { case Common::PageType::Unmapped: { LOG_ERROR(HW_Memory, "Unmapped ZeroBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})", @@ -421,10 +388,8 @@ struct Memory::Impl { break; } case Common::PageType::Memory: { - DEBUG_ASSERT(page_table.pointers[page_index]); - - u8* dest_ptr = - page_table.pointers[page_index] + page_offset + (page_index << PAGE_BITS); + DEBUG_ASSERT(pointer); + u8* const dest_ptr = pointer + page_offset + (page_index << PAGE_BITS); std::memset(dest_ptr, 0, copy_amount); break; } @@ -460,7 +425,8 @@ struct Memory::Impl { std::min(static_cast<std::size_t>(PAGE_SIZE) - page_offset, remaining_size); const auto current_vaddr = static_cast<VAddr>((page_index << PAGE_BITS) + page_offset); - switch (page_table.attributes[page_index]) { + const auto [pointer, type] = page_table.pointers[page_index].PointerType(); + switch (type) { case Common::PageType::Unmapped: { LOG_ERROR(HW_Memory, "Unmapped CopyBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})", @@ -469,9 +435,8 @@ struct Memory::Impl { break; } case Common::PageType::Memory: { - DEBUG_ASSERT(page_table.pointers[page_index]); - const u8* src_ptr = - page_table.pointers[page_index] + page_offset + (page_index << PAGE_BITS); + DEBUG_ASSERT(pointer); + const u8* src_ptr = pointer + page_offset + (page_index << PAGE_BITS); WriteBlock(process, dest_addr, src_ptr, copy_amount); break; } @@ -501,16 +466,15 @@ struct Memory::Impl { if (vaddr == 0) { return; } - // Iterate over a contiguous CPU address space, which corresponds to the specified GPU // address space, marking the region as un/cached. The region is marked un/cached at a // granularity of CPU pages, hence why we iterate on a CPU page basis (note: GPU page size // is different). This assumes the specified GPU address region is contiguous as well. - u64 num_pages = ((vaddr + size - 1) >> PAGE_BITS) - (vaddr >> PAGE_BITS) + 1; - for (unsigned i = 0; i < num_pages; ++i, vaddr += PAGE_SIZE) { - Common::PageType& page_type{current_page_table->attributes[vaddr >> PAGE_BITS]}; - + const u64 num_pages = ((vaddr + size - 1) >> PAGE_BITS) - (vaddr >> PAGE_BITS) + 1; + for (u64 i = 0; i < num_pages; ++i, vaddr += PAGE_SIZE) { + const Common::PageType page_type{ + current_page_table->pointers[vaddr >> PAGE_BITS].Type()}; if (cached) { // Switch page type to cached if now cached switch (page_type) { @@ -519,8 +483,8 @@ struct Memory::Impl { // space, for example, a system module need not have a VRAM mapping. break; case Common::PageType::Memory: - page_type = Common::PageType::RasterizerCachedMemory; - current_page_table->pointers[vaddr >> PAGE_BITS] = nullptr; + current_page_table->pointers[vaddr >> PAGE_BITS].Store( + nullptr, Common::PageType::RasterizerCachedMemory); break; case Common::PageType::RasterizerCachedMemory: // There can be more than one GPU region mapped per CPU region, so it's common @@ -541,16 +505,16 @@ struct Memory::Impl { // that this area is already unmarked as cached. break; case Common::PageType::RasterizerCachedMemory: { - u8* pointer{GetPointerFromRasterizerCachedMemory(vaddr & ~PAGE_MASK)}; + u8* const pointer{GetPointerFromRasterizerCachedMemory(vaddr & ~PAGE_MASK)}; if (pointer == nullptr) { // It's possible that this function has been called while updating the // pagetable after unmapping a VMA. In that case the underlying VMA will no // longer exist, and we should just leave the pagetable entry blank. - page_type = Common::PageType::Unmapped; + current_page_table->pointers[vaddr >> PAGE_BITS].Store( + nullptr, Common::PageType::Unmapped); } else { - current_page_table->pointers[vaddr >> PAGE_BITS] = - pointer - (vaddr & ~PAGE_MASK); - page_type = Common::PageType::Memory; + current_page_table->pointers[vaddr >> PAGE_BITS].Store( + pointer - (vaddr & ~PAGE_MASK), Common::PageType::Memory); } break; } @@ -580,7 +544,7 @@ struct Memory::Impl { auto& gpu = system.GPU(); for (u64 i = 0; i < size; i++) { const auto page = base + i; - if (page_table.attributes[page] == Common::PageType::RasterizerCachedMemory) { + if (page_table.pointers[page].Type() == Common::PageType::RasterizerCachedMemory) { gpu.FlushAndInvalidateRegion(page << PAGE_BITS, PAGE_SIZE); } } @@ -595,20 +559,18 @@ struct Memory::Impl { "Mapping memory page without a pointer @ {:016x}", base * PAGE_SIZE); while (base != end) { - page_table.attributes[base] = type; - page_table.pointers[base] = nullptr; + page_table.pointers[base].Store(nullptr, type); page_table.backing_addr[base] = 0; base += 1; } } else { while (base != end) { - page_table.pointers[base] = - system.DeviceMemory().GetPointer(target) - (base << PAGE_BITS); - page_table.attributes[base] = type; + page_table.pointers[base].Store( + system.DeviceMemory().GetPointer(target) - (base << PAGE_BITS), type); page_table.backing_addr[base] = target - (base << PAGE_BITS); - ASSERT_MSG(page_table.pointers[base], + ASSERT_MSG(page_table.pointers[base].Pointer(), "memory mapping base yield a nullptr within the table"); base += 1; @@ -630,16 +592,14 @@ struct Memory::Impl { */ template <typename T> T Read(const VAddr vaddr) { - const u8* const page_pointer = current_page_table->pointers[vaddr >> PAGE_BITS]; - if (page_pointer != nullptr) { - // NOTE: Avoid adding any extra logic to this fast-path block + // Avoid adding any extra logic to this fast-path block + const uintptr_t raw_pointer = current_page_table->pointers[vaddr >> PAGE_BITS].Raw(); + if (const u8* const pointer = Common::PageTable::PageInfo::ExtractPointer(raw_pointer)) { T value; - std::memcpy(&value, &page_pointer[vaddr], sizeof(T)); + std::memcpy(&value, &pointer[vaddr], sizeof(T)); return value; } - - const Common::PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; - switch (type) { + switch (Common::PageTable::PageInfo::ExtractType(raw_pointer)) { case Common::PageType::Unmapped: LOG_ERROR(HW_Memory, "Unmapped Read{} @ 0x{:08X}", sizeof(T) * 8, vaddr); return 0; @@ -667,20 +627,16 @@ struct Memory::Impl { * @tparam T The data type to write to memory. This type *must* be * trivially copyable, otherwise the behavior of this function * is undefined. - * - * @returns The instance of T write to the specified virtual address. */ template <typename T> void Write(const VAddr vaddr, const T data) { - u8* const page_pointer = current_page_table->pointers[vaddr >> PAGE_BITS]; - if (page_pointer != nullptr) { - // NOTE: Avoid adding any extra logic to this fast-path block - std::memcpy(&page_pointer[vaddr], &data, sizeof(T)); + // Avoid adding any extra logic to this fast-path block + const uintptr_t raw_pointer = current_page_table->pointers[vaddr >> PAGE_BITS].Raw(); + if (u8* const pointer = Common::PageTable::PageInfo::ExtractPointer(raw_pointer)) { + std::memcpy(&pointer[vaddr], &data, sizeof(T)); return; } - - const Common::PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; - switch (type) { + switch (Common::PageTable::PageInfo::ExtractType(raw_pointer)) { case Common::PageType::Unmapped: LOG_ERROR(HW_Memory, "Unmapped Write{} 0x{:08X} @ 0x{:016X}", sizeof(data) * 8, static_cast<u32>(data), vaddr); @@ -701,15 +657,13 @@ struct Memory::Impl { template <typename T> bool WriteExclusive(const VAddr vaddr, const T data, const T expected) { - u8* page_pointer = current_page_table->pointers[vaddr >> PAGE_BITS]; - if (page_pointer != nullptr) { + const uintptr_t raw_pointer = current_page_table->pointers[vaddr >> PAGE_BITS].Raw(); + if (u8* const pointer = Common::PageTable::PageInfo::ExtractPointer(raw_pointer)) { // NOTE: Avoid adding any extra logic to this fast-path block - auto* pointer = reinterpret_cast<volatile T*>(&page_pointer[vaddr]); - return Common::AtomicCompareAndSwap(pointer, data, expected); + const auto volatile_pointer = reinterpret_cast<volatile T*>(&pointer[vaddr]); + return Common::AtomicCompareAndSwap(volatile_pointer, data, expected); } - - const Common::PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; - switch (type) { + switch (Common::PageTable::PageInfo::ExtractType(raw_pointer)) { case Common::PageType::Unmapped: LOG_ERROR(HW_Memory, "Unmapped Write{} 0x{:08X} @ 0x{:016X}", sizeof(data) * 8, static_cast<u32>(data), vaddr); @@ -730,15 +684,13 @@ struct Memory::Impl { } bool WriteExclusive128(const VAddr vaddr, const u128 data, const u128 expected) { - u8* const page_pointer = current_page_table->pointers[vaddr >> PAGE_BITS]; - if (page_pointer != nullptr) { + const uintptr_t raw_pointer = current_page_table->pointers[vaddr >> PAGE_BITS].Raw(); + if (u8* const pointer = Common::PageTable::PageInfo::ExtractPointer(raw_pointer)) { // NOTE: Avoid adding any extra logic to this fast-path block - auto* pointer = reinterpret_cast<volatile u64*>(&page_pointer[vaddr]); - return Common::AtomicCompareAndSwap(pointer, data, expected); + const auto volatile_pointer = reinterpret_cast<volatile u64*>(&pointer[vaddr]); + return Common::AtomicCompareAndSwap(volatile_pointer, data, expected); } - - const Common::PageType type = current_page_table->attributes[vaddr >> PAGE_BITS]; - switch (type) { + switch (Common::PageTable::PageInfo::ExtractType(raw_pointer)) { case Common::PageType::Unmapped: LOG_ERROR(HW_Memory, "Unmapped Write{} 0x{:08X} @ 0x{:016X}{:016X}", sizeof(data) * 8, static_cast<u64>(data[1]), static_cast<u64>(data[0]), vaddr); @@ -773,25 +725,10 @@ void Memory::MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size impl->MapMemoryRegion(page_table, base, size, target); } -void Memory::MapIoRegion(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer mmio_handler) { - impl->MapIoRegion(page_table, base, size, std::move(mmio_handler)); -} - void Memory::UnmapRegion(Common::PageTable& page_table, VAddr base, u64 size) { impl->UnmapRegion(page_table, base, size); } -void Memory::AddDebugHook(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer hook) { - impl->AddDebugHook(page_table, base, size, std::move(hook)); -} - -void Memory::RemoveDebugHook(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer hook) { - impl->RemoveDebugHook(page_table, base, size, std::move(hook)); -} - bool Memory::IsValidVirtualAddress(const Kernel::Process& process, const VAddr vaddr) const { return impl->IsValidVirtualAddress(process, vaddr); } diff --git a/src/core/memory.h b/src/core/memory.h index 4a1cc63f4..705ebb23d 100644 --- a/src/core/memory.h +++ b/src/core/memory.h @@ -8,7 +8,6 @@ #include <memory> #include <string> #include "common/common_types.h" -#include "common/memory_hook.h" namespace Common { struct PageTable; @@ -78,17 +77,6 @@ public: void MapMemoryRegion(Common::PageTable& page_table, VAddr base, u64 size, PAddr target); /** - * Maps a region of the emulated process address space as a IO region. - * - * @param page_table The page table of the emulated process. - * @param base The address to start mapping at. Must be page-aligned. - * @param size The amount of bytes to map. Must be page-aligned. - * @param mmio_handler The handler that backs the mapping. - */ - void MapIoRegion(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer mmio_handler); - - /** * Unmaps a region of the emulated process address space. * * @param page_table The page table of the emulated process. @@ -98,28 +86,6 @@ public: void UnmapRegion(Common::PageTable& page_table, VAddr base, u64 size); /** - * Adds a memory hook to intercept reads and writes to given region of memory. - * - * @param page_table The page table of the emulated process - * @param base The starting address to apply the hook to. - * @param size The size of the memory region to apply the hook to, in bytes. - * @param hook The hook to apply to the region of memory. - */ - void AddDebugHook(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer hook); - - /** - * Removes a memory hook from a given range of memory. - * - * @param page_table The page table of the emulated process. - * @param base The starting address to remove the hook from. - * @param size The size of the memory region to remove the hook from, in bytes. - * @param hook The hook to remove from the specified region of memory. - */ - void RemoveDebugHook(Common::PageTable& page_table, VAddr base, u64 size, - Common::MemoryHookPointer hook); - - /** * Checks whether or not the supplied address is a valid virtual * address for the given process. * diff --git a/src/core/memory/cheat_engine.cpp b/src/core/memory/cheat_engine.cpp index 29284a42d..2dd0eb0f8 100644 --- a/src/core/memory/cheat_engine.cpp +++ b/src/core/memory/cheat_engine.cpp @@ -153,8 +153,9 @@ std::vector<CheatEntry> TextCheatParser::Parse(std::string_view data) const { return {}; } + const auto value = static_cast<u32>(std::stoul(hex, nullptr, 0x10)); out[*current_entry].definition.opcodes[out[*current_entry].definition.num_opcodes++] = - std::stoul(hex, nullptr, 0x10); + value; i += 8; } else { diff --git a/src/core/network/network.cpp b/src/core/network/network.cpp index 56d173b5e..681e93468 100644 --- a/src/core/network/network.cpp +++ b/src/core/network/network.cpp @@ -11,7 +11,7 @@ #ifdef _WIN32 #define _WINSOCK_DEPRECATED_NO_WARNINGS // gethostname #include <winsock2.h> -#elif __unix__ +#elif YUZU_UNIX #include <errno.h> #include <fcntl.h> #include <netdb.h> @@ -54,7 +54,7 @@ constexpr IPv4Address TranslateIPv4(in_addr addr) { sockaddr TranslateFromSockAddrIn(SockAddrIn input) { sockaddr_in result; -#ifdef __unix__ +#if YUZU_UNIX result.sin_len = sizeof(result); #endif @@ -63,7 +63,7 @@ sockaddr TranslateFromSockAddrIn(SockAddrIn input) { result.sin_family = AF_INET; break; default: - UNIMPLEMENTED_MSG("Unhandled sockaddr family={}", static_cast<int>(input.family)); + UNIMPLEMENTED_MSG("Unhandled sockaddr family={}", input.family); result.sin_family = AF_INET; break; } @@ -99,7 +99,7 @@ bool EnableNonBlock(SOCKET fd, bool enable) { return ioctlsocket(fd, FIONBIO, &value) != SOCKET_ERROR; } -#elif __unix__ // ^ _WIN32 v __unix__ +#elif YUZU_UNIX // ^ _WIN32 v YUZU_UNIX using SOCKET = int; using WSAPOLLFD = pollfd; @@ -133,7 +133,7 @@ sockaddr TranslateFromSockAddrIn(SockAddrIn input) { result.sin_family = AF_INET; break; default: - UNIMPLEMENTED_MSG("Unhandled sockaddr family={}", static_cast<int>(input.family)); + UNIMPLEMENTED_MSG("Unhandled sockaddr family={}", input.family); result.sin_family = AF_INET; break; } @@ -148,7 +148,7 @@ sockaddr TranslateFromSockAddrIn(SockAddrIn input) { } int WSAPoll(WSAPOLLFD* fds, ULONG nfds, int timeout) { - return poll(fds, nfds, timeout); + return poll(fds, static_cast<nfds_t>(nfds), timeout); } int closesocket(SOCKET fd) { @@ -186,7 +186,7 @@ int TranslateDomain(Domain domain) { case Domain::INET: return AF_INET; default: - UNIMPLEMENTED_MSG("Unimplemented domain={}", static_cast<int>(domain)); + UNIMPLEMENTED_MSG("Unimplemented domain={}", domain); return 0; } } @@ -198,7 +198,7 @@ int TranslateType(Type type) { case Type::DGRAM: return SOCK_DGRAM; default: - UNIMPLEMENTED_MSG("Unimplemented type={}", static_cast<int>(type)); + UNIMPLEMENTED_MSG("Unimplemented type={}", type); return 0; } } @@ -210,7 +210,7 @@ int TranslateProtocol(Protocol protocol) { case Protocol::UDP: return IPPROTO_UDP; default: - UNIMPLEMENTED_MSG("Unimplemented protocol={}", static_cast<int>(protocol)); + UNIMPLEMENTED_MSG("Unimplemented protocol={}", protocol); return 0; } } @@ -238,45 +238,45 @@ SockAddrIn TranslateToSockAddrIn(sockaddr input_) { return result; } -u16 TranslatePollEvents(u16 events) { - u16 result = 0; +short TranslatePollEvents(PollEvents events) { + short result = 0; - if (events & POLL_IN) { - events &= ~POLL_IN; + if (True(events & PollEvents::In)) { + events &= ~PollEvents::In; result |= POLLIN; } - if (events & POLL_PRI) { - events &= ~POLL_PRI; + if (True(events & PollEvents::Pri)) { + events &= ~PollEvents::Pri; #ifdef _WIN32 LOG_WARNING(Service, "Winsock doesn't support POLLPRI"); #else - result |= POLL_PRI; + result |= POLLPRI; #endif } - if (events & POLL_OUT) { - events &= ~POLL_OUT; + if (True(events & PollEvents::Out)) { + events &= ~PollEvents::Out; result |= POLLOUT; } - UNIMPLEMENTED_IF_MSG(events != 0, "Unhandled guest events=0x{:x}", events); + UNIMPLEMENTED_IF_MSG((u16)events != 0, "Unhandled guest events=0x{:x}", (u16)events); return result; } -u16 TranslatePollRevents(u16 revents) { - u16 result = 0; - const auto translate = [&result, &revents](int host, unsigned guest) { - if (revents & host) { - revents &= ~host; +PollEvents TranslatePollRevents(short revents) { + PollEvents result{}; + const auto translate = [&result, &revents](short host, PollEvents guest) { + if ((revents & host) != 0) { + revents &= static_cast<short>(~host); result |= guest; } }; - translate(POLLIN, POLL_IN); - translate(POLLPRI, POLL_PRI); - translate(POLLOUT, POLL_OUT); - translate(POLLERR, POLL_ERR); - translate(POLLHUP, POLL_HUP); + translate(POLLIN, PollEvents::In); + translate(POLLPRI, PollEvents::Pri); + translate(POLLOUT, PollEvents::Out); + translate(POLLERR, PollEvents::Err); + translate(POLLHUP, PollEvents::Hup); UNIMPLEMENTED_IF_MSG(revents != 0, "Unhandled host revents=0x{:x}", revents); @@ -408,7 +408,7 @@ std::pair<Socket::AcceptResult, Errno> Socket::Accept() { Errno Socket::Connect(SockAddrIn addr_in) { const sockaddr host_addr_in = TranslateFromSockAddrIn(addr_in); - if (connect(fd, &host_addr_in, sizeof(host_addr_in)) != INVALID_SOCKET) { + if (connect(fd, &host_addr_in, sizeof(host_addr_in)) != SOCKET_ERROR) { return Errno::SUCCESS; } @@ -482,7 +482,7 @@ Errno Socket::Shutdown(ShutdownHow how) { host_how = SD_BOTH; break; default: - UNIMPLEMENTED_MSG("Unimplemented flag how={}", static_cast<int>(how)); + UNIMPLEMENTED_MSG("Unimplemented flag how={}", how); return Errno::SUCCESS; } if (shutdown(fd, host_how) != SOCKET_ERROR) { @@ -503,10 +503,10 @@ std::pair<s32, Errno> Socket::Recv(int flags, std::vector<u8>& message) { ASSERT(flags == 0); ASSERT(message.size() < static_cast<size_t>(std::numeric_limits<int>::max())); - const int result = + const auto result = recv(fd, reinterpret_cast<char*>(message.data()), static_cast<int>(message.size()), 0); if (result != SOCKET_ERROR) { - return {result, Errno::SUCCESS}; + return {static_cast<s32>(result), Errno::SUCCESS}; } switch (const int ec = LastError()) { @@ -531,14 +531,14 @@ std::pair<s32, Errno> Socket::RecvFrom(int flags, std::vector<u8>& message, Sock socklen_t* const p_addrlen = addr ? &addrlen : nullptr; sockaddr* const p_addr_in = addr ? &addr_in : nullptr; - const int result = recvfrom(fd, reinterpret_cast<char*>(message.data()), - static_cast<int>(message.size()), 0, p_addr_in, p_addrlen); + const auto result = recvfrom(fd, reinterpret_cast<char*>(message.data()), + static_cast<int>(message.size()), 0, p_addr_in, p_addrlen); if (result != SOCKET_ERROR) { if (addr) { ASSERT(addrlen == sizeof(addr_in)); *addr = TranslateToSockAddrIn(addr_in); } - return {result, Errno::SUCCESS}; + return {static_cast<s32>(result), Errno::SUCCESS}; } switch (const int ec = LastError()) { @@ -558,10 +558,10 @@ std::pair<s32, Errno> Socket::Send(const std::vector<u8>& message, int flags) { ASSERT(message.size() < static_cast<size_t>(std::numeric_limits<int>::max())); ASSERT(flags == 0); - const int result = send(fd, reinterpret_cast<const char*>(message.data()), - static_cast<int>(message.size()), 0); + const auto result = send(fd, reinterpret_cast<const char*>(message.data()), + static_cast<int>(message.size()), 0); if (result != SOCKET_ERROR) { - return {result, Errno::SUCCESS}; + return {static_cast<s32>(result), Errno::SUCCESS}; } const int ec = LastError(); @@ -591,10 +591,10 @@ std::pair<s32, Errno> Socket::SendTo(u32 flags, const std::vector<u8>& message, to = &host_addr_in; } - const int result = sendto(fd, reinterpret_cast<const char*>(message.data()), - static_cast<int>(message.size()), 0, to, tolen); + const auto result = sendto(fd, reinterpret_cast<const char*>(message.data()), + static_cast<int>(message.size()), 0, to, tolen); if (result != SOCKET_ERROR) { - return {result, Errno::SUCCESS}; + return {static_cast<s32>(result), Errno::SUCCESS}; } const int ec = LastError(); diff --git a/src/core/network/network.h b/src/core/network/network.h index 0622e4593..76b2821f2 100644 --- a/src/core/network/network.h +++ b/src/core/network/network.h @@ -61,19 +61,25 @@ struct SockAddrIn { }; /// Cross-platform poll fd structure + +enum class PollEvents : u16 { + // Using Pascal case because IN is a macro on Windows. + In = 1 << 0, + Pri = 1 << 1, + Out = 1 << 2, + Err = 1 << 3, + Hup = 1 << 4, + Nval = 1 << 5, +}; + +DECLARE_ENUM_FLAG_OPERATORS(PollEvents); + struct PollFD { Socket* socket; - u16 events; - u16 revents; + PollEvents events; + PollEvents revents; }; -constexpr u16 POLL_IN = 1 << 0; -constexpr u16 POLL_PRI = 1 << 1; -constexpr u16 POLL_OUT = 1 << 2; -constexpr u16 POLL_ERR = 1 << 3; -constexpr u16 POLL_HUP = 1 << 4; -constexpr u16 POLL_NVAL = 1 << 5; - class NetworkInstance { public: explicit NetworkInstance(); diff --git a/src/core/network/sockets.h b/src/core/network/sockets.h index 7bdff0fe4..a44393325 100644 --- a/src/core/network/sockets.h +++ b/src/core/network/sockets.h @@ -9,7 +9,7 @@ #if defined(_WIN32) #include <winsock.h> -#elif !defined(__unix__) +#elif !YUZU_UNIX #error "Platform not implemented" #endif @@ -84,7 +84,7 @@ public: #if defined(_WIN32) SOCKET fd = INVALID_SOCKET; -#elif defined(__unix__) +#elif YUZU_UNIX int fd = -1; #endif }; diff --git a/src/core/settings.cpp b/src/core/settings.cpp index 28d3f9099..39306509a 100644 --- a/src/core/settings.cpp +++ b/src/core/settings.cpp @@ -4,9 +4,10 @@ #include <string_view> +#include "common/assert.h" #include "common/file_util.h" +#include "common/logging/log.h" #include "core/core.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/service/hid/hid.h" #include "core/settings.h" #include "video_core/renderer_base.h" @@ -14,7 +15,7 @@ namespace Settings { Values values = {}; -bool configuring_global = true; +static bool configuring_global = true; std::string GetTimeZoneString() { static constexpr std::array timezones{ @@ -31,13 +32,9 @@ std::string GetTimeZoneString() { return timezones[time_zone_index]; } -void Apply() { - GDBStub::SetServerPort(values.gdbstub_port); - GDBStub::ToggleServer(values.use_gdbstub); - - auto& system_instance = Core::System::GetInstance(); - if (system_instance.IsPoweredOn()) { - system_instance.Renderer().RefreshBaseSettings(); +void Apply(Core::System& system) { + if (system.IsPoweredOn()) { + system.Renderer().RefreshBaseSettings(); } Service::HID::ReloadInputDevices(); @@ -49,13 +46,14 @@ void LogSettings() { }; LOG_INFO(Config, "yuzu Configuration:"); - log_setting("Controls_UseDockedMode", values.use_docked_mode); + log_setting("Controls_UseDockedMode", values.use_docked_mode.GetValue()); log_setting("System_RngSeed", values.rng_seed.GetValue().value_or(0)); log_setting("System_CurrentUser", values.current_user); log_setting("System_LanguageIndex", values.language_index.GetValue()); log_setting("System_RegionIndex", values.region_index.GetValue()); log_setting("System_TimeZoneIndex", values.time_zone_index.GetValue()); log_setting("Core_UseMultiCore", values.use_multi_core.GetValue()); + log_setting("CPU_Accuracy", values.cpu_accuracy); log_setting("Renderer_UseResolutionFactor", values.resolution_factor.GetValue()); log_setting("Renderer_UseFrameLimit", values.use_frame_limit.GetValue()); log_setting("Renderer_FrameLimit", values.frame_limit.GetValue()); @@ -63,6 +61,7 @@ void LogSettings() { log_setting("Renderer_GPUAccuracyLevel", values.gpu_accuracy.GetValue()); log_setting("Renderer_UseAsynchronousGpuEmulation", values.use_asynchronous_gpu_emulation.GetValue()); + log_setting("Renderer_UseNvdecEmulation", values.use_nvdec_emulation.GetValue()); log_setting("Renderer_UseVsync", values.use_vsync.GetValue()); log_setting("Renderer_UseAssemblyShaders", values.use_assembly_shaders.GetValue()); log_setting("Renderer_UseAsynchronousShaders", values.use_asynchronous_shaders.GetValue()); @@ -73,18 +72,17 @@ void LogSettings() { log_setting("DataStorage_UseVirtualSd", values.use_virtual_sd); log_setting("DataStorage_NandDir", Common::FS::GetUserPath(Common::FS::UserPath::NANDDir)); log_setting("DataStorage_SdmcDir", Common::FS::GetUserPath(Common::FS::UserPath::SDMCDir)); - log_setting("Debugging_UseGdbstub", values.use_gdbstub); - log_setting("Debugging_GdbstubPort", values.gdbstub_port); log_setting("Debugging_ProgramArgs", values.program_args); log_setting("Services_BCATBackend", values.bcat_backend); log_setting("Services_BCATBoxcatLocal", values.bcat_boxcat_local); } -float Volume() { - if (values.audio_muted) { - return 0.0f; - } - return values.volume.GetValue(); +bool IsConfiguringGlobal() { + return configuring_global; +} + +void SetConfiguringGlobal(bool is_global) { + configuring_global = is_global; } bool IsGPULevelExtreme() { @@ -96,9 +94,16 @@ bool IsGPULevelHigh() { values.gpu_accuracy.GetValue() == GPUAccuracy::High; } -void RestoreGlobalState() { +float Volume() { + if (values.audio_muted) { + return 0.0f; + } + return values.volume.GetValue(); +} + +void RestoreGlobalState(bool is_powered_on) { // If a game is running, DO NOT restore the global settings state - if (Core::System::GetInstance().IsPoweredOn()) { + if (is_powered_on) { return; } @@ -119,6 +124,7 @@ void RestoreGlobalState() { values.use_disk_shader_cache.SetGlobal(true); values.gpu_accuracy.SetGlobal(true); values.use_asynchronous_gpu_emulation.SetGlobal(true); + values.use_nvdec_emulation.SetGlobal(true); values.use_vsync.SetGlobal(true); values.use_assembly_shaders.SetGlobal(true); values.use_asynchronous_shaders.SetGlobal(true); @@ -134,11 +140,12 @@ void RestoreGlobalState() { values.rng_seed.SetGlobal(true); values.custom_rtc.SetGlobal(true); values.sound_index.SetGlobal(true); -} -void Sanitize() { - values.use_asynchronous_gpu_emulation.SetValue( - values.use_asynchronous_gpu_emulation.GetValue() || values.use_multi_core.GetValue()); + // Controls + values.players.SetGlobal(true); + values.use_docked_mode.SetGlobal(true); + values.vibration_enabled.SetGlobal(true); + values.motion_enabled.SetGlobal(true); } } // namespace Settings diff --git a/src/core/settings.h b/src/core/settings.h index 9834f44bb..a324530bd 100644 --- a/src/core/settings.h +++ b/src/core/settings.h @@ -14,6 +14,10 @@ #include "common/common_types.h" #include "input_common/settings.h" +namespace Core { +class System; +} + namespace Settings { enum class RendererBackend { @@ -33,8 +37,6 @@ enum class CPUAccuracy { DebugMode = 2, }; -extern bool configuring_global; - template <typename Type> class Setting final { public: @@ -67,6 +69,38 @@ private: Type local{}; }; +/** + * The InputSetting class allows for getting a reference to either the global or local members. + * This is required as we cannot easily modify the values of user-defined types within containers + * using the SetValue() member function found in the Setting class. The primary purpose of this + * class is to store an array of 10 PlayerInput structs for both the global and local (per-game) + * setting and allows for easily accessing and modifying both settings. + */ +template <typename Type> +class InputSetting final { +public: + InputSetting() = default; + explicit InputSetting(Type val) : global{val} {} + ~InputSetting() = default; + void SetGlobal(bool to_global) { + use_global = to_global; + } + bool UsingGlobal() const { + return use_global; + } + Type& GetValue(bool need_global = false) { + if (use_global || need_global) { + return global; + } + return local; + } + +private: + bool use_global = true; + Type global{}; + Type local{}; +}; + struct TouchFromButtonMap { std::string name; std::vector<std::string> buttons; @@ -97,13 +131,14 @@ struct Values { bool cpuopt_unsafe_unfuse_fma; bool cpuopt_unsafe_reduce_fp_error; + bool cpuopt_unsafe_inaccurate_nan; // Renderer Setting<RendererBackend> renderer_backend; bool renderer_debug; Setting<int> vulkan_device; - Setting<u16> resolution_factor = Setting(static_cast<u16>(1)); + Setting<u16> resolution_factor{1}; Setting<int> aspect_ratio; Setting<int> max_anisotropy; Setting<bool> use_frame_limit; @@ -111,6 +146,7 @@ struct Values { Setting<bool> use_disk_shader_cache; Setting<GPUAccuracy> gpu_accuracy; Setting<bool> use_asynchronous_gpu_emulation; + Setting<bool> use_nvdec_emulation; Setting<bool> use_vsync; Setting<bool> use_assembly_shaders; Setting<bool> use_asynchronous_shaders; @@ -134,9 +170,18 @@ struct Values { Setting<s32> sound_index; // Controls - std::array<PlayerInput, 10> players; + InputSetting<std::array<PlayerInput, 10>> players; + + Setting<bool> use_docked_mode; + + Setting<bool> vibration_enabled; + Setting<bool> enable_accurate_vibrations; + + Setting<bool> motion_enabled; + std::string motion_device; + std::string udp_input_servers; - bool use_docked_mode; + bool emulate_analog_keyboard; bool mouse_enabled; std::string mouse_device; @@ -150,20 +195,15 @@ struct Values { ButtonsRaw debug_pad_buttons; AnalogsRaw debug_pad_analogs; - bool vibration_enabled; - - bool motion_enabled; - std::string motion_device; - std::string touch_device; TouchscreenInput touchscreen; - std::atomic_bool is_device_reload_pending{true}; + bool use_touch_from_button; + std::string touch_device; int touch_from_button_map_index; - std::string udp_input_address; - u16 udp_input_port; - u8 udp_pad_index; std::vector<TouchFromButtonMap> touch_from_button_maps; + std::atomic_bool is_device_reload_pending{true}; + // Data Storage bool use_virtual_sd; bool gamecard_inserted; @@ -180,8 +220,9 @@ struct Values { bool reporting_services; bool quest_flag; bool disable_macro_jit; + bool extended_logging; - // Misceallaneous + // Miscellaneous std::string log_filter; bool use_dev_keys; @@ -197,22 +238,24 @@ struct Values { // Add-Ons std::map<u64, std::vector<std::string>> disabled_addons; -} extern values; +}; -float Volume(); +extern Values values; + +bool IsConfiguringGlobal(); +void SetConfiguringGlobal(bool is_global); bool IsGPULevelExtreme(); bool IsGPULevelHigh(); +float Volume(); + std::string GetTimeZoneString(); -void Apply(); +void Apply(Core::System& system); void LogSettings(); // Restore the global state of all applicable settings in the Values struct -void RestoreGlobalState(); - -// Fixes settings that are known to cause issues with the emulator -void Sanitize(); +void RestoreGlobalState(bool is_powered_on); } // namespace Settings diff --git a/src/core/telemetry_session.cpp b/src/core/telemetry_session.cpp index da09c0dbc..d11b15f38 100644 --- a/src/core/telemetry_session.cpp +++ b/src/core/telemetry_session.cpp @@ -147,7 +147,9 @@ TelemetrySession::~TelemetrySession() { } } -void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) { +void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider) { // Log one-time top-level information AddField(Telemetry::FieldType::None, "TelemetryId", GetTelemetryId()); @@ -167,7 +169,10 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) { app_loader.ReadTitle(name); if (name.empty()) { - const auto metadata = FileSys::PatchManager(program_id).GetControlMetadata(); + const auto metadata = [&content_provider, &fsc, program_id] { + const FileSys::PatchManager pm{program_id, fsc, content_provider}; + return pm.GetControlMetadata(); + }(); if (metadata.first != nullptr) { name = metadata.first->GetApplicationName(); } @@ -206,12 +211,14 @@ void TelemetrySession::AddInitialInfo(Loader::AppLoader& app_loader) { TranslateGPUAccuracyLevel(Settings::values.gpu_accuracy.GetValue())); AddField(field_type, "Renderer_UseAsynchronousGpuEmulation", Settings::values.use_asynchronous_gpu_emulation.GetValue()); + AddField(field_type, "Renderer_UseNvdecEmulation", + Settings::values.use_nvdec_emulation.GetValue()); AddField(field_type, "Renderer_UseVsync", Settings::values.use_vsync.GetValue()); AddField(field_type, "Renderer_UseAssemblyShaders", Settings::values.use_assembly_shaders.GetValue()); AddField(field_type, "Renderer_UseAsynchronousShaders", Settings::values.use_asynchronous_shaders.GetValue()); - AddField(field_type, "System_UseDockedMode", Settings::values.use_docked_mode); + AddField(field_type, "System_UseDockedMode", Settings::values.use_docked_mode.GetValue()); } bool TelemetrySession::SubmitTestcase() { diff --git a/src/core/telemetry_session.h b/src/core/telemetry_session.h index 66789d4bd..6f3d45bea 100644 --- a/src/core/telemetry_session.h +++ b/src/core/telemetry_session.h @@ -7,10 +7,18 @@ #include <string> #include "common/telemetry.h" +namespace FileSys { +class ContentProvider; +} + namespace Loader { class AppLoader; } +namespace Service::FileSystem { +class FileSystemController; +} + namespace Core { /** @@ -40,10 +48,14 @@ public: * - Title file format * - Miscellaneous settings values. * - * @param app_loader The application loader to use to retrieve - * title-specific information. + * @param app_loader The application loader to use to retrieve + * title-specific information. + * @param fsc Filesystem controller to use to retrieve info. + * @param content_provider Content provider to use to retrieve info. */ - void AddInitialInfo(Loader::AppLoader& app_loader); + void AddInitialInfo(Loader::AppLoader& app_loader, + const Service::FileSystem::FileSystemController& fsc, + const FileSys::ContentProvider& content_provider); /** * Wrapper around the Telemetry::FieldCollection::AddField method. diff --git a/src/input_common/CMakeLists.txt b/src/input_common/CMakeLists.txt index 09361e37e..38ab31898 100644 --- a/src/input_common/CMakeLists.txt +++ b/src/input_common/CMakeLists.txt @@ -5,8 +5,8 @@ add_library(input_common STATIC keyboard.h main.cpp main.h - motion_emu.cpp - motion_emu.h + motion_from_button.cpp + motion_from_button.h motion_input.cpp motion_input.h settings.cpp @@ -17,6 +17,10 @@ add_library(input_common STATIC gcadapter/gc_adapter.h gcadapter/gc_poller.cpp gcadapter/gc_poller.h + mouse/mouse_input.cpp + mouse/mouse_input.h + mouse/mouse_poller.cpp + mouse/mouse_poller.h sdl/sdl.cpp sdl/sdl.h udp/client.cpp @@ -27,6 +31,39 @@ add_library(input_common STATIC udp/udp.h ) +if (MSVC) + target_compile_options(input_common PRIVATE + /W4 + /WX + + # 'expression' : signed/unsigned mismatch + /we4018 + # 'argument' : conversion from 'type1' to 'type2', possible loss of data (floating-point) + /we4244 + # 'conversion' : conversion from 'type1' to 'type2', signed/unsigned mismatch + /we4245 + # 'operator': conversion from 'type1:field_bits' to 'type2:field_bits', possible loss of data + /we4254 + # 'var' : conversion from 'size_t' to 'type', possible loss of data + /we4267 + # 'context' : truncation from 'type1' to 'type2' + /we4305 + ) +else() + target_compile_options(input_common PRIVATE + -Werror + -Werror=conversion + -Werror=ignored-qualifiers + -Werror=implicit-fallthrough + -Werror=reorder + -Werror=shadow + -Werror=sign-compare + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable> + -Werror=unused-variable + ) +endif() + if(SDL2_FOUND) target_sources(input_common PRIVATE sdl/sdl_impl.cpp diff --git a/src/input_common/analog_from_button.cpp b/src/input_common/analog_from_button.cpp index 6cabdaa3c..40b516f85 100755 --- a/src/input_common/analog_from_button.cpp +++ b/src/input_common/analog_from_button.cpp @@ -2,6 +2,11 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <chrono> +#include <cmath> +#include <thread> +#include "common/math_util.h" +#include "core/settings.h" #include "input_common/analog_from_button.h" namespace InputCommon { @@ -11,27 +16,123 @@ public: using Button = std::unique_ptr<Input::ButtonDevice>; Analog(Button up_, Button down_, Button left_, Button right_, Button modifier_, - float modifier_scale_) + float modifier_scale_, float modifier_angle_) : up(std::move(up_)), down(std::move(down_)), left(std::move(left_)), - right(std::move(right_)), modifier(std::move(modifier_)), - modifier_scale(modifier_scale_) {} + right(std::move(right_)), modifier(std::move(modifier_)), modifier_scale(modifier_scale_), + modifier_angle(modifier_angle_) { + update_thread = std::thread(&Analog::UpdateStatus, this); + } + + ~Analog() override { + update_thread_running = false; + if (update_thread.joinable()) { + update_thread.join(); + } + } + + void MoveToDirection(bool enable, float to_angle) { + if (!enable) { + return; + } + constexpr float TAU = Common::PI * 2.0f; + // Use wider angle to ease the transition. + constexpr float aperture = TAU * 0.15f; + const float top_limit = to_angle + aperture; + const float bottom_limit = to_angle - aperture; + + if ((angle > to_angle && angle <= top_limit) || + (angle + TAU > to_angle && angle + TAU <= top_limit)) { + angle -= modifier_angle; + if (angle < 0) { + angle += TAU; + } + } else if ((angle >= bottom_limit && angle < to_angle) || + (angle - TAU >= bottom_limit && angle - TAU < to_angle)) { + angle += modifier_angle; + if (angle >= TAU) { + angle -= TAU; + } + } else { + angle = to_angle; + } + } + + void UpdateStatus() { + while (update_thread_running) { + const float coef = modifier->GetStatus() ? modifier_scale : 1.0f; + + bool r = right->GetStatus(); + bool l = left->GetStatus(); + bool u = up->GetStatus(); + bool d = down->GetStatus(); + + // Eliminate contradictory movements + if (r && l) { + r = false; + l = false; + } + if (u && d) { + u = false; + d = false; + } + + // Move to the right + MoveToDirection(r && !u && !d, 0.0f); + + // Move to the upper right + MoveToDirection(r && u && !d, Common::PI * 0.25f); + + // Move up + MoveToDirection(u && !l && !r, Common::PI * 0.5f); + + // Move to the upper left + MoveToDirection(l && u && !d, Common::PI * 0.75f); + + // Move to the left + MoveToDirection(l && !u && !d, Common::PI); + + // Move to the bottom left + MoveToDirection(l && !u && d, Common::PI * 1.25f); + + // Move down + MoveToDirection(d && !l && !r, Common::PI * 1.5f); + + // Move to the bottom right + MoveToDirection(r && !u && d, Common::PI * 1.75f); + + // Move if a key is pressed + if (r || l || u || d) { + amplitude = coef; + } else { + amplitude = 0; + } + + // Delay the update rate to 100hz + std::this_thread::sleep_for(std::chrono::milliseconds(10)); + } + } std::tuple<float, float> GetStatus() const override { + if (Settings::values.emulate_analog_keyboard) { + return std::make_tuple(std::cos(angle) * amplitude, std::sin(angle) * amplitude); + } constexpr float SQRT_HALF = 0.707106781f; int x = 0, y = 0; - - if (right->GetStatus()) + if (right->GetStatus()) { ++x; - if (left->GetStatus()) + } + if (left->GetStatus()) { --x; - if (up->GetStatus()) + } + if (up->GetStatus()) { ++y; - if (down->GetStatus()) + } + if (down->GetStatus()) { --y; - - float coef = modifier->GetStatus() ? modifier_scale : 1.0f; - return std::make_tuple(x * coef * (y == 0 ? 1.0f : SQRT_HALF), - y * coef * (x == 0 ? 1.0f : SQRT_HALF)); + } + const float coef = modifier->GetStatus() ? modifier_scale : 1.0f; + return std::make_tuple(static_cast<float>(x) * coef * (y == 0 ? 1.0f : SQRT_HALF), + static_cast<float>(y) * coef * (x == 0 ? 1.0f : SQRT_HALF)); } bool GetAnalogDirectionStatus(Input::AnalogDirection direction) const override { @@ -55,6 +156,11 @@ private: Button right; Button modifier; float modifier_scale; + float modifier_angle; + float angle{}; + float amplitude{}; + std::thread update_thread; + bool update_thread_running{true}; }; std::unique_ptr<Input::AnalogDevice> AnalogFromButton::Create(const Common::ParamPackage& params) { @@ -65,8 +171,10 @@ std::unique_ptr<Input::AnalogDevice> AnalogFromButton::Create(const Common::Para auto right = Input::CreateDevice<Input::ButtonDevice>(params.Get("right", null_engine)); auto modifier = Input::CreateDevice<Input::ButtonDevice>(params.Get("modifier", null_engine)); auto modifier_scale = params.Get("modifier_scale", 0.5f); + auto modifier_angle = params.Get("modifier_angle", 0.035f); return std::make_unique<Analog>(std::move(up), std::move(down), std::move(left), - std::move(right), std::move(modifier), modifier_scale); + std::move(right), std::move(modifier), modifier_scale, + modifier_angle); } } // namespace InputCommon diff --git a/src/input_common/gcadapter/gc_adapter.cpp b/src/input_common/gcadapter/gc_adapter.cpp index 89c148aba..d80195c82 100644 --- a/src/input_common/gcadapter/gc_adapter.cpp +++ b/src/input_common/gcadapter/gc_adapter.cpp @@ -21,14 +21,6 @@ namespace GCAdapter { -/// Used to loop through and assign button in poller -constexpr std::array<PadButton, 12> PadButtonArray{ - PadButton::PAD_BUTTON_LEFT, PadButton::PAD_BUTTON_RIGHT, PadButton::PAD_BUTTON_DOWN, - PadButton::PAD_BUTTON_UP, PadButton::PAD_TRIGGER_Z, PadButton::PAD_TRIGGER_R, - PadButton::PAD_TRIGGER_L, PadButton::PAD_BUTTON_A, PadButton::PAD_BUTTON_B, - PadButton::PAD_BUTTON_X, PadButton::PAD_BUTTON_Y, PadButton::PAD_BUTTON_START, -}; - Adapter::Adapter() { if (usb_adapter_handle != nullptr) { return; @@ -37,179 +29,261 @@ Adapter::Adapter() { const int init_res = libusb_init(&libusb_ctx); if (init_res == LIBUSB_SUCCESS) { - Setup(); + adapter_scan_thread = std::thread(&Adapter::AdapterScanThread, this); } else { LOG_ERROR(Input, "libusb could not be initialized. failed with error = {}", init_res); } } -GCPadStatus Adapter::GetPadStatus(std::size_t port, const std::array<u8, 37>& adapter_payload) { - GCPadStatus pad = {}; - const std::size_t offset = 1 + (9 * port); +Adapter::~Adapter() { + Reset(); +} - adapter_controllers_status[port] = static_cast<ControllerTypes>(adapter_payload[offset] >> 4); +void Adapter::AdapterInputThread() { + LOG_DEBUG(Input, "GC Adapter input thread started"); + s32 payload_size{}; + AdapterPayload adapter_payload{}; - static constexpr std::array<PadButton, 8> b1_buttons{ - PadButton::PAD_BUTTON_A, PadButton::PAD_BUTTON_B, PadButton::PAD_BUTTON_X, - PadButton::PAD_BUTTON_Y, PadButton::PAD_BUTTON_LEFT, PadButton::PAD_BUTTON_RIGHT, - PadButton::PAD_BUTTON_DOWN, PadButton::PAD_BUTTON_UP, - }; + if (adapter_scan_thread.joinable()) { + adapter_scan_thread.join(); + } - static constexpr std::array<PadButton, 4> b2_buttons{ - PadButton::PAD_BUTTON_START, - PadButton::PAD_TRIGGER_Z, - PadButton::PAD_TRIGGER_R, - PadButton::PAD_TRIGGER_L, - }; + while (adapter_input_thread_running) { + libusb_interrupt_transfer(usb_adapter_handle, input_endpoint, adapter_payload.data(), + static_cast<s32>(adapter_payload.size()), &payload_size, 16); + if (IsPayloadCorrect(adapter_payload, payload_size)) { + UpdateControllers(adapter_payload); + UpdateVibrations(); + } + std::this_thread::yield(); + } - static constexpr std::array<PadAxes, 6> axes{ - PadAxes::StickX, PadAxes::StickY, PadAxes::SubstickX, - PadAxes::SubstickY, PadAxes::TriggerLeft, PadAxes::TriggerRight, - }; + if (restart_scan_thread) { + adapter_scan_thread = std::thread(&Adapter::AdapterScanThread, this); + restart_scan_thread = false; + } +} - if (adapter_controllers_status[port] == ControllerTypes::None && !get_origin[port]) { - // Controller may have been disconnected, recalibrate if reconnected. - get_origin[port] = true; +bool Adapter::IsPayloadCorrect(const AdapterPayload& adapter_payload, s32 payload_size) { + if (payload_size != static_cast<s32>(adapter_payload.size()) || + adapter_payload[0] != LIBUSB_DT_HID) { + LOG_DEBUG(Input, "Error reading payload (size: {}, type: {:02x})", payload_size, + adapter_payload[0]); + if (input_error_counter++ > 20) { + LOG_ERROR(Input, "GC adapter timeout, Is the adapter connected?"); + adapter_input_thread_running = false; + restart_scan_thread = true; + } + return false; } - if (adapter_controllers_status[port] != ControllerTypes::None) { - const u8 b1 = adapter_payload[offset + 1]; - const u8 b2 = adapter_payload[offset + 2]; + input_error_counter = 0; + return true; +} - for (std::size_t i = 0; i < b1_buttons.size(); ++i) { - if ((b1 & (1U << i)) != 0) { - pad.button |= static_cast<u16>(b1_buttons[i]); +void Adapter::UpdateControllers(const AdapterPayload& adapter_payload) { + for (std::size_t port = 0; port < pads.size(); ++port) { + const std::size_t offset = 1 + (9 * port); + const auto type = static_cast<ControllerTypes>(adapter_payload[offset] >> 4); + UpdatePadType(port, type); + if (DeviceConnected(port)) { + const u8 b1 = adapter_payload[offset + 1]; + const u8 b2 = adapter_payload[offset + 2]; + UpdateStateButtons(port, b1, b2); + UpdateStateAxes(port, adapter_payload); + if (configuring) { + UpdateYuzuSettings(port); } } + } +} - for (std::size_t j = 0; j < b2_buttons.size(); ++j) { - if ((b2 & (1U << j)) != 0) { - pad.button |= static_cast<u16>(b2_buttons[j]); - } - } - for (PadAxes axis : axes) { - const std::size_t index = static_cast<std::size_t>(axis); - pad.axis_values[index] = adapter_payload[offset + 3 + index]; +void Adapter::UpdatePadType(std::size_t port, ControllerTypes pad_type) { + if (pads[port].type == pad_type) { + return; + } + // Device changed reset device and set new type + ResetDevice(port); + pads[port].type = pad_type; +} + +void Adapter::UpdateStateButtons(std::size_t port, u8 b1, u8 b2) { + if (port >= pads.size()) { + return; + } + + static constexpr std::array<PadButton, 8> b1_buttons{ + PadButton::ButtonA, PadButton::ButtonB, PadButton::ButtonX, PadButton::ButtonY, + PadButton::ButtonLeft, PadButton::ButtonRight, PadButton::ButtonDown, PadButton::ButtonUp, + }; + + static constexpr std::array<PadButton, 4> b2_buttons{ + PadButton::ButtonStart, + PadButton::TriggerZ, + PadButton::TriggerR, + PadButton::TriggerL, + }; + pads[port].buttons = 0; + for (std::size_t i = 0; i < b1_buttons.size(); ++i) { + if ((b1 & (1U << i)) != 0) { + pads[port].buttons = + static_cast<u16>(pads[port].buttons | static_cast<u16>(b1_buttons[i])); + pads[port].last_button = b1_buttons[i]; } + } - if (get_origin[port]) { - origin_status[port].axis_values = pad.axis_values; - get_origin[port] = false; + for (std::size_t j = 0; j < b2_buttons.size(); ++j) { + if ((b2 & (1U << j)) != 0) { + pads[port].buttons = + static_cast<u16>(pads[port].buttons | static_cast<u16>(b2_buttons[j])); + pads[port].last_button = b2_buttons[j]; } } - return pad; } -void Adapter::PadToState(const GCPadStatus& pad, GCState& state) { - for (const auto& button : PadButtonArray) { - const u16 button_value = static_cast<u16>(button); - state.buttons.insert_or_assign(button_value, pad.button & button_value); +void Adapter::UpdateStateAxes(std::size_t port, const AdapterPayload& adapter_payload) { + if (port >= pads.size()) { + return; } - for (size_t i = 0; i < pad.axis_values.size(); ++i) { - state.axes.insert_or_assign(static_cast<u8>(i), pad.axis_values[i]); + const std::size_t offset = 1 + (9 * port); + static constexpr std::array<PadAxes, 6> axes{ + PadAxes::StickX, PadAxes::StickY, PadAxes::SubstickX, + PadAxes::SubstickY, PadAxes::TriggerLeft, PadAxes::TriggerRight, + }; + + for (const PadAxes axis : axes) { + const auto index = static_cast<std::size_t>(axis); + const u8 axis_value = adapter_payload[offset + 3 + index]; + if (pads[port].axis_origin[index] == 255) { + pads[port].axis_origin[index] = axis_value; + } + pads[port].axis_values[index] = + static_cast<s16>(axis_value - pads[port].axis_origin[index]); } } -void Adapter::Read() { - LOG_DEBUG(Input, "GC Adapter Read() thread started"); +void Adapter::UpdateYuzuSettings(std::size_t port) { + if (port >= pads.size()) { + return; + } - int payload_size; - std::array<u8, 37> adapter_payload; - std::array<GCPadStatus, 4> pads; + constexpr u8 axis_threshold = 50; + GCPadStatus pad_status = {.port = port}; - while (adapter_thread_running) { - libusb_interrupt_transfer(usb_adapter_handle, input_endpoint, adapter_payload.data(), - sizeof(adapter_payload), &payload_size, 16); - - if (payload_size != sizeof(adapter_payload) || adapter_payload[0] != LIBUSB_DT_HID) { - LOG_ERROR(Input, - "Error reading payload (size: {}, type: {:02x}) Is the adapter connected?", - payload_size, adapter_payload[0]); - adapter_thread_running = false; // error reading from adapter, stop reading. - break; - } - for (std::size_t port = 0; port < pads.size(); ++port) { - pads[port] = GetPadStatus(port, adapter_payload); - if (DeviceConnected(port) && configuring) { - if (pads[port].button != 0) { - pad_queue[port].Push(pads[port]); - } + if (pads[port].buttons != 0) { + pad_status.button = pads[port].last_button; + pad_queue.Push(pad_status); + } - // Accounting for a threshold here to ensure an intentional press - for (size_t i = 0; i < pads[port].axis_values.size(); ++i) { - const u8 value = pads[port].axis_values[i]; - const u8 origin = origin_status[port].axis_values[i]; - - if (value > origin + pads[port].THRESHOLD || - value < origin - pads[port].THRESHOLD) { - pads[port].axis = static_cast<PadAxes>(i); - pads[port].axis_value = pads[port].axis_values[i]; - pad_queue[port].Push(pads[port]); - } - } - } - PadToState(pads[port], state[port]); + // Accounting for a threshold here to ensure an intentional press + for (std::size_t i = 0; i < pads[port].axis_values.size(); ++i) { + const s16 value = pads[port].axis_values[i]; + + if (value > axis_threshold || value < -axis_threshold) { + pad_status.axis = static_cast<PadAxes>(i); + pad_status.axis_value = value; + pad_status.axis_threshold = axis_threshold; + pad_queue.Push(pad_status); } - std::this_thread::yield(); } } -void Adapter::Setup() { - // Initialize all controllers as unplugged - adapter_controllers_status.fill(ControllerTypes::None); - // Initialize all ports to store axis origin values - get_origin.fill(true); - - // pointer to list of connected usb devices - libusb_device** devices{}; - - // populate the list of devices, get the count - const ssize_t device_count = libusb_get_device_list(libusb_ctx, &devices); - if (device_count < 0) { - LOG_ERROR(Input, "libusb_get_device_list failed with error: {}", device_count); - return; +void Adapter::UpdateVibrations() { + // Use 8 states to keep the switching between on/off fast enough for + // a human to not notice the difference between switching from on/off + // More states = more rumble strengths = slower update time + constexpr u8 vibration_states = 8; + + vibration_counter = (vibration_counter + 1) % vibration_states; + + for (GCController& pad : pads) { + const bool vibrate = pad.rumble_amplitude > vibration_counter; + vibration_changed |= vibrate != pad.enable_vibration; + pad.enable_vibration = vibrate; } + SendVibrations(); +} - if (devices != nullptr) { - for (std::size_t index = 0; index < static_cast<std::size_t>(device_count); ++index) { - if (CheckDeviceAccess(devices[index])) { - // GC Adapter found and accessible, registering it - GetGCEndpoint(devices[index]); - break; - } +void Adapter::SendVibrations() { + if (!rumble_enabled || !vibration_changed) { + return; + } + s32 size{}; + constexpr u8 rumble_command = 0x11; + const u8 p1 = pads[0].enable_vibration; + const u8 p2 = pads[1].enable_vibration; + const u8 p3 = pads[2].enable_vibration; + const u8 p4 = pads[3].enable_vibration; + std::array<u8, 5> payload = {rumble_command, p1, p2, p3, p4}; + const int err = libusb_interrupt_transfer(usb_adapter_handle, output_endpoint, payload.data(), + static_cast<s32>(payload.size()), &size, 16); + if (err) { + LOG_DEBUG(Input, "Adapter libusb write failed: {}", libusb_error_name(err)); + if (output_error_counter++ > 5) { + LOG_ERROR(Input, "GC adapter output timeout, Rumble disabled"); + rumble_enabled = false; } - libusb_free_device_list(devices, 1); + return; } + output_error_counter = 0; + vibration_changed = false; } -bool Adapter::CheckDeviceAccess(libusb_device* device) { - libusb_device_descriptor desc; - const int get_descriptor_error = libusb_get_device_descriptor(device, &desc); - if (get_descriptor_error) { - // could not acquire the descriptor, no point in trying to use it. - LOG_ERROR(Input, "libusb_get_device_descriptor failed with error: {}", - get_descriptor_error); - return false; +bool Adapter::RumblePlay(std::size_t port, u8 amplitude) { + pads[port].rumble_amplitude = amplitude; + + return rumble_enabled; +} + +void Adapter::AdapterScanThread() { + adapter_scan_thread_running = true; + adapter_input_thread_running = false; + if (adapter_input_thread.joinable()) { + adapter_input_thread.join(); } + ClearLibusbHandle(); + ResetDevices(); + while (adapter_scan_thread_running && !adapter_input_thread_running) { + Setup(); + std::this_thread::sleep_for(std::chrono::seconds(1)); + } +} - if (desc.idVendor != 0x057e || desc.idProduct != 0x0337) { - // This isn't the device we are looking for. - return false; +void Adapter::Setup() { + usb_adapter_handle = libusb_open_device_with_vid_pid(libusb_ctx, 0x057e, 0x0337); + + if (usb_adapter_handle == NULL) { + return; + } + if (!CheckDeviceAccess()) { + ClearLibusbHandle(); + return; } - const int open_error = libusb_open(device, &usb_adapter_handle); - if (open_error == LIBUSB_ERROR_ACCESS) { - LOG_ERROR(Input, "Yuzu can not gain access to this device: ID {:04X}:{:04X}.", - desc.idVendor, desc.idProduct); - return false; + libusb_device* device = libusb_get_device(usb_adapter_handle); + + LOG_INFO(Input, "GC adapter is now connected"); + // GC Adapter found and accessible, registering it + if (GetGCEndpoint(device)) { + adapter_scan_thread_running = false; + adapter_input_thread_running = true; + rumble_enabled = true; + input_error_counter = 0; + output_error_counter = 0; + adapter_input_thread = std::thread(&Adapter::AdapterInputThread, this); } - if (open_error) { - LOG_ERROR(Input, "libusb_open failed to open device with error = {}", open_error); - return false; +} + +bool Adapter::CheckDeviceAccess() { + // This fixes payload problems from offbrand GCAdapters + const s32 control_transfer_error = + libusb_control_transfer(usb_adapter_handle, 0x21, 11, 0x0001, 0, nullptr, 0, 1000); + if (control_transfer_error < 0) { + LOG_ERROR(Input, "libusb_control_transfer failed with error= {}", control_transfer_error); } - int kernel_driver_error = libusb_kernel_driver_active(usb_adapter_handle, 0); + s32 kernel_driver_error = libusb_kernel_driver_active(usb_adapter_handle, 0); if (kernel_driver_error == 1) { kernel_driver_error = libusb_detach_kernel_driver(usb_adapter_handle, 0); if (kernel_driver_error != 0 && kernel_driver_error != LIBUSB_ERROR_NOT_SUPPORTED) { @@ -235,13 +309,13 @@ bool Adapter::CheckDeviceAccess(libusb_device* device) { return true; } -void Adapter::GetGCEndpoint(libusb_device* device) { +bool Adapter::GetGCEndpoint(libusb_device* device) { libusb_config_descriptor* config = nullptr; const int config_descriptor_return = libusb_get_config_descriptor(device, 0, &config); if (config_descriptor_return != LIBUSB_SUCCESS) { LOG_ERROR(Input, "libusb_get_config_descriptor failed with error = {}", config_descriptor_return); - return; + return false; } for (u8 ic = 0; ic < config->bNumInterfaces; ic++) { @@ -250,7 +324,7 @@ void Adapter::GetGCEndpoint(libusb_device* device) { const libusb_interface_descriptor* interface = &interfaceContainer->altsetting[i]; for (u8 e = 0; e < interface->bNumEndpoints; e++) { const libusb_endpoint_descriptor* endpoint = &interface->endpoint[e]; - if (endpoint->bEndpointAddress & LIBUSB_ENDPOINT_IN) { + if ((endpoint->bEndpointAddress & LIBUSB_ENDPOINT_IN) != 0) { input_endpoint = endpoint->bEndpointAddress; } else { output_endpoint = endpoint->bEndpointAddress; @@ -263,31 +337,51 @@ void Adapter::GetGCEndpoint(libusb_device* device) { unsigned char clear_payload = 0x13; libusb_interrupt_transfer(usb_adapter_handle, output_endpoint, &clear_payload, sizeof(clear_payload), nullptr, 16); - - adapter_thread_running = true; - adapter_input_thread = std::thread(&Adapter::Read, this); + return true; } -Adapter::~Adapter() { - Reset(); -} +void Adapter::JoinThreads() { + restart_scan_thread = false; + adapter_input_thread_running = false; + adapter_scan_thread_running = false; -void Adapter::Reset() { - if (adapter_thread_running) { - adapter_thread_running = false; + if (adapter_scan_thread.joinable()) { + adapter_scan_thread.join(); } + if (adapter_input_thread.joinable()) { adapter_input_thread.join(); } +} - adapter_controllers_status.fill(ControllerTypes::None); - get_origin.fill(true); - +void Adapter::ClearLibusbHandle() { if (usb_adapter_handle) { libusb_release_interface(usb_adapter_handle, 1); libusb_close(usb_adapter_handle); usb_adapter_handle = nullptr; } +} + +void Adapter::ResetDevices() { + for (std::size_t i = 0; i < pads.size(); ++i) { + ResetDevice(i); + } +} + +void Adapter::ResetDevice(std::size_t port) { + pads[port].type = ControllerTypes::None; + pads[port].enable_vibration = false; + pads[port].rumble_amplitude = 0; + pads[port].buttons = 0; + pads[port].last_button = PadButton::Undefined; + pads[port].axis_values.fill(0); + pads[port].axis_origin.fill(255); +} + +void Adapter::Reset() { + JoinThreads(); + ClearLibusbHandle(); + ResetDevices(); if (libusb_ctx) { libusb_exit(libusb_ctx); @@ -296,11 +390,11 @@ void Adapter::Reset() { std::vector<Common::ParamPackage> Adapter::GetInputDevices() const { std::vector<Common::ParamPackage> devices; - for (std::size_t port = 0; port < state.size(); ++port) { + for (std::size_t port = 0; port < pads.size(); ++port) { if (!DeviceConnected(port)) { continue; } - std::string name = fmt::format("Gamecube Controller {}", port); + std::string name = fmt::format("Gamecube Controller {}", port + 1); devices.emplace_back(Common::ParamPackage{ {"class", "gcpad"}, {"display", std::move(name)}, @@ -317,18 +411,18 @@ InputCommon::ButtonMapping Adapter::GetButtonMappingForDevice( // This list also excludes any button that can't be really mapped static constexpr std::array<std::pair<Settings::NativeButton::Values, PadButton>, 12> switch_to_gcadapter_button = { - std::pair{Settings::NativeButton::A, PadButton::PAD_BUTTON_A}, - {Settings::NativeButton::B, PadButton::PAD_BUTTON_B}, - {Settings::NativeButton::X, PadButton::PAD_BUTTON_X}, - {Settings::NativeButton::Y, PadButton::PAD_BUTTON_Y}, - {Settings::NativeButton::Plus, PadButton::PAD_BUTTON_START}, - {Settings::NativeButton::DLeft, PadButton::PAD_BUTTON_LEFT}, - {Settings::NativeButton::DUp, PadButton::PAD_BUTTON_UP}, - {Settings::NativeButton::DRight, PadButton::PAD_BUTTON_RIGHT}, - {Settings::NativeButton::DDown, PadButton::PAD_BUTTON_DOWN}, - {Settings::NativeButton::SL, PadButton::PAD_TRIGGER_L}, - {Settings::NativeButton::SR, PadButton::PAD_TRIGGER_R}, - {Settings::NativeButton::R, PadButton::PAD_TRIGGER_Z}, + std::pair{Settings::NativeButton::A, PadButton::ButtonA}, + {Settings::NativeButton::B, PadButton::ButtonB}, + {Settings::NativeButton::X, PadButton::ButtonX}, + {Settings::NativeButton::Y, PadButton::ButtonY}, + {Settings::NativeButton::Plus, PadButton::ButtonStart}, + {Settings::NativeButton::DLeft, PadButton::ButtonLeft}, + {Settings::NativeButton::DUp, PadButton::ButtonUp}, + {Settings::NativeButton::DRight, PadButton::ButtonRight}, + {Settings::NativeButton::DDown, PadButton::ButtonDown}, + {Settings::NativeButton::SL, PadButton::TriggerL}, + {Settings::NativeButton::SR, PadButton::TriggerR}, + {Settings::NativeButton::R, PadButton::TriggerZ}, }; if (!params.Has("port")) { return {}; @@ -351,8 +445,10 @@ InputCommon::ButtonMapping Adapter::GetButtonMappingForDevice( for (const auto& [switch_button, gcadapter_axis] : switch_to_gcadapter_axis) { Common::ParamPackage button_params({{"engine", "gcpad"}}); button_params.Set("port", params.Get("port", 0)); - button_params.Set("button", static_cast<int>(PadButton::PAD_STICK)); - button_params.Set("axis", static_cast<int>(gcadapter_axis)); + button_params.Set("button", static_cast<s32>(PadButton::Stick)); + button_params.Set("axis", static_cast<s32>(gcadapter_axis)); + button_params.Set("threshold", 0.5f); + button_params.Set("direction", "+"); mapping.insert_or_assign(switch_button, std::move(button_params)); } return mapping; @@ -381,46 +477,33 @@ InputCommon::AnalogMapping Adapter::GetAnalogMappingForDevice( } bool Adapter::DeviceConnected(std::size_t port) const { - return adapter_controllers_status[port] != ControllerTypes::None; -} - -void Adapter::ResetDeviceType(std::size_t port) { - adapter_controllers_status[port] = ControllerTypes::None; + return pads[port].type != ControllerTypes::None; } void Adapter::BeginConfiguration() { - get_origin.fill(true); - for (auto& pq : pad_queue) { - pq.Clear(); - } + pad_queue.Clear(); configuring = true; } void Adapter::EndConfiguration() { - for (auto& pq : pad_queue) { - pq.Clear(); - } + pad_queue.Clear(); configuring = false; } -std::array<Common::SPSCQueue<GCPadStatus>, 4>& Adapter::GetPadQueue() { +Common::SPSCQueue<GCPadStatus>& Adapter::GetPadQueue() { return pad_queue; } -const std::array<Common::SPSCQueue<GCPadStatus>, 4>& Adapter::GetPadQueue() const { +const Common::SPSCQueue<GCPadStatus>& Adapter::GetPadQueue() const { return pad_queue; } -std::array<GCState, 4>& Adapter::GetPadState() { - return state; -} - -const std::array<GCState, 4>& Adapter::GetPadState() const { - return state; +GCController& Adapter::GetPadState(std::size_t port) { + return pads.at(port); } -int Adapter::GetOriginValue(int port, int axis) const { - return origin_status[port].axis_values[axis]; +const GCController& Adapter::GetPadState(std::size_t port) const { + return pads.at(port); } } // namespace GCAdapter diff --git a/src/input_common/gcadapter/gc_adapter.h b/src/input_common/gcadapter/gc_adapter.h index 75bf9fe74..7a6c545bd 100644 --- a/src/input_common/gcadapter/gc_adapter.h +++ b/src/input_common/gcadapter/gc_adapter.h @@ -19,24 +19,23 @@ struct libusb_device_handle; namespace GCAdapter { enum class PadButton { - PAD_BUTTON_LEFT = 0x0001, - PAD_BUTTON_RIGHT = 0x0002, - PAD_BUTTON_DOWN = 0x0004, - PAD_BUTTON_UP = 0x0008, - PAD_TRIGGER_Z = 0x0010, - PAD_TRIGGER_R = 0x0020, - PAD_TRIGGER_L = 0x0040, - PAD_BUTTON_A = 0x0100, - PAD_BUTTON_B = 0x0200, - PAD_BUTTON_X = 0x0400, - PAD_BUTTON_Y = 0x0800, - PAD_BUTTON_START = 0x1000, + Undefined = 0x0000, + ButtonLeft = 0x0001, + ButtonRight = 0x0002, + ButtonDown = 0x0004, + ButtonUp = 0x0008, + TriggerZ = 0x0010, + TriggerR = 0x0020, + TriggerL = 0x0040, + ButtonA = 0x0100, + ButtonB = 0x0200, + ButtonX = 0x0400, + ButtonY = 0x0800, + ButtonStart = 0x1000, // Below is for compatibility with "AxisButton" type - PAD_STICK = 0x2000, + Stick = 0x2000, }; -extern const std::array<PadButton, 12> PadButtonArray; - enum class PadAxes : u8 { StickX, StickY, @@ -47,89 +46,122 @@ enum class PadAxes : u8 { Undefined, }; +enum class ControllerTypes { + None, + Wired, + Wireless, +}; + struct GCPadStatus { - u16 button{}; // Or-ed PAD_BUTTON_* and PAD_TRIGGER_* bits + std::size_t port{}; - std::array<u8, 6> axis_values{}; // Triggers and sticks, following indices defined in PadAxes - static constexpr u8 THRESHOLD = 50; // Threshold for axis press for polling + PadButton button{PadButton::Undefined}; // Or-ed PAD_BUTTON_* and PAD_TRIGGER_* bits - u8 port{}; PadAxes axis{PadAxes::Undefined}; - u8 axis_value{255}; + s16 axis_value{}; + u8 axis_threshold{50}; }; -struct GCState { - std::unordered_map<int, bool> buttons; - std::unordered_map<int, u16> axes; +struct GCController { + ControllerTypes type{}; + bool enable_vibration{}; + u8 rumble_amplitude{}; + u16 buttons{}; + PadButton last_button{}; + std::array<s16, 6> axis_values{}; + std::array<u8, 6> axis_origin{}; }; -enum class ControllerTypes { None, Wired, Wireless }; - class Adapter { public: - /// Initialize the GC Adapter capture and read sequence Adapter(); - - /// Close the adapter read thread and release the adapter ~Adapter(); + + /// Request a vibration for a controller + bool RumblePlay(std::size_t port, u8 amplitude); + /// Used for polling void BeginConfiguration(); void EndConfiguration(); + Common::SPSCQueue<GCPadStatus>& GetPadQueue(); + const Common::SPSCQueue<GCPadStatus>& GetPadQueue() const; + + GCController& GetPadState(std::size_t port); + const GCController& GetPadState(std::size_t port) const; + + /// Returns true if there is a device connected to port + bool DeviceConnected(std::size_t port) const; + + /// Used for automapping features std::vector<Common::ParamPackage> GetInputDevices() const; InputCommon::ButtonMapping GetButtonMappingForDevice(const Common::ParamPackage& params) const; InputCommon::AnalogMapping GetAnalogMappingForDevice(const Common::ParamPackage& params) const; - /// Returns true if there is a device connected to port - bool DeviceConnected(std::size_t port) const; +private: + using AdapterPayload = std::array<u8, 37>; - std::array<Common::SPSCQueue<GCPadStatus>, 4>& GetPadQueue(); - const std::array<Common::SPSCQueue<GCPadStatus>, 4>& GetPadQueue() const; + void UpdatePadType(std::size_t port, ControllerTypes pad_type); + void UpdateControllers(const AdapterPayload& adapter_payload); + void UpdateYuzuSettings(std::size_t port); + void UpdateStateButtons(std::size_t port, u8 b1, u8 b2); + void UpdateStateAxes(std::size_t port, const AdapterPayload& adapter_payload); + void UpdateVibrations(); - std::array<GCState, 4>& GetPadState(); - const std::array<GCState, 4>& GetPadState() const; + void AdapterInputThread(); - int GetOriginValue(int port, int axis) const; + void AdapterScanThread(); -private: - GCPadStatus GetPadStatus(std::size_t port, const std::array<u8, 37>& adapter_payload); + bool IsPayloadCorrect(const AdapterPayload& adapter_payload, s32 payload_size); + + // Updates vibration state of all controllers + void SendVibrations(); - void PadToState(const GCPadStatus& pad, GCState& state); + /// For use in initialization, querying devices to find the adapter + void Setup(); - void Read(); + /// Resets status of all GC controller devices to a disconnected state + void ResetDevices(); - /// Resets status of device connected to port - void ResetDeviceType(std::size_t port); + /// Resets status of device connected to a disconnected state + void ResetDevice(std::size_t port); /// Returns true if we successfully gain access to GC Adapter - bool CheckDeviceAccess(libusb_device* device); + bool CheckDeviceAccess(); - /// Captures GC Adapter endpoint address, - void GetGCEndpoint(libusb_device* device); + /// Captures GC Adapter endpoint address + /// Returns true if the endpoint was set correctly + bool GetGCEndpoint(libusb_device* device); /// For shutting down, clear all data, join all threads, release usb void Reset(); - /// For use in initialization, querying devices to find the adapter - void Setup(); + // Join all threads + void JoinThreads(); + + // Release usb handles + void ClearLibusbHandle(); libusb_device_handle* usb_adapter_handle = nullptr; + std::array<GCController, 4> pads; + Common::SPSCQueue<GCPadStatus> pad_queue; std::thread adapter_input_thread; - bool adapter_thread_running; + std::thread adapter_scan_thread; + bool adapter_input_thread_running; + bool adapter_scan_thread_running; + bool restart_scan_thread; libusb_context* libusb_ctx; - u8 input_endpoint = 0; - u8 output_endpoint = 0; - - bool configuring = false; + u8 input_endpoint{0}; + u8 output_endpoint{0}; + u8 input_error_counter{0}; + u8 output_error_counter{0}; + int vibration_counter{0}; - std::array<GCState, 4> state; - std::array<bool, 4> get_origin; - std::array<GCPadStatus, 4> origin_status; - std::array<Common::SPSCQueue<GCPadStatus>, 4> pad_queue; - std::array<ControllerTypes, 4> adapter_controllers_status{}; + bool configuring{false}; + bool rumble_enabled{true}; + bool vibration_changed{true}; }; - } // namespace GCAdapter diff --git a/src/input_common/gcadapter/gc_poller.cpp b/src/input_common/gcadapter/gc_poller.cpp index 92e9e8e89..9670bdeb2 100644 --- a/src/input_common/gcadapter/gc_poller.cpp +++ b/src/input_common/gcadapter/gc_poller.cpp @@ -15,36 +15,35 @@ namespace InputCommon { class GCButton final : public Input::ButtonDevice { public: - explicit GCButton(int port_, int button_, const GCAdapter::Adapter* adapter) + explicit GCButton(u32 port_, s32 button_, const GCAdapter::Adapter* adapter) : port(port_), button(button_), gcadapter(adapter) {} ~GCButton() override; bool GetStatus() const override { if (gcadapter->DeviceConnected(port)) { - return gcadapter->GetPadState()[port].buttons.at(button); + return (gcadapter->GetPadState(port).buttons & button) != 0; } return false; } private: - const int port; - const int button; + const u32 port; + const s32 button; const GCAdapter::Adapter* gcadapter; }; class GCAxisButton final : public Input::ButtonDevice { public: - explicit GCAxisButton(int port_, int axis_, float threshold_, bool trigger_if_greater_, + explicit GCAxisButton(u32 port_, u32 axis_, float threshold_, bool trigger_if_greater_, const GCAdapter::Adapter* adapter) : port(port_), axis(axis_), threshold(threshold_), trigger_if_greater(trigger_if_greater_), - gcadapter(adapter), - origin_value(static_cast<float>(adapter->GetOriginValue(port_, axis_))) {} + gcadapter(adapter) {} bool GetStatus() const override { if (gcadapter->DeviceConnected(port)) { - const float current_axis_value = gcadapter->GetPadState()[port].axes.at(axis); - const float axis_value = (current_axis_value - origin_value) / 128.0f; + const float current_axis_value = gcadapter->GetPadState(port).axis_values.at(axis); + const float axis_value = current_axis_value / 128.0f; if (trigger_if_greater) { // TODO: Might be worthwile to set a slider for the trigger threshold. It is // currently always set to 0.5 in configure_input_player.cpp ZL/ZR HandleClick @@ -56,12 +55,11 @@ public: } private: - const int port; - const int axis; + const u32 port; + const u32 axis; float threshold; bool trigger_if_greater; const GCAdapter::Adapter* gcadapter; - const float origin_value; }; GCButtonFactory::GCButtonFactory(std::shared_ptr<GCAdapter::Adapter> adapter_) @@ -70,10 +68,10 @@ GCButtonFactory::GCButtonFactory(std::shared_ptr<GCAdapter::Adapter> adapter_) GCButton::~GCButton() = default; std::unique_ptr<Input::ButtonDevice> GCButtonFactory::Create(const Common::ParamPackage& params) { - const int button_id = params.Get("button", 0); - const int port = params.Get("port", 0); + const auto button_id = params.Get("button", 0); + const auto port = static_cast<u32>(params.Get("port", 0)); - constexpr int PAD_STICK_ID = static_cast<u16>(GCAdapter::PadButton::PAD_STICK); + constexpr s32 PAD_STICK_ID = static_cast<s32>(GCAdapter::PadButton::Stick); // button is not an axis/stick button if (button_id != PAD_STICK_ID) { @@ -98,7 +96,6 @@ std::unique_ptr<Input::ButtonDevice> GCButtonFactory::Create(const Common::Param adapter.get()); } - UNREACHABLE(); return nullptr; } @@ -106,32 +103,25 @@ Common::ParamPackage GCButtonFactory::GetNextInput() const { Common::ParamPackage params; GCAdapter::GCPadStatus pad; auto& queue = adapter->GetPadQueue(); - for (std::size_t port = 0; port < queue.size(); ++port) { - while (queue[port].Pop(pad)) { - // This while loop will break on the earliest detected button - params.Set("engine", "gcpad"); - params.Set("port", static_cast<int>(port)); - for (const auto& button : GCAdapter::PadButtonArray) { - const u16 button_value = static_cast<u16>(button); - if (pad.button & button_value) { - params.Set("button", button_value); - break; - } - } + while (queue.Pop(pad)) { + // This while loop will break on the earliest detected button + params.Set("engine", "gcpad"); + params.Set("port", static_cast<s32>(pad.port)); + if (pad.button != GCAdapter::PadButton::Undefined) { + params.Set("button", static_cast<u16>(pad.button)); + } - // For Axis button implementation - if (pad.axis != GCAdapter::PadAxes::Undefined) { - params.Set("axis", static_cast<u8>(pad.axis)); - params.Set("button", static_cast<u16>(GCAdapter::PadButton::PAD_STICK)); - if (pad.axis_value > 128) { - params.Set("direction", "+"); - params.Set("threshold", "0.25"); - } else { - params.Set("direction", "-"); - params.Set("threshold", "-0.25"); - } - break; + // For Axis button implementation + if (pad.axis != GCAdapter::PadAxes::Undefined) { + params.Set("axis", static_cast<u8>(pad.axis)); + params.Set("button", static_cast<u16>(GCAdapter::PadButton::Stick)); + params.Set("threshold", "0.25"); + if (pad.axis_value > 0) { + params.Set("direction", "+"); + } else { + params.Set("direction", "-"); } + break; } } return params; @@ -149,26 +139,30 @@ void GCButtonFactory::EndConfiguration() { class GCAnalog final : public Input::AnalogDevice { public: - GCAnalog(int port_, int axis_x_, int axis_y_, float deadzone_, - const GCAdapter::Adapter* adapter, float range_) - : port(port_), axis_x(axis_x_), axis_y(axis_y_), deadzone(deadzone_), gcadapter(adapter), - origin_value_x(static_cast<float>(adapter->GetOriginValue(port_, axis_x_))), - origin_value_y(static_cast<float>(adapter->GetOriginValue(port_, axis_y_))), - range(range_) {} - - float GetAxis(int axis) const { + explicit GCAnalog(u32 port_, u32 axis_x_, u32 axis_y_, bool invert_x_, bool invert_y_, + float deadzone_, float range_, const GCAdapter::Adapter* adapter) + : port(port_), axis_x(axis_x_), axis_y(axis_y_), invert_x(invert_x_), invert_y(invert_y_), + deadzone(deadzone_), range(range_), gcadapter(adapter) {} + + float GetAxis(u32 axis) const { if (gcadapter->DeviceConnected(port)) { std::lock_guard lock{mutex}; - const auto origin_value = axis % 2 == 0 ? origin_value_x : origin_value_y; - return (gcadapter->GetPadState()[port].axes.at(axis) - origin_value) / (100.0f * range); + const auto axis_value = + static_cast<float>(gcadapter->GetPadState(port).axis_values.at(axis)); + return (axis_value) / (100.0f * range); } return 0.0f; } - std::pair<float, float> GetAnalog(int axis_x, int axis_y) const { - float x = GetAxis(axis_x); - float y = GetAxis(axis_y); - + std::pair<float, float> GetAnalog(u32 analog_axis_x, u32 analog_axis_y) const { + float x = GetAxis(analog_axis_x); + float y = GetAxis(analog_axis_y); + if (invert_x) { + x = -x; + } + if (invert_y) { + y = -y; + } // Make sure the coordinates are in the unit circle, // otherwise normalize it. float r = x * x + y * y; @@ -208,14 +202,14 @@ public: } private: - const int port; - const int axis_x; - const int axis_y; + const u32 port; + const u32 axis_x; + const u32 axis_y; + const bool invert_x; + const bool invert_y; const float deadzone; - const GCAdapter::Adapter* gcadapter; - const float origin_value_x; - const float origin_value_y; const float range; + const GCAdapter::Adapter* gcadapter; mutable std::mutex mutex; }; @@ -231,13 +225,18 @@ GCAnalogFactory::GCAnalogFactory(std::shared_ptr<GCAdapter::Adapter> adapter_) * - "axis_y": the index of the axis to be bind as y-axis */ std::unique_ptr<Input::AnalogDevice> GCAnalogFactory::Create(const Common::ParamPackage& params) { - const int port = params.Get("port", 0); - const int axis_x = params.Get("axis_x", 0); - const int axis_y = params.Get("axis_y", 1); - const float deadzone = std::clamp(params.Get("deadzone", 0.0f), 0.0f, 1.0f); - const float range = std::clamp(params.Get("range", 1.0f), 0.50f, 1.50f); - - return std::make_unique<GCAnalog>(port, axis_x, axis_y, deadzone, adapter.get(), range); + const auto port = static_cast<u32>(params.Get("port", 0)); + const auto axis_x = static_cast<u32>(params.Get("axis_x", 0)); + const auto axis_y = static_cast<u32>(params.Get("axis_y", 1)); + const auto deadzone = std::clamp(params.Get("deadzone", 0.0f), 0.0f, 1.0f); + const auto range = std::clamp(params.Get("range", 1.0f), 0.50f, 1.50f); + const std::string invert_x_value = params.Get("invert_x", "+"); + const std::string invert_y_value = params.Get("invert_y", "+"); + const bool invert_x = invert_x_value == "-"; + const bool invert_y = invert_y_value == "-"; + + return std::make_unique<GCAnalog>(port, axis_x, axis_y, invert_x, invert_y, deadzone, range, + adapter.get()); } void GCAnalogFactory::BeginConfiguration() { @@ -252,31 +251,51 @@ void GCAnalogFactory::EndConfiguration() { Common::ParamPackage GCAnalogFactory::GetNextInput() { GCAdapter::GCPadStatus pad; + Common::ParamPackage params; auto& queue = adapter->GetPadQueue(); - for (std::size_t port = 0; port < queue.size(); ++port) { - while (queue[port].Pop(pad)) { - if (pad.axis == GCAdapter::PadAxes::Undefined || - std::abs((pad.axis_value - 128.0f) / 128.0f) < 0.1) { - continue; - } - // An analog device needs two axes, so we need to store the axis for later and wait for - // a second input event. The axes also must be from the same joystick. - const u8 axis = static_cast<u8>(pad.axis); - if (analog_x_axis == -1) { - analog_x_axis = axis; - controller_number = static_cast<int>(port); - } else if (analog_y_axis == -1 && analog_x_axis != axis && - controller_number == static_cast<int>(port)) { - analog_y_axis = axis; - } + while (queue.Pop(pad)) { + if (pad.button != GCAdapter::PadButton::Undefined) { + params.Set("engine", "gcpad"); + params.Set("port", static_cast<s32>(pad.port)); + params.Set("button", static_cast<u16>(pad.button)); + return params; + } + if (pad.axis == GCAdapter::PadAxes::Undefined || + std::abs(static_cast<float>(pad.axis_value) / 128.0f) < 0.1f) { + continue; + } + // An analog device needs two axes, so we need to store the axis for later and wait for + // a second input event. The axes also must be from the same joystick. + const u8 axis = static_cast<u8>(pad.axis); + if (axis == 0 || axis == 1) { + analog_x_axis = 0; + analog_y_axis = 1; + controller_number = static_cast<s32>(pad.port); + break; + } + if (axis == 2 || axis == 3) { + analog_x_axis = 2; + analog_y_axis = 3; + controller_number = static_cast<s32>(pad.port); + break; + } + + if (analog_x_axis == -1) { + analog_x_axis = axis; + controller_number = static_cast<s32>(pad.port); + } else if (analog_y_axis == -1 && analog_x_axis != axis && + controller_number == static_cast<s32>(pad.port)) { + analog_y_axis = axis; + break; } } - Common::ParamPackage params; if (analog_x_axis != -1 && analog_y_axis != -1) { params.Set("engine", "gcpad"); params.Set("port", controller_number); params.Set("axis_x", analog_x_axis); params.Set("axis_y", analog_y_axis); + params.Set("invert_x", "+"); + params.Set("invert_y", "+"); analog_x_axis = -1; analog_y_axis = -1; controller_number = -1; @@ -285,4 +304,43 @@ Common::ParamPackage GCAnalogFactory::GetNextInput() { return params; } +class GCVibration final : public Input::VibrationDevice { +public: + explicit GCVibration(u32 port_, GCAdapter::Adapter* adapter) + : port(port_), gcadapter(adapter) {} + + u8 GetStatus() const override { + return gcadapter->RumblePlay(port, 0); + } + + bool SetRumblePlay(f32 amp_low, [[maybe_unused]] f32 freq_low, f32 amp_high, + [[maybe_unused]] f32 freq_high) const override { + const auto mean_amplitude = (amp_low + amp_high) * 0.5f; + const auto processed_amplitude = + static_cast<u8>((mean_amplitude + std::pow(mean_amplitude, 0.3f)) * 0.5f * 0x8); + + return gcadapter->RumblePlay(port, processed_amplitude); + } + +private: + const u32 port; + GCAdapter::Adapter* gcadapter; +}; + +/// An vibration device factory that creates vibration devices from GC Adapter +GCVibrationFactory::GCVibrationFactory(std::shared_ptr<GCAdapter::Adapter> adapter_) + : adapter(std::move(adapter_)) {} + +/** + * Creates a vibration device from a joystick + * @param params contains parameters for creating the device: + * - "port": the nth gcpad on the adapter + */ +std::unique_ptr<Input::VibrationDevice> GCVibrationFactory::Create( + const Common::ParamPackage& params) { + const auto port = static_cast<u32>(params.Get("port", 0)); + + return std::make_unique<GCVibration>(port, adapter.get()); +} + } // namespace InputCommon diff --git a/src/input_common/gcadapter/gc_poller.h b/src/input_common/gcadapter/gc_poller.h index 0527f328f..d1271e3ea 100644 --- a/src/input_common/gcadapter/gc_poller.h +++ b/src/input_common/gcadapter/gc_poller.h @@ -64,4 +64,15 @@ private: bool polling = false; }; +/// A vibration device factory creates vibration devices from GC Adapter +class GCVibrationFactory final : public Input::Factory<Input::VibrationDevice> { +public: + explicit GCVibrationFactory(std::shared_ptr<GCAdapter::Adapter> adapter_); + + std::unique_ptr<Input::VibrationDevice> Create(const Common::ParamPackage& params) override; + +private: + std::shared_ptr<GCAdapter::Adapter> adapter; +}; + } // namespace InputCommon diff --git a/src/input_common/keyboard.cpp b/src/input_common/keyboard.cpp index afb8e6612..24a6f7a33 100644 --- a/src/input_common/keyboard.cpp +++ b/src/input_common/keyboard.cpp @@ -49,8 +49,9 @@ public: void ChangeKeyStatus(int key_code, bool pressed) { std::lock_guard guard{mutex}; for (const KeyButtonPair& pair : list) { - if (pair.key_code == key_code) + if (pair.key_code == key_code) { pair.key_button->status.store(pressed); + } } } @@ -73,7 +74,7 @@ KeyButton::~KeyButton() { } std::unique_ptr<Input::ButtonDevice> Keyboard::Create(const Common::ParamPackage& params) { - int key_code = params.Get("code", 0); + const int key_code = params.Get("code", 0); std::unique_ptr<KeyButton> button = std::make_unique<KeyButton>(key_button_list); key_button_list->AddKeyButton(key_code, button.get()); return button; diff --git a/src/input_common/main.cpp b/src/input_common/main.cpp index 8da829132..7c4e7dd3b 100644 --- a/src/input_common/main.cpp +++ b/src/input_common/main.cpp @@ -10,7 +10,9 @@ #include "input_common/gcadapter/gc_poller.h" #include "input_common/keyboard.h" #include "input_common/main.h" -#include "input_common/motion_emu.h" +#include "input_common/motion_from_button.h" +#include "input_common/mouse/mouse_input.h" +#include "input_common/mouse/mouse_poller.h" #include "input_common/touch_from_button.h" #include "input_common/udp/client.h" #include "input_common/udp/udp.h" @@ -27,13 +29,15 @@ struct InputSubsystem::Impl { Input::RegisterFactory<Input::ButtonDevice>("gcpad", gcbuttons); gcanalog = std::make_shared<GCAnalogFactory>(gcadapter); Input::RegisterFactory<Input::AnalogDevice>("gcpad", gcanalog); + gcvibration = std::make_shared<GCVibrationFactory>(gcadapter); + Input::RegisterFactory<Input::VibrationDevice>("gcpad", gcvibration); keyboard = std::make_shared<Keyboard>(); Input::RegisterFactory<Input::ButtonDevice>("keyboard", keyboard); Input::RegisterFactory<Input::AnalogDevice>("analog_from_button", std::make_shared<AnalogFromButton>()); - motion_emu = std::make_shared<MotionEmu>(); - Input::RegisterFactory<Input::MotionDevice>("motion_emu", motion_emu); + Input::RegisterFactory<Input::MotionDevice>("keyboard", + std::make_shared<MotionFromButton>()); Input::RegisterFactory<Input::TouchDevice>("touch_from_button", std::make_shared<TouchFromButtonFactory>()); @@ -46,35 +50,56 @@ struct InputSubsystem::Impl { Input::RegisterFactory<Input::MotionDevice>("cemuhookudp", udpmotion); udptouch = std::make_shared<UDPTouchFactory>(udp); Input::RegisterFactory<Input::TouchDevice>("cemuhookudp", udptouch); + + mouse = std::make_shared<MouseInput::Mouse>(); + mousebuttons = std::make_shared<MouseButtonFactory>(mouse); + Input::RegisterFactory<Input::ButtonDevice>("mouse", mousebuttons); + mouseanalog = std::make_shared<MouseAnalogFactory>(mouse); + Input::RegisterFactory<Input::AnalogDevice>("mouse", mouseanalog); + mousemotion = std::make_shared<MouseMotionFactory>(mouse); + Input::RegisterFactory<Input::MotionDevice>("mouse", mousemotion); + mousetouch = std::make_shared<MouseTouchFactory>(mouse); + Input::RegisterFactory<Input::TouchDevice>("mouse", mousetouch); } void Shutdown() { Input::UnregisterFactory<Input::ButtonDevice>("keyboard"); + Input::UnregisterFactory<Input::MotionDevice>("keyboard"); keyboard.reset(); Input::UnregisterFactory<Input::AnalogDevice>("analog_from_button"); - Input::UnregisterFactory<Input::MotionDevice>("motion_emu"); - motion_emu.reset(); Input::UnregisterFactory<Input::TouchDevice>("touch_from_button"); #ifdef HAVE_SDL2 sdl.reset(); #endif Input::UnregisterFactory<Input::ButtonDevice>("gcpad"); Input::UnregisterFactory<Input::AnalogDevice>("gcpad"); + Input::UnregisterFactory<Input::VibrationDevice>("gcpad"); gcbuttons.reset(); gcanalog.reset(); + gcvibration.reset(); Input::UnregisterFactory<Input::MotionDevice>("cemuhookudp"); Input::UnregisterFactory<Input::TouchDevice>("cemuhookudp"); udpmotion.reset(); udptouch.reset(); + + Input::UnregisterFactory<Input::ButtonDevice>("mouse"); + Input::UnregisterFactory<Input::AnalogDevice>("mouse"); + Input::UnregisterFactory<Input::MotionDevice>("mouse"); + Input::UnregisterFactory<Input::TouchDevice>("mouse"); + + mousebuttons.reset(); + mouseanalog.reset(); + mousemotion.reset(); + mousetouch.reset(); } [[nodiscard]] std::vector<Common::ParamPackage> GetInputDevices() const { std::vector<Common::ParamPackage> devices = { Common::ParamPackage{{"display", "Any"}, {"class", "any"}}, - Common::ParamPackage{{"display", "Keyboard/Mouse"}, {"class", "key"}}, + Common::ParamPackage{{"display", "Keyboard/Mouse"}, {"class", "keyboard"}}, }; #ifdef HAVE_SDL2 auto sdl_devices = sdl->GetInputDevices(); @@ -92,10 +117,6 @@ struct InputSubsystem::Impl { if (!params.Has("class") || params.Get("class", "") == "any") { return {}; } - if (params.Get("class", "") == "key") { - // TODO consider returning the SDL key codes for the default keybindings - return {}; - } if (params.Get("class", "") == "gcpad") { return gcadapter->GetAnalogMappingForDevice(params); } @@ -112,10 +133,6 @@ struct InputSubsystem::Impl { if (!params.Has("class") || params.Get("class", "") == "any") { return {}; } - if (params.Get("class", "") == "key") { - // TODO consider returning the SDL key codes for the default keybindings - return {}; - } if (params.Get("class", "") == "gcpad") { return gcadapter->GetButtonMappingForDevice(params); } @@ -140,16 +157,21 @@ struct InputSubsystem::Impl { } std::shared_ptr<Keyboard> keyboard; - std::shared_ptr<MotionEmu> motion_emu; #ifdef HAVE_SDL2 std::unique_ptr<SDL::State> sdl; #endif std::shared_ptr<GCButtonFactory> gcbuttons; std::shared_ptr<GCAnalogFactory> gcanalog; + std::shared_ptr<GCVibrationFactory> gcvibration; std::shared_ptr<UDPMotionFactory> udpmotion; std::shared_ptr<UDPTouchFactory> udptouch; + std::shared_ptr<MouseButtonFactory> mousebuttons; + std::shared_ptr<MouseAnalogFactory> mouseanalog; + std::shared_ptr<MouseMotionFactory> mousemotion; + std::shared_ptr<MouseTouchFactory> mousetouch; std::shared_ptr<CemuhookUDP::Client> udp; std::shared_ptr<GCAdapter::Adapter> gcadapter; + std::shared_ptr<MouseInput::Mouse> mouse; }; InputSubsystem::InputSubsystem() : impl{std::make_unique<Impl>()} {} @@ -172,12 +194,12 @@ const Keyboard* InputSubsystem::GetKeyboard() const { return impl->keyboard.get(); } -MotionEmu* InputSubsystem::GetMotionEmu() { - return impl->motion_emu.get(); +MouseInput::Mouse* InputSubsystem::GetMouse() { + return impl->mouse.get(); } -const MotionEmu* InputSubsystem::GetMotionEmu() const { - return impl->motion_emu.get(); +const MouseInput::Mouse* InputSubsystem::GetMouse() const { + return impl->mouse.get(); } std::vector<Common::ParamPackage> InputSubsystem::GetInputDevices() const { @@ -192,6 +214,10 @@ ButtonMapping InputSubsystem::GetButtonMappingForDevice(const Common::ParamPacka return impl->GetButtonMappingForDevice(device); } +MotionMapping InputSubsystem::GetMotionMappingForDevice(const Common::ParamPackage& device) const { + return impl->GetMotionMappingForDevice(device); +} + GCAnalogFactory* InputSubsystem::GetGCAnalogs() { return impl->gcanalog.get(); } @@ -224,11 +250,43 @@ const UDPTouchFactory* InputSubsystem::GetUDPTouch() const { return impl->udptouch.get(); } +MouseButtonFactory* InputSubsystem::GetMouseButtons() { + return impl->mousebuttons.get(); +} + +const MouseButtonFactory* InputSubsystem::GetMouseButtons() const { + return impl->mousebuttons.get(); +} + +MouseAnalogFactory* InputSubsystem::GetMouseAnalogs() { + return impl->mouseanalog.get(); +} + +const MouseAnalogFactory* InputSubsystem::GetMouseAnalogs() const { + return impl->mouseanalog.get(); +} + +MouseMotionFactory* InputSubsystem::GetMouseMotions() { + return impl->mousemotion.get(); +} + +const MouseMotionFactory* InputSubsystem::GetMouseMotions() const { + return impl->mousemotion.get(); +} + +MouseTouchFactory* InputSubsystem::GetMouseTouch() { + return impl->mousetouch.get(); +} + +const MouseTouchFactory* InputSubsystem::GetMouseTouch() const { + return impl->mousetouch.get(); +} + void InputSubsystem::ReloadInputDevices() { if (!impl->udp) { return; } - impl->udp->ReloadUDPClient(); + impl->udp->ReloadSockets(); } std::vector<std::unique_ptr<Polling::DevicePoller>> InputSubsystem::GetPollers( diff --git a/src/input_common/main.h b/src/input_common/main.h index dded3f1ef..5d6f26385 100644 --- a/src/input_common/main.h +++ b/src/input_common/main.h @@ -25,6 +25,10 @@ namespace Settings::NativeMotion { enum Values : int; } +namespace MouseInput { +class Mouse; +} + namespace InputCommon { namespace Polling { @@ -56,8 +60,11 @@ class GCAnalogFactory; class GCButtonFactory; class UDPMotionFactory; class UDPTouchFactory; +class MouseButtonFactory; +class MouseAnalogFactory; +class MouseMotionFactory; +class MouseTouchFactory; class Keyboard; -class MotionEmu; /** * Given a ParamPackage for a Device returned from `GetInputDevices`, attempt to get the default @@ -90,11 +97,11 @@ public: /// Retrieves the underlying keyboard device. [[nodiscard]] const Keyboard* GetKeyboard() const; - /// Retrieves the underlying motion emulation factory. - [[nodiscard]] MotionEmu* GetMotionEmu(); + /// Retrieves the underlying mouse device. + [[nodiscard]] MouseInput::Mouse* GetMouse(); - /// Retrieves the underlying motion emulation factory. - [[nodiscard]] const MotionEmu* GetMotionEmu() const; + /// Retrieves the underlying mouse device. + [[nodiscard]] const MouseInput::Mouse* GetMouse() const; /** * Returns all available input devices that this Factory can create a new device with. @@ -137,6 +144,30 @@ public: /// Retrieves the underlying udp touch handler. [[nodiscard]] const UDPTouchFactory* GetUDPTouch() const; + /// Retrieves the underlying GameCube button handler. + [[nodiscard]] MouseButtonFactory* GetMouseButtons(); + + /// Retrieves the underlying GameCube button handler. + [[nodiscard]] const MouseButtonFactory* GetMouseButtons() const; + + /// Retrieves the underlying udp touch handler. + [[nodiscard]] MouseAnalogFactory* GetMouseAnalogs(); + + /// Retrieves the underlying udp touch handler. + [[nodiscard]] const MouseAnalogFactory* GetMouseAnalogs() const; + + /// Retrieves the underlying udp motion handler. + [[nodiscard]] MouseMotionFactory* GetMouseMotions(); + + /// Retrieves the underlying udp motion handler. + [[nodiscard]] const MouseMotionFactory* GetMouseMotions() const; + + /// Retrieves the underlying udp touch handler. + [[nodiscard]] MouseTouchFactory* GetMouseTouch(); + + /// Retrieves the underlying udp touch handler. + [[nodiscard]] const MouseTouchFactory* GetMouseTouch() const; + /// Reloads the input devices void ReloadInputDevices(); diff --git a/src/input_common/motion_emu.cpp b/src/input_common/motion_emu.cpp deleted file mode 100644 index 69fd3c1d2..000000000 --- a/src/input_common/motion_emu.cpp +++ /dev/null @@ -1,178 +0,0 @@ -// Copyright 2017 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <chrono> -#include <mutex> -#include <thread> -#include <tuple> -#include "common/math_util.h" -#include "common/quaternion.h" -#include "common/thread.h" -#include "common/vector_math.h" -#include "input_common/motion_emu.h" - -namespace InputCommon { - -// Implementation class of the motion emulation device -class MotionEmuDevice { -public: - MotionEmuDevice(int update_millisecond, float sensitivity) - : update_millisecond(update_millisecond), - update_duration(std::chrono::duration_cast<std::chrono::steady_clock::duration>( - std::chrono::milliseconds(update_millisecond))), - sensitivity(sensitivity), motion_emu_thread(&MotionEmuDevice::MotionEmuThread, this) {} - - ~MotionEmuDevice() { - if (motion_emu_thread.joinable()) { - shutdown_event.Set(); - motion_emu_thread.join(); - } - } - - void BeginTilt(int x, int y) { - mouse_origin = Common::MakeVec(x, y); - is_tilting = true; - } - - void Tilt(int x, int y) { - auto mouse_move = Common::MakeVec(x, y) - mouse_origin; - if (is_tilting) { - std::lock_guard guard{tilt_mutex}; - if (mouse_move.x == 0 && mouse_move.y == 0) { - tilt_angle = 0; - } else { - tilt_direction = mouse_move.Cast<float>(); - tilt_angle = - std::clamp(tilt_direction.Normalize() * sensitivity, 0.0f, Common::PI * 0.5f); - } - } - } - - void EndTilt() { - std::lock_guard guard{tilt_mutex}; - tilt_angle = 0; - is_tilting = false; - } - - Input::MotionStatus GetStatus() { - std::lock_guard guard{status_mutex}; - return status; - } - -private: - const int update_millisecond; - const std::chrono::steady_clock::duration update_duration; - const float sensitivity; - - Common::Vec2<int> mouse_origin; - - std::mutex tilt_mutex; - Common::Vec2<float> tilt_direction; - float tilt_angle = 0; - - bool is_tilting = false; - - Common::Event shutdown_event; - - Input::MotionStatus status; - std::mutex status_mutex; - - // Note: always keep the thread declaration at the end so that other objects are initialized - // before this! - std::thread motion_emu_thread; - - void MotionEmuThread() { - auto update_time = std::chrono::steady_clock::now(); - Common::Quaternion<float> q = Common::MakeQuaternion(Common::Vec3<float>(), 0); - Common::Quaternion<float> old_q; - - while (!shutdown_event.WaitUntil(update_time)) { - update_time += update_duration; - old_q = q; - - { - std::lock_guard guard{tilt_mutex}; - - // Find the quaternion describing current 3DS tilting - q = Common::MakeQuaternion( - Common::MakeVec(-tilt_direction.y, 0.0f, tilt_direction.x), tilt_angle); - } - - auto inv_q = q.Inverse(); - - // Set the gravity vector in world space - auto gravity = Common::MakeVec(0.0f, -1.0f, 0.0f); - - // Find the angular rate vector in world space - auto angular_rate = ((q - old_q) * inv_q).xyz * 2; - angular_rate *= 1000 / update_millisecond / Common::PI * 180; - - // Transform the two vectors from world space to 3DS space - gravity = QuaternionRotate(inv_q, gravity); - angular_rate = QuaternionRotate(inv_q, angular_rate); - - // TODO: Calculate the correct rotation vector and orientation matrix - const auto matrix4x4 = q.ToMatrix(); - const auto rotation = Common::MakeVec(0.0f, 0.0f, 0.0f); - const std::array orientation{ - Common::Vec3f(matrix4x4[0], matrix4x4[1], -matrix4x4[2]), - Common::Vec3f(matrix4x4[4], matrix4x4[5], -matrix4x4[6]), - Common::Vec3f(-matrix4x4[8], -matrix4x4[9], matrix4x4[10]), - }; - - // Update the sensor state - { - std::lock_guard guard{status_mutex}; - status = std::make_tuple(gravity, angular_rate, rotation, orientation); - } - } - } -}; - -// Interface wrapper held by input receiver as a unique_ptr. It holds the implementation class as -// a shared_ptr, which is also observed by the factory class as a weak_ptr. In this way the factory -// can forward all the inputs to the implementation only when it is valid. -class MotionEmuDeviceWrapper : public Input::MotionDevice { -public: - MotionEmuDeviceWrapper(int update_millisecond, float sensitivity) { - device = std::make_shared<MotionEmuDevice>(update_millisecond, sensitivity); - } - - Input::MotionStatus GetStatus() const override { - return device->GetStatus(); - } - - std::shared_ptr<MotionEmuDevice> device; -}; - -std::unique_ptr<Input::MotionDevice> MotionEmu::Create(const Common::ParamPackage& params) { - int update_period = params.Get("update_period", 100); - float sensitivity = params.Get("sensitivity", 0.01f); - auto device_wrapper = std::make_unique<MotionEmuDeviceWrapper>(update_period, sensitivity); - // Previously created device is disconnected here. Having two motion devices for 3DS is not - // expected. - current_device = device_wrapper->device; - return device_wrapper; -} - -void MotionEmu::BeginTilt(int x, int y) { - if (auto ptr = current_device.lock()) { - ptr->BeginTilt(x, y); - } -} - -void MotionEmu::Tilt(int x, int y) { - if (auto ptr = current_device.lock()) { - ptr->Tilt(x, y); - } -} - -void MotionEmu::EndTilt() { - if (auto ptr = current_device.lock()) { - ptr->EndTilt(); - } -} - -} // namespace InputCommon diff --git a/src/input_common/motion_emu.h b/src/input_common/motion_emu.h deleted file mode 100644 index 7a7e22467..000000000 --- a/src/input_common/motion_emu.h +++ /dev/null @@ -1,46 +0,0 @@ -// Copyright 2017 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "core/frontend/input.h" - -namespace InputCommon { - -class MotionEmuDevice; - -class MotionEmu : public Input::Factory<Input::MotionDevice> { -public: - /** - * Creates a motion device emulated from mouse input - * @param params contains parameters for creating the device: - * - "update_period": update period in milliseconds - * - "sensitivity": the coefficient converting mouse movement to tilting angle - */ - std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override; - - /** - * Signals that a motion sensor tilt has begun. - * @param x the x-coordinate of the cursor - * @param y the y-coordinate of the cursor - */ - void BeginTilt(int x, int y); - - /** - * Signals that a motion sensor tilt is occurring. - * @param x the x-coordinate of the cursor - * @param y the y-coordinate of the cursor - */ - void Tilt(int x, int y); - - /** - * Signals that a motion sensor tilt has ended. - */ - void EndTilt(); - -private: - std::weak_ptr<MotionEmuDevice> current_device; -}; - -} // namespace InputCommon diff --git a/src/input_common/motion_from_button.cpp b/src/input_common/motion_from_button.cpp new file mode 100644 index 000000000..29045a673 --- /dev/null +++ b/src/input_common/motion_from_button.cpp @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "input_common/motion_from_button.h" +#include "input_common/motion_input.h" + +namespace InputCommon { + +class MotionKey final : public Input::MotionDevice { +public: + using Button = std::unique_ptr<Input::ButtonDevice>; + + explicit MotionKey(Button key_) : key(std::move(key_)) {} + + Input::MotionStatus GetStatus() const override { + + if (key->GetStatus()) { + return motion.GetRandomMotion(2, 6); + } + return motion.GetRandomMotion(0, 0); + } + +private: + Button key; + InputCommon::MotionInput motion{0.0f, 0.0f, 0.0f}; +}; + +std::unique_ptr<Input::MotionDevice> MotionFromButton::Create(const Common::ParamPackage& params) { + auto key = Input::CreateDevice<Input::ButtonDevice>(params.Serialize()); + return std::make_unique<MotionKey>(std::move(key)); +} + +} // namespace InputCommon diff --git a/src/input_common/motion_from_button.h b/src/input_common/motion_from_button.h new file mode 100644 index 000000000..a959046fb --- /dev/null +++ b/src/input_common/motion_from_button.h @@ -0,0 +1,25 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "core/frontend/input.h" + +namespace InputCommon { + +/** + * An motion device factory that takes a keyboard button and uses it as a random + * motion device. + */ +class MotionFromButton final : public Input::Factory<Input::MotionDevice> { +public: + /** + * Creates an motion device from button devices + * @param params contains parameters for creating the device: + * - "key": a serialized ParamPackage for creating a button device + */ + std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override; +}; + +} // namespace InputCommon diff --git a/src/input_common/motion_input.cpp b/src/input_common/motion_input.cpp index 22a849866..6a65f175e 100644 --- a/src/input_common/motion_input.cpp +++ b/src/input_common/motion_input.cpp @@ -2,13 +2,13 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included +#include <random> #include "common/math_util.h" #include "input_common/motion_input.h" namespace InputCommon { -MotionInput::MotionInput(f32 new_kp, f32 new_ki, f32 new_kd) - : kp(new_kp), ki(new_ki), kd(new_kd), quat{{0, 0, -1}, 0} {} +MotionInput::MotionInput(f32 new_kp, f32 new_ki, f32 new_kd) : kp(new_kp), ki(new_ki), kd(new_kd) {} void MotionInput::SetAcceleration(const Common::Vec3f& acceleration) { accel = acceleration; @@ -16,8 +16,16 @@ void MotionInput::SetAcceleration(const Common::Vec3f& acceleration) { void MotionInput::SetGyroscope(const Common::Vec3f& gyroscope) { gyro = gyroscope - gyro_drift; + + // Auto adjust drift to minimize drift + if (!IsMoving(0.1f)) { + gyro_drift = (gyro_drift * 0.9999f) + (gyroscope * 0.0001f); + } + if (gyro.Length2() < gyro_threshold) { gyro = {}; + } else { + only_accelerometer = false; } } @@ -50,7 +58,7 @@ bool MotionInput::IsCalibrated(f32 sensitivity) const { } void MotionInput::UpdateRotation(u64 elapsed_time) { - const f32 sample_period = elapsed_time / 1000000.0f; + const auto sample_period = static_cast<f32>(elapsed_time) / 1000000.0f; if (sample_period > 0.1f) { return; } @@ -66,9 +74,9 @@ void MotionInput::UpdateOrientation(u64 elapsed_time) { f32 q2 = quat.xyz[0]; f32 q3 = quat.xyz[1]; f32 q4 = quat.xyz[2]; - const f32 sample_period = elapsed_time / 1000000.0f; + const auto sample_period = static_cast<f32>(elapsed_time) / 1000000.0f; - // ignore invalid elapsed time + // Ignore invalid elapsed time if (sample_period > 0.1f) { return; } @@ -80,6 +88,13 @@ void MotionInput::UpdateOrientation(u64 elapsed_time) { rad_gyro.y = -swap; rad_gyro.z = -rad_gyro.z; + // Clear gyro values if there is no gyro present + if (only_accelerometer) { + rad_gyro.x = 0; + rad_gyro.y = 0; + rad_gyro.z = 0; + } + // Ignore drift correction if acceleration is not reliable if (accel.Length() >= 0.75f && accel.Length() <= 1.25f) { const f32 ax = -normal_accel.x; @@ -92,8 +107,11 @@ void MotionInput::UpdateOrientation(u64 elapsed_time) { const f32 vz = q1 * q1 - q2 * q2 - q3 * q3 + q4 * q4; // Error is cross product between estimated direction and measured direction of gravity - const Common::Vec3f new_real_error = {az * vx - ax * vz, ay * vz - az * vy, - ax * vy - ay * vx}; + const Common::Vec3f new_real_error = { + az * vx - ax * vz, + ay * vz - az * vy, + ax * vy - ay * vx, + }; derivative_error = new_real_error - real_error; real_error = new_real_error; @@ -106,9 +124,22 @@ void MotionInput::UpdateOrientation(u64 elapsed_time) { } // Apply feedback terms - rad_gyro += kp * real_error; - rad_gyro += ki * integral_error; - rad_gyro += kd * derivative_error; + if (!only_accelerometer) { + rad_gyro += kp * real_error; + rad_gyro += ki * integral_error; + rad_gyro += kd * derivative_error; + } else { + // Give more weight to accelerometer values to compensate for the lack of gyro + rad_gyro += 35.0f * kp * real_error; + rad_gyro += 10.0f * ki * integral_error; + rad_gyro += 10.0f * kd * derivative_error; + + // Emulate gyro values for games that need them + gyro.x = -rad_gyro.y; + gyro.y = rad_gyro.x; + gyro.z = -rad_gyro.z; + UpdateRotation(elapsed_time); + } } const f32 gx = rad_gyro.y; @@ -159,18 +190,49 @@ Common::Vec3f MotionInput::GetRotations() const { return rotations; } +Input::MotionStatus MotionInput::GetMotion() const { + const Common::Vec3f gyroscope = GetGyroscope(); + const Common::Vec3f accelerometer = GetAcceleration(); + const Common::Vec3f rotation = GetRotations(); + const std::array<Common::Vec3f, 3> orientation = GetOrientation(); + return {accelerometer, gyroscope, rotation, orientation}; +} + +Input::MotionStatus MotionInput::GetRandomMotion(int accel_magnitude, int gyro_magnitude) const { + std::random_device device; + std::mt19937 gen(device()); + std::uniform_int_distribution<s16> distribution(-1000, 1000); + const Common::Vec3f gyroscope{ + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + }; + const Common::Vec3f accelerometer{ + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + static_cast<f32>(distribution(gen)) * 0.001f, + }; + constexpr Common::Vec3f rotation; + constexpr std::array orientation{ + Common::Vec3f{1.0f, 0.0f, 0.0f}, + Common::Vec3f{0.0f, 1.0f, 0.0f}, + Common::Vec3f{0.0f, 0.0f, 1.0f}, + }; + return {accelerometer * accel_magnitude, gyroscope * gyro_magnitude, rotation, orientation}; +} + void MotionInput::ResetOrientation() { - if (!reset_enabled) { + if (!reset_enabled || only_accelerometer) { return; } if (!IsMoving(0.5f) && accel.z <= -0.9f) { ++reset_counter; if (reset_counter > 900) { - // TODO: calculate quaternion from gravity vector quat.w = 0; quat.xyz[0] = 0; quat.xyz[1] = 0; quat.xyz[2] = -1; + SetOrientationFromAccelerometer(); integral_error = {}; reset_counter = 0; } @@ -178,4 +240,62 @@ void MotionInput::ResetOrientation() { reset_counter = 0; } } + +void MotionInput::SetOrientationFromAccelerometer() { + int iterations = 0; + const f32 sample_period = 0.015f; + + const auto normal_accel = accel.Normalized(); + + while (!IsCalibrated(0.01f) && ++iterations < 100) { + // Short name local variable for readability + f32 q1 = quat.w; + f32 q2 = quat.xyz[0]; + f32 q3 = quat.xyz[1]; + f32 q4 = quat.xyz[2]; + + Common::Vec3f rad_gyro; + const f32 ax = -normal_accel.x; + const f32 ay = normal_accel.y; + const f32 az = -normal_accel.z; + + // Estimated direction of gravity + const f32 vx = 2.0f * (q2 * q4 - q1 * q3); + const f32 vy = 2.0f * (q1 * q2 + q3 * q4); + const f32 vz = q1 * q1 - q2 * q2 - q3 * q3 + q4 * q4; + + // Error is cross product between estimated direction and measured direction of gravity + const Common::Vec3f new_real_error = { + az * vx - ax * vz, + ay * vz - az * vy, + ax * vy - ay * vx, + }; + + derivative_error = new_real_error - real_error; + real_error = new_real_error; + + rad_gyro += 10.0f * kp * real_error; + rad_gyro += 5.0f * ki * integral_error; + rad_gyro += 10.0f * kd * derivative_error; + + const f32 gx = rad_gyro.y; + const f32 gy = rad_gyro.x; + const f32 gz = rad_gyro.z; + + // Integrate rate of change of quaternion + const f32 pa = q2; + const f32 pb = q3; + const f32 pc = q4; + q1 = q1 + (-q2 * gx - q3 * gy - q4 * gz) * (0.5f * sample_period); + q2 = pa + (q1 * gx + pb * gz - pc * gy) * (0.5f * sample_period); + q3 = pb + (q1 * gy - pa * gz + pc * gx) * (0.5f * sample_period); + q4 = pc + (q1 * gz + pa * gy - pb * gx) * (0.5f * sample_period); + + quat.w = q1; + quat.xyz[0] = q2; + quat.xyz[1] = q3; + quat.xyz[2] = q4; + quat = quat.Normalized(); + } +} } // namespace InputCommon diff --git a/src/input_common/motion_input.h b/src/input_common/motion_input.h index 54b4439d9..efe74cf19 100644 --- a/src/input_common/motion_input.h +++ b/src/input_common/motion_input.h @@ -7,12 +7,13 @@ #include "common/common_types.h" #include "common/quaternion.h" #include "common/vector_math.h" +#include "core/frontend/input.h" namespace InputCommon { class MotionInput { public: - MotionInput(f32 new_kp, f32 new_ki, f32 new_kd); + explicit MotionInput(f32 new_kp, f32 new_ki, f32 new_kd); MotionInput(const MotionInput&) = default; MotionInput& operator=(const MotionInput&) = default; @@ -21,7 +22,7 @@ public: MotionInput& operator=(MotionInput&&) = default; void SetAcceleration(const Common::Vec3f& acceleration); - void SetGyroscope(const Common::Vec3f& acceleration); + void SetGyroscope(const Common::Vec3f& gyroscope); void SetQuaternion(const Common::Quaternion<f32>& quaternion); void SetGyroDrift(const Common::Vec3f& drift); void SetGyroThreshold(f32 threshold); @@ -32,29 +33,33 @@ public: void UpdateRotation(u64 elapsed_time); void UpdateOrientation(u64 elapsed_time); - std::array<Common::Vec3f, 3> GetOrientation() const; - Common::Vec3f GetAcceleration() const; - Common::Vec3f GetGyroscope() const; - Common::Vec3f GetRotations() const; - Common::Quaternion<f32> GetQuaternion() const; + [[nodiscard]] std::array<Common::Vec3f, 3> GetOrientation() const; + [[nodiscard]] Common::Vec3f GetAcceleration() const; + [[nodiscard]] Common::Vec3f GetGyroscope() const; + [[nodiscard]] Common::Vec3f GetRotations() const; + [[nodiscard]] Common::Quaternion<f32> GetQuaternion() const; + [[nodiscard]] Input::MotionStatus GetMotion() const; + [[nodiscard]] Input::MotionStatus GetRandomMotion(int accel_magnitude, + int gyro_magnitude) const; - bool IsMoving(f32 sensitivity) const; - bool IsCalibrated(f32 sensitivity) const; + [[nodiscard]] bool IsMoving(f32 sensitivity) const; + [[nodiscard]] bool IsCalibrated(f32 sensitivity) const; private: void ResetOrientation(); + void SetOrientationFromAccelerometer(); // PID constants - const f32 kp; - const f32 ki; - const f32 kd; + f32 kp; + f32 ki; + f32 kd; // PID errors Common::Vec3f real_error; Common::Vec3f integral_error; Common::Vec3f derivative_error; - Common::Quaternion<f32> quat; + Common::Quaternion<f32> quat{{0.0f, 0.0f, -1.0f}, 0.0f}; Common::Vec3f rotations; Common::Vec3f accel; Common::Vec3f gyro; @@ -63,6 +68,7 @@ private: f32 gyro_threshold = 0.0f; u32 reset_counter = 0; bool reset_enabled = true; + bool only_accelerometer = true; }; } // namespace InputCommon diff --git a/src/input_common/mouse/mouse_input.cpp b/src/input_common/mouse/mouse_input.cpp new file mode 100644 index 000000000..10786a541 --- /dev/null +++ b/src/input_common/mouse/mouse_input.cpp @@ -0,0 +1,129 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2+ +// Refer to the license.txt file included. + +#include "input_common/mouse/mouse_input.h" + +namespace MouseInput { + +Mouse::Mouse() { + update_thread = std::thread(&Mouse::UpdateThread, this); +} + +Mouse::~Mouse() { + update_thread_running = false; + if (update_thread.joinable()) { + update_thread.join(); + } +} + +void Mouse::UpdateThread() { + constexpr int update_time = 10; + while (update_thread_running) { + for (MouseInfo& info : mouse_info) { + const Common::Vec3f angular_direction{ + -info.tilt_direction.y, + 0.0f, + -info.tilt_direction.x, + }; + + info.motion.SetGyroscope(angular_direction * info.tilt_speed); + info.motion.UpdateRotation(update_time * 1000); + info.motion.UpdateOrientation(update_time * 1000); + info.tilt_speed = 0; + info.data.motion = info.motion.GetMotion(); + } + if (configuring) { + UpdateYuzuSettings(); + } + std::this_thread::sleep_for(std::chrono::milliseconds(update_time)); + } +} + +void Mouse::UpdateYuzuSettings() { + if (buttons == 0) { + return; + } + + mouse_queue.Push(MouseStatus{ + .button = last_button, + }); +} + +void Mouse::PressButton(int x, int y, int button_) { + const auto button_index = static_cast<std::size_t>(button_); + if (button_index >= mouse_info.size()) { + return; + } + + const auto button = 1U << button_index; + buttons |= static_cast<u16>(button); + last_button = static_cast<MouseButton>(button_index); + + mouse_info[button_index].mouse_origin = Common::MakeVec(x, y); + mouse_info[button_index].last_mouse_position = Common::MakeVec(x, y); + mouse_info[button_index].data.pressed = true; +} + +void Mouse::MouseMove(int x, int y) { + for (MouseInfo& info : mouse_info) { + if (info.data.pressed) { + const auto mouse_move = Common::MakeVec(x, y) - info.mouse_origin; + const auto mouse_change = Common::MakeVec(x, y) - info.last_mouse_position; + info.last_mouse_position = Common::MakeVec(x, y); + info.data.axis = {mouse_move.x, -mouse_move.y}; + + if (mouse_change.x == 0 && mouse_change.y == 0) { + info.tilt_speed = 0; + } else { + info.tilt_direction = mouse_change.Cast<float>(); + info.tilt_speed = info.tilt_direction.Normalize() * info.sensitivity; + } + } + } +} + +void Mouse::ReleaseButton(int button_) { + const auto button_index = static_cast<std::size_t>(button_); + if (button_index >= mouse_info.size()) { + return; + } + + const auto button = 1U << button_index; + buttons &= static_cast<u16>(0xFF - button); + + mouse_info[button_index].tilt_speed = 0; + mouse_info[button_index].data.pressed = false; + mouse_info[button_index].data.axis = {0, 0}; +} + +void Mouse::BeginConfiguration() { + buttons = 0; + last_button = MouseButton::Undefined; + mouse_queue.Clear(); + configuring = true; +} + +void Mouse::EndConfiguration() { + buttons = 0; + last_button = MouseButton::Undefined; + mouse_queue.Clear(); + configuring = false; +} + +Common::SPSCQueue<MouseStatus>& Mouse::GetMouseQueue() { + return mouse_queue; +} + +const Common::SPSCQueue<MouseStatus>& Mouse::GetMouseQueue() const { + return mouse_queue; +} + +MouseData& Mouse::GetMouseState(std::size_t button) { + return mouse_info[button].data; +} + +const MouseData& Mouse::GetMouseState(std::size_t button) const { + return mouse_info[button].data; +} +} // namespace MouseInput diff --git a/src/input_common/mouse/mouse_input.h b/src/input_common/mouse/mouse_input.h new file mode 100644 index 000000000..58803c1bf --- /dev/null +++ b/src/input_common/mouse/mouse_input.h @@ -0,0 +1,98 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <mutex> +#include <thread> + +#include "common/common_types.h" +#include "common/threadsafe_queue.h" +#include "common/vector_math.h" +#include "core/frontend/input.h" +#include "input_common/motion_input.h" + +namespace MouseInput { + +enum class MouseButton { + Left, + Wheel, + Right, + Forward, + Backward, + Undefined, +}; + +struct MouseStatus { + MouseButton button{MouseButton::Undefined}; +}; + +struct MouseData { + bool pressed{}; + std::array<int, 2> axis{}; + Input::MotionStatus motion{}; + Input::TouchStatus touch{}; +}; + +class Mouse { +public: + Mouse(); + ~Mouse(); + + /// Used for polling + void BeginConfiguration(); + void EndConfiguration(); + + /** + * Signals that a button is pressed. + * @param x the x-coordinate of the cursor + * @param y the y-coordinate of the cursor + * @param button_ the button pressed + */ + void PressButton(int x, int y, int button_); + + /** + * Signals that mouse has moved. + * @param x the x-coordinate of the cursor + * @param y the y-coordinate of the cursor + */ + void MouseMove(int x, int y); + + /** + * Signals that a motion sensor tilt has ended. + */ + void ReleaseButton(int button_); + + [[nodiscard]] Common::SPSCQueue<MouseStatus>& GetMouseQueue(); + [[nodiscard]] const Common::SPSCQueue<MouseStatus>& GetMouseQueue() const; + + [[nodiscard]] MouseData& GetMouseState(std::size_t button); + [[nodiscard]] const MouseData& GetMouseState(std::size_t button) const; + +private: + void UpdateThread(); + void UpdateYuzuSettings(); + + struct MouseInfo { + InputCommon::MotionInput motion{0.0f, 0.0f, 0.0f}; + Common::Vec2<int> mouse_origin; + Common::Vec2<int> last_mouse_position; + bool is_tilting = false; + float sensitivity{0.120f}; + + float tilt_speed = 0; + Common::Vec2<float> tilt_direction; + MouseData data; + }; + + u16 buttons{}; + std::thread update_thread; + MouseButton last_button{MouseButton::Undefined}; + std::array<MouseInfo, 5> mouse_info; + Common::SPSCQueue<MouseStatus> mouse_queue; + bool configuring{false}; + bool update_thread_running{true}; +}; +} // namespace MouseInput diff --git a/src/input_common/mouse/mouse_poller.cpp b/src/input_common/mouse/mouse_poller.cpp new file mode 100644 index 000000000..508eb0c7d --- /dev/null +++ b/src/input_common/mouse/mouse_poller.cpp @@ -0,0 +1,274 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <mutex> +#include <utility> + +#include "common/threadsafe_queue.h" +#include "input_common/mouse/mouse_input.h" +#include "input_common/mouse/mouse_poller.h" + +namespace InputCommon { + +class MouseButton final : public Input::ButtonDevice { +public: + explicit MouseButton(u32 button_, const MouseInput::Mouse* mouse_input_) + : button(button_), mouse_input(mouse_input_) {} + + bool GetStatus() const override { + return mouse_input->GetMouseState(button).pressed; + } + +private: + const u32 button; + const MouseInput::Mouse* mouse_input; +}; + +MouseButtonFactory::MouseButtonFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_) + : mouse_input(std::move(mouse_input_)) {} + +std::unique_ptr<Input::ButtonDevice> MouseButtonFactory::Create( + const Common::ParamPackage& params) { + const auto button_id = params.Get("button", 0); + + return std::make_unique<MouseButton>(button_id, mouse_input.get()); +} + +Common::ParamPackage MouseButtonFactory::GetNextInput() const { + MouseInput::MouseStatus pad; + Common::ParamPackage params; + auto& queue = mouse_input->GetMouseQueue(); + while (queue.Pop(pad)) { + // This while loop will break on the earliest detected button + if (pad.button != MouseInput::MouseButton::Undefined) { + params.Set("engine", "mouse"); + params.Set("button", static_cast<u16>(pad.button)); + return params; + } + } + return params; +} + +void MouseButtonFactory::BeginConfiguration() { + polling = true; + mouse_input->BeginConfiguration(); +} + +void MouseButtonFactory::EndConfiguration() { + polling = false; + mouse_input->EndConfiguration(); +} + +class MouseAnalog final : public Input::AnalogDevice { +public: + explicit MouseAnalog(u32 port_, u32 axis_x_, u32 axis_y_, bool invert_x_, bool invert_y_, + float deadzone_, float range_, const MouseInput::Mouse* mouse_input_) + : button(port_), axis_x(axis_x_), axis_y(axis_y_), invert_x(invert_x_), invert_y(invert_y_), + deadzone(deadzone_), range(range_), mouse_input(mouse_input_) {} + + float GetAxis(u32 axis) const { + std::lock_guard lock{mutex}; + const auto axis_value = + static_cast<float>(mouse_input->GetMouseState(button).axis.at(axis)); + return axis_value / (100.0f * range); + } + + std::pair<float, float> GetAnalog(u32 analog_axis_x, u32 analog_axis_y) const { + float x = GetAxis(analog_axis_x); + float y = GetAxis(analog_axis_y); + if (invert_x) { + x = -x; + } + if (invert_y) { + y = -y; + } + + // Make sure the coordinates are in the unit circle, + // otherwise normalize it. + float r = x * x + y * y; + if (r > 1.0f) { + r = std::sqrt(r); + x /= r; + y /= r; + } + + return {x, y}; + } + + std::tuple<float, float> GetStatus() const override { + const auto [x, y] = GetAnalog(axis_x, axis_y); + const float r = std::sqrt((x * x) + (y * y)); + if (r > deadzone) { + return {x / r * (r - deadzone) / (1 - deadzone), + y / r * (r - deadzone) / (1 - deadzone)}; + } + return {0.0f, 0.0f}; + } + +private: + const u32 button; + const u32 axis_x; + const u32 axis_y; + const bool invert_x; + const bool invert_y; + const float deadzone; + const float range; + const MouseInput::Mouse* mouse_input; + mutable std::mutex mutex; +}; + +/// An analog device factory that creates analog devices from GC Adapter +MouseAnalogFactory::MouseAnalogFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_) + : mouse_input(std::move(mouse_input_)) {} + +/** + * Creates analog device from joystick axes + * @param params contains parameters for creating the device: + * - "port": the nth gcpad on the adapter + * - "axis_x": the index of the axis to be bind as x-axis + * - "axis_y": the index of the axis to be bind as y-axis + */ +std::unique_ptr<Input::AnalogDevice> MouseAnalogFactory::Create( + const Common::ParamPackage& params) { + const auto port = static_cast<u32>(params.Get("port", 0)); + const auto axis_x = static_cast<u32>(params.Get("axis_x", 0)); + const auto axis_y = static_cast<u32>(params.Get("axis_y", 1)); + const auto deadzone = std::clamp(params.Get("deadzone", 0.0f), 0.0f, 1.0f); + const auto range = std::clamp(params.Get("range", 1.0f), 0.50f, 1.50f); + const std::string invert_x_value = params.Get("invert_x", "+"); + const std::string invert_y_value = params.Get("invert_y", "+"); + const bool invert_x = invert_x_value == "-"; + const bool invert_y = invert_y_value == "-"; + + return std::make_unique<MouseAnalog>(port, axis_x, axis_y, invert_x, invert_y, deadzone, range, + mouse_input.get()); +} + +void MouseAnalogFactory::BeginConfiguration() { + polling = true; + mouse_input->BeginConfiguration(); +} + +void MouseAnalogFactory::EndConfiguration() { + polling = false; + mouse_input->EndConfiguration(); +} + +Common::ParamPackage MouseAnalogFactory::GetNextInput() const { + MouseInput::MouseStatus pad; + Common::ParamPackage params; + auto& queue = mouse_input->GetMouseQueue(); + while (queue.Pop(pad)) { + // This while loop will break on the earliest detected button + if (pad.button != MouseInput::MouseButton::Undefined) { + params.Set("engine", "mouse"); + params.Set("port", static_cast<u16>(pad.button)); + params.Set("axis_x", 0); + params.Set("axis_y", 1); + params.Set("invert_x", "+"); + params.Set("invert_y", "+"); + return params; + } + } + return params; +} + +class MouseMotion final : public Input::MotionDevice { +public: + explicit MouseMotion(u32 button_, const MouseInput::Mouse* mouse_input_) + : button(button_), mouse_input(mouse_input_) {} + + Input::MotionStatus GetStatus() const override { + return mouse_input->GetMouseState(button).motion; + } + +private: + const u32 button; + const MouseInput::Mouse* mouse_input; +}; + +MouseMotionFactory::MouseMotionFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_) + : mouse_input(std::move(mouse_input_)) {} + +std::unique_ptr<Input::MotionDevice> MouseMotionFactory::Create( + const Common::ParamPackage& params) { + const auto button_id = params.Get("button", 0); + + return std::make_unique<MouseMotion>(button_id, mouse_input.get()); +} + +Common::ParamPackage MouseMotionFactory::GetNextInput() const { + MouseInput::MouseStatus pad; + Common::ParamPackage params; + auto& queue = mouse_input->GetMouseQueue(); + while (queue.Pop(pad)) { + // This while loop will break on the earliest detected button + if (pad.button != MouseInput::MouseButton::Undefined) { + params.Set("engine", "mouse"); + params.Set("button", static_cast<u16>(pad.button)); + return params; + } + } + return params; +} + +void MouseMotionFactory::BeginConfiguration() { + polling = true; + mouse_input->BeginConfiguration(); +} + +void MouseMotionFactory::EndConfiguration() { + polling = false; + mouse_input->EndConfiguration(); +} + +class MouseTouch final : public Input::TouchDevice { +public: + explicit MouseTouch(u32 button_, const MouseInput::Mouse* mouse_input_) + : button(button_), mouse_input(mouse_input_) {} + + Input::TouchStatus GetStatus() const override { + return mouse_input->GetMouseState(button).touch; + } + +private: + const u32 button; + const MouseInput::Mouse* mouse_input; +}; + +MouseTouchFactory::MouseTouchFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_) + : mouse_input(std::move(mouse_input_)) {} + +std::unique_ptr<Input::TouchDevice> MouseTouchFactory::Create(const Common::ParamPackage& params) { + const auto button_id = params.Get("button", 0); + + return std::make_unique<MouseTouch>(button_id, mouse_input.get()); +} + +Common::ParamPackage MouseTouchFactory::GetNextInput() const { + MouseInput::MouseStatus pad; + Common::ParamPackage params; + auto& queue = mouse_input->GetMouseQueue(); + while (queue.Pop(pad)) { + // This while loop will break on the earliest detected button + if (pad.button != MouseInput::MouseButton::Undefined) { + params.Set("engine", "mouse"); + params.Set("button", static_cast<u16>(pad.button)); + return params; + } + } + return params; +} + +void MouseTouchFactory::BeginConfiguration() { + polling = true; + mouse_input->BeginConfiguration(); +} + +void MouseTouchFactory::EndConfiguration() { + polling = false; + mouse_input->EndConfiguration(); +} + +} // namespace InputCommon diff --git a/src/input_common/mouse/mouse_poller.h b/src/input_common/mouse/mouse_poller.h new file mode 100644 index 000000000..cf331293b --- /dev/null +++ b/src/input_common/mouse/mouse_poller.h @@ -0,0 +1,109 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include "core/frontend/input.h" +#include "input_common/mouse/mouse_input.h" + +namespace InputCommon { + +/** + * A button device factory representing a mouse. It receives mouse events and forward them + * to all button devices it created. + */ +class MouseButtonFactory final : public Input::Factory<Input::ButtonDevice> { +public: + explicit MouseButtonFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_); + + /** + * Creates a button device from a button press + * @param params contains parameters for creating the device: + * - "code": the code of the key to bind with the button + */ + std::unique_ptr<Input::ButtonDevice> Create(const Common::ParamPackage& params) override; + + Common::ParamPackage GetNextInput() const; + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } + +private: + std::shared_ptr<MouseInput::Mouse> mouse_input; + bool polling = false; +}; + +/// An analog device factory that creates analog devices from mouse +class MouseAnalogFactory final : public Input::Factory<Input::AnalogDevice> { +public: + explicit MouseAnalogFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_); + + std::unique_ptr<Input::AnalogDevice> Create(const Common::ParamPackage& params) override; + + Common::ParamPackage GetNextInput() const; + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } + +private: + std::shared_ptr<MouseInput::Mouse> mouse_input; + bool polling = false; +}; + +/// A motion device factory that creates motion devices from mouse +class MouseMotionFactory final : public Input::Factory<Input::MotionDevice> { +public: + explicit MouseMotionFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_); + + std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override; + + Common::ParamPackage GetNextInput() const; + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } + +private: + std::shared_ptr<MouseInput::Mouse> mouse_input; + bool polling = false; +}; + +/// An touch device factory that creates touch devices from mouse +class MouseTouchFactory final : public Input::Factory<Input::TouchDevice> { +public: + explicit MouseTouchFactory(std::shared_ptr<MouseInput::Mouse> mouse_input_); + + std::unique_ptr<Input::TouchDevice> Create(const Common::ParamPackage& params) override; + + Common::ParamPackage GetNextInput() const; + + /// For device input configuration/polling + void BeginConfiguration(); + void EndConfiguration(); + + bool IsPolling() const { + return polling; + } + +private: + std::shared_ptr<MouseInput::Mouse> mouse_input; + bool polling = false; +}; + +} // namespace InputCommon diff --git a/src/input_common/sdl/sdl.h b/src/input_common/sdl/sdl.h index f3554be9a..42bbf14d4 100644 --- a/src/input_common/sdl/sdl.h +++ b/src/input_common/sdl/sdl.h @@ -23,7 +23,7 @@ public: /// Unregisters SDL device factories and shut them down. virtual ~State() = default; - virtual Pollers GetPollers(Polling::DeviceType type) { + virtual Pollers GetPollers(Polling::DeviceType) { return {}; } diff --git a/src/input_common/sdl/sdl_impl.cpp b/src/input_common/sdl/sdl_impl.cpp index a9e676f4b..d32eb732a 100644 --- a/src/input_common/sdl/sdl_impl.cpp +++ b/src/input_common/sdl/sdl_impl.cpp @@ -5,6 +5,7 @@ #include <algorithm> #include <array> #include <atomic> +#include <chrono> #include <cmath> #include <functional> #include <mutex> @@ -21,6 +22,7 @@ #include "common/param_package.h" #include "common/threadsafe_queue.h" #include "core/frontend/input.h" +#include "input_common/motion_input.h" #include "input_common/sdl/sdl_impl.h" #include "input_common/settings.h" @@ -54,9 +56,9 @@ static int SDLEventWatcher(void* user_data, SDL_Event* event) { class SDLJoystick { public: SDLJoystick(std::string guid_, int port_, SDL_Joystick* joystick, - SDL_GameController* gamecontroller) + SDL_GameController* game_controller) : guid{std::move(guid_)}, port{port_}, sdl_joystick{joystick, &SDL_JoystickClose}, - sdl_controller{gamecontroller, &SDL_GameControllerClose} {} + sdl_controller{game_controller, &SDL_GameControllerClose} {} void SetButton(int button, bool value) { std::lock_guard lock{mutex}; @@ -75,7 +77,17 @@ public: float GetAxis(int axis, float range) const { std::lock_guard lock{mutex}; - return state.axes.at(axis) / (32767.0f * range); + return static_cast<float>(state.axes.at(axis)) / (32767.0f * range); + } + + bool RumblePlay(u16 amp_low, u16 amp_high) { + if (sdl_controller) { + return SDL_GameControllerRumble(sdl_controller.get(), amp_low, amp_high, 0) == 0; + } else if (sdl_joystick) { + return SDL_JoystickRumble(sdl_joystick.get(), amp_low, amp_high, 0) == 0; + } + + return false; } std::tuple<float, float> GetAnalog(int axis_x, int axis_y, float range) const { @@ -95,6 +107,10 @@ public: return std::make_tuple(x, y); } + const MotionInput& GetMotion() const { + return motion; + } + void SetHat(int hat, Uint8 direction) { std::lock_guard lock{mutex}; state.hats.insert_or_assign(hat, direction); @@ -122,15 +138,15 @@ public: return sdl_joystick.get(); } - void SetSDLJoystick(SDL_Joystick* joystick, SDL_GameController* controller) { - sdl_controller.reset(controller); - sdl_joystick.reset(joystick); - } - SDL_GameController* GetSDLGameController() const { return sdl_controller.get(); } + void SetSDLJoystick(SDL_Joystick* joystick, SDL_GameController* controller) { + sdl_joystick.reset(joystick); + sdl_controller.reset(controller); + } + private: struct State { std::unordered_map<int, bool> buttons; @@ -142,74 +158,66 @@ private: std::unique_ptr<SDL_Joystick, decltype(&SDL_JoystickClose)> sdl_joystick; std::unique_ptr<SDL_GameController, decltype(&SDL_GameControllerClose)> sdl_controller; mutable std::mutex mutex; + + // Motion is initialized without PID values as motion input is not aviable for SDL2 + MotionInput motion{0.0f, 0.0f, 0.0f}; }; std::shared_ptr<SDLJoystick> SDLState::GetSDLJoystickByGUID(const std::string& guid, int port) { std::lock_guard lock{joystick_map_mutex}; const auto it = joystick_map.find(guid); + if (it != joystick_map.end()) { while (it->second.size() <= static_cast<std::size_t>(port)) { auto joystick = std::make_shared<SDLJoystick>(guid, static_cast<int>(it->second.size()), nullptr, nullptr); it->second.emplace_back(std::move(joystick)); } - return it->second[port]; + + return it->second[static_cast<std::size_t>(port)]; } + auto joystick = std::make_shared<SDLJoystick>(guid, 0, nullptr, nullptr); + return joystick_map[guid].emplace_back(std::move(joystick)); } std::shared_ptr<SDLJoystick> SDLState::GetSDLJoystickBySDLID(SDL_JoystickID sdl_id) { auto sdl_joystick = SDL_JoystickFromInstanceID(sdl_id); - auto sdl_controller = SDL_GameControllerFromInstanceID(sdl_id); const std::string guid = GetGUID(sdl_joystick); std::lock_guard lock{joystick_map_mutex}; const auto map_it = joystick_map.find(guid); - if (map_it != joystick_map.end()) { - const auto vec_it = - std::find_if(map_it->second.begin(), map_it->second.end(), - [&sdl_joystick](const std::shared_ptr<SDLJoystick>& joystick) { - return sdl_joystick == joystick->GetSDLJoystick(); - }); - if (vec_it != map_it->second.end()) { - // This is the common case: There is already an existing SDL_Joystick maped to a - // SDLJoystick. return the SDLJoystick - return *vec_it; - } - - // Search for a SDLJoystick without a mapped SDL_Joystick... - const auto nullptr_it = std::find_if(map_it->second.begin(), map_it->second.end(), - [](const std::shared_ptr<SDLJoystick>& joystick) { - return !joystick->GetSDLJoystick(); - }); - if (nullptr_it != map_it->second.end()) { - // ... and map it - (*nullptr_it)->SetSDLJoystick(sdl_joystick, sdl_controller); - return *nullptr_it; - } - - // There is no SDLJoystick without a mapped SDL_Joystick - // Create a new SDLJoystick - const int port = static_cast<int>(map_it->second.size()); - auto joystick = std::make_shared<SDLJoystick>(guid, port, sdl_joystick, sdl_controller); - return map_it->second.emplace_back(std::move(joystick)); - } - - auto joystick = std::make_shared<SDLJoystick>(guid, 0, sdl_joystick, sdl_controller); - return joystick_map[guid].emplace_back(std::move(joystick)); + + if (map_it == joystick_map.end()) { + return nullptr; + } + + const auto vec_it = std::find_if(map_it->second.begin(), map_it->second.end(), + [&sdl_joystick](const auto& joystick) { + return joystick->GetSDLJoystick() == sdl_joystick; + }); + + if (vec_it == map_it->second.end()) { + return nullptr; + } + + return *vec_it; } void SDLState::InitJoystick(int joystick_index) { SDL_Joystick* sdl_joystick = SDL_JoystickOpen(joystick_index); SDL_GameController* sdl_gamecontroller = nullptr; + if (SDL_IsGameController(joystick_index)) { sdl_gamecontroller = SDL_GameControllerOpen(joystick_index); } + if (!sdl_joystick) { - LOG_ERROR(Input, "failed to open joystick {}", joystick_index); + LOG_ERROR(Input, "Failed to open joystick {}", joystick_index); return; } + const std::string guid = GetGUID(sdl_joystick); std::lock_guard lock{joystick_map_mutex}; @@ -218,14 +226,17 @@ void SDLState::InitJoystick(int joystick_index) { joystick_map[guid].emplace_back(std::move(joystick)); return; } + auto& joystick_guid_list = joystick_map[guid]; - const auto it = std::find_if( - joystick_guid_list.begin(), joystick_guid_list.end(), - [](const std::shared_ptr<SDLJoystick>& joystick) { return !joystick->GetSDLJoystick(); }); - if (it != joystick_guid_list.end()) { - (*it)->SetSDLJoystick(sdl_joystick, sdl_gamecontroller); + const auto joystick_it = + std::find_if(joystick_guid_list.begin(), joystick_guid_list.end(), + [](const auto& joystick) { return !joystick->GetSDLJoystick(); }); + + if (joystick_it != joystick_guid_list.end()) { + (*joystick_it)->SetSDLJoystick(sdl_joystick, sdl_gamecontroller); return; } + const int port = static_cast<int>(joystick_guid_list.size()); auto joystick = std::make_shared<SDLJoystick>(guid, port, sdl_joystick, sdl_gamecontroller); joystick_guid_list.emplace_back(std::move(joystick)); @@ -234,22 +245,15 @@ void SDLState::InitJoystick(int joystick_index) { void SDLState::CloseJoystick(SDL_Joystick* sdl_joystick) { const std::string guid = GetGUID(sdl_joystick); - std::shared_ptr<SDLJoystick> joystick; - { - std::lock_guard lock{joystick_map_mutex}; - // This call to guid is safe since the joystick is guaranteed to be in the map - const auto& joystick_guid_list = joystick_map[guid]; - const auto joystick_it = - std::find_if(joystick_guid_list.begin(), joystick_guid_list.end(), - [&sdl_joystick](const std::shared_ptr<SDLJoystick>& joystick) { - return joystick->GetSDLJoystick() == sdl_joystick; - }); - joystick = *joystick_it; - } - - // Destruct SDL_Joystick outside the lock guard because SDL can internally call the - // event callback which locks the mutex again. - joystick->SetSDLJoystick(nullptr, nullptr); + std::lock_guard lock{joystick_map_mutex}; + // This call to guid is safe since the joystick is guaranteed to be in the map + const auto& joystick_guid_list = joystick_map[guid]; + const auto joystick_it = std::find_if(joystick_guid_list.begin(), joystick_guid_list.end(), + [&sdl_joystick](const auto& joystick) { + return joystick->GetSDLJoystick() == sdl_joystick; + }); + + (*joystick_it)->SetSDLJoystick(nullptr, nullptr); } void SDLState::HandleGameControllerEvent(const SDL_Event& event) { @@ -347,14 +351,21 @@ private: class SDLAnalog final : public Input::AnalogDevice { public: - SDLAnalog(std::shared_ptr<SDLJoystick> joystick_, int axis_x_, int axis_y_, float deadzone_, - float range_) - : joystick(std::move(joystick_)), axis_x(axis_x_), axis_y(axis_y_), deadzone(deadzone_), - range(range_) {} + explicit SDLAnalog(std::shared_ptr<SDLJoystick> joystick_, int axis_x_, int axis_y_, + bool invert_x_, bool invert_y_, float deadzone_, float range_) + : joystick(std::move(joystick_)), axis_x(axis_x_), axis_y(axis_y_), invert_x(invert_x_), + invert_y(invert_y_), deadzone(deadzone_), range(range_) {} std::tuple<float, float> GetStatus() const override { - const auto [x, y] = joystick->GetAnalog(axis_x, axis_y, range); + auto [x, y] = joystick->GetAnalog(axis_x, axis_y, range); const float r = std::sqrt((x * x) + (y * y)); + if (invert_x) { + x = -x; + } + if (invert_y) { + y = -y; + } + if (r > deadzone) { return std::make_tuple(x / r * (r - deadzone) / (1 - deadzone), y / r * (r - deadzone) / (1 - deadzone)); @@ -382,10 +393,100 @@ private: std::shared_ptr<SDLJoystick> joystick; const int axis_x; const int axis_y; + const bool invert_x; + const bool invert_y; const float deadzone; const float range; }; +class SDLVibration final : public Input::VibrationDevice { +public: + explicit SDLVibration(std::shared_ptr<SDLJoystick> joystick_) + : joystick(std::move(joystick_)) {} + + u8 GetStatus() const override { + joystick->RumblePlay(1, 1); + return joystick->RumblePlay(0, 0); + } + + bool SetRumblePlay(f32 amp_low, [[maybe_unused]] f32 freq_low, f32 amp_high, + [[maybe_unused]] f32 freq_high) const override { + const auto process_amplitude = [](f32 amplitude) { + return static_cast<u16>((amplitude + std::pow(amplitude, 0.3f)) * 0.5f * 0xFFFF); + }; + + const auto processed_amp_low = process_amplitude(amp_low); + const auto processed_amp_high = process_amplitude(amp_high); + + return joystick->RumblePlay(processed_amp_low, processed_amp_high); + } + +private: + std::shared_ptr<SDLJoystick> joystick; +}; + +class SDLDirectionMotion final : public Input::MotionDevice { +public: + explicit SDLDirectionMotion(std::shared_ptr<SDLJoystick> joystick_, int hat_, Uint8 direction_) + : joystick(std::move(joystick_)), hat(hat_), direction(direction_) {} + + Input::MotionStatus GetStatus() const override { + if (joystick->GetHatDirection(hat, direction)) { + return joystick->GetMotion().GetRandomMotion(2, 6); + } + return joystick->GetMotion().GetRandomMotion(0, 0); + } + +private: + std::shared_ptr<SDLJoystick> joystick; + int hat; + Uint8 direction; +}; + +class SDLAxisMotion final : public Input::MotionDevice { +public: + explicit SDLAxisMotion(std::shared_ptr<SDLJoystick> joystick_, int axis_, float threshold_, + bool trigger_if_greater_) + : joystick(std::move(joystick_)), axis(axis_), threshold(threshold_), + trigger_if_greater(trigger_if_greater_) {} + + Input::MotionStatus GetStatus() const override { + const float axis_value = joystick->GetAxis(axis, 1.0f); + bool trigger = axis_value < threshold; + if (trigger_if_greater) { + trigger = axis_value > threshold; + } + + if (trigger) { + return joystick->GetMotion().GetRandomMotion(2, 6); + } + return joystick->GetMotion().GetRandomMotion(0, 0); + } + +private: + std::shared_ptr<SDLJoystick> joystick; + int axis; + float threshold; + bool trigger_if_greater; +}; + +class SDLButtonMotion final : public Input::MotionDevice { +public: + explicit SDLButtonMotion(std::shared_ptr<SDLJoystick> joystick_, int button_) + : joystick(std::move(joystick_)), button(button_) {} + + Input::MotionStatus GetStatus() const override { + if (joystick->GetButton(button)) { + return joystick->GetMotion().GetRandomMotion(2, 6); + } + return joystick->GetMotion().GetRandomMotion(0, 0); + } + +private: + std::shared_ptr<SDLJoystick> joystick; + int button; +}; + /// A button device factory that creates button devices from SDL joystick class SDLButtonFactory final : public Input::Factory<Input::ButtonDevice> { public: @@ -466,7 +567,7 @@ class SDLAnalogFactory final : public Input::Factory<Input::AnalogDevice> { public: explicit SDLAnalogFactory(SDLState& state_) : state(state_) {} /** - * Creates analog device from joystick axes + * Creates an analog device from joystick axes * @param params contains parameters for creating the device: * - "guid": the guid of the joystick to bind * - "port": the nth joystick of the same type @@ -480,12 +581,101 @@ public: const int axis_y = params.Get("axis_y", 1); const float deadzone = std::clamp(params.Get("deadzone", 0.0f), 0.0f, 1.0f); const float range = std::clamp(params.Get("range", 1.0f), 0.50f, 1.50f); + const std::string invert_x_value = params.Get("invert_x", "+"); + const std::string invert_y_value = params.Get("invert_y", "+"); + const bool invert_x = invert_x_value == "-"; + const bool invert_y = invert_y_value == "-"; auto joystick = state.GetSDLJoystickByGUID(guid, port); // This is necessary so accessing GetAxis with axis_x and axis_y won't crash joystick->SetAxis(axis_x, 0); joystick->SetAxis(axis_y, 0); - return std::make_unique<SDLAnalog>(joystick, axis_x, axis_y, deadzone, range); + return std::make_unique<SDLAnalog>(joystick, axis_x, axis_y, invert_x, invert_y, deadzone, + range); + } + +private: + SDLState& state; +}; + +/// An vibration device factory that creates vibration devices from SDL joystick +class SDLVibrationFactory final : public Input::Factory<Input::VibrationDevice> { +public: + explicit SDLVibrationFactory(SDLState& state_) : state(state_) {} + /** + * Creates a vibration device from a joystick + * @param params contains parameters for creating the device: + * - "guid": the guid of the joystick to bind + * - "port": the nth joystick of the same type + */ + std::unique_ptr<Input::VibrationDevice> Create(const Common::ParamPackage& params) override { + const std::string guid = params.Get("guid", "0"); + const int port = params.Get("port", 0); + return std::make_unique<SDLVibration>(state.GetSDLJoystickByGUID(guid, port)); + } + +private: + SDLState& state; +}; + +/// A motion device factory that creates motion devices from SDL joystick +class SDLMotionFactory final : public Input::Factory<Input::MotionDevice> { +public: + explicit SDLMotionFactory(SDLState& state_) : state(state_) {} + /** + * Creates motion device from joystick axes + * @param params contains parameters for creating the device: + * - "guid": the guid of the joystick to bind + * - "port": the nth joystick of the same type + */ + std::unique_ptr<Input::MotionDevice> Create(const Common::ParamPackage& params) override { + const std::string guid = params.Get("guid", "0"); + const int port = params.Get("port", 0); + + auto joystick = state.GetSDLJoystickByGUID(guid, port); + + if (params.Has("hat")) { + const int hat = params.Get("hat", 0); + const std::string direction_name = params.Get("direction", ""); + Uint8 direction; + if (direction_name == "up") { + direction = SDL_HAT_UP; + } else if (direction_name == "down") { + direction = SDL_HAT_DOWN; + } else if (direction_name == "left") { + direction = SDL_HAT_LEFT; + } else if (direction_name == "right") { + direction = SDL_HAT_RIGHT; + } else { + direction = 0; + } + // This is necessary so accessing GetHat with hat won't crash + joystick->SetHat(hat, SDL_HAT_CENTERED); + return std::make_unique<SDLDirectionMotion>(joystick, hat, direction); + } + + if (params.Has("axis")) { + const int axis = params.Get("axis", 0); + const float threshold = params.Get("threshold", 0.5f); + const std::string direction_name = params.Get("direction", ""); + bool trigger_if_greater; + if (direction_name == "+") { + trigger_if_greater = true; + } else if (direction_name == "-") { + trigger_if_greater = false; + } else { + trigger_if_greater = true; + LOG_ERROR(Input, "Unknown direction {}", direction_name); + } + // This is necessary so accessing GetAxis with axis won't crash + joystick->SetAxis(axis, 0); + return std::make_unique<SDLAxisMotion>(joystick, axis, threshold, trigger_if_greater); + } + + const int button = params.Get("button", 0); + // This is necessary so accessing GetButton with button won't crash + joystick->SetButton(button, false); + return std::make_unique<SDLButtonMotion>(joystick, button); } private: @@ -494,18 +684,22 @@ private: SDLState::SDLState() { using namespace Input; - analog_factory = std::make_shared<SDLAnalogFactory>(*this); button_factory = std::make_shared<SDLButtonFactory>(*this); - RegisterFactory<AnalogDevice>("sdl", analog_factory); + analog_factory = std::make_shared<SDLAnalogFactory>(*this); + vibration_factory = std::make_shared<SDLVibrationFactory>(*this); + motion_factory = std::make_shared<SDLMotionFactory>(*this); RegisterFactory<ButtonDevice>("sdl", button_factory); + RegisterFactory<AnalogDevice>("sdl", analog_factory); + RegisterFactory<VibrationDevice>("sdl", vibration_factory); + RegisterFactory<MotionDevice>("sdl", motion_factory); - // If the frontend is going to manage the event loop, then we dont start one here - start_thread = !SDL_WasInit(SDL_INIT_JOYSTICK); + // If the frontend is going to manage the event loop, then we don't start one here + start_thread = SDL_WasInit(SDL_INIT_JOYSTICK) == 0; if (start_thread && SDL_Init(SDL_INIT_JOYSTICK) < 0) { LOG_CRITICAL(Input, "SDL_Init(SDL_INIT_JOYSTICK) failed with: {}", SDL_GetError()); return; } - has_gamecontroller = SDL_InitSubSystem(SDL_INIT_GAMECONTROLLER); + has_gamecontroller = SDL_InitSubSystem(SDL_INIT_GAMECONTROLLER) != 0; if (SDL_SetHint(SDL_HINT_JOYSTICK_ALLOW_BACKGROUND_EVENTS, "1") == SDL_FALSE) { LOG_ERROR(Input, "Failed to set hint for background events with: {}", SDL_GetError()); } @@ -518,7 +712,7 @@ SDLState::SDLState() { using namespace std::chrono_literals; while (initialized) { SDL_PumpEvents(); - std::this_thread::sleep_for(5ms); + std::this_thread::sleep_for(1ms); } }); } @@ -533,6 +727,8 @@ SDLState::~SDLState() { using namespace Input; UnregisterFactory<ButtonDevice>("sdl"); UnregisterFactory<AnalogDevice>("sdl"); + UnregisterFactory<VibrationDevice>("sdl"); + UnregisterFactory<MotionDevice>("sdl"); CloseJoysticks(); SDL_DelEventWatch(&SDLEventWatcher, this); @@ -549,8 +745,7 @@ std::vector<Common::ParamPackage> SDLState::GetInputDevices() { std::vector<Common::ParamPackage> devices; for (const auto& [key, value] : joystick_map) { for (const auto& joystick : value) { - auto joy = joystick->GetSDLJoystick(); - if (auto controller = joystick->GetSDLGameController()) { + if (auto* const controller = joystick->GetSDLGameController()) { std::string name = fmt::format("{} {}", SDL_GameControllerName(controller), joystick->GetPort()); devices.emplace_back(Common::ParamPackage{ @@ -559,7 +754,7 @@ std::vector<Common::ParamPackage> SDLState::GetInputDevices() { {"guid", joystick->GetGUID()}, {"port", std::to_string(joystick->GetPort())}, }); - } else if (joy) { + } else if (auto* const joy = joystick->GetSDLJoystick()) { std::string name = fmt::format("{} {}", SDL_JoystickName(joy), joystick->GetPort()); devices.emplace_back(Common::ParamPackage{ {"class", "sdl"}, @@ -574,7 +769,7 @@ std::vector<Common::ParamPackage> SDLState::GetInputDevices() { } namespace { -Common::ParamPackage BuildAnalogParamPackageForButton(int port, std::string guid, u8 axis, +Common::ParamPackage BuildAnalogParamPackageForButton(int port, std::string guid, s32 axis, float value = 0.1f) { Common::ParamPackage params({{"engine", "sdl"}}); params.Set("port", port); @@ -590,7 +785,7 @@ Common::ParamPackage BuildAnalogParamPackageForButton(int port, std::string guid return params; } -Common::ParamPackage BuildButtonParamPackageForButton(int port, std::string guid, u8 button) { +Common::ParamPackage BuildButtonParamPackageForButton(int port, std::string guid, s32 button) { Common::ParamPackage params({{"engine", "sdl"}}); params.Set("port", port); params.Set("guid", std::move(guid)); @@ -598,7 +793,7 @@ Common::ParamPackage BuildButtonParamPackageForButton(int port, std::string guid return params; } -Common::ParamPackage BuildHatParamPackageForButton(int port, std::string guid, u8 hat, u8 value) { +Common::ParamPackage BuildHatParamPackageForButton(int port, std::string guid, s32 hat, s32 value) { Common::ParamPackage params({{"engine", "sdl"}}); params.Set("port", port); @@ -626,19 +821,56 @@ Common::ParamPackage BuildHatParamPackageForButton(int port, std::string guid, u Common::ParamPackage SDLEventToButtonParamPackage(SDLState& state, const SDL_Event& event) { switch (event.type) { case SDL_JOYAXISMOTION: { - const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which); - return BuildAnalogParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), - event.jaxis.axis, event.jaxis.value); + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which)) { + return BuildAnalogParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jaxis.axis), + event.jaxis.value); + } + break; + } + case SDL_JOYBUTTONUP: { + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jbutton.which)) { + return BuildButtonParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jbutton.button)); + } + break; + } + case SDL_JOYHATMOTION: { + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jhat.which)) { + return BuildHatParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jhat.hat), + static_cast<s32>(event.jhat.value)); + } + break; + } + } + return {}; +} + +Common::ParamPackage SDLEventToMotionParamPackage(SDLState& state, const SDL_Event& event) { + switch (event.type) { + case SDL_JOYAXISMOTION: { + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which)) { + return BuildAnalogParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jaxis.axis), + event.jaxis.value); + } + break; } case SDL_JOYBUTTONUP: { - const auto joystick = state.GetSDLJoystickBySDLID(event.jbutton.which); - return BuildButtonParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), - event.jbutton.button); + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jbutton.which)) { + return BuildButtonParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jbutton.button)); + } + break; } case SDL_JOYHATMOTION: { - const auto joystick = state.GetSDLJoystickBySDLID(event.jhat.which); - return BuildHatParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), - event.jhat.hat, event.jhat.value); + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jhat.which)) { + return BuildHatParamPackageForButton(joystick->GetPort(), joystick->GetGUID(), + static_cast<s32>(event.jhat.hat), + static_cast<s32>(event.jhat.value)); + } + break; } } return {}; @@ -647,6 +879,8 @@ Common::ParamPackage SDLEventToButtonParamPackage(SDLState& state, const SDL_Eve Common::ParamPackage BuildParamPackageForBinding(int port, const std::string& guid, const SDL_GameControllerButtonBind& binding) { switch (binding.bindType) { + case SDL_CONTROLLER_BINDTYPE_NONE: + break; case SDL_CONTROLLER_BINDTYPE_AXIS: return BuildAnalogParamPackageForButton(port, guid, binding.value.axis); case SDL_CONTROLLER_BINDTYPE_BUTTON: @@ -666,6 +900,8 @@ Common::ParamPackage BuildParamPackageForAnalog(int port, const std::string& gui params.Set("guid", guid); params.Set("axis_x", axis_x); params.Set("axis_y", axis_y); + params.Set("invert_x", "+"); + params.Set("invert_y", "+"); return params; } } // Anonymous namespace @@ -767,7 +1003,7 @@ class SDLPoller : public InputCommon::Polling::DevicePoller { public: explicit SDLPoller(SDLState& state_) : state(state_) {} - void Start(const std::string& device_id) override { + void Start([[maybe_unused]] const std::string& device_id) override { state.event_queue.Clear(); state.polling = true; } @@ -794,6 +1030,78 @@ public: } return {}; } + [[nodiscard]] std::optional<Common::ParamPackage> FromEvent(SDL_Event& event) { + switch (event.type) { + case SDL_JOYAXISMOTION: + if (!axis_memory.count(event.jaxis.which) || + !axis_memory[event.jaxis.which].count(event.jaxis.axis)) { + axis_memory[event.jaxis.which][event.jaxis.axis] = event.jaxis.value; + axis_event_count[event.jaxis.which][event.jaxis.axis] = 1; + break; + } else { + axis_event_count[event.jaxis.which][event.jaxis.axis]++; + // The joystick and axis exist in our map if we take this branch, so no checks + // needed + if (std::abs( + (event.jaxis.value - axis_memory[event.jaxis.which][event.jaxis.axis]) / + 32767.0) < 0.5) { + break; + } else { + if (axis_event_count[event.jaxis.which][event.jaxis.axis] == 2 && + IsAxisAtPole(event.jaxis.value) && + IsAxisAtPole(axis_memory[event.jaxis.which][event.jaxis.axis])) { + // If we have exactly two events and both are near a pole, this is + // likely a digital input masquerading as an analog axis; Instead of + // trying to look at the direction the axis travelled, assume the first + // event was press and the second was release; This should handle most + // digital axes while deferring to the direction of travel for analog + // axes + event.jaxis.value = static_cast<Sint16>( + std::copysign(32767, axis_memory[event.jaxis.which][event.jaxis.axis])); + } else { + // There are more than two events, so this is likely a true analog axis, + // check the direction it travelled + event.jaxis.value = static_cast<Sint16>(std::copysign( + 32767, + event.jaxis.value - axis_memory[event.jaxis.which][event.jaxis.axis])); + } + axis_memory.clear(); + axis_event_count.clear(); + } + } + [[fallthrough]]; + case SDL_JOYBUTTONUP: + case SDL_JOYHATMOTION: + return {SDLEventToButtonParamPackage(state, event)}; + } + return std::nullopt; + } + +private: + // Determine whether an axis value is close to an extreme or center + // Some controllers have a digital D-Pad as a pair of analog sticks, with 3 possible values per + // axis, which is why the center must be considered a pole + bool IsAxisAtPole(int16_t value) const { + return std::abs(value) >= 32767 || std::abs(value) < 327; + } + std::unordered_map<SDL_JoystickID, std::unordered_map<uint8_t, int16_t>> axis_memory; + std::unordered_map<SDL_JoystickID, std::unordered_map<uint8_t, uint32_t>> axis_event_count; +}; + +class SDLMotionPoller final : public SDLPoller { +public: + explicit SDLMotionPoller(SDLState& state_) : SDLPoller(state_) {} + + Common::ParamPackage GetNextInput() override { + SDL_Event event; + while (state.event_queue.Pop(event)) { + const auto package = FromEvent(event); + if (package) { + return *package; + } + } + return {}; + } [[nodiscard]] std::optional<Common::ParamPackage> FromEvent(const SDL_Event& event) const { switch (event.type) { case SDL_JOYAXISMOTION: @@ -803,7 +1111,7 @@ public: [[fallthrough]]; case SDL_JOYBUTTONUP: case SDL_JOYHATMOTION: - return {SDLEventToButtonParamPackage(state, event)}; + return {SDLEventToMotionParamPackage(state, event)}; } return std::nullopt; } @@ -821,7 +1129,6 @@ public: void Start(const std::string& device_id) override { SDLPoller::Start(device_id); - // Load the game controller // Reset stored axes analog_x_axis = -1; analog_y_axis = -1; @@ -834,52 +1141,34 @@ public: if (event.type == SDL_JOYAXISMOTION && std::abs(event.jaxis.value / 32767.0) < 0.5) { continue; } - // Simplify controller config by testing if game controller support is enabled. if (event.type == SDL_JOYAXISMOTION) { const auto axis = event.jaxis.axis; - const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which); - const auto controller = joystick->GetSDLGameController(); - if (controller) { - const auto axis_left_x = - SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_LEFTX) - .value.axis; - const auto axis_left_y = - SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_LEFTY) - .value.axis; - const auto axis_right_x = - SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_RIGHTX) - .value.axis; - const auto axis_right_y = - SDL_GameControllerGetBindForAxis(controller, SDL_CONTROLLER_AXIS_RIGHTY) - .value.axis; - - if (axis == axis_left_x || axis == axis_left_y) { - analog_x_axis = axis_left_x; - analog_y_axis = axis_left_y; - break; - } else if (axis == axis_right_x || axis == axis_right_y) { - analog_x_axis = axis_right_x; - analog_y_axis = axis_right_y; - break; - } + // In order to return a complete analog param, we need inputs for both axes. + // First we take the x-axis (horizontal) input, then the y-axis (vertical) input. + if (analog_x_axis == -1) { + analog_x_axis = axis; + } else if (analog_y_axis == -1 && analog_x_axis != axis) { + analog_y_axis = axis; + } + } else { + // If the press wasn't accepted as a joy axis, check for a button press + auto button_press = button_poller.FromEvent(event); + if (button_press) { + return *button_press; } - } - - // If the press wasn't accepted as a joy axis, check for a button press - auto button_press = button_poller.FromEvent(event); - if (button_press) { - return *button_press; } } if (analog_x_axis != -1 && analog_y_axis != -1) { - const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which); - auto params = BuildParamPackageForAnalog(joystick->GetPort(), joystick->GetGUID(), - analog_x_axis, analog_y_axis); - analog_x_axis = -1; - analog_y_axis = -1; - return params; + if (const auto joystick = state.GetSDLJoystickBySDLID(event.jaxis.which)) { + auto params = BuildParamPackageForAnalog(joystick->GetPort(), joystick->GetGUID(), + analog_x_axis, analog_y_axis); + analog_x_axis = -1; + analog_y_axis = -1; + return params; + } } + return {}; } @@ -900,6 +1189,9 @@ SDLState::Pollers SDLState::GetPollers(InputCommon::Polling::DeviceType type) { case InputCommon::Polling::DeviceType::Button: pollers.emplace_back(std::make_unique<Polling::SDLButtonPoller>(*this)); break; + case InputCommon::Polling::DeviceType::Motion: + pollers.emplace_back(std::make_unique<Polling::SDLMotionPoller>(*this)); + break; } return pollers; diff --git a/src/input_common/sdl/sdl_impl.h b/src/input_common/sdl/sdl_impl.h index bd19ba61d..08044b00d 100644 --- a/src/input_common/sdl/sdl_impl.h +++ b/src/input_common/sdl/sdl_impl.h @@ -21,6 +21,8 @@ namespace InputCommon::SDL { class SDLAnalogFactory; class SDLButtonFactory; +class SDLMotionFactory; +class SDLVibrationFactory; class SDLJoystick; class SDLState : public State { @@ -71,6 +73,8 @@ private: std::shared_ptr<SDLButtonFactory> button_factory; std::shared_ptr<SDLAnalogFactory> analog_factory; + std::shared_ptr<SDLVibrationFactory> vibration_factory; + std::shared_ptr<SDLMotionFactory> motion_factory; bool start_thread = false; std::atomic<bool> initialized = false; diff --git a/src/input_common/settings.cpp b/src/input_common/settings.cpp index b66c05856..557e7a9a0 100644 --- a/src/input_common/settings.cpp +++ b/src/input_common/settings.cpp @@ -14,13 +14,6 @@ const std::array<const char*, NumButtons> mapping = {{ }}; } -namespace NativeMotion { -const std::array<const char*, NumMotions> mapping = {{ - "motionleft", - "motionright", -}}; -} - namespace NativeAnalog { const std::array<const char*, NumAnalogs> mapping = {{ "lstick", @@ -28,6 +21,20 @@ const std::array<const char*, NumAnalogs> mapping = {{ }}; } +namespace NativeVibration { +const std::array<const char*, NumVibrations> mapping = {{ + "left_vibration_device", + "right_vibration_device", +}}; +} + +namespace NativeMotion { +const std::array<const char*, NumMotions> mapping = {{ + "motionleft", + "motionright", +}}; +} + namespace NativeMouseButton { const std::array<const char*, NumMouseButtons> mapping = {{ "left", diff --git a/src/input_common/settings.h b/src/input_common/settings.h index ab0b95cf1..75486554b 100644 --- a/src/input_common/settings.h +++ b/src/input_common/settings.h @@ -66,17 +66,32 @@ constexpr int NUM_STICKS_HID = NumAnalogs; extern const std::array<const char*, NumAnalogs> mapping; } // namespace NativeAnalog +namespace NativeVibration { +enum Values : int { + LeftVibrationDevice, + RightVibrationDevice, + + NumVibrations, +}; + +constexpr int VIBRATION_HID_BEGIN = LeftVibrationDevice; +constexpr int VIBRATION_HID_END = NumVibrations; +constexpr int NUM_VIBRATIONS_HID = NumVibrations; + +extern const std::array<const char*, NumVibrations> mapping; +}; // namespace NativeVibration + namespace NativeMotion { enum Values : int { - MOTIONLEFT, - MOTIONRIGHT, + MotionLeft, + MotionRight, NumMotions, }; -constexpr int MOTION_HID_BEGIN = MOTIONLEFT; +constexpr int MOTION_HID_BEGIN = MotionLeft; constexpr int MOTION_HID_END = NumMotions; -constexpr int NUM_MOTION_HID = NumMotions; +constexpr int NUM_MOTIONS_HID = NumMotions; extern const std::array<const char*, NumMotions> mapping; } // namespace NativeMotion @@ -305,9 +320,11 @@ constexpr int NUM_KEYBOARD_MODS_HID = NumKeyboardMods; } // namespace NativeKeyboard -using ButtonsRaw = std::array<std::string, NativeButton::NumButtons>; using AnalogsRaw = std::array<std::string, NativeAnalog::NumAnalogs>; -using MotionRaw = std::array<std::string, NativeMotion::NumMotions>; +using ButtonsRaw = std::array<std::string, NativeButton::NumButtons>; +using MotionsRaw = std::array<std::string, NativeMotion::NumMotions>; +using VibrationsRaw = std::array<std::string, NativeVibration::NumVibrations>; + using MouseButtonsRaw = std::array<std::string, NativeMouseButton::NumMouseButtons>; using KeyboardKeysRaw = std::array<std::string, NativeKeyboard::NumKeyboardKeys>; using KeyboardModsRaw = std::array<std::string, NativeKeyboard::NumKeyboardMods>; @@ -330,9 +347,11 @@ struct PlayerInput { ControllerType controller_type; ButtonsRaw buttons; AnalogsRaw analogs; - MotionRaw motions; - std::string lstick_mod; - std::string rstick_mod; + VibrationsRaw vibrations; + MotionsRaw motions; + + bool vibration_enabled; + int vibration_strength; u32 body_color_left; u32 body_color_right; diff --git a/src/input_common/touch_from_button.cpp b/src/input_common/touch_from_button.cpp index 98da0ef1a..a07124a86 100644 --- a/src/input_common/touch_from_button.cpp +++ b/src/input_common/touch_from_button.cpp @@ -11,9 +11,11 @@ namespace InputCommon { class TouchFromButtonDevice final : public Input::TouchDevice { public: TouchFromButtonDevice() { - for (const auto& config_entry : - Settings::values.touch_from_button_maps[Settings::values.touch_from_button_map_index] - .buttons) { + const auto button_index = + static_cast<std::size_t>(Settings::values.touch_from_button_map_index); + const auto& buttons = Settings::values.touch_from_button_maps[button_index].buttons; + + for (const auto& config_entry : buttons) { const Common::ParamPackage package{config_entry}; map.emplace_back( Input::CreateDevice<Input::ButtonDevice>(config_entry), @@ -42,8 +44,7 @@ private: std::vector<std::tuple<std::unique_ptr<Input::ButtonDevice>, int, int>> map; }; -std::unique_ptr<Input::TouchDevice> TouchFromButtonFactory::Create( - const Common::ParamPackage& params) { +std::unique_ptr<Input::TouchDevice> TouchFromButtonFactory::Create(const Common::ParamPackage&) { return std::make_unique<TouchFromButtonDevice>(); } diff --git a/src/input_common/udp/client.cpp b/src/input_common/udp/client.cpp index 2b6a68d4b..412d57896 100644 --- a/src/input_common/udp/client.cpp +++ b/src/input_common/udp/client.cpp @@ -26,11 +26,11 @@ class Socket { public: using clock = std::chrono::system_clock; - explicit Socket(const std::string& host, u16 port, u8 pad_index, u32 client_id, - SocketCallback callback) - : callback(std::move(callback)), timer(io_service), - socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(client_id), - pad_index(pad_index) { + explicit Socket(const std::string& host, u16 port, std::size_t pad_index_, u32 client_id_, + SocketCallback callback_) + : callback(std::move(callback_)), timer(io_service), + socket(io_service, udp::endpoint(udp::v4(), 0)), client_id(client_id_), + pad_index(pad_index_) { boost::system::error_code ec{}; auto ipv4 = boost::asio::ip::make_address_v4(host, ec); if (ec.value() != boost::system::errc::success) { @@ -63,7 +63,7 @@ public: } private: - void HandleReceive(const boost::system::error_code& error, std::size_t bytes_transferred) { + void HandleReceive(const boost::system::error_code&, std::size_t bytes_transferred) { if (auto type = Response::Validate(receive_buffer.data(), bytes_transferred)) { switch (*type) { case Type::Version: { @@ -90,16 +90,20 @@ private: StartReceive(); } - void HandleSend(const boost::system::error_code& error) { + void HandleSend(const boost::system::error_code&) { boost::system::error_code _ignored{}; // Send a request for getting port info for the pad - Request::PortInfo port_info{1, {pad_index, 0, 0, 0}}; + const Request::PortInfo port_info{1, {static_cast<u8>(pad_index), 0, 0, 0}}; const auto port_message = Request::Create(port_info, client_id); std::memcpy(&send_buffer1, &port_message, PORT_INFO_SIZE); socket.send_to(boost::asio::buffer(send_buffer1), send_endpoint, {}, _ignored); // Send a request for getting pad data for the pad - Request::PadData pad_data{Request::PadData::Flags::Id, pad_index, EMPTY_MAC_ADDRESS}; + const Request::PadData pad_data{ + Request::PadData::Flags::Id, + static_cast<u8>(pad_index), + EMPTY_MAC_ADDRESS, + }; const auto pad_message = Request::Create(pad_data, client_id); std::memcpy(send_buffer2.data(), &pad_message, PAD_DATA_SIZE); socket.send_to(boost::asio::buffer(send_buffer2), send_endpoint, {}, _ignored); @@ -112,7 +116,7 @@ private: udp::socket socket; u32 client_id{}; - u8 pad_index{}; + std::size_t pad_index{}; static constexpr std::size_t PORT_INFO_SIZE = sizeof(Message<Request::PortInfo>); static constexpr std::size_t PAD_DATA_SIZE = sizeof(Message<Request::PadData>); @@ -132,15 +136,7 @@ static void SocketLoop(Socket* socket) { Client::Client() { LOG_INFO(Input, "Udp Initialization started"); - for (std::size_t client = 0; client < clients.size(); client++) { - u8 pad = client % 4; - StartCommunication(client, Settings::values.udp_input_address, - Settings::values.udp_input_port, pad, 24872); - // Set motion parameters - // SetGyroThreshold value should be dependent on GyroscopeZeroDriftMode - // Real HW values are unknown, 0.0001 is an approximate to Standard - clients[client].motion.SetGyroThreshold(0.0001f); - } + ReloadSockets(); } Client::~Client() { @@ -163,39 +159,77 @@ std::vector<Common::ParamPackage> Client::GetInputDevices() const { return devices; } -bool Client::DeviceConnected(std::size_t pad) const { +bool Client::DeviceConnected(std::size_t client) const { // Use last timestamp to detect if the socket has stopped sending data - const auto now = std::chrono::system_clock::now(); - u64 time_difference = - std::chrono::duration_cast<std::chrono::milliseconds>(now - clients[pad].last_motion_update) - .count(); - return time_difference < 1000 && clients[pad].active == 1; + const auto now = std::chrono::steady_clock::now(); + const auto time_difference = + static_cast<u64>(std::chrono::duration_cast<std::chrono::milliseconds>( + now - clients[client].last_motion_update) + .count()); + return time_difference < 1000 && clients[client].active == 1; } -void Client::ReloadUDPClient() { - for (std::size_t client = 0; client < clients.size(); client++) { - ReloadSocket(Settings::values.udp_input_address, Settings::values.udp_input_port, client); +void Client::ReloadSockets() { + Reset(); + + std::stringstream servers_ss(Settings::values.udp_input_servers); + std::string server_token; + std::size_t client = 0; + while (std::getline(servers_ss, server_token, ',')) { + if (client == max_udp_clients) { + break; + } + std::stringstream server_ss(server_token); + std::string token; + std::getline(server_ss, token, ':'); + std::string udp_input_address = token; + std::getline(server_ss, token, ':'); + char* temp; + const u16 udp_input_port = static_cast<u16>(std::strtol(token.c_str(), &temp, 0)); + if (*temp != '\0') { + LOG_ERROR(Input, "Port number is not valid {}", token); + continue; + } + + for (std::size_t pad = 0; pad < 4; ++pad) { + const std::size_t client_number = + GetClientNumber(udp_input_address, udp_input_port, pad); + if (client_number != max_udp_clients) { + LOG_ERROR(Input, "Duplicated UDP servers found"); + continue; + } + StartCommunication(client++, udp_input_address, udp_input_port, pad, 24872); + } } } -void Client::ReloadSocket(const std::string& host, u16 port, u8 pad_index, u32 client_id) { - // client number must be determined from host / port and pad index - std::size_t client = pad_index; - clients[client].socket->Stop(); - clients[client].thread.join(); - StartCommunication(client, host, port, pad_index, client_id); + +std::size_t Client::GetClientNumber(std::string_view host, u16 port, std::size_t pad) const { + for (std::size_t client = 0; client < clients.size(); client++) { + if (clients[client].active == -1) { + continue; + } + if (clients[client].host == host && clients[client].port == port && + clients[client].pad_index == pad) { + return client; + } + } + return max_udp_clients; } -void Client::OnVersion(Response::Version data) { +void Client::OnVersion([[maybe_unused]] Response::Version data) { LOG_TRACE(Input, "Version packet received: {}", data.version); } -void Client::OnPortInfo(Response::PortInfo data) { +void Client::OnPortInfo([[maybe_unused]] Response::PortInfo data) { LOG_TRACE(Input, "PortInfo packet received: {}", data.model); } -void Client::OnPadData(Response::PadData data) { - // client number must be determined from host / port and pad index - std::size_t client = data.info.id; +void Client::OnPadData(Response::PadData data, std::size_t client) { + // Accept packets only for the correct pad + if (static_cast<u8>(clients[client].pad_index) != data.info.id) { + return; + } + LOG_TRACE(Input, "PadData packet received"); if (data.packet_counter == clients[client].packet_sequence) { LOG_WARNING( @@ -204,14 +238,15 @@ void Client::OnPadData(Response::PadData data) { clients[client].packet_sequence, data.packet_counter); return; } - clients[client].active = data.info.is_pad_active; + clients[client].active = static_cast<s8>(data.info.is_pad_active); clients[client].packet_sequence = data.packet_counter; - const auto now = std::chrono::system_clock::now(); - u64 time_difference = std::chrono::duration_cast<std::chrono::microseconds>( - now - clients[client].last_motion_update) - .count(); + const auto now = std::chrono::steady_clock::now(); + const auto time_difference = + static_cast<u64>(std::chrono::duration_cast<std::chrono::microseconds>( + now - clients[client].last_motion_update) + .count()); clients[client].last_motion_update = now; - Common::Vec3f raw_gyroscope = {data.gyro.pitch, data.gyro.roll, -data.gyro.yaw}; + const Common::Vec3f raw_gyroscope = {data.gyro.pitch, data.gyro.roll, -data.gyro.yaw}; clients[client].motion.SetAcceleration({data.accel.x, -data.accel.z, data.accel.y}); // Gyroscope values are not it the correct scale from better joy. // Dividing by 312 allows us to make one full turn = 1 turn @@ -219,14 +254,10 @@ void Client::OnPadData(Response::PadData data) { clients[client].motion.SetGyroscope(raw_gyroscope / 312.0f); clients[client].motion.UpdateRotation(time_difference); clients[client].motion.UpdateOrientation(time_difference); - Common::Vec3f gyroscope = clients[client].motion.GetGyroscope(); - Common::Vec3f accelerometer = clients[client].motion.GetAcceleration(); - Common::Vec3f rotation = clients[client].motion.GetRotations(); - std::array<Common::Vec3f, 3> orientation = clients[client].motion.GetOrientation(); { std::lock_guard guard(clients[client].status.update_mutex); - clients[client].status.motion_status = {accelerometer, gyroscope, rotation, orientation}; + clients[client].status.motion_status = clients[client].motion.GetMotion(); // TODO: add a setting for "click" touch. Click touch refers to a device that differentiates // between a simple "tap" and a hard press that causes the touch screen to click. @@ -241,98 +272,129 @@ void Client::OnPadData(Response::PadData data) { const u16 min_y = clients[client].status.touch_calibration->min_y; const u16 max_y = clients[client].status.touch_calibration->max_y; - x = (std::clamp(static_cast<u16>(data.touch_1.x), min_x, max_x) - min_x) / + x = static_cast<float>(std::clamp(static_cast<u16>(data.touch_1.x), min_x, max_x) - + min_x) / static_cast<float>(max_x - min_x); - y = (std::clamp(static_cast<u16>(data.touch_1.y), min_y, max_y) - min_y) / + y = static_cast<float>(std::clamp(static_cast<u16>(data.touch_1.y), min_y, max_y) - + min_y) / static_cast<float>(max_y - min_y); } clients[client].status.touch_status = {x, y, is_active}; if (configuring) { + const Common::Vec3f gyroscope = clients[client].motion.GetGyroscope(); + const Common::Vec3f accelerometer = clients[client].motion.GetAcceleration(); UpdateYuzuSettings(client, accelerometer, gyroscope, is_active); } } } -void Client::StartCommunication(std::size_t client, const std::string& host, u16 port, u8 pad_index, - u32 client_id) { +void Client::StartCommunication(std::size_t client, const std::string& host, u16 port, + std::size_t pad_index, u32 client_id) { SocketCallback callback{[this](Response::Version version) { OnVersion(version); }, [this](Response::PortInfo info) { OnPortInfo(info); }, - [this](Response::PadData data) { OnPadData(data); }}; - LOG_INFO(Input, "Starting communication with UDP input server on {}:{}", host, port); + [this, client](Response::PadData data) { OnPadData(data, client); }}; + LOG_INFO(Input, "Starting communication with UDP input server on {}:{}:{}", host, port, + pad_index); + clients[client].host = host; + clients[client].port = port; + clients[client].pad_index = pad_index; + clients[client].active = 0; clients[client].socket = std::make_unique<Socket>(host, port, pad_index, client_id, callback); clients[client].thread = std::thread{SocketLoop, clients[client].socket.get()}; + // Set motion parameters + // SetGyroThreshold value should be dependent on GyroscopeZeroDriftMode + // Real HW values are unknown, 0.0001 is an approximate to Standard + clients[client].motion.SetGyroThreshold(0.0001f); } void Client::Reset() { - for (std::size_t client = 0; client < clients.size(); client++) { - clients[client].socket->Stop(); - clients[client].thread.join(); + for (auto& client : clients) { + if (client.thread.joinable()) { + client.active = -1; + client.socket->Stop(); + client.thread.join(); + } } } void Client::UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& acc, const Common::Vec3<float>& gyro, bool touch) { - UDPPadStatus pad; + if (gyro.Length() > 0.2f) { + LOG_DEBUG(Input, "UDP Controller {}: gyro=({}, {}, {}), accel=({}, {}, {}), touch={}", + client, gyro[0], gyro[1], gyro[2], acc[0], acc[1], acc[2], touch); + } + UDPPadStatus pad{ + .host = clients[client].host, + .port = clients[client].port, + .pad_index = clients[client].pad_index, + }; if (touch) { pad.touch = PadTouch::Click; - pad_queue[client].Push(pad); + pad_queue.Push(pad); } for (size_t i = 0; i < 3; ++i) { - if (gyro[i] > 6.0f || gyro[i] < -6.0f) { + if (gyro[i] > 5.0f || gyro[i] < -5.0f) { pad.motion = static_cast<PadMotion>(i); pad.motion_value = gyro[i]; - pad_queue[client].Push(pad); + pad_queue.Push(pad); } - if (acc[i] > 2.0f || acc[i] < -2.0f) { + if (acc[i] > 1.75f || acc[i] < -1.75f) { pad.motion = static_cast<PadMotion>(i + 3); pad.motion_value = acc[i]; - pad_queue[client].Push(pad); + pad_queue.Push(pad); } } } void Client::BeginConfiguration() { - for (auto& pq : pad_queue) { - pq.Clear(); - } + pad_queue.Clear(); configuring = true; } void Client::EndConfiguration() { - for (auto& pq : pad_queue) { - pq.Clear(); - } + pad_queue.Clear(); configuring = false; } -DeviceStatus& Client::GetPadState(std::size_t pad) { - return clients[pad].status; +DeviceStatus& Client::GetPadState(const std::string& host, u16 port, std::size_t pad) { + const std::size_t client_number = GetClientNumber(host, port, pad); + if (client_number == max_udp_clients) { + return clients[0].status; + } + return clients[client_number].status; } -const DeviceStatus& Client::GetPadState(std::size_t pad) const { - return clients[pad].status; +const DeviceStatus& Client::GetPadState(const std::string& host, u16 port, std::size_t pad) const { + const std::size_t client_number = GetClientNumber(host, port, pad); + if (client_number == max_udp_clients) { + return clients[0].status; + } + return clients[client_number].status; } -std::array<Common::SPSCQueue<UDPPadStatus>, 4>& Client::GetPadQueue() { +Common::SPSCQueue<UDPPadStatus>& Client::GetPadQueue() { return pad_queue; } -const std::array<Common::SPSCQueue<UDPPadStatus>, 4>& Client::GetPadQueue() const { +const Common::SPSCQueue<UDPPadStatus>& Client::GetPadQueue() const { return pad_queue; } -void TestCommunication(const std::string& host, u16 port, u8 pad_index, u32 client_id, - std::function<void()> success_callback, - std::function<void()> failure_callback) { +void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, u32 client_id, + const std::function<void()>& success_callback, + const std::function<void()>& failure_callback) { std::thread([=] { Common::Event success_event; - SocketCallback callback{[](Response::Version version) {}, [](Response::PortInfo info) {}, - [&](Response::PadData data) { success_event.Set(); }}; + SocketCallback callback{ + .version = [](Response::Version) {}, + .port_info = [](Response::PortInfo) {}, + .pad_data = [&](Response::PadData) { success_event.Set(); }, + }; Socket socket{host, port, pad_index, client_id, std::move(callback)}; std::thread worker_thread{SocketLoop, &socket}; - bool result = success_event.WaitFor(std::chrono::seconds(8)); + const bool result = success_event.WaitFor(std::chrono::seconds(5)); socket.Stop(); worker_thread.join(); if (result) { @@ -344,7 +406,7 @@ void TestCommunication(const std::string& host, u16 port, u8 pad_index, u32 clie } CalibrationConfigurationJob::CalibrationConfigurationJob( - const std::string& host, u16 port, u8 pad_index, u32 client_id, + const std::string& host, u16 port, std::size_t pad_index, u32 client_id, std::function<void(Status)> status_callback, std::function<void(u16, u16, u16, u16)> data_callback) { @@ -357,14 +419,14 @@ CalibrationConfigurationJob::CalibrationConfigurationJob( u16 max_y{}; Status current_status{Status::Initialized}; - SocketCallback callback{[](Response::Version version) {}, [](Response::PortInfo info) {}, + SocketCallback callback{[](Response::Version) {}, [](Response::PortInfo) {}, [&](Response::PadData data) { if (current_status == Status::Initialized) { // Receiving data means the communication is ready now current_status = Status::Ready; status_callback(current_status); } - if (!data.touch_1.is_active) { + if (data.touch_1.is_active == 0) { return; } LOG_DEBUG(Input, "Current touch: {} {}", data.touch_1.x, diff --git a/src/input_common/udp/client.h b/src/input_common/udp/client.h index 523dc6a7a..00c8b09f5 100644 --- a/src/input_common/udp/client.h +++ b/src/input_common/udp/client.h @@ -21,8 +21,7 @@ namespace InputCommon::CemuhookUDP { -constexpr u16 DEFAULT_PORT = 26760; -constexpr char DEFAULT_ADDR[] = "127.0.0.1"; +constexpr char DEFAULT_SRV[] = "127.0.0.1:26760"; class Socket; @@ -48,6 +47,9 @@ enum class PadTouch { }; struct UDPPadStatus { + std::string host{"127.0.0.1"}; + u16 port{26760}; + std::size_t pad_index{}; PadTouch touch{PadTouch::Undefined}; PadMotion motion{PadMotion::Undefined}; f32 motion_value{0.0f}; @@ -82,46 +84,52 @@ public: std::vector<Common::ParamPackage> GetInputDevices() const; - bool DeviceConnected(std::size_t pad) const; - void ReloadUDPClient(); - void ReloadSocket(const std::string& host = "127.0.0.1", u16 port = 26760, u8 pad_index = 0, - u32 client_id = 24872); + bool DeviceConnected(std::size_t client) const; + void ReloadSockets(); - std::array<Common::SPSCQueue<UDPPadStatus>, 4>& GetPadQueue(); - const std::array<Common::SPSCQueue<UDPPadStatus>, 4>& GetPadQueue() const; + Common::SPSCQueue<UDPPadStatus>& GetPadQueue(); + const Common::SPSCQueue<UDPPadStatus>& GetPadQueue() const; - DeviceStatus& GetPadState(std::size_t pad); - const DeviceStatus& GetPadState(std::size_t pad) const; + DeviceStatus& GetPadState(const std::string& host, u16 port, std::size_t pad); + const DeviceStatus& GetPadState(const std::string& host, u16 port, std::size_t pad) const; private: struct ClientData { + std::string host{"127.0.0.1"}; + u16 port{26760}; + std::size_t pad_index{}; std::unique_ptr<Socket> socket; DeviceStatus status; std::thread thread; - u64 packet_sequence = 0; - u8 active; + u64 packet_sequence{}; + s8 active{-1}; // Realtime values // motion is initalized with PID values for drift correction on joycons InputCommon::MotionInput motion{0.3f, 0.005f, 0.0f}; - std::chrono::time_point<std::chrono::system_clock> last_motion_update; + std::chrono::time_point<std::chrono::steady_clock> last_motion_update; }; // For shutting down, clear all data, join all threads, release usb void Reset(); + // Translates configuration to client number + std::size_t GetClientNumber(std::string_view host, u16 port, std::size_t pad) const; + void OnVersion(Response::Version); void OnPortInfo(Response::PortInfo); - void OnPadData(Response::PadData); - void StartCommunication(std::size_t client, const std::string& host, u16 port, u8 pad_index, - u32 client_id); + void OnPadData(Response::PadData, std::size_t client); + void StartCommunication(std::size_t client, const std::string& host, u16 port, + std::size_t pad_index, u32 client_id); void UpdateYuzuSettings(std::size_t client, const Common::Vec3<float>& acc, const Common::Vec3<float>& gyro, bool touch); bool configuring = false; - std::array<ClientData, 4> clients; - std::array<Common::SPSCQueue<UDPPadStatus>, 4> pad_queue; + // Allocate clients for 8 udp servers + const std::size_t max_udp_clients = 32; + std::array<ClientData, 4 * 8> clients; + Common::SPSCQueue<UDPPadStatus> pad_queue; }; /// An async job allowing configuration of the touchpad calibration. @@ -139,7 +147,7 @@ public: * @param status_callback Callback for job status updates * @param data_callback Called when calibration data is ready */ - explicit CalibrationConfigurationJob(const std::string& host, u16 port, u8 pad_index, + explicit CalibrationConfigurationJob(const std::string& host, u16 port, std::size_t pad_index, u32 client_id, std::function<void(Status)> status_callback, std::function<void(u16, u16, u16, u16)> data_callback); ~CalibrationConfigurationJob(); @@ -149,8 +157,8 @@ private: Common::Event complete_event; }; -void TestCommunication(const std::string& host, u16 port, u8 pad_index, u32 client_id, - std::function<void()> success_callback, - std::function<void()> failure_callback); +void TestCommunication(const std::string& host, u16 port, std::size_t pad_index, u32 client_id, + const std::function<void()>& success_callback, + const std::function<void()>& failure_callback); } // namespace InputCommon::CemuhookUDP diff --git a/src/input_common/udp/protocol.h b/src/input_common/udp/protocol.h index 3ba4d1fc8..fc1aea4b9 100644 --- a/src/input_common/udp/protocol.h +++ b/src/input_common/udp/protocol.h @@ -7,7 +7,16 @@ #include <array> #include <optional> #include <type_traits> + +#ifdef _MSC_VER +#pragma warning(push) +#pragma warning(disable : 4701) +#endif #include <boost/crc.hpp> +#ifdef _MSC_VER +#pragma warning(pop) +#endif + #include "common/bit_field.h" #include "common/swap.h" @@ -93,7 +102,7 @@ static_assert(std::is_trivially_copyable_v<PadData>, /** * Creates a message with the proper header data that can be sent to the server. - * @param T data Request body to send + * @param data Request body to send * @param client_id ID of the udp client (usually not checked on the server) */ template <typename T> diff --git a/src/input_common/udp/udp.cpp b/src/input_common/udp/udp.cpp index eba077a36..c5da27a38 100644 --- a/src/input_common/udp/udp.cpp +++ b/src/input_common/udp/udp.cpp @@ -2,8 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <atomic> -#include <list> #include <mutex> #include <utility> #include "common/assert.h" @@ -15,36 +13,36 @@ namespace InputCommon { class UDPMotion final : public Input::MotionDevice { public: - UDPMotion(std::string ip_, int port_, int pad_, CemuhookUDP::Client* client_) - : ip(ip_), port(port_), pad(pad_), client(client_) {} + explicit UDPMotion(std::string ip_, u16 port_, u16 pad_, CemuhookUDP::Client* client_) + : ip(std::move(ip_)), port(port_), pad(pad_), client(client_) {} Input::MotionStatus GetStatus() const override { - return client->GetPadState(pad).motion_status; + return client->GetPadState(ip, port, pad).motion_status; } private: const std::string ip; - const int port; - const int pad; + const u16 port; + const u16 pad; CemuhookUDP::Client* client; mutable std::mutex mutex; }; -/// A motion device factory that creates motion devices from JC Adapter +/// A motion device factory that creates motion devices from a UDP client UDPMotionFactory::UDPMotionFactory(std::shared_ptr<CemuhookUDP::Client> client_) : client(std::move(client_)) {} /** * Creates motion device * @param params contains parameters for creating the device: - * - "port": the nth jcpad on the adapter + * - "port": the UDP port number */ std::unique_ptr<Input::MotionDevice> UDPMotionFactory::Create(const Common::ParamPackage& params) { - const std::string ip = params.Get("ip", "127.0.0.1"); - const int port = params.Get("port", 26760); - const int pad = params.Get("pad_index", 0); + auto ip = params.Get("ip", "127.0.0.1"); + const auto port = static_cast<u16>(params.Get("port", 26760)); + const auto pad = static_cast<u16>(params.Get("pad_index", 0)); - return std::make_unique<UDPMotion>(ip, port, pad, client.get()); + return std::make_unique<UDPMotion>(std::move(ip), port, pad, client.get()); } void UDPMotionFactory::BeginConfiguration() { @@ -61,54 +59,52 @@ Common::ParamPackage UDPMotionFactory::GetNextInput() { Common::ParamPackage params; CemuhookUDP::UDPPadStatus pad; auto& queue = client->GetPadQueue(); - for (std::size_t pad_number = 0; pad_number < queue.size(); ++pad_number) { - while (queue[pad_number].Pop(pad)) { - if (pad.motion == CemuhookUDP::PadMotion::Undefined || std::abs(pad.motion_value) < 1) { - continue; - } - params.Set("engine", "cemuhookudp"); - params.Set("ip", "127.0.0.1"); - params.Set("port", 26760); - params.Set("pad_index", static_cast<int>(pad_number)); - params.Set("motion", static_cast<u16>(pad.motion)); - return params; + while (queue.Pop(pad)) { + if (pad.motion == CemuhookUDP::PadMotion::Undefined || std::abs(pad.motion_value) < 1) { + continue; } + params.Set("engine", "cemuhookudp"); + params.Set("ip", pad.host); + params.Set("port", static_cast<u16>(pad.port)); + params.Set("pad_index", static_cast<u16>(pad.pad_index)); + params.Set("motion", static_cast<u16>(pad.motion)); + return params; } return params; } class UDPTouch final : public Input::TouchDevice { public: - UDPTouch(std::string ip_, int port_, int pad_, CemuhookUDP::Client* client_) + explicit UDPTouch(std::string ip_, u16 port_, u16 pad_, CemuhookUDP::Client* client_) : ip(std::move(ip_)), port(port_), pad(pad_), client(client_) {} std::tuple<float, float, bool> GetStatus() const override { - return client->GetPadState(pad).touch_status; + return client->GetPadState(ip, port, pad).touch_status; } private: const std::string ip; - const int port; - const int pad; + const u16 port; + const u16 pad; CemuhookUDP::Client* client; mutable std::mutex mutex; }; -/// A motion device factory that creates motion devices from JC Adapter +/// A motion device factory that creates motion devices from a UDP client UDPTouchFactory::UDPTouchFactory(std::shared_ptr<CemuhookUDP::Client> client_) : client(std::move(client_)) {} /** * Creates motion device * @param params contains parameters for creating the device: - * - "port": the nth jcpad on the adapter + * - "port": the UDP port number */ std::unique_ptr<Input::TouchDevice> UDPTouchFactory::Create(const Common::ParamPackage& params) { - const std::string ip = params.Get("ip", "127.0.0.1"); - const int port = params.Get("port", 26760); - const int pad = params.Get("pad_index", 0); + auto ip = params.Get("ip", "127.0.0.1"); + const auto port = static_cast<u16>(params.Get("port", 26760)); + const auto pad = static_cast<u16>(params.Get("pad_index", 0)); - return std::make_unique<UDPTouch>(ip, port, pad, client.get()); + return std::make_unique<UDPTouch>(std::move(ip), port, pad, client.get()); } void UDPTouchFactory::BeginConfiguration() { @@ -125,18 +121,16 @@ Common::ParamPackage UDPTouchFactory::GetNextInput() { Common::ParamPackage params; CemuhookUDP::UDPPadStatus pad; auto& queue = client->GetPadQueue(); - for (std::size_t pad_number = 0; pad_number < queue.size(); ++pad_number) { - while (queue[pad_number].Pop(pad)) { - if (pad.touch == CemuhookUDP::PadTouch::Undefined) { - continue; - } - params.Set("engine", "cemuhookudp"); - params.Set("ip", "127.0.0.1"); - params.Set("port", 26760); - params.Set("pad_index", static_cast<int>(pad_number)); - params.Set("touch", static_cast<u16>(pad.touch)); - return params; + while (queue.Pop(pad)) { + if (pad.touch == CemuhookUDP::PadTouch::Undefined) { + continue; } + params.Set("engine", "cemuhookudp"); + params.Set("ip", pad.host); + params.Set("port", static_cast<u16>(pad.port)); + params.Set("pad_index", static_cast<u16>(pad.pad_index)); + params.Set("touch", static_cast<u16>(pad.touch)); + return params; } return params; } diff --git a/src/tests/CMakeLists.txt b/src/tests/CMakeLists.txt index 47ef30aa9..8a606b448 100644 --- a/src/tests/CMakeLists.txt +++ b/src/tests/CMakeLists.txt @@ -2,11 +2,8 @@ add_executable(tests common/bit_field.cpp common/bit_utils.cpp common/fibers.cpp - common/multi_level_queue.cpp common/param_package.cpp common/ring_buffer.cpp - core/arm/arm_test_common.cpp - core/arm/arm_test_common.h core/core_timing.cpp tests.cpp ) diff --git a/src/tests/common/bit_field.cpp b/src/tests/common/bit_field.cpp index 8ca1889f9..182638000 100644 --- a/src/tests/common/bit_field.cpp +++ b/src/tests/common/bit_field.cpp @@ -68,7 +68,7 @@ TEST_CASE("BitField", "[common]") { }}); // bit fields: 01101100111101'10101110'1011'101100 - REQUIRE(be_bitfield.raw == 0b01101100'11110110'10111010'11101100); + REQUIRE(be_bitfield.raw == 0b01101100'11110110'10111010'11101100U); REQUIRE(be_bitfield.a == 0b101100); REQUIRE(be_bitfield.b == -5); // 1011 as two's complement REQUIRE(be_bitfield.c == TestEnum::B); @@ -80,7 +80,7 @@ TEST_CASE("BitField", "[common]") { be_bitfield.d.Assign(0b01010101010101); std::memcpy(&raw, &be_bitfield, sizeof(raw)); // bit fields: 01010101010101'00001111'1111'000111 - REQUIRE(be_bitfield.raw == 0b01010101'01010100'00111111'11000111); + REQUIRE(be_bitfield.raw == 0b01010101'01010100'00111111'11000111U); REQUIRE(raw == std::array<u8, 4>{{ 0b01010101, 0b01010100, diff --git a/src/tests/common/fibers.cpp b/src/tests/common/fibers.cpp index 4fd92428f..d94492fc6 100644 --- a/src/tests/common/fibers.cpp +++ b/src/tests/common/fibers.cpp @@ -6,18 +6,40 @@ #include <cstdlib> #include <functional> #include <memory> +#include <mutex> +#include <stdexcept> #include <thread> #include <unordered_map> #include <vector> #include <catch2/catch.hpp> -#include <math.h> + #include "common/common_types.h" #include "common/fiber.h" -#include "common/spin_lock.h" namespace Common { +class ThreadIds { +public: + void Register(u32 id) { + const auto thread_id = std::this_thread::get_id(); + std::scoped_lock lock{mutex}; + if (ids.contains(thread_id)) { + throw std::logic_error{"Registering the same thread twice"}; + } + ids.emplace(thread_id, id); + } + + [[nodiscard]] u32 Get() const { + std::scoped_lock lock{mutex}; + return ids.at(std::this_thread::get_id()); + } + +private: + mutable std::mutex mutex; + std::unordered_map<std::thread::id, u32> ids; +}; + class TestControl1 { public: TestControl1() = default; @@ -26,7 +48,7 @@ public: void ExecuteThread(u32 id); - std::unordered_map<std::thread::id, u32> ids; + ThreadIds thread_ids; std::vector<std::shared_ptr<Common::Fiber>> thread_fibers; std::vector<std::shared_ptr<Common::Fiber>> work_fibers; std::vector<u32> items; @@ -39,8 +61,7 @@ static void WorkControl1(void* control) { } void TestControl1::DoWork() { - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); u32 value = items[id]; for (u32 i = 0; i < id; i++) { value++; @@ -50,8 +71,7 @@ void TestControl1::DoWork() { } void TestControl1::ExecuteThread(u32 id) { - std::thread::id this_id = std::this_thread::get_id(); - ids[this_id] = id; + thread_ids.Register(id); auto thread_fiber = Fiber::ThreadToFiber(); thread_fibers[id] = thread_fiber; work_fibers[id] = std::make_shared<Fiber>(std::function<void(void*)>{WorkControl1}, this); @@ -98,8 +118,7 @@ public: value1 += i; } Fiber::YieldTo(fiber1, fiber3); - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); assert1 = id == 1; value2 += 5000; Fiber::YieldTo(fiber1, thread_fibers[id]); @@ -115,8 +134,7 @@ public: } void DoWork3() { - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); assert2 = id == 0; value1 += 1000; Fiber::YieldTo(fiber3, thread_fibers[id]); @@ -125,14 +143,12 @@ public: void ExecuteThread(u32 id); void CallFiber1() { - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); Fiber::YieldTo(thread_fibers[id], fiber1); } void CallFiber2() { - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); Fiber::YieldTo(thread_fibers[id], fiber2); } @@ -145,7 +161,7 @@ public: u32 value2{}; std::atomic<bool> trap{true}; std::atomic<bool> trap2{true}; - std::unordered_map<std::thread::id, u32> ids; + ThreadIds thread_ids; std::vector<std::shared_ptr<Common::Fiber>> thread_fibers; std::shared_ptr<Common::Fiber> fiber1; std::shared_ptr<Common::Fiber> fiber2; @@ -168,15 +184,13 @@ static void WorkControl2_3(void* control) { } void TestControl2::ExecuteThread(u32 id) { - std::thread::id this_id = std::this_thread::get_id(); - ids[this_id] = id; + thread_ids.Register(id); auto thread_fiber = Fiber::ThreadToFiber(); thread_fibers[id] = thread_fiber; } void TestControl2::Exit() { - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); thread_fibers[id]->Exit(); } @@ -193,7 +207,7 @@ static void ThreadStart2_2(u32 id, TestControl2& test_control) { } /** This test checks for fiber thread exchange configuration and validates that fibers are - * that a fiber has been succesfully transfered from one thread to another and that the TLS + * that a fiber has been successfully transferred from one thread to another and that the TLS * region of the thread is kept while changing fibers. */ TEST_CASE("Fibers::InterExchange", "[common]") { @@ -228,24 +242,21 @@ public: void DoWork1() { value1 += 1; Fiber::YieldTo(fiber1, fiber2); - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); value3 += 1; Fiber::YieldTo(fiber1, thread_fibers[id]); } void DoWork2() { value2 += 1; - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); Fiber::YieldTo(fiber2, thread_fibers[id]); } void ExecuteThread(u32 id); void CallFiber1() { - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); Fiber::YieldTo(thread_fibers[id], fiber1); } @@ -254,7 +265,7 @@ public: u32 value1{}; u32 value2{}; u32 value3{}; - std::unordered_map<std::thread::id, u32> ids; + ThreadIds thread_ids; std::vector<std::shared_ptr<Common::Fiber>> thread_fibers; std::shared_ptr<Common::Fiber> fiber1; std::shared_ptr<Common::Fiber> fiber2; @@ -271,15 +282,13 @@ static void WorkControl3_2(void* control) { } void TestControl3::ExecuteThread(u32 id) { - std::thread::id this_id = std::this_thread::get_id(); - ids[this_id] = id; + thread_ids.Register(id); auto thread_fiber = Fiber::ThreadToFiber(); thread_fibers[id] = thread_fiber; } void TestControl3::Exit() { - std::thread::id this_id = std::this_thread::get_id(); - u32 id = ids[this_id]; + const u32 id = thread_ids.Get(); thread_fibers[id]->Exit(); } @@ -290,7 +299,7 @@ static void ThreadStart3(u32 id, TestControl3& test_control) { } /** This test checks for one two threads racing for starting the same fiber. - * It checks execution occured in an ordered manner and by no time there were + * It checks execution occurred in an ordered manner and by no time there were * two contexts at the same time. */ TEST_CASE("Fibers::StartRace", "[common]") { diff --git a/src/tests/common/multi_level_queue.cpp b/src/tests/common/multi_level_queue.cpp deleted file mode 100644 index cca7ec7da..000000000 --- a/src/tests/common/multi_level_queue.cpp +++ /dev/null @@ -1,55 +0,0 @@ -// Copyright 2019 Yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <catch2/catch.hpp> -#include <math.h> -#include "common/common_types.h" -#include "common/multi_level_queue.h" - -namespace Common { - -TEST_CASE("MultiLevelQueue", "[common]") { - std::array<f32, 8> values = {0.0, 5.0, 1.0, 9.0, 8.0, 2.0, 6.0, 7.0}; - Common::MultiLevelQueue<f32, 64> mlq; - REQUIRE(mlq.empty()); - mlq.add(values[2], 2); - mlq.add(values[7], 7); - mlq.add(values[3], 3); - mlq.add(values[4], 4); - mlq.add(values[0], 0); - mlq.add(values[5], 5); - mlq.add(values[6], 6); - mlq.add(values[1], 1); - u32 index = 0; - bool all_set = true; - for (auto& f : mlq) { - all_set &= (f == values[index]); - index++; - } - REQUIRE(all_set); - REQUIRE(!mlq.empty()); - f32 v = 8.0; - mlq.add(v, 2); - v = -7.0; - mlq.add(v, 2, false); - REQUIRE(mlq.front(2) == -7.0); - mlq.yield(2); - REQUIRE(mlq.front(2) == values[2]); - REQUIRE(mlq.back(2) == -7.0); - REQUIRE(mlq.empty(8)); - v = 10.0; - mlq.add(v, 8); - mlq.adjust(v, 8, 9); - REQUIRE(mlq.front(9) == v); - REQUIRE(mlq.empty(8)); - REQUIRE(!mlq.empty(9)); - mlq.adjust(values[0], 0, 9); - REQUIRE(mlq.highest_priority_set() == 1); - REQUIRE(mlq.lowest_priority_set() == 9); - mlq.remove(values[1], 1); - REQUIRE(mlq.highest_priority_set() == 2); - REQUIRE(mlq.empty(1)); -} - -} // namespace Common diff --git a/src/tests/common/ring_buffer.cpp b/src/tests/common/ring_buffer.cpp index c883c4d56..54def22da 100644 --- a/src/tests/common/ring_buffer.cpp +++ b/src/tests/common/ring_buffer.cpp @@ -20,60 +20,60 @@ TEST_CASE("RingBuffer: Basic Tests", "[common]") { for (std::size_t i = 0; i < 4; i++) { const char elem = static_cast<char>(i); const std::size_t count = buf.Push(&elem, 1); - REQUIRE(count == 1); + REQUIRE(count == 1U); } - REQUIRE(buf.Size() == 4); + REQUIRE(buf.Size() == 4U); // Pushing values into a full ring buffer should fail. { const char elem = static_cast<char>(42); const std::size_t count = buf.Push(&elem, 1); - REQUIRE(count == 0); + REQUIRE(count == 0U); } - REQUIRE(buf.Size() == 4); + REQUIRE(buf.Size() == 4U); // Popping multiple values from a ring buffer with values should succeed. { const std::vector<char> popped = buf.Pop(2); - REQUIRE(popped.size() == 2); + REQUIRE(popped.size() == 2U); REQUIRE(popped[0] == 0); REQUIRE(popped[1] == 1); } - REQUIRE(buf.Size() == 2); + REQUIRE(buf.Size() == 2U); // Popping a single value from a ring buffer with values should succeed. { const std::vector<char> popped = buf.Pop(1); - REQUIRE(popped.size() == 1); + REQUIRE(popped.size() == 1U); REQUIRE(popped[0] == 2); } - REQUIRE(buf.Size() == 1); + REQUIRE(buf.Size() == 1U); // Pushing more values than space available should partially suceed. { std::vector<char> to_push(6); std::iota(to_push.begin(), to_push.end(), 88); const std::size_t count = buf.Push(to_push); - REQUIRE(count == 3); + REQUIRE(count == 3U); } - REQUIRE(buf.Size() == 4); + REQUIRE(buf.Size() == 4U); // Doing an unlimited pop should pop all values. { const std::vector<char> popped = buf.Pop(); - REQUIRE(popped.size() == 4); + REQUIRE(popped.size() == 4U); REQUIRE(popped[0] == 3); REQUIRE(popped[1] == 88); REQUIRE(popped[2] == 89); REQUIRE(popped[3] == 90); } - REQUIRE(buf.Size() == 0); + REQUIRE(buf.Size() == 0U); } TEST_CASE("RingBuffer: Threaded Test", "[common]") { @@ -93,7 +93,7 @@ TEST_CASE("RingBuffer: Threaded Test", "[common]") { std::size_t i = 0; while (i < count) { if (const std::size_t c = buf.Push(&value[0], 1); c > 0) { - REQUIRE(c == 1); + REQUIRE(c == 1U); i++; next_value(value); } else { @@ -108,7 +108,7 @@ TEST_CASE("RingBuffer: Threaded Test", "[common]") { std::size_t i = 0; while (i < count) { if (const std::vector<char> v = buf.Pop(1); v.size() > 0) { - REQUIRE(v.size() == 2); + REQUIRE(v.size() == 2U); REQUIRE(v[0] == value[0]); REQUIRE(v[1] == value[1]); i++; @@ -123,7 +123,7 @@ TEST_CASE("RingBuffer: Threaded Test", "[common]") { producer.join(); consumer.join(); - REQUIRE(buf.Size() == 0); + REQUIRE(buf.Size() == 0U); printf("RingBuffer: Threaded Test: full: %zu, empty: %zu\n", full, empty); } diff --git a/src/tests/core/arm/arm_test_common.cpp b/src/tests/core/arm/arm_test_common.cpp deleted file mode 100644 index e54674d11..000000000 --- a/src/tests/core/arm/arm_test_common.cpp +++ /dev/null @@ -1,145 +0,0 @@ -// Copyright 2016 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> - -#include "common/page_table.h" -#include "core/core.h" -#include "core/hle/kernel/memory/page_table.h" -#include "core/hle/kernel/process.h" -#include "core/memory.h" -#include "tests/core/arm/arm_test_common.h" - -namespace ArmTests { - -TestEnvironment::TestEnvironment(bool mutable_memory_) - : mutable_memory(mutable_memory_), - test_memory(std::make_shared<TestMemory>(this)), kernel{Core::System::GetInstance()} { - auto& system = Core::System::GetInstance(); - - auto process = Kernel::Process::Create(system, "", Kernel::Process::ProcessType::Userland); - page_table = &process->PageTable().PageTableImpl(); - - system.Memory().MapIoRegion(*page_table, 0x00000000, 0x80000000, test_memory); - system.Memory().MapIoRegion(*page_table, 0x80000000, 0x80000000, test_memory); - - kernel.MakeCurrentProcess(process.get()); -} - -TestEnvironment::~TestEnvironment() { - auto& system = Core::System::GetInstance(); - system.Memory().UnmapRegion(*page_table, 0x80000000, 0x80000000); - system.Memory().UnmapRegion(*page_table, 0x00000000, 0x80000000); -} - -void TestEnvironment::SetMemory64(VAddr vaddr, u64 value) { - SetMemory32(vaddr + 0, static_cast<u32>(value)); - SetMemory32(vaddr + 4, static_cast<u32>(value >> 32)); -} - -void TestEnvironment::SetMemory32(VAddr vaddr, u32 value) { - SetMemory16(vaddr + 0, static_cast<u16>(value)); - SetMemory16(vaddr + 2, static_cast<u16>(value >> 16)); -} - -void TestEnvironment::SetMemory16(VAddr vaddr, u16 value) { - SetMemory8(vaddr + 0, static_cast<u8>(value)); - SetMemory8(vaddr + 1, static_cast<u8>(value >> 8)); -} - -void TestEnvironment::SetMemory8(VAddr vaddr, u8 value) { - test_memory->data[vaddr] = value; -} - -std::vector<WriteRecord> TestEnvironment::GetWriteRecords() const { - return write_records; -} - -void TestEnvironment::ClearWriteRecords() { - write_records.clear(); -} - -TestEnvironment::TestMemory::~TestMemory() {} - -std::optional<bool> TestEnvironment::TestMemory::IsValidAddress(VAddr addr) { - return true; -} - -std::optional<u8> TestEnvironment::TestMemory::Read8(VAddr addr) { - const auto iter = data.find(addr); - - if (iter == data.end()) { - // Some arbitrary data - return static_cast<u8>(addr); - } - - return iter->second; -} - -std::optional<u16> TestEnvironment::TestMemory::Read16(VAddr addr) { - return *Read8(addr) | static_cast<u16>(*Read8(addr + 1)) << 8; -} - -std::optional<u32> TestEnvironment::TestMemory::Read32(VAddr addr) { - return *Read16(addr) | static_cast<u32>(*Read16(addr + 2)) << 16; -} - -std::optional<u64> TestEnvironment::TestMemory::Read64(VAddr addr) { - return *Read32(addr) | static_cast<u64>(*Read32(addr + 4)) << 32; -} - -bool TestEnvironment::TestMemory::ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) { - VAddr addr = src_addr; - u8* data = static_cast<u8*>(dest_buffer); - - for (std::size_t i = 0; i < size; i++, addr++, data++) { - *data = *Read8(addr); - } - - return true; -} - -bool TestEnvironment::TestMemory::Write8(VAddr addr, u8 data) { - env->write_records.emplace_back(8, addr, data); - if (env->mutable_memory) - env->SetMemory8(addr, data); - return true; -} - -bool TestEnvironment::TestMemory::Write16(VAddr addr, u16 data) { - env->write_records.emplace_back(16, addr, data); - if (env->mutable_memory) - env->SetMemory16(addr, data); - return true; -} - -bool TestEnvironment::TestMemory::Write32(VAddr addr, u32 data) { - env->write_records.emplace_back(32, addr, data); - if (env->mutable_memory) - env->SetMemory32(addr, data); - return true; -} - -bool TestEnvironment::TestMemory::Write64(VAddr addr, u64 data) { - env->write_records.emplace_back(64, addr, data); - if (env->mutable_memory) - env->SetMemory64(addr, data); - return true; -} - -bool TestEnvironment::TestMemory::WriteBlock(VAddr dest_addr, const void* src_buffer, - std::size_t size) { - VAddr addr = dest_addr; - const u8* data = static_cast<const u8*>(src_buffer); - - for (std::size_t i = 0; i < size; i++, addr++, data++) { - env->write_records.emplace_back(8, addr, *data); - if (env->mutable_memory) - env->SetMemory8(addr, *data); - } - - return true; -} - -} // namespace ArmTests diff --git a/src/tests/core/arm/arm_test_common.h b/src/tests/core/arm/arm_test_common.h deleted file mode 100644 index d145dbfcc..000000000 --- a/src/tests/core/arm/arm_test_common.h +++ /dev/null @@ -1,93 +0,0 @@ -// Copyright 2016 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <tuple> -#include <unordered_map> -#include <vector> - -#include "common/common_types.h" -#include "common/memory_hook.h" -#include "core/hle/kernel/kernel.h" - -namespace Common { -struct PageTable; -} - -namespace ArmTests { - -struct WriteRecord { - WriteRecord(std::size_t size, VAddr addr, u64 data) : size(size), addr(addr), data(data) {} - std::size_t size; - VAddr addr; - u64 data; - bool operator==(const WriteRecord& o) const { - return std::tie(size, addr, data) == std::tie(o.size, o.addr, o.data); - } -}; - -class TestEnvironment final { -public: - /* - * Inititalise test environment - * @param mutable_memory If false, writes to memory can never be read back. - * (Memory is immutable.) - */ - explicit TestEnvironment(bool mutable_memory = false); - - /// Shutdown test environment - ~TestEnvironment(); - - /// Sets value at memory location vaddr. - void SetMemory8(VAddr vaddr, u8 value); - void SetMemory16(VAddr vaddr, u16 value); - void SetMemory32(VAddr vaddr, u32 value); - void SetMemory64(VAddr vaddr, u64 value); - - /** - * Whenever Memory::Write{8,16,32,64} is called within the test environment, - * a new write-record is made. - * @returns A vector of write records made since they were last cleared. - */ - std::vector<WriteRecord> GetWriteRecords() const; - - /// Empties the internal write-record store. - void ClearWriteRecords(); - -private: - friend struct TestMemory; - struct TestMemory final : Common::MemoryHook { - explicit TestMemory(TestEnvironment* env_) : env(env_) {} - TestEnvironment* env; - - ~TestMemory() override; - - std::optional<bool> IsValidAddress(VAddr addr) override; - - std::optional<u8> Read8(VAddr addr) override; - std::optional<u16> Read16(VAddr addr) override; - std::optional<u32> Read32(VAddr addr) override; - std::optional<u64> Read64(VAddr addr) override; - - bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) override; - - bool Write8(VAddr addr, u8 data) override; - bool Write16(VAddr addr, u16 data) override; - bool Write32(VAddr addr, u32 data) override; - bool Write64(VAddr addr, u64 data) override; - - bool WriteBlock(VAddr dest_addr, const void* src_buffer, std::size_t size) override; - - std::unordered_map<VAddr, u8> data; - }; - - bool mutable_memory; - std::shared_ptr<TestMemory> test_memory; - std::vector<WriteRecord> write_records; - Common::PageTable* page_table = nullptr; - Kernel::KernelCore kernel; -}; - -} // namespace ArmTests diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index da9e9fdda..f7b9d7f86 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -5,8 +5,27 @@ add_library(video_core STATIC buffer_cache/buffer_cache.h buffer_cache/map_interval.cpp buffer_cache/map_interval.h + cdma_pusher.cpp + cdma_pusher.h + command_classes/codecs/codec.cpp + command_classes/codecs/codec.h + command_classes/codecs/h264.cpp + command_classes/codecs/h264.h + command_classes/codecs/vp9.cpp + command_classes/codecs/vp9.h + command_classes/codecs/vp9_types.h + command_classes/host1x.cpp + command_classes/host1x.h + command_classes/nvdec.cpp + command_classes/nvdec.h + command_classes/nvdec_common.h + command_classes/sync_manager.cpp + command_classes/sync_manager.h + command_classes/vic.cpp + command_classes/vic.h compatible_formats.cpp compatible_formats.h + delayed_destruction_ring.h dirty_flags.cpp dirty_flags.h dma_pusher.cpp @@ -29,6 +48,7 @@ add_library(video_core STATIC engines/shader_bytecode.h engines/shader_header.h engines/shader_type.h + framebuffer_config.h macro/macro.cpp macro/macro.h macro/macro_hle.cpp @@ -40,10 +60,6 @@ add_library(video_core STATIC fence_manager.h gpu.cpp gpu.h - gpu_asynch.cpp - gpu_asynch.h - gpu_synch.cpp - gpu_synch.h gpu_thread.cpp gpu_thread.h guest_driver.cpp @@ -66,14 +82,10 @@ add_library(video_core STATIC renderer_opengl/gl_device.h renderer_opengl/gl_fence_manager.cpp renderer_opengl/gl_fence_manager.h - renderer_opengl/gl_framebuffer_cache.cpp - renderer_opengl/gl_framebuffer_cache.h renderer_opengl/gl_rasterizer.cpp renderer_opengl/gl_rasterizer.h renderer_opengl/gl_resource_manager.cpp renderer_opengl/gl_resource_manager.h - renderer_opengl/gl_sampler_cache.cpp - renderer_opengl/gl_sampler_cache.h renderer_opengl/gl_shader_cache.cpp renderer_opengl/gl_shader_cache.h renderer_opengl/gl_shader_decompiler.cpp @@ -95,10 +107,62 @@ add_library(video_core STATIC renderer_opengl/maxwell_to_gl.h renderer_opengl/renderer_opengl.cpp renderer_opengl/renderer_opengl.h - renderer_opengl/utils.cpp - renderer_opengl/utils.h - sampler_cache.cpp - sampler_cache.h + renderer_opengl/util_shaders.cpp + renderer_opengl/util_shaders.h + renderer_vulkan/blit_image.cpp + renderer_vulkan/blit_image.h + renderer_vulkan/fixed_pipeline_state.cpp + renderer_vulkan/fixed_pipeline_state.h + renderer_vulkan/maxwell_to_vk.cpp + renderer_vulkan/maxwell_to_vk.h + renderer_vulkan/renderer_vulkan.h + renderer_vulkan/renderer_vulkan.cpp + renderer_vulkan/vk_blit_screen.cpp + renderer_vulkan/vk_blit_screen.h + renderer_vulkan/vk_buffer_cache.cpp + renderer_vulkan/vk_buffer_cache.h + renderer_vulkan/vk_command_pool.cpp + renderer_vulkan/vk_command_pool.h + renderer_vulkan/vk_compute_pass.cpp + renderer_vulkan/vk_compute_pass.h + renderer_vulkan/vk_compute_pipeline.cpp + renderer_vulkan/vk_compute_pipeline.h + renderer_vulkan/vk_descriptor_pool.cpp + renderer_vulkan/vk_descriptor_pool.h + renderer_vulkan/vk_fence_manager.cpp + renderer_vulkan/vk_fence_manager.h + renderer_vulkan/vk_graphics_pipeline.cpp + renderer_vulkan/vk_graphics_pipeline.h + renderer_vulkan/vk_master_semaphore.cpp + renderer_vulkan/vk_master_semaphore.h + renderer_vulkan/vk_memory_manager.cpp + renderer_vulkan/vk_memory_manager.h + renderer_vulkan/vk_pipeline_cache.cpp + renderer_vulkan/vk_pipeline_cache.h + renderer_vulkan/vk_query_cache.cpp + renderer_vulkan/vk_query_cache.h + renderer_vulkan/vk_rasterizer.cpp + renderer_vulkan/vk_rasterizer.h + renderer_vulkan/vk_resource_pool.cpp + renderer_vulkan/vk_resource_pool.h + renderer_vulkan/vk_scheduler.cpp + renderer_vulkan/vk_scheduler.h + renderer_vulkan/vk_shader_decompiler.cpp + renderer_vulkan/vk_shader_decompiler.h + renderer_vulkan/vk_shader_util.cpp + renderer_vulkan/vk_shader_util.h + renderer_vulkan/vk_staging_buffer_pool.cpp + renderer_vulkan/vk_staging_buffer_pool.h + renderer_vulkan/vk_state_tracker.cpp + renderer_vulkan/vk_state_tracker.h + renderer_vulkan/vk_stream_buffer.cpp + renderer_vulkan/vk_stream_buffer.h + renderer_vulkan/vk_swapchain.cpp + renderer_vulkan/vk_swapchain.h + renderer_vulkan/vk_texture_cache.cpp + renderer_vulkan/vk_texture_cache.h + renderer_vulkan/vk_update_descriptor.cpp + renderer_vulkan/vk_update_descriptor.h shader_cache.h shader_notify.cpp shader_notify.h @@ -155,109 +219,71 @@ add_library(video_core STATIC shader/transform_feedback.h surface.cpp surface.h + texture_cache/accelerated_swizzle.cpp + texture_cache/accelerated_swizzle.h + texture_cache/decode_bc4.cpp + texture_cache/decode_bc4.h + texture_cache/descriptor_table.h + texture_cache/formatter.cpp + texture_cache/formatter.h texture_cache/format_lookup_table.cpp texture_cache/format_lookup_table.h - texture_cache/surface_base.cpp - texture_cache/surface_base.h - texture_cache/surface_params.cpp - texture_cache/surface_params.h - texture_cache/surface_view.cpp - texture_cache/surface_view.h + texture_cache/image_base.cpp + texture_cache/image_base.h + texture_cache/image_info.cpp + texture_cache/image_info.h + texture_cache/image_view_base.cpp + texture_cache/image_view_base.h + texture_cache/image_view_info.cpp + texture_cache/image_view_info.h + texture_cache/render_targets.h + texture_cache/samples_helper.h + texture_cache/slot_vector.h texture_cache/texture_cache.h + texture_cache/types.h + texture_cache/util.cpp + texture_cache/util.h textures/astc.cpp textures/astc.h - textures/convert.cpp - textures/convert.h textures/decoders.cpp textures/decoders.h textures/texture.cpp textures/texture.h video_core.cpp video_core.h + vulkan_common/vulkan_debug_callback.cpp + vulkan_common/vulkan_debug_callback.h + vulkan_common/vulkan_device.cpp + vulkan_common/vulkan_device.h + vulkan_common/vulkan_instance.cpp + vulkan_common/vulkan_instance.h + vulkan_common/vulkan_library.cpp + vulkan_common/vulkan_library.h + vulkan_common/vulkan_surface.cpp + vulkan_common/vulkan_surface.h + vulkan_common/vulkan_wrapper.cpp + vulkan_common/vulkan_wrapper.h + vulkan_common/nsight_aftermath_tracker.cpp + vulkan_common/nsight_aftermath_tracker.h ) -if (ENABLE_VULKAN) - target_sources(video_core PRIVATE - renderer_vulkan/fixed_pipeline_state.cpp - renderer_vulkan/fixed_pipeline_state.h - renderer_vulkan/maxwell_to_vk.cpp - renderer_vulkan/maxwell_to_vk.h - renderer_vulkan/nsight_aftermath_tracker.cpp - renderer_vulkan/nsight_aftermath_tracker.h - renderer_vulkan/renderer_vulkan.h - renderer_vulkan/renderer_vulkan.cpp - renderer_vulkan/vk_blit_screen.cpp - renderer_vulkan/vk_blit_screen.h - renderer_vulkan/vk_buffer_cache.cpp - renderer_vulkan/vk_buffer_cache.h - renderer_vulkan/vk_command_pool.cpp - renderer_vulkan/vk_command_pool.h - renderer_vulkan/vk_compute_pass.cpp - renderer_vulkan/vk_compute_pass.h - renderer_vulkan/vk_compute_pipeline.cpp - renderer_vulkan/vk_compute_pipeline.h - renderer_vulkan/vk_descriptor_pool.cpp - renderer_vulkan/vk_descriptor_pool.h - renderer_vulkan/vk_device.cpp - renderer_vulkan/vk_device.h - renderer_vulkan/vk_fence_manager.cpp - renderer_vulkan/vk_fence_manager.h - renderer_vulkan/vk_graphics_pipeline.cpp - renderer_vulkan/vk_graphics_pipeline.h - renderer_vulkan/vk_image.cpp - renderer_vulkan/vk_image.h - renderer_vulkan/vk_master_semaphore.cpp - renderer_vulkan/vk_master_semaphore.h - renderer_vulkan/vk_memory_manager.cpp - renderer_vulkan/vk_memory_manager.h - renderer_vulkan/vk_pipeline_cache.cpp - renderer_vulkan/vk_pipeline_cache.h - renderer_vulkan/vk_query_cache.cpp - renderer_vulkan/vk_query_cache.h - renderer_vulkan/vk_rasterizer.cpp - renderer_vulkan/vk_rasterizer.h - renderer_vulkan/vk_renderpass_cache.cpp - renderer_vulkan/vk_renderpass_cache.h - renderer_vulkan/vk_resource_pool.cpp - renderer_vulkan/vk_resource_pool.h - renderer_vulkan/vk_sampler_cache.cpp - renderer_vulkan/vk_sampler_cache.h - renderer_vulkan/vk_scheduler.cpp - renderer_vulkan/vk_scheduler.h - renderer_vulkan/vk_shader_decompiler.cpp - renderer_vulkan/vk_shader_decompiler.h - renderer_vulkan/vk_shader_util.cpp - renderer_vulkan/vk_shader_util.h - renderer_vulkan/vk_staging_buffer_pool.cpp - renderer_vulkan/vk_staging_buffer_pool.h - renderer_vulkan/vk_state_tracker.cpp - renderer_vulkan/vk_state_tracker.h - renderer_vulkan/vk_stream_buffer.cpp - renderer_vulkan/vk_stream_buffer.h - renderer_vulkan/vk_swapchain.cpp - renderer_vulkan/vk_swapchain.h - renderer_vulkan/vk_texture_cache.cpp - renderer_vulkan/vk_texture_cache.h - renderer_vulkan/vk_update_descriptor.cpp - renderer_vulkan/vk_update_descriptor.h - renderer_vulkan/wrapper.cpp - renderer_vulkan/wrapper.h - ) -endif() - create_target_directory_groups(video_core) target_link_libraries(video_core PUBLIC common core) target_link_libraries(video_core PRIVATE glad xbyak) +if (MSVC) + target_include_directories(video_core PRIVATE ${FFMPEG_INCLUDE_DIR}) + target_link_libraries(video_core PUBLIC ${FFMPEG_LIBRARY_DIR}/swscale.lib ${FFMPEG_LIBRARY_DIR}/avcodec.lib ${FFMPEG_LIBRARY_DIR}/avutil.lib) +else() + target_include_directories(video_core PRIVATE ${FFMPEG_INCLUDE_DIR}) + target_link_libraries(video_core PRIVATE ${FFMPEG_LIBRARIES}) +endif() + add_dependencies(video_core host_shaders) target_include_directories(video_core PRIVATE ${HOST_SHADERS_INCLUDE}) - -if (ENABLE_VULKAN) - target_include_directories(video_core PRIVATE sirit ../../externals/Vulkan-Headers/include) - target_compile_definitions(video_core PRIVATE HAS_VULKAN) - target_link_libraries(video_core PRIVATE sirit) -endif() +target_include_directories(video_core PRIVATE sirit ../../externals/Vulkan-Headers/include) +target_link_libraries(video_core PRIVATE sirit) if (ENABLE_NSIGHT_AFTERMATH) if (NOT DEFINED ENV{NSIGHT_AFTERMATH_SDK}) @@ -271,7 +297,27 @@ if (ENABLE_NSIGHT_AFTERMATH) endif() if (MSVC) - target_compile_options(video_core PRIVATE /we4267) + target_compile_options(video_core PRIVATE + /we4267 # 'var' : conversion from 'size_t' to 'type', possible loss of data + /we4456 # Declaration of 'identifier' hides previous local declaration + /we4457 # Declaration of 'identifier' hides function parameter + /we4458 # Declaration of 'identifier' hides class member + /we4459 # Declaration of 'identifier' hides global declaration + /we4715 # 'function' : not all control paths return a value + ) else() - target_compile_options(video_core PRIVATE -Werror=conversion -Wno-error=sign-conversion -Werror=switch) + target_compile_options(video_core PRIVATE + -Werror=conversion + -Wno-error=sign-conversion + -Werror=pessimizing-move + -Werror=redundant-move + -Werror=shadow + -Werror=switch + -Werror=type-limits + -Werror=unused-variable + + $<$<CXX_COMPILER_ID:GNU>:-Werror=class-memaccess> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-parameter> + $<$<CXX_COMPILER_ID:GNU>:-Werror=unused-but-set-variable> + ) endif() diff --git a/src/video_core/buffer_cache/buffer_block.h b/src/video_core/buffer_cache/buffer_block.h index e64170e66..e9306194a 100644 --- a/src/video_core/buffer_cache/buffer_block.h +++ b/src/video_core/buffer_cache/buffer_block.h @@ -4,34 +4,29 @@ #pragma once -#include <unordered_set> -#include <utility> - -#include "common/alignment.h" #include "common/common_types.h" -#include "video_core/gpu.h" namespace VideoCommon { class BufferBlock { public: - bool Overlaps(VAddr start, VAddr end) const { + [[nodiscard]] bool Overlaps(VAddr start, VAddr end) const { return (cpu_addr < end) && (cpu_addr_end > start); } - bool IsInside(VAddr other_start, VAddr other_end) const { + [[nodiscard]] bool IsInside(VAddr other_start, VAddr other_end) const { return cpu_addr <= other_start && other_end <= cpu_addr_end; } - std::size_t Offset(VAddr in_addr) const { + [[nodiscard]] std::size_t Offset(VAddr in_addr) const { return static_cast<std::size_t>(in_addr - cpu_addr); } - VAddr CpuAddr() const { + [[nodiscard]] VAddr CpuAddr() const { return cpu_addr; } - VAddr CpuAddrEnd() const { + [[nodiscard]] VAddr CpuAddrEnd() const { return cpu_addr_end; } @@ -40,11 +35,11 @@ public: cpu_addr_end = new_addr + size; } - std::size_t Size() const { + [[nodiscard]] std::size_t Size() const { return size; } - u64 Epoch() const { + [[nodiscard]] u64 Epoch() const { return epoch; } diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index e7edd733f..83b9ee871 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -118,20 +118,17 @@ public: /// Prepares the buffer cache for data uploading /// @param max_size Maximum number of bytes that will be uploaded /// @return True when a stream buffer invalidation was required, false otherwise - bool Map(std::size_t max_size) { + void Map(std::size_t max_size) { std::lock_guard lock{mutex}; - bool invalidated; - std::tie(buffer_ptr, buffer_offset_base, invalidated) = stream_buffer->Map(max_size, 4); + std::tie(buffer_ptr, buffer_offset_base) = stream_buffer.Map(max_size, 4); buffer_offset = buffer_offset_base; - - return invalidated; } /// Finishes the upload stream void Unmap() { std::lock_guard lock{mutex}; - stream_buffer->Unmap(buffer_offset - buffer_offset_base); + stream_buffer.Unmap(buffer_offset - buffer_offset_base); } /// Function called at the end of each frame, inteded for deferred operations @@ -261,9 +258,9 @@ public: protected: explicit BufferCache(VideoCore::RasterizerInterface& rasterizer_, Tegra::MemoryManager& gpu_memory_, Core::Memory::Memory& cpu_memory_, - std::unique_ptr<StreamBuffer> stream_buffer_) + StreamBuffer& stream_buffer_) : rasterizer{rasterizer_}, gpu_memory{gpu_memory_}, cpu_memory{cpu_memory_}, - stream_buffer{std::move(stream_buffer_)}, stream_buffer_handle{stream_buffer->Handle()} {} + stream_buffer{stream_buffer_} {} ~BufferCache() = default; @@ -441,7 +438,7 @@ private: buffer_ptr += size; buffer_offset += size; - return BufferInfo{stream_buffer->Handle(), uploaded_offset, stream_buffer->Address()}; + return BufferInfo{stream_buffer.Handle(), uploaded_offset, stream_buffer.Address()}; } void AlignBuffer(std::size_t alignment) { @@ -545,7 +542,7 @@ private: bool IsRegionWritten(VAddr start, VAddr end) const { const u64 page_end = end >> WRITE_PAGE_BIT; for (u64 page_start = start >> WRITE_PAGE_BIT; page_start <= page_end; ++page_start) { - if (written_pages.count(page_start) > 0) { + if (written_pages.contains(page_start)) { return true; } } @@ -567,9 +564,7 @@ private: VideoCore::RasterizerInterface& rasterizer; Tegra::MemoryManager& gpu_memory; Core::Memory::Memory& cpu_memory; - - std::unique_ptr<StreamBuffer> stream_buffer; - BufferType stream_buffer_handle; + StreamBuffer& stream_buffer; u8* buffer_ptr = nullptr; u64 buffer_offset = 0; diff --git a/src/video_core/buffer_cache/map_interval.h b/src/video_core/buffer_cache/map_interval.h index fe0bcd1d8..ef974b08a 100644 --- a/src/video_core/buffer_cache/map_interval.h +++ b/src/video_core/buffer_cache/map_interval.h @@ -84,9 +84,10 @@ private: void FillFreeList(Chunk& chunk); std::vector<MapInterval*> free_list; - std::unique_ptr<Chunk>* new_chunk = &first_chunk.next; Chunk first_chunk; + + std::unique_ptr<Chunk>* new_chunk = &first_chunk.next; }; } // namespace VideoCommon diff --git a/src/video_core/cdma_pusher.cpp b/src/video_core/cdma_pusher.cpp new file mode 100644 index 000000000..94679d5d1 --- /dev/null +++ b/src/video_core/cdma_pusher.cpp @@ -0,0 +1,170 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include "command_classes/host1x.h" +#include "command_classes/nvdec.h" +#include "command_classes/vic.h" +#include "common/bit_util.h" +#include "video_core/cdma_pusher.h" +#include "video_core/command_classes/nvdec_common.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +namespace Tegra { +CDmaPusher::CDmaPusher(GPU& gpu_) + : gpu{gpu_}, nvdec_processor(std::make_shared<Nvdec>(gpu)), + vic_processor(std::make_unique<Vic>(gpu, nvdec_processor)), + host1x_processor(std::make_unique<Host1x>(gpu)), + sync_manager(std::make_unique<SyncptIncrManager>(gpu)) {} + +CDmaPusher::~CDmaPusher() = default; + +void CDmaPusher::Push(ChCommandHeaderList&& entries) { + cdma_queue.push(std::move(entries)); +} + +void CDmaPusher::DispatchCalls() { + while (!cdma_queue.empty()) { + Step(); + } +} + +void CDmaPusher::Step() { + const auto entries{cdma_queue.front()}; + cdma_queue.pop(); + + std::vector<u32> values(entries.size()); + std::memcpy(values.data(), entries.data(), entries.size() * sizeof(u32)); + + for (const u32 value : values) { + if (mask != 0) { + const u32 lbs = Common::CountTrailingZeroes32(mask); + mask &= ~(1U << lbs); + ExecuteCommand(static_cast<u32>(offset + lbs), value); + continue; + } else if (count != 0) { + --count; + ExecuteCommand(static_cast<u32>(offset), value); + if (incrementing) { + ++offset; + } + continue; + } + const auto mode = static_cast<ChSubmissionMode>((value >> 28) & 0xf); + switch (mode) { + case ChSubmissionMode::SetClass: { + mask = value & 0x3f; + offset = (value >> 16) & 0xfff; + current_class = static_cast<ChClassId>((value >> 6) & 0x3ff); + break; + } + case ChSubmissionMode::Incrementing: + case ChSubmissionMode::NonIncrementing: + count = value & 0xffff; + offset = (value >> 16) & 0xfff; + incrementing = mode == ChSubmissionMode::Incrementing; + break; + case ChSubmissionMode::Mask: + mask = value & 0xffff; + offset = (value >> 16) & 0xfff; + break; + case ChSubmissionMode::Immediate: { + const u32 data = value & 0xfff; + offset = (value >> 16) & 0xfff; + ExecuteCommand(static_cast<u32>(offset), data); + break; + } + default: + UNIMPLEMENTED_MSG("ChSubmission mode {} is not implemented!", static_cast<u32>(mode)); + break; + } + } +} + +void CDmaPusher::ExecuteCommand(u32 state_offset, u32 data) { + switch (current_class) { + case ChClassId::NvDec: + ThiStateWrite(nvdec_thi_state, state_offset, {data}); + switch (static_cast<ThiMethod>(state_offset)) { + case ThiMethod::IncSyncpt: { + LOG_DEBUG(Service_NVDRV, "NVDEC Class IncSyncpt Method"); + const auto syncpoint_id = static_cast<u32>(data & 0xFF); + const auto cond = static_cast<u32>((data >> 8) & 0xFF); + if (cond == 0) { + sync_manager->Increment(syncpoint_id); + } else { + sync_manager->SignalDone( + sync_manager->IncrementWhenDone(static_cast<u32>(current_class), syncpoint_id)); + } + break; + } + case ThiMethod::SetMethod1: + LOG_DEBUG(Service_NVDRV, "NVDEC method 0x{:X}", + static_cast<u32>(nvdec_thi_state.method_0)); + nvdec_processor->ProcessMethod(static_cast<Nvdec::Method>(nvdec_thi_state.method_0), + {data}); + break; + default: + break; + } + break; + case ChClassId::GraphicsVic: + ThiStateWrite(vic_thi_state, static_cast<u32>(state_offset), {data}); + switch (static_cast<ThiMethod>(state_offset)) { + case ThiMethod::IncSyncpt: { + LOG_DEBUG(Service_NVDRV, "VIC Class IncSyncpt Method"); + const auto syncpoint_id = static_cast<u32>(data & 0xFF); + const auto cond = static_cast<u32>((data >> 8) & 0xFF); + if (cond == 0) { + sync_manager->Increment(syncpoint_id); + } else { + sync_manager->SignalDone( + sync_manager->IncrementWhenDone(static_cast<u32>(current_class), syncpoint_id)); + } + break; + } + case ThiMethod::SetMethod1: + LOG_DEBUG(Service_NVDRV, "VIC method 0x{:X}, Args=({})", + static_cast<u32>(vic_thi_state.method_0), data); + vic_processor->ProcessMethod(static_cast<Vic::Method>(vic_thi_state.method_0), {data}); + break; + default: + break; + } + break; + case ChClassId::Host1x: + // This device is mainly for syncpoint synchronization + LOG_DEBUG(Service_NVDRV, "Host1X Class Method"); + host1x_processor->ProcessMethod(static_cast<Host1x::Method>(state_offset), {data}); + break; + default: + UNIMPLEMENTED_MSG("Current class not implemented {:X}", static_cast<u32>(current_class)); + break; + } +} + +void CDmaPusher::ThiStateWrite(ThiRegisters& state, u32 state_offset, + const std::vector<u32>& arguments) { + u8* const state_offset_ptr = reinterpret_cast<u8*>(&state) + sizeof(u32) * state_offset; + std::memcpy(state_offset_ptr, arguments.data(), sizeof(u32) * arguments.size()); +} + +} // namespace Tegra diff --git a/src/video_core/cdma_pusher.h b/src/video_core/cdma_pusher.h new file mode 100644 index 000000000..8ca70b6dd --- /dev/null +++ b/src/video_core/cdma_pusher.h @@ -0,0 +1,136 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <unordered_map> +#include <vector> +#include <queue> + +#include "common/bit_field.h" +#include "common/common_types.h" +#include "video_core/command_classes/sync_manager.h" + +namespace Tegra { + +class GPU; +class Nvdec; +class Vic; +class Host1x; + +enum class ChSubmissionMode : u32 { + SetClass = 0, + Incrementing = 1, + NonIncrementing = 2, + Mask = 3, + Immediate = 4, + Restart = 5, + Gather = 6, +}; + +enum class ChClassId : u32 { + NoClass = 0x0, + Host1x = 0x1, + VideoEncodeMpeg = 0x20, + VideoEncodeNvEnc = 0x21, + VideoStreamingVi = 0x30, + VideoStreamingIsp = 0x32, + VideoStreamingIspB = 0x34, + VideoStreamingViI2c = 0x36, + GraphicsVic = 0x5d, + Graphics3D = 0x60, + GraphicsGpu = 0x61, + Tsec = 0xe0, + TsecB = 0xe1, + NvJpg = 0xc0, + NvDec = 0xf0 +}; + +enum class ChMethod : u32 { + Empty = 0, + SetMethod = 0x10, + SetData = 0x11, +}; + +union ChCommandHeader { + u32 raw; + BitField<0, 16, u32> value; + BitField<16, 12, ChMethod> method_offset; + BitField<28, 4, ChSubmissionMode> submission_mode; +}; +static_assert(sizeof(ChCommandHeader) == sizeof(u32), "ChCommand header is an invalid size"); + +struct ChCommand { + ChClassId class_id{}; + int method_offset{}; + std::vector<u32> arguments; +}; + +using ChCommandHeaderList = std::vector<ChCommandHeader>; +using ChCommandList = std::vector<ChCommand>; + +struct ThiRegisters { + u32_le increment_syncpt{}; + INSERT_PADDING_WORDS(1); + u32_le increment_syncpt_error{}; + u32_le ctx_switch_incremement_syncpt{}; + INSERT_PADDING_WORDS(4); + u32_le ctx_switch{}; + INSERT_PADDING_WORDS(1); + u32_le ctx_syncpt_eof{}; + INSERT_PADDING_WORDS(5); + u32_le method_0{}; + u32_le method_1{}; + INSERT_PADDING_WORDS(12); + u32_le int_status{}; + u32_le int_mask{}; +}; + +enum class ThiMethod : u32 { + IncSyncpt = offsetof(ThiRegisters, increment_syncpt) / sizeof(u32), + SetMethod0 = offsetof(ThiRegisters, method_0) / sizeof(u32), + SetMethod1 = offsetof(ThiRegisters, method_1) / sizeof(u32), +}; + +class CDmaPusher { +public: + explicit CDmaPusher(GPU& gpu_); + ~CDmaPusher(); + + /// Push NVDEC command buffer entries into queue + void Push(ChCommandHeaderList&& entries); + + /// Process queued command buffer entries + void DispatchCalls(); + + /// Process one queue element + void Step(); + + /// Invoke command class devices to execute the command based on the current state + void ExecuteCommand(u32 state_offset, u32 data); + +private: + /// Write arguments value to the ThiRegisters member at the specified offset + void ThiStateWrite(ThiRegisters& state, u32 state_offset, const std::vector<u32>& arguments); + + GPU& gpu; + std::shared_ptr<Tegra::Nvdec> nvdec_processor; + std::unique_ptr<Tegra::Vic> vic_processor; + std::unique_ptr<Tegra::Host1x> host1x_processor; + std::unique_ptr<SyncptIncrManager> sync_manager; + ChClassId current_class{}; + ThiRegisters vic_thi_state{}; + ThiRegisters nvdec_thi_state{}; + + s32 count{}; + s32 offset{}; + s32 mask{}; + bool incrementing{}; + + // Queue of command lists to be processed + std::queue<ChCommandHeaderList> cdma_queue; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/codec.cpp b/src/video_core/command_classes/codecs/codec.cpp new file mode 100644 index 000000000..39bc923a5 --- /dev/null +++ b/src/video_core/command_classes/codecs/codec.cpp @@ -0,0 +1,129 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> +#include <fstream> +#include <vector> +#include "common/assert.h" +#include "video_core/command_classes/codecs/codec.h" +#include "video_core/command_classes/codecs/h264.h" +#include "video_core/command_classes/codecs/vp9.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +extern "C" { +#include <libavutil/opt.h> +} + +namespace Tegra { + +void AVFrameDeleter(AVFrame* ptr) { + av_frame_unref(ptr); + av_free(ptr); +} + +Codec::Codec(GPU& gpu_) + : gpu(gpu_), h264_decoder(std::make_unique<Decoder::H264>(gpu)), + vp9_decoder(std::make_unique<Decoder::VP9>(gpu)) {} + +Codec::~Codec() { + if (!initialized) { + return; + } + // Free libav memory + AVFrame* av_frame{nullptr}; + avcodec_send_packet(av_codec_ctx, nullptr); + av_frame = av_frame_alloc(); + avcodec_receive_frame(av_codec_ctx, av_frame); + avcodec_flush_buffers(av_codec_ctx); + + av_frame_unref(av_frame); + av_free(av_frame); + avcodec_close(av_codec_ctx); +} + +void Codec::SetTargetCodec(NvdecCommon::VideoCodec codec) { + LOG_INFO(Service_NVDRV, "NVDEC video codec initialized to {}", codec); + current_codec = codec; +} + +void Codec::StateWrite(u32 offset, u64 arguments) { + u8* const state_offset = reinterpret_cast<u8*>(&state) + offset * sizeof(u64); + std::memcpy(state_offset, &arguments, sizeof(u64)); +} + +void Codec::Decode() { + bool is_first_frame = false; + + if (!initialized) { + if (current_codec == NvdecCommon::VideoCodec::H264) { + av_codec = avcodec_find_decoder(AV_CODEC_ID_H264); + } else if (current_codec == NvdecCommon::VideoCodec::Vp9) { + av_codec = avcodec_find_decoder(AV_CODEC_ID_VP9); + } else { + LOG_ERROR(Service_NVDRV, "Unknown video codec {}", current_codec); + return; + } + + av_codec_ctx = avcodec_alloc_context3(av_codec); + av_opt_set(av_codec_ctx->priv_data, "tune", "zerolatency", 0); + + // TODO(ameerj): libavcodec gpu hw acceleration + + const auto av_error = avcodec_open2(av_codec_ctx, av_codec, nullptr); + if (av_error < 0) { + LOG_ERROR(Service_NVDRV, "avcodec_open2() Failed."); + avcodec_close(av_codec_ctx); + return; + } + initialized = true; + is_first_frame = true; + } + bool vp9_hidden_frame = false; + + AVPacket packet{}; + av_init_packet(&packet); + std::vector<u8> frame_data; + + if (current_codec == NvdecCommon::VideoCodec::H264) { + frame_data = h264_decoder->ComposeFrameHeader(state, is_first_frame); + } else if (current_codec == NvdecCommon::VideoCodec::Vp9) { + frame_data = vp9_decoder->ComposeFrameHeader(state); + vp9_hidden_frame = vp9_decoder->WasFrameHidden(); + } + + packet.data = frame_data.data(); + packet.size = static_cast<int>(frame_data.size()); + + avcodec_send_packet(av_codec_ctx, &packet); + + if (!vp9_hidden_frame) { + // Only receive/store visible frames + AVFramePtr frame = AVFramePtr{av_frame_alloc(), AVFrameDeleter}; + avcodec_receive_frame(av_codec_ctx, frame.get()); + av_frames.push(std::move(frame)); + // Limit queue to 10 frames. Workaround for ZLA decode and queue spam + if (av_frames.size() > 10) { + av_frames.pop(); + } + } +} + +AVFramePtr Codec::GetCurrentFrame() { + // Sometimes VIC will request more frames than have been decoded. + // in this case, return a nullptr and don't overwrite previous frame data + if (av_frames.empty()) { + return AVFramePtr{nullptr, AVFrameDeleter}; + } + + AVFramePtr frame = std::move(av_frames.front()); + av_frames.pop(); + return frame; +} + +NvdecCommon::VideoCodec Codec::GetCurrentCodec() const { + return current_codec; +} + +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/codec.h b/src/video_core/command_classes/codecs/codec.h new file mode 100644 index 000000000..8a2a6c360 --- /dev/null +++ b/src/video_core/command_classes/codecs/codec.h @@ -0,0 +1,70 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <queue> +#include "common/common_types.h" +#include "video_core/command_classes/nvdec_common.h" + +extern "C" { +#if defined(__GNUC__) || defined(__clang__) +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wconversion" +#endif +#include <libavcodec/avcodec.h> +#if defined(__GNUC__) || defined(__clang__) +#pragma GCC diagnostic pop +#endif +} + +namespace Tegra { +class GPU; +struct VicRegisters; + +void AVFrameDeleter(AVFrame* ptr); +using AVFramePtr = std::unique_ptr<AVFrame, decltype(&AVFrameDeleter)>; + +namespace Decoder { +class H264; +class VP9; +} // namespace Decoder + +class Codec { +public: + explicit Codec(GPU& gpu); + ~Codec(); + + /// Sets NVDEC video stream codec + void SetTargetCodec(NvdecCommon::VideoCodec codec); + + /// Populate NvdecRegisters state with argument value at the provided offset + void StateWrite(u32 offset, u64 arguments); + + /// Call decoders to construct headers, decode AVFrame with ffmpeg + void Decode(); + + /// Returns next decoded frame + [[nodiscard]] AVFramePtr GetCurrentFrame(); + + /// Returns the value of current_codec + [[nodiscard]] NvdecCommon::VideoCodec GetCurrentCodec() const; + +private: + bool initialized{}; + NvdecCommon::VideoCodec current_codec{NvdecCommon::VideoCodec::None}; + + AVCodec* av_codec{nullptr}; + AVCodecContext* av_codec_ctx{nullptr}; + + GPU& gpu; + std::unique_ptr<Decoder::H264> h264_decoder; + std::unique_ptr<Decoder::VP9> vp9_decoder; + + NvdecCommon::NvdecRegisters state{}; + std::queue<AVFramePtr> av_frames{}; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/h264.cpp b/src/video_core/command_classes/codecs/h264.cpp new file mode 100644 index 000000000..65bbeac78 --- /dev/null +++ b/src/video_core/command_classes/codecs/h264.cpp @@ -0,0 +1,293 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include <array> +#include "common/bit_util.h" +#include "video_core/command_classes/codecs/h264.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +namespace Tegra::Decoder { +namespace { +// ZigZag LUTs from libavcodec. +constexpr std::array<u8, 64> zig_zag_direct{ + 0, 1, 8, 16, 9, 2, 3, 10, 17, 24, 32, 25, 18, 11, 4, 5, 12, 19, 26, 33, 40, 48, + 41, 34, 27, 20, 13, 6, 7, 14, 21, 28, 35, 42, 49, 56, 57, 50, 43, 36, 29, 22, 15, 23, + 30, 37, 44, 51, 58, 59, 52, 45, 38, 31, 39, 46, 53, 60, 61, 54, 47, 55, 62, 63, +}; + +constexpr std::array<u8, 16> zig_zag_scan{ + 0 + 0 * 4, 1 + 0 * 4, 0 + 1 * 4, 0 + 2 * 4, 1 + 1 * 4, 2 + 0 * 4, 3 + 0 * 4, 2 + 1 * 4, + 1 + 2 * 4, 0 + 3 * 4, 1 + 3 * 4, 2 + 2 * 4, 3 + 1 * 4, 3 + 2 * 4, 2 + 3 * 4, 3 + 3 * 4, +}; +} // Anonymous namespace + +H264::H264(GPU& gpu_) : gpu(gpu_) {} + +H264::~H264() = default; + +const std::vector<u8>& H264::ComposeFrameHeader(const NvdecCommon::NvdecRegisters& state, + bool is_first_frame) { + H264DecoderContext context{}; + gpu.MemoryManager().ReadBlock(state.picture_info_offset, &context, sizeof(H264DecoderContext)); + + const s32 frame_number = static_cast<s32>((context.h264_parameter_set.flags >> 46) & 0x1ffff); + if (!is_first_frame && frame_number != 0) { + frame.resize(context.frame_data_size); + + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, frame.data(), frame.size()); + } else { + /// Encode header + H264BitWriter writer{}; + writer.WriteU(1, 24); + writer.WriteU(0, 1); + writer.WriteU(3, 2); + writer.WriteU(7, 5); + writer.WriteU(100, 8); + writer.WriteU(0, 8); + writer.WriteU(31, 8); + writer.WriteUe(0); + const auto chroma_format_idc = + static_cast<u32>((context.h264_parameter_set.flags >> 12) & 3); + writer.WriteUe(chroma_format_idc); + if (chroma_format_idc == 3) { + writer.WriteBit(false); + } + + writer.WriteUe(0); + writer.WriteUe(0); + writer.WriteBit(false); // QpprimeYZeroTransformBypassFlag + writer.WriteBit(false); // Scaling matrix present flag + + const auto order_cnt_type = static_cast<u32>((context.h264_parameter_set.flags >> 14) & 3); + writer.WriteUe(static_cast<u32>((context.h264_parameter_set.flags >> 8) & 0xf)); + writer.WriteUe(order_cnt_type); + if (order_cnt_type == 0) { + writer.WriteUe(context.h264_parameter_set.log2_max_pic_order_cnt); + } else if (order_cnt_type == 1) { + writer.WriteBit(context.h264_parameter_set.delta_pic_order_always_zero_flag != 0); + + writer.WriteSe(0); + writer.WriteSe(0); + writer.WriteUe(0); + } + + const s32 pic_height = context.h264_parameter_set.pic_height_in_map_units / + (context.h264_parameter_set.frame_mbs_only_flag ? 1 : 2); + + writer.WriteUe(16); + writer.WriteBit(false); + writer.WriteUe(context.h264_parameter_set.pic_width_in_mbs - 1); + writer.WriteUe(pic_height - 1); + writer.WriteBit(context.h264_parameter_set.frame_mbs_only_flag != 0); + + if (!context.h264_parameter_set.frame_mbs_only_flag) { + writer.WriteBit(((context.h264_parameter_set.flags >> 0) & 1) != 0); + } + + writer.WriteBit(((context.h264_parameter_set.flags >> 1) & 1) != 0); + writer.WriteBit(false); // Frame cropping flag + writer.WriteBit(false); // VUI parameter present flag + + writer.End(); + + // H264 PPS + writer.WriteU(1, 24); + writer.WriteU(0, 1); + writer.WriteU(3, 2); + writer.WriteU(8, 5); + + writer.WriteUe(0); + writer.WriteUe(0); + + writer.WriteBit(context.h264_parameter_set.entropy_coding_mode_flag != 0); + writer.WriteBit(false); + writer.WriteUe(0); + writer.WriteUe(context.h264_parameter_set.num_refidx_l0_default_active); + writer.WriteUe(context.h264_parameter_set.num_refidx_l1_default_active); + writer.WriteBit(((context.h264_parameter_set.flags >> 2) & 1) != 0); + writer.WriteU(static_cast<s32>((context.h264_parameter_set.flags >> 32) & 0x3), 2); + s32 pic_init_qp = static_cast<s32>((context.h264_parameter_set.flags >> 16) & 0x3f); + pic_init_qp = (pic_init_qp << 26) >> 26; + writer.WriteSe(pic_init_qp); + writer.WriteSe(0); + s32 chroma_qp_index_offset = + static_cast<s32>((context.h264_parameter_set.flags >> 22) & 0x1f); + chroma_qp_index_offset = (chroma_qp_index_offset << 27) >> 27; + + writer.WriteSe(chroma_qp_index_offset); + writer.WriteBit(context.h264_parameter_set.deblocking_filter_control_flag != 0); + writer.WriteBit(((context.h264_parameter_set.flags >> 3) & 1) != 0); + writer.WriteBit(context.h264_parameter_set.redundant_pic_count_flag != 0); + writer.WriteBit(context.h264_parameter_set.transform_8x8_mode_flag != 0); + + writer.WriteBit(true); + + for (s32 index = 0; index < 6; index++) { + writer.WriteBit(true); + const auto matrix_x4 = + std::vector<u8>(context.scaling_matrix_4.begin(), context.scaling_matrix_4.end()); + writer.WriteScalingList(matrix_x4, index * 16, 16); + } + + if (context.h264_parameter_set.transform_8x8_mode_flag) { + for (s32 index = 0; index < 2; index++) { + writer.WriteBit(true); + const auto matrix_x8 = std::vector<u8>(context.scaling_matrix_8.begin(), + context.scaling_matrix_8.end()); + + writer.WriteScalingList(matrix_x8, index * 64, 64); + } + } + + s32 chroma_qp_index_offset2 = + static_cast<s32>((context.h264_parameter_set.flags >> 27) & 0x1f); + chroma_qp_index_offset2 = (chroma_qp_index_offset2 << 27) >> 27; + + writer.WriteSe(chroma_qp_index_offset2); + + writer.End(); + + const auto& encoded_header = writer.GetByteArray(); + frame.resize(encoded_header.size() + context.frame_data_size); + std::memcpy(frame.data(), encoded_header.data(), encoded_header.size()); + + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, + frame.data() + encoded_header.size(), + context.frame_data_size); + } + + return frame; +} + +H264BitWriter::H264BitWriter() = default; + +H264BitWriter::~H264BitWriter() = default; + +void H264BitWriter::WriteU(s32 value, s32 value_sz) { + WriteBits(value, value_sz); +} + +void H264BitWriter::WriteSe(s32 value) { + WriteExpGolombCodedInt(value); +} + +void H264BitWriter::WriteUe(u32 value) { + WriteExpGolombCodedUInt(value); +} + +void H264BitWriter::End() { + WriteBit(true); + Flush(); +} + +void H264BitWriter::WriteBit(bool state) { + WriteBits(state ? 1 : 0, 1); +} + +void H264BitWriter::WriteScalingList(const std::vector<u8>& list, s32 start, s32 count) { + std::vector<u8> scan(count); + if (count == 16) { + std::memcpy(scan.data(), zig_zag_scan.data(), scan.size()); + } else { + std::memcpy(scan.data(), zig_zag_direct.data(), scan.size()); + } + u8 last_scale = 8; + + for (s32 index = 0; index < count; index++) { + const u8 value = list[start + scan[index]]; + const s32 delta_scale = static_cast<s32>(value - last_scale); + + WriteSe(delta_scale); + + last_scale = value; + } +} + +std::vector<u8>& H264BitWriter::GetByteArray() { + return byte_array; +} + +const std::vector<u8>& H264BitWriter::GetByteArray() const { + return byte_array; +} + +void H264BitWriter::WriteBits(s32 value, s32 bit_count) { + s32 value_pos = 0; + + s32 remaining = bit_count; + + while (remaining > 0) { + s32 copy_size = remaining; + + const s32 free_bits = GetFreeBufferBits(); + + if (copy_size > free_bits) { + copy_size = free_bits; + } + + const s32 mask = (1 << copy_size) - 1; + + const s32 src_shift = (bit_count - value_pos) - copy_size; + const s32 dst_shift = (buffer_size - buffer_pos) - copy_size; + + buffer |= ((value >> src_shift) & mask) << dst_shift; + + value_pos += copy_size; + buffer_pos += copy_size; + remaining -= copy_size; + } +} + +void H264BitWriter::WriteExpGolombCodedInt(s32 value) { + const s32 sign = value <= 0 ? 0 : 1; + if (value < 0) { + value = -value; + } + value = (value << 1) - sign; + WriteExpGolombCodedUInt(value); +} + +void H264BitWriter::WriteExpGolombCodedUInt(u32 value) { + const s32 size = 32 - Common::CountLeadingZeroes32(static_cast<s32>(value + 1)); + WriteBits(1, size); + + value -= (1U << (size - 1)) - 1; + WriteBits(static_cast<s32>(value), size - 1); +} + +s32 H264BitWriter::GetFreeBufferBits() { + if (buffer_pos == buffer_size) { + Flush(); + } + + return buffer_size - buffer_pos; +} + +void H264BitWriter::Flush() { + if (buffer_pos == 0) { + return; + } + byte_array.push_back(static_cast<u8>(buffer)); + + buffer = 0; + buffer_pos = 0; +} +} // namespace Tegra::Decoder diff --git a/src/video_core/command_classes/codecs/h264.h b/src/video_core/command_classes/codecs/h264.h new file mode 100644 index 000000000..0f3a1d9f3 --- /dev/null +++ b/src/video_core/command_classes/codecs/h264.h @@ -0,0 +1,118 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#pragma once + +#include <vector> +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "video_core/command_classes/nvdec_common.h" + +namespace Tegra { +class GPU; +namespace Decoder { + +class H264BitWriter { +public: + H264BitWriter(); + ~H264BitWriter(); + + /// The following Write methods are based on clause 9.1 in the H.264 specification. + /// WriteSe and WriteUe write in the Exp-Golomb-coded syntax + void WriteU(s32 value, s32 value_sz); + void WriteSe(s32 value); + void WriteUe(u32 value); + + /// Finalize the bitstream + void End(); + + /// append a bit to the stream, equivalent value to the state parameter + void WriteBit(bool state); + + /// Based on section 7.3.2.1.1.1 and Table 7-4 in the H.264 specification + /// Writes the scaling matrices of the sream + void WriteScalingList(const std::vector<u8>& list, s32 start, s32 count); + + /// Return the bitstream as a vector. + [[nodiscard]] std::vector<u8>& GetByteArray(); + [[nodiscard]] const std::vector<u8>& GetByteArray() const; + +private: + void WriteBits(s32 value, s32 bit_count); + void WriteExpGolombCodedInt(s32 value); + void WriteExpGolombCodedUInt(u32 value); + [[nodiscard]] s32 GetFreeBufferBits(); + void Flush(); + + s32 buffer_size{8}; + + s32 buffer{}; + s32 buffer_pos{}; + std::vector<u8> byte_array; +}; + +class H264 { +public: + explicit H264(GPU& gpu); + ~H264(); + + /// Compose the H264 header of the frame for FFmpeg decoding + [[nodiscard]] const std::vector<u8>& ComposeFrameHeader( + const NvdecCommon::NvdecRegisters& state, bool is_first_frame = false); + +private: + struct H264ParameterSet { + u32 log2_max_pic_order_cnt{}; + u32 delta_pic_order_always_zero_flag{}; + u32 frame_mbs_only_flag{}; + u32 pic_width_in_mbs{}; + u32 pic_height_in_map_units{}; + INSERT_PADDING_WORDS(1); + u32 entropy_coding_mode_flag{}; + u32 bottom_field_pic_order_flag{}; + u32 num_refidx_l0_default_active{}; + u32 num_refidx_l1_default_active{}; + u32 deblocking_filter_control_flag{}; + u32 redundant_pic_count_flag{}; + u32 transform_8x8_mode_flag{}; + INSERT_PADDING_WORDS(9); + u64 flags{}; + u32 frame_number{}; + u32 frame_number2{}; + }; + static_assert(sizeof(H264ParameterSet) == 0x68, "H264ParameterSet is an invalid size"); + + struct H264DecoderContext { + INSERT_PADDING_BYTES(0x48); + u32 frame_data_size{}; + INSERT_PADDING_BYTES(0xc); + H264ParameterSet h264_parameter_set{}; + INSERT_PADDING_BYTES(0x100); + std::array<u8, 0x60> scaling_matrix_4; + std::array<u8, 0x80> scaling_matrix_8; + }; + static_assert(sizeof(H264DecoderContext) == 0x2a0, "H264DecoderContext is an invalid size"); + + std::vector<u8> frame; + GPU& gpu; +}; + +} // namespace Decoder +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/vp9.cpp b/src/video_core/command_classes/codecs/vp9.cpp new file mode 100644 index 000000000..59e586695 --- /dev/null +++ b/src/video_core/command_classes/codecs/vp9.cpp @@ -0,0 +1,989 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstring> // for std::memcpy +#include <numeric> +#include "video_core/command_classes/codecs/vp9.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" + +namespace Tegra::Decoder { +namespace { +// Default compressed header probabilities once frame context resets +constexpr Vp9EntropyProbs default_probs{ + .y_mode_prob{ + 65, 32, 18, 144, 162, 194, 41, 51, 98, 132, 68, 18, 165, 217, 196, 45, 40, 78, + 173, 80, 19, 176, 240, 193, 64, 35, 46, 221, 135, 38, 194, 248, 121, 96, 85, 29, + }, + .partition_prob{ + 199, 122, 141, 0, 147, 63, 159, 0, 148, 133, 118, 0, 121, 104, 114, 0, + 174, 73, 87, 0, 92, 41, 83, 0, 82, 99, 50, 0, 53, 39, 39, 0, + 177, 58, 59, 0, 68, 26, 63, 0, 52, 79, 25, 0, 17, 14, 12, 0, + 222, 34, 30, 0, 72, 16, 44, 0, 58, 32, 12, 0, 10, 7, 6, 0, + }, + .coef_probs{ + 195, 29, 183, 84, 49, 136, 8, 42, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 31, 107, 169, 35, 99, 159, 17, 82, 140, 8, 66, 114, 2, 44, 76, 1, 19, 32, + 40, 132, 201, 29, 114, 187, 13, 91, 157, 7, 75, 127, 3, 58, 95, 1, 28, 47, + 69, 142, 221, 42, 122, 201, 15, 91, 159, 6, 67, 121, 1, 42, 77, 1, 17, 31, + 102, 148, 228, 67, 117, 204, 17, 82, 154, 6, 59, 114, 2, 39, 75, 1, 15, 29, + 156, 57, 233, 119, 57, 212, 58, 48, 163, 29, 40, 124, 12, 30, 81, 3, 12, 31, + 191, 107, 226, 124, 117, 204, 25, 99, 155, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 29, 148, 210, 37, 126, 194, 8, 93, 157, 2, 68, 118, 1, 39, 69, 1, 17, 33, + 41, 151, 213, 27, 123, 193, 3, 82, 144, 1, 58, 105, 1, 32, 60, 1, 13, 26, + 59, 159, 220, 23, 126, 198, 4, 88, 151, 1, 66, 114, 1, 38, 71, 1, 18, 34, + 114, 136, 232, 51, 114, 207, 11, 83, 155, 3, 56, 105, 1, 33, 65, 1, 17, 34, + 149, 65, 234, 121, 57, 215, 61, 49, 166, 28, 36, 114, 12, 25, 76, 3, 16, 42, + 214, 49, 220, 132, 63, 188, 42, 65, 137, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 85, 137, 221, 104, 131, 216, 49, 111, 192, 21, 87, 155, 2, 49, 87, 1, 16, 28, + 89, 163, 230, 90, 137, 220, 29, 100, 183, 10, 70, 135, 2, 42, 81, 1, 17, 33, + 108, 167, 237, 55, 133, 222, 15, 97, 179, 4, 72, 135, 1, 45, 85, 1, 19, 38, + 124, 146, 240, 66, 124, 224, 17, 88, 175, 4, 58, 122, 1, 36, 75, 1, 18, 37, + 141, 79, 241, 126, 70, 227, 66, 58, 182, 30, 44, 136, 12, 34, 96, 2, 20, 47, + 229, 99, 249, 143, 111, 235, 46, 109, 192, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 82, 158, 236, 94, 146, 224, 25, 117, 191, 9, 87, 149, 3, 56, 99, 1, 33, 57, + 83, 167, 237, 68, 145, 222, 10, 103, 177, 2, 72, 131, 1, 41, 79, 1, 20, 39, + 99, 167, 239, 47, 141, 224, 10, 104, 178, 2, 73, 133, 1, 44, 85, 1, 22, 47, + 127, 145, 243, 71, 129, 228, 17, 93, 177, 3, 61, 124, 1, 41, 84, 1, 21, 52, + 157, 78, 244, 140, 72, 231, 69, 58, 184, 31, 44, 137, 14, 38, 105, 8, 23, 61, + 125, 34, 187, 52, 41, 133, 6, 31, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 37, 109, 153, 51, 102, 147, 23, 87, 128, 8, 67, 101, 1, 41, 63, 1, 19, 29, + 31, 154, 185, 17, 127, 175, 6, 96, 145, 2, 73, 114, 1, 51, 82, 1, 28, 45, + 23, 163, 200, 10, 131, 185, 2, 93, 148, 1, 67, 111, 1, 41, 69, 1, 14, 24, + 29, 176, 217, 12, 145, 201, 3, 101, 156, 1, 69, 111, 1, 39, 63, 1, 14, 23, + 57, 192, 233, 25, 154, 215, 6, 109, 167, 3, 78, 118, 1, 48, 69, 1, 21, 29, + 202, 105, 245, 108, 106, 216, 18, 90, 144, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 33, 172, 219, 64, 149, 206, 14, 117, 177, 5, 90, 141, 2, 61, 95, 1, 37, 57, + 33, 179, 220, 11, 140, 198, 1, 89, 148, 1, 60, 104, 1, 33, 57, 1, 12, 21, + 30, 181, 221, 8, 141, 198, 1, 87, 145, 1, 58, 100, 1, 31, 55, 1, 12, 20, + 32, 186, 224, 7, 142, 198, 1, 86, 143, 1, 58, 100, 1, 31, 55, 1, 12, 22, + 57, 192, 227, 20, 143, 204, 3, 96, 154, 1, 68, 112, 1, 42, 69, 1, 19, 32, + 212, 35, 215, 113, 47, 169, 29, 48, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 74, 129, 203, 106, 120, 203, 49, 107, 178, 19, 84, 144, 4, 50, 84, 1, 15, 25, + 71, 172, 217, 44, 141, 209, 15, 102, 173, 6, 76, 133, 2, 51, 89, 1, 24, 42, + 64, 185, 231, 31, 148, 216, 8, 103, 175, 3, 74, 131, 1, 46, 81, 1, 18, 30, + 65, 196, 235, 25, 157, 221, 5, 105, 174, 1, 67, 120, 1, 38, 69, 1, 15, 30, + 65, 204, 238, 30, 156, 224, 7, 107, 177, 2, 70, 124, 1, 42, 73, 1, 18, 34, + 225, 86, 251, 144, 104, 235, 42, 99, 181, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 85, 175, 239, 112, 165, 229, 29, 136, 200, 12, 103, 162, 6, 77, 123, 2, 53, 84, + 75, 183, 239, 30, 155, 221, 3, 106, 171, 1, 74, 128, 1, 44, 76, 1, 17, 28, + 73, 185, 240, 27, 159, 222, 2, 107, 172, 1, 75, 127, 1, 42, 73, 1, 17, 29, + 62, 190, 238, 21, 159, 222, 2, 107, 172, 1, 72, 122, 1, 40, 71, 1, 18, 32, + 61, 199, 240, 27, 161, 226, 4, 113, 180, 1, 76, 129, 1, 46, 80, 1, 23, 41, + 7, 27, 153, 5, 30, 95, 1, 16, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 50, 75, 127, 57, 75, 124, 27, 67, 108, 10, 54, 86, 1, 33, 52, 1, 12, 18, + 43, 125, 151, 26, 108, 148, 7, 83, 122, 2, 59, 89, 1, 38, 60, 1, 17, 27, + 23, 144, 163, 13, 112, 154, 2, 75, 117, 1, 50, 81, 1, 31, 51, 1, 14, 23, + 18, 162, 185, 6, 123, 171, 1, 78, 125, 1, 51, 86, 1, 31, 54, 1, 14, 23, + 15, 199, 227, 3, 150, 204, 1, 91, 146, 1, 55, 95, 1, 30, 53, 1, 11, 20, + 19, 55, 240, 19, 59, 196, 3, 52, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 41, 166, 207, 104, 153, 199, 31, 123, 181, 14, 101, 152, 5, 72, 106, 1, 36, 52, + 35, 176, 211, 12, 131, 190, 2, 88, 144, 1, 60, 101, 1, 36, 60, 1, 16, 28, + 28, 183, 213, 8, 134, 191, 1, 86, 142, 1, 56, 96, 1, 30, 53, 1, 12, 20, + 20, 190, 215, 4, 135, 192, 1, 84, 139, 1, 53, 91, 1, 28, 49, 1, 11, 20, + 13, 196, 216, 2, 137, 192, 1, 86, 143, 1, 57, 99, 1, 32, 56, 1, 13, 24, + 211, 29, 217, 96, 47, 156, 22, 43, 87, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 78, 120, 193, 111, 116, 186, 46, 102, 164, 15, 80, 128, 2, 49, 76, 1, 18, 28, + 71, 161, 203, 42, 132, 192, 10, 98, 150, 3, 69, 109, 1, 44, 70, 1, 18, 29, + 57, 186, 211, 30, 140, 196, 4, 93, 146, 1, 62, 102, 1, 38, 65, 1, 16, 27, + 47, 199, 217, 14, 145, 196, 1, 88, 142, 1, 57, 98, 1, 36, 62, 1, 15, 26, + 26, 219, 229, 5, 155, 207, 1, 94, 151, 1, 60, 104, 1, 36, 62, 1, 16, 28, + 233, 29, 248, 146, 47, 220, 43, 52, 140, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 100, 163, 232, 179, 161, 222, 63, 142, 204, 37, 113, 174, 26, 89, 137, 18, 68, 97, + 85, 181, 230, 32, 146, 209, 7, 100, 164, 3, 71, 121, 1, 45, 77, 1, 18, 30, + 65, 187, 230, 20, 148, 207, 2, 97, 159, 1, 68, 116, 1, 40, 70, 1, 14, 29, + 40, 194, 227, 8, 147, 204, 1, 94, 155, 1, 65, 112, 1, 39, 66, 1, 14, 26, + 16, 208, 228, 3, 151, 207, 1, 98, 160, 1, 67, 117, 1, 41, 74, 1, 17, 31, + 17, 38, 140, 7, 34, 80, 1, 17, 29, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 37, 75, 128, 41, 76, 128, 26, 66, 116, 12, 52, 94, 2, 32, 55, 1, 10, 16, + 50, 127, 154, 37, 109, 152, 16, 82, 121, 5, 59, 85, 1, 35, 54, 1, 13, 20, + 40, 142, 167, 17, 110, 157, 2, 71, 112, 1, 44, 72, 1, 27, 45, 1, 11, 17, + 30, 175, 188, 9, 124, 169, 1, 74, 116, 1, 48, 78, 1, 30, 49, 1, 11, 18, + 10, 222, 223, 2, 150, 194, 1, 83, 128, 1, 48, 79, 1, 27, 45, 1, 11, 17, + 36, 41, 235, 29, 36, 193, 10, 27, 111, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 85, 165, 222, 177, 162, 215, 110, 135, 195, 57, 113, 168, 23, 83, 120, 10, 49, 61, + 85, 190, 223, 36, 139, 200, 5, 90, 146, 1, 60, 103, 1, 38, 65, 1, 18, 30, + 72, 202, 223, 23, 141, 199, 2, 86, 140, 1, 56, 97, 1, 36, 61, 1, 16, 27, + 55, 218, 225, 13, 145, 200, 1, 86, 141, 1, 57, 99, 1, 35, 61, 1, 13, 22, + 15, 235, 212, 1, 132, 184, 1, 84, 139, 1, 57, 97, 1, 34, 56, 1, 14, 23, + 181, 21, 201, 61, 37, 123, 10, 38, 71, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 47, 106, 172, 95, 104, 173, 42, 93, 159, 18, 77, 131, 4, 50, 81, 1, 17, 23, + 62, 147, 199, 44, 130, 189, 28, 102, 154, 18, 75, 115, 2, 44, 65, 1, 12, 19, + 55, 153, 210, 24, 130, 194, 3, 93, 146, 1, 61, 97, 1, 31, 50, 1, 10, 16, + 49, 186, 223, 17, 148, 204, 1, 96, 142, 1, 53, 83, 1, 26, 44, 1, 11, 17, + 13, 217, 212, 2, 136, 180, 1, 78, 124, 1, 50, 83, 1, 29, 49, 1, 14, 23, + 197, 13, 247, 82, 17, 222, 25, 17, 162, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 126, 186, 247, 234, 191, 243, 176, 177, 234, 104, 158, 220, 66, 128, 186, 55, 90, 137, + 111, 197, 242, 46, 158, 219, 9, 104, 171, 2, 65, 125, 1, 44, 80, 1, 17, 91, + 104, 208, 245, 39, 168, 224, 3, 109, 162, 1, 79, 124, 1, 50, 102, 1, 43, 102, + 84, 220, 246, 31, 177, 231, 2, 115, 180, 1, 79, 134, 1, 55, 77, 1, 60, 79, + 43, 243, 240, 8, 180, 217, 1, 115, 166, 1, 84, 121, 1, 51, 67, 1, 16, 6, + }, + .switchable_interp_prob{235, 162, 36, 255, 34, 3, 149, 144}, + .inter_mode_prob{ + 2, 173, 34, 0, 7, 145, 85, 0, 7, 166, 63, 0, 7, 94, + 66, 0, 8, 64, 46, 0, 17, 81, 31, 0, 25, 29, 30, 0, + }, + .intra_inter_prob{9, 102, 187, 225}, + .comp_inter_prob{9, 102, 187, 225, 0}, + .single_ref_prob{33, 16, 77, 74, 142, 142, 172, 170, 238, 247}, + .comp_ref_prob{50, 126, 123, 221, 226}, + .tx_32x32_prob{3, 136, 37, 5, 52, 13}, + .tx_16x16_prob{20, 152, 15, 101}, + .tx_8x8_prob{100, 66}, + .skip_probs{192, 128, 64}, + .joints{32, 64, 96}, + .sign{128, 128}, + .classes{ + 224, 144, 192, 168, 192, 176, 192, 198, 198, 245, + 216, 128, 176, 160, 176, 176, 192, 198, 198, 208, + }, + .class_0{216, 208}, + .prob_bits{ + 136, 140, 148, 160, 176, 192, 224, 234, 234, 240, + 136, 140, 148, 160, 176, 192, 224, 234, 234, 240, + }, + .class_0_fr{128, 128, 64, 96, 112, 64, 128, 128, 64, 96, 112, 64}, + .fr{64, 96, 64, 64, 96, 64}, + .class_0_hp{160, 160}, + .high_precision{128, 128}, +}; + +constexpr std::array<s32, 256> norm_lut{ + 0, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, +}; + +constexpr std::array<s32, 254> map_lut{ + 20, 21, 22, 23, 24, 25, 0, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, + 1, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 2, 50, 51, 52, 53, 54, + 55, 56, 57, 58, 59, 60, 61, 3, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, + 73, 4, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 5, 86, 87, 88, 89, + 90, 91, 92, 93, 94, 95, 96, 97, 6, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, + 108, 109, 7, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 8, 122, 123, 124, + 125, 126, 127, 128, 129, 130, 131, 132, 133, 9, 134, 135, 136, 137, 138, 139, 140, 141, 142, + 143, 144, 145, 10, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 11, 158, 159, + 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 12, 170, 171, 172, 173, 174, 175, 176, 177, + 178, 179, 180, 181, 13, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 14, 194, + 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 15, 206, 207, 208, 209, 210, 211, 212, + 213, 214, 215, 216, 217, 16, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 17, + 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 18, 242, 243, 244, 245, 246, 247, + 248, 249, 250, 251, 252, 253, 19, +}; + +// 6.2.14 Tile size calculation + +[[nodiscard]] s32 CalcMinLog2TileCols(s32 frame_width) { + const s32 sb64_cols = (frame_width + 63) / 64; + s32 min_log2 = 0; + + while ((64 << min_log2) < sb64_cols) { + min_log2++; + } + + return min_log2; +} + +[[nodiscard]] s32 CalcMaxLog2TileCols(s32 frame_width) { + const s32 sb64_cols = (frame_width + 63) / 64; + s32 max_log2 = 1; + + while ((sb64_cols >> max_log2) >= 4) { + max_log2++; + } + + return max_log2 - 1; +} + +// Recenters probability. Based on section 6.3.6 of VP9 Specification +[[nodiscard]] s32 RecenterNonNeg(s32 new_prob, s32 old_prob) { + if (new_prob > old_prob * 2) { + return new_prob; + } + + if (new_prob >= old_prob) { + return (new_prob - old_prob) * 2; + } + + return (old_prob - new_prob) * 2 - 1; +} + +// Adjusts old_prob depending on new_prob. Based on section 6.3.5 of VP9 Specification +[[nodiscard]] s32 RemapProbability(s32 new_prob, s32 old_prob) { + new_prob--; + old_prob--; + + std::size_t index{}; + + if (old_prob * 2 <= 0xff) { + index = static_cast<std::size_t>(std::max(0, RecenterNonNeg(new_prob, old_prob) - 1)); + } else { + index = static_cast<std::size_t>( + std::max(0, RecenterNonNeg(0xff - 1 - new_prob, 0xff - 1 - old_prob) - 1)); + } + + return map_lut[index]; +} +} // Anonymous namespace + +VP9::VP9(GPU& gpu_) : gpu{gpu_} {} + +VP9::~VP9() = default; + +void VP9::WriteProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const bool update = new_prob != old_prob; + + writer.Write(update, diff_update_probability); + + if (update) { + WriteProbabilityDelta(writer, new_prob, old_prob); + } +} +template <typename T, std::size_t N> +void VP9::WriteProbabilityUpdate(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob) { + for (std::size_t offset = 0; offset < new_prob.size(); ++offset) { + WriteProbabilityUpdate(writer, new_prob[offset], old_prob[offset]); + } +} + +template <typename T, std::size_t N> +void VP9::WriteProbabilityUpdateAligned4(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob) { + for (std::size_t offset = 0; offset < new_prob.size(); offset += 4) { + WriteProbabilityUpdate(writer, new_prob[offset + 0], old_prob[offset + 0]); + WriteProbabilityUpdate(writer, new_prob[offset + 1], old_prob[offset + 1]); + WriteProbabilityUpdate(writer, new_prob[offset + 2], old_prob[offset + 2]); + } +} + +void VP9::WriteProbabilityDelta(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const int delta = RemapProbability(new_prob, old_prob); + + EncodeTermSubExp(writer, delta); +} + +void VP9::EncodeTermSubExp(VpxRangeEncoder& writer, s32 value) { + if (WriteLessThan(writer, value, 16)) { + writer.Write(value, 4); + } else if (WriteLessThan(writer, value, 32)) { + writer.Write(value - 16, 4); + } else if (WriteLessThan(writer, value, 64)) { + writer.Write(value - 32, 5); + } else { + value -= 64; + + constexpr s32 size = 8; + + const s32 mask = (1 << size) - 191; + + const s32 delta = value - mask; + + if (delta < 0) { + writer.Write(value, size - 1); + } else { + writer.Write(delta / 2 + mask, size - 1); + writer.Write(delta & 1, 1); + } + } +} + +bool VP9::WriteLessThan(VpxRangeEncoder& writer, s32 value, s32 test) { + const bool is_lt = value < test; + writer.Write(!is_lt); + return is_lt; +} + +void VP9::WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, + const std::array<u8, 1728>& new_prob, + const std::array<u8, 1728>& old_prob) { + constexpr u32 block_bytes = 2 * 2 * 6 * 6 * 3; + + const auto needs_update = [&](u32 base_index) { + return !std::equal(new_prob.begin() + base_index, + new_prob.begin() + base_index + block_bytes, + old_prob.begin() + base_index); + }; + + for (u32 block_index = 0; block_index < 4; block_index++) { + const u32 base_index = block_index * block_bytes; + const bool update = needs_update(base_index); + writer.Write(update); + + if (update) { + u32 index = base_index; + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 6; k++) { + for (s32 l = 0; l < 6; l++) { + if (k != 0 || l < 3) { + WriteProbabilityUpdate(writer, new_prob[index + 0], + old_prob[index + 0]); + WriteProbabilityUpdate(writer, new_prob[index + 1], + old_prob[index + 1]); + WriteProbabilityUpdate(writer, new_prob[index + 2], + old_prob[index + 2]); + } + index += 3; + } + } + } + } + } + if (block_index == static_cast<u32>(tx_mode)) { + break; + } + } +} + +void VP9::WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob) { + const bool update = new_prob != old_prob; + writer.Write(update, diff_update_probability); + + if (update) { + writer.Write(new_prob >> 1, 7); + } +} + +Vp9PictureInfo VP9::GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state) { + PictureInfo picture_info{}; + gpu.MemoryManager().ReadBlock(state.picture_info_offset, &picture_info, sizeof(PictureInfo)); + Vp9PictureInfo vp9_info = picture_info.Convert(); + + InsertEntropy(state.vp9_entropy_probs_offset, vp9_info.entropy); + + // surface_luma_offset[0:3] contains the address of the reference frame offsets in the following + // order: last, golden, altref, current. It may be worthwhile to track the updates done here + // to avoid buffering frame data needed for reference frame updating in the header composition. + std::memcpy(vp9_info.frame_offsets.data(), state.surface_luma_offset.data(), 4 * sizeof(u64)); + + return vp9_info; +} + +void VP9::InsertEntropy(u64 offset, Vp9EntropyProbs& dst) { + EntropyProbs entropy{}; + gpu.MemoryManager().ReadBlock(offset, &entropy, sizeof(EntropyProbs)); + entropy.Convert(dst); +} + +Vp9FrameContainer VP9::GetCurrentFrame(const NvdecCommon::NvdecRegisters& state) { + Vp9FrameContainer current_frame{}; + { + gpu.SyncGuestHost(); + current_frame.info = GetVp9PictureInfo(state); + current_frame.bit_stream.resize(current_frame.info.bitstream_size); + gpu.MemoryManager().ReadBlock(state.frame_bitstream_offset, current_frame.bit_stream.data(), + current_frame.info.bitstream_size); + } + // Buffer two frames, saving the last show frame info + if (!next_next_frame.bit_stream.empty()) { + Vp9FrameContainer temp{ + .info = current_frame.info, + .bit_stream = std::move(current_frame.bit_stream), + }; + next_next_frame.info.show_frame = current_frame.info.last_frame_shown; + current_frame.info = next_next_frame.info; + current_frame.bit_stream = std::move(next_next_frame.bit_stream); + next_next_frame = std::move(temp); + + if (!next_frame.bit_stream.empty()) { + Vp9FrameContainer temp2{ + .info = current_frame.info, + .bit_stream = std::move(current_frame.bit_stream), + }; + next_frame.info.show_frame = current_frame.info.last_frame_shown; + current_frame.info = next_frame.info; + current_frame.bit_stream = std::move(next_frame.bit_stream); + next_frame = std::move(temp2); + } else { + next_frame.info = current_frame.info; + next_frame.bit_stream = std::move(current_frame.bit_stream); + } + } else { + next_next_frame.info = current_frame.info; + next_next_frame.bit_stream = std::move(current_frame.bit_stream); + } + return current_frame; +} + +std::vector<u8> VP9::ComposeCompressedHeader() { + VpxRangeEncoder writer{}; + const bool update_probs = current_frame_info.show_frame && !current_frame_info.is_key_frame; + if (!current_frame_info.lossless) { + if (static_cast<u32>(current_frame_info.transform_mode) >= 3) { + writer.Write(3, 2); + writer.Write(current_frame_info.transform_mode == 4); + } else { + writer.Write(current_frame_info.transform_mode, 2); + } + } + + if (current_frame_info.transform_mode == 4) { + // tx_mode_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_8x8_prob, + prev_frame_probs.tx_8x8_prob); + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_16x16_prob, + prev_frame_probs.tx_16x16_prob); + WriteProbabilityUpdate(writer, current_frame_info.entropy.tx_32x32_prob, + prev_frame_probs.tx_32x32_prob); + if (update_probs) { + prev_frame_probs.tx_8x8_prob = current_frame_info.entropy.tx_8x8_prob; + prev_frame_probs.tx_16x16_prob = current_frame_info.entropy.tx_16x16_prob; + prev_frame_probs.tx_32x32_prob = current_frame_info.entropy.tx_32x32_prob; + } + } + // read_coef_probs() in the spec + WriteCoefProbabilityUpdate(writer, current_frame_info.transform_mode, + current_frame_info.entropy.coef_probs, prev_frame_probs.coef_probs); + // read_skip_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.skip_probs, + prev_frame_probs.skip_probs); + + if (update_probs) { + prev_frame_probs.coef_probs = current_frame_info.entropy.coef_probs; + prev_frame_probs.skip_probs = current_frame_info.entropy.skip_probs; + } + + if (!current_frame_info.intra_only) { + // read_inter_probs() in the spec + WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.inter_mode_prob, + prev_frame_probs.inter_mode_prob); + + if (current_frame_info.interp_filter == 4) { + // read_interp_filter_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.switchable_interp_prob, + prev_frame_probs.switchable_interp_prob); + if (update_probs) { + prev_frame_probs.switchable_interp_prob = + current_frame_info.entropy.switchable_interp_prob; + } + } + + // read_is_inter_probs() in the spec + WriteProbabilityUpdate(writer, current_frame_info.entropy.intra_inter_prob, + prev_frame_probs.intra_inter_prob); + + // frame_reference_mode() in the spec + if ((current_frame_info.ref_frame_sign_bias[1] & 1) != + (current_frame_info.ref_frame_sign_bias[2] & 1) || + (current_frame_info.ref_frame_sign_bias[1] & 1) != + (current_frame_info.ref_frame_sign_bias[3] & 1)) { + if (current_frame_info.reference_mode >= 1) { + writer.Write(1, 1); + writer.Write(current_frame_info.reference_mode == 2); + } else { + writer.Write(0, 1); + } + } + + // frame_reference_mode_probs() in the spec + if (current_frame_info.reference_mode == 2) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_inter_prob, + prev_frame_probs.comp_inter_prob); + if (update_probs) { + prev_frame_probs.comp_inter_prob = current_frame_info.entropy.comp_inter_prob; + } + } + + if (current_frame_info.reference_mode != 1) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.single_ref_prob, + prev_frame_probs.single_ref_prob); + if (update_probs) { + prev_frame_probs.single_ref_prob = current_frame_info.entropy.single_ref_prob; + } + } + + if (current_frame_info.reference_mode != 0) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.comp_ref_prob, + prev_frame_probs.comp_ref_prob); + if (update_probs) { + prev_frame_probs.comp_ref_prob = current_frame_info.entropy.comp_ref_prob; + } + } + + // read_y_mode_probs + for (std::size_t index = 0; index < current_frame_info.entropy.y_mode_prob.size(); + ++index) { + WriteProbabilityUpdate(writer, current_frame_info.entropy.y_mode_prob[index], + prev_frame_probs.y_mode_prob[index]); + } + + // read_partition_probs + WriteProbabilityUpdateAligned4(writer, current_frame_info.entropy.partition_prob, + prev_frame_probs.partition_prob); + + // mv_probs + for (s32 i = 0; i < 3; i++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.joints[i], + prev_frame_probs.joints[i]); + } + if (update_probs) { + prev_frame_probs.inter_mode_prob = current_frame_info.entropy.inter_mode_prob; + prev_frame_probs.intra_inter_prob = current_frame_info.entropy.intra_inter_prob; + prev_frame_probs.y_mode_prob = current_frame_info.entropy.y_mode_prob; + prev_frame_probs.partition_prob = current_frame_info.entropy.partition_prob; + prev_frame_probs.joints = current_frame_info.entropy.joints; + } + + for (s32 i = 0; i < 2; i++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.sign[i], + prev_frame_probs.sign[i]); + for (s32 j = 0; j < 10; j++) { + const int index = i * 10 + j; + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.classes[index], + prev_frame_probs.classes[index]); + } + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0[i], + prev_frame_probs.class_0[i]); + + for (s32 j = 0; j < 10; j++) { + const int index = i * 10 + j; + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.prob_bits[index], + prev_frame_probs.prob_bits[index]); + } + } + + for (s32 i = 0; i < 2; i++) { + for (s32 j = 0; j < 2; j++) { + for (s32 k = 0; k < 3; k++) { + const int index = i * 2 * 3 + j * 3 + k; + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0_fr[index], + prev_frame_probs.class_0_fr[index]); + } + } + + for (s32 j = 0; j < 3; j++) { + const int index = i * 3 + j; + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.fr[index], + prev_frame_probs.fr[index]); + } + } + + if (current_frame_info.allow_high_precision_mv) { + for (s32 index = 0; index < 2; index++) { + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.class_0_hp[index], + prev_frame_probs.class_0_hp[index]); + WriteMvProbabilityUpdate(writer, current_frame_info.entropy.high_precision[index], + prev_frame_probs.high_precision[index]); + } + } + + // save previous probs + if (update_probs) { + prev_frame_probs.sign = current_frame_info.entropy.sign; + prev_frame_probs.classes = current_frame_info.entropy.classes; + prev_frame_probs.class_0 = current_frame_info.entropy.class_0; + prev_frame_probs.prob_bits = current_frame_info.entropy.prob_bits; + prev_frame_probs.class_0_fr = current_frame_info.entropy.class_0_fr; + prev_frame_probs.fr = current_frame_info.entropy.fr; + prev_frame_probs.class_0_hp = current_frame_info.entropy.class_0_hp; + prev_frame_probs.high_precision = current_frame_info.entropy.high_precision; + } + } + writer.End(); + return writer.GetBuffer(); +} + +VpxBitStreamWriter VP9::ComposeUncompressedHeader() { + VpxBitStreamWriter uncomp_writer{}; + + uncomp_writer.WriteU(2, 2); // Frame marker. + uncomp_writer.WriteU(0, 2); // Profile. + uncomp_writer.WriteBit(false); // Show existing frame. + uncomp_writer.WriteBit(!current_frame_info.is_key_frame); // is key frame? + uncomp_writer.WriteBit(current_frame_info.show_frame); // show frame? + uncomp_writer.WriteBit(current_frame_info.error_resilient_mode); // error reslience + + if (current_frame_info.is_key_frame) { + uncomp_writer.WriteU(frame_sync_code, 24); + uncomp_writer.WriteU(0, 3); // Color space. + uncomp_writer.WriteU(0, 1); // Color range. + uncomp_writer.WriteU(current_frame_info.frame_size.width - 1, 16); + uncomp_writer.WriteU(current_frame_info.frame_size.height - 1, 16); + uncomp_writer.WriteBit(false); // Render and frame size different. + + // Reset context + prev_frame_probs = default_probs; + swap_next_golden = false; + loop_filter_ref_deltas.fill(0); + loop_filter_mode_deltas.fill(0); + + // allow frames offsets to stabilize before checking for golden frames + grace_period = 4; + + // On key frames, all frame slots are set to the current frame, + // so the value of the selected slot doesn't really matter. + frame_ctxs.fill({current_frame_number, false, default_probs}); + + // intra only, meaning the frame can be recreated with no other references + current_frame_info.intra_only = true; + + } else { + + if (!current_frame_info.show_frame) { + uncomp_writer.WriteBit(current_frame_info.intra_only); + if (!current_frame_info.last_frame_was_key) { + swap_next_golden = !swap_next_golden; + } + } else { + current_frame_info.intra_only = false; + } + if (!current_frame_info.error_resilient_mode) { + uncomp_writer.WriteU(0, 2); // Reset frame context. + } + + // Last, Golden, Altref frames + std::array<s32, 3> ref_frame_index{0, 1, 2}; + + // Set when next frame is hidden + // altref and golden references are swapped + if (swap_next_golden) { + ref_frame_index = std::array<s32, 3>{0, 2, 1}; + } + + // update Last Frame + u64 refresh_frame_flags = 1; + + // golden frame may refresh, determined if the next golden frame offset is changed + bool golden_refresh = false; + if (grace_period <= 0) { + for (s32 index = 1; index < 3; ++index) { + if (current_frame_info.frame_offsets[index] != + next_frame.info.frame_offsets[index]) { + current_frame_info.refresh_frame[index] = true; + golden_refresh = true; + grace_period = 3; + } + } + } + + if (current_frame_info.show_frame && + (!next_frame.info.show_frame || next_frame.info.is_key_frame)) { + // Update golden frame + refresh_frame_flags = swap_next_golden ? 2 : 4; + } + + if (!current_frame_info.show_frame) { + // Update altref + refresh_frame_flags = swap_next_golden ? 2 : 4; + } else if (golden_refresh) { + refresh_frame_flags = 3; + } + + if (current_frame_info.intra_only) { + uncomp_writer.WriteU(frame_sync_code, 24); + uncomp_writer.WriteU(static_cast<s32>(refresh_frame_flags), 8); + uncomp_writer.WriteU(current_frame_info.frame_size.width - 1, 16); + uncomp_writer.WriteU(current_frame_info.frame_size.height - 1, 16); + uncomp_writer.WriteBit(false); // Render and frame size different. + } else { + uncomp_writer.WriteU(static_cast<s32>(refresh_frame_flags), 8); + + for (s32 index = 1; index < 4; index++) { + uncomp_writer.WriteU(ref_frame_index[index - 1], 3); + uncomp_writer.WriteU(current_frame_info.ref_frame_sign_bias[index], 1); + } + + uncomp_writer.WriteBit(true); // Frame size with refs. + uncomp_writer.WriteBit(false); // Render and frame size different. + uncomp_writer.WriteBit(current_frame_info.allow_high_precision_mv); + uncomp_writer.WriteBit(current_frame_info.interp_filter == 4); + + if (current_frame_info.interp_filter != 4) { + uncomp_writer.WriteU(current_frame_info.interp_filter, 2); + } + } + } + + if (!current_frame_info.error_resilient_mode) { + uncomp_writer.WriteBit(true); // Refresh frame context. where do i get this info from? + uncomp_writer.WriteBit(true); // Frame parallel decoding mode. + } + + int frame_ctx_idx = 0; + if (!current_frame_info.show_frame) { + frame_ctx_idx = 1; + } + + uncomp_writer.WriteU(frame_ctx_idx, 2); // Frame context index. + prev_frame_probs = + frame_ctxs[frame_ctx_idx].probs; // reference probabilities for compressed header + frame_ctxs[frame_ctx_idx] = {current_frame_number, false, current_frame_info.entropy}; + + uncomp_writer.WriteU(current_frame_info.first_level, 6); + uncomp_writer.WriteU(current_frame_info.sharpness_level, 3); + uncomp_writer.WriteBit(current_frame_info.mode_ref_delta_enabled); + + if (current_frame_info.mode_ref_delta_enabled) { + // check if ref deltas are different, update accordingly + std::array<bool, 4> update_loop_filter_ref_deltas; + std::array<bool, 2> update_loop_filter_mode_deltas; + + bool loop_filter_delta_update = false; + + for (std::size_t index = 0; index < current_frame_info.ref_deltas.size(); index++) { + const s8 old_deltas = loop_filter_ref_deltas[index]; + const s8 new_deltas = current_frame_info.ref_deltas[index]; + const bool differing_delta = old_deltas != new_deltas; + + update_loop_filter_ref_deltas[index] = differing_delta; + loop_filter_delta_update |= differing_delta; + } + + for (std::size_t index = 0; index < current_frame_info.mode_deltas.size(); index++) { + const s8 old_deltas = loop_filter_mode_deltas[index]; + const s8 new_deltas = current_frame_info.mode_deltas[index]; + const bool differing_delta = old_deltas != new_deltas; + + update_loop_filter_mode_deltas[index] = differing_delta; + loop_filter_delta_update |= differing_delta; + } + + uncomp_writer.WriteBit(loop_filter_delta_update); + + if (loop_filter_delta_update) { + for (std::size_t index = 0; index < current_frame_info.ref_deltas.size(); index++) { + uncomp_writer.WriteBit(update_loop_filter_ref_deltas[index]); + + if (update_loop_filter_ref_deltas[index]) { + uncomp_writer.WriteS(current_frame_info.ref_deltas[index], 6); + } + } + + for (std::size_t index = 0; index < current_frame_info.mode_deltas.size(); index++) { + uncomp_writer.WriteBit(update_loop_filter_mode_deltas[index]); + + if (update_loop_filter_mode_deltas[index]) { + uncomp_writer.WriteS(current_frame_info.mode_deltas[index], 6); + } + } + // save new deltas + loop_filter_ref_deltas = current_frame_info.ref_deltas; + loop_filter_mode_deltas = current_frame_info.mode_deltas; + } + } + + uncomp_writer.WriteU(current_frame_info.base_q_index, 8); + + uncomp_writer.WriteDeltaQ(current_frame_info.y_dc_delta_q); + uncomp_writer.WriteDeltaQ(current_frame_info.uv_dc_delta_q); + uncomp_writer.WriteDeltaQ(current_frame_info.uv_ac_delta_q); + + uncomp_writer.WriteBit(false); // Segmentation enabled (TODO). + + const s32 min_tile_cols_log2 = CalcMinLog2TileCols(current_frame_info.frame_size.width); + const s32 max_tile_cols_log2 = CalcMaxLog2TileCols(current_frame_info.frame_size.width); + + const s32 tile_cols_log2_diff = current_frame_info.log2_tile_cols - min_tile_cols_log2; + const s32 tile_cols_log2_inc_mask = (1 << tile_cols_log2_diff) - 1; + + // If it's less than the maximum, we need to add an extra 0 on the bitstream + // to indicate that it should stop reading. + if (current_frame_info.log2_tile_cols < max_tile_cols_log2) { + uncomp_writer.WriteU(tile_cols_log2_inc_mask << 1, tile_cols_log2_diff + 1); + } else { + uncomp_writer.WriteU(tile_cols_log2_inc_mask, tile_cols_log2_diff); + } + + const bool tile_rows_log2_is_nonzero = current_frame_info.log2_tile_rows != 0; + + uncomp_writer.WriteBit(tile_rows_log2_is_nonzero); + + if (tile_rows_log2_is_nonzero) { + uncomp_writer.WriteBit(current_frame_info.log2_tile_rows > 1); + } + + return uncomp_writer; +} + +const std::vector<u8>& VP9::ComposeFrameHeader(const NvdecCommon::NvdecRegisters& state) { + std::vector<u8> bitstream; + { + Vp9FrameContainer curr_frame = GetCurrentFrame(state); + current_frame_info = curr_frame.info; + bitstream = std::move(curr_frame.bit_stream); + } + + // The uncompressed header routine sets PrevProb parameters needed for the compressed header + auto uncomp_writer = ComposeUncompressedHeader(); + std::vector<u8> compressed_header = ComposeCompressedHeader(); + + uncomp_writer.WriteU(static_cast<s32>(compressed_header.size()), 16); + uncomp_writer.Flush(); + std::vector<u8> uncompressed_header = uncomp_writer.GetByteArray(); + + // Write headers and frame to buffer + frame.resize(uncompressed_header.size() + compressed_header.size() + bitstream.size()); + std::memcpy(frame.data(), uncompressed_header.data(), uncompressed_header.size()); + std::memcpy(frame.data() + uncompressed_header.size(), compressed_header.data(), + compressed_header.size()); + std::memcpy(frame.data() + uncompressed_header.size() + compressed_header.size(), + bitstream.data(), bitstream.size()); + + // keep track of frame number + current_frame_number++; + grace_period--; + + // don't display hidden frames + hidden = !current_frame_info.show_frame; + return frame; +} + +VpxRangeEncoder::VpxRangeEncoder() { + Write(false); +} + +VpxRangeEncoder::~VpxRangeEncoder() = default; + +void VpxRangeEncoder::Write(s32 value, s32 value_size) { + for (s32 bit = value_size - 1; bit >= 0; bit--) { + Write(((value >> bit) & 1) != 0); + } +} + +void VpxRangeEncoder::Write(bool bit) { + Write(bit, half_probability); +} + +void VpxRangeEncoder::Write(bool bit, s32 probability) { + u32 local_range = range; + const u32 split = 1 + (((local_range - 1) * static_cast<u32>(probability)) >> 8); + local_range = split; + + if (bit) { + low_value += split; + local_range = range - split; + } + + s32 shift = norm_lut[local_range]; + local_range <<= shift; + count += shift; + + if (count >= 0) { + const s32 offset = shift - count; + + if (((low_value << (offset - 1)) >> 31) != 0) { + const s32 current_pos = static_cast<s32>(base_stream.GetPosition()); + base_stream.Seek(-1, Common::SeekOrigin::FromCurrentPos); + while (PeekByte() == 0xff) { + base_stream.WriteByte(0); + + base_stream.Seek(-2, Common::SeekOrigin::FromCurrentPos); + } + base_stream.WriteByte(static_cast<u8>((PeekByte() + 1))); + base_stream.Seek(current_pos, Common::SeekOrigin::SetOrigin); + } + base_stream.WriteByte(static_cast<u8>((low_value >> (24 - offset)))); + + low_value <<= offset; + shift = count; + low_value &= 0xffffff; + count -= 8; + } + + low_value <<= shift; + range = local_range; +} + +void VpxRangeEncoder::End() { + for (std::size_t index = 0; index < 32; ++index) { + Write(false); + } +} + +u8 VpxRangeEncoder::PeekByte() { + const u8 value = base_stream.ReadByte(); + base_stream.Seek(-1, Common::SeekOrigin::FromCurrentPos); + + return value; +} + +VpxBitStreamWriter::VpxBitStreamWriter() = default; + +VpxBitStreamWriter::~VpxBitStreamWriter() = default; + +void VpxBitStreamWriter::WriteU(u32 value, u32 value_size) { + WriteBits(value, value_size); +} + +void VpxBitStreamWriter::WriteS(s32 value, u32 value_size) { + const bool sign = value < 0; + if (sign) { + value = -value; + } + + WriteBits(static_cast<u32>(value << 1) | (sign ? 1 : 0), value_size + 1); +} + +void VpxBitStreamWriter::WriteDeltaQ(u32 value) { + const bool delta_coded = value != 0; + WriteBit(delta_coded); + + if (delta_coded) { + WriteBits(value, 4); + } +} + +void VpxBitStreamWriter::WriteBits(u32 value, u32 bit_count) { + s32 value_pos = 0; + s32 remaining = bit_count; + + while (remaining > 0) { + s32 copy_size = remaining; + + const s32 free = GetFreeBufferBits(); + + if (copy_size > free) { + copy_size = free; + } + + const s32 mask = (1 << copy_size) - 1; + + const s32 src_shift = (bit_count - value_pos) - copy_size; + const s32 dst_shift = (buffer_size - buffer_pos) - copy_size; + + buffer |= ((value >> src_shift) & mask) << dst_shift; + + value_pos += copy_size; + buffer_pos += copy_size; + remaining -= copy_size; + } +} + +void VpxBitStreamWriter::WriteBit(bool state) { + WriteBits(state ? 1 : 0, 1); +} + +s32 VpxBitStreamWriter::GetFreeBufferBits() { + if (buffer_pos == buffer_size) { + Flush(); + } + + return buffer_size - buffer_pos; +} + +void VpxBitStreamWriter::Flush() { + if (buffer_pos == 0) { + return; + } + byte_array.push_back(static_cast<u8>(buffer)); + buffer = 0; + buffer_pos = 0; +} + +std::vector<u8>& VpxBitStreamWriter::GetByteArray() { + return byte_array; +} + +const std::vector<u8>& VpxBitStreamWriter::GetByteArray() const { + return byte_array; +} + +} // namespace Tegra::Decoder diff --git a/src/video_core/command_classes/codecs/vp9.h b/src/video_core/command_classes/codecs/vp9.h new file mode 100644 index 000000000..8396c8105 --- /dev/null +++ b/src/video_core/command_classes/codecs/vp9.h @@ -0,0 +1,197 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <vector> + +#include "common/common_types.h" +#include "common/stream.h" +#include "video_core/command_classes/codecs/vp9_types.h" +#include "video_core/command_classes/nvdec_common.h" + +namespace Tegra { +class GPU; +enum class FrameType { KeyFrame = 0, InterFrame = 1 }; +namespace Decoder { + +/// The VpxRangeEncoder, and VpxBitStreamWriter classes are used to compose the +/// VP9 header bitstreams. + +class VpxRangeEncoder { +public: + VpxRangeEncoder(); + ~VpxRangeEncoder(); + + VpxRangeEncoder(const VpxRangeEncoder&) = delete; + VpxRangeEncoder& operator=(const VpxRangeEncoder&) = delete; + + VpxRangeEncoder(VpxRangeEncoder&&) = default; + VpxRangeEncoder& operator=(VpxRangeEncoder&&) = default; + + /// Writes the rightmost value_size bits from value into the stream + void Write(s32 value, s32 value_size); + + /// Writes a single bit with half probability + void Write(bool bit); + + /// Writes a bit to the base_stream encoded with probability + void Write(bool bit, s32 probability); + + /// Signal the end of the bitstream + void End(); + + [[nodiscard]] std::vector<u8>& GetBuffer() { + return base_stream.GetBuffer(); + } + + [[nodiscard]] const std::vector<u8>& GetBuffer() const { + return base_stream.GetBuffer(); + } + +private: + u8 PeekByte(); + Common::Stream base_stream{}; + u32 low_value{}; + u32 range{0xff}; + s32 count{-24}; + s32 half_probability{128}; +}; + +class VpxBitStreamWriter { +public: + VpxBitStreamWriter(); + ~VpxBitStreamWriter(); + + VpxBitStreamWriter(const VpxBitStreamWriter&) = delete; + VpxBitStreamWriter& operator=(const VpxBitStreamWriter&) = delete; + + VpxBitStreamWriter(VpxBitStreamWriter&&) = default; + VpxBitStreamWriter& operator=(VpxBitStreamWriter&&) = default; + + /// Write an unsigned integer value + void WriteU(u32 value, u32 value_size); + + /// Write a signed integer value + void WriteS(s32 value, u32 value_size); + + /// Based on 6.2.10 of VP9 Spec, writes a delta coded value + void WriteDeltaQ(u32 value); + + /// Write a single bit. + void WriteBit(bool state); + + /// Pushes current buffer into buffer_array, resets buffer + void Flush(); + + /// Returns byte_array + [[nodiscard]] std::vector<u8>& GetByteArray(); + + /// Returns const byte_array + [[nodiscard]] const std::vector<u8>& GetByteArray() const; + +private: + /// Write bit_count bits from value into buffer + void WriteBits(u32 value, u32 bit_count); + + /// Gets next available position in buffer, invokes Flush() if buffer is full + s32 GetFreeBufferBits(); + + s32 buffer_size{8}; + + s32 buffer{}; + s32 buffer_pos{}; + std::vector<u8> byte_array; +}; + +class VP9 { +public: + explicit VP9(GPU& gpu_); + ~VP9(); + + VP9(const VP9&) = delete; + VP9& operator=(const VP9&) = delete; + + VP9(VP9&&) = default; + VP9& operator=(VP9&&) = delete; + + /// Composes the VP9 frame from the GPU state information. Based on the official VP9 spec + /// documentation + [[nodiscard]] const std::vector<u8>& ComposeFrameHeader( + const NvdecCommon::NvdecRegisters& state); + + /// Returns true if the most recent frame was a hidden frame. + [[nodiscard]] bool WasFrameHidden() const { + return hidden; + } + +private: + /// Generates compressed header probability updates in the bitstream writer + template <typename T, std::size_t N> + void WriteProbabilityUpdate(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob); + + /// Generates compressed header probability updates in the bitstream writer + /// If probs are not equal, WriteProbabilityDelta is invoked + void WriteProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); + + /// Generates compressed header probability deltas in the bitstream writer + void WriteProbabilityDelta(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); + + /// Inverse of 6.3.4 Decode term subexp + void EncodeTermSubExp(VpxRangeEncoder& writer, s32 value); + + /// Writes if the value is less than the test value + bool WriteLessThan(VpxRangeEncoder& writer, s32 value, s32 test); + + /// Writes probability updates for the Coef probabilities + void WriteCoefProbabilityUpdate(VpxRangeEncoder& writer, s32 tx_mode, + const std::array<u8, 1728>& new_prob, + const std::array<u8, 1728>& old_prob); + + /// Write probabilities for 4-byte aligned structures + template <typename T, std::size_t N> + void WriteProbabilityUpdateAligned4(VpxRangeEncoder& writer, const std::array<T, N>& new_prob, + const std::array<T, N>& old_prob); + + /// Write motion vector probability updates. 6.3.17 in the spec + void WriteMvProbabilityUpdate(VpxRangeEncoder& writer, u8 new_prob, u8 old_prob); + + /// Returns VP9 information from NVDEC provided offset and size + [[nodiscard]] Vp9PictureInfo GetVp9PictureInfo(const NvdecCommon::NvdecRegisters& state); + + /// Read and convert NVDEC provided entropy probs to Vp9EntropyProbs struct + void InsertEntropy(u64 offset, Vp9EntropyProbs& dst); + + /// Returns frame to be decoded after buffering + [[nodiscard]] Vp9FrameContainer GetCurrentFrame(const NvdecCommon::NvdecRegisters& state); + + /// Use NVDEC providied information to compose the headers for the current frame + [[nodiscard]] std::vector<u8> ComposeCompressedHeader(); + [[nodiscard]] VpxBitStreamWriter ComposeUncompressedHeader(); + + GPU& gpu; + std::vector<u8> frame; + + std::array<s8, 4> loop_filter_ref_deltas{}; + std::array<s8, 2> loop_filter_mode_deltas{}; + + bool hidden = false; + s64 current_frame_number = -2; // since we buffer 2 frames + s32 grace_period = 6; // frame offsets need to stabilize + std::array<FrameContexts, 4> frame_ctxs{}; + Vp9FrameContainer next_frame{}; + Vp9FrameContainer next_next_frame{}; + bool swap_next_golden{}; + + Vp9PictureInfo current_frame_info{}; + Vp9EntropyProbs prev_frame_probs{}; + + s32 diff_update_probability = 252; + s32 frame_sync_code = 0x498342; +}; + +} // namespace Decoder +} // namespace Tegra diff --git a/src/video_core/command_classes/codecs/vp9_types.h b/src/video_core/command_classes/codecs/vp9_types.h new file mode 100644 index 000000000..139501a1c --- /dev/null +++ b/src/video_core/command_classes/codecs/vp9_types.h @@ -0,0 +1,302 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <cstring> +#include <vector> +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Tegra { +class GPU; + +namespace Decoder { +struct Vp9FrameDimensions { + s16 width{}; + s16 height{}; + s16 luma_pitch{}; + s16 chroma_pitch{}; +}; +static_assert(sizeof(Vp9FrameDimensions) == 0x8, "Vp9 Vp9FrameDimensions is an invalid size"); + +enum FrameFlags : u32 { + IsKeyFrame = 1 << 0, + LastFrameIsKeyFrame = 1 << 1, + FrameSizeChanged = 1 << 2, + ErrorResilientMode = 1 << 3, + LastShowFrame = 1 << 4, + IntraOnly = 1 << 5, +}; + +enum class TxSize { + Tx4x4 = 0, // 4x4 transform + Tx8x8 = 1, // 8x8 transform + Tx16x16 = 2, // 16x16 transform + Tx32x32 = 3, // 32x32 transform + TxSizes = 4 +}; + +enum class TxMode { + Only4X4 = 0, // Only 4x4 transform used + Allow8X8 = 1, // Allow block transform size up to 8x8 + Allow16X16 = 2, // Allow block transform size up to 16x16 + Allow32X32 = 3, // Allow block transform size up to 32x32 + TxModeSelect = 4, // Transform specified for each block + TxModes = 5 +}; + +struct Segmentation { + u8 enabled{}; + u8 update_map{}; + u8 temporal_update{}; + u8 abs_delta{}; + std::array<u32, 8> feature_mask{}; + std::array<std::array<s16, 4>, 8> feature_data{}; +}; +static_assert(sizeof(Segmentation) == 0x64, "Segmentation is an invalid size"); + +struct LoopFilter { + u8 mode_ref_delta_enabled{}; + std::array<s8, 4> ref_deltas{}; + std::array<s8, 2> mode_deltas{}; +}; +static_assert(sizeof(LoopFilter) == 0x7, "LoopFilter is an invalid size"); + +struct Vp9EntropyProbs { + std::array<u8, 36> y_mode_prob{}; + std::array<u8, 64> partition_prob{}; + std::array<u8, 1728> coef_probs{}; + std::array<u8, 8> switchable_interp_prob{}; + std::array<u8, 28> inter_mode_prob{}; + std::array<u8, 4> intra_inter_prob{}; + std::array<u8, 5> comp_inter_prob{}; + std::array<u8, 10> single_ref_prob{}; + std::array<u8, 5> comp_ref_prob{}; + std::array<u8, 6> tx_32x32_prob{}; + std::array<u8, 4> tx_16x16_prob{}; + std::array<u8, 2> tx_8x8_prob{}; + std::array<u8, 3> skip_probs{}; + std::array<u8, 3> joints{}; + std::array<u8, 2> sign{}; + std::array<u8, 20> classes{}; + std::array<u8, 2> class_0{}; + std::array<u8, 20> prob_bits{}; + std::array<u8, 12> class_0_fr{}; + std::array<u8, 6> fr{}; + std::array<u8, 2> class_0_hp{}; + std::array<u8, 2> high_precision{}; +}; +static_assert(sizeof(Vp9EntropyProbs) == 0x7B4, "Vp9EntropyProbs is an invalid size"); + +struct Vp9PictureInfo { + bool is_key_frame{}; + bool intra_only{}; + bool last_frame_was_key{}; + bool frame_size_changed{}; + bool error_resilient_mode{}; + bool last_frame_shown{}; + bool show_frame{}; + std::array<s8, 4> ref_frame_sign_bias{}; + s32 base_q_index{}; + s32 y_dc_delta_q{}; + s32 uv_dc_delta_q{}; + s32 uv_ac_delta_q{}; + bool lossless{}; + s32 transform_mode{}; + bool allow_high_precision_mv{}; + s32 interp_filter{}; + s32 reference_mode{}; + s8 comp_fixed_ref{}; + std::array<s8, 2> comp_var_ref{}; + s32 log2_tile_cols{}; + s32 log2_tile_rows{}; + bool segment_enabled{}; + bool segment_map_update{}; + bool segment_map_temporal_update{}; + s32 segment_abs_delta{}; + std::array<u32, 8> segment_feature_enable{}; + std::array<std::array<s16, 4>, 8> segment_feature_data{}; + bool mode_ref_delta_enabled{}; + bool use_prev_in_find_mv_refs{}; + std::array<s8, 4> ref_deltas{}; + std::array<s8, 2> mode_deltas{}; + Vp9EntropyProbs entropy{}; + Vp9FrameDimensions frame_size{}; + u8 first_level{}; + u8 sharpness_level{}; + u32 bitstream_size{}; + std::array<u64, 4> frame_offsets{}; + std::array<bool, 4> refresh_frame{}; +}; + +struct Vp9FrameContainer { + Vp9PictureInfo info{}; + std::vector<u8> bit_stream; +}; + +struct PictureInfo { + INSERT_PADDING_WORDS(12); + u32 bitstream_size{}; + INSERT_PADDING_WORDS(5); + Vp9FrameDimensions last_frame_size{}; + Vp9FrameDimensions golden_frame_size{}; + Vp9FrameDimensions alt_frame_size{}; + Vp9FrameDimensions current_frame_size{}; + u32 vp9_flags{}; + std::array<s8, 4> ref_frame_sign_bias{}; + u8 first_level{}; + u8 sharpness_level{}; + u8 base_q_index{}; + u8 y_dc_delta_q{}; + u8 uv_ac_delta_q{}; + u8 uv_dc_delta_q{}; + u8 lossless{}; + u8 tx_mode{}; + u8 allow_high_precision_mv{}; + u8 interp_filter{}; + u8 reference_mode{}; + s8 comp_fixed_ref{}; + std::array<s8, 2> comp_var_ref{}; + u8 log2_tile_cols{}; + u8 log2_tile_rows{}; + Segmentation segmentation{}; + LoopFilter loop_filter{}; + INSERT_PADDING_BYTES(5); + u32 surface_params{}; + INSERT_PADDING_WORDS(3); + + [[nodiscard]] Vp9PictureInfo Convert() const { + return { + .is_key_frame = (vp9_flags & FrameFlags::IsKeyFrame) != 0, + .intra_only = (vp9_flags & FrameFlags::IntraOnly) != 0, + .last_frame_was_key = (vp9_flags & FrameFlags::LastFrameIsKeyFrame) != 0, + .frame_size_changed = (vp9_flags & FrameFlags::FrameSizeChanged) != 0, + .error_resilient_mode = (vp9_flags & FrameFlags::ErrorResilientMode) != 0, + .last_frame_shown = (vp9_flags & FrameFlags::LastShowFrame) != 0, + .ref_frame_sign_bias = ref_frame_sign_bias, + .base_q_index = base_q_index, + .y_dc_delta_q = y_dc_delta_q, + .uv_dc_delta_q = uv_dc_delta_q, + .uv_ac_delta_q = uv_ac_delta_q, + .lossless = lossless != 0, + .transform_mode = tx_mode, + .allow_high_precision_mv = allow_high_precision_mv != 0, + .interp_filter = interp_filter, + .reference_mode = reference_mode, + .comp_fixed_ref = comp_fixed_ref, + .comp_var_ref = comp_var_ref, + .log2_tile_cols = log2_tile_cols, + .log2_tile_rows = log2_tile_rows, + .segment_enabled = segmentation.enabled != 0, + .segment_map_update = segmentation.update_map != 0, + .segment_map_temporal_update = segmentation.temporal_update != 0, + .segment_abs_delta = segmentation.abs_delta, + .segment_feature_enable = segmentation.feature_mask, + .segment_feature_data = segmentation.feature_data, + .mode_ref_delta_enabled = loop_filter.mode_ref_delta_enabled != 0, + .use_prev_in_find_mv_refs = !(vp9_flags == (FrameFlags::ErrorResilientMode)) && + !(vp9_flags == (FrameFlags::FrameSizeChanged)) && + !(vp9_flags == (FrameFlags::IntraOnly)) && + (vp9_flags == (FrameFlags::LastShowFrame)) && + !(vp9_flags == (FrameFlags::LastFrameIsKeyFrame)), + .ref_deltas = loop_filter.ref_deltas, + .mode_deltas = loop_filter.mode_deltas, + .frame_size = current_frame_size, + .first_level = first_level, + .sharpness_level = sharpness_level, + .bitstream_size = bitstream_size, + }; + } +}; +static_assert(sizeof(PictureInfo) == 0x100, "PictureInfo is an invalid size"); + +struct EntropyProbs { + INSERT_PADDING_BYTES(1024); + std::array<u8, 28> inter_mode_prob{}; + std::array<u8, 4> intra_inter_prob{}; + INSERT_PADDING_BYTES(80); + std::array<u8, 2> tx_8x8_prob{}; + std::array<u8, 4> tx_16x16_prob{}; + std::array<u8, 6> tx_32x32_prob{}; + std::array<u8, 4> y_mode_prob_e8{}; + std::array<std::array<u8, 8>, 4> y_mode_prob_e0e7{}; + INSERT_PADDING_BYTES(64); + std::array<u8, 64> partition_prob{}; + INSERT_PADDING_BYTES(10); + std::array<u8, 8> switchable_interp_prob{}; + std::array<u8, 5> comp_inter_prob{}; + std::array<u8, 3> skip_probs{}; + INSERT_PADDING_BYTES(1); + std::array<u8, 3> joints{}; + std::array<u8, 2> sign{}; + std::array<u8, 2> class_0{}; + std::array<u8, 6> fr{}; + std::array<u8, 2> class_0_hp{}; + std::array<u8, 2> high_precision{}; + std::array<u8, 20> classes{}; + std::array<u8, 12> class_0_fr{}; + std::array<u8, 20> pred_bits{}; + std::array<u8, 10> single_ref_prob{}; + std::array<u8, 5> comp_ref_prob{}; + INSERT_PADDING_BYTES(17); + std::array<u8, 2304> coef_probs{}; + + void Convert(Vp9EntropyProbs& fc) { + fc.inter_mode_prob = inter_mode_prob; + fc.intra_inter_prob = intra_inter_prob; + fc.tx_8x8_prob = tx_8x8_prob; + fc.tx_16x16_prob = tx_16x16_prob; + fc.tx_32x32_prob = tx_32x32_prob; + + for (std::size_t i = 0; i < 4; i++) { + for (std::size_t j = 0; j < 9; j++) { + fc.y_mode_prob[j + 9 * i] = j < 8 ? y_mode_prob_e0e7[i][j] : y_mode_prob_e8[i]; + } + } + + fc.partition_prob = partition_prob; + fc.switchable_interp_prob = switchable_interp_prob; + fc.comp_inter_prob = comp_inter_prob; + fc.skip_probs = skip_probs; + fc.joints = joints; + fc.sign = sign; + fc.class_0 = class_0; + fc.fr = fr; + fc.class_0_hp = class_0_hp; + fc.high_precision = high_precision; + fc.classes = classes; + fc.class_0_fr = class_0_fr; + fc.prob_bits = pred_bits; + fc.single_ref_prob = single_ref_prob; + fc.comp_ref_prob = comp_ref_prob; + + // Skip the 4th element as it goes unused + for (std::size_t i = 0; i < coef_probs.size(); i += 4) { + const std::size_t j = i - i / 4; + fc.coef_probs[j] = coef_probs[i]; + fc.coef_probs[j + 1] = coef_probs[i + 1]; + fc.coef_probs[j + 2] = coef_probs[i + 2]; + } + } +}; +static_assert(sizeof(EntropyProbs) == 0xEA0, "EntropyProbs is an invalid size"); + +enum class Ref { Last, Golden, AltRef }; + +struct RefPoolElement { + s64 frame{}; + Ref ref{}; + bool refresh{}; +}; + +struct FrameContexts { + s64 from{}; + bool adapted{}; + Vp9EntropyProbs probs{}; +}; + +}; // namespace Decoder +}; // namespace Tegra diff --git a/src/video_core/command_classes/host1x.cpp b/src/video_core/command_classes/host1x.cpp new file mode 100644 index 000000000..b12494528 --- /dev/null +++ b/src/video_core/command_classes/host1x.cpp @@ -0,0 +1,30 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "video_core/command_classes/host1x.h" +#include "video_core/gpu.h" + +Tegra::Host1x::Host1x(GPU& gpu_) : gpu(gpu_) {} + +Tegra::Host1x::~Host1x() = default; + +void Tegra::Host1x::ProcessMethod(Method method, u32 argument) { + switch (method) { + case Method::LoadSyncptPayload32: + syncpoint_value = argument; + break; + case Method::WaitSyncpt: + case Method::WaitSyncpt32: + Execute(argument); + break; + default: + UNIMPLEMENTED_MSG("Host1x method 0x{:X}", static_cast<u32>(method)); + break; + } +} + +void Tegra::Host1x::Execute(u32 data) { + gpu.WaitFence(data, syncpoint_value); +} diff --git a/src/video_core/command_classes/host1x.h b/src/video_core/command_classes/host1x.h new file mode 100644 index 000000000..7e94799dd --- /dev/null +++ b/src/video_core/command_classes/host1x.h @@ -0,0 +1,37 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <vector> +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Tegra { +class GPU; +class Nvdec; + +class Host1x { +public: + enum class Method : u32 { + WaitSyncpt = 0x8, + LoadSyncptPayload32 = 0x4e, + WaitSyncpt32 = 0x50, + }; + + explicit Host1x(GPU& gpu); + ~Host1x(); + + /// Writes the method into the state, Invoke Execute() if encountered + void ProcessMethod(Method method, u32 argument); + +private: + /// For Host1x, execute is waiting on a syncpoint previously written into the state + void Execute(u32 data); + + u32 syncpoint_value{}; + GPU& gpu; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/nvdec.cpp b/src/video_core/command_classes/nvdec.cpp new file mode 100644 index 000000000..79e1f4e13 --- /dev/null +++ b/src/video_core/command_classes/nvdec.cpp @@ -0,0 +1,48 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "video_core/command_classes/nvdec.h" +#include "video_core/gpu.h" + +namespace Tegra { + +Nvdec::Nvdec(GPU& gpu_) : gpu(gpu_), codec(std::make_unique<Codec>(gpu)) {} + +Nvdec::~Nvdec() = default; + +void Nvdec::ProcessMethod(Method method, const std::vector<u32>& arguments) { + if (method == Method::SetVideoCodec) { + codec->StateWrite(static_cast<u32>(method), arguments[0]); + } else { + codec->StateWrite(static_cast<u32>(method), static_cast<u64>(arguments[0]) << 8); + } + + switch (method) { + case Method::SetVideoCodec: + codec->SetTargetCodec(static_cast<NvdecCommon::VideoCodec>(arguments[0])); + break; + case Method::Execute: + Execute(); + break; + } +} + +AVFramePtr Nvdec::GetFrame() { + return codec->GetCurrentFrame(); +} + +void Nvdec::Execute() { + switch (codec->GetCurrentCodec()) { + case NvdecCommon::VideoCodec::H264: + case NvdecCommon::VideoCodec::Vp9: + codec->Decode(); + break; + default: + UNIMPLEMENTED_MSG("Unknown codec {}", static_cast<u32>(codec->GetCurrentCodec())); + break; + } +} + +} // namespace Tegra diff --git a/src/video_core/command_classes/nvdec.h b/src/video_core/command_classes/nvdec.h new file mode 100644 index 000000000..e4877c533 --- /dev/null +++ b/src/video_core/command_classes/nvdec.h @@ -0,0 +1,38 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <vector> +#include "common/common_types.h" +#include "video_core/command_classes/codecs/codec.h" + +namespace Tegra { +class GPU; + +class Nvdec { +public: + enum class Method : u32 { + SetVideoCodec = 0x80, + Execute = 0xc0, + }; + + explicit Nvdec(GPU& gpu); + ~Nvdec(); + + /// Writes the method into the state, Invoke Execute() if encountered + void ProcessMethod(Method method, const std::vector<u32>& arguments); + + /// Return most recently decoded frame + [[nodiscard]] AVFramePtr GetFrame(); + +private: + /// Invoke codec to decode a frame + void Execute(); + + GPU& gpu; + std::unique_ptr<Codec> codec; +}; +} // namespace Tegra diff --git a/src/video_core/command_classes/nvdec_common.h b/src/video_core/command_classes/nvdec_common.h new file mode 100644 index 000000000..01b5e086d --- /dev/null +++ b/src/video_core/command_classes/nvdec_common.h @@ -0,0 +1,48 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "common/common_types.h" + +namespace Tegra::NvdecCommon { + +struct NvdecRegisters { + INSERT_PADDING_WORDS(256); + u64 set_codec_id{}; + INSERT_PADDING_WORDS(254); + u64 set_platform_id{}; + u64 picture_info_offset{}; + u64 frame_bitstream_offset{}; + u64 frame_number{}; + u64 h264_slice_data_offsets{}; + u64 h264_mv_dump_offset{}; + INSERT_PADDING_WORDS(6); + u64 frame_stats_offset{}; + u64 h264_last_surface_luma_offset{}; + u64 h264_last_surface_chroma_offset{}; + std::array<u64, 17> surface_luma_offset{}; + std::array<u64, 17> surface_chroma_offset{}; + INSERT_PADDING_WORDS(132); + u64 vp9_entropy_probs_offset{}; + u64 vp9_backward_updates_offset{}; + u64 vp9_last_frame_segmap_offset{}; + u64 vp9_curr_frame_segmap_offset{}; + INSERT_PADDING_WORDS(2); + u64 vp9_last_frame_mvs_offset{}; + u64 vp9_curr_frame_mvs_offset{}; + INSERT_PADDING_WORDS(2); +}; +static_assert(sizeof(NvdecRegisters) == (0xBC0), "NvdecRegisters is incorrect size"); + +enum class VideoCodec : u32 { + None = 0x0, + H264 = 0x3, + Vp8 = 0x5, + H265 = 0x7, + Vp9 = 0x9, +}; + +} // namespace Tegra::NvdecCommon diff --git a/src/video_core/command_classes/sync_manager.cpp b/src/video_core/command_classes/sync_manager.cpp new file mode 100644 index 000000000..19dc9e0ab --- /dev/null +++ b/src/video_core/command_classes/sync_manager.cpp @@ -0,0 +1,60 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include <algorithm> +#include "sync_manager.h" +#include "video_core/gpu.h" + +namespace Tegra { +SyncptIncrManager::SyncptIncrManager(GPU& gpu_) : gpu(gpu_) {} +SyncptIncrManager::~SyncptIncrManager() = default; + +void SyncptIncrManager::Increment(u32 id) { + increments.emplace_back(0, 0, id, true); + IncrementAllDone(); +} + +u32 SyncptIncrManager::IncrementWhenDone(u32 class_id, u32 id) { + const u32 handle = current_id++; + increments.emplace_back(handle, class_id, id); + return handle; +} + +void SyncptIncrManager::SignalDone(u32 handle) { + const auto done_incr = + std::find_if(increments.begin(), increments.end(), + [handle](const SyncptIncr& incr) { return incr.id == handle; }); + if (done_incr != increments.cend()) { + done_incr->complete = true; + } + IncrementAllDone(); +} + +void SyncptIncrManager::IncrementAllDone() { + std::size_t done_count = 0; + for (; done_count < increments.size(); ++done_count) { + if (!increments[done_count].complete) { + break; + } + gpu.IncrementSyncPoint(increments[done_count].syncpt_id); + } + increments.erase(increments.begin(), increments.begin() + done_count); +} +} // namespace Tegra diff --git a/src/video_core/command_classes/sync_manager.h b/src/video_core/command_classes/sync_manager.h new file mode 100644 index 000000000..2c321ec58 --- /dev/null +++ b/src/video_core/command_classes/sync_manager.h @@ -0,0 +1,64 @@ +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#pragma once + +#include <mutex> +#include <vector> +#include "common/common_types.h" + +namespace Tegra { +class GPU; +struct SyncptIncr { + u32 id; + u32 class_id; + u32 syncpt_id; + bool complete; + + SyncptIncr(u32 id_, u32 class_id_, u32 syncpt_id_, bool done = false) + : id(id_), class_id(class_id_), syncpt_id(syncpt_id_), complete(done) {} +}; + +class SyncptIncrManager { +public: + explicit SyncptIncrManager(GPU& gpu); + ~SyncptIncrManager(); + + /// Add syncpoint id and increment all + void Increment(u32 id); + + /// Returns a handle to increment later + u32 IncrementWhenDone(u32 class_id, u32 id); + + /// IncrememntAllDone, including handle + void SignalDone(u32 handle); + + /// Increment all sequential pending increments that are already done. + void IncrementAllDone(); + +private: + std::vector<SyncptIncr> increments; + std::mutex increment_lock; + u32 current_id{}; + + GPU& gpu; +}; + +} // namespace Tegra diff --git a/src/video_core/command_classes/vic.cpp b/src/video_core/command_classes/vic.cpp new file mode 100644 index 000000000..55e632346 --- /dev/null +++ b/src/video_core/command_classes/vic.cpp @@ -0,0 +1,175 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <array> +#include "common/assert.h" +#include "video_core/command_classes/nvdec.h" +#include "video_core/command_classes/vic.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/gpu.h" +#include "video_core/memory_manager.h" +#include "video_core/textures/decoders.h" + +extern "C" { +#include <libswscale/swscale.h> +} + +namespace Tegra { + +Vic::Vic(GPU& gpu_, std::shared_ptr<Nvdec> nvdec_processor_) + : gpu(gpu_), nvdec_processor(std::move(nvdec_processor_)) {} +Vic::~Vic() = default; + +void Vic::VicStateWrite(u32 offset, u32 arguments) { + u8* const state_offset = reinterpret_cast<u8*>(&vic_state) + offset * sizeof(u32); + std::memcpy(state_offset, &arguments, sizeof(u32)); +} + +void Vic::ProcessMethod(Method method, const std::vector<u32>& arguments) { + LOG_DEBUG(HW_GPU, "Vic method 0x{:X}", method); + VicStateWrite(static_cast<u32>(method), arguments[0]); + const u64 arg = static_cast<u64>(arguments[0]) << 8; + switch (method) { + case Method::Execute: + Execute(); + break; + case Method::SetConfigStructOffset: + config_struct_address = arg; + break; + case Method::SetOutputSurfaceLumaOffset: + output_surface_luma_address = arg; + break; + case Method::SetOutputSurfaceChromaUOffset: + output_surface_chroma_u_address = arg; + break; + case Method::SetOutputSurfaceChromaVOffset: + output_surface_chroma_v_address = arg; + break; + default: + break; + } +} + +void Vic::Execute() { + if (output_surface_luma_address == 0) { + LOG_ERROR(Service_NVDRV, "VIC Luma address not set. Received 0x{:X}", + vic_state.output_surface.luma_offset); + return; + } + const VicConfig config{gpu.MemoryManager().Read<u64>(config_struct_address + 0x20)}; + const AVFramePtr frame_ptr = nvdec_processor->GetFrame(); + const auto* frame = frame_ptr.get(); + if (!frame || frame->width == 0 || frame->height == 0) { + return; + } + const VideoPixelFormat pixel_format = + static_cast<VideoPixelFormat>(config.pixel_format.Value()); + switch (pixel_format) { + case VideoPixelFormat::BGRA8: + case VideoPixelFormat::RGBA8: { + LOG_TRACE(Service_NVDRV, "Writing RGB Frame"); + + if (scaler_ctx == nullptr || frame->width != scaler_width || + frame->height != scaler_height) { + const AVPixelFormat target_format = + (pixel_format == VideoPixelFormat::RGBA8) ? AV_PIX_FMT_RGBA : AV_PIX_FMT_BGRA; + + sws_freeContext(scaler_ctx); + scaler_ctx = nullptr; + + // FFmpeg returns all frames in YUV420, convert it into expected format + scaler_ctx = + sws_getContext(frame->width, frame->height, AV_PIX_FMT_YUV420P, frame->width, + frame->height, target_format, 0, nullptr, nullptr, nullptr); + + scaler_width = frame->width; + scaler_height = frame->height; + } + // Get Converted frame + const std::size_t linear_size = frame->width * frame->height * 4; + + using AVMallocPtr = std::unique_ptr<u8, decltype(&av_free)>; + AVMallocPtr converted_frame_buffer{static_cast<u8*>(av_malloc(linear_size)), av_free}; + + const int converted_stride{frame->width * 4}; + u8* const converted_frame_buf_addr{converted_frame_buffer.get()}; + + sws_scale(scaler_ctx, frame->data, frame->linesize, 0, frame->height, + &converted_frame_buf_addr, &converted_stride); + + const u32 blk_kind = static_cast<u32>(config.block_linear_kind); + if (blk_kind != 0) { + // swizzle pitch linear to block linear + const u32 block_height = static_cast<u32>(config.block_linear_height_log2); + const auto size = Tegra::Texture::CalculateSize(true, 4, frame->width, frame->height, 1, + block_height, 0); + std::vector<u8> swizzled_data(size); + Tegra::Texture::SwizzleSubrect(frame->width, frame->height, frame->width * 4, + frame->width, 4, swizzled_data.data(), + converted_frame_buffer.get(), block_height, 0, 0); + + gpu.MemoryManager().WriteBlock(output_surface_luma_address, swizzled_data.data(), size); + gpu.Maxwell3D().OnMemoryWrite(); + } else { + // send pitch linear frame + gpu.MemoryManager().WriteBlock(output_surface_luma_address, converted_frame_buf_addr, + linear_size); + gpu.Maxwell3D().OnMemoryWrite(); + } + break; + } + case VideoPixelFormat::Yuv420: { + LOG_TRACE(Service_NVDRV, "Writing YUV420 Frame"); + + const std::size_t surface_width = config.surface_width_minus1 + 1; + const std::size_t surface_height = config.surface_height_minus1 + 1; + const std::size_t half_width = surface_width / 2; + const std::size_t half_height = config.surface_height_minus1 / 2; + const std::size_t aligned_width = (surface_width + 0xff) & ~0xff; + + const auto* luma_ptr = frame->data[0]; + const auto* chroma_b_ptr = frame->data[1]; + const auto* chroma_r_ptr = frame->data[2]; + const auto stride = frame->linesize[0]; + const auto half_stride = frame->linesize[1]; + + std::vector<u8> luma_buffer(aligned_width * surface_height); + std::vector<u8> chroma_buffer(aligned_width * half_height); + + // Populate luma buffer + for (std::size_t y = 0; y < surface_height - 1; ++y) { + std::size_t src = y * stride; + std::size_t dst = y * aligned_width; + + std::size_t size = surface_width; + + for (std::size_t offset = 0; offset < size; ++offset) { + luma_buffer[dst + offset] = luma_ptr[src + offset]; + } + } + gpu.MemoryManager().WriteBlock(output_surface_luma_address, luma_buffer.data(), + luma_buffer.size()); + + // Populate chroma buffer from both channels with interleaving. + for (std::size_t y = 0; y < half_height; ++y) { + std::size_t src = y * half_stride; + std::size_t dst = y * aligned_width; + + for (std::size_t x = 0; x < half_width; ++x) { + chroma_buffer[dst + x * 2] = chroma_b_ptr[src + x]; + chroma_buffer[dst + x * 2 + 1] = chroma_r_ptr[src + x]; + } + } + gpu.MemoryManager().WriteBlock(output_surface_chroma_u_address, chroma_buffer.data(), + chroma_buffer.size()); + gpu.Maxwell3D().OnMemoryWrite(); + break; + } + default: + UNIMPLEMENTED_MSG("Unknown video pixel format {}", config.pixel_format.Value()); + break; + } +} + +} // namespace Tegra diff --git a/src/video_core/command_classes/vic.h b/src/video_core/command_classes/vic.h new file mode 100644 index 000000000..8c4e284a1 --- /dev/null +++ b/src/video_core/command_classes/vic.h @@ -0,0 +1,110 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <vector> +#include "common/bit_field.h" +#include "common/common_types.h" + +struct SwsContext; + +namespace Tegra { +class GPU; +class Nvdec; + +struct PlaneOffsets { + u32 luma_offset{}; + u32 chroma_u_offset{}; + u32 chroma_v_offset{}; +}; + +struct VicRegisters { + INSERT_PADDING_WORDS(64); + u32 nop{}; + INSERT_PADDING_WORDS(15); + u32 pm_trigger{}; + INSERT_PADDING_WORDS(47); + u32 set_application_id{}; + u32 set_watchdog_timer{}; + INSERT_PADDING_WORDS(17); + u32 context_save_area{}; + u32 context_switch{}; + INSERT_PADDING_WORDS(43); + u32 execute{}; + INSERT_PADDING_WORDS(63); + std::array<std::array<PlaneOffsets, 8>, 8> surfacex_slots{}; + u32 picture_index{}; + u32 control_params{}; + u32 config_struct_offset{}; + u32 filter_struct_offset{}; + u32 palette_offset{}; + u32 hist_offset{}; + u32 context_id{}; + u32 fce_ucode_size{}; + PlaneOffsets output_surface{}; + u32 fce_ucode_offset{}; + INSERT_PADDING_WORDS(4); + std::array<u32, 8> slot_context_id{}; + INSERT_PADDING_WORDS(16); +}; +static_assert(sizeof(VicRegisters) == 0x7A0, "VicRegisters is an invalid size"); + +class Vic { +public: + enum class Method : u32 { + Execute = 0xc0, + SetControlParams = 0x1c1, + SetConfigStructOffset = 0x1c2, + SetOutputSurfaceLumaOffset = 0x1c8, + SetOutputSurfaceChromaUOffset = 0x1c9, + SetOutputSurfaceChromaVOffset = 0x1ca + }; + + explicit Vic(GPU& gpu, std::shared_ptr<Nvdec> nvdec_processor); + ~Vic(); + + /// Write to the device state. + void ProcessMethod(Method method, const std::vector<u32>& arguments); + +private: + void Execute(); + + void VicStateWrite(u32 offset, u32 arguments); + VicRegisters vic_state{}; + + enum class VideoPixelFormat : u64_le { + RGBA8 = 0x1f, + BGRA8 = 0x20, + Yuv420 = 0x44, + }; + + union VicConfig { + u64_le raw{}; + BitField<0, 7, u64_le> pixel_format; + BitField<7, 2, u64_le> chroma_loc_horiz; + BitField<9, 2, u64_le> chroma_loc_vert; + BitField<11, 4, u64_le> block_linear_kind; + BitField<15, 4, u64_le> block_linear_height_log2; + BitField<19, 3, u64_le> reserved0; + BitField<22, 10, u64_le> reserved1; + BitField<32, 14, u64_le> surface_width_minus1; + BitField<46, 14, u64_le> surface_height_minus1; + }; + + GPU& gpu; + std::shared_ptr<Tegra::Nvdec> nvdec_processor; + + GPUVAddr config_struct_address{}; + GPUVAddr output_surface_luma_address{}; + GPUVAddr output_surface_chroma_u_address{}; + GPUVAddr output_surface_chroma_v_address{}; + + SwsContext* scaler_ctx{}; + s32 scaler_width{}; + s32 scaler_height{}; +}; + +} // namespace Tegra diff --git a/src/video_core/compatible_formats.cpp b/src/video_core/compatible_formats.cpp index b06c32c84..acf2668dc 100644 --- a/src/video_core/compatible_formats.cpp +++ b/src/video_core/compatible_formats.cpp @@ -3,33 +3,33 @@ // Refer to the license.txt file included. #include <array> -#include <bitset> #include <cstddef> +#include "common/common_types.h" #include "video_core/compatible_formats.h" #include "video_core/surface.h" namespace VideoCore::Surface { - namespace { +using Table = std::array<std::array<u64, 2>, MaxPixelFormat>; // Compatibility table taken from Table 3.X.2 in: // https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_texture_view.txt -constexpr std::array VIEW_CLASS_128_BITS = { +constexpr std::array VIEW_CLASS_128_BITS{ PixelFormat::R32G32B32A32_FLOAT, PixelFormat::R32G32B32A32_UINT, PixelFormat::R32G32B32A32_SINT, }; -constexpr std::array VIEW_CLASS_96_BITS = { +constexpr std::array VIEW_CLASS_96_BITS{ PixelFormat::R32G32B32_FLOAT, }; // Missing formats: // PixelFormat::RGB32UI, // PixelFormat::RGB32I, -constexpr std::array VIEW_CLASS_64_BITS = { +constexpr std::array VIEW_CLASS_64_BITS{ PixelFormat::R32G32_FLOAT, PixelFormat::R32G32_UINT, PixelFormat::R32G32_SINT, PixelFormat::R16G16B16A16_FLOAT, PixelFormat::R16G16B16A16_UNORM, PixelFormat::R16G16B16A16_SNORM, @@ -38,7 +38,7 @@ constexpr std::array VIEW_CLASS_64_BITS = { // TODO: How should we handle 48 bits? -constexpr std::array VIEW_CLASS_32_BITS = { +constexpr std::array VIEW_CLASS_32_BITS{ PixelFormat::R16G16_FLOAT, PixelFormat::B10G11R11_FLOAT, PixelFormat::R32_FLOAT, PixelFormat::A2B10G10R10_UNORM, PixelFormat::R16G16_UINT, PixelFormat::R32_UINT, PixelFormat::R16G16_SINT, PixelFormat::R32_SINT, PixelFormat::A8B8G8R8_UNORM, @@ -50,43 +50,105 @@ constexpr std::array VIEW_CLASS_32_BITS = { // TODO: How should we handle 24 bits? -constexpr std::array VIEW_CLASS_16_BITS = { +constexpr std::array VIEW_CLASS_16_BITS{ PixelFormat::R16_FLOAT, PixelFormat::R8G8_UINT, PixelFormat::R16_UINT, PixelFormat::R16_SINT, PixelFormat::R8G8_UNORM, PixelFormat::R16_UNORM, PixelFormat::R8G8_SNORM, PixelFormat::R16_SNORM, PixelFormat::R8G8_SINT, }; -constexpr std::array VIEW_CLASS_8_BITS = { +constexpr std::array VIEW_CLASS_8_BITS{ PixelFormat::R8_UINT, PixelFormat::R8_UNORM, PixelFormat::R8_SINT, PixelFormat::R8_SNORM, }; -constexpr std::array VIEW_CLASS_RGTC1_RED = { +constexpr std::array VIEW_CLASS_RGTC1_RED{ PixelFormat::BC4_UNORM, PixelFormat::BC4_SNORM, }; -constexpr std::array VIEW_CLASS_RGTC2_RG = { +constexpr std::array VIEW_CLASS_RGTC2_RG{ PixelFormat::BC5_UNORM, PixelFormat::BC5_SNORM, }; -constexpr std::array VIEW_CLASS_BPTC_UNORM = { +constexpr std::array VIEW_CLASS_BPTC_UNORM{ PixelFormat::BC7_UNORM, PixelFormat::BC7_SRGB, }; -constexpr std::array VIEW_CLASS_BPTC_FLOAT = { +constexpr std::array VIEW_CLASS_BPTC_FLOAT{ PixelFormat::BC6H_SFLOAT, PixelFormat::BC6H_UFLOAT, }; +constexpr std::array VIEW_CLASS_ASTC_4x4_RGBA{ + PixelFormat::ASTC_2D_4X4_UNORM, + PixelFormat::ASTC_2D_4X4_SRGB, +}; + +constexpr std::array VIEW_CLASS_ASTC_5x4_RGBA{ + PixelFormat::ASTC_2D_5X4_UNORM, + PixelFormat::ASTC_2D_5X4_SRGB, +}; + +constexpr std::array VIEW_CLASS_ASTC_5x5_RGBA{ + PixelFormat::ASTC_2D_5X5_UNORM, + PixelFormat::ASTC_2D_5X5_SRGB, +}; + +constexpr std::array VIEW_CLASS_ASTC_6x5_RGBA{ + PixelFormat::ASTC_2D_6X5_UNORM, + PixelFormat::ASTC_2D_6X5_SRGB, +}; + +constexpr std::array VIEW_CLASS_ASTC_6x6_RGBA{ + PixelFormat::ASTC_2D_6X6_UNORM, + PixelFormat::ASTC_2D_6X6_SRGB, +}; + +constexpr std::array VIEW_CLASS_ASTC_8x5_RGBA{ + PixelFormat::ASTC_2D_8X5_UNORM, + PixelFormat::ASTC_2D_8X5_SRGB, +}; + +constexpr std::array VIEW_CLASS_ASTC_8x8_RGBA{ + PixelFormat::ASTC_2D_8X8_UNORM, + PixelFormat::ASTC_2D_8X8_SRGB, +}; + +// Missing formats: +// PixelFormat::ASTC_2D_10X5_UNORM +// PixelFormat::ASTC_2D_10X5_SRGB + +// Missing formats: +// PixelFormat::ASTC_2D_10X6_UNORM +// PixelFormat::ASTC_2D_10X6_SRGB + +constexpr std::array VIEW_CLASS_ASTC_10x8_RGBA{ + PixelFormat::ASTC_2D_10X8_UNORM, + PixelFormat::ASTC_2D_10X8_SRGB, +}; + +constexpr std::array VIEW_CLASS_ASTC_10x10_RGBA{ + PixelFormat::ASTC_2D_10X10_UNORM, + PixelFormat::ASTC_2D_10X10_SRGB, +}; + +// Missing formats +// ASTC_2D_12X10_UNORM, +// ASTC_2D_12X10_SRGB, + +constexpr std::array VIEW_CLASS_ASTC_12x12_RGBA{ + PixelFormat::ASTC_2D_12X12_UNORM, + PixelFormat::ASTC_2D_12X12_SRGB, +}; + // Compatibility table taken from Table 4.X.1 in: // https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_copy_image.txt -constexpr std::array COPY_CLASS_128_BITS = { +constexpr std::array COPY_CLASS_128_BITS{ PixelFormat::R32G32B32A32_UINT, PixelFormat::R32G32B32A32_FLOAT, PixelFormat::R32G32B32A32_SINT, PixelFormat::BC2_UNORM, PixelFormat::BC2_SRGB, PixelFormat::BC3_UNORM, PixelFormat::BC3_SRGB, PixelFormat::BC5_UNORM, PixelFormat::BC5_SNORM, @@ -97,7 +159,7 @@ constexpr std::array COPY_CLASS_128_BITS = { // PixelFormat::RGBA32I // COMPRESSED_RG_RGTC2 -constexpr std::array COPY_CLASS_64_BITS = { +constexpr std::array COPY_CLASS_64_BITS{ PixelFormat::R16G16B16A16_FLOAT, PixelFormat::R16G16B16A16_UINT, PixelFormat::R16G16B16A16_UNORM, PixelFormat::R16G16B16A16_SNORM, PixelFormat::R16G16B16A16_SINT, PixelFormat::R32G32_UINT, @@ -110,32 +172,36 @@ constexpr std::array COPY_CLASS_64_BITS = { // COMPRESSED_RGBA_S3TC_DXT1_EXT // COMPRESSED_SIGNED_RED_RGTC1 -void Enable(FormatCompatibility::Table& compatiblity, size_t format_a, size_t format_b) { - compatiblity[format_a][format_b] = true; - compatiblity[format_b][format_a] = true; +constexpr void Enable(Table& table, size_t format_a, size_t format_b) { + table[format_a][format_b / 64] |= u64(1) << (format_b % 64); + table[format_b][format_a / 64] |= u64(1) << (format_a % 64); } -void Enable(FormatCompatibility::Table& compatibility, PixelFormat format_a, PixelFormat format_b) { - Enable(compatibility, static_cast<size_t>(format_a), static_cast<size_t>(format_b)); +constexpr void Enable(Table& table, PixelFormat format_a, PixelFormat format_b) { + Enable(table, static_cast<size_t>(format_a), static_cast<size_t>(format_b)); } template <typename Range> -void EnableRange(FormatCompatibility::Table& compatibility, const Range& range) { +constexpr void EnableRange(Table& table, const Range& range) { for (auto it_a = range.begin(); it_a != range.end(); ++it_a) { for (auto it_b = it_a; it_b != range.end(); ++it_b) { - Enable(compatibility, *it_a, *it_b); + Enable(table, *it_a, *it_b); } } } -} // Anonymous namespace +constexpr bool IsSupported(const Table& table, PixelFormat format_a, PixelFormat format_b) { + const size_t a = static_cast<size_t>(format_a); + const size_t b = static_cast<size_t>(format_b); + return ((table[a][b / 64] >> (b % 64)) & 1) != 0; +} -FormatCompatibility::FormatCompatibility() { +constexpr Table MakeViewTable() { + Table view{}; for (size_t i = 0; i < MaxPixelFormat; ++i) { // Identity is allowed Enable(view, i, i); } - EnableRange(view, VIEW_CLASS_128_BITS); EnableRange(view, VIEW_CLASS_96_BITS); EnableRange(view, VIEW_CLASS_64_BITS); @@ -146,10 +212,39 @@ FormatCompatibility::FormatCompatibility() { EnableRange(view, VIEW_CLASS_RGTC2_RG); EnableRange(view, VIEW_CLASS_BPTC_UNORM); EnableRange(view, VIEW_CLASS_BPTC_FLOAT); + EnableRange(view, VIEW_CLASS_ASTC_4x4_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_5x4_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_5x5_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_6x5_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_6x6_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_8x5_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_8x8_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_10x8_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_10x10_RGBA); + EnableRange(view, VIEW_CLASS_ASTC_12x12_RGBA); + return view; +} - copy = view; +constexpr Table MakeCopyTable() { + Table copy = MakeViewTable(); EnableRange(copy, COPY_CLASS_128_BITS); EnableRange(copy, COPY_CLASS_64_BITS); + return copy; +} +} // Anonymous namespace + +bool IsViewCompatible(PixelFormat format_a, PixelFormat format_b, bool broken_views) { + if (broken_views) { + // If format views are broken, only accept formats that are identical. + return format_a == format_b; + } + static constexpr Table TABLE = MakeViewTable(); + return IsSupported(TABLE, format_a, format_b); +} + +bool IsCopyCompatible(PixelFormat format_a, PixelFormat format_b) { + static constexpr Table TABLE = MakeCopyTable(); + return IsSupported(TABLE, format_a, format_b); } } // namespace VideoCore::Surface diff --git a/src/video_core/compatible_formats.h b/src/video_core/compatible_formats.h index 51766349b..9a0522988 100644 --- a/src/video_core/compatible_formats.h +++ b/src/video_core/compatible_formats.h @@ -4,31 +4,12 @@ #pragma once -#include <array> -#include <bitset> -#include <cstddef> - #include "video_core/surface.h" namespace VideoCore::Surface { -class FormatCompatibility { -public: - using Table = std::array<std::bitset<MaxPixelFormat>, MaxPixelFormat>; - - explicit FormatCompatibility(); - - bool TestView(PixelFormat format_a, PixelFormat format_b) const noexcept { - return view[static_cast<size_t>(format_a)][static_cast<size_t>(format_b)]; - } - - bool TestCopy(PixelFormat format_a, PixelFormat format_b) const noexcept { - return copy[static_cast<size_t>(format_a)][static_cast<size_t>(format_b)]; - } +bool IsViewCompatible(PixelFormat format_a, PixelFormat format_b, bool broken_views); -private: - Table view; - Table copy; -}; +bool IsCopyCompatible(PixelFormat format_a, PixelFormat format_b); } // namespace VideoCore::Surface diff --git a/src/video_core/delayed_destruction_ring.h b/src/video_core/delayed_destruction_ring.h new file mode 100644 index 000000000..4f1d29c04 --- /dev/null +++ b/src/video_core/delayed_destruction_ring.h @@ -0,0 +1,32 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <cstddef> +#include <utility> +#include <vector> + +namespace VideoCommon { + +/// Container to push objects to be destroyed a few ticks in the future +template <typename T, size_t TICKS_TO_DESTROY> +class DelayedDestructionRing { +public: + void Tick() { + index = (index + 1) % TICKS_TO_DESTROY; + elements[index].clear(); + } + + void Push(T&& object) { + elements[index].push_back(std::move(object)); + } + +private: + size_t index = 0; + std::array<std::vector<T>, TICKS_TO_DESTROY> elements; +}; + +} // namespace VideoCommon diff --git a/src/video_core/dirty_flags.cpp b/src/video_core/dirty_flags.cpp index e16075993..b1eaac00c 100644 --- a/src/video_core/dirty_flags.cpp +++ b/src/video_core/dirty_flags.cpp @@ -9,13 +9,16 @@ #include "video_core/dirty_flags.h" #define OFF(field_name) MAXWELL3D_REG_INDEX(field_name) -#define NUM(field_name) (sizeof(::Tegra::Engines::Maxwell3D::Regs::field_name) / sizeof(u32)) +#define NUM(field_name) (sizeof(::Tegra::Engines::Maxwell3D::Regs::field_name) / (sizeof(u32))) namespace VideoCommon::Dirty { using Tegra::Engines::Maxwell3D; void SetupDirtyRenderTargets(Tegra::Engines::Maxwell3D::DirtyState::Tables& tables) { + FillBlock(tables[0], OFF(tic), NUM(tic), Descriptors); + FillBlock(tables[0], OFF(tsc), NUM(tsc), Descriptors); + static constexpr std::size_t num_per_rt = NUM(rt[0]); static constexpr std::size_t begin = OFF(rt); static constexpr std::size_t num = num_per_rt * Maxwell3D::Regs::NumRenderTargets; @@ -23,6 +26,10 @@ void SetupDirtyRenderTargets(Tegra::Engines::Maxwell3D::DirtyState::Tables& tabl FillBlock(tables[0], begin + rt * num_per_rt, num_per_rt, ColorBuffer0 + rt); } FillBlock(tables[1], begin, num, RenderTargets); + FillBlock(tables[0], OFF(render_area), NUM(render_area), RenderTargets); + + tables[0][OFF(rt_control)] = RenderTargets; + tables[1][OFF(rt_control)] = RenderTargetControl; static constexpr std::array zeta_flags{ZetaBuffer, RenderTargets}; for (std::size_t i = 0; i < std::size(zeta_flags); ++i) { diff --git a/src/video_core/dirty_flags.h b/src/video_core/dirty_flags.h index 3f6c1d83a..875527ddd 100644 --- a/src/video_core/dirty_flags.h +++ b/src/video_core/dirty_flags.h @@ -16,7 +16,10 @@ namespace VideoCommon::Dirty { enum : u8 { NullEntry = 0, + Descriptors, + RenderTargets, + RenderTargetControl, ColorBuffer0, ColorBuffer1, ColorBuffer2, diff --git a/src/video_core/dma_pusher.cpp b/src/video_core/dma_pusher.cpp index f2f96ac33..2c8b20024 100644 --- a/src/video_core/dma_pusher.cpp +++ b/src/video_core/dma_pusher.cpp @@ -2,6 +2,7 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include "common/cityhash.h" #include "common/microprofile.h" #include "core/core.h" #include "core/memory.h" @@ -12,7 +13,7 @@ namespace Tegra { -DmaPusher::DmaPusher(Core::System& system, GPU& gpu) : gpu{gpu}, system{system} {} +DmaPusher::DmaPusher(Core::System& system_, GPU& gpu_) : gpu{gpu_}, system{system_} {} DmaPusher::~DmaPusher() = default; @@ -45,32 +46,41 @@ bool DmaPusher::Step() { return false; } - const CommandList& command_list{dma_pushbuffer.front()}; - ASSERT_OR_EXECUTE(!command_list.empty(), { - // Somehow the command_list is empty, in order to avoid a crash - // We ignore it and assume its size is 0. - dma_pushbuffer.pop(); - dma_pushbuffer_subindex = 0; - return true; - }); - const CommandListHeader command_list_header{command_list[dma_pushbuffer_subindex++]}; - const GPUVAddr dma_get = command_list_header.addr; - - if (dma_pushbuffer_subindex >= command_list.size()) { - // We've gone through the current list, remove it from the queue - dma_pushbuffer.pop(); - dma_pushbuffer_subindex = 0; - } + CommandList& command_list{dma_pushbuffer.front()}; - if (command_list_header.size == 0) { - return true; - } + ASSERT_OR_EXECUTE( + command_list.command_lists.size() || command_list.prefetch_command_list.size(), { + // Somehow the command_list is empty, in order to avoid a crash + // We ignore it and assume its size is 0. + dma_pushbuffer.pop(); + dma_pushbuffer_subindex = 0; + return true; + }); + + if (command_list.prefetch_command_list.size()) { + // Prefetched command list from nvdrv, used for things like synchronization + command_headers = std::move(command_list.prefetch_command_list); + dma_pushbuffer.pop(); + } else { + const CommandListHeader command_list_header{ + command_list.command_lists[dma_pushbuffer_subindex++]}; + const GPUVAddr dma_get = command_list_header.addr; + + if (dma_pushbuffer_subindex >= command_list.command_lists.size()) { + // We've gone through the current list, remove it from the queue + dma_pushbuffer.pop(); + dma_pushbuffer_subindex = 0; + } - // Push buffer non-empty, read a word - command_headers.resize(command_list_header.size); - gpu.MemoryManager().ReadBlockUnsafe(dma_get, command_headers.data(), - command_list_header.size * sizeof(u32)); + if (command_list_header.size == 0) { + return true; + } + // Push buffer non-empty, read a word + command_headers.resize(command_list_header.size); + gpu.MemoryManager().ReadBlockUnsafe(dma_get, command_headers.data(), + command_list_header.size * sizeof(u32)); + } for (std::size_t index = 0; index < command_headers.size();) { const CommandHeader& command_header = command_headers[index]; @@ -142,7 +152,12 @@ void DmaPusher::SetState(const CommandHeader& command_header) { void DmaPusher::CallMethod(u32 argument) const { if (dma_state.method < non_puller_methods) { - gpu.CallMethod({dma_state.method, argument, dma_state.subchannel, dma_state.method_count}); + gpu.CallMethod(GPU::MethodCall{ + dma_state.method, + argument, + dma_state.subchannel, + dma_state.method_count, + }); } else { subchannels[dma_state.subchannel]->CallMethod(dma_state.method, argument, dma_state.is_last_call); diff --git a/src/video_core/dma_pusher.h b/src/video_core/dma_pusher.h index efa90d170..19f286fa7 100644 --- a/src/video_core/dma_pusher.h +++ b/src/video_core/dma_pusher.h @@ -18,6 +18,8 @@ class System; namespace Tegra { +class GPU; + enum class SubmissionMode : u32 { IncreasingOld = 0, Increasing = 1, @@ -27,6 +29,31 @@ enum class SubmissionMode : u32 { IncreaseOnce = 5 }; +// Note that, traditionally, methods are treated as 4-byte addressable locations, and hence +// their numbers are written down multiplied by 4 in Docs. Here we are not multiply by 4. +// So the values you see in docs might be multiplied by 4. +enum class BufferMethods : u32 { + BindObject = 0x0, + Nop = 0x2, + SemaphoreAddressHigh = 0x4, + SemaphoreAddressLow = 0x5, + SemaphoreSequence = 0x6, + SemaphoreTrigger = 0x7, + NotifyIntr = 0x8, + WrcacheFlush = 0x9, + Unk28 = 0xA, + UnkCacheFlush = 0xB, + RefCnt = 0x14, + SemaphoreAcquire = 0x1A, + SemaphoreRelease = 0x1B, + FenceValue = 0x1C, + FenceAction = 0x1D, + WaitForInterrupt = 0x1E, + Unk7c = 0x1F, + Yield = 0x20, + NonPullerMethods = 0x40, +}; + struct CommandListHeader { union { u64 raw; @@ -49,9 +76,23 @@ union CommandHeader { static_assert(std::is_standard_layout_v<CommandHeader>, "CommandHeader is not standard layout"); static_assert(sizeof(CommandHeader) == sizeof(u32), "CommandHeader has incorrect size!"); -class GPU; +inline CommandHeader BuildCommandHeader(BufferMethods method, u32 arg_count, SubmissionMode mode) { + CommandHeader result{}; + result.method.Assign(static_cast<u32>(method)); + result.arg_count.Assign(arg_count); + result.mode.Assign(mode); + return result; +} -using CommandList = std::vector<Tegra::CommandListHeader>; +struct CommandList final { + CommandList() = default; + explicit CommandList(std::size_t size) : command_lists(size) {} + explicit CommandList(std::vector<CommandHeader>&& prefetch_command_list_) + : prefetch_command_list{std::move(prefetch_command_list_)} {} + + std::vector<CommandListHeader> command_lists; + std::vector<CommandHeader> prefetch_command_list; +}; /** * The DmaPusher class implements DMA submission to FIFOs, providing an area of memory that the @@ -60,9 +101,9 @@ using CommandList = std::vector<Tegra::CommandListHeader>; * See https://envytools.readthedocs.io/en/latest/hw/fifo/dma-pusher.html#fifo-dma-pusher for * details on this implementation. */ -class DmaPusher { +class DmaPusher final { public: - explicit DmaPusher(Core::System& system, GPU& gpu); + explicit DmaPusher(Core::System& system_, GPU& gpu_); ~DmaPusher(); void Push(CommandList&& entries) { @@ -71,7 +112,7 @@ public: void DispatchCalls(); - void BindSubchannel(Tegra::Engines::EngineInterface* engine, u32 subchannel_id) { + void BindSubchannel(Engines::EngineInterface* engine, u32 subchannel_id) { subchannels[subchannel_id] = engine; } @@ -104,7 +145,7 @@ private: bool ib_enable{true}; ///< IB mode enabled - std::array<Tegra::Engines::EngineInterface*, max_subchannels> subchannels{}; + std::array<Engines::EngineInterface*, max_subchannels> subchannels{}; GPU& gpu; Core::System& system; diff --git a/src/video_core/engines/engine_upload.cpp b/src/video_core/engines/engine_upload.cpp index d44ad0cd8..71d7e1473 100644 --- a/src/video_core/engines/engine_upload.cpp +++ b/src/video_core/engines/engine_upload.cpp @@ -11,16 +11,16 @@ namespace Tegra::Engines::Upload { -State::State(MemoryManager& memory_manager, Registers& regs) - : regs{regs}, memory_manager{memory_manager} {} +State::State(MemoryManager& memory_manager_, Registers& regs_) + : regs{regs_}, memory_manager{memory_manager_} {} State::~State() = default; -void State::ProcessExec(const bool is_linear) { +void State::ProcessExec(const bool is_linear_) { write_offset = 0; copy_size = regs.line_length_in * regs.line_count; inner_buffer.resize(copy_size); - this->is_linear = is_linear; + is_linear = is_linear_; } void State::ProcessData(const u32 data, const bool is_last_call) { diff --git a/src/video_core/engines/engine_upload.h b/src/video_core/engines/engine_upload.h index 462da419e..1c7f1effa 100644 --- a/src/video_core/engines/engine_upload.h +++ b/src/video_core/engines/engine_upload.h @@ -54,10 +54,10 @@ struct Registers { class State { public: - State(MemoryManager& memory_manager, Registers& regs); + explicit State(MemoryManager& memory_manager_, Registers& regs_); ~State(); - void ProcessExec(bool is_linear); + void ProcessExec(bool is_linear_); void ProcessData(u32 data, bool is_last_call); private: diff --git a/src/video_core/engines/fermi_2d.cpp b/src/video_core/engines/fermi_2d.cpp index 9409c4075..a01d334ad 100644 --- a/src/video_core/engines/fermi_2d.cpp +++ b/src/video_core/engines/fermi_2d.cpp @@ -10,7 +10,11 @@ namespace Tegra::Engines { -Fermi2D::Fermi2D() = default; +Fermi2D::Fermi2D() { + // Nvidia's OpenGL driver seems to assume these values + regs.src.depth = 1; + regs.dst.depth = 1; +} Fermi2D::~Fermi2D() = default; @@ -21,79 +25,43 @@ void Fermi2D::BindRasterizer(VideoCore::RasterizerInterface& rasterizer_) { void Fermi2D::CallMethod(u32 method, u32 method_argument, bool is_last_call) { ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Fermi2D register, increase the size of the Regs structure"); - regs.reg_array[method] = method_argument; - switch (method) { - // Trigger the surface copy on the last register write. This is blit_src_y, but this is 64-bit, - // so trigger on the second 32-bit write. - case FERMI2D_REG_INDEX(blit_src_y) + 1: { - HandleSurfaceCopy(); - break; - } + if (method == FERMI2D_REG_INDEX(pixels_from_memory.src_y0) + 1) { + Blit(); } } void Fermi2D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) { - for (std::size_t i = 0; i < amount; i++) { - CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); + for (u32 i = 0; i < amount; ++i) { + CallMethod(method, base_start[i], methods_pending - i <= 1); } } -static std::pair<u32, u32> DelimitLine(u32 src_1, u32 src_2, u32 dst_1, u32 dst_2, u32 src_line) { - const u32 line_a = src_2 - src_1; - const u32 line_b = dst_2 - dst_1; - const u32 excess = std::max<s32>(0, line_a - src_line + src_1); - return {line_b - (excess * line_b) / line_a, excess}; -} - -void Fermi2D::HandleSurfaceCopy() { - LOG_DEBUG(HW_GPU, "Requested a surface copy with operation {}", - static_cast<u32>(regs.operation)); +void Fermi2D::Blit() { + LOG_DEBUG(HW_GPU, "called. source address=0x{:x}, destination address=0x{:x}", + regs.src.Address(), regs.dst.Address()); - // TODO(Subv): Only raw copies are implemented. - ASSERT(regs.operation == Operation::SrcCopy); + UNIMPLEMENTED_IF_MSG(regs.operation != Operation::SrcCopy, "Operation is not copy"); + UNIMPLEMENTED_IF_MSG(regs.src.layer != 0, "Source layer is not zero"); + UNIMPLEMENTED_IF_MSG(regs.dst.layer != 0, "Destination layer is not zero"); + UNIMPLEMENTED_IF_MSG(regs.src.depth != 1, "Source depth is not one"); + UNIMPLEMENTED_IF_MSG(regs.clip_enable != 0, "Clipped blit enabled"); - const u32 src_blit_x1{static_cast<u32>(regs.blit_src_x >> 32)}; - const u32 src_blit_y1{static_cast<u32>(regs.blit_src_y >> 32)}; - u32 src_blit_x2, src_blit_y2; - if (regs.blit_control.origin == Origin::Corner) { - src_blit_x2 = - static_cast<u32>((regs.blit_src_x + (regs.blit_du_dx * regs.blit_dst_width)) >> 32); - src_blit_y2 = - static_cast<u32>((regs.blit_src_y + (regs.blit_dv_dy * regs.blit_dst_height)) >> 32); - } else { - src_blit_x2 = static_cast<u32>((regs.blit_src_x >> 32) + regs.blit_dst_width); - src_blit_y2 = static_cast<u32>((regs.blit_src_y >> 32) + regs.blit_dst_height); - } - u32 dst_blit_x2 = regs.blit_dst_x + regs.blit_dst_width; - u32 dst_blit_y2 = regs.blit_dst_y + regs.blit_dst_height; - const auto [new_dst_w, src_excess_x] = - DelimitLine(src_blit_x1, src_blit_x2, regs.blit_dst_x, dst_blit_x2, regs.src.width); - const auto [new_dst_h, src_excess_y] = - DelimitLine(src_blit_y1, src_blit_y2, regs.blit_dst_y, dst_blit_y2, regs.src.height); - dst_blit_x2 = new_dst_w + regs.blit_dst_x; - src_blit_x2 = src_blit_x2 - src_excess_x; - dst_blit_y2 = new_dst_h + regs.blit_dst_y; - src_blit_y2 = src_blit_y2 - src_excess_y; - const auto [new_src_w, dst_excess_x] = - DelimitLine(regs.blit_dst_x, dst_blit_x2, src_blit_x1, src_blit_x2, regs.dst.width); - const auto [new_src_h, dst_excess_y] = - DelimitLine(regs.blit_dst_y, dst_blit_y2, src_blit_y1, src_blit_y2, regs.dst.height); - src_blit_x2 = new_src_w + src_blit_x1; - dst_blit_x2 = dst_blit_x2 - dst_excess_x; - src_blit_y2 = new_src_h + src_blit_y1; - dst_blit_y2 = dst_blit_y2 - dst_excess_y; - const Common::Rectangle<u32> src_rect{src_blit_x1, src_blit_y1, src_blit_x2, src_blit_y2}; - const Common::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y, dst_blit_x2, - dst_blit_y2}; - const Config copy_config{ + const auto& args = regs.pixels_from_memory; + const Config config{ .operation = regs.operation, - .filter = regs.blit_control.filter, - .src_rect = src_rect, - .dst_rect = dst_rect, + .filter = args.sample_mode.filter, + .dst_x0 = args.dst_x0, + .dst_y0 = args.dst_y0, + .dst_x1 = args.dst_x0 + args.dst_width, + .dst_y1 = args.dst_y0 + args.dst_height, + .src_x0 = static_cast<s32>(args.src_x0 >> 32), + .src_y0 = static_cast<s32>(args.src_y0 >> 32), + .src_x1 = static_cast<s32>((args.du_dx * args.dst_width + args.src_x0) >> 32), + .src_y1 = static_cast<s32>((args.dv_dy * args.dst_height + args.src_y0) >> 32), }; - if (!rasterizer->AccelerateSurfaceCopy(regs.src, regs.dst, copy_config)) { + if (!rasterizer->AccelerateSurfaceCopy(regs.src, regs.dst, config)) { UNIMPLEMENTED(); } } diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h index 0909709ec..81522988e 100644 --- a/src/video_core/engines/fermi_2d.h +++ b/src/video_core/engines/fermi_2d.h @@ -53,8 +53,8 @@ public: }; enum class Filter : u32 { - PointSample = 0, // Nearest - Linear = 1, + Point = 0, + Bilinear = 1, }; enum class Operation : u32 { @@ -67,88 +67,235 @@ public: BlendPremult = 6, }; - struct Regs { - static constexpr std::size_t NUM_REGS = 0x258; + enum class MemoryLayout : u32 { + BlockLinear = 0, + Pitch = 1, + }; - struct Surface { - RenderTargetFormat format; - BitField<0, 1, u32> linear; - union { - BitField<0, 4, u32> block_width; - BitField<4, 4, u32> block_height; - BitField<8, 4, u32> block_depth; - }; - u32 depth; - u32 layer; - u32 pitch; - u32 width; - u32 height; - u32 address_high; - u32 address_low; - - GPUVAddr Address() const { - return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | - address_low); - } - - u32 BlockWidth() const { - return block_width.Value(); - } - - u32 BlockHeight() const { - return block_height.Value(); - } - - u32 BlockDepth() const { - return block_depth.Value(); - } - }; - static_assert(sizeof(Surface) == 0x28, "Surface has incorrect size"); + enum class CpuIndexWrap : u32 { + Wrap = 0, + NoWrap = 1, + }; + struct Surface { + RenderTargetFormat format; + MemoryLayout linear; union { - struct { - INSERT_UNION_PADDING_WORDS(0x80); + BitField<0, 4, u32> block_width; + BitField<4, 4, u32> block_height; + BitField<8, 4, u32> block_depth; + }; + u32 depth; + u32 layer; + u32 pitch; + u32 width; + u32 height; + u32 addr_upper; + u32 addr_lower; + + [[nodiscard]] constexpr GPUVAddr Address() const noexcept { + return (static_cast<GPUVAddr>(addr_upper) << 32) | static_cast<GPUVAddr>(addr_lower); + } + }; + static_assert(sizeof(Surface) == 0x28, "Surface has incorrect size"); - Surface dst; + enum class SectorPromotion : u32 { + NoPromotion = 0, + PromoteTo2V = 1, + PromoteTo2H = 2, + PromoteTo4 = 3, + }; + + enum class NumTpcs : u32 { + All = 0, + One = 1, + }; - INSERT_UNION_PADDING_WORDS(2); + enum class RenderEnableMode : u32 { + False = 0, + True = 1, + Conditional = 2, + RenderIfEqual = 3, + RenderIfNotEqual = 4, + }; - Surface src; + enum class ColorKeyFormat : u32 { + A16R56G6B5 = 0, + A1R5G55B5 = 1, + A8R8G8B8 = 2, + A2R10G10B10 = 3, + Y8 = 4, + Y16 = 5, + Y32 = 6, + }; - INSERT_UNION_PADDING_WORDS(0x15); + union Beta4 { + BitField<0, 8, u32> b; + BitField<8, 8, u32> g; + BitField<16, 8, u32> r; + BitField<24, 8, u32> a; + }; - Operation operation; + struct Point { + u32 x; + u32 y; + }; - INSERT_UNION_PADDING_WORDS(0x177); + enum class PatternSelect : u32 { + MonoChrome8x8 = 0, + MonoChrome64x1 = 1, + MonoChrome1x64 = 2, + Color = 3, + }; + enum class NotifyType : u32 { + WriteOnly = 0, + WriteThenAwaken = 1, + }; + + enum class MonochromePatternColorFormat : u32 { + A8X8R8G6B5 = 0, + A1R5G5B5 = 1, + A8R8G8B8 = 2, + A8Y8 = 3, + A8X8Y16 = 4, + Y32 = 5, + }; + + enum class MonochromePatternFormat : u32 { + CGA6_M1 = 0, + LE_M1 = 1, + }; + + union Regs { + static constexpr std::size_t NUM_REGS = 0x258; + struct { + u32 object; + INSERT_UNION_PADDING_WORDS(0x3F); + u32 no_operation; + NotifyType notify; + INSERT_UNION_PADDING_WORDS(0x2); + u32 wait_for_idle; + INSERT_UNION_PADDING_WORDS(0xB); + u32 pm_trigger; + INSERT_UNION_PADDING_WORDS(0xF); + u32 context_dma_notify; + u32 dst_context_dma; + u32 src_context_dma; + u32 semaphore_context_dma; + INSERT_UNION_PADDING_WORDS(0x1C); + Surface dst; + CpuIndexWrap pixels_from_cpu_index_wrap; + u32 kind2d_check_enable; + Surface src; + SectorPromotion pixels_from_memory_sector_promotion; + INSERT_UNION_PADDING_WORDS(0x1); + NumTpcs num_tpcs; + u32 render_enable_addr_upper; + u32 render_enable_addr_lower; + RenderEnableMode render_enable_mode; + INSERT_UNION_PADDING_WORDS(0x4); + u32 clip_x0; + u32 clip_y0; + u32 clip_width; + u32 clip_height; + BitField<0, 1, u32> clip_enable; + BitField<0, 3, ColorKeyFormat> color_key_format; + u32 color_key; + BitField<0, 1, u32> color_key_enable; + BitField<0, 8, u32> rop; + u32 beta1; + Beta4 beta4; + Operation operation; + union { + BitField<0, 6, u32> x; + BitField<8, 6, u32> y; + } pattern_offset; + BitField<0, 2, PatternSelect> pattern_select; + INSERT_UNION_PADDING_WORDS(0xC); + struct { + BitField<0, 3, MonochromePatternColorFormat> color_format; + BitField<0, 1, MonochromePatternFormat> format; + u32 color0; + u32 color1; + u32 pattern0; + u32 pattern1; + } monochrome_pattern; + struct { + std::array<u32, 0x40> X8R8G8B8; + std::array<u32, 0x20> R5G6B5; + std::array<u32, 0x20> X1R5G5B5; + std::array<u32, 0x10> Y8; + } color_pattern; + INSERT_UNION_PADDING_WORDS(0x10); + struct { + u32 prim_mode; + u32 prim_color_format; + u32 prim_color; + u32 line_tie_break_bits; + INSERT_UNION_PADDING_WORDS(0x14); + u32 prim_point_xy; + INSERT_UNION_PADDING_WORDS(0x7); + std::array<Point, 0x40> prim_point; + } render_solid; + struct { + u32 data_type; + u32 color_format; + u32 index_format; + u32 mono_format; + u32 wrap; + u32 color0; + u32 color1; + u32 mono_opacity; + INSERT_UNION_PADDING_WORDS(0x6); + u32 src_width; + u32 src_height; + u32 dx_du_frac; + u32 dx_du_int; + u32 dx_dv_frac; + u32 dy_dv_int; + u32 dst_x0_frac; + u32 dst_x0_int; + u32 dst_y0_frac; + u32 dst_y0_int; + u32 data; + } pixels_from_cpu; + INSERT_UNION_PADDING_WORDS(0x3); + u32 big_endian_control; + INSERT_UNION_PADDING_WORDS(0x3); + struct { + BitField<0, 3, u32> block_shape; + BitField<0, 5, u32> corral_size; + BitField<0, 1, u32> safe_overlap; union { - u32 raw; BitField<0, 1, Origin> origin; BitField<4, 1, Filter> filter; - } blit_control; - + } sample_mode; INSERT_UNION_PADDING_WORDS(0x8); - - u32 blit_dst_x; - u32 blit_dst_y; - u32 blit_dst_width; - u32 blit_dst_height; - u64 blit_du_dx; - u64 blit_dv_dy; - u64 blit_src_x; - u64 blit_src_y; - - INSERT_UNION_PADDING_WORDS(0x21); - }; - std::array<u32, NUM_REGS> reg_array; + s32 dst_x0; + s32 dst_y0; + s32 dst_width; + s32 dst_height; + s64 du_dx; + s64 dv_dy; + s64 src_x0; + s64 src_y0; + } pixels_from_memory; }; + std::array<u32, NUM_REGS> reg_array; } regs{}; struct Config { - Operation operation{}; - Filter filter{}; - Common::Rectangle<u32> src_rect; - Common::Rectangle<u32> dst_rect; + Operation operation; + Filter filter; + s32 dst_x0; + s32 dst_y0; + s32 dst_x1; + s32 dst_y1; + s32 src_x0; + s32 src_y0; + s32 src_x1; + s32 src_y1; }; private: @@ -156,25 +303,49 @@ private: /// Performs the copy from the source surface to the destination surface as configured in the /// registers. - void HandleSurfaceCopy(); + void Blit(); }; #define ASSERT_REG_POSITION(field_name, position) \ - static_assert(offsetof(Fermi2D::Regs, field_name) == position * 4, \ + static_assert(offsetof(Fermi2D::Regs, field_name) == position, \ "Field " #field_name " has invalid position") -ASSERT_REG_POSITION(dst, 0x80); -ASSERT_REG_POSITION(src, 0x8C); -ASSERT_REG_POSITION(operation, 0xAB); -ASSERT_REG_POSITION(blit_control, 0x223); -ASSERT_REG_POSITION(blit_dst_x, 0x22c); -ASSERT_REG_POSITION(blit_dst_y, 0x22d); -ASSERT_REG_POSITION(blit_dst_width, 0x22e); -ASSERT_REG_POSITION(blit_dst_height, 0x22f); -ASSERT_REG_POSITION(blit_du_dx, 0x230); -ASSERT_REG_POSITION(blit_dv_dy, 0x232); -ASSERT_REG_POSITION(blit_src_x, 0x234); -ASSERT_REG_POSITION(blit_src_y, 0x236); +ASSERT_REG_POSITION(object, 0x0); +ASSERT_REG_POSITION(no_operation, 0x100); +ASSERT_REG_POSITION(notify, 0x104); +ASSERT_REG_POSITION(wait_for_idle, 0x110); +ASSERT_REG_POSITION(pm_trigger, 0x140); +ASSERT_REG_POSITION(context_dma_notify, 0x180); +ASSERT_REG_POSITION(dst_context_dma, 0x184); +ASSERT_REG_POSITION(src_context_dma, 0x188); +ASSERT_REG_POSITION(semaphore_context_dma, 0x18C); +ASSERT_REG_POSITION(dst, 0x200); +ASSERT_REG_POSITION(pixels_from_cpu_index_wrap, 0x228); +ASSERT_REG_POSITION(kind2d_check_enable, 0x22C); +ASSERT_REG_POSITION(src, 0x230); +ASSERT_REG_POSITION(pixels_from_memory_sector_promotion, 0x258); +ASSERT_REG_POSITION(num_tpcs, 0x260); +ASSERT_REG_POSITION(render_enable_addr_upper, 0x264); +ASSERT_REG_POSITION(render_enable_addr_lower, 0x268); +ASSERT_REG_POSITION(clip_x0, 0x280); +ASSERT_REG_POSITION(clip_y0, 0x284); +ASSERT_REG_POSITION(clip_width, 0x288); +ASSERT_REG_POSITION(clip_height, 0x28c); +ASSERT_REG_POSITION(clip_enable, 0x290); +ASSERT_REG_POSITION(color_key_format, 0x294); +ASSERT_REG_POSITION(color_key, 0x298); +ASSERT_REG_POSITION(rop, 0x2A0); +ASSERT_REG_POSITION(beta1, 0x2A4); +ASSERT_REG_POSITION(beta4, 0x2A8); +ASSERT_REG_POSITION(operation, 0x2AC); +ASSERT_REG_POSITION(pattern_offset, 0x2B0); +ASSERT_REG_POSITION(pattern_select, 0x2B4); +ASSERT_REG_POSITION(monochrome_pattern, 0x2E8); +ASSERT_REG_POSITION(color_pattern, 0x300); +ASSERT_REG_POSITION(render_solid, 0x580); +ASSERT_REG_POSITION(pixels_from_cpu, 0x800); +ASSERT_REG_POSITION(big_endian_control, 0x870); +ASSERT_REG_POSITION(pixels_from_memory, 0x880); #undef ASSERT_REG_POSITION diff --git a/src/video_core/engines/kepler_compute.cpp b/src/video_core/engines/kepler_compute.cpp index 898370739..ba387506e 100644 --- a/src/video_core/engines/kepler_compute.cpp +++ b/src/video_core/engines/kepler_compute.cpp @@ -58,24 +58,6 @@ void KeplerCompute::CallMultiMethod(u32 method, const u32* base_start, u32 amoun } } -Texture::FullTextureInfo KeplerCompute::GetTexture(std::size_t offset) const { - const std::bitset<8> cbuf_mask = launch_description.const_buffer_enable_mask.Value(); - ASSERT(cbuf_mask[regs.tex_cb_index]); - - const auto& texinfo = launch_description.const_buffer_config[regs.tex_cb_index]; - ASSERT(texinfo.Address() != 0); - - const GPUVAddr address = texinfo.Address() + offset * sizeof(Texture::TextureHandle); - ASSERT(address < texinfo.Address() + texinfo.size); - - const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(address)}; - return GetTextureInfo(tex_handle); -} - -Texture::FullTextureInfo KeplerCompute::GetTextureInfo(Texture::TextureHandle tex_handle) const { - return Texture::FullTextureInfo{GetTICEntry(tex_handle.tic_id), GetTSCEntry(tex_handle.tsc_id)}; -} - u32 KeplerCompute::AccessConstBuffer32(ShaderType stage, u64 const_buffer, u64 offset) const { ASSERT(stage == ShaderType::Compute); const auto& buffer = launch_description.const_buffer_config[const_buffer]; @@ -98,9 +80,11 @@ SamplerDescriptor KeplerCompute::AccessBindlessSampler(ShaderType stage, u64 con SamplerDescriptor KeplerCompute::AccessSampler(u32 handle) const { const Texture::TextureHandle tex_handle{handle}; - const Texture::FullTextureInfo tex_info = GetTextureInfo(tex_handle); - SamplerDescriptor result = SamplerDescriptor::FromTIC(tex_info.tic); - result.is_shadow.Assign(tex_info.tsc.depth_compare_enabled.Value()); + const Texture::TICEntry tic = GetTICEntry(tex_handle.tic_id); + const Texture::TSCEntry tsc = GetTSCEntry(tex_handle.tsc_id); + + SamplerDescriptor result = SamplerDescriptor::FromTIC(tic); + result.is_shadow.Assign(tsc.depth_compare_enabled.Value()); return result; } diff --git a/src/video_core/engines/kepler_compute.h b/src/video_core/engines/kepler_compute.h index 7f2500aab..51a041202 100644 --- a/src/video_core/engines/kepler_compute.h +++ b/src/video_core/engines/kepler_compute.h @@ -209,11 +209,6 @@ public: void CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32 methods_pending) override; - Texture::FullTextureInfo GetTexture(std::size_t offset) const; - - /// Given a texture handle, returns the TSC and TIC entries. - Texture::FullTextureInfo GetTextureInfo(Texture::TextureHandle tex_handle) const; - u32 AccessConstBuffer32(ShaderType stage, u64 const_buffer, u64 offset) const override; SamplerDescriptor AccessBoundSampler(ShaderType stage, u64 offset) const override; diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp index dc71b2eec..9911140e9 100644 --- a/src/video_core/engines/kepler_memory.cpp +++ b/src/video_core/engines/kepler_memory.cpp @@ -14,8 +14,8 @@ namespace Tegra::Engines { -KeplerMemory::KeplerMemory(Core::System& system, MemoryManager& memory_manager) - : system{system}, upload_state{memory_manager, regs.upload} {} +KeplerMemory::KeplerMemory(Core::System& system_, MemoryManager& memory_manager) + : system{system_}, upload_state{memory_manager, regs.upload} {} KeplerMemory::~KeplerMemory() = default; diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h index 5b7f71a00..62483589e 100644 --- a/src/video_core/engines/kepler_memory.h +++ b/src/video_core/engines/kepler_memory.h @@ -35,7 +35,7 @@ namespace Tegra::Engines { class KeplerMemory final : public EngineInterface { public: - KeplerMemory(Core::System& system, MemoryManager& memory_manager); + explicit KeplerMemory(Core::System& system_, MemoryManager& memory_manager); ~KeplerMemory(); /// Write the value to the register identified by method. diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index 57ebc785f..9be651e24 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -2,7 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <cinttypes> #include <cstring> #include <optional> #include "common/assert.h" @@ -124,6 +123,116 @@ void Maxwell3D::InitializeRegisterDefaults() { mme_inline[MAXWELL3D_REG_INDEX(index_array.count)] = true; } +void Maxwell3D::ProcessMacro(u32 method, const u32* base_start, u32 amount, bool is_last_call) { + if (executing_macro == 0) { + // A macro call must begin by writing the macro method's register, not its argument. + ASSERT_MSG((method % 2) == 0, + "Can't start macro execution by writing to the ARGS register"); + executing_macro = method; + } + + macro_params.insert(macro_params.end(), base_start, base_start + amount); + + // Call the macro when there are no more parameters in the command buffer + if (is_last_call) { + CallMacroMethod(executing_macro, macro_params); + macro_params.clear(); + } +} + +u32 Maxwell3D::ProcessShadowRam(u32 method, u32 argument) { + // Keep track of the register value in shadow_state when requested. + const auto control = shadow_state.shadow_ram_control; + if (control == Regs::ShadowRamControl::Track || + control == Regs::ShadowRamControl::TrackWithFilter) { + shadow_state.reg_array[method] = argument; + return argument; + } + if (control == Regs::ShadowRamControl::Replay) { + return shadow_state.reg_array[method]; + } + return argument; +} + +void Maxwell3D::ProcessDirtyRegisters(u32 method, u32 argument) { + if (regs.reg_array[method] == argument) { + return; + } + regs.reg_array[method] = argument; + + for (const auto& table : dirty.tables) { + dirty.flags[table[method]] = true; + } +} + +void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argument, + bool is_last_call) { + switch (method) { + case MAXWELL3D_REG_INDEX(wait_for_idle): + return rasterizer->WaitForIdle(); + case MAXWELL3D_REG_INDEX(shadow_ram_control): + shadow_state.shadow_ram_control = static_cast<Regs::ShadowRamControl>(nonshadow_argument); + return; + case MAXWELL3D_REG_INDEX(macros.data): + return macro_engine->AddCode(regs.macros.upload_address, argument); + case MAXWELL3D_REG_INDEX(macros.bind): + return ProcessMacroBind(argument); + case MAXWELL3D_REG_INDEX(firmware[4]): + return ProcessFirmwareCall4(); + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[3]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[4]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[5]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[6]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[7]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[8]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[9]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[10]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[11]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[12]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[13]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[14]): + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): + return StartCBData(method); + case MAXWELL3D_REG_INDEX(cb_bind[0]): + return ProcessCBBind(0); + case MAXWELL3D_REG_INDEX(cb_bind[1]): + return ProcessCBBind(1); + case MAXWELL3D_REG_INDEX(cb_bind[2]): + return ProcessCBBind(2); + case MAXWELL3D_REG_INDEX(cb_bind[3]): + return ProcessCBBind(3); + case MAXWELL3D_REG_INDEX(cb_bind[4]): + return ProcessCBBind(4); + case MAXWELL3D_REG_INDEX(draw.vertex_end_gl): + return DrawArrays(); + case MAXWELL3D_REG_INDEX(clear_buffers): + return ProcessClearBuffers(); + case MAXWELL3D_REG_INDEX(query.query_get): + return ProcessQueryGet(); + case MAXWELL3D_REG_INDEX(condition.mode): + return ProcessQueryCondition(); + case MAXWELL3D_REG_INDEX(counter_reset): + return ProcessCounterReset(); + case MAXWELL3D_REG_INDEX(sync_info): + return ProcessSyncPoint(); + case MAXWELL3D_REG_INDEX(exec_upload): + return upload_state.ProcessExec(regs.exec_upload.linear != 0); + case MAXWELL3D_REG_INDEX(data_upload): + upload_state.ProcessData(argument, is_last_call); + if (is_last_call) { + OnMemoryWrite(); + } + return; + case MAXWELL3D_REG_INDEX(fragment_barrier): + return rasterizer->FragmentBarrier(); + case MAXWELL3D_REG_INDEX(tiled_cache_barrier): + return rasterizer->TiledCacheBarrier(); + } +} + void Maxwell3D::CallMacroMethod(u32 method, const std::vector<u32>& parameters) { // Reset the current macro. executing_macro = 0; @@ -157,142 +266,16 @@ void Maxwell3D::CallMethod(u32 method, u32 method_argument, bool is_last_call) { // Methods after 0xE00 are special, they're actually triggers for some microcode that was // uploaded to the GPU during initialization. if (method >= MacroRegistersStart) { - // We're trying to execute a macro - if (executing_macro == 0) { - // A macro call must begin by writing the macro method's register, not its argument. - ASSERT_MSG((method % 2) == 0, - "Can't start macro execution by writing to the ARGS register"); - executing_macro = method; - } - - macro_params.push_back(method_argument); - - // Call the macro when there are no more parameters in the command buffer - if (is_last_call) { - CallMacroMethod(executing_macro, macro_params); - macro_params.clear(); - } + ProcessMacro(method, &method_argument, 1, is_last_call); return; } ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Maxwell3D register, increase the size of the Regs structure"); - u32 arg = method_argument; - // Keep track of the register value in shadow_state when requested. - if (shadow_state.shadow_ram_control == Regs::ShadowRamControl::Track || - shadow_state.shadow_ram_control == Regs::ShadowRamControl::TrackWithFilter) { - shadow_state.reg_array[method] = arg; - } else if (shadow_state.shadow_ram_control == Regs::ShadowRamControl::Replay) { - arg = shadow_state.reg_array[method]; - } - - if (regs.reg_array[method] != arg) { - regs.reg_array[method] = arg; - - for (const auto& table : dirty.tables) { - dirty.flags[table[method]] = true; - } - } - - switch (method) { - case MAXWELL3D_REG_INDEX(wait_for_idle): { - rasterizer->WaitForIdle(); - break; - } - case MAXWELL3D_REG_INDEX(shadow_ram_control): { - shadow_state.shadow_ram_control = static_cast<Regs::ShadowRamControl>(method_argument); - break; - } - case MAXWELL3D_REG_INDEX(macros.data): { - macro_engine->AddCode(regs.macros.upload_address, arg); - break; - } - case MAXWELL3D_REG_INDEX(macros.bind): { - ProcessMacroBind(arg); - break; - } - case MAXWELL3D_REG_INDEX(firmware[4]): { - ProcessFirmwareCall4(); - break; - } - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[0]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[1]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[2]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[3]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[4]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[5]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[6]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[7]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[8]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[9]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[10]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[11]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[12]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[13]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[14]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): { - StartCBData(method); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[0]): { - ProcessCBBind(0); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[1]): { - ProcessCBBind(1); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[2]): { - ProcessCBBind(2); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[3]): { - ProcessCBBind(3); - break; - } - case MAXWELL3D_REG_INDEX(cb_bind[4]): { - ProcessCBBind(4); - break; - } - case MAXWELL3D_REG_INDEX(draw.vertex_end_gl): { - DrawArrays(); - break; - } - case MAXWELL3D_REG_INDEX(clear_buffers): { - ProcessClearBuffers(); - break; - } - case MAXWELL3D_REG_INDEX(query.query_get): { - ProcessQueryGet(); - break; - } - case MAXWELL3D_REG_INDEX(condition.mode): { - ProcessQueryCondition(); - break; - } - case MAXWELL3D_REG_INDEX(counter_reset): { - ProcessCounterReset(); - break; - } - case MAXWELL3D_REG_INDEX(sync_info): { - ProcessSyncPoint(); - break; - } - case MAXWELL3D_REG_INDEX(exec_upload): { - upload_state.ProcessExec(regs.exec_upload.linear != 0); - break; - } - case MAXWELL3D_REG_INDEX(data_upload): { - upload_state.ProcessData(arg, is_last_call); - if (is_last_call) { - OnMemoryWrite(); - } - break; - } - default: - break; - } + const u32 argument = ProcessShadowRam(method, method_argument); + ProcessDirtyRegisters(method, argument); + ProcessMethodCall(method, argument, method_argument, is_last_call); } void Maxwell3D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, @@ -300,23 +283,7 @@ void Maxwell3D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, // Methods after 0xE00 are special, they're actually triggers for some microcode that was // uploaded to the GPU during initialization. if (method >= MacroRegistersStart) { - // We're trying to execute a macro - if (executing_macro == 0) { - // A macro call must begin by writing the macro method's register, not its argument. - ASSERT_MSG((method % 2) == 0, - "Can't start macro execution by writing to the ARGS register"); - executing_macro = method; - } - - for (std::size_t i = 0; i < amount; i++) { - macro_params.push_back(base_start[i]); - } - - // Call the macro when there are no more parameters in the command buffer - if (amount == methods_pending) { - CallMacroMethod(executing_macro, macro_params); - macro_params.clear(); - } + ProcessMacro(method, base_start, amount, amount == methods_pending); return; } switch (method) { @@ -335,15 +302,14 @@ void Maxwell3D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, case MAXWELL3D_REG_INDEX(const_buffer.cb_data[12]): case MAXWELL3D_REG_INDEX(const_buffer.cb_data[13]): case MAXWELL3D_REG_INDEX(const_buffer.cb_data[14]): - case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): { + case MAXWELL3D_REG_INDEX(const_buffer.cb_data[15]): ProcessCBMultiData(method, base_start, amount); break; - } - default: { + default: for (std::size_t i = 0; i < amount; i++) { CallMethod(method, base_start[i], methods_pending - static_cast<u32>(i) <= 1); } - } + break; } } @@ -396,7 +362,7 @@ void Maxwell3D::CallMethodFromMME(u32 method, u32 method_argument) { } void Maxwell3D::FlushMMEInlineDraw() { - LOG_TRACE(HW_GPU, "called, topology={}, count={}", static_cast<u32>(regs.draw.topology.Value()), + LOG_TRACE(HW_GPU, "called, topology={}, count={}", regs.draw.topology.Value(), regs.vertex_buffer.count); ASSERT_MSG(!(regs.index_array.count && regs.vertex_buffer.count), "Both indexed and direct?"); ASSERT(mme_draw.instance_count == mme_draw.gl_end_count); @@ -541,8 +507,7 @@ void Maxwell3D::ProcessCounterReset() { rasterizer->ResetCounter(QueryType::SamplesPassed); break; default: - LOG_DEBUG(Render_OpenGL, "Unimplemented counter reset={}", - static_cast<int>(regs.counter_reset)); + LOG_DEBUG(Render_OpenGL, "Unimplemented counter reset={}", regs.counter_reset); break; } } @@ -557,7 +522,7 @@ void Maxwell3D::ProcessSyncPoint() { } void Maxwell3D::DrawArrays() { - LOG_TRACE(HW_GPU, "called, topology={}, count={}", static_cast<u32>(regs.draw.topology.Value()), + LOG_TRACE(HW_GPU, "called, topology={}, count={}", regs.draw.topology.Value(), regs.vertex_buffer.count); ASSERT_MSG(!(regs.index_array.count && regs.vertex_buffer.count), "Both indexed and direct?"); @@ -595,12 +560,12 @@ std::optional<u64> Maxwell3D::GetQueryResult() { return 0; case Regs::QuerySelect::SamplesPassed: // Deferred. - rasterizer->Query(regs.query.QueryAddress(), VideoCore::QueryType::SamplesPassed, + rasterizer->Query(regs.query.QueryAddress(), QueryType::SamplesPassed, system.GPU().GetTicks()); return std::nullopt; default: LOG_DEBUG(HW_GPU, "Unimplemented query select type {}", - static_cast<u32>(regs.query.query_get.select.Value())); + regs.query.query_get.select.Value()); return 1; } } @@ -677,7 +642,7 @@ void Maxwell3D::FinishCBData() { } Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const { - const GPUVAddr tic_address_gpu{regs.tic.TICAddress() + tic_index * sizeof(Texture::TICEntry)}; + const GPUVAddr tic_address_gpu{regs.tic.Address() + tic_index * sizeof(Texture::TICEntry)}; Texture::TICEntry tic_entry; memory_manager.ReadBlockUnsafe(tic_address_gpu, &tic_entry, sizeof(Texture::TICEntry)); @@ -686,43 +651,19 @@ Texture::TICEntry Maxwell3D::GetTICEntry(u32 tic_index) const { } Texture::TSCEntry Maxwell3D::GetTSCEntry(u32 tsc_index) const { - const GPUVAddr tsc_address_gpu{regs.tsc.TSCAddress() + tsc_index * sizeof(Texture::TSCEntry)}; + const GPUVAddr tsc_address_gpu{regs.tsc.Address() + tsc_index * sizeof(Texture::TSCEntry)}; Texture::TSCEntry tsc_entry; memory_manager.ReadBlockUnsafe(tsc_address_gpu, &tsc_entry, sizeof(Texture::TSCEntry)); return tsc_entry; } -Texture::FullTextureInfo Maxwell3D::GetTextureInfo(Texture::TextureHandle tex_handle) const { - return Texture::FullTextureInfo{GetTICEntry(tex_handle.tic_id), GetTSCEntry(tex_handle.tsc_id)}; -} - -Texture::FullTextureInfo Maxwell3D::GetStageTexture(ShaderType stage, std::size_t offset) const { - const auto stage_index = static_cast<std::size_t>(stage); - const auto& shader = state.shader_stages[stage_index]; - const auto& tex_info_buffer = shader.const_buffers[regs.tex_cb_index]; - ASSERT(tex_info_buffer.enabled && tex_info_buffer.address != 0); - - const GPUVAddr tex_info_address = - tex_info_buffer.address + offset * sizeof(Texture::TextureHandle); - - ASSERT(tex_info_address < tex_info_buffer.address + tex_info_buffer.size); - - const Texture::TextureHandle tex_handle{memory_manager.Read<u32>(tex_info_address)}; - - return GetTextureInfo(tex_handle); -} - u32 Maxwell3D::GetRegisterValue(u32 method) const { ASSERT_MSG(method < Regs::NUM_REGS, "Invalid Maxwell3D register"); return regs.reg_array[method]; } void Maxwell3D::ProcessClearBuffers() { - ASSERT(regs.clear_buffers.R == regs.clear_buffers.G && - regs.clear_buffers.R == regs.clear_buffers.B && - regs.clear_buffers.R == regs.clear_buffers.A); - rasterizer->Clear(); } @@ -730,9 +671,7 @@ u32 Maxwell3D::AccessConstBuffer32(ShaderType stage, u64 const_buffer, u64 offse ASSERT(stage != ShaderType::Compute); const auto& shader_stage = state.shader_stages[static_cast<std::size_t>(stage)]; const auto& buffer = shader_stage.const_buffers[const_buffer]; - u32 result; - std::memcpy(&result, memory_manager.GetPointer(buffer.address + offset), sizeof(u32)); - return result; + return memory_manager.Read<u32>(buffer.address + offset); } SamplerDescriptor Maxwell3D::AccessBoundSampler(ShaderType stage, u64 offset) const { @@ -750,9 +689,11 @@ SamplerDescriptor Maxwell3D::AccessBindlessSampler(ShaderType stage, u64 const_b SamplerDescriptor Maxwell3D::AccessSampler(u32 handle) const { const Texture::TextureHandle tex_handle{handle}; - const Texture::FullTextureInfo tex_info = GetTextureInfo(tex_handle); - SamplerDescriptor result = SamplerDescriptor::FromTIC(tex_info.tic); - result.is_shadow.Assign(tex_info.tsc.depth_compare_enabled.Value()); + const Texture::TICEntry tic = GetTICEntry(tex_handle.tic_id); + const Texture::TSCEntry tsc = GetTSCEntry(tex_handle.tsc_id); + + SamplerDescriptor result = SamplerDescriptor::FromTIC(tic); + result.is_shadow.Assign(tsc.depth_compare_enabled.Value()); return result; } diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index bc289c55d..bf9e07c9b 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -438,16 +438,6 @@ public: DecrWrapOGL = 0x8508, }; - enum class MemoryLayout : u32 { - Linear = 0, - BlockLinear = 1, - }; - - enum class InvMemoryLayout : u32 { - BlockLinear = 0, - Linear = 1, - }; - enum class CounterReset : u32 { SampleCnt = 0x01, Unk02 = 0x02, @@ -589,21 +579,31 @@ public: NegativeW = 7, }; + enum class SamplerIndex : u32 { + Independently = 0, + ViaHeaderIndex = 1, + }; + + struct TileMode { + union { + BitField<0, 4, u32> block_width; + BitField<4, 4, u32> block_height; + BitField<8, 4, u32> block_depth; + BitField<12, 1, u32> is_pitch_linear; + BitField<16, 1, u32> is_3d; + }; + }; + static_assert(sizeof(TileMode) == 4); + struct RenderTargetConfig { u32 address_high; u32 address_low; u32 width; u32 height; Tegra::RenderTargetFormat format; + TileMode tile_mode; union { - BitField<0, 3, u32> block_width; - BitField<4, 3, u32> block_height; - BitField<8, 3, u32> block_depth; - BitField<12, 1, InvMemoryLayout> type; - BitField<16, 1, u32> is_3d; - } memory_layout; - union { - BitField<0, 16, u32> layers; + BitField<0, 16, u32> depth; BitField<16, 1, u32> volume; }; u32 layer_stride; @@ -755,7 +755,11 @@ public: u32 data_upload; - INSERT_UNION_PADDING_WORDS(0x44); + INSERT_UNION_PADDING_WORDS(0x16); + + u32 force_early_fragment_tests; + + INSERT_UNION_PADDING_WORDS(0x2D); struct { union { @@ -828,7 +832,11 @@ public: u32 patch_vertices; - INSERT_UNION_PADDING_WORDS(0xC); + INSERT_UNION_PADDING_WORDS(0x4); + + u32 fragment_barrier; + + INSERT_UNION_PADDING_WORDS(0x7); std::array<ScissorTest, NumViewports> scissor_test; @@ -838,7 +846,15 @@ public: u32 stencil_back_mask; u32 stencil_back_func_mask; - INSERT_UNION_PADDING_WORDS(0xC); + INSERT_UNION_PADDING_WORDS(0x5); + + u32 invalidate_texture_data_cache; + + INSERT_UNION_PADDING_WORDS(0x1); + + u32 tiled_cache_barrier; + + INSERT_UNION_PADDING_WORDS(0x4); u32 color_mask_common; @@ -862,12 +878,7 @@ public: u32 address_high; u32 address_low; Tegra::DepthFormat format; - union { - BitField<0, 4, u32> block_width; - BitField<4, 4, u32> block_height; - BitField<8, 4, u32> block_depth; - BitField<20, 1, InvMemoryLayout> type; - } memory_layout; + TileMode tile_mode; u32 layer_stride; GPUVAddr Address() const { @@ -876,7 +887,18 @@ public: } } zeta; - INSERT_UNION_PADDING_WORDS(0x41); + struct { + union { + BitField<0, 16, u32> x; + BitField<16, 16, u32> width; + }; + union { + BitField<0, 16, u32> y; + BitField<16, 16, u32> height; + }; + } render_area; + + INSERT_UNION_PADDING_WORDS(0x3F); union { BitField<0, 4, u32> stencil; @@ -917,7 +939,7 @@ public: BitField<25, 3, u32> map_7; }; - u32 GetMap(std::size_t index) const { + u32 Map(std::size_t index) const { const std::array<u32, NumRenderTargets> maps{map_0, map_1, map_2, map_3, map_4, map_5, map_6, map_7}; ASSERT(index < maps.size()); @@ -930,11 +952,13 @@ public: u32 zeta_width; u32 zeta_height; union { - BitField<0, 16, u32> zeta_layers; + BitField<0, 16, u32> zeta_depth; BitField<16, 1, u32> zeta_volume; }; - INSERT_UNION_PADDING_WORDS(0x26); + SamplerIndex sampler_index; + + INSERT_UNION_PADDING_WORDS(0x25); u32 depth_test_enable; @@ -960,6 +984,7 @@ public: float b; float a; } blend_color; + INSERT_UNION_PADDING_WORDS(0x4); struct { @@ -997,7 +1022,12 @@ public: float line_width_smooth; float line_width_aliased; - INSERT_UNION_PADDING_WORDS(0x1F); + INSERT_UNION_PADDING_WORDS(0x1B); + + u32 invalidate_sampler_cache_no_wfi; + u32 invalidate_texture_header_cache_no_wfi; + + INSERT_UNION_PADDING_WORDS(0x2); u32 vb_element_base; u32 vb_base_instance; @@ -1041,13 +1071,13 @@ public: } condition; struct { - u32 tsc_address_high; - u32 tsc_address_low; - u32 tsc_limit; + u32 address_high; + u32 address_low; + u32 limit; - GPUVAddr TSCAddress() const { - return static_cast<GPUVAddr>( - (static_cast<GPUVAddr>(tsc_address_high) << 32) | tsc_address_low); + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); } } tsc; @@ -1058,13 +1088,13 @@ public: u32 line_smooth_enable; struct { - u32 tic_address_high; - u32 tic_address_low; - u32 tic_limit; + u32 address_high; + u32 address_low; + u32 limit; - GPUVAddr TICAddress() const { - return static_cast<GPUVAddr>( - (static_cast<GPUVAddr>(tic_address_high) << 32) | tic_address_low); + GPUVAddr Address() const { + return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | + address_low); } } tic; @@ -1393,12 +1423,6 @@ public: void FlushMMEInlineDraw(); - /// Given a texture handle, returns the TSC and TIC entries. - Texture::FullTextureInfo GetTextureInfo(Texture::TextureHandle tex_handle) const; - - /// Returns the texture information for a specific texture in a specific shader stage. - Texture::FullTextureInfo GetStageTexture(ShaderType stage, std::size_t offset) const; - u32 AccessConstBuffer32(ShaderType stage, u64 const_buffer, u64 offset) const override; SamplerDescriptor AccessBoundSampler(ShaderType stage, u64 offset) const override; @@ -1461,38 +1485,13 @@ public: private: void InitializeRegisterDefaults(); - Core::System& system; - MemoryManager& memory_manager; - - VideoCore::RasterizerInterface* rasterizer = nullptr; - - /// Start offsets of each macro in macro_memory - std::array<u32, 0x80> macro_positions = {}; - - std::array<bool, Regs::NUM_REGS> mme_inline{}; - - /// Macro method that is currently being executed / being fed parameters. - u32 executing_macro = 0; - /// Parameters that have been submitted to the macro call so far. - std::vector<u32> macro_params; - - /// Interpreter for the macro codes uploaded to the GPU. - std::unique_ptr<MacroEngine> macro_engine; - - static constexpr u32 null_cb_data = 0xFFFFFFFF; - struct { - std::array<std::array<u32, 0x4000>, 16> buffer; - u32 current{null_cb_data}; - u32 id{null_cb_data}; - u32 start_pos{}; - u32 counter{}; - } cb_data_state; + void ProcessMacro(u32 method, const u32* base_start, u32 amount, bool is_last_call); - Upload::State upload_state; + u32 ProcessShadowRam(u32 method, u32 argument); - bool execute_on{true}; + void ProcessDirtyRegisters(u32 method, u32 argument); - std::array<u8, Regs::NUM_REGS> dirty_pointers{}; + void ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argument, bool is_last_call); /// Retrieves information about a specific TIC entry from the TIC buffer. Texture::TICEntry GetTICEntry(u32 tic_index) const; @@ -1502,8 +1501,8 @@ private: /** * Call a macro on this engine. + * * @param method Method to call - * @param num_parameters Number of arguments * @param parameters Arguments to the method call */ void CallMacroMethod(u32 method, const std::vector<u32>& parameters); @@ -1552,6 +1551,38 @@ private: /// Returns a query's value or an empty object if the value will be deferred through a cache. std::optional<u64> GetQueryResult(); + + Core::System& system; + MemoryManager& memory_manager; + + VideoCore::RasterizerInterface* rasterizer = nullptr; + + /// Start offsets of each macro in macro_memory + std::array<u32, 0x80> macro_positions{}; + + std::array<bool, Regs::NUM_REGS> mme_inline{}; + + /// Macro method that is currently being executed / being fed parameters. + u32 executing_macro = 0; + /// Parameters that have been submitted to the macro call so far. + std::vector<u32> macro_params; + + /// Interpreter for the macro codes uploaded to the GPU. + std::unique_ptr<MacroEngine> macro_engine; + + static constexpr u32 null_cb_data = 0xFFFFFFFF; + struct CBDataState { + std::array<std::array<u32, 0x4000>, 16> buffer; + u32 current{null_cb_data}; + u32 id{null_cb_data}; + u32 start_pos{}; + u32 counter{}; + }; + CBDataState cb_data_state; + + Upload::State upload_state; + + bool execute_on{true}; }; #define ASSERT_REG_POSITION(field_name, position) \ @@ -1564,6 +1595,7 @@ ASSERT_REG_POSITION(shadow_ram_control, 0x49); ASSERT_REG_POSITION(upload, 0x60); ASSERT_REG_POSITION(exec_upload, 0x6C); ASSERT_REG_POSITION(data_upload, 0x6D); +ASSERT_REG_POSITION(force_early_fragment_tests, 0x84); ASSERT_REG_POSITION(sync_info, 0xB2); ASSERT_REG_POSITION(tess_mode, 0xC8); ASSERT_REG_POSITION(tess_level_outer, 0xC9); @@ -1586,10 +1618,13 @@ ASSERT_REG_POSITION(polygon_offset_point_enable, 0x370); ASSERT_REG_POSITION(polygon_offset_line_enable, 0x371); ASSERT_REG_POSITION(polygon_offset_fill_enable, 0x372); ASSERT_REG_POSITION(patch_vertices, 0x373); +ASSERT_REG_POSITION(fragment_barrier, 0x378); ASSERT_REG_POSITION(scissor_test, 0x380); ASSERT_REG_POSITION(stencil_back_func_ref, 0x3D5); ASSERT_REG_POSITION(stencil_back_mask, 0x3D6); ASSERT_REG_POSITION(stencil_back_func_mask, 0x3D7); +ASSERT_REG_POSITION(invalidate_texture_data_cache, 0x3DD); +ASSERT_REG_POSITION(tiled_cache_barrier, 0x3DF); ASSERT_REG_POSITION(color_mask_common, 0x3E4); ASSERT_REG_POSITION(depth_bounds, 0x3E7); ASSERT_REG_POSITION(rt_separate_frag_data, 0x3EB); @@ -1597,6 +1632,7 @@ ASSERT_REG_POSITION(multisample_raster_enable, 0x3ED); ASSERT_REG_POSITION(multisample_raster_samples, 0x3EE); ASSERT_REG_POSITION(multisample_sample_mask, 0x3EF); ASSERT_REG_POSITION(zeta, 0x3F8); +ASSERT_REG_POSITION(render_area, 0x3FD); ASSERT_REG_POSITION(clear_flags, 0x43E); ASSERT_REG_POSITION(fill_rectangle, 0x44F); ASSERT_REG_POSITION(vertex_attrib_format, 0x458); @@ -1605,7 +1641,8 @@ ASSERT_REG_POSITION(multisample_coverage_to_color, 0x47E); ASSERT_REG_POSITION(rt_control, 0x487); ASSERT_REG_POSITION(zeta_width, 0x48a); ASSERT_REG_POSITION(zeta_height, 0x48b); -ASSERT_REG_POSITION(zeta_layers, 0x48c); +ASSERT_REG_POSITION(zeta_depth, 0x48c); +ASSERT_REG_POSITION(sampler_index, 0x48D); ASSERT_REG_POSITION(depth_test_enable, 0x4B3); ASSERT_REG_POSITION(independent_blend_enable, 0x4B9); ASSERT_REG_POSITION(depth_write_enabled, 0x4BA); @@ -1629,6 +1666,8 @@ ASSERT_REG_POSITION(frag_color_clamp, 0x4EA); ASSERT_REG_POSITION(screen_y_control, 0x4EB); ASSERT_REG_POSITION(line_width_smooth, 0x4EC); ASSERT_REG_POSITION(line_width_aliased, 0x4ED); +ASSERT_REG_POSITION(invalidate_sampler_cache_no_wfi, 0x509); +ASSERT_REG_POSITION(invalidate_texture_header_cache_no_wfi, 0x50A); ASSERT_REG_POSITION(vb_element_base, 0x50D); ASSERT_REG_POSITION(vb_base_instance, 0x50E); ASSERT_REG_POSITION(clip_distance_enabled, 0x544); diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp index e88290754..ba750748c 100644 --- a/src/video_core/engines/maxwell_dma.cpp +++ b/src/video_core/engines/maxwell_dma.cpp @@ -16,8 +16,10 @@ namespace Tegra::Engines { using namespace Texture; -MaxwellDMA::MaxwellDMA(Core::System& system, MemoryManager& memory_manager) - : system{system}, memory_manager{memory_manager} {} +MaxwellDMA::MaxwellDMA(Core::System& system_, MemoryManager& memory_manager_) + : system{system_}, memory_manager{memory_manager_} {} + +MaxwellDMA::~MaxwellDMA() = default; void MaxwellDMA::CallMethod(u32 method, u32 method_argument, bool is_last_call) { ASSERT_MSG(method < NUM_REGS, "Invalid MaxwellDMA register"); @@ -94,6 +96,7 @@ void MaxwellDMA::CopyPitchToPitch() { } void MaxwellDMA::CopyBlockLinearToPitch() { + UNIMPLEMENTED_IF(regs.src_params.block_size.width != 0); UNIMPLEMENTED_IF(regs.src_params.block_size.depth != 0); UNIMPLEMENTED_IF(regs.src_params.layer != 0); @@ -114,8 +117,6 @@ void MaxwellDMA::CopyBlockLinearToPitch() { const u32 block_depth = src_params.block_size.depth; const size_t src_size = CalculateSize(true, bytes_per_pixel, width, height, depth, block_height, block_depth); - const size_t src_layer_size = - CalculateSize(true, bytes_per_pixel, width, height, 1, block_height, block_depth); if (read_buffer.size() < src_size) { read_buffer.resize(src_size); @@ -135,6 +136,8 @@ void MaxwellDMA::CopyBlockLinearToPitch() { } void MaxwellDMA::CopyPitchToBlockLinear() { + UNIMPLEMENTED_IF_MSG(regs.dst_params.block_size.width != 0, "Block width is not one"); + const auto& dst_params = regs.dst_params; const u32 bytes_per_pixel = regs.pitch_in / regs.line_length_in; const u32 width = dst_params.width; diff --git a/src/video_core/engines/maxwell_dma.h b/src/video_core/engines/maxwell_dma.h index 50f445efc..3c59eeb13 100644 --- a/src/video_core/engines/maxwell_dma.h +++ b/src/video_core/engines/maxwell_dma.h @@ -72,11 +72,13 @@ public: struct RenderEnable { enum class Mode : u32 { - FALSE = 0, - TRUE = 1, - CONDITIONAL = 2, - RENDER_IF_EQUAL = 3, - RENDER_IF_NOT_EQUAL = 4, + // Note: This uses Pascal case in order to avoid the identifiers + // FALSE and TRUE, which are reserved on Darwin. + False = 0, + True = 1, + Conditional = 2, + RenderIfEqual = 3, + RenderIfNotEqual = 4, }; PackedGPUVAddr address; @@ -185,8 +187,8 @@ public: }; static_assert(sizeof(RemapConst) == 12); - explicit MaxwellDMA(Core::System& system, MemoryManager& memory_manager); - ~MaxwellDMA() = default; + explicit MaxwellDMA(Core::System& system_, MemoryManager& memory_manager_); + ~MaxwellDMA(); /// Write the value to the register identified by method. void CallMethod(u32 method, u32 method_argument, bool is_last_call) override; diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index d374b73cf..8b45f1b62 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -32,31 +32,31 @@ struct Register { constexpr Register() = default; - constexpr Register(u64 value) : value(value) {} + constexpr Register(u64 value_) : value(value_) {} - constexpr operator u64() const { + [[nodiscard]] constexpr operator u64() const { return value; } template <typename T> - constexpr u64 operator-(const T& oth) const { + [[nodiscard]] constexpr u64 operator-(const T& oth) const { return value - oth; } template <typename T> - constexpr u64 operator&(const T& oth) const { + [[nodiscard]] constexpr u64 operator&(const T& oth) const { return value & oth; } - constexpr u64 operator&(const Register& oth) const { + [[nodiscard]] constexpr u64 operator&(const Register& oth) const { return value & oth.value; } - constexpr u64 operator~() const { + [[nodiscard]] constexpr u64 operator~() const { return ~value; } - u64 GetSwizzledIndex(u64 elem) const { + [[nodiscard]] u64 GetSwizzledIndex(u64 elem) const { elem = (value + elem) & 3; return (value & ~3) + elem; } @@ -75,7 +75,7 @@ enum class AttributeSize : u64 { union Attribute { Attribute() = default; - constexpr explicit Attribute(u64 value) : value(value) {} + constexpr explicit Attribute(u64 value_) : value(value_) {} enum class Index : u64 { LayerViewportPointSize = 6, @@ -107,7 +107,7 @@ union Attribute { BitField<31, 1, u64> patch; BitField<47, 3, AttributeSize> size; - bool IsPhysical() const { + [[nodiscard]] bool IsPhysical() const { return patch == 0 && element == 0 && static_cast<u64>(index.Value()) == 0; } } fmt20; @@ -124,7 +124,7 @@ union Attribute { union Sampler { Sampler() = default; - constexpr explicit Sampler(u64 value) : value(value) {} + constexpr explicit Sampler(u64 value_) : value(value_) {} enum class Index : u64 { Sampler_0 = 8, @@ -137,7 +137,7 @@ union Sampler { union Image { Image() = default; - constexpr explicit Image(u64 value) : value{value} {} + constexpr explicit Image(u64 value_) : value{value_} {} BitField<36, 13, u64> index; u64 value; @@ -505,14 +505,14 @@ struct IpaMode { IpaInterpMode interpolation_mode; IpaSampleMode sampling_mode; - bool operator==(const IpaMode& a) const { + [[nodiscard]] bool operator==(const IpaMode& a) const { return std::tie(interpolation_mode, sampling_mode) == std::tie(a.interpolation_mode, a.sampling_mode); } - bool operator!=(const IpaMode& a) const { + [[nodiscard]] bool operator!=(const IpaMode& a) const { return !operator==(a); } - bool operator<(const IpaMode& a) const { + [[nodiscard]] bool operator<(const IpaMode& a) const { return std::tie(interpolation_mode, sampling_mode) < std::tie(a.interpolation_mode, a.sampling_mode); } @@ -658,10 +658,10 @@ union Instruction { return *this; } - constexpr Instruction(u64 value) : value{value} {} + constexpr Instruction(u64 value_) : value{value_} {} constexpr Instruction(const Instruction& instr) : value(instr.value) {} - constexpr bool Bit(u64 offset) const { + [[nodiscard]] constexpr bool Bit(u64 offset) const { return ((value >> offset) & 1) != 0; } @@ -746,34 +746,34 @@ union Instruction { BitField<28, 8, u64> imm_lut28; BitField<48, 8, u64> imm_lut48; - u32 GetImmLut28() const { + [[nodiscard]] u32 GetImmLut28() const { return static_cast<u32>(imm_lut28); } - u32 GetImmLut48() const { + [[nodiscard]] u32 GetImmLut48() const { return static_cast<u32>(imm_lut48); } } lop3; - u16 GetImm20_16() const { + [[nodiscard]] u16 GetImm20_16() const { return static_cast<u16>(imm20_16); } - u32 GetImm20_19() const { + [[nodiscard]] u32 GetImm20_19() const { u32 imm{static_cast<u32>(imm20_19)}; imm <<= 12; imm |= negate_imm ? 0x80000000 : 0; return imm; } - u32 GetImm20_32() const { + [[nodiscard]] u32 GetImm20_32() const { return static_cast<u32>(imm20_32); } - s32 GetSignedImm20_20() const { - u32 immediate = static_cast<u32>(imm20_19 | (negate_imm << 19)); + [[nodiscard]] s32 GetSignedImm20_20() const { + const auto immediate = static_cast<u32>(imm20_19 | (negate_imm << 19)); // Sign extend the 20-bit value. - u32 mask = 1U << (20 - 1); + const auto mask = 1U << (20 - 1); return static_cast<s32>((immediate ^ mask) - mask); } } alu; @@ -857,7 +857,7 @@ union Instruction { BitField<56, 1, u64> second_negate; BitField<30, 9, u64> second; - u32 PackImmediates() const { + [[nodiscard]] u32 PackImmediates() const { // Immediates are half floats shifted. constexpr u32 imm_shift = 6; return static_cast<u32>((first << imm_shift) | (second << (16 + imm_shift))); @@ -1033,7 +1033,7 @@ union Instruction { BitField<28, 2, AtomicType> type; BitField<30, 22, s64> offset; - s32 GetImmediateOffset() const { + [[nodiscard]] s32 GetImmediateOffset() const { return static_cast<s32>(offset << 2); } } atoms; @@ -1215,7 +1215,7 @@ union Instruction { BitField<39, 4, u64> rounding; // H0, H1 extract for F16 missing BitField<41, 1, u64> selector; // Guessed as some games set it, TODO: reverse this value - F2fRoundingOp GetRoundingMode() const { + [[nodiscard]] F2fRoundingOp GetRoundingMode() const { constexpr u64 rounding_mask = 0x0B; return static_cast<F2fRoundingOp>(rounding.Value() & rounding_mask); } @@ -1239,15 +1239,15 @@ union Instruction { BitField<54, 1, u64> aoffi_flag; BitField<55, 3, TextureProcessMode> process_mode; - bool IsComponentEnabled(std::size_t component) const { - return ((1ull << component) & component_mask) != 0; + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { + return ((1ULL << component) & component_mask) != 0; } - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { return process_mode; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return dc_flag != 0; @@ -1271,15 +1271,15 @@ union Instruction { BitField<36, 1, u64> aoffi_flag; BitField<37, 3, TextureProcessMode> process_mode; - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { return ((1ULL << component) & component_mask) != 0; } - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { return process_mode; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return dc_flag != 0; @@ -1299,7 +1299,7 @@ union Instruction { BitField<31, 4, u64> component_mask; BitField<49, 1, u64> nodep_flag; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NODEP: return nodep_flag != 0; @@ -1309,7 +1309,7 @@ union Instruction { return false; } - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { return ((1ULL << component) & component_mask) != 0; } } txq; @@ -1321,11 +1321,11 @@ union Instruction { BitField<35, 1, u64> ndv_flag; BitField<49, 1, u64> nodep_flag; - bool IsComponentEnabled(std::size_t component) const { - return ((1ull << component) & component_mask) != 0; + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { + return ((1ULL << component) & component_mask) != 0; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NDV: return (ndv_flag != 0); @@ -1347,7 +1347,7 @@ union Instruction { BitField<54, 2, u64> offset_mode; BitField<56, 2, u64> component; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NDV: return ndv_flag != 0; @@ -1373,7 +1373,7 @@ union Instruction { BitField<33, 2, u64> offset_mode; BitField<37, 2, u64> component; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::NDV: return ndv_flag != 0; @@ -1399,7 +1399,7 @@ union Instruction { BitField<52, 2, u64> component; BitField<55, 1, u64> fp16_flag; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return dc_flag != 0; @@ -1422,24 +1422,27 @@ union Instruction { BitField<53, 4, u64> texture_info; BitField<59, 1, u64> fp32_flag; - TextureType GetTextureType() const { + [[nodiscard]] TextureType GetTextureType() const { // The TEXS instruction has a weird encoding for the texture type. - if (texture_info == 0) + if (texture_info == 0) { return TextureType::Texture1D; - if (texture_info >= 1 && texture_info <= 9) + } + if (texture_info >= 1 && texture_info <= 9) { return TextureType::Texture2D; - if (texture_info >= 10 && texture_info <= 11) + } + if (texture_info >= 10 && texture_info <= 11) { return TextureType::Texture3D; - if (texture_info >= 12 && texture_info <= 13) + } + if (texture_info >= 12 && texture_info <= 13) { return TextureType::TextureCube; + } - LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}", - static_cast<u32>(texture_info.Value())); + LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}", texture_info.Value()); UNREACHABLE(); return TextureType::Texture1D; } - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { switch (texture_info) { case 0: case 2: @@ -1458,7 +1461,7 @@ union Instruction { return TextureProcessMode::None; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::DC: return (texture_info >= 4 && texture_info <= 6) || texture_info == 9; @@ -1470,16 +1473,16 @@ union Instruction { return false; } - bool IsArrayTexture() const { + [[nodiscard]] bool IsArrayTexture() const { // TEXS only supports Texture2D arrays. return texture_info >= 7 && texture_info <= 9; } - bool HasTwoDestinations() const { + [[nodiscard]] bool HasTwoDestinations() const { return gpr28.Value() != Register::ZeroIndex; } - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { static constexpr std::array<std::array<u32, 8>, 4> mask_lut{{ {}, {0x1, 0x2, 0x4, 0x8, 0x3, 0x9, 0xa, 0xc}, @@ -1506,7 +1509,7 @@ union Instruction { BitField<54, 1, u64> cl; BitField<55, 1, u64> process_mode; - TextureProcessMode GetTextureProcessMode() const { + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { return process_mode == 0 ? TextureProcessMode::LZ : TextureProcessMode::LL; } } tld; @@ -1516,7 +1519,7 @@ union Instruction { BitField<53, 4, u64> texture_info; BitField<59, 1, u64> fp32_flag; - TextureType GetTextureType() const { + [[nodiscard]] TextureType GetTextureType() const { // The TLDS instruction has a weird encoding for the texture type. if (texture_info <= 1) { return TextureType::Texture1D; @@ -1529,19 +1532,19 @@ union Instruction { return TextureType::Texture3D; } - LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}", - static_cast<u32>(texture_info.Value())); + LOG_CRITICAL(HW_GPU, "Unhandled texture_info: {}", texture_info.Value()); UNREACHABLE(); return TextureType::Texture1D; } - TextureProcessMode GetTextureProcessMode() const { - if (texture_info == 1 || texture_info == 5 || texture_info == 12) + [[nodiscard]] TextureProcessMode GetTextureProcessMode() const { + if (texture_info == 1 || texture_info == 5 || texture_info == 12) { return TextureProcessMode::LL; + } return TextureProcessMode::LZ; } - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::AOFFI: return texture_info == 12 || texture_info == 4; @@ -1555,7 +1558,7 @@ union Instruction { return false; } - bool IsArrayTexture() const { + [[nodiscard]] bool IsArrayTexture() const { // TEXS only supports Texture2D arrays. return texture_info == 8; } @@ -1567,7 +1570,7 @@ union Instruction { BitField<35, 1, u64> aoffi_flag; BitField<49, 1, u64> nodep_flag; - bool UsesMiscMode(TextureMiscMode mode) const { + [[nodiscard]] bool UsesMiscMode(TextureMiscMode mode) const { switch (mode) { case TextureMiscMode::AOFFI: return aoffi_flag != 0; @@ -1591,7 +1594,7 @@ union Instruction { BitField<20, 3, StoreType> store_data_layout; BitField<20, 4, u64> component_mask_selector; - bool IsComponentEnabled(std::size_t component) const { + [[nodiscard]] bool IsComponentEnabled(std::size_t component) const { ASSERT(mode == SurfaceDataMode::P); constexpr u8 R = 0b0001; constexpr u8 G = 0b0010; @@ -1604,7 +1607,7 @@ union Instruction { return std::bitset<4>{mask.at(component_mask_selector)}.test(component); } - StoreType GetStoreDataLayout() const { + [[nodiscard]] StoreType GetStoreDataLayout() const { ASSERT(mode == SurfaceDataMode::D_BA); return store_data_layout; } @@ -1622,14 +1625,15 @@ union Instruction { BitField<20, 24, u64> target; BitField<5, 1, u64> constant_buffer; - s32 GetBranchTarget() const { + [[nodiscard]] s32 GetBranchTarget() const { // Sign extend the branch target offset - u32 mask = 1U << (24 - 1); - u32 value = static_cast<u32>(target); + const auto mask = 1U << (24 - 1); + const auto target_value = static_cast<u32>(target); + constexpr auto instruction_size = static_cast<s32>(sizeof(Instruction)); + // The branch offset is relative to the next instruction and is stored in bytes, so // divide it by the size of an instruction and add 1 to it. - return static_cast<s32>((value ^ mask) - mask) / static_cast<s32>(sizeof(Instruction)) + - 1; + return static_cast<s32>((target_value ^ mask) - mask) / instruction_size + 1; } } bra; @@ -1637,14 +1641,15 @@ union Instruction { BitField<20, 24, u64> target; BitField<5, 1, u64> constant_buffer; - s32 GetBranchExtend() const { + [[nodiscard]] s32 GetBranchExtend() const { // Sign extend the branch target offset - u32 mask = 1U << (24 - 1); - u32 value = static_cast<u32>(target); + const auto mask = 1U << (24 - 1); + const auto target_value = static_cast<u32>(target); + constexpr auto instruction_size = static_cast<s32>(sizeof(Instruction)); + // The branch offset is relative to the next instruction and is stored in bytes, so // divide it by the size of an instruction and add 1 to it. - return static_cast<s32>((value ^ mask) - mask) / static_cast<s32>(sizeof(Instruction)) + - 1; + return static_cast<s32>((target_value ^ mask) - mask) / instruction_size + 1; } } brx; @@ -1697,7 +1702,7 @@ union Instruction { BitField<50, 1, u64> is_op_b_register; BitField<51, 3, VmnmxOperation> operation; - VmnmxType SourceFormatA() const { + [[nodiscard]] VmnmxType SourceFormatA() const { switch (src_format_a) { case 0b11: return VmnmxType::Bits32; @@ -1708,7 +1713,7 @@ union Instruction { } } - VmnmxType SourceFormatB() const { + [[nodiscard]] VmnmxType SourceFormatB() const { switch (src_format_b) { case 0b11: return VmnmxType::Bits32; @@ -1739,7 +1744,7 @@ union Instruction { BitField<20, 14, u64> shifted_offset; BitField<34, 5, u64> index; - u64 GetOffset() const { + [[nodiscard]] u64 GetOffset() const { return shifted_offset * 4; } } cbuf34; @@ -1748,7 +1753,7 @@ union Instruction { BitField<20, 16, s64> offset; BitField<36, 5, u64> index; - s64 GetOffset() const { + [[nodiscard]] s64 GetOffset() const { return offset; } } cbuf36; @@ -1893,6 +1898,7 @@ public: ICMP_IMM, FCMP_RR, FCMP_RC, + FCMP_IMMR, MUFU, // Multi-Function Operator RRO_C, // Range Reduction Operator RRO_R, @@ -1996,29 +2002,29 @@ public: /// Returns whether an opcode has an execution predicate field or not (ie, whether it can be /// conditionally executed). - static bool IsPredicatedInstruction(Id opcode) { + [[nodiscard]] static bool IsPredicatedInstruction(Id opcode) { // TODO(Subv): Add the rest of unpredicated instructions. return opcode != Id::SSY && opcode != Id::PBK; } class Matcher { public: - constexpr Matcher(const char* const name, u16 mask, u16 expected, Id id, Type type) - : name{name}, mask{mask}, expected{expected}, id{id}, type{type} {} + constexpr Matcher(const char* const name_, u16 mask_, u16 expected_, Id id_, Type type_) + : name{name_}, mask{mask_}, expected{expected_}, id{id_}, type{type_} {} - constexpr const char* GetName() const { + [[nodiscard]] constexpr const char* GetName() const { return name; } - constexpr u16 GetMask() const { + [[nodiscard]] constexpr u16 GetMask() const { return mask; } - constexpr Id GetId() const { + [[nodiscard]] constexpr Id GetId() const { return id; } - constexpr Type GetType() const { + [[nodiscard]] constexpr Type GetType() const { return type; } @@ -2027,7 +2033,7 @@ public: * @param instruction The instruction to test * @returns true if the given instruction matches. */ - constexpr bool Matches(u16 instruction) const { + [[nodiscard]] constexpr bool Matches(u16 instruction) const { return (instruction & mask) == expected; } @@ -2039,7 +2045,8 @@ public: Type type; }; - static std::optional<std::reference_wrapper<const Matcher>> Decode(Instruction instr) { + using DecodeResult = std::optional<std::reference_wrapper<const Matcher>>; + [[nodiscard]] static DecodeResult Decode(Instruction instr) { static const auto table{GetDecodeTable()}; const auto matches_instruction = [instr](const auto& matcher) { @@ -2061,7 +2068,7 @@ private: * A '0' in a bitstring indicates that a zero must be present at that bit position. * A '1' in a bitstring indicates that a one must be present at that bit position. */ - static constexpr auto GetMaskAndExpect(const char* const bitstring) { + [[nodiscard]] static constexpr auto GetMaskAndExpect(const char* const bitstring) { u16 mask = 0, expect = 0; for (std::size_t i = 0; i < opcode_bitsize; i++) { const std::size_t bit_position = opcode_bitsize - i - 1; @@ -2083,14 +2090,14 @@ private: public: /// Creates a matcher that can match and parse instructions based on bitstring. - static constexpr auto GetMatcher(const char* const bitstring, Id op, Type type, - const char* const name) { + [[nodiscard]] static constexpr auto GetMatcher(const char* const bitstring, Id op, + Type type, const char* const name) { const auto [mask, expected] = GetMaskAndExpect(bitstring); return Matcher(name, mask, expected, op, type); } }; - static std::vector<Matcher> GetDecodeTable() { + [[nodiscard]] static std::vector<Matcher> GetDecodeTable() { std::vector<Matcher> table = { #define INST(bitstring, op, type, name) Detail::GetMatcher(bitstring, op, type, name) INST("111000110011----", Id::KIL, Type::Flow, "KIL"), @@ -2205,6 +2212,7 @@ private: INST("0111110-0-------", Id::HSET2_IMM, Type::HalfSet, "HSET2_IMM"), INST("010110111010----", Id::FCMP_RR, Type::Arithmetic, "FCMP_RR"), INST("010010111010----", Id::FCMP_RC, Type::Arithmetic, "FCMP_RC"), + INST("0011011-1010----", Id::FCMP_IMMR, Type::Arithmetic, "FCMP_IMMR"), INST("0101000010000---", Id::MUFU, Type::Arithmetic, "MUFU"), INST("0100110010010---", Id::RRO_C, Type::Arithmetic, "RRO_C"), INST("0101110010010---", Id::RRO_R, Type::Arithmetic, "RRO_R"), diff --git a/src/video_core/engines/shader_header.h b/src/video_core/engines/shader_header.h index 72e2a33d5..ceec05459 100644 --- a/src/video_core/engines/shader_header.h +++ b/src/video_core/engines/shader_header.h @@ -41,30 +41,30 @@ struct Header { BitField<26, 1, u32> does_load_or_store; BitField<27, 1, u32> does_fp64; BitField<28, 4, u32> stream_out_mask; - } common0{}; + } common0; union { BitField<0, 24, u32> shader_local_memory_low_size; BitField<24, 8, u32> per_patch_attribute_count; - } common1{}; + } common1; union { BitField<0, 24, u32> shader_local_memory_high_size; BitField<24, 8, u32> threads_per_input_primitive; - } common2{}; + } common2; union { BitField<0, 24, u32> shader_local_memory_crs_size; BitField<24, 4, OutputTopology> output_topology; BitField<28, 4, u32> reserved; - } common3{}; + } common3; union { BitField<0, 12, u32> max_output_vertices; BitField<12, 8, u32> store_req_start; // NOTE: not used by geometry shaders. BitField<20, 4, u32> reserved; BitField<24, 8, u32> store_req_end; // NOTE: not used by geometry shaders. - } common4{}; + } common4; union { struct { @@ -145,7 +145,7 @@ struct Header { } } ps; - std::array<u32, 0xF> raw{}; + std::array<u32, 0xF> raw; }; u64 GetLocalMemorySize() const { @@ -153,7 +153,6 @@ struct Header { (common2.shader_local_memory_high_size << 24)); } }; - static_assert(sizeof(Header) == 0x50, "Incorrect structure size"); } // namespace Tegra::Shader diff --git a/src/video_core/fence_manager.h b/src/video_core/fence_manager.h index de6991ef6..3512283ff 100644 --- a/src/video_core/fence_manager.h +++ b/src/video_core/fence_manager.h @@ -9,6 +9,7 @@ #include "common/common_types.h" #include "core/core.h" +#include "video_core/delayed_destruction_ring.h" #include "video_core/gpu.h" #include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" @@ -17,11 +18,11 @@ namespace VideoCommon { class FenceBase { public: - FenceBase(u32 payload, bool is_stubbed) - : address{}, payload{payload}, is_semaphore{false}, is_stubbed{is_stubbed} {} + explicit FenceBase(u32 payload_, bool is_stubbed_) + : address{}, payload{payload_}, is_semaphore{false}, is_stubbed{is_stubbed_} {} - FenceBase(GPUVAddr address, u32 payload, bool is_stubbed) - : address{address}, payload{payload}, is_semaphore{true}, is_stubbed{is_stubbed} {} + explicit FenceBase(GPUVAddr address_, u32 payload_, bool is_stubbed_) + : address{address_}, payload{payload_}, is_semaphore{true}, is_stubbed{is_stubbed_} {} GPUVAddr GetAddress() const { return address; @@ -47,6 +48,11 @@ protected: template <typename TFence, typename TTextureCache, typename TTBufferCache, typename TQueryCache> class FenceManager { public: + /// Notify the fence manager about a new frame + void TickFrame() { + delayed_destruction_ring.Tick(); + } + void SignalSemaphore(GPUVAddr addr, u32 value) { TryReleasePendingFences(); const bool should_flush = ShouldFlush(); @@ -86,7 +92,7 @@ public: } else { gpu.IncrementSyncPoint(current_fence->GetPayload()); } - fences.pop(); + PopFence(); } } @@ -132,7 +138,7 @@ private: } else { gpu.IncrementSyncPoint(current_fence->GetPayload()); } - fences.pop(); + PopFence(); } } @@ -158,7 +164,14 @@ private: query_cache.CommitAsyncFlushes(); } + void PopFence() { + delayed_destruction_ring.Push(std::move(fences.front())); + fences.pop(); + } + std::queue<TFence> fences; + + DelayedDestructionRing<TFence, 6> delayed_destruction_ring; }; } // namespace VideoCommon diff --git a/src/video_core/framebuffer_config.h b/src/video_core/framebuffer_config.h new file mode 100644 index 000000000..b86c3a757 --- /dev/null +++ b/src/video_core/framebuffer_config.h @@ -0,0 +1,31 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +namespace Tegra { + +/** + * Struct describing framebuffer configuration + */ +struct FramebufferConfig { + enum class PixelFormat : u32 { + A8B8G8R8_UNORM = 1, + RGB565_UNORM = 4, + B8G8R8A8_UNORM = 5, + }; + + VAddr address{}; + u32 offset{}; + u32 width{}; + u32 height{}; + u32 stride{}; + PixelFormat pixel_format{}; + + using TransformFlags = Service::NVFlinger::BufferQueue::BufferTransformFlags; + TransformFlags transform_flags{}; + Common::Rectangle<int> crop_rect; +}; + +} // namespace Tegra diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp index 4bb9256e9..6ab06775f 100644 --- a/src/video_core/gpu.cpp +++ b/src/video_core/gpu.cpp @@ -10,6 +10,7 @@ #include "core/core_timing.h" #include "core/core_timing_util.h" #include "core/frontend/emu_window.h" +#include "core/hardware_interrupt_manager.h" #include "core/memory.h" #include "core/settings.h" #include "video_core/engines/fermi_2d.h" @@ -27,15 +28,17 @@ namespace Tegra { MICROPROFILE_DEFINE(GPU_wait, "GPU", "Wait for the GPU", MP_RGB(128, 128, 192)); -GPU::GPU(Core::System& system_, bool is_async_) +GPU::GPU(Core::System& system_, bool is_async_, bool use_nvdec_) : system{system_}, memory_manager{std::make_unique<Tegra::MemoryManager>(system)}, dma_pusher{std::make_unique<Tegra::DmaPusher>(system, *this)}, + cdma_pusher{std::make_unique<Tegra::CDmaPusher>(*this)}, use_nvdec{use_nvdec_}, maxwell_3d{std::make_unique<Engines::Maxwell3D>(system, *memory_manager)}, fermi_2d{std::make_unique<Engines::Fermi2D>()}, kepler_compute{std::make_unique<Engines::KeplerCompute>(system, *memory_manager)}, maxwell_dma{std::make_unique<Engines::MaxwellDMA>(system, *memory_manager)}, kepler_memory{std::make_unique<Engines::KeplerMemory>(system, *memory_manager)}, - shader_notify{std::make_unique<VideoCore::ShaderNotify>()}, is_async{is_async_} {} + shader_notify{std::make_unique<VideoCore::ShaderNotify>()}, is_async{is_async_}, + gpu_thread{system_, is_async_} {} GPU::~GPU() = default; @@ -77,31 +80,46 @@ DmaPusher& GPU::DmaPusher() { return *dma_pusher; } +Tegra::CDmaPusher& GPU::CDmaPusher() { + return *cdma_pusher; +} + const DmaPusher& GPU::DmaPusher() const { return *dma_pusher; } +const Tegra::CDmaPusher& GPU::CDmaPusher() const { + return *cdma_pusher; +} + void GPU::WaitFence(u32 syncpoint_id, u32 value) { // Synced GPU, is always in sync if (!is_async) { return; } + if (syncpoint_id == UINT32_MAX) { + // TODO: Research what this does. + LOG_ERROR(HW_GPU, "Waiting for syncpoint -1 not implemented"); + return; + } MICROPROFILE_SCOPE(GPU_wait); std::unique_lock lock{sync_mutex}; - sync_cv.wait(lock, [=, this] { return syncpoints[syncpoint_id].load() >= value; }); + sync_cv.wait(lock, [=, this] { return syncpoints.at(syncpoint_id).load() >= value; }); } void GPU::IncrementSyncPoint(const u32 syncpoint_id) { - syncpoints[syncpoint_id]++; + auto& syncpoint = syncpoints.at(syncpoint_id); + syncpoint++; std::lock_guard lock{sync_mutex}; sync_cv.notify_all(); - if (!syncpt_interrupts[syncpoint_id].empty()) { - u32 value = syncpoints[syncpoint_id].load(); - auto it = syncpt_interrupts[syncpoint_id].begin(); - while (it != syncpt_interrupts[syncpoint_id].end()) { + auto& interrupt = syncpt_interrupts.at(syncpoint_id); + if (!interrupt.empty()) { + u32 value = syncpoint.load(); + auto it = interrupt.begin(); + while (it != interrupt.end()) { if (value >= *it) { TriggerCpuInterrupt(syncpoint_id, *it); - it = syncpt_interrupts[syncpoint_id].erase(it); + it = interrupt.erase(it); continue; } it++; @@ -110,22 +128,22 @@ void GPU::IncrementSyncPoint(const u32 syncpoint_id) { } u32 GPU::GetSyncpointValue(const u32 syncpoint_id) const { - return syncpoints[syncpoint_id].load(); + return syncpoints.at(syncpoint_id).load(); } void GPU::RegisterSyncptInterrupt(const u32 syncpoint_id, const u32 value) { - auto& interrupt = syncpt_interrupts[syncpoint_id]; + auto& interrupt = syncpt_interrupts.at(syncpoint_id); bool contains = std::any_of(interrupt.begin(), interrupt.end(), [value](u32 in_value) { return in_value == value; }); if (contains) { return; } - syncpt_interrupts[syncpoint_id].emplace_back(value); + interrupt.emplace_back(value); } bool GPU::CancelSyncptInterrupt(const u32 syncpoint_id, const u32 value) { std::lock_guard lock{sync_mutex}; - auto& interrupt = syncpt_interrupts[syncpoint_id]; + auto& interrupt = syncpt_interrupts.at(syncpoint_id); const auto iter = std::find_if(interrupt.begin(), interrupt.end(), [value](u32 interrupt_value) { return value == interrupt_value; }); @@ -182,34 +200,6 @@ void GPU::SyncGuestHost() { renderer->Rasterizer().SyncGuestHost(); } -void GPU::OnCommandListEnd() { - renderer->Rasterizer().ReleaseFences(); -} -// Note that, traditionally, methods are treated as 4-byte addressable locations, and hence -// their numbers are written down multiplied by 4 in Docs. Here we are not multiply by 4. -// So the values you see in docs might be multiplied by 4. -enum class BufferMethods { - BindObject = 0x0, - Nop = 0x2, - SemaphoreAddressHigh = 0x4, - SemaphoreAddressLow = 0x5, - SemaphoreSequence = 0x6, - SemaphoreTrigger = 0x7, - NotifyIntr = 0x8, - WrcacheFlush = 0x9, - Unk28 = 0xA, - UnkCacheFlush = 0xB, - RefCnt = 0x14, - SemaphoreAcquire = 0x1A, - SemaphoreRelease = 0x1B, - FenceValue = 0x1C, - FenceAction = 0x1D, - Unk78 = 0x1E, - Unk7c = 0x1F, - Yield = 0x20, - NonPullerMethods = 0x40, -}; - enum class GpuSemaphoreOperation { AcquireEqual = 0x1, WriteLong = 0x2, @@ -240,8 +230,12 @@ void GPU::CallMultiMethod(u32 method, u32 subchannel, const u32* base_start, u32 CallEngineMultiMethod(method, subchannel, base_start, amount, methods_pending); } else { for (std::size_t i = 0; i < amount; i++) { - CallPullerMethod( - {method, base_start[i], subchannel, methods_pending - static_cast<u32>(i)}); + CallPullerMethod(MethodCall{ + method, + base_start[i], + subchannel, + methods_pending - static_cast<u32>(i), + }); } } } @@ -268,7 +262,12 @@ void GPU::CallPullerMethod(const MethodCall& method_call) { case BufferMethods::UnkCacheFlush: case BufferMethods::WrcacheFlush: case BufferMethods::FenceValue: + break; case BufferMethods::FenceAction: + ProcessFenceActionMethod(); + break; + case BufferMethods::WaitForInterrupt: + ProcessWaitForInterruptMethod(); break; case BufferMethods::SemaphoreTrigger: { ProcessSemaphoreTriggerMethod(); @@ -298,8 +297,7 @@ void GPU::CallPullerMethod(const MethodCall& method_call) { break; } default: - LOG_ERROR(HW_GPU, "Special puller engine method {:X} not implemented", - static_cast<u32>(method)); + LOG_ERROR(HW_GPU, "Special puller engine method {:X} not implemented", method); break; } } @@ -378,10 +376,28 @@ void GPU::ProcessBindMethod(const MethodCall& method_call) { dma_pusher->BindSubchannel(kepler_memory.get(), method_call.subchannel); break; default: - UNIMPLEMENTED_MSG("Unimplemented engine {:04X}", static_cast<u32>(engine_id)); + UNIMPLEMENTED_MSG("Unimplemented engine {:04X}", engine_id); + } +} + +void GPU::ProcessFenceActionMethod() { + switch (regs.fence_action.op) { + case FenceOperation::Acquire: + WaitFence(regs.fence_action.syncpoint_id, regs.fence_value); + break; + case FenceOperation::Increment: + IncrementSyncPoint(regs.fence_action.syncpoint_id); + break; + default: + UNIMPLEMENTED_MSG("Unimplemented operation {}", regs.fence_action.op.Value()); } } +void GPU::ProcessWaitForInterruptMethod() { + // TODO(bunnei) ImplementMe + LOG_WARNING(HW_GPU, "(STUBBED) called"); +} + void GPU::ProcessSemaphoreTriggerMethod() { const auto semaphoreOperationMask = 0xF; const auto op = @@ -443,4 +459,75 @@ void GPU::ProcessSemaphoreAcquire() { } } +void GPU::Start() { + gpu_thread.StartThread(*renderer, renderer->Context(), *dma_pusher, *cdma_pusher); + cpu_context = renderer->GetRenderWindow().CreateSharedContext(); + cpu_context->MakeCurrent(); +} + +void GPU::ObtainContext() { + cpu_context->MakeCurrent(); +} + +void GPU::ReleaseContext() { + cpu_context->DoneCurrent(); +} + +void GPU::PushGPUEntries(Tegra::CommandList&& entries) { + gpu_thread.SubmitList(std::move(entries)); +} + +void GPU::PushCommandBuffer(Tegra::ChCommandHeaderList& entries) { + if (!use_nvdec) { + return; + } + // This condition fires when a video stream ends, clear all intermediary data + if (entries[0].raw == 0xDEADB33F) { + cdma_pusher.reset(); + return; + } + if (!cdma_pusher) { + cdma_pusher = std::make_unique<Tegra::CDmaPusher>(*this); + } + + // SubmitCommandBuffer would make the nvdec operations async, this is not currently working + // TODO(ameerj): RE proper async nvdec operation + // gpu_thread.SubmitCommandBuffer(std::move(entries)); + + cdma_pusher->Push(std::move(entries)); + cdma_pusher->DispatchCalls(); +} + +void GPU::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { + gpu_thread.SwapBuffers(framebuffer); +} + +void GPU::FlushRegion(VAddr addr, u64 size) { + gpu_thread.FlushRegion(addr, size); +} + +void GPU::InvalidateRegion(VAddr addr, u64 size) { + gpu_thread.InvalidateRegion(addr, size); +} + +void GPU::FlushAndInvalidateRegion(VAddr addr, u64 size) { + gpu_thread.FlushAndInvalidateRegion(addr, size); +} + +void GPU::TriggerCpuInterrupt(const u32 syncpoint_id, const u32 value) const { + auto& interrupt_manager = system.InterruptManager(); + interrupt_manager.GPUInterruptSyncpt(syncpoint_id, value); +} + +void GPU::WaitIdle() const { + gpu_thread.WaitIdle(); +} + +void GPU::OnCommandListEnd() { + if (is_async) { + // This command only applies to asynchronous GPU mode + gpu_thread.OnCommandListEnd(); + } +} + } // namespace Tegra diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h index 2d15d1c6f..d81e38680 100644 --- a/src/video_core/gpu.h +++ b/src/video_core/gpu.h @@ -13,14 +13,17 @@ #include "common/common_types.h" #include "core/hle/service/nvdrv/nvdata.h" #include "core/hle/service/nvflinger/buffer_queue.h" +#include "video_core/cdma_pusher.h" #include "video_core/dma_pusher.h" +#include "video_core/framebuffer_config.h" +#include "video_core/gpu_thread.h" using CacheAddr = std::uintptr_t; -inline CacheAddr ToCacheAddr(const void* host_ptr) { +[[nodiscard]] inline CacheAddr ToCacheAddr(const void* host_ptr) { return reinterpret_cast<CacheAddr>(host_ptr); } -inline u8* FromCacheAddr(CacheAddr cache_addr) { +[[nodiscard]] inline u8* FromCacheAddr(CacheAddr cache_addr) { return reinterpret_cast<u8*>(cache_addr); } @@ -100,28 +103,6 @@ enum class DepthFormat : u32 { struct CommandListHeader; class DebugContext; -/** - * Struct describing framebuffer configuration - */ -struct FramebufferConfig { - enum class PixelFormat : u32 { - A8B8G8R8_UNORM = 1, - RGB565_UNORM = 4, - B8G8R8A8_UNORM = 5, - }; - - VAddr address; - u32 offset; - u32 width; - u32 height; - u32 stride; - PixelFormat pixel_format; - - using TransformFlags = Service::NVFlinger::BufferQueue::BufferTransformFlags; - TransformFlags transform_flags; - Common::Rectangle<int> crop_rect; -}; - namespace Engines { class Fermi2D; class Maxwell3D; @@ -140,7 +121,7 @@ enum class EngineID { class MemoryManager; -class GPU { +class GPU final { public: struct MethodCall { u32 method{}; @@ -148,17 +129,17 @@ public: u32 subchannel{}; u32 method_count{}; - bool IsLastCall() const { + explicit MethodCall(u32 method_, u32 argument_, u32 subchannel_ = 0, u32 method_count_ = 0) + : method(method_), argument(argument_), subchannel(subchannel_), + method_count(method_count_) {} + + [[nodiscard]] bool IsLastCall() const { return method_count <= 1; } - - MethodCall(u32 method, u32 argument, u32 subchannel = 0, u32 method_count = 0) - : method(method), argument(argument), subchannel(subchannel), - method_count(method_count) {} }; - explicit GPU(Core::System& system, bool is_async); - virtual ~GPU(); + explicit GPU(Core::System& system_, bool is_async_, bool use_nvdec_); + ~GPU(); /// Binds a renderer to the GPU. void BindRenderer(std::unique_ptr<VideoCore::RendererBase> renderer); @@ -175,13 +156,13 @@ public: /// Synchronizes CPU writes with Host GPU memory. void SyncGuestHost(); /// Signal the ending of command list. - virtual void OnCommandListEnd(); + void OnCommandListEnd(); /// Request a host GPU memory flush from the CPU. - u64 RequestFlush(VAddr addr, std::size_t size); + [[nodiscard]] u64 RequestFlush(VAddr addr, std::size_t size); /// Obtains current flush request fence id. - u64 CurrentFlushRequestFence() const { + [[nodiscard]] u64 CurrentFlushRequestFence() const { return current_flush_fence.load(std::memory_order_relaxed); } @@ -189,68 +170,100 @@ public: void TickWork(); /// Returns a reference to the Maxwell3D GPU engine. - Engines::Maxwell3D& Maxwell3D(); + [[nodiscard]] Engines::Maxwell3D& Maxwell3D(); /// Returns a const reference to the Maxwell3D GPU engine. - const Engines::Maxwell3D& Maxwell3D() const; + [[nodiscard]] const Engines::Maxwell3D& Maxwell3D() const; /// Returns a reference to the KeplerCompute GPU engine. - Engines::KeplerCompute& KeplerCompute(); + [[nodiscard]] Engines::KeplerCompute& KeplerCompute(); /// Returns a reference to the KeplerCompute GPU engine. - const Engines::KeplerCompute& KeplerCompute() const; + [[nodiscard]] const Engines::KeplerCompute& KeplerCompute() const; /// Returns a reference to the GPU memory manager. - Tegra::MemoryManager& MemoryManager(); + [[nodiscard]] Tegra::MemoryManager& MemoryManager(); /// Returns a const reference to the GPU memory manager. - const Tegra::MemoryManager& MemoryManager() const; + [[nodiscard]] const Tegra::MemoryManager& MemoryManager() const; /// Returns a reference to the GPU DMA pusher. - Tegra::DmaPusher& DmaPusher(); + [[nodiscard]] Tegra::DmaPusher& DmaPusher(); + + /// Returns a const reference to the GPU DMA pusher. + [[nodiscard]] const Tegra::DmaPusher& DmaPusher() const; + + /// Returns a reference to the GPU CDMA pusher. + [[nodiscard]] Tegra::CDmaPusher& CDmaPusher(); + + /// Returns a const reference to the GPU CDMA pusher. + [[nodiscard]] const Tegra::CDmaPusher& CDmaPusher() const; - VideoCore::RendererBase& Renderer() { + /// Returns a reference to the underlying renderer. + [[nodiscard]] VideoCore::RendererBase& Renderer() { return *renderer; } - const VideoCore::RendererBase& Renderer() const { + /// Returns a const reference to the underlying renderer. + [[nodiscard]] const VideoCore::RendererBase& Renderer() const { return *renderer; } - VideoCore::ShaderNotify& ShaderNotify() { + /// Returns a reference to the shader notifier. + [[nodiscard]] VideoCore::ShaderNotify& ShaderNotify() { return *shader_notify; } - const VideoCore::ShaderNotify& ShaderNotify() const { + /// Returns a const reference to the shader notifier. + [[nodiscard]] const VideoCore::ShaderNotify& ShaderNotify() const { return *shader_notify; } // Waits for the GPU to finish working - virtual void WaitIdle() const = 0; + void WaitIdle() const; /// Allows the CPU/NvFlinger to wait on the GPU before presenting a frame. void WaitFence(u32 syncpoint_id, u32 value); void IncrementSyncPoint(u32 syncpoint_id); - u32 GetSyncpointValue(u32 syncpoint_id) const; + [[nodiscard]] u32 GetSyncpointValue(u32 syncpoint_id) const; void RegisterSyncptInterrupt(u32 syncpoint_id, u32 value); - bool CancelSyncptInterrupt(u32 syncpoint_id, u32 value); + [[nodiscard]] bool CancelSyncptInterrupt(u32 syncpoint_id, u32 value); - u64 GetTicks() const; + [[nodiscard]] u64 GetTicks() const; - std::unique_lock<std::mutex> LockSync() { + [[nodiscard]] std::unique_lock<std::mutex> LockSync() { return std::unique_lock{sync_mutex}; } - bool IsAsync() const { + [[nodiscard]] bool IsAsync() const { return is_async; } - /// Returns a const reference to the GPU DMA pusher. - const Tegra::DmaPusher& DmaPusher() const; + [[nodiscard]] bool UseNvdec() const { + return use_nvdec; + } + + enum class FenceOperation : u32 { + Acquire = 0, + Increment = 1, + }; + + union FenceAction { + u32 raw; + BitField<0, 1, FenceOperation> op; + BitField<8, 24, u32> syncpoint_id; + + [[nodiscard]] static CommandHeader Build(FenceOperation op, u32 syncpoint_id) { + FenceAction result{}; + result.op.Assign(op); + result.syncpoint_id.Assign(syncpoint_id); + return {result.raw}; + } + }; struct Regs { static constexpr size_t NUM_REGS = 0x40; @@ -262,7 +275,7 @@ public: u32 address_high; u32 address_low; - GPUVAddr SemaphoreAddress() const { + [[nodiscard]] GPUVAddr SemaphoreAddress() const { return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low); } @@ -280,10 +293,7 @@ public: u32 semaphore_acquire; u32 semaphore_release; u32 fence_value; - union { - BitField<4, 4, u32> operation; - BitField<8, 8, u32> id; - } fence_action; + FenceAction fence_action; INSERT_UNION_PADDING_WORDS(0xE2); // Puller state @@ -300,34 +310,39 @@ public: /// Performs any additional setup necessary in order to begin GPU emulation. /// This can be used to launch any necessary threads and register any necessary /// core timing events. - virtual void Start() = 0; + void Start(); /// Obtain the CPU Context - virtual void ObtainContext() = 0; + void ObtainContext(); /// Release the CPU Context - virtual void ReleaseContext() = 0; + void ReleaseContext(); /// Push GPU command entries to be processed - virtual void PushGPUEntries(Tegra::CommandList&& entries) = 0; + void PushGPUEntries(Tegra::CommandList&& entries); + + /// Push GPU command buffer entries to be processed + void PushCommandBuffer(Tegra::ChCommandHeaderList& entries); /// Swap buffers (render frame) - virtual void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) = 0; + void SwapBuffers(const Tegra::FramebufferConfig* framebuffer); /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory - virtual void FlushRegion(VAddr addr, u64 size) = 0; + void FlushRegion(VAddr addr, u64 size); /// Notify rasterizer that any caches of the specified region should be invalidated - virtual void InvalidateRegion(VAddr addr, u64 size) = 0; + void InvalidateRegion(VAddr addr, u64 size); /// Notify rasterizer that any caches of the specified region should be flushed and invalidated - virtual void FlushAndInvalidateRegion(VAddr addr, u64 size) = 0; + void FlushAndInvalidateRegion(VAddr addr, u64 size); protected: - virtual void TriggerCpuInterrupt(u32 syncpoint_id, u32 value) const = 0; + void TriggerCpuInterrupt(u32 syncpoint_id, u32 value) const; private: void ProcessBindMethod(const MethodCall& method_call); + void ProcessFenceActionMethod(); + void ProcessWaitForInterruptMethod(); void ProcessSemaphoreTriggerMethod(); void ProcessSemaphoreRelease(); void ProcessSemaphoreAcquire(); @@ -343,13 +358,15 @@ private: u32 methods_pending); /// Determines where the method should be executed. - bool ExecuteMethodOnEngine(u32 method); + [[nodiscard]] bool ExecuteMethodOnEngine(u32 method); protected: Core::System& system; std::unique_ptr<Tegra::MemoryManager> memory_manager; std::unique_ptr<Tegra::DmaPusher> dma_pusher; + std::unique_ptr<Tegra::CDmaPusher> cdma_pusher; std::unique_ptr<VideoCore::RendererBase> renderer; + const bool use_nvdec; private: /// Mapping of command subchannels to their bound engine ids @@ -372,12 +389,13 @@ private: std::array<std::list<u32>, Service::Nvidia::MaxSyncPoints> syncpt_interrupts; std::mutex sync_mutex; + std::mutex device_mutex; std::condition_variable sync_cv; struct FlushRequest { - FlushRequest(u64 fence, VAddr addr, std::size_t size) - : fence{fence}, addr{addr}, size{size} {} + explicit FlushRequest(u64 fence_, VAddr addr_, std::size_t size_) + : fence{fence_}, addr{addr_}, size{size_} {} u64 fence; VAddr addr; std::size_t size; @@ -389,6 +407,9 @@ private: std::mutex flush_request_mutex; const bool is_async; + + VideoCommon::GPUThread::ThreadManager gpu_thread; + std::unique_ptr<Core::Frontend::GraphicsContext> cpu_context; }; #define ASSERT_REG_POSITION(field_name, position) \ diff --git a/src/video_core/gpu_asynch.cpp b/src/video_core/gpu_asynch.cpp deleted file mode 100644 index 70a3d5738..000000000 --- a/src/video_core/gpu_asynch.cpp +++ /dev/null @@ -1,64 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "core/core.h" -#include "core/hardware_interrupt_manager.h" -#include "video_core/gpu_asynch.h" -#include "video_core/gpu_thread.h" -#include "video_core/renderer_base.h" - -namespace VideoCommon { - -GPUAsynch::GPUAsynch(Core::System& system) : GPU{system, true}, gpu_thread{system} {} - -GPUAsynch::~GPUAsynch() = default; - -void GPUAsynch::Start() { - gpu_thread.StartThread(*renderer, renderer->Context(), *dma_pusher); - cpu_context = renderer->GetRenderWindow().CreateSharedContext(); - cpu_context->MakeCurrent(); -} - -void GPUAsynch::ObtainContext() { - cpu_context->MakeCurrent(); -} - -void GPUAsynch::ReleaseContext() { - cpu_context->DoneCurrent(); -} - -void GPUAsynch::PushGPUEntries(Tegra::CommandList&& entries) { - gpu_thread.SubmitList(std::move(entries)); -} - -void GPUAsynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { - gpu_thread.SwapBuffers(framebuffer); -} - -void GPUAsynch::FlushRegion(VAddr addr, u64 size) { - gpu_thread.FlushRegion(addr, size); -} - -void GPUAsynch::InvalidateRegion(VAddr addr, u64 size) { - gpu_thread.InvalidateRegion(addr, size); -} - -void GPUAsynch::FlushAndInvalidateRegion(VAddr addr, u64 size) { - gpu_thread.FlushAndInvalidateRegion(addr, size); -} - -void GPUAsynch::TriggerCpuInterrupt(const u32 syncpoint_id, const u32 value) const { - auto& interrupt_manager = system.InterruptManager(); - interrupt_manager.GPUInterruptSyncpt(syncpoint_id, value); -} - -void GPUAsynch::WaitIdle() const { - gpu_thread.WaitIdle(); -} - -void GPUAsynch::OnCommandListEnd() { - gpu_thread.OnCommandListEnd(); -} - -} // namespace VideoCommon diff --git a/src/video_core/gpu_asynch.h b/src/video_core/gpu_asynch.h deleted file mode 100644 index f89c855a5..000000000 --- a/src/video_core/gpu_asynch.h +++ /dev/null @@ -1,46 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "video_core/gpu.h" -#include "video_core/gpu_thread.h" - -namespace Core::Frontend { -class GraphicsContext; -} - -namespace VideoCore { -class RendererBase; -} // namespace VideoCore - -namespace VideoCommon { - -/// Implementation of GPU interface that runs the GPU asynchronously -class GPUAsynch final : public Tegra::GPU { -public: - explicit GPUAsynch(Core::System& system); - ~GPUAsynch() override; - - void Start() override; - void ObtainContext() override; - void ReleaseContext() override; - void PushGPUEntries(Tegra::CommandList&& entries) override; - void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override; - void FlushRegion(VAddr addr, u64 size) override; - void InvalidateRegion(VAddr addr, u64 size) override; - void FlushAndInvalidateRegion(VAddr addr, u64 size) override; - void WaitIdle() const override; - - void OnCommandListEnd() override; - -protected: - void TriggerCpuInterrupt(u32 syncpoint_id, u32 value) const override; - -private: - GPUThread::ThreadManager gpu_thread; - std::unique_ptr<Core::Frontend::GraphicsContext> cpu_context; -}; - -} // namespace VideoCommon diff --git a/src/video_core/gpu_synch.cpp b/src/video_core/gpu_synch.cpp deleted file mode 100644 index 1ca47ddef..000000000 --- a/src/video_core/gpu_synch.cpp +++ /dev/null @@ -1,45 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "video_core/gpu_synch.h" -#include "video_core/renderer_base.h" - -namespace VideoCommon { - -GPUSynch::GPUSynch(Core::System& system) : GPU{system, false} {} - -GPUSynch::~GPUSynch() = default; - -void GPUSynch::Start() {} - -void GPUSynch::ObtainContext() { - renderer->Context().MakeCurrent(); -} - -void GPUSynch::ReleaseContext() { - renderer->Context().DoneCurrent(); -} - -void GPUSynch::PushGPUEntries(Tegra::CommandList&& entries) { - dma_pusher->Push(std::move(entries)); - dma_pusher->DispatchCalls(); -} - -void GPUSynch::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { - renderer->SwapBuffers(framebuffer); -} - -void GPUSynch::FlushRegion(VAddr addr, u64 size) { - renderer->Rasterizer().FlushRegion(addr, size); -} - -void GPUSynch::InvalidateRegion(VAddr addr, u64 size) { - renderer->Rasterizer().InvalidateRegion(addr, size); -} - -void GPUSynch::FlushAndInvalidateRegion(VAddr addr, u64 size) { - renderer->Rasterizer().FlushAndInvalidateRegion(addr, size); -} - -} // namespace VideoCommon diff --git a/src/video_core/gpu_synch.h b/src/video_core/gpu_synch.h deleted file mode 100644 index 297258cb1..000000000 --- a/src/video_core/gpu_synch.h +++ /dev/null @@ -1,40 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "video_core/gpu.h" - -namespace Core::Frontend { -class GraphicsContext; -} - -namespace VideoCore { -class RendererBase; -} // namespace VideoCore - -namespace VideoCommon { - -/// Implementation of GPU interface that runs the GPU synchronously -class GPUSynch final : public Tegra::GPU { -public: - explicit GPUSynch(Core::System& system); - ~GPUSynch() override; - - void Start() override; - void ObtainContext() override; - void ReleaseContext() override; - void PushGPUEntries(Tegra::CommandList&& entries) override; - void SwapBuffers(const Tegra::FramebufferConfig* framebuffer) override; - void FlushRegion(VAddr addr, u64 size) override; - void InvalidateRegion(VAddr addr, u64 size) override; - void FlushAndInvalidateRegion(VAddr addr, u64 size) override; - void WaitIdle() const override {} - -protected: - void TriggerCpuInterrupt([[maybe_unused]] u32 syncpoint_id, - [[maybe_unused]] u32 value) const override {} -}; - -} // namespace VideoCommon diff --git a/src/video_core/gpu_thread.cpp b/src/video_core/gpu_thread.cpp index bf761abf2..7e490bcc3 100644 --- a/src/video_core/gpu_thread.cpp +++ b/src/video_core/gpu_thread.cpp @@ -4,6 +4,7 @@ #include "common/assert.h" #include "common/microprofile.h" +#include "common/scope_exit.h" #include "common/thread.h" #include "core/core.h" #include "core/frontend/emu_window.h" @@ -18,9 +19,11 @@ namespace VideoCommon::GPUThread { /// Runs the GPU thread static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, Tegra::DmaPusher& dma_pusher, - SynchState& state) { + SynchState& state, Tegra::CDmaPusher& cdma_pusher) { std::string name = "yuzu:GPU"; MicroProfileOnThreadCreate(name.c_str()); + SCOPE_EXIT({ MicroProfileOnThreadExit(); }); + Common::SetCurrentThreadName(name.c_str()); Common::SetCurrentThreadPriority(Common::ThreadPriority::High); system.RegisterHostThread(); @@ -39,19 +42,23 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, CommandDataContainer next; while (state.is_running) { next = state.queue.PopWait(); - if (const auto submit_list = std::get_if<SubmitListCommand>(&next.data)) { + if (auto* submit_list = std::get_if<SubmitListCommand>(&next.data)) { dma_pusher.Push(std::move(submit_list->entries)); dma_pusher.DispatchCalls(); - } else if (const auto data = std::get_if<SwapBuffersCommand>(&next.data)) { + } else if (auto* command_list = std::get_if<SubmitChCommandEntries>(&next.data)) { + // NVDEC + cdma_pusher.Push(std::move(command_list->entries)); + cdma_pusher.DispatchCalls(); + } else if (const auto* data = std::get_if<SwapBuffersCommand>(&next.data)) { renderer.SwapBuffers(data->framebuffer ? &*data->framebuffer : nullptr); } else if (std::holds_alternative<OnCommandListEndCommand>(next.data)) { renderer.Rasterizer().ReleaseFences(); } else if (std::holds_alternative<GPUTickCommand>(next.data)) { system.GPU().TickWork(); - } else if (const auto data = std::get_if<FlushRegionCommand>(&next.data)) { - renderer.Rasterizer().FlushRegion(data->addr, data->size); - } else if (const auto data = std::get_if<InvalidateRegionCommand>(&next.data)) { - renderer.Rasterizer().OnCPUWrite(data->addr, data->size); + } else if (const auto* flush = std::get_if<FlushRegionCommand>(&next.data)) { + renderer.Rasterizer().FlushRegion(flush->addr, flush->size); + } else if (const auto* invalidate = std::get_if<InvalidateRegionCommand>(&next.data)) { + renderer.Rasterizer().OnCPUWrite(invalidate->addr, invalidate->size); } else if (std::holds_alternative<EndProcessingCommand>(next.data)) { return; } else { @@ -61,7 +68,8 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer, } } -ThreadManager::ThreadManager(Core::System& system) : system{system} {} +ThreadManager::ThreadManager(Core::System& system_, bool is_async_) + : system{system_}, is_async{is_async_} {} ThreadManager::~ThreadManager() { if (!thread.joinable()) { @@ -75,33 +83,48 @@ ThreadManager::~ThreadManager() { void ThreadManager::StartThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, - Tegra::DmaPusher& dma_pusher) { - thread = std::thread{RunThread, std::ref(system), std::ref(renderer), - std::ref(context), std::ref(dma_pusher), std::ref(state)}; + Tegra::DmaPusher& dma_pusher, Tegra::CDmaPusher& cdma_pusher) { + thread = std::thread(RunThread, std::ref(system), std::ref(renderer), std::ref(context), + std::ref(dma_pusher), std::ref(state), std::ref(cdma_pusher)); } void ThreadManager::SubmitList(Tegra::CommandList&& entries) { PushCommand(SubmitListCommand(std::move(entries))); } +void ThreadManager::SubmitCommandBuffer(Tegra::ChCommandHeaderList&& entries) { + PushCommand(SubmitChCommandEntries(std::move(entries))); +} + void ThreadManager::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { PushCommand(SwapBuffersCommand(framebuffer ? std::make_optional(*framebuffer) : std::nullopt)); } void ThreadManager::FlushRegion(VAddr addr, u64 size) { - if (!Settings::IsGPULevelHigh()) { + if (!is_async) { + // Always flush with synchronous GPU mode PushCommand(FlushRegionCommand(addr, size)); return; } - if (!Settings::IsGPULevelExtreme()) { - return; - } - if (system.Renderer().Rasterizer().MustFlushRegion(addr, size)) { + + // Asynchronous GPU mode + switch (Settings::values.gpu_accuracy.GetValue()) { + case Settings::GPUAccuracy::Normal: + PushCommand(FlushRegionCommand(addr, size)); + break; + case Settings::GPUAccuracy::High: + // TODO(bunnei): Is this right? Preserving existing behavior for now + break; + case Settings::GPUAccuracy::Extreme: { auto& gpu = system.GPU(); u64 fence = gpu.RequestFlush(addr, size); PushCommand(GPUTickCommand()); while (fence > gpu.CurrentFlushRequestFence()) { } + break; + } + default: + UNIMPLEMENTED_MSG("Unsupported gpu_accuracy {}", Settings::values.gpu_accuracy.GetValue()); } } @@ -115,7 +138,8 @@ void ThreadManager::FlushAndInvalidateRegion(VAddr addr, u64 size) { } void ThreadManager::WaitIdle() const { - while (state.last_fence > state.signaled_fence.load(std::memory_order_relaxed)) { + while (state.last_fence > state.signaled_fence.load(std::memory_order_relaxed) && + system.IsPoweredOn()) { } } @@ -126,6 +150,12 @@ void ThreadManager::OnCommandListEnd() { u64 ThreadManager::PushCommand(CommandData&& command_data) { const u64 fence{++state.last_fence}; state.queue.Push(CommandDataContainer(std::move(command_data), fence)); + + if (!is_async) { + // In synchronous GPU mode, block the caller until the command has executed + WaitIdle(); + } + return fence; } diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index 5a28335d6..2775629e7 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -10,8 +10,9 @@ #include <optional> #include <thread> #include <variant> + #include "common/threadsafe_queue.h" -#include "video_core/gpu.h" +#include "video_core/framebuffer_config.h" namespace Tegra { struct FramebufferConfig; @@ -25,6 +26,10 @@ class GraphicsContext; class System; } // namespace Core +namespace VideoCore { +class RendererBase; +} // namespace VideoCore + namespace VideoCommon::GPUThread { /// Command to signal to the GPU thread that processing has ended @@ -32,22 +37,30 @@ struct EndProcessingCommand final {}; /// Command to signal to the GPU thread that a command list is ready for processing struct SubmitListCommand final { - explicit SubmitListCommand(Tegra::CommandList&& entries) : entries{std::move(entries)} {} + explicit SubmitListCommand(Tegra::CommandList&& entries_) : entries{std::move(entries_)} {} Tegra::CommandList entries; }; +/// Command to signal to the GPU thread that a cdma command list is ready for processing +struct SubmitChCommandEntries final { + explicit SubmitChCommandEntries(Tegra::ChCommandHeaderList&& entries_) + : entries{std::move(entries_)} {} + + Tegra::ChCommandHeaderList entries; +}; + /// Command to signal to the GPU thread that a swap buffers is pending struct SwapBuffersCommand final { - explicit SwapBuffersCommand(std::optional<const Tegra::FramebufferConfig> framebuffer) - : framebuffer{std::move(framebuffer)} {} + explicit SwapBuffersCommand(std::optional<const Tegra::FramebufferConfig> framebuffer_) + : framebuffer{std::move(framebuffer_)} {} std::optional<Tegra::FramebufferConfig> framebuffer; }; /// Command to signal to the GPU thread to flush a region struct FlushRegionCommand final { - explicit constexpr FlushRegionCommand(VAddr addr, u64 size) : addr{addr}, size{size} {} + explicit constexpr FlushRegionCommand(VAddr addr_, u64 size_) : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -55,7 +68,7 @@ struct FlushRegionCommand final { /// Command to signal to the GPU thread to invalidate a region struct InvalidateRegionCommand final { - explicit constexpr InvalidateRegionCommand(VAddr addr, u64 size) : addr{addr}, size{size} {} + explicit constexpr InvalidateRegionCommand(VAddr addr_, u64 size_) : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -63,8 +76,8 @@ struct InvalidateRegionCommand final { /// Command to signal to the GPU thread to flush and invalidate a region struct FlushAndInvalidateRegionCommand final { - explicit constexpr FlushAndInvalidateRegionCommand(VAddr addr, u64 size) - : addr{addr}, size{size} {} + explicit constexpr FlushAndInvalidateRegionCommand(VAddr addr_, u64 size_) + : addr{addr_}, size{size_} {} VAddr addr; u64 size; @@ -77,15 +90,15 @@ struct OnCommandListEndCommand final {}; struct GPUTickCommand final {}; using CommandData = - std::variant<EndProcessingCommand, SubmitListCommand, SwapBuffersCommand, FlushRegionCommand, - InvalidateRegionCommand, FlushAndInvalidateRegionCommand, OnCommandListEndCommand, - GPUTickCommand>; + std::variant<EndProcessingCommand, SubmitListCommand, SubmitChCommandEntries, + SwapBuffersCommand, FlushRegionCommand, InvalidateRegionCommand, + FlushAndInvalidateRegionCommand, OnCommandListEndCommand, GPUTickCommand>; struct CommandDataContainer { CommandDataContainer() = default; - CommandDataContainer(CommandData&& data, u64 next_fence) - : data{std::move(data)}, fence{next_fence} {} + explicit CommandDataContainer(CommandData&& data_, u64 next_fence_) + : data{std::move(data_)}, fence{next_fence_} {} CommandData data; u64 fence{}; @@ -104,16 +117,19 @@ struct SynchState final { /// Class used to manage the GPU thread class ThreadManager final { public: - explicit ThreadManager(Core::System& system); + explicit ThreadManager(Core::System& system_, bool is_async_); ~ThreadManager(); /// Creates and starts the GPU thread. void StartThread(VideoCore::RendererBase& renderer, Core::Frontend::GraphicsContext& context, - Tegra::DmaPusher& dma_pusher); + Tegra::DmaPusher& dma_pusher, Tegra::CDmaPusher& cdma_pusher); /// Push GPU command entries to be processed void SubmitList(Tegra::CommandList&& entries); + /// Push GPU CDMA command buffer entries to be processed + void SubmitCommandBuffer(Tegra::ChCommandHeaderList&& entries); + /// Swap buffers (render frame) void SwapBuffers(const Tegra::FramebufferConfig* framebuffer); @@ -135,11 +151,11 @@ private: /// Pushes a command to be executed by the GPU thread u64 PushCommand(CommandData&& command_data); -private: SynchState state; Core::System& system; std::thread thread; std::thread::id thread_id; + const bool is_async; }; } // namespace VideoCommon::GPUThread diff --git a/src/video_core/guest_driver.h b/src/video_core/guest_driver.h index 99450777e..21e569ba1 100644 --- a/src/video_core/guest_driver.h +++ b/src/video_core/guest_driver.h @@ -19,8 +19,8 @@ namespace VideoCore { class GuestDriverProfile { public: explicit GuestDriverProfile() = default; - explicit GuestDriverProfile(std::optional<u32> texture_handler_size) - : texture_handler_size{texture_handler_size} {} + explicit GuestDriverProfile(std::optional<u32> texture_handler_size_) + : texture_handler_size{texture_handler_size_} {} void DeduceTextureHandlerSize(std::vector<u32> bound_offsets); diff --git a/src/video_core/host_shaders/CMakeLists.txt b/src/video_core/host_shaders/CMakeLists.txt index aa62363a7..4c7399d5a 100644 --- a/src/video_core/host_shaders/CMakeLists.txt +++ b/src/video_core/host_shaders/CMakeLists.txt @@ -1,18 +1,29 @@ set(SHADER_FILES + block_linear_unswizzle_2d.comp + block_linear_unswizzle_3d.comp + convert_depth_to_float.frag + convert_float_to_depth.frag + full_screen_triangle.vert + opengl_copy_bc4.comp opengl_present.frag opengl_present.vert + pitch_unswizzle.comp + vulkan_blit_color_float.frag + vulkan_blit_depth_stencil.frag + vulkan_present.frag + vulkan_present.vert + vulkan_quad_array.comp + vulkan_quad_indexed.comp + vulkan_uint8.comp ) -set(SHADER_INCLUDE ${CMAKE_CURRENT_BINARY_DIR}/include) -set(HOST_SHADERS_INCLUDE ${SHADER_INCLUDE} PARENT_SCOPE) +find_program(GLSLANGVALIDATOR "glslangValidator" REQUIRED) + +set(GLSL_FLAGS "") +set(SHADER_INCLUDE ${CMAKE_CURRENT_BINARY_DIR}/include) set(SHADER_DIR ${SHADER_INCLUDE}/video_core/host_shaders) -add_custom_command( - OUTPUT - ${SHADER_DIR} - COMMAND - ${CMAKE_COMMAND} -E make_directory ${SHADER_DIR} -) +set(HOST_SHADERS_INCLUDE ${SHADER_INCLUDE} PARENT_SCOPE) set(INPUT_FILE ${CMAKE_CURRENT_SOURCE_DIR}/source_shader.h.in) set(HEADER_GENERATOR ${CMAKE_CURRENT_SOURCE_DIR}/StringShaderHeader.cmake) @@ -20,19 +31,36 @@ set(HEADER_GENERATOR ${CMAKE_CURRENT_SOURCE_DIR}/StringShaderHeader.cmake) foreach(FILENAME IN ITEMS ${SHADER_FILES}) string(REPLACE "." "_" SHADER_NAME ${FILENAME}) set(SOURCE_FILE ${CMAKE_CURRENT_SOURCE_DIR}/${FILENAME}) - set(HEADER_FILE ${SHADER_DIR}/${SHADER_NAME}.h) - add_custom_command( - OUTPUT - ${HEADER_FILE} - COMMAND - ${CMAKE_COMMAND} -P ${HEADER_GENERATOR} ${SOURCE_FILE} ${HEADER_FILE} ${INPUT_FILE} - MAIN_DEPENDENCY - ${SOURCE_FILE} - DEPENDS - ${HEADER_GENERATOR} - ${INPUT_FILE} - ) - set(SHADER_HEADERS ${SHADER_HEADERS} ${HEADER_FILE}) + # Skip generating source headers on Vulkan exclusive files + if (NOT ${FILENAME} MATCHES "vulkan.*") + set(SOURCE_HEADER_FILE ${SHADER_DIR}/${SHADER_NAME}.h) + add_custom_command( + OUTPUT + ${SOURCE_HEADER_FILE} + COMMAND + ${CMAKE_COMMAND} -P ${HEADER_GENERATOR} ${SOURCE_FILE} ${SOURCE_HEADER_FILE} ${INPUT_FILE} + MAIN_DEPENDENCY + ${SOURCE_FILE} + DEPENDS + ${INPUT_FILE} + # HEADER_GENERATOR should be included here but msbuild seems to assume it's always modified + ) + set(SHADER_HEADERS ${SHADER_HEADERS} ${SOURCE_HEADER_FILE}) + endif() + # Skip compiling to SPIR-V OpenGL exclusive files + if (NOT ${FILENAME} MATCHES "opengl.*") + string(TOUPPER ${SHADER_NAME}_SPV SPIRV_VARIABLE_NAME) + set(SPIRV_HEADER_FILE ${SHADER_DIR}/${SHADER_NAME}_spv.h) + add_custom_command( + OUTPUT + ${SPIRV_HEADER_FILE} + COMMAND + ${GLSLANGVALIDATOR} -V ${GLSL_FLAGS} --variable-name ${SPIRV_VARIABLE_NAME} -o ${SPIRV_HEADER_FILE} ${SOURCE_FILE} + MAIN_DEPENDENCY + ${SOURCE_FILE} + ) + set(SHADER_HEADERS ${SHADER_HEADERS} ${SPIRV_HEADER_FILE}) + endif() endforeach() add_custom_target(host_shaders diff --git a/src/video_core/host_shaders/StringShaderHeader.cmake b/src/video_core/host_shaders/StringShaderHeader.cmake index 368bce0ed..c0fc49768 100644 --- a/src/video_core/host_shaders/StringShaderHeader.cmake +++ b/src/video_core/host_shaders/StringShaderHeader.cmake @@ -8,4 +8,6 @@ string(TOUPPER ${CONTENTS_NAME} CONTENTS_NAME) file(READ ${SOURCE_FILE} CONTENTS) +get_filename_component(OUTPUT_DIR ${HEADER_FILE} DIRECTORY) +make_directory(${OUTPUT_DIR}) configure_file(${INPUT_FILE} ${HEADER_FILE} @ONLY) diff --git a/src/video_core/host_shaders/block_linear_unswizzle_2d.comp b/src/video_core/host_shaders/block_linear_unswizzle_2d.comp new file mode 100644 index 000000000..a131be79e --- /dev/null +++ b/src/video_core/host_shaders/block_linear_unswizzle_2d.comp @@ -0,0 +1,122 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 430 + +#ifdef VULKAN + +#extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_shader_8bit_storage : require +#define HAS_EXTENDED_TYPES 1 +#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants { +#define END_PUSH_CONSTANTS }; +#define UNIFORM(n) +#define BINDING_SWIZZLE_BUFFER 0 +#define BINDING_INPUT_BUFFER 1 +#define BINDING_OUTPUT_IMAGE 2 + +#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv + +#extension GL_NV_gpu_shader5 : enable +#ifdef GL_NV_gpu_shader5 +#define HAS_EXTENDED_TYPES 1 +#else +#define HAS_EXTENDED_TYPES 0 +#endif +#define BEGIN_PUSH_CONSTANTS +#define END_PUSH_CONSTANTS +#define UNIFORM(n) layout (location = n) uniform +#define BINDING_SWIZZLE_BUFFER 0 +#define BINDING_INPUT_BUFFER 1 +#define BINDING_OUTPUT_IMAGE 0 + +#endif + +BEGIN_PUSH_CONSTANTS +UNIFORM(0) uvec3 origin; +UNIFORM(1) ivec3 destination; +UNIFORM(2) uint bytes_per_block_log2; +UNIFORM(3) uint layer_stride; +UNIFORM(4) uint block_size; +UNIFORM(5) uint x_shift; +UNIFORM(6) uint block_height; +UNIFORM(7) uint block_height_mask; +END_PUSH_CONSTANTS + +layout(binding = BINDING_SWIZZLE_BUFFER, std430) readonly buffer SwizzleTable { + uint swizzle_table[]; +}; + +#if HAS_EXTENDED_TYPES +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU8 { uint8_t u8data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU16 { uint16_t u16data[]; }; +#endif +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU32 { uint u32data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU64 { uvec2 u64data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU128 { uvec4 u128data[]; }; + +layout(binding = BINDING_OUTPUT_IMAGE) uniform writeonly uimage2DArray output_image; + +layout(local_size_x = 32, local_size_y = 32, local_size_z = 1) in; + +const uint GOB_SIZE_X = 64; +const uint GOB_SIZE_Y = 8; +const uint GOB_SIZE_Z = 1; +const uint GOB_SIZE = GOB_SIZE_X * GOB_SIZE_Y * GOB_SIZE_Z; + +const uint GOB_SIZE_X_SHIFT = 6; +const uint GOB_SIZE_Y_SHIFT = 3; +const uint GOB_SIZE_Z_SHIFT = 0; +const uint GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHIFT; + +const uvec2 SWIZZLE_MASK = uvec2(GOB_SIZE_X - 1, GOB_SIZE_Y - 1); + +uint SwizzleOffset(uvec2 pos) { + pos = pos & SWIZZLE_MASK; + return swizzle_table[pos.y * 64 + pos.x]; +} + +uvec4 ReadTexel(uint offset) { + switch (bytes_per_block_log2) { +#if HAS_EXTENDED_TYPES + case 0: + return uvec4(u8data[offset], 0, 0, 0); + case 1: + return uvec4(u16data[offset / 2], 0, 0, 0); +#else + case 0: + return uvec4(bitfieldExtract(u32data[offset / 4], int((offset * 8) & 24), 8), 0, 0, 0); + case 1: + return uvec4(bitfieldExtract(u32data[offset / 4], int((offset * 8) & 16), 16), 0, 0, 0); +#endif + case 2: + return uvec4(u32data[offset / 4], 0, 0, 0); + case 3: + return uvec4(u64data[offset / 8], 0, 0); + case 4: + return u128data[offset / 16]; + } + return uvec4(0); +} + +void main() { + uvec3 pos = gl_GlobalInvocationID + origin; + pos.x <<= bytes_per_block_log2; + + // Read as soon as possible due to its latency + const uint swizzle = SwizzleOffset(pos.xy); + + const uint block_y = pos.y >> GOB_SIZE_Y_SHIFT; + + uint offset = 0; + offset += pos.z * layer_stride; + offset += (block_y >> block_height) * block_size; + offset += (block_y & block_height_mask) << GOB_SIZE_SHIFT; + offset += (pos.x >> GOB_SIZE_X_SHIFT) << x_shift; + offset += swizzle; + + const uvec4 texel = ReadTexel(offset); + const ivec3 coord = ivec3(gl_GlobalInvocationID) + destination; + imageStore(output_image, coord, texel); +} diff --git a/src/video_core/host_shaders/block_linear_unswizzle_3d.comp b/src/video_core/host_shaders/block_linear_unswizzle_3d.comp new file mode 100644 index 000000000..bb6872e6b --- /dev/null +++ b/src/video_core/host_shaders/block_linear_unswizzle_3d.comp @@ -0,0 +1,125 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 430 + +#ifdef VULKAN + +#extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_shader_8bit_storage : require +#define HAS_EXTENDED_TYPES 1 +#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants { +#define END_PUSH_CONSTANTS }; +#define UNIFORM(n) +#define BINDING_SWIZZLE_BUFFER 0 +#define BINDING_INPUT_BUFFER 1 +#define BINDING_OUTPUT_IMAGE 2 + +#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv + +#extension GL_NV_gpu_shader5 : enable +#ifdef GL_NV_gpu_shader5 +#define HAS_EXTENDED_TYPES 1 +#else +#define HAS_EXTENDED_TYPES 0 +#endif +#define BEGIN_PUSH_CONSTANTS +#define END_PUSH_CONSTANTS +#define UNIFORM(n) layout (location = n) uniform +#define BINDING_SWIZZLE_BUFFER 0 +#define BINDING_INPUT_BUFFER 1 +#define BINDING_OUTPUT_IMAGE 0 + +#endif + +BEGIN_PUSH_CONSTANTS +UNIFORM(0) uvec3 origin; +UNIFORM(1) ivec3 destination; +UNIFORM(2) uint bytes_per_block_log2; +UNIFORM(3) uint slice_size; +UNIFORM(4) uint block_size; +UNIFORM(5) uint x_shift; +UNIFORM(6) uint block_height; +UNIFORM(7) uint block_height_mask; +UNIFORM(8) uint block_depth; +UNIFORM(9) uint block_depth_mask; +END_PUSH_CONSTANTS + +layout(binding = BINDING_SWIZZLE_BUFFER, std430) readonly buffer SwizzleTable { + uint swizzle_table[]; +}; + +#if HAS_EXTENDED_TYPES +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU8 { uint8_t u8data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU16 { uint16_t u16data[]; }; +#endif +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU32 { uint u32data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU64 { uvec2 u64data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) buffer InputBufferU128 { uvec4 u128data[]; }; + +layout(binding = BINDING_OUTPUT_IMAGE) uniform writeonly uimage3D output_image; + +layout(local_size_x = 16, local_size_y = 8, local_size_z = 8) in; + +const uint GOB_SIZE_X = 64; +const uint GOB_SIZE_Y = 8; +const uint GOB_SIZE_Z = 1; +const uint GOB_SIZE = GOB_SIZE_X * GOB_SIZE_Y * GOB_SIZE_Z; + +const uint GOB_SIZE_X_SHIFT = 6; +const uint GOB_SIZE_Y_SHIFT = 3; +const uint GOB_SIZE_Z_SHIFT = 0; +const uint GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHIFT; + +const uvec2 SWIZZLE_MASK = uvec2(GOB_SIZE_X - 1, GOB_SIZE_Y - 1); + +uint SwizzleOffset(uvec2 pos) { + pos = pos & SWIZZLE_MASK; + return swizzle_table[pos.y * 64 + pos.x]; +} + +uvec4 ReadTexel(uint offset) { + switch (bytes_per_block_log2) { +#if HAS_EXTENDED_TYPES + case 0: + return uvec4(u8data[offset], 0, 0, 0); + case 1: + return uvec4(u16data[offset / 2], 0, 0, 0); +#else + case 0: + return uvec4(bitfieldExtract(u32data[offset / 4], int((offset * 8) & 24), 8), 0, 0, 0); + case 1: + return uvec4(bitfieldExtract(u32data[offset / 4], int((offset * 8) & 16), 16), 0, 0, 0); +#endif + case 2: + return uvec4(u32data[offset / 4], 0, 0, 0); + case 3: + return uvec4(u64data[offset / 8], 0, 0); + case 4: + return u128data[offset / 16]; + } + return uvec4(0); +} + +void main() { + uvec3 pos = gl_GlobalInvocationID + origin; + pos.x <<= bytes_per_block_log2; + + // Read as soon as possible due to its latency + const uint swizzle = SwizzleOffset(pos.xy); + + const uint block_y = pos.y >> GOB_SIZE_Y_SHIFT; + + uint offset = 0; + offset += (pos.z >> block_depth) * slice_size; + offset += (pos.z & block_depth_mask) << (GOB_SIZE_SHIFT + block_height); + offset += (block_y >> block_height) * block_size; + offset += (block_y & block_height_mask) << GOB_SIZE_SHIFT; + offset += (pos.x >> GOB_SIZE_X_SHIFT) << x_shift; + offset += swizzle; + + const uvec4 texel = ReadTexel(offset); + const ivec3 coord = ivec3(gl_GlobalInvocationID) + destination; + imageStore(output_image, coord, texel); +} diff --git a/src/video_core/host_shaders/convert_depth_to_float.frag b/src/video_core/host_shaders/convert_depth_to_float.frag new file mode 100644 index 000000000..624c58509 --- /dev/null +++ b/src/video_core/host_shaders/convert_depth_to_float.frag @@ -0,0 +1,13 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +layout(binding = 0) uniform sampler2D depth_texture; +layout(location = 0) out float output_color; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + output_color = texelFetch(depth_texture, coord, 0).r; +} diff --git a/src/video_core/host_shaders/convert_float_to_depth.frag b/src/video_core/host_shaders/convert_float_to_depth.frag new file mode 100644 index 000000000..d86c795f4 --- /dev/null +++ b/src/video_core/host_shaders/convert_float_to_depth.frag @@ -0,0 +1,13 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +layout(binding = 0) uniform sampler2D color_texture; + +void main() { + ivec2 coord = ivec2(gl_FragCoord.xy); + float color = texelFetch(color_texture, coord, 0).r; + gl_FragDepth = color; +} diff --git a/src/video_core/host_shaders/full_screen_triangle.vert b/src/video_core/host_shaders/full_screen_triangle.vert new file mode 100644 index 000000000..452ad6502 --- /dev/null +++ b/src/video_core/host_shaders/full_screen_triangle.vert @@ -0,0 +1,29 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +#ifdef VULKAN +#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants { +#define END_PUSH_CONSTANTS }; +#define UNIFORM(n) +#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv +#define BEGIN_PUSH_CONSTANTS +#define END_PUSH_CONSTANTS +#define UNIFORM(n) layout (location = n) uniform +#endif + +BEGIN_PUSH_CONSTANTS +UNIFORM(0) vec2 tex_scale; +UNIFORM(1) vec2 tex_offset; +END_PUSH_CONSTANTS + +layout(location = 0) out vec2 texcoord; + +void main() { + float x = float((gl_VertexIndex & 1) << 2); + float y = float((gl_VertexIndex & 2) << 1); + gl_Position = vec4(x - 1.0, y - 1.0, 0.0, 1.0); + texcoord = fma(vec2(x, y) / 2.0, tex_scale, tex_offset); +} diff --git a/src/video_core/host_shaders/opengl_copy_bc4.comp b/src/video_core/host_shaders/opengl_copy_bc4.comp new file mode 100644 index 000000000..7b8e20fbe --- /dev/null +++ b/src/video_core/host_shaders/opengl_copy_bc4.comp @@ -0,0 +1,70 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 430 core +#extension GL_ARB_gpu_shader_int64 : require + +layout (local_size_x = 4, local_size_y = 4) in; + +layout(binding = 0, rg32ui) readonly uniform uimage3D bc4_input; +layout(binding = 1, rgba8ui) writeonly uniform uimage3D bc4_output; + +layout(location = 0) uniform uvec3 src_offset; +layout(location = 1) uniform uvec3 dst_offset; + +// https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_texture_compression_rgtc.txt +uint DecompressBlock(uint64_t bits, uvec2 coord) { + const uint code_offset = 16 + 3 * (4 * coord.y + coord.x); + const uint code = uint(bits >> code_offset) & 7; + const uint red0 = uint(bits >> 0) & 0xff; + const uint red1 = uint(bits >> 8) & 0xff; + if (red0 > red1) { + switch (code) { + case 0: + return red0; + case 1: + return red1; + case 2: + return (6 * red0 + 1 * red1) / 7; + case 3: + return (5 * red0 + 2 * red1) / 7; + case 4: + return (4 * red0 + 3 * red1) / 7; + case 5: + return (3 * red0 + 4 * red1) / 7; + case 6: + return (2 * red0 + 5 * red1) / 7; + case 7: + return (1 * red0 + 6 * red1) / 7; + } + } else { + switch (code) { + case 0: + return red0; + case 1: + return red1; + case 2: + return (4 * red0 + 1 * red1) / 5; + case 3: + return (3 * red0 + 2 * red1) / 5; + case 4: + return (2 * red0 + 3 * red1) / 5; + case 5: + return (1 * red0 + 4 * red1) / 5; + case 6: + return 0; + case 7: + return 0xff; + } + } + return 0; +} + +void main() { + uvec2 packed_bits = imageLoad(bc4_input, ivec3(gl_WorkGroupID + src_offset)).rg; + uint64_t bits = packUint2x32(packed_bits); + uint red = DecompressBlock(bits, gl_LocalInvocationID.xy); + uvec4 color = uvec4(red & 0xff, 0, 0, 0xff); + imageStore(bc4_output, ivec3(gl_GlobalInvocationID + dst_offset), color); +} diff --git a/src/video_core/host_shaders/opengl_present.frag b/src/video_core/host_shaders/opengl_present.frag index 8a4cb024b..84b818227 100644 --- a/src/video_core/host_shaders/opengl_present.frag +++ b/src/video_core/host_shaders/opengl_present.frag @@ -1,3 +1,7 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + #version 430 core layout (location = 0) in vec2 frag_tex_coord; diff --git a/src/video_core/host_shaders/opengl_present.vert b/src/video_core/host_shaders/opengl_present.vert index 2235d31a4..c3b5adbba 100644 --- a/src/video_core/host_shaders/opengl_present.vert +++ b/src/video_core/host_shaders/opengl_present.vert @@ -1,3 +1,7 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + #version 430 core out gl_PerVertex { diff --git a/src/video_core/host_shaders/pitch_unswizzle.comp b/src/video_core/host_shaders/pitch_unswizzle.comp new file mode 100644 index 000000000..cb48ec170 --- /dev/null +++ b/src/video_core/host_shaders/pitch_unswizzle.comp @@ -0,0 +1,86 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 430 + +#ifdef VULKAN + +#extension GL_EXT_shader_16bit_storage : require +#extension GL_EXT_shader_8bit_storage : require +#define HAS_EXTENDED_TYPES 1 +#define BEGIN_PUSH_CONSTANTS layout(push_constant) uniform PushConstants { +#define END_PUSH_CONSTANTS }; +#define UNIFORM(n) +#define BINDING_INPUT_BUFFER 0 +#define BINDING_OUTPUT_IMAGE 1 + +#else // ^^^ Vulkan ^^^ // vvv OpenGL vvv + +#extension GL_NV_gpu_shader5 : enable +#ifdef GL_NV_gpu_shader5 +#define HAS_EXTENDED_TYPES 1 +#else +#define HAS_EXTENDED_TYPES 0 +#endif +#define BEGIN_PUSH_CONSTANTS +#define END_PUSH_CONSTANTS +#define UNIFORM(n) layout (location = n) uniform +#define BINDING_INPUT_BUFFER 0 +#define BINDING_OUTPUT_IMAGE 0 + +#endif + +BEGIN_PUSH_CONSTANTS +UNIFORM(0) uvec2 origin; +UNIFORM(1) ivec2 destination; +UNIFORM(2) uint bytes_per_block; +UNIFORM(3) uint pitch; +END_PUSH_CONSTANTS + +#if HAS_EXTENDED_TYPES +layout(binding = BINDING_INPUT_BUFFER, std430) readonly buffer InputBufferU8 { uint8_t u8data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) readonly buffer InputBufferU16 { uint16_t u16data[]; }; +#endif +layout(binding = BINDING_INPUT_BUFFER, std430) readonly buffer InputBufferU32 { uint u32data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) readonly buffer InputBufferU64 { uvec2 u64data[]; }; +layout(binding = BINDING_INPUT_BUFFER, std430) readonly buffer InputBufferU128 { uvec4 u128data[]; }; + +layout(binding = BINDING_OUTPUT_IMAGE) writeonly uniform uimage2D output_image; + +layout(local_size_x = 32, local_size_y = 32, local_size_z = 1) in; + +uvec4 ReadTexel(uint offset) { + switch (bytes_per_block) { +#if HAS_EXTENDED_TYPES + case 1: + return uvec4(u8data[offset], 0, 0, 0); + case 2: + return uvec4(u16data[offset / 2], 0, 0, 0); +#else + case 1: + return uvec4(bitfieldExtract(u32data[offset / 4], int((offset * 8) & 24), 8), 0, 0, 0); + case 2: + return uvec4(bitfieldExtract(u32data[offset / 4], int((offset * 8) & 16), 16), 0, 0, 0); +#endif + case 4: + return uvec4(u32data[offset / 4], 0, 0, 0); + case 8: + return uvec4(u64data[offset / 8], 0, 0); + case 16: + return u128data[offset / 16]; + } + return uvec4(0); +} + +void main() { + uvec2 pos = gl_GlobalInvocationID.xy + origin; + + uint offset = 0; + offset += pos.x * bytes_per_block; + offset += pos.y * pitch; + + const uvec4 texel = ReadTexel(offset); + const ivec2 coord = ivec2(gl_GlobalInvocationID.xy) + destination; + imageStore(output_image, coord, texel); +} diff --git a/src/video_core/host_shaders/vulkan_blit_color_float.frag b/src/video_core/host_shaders/vulkan_blit_color_float.frag new file mode 100644 index 000000000..4a6aae410 --- /dev/null +++ b/src/video_core/host_shaders/vulkan_blit_color_float.frag @@ -0,0 +1,14 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 + +layout(binding = 0) uniform sampler2D tex; + +layout(location = 0) in vec2 texcoord; +layout(location = 0) out vec4 color; + +void main() { + color = textureLod(tex, texcoord, 0); +} diff --git a/src/video_core/host_shaders/vulkan_blit_depth_stencil.frag b/src/video_core/host_shaders/vulkan_blit_depth_stencil.frag new file mode 100644 index 000000000..19bb23a5a --- /dev/null +++ b/src/video_core/host_shaders/vulkan_blit_depth_stencil.frag @@ -0,0 +1,16 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#version 450 +#extension GL_ARB_shader_stencil_export : require + +layout(binding = 0) uniform sampler2D depth_tex; +layout(binding = 1) uniform isampler2D stencil_tex; + +layout(location = 0) in vec2 texcoord; + +void main() { + gl_FragDepth = textureLod(depth_tex, texcoord, 0).r; + gl_FragStencilRefARB = textureLod(stencil_tex, texcoord, 0).r; +} diff --git a/src/video_core/renderer_vulkan/shaders/blit.frag b/src/video_core/host_shaders/vulkan_present.frag index a06ecd24a..0979ff3e6 100644 --- a/src/video_core/renderer_vulkan/shaders/blit.frag +++ b/src/video_core/host_shaders/vulkan_present.frag @@ -2,15 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -/* - * Build instructions: - * $ glslangValidator -V $THIS_FILE -o output.spv - * $ spirv-opt -O --strip-debug output.spv -o optimized.spv - * $ xxd -i optimized.spv - * - * Then copy that bytecode to the C++ file - */ - #version 460 core layout (location = 0) in vec2 frag_tex_coord; diff --git a/src/video_core/renderer_vulkan/shaders/blit.vert b/src/video_core/host_shaders/vulkan_present.vert index c64d9235a..00b868958 100644 --- a/src/video_core/renderer_vulkan/shaders/blit.vert +++ b/src/video_core/host_shaders/vulkan_present.vert @@ -2,15 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -/* - * Build instructions: - * $ glslangValidator -V $THIS_FILE -o output.spv - * $ spirv-opt -O --strip-debug output.spv -o optimized.spv - * $ xxd -i optimized.spv - * - * Then copy that bytecode to the C++ file - */ - #version 460 core layout (location = 0) in vec2 vert_position; diff --git a/src/video_core/renderer_vulkan/shaders/quad_array.comp b/src/video_core/host_shaders/vulkan_quad_array.comp index 5a5703308..212f4e998 100644 --- a/src/video_core/renderer_vulkan/shaders/quad_array.comp +++ b/src/video_core/host_shaders/vulkan_quad_array.comp @@ -2,15 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -/* - * Build instructions: - * $ glslangValidator -V $THIS_FILE -o output.spv - * $ spirv-opt -O --strip-debug output.spv -o optimized.spv - * $ xxd -i optimized.spv - * - * Then copy that bytecode to the C++ file - */ - #version 460 core layout (local_size_x = 1024) in; diff --git a/src/video_core/renderer_vulkan/shaders/quad_indexed.comp b/src/video_core/host_shaders/vulkan_quad_indexed.comp index 5a472ba9b..8655591d0 100644 --- a/src/video_core/renderer_vulkan/shaders/quad_indexed.comp +++ b/src/video_core/host_shaders/vulkan_quad_indexed.comp @@ -2,15 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -/* - * Build instructions: - * $ glslangValidator -V quad_indexed.comp -o output.spv - * $ spirv-opt -O --strip-debug output.spv -o optimized.spv - * $ xxd -i optimized.spv - * - * Then copy that bytecode to the C++ file - */ - #version 460 core layout (local_size_x = 1024) in; diff --git a/src/video_core/renderer_vulkan/shaders/uint8.comp b/src/video_core/host_shaders/vulkan_uint8.comp index a320f3ae0..ad74d7af9 100644 --- a/src/video_core/renderer_vulkan/shaders/uint8.comp +++ b/src/video_core/host_shaders/vulkan_uint8.comp @@ -2,15 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -/* - * Build instructions: - * $ glslangValidator -V $THIS_FILE -o output.spv - * $ spirv-opt -O --strip-debug output.spv -o optimized.spv - * $ xxd -i optimized.spv - * - * Then copy that bytecode to the C++ file - */ - #version 460 core #extension GL_EXT_shader_16bit_storage : require #extension GL_EXT_shader_8bit_storage : require diff --git a/src/video_core/macro/macro_hle.cpp b/src/video_core/macro/macro_hle.cpp index df00b57df..70ac7c620 100644 --- a/src/video_core/macro/macro_hle.cpp +++ b/src/video_core/macro/macro_hle.cpp @@ -85,7 +85,7 @@ constexpr std::array<std::pair<u64, HLEFunction>, 3> hle_funcs{{ {0x0217920100488FF7, &HLE_0217920100488FF7}, }}; -HLEMacro::HLEMacro(Engines::Maxwell3D& maxwell3d) : maxwell3d(maxwell3d) {} +HLEMacro::HLEMacro(Engines::Maxwell3D& maxwell3d_) : maxwell3d{maxwell3d_} {} HLEMacro::~HLEMacro() = default; std::optional<std::unique_ptr<CachedMacro>> HLEMacro::GetHLEProgram(u64 hash) const { @@ -99,8 +99,8 @@ std::optional<std::unique_ptr<CachedMacro>> HLEMacro::GetHLEProgram(u64 hash) co HLEMacroImpl::~HLEMacroImpl() = default; -HLEMacroImpl::HLEMacroImpl(Engines::Maxwell3D& maxwell3d, HLEFunction func) - : maxwell3d(maxwell3d), func(func) {} +HLEMacroImpl::HLEMacroImpl(Engines::Maxwell3D& maxwell3d_, HLEFunction func_) + : maxwell3d{maxwell3d_}, func{func_} {} void HLEMacroImpl::Execute(const std::vector<u32>& parameters, u32 method) { func(maxwell3d, parameters); diff --git a/src/video_core/macro/macro_hle.h b/src/video_core/macro/macro_hle.h index 37af875a0..cb3bd1600 100644 --- a/src/video_core/macro/macro_hle.h +++ b/src/video_core/macro/macro_hle.h @@ -20,7 +20,7 @@ using HLEFunction = void (*)(Engines::Maxwell3D& maxwell3d, const std::vector<u3 class HLEMacro { public: - explicit HLEMacro(Engines::Maxwell3D& maxwell3d); + explicit HLEMacro(Engines::Maxwell3D& maxwell3d_); ~HLEMacro(); std::optional<std::unique_ptr<CachedMacro>> GetHLEProgram(u64 hash) const; diff --git a/src/video_core/macro/macro_interpreter.cpp b/src/video_core/macro/macro_interpreter.cpp index bd01fd1f2..8da26fd59 100644 --- a/src/video_core/macro/macro_interpreter.cpp +++ b/src/video_core/macro/macro_interpreter.cpp @@ -11,29 +11,29 @@ MICROPROFILE_DEFINE(MacroInterp, "GPU", "Execute macro interpreter", MP_RGB(128, 128, 192)); namespace Tegra { -MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d) - : MacroEngine::MacroEngine(maxwell3d), maxwell3d(maxwell3d) {} +MacroInterpreter::MacroInterpreter(Engines::Maxwell3D& maxwell3d_) + : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {} std::unique_ptr<CachedMacro> MacroInterpreter::Compile(const std::vector<u32>& code) { return std::make_unique<MacroInterpreterImpl>(maxwell3d, code); } -MacroInterpreterImpl::MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d, - const std::vector<u32>& code) - : maxwell3d(maxwell3d), code(code) {} +MacroInterpreterImpl::MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d_, + const std::vector<u32>& code_) + : maxwell3d{maxwell3d_}, code{code_} {} -void MacroInterpreterImpl::Execute(const std::vector<u32>& parameters, u32 method) { +void MacroInterpreterImpl::Execute(const std::vector<u32>& params, u32 method) { MICROPROFILE_SCOPE(MacroInterp); Reset(); - registers[1] = parameters[0]; - num_parameters = parameters.size(); + registers[1] = params[0]; + num_parameters = params.size(); if (num_parameters > parameters_capacity) { parameters_capacity = num_parameters; - this->parameters = std::make_unique<u32[]>(num_parameters); + parameters = std::make_unique<u32[]>(num_parameters); } - std::memcpy(this->parameters.get(), parameters.data(), num_parameters * sizeof(u32)); + std::memcpy(parameters.get(), params.data(), num_parameters * sizeof(u32)); // Execute the code until we hit an exit condition. bool keep_executing = true; @@ -133,8 +133,7 @@ bool MacroInterpreterImpl::Step(bool is_delay_slot) { break; } default: - UNIMPLEMENTED_MSG("Unimplemented macro operation {}", - static_cast<u32>(opcode.operation.Value())); + UNIMPLEMENTED_MSG("Unimplemented macro operation {}", opcode.operation.Value()); } // An instruction with the Exit flag will not actually @@ -182,7 +181,7 @@ u32 MacroInterpreterImpl::GetALUResult(Macro::ALUOperation operation, u32 src_a, return ~(src_a & src_b); default: - UNIMPLEMENTED_MSG("Unimplemented ALU operation {}", static_cast<u32>(operation)); + UNIMPLEMENTED_MSG("Unimplemented ALU operation {}", operation); return 0; } } @@ -230,7 +229,7 @@ void MacroInterpreterImpl::ProcessResult(Macro::ResultOperation operation, u32 r Send((result >> 12) & 0b111111); break; default: - UNIMPLEMENTED_MSG("Unimplemented result operation {}", static_cast<u32>(operation)); + UNIMPLEMENTED_MSG("Unimplemented result operation {}", operation); } } diff --git a/src/video_core/macro/macro_interpreter.h b/src/video_core/macro/macro_interpreter.h index 90217fc89..d50c619ce 100644 --- a/src/video_core/macro/macro_interpreter.h +++ b/src/video_core/macro/macro_interpreter.h @@ -17,7 +17,7 @@ class Maxwell3D; class MacroInterpreter final : public MacroEngine { public: - explicit MacroInterpreter(Engines::Maxwell3D& maxwell3d); + explicit MacroInterpreter(Engines::Maxwell3D& maxwell3d_); protected: std::unique_ptr<CachedMacro> Compile(const std::vector<u32>& code) override; @@ -28,8 +28,8 @@ private: class MacroInterpreterImpl : public CachedMacro { public: - MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& code); - void Execute(const std::vector<u32>& parameters, u32 method) override; + explicit MacroInterpreterImpl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_); + void Execute(const std::vector<u32>& params, u32 method) override; private: /// Resets the execution engine state, zeroing registers, etc. @@ -38,9 +38,9 @@ private: /** * Executes a single macro instruction located at the current program counter. Returns whether * the interpreter should keep running. - * @param offset Offset to start execution at. + * * @param is_delay_slot Whether the current step is being executed due to a delay slot in a - * previous instruction. + * previous instruction. */ bool Step(bool is_delay_slot); diff --git a/src/video_core/macro/macro_jit_x64.cpp b/src/video_core/macro/macro_jit_x64.cpp index 954b87515..c6b2b2109 100644 --- a/src/video_core/macro/macro_jit_x64.cpp +++ b/src/video_core/macro/macro_jit_x64.cpp @@ -28,15 +28,15 @@ static const std::bitset<32> PERSISTENT_REGISTERS = Common::X64::BuildRegSet({ BRANCH_HOLDER, }); -MacroJITx64::MacroJITx64(Engines::Maxwell3D& maxwell3d) - : MacroEngine::MacroEngine(maxwell3d), maxwell3d(maxwell3d) {} +MacroJITx64::MacroJITx64(Engines::Maxwell3D& maxwell3d_) + : MacroEngine{maxwell3d_}, maxwell3d{maxwell3d_} {} std::unique_ptr<CachedMacro> MacroJITx64::Compile(const std::vector<u32>& code) { return std::make_unique<MacroJITx64Impl>(maxwell3d, code); } -MacroJITx64Impl::MacroJITx64Impl(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& code) - : Xbyak::CodeGenerator(MAX_CODE_SIZE), code(code), maxwell3d(maxwell3d) { +MacroJITx64Impl::MacroJITx64Impl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_) + : CodeGenerator{MAX_CODE_SIZE}, code{code_}, maxwell3d{maxwell3d_} { Compile(); } @@ -165,8 +165,7 @@ void MacroJITx64Impl::Compile_ALU(Macro::Opcode opcode) { } break; default: - UNIMPLEMENTED_MSG("Unimplemented ALU operation {}", - static_cast<std::size_t>(opcode.alu_operation.Value())); + UNIMPLEMENTED_MSG("Unimplemented ALU operation {}", opcode.alu_operation.Value()); break; } Compile_ProcessResult(opcode.result_operation, opcode.dst); @@ -553,15 +552,15 @@ Xbyak::Reg32 MacroJITx64Impl::Compile_GetRegister(u32 index, Xbyak::Reg32 dst) { } void MacroJITx64Impl::Compile_ProcessResult(Macro::ResultOperation operation, u32 reg) { - const auto SetRegister = [this](u32 reg, const Xbyak::Reg32& result) { + const auto SetRegister = [this](u32 reg_index, const Xbyak::Reg32& result) { // Register 0 is supposed to always return 0. NOP is implemented as a store to the zero // register. - if (reg == 0) { + if (reg_index == 0) { return; } - mov(dword[STATE + offsetof(JITState, registers) + reg * sizeof(u32)], result); + mov(dword[STATE + offsetof(JITState, registers) + reg_index * sizeof(u32)], result); }; - const auto SetMethodAddress = [this](const Xbyak::Reg32& reg) { mov(METHOD_ADDRESS, reg); }; + const auto SetMethodAddress = [this](const Xbyak::Reg32& reg32) { mov(METHOD_ADDRESS, reg32); }; switch (operation) { case Macro::ResultOperation::IgnoreAndFetch: @@ -604,7 +603,7 @@ void MacroJITx64Impl::Compile_ProcessResult(Macro::ResultOperation operation, u3 Compile_Send(RESULT); break; default: - UNIMPLEMENTED_MSG("Unimplemented macro operation {}", static_cast<std::size_t>(operation)); + UNIMPLEMENTED_MSG("Unimplemented macro operation {}", operation); } } diff --git a/src/video_core/macro/macro_jit_x64.h b/src/video_core/macro/macro_jit_x64.h index a180e7428..7f50ac2f8 100644 --- a/src/video_core/macro/macro_jit_x64.h +++ b/src/video_core/macro/macro_jit_x64.h @@ -23,7 +23,7 @@ constexpr size_t MAX_CODE_SIZE = 0x10000; class MacroJITx64 final : public MacroEngine { public: - explicit MacroJITx64(Engines::Maxwell3D& maxwell3d); + explicit MacroJITx64(Engines::Maxwell3D& maxwell3d_); protected: std::unique_ptr<CachedMacro> Compile(const std::vector<u32>& code) override; @@ -34,7 +34,7 @@ private: class MacroJITx64Impl : public Xbyak::CodeGenerator, public CachedMacro { public: - MacroJITx64Impl(Engines::Maxwell3D& maxwell3d, const std::vector<u32>& code); + explicit MacroJITx64Impl(Engines::Maxwell3D& maxwell3d_, const std::vector<u32>& code_); ~MacroJITx64Impl(); void Execute(const std::vector<u32>& parameters, u32 method) override; diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index 02cf53d15..65feff588 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -11,6 +11,7 @@ #include "video_core/gpu.h" #include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" +#include "video_core/renderer_base.h" namespace Tegra { @@ -44,13 +45,22 @@ GPUVAddr MemoryManager::MapAllocate(VAddr cpu_addr, std::size_t size, std::size_ return Map(cpu_addr, *FindFreeRange(size, align), size); } +GPUVAddr MemoryManager::MapAllocate32(VAddr cpu_addr, std::size_t size) { + const std::optional<GPUVAddr> gpu_addr = FindFreeRange(size, 1, true); + ASSERT(gpu_addr); + return Map(cpu_addr, *gpu_addr, size); +} + void MemoryManager::Unmap(GPUVAddr gpu_addr, std::size_t size) { if (!size) { return; } // Flush and invalidate through the GPU interface, to be asynchronous if possible. - system.GPU().FlushAndInvalidateRegion(*GpuToCpuAddress(gpu_addr), size); + const std::optional<VAddr> cpu_addr = GpuToCpuAddress(gpu_addr); + ASSERT(cpu_addr); + + rasterizer->UnmapMemory(*cpu_addr, size); UpdateRange(gpu_addr, PageEntry::State::Unmapped, size); } @@ -108,7 +118,8 @@ void MemoryManager::SetPageEntry(GPUVAddr gpu_addr, PageEntry page_entry, std::s page_table[PageEntryIndex(gpu_addr)] = page_entry; } -std::optional<GPUVAddr> MemoryManager::FindFreeRange(std::size_t size, std::size_t align) const { +std::optional<GPUVAddr> MemoryManager::FindFreeRange(std::size_t size, std::size_t align, + bool start_32bit_address) const { if (!align) { align = page_size; } else { @@ -116,7 +127,7 @@ std::optional<GPUVAddr> MemoryManager::FindFreeRange(std::size_t size, std::size } u64 available_size{}; - GPUVAddr gpu_addr{address_space_start}; + GPUVAddr gpu_addr{start_32bit_address ? address_space_start_low : address_space_start}; while (gpu_addr + available_size < address_space_size) { if (GetPageEntry(gpu_addr + available_size).IsUnmapped()) { available_size += page_size; diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index 53c8d122a..c35e57689 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -28,7 +28,7 @@ public: }; constexpr PageEntry() = default; - constexpr PageEntry(State state) : state{state} {} + constexpr PageEntry(State state_) : state{state_} {} constexpr PageEntry(VAddr addr) : state{static_cast<State>(addr >> ShiftBits)} {} [[nodiscard]] constexpr bool IsUnmapped() const { @@ -68,7 +68,7 @@ static_assert(sizeof(PageEntry) == 4, "PageEntry is too large"); class MemoryManager final { public: - explicit MemoryManager(Core::System& system); + explicit MemoryManager(Core::System& system_); ~MemoryManager(); /// Binds a renderer to the memory manager. @@ -116,6 +116,7 @@ public: [[nodiscard]] GPUVAddr Map(VAddr cpu_addr, GPUVAddr gpu_addr, std::size_t size); [[nodiscard]] GPUVAddr MapAllocate(VAddr cpu_addr, std::size_t size, std::size_t align); + [[nodiscard]] GPUVAddr MapAllocate32(VAddr cpu_addr, std::size_t size); [[nodiscard]] std::optional<GPUVAddr> AllocateFixed(GPUVAddr gpu_addr, std::size_t size); [[nodiscard]] GPUVAddr Allocate(std::size_t size, std::size_t align); void Unmap(GPUVAddr gpu_addr, std::size_t size); @@ -124,7 +125,8 @@ private: [[nodiscard]] PageEntry GetPageEntry(GPUVAddr gpu_addr) const; void SetPageEntry(GPUVAddr gpu_addr, PageEntry page_entry, std::size_t size = page_size); GPUVAddr UpdateRange(GPUVAddr gpu_addr, PageEntry page_entry, std::size_t size); - [[nodiscard]] std::optional<GPUVAddr> FindFreeRange(std::size_t size, std::size_t align) const; + [[nodiscard]] std::optional<GPUVAddr> FindFreeRange(std::size_t size, std::size_t align, + bool start_32bit_address = false) const; void TryLockPage(PageEntry page_entry, std::size_t size); void TryUnlockPage(PageEntry page_entry, std::size_t size); @@ -135,6 +137,7 @@ private: static constexpr u64 address_space_size = 1ULL << 40; static constexpr u64 address_space_start = 1ULL << 32; + static constexpr u64 address_space_start_low = 1ULL << 16; static constexpr u64 page_bits{16}; static constexpr u64 page_size{1 << page_bits}; static constexpr u64 page_mask{page_size - 1}; diff --git a/src/video_core/morton.cpp b/src/video_core/morton.cpp index 9da9fb4ff..e69de29bb 100644 --- a/src/video_core/morton.cpp +++ b/src/video_core/morton.cpp @@ -1,250 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <array> -#include <cstring> -#include "common/assert.h" -#include "common/common_types.h" -#include "video_core/morton.h" -#include "video_core/surface.h" -#include "video_core/textures/decoders.h" - -namespace VideoCore { - -using Surface::GetBytesPerPixel; -using Surface::PixelFormat; - -using MortonCopyFn = void (*)(u32, u32, u32, u32, u32, u32, u8*, u8*); -using ConversionArray = std::array<MortonCopyFn, Surface::MaxPixelFormat>; - -template <bool morton_to_linear, PixelFormat format> -static void MortonCopy(u32 stride, u32 block_height, u32 height, u32 block_depth, u32 depth, - u32 tile_width_spacing, u8* buffer, u8* addr) { - constexpr u32 bytes_per_pixel = GetBytesPerPixel(format); - - // With the BCn formats (DXT and DXN), each 4x4 tile is swizzled instead of just individual - // pixel values. - constexpr u32 tile_size_x{GetDefaultBlockWidth(format)}; - constexpr u32 tile_size_y{GetDefaultBlockHeight(format)}; - - if constexpr (morton_to_linear) { - Tegra::Texture::UnswizzleTexture(buffer, addr, tile_size_x, tile_size_y, bytes_per_pixel, - stride, height, depth, block_height, block_depth, - tile_width_spacing); - } else { - Tegra::Texture::CopySwizzledData((stride + tile_size_x - 1) / tile_size_x, - (height + tile_size_y - 1) / tile_size_y, depth, - bytes_per_pixel, bytes_per_pixel, addr, buffer, false, - block_height, block_depth, tile_width_spacing); - } -} - -static constexpr ConversionArray morton_to_linear_fns = { - MortonCopy<true, PixelFormat::A8B8G8R8_UNORM>, - MortonCopy<true, PixelFormat::A8B8G8R8_SNORM>, - MortonCopy<true, PixelFormat::A8B8G8R8_SINT>, - MortonCopy<true, PixelFormat::A8B8G8R8_UINT>, - MortonCopy<true, PixelFormat::R5G6B5_UNORM>, - MortonCopy<true, PixelFormat::B5G6R5_UNORM>, - MortonCopy<true, PixelFormat::A1R5G5B5_UNORM>, - MortonCopy<true, PixelFormat::A2B10G10R10_UNORM>, - MortonCopy<true, PixelFormat::A2B10G10R10_UINT>, - MortonCopy<true, PixelFormat::A1B5G5R5_UNORM>, - MortonCopy<true, PixelFormat::R8_UNORM>, - MortonCopy<true, PixelFormat::R8_SNORM>, - MortonCopy<true, PixelFormat::R8_SINT>, - MortonCopy<true, PixelFormat::R8_UINT>, - MortonCopy<true, PixelFormat::R16G16B16A16_FLOAT>, - MortonCopy<true, PixelFormat::R16G16B16A16_UNORM>, - MortonCopy<true, PixelFormat::R16G16B16A16_SNORM>, - MortonCopy<true, PixelFormat::R16G16B16A16_SINT>, - MortonCopy<true, PixelFormat::R16G16B16A16_UINT>, - MortonCopy<true, PixelFormat::B10G11R11_FLOAT>, - MortonCopy<true, PixelFormat::R32G32B32A32_UINT>, - MortonCopy<true, PixelFormat::BC1_RGBA_UNORM>, - MortonCopy<true, PixelFormat::BC2_UNORM>, - MortonCopy<true, PixelFormat::BC3_UNORM>, - MortonCopy<true, PixelFormat::BC4_UNORM>, - MortonCopy<true, PixelFormat::BC4_SNORM>, - MortonCopy<true, PixelFormat::BC5_UNORM>, - MortonCopy<true, PixelFormat::BC5_SNORM>, - MortonCopy<true, PixelFormat::BC7_UNORM>, - MortonCopy<true, PixelFormat::BC6H_UFLOAT>, - MortonCopy<true, PixelFormat::BC6H_SFLOAT>, - MortonCopy<true, PixelFormat::ASTC_2D_4X4_UNORM>, - MortonCopy<true, PixelFormat::B8G8R8A8_UNORM>, - MortonCopy<true, PixelFormat::R32G32B32A32_FLOAT>, - MortonCopy<true, PixelFormat::R32G32B32A32_SINT>, - MortonCopy<true, PixelFormat::R32G32_FLOAT>, - MortonCopy<true, PixelFormat::R32G32_SINT>, - MortonCopy<true, PixelFormat::R32_FLOAT>, - MortonCopy<true, PixelFormat::R16_FLOAT>, - MortonCopy<true, PixelFormat::R16_UNORM>, - MortonCopy<true, PixelFormat::R16_SNORM>, - MortonCopy<true, PixelFormat::R16_UINT>, - MortonCopy<true, PixelFormat::R16_SINT>, - MortonCopy<true, PixelFormat::R16G16_UNORM>, - MortonCopy<true, PixelFormat::R16G16_FLOAT>, - MortonCopy<true, PixelFormat::R16G16_UINT>, - MortonCopy<true, PixelFormat::R16G16_SINT>, - MortonCopy<true, PixelFormat::R16G16_SNORM>, - MortonCopy<true, PixelFormat::R32G32B32_FLOAT>, - MortonCopy<true, PixelFormat::A8B8G8R8_SRGB>, - MortonCopy<true, PixelFormat::R8G8_UNORM>, - MortonCopy<true, PixelFormat::R8G8_SNORM>, - MortonCopy<true, PixelFormat::R8G8_SINT>, - MortonCopy<true, PixelFormat::R8G8_UINT>, - MortonCopy<true, PixelFormat::R32G32_UINT>, - MortonCopy<true, PixelFormat::R16G16B16X16_FLOAT>, - MortonCopy<true, PixelFormat::R32_UINT>, - MortonCopy<true, PixelFormat::R32_SINT>, - MortonCopy<true, PixelFormat::ASTC_2D_8X8_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_8X5_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_5X4_UNORM>, - MortonCopy<true, PixelFormat::B8G8R8A8_SRGB>, - MortonCopy<true, PixelFormat::BC1_RGBA_SRGB>, - MortonCopy<true, PixelFormat::BC2_SRGB>, - MortonCopy<true, PixelFormat::BC3_SRGB>, - MortonCopy<true, PixelFormat::BC7_SRGB>, - MortonCopy<true, PixelFormat::A4B4G4R4_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_4X4_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_8X8_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_8X5_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_5X4_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_5X5_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_5X5_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_10X8_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_10X8_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_6X6_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_6X6_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_10X10_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_10X10_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_12X12_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_12X12_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_8X6_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_8X6_SRGB>, - MortonCopy<true, PixelFormat::ASTC_2D_6X5_UNORM>, - MortonCopy<true, PixelFormat::ASTC_2D_6X5_SRGB>, - MortonCopy<true, PixelFormat::E5B9G9R9_FLOAT>, - MortonCopy<true, PixelFormat::D32_FLOAT>, - MortonCopy<true, PixelFormat::D16_UNORM>, - MortonCopy<true, PixelFormat::D24_UNORM_S8_UINT>, - MortonCopy<true, PixelFormat::S8_UINT_D24_UNORM>, - MortonCopy<true, PixelFormat::D32_FLOAT_S8_UINT>, -}; - -static constexpr ConversionArray linear_to_morton_fns = { - MortonCopy<false, PixelFormat::A8B8G8R8_UNORM>, - MortonCopy<false, PixelFormat::A8B8G8R8_SNORM>, - MortonCopy<false, PixelFormat::A8B8G8R8_SINT>, - MortonCopy<false, PixelFormat::A8B8G8R8_UINT>, - MortonCopy<false, PixelFormat::R5G6B5_UNORM>, - MortonCopy<false, PixelFormat::B5G6R5_UNORM>, - MortonCopy<false, PixelFormat::A1R5G5B5_UNORM>, - MortonCopy<false, PixelFormat::A2B10G10R10_UNORM>, - MortonCopy<false, PixelFormat::A2B10G10R10_UINT>, - MortonCopy<false, PixelFormat::A1B5G5R5_UNORM>, - MortonCopy<false, PixelFormat::R8_UNORM>, - MortonCopy<false, PixelFormat::R8_SNORM>, - MortonCopy<false, PixelFormat::R8_SINT>, - MortonCopy<false, PixelFormat::R8_UINT>, - MortonCopy<false, PixelFormat::R16G16B16A16_FLOAT>, - MortonCopy<false, PixelFormat::R16G16B16A16_SNORM>, - MortonCopy<false, PixelFormat::R16G16B16A16_SINT>, - MortonCopy<false, PixelFormat::R16G16B16A16_UNORM>, - MortonCopy<false, PixelFormat::R16G16B16A16_UINT>, - MortonCopy<false, PixelFormat::B10G11R11_FLOAT>, - MortonCopy<false, PixelFormat::R32G32B32A32_UINT>, - MortonCopy<false, PixelFormat::BC1_RGBA_UNORM>, - MortonCopy<false, PixelFormat::BC2_UNORM>, - MortonCopy<false, PixelFormat::BC3_UNORM>, - MortonCopy<false, PixelFormat::BC4_UNORM>, - MortonCopy<false, PixelFormat::BC4_SNORM>, - MortonCopy<false, PixelFormat::BC5_UNORM>, - MortonCopy<false, PixelFormat::BC5_SNORM>, - MortonCopy<false, PixelFormat::BC7_UNORM>, - MortonCopy<false, PixelFormat::BC6H_UFLOAT>, - MortonCopy<false, PixelFormat::BC6H_SFLOAT>, - // TODO(Subv): Swizzling ASTC formats are not supported - nullptr, - MortonCopy<false, PixelFormat::B8G8R8A8_UNORM>, - MortonCopy<false, PixelFormat::R32G32B32A32_FLOAT>, - MortonCopy<false, PixelFormat::R32G32B32A32_SINT>, - MortonCopy<false, PixelFormat::R32G32_FLOAT>, - MortonCopy<false, PixelFormat::R32G32_SINT>, - MortonCopy<false, PixelFormat::R32_FLOAT>, - MortonCopy<false, PixelFormat::R16_FLOAT>, - MortonCopy<false, PixelFormat::R16_UNORM>, - MortonCopy<false, PixelFormat::R16_SNORM>, - MortonCopy<false, PixelFormat::R16_UINT>, - MortonCopy<false, PixelFormat::R16_SINT>, - MortonCopy<false, PixelFormat::R16G16_UNORM>, - MortonCopy<false, PixelFormat::R16G16_FLOAT>, - MortonCopy<false, PixelFormat::R16G16_UINT>, - MortonCopy<false, PixelFormat::R16G16_SINT>, - MortonCopy<false, PixelFormat::R16G16_SNORM>, - MortonCopy<false, PixelFormat::R32G32B32_FLOAT>, - MortonCopy<false, PixelFormat::A8B8G8R8_SRGB>, - MortonCopy<false, PixelFormat::R8G8_UNORM>, - MortonCopy<false, PixelFormat::R8G8_SNORM>, - MortonCopy<false, PixelFormat::R8G8_SINT>, - MortonCopy<false, PixelFormat::R8G8_UINT>, - MortonCopy<false, PixelFormat::R32G32_UINT>, - MortonCopy<false, PixelFormat::R16G16B16X16_FLOAT>, - MortonCopy<false, PixelFormat::R32_UINT>, - MortonCopy<false, PixelFormat::R32_SINT>, - nullptr, - nullptr, - nullptr, - MortonCopy<false, PixelFormat::B8G8R8A8_SRGB>, - MortonCopy<false, PixelFormat::BC1_RGBA_SRGB>, - MortonCopy<false, PixelFormat::BC2_SRGB>, - MortonCopy<false, PixelFormat::BC3_SRGB>, - MortonCopy<false, PixelFormat::BC7_SRGB>, - MortonCopy<false, PixelFormat::A4B4G4R4_UNORM>, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - nullptr, - MortonCopy<false, PixelFormat::E5B9G9R9_FLOAT>, - MortonCopy<false, PixelFormat::D32_FLOAT>, - MortonCopy<false, PixelFormat::D16_UNORM>, - MortonCopy<false, PixelFormat::D24_UNORM_S8_UINT>, - MortonCopy<false, PixelFormat::S8_UINT_D24_UNORM>, - MortonCopy<false, PixelFormat::D32_FLOAT_S8_UINT>, -}; - -static MortonCopyFn GetSwizzleFunction(MortonSwizzleMode mode, Surface::PixelFormat format) { - switch (mode) { - case MortonSwizzleMode::MortonToLinear: - return morton_to_linear_fns[static_cast<std::size_t>(format)]; - case MortonSwizzleMode::LinearToMorton: - return linear_to_morton_fns[static_cast<std::size_t>(format)]; - } - UNREACHABLE(); - return morton_to_linear_fns[static_cast<std::size_t>(format)]; -} - -void MortonSwizzle(MortonSwizzleMode mode, Surface::PixelFormat format, u32 stride, - u32 block_height, u32 height, u32 block_depth, u32 depth, u32 tile_width_spacing, - u8* buffer, u8* addr) { - GetSwizzleFunction(mode, format)(stride, block_height, height, block_depth, depth, - tile_width_spacing, buffer, addr); -} - -} // namespace VideoCore diff --git a/src/video_core/morton.h b/src/video_core/morton.h index b714a7e3f..e69de29bb 100644 --- a/src/video_core/morton.h +++ b/src/video_core/morton.h @@ -1,18 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "common/common_types.h" -#include "video_core/surface.h" - -namespace VideoCore { - -enum class MortonSwizzleMode { MortonToLinear, LinearToMorton }; - -void MortonSwizzle(MortonSwizzleMode mode, VideoCore::Surface::PixelFormat format, u32 stride, - u32 block_height, u32 height, u32 block_depth, u32 depth, u32 tile_width_spacing, - u8* buffer, u8* addr); - -} // namespace VideoCore diff --git a/src/video_core/query_cache.h b/src/video_core/query_cache.h index fc54ca0ef..203f2af05 100644 --- a/src/video_core/query_cache.h +++ b/src/video_core/query_cache.h @@ -28,8 +28,8 @@ namespace VideoCommon { template <class QueryCache, class HostCounter> class CounterStreamBase { public: - explicit CounterStreamBase(QueryCache& cache, VideoCore::QueryType type) - : cache{cache}, type{type} {} + explicit CounterStreamBase(QueryCache& cache_, VideoCore::QueryType type_) + : cache{cache_}, type{type_} {} /// Updates the state of the stream, enabling or disabling as needed. void Update(bool enabled) { @@ -334,8 +334,8 @@ private: template <class HostCounter> class CachedQueryBase { public: - explicit CachedQueryBase(VAddr cpu_addr, u8* host_ptr) - : cpu_addr{cpu_addr}, host_ptr{host_ptr} {} + explicit CachedQueryBase(VAddr cpu_addr_, u8* host_ptr_) + : cpu_addr{cpu_addr_}, host_ptr{host_ptr_} {} virtual ~CachedQueryBase() = default; CachedQueryBase(CachedQueryBase&&) noexcept = default; diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h index b3e0919f8..0cb0f387d 100644 --- a/src/video_core/rasterizer_interface.h +++ b/src/video_core/rasterizer_interface.h @@ -32,7 +32,7 @@ using DiskResourceLoadCallback = std::function<void(LoadCallbackStage, std::size class RasterizerInterface { public: - virtual ~RasterizerInterface() {} + virtual ~RasterizerInterface() = default; /// Dispatches a draw invocation virtual void Draw(bool is_indexed, bool is_instanced) = 0; @@ -76,6 +76,9 @@ public: /// Sync memory between guest and host. virtual void SyncGuestHost() = 0; + /// Unmap memory range + virtual void UnmapMemory(VAddr addr, u64 size) = 0; + /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory /// and invalidated virtual void FlushAndInvalidateRegion(VAddr addr, u64 size) = 0; @@ -83,6 +86,12 @@ public: /// Notify the host renderer to wait for previous primitive and compute operations. virtual void WaitForIdle() = 0; + /// Notify the host renderer to wait for reads and writes to render targets and flush caches. + virtual void FragmentBarrier() = 0; + + /// Notify the host renderer to make available previous render target writes. + virtual void TiledCacheBarrier() = 0; + /// Notify the rasterizer to send all written commands to the host GPU. virtual void FlushCommands() = 0; @@ -90,15 +99,15 @@ public: virtual void TickFrame() = 0; /// Attempt to use a faster method to perform a surface copy - virtual bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, - const Tegra::Engines::Fermi2D::Regs::Surface& dst, - const Tegra::Engines::Fermi2D::Config& copy_config) { + [[nodiscard]] virtual bool AccelerateSurfaceCopy( + const Tegra::Engines::Fermi2D::Surface& src, const Tegra::Engines::Fermi2D::Surface& dst, + const Tegra::Engines::Fermi2D::Config& copy_config) { return false; } /// Attempt to use a faster method to display the framebuffer to screen - virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr, - u32 pixel_stride) { + [[nodiscard]] virtual bool AccelerateDisplay(const Tegra::FramebufferConfig& config, + VAddr framebuffer_addr, u32 pixel_stride) { return false; } @@ -110,12 +119,12 @@ public: const DiskResourceLoadCallback& callback) {} /// Grant access to the Guest Driver Profile for recording/obtaining info on the guest driver. - GuestDriverProfile& AccessGuestDriverProfile() { + [[nodiscard]] GuestDriverProfile& AccessGuestDriverProfile() { return guest_driver_profile; } /// Grant access to the Guest Driver Profile for recording/obtaining info on the guest driver. - const GuestDriverProfile& AccessGuestDriverProfile() const { + [[nodiscard]] const GuestDriverProfile& AccessGuestDriverProfile() const { return guest_driver_profile; } diff --git a/src/video_core/renderer_base.h b/src/video_core/renderer_base.h index 5c650808b..51dde8eb5 100644 --- a/src/video_core/renderer_base.h +++ b/src/video_core/renderer_base.h @@ -38,7 +38,7 @@ public: virtual ~RendererBase(); /// Initialize the renderer - virtual bool Init() = 0; + [[nodiscard]] virtual bool Init() = 0; /// Shutdown the renderer virtual void ShutDown() = 0; @@ -49,43 +49,43 @@ public: // Getter/setter functions: // ------------------------ - f32 GetCurrentFPS() const { + [[nodiscard]] f32 GetCurrentFPS() const { return m_current_fps; } - int GetCurrentFrame() const { + [[nodiscard]] int GetCurrentFrame() const { return m_current_frame; } - RasterizerInterface& Rasterizer() { + [[nodiscard]] RasterizerInterface& Rasterizer() { return *rasterizer; } - const RasterizerInterface& Rasterizer() const { + [[nodiscard]] const RasterizerInterface& Rasterizer() const { return *rasterizer; } - Core::Frontend::GraphicsContext& Context() { + [[nodiscard]] Core::Frontend::GraphicsContext& Context() { return *context; } - const Core::Frontend::GraphicsContext& Context() const { + [[nodiscard]] const Core::Frontend::GraphicsContext& Context() const { return *context; } - Core::Frontend::EmuWindow& GetRenderWindow() { + [[nodiscard]] Core::Frontend::EmuWindow& GetRenderWindow() { return render_window; } - const Core::Frontend::EmuWindow& GetRenderWindow() const { + [[nodiscard]] const Core::Frontend::EmuWindow& GetRenderWindow() const { return render_window; } - RendererSettings& Settings() { + [[nodiscard]] RendererSettings& Settings() { return renderer_settings; } - const RendererSettings& Settings() const { + [[nodiscard]] const RendererSettings& Settings() const { return renderer_settings; } diff --git a/src/video_core/renderer_opengl/gl_arb_decompiler.cpp b/src/video_core/renderer_opengl/gl_arb_decompiler.cpp index b7e9ed2e9..3e4d88c30 100644 --- a/src/video_core/renderer_opengl/gl_arb_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_arb_decompiler.cpp @@ -39,8 +39,8 @@ using Operation = const OperationNode&; constexpr std::array INTERNAL_FLAG_NAMES = {"ZERO", "SIGN", "CARRY", "OVERFLOW"}; char Swizzle(std::size_t component) { - ASSERT(component < 4); - return component["xyzw"]; + static constexpr std::string_view SWIZZLE{"xyzw"}; + return SWIZZLE.at(component); } constexpr bool IsGenericAttribute(Attribute::Index index) { @@ -71,7 +71,7 @@ std::string_view GetInputFlags(PixelImap attribute) { case PixelImap::Unused: break; } - UNIMPLEMENTED_MSG("Unknown attribute usage index={}", static_cast<int>(attribute)); + UNIMPLEMENTED_MSG("Unknown attribute usage index={}", attribute); return {}; } @@ -123,7 +123,7 @@ std::string_view PrimitiveDescription(Tegra::Engines::Maxwell3D::Regs::Primitive case Tegra::Engines::Maxwell3D::Regs::PrimitiveTopology::TriangleStripAdjacency: return "TRIANGLES_ADJACENCY"; default: - UNIMPLEMENTED_MSG("topology={}", static_cast<int>(topology)); + UNIMPLEMENTED_MSG("topology={}", topology); return "POINTS"; } } @@ -137,7 +137,7 @@ std::string_view TopologyName(Tegra::Shader::OutputTopology topology) { case Tegra::Shader::OutputTopology::TriangleStrip: return "TRIANGLE_STRIP"; default: - UNIMPLEMENTED_MSG("Unknown output topology: {}", static_cast<u32>(topology)); + UNIMPLEMENTED_MSG("Unknown output topology: {}", topology); return "points"; } } @@ -187,8 +187,8 @@ std::string TextureType(const MetaTexture& meta) { class ARBDecompiler final { public: - explicit ARBDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, - ShaderType stage, std::string_view identifier); + explicit ARBDecompiler(const Device& device_, const ShaderIR& ir_, const Registry& registry_, + ShaderType stage_, std::string_view identifier); std::string Code() const { return shader_source; @@ -224,7 +224,7 @@ private: std::string Visit(const Node& node); - std::pair<std::string, std::size_t> BuildCoords(Operation); + std::tuple<std::string, std::string, std::size_t> BuildCoords(Operation); std::string BuildAoffi(Operation); std::string GlobalMemoryPointer(const GmemNode& gmem); void Exit(); @@ -376,9 +376,11 @@ private: std::string temporary = AllocTemporary(); std::string address; std::string_view opname; + bool robust = false; if (const auto gmem = std::get_if<GmemNode>(&*operation[0])) { address = GlobalMemoryPointer(*gmem); opname = "ATOM"; + robust = true; } else if (const auto smem = std::get_if<SmemNode>(&*operation[0])) { address = fmt::format("shared_mem[{}]", Visit(smem->GetAddress())); opname = "ATOMS"; @@ -386,7 +388,15 @@ private: UNREACHABLE(); return "{0, 0, 0, 0}"; } + if (robust) { + AddLine("IF NE.x;"); + } AddLine("{}.{}.{} {}, {}, {};", opname, op, type, temporary, Visit(operation[1]), address); + if (robust) { + AddLine("ELSE;"); + AddLine("MOV.S {}, 0;", temporary); + AddLine("ENDIF;"); + } return temporary; } @@ -792,9 +802,9 @@ private: }; }; -ARBDecompiler::ARBDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, - ShaderType stage, std::string_view identifier) - : device{device}, ir{ir}, registry{registry}, stage{stage} { +ARBDecompiler::ARBDecompiler(const Device& device_, const ShaderIR& ir_, const Registry& registry_, + ShaderType stage_, std::string_view identifier) + : device{device_}, ir{ir_}, registry{registry_}, stage{stage_} { DefineGlobalMemory(); AddLine("TEMP RC;"); @@ -980,10 +990,9 @@ void ARBDecompiler::DeclareLocalMemory() { } void ARBDecompiler::DeclareGlobalMemory() { - const std::size_t num_entries = ir.GetGlobalMemory().size(); + const size_t num_entries = ir.GetGlobalMemory().size(); if (num_entries > 0) { - const std::size_t num_vectors = Common::AlignUp(num_entries, 2) / 2; - AddLine("PARAM c[{}] = {{ program.local[0..{}] }};", num_vectors, num_vectors - 1); + AddLine("PARAM c[{}] = {{ program.local[0..{}] }};", num_entries, num_entries - 1); } } @@ -1125,44 +1134,44 @@ void ARBDecompiler::VisitAST(const ASTNode& node) { for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } - } else if (const auto ast = std::get_if<ASTIfThen>(&*node->GetInnerData())) { - const std::string condition = VisitExpression(ast->condition); + } else if (const auto if_then = std::get_if<ASTIfThen>(&*node->GetInnerData())) { + const std::string condition = VisitExpression(if_then->condition); ResetTemporaries(); AddLine("MOVC.U RC.x, {};", condition); AddLine("IF NE.x;"); - for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + for (ASTNode current = if_then->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } AddLine("ENDIF;"); - } else if (const auto ast = std::get_if<ASTIfElse>(&*node->GetInnerData())) { + } else if (const auto if_else = std::get_if<ASTIfElse>(&*node->GetInnerData())) { AddLine("ELSE;"); - for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + for (ASTNode current = if_else->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } - } else if (const auto ast = std::get_if<ASTBlockDecoded>(&*node->GetInnerData())) { - VisitBlock(ast->nodes); - } else if (const auto ast = std::get_if<ASTVarSet>(&*node->GetInnerData())) { - AddLine("MOV.U F{}, {};", ast->index, VisitExpression(ast->condition)); + } else if (const auto decoded = std::get_if<ASTBlockDecoded>(&*node->GetInnerData())) { + VisitBlock(decoded->nodes); + } else if (const auto var_set = std::get_if<ASTVarSet>(&*node->GetInnerData())) { + AddLine("MOV.U F{}, {};", var_set->index, VisitExpression(var_set->condition)); ResetTemporaries(); - } else if (const auto ast = std::get_if<ASTDoWhile>(&*node->GetInnerData())) { - const std::string condition = VisitExpression(ast->condition); + } else if (const auto do_while = std::get_if<ASTDoWhile>(&*node->GetInnerData())) { + const std::string condition = VisitExpression(do_while->condition); ResetTemporaries(); AddLine("REP;"); - for (ASTNode current = ast->nodes.GetFirst(); current; current = current->GetNext()) { + for (ASTNode current = do_while->nodes.GetFirst(); current; current = current->GetNext()) { VisitAST(current); } AddLine("MOVC.U RC.x, {};", condition); AddLine("BRK (NE.x);"); AddLine("ENDREP;"); - } else if (const auto ast = std::get_if<ASTReturn>(&*node->GetInnerData())) { - const bool is_true = ExprIsTrue(ast->condition); + } else if (const auto ast_return = std::get_if<ASTReturn>(&*node->GetInnerData())) { + const bool is_true = ExprIsTrue(ast_return->condition); if (!is_true) { - AddLine("MOVC.U RC.x, {};", VisitExpression(ast->condition)); + AddLine("MOVC.U RC.x, {};", VisitExpression(ast_return->condition)); AddLine("IF NE.x;"); ResetTemporaries(); } - if (ast->kills) { + if (ast_return->kills) { AddLine("KIL TR;"); } else { Exit(); @@ -1170,11 +1179,11 @@ void ARBDecompiler::VisitAST(const ASTNode& node) { if (!is_true) { AddLine("ENDIF;"); } - } else if (const auto ast = std::get_if<ASTBreak>(&*node->GetInnerData())) { - if (ExprIsTrue(ast->condition)) { + } else if (const auto ast_break = std::get_if<ASTBreak>(&*node->GetInnerData())) { + if (ExprIsTrue(ast_break->condition)) { AddLine("BRK;"); } else { - AddLine("MOVC.U RC.x, {};", VisitExpression(ast->condition)); + AddLine("MOVC.U RC.x, {};", VisitExpression(ast_break->condition)); AddLine("BRK (NE.x);"); ResetTemporaries(); } @@ -1342,7 +1351,7 @@ std::string ARBDecompiler::Visit(const Node& node) { GetGenericAttributeIndex(index), swizzle); } } - UNIMPLEMENTED_MSG("Unimplemented input attribute={}", static_cast<int>(index)); + UNIMPLEMENTED_MSG("Unimplemented input attribute={}", index); break; } return "{0, 0, 0, 0}.x"; @@ -1363,7 +1372,8 @@ std::string ARBDecompiler::Visit(const Node& node) { if (const auto gmem = std::get_if<GmemNode>(&*node)) { std::string temporary = AllocTemporary(); - AddLine("LOAD.U32 {}, {};", temporary, GlobalMemoryPointer(*gmem)); + AddLine("MOV {}, 0;", temporary); + AddLine("LOAD.U32 {} (NE.x), {};", temporary, GlobalMemoryPointer(*gmem)); return temporary; } @@ -1406,12 +1416,12 @@ std::string ARBDecompiler::Visit(const Node& node) { return {}; } -std::pair<std::string, std::size_t> ARBDecompiler::BuildCoords(Operation operation) { +std::tuple<std::string, std::string, std::size_t> ARBDecompiler::BuildCoords(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); UNIMPLEMENTED_IF(meta.sampler.is_indexed); - UNIMPLEMENTED_IF(meta.sampler.is_shadow && meta.sampler.is_array && - meta.sampler.type == Tegra::Shader::TextureType::TextureCube); + const bool is_extended = meta.sampler.is_shadow && meta.sampler.is_array && + meta.sampler.type == Tegra::Shader::TextureType::TextureCube; const std::size_t count = operation.GetOperandsCount(); std::string temporary = AllocVectorTemporary(); std::size_t i = 0; @@ -1419,12 +1429,21 @@ std::pair<std::string, std::size_t> ARBDecompiler::BuildCoords(Operation operati AddLine("MOV.F {}.{}, {};", temporary, Swizzle(i), Visit(operation[i])); } if (meta.sampler.is_array) { - AddLine("I2F.S {}.{}, {};", temporary, Swizzle(i++), Visit(meta.array)); + AddLine("I2F.S {}.{}, {};", temporary, Swizzle(i), Visit(meta.array)); + ++i; } if (meta.sampler.is_shadow) { - AddLine("MOV.F {}.{}, {};", temporary, Swizzle(i++), Visit(meta.depth_compare)); + std::string compare = Visit(meta.depth_compare); + if (is_extended) { + ASSERT(i == 4); + std::string extra_coord = AllocVectorTemporary(); + AddLine("MOV.F {}.x, {};", extra_coord, compare); + return {fmt::format("{}, {}", temporary, extra_coord), extra_coord, 0}; + } + AddLine("MOV.F {}.{}, {};", temporary, Swizzle(i), compare); + ++i; } - return {std::move(temporary), i}; + return {temporary, temporary, i}; } std::string ARBDecompiler::BuildAoffi(Operation operation) { @@ -1441,18 +1460,21 @@ std::string ARBDecompiler::BuildAoffi(Operation operation) { } std::string ARBDecompiler::GlobalMemoryPointer(const GmemNode& gmem) { + // Read a bindless SSBO, return its address and set CC accordingly + // address = c[binding].xy + // length = c[binding].z const u32 binding = global_memory_names.at(gmem.GetDescriptor()); - const char result_swizzle = binding % 2 == 0 ? 'x' : 'y'; const std::string pointer = AllocLongVectorTemporary(); std::string temporary = AllocTemporary(); - const u32 local_index = binding / 2; - AddLine("PK64.U {}, c[{}];", pointer, local_index); + AddLine("PK64.U {}, c[{}];", pointer, binding); AddLine("SUB.U {}, {}, {};", temporary, Visit(gmem.GetRealAddress()), Visit(gmem.GetBaseAddress())); AddLine("CVT.U64.U32 {}.z, {};", pointer, temporary); - AddLine("ADD.U64 {}.x, {}.{}, {}.z;", pointer, pointer, result_swizzle, pointer); + AddLine("ADD.U64 {}.x, {}.x, {}.z;", pointer, pointer, pointer); + // Compare offset to length and set CC + AddLine("SLT.U.CC RC.x, {}, c[{}].z;", temporary, binding); return fmt::format("{}.x", pointer); } @@ -1463,9 +1485,7 @@ void ARBDecompiler::Exit() { } const auto safe_get_register = [this](u32 reg) -> std::string { - // TODO(Rodrigo): Replace with contains once C++20 releases - const auto& used_registers = ir.GetRegisters(); - if (used_registers.find(reg) != used_registers.end()) { + if (ir.GetRegisters().contains(reg)) { return fmt::format("R{}.x", reg); } return "{0, 0, 0, 0}.x"; @@ -1552,7 +1572,9 @@ std::string ARBDecompiler::Assign(Operation operation) { ResetTemporaries(); return {}; } else if (const auto gmem = std::get_if<GmemNode>(&*dest)) { + AddLine("IF NE.x;"); AddLine("STORE.U32 {}, {};", Visit(src), GlobalMemoryPointer(*gmem)); + AddLine("ENDIF;"); ResetTemporaries(); return {}; } else { @@ -1844,7 +1866,7 @@ std::string ARBDecompiler::LogicalAddCarry(Operation operation) { std::string ARBDecompiler::Texture(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; - const auto [temporary, swizzle] = BuildCoords(operation); + const auto [coords, temporary, swizzle] = BuildCoords(operation); std::string_view opcode = "TEX"; std::string extra; @@ -1873,7 +1895,7 @@ std::string ARBDecompiler::Texture(Operation operation) { } } - AddLine("{}.F {}, {},{} texture[{}], {}{};", opcode, temporary, temporary, extra, sampler_id, + AddLine("{}.F {}, {},{} texture[{}], {}{};", opcode, temporary, coords, extra, sampler_id, TextureType(meta), BuildAoffi(operation)); AddLine("MOV.U {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); return fmt::format("{}.x", temporary); @@ -1882,7 +1904,7 @@ std::string ARBDecompiler::Texture(Operation operation) { std::string ARBDecompiler::TextureGather(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; - const auto [temporary, swizzle] = BuildCoords(operation); + const auto [coords, temporary, swizzle] = BuildCoords(operation); std::string comp; if (!meta.sampler.is_shadow) { @@ -1892,7 +1914,7 @@ std::string ARBDecompiler::TextureGather(Operation operation) { AddLine("TXG.F {}, {}, texture[{}]{}, {}{};", temporary, temporary, sampler_id, comp, TextureType(meta), BuildAoffi(operation)); - AddLine("MOV.U {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); + AddLine("MOV.U {}.x, {}.{};", temporary, coords, Swizzle(meta.element)); return fmt::format("{}.x", temporary); } @@ -1930,13 +1952,13 @@ std::string ARBDecompiler::TextureQueryLod(Operation operation) { std::string ARBDecompiler::TexelFetch(Operation operation) { const auto& meta = std::get<MetaTexture>(operation.GetMeta()); const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; - const auto [temporary, swizzle] = BuildCoords(operation); + const auto [coords, temporary, swizzle] = BuildCoords(operation); if (!meta.sampler.is_buffer) { ASSERT(swizzle < 4); AddLine("MOV.F {}.w, {};", temporary, Visit(meta.lod)); } - AddLine("TXF.F {}, {}, texture[{}], {}{};", temporary, temporary, sampler_id, TextureType(meta), + AddLine("TXF.F {}, {}, texture[{}], {}{};", temporary, coords, sampler_id, TextureType(meta), BuildAoffi(operation)); AddLine("MOV.U {}.x, {}.{};", temporary, temporary, Swizzle(meta.element)); return fmt::format("{}.x", temporary); @@ -1947,7 +1969,7 @@ std::string ARBDecompiler::TextureGradient(Operation operation) { const u32 sampler_id = device.GetBaseBindings(stage).sampler + meta.sampler.index; const std::string ddx = AllocVectorTemporary(); const std::string ddy = AllocVectorTemporary(); - const std::string coord = BuildCoords(operation).first; + const std::string coord = std::get<1>(BuildCoords(operation)); const std::size_t num_components = meta.derivates.size() / 2; for (std::size_t index = 0; index < num_components; ++index) { diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index b1c4cd62f..5772cad87 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -22,11 +22,11 @@ using Maxwell = Tegra::Engines::Maxwell3D::Regs; MICROPROFILE_DEFINE(OpenGL_Buffer_Download, "OpenGL", "Buffer Download", MP_RGB(192, 192, 128)); -Buffer::Buffer(const Device& device, VAddr cpu_addr, std::size_t size) - : VideoCommon::BufferBlock{cpu_addr, size} { +Buffer::Buffer(const Device& device_, VAddr cpu_addr_, std::size_t size_) + : BufferBlock{cpu_addr_, size_} { gl_buffer.Create(); - glNamedBufferData(gl_buffer.handle, static_cast<GLsizeiptr>(size), nullptr, GL_DYNAMIC_DRAW); - if (device.UseAssemblyShaders() || device.HasVertexBufferUnifiedMemory()) { + glNamedBufferData(gl_buffer.handle, static_cast<GLsizeiptr>(size_), nullptr, GL_DYNAMIC_DRAW); + if (device_.UseAssemblyShaders() || device_.HasVertexBufferUnifiedMemory()) { glMakeNamedBufferResidentNV(gl_buffer.handle, GL_READ_WRITE); glGetNamedBufferParameterui64vNV(gl_buffer.handle, GL_BUFFER_GPU_ADDRESS_NV, &gpu_address); } @@ -34,14 +34,14 @@ Buffer::Buffer(const Device& device, VAddr cpu_addr, std::size_t size) Buffer::~Buffer() = default; -void Buffer::Upload(std::size_t offset, std::size_t size, const u8* data) { - glNamedBufferSubData(Handle(), static_cast<GLintptr>(offset), static_cast<GLsizeiptr>(size), - data); +void Buffer::Upload(std::size_t offset, std::size_t data_size, const u8* data) { + glNamedBufferSubData(Handle(), static_cast<GLintptr>(offset), + static_cast<GLsizeiptr>(data_size), data); } -void Buffer::Download(std::size_t offset, std::size_t size, u8* data) { +void Buffer::Download(std::size_t offset, std::size_t data_size, u8* data) { MICROPROFILE_SCOPE(OpenGL_Buffer_Download); - const GLsizeiptr gl_size = static_cast<GLsizeiptr>(size); + const GLsizeiptr gl_size = static_cast<GLsizeiptr>(data_size); const GLintptr gl_offset = static_cast<GLintptr>(offset); if (read_buffer.handle == 0) { read_buffer.Create(); @@ -54,17 +54,16 @@ void Buffer::Download(std::size_t offset, std::size_t size, u8* data) { } void Buffer::CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, - std::size_t size) { + std::size_t copy_size) { glCopyNamedBufferSubData(src.Handle(), Handle(), static_cast<GLintptr>(src_offset), - static_cast<GLintptr>(dst_offset), static_cast<GLsizeiptr>(size)); + static_cast<GLintptr>(dst_offset), static_cast<GLsizeiptr>(copy_size)); } -OGLBufferCache::OGLBufferCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, - const Device& device_, std::size_t stream_size) - : GenericBufferCache{rasterizer, gpu_memory, cpu_memory, - std::make_unique<OGLStreamBuffer>(device_, stream_size, true)}, - device{device_} { +OGLBufferCache::OGLBufferCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::MemoryManager& gpu_memory_, Core::Memory::Memory& cpu_memory_, + const Device& device_, OGLStreamBuffer& stream_buffer_, + StateTracker& state_tracker) + : GenericBufferCache{rasterizer_, gpu_memory_, cpu_memory_, stream_buffer_}, device{device_} { if (!device.HasFastBufferSubData()) { return; } diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h index f75b32e31..17ee90316 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.h +++ b/src/video_core/renderer_opengl/gl_buffer_cache.h @@ -22,18 +22,19 @@ namespace OpenGL { class Device; class OGLStreamBuffer; class RasterizerOpenGL; +class StateTracker; class Buffer : public VideoCommon::BufferBlock { public: - explicit Buffer(const Device& device, VAddr cpu_addr, std::size_t size); + explicit Buffer(const Device& device_, VAddr cpu_addr_, std::size_t size_); ~Buffer(); - void Upload(std::size_t offset, std::size_t size, const u8* data); + void Upload(std::size_t offset, std::size_t data_size, const u8* data); - void Download(std::size_t offset, std::size_t size, u8* data); + void Download(std::size_t offset, std::size_t data_size, u8* data); void CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, - std::size_t size); + std::size_t copy_size); GLuint Handle() const noexcept { return gl_buffer.handle; @@ -54,7 +55,8 @@ class OGLBufferCache final : public GenericBufferCache { public: explicit OGLBufferCache(VideoCore::RasterizerInterface& rasterizer, Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, - const Device& device, std::size_t stream_size); + const Device& device, OGLStreamBuffer& stream_buffer, + StateTracker& state_tracker); ~OGLBufferCache(); BufferInfo GetEmptyBuffer(std::size_t) override; diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp index e7d95149f..81b71edfb 100644 --- a/src/video_core/renderer_opengl/gl_device.cpp +++ b/src/video_core/renderer_opengl/gl_device.cpp @@ -5,9 +5,11 @@ #include <algorithm> #include <array> #include <cstddef> +#include <cstdlib> #include <cstring> #include <limits> #include <optional> +#include <span> #include <vector> #include <glad/glad.h> @@ -27,27 +29,29 @@ constexpr u32 ReservedUniformBlocks = 1; constexpr u32 NumStages = 5; -constexpr std::array LimitUBOs = { +constexpr std::array LIMIT_UBOS = { GL_MAX_VERTEX_UNIFORM_BLOCKS, GL_MAX_TESS_CONTROL_UNIFORM_BLOCKS, GL_MAX_TESS_EVALUATION_UNIFORM_BLOCKS, GL_MAX_GEOMETRY_UNIFORM_BLOCKS, - GL_MAX_FRAGMENT_UNIFORM_BLOCKS, GL_MAX_COMPUTE_UNIFORM_BLOCKS}; - -constexpr std::array LimitSSBOs = { + GL_MAX_FRAGMENT_UNIFORM_BLOCKS, GL_MAX_COMPUTE_UNIFORM_BLOCKS, +}; +constexpr std::array LIMIT_SSBOS = { GL_MAX_VERTEX_SHADER_STORAGE_BLOCKS, GL_MAX_TESS_CONTROL_SHADER_STORAGE_BLOCKS, GL_MAX_TESS_EVALUATION_SHADER_STORAGE_BLOCKS, GL_MAX_GEOMETRY_SHADER_STORAGE_BLOCKS, - GL_MAX_FRAGMENT_SHADER_STORAGE_BLOCKS, GL_MAX_COMPUTE_SHADER_STORAGE_BLOCKS}; - -constexpr std::array LimitSamplers = {GL_MAX_VERTEX_TEXTURE_IMAGE_UNITS, - GL_MAX_TESS_CONTROL_TEXTURE_IMAGE_UNITS, - GL_MAX_TESS_EVALUATION_TEXTURE_IMAGE_UNITS, - GL_MAX_GEOMETRY_TEXTURE_IMAGE_UNITS, - GL_MAX_TEXTURE_IMAGE_UNITS, - GL_MAX_COMPUTE_TEXTURE_IMAGE_UNITS}; - -constexpr std::array LimitImages = { + GL_MAX_FRAGMENT_SHADER_STORAGE_BLOCKS, GL_MAX_COMPUTE_SHADER_STORAGE_BLOCKS, +}; +constexpr std::array LIMIT_SAMPLERS = { + GL_MAX_VERTEX_TEXTURE_IMAGE_UNITS, + GL_MAX_TESS_CONTROL_TEXTURE_IMAGE_UNITS, + GL_MAX_TESS_EVALUATION_TEXTURE_IMAGE_UNITS, + GL_MAX_GEOMETRY_TEXTURE_IMAGE_UNITS, + GL_MAX_TEXTURE_IMAGE_UNITS, + GL_MAX_COMPUTE_TEXTURE_IMAGE_UNITS, +}; +constexpr std::array LIMIT_IMAGES = { GL_MAX_VERTEX_IMAGE_UNIFORMS, GL_MAX_TESS_CONTROL_IMAGE_UNIFORMS, GL_MAX_TESS_EVALUATION_IMAGE_UNIFORMS, GL_MAX_GEOMETRY_IMAGE_UNIFORMS, - GL_MAX_FRAGMENT_IMAGE_UNIFORMS, GL_MAX_COMPUTE_IMAGE_UNIFORMS}; + GL_MAX_FRAGMENT_IMAGE_UNIFORMS, GL_MAX_COMPUTE_IMAGE_UNIFORMS, +}; template <typename T> T GetInteger(GLenum pname) { @@ -76,8 +80,8 @@ std::vector<std::string_view> GetExtensions() { return extensions; } -bool HasExtension(const std::vector<std::string_view>& images, std::string_view extension) { - return std::find(images.begin(), images.end(), extension) != images.end(); +bool HasExtension(std::span<const std::string_view> extensions, std::string_view extension) { + return std::ranges::find(extensions, extension) != extensions.end(); } u32 Extract(u32& base, u32& num, u32 amount, std::optional<GLenum> limit = {}) { @@ -91,8 +95,8 @@ u32 Extract(u32& base, u32& num, u32 amount, std::optional<GLenum> limit = {}) { std::array<u32, Tegra::Engines::MaxShaderTypes> BuildMaxUniformBuffers() noexcept { std::array<u32, Tegra::Engines::MaxShaderTypes> max; - std::transform(LimitUBOs.begin(), LimitUBOs.end(), max.begin(), - [](GLenum pname) { return GetInteger<u32>(pname); }); + std::ranges::transform(LIMIT_UBOS, max.begin(), + [](GLenum pname) { return GetInteger<u32>(pname); }); return max; } @@ -115,9 +119,10 @@ std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> BuildBaseBindin for (std::size_t i = 0; i < NumStages; ++i) { const std::size_t stage = stage_swizzle[i]; bindings[stage] = { - Extract(base_ubo, num_ubos, total_ubos / NumStages, LimitUBOs[stage]), - Extract(base_ssbo, num_ssbos, total_ssbos / NumStages, LimitSSBOs[stage]), - Extract(base_samplers, num_samplers, total_samplers / NumStages, LimitSamplers[stage])}; + Extract(base_ubo, num_ubos, total_ubos / NumStages, LIMIT_UBOS[stage]), + Extract(base_ssbo, num_ssbos, total_ssbos / NumStages, LIMIT_SSBOS[stage]), + Extract(base_samplers, num_samplers, total_samplers / NumStages, + LIMIT_SAMPLERS[stage])}; } u32 num_images = GetInteger<u32>(GL_MAX_IMAGE_UNITS); @@ -130,7 +135,7 @@ std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> BuildBaseBindin // Reserve at least 4 image bindings on the fragment stage. bindings[4].image = - Extract(base_images, num_images, std::max(4U, num_images / NumStages), LimitImages[4]); + Extract(base_images, num_images, std::max(4U, num_images / NumStages), LIMIT_IMAGES[4]); // This is guaranteed to be at least 1. const u32 total_extracted_images = num_images / (NumStages - 1); @@ -142,7 +147,7 @@ std::array<Device::BaseBindings, Tegra::Engines::MaxShaderTypes> BuildBaseBindin continue; } bindings[stage].image = - Extract(base_images, num_images, total_extracted_images, LimitImages[stage]); + Extract(base_images, num_images, total_extracted_images, LIMIT_IMAGES[stage]); } // Compute doesn't care about any of this. @@ -188,17 +193,22 @@ bool IsASTCSupported() { return true; } +[[nodiscard]] bool IsDebugToolAttached(std::span<const std::string_view> extensions) { + const bool nsight = std::getenv("NVTX_INJECTION64_PATH") || std::getenv("NSIGHT_LAUNCHED"); + return nsight || HasExtension(extensions, "GL_EXT_debug_tool"); +} + } // Anonymous namespace Device::Device() : max_uniform_buffers{BuildMaxUniformBuffers()}, base_bindings{BuildBaseBindings()} { const std::string_view vendor = reinterpret_cast<const char*>(glGetString(GL_VENDOR)); - const std::string_view renderer = reinterpret_cast<const char*>(glGetString(GL_RENDERER)); const std::string_view version = reinterpret_cast<const char*>(glGetString(GL_VERSION)); const std::vector extensions = GetExtensions(); const bool is_nvidia = vendor == "NVIDIA Corporation"; const bool is_amd = vendor == "ATI Technologies Inc."; + const bool is_intel = vendor == "Intel"; bool disable_fast_buffer_sub_data = false; if (is_nvidia && version == "4.6.0 NVIDIA 443.24") { @@ -207,9 +217,8 @@ Device::Device() "Beta driver 443.24 is known to have issues. There might be performance issues."); disable_fast_buffer_sub_data = true; } - - uniform_buffer_alignment = GetInteger<std::size_t>(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT); - shader_storage_alignment = GetInteger<std::size_t>(GL_SHADER_STORAGE_BUFFER_OFFSET_ALIGNMENT); + uniform_buffer_alignment = GetInteger<size_t>(GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT); + shader_storage_alignment = GetInteger<size_t>(GL_SHADER_STORAGE_BUFFER_OFFSET_ALIGNMENT); max_vertex_attributes = GetInteger<u32>(GL_MAX_VERTEX_ATTRIBS); max_varyings = GetInteger<u32>(GL_MAX_VARYING_VECTORS); max_compute_shared_memory_size = GetInteger<u32>(GL_MAX_COMPUTE_SHARED_MEMORY_SIZE); @@ -223,8 +232,10 @@ Device::Device() has_variable_aoffi = TestVariableAoffi(); has_component_indexing_bug = is_amd; has_precise_bug = TestPreciseBug(); + has_broken_texture_view_formats = is_amd || is_intel; has_nv_viewport_array2 = GLAD_GL_NV_viewport_array2; has_vertex_buffer_unified_memory = GLAD_GL_NV_vertex_buffer_unified_memory; + has_debugging_tool_attached = IsDebugToolAttached(extensions); // At the moment of writing this, only Nvidia's driver optimizes BufferSubData on exclusive // uniform buffers as "push constants" @@ -239,6 +250,8 @@ Device::Device() LOG_INFO(Render_OpenGL, "Renderer_VariableAOFFI: {}", has_variable_aoffi); LOG_INFO(Render_OpenGL, "Renderer_ComponentIndexingBug: {}", has_component_indexing_bug); LOG_INFO(Render_OpenGL, "Renderer_PreciseBug: {}", has_precise_bug); + LOG_INFO(Render_OpenGL, "Renderer_BrokenTextureViewFormats: {}", + has_broken_texture_view_formats); if (Settings::values.use_assembly_shaders.GetValue() && !use_assembly_shaders) { LOG_ERROR(Render_OpenGL, "Assembly shaders enabled but not supported"); diff --git a/src/video_core/renderer_opengl/gl_device.h b/src/video_core/renderer_opengl/gl_device.h index 8a4b6b9fc..3e79d1e37 100644 --- a/src/video_core/renderer_opengl/gl_device.h +++ b/src/video_core/renderer_opengl/gl_device.h @@ -36,11 +36,11 @@ public: return GetBaseBindings(static_cast<std::size_t>(shader_type)); } - std::size_t GetUniformBufferAlignment() const { + size_t GetUniformBufferAlignment() const { return uniform_buffer_alignment; } - std::size_t GetShaderStorageBufferAlignment() const { + size_t GetShaderStorageBufferAlignment() const { return shader_storage_alignment; } @@ -96,6 +96,10 @@ public: return has_precise_bug; } + bool HasBrokenTextureViewFormats() const { + return has_broken_texture_view_formats; + } + bool HasFastBufferSubData() const { return has_fast_buffer_sub_data; } @@ -104,6 +108,10 @@ public: return has_nv_viewport_array2; } + bool HasDebuggingToolAttached() const { + return has_debugging_tool_attached; + } + bool UseAssemblyShaders() const { return use_assembly_shaders; } @@ -118,8 +126,8 @@ private: std::array<u32, Tegra::Engines::MaxShaderTypes> max_uniform_buffers{}; std::array<BaseBindings, Tegra::Engines::MaxShaderTypes> base_bindings{}; - std::size_t uniform_buffer_alignment{}; - std::size_t shader_storage_alignment{}; + size_t uniform_buffer_alignment{}; + size_t shader_storage_alignment{}; u32 max_vertex_attributes{}; u32 max_varyings{}; u32 max_compute_shared_memory_size{}; @@ -133,8 +141,10 @@ private: bool has_variable_aoffi{}; bool has_component_indexing_bug{}; bool has_precise_bug{}; + bool has_broken_texture_view_formats{}; bool has_fast_buffer_sub_data{}; bool has_nv_viewport_array2{}; + bool has_debugging_tool_attached{}; bool use_assembly_shaders{}; bool use_asynchronous_shaders{}; }; diff --git a/src/video_core/renderer_opengl/gl_fence_manager.cpp b/src/video_core/renderer_opengl/gl_fence_manager.cpp index b532fdcc2..3e9c922f5 100644 --- a/src/video_core/renderer_opengl/gl_fence_manager.cpp +++ b/src/video_core/renderer_opengl/gl_fence_manager.cpp @@ -11,10 +11,10 @@ namespace OpenGL { -GLInnerFence::GLInnerFence(u32 payload, bool is_stubbed) : FenceBase(payload, is_stubbed) {} +GLInnerFence::GLInnerFence(u32 payload_, bool is_stubbed_) : FenceBase{payload_, is_stubbed_} {} -GLInnerFence::GLInnerFence(GPUVAddr address, u32 payload, bool is_stubbed) - : FenceBase(address, payload, is_stubbed) {} +GLInnerFence::GLInnerFence(GPUVAddr address_, u32 payload_, bool is_stubbed_) + : FenceBase{address_, payload_, is_stubbed_} {} GLInnerFence::~GLInnerFence() = default; @@ -45,10 +45,10 @@ void GLInnerFence::Wait() { glClientWaitSync(sync_object.handle, 0, GL_TIMEOUT_IGNORED); } -FenceManagerOpenGL::FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - TextureCacheOpenGL& texture_cache, - OGLBufferCache& buffer_cache, QueryCache& query_cache) - : GenericFenceManager{rasterizer, gpu, texture_cache, buffer_cache, query_cache} {} +FenceManagerOpenGL::FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer_, + Tegra::GPU& gpu_, TextureCache& texture_cache_, + OGLBufferCache& buffer_cache_, QueryCache& query_cache_) + : GenericFenceManager{rasterizer_, gpu_, texture_cache_, buffer_cache_, query_cache_} {} Fence FenceManagerOpenGL::CreateFence(u32 value, bool is_stubbed) { return std::make_shared<GLInnerFence>(value, is_stubbed); diff --git a/src/video_core/renderer_opengl/gl_fence_manager.h b/src/video_core/renderer_opengl/gl_fence_manager.h index da1dcdace..30dbee613 100644 --- a/src/video_core/renderer_opengl/gl_fence_manager.h +++ b/src/video_core/renderer_opengl/gl_fence_manager.h @@ -17,8 +17,8 @@ namespace OpenGL { class GLInnerFence : public VideoCommon::FenceBase { public: - GLInnerFence(u32 payload, bool is_stubbed); - GLInnerFence(GPUVAddr address, u32 payload, bool is_stubbed); + explicit GLInnerFence(u32 payload_, bool is_stubbed_); + explicit GLInnerFence(GPUVAddr address_, u32 payload_, bool is_stubbed_); ~GLInnerFence(); void Queue(); @@ -33,13 +33,13 @@ private: using Fence = std::shared_ptr<GLInnerFence>; using GenericFenceManager = - VideoCommon::FenceManager<Fence, TextureCacheOpenGL, OGLBufferCache, QueryCache>; + VideoCommon::FenceManager<Fence, TextureCache, OGLBufferCache, QueryCache>; class FenceManagerOpenGL final : public GenericFenceManager { public: - explicit FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - TextureCacheOpenGL& texture_cache, OGLBufferCache& buffer_cache, - QueryCache& query_cache); + explicit FenceManagerOpenGL(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, + TextureCache& texture_cache_, OGLBufferCache& buffer_cache_, + QueryCache& query_cache_); protected: Fence CreateFence(u32 value, bool is_stubbed) override; diff --git a/src/video_core/renderer_opengl/gl_framebuffer_cache.cpp b/src/video_core/renderer_opengl/gl_framebuffer_cache.cpp deleted file mode 100644 index b8a512cb6..000000000 --- a/src/video_core/renderer_opengl/gl_framebuffer_cache.cpp +++ /dev/null @@ -1,85 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <tuple> -#include <unordered_map> -#include <utility> - -#include <glad/glad.h> - -#include "common/common_types.h" -#include "video_core/engines/maxwell_3d.h" -#include "video_core/renderer_opengl/gl_framebuffer_cache.h" - -namespace OpenGL { - -using Maxwell = Tegra::Engines::Maxwell3D::Regs; -using VideoCore::Surface::SurfaceType; - -FramebufferCacheOpenGL::FramebufferCacheOpenGL() = default; - -FramebufferCacheOpenGL::~FramebufferCacheOpenGL() = default; - -GLuint FramebufferCacheOpenGL::GetFramebuffer(const FramebufferCacheKey& key) { - const auto [entry, is_cache_miss] = cache.try_emplace(key); - auto& framebuffer{entry->second}; - if (is_cache_miss) { - framebuffer = CreateFramebuffer(key); - } - return framebuffer.handle; -} - -OGLFramebuffer FramebufferCacheOpenGL::CreateFramebuffer(const FramebufferCacheKey& key) { - OGLFramebuffer framebuffer; - framebuffer.Create(); - - // TODO(Rodrigo): Use DSA here after Nvidia fixes their framebuffer DSA bugs. - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer.handle); - - if (key.zeta) { - const bool stencil = key.zeta->GetSurfaceParams().type == SurfaceType::DepthStencil; - const GLenum attach_target = stencil ? GL_DEPTH_STENCIL_ATTACHMENT : GL_DEPTH_ATTACHMENT; - key.zeta->Attach(attach_target, GL_DRAW_FRAMEBUFFER); - } - - std::size_t num_buffers = 0; - std::array<GLenum, Maxwell::NumRenderTargets> targets; - - for (std::size_t index = 0; index < Maxwell::NumRenderTargets; ++index) { - if (!key.colors[index]) { - targets[index] = GL_NONE; - continue; - } - const GLenum attach_target = GL_COLOR_ATTACHMENT0 + static_cast<GLenum>(index); - key.colors[index]->Attach(attach_target, GL_DRAW_FRAMEBUFFER); - - const u32 attachment = (key.color_attachments >> (BitsPerAttachment * index)) & 0b1111; - targets[index] = GL_COLOR_ATTACHMENT0 + attachment; - num_buffers = index + 1; - } - - if (num_buffers > 0) { - glDrawBuffers(static_cast<GLsizei>(num_buffers), std::data(targets)); - } else { - glDrawBuffer(GL_NONE); - } - - return framebuffer; -} - -std::size_t FramebufferCacheKey::Hash() const noexcept { - std::size_t hash = std::hash<View>{}(zeta); - for (const auto& color : colors) { - hash ^= std::hash<View>{}(color); - } - hash ^= static_cast<std::size_t>(color_attachments) << 16; - return hash; -} - -bool FramebufferCacheKey::operator==(const FramebufferCacheKey& rhs) const noexcept { - return std::tie(colors, zeta, color_attachments) == - std::tie(rhs.colors, rhs.zeta, rhs.color_attachments); -} - -} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_framebuffer_cache.h b/src/video_core/renderer_opengl/gl_framebuffer_cache.h deleted file mode 100644 index 8f698fee0..000000000 --- a/src/video_core/renderer_opengl/gl_framebuffer_cache.h +++ /dev/null @@ -1,68 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <array> -#include <cstddef> -#include <unordered_map> - -#include <glad/glad.h> - -#include "common/common_types.h" -#include "video_core/engines/maxwell_3d.h" -#include "video_core/renderer_opengl/gl_resource_manager.h" -#include "video_core/renderer_opengl/gl_texture_cache.h" - -namespace OpenGL { - -constexpr std::size_t BitsPerAttachment = 4; - -struct FramebufferCacheKey { - View zeta; - std::array<View, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> colors; - u32 color_attachments = 0; - - std::size_t Hash() const noexcept; - - bool operator==(const FramebufferCacheKey& rhs) const noexcept; - - bool operator!=(const FramebufferCacheKey& rhs) const noexcept { - return !operator==(rhs); - } - - void SetAttachment(std::size_t index, u32 attachment) { - color_attachments |= attachment << (BitsPerAttachment * index); - } -}; - -} // namespace OpenGL - -namespace std { - -template <> -struct hash<OpenGL::FramebufferCacheKey> { - std::size_t operator()(const OpenGL::FramebufferCacheKey& k) const noexcept { - return k.Hash(); - } -}; - -} // namespace std - -namespace OpenGL { - -class FramebufferCacheOpenGL { -public: - FramebufferCacheOpenGL(); - ~FramebufferCacheOpenGL(); - - GLuint GetFramebuffer(const FramebufferCacheKey& key); - -private: - OGLFramebuffer CreateFramebuffer(const FramebufferCacheKey& key); - - std::unordered_map<FramebufferCacheKey, OGLFramebuffer> cache; -}; - -} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_query_cache.cpp b/src/video_core/renderer_opengl/gl_query_cache.cpp index 1a3d9720e..acebbf5f4 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.cpp +++ b/src/video_core/renderer_opengl/gl_query_cache.cpp @@ -30,11 +30,9 @@ constexpr GLenum GetTarget(VideoCore::QueryType type) { } // Anonymous namespace -QueryCache::QueryCache(RasterizerOpenGL& rasterizer, Tegra::Engines::Maxwell3D& maxwell3d, - Tegra::MemoryManager& gpu_memory) - : VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter>( - rasterizer, maxwell3d, gpu_memory), - gl_rasterizer{rasterizer} {} +QueryCache::QueryCache(RasterizerOpenGL& rasterizer_, Tegra::Engines::Maxwell3D& maxwell3d_, + Tegra::MemoryManager& gpu_memory_) + : QueryCacheBase(rasterizer_, maxwell3d_, gpu_memory_), gl_rasterizer{rasterizer_} {} QueryCache::~QueryCache() = default; @@ -59,10 +57,11 @@ bool QueryCache::AnyCommandQueued() const noexcept { return gl_rasterizer.AnyCommandQueued(); } -HostCounter::HostCounter(QueryCache& cache, std::shared_ptr<HostCounter> dependency, - VideoCore::QueryType type) - : VideoCommon::HostCounterBase<QueryCache, HostCounter>{std::move(dependency)}, cache{cache}, - type{type}, query{cache.AllocateQuery(type)} { +HostCounter::HostCounter(QueryCache& cache_, std::shared_ptr<HostCounter> dependency_, + VideoCore::QueryType type_) + : HostCounterBase{std::move(dependency_)}, cache{cache_}, type{type_}, query{ + cache.AllocateQuery( + type)} { glBeginQuery(GetTarget(type), query.handle); } @@ -86,13 +85,14 @@ u64 HostCounter::BlockingQuery() const { return static_cast<u64>(value); } -CachedQuery::CachedQuery(QueryCache& cache, VideoCore::QueryType type, VAddr cpu_addr, u8* host_ptr) - : VideoCommon::CachedQueryBase<HostCounter>{cpu_addr, host_ptr}, cache{&cache}, type{type} {} +CachedQuery::CachedQuery(QueryCache& cache_, VideoCore::QueryType type_, VAddr cpu_addr_, + u8* host_ptr_) + : CachedQueryBase{cpu_addr_, host_ptr_}, cache{&cache_}, type{type_} {} CachedQuery::~CachedQuery() = default; CachedQuery::CachedQuery(CachedQuery&& rhs) noexcept - : VideoCommon::CachedQueryBase<HostCounter>(std::move(rhs)), cache{rhs.cache}, type{rhs.type} {} + : CachedQueryBase(std::move(rhs)), cache{rhs.cache}, type{rhs.type} {} CachedQuery& CachedQuery::operator=(CachedQuery&& rhs) noexcept { cache = rhs.cache; diff --git a/src/video_core/renderer_opengl/gl_query_cache.h b/src/video_core/renderer_opengl/gl_query_cache.h index 82cac51ee..7bbe5cfe9 100644 --- a/src/video_core/renderer_opengl/gl_query_cache.h +++ b/src/video_core/renderer_opengl/gl_query_cache.h @@ -29,8 +29,8 @@ using CounterStream = VideoCommon::CounterStreamBase<QueryCache, HostCounter>; class QueryCache final : public VideoCommon::QueryCacheBase<QueryCache, CachedQuery, CounterStream, HostCounter> { public: - explicit QueryCache(RasterizerOpenGL& rasterizer, Tegra::Engines::Maxwell3D& maxwell3d, - Tegra::MemoryManager& gpu_memory); + explicit QueryCache(RasterizerOpenGL& rasterizer_, Tegra::Engines::Maxwell3D& maxwell3d_, + Tegra::MemoryManager& gpu_memory_); ~QueryCache(); OGLQuery AllocateQuery(VideoCore::QueryType type); @@ -46,8 +46,8 @@ private: class HostCounter final : public VideoCommon::HostCounterBase<QueryCache, HostCounter> { public: - explicit HostCounter(QueryCache& cache, std::shared_ptr<HostCounter> dependency, - VideoCore::QueryType type); + explicit HostCounter(QueryCache& cache_, std::shared_ptr<HostCounter> dependency_, + VideoCore::QueryType type_); ~HostCounter(); void EndQuery(); @@ -62,8 +62,8 @@ private: class CachedQuery final : public VideoCommon::CachedQueryBase<HostCounter> { public: - explicit CachedQuery(QueryCache& cache, VideoCore::QueryType type, VAddr cpu_addr, - u8* host_ptr); + explicit CachedQuery(QueryCache& cache_, VideoCore::QueryType type_, VAddr cpu_addr_, + u8* host_ptr_); ~CachedQuery() override; CachedQuery(CachedQuery&& rhs) noexcept; diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index bbb2eb17c..8aa63d329 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -25,12 +25,15 @@ #include "video_core/engines/maxwell_3d.h" #include "video_core/engines/shader_type.h" #include "video_core/memory_manager.h" +#include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_query_cache.h" #include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_shader_cache.h" +#include "video_core/renderer_opengl/gl_texture_cache.h" #include "video_core/renderer_opengl/maxwell_to_gl.h" #include "video_core/renderer_opengl/renderer_opengl.h" #include "video_core/shader_cache.h" +#include "video_core/texture_cache/texture_cache.h" namespace OpenGL { @@ -55,18 +58,32 @@ MICROPROFILE_DEFINE(OpenGL_PrimitiveAssembly, "OpenGL", "Prim Asmbl", MP_RGB(255 namespace { -constexpr std::size_t NUM_CONST_BUFFERS_PER_STAGE = 18; -constexpr std::size_t NUM_CONST_BUFFERS_BYTES_PER_STAGE = +constexpr size_t NUM_CONST_BUFFERS_PER_STAGE = 18; +constexpr size_t NUM_CONST_BUFFERS_BYTES_PER_STAGE = NUM_CONST_BUFFERS_PER_STAGE * Maxwell::MaxConstBufferSize; -constexpr std::size_t TOTAL_CONST_BUFFER_BYTES = +constexpr size_t TOTAL_CONST_BUFFER_BYTES = NUM_CONST_BUFFERS_BYTES_PER_STAGE * Maxwell::MaxShaderStage; -constexpr std::size_t NUM_SUPPORTED_VERTEX_ATTRIBUTES = 16; -constexpr std::size_t NUM_SUPPORTED_VERTEX_BINDINGS = 16; +constexpr size_t NUM_SUPPORTED_VERTEX_ATTRIBUTES = 16; +constexpr size_t NUM_SUPPORTED_VERTEX_BINDINGS = 16; + +constexpr size_t MAX_TEXTURES = 192; +constexpr size_t MAX_IMAGES = 48; + +struct TextureHandle { + constexpr TextureHandle(u32 data, bool via_header_index) { + const Tegra::Texture::TextureHandle handle{data}; + image = handle.tic_id; + sampler = via_header_index ? image : handle.tsc_id.Value(); + } + + u32 image; + u32 sampler; +}; template <typename Engine, typename Entry> -Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry, - ShaderType shader_type, std::size_t index = 0) { +TextureHandle GetTextureInfo(const Engine& engine, bool via_header_index, const Entry& entry, + ShaderType shader_type, size_t index = 0) { if constexpr (std::is_same_v<Entry, SamplerEntry>) { if (entry.is_separated) { const u32 buffer_1 = entry.buffer; @@ -75,21 +92,16 @@ Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry const u32 offset_2 = entry.secondary_offset; const u32 handle_1 = engine.AccessConstBuffer32(shader_type, buffer_1, offset_1); const u32 handle_2 = engine.AccessConstBuffer32(shader_type, buffer_2, offset_2); - return engine.GetTextureInfo(handle_1 | handle_2); + return TextureHandle(handle_1 | handle_2, via_header_index); } } if (entry.is_bindless) { - const u32 handle = engine.AccessConstBuffer32(shader_type, entry.buffer, entry.offset); - return engine.GetTextureInfo(handle); - } - - const auto& gpu_profile = engine.AccessGuestDriverProfile(); - const u32 offset = entry.offset + static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); - if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) { - return engine.GetStageTexture(shader_type, offset); - } else { - return engine.GetTexture(offset); + const u32 raw = engine.AccessConstBuffer32(shader_type, entry.buffer, entry.offset); + return TextureHandle(raw, via_header_index); } + const u32 buffer = engine.GetBoundBuffer(); + const u64 offset = (entry.offset + index) * sizeof(u32); + return TextureHandle(engine.AccessConstBuffer32(shader_type, buffer, offset), via_header_index); } std::size_t GetConstBufferSize(const Tegra::Engines::ConstBufferInfo& buffer, @@ -97,7 +109,6 @@ std::size_t GetConstBufferSize(const Tegra::Engines::ConstBufferInfo& buffer, if (!entry.IsIndirect()) { return entry.GetSize(); } - if (buffer.size > Maxwell::MaxConstBufferSize) { LOG_WARNING(Render_OpenGL, "Indirect constbuffer size {} exceeds maximum {}", buffer.size, Maxwell::MaxConstBufferSize); @@ -131,7 +142,7 @@ std::pair<GLint, GLint> TransformFeedbackEnum(u8 location) { case 43: return {GL_BACK_SECONDARY_COLOR_NV, 0}; } - UNIMPLEMENTED_MSG("index={}", static_cast<int>(index)); + UNIMPLEMENTED_MSG("index={}", index); return {GL_POSITION, 0}; } @@ -139,35 +150,68 @@ void oglEnable(GLenum cap, bool state) { (state ? glEnable : glDisable)(cap); } -void UpdateBindlessPointers(GLenum target, GLuint64EXT* pointers, std::size_t num_entries) { - if (num_entries == 0) { +void UpdateBindlessSSBOs(GLenum target, const BindlessSSBO* ssbos, size_t num_ssbos) { + if (num_ssbos == 0) { return; } - if (num_entries % 2 == 1) { - pointers[num_entries] = 0; + glProgramLocalParametersI4uivNV(target, 0, static_cast<GLsizei>(num_ssbos), + reinterpret_cast<const GLuint*>(ssbos)); +} + +ImageViewType ImageViewTypeFromEntry(const SamplerEntry& entry) { + if (entry.is_buffer) { + return ImageViewType::Buffer; + } + switch (entry.type) { + case Tegra::Shader::TextureType::Texture1D: + return entry.is_array ? ImageViewType::e1DArray : ImageViewType::e1D; + case Tegra::Shader::TextureType::Texture2D: + return entry.is_array ? ImageViewType::e2DArray : ImageViewType::e2D; + case Tegra::Shader::TextureType::Texture3D: + return ImageViewType::e3D; + case Tegra::Shader::TextureType::TextureCube: + return entry.is_array ? ImageViewType::CubeArray : ImageViewType::Cube; + } + UNREACHABLE(); + return ImageViewType::e2D; +} + +ImageViewType ImageViewTypeFromEntry(const ImageEntry& entry) { + switch (entry.type) { + case Tegra::Shader::ImageType::Texture1D: + return ImageViewType::e1D; + case Tegra::Shader::ImageType::Texture1DArray: + return ImageViewType::e1DArray; + case Tegra::Shader::ImageType::Texture2D: + return ImageViewType::e2D; + case Tegra::Shader::ImageType::Texture2DArray: + return ImageViewType::e2DArray; + case Tegra::Shader::ImageType::Texture3D: + return ImageViewType::e3D; + case Tegra::Shader::ImageType::TextureBuffer: + return ImageViewType::Buffer; } - const GLsizei num_vectors = static_cast<GLsizei>((num_entries + 1) / 2); - glProgramLocalParametersI4uivNV(target, 0, num_vectors, - reinterpret_cast<const GLuint*>(pointers)); + UNREACHABLE(); + return ImageViewType::e2D; } } // Anonymous namespace -RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu_, - Core::Memory::Memory& cpu_memory, const Device& device_, +RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_, + Core::Memory::Memory& cpu_memory_, const Device& device_, ScreenInfo& screen_info_, ProgramManager& program_manager_, StateTracker& state_tracker_) - : RasterizerAccelerated{cpu_memory}, gpu(gpu_), maxwell3d(gpu.Maxwell3D()), + : RasterizerAccelerated(cpu_memory_), gpu(gpu_), maxwell3d(gpu.Maxwell3D()), kepler_compute(gpu.KeplerCompute()), gpu_memory(gpu.MemoryManager()), device(device_), screen_info(screen_info_), program_manager(program_manager_), state_tracker(state_tracker_), - texture_cache(*this, maxwell3d, gpu_memory, device, state_tracker), - shader_cache(*this, emu_window, gpu, maxwell3d, kepler_compute, gpu_memory, device), + stream_buffer(device, state_tracker), + texture_cache_runtime(device, program_manager, state_tracker), + texture_cache(texture_cache_runtime, *this, maxwell3d, kepler_compute, gpu_memory), + shader_cache(*this, emu_window_, gpu, maxwell3d, kepler_compute, gpu_memory, device), query_cache(*this, maxwell3d, gpu_memory), - buffer_cache(*this, gpu_memory, cpu_memory, device, STREAM_BUFFER_SIZE), + buffer_cache(*this, gpu_memory, cpu_memory_, device, stream_buffer, state_tracker), fence_manager(*this, gpu, texture_cache, buffer_cache, query_cache), - async_shaders(emu_window) { - CheckExtensions(); - + async_shaders(emu_window_) { unified_uniform_buffer.Create(); glNamedBufferStorage(unified_uniform_buffer.handle, TOTAL_CONST_BUFFER_BYTES, nullptr, 0); @@ -178,7 +222,6 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window, Tegra: nullptr, 0); } } - if (device.UseAsynchronousShaders()) { async_shaders.AllocateWorkers(); } @@ -190,14 +233,6 @@ RasterizerOpenGL::~RasterizerOpenGL() { } } -void RasterizerOpenGL::CheckExtensions() { - if (!GLAD_GL_ARB_texture_filter_anisotropic && !GLAD_GL_EXT_texture_filter_anisotropic) { - LOG_WARNING( - Render_OpenGL, - "Anisotropic filter is not supported! This can cause graphical issues in some games."); - } -} - void RasterizerOpenGL::SetupVertexFormat() { auto& flags = maxwell3d.dirty.flags; if (!flags[Dirty::VertexFormats]) { @@ -320,10 +355,16 @@ GLintptr RasterizerOpenGL::SetupIndexBuffer() { return info.offset; } -void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { +void RasterizerOpenGL::SetupShaders() { MICROPROFILE_SCOPE(OpenGL_Shader); u32 clip_distances = 0; + std::array<Shader*, Maxwell::MaxShaderStage> shaders{}; + image_view_indices.clear(); + sampler_handles.clear(); + + texture_cache.SynchronizeGraphicsDescriptors(); + for (std::size_t index = 0; index < Maxwell::MaxShaderProgram; ++index) { const auto& shader_config = maxwell3d.regs.shader_config[index]; const auto program{static_cast<Maxwell::ShaderProgram>(index)}; @@ -342,7 +383,6 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { } continue; } - // Currently this stages are not supported in the OpenGL backend. // TODO(Blinkhawk): Port tesselation shaders from Vulkan to OpenGL if (program == Maxwell::ShaderProgram::TesselationControl || @@ -351,7 +391,6 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { } Shader* const shader = shader_cache.GetStageProgram(program, async_shaders); - const GLuint program_handle = shader->IsBuilt() ? shader->GetHandle() : 0; switch (program) { case Maxwell::ShaderProgram::VertexA: @@ -367,14 +406,17 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { default: UNIMPLEMENTED_MSG("Unimplemented shader index={}, enable={}, offset=0x{:08X}", index, shader_config.enable.Value(), shader_config.offset); + break; } // Stage indices are 0 - 5 - const std::size_t stage = index == 0 ? 0 : index - 1; + const size_t stage = index == 0 ? 0 : index - 1; + shaders[stage] = shader; + SetupDrawConstBuffers(stage, shader); SetupDrawGlobalMemory(stage, shader); - SetupDrawTextures(stage, shader); - SetupDrawImages(stage, shader); + SetupDrawTextures(shader, stage); + SetupDrawImages(shader, stage); // Workaround for Intel drivers. // When a clip distance is enabled but not set in the shader it crops parts of the screen @@ -388,9 +430,23 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { ++index; } } - SyncClipEnabled(clip_distances); maxwell3d.dirty.flags[Dirty::Shaders] = false; + + const std::span indices_span(image_view_indices.data(), image_view_indices.size()); + texture_cache.FillGraphicsImageViews(indices_span, image_view_ids); + + size_t image_view_index = 0; + size_t texture_index = 0; + size_t image_index = 0; + for (size_t stage = 0; stage < Maxwell::MaxShaderStage; ++stage) { + const Shader* const shader = shaders[stage]; + if (shader) { + const auto base = device.GetBaseBindings(stage); + BindTextures(shader->GetEntries(), base.sampler, base.image, image_view_index, + texture_index, image_index); + } + } } std::size_t RasterizerOpenGL::CalculateVertexArraysSize() const { @@ -421,98 +477,6 @@ void RasterizerOpenGL::LoadDiskResources(u64 title_id, const std::atomic_bool& s shader_cache.LoadDiskCache(title_id, stop_loading, callback); } -void RasterizerOpenGL::ConfigureFramebuffers() { - MICROPROFILE_SCOPE(OpenGL_Framebuffer); - if (!maxwell3d.dirty.flags[VideoCommon::Dirty::RenderTargets]) { - return; - } - maxwell3d.dirty.flags[VideoCommon::Dirty::RenderTargets] = false; - - texture_cache.GuardRenderTargets(true); - - View depth_surface = texture_cache.GetDepthBufferSurface(true); - - const auto& regs = maxwell3d.regs; - UNIMPLEMENTED_IF(regs.rt_separate_frag_data == 0); - - // Bind the framebuffer surfaces - FramebufferCacheKey key; - const auto colors_count = static_cast<std::size_t>(regs.rt_control.count); - for (std::size_t index = 0; index < colors_count; ++index) { - View color_surface{texture_cache.GetColorBufferSurface(index, true)}; - if (!color_surface) { - continue; - } - // Assume that a surface will be written to if it is used as a framebuffer, even - // if the shader doesn't actually write to it. - texture_cache.MarkColorBufferInUse(index); - - key.SetAttachment(index, regs.rt_control.GetMap(index)); - key.colors[index] = std::move(color_surface); - } - - if (depth_surface) { - // Assume that a surface will be written to if it is used as a framebuffer, even if - // the shader doesn't actually write to it. - texture_cache.MarkDepthBufferInUse(); - key.zeta = std::move(depth_surface); - } - - texture_cache.GuardRenderTargets(false); - - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer_cache.GetFramebuffer(key)); -} - -void RasterizerOpenGL::ConfigureClearFramebuffer(bool using_color, bool using_depth_stencil) { - const auto& regs = maxwell3d.regs; - - texture_cache.GuardRenderTargets(true); - View color_surface; - - if (using_color) { - // Determine if we have to preserve the contents. - // First we have to make sure all clear masks are enabled. - bool preserve_contents = !regs.clear_buffers.R || !regs.clear_buffers.G || - !regs.clear_buffers.B || !regs.clear_buffers.A; - const std::size_t index = regs.clear_buffers.RT; - if (regs.clear_flags.scissor) { - // Then we have to confirm scissor testing clears the whole image. - const auto& scissor = regs.scissor_test[0]; - preserve_contents |= scissor.min_x > 0; - preserve_contents |= scissor.min_y > 0; - preserve_contents |= scissor.max_x < regs.rt[index].width; - preserve_contents |= scissor.max_y < regs.rt[index].height; - } - - color_surface = texture_cache.GetColorBufferSurface(index, preserve_contents); - texture_cache.MarkColorBufferInUse(index); - } - - View depth_surface; - if (using_depth_stencil) { - bool preserve_contents = false; - if (regs.clear_flags.scissor) { - // For depth stencil clears we only have to confirm scissor test covers the whole image. - const auto& scissor = regs.scissor_test[0]; - preserve_contents |= scissor.min_x > 0; - preserve_contents |= scissor.min_y > 0; - preserve_contents |= scissor.max_x < regs.zeta_width; - preserve_contents |= scissor.max_y < regs.zeta_height; - } - - depth_surface = texture_cache.GetDepthBufferSurface(preserve_contents); - texture_cache.MarkDepthBufferInUse(); - } - texture_cache.GuardRenderTargets(false); - - FramebufferCacheKey key; - key.colors[0] = std::move(color_surface); - key.zeta = std::move(depth_surface); - - state_tracker.NotifyFramebuffer(); - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer_cache.GetFramebuffer(key)); -} - void RasterizerOpenGL::Clear() { if (!maxwell3d.ShouldExecute()) { return; @@ -527,8 +491,9 @@ void RasterizerOpenGL::Clear() { regs.clear_buffers.A) { use_color = true; - state_tracker.NotifyColorMask0(); - glColorMaski(0, regs.clear_buffers.R != 0, regs.clear_buffers.G != 0, + const GLuint index = regs.clear_buffers.RT; + state_tracker.NotifyColorMask(index); + glColorMaski(index, regs.clear_buffers.R != 0, regs.clear_buffers.G != 0, regs.clear_buffers.B != 0, regs.clear_buffers.A != 0); // TODO(Rodrigo): Determine if clamping is used on clears @@ -561,15 +526,17 @@ void RasterizerOpenGL::Clear() { state_tracker.NotifyScissor0(); glDisablei(GL_SCISSOR_TEST, 0); } - UNIMPLEMENTED_IF(regs.clear_flags.viewport); - ConfigureClearFramebuffer(use_color, use_depth || use_stencil); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.UpdateRenderTargets(true); + state_tracker.BindFramebuffer(texture_cache.GetFramebuffer()->Handle()); + } if (use_color) { - glClearBufferfv(GL_COLOR, 0, regs.clear_color); + glClearBufferfv(GL_COLOR, regs.clear_buffers.RT, regs.clear_color); } - if (use_depth && use_stencil) { glClearBufferfi(GL_DEPTH_STENCIL, 0, regs.clear_depth, regs.clear_stencil); } else if (use_depth) { @@ -626,16 +593,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { (Maxwell::MaxConstBufferSize + device.GetUniformBufferAlignment()); // Prepare the vertex array. - const bool invalidated = buffer_cache.Map(buffer_size); - - if (invalidated) { - // When the stream buffer has been invalidated, we have to consider vertex buffers as dirty - auto& dirty = maxwell3d.dirty.flags; - dirty[Dirty::VertexBuffers] = true; - for (int index = Dirty::VertexBuffer0; index <= Dirty::VertexBuffer31; ++index) { - dirty[index] = true; - } - } + buffer_cache.Map(buffer_size); // Prepare vertex array format. SetupVertexFormat(); @@ -659,22 +617,16 @@ void RasterizerOpenGL::Draw(bool is_indexed, bool is_instanced) { } // Setup shaders and their used resources. - texture_cache.GuardSamplers(true); - const GLenum primitive_mode = MaxwellToGL::PrimitiveTopology(maxwell3d.regs.draw.topology); - SetupShaders(primitive_mode); - texture_cache.GuardSamplers(false); - - ConfigureFramebuffers(); + auto lock = texture_cache.AcquireLock(); + SetupShaders(); // Signal the buffer cache that we are not going to upload more things. buffer_cache.Unmap(); - + texture_cache.UpdateRenderTargets(false); + state_tracker.BindFramebuffer(texture_cache.GetFramebuffer()->Handle()); program_manager.BindGraphicsPipeline(); - if (texture_cache.TextureBarrier()) { - glTextureBarrier(); - } - + const GLenum primitive_mode = MaxwellToGL::PrimitiveTopology(maxwell3d.regs.draw.topology); BeginTransformFeedback(primitive_mode); const GLuint base_instance = static_cast<GLuint>(maxwell3d.regs.vb_base_instance); @@ -726,15 +678,13 @@ void RasterizerOpenGL::DispatchCompute(GPUVAddr code_addr) { buffer_cache.Acquire(); current_cbuf = 0; - auto kernel = shader_cache.GetComputeKernel(code_addr); - program_manager.BindCompute(kernel->GetHandle()); + Shader* const kernel = shader_cache.GetComputeKernel(code_addr); - SetupComputeTextures(kernel); - SetupComputeImages(kernel); + auto lock = texture_cache.AcquireLock(); + BindComputeTextures(kernel); - const std::size_t buffer_size = - Tegra::Engines::KeplerCompute::NumConstBuffers * - (Maxwell::MaxConstBufferSize + device.GetUniformBufferAlignment()); + const size_t buffer_size = Tegra::Engines::KeplerCompute::NumConstBuffers * + (Maxwell::MaxConstBufferSize + device.GetUniformBufferAlignment()); buffer_cache.Map(buffer_size); SetupComputeConstBuffers(kernel); @@ -743,7 +693,6 @@ void RasterizerOpenGL::DispatchCompute(GPUVAddr code_addr) { buffer_cache.Unmap(); const auto& launch_desc = kepler_compute.launch_description; - program_manager.BindCompute(kernel->GetHandle()); glDispatchCompute(launch_desc.grid_dim_x, launch_desc.grid_dim_y, launch_desc.grid_dim_z); ++num_queued_commands; } @@ -764,7 +713,10 @@ void RasterizerOpenGL::FlushRegion(VAddr addr, u64 size) { if (addr == 0 || size == 0) { return; } - texture_cache.FlushRegion(addr, size); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.DownloadMemory(addr, size); + } buffer_cache.FlushRegion(addr, size); query_cache.FlushRegion(addr, size); } @@ -773,7 +725,8 @@ bool RasterizerOpenGL::MustFlushRegion(VAddr addr, u64 size) { if (!Settings::IsGPULevelHigh()) { return buffer_cache.MustFlushRegion(addr, size); } - return texture_cache.MustFlushRegion(addr, size) || buffer_cache.MustFlushRegion(addr, size); + return texture_cache.IsRegionGpuModified(addr, size) || + buffer_cache.MustFlushRegion(addr, size); } void RasterizerOpenGL::InvalidateRegion(VAddr addr, u64 size) { @@ -781,7 +734,10 @@ void RasterizerOpenGL::InvalidateRegion(VAddr addr, u64 size) { if (addr == 0 || size == 0) { return; } - texture_cache.InvalidateRegion(addr, size); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.WriteMemory(addr, size); + } shader_cache.InvalidateRegion(addr, size); buffer_cache.InvalidateRegion(addr, size); query_cache.InvalidateRegion(addr, size); @@ -792,18 +748,29 @@ void RasterizerOpenGL::OnCPUWrite(VAddr addr, u64 size) { if (addr == 0 || size == 0) { return; } - texture_cache.OnCPUWrite(addr, size); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.WriteMemory(addr, size); + } shader_cache.OnCPUWrite(addr, size); buffer_cache.OnCPUWrite(addr, size); } void RasterizerOpenGL::SyncGuestHost() { MICROPROFILE_SCOPE(OpenGL_CacheManagement); - texture_cache.SyncGuestHost(); buffer_cache.SyncGuestHost(); shader_cache.SyncGuestHost(); } +void RasterizerOpenGL::UnmapMemory(VAddr addr, u64 size) { + { + auto lock = texture_cache.AcquireLock(); + texture_cache.UnmapMemory(addr, size); + } + buffer_cache.OnCPUWrite(addr, size); + shader_cache.OnCPUWrite(addr, size); +} + void RasterizerOpenGL::SignalSemaphore(GPUVAddr addr, u32 value) { if (!gpu.IsAsync()) { gpu_memory.Write<u32>(addr, value); @@ -845,6 +812,14 @@ void RasterizerOpenGL::WaitForIdle() { GL_SHADER_STORAGE_BARRIER_BIT | GL_QUERY_BUFFER_BARRIER_BIT); } +void RasterizerOpenGL::FragmentBarrier() { + glMemoryBarrier(GL_FRAMEBUFFER_BARRIER_BIT); +} + +void RasterizerOpenGL::TiledCacheBarrier() { + glTextureBarrier(); +} + void RasterizerOpenGL::FlushCommands() { // Only flush when we have commands queued to OpenGL. if (num_queued_commands == 0) { @@ -858,53 +833,103 @@ void RasterizerOpenGL::TickFrame() { // Ticking a frame means that buffers will be swapped, calling glFlush implicitly. num_queued_commands = 0; + fence_manager.TickFrame(); buffer_cache.TickFrame(); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.TickFrame(); + } } -bool RasterizerOpenGL::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, - const Tegra::Engines::Fermi2D::Regs::Surface& dst, +bool RasterizerOpenGL::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Surface& src, + const Tegra::Engines::Fermi2D::Surface& dst, const Tegra::Engines::Fermi2D::Config& copy_config) { MICROPROFILE_SCOPE(OpenGL_Blits); - texture_cache.DoFermiCopy(src, dst, copy_config); + auto lock = texture_cache.AcquireLock(); + texture_cache.BlitImage(dst, src, copy_config); return true; } bool RasterizerOpenGL::AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr, u32 pixel_stride) { - if (!framebuffer_addr) { - return {}; + if (framebuffer_addr == 0) { + return false; } - MICROPROFILE_SCOPE(OpenGL_CacheManagement); - const auto surface{texture_cache.TryFindFramebufferSurface(framebuffer_addr)}; - if (!surface) { - return {}; + auto lock = texture_cache.AcquireLock(); + ImageView* const image_view{texture_cache.TryFindFramebufferImageView(framebuffer_addr)}; + if (!image_view) { + return false; } - // Verify that the cached surface is the same size and format as the requested framebuffer - const auto& params{surface->GetSurfaceParams()}; - const auto& pixel_format{ - VideoCore::Surface::PixelFormatFromGPUPixelFormat(config.pixel_format)}; - ASSERT_MSG(params.width == config.width, "Framebuffer width is different"); - ASSERT_MSG(params.height == config.height, "Framebuffer height is different"); + // ASSERT_MSG(image_view->size.width == config.width, "Framebuffer width is different"); + // ASSERT_MSG(image_view->size.height == config.height, "Framebuffer height is different"); - if (params.pixel_format != pixel_format) { - LOG_DEBUG(Render_OpenGL, "Framebuffer pixel_format is different"); - } + screen_info.display_texture = image_view->Handle(ImageViewType::e2D); + screen_info.display_srgb = VideoCore::Surface::IsPixelFormatSRGB(image_view->format); + return true; +} - screen_info.display_texture = surface->GetTexture(); - screen_info.display_srgb = surface->GetSurfaceParams().srgb_conversion; +void RasterizerOpenGL::BindComputeTextures(Shader* kernel) { + image_view_indices.clear(); + sampler_handles.clear(); - return true; + texture_cache.SynchronizeComputeDescriptors(); + + SetupComputeTextures(kernel); + SetupComputeImages(kernel); + + const std::span indices_span(image_view_indices.data(), image_view_indices.size()); + texture_cache.FillComputeImageViews(indices_span, image_view_ids); + + program_manager.BindCompute(kernel->GetHandle()); + size_t image_view_index = 0; + size_t texture_index = 0; + size_t image_index = 0; + BindTextures(kernel->GetEntries(), 0, 0, image_view_index, texture_index, image_index); +} + +void RasterizerOpenGL::BindTextures(const ShaderEntries& entries, GLuint base_texture, + GLuint base_image, size_t& image_view_index, + size_t& texture_index, size_t& image_index) { + const GLuint* const samplers = sampler_handles.data() + texture_index; + const GLuint* const textures = texture_handles.data() + texture_index; + const GLuint* const images = image_handles.data() + image_index; + + const size_t num_samplers = entries.samplers.size(); + for (const auto& sampler : entries.samplers) { + for (size_t i = 0; i < sampler.size; ++i) { + const ImageViewId image_view_id = image_view_ids[image_view_index++]; + const ImageView& image_view = texture_cache.GetImageView(image_view_id); + const GLuint handle = image_view.Handle(ImageViewTypeFromEntry(sampler)); + texture_handles[texture_index++] = handle; + } + } + const size_t num_images = entries.images.size(); + for (size_t unit = 0; unit < num_images; ++unit) { + // TODO: Mark as modified + const ImageViewId image_view_id = image_view_ids[image_view_index++]; + const ImageView& image_view = texture_cache.GetImageView(image_view_id); + const GLuint handle = image_view.Handle(ImageViewTypeFromEntry(entries.images[unit])); + image_handles[image_index] = handle; + ++image_index; + } + if (num_samplers > 0) { + glBindSamplers(base_texture, static_cast<GLsizei>(num_samplers), samplers); + glBindTextures(base_texture, static_cast<GLsizei>(num_samplers), textures); + } + if (num_images > 0) { + glBindImageTextures(base_image, static_cast<GLsizei>(num_images), images); + } } void RasterizerOpenGL::SetupDrawConstBuffers(std::size_t stage_index, Shader* shader) { - static constexpr std::array PARAMETER_LUT = { - GL_VERTEX_PROGRAM_PARAMETER_BUFFER_NV, GL_TESS_CONTROL_PROGRAM_PARAMETER_BUFFER_NV, + static constexpr std::array PARAMETER_LUT{ + GL_VERTEX_PROGRAM_PARAMETER_BUFFER_NV, GL_TESS_CONTROL_PROGRAM_PARAMETER_BUFFER_NV, GL_TESS_EVALUATION_PROGRAM_PARAMETER_BUFFER_NV, GL_GEOMETRY_PROGRAM_PARAMETER_BUFFER_NV, - GL_FRAGMENT_PROGRAM_PARAMETER_BUFFER_NV}; - + GL_FRAGMENT_PROGRAM_PARAMETER_BUFFER_NV, + }; MICROPROFILE_SCOPE(OpenGL_UBO); const auto& stages = maxwell3d.state.shader_stages; const auto& shader_stage = stages[stage_index]; @@ -1003,12 +1028,11 @@ void RasterizerOpenGL::SetupDrawGlobalMemory(std::size_t stage_index, Shader* sh GL_VERTEX_PROGRAM_NV, GL_TESS_CONTROL_PROGRAM_NV, GL_TESS_EVALUATION_PROGRAM_NV, GL_GEOMETRY_PROGRAM_NV, GL_FRAGMENT_PROGRAM_NV, }; - const auto& cbufs{maxwell3d.state.shader_stages[stage_index]}; const auto& entries{shader->GetEntries().global_memory_entries}; - std::array<GLuint64EXT, 32> pointers; - ASSERT(entries.size() < pointers.size()); + std::array<BindlessSSBO, 32> ssbos; + ASSERT(entries.size() < ssbos.size()); const bool assembly_shaders = device.UseAssemblyShaders(); u32 binding = assembly_shaders ? 0 : device.GetBaseBindings(stage_index).shader_storage_buffer; @@ -1016,11 +1040,11 @@ void RasterizerOpenGL::SetupDrawGlobalMemory(std::size_t stage_index, Shader* sh const GPUVAddr addr{cbufs.const_buffers[entry.cbuf_index].address + entry.cbuf_offset}; const GPUVAddr gpu_addr{gpu_memory.Read<u64>(addr)}; const u32 size{gpu_memory.Read<u32>(addr + 8)}; - SetupGlobalMemory(binding, entry, gpu_addr, size, &pointers[binding]); + SetupGlobalMemory(binding, entry, gpu_addr, size, &ssbos[binding]); ++binding; } if (assembly_shaders) { - UpdateBindlessPointers(TARGET_LUT[stage_index], pointers.data(), entries.size()); + UpdateBindlessSSBOs(TARGET_LUT[stage_index], ssbos.data(), entries.size()); } } @@ -1028,106 +1052,85 @@ void RasterizerOpenGL::SetupComputeGlobalMemory(Shader* kernel) { const auto& cbufs{kepler_compute.launch_description.const_buffer_config}; const auto& entries{kernel->GetEntries().global_memory_entries}; - std::array<GLuint64EXT, 32> pointers; - ASSERT(entries.size() < pointers.size()); + std::array<BindlessSSBO, 32> ssbos; + ASSERT(entries.size() < ssbos.size()); u32 binding = 0; for (const auto& entry : entries) { const GPUVAddr addr{cbufs[entry.cbuf_index].Address() + entry.cbuf_offset}; const GPUVAddr gpu_addr{gpu_memory.Read<u64>(addr)}; const u32 size{gpu_memory.Read<u32>(addr + 8)}; - SetupGlobalMemory(binding, entry, gpu_addr, size, &pointers[binding]); + SetupGlobalMemory(binding, entry, gpu_addr, size, &ssbos[binding]); ++binding; } if (device.UseAssemblyShaders()) { - UpdateBindlessPointers(GL_COMPUTE_PROGRAM_NV, pointers.data(), entries.size()); + UpdateBindlessSSBOs(GL_COMPUTE_PROGRAM_NV, ssbos.data(), ssbos.size()); } } void RasterizerOpenGL::SetupGlobalMemory(u32 binding, const GlobalMemoryEntry& entry, - GPUVAddr gpu_addr, std::size_t size, - GLuint64EXT* pointer) { - const std::size_t alignment{device.GetShaderStorageBufferAlignment()}; + GPUVAddr gpu_addr, size_t size, BindlessSSBO* ssbo) { + const size_t alignment{device.GetShaderStorageBufferAlignment()}; const auto info = buffer_cache.UploadMemory(gpu_addr, size, alignment, entry.is_written); if (device.UseAssemblyShaders()) { - *pointer = info.address + info.offset; + *ssbo = BindlessSSBO{ + .address = static_cast<GLuint64EXT>(info.address + info.offset), + .length = static_cast<GLsizei>(size), + .padding = 0, + }; } else { glBindBufferRange(GL_SHADER_STORAGE_BUFFER, binding, info.handle, info.offset, static_cast<GLsizeiptr>(size)); } } -void RasterizerOpenGL::SetupDrawTextures(std::size_t stage_index, Shader* shader) { - MICROPROFILE_SCOPE(OpenGL_Texture); - u32 binding = device.GetBaseBindings(stage_index).sampler; +void RasterizerOpenGL::SetupDrawTextures(const Shader* shader, size_t stage_index) { + const bool via_header_index = + maxwell3d.regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex; for (const auto& entry : shader->GetEntries().samplers) { const auto shader_type = static_cast<ShaderType>(stage_index); - for (std::size_t i = 0; i < entry.size; ++i) { - const auto texture = GetTextureInfo(maxwell3d, entry, shader_type, i); - SetupTexture(binding++, texture, entry); + for (size_t index = 0; index < entry.size; ++index) { + const auto handle = + GetTextureInfo(maxwell3d, via_header_index, entry, shader_type, index); + const Sampler* const sampler = texture_cache.GetGraphicsSampler(handle.sampler); + sampler_handles.push_back(sampler->Handle()); + image_view_indices.push_back(handle.image); } } } -void RasterizerOpenGL::SetupComputeTextures(Shader* kernel) { - MICROPROFILE_SCOPE(OpenGL_Texture); - u32 binding = 0; +void RasterizerOpenGL::SetupComputeTextures(const Shader* kernel) { + const bool via_header_index = kepler_compute.launch_description.linked_tsc; for (const auto& entry : kernel->GetEntries().samplers) { - for (std::size_t i = 0; i < entry.size; ++i) { - const auto texture = GetTextureInfo(kepler_compute, entry, ShaderType::Compute, i); - SetupTexture(binding++, texture, entry); + for (size_t i = 0; i < entry.size; ++i) { + const auto handle = + GetTextureInfo(kepler_compute, via_header_index, entry, ShaderType::Compute, i); + const Sampler* const sampler = texture_cache.GetComputeSampler(handle.sampler); + sampler_handles.push_back(sampler->Handle()); + image_view_indices.push_back(handle.image); } } } -void RasterizerOpenGL::SetupTexture(u32 binding, const Tegra::Texture::FullTextureInfo& texture, - const SamplerEntry& entry) { - const auto view = texture_cache.GetTextureSurface(texture.tic, entry); - if (!view) { - // Can occur when texture addr is null or its memory is unmapped/invalid - glBindSampler(binding, 0); - glBindTextureUnit(binding, 0); - return; - } - const GLuint handle = view->GetTexture(texture.tic.x_source, texture.tic.y_source, - texture.tic.z_source, texture.tic.w_source); - glBindTextureUnit(binding, handle); - if (!view->GetSurfaceParams().IsBuffer()) { - glBindSampler(binding, sampler_cache.GetSampler(texture.tsc)); - } -} - -void RasterizerOpenGL::SetupDrawImages(std::size_t stage_index, Shader* shader) { - u32 binding = device.GetBaseBindings(stage_index).image; +void RasterizerOpenGL::SetupDrawImages(const Shader* shader, size_t stage_index) { + const bool via_header_index = + maxwell3d.regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex; for (const auto& entry : shader->GetEntries().images) { const auto shader_type = static_cast<ShaderType>(stage_index); - const auto tic = GetTextureInfo(maxwell3d, entry, shader_type).tic; - SetupImage(binding++, tic, entry); + const auto handle = GetTextureInfo(maxwell3d, via_header_index, entry, shader_type); + image_view_indices.push_back(handle.image); } } -void RasterizerOpenGL::SetupComputeImages(Shader* shader) { - u32 binding = 0; +void RasterizerOpenGL::SetupComputeImages(const Shader* shader) { + const bool via_header_index = kepler_compute.launch_description.linked_tsc; for (const auto& entry : shader->GetEntries().images) { - const auto tic = GetTextureInfo(kepler_compute, entry, ShaderType::Compute).tic; - SetupImage(binding++, tic, entry); + const auto handle = + GetTextureInfo(kepler_compute, via_header_index, entry, ShaderType::Compute); + image_view_indices.push_back(handle.image); } } -void RasterizerOpenGL::SetupImage(u32 binding, const Tegra::Texture::TICEntry& tic, - const ImageEntry& entry) { - const auto view = texture_cache.GetImageSurface(tic, entry); - if (!view) { - glBindImageTexture(binding, 0, 0, GL_FALSE, 0, GL_READ_ONLY, GL_R8); - return; - } - if (entry.is_written) { - view->MarkAsModified(texture_cache.Tick()); - } - const GLuint handle = view->GetTexture(tic.x_source, tic.y_source, tic.z_source, tic.w_source); - glBindImageTexture(binding, handle, 0, GL_TRUE, 0, GL_READ_WRITE, view->GetFormat()); -} - void RasterizerOpenGL::SyncViewport() { auto& flags = maxwell3d.dirty.flags; const auto& regs = maxwell3d.regs; @@ -1157,7 +1160,7 @@ void RasterizerOpenGL::SyncViewport() { flags[Dirty::ClipControl] = false; bool flip_y = false; - if (regs.viewport_transform[0].scale_y < 0.0) { + if (regs.viewport_transform[0].scale_y < 0.0f) { flip_y = !flip_y; } if (regs.screen_y_control.y_negate != 0) { @@ -1527,17 +1530,9 @@ void RasterizerOpenGL::SyncPointState() { flags[Dirty::PointSize] = false; oglEnable(GL_POINT_SPRITE, maxwell3d.regs.point_sprite_enable); + oglEnable(GL_PROGRAM_POINT_SIZE, maxwell3d.regs.vp_point_size.enable); - if (maxwell3d.regs.vp_point_size.enable) { - // By definition of GL_POINT_SIZE, it only matters if GL_PROGRAM_POINT_SIZE is disabled. - glEnable(GL_PROGRAM_POINT_SIZE); - return; - } - - // Limit the point size to 1 since nouveau sometimes sets a point size of 0 (and that's invalid - // in OpenGL). glPointSize(std::max(1.0f, maxwell3d.regs.point_size)); - glDisable(GL_PROGRAM_POINT_SIZE); } void RasterizerOpenGL::SyncLineState() { @@ -1580,10 +1575,6 @@ void RasterizerOpenGL::SyncAlphaTest() { flags[Dirty::AlphaTest] = false; const auto& regs = maxwell3d.regs; - if (regs.alpha_test_enabled && regs.rt_control.count > 1) { - LOG_WARNING(Render_OpenGL, "Alpha testing with more than one render target is not tested"); - } - if (regs.alpha_test_enabled) { glEnable(GL_ALPHA_TEST); glAlphaFunc(MaxwellToGL::ComparisonOp(regs.alpha_test_func), regs.alpha_test_ref); diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index f451404b2..82e03e677 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -7,12 +7,13 @@ #include <array> #include <atomic> #include <cstddef> -#include <map> #include <memory> #include <optional> #include <tuple> #include <utility> +#include <boost/container/static_vector.hpp> + #include <glad/glad.h> #include "common/common_types.h" @@ -23,16 +24,14 @@ #include "video_core/renderer_opengl/gl_buffer_cache.h" #include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_fence_manager.h" -#include "video_core/renderer_opengl/gl_framebuffer_cache.h" #include "video_core/renderer_opengl/gl_query_cache.h" #include "video_core/renderer_opengl/gl_resource_manager.h" -#include "video_core/renderer_opengl/gl_sampler_cache.h" #include "video_core/renderer_opengl/gl_shader_cache.h" #include "video_core/renderer_opengl/gl_shader_decompiler.h" #include "video_core/renderer_opengl/gl_shader_manager.h" #include "video_core/renderer_opengl/gl_state_tracker.h" +#include "video_core/renderer_opengl/gl_stream_buffer.h" #include "video_core/renderer_opengl/gl_texture_cache.h" -#include "video_core/renderer_opengl/utils.h" #include "video_core/shader/async_shaders.h" #include "video_core/textures/texture.h" @@ -51,14 +50,21 @@ class MemoryManager; namespace OpenGL { struct ScreenInfo; -struct DrawParameters; +struct ShaderEntries; + +struct BindlessSSBO { + GLuint64EXT address; + GLsizei length; + GLsizei padding; +}; +static_assert(sizeof(BindlessSSBO) * CHAR_BIT == 128); class RasterizerOpenGL : public VideoCore::RasterizerAccelerated { public: - explicit RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu, - Core::Memory::Memory& cpu_memory, const Device& device, - ScreenInfo& screen_info, ProgramManager& program_manager, - StateTracker& state_tracker); + explicit RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_, + Core::Memory::Memory& cpu_memory_, const Device& device_, + ScreenInfo& screen_info_, ProgramManager& program_manager_, + StateTracker& state_tracker_); ~RasterizerOpenGL() override; void Draw(bool is_indexed, bool is_instanced) override; @@ -72,15 +78,18 @@ public: void InvalidateRegion(VAddr addr, u64 size) override; void OnCPUWrite(VAddr addr, u64 size) override; void SyncGuestHost() override; + void UnmapMemory(VAddr addr, u64 size) override; void SignalSemaphore(GPUVAddr addr, u32 value) override; void SignalSyncPoint(u32 value) override; void ReleaseFences() override; void FlushAndInvalidateRegion(VAddr addr, u64 size) override; void WaitForIdle() override; + void FragmentBarrier() override; + void TiledCacheBarrier() override; void FlushCommands() override; void TickFrame() override; - bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, - const Tegra::Engines::Fermi2D::Regs::Surface& dst, + bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Surface& src, + const Tegra::Engines::Fermi2D::Surface& dst, const Tegra::Engines::Fermi2D::Config& copy_config) override; bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr, u32 pixel_stride) override; @@ -101,11 +110,14 @@ public: } private: - /// Configures the color and depth framebuffer states. - void ConfigureFramebuffers(); + static constexpr size_t MAX_TEXTURES = 192; + static constexpr size_t MAX_IMAGES = 48; + static constexpr size_t MAX_IMAGE_VIEWS = MAX_TEXTURES + MAX_IMAGES; + + void BindComputeTextures(Shader* kernel); - /// Configures the color and depth framebuffer for clearing. - void ConfigureClearFramebuffer(bool using_color, bool using_depth_stencil); + void BindTextures(const ShaderEntries& entries, GLuint base_texture, GLuint base_image, + size_t& image_view_index, size_t& texture_index, size_t& image_index); /// Configures the current constbuffers to use for the draw command. void SetupDrawConstBuffers(std::size_t stage_index, Shader* shader); @@ -126,26 +138,19 @@ private: /// Configures a global memory buffer. void SetupGlobalMemory(u32 binding, const GlobalMemoryEntry& entry, GPUVAddr gpu_addr, - std::size_t size, GLuint64EXT* pointer); + size_t size, BindlessSSBO* ssbo); /// Configures the current textures to use for the draw command. - void SetupDrawTextures(std::size_t stage_index, Shader* shader); + void SetupDrawTextures(const Shader* shader, size_t stage_index); /// Configures the textures used in a compute shader. - void SetupComputeTextures(Shader* kernel); - - /// Configures a texture. - void SetupTexture(u32 binding, const Tegra::Texture::FullTextureInfo& texture, - const SamplerEntry& entry); + void SetupComputeTextures(const Shader* kernel); /// Configures images in a graphics shader. - void SetupDrawImages(std::size_t stage_index, Shader* shader); + void SetupDrawImages(const Shader* shader, size_t stage_index); /// Configures images in a compute shader. - void SetupComputeImages(Shader* shader); - - /// Configures an image. - void SetupImage(u32 binding, const Tegra::Texture::TICEntry& tic, const ImageEntry& entry); + void SetupComputeImages(const Shader* shader); /// Syncs the viewport and depth range to match the guest state void SyncViewport(); @@ -220,9 +225,6 @@ private: /// End a transform feedback void EndTransformFeedback(); - /// Check for extension that are not strictly required but are needed for correct emulation - void CheckExtensions(); - std::size_t CalculateVertexArraysSize() const; std::size_t CalculateIndexBufferSize() const; @@ -235,7 +237,7 @@ private: GLintptr SetupIndexBuffer(); - void SetupShaders(GLenum primitive_mode); + void SetupShaders(); Tegra::GPU& gpu; Tegra::Engines::Maxwell3D& maxwell3d; @@ -247,19 +249,21 @@ private: ProgramManager& program_manager; StateTracker& state_tracker; - TextureCacheOpenGL texture_cache; + OGLStreamBuffer stream_buffer; + TextureCacheRuntime texture_cache_runtime; + TextureCache texture_cache; ShaderCacheOpenGL shader_cache; - SamplerCacheOpenGL sampler_cache; - FramebufferCacheOpenGL framebuffer_cache; QueryCache query_cache; OGLBufferCache buffer_cache; FenceManagerOpenGL fence_manager; VideoCommon::Shader::AsyncShaders async_shaders; - static constexpr std::size_t STREAM_BUFFER_SIZE = 128 * 1024 * 1024; - - GLint vertex_binding = 0; + boost::container::static_vector<u32, MAX_IMAGE_VIEWS> image_view_indices; + std::array<ImageViewId, MAX_IMAGE_VIEWS> image_view_ids; + boost::container::static_vector<GLuint, MAX_TEXTURES> sampler_handles; + std::array<GLuint, MAX_TEXTURES> texture_handles; + std::array<GLuint, MAX_IMAGES> image_handles; std::array<OGLBuffer, Tegra::Engines::Maxwell3D::Regs::NumTransformFeedbackBuffers> transform_feedback_buffers; @@ -273,7 +277,7 @@ private: std::size_t current_cbuf = 0; OGLBuffer unified_uniform_buffer; - /// Number of commands queued to the OpenGL driver. Reseted on flush. + /// Number of commands queued to the OpenGL driver. Resetted on flush. std::size_t num_queued_commands = 0; u32 last_clip_distance_mask = 0; diff --git a/src/video_core/renderer_opengl/gl_resource_manager.cpp b/src/video_core/renderer_opengl/gl_resource_manager.cpp index 0ebcec427..0e34a0f20 100644 --- a/src/video_core/renderer_opengl/gl_resource_manager.cpp +++ b/src/video_core/renderer_opengl/gl_resource_manager.cpp @@ -71,7 +71,7 @@ void OGLSampler::Create() { return; MICROPROFILE_SCOPE(OpenGL_ResourceCreation); - glGenSamplers(1, &handle); + glCreateSamplers(1, &handle); } void OGLSampler::Release() { diff --git a/src/video_core/renderer_opengl/gl_sampler_cache.cpp b/src/video_core/renderer_opengl/gl_sampler_cache.cpp deleted file mode 100644 index 5c174879a..000000000 --- a/src/video_core/renderer_opengl/gl_sampler_cache.cpp +++ /dev/null @@ -1,52 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "common/logging/log.h" -#include "video_core/renderer_opengl/gl_resource_manager.h" -#include "video_core/renderer_opengl/gl_sampler_cache.h" -#include "video_core/renderer_opengl/maxwell_to_gl.h" - -namespace OpenGL { - -SamplerCacheOpenGL::SamplerCacheOpenGL() = default; - -SamplerCacheOpenGL::~SamplerCacheOpenGL() = default; - -OGLSampler SamplerCacheOpenGL::CreateSampler(const Tegra::Texture::TSCEntry& tsc) const { - OGLSampler sampler; - sampler.Create(); - - const GLuint sampler_id{sampler.handle}; - glSamplerParameteri( - sampler_id, GL_TEXTURE_MAG_FILTER, - MaxwellToGL::TextureFilterMode(tsc.mag_filter, Tegra::Texture::TextureMipmapFilter::None)); - glSamplerParameteri(sampler_id, GL_TEXTURE_MIN_FILTER, - MaxwellToGL::TextureFilterMode(tsc.min_filter, tsc.mipmap_filter)); - glSamplerParameteri(sampler_id, GL_TEXTURE_WRAP_S, MaxwellToGL::WrapMode(tsc.wrap_u)); - glSamplerParameteri(sampler_id, GL_TEXTURE_WRAP_T, MaxwellToGL::WrapMode(tsc.wrap_v)); - glSamplerParameteri(sampler_id, GL_TEXTURE_WRAP_R, MaxwellToGL::WrapMode(tsc.wrap_p)); - glSamplerParameteri(sampler_id, GL_TEXTURE_COMPARE_MODE, - tsc.depth_compare_enabled == 1 ? GL_COMPARE_REF_TO_TEXTURE : GL_NONE); - glSamplerParameteri(sampler_id, GL_TEXTURE_COMPARE_FUNC, - MaxwellToGL::DepthCompareFunc(tsc.depth_compare_func)); - glSamplerParameterfv(sampler_id, GL_TEXTURE_BORDER_COLOR, tsc.GetBorderColor().data()); - glSamplerParameterf(sampler_id, GL_TEXTURE_MIN_LOD, tsc.GetMinLod()); - glSamplerParameterf(sampler_id, GL_TEXTURE_MAX_LOD, tsc.GetMaxLod()); - glSamplerParameterf(sampler_id, GL_TEXTURE_LOD_BIAS, tsc.GetLodBias()); - if (GLAD_GL_ARB_texture_filter_anisotropic) { - glSamplerParameterf(sampler_id, GL_TEXTURE_MAX_ANISOTROPY, tsc.GetMaxAnisotropy()); - } else if (GLAD_GL_EXT_texture_filter_anisotropic) { - glSamplerParameterf(sampler_id, GL_TEXTURE_MAX_ANISOTROPY_EXT, tsc.GetMaxAnisotropy()); - } else { - LOG_WARNING(Render_OpenGL, "Anisotropy not supported by host GPU driver"); - } - - return sampler; -} - -GLuint SamplerCacheOpenGL::ToSamplerType(const OGLSampler& sampler) const { - return sampler.handle; -} - -} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_sampler_cache.h b/src/video_core/renderer_opengl/gl_sampler_cache.h deleted file mode 100644 index 34ee37f00..000000000 --- a/src/video_core/renderer_opengl/gl_sampler_cache.h +++ /dev/null @@ -1,25 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <glad/glad.h> - -#include "video_core/renderer_opengl/gl_resource_manager.h" -#include "video_core/sampler_cache.h" - -namespace OpenGL { - -class SamplerCacheOpenGL final : public VideoCommon::SamplerCache<GLuint, OGLSampler> { -public: - explicit SamplerCacheOpenGL(); - ~SamplerCacheOpenGL(); - -protected: - OGLSampler CreateSampler(const Tegra::Texture::TSCEntry& tsc) const override; - - GLuint ToSamplerType(const OGLSampler& sampler) const override; -}; - -} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index bd56bed0c..d4841fdb7 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -27,7 +27,6 @@ #include "video_core/renderer_opengl/gl_shader_decompiler.h" #include "video_core/renderer_opengl/gl_shader_disk_cache.h" #include "video_core/renderer_opengl/gl_state_tracker.h" -#include "video_core/renderer_opengl/utils.h" #include "video_core/shader/memory_util.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" @@ -198,10 +197,10 @@ ProgramSharedPtr BuildShader(const Device& device, ShaderType shader_type, u64 u return program; } -Shader::Shader(std::shared_ptr<VideoCommon::Shader::Registry> registry_, ShaderEntries entries_, - ProgramSharedPtr program_, bool is_built) +Shader::Shader(std::shared_ptr<Registry> registry_, ShaderEntries entries_, + ProgramSharedPtr program_, bool is_built_) : registry{std::move(registry_)}, entries{std::move(entries_)}, program{std::move(program_)}, - is_built(is_built) { + is_built{is_built_} { handle = program->assembly_program.handle; if (handle == 0) { handle = program->source_program.handle; @@ -318,14 +317,13 @@ std::unique_ptr<Shader> Shader::CreateFromCache(const ShaderParameters& params, precompiled_shader.registry, precompiled_shader.entries, precompiled_shader.program)); } -ShaderCacheOpenGL::ShaderCacheOpenGL(RasterizerOpenGL& rasterizer, +ShaderCacheOpenGL::ShaderCacheOpenGL(RasterizerOpenGL& rasterizer_, Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_, Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::Engines::KeplerCompute& kepler_compute_, Tegra::MemoryManager& gpu_memory_, const Device& device_) - : VideoCommon::ShaderCache<Shader>{rasterizer}, emu_window{emu_window_}, gpu{gpu_}, - gpu_memory{gpu_memory_}, maxwell3d{maxwell3d_}, - kepler_compute{kepler_compute_}, device{device_} {} + : ShaderCache{rasterizer_}, emu_window{emu_window_}, gpu{gpu_}, gpu_memory{gpu_memory_}, + maxwell3d{maxwell3d_}, kepler_compute{kepler_compute_}, device{device_} {} ShaderCacheOpenGL::~ShaderCacheOpenGL() = default; @@ -460,7 +458,7 @@ void ShaderCacheOpenGL::LoadDiskCache(u64 title_id, const std::atomic_bool& stop ProgramSharedPtr ShaderCacheOpenGL::GeneratePrecompiledProgram( const ShaderDiskCacheEntry& entry, const ShaderDiskCachePrecompiled& precompiled_entry, const std::unordered_set<GLenum>& supported_formats) { - if (supported_formats.find(precompiled_entry.binary_format) == supported_formats.end()) { + if (!supported_formats.contains(precompiled_entry.binary_format)) { LOG_INFO(Render_OpenGL, "Precompiled cache entry with unsupported format, removing"); return {}; } diff --git a/src/video_core/renderer_opengl/gl_shader_cache.h b/src/video_core/renderer_opengl/gl_shader_cache.h index 1708af06a..2aed0697e 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.h +++ b/src/video_core/renderer_opengl/gl_shader_cache.h @@ -108,7 +108,7 @@ public: private: explicit Shader(std::shared_ptr<VideoCommon::Shader::Registry> registry, ShaderEntries entries, - ProgramSharedPtr program, bool is_built = true); + ProgramSharedPtr program, bool is_built_ = true); std::shared_ptr<VideoCommon::Shader::Registry> registry; ShaderEntries entries; @@ -119,10 +119,11 @@ private: class ShaderCacheOpenGL final : public VideoCommon::ShaderCache<Shader> { public: - explicit ShaderCacheOpenGL(RasterizerOpenGL& rasterizer, Core::Frontend::EmuWindow& emu_window, - Tegra::GPU& gpu, Tegra::Engines::Maxwell3D& maxwell3d, - Tegra::Engines::KeplerCompute& kepler_compute, - Tegra::MemoryManager& gpu_memory, const Device& device); + explicit ShaderCacheOpenGL(RasterizerOpenGL& rasterizer_, + Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu, + Tegra::Engines::Maxwell3D& maxwell3d_, + Tegra::Engines::KeplerCompute& kepler_compute_, + Tegra::MemoryManager& gpu_memory_, const Device& device_); ~ShaderCacheOpenGL() override; /// Loads disk cache for the current game diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index bbb8fb095..2e1fa252d 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -38,11 +38,9 @@ using Tegra::Shader::IpaSampleMode; using Tegra::Shader::PixelImap; using Tegra::Shader::Register; using Tegra::Shader::TextureType; -using VideoCommon::Shader::BuildTransformFeedback; -using VideoCommon::Shader::Registry; -using namespace std::string_literals; using namespace VideoCommon::Shader; +using namespace std::string_literals; using Maxwell = Tegra::Engines::Maxwell3D::Regs; using Operation = const OperationNode&; @@ -131,7 +129,7 @@ private: class Expression final { public: - Expression(std::string code, Type type) : code{std::move(code)}, type{type} { + Expression(std::string code_, Type type_) : code{std::move(code_)}, type{type_} { ASSERT(type != Type::Void); } Expression() : type{Type::Void} {} @@ -148,8 +146,8 @@ public: ASSERT(type == Type::Void); } - std::string As(Type type) const { - switch (type) { + std::string As(Type type_) const { + switch (type_) { case Type::Bool: return AsBool(); case Type::Bool2: @@ -316,7 +314,7 @@ std::pair<const char*, u32> GetPrimitiveDescription(Maxwell::PrimitiveTopology t case Maxwell::PrimitiveTopology::TriangleStripAdjacency: return {"triangles_adjacency", 6}; default: - UNIMPLEMENTED_MSG("topology={}", static_cast<int>(topology)); + UNIMPLEMENTED_MSG("topology={}", topology); return {"points", 1}; } } @@ -342,7 +340,7 @@ std::string GetTopologyName(Tegra::Shader::OutputTopology topology) { case Tegra::Shader::OutputTopology::TriangleStrip: return "triangle_strip"; default: - UNIMPLEMENTED_MSG("Unknown output topology: {}", static_cast<u32>(topology)); + UNIMPLEMENTED_MSG("Unknown output topology: {}", topology); return "points"; } } @@ -418,11 +416,12 @@ struct GenericVaryingDescription { class GLSLDecompiler final { public: - explicit GLSLDecompiler(const Device& device, const ShaderIR& ir, const Registry& registry, - ShaderType stage, std::string_view identifier, std::string_view suffix) - : device{device}, ir{ir}, registry{registry}, stage{stage}, identifier{identifier}, - suffix{suffix}, header{ir.GetHeader()}, use_unified_uniforms{ - UseUnifiedUniforms(device, ir, stage)} { + explicit GLSLDecompiler(const Device& device_, const ShaderIR& ir_, const Registry& registry_, + ShaderType stage_, std::string_view identifier_, + std::string_view suffix_) + : device{device_}, ir{ir_}, registry{registry_}, stage{stage_}, identifier{identifier_}, + suffix{suffix_}, header{ir.GetHeader()}, use_unified_uniforms{ + UseUnifiedUniforms(device_, ir_, stage_)} { if (stage != ShaderType::Compute) { transform_feedback = BuildTransformFeedback(registry.GetGraphicsInfo()); } @@ -744,7 +743,7 @@ private: case PixelImap::Unused: break; } - UNIMPLEMENTED_MSG("Unknown attribute usage index={}", static_cast<int>(attribute)); + UNIMPLEMENTED_MSG("Unknown attribute usage index={}", attribute); return {}; } @@ -777,16 +776,16 @@ private: name = "gs_" + name + "[]"; } - std::string suffix; + std::string suffix_; if (stage == ShaderType::Fragment) { const auto input_mode{header.ps.GetPixelImap(location)}; if (input_mode == PixelImap::Unused) { return; } - suffix = GetInputFlags(input_mode); + suffix_ = GetInputFlags(input_mode); } - code.AddLine("layout (location = {}) {} in vec4 {};", location, suffix, name); + code.AddLine("layout (location = {}) {} in vec4 {};", location, suffix_, name); } void DeclareOutputAttributes() { @@ -877,7 +876,7 @@ private: } u32 binding = device.GetBaseBindings(stage).uniform_buffer; - for (const auto [index, info] : ir.GetConstantBuffers()) { + for (const auto& [index, info] : ir.GetConstantBuffers()) { const u32 num_elements = Common::AlignUp(info.GetSize(), 4) / 4; const u32 size = info.IsIndirect() ? MAX_CONSTBUFFER_ELEMENTS : num_elements; code.AddLine("layout (std140, binding = {}) uniform {} {{", binding++, @@ -1251,7 +1250,7 @@ private: } break; } - UNIMPLEMENTED_MSG("Unhandled input attribute: {}", static_cast<u32>(attribute)); + UNIMPLEMENTED_MSG("Unhandled input attribute: {}", attribute); return {"0", Type::Int}; } @@ -1331,7 +1330,7 @@ private: GetSwizzle(element)), Type::Float}}; } - UNIMPLEMENTED_MSG("Unhandled output attribute: {}", static_cast<u32>(attribute)); + UNIMPLEMENTED_MSG("Unhandled output attribute: {}", attribute); return std::nullopt; } } @@ -2056,15 +2055,19 @@ private: } Expression Texture(Operation operation) { - const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); - ASSERT(meta); - - std::string expr = GenerateTexture( - operation, "", {TextureOffset{}, TextureArgument{Type::Float, meta->bias}}); - if (meta->sampler.is_shadow) { - expr = "vec4(" + expr + ')'; + const auto meta = std::get<MetaTexture>(operation.GetMeta()); + const bool separate_dc = meta.sampler.type == TextureType::TextureCube && + meta.sampler.is_array && meta.sampler.is_shadow; + // TODO: Replace this with an array and make GenerateTexture use C++20 std::span + const std::vector<TextureIR> extras{ + TextureOffset{}, + TextureArgument{Type::Float, meta.bias}, + }; + std::string expr = GenerateTexture(operation, "", extras, separate_dc); + if (meta.sampler.is_shadow) { + expr = fmt::format("vec4({})", expr); } - return {expr + GetSwizzle(meta->element), Type::Float}; + return {expr + GetSwizzle(meta.element), Type::Float}; } Expression TextureLod(Operation operation) { @@ -2096,13 +2099,13 @@ private: const auto type = meta.sampler.is_shadow ? Type::Float : Type::Int; const bool separate_dc = meta.sampler.is_shadow; - std::vector<TextureIR> ir; + std::vector<TextureIR> ir_; if (meta.sampler.is_shadow) { - ir = {TextureOffset{}}; + ir_ = {TextureOffset{}}; } else { - ir = {TextureOffset{}, TextureArgument{type, meta.component}}; + ir_ = {TextureOffset{}, TextureArgument{type, meta.component}}; } - return {GenerateTexture(operation, "Gather", ir, separate_dc) + GetSwizzle(meta.element), + return {GenerateTexture(operation, "Gather", ir_, separate_dc) + GetSwizzle(meta.element), Type::Float}; } @@ -2748,11 +2751,11 @@ private: } } - std::string GetSampler(const Sampler& sampler) const { + std::string GetSampler(const SamplerEntry& sampler) const { return AppendSuffix(sampler.index, "sampler"); } - std::string GetImage(const Image& image) const { + std::string GetImage(const ImageEntry& image) const { return AppendSuffix(image.index, "image"); } @@ -2797,7 +2800,7 @@ std::string GetFlowVariable(u32 index) { class ExprDecompiler { public: - explicit ExprDecompiler(GLSLDecompiler& decomp) : decomp{decomp} {} + explicit ExprDecompiler(GLSLDecompiler& decomp_) : decomp{decomp_} {} void operator()(const ExprAnd& expr) { inner += '('; @@ -2852,7 +2855,7 @@ private: class ASTDecompiler { public: - explicit ASTDecompiler(GLSLDecompiler& decomp) : decomp{decomp} {} + explicit ASTDecompiler(GLSLDecompiler& decomp_) : decomp{decomp_} {} void operator()(const ASTProgram& ast) { ASTNode current = ast.nodes.GetFirst(); diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.h b/src/video_core/renderer_opengl/gl_shader_decompiler.h index 451c9689a..be68994bb 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.h +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.h @@ -20,13 +20,13 @@ namespace OpenGL { class Device; using Maxwell = Tegra::Engines::Maxwell3D::Regs; -using SamplerEntry = VideoCommon::Shader::Sampler; -using ImageEntry = VideoCommon::Shader::Image; +using SamplerEntry = VideoCommon::Shader::SamplerEntry; +using ImageEntry = VideoCommon::Shader::ImageEntry; class ConstBufferEntry : public VideoCommon::Shader::ConstBuffer { public: - explicit ConstBufferEntry(u32 max_offset, bool is_indirect, u32 index) - : VideoCommon::Shader::ConstBuffer{max_offset, is_indirect}, index{index} {} + explicit ConstBufferEntry(u32 max_offset_, bool is_indirect_, u32 index_) + : ConstBuffer{max_offset_, is_indirect_}, index{index_} {} u32 GetIndex() const { return index; @@ -37,10 +37,10 @@ private: }; struct GlobalMemoryEntry { - constexpr explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset, bool is_read, - bool is_written) - : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset}, is_read{is_read}, is_written{ - is_written} {} + constexpr explicit GlobalMemoryEntry(u32 cbuf_index_, u32 cbuf_offset_, bool is_read_, + bool is_written_) + : cbuf_index{cbuf_index_}, cbuf_offset{cbuf_offset_}, is_read{is_read_}, is_written{ + is_written_} {} u32 cbuf_index = 0; u32 cbuf_offset = 0; diff --git a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp index 166ee34e1..955b2abc4 100644 --- a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp @@ -317,8 +317,7 @@ std::optional<std::vector<ShaderDiskCachePrecompiled>> ShaderDiskCacheOpenGL::Lo return std::nullopt; } } - - return std::move(entries); + return entries; } void ShaderDiskCacheOpenGL::InvalidateTransferable() { @@ -344,7 +343,7 @@ void ShaderDiskCacheOpenGL::SaveEntry(const ShaderDiskCacheEntry& entry) { } const u64 id = entry.unique_identifier; - if (stored_transferable.find(id) != stored_transferable.end()) { + if (stored_transferable.contains(id)) { // The shader already exists return; } diff --git a/src/video_core/renderer_opengl/gl_shader_manager.cpp b/src/video_core/renderer_opengl/gl_shader_manager.cpp index 691c6c79b..553e6e8d6 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.cpp +++ b/src/video_core/renderer_opengl/gl_shader_manager.cpp @@ -83,6 +83,21 @@ void ProgramManager::RestoreGuestPipeline() { } } +void ProgramManager::BindHostCompute(GLuint program) { + if (use_assembly_programs) { + glDisable(GL_COMPUTE_PROGRAM_NV); + } + glUseProgram(program); + is_graphics_bound = false; +} + +void ProgramManager::RestoreGuestCompute() { + if (use_assembly_programs) { + glEnable(GL_COMPUTE_PROGRAM_NV); + glUseProgram(0); + } +} + void ProgramManager::UseVertexShader(GLuint program) { if (use_assembly_programs) { BindProgram(GL_VERTEX_PROGRAM_NV, program, current_state.vertex, vertex_enabled); diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h index 950e0dfcb..ad42cce74 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.h +++ b/src/video_core/renderer_opengl/gl_shader_manager.h @@ -45,6 +45,12 @@ public: /// Rewinds BindHostPipeline state changes. void RestoreGuestPipeline(); + /// Binds an OpenGL GLSL program object unsynchronized with the guest state. + void BindHostCompute(GLuint program); + + /// Rewinds BindHostCompute state changes. + void RestoreGuestCompute(); + void UseVertexShader(GLuint program); void UseGeometryShader(GLuint program); void UseFragmentShader(GLuint program); diff --git a/src/video_core/renderer_opengl/gl_state_tracker.cpp b/src/video_core/renderer_opengl/gl_state_tracker.cpp index 6bcf831f2..60e6fa39f 100644 --- a/src/video_core/renderer_opengl/gl_state_tracker.cpp +++ b/src/video_core/renderer_opengl/gl_state_tracker.cpp @@ -13,7 +13,7 @@ #include "video_core/renderer_opengl/gl_state_tracker.h" #define OFF(field_name) MAXWELL3D_REG_INDEX(field_name) -#define NUM(field_name) (sizeof(Maxwell3D::Regs::field_name) / sizeof(u32)) +#define NUM(field_name) (sizeof(Maxwell3D::Regs::field_name) / (sizeof(u32))) namespace OpenGL { @@ -249,4 +249,11 @@ StateTracker::StateTracker(Tegra::GPU& gpu) : flags{gpu.Maxwell3D().dirty.flags} } } +void StateTracker::InvalidateStreamBuffer() { + flags[Dirty::VertexBuffers] = true; + for (int index = Dirty::VertexBuffer0; index <= Dirty::VertexBuffer31; ++index) { + flags[index] = true; + } +} + } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_state_tracker.h b/src/video_core/renderer_opengl/gl_state_tracker.h index 9d127548f..574615d3c 100644 --- a/src/video_core/renderer_opengl/gl_state_tracker.h +++ b/src/video_core/renderer_opengl/gl_state_tracker.h @@ -92,6 +92,8 @@ class StateTracker { public: explicit StateTracker(Tegra::GPU& gpu); + void InvalidateStreamBuffer(); + void BindIndexBuffer(GLuint new_index_buffer) { if (index_buffer == new_index_buffer) { return; @@ -100,6 +102,14 @@ public: glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, new_index_buffer); } + void BindFramebuffer(GLuint new_framebuffer) { + if (framebuffer == new_framebuffer) { + return; + } + framebuffer = new_framebuffer; + glBindFramebuffer(GL_DRAW_FRAMEBUFFER, framebuffer); + } + void NotifyScreenDrawVertexArray() { flags[OpenGL::Dirty::VertexFormats] = true; flags[OpenGL::Dirty::VertexFormat0 + 0] = true; @@ -129,9 +139,9 @@ public: flags[OpenGL::Dirty::Scissor0] = true; } - void NotifyColorMask0() { + void NotifyColorMask(size_t index) { flags[OpenGL::Dirty::ColorMasks] = true; - flags[OpenGL::Dirty::ColorMask0] = true; + flags[OpenGL::Dirty::ColorMask0 + index] = true; } void NotifyBlend0() { @@ -190,6 +200,7 @@ public: private: Tegra::Engines::Maxwell3D::DirtyState::Flags& flags; + GLuint framebuffer = 0; GLuint index_buffer = 0; }; diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.cpp b/src/video_core/renderer_opengl/gl_stream_buffer.cpp index 887995cf4..e0819cdf2 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.cpp +++ b/src/video_core/renderer_opengl/gl_stream_buffer.cpp @@ -9,6 +9,7 @@ #include "common/assert.h" #include "common/microprofile.h" #include "video_core/renderer_opengl/gl_device.h" +#include "video_core/renderer_opengl/gl_state_tracker.h" #include "video_core/renderer_opengl/gl_stream_buffer.h" MICROPROFILE_DEFINE(OpenGL_StreamBuffer, "OpenGL", "Stream Buffer Orphaning", @@ -16,24 +17,14 @@ MICROPROFILE_DEFINE(OpenGL_StreamBuffer, "OpenGL", "Stream Buffer Orphaning", namespace OpenGL { -OGLStreamBuffer::OGLStreamBuffer(const Device& device, GLsizeiptr size, bool vertex_data_usage) - : buffer_size(size) { +OGLStreamBuffer::OGLStreamBuffer(const Device& device, StateTracker& state_tracker_) + : state_tracker{state_tracker_} { gl_buffer.Create(); - GLsizeiptr allocate_size = size; - if (vertex_data_usage) { - // On AMD GPU there is a strange crash in indexed drawing. The crash happens when the buffer - // read position is near the end and is an out-of-bound access to the vertex buffer. This is - // probably a bug in the driver and is related to the usage of vec3<byte> attributes in the - // vertex array. Doubling the allocation size for the vertex buffer seems to avoid the - // crash. - allocate_size *= 2; - } - static constexpr GLbitfield flags = GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT; - glNamedBufferStorage(gl_buffer.handle, allocate_size, nullptr, flags); + glNamedBufferStorage(gl_buffer.handle, BUFFER_SIZE, nullptr, flags); mapped_ptr = static_cast<u8*>( - glMapNamedBufferRange(gl_buffer.handle, 0, buffer_size, flags | GL_MAP_FLUSH_EXPLICIT_BIT)); + glMapNamedBufferRange(gl_buffer.handle, 0, BUFFER_SIZE, flags | GL_MAP_FLUSH_EXPLICIT_BIT)); if (device.UseAssemblyShaders() || device.HasVertexBufferUnifiedMemory()) { glMakeNamedBufferResidentNV(gl_buffer.handle, GL_READ_ONLY); @@ -46,25 +37,24 @@ OGLStreamBuffer::~OGLStreamBuffer() { gl_buffer.Release(); } -std::tuple<u8*, GLintptr, bool> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr alignment) { - ASSERT(size <= buffer_size); - ASSERT(alignment <= buffer_size); +std::pair<u8*, GLintptr> OGLStreamBuffer::Map(GLsizeiptr size, GLintptr alignment) { + ASSERT(size <= BUFFER_SIZE); + ASSERT(alignment <= BUFFER_SIZE); mapped_size = size; if (alignment > 0) { buffer_pos = Common::AlignUp<std::size_t>(buffer_pos, alignment); } - bool invalidate = false; - if (buffer_pos + size > buffer_size) { + if (buffer_pos + size > BUFFER_SIZE) { MICROPROFILE_SCOPE(OpenGL_StreamBuffer); glInvalidateBufferData(gl_buffer.handle); + state_tracker.InvalidateStreamBuffer(); buffer_pos = 0; - invalidate = true; } - return std::make_tuple(mapped_ptr + buffer_pos, buffer_pos, invalidate); + return std::make_pair(mapped_ptr + buffer_pos, buffer_pos); } void OGLStreamBuffer::Unmap(GLsizeiptr size) { diff --git a/src/video_core/renderer_opengl/gl_stream_buffer.h b/src/video_core/renderer_opengl/gl_stream_buffer.h index 307a67113..dd9cf67eb 100644 --- a/src/video_core/renderer_opengl/gl_stream_buffer.h +++ b/src/video_core/renderer_opengl/gl_stream_buffer.h @@ -4,29 +4,31 @@ #pragma once -#include <tuple> +#include <utility> + #include <glad/glad.h> + #include "common/common_types.h" #include "video_core/renderer_opengl/gl_resource_manager.h" namespace OpenGL { class Device; +class StateTracker; class OGLStreamBuffer : private NonCopyable { public: - explicit OGLStreamBuffer(const Device& device, GLsizeiptr size, bool vertex_data_usage); + explicit OGLStreamBuffer(const Device& device, StateTracker& state_tracker_); ~OGLStreamBuffer(); /* * Allocates a linear chunk of memory in the GPU buffer with at least "size" bytes * and the optional alignment requirement. * If the buffer is full, the whole buffer is reallocated which invalidates old chunks. - * The return values are the pointer to the new chunk, the offset within the buffer, - * and the invalidation flag for previous chunks. + * The return values are the pointer to the new chunk, and the offset within the buffer. * The actual used size must be specified on unmapping the chunk. */ - std::tuple<u8*, GLintptr, bool> Map(GLsizeiptr size, GLintptr alignment = 0); + std::pair<u8*, GLintptr> Map(GLsizeiptr size, GLintptr alignment = 0); void Unmap(GLsizeiptr size); @@ -39,15 +41,18 @@ public: } GLsizeiptr Size() const noexcept { - return buffer_size; + return BUFFER_SIZE; } private: + static constexpr GLsizeiptr BUFFER_SIZE = 256 * 1024 * 1024; + + StateTracker& state_tracker; + OGLBuffer gl_buffer; GLuint64EXT gpu_address = 0; GLintptr buffer_pos = 0; - GLsizeiptr buffer_size = 0; GLsizeiptr mapped_size = 0; u8* mapped_ptr = nullptr; }; diff --git a/src/video_core/renderer_opengl/gl_texture_cache.cpp b/src/video_core/renderer_opengl/gl_texture_cache.cpp index a863ef218..546cb6d00 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.cpp +++ b/src/video_core/renderer_opengl/gl_texture_cache.cpp @@ -2,37 +2,60 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "common/assert.h" -#include "common/bit_util.h" -#include "common/common_types.h" -#include "common/microprofile.h" -#include "common/scope_exit.h" -#include "core/core.h" -#include "video_core/morton.h" -#include "video_core/renderer_opengl/gl_resource_manager.h" +#include <algorithm> +#include <array> +#include <bit> +#include <string> + +#include <glad/glad.h> + +#include "video_core/renderer_opengl/gl_device.h" +#include "video_core/renderer_opengl/gl_shader_manager.h" #include "video_core/renderer_opengl/gl_state_tracker.h" #include "video_core/renderer_opengl/gl_texture_cache.h" -#include "video_core/renderer_opengl/utils.h" -#include "video_core/texture_cache/surface_base.h" +#include "video_core/renderer_opengl/maxwell_to_gl.h" +#include "video_core/renderer_opengl/util_shaders.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/format_lookup_table.h" +#include "video_core/texture_cache/samples_helper.h" #include "video_core/texture_cache/texture_cache.h" -#include "video_core/textures/convert.h" -#include "video_core/textures/texture.h" +#include "video_core/textures/decoders.h" namespace OpenGL { -using Tegra::Texture::SwizzleSource; -using VideoCore::MortonSwizzleMode; +namespace { +using Tegra::Texture::SwizzleSource; +using Tegra::Texture::TextureMipmapFilter; +using Tegra::Texture::TextureType; +using Tegra::Texture::TICEntry; +using Tegra::Texture::TSCEntry; +using VideoCommon::CalculateLevelStrideAlignment; +using VideoCommon::ImageCopy; +using VideoCommon::ImageFlagBits; +using VideoCommon::ImageType; +using VideoCommon::NUM_RT; +using VideoCommon::SamplesLog2; +using VideoCommon::SwizzleParameters; +using VideoCore::Surface::BytesPerBlock; +using VideoCore::Surface::IsPixelFormatASTC; +using VideoCore::Surface::IsPixelFormatSRGB; +using VideoCore::Surface::MaxPixelFormat; using VideoCore::Surface::PixelFormat; -using VideoCore::Surface::SurfaceTarget; using VideoCore::Surface::SurfaceType; -MICROPROFILE_DEFINE(OpenGL_Texture_Upload, "OpenGL", "Texture Upload", MP_RGB(128, 192, 128)); -MICROPROFILE_DEFINE(OpenGL_Texture_Download, "OpenGL", "Texture Download", MP_RGB(128, 192, 128)); -MICROPROFILE_DEFINE(OpenGL_Texture_Buffer_Copy, "OpenGL", "Texture Buffer Copy", - MP_RGB(128, 192, 128)); +struct CopyOrigin { + GLint level; + GLint x; + GLint y; + GLint z; +}; -namespace { +struct CopyRegion { + GLsizei width; + GLsizei height; + GLsizei depth; +}; struct FormatTuple { GLenum internal_format; @@ -40,7 +63,7 @@ struct FormatTuple { GLenum type = GL_NONE; }; -constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format_tuples = {{ +constexpr std::array<FormatTuple, MaxPixelFormat> FORMAT_TABLE = {{ {GL_RGBA8, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8_REV}, // A8B8G8R8_UNORM {GL_RGBA8_SNORM, GL_RGBA, GL_BYTE}, // A8B8G8R8_SNORM {GL_RGBA8I, GL_RGBA_INTEGER, GL_BYTE}, // A8B8G8R8_SINT @@ -103,72 +126,113 @@ constexpr std::array<FormatTuple, VideoCore::Surface::MaxPixelFormat> tex_format {GL_COMPRESSED_RGBA_ASTC_8x5_KHR}, // ASTC_2D_8X5_UNORM {GL_COMPRESSED_RGBA_ASTC_5x4_KHR}, // ASTC_2D_5X4_UNORM {GL_SRGB8_ALPHA8, GL_BGRA, GL_UNSIGNED_BYTE}, // B8G8R8A8_UNORM - // Compressed sRGB formats - {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT}, // BC1_RGBA_SRGB - {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT}, // BC2_SRGB - {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT}, // BC3_SRGB - {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM}, // BC7_SRGB - {GL_RGBA4, GL_RGBA, GL_UNSIGNED_SHORT_4_4_4_4_REV}, // A4B4G4R4_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_4x4_KHR}, // ASTC_2D_4X4_SRGB - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x8_KHR}, // ASTC_2D_8X8_SRGB - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x5_KHR}, // ASTC_2D_8X5_SRGB - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x4_KHR}, // ASTC_2D_5X4_SRGB - {GL_COMPRESSED_RGBA_ASTC_5x5_KHR}, // ASTC_2D_5X5_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x5_KHR}, // ASTC_2D_5X5_SRGB - {GL_COMPRESSED_RGBA_ASTC_10x8_KHR}, // ASTC_2D_10X8_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x8_KHR}, // ASTC_2D_10X8_SRGB - {GL_COMPRESSED_RGBA_ASTC_6x6_KHR}, // ASTC_2D_6X6_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x6_KHR}, // ASTC_2D_6X6_SRGB - {GL_COMPRESSED_RGBA_ASTC_10x10_KHR}, // ASTC_2D_10X10_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR}, // ASTC_2D_10X10_SRGB - {GL_COMPRESSED_RGBA_ASTC_12x12_KHR}, // ASTC_2D_12X12_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR}, // ASTC_2D_12X12_SRGB - {GL_COMPRESSED_RGBA_ASTC_8x6_KHR}, // ASTC_2D_8X6_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x6_KHR}, // ASTC_2D_8X6_SRGB - {GL_COMPRESSED_RGBA_ASTC_6x5_KHR}, // ASTC_2D_6X5_UNORM - {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x5_KHR}, // ASTC_2D_6X5_SRGB - {GL_RGB9_E5, GL_RGB, GL_UNSIGNED_INT_5_9_9_9_REV}, // E5B9G9R9_FLOAT - - // Depth formats - {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT}, // D32_FLOAT - {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT}, // D16_UNORM - - // DepthStencil formats - {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // D24_UNORM_S8_UINT - {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // S8_UINT_D24_UNORM + {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT}, // BC1_RGBA_SRGB + {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT}, // BC2_SRGB + {GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT}, // BC3_SRGB + {GL_COMPRESSED_SRGB_ALPHA_BPTC_UNORM}, // BC7_SRGB + {GL_RGBA4, GL_RGBA, GL_UNSIGNED_SHORT_4_4_4_4_REV}, // A4B4G4R4_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_4x4_KHR}, // ASTC_2D_4X4_SRGB + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x8_KHR}, // ASTC_2D_8X8_SRGB + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x5_KHR}, // ASTC_2D_8X5_SRGB + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x4_KHR}, // ASTC_2D_5X4_SRGB + {GL_COMPRESSED_RGBA_ASTC_5x5_KHR}, // ASTC_2D_5X5_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_5x5_KHR}, // ASTC_2D_5X5_SRGB + {GL_COMPRESSED_RGBA_ASTC_10x8_KHR}, // ASTC_2D_10X8_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x8_KHR}, // ASTC_2D_10X8_SRGB + {GL_COMPRESSED_RGBA_ASTC_6x6_KHR}, // ASTC_2D_6X6_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x6_KHR}, // ASTC_2D_6X6_SRGB + {GL_COMPRESSED_RGBA_ASTC_10x10_KHR}, // ASTC_2D_10X10_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_10x10_KHR}, // ASTC_2D_10X10_SRGB + {GL_COMPRESSED_RGBA_ASTC_12x12_KHR}, // ASTC_2D_12X12_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_12x12_KHR}, // ASTC_2D_12X12_SRGB + {GL_COMPRESSED_RGBA_ASTC_8x6_KHR}, // ASTC_2D_8X6_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_8x6_KHR}, // ASTC_2D_8X6_SRGB + {GL_COMPRESSED_RGBA_ASTC_6x5_KHR}, // ASTC_2D_6X5_UNORM + {GL_COMPRESSED_SRGB8_ALPHA8_ASTC_6x5_KHR}, // ASTC_2D_6X5_SRGB + {GL_RGB9_E5, GL_RGB, GL_UNSIGNED_INT_5_9_9_9_REV}, // E5B9G9R9_FLOAT + {GL_DEPTH_COMPONENT32F, GL_DEPTH_COMPONENT, GL_FLOAT}, // D32_FLOAT + {GL_DEPTH_COMPONENT16, GL_DEPTH_COMPONENT, GL_UNSIGNED_SHORT}, // D16_UNORM + {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // D24_UNORM_S8_UINT + {GL_DEPTH24_STENCIL8, GL_DEPTH_STENCIL, GL_UNSIGNED_INT_24_8}, // S8_UINT_D24_UNORM {GL_DEPTH32F_STENCIL8, GL_DEPTH_STENCIL, GL_FLOAT_32_UNSIGNED_INT_24_8_REV}, // D32_FLOAT_S8_UINT }}; +constexpr std::array ACCELERATED_FORMATS{ + GL_RGBA32F, GL_RGBA16F, GL_RG32F, GL_RG16F, GL_R11F_G11F_B10F, GL_R32F, + GL_R16F, GL_RGBA32UI, GL_RGBA16UI, GL_RGB10_A2UI, GL_RGBA8UI, GL_RG32UI, + GL_RG16UI, GL_RG8UI, GL_R32UI, GL_R16UI, GL_R8UI, GL_RGBA32I, + GL_RGBA16I, GL_RGBA8I, GL_RG32I, GL_RG16I, GL_RG8I, GL_R32I, + GL_R16I, GL_R8I, GL_RGBA16, GL_RGB10_A2, GL_RGBA8, GL_RG16, + GL_RG8, GL_R16, GL_R8, GL_RGBA16_SNORM, GL_RGBA8_SNORM, GL_RG16_SNORM, + GL_RG8_SNORM, GL_R16_SNORM, GL_R8_SNORM, +}; + const FormatTuple& GetFormatTuple(PixelFormat pixel_format) { - ASSERT(static_cast<std::size_t>(pixel_format) < tex_format_tuples.size()); - return tex_format_tuples[static_cast<std::size_t>(pixel_format)]; + ASSERT(static_cast<size_t>(pixel_format) < FORMAT_TABLE.size()); + return FORMAT_TABLE[static_cast<size_t>(pixel_format)]; } -GLenum GetTextureTarget(const SurfaceTarget& target) { - switch (target) { - case SurfaceTarget::TextureBuffer: +GLenum ImageTarget(const VideoCommon::ImageInfo& info) { + switch (info.type) { + case ImageType::e1D: + return GL_TEXTURE_1D_ARRAY; + case ImageType::e2D: + if (info.num_samples > 1) { + return GL_TEXTURE_2D_MULTISAMPLE_ARRAY; + } + return GL_TEXTURE_2D_ARRAY; + case ImageType::e3D: + return GL_TEXTURE_3D; + case ImageType::Linear: + return GL_TEXTURE_2D_ARRAY; + case ImageType::Buffer: return GL_TEXTURE_BUFFER; - case SurfaceTarget::Texture1D: + } + UNREACHABLE_MSG("Invalid image type={}", info.type); + return GL_NONE; +} + +GLenum ImageTarget(ImageViewType type, int num_samples = 1) { + const bool is_multisampled = num_samples > 1; + switch (type) { + case ImageViewType::e1D: return GL_TEXTURE_1D; - case SurfaceTarget::Texture2D: - return GL_TEXTURE_2D; - case SurfaceTarget::Texture3D: + case ImageViewType::e2D: + return is_multisampled ? GL_TEXTURE_2D_MULTISAMPLE : GL_TEXTURE_2D; + case ImageViewType::Cube: + return GL_TEXTURE_CUBE_MAP; + case ImageViewType::e3D: return GL_TEXTURE_3D; - case SurfaceTarget::Texture1DArray: + case ImageViewType::e1DArray: return GL_TEXTURE_1D_ARRAY; - case SurfaceTarget::Texture2DArray: - return GL_TEXTURE_2D_ARRAY; - case SurfaceTarget::TextureCubemap: - return GL_TEXTURE_CUBE_MAP; - case SurfaceTarget::TextureCubeArray: + case ImageViewType::e2DArray: + return is_multisampled ? GL_TEXTURE_2D_MULTISAMPLE_ARRAY : GL_TEXTURE_2D_ARRAY; + case ImageViewType::CubeArray: return GL_TEXTURE_CUBE_MAP_ARRAY; + case ImageViewType::Rect: + return GL_TEXTURE_RECTANGLE; + case ImageViewType::Buffer: + return GL_TEXTURE_BUFFER; } - UNREACHABLE(); - return {}; + UNREACHABLE_MSG("Invalid image view type={}", type); + return GL_NONE; } -GLint GetSwizzleSource(SwizzleSource source) { +GLenum TextureMode(PixelFormat format, bool is_first) { + switch (format) { + case PixelFormat::D24_UNORM_S8_UINT: + case PixelFormat::D32_FLOAT_S8_UINT: + return is_first ? GL_DEPTH_COMPONENT : GL_STENCIL_INDEX; + case PixelFormat::S8_UINT_D24_UNORM: + return is_first ? GL_STENCIL_INDEX : GL_DEPTH_COMPONENT; + default: + UNREACHABLE(); + return GL_DEPTH_COMPONENT; + } +} + +GLint Swizzle(SwizzleSource source) { switch (source) { case SwizzleSource::Zero: return GL_ZERO; @@ -184,531 +248,813 @@ GLint GetSwizzleSource(SwizzleSource source) { case SwizzleSource::OneFloat: return GL_ONE; } - UNREACHABLE(); + UNREACHABLE_MSG("Invalid swizzle source={}", source); return GL_NONE; } -GLenum GetComponent(PixelFormat format, bool is_first) { - switch (format) { - case PixelFormat::D24_UNORM_S8_UINT: - case PixelFormat::D32_FLOAT_S8_UINT: - return is_first ? GL_DEPTH_COMPONENT : GL_STENCIL_INDEX; - case PixelFormat::S8_UINT_D24_UNORM: - return is_first ? GL_STENCIL_INDEX : GL_DEPTH_COMPONENT; +GLenum AttachmentType(PixelFormat format) { + switch (const SurfaceType type = VideoCore::Surface::GetFormatType(format); type) { + case SurfaceType::Depth: + return GL_DEPTH_ATTACHMENT; + case SurfaceType::DepthStencil: + return GL_DEPTH_STENCIL_ATTACHMENT; default: - UNREACHABLE(); - return GL_DEPTH_COMPONENT; + UNIMPLEMENTED_MSG("Unimplemented type={}", type); + return GL_NONE; } } -void ApplyTextureDefaults(const SurfaceParams& params, GLuint texture) { - if (params.IsBuffer()) { - return; +[[nodiscard]] bool IsConverted(const Device& device, PixelFormat format, ImageType type) { + if (!device.HasASTC() && IsPixelFormatASTC(format)) { + return true; } - glTextureParameteri(texture, GL_TEXTURE_MIN_FILTER, GL_LINEAR); - glTextureParameteri(texture, GL_TEXTURE_MAG_FILTER, GL_LINEAR); - glTextureParameteri(texture, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE); - glTextureParameteri(texture, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE); - glTextureParameteri(texture, GL_TEXTURE_MAX_LEVEL, static_cast<GLint>(params.num_levels - 1)); - if (params.num_levels == 1) { - glTextureParameterf(texture, GL_TEXTURE_LOD_BIAS, 1000.0f); + switch (format) { + case PixelFormat::BC4_UNORM: + case PixelFormat::BC5_UNORM: + return type == ImageType::e3D; + default: + break; } + return false; } -OGLTexture CreateTexture(const SurfaceParams& params, GLenum target, GLenum internal_format, - OGLBuffer& texture_buffer) { - OGLTexture texture; - texture.Create(target); +[[nodiscard]] constexpr SwizzleSource ConvertGreenRed(SwizzleSource value) { + switch (value) { + case SwizzleSource::G: + return SwizzleSource::R; + default: + return value; + } +} - switch (params.target) { - case SurfaceTarget::Texture1D: - glTextureStorage1D(texture.handle, params.emulated_levels, internal_format, params.width); - break; - case SurfaceTarget::TextureBuffer: - texture_buffer.Create(); - glNamedBufferStorage(texture_buffer.handle, params.width * params.GetBytesPerPixel(), - nullptr, GL_DYNAMIC_STORAGE_BIT); - glTextureBuffer(texture.handle, internal_format, texture_buffer.handle); +void ApplySwizzle(GLuint handle, PixelFormat format, std::array<SwizzleSource, 4> swizzle) { + switch (format) { + case PixelFormat::D24_UNORM_S8_UINT: + case PixelFormat::D32_FLOAT_S8_UINT: + case PixelFormat::S8_UINT_D24_UNORM: + UNIMPLEMENTED_IF(swizzle[0] != SwizzleSource::R && swizzle[0] != SwizzleSource::G); + glTextureParameteri(handle, GL_DEPTH_STENCIL_TEXTURE_MODE, + TextureMode(format, swizzle[0] == SwizzleSource::R)); + std::ranges::transform(swizzle, swizzle.begin(), ConvertGreenRed); break; - case SurfaceTarget::Texture2D: - case SurfaceTarget::TextureCubemap: - glTextureStorage2D(texture.handle, params.emulated_levels, internal_format, params.width, - params.height); + default: break; - case SurfaceTarget::Texture3D: - case SurfaceTarget::Texture2DArray: - case SurfaceTarget::TextureCubeArray: - glTextureStorage3D(texture.handle, params.emulated_levels, internal_format, params.width, - params.height, params.depth); + } + std::array<GLint, 4> gl_swizzle; + std::ranges::transform(swizzle, gl_swizzle.begin(), Swizzle); + glTextureParameteriv(handle, GL_TEXTURE_SWIZZLE_RGBA, gl_swizzle.data()); +} + +[[nodiscard]] bool CanBeAccelerated(const TextureCacheRuntime& runtime, + const VideoCommon::ImageInfo& info) { + // Disable accelerated uploads for now as they don't implement swizzled uploads + return false; + switch (info.type) { + case ImageType::e2D: + case ImageType::e3D: + case ImageType::Linear: break; default: - UNREACHABLE(); + return false; + } + const GLenum internal_format = GetFormatTuple(info.format).internal_format; + const auto& format_info = runtime.FormatInfo(info.type, internal_format); + if (format_info.is_compressed) { + return false; + } + if (std::ranges::find(ACCELERATED_FORMATS, internal_format) == ACCELERATED_FORMATS.end()) { + return false; } + if (format_info.compatibility_by_size) { + return true; + } + const GLenum store_format = StoreFormat(BytesPerBlock(info.format)); + const GLenum store_class = runtime.FormatInfo(info.type, store_format).compatibility_class; + return format_info.compatibility_class == store_class; +} - ApplyTextureDefaults(params, texture.handle); +[[nodiscard]] CopyOrigin MakeCopyOrigin(VideoCommon::Offset3D offset, + VideoCommon::SubresourceLayers subresource, GLenum target) { + switch (target) { + case GL_TEXTURE_2D_ARRAY: + case GL_TEXTURE_2D_MULTISAMPLE_ARRAY: + return CopyOrigin{ + .level = static_cast<GLint>(subresource.base_level), + .x = static_cast<GLint>(offset.x), + .y = static_cast<GLint>(offset.y), + .z = static_cast<GLint>(subresource.base_layer), + }; + case GL_TEXTURE_3D: + return CopyOrigin{ + .level = static_cast<GLint>(subresource.base_level), + .x = static_cast<GLint>(offset.x), + .y = static_cast<GLint>(offset.y), + .z = static_cast<GLint>(offset.z), + }; + default: + UNIMPLEMENTED_MSG("Unimplemented copy target={}", target); + return CopyOrigin{.level = 0, .x = 0, .y = 0, .z = 0}; + } +} - return texture; +[[nodiscard]] CopyRegion MakeCopyRegion(VideoCommon::Extent3D extent, + VideoCommon::SubresourceLayers dst_subresource, + GLenum target) { + switch (target) { + case GL_TEXTURE_2D_ARRAY: + case GL_TEXTURE_2D_MULTISAMPLE_ARRAY: + return CopyRegion{ + .width = static_cast<GLsizei>(extent.width), + .height = static_cast<GLsizei>(extent.height), + .depth = static_cast<GLsizei>(dst_subresource.num_layers), + }; + case GL_TEXTURE_3D: + return CopyRegion{ + .width = static_cast<GLsizei>(extent.width), + .height = static_cast<GLsizei>(extent.height), + .depth = static_cast<GLsizei>(extent.depth), + }; + default: + UNIMPLEMENTED_MSG("Unimplemented copy target={}", target); + return CopyRegion{.width = 0, .height = 0, .depth = 0}; + } } -constexpr u32 EncodeSwizzle(SwizzleSource x_source, SwizzleSource y_source, SwizzleSource z_source, - SwizzleSource w_source) { - return (static_cast<u32>(x_source) << 24) | (static_cast<u32>(y_source) << 16) | - (static_cast<u32>(z_source) << 8) | static_cast<u32>(w_source); +void AttachTexture(GLuint fbo, GLenum attachment, const ImageView* image_view) { + if (False(image_view->flags & VideoCommon::ImageViewFlagBits::Slice)) { + const GLuint texture = image_view->DefaultHandle(); + glNamedFramebufferTexture(fbo, attachment, texture, 0); + return; + } + const GLuint texture = image_view->Handle(ImageViewType::e3D); + if (image_view->range.extent.layers > 1) { + // TODO: OpenGL doesn't support rendering to a fixed number of slices + glNamedFramebufferTexture(fbo, attachment, texture, 0); + } else { + const u32 slice = image_view->range.base.layer; + glNamedFramebufferTextureLayer(fbo, attachment, texture, 0, slice); + } } } // Anonymous namespace -CachedSurface::CachedSurface(const GPUVAddr gpu_addr, const SurfaceParams& params, - bool is_astc_supported) - : VideoCommon::SurfaceBase<View>(gpu_addr, params, is_astc_supported) { - if (is_converted) { - internal_format = params.srgb_conversion ? GL_SRGB8_ALPHA8 : GL_RGBA8; - format = GL_RGBA; - type = GL_UNSIGNED_BYTE; - } else { - const auto& tuple{GetFormatTuple(params.pixel_format)}; - internal_format = tuple.internal_format; - format = tuple.format; - type = tuple.type; - is_compressed = params.IsCompressed(); - } - target = GetTextureTarget(params.target); - texture = CreateTexture(params, target, internal_format, texture_buffer); - DecorateSurfaceName(); +ImageBufferMap::ImageBufferMap(GLuint handle_, u8* map, size_t size, OGLSync* sync_) + : span(map, size), sync{sync_}, handle{handle_} {} - u32 num_layers = 1; - if (params.is_layered || params.target == SurfaceTarget::Texture3D) { - num_layers = params.depth; +ImageBufferMap::~ImageBufferMap() { + if (sync) { + sync->Create(); } - - main_view = - CreateViewInner(ViewParams(params.target, 0, num_layers, 0, params.num_levels), true); } -CachedSurface::~CachedSurface() = default; +TextureCacheRuntime::TextureCacheRuntime(const Device& device_, ProgramManager& program_manager, + StateTracker& state_tracker_) + : device{device_}, state_tracker{state_tracker_}, util_shaders(program_manager) { + static constexpr std::array TARGETS{GL_TEXTURE_1D_ARRAY, GL_TEXTURE_2D_ARRAY, GL_TEXTURE_3D}; + for (size_t i = 0; i < TARGETS.size(); ++i) { + const GLenum target = TARGETS[i]; + for (const FormatTuple& tuple : FORMAT_TABLE) { + const GLenum format = tuple.internal_format; + GLint compat_class; + GLint compat_type; + GLint is_compressed; + glGetInternalformativ(target, format, GL_IMAGE_COMPATIBILITY_CLASS, 1, &compat_class); + glGetInternalformativ(target, format, GL_IMAGE_FORMAT_COMPATIBILITY_TYPE, 1, + &compat_type); + glGetInternalformativ(target, format, GL_TEXTURE_COMPRESSED, 1, &is_compressed); + const FormatProperties properties{ + .compatibility_class = static_cast<GLenum>(compat_class), + .compatibility_by_size = compat_type == GL_IMAGE_FORMAT_COMPATIBILITY_BY_SIZE, + .is_compressed = is_compressed == GL_TRUE, + }; + format_properties[i].emplace(format, properties); + } + } + has_broken_texture_view_formats = device.HasBrokenTextureViewFormats(); + + null_image_1d_array.Create(GL_TEXTURE_1D_ARRAY); + null_image_cube_array.Create(GL_TEXTURE_CUBE_MAP_ARRAY); + null_image_3d.Create(GL_TEXTURE_3D); + null_image_rect.Create(GL_TEXTURE_RECTANGLE); + glTextureStorage2D(null_image_1d_array.handle, 1, GL_R8, 1, 1); + glTextureStorage3D(null_image_cube_array.handle, 1, GL_R8, 1, 1, 6); + glTextureStorage3D(null_image_3d.handle, 1, GL_R8, 1, 1, 1); + glTextureStorage2D(null_image_rect.handle, 1, GL_R8, 1, 1); + + std::array<GLuint, 4> new_handles; + glGenTextures(static_cast<GLsizei>(new_handles.size()), new_handles.data()); + null_image_view_1d.handle = new_handles[0]; + null_image_view_2d.handle = new_handles[1]; + null_image_view_2d_array.handle = new_handles[2]; + null_image_view_cube.handle = new_handles[3]; + glTextureView(null_image_view_1d.handle, GL_TEXTURE_1D, null_image_1d_array.handle, GL_R8, 0, 1, + 0, 1); + glTextureView(null_image_view_2d.handle, GL_TEXTURE_2D, null_image_cube_array.handle, GL_R8, 0, + 1, 0, 1); + glTextureView(null_image_view_2d_array.handle, GL_TEXTURE_2D_ARRAY, + null_image_cube_array.handle, GL_R8, 0, 1, 0, 1); + glTextureView(null_image_view_cube.handle, GL_TEXTURE_CUBE_MAP, null_image_cube_array.handle, + GL_R8, 0, 1, 0, 6); + const std::array texture_handles{ + null_image_1d_array.handle, null_image_cube_array.handle, null_image_3d.handle, + null_image_rect.handle, null_image_view_1d.handle, null_image_view_2d.handle, + null_image_view_2d_array.handle, null_image_view_cube.handle, + }; + for (const GLuint handle : texture_handles) { + static constexpr std::array NULL_SWIZZLE{GL_ZERO, GL_ZERO, GL_ZERO, GL_ZERO}; + glTextureParameteriv(handle, GL_TEXTURE_SWIZZLE_RGBA, NULL_SWIZZLE.data()); + } + const auto set_view = [this](ImageViewType type, GLuint handle) { + if (device.HasDebuggingToolAttached()) { + const std::string name = fmt::format("NullImage {}", type); + glObjectLabel(GL_TEXTURE, handle, static_cast<GLsizei>(name.size()), name.data()); + } + null_image_views[static_cast<size_t>(type)] = handle; + }; + set_view(ImageViewType::e1D, null_image_view_1d.handle); + set_view(ImageViewType::e2D, null_image_view_2d.handle); + set_view(ImageViewType::Cube, null_image_view_cube.handle); + set_view(ImageViewType::e3D, null_image_3d.handle); + set_view(ImageViewType::e1DArray, null_image_1d_array.handle); + set_view(ImageViewType::e2DArray, null_image_view_2d_array.handle); + set_view(ImageViewType::CubeArray, null_image_cube_array.handle); + set_view(ImageViewType::Rect, null_image_rect.handle); +} -void CachedSurface::DownloadTexture(std::vector<u8>& staging_buffer) { - MICROPROFILE_SCOPE(OpenGL_Texture_Download); +TextureCacheRuntime::~TextureCacheRuntime() = default; - if (params.IsBuffer()) { - glGetNamedBufferSubData(texture_buffer.handle, 0, - static_cast<GLsizeiptr>(params.GetHostSizeInBytes(false)), - staging_buffer.data()); - return; - } +void TextureCacheRuntime::Finish() { + glFinish(); +} - SCOPE_EXIT({ glPixelStorei(GL_PACK_ROW_LENGTH, 0); }); +ImageBufferMap TextureCacheRuntime::MapUploadBuffer(size_t size) { + return upload_buffers.RequestMap(size, true); +} - for (u32 level = 0; level < params.emulated_levels; ++level) { - glPixelStorei(GL_PACK_ALIGNMENT, std::min(8U, params.GetRowAlignment(level, is_converted))); - glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.GetMipWidth(level))); - const std::size_t mip_offset = params.GetHostMipmapLevelOffset(level, is_converted); +ImageBufferMap TextureCacheRuntime::MapDownloadBuffer(size_t size) { + return download_buffers.RequestMap(size, false); +} - u8* const mip_data = staging_buffer.data() + mip_offset; - const GLsizei size = static_cast<GLsizei>(params.GetHostMipmapSize(level)); - if (is_compressed) { - glGetCompressedTextureImage(texture.handle, level, size, mip_data); - } else { - glGetTextureImage(texture.handle, level, format, type, size, mip_data); - } +void TextureCacheRuntime::CopyImage(Image& dst_image, Image& src_image, + std::span<const ImageCopy> copies) { + const GLuint dst_name = dst_image.Handle(); + const GLuint src_name = src_image.Handle(); + const GLenum dst_target = ImageTarget(dst_image.info); + const GLenum src_target = ImageTarget(src_image.info); + for (const ImageCopy& copy : copies) { + const auto src_origin = MakeCopyOrigin(copy.src_offset, copy.src_subresource, src_target); + const auto dst_origin = MakeCopyOrigin(copy.dst_offset, copy.dst_subresource, dst_target); + const auto region = MakeCopyRegion(copy.extent, copy.dst_subresource, dst_target); + glCopyImageSubData(src_name, src_target, src_origin.level, src_origin.x, src_origin.y, + src_origin.z, dst_name, dst_target, dst_origin.level, dst_origin.x, + dst_origin.y, dst_origin.z, region.width, region.height, region.depth); } } -void CachedSurface::UploadTexture(const std::vector<u8>& staging_buffer) { - MICROPROFILE_SCOPE(OpenGL_Texture_Upload); - SCOPE_EXIT({ glPixelStorei(GL_UNPACK_ROW_LENGTH, 0); }); - for (u32 level = 0; level < params.emulated_levels; ++level) { - UploadTextureMipmap(level, staging_buffer); +bool TextureCacheRuntime::CanImageBeCopied(const Image& dst, const Image& src) { + if (dst.info.type == ImageType::e3D && dst.info.format == PixelFormat::BC4_UNORM) { + return false; } + return true; } -void CachedSurface::UploadTextureMipmap(u32 level, const std::vector<u8>& staging_buffer) { - glPixelStorei(GL_UNPACK_ALIGNMENT, std::min(8U, params.GetRowAlignment(level, is_converted))); - glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.GetMipWidth(level))); - - const std::size_t mip_offset = params.GetHostMipmapLevelOffset(level, is_converted); - const u8* buffer{staging_buffer.data() + mip_offset}; - if (is_compressed) { - const auto image_size{static_cast<GLsizei>(params.GetHostMipmapSize(level))}; - switch (params.target) { - case SurfaceTarget::Texture2D: - glCompressedTextureSubImage2D(texture.handle, level, 0, 0, - static_cast<GLsizei>(params.GetMipWidth(level)), - static_cast<GLsizei>(params.GetMipHeight(level)), - internal_format, image_size, buffer); - break; - case SurfaceTarget::Texture3D: - case SurfaceTarget::Texture2DArray: - case SurfaceTarget::TextureCubeArray: - glCompressedTextureSubImage3D(texture.handle, level, 0, 0, 0, - static_cast<GLsizei>(params.GetMipWidth(level)), - static_cast<GLsizei>(params.GetMipHeight(level)), - static_cast<GLsizei>(params.GetMipDepth(level)), - internal_format, image_size, buffer); - break; - case SurfaceTarget::TextureCubemap: { - const std::size_t layer_size{params.GetHostLayerSize(level)}; - for (std::size_t face = 0; face < params.depth; ++face) { - glCompressedTextureSubImage3D(texture.handle, level, 0, 0, static_cast<GLint>(face), - static_cast<GLsizei>(params.GetMipWidth(level)), - static_cast<GLsizei>(params.GetMipHeight(level)), 1, - internal_format, static_cast<GLsizei>(layer_size), - buffer); - buffer += layer_size; - } - break; - } - default: - UNREACHABLE(); - } +void TextureCacheRuntime::EmulateCopyImage(Image& dst, Image& src, + std::span<const ImageCopy> copies) { + if (dst.info.type == ImageType::e3D && dst.info.format == PixelFormat::BC4_UNORM) { + ASSERT(src.info.type == ImageType::e3D); + util_shaders.CopyBC4(dst, src, copies); } else { - switch (params.target) { - case SurfaceTarget::Texture1D: - glTextureSubImage1D(texture.handle, level, 0, params.GetMipWidth(level), format, type, - buffer); - break; - case SurfaceTarget::TextureBuffer: - ASSERT(level == 0); - glNamedBufferSubData(texture_buffer.handle, 0, - params.GetMipWidth(level) * params.GetBytesPerPixel(), buffer); - break; - case SurfaceTarget::Texture1DArray: - case SurfaceTarget::Texture2D: - glTextureSubImage2D(texture.handle, level, 0, 0, params.GetMipWidth(level), - params.GetMipHeight(level), format, type, buffer); - break; - case SurfaceTarget::Texture3D: - case SurfaceTarget::Texture2DArray: - case SurfaceTarget::TextureCubeArray: - glTextureSubImage3D( - texture.handle, level, 0, 0, 0, static_cast<GLsizei>(params.GetMipWidth(level)), - static_cast<GLsizei>(params.GetMipHeight(level)), - static_cast<GLsizei>(params.GetMipDepth(level)), format, type, buffer); - break; - case SurfaceTarget::TextureCubemap: - for (std::size_t face = 0; face < params.depth; ++face) { - glTextureSubImage3D(texture.handle, level, 0, 0, static_cast<GLint>(face), - params.GetMipWidth(level), params.GetMipHeight(level), 1, - format, type, buffer); - buffer += params.GetHostLayerSize(level); - } - break; - default: - UNREACHABLE(); - } + UNREACHABLE(); } } -void CachedSurface::DecorateSurfaceName() { - LabelGLObject(GL_TEXTURE, texture.handle, GetGpuAddr(), params.TargetName()); -} +void TextureCacheRuntime::BlitFramebuffer(Framebuffer* dst, Framebuffer* src, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation) { + state_tracker.NotifyScissor0(); + state_tracker.NotifyRasterizeEnable(); + state_tracker.NotifyFramebufferSRGB(); -void CachedSurfaceView::DecorateViewName(GPUVAddr gpu_addr, const std::string& prefix) { - LabelGLObject(GL_TEXTURE, main_view.handle, gpu_addr, prefix); + ASSERT(dst->BufferBits() == src->BufferBits()); + + glEnable(GL_FRAMEBUFFER_SRGB); + glDisable(GL_RASTERIZER_DISCARD); + glDisablei(GL_SCISSOR_TEST, 0); + + const GLbitfield buffer_bits = dst->BufferBits(); + const bool has_depth = (buffer_bits & ~GL_COLOR_BUFFER_BIT) != 0; + const bool is_linear = !has_depth && filter == Tegra::Engines::Fermi2D::Filter::Bilinear; + glBlitNamedFramebuffer(src->Handle(), dst->Handle(), src_region[0].x, src_region[0].y, + src_region[1].x, src_region[1].y, dst_region[0].x, dst_region[0].y, + dst_region[1].x, dst_region[1].y, buffer_bits, + is_linear ? GL_LINEAR : GL_NEAREST); } -View CachedSurface::CreateView(const ViewParams& view_key) { - return CreateViewInner(view_key, false); +void TextureCacheRuntime::AccelerateImageUpload(Image& image, const ImageBufferMap& map, + size_t buffer_offset, + std::span<const SwizzleParameters> swizzles) { + switch (image.info.type) { + case ImageType::e2D: + return util_shaders.BlockLinearUpload2D(image, map, buffer_offset, swizzles); + case ImageType::e3D: + return util_shaders.BlockLinearUpload3D(image, map, buffer_offset, swizzles); + case ImageType::Linear: + return util_shaders.PitchUpload(image, map, buffer_offset, swizzles); + default: + UNREACHABLE(); + break; + } } -View CachedSurface::CreateViewInner(const ViewParams& view_key, const bool is_proxy) { - auto view = std::make_shared<CachedSurfaceView>(*this, view_key, is_proxy); - views[view_key] = view; - if (!is_proxy) - view->DecorateViewName(gpu_addr, params.TargetName() + "V:" + std::to_string(view_count++)); - return view; +void TextureCacheRuntime::InsertUploadMemoryBarrier() { + glMemoryBarrier(GL_TEXTURE_FETCH_BARRIER_BIT | GL_SHADER_IMAGE_ACCESS_BARRIER_BIT); } -CachedSurfaceView::CachedSurfaceView(CachedSurface& surface, const ViewParams& params, - bool is_proxy) - : VideoCommon::ViewBase(params), surface{surface}, format{surface.internal_format}, - target{GetTextureTarget(params.target)}, is_proxy{is_proxy} { - if (!is_proxy) { - main_view = CreateTextureView(); +FormatProperties TextureCacheRuntime::FormatInfo(ImageType type, GLenum internal_format) const { + switch (type) { + case ImageType::e1D: + return format_properties[0].at(internal_format); + case ImageType::e2D: + case ImageType::Linear: + return format_properties[1].at(internal_format); + case ImageType::e3D: + return format_properties[2].at(internal_format); + default: + UNREACHABLE(); + return FormatProperties{}; } } -CachedSurfaceView::~CachedSurfaceView() = default; +TextureCacheRuntime::StagingBuffers::StagingBuffers(GLenum storage_flags_, GLenum map_flags_) + : storage_flags{storage_flags_}, map_flags{map_flags_} {} -void CachedSurfaceView::Attach(GLenum attachment, GLenum fb_target) const { - ASSERT(params.num_levels == 1); +TextureCacheRuntime::StagingBuffers::~StagingBuffers() = default; - if (params.target == SurfaceTarget::Texture3D) { - if (params.num_layers > 1) { - ASSERT(params.base_layer == 0); - glFramebufferTexture(fb_target, attachment, surface.texture.handle, params.base_level); - } else { - glFramebufferTexture3D(fb_target, attachment, target, surface.texture.handle, - params.base_level, params.base_layer); - } - return; +ImageBufferMap TextureCacheRuntime::StagingBuffers::RequestMap(size_t requested_size, + bool insert_fence) { + const size_t index = RequestBuffer(requested_size); + OGLSync* const sync = insert_fence ? &syncs[index] : nullptr; + return ImageBufferMap(buffers[index].handle, maps[index], requested_size, sync); +} + +size_t TextureCacheRuntime::StagingBuffers::RequestBuffer(size_t requested_size) { + if (const std::optional<size_t> index = FindBuffer(requested_size); index) { + return *index; } - if (params.num_layers > 1) { - UNIMPLEMENTED_IF(params.base_layer != 0); - glFramebufferTexture(fb_target, attachment, GetTexture(), 0); - return; + OGLBuffer& buffer = buffers.emplace_back(); + buffer.Create(); + glNamedBufferStorage(buffer.handle, requested_size, nullptr, + storage_flags | GL_MAP_PERSISTENT_BIT); + maps.push_back(static_cast<u8*>(glMapNamedBufferRange(buffer.handle, 0, requested_size, + map_flags | GL_MAP_PERSISTENT_BIT))); + + syncs.emplace_back(); + sizes.push_back(requested_size); + + ASSERT(syncs.size() == buffers.size() && buffers.size() == maps.size() && + maps.size() == sizes.size()); + + return buffers.size() - 1; +} + +std::optional<size_t> TextureCacheRuntime::StagingBuffers::FindBuffer(size_t requested_size) { + size_t smallest_buffer = std::numeric_limits<size_t>::max(); + std::optional<size_t> found; + const size_t num_buffers = sizes.size(); + for (size_t index = 0; index < num_buffers; ++index) { + const size_t buffer_size = sizes[index]; + if (buffer_size < requested_size || buffer_size >= smallest_buffer) { + continue; + } + if (syncs[index].handle != 0) { + GLint status; + glGetSynciv(syncs[index].handle, GL_SYNC_STATUS, 1, nullptr, &status); + if (status != GL_SIGNALED) { + continue; + } + syncs[index].Release(); + } + smallest_buffer = buffer_size; + found = index; } + return found; +} - const GLenum view_target = surface.GetTarget(); - const GLuint texture = surface.GetTexture(); - switch (surface.GetSurfaceParams().target) { - case SurfaceTarget::Texture1D: - glFramebufferTexture1D(fb_target, attachment, view_target, texture, params.base_level); +Image::Image(TextureCacheRuntime& runtime, const VideoCommon::ImageInfo& info_, GPUVAddr gpu_addr_, + VAddr cpu_addr_) + : VideoCommon::ImageBase(info_, gpu_addr_, cpu_addr_) { + if (CanBeAccelerated(runtime, info)) { + flags |= ImageFlagBits::AcceleratedUpload; + } + if (IsConverted(runtime.device, info.format, info.type)) { + flags |= ImageFlagBits::Converted; + gl_internal_format = IsPixelFormatSRGB(info.format) ? GL_SRGB8_ALPHA8 : GL_RGBA8; + gl_format = GL_RGBA; + gl_type = GL_UNSIGNED_INT_8_8_8_8_REV; + } else { + const auto& tuple = GetFormatTuple(info.format); + gl_internal_format = tuple.internal_format; + gl_format = tuple.format; + gl_type = tuple.type; + } + const GLenum target = ImageTarget(info); + const GLsizei width = info.size.width; + const GLsizei height = info.size.height; + const GLsizei depth = info.size.depth; + const int max_host_mip_levels = std::bit_width(info.size.width); + const GLsizei num_levels = std::min(info.resources.levels, max_host_mip_levels); + const GLsizei num_layers = info.resources.layers; + const GLsizei num_samples = info.num_samples; + + GLuint handle = 0; + if (target != GL_TEXTURE_BUFFER) { + texture.Create(target); + handle = texture.handle; + } + switch (target) { + case GL_TEXTURE_1D_ARRAY: + glTextureStorage2D(handle, num_levels, gl_internal_format, width, num_layers); break; - case SurfaceTarget::Texture2D: - glFramebufferTexture2D(fb_target, attachment, view_target, texture, params.base_level); + case GL_TEXTURE_2D_ARRAY: + glTextureStorage3D(handle, num_levels, gl_internal_format, width, height, num_layers); break; - case SurfaceTarget::Texture1DArray: - case SurfaceTarget::Texture2DArray: - case SurfaceTarget::TextureCubemap: - case SurfaceTarget::TextureCubeArray: - glFramebufferTextureLayer(fb_target, attachment, texture, params.base_level, - params.base_layer); + case GL_TEXTURE_2D_MULTISAMPLE_ARRAY: { + // TODO: Where should 'fixedsamplelocations' come from? + const auto [samples_x, samples_y] = SamplesLog2(info.num_samples); + glTextureStorage3DMultisample(handle, num_samples, gl_internal_format, width >> samples_x, + height >> samples_y, num_layers, GL_FALSE); + break; + } + case GL_TEXTURE_RECTANGLE: + glTextureStorage2D(handle, num_levels, gl_internal_format, width, height); + break; + case GL_TEXTURE_3D: + glTextureStorage3D(handle, num_levels, gl_internal_format, width, height, depth); + break; + case GL_TEXTURE_BUFFER: + buffer.Create(); + glNamedBufferStorage(buffer.handle, guest_size_bytes, nullptr, 0); break; default: - UNIMPLEMENTED(); + UNREACHABLE_MSG("Invalid target=0x{:x}", target); + break; + } + if (runtime.device.HasDebuggingToolAttached()) { + const std::string name = VideoCommon::Name(*this); + glObjectLabel(target == GL_TEXTURE_BUFFER ? GL_BUFFER : GL_TEXTURE, handle, + static_cast<GLsizei>(name.size()), name.data()); } } -GLuint CachedSurfaceView::GetTexture(SwizzleSource x_source, SwizzleSource y_source, - SwizzleSource z_source, SwizzleSource w_source) { - if (GetSurfaceParams().IsBuffer()) { - return GetTexture(); - } - const u32 new_swizzle = EncodeSwizzle(x_source, y_source, z_source, w_source); - if (current_swizzle == new_swizzle) { - return current_view; - } - current_swizzle = new_swizzle; +void Image::UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferImageCopy> copies) { + glBindBuffer(GL_PIXEL_UNPACK_BUFFER, map.Handle()); + glFlushMappedBufferRange(GL_PIXEL_UNPACK_BUFFER, buffer_offset, unswizzled_size_bytes); - const auto [entry, is_cache_miss] = view_cache.try_emplace(new_swizzle); - OGLTextureView& view = entry->second; - if (!is_cache_miss) { - current_view = view.handle; - return view.handle; - } - view = CreateTextureView(); - current_view = view.handle; + glPixelStorei(GL_UNPACK_ALIGNMENT, 1); - std::array swizzle{x_source, y_source, z_source, w_source}; + u32 current_row_length = std::numeric_limits<u32>::max(); + u32 current_image_height = std::numeric_limits<u32>::max(); - switch (const PixelFormat format = GetSurfaceParams().pixel_format) { - case PixelFormat::D24_UNORM_S8_UINT: - case PixelFormat::D32_FLOAT_S8_UINT: - case PixelFormat::S8_UINT_D24_UNORM: - UNIMPLEMENTED_IF(x_source != SwizzleSource::R && x_source != SwizzleSource::G); - glTextureParameteri(view.handle, GL_DEPTH_STENCIL_TEXTURE_MODE, - GetComponent(format, x_source == SwizzleSource::R)); - - // Make sure we sample the first component - std::transform(swizzle.begin(), swizzle.end(), swizzle.begin(), [](SwizzleSource value) { - return value == SwizzleSource::G ? SwizzleSource::R : value; - }); - [[fallthrough]]; - default: { - const std::array gl_swizzle = {GetSwizzleSource(swizzle[0]), GetSwizzleSource(swizzle[1]), - GetSwizzleSource(swizzle[2]), GetSwizzleSource(swizzle[3])}; - glTextureParameteriv(view.handle, GL_TEXTURE_SWIZZLE_RGBA, gl_swizzle.data()); - break; - } + for (const VideoCommon::BufferImageCopy& copy : copies) { + if (current_row_length != copy.buffer_row_length) { + current_row_length = copy.buffer_row_length; + glPixelStorei(GL_UNPACK_ROW_LENGTH, current_row_length); + } + if (current_image_height != copy.buffer_image_height) { + current_image_height = copy.buffer_image_height; + glPixelStorei(GL_UNPACK_IMAGE_HEIGHT, current_image_height); + } + CopyBufferToImage(copy, buffer_offset); } - return view.handle; } -OGLTextureView CachedSurfaceView::CreateTextureView() const { - OGLTextureView texture_view; - texture_view.Create(); - - if (target == GL_TEXTURE_3D) { - glTextureView(texture_view.handle, target, surface.texture.handle, format, - params.base_level, params.num_levels, 0, 1); - } else { - glTextureView(texture_view.handle, target, surface.texture.handle, format, - params.base_level, params.num_levels, params.base_layer, params.num_layers); +void Image::UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferCopy> copies) { + for (const VideoCommon::BufferCopy& copy : copies) { + glCopyNamedBufferSubData(map.Handle(), buffer.handle, copy.src_offset + buffer_offset, + copy.dst_offset, copy.size); } - ApplyTextureDefaults(surface.GetSurfaceParams(), texture_view.handle); - - return texture_view; } -TextureCacheOpenGL::TextureCacheOpenGL(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, - Tegra::MemoryManager& gpu_memory, const Device& device, - StateTracker& state_tracker_) - : TextureCacheBase{rasterizer, maxwell3d, gpu_memory, device.HasASTC()}, state_tracker{ - state_tracker_} { - src_framebuffer.Create(); - dst_framebuffer.Create(); -} +void Image::DownloadMemory(ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferImageCopy> copies) { + glMemoryBarrier(GL_PIXEL_BUFFER_BARRIER_BIT); // TODO: Move this to its own API -TextureCacheOpenGL::~TextureCacheOpenGL() = default; + glBindBuffer(GL_PIXEL_PACK_BUFFER, map.Handle()); + glPixelStorei(GL_PACK_ALIGNMENT, 1); -Surface TextureCacheOpenGL::CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) { - return std::make_shared<CachedSurface>(gpu_addr, params, is_astc_supported); -} + u32 current_row_length = std::numeric_limits<u32>::max(); + u32 current_image_height = std::numeric_limits<u32>::max(); -void TextureCacheOpenGL::ImageCopy(Surface& src_surface, Surface& dst_surface, - const VideoCommon::CopyParams& copy_params) { - const auto& src_params = src_surface->GetSurfaceParams(); - const auto& dst_params = dst_surface->GetSurfaceParams(); - if (src_params.type != dst_params.type) { - // A fallback is needed - return; + for (const VideoCommon::BufferImageCopy& copy : copies) { + if (current_row_length != copy.buffer_row_length) { + current_row_length = copy.buffer_row_length; + glPixelStorei(GL_PACK_ROW_LENGTH, current_row_length); + } + if (current_image_height != copy.buffer_image_height) { + current_image_height = copy.buffer_image_height; + glPixelStorei(GL_PACK_IMAGE_HEIGHT, current_image_height); + } + CopyImageToBuffer(copy, buffer_offset); } - const auto src_handle = src_surface->GetTexture(); - const auto src_target = src_surface->GetTarget(); - const auto dst_handle = dst_surface->GetTexture(); - const auto dst_target = dst_surface->GetTarget(); - glCopyImageSubData(src_handle, src_target, copy_params.source_level, copy_params.source_x, - copy_params.source_y, copy_params.source_z, dst_handle, dst_target, - copy_params.dest_level, copy_params.dest_x, copy_params.dest_y, - copy_params.dest_z, copy_params.width, copy_params.height, - copy_params.depth); } -void TextureCacheOpenGL::ImageBlit(View& src_view, View& dst_view, - const Tegra::Engines::Fermi2D::Config& copy_config) { - const auto& src_params{src_view->GetSurfaceParams()}; - const auto& dst_params{dst_view->GetSurfaceParams()}; - UNIMPLEMENTED_IF(src_params.depth != 1); - UNIMPLEMENTED_IF(dst_params.depth != 1); - - state_tracker.NotifyScissor0(); - state_tracker.NotifyFramebuffer(); - state_tracker.NotifyRasterizeEnable(); - state_tracker.NotifyFramebufferSRGB(); +void Image::CopyBufferToImage(const VideoCommon::BufferImageCopy& copy, size_t buffer_offset) { + // Compressed formats don't have a pixel format or type + const bool is_compressed = gl_format == GL_NONE; + const void* const offset = reinterpret_cast<const void*>(copy.buffer_offset + buffer_offset); - if (dst_params.srgb_conversion) { - glEnable(GL_FRAMEBUFFER_SRGB); - } else { - glDisable(GL_FRAMEBUFFER_SRGB); + switch (info.type) { + case ImageType::e1D: + if (is_compressed) { + glCompressedTextureSubImage2D(texture.handle, copy.image_subresource.base_level, + copy.image_offset.x, copy.image_subresource.base_layer, + copy.image_extent.width, + copy.image_subresource.num_layers, gl_internal_format, + static_cast<GLsizei>(copy.buffer_size), offset); + } else { + glTextureSubImage2D(texture.handle, copy.image_subresource.base_level, + copy.image_offset.x, copy.image_subresource.base_layer, + copy.image_extent.width, copy.image_subresource.num_layers, + gl_format, gl_type, offset); + } + break; + case ImageType::e2D: + case ImageType::Linear: + if (is_compressed) { + glCompressedTextureSubImage3D( + texture.handle, copy.image_subresource.base_level, copy.image_offset.x, + copy.image_offset.y, copy.image_subresource.base_layer, copy.image_extent.width, + copy.image_extent.height, copy.image_subresource.num_layers, gl_internal_format, + static_cast<GLsizei>(copy.buffer_size), offset); + } else { + glTextureSubImage3D(texture.handle, copy.image_subresource.base_level, + copy.image_offset.x, copy.image_offset.y, + copy.image_subresource.base_layer, copy.image_extent.width, + copy.image_extent.height, copy.image_subresource.num_layers, + gl_format, gl_type, offset); + } + break; + case ImageType::e3D: + if (is_compressed) { + glCompressedTextureSubImage3D( + texture.handle, copy.image_subresource.base_level, copy.image_offset.x, + copy.image_offset.y, copy.image_offset.z, copy.image_extent.width, + copy.image_extent.height, copy.image_extent.depth, gl_internal_format, + static_cast<GLsizei>(copy.buffer_size), offset); + } else { + glTextureSubImage3D(texture.handle, copy.image_subresource.base_level, + copy.image_offset.x, copy.image_offset.y, copy.image_offset.z, + copy.image_extent.width, copy.image_extent.height, + copy.image_extent.depth, gl_format, gl_type, offset); + } + break; + default: + UNREACHABLE(); } - glDisable(GL_RASTERIZER_DISCARD); - glDisablei(GL_SCISSOR_TEST, 0); - - glBindFramebuffer(GL_READ_FRAMEBUFFER, src_framebuffer.handle); - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, dst_framebuffer.handle); - - GLenum buffers = 0; - if (src_params.type == SurfaceType::ColorTexture) { - src_view->Attach(GL_COLOR_ATTACHMENT0, GL_READ_FRAMEBUFFER); - glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, - 0); - - dst_view->Attach(GL_COLOR_ATTACHMENT0, GL_DRAW_FRAMEBUFFER); - glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_DEPTH_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, - 0); - - buffers = GL_COLOR_BUFFER_BIT; - } else if (src_params.type == SurfaceType::Depth) { - glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0); - src_view->Attach(GL_DEPTH_ATTACHMENT, GL_READ_FRAMEBUFFER); - glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0); +} - glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0); - dst_view->Attach(GL_DEPTH_ATTACHMENT, GL_DRAW_FRAMEBUFFER); - glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_STENCIL_ATTACHMENT, GL_TEXTURE_2D, 0, 0); +void Image::CopyImageToBuffer(const VideoCommon::BufferImageCopy& copy, size_t buffer_offset) { + const GLint x_offset = copy.image_offset.x; + const GLsizei width = copy.image_extent.width; - buffers = GL_DEPTH_BUFFER_BIT; - } else if (src_params.type == SurfaceType::DepthStencil) { - glFramebufferTexture2D(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0); - src_view->Attach(GL_DEPTH_STENCIL_ATTACHMENT, GL_READ_FRAMEBUFFER); + const GLint level = copy.image_subresource.base_level; + const GLsizei buffer_size = static_cast<GLsizei>(copy.buffer_size); + void* const offset = reinterpret_cast<void*>(copy.buffer_offset + buffer_offset); - glFramebufferTexture2D(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, 0, 0); - dst_view->Attach(GL_DEPTH_STENCIL_ATTACHMENT, GL_DRAW_FRAMEBUFFER); + GLint y_offset = 0; + GLint z_offset = 0; + GLsizei height = 1; + GLsizei depth = 1; - buffers = GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT; + switch (info.type) { + case ImageType::e1D: + y_offset = copy.image_subresource.base_layer; + height = copy.image_subresource.num_layers; + break; + case ImageType::e2D: + case ImageType::Linear: + y_offset = copy.image_offset.y; + z_offset = copy.image_subresource.base_layer; + height = copy.image_extent.height; + depth = copy.image_subresource.num_layers; + break; + case ImageType::e3D: + y_offset = copy.image_offset.y; + z_offset = copy.image_offset.z; + height = copy.image_extent.height; + depth = copy.image_extent.depth; + break; + default: + UNREACHABLE(); + } + // Compressed formats don't have a pixel format or type + const bool is_compressed = gl_format == GL_NONE; + if (is_compressed) { + glGetCompressedTextureSubImage(texture.handle, level, x_offset, y_offset, z_offset, width, + height, depth, buffer_size, offset); + } else { + glGetTextureSubImage(texture.handle, level, x_offset, y_offset, z_offset, width, height, + depth, gl_format, gl_type, buffer_size, offset); } - - const Common::Rectangle<u32>& src_rect = copy_config.src_rect; - const Common::Rectangle<u32>& dst_rect = copy_config.dst_rect; - const bool is_linear = copy_config.filter == Tegra::Engines::Fermi2D::Filter::Linear; - - glBlitFramebuffer(static_cast<GLint>(src_rect.left), static_cast<GLint>(src_rect.top), - static_cast<GLint>(src_rect.right), static_cast<GLint>(src_rect.bottom), - static_cast<GLint>(dst_rect.left), static_cast<GLint>(dst_rect.top), - static_cast<GLint>(dst_rect.right), static_cast<GLint>(dst_rect.bottom), - buffers, - is_linear && (buffers == GL_COLOR_BUFFER_BIT) ? GL_LINEAR : GL_NEAREST); } -void TextureCacheOpenGL::BufferCopy(Surface& src_surface, Surface& dst_surface) { - MICROPROFILE_SCOPE(OpenGL_Texture_Buffer_Copy); - const auto& src_params = src_surface->GetSurfaceParams(); - const auto& dst_params = dst_surface->GetSurfaceParams(); - UNIMPLEMENTED_IF(src_params.num_levels > 1 || dst_params.num_levels > 1); +ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::ImageViewInfo& info, + ImageId image_id_, Image& image) + : VideoCommon::ImageViewBase{info, image.info, image_id_}, views{runtime.null_image_views} { + const Device& device = runtime.device; + if (True(image.flags & ImageFlagBits::Converted)) { + internal_format = IsPixelFormatSRGB(info.format) ? GL_SRGB8_ALPHA8 : GL_RGBA8; + } else { + internal_format = GetFormatTuple(format).internal_format; + } + VideoCommon::SubresourceRange flatten_range = info.range; + std::array<GLuint, 2> handles; + stored_views.reserve(2); - const auto source_format = GetFormatTuple(src_params.pixel_format); - const auto dest_format = GetFormatTuple(dst_params.pixel_format); + switch (info.type) { + case ImageViewType::e1DArray: + flatten_range.extent.layers = 1; + [[fallthrough]]; + case ImageViewType::e1D: + glGenTextures(2, handles.data()); + SetupView(device, image, ImageViewType::e1D, handles[0], info, flatten_range); + SetupView(device, image, ImageViewType::e1DArray, handles[1], info, info.range); + break; + case ImageViewType::e2DArray: + flatten_range.extent.layers = 1; + [[fallthrough]]; + case ImageViewType::e2D: + if (True(flags & VideoCommon::ImageViewFlagBits::Slice)) { + // 2D and 2D array views on a 3D textures are used exclusively for render targets + ASSERT(info.range.extent.levels == 1); + const VideoCommon::SubresourceRange slice_range{ + .base = {.level = info.range.base.level, .layer = 0}, + .extent = {.levels = 1, .layers = 1}, + }; + glGenTextures(1, handles.data()); + SetupView(device, image, ImageViewType::e3D, handles[0], info, slice_range); + break; + } + glGenTextures(2, handles.data()); + SetupView(device, image, ImageViewType::e2D, handles[0], info, flatten_range); + SetupView(device, image, ImageViewType::e2DArray, handles[1], info, info.range); + break; + case ImageViewType::e3D: + glGenTextures(1, handles.data()); + SetupView(device, image, ImageViewType::e3D, handles[0], info, info.range); + break; + case ImageViewType::CubeArray: + flatten_range.extent.layers = 6; + [[fallthrough]]; + case ImageViewType::Cube: + glGenTextures(2, handles.data()); + SetupView(device, image, ImageViewType::Cube, handles[0], info, flatten_range); + SetupView(device, image, ImageViewType::CubeArray, handles[1], info, info.range); + break; + case ImageViewType::Rect: + glGenTextures(1, handles.data()); + SetupView(device, image, ImageViewType::Rect, handles[0], info, info.range); + break; + case ImageViewType::Buffer: + glCreateTextures(GL_TEXTURE_BUFFER, 1, handles.data()); + SetupView(device, image, ImageViewType::Buffer, handles[0], info, info.range); + break; + } + default_handle = Handle(info.type); +} - const std::size_t source_size = src_surface->GetHostSizeInBytes(); - const std::size_t dest_size = dst_surface->GetHostSizeInBytes(); +ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::NullImageParams& params) + : VideoCommon::ImageViewBase{params}, views{runtime.null_image_views} {} - const std::size_t buffer_size = std::max(source_size, dest_size); +void ImageView::SetupView(const Device& device, Image& image, ImageViewType view_type, + GLuint handle, const VideoCommon::ImageViewInfo& info, + VideoCommon::SubresourceRange view_range) { + if (info.type == ImageViewType::Buffer) { + // TODO: Take offset from buffer cache + glTextureBufferRange(handle, internal_format, image.buffer.handle, 0, + image.guest_size_bytes); + } else { + const GLuint parent = image.texture.handle; + const GLenum target = ImageTarget(view_type, image.info.num_samples); + glTextureView(handle, target, parent, internal_format, view_range.base.level, + view_range.extent.levels, view_range.base.layer, view_range.extent.layers); + if (!info.IsRenderTarget()) { + ApplySwizzle(handle, format, info.Swizzle()); + } + } + if (device.HasDebuggingToolAttached()) { + const std::string name = VideoCommon::Name(*this, view_type); + glObjectLabel(GL_TEXTURE, handle, static_cast<GLsizei>(name.size()), name.data()); + } + stored_views.emplace_back().handle = handle; + views[static_cast<size_t>(view_type)] = handle; +} - GLuint copy_pbo_handle = FetchPBO(buffer_size); +Sampler::Sampler(TextureCacheRuntime& runtime, const TSCEntry& config) { + const GLenum compare_mode = config.depth_compare_enabled ? GL_COMPARE_REF_TO_TEXTURE : GL_NONE; + const GLenum compare_func = MaxwellToGL::DepthCompareFunc(config.depth_compare_func); + const GLenum mag = MaxwellToGL::TextureFilterMode(config.mag_filter, TextureMipmapFilter::None); + const GLenum min = MaxwellToGL::TextureFilterMode(config.min_filter, config.mipmap_filter); + const GLenum reduction_filter = MaxwellToGL::ReductionFilter(config.reduction_filter); + const GLint seamless = config.cubemap_interface_filtering ? GL_TRUE : GL_FALSE; + + UNIMPLEMENTED_IF(config.cubemap_anisotropy != 1); + UNIMPLEMENTED_IF(config.float_coord_normalization != 0); + + sampler.Create(); + const GLuint handle = sampler.handle; + glSamplerParameteri(handle, GL_TEXTURE_WRAP_S, MaxwellToGL::WrapMode(config.wrap_u)); + glSamplerParameteri(handle, GL_TEXTURE_WRAP_T, MaxwellToGL::WrapMode(config.wrap_v)); + glSamplerParameteri(handle, GL_TEXTURE_WRAP_R, MaxwellToGL::WrapMode(config.wrap_p)); + glSamplerParameteri(handle, GL_TEXTURE_COMPARE_MODE, compare_mode); + glSamplerParameteri(handle, GL_TEXTURE_COMPARE_FUNC, compare_func); + glSamplerParameteri(handle, GL_TEXTURE_MAG_FILTER, mag); + glSamplerParameteri(handle, GL_TEXTURE_MIN_FILTER, min); + glSamplerParameterf(handle, GL_TEXTURE_LOD_BIAS, config.LodBias()); + glSamplerParameterf(handle, GL_TEXTURE_MIN_LOD, config.MinLod()); + glSamplerParameterf(handle, GL_TEXTURE_MAX_LOD, config.MaxLod()); + glSamplerParameterfv(handle, GL_TEXTURE_BORDER_COLOR, config.BorderColor().data()); + + if (GLAD_GL_ARB_texture_filter_anisotropic || GLAD_GL_EXT_texture_filter_anisotropic) { + glSamplerParameterf(handle, GL_TEXTURE_MAX_ANISOTROPY, config.MaxAnisotropy()); + } else { + LOG_WARNING(Render_OpenGL, "GL_ARB_texture_filter_anisotropic is required"); + } + if (GLAD_GL_ARB_texture_filter_minmax || GLAD_GL_EXT_texture_filter_minmax) { + glSamplerParameteri(handle, GL_TEXTURE_REDUCTION_MODE_ARB, reduction_filter); + } else if (reduction_filter != GL_WEIGHTED_AVERAGE_ARB) { + LOG_WARNING(Render_OpenGL, "GL_ARB_texture_filter_minmax is required"); + } + if (GLAD_GL_ARB_seamless_cubemap_per_texture || GLAD_GL_AMD_seamless_cubemap_per_texture) { + glSamplerParameteri(handle, GL_TEXTURE_CUBE_MAP_SEAMLESS, seamless); + } else if (seamless == GL_FALSE) { + // We default to false because it's more common + LOG_WARNING(Render_OpenGL, "GL_ARB_seamless_cubemap_per_texture is required"); + } +} - glBindBuffer(GL_PIXEL_PACK_BUFFER, copy_pbo_handle); +Framebuffer::Framebuffer(TextureCacheRuntime& runtime, std::span<ImageView*, NUM_RT> color_buffers, + ImageView* depth_buffer, const VideoCommon::RenderTargets& key) { + // Bind to READ_FRAMEBUFFER to stop Nvidia's driver from creating an EXT_framebuffer instead of + // a core framebuffer. EXT framebuffer attachments have to match in size and can be shared + // across contexts. yuzu doesn't share framebuffers across contexts and we need attachments with + // mismatching size, this is why core framebuffers are preferred. + GLuint handle; + glGenFramebuffers(1, &handle); + glBindFramebuffer(GL_READ_FRAMEBUFFER, handle); + + GLsizei num_buffers = 0; + std::array<GLenum, NUM_RT> gl_draw_buffers; + gl_draw_buffers.fill(GL_NONE); + + for (size_t index = 0; index < color_buffers.size(); ++index) { + const ImageView* const image_view = color_buffers[index]; + if (!image_view) { + continue; + } + buffer_bits |= GL_COLOR_BUFFER_BIT; + gl_draw_buffers[index] = GL_COLOR_ATTACHMENT0 + key.draw_buffers[index]; + num_buffers = static_cast<GLsizei>(index + 1); - if (src_surface->IsCompressed()) { - glGetCompressedTextureImage(src_surface->GetTexture(), 0, static_cast<GLsizei>(source_size), - nullptr); - } else { - glGetTextureImage(src_surface->GetTexture(), 0, source_format.format, source_format.type, - static_cast<GLsizei>(source_size), nullptr); + const GLenum attachment = static_cast<GLenum>(GL_COLOR_ATTACHMENT0 + index); + AttachTexture(handle, attachment, image_view); } - glBindBuffer(GL_PIXEL_PACK_BUFFER, 0); - glBindBuffer(GL_PIXEL_UNPACK_BUFFER, copy_pbo_handle); + if (const ImageView* const image_view = depth_buffer; image_view) { + if (GetFormatType(image_view->format) == SurfaceType::DepthStencil) { + buffer_bits |= GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT; + } else { + buffer_bits |= GL_DEPTH_BUFFER_BIT; + } + const GLenum attachment = AttachmentType(image_view->format); + AttachTexture(handle, attachment, image_view); + } - const GLsizei width = static_cast<GLsizei>(dst_params.width); - const GLsizei height = static_cast<GLsizei>(dst_params.height); - const GLsizei depth = static_cast<GLsizei>(dst_params.depth); - if (dst_surface->IsCompressed()) { - LOG_CRITICAL(HW_GPU, "Compressed buffer copy is unimplemented!"); - UNREACHABLE(); + if (num_buffers > 1) { + glNamedFramebufferDrawBuffers(handle, num_buffers, gl_draw_buffers.data()); + } else if (num_buffers > 0) { + glNamedFramebufferDrawBuffer(handle, gl_draw_buffers[0]); } else { - switch (dst_params.target) { - case SurfaceTarget::Texture1D: - glTextureSubImage1D(dst_surface->GetTexture(), 0, 0, width, dest_format.format, - dest_format.type, nullptr); - break; - case SurfaceTarget::Texture2D: - glTextureSubImage2D(dst_surface->GetTexture(), 0, 0, 0, width, height, - dest_format.format, dest_format.type, nullptr); - break; - case SurfaceTarget::Texture3D: - case SurfaceTarget::Texture2DArray: - case SurfaceTarget::TextureCubeArray: - glTextureSubImage3D(dst_surface->GetTexture(), 0, 0, 0, 0, width, height, depth, - dest_format.format, dest_format.type, nullptr); - break; - case SurfaceTarget::TextureCubemap: - glTextureSubImage3D(dst_surface->GetTexture(), 0, 0, 0, 0, width, height, depth, - dest_format.format, dest_format.type, nullptr); - break; - default: - LOG_CRITICAL(Render_OpenGL, "Unimplemented surface target={}", - static_cast<u32>(dst_params.target)); - UNREACHABLE(); - } + glNamedFramebufferDrawBuffer(handle, GL_NONE); } - glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0); - glTextureBarrier(); -} + glNamedFramebufferParameteri(handle, GL_FRAMEBUFFER_DEFAULT_WIDTH, key.size.width); + glNamedFramebufferParameteri(handle, GL_FRAMEBUFFER_DEFAULT_HEIGHT, key.size.height); + // TODO + // glNamedFramebufferParameteri(handle, GL_FRAMEBUFFER_DEFAULT_LAYERS, ...); + // glNamedFramebufferParameteri(handle, GL_FRAMEBUFFER_DEFAULT_SAMPLES, ...); + // glNamedFramebufferParameteri(handle, GL_FRAMEBUFFER_DEFAULT_FIXED_SAMPLE_LOCATIONS, ...); -GLuint TextureCacheOpenGL::FetchPBO(std::size_t buffer_size) { - ASSERT_OR_EXECUTE(buffer_size > 0, { return 0; }); - const u32 l2 = Common::Log2Ceil64(static_cast<u64>(buffer_size)); - OGLBuffer& cp = copy_pbo_cache[l2]; - if (cp.handle == 0) { - const std::size_t ceil_size = 1ULL << l2; - cp.Create(); - cp.MakeStreamCopy(ceil_size); + if (runtime.device.HasDebuggingToolAttached()) { + const std::string name = VideoCommon::Name(key); + glObjectLabel(GL_FRAMEBUFFER, handle, static_cast<GLsizei>(name.size()), name.data()); } - return cp.handle; + framebuffer.handle = handle; } } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_texture_cache.h b/src/video_core/renderer_opengl/gl_texture_cache.h index 7787134fc..15b7c3676 100644 --- a/src/video_core/renderer_opengl/gl_texture_cache.h +++ b/src/video_core/renderer_opengl/gl_texture_cache.h @@ -4,156 +4,251 @@ #pragma once -#include <array> -#include <functional> #include <memory> -#include <unordered_map> -#include <utility> -#include <vector> +#include <span> #include <glad/glad.h> -#include "common/common_types.h" -#include "video_core/engines/shader_bytecode.h" -#include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_resource_manager.h" +#include "video_core/renderer_opengl/util_shaders.h" #include "video_core/texture_cache/texture_cache.h" namespace OpenGL { -using VideoCommon::SurfaceParams; -using VideoCommon::ViewParams; - -class CachedSurfaceView; -class CachedSurface; -class TextureCacheOpenGL; +class Device; +class ProgramManager; class StateTracker; -using Surface = std::shared_ptr<CachedSurface>; -using View = std::shared_ptr<CachedSurfaceView>; -using TextureCacheBase = VideoCommon::TextureCache<Surface, View>; +class Framebuffer; +class Image; +class ImageView; +class Sampler; -class CachedSurface final : public VideoCommon::SurfaceBase<View> { - friend CachedSurfaceView; +using VideoCommon::ImageId; +using VideoCommon::ImageViewId; +using VideoCommon::ImageViewType; +using VideoCommon::NUM_RT; +using VideoCommon::Offset2D; +using VideoCommon::RenderTargets; +class ImageBufferMap { public: - explicit CachedSurface(GPUVAddr gpu_addr, const SurfaceParams& params, bool is_astc_supported); - ~CachedSurface(); - - void UploadTexture(const std::vector<u8>& staging_buffer) override; - void DownloadTexture(std::vector<u8>& staging_buffer) override; + explicit ImageBufferMap(GLuint handle, u8* map, size_t size, OGLSync* sync); + ~ImageBufferMap(); - GLenum GetTarget() const { - return target; + GLuint Handle() const noexcept { + return handle; } - GLuint GetTexture() const { - return texture.handle; + std::span<u8> Span() const noexcept { + return span; } - bool IsCompressed() const { - return is_compressed; +private: + std::span<u8> span; + OGLSync* sync; + GLuint handle; +}; + +struct FormatProperties { + GLenum compatibility_class; + bool compatibility_by_size; + bool is_compressed; +}; + +class TextureCacheRuntime { + friend Framebuffer; + friend Image; + friend ImageView; + friend Sampler; + +public: + explicit TextureCacheRuntime(const Device& device, ProgramManager& program_manager, + StateTracker& state_tracker); + ~TextureCacheRuntime(); + + void Finish(); + + ImageBufferMap MapUploadBuffer(size_t size); + + ImageBufferMap MapDownloadBuffer(size_t size); + + void CopyImage(Image& dst, Image& src, std::span<const VideoCommon::ImageCopy> copies); + + void ConvertImage(Framebuffer* dst, ImageView& dst_view, ImageView& src_view) { + UNIMPLEMENTED(); } -protected: - void DecorateSurfaceName() override; + bool CanImageBeCopied(const Image& dst, const Image& src); + + void EmulateCopyImage(Image& dst, Image& src, std::span<const VideoCommon::ImageCopy> copies); + + void BlitFramebuffer(Framebuffer* dst, Framebuffer* src, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation); + + void AccelerateImageUpload(Image& image, const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::SwizzleParameters> swizzles); - View CreateView(const ViewParams& view_key) override; - View CreateViewInner(const ViewParams& view_key, bool is_proxy); + void InsertUploadMemoryBarrier(); + + FormatProperties FormatInfo(VideoCommon::ImageType type, GLenum internal_format) const; + + bool HasBrokenTextureViewFormats() const noexcept { + return has_broken_texture_view_formats; + } private: - void UploadTextureMipmap(u32 level, const std::vector<u8>& staging_buffer); + struct StagingBuffers { + explicit StagingBuffers(GLenum storage_flags_, GLenum map_flags_); + ~StagingBuffers(); - GLenum internal_format{}; - GLenum format{}; - GLenum type{}; - bool is_compressed{}; - GLenum target{}; - u32 view_count{}; + ImageBufferMap RequestMap(size_t requested_size, bool insert_fence); - OGLTexture texture; - OGLBuffer texture_buffer; + size_t RequestBuffer(size_t requested_size); + + std::optional<size_t> FindBuffer(size_t requested_size); + + std::vector<OGLSync> syncs; + std::vector<OGLBuffer> buffers; + std::vector<u8*> maps; + std::vector<size_t> sizes; + GLenum storage_flags; + GLenum map_flags; + }; + + const Device& device; + StateTracker& state_tracker; + UtilShaders util_shaders; + + std::array<std::unordered_map<GLenum, FormatProperties>, 3> format_properties; + bool has_broken_texture_view_formats = false; + + StagingBuffers upload_buffers{GL_MAP_WRITE_BIT, GL_MAP_WRITE_BIT | GL_MAP_FLUSH_EXPLICIT_BIT}; + StagingBuffers download_buffers{GL_MAP_READ_BIT, GL_MAP_READ_BIT}; + + OGLTexture null_image_1d_array; + OGLTexture null_image_cube_array; + OGLTexture null_image_3d; + OGLTexture null_image_rect; + OGLTextureView null_image_view_1d; + OGLTextureView null_image_view_2d; + OGLTextureView null_image_view_2d_array; + OGLTextureView null_image_view_cube; + + std::array<GLuint, VideoCommon::NUM_IMAGE_VIEW_TYPES> null_image_views; }; -class CachedSurfaceView final : public VideoCommon::ViewBase { +class Image : public VideoCommon::ImageBase { + friend ImageView; + public: - explicit CachedSurfaceView(CachedSurface& surface, const ViewParams& params, bool is_proxy); - ~CachedSurfaceView(); + explicit Image(TextureCacheRuntime&, const VideoCommon::ImageInfo& info, GPUVAddr gpu_addr, + VAddr cpu_addr); - /// @brief Attaches this texture view to the currently bound fb_target framebuffer - /// @param attachment Attachment to bind textures to - /// @param fb_target Framebuffer target to attach to (e.g. DRAW_FRAMEBUFFER) - void Attach(GLenum attachment, GLenum fb_target) const; + void UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferImageCopy> copies); - GLuint GetTexture(Tegra::Texture::SwizzleSource x_source, - Tegra::Texture::SwizzleSource y_source, - Tegra::Texture::SwizzleSource z_source, - Tegra::Texture::SwizzleSource w_source); + void UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferCopy> copies); - void DecorateViewName(GPUVAddr gpu_addr, const std::string& prefix); + void DownloadMemory(ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferImageCopy> copies); - void MarkAsModified(u64 tick) { - surface.MarkAsModified(true, tick); + GLuint Handle() const noexcept { + return texture.handle; } - GLuint GetTexture() const { - if (is_proxy) { - return surface.GetTexture(); - } - return main_view.handle; +private: + void CopyBufferToImage(const VideoCommon::BufferImageCopy& copy, size_t buffer_offset); + + void CopyImageToBuffer(const VideoCommon::BufferImageCopy& copy, size_t buffer_offset); + + OGLTexture texture; + OGLTextureView store_view; + OGLBuffer buffer; + GLenum gl_internal_format = GL_NONE; + GLenum gl_format = GL_NONE; + GLenum gl_type = GL_NONE; +}; + +class ImageView : public VideoCommon::ImageViewBase { + friend Image; + +public: + explicit ImageView(TextureCacheRuntime&, const VideoCommon::ImageViewInfo&, ImageId, Image&); + explicit ImageView(TextureCacheRuntime&, const VideoCommon::NullImageParams&); + + [[nodiscard]] GLuint Handle(ImageViewType query_type) const noexcept { + return views[static_cast<size_t>(query_type)]; } - GLenum GetFormat() const { - return format; + [[nodiscard]] GLuint DefaultHandle() const noexcept { + return default_handle; } - const SurfaceParams& GetSurfaceParams() const { - return surface.GetSurfaceParams(); + [[nodiscard]] GLenum Format() const noexcept { + return internal_format; } private: - OGLTextureView CreateTextureView() const; + void SetupView(const Device& device, Image& image, ImageViewType view_type, GLuint handle, + const VideoCommon::ImageViewInfo& info, + VideoCommon::SubresourceRange view_range); + + std::array<GLuint, VideoCommon::NUM_IMAGE_VIEW_TYPES> views{}; + std::vector<OGLTextureView> stored_views; + GLuint default_handle = 0; + GLenum internal_format = GL_NONE; +}; + +class ImageAlloc : public VideoCommon::ImageAllocBase {}; - CachedSurface& surface; - const GLenum format; - const GLenum target; - const bool is_proxy; +class Sampler { +public: + explicit Sampler(TextureCacheRuntime&, const Tegra::Texture::TSCEntry&); - std::unordered_map<u32, OGLTextureView> view_cache; - OGLTextureView main_view; + GLuint Handle() const noexcept { + return sampler.handle; + } - // Use an invalid default so it always fails the comparison test - u32 current_swizzle = 0xffffffff; - GLuint current_view = 0; +private: + OGLSampler sampler; }; -class TextureCacheOpenGL final : public TextureCacheBase { +class Framebuffer { public: - explicit TextureCacheOpenGL(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, - Tegra::MemoryManager& gpu_memory, const Device& device, - StateTracker& state_tracker); - ~TextureCacheOpenGL(); - -protected: - Surface CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) override; - - void ImageCopy(Surface& src_surface, Surface& dst_surface, - const VideoCommon::CopyParams& copy_params) override; + explicit Framebuffer(TextureCacheRuntime&, std::span<ImageView*, NUM_RT> color_buffers, + ImageView* depth_buffer, const VideoCommon::RenderTargets& key); - void ImageBlit(View& src_view, View& dst_view, - const Tegra::Engines::Fermi2D::Config& copy_config) override; + [[nodiscard]] GLuint Handle() const noexcept { + return framebuffer.handle; + } - void BufferCopy(Surface& src_surface, Surface& dst_surface) override; + [[nodiscard]] GLbitfield BufferBits() const noexcept { + return buffer_bits; + } private: - GLuint FetchPBO(std::size_t buffer_size); - - StateTracker& state_tracker; + OGLFramebuffer framebuffer; + GLbitfield buffer_bits = GL_NONE; +}; - OGLFramebuffer src_framebuffer; - OGLFramebuffer dst_framebuffer; - std::unordered_map<u32, OGLBuffer> copy_pbo_cache; +struct TextureCacheParams { + static constexpr bool ENABLE_VALIDATION = true; + static constexpr bool FRAMEBUFFER_BLITS = true; + static constexpr bool HAS_EMULATED_COPIES = true; + + using Runtime = OpenGL::TextureCacheRuntime; + using Image = OpenGL::Image; + using ImageAlloc = OpenGL::ImageAlloc; + using ImageView = OpenGL::ImageView; + using Sampler = OpenGL::Sampler; + using Framebuffer = OpenGL::Framebuffer; }; +using TextureCache = VideoCommon::TextureCache<TextureCacheParams>; + } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/maxwell_to_gl.h b/src/video_core/renderer_opengl/maxwell_to_gl.h index a8be2aa37..cbccfdeb4 100644 --- a/src/video_core/renderer_opengl/maxwell_to_gl.h +++ b/src/video_core/renderer_opengl/maxwell_to_gl.h @@ -107,7 +107,7 @@ inline GLenum IndexFormat(Maxwell::IndexFormat index_format) { case Maxwell::IndexFormat::UnsignedInt: return GL_UNSIGNED_INT; } - UNREACHABLE_MSG("Invalid index_format={}", static_cast<u32>(index_format)); + UNREACHABLE_MSG("Invalid index_format={}", index_format); return {}; } @@ -144,7 +144,7 @@ inline GLenum PrimitiveTopology(Maxwell::PrimitiveTopology topology) { case Maxwell::PrimitiveTopology::Patches: return GL_PATCHES; } - UNREACHABLE_MSG("Invalid topology={}", static_cast<int>(topology)); + UNREACHABLE_MSG("Invalid topology={}", topology); return GL_POINTS; } @@ -172,8 +172,8 @@ inline GLenum TextureFilterMode(Tegra::Texture::TextureFilter filter_mode, } break; } - UNREACHABLE_MSG("Invalid texture filter mode={} and mipmap filter mode={}", - static_cast<u32>(filter_mode), static_cast<u32>(mipmap_filter_mode)); + UNREACHABLE_MSG("Invalid texture filter mode={} and mipmap filter mode={}", filter_mode, + mipmap_filter_mode); return GL_NEAREST; } @@ -204,7 +204,7 @@ inline GLenum WrapMode(Tegra::Texture::WrapMode wrap_mode) { return GL_MIRROR_CLAMP_TO_EDGE; } } - UNIMPLEMENTED_MSG("Unimplemented texture wrap mode={}", static_cast<u32>(wrap_mode)); + UNIMPLEMENTED_MSG("Unimplemented texture wrap mode={}", wrap_mode); return GL_REPEAT; } @@ -227,7 +227,7 @@ inline GLenum DepthCompareFunc(Tegra::Texture::DepthCompareFunc func) { case Tegra::Texture::DepthCompareFunc::Always: return GL_ALWAYS; } - UNIMPLEMENTED_MSG("Unimplemented texture depth compare function={}", static_cast<u32>(func)); + UNIMPLEMENTED_MSG("Unimplemented texture depth compare function={}", func); return GL_GREATER; } @@ -249,7 +249,7 @@ inline GLenum BlendEquation(Maxwell::Blend::Equation equation) { case Maxwell::Blend::Equation::MaxGL: return GL_MAX; } - UNIMPLEMENTED_MSG("Unimplemented blend equation={}", static_cast<u32>(equation)); + UNIMPLEMENTED_MSG("Unimplemented blend equation={}", equation); return GL_FUNC_ADD; } @@ -313,7 +313,7 @@ inline GLenum BlendFunc(Maxwell::Blend::Factor factor) { case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: return GL_ONE_MINUS_CONSTANT_ALPHA; } - UNIMPLEMENTED_MSG("Unimplemented blend factor={}", static_cast<u32>(factor)); + UNIMPLEMENTED_MSG("Unimplemented blend factor={}", factor); return GL_ZERO; } @@ -333,7 +333,7 @@ inline GLenum SwizzleSource(Tegra::Texture::SwizzleSource source) { case Tegra::Texture::SwizzleSource::OneFloat: return GL_ONE; } - UNIMPLEMENTED_MSG("Unimplemented swizzle source={}", static_cast<u32>(source)); + UNIMPLEMENTED_MSG("Unimplemented swizzle source={}", source); return GL_ZERO; } @@ -364,7 +364,7 @@ inline GLenum ComparisonOp(Maxwell::ComparisonOp comparison) { case Maxwell::ComparisonOp::AlwaysOld: return GL_ALWAYS; } - UNIMPLEMENTED_MSG("Unimplemented comparison op={}", static_cast<u32>(comparison)); + UNIMPLEMENTED_MSG("Unimplemented comparison op={}", comparison); return GL_ALWAYS; } @@ -395,7 +395,7 @@ inline GLenum StencilOp(Maxwell::StencilOp stencil) { case Maxwell::StencilOp::DecrWrapOGL: return GL_DECR_WRAP; } - UNIMPLEMENTED_MSG("Unimplemented stencil op={}", static_cast<u32>(stencil)); + UNIMPLEMENTED_MSG("Unimplemented stencil op={}", stencil); return GL_KEEP; } @@ -406,7 +406,7 @@ inline GLenum FrontFace(Maxwell::FrontFace front_face) { case Maxwell::FrontFace::CounterClockWise: return GL_CCW; } - UNIMPLEMENTED_MSG("Unimplemented front face cull={}", static_cast<u32>(front_face)); + UNIMPLEMENTED_MSG("Unimplemented front face cull={}", front_face); return GL_CCW; } @@ -419,7 +419,7 @@ inline GLenum CullFace(Maxwell::CullFace cull_face) { case Maxwell::CullFace::FrontAndBack: return GL_FRONT_AND_BACK; } - UNIMPLEMENTED_MSG("Unimplemented cull face={}", static_cast<u32>(cull_face)); + UNIMPLEMENTED_MSG("Unimplemented cull face={}", cull_face); return GL_BACK; } @@ -458,7 +458,7 @@ inline GLenum LogicOp(Maxwell::LogicOperation operation) { case Maxwell::LogicOperation::Set: return GL_SET; } - UNIMPLEMENTED_MSG("Unimplemented logic operation={}", static_cast<u32>(operation)); + UNIMPLEMENTED_MSG("Unimplemented logic operation={}", operation); return GL_COPY; } @@ -471,10 +471,23 @@ inline GLenum PolygonMode(Maxwell::PolygonMode polygon_mode) { case Maxwell::PolygonMode::Fill: return GL_FILL; } - UNREACHABLE_MSG("Invalid polygon mode={}", static_cast<int>(polygon_mode)); + UNREACHABLE_MSG("Invalid polygon mode={}", polygon_mode); return GL_FILL; } +inline GLenum ReductionFilter(Tegra::Texture::SamplerReduction filter) { + switch (filter) { + case Tegra::Texture::SamplerReduction::WeightedAverage: + return GL_WEIGHTED_AVERAGE_ARB; + case Tegra::Texture::SamplerReduction::Min: + return GL_MIN; + case Tegra::Texture::SamplerReduction::Max: + return GL_MAX; + } + UNREACHABLE_MSG("Invalid reduction filter={}", static_cast<int>(filter)); + return GL_WEIGHTED_AVERAGE_ARB; +} + inline GLenum ViewportSwizzle(Maxwell::ViewportSwizzle swizzle) { // Enumeration order matches register order. We can convert it arithmetically. return GL_VIEWPORT_SWIZZLE_POSITIVE_X_NV + static_cast<GLenum>(swizzle); diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index 2ccca1993..dd77a543c 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -23,10 +23,10 @@ #include "core/telemetry_session.h" #include "video_core/host_shaders/opengl_present_frag.h" #include "video_core/host_shaders/opengl_present_vert.h" -#include "video_core/morton.h" #include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_shader_manager.h" #include "video_core/renderer_opengl/renderer_opengl.h" +#include "video_core/textures/decoders.h" namespace OpenGL { @@ -130,8 +130,8 @@ void APIENTRY DebugHandler(GLenum source, GLenum type, GLuint id, GLenum severit RendererOpenGL::RendererOpenGL(Core::TelemetrySession& telemetry_session_, Core::Frontend::EmuWindow& emu_window_, Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, - std::unique_ptr<Core::Frontend::GraphicsContext> context) - : RendererBase{emu_window_, std::move(context)}, telemetry_session{telemetry_session_}, + std::unique_ptr<Core::Frontend::GraphicsContext> context_) + : RendererBase{emu_window_, std::move(context_)}, telemetry_session{telemetry_session_}, emu_window{emu_window_}, cpu_memory{cpu_memory_}, gpu{gpu_}, program_manager{device} {} RendererOpenGL::~RendererOpenGL() = default; @@ -140,19 +140,18 @@ void RendererOpenGL::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { if (!framebuffer) { return; } - PrepareRendertarget(framebuffer); RenderScreenshot(); - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, 0); + state_tracker.BindFramebuffer(0); DrawScreen(emu_window.GetFramebufferLayout()); ++m_current_frame; rasterizer->TickFrame(); - render_window.PollEvents(); context->SwapBuffers(); + render_window.OnFrameDisplayed(); } void RendererOpenGL::PrepareRendertarget(const Tegra::FramebufferConfig* framebuffer) { @@ -187,19 +186,20 @@ void RendererOpenGL::LoadFBToScreenInfo(const Tegra::FramebufferConfig& framebuf // Reset the screen info's display texture to its own permanent texture screen_info.display_texture = screen_info.texture.resource.handle; - const auto pixel_format{ - VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)}; - const u32 bytes_per_pixel{VideoCore::Surface::GetBytesPerPixel(pixel_format)}; - const u64 size_in_bytes{framebuffer.stride * framebuffer.height * bytes_per_pixel}; - u8* const host_ptr{cpu_memory.GetPointer(framebuffer_addr)}; - rasterizer->FlushRegion(ToCacheAddr(host_ptr), size_in_bytes); - // TODO(Rodrigo): Read this from HLE constexpr u32 block_height_log2 = 4; - VideoCore::MortonSwizzle(VideoCore::MortonSwizzleMode::MortonToLinear, pixel_format, - framebuffer.stride, block_height_log2, framebuffer.height, 0, 1, 1, - gl_framebuffer_data.data(), host_ptr); - + const auto pixel_format{ + VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)}; + const u32 bytes_per_pixel{VideoCore::Surface::BytesPerBlock(pixel_format)}; + const u64 size_in_bytes{Tegra::Texture::CalculateSize( + true, bytes_per_pixel, framebuffer.stride, framebuffer.height, 1, block_height_log2, 0)}; + const u8* const host_ptr{cpu_memory.GetPointer(framebuffer_addr)}; + const std::span<const u8> input_data(host_ptr, size_in_bytes); + Tegra::Texture::UnswizzleTexture(gl_framebuffer_data, input_data, bytes_per_pixel, + framebuffer.width, framebuffer.height, 1, block_height_log2, + 0); + + glBindBuffer(GL_PIXEL_UNPACK_BUFFER, 0); glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(framebuffer.stride)); // Update existing texture @@ -238,6 +238,10 @@ void RendererOpenGL::InitOpenGLObjects() { glUseProgramStages(pipeline.handle, GL_VERTEX_SHADER_BIT, vertex_program.handle); glUseProgramStages(pipeline.handle, GL_FRAGMENT_SHADER_BIT, fragment_program.handle); + // Generate presentation sampler + present_sampler.Create(); + glSamplerParameteri(present_sampler.handle, GL_TEXTURE_MIN_FILTER, GL_LINEAR); + // Generate VBO handle for drawing vertex_buffer.Create(); @@ -255,6 +259,11 @@ void RendererOpenGL::InitOpenGLObjects() { // Clear screen to black LoadColorToActiveGLTexture(0, 0, 0, 0, screen_info.texture); + // Enable seamless cubemaps when per texture parameters are not available + if (!GLAD_GL_ARB_seamless_cubemap_per_texture && !GLAD_GL_AMD_seamless_cubemap_per_texture) { + glEnable(GL_TEXTURE_CUBE_MAP_SEAMLESS); + } + // Enable unified vertex attributes and query vertex buffer address when the driver supports it if (device.HasVertexBufferUnifiedMemory()) { glEnableClientState(GL_VERTEX_ATTRIB_ARRAY_UNIFIED_NV); @@ -275,9 +284,9 @@ void RendererOpenGL::AddTelemetryFields() { LOG_INFO(Render_OpenGL, "GL_RENDERER: {}", gpu_model); constexpr auto user_system = Common::Telemetry::FieldType::UserSystem; - telemetry_session.AddField(user_system, "GPU_Vendor", gpu_vendor); - telemetry_session.AddField(user_system, "GPU_Model", gpu_model); - telemetry_session.AddField(user_system, "GPU_OpenGL_Version", gl_version); + telemetry_session.AddField(user_system, "GPU_Vendor", std::string(gpu_vendor)); + telemetry_session.AddField(user_system, "GPU_Model", std::string(gpu_model)); + telemetry_session.AddField(user_system, "GPU_OpenGL_Version", std::string(gl_version)); } void RendererOpenGL::CreateRasterizer() { @@ -296,7 +305,7 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture, const auto pixel_format{ VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)}; - const u32 bytes_per_pixel{VideoCore::Surface::GetBytesPerPixel(pixel_format)}; + const u32 bytes_per_pixel{VideoCore::Surface::BytesPerBlock(pixel_format)}; gl_framebuffer_data.resize(texture.width * texture.height * bytes_per_pixel); GLint internal_format; @@ -315,8 +324,8 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture, internal_format = GL_RGBA8; texture.gl_format = GL_RGBA; texture.gl_type = GL_UNSIGNED_INT_8_8_8_8_REV; - UNIMPLEMENTED_MSG("Unknown framebuffer pixel format: {}", - static_cast<u32>(framebuffer.pixel_format)); + // UNIMPLEMENTED_MSG("Unknown framebuffer pixel format: {}", + // static_cast<u32>(framebuffer.pixel_format)); } texture.resource.Release(); @@ -348,7 +357,7 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) { } else { // Other transformations are unsupported LOG_CRITICAL(Render_OpenGL, "Unsupported framebuffer_transform_flags={}", - static_cast<u32>(framebuffer_transform_flags)); + framebuffer_transform_flags); UNIMPLEMENTED(); } } @@ -382,7 +391,7 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) { state_tracker.NotifyPolygonModes(); state_tracker.NotifyViewport0(); state_tracker.NotifyScissor0(); - state_tracker.NotifyColorMask0(); + state_tracker.NotifyColorMask(0); state_tracker.NotifyBlend0(); state_tracker.NotifyFramebuffer(); state_tracker.NotifyFrontFace(); @@ -440,7 +449,7 @@ void RendererOpenGL::DrawScreen(const Layout::FramebufferLayout& layout) { } glBindTextureUnit(0, screen_info.display_texture); - glBindSampler(0, 0); + glBindSampler(0, present_sampler.handle); glClear(GL_COLOR_BUFFER_BIT); glDrawArrays(GL_TRIANGLE_STRIP, 0, 4); @@ -473,6 +482,8 @@ void RendererOpenGL::RenderScreenshot() { DrawScreen(layout); + glBindBuffer(GL_PIXEL_PACK_BUFFER, 0); + glPixelStorei(GL_PACK_ROW_LENGTH, 0); glReadPixels(0, 0, layout.width, layout.height, GL_BGRA, GL_UNSIGNED_INT_8_8_8_8_REV, renderer_settings.screenshot_bits); diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h index 9ef181f95..44e109794 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.h +++ b/src/video_core/renderer_opengl/renderer_opengl.h @@ -57,10 +57,10 @@ struct ScreenInfo { class RendererOpenGL final : public VideoCore::RendererBase { public: - explicit RendererOpenGL(Core::TelemetrySession& telemetry_session, - Core::Frontend::EmuWindow& emu_window, Core::Memory::Memory& cpu_memory, - Tegra::GPU& gpu, - std::unique_ptr<Core::Frontend::GraphicsContext> context); + explicit RendererOpenGL(Core::TelemetrySession& telemetry_session_, + Core::Frontend::EmuWindow& emu_window_, + Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, + std::unique_ptr<Core::Frontend::GraphicsContext> context_); ~RendererOpenGL() override; bool Init() override; @@ -102,6 +102,7 @@ private: StateTracker state_tracker{gpu}; // OpenGL object IDs + OGLSampler present_sampler; OGLBuffer vertex_buffer; OGLProgram vertex_program; OGLProgram fragment_program; diff --git a/src/video_core/renderer_opengl/util_shaders.cpp b/src/video_core/renderer_opengl/util_shaders.cpp new file mode 100644 index 000000000..eb849cbf2 --- /dev/null +++ b/src/video_core/renderer_opengl/util_shaders.cpp @@ -0,0 +1,224 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <bit> +#include <span> +#include <string_view> + +#include <glad/glad.h> + +#include "common/assert.h" +#include "common/common_types.h" +#include "common/div_ceil.h" +#include "video_core/host_shaders/block_linear_unswizzle_2d_comp.h" +#include "video_core/host_shaders/block_linear_unswizzle_3d_comp.h" +#include "video_core/host_shaders/opengl_copy_bc4_comp.h" +#include "video_core/host_shaders/pitch_unswizzle_comp.h" +#include "video_core/renderer_opengl/gl_resource_manager.h" +#include "video_core/renderer_opengl/gl_shader_manager.h" +#include "video_core/renderer_opengl/gl_texture_cache.h" +#include "video_core/renderer_opengl/util_shaders.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/accelerated_swizzle.h" +#include "video_core/texture_cache/types.h" +#include "video_core/texture_cache/util.h" +#include "video_core/textures/decoders.h" + +namespace OpenGL { + +using namespace HostShaders; + +using VideoCommon::Extent3D; +using VideoCommon::ImageCopy; +using VideoCommon::ImageType; +using VideoCommon::SwizzleParameters; +using VideoCommon::Accelerated::MakeBlockLinearSwizzle2DParams; +using VideoCommon::Accelerated::MakeBlockLinearSwizzle3DParams; +using VideoCore::Surface::BytesPerBlock; + +namespace { + +OGLProgram MakeProgram(std::string_view source) { + OGLShader shader; + shader.Create(source, GL_COMPUTE_SHADER); + + OGLProgram program; + program.Create(true, false, shader.handle); + return program; +} + +} // Anonymous namespace + +UtilShaders::UtilShaders(ProgramManager& program_manager_) + : program_manager{program_manager_}, + block_linear_unswizzle_2d_program(MakeProgram(BLOCK_LINEAR_UNSWIZZLE_2D_COMP)), + block_linear_unswizzle_3d_program(MakeProgram(BLOCK_LINEAR_UNSWIZZLE_3D_COMP)), + pitch_unswizzle_program(MakeProgram(PITCH_UNSWIZZLE_COMP)), + copy_bc4_program(MakeProgram(OPENGL_COPY_BC4_COMP)) { + const auto swizzle_table = Tegra::Texture::MakeSwizzleTable(); + swizzle_table_buffer.Create(); + glNamedBufferStorage(swizzle_table_buffer.handle, sizeof(swizzle_table), &swizzle_table, 0); +} + +UtilShaders::~UtilShaders() = default; + +void UtilShaders::BlockLinearUpload2D(Image& image, const ImageBufferMap& map, size_t buffer_offset, + std::span<const SwizzleParameters> swizzles) { + static constexpr Extent3D WORKGROUP_SIZE{32, 32, 1}; + static constexpr GLuint BINDING_SWIZZLE_BUFFER = 0; + static constexpr GLuint BINDING_INPUT_BUFFER = 1; + static constexpr GLuint BINDING_OUTPUT_IMAGE = 0; + + program_manager.BindHostCompute(block_linear_unswizzle_2d_program.handle); + glFlushMappedNamedBufferRange(map.Handle(), buffer_offset, image.guest_size_bytes); + glBindBufferBase(GL_SHADER_STORAGE_BUFFER, BINDING_SWIZZLE_BUFFER, swizzle_table_buffer.handle); + + const GLenum store_format = StoreFormat(BytesPerBlock(image.info.format)); + for (const SwizzleParameters& swizzle : swizzles) { + const Extent3D num_tiles = swizzle.num_tiles; + const size_t input_offset = swizzle.buffer_offset + buffer_offset; + + const u32 num_dispatches_x = Common::DivCeil(num_tiles.width, WORKGROUP_SIZE.width); + const u32 num_dispatches_y = Common::DivCeil(num_tiles.height, WORKGROUP_SIZE.height); + + const auto params = MakeBlockLinearSwizzle2DParams(swizzle, image.info); + glUniform3uiv(0, 1, params.origin.data()); + glUniform3iv(1, 1, params.destination.data()); + glUniform1ui(2, params.bytes_per_block_log2); + glUniform1ui(3, params.layer_stride); + glUniform1ui(4, params.block_size); + glUniform1ui(5, params.x_shift); + glUniform1ui(6, params.block_height); + glUniform1ui(7, params.block_height_mask); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_INPUT_BUFFER, map.Handle(), + input_offset, image.guest_size_bytes - swizzle.buffer_offset); + glBindImageTexture(BINDING_OUTPUT_IMAGE, image.Handle(), swizzle.level, GL_TRUE, 0, + GL_WRITE_ONLY, store_format); + glDispatchCompute(num_dispatches_x, num_dispatches_y, image.info.resources.layers); + } + program_manager.RestoreGuestCompute(); +} + +void UtilShaders::BlockLinearUpload3D(Image& image, const ImageBufferMap& map, size_t buffer_offset, + std::span<const SwizzleParameters> swizzles) { + static constexpr Extent3D WORKGROUP_SIZE{16, 8, 8}; + + static constexpr GLuint BINDING_SWIZZLE_BUFFER = 0; + static constexpr GLuint BINDING_INPUT_BUFFER = 1; + static constexpr GLuint BINDING_OUTPUT_IMAGE = 0; + + glFlushMappedNamedBufferRange(map.Handle(), buffer_offset, image.guest_size_bytes); + program_manager.BindHostCompute(block_linear_unswizzle_3d_program.handle); + glBindBufferBase(GL_SHADER_STORAGE_BUFFER, BINDING_SWIZZLE_BUFFER, swizzle_table_buffer.handle); + + const GLenum store_format = StoreFormat(BytesPerBlock(image.info.format)); + for (const SwizzleParameters& swizzle : swizzles) { + const Extent3D num_tiles = swizzle.num_tiles; + const size_t input_offset = swizzle.buffer_offset + buffer_offset; + + const u32 num_dispatches_x = Common::DivCeil(num_tiles.width, WORKGROUP_SIZE.width); + const u32 num_dispatches_y = Common::DivCeil(num_tiles.height, WORKGROUP_SIZE.height); + const u32 num_dispatches_z = Common::DivCeil(num_tiles.depth, WORKGROUP_SIZE.depth); + + const auto params = MakeBlockLinearSwizzle3DParams(swizzle, image.info); + glUniform3uiv(0, 1, params.origin.data()); + glUniform3iv(1, 1, params.destination.data()); + glUniform1ui(2, params.bytes_per_block_log2); + glUniform1ui(3, params.slice_size); + glUniform1ui(4, params.block_size); + glUniform1ui(5, params.x_shift); + glUniform1ui(6, params.block_height); + glUniform1ui(7, params.block_height_mask); + glUniform1ui(8, params.block_depth); + glUniform1ui(9, params.block_depth_mask); + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_INPUT_BUFFER, map.Handle(), + input_offset, image.guest_size_bytes - swizzle.buffer_offset); + glBindImageTexture(BINDING_OUTPUT_IMAGE, image.Handle(), swizzle.level, GL_TRUE, 0, + GL_WRITE_ONLY, store_format); + glDispatchCompute(num_dispatches_x, num_dispatches_y, num_dispatches_z); + } + program_manager.RestoreGuestCompute(); +} + +void UtilShaders::PitchUpload(Image& image, const ImageBufferMap& map, size_t buffer_offset, + std::span<const SwizzleParameters> swizzles) { + static constexpr Extent3D WORKGROUP_SIZE{32, 32, 1}; + static constexpr GLuint BINDING_INPUT_BUFFER = 0; + static constexpr GLuint BINDING_OUTPUT_IMAGE = 0; + static constexpr GLuint LOC_ORIGIN = 0; + static constexpr GLuint LOC_DESTINATION = 1; + static constexpr GLuint LOC_BYTES_PER_BLOCK = 2; + static constexpr GLuint LOC_PITCH = 3; + + const u32 bytes_per_block = BytesPerBlock(image.info.format); + const GLenum format = StoreFormat(bytes_per_block); + const u32 pitch = image.info.pitch; + + UNIMPLEMENTED_IF_MSG(!std::has_single_bit(bytes_per_block), + "Non-power of two images are not implemented"); + + program_manager.BindHostCompute(pitch_unswizzle_program.handle); + glFlushMappedNamedBufferRange(map.Handle(), buffer_offset, image.guest_size_bytes); + glUniform2ui(LOC_ORIGIN, 0, 0); + glUniform2i(LOC_DESTINATION, 0, 0); + glUniform1ui(LOC_BYTES_PER_BLOCK, bytes_per_block); + glUniform1ui(LOC_PITCH, pitch); + glBindImageTexture(BINDING_OUTPUT_IMAGE, image.Handle(), 0, GL_FALSE, 0, GL_WRITE_ONLY, format); + for (const SwizzleParameters& swizzle : swizzles) { + const Extent3D num_tiles = swizzle.num_tiles; + const size_t input_offset = swizzle.buffer_offset + buffer_offset; + + const u32 num_dispatches_x = Common::DivCeil(num_tiles.width, WORKGROUP_SIZE.width); + const u32 num_dispatches_y = Common::DivCeil(num_tiles.height, WORKGROUP_SIZE.height); + + glBindBufferRange(GL_SHADER_STORAGE_BUFFER, BINDING_INPUT_BUFFER, map.Handle(), + input_offset, image.guest_size_bytes - swizzle.buffer_offset); + glDispatchCompute(num_dispatches_x, num_dispatches_y, 1); + } + program_manager.RestoreGuestCompute(); +} + +void UtilShaders::CopyBC4(Image& dst_image, Image& src_image, std::span<const ImageCopy> copies) { + static constexpr GLuint BINDING_INPUT_IMAGE = 0; + static constexpr GLuint BINDING_OUTPUT_IMAGE = 1; + static constexpr GLuint LOC_SRC_OFFSET = 0; + static constexpr GLuint LOC_DST_OFFSET = 1; + + program_manager.BindHostCompute(copy_bc4_program.handle); + + for (const ImageCopy& copy : copies) { + ASSERT(copy.src_subresource.base_layer == 0); + ASSERT(copy.src_subresource.num_layers == 1); + ASSERT(copy.dst_subresource.base_layer == 0); + ASSERT(copy.dst_subresource.num_layers == 1); + + glUniform3ui(LOC_SRC_OFFSET, copy.src_offset.x, copy.src_offset.y, copy.src_offset.z); + glUniform3ui(LOC_DST_OFFSET, copy.dst_offset.x, copy.dst_offset.y, copy.dst_offset.z); + glBindImageTexture(BINDING_INPUT_IMAGE, src_image.Handle(), copy.src_subresource.base_level, + GL_FALSE, 0, GL_READ_ONLY, GL_RG32UI); + glBindImageTexture(BINDING_OUTPUT_IMAGE, dst_image.Handle(), + copy.dst_subresource.base_level, GL_FALSE, 0, GL_WRITE_ONLY, GL_RGBA8UI); + glDispatchCompute(copy.extent.width, copy.extent.height, copy.extent.depth); + } + program_manager.RestoreGuestCompute(); +} + +GLenum StoreFormat(u32 bytes_per_block) { + switch (bytes_per_block) { + case 1: + return GL_R8UI; + case 2: + return GL_R16UI; + case 4: + return GL_R32UI; + case 8: + return GL_RG32UI; + case 16: + return GL_RGBA32UI; + } + UNREACHABLE(); + return GL_R8UI; +} + +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/util_shaders.h b/src/video_core/renderer_opengl/util_shaders.h new file mode 100644 index 000000000..359997255 --- /dev/null +++ b/src/video_core/renderer_opengl/util_shaders.h @@ -0,0 +1,51 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <span> + +#include <glad/glad.h> + +#include "common/common_types.h" +#include "video_core/renderer_opengl/gl_resource_manager.h" +#include "video_core/texture_cache/types.h" + +namespace OpenGL { + +class Image; +class ImageBufferMap; +class ProgramManager; + +class UtilShaders { +public: + explicit UtilShaders(ProgramManager& program_manager); + ~UtilShaders(); + + void BlockLinearUpload2D(Image& image, const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::SwizzleParameters> swizzles); + + void BlockLinearUpload3D(Image& image, const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::SwizzleParameters> swizzles); + + void PitchUpload(Image& image, const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::SwizzleParameters> swizzles); + + void CopyBC4(Image& dst_image, Image& src_image, + std::span<const VideoCommon::ImageCopy> copies); + +private: + ProgramManager& program_manager; + + OGLBuffer swizzle_table_buffer; + + OGLProgram block_linear_unswizzle_2d_program; + OGLProgram block_linear_unswizzle_3d_program; + OGLProgram pitch_unswizzle_program; + OGLProgram copy_bc4_program; +}; + +GLenum StoreFormat(u32 bytes_per_block); + +} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/utils.cpp b/src/video_core/renderer_opengl/utils.cpp deleted file mode 100644 index 6d7bb16b2..000000000 --- a/src/video_core/renderer_opengl/utils.cpp +++ /dev/null @@ -1,42 +0,0 @@ -// Copyright 2014 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <string> -#include <vector> - -#include <fmt/format.h> -#include <glad/glad.h> - -#include "common/common_types.h" -#include "video_core/renderer_opengl/gl_state_tracker.h" -#include "video_core/renderer_opengl/utils.h" - -namespace OpenGL { - -void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string_view extra_info) { - if (!GLAD_GL_KHR_debug) { - // We don't need to throw an error as this is just for debugging - return; - } - - std::string object_label; - if (extra_info.empty()) { - switch (identifier) { - case GL_TEXTURE: - object_label = fmt::format("Texture@0x{:016X}", addr); - break; - case GL_PROGRAM: - object_label = fmt::format("Shader@0x{:016X}", addr); - break; - default: - object_label = fmt::format("Object(0x{:X})@0x{:016X}", identifier, addr); - break; - } - } else { - object_label = fmt::format("{}@0x{:016X}", extra_info, addr); - } - glObjectLabel(identifier, handle, -1, static_cast<const GLchar*>(object_label.c_str())); -} - -} // namespace OpenGL diff --git a/src/video_core/renderer_opengl/utils.h b/src/video_core/renderer_opengl/utils.h deleted file mode 100644 index 9c09ee12c..000000000 --- a/src/video_core/renderer_opengl/utils.h +++ /dev/null @@ -1,16 +0,0 @@ -// Copyright 2014 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <string_view> -#include <vector> -#include <glad/glad.h> -#include "common/common_types.h" - -namespace OpenGL { - -void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string_view extra_info = {}); - -} // namespace OpenGL diff --git a/src/video_core/renderer_vulkan/blit_image.cpp b/src/video_core/renderer_vulkan/blit_image.cpp new file mode 100644 index 000000000..1f6a169ae --- /dev/null +++ b/src/video_core/renderer_vulkan/blit_image.cpp @@ -0,0 +1,624 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> + +#include "video_core/host_shaders/convert_depth_to_float_frag_spv.h" +#include "video_core/host_shaders/convert_float_to_depth_frag_spv.h" +#include "video_core/host_shaders/full_screen_triangle_vert_spv.h" +#include "video_core/host_shaders/vulkan_blit_color_float_frag_spv.h" +#include "video_core/host_shaders/vulkan_blit_depth_stencil_frag_spv.h" +#include "video_core/renderer_vulkan/blit_image.h" +#include "video_core/renderer_vulkan/maxwell_to_vk.h" +#include "video_core/renderer_vulkan/vk_scheduler.h" +#include "video_core/renderer_vulkan/vk_shader_util.h" +#include "video_core/renderer_vulkan/vk_state_tracker.h" +#include "video_core/renderer_vulkan/vk_texture_cache.h" +#include "video_core/renderer_vulkan/vk_update_descriptor.h" +#include "video_core/surface.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" + +namespace Vulkan { + +using VideoCommon::ImageViewType; + +namespace { +struct PushConstants { + std::array<float, 2> tex_scale; + std::array<float, 2> tex_offset; +}; + +template <u32 binding> +inline constexpr VkDescriptorSetLayoutBinding TEXTURE_DESCRIPTOR_SET_LAYOUT_BINDING{ + .binding = binding, + .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, + .descriptorCount = 1, + .stageFlags = VK_SHADER_STAGE_FRAGMENT_BIT, + .pImmutableSamplers = nullptr, +}; +constexpr std::array TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_BINDINGS{ + TEXTURE_DESCRIPTOR_SET_LAYOUT_BINDING<0>, + TEXTURE_DESCRIPTOR_SET_LAYOUT_BINDING<1>, +}; +constexpr VkDescriptorSetLayoutCreateInfo ONE_TEXTURE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .bindingCount = 1, + .pBindings = &TEXTURE_DESCRIPTOR_SET_LAYOUT_BINDING<0>, +}; +constexpr VkDescriptorSetLayoutCreateInfo TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .bindingCount = static_cast<u32>(TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_BINDINGS.size()), + .pBindings = TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_BINDINGS.data(), +}; +constexpr VkPushConstantRange PUSH_CONSTANT_RANGE{ + .stageFlags = VK_SHADER_STAGE_VERTEX_BIT, + .offset = 0, + .size = sizeof(PushConstants), +}; +constexpr VkPipelineVertexInputStateCreateInfo PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .vertexBindingDescriptionCount = 0, + .pVertexBindingDescriptions = nullptr, + .vertexAttributeDescriptionCount = 0, + .pVertexAttributeDescriptions = nullptr, +}; +constexpr VkPipelineInputAssemblyStateCreateInfo PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .topology = VK_PRIMITIVE_TOPOLOGY_TRIANGLE_LIST, + .primitiveRestartEnable = VK_FALSE, +}; +constexpr VkPipelineViewportStateCreateInfo PIPELINE_VIEWPORT_STATE_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .viewportCount = 1, + .pViewports = nullptr, + .scissorCount = 1, + .pScissors = nullptr, +}; +constexpr VkPipelineRasterizationStateCreateInfo PIPELINE_RASTERIZATION_STATE_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .depthClampEnable = VK_FALSE, + .rasterizerDiscardEnable = VK_FALSE, + .polygonMode = VK_POLYGON_MODE_FILL, + .cullMode = VK_CULL_MODE_BACK_BIT, + .frontFace = VK_FRONT_FACE_CLOCKWISE, + .depthBiasEnable = VK_FALSE, + .depthBiasConstantFactor = 0.0f, + .depthBiasClamp = 0.0f, + .depthBiasSlopeFactor = 0.0f, + .lineWidth = 1.0f, +}; +constexpr VkPipelineMultisampleStateCreateInfo PIPELINE_MULTISAMPLE_STATE_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .rasterizationSamples = VK_SAMPLE_COUNT_1_BIT, + .sampleShadingEnable = VK_FALSE, + .minSampleShading = 0.0f, + .pSampleMask = nullptr, + .alphaToCoverageEnable = VK_FALSE, + .alphaToOneEnable = VK_FALSE, +}; +constexpr std::array DYNAMIC_STATES{ + VK_DYNAMIC_STATE_VIEWPORT, + VK_DYNAMIC_STATE_SCISSOR, +}; +constexpr VkPipelineDynamicStateCreateInfo PIPELINE_DYNAMIC_STATE_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .dynamicStateCount = static_cast<u32>(DYNAMIC_STATES.size()), + .pDynamicStates = DYNAMIC_STATES.data(), +}; +constexpr VkPipelineColorBlendStateCreateInfo PIPELINE_COLOR_BLEND_STATE_EMPTY_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .logicOpEnable = VK_FALSE, + .logicOp = VK_LOGIC_OP_CLEAR, + .attachmentCount = 0, + .pAttachments = nullptr, + .blendConstants = {0.0f, 0.0f, 0.0f, 0.0f}, +}; +constexpr VkPipelineColorBlendAttachmentState PIPELINE_COLOR_BLEND_ATTACHMENT_STATE{ + .blendEnable = VK_FALSE, + .srcColorBlendFactor = VK_BLEND_FACTOR_ZERO, + .dstColorBlendFactor = VK_BLEND_FACTOR_ZERO, + .colorBlendOp = VK_BLEND_OP_ADD, + .srcAlphaBlendFactor = VK_BLEND_FACTOR_ZERO, + .dstAlphaBlendFactor = VK_BLEND_FACTOR_ZERO, + .alphaBlendOp = VK_BLEND_OP_ADD, + .colorWriteMask = VK_COLOR_COMPONENT_R_BIT | VK_COLOR_COMPONENT_G_BIT | + VK_COLOR_COMPONENT_B_BIT | VK_COLOR_COMPONENT_A_BIT, +}; +constexpr VkPipelineColorBlendStateCreateInfo PIPELINE_COLOR_BLEND_STATE_GENERIC_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .logicOpEnable = VK_FALSE, + .logicOp = VK_LOGIC_OP_CLEAR, + .attachmentCount = 1, + .pAttachments = &PIPELINE_COLOR_BLEND_ATTACHMENT_STATE, + .blendConstants = {0.0f, 0.0f, 0.0f, 0.0f}, +}; +constexpr VkPipelineDepthStencilStateCreateInfo PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .depthTestEnable = VK_TRUE, + .depthWriteEnable = VK_TRUE, + .depthCompareOp = VK_COMPARE_OP_ALWAYS, + .depthBoundsTestEnable = VK_FALSE, + .stencilTestEnable = VK_FALSE, + .front = VkStencilOpState{}, + .back = VkStencilOpState{}, + .minDepthBounds = 0.0f, + .maxDepthBounds = 0.0f, +}; + +template <VkFilter filter> +inline constexpr VkSamplerCreateInfo SAMPLER_CREATE_INFO{ + .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .magFilter = filter, + .minFilter = filter, + .mipmapMode = VK_SAMPLER_MIPMAP_MODE_NEAREST, + .addressModeU = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER, + .addressModeV = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER, + .addressModeW = VK_SAMPLER_ADDRESS_MODE_CLAMP_TO_BORDER, + .mipLodBias = 0.0f, + .anisotropyEnable = VK_FALSE, + .maxAnisotropy = 0.0f, + .compareEnable = VK_FALSE, + .compareOp = VK_COMPARE_OP_NEVER, + .minLod = 0.0f, + .maxLod = 0.0f, + .borderColor = VK_BORDER_COLOR_FLOAT_OPAQUE_WHITE, + .unnormalizedCoordinates = VK_TRUE, +}; + +constexpr VkPipelineLayoutCreateInfo PipelineLayoutCreateInfo( + const VkDescriptorSetLayout* set_layout) { + return VkPipelineLayoutCreateInfo{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .setLayoutCount = 1, + .pSetLayouts = set_layout, + .pushConstantRangeCount = 1, + .pPushConstantRanges = &PUSH_CONSTANT_RANGE, + }; +} + +constexpr VkPipelineShaderStageCreateInfo PipelineShaderStageCreateInfo(VkShaderStageFlagBits stage, + VkShaderModule shader) { + return VkPipelineShaderStageCreateInfo{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_SHADER_STAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stage = stage, + .module = shader, + .pName = "main", + .pSpecializationInfo = nullptr, + }; +} + +constexpr std::array<VkPipelineShaderStageCreateInfo, 2> MakeStages( + VkShaderModule vertex_shader, VkShaderModule fragment_shader) { + return std::array{ + PipelineShaderStageCreateInfo(VK_SHADER_STAGE_VERTEX_BIT, vertex_shader), + PipelineShaderStageCreateInfo(VK_SHADER_STAGE_FRAGMENT_BIT, fragment_shader), + }; +} + +void UpdateOneTextureDescriptorSet(const Device& device, VkDescriptorSet descriptor_set, + VkSampler sampler, VkImageView image_view) { + const VkDescriptorImageInfo image_info{ + .sampler = sampler, + .imageView = image_view, + .imageLayout = VK_IMAGE_LAYOUT_GENERAL, + }; + const VkWriteDescriptorSet write_descriptor_set{ + .sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET, + .pNext = nullptr, + .dstSet = descriptor_set, + .dstBinding = 0, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, + .pImageInfo = &image_info, + .pBufferInfo = nullptr, + .pTexelBufferView = nullptr, + }; + device.GetLogical().UpdateDescriptorSets(write_descriptor_set, nullptr); +} + +void UpdateTwoTexturesDescriptorSet(const Device& device, VkDescriptorSet descriptor_set, + VkSampler sampler, VkImageView image_view_0, + VkImageView image_view_1) { + const VkDescriptorImageInfo image_info_0{ + .sampler = sampler, + .imageView = image_view_0, + .imageLayout = VK_IMAGE_LAYOUT_GENERAL, + }; + const VkDescriptorImageInfo image_info_1{ + .sampler = sampler, + .imageView = image_view_1, + .imageLayout = VK_IMAGE_LAYOUT_GENERAL, + }; + const std::array write_descriptor_sets{ + VkWriteDescriptorSet{ + .sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET, + .pNext = nullptr, + .dstSet = descriptor_set, + .dstBinding = 0, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, + .pImageInfo = &image_info_0, + .pBufferInfo = nullptr, + .pTexelBufferView = nullptr, + }, + VkWriteDescriptorSet{ + .sType = VK_STRUCTURE_TYPE_WRITE_DESCRIPTOR_SET, + .pNext = nullptr, + .dstSet = descriptor_set, + .dstBinding = 1, + .dstArrayElement = 0, + .descriptorCount = 1, + .descriptorType = VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER, + .pImageInfo = &image_info_1, + .pBufferInfo = nullptr, + .pTexelBufferView = nullptr, + }, + }; + device.GetLogical().UpdateDescriptorSets(write_descriptor_sets, nullptr); +} + +void BindBlitState(vk::CommandBuffer cmdbuf, VkPipelineLayout layout, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region) { + const VkOffset2D offset{ + .x = std::min(dst_region[0].x, dst_region[1].x), + .y = std::min(dst_region[0].y, dst_region[1].y), + }; + const VkExtent2D extent{ + .width = static_cast<u32>(std::abs(dst_region[1].x - dst_region[0].x)), + .height = static_cast<u32>(std::abs(dst_region[1].y - dst_region[0].y)), + }; + const VkViewport viewport{ + .x = static_cast<float>(offset.x), + .y = static_cast<float>(offset.y), + .width = static_cast<float>(extent.width), + .height = static_cast<float>(extent.height), + .minDepth = 0.0f, + .maxDepth = 1.0f, + }; + // TODO: Support scissored blits + const VkRect2D scissor{ + .offset = offset, + .extent = extent, + }; + const float scale_x = static_cast<float>(src_region[1].x - src_region[0].x); + const float scale_y = static_cast<float>(src_region[1].y - src_region[0].y); + const PushConstants push_constants{ + .tex_scale = {scale_x, scale_y}, + .tex_offset = {static_cast<float>(src_region[0].x), static_cast<float>(src_region[0].y)}, + }; + cmdbuf.SetViewport(0, viewport); + cmdbuf.SetScissor(0, scissor); + cmdbuf.PushConstants(layout, VK_SHADER_STAGE_VERTEX_BIT, push_constants); +} + +} // Anonymous namespace + +BlitImageHelper::BlitImageHelper(const Device& device_, VKScheduler& scheduler_, + StateTracker& state_tracker_, VKDescriptorPool& descriptor_pool) + : device{device_}, scheduler{scheduler_}, state_tracker{state_tracker_}, + one_texture_set_layout(device.GetLogical().CreateDescriptorSetLayout( + ONE_TEXTURE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO)), + two_textures_set_layout(device.GetLogical().CreateDescriptorSetLayout( + TWO_TEXTURES_DESCRIPTOR_SET_LAYOUT_CREATE_INFO)), + one_texture_descriptor_allocator(descriptor_pool, *one_texture_set_layout), + two_textures_descriptor_allocator(descriptor_pool, *two_textures_set_layout), + one_texture_pipeline_layout(device.GetLogical().CreatePipelineLayout( + PipelineLayoutCreateInfo(one_texture_set_layout.address()))), + two_textures_pipeline_layout(device.GetLogical().CreatePipelineLayout( + PipelineLayoutCreateInfo(two_textures_set_layout.address()))), + full_screen_vert(BuildShader(device, FULL_SCREEN_TRIANGLE_VERT_SPV)), + blit_color_to_color_frag(BuildShader(device, VULKAN_BLIT_COLOR_FLOAT_FRAG_SPV)), + convert_depth_to_float_frag(BuildShader(device, CONVERT_DEPTH_TO_FLOAT_FRAG_SPV)), + convert_float_to_depth_frag(BuildShader(device, CONVERT_FLOAT_TO_DEPTH_FRAG_SPV)), + linear_sampler(device.GetLogical().CreateSampler(SAMPLER_CREATE_INFO<VK_FILTER_LINEAR>)), + nearest_sampler(device.GetLogical().CreateSampler(SAMPLER_CREATE_INFO<VK_FILTER_NEAREST>)) { + if (device.IsExtShaderStencilExportSupported()) { + blit_depth_stencil_frag = BuildShader(device, VULKAN_BLIT_DEPTH_STENCIL_FRAG_SPV); + } +} + +BlitImageHelper::~BlitImageHelper() = default; + +void BlitImageHelper::BlitColor(const Framebuffer* dst_framebuffer, const ImageView& src_image_view, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation) { + const bool is_linear = filter == Tegra::Engines::Fermi2D::Filter::Bilinear; + const BlitImagePipelineKey key{ + .renderpass = dst_framebuffer->RenderPass(), + .operation = operation, + }; + const VkPipelineLayout layout = *one_texture_pipeline_layout; + const VkImageView src_view = src_image_view.Handle(ImageViewType::e2D); + const VkSampler sampler = is_linear ? *linear_sampler : *nearest_sampler; + const VkPipeline pipeline = FindOrEmplacePipeline(key); + const VkDescriptorSet descriptor_set = one_texture_descriptor_allocator.Commit(); + scheduler.RequestRenderpass(dst_framebuffer); + scheduler.Record([dst_region, src_region, pipeline, layout, sampler, src_view, descriptor_set, + &device = device](vk::CommandBuffer cmdbuf) { + // TODO: Barriers + UpdateOneTextureDescriptorSet(device, descriptor_set, sampler, src_view); + cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline); + cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, layout, 0, descriptor_set, + nullptr); + BindBlitState(cmdbuf, layout, dst_region, src_region); + cmdbuf.Draw(3, 1, 0, 0); + }); + scheduler.InvalidateState(); +} + +void BlitImageHelper::BlitDepthStencil(const Framebuffer* dst_framebuffer, + VkImageView src_depth_view, VkImageView src_stencil_view, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation) { + ASSERT(filter == Tegra::Engines::Fermi2D::Filter::Point); + ASSERT(operation == Tegra::Engines::Fermi2D::Operation::SrcCopy); + + const VkPipelineLayout layout = *two_textures_pipeline_layout; + const VkSampler sampler = *nearest_sampler; + const VkPipeline pipeline = BlitDepthStencilPipeline(dst_framebuffer->RenderPass()); + const VkDescriptorSet descriptor_set = two_textures_descriptor_allocator.Commit(); + scheduler.RequestRenderpass(dst_framebuffer); + scheduler.Record([dst_region, src_region, pipeline, layout, sampler, src_depth_view, + src_stencil_view, descriptor_set, + &device = device](vk::CommandBuffer cmdbuf) { + // TODO: Barriers + UpdateTwoTexturesDescriptorSet(device, descriptor_set, sampler, src_depth_view, + src_stencil_view); + cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline); + cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, layout, 0, descriptor_set, + nullptr); + BindBlitState(cmdbuf, layout, dst_region, src_region); + cmdbuf.Draw(3, 1, 0, 0); + }); + scheduler.InvalidateState(); +} + +void BlitImageHelper::ConvertD32ToR32(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view) { + ConvertDepthToColorPipeline(convert_d32_to_r32_pipeline, dst_framebuffer->RenderPass()); + Convert(*convert_d32_to_r32_pipeline, dst_framebuffer, src_image_view); +} + +void BlitImageHelper::ConvertR32ToD32(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view) { + + ConvertColorToDepthPipeline(convert_r32_to_d32_pipeline, dst_framebuffer->RenderPass()); + Convert(*convert_r32_to_d32_pipeline, dst_framebuffer, src_image_view); +} + +void BlitImageHelper::ConvertD16ToR16(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view) { + ConvertDepthToColorPipeline(convert_d16_to_r16_pipeline, dst_framebuffer->RenderPass()); + Convert(*convert_d16_to_r16_pipeline, dst_framebuffer, src_image_view); +} + +void BlitImageHelper::ConvertR16ToD16(const Framebuffer* dst_framebuffer, + const ImageView& src_image_view) { + ConvertColorToDepthPipeline(convert_r16_to_d16_pipeline, dst_framebuffer->RenderPass()); + Convert(*convert_r16_to_d16_pipeline, dst_framebuffer, src_image_view); +} + +void BlitImageHelper::Convert(VkPipeline pipeline, const Framebuffer* dst_framebuffer, + const ImageView& src_image_view) { + const VkPipelineLayout layout = *one_texture_pipeline_layout; + const VkImageView src_view = src_image_view.Handle(ImageViewType::e2D); + const VkSampler sampler = *nearest_sampler; + const VkDescriptorSet descriptor_set = one_texture_descriptor_allocator.Commit(); + const VkExtent2D extent{ + .width = src_image_view.size.width, + .height = src_image_view.size.height, + }; + scheduler.RequestRenderpass(dst_framebuffer); + scheduler.Record([pipeline, layout, sampler, src_view, descriptor_set, extent, + &device = device](vk::CommandBuffer cmdbuf) { + const VkOffset2D offset{ + .x = 0, + .y = 0, + }; + const VkViewport viewport{ + .x = 0.0f, + .y = 0.0f, + .width = static_cast<float>(extent.width), + .height = static_cast<float>(extent.height), + .minDepth = 0.0f, + .maxDepth = 0.0f, + }; + const VkRect2D scissor{ + .offset = offset, + .extent = extent, + }; + const PushConstants push_constants{ + .tex_scale = {viewport.width, viewport.height}, + .tex_offset = {0.0f, 0.0f}, + }; + UpdateOneTextureDescriptorSet(device, descriptor_set, sampler, src_view); + + // TODO: Barriers + cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline); + cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_GRAPHICS, layout, 0, descriptor_set, + nullptr); + cmdbuf.SetViewport(0, viewport); + cmdbuf.SetScissor(0, scissor); + cmdbuf.PushConstants(layout, VK_SHADER_STAGE_VERTEX_BIT, push_constants); + cmdbuf.Draw(3, 1, 0, 0); + }); + scheduler.InvalidateState(); +} + +VkPipeline BlitImageHelper::FindOrEmplacePipeline(const BlitImagePipelineKey& key) { + const auto it = std::ranges::find(blit_color_keys, key); + if (it != blit_color_keys.end()) { + return *blit_color_pipelines[std::distance(blit_color_keys.begin(), it)]; + } + blit_color_keys.push_back(key); + + const std::array stages = MakeStages(*full_screen_vert, *blit_color_to_color_frag); + const VkPipelineColorBlendAttachmentState blend_attachment{ + .blendEnable = VK_FALSE, + .srcColorBlendFactor = VK_BLEND_FACTOR_ZERO, + .dstColorBlendFactor = VK_BLEND_FACTOR_ZERO, + .colorBlendOp = VK_BLEND_OP_ADD, + .srcAlphaBlendFactor = VK_BLEND_FACTOR_ZERO, + .dstAlphaBlendFactor = VK_BLEND_FACTOR_ZERO, + .alphaBlendOp = VK_BLEND_OP_ADD, + .colorWriteMask = VK_COLOR_COMPONENT_R_BIT | VK_COLOR_COMPONENT_G_BIT | + VK_COLOR_COMPONENT_B_BIT | VK_COLOR_COMPONENT_A_BIT, + }; + // TODO: programmable blending + const VkPipelineColorBlendStateCreateInfo color_blend_create_info{ + .sType = VK_STRUCTURE_TYPE_PIPELINE_COLOR_BLEND_STATE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .logicOpEnable = VK_FALSE, + .logicOp = VK_LOGIC_OP_CLEAR, + .attachmentCount = 1, + .pAttachments = &blend_attachment, + .blendConstants = {0.0f, 0.0f, 0.0f, 0.0f}, + }; + blit_color_pipelines.push_back(device.GetLogical().CreateGraphicsPipeline({ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(stages.size()), + .pStages = stages.data(), + .pVertexInputState = &PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pInputAssemblyState = &PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pTessellationState = nullptr, + .pViewportState = &PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pRasterizationState = &PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pMultisampleState = &PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pDepthStencilState = nullptr, + .pColorBlendState = &color_blend_create_info, + .pDynamicState = &PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .layout = *one_texture_pipeline_layout, + .renderPass = key.renderpass, + .subpass = 0, + .basePipelineHandle = VK_NULL_HANDLE, + .basePipelineIndex = 0, + })); + return *blit_color_pipelines.back(); +} + +VkPipeline BlitImageHelper::BlitDepthStencilPipeline(VkRenderPass renderpass) { + if (blit_depth_stencil_pipeline) { + return *blit_depth_stencil_pipeline; + } + const std::array stages = MakeStages(*full_screen_vert, *blit_depth_stencil_frag); + blit_depth_stencil_pipeline = device.GetLogical().CreateGraphicsPipeline({ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(stages.size()), + .pStages = stages.data(), + .pVertexInputState = &PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pInputAssemblyState = &PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pTessellationState = nullptr, + .pViewportState = &PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pRasterizationState = &PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pMultisampleState = &PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pDepthStencilState = &PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO, + .pColorBlendState = &PIPELINE_COLOR_BLEND_STATE_EMPTY_CREATE_INFO, + .pDynamicState = &PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .layout = *two_textures_pipeline_layout, + .renderPass = renderpass, + .subpass = 0, + .basePipelineHandle = VK_NULL_HANDLE, + .basePipelineIndex = 0, + }); + return *blit_depth_stencil_pipeline; +} + +void BlitImageHelper::ConvertDepthToColorPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass) { + if (pipeline) { + return; + } + const std::array stages = MakeStages(*full_screen_vert, *convert_depth_to_float_frag); + pipeline = device.GetLogical().CreateGraphicsPipeline({ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(stages.size()), + .pStages = stages.data(), + .pVertexInputState = &PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pInputAssemblyState = &PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pTessellationState = nullptr, + .pViewportState = &PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pRasterizationState = &PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pMultisampleState = &PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pDepthStencilState = nullptr, + .pColorBlendState = &PIPELINE_COLOR_BLEND_STATE_GENERIC_CREATE_INFO, + .pDynamicState = &PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .layout = *one_texture_pipeline_layout, + .renderPass = renderpass, + .subpass = 0, + .basePipelineHandle = VK_NULL_HANDLE, + .basePipelineIndex = 0, + }); +} + +void BlitImageHelper::ConvertColorToDepthPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass) { + if (pipeline) { + return; + } + const std::array stages = MakeStages(*full_screen_vert, *convert_float_to_depth_frag); + pipeline = device.GetLogical().CreateGraphicsPipeline({ + .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .stageCount = static_cast<u32>(stages.size()), + .pStages = stages.data(), + .pVertexInputState = &PIPELINE_VERTEX_INPUT_STATE_CREATE_INFO, + .pInputAssemblyState = &PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, + .pTessellationState = nullptr, + .pViewportState = &PIPELINE_VIEWPORT_STATE_CREATE_INFO, + .pRasterizationState = &PIPELINE_RASTERIZATION_STATE_CREATE_INFO, + .pMultisampleState = &PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, + .pDepthStencilState = &PIPELINE_DEPTH_STENCIL_STATE_CREATE_INFO, + .pColorBlendState = &PIPELINE_COLOR_BLEND_STATE_EMPTY_CREATE_INFO, + .pDynamicState = &PIPELINE_DYNAMIC_STATE_CREATE_INFO, + .layout = *one_texture_pipeline_layout, + .renderPass = renderpass, + .subpass = 0, + .basePipelineHandle = VK_NULL_HANDLE, + .basePipelineIndex = 0, + }); +} + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/blit_image.h b/src/video_core/renderer_vulkan/blit_image.h new file mode 100644 index 000000000..43fd3d737 --- /dev/null +++ b/src/video_core/renderer_vulkan/blit_image.h @@ -0,0 +1,96 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <compare> + +#include "video_core/engines/fermi_2d.h" +#include "video_core/renderer_vulkan/vk_descriptor_pool.h" +#include "video_core/texture_cache/types.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" + +namespace Vulkan { + +using VideoCommon::Offset2D; + +class Device; +class Framebuffer; +class ImageView; +class StateTracker; +class VKScheduler; + +struct BlitImagePipelineKey { + constexpr auto operator<=>(const BlitImagePipelineKey&) const noexcept = default; + + VkRenderPass renderpass; + Tegra::Engines::Fermi2D::Operation operation; +}; + +class BlitImageHelper { +public: + explicit BlitImageHelper(const Device& device, VKScheduler& scheduler, + StateTracker& state_tracker, VKDescriptorPool& descriptor_pool); + ~BlitImageHelper(); + + void BlitColor(const Framebuffer* dst_framebuffer, const ImageView& src_image_view, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation); + + void BlitDepthStencil(const Framebuffer* dst_framebuffer, VkImageView src_depth_view, + VkImageView src_stencil_view, const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation); + + void ConvertD32ToR32(const Framebuffer* dst_framebuffer, const ImageView& src_image_view); + + void ConvertR32ToD32(const Framebuffer* dst_framebuffer, const ImageView& src_image_view); + + void ConvertD16ToR16(const Framebuffer* dst_framebuffer, const ImageView& src_image_view); + + void ConvertR16ToD16(const Framebuffer* dst_framebuffer, const ImageView& src_image_view); + +private: + void Convert(VkPipeline pipeline, const Framebuffer* dst_framebuffer, + const ImageView& src_image_view); + + [[nodiscard]] VkPipeline FindOrEmplacePipeline(const BlitImagePipelineKey& key); + + [[nodiscard]] VkPipeline BlitDepthStencilPipeline(VkRenderPass renderpass); + + void ConvertDepthToColorPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass); + + void ConvertColorToDepthPipeline(vk::Pipeline& pipeline, VkRenderPass renderpass); + + const Device& device; + VKScheduler& scheduler; + StateTracker& state_tracker; + + vk::DescriptorSetLayout one_texture_set_layout; + vk::DescriptorSetLayout two_textures_set_layout; + DescriptorAllocator one_texture_descriptor_allocator; + DescriptorAllocator two_textures_descriptor_allocator; + vk::PipelineLayout one_texture_pipeline_layout; + vk::PipelineLayout two_textures_pipeline_layout; + vk::ShaderModule full_screen_vert; + vk::ShaderModule blit_color_to_color_frag; + vk::ShaderModule blit_depth_stencil_frag; + vk::ShaderModule convert_depth_to_float_frag; + vk::ShaderModule convert_float_to_depth_frag; + vk::Sampler linear_sampler; + vk::Sampler nearest_sampler; + + std::vector<BlitImagePipelineKey> blit_color_keys; + std::vector<vk::Pipeline> blit_color_pipelines; + vk::Pipeline blit_depth_stencil_pipeline; + vk::Pipeline convert_d32_to_r32_pipeline; + vk::Pipeline convert_r32_to_d32_pipeline; + vk::Pipeline convert_d16_to_r16_pipeline; + vk::Pipeline convert_r16_to_d16_pipeline; +}; + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp index 81a39a3b8..5be6dabd9 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp @@ -8,6 +8,7 @@ #include <boost/functional/hash.hpp> +#include "common/bit_cast.h" #include "common/cityhash.h" #include "common/common_types.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" @@ -45,7 +46,7 @@ void FixedPipelineState::Fill(const Maxwell& regs, bool has_extended_dynamic_sta regs.polygon_offset_fill_enable}; const u32 topology_index = static_cast<u32>(regs.draw.topology.Value()); - raw = 0; + raw1 = 0; primitive_restart_enable.Assign(regs.primitive_restart.enabled != 0 ? 1 : 0); depth_bias_enable.Assign(enabled_lut[POLYGON_OFFSET_ENABLE_LUT[topology_index]] != 0 ? 1 : 0); depth_clamp_disabled.Assign(regs.view_volume_clip_control.depth_clamp_disabled.Value()); @@ -58,15 +59,24 @@ void FixedPipelineState::Fill(const Maxwell& regs, bool has_extended_dynamic_sta logic_op_enable.Assign(regs.logic_op.enable != 0 ? 1 : 0); logic_op.Assign(PackLogicOp(regs.logic_op.operation)); rasterize_enable.Assign(regs.rasterize_enable != 0 ? 1 : 0); + topology.Assign(regs.draw.topology); + msaa_mode.Assign(regs.multisample_mode); + + raw2 = 0; + const auto test_func = + regs.alpha_test_enabled == 1 ? regs.alpha_test_func : Maxwell::ComparisonOp::Always; + alpha_test_func.Assign(PackComparisonOp(test_func)); + early_z.Assign(regs.force_early_fragment_tests != 0 ? 1 : 0); - std::memcpy(&point_size, ®s.point_size, sizeof(point_size)); // TODO: C++20 std::bit_cast + alpha_test_ref = Common::BitCast<u32>(regs.alpha_test_ref); + point_size = Common::BitCast<u32>(regs.point_size); for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { binding_divisors[index] = regs.instanced_arrays.IsInstancingEnabled(index) ? regs.vertex_array[index].divisor : 0; } - for (std::size_t index = 0; index < Maxwell::NumVertexAttributes; ++index) { + for (size_t index = 0; index < Maxwell::NumVertexAttributes; ++index) { const auto& input = regs.vertex_attrib_format[index]; auto& attribute = attributes[index]; attribute.raw = 0; @@ -75,6 +85,7 @@ void FixedPipelineState::Fill(const Maxwell& regs, bool has_extended_dynamic_sta attribute.offset.Assign(input.offset); attribute.type.Assign(static_cast<u32>(input.type.Value())); attribute.size.Assign(static_cast<u32>(input.size.Value())); + attribute.binding_index_enabled.Assign(regs.vertex_array[index].IsEnabled() ? 1 : 0); } for (std::size_t index = 0; index < std::size(attachments); ++index) { @@ -131,7 +142,6 @@ void FixedPipelineState::BlendingAttachment::Fill(const Maxwell& regs, std::size } void FixedPipelineState::DynamicState::Fill(const Maxwell& regs) { - const u32 topology_index = static_cast<u32>(regs.draw.topology.Value()); u32 packed_front_face = PackFrontFace(regs.front_face); if (regs.screen_y_control.triangle_rast_flip != 0) { // Flip front face @@ -161,17 +171,11 @@ void FixedPipelineState::DynamicState::Fill(const Maxwell& regs) { depth_test_enable.Assign(regs.depth_test_enable); front_face.Assign(packed_front_face); depth_test_func.Assign(PackComparisonOp(regs.depth_test_func)); - topology.Assign(topology_index); cull_face.Assign(PackCullFace(regs.cull_face)); cull_enable.Assign(regs.cull_test_enabled != 0 ? 1 : 0); - - for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { - const auto& input = regs.vertex_array[index]; - VertexBinding& binding = vertex_bindings[index]; - binding.raw = 0; - binding.enabled.Assign(input.IsEnabled() ? 1 : 0); - binding.stride.Assign(static_cast<u16>(input.stride.Value())); - } + std::ranges::transform(regs.vertex_array, vertex_strides.begin(), [](const auto& array) { + return static_cast<u16>(array.stride.Value()); + }); } std::size_t FixedPipelineState::Hash() const noexcept { diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.h b/src/video_core/renderer_vulkan/fixed_pipeline_state.h index cdcbb65f5..465a55fdb 100644 --- a/src/video_core/renderer_vulkan/fixed_pipeline_state.h +++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.h @@ -96,6 +96,8 @@ struct FixedPipelineState { BitField<6, 14, u32> offset; BitField<20, 3, u32> type; BitField<23, 6, u32> size; + // Not really an element of a vertex attribute, but it can be packed here + BitField<29, 1, u32> binding_index_enabled; constexpr Maxwell::VertexAttribute::Type Type() const noexcept { return static_cast<Maxwell::VertexAttribute::Type>(type.Value()); @@ -130,12 +132,6 @@ struct FixedPipelineState { } }; - union VertexBinding { - u16 raw; - BitField<0, 12, u16> stride; - BitField<12, 1, u16> enabled; - }; - struct DynamicState { union { u32 raw1; @@ -150,11 +146,11 @@ struct FixedPipelineState { }; union { u32 raw2; - BitField<0, 4, u32> topology; - BitField<4, 2, u32> cull_face; - BitField<6, 1, u32> cull_enable; + BitField<0, 2, u32> cull_face; + BitField<2, 1, u32> cull_enable; }; - std::array<VertexBinding, Maxwell::NumVertexArrays> vertex_bindings; + // Vertex stride is a 12 bits value, we have 4 bits to spare per element + std::array<u16, Maxwell::NumVertexArrays> vertex_strides; void Fill(const Maxwell& regs); @@ -169,14 +165,10 @@ struct FixedPipelineState { Maxwell::FrontFace FrontFace() const noexcept { return UnpackFrontFace(front_face.Value()); } - - constexpr Maxwell::PrimitiveTopology Topology() const noexcept { - return static_cast<Maxwell::PrimitiveTopology>(topology.Value()); - } }; union { - u32 raw; + u32 raw1; BitField<0, 1, u32> no_extended_dynamic_state; BitField<2, 1, u32> primitive_restart_enable; BitField<3, 1, u32> depth_bias_enable; @@ -190,7 +182,16 @@ struct FixedPipelineState { BitField<18, 1, u32> logic_op_enable; BitField<19, 4, u32> logic_op; BitField<23, 1, u32> rasterize_enable; + BitField<24, 4, Maxwell::PrimitiveTopology> topology; + BitField<28, 4, Tegra::Texture::MsaaMode> msaa_mode; + }; + union { + u32 raw2; + BitField<0, 3, u32> alpha_test_func; + BitField<3, 1, u32> early_z; }; + + u32 alpha_test_ref; u32 point_size; std::array<u32, Maxwell::NumVertexArrays> binding_divisors; std::array<VertexAttribute, Maxwell::NumVertexAttributes> attributes; diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp index d22de1d81..ca7c2c579 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.cpp +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.cpp @@ -9,9 +9,9 @@ #include "common/logging/log.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/wrapper.h" #include "video_core/surface.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan::MaxwellToVK { @@ -26,7 +26,7 @@ VkFilter Filter(Tegra::Texture::TextureFilter filter) { case Tegra::Texture::TextureFilter::Linear: return VK_FILTER_LINEAR; } - UNREACHABLE_MSG("Invalid sampler filter={}", static_cast<u32>(filter)); + UNREACHABLE_MSG("Invalid sampler filter={}", filter); return {}; } @@ -43,11 +43,11 @@ VkSamplerMipmapMode MipmapMode(Tegra::Texture::TextureMipmapFilter mipmap_filter case Tegra::Texture::TextureMipmapFilter::Linear: return VK_SAMPLER_MIPMAP_MODE_LINEAR; } - UNREACHABLE_MSG("Invalid sampler mipmap mode={}", static_cast<u32>(mipmap_filter)); + UNREACHABLE_MSG("Invalid sampler mipmap mode={}", mipmap_filter); return {}; } -VkSamplerAddressMode WrapMode(const VKDevice& device, Tegra::Texture::WrapMode wrap_mode, +VkSamplerAddressMode WrapMode(const Device& device, Tegra::Texture::WrapMode wrap_mode, Tegra::Texture::TextureFilter filter) { switch (wrap_mode) { case Tegra::Texture::WrapMode::Wrap: @@ -79,7 +79,7 @@ VkSamplerAddressMode WrapMode(const VKDevice& device, Tegra::Texture::WrapMode w UNIMPLEMENTED(); return VK_SAMPLER_ADDRESS_MODE_MIRROR_CLAMP_TO_EDGE; default: - UNIMPLEMENTED_MSG("Unimplemented wrap mode={}", static_cast<u32>(wrap_mode)); + UNIMPLEMENTED_MSG("Unimplemented wrap mode={}", wrap_mode); return {}; } } @@ -103,8 +103,7 @@ VkCompareOp DepthCompareFunction(Tegra::Texture::DepthCompareFunc depth_compare_ case Tegra::Texture::DepthCompareFunc::Always: return VK_COMPARE_OP_ALWAYS; } - UNIMPLEMENTED_MSG("Unimplemented sampler depth compare function={}", - static_cast<u32>(depth_compare_func)); + UNIMPLEMENTED_MSG("Unimplemented sampler depth compare function={}", depth_compare_func); return {}; } @@ -123,7 +122,7 @@ struct FormatTuple { {VK_FORMAT_A8B8G8R8_SINT_PACK32, Attachable | Storage}, // A8B8G8R8_SINT {VK_FORMAT_A8B8G8R8_UINT_PACK32, Attachable | Storage}, // A8B8G8R8_UINT {VK_FORMAT_R5G6B5_UNORM_PACK16, Attachable}, // R5G6B5_UNORM - {VK_FORMAT_B5G6R5_UNORM_PACK16, Attachable}, // B5G6R5_UNORM + {VK_FORMAT_B5G6R5_UNORM_PACK16}, // B5G6R5_UNORM {VK_FORMAT_A1R5G5B5_UNORM_PACK16, Attachable}, // A1R5G5B5_UNORM {VK_FORMAT_A2B10G10R10_UNORM_PACK32, Attachable | Storage}, // A2B10G10R10_UNORM {VK_FORMAT_A2B10G10R10_UINT_PACK32, Attachable | Storage}, // A2B10G10R10_UINT @@ -164,7 +163,7 @@ struct FormatTuple { {VK_FORMAT_R16G16_UNORM, Attachable | Storage}, // R16G16_UNORM {VK_FORMAT_R16G16_SFLOAT, Attachable | Storage}, // R16G16_FLOAT {VK_FORMAT_UNDEFINED}, // R16G16_UINT - {VK_FORMAT_UNDEFINED}, // R16G16_SINT + {VK_FORMAT_R16G16_SINT, Attachable | Storage}, // R16G16_SINT {VK_FORMAT_R16G16_SNORM, Attachable | Storage}, // R16G16_SNORM {VK_FORMAT_UNDEFINED}, // R32G32B32_FLOAT {VK_FORMAT_R8G8B8A8_SRGB, Attachable}, // A8B8G8R8_SRGB @@ -223,30 +222,31 @@ constexpr bool IsZetaFormat(PixelFormat pixel_format) { } // Anonymous namespace -FormatInfo SurfaceFormat(const VKDevice& device, FormatType format_type, PixelFormat pixel_format) { +FormatInfo SurfaceFormat(const Device& device, FormatType format_type, PixelFormat pixel_format) { ASSERT(static_cast<std::size_t>(pixel_format) < std::size(tex_format_tuples)); auto tuple = tex_format_tuples[static_cast<std::size_t>(pixel_format)]; if (tuple.format == VK_FORMAT_UNDEFINED) { - UNIMPLEMENTED_MSG("Unimplemented texture format with pixel format={}", - static_cast<u32>(pixel_format)); + UNIMPLEMENTED_MSG("Unimplemented texture format with pixel format={}", pixel_format); return {VK_FORMAT_A8B8G8R8_UNORM_PACK32, true, true}; } // Use A8B8G8R8_UNORM on hardware that doesn't support ASTC natively if (!device.IsOptimalAstcSupported() && VideoCore::Surface::IsPixelFormatASTC(pixel_format)) { - tuple.format = VideoCore::Surface::IsPixelFormatSRGB(pixel_format) - ? VK_FORMAT_A8B8G8R8_SRGB_PACK32 - : VK_FORMAT_A8B8G8R8_UNORM_PACK32; + const bool is_srgb = VideoCore::Surface::IsPixelFormatSRGB(pixel_format); + tuple.format = is_srgb ? VK_FORMAT_A8B8G8R8_SRGB_PACK32 : VK_FORMAT_A8B8G8R8_UNORM_PACK32; } const bool attachable = tuple.usage & Attachable; const bool storage = tuple.usage & Storage; - VkFormatFeatureFlags usage; - if (format_type == FormatType::Buffer) { + VkFormatFeatureFlags usage{}; + switch (format_type) { + case FormatType::Buffer: usage = VK_FORMAT_FEATURE_STORAGE_TEXEL_BUFFER_BIT | VK_FORMAT_FEATURE_UNIFORM_TEXEL_BUFFER_BIT; - } else { + break; + case FormatType::Linear: + case FormatType::Optimal: usage = VK_FORMAT_FEATURE_SAMPLED_IMAGE_BIT | VK_FORMAT_FEATURE_TRANSFER_DST_BIT | VK_FORMAT_FEATURE_TRANSFER_SRC_BIT; if (attachable) { @@ -256,6 +256,7 @@ FormatInfo SurfaceFormat(const VKDevice& device, FormatType format_type, PixelFo if (storage) { usage |= VK_FORMAT_FEATURE_STORAGE_IMAGE_BIT; } + break; } return {device.GetSupportedFormat(tuple.format, usage, format_type), attachable, storage}; } @@ -275,11 +276,11 @@ VkShaderStageFlagBits ShaderStage(Tegra::Engines::ShaderType stage) { case Tegra::Engines::ShaderType::Compute: return VK_SHADER_STAGE_COMPUTE_BIT; } - UNIMPLEMENTED_MSG("Unimplemented shader stage={}", static_cast<u32>(stage)); + UNIMPLEMENTED_MSG("Unimplemented shader stage={}", stage); return {}; } -VkPrimitiveTopology PrimitiveTopology([[maybe_unused]] const VKDevice& device, +VkPrimitiveTopology PrimitiveTopology([[maybe_unused]] const Device& device, Maxwell::PrimitiveTopology topology) { switch (topology) { case Maxwell::PrimitiveTopology::Points: @@ -300,7 +301,7 @@ VkPrimitiveTopology PrimitiveTopology([[maybe_unused]] const VKDevice& device, case Maxwell::PrimitiveTopology::Patches: return VK_PRIMITIVE_TOPOLOGY_PATCH_LIST; default: - UNIMPLEMENTED_MSG("Unimplemented topology={}", static_cast<u32>(topology)); + UNIMPLEMENTED_MSG("Unimplemented topology={}", topology); return {}; } } @@ -490,8 +491,7 @@ VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttrib } break; } - UNIMPLEMENTED_MSG("Unimplemented vertex format of type={} and size={}", static_cast<u32>(type), - static_cast<u32>(size)); + UNIMPLEMENTED_MSG("Unimplemented vertex format of type={} and size={}", type, size); return {}; } @@ -522,11 +522,11 @@ VkCompareOp ComparisonOp(Maxwell::ComparisonOp comparison) { case Maxwell::ComparisonOp::AlwaysOld: return VK_COMPARE_OP_ALWAYS; } - UNIMPLEMENTED_MSG("Unimplemented comparison op={}", static_cast<u32>(comparison)); + UNIMPLEMENTED_MSG("Unimplemented comparison op={}", comparison); return {}; } -VkIndexType IndexFormat(const VKDevice& device, Maxwell::IndexFormat index_format) { +VkIndexType IndexFormat(const Device& device, Maxwell::IndexFormat index_format) { switch (index_format) { case Maxwell::IndexFormat::UnsignedByte: if (!device.IsExtIndexTypeUint8Supported()) { @@ -539,7 +539,7 @@ VkIndexType IndexFormat(const VKDevice& device, Maxwell::IndexFormat index_forma case Maxwell::IndexFormat::UnsignedInt: return VK_INDEX_TYPE_UINT32; } - UNIMPLEMENTED_MSG("Unimplemented index_format={}", static_cast<u32>(index_format)); + UNIMPLEMENTED_MSG("Unimplemented index_format={}", index_format); return {}; } @@ -570,7 +570,7 @@ VkStencilOp StencilOp(Maxwell::StencilOp stencil_op) { case Maxwell::StencilOp::DecrWrapOGL: return VK_STENCIL_OP_DECREMENT_AND_WRAP; } - UNIMPLEMENTED_MSG("Unimplemented stencil op={}", static_cast<u32>(stencil_op)); + UNIMPLEMENTED_MSG("Unimplemented stencil op={}", stencil_op); return {}; } @@ -592,7 +592,7 @@ VkBlendOp BlendEquation(Maxwell::Blend::Equation equation) { case Maxwell::Blend::Equation::MaxGL: return VK_BLEND_OP_MAX; } - UNIMPLEMENTED_MSG("Unimplemented blend equation={}", static_cast<u32>(equation)); + UNIMPLEMENTED_MSG("Unimplemented blend equation={}", equation); return {}; } @@ -656,7 +656,7 @@ VkBlendFactor BlendFactor(Maxwell::Blend::Factor factor) { case Maxwell::Blend::Factor::OneMinusConstantAlphaGL: return VK_BLEND_FACTOR_ONE_MINUS_CONSTANT_ALPHA; } - UNIMPLEMENTED_MSG("Unimplemented blend factor={}", static_cast<u32>(factor)); + UNIMPLEMENTED_MSG("Unimplemented blend factor={}", factor); return {}; } @@ -667,7 +667,7 @@ VkFrontFace FrontFace(Maxwell::FrontFace front_face) { case Maxwell::FrontFace::CounterClockWise: return VK_FRONT_FACE_COUNTER_CLOCKWISE; } - UNIMPLEMENTED_MSG("Unimplemented front face={}", static_cast<u32>(front_face)); + UNIMPLEMENTED_MSG("Unimplemented front face={}", front_face); return {}; } @@ -680,7 +680,7 @@ VkCullModeFlags CullFace(Maxwell::CullFace cull_face) { case Maxwell::CullFace::FrontAndBack: return VK_CULL_MODE_FRONT_AND_BACK; } - UNIMPLEMENTED_MSG("Unimplemented cull face={}", static_cast<u32>(cull_face)); + UNIMPLEMENTED_MSG("Unimplemented cull face={}", cull_face); return {}; } @@ -700,7 +700,7 @@ VkComponentSwizzle SwizzleSource(Tegra::Texture::SwizzleSource swizzle) { case Tegra::Texture::SwizzleSource::OneFloat: return VK_COMPONENT_SWIZZLE_ONE; } - UNIMPLEMENTED_MSG("Unimplemented swizzle source={}", static_cast<u32>(swizzle)); + UNIMPLEMENTED_MSG("Unimplemented swizzle source={}", swizzle); return {}; } @@ -723,8 +723,21 @@ VkViewportCoordinateSwizzleNV ViewportSwizzle(Maxwell::ViewportSwizzle swizzle) case Maxwell::ViewportSwizzle::NegativeW: return VK_VIEWPORT_COORDINATE_SWIZZLE_NEGATIVE_W_NV; } - UNREACHABLE_MSG("Invalid swizzle={}", static_cast<int>(swizzle)); + UNREACHABLE_MSG("Invalid swizzle={}", swizzle); return {}; } +VkSamplerReductionMode SamplerReduction(Tegra::Texture::SamplerReduction reduction) { + switch (reduction) { + case Tegra::Texture::SamplerReduction::WeightedAverage: + return VK_SAMPLER_REDUCTION_MODE_WEIGHTED_AVERAGE_EXT; + case Tegra::Texture::SamplerReduction::Min: + return VK_SAMPLER_REDUCTION_MODE_MIN_EXT; + case Tegra::Texture::SamplerReduction::Max: + return VK_SAMPLER_REDUCTION_MODE_MAX_EXT; + } + UNREACHABLE_MSG("Invalid sampler mode={}", static_cast<int>(reduction)); + return VK_SAMPLER_REDUCTION_MODE_WEIGHTED_AVERAGE_EXT; +} + } // namespace Vulkan::MaxwellToVK diff --git a/src/video_core/renderer_vulkan/maxwell_to_vk.h b/src/video_core/renderer_vulkan/maxwell_to_vk.h index 7e213452f..537969840 100644 --- a/src/video_core/renderer_vulkan/maxwell_to_vk.h +++ b/src/video_core/renderer_vulkan/maxwell_to_vk.h @@ -6,10 +6,10 @@ #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/wrapper.h" #include "video_core/surface.h" #include "video_core/textures/texture.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan::MaxwellToVK { @@ -22,7 +22,7 @@ VkFilter Filter(Tegra::Texture::TextureFilter filter); VkSamplerMipmapMode MipmapMode(Tegra::Texture::TextureMipmapFilter mipmap_filter); -VkSamplerAddressMode WrapMode(const VKDevice& device, Tegra::Texture::WrapMode wrap_mode, +VkSamplerAddressMode WrapMode(const Device& device, Tegra::Texture::WrapMode wrap_mode, Tegra::Texture::TextureFilter filter); VkCompareOp DepthCompareFunction(Tegra::Texture::DepthCompareFunc depth_compare_func); @@ -35,17 +35,17 @@ struct FormatInfo { bool storage; }; -FormatInfo SurfaceFormat(const VKDevice& device, FormatType format_type, PixelFormat pixel_format); +FormatInfo SurfaceFormat(const Device& device, FormatType format_type, PixelFormat pixel_format); VkShaderStageFlagBits ShaderStage(Tegra::Engines::ShaderType stage); -VkPrimitiveTopology PrimitiveTopology(const VKDevice& device, Maxwell::PrimitiveTopology topology); +VkPrimitiveTopology PrimitiveTopology(const Device& device, Maxwell::PrimitiveTopology topology); VkFormat VertexFormat(Maxwell::VertexAttribute::Type type, Maxwell::VertexAttribute::Size size); VkCompareOp ComparisonOp(Maxwell::ComparisonOp comparison); -VkIndexType IndexFormat(const VKDevice& device, Maxwell::IndexFormat index_format); +VkIndexType IndexFormat(const Device& device, Maxwell::IndexFormat index_format); VkStencilOp StencilOp(Maxwell::StencilOp stencil_op); @@ -61,4 +61,6 @@ VkComponentSwizzle SwizzleSource(Tegra::Texture::SwizzleSource swizzle); VkViewportCoordinateSwizzleNV ViewportSwizzle(Maxwell::ViewportSwizzle swizzle); +VkSamplerReductionMode SamplerReduction(Tegra::Texture::SamplerReduction reduction); + } // namespace Vulkan::MaxwellToVK diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.cpp b/src/video_core/renderer_vulkan/renderer_vulkan.cpp index 715182b3b..d7437e185 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp +++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp @@ -12,8 +12,6 @@ #include <fmt/format.h> -#include "common/dynamic_library.h" -#include "common/file_util.h" #include "common/logging/log.h" #include "common/telemetry.h" #include "core/core.h" @@ -24,179 +22,27 @@ #include "video_core/gpu.h" #include "video_core/renderer_vulkan/renderer_vulkan.h" #include "video_core/renderer_vulkan/vk_blit_screen.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_master_semaphore.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_state_tracker.h" #include "video_core/renderer_vulkan/vk_swapchain.h" -#include "video_core/renderer_vulkan/wrapper.h" - -// Include these late to avoid polluting previous headers -#ifdef _WIN32 -#include <windows.h> -// ensure include order -#include <vulkan/vulkan_win32.h> -#endif - -#if !defined(_WIN32) && !defined(__APPLE__) -#include <X11/Xlib.h> -#include <vulkan/vulkan_wayland.h> -#include <vulkan/vulkan_xlib.h> -#endif +#include "video_core/vulkan_common/vulkan_debug_callback.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_instance.h" +#include "video_core/vulkan_common/vulkan_library.h" +#include "video_core/vulkan_common/vulkan_surface.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { - namespace { - -using Core::Frontend::WindowSystemType; - -VkBool32 DebugCallback(VkDebugUtilsMessageSeverityFlagBitsEXT severity, - VkDebugUtilsMessageTypeFlagsEXT type, - const VkDebugUtilsMessengerCallbackDataEXT* data, - [[maybe_unused]] void* user_data) { - const char* const message{data->pMessage}; - - if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT) { - LOG_CRITICAL(Render_Vulkan, "{}", message); - } else if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT) { - LOG_WARNING(Render_Vulkan, "{}", message); - } else if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT) { - LOG_INFO(Render_Vulkan, "{}", message); - } else if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT) { - LOG_DEBUG(Render_Vulkan, "{}", message); - } - return VK_FALSE; -} - -Common::DynamicLibrary OpenVulkanLibrary() { - Common::DynamicLibrary library; -#ifdef __APPLE__ - // Check if a path to a specific Vulkan library has been specified. - char* libvulkan_env = getenv("LIBVULKAN_PATH"); - if (!libvulkan_env || !library.Open(libvulkan_env)) { - // Use the libvulkan.dylib from the application bundle. - const std::string filename = - Common::FS::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib"; - library.Open(filename.c_str()); - } -#else - std::string filename = Common::DynamicLibrary::GetVersionedFilename("vulkan", 1); - if (!library.Open(filename.c_str())) { - // Android devices may not have libvulkan.so.1, only libvulkan.so. - filename = Common::DynamicLibrary::GetVersionedFilename("vulkan"); - (void)library.Open(filename.c_str()); - } -#endif - return library; -} - -vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatch& dld, - WindowSystemType window_type = WindowSystemType::Headless, - bool enable_layers = false) { - if (!library.IsOpen()) { - LOG_ERROR(Render_Vulkan, "Vulkan library not available"); - return {}; - } - if (!library.GetSymbol("vkGetInstanceProcAddr", &dld.vkGetInstanceProcAddr)) { - LOG_ERROR(Render_Vulkan, "vkGetInstanceProcAddr not present in Vulkan"); - return {}; - } - if (!vk::Load(dld)) { - LOG_ERROR(Render_Vulkan, "Failed to load Vulkan function pointers"); - return {}; - } - - std::vector<const char*> extensions; - extensions.reserve(6); - switch (window_type) { - case Core::Frontend::WindowSystemType::Headless: - break; -#ifdef _WIN32 - case Core::Frontend::WindowSystemType::Windows: - extensions.push_back(VK_KHR_WIN32_SURFACE_EXTENSION_NAME); - break; -#endif -#if !defined(_WIN32) && !defined(__APPLE__) - case Core::Frontend::WindowSystemType::X11: - extensions.push_back(VK_KHR_XLIB_SURFACE_EXTENSION_NAME); - break; - case Core::Frontend::WindowSystemType::Wayland: - extensions.push_back(VK_KHR_WAYLAND_SURFACE_EXTENSION_NAME); - break; -#endif - default: - LOG_ERROR(Render_Vulkan, "Presentation not supported on this platform"); - break; - } - if (window_type != Core::Frontend::WindowSystemType::Headless) { - extensions.push_back(VK_KHR_SURFACE_EXTENSION_NAME); - } - if (enable_layers) { - extensions.push_back(VK_EXT_DEBUG_UTILS_EXTENSION_NAME); - } - extensions.push_back(VK_KHR_GET_PHYSICAL_DEVICE_PROPERTIES_2_EXTENSION_NAME); - - const std::optional properties = vk::EnumerateInstanceExtensionProperties(dld); - if (!properties) { - LOG_ERROR(Render_Vulkan, "Failed to query extension properties"); - return {}; - } - - for (const char* extension : extensions) { - const auto it = - std::find_if(properties->begin(), properties->end(), [extension](const auto& prop) { - return !std::strcmp(extension, prop.extensionName); - }); - if (it == properties->end()) { - LOG_ERROR(Render_Vulkan, "Required instance extension {} is not available", extension); - return {}; - } - } - - std::vector<const char*> layers; - layers.reserve(1); - if (enable_layers) { - layers.push_back("VK_LAYER_KHRONOS_validation"); - } - - const std::optional layer_properties = vk::EnumerateInstanceLayerProperties(dld); - if (!layer_properties) { - LOG_ERROR(Render_Vulkan, "Failed to query layer properties, disabling layers"); - layers.clear(); - } - - for (auto layer_it = layers.begin(); layer_it != layers.end();) { - const char* const layer = *layer_it; - const auto it = std::find_if( - layer_properties->begin(), layer_properties->end(), - [layer](const VkLayerProperties& prop) { return !std::strcmp(layer, prop.layerName); }); - if (it == layer_properties->end()) { - LOG_ERROR(Render_Vulkan, "Layer {} not available, removing it", layer); - layer_it = layers.erase(layer_it); - } else { - ++layer_it; - } - } - - vk::Instance instance = vk::Instance::Create(layers, extensions, dld); - if (!instance) { - LOG_ERROR(Render_Vulkan, "Failed to create Vulkan instance"); - return {}; - } - if (!vk::Load(*instance, dld)) { - LOG_ERROR(Render_Vulkan, "Failed to load Vulkan instance function pointers"); - } - return instance; -} - std::string GetReadableVersion(u32 version) { return fmt::format("{}.{}.{}", VK_VERSION_MAJOR(version), VK_VERSION_MINOR(version), VK_VERSION_PATCH(version)); } -std::string GetDriverVersion(const VKDevice& device) { +std::string GetDriverVersion(const Device& device) { // Extracted from // https://github.com/SaschaWillems/vulkan.gpuinfo.org/blob/5dddea46ea1120b0df14eef8f15ff8e318e35462/functions.php#L308-L314 const u32 version = device.GetDriverVersion(); @@ -213,7 +59,6 @@ std::string GetDriverVersion(const VKDevice& device) { const u32 minor = version & 0x3fff; return fmt::format("{}.{}", major, minor); } - return GetReadableVersion(version); } @@ -240,8 +85,8 @@ std::string BuildCommaSeparatedExtensions(std::vector<std::string> available_ext RendererVulkan::RendererVulkan(Core::TelemetrySession& telemetry_session_, Core::Frontend::EmuWindow& emu_window, Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, - std::unique_ptr<Core::Frontend::GraphicsContext> context) - : RendererBase{emu_window, std::move(context)}, telemetry_session{telemetry_session_}, + std::unique_ptr<Core::Frontend::GraphicsContext> context_) + : RendererBase{emu_window, std::move(context_)}, telemetry_session{telemetry_session_}, cpu_memory{cpu_memory_}, gpu{gpu_} {} RendererVulkan::~RendererVulkan() { @@ -249,12 +94,9 @@ RendererVulkan::~RendererVulkan() { } void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { - render_window.PollEvents(); - if (!framebuffer) { return; } - const auto& layout = render_window.GetFramebufferLayout(); if (layout.width > 0 && layout.height > 0 && render_window.IsShown()) { const VAddr framebuffer_addr = framebuffer->address + framebuffer->offset; @@ -280,17 +122,19 @@ void RendererVulkan::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) { rasterizer->TickFrame(); } - render_window.PollEvents(); + render_window.OnFrameDisplayed(); } -bool RendererVulkan::Init() { - library = OpenVulkanLibrary(); - instance = CreateInstance(library, dld, render_window.GetWindowInfo().type, - Settings::values.renderer_debug); - if (!instance || !CreateDebugCallback() || !CreateSurface() || !PickDevices()) { - return false; +bool RendererVulkan::Init() try { + library = OpenLibrary(); + instance = CreateInstance(library, dld, VK_API_VERSION_1_1, render_window.GetWindowInfo().type, + true, Settings::values.renderer_debug); + if (Settings::values.renderer_debug) { + debug_callback = CreateDebugCallback(instance); } + surface = CreateSurface(instance, render_window); + InitializeDevice(); Report(); memory_manager = std::make_unique<VKMemoryManager>(*device); @@ -310,8 +154,11 @@ bool RendererVulkan::Init() { blit_screen = std::make_unique<VKBlitScreen>(cpu_memory, render_window, *rasterizer, *device, *memory_manager, *swapchain, *scheduler, screen_info); - return true; + +} catch (const vk::Exception& exception) { + LOG_ERROR(Render_Vulkan, "Vulkan initialization failed with error: {}", exception.what()); + return false; } void RendererVulkan::ShutDown() { @@ -321,7 +168,6 @@ void RendererVulkan::ShutDown() { if (const auto& dev = device->GetLogical()) { dev.WaitIdle(); } - rasterizer.reset(); blit_screen.reset(); scheduler.reset(); @@ -330,94 +176,15 @@ void RendererVulkan::ShutDown() { device.reset(); } -bool RendererVulkan::CreateDebugCallback() { - if (!Settings::values.renderer_debug) { - return true; - } - debug_callback = instance.TryCreateDebugCallback(DebugCallback); - if (!debug_callback) { - LOG_ERROR(Render_Vulkan, "Failed to create debug callback"); - return false; - } - return true; -} - -bool RendererVulkan::CreateSurface() { - [[maybe_unused]] const auto& window_info = render_window.GetWindowInfo(); - VkSurfaceKHR unsafe_surface = nullptr; - -#ifdef _WIN32 - if (window_info.type == Core::Frontend::WindowSystemType::Windows) { - const HWND hWnd = static_cast<HWND>(window_info.render_surface); - const VkWin32SurfaceCreateInfoKHR win32_ci{VK_STRUCTURE_TYPE_WIN32_SURFACE_CREATE_INFO_KHR, - nullptr, 0, nullptr, hWnd}; - const auto vkCreateWin32SurfaceKHR = reinterpret_cast<PFN_vkCreateWin32SurfaceKHR>( - dld.vkGetInstanceProcAddr(*instance, "vkCreateWin32SurfaceKHR")); - if (!vkCreateWin32SurfaceKHR || - vkCreateWin32SurfaceKHR(*instance, &win32_ci, nullptr, &unsafe_surface) != VK_SUCCESS) { - LOG_ERROR(Render_Vulkan, "Failed to initialize Win32 surface"); - return false; - } - } -#endif -#if !defined(_WIN32) && !defined(__APPLE__) - if (window_info.type == Core::Frontend::WindowSystemType::X11) { - const VkXlibSurfaceCreateInfoKHR xlib_ci{ - VK_STRUCTURE_TYPE_XLIB_SURFACE_CREATE_INFO_KHR, nullptr, 0, - static_cast<Display*>(window_info.display_connection), - reinterpret_cast<Window>(window_info.render_surface)}; - const auto vkCreateXlibSurfaceKHR = reinterpret_cast<PFN_vkCreateXlibSurfaceKHR>( - dld.vkGetInstanceProcAddr(*instance, "vkCreateXlibSurfaceKHR")); - if (!vkCreateXlibSurfaceKHR || - vkCreateXlibSurfaceKHR(*instance, &xlib_ci, nullptr, &unsafe_surface) != VK_SUCCESS) { - LOG_ERROR(Render_Vulkan, "Failed to initialize Xlib surface"); - return false; - } - } - if (window_info.type == Core::Frontend::WindowSystemType::Wayland) { - const VkWaylandSurfaceCreateInfoKHR wayland_ci{ - VK_STRUCTURE_TYPE_WAYLAND_SURFACE_CREATE_INFO_KHR, nullptr, 0, - static_cast<wl_display*>(window_info.display_connection), - static_cast<wl_surface*>(window_info.render_surface)}; - const auto vkCreateWaylandSurfaceKHR = reinterpret_cast<PFN_vkCreateWaylandSurfaceKHR>( - dld.vkGetInstanceProcAddr(*instance, "vkCreateWaylandSurfaceKHR")); - if (!vkCreateWaylandSurfaceKHR || - vkCreateWaylandSurfaceKHR(*instance, &wayland_ci, nullptr, &unsafe_surface) != - VK_SUCCESS) { - LOG_ERROR(Render_Vulkan, "Failed to initialize Wayland surface"); - return false; - } - } -#endif - if (!unsafe_surface) { - LOG_ERROR(Render_Vulkan, "Presentation not supported on this platform"); - return false; - } - - surface = vk::SurfaceKHR(unsafe_surface, *instance, dld); - return true; -} - -bool RendererVulkan::PickDevices() { - const auto devices = instance.EnumeratePhysicalDevices(); - if (!devices) { - LOG_ERROR(Render_Vulkan, "Failed to enumerate physical devices"); - return false; - } - +void RendererVulkan::InitializeDevice() { + const std::vector<VkPhysicalDevice> devices = instance.EnumeratePhysicalDevices(); const s32 device_index = Settings::values.vulkan_device.GetValue(); - if (device_index < 0 || device_index >= static_cast<s32>(devices->size())) { + if (device_index < 0 || device_index >= static_cast<s32>(devices.size())) { LOG_ERROR(Render_Vulkan, "Invalid device index {}!", device_index); - return false; - } - const vk::PhysicalDevice physical_device((*devices)[static_cast<std::size_t>(device_index)], - dld); - if (!VKDevice::IsSuitable(physical_device, *surface)) { - return false; + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); } - - device = std::make_unique<VKDevice>(*instance, physical_device, *surface, dld); - return device->Create(); + const vk::PhysicalDevice physical_device(devices[static_cast<size_t>(device_index)], dld); + device = std::make_unique<Device>(*instance, physical_device, *surface, dld); } void RendererVulkan::Report() const { @@ -426,7 +193,7 @@ void RendererVulkan::Report() const { const std::string driver_version = GetDriverVersion(*device); const std::string driver_name = fmt::format("{} {}", vendor_name, driver_version); - const std::string api_version = GetReadableVersion(device->GetApiVersion()); + const std::string api_version = GetReadableVersion(device->ApiVersion()); const std::string extensions = BuildCommaSeparatedExtensions(device->GetAvailableExtensions()); @@ -442,25 +209,21 @@ void RendererVulkan::Report() const { telemetry_session.AddField(field, "GPU_Vulkan_Extensions", extensions); } -std::vector<std::string> RendererVulkan::EnumerateDevices() { +std::vector<std::string> RendererVulkan::EnumerateDevices() try { vk::InstanceDispatch dld; - Common::DynamicLibrary library = OpenVulkanLibrary(); - vk::Instance instance = CreateInstance(library, dld); - if (!instance) { - return {}; - } - - const std::optional physical_devices = instance.EnumeratePhysicalDevices(); - if (!physical_devices) { - return {}; - } - + const Common::DynamicLibrary library = OpenLibrary(); + const vk::Instance instance = CreateInstance(library, dld, VK_API_VERSION_1_0); + const std::vector<VkPhysicalDevice> physical_devices = instance.EnumeratePhysicalDevices(); std::vector<std::string> names; - names.reserve(physical_devices->size()); - for (const auto& device : *physical_devices) { + names.reserve(physical_devices.size()); + for (const VkPhysicalDevice device : physical_devices) { names.push_back(vk::PhysicalDevice(device, dld).GetProperties().deviceName); } return names; + +} catch (const vk::Exception& exception) { + LOG_ERROR(Render_Vulkan, "Failed to enumerate devices with error: {}", exception.what()); + return {}; } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.h b/src/video_core/renderer_vulkan/renderer_vulkan.h index 49a4141ec..5575ffc54 100644 --- a/src/video_core/renderer_vulkan/renderer_vulkan.h +++ b/src/video_core/renderer_vulkan/renderer_vulkan.h @@ -11,7 +11,7 @@ #include "common/dynamic_library.h" #include "video_core/renderer_base.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Core { class TelemetrySession; @@ -27,16 +27,15 @@ class GPU; namespace Vulkan { +class Device; class StateTracker; class VKBlitScreen; -class VKDevice; class VKMemoryManager; class VKSwapchain; class VKScheduler; -class VKImage; struct VKScreenInfo { - VKImage* image{}; + VkImageView image_view{}; u32 width{}; u32 height{}; bool is_srgb{}; @@ -45,9 +44,9 @@ struct VKScreenInfo { class RendererVulkan final : public VideoCore::RendererBase { public: explicit RendererVulkan(Core::TelemetrySession& telemtry_session, - Core::Frontend::EmuWindow& emu_window, Core::Memory::Memory& cpu_memory, - Tegra::GPU& gpu, - std::unique_ptr<Core::Frontend::GraphicsContext> context); + Core::Frontend::EmuWindow& emu_window, + Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_, + std::unique_ptr<Core::Frontend::GraphicsContext> context_); ~RendererVulkan() override; bool Init() override; @@ -57,11 +56,7 @@ public: static std::vector<std::string> EnumerateDevices(); private: - bool CreateDebugCallback(); - - bool CreateSurface(); - - bool PickDevices(); + void InitializeDevice(); void Report() const; @@ -73,12 +68,13 @@ private: vk::InstanceDispatch dld; vk::Instance instance; + vk::SurfaceKHR surface; VKScreenInfo screen_info; - vk::DebugCallback debug_callback; - std::unique_ptr<VKDevice> device; + vk::DebugUtilsMessenger debug_callback; + std::unique_ptr<Device> device; std::unique_ptr<VKMemoryManager> memory_manager; std::unique_ptr<StateTracker> state_tracker; std::unique_ptr<VKScheduler> scheduler; diff --git a/src/video_core/renderer_vulkan/vk_blit_screen.cpp b/src/video_core/renderer_vulkan/vk_blit_screen.cpp index b5b60309e..5e184eb42 100644 --- a/src/video_core/renderer_vulkan/vk_blit_screen.cpp +++ b/src/video_core/renderer_vulkan/vk_blit_screen.cpp @@ -16,121 +16,25 @@ #include "core/frontend/emu_window.h" #include "core/memory.h" #include "video_core/gpu.h" -#include "video_core/morton.h" +#include "video_core/host_shaders/vulkan_present_frag_spv.h" +#include "video_core/host_shaders/vulkan_present_vert_spv.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_vulkan/renderer_vulkan.h" #include "video_core/renderer_vulkan/vk_blit_screen.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_image.h" #include "video_core/renderer_vulkan/vk_master_semaphore.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_shader_util.h" #include "video_core/renderer_vulkan/vk_swapchain.h" -#include "video_core/renderer_vulkan/wrapper.h" #include "video_core/surface.h" +#include "video_core/textures/decoders.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { namespace { -// Generated from the "shaders/" directory, read the instructions there. -constexpr u8 blit_vertex_code[] = { - 0x03, 0x02, 0x23, 0x07, 0x00, 0x00, 0x01, 0x00, 0x07, 0x00, 0x08, 0x00, 0x27, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x06, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x47, 0x4c, 0x53, 0x4c, 0x2e, 0x73, 0x74, 0x64, 0x2e, 0x34, 0x35, 0x30, - 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x0f, 0x00, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, - 0x00, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, - 0x25, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x0b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, - 0x0b, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x05, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x04, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x04, 0x00, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x05, 0x00, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x07, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x11, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x13, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x13, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x19, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x24, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x25, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x13, 0x00, 0x02, 0x00, 0x02, 0x00, 0x00, 0x00, 0x21, 0x00, 0x03, 0x00, - 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x16, 0x00, 0x03, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x20, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, 0x07, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x04, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, 0x08, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x08, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x04, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x06, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, - 0x0c, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, - 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, - 0x0e, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, - 0x0e, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x00, 0x04, 0x00, - 0x10, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x12, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x12, 0x00, 0x00, 0x00, - 0x13, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x14, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, - 0x19, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x1b, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x1c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x80, 0x3f, 0x20, 0x00, 0x04, 0x00, 0x21, 0x00, 0x00, 0x00, - 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x03, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x24, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, - 0x25, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x36, 0x00, 0x05, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, - 0x05, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, - 0x13, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x10, 0x00, 0x00, 0x00, - 0x16, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, - 0x1a, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x51, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x1d, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x51, 0x00, 0x05, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x50, 0x00, 0x07, 0x00, 0x07, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x91, 0x00, 0x05, 0x00, - 0x07, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x1f, 0x00, 0x00, 0x00, - 0x41, 0x00, 0x05, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, - 0x0f, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x03, 0x00, 0x22, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, - 0x3d, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, - 0x3e, 0x00, 0x03, 0x00, 0x24, 0x00, 0x00, 0x00, 0x26, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x01, 0x00, - 0x38, 0x00, 0x01, 0x00}; - -constexpr u8 blit_fragment_code[] = { - 0x03, 0x02, 0x23, 0x07, 0x00, 0x00, 0x01, 0x00, 0x07, 0x00, 0x08, 0x00, 0x14, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x06, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x47, 0x4c, 0x53, 0x4c, 0x2e, 0x73, 0x74, 0x64, 0x2e, 0x34, 0x35, 0x30, - 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x0f, 0x00, 0x07, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, - 0x00, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x10, 0x00, 0x03, 0x00, - 0x04, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x0d, 0x00, 0x00, 0x00, - 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x0d, 0x00, 0x00, 0x00, - 0x21, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x11, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x13, 0x00, 0x02, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x21, 0x00, 0x03, 0x00, 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x16, 0x00, 0x03, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, 0x07, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x08, 0x00, 0x00, 0x00, - 0x03, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x08, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x19, 0x00, 0x09, 0x00, 0x0a, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x03, 0x00, - 0x0b, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x0c, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x0c, 0x00, 0x00, 0x00, - 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, 0x0f, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x10, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x10, 0x00, 0x00, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x36, 0x00, 0x05, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, - 0x05, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, - 0x0d, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x57, 0x00, 0x05, 0x00, 0x07, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, - 0x0e, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x03, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x13, 0x00, 0x00, 0x00, 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00}; - struct ScreenRectVertex { ScreenRectVertex() = default; explicit ScreenRectVertex(f32 x, f32 y, f32 u, f32 v) : position{{x, y}}, tex_coord{{u, v}} {} @@ -173,9 +77,9 @@ constexpr std::array<f32, 4 * 4> MakeOrthographicMatrix(f32 width, f32 height) { // clang-format on } -std::size_t GetBytesPerPixel(const Tegra::FramebufferConfig& framebuffer) { +u32 GetBytesPerPixel(const Tegra::FramebufferConfig& framebuffer) { using namespace VideoCore::Surface; - return GetBytesPerPixel(PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)); + return BytesPerBlock(PixelFormatFromGPUPixelFormat(framebuffer.pixel_format)); } std::size_t GetSizeInBytes(const Tegra::FramebufferConfig& framebuffer) { @@ -210,7 +114,7 @@ struct VKBlitScreen::BufferData { VKBlitScreen::VKBlitScreen(Core::Memory::Memory& cpu_memory_, Core::Frontend::EmuWindow& render_window_, - VideoCore::RasterizerInterface& rasterizer_, const VKDevice& device_, + VideoCore::RasterizerInterface& rasterizer_, const Device& device_, VKMemoryManager& memory_manager_, VKSwapchain& swapchain_, VKScheduler& scheduler_, const VKScreenInfo& screen_info_) : cpu_memory{cpu_memory_}, render_window{render_window_}, rasterizer{rasterizer_}, @@ -239,34 +143,30 @@ VkSemaphore VKBlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer, bool scheduler.Wait(resource_ticks[image_index]); resource_ticks[image_index] = scheduler.CurrentTick(); - VKImage* blit_image = use_accelerated ? screen_info.image : raw_images[image_index].get(); - - UpdateDescriptorSet(image_index, blit_image->GetPresentView()); + UpdateDescriptorSet(image_index, + use_accelerated ? screen_info.image_view : *raw_image_views[image_index]); BufferData data; SetUniformData(data, framebuffer); SetVertexData(data, framebuffer); auto map = buffer_commit->Map(); - std::memcpy(map.GetAddress(), &data, sizeof(data)); + std::memcpy(map.Address(), &data, sizeof(data)); if (!use_accelerated) { const u64 image_offset = GetRawImageOffset(framebuffer, image_index); - const auto pixel_format = - VideoCore::Surface::PixelFormatFromGPUPixelFormat(framebuffer.pixel_format); const VAddr framebuffer_addr = framebuffer.address + framebuffer.offset; - const auto host_ptr = cpu_memory.GetPointer(framebuffer_addr); - rasterizer.FlushRegion(ToCacheAddr(host_ptr), GetSizeInBytes(framebuffer)); + const u8* const host_ptr = cpu_memory.GetPointer(framebuffer_addr); + const size_t size_bytes = GetSizeInBytes(framebuffer); + rasterizer.FlushRegion(ToCacheAddr(host_ptr), size_bytes); // TODO(Rodrigo): Read this from HLE constexpr u32 block_height_log2 = 4; - VideoCore::MortonSwizzle(VideoCore::MortonSwizzleMode::MortonToLinear, pixel_format, - framebuffer.stride, block_height_log2, framebuffer.height, 0, 1, 1, - map.GetAddress() + image_offset, host_ptr); - - blit_image->Transition(0, 1, 0, 1, VK_PIPELINE_STAGE_TRANSFER_BIT, - VK_ACCESS_TRANSFER_WRITE_BIT, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL); + const u32 bytes_per_pixel = GetBytesPerPixel(framebuffer); + Tegra::Texture::UnswizzleTexture( + std::span(map.Address() + image_offset, size_bytes), std::span(host_ptr, size_bytes), + bytes_per_pixel, framebuffer.width, framebuffer.height, 1, block_height_log2, 0); const VkBufferImageCopy copy{ .bufferOffset = image_offset, @@ -288,15 +188,44 @@ VkSemaphore VKBlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer, bool }, }; scheduler.Record( - [buffer = *buffer, image = *blit_image->GetHandle(), copy](vk::CommandBuffer cmdbuf) { - cmdbuf.CopyBufferToImage(buffer, image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, copy); + [buffer = *buffer, image = *raw_images[image_index], copy](vk::CommandBuffer cmdbuf) { + const VkImageMemoryBarrier base_barrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = 0, + .dstAccessMask = 0, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = image, + .subresourceRange = + { + .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT, + .baseMipLevel = 0, + .levelCount = 1, + .baseArrayLayer = 0, + .layerCount = 1, + }, + }; + VkImageMemoryBarrier read_barrier = base_barrier; + read_barrier.srcAccessMask = VK_ACCESS_HOST_WRITE_BIT; + read_barrier.dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT; + read_barrier.oldLayout = VK_IMAGE_LAYOUT_UNDEFINED; + + VkImageMemoryBarrier write_barrier = base_barrier; + write_barrier.srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT; + write_barrier.dstAccessMask = VK_ACCESS_SHADER_READ_BIT; + + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_HOST_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, + 0, read_barrier); + cmdbuf.CopyBufferToImage(buffer, image, VK_IMAGE_LAYOUT_GENERAL, copy); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, + VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, 0, write_barrier); }); } map.Release(); - blit_image->Transition(0, 1, 0, 1, VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT, - VK_ACCESS_SHADER_READ_BIT, VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL); - scheduler.Record([renderpass = *renderpass, framebuffer = *framebuffers[image_index], descriptor_set = descriptor_sets[image_index], buffer = *buffer, size = swapchain.GetSize(), pipeline = *pipeline, @@ -304,31 +233,31 @@ VkSemaphore VKBlitScreen::Draw(const Tegra::FramebufferConfig& framebuffer, bool const VkClearValue clear_color{ .color = {.float32 = {0.0f, 0.0f, 0.0f, 0.0f}}, }; - - VkRenderPassBeginInfo renderpass_bi; - renderpass_bi.sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO; - renderpass_bi.pNext = nullptr; - renderpass_bi.renderPass = renderpass; - renderpass_bi.framebuffer = framebuffer; - renderpass_bi.renderArea.offset.x = 0; - renderpass_bi.renderArea.offset.y = 0; - renderpass_bi.renderArea.extent = size; - renderpass_bi.clearValueCount = 1; - renderpass_bi.pClearValues = &clear_color; - - VkViewport viewport; - viewport.x = 0.0f; - viewport.y = 0.0f; - viewport.width = static_cast<float>(size.width); - viewport.height = static_cast<float>(size.height); - viewport.minDepth = 0.0f; - viewport.maxDepth = 1.0f; - - VkRect2D scissor; - scissor.offset.x = 0; - scissor.offset.y = 0; - scissor.extent = size; - + const VkRenderPassBeginInfo renderpass_bi{ + .sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO, + .pNext = nullptr, + .renderPass = renderpass, + .framebuffer = framebuffer, + .renderArea = + { + .offset = {0, 0}, + .extent = size, + }, + .clearValueCount = 1, + .pClearValues = &clear_color, + }; + const VkViewport viewport{ + .x = 0.0f, + .y = 0.0f, + .width = static_cast<float>(size.width), + .height = static_cast<float>(size.height), + .minDepth = 0.0f, + .maxDepth = 1.0f, + }; + const VkRect2D scissor{ + .offset = {0, 0}, + .extent = size, + }; cmdbuf.BeginRenderPass(renderpass_bi, VK_SUBPASS_CONTENTS_INLINE); cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_GRAPHICS, pipeline); cmdbuf.SetViewport(0, viewport); @@ -372,8 +301,8 @@ void VKBlitScreen::RefreshResources(const Tegra::FramebufferConfig& framebuffer) } void VKBlitScreen::CreateShaders() { - vertex_shader = BuildShader(device, sizeof(blit_vertex_code), blit_vertex_code); - fragment_shader = BuildShader(device, sizeof(blit_fragment_code), blit_fragment_code); + vertex_shader = BuildShader(device, VULKAN_PRESENT_VERT_SPV); + fragment_shader = BuildShader(device, VULKAN_PRESENT_FRAG_SPV); } void VKBlitScreen::CreateSemaphores() { @@ -420,7 +349,7 @@ void VKBlitScreen::CreateRenderPass() { const VkAttachmentReference color_attachment_ref{ .attachment = 0, - .layout = VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL, + .layout = VK_IMAGE_LAYOUT_GENERAL, }; const VkSubpassDescription subpass_description{ @@ -735,34 +664,56 @@ void VKBlitScreen::CreateStagingBuffer(const Tegra::FramebufferConfig& framebuff void VKBlitScreen::CreateRawImages(const Tegra::FramebufferConfig& framebuffer) { raw_images.resize(image_count); + raw_image_views.resize(image_count); raw_buffer_commits.resize(image_count); - const VkImageCreateInfo ci{ - .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO, - .pNext = nullptr, - .flags = 0, - .imageType = VK_IMAGE_TYPE_2D, - .format = GetFormat(framebuffer), - .extent = - { - .width = framebuffer.width, - .height = framebuffer.height, - .depth = 1, - }, - .mipLevels = 1, - .arrayLayers = 1, - .samples = VK_SAMPLE_COUNT_1_BIT, - .tiling = VK_IMAGE_TILING_LINEAR, - .usage = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_SAMPLED_BIT, - .sharingMode = VK_SHARING_MODE_EXCLUSIVE, - .queueFamilyIndexCount = 0, - .pQueueFamilyIndices = nullptr, - .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, - }; - - for (std::size_t i = 0; i < image_count; ++i) { - raw_images[i] = std::make_unique<VKImage>(device, scheduler, ci, VK_IMAGE_ASPECT_COLOR_BIT); - raw_buffer_commits[i] = memory_manager.Commit(raw_images[i]->GetHandle(), false); + for (size_t i = 0; i < image_count; ++i) { + raw_images[i] = device.GetLogical().CreateImage(VkImageCreateInfo{ + .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .imageType = VK_IMAGE_TYPE_2D, + .format = GetFormat(framebuffer), + .extent = + { + .width = framebuffer.width, + .height = framebuffer.height, + .depth = 1, + }, + .mipLevels = 1, + .arrayLayers = 1, + .samples = VK_SAMPLE_COUNT_1_BIT, + .tiling = VK_IMAGE_TILING_LINEAR, + .usage = VK_IMAGE_USAGE_TRANSFER_DST_BIT | VK_IMAGE_USAGE_SAMPLED_BIT, + .sharingMode = VK_SHARING_MODE_EXCLUSIVE, + .queueFamilyIndexCount = 0, + .pQueueFamilyIndices = nullptr, + .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, + }); + raw_buffer_commits[i] = memory_manager.Commit(raw_images[i], false); + raw_image_views[i] = device.GetLogical().CreateImageView(VkImageViewCreateInfo{ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .image = *raw_images[i], + .viewType = VK_IMAGE_VIEW_TYPE_2D, + .format = GetFormat(framebuffer), + .components = + { + .r = VK_COMPONENT_SWIZZLE_IDENTITY, + .g = VK_COMPONENT_SWIZZLE_IDENTITY, + .b = VK_COMPONENT_SWIZZLE_IDENTITY, + .a = VK_COMPONENT_SWIZZLE_IDENTITY, + }, + .subresourceRange = + { + .aspectMask = VK_IMAGE_ASPECT_COLOR_BIT, + .baseMipLevel = 0, + .levelCount = 1, + .baseArrayLayer = 0, + .layerCount = 1, + }, + }); } } @@ -789,7 +740,7 @@ void VKBlitScreen::UpdateDescriptorSet(std::size_t image_index, VkImageView imag const VkDescriptorImageInfo image_info{ .sampler = *sampler, .imageView = image_view, - .imageLayout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL, + .imageLayout = VK_IMAGE_LAYOUT_GENERAL, }; const VkWriteDescriptorSet sampler_write{ diff --git a/src/video_core/renderer_vulkan/vk_blit_screen.h b/src/video_core/renderer_vulkan/vk_blit_screen.h index 8f2839214..69ed61770 100644 --- a/src/video_core/renderer_vulkan/vk_blit_screen.h +++ b/src/video_core/renderer_vulkan/vk_blit_screen.h @@ -7,7 +7,7 @@ #include <memory> #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Core { class System; @@ -33,9 +33,8 @@ namespace Vulkan { struct ScreenInfo; +class Device; class RasterizerVulkan; -class VKDevice; -class VKImage; class VKScheduler; class VKSwapchain; @@ -43,7 +42,7 @@ class VKBlitScreen final { public: explicit VKBlitScreen(Core::Memory::Memory& cpu_memory, Core::Frontend::EmuWindow& render_window, - VideoCore::RasterizerInterface& rasterizer, const VKDevice& device, + VideoCore::RasterizerInterface& rasterizer, const Device& device, VKMemoryManager& memory_manager, VKSwapchain& swapchain, VKScheduler& scheduler, const VKScreenInfo& screen_info); ~VKBlitScreen(); @@ -86,7 +85,7 @@ private: Core::Memory::Memory& cpu_memory; Core::Frontend::EmuWindow& render_window; VideoCore::RasterizerInterface& rasterizer; - const VKDevice& device; + const Device& device; VKMemoryManager& memory_manager; VKSwapchain& swapchain; VKScheduler& scheduler; @@ -110,7 +109,8 @@ private: std::vector<u64> resource_ticks; std::vector<vk::Semaphore> semaphores; - std::vector<std::unique_ptr<VKImage>> raw_images; + std::vector<vk::Image> raw_images; + std::vector<vk::ImageView> raw_image_views; std::vector<VKMemoryCommit> raw_buffer_commits; u32 raw_width = 0; u32 raw_height = 0; diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp index d9d3da9ea..4d517c547 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp @@ -9,10 +9,10 @@ #include "core/core.h" #include "video_core/buffer_cache/buffer_cache.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_stream_buffer.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -31,20 +31,24 @@ constexpr VkAccessFlags UPLOAD_ACCESS_BARRIERS = VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_UNIFORM_READ_BIT | VK_ACCESS_VERTEX_ATTRIBUTE_READ_BIT | VK_ACCESS_INDEX_READ_BIT; -std::unique_ptr<VKStreamBuffer> CreateStreamBuffer(const VKDevice& device, VKScheduler& scheduler) { - return std::make_unique<VKStreamBuffer>(device, scheduler, BUFFER_USAGE); +constexpr VkAccessFlags TRANSFORM_FEEDBACK_WRITE_ACCESS = + VK_ACCESS_TRANSFORM_FEEDBACK_WRITE_BIT_EXT | VK_ACCESS_TRANSFORM_FEEDBACK_COUNTER_WRITE_BIT_EXT; + +std::unique_ptr<VKStreamBuffer> CreateStreamBuffer(const Device& device, VKScheduler& scheduler) { + return std::make_unique<VKStreamBuffer>(device, scheduler); } } // Anonymous namespace -Buffer::Buffer(const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler_, - VKStagingBufferPool& staging_pool_, VAddr cpu_addr, std::size_t size) - : BufferBlock{cpu_addr, size}, scheduler{scheduler_}, staging_pool{staging_pool_} { +Buffer::Buffer(const Device& device_, VKMemoryManager& memory_manager, VKScheduler& scheduler_, + VKStagingBufferPool& staging_pool_, VAddr cpu_addr_, std::size_t size_) + : BufferBlock{cpu_addr_, size_}, device{device_}, scheduler{scheduler_}, staging_pool{ + staging_pool_} { const VkBufferCreateInfo ci{ .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, .pNext = nullptr, .flags = 0, - .size = static_cast<VkDeviceSize>(size), + .size = static_cast<VkDeviceSize>(size_), .usage = BUFFER_USAGE | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT, .sharingMode = VK_SHARING_MODE_EXCLUSIVE, .queueFamilyIndexCount = 0, @@ -57,69 +61,86 @@ Buffer::Buffer(const VKDevice& device, VKMemoryManager& memory_manager, VKSchedu Buffer::~Buffer() = default; -void Buffer::Upload(std::size_t offset, std::size_t size, const u8* data) { - const auto& staging = staging_pool.GetUnusedBuffer(size, true); - std::memcpy(staging.commit->Map(size), data, size); +void Buffer::Upload(std::size_t offset, std::size_t data_size, const u8* data) { + const auto& staging = staging_pool.GetUnusedBuffer(data_size, true); + std::memcpy(staging.commit->Map(data_size), data, data_size); scheduler.RequestOutsideRenderPassOperationContext(); const VkBuffer handle = Handle(); - scheduler.Record([staging = *staging.handle, handle, offset, size](vk::CommandBuffer cmdbuf) { - cmdbuf.CopyBuffer(staging, handle, VkBufferCopy{0, offset, size}); - - const VkBufferMemoryBarrier barrier{ + scheduler.Record([staging = *staging.handle, handle, offset, data_size, + &device = device](vk::CommandBuffer cmdbuf) { + const VkBufferMemoryBarrier read_barrier{ .sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER, .pNext = nullptr, - .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, - .dstAccessMask = UPLOAD_ACCESS_BARRIERS, + .srcAccessMask = + VK_ACCESS_SHADER_WRITE_BIT | VK_ACCESS_TRANSFER_WRITE_BIT | + VK_ACCESS_HOST_WRITE_BIT | + (device.IsExtTransformFeedbackSupported() ? TRANSFORM_FEEDBACK_WRITE_ACCESS : 0), + .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT, .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, .buffer = handle, .offset = offset, - .size = size, + .size = data_size, }; - cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, UPLOAD_PIPELINE_STAGE, 0, {}, - barrier, {}); - }); -} - -void Buffer::Download(std::size_t offset, std::size_t size, u8* data) { - const auto& staging = staging_pool.GetUnusedBuffer(size, true); - scheduler.RequestOutsideRenderPassOperationContext(); - - const VkBuffer handle = Handle(); - scheduler.Record([staging = *staging.handle, handle, offset, size](vk::CommandBuffer cmdbuf) { - const VkBufferMemoryBarrier barrier{ + const VkBufferMemoryBarrier write_barrier{ .sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER, .pNext = nullptr, - .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT, - .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT, + .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = UPLOAD_ACCESS_BARRIERS, .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, .buffer = handle, .offset = offset, - .size = size, + .size = data_size, }; - - cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | - VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | - VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, - VK_PIPELINE_STAGE_TRANSFER_BIT, 0, {}, barrier, {}); - cmdbuf.CopyBuffer(handle, staging, VkBufferCopy{offset, 0, size}); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, + 0, read_barrier); + cmdbuf.CopyBuffer(staging, handle, VkBufferCopy{0, offset, data_size}); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, UPLOAD_PIPELINE_STAGE, 0, + write_barrier); }); +} + +void Buffer::Download(std::size_t offset, std::size_t data_size, u8* data) { + const auto& staging = staging_pool.GetUnusedBuffer(data_size, true); + scheduler.RequestOutsideRenderPassOperationContext(); + + const VkBuffer handle = Handle(); + scheduler.Record( + [staging = *staging.handle, handle, offset, data_size](vk::CommandBuffer cmdbuf) { + const VkBufferMemoryBarrier barrier{ + .sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .buffer = handle, + .offset = offset, + .size = data_size, + }; + + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_VERTEX_SHADER_BIT | + VK_PIPELINE_STAGE_FRAGMENT_SHADER_BIT | + VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, + VK_PIPELINE_STAGE_TRANSFER_BIT, 0, {}, barrier, {}); + cmdbuf.CopyBuffer(handle, staging, VkBufferCopy{offset, 0, data_size}); + }); scheduler.Finish(); - std::memcpy(data, staging.commit->Map(size), size); + std::memcpy(data, staging.commit->Map(data_size), data_size); } void Buffer::CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, - std::size_t size) { + std::size_t copy_size) { scheduler.RequestOutsideRenderPassOperationContext(); const VkBuffer dst_buffer = Handle(); scheduler.Record([src_buffer = src.Handle(), dst_buffer, src_offset, dst_offset, - size](vk::CommandBuffer cmdbuf) { - cmdbuf.CopyBuffer(src_buffer, dst_buffer, VkBufferCopy{src_offset, dst_offset, size}); + copy_size](vk::CommandBuffer cmdbuf) { + cmdbuf.CopyBuffer(src_buffer, dst_buffer, VkBufferCopy{src_offset, dst_offset, copy_size}); std::array<VkBufferMemoryBarrier, 2> barriers; barriers[0].sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER; @@ -130,7 +151,7 @@ void Buffer::CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst barriers[0].dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; barriers[0].buffer = src_buffer; barriers[0].offset = src_offset; - barriers[0].size = size; + barriers[0].size = copy_size; barriers[1].sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER; barriers[1].pNext = nullptr; barriers[1].srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT; @@ -139,19 +160,19 @@ void Buffer::CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst barriers[1].dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; barriers[1].buffer = dst_buffer; barriers[1].offset = dst_offset; - barriers[1].size = size; + barriers[1].size = copy_size; cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, UPLOAD_PIPELINE_STAGE, 0, {}, barriers, {}); }); } -VKBufferCache::VKBufferCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, - const VKDevice& device_, VKMemoryManager& memory_manager_, - VKScheduler& scheduler_, VKStagingBufferPool& staging_pool_) - : VideoCommon::BufferCache<Buffer, VkBuffer, VKStreamBuffer>{rasterizer, gpu_memory, cpu_memory, - CreateStreamBuffer(device_, - scheduler_)}, +VKBufferCache::VKBufferCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::MemoryManager& gpu_memory_, Core::Memory::Memory& cpu_memory_, + const Device& device_, VKMemoryManager& memory_manager_, + VKScheduler& scheduler_, VKStreamBuffer& stream_buffer_, + VKStagingBufferPool& staging_pool_) + : VideoCommon::BufferCache<Buffer, VkBuffer, VKStreamBuffer>{rasterizer_, gpu_memory_, + cpu_memory_, stream_buffer_}, device{device_}, memory_manager{memory_manager_}, scheduler{scheduler_}, staging_pool{ staging_pool_} {} diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h index 7fb5ceedf..1c39aed34 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.h +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h @@ -11,26 +11,26 @@ #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/vk_stream_buffer.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class VKMemoryManager; class VKScheduler; class Buffer final : public VideoCommon::BufferBlock { public: - explicit Buffer(const VKDevice& device, VKMemoryManager& memory_manager, VKScheduler& scheduler, - VKStagingBufferPool& staging_pool, VAddr cpu_addr, std::size_t size); + explicit Buffer(const Device& device, VKMemoryManager& memory_manager, VKScheduler& scheduler, + VKStagingBufferPool& staging_pool, VAddr cpu_addr_, std::size_t size_); ~Buffer(); - void Upload(std::size_t offset, std::size_t size, const u8* data); + void Upload(std::size_t offset, std::size_t data_size, const u8* data); - void Download(std::size_t offset, std::size_t size, u8* data); + void Download(std::size_t offset, std::size_t data_size, u8* data); void CopyFrom(const Buffer& src, std::size_t src_offset, std::size_t dst_offset, - std::size_t size); + std::size_t copy_size); VkBuffer Handle() const { return *buffer.handle; @@ -41,6 +41,7 @@ public: } private: + const Device& device; VKScheduler& scheduler; VKStagingBufferPool& staging_pool; @@ -51,8 +52,9 @@ class VKBufferCache final : public VideoCommon::BufferCache<Buffer, VkBuffer, VK public: explicit VKBufferCache(VideoCore::RasterizerInterface& rasterizer, Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, - const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, VKStagingBufferPool& staging_pool); + const Device& device, VKMemoryManager& memory_manager, + VKScheduler& scheduler, VKStreamBuffer& stream_buffer, + VKStagingBufferPool& staging_pool); ~VKBufferCache(); BufferInfo GetEmptyBuffer(std::size_t size) override; @@ -61,7 +63,7 @@ protected: std::shared_ptr<Buffer> CreateBlock(VAddr cpu_addr, std::size_t size) override; private: - const VKDevice& device; + const Device& device; VKMemoryManager& memory_manager; VKScheduler& scheduler; VKStagingBufferPool& staging_pool; diff --git a/src/video_core/renderer_vulkan/vk_command_pool.cpp b/src/video_core/renderer_vulkan/vk_command_pool.cpp index f1abd4b1a..a99df9323 100644 --- a/src/video_core/renderer_vulkan/vk_command_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_command_pool.cpp @@ -5,15 +5,20 @@ #include <cstddef> #include "video_core/renderer_vulkan/vk_command_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { constexpr size_t COMMAND_BUFFER_POOL_SIZE = 0x1000; -CommandPool::CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device) - : ResourcePool(master_semaphore, COMMAND_BUFFER_POOL_SIZE), device{device} {} +struct CommandPool::Pool { + vk::CommandPool handle; + vk::CommandBuffers cmdbufs; +}; + +CommandPool::CommandPool(MasterSemaphore& master_semaphore_, const Device& device_) + : ResourcePool(master_semaphore_, COMMAND_BUFFER_POOL_SIZE), device{device_} {} CommandPool::~CommandPool() = default; diff --git a/src/video_core/renderer_vulkan/vk_command_pool.h b/src/video_core/renderer_vulkan/vk_command_pool.h index 3aee239b9..61c26a22a 100644 --- a/src/video_core/renderer_vulkan/vk_command_pool.h +++ b/src/video_core/renderer_vulkan/vk_command_pool.h @@ -2,33 +2,32 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#pragma once + #include <cstddef> #include <vector> #include "video_core/renderer_vulkan/vk_resource_pool.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { +class Device; class MasterSemaphore; -class VKDevice; class CommandPool final : public ResourcePool { public: - explicit CommandPool(MasterSemaphore& master_semaphore, const VKDevice& device); - virtual ~CommandPool(); + explicit CommandPool(MasterSemaphore& master_semaphore_, const Device& device_); + ~CommandPool() override; void Allocate(size_t begin, size_t end) override; VkCommandBuffer Commit(); private: - struct Pool { - vk::CommandPool handle; - vk::CommandBuffers cmdbufs; - }; + struct Pool; - const VKDevice& device; + const Device& device; std::vector<Pool> pools; }; diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.cpp b/src/video_core/renderer_vulkan/vk_compute_pass.cpp index 9637c6059..02a6d54b7 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pass.cpp @@ -10,111 +10,21 @@ #include "common/alignment.h" #include "common/assert.h" #include "common/common_types.h" +#include "video_core/host_shaders/vulkan_quad_array_comp_spv.h" +#include "video_core/host_shaders/vulkan_quad_indexed_comp_spv.h" +#include "video_core/host_shaders/vulkan_uint8_comp_spv.h" #include "video_core/renderer_vulkan/vk_compute_pass.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { namespace { -// Quad array SPIR-V module. Generated from the "shaders/" directory, read the instructions there. -constexpr u8 quad_array[] = { - 0x03, 0x02, 0x23, 0x07, 0x00, 0x00, 0x01, 0x00, 0x07, 0x00, 0x08, 0x00, 0x54, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x06, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x47, 0x4c, 0x53, 0x4c, 0x2e, 0x73, 0x74, 0x64, 0x2e, 0x34, 0x35, 0x30, - 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x0f, 0x00, 0x06, 0x00, 0x05, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, - 0x00, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x10, 0x00, 0x06, 0x00, 0x04, 0x00, 0x00, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x13, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x05, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x14, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x16, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x16, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x05, 0x00, 0x29, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x29, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x4a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, - 0x13, 0x00, 0x02, 0x00, 0x02, 0x00, 0x00, 0x00, 0x21, 0x00, 0x03, 0x00, 0x03, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x03, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x03, 0x00, 0x13, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x03, 0x00, 0x14, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, - 0x15, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, - 0x15, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, - 0x18, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x02, 0x00, - 0x1b, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, 0x29, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x20, 0x00, 0x04, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, - 0x3b, 0x00, 0x04, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x20, 0x00, 0x04, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, - 0x1c, 0x00, 0x04, 0x00, 0x34, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, - 0x2c, 0x00, 0x09, 0x00, 0x34, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, - 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, - 0x37, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x3a, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, - 0x34, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x44, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, 0x47, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, - 0x00, 0x04, 0x00, 0x00, 0x2c, 0x00, 0x06, 0x00, 0x09, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, - 0x49, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x35, 0x00, 0x00, 0x00, 0x36, 0x00, 0x05, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x05, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x3a, 0x00, 0x00, 0x00, - 0x3b, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x4c, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x4c, 0x00, 0x00, 0x00, 0xf6, 0x00, 0x04, 0x00, 0x4b, 0x00, 0x00, 0x00, - 0x4e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x4d, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x4d, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x0d, 0x00, 0x00, 0x00, - 0x0e, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x12, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, - 0x44, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x17, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, - 0x17, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, - 0x19, 0x00, 0x00, 0x00, 0xae, 0x00, 0x05, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, - 0x12, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0xf7, 0x00, 0x03, 0x00, 0x1e, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0xfa, 0x00, 0x04, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1d, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, - 0x4b, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1e, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, - 0x21, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x21, 0x00, 0x00, 0x00, 0xf5, 0x00, 0x07, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0xb0, 0x00, 0x05, 0x00, 0x1b, 0x00, 0x00, 0x00, - 0x27, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0xf6, 0x00, 0x04, 0x00, - 0x23, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfa, 0x00, 0x04, 0x00, - 0x27, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, - 0x22, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x2f, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x32, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, 0x2f, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, - 0x3e, 0x00, 0x03, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x38, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, - 0x07, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, - 0x3d, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, - 0x80, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, 0x33, 0x00, 0x00, 0x00, - 0x3d, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, - 0x12, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, 0x41, 0x00, 0x06, 0x00, 0x44, 0x00, 0x00, 0x00, - 0x45, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, - 0x3e, 0x00, 0x03, 0x00, 0x45, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x48, 0x00, 0x00, 0x00, 0x53, 0x00, 0x00, 0x00, 0x47, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x21, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x23, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x4b, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x4e, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x4c, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x4b, 0x00, 0x00, 0x00, - 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00, -}; - VkDescriptorSetLayoutBinding BuildQuadArrayPassDescriptorSetLayoutBinding() { return { .binding = 0, @@ -144,208 +54,6 @@ VkPushConstantRange BuildComputePushConstantRange(std::size_t size) { }; } -// Uint8 SPIR-V module. Generated from the "shaders/" directory. -constexpr u8 uint8_pass[] = { - 0x03, 0x02, 0x23, 0x07, 0x00, 0x00, 0x01, 0x00, 0x07, 0x00, 0x08, 0x00, 0x2f, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, 0x01, 0x00, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, - 0x51, 0x11, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, 0x61, 0x11, 0x00, 0x00, 0x0a, 0x00, 0x07, 0x00, - 0x53, 0x50, 0x56, 0x5f, 0x4b, 0x48, 0x52, 0x5f, 0x31, 0x36, 0x62, 0x69, 0x74, 0x5f, 0x73, 0x74, - 0x6f, 0x72, 0x61, 0x67, 0x65, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x07, 0x00, 0x53, 0x50, 0x56, 0x5f, - 0x4b, 0x48, 0x52, 0x5f, 0x38, 0x62, 0x69, 0x74, 0x5f, 0x73, 0x74, 0x6f, 0x72, 0x61, 0x67, 0x65, - 0x00, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x06, 0x00, 0x01, 0x00, 0x00, 0x00, 0x47, 0x4c, 0x53, 0x4c, - 0x2e, 0x73, 0x74, 0x64, 0x2e, 0x34, 0x35, 0x30, 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x03, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x06, 0x00, 0x05, 0x00, 0x00, 0x00, - 0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, 0x00, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x10, 0x00, 0x06, 0x00, 0x04, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x0b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x12, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x48, 0x00, 0x04, 0x00, 0x13, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x13, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, - 0x13, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x15, 0x00, 0x00, 0x00, - 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x15, 0x00, 0x00, 0x00, - 0x21, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x1f, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x48, 0x00, 0x04, 0x00, 0x20, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x20, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, - 0x20, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x22, 0x00, 0x00, 0x00, - 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x22, 0x00, 0x00, 0x00, - 0x21, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x2e, 0x00, 0x00, 0x00, - 0x0b, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x13, 0x00, 0x02, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x21, 0x00, 0x03, 0x00, 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, - 0x07, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, - 0x0a, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, - 0x0a, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, - 0x0d, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x03, 0x00, - 0x12, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, 0x13, 0x00, 0x00, 0x00, - 0x12, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x14, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x13, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x14, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, 0x20, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x02, 0x00, 0x1a, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x03, 0x00, - 0x1f, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, 0x20, 0x00, 0x00, 0x00, - 0x1f, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x21, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x20, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x21, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x26, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x00, 0x00, - 0x00, 0x04, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x06, 0x00, 0x09, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, - 0x2c, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x2d, 0x00, 0x00, 0x00, 0x36, 0x00, 0x05, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x05, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x07, 0x00, 0x00, 0x00, - 0x08, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x0d, 0x00, 0x00, 0x00, - 0x0e, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x03, 0x00, - 0x08, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x10, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x44, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x16, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, - 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0xb0, 0x00, 0x05, 0x00, - 0x1a, 0x00, 0x00, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, - 0xf7, 0x00, 0x03, 0x00, 0x1d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfa, 0x00, 0x04, 0x00, - 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, - 0x1c, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, - 0x08, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, - 0x08, 0x00, 0x00, 0x00, 0x41, 0x00, 0x06, 0x00, 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, - 0x15, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x71, 0x00, 0x04, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x41, 0x00, 0x06, 0x00, - 0x2a, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x24, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x03, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x1d, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1d, 0x00, 0x00, 0x00, - 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00, -}; - -// Quad indexed SPIR-V module. Generated from the "shaders/" directory. -constexpr u8 QUAD_INDEXED_SPV[] = { - 0x03, 0x02, 0x23, 0x07, 0x00, 0x00, 0x01, 0x00, 0x07, 0x00, 0x08, 0x00, 0x7c, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x11, 0x00, 0x02, 0x00, 0x01, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x06, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x47, 0x4c, 0x53, 0x4c, 0x2e, 0x73, 0x74, 0x64, 0x2e, 0x34, 0x35, 0x30, - 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x0f, 0x00, 0x06, 0x00, 0x05, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, - 0x00, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x10, 0x00, 0x06, 0x00, 0x04, 0x00, 0x00, 0x00, - 0x11, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x15, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x04, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x05, 0x00, 0x16, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x16, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x04, 0x00, 0x18, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x05, 0x00, 0x22, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x22, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, - 0x23, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x22, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x56, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x04, 0x00, 0x00, 0x00, 0x48, 0x00, 0x04, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x18, 0x00, 0x00, 0x00, 0x48, 0x00, 0x05, 0x00, 0x57, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0x23, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x03, 0x00, 0x57, 0x00, 0x00, 0x00, - 0x03, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x59, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x59, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x47, 0x00, 0x04, 0x00, 0x72, 0x00, 0x00, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x19, 0x00, 0x00, 0x00, 0x13, 0x00, 0x02, 0x00, 0x02, 0x00, 0x00, 0x00, 0x21, 0x00, 0x03, 0x00, - 0x03, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x20, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x07, 0x00, 0x00, 0x00, - 0x07, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x15, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x20, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x17, 0x00, 0x04, 0x00, 0x0a, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x0b, 0x00, 0x00, 0x00, - 0x0c, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x0d, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x0e, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x13, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x03, 0x00, 0x15, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, 0x16, 0x00, 0x00, 0x00, 0x15, 0x00, 0x00, 0x00, - 0x20, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x16, 0x00, 0x00, 0x00, - 0x3b, 0x00, 0x04, 0x00, 0x17, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x14, 0x00, 0x02, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x21, 0x00, 0x00, 0x00, 0x08, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x04, 0x00, 0x22, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x22, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x23, 0x00, 0x00, 0x00, - 0x24, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x25, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x26, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x3b, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x3f, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x04, 0x00, 0x41, 0x00, 0x00, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x42, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x43, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x09, 0x00, 0x41, 0x00, 0x00, 0x00, - 0x44, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, - 0x42, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x00, 0x00, 0x43, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, - 0x46, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0x41, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x03, 0x00, - 0x56, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, 0x1e, 0x00, 0x03, 0x00, 0x57, 0x00, 0x00, 0x00, - 0x56, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x58, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x57, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x58, 0x00, 0x00, 0x00, 0x59, 0x00, 0x00, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x5b, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x20, 0x00, 0x04, 0x00, 0x69, 0x00, 0x00, 0x00, 0x09, 0x00, 0x00, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x70, 0x00, 0x00, 0x00, - 0x00, 0x04, 0x00, 0x00, 0x2b, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x71, 0x00, 0x00, 0x00, - 0x01, 0x00, 0x00, 0x00, 0x2c, 0x00, 0x06, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x72, 0x00, 0x00, 0x00, - 0x70, 0x00, 0x00, 0x00, 0x71, 0x00, 0x00, 0x00, 0x71, 0x00, 0x00, 0x00, 0x36, 0x00, 0x05, 0x00, - 0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x05, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x04, 0x00, 0x46, 0x00, 0x00, 0x00, - 0x47, 0x00, 0x00, 0x00, 0x07, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x74, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x74, 0x00, 0x00, 0x00, 0xf6, 0x00, 0x04, 0x00, 0x73, 0x00, 0x00, 0x00, - 0x76, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, 0x75, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x75, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x0e, 0x00, 0x00, 0x00, - 0x0f, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x0f, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x13, 0x00, 0x00, 0x00, - 0x44, 0x00, 0x05, 0x00, 0x09, 0x00, 0x00, 0x00, 0x19, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, - 0x19, 0x00, 0x00, 0x00, 0xaf, 0x00, 0x05, 0x00, 0x1b, 0x00, 0x00, 0x00, 0x1c, 0x00, 0x00, 0x00, - 0x14, 0x00, 0x00, 0x00, 0x1a, 0x00, 0x00, 0x00, 0xf7, 0x00, 0x03, 0x00, 0x1e, 0x00, 0x00, 0x00, - 0x00, 0x00, 0x00, 0x00, 0xfa, 0x00, 0x04, 0x00, 0x1c, 0x00, 0x00, 0x00, 0x1d, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1d, 0x00, 0x00, 0x00, 0xf9, 0x00, 0x02, 0x00, - 0x73, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x1e, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, - 0x26, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, - 0x3d, 0x00, 0x04, 0x00, 0x06, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0x27, 0x00, 0x00, 0x00, - 0xc4, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x21, 0x00, 0x00, 0x00, - 0x28, 0x00, 0x00, 0x00, 0x82, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, - 0x2b, 0x00, 0x00, 0x00, 0x28, 0x00, 0x00, 0x00, 0xc4, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x31, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, 0x2e, 0x00, 0x00, 0x00, 0x82, 0x00, 0x05, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x31, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x35, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x35, 0x00, 0x00, 0x00, - 0xf5, 0x00, 0x07, 0x00, 0x09, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x00, 0x00, - 0x1e, 0x00, 0x00, 0x00, 0x6f, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0xb0, 0x00, 0x05, 0x00, - 0x1b, 0x00, 0x00, 0x00, 0x3c, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x3b, 0x00, 0x00, 0x00, - 0xf6, 0x00, 0x04, 0x00, 0x37, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, - 0xfa, 0x00, 0x04, 0x00, 0x3c, 0x00, 0x00, 0x00, 0x36, 0x00, 0x00, 0x00, 0x37, 0x00, 0x00, 0x00, - 0xf8, 0x00, 0x02, 0x00, 0x36, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x40, 0x00, 0x00, 0x00, 0x11, 0x00, 0x00, 0x00, 0x3f, 0x00, 0x00, 0x00, 0x3e, 0x00, 0x03, 0x00, - 0x47, 0x00, 0x00, 0x00, 0x44, 0x00, 0x00, 0x00, 0x41, 0x00, 0x05, 0x00, 0x07, 0x00, 0x00, 0x00, - 0x48, 0x00, 0x00, 0x00, 0x47, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, 0x48, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, - 0x06, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x49, 0x00, 0x00, 0x00, - 0xc3, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x4e, 0x00, 0x00, 0x00, 0x4a, 0x00, 0x00, 0x00, - 0x2e, 0x00, 0x00, 0x00, 0xc7, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, - 0x4a, 0x00, 0x00, 0x00, 0x32, 0x00, 0x00, 0x00, 0x84, 0x00, 0x05, 0x00, 0x06, 0x00, 0x00, 0x00, - 0x54, 0x00, 0x00, 0x00, 0x52, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x41, 0x00, 0x06, 0x00, - 0x5b, 0x00, 0x00, 0x00, 0x5c, 0x00, 0x00, 0x00, 0x59, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, - 0x4e, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x5d, 0x00, 0x00, 0x00, - 0x5c, 0x00, 0x00, 0x00, 0xcb, 0x00, 0x06, 0x00, 0x09, 0x00, 0x00, 0x00, 0x62, 0x00, 0x00, 0x00, - 0x5d, 0x00, 0x00, 0x00, 0x54, 0x00, 0x00, 0x00, 0x29, 0x00, 0x00, 0x00, 0x7c, 0x00, 0x04, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x65, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x67, 0x00, 0x00, 0x00, 0x65, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, - 0x41, 0x00, 0x05, 0x00, 0x69, 0x00, 0x00, 0x00, 0x6a, 0x00, 0x00, 0x00, 0x24, 0x00, 0x00, 0x00, - 0x42, 0x00, 0x00, 0x00, 0x3d, 0x00, 0x04, 0x00, 0x09, 0x00, 0x00, 0x00, 0x6b, 0x00, 0x00, 0x00, - 0x6a, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, 0x09, 0x00, 0x00, 0x00, 0x6c, 0x00, 0x00, 0x00, - 0x62, 0x00, 0x00, 0x00, 0x6b, 0x00, 0x00, 0x00, 0x41, 0x00, 0x06, 0x00, 0x5b, 0x00, 0x00, 0x00, - 0x6d, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x42, 0x00, 0x00, 0x00, 0x67, 0x00, 0x00, 0x00, - 0x3e, 0x00, 0x03, 0x00, 0x6d, 0x00, 0x00, 0x00, 0x6c, 0x00, 0x00, 0x00, 0x80, 0x00, 0x05, 0x00, - 0x09, 0x00, 0x00, 0x00, 0x6f, 0x00, 0x00, 0x00, 0x7b, 0x00, 0x00, 0x00, 0x25, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x35, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x37, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x73, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x76, 0x00, 0x00, 0x00, - 0xf9, 0x00, 0x02, 0x00, 0x74, 0x00, 0x00, 0x00, 0xf8, 0x00, 0x02, 0x00, 0x73, 0x00, 0x00, 0x00, - 0xfd, 0x00, 0x01, 0x00, 0x38, 0x00, 0x01, 0x00, -}; - std::array<VkDescriptorSetLayoutBinding, 2> BuildInputOutputDescriptorSetBindings() { return {{ { @@ -378,11 +86,11 @@ VkDescriptorUpdateTemplateEntryKHR BuildInputOutputDescriptorUpdateTemplate() { } // Anonymous namespace -VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descriptor_pool, +VKComputePass::VKComputePass(const Device& device, VKDescriptorPool& descriptor_pool, vk::Span<VkDescriptorSetLayoutBinding> bindings, vk::Span<VkDescriptorUpdateTemplateEntryKHR> templates, - vk::Span<VkPushConstantRange> push_constants, std::size_t code_size, - const u8* code) { + vk::Span<VkPushConstantRange> push_constants, + std::span<const u32> code) { descriptor_set_layout = device.GetLogical().CreateDescriptorSetLayout({ .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_SET_LAYOUT_CREATE_INFO, .pNext = nullptr, @@ -390,7 +98,6 @@ VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descripto .bindingCount = bindings.size(), .pBindings = bindings.data(), }); - layout = device.GetLogical().CreatePipelineLayout({ .sType = VK_STRUCTURE_TYPE_PIPELINE_LAYOUT_CREATE_INFO, .pNext = nullptr, @@ -400,7 +107,6 @@ VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descripto .pushConstantRangeCount = push_constants.size(), .pPushConstantRanges = push_constants.data(), }); - if (!templates.empty()) { descriptor_template = device.GetLogical().CreateDescriptorUpdateTemplateKHR({ .sType = VK_STRUCTURE_TYPE_DESCRIPTOR_UPDATE_TEMPLATE_CREATE_INFO_KHR, @@ -417,18 +123,13 @@ VKComputePass::VKComputePass(const VKDevice& device, VKDescriptorPool& descripto descriptor_allocator.emplace(descriptor_pool, *descriptor_set_layout); } - - auto code_copy = std::make_unique<u32[]>(code_size / sizeof(u32) + 1); - std::memcpy(code_copy.get(), code, code_size); - module = device.GetLogical().CreateShaderModule({ .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, .pNext = nullptr, .flags = 0, - .codeSize = code_size, - .pCode = code_copy.get(), + .codeSize = static_cast<u32>(code.size_bytes()), + .pCode = code.data(), }); - pipeline = device.GetLogical().CreateComputePipeline({ .sType = VK_STRUCTURE_TYPE_COMPUTE_PIPELINE_CREATE_INFO, .pNext = nullptr, @@ -461,15 +162,15 @@ VkDescriptorSet VKComputePass::CommitDescriptorSet( return set; } -QuadArrayPass::QuadArrayPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue) - : VKComputePass(device, descriptor_pool, BuildQuadArrayPassDescriptorSetLayoutBinding(), +QuadArrayPass::QuadArrayPass(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_) + : VKComputePass(device_, descriptor_pool_, BuildQuadArrayPassDescriptorSetLayoutBinding(), BuildQuadArrayPassDescriptorUpdateTemplateEntry(), - BuildComputePushConstantRange(sizeof(u32)), std::size(quad_array), quad_array), - scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, - update_descriptor_queue{update_descriptor_queue} {} + BuildComputePushConstantRange(sizeof(u32)), VULKAN_QUAD_ARRAY_COMP_SPV), + scheduler{scheduler_}, staging_buffer_pool{staging_buffer_pool_}, + update_descriptor_queue{update_descriptor_queue_} {} QuadArrayPass::~QuadArrayPass() = default; @@ -510,14 +211,13 @@ std::pair<VkBuffer, VkDeviceSize> QuadArrayPass::Assemble(u32 num_vertices, u32 return {*buffer.handle, 0}; } -Uint8Pass::Uint8Pass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue) +Uint8Pass::Uint8Pass(const Device& device, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool, VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_) : VKComputePass(device, descriptor_pool, BuildInputOutputDescriptorSetBindings(), - BuildInputOutputDescriptorUpdateTemplate(), {}, std::size(uint8_pass), - uint8_pass), - scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, - update_descriptor_queue{update_descriptor_queue} {} + BuildInputOutputDescriptorUpdateTemplate(), {}, VULKAN_UINT8_COMP_SPV), + scheduler{scheduler_}, staging_buffer_pool{staging_buffer_pool_}, + update_descriptor_queue{update_descriptor_queue_} {} Uint8Pass::~Uint8Pass() = default; @@ -555,16 +255,15 @@ std::pair<VkBuffer, u64> Uint8Pass::Assemble(u32 num_vertices, VkBuffer src_buff return {*buffer.handle, 0}; } -QuadIndexedPass::QuadIndexedPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue) - : VKComputePass(device, descriptor_pool, BuildInputOutputDescriptorSetBindings(), +QuadIndexedPass::QuadIndexedPass(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_) + : VKComputePass(device_, descriptor_pool_, BuildInputOutputDescriptorSetBindings(), BuildInputOutputDescriptorUpdateTemplate(), - BuildComputePushConstantRange(sizeof(u32) * 2), std::size(QUAD_INDEXED_SPV), - QUAD_INDEXED_SPV), - scheduler{scheduler}, staging_buffer_pool{staging_buffer_pool}, - update_descriptor_queue{update_descriptor_queue} {} + BuildComputePushConstantRange(sizeof(u32) * 2), VULKAN_QUAD_INDEXED_COMP_SPV), + scheduler{scheduler_}, staging_buffer_pool{staging_buffer_pool_}, + update_descriptor_queue{update_descriptor_queue_} {} QuadIndexedPass::~QuadIndexedPass() = default; diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.h b/src/video_core/renderer_vulkan/vk_compute_pass.h index acc94f27e..7ddb09afb 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.h +++ b/src/video_core/renderer_vulkan/vk_compute_pass.h @@ -5,27 +5,27 @@ #pragma once #include <optional> +#include <span> #include <utility> #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class VKScheduler; class VKStagingBufferPool; class VKUpdateDescriptorQueue; class VKComputePass { public: - explicit VKComputePass(const VKDevice& device, VKDescriptorPool& descriptor_pool, + explicit VKComputePass(const Device& device, VKDescriptorPool& descriptor_pool, vk::Span<VkDescriptorSetLayoutBinding> bindings, vk::Span<VkDescriptorUpdateTemplateEntryKHR> templates, - vk::Span<VkPushConstantRange> push_constants, std::size_t code_size, - const u8* code); + vk::Span<VkPushConstantRange> push_constants, std::span<const u32> code); ~VKComputePass(); protected: @@ -43,10 +43,10 @@ private: class QuadArrayPass final : public VKComputePass { public: - explicit QuadArrayPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue); + explicit QuadArrayPass(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_); ~QuadArrayPass(); std::pair<VkBuffer, VkDeviceSize> Assemble(u32 num_vertices, u32 first); @@ -59,9 +59,10 @@ private: class Uint8Pass final : public VKComputePass { public: - explicit Uint8Pass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue); + explicit Uint8Pass(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_); ~Uint8Pass(); std::pair<VkBuffer, u64> Assemble(u32 num_vertices, VkBuffer src_buffer, u64 src_offset); @@ -74,10 +75,10 @@ private: class QuadIndexedPass final : public VKComputePass { public: - explicit QuadIndexedPass(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKStagingBufferPool& staging_buffer_pool, - VKUpdateDescriptorQueue& update_descriptor_queue); + explicit QuadIndexedPass(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKStagingBufferPool& staging_buffer_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_); ~QuadIndexedPass(); std::pair<VkBuffer, u64> Assemble(Tegra::Engines::Maxwell3D::Regs::IndexFormat index_format, diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp index 9be72dc9b..3a48219b7 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.cpp @@ -6,25 +6,25 @@ #include "video_core/renderer_vulkan/vk_compute_pipeline.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -VKComputePipeline::VKComputePipeline(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - const SPIRVShader& shader) - : device{device}, scheduler{scheduler}, entries{shader.entries}, +VKComputePipeline::VKComputePipeline(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + const SPIRVShader& shader_) + : device{device_}, scheduler{scheduler_}, entries{shader_.entries}, descriptor_set_layout{CreateDescriptorSetLayout()}, - descriptor_allocator{descriptor_pool, *descriptor_set_layout}, - update_descriptor_queue{update_descriptor_queue}, layout{CreatePipelineLayout()}, + descriptor_allocator{descriptor_pool_, *descriptor_set_layout}, + update_descriptor_queue{update_descriptor_queue_}, layout{CreatePipelineLayout()}, descriptor_template{CreateDescriptorUpdateTemplate()}, - shader_module{CreateShaderModule(shader.code)}, pipeline{CreatePipeline()} {} + shader_module{CreateShaderModule(shader_.code)}, pipeline{CreatePipeline()} {} VKComputePipeline::~VKComputePipeline() = default; diff --git a/src/video_core/renderer_vulkan/vk_compute_pipeline.h b/src/video_core/renderer_vulkan/vk_compute_pipeline.h index 6e2f22a4a..7e16575ac 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_compute_pipeline.h @@ -7,20 +7,20 @@ #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class VKScheduler; class VKUpdateDescriptorQueue; class VKComputePipeline final { public: - explicit VKComputePipeline(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - const SPIRVShader& shader); + explicit VKComputePipeline(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, + const SPIRVShader& shader_); ~VKComputePipeline(); VkDescriptorSet CommitDescriptorSet(); @@ -48,7 +48,7 @@ private: vk::Pipeline CreatePipeline() const; - const VKDevice& device; + const Device& device; VKScheduler& scheduler; ShaderEntries entries; diff --git a/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp b/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp index f38e089d5..ef9fb5910 100644 --- a/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_descriptor_pool.cpp @@ -6,10 +6,10 @@ #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_resource_pool.h" #include "video_core/renderer_vulkan/vk_scheduler.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -32,7 +32,7 @@ void DescriptorAllocator::Allocate(std::size_t begin, std::size_t end) { descriptors_allocations.push_back(descriptor_pool.AllocateDescriptors(layout, end - begin)); } -VKDescriptorPool::VKDescriptorPool(const VKDevice& device_, VKScheduler& scheduler) +VKDescriptorPool::VKDescriptorPool(const Device& device_, VKScheduler& scheduler) : device{device_}, master_semaphore{scheduler.GetMasterSemaphore()}, active_pool{ AllocateNewPool()} {} diff --git a/src/video_core/renderer_vulkan/vk_descriptor_pool.h b/src/video_core/renderer_vulkan/vk_descriptor_pool.h index 544f32a20..f892be7be 100644 --- a/src/video_core/renderer_vulkan/vk_descriptor_pool.h +++ b/src/video_core/renderer_vulkan/vk_descriptor_pool.h @@ -7,11 +7,11 @@ #include <vector> #include "video_core/renderer_vulkan/vk_resource_pool.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class VKDescriptorPool; class VKScheduler; @@ -39,7 +39,7 @@ class VKDescriptorPool final { friend DescriptorAllocator; public: - explicit VKDescriptorPool(const VKDevice& device, VKScheduler& scheduler); + explicit VKDescriptorPool(const Device& device, VKScheduler& scheduler); ~VKDescriptorPool(); VKDescriptorPool(const VKDescriptorPool&) = delete; @@ -50,7 +50,7 @@ private: vk::DescriptorSets AllocateDescriptors(VkDescriptorSetLayout layout, std::size_t count); - const VKDevice& device; + const Device& device; MasterSemaphore& master_semaphore; std::vector<vk::DescriptorPool> pools; diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.cpp b/src/video_core/renderer_vulkan/vk_fence_manager.cpp index 5babbdd0b..4c5bc0aa1 100644 --- a/src/video_core/renderer_vulkan/vk_fence_manager.cpp +++ b/src/video_core/renderer_vulkan/vk_fence_manager.cpp @@ -6,20 +6,21 @@ #include <thread> #include "video_core/renderer_vulkan/vk_buffer_cache.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_fence_manager.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_texture_cache.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -InnerFence::InnerFence(const VKDevice& device, VKScheduler& scheduler, u32 payload, bool is_stubbed) - : VideoCommon::FenceBase(payload, is_stubbed), device{device}, scheduler{scheduler} {} +InnerFence::InnerFence(const Device& device_, VKScheduler& scheduler_, u32 payload_, + bool is_stubbed_) + : FenceBase{payload_, is_stubbed_}, device{device_}, scheduler{scheduler_} {} -InnerFence::InnerFence(const VKDevice& device, VKScheduler& scheduler, GPUVAddr address, - u32 payload, bool is_stubbed) - : VideoCommon::FenceBase(address, payload, is_stubbed), device{device}, scheduler{scheduler} {} +InnerFence::InnerFence(const Device& device_, VKScheduler& scheduler_, GPUVAddr address_, + u32 payload_, bool is_stubbed_) + : FenceBase{address_, payload_, is_stubbed_}, device{device_}, scheduler{scheduler_} {} InnerFence::~InnerFence() = default; @@ -71,11 +72,11 @@ bool InnerFence::IsEventSignalled() const { } } -VKFenceManager::VKFenceManager(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - Tegra::MemoryManager& memory_manager, VKTextureCache& texture_cache, - VKBufferCache& buffer_cache, VKQueryCache& query_cache, - const VKDevice& device_, VKScheduler& scheduler_) - : GenericFenceManager(rasterizer, gpu, texture_cache, buffer_cache, query_cache), +VKFenceManager::VKFenceManager(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, + Tegra::MemoryManager& memory_manager_, TextureCache& texture_cache_, + VKBufferCache& buffer_cache_, VKQueryCache& query_cache_, + const Device& device_, VKScheduler& scheduler_) + : GenericFenceManager{rasterizer_, gpu_, texture_cache_, buffer_cache_, query_cache_}, device{device_}, scheduler{scheduler_} {} Fence VKFenceManager::CreateFence(u32 value, bool is_stubbed) { diff --git a/src/video_core/renderer_vulkan/vk_fence_manager.h b/src/video_core/renderer_vulkan/vk_fence_manager.h index 1547d6d30..6b51e4587 100644 --- a/src/video_core/renderer_vulkan/vk_fence_manager.h +++ b/src/video_core/renderer_vulkan/vk_fence_manager.h @@ -8,7 +8,8 @@ #include "video_core/fence_manager.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/renderer_vulkan/vk_texture_cache.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Core { class System; @@ -20,18 +21,17 @@ class RasterizerInterface; namespace Vulkan { +class Device; class VKBufferCache; -class VKDevice; class VKQueryCache; class VKScheduler; -class VKTextureCache; class InnerFence : public VideoCommon::FenceBase { public: - explicit InnerFence(const VKDevice& device, VKScheduler& scheduler, u32 payload, - bool is_stubbed); - explicit InnerFence(const VKDevice& device, VKScheduler& scheduler, GPUVAddr address, - u32 payload, bool is_stubbed); + explicit InnerFence(const Device& device_, VKScheduler& scheduler_, u32 payload_, + bool is_stubbed_); + explicit InnerFence(const Device& device_, VKScheduler& scheduler_, GPUVAddr address_, + u32 payload_, bool is_stubbed_); ~InnerFence(); void Queue(); @@ -43,7 +43,7 @@ public: private: bool IsEventSignalled() const; - const VKDevice& device; + const Device& device; VKScheduler& scheduler; vk::Event event; u64 ticks = 0; @@ -51,14 +51,14 @@ private: using Fence = std::shared_ptr<InnerFence>; using GenericFenceManager = - VideoCommon::FenceManager<Fence, VKTextureCache, VKBufferCache, VKQueryCache>; + VideoCommon::FenceManager<Fence, TextureCache, VKBufferCache, VKQueryCache>; class VKFenceManager final : public GenericFenceManager { public: - explicit VKFenceManager(VideoCore::RasterizerInterface& rasterizer, Tegra::GPU& gpu, - Tegra::MemoryManager& memory_manager, VKTextureCache& texture_cache, - VKBufferCache& buffer_cache, VKQueryCache& query_cache, - const VKDevice& device, VKScheduler& scheduler); + explicit VKFenceManager(VideoCore::RasterizerInterface& rasterizer_, Tegra::GPU& gpu_, + Tegra::MemoryManager& memory_manager_, TextureCache& texture_cache_, + VKBufferCache& buffer_cache_, VKQueryCache& query_cache_, + const Device& device_, VKScheduler& scheduler_); protected: Fence CreateFence(u32 value, bool is_stubbed) override; @@ -68,7 +68,7 @@ protected: void WaitFence(Fence& fence) override; private: - const VKDevice& device; + const Device& device; VKScheduler& scheduler; }; diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp index a4b9e7ef5..a5214d0bc 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp @@ -12,13 +12,12 @@ #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_graphics_pipeline.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" -#include "video_core/renderer_vulkan/vk_renderpass_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -69,23 +68,45 @@ VkViewportSwizzleNV UnpackViewportSwizzle(u16 swizzle) { }; } +VkSampleCountFlagBits ConvertMsaaMode(Tegra::Texture::MsaaMode msaa_mode) { + switch (msaa_mode) { + case Tegra::Texture::MsaaMode::Msaa1x1: + return VK_SAMPLE_COUNT_1_BIT; + case Tegra::Texture::MsaaMode::Msaa2x1: + case Tegra::Texture::MsaaMode::Msaa2x1_D3D: + return VK_SAMPLE_COUNT_2_BIT; + case Tegra::Texture::MsaaMode::Msaa2x2: + case Tegra::Texture::MsaaMode::Msaa2x2_VC4: + case Tegra::Texture::MsaaMode::Msaa2x2_VC12: + return VK_SAMPLE_COUNT_4_BIT; + case Tegra::Texture::MsaaMode::Msaa4x2: + case Tegra::Texture::MsaaMode::Msaa4x2_D3D: + case Tegra::Texture::MsaaMode::Msaa4x2_VC8: + case Tegra::Texture::MsaaMode::Msaa4x2_VC24: + return VK_SAMPLE_COUNT_8_BIT; + case Tegra::Texture::MsaaMode::Msaa4x4: + return VK_SAMPLE_COUNT_16_BIT; + default: + UNREACHABLE_MSG("Invalid msaa_mode={}", static_cast<int>(msaa_mode)); + return VK_SAMPLE_COUNT_1_BIT; + } +} + } // Anonymous namespace -VKGraphicsPipeline::VKGraphicsPipeline(const VKDevice& device, VKScheduler& scheduler, - VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - VKRenderPassCache& renderpass_cache, +VKGraphicsPipeline::VKGraphicsPipeline(const Device& device_, VKScheduler& scheduler_, + VKDescriptorPool& descriptor_pool_, + VKUpdateDescriptorQueue& update_descriptor_queue_, const GraphicsPipelineCacheKey& key, vk::Span<VkDescriptorSetLayoutBinding> bindings, - const SPIRVProgram& program) - : device{device}, scheduler{scheduler}, cache_key{key}, hash{cache_key.Hash()}, + const SPIRVProgram& program, u32 num_color_buffers) + : device{device_}, scheduler{scheduler_}, cache_key{key}, hash{cache_key.Hash()}, descriptor_set_layout{CreateDescriptorSetLayout(bindings)}, - descriptor_allocator{descriptor_pool, *descriptor_set_layout}, - update_descriptor_queue{update_descriptor_queue}, layout{CreatePipelineLayout()}, - descriptor_template{CreateDescriptorUpdateTemplate(program)}, modules{CreateShaderModules( - program)}, - renderpass{renderpass_cache.GetRenderPass(cache_key.renderpass_params)}, - pipeline{CreatePipeline(cache_key.renderpass_params, program)} {} + descriptor_allocator{descriptor_pool_, *descriptor_set_layout}, + update_descriptor_queue{update_descriptor_queue_}, layout{CreatePipelineLayout()}, + descriptor_template{CreateDescriptorUpdateTemplate(program)}, + modules(CreateShaderModules(program)), + pipeline(CreatePipeline(program, cache_key.renderpass, num_color_buffers)) {} VKGraphicsPipeline::~VKGraphicsPipeline() = default; @@ -159,10 +180,11 @@ std::vector<vk::ShaderModule> VKGraphicsPipeline::CreateShaderModules( .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, .pNext = nullptr, .flags = 0, + .codeSize = 0, }; - std::vector<vk::ShaderModule> modules; - modules.reserve(Maxwell::MaxShaderStage); + std::vector<vk::ShaderModule> shader_modules; + shader_modules.reserve(Maxwell::MaxShaderStage); for (std::size_t i = 0; i < Maxwell::MaxShaderStage; ++i) { const auto& stage = program[i]; if (!stage) { @@ -173,13 +195,14 @@ std::vector<vk::ShaderModule> VKGraphicsPipeline::CreateShaderModules( ci.codeSize = stage->code.size() * sizeof(u32); ci.pCode = stage->code.data(); - modules.push_back(device.GetLogical().CreateShaderModule(ci)); + shader_modules.push_back(device.GetLogical().CreateShaderModule(ci)); } - return modules; + return shader_modules; } -vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpass_params, - const SPIRVProgram& program) const { +vk::Pipeline VKGraphicsPipeline::CreatePipeline(const SPIRVProgram& program, + VkRenderPass renderpass, + u32 num_color_buffers) const { const auto& state = cache_key.fixed_state; const auto& viewport_swizzles = state.viewport_swizzles; @@ -189,11 +212,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa // state is ignored dynamic.raw1 = 0; dynamic.raw2 = 0; - for (FixedPipelineState::VertexBinding& binding : dynamic.vertex_bindings) { - // Enable all vertex bindings - binding.raw = 0; - binding.enabled.Assign(1); - } + dynamic.vertex_strides.fill(0); } else { dynamic = state.dynamic_state; } @@ -201,19 +220,16 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa std::vector<VkVertexInputBindingDescription> vertex_bindings; std::vector<VkVertexInputBindingDivisorDescriptionEXT> vertex_binding_divisors; for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { - const auto& binding = dynamic.vertex_bindings[index]; - if (!binding.enabled) { + if (state.attributes[index].binding_index_enabled == 0) { continue; } const bool instanced = state.binding_divisors[index] != 0; const auto rate = instanced ? VK_VERTEX_INPUT_RATE_INSTANCE : VK_VERTEX_INPUT_RATE_VERTEX; - vertex_bindings.push_back({ .binding = static_cast<u32>(index), - .stride = binding.stride, + .stride = dynamic.vertex_strides[index], .inputRate = rate, }); - if (instanced) { vertex_binding_divisors.push_back({ .binding = static_cast<u32>(index), @@ -229,7 +245,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa if (!attribute.enabled) { continue; } - if (input_attributes.find(static_cast<u32>(index)) == input_attributes.end()) { + if (!input_attributes.contains(static_cast<u32>(index))) { // Skip attributes not used by the vertex shaders. continue; } @@ -261,12 +277,12 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa vertex_input_ci.pNext = &input_divisor_ci; } - const auto input_assembly_topology = MaxwellToVK::PrimitiveTopology(device, dynamic.Topology()); + const auto input_assembly_topology = MaxwellToVK::PrimitiveTopology(device, state.topology); const VkPipelineInputAssemblyStateCreateInfo input_assembly_ci{ .sType = VK_STRUCTURE_TYPE_PIPELINE_INPUT_ASSEMBLY_STATE_CREATE_INFO, .pNext = nullptr, .flags = 0, - .topology = MaxwellToVK::PrimitiveTopology(device, dynamic.Topology()), + .topology = MaxwellToVK::PrimitiveTopology(device, state.topology), .primitiveRestartEnable = state.primitive_restart_enable != 0 && SupportsPrimitiveRestart(input_assembly_topology), }; @@ -289,8 +305,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa }; std::array<VkViewportSwizzleNV, Maxwell::NumViewports> swizzles; - std::transform(viewport_swizzles.begin(), viewport_swizzles.end(), swizzles.begin(), - UnpackViewportSwizzle); + std::ranges::transform(viewport_swizzles, swizzles.begin(), UnpackViewportSwizzle); VkPipelineViewportSwizzleStateCreateInfoNV swizzle_ci{ .sType = VK_STRUCTURE_TYPE_PIPELINE_VIEWPORT_SWIZZLE_STATE_CREATE_INFO_NV, .pNext = nullptr, @@ -325,7 +340,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa .sType = VK_STRUCTURE_TYPE_PIPELINE_MULTISAMPLE_STATE_CREATE_INFO, .pNext = nullptr, .flags = 0, - .rasterizationSamples = VK_SAMPLE_COUNT_1_BIT, + .rasterizationSamples = ConvertMsaaMode(state.msaa_mode), .sampleShadingEnable = VK_FALSE, .minSampleShading = 0.0f, .pSampleMask = nullptr, @@ -351,8 +366,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa }; std::array<VkPipelineColorBlendAttachmentState, Maxwell::NumRenderTargets> cb_attachments; - const auto num_attachments = static_cast<std::size_t>(renderpass_params.num_color_attachments); - for (std::size_t index = 0; index < num_attachments; ++index) { + for (std::size_t index = 0; index < num_color_buffers; ++index) { static constexpr std::array COMPONENT_TABLE{ VK_COLOR_COMPONENT_R_BIT, VK_COLOR_COMPONENT_G_BIT, @@ -386,8 +400,9 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa .flags = 0, .logicOpEnable = VK_FALSE, .logicOp = VK_LOGIC_OP_COPY, - .attachmentCount = static_cast<u32>(num_attachments), + .attachmentCount = num_color_buffers, .pAttachments = cb_attachments.data(), + .blendConstants = {}, }; std::vector dynamic_states{ @@ -400,7 +415,6 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa static constexpr std::array extended{ VK_DYNAMIC_STATE_CULL_MODE_EXT, VK_DYNAMIC_STATE_FRONT_FACE_EXT, - VK_DYNAMIC_STATE_PRIMITIVE_TOPOLOGY_EXT, VK_DYNAMIC_STATE_VERTEX_INPUT_BINDING_STRIDE_EXT, VK_DYNAMIC_STATE_DEPTH_TEST_ENABLE_EXT, VK_DYNAMIC_STATE_DEPTH_WRITE_ENABLE_EXT, @@ -446,8 +460,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa stage_ci.pNext = &subgroup_size_ci; } } - - const VkGraphicsPipelineCreateInfo ci{ + return device.GetLogical().CreateGraphicsPipeline(VkGraphicsPipelineCreateInfo{ .sType = VK_STRUCTURE_TYPE_GRAPHICS_PIPELINE_CREATE_INFO, .pNext = nullptr, .flags = 0, @@ -467,8 +480,7 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const RenderPassParams& renderpa .subpass = 0, .basePipelineHandle = nullptr, .basePipelineIndex = 0, - }; - return device.GetLogical().CreateGraphicsPipeline(ci); + }); } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h index 58aa35efd..8b6a98fe0 100644 --- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.h +++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.h @@ -8,20 +8,19 @@ #include <optional> #include <vector> +#include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/vk_renderpass_cache.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { using Maxwell = Tegra::Engines::Maxwell3D::Regs; struct GraphicsPipelineCacheKey { - RenderPassParams renderpass_params; - u32 padding; + VkRenderPass renderpass; std::array<GPUVAddr, Maxwell::MaxShaderProgram> shaders; FixedPipelineState fixed_state; @@ -34,16 +33,15 @@ struct GraphicsPipelineCacheKey { } std::size_t Size() const noexcept { - return sizeof(renderpass_params) + sizeof(padding) + sizeof(shaders) + fixed_state.Size(); + return sizeof(renderpass) + sizeof(shaders) + fixed_state.Size(); } }; static_assert(std::has_unique_object_representations_v<GraphicsPipelineCacheKey>); static_assert(std::is_trivially_copyable_v<GraphicsPipelineCacheKey>); static_assert(std::is_trivially_constructible_v<GraphicsPipelineCacheKey>); +class Device; class VKDescriptorPool; -class VKDevice; -class VKRenderPassCache; class VKScheduler; class VKUpdateDescriptorQueue; @@ -51,13 +49,12 @@ using SPIRVProgram = std::array<std::optional<SPIRVShader>, Maxwell::MaxShaderSt class VKGraphicsPipeline final { public: - explicit VKGraphicsPipeline(const VKDevice& device, VKScheduler& scheduler, + explicit VKGraphicsPipeline(const Device& device_, VKScheduler& scheduler_, VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - VKRenderPassCache& renderpass_cache, + VKUpdateDescriptorQueue& update_descriptor_queue_, const GraphicsPipelineCacheKey& key, vk::Span<VkDescriptorSetLayoutBinding> bindings, - const SPIRVProgram& program); + const SPIRVProgram& program, u32 num_color_buffers); ~VKGraphicsPipeline(); VkDescriptorSet CommitDescriptorSet(); @@ -70,10 +67,6 @@ public: return *layout; } - VkRenderPass GetRenderPass() const { - return renderpass; - } - GraphicsPipelineCacheKey GetCacheKey() const { return cache_key; } @@ -89,10 +82,10 @@ private: std::vector<vk::ShaderModule> CreateShaderModules(const SPIRVProgram& program) const; - vk::Pipeline CreatePipeline(const RenderPassParams& renderpass_params, - const SPIRVProgram& program) const; + vk::Pipeline CreatePipeline(const SPIRVProgram& program, VkRenderPass renderpass, + u32 num_color_buffers) const; - const VKDevice& device; + const Device& device; VKScheduler& scheduler; const GraphicsPipelineCacheKey cache_key; const u64 hash; @@ -104,7 +97,6 @@ private: vk::DescriptorUpdateTemplateKHR descriptor_template; std::vector<vk::ShaderModule> modules; - VkRenderPass renderpass; vk::Pipeline pipeline; }; diff --git a/src/video_core/renderer_vulkan/vk_image.cpp b/src/video_core/renderer_vulkan/vk_image.cpp deleted file mode 100644 index 1c418ea17..000000000 --- a/src/video_core/renderer_vulkan/vk_image.cpp +++ /dev/null @@ -1,135 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <memory> -#include <vector> - -#include "common/assert.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_image.h" -#include "video_core/renderer_vulkan/vk_scheduler.h" -#include "video_core/renderer_vulkan/wrapper.h" - -namespace Vulkan { - -VKImage::VKImage(const VKDevice& device, VKScheduler& scheduler, const VkImageCreateInfo& image_ci, - VkImageAspectFlags aspect_mask) - : device{device}, scheduler{scheduler}, format{image_ci.format}, aspect_mask{aspect_mask}, - image_num_layers{image_ci.arrayLayers}, image_num_levels{image_ci.mipLevels} { - UNIMPLEMENTED_IF_MSG(image_ci.queueFamilyIndexCount != 0, - "Queue family tracking is not implemented"); - - image = device.GetLogical().CreateImage(image_ci); - - const u32 num_ranges = image_num_layers * image_num_levels; - barriers.resize(num_ranges); - subrange_states.resize(num_ranges, {{}, image_ci.initialLayout}); -} - -VKImage::~VKImage() = default; - -void VKImage::Transition(u32 base_layer, u32 num_layers, u32 base_level, u32 num_levels, - VkPipelineStageFlags new_stage_mask, VkAccessFlags new_access, - VkImageLayout new_layout) { - if (!HasChanged(base_layer, num_layers, base_level, num_levels, new_access, new_layout)) { - return; - } - - std::size_t cursor = 0; - for (u32 layer_it = 0; layer_it < num_layers; ++layer_it) { - for (u32 level_it = 0; level_it < num_levels; ++level_it, ++cursor) { - const u32 layer = base_layer + layer_it; - const u32 level = base_level + level_it; - auto& state = GetSubrangeState(layer, level); - auto& barrier = barriers[cursor]; - barrier.sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER; - barrier.pNext = nullptr; - barrier.srcAccessMask = state.access; - barrier.dstAccessMask = new_access; - barrier.oldLayout = state.layout; - barrier.newLayout = new_layout; - barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; - barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; - barrier.image = *image; - barrier.subresourceRange.aspectMask = aspect_mask; - barrier.subresourceRange.baseMipLevel = level; - barrier.subresourceRange.levelCount = 1; - barrier.subresourceRange.baseArrayLayer = layer; - barrier.subresourceRange.layerCount = 1; - state.access = new_access; - state.layout = new_layout; - } - } - - scheduler.RequestOutsideRenderPassOperationContext(); - - scheduler.Record([barriers = barriers, cursor](vk::CommandBuffer cmdbuf) { - // TODO(Rodrigo): Implement a way to use the latest stage across subresources. - cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, - VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, 0, {}, {}, - vk::Span(barriers.data(), cursor)); - }); -} - -bool VKImage::HasChanged(u32 base_layer, u32 num_layers, u32 base_level, u32 num_levels, - VkAccessFlags new_access, VkImageLayout new_layout) noexcept { - const bool is_full_range = base_layer == 0 && num_layers == image_num_layers && - base_level == 0 && num_levels == image_num_levels; - if (!is_full_range) { - state_diverged = true; - } - - if (!state_diverged) { - auto& state = GetSubrangeState(0, 0); - if (state.access != new_access || state.layout != new_layout) { - return true; - } - } - - for (u32 layer_it = 0; layer_it < num_layers; ++layer_it) { - for (u32 level_it = 0; level_it < num_levels; ++level_it) { - const u32 layer = base_layer + layer_it; - const u32 level = base_level + level_it; - auto& state = GetSubrangeState(layer, level); - if (state.access != new_access || state.layout != new_layout) { - return true; - } - } - } - return false; -} - -void VKImage::CreatePresentView() { - // Image type has to be 2D to be presented. - present_view = device.GetLogical().CreateImageView({ - .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, - .pNext = nullptr, - .flags = 0, - .image = *image, - .viewType = VK_IMAGE_VIEW_TYPE_2D, - .format = format, - .components = - { - .r = VK_COMPONENT_SWIZZLE_IDENTITY, - .g = VK_COMPONENT_SWIZZLE_IDENTITY, - .b = VK_COMPONENT_SWIZZLE_IDENTITY, - .a = VK_COMPONENT_SWIZZLE_IDENTITY, - }, - .subresourceRange = - { - .aspectMask = aspect_mask, - .baseMipLevel = 0, - .levelCount = 1, - .baseArrayLayer = 0, - .layerCount = 1, - }, - }); -} - -VKImage::SubrangeState& VKImage::GetSubrangeState(u32 layer, u32 level) noexcept { - return subrange_states[static_cast<std::size_t>(layer * image_num_levels) + - static_cast<std::size_t>(level)]; -} - -} // namespace Vulkan
\ No newline at end of file diff --git a/src/video_core/renderer_vulkan/vk_image.h b/src/video_core/renderer_vulkan/vk_image.h deleted file mode 100644 index b4d7229e5..000000000 --- a/src/video_core/renderer_vulkan/vk_image.h +++ /dev/null @@ -1,84 +0,0 @@ -// Copyright 2018 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <memory> -#include <vector> - -#include "common/common_types.h" -#include "video_core/renderer_vulkan/wrapper.h" - -namespace Vulkan { - -class VKDevice; -class VKScheduler; - -class VKImage { -public: - explicit VKImage(const VKDevice& device, VKScheduler& scheduler, - const VkImageCreateInfo& image_ci, VkImageAspectFlags aspect_mask); - ~VKImage(); - - /// Records in the passed command buffer an image transition and updates the state of the image. - void Transition(u32 base_layer, u32 num_layers, u32 base_level, u32 num_levels, - VkPipelineStageFlags new_stage_mask, VkAccessFlags new_access, - VkImageLayout new_layout); - - /// Returns a view compatible with presentation, the image has to be 2D. - VkImageView GetPresentView() { - if (!present_view) { - CreatePresentView(); - } - return *present_view; - } - - /// Returns the Vulkan image handler. - const vk::Image& GetHandle() const { - return image; - } - - /// Returns the Vulkan format for this image. - VkFormat GetFormat() const { - return format; - } - - /// Returns the Vulkan aspect mask. - VkImageAspectFlags GetAspectMask() const { - return aspect_mask; - } - -private: - struct SubrangeState final { - VkAccessFlags access = 0; ///< Current access bits. - VkImageLayout layout = VK_IMAGE_LAYOUT_UNDEFINED; ///< Current image layout. - }; - - bool HasChanged(u32 base_layer, u32 num_layers, u32 base_level, u32 num_levels, - VkAccessFlags new_access, VkImageLayout new_layout) noexcept; - - /// Creates a presentation view. - void CreatePresentView(); - - /// Returns the subrange state for a layer and layer. - SubrangeState& GetSubrangeState(u32 layer, u32 level) noexcept; - - const VKDevice& device; ///< Device handler. - VKScheduler& scheduler; ///< Device scheduler. - - const VkFormat format; ///< Vulkan format. - const VkImageAspectFlags aspect_mask; ///< Vulkan aspect mask. - const u32 image_num_layers; ///< Number of layers. - const u32 image_num_levels; ///< Number of mipmap levels. - - vk::Image image; ///< Image handle. - vk::ImageView present_view; ///< Image view compatible with presentation. - - std::vector<VkImageMemoryBarrier> barriers; ///< Pool of barriers. - std::vector<SubrangeState> subrange_states; ///< Current subrange state. - - bool state_diverged = false; ///< True when subresources mismatch in layout. -}; - -} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_master_semaphore.cpp b/src/video_core/renderer_vulkan/vk_master_semaphore.cpp index ae26e558d..56ec5e380 100644 --- a/src/video_core/renderer_vulkan/vk_master_semaphore.cpp +++ b/src/video_core/renderer_vulkan/vk_master_semaphore.cpp @@ -6,15 +6,15 @@ #include <chrono> #include "core/settings.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_master_semaphore.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { using namespace std::chrono_literals; -MasterSemaphore::MasterSemaphore(const VKDevice& device) { +MasterSemaphore::MasterSemaphore(const Device& device) { static constexpr VkSemaphoreTypeCreateInfoKHR semaphore_type_ci{ .sType = VK_STRUCTURE_TYPE_SEMAPHORE_TYPE_CREATE_INFO_KHR, .pNext = nullptr, diff --git a/src/video_core/renderer_vulkan/vk_master_semaphore.h b/src/video_core/renderer_vulkan/vk_master_semaphore.h index 0e93706d7..f336f1862 100644 --- a/src/video_core/renderer_vulkan/vk_master_semaphore.h +++ b/src/video_core/renderer_vulkan/vk_master_semaphore.h @@ -8,15 +8,15 @@ #include <thread> #include "common/common_types.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class MasterSemaphore { public: - explicit MasterSemaphore(const VKDevice& device); + explicit MasterSemaphore(const Device& device); ~MasterSemaphore(); /// Returns the current logical tick. diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.cpp b/src/video_core/renderer_vulkan/vk_memory_manager.cpp index 24c8960ac..a6abd0eee 100644 --- a/src/video_core/renderer_vulkan/vk_memory_manager.cpp +++ b/src/video_core/renderer_vulkan/vk_memory_manager.cpp @@ -11,9 +11,9 @@ #include "common/assert.h" #include "common/common_types.h" #include "common/logging/log.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -29,10 +29,10 @@ u64 GetAllocationChunkSize(u64 required_size) { class VKMemoryAllocation final { public: - explicit VKMemoryAllocation(const VKDevice& device, vk::DeviceMemory memory, - VkMemoryPropertyFlags properties, u64 allocation_size, u32 type) - : device{device}, memory{std::move(memory)}, properties{properties}, - allocation_size{allocation_size}, shifted_type{ShiftType(type)} {} + explicit VKMemoryAllocation(const Device& device_, vk::DeviceMemory memory_, + VkMemoryPropertyFlags properties_, u64 allocation_size_, u32 type_) + : device{device_}, memory{std::move(memory_)}, properties{properties_}, + allocation_size{allocation_size_}, shifted_type{ShiftType(type_)} {} VKMemoryCommit Commit(VkDeviceSize commit_size, VkDeviceSize alignment) { auto found = TryFindFreeSection(free_iterator, allocation_size, @@ -104,7 +104,7 @@ private: return std::nullopt; } - const VKDevice& device; ///< Vulkan device. + const Device& device; ///< Vulkan device. const vk::DeviceMemory memory; ///< Vulkan memory allocation handler. const VkMemoryPropertyFlags properties; ///< Vulkan properties. const u64 allocation_size; ///< Size of this allocation. @@ -117,8 +117,8 @@ private: std::vector<const VKMemoryCommitImpl*> commits; }; -VKMemoryManager::VKMemoryManager(const VKDevice& device) - : device{device}, properties{device.GetPhysical().GetMemoryProperties()} {} +VKMemoryManager::VKMemoryManager(const Device& device_) + : device{device_}, properties{device_.GetPhysical().GetMemoryProperties()} {} VKMemoryManager::~VKMemoryManager() = default; @@ -207,16 +207,16 @@ VKMemoryCommit VKMemoryManager::TryAllocCommit(const VkMemoryRequirements& requi return {}; } -VKMemoryCommitImpl::VKMemoryCommitImpl(const VKDevice& device, VKMemoryAllocation* allocation, - const vk::DeviceMemory& memory, u64 begin, u64 end) - : device{device}, memory{memory}, interval{begin, end}, allocation{allocation} {} +VKMemoryCommitImpl::VKMemoryCommitImpl(const Device& device_, VKMemoryAllocation* allocation_, + const vk::DeviceMemory& memory_, u64 begin_, u64 end_) + : device{device_}, memory{memory_}, interval{begin_, end_}, allocation{allocation_} {} VKMemoryCommitImpl::~VKMemoryCommitImpl() { allocation->Free(this); } MemoryMap VKMemoryCommitImpl::Map(u64 size, u64 offset_) const { - return MemoryMap{this, memory.Map(interval.first + offset_, size)}; + return MemoryMap(this, std::span<u8>(memory.Map(interval.first + offset_, size), size)); } void VKMemoryCommitImpl::Unmap() const { diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.h b/src/video_core/renderer_vulkan/vk_memory_manager.h index 1af88e3d4..2452bca4e 100644 --- a/src/video_core/renderer_vulkan/vk_memory_manager.h +++ b/src/video_core/renderer_vulkan/vk_memory_manager.h @@ -5,15 +5,16 @@ #pragma once #include <memory> +#include <span> #include <utility> #include <vector> #include "common/common_types.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { +class Device; class MemoryMap; -class VKDevice; class VKMemoryAllocation; class VKMemoryCommitImpl; @@ -21,7 +22,7 @@ using VKMemoryCommit = std::unique_ptr<VKMemoryCommitImpl>; class VKMemoryManager final { public: - explicit VKMemoryManager(const VKDevice& device); + explicit VKMemoryManager(const Device& device_); VKMemoryManager(const VKMemoryManager&) = delete; ~VKMemoryManager(); @@ -48,7 +49,7 @@ private: VKMemoryCommit TryAllocCommit(const VkMemoryRequirements& requirements, VkMemoryPropertyFlags wanted_properties); - const VKDevice& device; ///< Device handler. + const Device& device; ///< Device handler. const VkPhysicalDeviceMemoryProperties properties; ///< Physical device properties. std::vector<std::unique_ptr<VKMemoryAllocation>> allocations; ///< Current allocations. }; @@ -58,8 +59,8 @@ class VKMemoryCommitImpl final { friend MemoryMap; public: - explicit VKMemoryCommitImpl(const VKDevice& device, VKMemoryAllocation* allocation, - const vk::DeviceMemory& memory, u64 begin, u64 end); + explicit VKMemoryCommitImpl(const Device& device_, VKMemoryAllocation* allocation_, + const vk::DeviceMemory& memory_, u64 begin_, u64 end_); ~VKMemoryCommitImpl(); /// Maps a memory region and returns a pointer to it. @@ -84,7 +85,7 @@ private: /// Unmaps memory. void Unmap() const; - const VKDevice& device; ///< Vulkan device. + const Device& device; ///< Vulkan device. const vk::DeviceMemory& memory; ///< Vulkan device memory handler. std::pair<u64, u64> interval{}; ///< Interval where the commit exists. VKMemoryAllocation* allocation{}; ///< Pointer to the large memory allocation. @@ -93,8 +94,8 @@ private: /// Holds ownership of a memory map. class MemoryMap final { public: - explicit MemoryMap(const VKMemoryCommitImpl* commit, u8* address) - : commit{commit}, address{address} {} + explicit MemoryMap(const VKMemoryCommitImpl* commit_, std::span<u8> span_) + : commit{commit_}, span{span_} {} ~MemoryMap() { if (commit) { @@ -108,19 +109,24 @@ public: commit = nullptr; } + /// Returns a span to the memory map. + [[nodiscard]] std::span<u8> Span() const noexcept { + return span; + } + /// Returns the address of the memory map. - u8* GetAddress() const { - return address; + [[nodiscard]] u8* Address() const noexcept { + return span.data(); } /// Returns the address of the memory map; - operator u8*() const { - return address; + [[nodiscard]] operator u8*() const noexcept { + return span.data(); } private: const VKMemoryCommitImpl* commit{}; ///< Mapped memory commit. - u8* address{}; ///< Address to the mapped memory. + std::span<u8> span; ///< Address to the mapped memory. }; } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp index 5c038f4bc..02282e36f 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.cpp @@ -7,6 +7,8 @@ #include <memory> #include <vector> +#include "common/bit_cast.h" +#include "common/cityhash.h" #include "common/microprofile.h" #include "core/core.h" #include "core/memory.h" @@ -17,18 +19,17 @@ #include "video_core/renderer_vulkan/maxwell_to_vk.h" #include "video_core/renderer_vulkan/vk_compute_pipeline.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_graphics_pipeline.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" -#include "video_core/renderer_vulkan/vk_renderpass_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" -#include "video_core/renderer_vulkan/wrapper.h" #include "video_core/shader/compiler_settings.h" #include "video_core/shader/memory_util.h" #include "video_core/shader_cache.h" #include "video_core/shader_notify.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -51,7 +52,9 @@ constexpr VkDescriptorType STORAGE_TEXEL_BUFFER = VK_DESCRIPTOR_TYPE_STORAGE_TEX constexpr VkDescriptorType STORAGE_IMAGE = VK_DESCRIPTOR_TYPE_STORAGE_IMAGE; constexpr VideoCommon::Shader::CompilerSettings compiler_settings{ - VideoCommon::Shader::CompileDepth::FullDecompile}; + .depth = VideoCommon::Shader::CompileDepth::FullDecompile, + .disable_else_derivation = true, +}; constexpr std::size_t GetStageFromProgram(std::size_t program) { return program == 0 ? 0 : program - 1; @@ -74,7 +77,7 @@ ShaderType GetShaderType(Maxwell::ShaderProgram program) { case Maxwell::ShaderProgram::Fragment: return ShaderType::Fragment; default: - UNIMPLEMENTED_MSG("program={}", static_cast<u32>(program)); + UNIMPLEMENTED_MSG("program={}", program); return ShaderType::Vertex; } } @@ -135,26 +138,24 @@ bool ComputePipelineCacheKey::operator==(const ComputePipelineCacheKey& rhs) con return std::memcmp(&rhs, this, sizeof *this) == 0; } -Shader::Shader(Tegra::Engines::ConstBufferEngineInterface& engine, Tegra::Engines::ShaderType stage, - GPUVAddr gpu_addr_, VAddr cpu_addr, VideoCommon::Shader::ProgramCode program_code_, - u32 main_offset) - : gpu_addr(gpu_addr_), program_code(std::move(program_code_)), registry(stage, engine), - shader_ir(program_code, main_offset, compiler_settings, registry), +Shader::Shader(Tegra::Engines::ConstBufferEngineInterface& engine_, ShaderType stage_, + GPUVAddr gpu_addr_, VAddr cpu_addr_, ProgramCode program_code_, u32 main_offset_) + : gpu_addr(gpu_addr_), program_code(std::move(program_code_)), registry(stage_, engine_), + shader_ir(program_code, main_offset_, compiler_settings, registry), entries(GenerateShaderEntries(shader_ir)) {} Shader::~Shader() = default; -VKPipelineCache::VKPipelineCache(RasterizerVulkan& rasterizer, Tegra::GPU& gpu_, +VKPipelineCache::VKPipelineCache(RasterizerVulkan& rasterizer_, Tegra::GPU& gpu_, Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::Engines::KeplerCompute& kepler_compute_, - Tegra::MemoryManager& gpu_memory_, const VKDevice& device_, + Tegra::MemoryManager& gpu_memory_, const Device& device_, VKScheduler& scheduler_, VKDescriptorPool& descriptor_pool_, - VKUpdateDescriptorQueue& update_descriptor_queue_, - VKRenderPassCache& renderpass_cache_) - : VideoCommon::ShaderCache<Shader>{rasterizer}, gpu{gpu_}, maxwell3d{maxwell3d_}, + VKUpdateDescriptorQueue& update_descriptor_queue_) + : VideoCommon::ShaderCache<Shader>{rasterizer_}, gpu{gpu_}, maxwell3d{maxwell3d_}, kepler_compute{kepler_compute_}, gpu_memory{gpu_memory_}, device{device_}, - scheduler{scheduler_}, descriptor_pool{descriptor_pool_}, - update_descriptor_queue{update_descriptor_queue_}, renderpass_cache{renderpass_cache_} {} + scheduler{scheduler_}, descriptor_pool{descriptor_pool_}, update_descriptor_queue{ + update_descriptor_queue_} {} VKPipelineCache::~VKPipelineCache() = default; @@ -199,7 +200,8 @@ std::array<Shader*, Maxwell::MaxShaderProgram> VKPipelineCache::GetShaders() { } VKGraphicsPipeline* VKPipelineCache::GetGraphicsPipeline( - const GraphicsPipelineCacheKey& key, VideoCommon::Shader::AsyncShaders& async_shaders) { + const GraphicsPipelineCacheKey& key, u32 num_color_buffers, + VideoCommon::Shader::AsyncShaders& async_shaders) { MICROPROFILE_SCOPE(Vulkan_PipelineCache); if (last_graphics_pipeline && last_graphics_key == key) { @@ -215,8 +217,8 @@ VKGraphicsPipeline* VKPipelineCache::GetGraphicsPipeline( LOG_INFO(Render_Vulkan, "Compile 0x{:016X}", key.Hash()); const auto [program, bindings] = DecompileShaders(key.fixed_state); async_shaders.QueueVulkanShader(this, device, scheduler, descriptor_pool, - update_descriptor_queue, renderpass_cache, bindings, - program, key); + update_descriptor_queue, bindings, program, key, + num_color_buffers); } last_graphics_pipeline = pair->second.get(); return last_graphics_pipeline; @@ -229,8 +231,8 @@ VKGraphicsPipeline* VKPipelineCache::GetGraphicsPipeline( LOG_INFO(Render_Vulkan, "Compile 0x{:016X}", key.Hash()); const auto [program, bindings] = DecompileShaders(key.fixed_state); entry = std::make_unique<VKGraphicsPipeline>(device, scheduler, descriptor_pool, - update_descriptor_queue, renderpass_cache, key, - bindings, program); + update_descriptor_queue, key, bindings, + program, num_color_buffers); gpu.ShaderNotify().MarkShaderComplete(); } last_graphics_pipeline = entry.get(); @@ -331,8 +333,7 @@ void VKPipelineCache::OnShaderRemoval(Shader* shader) { std::pair<SPIRVProgram, std::vector<VkDescriptorSetLayoutBinding>> VKPipelineCache::DecompileShaders(const FixedPipelineState& fixed_state) { Specialization specialization; - if (fixed_state.dynamic_state.Topology() == Maxwell::PrimitiveTopology::Points || - device.IsExtExtendedDynamicStateSupported()) { + if (fixed_state.topology == Maxwell::PrimitiveTopology::Points) { float point_size; std::memcpy(&point_size, &fixed_state.point_size, sizeof(float)); specialization.point_size = point_size; @@ -344,6 +345,12 @@ VKPipelineCache::DecompileShaders(const FixedPipelineState& fixed_state) { specialization.attribute_types[i] = attribute.Type(); } specialization.ndc_minus_one_to_one = fixed_state.ndc_minus_one_to_one; + specialization.early_fragment_tests = fixed_state.early_z; + + // Alpha test + specialization.alpha_test_func = + FixedPipelineState::UnpackComparisonOp(fixed_state.alpha_test_func.Value()); + specialization.alpha_test_ref = Common::BitCast<float>(fixed_state.alpha_test_ref); SPIRVProgram program; std::vector<VkDescriptorSetLayoutBinding> bindings; diff --git a/src/video_core/renderer_vulkan/vk_pipeline_cache.h b/src/video_core/renderer_vulkan/vk_pipeline_cache.h index e558e6658..89d635a3d 100644 --- a/src/video_core/renderer_vulkan/vk_pipeline_cache.h +++ b/src/video_core/renderer_vulkan/vk_pipeline_cache.h @@ -19,14 +19,13 @@ #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_graphics_pipeline.h" -#include "video_core/renderer_vulkan/vk_renderpass_cache.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" -#include "video_core/renderer_vulkan/wrapper.h" #include "video_core/shader/async_shaders.h" #include "video_core/shader/memory_util.h" #include "video_core/shader/registry.h" #include "video_core/shader/shader_ir.h" #include "video_core/shader_cache.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Core { class System; @@ -34,10 +33,10 @@ class System; namespace Vulkan { +class Device; class RasterizerVulkan; class VKComputePipeline; class VKDescriptorPool; -class VKDevice; class VKScheduler; class VKUpdateDescriptorQueue; @@ -84,9 +83,9 @@ namespace Vulkan { class Shader { public: - explicit Shader(Tegra::Engines::ConstBufferEngineInterface& engine, - Tegra::Engines::ShaderType stage, GPUVAddr gpu_addr, VAddr cpu_addr, - VideoCommon::Shader::ProgramCode program_code, u32 main_offset); + explicit Shader(Tegra::Engines::ConstBufferEngineInterface& engine_, + Tegra::Engines::ShaderType stage_, GPUVAddr gpu_addr, VAddr cpu_addr_, + VideoCommon::Shader::ProgramCode program_code, u32 main_offset_); ~Shader(); GPUVAddr GetGpuAddr() const { @@ -122,15 +121,15 @@ public: explicit VKPipelineCache(RasterizerVulkan& rasterizer, Tegra::GPU& gpu, Tegra::Engines::Maxwell3D& maxwell3d, Tegra::Engines::KeplerCompute& kepler_compute, - Tegra::MemoryManager& gpu_memory, const VKDevice& device, + Tegra::MemoryManager& gpu_memory, const Device& device, VKScheduler& scheduler, VKDescriptorPool& descriptor_pool, - VKUpdateDescriptorQueue& update_descriptor_queue, - VKRenderPassCache& renderpass_cache); + VKUpdateDescriptorQueue& update_descriptor_queue); ~VKPipelineCache() override; std::array<Shader*, Maxwell::MaxShaderProgram> GetShaders(); VKGraphicsPipeline* GetGraphicsPipeline(const GraphicsPipelineCacheKey& key, + u32 num_color_buffers, VideoCommon::Shader::AsyncShaders& async_shaders); VKComputePipeline& GetComputePipeline(const ComputePipelineCacheKey& key); @@ -149,11 +148,10 @@ private: Tegra::Engines::KeplerCompute& kepler_compute; Tegra::MemoryManager& gpu_memory; - const VKDevice& device; + const Device& device; VKScheduler& scheduler; VKDescriptorPool& descriptor_pool; VKUpdateDescriptorQueue& update_descriptor_queue; - VKRenderPassCache& renderpass_cache; std::unique_ptr<Shader> null_shader; std::unique_ptr<Shader> null_kernel; diff --git a/src/video_core/renderer_vulkan/vk_query_cache.cpp b/src/video_core/renderer_vulkan/vk_query_cache.cpp index ee2d871e3..7cadd5147 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_query_cache.cpp @@ -7,11 +7,11 @@ #include <utility> #include <vector> -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_query_cache.h" #include "video_core/renderer_vulkan/vk_resource_pool.h" #include "video_core/renderer_vulkan/vk_scheduler.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -27,7 +27,7 @@ constexpr VkQueryType GetTarget(QueryType type) { } // Anonymous namespace -QueryPool::QueryPool(const VKDevice& device_, VKScheduler& scheduler, QueryType type_) +QueryPool::QueryPool(const Device& device_, VKScheduler& scheduler, QueryType type_) : ResourcePool{scheduler.GetMasterSemaphore(), GROW_STEP}, device{device_}, type{type_} {} QueryPool::~QueryPool() = default; @@ -66,15 +66,13 @@ void QueryPool::Reserve(std::pair<VkQueryPool, u32> query) { usage[pool_index * GROW_STEP + static_cast<std::ptrdiff_t>(query.second)] = false; } -VKQueryCache::VKQueryCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, - const VKDevice& device, VKScheduler& scheduler) - : VideoCommon::QueryCacheBase<VKQueryCache, CachedQuery, CounterStream, - HostCounter>{rasterizer, maxwell3d, gpu_memory}, - device{device}, scheduler{scheduler}, query_pools{ - QueryPool{device, scheduler, - QueryType::SamplesPassed}, - } {} +VKQueryCache::VKQueryCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, + const Device& device_, VKScheduler& scheduler_) + : QueryCacheBase{rasterizer_, maxwell3d_, gpu_memory_}, device{device_}, scheduler{scheduler_}, + query_pools{ + QueryPool{device_, scheduler_, QueryType::SamplesPassed}, + } {} VKQueryCache::~VKQueryCache() { // TODO(Rodrigo): This is a hack to destroy all HostCounter instances before the base class @@ -95,12 +93,12 @@ void VKQueryCache::Reserve(QueryType type, std::pair<VkQueryPool, u32> query) { query_pools[static_cast<std::size_t>(type)].Reserve(query); } -HostCounter::HostCounter(VKQueryCache& cache, std::shared_ptr<HostCounter> dependency, - QueryType type) - : VideoCommon::HostCounterBase<VKQueryCache, HostCounter>{std::move(dependency)}, cache{cache}, - type{type}, query{cache.AllocateQuery(type)}, tick{cache.Scheduler().CurrentTick()} { - const vk::Device* logical = &cache.Device().GetLogical(); - cache.Scheduler().Record([logical, query = query](vk::CommandBuffer cmdbuf) { +HostCounter::HostCounter(VKQueryCache& cache_, std::shared_ptr<HostCounter> dependency_, + QueryType type_) + : HostCounterBase{std::move(dependency_)}, cache{cache_}, type{type_}, + query{cache_.AllocateQuery(type_)}, tick{cache_.GetScheduler().CurrentTick()} { + const vk::Device* logical = &cache.GetDevice().GetLogical(); + cache.GetScheduler().Record([logical, query = query](vk::CommandBuffer cmdbuf) { logical->ResetQueryPoolEXT(query.first, query.second, 1); cmdbuf.BeginQuery(query.first, query.second, VK_QUERY_CONTROL_PRECISE_BIT); }); @@ -111,26 +109,28 @@ HostCounter::~HostCounter() { } void HostCounter::EndQuery() { - cache.Scheduler().Record( + cache.GetScheduler().Record( [query = query](vk::CommandBuffer cmdbuf) { cmdbuf.EndQuery(query.first, query.second); }); } u64 HostCounter::BlockingQuery() const { - if (tick >= cache.Scheduler().CurrentTick()) { - cache.Scheduler().Flush(); + if (tick >= cache.GetScheduler().CurrentTick()) { + cache.GetScheduler().Flush(); } + u64 data; - const VkResult result = cache.Device().GetLogical().GetQueryResults( + const VkResult query_result = cache.GetDevice().GetLogical().GetQueryResults( query.first, query.second, 1, sizeof(data), &data, sizeof(data), VK_QUERY_RESULT_64_BIT | VK_QUERY_RESULT_WAIT_BIT); - switch (result) { + + switch (query_result) { case VK_SUCCESS: return data; case VK_ERROR_DEVICE_LOST: - cache.Device().ReportLoss(); + cache.GetDevice().ReportLoss(); [[fallthrough]]; default: - throw vk::Exception(result); + throw vk::Exception(query_result); } } diff --git a/src/video_core/renderer_vulkan/vk_query_cache.h b/src/video_core/renderer_vulkan/vk_query_cache.h index 2e57fb75d..7190946b9 100644 --- a/src/video_core/renderer_vulkan/vk_query_cache.h +++ b/src/video_core/renderer_vulkan/vk_query_cache.h @@ -12,7 +12,7 @@ #include "common/common_types.h" #include "video_core/query_cache.h" #include "video_core/renderer_vulkan/vk_resource_pool.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace VideoCore { class RasterizerInterface; @@ -21,8 +21,8 @@ class RasterizerInterface; namespace Vulkan { class CachedQuery; +class Device; class HostCounter; -class VKDevice; class VKQueryCache; class VKScheduler; @@ -30,7 +30,7 @@ using CounterStream = VideoCommon::CounterStreamBase<VKQueryCache, HostCounter>; class QueryPool final : public ResourcePool { public: - explicit QueryPool(const VKDevice& device, VKScheduler& scheduler, VideoCore::QueryType type); + explicit QueryPool(const Device& device, VKScheduler& scheduler, VideoCore::QueryType type); ~QueryPool() override; std::pair<VkQueryPool, u32> Commit(); @@ -43,7 +43,7 @@ protected: private: static constexpr std::size_t GROW_STEP = 512; - const VKDevice& device; + const Device& device; const VideoCore::QueryType type; std::vector<vk::QueryPool> pools; @@ -53,33 +53,33 @@ private: class VKQueryCache final : public VideoCommon::QueryCacheBase<VKQueryCache, CachedQuery, CounterStream, HostCounter> { public: - explicit VKQueryCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, - const VKDevice& device, VKScheduler& scheduler); + explicit VKQueryCache(VideoCore::RasterizerInterface& rasterizer_, + Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, + const Device& device_, VKScheduler& scheduler_); ~VKQueryCache(); std::pair<VkQueryPool, u32> AllocateQuery(VideoCore::QueryType type); void Reserve(VideoCore::QueryType type, std::pair<VkQueryPool, u32> query); - const VKDevice& Device() const noexcept { + const Device& GetDevice() const noexcept { return device; } - VKScheduler& Scheduler() const noexcept { + VKScheduler& GetScheduler() const noexcept { return scheduler; } private: - const VKDevice& device; + const Device& device; VKScheduler& scheduler; std::array<QueryPool, VideoCore::NumQueryTypes> query_pools; }; class HostCounter final : public VideoCommon::HostCounterBase<VKQueryCache, HostCounter> { public: - explicit HostCounter(VKQueryCache& cache, std::shared_ptr<HostCounter> dependency, - VideoCore::QueryType type); + explicit HostCounter(VKQueryCache& cache_, std::shared_ptr<HostCounter> dependency_, + VideoCore::QueryType type_); ~HostCounter(); void EndQuery(); @@ -95,8 +95,8 @@ private: class CachedQuery : public VideoCommon::CachedQueryBase<HostCounter> { public: - explicit CachedQuery(VKQueryCache&, VideoCore::QueryType, VAddr cpu_addr, u8* host_ptr) - : VideoCommon::CachedQueryBase<HostCounter>{cpu_addr, host_ptr} {} + explicit CachedQuery(VKQueryCache&, VideoCore::QueryType, VAddr cpu_addr_, u8* host_ptr_) + : CachedQueryBase{cpu_addr_, host_ptr_} {} }; } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp index f3c2483c8..93fbea510 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp +++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp @@ -19,6 +19,7 @@ #include "core/settings.h" #include "video_core/engines/kepler_compute.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/renderer_vulkan/blit_image.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" #include "video_core/renderer_vulkan/renderer_vulkan.h" @@ -26,23 +27,24 @@ #include "video_core/renderer_vulkan/vk_compute_pass.h" #include "video_core/renderer_vulkan/vk_compute_pipeline.h" #include "video_core/renderer_vulkan/vk_descriptor_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_graphics_pipeline.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" -#include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_sampler_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/vk_state_tracker.h" #include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" -#include "video_core/renderer_vulkan/wrapper.h" #include "video_core/shader_cache.h" +#include "video_core/texture_cache/texture_cache.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { using Maxwell = Tegra::Engines::Maxwell3D::Regs; +using VideoCommon::ImageViewId; +using VideoCommon::ImageViewType; MICROPROFILE_DEFINE(Vulkan_WaitForWorker, "Vulkan", "Wait for worker", MP_RGB(255, 192, 192)); MICROPROFILE_DEFINE(Vulkan_Drawing, "Vulkan", "Record drawing", MP_RGB(192, 128, 128)); @@ -58,9 +60,9 @@ MICROPROFILE_DEFINE(Vulkan_PipelineCache, "Vulkan", "Pipeline cache", MP_RGB(192 namespace { -constexpr auto ComputeShaderIndex = static_cast<std::size_t>(Tegra::Engines::ShaderType::Compute); +constexpr auto COMPUTE_SHADER_INDEX = static_cast<size_t>(Tegra::Engines::ShaderType::Compute); -VkViewport GetViewportState(const VKDevice& device, const Maxwell& regs, std::size_t index) { +VkViewport GetViewportState(const Device& device, const Maxwell& regs, size_t index) { const auto& src = regs.viewport_transform[index]; const float width = src.scale_x * 2.0f; const float height = src.scale_y * 2.0f; @@ -83,7 +85,7 @@ VkViewport GetViewportState(const VKDevice& device, const Maxwell& regs, std::si return viewport; } -VkRect2D GetScissorState(const Maxwell& regs, std::size_t index) { +VkRect2D GetScissorState(const Maxwell& regs, size_t index) { const auto& src = regs.scissor_test[index]; VkRect2D scissor; if (src.enable) { @@ -103,98 +105,122 @@ VkRect2D GetScissorState(const Maxwell& regs, std::size_t index) { std::array<GPUVAddr, Maxwell::MaxShaderProgram> GetShaderAddresses( const std::array<Shader*, Maxwell::MaxShaderProgram>& shaders) { std::array<GPUVAddr, Maxwell::MaxShaderProgram> addresses; - for (std::size_t i = 0; i < std::size(addresses); ++i) { + for (size_t i = 0; i < std::size(addresses); ++i) { addresses[i] = shaders[i] ? shaders[i]->GetGpuAddr() : 0; } return addresses; } -void TransitionImages(const std::vector<ImageView>& views, VkPipelineStageFlags pipeline_stage, - VkAccessFlags access) { - for (auto& [view, layout] : views) { - view->Transition(*layout, pipeline_stage, access); +struct TextureHandle { + constexpr TextureHandle(u32 data, bool via_header_index) { + const Tegra::Texture::TextureHandle handle{data}; + image = handle.tic_id; + sampler = via_header_index ? image : handle.tsc_id.Value(); } -} + + u32 image; + u32 sampler; +}; template <typename Engine, typename Entry> -Tegra::Texture::FullTextureInfo GetTextureInfo(const Engine& engine, const Entry& entry, - std::size_t stage, std::size_t index = 0) { - const auto stage_type = static_cast<Tegra::Engines::ShaderType>(stage); +TextureHandle GetTextureInfo(const Engine& engine, bool via_header_index, const Entry& entry, + size_t stage, size_t index = 0) { + const auto shader_type = static_cast<Tegra::Engines::ShaderType>(stage); if constexpr (std::is_same_v<Entry, SamplerEntry>) { if (entry.is_separated) { const u32 buffer_1 = entry.buffer; const u32 buffer_2 = entry.secondary_buffer; const u32 offset_1 = entry.offset; const u32 offset_2 = entry.secondary_offset; - const u32 handle_1 = engine.AccessConstBuffer32(stage_type, buffer_1, offset_1); - const u32 handle_2 = engine.AccessConstBuffer32(stage_type, buffer_2, offset_2); - return engine.GetTextureInfo(handle_1 | handle_2); + const u32 handle_1 = engine.AccessConstBuffer32(shader_type, buffer_1, offset_1); + const u32 handle_2 = engine.AccessConstBuffer32(shader_type, buffer_2, offset_2); + return TextureHandle(handle_1 | handle_2, via_header_index); } } if (entry.is_bindless) { - const auto tex_handle = engine.AccessConstBuffer32(stage_type, entry.buffer, entry.offset); - return engine.GetTextureInfo(tex_handle); - } - const auto& gpu_profile = engine.AccessGuestDriverProfile(); - const u32 entry_offset = static_cast<u32>(index * gpu_profile.GetTextureHandlerSize()); - const u32 offset = entry.offset + entry_offset; - if constexpr (std::is_same_v<Engine, Tegra::Engines::Maxwell3D>) { - return engine.GetStageTexture(stage_type, offset); - } else { - return engine.GetTexture(offset); - } -} - -/// @brief Determine if an attachment to be updated has to preserve contents -/// @param is_clear True when a clear is being executed -/// @param regs 3D registers -/// @return True when the contents have to be preserved -bool HasToPreserveColorContents(bool is_clear, const Maxwell& regs) { - if (!is_clear) { - return true; - } - // First we have to make sure all clear masks are enabled. - if (!regs.clear_buffers.R || !regs.clear_buffers.G || !regs.clear_buffers.B || - !regs.clear_buffers.A) { - return true; - } - // If scissors are disabled, the whole screen is cleared - if (!regs.clear_flags.scissor) { - return false; + const u32 raw = engine.AccessConstBuffer32(shader_type, entry.buffer, entry.offset); + return TextureHandle(raw, via_header_index); } - // Then we have to confirm scissor testing clears the whole image - const std::size_t index = regs.clear_buffers.RT; - const auto& scissor = regs.scissor_test[0]; - return scissor.min_x > 0 || scissor.min_y > 0 || scissor.max_x < regs.rt[index].width || - scissor.max_y < regs.rt[index].height; + const u32 buffer = engine.GetBoundBuffer(); + const u64 offset = (entry.offset + index) * sizeof(u32); + return TextureHandle(engine.AccessConstBuffer32(shader_type, buffer, offset), via_header_index); } -/// @brief Determine if an attachment to be updated has to preserve contents -/// @param is_clear True when a clear is being executed -/// @param regs 3D registers -/// @return True when the contents have to be preserved -bool HasToPreserveDepthContents(bool is_clear, const Maxwell& regs) { - // If we are not clearing, the contents have to be preserved - if (!is_clear) { - return true; - } - // For depth stencil clears we only have to confirm scissor test covers the whole image - if (!regs.clear_flags.scissor) { - return false; - } - // Make sure the clear cover the whole image - const auto& scissor = regs.scissor_test[0]; - return scissor.min_x > 0 || scissor.min_y > 0 || scissor.max_x < regs.zeta_width || - scissor.max_y < regs.zeta_height; -} - -template <std::size_t N> +template <size_t N> std::array<VkDeviceSize, N> ExpandStrides(const std::array<u16, N>& strides) { std::array<VkDeviceSize, N> expanded; std::copy(strides.begin(), strides.end(), expanded.begin()); return expanded; } +ImageViewType ImageViewTypeFromEntry(const SamplerEntry& entry) { + if (entry.is_buffer) { + return ImageViewType::e2D; + } + switch (entry.type) { + case Tegra::Shader::TextureType::Texture1D: + return entry.is_array ? ImageViewType::e1DArray : ImageViewType::e1D; + case Tegra::Shader::TextureType::Texture2D: + return entry.is_array ? ImageViewType::e2DArray : ImageViewType::e2D; + case Tegra::Shader::TextureType::Texture3D: + return ImageViewType::e3D; + case Tegra::Shader::TextureType::TextureCube: + return entry.is_array ? ImageViewType::CubeArray : ImageViewType::Cube; + } + UNREACHABLE(); + return ImageViewType::e2D; +} + +ImageViewType ImageViewTypeFromEntry(const ImageEntry& entry) { + switch (entry.type) { + case Tegra::Shader::ImageType::Texture1D: + return ImageViewType::e1D; + case Tegra::Shader::ImageType::Texture1DArray: + return ImageViewType::e1DArray; + case Tegra::Shader::ImageType::Texture2D: + return ImageViewType::e2D; + case Tegra::Shader::ImageType::Texture2DArray: + return ImageViewType::e2DArray; + case Tegra::Shader::ImageType::Texture3D: + return ImageViewType::e3D; + case Tegra::Shader::ImageType::TextureBuffer: + return ImageViewType::Buffer; + } + UNREACHABLE(); + return ImageViewType::e2D; +} + +void PushImageDescriptors(const ShaderEntries& entries, TextureCache& texture_cache, + VKUpdateDescriptorQueue& update_descriptor_queue, + ImageViewId*& image_view_id_ptr, VkSampler*& sampler_ptr) { + for ([[maybe_unused]] const auto& entry : entries.uniform_texels) { + const ImageViewId image_view_id = *image_view_id_ptr++; + const ImageView& image_view = texture_cache.GetImageView(image_view_id); + update_descriptor_queue.AddTexelBuffer(image_view.BufferView()); + } + for (const auto& entry : entries.samplers) { + for (size_t i = 0; i < entry.size; ++i) { + const VkSampler sampler = *sampler_ptr++; + const ImageViewId image_view_id = *image_view_id_ptr++; + const ImageView& image_view = texture_cache.GetImageView(image_view_id); + const VkImageView handle = image_view.Handle(ImageViewTypeFromEntry(entry)); + update_descriptor_queue.AddSampledImage(handle, sampler); + } + } + for ([[maybe_unused]] const auto& entry : entries.storage_texels) { + const ImageViewId image_view_id = *image_view_id_ptr++; + const ImageView& image_view = texture_cache.GetImageView(image_view_id); + update_descriptor_queue.AddTexelBuffer(image_view.BufferView()); + } + for (const auto& entry : entries.images) { + // TODO: Mark as modified + const ImageViewId image_view_id = *image_view_id_ptr++; + const ImageView& image_view = texture_cache.GetImageView(image_view_id); + const VkImageView handle = image_view.Handle(ImageViewTypeFromEntry(entry)); + update_descriptor_queue.AddImage(handle); + } +} + } // Anonymous namespace class BufferBindings final { @@ -213,7 +239,7 @@ public: index.type = type; } - void Bind(const VKDevice& device, VKScheduler& scheduler) const { + void Bind(const Device& device, VKScheduler& scheduler) const { // Use this large switch case to avoid dispatching more memory in the record lambda than // what we need. It looks horrible, but it's the best we can do on standard C++. switch (vertex.num_buffers) { @@ -290,7 +316,7 @@ public: private: // Some of these fields are intentionally left uninitialized to avoid initializing them twice. struct { - std::size_t num_buffers = 0; + size_t num_buffers = 0; std::array<VkBuffer, Maxwell::NumVertexArrays> buffers; std::array<VkDeviceSize, Maxwell::NumVertexArrays> offsets; std::array<VkDeviceSize, Maxwell::NumVertexArrays> sizes; @@ -303,8 +329,8 @@ private: VkIndexType type; } index; - template <std::size_t N> - void BindStatic(const VKDevice& device, VKScheduler& scheduler) const { + template <size_t N> + void BindStatic(const Device& device, VKScheduler& scheduler) const { if (device.IsExtExtendedDynamicStateSupported()) { if (index.buffer) { BindStatic<N, true, true>(scheduler); @@ -320,7 +346,7 @@ private: } } - template <std::size_t N, bool is_indexed, bool has_extended_dynamic_state> + template <size_t N, bool is_indexed, bool has_extended_dynamic_state> void BindStatic(VKScheduler& scheduler) const { static_assert(N <= Maxwell::NumVertexArrays); if constexpr (N == 0) { @@ -380,28 +406,31 @@ void RasterizerVulkan::DrawParameters::Draw(vk::CommandBuffer cmdbuf) const { } } -RasterizerVulkan::RasterizerVulkan(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu_, +RasterizerVulkan::RasterizerVulkan(Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_, Tegra::MemoryManager& gpu_memory_, - Core::Memory::Memory& cpu_memory, VKScreenInfo& screen_info_, - const VKDevice& device_, VKMemoryManager& memory_manager_, + Core::Memory::Memory& cpu_memory_, VKScreenInfo& screen_info_, + const Device& device_, VKMemoryManager& memory_manager_, StateTracker& state_tracker_, VKScheduler& scheduler_) - : RasterizerAccelerated(cpu_memory), gpu(gpu_), gpu_memory(gpu_memory_), - maxwell3d(gpu.Maxwell3D()), kepler_compute(gpu.KeplerCompute()), screen_info(screen_info_), - device(device_), memory_manager(memory_manager_), state_tracker(state_tracker_), - scheduler(scheduler_), staging_pool(device, memory_manager, scheduler), - descriptor_pool(device, scheduler_), update_descriptor_queue(device, scheduler), - renderpass_cache(device), + : RasterizerAccelerated{cpu_memory_}, gpu{gpu_}, + gpu_memory{gpu_memory_}, maxwell3d{gpu.Maxwell3D()}, kepler_compute{gpu.KeplerCompute()}, + screen_info{screen_info_}, device{device_}, memory_manager{memory_manager_}, + state_tracker{state_tracker_}, scheduler{scheduler_}, stream_buffer(device, scheduler), + staging_pool(device, memory_manager, scheduler), descriptor_pool(device, scheduler), + update_descriptor_queue(device, scheduler), + blit_image(device, scheduler, state_tracker, descriptor_pool), quad_array_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue), quad_indexed_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue), uint8_pass(device, scheduler, descriptor_pool, staging_pool, update_descriptor_queue), - texture_cache(*this, maxwell3d, gpu_memory, device, memory_manager, scheduler, staging_pool), + texture_cache_runtime{device, scheduler, memory_manager, staging_pool, blit_image}, + texture_cache(texture_cache_runtime, *this, maxwell3d, kepler_compute, gpu_memory), pipeline_cache(*this, gpu, maxwell3d, kepler_compute, gpu_memory, device, scheduler, - descriptor_pool, update_descriptor_queue, renderpass_cache), - buffer_cache(*this, gpu_memory, cpu_memory, device, memory_manager, scheduler, staging_pool), - sampler_cache(device), query_cache(*this, maxwell3d, gpu_memory, device, scheduler), + descriptor_pool, update_descriptor_queue), + buffer_cache(*this, gpu_memory, cpu_memory_, device, memory_manager, scheduler, stream_buffer, + staging_pool), + query_cache{*this, maxwell3d, gpu_memory, device, scheduler}, fence_manager(*this, gpu, gpu_memory, texture_cache, buffer_cache, query_cache, device, scheduler), - wfi_event(device.GetLogical().CreateEvent()), async_shaders(emu_window) { + wfi_event(device.GetLogical().CreateEvent()), async_shaders(emu_window_) { scheduler.SetQueryCache(query_cache); if (device.UseAsynchronousShaders()) { async_shaders.AllocateWorkers(); @@ -427,9 +456,10 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) { const DrawParameters draw_params = SetupGeometry(key.fixed_state, buffer_bindings, is_indexed, is_instanced); - update_descriptor_queue.Acquire(); - sampled_views.clear(); - image_views.clear(); + auto lock = texture_cache.AcquireLock(); + texture_cache.SynchronizeGraphicsDescriptors(); + + texture_cache.UpdateRenderTargets(false); const auto shaders = pipeline_cache.GetShaders(); key.shaders = GetShaderAddresses(shaders); @@ -437,30 +467,24 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) { buffer_cache.Unmap(); - const Texceptions texceptions = UpdateAttachments(false); - SetupImageTransitions(texceptions, color_attachments, zeta_attachment); - - key.renderpass_params = GetRenderPassParams(texceptions); - key.padding = 0; + const Framebuffer* const framebuffer = texture_cache.GetFramebuffer(); + key.renderpass = framebuffer->RenderPass(); - auto* pipeline = pipeline_cache.GetGraphicsPipeline(key, async_shaders); + auto* const pipeline = + pipeline_cache.GetGraphicsPipeline(key, framebuffer->NumColorBuffers(), async_shaders); if (pipeline == nullptr || pipeline->GetHandle() == VK_NULL_HANDLE) { // Async graphics pipeline was not ready. return; } - scheduler.BindGraphicsPipeline(pipeline->GetHandle()); - - const auto renderpass = pipeline->GetRenderPass(); - const auto [framebuffer, render_area] = ConfigureFramebuffers(renderpass); - scheduler.RequestRenderpass(renderpass, framebuffer, render_area); - - UpdateDynamicStates(); - buffer_bindings.Bind(device, scheduler); BeginTransformFeedback(); + scheduler.RequestRenderpass(framebuffer); + scheduler.BindGraphicsPipeline(pipeline->GetHandle()); + UpdateDynamicStates(); + const auto pipeline_layout = pipeline->GetLayout(); const auto descriptor_set = pipeline->CommitDescriptorSet(); scheduler.Record([pipeline_layout, descriptor_set, draw_params](vk::CommandBuffer cmdbuf) { @@ -481,9 +505,6 @@ void RasterizerVulkan::Clear() { return; } - sampled_views.clear(); - image_views.clear(); - query_cache.UpdateCounters(); const auto& regs = maxwell3d.regs; @@ -495,20 +516,24 @@ void RasterizerVulkan::Clear() { return; } - [[maybe_unused]] const auto texceptions = UpdateAttachments(true); - DEBUG_ASSERT(texceptions.none()); - SetupImageTransitions(0, color_attachments, zeta_attachment); - - const VkRenderPass renderpass = renderpass_cache.GetRenderPass(GetRenderPassParams(0)); - const auto [framebuffer, render_area] = ConfigureFramebuffers(renderpass); - scheduler.RequestRenderpass(renderpass, framebuffer, render_area); + auto lock = texture_cache.AcquireLock(); + texture_cache.UpdateRenderTargets(true); + const Framebuffer* const framebuffer = texture_cache.GetFramebuffer(); + const VkExtent2D render_area = framebuffer->RenderArea(); + scheduler.RequestRenderpass(framebuffer); - VkClearRect clear_rect; - clear_rect.baseArrayLayer = regs.clear_buffers.layer; - clear_rect.layerCount = 1; - clear_rect.rect = GetScissorState(regs, 0); - clear_rect.rect.extent.width = std::min(clear_rect.rect.extent.width, render_area.width); - clear_rect.rect.extent.height = std::min(clear_rect.rect.extent.height, render_area.height); + VkClearRect clear_rect{ + .rect = GetScissorState(regs, 0), + .baseArrayLayer = regs.clear_buffers.layer, + .layerCount = 1, + }; + if (clear_rect.rect.extent.width == 0 || clear_rect.rect.extent.height == 0) { + return; + } + clear_rect.rect.extent = VkExtent2D{ + .width = std::min(clear_rect.rect.extent.width, render_area.width), + .height = std::min(clear_rect.rect.extent.height, render_area.height), + }; if (use_color) { VkClearValue clear_value; @@ -549,9 +574,6 @@ void RasterizerVulkan::Clear() { void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) { MICROPROFILE_SCOPE(Vulkan_Compute); - update_descriptor_queue.Acquire(); - sampled_views.clear(); - image_views.clear(); query_cache.UpdateCounters(); @@ -570,30 +592,46 @@ void RasterizerVulkan::DispatchCompute(GPUVAddr code_addr) { // Compute dispatches can't be executed inside a renderpass scheduler.RequestOutsideRenderPassOperationContext(); - buffer_cache.Map(CalculateComputeStreamBufferSize()); + image_view_indices.clear(); + sampler_handles.clear(); + + auto lock = texture_cache.AcquireLock(); + texture_cache.SynchronizeComputeDescriptors(); const auto& entries = pipeline.GetEntries(); - SetupComputeConstBuffers(entries); - SetupComputeGlobalBuffers(entries); SetupComputeUniformTexels(entries); SetupComputeTextures(entries); SetupComputeStorageTexels(entries); SetupComputeImages(entries); - buffer_cache.Unmap(); + const std::span indices_span(image_view_indices.data(), image_view_indices.size()); + texture_cache.FillComputeImageViews(indices_span, image_view_ids); + + buffer_cache.Map(CalculateComputeStreamBufferSize()); + + update_descriptor_queue.Acquire(); + + SetupComputeConstBuffers(entries); + SetupComputeGlobalBuffers(entries); + + ImageViewId* image_view_id_ptr = image_view_ids.data(); + VkSampler* sampler_ptr = sampler_handles.data(); + PushImageDescriptors(entries, texture_cache, update_descriptor_queue, image_view_id_ptr, + sampler_ptr); - TransitionImages(sampled_views, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, - VK_ACCESS_SHADER_READ_BIT); - TransitionImages(image_views, VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, - VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT); + buffer_cache.Unmap(); + const VkPipeline pipeline_handle = pipeline.GetHandle(); + const VkPipelineLayout pipeline_layout = pipeline.GetLayout(); + const VkDescriptorSet descriptor_set = pipeline.CommitDescriptorSet(); scheduler.Record([grid_x = launch_desc.grid_dim_x, grid_y = launch_desc.grid_dim_y, - grid_z = launch_desc.grid_dim_z, pipeline_handle = pipeline.GetHandle(), - layout = pipeline.GetLayout(), - descriptor_set = pipeline.CommitDescriptorSet()](vk::CommandBuffer cmdbuf) { + grid_z = launch_desc.grid_dim_z, pipeline_handle, pipeline_layout, + descriptor_set](vk::CommandBuffer cmdbuf) { cmdbuf.BindPipeline(VK_PIPELINE_BIND_POINT_COMPUTE, pipeline_handle); - cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_COMPUTE, layout, DESCRIPTOR_SET, - descriptor_set, {}); + if (descriptor_set) { + cmdbuf.BindDescriptorSets(VK_PIPELINE_BIND_POINT_COMPUTE, pipeline_layout, + DESCRIPTOR_SET, descriptor_set, nullptr); + } cmdbuf.Dispatch(grid_x, grid_y, grid_z); }); } @@ -613,7 +651,10 @@ void RasterizerVulkan::FlushRegion(VAddr addr, u64 size) { if (addr == 0 || size == 0) { return; } - texture_cache.FlushRegion(addr, size); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.DownloadMemory(addr, size); + } buffer_cache.FlushRegion(addr, size); query_cache.FlushRegion(addr, size); } @@ -622,14 +663,18 @@ bool RasterizerVulkan::MustFlushRegion(VAddr addr, u64 size) { if (!Settings::IsGPULevelHigh()) { return buffer_cache.MustFlushRegion(addr, size); } - return texture_cache.MustFlushRegion(addr, size) || buffer_cache.MustFlushRegion(addr, size); + return texture_cache.IsRegionGpuModified(addr, size) || + buffer_cache.MustFlushRegion(addr, size); } void RasterizerVulkan::InvalidateRegion(VAddr addr, u64 size) { if (addr == 0 || size == 0) { return; } - texture_cache.InvalidateRegion(addr, size); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.WriteMemory(addr, size); + } pipeline_cache.InvalidateRegion(addr, size); buffer_cache.InvalidateRegion(addr, size); query_cache.InvalidateRegion(addr, size); @@ -639,17 +684,28 @@ void RasterizerVulkan::OnCPUWrite(VAddr addr, u64 size) { if (addr == 0 || size == 0) { return; } - texture_cache.OnCPUWrite(addr, size); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.WriteMemory(addr, size); + } pipeline_cache.OnCPUWrite(addr, size); buffer_cache.OnCPUWrite(addr, size); } void RasterizerVulkan::SyncGuestHost() { - texture_cache.SyncGuestHost(); buffer_cache.SyncGuestHost(); pipeline_cache.SyncGuestHost(); } +void RasterizerVulkan::UnmapMemory(VAddr addr, u64 size) { + { + auto lock = texture_cache.AcquireLock(); + texture_cache.UnmapMemory(addr, size); + } + buffer_cache.OnCPUWrite(addr, size); + pipeline_cache.OnCPUWrite(addr, size); +} + void RasterizerVulkan::SignalSemaphore(GPUVAddr addr, u32 value) { if (!gpu.IsAsync()) { gpu_memory.Write<u32>(addr, value); @@ -700,6 +756,14 @@ void RasterizerVulkan::WaitForIdle() { }); } +void RasterizerVulkan::FragmentBarrier() { + // We already put barriers when a render pass finishes +} + +void RasterizerVulkan::TiledCacheBarrier() { + // TODO: Implementing tiled barriers requires rewriting a good chunk of the Vulkan backend +} + void RasterizerVulkan::FlushCommands() { if (draw_counter > 0) { draw_counter = 0; @@ -710,14 +774,20 @@ void RasterizerVulkan::FlushCommands() { void RasterizerVulkan::TickFrame() { draw_counter = 0; update_descriptor_queue.TickFrame(); + fence_manager.TickFrame(); buffer_cache.TickFrame(); staging_pool.TickFrame(); + { + auto lock = texture_cache.AcquireLock(); + texture_cache.TickFrame(); + } } -bool RasterizerVulkan::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, - const Tegra::Engines::Fermi2D::Regs::Surface& dst, +bool RasterizerVulkan::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Surface& src, + const Tegra::Engines::Fermi2D::Surface& dst, const Tegra::Engines::Fermi2D::Config& copy_config) { - texture_cache.DoFermiCopy(src, dst, copy_config); + auto lock = texture_cache.AcquireLock(); + texture_cache.BlitImage(dst, src, copy_config); return true; } @@ -727,20 +797,16 @@ bool RasterizerVulkan::AccelerateDisplay(const Tegra::FramebufferConfig& config, return false; } - const auto surface{texture_cache.TryFindFramebufferSurface(framebuffer_addr)}; - if (!surface) { + auto lock = texture_cache.AcquireLock(); + ImageView* const image_view = texture_cache.TryFindFramebufferImageView(framebuffer_addr); + if (!image_view) { return false; } - // Verify that the cached surface is the same size and format as the requested framebuffer - const auto& params{surface->GetSurfaceParams()}; - ASSERT_MSG(params.width == config.width, "Framebuffer width is different"); - ASSERT_MSG(params.height == config.height, "Framebuffer height is different"); - - screen_info.image = &surface->GetImage(); - screen_info.width = params.width; - screen_info.height = params.height; - screen_info.is_srgb = surface->GetSurfaceParams().srgb_conversion; + screen_info.image_view = image_view->Handle(VideoCommon::ImageViewType::e2D); + screen_info.width = image_view->size.width; + screen_info.height = image_view->size.height; + screen_info.is_srgb = VideoCore::Surface::IsPixelFormatSRGB(image_view->format); return true; } @@ -765,103 +831,6 @@ void RasterizerVulkan::FlushWork() { draw_counter = 0; } -RasterizerVulkan::Texceptions RasterizerVulkan::UpdateAttachments(bool is_clear) { - MICROPROFILE_SCOPE(Vulkan_RenderTargets); - - const auto& regs = maxwell3d.regs; - auto& dirty = maxwell3d.dirty.flags; - const bool update_rendertargets = dirty[VideoCommon::Dirty::RenderTargets]; - dirty[VideoCommon::Dirty::RenderTargets] = false; - - texture_cache.GuardRenderTargets(true); - - Texceptions texceptions; - for (std::size_t rt = 0; rt < Maxwell::NumRenderTargets; ++rt) { - if (update_rendertargets) { - const bool preserve_contents = HasToPreserveColorContents(is_clear, regs); - color_attachments[rt] = texture_cache.GetColorBufferSurface(rt, preserve_contents); - } - if (color_attachments[rt] && WalkAttachmentOverlaps(*color_attachments[rt])) { - texceptions[rt] = true; - } - } - - if (update_rendertargets) { - const bool preserve_contents = HasToPreserveDepthContents(is_clear, regs); - zeta_attachment = texture_cache.GetDepthBufferSurface(preserve_contents); - } - if (zeta_attachment && WalkAttachmentOverlaps(*zeta_attachment)) { - texceptions[ZETA_TEXCEPTION_INDEX] = true; - } - - texture_cache.GuardRenderTargets(false); - - return texceptions; -} - -bool RasterizerVulkan::WalkAttachmentOverlaps(const CachedSurfaceView& attachment) { - bool overlap = false; - for (auto& [view, layout] : sampled_views) { - if (!attachment.IsSameSurface(*view)) { - continue; - } - overlap = true; - *layout = VK_IMAGE_LAYOUT_GENERAL; - } - return overlap; -} - -std::tuple<VkFramebuffer, VkExtent2D> RasterizerVulkan::ConfigureFramebuffers( - VkRenderPass renderpass) { - FramebufferCacheKey key{ - .renderpass = renderpass, - .width = std::numeric_limits<u32>::max(), - .height = std::numeric_limits<u32>::max(), - .layers = std::numeric_limits<u32>::max(), - .views = {}, - }; - - const auto try_push = [&key](const View& view) { - if (!view) { - return false; - } - key.views.push_back(view->GetAttachment()); - key.width = std::min(key.width, view->GetWidth()); - key.height = std::min(key.height, view->GetHeight()); - key.layers = std::min(key.layers, view->GetNumLayers()); - return true; - }; - - const auto& regs = maxwell3d.regs; - const std::size_t num_attachments = static_cast<std::size_t>(regs.rt_control.count); - for (std::size_t index = 0; index < num_attachments; ++index) { - if (try_push(color_attachments[index])) { - texture_cache.MarkColorBufferInUse(index); - } - } - if (try_push(zeta_attachment)) { - texture_cache.MarkDepthBufferInUse(); - } - - const auto [fbentry, is_cache_miss] = framebuffer_cache.try_emplace(key); - auto& framebuffer = fbentry->second; - if (is_cache_miss) { - framebuffer = device.GetLogical().CreateFramebuffer({ - .sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO, - .pNext = nullptr, - .flags = 0, - .renderPass = key.renderpass, - .attachmentCount = static_cast<u32>(key.views.size()), - .pAttachments = key.views.data(), - .width = key.width, - .height = key.height, - .layers = key.layers, - }); - } - - return {*framebuffer, VkExtent2D{key.width, key.height}}; -} - RasterizerVulkan::DrawParameters RasterizerVulkan::SetupGeometry(FixedPipelineState& fixed_state, BufferBindings& buffer_bindings, bool is_indexed, @@ -885,51 +854,37 @@ RasterizerVulkan::DrawParameters RasterizerVulkan::SetupGeometry(FixedPipelineSt void RasterizerVulkan::SetupShaderDescriptors( const std::array<Shader*, Maxwell::MaxShaderProgram>& shaders) { - texture_cache.GuardSamplers(true); - - for (std::size_t stage = 0; stage < Maxwell::MaxShaderStage; ++stage) { - // Skip VertexA stage + image_view_indices.clear(); + sampler_handles.clear(); + for (size_t stage = 0; stage < Maxwell::MaxShaderStage; ++stage) { Shader* const shader = shaders[stage + 1]; if (!shader) { continue; } const auto& entries = shader->GetEntries(); - SetupGraphicsConstBuffers(entries, stage); - SetupGraphicsGlobalBuffers(entries, stage); SetupGraphicsUniformTexels(entries, stage); SetupGraphicsTextures(entries, stage); SetupGraphicsStorageTexels(entries, stage); SetupGraphicsImages(entries, stage); } - texture_cache.GuardSamplers(false); -} + const std::span indices_span(image_view_indices.data(), image_view_indices.size()); + texture_cache.FillGraphicsImageViews(indices_span, image_view_ids); -void RasterizerVulkan::SetupImageTransitions( - Texceptions texceptions, const std::array<View, Maxwell::NumRenderTargets>& color_attachments, - const View& zeta_attachment) { - TransitionImages(sampled_views, VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT, VK_ACCESS_SHADER_READ_BIT); - TransitionImages(image_views, VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT, - VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT); + update_descriptor_queue.Acquire(); - for (std::size_t rt = 0; rt < std::size(color_attachments); ++rt) { - const auto color_attachment = color_attachments[rt]; - if (color_attachment == nullptr) { + ImageViewId* image_view_id_ptr = image_view_ids.data(); + VkSampler* sampler_ptr = sampler_handles.data(); + for (size_t stage = 0; stage < Maxwell::MaxShaderStage; ++stage) { + // Skip VertexA stage + Shader* const shader = shaders[stage + 1]; + if (!shader) { continue; } - const auto image_layout = - texceptions[rt] ? VK_IMAGE_LAYOUT_GENERAL : VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL; - color_attachment->Transition(image_layout, VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT, - VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | - VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT); - } - - if (zeta_attachment != nullptr) { - const auto image_layout = texceptions[ZETA_TEXCEPTION_INDEX] - ? VK_IMAGE_LAYOUT_GENERAL - : VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL; - zeta_attachment->Transition(image_layout, VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT, - VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | - VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT); + const auto& entries = shader->GetEntries(); + SetupGraphicsConstBuffers(entries, stage); + SetupGraphicsGlobalBuffers(entries, stage); + PushImageDescriptors(entries, texture_cache, update_descriptor_queue, image_view_id_ptr, + sampler_ptr); } } @@ -948,7 +903,6 @@ void RasterizerVulkan::UpdateDynamicStates() { UpdateDepthWriteEnable(regs); UpdateDepthCompareOp(regs); UpdateFrontFace(regs); - UpdatePrimitiveTopology(regs); UpdateStencilOp(regs); UpdateStencilTestEnable(regs); } @@ -1002,7 +956,7 @@ void RasterizerVulkan::EndTransformFeedback() { void RasterizerVulkan::SetupVertexArrays(BufferBindings& buffer_bindings) { const auto& regs = maxwell3d.regs; - for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { + for (size_t index = 0; index < Maxwell::NumVertexArrays; ++index) { const auto& vertex_array = regs.vertex_array[index]; if (!vertex_array.IsEnabled()) { continue; @@ -1011,7 +965,7 @@ void RasterizerVulkan::SetupVertexArrays(BufferBindings& buffer_bindings) { const GPUVAddr end{regs.vertex_array_limit[index].LimitAddress()}; ASSERT(end >= start); - const std::size_t size = end - start; + const size_t size = end - start; if (size == 0) { buffer_bindings.AddVertexBinding(DefaultBuffer(), 0, DEFAULT_BUFFER_SIZE, 0); continue; @@ -1072,7 +1026,7 @@ void RasterizerVulkan::SetupIndexBuffer(BufferBindings& buffer_bindings, DrawPar } } -void RasterizerVulkan::SetupGraphicsConstBuffers(const ShaderEntries& entries, std::size_t stage) { +void RasterizerVulkan::SetupGraphicsConstBuffers(const ShaderEntries& entries, size_t stage) { MICROPROFILE_SCOPE(Vulkan_ConstBuffers); const auto& shader_stage = maxwell3d.state.shader_stages[stage]; for (const auto& entry : entries.const_buffers) { @@ -1080,7 +1034,7 @@ void RasterizerVulkan::SetupGraphicsConstBuffers(const ShaderEntries& entries, s } } -void RasterizerVulkan::SetupGraphicsGlobalBuffers(const ShaderEntries& entries, std::size_t stage) { +void RasterizerVulkan::SetupGraphicsGlobalBuffers(const ShaderEntries& entries, size_t stage) { MICROPROFILE_SCOPE(Vulkan_GlobalBuffers); const auto& cbufs{maxwell3d.state.shader_stages[stage]}; @@ -1090,37 +1044,49 @@ void RasterizerVulkan::SetupGraphicsGlobalBuffers(const ShaderEntries& entries, } } -void RasterizerVulkan::SetupGraphicsUniformTexels(const ShaderEntries& entries, std::size_t stage) { +void RasterizerVulkan::SetupGraphicsUniformTexels(const ShaderEntries& entries, size_t stage) { MICROPROFILE_SCOPE(Vulkan_Textures); + const auto& regs = maxwell3d.regs; + const bool via_header_index = regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex; for (const auto& entry : entries.uniform_texels) { - const auto image = GetTextureInfo(maxwell3d, entry, stage).tic; - SetupUniformTexels(image, entry); + const TextureHandle handle = GetTextureInfo(maxwell3d, via_header_index, entry, stage); + image_view_indices.push_back(handle.image); } } -void RasterizerVulkan::SetupGraphicsTextures(const ShaderEntries& entries, std::size_t stage) { +void RasterizerVulkan::SetupGraphicsTextures(const ShaderEntries& entries, size_t stage) { MICROPROFILE_SCOPE(Vulkan_Textures); + const auto& regs = maxwell3d.regs; + const bool via_header_index = regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex; for (const auto& entry : entries.samplers) { - for (std::size_t i = 0; i < entry.size; ++i) { - const auto texture = GetTextureInfo(maxwell3d, entry, stage, i); - SetupTexture(texture, entry); + for (size_t index = 0; index < entry.size; ++index) { + const TextureHandle handle = + GetTextureInfo(maxwell3d, via_header_index, entry, stage, index); + image_view_indices.push_back(handle.image); + + Sampler* const sampler = texture_cache.GetGraphicsSampler(handle.sampler); + sampler_handles.push_back(sampler->Handle()); } } } -void RasterizerVulkan::SetupGraphicsStorageTexels(const ShaderEntries& entries, std::size_t stage) { +void RasterizerVulkan::SetupGraphicsStorageTexels(const ShaderEntries& entries, size_t stage) { MICROPROFILE_SCOPE(Vulkan_Textures); + const auto& regs = maxwell3d.regs; + const bool via_header_index = regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex; for (const auto& entry : entries.storage_texels) { - const auto image = GetTextureInfo(maxwell3d, entry, stage).tic; - SetupStorageTexel(image, entry); + const TextureHandle handle = GetTextureInfo(maxwell3d, via_header_index, entry, stage); + image_view_indices.push_back(handle.image); } } -void RasterizerVulkan::SetupGraphicsImages(const ShaderEntries& entries, std::size_t stage) { +void RasterizerVulkan::SetupGraphicsImages(const ShaderEntries& entries, size_t stage) { MICROPROFILE_SCOPE(Vulkan_Images); + const auto& regs = maxwell3d.regs; + const bool via_header_index = regs.sampler_index == Maxwell::SamplerIndex::ViaHeaderIndex; for (const auto& entry : entries.images) { - const auto tic = GetTextureInfo(maxwell3d, entry, stage).tic; - SetupImage(tic, entry); + const TextureHandle handle = GetTextureInfo(maxwell3d, via_header_index, entry, stage); + image_view_indices.push_back(handle.image); } } @@ -1130,11 +1096,12 @@ void RasterizerVulkan::SetupComputeConstBuffers(const ShaderEntries& entries) { for (const auto& entry : entries.const_buffers) { const auto& config = launch_desc.const_buffer_config[entry.GetIndex()]; const std::bitset<8> mask = launch_desc.const_buffer_enable_mask.Value(); - Tegra::Engines::ConstBufferInfo buffer; - buffer.address = config.Address(); - buffer.size = config.size; - buffer.enabled = mask[entry.GetIndex()]; - SetupConstBuffer(entry, buffer); + const Tegra::Engines::ConstBufferInfo info{ + .address = config.Address(), + .size = config.size, + .enabled = mask[entry.GetIndex()], + }; + SetupConstBuffer(entry, info); } } @@ -1149,35 +1116,46 @@ void RasterizerVulkan::SetupComputeGlobalBuffers(const ShaderEntries& entries) { void RasterizerVulkan::SetupComputeUniformTexels(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Textures); + const bool via_header_index = kepler_compute.launch_description.linked_tsc; for (const auto& entry : entries.uniform_texels) { - const auto image = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex).tic; - SetupUniformTexels(image, entry); + const TextureHandle handle = + GetTextureInfo(kepler_compute, via_header_index, entry, COMPUTE_SHADER_INDEX); + image_view_indices.push_back(handle.image); } } void RasterizerVulkan::SetupComputeTextures(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Textures); + const bool via_header_index = kepler_compute.launch_description.linked_tsc; for (const auto& entry : entries.samplers) { - for (std::size_t i = 0; i < entry.size; ++i) { - const auto texture = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex, i); - SetupTexture(texture, entry); + for (size_t index = 0; index < entry.size; ++index) { + const TextureHandle handle = GetTextureInfo(kepler_compute, via_header_index, entry, + COMPUTE_SHADER_INDEX, index); + image_view_indices.push_back(handle.image); + + Sampler* const sampler = texture_cache.GetComputeSampler(handle.sampler); + sampler_handles.push_back(sampler->Handle()); } } } void RasterizerVulkan::SetupComputeStorageTexels(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Textures); + const bool via_header_index = kepler_compute.launch_description.linked_tsc; for (const auto& entry : entries.storage_texels) { - const auto image = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex).tic; - SetupStorageTexel(image, entry); + const TextureHandle handle = + GetTextureInfo(kepler_compute, via_header_index, entry, COMPUTE_SHADER_INDEX); + image_view_indices.push_back(handle.image); } } void RasterizerVulkan::SetupComputeImages(const ShaderEntries& entries) { MICROPROFILE_SCOPE(Vulkan_Images); + const bool via_header_index = kepler_compute.launch_description.linked_tsc; for (const auto& entry : entries.images) { - const auto tic = GetTextureInfo(kepler_compute, entry, ComputeShaderIndex).tic; - SetupImage(tic, entry); + const TextureHandle handle = + GetTextureInfo(kepler_compute, via_header_index, entry, COMPUTE_SHADER_INDEX); + image_view_indices.push_back(handle.image); } } @@ -1188,14 +1166,12 @@ void RasterizerVulkan::SetupConstBuffer(const ConstBufferEntry& entry, update_descriptor_queue.AddBuffer(DefaultBuffer(), 0, DEFAULT_BUFFER_SIZE); return; } - // Align the size to avoid bad std140 interactions - const std::size_t size = - Common::AlignUp(CalculateConstBufferSize(entry, buffer), 4 * sizeof(float)); + const size_t size = Common::AlignUp(CalculateConstBufferSize(entry, buffer), 4 * sizeof(float)); ASSERT(size <= MaxConstbufferSize); - const auto info = - buffer_cache.UploadMemory(buffer.address, size, device.GetUniformBufferAlignment()); + const u64 alignment = device.GetUniformBufferAlignment(); + const auto info = buffer_cache.UploadMemory(buffer.address, size, alignment); update_descriptor_queue.AddBuffer(info.handle, info.offset, size); } @@ -1208,7 +1184,7 @@ void RasterizerVulkan::SetupGlobalBuffer(const GlobalBufferEntry& entry, GPUVAdd // because Vulkan doesn't like empty buffers. // Note: Do *not* use DefaultBuffer() here, storage buffers can be written breaking the // default buffer. - static constexpr std::size_t dummy_size = 4; + static constexpr size_t dummy_size = 4; const auto info = buffer_cache.GetEmptyBuffer(dummy_size); update_descriptor_queue.AddBuffer(info.handle, info.offset, dummy_size); return; @@ -1219,55 +1195,6 @@ void RasterizerVulkan::SetupGlobalBuffer(const GlobalBufferEntry& entry, GPUVAdd update_descriptor_queue.AddBuffer(info.handle, info.offset, size); } -void RasterizerVulkan::SetupUniformTexels(const Tegra::Texture::TICEntry& tic, - const UniformTexelEntry& entry) { - const auto view = texture_cache.GetTextureSurface(tic, entry); - ASSERT(view->IsBufferView()); - - update_descriptor_queue.AddTexelBuffer(view->GetBufferView()); -} - -void RasterizerVulkan::SetupTexture(const Tegra::Texture::FullTextureInfo& texture, - const SamplerEntry& entry) { - auto view = texture_cache.GetTextureSurface(texture.tic, entry); - ASSERT(!view->IsBufferView()); - - const VkImageView image_view = view->GetImageView(texture.tic.x_source, texture.tic.y_source, - texture.tic.z_source, texture.tic.w_source); - const auto sampler = sampler_cache.GetSampler(texture.tsc); - update_descriptor_queue.AddSampledImage(sampler, image_view); - - VkImageLayout* const image_layout = update_descriptor_queue.LastImageLayout(); - *image_layout = VK_IMAGE_LAYOUT_SHADER_READ_ONLY_OPTIMAL; - sampled_views.push_back(ImageView{std::move(view), image_layout}); -} - -void RasterizerVulkan::SetupStorageTexel(const Tegra::Texture::TICEntry& tic, - const StorageTexelEntry& entry) { - const auto view = texture_cache.GetImageSurface(tic, entry); - ASSERT(view->IsBufferView()); - - update_descriptor_queue.AddTexelBuffer(view->GetBufferView()); -} - -void RasterizerVulkan::SetupImage(const Tegra::Texture::TICEntry& tic, const ImageEntry& entry) { - auto view = texture_cache.GetImageSurface(tic, entry); - - if (entry.is_written) { - view->MarkAsModified(texture_cache.Tick()); - } - - UNIMPLEMENTED_IF(tic.IsBuffer()); - - const VkImageView image_view = - view->GetImageView(tic.x_source, tic.y_source, tic.z_source, tic.w_source); - update_descriptor_queue.AddImage(image_view); - - VkImageLayout* const image_layout = update_descriptor_queue.LastImageLayout(); - *image_layout = VK_IMAGE_LAYOUT_GENERAL; - image_views.push_back(ImageView{std::move(view), image_layout}); -} - void RasterizerVulkan::UpdateViewportsState(Tegra::Engines::Maxwell3D::Regs& regs) { if (!state_tracker.TouchViewports()) { return; @@ -1418,16 +1345,6 @@ void RasterizerVulkan::UpdateFrontFace(Tegra::Engines::Maxwell3D::Regs& regs) { [front_face](vk::CommandBuffer cmdbuf) { cmdbuf.SetFrontFaceEXT(front_face); }); } -void RasterizerVulkan::UpdatePrimitiveTopology(Tegra::Engines::Maxwell3D::Regs& regs) { - const Maxwell::PrimitiveTopology primitive_topology = regs.draw.topology.Value(); - if (!state_tracker.ChangePrimitiveTopology(primitive_topology)) { - return; - } - scheduler.Record([this, primitive_topology](vk::CommandBuffer cmdbuf) { - cmdbuf.SetPrimitiveTopologyEXT(MaxwellToVK::PrimitiveTopology(device, primitive_topology)); - }); -} - void RasterizerVulkan::UpdateStencilOp(Tegra::Engines::Maxwell3D::Regs& regs) { if (!state_tracker.TouchStencilOp()) { return; @@ -1469,8 +1386,8 @@ void RasterizerVulkan::UpdateStencilTestEnable(Tegra::Engines::Maxwell3D::Regs& }); } -std::size_t RasterizerVulkan::CalculateGraphicsStreamBufferSize(bool is_indexed) const { - std::size_t size = CalculateVertexArraysSize(); +size_t RasterizerVulkan::CalculateGraphicsStreamBufferSize(bool is_indexed) const { + size_t size = CalculateVertexArraysSize(); if (is_indexed) { size = Common::AlignUp(size, 4) + CalculateIndexBufferSize(); } @@ -1478,15 +1395,15 @@ std::size_t RasterizerVulkan::CalculateGraphicsStreamBufferSize(bool is_indexed) return size; } -std::size_t RasterizerVulkan::CalculateComputeStreamBufferSize() const { +size_t RasterizerVulkan::CalculateComputeStreamBufferSize() const { return Tegra::Engines::KeplerCompute::NumConstBuffers * (Maxwell::MaxConstBufferSize + device.GetUniformBufferAlignment()); } -std::size_t RasterizerVulkan::CalculateVertexArraysSize() const { +size_t RasterizerVulkan::CalculateVertexArraysSize() const { const auto& regs = maxwell3d.regs; - std::size_t size = 0; + size_t size = 0; for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) { // This implementation assumes that all attributes are used in the shader. const GPUVAddr start{regs.vertex_array[index].StartAddress()}; @@ -1498,12 +1415,12 @@ std::size_t RasterizerVulkan::CalculateVertexArraysSize() const { return size; } -std::size_t RasterizerVulkan::CalculateIndexBufferSize() const { - return static_cast<std::size_t>(maxwell3d.regs.index_array.count) * - static_cast<std::size_t>(maxwell3d.regs.index_array.FormatSizeInBytes()); +size_t RasterizerVulkan::CalculateIndexBufferSize() const { + return static_cast<size_t>(maxwell3d.regs.index_array.count) * + static_cast<size_t>(maxwell3d.regs.index_array.FormatSizeInBytes()); } -std::size_t RasterizerVulkan::CalculateConstBufferSize( +size_t RasterizerVulkan::CalculateConstBufferSize( const ConstBufferEntry& entry, const Tegra::Engines::ConstBufferInfo& buffer) const { if (entry.IsIndirect()) { // Buffer is accessed indirectly, so upload the entire thing @@ -1514,37 +1431,10 @@ std::size_t RasterizerVulkan::CalculateConstBufferSize( } } -RenderPassParams RasterizerVulkan::GetRenderPassParams(Texceptions texceptions) const { - const auto& regs = maxwell3d.regs; - const std::size_t num_attachments = static_cast<std::size_t>(regs.rt_control.count); - - RenderPassParams params; - params.color_formats = {}; - std::size_t color_texceptions = 0; - - std::size_t index = 0; - for (std::size_t rt = 0; rt < num_attachments; ++rt) { - const auto& rendertarget = regs.rt[rt]; - if (rendertarget.Address() == 0 || rendertarget.format == Tegra::RenderTargetFormat::NONE) { - continue; - } - params.color_formats[index] = static_cast<u8>(rendertarget.format); - color_texceptions |= (texceptions[rt] ? 1ULL : 0ULL) << index; - ++index; - } - params.num_color_attachments = static_cast<u8>(index); - params.texceptions = static_cast<u8>(color_texceptions); - - params.zeta_format = regs.zeta_enable ? static_cast<u8>(regs.zeta.format) : 0; - params.zeta_texception = texceptions[ZETA_TEXCEPTION_INDEX]; - return params; -} - VkBuffer RasterizerVulkan::DefaultBuffer() { if (default_buffer) { return *default_buffer; } - default_buffer = device.GetLogical().CreateBuffer({ .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, .pNext = nullptr, diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h index b47c8fc13..4695718e9 100644 --- a/src/video_core/renderer_vulkan/vk_rasterizer.h +++ b/src/video_core/renderer_vulkan/vk_rasterizer.h @@ -11,11 +11,11 @@ #include <vector> #include <boost/container/static_vector.hpp> -#include <boost/functional/hash.hpp> #include "common/common_types.h" #include "video_core/rasterizer_accelerated.h" #include "video_core/rasterizer_interface.h" +#include "video_core/renderer_vulkan/blit_image.h" #include "video_core/renderer_vulkan/fixed_pipeline_state.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" #include "video_core/renderer_vulkan/vk_compute_pass.h" @@ -24,14 +24,13 @@ #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_query_cache.h" -#include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/vk_sampler_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" +#include "video_core/renderer_vulkan/vk_stream_buffer.h" #include "video_core/renderer_vulkan/vk_texture_cache.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" -#include "video_core/renderer_vulkan/wrapper.h" #include "video_core/shader/async_shaders.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Core { class System; @@ -49,67 +48,16 @@ namespace Vulkan { struct VKScreenInfo; -using ImageViewsPack = boost::container::static_vector<VkImageView, Maxwell::NumRenderTargets + 1>; - -struct FramebufferCacheKey { - VkRenderPass renderpass{}; - u32 width = 0; - u32 height = 0; - u32 layers = 0; - ImageViewsPack views; - - std::size_t Hash() const noexcept { - std::size_t hash = 0; - boost::hash_combine(hash, static_cast<VkRenderPass>(renderpass)); - for (const auto& view : views) { - boost::hash_combine(hash, static_cast<VkImageView>(view)); - } - boost::hash_combine(hash, width); - boost::hash_combine(hash, height); - boost::hash_combine(hash, layers); - return hash; - } - - bool operator==(const FramebufferCacheKey& rhs) const noexcept { - return std::tie(renderpass, views, width, height, layers) == - std::tie(rhs.renderpass, rhs.views, rhs.width, rhs.height, rhs.layers); - } - - bool operator!=(const FramebufferCacheKey& rhs) const noexcept { - return !operator==(rhs); - } -}; - -} // namespace Vulkan - -namespace std { - -template <> -struct hash<Vulkan::FramebufferCacheKey> { - std::size_t operator()(const Vulkan::FramebufferCacheKey& k) const noexcept { - return k.Hash(); - } -}; - -} // namespace std - -namespace Vulkan { - class StateTracker; class BufferBindings; -struct ImageView { - View view; - VkImageLayout* layout = nullptr; -}; - class RasterizerVulkan final : public VideoCore::RasterizerAccelerated { public: - explicit RasterizerVulkan(Core::Frontend::EmuWindow& emu_window, Tegra::GPU& gpu, - Tegra::MemoryManager& gpu_memory, Core::Memory::Memory& cpu_memory, - VKScreenInfo& screen_info, const VKDevice& device, - VKMemoryManager& memory_manager, StateTracker& state_tracker, - VKScheduler& scheduler); + explicit RasterizerVulkan(Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_, + Tegra::MemoryManager& gpu_memory_, Core::Memory::Memory& cpu_memory_, + VKScreenInfo& screen_info_, const Device& device_, + VKMemoryManager& memory_manager_, StateTracker& state_tracker_, + VKScheduler& scheduler_); ~RasterizerVulkan() override; void Draw(bool is_indexed, bool is_instanced) override; @@ -123,15 +71,18 @@ public: void InvalidateRegion(VAddr addr, u64 size) override; void OnCPUWrite(VAddr addr, u64 size) override; void SyncGuestHost() override; + void UnmapMemory(VAddr addr, u64 size) override; void SignalSemaphore(GPUVAddr addr, u32 value) override; void SignalSyncPoint(u32 value) override; void ReleaseFences() override; void FlushAndInvalidateRegion(VAddr addr, u64 size) override; void WaitForIdle() override; + void FragmentBarrier() override; + void TiledCacheBarrier() override; void FlushCommands() override; void TickFrame() override; - bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src, - const Tegra::Engines::Fermi2D::Regs::Surface& dst, + bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Surface& src, + const Tegra::Engines::Fermi2D::Surface& dst, const Tegra::Engines::Fermi2D::Config& copy_config) override; bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr, u32 pixel_stride) override; @@ -145,11 +96,17 @@ public: } /// Maximum supported size that a constbuffer can have in bytes. - static constexpr std::size_t MaxConstbufferSize = 0x10000; + static constexpr size_t MaxConstbufferSize = 0x10000; static_assert(MaxConstbufferSize % (4 * sizeof(float)) == 0, "The maximum size of a constbuffer must be a multiple of the size of GLvec4"); private: + static constexpr size_t MAX_TEXTURES = 192; + static constexpr size_t MAX_IMAGES = 48; + static constexpr size_t MAX_IMAGE_VIEWS = MAX_TEXTURES + MAX_IMAGES; + + static constexpr VkDeviceSize DEFAULT_BUFFER_SIZE = 4 * sizeof(float); + struct DrawParameters { void Draw(vk::CommandBuffer cmdbuf) const; @@ -160,20 +117,8 @@ private: bool is_indexed = 0; }; - using Texceptions = std::bitset<Maxwell::NumRenderTargets + 1>; - - static constexpr std::size_t ZETA_TEXCEPTION_INDEX = 8; - static constexpr VkDeviceSize DEFAULT_BUFFER_SIZE = 4 * sizeof(float); - void FlushWork(); - /// @brief Updates the currently bound attachments - /// @param is_clear True when the framebuffer is updated as a clear - /// @return Bitfield of attachments being used as sampled textures - Texceptions UpdateAttachments(bool is_clear); - - std::tuple<VkFramebuffer, VkExtent2D> ConfigureFramebuffers(VkRenderPass renderpass); - /// Setups geometry buffers and state. DrawParameters SetupGeometry(FixedPipelineState& fixed_state, BufferBindings& buffer_bindings, bool is_indexed, bool is_instanced); @@ -181,18 +126,12 @@ private: /// Setup descriptors in the graphics pipeline. void SetupShaderDescriptors(const std::array<Shader*, Maxwell::MaxShaderProgram>& shaders); - void SetupImageTransitions(Texceptions texceptions, - const std::array<View, Maxwell::NumRenderTargets>& color_attachments, - const View& zeta_attachment); - void UpdateDynamicStates(); void BeginTransformFeedback(); void EndTransformFeedback(); - bool WalkAttachmentOverlaps(const CachedSurfaceView& attachment); - void SetupVertexArrays(BufferBindings& buffer_bindings); void SetupIndexBuffer(BufferBindings& buffer_bindings, DrawParameters& params, bool is_indexed); @@ -238,14 +177,6 @@ private: void SetupGlobalBuffer(const GlobalBufferEntry& entry, GPUVAddr address); - void SetupUniformTexels(const Tegra::Texture::TICEntry& image, const UniformTexelEntry& entry); - - void SetupTexture(const Tegra::Texture::FullTextureInfo& texture, const SamplerEntry& entry); - - void SetupStorageTexel(const Tegra::Texture::TICEntry& tic, const StorageTexelEntry& entry); - - void SetupImage(const Tegra::Texture::TICEntry& tic, const ImageEntry& entry); - void UpdateViewportsState(Tegra::Engines::Maxwell3D::Regs& regs); void UpdateScissorsState(Tegra::Engines::Maxwell3D::Regs& regs); void UpdateDepthBias(Tegra::Engines::Maxwell3D::Regs& regs); @@ -259,22 +190,19 @@ private: void UpdateDepthWriteEnable(Tegra::Engines::Maxwell3D::Regs& regs); void UpdateDepthCompareOp(Tegra::Engines::Maxwell3D::Regs& regs); void UpdateFrontFace(Tegra::Engines::Maxwell3D::Regs& regs); - void UpdatePrimitiveTopology(Tegra::Engines::Maxwell3D::Regs& regs); void UpdateStencilOp(Tegra::Engines::Maxwell3D::Regs& regs); void UpdateStencilTestEnable(Tegra::Engines::Maxwell3D::Regs& regs); - std::size_t CalculateGraphicsStreamBufferSize(bool is_indexed) const; - - std::size_t CalculateComputeStreamBufferSize() const; + size_t CalculateGraphicsStreamBufferSize(bool is_indexed) const; - std::size_t CalculateVertexArraysSize() const; + size_t CalculateComputeStreamBufferSize() const; - std::size_t CalculateIndexBufferSize() const; + size_t CalculateVertexArraysSize() const; - std::size_t CalculateConstBufferSize(const ConstBufferEntry& entry, - const Tegra::Engines::ConstBufferInfo& buffer) const; + size_t CalculateIndexBufferSize() const; - RenderPassParams GetRenderPassParams(Texceptions texceptions) const; + size_t CalculateConstBufferSize(const ConstBufferEntry& entry, + const Tegra::Engines::ConstBufferInfo& buffer) const; VkBuffer DefaultBuffer(); @@ -284,23 +212,24 @@ private: Tegra::Engines::KeplerCompute& kepler_compute; VKScreenInfo& screen_info; - const VKDevice& device; + const Device& device; VKMemoryManager& memory_manager; StateTracker& state_tracker; VKScheduler& scheduler; + VKStreamBuffer stream_buffer; VKStagingBufferPool staging_pool; VKDescriptorPool descriptor_pool; VKUpdateDescriptorQueue update_descriptor_queue; - VKRenderPassCache renderpass_cache; + BlitImageHelper blit_image; QuadArrayPass quad_array_pass; QuadIndexedPass quad_indexed_pass; Uint8Pass uint8_pass; - VKTextureCache texture_cache; + TextureCacheRuntime texture_cache_runtime; + TextureCache texture_cache; VKPipelineCache pipeline_cache; VKBufferCache buffer_cache; - VKSamplerCache sampler_cache; VKQueryCache query_cache; VKFenceManager fence_manager; @@ -309,16 +238,11 @@ private: vk::Event wfi_event; VideoCommon::Shader::AsyncShaders async_shaders; - std::array<View, Maxwell::NumRenderTargets> color_attachments; - View zeta_attachment; - - std::vector<ImageView> sampled_views; - std::vector<ImageView> image_views; + boost::container::static_vector<u32, MAX_IMAGE_VIEWS> image_view_indices; + std::array<VideoCommon::ImageViewId, MAX_IMAGE_VIEWS> image_view_ids; + boost::container::static_vector<VkSampler, MAX_TEXTURES> sampler_handles; u32 draw_counter = 0; - - // TODO(Rodrigo): Invalidate on image destruction - std::unordered_map<FramebufferCacheKey, vk::Framebuffer> framebuffer_cache; }; } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp b/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp deleted file mode 100644 index 80284cf92..000000000 --- a/src/video_core/renderer_vulkan/vk_renderpass_cache.cpp +++ /dev/null @@ -1,158 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <cstring> -#include <memory> -#include <vector> - -#include "common/cityhash.h" -#include "video_core/engines/maxwell_3d.h" -#include "video_core/renderer_vulkan/maxwell_to_vk.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/vk_renderpass_cache.h" -#include "video_core/renderer_vulkan/wrapper.h" - -namespace Vulkan { - -std::size_t RenderPassParams::Hash() const noexcept { - const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), sizeof *this); - return static_cast<std::size_t>(hash); -} - -bool RenderPassParams::operator==(const RenderPassParams& rhs) const noexcept { - return std::memcmp(&rhs, this, sizeof *this) == 0; -} - -VKRenderPassCache::VKRenderPassCache(const VKDevice& device) : device{device} {} - -VKRenderPassCache::~VKRenderPassCache() = default; - -VkRenderPass VKRenderPassCache::GetRenderPass(const RenderPassParams& params) { - const auto [pair, is_cache_miss] = cache.try_emplace(params); - auto& entry = pair->second; - if (is_cache_miss) { - entry = CreateRenderPass(params); - } - return *entry; -} - -vk::RenderPass VKRenderPassCache::CreateRenderPass(const RenderPassParams& params) const { - using namespace VideoCore::Surface; - const std::size_t num_attachments = static_cast<std::size_t>(params.num_color_attachments); - - std::vector<VkAttachmentDescription> descriptors; - descriptors.reserve(num_attachments); - - std::vector<VkAttachmentReference> color_references; - color_references.reserve(num_attachments); - - for (std::size_t rt = 0; rt < num_attachments; ++rt) { - const auto guest_format = static_cast<Tegra::RenderTargetFormat>(params.color_formats[rt]); - const PixelFormat pixel_format = PixelFormatFromRenderTargetFormat(guest_format); - const auto format = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, pixel_format); - ASSERT_MSG(format.attachable, "Trying to attach a non-attachable format with format={}", - static_cast<int>(pixel_format)); - - // TODO(Rodrigo): Add MAY_ALIAS_BIT when it's needed. - const VkImageLayout color_layout = ((params.texceptions >> rt) & 1) != 0 - ? VK_IMAGE_LAYOUT_GENERAL - : VK_IMAGE_LAYOUT_COLOR_ATTACHMENT_OPTIMAL; - descriptors.push_back({ - .flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT, - .format = format.format, - .samples = VK_SAMPLE_COUNT_1_BIT, - .loadOp = VK_ATTACHMENT_LOAD_OP_LOAD, - .storeOp = VK_ATTACHMENT_STORE_OP_STORE, - .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_DONT_CARE, - .stencilStoreOp = VK_ATTACHMENT_STORE_OP_DONT_CARE, - .initialLayout = color_layout, - .finalLayout = color_layout, - }); - - color_references.push_back({ - .attachment = static_cast<u32>(rt), - .layout = color_layout, - }); - } - - VkAttachmentReference zeta_attachment_ref; - const bool has_zeta = params.zeta_format != 0; - if (has_zeta) { - const auto guest_format = static_cast<Tegra::DepthFormat>(params.zeta_format); - const PixelFormat pixel_format = PixelFormatFromDepthFormat(guest_format); - const auto format = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, pixel_format); - ASSERT_MSG(format.attachable, "Trying to attach a non-attachable format with format={}", - static_cast<int>(pixel_format)); - - const VkImageLayout zeta_layout = params.zeta_texception != 0 - ? VK_IMAGE_LAYOUT_GENERAL - : VK_IMAGE_LAYOUT_DEPTH_STENCIL_ATTACHMENT_OPTIMAL; - descriptors.push_back({ - .flags = 0, - .format = format.format, - .samples = VK_SAMPLE_COUNT_1_BIT, - .loadOp = VK_ATTACHMENT_LOAD_OP_LOAD, - .storeOp = VK_ATTACHMENT_STORE_OP_STORE, - .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD, - .stencilStoreOp = VK_ATTACHMENT_STORE_OP_STORE, - .initialLayout = zeta_layout, - .finalLayout = zeta_layout, - }); - - zeta_attachment_ref = { - .attachment = static_cast<u32>(num_attachments), - .layout = zeta_layout, - }; - } - - const VkSubpassDescription subpass_description{ - .flags = 0, - .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS, - .inputAttachmentCount = 0, - .pInputAttachments = nullptr, - .colorAttachmentCount = static_cast<u32>(color_references.size()), - .pColorAttachments = color_references.data(), - .pResolveAttachments = nullptr, - .pDepthStencilAttachment = has_zeta ? &zeta_attachment_ref : nullptr, - .preserveAttachmentCount = 0, - .pPreserveAttachments = nullptr, - }; - - VkAccessFlags access = 0; - VkPipelineStageFlags stage = 0; - if (!color_references.empty()) { - access |= VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT; - stage |= VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT; - } - - if (has_zeta) { - access |= VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | - VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT; - stage |= VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT; - } - - const VkSubpassDependency subpass_dependency{ - .srcSubpass = VK_SUBPASS_EXTERNAL, - .dstSubpass = 0, - .srcStageMask = stage, - .dstStageMask = stage, - .srcAccessMask = 0, - .dstAccessMask = access, - .dependencyFlags = 0, - }; - - return device.GetLogical().CreateRenderPass({ - .sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO, - .pNext = nullptr, - .flags = 0, - .attachmentCount = static_cast<u32>(descriptors.size()), - .pAttachments = descriptors.data(), - .subpassCount = 1, - .pSubpasses = &subpass_description, - .dependencyCount = 1, - .pDependencies = &subpass_dependency, - }); -} - -} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_renderpass_cache.h b/src/video_core/renderer_vulkan/vk_renderpass_cache.h deleted file mode 100644 index 8b0fec720..000000000 --- a/src/video_core/renderer_vulkan/vk_renderpass_cache.h +++ /dev/null @@ -1,70 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <type_traits> -#include <unordered_map> - -#include <boost/container/static_vector.hpp> -#include <boost/functional/hash.hpp> - -#include "video_core/engines/maxwell_3d.h" -#include "video_core/renderer_vulkan/wrapper.h" -#include "video_core/surface.h" - -namespace Vulkan { - -class VKDevice; - -struct RenderPassParams { - std::array<u8, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> color_formats; - u8 num_color_attachments; - u8 texceptions; - - u8 zeta_format; - u8 zeta_texception; - - std::size_t Hash() const noexcept; - - bool operator==(const RenderPassParams& rhs) const noexcept; - - bool operator!=(const RenderPassParams& rhs) const noexcept { - return !operator==(rhs); - } -}; -static_assert(std::has_unique_object_representations_v<RenderPassParams>); -static_assert(std::is_trivially_copyable_v<RenderPassParams>); -static_assert(std::is_trivially_constructible_v<RenderPassParams>); - -} // namespace Vulkan - -namespace std { - -template <> -struct hash<Vulkan::RenderPassParams> { - std::size_t operator()(const Vulkan::RenderPassParams& k) const noexcept { - return k.Hash(); - } -}; - -} // namespace std - -namespace Vulkan { - -class VKRenderPassCache final { -public: - explicit VKRenderPassCache(const VKDevice& device); - ~VKRenderPassCache(); - - VkRenderPass GetRenderPass(const RenderPassParams& params); - -private: - vk::RenderPass CreateRenderPass(const RenderPassParams& params) const; - - const VKDevice& device; - std::unordered_map<RenderPassParams, vk::RenderPass> cache; -}; - -} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp b/src/video_core/renderer_vulkan/vk_sampler_cache.cpp deleted file mode 100644 index b068888f9..000000000 --- a/src/video_core/renderer_vulkan/vk_sampler_cache.cpp +++ /dev/null @@ -1,83 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <unordered_map> - -#include "video_core/renderer_vulkan/maxwell_to_vk.h" -#include "video_core/renderer_vulkan/vk_sampler_cache.h" -#include "video_core/renderer_vulkan/wrapper.h" -#include "video_core/textures/texture.h" - -using Tegra::Texture::TextureMipmapFilter; - -namespace Vulkan { - -namespace { - -VkBorderColor ConvertBorderColor(std::array<float, 4> color) { - // TODO(Rodrigo): Manage integer border colors - if (color == std::array<float, 4>{0, 0, 0, 0}) { - return VK_BORDER_COLOR_FLOAT_TRANSPARENT_BLACK; - } else if (color == std::array<float, 4>{0, 0, 0, 1}) { - return VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK; - } else if (color == std::array<float, 4>{1, 1, 1, 1}) { - return VK_BORDER_COLOR_FLOAT_OPAQUE_WHITE; - } - if (color[0] + color[1] + color[2] > 1.35f) { - // If color elements are brighter than roughly 0.5 average, use white border - return VK_BORDER_COLOR_FLOAT_OPAQUE_WHITE; - } else if (color[3] > 0.5f) { - return VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK; - } else { - return VK_BORDER_COLOR_FLOAT_TRANSPARENT_BLACK; - } -} - -} // Anonymous namespace - -VKSamplerCache::VKSamplerCache(const VKDevice& device) : device{device} {} - -VKSamplerCache::~VKSamplerCache() = default; - -vk::Sampler VKSamplerCache::CreateSampler(const Tegra::Texture::TSCEntry& tsc) const { - const bool arbitrary_borders = device.IsExtCustomBorderColorSupported(); - const std::array color = tsc.GetBorderColor(); - - VkSamplerCustomBorderColorCreateInfoEXT border{ - .sType = VK_STRUCTURE_TYPE_SAMPLER_CUSTOM_BORDER_COLOR_CREATE_INFO_EXT, - .pNext = nullptr, - .customBorderColor = {}, - .format = VK_FORMAT_UNDEFINED, - }; - std::memcpy(&border.customBorderColor, color.data(), sizeof(color)); - - return device.GetLogical().CreateSampler({ - .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO, - .pNext = arbitrary_borders ? &border : nullptr, - .flags = 0, - .magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter), - .minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter), - .mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter), - .addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter), - .addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter), - .addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter), - .mipLodBias = tsc.GetLodBias(), - .anisotropyEnable = - static_cast<VkBool32>(tsc.GetMaxAnisotropy() > 1.0f ? VK_TRUE : VK_FALSE), - .maxAnisotropy = tsc.GetMaxAnisotropy(), - .compareEnable = tsc.depth_compare_enabled, - .compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func), - .minLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.0f : tsc.GetMinLod(), - .maxLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.25f : tsc.GetMaxLod(), - .borderColor = - arbitrary_borders ? VK_BORDER_COLOR_INT_CUSTOM_EXT : ConvertBorderColor(color), - .unnormalizedCoordinates = VK_FALSE, - }); -} - -VkSampler VKSamplerCache::ToSamplerType(const vk::Sampler& sampler) const { - return *sampler; -} - -} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_sampler_cache.h b/src/video_core/renderer_vulkan/vk_sampler_cache.h deleted file mode 100644 index a33d1c0ee..000000000 --- a/src/video_core/renderer_vulkan/vk_sampler_cache.h +++ /dev/null @@ -1,29 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "video_core/renderer_vulkan/wrapper.h" -#include "video_core/sampler_cache.h" -#include "video_core/textures/texture.h" - -namespace Vulkan { - -class VKDevice; - -class VKSamplerCache final : public VideoCommon::SamplerCache<VkSampler, vk::Sampler> { -public: - explicit VKSamplerCache(const VKDevice& device); - ~VKSamplerCache(); - -protected: - vk::Sampler CreateSampler(const Tegra::Texture::TSCEntry& tsc) const override; - - VkSampler ToSamplerType(const vk::Sampler& sampler) const override; - -private: - const VKDevice& device; -}; - -} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_scheduler.cpp b/src/video_core/renderer_vulkan/vk_scheduler.cpp index 1a483dc71..66004f9c0 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.cpp +++ b/src/video_core/renderer_vulkan/vk_scheduler.cpp @@ -11,12 +11,13 @@ #include "common/microprofile.h" #include "common/thread.h" #include "video_core/renderer_vulkan/vk_command_pool.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_master_semaphore.h" #include "video_core/renderer_vulkan/vk_query_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_state_tracker.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/renderer_vulkan/vk_texture_cache.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -36,7 +37,7 @@ void VKScheduler::CommandChunk::ExecuteAll(vk::CommandBuffer cmdbuf) { last = nullptr; } -VKScheduler::VKScheduler(const VKDevice& device_, StateTracker& state_tracker_) +VKScheduler::VKScheduler(const Device& device_, StateTracker& state_tracker_) : device{device_}, state_tracker{state_tracker_}, master_semaphore{std::make_unique<MasterSemaphore>(device)}, command_pool{std::make_unique<CommandPool>(*master_semaphore, device)} { @@ -96,38 +97,39 @@ void VKScheduler::DispatchWork() { AcquireNewChunk(); } -void VKScheduler::RequestRenderpass(VkRenderPass renderpass, VkFramebuffer framebuffer, - VkExtent2D render_area) { - if (renderpass == state.renderpass && framebuffer == state.framebuffer && +void VKScheduler::RequestRenderpass(const Framebuffer* framebuffer) { + const VkRenderPass renderpass = framebuffer->RenderPass(); + const VkFramebuffer framebuffer_handle = framebuffer->Handle(); + const VkExtent2D render_area = framebuffer->RenderArea(); + if (renderpass == state.renderpass && framebuffer_handle == state.framebuffer && render_area.width == state.render_area.width && render_area.height == state.render_area.height) { return; } - const bool end_renderpass = state.renderpass != nullptr; + EndRenderPass(); state.renderpass = renderpass; - state.framebuffer = framebuffer; + state.framebuffer = framebuffer_handle; state.render_area = render_area; - const VkRenderPassBeginInfo renderpass_bi{ - .sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO, - .pNext = nullptr, - .renderPass = renderpass, - .framebuffer = framebuffer, - .renderArea = - { - .offset = {.x = 0, .y = 0}, - .extent = render_area, - }, - .clearValueCount = 0, - .pClearValues = nullptr, - }; - - Record([renderpass_bi, end_renderpass](vk::CommandBuffer cmdbuf) { - if (end_renderpass) { - cmdbuf.EndRenderPass(); - } + Record([renderpass, framebuffer_handle, render_area](vk::CommandBuffer cmdbuf) { + const VkRenderPassBeginInfo renderpass_bi{ + .sType = VK_STRUCTURE_TYPE_RENDER_PASS_BEGIN_INFO, + .pNext = nullptr, + .renderPass = renderpass, + .framebuffer = framebuffer_handle, + .renderArea = + { + .offset = {.x = 0, .y = 0}, + .extent = render_area, + }, + .clearValueCount = 0, + .pClearValues = nullptr, + }; cmdbuf.BeginRenderPass(renderpass_bi, VK_SUBPASS_CONTENTS_INLINE); }); + num_renderpass_images = framebuffer->NumImages(); + renderpass_images = framebuffer->Images(); + renderpass_image_ranges = framebuffer->ImageRanges(); } void VKScheduler::RequestOutsideRenderPassOperationContext() { @@ -241,8 +243,37 @@ void VKScheduler::EndRenderPass() { if (!state.renderpass) { return; } + Record([num_images = num_renderpass_images, images = renderpass_images, + ranges = renderpass_image_ranges](vk::CommandBuffer cmdbuf) { + std::array<VkImageMemoryBarrier, 9> barriers; + for (size_t i = 0; i < num_images; ++i) { + barriers[i] = VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT, + .dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | + VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = images[i], + .subresourceRange = ranges[i], + }; + } + cmdbuf.EndRenderPass(); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_EARLY_FRAGMENT_TESTS_BIT | + VK_PIPELINE_STAGE_LATE_FRAGMENT_TESTS_BIT | + VK_PIPELINE_STAGE_COLOR_ATTACHMENT_OUTPUT_BIT, + VK_PIPELINE_STAGE_ALL_GRAPHICS_BIT, 0, nullptr, nullptr, + vk::Span(barriers.data(), num_images)); + }); state.renderpass = nullptr; - Record([](vk::CommandBuffer cmdbuf) { cmdbuf.EndRenderPass(); }); + num_renderpass_images = 0; } void VKScheduler::AcquireNewChunk() { diff --git a/src/video_core/renderer_vulkan/vk_scheduler.h b/src/video_core/renderer_vulkan/vk_scheduler.h index 7be8a19f0..4cd43e425 100644 --- a/src/video_core/renderer_vulkan/vk_scheduler.h +++ b/src/video_core/renderer_vulkan/vk_scheduler.h @@ -12,21 +12,22 @@ #include <utility> #include "common/common_types.h" #include "common/threadsafe_queue.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { class CommandPool; +class Device; +class Framebuffer; class MasterSemaphore; class StateTracker; -class VKDevice; class VKQueryCache; /// The scheduler abstracts command buffer and fence management with an interface that's able to do /// OpenGL-like operations on Vulkan command buffers. class VKScheduler { public: - explicit VKScheduler(const VKDevice& device, StateTracker& state_tracker); + explicit VKScheduler(const Device& device, StateTracker& state_tracker); ~VKScheduler(); /// Returns the current command buffer tick. @@ -52,8 +53,7 @@ public: void DispatchWork(); /// Requests to begin a renderpass. - void RequestRenderpass(VkRenderPass renderpass, VkFramebuffer framebuffer, - VkExtent2D render_area); + void RequestRenderpass(const Framebuffer* framebuffer); /// Requests the current executino context to be able to execute operations only allowed outside /// of a renderpass. @@ -62,6 +62,9 @@ public: /// Binds a pipeline to the current execution context. void BindGraphicsPipeline(VkPipeline pipeline); + /// Invalidates current command buffer state except for render passes + void InvalidateState(); + /// Assigns the query cache. void SetQueryCache(VKQueryCache& query_cache_) { query_cache = &query_cache_; @@ -104,7 +107,7 @@ private: template <typename T> class TypedCommand final : public Command { public: - explicit TypedCommand(T&& command) : command{std::move(command)} {} + explicit TypedCommand(T&& command_) : command{std::move(command_)} {} ~TypedCommand() override = default; TypedCommand(TypedCommand&&) = delete; @@ -170,15 +173,13 @@ private: void AllocateNewContext(); - void InvalidateState(); - void EndPendingOperations(); void EndRenderPass(); void AcquireNewChunk(); - const VKDevice& device; + const Device& device; StateTracker& state_tracker; std::unique_ptr<MasterSemaphore> master_semaphore; @@ -192,6 +193,11 @@ private: std::thread worker_thread; State state; + + u32 num_renderpass_images = 0; + std::array<VkImage, 9> renderpass_images{}; + std::array<VkImageSubresourceRange, 9> renderpass_image_ranges{}; + Common::SPSCQueue<std::unique_ptr<CommandChunk>> chunk_queue; Common::SPSCQueue<std::unique_ptr<CommandChunk>> chunk_reserve; std::mutex mutex; diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp index cd7d7a4e4..89cbe01ad 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp @@ -22,11 +22,11 @@ #include "video_core/engines/shader_bytecode.h" #include "video_core/engines/shader_header.h" #include "video_core/engines/shader_type.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_shader_decompiler.h" #include "video_core/shader/node.h" #include "video_core/shader/shader_ir.h" #include "video_core/shader/transform_feedback.h" +#include "video_core/vulkan_common/vulkan_device.h" namespace Vulkan { @@ -55,8 +55,8 @@ enum class Type { Void, Bool, Bool2, Float, Int, Uint, HalfFloat }; class Expression final { public: - Expression(Id id, Type type) : id{id}, type{type} { - ASSERT(type != Type::Void); + Expression(Id id_, Type type_) : id{id_}, type{type_} { + ASSERT(type_ != Type::Void); } Expression() : type{Type::Void} {} @@ -102,7 +102,7 @@ struct GenericVaryingDescription { bool is_scalar = false; }; -spv::Dim GetSamplerDim(const Sampler& sampler) { +spv::Dim GetSamplerDim(const SamplerEntry& sampler) { ASSERT(!sampler.is_buffer); switch (sampler.type) { case Tegra::Shader::TextureType::Texture1D: @@ -114,12 +114,12 @@ spv::Dim GetSamplerDim(const Sampler& sampler) { case Tegra::Shader::TextureType::TextureCube: return spv::Dim::Cube; default: - UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<int>(sampler.type)); + UNIMPLEMENTED_MSG("Unimplemented sampler type={}", sampler.type); return spv::Dim::Dim2D; } } -std::pair<spv::Dim, bool> GetImageDim(const Image& image) { +std::pair<spv::Dim, bool> GetImageDim(const ImageEntry& image) { switch (image.type) { case Tegra::Shader::ImageType::Texture1D: return {spv::Dim::Dim1D, false}; @@ -134,7 +134,7 @@ std::pair<spv::Dim, bool> GetImageDim(const Image& image) { case Tegra::Shader::ImageType::Texture3D: return {spv::Dim::Dim3D, false}; default: - UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<int>(image.type)); + UNIMPLEMENTED_MSG("Unimplemented image type={}", image.type); return {spv::Dim::Dim2D, false}; } } @@ -274,12 +274,12 @@ bool IsPrecise(Operation operand) { class SPIRVDecompiler final : public Sirit::Module { public: - explicit SPIRVDecompiler(const VKDevice& device, const ShaderIR& ir, ShaderType stage, - const Registry& registry, const Specialization& specialization) - : Module(0x00010300), device{device}, ir{ir}, stage{stage}, header{ir.GetHeader()}, - registry{registry}, specialization{specialization} { - if (stage != ShaderType::Compute) { - transform_feedback = BuildTransformFeedback(registry.GetGraphicsInfo()); + explicit SPIRVDecompiler(const Device& device_, const ShaderIR& ir_, ShaderType stage_, + const Registry& registry_, const Specialization& specialization_) + : Module(0x00010300), device{device_}, ir{ir_}, stage{stage_}, header{ir_.GetHeader()}, + registry{registry_}, specialization{specialization_} { + if (stage_ != ShaderType::Compute) { + transform_feedback = BuildTransformFeedback(registry_.GetGraphicsInfo()); } AddCapability(spv::Capability::Shader); @@ -293,6 +293,7 @@ public: AddCapability(spv::Capability::DrawParameters); AddCapability(spv::Capability::SubgroupBallotKHR); AddCapability(spv::Capability::SubgroupVoteKHR); + AddExtension("SPV_KHR_16bit_storage"); AddExtension("SPV_KHR_shader_ballot"); AddExtension("SPV_KHR_subgroup_vote"); AddExtension("SPV_KHR_storage_buffer_storage_class"); @@ -307,7 +308,6 @@ public: "supported on this device"); } } - if (ir.UsesLayer() || ir.UsesViewportIndex()) { if (ir.UsesViewportIndex()) { AddCapability(spv::Capability::MultiViewport); @@ -317,15 +317,13 @@ public: AddCapability(spv::Capability::ShaderViewportIndexLayerEXT); } } - if (device.IsFormatlessImageLoadSupported()) { AddCapability(spv::Capability::StorageImageReadWithoutFormat); } - if (device.IsFloat16Supported()) { AddCapability(spv::Capability::Float16); } - t_scalar_half = Name(TypeFloat(device.IsFloat16Supported() ? 16 : 32), "scalar_half"); + t_scalar_half = Name(TypeFloat(device_.IsFloat16Supported() ? 16 : 32), "scalar_half"); t_half = Name(TypeVector(t_scalar_half, 2), "half"); const Id main = Decompile(); @@ -369,6 +367,9 @@ public: if (header.ps.omap.depth) { AddExecutionMode(main, spv::ExecutionMode::DepthReplacing); } + if (specialization.early_fragment_tests) { + AddExecutionMode(main, spv::ExecutionMode::EarlyFragmentTests); + } break; case ShaderType::Compute: const auto workgroup_size = specialization.workgroup_size; @@ -972,7 +973,7 @@ private: return binding; } - void DeclareImage(const Image& image, u32& binding) { + void DeclareImage(const ImageEntry& image, u32& binding) { const auto [dim, arrayed] = GetImageDim(image); constexpr int depth = 0; constexpr bool ms = false; @@ -1080,9 +1081,9 @@ private: indices.point_size = AddBuiltIn(t_float, spv::BuiltIn::PointSize, "point_size"); } - const auto& output_attributes = ir.GetOutputAttributes(); - const bool declare_clip_distances = - std::any_of(output_attributes.begin(), output_attributes.end(), [](const auto& index) { + const auto& ir_output_attributes = ir.GetOutputAttributes(); + const bool declare_clip_distances = std::any_of( + ir_output_attributes.begin(), ir_output_attributes.end(), [](const auto& index) { return index == Attribute::Index::ClipDistances0123 || index == Attribute::Index::ClipDistances4567; }); @@ -1246,7 +1247,7 @@ private: const Id pointer = ArrayPass(type_descriptor.scalar, attribute_id, elements); return {OpLoad(GetTypeDefinition(type), pointer), type}; } - UNIMPLEMENTED_MSG("Unhandled input attribute: {}", static_cast<u32>(attribute)); + UNIMPLEMENTED_MSG("Unhandled input attribute: {}", attribute); return {v_float_zero, Type::Float}; } @@ -1882,7 +1883,7 @@ private: case Tegra::Shader::TextureType::Texture3D: return 3; default: - UNREACHABLE_MSG("Invalid texture type={}", static_cast<int>(type)); + UNREACHABLE_MSG("Invalid texture type={}", type); return 2; } }(); @@ -2067,6 +2068,46 @@ private: return {}; } + Id MaxwellToSpirvComparison(Maxwell::ComparisonOp compare_op, Id operand_1, Id operand_2) { + using Compare = Maxwell::ComparisonOp; + switch (compare_op) { + case Compare::NeverOld: + return v_false; // Never let the test pass + case Compare::LessOld: + return OpFOrdLessThan(t_bool, operand_1, operand_2); + case Compare::EqualOld: + return OpFOrdEqual(t_bool, operand_1, operand_2); + case Compare::LessEqualOld: + return OpFOrdLessThanEqual(t_bool, operand_1, operand_2); + case Compare::GreaterOld: + return OpFOrdGreaterThan(t_bool, operand_1, operand_2); + case Compare::NotEqualOld: + return OpFOrdNotEqual(t_bool, operand_1, operand_2); + case Compare::GreaterEqualOld: + return OpFOrdGreaterThanEqual(t_bool, operand_1, operand_2); + default: + UNREACHABLE(); + return v_true; + } + } + + void AlphaTest(Id pointer) { + if (specialization.alpha_test_func == Maxwell::ComparisonOp::AlwaysOld) { + return; + } + const Id true_label = OpLabel(); + const Id discard_label = OpLabel(); + const Id alpha_reference = Constant(t_float, specialization.alpha_test_ref); + const Id alpha_value = OpLoad(t_float, pointer); + const Id condition = + MaxwellToSpirvComparison(specialization.alpha_test_func, alpha_value, alpha_reference); + + OpBranchConditional(condition, true_label, discard_label); + AddLabel(discard_label); + OpKill(); + AddLabel(true_label); + } + void PreExit() { if (stage == ShaderType::Vertex && specialization.ndc_minus_one_to_one) { const u32 position_index = out_indices.position.value(); @@ -2078,8 +2119,7 @@ private: OpStore(z_pointer, depth); } if (stage == ShaderType::Fragment) { - const auto SafeGetRegister = [&](u32 reg) { - // TODO(Rodrigo): Replace with contains once C++20 releases + const auto SafeGetRegister = [this](u32 reg) { if (const auto it = registers.find(reg); it != registers.end()) { return OpLoad(t_float, it->second); } @@ -2089,8 +2129,6 @@ private: UNIMPLEMENTED_IF_MSG(header.ps.omap.sample_mask != 0, "Sample mask write is unimplemented"); - // TODO(Rodrigo): Alpha testing - // Write the color outputs using the data in the shader registers, disabled // rendertargets/components are skipped in the register assignment. u32 current_reg = 0; @@ -2102,6 +2140,9 @@ private: } const Id pointer = AccessElement(t_out_float, frag_colors[rt], component); OpStore(pointer, SafeGetRegister(current_reg)); + if (rt == 0 && component == 3) { + AlphaTest(pointer); + } ++current_reg; } } @@ -2701,7 +2742,7 @@ private: }; static_assert(operation_decompilers.size() == static_cast<std::size_t>(OperationCode::Amount)); - const VKDevice& device; + const Device& device; const ShaderIR& ir; const ShaderType stage; const Tegra::Shader::Header header; @@ -2843,7 +2884,7 @@ private: class ExprDecompiler { public: - explicit ExprDecompiler(SPIRVDecompiler& decomp) : decomp{decomp} {} + explicit ExprDecompiler(SPIRVDecompiler& decomp_) : decomp{decomp_} {} Id operator()(const ExprAnd& expr) { const Id type_def = decomp.GetTypeDefinition(Type::Bool); @@ -2899,7 +2940,7 @@ private: class ASTDecompiler { public: - explicit ASTDecompiler(SPIRVDecompiler& decomp) : decomp{decomp} {} + explicit ASTDecompiler(SPIRVDecompiler& decomp_) : decomp{decomp_} {} void operator()(const ASTProgram& ast) { ASTNode current = ast.nodes.GetFirst(); @@ -3089,7 +3130,7 @@ ShaderEntries GenerateShaderEntries(const VideoCommon::Shader::ShaderIR& ir) { return entries; } -std::vector<u32> Decompile(const VKDevice& device, const VideoCommon::Shader::ShaderIR& ir, +std::vector<u32> Decompile(const Device& device, const VideoCommon::Shader::ShaderIR& ir, ShaderType stage, const VideoCommon::Shader::Registry& registry, const Specialization& specialization) { return SPIRVDecompiler(device, ir, stage, registry, specialization).Assemble(); diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.h b/src/video_core/renderer_vulkan/vk_shader_decompiler.h index 2b0e90396..26381e444 100644 --- a/src/video_core/renderer_vulkan/vk_shader_decompiler.h +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.h @@ -15,23 +15,21 @@ #include "video_core/shader/shader_ir.h" namespace Vulkan { -class VKDevice; -} -namespace Vulkan { +class Device; using Maxwell = Tegra::Engines::Maxwell3D::Regs; -using UniformTexelEntry = VideoCommon::Shader::Sampler; -using SamplerEntry = VideoCommon::Shader::Sampler; -using StorageTexelEntry = VideoCommon::Shader::Image; -using ImageEntry = VideoCommon::Shader::Image; +using UniformTexelEntry = VideoCommon::Shader::SamplerEntry; +using SamplerEntry = VideoCommon::Shader::SamplerEntry; +using StorageTexelEntry = VideoCommon::Shader::ImageEntry; +using ImageEntry = VideoCommon::Shader::ImageEntry; constexpr u32 DESCRIPTOR_SET = 0; class ConstBufferEntry : public VideoCommon::Shader::ConstBuffer { public: - explicit constexpr ConstBufferEntry(const VideoCommon::Shader::ConstBuffer& entry, u32 index) - : VideoCommon::Shader::ConstBuffer{entry}, index{index} {} + explicit constexpr ConstBufferEntry(const ConstBuffer& entry_, u32 index_) + : ConstBuffer{entry_}, index{index_} {} constexpr u32 GetIndex() const { return index; @@ -43,8 +41,8 @@ private: class GlobalBufferEntry { public: - constexpr explicit GlobalBufferEntry(u32 cbuf_index, u32 cbuf_offset, bool is_written) - : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset}, is_written{is_written} {} + constexpr explicit GlobalBufferEntry(u32 cbuf_index_, u32 cbuf_offset_, bool is_written_) + : cbuf_index{cbuf_index_}, cbuf_offset{cbuf_offset_}, is_written{is_written_} {} constexpr u32 GetCbufIndex() const { return cbuf_index; @@ -95,6 +93,9 @@ struct Specialization final { std::bitset<Maxwell::NumVertexAttributes> enabled_attributes; std::array<Maxwell::VertexAttribute::Type, Maxwell::NumVertexAttributes> attribute_types{}; bool ndc_minus_one_to_one{}; + bool early_fragment_tests{}; + float alpha_test_ref{}; + Maxwell::ComparisonOp alpha_test_func{}; }; // Old gcc versions don't consider this trivially copyable. // static_assert(std::is_trivially_copyable_v<Specialization>); @@ -106,7 +107,7 @@ struct SPIRVShader { ShaderEntries GenerateShaderEntries(const VideoCommon::Shader::ShaderIR& ir); -std::vector<u32> Decompile(const VKDevice& device, const VideoCommon::Shader::ShaderIR& ir, +std::vector<u32> Decompile(const Device& device, const VideoCommon::Shader::ShaderIR& ir, Tegra::Engines::ShaderType stage, const VideoCommon::Shader::Registry& registry, const Specialization& specialization); diff --git a/src/video_core/renderer_vulkan/vk_shader_util.cpp b/src/video_core/renderer_vulkan/vk_shader_util.cpp index c1a218d76..aaad4f292 100644 --- a/src/video_core/renderer_vulkan/vk_shader_util.cpp +++ b/src/video_core/renderer_vulkan/vk_shader_util.cpp @@ -7,24 +7,19 @@ #include "common/assert.h" #include "common/common_types.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_shader_util.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -vk::ShaderModule BuildShader(const VKDevice& device, std::size_t code_size, const u8* code_data) { - // Avoid undefined behavior by copying to a staging allocation - ASSERT(code_size % sizeof(u32) == 0); - const auto data = std::make_unique<u32[]>(code_size / sizeof(u32)); - std::memcpy(data.get(), code_data, code_size); - +vk::ShaderModule BuildShader(const Device& device, std::span<const u32> code) { return device.GetLogical().CreateShaderModule({ .sType = VK_STRUCTURE_TYPE_SHADER_MODULE_CREATE_INFO, .pNext = nullptr, .flags = 0, - .codeSize = code_size, - .pCode = data.get(), + .codeSize = static_cast<u32>(code.size_bytes()), + .pCode = code.data(), }); } diff --git a/src/video_core/renderer_vulkan/vk_shader_util.h b/src/video_core/renderer_vulkan/vk_shader_util.h index d1d3f3cae..9517cbe84 100644 --- a/src/video_core/renderer_vulkan/vk_shader_util.h +++ b/src/video_core/renderer_vulkan/vk_shader_util.h @@ -4,13 +4,15 @@ #pragma once +#include <span> + #include "common/common_types.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; -vk::ShaderModule BuildShader(const VKDevice& device, std::size_t code_size, const u8* code_data); +vk::ShaderModule BuildShader(const Device& device, std::span<const u32> code); } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp index 2fd3b7f39..1e0b8b922 100644 --- a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp +++ b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.cpp @@ -9,17 +9,17 @@ #include "common/bit_util.h" #include "common/common_types.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { VKStagingBufferPool::StagingBuffer::StagingBuffer(std::unique_ptr<VKBuffer> buffer_) : buffer{std::move(buffer_)} {} -VKStagingBufferPool::VKStagingBufferPool(const VKDevice& device_, VKMemoryManager& memory_manager_, +VKStagingBufferPool::VKStagingBufferPool(const Device& device_, VKMemoryManager& memory_manager_, VKScheduler& scheduler_) : device{device_}, memory_manager{memory_manager_}, scheduler{scheduler_} {} diff --git a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h index 2dd5049ac..90dadcbbe 100644 --- a/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h +++ b/src/video_core/renderer_vulkan/vk_staging_buffer_pool.h @@ -10,11 +10,11 @@ #include "common/common_types.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class VKScheduler; struct VKBuffer final { @@ -24,7 +24,7 @@ struct VKBuffer final { class VKStagingBufferPool final { public: - explicit VKStagingBufferPool(const VKDevice& device, VKMemoryManager& memory_manager, + explicit VKStagingBufferPool(const Device& device, VKMemoryManager& memory_manager, VKScheduler& scheduler); ~VKStagingBufferPool(); @@ -58,7 +58,7 @@ private: u64 ReleaseLevel(StagingBuffersCache& cache, std::size_t log2); - const VKDevice& device; + const Device& device; VKMemoryManager& memory_manager; VKScheduler& scheduler; diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.cpp b/src/video_core/renderer_vulkan/vk_state_tracker.cpp index 5d2c4a796..1779a2e30 100644 --- a/src/video_core/renderer_vulkan/vk_state_tracker.cpp +++ b/src/video_core/renderer_vulkan/vk_state_tracker.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include <algorithm> +#include <array> #include <cstddef> #include <iterator> @@ -14,7 +15,7 @@ #include "video_core/renderer_vulkan/vk_state_tracker.h" #define OFF(field_name) MAXWELL3D_REG_INDEX(field_name) -#define NUM(field_name) (sizeof(Maxwell3D::Regs::field_name) / sizeof(u32)) +#define NUM(field_name) (sizeof(Maxwell3D::Regs::field_name) / (sizeof(u32))) namespace Vulkan { @@ -29,21 +30,15 @@ using Table = Maxwell3D::DirtyState::Table; using Flags = Maxwell3D::DirtyState::Flags; Flags MakeInvalidationFlags() { + static constexpr std::array INVALIDATION_FLAGS{ + Viewports, Scissors, DepthBias, BlendConstants, DepthBounds, + StencilProperties, CullMode, DepthBoundsEnable, DepthTestEnable, DepthWriteEnable, + DepthCompareOp, FrontFace, StencilOp, StencilTestEnable, + }; Flags flags{}; - flags[Viewports] = true; - flags[Scissors] = true; - flags[DepthBias] = true; - flags[BlendConstants] = true; - flags[DepthBounds] = true; - flags[StencilProperties] = true; - flags[CullMode] = true; - flags[DepthBoundsEnable] = true; - flags[DepthTestEnable] = true; - flags[DepthWriteEnable] = true; - flags[DepthCompareOp] = true; - flags[FrontFace] = true; - flags[StencilOp] = true; - flags[StencilTestEnable] = true; + for (const int flag : INVALIDATION_FLAGS) { + flags[flag] = true; + } return flags; } diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.h b/src/video_core/renderer_vulkan/vk_state_tracker.h index 1de789e57..c335d2bdf 100644 --- a/src/video_core/renderer_vulkan/vk_state_tracker.h +++ b/src/video_core/renderer_vulkan/vk_state_tracker.h @@ -52,6 +52,14 @@ public: current_topology = INVALID_TOPOLOGY; } + void InvalidateViewports() { + flags[Dirty::Viewports] = true; + } + + void InvalidateScissors() { + flags[Dirty::Scissors] = true; + } + bool TouchViewports() { return Exchange(Dirty::Viewports, false); } diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp index 5218c875b..a09fe084e 100644 --- a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp +++ b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp @@ -10,15 +10,19 @@ #include "common/alignment.h" #include "common/assert.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_stream_buffer.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { namespace { +constexpr VkBufferUsageFlags BUFFER_USAGE = + VK_BUFFER_USAGE_VERTEX_BUFFER_BIT | VK_BUFFER_USAGE_INDEX_BUFFER_BIT | + VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT; + constexpr u64 WATCHES_INITIAL_RESERVE = 0x4000; constexpr u64 WATCHES_RESERVE_CHUNK = 0x1000; @@ -56,17 +60,16 @@ u32 GetMemoryType(const VkPhysicalDeviceMemoryProperties& properties, } // Anonymous namespace -VKStreamBuffer::VKStreamBuffer(const VKDevice& device_, VKScheduler& scheduler_, - VkBufferUsageFlags usage) +VKStreamBuffer::VKStreamBuffer(const Device& device_, VKScheduler& scheduler_) : device{device_}, scheduler{scheduler_} { - CreateBuffers(usage); + CreateBuffers(); ReserveWatches(current_watches, WATCHES_INITIAL_RESERVE); ReserveWatches(previous_watches, WATCHES_INITIAL_RESERVE); } VKStreamBuffer::~VKStreamBuffer() = default; -std::tuple<u8*, u64, bool> VKStreamBuffer::Map(u64 size, u64 alignment) { +std::pair<u8*, u64> VKStreamBuffer::Map(u64 size, u64 alignment) { ASSERT(size <= stream_buffer_size); mapped_size = size; @@ -76,7 +79,6 @@ std::tuple<u8*, u64, bool> VKStreamBuffer::Map(u64 size, u64 alignment) { WaitPendingOperations(offset); - bool invalidated = false; if (offset + size > stream_buffer_size) { // The buffer would overflow, save the amount of used watches and reset the state. invalidation_mark = current_watch_cursor; @@ -90,11 +92,9 @@ std::tuple<u8*, u64, bool> VKStreamBuffer::Map(u64 size, u64 alignment) { // Ensure that we don't wait for uncommitted fences. scheduler.Flush(); - - invalidated = true; } - return {memory.Map(offset, size), offset, invalidated}; + return std::make_pair(memory.Map(offset, size), offset); } void VKStreamBuffer::Unmap(u64 size) { @@ -113,20 +113,21 @@ void VKStreamBuffer::Unmap(u64 size) { watch.tick = scheduler.CurrentTick(); } -void VKStreamBuffer::CreateBuffers(VkBufferUsageFlags usage) { +void VKStreamBuffer::CreateBuffers() { const auto memory_properties = device.GetPhysical().GetMemoryProperties(); const u32 preferred_type = GetMemoryType(memory_properties); const u32 preferred_heap = memory_properties.memoryTypes[preferred_type].heapIndex; // Substract from the preferred heap size some bytes to avoid getting out of memory. const VkDeviceSize heap_size = memory_properties.memoryHeaps[preferred_heap].size; - const VkDeviceSize allocable_size = heap_size - 9 * 1024 * 1024; + // As per DXVK's example, using `heap_size / 2` + const VkDeviceSize allocable_size = heap_size / 2; buffer = device.GetLogical().CreateBuffer({ .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, .pNext = nullptr, .flags = 0, .size = std::min(PREFERRED_STREAM_BUFFER_SIZE, allocable_size), - .usage = usage, + .usage = BUFFER_USAGE, .sharingMode = VK_SHARING_MODE_EXCLUSIVE, .queueFamilyIndexCount = 0, .pQueueFamilyIndices = nullptr, diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.h b/src/video_core/renderer_vulkan/vk_stream_buffer.h index 5e15ad78f..2e9c8cb46 100644 --- a/src/video_core/renderer_vulkan/vk_stream_buffer.h +++ b/src/video_core/renderer_vulkan/vk_stream_buffer.h @@ -5,31 +5,29 @@ #pragma once #include <optional> -#include <tuple> +#include <utility> #include <vector> #include "common/common_types.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class VKFenceWatch; class VKScheduler; class VKStreamBuffer final { public: - explicit VKStreamBuffer(const VKDevice& device, VKScheduler& scheduler, - VkBufferUsageFlags usage); + explicit VKStreamBuffer(const Device& device, VKScheduler& scheduler); ~VKStreamBuffer(); /** * Reserves a region of memory from the stream buffer. * @param size Size to reserve. - * @returns A tuple in the following order: Raw memory pointer (with offset added), buffer - * offset and a boolean that's true when buffer has been invalidated. + * @returns A pair of a raw memory pointer (with offset added), and the buffer offset */ - std::tuple<u8*, u64, bool> Map(u64 size, u64 alignment); + std::pair<u8*, u64> Map(u64 size, u64 alignment); /// Ensures that "size" bytes of memory are available to the GPU, potentially recording a copy. void Unmap(u64 size); @@ -49,14 +47,14 @@ private: }; /// Creates Vulkan buffer handles committing the required the required memory. - void CreateBuffers(VkBufferUsageFlags usage); + void CreateBuffers(); /// Increases the amount of watches available. void ReserveWatches(std::vector<Watch>& watches, std::size_t grow_size); void WaitPendingOperations(u64 requested_upper_bound); - const VKDevice& device; ///< Vulkan device manager. + const Device& device; ///< Vulkan device manager. VKScheduler& scheduler; ///< Command scheduler. vk::Buffer buffer; ///< Mapped buffer. diff --git a/src/video_core/renderer_vulkan/vk_swapchain.cpp b/src/video_core/renderer_vulkan/vk_swapchain.cpp index 9636a7c65..725a2a05d 100644 --- a/src/video_core/renderer_vulkan/vk_swapchain.cpp +++ b/src/video_core/renderer_vulkan/vk_swapchain.cpp @@ -11,10 +11,10 @@ #include "common/logging/log.h" #include "core/core.h" #include "core/frontend/framebuffer_layout.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_swapchain.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -56,7 +56,7 @@ VkExtent2D ChooseSwapExtent(const VkSurfaceCapabilitiesKHR& capabilities, u32 wi } // Anonymous namespace -VKSwapchain::VKSwapchain(VkSurfaceKHR surface_, const VKDevice& device_, VKScheduler& scheduler_) +VKSwapchain::VKSwapchain(VkSurfaceKHR surface_, const Device& device_, VKScheduler& scheduler_) : surface{surface_}, device{device_}, scheduler{scheduler_} {} VKSwapchain::~VKSwapchain() = default; diff --git a/src/video_core/renderer_vulkan/vk_swapchain.h b/src/video_core/renderer_vulkan/vk_swapchain.h index 6b39befdf..2eadd62b3 100644 --- a/src/video_core/renderer_vulkan/vk_swapchain.h +++ b/src/video_core/renderer_vulkan/vk_swapchain.h @@ -7,7 +7,7 @@ #include <vector> #include "common/common_types.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Layout { struct FramebufferLayout; @@ -15,12 +15,12 @@ struct FramebufferLayout; namespace Vulkan { -class VKDevice; +class Device; class VKScheduler; class VKSwapchain { public: - explicit VKSwapchain(VkSurfaceKHR surface, const VKDevice& device, VKScheduler& scheduler); + explicit VKSwapchain(VkSurfaceKHR surface, const Device& device, VKScheduler& scheduler); ~VKSwapchain(); /// Creates (or recreates) the swapchain with a given size. @@ -73,7 +73,7 @@ private: void Destroy(); const VkSurfaceKHR surface; - const VKDevice& device; + const Device& device; VKScheduler& scheduler; vk::SwapchainKHR swapchain; diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index f2c8f2ae1..bd11de012 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -4,613 +4,1105 @@ #include <algorithm> #include <array> -#include <cstddef> -#include <cstring> -#include <memory> -#include <variant> +#include <span> #include <vector> -#include "common/assert.h" -#include "common/common_types.h" -#include "core/core.h" -#include "video_core/engines/maxwell_3d.h" -#include "video_core/morton.h" +#include "video_core/engines/fermi_2d.h" +#include "video_core/renderer_vulkan/blit_image.h" #include "video_core/renderer_vulkan/maxwell_to_vk.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" #include "video_core/renderer_vulkan/vk_rasterizer.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h" #include "video_core/renderer_vulkan/vk_texture_cache.h" -#include "video_core/renderer_vulkan/wrapper.h" -#include "video_core/surface.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -using VideoCore::MortonSwizzle; -using VideoCore::MortonSwizzleMode; - +using Tegra::Engines::Fermi2D; using Tegra::Texture::SwizzleSource; -using VideoCore::Surface::PixelFormat; -using VideoCore::Surface::SurfaceTarget; +using Tegra::Texture::TextureMipmapFilter; +using VideoCommon::BufferImageCopy; +using VideoCommon::ImageInfo; +using VideoCommon::ImageType; +using VideoCommon::SubresourceRange; +using VideoCore::Surface::IsPixelFormatASTC; namespace { -VkImageType SurfaceTargetToImage(SurfaceTarget target) { - switch (target) { - case SurfaceTarget::Texture1D: - case SurfaceTarget::Texture1DArray: +constexpr std::array ATTACHMENT_REFERENCES{ + VkAttachmentReference{0, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{1, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{2, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{3, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{4, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{5, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{6, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{7, VK_IMAGE_LAYOUT_GENERAL}, + VkAttachmentReference{8, VK_IMAGE_LAYOUT_GENERAL}, +}; + +constexpr VkBorderColor ConvertBorderColor(const std::array<float, 4>& color) { + if (color == std::array<float, 4>{0, 0, 0, 0}) { + return VK_BORDER_COLOR_FLOAT_TRANSPARENT_BLACK; + } else if (color == std::array<float, 4>{0, 0, 0, 1}) { + return VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK; + } else if (color == std::array<float, 4>{1, 1, 1, 1}) { + return VK_BORDER_COLOR_FLOAT_OPAQUE_WHITE; + } + if (color[0] + color[1] + color[2] > 1.35f) { + // If color elements are brighter than roughly 0.5 average, use white border + return VK_BORDER_COLOR_FLOAT_OPAQUE_WHITE; + } else if (color[3] > 0.5f) { + return VK_BORDER_COLOR_FLOAT_OPAQUE_BLACK; + } else { + return VK_BORDER_COLOR_FLOAT_TRANSPARENT_BLACK; + } +} + +[[nodiscard]] VkImageType ConvertImageType(const ImageType type) { + switch (type) { + case ImageType::e1D: return VK_IMAGE_TYPE_1D; - case SurfaceTarget::Texture2D: - case SurfaceTarget::Texture2DArray: - case SurfaceTarget::TextureCubemap: - case SurfaceTarget::TextureCubeArray: + case ImageType::e2D: + case ImageType::Linear: return VK_IMAGE_TYPE_2D; - case SurfaceTarget::Texture3D: + case ImageType::e3D: return VK_IMAGE_TYPE_3D; - case SurfaceTarget::TextureBuffer: - UNREACHABLE(); - return {}; + case ImageType::Buffer: + break; } - UNREACHABLE_MSG("Unknown texture target={}", static_cast<u32>(target)); + UNREACHABLE_MSG("Invalid image type={}", type); return {}; } -VkImageAspectFlags PixelFormatToImageAspect(PixelFormat pixel_format) { - if (pixel_format < PixelFormat::MaxColorFormat) { - return VK_IMAGE_ASPECT_COLOR_BIT; - } else if (pixel_format < PixelFormat::MaxDepthFormat) { - return VK_IMAGE_ASPECT_DEPTH_BIT; - } else if (pixel_format < PixelFormat::MaxDepthStencilFormat) { - return VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT; - } else { - UNREACHABLE_MSG("Invalid pixel format={}", static_cast<int>(pixel_format)); - return VK_IMAGE_ASPECT_COLOR_BIT; +[[nodiscard]] VkSampleCountFlagBits ConvertSampleCount(u32 num_samples) { + switch (num_samples) { + case 1: + return VK_SAMPLE_COUNT_1_BIT; + case 2: + return VK_SAMPLE_COUNT_2_BIT; + case 4: + return VK_SAMPLE_COUNT_4_BIT; + case 8: + return VK_SAMPLE_COUNT_8_BIT; + case 16: + return VK_SAMPLE_COUNT_16_BIT; + default: + UNREACHABLE_MSG("Invalid number of samples={}", num_samples); + return VK_SAMPLE_COUNT_1_BIT; } } -VkImageViewType GetImageViewType(SurfaceTarget target) { - switch (target) { - case SurfaceTarget::Texture1D: - return VK_IMAGE_VIEW_TYPE_1D; - case SurfaceTarget::Texture2D: - return VK_IMAGE_VIEW_TYPE_2D; - case SurfaceTarget::Texture3D: - return VK_IMAGE_VIEW_TYPE_3D; - case SurfaceTarget::Texture1DArray: - return VK_IMAGE_VIEW_TYPE_1D_ARRAY; - case SurfaceTarget::Texture2DArray: - return VK_IMAGE_VIEW_TYPE_2D_ARRAY; - case SurfaceTarget::TextureCubemap: - return VK_IMAGE_VIEW_TYPE_CUBE; - case SurfaceTarget::TextureCubeArray: - return VK_IMAGE_VIEW_TYPE_CUBE_ARRAY; - case SurfaceTarget::TextureBuffer: - break; +[[nodiscard]] VkImageCreateInfo MakeImageCreateInfo(const Device& device, const ImageInfo& info) { + const auto format_info = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, info.format); + VkImageCreateFlags flags = VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT; + if (info.type == ImageType::e2D && info.resources.layers >= 6 && + info.size.width == info.size.height) { + flags |= VK_IMAGE_CREATE_CUBE_COMPATIBLE_BIT; } - UNREACHABLE(); - return {}; -} - -vk::Buffer CreateBuffer(const VKDevice& device, const SurfaceParams& params, - std::size_t host_memory_size) { - // TODO(Rodrigo): Move texture buffer creation to the buffer cache - return device.GetLogical().CreateBuffer({ - .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, + if (info.type == ImageType::e3D) { + flags |= VK_IMAGE_CREATE_2D_ARRAY_COMPATIBLE_BIT; + } + VkImageUsageFlags usage = VK_IMAGE_USAGE_TRANSFER_SRC_BIT | VK_IMAGE_USAGE_TRANSFER_DST_BIT | + VK_IMAGE_USAGE_SAMPLED_BIT; + if (format_info.attachable) { + switch (VideoCore::Surface::GetFormatType(info.format)) { + case VideoCore::Surface::SurfaceType::ColorTexture: + usage |= VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT; + break; + case VideoCore::Surface::SurfaceType::Depth: + case VideoCore::Surface::SurfaceType::DepthStencil: + usage |= VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT; + break; + default: + UNREACHABLE_MSG("Invalid surface type"); + } + } + if (format_info.storage) { + usage |= VK_IMAGE_USAGE_STORAGE_BIT; + } + const auto [samples_x, samples_y] = VideoCommon::SamplesLog2(info.num_samples); + return VkImageCreateInfo{ + .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO, .pNext = nullptr, - .flags = 0, - .size = static_cast<VkDeviceSize>(host_memory_size), - .usage = VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT | - VK_BUFFER_USAGE_STORAGE_TEXEL_BUFFER_BIT | VK_BUFFER_USAGE_TRANSFER_SRC_BIT | - VK_BUFFER_USAGE_TRANSFER_DST_BIT, + .flags = flags, + .imageType = ConvertImageType(info.type), + .format = format_info.format, + .extent = + { + .width = info.size.width >> samples_x, + .height = info.size.height >> samples_y, + .depth = info.size.depth, + }, + .mipLevels = static_cast<u32>(info.resources.levels), + .arrayLayers = static_cast<u32>(info.resources.layers), + .samples = ConvertSampleCount(info.num_samples), + .tiling = VK_IMAGE_TILING_OPTIMAL, + .usage = usage, .sharingMode = VK_SHARING_MODE_EXCLUSIVE, .queueFamilyIndexCount = 0, .pQueueFamilyIndices = nullptr, - }); -} - -VkBufferViewCreateInfo GenerateBufferViewCreateInfo(const VKDevice& device, - const SurfaceParams& params, VkBuffer buffer, - std::size_t host_memory_size) { - ASSERT(params.IsBuffer()); - - return { - .sType = VK_STRUCTURE_TYPE_BUFFER_VIEW_CREATE_INFO, - .pNext = nullptr, - .flags = 0, - .buffer = buffer, - .format = - MaxwellToVK::SurfaceFormat(device, FormatType::Buffer, params.pixel_format).format, - .offset = 0, - .range = static_cast<VkDeviceSize>(host_memory_size), + .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, }; } -VkImageCreateInfo GenerateImageCreateInfo(const VKDevice& device, const SurfaceParams& params) { - ASSERT(!params.IsBuffer()); - - const auto [format, attachable, storage] = - MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, params.pixel_format); +[[nodiscard]] vk::Image MakeImage(const Device& device, const ImageInfo& info) { + if (info.type == ImageType::Buffer) { + return vk::Image{}; + } + return device.GetLogical().CreateImage(MakeImageCreateInfo(device, info)); +} - VkImageCreateInfo ci{ - .sType = VK_STRUCTURE_TYPE_IMAGE_CREATE_INFO, +[[nodiscard]] vk::Buffer MakeBuffer(const Device& device, const ImageInfo& info) { + if (info.type != ImageType::Buffer) { + return vk::Buffer{}; + } + const size_t bytes_per_block = VideoCore::Surface::BytesPerBlock(info.format); + return device.GetLogical().CreateBuffer(VkBufferCreateInfo{ + .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO, .pNext = nullptr, .flags = 0, - .imageType = SurfaceTargetToImage(params.target), - .format = format, - .extent = {}, - .mipLevels = params.num_levels, - .arrayLayers = static_cast<u32>(params.GetNumLayers()), - .samples = VK_SAMPLE_COUNT_1_BIT, - .tiling = VK_IMAGE_TILING_OPTIMAL, - .usage = VK_IMAGE_USAGE_SAMPLED_BIT | VK_IMAGE_USAGE_TRANSFER_DST_BIT | - VK_IMAGE_USAGE_TRANSFER_SRC_BIT, + .size = info.size.width * bytes_per_block, + .usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT | + VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT | + VK_BUFFER_USAGE_STORAGE_TEXEL_BUFFER_BIT, .sharingMode = VK_SHARING_MODE_EXCLUSIVE, .queueFamilyIndexCount = 0, .pQueueFamilyIndices = nullptr, - .initialLayout = VK_IMAGE_LAYOUT_UNDEFINED, - }; - if (attachable) { - ci.usage |= params.IsPixelFormatZeta() ? VK_IMAGE_USAGE_DEPTH_STENCIL_ATTACHMENT_BIT - : VK_IMAGE_USAGE_COLOR_ATTACHMENT_BIT; - } - if (storage) { - ci.usage |= VK_IMAGE_USAGE_STORAGE_BIT; - } - - switch (params.target) { - case SurfaceTarget::TextureCubemap: - case SurfaceTarget::TextureCubeArray: - ci.flags |= VK_IMAGE_CREATE_CUBE_COMPATIBLE_BIT; - [[fallthrough]]; - case SurfaceTarget::Texture1D: - case SurfaceTarget::Texture1DArray: - case SurfaceTarget::Texture2D: - case SurfaceTarget::Texture2DArray: - ci.extent = {params.width, params.height, 1}; - break; - case SurfaceTarget::Texture3D: - ci.flags |= VK_IMAGE_CREATE_2D_ARRAY_COMPATIBLE_BIT; - ci.extent = {params.width, params.height, params.depth}; - break; - case SurfaceTarget::TextureBuffer: - UNREACHABLE(); - } - - return ci; + }); } -u32 EncodeSwizzle(Tegra::Texture::SwizzleSource x_source, Tegra::Texture::SwizzleSource y_source, - Tegra::Texture::SwizzleSource z_source, Tegra::Texture::SwizzleSource w_source) { - return (static_cast<u32>(x_source) << 24) | (static_cast<u32>(y_source) << 16) | - (static_cast<u32>(z_source) << 8) | static_cast<u32>(w_source); +[[nodiscard]] VkImageAspectFlags ImageAspectMask(PixelFormat format) { + switch (VideoCore::Surface::GetFormatType(format)) { + case VideoCore::Surface::SurfaceType::ColorTexture: + return VK_IMAGE_ASPECT_COLOR_BIT; + case VideoCore::Surface::SurfaceType::Depth: + return VK_IMAGE_ASPECT_DEPTH_BIT; + case VideoCore::Surface::SurfaceType::DepthStencil: + return VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT; + default: + UNREACHABLE_MSG("Invalid surface type"); + return VkImageAspectFlags{}; + } } -} // Anonymous namespace - -CachedSurface::CachedSurface(const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, VKStagingBufferPool& staging_pool, - GPUVAddr gpu_addr, const SurfaceParams& params) - : SurfaceBase<View>{gpu_addr, params, device.IsOptimalAstcSupported()}, device{device}, - memory_manager{memory_manager}, scheduler{scheduler}, staging_pool{staging_pool} { - if (params.IsBuffer()) { - buffer = CreateBuffer(device, params, host_memory_size); - commit = memory_manager.Commit(buffer, false); - - const auto buffer_view_ci = - GenerateBufferViewCreateInfo(device, params, *buffer, host_memory_size); - format = buffer_view_ci.format; - - buffer_view = device.GetLogical().CreateBufferView(buffer_view_ci); - } else { - const auto image_ci = GenerateImageCreateInfo(device, params); - format = image_ci.format; - - image.emplace(device, scheduler, image_ci, PixelFormatToImageAspect(params.pixel_format)); - commit = memory_manager.Commit(image->GetHandle(), false); +[[nodiscard]] VkImageAspectFlags ImageViewAspectMask(const VideoCommon::ImageViewInfo& info) { + if (info.IsRenderTarget()) { + return ImageAspectMask(info.format); } - - // TODO(Rodrigo): Move this to a virtual function. - u32 num_layers = 1; - if (params.is_layered || params.target == SurfaceTarget::Texture3D) { - num_layers = params.depth; + const bool is_first = info.Swizzle()[0] == SwizzleSource::R; + switch (info.format) { + case PixelFormat::D24_UNORM_S8_UINT: + case PixelFormat::D32_FLOAT_S8_UINT: + return is_first ? VK_IMAGE_ASPECT_DEPTH_BIT : VK_IMAGE_ASPECT_STENCIL_BIT; + case PixelFormat::S8_UINT_D24_UNORM: + return is_first ? VK_IMAGE_ASPECT_STENCIL_BIT : VK_IMAGE_ASPECT_DEPTH_BIT; + case PixelFormat::D16_UNORM: + case PixelFormat::D32_FLOAT: + return VK_IMAGE_ASPECT_DEPTH_BIT; + default: + return VK_IMAGE_ASPECT_COLOR_BIT; } - main_view = CreateView(ViewParams(params.target, 0, num_layers, 0, params.num_levels)); } -CachedSurface::~CachedSurface() = default; - -void CachedSurface::UploadTexture(const std::vector<u8>& staging_buffer) { - // To upload data we have to be outside of a renderpass - scheduler.RequestOutsideRenderPassOperationContext(); +[[nodiscard]] VkAttachmentDescription AttachmentDescription(const Device& device, + const ImageView* image_view) { + const auto pixel_format = image_view->format; + return VkAttachmentDescription{ + .flags = VK_ATTACHMENT_DESCRIPTION_MAY_ALIAS_BIT, + .format = MaxwellToVK::SurfaceFormat(device, FormatType::Optimal, pixel_format).format, + .samples = image_view->Samples(), + .loadOp = VK_ATTACHMENT_LOAD_OP_LOAD, + .storeOp = VK_ATTACHMENT_STORE_OP_STORE, + .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD, + .stencilStoreOp = VK_ATTACHMENT_STORE_OP_STORE, + .initialLayout = VK_IMAGE_LAYOUT_GENERAL, + .finalLayout = VK_IMAGE_LAYOUT_GENERAL, + }; +} - if (params.IsBuffer()) { - UploadBuffer(staging_buffer); - } else { - UploadImage(staging_buffer); +[[nodiscard]] VkComponentSwizzle ComponentSwizzle(SwizzleSource swizzle) { + switch (swizzle) { + case SwizzleSource::Zero: + return VK_COMPONENT_SWIZZLE_ZERO; + case SwizzleSource::R: + return VK_COMPONENT_SWIZZLE_R; + case SwizzleSource::G: + return VK_COMPONENT_SWIZZLE_G; + case SwizzleSource::B: + return VK_COMPONENT_SWIZZLE_B; + case SwizzleSource::A: + return VK_COMPONENT_SWIZZLE_A; + case SwizzleSource::OneFloat: + case SwizzleSource::OneInt: + return VK_COMPONENT_SWIZZLE_ONE; } + UNREACHABLE_MSG("Invalid swizzle={}", swizzle); + return VK_COMPONENT_SWIZZLE_ZERO; } -void CachedSurface::DownloadTexture(std::vector<u8>& staging_buffer) { - UNIMPLEMENTED_IF(params.IsBuffer()); - - if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5_UNORM) { - LOG_WARNING(Render_Vulkan, "A1B5G5R5 flushing is stubbed"); +[[nodiscard]] VkImageViewType ImageViewType(VideoCommon::ImageViewType type) { + switch (type) { + case VideoCommon::ImageViewType::e1D: + return VK_IMAGE_VIEW_TYPE_1D; + case VideoCommon::ImageViewType::e2D: + return VK_IMAGE_VIEW_TYPE_2D; + case VideoCommon::ImageViewType::Cube: + return VK_IMAGE_VIEW_TYPE_CUBE; + case VideoCommon::ImageViewType::e3D: + return VK_IMAGE_VIEW_TYPE_3D; + case VideoCommon::ImageViewType::e1DArray: + return VK_IMAGE_VIEW_TYPE_1D_ARRAY; + case VideoCommon::ImageViewType::e2DArray: + return VK_IMAGE_VIEW_TYPE_2D_ARRAY; + case VideoCommon::ImageViewType::CubeArray: + return VK_IMAGE_VIEW_TYPE_CUBE_ARRAY; + case VideoCommon::ImageViewType::Rect: + LOG_WARNING(Render_Vulkan, "Unnormalized image view type not supported"); + return VK_IMAGE_VIEW_TYPE_2D; + case VideoCommon::ImageViewType::Buffer: + UNREACHABLE_MSG("Texture buffers can't be image views"); + return VK_IMAGE_VIEW_TYPE_1D; } + UNREACHABLE_MSG("Invalid image view type={}", type); + return VK_IMAGE_VIEW_TYPE_2D; +} - // We can't copy images to buffers inside a renderpass - scheduler.RequestOutsideRenderPassOperationContext(); +[[nodiscard]] VkImageSubresourceLayers MakeImageSubresourceLayers( + VideoCommon::SubresourceLayers subresource, VkImageAspectFlags aspect_mask) { + return VkImageSubresourceLayers{ + .aspectMask = aspect_mask, + .mipLevel = static_cast<u32>(subresource.base_level), + .baseArrayLayer = static_cast<u32>(subresource.base_layer), + .layerCount = static_cast<u32>(subresource.num_layers), + }; +} - FullTransition(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_READ_BIT, - VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL); +[[nodiscard]] VkOffset3D MakeOffset3D(VideoCommon::Offset3D offset3d) { + return VkOffset3D{ + .x = offset3d.x, + .y = offset3d.y, + .z = offset3d.z, + }; +} - const auto& buffer = staging_pool.GetUnusedBuffer(host_memory_size, true); - // TODO(Rodrigo): Do this in a single copy - for (u32 level = 0; level < params.num_levels; ++level) { - scheduler.Record([image = *image->GetHandle(), buffer = *buffer.handle, - copy = GetBufferImageCopy(level)](vk::CommandBuffer cmdbuf) { - cmdbuf.CopyImageToBuffer(image, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, buffer, copy); - }); - } - scheduler.Finish(); +[[nodiscard]] VkExtent3D MakeExtent3D(VideoCommon::Extent3D extent3d) { + return VkExtent3D{ + .width = static_cast<u32>(extent3d.width), + .height = static_cast<u32>(extent3d.height), + .depth = static_cast<u32>(extent3d.depth), + }; +} - // TODO(Rodrigo): Use an intern buffer for staging buffers and avoid this unnecessary memcpy. - std::memcpy(staging_buffer.data(), buffer.commit->Map(host_memory_size), host_memory_size); +[[nodiscard]] VkImageCopy MakeImageCopy(const VideoCommon::ImageCopy& copy, + VkImageAspectFlags aspect_mask) noexcept { + return VkImageCopy{ + .srcSubresource = MakeImageSubresourceLayers(copy.src_subresource, aspect_mask), + .srcOffset = MakeOffset3D(copy.src_offset), + .dstSubresource = MakeImageSubresourceLayers(copy.dst_subresource, aspect_mask), + .dstOffset = MakeOffset3D(copy.dst_offset), + .extent = MakeExtent3D(copy.extent), + }; } -void CachedSurface::DecorateSurfaceName() { - // TODO(Rodrigo): Add name decorations +[[nodiscard]] std::vector<VkBufferCopy> TransformBufferCopies( + std::span<const VideoCommon::BufferCopy> copies, size_t buffer_offset) { + std::vector<VkBufferCopy> result(copies.size()); + std::ranges::transform( + copies, result.begin(), [buffer_offset](const VideoCommon::BufferCopy& copy) { + return VkBufferCopy{ + .srcOffset = static_cast<VkDeviceSize>(copy.src_offset + buffer_offset), + .dstOffset = static_cast<VkDeviceSize>(copy.dst_offset), + .size = static_cast<VkDeviceSize>(copy.size), + }; + }); + return result; } -View CachedSurface::CreateView(const ViewParams& params) { - // TODO(Rodrigo): Add name decorations - return views[params] = std::make_shared<CachedSurfaceView>(device, *this, params); +[[nodiscard]] std::vector<VkBufferImageCopy> TransformBufferImageCopies( + std::span<const BufferImageCopy> copies, size_t buffer_offset, VkImageAspectFlags aspect_mask) { + struct Maker { + VkBufferImageCopy operator()(const BufferImageCopy& copy) const { + return VkBufferImageCopy{ + .bufferOffset = copy.buffer_offset + buffer_offset, + .bufferRowLength = copy.buffer_row_length, + .bufferImageHeight = copy.buffer_image_height, + .imageSubresource = + { + .aspectMask = aspect_mask, + .mipLevel = static_cast<u32>(copy.image_subresource.base_level), + .baseArrayLayer = static_cast<u32>(copy.image_subresource.base_layer), + .layerCount = static_cast<u32>(copy.image_subresource.num_layers), + }, + .imageOffset = + { + .x = copy.image_offset.x, + .y = copy.image_offset.y, + .z = copy.image_offset.z, + }, + .imageExtent = + { + .width = copy.image_extent.width, + .height = copy.image_extent.height, + .depth = copy.image_extent.depth, + }, + }; + } + size_t buffer_offset; + VkImageAspectFlags aspect_mask; + }; + if (aspect_mask == (VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT)) { + std::vector<VkBufferImageCopy> result(copies.size() * 2); + std::ranges::transform(copies, result.begin(), + Maker{buffer_offset, VK_IMAGE_ASPECT_DEPTH_BIT}); + std::ranges::transform(copies, result.begin() + copies.size(), + Maker{buffer_offset, VK_IMAGE_ASPECT_STENCIL_BIT}); + return result; + } else { + std::vector<VkBufferImageCopy> result(copies.size()); + std::ranges::transform(copies, result.begin(), Maker{buffer_offset, aspect_mask}); + return result; + } } -void CachedSurface::UploadBuffer(const std::vector<u8>& staging_buffer) { - const auto& src_buffer = staging_pool.GetUnusedBuffer(host_memory_size, true); - std::memcpy(src_buffer.commit->Map(host_memory_size), staging_buffer.data(), host_memory_size); +[[nodiscard]] VkImageSubresourceRange MakeSubresourceRange(VkImageAspectFlags aspect_mask, + const SubresourceRange& range) { + return VkImageSubresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = static_cast<u32>(range.base.level), + .levelCount = static_cast<u32>(range.extent.levels), + .baseArrayLayer = static_cast<u32>(range.base.layer), + .layerCount = static_cast<u32>(range.extent.layers), + }; +} - scheduler.Record([src_buffer = *src_buffer.handle, dst_buffer = *buffer, - size = host_memory_size](vk::CommandBuffer cmdbuf) { - VkBufferCopy copy; - copy.srcOffset = 0; - copy.dstOffset = 0; - copy.size = size; - cmdbuf.CopyBuffer(src_buffer, dst_buffer, copy); +[[nodiscard]] VkImageSubresourceRange MakeSubresourceRange(const ImageView* image_view) { + SubresourceRange range = image_view->range; + if (True(image_view->flags & VideoCommon::ImageViewFlagBits::Slice)) { + // Slice image views always affect a single layer, but their subresource range corresponds + // to the slice. Override the value to affect a single layer. + range.base.layer = 0; + range.extent.layers = 1; + } + return MakeSubresourceRange(ImageAspectMask(image_view->format), range); +} - VkBufferMemoryBarrier barrier; - barrier.sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_BARRIER; - barrier.pNext = nullptr; - barrier.srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT; - barrier.dstAccessMask = VK_ACCESS_SHADER_READ_BIT; - barrier.srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; // They'll be ignored anyway - barrier.dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED; - barrier.buffer = dst_buffer; - barrier.offset = 0; - barrier.size = size; - cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_VERTEX_SHADER_BIT, - 0, {}, barrier, {}); - }); +[[nodiscard]] VkImageSubresourceLayers MakeSubresourceLayers(const ImageView* image_view) { + return VkImageSubresourceLayers{ + .aspectMask = ImageAspectMask(image_view->format), + .mipLevel = static_cast<u32>(image_view->range.base.level), + .baseArrayLayer = static_cast<u32>(image_view->range.base.layer), + .layerCount = static_cast<u32>(image_view->range.extent.layers), + }; } -void CachedSurface::UploadImage(const std::vector<u8>& staging_buffer) { - const auto& src_buffer = staging_pool.GetUnusedBuffer(host_memory_size, true); - std::memcpy(src_buffer.commit->Map(host_memory_size), staging_buffer.data(), host_memory_size); - - FullTransition(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_WRITE_BIT, - VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL); - - for (u32 level = 0; level < params.num_levels; ++level) { - const VkBufferImageCopy copy = GetBufferImageCopy(level); - if (image->GetAspectMask() == (VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT)) { - scheduler.Record([buffer = *src_buffer.handle, image = *image->GetHandle(), - copy](vk::CommandBuffer cmdbuf) { - std::array<VkBufferImageCopy, 2> copies = {copy, copy}; - copies[0].imageSubresource.aspectMask = VK_IMAGE_ASPECT_DEPTH_BIT; - copies[1].imageSubresource.aspectMask = VK_IMAGE_ASPECT_STENCIL_BIT; - cmdbuf.CopyBufferToImage(buffer, image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, - copies); - }); - } else { - scheduler.Record([buffer = *src_buffer.handle, image = *image->GetHandle(), - copy](vk::CommandBuffer cmdbuf) { - cmdbuf.CopyBufferToImage(buffer, image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, copy); - }); - } +[[nodiscard]] constexpr SwizzleSource ConvertGreenRed(SwizzleSource value) { + switch (value) { + case SwizzleSource::G: + return SwizzleSource::R; + default: + return value; } } -VkBufferImageCopy CachedSurface::GetBufferImageCopy(u32 level) const { - return { - .bufferOffset = params.GetHostMipmapLevelOffset(level, is_converted), - .bufferRowLength = 0, - .bufferImageHeight = 0, - .imageSubresource = +void CopyBufferToImage(vk::CommandBuffer cmdbuf, VkBuffer src_buffer, VkImage image, + VkImageAspectFlags aspect_mask, bool is_initialized, + std::span<const VkBufferImageCopy> copies) { + static constexpr VkAccessFlags ACCESS_FLAGS = VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT; + const VkImageMemoryBarrier read_barrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = ACCESS_FLAGS, + .dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = is_initialized ? VK_IMAGE_LAYOUT_GENERAL : VK_IMAGE_LAYOUT_UNDEFINED, + .newLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = image, + .subresourceRange = { - .aspectMask = image->GetAspectMask(), - .mipLevel = level, + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = VK_REMAINING_MIP_LEVELS, .baseArrayLayer = 0, - .layerCount = static_cast<u32>(params.GetNumLayers()), + .layerCount = VK_REMAINING_ARRAY_LAYERS, }, - .imageOffset = {.x = 0, .y = 0, .z = 0}, - .imageExtent = + }; + const VkImageMemoryBarrier write_barrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = ACCESS_FLAGS, + .oldLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = image, + .subresourceRange = { - .width = params.GetMipWidth(level), - .height = params.GetMipHeight(level), - .depth = params.target == SurfaceTarget::Texture3D ? params.GetMipDepth(level) : 1U, + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = VK_REMAINING_MIP_LEVELS, + .baseArrayLayer = 0, + .layerCount = VK_REMAINING_ARRAY_LAYERS, }, }; + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, 0, + read_barrier); + cmdbuf.CopyBufferToImage(src_buffer, image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, copies); + // TODO: Move this to another API + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, 0, + write_barrier); } -VkImageSubresourceRange CachedSurface::GetImageSubresourceRange() const { - return {image->GetAspectMask(), 0, params.num_levels, 0, - static_cast<u32>(params.GetNumLayers())}; +[[nodiscard]] VkImageBlit MakeImageBlit(const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + const VkImageSubresourceLayers& dst_layers, + const VkImageSubresourceLayers& src_layers) { + return VkImageBlit{ + .srcSubresource = src_layers, + .srcOffsets = + { + { + .x = src_region[0].x, + .y = src_region[0].y, + .z = 0, + }, + { + .x = src_region[1].x, + .y = src_region[1].y, + .z = 1, + }, + }, + .dstSubresource = dst_layers, + .dstOffsets = + { + { + .x = dst_region[0].x, + .y = dst_region[0].y, + .z = 0, + }, + { + .x = dst_region[1].x, + .y = dst_region[1].y, + .z = 1, + }, + }, + }; } -CachedSurfaceView::CachedSurfaceView(const VKDevice& device, CachedSurface& surface, - const ViewParams& params) - : VideoCommon::ViewBase{params}, params{surface.GetSurfaceParams()}, - image{surface.GetImageHandle()}, buffer_view{surface.GetBufferViewHandle()}, - aspect_mask{surface.GetAspectMask()}, device{device}, surface{surface}, - base_level{params.base_level}, num_levels{params.num_levels}, - image_view_type{image ? GetImageViewType(params.target) : VK_IMAGE_VIEW_TYPE_1D} { - if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { - base_layer = 0; - num_layers = 1; - base_slice = params.base_layer; - num_slices = params.num_layers; - } else { - base_layer = params.base_layer; - num_layers = params.num_layers; - } +[[nodiscard]] VkImageResolve MakeImageResolve(const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + const VkImageSubresourceLayers& dst_layers, + const VkImageSubresourceLayers& src_layers) { + return VkImageResolve{ + .srcSubresource = src_layers, + .srcOffset = + { + .x = src_region[0].x, + .y = src_region[0].y, + .z = 0, + }, + .dstSubresource = dst_layers, + .dstOffset = + { + .x = dst_region[0].x, + .y = dst_region[0].y, + .z = 0, + }, + .extent = + { + .width = static_cast<u32>(dst_region[1].x - dst_region[0].x), + .height = static_cast<u32>(dst_region[1].y - dst_region[0].y), + .depth = 1, + }, + }; } -CachedSurfaceView::~CachedSurfaceView() = default; - -VkImageView CachedSurfaceView::GetImageView(SwizzleSource x_source, SwizzleSource y_source, - SwizzleSource z_source, SwizzleSource w_source) { - const u32 new_swizzle = EncodeSwizzle(x_source, y_source, z_source, w_source); - if (last_image_view && last_swizzle == new_swizzle) { - return last_image_view; +struct RangedBarrierRange { + u32 min_mip = std::numeric_limits<u32>::max(); + u32 max_mip = std::numeric_limits<u32>::min(); + u32 min_layer = std::numeric_limits<u32>::max(); + u32 max_layer = std::numeric_limits<u32>::min(); + + void AddLayers(const VkImageSubresourceLayers& layers) { + min_mip = std::min(min_mip, layers.mipLevel); + max_mip = std::max(max_mip, layers.mipLevel + 1); + min_layer = std::min(min_layer, layers.baseArrayLayer); + max_layer = std::max(max_layer, layers.baseArrayLayer + layers.layerCount); } - last_swizzle = new_swizzle; - const auto [entry, is_cache_miss] = view_cache.try_emplace(new_swizzle); - auto& image_view = entry->second; - if (!is_cache_miss) { - return last_image_view = *image_view; + VkImageSubresourceRange SubresourceRange(VkImageAspectFlags aspect_mask) const noexcept { + return VkImageSubresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = min_mip, + .levelCount = max_mip - min_mip, + .baseArrayLayer = min_layer, + .layerCount = max_layer - min_layer, + }; } +}; - std::array swizzle{MaxwellToVK::SwizzleSource(x_source), MaxwellToVK::SwizzleSource(y_source), - MaxwellToVK::SwizzleSource(z_source), MaxwellToVK::SwizzleSource(w_source)}; - if (params.pixel_format == VideoCore::Surface::PixelFormat::A1B5G5R5_UNORM) { - // A1B5G5R5 is implemented as A1R5G5B5, we have to change the swizzle here. - std::swap(swizzle[0], swizzle[2]); - } +} // Anonymous namespace - // Games can sample depth or stencil values on textures. This is decided by the swizzle value on - // hardware. To emulate this on Vulkan we specify it in the aspect. - VkImageAspectFlags aspect = aspect_mask; - if (aspect == (VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT)) { - UNIMPLEMENTED_IF(x_source != SwizzleSource::R && x_source != SwizzleSource::G); - const bool is_first = x_source == SwizzleSource::R; - switch (params.pixel_format) { - case VideoCore::Surface::PixelFormat::D24_UNORM_S8_UINT: - case VideoCore::Surface::PixelFormat::D32_FLOAT_S8_UINT: - aspect = is_first ? VK_IMAGE_ASPECT_DEPTH_BIT : VK_IMAGE_ASPECT_STENCIL_BIT; - break; - case VideoCore::Surface::PixelFormat::S8_UINT_D24_UNORM: - aspect = is_first ? VK_IMAGE_ASPECT_STENCIL_BIT : VK_IMAGE_ASPECT_DEPTH_BIT; - break; - default: - aspect = VK_IMAGE_ASPECT_DEPTH_BIT; - UNIMPLEMENTED(); - } +void TextureCacheRuntime::Finish() { + scheduler.Finish(); +} - // Make sure we sample the first component - std::transform( - swizzle.begin(), swizzle.end(), swizzle.begin(), [](VkComponentSwizzle component) { - return component == VK_COMPONENT_SWIZZLE_G ? VK_COMPONENT_SWIZZLE_R : component; - }); - } +ImageBufferMap TextureCacheRuntime::MapUploadBuffer(size_t size) { + const auto& buffer = staging_buffer_pool.GetUnusedBuffer(size, true); + return ImageBufferMap{ + .handle = *buffer.handle, + .map = buffer.commit->Map(size), + }; +} - if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { - ASSERT(base_slice == 0); - ASSERT(num_slices == params.depth); +void TextureCacheRuntime::BlitImage(Framebuffer* dst_framebuffer, ImageView& dst, ImageView& src, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation) { + const VkImageAspectFlags aspect_mask = ImageAspectMask(src.format); + const bool is_dst_msaa = dst.Samples() != VK_SAMPLE_COUNT_1_BIT; + const bool is_src_msaa = src.Samples() != VK_SAMPLE_COUNT_1_BIT; + ASSERT(aspect_mask == ImageAspectMask(dst.format)); + if (aspect_mask == VK_IMAGE_ASPECT_COLOR_BIT && !is_src_msaa && !is_dst_msaa) { + blit_image_helper.BlitColor(dst_framebuffer, src, dst_region, src_region, filter, + operation); + return; } - - image_view = device.GetLogical().CreateImageView({ - .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, - .pNext = nullptr, - .flags = 0, - .image = surface.GetImageHandle(), - .viewType = image_view_type, - .format = surface.GetImage().GetFormat(), - .components = - { - .r = swizzle[0], - .g = swizzle[1], - .b = swizzle[2], - .a = swizzle[3], + if (aspect_mask == (VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT)) { + if (!device.IsBlitDepthStencilSupported()) { + UNIMPLEMENTED_IF(is_src_msaa || is_dst_msaa); + blit_image_helper.BlitDepthStencil(dst_framebuffer, src.DepthView(), src.StencilView(), + dst_region, src_region, filter, operation); + return; + } + } + ASSERT(src.ImageFormat() == dst.ImageFormat()); + ASSERT(!(is_dst_msaa && !is_src_msaa)); + ASSERT(operation == Fermi2D::Operation::SrcCopy); + + const VkImage dst_image = dst.ImageHandle(); + const VkImage src_image = src.ImageHandle(); + const VkImageSubresourceLayers dst_layers = MakeSubresourceLayers(&dst); + const VkImageSubresourceLayers src_layers = MakeSubresourceLayers(&src); + const bool is_resolve = is_src_msaa && !is_dst_msaa; + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([filter, dst_region, src_region, dst_image, src_image, dst_layers, src_layers, + aspect_mask, is_resolve](vk::CommandBuffer cmdbuf) { + const std::array read_barriers{ + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_READ_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = src_image, + .subresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = VK_REMAINING_MIP_LEVELS, + .baseArrayLayer = 0, + .layerCount = VK_REMAINING_ARRAY_LAYERS, + }, }, - .subresourceRange = - { - .aspectMask = aspect, - .baseMipLevel = base_level, - .levelCount = num_levels, - .baseArrayLayer = base_layer, - .layerCount = num_layers, + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = dst_image, + .subresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = VK_REMAINING_MIP_LEVELS, + .baseArrayLayer = 0, + .layerCount = VK_REMAINING_ARRAY_LAYERS, + }, + }, + }; + VkImageMemoryBarrier write_barrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = dst_image, + .subresourceRange{ + .aspectMask = aspect_mask, + .baseMipLevel = 0, + .levelCount = VK_REMAINING_MIP_LEVELS, + .baseArrayLayer = 0, + .layerCount = VK_REMAINING_ARRAY_LAYERS, }, + }; + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, + 0, nullptr, nullptr, read_barriers); + if (is_resolve) { + cmdbuf.ResolveImage(src_image, VK_IMAGE_LAYOUT_GENERAL, dst_image, + VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + MakeImageResolve(dst_region, src_region, dst_layers, src_layers)); + } else { + const bool is_linear = filter == Fermi2D::Filter::Bilinear; + const VkFilter vk_filter = is_linear ? VK_FILTER_LINEAR : VK_FILTER_NEAREST; + cmdbuf.BlitImage( + src_image, VK_IMAGE_LAYOUT_GENERAL, dst_image, VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + MakeImageBlit(dst_region, src_region, dst_layers, src_layers), vk_filter); + } + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, + 0, write_barrier); }); - - return last_image_view = *image_view; } -VkImageView CachedSurfaceView::GetAttachment() { - if (render_target) { - return *render_target; +void TextureCacheRuntime::ConvertImage(Framebuffer* dst, ImageView& dst_view, ImageView& src_view) { + switch (dst_view.format) { + case PixelFormat::R16_UNORM: + if (src_view.format == PixelFormat::D16_UNORM) { + return blit_image_helper.ConvertD16ToR16(dst, src_view); + } + break; + case PixelFormat::R32_FLOAT: + if (src_view.format == PixelFormat::D32_FLOAT) { + return blit_image_helper.ConvertD32ToR32(dst, src_view); + } + break; + case PixelFormat::D16_UNORM: + if (src_view.format == PixelFormat::R16_UNORM) { + return blit_image_helper.ConvertR16ToD16(dst, src_view); + } + break; + case PixelFormat::D32_FLOAT: + if (src_view.format == PixelFormat::R32_FLOAT) { + return blit_image_helper.ConvertR32ToD32(dst, src_view); + } + break; + default: + break; } + UNIMPLEMENTED_MSG("Unimplemented format copy from {} to {}", src_view.format, dst_view.format); +} - VkImageViewCreateInfo ci{ - .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, - .pNext = nullptr, - .flags = 0, - .image = surface.GetImageHandle(), - .viewType = VK_IMAGE_VIEW_TYPE_1D, - .format = surface.GetImage().GetFormat(), - .components = - { - .r = VK_COMPONENT_SWIZZLE_IDENTITY, - .g = VK_COMPONENT_SWIZZLE_IDENTITY, - .b = VK_COMPONENT_SWIZZLE_IDENTITY, - .a = VK_COMPONENT_SWIZZLE_IDENTITY, +void TextureCacheRuntime::CopyImage(Image& dst, Image& src, + std::span<const VideoCommon::ImageCopy> copies) { + std::vector<VkImageCopy> vk_copies(copies.size()); + const VkImageAspectFlags aspect_mask = dst.AspectMask(); + ASSERT(aspect_mask == src.AspectMask()); + + std::ranges::transform(copies, vk_copies.begin(), [aspect_mask](const auto& copy) { + return MakeImageCopy(copy, aspect_mask); + }); + const VkImage dst_image = dst.Handle(); + const VkImage src_image = src.Handle(); + scheduler.RequestOutsideRenderPassOperationContext(); + scheduler.Record([dst_image, src_image, aspect_mask, vk_copies](vk::CommandBuffer cmdbuf) { + RangedBarrierRange dst_range; + RangedBarrierRange src_range; + for (const VkImageCopy& copy : vk_copies) { + dst_range.AddLayers(copy.dstSubresource); + src_range.AddLayers(copy.srcSubresource); + } + const std::array read_barriers{ + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_WRITE_BIT | VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = src_image, + .subresourceRange = src_range.SubresourceRange(aspect_mask), }, - .subresourceRange = - { - .aspectMask = aspect_mask, - .baseMipLevel = base_level, - .levelCount = num_levels, - .baseArrayLayer = 0, - .layerCount = 0, + VkImageMemoryBarrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | + VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_GENERAL, + .newLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = dst_image, + .subresourceRange = dst_range.SubresourceRange(aspect_mask), }, - }; - if (image_view_type == VK_IMAGE_VIEW_TYPE_3D) { - ci.viewType = num_slices > 1 ? VK_IMAGE_VIEW_TYPE_2D_ARRAY : VK_IMAGE_VIEW_TYPE_2D; - ci.subresourceRange.baseArrayLayer = base_slice; - ci.subresourceRange.layerCount = num_slices; + }; + const VkImageMemoryBarrier write_barrier{ + .sType = VK_STRUCTURE_TYPE_IMAGE_MEMORY_BARRIER, + .pNext = nullptr, + .srcAccessMask = VK_ACCESS_TRANSFER_WRITE_BIT, + .dstAccessMask = VK_ACCESS_SHADER_READ_BIT | VK_ACCESS_SHADER_WRITE_BIT | + VK_ACCESS_COLOR_ATTACHMENT_READ_BIT | + VK_ACCESS_COLOR_ATTACHMENT_WRITE_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_READ_BIT | + VK_ACCESS_DEPTH_STENCIL_ATTACHMENT_WRITE_BIT | + VK_ACCESS_TRANSFER_READ_BIT | VK_ACCESS_TRANSFER_WRITE_BIT, + .oldLayout = VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, + .newLayout = VK_IMAGE_LAYOUT_GENERAL, + .srcQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .dstQueueFamilyIndex = VK_QUEUE_FAMILY_IGNORED, + .image = dst_image, + .subresourceRange = dst_range.SubresourceRange(aspect_mask), + }; + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, + 0, {}, {}, read_barriers); + cmdbuf.CopyImage(src_image, VK_IMAGE_LAYOUT_GENERAL, dst_image, + VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, vk_copies); + cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_ALL_COMMANDS_BIT, + 0, write_barrier); + }); +} + +Image::Image(TextureCacheRuntime& runtime, const ImageInfo& info_, GPUVAddr gpu_addr_, + VAddr cpu_addr_) + : VideoCommon::ImageBase(info_, gpu_addr_, cpu_addr_), scheduler{&runtime.scheduler}, + image(MakeImage(runtime.device, info)), buffer(MakeBuffer(runtime.device, info)), + aspect_mask(ImageAspectMask(info.format)) { + if (image) { + commit = runtime.memory_manager.Commit(image, false); } else { - ci.viewType = image_view_type; - ci.subresourceRange.baseArrayLayer = base_layer; - ci.subresourceRange.layerCount = num_layers; + commit = runtime.memory_manager.Commit(buffer, false); + } + if (IsPixelFormatASTC(info.format) && !runtime.device.IsOptimalAstcSupported()) { + flags |= VideoCommon::ImageFlagBits::Converted; + } + if (runtime.device.HasDebuggingToolAttached()) { + if (image) { + image.SetObjectNameEXT(VideoCommon::Name(*this).c_str()); + } else { + buffer.SetObjectNameEXT(VideoCommon::Name(*this).c_str()); + } } - render_target = device.GetLogical().CreateImageView(ci); - return *render_target; } -VKTextureCache::VKTextureCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, - Tegra::MemoryManager& gpu_memory, const VKDevice& device_, - VKMemoryManager& memory_manager_, VKScheduler& scheduler_, - VKStagingBufferPool& staging_pool_) - : TextureCache(rasterizer, maxwell3d, gpu_memory, device_.IsOptimalAstcSupported()), - device{device_}, memory_manager{memory_manager_}, scheduler{scheduler_}, staging_pool{ - staging_pool_} {} - -VKTextureCache::~VKTextureCache() = default; - -Surface VKTextureCache::CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) { - return std::make_shared<CachedSurface>(device, memory_manager, scheduler, staging_pool, - gpu_addr, params); +void Image::UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const BufferImageCopy> copies) { + // TODO: Move this to another API + scheduler->RequestOutsideRenderPassOperationContext(); + std::vector vk_copies = TransformBufferImageCopies(copies, buffer_offset, aspect_mask); + const VkBuffer src_buffer = map.handle; + const VkImage vk_image = *image; + const VkImageAspectFlags vk_aspect_mask = aspect_mask; + const bool is_initialized = std::exchange(initialized, true); + scheduler->Record([src_buffer, vk_image, vk_aspect_mask, is_initialized, + vk_copies](vk::CommandBuffer cmdbuf) { + CopyBufferToImage(cmdbuf, src_buffer, vk_image, vk_aspect_mask, is_initialized, vk_copies); + }); } -void VKTextureCache::ImageCopy(Surface& src_surface, Surface& dst_surface, - const VideoCommon::CopyParams& copy_params) { - const bool src_3d = src_surface->GetSurfaceParams().target == SurfaceTarget::Texture3D; - const bool dst_3d = dst_surface->GetSurfaceParams().target == SurfaceTarget::Texture3D; - UNIMPLEMENTED_IF(src_3d); +void Image::UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferCopy> copies) { + // TODO: Move this to another API + scheduler->RequestOutsideRenderPassOperationContext(); + std::vector vk_copies = TransformBufferCopies(copies, buffer_offset); + const VkBuffer src_buffer = map.handle; + const VkBuffer dst_buffer = *buffer; + scheduler->Record([src_buffer, dst_buffer, vk_copies](vk::CommandBuffer cmdbuf) { + // TODO: Barriers + cmdbuf.CopyBuffer(src_buffer, dst_buffer, vk_copies); + }); +} - // The texture cache handles depth in OpenGL terms, we have to handle it as subresource and - // dimension respectively. - const u32 dst_base_layer = dst_3d ? 0 : copy_params.dest_z; - const u32 dst_offset_z = dst_3d ? copy_params.dest_z : 0; +void Image::DownloadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const BufferImageCopy> copies) { + std::vector vk_copies = TransformBufferImageCopies(copies, buffer_offset, aspect_mask); + scheduler->Record([buffer = map.handle, image = *image, aspect_mask = aspect_mask, + vk_copies](vk::CommandBuffer cmdbuf) { + // TODO: Barriers + cmdbuf.CopyImageToBuffer(image, VK_IMAGE_LAYOUT_GENERAL, buffer, vk_copies); + }); +} - const u32 extent_z = dst_3d ? copy_params.depth : 1; - const u32 num_layers = dst_3d ? 1 : copy_params.depth; +ImageView::ImageView(TextureCacheRuntime& runtime, const VideoCommon::ImageViewInfo& info, + ImageId image_id_, Image& image) + : VideoCommon::ImageViewBase{info, image.info, image_id_}, device{&runtime.device}, + image_handle{image.Handle()}, image_format{image.info.format}, samples{ConvertSampleCount( + image.info.num_samples)} { + const VkImageAspectFlags aspect_mask = ImageViewAspectMask(info); + std::array<SwizzleSource, 4> swizzle{ + SwizzleSource::R, + SwizzleSource::G, + SwizzleSource::B, + SwizzleSource::A, + }; + if (!info.IsRenderTarget()) { + swizzle = info.Swizzle(); + if ((aspect_mask & (VK_IMAGE_ASPECT_DEPTH_BIT | VK_IMAGE_ASPECT_STENCIL_BIT)) != 0) { + std::ranges::transform(swizzle, swizzle.begin(), ConvertGreenRed); + } + } + const VkFormat vk_format = + MaxwellToVK::SurfaceFormat(*device, FormatType::Optimal, format).format; + const VkImageViewCreateInfo create_info{ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .image = image.Handle(), + .viewType = VkImageViewType{}, + .format = vk_format, + .components{ + .r = ComponentSwizzle(swizzle[0]), + .g = ComponentSwizzle(swizzle[1]), + .b = ComponentSwizzle(swizzle[2]), + .a = ComponentSwizzle(swizzle[3]), + }, + .subresourceRange = MakeSubresourceRange(aspect_mask, info.range), + }; + const auto create = [&](VideoCommon::ImageViewType view_type, std::optional<u32> num_layers) { + VkImageViewCreateInfo ci{create_info}; + ci.viewType = ImageViewType(view_type); + if (num_layers) { + ci.subresourceRange.layerCount = *num_layers; + } + vk::ImageView handle = device->GetLogical().CreateImageView(ci); + if (device->HasDebuggingToolAttached()) { + handle.SetObjectNameEXT(VideoCommon::Name(*this, view_type).c_str()); + } + image_views[static_cast<size_t>(view_type)] = std::move(handle); + }; + switch (info.type) { + case VideoCommon::ImageViewType::e1D: + case VideoCommon::ImageViewType::e1DArray: + create(VideoCommon::ImageViewType::e1D, 1); + create(VideoCommon::ImageViewType::e1DArray, std::nullopt); + render_target = Handle(VideoCommon::ImageViewType::e1DArray); + break; + case VideoCommon::ImageViewType::e2D: + case VideoCommon::ImageViewType::e2DArray: + create(VideoCommon::ImageViewType::e2D, 1); + create(VideoCommon::ImageViewType::e2DArray, std::nullopt); + render_target = Handle(VideoCommon::ImageViewType::e2DArray); + break; + case VideoCommon::ImageViewType::e3D: + create(VideoCommon::ImageViewType::e3D, std::nullopt); + render_target = Handle(VideoCommon::ImageViewType::e3D); + break; + case VideoCommon::ImageViewType::Cube: + case VideoCommon::ImageViewType::CubeArray: + create(VideoCommon::ImageViewType::Cube, 6); + create(VideoCommon::ImageViewType::CubeArray, std::nullopt); + break; + case VideoCommon::ImageViewType::Rect: + UNIMPLEMENTED(); + break; + case VideoCommon::ImageViewType::Buffer: + buffer_view = device->GetLogical().CreateBufferView(VkBufferViewCreateInfo{ + .sType = VK_STRUCTURE_TYPE_BUFFER_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .buffer = image.Buffer(), + .format = vk_format, + .offset = 0, // TODO: Redesign buffer cache to support this + .range = image.guest_size_bytes, + }); + break; + } +} - // We can't copy inside a renderpass - scheduler.RequestOutsideRenderPassOperationContext(); +ImageView::ImageView(TextureCacheRuntime&, const VideoCommon::NullImageParams& params) + : VideoCommon::ImageViewBase{params} {} - src_surface->Transition(copy_params.source_z, copy_params.depth, copy_params.source_level, 1, - VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_READ_BIT, - VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL); - dst_surface->Transition(dst_base_layer, num_layers, copy_params.dest_level, 1, - VK_PIPELINE_STAGE_TRANSFER_BIT, VK_ACCESS_TRANSFER_WRITE_BIT, - VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL); +VkImageView ImageView::DepthView() { + if (depth_view) { + return *depth_view; + } + depth_view = MakeDepthStencilView(VK_IMAGE_ASPECT_DEPTH_BIT); + return *depth_view; +} - const VkImageCopy copy{ - .srcSubresource = - { - .aspectMask = src_surface->GetAspectMask(), - .mipLevel = copy_params.source_level, - .baseArrayLayer = copy_params.source_z, - .layerCount = num_layers, - }, - .srcOffset = - { - .x = static_cast<s32>(copy_params.source_x), - .y = static_cast<s32>(copy_params.source_y), - .z = 0, - }, - .dstSubresource = - { - .aspectMask = dst_surface->GetAspectMask(), - .mipLevel = copy_params.dest_level, - .baseArrayLayer = dst_base_layer, - .layerCount = num_layers, - }, - .dstOffset = - { - .x = static_cast<s32>(copy_params.dest_x), - .y = static_cast<s32>(copy_params.dest_y), - .z = static_cast<s32>(dst_offset_z), - }, - .extent = - { - .width = copy_params.width, - .height = copy_params.height, - .depth = extent_z, - }, - }; +VkImageView ImageView::StencilView() { + if (stencil_view) { + return *stencil_view; + } + stencil_view = MakeDepthStencilView(VK_IMAGE_ASPECT_STENCIL_BIT); + return *stencil_view; +} - const VkImage src_image = src_surface->GetImageHandle(); - const VkImage dst_image = dst_surface->GetImageHandle(); - scheduler.Record([src_image, dst_image, copy](vk::CommandBuffer cmdbuf) { - cmdbuf.CopyImage(src_image, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, dst_image, - VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, copy); +vk::ImageView ImageView::MakeDepthStencilView(VkImageAspectFlags aspect_mask) { + return device->GetLogical().CreateImageView({ + .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .image = image_handle, + .viewType = ImageViewType(type), + .format = MaxwellToVK::SurfaceFormat(*device, FormatType::Optimal, format).format, + .components{ + .r = VK_COMPONENT_SWIZZLE_IDENTITY, + .g = VK_COMPONENT_SWIZZLE_IDENTITY, + .b = VK_COMPONENT_SWIZZLE_IDENTITY, + .a = VK_COMPONENT_SWIZZLE_IDENTITY, + }, + .subresourceRange = MakeSubresourceRange(aspect_mask, range), }); } -void VKTextureCache::ImageBlit(View& src_view, View& dst_view, - const Tegra::Engines::Fermi2D::Config& copy_config) { - // We can't blit inside a renderpass - scheduler.RequestOutsideRenderPassOperationContext(); - - src_view->Transition(VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, VK_PIPELINE_STAGE_TRANSFER_BIT, - VK_ACCESS_TRANSFER_READ_BIT); - dst_view->Transition(VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, VK_PIPELINE_STAGE_TRANSFER_BIT, - VK_ACCESS_TRANSFER_WRITE_BIT); - - VkImageBlit blit; - blit.srcSubresource = src_view->GetImageSubresourceLayers(); - blit.srcOffsets[0].x = copy_config.src_rect.left; - blit.srcOffsets[0].y = copy_config.src_rect.top; - blit.srcOffsets[0].z = 0; - blit.srcOffsets[1].x = copy_config.src_rect.right; - blit.srcOffsets[1].y = copy_config.src_rect.bottom; - blit.srcOffsets[1].z = 1; - blit.dstSubresource = dst_view->GetImageSubresourceLayers(); - blit.dstOffsets[0].x = copy_config.dst_rect.left; - blit.dstOffsets[0].y = copy_config.dst_rect.top; - blit.dstOffsets[0].z = 0; - blit.dstOffsets[1].x = copy_config.dst_rect.right; - blit.dstOffsets[1].y = copy_config.dst_rect.bottom; - blit.dstOffsets[1].z = 1; - - const bool is_linear = copy_config.filter == Tegra::Engines::Fermi2D::Filter::Linear; - - scheduler.Record([src_image = src_view->GetImage(), dst_image = dst_view->GetImage(), blit, - is_linear](vk::CommandBuffer cmdbuf) { - cmdbuf.BlitImage(src_image, VK_IMAGE_LAYOUT_TRANSFER_SRC_OPTIMAL, dst_image, - VK_IMAGE_LAYOUT_TRANSFER_DST_OPTIMAL, blit, - is_linear ? VK_FILTER_LINEAR : VK_FILTER_NEAREST); +Sampler::Sampler(TextureCacheRuntime& runtime, const Tegra::Texture::TSCEntry& tsc) { + const auto& device = runtime.device; + const bool arbitrary_borders = runtime.device.IsExtCustomBorderColorSupported(); + const std::array<float, 4> color = tsc.BorderColor(); + // C++20 bit_cast + VkClearColorValue border_color; + std::memcpy(&border_color, &color, sizeof(color)); + const VkSamplerCustomBorderColorCreateInfoEXT border_ci{ + .sType = VK_STRUCTURE_TYPE_SAMPLER_CUSTOM_BORDER_COLOR_CREATE_INFO_EXT, + .pNext = nullptr, + .customBorderColor = border_color, + .format = VK_FORMAT_UNDEFINED, + }; + const void* pnext = nullptr; + if (arbitrary_borders) { + pnext = &border_ci; + } + const VkSamplerReductionModeCreateInfoEXT reduction_ci{ + .sType = VK_STRUCTURE_TYPE_SAMPLER_REDUCTION_MODE_CREATE_INFO_EXT, + .pNext = pnext, + .reductionMode = MaxwellToVK::SamplerReduction(tsc.reduction_filter), + }; + if (runtime.device.IsExtSamplerFilterMinmaxSupported()) { + pnext = &reduction_ci; + } else if (reduction_ci.reductionMode != VK_SAMPLER_REDUCTION_MODE_WEIGHTED_AVERAGE_EXT) { + LOG_WARNING(Render_Vulkan, "VK_EXT_sampler_filter_minmax is required"); + } + // Some games have samplers with garbage. Sanitize them here. + const float max_anisotropy = std::clamp(tsc.MaxAnisotropy(), 1.0f, 16.0f); + sampler = device.GetLogical().CreateSampler(VkSamplerCreateInfo{ + .sType = VK_STRUCTURE_TYPE_SAMPLER_CREATE_INFO, + .pNext = pnext, + .flags = 0, + .magFilter = MaxwellToVK::Sampler::Filter(tsc.mag_filter), + .minFilter = MaxwellToVK::Sampler::Filter(tsc.min_filter), + .mipmapMode = MaxwellToVK::Sampler::MipmapMode(tsc.mipmap_filter), + .addressModeU = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_u, tsc.mag_filter), + .addressModeV = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_v, tsc.mag_filter), + .addressModeW = MaxwellToVK::Sampler::WrapMode(device, tsc.wrap_p, tsc.mag_filter), + .mipLodBias = tsc.LodBias(), + .anisotropyEnable = static_cast<VkBool32>(max_anisotropy > 1.0f ? VK_TRUE : VK_FALSE), + .maxAnisotropy = max_anisotropy, + .compareEnable = tsc.depth_compare_enabled, + .compareOp = MaxwellToVK::Sampler::DepthCompareFunction(tsc.depth_compare_func), + .minLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.0f : tsc.MinLod(), + .maxLod = tsc.mipmap_filter == TextureMipmapFilter::None ? 0.25f : tsc.MaxLod(), + .borderColor = + arbitrary_borders ? VK_BORDER_COLOR_INT_CUSTOM_EXT : ConvertBorderColor(color), + .unnormalizedCoordinates = VK_FALSE, }); } -void VKTextureCache::BufferCopy(Surface& src_surface, Surface& dst_surface) { - // Currently unimplemented. PBO copies should be dropped and we should use a render pass to - // convert from color to depth and viceversa. - LOG_WARNING(Render_Vulkan, "Unimplemented"); +Framebuffer::Framebuffer(TextureCacheRuntime& runtime, std::span<ImageView*, NUM_RT> color_buffers, + ImageView* depth_buffer, const VideoCommon::RenderTargets& key) { + std::vector<VkAttachmentDescription> descriptions; + std::vector<VkImageView> attachments; + RenderPassKey renderpass_key{}; + s32 num_layers = 1; + + for (size_t index = 0; index < NUM_RT; ++index) { + const ImageView* const color_buffer = color_buffers[index]; + if (!color_buffer) { + renderpass_key.color_formats[index] = PixelFormat::Invalid; + continue; + } + descriptions.push_back(AttachmentDescription(runtime.device, color_buffer)); + attachments.push_back(color_buffer->RenderTarget()); + renderpass_key.color_formats[index] = color_buffer->format; + num_layers = std::max(num_layers, color_buffer->range.extent.layers); + images[num_images] = color_buffer->ImageHandle(); + image_ranges[num_images] = MakeSubresourceRange(color_buffer); + samples = color_buffer->Samples(); + ++num_images; + } + const size_t num_colors = attachments.size(); + const VkAttachmentReference* depth_attachment = + depth_buffer ? &ATTACHMENT_REFERENCES[num_colors] : nullptr; + if (depth_buffer) { + descriptions.push_back(AttachmentDescription(runtime.device, depth_buffer)); + attachments.push_back(depth_buffer->RenderTarget()); + renderpass_key.depth_format = depth_buffer->format; + num_layers = std::max(num_layers, depth_buffer->range.extent.layers); + images[num_images] = depth_buffer->ImageHandle(); + image_ranges[num_images] = MakeSubresourceRange(depth_buffer); + samples = depth_buffer->Samples(); + ++num_images; + } else { + renderpass_key.depth_format = PixelFormat::Invalid; + } + renderpass_key.samples = samples; + + const auto& device = runtime.device.GetLogical(); + const auto [cache_pair, is_new] = runtime.renderpass_cache.try_emplace(renderpass_key); + if (is_new) { + const VkSubpassDescription subpass{ + .flags = 0, + .pipelineBindPoint = VK_PIPELINE_BIND_POINT_GRAPHICS, + .inputAttachmentCount = 0, + .pInputAttachments = nullptr, + .colorAttachmentCount = static_cast<u32>(num_colors), + .pColorAttachments = num_colors != 0 ? ATTACHMENT_REFERENCES.data() : nullptr, + .pResolveAttachments = nullptr, + .pDepthStencilAttachment = depth_attachment, + .preserveAttachmentCount = 0, + .pPreserveAttachments = nullptr, + }; + cache_pair->second = device.CreateRenderPass(VkRenderPassCreateInfo{ + .sType = VK_STRUCTURE_TYPE_RENDER_PASS_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .attachmentCount = static_cast<u32>(descriptions.size()), + .pAttachments = descriptions.data(), + .subpassCount = 1, + .pSubpasses = &subpass, + .dependencyCount = 0, + .pDependencies = nullptr, + }); + } + renderpass = *cache_pair->second; + render_area = VkExtent2D{ + .width = key.size.width, + .height = key.size.height, + }; + num_color_buffers = static_cast<u32>(num_colors); + framebuffer = device.CreateFramebuffer(VkFramebufferCreateInfo{ + .sType = VK_STRUCTURE_TYPE_FRAMEBUFFER_CREATE_INFO, + .pNext = nullptr, + .flags = 0, + .renderPass = renderpass, + .attachmentCount = static_cast<u32>(attachments.size()), + .pAttachments = attachments.data(), + .width = key.size.width, + .height = key.size.height, + .layers = static_cast<u32>(num_layers), + }); + if (runtime.device.HasDebuggingToolAttached()) { + framebuffer.SetObjectNameEXT(VideoCommon::Name(key).c_str()); + } } } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h index 39202feba..92a7aad8b 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.h +++ b/src/video_core/renderer_vulkan/vk_texture_cache.h @@ -4,216 +4,270 @@ #pragma once -#include <memory> -#include <unordered_map> +#include <compare> +#include <span> -#include "common/common_types.h" -#include "video_core/renderer_vulkan/vk_image.h" #include "video_core/renderer_vulkan/vk_memory_manager.h" -#include "video_core/renderer_vulkan/vk_scheduler.h" -#include "video_core/renderer_vulkan/wrapper.h" -#include "video_core/texture_cache/surface_base.h" #include "video_core/texture_cache/texture_cache.h" - -namespace VideoCore { -class RasterizerInterface; -} +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class RasterizerVulkan; -class VKDevice; +using VideoCommon::ImageId; +using VideoCommon::NUM_RT; +using VideoCommon::Offset2D; +using VideoCommon::RenderTargets; +using VideoCore::Surface::PixelFormat; + class VKScheduler; class VKStagingBufferPool; -class CachedSurfaceView; -class CachedSurface; +class BlitImageHelper; +class Device; +class Image; +class ImageView; +class Framebuffer; -using Surface = std::shared_ptr<CachedSurface>; -using View = std::shared_ptr<CachedSurfaceView>; -using TextureCacheBase = VideoCommon::TextureCache<Surface, View>; +struct RenderPassKey { + constexpr auto operator<=>(const RenderPassKey&) const noexcept = default; -using VideoCommon::SurfaceParams; -using VideoCommon::ViewParams; + std::array<PixelFormat, NUM_RT> color_formats; + PixelFormat depth_format; + VkSampleCountFlagBits samples; +}; -class CachedSurface final : public VideoCommon::SurfaceBase<View> { - friend CachedSurfaceView; +} // namespace Vulkan -public: - explicit CachedSurface(const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, VKStagingBufferPool& staging_pool, - GPUVAddr gpu_addr, const SurfaceParams& params); - ~CachedSurface(); +namespace std { +template <> +struct hash<Vulkan::RenderPassKey> { + [[nodiscard]] constexpr size_t operator()(const Vulkan::RenderPassKey& key) const noexcept { + size_t value = static_cast<size_t>(key.depth_format) << 48; + value ^= static_cast<size_t>(key.samples) << 52; + for (size_t i = 0; i < key.color_formats.size(); ++i) { + value ^= static_cast<size_t>(key.color_formats[i]) << (i * 6); + } + return value; + } +}; +} // namespace std - void UploadTexture(const std::vector<u8>& staging_buffer) override; - void DownloadTexture(std::vector<u8>& staging_buffer) override; +namespace Vulkan { - void FullTransition(VkPipelineStageFlags new_stage_mask, VkAccessFlags new_access, - VkImageLayout new_layout) { - image->Transition(0, static_cast<u32>(params.GetNumLayers()), 0, params.num_levels, - new_stage_mask, new_access, new_layout); +struct ImageBufferMap { + [[nodiscard]] VkBuffer Handle() const noexcept { + return handle; } - void Transition(u32 base_layer, u32 num_layers, u32 base_level, u32 num_levels, - VkPipelineStageFlags new_stage_mask, VkAccessFlags new_access, - VkImageLayout new_layout) { - image->Transition(base_layer, num_layers, base_level, num_levels, new_stage_mask, - new_access, new_layout); + [[nodiscard]] std::span<u8> Span() const noexcept { + return map.Span(); } - VKImage& GetImage() { - return *image; - } + VkBuffer handle; + MemoryMap map; +}; - const VKImage& GetImage() const { - return *image; +struct TextureCacheRuntime { + const Device& device; + VKScheduler& scheduler; + VKMemoryManager& memory_manager; + VKStagingBufferPool& staging_buffer_pool; + BlitImageHelper& blit_image_helper; + std::unordered_map<RenderPassKey, vk::RenderPass> renderpass_cache; + + void Finish(); + + [[nodiscard]] ImageBufferMap MapUploadBuffer(size_t size); + + [[nodiscard]] ImageBufferMap MapDownloadBuffer(size_t size) { + // TODO: Have a special function for this + return MapUploadBuffer(size); } - VkImage GetImageHandle() const { - return *image->GetHandle(); + void BlitImage(Framebuffer* dst_framebuffer, ImageView& dst, ImageView& src, + const std::array<Offset2D, 2>& dst_region, + const std::array<Offset2D, 2>& src_region, + Tegra::Engines::Fermi2D::Filter filter, + Tegra::Engines::Fermi2D::Operation operation); + + void CopyImage(Image& dst, Image& src, std::span<const VideoCommon::ImageCopy> copies); + + void ConvertImage(Framebuffer* dst, ImageView& dst_view, ImageView& src_view); + + [[nodiscard]] bool CanAccelerateImageUpload(Image&) const noexcept { + return false; } - VkImageAspectFlags GetAspectMask() const { - return image->GetAspectMask(); + void AccelerateImageUpload(Image&, const ImageBufferMap&, size_t, + std::span<const VideoCommon::SwizzleParameters>) { + UNREACHABLE(); } - VkBufferView GetBufferViewHandle() const { - return *buffer_view; + void InsertUploadMemoryBarrier() {} + + bool HasBrokenTextureViewFormats() const noexcept { + // No known Vulkan driver has broken image views + return false; } +}; -protected: - void DecorateSurfaceName(); +class Image : public VideoCommon::ImageBase { +public: + explicit Image(TextureCacheRuntime&, const VideoCommon::ImageInfo& info, GPUVAddr gpu_addr, + VAddr cpu_addr); - View CreateView(const ViewParams& params) override; + void UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferImageCopy> copies); -private: - void UploadBuffer(const std::vector<u8>& staging_buffer); + void UploadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferCopy> copies); - void UploadImage(const std::vector<u8>& staging_buffer); + void DownloadMemory(const ImageBufferMap& map, size_t buffer_offset, + std::span<const VideoCommon::BufferImageCopy> copies); - VkBufferImageCopy GetBufferImageCopy(u32 level) const; + [[nodiscard]] VkImage Handle() const noexcept { + return *image; + } - VkImageSubresourceRange GetImageSubresourceRange() const; + [[nodiscard]] VkBuffer Buffer() const noexcept { + return *buffer; + } - const VKDevice& device; - VKMemoryManager& memory_manager; - VKScheduler& scheduler; - VKStagingBufferPool& staging_pool; + [[nodiscard]] VkImageCreateFlags AspectMask() const noexcept { + return aspect_mask; + } - std::optional<VKImage> image; +private: + VKScheduler* scheduler; + vk::Image image; vk::Buffer buffer; - vk::BufferView buffer_view; VKMemoryCommit commit; - - VkFormat format = VK_FORMAT_UNDEFINED; + VkImageAspectFlags aspect_mask = 0; + bool initialized = false; }; -class CachedSurfaceView final : public VideoCommon::ViewBase { +class ImageView : public VideoCommon::ImageViewBase { public: - explicit CachedSurfaceView(const VKDevice& device, CachedSurface& surface, - const ViewParams& params); - ~CachedSurfaceView(); + explicit ImageView(TextureCacheRuntime&, const VideoCommon::ImageViewInfo&, ImageId, Image&); + explicit ImageView(TextureCacheRuntime&, const VideoCommon::NullImageParams&); - VkImageView GetImageView(Tegra::Texture::SwizzleSource x_source, - Tegra::Texture::SwizzleSource y_source, - Tegra::Texture::SwizzleSource z_source, - Tegra::Texture::SwizzleSource w_source); + [[nodiscard]] VkImageView DepthView(); - VkImageView GetAttachment(); + [[nodiscard]] VkImageView StencilView(); - bool IsSameSurface(const CachedSurfaceView& rhs) const { - return &surface == &rhs.surface; + [[nodiscard]] VkImageView Handle(VideoCommon::ImageViewType query_type) const noexcept { + return *image_views[static_cast<size_t>(query_type)]; } - u32 GetWidth() const { - return params.GetMipWidth(base_level); + [[nodiscard]] VkBufferView BufferView() const noexcept { + return *buffer_view; } - u32 GetHeight() const { - return params.GetMipHeight(base_level); + [[nodiscard]] VkImage ImageHandle() const noexcept { + return image_handle; } - u32 GetNumLayers() const { - return num_layers; + [[nodiscard]] VkImageView RenderTarget() const noexcept { + return render_target; } - bool IsBufferView() const { - return buffer_view; + [[nodiscard]] PixelFormat ImageFormat() const noexcept { + return image_format; } - VkImage GetImage() const { - return image; + [[nodiscard]] VkSampleCountFlagBits Samples() const noexcept { + return samples; } - VkBufferView GetBufferView() const { - return buffer_view; - } +private: + [[nodiscard]] vk::ImageView MakeDepthStencilView(VkImageAspectFlags aspect_mask); - VkImageSubresourceRange GetImageSubresourceRange() const { - return {aspect_mask, base_level, num_levels, base_layer, num_layers}; - } + const Device* device = nullptr; + std::array<vk::ImageView, VideoCommon::NUM_IMAGE_VIEW_TYPES> image_views; + vk::ImageView depth_view; + vk::ImageView stencil_view; + vk::BufferView buffer_view; + VkImage image_handle = VK_NULL_HANDLE; + VkImageView render_target = VK_NULL_HANDLE; + PixelFormat image_format = PixelFormat::Invalid; + VkSampleCountFlagBits samples = VK_SAMPLE_COUNT_1_BIT; +}; - VkImageSubresourceLayers GetImageSubresourceLayers() const { - return {surface.GetAspectMask(), base_level, base_layer, num_layers}; - } +class ImageAlloc : public VideoCommon::ImageAllocBase {}; - void Transition(VkImageLayout new_layout, VkPipelineStageFlags new_stage_mask, - VkAccessFlags new_access) const { - surface.Transition(base_layer, num_layers, base_level, num_levels, new_stage_mask, - new_access, new_layout); - } +class Sampler { +public: + explicit Sampler(TextureCacheRuntime&, const Tegra::Texture::TSCEntry&); - void MarkAsModified(u64 tick) { - surface.MarkAsModified(true, tick); + [[nodiscard]] VkSampler Handle() const noexcept { + return *sampler; } private: - // Store a copy of these values to avoid double dereference when reading them - const SurfaceParams params; - const VkImage image; - const VkBufferView buffer_view; - const VkImageAspectFlags aspect_mask; - - const VKDevice& device; - CachedSurface& surface; - const u32 base_level; - const u32 num_levels; - const VkImageViewType image_view_type; - u32 base_layer = 0; - u32 num_layers = 0; - u32 base_slice = 0; - u32 num_slices = 0; - - VkImageView last_image_view = nullptr; - u32 last_swizzle = 0; - - vk::ImageView render_target; - std::unordered_map<u32, vk::ImageView> view_cache; + vk::Sampler sampler; }; -class VKTextureCache final : public TextureCacheBase { +class Framebuffer { public: - explicit VKTextureCache(VideoCore::RasterizerInterface& rasterizer, - Tegra::Engines::Maxwell3D& maxwell3d, Tegra::MemoryManager& gpu_memory, - const VKDevice& device, VKMemoryManager& memory_manager, - VKScheduler& scheduler, VKStagingBufferPool& staging_pool); - ~VKTextureCache(); + explicit Framebuffer(TextureCacheRuntime&, std::span<ImageView*, NUM_RT> color_buffers, + ImageView* depth_buffer, const VideoCommon::RenderTargets& key); -private: - Surface CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) override; + [[nodiscard]] VkFramebuffer Handle() const noexcept { + return *framebuffer; + } - void ImageCopy(Surface& src_surface, Surface& dst_surface, - const VideoCommon::CopyParams& copy_params) override; + [[nodiscard]] VkRenderPass RenderPass() const noexcept { + return renderpass; + } - void ImageBlit(View& src_view, View& dst_view, - const Tegra::Engines::Fermi2D::Config& copy_config) override; + [[nodiscard]] VkExtent2D RenderArea() const noexcept { + return render_area; + } - void BufferCopy(Surface& src_surface, Surface& dst_surface) override; + [[nodiscard]] VkSampleCountFlagBits Samples() const noexcept { + return samples; + } - const VKDevice& device; - VKMemoryManager& memory_manager; - VKScheduler& scheduler; - VKStagingBufferPool& staging_pool; + [[nodiscard]] u32 NumColorBuffers() const noexcept { + return num_color_buffers; + } + + [[nodiscard]] u32 NumImages() const noexcept { + return num_images; + } + + [[nodiscard]] const std::array<VkImage, 9>& Images() const noexcept { + return images; + } + + [[nodiscard]] const std::array<VkImageSubresourceRange, 9>& ImageRanges() const noexcept { + return image_ranges; + } + +private: + vk::Framebuffer framebuffer; + VkRenderPass renderpass{}; + VkExtent2D render_area{}; + VkSampleCountFlagBits samples = VK_SAMPLE_COUNT_1_BIT; + u32 num_color_buffers = 0; + u32 num_images = 0; + std::array<VkImage, 9> images{}; + std::array<VkImageSubresourceRange, 9> image_ranges{}; }; +struct TextureCacheParams { + static constexpr bool ENABLE_VALIDATION = true; + static constexpr bool FRAMEBUFFER_BLITS = false; + static constexpr bool HAS_EMULATED_COPIES = false; + + using Runtime = Vulkan::TextureCacheRuntime; + using Image = Vulkan::Image; + using ImageAlloc = Vulkan::ImageAlloc; + using ImageView = Vulkan::ImageView; + using Sampler = Vulkan::Sampler; + using Framebuffer = Vulkan::Framebuffer; +}; + +using TextureCache = VideoCommon::TextureCache<TextureCacheParams>; + } // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp index 351c048d2..f99273c6a 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp @@ -7,15 +7,15 @@ #include "common/assert.h" #include "common/logging/log.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_scheduler.h" #include "video_core/renderer_vulkan/vk_update_descriptor.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -VKUpdateDescriptorQueue::VKUpdateDescriptorQueue(const VKDevice& device, VKScheduler& scheduler) - : device{device}, scheduler{scheduler} {} +VKUpdateDescriptorQueue::VKUpdateDescriptorQueue(const Device& device_, VKScheduler& scheduler_) + : device{device_}, scheduler{scheduler_} {} VKUpdateDescriptorQueue::~VKUpdateDescriptorQueue() = default; diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.h b/src/video_core/renderer_vulkan/vk_update_descriptor.h index 945320c72..e214f7195 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.h +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.h @@ -8,11 +8,11 @@ #include <boost/container/static_vector.hpp> #include "common/common_types.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { -class VKDevice; +class Device; class VKScheduler; struct DescriptorUpdateEntry { @@ -31,7 +31,7 @@ struct DescriptorUpdateEntry { class VKUpdateDescriptorQueue final { public: - explicit VKUpdateDescriptorQueue(const VKDevice& device, VKScheduler& scheduler); + explicit VKUpdateDescriptorQueue(const Device& device_, VKScheduler& scheduler_); ~VKUpdateDescriptorQueue(); void TickFrame(); @@ -40,32 +40,36 @@ public: void Send(VkDescriptorUpdateTemplateKHR update_template, VkDescriptorSet set); - void AddSampledImage(VkSampler sampler, VkImageView image_view) { - payload.emplace_back(VkDescriptorImageInfo{sampler, image_view, {}}); + void AddSampledImage(VkImageView image_view, VkSampler sampler) { + payload.emplace_back(VkDescriptorImageInfo{ + .sampler = sampler, + .imageView = image_view, + .imageLayout = VK_IMAGE_LAYOUT_GENERAL, + }); } void AddImage(VkImageView image_view) { - payload.emplace_back(VkDescriptorImageInfo{{}, image_view, {}}); + payload.emplace_back(VkDescriptorImageInfo{ + .sampler = VK_NULL_HANDLE, + .imageView = image_view, + .imageLayout = VK_IMAGE_LAYOUT_GENERAL, + }); } - void AddBuffer(VkBuffer buffer, u64 offset, std::size_t size) { - payload.emplace_back(VkDescriptorBufferInfo{buffer, offset, size}); + void AddBuffer(VkBuffer buffer, u64 offset, size_t size) { + payload.emplace_back(VkDescriptorBufferInfo{ + .buffer = buffer, + .offset = offset, + .range = size, + }); } void AddTexelBuffer(VkBufferView texel_buffer) { payload.emplace_back(texel_buffer); } - VkImageLayout* LastImageLayout() { - return &payload.back().image.imageLayout; - } - - const VkImageLayout* LastImageLayout() const { - return &payload.back().image.imageLayout; - } - private: - const VKDevice& device; + const Device& device; VKScheduler& scheduler; const DescriptorUpdateEntry* upload_start = nullptr; diff --git a/src/video_core/sampler_cache.cpp b/src/video_core/sampler_cache.cpp deleted file mode 100644 index 53c7ef12d..000000000 --- a/src/video_core/sampler_cache.cpp +++ /dev/null @@ -1,21 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "common/cityhash.h" -#include "common/common_types.h" -#include "video_core/sampler_cache.h" - -namespace VideoCommon { - -std::size_t SamplerCacheKey::Hash() const { - static_assert(sizeof(raw) % sizeof(u64) == 0); - return static_cast<std::size_t>( - Common::CityHash64(reinterpret_cast<const char*>(raw.data()), sizeof(raw) / sizeof(u64))); -} - -bool SamplerCacheKey::operator==(const SamplerCacheKey& rhs) const { - return raw == rhs.raw; -} - -} // namespace VideoCommon diff --git a/src/video_core/sampler_cache.h b/src/video_core/sampler_cache.h deleted file mode 100644 index cbe3ad071..000000000 --- a/src/video_core/sampler_cache.h +++ /dev/null @@ -1,60 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <cstddef> -#include <unordered_map> - -#include "video_core/textures/texture.h" - -namespace VideoCommon { - -struct SamplerCacheKey final : public Tegra::Texture::TSCEntry { - std::size_t Hash() const; - - bool operator==(const SamplerCacheKey& rhs) const; - - bool operator!=(const SamplerCacheKey& rhs) const { - return !operator==(rhs); - } -}; - -} // namespace VideoCommon - -namespace std { - -template <> -struct hash<VideoCommon::SamplerCacheKey> { - std::size_t operator()(const VideoCommon::SamplerCacheKey& k) const noexcept { - return k.Hash(); - } -}; - -} // namespace std - -namespace VideoCommon { - -template <typename SamplerType, typename SamplerStorageType> -class SamplerCache { -public: - SamplerType GetSampler(const Tegra::Texture::TSCEntry& tsc) { - const auto [entry, is_cache_miss] = cache.try_emplace(SamplerCacheKey{tsc}); - auto& sampler = entry->second; - if (is_cache_miss) { - sampler = CreateSampler(tsc); - } - return ToSamplerType(sampler); - } - -protected: - virtual SamplerStorageType CreateSampler(const Tegra::Texture::TSCEntry& tsc) const = 0; - - virtual SamplerType ToSamplerType(const SamplerStorageType& sampler) const = 0; - -private: - std::unordered_map<SamplerCacheKey, SamplerStorageType> cache; -}; - -} // namespace VideoCommon
\ No newline at end of file diff --git a/src/video_core/shader/ast.cpp b/src/video_core/shader/ast.cpp index 3f96d9076..db11144c7 100644 --- a/src/video_core/shader/ast.cpp +++ b/src/video_core/shader/ast.cpp @@ -212,16 +212,15 @@ public: } void operator()(const ExprPredicate& expr) { - inner += "P" + std::to_string(expr.predicate); + inner += fmt::format("P{}", expr.predicate); } void operator()(const ExprCondCode& expr) { - u32 cc = static_cast<u32>(expr.cc); - inner += "CC" + std::to_string(cc); + inner += fmt::format("CC{}", expr.cc); } void operator()(const ExprVar& expr) { - inner += "V" + std::to_string(expr.var_index); + inner += fmt::format("V{}", expr.var_index); } void operator()(const ExprBoolean& expr) { @@ -229,7 +228,7 @@ public: } void operator()(const ExprGprEqual& expr) { - inner += "( gpr_" + std::to_string(expr.gpr) + " == " + std::to_string(expr.value) + ')'; + inner += fmt::format("(gpr_{} == {})", expr.gpr, expr.value); } const std::string& GetResult() const { @@ -374,8 +373,8 @@ std::string ASTManager::Print() const { return printer.GetResult(); } -ASTManager::ASTManager(bool full_decompile, bool disable_else_derivation) - : full_decompile{full_decompile}, disable_else_derivation{disable_else_derivation} {}; +ASTManager::ASTManager(bool do_full_decompile, bool disable_else_derivation_) + : full_decompile{do_full_decompile}, disable_else_derivation{disable_else_derivation_} {} ASTManager::~ASTManager() { Clear(); diff --git a/src/video_core/shader/ast.h b/src/video_core/shader/ast.h index 8e5a22ab3..dc49b369e 100644 --- a/src/video_core/shader/ast.h +++ b/src/video_core/shader/ast.h @@ -76,7 +76,7 @@ public: class ASTIfThen { public: - explicit ASTIfThen(Expr condition) : condition{std::move(condition)} {} + explicit ASTIfThen(Expr condition_) : condition{std::move(condition_)} {} Expr condition; ASTZipper nodes{}; }; @@ -88,63 +88,68 @@ public: class ASTBlockEncoded { public: - explicit ASTBlockEncoded(u32 start, u32 end) : start{start}, end{end} {} + explicit ASTBlockEncoded(u32 start_, u32 _) : start{start_}, end{_} {} u32 start; u32 end; }; class ASTBlockDecoded { public: - explicit ASTBlockDecoded(NodeBlock&& new_nodes) : nodes(std::move(new_nodes)) {} + explicit ASTBlockDecoded(NodeBlock&& new_nodes_) : nodes(std::move(new_nodes_)) {} NodeBlock nodes; }; class ASTVarSet { public: - explicit ASTVarSet(u32 index, Expr condition) : index{index}, condition{std::move(condition)} {} + explicit ASTVarSet(u32 index_, Expr condition_) + : index{index_}, condition{std::move(condition_)} {} + u32 index; Expr condition; }; class ASTLabel { public: - explicit ASTLabel(u32 index) : index{index} {} + explicit ASTLabel(u32 index_) : index{index_} {} u32 index; bool unused{}; }; class ASTGoto { public: - explicit ASTGoto(Expr condition, u32 label) : condition{std::move(condition)}, label{label} {} + explicit ASTGoto(Expr condition_, u32 label_) + : condition{std::move(condition_)}, label{label_} {} + Expr condition; u32 label; }; class ASTDoWhile { public: - explicit ASTDoWhile(Expr condition) : condition{std::move(condition)} {} + explicit ASTDoWhile(Expr condition_) : condition{std::move(condition_)} {} Expr condition; ASTZipper nodes{}; }; class ASTReturn { public: - explicit ASTReturn(Expr condition, bool kills) - : condition{std::move(condition)}, kills{kills} {} + explicit ASTReturn(Expr condition_, bool kills_) + : condition{std::move(condition_)}, kills{kills_} {} + Expr condition; bool kills; }; class ASTBreak { public: - explicit ASTBreak(Expr condition) : condition{std::move(condition)} {} + explicit ASTBreak(Expr condition_) : condition{std::move(condition_)} {} Expr condition; }; class ASTBase { public: - explicit ASTBase(ASTNode parent, ASTData data) - : data{std::move(data)}, parent{std::move(parent)} {} + explicit ASTBase(ASTNode parent_, ASTData data_) + : data{std::move(data_)}, parent{std::move(parent_)} {} template <class U, class... Args> static ASTNode Make(ASTNode parent, Args&&... args) { @@ -300,7 +305,7 @@ private: class ASTManager final { public: - ASTManager(bool full_decompile, bool disable_else_derivation); + explicit ASTManager(bool do_full_decompile, bool disable_else_derivation_); ~ASTManager(); ASTManager(const ASTManager& o) = delete; diff --git a/src/video_core/shader/async_shaders.cpp b/src/video_core/shader/async_shaders.cpp index aabd62c5c..9707136e9 100644 --- a/src/video_core/shader/async_shaders.cpp +++ b/src/video_core/shader/async_shaders.cpp @@ -13,21 +13,22 @@ namespace VideoCommon::Shader { -AsyncShaders::AsyncShaders(Core::Frontend::EmuWindow& emu_window) : emu_window(emu_window) {} +AsyncShaders::AsyncShaders(Core::Frontend::EmuWindow& emu_window_) : emu_window(emu_window_) {} AsyncShaders::~AsyncShaders() { KillWorkers(); } void AsyncShaders::AllocateWorkers() { - // Max worker threads we should allow - constexpr u32 MAX_THREADS = 4; - // Deduce how many threads we can use - const u32 threads_used = std::thread::hardware_concurrency() / 4; - // Always allow at least 1 thread regardless of our settings - const auto max_worker_count = std::max(1U, threads_used); - // Don't use more than MAX_THREADS - const auto num_workers = std::min(max_worker_count, MAX_THREADS); + // Use at least one thread + u32 num_workers = 1; + + // Deduce how many more threads we can use + const u32 thread_count = std::thread::hardware_concurrency(); + if (thread_count >= 8) { + // Increase async workers by 1 for every 2 threads >= 8 + num_workers += 1 + (thread_count - 8) / 2; + } // If we already have workers queued, ignore if (num_workers == worker_threads.size()) { @@ -42,8 +43,8 @@ void AsyncShaders::AllocateWorkers() { // Create workers for (std::size_t i = 0; i < num_workers; i++) { context_list.push_back(emu_window.CreateSharedContext()); - worker_threads.push_back( - std::thread(&AsyncShaders::ShaderCompilerThread, this, context_list[i].get())); + worker_threads.emplace_back(&AsyncShaders::ShaderCompilerThread, this, + context_list[i].get()); } } @@ -105,8 +106,7 @@ std::vector<AsyncShaders::Result> AsyncShaders::GetCompletedWork() { std::vector<Result> results; { std::unique_lock lock{completed_mutex}; - results.assign(std::make_move_iterator(finished_work.begin()), - std::make_move_iterator(finished_work.end())); + results = std::move(finished_work); finished_work.clear(); } return results; @@ -115,11 +115,10 @@ std::vector<AsyncShaders::Result> AsyncShaders::GetCompletedWork() { void AsyncShaders::QueueOpenGLShader(const OpenGL::Device& device, Tegra::Engines::ShaderType shader_type, u64 uid, std::vector<u64> code, std::vector<u64> code_b, - u32 main_offset, - VideoCommon::Shader::CompilerSettings compiler_settings, - const VideoCommon::Shader::Registry& registry, - VAddr cpu_addr) { - WorkerParams params{ + u32 main_offset, CompilerSettings compiler_settings, + const Registry& registry, VAddr cpu_addr) { + std::unique_lock lock(queue_mutex); + pending_queue.push({ .backend = device.UseAssemblyShaders() ? Backend::GLASM : Backend::OpenGL, .device = &device, .shader_type = shader_type, @@ -130,35 +129,30 @@ void AsyncShaders::QueueOpenGLShader(const OpenGL::Device& device, .compiler_settings = compiler_settings, .registry = registry, .cpu_address = cpu_addr, - }; - std::unique_lock lock(queue_mutex); - pending_queue.push(std::move(params)); + }); cv.notify_one(); } void AsyncShaders::QueueVulkanShader(Vulkan::VKPipelineCache* pp_cache, - const Vulkan::VKDevice& device, Vulkan::VKScheduler& scheduler, + const Vulkan::Device& device, Vulkan::VKScheduler& scheduler, Vulkan::VKDescriptorPool& descriptor_pool, Vulkan::VKUpdateDescriptorQueue& update_descriptor_queue, - Vulkan::VKRenderPassCache& renderpass_cache, std::vector<VkDescriptorSetLayoutBinding> bindings, Vulkan::SPIRVProgram program, - Vulkan::GraphicsPipelineCacheKey key) { - WorkerParams params{ + Vulkan::GraphicsPipelineCacheKey key, u32 num_color_buffers) { + std::unique_lock lock(queue_mutex); + pending_queue.push({ .backend = Backend::Vulkan, .pp_cache = pp_cache, .vk_device = &device, .scheduler = &scheduler, .descriptor_pool = &descriptor_pool, .update_descriptor_queue = &update_descriptor_queue, - .renderpass_cache = &renderpass_cache, - .bindings = bindings, - .program = program, + .bindings = std::move(bindings), + .program = std::move(program), .key = key, - }; - - std::unique_lock lock(queue_mutex); - pending_queue.push(std::move(params)); + .num_color_buffers = num_color_buffers, + }); cv.notify_one(); } @@ -210,8 +204,8 @@ void AsyncShaders::ShaderCompilerThread(Core::Frontend::GraphicsContext* context } else if (work.backend == Backend::Vulkan) { auto pipeline = std::make_unique<Vulkan::VKGraphicsPipeline>( *work.vk_device, *work.scheduler, *work.descriptor_pool, - *work.update_descriptor_queue, *work.renderpass_cache, work.key, work.bindings, - work.program); + *work.update_descriptor_queue, work.key, work.bindings, work.program, + work.num_color_buffers); work.pp_cache->EmplacePipeline(std::move(pipeline)); } diff --git a/src/video_core/shader/async_shaders.h b/src/video_core/shader/async_shaders.h index 7a99e1dc5..0dbb1a31f 100644 --- a/src/video_core/shader/async_shaders.h +++ b/src/video_core/shader/async_shaders.h @@ -24,9 +24,9 @@ #include "video_core/renderer_opengl/gl_device.h" #include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_shader_decompiler.h" -#include "video_core/renderer_vulkan/vk_device.h" #include "video_core/renderer_vulkan/vk_pipeline_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" +#include "video_core/vulkan_common/vulkan_device.h" namespace Core::Frontend { class EmuWindow; @@ -66,7 +66,7 @@ public: Tegra::Engines::ShaderType shader_type; }; - explicit AsyncShaders(Core::Frontend::EmuWindow& emu_window); + explicit AsyncShaders(Core::Frontend::EmuWindow& emu_window_); ~AsyncShaders(); /// Start up shader worker threads @@ -94,13 +94,13 @@ public: CompilerSettings compiler_settings, const Registry& registry, VAddr cpu_addr); - void QueueVulkanShader(Vulkan::VKPipelineCache* pp_cache, const Vulkan::VKDevice& device, + void QueueVulkanShader(Vulkan::VKPipelineCache* pp_cache, const Vulkan::Device& device, Vulkan::VKScheduler& scheduler, Vulkan::VKDescriptorPool& descriptor_pool, Vulkan::VKUpdateDescriptorQueue& update_descriptor_queue, - Vulkan::VKRenderPassCache& renderpass_cache, std::vector<VkDescriptorSetLayoutBinding> bindings, - Vulkan::SPIRVProgram program, Vulkan::GraphicsPipelineCacheKey key); + Vulkan::SPIRVProgram program, Vulkan::GraphicsPipelineCacheKey key, + u32 num_color_buffers); private: void ShaderCompilerThread(Core::Frontend::GraphicsContext* context); @@ -123,14 +123,14 @@ private: // For Vulkan Vulkan::VKPipelineCache* pp_cache; - const Vulkan::VKDevice* vk_device; + const Vulkan::Device* vk_device; Vulkan::VKScheduler* scheduler; Vulkan::VKDescriptorPool* descriptor_pool; Vulkan::VKUpdateDescriptorQueue* update_descriptor_queue; - Vulkan::VKRenderPassCache* renderpass_cache; std::vector<VkDescriptorSetLayoutBinding> bindings; Vulkan::SPIRVProgram program; Vulkan::GraphicsPipelineCacheKey key; + u32 num_color_buffers; }; std::condition_variable cv; diff --git a/src/video_core/shader/control_flow.cpp b/src/video_core/shader/control_flow.cpp index 4c8971615..43d965f2f 100644 --- a/src/video_core/shader/control_flow.cpp +++ b/src/video_core/shader/control_flow.cpp @@ -66,8 +66,8 @@ struct BlockInfo { }; struct CFGRebuildState { - explicit CFGRebuildState(const ProgramCode& program_code, u32 start, Registry& registry) - : program_code{program_code}, registry{registry}, start{start} {} + explicit CFGRebuildState(const ProgramCode& program_code_, u32 start_, Registry& registry_) + : program_code{program_code_}, registry{registry_}, start{start_} {} const ProgramCode& program_code; Registry& registry; @@ -241,10 +241,10 @@ std::pair<ParseResult, ParseInfo> ParseCode(CFGRebuildState& state, u32 address) ParseInfo parse_info{}; SingleBranch single_branch{}; - const auto insert_label = [](CFGRebuildState& state, u32 address) { - const auto pair = state.labels.emplace(address); + const auto insert_label = [](CFGRebuildState& rebuild_state, u32 label_address) { + const auto pair = rebuild_state.labels.emplace(label_address); if (pair.second) { - state.inspect_queries.push_back(address); + rebuild_state.inspect_queries.push_back(label_address); } }; @@ -257,7 +257,7 @@ std::pair<ParseResult, ParseInfo> ParseCode(CFGRebuildState& state, u32 address) single_branch.ignore = false; break; } - if (state.registered.count(offset) != 0) { + if (state.registered.contains(offset)) { single_branch.address = offset; single_branch.ignore = true; break; @@ -632,12 +632,12 @@ void DecompileShader(CFGRebuildState& state) { for (auto label : state.labels) { state.manager->DeclareLabel(label); } - for (auto& block : state.block_info) { - if (state.labels.count(block.start) != 0) { + for (const auto& block : state.block_info) { + if (state.labels.contains(block.start)) { state.manager->InsertLabel(block.start); } const bool ignore = BlockBranchIsIgnored(block.branch); - u32 end = ignore ? block.end + 1 : block.end; + const u32 end = ignore ? block.end + 1 : block.end; state.manager->InsertBlock(block.start, end); if (!ignore) { InsertBranch(*state.manager, block.branch); @@ -737,7 +737,7 @@ std::unique_ptr<ShaderCharacteristics> ScanFlow(const ProgramCode& program_code, auto back = result_out->blocks.begin(); auto next = std::next(back); while (next != result_out->blocks.end()) { - if (state.labels.count(next->start) == 0 && next->start == back->end + 1) { + if (!state.labels.contains(next->start) && next->start == back->end + 1) { back->end = next->end; next = result_out->blocks.erase(next); continue; diff --git a/src/video_core/shader/control_flow.h b/src/video_core/shader/control_flow.h index 62a3510d8..37bf96492 100644 --- a/src/video_core/shader/control_flow.h +++ b/src/video_core/shader/control_flow.h @@ -42,10 +42,10 @@ struct Condition { class SingleBranch { public: SingleBranch() = default; - SingleBranch(Condition condition, s32 address, bool kill, bool is_sync, bool is_brk, - bool ignore) - : condition{condition}, address{address}, kill{kill}, is_sync{is_sync}, is_brk{is_brk}, - ignore{ignore} {} + explicit SingleBranch(Condition condition_, s32 address_, bool kill_, bool is_sync_, + bool is_brk_, bool ignore_) + : condition{condition_}, address{address_}, kill{kill_}, is_sync{is_sync_}, is_brk{is_brk_}, + ignore{ignore_} {} bool operator==(const SingleBranch& b) const { return std::tie(condition, address, kill, is_sync, is_brk, ignore) == @@ -65,15 +65,15 @@ public: }; struct CaseBranch { - CaseBranch(u32 cmp_value, u32 address) : cmp_value{cmp_value}, address{address} {} + explicit CaseBranch(u32 cmp_value_, u32 address_) : cmp_value{cmp_value_}, address{address_} {} u32 cmp_value; u32 address; }; class MultiBranch { public: - MultiBranch(u32 gpr, std::vector<CaseBranch>&& branches) - : gpr{gpr}, branches{std::move(branches)} {} + explicit MultiBranch(u32 gpr_, std::vector<CaseBranch>&& branches_) + : gpr{gpr_}, branches{std::move(branches_)} {} u32 gpr{}; std::vector<CaseBranch> branches{}; diff --git a/src/video_core/shader/decode.cpp b/src/video_core/shader/decode.cpp index eeac328a6..6576d1208 100644 --- a/src/video_core/shader/decode.cpp +++ b/src/video_core/shader/decode.cpp @@ -25,7 +25,7 @@ using Tegra::Shader::OpCode; namespace { void DeduceTextureHandlerSize(VideoCore::GuestDriverProfile& gpu_driver, - const std::list<Sampler>& used_samplers) { + const std::list<SamplerEntry>& used_samplers) { if (gpu_driver.IsTextureHandlerSizeKnown() || used_samplers.size() <= 1) { return; } @@ -43,9 +43,9 @@ void DeduceTextureHandlerSize(VideoCore::GuestDriverProfile& gpu_driver, } } -std::optional<u32> TryDeduceSamplerSize(const Sampler& sampler_to_deduce, +std::optional<u32> TryDeduceSamplerSize(const SamplerEntry& sampler_to_deduce, VideoCore::GuestDriverProfile& gpu_driver, - const std::list<Sampler>& used_samplers) { + const std::list<SamplerEntry>& used_samplers) { const u32 base_offset = sampler_to_deduce.offset; u32 max_offset{std::numeric_limits<u32>::max()}; for (const auto& sampler : used_samplers) { @@ -66,7 +66,7 @@ std::optional<u32> TryDeduceSamplerSize(const Sampler& sampler_to_deduce, class ASTDecoder { public: - ASTDecoder(ShaderIR& ir) : ir(ir) {} + explicit ASTDecoder(ShaderIR& ir_) : ir(ir_) {} void operator()(ASTProgram& ast) { ASTNode current = ast.nodes.GetFirst(); @@ -153,8 +153,8 @@ void ShaderIR::Decode() { const auto& blocks = shader_info.blocks; NodeBlock current_block; u32 current_label = static_cast<u32>(exit_branch); - for (auto& block : blocks) { - if (shader_info.labels.count(block.start) != 0) { + for (const auto& block : blocks) { + if (shader_info.labels.contains(block.start)) { insert_block(current_block, current_label); current_block.clear(); current_label = block.start; diff --git a/src/video_core/shader/decode/arithmetic.cpp b/src/video_core/shader/decode/arithmetic.cpp index 4db329fa5..15eb700e7 100644 --- a/src/video_core/shader/decode/arithmetic.cpp +++ b/src/video_core/shader/decode/arithmetic.cpp @@ -110,8 +110,7 @@ u32 ShaderIR::DecodeArithmetic(NodeBlock& bb, u32 pc) { case SubOp::Sqrt: return Operation(OperationCode::FSqrt, PRECISE, op_a); default: - UNIMPLEMENTED_MSG("Unhandled MUFU sub op={0:x}", - static_cast<unsigned>(instr.sub_op.Value())); + UNIMPLEMENTED_MSG("Unhandled MUFU sub op={0:x}", instr.sub_op.Value()); return Immediate(0); } }(); @@ -137,7 +136,8 @@ u32 ShaderIR::DecodeArithmetic(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::FCMP_RR: - case OpCode::Id::FCMP_RC: { + case OpCode::Id::FCMP_RC: + case OpCode::Id::FCMP_IMMR: { UNIMPLEMENTED_IF(instr.fcmp.ftz == 0); Node op_c = GetRegister(instr.gpr39); Node comp = GetPredicateComparisonFloat(instr.fcmp.cond, std::move(op_c), Immediate(0.0f)); diff --git a/src/video_core/shader/decode/arithmetic_integer.cpp b/src/video_core/shader/decode/arithmetic_integer.cpp index 73155966f..7b5bb7003 100644 --- a/src/video_core/shader/decode/arithmetic_integer.cpp +++ b/src/video_core/shader/decode/arithmetic_integer.cpp @@ -83,7 +83,7 @@ u32 ShaderIR::DecodeArithmeticInteger(NodeBlock& bb, u32 pc) { case IAdd3Height::UpperHalfWord: return BitfieldExtract(value, 16, 16); default: - UNIMPLEMENTED_MSG("Unhandled IADD3 height: {}", static_cast<u32>(height)); + UNIMPLEMENTED_MSG("Unhandled IADD3 height: {}", height); return Immediate(0); } }; @@ -258,7 +258,7 @@ u32 ShaderIR::DecodeArithmeticInteger(NodeBlock& bb, u32 pc) { case OpCode::Id::LEA_IMM: case OpCode::Id::LEA_RZ: case OpCode::Id::LEA_HI: { - auto [op_a, op_b, op_c] = [&]() -> std::tuple<Node, Node, Node> { + auto [op_a_, op_b_, op_c_] = [&]() -> std::tuple<Node, Node, Node> { switch (opcode->get().GetId()) { case OpCode::Id::LEA_R2: { return {GetRegister(instr.gpr20), GetRegister(instr.gpr39), @@ -294,8 +294,9 @@ u32 ShaderIR::DecodeArithmeticInteger(NodeBlock& bb, u32 pc) { UNIMPLEMENTED_IF_MSG(instr.lea.pred48 != static_cast<u64>(Pred::UnusedIndex), "Unhandled LEA Predicate"); - Node value = Operation(OperationCode::ILogicalShiftLeft, std::move(op_a), std::move(op_c)); - value = Operation(OperationCode::IAdd, std::move(op_b), std::move(value)); + Node value = + Operation(OperationCode::ILogicalShiftLeft, std::move(op_a_), std::move(op_c_)); + value = Operation(OperationCode::IAdd, std::move(op_b_), std::move(value)); SetRegister(bb, instr.gpr0, std::move(value)); break; diff --git a/src/video_core/shader/decode/arithmetic_integer_immediate.cpp b/src/video_core/shader/decode/arithmetic_integer_immediate.cpp index 73880db0e..73580277a 100644 --- a/src/video_core/shader/decode/arithmetic_integer_immediate.cpp +++ b/src/video_core/shader/decode/arithmetic_integer_immediate.cpp @@ -28,23 +28,26 @@ u32 ShaderIR::DecodeArithmeticIntegerImmediate(NodeBlock& bb, u32 pc) { case OpCode::Id::IADD32I: { UNIMPLEMENTED_IF_MSG(instr.iadd32i.saturate, "IADD32I saturation is not implemented"); - op_a = GetOperandAbsNegInteger(op_a, false, instr.iadd32i.negate_a, true); + op_a = GetOperandAbsNegInteger(std::move(op_a), false, instr.iadd32i.negate_a != 0, true); - const Node value = Operation(OperationCode::IAdd, PRECISE, op_a, op_b); + Node value = Operation(OperationCode::IAdd, PRECISE, std::move(op_a), std::move(op_b)); - SetInternalFlagsFromInteger(bb, value, instr.op_32.generates_cc); - SetRegister(bb, instr.gpr0, value); + SetInternalFlagsFromInteger(bb, value, instr.op_32.generates_cc != 0); + SetRegister(bb, instr.gpr0, std::move(value)); break; } case OpCode::Id::LOP32I: { - if (instr.alu.lop32i.invert_a) - op_a = Operation(OperationCode::IBitwiseNot, NO_PRECISE, op_a); + if (instr.alu.lop32i.invert_a) { + op_a = Operation(OperationCode::IBitwiseNot, NO_PRECISE, std::move(op_a)); + } - if (instr.alu.lop32i.invert_b) - op_b = Operation(OperationCode::IBitwiseNot, NO_PRECISE, op_b); + if (instr.alu.lop32i.invert_b) { + op_b = Operation(OperationCode::IBitwiseNot, NO_PRECISE, std::move(op_b)); + } - WriteLogicOperation(bb, instr.gpr0, instr.alu.lop32i.operation, op_a, op_b, - PredicateResultMode::None, Pred::UnusedIndex, instr.op_32.generates_cc); + WriteLogicOperation(bb, instr.gpr0, instr.alu.lop32i.operation, std::move(op_a), + std::move(op_b), PredicateResultMode::None, Pred::UnusedIndex, + instr.op_32.generates_cc != 0); break; } default: @@ -58,18 +61,18 @@ u32 ShaderIR::DecodeArithmeticIntegerImmediate(NodeBlock& bb, u32 pc) { void ShaderIR::WriteLogicOperation(NodeBlock& bb, Register dest, LogicOperation logic_op, Node op_a, Node op_b, PredicateResultMode predicate_mode, Pred predicate, bool sets_cc) { - const Node result = [&]() { + Node result = [&] { switch (logic_op) { case LogicOperation::And: - return Operation(OperationCode::IBitwiseAnd, PRECISE, op_a, op_b); + return Operation(OperationCode::IBitwiseAnd, PRECISE, std::move(op_a), std::move(op_b)); case LogicOperation::Or: - return Operation(OperationCode::IBitwiseOr, PRECISE, op_a, op_b); + return Operation(OperationCode::IBitwiseOr, PRECISE, std::move(op_a), std::move(op_b)); case LogicOperation::Xor: - return Operation(OperationCode::IBitwiseXor, PRECISE, op_a, op_b); + return Operation(OperationCode::IBitwiseXor, PRECISE, std::move(op_a), std::move(op_b)); case LogicOperation::PassB: return op_b; default: - UNIMPLEMENTED_MSG("Unimplemented logic operation={}", static_cast<u32>(logic_op)); + UNIMPLEMENTED_MSG("Unimplemented logic operation={}", logic_op); return Immediate(0); } }(); @@ -84,13 +87,12 @@ void ShaderIR::WriteLogicOperation(NodeBlock& bb, Register dest, LogicOperation return; case PredicateResultMode::NotZero: { // Set the predicate to true if the result is not zero. - const Node compare = Operation(OperationCode::LogicalINotEqual, result, Immediate(0)); - SetPredicate(bb, static_cast<u64>(predicate), compare); + Node compare = Operation(OperationCode::LogicalINotEqual, std::move(result), Immediate(0)); + SetPredicate(bb, static_cast<u64>(predicate), std::move(compare)); break; } default: - UNIMPLEMENTED_MSG("Unimplemented predicate result mode: {}", - static_cast<u32>(predicate_mode)); + UNIMPLEMENTED_MSG("Unimplemented predicate result mode: {}", predicate_mode); } } diff --git a/src/video_core/shader/decode/conversion.cpp b/src/video_core/shader/decode/conversion.cpp index b9989c88c..fea7a54df 100644 --- a/src/video_core/shader/decode/conversion.cpp +++ b/src/video_core/shader/decode/conversion.cpp @@ -244,7 +244,7 @@ u32 ShaderIR::DecodeConversion(NodeBlock& bb, u32 pc) { return Operation(OperationCode::FTrunc, value); default: UNIMPLEMENTED_MSG("Unimplemented F2F rounding mode {}", - static_cast<u32>(instr.conversion.f2f.rounding.Value())); + instr.conversion.f2f.rounding.Value()); return value; } }(); @@ -300,7 +300,7 @@ u32 ShaderIR::DecodeConversion(NodeBlock& bb, u32 pc) { return Operation(OperationCode::FTrunc, PRECISE, value); default: UNIMPLEMENTED_MSG("Unimplemented F2I rounding mode {}", - static_cast<u32>(instr.conversion.f2i.rounding.Value())); + instr.conversion.f2i.rounding.Value()); return Immediate(0); } }(); diff --git a/src/video_core/shader/decode/half_set.cpp b/src/video_core/shader/decode/half_set.cpp index b2e88fa20..fa83108cd 100644 --- a/src/video_core/shader/decode/half_set.cpp +++ b/src/video_core/shader/decode/half_set.cpp @@ -22,13 +22,13 @@ u32 ShaderIR::DecodeHalfSet(NodeBlock& bb, u32 pc) { const Instruction instr = {program_code[pc]}; const auto opcode = OpCode::Decode(instr); - PredCondition cond; - bool bf; - bool ftz; - bool neg_a; - bool abs_a; - bool neg_b; - bool abs_b; + PredCondition cond{}; + bool bf = false; + bool ftz = false; + bool neg_a = false; + bool abs_a = false; + bool neg_b = false; + bool abs_b = false; switch (opcode->get().GetId()) { case OpCode::Id::HSET2_C: case OpCode::Id::HSET2_IMM: diff --git a/src/video_core/shader/decode/image.cpp b/src/video_core/shader/decode/image.cpp index 618d309d2..5470e8cf4 100644 --- a/src/video_core/shader/decode/image.cpp +++ b/src/video_core/shader/decode/image.cpp @@ -212,10 +212,10 @@ u32 GetComponentSize(TextureFormat format, std::size_t component) { return 0; case TextureFormat::R8G24: if (component == 0) { - return 8; + return 24; } if (component == 1) { - return 24; + return 8; } return 0; case TextureFormat::R8G8: @@ -358,9 +358,9 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { instr.suldst.GetStoreDataLayout() != StoreType::Bits64); auto descriptor = [this, instr] { - std::optional<Tegra::Engines::SamplerDescriptor> descriptor; + std::optional<Tegra::Engines::SamplerDescriptor> sampler_descriptor; if (instr.suldst.is_immediate) { - descriptor = + sampler_descriptor = registry.ObtainBoundSampler(static_cast<u32>(instr.image.index.Value())); } else { const Node image_register = GetRegister(instr.gpr39); @@ -368,12 +368,12 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { static_cast<s64>(global_code.size())); const auto buffer = std::get<1>(result); const auto offset = std::get<2>(result); - descriptor = registry.ObtainBindlessSampler(buffer, offset); + sampler_descriptor = registry.ObtainBindlessSampler(buffer, offset); } - if (!descriptor) { + if (!sampler_descriptor) { UNREACHABLE_MSG("Failed to obtain image descriptor"); } - return *descriptor; + return *sampler_descriptor; }(); const auto comp_mask = GetImageComponentMask(descriptor.format); @@ -497,11 +497,12 @@ u32 ShaderIR::DecodeImage(NodeBlock& bb, u32 pc) { return pc; } -Image& ShaderIR::GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type) { +ImageEntry& ShaderIR::GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type) { const auto offset = static_cast<u32>(image.index.Value()); - const auto it = std::find_if(std::begin(used_images), std::end(used_images), - [offset](const Image& entry) { return entry.offset == offset; }); + const auto it = + std::find_if(std::begin(used_images), std::end(used_images), + [offset](const ImageEntry& entry) { return entry.offset == offset; }); if (it != std::end(used_images)) { ASSERT(!it->is_bindless && it->type == type); return *it; @@ -511,7 +512,7 @@ Image& ShaderIR::GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType t return used_images.emplace_back(next_index, offset, type); } -Image& ShaderIR::GetBindlessImage(Tegra::Shader::Register reg, Tegra::Shader::ImageType type) { +ImageEntry& ShaderIR::GetBindlessImage(Tegra::Shader::Register reg, Tegra::Shader::ImageType type) { const Node image_register = GetRegister(reg); const auto result = TrackCbuf(image_register, global_code, static_cast<s64>(global_code.size())); @@ -520,7 +521,7 @@ Image& ShaderIR::GetBindlessImage(Tegra::Shader::Register reg, Tegra::Shader::Im const auto offset = std::get<2>(result); const auto it = std::find_if(std::begin(used_images), std::end(used_images), - [buffer, offset](const Image& entry) { + [buffer, offset](const ImageEntry& entry) { return entry.buffer == buffer && entry.offset == offset; }); if (it != std::end(used_images)) { diff --git a/src/video_core/shader/decode/memory.cpp b/src/video_core/shader/decode/memory.cpp index e2bba88dd..50f4e7d35 100644 --- a/src/video_core/shader/decode/memory.cpp +++ b/src/video_core/shader/decode/memory.cpp @@ -47,7 +47,7 @@ OperationCode GetAtomOperation(AtomicOp op) { case AtomicOp::Exch: return OperationCode::AtomicIExchange; default: - UNIMPLEMENTED_MSG("op={}", static_cast<int>(op)); + UNIMPLEMENTED_MSG("op={}", op); return OperationCode::AtomicIAdd; } } @@ -83,7 +83,7 @@ u32 GetMemorySize(Tegra::Shader::UniformType uniform_type) { case Tegra::Shader::UniformType::UnsignedQuad: return 128; default: - UNIMPLEMENTED_MSG("Unimplemented size={}!", static_cast<u32>(uniform_type)); + UNIMPLEMENTED_MSG("Unimplemented size={}!", uniform_type); return 32; } } @@ -175,12 +175,12 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { break; } default: - UNIMPLEMENTED_MSG("Unhandled type: {}", static_cast<unsigned>(instr.ld_c.type.Value())); + UNIMPLEMENTED_MSG("Unhandled type: {}", instr.ld_c.type.Value()); } break; } case OpCode::Id::LD_L: - LOG_DEBUG(HW_GPU, "LD_L cache management mode: {}", static_cast<u64>(instr.ld_l.unknown)); + LOG_DEBUG(HW_GPU, "LD_L cache management mode: {}", instr.ld_l.unknown); [[fallthrough]]; case OpCode::Id::LD_S: { const auto GetAddress = [&](s32 offset) { @@ -224,7 +224,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { } default: UNIMPLEMENTED_MSG("{} Unhandled type: {}", opcode->get().GetName(), - static_cast<u32>(instr.ldst_sl.type.Value())); + instr.ldst_sl.type.Value()); } break; } @@ -306,8 +306,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::ST_L: - LOG_DEBUG(HW_GPU, "ST_L cache management mode: {}", - static_cast<u64>(instr.st_l.cache_management.Value())); + LOG_DEBUG(HW_GPU, "ST_L cache management mode: {}", instr.st_l.cache_management.Value()); [[fallthrough]]; case OpCode::Id::ST_S: { const auto GetAddress = [&](s32 offset) { @@ -340,7 +339,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { } default: UNIMPLEMENTED_MSG("{} unhandled type: {}", opcode->get().GetName(), - static_cast<u32>(instr.ldst_sl.type.Value())); + instr.ldst_sl.type.Value()); } break; } @@ -387,7 +386,7 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { } case OpCode::Id::RED: { UNIMPLEMENTED_IF_MSG(instr.red.type != GlobalAtomicType::U32, "type={}", - static_cast<int>(instr.red.type.Value())); + instr.red.type.Value()); const auto [real_address, base_address, descriptor] = TrackGlobalMemory(bb, instr, true, true); if (!real_address || !base_address) { @@ -403,12 +402,12 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { UNIMPLEMENTED_IF_MSG(instr.atom.operation == AtomicOp::Inc || instr.atom.operation == AtomicOp::Dec || instr.atom.operation == AtomicOp::SafeAdd, - "operation={}", static_cast<int>(instr.atom.operation.Value())); + "operation={}", instr.atom.operation.Value()); UNIMPLEMENTED_IF_MSG(instr.atom.type == GlobalAtomicType::S64 || instr.atom.type == GlobalAtomicType::U64 || instr.atom.type == GlobalAtomicType::F16x2_FTZ_RN || instr.atom.type == GlobalAtomicType::F32_FTZ_RN, - "type={}", static_cast<int>(instr.atom.type.Value())); + "type={}", instr.atom.type.Value()); const auto [real_address, base_address, descriptor] = TrackGlobalMemory(bb, instr, true, true); @@ -428,10 +427,10 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { case OpCode::Id::ATOMS: { UNIMPLEMENTED_IF_MSG(instr.atoms.operation == AtomicOp::Inc || instr.atoms.operation == AtomicOp::Dec, - "operation={}", static_cast<int>(instr.atoms.operation.Value())); + "operation={}", instr.atoms.operation.Value()); UNIMPLEMENTED_IF_MSG(instr.atoms.type == AtomicType::S64 || instr.atoms.type == AtomicType::U64, - "type={}", static_cast<int>(instr.atoms.type.Value())); + "type={}", instr.atoms.type.Value()); const bool is_signed = instr.atoms.type == AtomicType::S32 || instr.atoms.type == AtomicType::S64; const s32 offset = instr.atoms.GetImmediateOffset(); diff --git a/src/video_core/shader/decode/other.cpp b/src/video_core/shader/decode/other.cpp index 29a7cfbfe..d3ea07aac 100644 --- a/src/video_core/shader/decode/other.cpp +++ b/src/video_core/shader/decode/other.cpp @@ -34,14 +34,13 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::EXIT: { - const Tegra::Shader::ConditionCode cc = instr.flow_condition_code; - UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T, "EXIT condition code used: {}", - static_cast<u32>(cc)); + const ConditionCode cc = instr.flow_condition_code; + UNIMPLEMENTED_IF_MSG(cc != ConditionCode::T, "EXIT condition code used: {}", cc); switch (instr.flow.cond) { case Tegra::Shader::FlowCondition::Always: bb.push_back(Operation(OperationCode::Exit)); - if (instr.pred.pred_index == static_cast<u64>(Tegra::Shader::Pred::UnusedIndex)) { + if (instr.pred.pred_index == static_cast<u64>(Pred::UnusedIndex)) { // If this is an unconditional exit then just end processing here, // otherwise we have to account for the possibility of the condition // not being met, so continue processing the next instruction. @@ -56,17 +55,15 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { break; default: - UNIMPLEMENTED_MSG("Unhandled flow condition: {}", - static_cast<u32>(instr.flow.cond.Value())); + UNIMPLEMENTED_MSG("Unhandled flow condition: {}", instr.flow.cond.Value()); } break; } case OpCode::Id::KIL: { UNIMPLEMENTED_IF(instr.flow.cond != Tegra::Shader::FlowCondition::Always); - const Tegra::Shader::ConditionCode cc = instr.flow_condition_code; - UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T, "KIL condition code used: {}", - static_cast<u32>(cc)); + const ConditionCode cc = instr.flow_condition_code; + UNIMPLEMENTED_IF_MSG(cc != ConditionCode::T, "KIL condition code used: {}", cc); bb.push_back(Operation(OperationCode::Discard)); break; @@ -90,11 +87,11 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { UNIMPLEMENTED_MSG("S2R WscaleFactorZ is not implemented"); return Immediate(0U); case SystemVariable::Tid: { - Node value = Immediate(0); - value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdX), 0, 9); - value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdY), 16, 9); - value = BitfieldInsert(value, Operation(OperationCode::LocalInvocationIdZ), 26, 5); - return value; + Node val = Immediate(0); + val = BitfieldInsert(val, Operation(OperationCode::LocalInvocationIdX), 0, 9); + val = BitfieldInsert(val, Operation(OperationCode::LocalInvocationIdY), 16, 9); + val = BitfieldInsert(val, Operation(OperationCode::LocalInvocationIdZ), 26, 5); + return val; } case SystemVariable::TidX: return Operation(OperationCode::LocalInvocationIdX); @@ -130,8 +127,7 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { return Immediate(0u); } default: - UNIMPLEMENTED_MSG("Unhandled system move: {}", - static_cast<u32>(instr.sys20.Value())); + UNIMPLEMENTED_MSG("Unhandled system move: {}", instr.sys20.Value()); return Immediate(0u); } }(); @@ -181,8 +177,8 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { } const Node branch = Operation(OperationCode::BranchIndirect, operand); - const Tegra::Shader::ConditionCode cc = instr.flow_condition_code; - if (cc != Tegra::Shader::ConditionCode::T) { + const ConditionCode cc = instr.flow_condition_code; + if (cc != ConditionCode::T) { bb.push_back(Conditional(GetConditionCode(cc), {branch})); } else { bb.push_back(branch); @@ -218,9 +214,8 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::SYNC: { - const Tegra::Shader::ConditionCode cc = instr.flow_condition_code; - UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T, "SYNC condition code used: {}", - static_cast<u32>(cc)); + const ConditionCode cc = instr.flow_condition_code; + UNIMPLEMENTED_IF_MSG(cc != ConditionCode::T, "SYNC condition code used: {}", cc); if (decompiled) { break; @@ -231,9 +226,8 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::BRK: { - const Tegra::Shader::ConditionCode cc = instr.flow_condition_code; - UNIMPLEMENTED_IF_MSG(cc != Tegra::Shader::ConditionCode::T, "BRK condition code used: {}", - static_cast<u32>(cc)); + const ConditionCode cc = instr.flow_condition_code; + UNIMPLEMENTED_IF_MSG(cc != ConditionCode::T, "BRK condition code used: {}", cc); if (decompiled) { break; } @@ -306,7 +300,7 @@ u32 ShaderIR::DecodeOther(NodeBlock& bb, u32 pc) { case Tegra::Shader::MembarType::GL: return OperationCode::MemoryBarrierGlobal; default: - UNIMPLEMENTED_MSG("MEMBAR type={}", static_cast<int>(instr.membar.type.Value())); + UNIMPLEMENTED_MSG("MEMBAR type={}", instr.membar.type.Value()); return OperationCode::MemoryBarrierGlobal; } }(); diff --git a/src/video_core/shader/decode/shift.cpp b/src/video_core/shader/decode/shift.cpp index d4ffa8014..a53819c15 100644 --- a/src/video_core/shader/decode/shift.cpp +++ b/src/video_core/shader/decode/shift.cpp @@ -125,7 +125,7 @@ u32 ShaderIR::DecodeShift(NodeBlock& bb, u32 pc) { case OpCode::Id::SHF_LEFT_IMM: { UNIMPLEMENTED_IF(instr.generates_cc); UNIMPLEMENTED_IF_MSG(instr.shf.xmode != ShfXmode::None, "xmode={}", - static_cast<int>(instr.shf.xmode.Value())); + instr.shf.xmode.Value()); if (instr.is_b_imm) { op_b = Immediate(static_cast<u32>(instr.shf.immediate)); diff --git a/src/video_core/shader/decode/texture.cpp b/src/video_core/shader/decode/texture.cpp index a03b50e39..833fa2a39 100644 --- a/src/video_core/shader/decode/texture.cpp +++ b/src/video_core/shader/decode/texture.cpp @@ -34,7 +34,7 @@ static std::size_t GetCoordCount(TextureType texture_type) { case TextureType::TextureCube: return 3; default: - UNIMPLEMENTED_MSG("Unhandled texture type: {}", static_cast<u32>(texture_type)); + UNIMPLEMENTED_MSG("Unhandled texture type: {}", texture_type); return 0; } } @@ -141,7 +141,7 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { SamplerInfo info; info.is_shadow = is_depth_compare; - const std::optional<Sampler> sampler = GetSampler(instr.sampler, info); + const std::optional<SamplerEntry> sampler = GetSampler(instr.sampler, info); Node4 values; for (u32 element = 0; element < values.size(); ++element) { @@ -173,9 +173,9 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { SamplerInfo info; info.type = texture_type; info.is_array = is_array; - const std::optional<Sampler> sampler = is_bindless - ? GetBindlessSampler(base_reg, info, index_var) - : GetSampler(instr.sampler, info); + const std::optional<SamplerEntry> sampler = + is_bindless ? GetBindlessSampler(base_reg, info, index_var) + : GetSampler(instr.sampler, info); Node4 values; if (!sampler) { std::generate(values.begin(), values.end(), [this] { return Immediate(0); }); @@ -217,9 +217,9 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { [[fallthrough]]; case OpCode::Id::TXQ: { Node index_var; - const std::optional<Sampler> sampler = is_bindless - ? GetBindlessSampler(instr.gpr8, {}, index_var) - : GetSampler(instr.sampler, {}); + const std::optional<SamplerEntry> sampler = + is_bindless ? GetBindlessSampler(instr.gpr8, {}, index_var) + : GetSampler(instr.sampler, {}); if (!sampler) { u32 indexer = 0; @@ -255,8 +255,7 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { break; } default: - UNIMPLEMENTED_MSG("Unhandled texture query type: {}", - static_cast<u32>(instr.txq.query_type.Value())); + UNIMPLEMENTED_MSG("Unhandled texture query type: {}", instr.txq.query_type.Value()); } break; } @@ -273,7 +272,7 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { info.type = texture_type; info.is_array = is_array; Node index_var; - const std::optional<Sampler> sampler = + const std::optional<SamplerEntry> sampler = is_bindless ? GetBindlessSampler(instr.gpr20, info, index_var) : GetSampler(instr.sampler, info); @@ -292,33 +291,36 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { break; } - std::vector<Node> coords; - - // TODO: Add coordinates for different samplers once other texture types are implemented. - switch (texture_type) { - case TextureType::Texture1D: - coords.push_back(GetRegister(instr.gpr8)); - break; - case TextureType::Texture2D: - coords.push_back(GetRegister(instr.gpr8.Value() + 0)); - coords.push_back(GetRegister(instr.gpr8.Value() + 1)); - break; - default: - UNIMPLEMENTED_MSG("Unhandled texture type {}", static_cast<int>(texture_type)); + const u64 base_index = is_array ? 1 : 0; + const u64 num_components = [texture_type] { + switch (texture_type) { + case TextureType::Texture1D: + return 1; + case TextureType::Texture2D: + return 2; + case TextureType::TextureCube: + return 3; + default: + UNIMPLEMENTED_MSG("Unhandled texture type {}", texture_type); + return 2; + } + }(); + // TODO: What's the array component used for? - // Fallback to interpreting as a 2D texture for now - coords.push_back(GetRegister(instr.gpr8.Value() + 0)); - coords.push_back(GetRegister(instr.gpr8.Value() + 1)); + std::vector<Node> coords; + coords.reserve(num_components); + for (u64 component = 0; component < num_components; ++component) { + coords.push_back(GetRegister(instr.gpr8.Value() + base_index + component)); } + u32 indexer = 0; for (u32 element = 0; element < 2; ++element) { if (!instr.tmml.IsComponentEnabled(element)) { continue; } - auto params = coords; MetaTexture meta{*sampler, {}, {}, {}, {}, {}, {}, {}, {}, element, index_var}; - const Node value = Operation(OperationCode::TextureQueryLod, meta, std::move(params)); - SetTemporary(bb, indexer++, value); + Node value = Operation(OperationCode::TextureQueryLod, meta, coords); + SetTemporary(bb, indexer++, std::move(value)); } for (u32 i = 0; i < indexer; ++i) { SetRegister(bb, instr.gpr0.Value() + i, GetTemporary(i)); @@ -377,14 +379,15 @@ ShaderIR::SamplerInfo ShaderIR::GetSamplerInfo( return info; } -std::optional<Sampler> ShaderIR::GetSampler(Tegra::Shader::Sampler sampler, - SamplerInfo sampler_info) { +std::optional<SamplerEntry> ShaderIR::GetSampler(Tegra::Shader::Sampler sampler, + SamplerInfo sampler_info) { const u32 offset = static_cast<u32>(sampler.index.Value()); const auto info = GetSamplerInfo(sampler_info, registry.ObtainBoundSampler(offset)); // If this sampler has already been used, return the existing mapping. - const auto it = std::find_if(used_samplers.begin(), used_samplers.end(), - [offset](const Sampler& entry) { return entry.offset == offset; }); + const auto it = + std::find_if(used_samplers.begin(), used_samplers.end(), + [offset](const SamplerEntry& entry) { return entry.offset == offset; }); if (it != used_samplers.end()) { ASSERT(!it->is_bindless && it->type == info.type && it->is_array == info.is_array && it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer); @@ -397,8 +400,8 @@ std::optional<Sampler> ShaderIR::GetSampler(Tegra::Shader::Sampler sampler, *info.is_shadow, *info.is_buffer, false); } -std::optional<Sampler> ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, SamplerInfo info, - Node& index_var) { +std::optional<SamplerEntry> ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, + SamplerInfo info, Node& index_var) { const Node sampler_register = GetRegister(reg); const auto [base_node, tracked_sampler_info] = TrackBindlessSampler(sampler_register, global_code, static_cast<s64>(global_code.size())); @@ -414,7 +417,7 @@ std::optional<Sampler> ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, // If this sampler has already been used, return the existing mapping. const auto it = std::find_if(used_samplers.begin(), used_samplers.end(), - [buffer, offset](const Sampler& entry) { + [buffer, offset](const SamplerEntry& entry) { return entry.buffer == buffer && entry.offset == offset; }); if (it != used_samplers.end()) { @@ -434,11 +437,12 @@ std::optional<Sampler> ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, info = GetSamplerInfo(info, registry.ObtainSeparateSampler(indices, offsets)); // Try to use an already created sampler if it exists - const auto it = std::find_if( - used_samplers.begin(), used_samplers.end(), [indices, offsets](const Sampler& entry) { - return offsets == std::pair{entry.offset, entry.secondary_offset} && - indices == std::pair{entry.buffer, entry.secondary_buffer}; - }); + const auto it = + std::find_if(used_samplers.begin(), used_samplers.end(), + [indices, offsets](const SamplerEntry& entry) { + return offsets == std::pair{entry.offset, entry.secondary_offset} && + indices == std::pair{entry.buffer, entry.secondary_buffer}; + }); if (it != used_samplers.end()) { ASSERT(it->is_separated && it->type == info.type && it->is_array == info.is_array && it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer); @@ -458,7 +462,7 @@ std::optional<Sampler> ShaderIR::GetBindlessSampler(Tegra::Shader::Register reg, // If this sampler has already been used, return the existing mapping. const auto it = std::find_if( used_samplers.begin(), used_samplers.end(), - [base_offset](const Sampler& entry) { return entry.offset == base_offset; }); + [base_offset](const SamplerEntry& entry) { return entry.offset == base_offset; }); if (it != used_samplers.end()) { ASSERT(!it->is_bindless && it->type == info.type && it->is_array == info.is_array && it->is_shadow == info.is_shadow && it->is_buffer == info.is_buffer && @@ -553,7 +557,6 @@ Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, const bool is_shadow = depth_compare != nullptr; const bool is_bindless = bindless_reg.has_value(); - UNIMPLEMENTED_IF(texture_type == TextureType::TextureCube && is_array && is_shadow); ASSERT_MSG(texture_type != TextureType::Texture3D || !is_array || !is_shadow, "Illegal texture type"); @@ -564,9 +567,9 @@ Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, info.is_buffer = false; Node index_var; - const std::optional<Sampler> sampler = is_bindless - ? GetBindlessSampler(*bindless_reg, info, index_var) - : GetSampler(instr.sampler, info); + const std::optional<SamplerEntry> sampler = + is_bindless ? GetBindlessSampler(*bindless_reg, info, index_var) + : GetSampler(instr.sampler, info); if (!sampler) { return {Immediate(0), Immediate(0), Immediate(0), Immediate(0)}; } @@ -593,7 +596,7 @@ Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, lod = GetRegister(instr.gpr20.Value() + bias_offset); break; default: - UNIMPLEMENTED_MSG("Unimplemented process mode={}", static_cast<u32>(process_mode)); + UNIMPLEMENTED_MSG("Unimplemented process mode={}", process_mode); break; } @@ -723,7 +726,7 @@ Node4 ShaderIR::GetTld4Code(Instruction instr, TextureType texture_type, bool de info.is_shadow = depth_compare; Node index_var; - const std::optional<Sampler> sampler = + const std::optional<SamplerEntry> sampler = is_bindless ? GetBindlessSampler(parameter_register++, info, index_var) : GetSampler(instr.sampler, info); Node4 values; @@ -782,7 +785,7 @@ Node4 ShaderIR::GetTldCode(Tegra::Shader::Instruction instr) { // const Node aoffi_register{is_aoffi ? GetRegister(gpr20_cursor++) : nullptr}; // const Node multisample{is_multisample ? GetRegister(gpr20_cursor++) : nullptr}; - const std::optional<Sampler> sampler = GetSampler(instr.sampler, {}); + const std::optional<SamplerEntry> sampler = GetSampler(instr.sampler, {}); Node4 values; for (u32 element = 0; element < values.size(); ++element) { @@ -799,7 +802,7 @@ Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is info.type = texture_type; info.is_array = is_array; info.is_shadow = false; - const std::optional<Sampler> sampler = GetSampler(instr.sampler, info); + const std::optional<SamplerEntry> sampler = GetSampler(instr.sampler, info); const std::size_t type_coord_count = GetCoordCount(texture_type); const bool lod_enabled = instr.tlds.GetTextureProcessMode() == TextureProcessMode::LL; diff --git a/src/video_core/shader/decode/warp.cpp b/src/video_core/shader/decode/warp.cpp index 11b77f795..37433d783 100644 --- a/src/video_core/shader/decode/warp.cpp +++ b/src/video_core/shader/decode/warp.cpp @@ -27,7 +27,7 @@ OperationCode GetOperationCode(VoteOperation vote_op) { case VoteOperation::Eq: return OperationCode::VoteEqual; default: - UNREACHABLE_MSG("Invalid vote operation={}", static_cast<u64>(vote_op)); + UNREACHABLE_MSG("Invalid vote operation={}", vote_op); return OperationCode::VoteAll; } } diff --git a/src/video_core/shader/expr.h b/src/video_core/shader/expr.h index 4e8264367..cda284c72 100644 --- a/src/video_core/shader/expr.h +++ b/src/video_core/shader/expr.h @@ -76,7 +76,7 @@ public: class ExprPredicate final { public: - explicit ExprPredicate(u32 predicate) : predicate{predicate} {} + explicit ExprPredicate(u32 predicate_) : predicate{predicate_} {} bool operator==(const ExprPredicate& b) const { return predicate == b.predicate; @@ -91,7 +91,7 @@ public: class ExprCondCode final { public: - explicit ExprCondCode(ConditionCode cc) : cc{cc} {} + explicit ExprCondCode(ConditionCode condition_code) : cc{condition_code} {} bool operator==(const ExprCondCode& b) const { return cc == b.cc; @@ -121,7 +121,7 @@ public: class ExprGprEqual final { public: - ExprGprEqual(u32 gpr, u32 value) : gpr{gpr}, value{value} {} + explicit ExprGprEqual(u32 gpr_, u32 value_) : gpr{gpr_}, value{value_} {} bool operator==(const ExprGprEqual& b) const { return gpr == b.gpr && value == b.value; diff --git a/src/video_core/shader/node.h b/src/video_core/shader/node.h index 1b19a0673..c9840b75e 100644 --- a/src/video_core/shader/node.h +++ b/src/video_core/shader/node.h @@ -282,26 +282,27 @@ struct SeparateSamplerNode; using TrackSamplerData = std::variant<BindlessSamplerNode, SeparateSamplerNode, ArraySamplerNode>; using TrackSampler = std::shared_ptr<TrackSamplerData>; -struct Sampler { +struct SamplerEntry { /// Bound samplers constructor - constexpr explicit Sampler(u32 index, u32 offset, Tegra::Shader::TextureType type, - bool is_array, bool is_shadow, bool is_buffer, bool is_indexed) - : index{index}, offset{offset}, type{type}, is_array{is_array}, is_shadow{is_shadow}, - is_buffer{is_buffer}, is_indexed{is_indexed} {} + explicit SamplerEntry(u32 index_, u32 offset_, Tegra::Shader::TextureType type_, bool is_array_, + bool is_shadow_, bool is_buffer_, bool is_indexed_) + : index{index_}, offset{offset_}, type{type_}, is_array{is_array_}, is_shadow{is_shadow_}, + is_buffer{is_buffer_}, is_indexed{is_indexed_} {} /// Separate sampler constructor - constexpr explicit Sampler(u32 index, std::pair<u32, u32> offsets, std::pair<u32, u32> buffers, - Tegra::Shader::TextureType type, bool is_array, bool is_shadow, - bool is_buffer) - : index{index}, offset{offsets.first}, secondary_offset{offsets.second}, - buffer{buffers.first}, secondary_buffer{buffers.second}, type{type}, is_array{is_array}, - is_shadow{is_shadow}, is_buffer{is_buffer}, is_separated{true} {} + explicit SamplerEntry(u32 index_, std::pair<u32, u32> offsets, std::pair<u32, u32> buffers, + Tegra::Shader::TextureType type_, bool is_array_, bool is_shadow_, + bool is_buffer_) + : index{index_}, offset{offsets.first}, secondary_offset{offsets.second}, + buffer{buffers.first}, secondary_buffer{buffers.second}, type{type_}, is_array{is_array_}, + is_shadow{is_shadow_}, is_buffer{is_buffer_}, is_separated{true} {} /// Bindless samplers constructor - constexpr explicit Sampler(u32 index, u32 offset, u32 buffer, Tegra::Shader::TextureType type, - bool is_array, bool is_shadow, bool is_buffer, bool is_indexed) - : index{index}, offset{offset}, buffer{buffer}, type{type}, is_array{is_array}, - is_shadow{is_shadow}, is_buffer{is_buffer}, is_bindless{true}, is_indexed{is_indexed} {} + explicit SamplerEntry(u32 index_, u32 offset_, u32 buffer_, Tegra::Shader::TextureType type_, + bool is_array_, bool is_shadow_, bool is_buffer_, bool is_indexed_) + : index{index_}, offset{offset_}, buffer{buffer_}, type{type_}, is_array{is_array_}, + is_shadow{is_shadow_}, is_buffer{is_buffer_}, is_bindless{true}, is_indexed{is_indexed_} { + } u32 index = 0; ///< Emulated index given for the this sampler. u32 offset = 0; ///< Offset in the const buffer from where the sampler is being read. @@ -338,15 +339,15 @@ struct BindlessSamplerNode { u32 offset; }; -struct Image { +struct ImageEntry { public: /// Bound images constructor - constexpr explicit Image(u32 index, u32 offset, Tegra::Shader::ImageType type) - : index{index}, offset{offset}, type{type} {} + explicit ImageEntry(u32 index_, u32 offset_, Tegra::Shader::ImageType type_) + : index{index_}, offset{offset_}, type{type_} {} /// Bindless samplers constructor - constexpr explicit Image(u32 index, u32 offset, u32 buffer, Tegra::Shader::ImageType type) - : index{index}, offset{offset}, buffer{buffer}, type{type}, is_bindless{true} {} + explicit ImageEntry(u32 index_, u32 offset_, u32 buffer_, Tegra::Shader::ImageType type_) + : index{index_}, offset{offset_}, buffer{buffer_}, type{type_}, is_bindless{true} {} void MarkWrite() { is_written = true; @@ -377,7 +378,7 @@ struct GlobalMemoryBase { u32 cbuf_index = 0; u32 cbuf_offset = 0; - bool operator<(const GlobalMemoryBase& rhs) const { + [[nodiscard]] bool operator<(const GlobalMemoryBase& rhs) const { return std::tie(cbuf_index, cbuf_offset) < std::tie(rhs.cbuf_index, rhs.cbuf_offset); } }; @@ -389,7 +390,7 @@ struct MetaArithmetic { /// Parameters describing a texture sampler struct MetaTexture { - Sampler sampler; + SamplerEntry sampler; Node array; Node depth_compare; std::vector<Node> aoffi; @@ -403,7 +404,7 @@ struct MetaTexture { }; struct MetaImage { - const Image& image; + const ImageEntry& image; std::vector<Node> values; u32 element{}; }; @@ -414,7 +415,7 @@ using Meta = class AmendNode { public: - std::optional<std::size_t> GetAmendIndex() const { + [[nodiscard]] std::optional<std::size_t> GetAmendIndex() const { if (amend_index == amend_null_index) { return std::nullopt; } @@ -437,30 +438,30 @@ private: /// Holds any kind of operation that can be done in the IR class OperationNode final : public AmendNode { public: - explicit OperationNode(OperationCode code) : OperationNode(code, Meta{}) {} + explicit OperationNode(OperationCode code_) : OperationNode(code_, Meta{}) {} - explicit OperationNode(OperationCode code, Meta meta) - : OperationNode(code, std::move(meta), std::vector<Node>{}) {} + explicit OperationNode(OperationCode code_, Meta meta_) + : OperationNode(code_, std::move(meta_), std::vector<Node>{}) {} - explicit OperationNode(OperationCode code, std::vector<Node> operands) - : OperationNode(code, Meta{}, std::move(operands)) {} + explicit OperationNode(OperationCode code_, std::vector<Node> operands_) + : OperationNode(code_, Meta{}, std::move(operands_)) {} - explicit OperationNode(OperationCode code, Meta meta, std::vector<Node> operands) - : code{code}, meta{std::move(meta)}, operands{std::move(operands)} {} + explicit OperationNode(OperationCode code_, Meta meta_, std::vector<Node> operands_) + : code{code_}, meta{std::move(meta_)}, operands{std::move(operands_)} {} template <typename... Args> - explicit OperationNode(OperationCode code, Meta meta, Args&&... operands) - : code{code}, meta{std::move(meta)}, operands{operands...} {} + explicit OperationNode(OperationCode code_, Meta meta_, Args&&... operands_) + : code{code_}, meta{std::move(meta_)}, operands{operands_...} {} - OperationCode GetCode() const { + [[nodiscard]] OperationCode GetCode() const { return code; } - const Meta& GetMeta() const { + [[nodiscard]] const Meta& GetMeta() const { return meta; } - std::size_t GetOperandsCount() const { + [[nodiscard]] std::size_t GetOperandsCount() const { return operands.size(); } @@ -472,7 +473,7 @@ public: return operands; } - const Node& operator[](std::size_t operand_index) const { + [[nodiscard]] const Node& operator[](std::size_t operand_index) const { return operands.at(operand_index); } @@ -485,14 +486,14 @@ private: /// Encloses inside any kind of node that returns a boolean conditionally-executed code class ConditionalNode final : public AmendNode { public: - explicit ConditionalNode(Node condition, std::vector<Node>&& code) - : condition{std::move(condition)}, code{std::move(code)} {} + explicit ConditionalNode(Node condition_, std::vector<Node>&& code_) + : condition{std::move(condition_)}, code{std::move(code_)} {} - const Node& GetCondition() const { + [[nodiscard]] const Node& GetCondition() const { return condition; } - const std::vector<Node>& GetCode() const { + [[nodiscard]] const std::vector<Node>& GetCode() const { return code; } @@ -504,9 +505,9 @@ private: /// A general purpose register class GprNode final { public: - explicit constexpr GprNode(Tegra::Shader::Register index) : index{index} {} + explicit constexpr GprNode(Tegra::Shader::Register index_) : index{index_} {} - u32 GetIndex() const { + [[nodiscard]] constexpr u32 GetIndex() const { return static_cast<u32>(index); } @@ -517,9 +518,9 @@ private: /// A custom variable class CustomVarNode final { public: - explicit constexpr CustomVarNode(u32 index) : index{index} {} + explicit constexpr CustomVarNode(u32 index_) : index{index_} {} - constexpr u32 GetIndex() const { + [[nodiscard]] constexpr u32 GetIndex() const { return index; } @@ -530,9 +531,9 @@ private: /// A 32-bits value that represents an immediate value class ImmediateNode final { public: - explicit constexpr ImmediateNode(u32 value) : value{value} {} + explicit constexpr ImmediateNode(u32 value_) : value{value_} {} - u32 GetValue() const { + [[nodiscard]] constexpr u32 GetValue() const { return value; } @@ -543,9 +544,9 @@ private: /// One of Maxwell's internal flags class InternalFlagNode final { public: - explicit constexpr InternalFlagNode(InternalFlag flag) : flag{flag} {} + explicit constexpr InternalFlagNode(InternalFlag flag_) : flag{flag_} {} - InternalFlag GetFlag() const { + [[nodiscard]] constexpr InternalFlag GetFlag() const { return flag; } @@ -556,14 +557,14 @@ private: /// A predicate register, it can be negated without additional nodes class PredicateNode final { public: - explicit constexpr PredicateNode(Tegra::Shader::Pred index, bool negated) - : index{index}, negated{negated} {} + explicit constexpr PredicateNode(Tegra::Shader::Pred index_, bool negated_) + : index{index_}, negated{negated_} {} - Tegra::Shader::Pred GetIndex() const { + [[nodiscard]] constexpr Tegra::Shader::Pred GetIndex() const { return index; } - bool IsNegated() const { + [[nodiscard]] constexpr bool IsNegated() const { return negated; } @@ -576,30 +577,30 @@ private: class AbufNode final { public: // Initialize for standard attributes (index is explicit). - explicit AbufNode(Tegra::Shader::Attribute::Index index, u32 element, Node buffer = {}) - : buffer{std::move(buffer)}, index{index}, element{element} {} + explicit AbufNode(Tegra::Shader::Attribute::Index index_, u32 element_, Node buffer_ = {}) + : buffer{std::move(buffer_)}, index{index_}, element{element_} {} // Initialize for physical attributes (index is a variable value). - explicit AbufNode(Node physical_address, Node buffer = {}) - : physical_address{std::move(physical_address)}, buffer{std::move(buffer)} {} + explicit AbufNode(Node physical_address_, Node buffer_ = {}) + : physical_address{std::move(physical_address_)}, buffer{std::move(buffer_)} {} - Tegra::Shader::Attribute::Index GetIndex() const { + [[nodiscard]] Tegra::Shader::Attribute::Index GetIndex() const { return index; } - u32 GetElement() const { + [[nodiscard]] u32 GetElement() const { return element; } - const Node& GetBuffer() const { + [[nodiscard]] const Node& GetBuffer() const { return buffer; } - bool IsPhysicalBuffer() const { + [[nodiscard]] bool IsPhysicalBuffer() const { return static_cast<bool>(physical_address); } - const Node& GetPhysicalAddress() const { + [[nodiscard]] const Node& GetPhysicalAddress() const { return physical_address; } @@ -613,9 +614,9 @@ private: /// Patch memory (used to communicate tessellation stages). class PatchNode final { public: - explicit PatchNode(u32 offset) : offset{offset} {} + explicit constexpr PatchNode(u32 offset_) : offset{offset_} {} - u32 GetOffset() const { + [[nodiscard]] constexpr u32 GetOffset() const { return offset; } @@ -626,13 +627,13 @@ private: /// Constant buffer node, usually mapped to uniform buffers in GLSL class CbufNode final { public: - explicit CbufNode(u32 index, Node offset) : index{index}, offset{std::move(offset)} {} + explicit CbufNode(u32 index_, Node offset_) : index{index_}, offset{std::move(offset_)} {} - u32 GetIndex() const { + [[nodiscard]] u32 GetIndex() const { return index; } - const Node& GetOffset() const { + [[nodiscard]] const Node& GetOffset() const { return offset; } @@ -644,9 +645,9 @@ private: /// Local memory node class LmemNode final { public: - explicit LmemNode(Node address) : address{std::move(address)} {} + explicit LmemNode(Node address_) : address{std::move(address_)} {} - const Node& GetAddress() const { + [[nodiscard]] const Node& GetAddress() const { return address; } @@ -657,9 +658,9 @@ private: /// Shared memory node class SmemNode final { public: - explicit SmemNode(Node address) : address{std::move(address)} {} + explicit SmemNode(Node address_) : address{std::move(address_)} {} - const Node& GetAddress() const { + [[nodiscard]] const Node& GetAddress() const { return address; } @@ -670,19 +671,19 @@ private: /// Global memory node class GmemNode final { public: - explicit GmemNode(Node real_address, Node base_address, const GlobalMemoryBase& descriptor) - : real_address{std::move(real_address)}, base_address{std::move(base_address)}, - descriptor{descriptor} {} + explicit GmemNode(Node real_address_, Node base_address_, const GlobalMemoryBase& descriptor_) + : real_address{std::move(real_address_)}, base_address{std::move(base_address_)}, + descriptor{descriptor_} {} - const Node& GetRealAddress() const { + [[nodiscard]] const Node& GetRealAddress() const { return real_address; } - const Node& GetBaseAddress() const { + [[nodiscard]] const Node& GetBaseAddress() const { return base_address; } - const GlobalMemoryBase& GetDescriptor() const { + [[nodiscard]] const GlobalMemoryBase& GetDescriptor() const { return descriptor; } @@ -695,9 +696,9 @@ private: /// Commentary, can be dropped class CommentNode final { public: - explicit CommentNode(std::string text) : text{std::move(text)} {} + explicit CommentNode(std::string text_) : text{std::move(text_)} {} - const std::string& GetText() const { + [[nodiscard]] const std::string& GetText() const { return text; } diff --git a/src/video_core/shader/node_helper.cpp b/src/video_core/shader/node_helper.cpp index 7bf4ff387..6a5b6940d 100644 --- a/src/video_core/shader/node_helper.cpp +++ b/src/video_core/shader/node_helper.cpp @@ -107,7 +107,7 @@ OperationCode SignedToUnsignedCode(OperationCode operation_code, bool is_signed) UNREACHABLE_MSG("Can't apply absolute to an unsigned integer"); return {}; default: - UNREACHABLE_MSG("Unknown signed operation with code={}", static_cast<u32>(operation_code)); + UNREACHABLE_MSG("Unknown signed operation with code={}", operation_code); return {}; } } diff --git a/src/video_core/shader/registry.cpp b/src/video_core/shader/registry.cpp index cdf274e54..148d91fcb 100644 --- a/src/video_core/shader/registry.cpp +++ b/src/video_core/shader/registry.cpp @@ -24,44 +24,45 @@ GraphicsInfo MakeGraphicsInfo(ShaderType shader_stage, ConstBufferEngineInterfac if (shader_stage == ShaderType::Compute) { return {}; } - auto& graphics = static_cast<Tegra::Engines::Maxwell3D&>(engine); - - GraphicsInfo info; - info.tfb_layouts = graphics.regs.tfb_layouts; - info.tfb_varying_locs = graphics.regs.tfb_varying_locs; - info.primitive_topology = graphics.regs.draw.topology; - info.tessellation_primitive = graphics.regs.tess_mode.prim; - info.tessellation_spacing = graphics.regs.tess_mode.spacing; - info.tfb_enabled = graphics.regs.tfb_enabled; - info.tessellation_clockwise = graphics.regs.tess_mode.cw; - return info; + + auto& graphics = dynamic_cast<Tegra::Engines::Maxwell3D&>(engine); + + return { + .tfb_layouts = graphics.regs.tfb_layouts, + .tfb_varying_locs = graphics.regs.tfb_varying_locs, + .primitive_topology = graphics.regs.draw.topology, + .tessellation_primitive = graphics.regs.tess_mode.prim, + .tessellation_spacing = graphics.regs.tess_mode.spacing, + .tfb_enabled = graphics.regs.tfb_enabled != 0, + .tessellation_clockwise = graphics.regs.tess_mode.cw.Value() != 0, + }; } ComputeInfo MakeComputeInfo(ShaderType shader_stage, ConstBufferEngineInterface& engine) { if (shader_stage != ShaderType::Compute) { return {}; } - auto& compute = static_cast<Tegra::Engines::KeplerCompute&>(engine); + + auto& compute = dynamic_cast<Tegra::Engines::KeplerCompute&>(engine); const auto& launch = compute.launch_description; - ComputeInfo info; - info.workgroup_size = {launch.block_dim_x, launch.block_dim_y, launch.block_dim_z}; - info.local_memory_size_in_words = launch.local_pos_alloc; - info.shared_memory_size_in_words = launch.shared_alloc; - return info; + return { + .workgroup_size = {launch.block_dim_x, launch.block_dim_y, launch.block_dim_z}, + .shared_memory_size_in_words = launch.shared_alloc, + .local_memory_size_in_words = launch.local_pos_alloc, + }; } } // Anonymous namespace -Registry::Registry(Tegra::Engines::ShaderType shader_stage, const SerializedRegistryInfo& info) +Registry::Registry(ShaderType shader_stage, const SerializedRegistryInfo& info) : stage{shader_stage}, stored_guest_driver_profile{info.guest_driver_profile}, bound_buffer{info.bound_buffer}, graphics_info{info.graphics}, compute_info{info.compute} {} -Registry::Registry(Tegra::Engines::ShaderType shader_stage, - Tegra::Engines::ConstBufferEngineInterface& engine) - : stage{shader_stage}, engine{&engine}, bound_buffer{engine.GetBoundBuffer()}, - graphics_info{MakeGraphicsInfo(shader_stage, engine)}, compute_info{MakeComputeInfo( - shader_stage, engine)} {} +Registry::Registry(ShaderType shader_stage, ConstBufferEngineInterface& engine_) + : stage{shader_stage}, engine{&engine_}, bound_buffer{engine_.GetBoundBuffer()}, + graphics_info{MakeGraphicsInfo(shader_stage, engine_)}, compute_info{MakeComputeInfo( + shader_stage, engine_)} {} Registry::~Registry() = default; @@ -113,8 +114,7 @@ std::optional<Tegra::Engines::SamplerDescriptor> Registry::ObtainSeparateSampler return value; } -std::optional<Tegra::Engines::SamplerDescriptor> Registry::ObtainBindlessSampler(u32 buffer, - u32 offset) { +std::optional<SamplerDescriptor> Registry::ObtainBindlessSampler(u32 buffer, u32 offset) { const std::pair key = {buffer, offset}; const auto iter = bindless_samplers.find(key); if (iter != bindless_samplers.end()) { diff --git a/src/video_core/shader/registry.h b/src/video_core/shader/registry.h index 231206765..4bebefdde 100644 --- a/src/video_core/shader/registry.h +++ b/src/video_core/shader/registry.h @@ -94,7 +94,7 @@ public: explicit Registry(Tegra::Engines::ShaderType shader_stage, const SerializedRegistryInfo& info); explicit Registry(Tegra::Engines::ShaderType shader_stage, - Tegra::Engines::ConstBufferEngineInterface& engine); + Tegra::Engines::ConstBufferEngineInterface& engine_); ~Registry(); diff --git a/src/video_core/shader/shader_ir.cpp b/src/video_core/shader/shader_ir.cpp index f207bbfbf..caf5ff362 100644 --- a/src/video_core/shader/shader_ir.cpp +++ b/src/video_core/shader/shader_ir.cpp @@ -25,9 +25,10 @@ using Tegra::Shader::PredCondition; using Tegra::Shader::PredOperation; using Tegra::Shader::Register; -ShaderIR::ShaderIR(const ProgramCode& program_code, u32 main_offset, CompilerSettings settings, - Registry& registry) - : program_code{program_code}, main_offset{main_offset}, settings{settings}, registry{registry} { +ShaderIR::ShaderIR(const ProgramCode& program_code_, u32 main_offset_, CompilerSettings settings_, + Registry& registry_) + : program_code{program_code_}, main_offset{main_offset_}, settings{settings_}, registry{ + registry_} { Decode(); PostDecode(); } @@ -170,7 +171,7 @@ Node ShaderIR::ConvertIntegerSize(Node value, Register::Size size, bool is_signe // Default - do nothing return value; default: - UNREACHABLE_MSG("Unimplemented conversion size: {}", static_cast<u32>(size)); + UNREACHABLE_MSG("Unimplemented conversion size: {}", size); return value; } } @@ -335,15 +336,15 @@ OperationCode ShaderIR::GetPredicateCombiner(PredOperation operation) { return operation_table[index]; } -Node ShaderIR::GetConditionCode(Tegra::Shader::ConditionCode cc) const { +Node ShaderIR::GetConditionCode(ConditionCode cc) const { switch (cc) { - case Tegra::Shader::ConditionCode::NEU: + case ConditionCode::NEU: return GetInternalFlag(InternalFlag::Zero, true); - case Tegra::Shader::ConditionCode::FCSM_TR: + case ConditionCode::FCSM_TR: UNIMPLEMENTED_MSG("EXIT.FCSM_TR is not implemented"); return MakeNode<PredicateNode>(Pred::NeverExecute, false); default: - UNIMPLEMENTED_MSG("Unimplemented condition code: {}", static_cast<u32>(cc)); + UNIMPLEMENTED_MSG("Unimplemented condition code: {}", cc); return MakeNode<PredicateNode>(Pred::NeverExecute, false); } } @@ -496,8 +497,8 @@ void ShaderIR::MarkAttributeUsage(Attribute::Index index, u64 element) { } std::size_t ShaderIR::DeclareAmend(Node new_amend) { - const std::size_t id = amend_code.size(); - amend_code.push_back(new_amend); + const auto id = amend_code.size(); + amend_code.push_back(std::move(new_amend)); return id; } diff --git a/src/video_core/shader/shader_ir.h b/src/video_core/shader/shader_ir.h index b450f3b8a..0afa39531 100644 --- a/src/video_core/shader/shader_ir.h +++ b/src/video_core/shader/shader_ir.h @@ -29,8 +29,8 @@ struct ShaderBlock; constexpr u32 MAX_PROGRAM_LENGTH = 0x1000; struct ConstBuffer { - constexpr explicit ConstBuffer(u32 max_offset, bool is_indirect) - : max_offset{max_offset}, is_indirect{is_indirect} {} + constexpr explicit ConstBuffer(u32 max_offset_, bool is_indirect_) + : max_offset{max_offset_}, is_indirect{is_indirect_} {} constexpr ConstBuffer() = default; @@ -66,8 +66,8 @@ struct GlobalMemoryUsage { class ShaderIR final { public: - explicit ShaderIR(const ProgramCode& program_code, u32 main_offset, CompilerSettings settings, - Registry& registry); + explicit ShaderIR(const ProgramCode& program_code_, u32 main_offset_, + CompilerSettings settings_, Registry& registry_); ~ShaderIR(); const std::map<u32, NodeBlock>& GetBasicBlocks() const { @@ -94,11 +94,11 @@ public: return used_cbufs; } - const std::list<Sampler>& GetSamplers() const { + const std::list<SamplerEntry>& GetSamplers() const { return used_samplers; } - const std::list<Image>& GetImages() const { + const std::list<ImageEntry>& GetImages() const { return used_images; } @@ -334,17 +334,17 @@ private: std::optional<Tegra::Engines::SamplerDescriptor> sampler); /// Accesses a texture sampler. - std::optional<Sampler> GetSampler(Tegra::Shader::Sampler sampler, SamplerInfo info); + std::optional<SamplerEntry> GetSampler(Tegra::Shader::Sampler sampler, SamplerInfo info); /// Accesses a texture sampler for a bindless texture. - std::optional<Sampler> GetBindlessSampler(Tegra::Shader::Register reg, SamplerInfo info, - Node& index_var); + std::optional<SamplerEntry> GetBindlessSampler(Tegra::Shader::Register reg, SamplerInfo info, + Node& index_var); /// Accesses an image. - Image& GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type); + ImageEntry& GetImage(Tegra::Shader::Image image, Tegra::Shader::ImageType type); /// Access a bindless image sampler. - Image& GetBindlessImage(Tegra::Shader::Register reg, Tegra::Shader::ImageType type); + ImageEntry& GetBindlessImage(Tegra::Shader::Register reg, Tegra::Shader::ImageType type); /// Recursive Iteration over the OperationNode operands, searching for GprNodes. void SearchOperands(NodeBlock& nb, Node var); @@ -457,8 +457,8 @@ private: std::set<Tegra::Shader::Attribute::Index> used_input_attributes; std::set<Tegra::Shader::Attribute::Index> used_output_attributes; std::map<u32, ConstBuffer> used_cbufs; - std::list<Sampler> used_samplers; - std::list<Image> used_images; + std::list<SamplerEntry> used_samplers; + std::list<ImageEntry> used_images; std::array<bool, Tegra::Engines::Maxwell3D::Regs::NumClipDistances> used_clip_distances{}; std::map<GlobalMemoryBase, GlobalMemoryUsage> used_global_memory; bool uses_layer{}; diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp index 1688267bb..6308aef94 100644 --- a/src/video_core/surface.cpp +++ b/src/video_core/surface.cpp @@ -28,7 +28,7 @@ SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_t case Tegra::Texture::TextureType::Texture2DArray: return SurfaceTarget::Texture2DArray; default: - LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", static_cast<u32>(texture_type)); + LOG_CRITICAL(HW_GPU, "Unimplemented texture_type={}", texture_type); UNREACHABLE(); return SurfaceTarget::Texture2D; } @@ -47,7 +47,7 @@ bool SurfaceTargetIsLayered(SurfaceTarget target) { case SurfaceTarget::TextureCubeArray: return true; default: - LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target)); + LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", target); UNREACHABLE(); return false; } @@ -66,7 +66,7 @@ bool SurfaceTargetIsArray(SurfaceTarget target) { case SurfaceTarget::TextureCubeArray: return true; default: - LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target)); + LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", target); UNREACHABLE(); return false; } @@ -85,7 +85,7 @@ PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) { case Tegra::DepthFormat::D32_FLOAT_S8X24_UINT: return PixelFormat::D32_FLOAT_S8_UINT; default: - UNIMPLEMENTED_MSG("Unimplemented format={}", static_cast<u32>(format)); + UNIMPLEMENTED_MSG("Unimplemented format={}", format); return PixelFormat::S8_UINT_D24_UNORM; } } @@ -183,7 +183,7 @@ PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) case Tegra::RenderTargetFormat::R8_UINT: return PixelFormat::R8_UINT; default: - UNIMPLEMENTED_MSG("Unimplemented format={}", static_cast<int>(format)); + UNIMPLEMENTED_MSG("Unimplemented format={}", format); return PixelFormat::A8B8G8R8_UNORM; } } @@ -197,7 +197,7 @@ PixelFormat PixelFormatFromGPUPixelFormat(Tegra::FramebufferConfig::PixelFormat case Tegra::FramebufferConfig::PixelFormat::B8G8R8A8_UNORM: return PixelFormat::B8G8R8A8_UNORM; default: - UNIMPLEMENTED_MSG("Unimplemented format={}", static_cast<u32>(format)); + UNIMPLEMENTED_MSG("Unimplemented format={}", format); return PixelFormat::A8B8G8R8_UNORM; } } @@ -280,7 +280,7 @@ bool IsPixelFormatSRGB(PixelFormat format) { } std::pair<u32, u32> GetASTCBlockSize(PixelFormat format) { - return {GetDefaultBlockWidth(format), GetDefaultBlockHeight(format)}; + return {DefaultBlockWidth(format), DefaultBlockHeight(format)}; } } // namespace VideoCore::Surface diff --git a/src/video_core/surface.h b/src/video_core/surface.h index cfd12fa61..c40ab89d0 100644 --- a/src/video_core/surface.h +++ b/src/video_core/surface.h @@ -120,7 +120,7 @@ enum class PixelFormat { Max = MaxDepthStencilFormat, Invalid = 255, }; -static constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max); +constexpr std::size_t MaxPixelFormat = static_cast<std::size_t>(PixelFormat::Max); enum class SurfaceType { ColorTexture = 0, @@ -140,117 +140,7 @@ enum class SurfaceTarget { TextureCubeArray, }; -constexpr std::array<u32, MaxPixelFormat> compression_factor_shift_table = {{ - 0, // A8B8G8R8_UNORM - 0, // A8B8G8R8_SNORM - 0, // A8B8G8R8_SINT - 0, // A8B8G8R8_UINT - 0, // R5G6B5_UNORM - 0, // B5G6R5_UNORM - 0, // A1R5G5B5_UNORM - 0, // A2B10G10R10_UNORM - 0, // A2B10G10R10_UINT - 0, // A1B5G5R5_UNORM - 0, // R8_UNORM - 0, // R8_SNORM - 0, // R8_SINT - 0, // R8_UINT - 0, // R16G16B16A16_FLOAT - 0, // R16G16B16A16_UNORM - 0, // R16G16B16A16_SNORM - 0, // R16G16B16A16_SINT - 0, // R16G16B16A16_UINT - 0, // B10G11R11_FLOAT - 0, // R32G32B32A32_UINT - 2, // BC1_RGBA_UNORM - 2, // BC2_UNORM - 2, // BC3_UNORM - 2, // BC4_UNORM - 2, // BC4_SNORM - 2, // BC5_UNORM - 2, // BC5_SNORM - 2, // BC7_UNORM - 2, // BC6H_UFLOAT - 2, // BC6H_SFLOAT - 2, // ASTC_2D_4X4_UNORM - 0, // B8G8R8A8_UNORM - 0, // R32G32B32A32_FLOAT - 0, // R32G32B32A32_SINT - 0, // R32G32_FLOAT - 0, // R32G32_SINT - 0, // R32_FLOAT - 0, // R16_FLOAT - 0, // R16_UNORM - 0, // R16_SNORM - 0, // R16_UINT - 0, // R16_SINT - 0, // R16G16_UNORM - 0, // R16G16_FLOAT - 0, // R16G16_UINT - 0, // R16G16_SINT - 0, // R16G16_SNORM - 0, // R32G32B32_FLOAT - 0, // A8B8G8R8_SRGB - 0, // R8G8_UNORM - 0, // R8G8_SNORM - 0, // R8G8_SINT - 0, // R8G8_UINT - 0, // R32G32_UINT - 0, // R16G16B16X16_FLOAT - 0, // R32_UINT - 0, // R32_SINT - 2, // ASTC_2D_8X8_UNORM - 2, // ASTC_2D_8X5_UNORM - 2, // ASTC_2D_5X4_UNORM - 0, // B8G8R8A8_SRGB - 2, // BC1_RGBA_SRGB - 2, // BC2_SRGB - 2, // BC3_SRGB - 2, // BC7_SRGB - 0, // A4B4G4R4_UNORM - 2, // ASTC_2D_4X4_SRGB - 2, // ASTC_2D_8X8_SRGB - 2, // ASTC_2D_8X5_SRGB - 2, // ASTC_2D_5X4_SRGB - 2, // ASTC_2D_5X5_UNORM - 2, // ASTC_2D_5X5_SRGB - 2, // ASTC_2D_10X8_UNORM - 2, // ASTC_2D_10X8_SRGB - 2, // ASTC_2D_6X6_UNORM - 2, // ASTC_2D_6X6_SRGB - 2, // ASTC_2D_10X10_UNORM - 2, // ASTC_2D_10X10_SRGB - 2, // ASTC_2D_12X12_UNORM - 2, // ASTC_2D_12X12_SRGB - 2, // ASTC_2D_8X6_UNORM - 2, // ASTC_2D_8X6_SRGB - 2, // ASTC_2D_6X5_UNORM - 2, // ASTC_2D_6X5_SRGB - 0, // E5B9G9R9_FLOAT - 0, // D32_FLOAT - 0, // D16_UNORM - 0, // D24_UNORM_S8_UINT - 0, // S8_UINT_D24_UNORM - 0, // D32_FLOAT_S8_UINT -}}; - -/** - * Gets the compression factor for the specified PixelFormat. This applies to just the - * "compressed width" and "compressed height", not the overall compression factor of a - * compressed image. This is used for maintaining proper surface sizes for compressed - * texture formats. - */ -inline constexpr u32 GetCompressionFactorShift(PixelFormat format) { - DEBUG_ASSERT(format != PixelFormat::Invalid); - DEBUG_ASSERT(static_cast<std::size_t>(format) < compression_factor_shift_table.size()); - return compression_factor_shift_table[static_cast<std::size_t>(format)]; -} - -inline constexpr u32 GetCompressionFactor(PixelFormat format) { - return 1U << GetCompressionFactorShift(format); -} - -constexpr std::array<u32, MaxPixelFormat> block_width_table = {{ +constexpr std::array<u32, MaxPixelFormat> BLOCK_WIDTH_TABLE = {{ 1, // A8B8G8R8_UNORM 1, // A8B8G8R8_SNORM 1, // A8B8G8R8_SINT @@ -344,15 +234,12 @@ constexpr std::array<u32, MaxPixelFormat> block_width_table = {{ 1, // D32_FLOAT_S8_UINT }}; -static constexpr u32 GetDefaultBlockWidth(PixelFormat format) { - if (format == PixelFormat::Invalid) - return 0; - - ASSERT(static_cast<std::size_t>(format) < block_width_table.size()); - return block_width_table[static_cast<std::size_t>(format)]; +constexpr u32 DefaultBlockWidth(PixelFormat format) { + ASSERT(static_cast<std::size_t>(format) < BLOCK_WIDTH_TABLE.size()); + return BLOCK_WIDTH_TABLE[static_cast<std::size_t>(format)]; } -constexpr std::array<u32, MaxPixelFormat> block_height_table = {{ +constexpr std::array<u32, MaxPixelFormat> BLOCK_HEIGHT_TABLE = {{ 1, // A8B8G8R8_UNORM 1, // A8B8G8R8_SNORM 1, // A8B8G8R8_SINT @@ -446,15 +333,12 @@ constexpr std::array<u32, MaxPixelFormat> block_height_table = {{ 1, // D32_FLOAT_S8_UINT }}; -static constexpr u32 GetDefaultBlockHeight(PixelFormat format) { - if (format == PixelFormat::Invalid) - return 0; - - ASSERT(static_cast<std::size_t>(format) < block_height_table.size()); - return block_height_table[static_cast<std::size_t>(format)]; +constexpr u32 DefaultBlockHeight(PixelFormat format) { + ASSERT(static_cast<std::size_t>(format) < BLOCK_HEIGHT_TABLE.size()); + return BLOCK_HEIGHT_TABLE[static_cast<std::size_t>(format)]; } -constexpr std::array<u32, MaxPixelFormat> bpp_table = {{ +constexpr std::array<u32, MaxPixelFormat> BITS_PER_BLOCK_TABLE = {{ 32, // A8B8G8R8_UNORM 32, // A8B8G8R8_SNORM 32, // A8B8G8R8_SINT @@ -548,20 +432,14 @@ constexpr std::array<u32, MaxPixelFormat> bpp_table = {{ 64, // D32_FLOAT_S8_UINT }}; -static constexpr u32 GetFormatBpp(PixelFormat format) { - if (format == PixelFormat::Invalid) - return 0; - - ASSERT(static_cast<std::size_t>(format) < bpp_table.size()); - return bpp_table[static_cast<std::size_t>(format)]; +constexpr u32 BitsPerBlock(PixelFormat format) { + ASSERT(static_cast<std::size_t>(format) < BITS_PER_BLOCK_TABLE.size()); + return BITS_PER_BLOCK_TABLE[static_cast<std::size_t>(format)]; } /// Returns the sizer in bytes of the specified pixel format -static constexpr u32 GetBytesPerPixel(PixelFormat pixel_format) { - if (pixel_format == PixelFormat::Invalid) { - return 0; - } - return GetFormatBpp(pixel_format) / CHAR_BIT; +constexpr u32 BytesPerBlock(PixelFormat pixel_format) { + return BitsPerBlock(pixel_format) / CHAR_BIT; } SurfaceTarget SurfaceTargetFromTextureType(Tegra::Texture::TextureType texture_type); diff --git a/src/video_core/texture_cache/accelerated_swizzle.cpp b/src/video_core/texture_cache/accelerated_swizzle.cpp new file mode 100644 index 000000000..a4fc1184b --- /dev/null +++ b/src/video_core/texture_cache/accelerated_swizzle.cpp @@ -0,0 +1,70 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <array> +#include <bit> + +#include "common/alignment.h" +#include "common/common_types.h" +#include "common/div_ceil.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/accelerated_swizzle.h" +#include "video_core/texture_cache/util.h" +#include "video_core/textures/decoders.h" + +namespace VideoCommon::Accelerated { + +using Tegra::Texture::GOB_SIZE_SHIFT; +using Tegra::Texture::GOB_SIZE_X; +using Tegra::Texture::GOB_SIZE_X_SHIFT; +using Tegra::Texture::GOB_SIZE_Y_SHIFT; +using VideoCore::Surface::BytesPerBlock; + +BlockLinearSwizzle2DParams MakeBlockLinearSwizzle2DParams(const SwizzleParameters& swizzle, + const ImageInfo& info) { + const Extent3D block = swizzle.block; + const Extent3D num_tiles = swizzle.num_tiles; + const u32 bytes_per_block = BytesPerBlock(info.format); + const u32 stride_alignment = CalculateLevelStrideAlignment(info, swizzle.level); + const u32 stride = Common::AlignBits(num_tiles.width, stride_alignment) * bytes_per_block; + const u32 gobs_in_x = Common::DivCeilLog2(stride, GOB_SIZE_X_SHIFT); + return BlockLinearSwizzle2DParams{ + .origin{0, 0, 0}, + .destination{0, 0, 0}, + .bytes_per_block_log2 = static_cast<u32>(std::countr_zero(bytes_per_block)), + .layer_stride = info.layer_stride, + .block_size = gobs_in_x << (GOB_SIZE_SHIFT + block.height + block.depth), + .x_shift = GOB_SIZE_SHIFT + block.height + block.depth, + .block_height = block.height, + .block_height_mask = (1U << block.height) - 1, + }; +} + +BlockLinearSwizzle3DParams MakeBlockLinearSwizzle3DParams(const SwizzleParameters& swizzle, + const ImageInfo& info) { + const Extent3D block = swizzle.block; + const Extent3D num_tiles = swizzle.num_tiles; + const u32 bytes_per_block = BytesPerBlock(info.format); + const u32 stride_alignment = CalculateLevelStrideAlignment(info, swizzle.level); + const u32 stride = Common::AlignBits(num_tiles.width, stride_alignment) * bytes_per_block; + + const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) >> GOB_SIZE_X_SHIFT; + const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block.height + block.depth); + const u32 slice_size = + Common::DivCeilLog2(num_tiles.height, block.height + GOB_SIZE_Y_SHIFT) * block_size; + return BlockLinearSwizzle3DParams{ + .origin{0, 0, 0}, + .destination{0, 0, 0}, + .bytes_per_block_log2 = static_cast<u32>(std::countr_zero(bytes_per_block)), + .slice_size = slice_size, + .block_size = block_size, + .x_shift = GOB_SIZE_SHIFT + block.height + block.depth, + .block_height = block.height, + .block_height_mask = (1U << block.height) - 1, + .block_depth = block.depth, + .block_depth_mask = (1U << block.depth) - 1, + }; +} + +} // namespace VideoCommon::Accelerated
\ No newline at end of file diff --git a/src/video_core/texture_cache/accelerated_swizzle.h b/src/video_core/texture_cache/accelerated_swizzle.h new file mode 100644 index 000000000..6ec5c78c4 --- /dev/null +++ b/src/video_core/texture_cache/accelerated_swizzle.h @@ -0,0 +1,45 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> + +#include "common/common_types.h" +#include "video_core/texture_cache/image_info.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon::Accelerated { + +struct BlockLinearSwizzle2DParams { + std::array<u32, 3> origin; + std::array<s32, 3> destination; + u32 bytes_per_block_log2; + u32 layer_stride; + u32 block_size; + u32 x_shift; + u32 block_height; + u32 block_height_mask; +}; + +struct BlockLinearSwizzle3DParams { + std::array<u32, 3> origin; + std::array<s32, 3> destination; + u32 bytes_per_block_log2; + u32 slice_size; + u32 block_size; + u32 x_shift; + u32 block_height; + u32 block_height_mask; + u32 block_depth; + u32 block_depth_mask; +}; + +[[nodiscard]] BlockLinearSwizzle2DParams MakeBlockLinearSwizzle2DParams( + const SwizzleParameters& swizzle, const ImageInfo& info); + +[[nodiscard]] BlockLinearSwizzle3DParams MakeBlockLinearSwizzle3DParams( + const SwizzleParameters& swizzle, const ImageInfo& info); + +} // namespace VideoCommon::Accelerated diff --git a/src/video_core/texture_cache/copy_params.h b/src/video_core/texture_cache/copy_params.h deleted file mode 100644 index 9c21a0649..000000000 --- a/src/video_core/texture_cache/copy_params.h +++ /dev/null @@ -1,36 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "common/common_types.h" - -namespace VideoCommon { - -struct CopyParams { - constexpr CopyParams(u32 source_x, u32 source_y, u32 source_z, u32 dest_x, u32 dest_y, - u32 dest_z, u32 source_level, u32 dest_level, u32 width, u32 height, - u32 depth) - : source_x{source_x}, source_y{source_y}, source_z{source_z}, dest_x{dest_x}, - dest_y{dest_y}, dest_z{dest_z}, source_level{source_level}, - dest_level{dest_level}, width{width}, height{height}, depth{depth} {} - - constexpr CopyParams(u32 width, u32 height, u32 depth, u32 level) - : source_x{}, source_y{}, source_z{}, dest_x{}, dest_y{}, dest_z{}, source_level{level}, - dest_level{level}, width{width}, height{height}, depth{depth} {} - - u32 source_x; - u32 source_y; - u32 source_z; - u32 dest_x; - u32 dest_y; - u32 dest_z; - u32 source_level; - u32 dest_level; - u32 width; - u32 height; - u32 depth; -}; - -} // namespace VideoCommon diff --git a/src/video_core/texture_cache/decode_bc4.cpp b/src/video_core/texture_cache/decode_bc4.cpp new file mode 100644 index 000000000..017327975 --- /dev/null +++ b/src/video_core/texture_cache/decode_bc4.cpp @@ -0,0 +1,97 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <array> +#include <span> + +#include "common/assert.h" +#include "common/common_types.h" +#include "video_core/texture_cache/decode_bc4.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon { + +// https://www.khronos.org/registry/OpenGL/extensions/ARB/ARB_texture_compression_rgtc.txt +[[nodiscard]] constexpr u32 DecompressBlock(u64 bits, u32 x, u32 y) { + const u32 code_offset = 16 + 3 * (4 * y + x); + const u32 code = (bits >> code_offset) & 7; + const u32 red0 = (bits >> 0) & 0xff; + const u32 red1 = (bits >> 8) & 0xff; + if (red0 > red1) { + switch (code) { + case 0: + return red0; + case 1: + return red1; + case 2: + return (6 * red0 + 1 * red1) / 7; + case 3: + return (5 * red0 + 2 * red1) / 7; + case 4: + return (4 * red0 + 3 * red1) / 7; + case 5: + return (3 * red0 + 4 * red1) / 7; + case 6: + return (2 * red0 + 5 * red1) / 7; + case 7: + return (1 * red0 + 6 * red1) / 7; + } + } else { + switch (code) { + case 0: + return red0; + case 1: + return red1; + case 2: + return (4 * red0 + 1 * red1) / 5; + case 3: + return (3 * red0 + 2 * red1) / 5; + case 4: + return (2 * red0 + 3 * red1) / 5; + case 5: + return (1 * red0 + 4 * red1) / 5; + case 6: + return 0; + case 7: + return 0xff; + } + } + return 0; +} + +void DecompressBC4(std::span<const u8> input, Extent3D extent, std::span<u8> output) { + UNIMPLEMENTED_IF_MSG(extent.width % 4 != 0, "Unaligned width={}", extent.width); + UNIMPLEMENTED_IF_MSG(extent.height % 4 != 0, "Unaligned height={}", extent.height); + static constexpr u32 BLOCK_SIZE = 4; + size_t input_offset = 0; + for (u32 slice = 0; slice < extent.depth; ++slice) { + for (u32 block_y = 0; block_y < extent.height / 4; ++block_y) { + for (u32 block_x = 0; block_x < extent.width / 4; ++block_x) { + u64 bits; + std::memcpy(&bits, &input[input_offset], sizeof(bits)); + input_offset += sizeof(bits); + + for (u32 y = 0; y < BLOCK_SIZE; ++y) { + for (u32 x = 0; x < BLOCK_SIZE; ++x) { + const u32 linear_z = slice; + const u32 linear_y = block_y * BLOCK_SIZE + y; + const u32 linear_x = block_x * BLOCK_SIZE + x; + const u32 offset_z = linear_z * extent.width * extent.height; + const u32 offset_y = linear_y * extent.width; + const u32 offset_x = linear_x; + const u32 output_offset = (offset_z + offset_y + offset_x) * 4ULL; + const u32 color = DecompressBlock(bits, x, y); + output[output_offset + 0] = static_cast<u8>(color); + output[output_offset + 1] = 0; + output[output_offset + 2] = 0; + output[output_offset + 3] = 0xff; + } + } + } + } + } +} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/decode_bc4.h b/src/video_core/texture_cache/decode_bc4.h new file mode 100644 index 000000000..63fb23508 --- /dev/null +++ b/src/video_core/texture_cache/decode_bc4.h @@ -0,0 +1,16 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <span> + +#include "common/common_types.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon { + +void DecompressBC4(std::span<const u8> data, Extent3D extent, std::span<u8> output); + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/descriptor_table.h b/src/video_core/texture_cache/descriptor_table.h new file mode 100644 index 000000000..3a03b786f --- /dev/null +++ b/src/video_core/texture_cache/descriptor_table.h @@ -0,0 +1,82 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <algorithm> +#include <vector> + +#include "common/common_types.h" +#include "common/div_ceil.h" +#include "common/logging/log.h" +#include "video_core/memory_manager.h" +#include "video_core/rasterizer_interface.h" + +namespace VideoCommon { + +template <typename Descriptor> +class DescriptorTable { +public: + explicit DescriptorTable(Tegra::MemoryManager& gpu_memory_) : gpu_memory{gpu_memory_} {} + + [[nodiscard]] bool Synchornize(GPUVAddr gpu_addr, u32 limit) { + [[likely]] if (current_gpu_addr == gpu_addr && current_limit == limit) { + return false; + } + Refresh(gpu_addr, limit); + return true; + } + + void Invalidate() noexcept { + std::ranges::fill(read_descriptors, 0); + } + + [[nodiscard]] std::pair<Descriptor, bool> Read(u32 index) { + DEBUG_ASSERT(index <= current_limit); + const GPUVAddr gpu_addr = current_gpu_addr + index * sizeof(Descriptor); + std::pair<Descriptor, bool> result; + gpu_memory.ReadBlockUnsafe(gpu_addr, &result.first, sizeof(Descriptor)); + if (IsDescriptorRead(index)) { + result.second = result.first != descriptors[index]; + } else { + MarkDescriptorAsRead(index); + result.second = true; + } + if (result.second) { + descriptors[index] = result.first; + } + return result; + } + + [[nodiscard]] u32 Limit() const noexcept { + return current_limit; + } + +private: + void Refresh(GPUVAddr gpu_addr, u32 limit) { + current_gpu_addr = gpu_addr; + current_limit = limit; + + const size_t num_descriptors = static_cast<size_t>(limit) + 1; + read_descriptors.clear(); + read_descriptors.resize(Common::DivCeil(num_descriptors, 64U), 0); + descriptors.resize(num_descriptors); + } + + void MarkDescriptorAsRead(u32 index) noexcept { + read_descriptors[index / 64] |= 1ULL << (index % 64); + } + + [[nodiscard]] bool IsDescriptorRead(u32 index) const noexcept { + return (read_descriptors[index / 64] & (1ULL << (index % 64))) != 0; + } + + Tegra::MemoryManager& gpu_memory; + GPUVAddr current_gpu_addr{}; + u32 current_limit{}; + std::vector<u64> read_descriptors; + std::vector<Descriptor> descriptors; +}; + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/format_lookup_table.cpp b/src/video_core/texture_cache/format_lookup_table.cpp index 7d5a75648..ddfb726fe 100644 --- a/src/video_core/texture_cache/format_lookup_table.cpp +++ b/src/video_core/texture_cache/format_lookup_table.cpp @@ -2,7 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <array> #include "common/common_types.h" #include "common/logging/log.h" #include "video_core/texture_cache/format_lookup_table.h" @@ -20,198 +19,207 @@ constexpr auto UNORM = ComponentType::UNORM; constexpr auto SINT = ComponentType::SINT; constexpr auto UINT = ComponentType::UINT; constexpr auto FLOAT = ComponentType::FLOAT; -constexpr bool C = false; // Normal color -constexpr bool S = true; // Srgb - -struct Table { - constexpr Table(TextureFormat texture_format, bool is_srgb, ComponentType red_component, - ComponentType green_component, ComponentType blue_component, - ComponentType alpha_component, PixelFormat pixel_format) - : texture_format{texture_format}, pixel_format{pixel_format}, red_component{red_component}, - green_component{green_component}, blue_component{blue_component}, - alpha_component{alpha_component}, is_srgb{is_srgb} {} - - TextureFormat texture_format; - PixelFormat pixel_format; - ComponentType red_component; - ComponentType green_component; - ComponentType blue_component; - ComponentType alpha_component; - bool is_srgb; -}; -constexpr std::array<Table, 86> DefinitionTable = {{ - {TextureFormat::A8R8G8B8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A8B8G8R8_UNORM}, - {TextureFormat::A8R8G8B8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::A8B8G8R8_SNORM}, - {TextureFormat::A8R8G8B8, C, UINT, UINT, UINT, UINT, PixelFormat::A8B8G8R8_UINT}, - {TextureFormat::A8R8G8B8, C, SINT, SINT, SINT, SINT, PixelFormat::A8B8G8R8_SINT}, - {TextureFormat::A8R8G8B8, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::A8B8G8R8_SRGB}, - - {TextureFormat::B5G6R5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::B5G6R5_UNORM}, - - {TextureFormat::A2B10G10R10, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A2B10G10R10_UNORM}, - {TextureFormat::A2B10G10R10, C, UINT, UINT, UINT, UINT, PixelFormat::A2B10G10R10_UINT}, - - {TextureFormat::A1B5G5R5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A1B5G5R5_UNORM}, - - {TextureFormat::A4B4G4R4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::A4B4G4R4_UNORM}, - - {TextureFormat::R8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R8_UNORM}, - {TextureFormat::R8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R8_SNORM}, - {TextureFormat::R8, C, UINT, UINT, UINT, UINT, PixelFormat::R8_UINT}, - {TextureFormat::R8, C, SINT, SINT, SINT, SINT, PixelFormat::R8_SINT}, - - {TextureFormat::R8G8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R8G8_UNORM}, - {TextureFormat::R8G8, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R8G8_SNORM}, - {TextureFormat::R8G8, C, UINT, UINT, UINT, UINT, PixelFormat::R8G8_UINT}, - {TextureFormat::R8G8, C, SINT, SINT, SINT, SINT, PixelFormat::R8G8_SINT}, - - {TextureFormat::R16G16B16A16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R16G16B16A16_SNORM}, - {TextureFormat::R16G16B16A16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R16G16B16A16_UNORM}, - {TextureFormat::R16G16B16A16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R16G16B16A16_FLOAT}, - {TextureFormat::R16G16B16A16, C, UINT, UINT, UINT, UINT, PixelFormat::R16G16B16A16_UINT}, - {TextureFormat::R16G16B16A16, C, SINT, SINT, SINT, SINT, PixelFormat::R16G16B16A16_SINT}, - - {TextureFormat::R16G16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R16G16_FLOAT}, - {TextureFormat::R16G16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R16G16_UNORM}, - {TextureFormat::R16G16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R16G16_SNORM}, - {TextureFormat::R16G16, C, UINT, UINT, UINT, UINT, PixelFormat::R16G16_UINT}, - {TextureFormat::R16G16, C, SINT, SINT, SINT, SINT, PixelFormat::R16G16_SINT}, - - {TextureFormat::R16, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R16_FLOAT}, - {TextureFormat::R16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::R16_UNORM}, - {TextureFormat::R16, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::R16_SNORM}, - {TextureFormat::R16, C, UINT, UINT, UINT, UINT, PixelFormat::R16_UINT}, - {TextureFormat::R16, C, SINT, SINT, SINT, SINT, PixelFormat::R16_SINT}, - - {TextureFormat::B10G11R11, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::B10G11R11_FLOAT}, - - {TextureFormat::R32G32B32A32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32G32B32A32_FLOAT}, - {TextureFormat::R32G32B32A32, C, UINT, UINT, UINT, UINT, PixelFormat::R32G32B32A32_UINT}, - {TextureFormat::R32G32B32A32, C, SINT, SINT, SINT, SINT, PixelFormat::R32G32B32A32_SINT}, - - {TextureFormat::R32G32B32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32G32B32_FLOAT}, - - {TextureFormat::R32G32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32G32_FLOAT}, - {TextureFormat::R32G32, C, UINT, UINT, UINT, UINT, PixelFormat::R32G32_UINT}, - {TextureFormat::R32G32, C, SINT, SINT, SINT, SINT, PixelFormat::R32G32_SINT}, - - {TextureFormat::R32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::R32_FLOAT}, - {TextureFormat::R32, C, UINT, UINT, UINT, UINT, PixelFormat::R32_UINT}, - {TextureFormat::R32, C, SINT, SINT, SINT, SINT, PixelFormat::R32_SINT}, - - {TextureFormat::E5B9G9R9, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::E5B9G9R9_FLOAT}, - - {TextureFormat::D32, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::D32_FLOAT}, - {TextureFormat::D16, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::D16_UNORM}, - {TextureFormat::S8D24, C, UINT, UNORM, UNORM, UNORM, PixelFormat::S8_UINT_D24_UNORM}, - {TextureFormat::R8G24, C, UINT, UNORM, UNORM, UNORM, PixelFormat::S8_UINT_D24_UNORM}, - {TextureFormat::D32S8, C, FLOAT, UINT, UNORM, UNORM, PixelFormat::D32_FLOAT_S8_UINT}, - - {TextureFormat::BC1_RGBA, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC1_RGBA_UNORM}, - {TextureFormat::BC1_RGBA, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC1_RGBA_SRGB}, - - {TextureFormat::BC2, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC2_UNORM}, - {TextureFormat::BC2, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC2_SRGB}, - - {TextureFormat::BC3, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC3_UNORM}, - {TextureFormat::BC3, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC3_SRGB}, - - {TextureFormat::BC4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC4_UNORM}, - {TextureFormat::BC4, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::BC4_SNORM}, - - {TextureFormat::BC5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC5_UNORM}, - {TextureFormat::BC5, C, SNORM, SNORM, SNORM, SNORM, PixelFormat::BC5_SNORM}, - - {TextureFormat::BC7, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC7_UNORM}, - {TextureFormat::BC7, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::BC7_SRGB}, - - {TextureFormat::BC6H_SFLOAT, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::BC6H_SFLOAT}, - {TextureFormat::BC6H_UFLOAT, C, FLOAT, FLOAT, FLOAT, FLOAT, PixelFormat::BC6H_UFLOAT}, - - {TextureFormat::ASTC_2D_4X4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_4X4_UNORM}, - {TextureFormat::ASTC_2D_4X4, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_4X4_SRGB}, - - {TextureFormat::ASTC_2D_5X4, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X4_UNORM}, - {TextureFormat::ASTC_2D_5X4, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X4_SRGB}, - - {TextureFormat::ASTC_2D_5X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X5_UNORM}, - {TextureFormat::ASTC_2D_5X5, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_5X5_SRGB}, - - {TextureFormat::ASTC_2D_8X8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X8_UNORM}, - {TextureFormat::ASTC_2D_8X8, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X8_SRGB}, - - {TextureFormat::ASTC_2D_8X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X5_UNORM}, - {TextureFormat::ASTC_2D_8X5, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X5_SRGB}, - - {TextureFormat::ASTC_2D_10X8, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X8_UNORM}, - {TextureFormat::ASTC_2D_10X8, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X8_SRGB}, - - {TextureFormat::ASTC_2D_6X6, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X6_UNORM}, - {TextureFormat::ASTC_2D_6X6, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X6_SRGB}, - - {TextureFormat::ASTC_2D_10X10, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X10_UNORM}, - {TextureFormat::ASTC_2D_10X10, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_10X10_SRGB}, - - {TextureFormat::ASTC_2D_12X12, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_12X12_UNORM}, - {TextureFormat::ASTC_2D_12X12, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_12X12_SRGB}, - - {TextureFormat::ASTC_2D_8X6, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X6_UNORM}, - {TextureFormat::ASTC_2D_8X6, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_8X6_SRGB}, +constexpr bool LINEAR = false; +constexpr bool SRGB = true; + +constexpr u32 Hash(TextureFormat format, ComponentType red_component, ComponentType green_component, + ComponentType blue_component, ComponentType alpha_component, bool is_srgb) { + u32 hash = is_srgb ? 1 : 0; + hash |= static_cast<u32>(red_component) << 1; + hash |= static_cast<u32>(green_component) << 4; + hash |= static_cast<u32>(blue_component) << 7; + hash |= static_cast<u32>(alpha_component) << 10; + hash |= static_cast<u32>(format) << 13; + return hash; +} - {TextureFormat::ASTC_2D_6X5, C, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X5_UNORM}, - {TextureFormat::ASTC_2D_6X5, S, UNORM, UNORM, UNORM, UNORM, PixelFormat::ASTC_2D_6X5_SRGB}, -}}; +constexpr u32 Hash(TextureFormat format, ComponentType component, bool is_srgb = LINEAR) { + return Hash(format, component, component, component, component, is_srgb); +} } // Anonymous namespace -FormatLookupTable::FormatLookupTable() { - table.fill(static_cast<u8>(PixelFormat::Invalid)); - - for (const auto& entry : DefinitionTable) { - table[CalculateIndex(entry.texture_format, entry.is_srgb != 0, entry.red_component, - entry.green_component, entry.blue_component, entry.alpha_component)] = - static_cast<u8>(entry.pixel_format); - } -} - -PixelFormat FormatLookupTable::GetPixelFormat(TextureFormat format, bool is_srgb, - ComponentType red_component, - ComponentType green_component, - ComponentType blue_component, - ComponentType alpha_component) const noexcept { - const auto pixel_format = static_cast<PixelFormat>(table[CalculateIndex( - format, is_srgb, red_component, green_component, blue_component, alpha_component)]); - // [[likely]] - if (pixel_format != PixelFormat::Invalid) { - return pixel_format; +PixelFormat PixelFormatFromTextureInfo(TextureFormat format, ComponentType red, ComponentType green, + ComponentType blue, ComponentType alpha, + bool is_srgb) noexcept { + switch (Hash(format, red, green, blue, alpha, is_srgb)) { + case Hash(TextureFormat::A8R8G8B8, UNORM): + return PixelFormat::A8B8G8R8_UNORM; + case Hash(TextureFormat::A8R8G8B8, SNORM): + return PixelFormat::A8B8G8R8_SNORM; + case Hash(TextureFormat::A8R8G8B8, UINT): + return PixelFormat::A8B8G8R8_UINT; + case Hash(TextureFormat::A8R8G8B8, SINT): + return PixelFormat::A8B8G8R8_SINT; + case Hash(TextureFormat::A8R8G8B8, UNORM, SRGB): + return PixelFormat::A8B8G8R8_SRGB; + case Hash(TextureFormat::B5G6R5, UNORM): + return PixelFormat::B5G6R5_UNORM; + case Hash(TextureFormat::A2B10G10R10, UNORM): + return PixelFormat::A2B10G10R10_UNORM; + case Hash(TextureFormat::A2B10G10R10, UINT): + return PixelFormat::A2B10G10R10_UINT; + case Hash(TextureFormat::A1B5G5R5, UNORM): + return PixelFormat::A1B5G5R5_UNORM; + case Hash(TextureFormat::A4B4G4R4, UNORM): + return PixelFormat::A4B4G4R4_UNORM; + case Hash(TextureFormat::R8, UNORM): + return PixelFormat::R8_UNORM; + case Hash(TextureFormat::R8, SNORM): + return PixelFormat::R8_SNORM; + case Hash(TextureFormat::R8, UINT): + return PixelFormat::R8_UINT; + case Hash(TextureFormat::R8, SINT): + return PixelFormat::R8_SINT; + case Hash(TextureFormat::R8G8, UNORM): + return PixelFormat::R8G8_UNORM; + case Hash(TextureFormat::R8G8, SNORM): + return PixelFormat::R8G8_SNORM; + case Hash(TextureFormat::R8G8, UINT): + return PixelFormat::R8G8_UINT; + case Hash(TextureFormat::R8G8, SINT): + return PixelFormat::R8G8_SINT; + case Hash(TextureFormat::R16G16B16A16, FLOAT): + return PixelFormat::R16G16B16A16_FLOAT; + case Hash(TextureFormat::R16G16B16A16, UNORM): + return PixelFormat::R16G16B16A16_UNORM; + case Hash(TextureFormat::R16G16B16A16, SNORM): + return PixelFormat::R16G16B16A16_SNORM; + case Hash(TextureFormat::R16G16B16A16, UINT): + return PixelFormat::R16G16B16A16_UINT; + case Hash(TextureFormat::R16G16B16A16, SINT): + return PixelFormat::R16G16B16A16_SINT; + case Hash(TextureFormat::R16G16, FLOAT): + return PixelFormat::R16G16_FLOAT; + case Hash(TextureFormat::R16G16, UNORM): + return PixelFormat::R16G16_UNORM; + case Hash(TextureFormat::R16G16, SNORM): + return PixelFormat::R16G16_SNORM; + case Hash(TextureFormat::R16G16, UINT): + return PixelFormat::R16G16_UINT; + case Hash(TextureFormat::R16G16, SINT): + return PixelFormat::R16G16_SINT; + case Hash(TextureFormat::R16, FLOAT): + return PixelFormat::R16_FLOAT; + case Hash(TextureFormat::R16, UNORM): + return PixelFormat::R16_UNORM; + case Hash(TextureFormat::R16, SNORM): + return PixelFormat::R16_SNORM; + case Hash(TextureFormat::R16, UINT): + return PixelFormat::R16_UINT; + case Hash(TextureFormat::R16, SINT): + return PixelFormat::R16_SINT; + case Hash(TextureFormat::B10G11R11, FLOAT): + return PixelFormat::B10G11R11_FLOAT; + case Hash(TextureFormat::R32G32B32A32, FLOAT): + return PixelFormat::R32G32B32A32_FLOAT; + case Hash(TextureFormat::R32G32B32A32, UINT): + return PixelFormat::R32G32B32A32_UINT; + case Hash(TextureFormat::R32G32B32A32, SINT): + return PixelFormat::R32G32B32A32_SINT; + case Hash(TextureFormat::R32G32B32, FLOAT): + return PixelFormat::R32G32B32_FLOAT; + case Hash(TextureFormat::R32G32, FLOAT): + return PixelFormat::R32G32_FLOAT; + case Hash(TextureFormat::R32G32, UINT): + return PixelFormat::R32G32_UINT; + case Hash(TextureFormat::R32G32, SINT): + return PixelFormat::R32G32_SINT; + case Hash(TextureFormat::R32, FLOAT): + return PixelFormat::R32_FLOAT; + case Hash(TextureFormat::R32, UINT): + return PixelFormat::R32_UINT; + case Hash(TextureFormat::R32, SINT): + return PixelFormat::R32_SINT; + case Hash(TextureFormat::E5B9G9R9, FLOAT): + return PixelFormat::E5B9G9R9_FLOAT; + case Hash(TextureFormat::D32, FLOAT): + return PixelFormat::D32_FLOAT; + case Hash(TextureFormat::D16, UNORM): + return PixelFormat::D16_UNORM; + case Hash(TextureFormat::S8D24, UINT, UNORM, UNORM, UNORM, LINEAR): + return PixelFormat::S8_UINT_D24_UNORM; + case Hash(TextureFormat::R8G24, UINT, UNORM, UNORM, UNORM, LINEAR): + return PixelFormat::S8_UINT_D24_UNORM; + case Hash(TextureFormat::D32S8, FLOAT, UINT, UNORM, UNORM, LINEAR): + return PixelFormat::D32_FLOAT_S8_UINT; + case Hash(TextureFormat::BC1_RGBA, UNORM, LINEAR): + return PixelFormat::BC1_RGBA_UNORM; + case Hash(TextureFormat::BC1_RGBA, UNORM, SRGB): + return PixelFormat::BC1_RGBA_SRGB; + case Hash(TextureFormat::BC2, UNORM, LINEAR): + return PixelFormat::BC2_UNORM; + case Hash(TextureFormat::BC2, UNORM, SRGB): + return PixelFormat::BC2_SRGB; + case Hash(TextureFormat::BC3, UNORM, LINEAR): + return PixelFormat::BC3_UNORM; + case Hash(TextureFormat::BC3, UNORM, SRGB): + return PixelFormat::BC3_SRGB; + case Hash(TextureFormat::BC4, UNORM): + return PixelFormat::BC4_UNORM; + case Hash(TextureFormat::BC4, SNORM): + return PixelFormat::BC4_SNORM; + case Hash(TextureFormat::BC5, UNORM): + return PixelFormat::BC5_UNORM; + case Hash(TextureFormat::BC5, SNORM): + return PixelFormat::BC5_SNORM; + case Hash(TextureFormat::BC7, UNORM, LINEAR): + return PixelFormat::BC7_UNORM; + case Hash(TextureFormat::BC7, UNORM, SRGB): + return PixelFormat::BC7_SRGB; + case Hash(TextureFormat::BC6H_SFLOAT, FLOAT): + return PixelFormat::BC6H_SFLOAT; + case Hash(TextureFormat::BC6H_UFLOAT, FLOAT): + return PixelFormat::BC6H_UFLOAT; + case Hash(TextureFormat::ASTC_2D_4X4, UNORM, LINEAR): + return PixelFormat::ASTC_2D_4X4_UNORM; + case Hash(TextureFormat::ASTC_2D_4X4, UNORM, SRGB): + return PixelFormat::ASTC_2D_4X4_SRGB; + case Hash(TextureFormat::ASTC_2D_5X4, UNORM, LINEAR): + return PixelFormat::ASTC_2D_5X4_UNORM; + case Hash(TextureFormat::ASTC_2D_5X4, UNORM, SRGB): + return PixelFormat::ASTC_2D_5X4_SRGB; + case Hash(TextureFormat::ASTC_2D_5X5, UNORM, LINEAR): + return PixelFormat::ASTC_2D_5X5_UNORM; + case Hash(TextureFormat::ASTC_2D_5X5, UNORM, SRGB): + return PixelFormat::ASTC_2D_5X5_SRGB; + case Hash(TextureFormat::ASTC_2D_8X8, UNORM, LINEAR): + return PixelFormat::ASTC_2D_8X8_UNORM; + case Hash(TextureFormat::ASTC_2D_8X8, UNORM, SRGB): + return PixelFormat::ASTC_2D_8X8_SRGB; + case Hash(TextureFormat::ASTC_2D_8X5, UNORM, LINEAR): + return PixelFormat::ASTC_2D_8X5_UNORM; + case Hash(TextureFormat::ASTC_2D_8X5, UNORM, SRGB): + return PixelFormat::ASTC_2D_8X5_SRGB; + case Hash(TextureFormat::ASTC_2D_10X8, UNORM, LINEAR): + return PixelFormat::ASTC_2D_10X8_UNORM; + case Hash(TextureFormat::ASTC_2D_10X8, UNORM, SRGB): + return PixelFormat::ASTC_2D_10X8_SRGB; + case Hash(TextureFormat::ASTC_2D_6X6, UNORM, LINEAR): + return PixelFormat::ASTC_2D_6X6_UNORM; + case Hash(TextureFormat::ASTC_2D_6X6, UNORM, SRGB): + return PixelFormat::ASTC_2D_6X6_SRGB; + case Hash(TextureFormat::ASTC_2D_10X10, UNORM, LINEAR): + return PixelFormat::ASTC_2D_10X10_UNORM; + case Hash(TextureFormat::ASTC_2D_10X10, UNORM, SRGB): + return PixelFormat::ASTC_2D_10X10_SRGB; + case Hash(TextureFormat::ASTC_2D_12X12, UNORM, LINEAR): + return PixelFormat::ASTC_2D_12X12_UNORM; + case Hash(TextureFormat::ASTC_2D_12X12, UNORM, SRGB): + return PixelFormat::ASTC_2D_12X12_SRGB; + case Hash(TextureFormat::ASTC_2D_8X6, UNORM, LINEAR): + return PixelFormat::ASTC_2D_8X6_UNORM; + case Hash(TextureFormat::ASTC_2D_8X6, UNORM, SRGB): + return PixelFormat::ASTC_2D_8X6_SRGB; + case Hash(TextureFormat::ASTC_2D_6X5, UNORM, LINEAR): + return PixelFormat::ASTC_2D_6X5_UNORM; + case Hash(TextureFormat::ASTC_2D_6X5, UNORM, SRGB): + return PixelFormat::ASTC_2D_6X5_SRGB; } UNIMPLEMENTED_MSG("texture format={} srgb={} components={{{} {} {} {}}}", - static_cast<int>(format), is_srgb, static_cast<int>(red_component), - static_cast<int>(green_component), static_cast<int>(blue_component), - static_cast<int>(alpha_component)); + static_cast<int>(format), is_srgb, static_cast<int>(red), + static_cast<int>(green), static_cast<int>(blue), static_cast<int>(alpha)); return PixelFormat::A8B8G8R8_UNORM; } -void FormatLookupTable::Set(TextureFormat format, bool is_srgb, ComponentType red_component, - ComponentType green_component, ComponentType blue_component, - ComponentType alpha_component, PixelFormat pixel_format) {} - -std::size_t FormatLookupTable::CalculateIndex(TextureFormat format, bool is_srgb, - ComponentType red_component, - ComponentType green_component, - ComponentType blue_component, - ComponentType alpha_component) noexcept { - const auto format_index = static_cast<std::size_t>(format); - const auto red_index = static_cast<std::size_t>(red_component); - const auto green_index = static_cast<std::size_t>(green_component); - const auto blue_index = static_cast<std::size_t>(blue_component); - const auto alpha_index = static_cast<std::size_t>(alpha_component); - const std::size_t srgb_index = is_srgb ? 1 : 0; - - return format_index * PerFormat + - srgb_index * PerComponent * PerComponent * PerComponent * PerComponent + - alpha_index * PerComponent * PerComponent * PerComponent + - blue_index * PerComponent * PerComponent + green_index * PerComponent + red_index; -} - } // namespace VideoCommon diff --git a/src/video_core/texture_cache/format_lookup_table.h b/src/video_core/texture_cache/format_lookup_table.h index aa77e0a5a..729533999 100644 --- a/src/video_core/texture_cache/format_lookup_table.h +++ b/src/video_core/texture_cache/format_lookup_table.h @@ -4,48 +4,14 @@ #pragma once -#include <array> -#include <limits> #include "video_core/surface.h" #include "video_core/textures/texture.h" namespace VideoCommon { -class FormatLookupTable { -public: - explicit FormatLookupTable(); - - VideoCore::Surface::PixelFormat GetPixelFormat( - Tegra::Texture::TextureFormat format, bool is_srgb, - Tegra::Texture::ComponentType red_component, Tegra::Texture::ComponentType green_component, - Tegra::Texture::ComponentType blue_component, - Tegra::Texture::ComponentType alpha_component) const noexcept; - -private: - static_assert(VideoCore::Surface::MaxPixelFormat <= std::numeric_limits<u8>::max()); - - static constexpr std::size_t NumTextureFormats = 128; - - static constexpr std::size_t PerComponent = 8; - static constexpr std::size_t PerComponents2 = PerComponent * PerComponent; - static constexpr std::size_t PerComponents3 = PerComponents2 * PerComponent; - static constexpr std::size_t PerComponents4 = PerComponents3 * PerComponent; - static constexpr std::size_t PerFormat = PerComponents4 * 2; - - static std::size_t CalculateIndex(Tegra::Texture::TextureFormat format, bool is_srgb, - Tegra::Texture::ComponentType red_component, - Tegra::Texture::ComponentType green_component, - Tegra::Texture::ComponentType blue_component, - Tegra::Texture::ComponentType alpha_component) noexcept; - - void Set(Tegra::Texture::TextureFormat format, bool is_srgb, - Tegra::Texture::ComponentType red_component, - Tegra::Texture::ComponentType green_component, - Tegra::Texture::ComponentType blue_component, - Tegra::Texture::ComponentType alpha_component, - VideoCore::Surface::PixelFormat pixel_format); - - std::array<u8, NumTextureFormats * PerFormat> table; -}; +VideoCore::Surface::PixelFormat PixelFormatFromTextureInfo( + Tegra::Texture::TextureFormat format, Tegra::Texture::ComponentType red_component, + Tegra::Texture::ComponentType green_component, Tegra::Texture::ComponentType blue_component, + Tegra::Texture::ComponentType alpha_component, bool is_srgb) noexcept; } // namespace VideoCommon diff --git a/src/video_core/texture_cache/formatter.cpp b/src/video_core/texture_cache/formatter.cpp new file mode 100644 index 000000000..d10ba4ccd --- /dev/null +++ b/src/video_core/texture_cache/formatter.cpp @@ -0,0 +1,95 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <string> + +#include "video_core/texture_cache/formatter.h" +#include "video_core/texture_cache/image_base.h" +#include "video_core/texture_cache/image_info.h" +#include "video_core/texture_cache/image_view_base.h" +#include "video_core/texture_cache/render_targets.h" + +namespace VideoCommon { + +std::string Name(const ImageBase& image) { + const GPUVAddr gpu_addr = image.gpu_addr; + const ImageInfo& info = image.info; + const u32 width = info.size.width; + const u32 height = info.size.height; + const u32 depth = info.size.depth; + const u32 num_layers = image.info.resources.layers; + const u32 num_levels = image.info.resources.levels; + std::string resource; + if (num_layers > 1) { + resource += fmt::format(":L{}", num_layers); + } + if (num_levels > 1) { + resource += fmt::format(":M{}", num_levels); + } + switch (image.info.type) { + case ImageType::e1D: + return fmt::format("Image 1D 0x{:x} {}{}", gpu_addr, width, resource); + case ImageType::e2D: + return fmt::format("Image 2D 0x{:x} {}x{}{}", gpu_addr, width, height, resource); + case ImageType::e3D: + return fmt::format("Image 2D 0x{:x} {}x{}x{}{}", gpu_addr, width, height, depth, resource); + case ImageType::Linear: + return fmt::format("Image Linear 0x{:x} {}x{}", gpu_addr, width, height); + case ImageType::Buffer: + return fmt::format("Buffer 0x{:x} {}", image.gpu_addr, image.info.size.width); + } + return "Invalid"; +} + +std::string Name(const ImageViewBase& image_view, std::optional<ImageViewType> type) { + const u32 width = image_view.size.width; + const u32 height = image_view.size.height; + const u32 depth = image_view.size.depth; + const u32 num_levels = image_view.range.extent.levels; + const u32 num_layers = image_view.range.extent.layers; + + const std::string level = num_levels > 1 ? fmt::format(":{}", num_levels) : ""; + switch (type.value_or(image_view.type)) { + case ImageViewType::e1D: + return fmt::format("ImageView 1D {}{}", width, level); + case ImageViewType::e2D: + return fmt::format("ImageView 2D {}x{}{}", width, height, level); + case ImageViewType::Cube: + return fmt::format("ImageView Cube {}x{}{}", width, height, level); + case ImageViewType::e3D: + return fmt::format("ImageView 3D {}x{}x{}{}", width, height, depth, level); + case ImageViewType::e1DArray: + return fmt::format("ImageView 1DArray {}{}|{}", width, level, num_layers); + case ImageViewType::e2DArray: + return fmt::format("ImageView 2DArray {}x{}{}|{}", width, height, level, num_layers); + case ImageViewType::CubeArray: + return fmt::format("ImageView CubeArray {}x{}{}|{}", width, height, level, num_layers); + case ImageViewType::Rect: + return fmt::format("ImageView Rect {}x{}{}", width, height, level); + case ImageViewType::Buffer: + return fmt::format("BufferView {}", width); + } + return "Invalid"; +} + +std::string Name(const RenderTargets& render_targets) { + std::string_view debug_prefix; + const auto num_color = std::ranges::count_if( + render_targets.color_buffer_ids, [](ImageViewId id) { return static_cast<bool>(id); }); + if (render_targets.depth_buffer_id) { + debug_prefix = num_color > 0 ? "R" : "Z"; + } else { + debug_prefix = num_color > 0 ? "C" : "X"; + } + const Extent2D size = render_targets.size; + if (num_color > 0) { + return fmt::format("Framebuffer {}{} {}x{}", debug_prefix, num_color, size.width, + size.height); + } else { + return fmt::format("Framebuffer {} {}x{}", debug_prefix, size.width, size.height); + } +} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/formatter.h b/src/video_core/texture_cache/formatter.h new file mode 100644 index 000000000..a48413983 --- /dev/null +++ b/src/video_core/texture_cache/formatter.h @@ -0,0 +1,263 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <string> + +#include <fmt/format.h> + +#include "video_core/surface.h" +#include "video_core/texture_cache/types.h" + +template <> +struct fmt::formatter<VideoCore::Surface::PixelFormat> : fmt::formatter<fmt::string_view> { + template <typename FormatContext> + auto format(VideoCore::Surface::PixelFormat format, FormatContext& ctx) { + using VideoCore::Surface::PixelFormat; + const string_view name = [format] { + switch (format) { + case PixelFormat::A8B8G8R8_UNORM: + return "A8B8G8R8_UNORM"; + case PixelFormat::A8B8G8R8_SNORM: + return "A8B8G8R8_SNORM"; + case PixelFormat::A8B8G8R8_SINT: + return "A8B8G8R8_SINT"; + case PixelFormat::A8B8G8R8_UINT: + return "A8B8G8R8_UINT"; + case PixelFormat::R5G6B5_UNORM: + return "R5G6B5_UNORM"; + case PixelFormat::B5G6R5_UNORM: + return "B5G6R5_UNORM"; + case PixelFormat::A1R5G5B5_UNORM: + return "A1R5G5B5_UNORM"; + case PixelFormat::A2B10G10R10_UNORM: + return "A2B10G10R10_UNORM"; + case PixelFormat::A2B10G10R10_UINT: + return "A2B10G10R10_UINT"; + case PixelFormat::A1B5G5R5_UNORM: + return "A1B5G5R5_UNORM"; + case PixelFormat::R8_UNORM: + return "R8_UNORM"; + case PixelFormat::R8_SNORM: + return "R8_SNORM"; + case PixelFormat::R8_SINT: + return "R8_SINT"; + case PixelFormat::R8_UINT: + return "R8_UINT"; + case PixelFormat::R16G16B16A16_FLOAT: + return "R16G16B16A16_FLOAT"; + case PixelFormat::R16G16B16A16_UNORM: + return "R16G16B16A16_UNORM"; + case PixelFormat::R16G16B16A16_SNORM: + return "R16G16B16A16_SNORM"; + case PixelFormat::R16G16B16A16_SINT: + return "R16G16B16A16_SINT"; + case PixelFormat::R16G16B16A16_UINT: + return "R16G16B16A16_UINT"; + case PixelFormat::B10G11R11_FLOAT: + return "B10G11R11_FLOAT"; + case PixelFormat::R32G32B32A32_UINT: + return "R32G32B32A32_UINT"; + case PixelFormat::BC1_RGBA_UNORM: + return "BC1_RGBA_UNORM"; + case PixelFormat::BC2_UNORM: + return "BC2_UNORM"; + case PixelFormat::BC3_UNORM: + return "BC3_UNORM"; + case PixelFormat::BC4_UNORM: + return "BC4_UNORM"; + case PixelFormat::BC4_SNORM: + return "BC4_SNORM"; + case PixelFormat::BC5_UNORM: + return "BC5_UNORM"; + case PixelFormat::BC5_SNORM: + return "BC5_SNORM"; + case PixelFormat::BC7_UNORM: + return "BC7_UNORM"; + case PixelFormat::BC6H_UFLOAT: + return "BC6H_UFLOAT"; + case PixelFormat::BC6H_SFLOAT: + return "BC6H_SFLOAT"; + case PixelFormat::ASTC_2D_4X4_UNORM: + return "ASTC_2D_4X4_UNORM"; + case PixelFormat::B8G8R8A8_UNORM: + return "B8G8R8A8_UNORM"; + case PixelFormat::R32G32B32A32_FLOAT: + return "R32G32B32A32_FLOAT"; + case PixelFormat::R32G32B32A32_SINT: + return "R32G32B32A32_SINT"; + case PixelFormat::R32G32_FLOAT: + return "R32G32_FLOAT"; + case PixelFormat::R32G32_SINT: + return "R32G32_SINT"; + case PixelFormat::R32_FLOAT: + return "R32_FLOAT"; + case PixelFormat::R16_FLOAT: + return "R16_FLOAT"; + case PixelFormat::R16_UNORM: + return "R16_UNORM"; + case PixelFormat::R16_SNORM: + return "R16_SNORM"; + case PixelFormat::R16_UINT: + return "R16_UINT"; + case PixelFormat::R16_SINT: + return "R16_SINT"; + case PixelFormat::R16G16_UNORM: + return "R16G16_UNORM"; + case PixelFormat::R16G16_FLOAT: + return "R16G16_FLOAT"; + case PixelFormat::R16G16_UINT: + return "R16G16_UINT"; + case PixelFormat::R16G16_SINT: + return "R16G16_SINT"; + case PixelFormat::R16G16_SNORM: + return "R16G16_SNORM"; + case PixelFormat::R32G32B32_FLOAT: + return "R32G32B32_FLOAT"; + case PixelFormat::A8B8G8R8_SRGB: + return "A8B8G8R8_SRGB"; + case PixelFormat::R8G8_UNORM: + return "R8G8_UNORM"; + case PixelFormat::R8G8_SNORM: + return "R8G8_SNORM"; + case PixelFormat::R8G8_SINT: + return "R8G8_SINT"; + case PixelFormat::R8G8_UINT: + return "R8G8_UINT"; + case PixelFormat::R32G32_UINT: + return "R32G32_UINT"; + case PixelFormat::R16G16B16X16_FLOAT: + return "R16G16B16X16_FLOAT"; + case PixelFormat::R32_UINT: + return "R32_UINT"; + case PixelFormat::R32_SINT: + return "R32_SINT"; + case PixelFormat::ASTC_2D_8X8_UNORM: + return "ASTC_2D_8X8_UNORM"; + case PixelFormat::ASTC_2D_8X5_UNORM: + return "ASTC_2D_8X5_UNORM"; + case PixelFormat::ASTC_2D_5X4_UNORM: + return "ASTC_2D_5X4_UNORM"; + case PixelFormat::B8G8R8A8_SRGB: + return "B8G8R8A8_SRGB"; + case PixelFormat::BC1_RGBA_SRGB: + return "BC1_RGBA_SRGB"; + case PixelFormat::BC2_SRGB: + return "BC2_SRGB"; + case PixelFormat::BC3_SRGB: + return "BC3_SRGB"; + case PixelFormat::BC7_SRGB: + return "BC7_SRGB"; + case PixelFormat::A4B4G4R4_UNORM: + return "A4B4G4R4_UNORM"; + case PixelFormat::ASTC_2D_4X4_SRGB: + return "ASTC_2D_4X4_SRGB"; + case PixelFormat::ASTC_2D_8X8_SRGB: + return "ASTC_2D_8X8_SRGB"; + case PixelFormat::ASTC_2D_8X5_SRGB: + return "ASTC_2D_8X5_SRGB"; + case PixelFormat::ASTC_2D_5X4_SRGB: + return "ASTC_2D_5X4_SRGB"; + case PixelFormat::ASTC_2D_5X5_UNORM: + return "ASTC_2D_5X5_UNORM"; + case PixelFormat::ASTC_2D_5X5_SRGB: + return "ASTC_2D_5X5_SRGB"; + case PixelFormat::ASTC_2D_10X8_UNORM: + return "ASTC_2D_10X8_UNORM"; + case PixelFormat::ASTC_2D_10X8_SRGB: + return "ASTC_2D_10X8_SRGB"; + case PixelFormat::ASTC_2D_6X6_UNORM: + return "ASTC_2D_6X6_UNORM"; + case PixelFormat::ASTC_2D_6X6_SRGB: + return "ASTC_2D_6X6_SRGB"; + case PixelFormat::ASTC_2D_10X10_UNORM: + return "ASTC_2D_10X10_UNORM"; + case PixelFormat::ASTC_2D_10X10_SRGB: + return "ASTC_2D_10X10_SRGB"; + case PixelFormat::ASTC_2D_12X12_UNORM: + return "ASTC_2D_12X12_UNORM"; + case PixelFormat::ASTC_2D_12X12_SRGB: + return "ASTC_2D_12X12_SRGB"; + case PixelFormat::ASTC_2D_8X6_UNORM: + return "ASTC_2D_8X6_UNORM"; + case PixelFormat::ASTC_2D_8X6_SRGB: + return "ASTC_2D_8X6_SRGB"; + case PixelFormat::ASTC_2D_6X5_UNORM: + return "ASTC_2D_6X5_UNORM"; + case PixelFormat::ASTC_2D_6X5_SRGB: + return "ASTC_2D_6X5_SRGB"; + case PixelFormat::E5B9G9R9_FLOAT: + return "E5B9G9R9_FLOAT"; + case PixelFormat::D32_FLOAT: + return "D32_FLOAT"; + case PixelFormat::D16_UNORM: + return "D16_UNORM"; + case PixelFormat::D24_UNORM_S8_UINT: + return "D24_UNORM_S8_UINT"; + case PixelFormat::S8_UINT_D24_UNORM: + return "S8_UINT_D24_UNORM"; + case PixelFormat::D32_FLOAT_S8_UINT: + return "D32_FLOAT_S8_UINT"; + case PixelFormat::MaxDepthStencilFormat: + case PixelFormat::Invalid: + return "Invalid"; + } + return "Invalid"; + }(); + return formatter<string_view>::format(name, ctx); + } +}; + +template <> +struct fmt::formatter<VideoCommon::ImageType> : fmt::formatter<fmt::string_view> { + template <typename FormatContext> + auto format(VideoCommon::ImageType type, FormatContext& ctx) { + const string_view name = [type] { + using VideoCommon::ImageType; + switch (type) { + case ImageType::e1D: + return "1D"; + case ImageType::e2D: + return "2D"; + case ImageType::e3D: + return "3D"; + case ImageType::Linear: + return "Linear"; + case ImageType::Buffer: + return "Buffer"; + } + return "Invalid"; + }(); + return formatter<string_view>::format(name, ctx); + } +}; + +template <> +struct fmt::formatter<VideoCommon::Extent3D> { + constexpr auto parse(fmt::format_parse_context& ctx) { + return ctx.begin(); + } + + template <typename FormatContext> + auto format(const VideoCommon::Extent3D& extent, FormatContext& ctx) { + return fmt::format_to(ctx.out(), "{{{}, {}, {}}}", extent.width, extent.height, + extent.depth); + } +}; + +namespace VideoCommon { + +struct ImageBase; +struct ImageViewBase; +struct RenderTargets; + +[[nodiscard]] std::string Name(const ImageBase& image); + +[[nodiscard]] std::string Name(const ImageViewBase& image_view, + std::optional<ImageViewType> type = std::nullopt); + +[[nodiscard]] std::string Name(const RenderTargets& render_targets); + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_base.cpp b/src/video_core/texture_cache/image_base.cpp new file mode 100644 index 000000000..959b3f115 --- /dev/null +++ b/src/video_core/texture_cache/image_base.cpp @@ -0,0 +1,218 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <optional> +#include <utility> +#include <vector> + +#include "common/common_types.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/formatter.h" +#include "video_core/texture_cache/image_base.h" +#include "video_core/texture_cache/image_view_info.h" +#include "video_core/texture_cache/util.h" + +namespace VideoCommon { + +using VideoCore::Surface::DefaultBlockHeight; +using VideoCore::Surface::DefaultBlockWidth; + +namespace { +/// Returns the base layer and mip level offset +[[nodiscard]] std::pair<s32, s32> LayerMipOffset(s32 diff, u32 layer_stride) { + if (layer_stride == 0) { + return {0, diff}; + } else { + return {diff / layer_stride, diff % layer_stride}; + } +} + +[[nodiscard]] bool ValidateLayers(const SubresourceLayers& layers, const ImageInfo& info) { + return layers.base_level < info.resources.levels && + layers.base_layer + layers.num_layers <= info.resources.layers; +} + +[[nodiscard]] bool ValidateCopy(const ImageCopy& copy, const ImageInfo& dst, const ImageInfo& src) { + const Extent3D src_size = MipSize(src.size, copy.src_subresource.base_level); + const Extent3D dst_size = MipSize(dst.size, copy.dst_subresource.base_level); + if (!ValidateLayers(copy.src_subresource, src)) { + return false; + } + if (!ValidateLayers(copy.dst_subresource, dst)) { + return false; + } + if (copy.src_offset.x + copy.extent.width > src_size.width || + copy.src_offset.y + copy.extent.height > src_size.height || + copy.src_offset.z + copy.extent.depth > src_size.depth) { + return false; + } + if (copy.dst_offset.x + copy.extent.width > dst_size.width || + copy.dst_offset.y + copy.extent.height > dst_size.height || + copy.dst_offset.z + copy.extent.depth > dst_size.depth) { + return false; + } + return true; +} +} // Anonymous namespace + +ImageBase::ImageBase(const ImageInfo& info_, GPUVAddr gpu_addr_, VAddr cpu_addr_) + : info{info_}, guest_size_bytes{CalculateGuestSizeInBytes(info)}, + unswizzled_size_bytes{CalculateUnswizzledSizeBytes(info)}, + converted_size_bytes{CalculateConvertedSizeBytes(info)}, gpu_addr{gpu_addr_}, + cpu_addr{cpu_addr_}, cpu_addr_end{cpu_addr + guest_size_bytes}, + mip_level_offsets{CalculateMipLevelOffsets(info)} { + if (info.type == ImageType::e3D) { + slice_offsets = CalculateSliceOffsets(info); + slice_subresources = CalculateSliceSubresources(info); + } +} + +std::optional<SubresourceBase> ImageBase::TryFindBase(GPUVAddr other_addr) const noexcept { + if (other_addr < gpu_addr) { + // Subresource address can't be lower than the base + return std::nullopt; + } + const u32 diff = static_cast<u32>(other_addr - gpu_addr); + if (diff > guest_size_bytes) { + // This can happen when two CPU addresses are used for different GPU addresses + return std::nullopt; + } + if (info.type != ImageType::e3D) { + const auto [layer, mip_offset] = LayerMipOffset(diff, info.layer_stride); + const auto end = mip_level_offsets.begin() + info.resources.levels; + const auto it = std::find(mip_level_offsets.begin(), end, mip_offset); + if (layer > info.resources.layers || it == end) { + return std::nullopt; + } + return SubresourceBase{ + .level = static_cast<s32>(std::distance(mip_level_offsets.begin(), it)), + .layer = layer, + }; + } else { + // TODO: Consider using binary_search after a threshold + const auto it = std::ranges::find(slice_offsets, diff); + if (it == slice_offsets.cend()) { + return std::nullopt; + } + return slice_subresources[std::distance(slice_offsets.begin(), it)]; + } +} + +ImageViewId ImageBase::FindView(const ImageViewInfo& view_info) const noexcept { + const auto it = std::ranges::find(image_view_infos, view_info); + if (it == image_view_infos.end()) { + return ImageViewId{}; + } + return image_view_ids[std::distance(image_view_infos.begin(), it)]; +} + +void ImageBase::InsertView(const ImageViewInfo& view_info, ImageViewId image_view_id) { + image_view_infos.push_back(view_info); + image_view_ids.push_back(image_view_id); +} + +void AddImageAlias(ImageBase& lhs, ImageBase& rhs, ImageId lhs_id, ImageId rhs_id) { + static constexpr auto OPTIONS = RelaxedOptions::Size | RelaxedOptions::Format; + ASSERT(lhs.info.type == rhs.info.type); + std::optional<SubresourceBase> base; + if (lhs.info.type == ImageType::Linear) { + base = SubresourceBase{.level = 0, .layer = 0}; + } else { + // We are passing relaxed formats as an option, having broken views or not won't matter + static constexpr bool broken_views = false; + base = FindSubresource(rhs.info, lhs, rhs.gpu_addr, OPTIONS, broken_views); + } + if (!base) { + LOG_ERROR(HW_GPU, "Image alias should have been flipped"); + return; + } + const PixelFormat lhs_format = lhs.info.format; + const PixelFormat rhs_format = rhs.info.format; + const Extent2D lhs_block{ + .width = DefaultBlockWidth(lhs_format), + .height = DefaultBlockHeight(lhs_format), + }; + const Extent2D rhs_block{ + .width = DefaultBlockWidth(rhs_format), + .height = DefaultBlockHeight(rhs_format), + }; + const bool is_lhs_compressed = lhs_block.width > 1 || lhs_block.height > 1; + const bool is_rhs_compressed = rhs_block.width > 1 || rhs_block.height > 1; + if (is_lhs_compressed && is_rhs_compressed) { + LOG_ERROR(HW_GPU, "Compressed to compressed image aliasing is not implemented"); + return; + } + const s32 lhs_mips = lhs.info.resources.levels; + const s32 rhs_mips = rhs.info.resources.levels; + const s32 num_mips = std::min(lhs_mips - base->level, rhs_mips); + AliasedImage lhs_alias; + AliasedImage rhs_alias; + lhs_alias.id = rhs_id; + rhs_alias.id = lhs_id; + lhs_alias.copies.reserve(num_mips); + rhs_alias.copies.reserve(num_mips); + for (s32 mip_level = 0; mip_level < num_mips; ++mip_level) { + Extent3D lhs_size = MipSize(lhs.info.size, base->level + mip_level); + Extent3D rhs_size = MipSize(rhs.info.size, mip_level); + if (is_lhs_compressed) { + lhs_size.width /= lhs_block.width; + lhs_size.height /= lhs_block.height; + } + if (is_rhs_compressed) { + rhs_size.width /= rhs_block.width; + rhs_size.height /= rhs_block.height; + } + const Extent3D copy_size{ + .width = std::min(lhs_size.width, rhs_size.width), + .height = std::min(lhs_size.height, rhs_size.height), + .depth = std::min(lhs_size.depth, rhs_size.depth), + }; + if (copy_size.width == 0 || copy_size.height == 0) { + LOG_WARNING(HW_GPU, "Copy size is smaller than block size. Mip cannot be aliased."); + continue; + } + const bool is_lhs_3d = lhs.info.type == ImageType::e3D; + const bool is_rhs_3d = rhs.info.type == ImageType::e3D; + const Offset3D lhs_offset{0, 0, 0}; + const Offset3D rhs_offset{0, 0, is_rhs_3d ? base->layer : 0}; + const s32 lhs_layers = is_lhs_3d ? 1 : lhs.info.resources.layers - base->layer; + const s32 rhs_layers = is_rhs_3d ? 1 : rhs.info.resources.layers; + const s32 num_layers = std::min(lhs_layers, rhs_layers); + const SubresourceLayers lhs_subresource{ + .base_level = mip_level, + .base_layer = 0, + .num_layers = num_layers, + }; + const SubresourceLayers rhs_subresource{ + .base_level = base->level + mip_level, + .base_layer = is_rhs_3d ? 0 : base->layer, + .num_layers = num_layers, + }; + [[maybe_unused]] const ImageCopy& to_lhs_copy = lhs_alias.copies.emplace_back(ImageCopy{ + .src_subresource = lhs_subresource, + .dst_subresource = rhs_subresource, + .src_offset = lhs_offset, + .dst_offset = rhs_offset, + .extent = copy_size, + }); + [[maybe_unused]] const ImageCopy& to_rhs_copy = rhs_alias.copies.emplace_back(ImageCopy{ + .src_subresource = rhs_subresource, + .dst_subresource = lhs_subresource, + .src_offset = rhs_offset, + .dst_offset = lhs_offset, + .extent = copy_size, + }); + ASSERT_MSG(ValidateCopy(to_lhs_copy, lhs.info, rhs.info), "Invalid RHS to LHS copy"); + ASSERT_MSG(ValidateCopy(to_rhs_copy, rhs.info, lhs.info), "Invalid LHS to RHS copy"); + } + ASSERT(lhs_alias.copies.empty() == rhs_alias.copies.empty()); + if (lhs_alias.copies.empty()) { + return; + } + lhs.aliased_images.push_back(std::move(lhs_alias)); + rhs.aliased_images.push_back(std::move(rhs_alias)); +} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_base.h b/src/video_core/texture_cache/image_base.h new file mode 100644 index 000000000..b7f3b7e43 --- /dev/null +++ b/src/video_core/texture_cache/image_base.h @@ -0,0 +1,83 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <optional> +#include <vector> + +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "video_core/texture_cache/image_info.h" +#include "video_core/texture_cache/image_view_info.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon { + +enum class ImageFlagBits : u32 { + AcceleratedUpload = 1 << 0, ///< Upload can be accelerated in the GPU + Converted = 1 << 1, ///< Guest format is not supported natively and it has to be converted + CpuModified = 1 << 2, ///< Contents have been modified from the CPU + GpuModified = 1 << 3, ///< Contents have been modified from the GPU + Tracked = 1 << 4, ///< Writes and reads are being hooked from the CPU JIT + Strong = 1 << 5, ///< Exists in the image table, the dimensions are can be trusted + Registered = 1 << 6, ///< True when the image is registered + Picked = 1 << 7, ///< Temporary flag to mark the image as picked +}; +DECLARE_ENUM_FLAG_OPERATORS(ImageFlagBits) + +struct ImageViewInfo; + +struct AliasedImage { + std::vector<ImageCopy> copies; + ImageId id; +}; + +struct ImageBase { + explicit ImageBase(const ImageInfo& info, GPUVAddr gpu_addr, VAddr cpu_addr); + + [[nodiscard]] std::optional<SubresourceBase> TryFindBase(GPUVAddr other_addr) const noexcept; + + [[nodiscard]] ImageViewId FindView(const ImageViewInfo& view_info) const noexcept; + + void InsertView(const ImageViewInfo& view_info, ImageViewId image_view_id); + + [[nodiscard]] bool Overlaps(VAddr overlap_cpu_addr, size_t overlap_size) const noexcept { + const VAddr overlap_end = overlap_cpu_addr + overlap_size; + return cpu_addr < overlap_end && overlap_cpu_addr < cpu_addr_end; + } + + ImageInfo info; + + u32 guest_size_bytes = 0; + u32 unswizzled_size_bytes = 0; + u32 converted_size_bytes = 0; + ImageFlagBits flags = ImageFlagBits::CpuModified; + + GPUVAddr gpu_addr = 0; + VAddr cpu_addr = 0; + VAddr cpu_addr_end = 0; + + u64 modification_tick = 0; + u64 frame_tick = 0; + + std::array<u32, MAX_MIP_LEVELS> mip_level_offsets{}; + + std::vector<ImageViewInfo> image_view_infos; + std::vector<ImageViewId> image_view_ids; + + std::vector<u32> slice_offsets; + std::vector<SubresourceBase> slice_subresources; + + std::vector<AliasedImage> aliased_images; +}; + +struct ImageAllocBase { + std::vector<ImageId> images; +}; + +void AddImageAlias(ImageBase& lhs, ImageBase& rhs, ImageId lhs_id, ImageId rhs_id); + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_info.cpp b/src/video_core/texture_cache/image_info.cpp new file mode 100644 index 000000000..64fd7010a --- /dev/null +++ b/src/video_core/texture_cache/image_info.cpp @@ -0,0 +1,189 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/assert.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/format_lookup_table.h" +#include "video_core/texture_cache/image_info.h" +#include "video_core/texture_cache/samples_helper.h" +#include "video_core/texture_cache/types.h" +#include "video_core/texture_cache/util.h" +#include "video_core/textures/texture.h" + +namespace VideoCommon { + +using Tegra::Texture::TextureType; +using Tegra::Texture::TICEntry; +using VideoCore::Surface::PixelFormat; + +ImageInfo::ImageInfo(const TICEntry& config) noexcept { + format = PixelFormatFromTextureInfo(config.format, config.r_type, config.g_type, config.b_type, + config.a_type, config.srgb_conversion); + num_samples = NumSamples(config.msaa_mode); + resources.levels = config.max_mip_level + 1; + if (config.IsPitchLinear()) { + pitch = config.Pitch(); + } else if (config.IsBlockLinear()) { + block = Extent3D{ + .width = config.block_width, + .height = config.block_height, + .depth = config.block_depth, + }; + } + tile_width_spacing = config.tile_width_spacing; + if (config.texture_type != TextureType::Texture2D && + config.texture_type != TextureType::Texture2DNoMipmap) { + ASSERT(!config.IsPitchLinear()); + } + switch (config.texture_type) { + case TextureType::Texture1D: + ASSERT(config.BaseLayer() == 0); + type = ImageType::e1D; + size.width = config.Width(); + break; + case TextureType::Texture1DArray: + UNIMPLEMENTED_IF(config.BaseLayer() != 0); + type = ImageType::e1D; + size.width = config.Width(); + resources.layers = config.Depth(); + break; + case TextureType::Texture2D: + case TextureType::Texture2DNoMipmap: + ASSERT(config.Depth() == 1); + type = config.IsPitchLinear() ? ImageType::Linear : ImageType::e2D; + size.width = config.Width(); + size.height = config.Height(); + resources.layers = config.BaseLayer() + 1; + break; + case TextureType::Texture2DArray: + type = ImageType::e2D; + size.width = config.Width(); + size.height = config.Height(); + resources.layers = config.BaseLayer() + config.Depth(); + break; + case TextureType::TextureCubemap: + ASSERT(config.Depth() == 1); + type = ImageType::e2D; + size.width = config.Width(); + size.height = config.Height(); + resources.layers = config.BaseLayer() + 6; + break; + case TextureType::TextureCubeArray: + UNIMPLEMENTED_IF(config.load_store_hint != 0); + type = ImageType::e2D; + size.width = config.Width(); + size.height = config.Height(); + resources.layers = config.BaseLayer() + config.Depth() * 6; + break; + case TextureType::Texture3D: + ASSERT(config.BaseLayer() == 0); + type = ImageType::e3D; + size.width = config.Width(); + size.height = config.Height(); + size.depth = config.Depth(); + break; + case TextureType::Texture1DBuffer: + type = ImageType::Buffer; + size.width = config.Width(); + break; + default: + UNREACHABLE_MSG("Invalid texture_type={}", static_cast<int>(config.texture_type.Value())); + break; + } + if (type != ImageType::Linear) { + // FIXME: Call this without passing *this + layer_stride = CalculateLayerStride(*this); + maybe_unaligned_layer_stride = CalculateLayerSize(*this); + } +} + +ImageInfo::ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs, size_t index) noexcept { + const auto& rt = regs.rt[index]; + format = VideoCore::Surface::PixelFormatFromRenderTargetFormat(rt.format); + if (rt.tile_mode.is_pitch_linear) { + ASSERT(rt.tile_mode.is_3d == 0); + type = ImageType::Linear; + pitch = rt.width; + size = Extent3D{ + .width = pitch / BytesPerBlock(format), + .height = rt.height, + .depth = 1, + }; + return; + } + size.width = rt.width; + size.height = rt.height; + layer_stride = rt.layer_stride * 4; + maybe_unaligned_layer_stride = layer_stride; + num_samples = NumSamples(regs.multisample_mode); + block = Extent3D{ + .width = rt.tile_mode.block_width, + .height = rt.tile_mode.block_height, + .depth = rt.tile_mode.block_depth, + }; + if (rt.tile_mode.is_3d) { + type = ImageType::e3D; + size.depth = rt.depth; + } else { + type = ImageType::e2D; + resources.layers = rt.depth; + } +} + +ImageInfo::ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs) noexcept { + format = VideoCore::Surface::PixelFormatFromDepthFormat(regs.zeta.format); + size.width = regs.zeta_width; + size.height = regs.zeta_height; + resources.levels = 1; + layer_stride = regs.zeta.layer_stride * 4; + maybe_unaligned_layer_stride = layer_stride; + num_samples = NumSamples(regs.multisample_mode); + block = Extent3D{ + .width = regs.zeta.tile_mode.block_width, + .height = regs.zeta.tile_mode.block_height, + .depth = regs.zeta.tile_mode.block_depth, + }; + if (regs.zeta.tile_mode.is_pitch_linear) { + ASSERT(regs.zeta.tile_mode.is_3d == 0); + type = ImageType::Linear; + pitch = size.width * BytesPerBlock(format); + } else if (regs.zeta.tile_mode.is_3d) { + ASSERT(regs.zeta.tile_mode.is_pitch_linear == 0); + type = ImageType::e3D; + size.depth = regs.zeta_depth; + } else { + type = ImageType::e2D; + resources.layers = regs.zeta_depth; + } +} + +ImageInfo::ImageInfo(const Tegra::Engines::Fermi2D::Surface& config) noexcept { + UNIMPLEMENTED_IF_MSG(config.layer != 0, "Surface layer is not zero"); + format = VideoCore::Surface::PixelFormatFromRenderTargetFormat(config.format); + if (config.linear == Tegra::Engines::Fermi2D::MemoryLayout::Pitch) { + type = ImageType::Linear; + size = Extent3D{ + .width = config.pitch / VideoCore::Surface::BytesPerBlock(format), + .height = config.height, + .depth = 1, + }; + pitch = config.pitch; + } else { + type = config.block_depth > 0 ? ImageType::e3D : ImageType::e2D; + block = Extent3D{ + .width = config.block_width, + .height = config.block_height, + .depth = config.block_depth, + }; + // 3D blits with more than once slice are not implemented for now + // Render to individual slices + size = Extent3D{ + .width = config.width, + .height = config.height, + .depth = 1, + }; + } +} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_info.h b/src/video_core/texture_cache/image_info.h new file mode 100644 index 000000000..5049fc36e --- /dev/null +++ b/src/video_core/texture_cache/image_info.h @@ -0,0 +1,38 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "video_core/engines/fermi_2d.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon { + +using Tegra::Texture::TICEntry; +using VideoCore::Surface::PixelFormat; + +struct ImageInfo { + explicit ImageInfo() = default; + explicit ImageInfo(const TICEntry& config) noexcept; + explicit ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs, size_t index) noexcept; + explicit ImageInfo(const Tegra::Engines::Maxwell3D::Regs& regs) noexcept; + explicit ImageInfo(const Tegra::Engines::Fermi2D::Surface& config) noexcept; + + PixelFormat format = PixelFormat::Invalid; + ImageType type = ImageType::e1D; + SubresourceExtent resources; + Extent3D size{1, 1, 1}; + union { + Extent3D block{0, 0, 0}; + u32 pitch; + }; + u32 layer_stride = 0; + u32 maybe_unaligned_layer_stride = 0; + u32 num_samples = 1; + u32 tile_width_spacing = 0; +}; + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_view_base.cpp b/src/video_core/texture_cache/image_view_base.cpp new file mode 100644 index 000000000..18f72e508 --- /dev/null +++ b/src/video_core/texture_cache/image_view_base.cpp @@ -0,0 +1,41 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> + +#include "common/assert.h" +#include "core/settings.h" +#include "video_core/compatible_formats.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/formatter.h" +#include "video_core/texture_cache/image_info.h" +#include "video_core/texture_cache/image_view_base.h" +#include "video_core/texture_cache/image_view_info.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon { + +ImageViewBase::ImageViewBase(const ImageViewInfo& info, const ImageInfo& image_info, + ImageId image_id_) + : image_id{image_id_}, format{info.format}, type{info.type}, range{info.range}, + size{ + .width = std::max(image_info.size.width >> range.base.level, 1u), + .height = std::max(image_info.size.height >> range.base.level, 1u), + .depth = std::max(image_info.size.depth >> range.base.level, 1u), + } { + ASSERT_MSG(VideoCore::Surface::IsViewCompatible(image_info.format, info.format, false), + "Image view format {} is incompatible with image format {}", info.format, + image_info.format); + const bool is_async = Settings::values.use_asynchronous_gpu_emulation.GetValue(); + if (image_info.type == ImageType::Linear && is_async) { + flags |= ImageViewFlagBits::PreemtiveDownload; + } + if (image_info.type == ImageType::e3D && info.type != ImageViewType::e3D) { + flags |= ImageViewFlagBits::Slice; + } +} + +ImageViewBase::ImageViewBase(const NullImageParams&) {} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_view_base.h b/src/video_core/texture_cache/image_view_base.h new file mode 100644 index 000000000..73954167e --- /dev/null +++ b/src/video_core/texture_cache/image_view_base.h @@ -0,0 +1,47 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon { + +using VideoCore::Surface::PixelFormat; + +struct ImageViewInfo; +struct ImageInfo; + +struct NullImageParams {}; + +enum class ImageViewFlagBits : u16 { + PreemtiveDownload = 1 << 0, + Strong = 1 << 1, + Slice = 1 << 2, +}; +DECLARE_ENUM_FLAG_OPERATORS(ImageViewFlagBits) + +struct ImageViewBase { + explicit ImageViewBase(const ImageViewInfo& info, const ImageInfo& image_info, + ImageId image_id); + explicit ImageViewBase(const NullImageParams&); + + [[nodiscard]] bool IsBuffer() const noexcept { + return type == ImageViewType::Buffer; + } + + ImageId image_id{}; + PixelFormat format{}; + ImageViewType type{}; + SubresourceRange range; + Extent3D size{0, 0, 0}; + ImageViewFlagBits flags{}; + + u64 invalidation_tick = 0; + u64 modification_tick = 0; +}; + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_view_info.cpp b/src/video_core/texture_cache/image_view_info.cpp new file mode 100644 index 000000000..faf5b151f --- /dev/null +++ b/src/video_core/texture_cache/image_view_info.cpp @@ -0,0 +1,88 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <limits> + +#include "common/assert.h" +#include "video_core/texture_cache/image_view_info.h" +#include "video_core/texture_cache/texture_cache.h" +#include "video_core/texture_cache/types.h" +#include "video_core/textures/texture.h" + +namespace VideoCommon { + +namespace { + +constexpr u8 RENDER_TARGET_SWIZZLE = std::numeric_limits<u8>::max(); + +[[nodiscard]] u8 CastSwizzle(SwizzleSource source) { + const u8 casted = static_cast<u8>(source); + ASSERT(static_cast<SwizzleSource>(casted) == source); + return casted; +} + +} // Anonymous namespace + +ImageViewInfo::ImageViewInfo(const TICEntry& config, s32 base_layer) noexcept + : format{PixelFormatFromTIC(config)}, x_source{CastSwizzle(config.x_source)}, + y_source{CastSwizzle(config.y_source)}, z_source{CastSwizzle(config.z_source)}, + w_source{CastSwizzle(config.w_source)} { + range.base = SubresourceBase{ + .level = static_cast<s32>(config.res_min_mip_level), + .layer = base_layer, + }; + range.extent.levels = config.res_max_mip_level - config.res_min_mip_level + 1; + + switch (config.texture_type) { + case TextureType::Texture1D: + ASSERT(config.Height() == 1); + ASSERT(config.Depth() == 1); + type = ImageViewType::e1D; + break; + case TextureType::Texture2D: + case TextureType::Texture2DNoMipmap: + ASSERT(config.Depth() == 1); + type = config.normalized_coords ? ImageViewType::e2D : ImageViewType::Rect; + break; + case TextureType::Texture3D: + type = ImageViewType::e3D; + break; + case TextureType::TextureCubemap: + ASSERT(config.Depth() == 1); + type = ImageViewType::Cube; + range.extent.layers = 6; + break; + case TextureType::Texture1DArray: + type = ImageViewType::e1DArray; + range.extent.layers = config.Depth(); + break; + case TextureType::Texture2DArray: + type = ImageViewType::e2DArray; + range.extent.layers = config.Depth(); + break; + case TextureType::Texture1DBuffer: + type = ImageViewType::Buffer; + break; + case TextureType::TextureCubeArray: + type = ImageViewType::CubeArray; + range.extent.layers = config.Depth() * 6; + break; + default: + UNREACHABLE_MSG("Invalid texture_type={}", static_cast<int>(config.texture_type.Value())); + break; + } +} + +ImageViewInfo::ImageViewInfo(ImageViewType type_, PixelFormat format_, + SubresourceRange range_) noexcept + : type{type_}, format{format_}, range{range_}, x_source{RENDER_TARGET_SWIZZLE}, + y_source{RENDER_TARGET_SWIZZLE}, z_source{RENDER_TARGET_SWIZZLE}, + w_source{RENDER_TARGET_SWIZZLE} {} + +bool ImageViewInfo::IsRenderTarget() const noexcept { + return x_source == RENDER_TARGET_SWIZZLE && y_source == RENDER_TARGET_SWIZZLE && + z_source == RENDER_TARGET_SWIZZLE && w_source == RENDER_TARGET_SWIZZLE; +} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/image_view_info.h b/src/video_core/texture_cache/image_view_info.h new file mode 100644 index 000000000..0c1f99117 --- /dev/null +++ b/src/video_core/texture_cache/image_view_info.h @@ -0,0 +1,50 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <type_traits> + +#include "video_core/surface.h" +#include "video_core/texture_cache/types.h" +#include "video_core/textures/texture.h" + +namespace VideoCommon { + +using Tegra::Texture::SwizzleSource; +using Tegra::Texture::TICEntry; +using VideoCore::Surface::PixelFormat; + +/// Properties used to determine a image view +struct ImageViewInfo { + explicit ImageViewInfo() noexcept = default; + explicit ImageViewInfo(const TICEntry& config, s32 base_layer) noexcept; + explicit ImageViewInfo(ImageViewType type, PixelFormat format, + SubresourceRange range = {}) noexcept; + + auto operator<=>(const ImageViewInfo&) const noexcept = default; + + [[nodiscard]] bool IsRenderTarget() const noexcept; + + [[nodiscard]] std::array<SwizzleSource, 4> Swizzle() const noexcept { + return std::array{ + static_cast<SwizzleSource>(x_source), + static_cast<SwizzleSource>(y_source), + static_cast<SwizzleSource>(z_source), + static_cast<SwizzleSource>(w_source), + }; + } + + ImageViewType type{}; + PixelFormat format{}; + SubresourceRange range; + u8 x_source = static_cast<u8>(SwizzleSource::R); + u8 y_source = static_cast<u8>(SwizzleSource::G); + u8 z_source = static_cast<u8>(SwizzleSource::B); + u8 w_source = static_cast<u8>(SwizzleSource::A); +}; +static_assert(std::has_unique_object_representations_v<ImageViewInfo>); + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/render_targets.h b/src/video_core/texture_cache/render_targets.h new file mode 100644 index 000000000..9b9544b07 --- /dev/null +++ b/src/video_core/texture_cache/render_targets.h @@ -0,0 +1,51 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <algorithm> +#include <span> +#include <utility> + +#include "common/bit_cast.h" +#include "video_core/texture_cache/types.h" + +namespace VideoCommon { + +/// Framebuffer properties used to lookup a framebuffer +struct RenderTargets { + constexpr auto operator<=>(const RenderTargets&) const noexcept = default; + + constexpr bool Contains(std::span<const ImageViewId> elements) const noexcept { + const auto contains = [elements](ImageViewId item) { + return std::ranges::find(elements, item) != elements.end(); + }; + return std::ranges::any_of(color_buffer_ids, contains) || contains(depth_buffer_id); + } + + std::array<ImageViewId, NUM_RT> color_buffer_ids; + ImageViewId depth_buffer_id; + std::array<u8, NUM_RT> draw_buffers{}; + Extent2D size; +}; + +} // namespace VideoCommon + +namespace std { + +template <> +struct hash<VideoCommon::RenderTargets> { + size_t operator()(const VideoCommon::RenderTargets& rt) const noexcept { + using VideoCommon::ImageViewId; + size_t value = std::hash<ImageViewId>{}(rt.depth_buffer_id); + for (const ImageViewId color_buffer_id : rt.color_buffer_ids) { + value ^= std::hash<ImageViewId>{}(color_buffer_id); + } + value ^= Common::BitCast<u64>(rt.draw_buffers); + value ^= Common::BitCast<u64>(rt.size); + return value; + } +}; + +} // namespace std diff --git a/src/video_core/texture_cache/samples_helper.h b/src/video_core/texture_cache/samples_helper.h new file mode 100644 index 000000000..04539a43c --- /dev/null +++ b/src/video_core/texture_cache/samples_helper.h @@ -0,0 +1,55 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <utility> + +#include "common/assert.h" +#include "video_core/textures/texture.h" + +namespace VideoCommon { + +[[nodiscard]] inline std::pair<int, int> SamplesLog2(int num_samples) { + switch (num_samples) { + case 1: + return {0, 0}; + case 2: + return {1, 0}; + case 4: + return {1, 1}; + case 8: + return {2, 1}; + case 16: + return {2, 2}; + } + UNREACHABLE_MSG("Invalid number of samples={}", num_samples); + return {1, 1}; +} + +[[nodiscard]] inline int NumSamples(Tegra::Texture::MsaaMode msaa_mode) { + using Tegra::Texture::MsaaMode; + switch (msaa_mode) { + case MsaaMode::Msaa1x1: + return 1; + case MsaaMode::Msaa2x1: + case MsaaMode::Msaa2x1_D3D: + return 2; + case MsaaMode::Msaa2x2: + case MsaaMode::Msaa2x2_VC4: + case MsaaMode::Msaa2x2_VC12: + return 4; + case MsaaMode::Msaa4x2: + case MsaaMode::Msaa4x2_D3D: + case MsaaMode::Msaa4x2_VC8: + case MsaaMode::Msaa4x2_VC24: + return 8; + case MsaaMode::Msaa4x4: + return 16; + } + UNREACHABLE_MSG("Invalid MSAA mode={}", static_cast<int>(msaa_mode)); + return 1; +} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/slot_vector.h b/src/video_core/texture_cache/slot_vector.h new file mode 100644 index 000000000..eae3be6ea --- /dev/null +++ b/src/video_core/texture_cache/slot_vector.h @@ -0,0 +1,156 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <concepts> +#include <numeric> +#include <type_traits> +#include <utility> +#include <vector> + +#include "common/assert.h" +#include "common/common_types.h" + +namespace VideoCommon { + +struct SlotId { + static constexpr u32 INVALID_INDEX = std::numeric_limits<u32>::max(); + + constexpr auto operator<=>(const SlotId&) const noexcept = default; + + constexpr explicit operator bool() const noexcept { + return index != INVALID_INDEX; + } + + u32 index = INVALID_INDEX; +}; + +template <class T> +requires std::is_nothrow_move_assignable_v<T>&& + std::is_nothrow_move_constructible_v<T> class SlotVector { +public: + ~SlotVector() noexcept { + size_t index = 0; + for (u64 bits : stored_bitset) { + for (size_t bit = 0; bits; ++bit, bits >>= 1) { + if ((bits & 1) != 0) { + values[index + bit].object.~T(); + } + } + index += 64; + } + delete[] values; + } + + [[nodiscard]] T& operator[](SlotId id) noexcept { + ValidateIndex(id); + return values[id.index].object; + } + + [[nodiscard]] const T& operator[](SlotId id) const noexcept { + ValidateIndex(id); + return values[id.index].object; + } + + template <typename... Args> + [[nodiscard]] SlotId insert(Args&&... args) noexcept { + const u32 index = FreeValueIndex(); + new (&values[index].object) T(std::forward<Args>(args)...); + SetStorageBit(index); + + return SlotId{index}; + } + + void erase(SlotId id) noexcept { + values[id.index].object.~T(); + free_list.push_back(id.index); + ResetStorageBit(id.index); + } + +private: + struct NonTrivialDummy { + NonTrivialDummy() noexcept {} + }; + + union Entry { + Entry() noexcept : dummy{} {} + ~Entry() noexcept {} + + NonTrivialDummy dummy; + T object; + }; + + void SetStorageBit(u32 index) noexcept { + stored_bitset[index / 64] |= u64(1) << (index % 64); + } + + void ResetStorageBit(u32 index) noexcept { + stored_bitset[index / 64] &= ~(u64(1) << (index % 64)); + } + + bool ReadStorageBit(u32 index) noexcept { + return ((stored_bitset[index / 64] >> (index % 64)) & 1) != 0; + } + + void ValidateIndex(SlotId id) const noexcept { + DEBUG_ASSERT(id); + DEBUG_ASSERT(id.index / 64 < stored_bitset.size()); + DEBUG_ASSERT(((stored_bitset[id.index / 64] >> (id.index % 64)) & 1) != 0); + } + + [[nodiscard]] u32 FreeValueIndex() noexcept { + if (free_list.empty()) { + Reserve(values_capacity ? (values_capacity << 1) : 1); + } + const u32 free_index = free_list.back(); + free_list.pop_back(); + return free_index; + } + + void Reserve(size_t new_capacity) noexcept { + Entry* const new_values = new Entry[new_capacity]; + size_t index = 0; + for (u64 bits : stored_bitset) { + for (size_t bit = 0; bits; ++bit, bits >>= 1) { + const size_t i = index + bit; + if ((bits & 1) == 0) { + continue; + } + T& old_value = values[i].object; + new (&new_values[i].object) T(std::move(old_value)); + old_value.~T(); + } + index += 64; + } + + stored_bitset.resize((new_capacity + 63) / 64); + + const size_t old_free_size = free_list.size(); + free_list.resize(old_free_size + (new_capacity - values_capacity)); + std::iota(free_list.begin() + old_free_size, free_list.end(), + static_cast<u32>(values_capacity)); + + delete[] values; + values = new_values; + values_capacity = new_capacity; + } + + Entry* values = nullptr; + size_t values_capacity = 0; + size_t values_size = 0; + + std::vector<u64> stored_bitset; + std::vector<u32> free_list; +}; + +} // namespace VideoCommon + +template <> +struct std::hash<VideoCommon::SlotId> { + size_t operator()(const VideoCommon::SlotId& id) const noexcept { + return std::hash<u32>{}(id.index); + } +}; diff --git a/src/video_core/texture_cache/surface_base.cpp b/src/video_core/texture_cache/surface_base.cpp deleted file mode 100644 index b44c09d71..000000000 --- a/src/video_core/texture_cache/surface_base.cpp +++ /dev/null @@ -1,298 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "common/algorithm.h" -#include "common/assert.h" -#include "common/common_types.h" -#include "common/microprofile.h" -#include "video_core/memory_manager.h" -#include "video_core/texture_cache/surface_base.h" -#include "video_core/texture_cache/surface_params.h" -#include "video_core/textures/convert.h" - -namespace VideoCommon { - -MICROPROFILE_DEFINE(GPU_Load_Texture, "GPU", "Texture Load", MP_RGB(128, 192, 128)); -MICROPROFILE_DEFINE(GPU_Flush_Texture, "GPU", "Texture Flush", MP_RGB(128, 192, 128)); - -using Tegra::Texture::ConvertFromGuestToHost; -using VideoCore::MortonSwizzleMode; -using VideoCore::Surface::IsPixelFormatASTC; -using VideoCore::Surface::PixelFormat; - -StagingCache::StagingCache() = default; - -StagingCache::~StagingCache() = default; - -SurfaceBaseImpl::SurfaceBaseImpl(GPUVAddr gpu_addr, const SurfaceParams& params, - bool is_astc_supported) - : params{params}, gpu_addr{gpu_addr}, mipmap_sizes(params.num_levels), - mipmap_offsets(params.num_levels) { - is_converted = IsPixelFormatASTC(params.pixel_format) && !is_astc_supported; - host_memory_size = params.GetHostSizeInBytes(is_converted); - - std::size_t offset = 0; - for (u32 level = 0; level < params.num_levels; ++level) { - const std::size_t mipmap_size{params.GetGuestMipmapSize(level)}; - mipmap_sizes[level] = mipmap_size; - mipmap_offsets[level] = offset; - offset += mipmap_size; - } - layer_size = offset; - if (params.is_layered) { - if (params.is_tiled) { - layer_size = - SurfaceParams::AlignLayered(layer_size, params.block_height, params.block_depth); - } - guest_memory_size = layer_size * params.depth; - } else { - guest_memory_size = layer_size; - } -} - -MatchTopologyResult SurfaceBaseImpl::MatchesTopology(const SurfaceParams& rhs) const { - const u32 src_bpp{params.GetBytesPerPixel()}; - const u32 dst_bpp{rhs.GetBytesPerPixel()}; - const bool ib1 = params.IsBuffer(); - const bool ib2 = rhs.IsBuffer(); - if (std::tie(src_bpp, params.is_tiled, ib1) == std::tie(dst_bpp, rhs.is_tiled, ib2)) { - const bool cb1 = params.IsCompressed(); - const bool cb2 = rhs.IsCompressed(); - if (cb1 == cb2) { - return MatchTopologyResult::FullMatch; - } - return MatchTopologyResult::CompressUnmatch; - } - return MatchTopologyResult::None; -} - -MatchStructureResult SurfaceBaseImpl::MatchesStructure(const SurfaceParams& rhs) const { - // Buffer surface Check - if (params.IsBuffer()) { - const std::size_t wd1 = params.width * params.GetBytesPerPixel(); - const std::size_t wd2 = rhs.width * rhs.GetBytesPerPixel(); - if (wd1 == wd2) { - return MatchStructureResult::FullMatch; - } - return MatchStructureResult::None; - } - - // Linear Surface check - if (!params.is_tiled) { - if (std::tie(params.height, params.pitch) == std::tie(rhs.height, rhs.pitch)) { - if (params.width == rhs.width) { - return MatchStructureResult::FullMatch; - } else { - return MatchStructureResult::SemiMatch; - } - } - return MatchStructureResult::None; - } - - // Tiled Surface check - if (std::tie(params.depth, params.block_width, params.block_height, params.block_depth, - params.tile_width_spacing, params.num_levels) == - std::tie(rhs.depth, rhs.block_width, rhs.block_height, rhs.block_depth, - rhs.tile_width_spacing, rhs.num_levels)) { - if (std::tie(params.width, params.height) == std::tie(rhs.width, rhs.height)) { - return MatchStructureResult::FullMatch; - } - const u32 ws = SurfaceParams::ConvertWidth(rhs.GetBlockAlignedWidth(), params.pixel_format, - rhs.pixel_format); - const u32 hs = - SurfaceParams::ConvertHeight(rhs.height, params.pixel_format, rhs.pixel_format); - const u32 w1 = params.GetBlockAlignedWidth(); - if (std::tie(w1, params.height) == std::tie(ws, hs)) { - return MatchStructureResult::SemiMatch; - } - } - return MatchStructureResult::None; -} - -std::optional<std::pair<u32, u32>> SurfaceBaseImpl::GetLayerMipmap( - const GPUVAddr candidate_gpu_addr) const { - if (gpu_addr == candidate_gpu_addr) { - return {{0, 0}}; - } - - if (candidate_gpu_addr < gpu_addr) { - return std::nullopt; - } - - const auto relative_address{static_cast<GPUVAddr>(candidate_gpu_addr - gpu_addr)}; - const auto layer{static_cast<u32>(relative_address / layer_size)}; - if (layer >= params.depth) { - return std::nullopt; - } - - const GPUVAddr mipmap_address = relative_address - layer_size * layer; - const auto mipmap_it = - Common::BinaryFind(mipmap_offsets.begin(), mipmap_offsets.end(), mipmap_address); - if (mipmap_it == mipmap_offsets.end()) { - return std::nullopt; - } - - const auto level{static_cast<u32>(std::distance(mipmap_offsets.begin(), mipmap_it))}; - return std::make_pair(layer, level); -} - -std::vector<CopyParams> SurfaceBaseImpl::BreakDownLayered(const SurfaceParams& in_params) const { - const u32 layers{params.depth}; - const u32 mipmaps{params.num_levels}; - std::vector<CopyParams> result; - result.reserve(static_cast<std::size_t>(layers) * static_cast<std::size_t>(mipmaps)); - - for (u32 layer = 0; layer < layers; layer++) { - for (u32 level = 0; level < mipmaps; level++) { - const u32 width = SurfaceParams::IntersectWidth(params, in_params, level, level); - const u32 height = SurfaceParams::IntersectHeight(params, in_params, level, level); - result.emplace_back(0, 0, layer, 0, 0, layer, level, level, width, height, 1); - } - } - return result; -} - -std::vector<CopyParams> SurfaceBaseImpl::BreakDownNonLayered(const SurfaceParams& in_params) const { - const u32 mipmaps{params.num_levels}; - std::vector<CopyParams> result; - result.reserve(mipmaps); - - for (u32 level = 0; level < mipmaps; level++) { - const u32 width = SurfaceParams::IntersectWidth(params, in_params, level, level); - const u32 height = SurfaceParams::IntersectHeight(params, in_params, level, level); - const u32 depth{std::min(params.GetMipDepth(level), in_params.GetMipDepth(level))}; - result.emplace_back(width, height, depth, level); - } - return result; -} - -void SurfaceBaseImpl::SwizzleFunc(MortonSwizzleMode mode, u8* memory, const SurfaceParams& params, - u8* buffer, u32 level) { - const u32 width{params.GetMipWidth(level)}; - const u32 height{params.GetMipHeight(level)}; - const u32 block_height{params.GetMipBlockHeight(level)}; - const u32 block_depth{params.GetMipBlockDepth(level)}; - - std::size_t guest_offset{mipmap_offsets[level]}; - if (params.is_layered) { - std::size_t host_offset = 0; - const std::size_t guest_stride = layer_size; - const std::size_t host_stride = params.GetHostLayerSize(level); - for (u32 layer = 0; layer < params.depth; ++layer) { - MortonSwizzle(mode, params.pixel_format, width, block_height, height, block_depth, 1, - params.tile_width_spacing, buffer + host_offset, memory + guest_offset); - guest_offset += guest_stride; - host_offset += host_stride; - } - } else { - MortonSwizzle(mode, params.pixel_format, width, block_height, height, block_depth, - params.GetMipDepth(level), params.tile_width_spacing, buffer, - memory + guest_offset); - } -} - -void SurfaceBaseImpl::LoadBuffer(Tegra::MemoryManager& memory_manager, - StagingCache& staging_cache) { - MICROPROFILE_SCOPE(GPU_Load_Texture); - auto& staging_buffer = staging_cache.GetBuffer(0); - u8* host_ptr; - // Use an extra temporal buffer - auto& tmp_buffer = staging_cache.GetBuffer(1); - tmp_buffer.resize(guest_memory_size); - host_ptr = tmp_buffer.data(); - memory_manager.ReadBlockUnsafe(gpu_addr, host_ptr, guest_memory_size); - - if (params.is_tiled) { - ASSERT_MSG(params.block_width == 0, "Block width is defined as {} on texture target {}", - params.block_width, static_cast<u32>(params.target)); - for (u32 level = 0; level < params.num_levels; ++level) { - const std::size_t host_offset{params.GetHostMipmapLevelOffset(level, false)}; - SwizzleFunc(MortonSwizzleMode::MortonToLinear, host_ptr, params, - staging_buffer.data() + host_offset, level); - } - } else { - ASSERT_MSG(params.num_levels == 1, "Linear mipmap loading is not implemented"); - const u32 bpp{params.GetBytesPerPixel()}; - const u32 block_width{params.GetDefaultBlockWidth()}; - const u32 block_height{params.GetDefaultBlockHeight()}; - const u32 width{(params.width + block_width - 1) / block_width}; - const u32 height{(params.height + block_height - 1) / block_height}; - const u32 copy_size{width * bpp}; - if (params.pitch == copy_size) { - std::memcpy(staging_buffer.data(), host_ptr, params.GetHostSizeInBytes(false)); - } else { - const u8* start{host_ptr}; - u8* write_to{staging_buffer.data()}; - for (u32 h = height; h > 0; --h) { - std::memcpy(write_to, start, copy_size); - start += params.pitch; - write_to += copy_size; - } - } - } - - if (!is_converted && params.pixel_format != PixelFormat::S8_UINT_D24_UNORM) { - return; - } - - for (u32 level = params.num_levels; level--;) { - const std::size_t in_host_offset{params.GetHostMipmapLevelOffset(level, false)}; - const std::size_t out_host_offset{params.GetHostMipmapLevelOffset(level, is_converted)}; - u8* const in_buffer = staging_buffer.data() + in_host_offset; - u8* const out_buffer = staging_buffer.data() + out_host_offset; - ConvertFromGuestToHost(in_buffer, out_buffer, params.pixel_format, - params.GetMipWidth(level), params.GetMipHeight(level), - params.GetMipDepth(level), true, true); - } -} - -void SurfaceBaseImpl::FlushBuffer(Tegra::MemoryManager& memory_manager, - StagingCache& staging_cache) { - MICROPROFILE_SCOPE(GPU_Flush_Texture); - auto& staging_buffer = staging_cache.GetBuffer(0); - u8* host_ptr; - - // Use an extra temporal buffer - auto& tmp_buffer = staging_cache.GetBuffer(1); - tmp_buffer.resize(guest_memory_size); - host_ptr = tmp_buffer.data(); - - if (params.target == SurfaceTarget::Texture3D) { - // Special case for 3D texture segments - memory_manager.ReadBlockUnsafe(gpu_addr, host_ptr, guest_memory_size); - } - - if (params.is_tiled) { - ASSERT_MSG(params.block_width == 0, "Block width is defined as {}", params.block_width); - for (u32 level = 0; level < params.num_levels; ++level) { - const std::size_t host_offset{params.GetHostMipmapLevelOffset(level, false)}; - SwizzleFunc(MortonSwizzleMode::LinearToMorton, host_ptr, params, - staging_buffer.data() + host_offset, level); - } - } else if (params.IsBuffer()) { - // Buffers don't have pitch or any fancy layout property. We can just memcpy them to guest - // memory. - std::memcpy(host_ptr, staging_buffer.data(), guest_memory_size); - } else { - ASSERT(params.target == SurfaceTarget::Texture2D); - ASSERT(params.num_levels == 1); - - const u32 bpp{params.GetBytesPerPixel()}; - const u32 copy_size{params.width * bpp}; - if (params.pitch == copy_size) { - std::memcpy(host_ptr, staging_buffer.data(), guest_memory_size); - } else { - u8* start{host_ptr}; - const u8* read_to{staging_buffer.data()}; - for (u32 h = params.height; h > 0; --h) { - std::memcpy(start, read_to, copy_size); - start += params.pitch; - read_to += copy_size; - } - } - } - memory_manager.WriteBlockUnsafe(gpu_addr, host_ptr, guest_memory_size); -} - -} // namespace VideoCommon diff --git a/src/video_core/texture_cache/surface_base.h b/src/video_core/texture_cache/surface_base.h deleted file mode 100644 index 173f2edba..000000000 --- a/src/video_core/texture_cache/surface_base.h +++ /dev/null @@ -1,333 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <optional> -#include <tuple> -#include <unordered_map> -#include <vector> - -#include "common/common_types.h" -#include "video_core/gpu.h" -#include "video_core/morton.h" -#include "video_core/texture_cache/copy_params.h" -#include "video_core/texture_cache/surface_params.h" -#include "video_core/texture_cache/surface_view.h" - -namespace Tegra { -class MemoryManager; -} - -namespace VideoCommon { - -using VideoCore::MortonSwizzleMode; -using VideoCore::Surface::SurfaceTarget; - -enum class MatchStructureResult : u32 { - FullMatch = 0, - SemiMatch = 1, - None = 2, -}; - -enum class MatchTopologyResult : u32 { - FullMatch = 0, - CompressUnmatch = 1, - None = 2, -}; - -class StagingCache { -public: - explicit StagingCache(); - ~StagingCache(); - - std::vector<u8>& GetBuffer(std::size_t index) { - return staging_buffer[index]; - } - - const std::vector<u8>& GetBuffer(std::size_t index) const { - return staging_buffer[index]; - } - - void SetSize(std::size_t size) { - staging_buffer.resize(size); - } - -private: - std::vector<std::vector<u8>> staging_buffer; -}; - -class SurfaceBaseImpl { -public: - void LoadBuffer(Tegra::MemoryManager& memory_manager, StagingCache& staging_cache); - - void FlushBuffer(Tegra::MemoryManager& memory_manager, StagingCache& staging_cache); - - GPUVAddr GetGpuAddr() const { - return gpu_addr; - } - - bool Overlaps(const VAddr start, const VAddr end) const { - return (cpu_addr < end) && (cpu_addr_end > start); - } - - bool IsInside(const GPUVAddr other_start, const GPUVAddr other_end) const { - const GPUVAddr gpu_addr_end = gpu_addr + guest_memory_size; - return gpu_addr <= other_start && other_end <= gpu_addr_end; - } - - // Use only when recycling a surface - void SetGpuAddr(const GPUVAddr new_addr) { - gpu_addr = new_addr; - } - - VAddr GetCpuAddr() const { - return cpu_addr; - } - - VAddr GetCpuAddrEnd() const { - return cpu_addr_end; - } - - void SetCpuAddr(const VAddr new_addr) { - cpu_addr = new_addr; - cpu_addr_end = new_addr + guest_memory_size; - } - - const SurfaceParams& GetSurfaceParams() const { - return params; - } - - std::size_t GetSizeInBytes() const { - return guest_memory_size; - } - - std::size_t GetHostSizeInBytes() const { - return host_memory_size; - } - - std::size_t GetMipmapSize(const u32 level) const { - return mipmap_sizes[level]; - } - - bool IsLinear() const { - return !params.is_tiled; - } - - bool IsConverted() const { - return is_converted; - } - - bool MatchFormat(VideoCore::Surface::PixelFormat pixel_format) const { - return params.pixel_format == pixel_format; - } - - VideoCore::Surface::PixelFormat GetFormat() const { - return params.pixel_format; - } - - bool MatchTarget(VideoCore::Surface::SurfaceTarget target) const { - return params.target == target; - } - - MatchTopologyResult MatchesTopology(const SurfaceParams& rhs) const; - - MatchStructureResult MatchesStructure(const SurfaceParams& rhs) const; - - bool MatchesSubTexture(const SurfaceParams& rhs, const GPUVAddr other_gpu_addr) const { - return std::tie(gpu_addr, params.target, params.num_levels) == - std::tie(other_gpu_addr, rhs.target, rhs.num_levels) && - params.target == SurfaceTarget::Texture2D && params.num_levels == 1; - } - - std::optional<std::pair<u32, u32>> GetLayerMipmap(const GPUVAddr candidate_gpu_addr) const; - - std::vector<CopyParams> BreakDown(const SurfaceParams& in_params) const { - return params.is_layered ? BreakDownLayered(in_params) : BreakDownNonLayered(in_params); - } - -protected: - explicit SurfaceBaseImpl(GPUVAddr gpu_addr, const SurfaceParams& params, - bool is_astc_supported); - ~SurfaceBaseImpl() = default; - - virtual void DecorateSurfaceName() = 0; - - const SurfaceParams params; - std::size_t layer_size; - std::size_t guest_memory_size; - std::size_t host_memory_size; - GPUVAddr gpu_addr{}; - VAddr cpu_addr{}; - VAddr cpu_addr_end{}; - bool is_converted{}; - - std::vector<std::size_t> mipmap_sizes; - std::vector<std::size_t> mipmap_offsets; - -private: - void SwizzleFunc(MortonSwizzleMode mode, u8* memory, const SurfaceParams& params, u8* buffer, - u32 level); - - std::vector<CopyParams> BreakDownLayered(const SurfaceParams& in_params) const; - - std::vector<CopyParams> BreakDownNonLayered(const SurfaceParams& in_params) const; -}; - -template <typename TView> -class SurfaceBase : public SurfaceBaseImpl { -public: - virtual void UploadTexture(const std::vector<u8>& staging_buffer) = 0; - - virtual void DownloadTexture(std::vector<u8>& staging_buffer) = 0; - - void MarkAsModified(bool is_modified_, u64 tick) { - is_modified = is_modified_ || is_target; - modification_tick = tick; - } - - void MarkAsRenderTarget(bool is_target_, u32 index_) { - is_target = is_target_; - index = index_; - } - - void SetMemoryMarked(bool is_memory_marked_) { - is_memory_marked = is_memory_marked_; - } - - bool IsMemoryMarked() const { - return is_memory_marked; - } - - void SetSyncPending(bool is_sync_pending_) { - is_sync_pending = is_sync_pending_; - } - - bool IsSyncPending() const { - return is_sync_pending; - } - - void MarkAsPicked(bool is_picked_) { - is_picked = is_picked_; - } - - bool IsModified() const { - return is_modified; - } - - bool IsProtected() const { - // Only 3D slices are to be protected - return is_target && params.target == SurfaceTarget::Texture3D; - } - - bool IsRenderTarget() const { - return is_target; - } - - u32 GetRenderTarget() const { - return index; - } - - bool IsRegistered() const { - return is_registered; - } - - bool IsPicked() const { - return is_picked; - } - - void MarkAsRegistered(bool is_reg) { - is_registered = is_reg; - } - - u64 GetModificationTick() const { - return modification_tick; - } - - TView EmplaceOverview(const SurfaceParams& overview_params) { - const u32 num_layers{(params.is_layered && !overview_params.is_layered) ? 1 : params.depth}; - return GetView(ViewParams(overview_params.target, 0, num_layers, 0, params.num_levels)); - } - - TView Emplace3DView(u32 slice, u32 depth, u32 base_level, u32 num_levels) { - return GetView(ViewParams(VideoCore::Surface::SurfaceTarget::Texture3D, slice, depth, - base_level, num_levels)); - } - - std::optional<TView> EmplaceIrregularView(const SurfaceParams& view_params, - const GPUVAddr view_addr, - const std::size_t candidate_size, const u32 mipmap, - const u32 layer) { - const auto layer_mipmap{GetLayerMipmap(view_addr + candidate_size)}; - if (!layer_mipmap) { - return {}; - } - const auto [end_layer, end_mipmap] = *layer_mipmap; - if (layer != end_layer) { - if (mipmap == 0 && end_mipmap == 0) { - return GetView(ViewParams(view_params.target, layer, end_layer - layer, 0, 1)); - } - return {}; - } else { - return GetView(ViewParams(view_params.target, layer, 1, mipmap, end_mipmap - mipmap)); - } - } - - std::optional<TView> EmplaceView(const SurfaceParams& view_params, const GPUVAddr view_addr, - const std::size_t candidate_size) { - if (params.target == SurfaceTarget::Texture3D || - view_params.target == SurfaceTarget::Texture3D || - (params.num_levels == 1 && !params.is_layered)) { - return {}; - } - const auto layer_mipmap{GetLayerMipmap(view_addr)}; - if (!layer_mipmap) { - return {}; - } - const auto [layer, mipmap] = *layer_mipmap; - if (GetMipmapSize(mipmap) != candidate_size) { - return EmplaceIrregularView(view_params, view_addr, candidate_size, mipmap, layer); - } - return GetView(ViewParams(view_params.target, layer, 1, mipmap, 1)); - } - - TView GetMainView() const { - return main_view; - } - -protected: - explicit SurfaceBase(const GPUVAddr gpu_addr, const SurfaceParams& params, - bool is_astc_supported) - : SurfaceBaseImpl(gpu_addr, params, is_astc_supported) {} - - ~SurfaceBase() = default; - - virtual TView CreateView(const ViewParams& view_key) = 0; - - TView main_view; - std::unordered_map<ViewParams, TView> views; - -private: - TView GetView(const ViewParams& key) { - const auto [entry, is_cache_miss] = views.try_emplace(key); - auto& view{entry->second}; - if (is_cache_miss) { - view = CreateView(key); - } - return view; - } - - static constexpr u32 NO_RT = 0xFFFFFFFF; - - bool is_modified{}; - bool is_target{}; - bool is_registered{}; - bool is_picked{}; - bool is_memory_marked{}; - bool is_sync_pending{}; - u32 index{NO_RT}; - u64 modification_tick{}; -}; - -} // namespace VideoCommon diff --git a/src/video_core/texture_cache/surface_params.cpp b/src/video_core/texture_cache/surface_params.cpp deleted file mode 100644 index e8515321b..000000000 --- a/src/video_core/texture_cache/surface_params.cpp +++ /dev/null @@ -1,444 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <string> -#include <tuple> - -#include "common/alignment.h" -#include "common/bit_util.h" -#include "core/core.h" -#include "video_core/engines/shader_bytecode.h" -#include "video_core/surface.h" -#include "video_core/texture_cache/format_lookup_table.h" -#include "video_core/texture_cache/surface_params.h" - -namespace VideoCommon { - -using VideoCore::Surface::PixelFormat; -using VideoCore::Surface::PixelFormatFromDepthFormat; -using VideoCore::Surface::PixelFormatFromRenderTargetFormat; -using VideoCore::Surface::SurfaceTarget; -using VideoCore::Surface::SurfaceTargetFromTextureType; -using VideoCore::Surface::SurfaceType; - -namespace { - -SurfaceTarget TextureTypeToSurfaceTarget(Tegra::Shader::TextureType type, bool is_array) { - switch (type) { - case Tegra::Shader::TextureType::Texture1D: - return is_array ? SurfaceTarget::Texture1DArray : SurfaceTarget::Texture1D; - case Tegra::Shader::TextureType::Texture2D: - return is_array ? SurfaceTarget::Texture2DArray : SurfaceTarget::Texture2D; - case Tegra::Shader::TextureType::Texture3D: - ASSERT(!is_array); - return SurfaceTarget::Texture3D; - case Tegra::Shader::TextureType::TextureCube: - return is_array ? SurfaceTarget::TextureCubeArray : SurfaceTarget::TextureCubemap; - default: - UNREACHABLE(); - return SurfaceTarget::Texture2D; - } -} - -SurfaceTarget ImageTypeToSurfaceTarget(Tegra::Shader::ImageType type) { - switch (type) { - case Tegra::Shader::ImageType::Texture1D: - return SurfaceTarget::Texture1D; - case Tegra::Shader::ImageType::TextureBuffer: - return SurfaceTarget::TextureBuffer; - case Tegra::Shader::ImageType::Texture1DArray: - return SurfaceTarget::Texture1DArray; - case Tegra::Shader::ImageType::Texture2D: - return SurfaceTarget::Texture2D; - case Tegra::Shader::ImageType::Texture2DArray: - return SurfaceTarget::Texture2DArray; - case Tegra::Shader::ImageType::Texture3D: - return SurfaceTarget::Texture3D; - default: - UNREACHABLE(); - return SurfaceTarget::Texture2D; - } -} - -constexpr u32 GetMipmapSize(bool uncompressed, u32 mip_size, u32 tile) { - return uncompressed ? mip_size : std::max(1U, (mip_size + tile - 1) / tile); -} - -} // Anonymous namespace - -SurfaceParams SurfaceParams::CreateForTexture(const FormatLookupTable& lookup_table, - const Tegra::Texture::TICEntry& tic, - const VideoCommon::Shader::Sampler& entry) { - SurfaceParams params; - params.is_tiled = tic.IsTiled(); - params.srgb_conversion = tic.IsSrgbConversionEnabled(); - params.block_width = params.is_tiled ? tic.BlockWidth() : 0; - params.block_height = params.is_tiled ? tic.BlockHeight() : 0; - params.block_depth = params.is_tiled ? tic.BlockDepth() : 0; - params.tile_width_spacing = params.is_tiled ? (1 << tic.tile_width_spacing.Value()) : 1; - params.pixel_format = lookup_table.GetPixelFormat( - tic.format, params.srgb_conversion, tic.r_type, tic.g_type, tic.b_type, tic.a_type); - params.type = GetFormatType(params.pixel_format); - if (entry.is_shadow && params.type == SurfaceType::ColorTexture) { - switch (params.pixel_format) { - case PixelFormat::R16_UNORM: - case PixelFormat::R16_FLOAT: - params.pixel_format = PixelFormat::D16_UNORM; - break; - case PixelFormat::R32_FLOAT: - params.pixel_format = PixelFormat::D32_FLOAT; - break; - default: - UNIMPLEMENTED_MSG("Unimplemented shadow convert format: {}", - static_cast<u32>(params.pixel_format)); - } - params.type = GetFormatType(params.pixel_format); - } - // TODO: on 1DBuffer we should use the tic info. - if (tic.IsBuffer()) { - params.target = SurfaceTarget::TextureBuffer; - params.width = tic.Width(); - params.pitch = params.width * params.GetBytesPerPixel(); - params.height = 1; - params.depth = 1; - params.num_levels = 1; - params.emulated_levels = 1; - params.is_layered = false; - } else { - params.target = TextureTypeToSurfaceTarget(entry.type, entry.is_array); - params.width = tic.Width(); - params.height = tic.Height(); - params.depth = tic.Depth(); - params.pitch = params.is_tiled ? 0 : tic.Pitch(); - if (params.target == SurfaceTarget::TextureCubemap || - params.target == SurfaceTarget::TextureCubeArray) { - params.depth *= 6; - } - params.num_levels = tic.max_mip_level + 1; - params.emulated_levels = std::min(params.num_levels, params.MaxPossibleMipmap()); - params.is_layered = params.IsLayered(); - } - return params; -} - -SurfaceParams SurfaceParams::CreateForImage(const FormatLookupTable& lookup_table, - const Tegra::Texture::TICEntry& tic, - const VideoCommon::Shader::Image& entry) { - SurfaceParams params; - params.is_tiled = tic.IsTiled(); - params.srgb_conversion = tic.IsSrgbConversionEnabled(); - params.block_width = params.is_tiled ? tic.BlockWidth() : 0; - params.block_height = params.is_tiled ? tic.BlockHeight() : 0; - params.block_depth = params.is_tiled ? tic.BlockDepth() : 0; - params.tile_width_spacing = params.is_tiled ? (1 << tic.tile_width_spacing.Value()) : 1; - params.pixel_format = lookup_table.GetPixelFormat( - tic.format, params.srgb_conversion, tic.r_type, tic.g_type, tic.b_type, tic.a_type); - params.type = GetFormatType(params.pixel_format); - params.target = ImageTypeToSurfaceTarget(entry.type); - // TODO: on 1DBuffer we should use the tic info. - if (tic.IsBuffer()) { - params.target = SurfaceTarget::TextureBuffer; - params.width = tic.Width(); - params.pitch = params.width * params.GetBytesPerPixel(); - params.height = 1; - params.depth = 1; - params.num_levels = 1; - params.emulated_levels = 1; - params.is_layered = false; - } else { - params.width = tic.Width(); - params.height = tic.Height(); - params.depth = tic.Depth(); - params.pitch = params.is_tiled ? 0 : tic.Pitch(); - if (params.target == SurfaceTarget::TextureCubemap || - params.target == SurfaceTarget::TextureCubeArray) { - params.depth *= 6; - } - params.num_levels = tic.max_mip_level + 1; - params.emulated_levels = std::min(params.num_levels, params.MaxPossibleMipmap()); - params.is_layered = params.IsLayered(); - } - return params; -} - -SurfaceParams SurfaceParams::CreateForDepthBuffer(Tegra::Engines::Maxwell3D& maxwell3d) { - const auto& regs = maxwell3d.regs; - const auto block_depth = std::min(regs.zeta.memory_layout.block_depth.Value(), 5U); - const bool is_layered = regs.zeta_layers > 1 && block_depth == 0; - const auto pixel_format = PixelFormatFromDepthFormat(regs.zeta.format); - return { - .is_tiled = regs.zeta.memory_layout.type == - Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear, - .srgb_conversion = false, - .is_layered = is_layered, - .block_width = std::min(regs.zeta.memory_layout.block_width.Value(), 5U), - .block_height = std::min(regs.zeta.memory_layout.block_height.Value(), 5U), - .block_depth = block_depth, - .tile_width_spacing = 1, - .width = regs.zeta_width, - .height = regs.zeta_height, - .depth = is_layered ? regs.zeta_layers.Value() : 1U, - .pitch = 0, - .num_levels = 1, - .emulated_levels = 1, - .pixel_format = pixel_format, - .type = GetFormatType(pixel_format), - .target = is_layered ? SurfaceTarget::Texture2DArray : SurfaceTarget::Texture2D, - }; -} - -SurfaceParams SurfaceParams::CreateForFramebuffer(Tegra::Engines::Maxwell3D& maxwell3d, - std::size_t index) { - const auto& config{maxwell3d.regs.rt[index]}; - SurfaceParams params; - params.is_tiled = - config.memory_layout.type == Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear; - params.srgb_conversion = config.format == Tegra::RenderTargetFormat::B8G8R8A8_SRGB || - config.format == Tegra::RenderTargetFormat::A8B8G8R8_SRGB; - params.block_width = config.memory_layout.block_width; - params.block_height = config.memory_layout.block_height; - params.block_depth = config.memory_layout.block_depth; - params.tile_width_spacing = 1; - params.pixel_format = PixelFormatFromRenderTargetFormat(config.format); - params.type = GetFormatType(params.pixel_format); - if (params.is_tiled) { - params.pitch = 0; - params.width = config.width; - } else { - const u32 bpp = GetFormatBpp(params.pixel_format) / CHAR_BIT; - params.pitch = config.width; - params.width = params.pitch / bpp; - } - params.height = config.height; - params.num_levels = 1; - params.emulated_levels = 1; - - if (config.memory_layout.is_3d != 0) { - params.depth = config.layers.Value(); - params.is_layered = false; - params.target = SurfaceTarget::Texture3D; - } else if (config.layers > 1) { - params.depth = config.layers.Value(); - params.is_layered = true; - params.target = SurfaceTarget::Texture2DArray; - } else { - params.depth = 1; - params.is_layered = false; - params.target = SurfaceTarget::Texture2D; - } - return params; -} - -SurfaceParams SurfaceParams::CreateForFermiCopySurface( - const Tegra::Engines::Fermi2D::Regs::Surface& config) { - const bool is_tiled = !config.linear; - const auto pixel_format = PixelFormatFromRenderTargetFormat(config.format); - - SurfaceParams params{ - .is_tiled = is_tiled, - .srgb_conversion = config.format == Tegra::RenderTargetFormat::B8G8R8A8_SRGB || - config.format == Tegra::RenderTargetFormat::A8B8G8R8_SRGB, - .block_width = is_tiled ? std::min(config.BlockWidth(), 5U) : 0U, - .block_height = is_tiled ? std::min(config.BlockHeight(), 5U) : 0U, - .block_depth = is_tiled ? std::min(config.BlockDepth(), 5U) : 0U, - .tile_width_spacing = 1, - .width = config.width, - .height = config.height, - .depth = 1, - .pitch = config.pitch, - .num_levels = 1, - .emulated_levels = 1, - .pixel_format = pixel_format, - .type = GetFormatType(pixel_format), - // TODO(Rodrigo): Try to guess texture arrays from parameters - .target = SurfaceTarget::Texture2D, - }; - - params.is_layered = params.IsLayered(); - return params; -} - -VideoCore::Surface::SurfaceTarget SurfaceParams::ExpectedTarget( - const VideoCommon::Shader::Sampler& entry) { - return TextureTypeToSurfaceTarget(entry.type, entry.is_array); -} - -VideoCore::Surface::SurfaceTarget SurfaceParams::ExpectedTarget( - const VideoCommon::Shader::Image& entry) { - return ImageTypeToSurfaceTarget(entry.type); -} - -bool SurfaceParams::IsLayered() const { - switch (target) { - case SurfaceTarget::Texture1DArray: - case SurfaceTarget::Texture2DArray: - case SurfaceTarget::TextureCubemap: - case SurfaceTarget::TextureCubeArray: - return true; - default: - return false; - } -} - -// Auto block resizing algorithm from: -// https://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/nouveau/nv50/nv50_miptree.c -u32 SurfaceParams::GetMipBlockHeight(u32 level) const { - if (level == 0) { - return this->block_height; - } - - const u32 height_new{GetMipHeight(level)}; - const u32 default_block_height{GetDefaultBlockHeight()}; - const u32 blocks_in_y{(height_new + default_block_height - 1) / default_block_height}; - const u32 block_height_new = Common::Log2Ceil32(blocks_in_y); - return std::clamp(block_height_new, 3U, 7U) - 3U; -} - -u32 SurfaceParams::GetMipBlockDepth(u32 level) const { - if (level == 0) { - return this->block_depth; - } - if (is_layered) { - return 0; - } - - const u32 depth_new{GetMipDepth(level)}; - const u32 block_depth_new = Common::Log2Ceil32(depth_new); - if (block_depth_new > 4) { - return 5 - (GetMipBlockHeight(level) >= 2); - } - return block_depth_new; -} - -std::size_t SurfaceParams::GetGuestMipmapLevelOffset(u32 level) const { - std::size_t offset = 0; - for (u32 i = 0; i < level; i++) { - offset += GetInnerMipmapMemorySize(i, false, false); - } - return offset; -} - -std::size_t SurfaceParams::GetHostMipmapLevelOffset(u32 level, bool is_converted) const { - std::size_t offset = 0; - if (is_converted) { - for (u32 i = 0; i < level; ++i) { - offset += GetConvertedMipmapSize(i) * GetNumLayers(); - } - } else { - for (u32 i = 0; i < level; ++i) { - offset += GetInnerMipmapMemorySize(i, true, false) * GetNumLayers(); - } - } - return offset; -} - -std::size_t SurfaceParams::GetConvertedMipmapSize(u32 level) const { - constexpr std::size_t rgba8_bpp = 4ULL; - const std::size_t mip_width = GetMipWidth(level); - const std::size_t mip_height = GetMipHeight(level); - const std::size_t mip_depth = is_layered ? 1 : GetMipDepth(level); - return mip_width * mip_height * mip_depth * rgba8_bpp; -} - -std::size_t SurfaceParams::GetLayerSize(bool as_host_size, bool uncompressed) const { - std::size_t size = 0; - for (u32 level = 0; level < num_levels; ++level) { - size += GetInnerMipmapMemorySize(level, as_host_size, uncompressed); - } - if (is_tiled && is_layered) { - return Common::AlignBits(size, Tegra::Texture::GOB_SIZE_SHIFT + block_height + block_depth); - } - return size; -} - -std::size_t SurfaceParams::GetInnerMipmapMemorySize(u32 level, bool as_host_size, - bool uncompressed) const { - const u32 width{GetMipmapSize(uncompressed, GetMipWidth(level), GetDefaultBlockWidth())}; - const u32 height{GetMipmapSize(uncompressed, GetMipHeight(level), GetDefaultBlockHeight())}; - const u32 depth{is_layered ? 1U : GetMipDepth(level)}; - if (is_tiled) { - return Tegra::Texture::CalculateSize(!as_host_size, GetBytesPerPixel(), width, height, - depth, GetMipBlockHeight(level), - GetMipBlockDepth(level)); - } else if (as_host_size || IsBuffer()) { - return GetBytesPerPixel() * width * height * depth; - } else { - // Linear Texture Case - return pitch * height * depth; - } -} - -bool SurfaceParams::operator==(const SurfaceParams& rhs) const { - return std::tie(is_tiled, block_width, block_height, block_depth, tile_width_spacing, width, - height, depth, pitch, num_levels, pixel_format, type, target) == - std::tie(rhs.is_tiled, rhs.block_width, rhs.block_height, rhs.block_depth, - rhs.tile_width_spacing, rhs.width, rhs.height, rhs.depth, rhs.pitch, - rhs.num_levels, rhs.pixel_format, rhs.type, rhs.target); -} - -std::string SurfaceParams::TargetName() const { - switch (target) { - case SurfaceTarget::Texture1D: - return "1D"; - case SurfaceTarget::TextureBuffer: - return "TexBuffer"; - case SurfaceTarget::Texture2D: - return "2D"; - case SurfaceTarget::Texture3D: - return "3D"; - case SurfaceTarget::Texture1DArray: - return "1DArray"; - case SurfaceTarget::Texture2DArray: - return "2DArray"; - case SurfaceTarget::TextureCubemap: - return "Cube"; - case SurfaceTarget::TextureCubeArray: - return "CubeArray"; - default: - LOG_CRITICAL(HW_GPU, "Unimplemented surface_target={}", static_cast<u32>(target)); - UNREACHABLE(); - return fmt::format("TUK({})", static_cast<u32>(target)); - } -} - -u32 SurfaceParams::GetBlockSize() const { - const u32 x = 64U << block_width; - const u32 y = 8U << block_height; - const u32 z = 1U << block_depth; - return x * y * z; -} - -std::pair<u32, u32> SurfaceParams::GetBlockXY() const { - const u32 x_pixels = 64U / GetBytesPerPixel(); - const u32 x = x_pixels << block_width; - const u32 y = 8U << block_height; - return {x, y}; -} - -std::tuple<u32, u32, u32> SurfaceParams::GetBlockOffsetXYZ(u32 offset) const { - const auto div_ceil = [](const u32 x, const u32 y) { return ((x + y - 1) / y); }; - const u32 block_size = GetBlockSize(); - const u32 block_index = offset / block_size; - const u32 gob_offset = offset % block_size; - const u32 gob_index = gob_offset / static_cast<u32>(Tegra::Texture::GOB_SIZE); - const u32 x_gob_pixels = 64U / GetBytesPerPixel(); - const u32 x_block_pixels = x_gob_pixels << block_width; - const u32 y_block_pixels = 8U << block_height; - const u32 z_block_pixels = 1U << block_depth; - const u32 x_blocks = div_ceil(width, x_block_pixels); - const u32 y_blocks = div_ceil(height, y_block_pixels); - const u32 z_blocks = div_ceil(depth, z_block_pixels); - const u32 base_x = block_index % x_blocks; - const u32 base_y = (block_index / x_blocks) % y_blocks; - const u32 base_z = (block_index / (x_blocks * y_blocks)) % z_blocks; - u32 x = base_x * x_block_pixels; - u32 y = base_y * y_block_pixels; - u32 z = base_z * z_block_pixels; - z += gob_index >> block_height; - y += (gob_index * 8U) % y_block_pixels; - return {x, y, z}; -} - -} // namespace VideoCommon diff --git a/src/video_core/texture_cache/surface_params.h b/src/video_core/texture_cache/surface_params.h deleted file mode 100644 index 4466c3c34..000000000 --- a/src/video_core/texture_cache/surface_params.h +++ /dev/null @@ -1,294 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <utility> - -#include "common/alignment.h" -#include "common/bit_util.h" -#include "common/cityhash.h" -#include "common/common_types.h" -#include "video_core/engines/fermi_2d.h" -#include "video_core/engines/maxwell_3d.h" -#include "video_core/shader/shader_ir.h" -#include "video_core/surface.h" -#include "video_core/textures/decoders.h" - -namespace VideoCommon { - -class FormatLookupTable; - -class SurfaceParams { -public: - /// Creates SurfaceCachedParams from a texture configuration. - static SurfaceParams CreateForTexture(const FormatLookupTable& lookup_table, - const Tegra::Texture::TICEntry& tic, - const VideoCommon::Shader::Sampler& entry); - - /// Creates SurfaceCachedParams from an image configuration. - static SurfaceParams CreateForImage(const FormatLookupTable& lookup_table, - const Tegra::Texture::TICEntry& tic, - const VideoCommon::Shader::Image& entry); - - /// Creates SurfaceCachedParams for a depth buffer configuration. - static SurfaceParams CreateForDepthBuffer(Tegra::Engines::Maxwell3D& maxwell3d); - - /// Creates SurfaceCachedParams from a framebuffer configuration. - static SurfaceParams CreateForFramebuffer(Tegra::Engines::Maxwell3D& maxwell3d, - std::size_t index); - - /// Creates SurfaceCachedParams from a Fermi2D surface configuration. - static SurfaceParams CreateForFermiCopySurface( - const Tegra::Engines::Fermi2D::Regs::Surface& config); - - /// Obtains the texture target from a shader's sampler entry. - static VideoCore::Surface::SurfaceTarget ExpectedTarget( - const VideoCommon::Shader::Sampler& entry); - - /// Obtains the texture target from a shader's sampler entry. - static VideoCore::Surface::SurfaceTarget ExpectedTarget( - const VideoCommon::Shader::Image& entry); - - std::size_t Hash() const { - return static_cast<std::size_t>( - Common::CityHash64(reinterpret_cast<const char*>(this), sizeof(*this))); - } - - bool operator==(const SurfaceParams& rhs) const; - - bool operator!=(const SurfaceParams& rhs) const { - return !operator==(rhs); - } - - std::size_t GetGuestSizeInBytes() const { - return GetInnerMemorySize(false, false, false); - } - - std::size_t GetHostSizeInBytes(bool is_converted) const { - if (!is_converted) { - return GetInnerMemorySize(true, false, false); - } - // ASTC is uncompressed in software, in emulated as RGBA8 - std::size_t host_size_in_bytes = 0; - for (u32 level = 0; level < num_levels; ++level) { - host_size_in_bytes += GetConvertedMipmapSize(level) * GetNumLayers(); - } - return host_size_in_bytes; - } - - u32 GetBlockAlignedWidth() const { - return Common::AlignUp(width, 64 / GetBytesPerPixel()); - } - - /// Returns the width of a given mipmap level. - u32 GetMipWidth(u32 level) const { - return std::max(1U, width >> level); - } - - /// Returns the height of a given mipmap level. - u32 GetMipHeight(u32 level) const { - return std::max(1U, height >> level); - } - - /// Returns the depth of a given mipmap level. - u32 GetMipDepth(u32 level) const { - return is_layered ? depth : std::max(1U, depth >> level); - } - - /// Returns the block height of a given mipmap level. - u32 GetMipBlockHeight(u32 level) const; - - /// Returns the block depth of a given mipmap level. - u32 GetMipBlockDepth(u32 level) const; - - /// Returns the best possible row/pitch alignment for the surface. - u32 GetRowAlignment(u32 level, bool is_converted) const { - const u32 bpp = is_converted ? 4 : GetBytesPerPixel(); - return 1U << Common::CountTrailingZeroes32(GetMipWidth(level) * bpp); - } - - /// Returns the offset in bytes in guest memory of a given mipmap level. - std::size_t GetGuestMipmapLevelOffset(u32 level) const; - - /// Returns the offset in bytes in host memory (linear) of a given mipmap level. - std::size_t GetHostMipmapLevelOffset(u32 level, bool is_converted) const; - - /// Returns the size in bytes in guest memory of a given mipmap level. - std::size_t GetGuestMipmapSize(u32 level) const { - return GetInnerMipmapMemorySize(level, false, false); - } - - /// Returns the size in bytes in host memory (linear) of a given mipmap level. - std::size_t GetHostMipmapSize(u32 level) const { - return GetInnerMipmapMemorySize(level, true, false) * GetNumLayers(); - } - - std::size_t GetConvertedMipmapSize(u32 level) const; - - /// Get this texture Tegra Block size in guest memory layout - u32 GetBlockSize() const; - - /// Get X, Y coordinates max sizes of a single block. - std::pair<u32, u32> GetBlockXY() const; - - /// Get the offset in x, y, z coordinates from a memory offset - std::tuple<u32, u32, u32> GetBlockOffsetXYZ(u32 offset) const; - - /// Returns the size of a layer in bytes in guest memory. - std::size_t GetGuestLayerSize() const { - return GetLayerSize(false, false); - } - - /// Returns the size of a layer in bytes in host memory for a given mipmap level. - std::size_t GetHostLayerSize(u32 level) const { - ASSERT(target != VideoCore::Surface::SurfaceTarget::Texture3D); - return GetInnerMipmapMemorySize(level, true, false); - } - - /// Returns the max possible mipmap that the texture can have in host gpu - u32 MaxPossibleMipmap() const { - const u32 max_mipmap_w = Common::Log2Ceil32(width) + 1U; - const u32 max_mipmap_h = Common::Log2Ceil32(height) + 1U; - const u32 max_mipmap = std::max(max_mipmap_w, max_mipmap_h); - if (target != VideoCore::Surface::SurfaceTarget::Texture3D) - return max_mipmap; - return std::max(max_mipmap, Common::Log2Ceil32(depth) + 1U); - } - - /// Returns if the guest surface is a compressed surface. - bool IsCompressed() const { - return GetDefaultBlockHeight() > 1 || GetDefaultBlockWidth() > 1; - } - - /// Returns the default block width. - u32 GetDefaultBlockWidth() const { - return VideoCore::Surface::GetDefaultBlockWidth(pixel_format); - } - - /// Returns the default block height. - u32 GetDefaultBlockHeight() const { - return VideoCore::Surface::GetDefaultBlockHeight(pixel_format); - } - - /// Returns the bits per pixel. - u32 GetBitsPerPixel() const { - return VideoCore::Surface::GetFormatBpp(pixel_format); - } - - /// Returns the bytes per pixel. - u32 GetBytesPerPixel() const { - return VideoCore::Surface::GetBytesPerPixel(pixel_format); - } - - /// Returns true if the pixel format is a depth and/or stencil format. - bool IsPixelFormatZeta() const { - return pixel_format >= VideoCore::Surface::PixelFormat::MaxColorFormat && - pixel_format < VideoCore::Surface::PixelFormat::MaxDepthStencilFormat; - } - - /// Returns is the surface is a TextureBuffer type of surface. - bool IsBuffer() const { - return target == VideoCore::Surface::SurfaceTarget::TextureBuffer; - } - - /// Returns the number of layers in the surface. - std::size_t GetNumLayers() const { - return is_layered ? depth : 1; - } - - /// Returns the debug name of the texture for use in graphic debuggers. - std::string TargetName() const; - - // Helper used for out of class size calculations - static std::size_t AlignLayered(const std::size_t out_size, const u32 block_height, - const u32 block_depth) { - return Common::AlignBits(out_size, - Tegra::Texture::GOB_SIZE_SHIFT + block_height + block_depth); - } - - /// Converts a width from a type of surface into another. This helps represent the - /// equivalent value between compressed/non-compressed textures. - static u32 ConvertWidth(u32 width, VideoCore::Surface::PixelFormat pixel_format_from, - VideoCore::Surface::PixelFormat pixel_format_to) { - const u32 bw1 = VideoCore::Surface::GetDefaultBlockWidth(pixel_format_from); - const u32 bw2 = VideoCore::Surface::GetDefaultBlockWidth(pixel_format_to); - return (width * bw2 + bw1 - 1) / bw1; - } - - /// Converts a height from a type of surface into another. This helps represent the - /// equivalent value between compressed/non-compressed textures. - static u32 ConvertHeight(u32 height, VideoCore::Surface::PixelFormat pixel_format_from, - VideoCore::Surface::PixelFormat pixel_format_to) { - const u32 bh1 = VideoCore::Surface::GetDefaultBlockHeight(pixel_format_from); - const u32 bh2 = VideoCore::Surface::GetDefaultBlockHeight(pixel_format_to); - return (height * bh2 + bh1 - 1) / bh1; - } - - // Finds the maximun possible width between 2 2D layers of different formats - static u32 IntersectWidth(const SurfaceParams& src_params, const SurfaceParams& dst_params, - const u32 src_level, const u32 dst_level) { - const u32 bw1 = src_params.GetDefaultBlockWidth(); - const u32 bw2 = dst_params.GetDefaultBlockWidth(); - const u32 t_src_width = (src_params.GetMipWidth(src_level) * bw2 + bw1 - 1) / bw1; - const u32 t_dst_width = (dst_params.GetMipWidth(dst_level) * bw1 + bw2 - 1) / bw2; - return std::min(t_src_width, t_dst_width); - } - - // Finds the maximun possible height between 2 2D layers of different formats - static u32 IntersectHeight(const SurfaceParams& src_params, const SurfaceParams& dst_params, - const u32 src_level, const u32 dst_level) { - const u32 bh1 = src_params.GetDefaultBlockHeight(); - const u32 bh2 = dst_params.GetDefaultBlockHeight(); - const u32 t_src_height = (src_params.GetMipHeight(src_level) * bh2 + bh1 - 1) / bh1; - const u32 t_dst_height = (dst_params.GetMipHeight(dst_level) * bh1 + bh2 - 1) / bh2; - return std::min(t_src_height, t_dst_height); - } - - bool is_tiled; - bool srgb_conversion; - bool is_layered; - u32 block_width; - u32 block_height; - u32 block_depth; - u32 tile_width_spacing; - u32 width; - u32 height; - u32 depth; - u32 pitch; - u32 num_levels; - u32 emulated_levels; - VideoCore::Surface::PixelFormat pixel_format; - VideoCore::Surface::SurfaceType type; - VideoCore::Surface::SurfaceTarget target; - -private: - /// Returns the size of a given mipmap level inside a layer. - std::size_t GetInnerMipmapMemorySize(u32 level, bool as_host_size, bool uncompressed) const; - - /// Returns the size of all mipmap levels and aligns as needed. - std::size_t GetInnerMemorySize(bool as_host_size, bool layer_only, bool uncompressed) const { - return GetLayerSize(as_host_size, uncompressed) * - (layer_only ? 1U : (is_layered ? depth : 1U)); - } - - /// Returns the size of a layer - std::size_t GetLayerSize(bool as_host_size, bool uncompressed) const; - - /// Returns true if these parameters are from a layered surface. - bool IsLayered() const; -}; - -} // namespace VideoCommon - -namespace std { - -template <> -struct hash<VideoCommon::SurfaceParams> { - std::size_t operator()(const VideoCommon::SurfaceParams& k) const noexcept { - return k.Hash(); - } -}; - -} // namespace std diff --git a/src/video_core/texture_cache/surface_view.cpp b/src/video_core/texture_cache/surface_view.cpp deleted file mode 100644 index 6b5f5984b..000000000 --- a/src/video_core/texture_cache/surface_view.cpp +++ /dev/null @@ -1,27 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <tuple> - -#include "common/common_types.h" -#include "video_core/texture_cache/surface_view.h" - -namespace VideoCommon { - -std::size_t ViewParams::Hash() const { - return static_cast<std::size_t>(base_layer) ^ (static_cast<std::size_t>(num_layers) << 16) ^ - (static_cast<std::size_t>(base_level) << 24) ^ - (static_cast<std::size_t>(num_levels) << 32) ^ (static_cast<std::size_t>(target) << 36); -} - -bool ViewParams::operator==(const ViewParams& rhs) const { - return std::tie(base_layer, num_layers, base_level, num_levels, target) == - std::tie(rhs.base_layer, rhs.num_layers, rhs.base_level, rhs.num_levels, rhs.target); -} - -bool ViewParams::operator!=(const ViewParams& rhs) const { - return !operator==(rhs); -} - -} // namespace VideoCommon diff --git a/src/video_core/texture_cache/surface_view.h b/src/video_core/texture_cache/surface_view.h deleted file mode 100644 index 90a8bb0ae..000000000 --- a/src/video_core/texture_cache/surface_view.h +++ /dev/null @@ -1,68 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <functional> - -#include "common/common_types.h" -#include "video_core/surface.h" -#include "video_core/texture_cache/surface_params.h" - -namespace VideoCommon { - -struct ViewParams { - constexpr explicit ViewParams(VideoCore::Surface::SurfaceTarget target, u32 base_layer, - u32 num_layers, u32 base_level, u32 num_levels) - : target{target}, base_layer{base_layer}, num_layers{num_layers}, base_level{base_level}, - num_levels{num_levels} {} - - std::size_t Hash() const; - - bool operator==(const ViewParams& rhs) const; - bool operator!=(const ViewParams& rhs) const; - - bool IsLayered() const { - switch (target) { - case VideoCore::Surface::SurfaceTarget::Texture1DArray: - case VideoCore::Surface::SurfaceTarget::Texture2DArray: - case VideoCore::Surface::SurfaceTarget::TextureCubemap: - case VideoCore::Surface::SurfaceTarget::TextureCubeArray: - return true; - default: - return false; - } - } - - VideoCore::Surface::SurfaceTarget target{}; - u32 base_layer{}; - u32 num_layers{}; - u32 base_level{}; - u32 num_levels{}; -}; - -class ViewBase { -public: - constexpr explicit ViewBase(const ViewParams& params) : params{params} {} - - constexpr const ViewParams& GetViewParams() const { - return params; - } - -protected: - ViewParams params; -}; - -} // namespace VideoCommon - -namespace std { - -template <> -struct hash<VideoCommon::ViewParams> { - std::size_t operator()(const VideoCommon::ViewParams& k) const noexcept { - return k.Hash(); - } -}; - -} // namespace std diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index ea835c59f..d1080300f 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -6,1299 +6,1454 @@ #include <algorithm> #include <array> -#include <list> +#include <bit> #include <memory> #include <mutex> -#include <set> -#include <tuple> +#include <optional> +#include <span> +#include <type_traits> #include <unordered_map> +#include <utility> #include <vector> #include <boost/container/small_vector.hpp> -#include <boost/icl/interval_map.hpp> -#include <boost/range/iterator_range.hpp> -#include "common/assert.h" +#include "common/alignment.h" +#include "common/common_funcs.h" #include "common/common_types.h" -#include "common/math_util.h" -#include "core/core.h" -#include "core/memory.h" -#include "core/settings.h" +#include "common/logging/log.h" #include "video_core/compatible_formats.h" +#include "video_core/delayed_destruction_ring.h" #include "video_core/dirty_flags.h" #include "video_core/engines/fermi_2d.h" +#include "video_core/engines/kepler_compute.h" #include "video_core/engines/maxwell_3d.h" -#include "video_core/gpu.h" #include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" #include "video_core/surface.h" -#include "video_core/texture_cache/copy_params.h" +#include "video_core/texture_cache/descriptor_table.h" #include "video_core/texture_cache/format_lookup_table.h" -#include "video_core/texture_cache/surface_base.h" -#include "video_core/texture_cache/surface_params.h" -#include "video_core/texture_cache/surface_view.h" - -namespace Tegra::Texture { -struct FullTextureInfo; -} - -namespace VideoCore { -class RasterizerInterface; -} +#include "video_core/texture_cache/formatter.h" +#include "video_core/texture_cache/image_base.h" +#include "video_core/texture_cache/image_info.h" +#include "video_core/texture_cache/image_view_base.h" +#include "video_core/texture_cache/image_view_info.h" +#include "video_core/texture_cache/render_targets.h" +#include "video_core/texture_cache/samples_helper.h" +#include "video_core/texture_cache/slot_vector.h" +#include "video_core/texture_cache/types.h" +#include "video_core/texture_cache/util.h" +#include "video_core/textures/texture.h" namespace VideoCommon { -using VideoCore::Surface::FormatCompatibility; +using Tegra::Texture::SwizzleSource; +using Tegra::Texture::TextureType; +using Tegra::Texture::TICEntry; +using Tegra::Texture::TSCEntry; +using VideoCore::Surface::GetFormatType; +using VideoCore::Surface::IsCopyCompatible; using VideoCore::Surface::PixelFormat; -using VideoCore::Surface::SurfaceTarget; -using RenderTargetConfig = Tegra::Engines::Maxwell3D::Regs::RenderTargetConfig; +using VideoCore::Surface::PixelFormatFromDepthFormat; +using VideoCore::Surface::PixelFormatFromRenderTargetFormat; +using VideoCore::Surface::SurfaceType; -template <typename TSurface, typename TView> +template <class P> class TextureCache { - using VectorSurface = boost::container::small_vector<TSurface, 1>; + /// Address shift for caching images into a hash table + static constexpr u64 PAGE_BITS = 20; + + /// Enables debugging features to the texture cache + static constexpr bool ENABLE_VALIDATION = P::ENABLE_VALIDATION; + /// Implement blits as copies between framebuffers + static constexpr bool FRAMEBUFFER_BLITS = P::FRAMEBUFFER_BLITS; + /// True when some copies have to be emulated + static constexpr bool HAS_EMULATED_COPIES = P::HAS_EMULATED_COPIES; + + /// Image view ID for null descriptors + static constexpr ImageViewId NULL_IMAGE_VIEW_ID{0}; + /// Sampler ID for bugged sampler ids + static constexpr SamplerId NULL_SAMPLER_ID{0}; + + using Runtime = typename P::Runtime; + using Image = typename P::Image; + using ImageAlloc = typename P::ImageAlloc; + using ImageView = typename P::ImageView; + using Sampler = typename P::Sampler; + using Framebuffer = typename P::Framebuffer; + + struct BlitImages { + ImageId dst_id; + ImageId src_id; + PixelFormat dst_format; + PixelFormat src_format; + }; + + template <typename T> + struct IdentityHash { + [[nodiscard]] size_t operator()(T value) const noexcept { + return static_cast<size_t>(value); + } + }; public: - void InvalidateRegion(VAddr addr, std::size_t size) { - std::lock_guard lock{mutex}; + explicit TextureCache(Runtime&, VideoCore::RasterizerInterface&, Tegra::Engines::Maxwell3D&, + Tegra::Engines::KeplerCompute&, Tegra::MemoryManager&); - for (const auto& surface : GetSurfacesInRegion(addr, size)) { - Unregister(surface); - } - } + /// Notify the cache that a new frame has been queued + void TickFrame(); - void OnCPUWrite(VAddr addr, std::size_t size) { - std::lock_guard lock{mutex}; + /// Return an unique mutually exclusive lock for the cache + [[nodiscard]] std::unique_lock<std::mutex> AcquireLock(); - for (const auto& surface : GetSurfacesInRegion(addr, size)) { - if (surface->IsMemoryMarked()) { - UnmarkMemory(surface); - surface->SetSyncPending(true); - marked_for_unregister.emplace_back(surface); - } - } - } + /// Return a constant reference to the given image view id + [[nodiscard]] const ImageView& GetImageView(ImageViewId id) const noexcept; - void SyncGuestHost() { - std::lock_guard lock{mutex}; + /// Return a reference to the given image view id + [[nodiscard]] ImageView& GetImageView(ImageViewId id) noexcept; - for (const auto& surface : marked_for_unregister) { - if (surface->IsRegistered()) { - surface->SetSyncPending(false); - Unregister(surface); - } - } - marked_for_unregister.clear(); - } + /// Fill image_view_ids with the graphics images in indices + void FillGraphicsImageViews(std::span<const u32> indices, + std::span<ImageViewId> image_view_ids); - /** - * Guarantees that rendertargets don't unregister themselves if the - * collide. Protection is currently only done on 3D slices. - */ - void GuardRenderTargets(bool new_guard) { - guard_render_targets = new_guard; - } + /// Fill image_view_ids with the compute images in indices + void FillComputeImageViews(std::span<const u32> indices, std::span<ImageViewId> image_view_ids); - void GuardSamplers(bool new_guard) { - guard_samplers = new_guard; - } + /// Get the sampler from the graphics descriptor table in the specified index + Sampler* GetGraphicsSampler(u32 index); - void FlushRegion(VAddr addr, std::size_t size) { - std::lock_guard lock{mutex}; + /// Get the sampler from the compute descriptor table in the specified index + Sampler* GetComputeSampler(u32 index); - auto surfaces = GetSurfacesInRegion(addr, size); - if (surfaces.empty()) { - return; - } - std::sort(surfaces.begin(), surfaces.end(), [](const TSurface& a, const TSurface& b) { - return a->GetModificationTick() < b->GetModificationTick(); - }); - for (const auto& surface : surfaces) { - mutex.unlock(); - FlushSurface(surface); - mutex.lock(); - } - } + /// Refresh the state for graphics image view and sampler descriptors + void SynchronizeGraphicsDescriptors(); - bool MustFlushRegion(VAddr addr, std::size_t size) { - std::lock_guard lock{mutex}; + /// Refresh the state for compute image view and sampler descriptors + void SynchronizeComputeDescriptors(); - const auto surfaces = GetSurfacesInRegion(addr, size); - return std::any_of(surfaces.cbegin(), surfaces.cend(), - [](const TSurface& surface) { return surface->IsModified(); }); - } + /// Update bound render targets and upload memory if necessary + /// @param is_clear True when the render targets are being used for clears + void UpdateRenderTargets(bool is_clear); - TView GetTextureSurface(const Tegra::Texture::TICEntry& tic, - const VideoCommon::Shader::Sampler& entry) { - std::lock_guard lock{mutex}; - const auto gpu_addr{tic.Address()}; - if (!gpu_addr) { - return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); - } + /// Find a framebuffer with the currently bound render targets + /// UpdateRenderTargets should be called before this + Framebuffer* GetFramebuffer(); - const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); - if (!cpu_addr) { - return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); - } + /// Mark images in a range as modified from the CPU + void WriteMemory(VAddr cpu_addr, size_t size); - if (!IsTypeCompatible(tic.texture_type, entry)) { - return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); - } + /// Download contents of host images to guest memory in a region + void DownloadMemory(VAddr cpu_addr, size_t size); - const auto params{SurfaceParams::CreateForTexture(format_lookup_table, tic, entry)}; - const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, true, false); - if (guard_samplers) { - sampled_textures.push_back(surface); - } - return view; - } + /// Remove images in a region + void UnmapMemory(VAddr cpu_addr, size_t size); - TView GetImageSurface(const Tegra::Texture::TICEntry& tic, - const VideoCommon::Shader::Image& entry) { - std::lock_guard lock{mutex}; - const auto gpu_addr{tic.Address()}; - if (!gpu_addr) { - return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); - } - const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); - if (!cpu_addr) { - return GetNullSurface(SurfaceParams::ExpectedTarget(entry)); - } - const auto params{SurfaceParams::CreateForImage(format_lookup_table, tic, entry)}; - const auto [surface, view] = GetSurface(gpu_addr, *cpu_addr, params, true, false); - if (guard_samplers) { - sampled_textures.push_back(surface); - } - return view; - } + /// Blit an image with the given parameters + void BlitImage(const Tegra::Engines::Fermi2D::Surface& dst, + const Tegra::Engines::Fermi2D::Surface& src, + const Tegra::Engines::Fermi2D::Config& copy); - bool TextureBarrier() { - const bool any_rt = - std::any_of(sampled_textures.begin(), sampled_textures.end(), - [](const auto& surface) { return surface->IsRenderTarget(); }); - sampled_textures.clear(); - return any_rt; - } + /// Invalidate the contents of the color buffer index + /// These contents become unspecified, the cache can assume aggressive optimizations. + void InvalidateColorBuffer(size_t index); - TView GetDepthBufferSurface(bool preserve_contents) { - std::lock_guard lock{mutex}; - auto& dirty = maxwell3d.dirty; - if (!dirty.flags[VideoCommon::Dirty::ZetaBuffer]) { - return depth_buffer.view; - } - dirty.flags[VideoCommon::Dirty::ZetaBuffer] = false; + /// Invalidate the contents of the depth buffer + /// These contents become unspecified, the cache can assume aggressive optimizations. + void InvalidateDepthBuffer(); - const auto& regs{maxwell3d.regs}; - const auto gpu_addr{regs.zeta.Address()}; - if (!gpu_addr || !regs.zeta_enable) { - SetEmptyDepthBuffer(); - return {}; - } - const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); - if (!cpu_addr) { - SetEmptyDepthBuffer(); - return {}; - } - const auto depth_params{SurfaceParams::CreateForDepthBuffer(maxwell3d)}; - auto surface_view = GetSurface(gpu_addr, *cpu_addr, depth_params, preserve_contents, true); - if (depth_buffer.target) - depth_buffer.target->MarkAsRenderTarget(false, NO_RT); - depth_buffer.target = surface_view.first; - depth_buffer.view = surface_view.second; - if (depth_buffer.target) - depth_buffer.target->MarkAsRenderTarget(true, DEPTH_RT); - return surface_view.second; - } - - TView GetColorBufferSurface(std::size_t index, bool preserve_contents) { - std::lock_guard lock{mutex}; - ASSERT(index < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets); - if (!maxwell3d.dirty.flags[VideoCommon::Dirty::ColorBuffer0 + index]) { - return render_targets[index].view; - } - maxwell3d.dirty.flags[VideoCommon::Dirty::ColorBuffer0 + index] = false; + /// Try to find a cached image view in the given CPU address + [[nodiscard]] ImageView* TryFindFramebufferImageView(VAddr cpu_addr); - const auto& regs{maxwell3d.regs}; - if (index >= regs.rt_control.count || regs.rt[index].Address() == 0 || - regs.rt[index].format == Tegra::RenderTargetFormat::NONE) { - SetEmptyColorBuffer(index); - return {}; - } + /// Return true when there are uncommitted images to be downloaded + [[nodiscard]] bool HasUncommittedFlushes() const noexcept; - const auto& config{regs.rt[index]}; - const auto gpu_addr{config.Address()}; - if (!gpu_addr) { - SetEmptyColorBuffer(index); - return {}; - } + /// Return true when the caller should wait for async downloads + [[nodiscard]] bool ShouldWaitAsyncFlushes() const noexcept; - const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); - if (!cpu_addr) { - SetEmptyColorBuffer(index); - return {}; - } + /// Commit asynchronous downloads + void CommitAsyncFlushes(); + + /// Pop asynchronous downloads + void PopAsyncFlushes(); + + /// Return true when a CPU region is modified from the GPU + [[nodiscard]] bool IsRegionGpuModified(VAddr addr, size_t size); - auto surface_view = - GetSurface(gpu_addr, *cpu_addr, SurfaceParams::CreateForFramebuffer(maxwell3d, index), - preserve_contents, true); - if (render_targets[index].target) { - auto& surface = render_targets[index].target; - surface->MarkAsRenderTarget(false, NO_RT); - const auto& cr_params = surface->GetSurfaceParams(); - if (!cr_params.is_tiled && Settings::values.use_asynchronous_gpu_emulation.GetValue()) { - AsyncFlushSurface(surface); +private: + /// Iterate over all page indices in a range + template <typename Func> + static void ForEachPage(VAddr addr, size_t size, Func&& func) { + static constexpr bool RETURNS_BOOL = std::is_same_v<std::invoke_result<Func, u64>, bool>; + const u64 page_end = (addr + size - 1) >> PAGE_BITS; + for (u64 page = addr >> PAGE_BITS; page <= page_end; ++page) { + if constexpr (RETURNS_BOOL) { + if (func(page)) { + break; + } + } else { + func(page); } } - render_targets[index].target = surface_view.first; - render_targets[index].view = surface_view.second; - if (render_targets[index].target) - render_targets[index].target->MarkAsRenderTarget(true, static_cast<u32>(index)); - return surface_view.second; } - void MarkColorBufferInUse(std::size_t index) { - if (auto& render_target = render_targets[index].target) { - render_target->MarkAsModified(true, Tick()); - } - } + /// Fills image_view_ids in the image views in indices + void FillImageViews(DescriptorTable<TICEntry>& table, + std::span<ImageViewId> cached_image_view_ids, std::span<const u32> indices, + std::span<ImageViewId> image_view_ids); - void MarkDepthBufferInUse() { - if (depth_buffer.target) { - depth_buffer.target->MarkAsModified(true, Tick()); - } - } + /// Find or create an image view in the guest descriptor table + ImageViewId VisitImageView(DescriptorTable<TICEntry>& table, + std::span<ImageViewId> cached_image_view_ids, u32 index); - void SetEmptyDepthBuffer() { - if (depth_buffer.target == nullptr) { - return; - } - depth_buffer.target->MarkAsRenderTarget(false, NO_RT); - depth_buffer.target = nullptr; - depth_buffer.view = nullptr; - } + /// Find or create a framebuffer with the given render target parameters + FramebufferId GetFramebufferId(const RenderTargets& key); - void SetEmptyColorBuffer(std::size_t index) { - if (render_targets[index].target == nullptr) { - return; - } - render_targets[index].target->MarkAsRenderTarget(false, NO_RT); - render_targets[index].target = nullptr; - render_targets[index].view = nullptr; - } - - void DoFermiCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src_config, - const Tegra::Engines::Fermi2D::Regs::Surface& dst_config, - const Tegra::Engines::Fermi2D::Config& copy_config) { - std::lock_guard lock{mutex}; - SurfaceParams src_params = SurfaceParams::CreateForFermiCopySurface(src_config); - SurfaceParams dst_params = SurfaceParams::CreateForFermiCopySurface(dst_config); - const GPUVAddr src_gpu_addr = src_config.Address(); - const GPUVAddr dst_gpu_addr = dst_config.Address(); - DeduceBestBlit(src_params, dst_params, src_gpu_addr, dst_gpu_addr); - - const std::optional<VAddr> dst_cpu_addr = gpu_memory.GpuToCpuAddress(dst_gpu_addr); - const std::optional<VAddr> src_cpu_addr = gpu_memory.GpuToCpuAddress(src_gpu_addr); - std::pair dst_surface = GetSurface(dst_gpu_addr, *dst_cpu_addr, dst_params, true, false); - TView src_surface = GetSurface(src_gpu_addr, *src_cpu_addr, src_params, true, false).second; - ImageBlit(src_surface, dst_surface.second, copy_config); - dst_surface.first->MarkAsModified(true, Tick()); - } - - TSurface TryFindFramebufferSurface(VAddr addr) const { - if (!addr) { - return nullptr; - } - const VAddr page = addr >> registry_page_bits; - const auto it = registry.find(page); - if (it == registry.end()) { - return nullptr; - } - const auto& list = it->second; - const auto found = std::find_if(list.begin(), list.end(), [addr](const auto& surface) { - return surface->GetCpuAddr() == addr; - }); - return found != list.end() ? *found : nullptr; - } + /// Refresh the contents (pixel data) of an image + void RefreshContents(Image& image); - u64 Tick() { - return ++ticks; - } + /// Upload data from guest to an image + template <typename MapBuffer> + void UploadImageContents(Image& image, MapBuffer& map, size_t buffer_offset); - void CommitAsyncFlushes() { - committed_flushes.push_back(uncommitted_flushes); - uncommitted_flushes.reset(); - } + /// Find or create an image view from a guest descriptor + [[nodiscard]] ImageViewId FindImageView(const TICEntry& config); - bool HasUncommittedFlushes() const { - return uncommitted_flushes != nullptr; - } + /// Create a new image view from a guest descriptor + [[nodiscard]] ImageViewId CreateImageView(const TICEntry& config); - bool ShouldWaitAsyncFlushes() const { - return !committed_flushes.empty() && committed_flushes.front() != nullptr; - } + /// Find or create an image from the given parameters + [[nodiscard]] ImageId FindOrInsertImage(const ImageInfo& info, GPUVAddr gpu_addr, + RelaxedOptions options = RelaxedOptions{}); - void PopAsyncFlushes() { - if (committed_flushes.empty()) { - return; - } - auto& flush_list = committed_flushes.front(); - if (!flush_list) { - committed_flushes.pop_front(); - return; - } - for (TSurface& surface : *flush_list) { - FlushSurface(surface); - } - committed_flushes.pop_front(); - } + /// Find an image from the given parameters + [[nodiscard]] ImageId FindImage(const ImageInfo& info, GPUVAddr gpu_addr, + RelaxedOptions options); -protected: - explicit TextureCache(VideoCore::RasterizerInterface& rasterizer_, - Tegra::Engines::Maxwell3D& maxwell3d_, Tegra::MemoryManager& gpu_memory_, - bool is_astc_supported_) - : is_astc_supported{is_astc_supported_}, rasterizer{rasterizer_}, maxwell3d{maxwell3d_}, - gpu_memory{gpu_memory_} { - for (std::size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { - SetEmptyColorBuffer(i); - } + /// Create an image from the given parameters + [[nodiscard]] ImageId InsertImage(const ImageInfo& info, GPUVAddr gpu_addr, + RelaxedOptions options); - SetEmptyDepthBuffer(); - staging_cache.SetSize(2); + /// Create a new image and join perfectly matching existing images + /// Remove joined images from the cache + [[nodiscard]] ImageId JoinImages(const ImageInfo& info, GPUVAddr gpu_addr, VAddr cpu_addr); - const auto make_siblings = [this](PixelFormat a, PixelFormat b) { - siblings_table[static_cast<std::size_t>(a)] = b; - siblings_table[static_cast<std::size_t>(b)] = a; - }; - std::fill(siblings_table.begin(), siblings_table.end(), PixelFormat::Invalid); - make_siblings(PixelFormat::D16_UNORM, PixelFormat::R16_UNORM); - make_siblings(PixelFormat::D32_FLOAT, PixelFormat::R32_FLOAT); - make_siblings(PixelFormat::D32_FLOAT_S8_UINT, PixelFormat::R32G32_FLOAT); + /// Return a blit image pair from the given guest blit parameters + [[nodiscard]] BlitImages GetBlitImages(const Tegra::Engines::Fermi2D::Surface& dst, + const Tegra::Engines::Fermi2D::Surface& src); - sampled_textures.reserve(64); - } + /// Find or create a sampler from a guest descriptor sampler + [[nodiscard]] SamplerId FindSampler(const TSCEntry& config); - ~TextureCache() = default; + /// Find or create an image view for the given color buffer index + [[nodiscard]] ImageViewId FindColorBuffer(size_t index, bool is_clear); - virtual TSurface CreateSurface(GPUVAddr gpu_addr, const SurfaceParams& params) = 0; + /// Find or create an image view for the depth buffer + [[nodiscard]] ImageViewId FindDepthBuffer(bool is_clear); - virtual void ImageCopy(TSurface& src_surface, TSurface& dst_surface, - const CopyParams& copy_params) = 0; + /// Find or create a view for a render target with the given image parameters + [[nodiscard]] ImageViewId FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr, + bool is_clear); - virtual void ImageBlit(TView& src_view, TView& dst_view, - const Tegra::Engines::Fermi2D::Config& copy_config) = 0; + /// Iterates over all the images in a region calling func + template <typename Func> + void ForEachImageInRegion(VAddr cpu_addr, size_t size, Func&& func); - // Depending on the backend, a buffer copy can be slow as it means deoptimizing the texture - // and reading it from a separate buffer. - virtual void BufferCopy(TSurface& src_surface, TSurface& dst_surface) = 0; + /// Find or create an image view in the given image with the passed parameters + [[nodiscard]] ImageViewId FindOrEmplaceImageView(ImageId image_id, const ImageViewInfo& info); - void ManageRenderTargetUnregister(TSurface& surface) { - auto& dirty = maxwell3d.dirty; - const u32 index = surface->GetRenderTarget(); - if (index == DEPTH_RT) { - dirty.flags[VideoCommon::Dirty::ZetaBuffer] = true; - } else { - dirty.flags[VideoCommon::Dirty::ColorBuffer0 + index] = true; - } - dirty.flags[VideoCommon::Dirty::RenderTargets] = true; + /// Register image in the page table + void RegisterImage(ImageId image); + + /// Unregister image from the page table + void UnregisterImage(ImageId image); + + /// Track CPU reads and writes for image + void TrackImage(ImageBase& image); + + /// Stop tracking CPU reads and writes for image + void UntrackImage(ImageBase& image); + + /// Delete image from the cache + void DeleteImage(ImageId image); + + /// Remove image views references from the cache + void RemoveImageViewReferences(std::span<const ImageViewId> removed_views); + + /// Remove framebuffers using the given image views from the cache + void RemoveFramebuffers(std::span<const ImageViewId> removed_views); + + /// Mark an image as modified from the GPU + void MarkModification(ImageBase& image) noexcept; + + /// Synchronize image aliases, copying data if needed + void SynchronizeAliases(ImageId image_id); + + /// Prepare an image to be used + void PrepareImage(ImageId image_id, bool is_modification, bool invalidate); + + /// Prepare an image view to be used + void PrepareImageView(ImageViewId image_view_id, bool is_modification, bool invalidate); + + /// Execute copies from one image to the other, even if they are incompatible + void CopyImage(ImageId dst_id, ImageId src_id, std::span<const ImageCopy> copies); + + /// Bind an image view as render target, downloading resources preemtively if needed + void BindRenderTarget(ImageViewId* old_id, ImageViewId new_id); + + /// Create a render target from a given image and image view parameters + [[nodiscard]] std::pair<FramebufferId, ImageViewId> RenderTargetFromImage( + ImageId, const ImageViewInfo& view_info); + + /// Returns true if the current clear parameters clear the whole image of a given image view + [[nodiscard]] bool IsFullClear(ImageViewId id); + + Runtime& runtime; + VideoCore::RasterizerInterface& rasterizer; + Tegra::Engines::Maxwell3D& maxwell3d; + Tegra::Engines::KeplerCompute& kepler_compute; + Tegra::MemoryManager& gpu_memory; + + DescriptorTable<TICEntry> graphics_image_table{gpu_memory}; + DescriptorTable<TSCEntry> graphics_sampler_table{gpu_memory}; + std::vector<SamplerId> graphics_sampler_ids; + std::vector<ImageViewId> graphics_image_view_ids; + + DescriptorTable<TICEntry> compute_image_table{gpu_memory}; + DescriptorTable<TSCEntry> compute_sampler_table{gpu_memory}; + std::vector<SamplerId> compute_sampler_ids; + std::vector<ImageViewId> compute_image_view_ids; + + RenderTargets render_targets; + + std::mutex mutex; + + std::unordered_map<TICEntry, ImageViewId> image_views; + std::unordered_map<TSCEntry, SamplerId> samplers; + std::unordered_map<RenderTargets, FramebufferId> framebuffers; + + std::unordered_map<u64, std::vector<ImageId>, IdentityHash<u64>> page_table; + + bool has_deleted_images = false; + + SlotVector<Image> slot_images; + SlotVector<ImageView> slot_image_views; + SlotVector<ImageAlloc> slot_image_allocs; + SlotVector<Sampler> slot_samplers; + SlotVector<Framebuffer> slot_framebuffers; + + // TODO: This data structure is not optimal and it should be reworked + std::vector<ImageId> uncommitted_downloads; + std::queue<std::vector<ImageId>> committed_downloads; + + static constexpr size_t TICKS_TO_DESTROY = 6; + DelayedDestructionRing<Image, TICKS_TO_DESTROY> sentenced_images; + DelayedDestructionRing<ImageView, TICKS_TO_DESTROY> sentenced_image_view; + DelayedDestructionRing<Framebuffer, TICKS_TO_DESTROY> sentenced_framebuffers; + + std::unordered_map<GPUVAddr, ImageAllocId> image_allocs_table; + + u64 modification_tick = 0; + u64 frame_tick = 0; +}; + +template <class P> +TextureCache<P>::TextureCache(Runtime& runtime_, VideoCore::RasterizerInterface& rasterizer_, + Tegra::Engines::Maxwell3D& maxwell3d_, + Tegra::Engines::KeplerCompute& kepler_compute_, + Tegra::MemoryManager& gpu_memory_) + : runtime{runtime_}, rasterizer{rasterizer_}, maxwell3d{maxwell3d_}, + kepler_compute{kepler_compute_}, gpu_memory{gpu_memory_} { + // Configure null sampler + TSCEntry sampler_descriptor{}; + sampler_descriptor.min_filter.Assign(Tegra::Texture::TextureFilter::Linear); + sampler_descriptor.mag_filter.Assign(Tegra::Texture::TextureFilter::Linear); + sampler_descriptor.mipmap_filter.Assign(Tegra::Texture::TextureMipmapFilter::Linear); + sampler_descriptor.cubemap_anisotropy.Assign(1); + + // Make sure the first index is reserved for the null resources + // This way the null resource becomes a compile time constant + void(slot_image_views.insert(runtime, NullImageParams{})); + void(slot_samplers.insert(runtime, sampler_descriptor)); +} + +template <class P> +void TextureCache<P>::TickFrame() { + // Tick sentenced resources in this order to ensure they are destroyed in the right order + sentenced_images.Tick(); + sentenced_framebuffers.Tick(); + sentenced_image_view.Tick(); + ++frame_tick; +} + +template <class P> +std::unique_lock<std::mutex> TextureCache<P>::AcquireLock() { + return std::unique_lock{mutex}; +} + +template <class P> +const typename P::ImageView& TextureCache<P>::GetImageView(ImageViewId id) const noexcept { + return slot_image_views[id]; +} + +template <class P> +typename P::ImageView& TextureCache<P>::GetImageView(ImageViewId id) noexcept { + return slot_image_views[id]; +} + +template <class P> +void TextureCache<P>::FillGraphicsImageViews(std::span<const u32> indices, + std::span<ImageViewId> image_view_ids) { + FillImageViews(graphics_image_table, graphics_image_view_ids, indices, image_view_ids); +} + +template <class P> +void TextureCache<P>::FillComputeImageViews(std::span<const u32> indices, + std::span<ImageViewId> image_view_ids) { + FillImageViews(compute_image_table, compute_image_view_ids, indices, image_view_ids); +} + +template <class P> +typename P::Sampler* TextureCache<P>::GetGraphicsSampler(u32 index) { + [[unlikely]] if (index > graphics_sampler_table.Limit()) { + LOG_ERROR(HW_GPU, "Invalid sampler index={}", index); + return &slot_samplers[NULL_SAMPLER_ID]; + } + const auto [descriptor, is_new] = graphics_sampler_table.Read(index); + SamplerId& id = graphics_sampler_ids[index]; + [[unlikely]] if (is_new) { + id = FindSampler(descriptor); } + return &slot_samplers[id]; +} - void Register(TSurface surface) { - const GPUVAddr gpu_addr = surface->GetGpuAddr(); - const std::size_t size = surface->GetSizeInBytes(); - const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); - if (!cpu_addr) { - LOG_CRITICAL(HW_GPU, "Failed to register surface with unmapped gpu_address 0x{:016x}", - gpu_addr); - return; - } - surface->SetCpuAddr(*cpu_addr); - RegisterInnerCache(surface); - surface->MarkAsRegistered(true); - surface->SetMemoryMarked(true); - rasterizer.UpdatePagesCachedCount(*cpu_addr, size, 1); +template <class P> +typename P::Sampler* TextureCache<P>::GetComputeSampler(u32 index) { + [[unlikely]] if (index > compute_sampler_table.Limit()) { + LOG_ERROR(HW_GPU, "Invalid sampler index={}", index); + return &slot_samplers[NULL_SAMPLER_ID]; + } + const auto [descriptor, is_new] = compute_sampler_table.Read(index); + SamplerId& id = compute_sampler_ids[index]; + [[unlikely]] if (is_new) { + id = FindSampler(descriptor); } + return &slot_samplers[id]; +} - void UnmarkMemory(TSurface surface) { - if (!surface->IsMemoryMarked()) { - return; - } - const std::size_t size = surface->GetSizeInBytes(); - const VAddr cpu_addr = surface->GetCpuAddr(); - rasterizer.UpdatePagesCachedCount(cpu_addr, size, -1); - surface->SetMemoryMarked(false); +template <class P> +void TextureCache<P>::SynchronizeGraphicsDescriptors() { + using SamplerIndex = Tegra::Engines::Maxwell3D::Regs::SamplerIndex; + const bool linked_tsc = maxwell3d.regs.sampler_index == SamplerIndex::ViaHeaderIndex; + const u32 tic_limit = maxwell3d.regs.tic.limit; + const u32 tsc_limit = linked_tsc ? tic_limit : maxwell3d.regs.tsc.limit; + if (graphics_sampler_table.Synchornize(maxwell3d.regs.tsc.Address(), tsc_limit)) { + graphics_sampler_ids.resize(tsc_limit + 1, CORRUPT_ID); } + if (graphics_image_table.Synchornize(maxwell3d.regs.tic.Address(), tic_limit)) { + graphics_image_view_ids.resize(tic_limit + 1, CORRUPT_ID); + } +} - void Unregister(TSurface surface) { - if (guard_render_targets && surface->IsProtected()) { - return; - } - if (!guard_render_targets && surface->IsRenderTarget()) { - ManageRenderTargetUnregister(surface); - } - UnmarkMemory(surface); - if (surface->IsSyncPending()) { - marked_for_unregister.remove(surface); - surface->SetSyncPending(false); - } - UnregisterInnerCache(surface); - surface->MarkAsRegistered(false); - ReserveSurface(surface->GetSurfaceParams(), surface); +template <class P> +void TextureCache<P>::SynchronizeComputeDescriptors() { + const bool linked_tsc = kepler_compute.launch_description.linked_tsc; + const u32 tic_limit = kepler_compute.regs.tic.limit; + const u32 tsc_limit = linked_tsc ? tic_limit : kepler_compute.regs.tsc.limit; + const GPUVAddr tsc_gpu_addr = kepler_compute.regs.tsc.Address(); + if (compute_sampler_table.Synchornize(tsc_gpu_addr, tsc_limit)) { + compute_sampler_ids.resize(tsc_limit + 1, CORRUPT_ID); } + if (compute_image_table.Synchornize(kepler_compute.regs.tic.Address(), tic_limit)) { + compute_image_view_ids.resize(tic_limit + 1, CORRUPT_ID); + } +} - TSurface GetUncachedSurface(const GPUVAddr gpu_addr, const SurfaceParams& params) { - if (const auto surface = TryGetReservedSurface(params); surface) { - surface->SetGpuAddr(gpu_addr); - return surface; - } - // No reserved surface available, create a new one and reserve it - auto new_surface{CreateSurface(gpu_addr, params)}; - return new_surface; +template <class P> +void TextureCache<P>::UpdateRenderTargets(bool is_clear) { + using namespace VideoCommon::Dirty; + auto& flags = maxwell3d.dirty.flags; + if (!flags[Dirty::RenderTargets]) { + return; } + flags[Dirty::RenderTargets] = false; - const bool is_astc_supported; + // Render target control is used on all render targets, so force look ups when this one is up + const bool force = flags[Dirty::RenderTargetControl]; + flags[Dirty::RenderTargetControl] = false; -private: - enum class RecycleStrategy : u32 { - Ignore = 0, - Flush = 1, - BufferCopy = 3, - }; + for (size_t index = 0; index < NUM_RT; ++index) { + ImageViewId& color_buffer_id = render_targets.color_buffer_ids[index]; + if (flags[Dirty::ColorBuffer0 + index] || force) { + flags[Dirty::ColorBuffer0 + index] = false; + BindRenderTarget(&color_buffer_id, FindColorBuffer(index, is_clear)); + } + PrepareImageView(color_buffer_id, true, is_clear && IsFullClear(color_buffer_id)); + } + if (flags[Dirty::ZetaBuffer] || force) { + flags[Dirty::ZetaBuffer] = false; + BindRenderTarget(&render_targets.depth_buffer_id, FindDepthBuffer(is_clear)); + } + const ImageViewId depth_buffer_id = render_targets.depth_buffer_id; + PrepareImageView(depth_buffer_id, true, is_clear && IsFullClear(depth_buffer_id)); - enum class DeductionType : u32 { - DeductionComplete, - DeductionIncomplete, - DeductionFailed, + for (size_t index = 0; index < NUM_RT; ++index) { + render_targets.draw_buffers[index] = static_cast<u8>(maxwell3d.regs.rt_control.Map(index)); + } + render_targets.size = Extent2D{ + maxwell3d.regs.render_area.width, + maxwell3d.regs.render_area.height, }; +} - struct Deduction { - DeductionType type{DeductionType::DeductionFailed}; - TSurface surface{}; +template <class P> +typename P::Framebuffer* TextureCache<P>::GetFramebuffer() { + return &slot_framebuffers[GetFramebufferId(render_targets)]; +} - bool Failed() const { - return type == DeductionType::DeductionFailed; - } +template <class P> +void TextureCache<P>::FillImageViews(DescriptorTable<TICEntry>& table, + std::span<ImageViewId> cached_image_view_ids, + std::span<const u32> indices, + std::span<ImageViewId> image_view_ids) { + ASSERT(indices.size() <= image_view_ids.size()); + do { + has_deleted_images = false; + std::ranges::transform(indices, image_view_ids.begin(), [&](u32 index) { + return VisitImageView(table, cached_image_view_ids, index); + }); + } while (has_deleted_images); +} - bool Incomplete() const { - return type == DeductionType::DeductionIncomplete; - } +template <class P> +ImageViewId TextureCache<P>::VisitImageView(DescriptorTable<TICEntry>& table, + std::span<ImageViewId> cached_image_view_ids, + u32 index) { + if (index > table.Limit()) { + LOG_ERROR(HW_GPU, "Invalid image view index={}", index); + return NULL_IMAGE_VIEW_ID; + } + const auto [descriptor, is_new] = table.Read(index); + ImageViewId& image_view_id = cached_image_view_ids[index]; + if (is_new) { + image_view_id = FindImageView(descriptor); + } + if (image_view_id != NULL_IMAGE_VIEW_ID) { + PrepareImageView(image_view_id, false, false); + } + return image_view_id; +} - bool IsDepth() const { - return surface->GetSurfaceParams().IsPixelFormatZeta(); - } - }; +template <class P> +FramebufferId TextureCache<P>::GetFramebufferId(const RenderTargets& key) { + const auto [pair, is_new] = framebuffers.try_emplace(key); + FramebufferId& framebuffer_id = pair->second; + if (!is_new) { + return framebuffer_id; + } + std::array<ImageView*, NUM_RT> color_buffers; + std::ranges::transform(key.color_buffer_ids, color_buffers.begin(), + [this](ImageViewId id) { return id ? &slot_image_views[id] : nullptr; }); + ImageView* const depth_buffer = + key.depth_buffer_id ? &slot_image_views[key.depth_buffer_id] : nullptr; + framebuffer_id = slot_framebuffers.insert(runtime, color_buffers, depth_buffer, key); + return framebuffer_id; +} - /** - * Takes care of selecting a proper strategy to deal with a texture recycle. - * - * @param overlaps The overlapping surfaces registered in the cache. - * @param params The parameters on the new surface. - * @param gpu_addr The starting address of the new surface. - * @param untopological Indicates to the recycler that the texture has no way - * to match the overlaps due to topological reasons. - **/ - RecycleStrategy PickStrategy(VectorSurface& overlaps, const SurfaceParams& params, - const GPUVAddr gpu_addr, const MatchTopologyResult untopological) { - if (Settings::IsGPULevelExtreme()) { - return RecycleStrategy::Flush; - } - // 3D Textures decision - if (params.target == SurfaceTarget::Texture3D) { - return RecycleStrategy::Flush; - } - for (const auto& s : overlaps) { - const auto& s_params = s->GetSurfaceParams(); - if (s_params.target == SurfaceTarget::Texture3D) { - return RecycleStrategy::Flush; - } - } - // Untopological decision - if (untopological == MatchTopologyResult::CompressUnmatch) { - return RecycleStrategy::Flush; - } - if (untopological == MatchTopologyResult::FullMatch && !params.is_tiled) { - return RecycleStrategy::Flush; - } - return RecycleStrategy::Ignore; - } - - /** - * Used to decide what to do with textures we can't resolve in the cache It has 2 implemented - * strategies: Ignore and Flush. - * - * - Ignore: Just unregisters all the overlaps and loads the new texture. - * - Flush: Flushes all the overlaps into memory and loads the new surface from that data. - * - * @param overlaps The overlapping surfaces registered in the cache. - * @param params The parameters for the new surface. - * @param gpu_addr The starting address of the new surface. - * @param preserve_contents Indicates that the new surface should be loaded from memory or left - * blank. - * @param untopological Indicates to the recycler that the texture has no way to match the - * overlaps due to topological reasons. - **/ - std::pair<TSurface, TView> RecycleSurface(VectorSurface& overlaps, const SurfaceParams& params, - const GPUVAddr gpu_addr, const bool preserve_contents, - const MatchTopologyResult untopological) { - const bool do_load = preserve_contents && Settings::IsGPULevelExtreme(); - for (auto& surface : overlaps) { - Unregister(surface); - } - switch (PickStrategy(overlaps, params, gpu_addr, untopological)) { - case RecycleStrategy::Ignore: { - return InitializeSurface(gpu_addr, params, do_load); - } - case RecycleStrategy::Flush: { - std::sort(overlaps.begin(), overlaps.end(), - [](const TSurface& a, const TSurface& b) -> bool { - return a->GetModificationTick() < b->GetModificationTick(); - }); - for (auto& surface : overlaps) { - FlushSurface(surface); - } - return InitializeSurface(gpu_addr, params, preserve_contents); +template <class P> +void TextureCache<P>::WriteMemory(VAddr cpu_addr, size_t size) { + ForEachImageInRegion(cpu_addr, size, [this](ImageId image_id, Image& image) { + if (True(image.flags & ImageFlagBits::CpuModified)) { + return; } - case RecycleStrategy::BufferCopy: { - auto new_surface = GetUncachedSurface(gpu_addr, params); - BufferCopy(overlaps[0], new_surface); - return {new_surface, new_surface->GetMainView()}; + image.flags |= ImageFlagBits::CpuModified; + UntrackImage(image); + }); +} + +template <class P> +void TextureCache<P>::DownloadMemory(VAddr cpu_addr, size_t size) { + std::vector<ImageId> images; + ForEachImageInRegion(cpu_addr, size, [this, &images](ImageId image_id, ImageBase& image) { + // Skip images that were not modified from the GPU + if (False(image.flags & ImageFlagBits::GpuModified)) { + return; } - default: { - UNIMPLEMENTED_MSG("Unimplemented Texture Cache Recycling Strategy!"); - return InitializeSurface(gpu_addr, params, do_load); + // Skip images that .are. modified from the CPU + // We don't want to write sensitive data from the guest + if (True(image.flags & ImageFlagBits::CpuModified)) { + return; } + if (image.info.num_samples > 1) { + LOG_WARNING(HW_GPU, "MSAA image downloads are not implemented"); + return; } + image.flags &= ~ImageFlagBits::GpuModified; + images.push_back(image_id); + }); + if (images.empty()) { + return; + } + std::ranges::sort(images, [this](ImageId lhs, ImageId rhs) { + return slot_images[lhs].modification_tick < slot_images[rhs].modification_tick; + }); + for (const ImageId image_id : images) { + Image& image = slot_images[image_id]; + auto map = runtime.MapDownloadBuffer(image.unswizzled_size_bytes); + const auto copies = FullDownloadCopies(image.info); + image.DownloadMemory(map, 0, copies); + runtime.Finish(); + SwizzleImage(gpu_memory, image.gpu_addr, image.info, copies, map.Span()); } +} - /** - * Takes a single surface and recreates into another that may differ in - * format, target or width alignment. - * - * @param current_surface The registered surface in the cache which we want to convert. - * @param params The new surface params which we'll use to recreate the surface. - * @param is_render Whether or not the surface is a render target. - **/ - std::pair<TSurface, TView> RebuildSurface(TSurface current_surface, const SurfaceParams& params, - bool is_render) { - const auto gpu_addr = current_surface->GetGpuAddr(); - const auto& cr_params = current_surface->GetSurfaceParams(); - TSurface new_surface; - if (cr_params.pixel_format != params.pixel_format && !is_render && - GetSiblingFormat(cr_params.pixel_format) == params.pixel_format) { - SurfaceParams new_params = params; - new_params.pixel_format = cr_params.pixel_format; - new_params.type = cr_params.type; - new_surface = GetUncachedSurface(gpu_addr, new_params); - } else { - new_surface = GetUncachedSurface(gpu_addr, params); - } - const SurfaceParams& final_params = new_surface->GetSurfaceParams(); - if (cr_params.type != final_params.type) { - if (Settings::IsGPULevelExtreme()) { - BufferCopy(current_surface, new_surface); - } - } else { - std::vector<CopyParams> bricks = current_surface->BreakDown(final_params); - for (auto& brick : bricks) { - TryCopyImage(current_surface, new_surface, brick); - } - } - Unregister(current_surface); - Register(new_surface); - new_surface->MarkAsModified(current_surface->IsModified(), Tick()); - return {new_surface, new_surface->GetMainView()}; - } - - /** - * Takes a single surface and checks with the new surface's params if it's an exact - * match, we return the main view of the registered surface. If its formats don't - * match, we rebuild the surface. We call this last method a `Mirage`. If formats - * match but the targets don't, we create an overview View of the registered surface. - * - * @param current_surface The registered surface in the cache which we want to convert. - * @param params The new surface params which we want to check. - * @param is_render Whether or not the surface is a render target. - **/ - std::pair<TSurface, TView> ManageStructuralMatch(TSurface current_surface, - const SurfaceParams& params, bool is_render) { - const bool is_mirage = !current_surface->MatchFormat(params.pixel_format); - const bool matches_target = current_surface->MatchTarget(params.target); - const auto match_check = [&]() -> std::pair<TSurface, TView> { - if (matches_target) { - return {current_surface, current_surface->GetMainView()}; - } - return {current_surface, current_surface->EmplaceOverview(params)}; - }; - if (!is_mirage) { - return match_check(); - } - if (!is_render && GetSiblingFormat(current_surface->GetFormat()) == params.pixel_format) { - return match_check(); - } - return RebuildSurface(current_surface, params, is_render); - } - - /** - * Unlike RebuildSurface where we know whether or not registered surfaces match the candidate - * in some way, we have no guarantees here. We try to see if the overlaps are sublayers/mipmaps - * of the new surface, if they all match we end up recreating a surface for them, - * else we return nothing. - * - * @param overlaps The overlapping surfaces registered in the cache. - * @param params The parameters on the new surface. - * @param gpu_addr The starting address of the new surface. - **/ - std::optional<std::pair<TSurface, TView>> TryReconstructSurface(VectorSurface& overlaps, - const SurfaceParams& params, - GPUVAddr gpu_addr) { - if (params.target == SurfaceTarget::Texture3D) { - return std::nullopt; - } - const auto test_modified = [](TSurface& surface) { return surface->IsModified(); }; - TSurface new_surface = GetUncachedSurface(gpu_addr, params); +template <class P> +void TextureCache<P>::UnmapMemory(VAddr cpu_addr, size_t size) { + std::vector<ImageId> deleted_images; + ForEachImageInRegion(cpu_addr, size, [&](ImageId id, Image&) { deleted_images.push_back(id); }); + for (const ImageId id : deleted_images) { + Image& image = slot_images[id]; + if (True(image.flags & ImageFlagBits::Tracked)) { + UntrackImage(image); + } + UnregisterImage(id); + DeleteImage(id); + } +} - if (std::none_of(overlaps.begin(), overlaps.end(), test_modified)) { - LoadSurface(new_surface); - for (const auto& surface : overlaps) { - Unregister(surface); - } - Register(new_surface); - return {{new_surface, new_surface->GetMainView()}}; - } +template <class P> +void TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst, + const Tegra::Engines::Fermi2D::Surface& src, + const Tegra::Engines::Fermi2D::Config& copy) { + const BlitImages images = GetBlitImages(dst, src); + const ImageId dst_id = images.dst_id; + const ImageId src_id = images.src_id; + PrepareImage(src_id, false, false); + PrepareImage(dst_id, true, false); + + ImageBase& dst_image = slot_images[dst_id]; + const ImageBase& src_image = slot_images[src_id]; + + // TODO: Deduplicate + const std::optional dst_base = dst_image.TryFindBase(dst.Address()); + const SubresourceRange dst_range{.base = dst_base.value(), .extent = {1, 1}}; + const ImageViewInfo dst_view_info(ImageViewType::e2D, images.dst_format, dst_range); + const auto [dst_framebuffer_id, dst_view_id] = RenderTargetFromImage(dst_id, dst_view_info); + const auto [src_samples_x, src_samples_y] = SamplesLog2(src_image.info.num_samples); + const std::array src_region{ + Offset2D{.x = copy.src_x0 >> src_samples_x, .y = copy.src_y0 >> src_samples_y}, + Offset2D{.x = copy.src_x1 >> src_samples_x, .y = copy.src_y1 >> src_samples_y}, + }; - std::size_t passed_tests = 0; - for (auto& surface : overlaps) { - const SurfaceParams& src_params = surface->GetSurfaceParams(); - const auto mipmap_layer{new_surface->GetLayerMipmap(surface->GetGpuAddr())}; - if (!mipmap_layer) { - continue; - } - const auto [base_layer, base_mipmap] = *mipmap_layer; - if (new_surface->GetMipmapSize(base_mipmap) != surface->GetMipmapSize(0)) { - continue; - } - ++passed_tests; - - // Copy all mipmaps and layers - const u32 block_width = params.GetDefaultBlockWidth(); - const u32 block_height = params.GetDefaultBlockHeight(); - for (u32 mipmap = base_mipmap; mipmap < base_mipmap + src_params.num_levels; ++mipmap) { - const u32 width = SurfaceParams::IntersectWidth(src_params, params, 0, mipmap); - const u32 height = SurfaceParams::IntersectHeight(src_params, params, 0, mipmap); - if (width < block_width || height < block_height) { - // Current APIs forbid copying small compressed textures, avoid errors - break; - } - const CopyParams copy_params(0, 0, 0, 0, 0, base_layer, 0, mipmap, width, height, - src_params.depth); - TryCopyImage(surface, new_surface, copy_params); - } - } - if (passed_tests == 0) { - return std::nullopt; - } - if (Settings::IsGPULevelExtreme() && passed_tests != overlaps.size()) { - // In Accurate GPU all tests should pass, else we recycle - return std::nullopt; - } + const std::optional src_base = src_image.TryFindBase(src.Address()); + const SubresourceRange src_range{.base = src_base.value(), .extent = {1, 1}}; + const ImageViewInfo src_view_info(ImageViewType::e2D, images.src_format, src_range); + const auto [src_framebuffer_id, src_view_id] = RenderTargetFromImage(src_id, src_view_info); + const auto [dst_samples_x, dst_samples_y] = SamplesLog2(dst_image.info.num_samples); + const std::array dst_region{ + Offset2D{.x = copy.dst_x0 >> dst_samples_x, .y = copy.dst_y0 >> dst_samples_y}, + Offset2D{.x = copy.dst_x1 >> dst_samples_x, .y = copy.dst_y1 >> dst_samples_y}, + }; - const bool modified = std::any_of(overlaps.begin(), overlaps.end(), test_modified); - for (const auto& surface : overlaps) { - Unregister(surface); - } + // Always call this after src_framebuffer_id was queried, as the address might be invalidated. + Framebuffer* const dst_framebuffer = &slot_framebuffers[dst_framebuffer_id]; + if constexpr (FRAMEBUFFER_BLITS) { + // OpenGL blits from framebuffers, not images + Framebuffer* const src_framebuffer = &slot_framebuffers[src_framebuffer_id]; + runtime.BlitFramebuffer(dst_framebuffer, src_framebuffer, dst_region, src_region, + copy.filter, copy.operation); + } else { + // Vulkan can blit images, but it lacks format reinterpretations + // Provide a framebuffer in case it's necessary + ImageView& dst_view = slot_image_views[dst_view_id]; + ImageView& src_view = slot_image_views[src_view_id]; + runtime.BlitImage(dst_framebuffer, dst_view, src_view, dst_region, src_region, copy.filter, + copy.operation); + } +} - new_surface->MarkAsModified(modified, Tick()); - Register(new_surface); - return {{new_surface, new_surface->GetMainView()}}; - } - - /** - * Takes care of managing 3D textures and its slices. Does HLE methods for reconstructing the 3D - * textures within the GPU if possible. Falls back to LLE when it isn't possible to use any of - * the HLE methods. - * - * @param overlaps The overlapping surfaces registered in the cache. - * @param params The parameters on the new surface. - * @param gpu_addr The starting address of the new surface. - * @param cpu_addr The starting address of the new surface on physical memory. - * @param preserve_contents Indicates that the new surface should be loaded from memory or - * left blank. - */ - std::optional<std::pair<TSurface, TView>> Manage3DSurfaces(VectorSurface& overlaps, - const SurfaceParams& params, - GPUVAddr gpu_addr, VAddr cpu_addr, - bool preserve_contents) { - if (params.target != SurfaceTarget::Texture3D) { - for (const auto& surface : overlaps) { - if (!surface->MatchTarget(params.target)) { - if (overlaps.size() == 1 && surface->GetCpuAddr() == cpu_addr) { - if (Settings::IsGPULevelExtreme()) { - return std::nullopt; - } - Unregister(surface); - return InitializeSurface(gpu_addr, params, preserve_contents); - } - return std::nullopt; - } - if (surface->GetCpuAddr() != cpu_addr) { - continue; - } - if (surface->MatchesStructure(params) == MatchStructureResult::FullMatch) { - return std::make_pair(surface, surface->GetMainView()); - } - } - return InitializeSurface(gpu_addr, params, preserve_contents); - } +template <class P> +void TextureCache<P>::InvalidateColorBuffer(size_t index) { + ImageViewId& color_buffer_id = render_targets.color_buffer_ids[index]; + color_buffer_id = FindColorBuffer(index, false); + if (!color_buffer_id) { + LOG_ERROR(HW_GPU, "Invalidating invalid color buffer in index={}", index); + return; + } + // When invalidating a color buffer, the old contents are no longer relevant + ImageView& color_buffer = slot_image_views[color_buffer_id]; + Image& image = slot_images[color_buffer.image_id]; + image.flags &= ~ImageFlagBits::CpuModified; + image.flags &= ~ImageFlagBits::GpuModified; - if (params.num_levels > 1) { - // We can't handle mipmaps in 3D textures yet, better fallback to LLE approach - return std::nullopt; - } + runtime.InvalidateColorBuffer(color_buffer, index); +} - if (overlaps.size() == 1) { - const auto& surface = overlaps[0]; - const SurfaceParams& overlap_params = surface->GetSurfaceParams(); - // Don't attempt to render to textures with more than one level for now - // The texture has to be to the right or the sample address if we want to render to it - if (overlap_params.num_levels == 1 && cpu_addr >= surface->GetCpuAddr()) { - const u32 offset = static_cast<u32>(cpu_addr - surface->GetCpuAddr()); - const u32 slice = std::get<2>(params.GetBlockOffsetXYZ(offset)); - if (slice < overlap_params.depth) { - auto view = surface->Emplace3DView(slice, params.depth, 0, 1); - return std::make_pair(std::move(surface), std::move(view)); - } - } - } +template <class P> +void TextureCache<P>::InvalidateDepthBuffer() { + ImageViewId& depth_buffer_id = render_targets.depth_buffer_id; + depth_buffer_id = FindDepthBuffer(false); + if (!depth_buffer_id) { + LOG_ERROR(HW_GPU, "Invalidating invalid depth buffer"); + return; + } + // When invalidating the depth buffer, the old contents are no longer relevant + ImageBase& image = slot_images[slot_image_views[depth_buffer_id].image_id]; + image.flags &= ~ImageFlagBits::CpuModified; + image.flags &= ~ImageFlagBits::GpuModified; - TSurface new_surface = GetUncachedSurface(gpu_addr, params); - bool modified = false; + ImageView& depth_buffer = slot_image_views[depth_buffer_id]; + runtime.InvalidateDepthBuffer(depth_buffer); +} - for (auto& surface : overlaps) { - const SurfaceParams& src_params = surface->GetSurfaceParams(); - if (src_params.target != SurfaceTarget::Texture2D || - src_params.height != params.height || - src_params.block_depth != params.block_depth || - src_params.block_height != params.block_height) { - return std::nullopt; - } - modified |= surface->IsModified(); - - const u32 offset = static_cast<u32>(surface->GetCpuAddr() - cpu_addr); - const u32 slice = std::get<2>(params.GetBlockOffsetXYZ(offset)); - const u32 width = params.width; - const u32 height = params.height; - const CopyParams copy_params(0, 0, 0, 0, 0, slice, 0, 0, width, height, 1); - TryCopyImage(surface, new_surface, copy_params); +template <class P> +typename P::ImageView* TextureCache<P>::TryFindFramebufferImageView(VAddr cpu_addr) { + // TODO: Properly implement this + const auto it = page_table.find(cpu_addr >> PAGE_BITS); + if (it == page_table.end()) { + return nullptr; + } + const auto& image_ids = it->second; + for (const ImageId image_id : image_ids) { + const ImageBase& image = slot_images[image_id]; + if (image.cpu_addr != cpu_addr) { + continue; } - for (const auto& surface : overlaps) { - Unregister(surface); + if (image.image_view_ids.empty()) { + continue; } - new_surface->MarkAsModified(modified, Tick()); - Register(new_surface); - - TView view = new_surface->GetMainView(); - return std::make_pair(std::move(new_surface), std::move(view)); - } - - /** - * Gets the starting address and parameters of a candidate surface and tries - * to find a matching surface within the cache. This is done in 3 big steps: - * - * 1. Check the 1st Level Cache in order to find an exact match, if we fail, we move to step 2. - * - * 2. Check if there are any overlaps at all, if there are none, we just load the texture from - * memory else we move to step 3. - * - * 3. Consists of figuring out the relationship between the candidate texture and the - * overlaps. We divide the scenarios depending if there's 1 or many overlaps. If - * there's many, we just try to reconstruct a new surface out of them based on the - * candidate's parameters, if we fail, we recycle. When there's only 1 overlap then we - * have to check if the candidate is a view (layer/mipmap) of the overlap or if the - * registered surface is a mipmap/layer of the candidate. In this last case we reconstruct - * a new surface. - * - * @param gpu_addr The starting address of the candidate surface. - * @param params The parameters on the candidate surface. - * @param preserve_contents Indicates that the new surface should be loaded from memory or - * left blank. - * @param is_render Whether or not the surface is a render target. - **/ - std::pair<TSurface, TView> GetSurface(const GPUVAddr gpu_addr, const VAddr cpu_addr, - const SurfaceParams& params, bool preserve_contents, - bool is_render) { - // Step 1 - // Check Level 1 Cache for a fast structural match. If candidate surface - // matches at certain level we are pretty much done. - if (const auto iter = l1_cache.find(cpu_addr); iter != l1_cache.end()) { - TSurface& current_surface = iter->second; - const auto topological_result = current_surface->MatchesTopology(params); - if (topological_result != MatchTopologyResult::FullMatch) { - VectorSurface overlaps{current_surface}; - return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, - topological_result); - } + return &slot_image_views[image.image_view_ids.at(0)]; + } + return nullptr; +} - const auto struct_result = current_surface->MatchesStructure(params); - if (struct_result != MatchStructureResult::None) { - const auto& old_params = current_surface->GetSurfaceParams(); - const bool not_3d = params.target != SurfaceTarget::Texture3D && - old_params.target != SurfaceTarget::Texture3D; - if (not_3d || current_surface->MatchTarget(params.target)) { - if (struct_result == MatchStructureResult::FullMatch) { - return ManageStructuralMatch(current_surface, params, is_render); - } else { - return RebuildSurface(current_surface, params, is_render); - } - } - } - } +template <class P> +bool TextureCache<P>::HasUncommittedFlushes() const noexcept { + return !uncommitted_downloads.empty(); +} - // Step 2 - // Obtain all possible overlaps in the memory region - const std::size_t candidate_size = params.GetGuestSizeInBytes(); - auto overlaps{GetSurfacesInRegion(cpu_addr, candidate_size)}; +template <class P> +bool TextureCache<P>::ShouldWaitAsyncFlushes() const noexcept { + return !committed_downloads.empty() && !committed_downloads.front().empty(); +} - // If none are found, we are done. we just load the surface and create it. - if (overlaps.empty()) { - return InitializeSurface(gpu_addr, params, preserve_contents); - } +template <class P> +void TextureCache<P>::CommitAsyncFlushes() { + // This is intentionally passing the value by copy + committed_downloads.push(uncommitted_downloads); + uncommitted_downloads.clear(); +} - // Step 3 - // Now we need to figure the relationship between the texture and its overlaps - // we do a topological test to ensure we can find some relationship. If it fails - // immediately recycle the texture - for (const auto& surface : overlaps) { - const auto topological_result = surface->MatchesTopology(params); - if (topological_result != MatchTopologyResult::FullMatch) { - return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, - topological_result); - } - } +template <class P> +void TextureCache<P>::PopAsyncFlushes() { + if (committed_downloads.empty()) { + return; + } + const std::span<const ImageId> download_ids = committed_downloads.front(); + if (download_ids.empty()) { + committed_downloads.pop(); + return; + } + size_t total_size_bytes = 0; + for (const ImageId image_id : download_ids) { + total_size_bytes += slot_images[image_id].unswizzled_size_bytes; + } + auto download_map = runtime.MapDownloadBuffer(total_size_bytes); + size_t buffer_offset = 0; + for (const ImageId image_id : download_ids) { + Image& image = slot_images[image_id]; + const auto copies = FullDownloadCopies(image.info); + image.DownloadMemory(download_map, buffer_offset, copies); + buffer_offset += image.unswizzled_size_bytes; + } + // Wait for downloads to finish + runtime.Finish(); + + buffer_offset = 0; + const std::span<u8> download_span = download_map.Span(); + for (const ImageId image_id : download_ids) { + const ImageBase& image = slot_images[image_id]; + const auto copies = FullDownloadCopies(image.info); + const std::span<u8> image_download_span = download_span.subspan(buffer_offset); + SwizzleImage(gpu_memory, image.gpu_addr, image.info, copies, image_download_span); + buffer_offset += image.unswizzled_size_bytes; + } + committed_downloads.pop(); +} - // Manage 3D textures - if (params.block_depth > 0) { - auto surface = - Manage3DSurfaces(overlaps, params, gpu_addr, cpu_addr, preserve_contents); - if (surface) { - return *surface; - } +template <class P> +bool TextureCache<P>::IsRegionGpuModified(VAddr addr, size_t size) { + bool is_modified = false; + ForEachImageInRegion(addr, size, [&is_modified](ImageId, ImageBase& image) { + if (False(image.flags & ImageFlagBits::GpuModified)) { + return false; } + is_modified = true; + return true; + }); + return is_modified; +} - // Split cases between 1 overlap or many. - if (overlaps.size() == 1) { - TSurface current_surface = overlaps[0]; - // First check if the surface is within the overlap. If not, it means - // two things either the candidate surface is a supertexture of the overlap - // or they don't match in any known way. - if (!current_surface->IsInside(gpu_addr, gpu_addr + candidate_size)) { - const std::optional view = TryReconstructSurface(overlaps, params, gpu_addr); - if (view) { - return *view; - } - return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, - MatchTopologyResult::FullMatch); - } - // Now we check if the candidate is a mipmap/layer of the overlap - std::optional<TView> view = - current_surface->EmplaceView(params, gpu_addr, candidate_size); - if (view) { - const bool is_mirage = !current_surface->MatchFormat(params.pixel_format); - if (is_mirage) { - // On a mirage view, we need to recreate the surface under this new view - // and then obtain a view again. - SurfaceParams new_params = current_surface->GetSurfaceParams(); - const u32 wh = SurfaceParams::ConvertWidth( - new_params.width, new_params.pixel_format, params.pixel_format); - const u32 hh = SurfaceParams::ConvertHeight( - new_params.height, new_params.pixel_format, params.pixel_format); - new_params.width = wh; - new_params.height = hh; - new_params.pixel_format = params.pixel_format; - std::pair<TSurface, TView> pair = - RebuildSurface(current_surface, new_params, is_render); - std::optional<TView> mirage_view = - pair.first->EmplaceView(params, gpu_addr, candidate_size); - if (mirage_view) - return {pair.first, *mirage_view}; - return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, - MatchTopologyResult::FullMatch); - } - return {current_surface, *view}; - } - } else { - // If there are many overlaps, odds are they are subtextures of the candidate - // surface. We try to construct a new surface based on the candidate parameters, - // using the overlaps. If a single overlap fails, this will fail. - std::optional<std::pair<TSurface, TView>> view = - TryReconstructSurface(overlaps, params, gpu_addr); - if (view) { - return *view; - } - } - // We failed all the tests, recycle the overlaps into a new texture. - return RecycleSurface(overlaps, params, gpu_addr, preserve_contents, - MatchTopologyResult::FullMatch); - } - - /** - * Gets the starting address and parameters of a candidate surface and tries to find a - * matching surface within the cache that's similar to it. If there are many textures - * or the texture found if entirely incompatible, it will fail. If no texture is found, the - * blit will be unsuccessful. - * - * @param gpu_addr The starting address of the candidate surface. - * @param params The parameters on the candidate surface. - **/ - Deduction DeduceSurface(const GPUVAddr gpu_addr, const SurfaceParams& params) { - const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); - - if (!cpu_addr) { - Deduction result{}; - result.type = DeductionType::DeductionFailed; - return result; - } +template <class P> +void TextureCache<P>::RefreshContents(Image& image) { + if (False(image.flags & ImageFlagBits::CpuModified)) { + // Only upload modified images + return; + } + image.flags &= ~ImageFlagBits::CpuModified; + TrackImage(image); - if (const auto iter = l1_cache.find(*cpu_addr); iter != l1_cache.end()) { - TSurface& current_surface = iter->second; - const auto topological_result = current_surface->MatchesTopology(params); - if (topological_result != MatchTopologyResult::FullMatch) { - Deduction result{}; - result.type = DeductionType::DeductionFailed; - return result; - } - const auto struct_result = current_surface->MatchesStructure(params); - if (struct_result != MatchStructureResult::None && - current_surface->MatchTarget(params.target)) { - Deduction result{}; - result.type = DeductionType::DeductionComplete; - result.surface = current_surface; - return result; - } - } + if (image.info.num_samples > 1) { + LOG_WARNING(HW_GPU, "MSAA image uploads are not implemented"); + return; + } + auto map = runtime.MapUploadBuffer(MapSizeBytes(image)); + UploadImageContents(image, map, 0); + runtime.InsertUploadMemoryBarrier(); +} - const std::size_t candidate_size = params.GetGuestSizeInBytes(); - auto overlaps{GetSurfacesInRegion(*cpu_addr, candidate_size)}; +template <class P> +template <typename MapBuffer> +void TextureCache<P>::UploadImageContents(Image& image, MapBuffer& map, size_t buffer_offset) { + const std::span<u8> mapped_span = map.Span().subspan(buffer_offset); + const GPUVAddr gpu_addr = image.gpu_addr; + + if (True(image.flags & ImageFlagBits::AcceleratedUpload)) { + gpu_memory.ReadBlockUnsafe(gpu_addr, mapped_span.data(), mapped_span.size_bytes()); + const auto uploads = FullUploadSwizzles(image.info); + runtime.AccelerateImageUpload(image, map, buffer_offset, uploads); + } else if (True(image.flags & ImageFlagBits::Converted)) { + std::vector<u8> unswizzled_data(image.unswizzled_size_bytes); + auto copies = UnswizzleImage(gpu_memory, gpu_addr, image.info, unswizzled_data); + ConvertImage(unswizzled_data, image.info, mapped_span, copies); + image.UploadMemory(map, buffer_offset, copies); + } else if (image.info.type == ImageType::Buffer) { + const std::array copies{UploadBufferCopy(gpu_memory, gpu_addr, image, mapped_span)}; + image.UploadMemory(map, buffer_offset, copies); + } else { + const auto copies = UnswizzleImage(gpu_memory, gpu_addr, image.info, mapped_span); + image.UploadMemory(map, buffer_offset, copies); + } +} - if (overlaps.empty()) { - Deduction result{}; - result.type = DeductionType::DeductionIncomplete; - return result; - } +template <class P> +ImageViewId TextureCache<P>::FindImageView(const TICEntry& config) { + if (!IsValidAddress(gpu_memory, config)) { + return NULL_IMAGE_VIEW_ID; + } + const auto [pair, is_new] = image_views.try_emplace(config); + ImageViewId& image_view_id = pair->second; + if (is_new) { + image_view_id = CreateImageView(config); + } + return image_view_id; +} - if (overlaps.size() > 1) { - Deduction result{}; - result.type = DeductionType::DeductionFailed; - return result; - } else { - Deduction result{}; - result.type = DeductionType::DeductionComplete; - result.surface = overlaps[0]; - return result; - } +template <class P> +ImageViewId TextureCache<P>::CreateImageView(const TICEntry& config) { + const ImageInfo info(config); + const GPUVAddr image_gpu_addr = config.Address() - config.BaseLayer() * info.layer_stride; + const ImageId image_id = FindOrInsertImage(info, image_gpu_addr); + if (!image_id) { + return NULL_IMAGE_VIEW_ID; } + ImageBase& image = slot_images[image_id]; + const SubresourceBase base = image.TryFindBase(config.Address()).value(); + ASSERT(base.level == 0); + const ImageViewInfo view_info(config, base.layer); + const ImageViewId image_view_id = FindOrEmplaceImageView(image_id, view_info); + ImageViewBase& image_view = slot_image_views[image_view_id]; + image_view.flags |= ImageViewFlagBits::Strong; + image.flags |= ImageFlagBits::Strong; + return image_view_id; +} - /** - * Gets a null surface based on a target texture. - * @param target The target of the null surface. - */ - TView GetNullSurface(SurfaceTarget target) { - const u32 i_target = static_cast<u32>(target); - if (const auto it = invalid_cache.find(i_target); it != invalid_cache.end()) { - return it->second->GetMainView(); - } - SurfaceParams params{}; - params.target = target; - params.is_tiled = false; - params.srgb_conversion = false; - params.is_layered = - target == SurfaceTarget::Texture1DArray || target == SurfaceTarget::Texture2DArray || - target == SurfaceTarget::TextureCubemap || target == SurfaceTarget::TextureCubeArray; - params.block_width = 0; - params.block_height = 0; - params.block_depth = 0; - params.tile_width_spacing = 1; - params.width = 1; - params.height = 1; - params.depth = 1; - if (target == SurfaceTarget::TextureCubemap || target == SurfaceTarget::TextureCubeArray) { - params.depth = 6; - } - params.pitch = 4; - params.num_levels = 1; - params.emulated_levels = 1; - params.pixel_format = VideoCore::Surface::PixelFormat::R8_UNORM; - params.type = VideoCore::Surface::SurfaceType::ColorTexture; - auto surface = CreateSurface(0ULL, params); - invalid_memory.resize(surface->GetHostSizeInBytes(), 0U); - surface->UploadTexture(invalid_memory); - surface->MarkAsModified(false, Tick()); - invalid_cache.emplace(i_target, surface); - return surface->GetMainView(); - } - - /** - * Gets the a source and destination starting address and parameters, - * and tries to deduce if they are supposed to be depth textures. If so, their - * parameters are modified and fixed into so. - * - * @param src_params The parameters of the candidate surface. - * @param dst_params The parameters of the destination surface. - * @param src_gpu_addr The starting address of the candidate surface. - * @param dst_gpu_addr The starting address of the destination surface. - **/ - void DeduceBestBlit(SurfaceParams& src_params, SurfaceParams& dst_params, - const GPUVAddr src_gpu_addr, const GPUVAddr dst_gpu_addr) { - auto deduced_src = DeduceSurface(src_gpu_addr, src_params); - auto deduced_dst = DeduceSurface(dst_gpu_addr, dst_params); - if (deduced_src.Failed() || deduced_dst.Failed()) { - return; +template <class P> +ImageId TextureCache<P>::FindOrInsertImage(const ImageInfo& info, GPUVAddr gpu_addr, + RelaxedOptions options) { + if (const ImageId image_id = FindImage(info, gpu_addr, options); image_id) { + return image_id; + } + return InsertImage(info, gpu_addr, options); +} + +template <class P> +ImageId TextureCache<P>::FindImage(const ImageInfo& info, GPUVAddr gpu_addr, + RelaxedOptions options) { + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); + if (!cpu_addr) { + return ImageId{}; + } + const bool broken_views = runtime.HasBrokenTextureViewFormats(); + ImageId image_id; + const auto lambda = [&](ImageId existing_image_id, ImageBase& existing_image) { + if (info.type == ImageType::Linear || existing_image.info.type == ImageType::Linear) { + const bool strict_size = False(options & RelaxedOptions::Size) && + True(existing_image.flags & ImageFlagBits::Strong); + const ImageInfo& existing = existing_image.info; + if (existing_image.gpu_addr == gpu_addr && existing.type == info.type && + existing.pitch == info.pitch && + IsPitchLinearSameSize(existing, info, strict_size) && + IsViewCompatible(existing.format, info.format, broken_views)) { + image_id = existing_image_id; + return true; + } + } else if (IsSubresource(info, existing_image, gpu_addr, options, broken_views)) { + image_id = existing_image_id; + return true; } + return false; + }; + ForEachImageInRegion(*cpu_addr, CalculateGuestSizeInBytes(info), lambda); + return image_id; +} - const bool incomplete_src = deduced_src.Incomplete(); - const bool incomplete_dst = deduced_dst.Incomplete(); +template <class P> +ImageId TextureCache<P>::InsertImage(const ImageInfo& info, GPUVAddr gpu_addr, + RelaxedOptions options) { + const std::optional<VAddr> cpu_addr = gpu_memory.GpuToCpuAddress(gpu_addr); + ASSERT_MSG(cpu_addr, "Tried to insert an image to an invalid gpu_addr=0x{:x}", gpu_addr); + const ImageId image_id = JoinImages(info, gpu_addr, *cpu_addr); + const Image& image = slot_images[image_id]; + // Using "image.gpu_addr" instead of "gpu_addr" is important because it might be different + const auto [it, is_new] = image_allocs_table.try_emplace(image.gpu_addr); + if (is_new) { + it->second = slot_image_allocs.insert(); + } + slot_image_allocs[it->second].images.push_back(image_id); + return image_id; +} - if (incomplete_src && incomplete_dst) { +template <class P> +ImageId TextureCache<P>::JoinImages(const ImageInfo& info, GPUVAddr gpu_addr, VAddr cpu_addr) { + ImageInfo new_info = info; + const size_t size_bytes = CalculateGuestSizeInBytes(new_info); + const bool broken_views = runtime.HasBrokenTextureViewFormats(); + std::vector<ImageId> overlap_ids; + std::vector<ImageId> left_aliased_ids; + std::vector<ImageId> right_aliased_ids; + ForEachImageInRegion(cpu_addr, size_bytes, [&](ImageId overlap_id, ImageBase& overlap) { + if (info.type != overlap.info.type) { return; } - - const bool any_incomplete = incomplete_src || incomplete_dst; - - if (!any_incomplete) { - if (!(deduced_src.IsDepth() && deduced_dst.IsDepth())) { - return; - } - } else { - if (incomplete_src && !(deduced_dst.IsDepth())) { - return; - } - - if (incomplete_dst && !(deduced_src.IsDepth())) { - return; + if (info.type == ImageType::Linear) { + if (info.pitch == overlap.info.pitch && gpu_addr == overlap.gpu_addr) { + // Alias linear images with the same pitch + left_aliased_ids.push_back(overlap_id); } + return; } - - const auto inherit_format = [](SurfaceParams& to, TSurface from) { - const SurfaceParams& params = from->GetSurfaceParams(); - to.pixel_format = params.pixel_format; - to.type = params.type; - }; - // Now we got the cases where one or both is Depth and the other is not known - if (!incomplete_src) { - inherit_format(src_params, deduced_src.surface); - } else { - inherit_format(src_params, deduced_dst.surface); + static constexpr bool strict_size = true; + const std::optional<OverlapResult> solution = + ResolveOverlap(new_info, gpu_addr, cpu_addr, overlap, strict_size, broken_views); + if (solution) { + gpu_addr = solution->gpu_addr; + cpu_addr = solution->cpu_addr; + new_info.resources = solution->resources; + overlap_ids.push_back(overlap_id); + return; } - if (!incomplete_dst) { - inherit_format(dst_params, deduced_dst.surface); + static constexpr auto options = RelaxedOptions::Size | RelaxedOptions::Format; + const ImageBase new_image_base(new_info, gpu_addr, cpu_addr); + if (IsSubresource(new_info, overlap, gpu_addr, options, broken_views)) { + left_aliased_ids.push_back(overlap_id); + } else if (IsSubresource(overlap.info, new_image_base, overlap.gpu_addr, options, + broken_views)) { + right_aliased_ids.push_back(overlap_id); + } + }); + const ImageId new_image_id = slot_images.insert(runtime, new_info, gpu_addr, cpu_addr); + Image& new_image = slot_images[new_image_id]; + + // TODO: Only upload what we need + RefreshContents(new_image); + + for (const ImageId overlap_id : overlap_ids) { + Image& overlap = slot_images[overlap_id]; + if (overlap.info.num_samples != new_image.info.num_samples) { + LOG_WARNING(HW_GPU, "Copying between images with different samples is not implemented"); } else { - inherit_format(dst_params, deduced_src.surface); + const SubresourceBase base = new_image.TryFindBase(overlap.gpu_addr).value(); + const auto copies = MakeShrinkImageCopies(new_info, overlap.info, base); + runtime.CopyImage(new_image, overlap, copies); } + if (True(overlap.flags & ImageFlagBits::Tracked)) { + UntrackImage(overlap); + } + UnregisterImage(overlap_id); + DeleteImage(overlap_id); + } + ImageBase& new_image_base = new_image; + for (const ImageId aliased_id : right_aliased_ids) { + ImageBase& aliased = slot_images[aliased_id]; + AddImageAlias(new_image_base, aliased, new_image_id, aliased_id); + } + for (const ImageId aliased_id : left_aliased_ids) { + ImageBase& aliased = slot_images[aliased_id]; + AddImageAlias(aliased, new_image_base, aliased_id, new_image_id); } + RegisterImage(new_image_id); + return new_image_id; +} - std::pair<TSurface, TView> InitializeSurface(GPUVAddr gpu_addr, const SurfaceParams& params, - bool preserve_contents) { - auto new_surface{GetUncachedSurface(gpu_addr, params)}; - Register(new_surface); - if (preserve_contents) { - LoadSurface(new_surface); - } - return {new_surface, new_surface->GetMainView()}; +template <class P> +typename TextureCache<P>::BlitImages TextureCache<P>::GetBlitImages( + const Tegra::Engines::Fermi2D::Surface& dst, const Tegra::Engines::Fermi2D::Surface& src) { + static constexpr auto FIND_OPTIONS = RelaxedOptions::Format | RelaxedOptions::Samples; + const GPUVAddr dst_addr = dst.Address(); + const GPUVAddr src_addr = src.Address(); + ImageInfo dst_info(dst); + ImageInfo src_info(src); + ImageId dst_id; + ImageId src_id; + do { + has_deleted_images = false; + dst_id = FindImage(dst_info, dst_addr, FIND_OPTIONS); + src_id = FindImage(src_info, src_addr, FIND_OPTIONS); + const ImageBase* const dst_image = dst_id ? &slot_images[dst_id] : nullptr; + const ImageBase* const src_image = src_id ? &slot_images[src_id] : nullptr; + DeduceBlitImages(dst_info, src_info, dst_image, src_image); + if (GetFormatType(dst_info.format) != GetFormatType(src_info.format)) { + continue; + } + if (!dst_id) { + dst_id = InsertImage(dst_info, dst_addr, RelaxedOptions{}); + } + if (!src_id) { + src_id = InsertImage(src_info, src_addr, RelaxedOptions{}); + } + } while (has_deleted_images); + return BlitImages{ + .dst_id = dst_id, + .src_id = src_id, + .dst_format = dst_info.format, + .src_format = src_info.format, + }; +} + +template <class P> +SamplerId TextureCache<P>::FindSampler(const TSCEntry& config) { + if (std::ranges::all_of(config.raw, [](u64 value) { return value == 0; })) { + return NULL_SAMPLER_ID; + } + const auto [pair, is_new] = samplers.try_emplace(config); + if (is_new) { + pair->second = slot_samplers.insert(runtime, config); } + return pair->second; +} - void LoadSurface(const TSurface& surface) { - staging_cache.GetBuffer(0).resize(surface->GetHostSizeInBytes()); - surface->LoadBuffer(gpu_memory, staging_cache); - surface->UploadTexture(staging_cache.GetBuffer(0)); - surface->MarkAsModified(false, Tick()); +template <class P> +ImageViewId TextureCache<P>::FindColorBuffer(size_t index, bool is_clear) { + const auto& regs = maxwell3d.regs; + if (index >= regs.rt_control.count) { + return ImageViewId{}; + } + const auto& rt = regs.rt[index]; + const GPUVAddr gpu_addr = rt.Address(); + if (gpu_addr == 0) { + return ImageViewId{}; + } + if (rt.format == Tegra::RenderTargetFormat::NONE) { + return ImageViewId{}; } + const ImageInfo info(regs, index); + return FindRenderTargetView(info, gpu_addr, is_clear); +} - void FlushSurface(const TSurface& surface) { - if (!surface->IsModified()) { - return; - } - staging_cache.GetBuffer(0).resize(surface->GetHostSizeInBytes()); - surface->DownloadTexture(staging_cache.GetBuffer(0)); - surface->FlushBuffer(gpu_memory, staging_cache); - surface->MarkAsModified(false, Tick()); - } - - void RegisterInnerCache(TSurface& surface) { - const VAddr cpu_addr = surface->GetCpuAddr(); - VAddr start = cpu_addr >> registry_page_bits; - const VAddr end = (surface->GetCpuAddrEnd() - 1) >> registry_page_bits; - l1_cache[cpu_addr] = surface; - while (start <= end) { - registry[start].push_back(surface); - start++; - } +template <class P> +ImageViewId TextureCache<P>::FindDepthBuffer(bool is_clear) { + const auto& regs = maxwell3d.regs; + if (!regs.zeta_enable) { + return ImageViewId{}; + } + const GPUVAddr gpu_addr = regs.zeta.Address(); + if (gpu_addr == 0) { + return ImageViewId{}; } + const ImageInfo info(regs); + return FindRenderTargetView(info, gpu_addr, is_clear); +} - void UnregisterInnerCache(TSurface& surface) { - const VAddr cpu_addr = surface->GetCpuAddr(); - VAddr start = cpu_addr >> registry_page_bits; - const VAddr end = (surface->GetCpuAddrEnd() - 1) >> registry_page_bits; - l1_cache.erase(cpu_addr); - while (start <= end) { - auto& reg{registry[start]}; - reg.erase(std::find(reg.begin(), reg.end(), surface)); - start++; - } +template <class P> +ImageViewId TextureCache<P>::FindRenderTargetView(const ImageInfo& info, GPUVAddr gpu_addr, + bool is_clear) { + const auto options = is_clear ? RelaxedOptions::Samples : RelaxedOptions{}; + const ImageId image_id = FindOrInsertImage(info, gpu_addr, options); + if (!image_id) { + return NULL_IMAGE_VIEW_ID; + } + Image& image = slot_images[image_id]; + const ImageViewType view_type = RenderTargetImageViewType(info); + SubresourceBase base; + if (image.info.type == ImageType::Linear) { + base = SubresourceBase{.level = 0, .layer = 0}; + } else { + base = image.TryFindBase(gpu_addr).value(); } + const s32 layers = image.info.type == ImageType::e3D ? info.size.depth : info.resources.layers; + const SubresourceRange range{ + .base = base, + .extent = {.levels = 1, .layers = layers}, + }; + return FindOrEmplaceImageView(image_id, ImageViewInfo(view_type, info.format, range)); +} - VectorSurface GetSurfacesInRegion(const VAddr cpu_addr, const std::size_t size) { - if (size == 0) { - return {}; +template <class P> +template <typename Func> +void TextureCache<P>::ForEachImageInRegion(VAddr cpu_addr, size_t size, Func&& func) { + using FuncReturn = typename std::invoke_result<Func, ImageId, Image&>::type; + static constexpr bool BOOL_BREAK = std::is_same_v<FuncReturn, bool>; + boost::container::small_vector<ImageId, 32> images; + ForEachPage(cpu_addr, size, [this, &images, cpu_addr, size, func](u64 page) { + const auto it = page_table.find(page); + if (it == page_table.end()) { + if constexpr (BOOL_BREAK) { + return false; + } else { + return; + } } - const VAddr cpu_addr_end = cpu_addr + size; - const VAddr end = (cpu_addr_end - 1) >> registry_page_bits; - VectorSurface surfaces; - for (VAddr start = cpu_addr >> registry_page_bits; start <= end; ++start) { - const auto it = registry.find(start); - if (it == registry.end()) { + for (const ImageId image_id : it->second) { + Image& image = slot_images[image_id]; + if (True(image.flags & ImageFlagBits::Picked)) { continue; } - for (auto& surface : it->second) { - if (surface->IsPicked() || !surface->Overlaps(cpu_addr, cpu_addr_end)) { - continue; + if (!image.Overlaps(cpu_addr, size)) { + continue; + } + image.flags |= ImageFlagBits::Picked; + images.push_back(image_id); + if constexpr (BOOL_BREAK) { + if (func(image_id, image)) { + return true; } - surface->MarkAsPicked(true); - surfaces.push_back(surface); + } else { + func(image_id, image); } } - for (auto& surface : surfaces) { - surface->MarkAsPicked(false); + if constexpr (BOOL_BREAK) { + return false; } - return surfaces; + }); + for (const ImageId image_id : images) { + slot_images[image_id].flags &= ~ImageFlagBits::Picked; } +} - void ReserveSurface(const SurfaceParams& params, TSurface surface) { - surface_reserve[params].push_back(std::move(surface)); +template <class P> +ImageViewId TextureCache<P>::FindOrEmplaceImageView(ImageId image_id, const ImageViewInfo& info) { + Image& image = slot_images[image_id]; + if (const ImageViewId image_view_id = image.FindView(info); image_view_id) { + return image_view_id; } + const ImageViewId image_view_id = slot_image_views.insert(runtime, info, image_id, image); + image.InsertView(info, image_view_id); + return image_view_id; +} + +template <class P> +void TextureCache<P>::RegisterImage(ImageId image_id) { + ImageBase& image = slot_images[image_id]; + ASSERT_MSG(False(image.flags & ImageFlagBits::Registered), + "Trying to register an already registered image"); + image.flags |= ImageFlagBits::Registered; + ForEachPage(image.cpu_addr, image.guest_size_bytes, + [this, image_id](u64 page) { page_table[page].push_back(image_id); }); +} - TSurface TryGetReservedSurface(const SurfaceParams& params) { - auto search{surface_reserve.find(params)}; - if (search == surface_reserve.end()) { - return {}; +template <class P> +void TextureCache<P>::UnregisterImage(ImageId image_id) { + Image& image = slot_images[image_id]; + ASSERT_MSG(True(image.flags & ImageFlagBits::Registered), + "Trying to unregister an already registered image"); + image.flags &= ~ImageFlagBits::Registered; + ForEachPage(image.cpu_addr, image.guest_size_bytes, [this, image_id](u64 page) { + const auto page_it = page_table.find(page); + if (page_it == page_table.end()) { + UNREACHABLE_MSG("Unregistering unregistered page=0x{:x}", page << PAGE_BITS); + return; } - for (auto& surface : search->second) { - if (!surface->IsRegistered()) { - return surface; - } + std::vector<ImageId>& image_ids = page_it->second; + const auto vector_it = std::ranges::find(image_ids, image_id); + if (vector_it == image_ids.end()) { + UNREACHABLE_MSG("Unregistering unregistered image in page=0x{:x}", page << PAGE_BITS); + return; } - return {}; - } + image_ids.erase(vector_it); + }); +} - /// Try to do an image copy logging when formats are incompatible. - void TryCopyImage(TSurface& src, TSurface& dst, const CopyParams& copy) { - const SurfaceParams& src_params = src->GetSurfaceParams(); - const SurfaceParams& dst_params = dst->GetSurfaceParams(); - if (!format_compatibility.TestCopy(src_params.pixel_format, dst_params.pixel_format)) { - LOG_ERROR(HW_GPU, "Illegal copy between formats={{{}, {}}}", - static_cast<int>(dst_params.pixel_format), - static_cast<int>(src_params.pixel_format)); - return; +template <class P> +void TextureCache<P>::TrackImage(ImageBase& image) { + ASSERT(False(image.flags & ImageFlagBits::Tracked)); + image.flags |= ImageFlagBits::Tracked; + rasterizer.UpdatePagesCachedCount(image.cpu_addr, image.guest_size_bytes, 1); +} + +template <class P> +void TextureCache<P>::UntrackImage(ImageBase& image) { + ASSERT(True(image.flags & ImageFlagBits::Tracked)); + image.flags &= ~ImageFlagBits::Tracked; + rasterizer.UpdatePagesCachedCount(image.cpu_addr, image.guest_size_bytes, -1); +} + +template <class P> +void TextureCache<P>::DeleteImage(ImageId image_id) { + ImageBase& image = slot_images[image_id]; + const GPUVAddr gpu_addr = image.gpu_addr; + const auto alloc_it = image_allocs_table.find(gpu_addr); + if (alloc_it == image_allocs_table.end()) { + UNREACHABLE_MSG("Trying to delete an image alloc that does not exist in address 0x{:x}", + gpu_addr); + return; + } + const ImageAllocId alloc_id = alloc_it->second; + std::vector<ImageId>& alloc_images = slot_image_allocs[alloc_id].images; + const auto alloc_image_it = std::ranges::find(alloc_images, image_id); + if (alloc_image_it == alloc_images.end()) { + UNREACHABLE_MSG("Trying to delete an image that does not exist"); + return; + } + ASSERT_MSG(False(image.flags & ImageFlagBits::Tracked), "Image was not untracked"); + ASSERT_MSG(False(image.flags & ImageFlagBits::Registered), "Image was not unregistered"); + + // Mark render targets as dirty + auto& dirty = maxwell3d.dirty.flags; + dirty[Dirty::RenderTargets] = true; + dirty[Dirty::ZetaBuffer] = true; + for (size_t rt = 0; rt < NUM_RT; ++rt) { + dirty[Dirty::ColorBuffer0 + rt] = true; + } + const std::span<const ImageViewId> image_view_ids = image.image_view_ids; + for (const ImageViewId image_view_id : image_view_ids) { + std::ranges::replace(render_targets.color_buffer_ids, image_view_id, ImageViewId{}); + if (render_targets.depth_buffer_id == image_view_id) { + render_targets.depth_buffer_id = ImageViewId{}; } - ImageCopy(src, dst, copy); } + RemoveImageViewReferences(image_view_ids); + RemoveFramebuffers(image_view_ids); + + for (const AliasedImage& alias : image.aliased_images) { + ImageBase& other_image = slot_images[alias.id]; + [[maybe_unused]] const size_t num_removed_aliases = + std::erase_if(other_image.aliased_images, [image_id](const AliasedImage& other_alias) { + return other_alias.id == image_id; + }); + ASSERT_MSG(num_removed_aliases == 1, "Invalid number of removed aliases: {}", + num_removed_aliases); + } + for (const ImageViewId image_view_id : image_view_ids) { + sentenced_image_view.Push(std::move(slot_image_views[image_view_id])); + slot_image_views.erase(image_view_id); + } + sentenced_images.Push(std::move(slot_images[image_id])); + slot_images.erase(image_id); - constexpr PixelFormat GetSiblingFormat(PixelFormat format) const { - return siblings_table[static_cast<std::size_t>(format)]; + alloc_images.erase(alloc_image_it); + if (alloc_images.empty()) { + image_allocs_table.erase(alloc_it); } + if constexpr (ENABLE_VALIDATION) { + std::ranges::fill(graphics_image_view_ids, CORRUPT_ID); + std::ranges::fill(compute_image_view_ids, CORRUPT_ID); + } + graphics_image_table.Invalidate(); + compute_image_table.Invalidate(); + has_deleted_images = true; +} - /// Returns true the shader sampler entry is compatible with the TIC texture type. - static bool IsTypeCompatible(Tegra::Texture::TextureType tic_type, - const VideoCommon::Shader::Sampler& entry) { - const auto shader_type = entry.type; - switch (tic_type) { - case Tegra::Texture::TextureType::Texture1D: - case Tegra::Texture::TextureType::Texture1DArray: - return shader_type == Tegra::Shader::TextureType::Texture1D; - case Tegra::Texture::TextureType::Texture1DBuffer: - // TODO(Rodrigo): Assume as valid for now - return true; - case Tegra::Texture::TextureType::Texture2D: - case Tegra::Texture::TextureType::Texture2DNoMipmap: - return shader_type == Tegra::Shader::TextureType::Texture2D; - case Tegra::Texture::TextureType::Texture2DArray: - return shader_type == Tegra::Shader::TextureType::Texture2D || - shader_type == Tegra::Shader::TextureType::TextureCube; - case Tegra::Texture::TextureType::Texture3D: - return shader_type == Tegra::Shader::TextureType::Texture3D; - case Tegra::Texture::TextureType::TextureCubeArray: - case Tegra::Texture::TextureType::TextureCubemap: - if (shader_type == Tegra::Shader::TextureType::TextureCube) { - return true; - } - return shader_type == Tegra::Shader::TextureType::Texture2D && entry.is_array; +template <class P> +void TextureCache<P>::RemoveImageViewReferences(std::span<const ImageViewId> removed_views) { + auto it = image_views.begin(); + while (it != image_views.end()) { + const auto found = std::ranges::find(removed_views, it->second); + if (found != removed_views.end()) { + it = image_views.erase(it); + } else { + ++it; } - UNREACHABLE(); - return true; } +} - struct FramebufferTargetInfo { - TSurface target; - TView view; - }; - - void AsyncFlushSurface(TSurface& surface) { - if (!uncommitted_flushes) { - uncommitted_flushes = std::make_shared<std::list<TSurface>>(); +template <class P> +void TextureCache<P>::RemoveFramebuffers(std::span<const ImageViewId> removed_views) { + auto it = framebuffers.begin(); + while (it != framebuffers.end()) { + if (it->first.Contains(removed_views)) { + it = framebuffers.erase(it); + } else { + ++it; } - uncommitted_flushes->push_back(surface); } +} - VideoCore::RasterizerInterface& rasterizer; - Tegra::Engines::Maxwell3D& maxwell3d; - Tegra::MemoryManager& gpu_memory; - - FormatLookupTable format_lookup_table; - FormatCompatibility format_compatibility; - - u64 ticks{}; - - // Guards the cache for protection conflicts. - bool guard_render_targets{}; - bool guard_samplers{}; - - // The siblings table is for formats that can inter exchange with one another - // without causing issues. This is only valid when a conflict occurs on a non - // rendering use. - std::array<PixelFormat, static_cast<std::size_t>(PixelFormat::Max)> siblings_table; - - // The internal Cache is different for the Texture Cache. It's based on buckets - // of 1MB. This fits better for the purpose of this cache as textures are normaly - // large in size. - static constexpr u64 registry_page_bits{20}; - static constexpr u64 registry_page_size{1 << registry_page_bits}; - std::unordered_map<VAddr, std::vector<TSurface>> registry; +template <class P> +void TextureCache<P>::MarkModification(ImageBase& image) noexcept { + image.flags |= ImageFlagBits::GpuModified; + image.modification_tick = ++modification_tick; +} - static constexpr u32 DEPTH_RT = 8; - static constexpr u32 NO_RT = 0xFFFFFFFF; +template <class P> +void TextureCache<P>::SynchronizeAliases(ImageId image_id) { + boost::container::small_vector<const AliasedImage*, 1> aliased_images; + ImageBase& image = slot_images[image_id]; + u64 most_recent_tick = image.modification_tick; + for (const AliasedImage& aliased : image.aliased_images) { + ImageBase& aliased_image = slot_images[aliased.id]; + if (image.modification_tick < aliased_image.modification_tick) { + most_recent_tick = std::max(most_recent_tick, aliased_image.modification_tick); + aliased_images.push_back(&aliased); + } + } + if (aliased_images.empty()) { + return; + } + image.modification_tick = most_recent_tick; + std::ranges::sort(aliased_images, [this](const AliasedImage* lhs, const AliasedImage* rhs) { + const ImageBase& lhs_image = slot_images[lhs->id]; + const ImageBase& rhs_image = slot_images[rhs->id]; + return lhs_image.modification_tick < rhs_image.modification_tick; + }); + for (const AliasedImage* const aliased : aliased_images) { + CopyImage(image_id, aliased->id, aliased->copies); + } +} - // The L1 Cache is used for fast texture lookup before checking the overlaps - // This avoids calculating size and other stuffs. - std::unordered_map<VAddr, TSurface> l1_cache; +template <class P> +void TextureCache<P>::PrepareImage(ImageId image_id, bool is_modification, bool invalidate) { + Image& image = slot_images[image_id]; + if (invalidate) { + image.flags &= ~(ImageFlagBits::CpuModified | ImageFlagBits::GpuModified); + if (False(image.flags & ImageFlagBits::Tracked)) { + TrackImage(image); + } + } else { + RefreshContents(image); + SynchronizeAliases(image_id); + } + if (is_modification) { + MarkModification(image); + } + image.frame_tick = frame_tick; +} - /// The surface reserve is a "backup" cache, this is where we put unique surfaces that have - /// previously been used. This is to prevent surfaces from being constantly created and - /// destroyed when used with different surface parameters. - std::unordered_map<SurfaceParams, std::vector<TSurface>> surface_reserve; - std::array<FramebufferTargetInfo, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> - render_targets; - FramebufferTargetInfo depth_buffer; +template <class P> +void TextureCache<P>::PrepareImageView(ImageViewId image_view_id, bool is_modification, + bool invalidate) { + if (!image_view_id) { + return; + } + const ImageViewBase& image_view = slot_image_views[image_view_id]; + PrepareImage(image_view.image_id, is_modification, invalidate); +} - std::vector<TSurface> sampled_textures; +template <class P> +void TextureCache<P>::CopyImage(ImageId dst_id, ImageId src_id, std::span<const ImageCopy> copies) { + Image& dst = slot_images[dst_id]; + Image& src = slot_images[src_id]; + const auto dst_format_type = GetFormatType(dst.info.format); + const auto src_format_type = GetFormatType(src.info.format); + if (src_format_type == dst_format_type) { + if constexpr (HAS_EMULATED_COPIES) { + if (!runtime.CanImageBeCopied(dst, src)) { + return runtime.EmulateCopyImage(dst, src, copies); + } + } + return runtime.CopyImage(dst, src, copies); + } + UNIMPLEMENTED_IF(dst.info.type != ImageType::e2D); + UNIMPLEMENTED_IF(src.info.type != ImageType::e2D); + for (const ImageCopy& copy : copies) { + UNIMPLEMENTED_IF(copy.dst_subresource.num_layers != 1); + UNIMPLEMENTED_IF(copy.src_subresource.num_layers != 1); + UNIMPLEMENTED_IF(copy.src_offset != Offset3D{}); + UNIMPLEMENTED_IF(copy.dst_offset != Offset3D{}); + + const SubresourceBase dst_base{ + .level = copy.dst_subresource.base_level, + .layer = copy.dst_subresource.base_layer, + }; + const SubresourceBase src_base{ + .level = copy.src_subresource.base_level, + .layer = copy.src_subresource.base_layer, + }; + const SubresourceExtent dst_extent{.levels = 1, .layers = 1}; + const SubresourceExtent src_extent{.levels = 1, .layers = 1}; + const SubresourceRange dst_range{.base = dst_base, .extent = dst_extent}; + const SubresourceRange src_range{.base = src_base, .extent = src_extent}; + const ImageViewInfo dst_view_info(ImageViewType::e2D, dst.info.format, dst_range); + const ImageViewInfo src_view_info(ImageViewType::e2D, src.info.format, src_range); + const auto [dst_framebuffer_id, dst_view_id] = RenderTargetFromImage(dst_id, dst_view_info); + Framebuffer* const dst_framebuffer = &slot_framebuffers[dst_framebuffer_id]; + const ImageViewId src_view_id = FindOrEmplaceImageView(src_id, src_view_info); + ImageView& dst_view = slot_image_views[dst_view_id]; + ImageView& src_view = slot_image_views[src_view_id]; + [[maybe_unused]] const Extent3D expected_size{ + .width = std::min(dst_view.size.width, src_view.size.width), + .height = std::min(dst_view.size.height, src_view.size.height), + .depth = std::min(dst_view.size.depth, src_view.size.depth), + }; + UNIMPLEMENTED_IF(copy.extent != expected_size); - /// This cache stores null surfaces in order to be used as a placeholder - /// for invalid texture calls. - std::unordered_map<u32, TSurface> invalid_cache; - std::vector<u8> invalid_memory; + runtime.ConvertImage(dst_framebuffer, dst_view, src_view); + } +} - std::list<TSurface> marked_for_unregister; +template <class P> +void TextureCache<P>::BindRenderTarget(ImageViewId* old_id, ImageViewId new_id) { + if (*old_id == new_id) { + return; + } + if (*old_id) { + const ImageViewBase& old_view = slot_image_views[*old_id]; + if (True(old_view.flags & ImageViewFlagBits::PreemtiveDownload)) { + uncommitted_downloads.push_back(old_view.image_id); + } + } + *old_id = new_id; +} - std::shared_ptr<std::list<TSurface>> uncommitted_flushes{}; - std::list<std::shared_ptr<std::list<TSurface>>> committed_flushes; +template <class P> +std::pair<FramebufferId, ImageViewId> TextureCache<P>::RenderTargetFromImage( + ImageId image_id, const ImageViewInfo& view_info) { + const ImageViewId view_id = FindOrEmplaceImageView(image_id, view_info); + const ImageBase& image = slot_images[image_id]; + const bool is_color = GetFormatType(image.info.format) == SurfaceType::ColorTexture; + const ImageViewId color_view_id = is_color ? view_id : ImageViewId{}; + const ImageViewId depth_view_id = is_color ? ImageViewId{} : view_id; + const Extent3D extent = MipSize(image.info.size, view_info.range.base.level); + const u32 num_samples = image.info.num_samples; + const auto [samples_x, samples_y] = SamplesLog2(num_samples); + const FramebufferId framebuffer_id = GetFramebufferId(RenderTargets{ + .color_buffer_ids = {color_view_id}, + .depth_buffer_id = depth_view_id, + .size = {extent.width >> samples_x, extent.height >> samples_y}, + }); + return {framebuffer_id, view_id}; +} - StagingCache staging_cache; - std::recursive_mutex mutex; -}; +template <class P> +bool TextureCache<P>::IsFullClear(ImageViewId id) { + if (!id) { + return true; + } + const ImageViewBase& image_view = slot_image_views[id]; + const ImageBase& image = slot_images[image_view.image_id]; + const Extent3D size = image_view.size; + const auto& regs = maxwell3d.regs; + const auto& scissor = regs.scissor_test[0]; + if (image.info.resources.levels > 1 || image.info.resources.layers > 1) { + // Images with multiple resources can't be cleared in a single call + return false; + } + if (regs.clear_flags.scissor == 0) { + // If scissor testing is disabled, the clear is always full + return true; + } + // Make sure the clear covers all texels in the subresource + return scissor.min_x == 0 && scissor.min_y == 0 && scissor.max_x >= size.width && + scissor.max_y >= size.height; +} } // namespace VideoCommon diff --git a/src/video_core/texture_cache/types.h b/src/video_core/texture_cache/types.h new file mode 100644 index 000000000..2ad2d72a6 --- /dev/null +++ b/src/video_core/texture_cache/types.h @@ -0,0 +1,140 @@ +// Copyright 2019 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_funcs.h" +#include "common/common_types.h" +#include "video_core/texture_cache/slot_vector.h" + +namespace VideoCommon { + +constexpr size_t NUM_RT = 8; +constexpr size_t MAX_MIP_LEVELS = 14; + +constexpr SlotId CORRUPT_ID{0xfffffffe}; + +using ImageId = SlotId; +using ImageViewId = SlotId; +using ImageAllocId = SlotId; +using SamplerId = SlotId; +using FramebufferId = SlotId; + +enum class ImageType : u32 { + e1D, + e2D, + e3D, + Linear, + Buffer, +}; + +enum class ImageViewType : u32 { + e1D, + e2D, + Cube, + e3D, + e1DArray, + e2DArray, + CubeArray, + Rect, + Buffer, +}; +constexpr size_t NUM_IMAGE_VIEW_TYPES = 9; + +enum class RelaxedOptions : u32 { + Size = 1 << 0, + Format = 1 << 1, + Samples = 1 << 2, +}; +DECLARE_ENUM_FLAG_OPERATORS(RelaxedOptions) + +struct Offset2D { + constexpr auto operator<=>(const Offset2D&) const noexcept = default; + + s32 x; + s32 y; +}; + +struct Offset3D { + constexpr auto operator<=>(const Offset3D&) const noexcept = default; + + s32 x; + s32 y; + s32 z; +}; + +struct Extent2D { + constexpr auto operator<=>(const Extent2D&) const noexcept = default; + + u32 width; + u32 height; +}; + +struct Extent3D { + constexpr auto operator<=>(const Extent3D&) const noexcept = default; + + u32 width; + u32 height; + u32 depth; +}; + +struct SubresourceLayers { + s32 base_level = 0; + s32 base_layer = 0; + s32 num_layers = 1; +}; + +struct SubresourceBase { + constexpr auto operator<=>(const SubresourceBase&) const noexcept = default; + + s32 level = 0; + s32 layer = 0; +}; + +struct SubresourceExtent { + constexpr auto operator<=>(const SubresourceExtent&) const noexcept = default; + + s32 levels = 1; + s32 layers = 1; +}; + +struct SubresourceRange { + constexpr auto operator<=>(const SubresourceRange&) const noexcept = default; + + SubresourceBase base; + SubresourceExtent extent; +}; + +struct ImageCopy { + SubresourceLayers src_subresource; + SubresourceLayers dst_subresource; + Offset3D src_offset; + Offset3D dst_offset; + Extent3D extent; +}; + +struct BufferImageCopy { + size_t buffer_offset; + size_t buffer_size; + u32 buffer_row_length; + u32 buffer_image_height; + SubresourceLayers image_subresource; + Offset3D image_offset; + Extent3D image_extent; +}; + +struct BufferCopy { + size_t src_offset; + size_t dst_offset; + size_t size; +}; + +struct SwizzleParameters { + Extent3D num_tiles; + Extent3D block; + size_t buffer_offset; + s32 level; +}; + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/util.cpp b/src/video_core/texture_cache/util.cpp new file mode 100644 index 000000000..279932778 --- /dev/null +++ b/src/video_core/texture_cache/util.cpp @@ -0,0 +1,1233 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +// This files contains code from Ryujinx +// A copy of the code can be obtained from https://github.com/Ryujinx/Ryujinx +// The sections using code from Ryujinx are marked with a link to the original version + +// MIT License +// +// Copyright (c) Ryujinx Team and Contributors +// +// Permission is hereby granted, free of charge, to any person obtaining a copy of this software and +// associated documentation files (the "Software"), to deal in the Software without restriction, +// including without limitation the rights to use, copy, modify, merge, publish, distribute, +// sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is +// furnished to do so, subject to the following conditions: +// +// The above copyright notice and this permission notice shall be included in all copies or +// substantial portions of the Software. +// +// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT +// NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND +// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, +// DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +// + +#include <algorithm> +#include <array> +#include <numeric> +#include <optional> +#include <span> +#include <vector> + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/bit_util.h" +#include "common/common_types.h" +#include "common/div_ceil.h" +#include "video_core/compatible_formats.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/memory_manager.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/decode_bc4.h" +#include "video_core/texture_cache/format_lookup_table.h" +#include "video_core/texture_cache/formatter.h" +#include "video_core/texture_cache/samples_helper.h" +#include "video_core/texture_cache/util.h" +#include "video_core/textures/astc.h" +#include "video_core/textures/decoders.h" + +namespace VideoCommon { + +namespace { + +using Tegra::Texture::GOB_SIZE; +using Tegra::Texture::GOB_SIZE_SHIFT; +using Tegra::Texture::GOB_SIZE_X; +using Tegra::Texture::GOB_SIZE_X_SHIFT; +using Tegra::Texture::GOB_SIZE_Y; +using Tegra::Texture::GOB_SIZE_Y_SHIFT; +using Tegra::Texture::GOB_SIZE_Z; +using Tegra::Texture::GOB_SIZE_Z_SHIFT; +using Tegra::Texture::MsaaMode; +using Tegra::Texture::SwizzleTexture; +using Tegra::Texture::TextureFormat; +using Tegra::Texture::TextureType; +using Tegra::Texture::TICEntry; +using Tegra::Texture::UnswizzleTexture; +using VideoCore::Surface::BytesPerBlock; +using VideoCore::Surface::DefaultBlockHeight; +using VideoCore::Surface::DefaultBlockWidth; +using VideoCore::Surface::IsCopyCompatible; +using VideoCore::Surface::IsPixelFormatASTC; +using VideoCore::Surface::IsViewCompatible; +using VideoCore::Surface::PixelFormatFromDepthFormat; +using VideoCore::Surface::PixelFormatFromRenderTargetFormat; +using VideoCore::Surface::SurfaceType; + +constexpr u32 CONVERTED_BYTES_PER_BLOCK = BytesPerBlock(PixelFormat::A8B8G8R8_UNORM); + +struct LevelInfo { + Extent3D size; + Extent3D block; + Extent2D tile_size; + u32 bpp_log2; + u32 tile_width_spacing; +}; + +[[nodiscard]] constexpr u32 AdjustTileSize(u32 shift, u32 unit_factor, u32 dimension) { + if (shift == 0) { + return 0; + } + u32 x = unit_factor << (shift - 1); + if (x >= dimension) { + while (--shift) { + x >>= 1; + if (x < dimension) { + break; + } + } + } + return shift; +} + +[[nodiscard]] constexpr u32 AdjustMipSize(u32 size, u32 level) { + return std::max<u32>(size >> level, 1); +} + +[[nodiscard]] constexpr Extent3D AdjustMipSize(Extent3D size, s32 level) { + return Extent3D{ + .width = AdjustMipSize(size.width, level), + .height = AdjustMipSize(size.height, level), + .depth = AdjustMipSize(size.depth, level), + }; +} + +[[nodiscard]] Extent3D AdjustSamplesSize(Extent3D size, s32 num_samples) { + const auto [samples_x, samples_y] = SamplesLog2(num_samples); + return Extent3D{ + .width = size.width >> samples_x, + .height = size.height >> samples_y, + .depth = size.depth, + }; +} + +template <u32 GOB_EXTENT> +[[nodiscard]] constexpr u32 AdjustMipBlockSize(u32 num_tiles, u32 block_size, u32 level) { + do { + while (block_size > 0 && num_tiles <= (1U << (block_size - 1)) * GOB_EXTENT) { + --block_size; + } + } while (level--); + return block_size; +} + +[[nodiscard]] constexpr Extent3D AdjustMipBlockSize(Extent3D num_tiles, Extent3D block_size, + u32 level) { + return { + .width = AdjustMipBlockSize<GOB_SIZE_X>(num_tiles.width, block_size.width, level), + .height = AdjustMipBlockSize<GOB_SIZE_Y>(num_tiles.height, block_size.height, level), + .depth = AdjustMipBlockSize<GOB_SIZE_Z>(num_tiles.depth, block_size.depth, level), + }; +} + +[[nodiscard]] constexpr Extent3D AdjustTileSize(Extent3D size, Extent2D tile_size) { + return { + .width = Common::DivCeil(size.width, tile_size.width), + .height = Common::DivCeil(size.height, tile_size.height), + .depth = size.depth, + }; +} + +[[nodiscard]] constexpr u32 BytesPerBlockLog2(u32 bytes_per_block) { + return std::countl_zero(bytes_per_block) ^ 0x1F; +} + +[[nodiscard]] constexpr u32 BytesPerBlockLog2(PixelFormat format) { + return BytesPerBlockLog2(BytesPerBlock(format)); +} + +[[nodiscard]] constexpr u32 NumBlocks(Extent3D size, Extent2D tile_size) { + const Extent3D num_blocks = AdjustTileSize(size, tile_size); + return num_blocks.width * num_blocks.height * num_blocks.depth; +} + +[[nodiscard]] constexpr u32 AdjustSize(u32 size, u32 level, u32 block_size) { + return Common::DivCeil(AdjustMipSize(size, level), block_size); +} + +[[nodiscard]] constexpr u32 LayerSize(const TICEntry& config, PixelFormat format) { + return config.Width() * config.Height() * BytesPerBlock(format); +} + +[[nodiscard]] constexpr bool HasTwoDimsPerLayer(TextureType type) { + switch (type) { + case TextureType::Texture2D: + case TextureType::Texture2DArray: + case TextureType::Texture2DNoMipmap: + case TextureType::Texture3D: + case TextureType::TextureCubeArray: + case TextureType::TextureCubemap: + return true; + case TextureType::Texture1D: + case TextureType::Texture1DArray: + case TextureType::Texture1DBuffer: + return false; + } + return false; +} + +[[nodiscard]] constexpr bool HasTwoDimsPerLayer(ImageType type) { + switch (type) { + case ImageType::e2D: + case ImageType::e3D: + case ImageType::Linear: + return true; + case ImageType::e1D: + case ImageType::Buffer: + return false; + } + UNREACHABLE_MSG("Invalid image type={}", static_cast<int>(type)); +} + +[[nodiscard]] constexpr std::pair<int, int> Samples(int num_samples) { + switch (num_samples) { + case 1: + return {1, 1}; + case 2: + return {2, 1}; + case 4: + return {2, 2}; + case 8: + return {4, 2}; + case 16: + return {4, 4}; + } + UNREACHABLE_MSG("Invalid number of samples={}", num_samples); + return {1, 1}; +} + +[[nodiscard]] constexpr Extent2D DefaultBlockSize(PixelFormat format) { + return {DefaultBlockWidth(format), DefaultBlockHeight(format)}; +} + +[[nodiscard]] constexpr Extent3D NumLevelBlocks(const LevelInfo& info, u32 level) { + return Extent3D{ + .width = AdjustSize(info.size.width, level, info.tile_size.width) << info.bpp_log2, + .height = AdjustSize(info.size.height, level, info.tile_size.height), + .depth = AdjustMipSize(info.size.depth, level), + }; +} + +[[nodiscard]] constexpr Extent3D TileShift(const LevelInfo& info, u32 level) { + const Extent3D blocks = NumLevelBlocks(info, level); + return Extent3D{ + .width = AdjustTileSize(info.block.width, GOB_SIZE_X, blocks.width), + .height = AdjustTileSize(info.block.height, GOB_SIZE_Y, blocks.height), + .depth = AdjustTileSize(info.block.depth, GOB_SIZE_Z, blocks.depth), + }; +} + +[[nodiscard]] constexpr Extent2D GobSize(u32 bpp_log2, u32 block_height, u32 tile_width_spacing) { + return Extent2D{ + .width = GOB_SIZE_X_SHIFT - bpp_log2 + tile_width_spacing, + .height = GOB_SIZE_Y_SHIFT + block_height, + }; +} + +[[nodiscard]] constexpr bool IsSmallerThanGobSize(Extent3D num_tiles, Extent2D gob, + u32 block_depth) { + return num_tiles.width <= (1U << gob.width) || num_tiles.height <= (1U << gob.height) || + num_tiles.depth < (1U << block_depth); +} + +[[nodiscard]] constexpr u32 StrideAlignment(Extent3D num_tiles, Extent3D block, Extent2D gob, + u32 bpp_log2) { + if (IsSmallerThanGobSize(num_tiles, gob, block.depth)) { + return GOB_SIZE_X_SHIFT - bpp_log2; + } else { + return gob.width; + } +} + +[[nodiscard]] constexpr u32 StrideAlignment(Extent3D num_tiles, Extent3D block, u32 bpp_log2, + u32 tile_width_spacing) { + const Extent2D gob = GobSize(bpp_log2, block.height, tile_width_spacing); + return StrideAlignment(num_tiles, block, gob, bpp_log2); +} + +[[nodiscard]] constexpr Extent2D NumGobs(const LevelInfo& info, u32 level) { + const Extent3D blocks = NumLevelBlocks(info, level); + const Extent2D gobs{ + .width = Common::DivCeilLog2(blocks.width, GOB_SIZE_X_SHIFT), + .height = Common::DivCeilLog2(blocks.height, GOB_SIZE_Y_SHIFT), + }; + const Extent2D gob = GobSize(info.bpp_log2, info.block.height, info.tile_width_spacing); + const bool is_small = IsSmallerThanGobSize(blocks, gob, info.block.depth); + const u32 alignment = is_small ? 0 : info.tile_width_spacing; + return Extent2D{ + .width = Common::AlignBits(gobs.width, alignment), + .height = gobs.height, + }; +} + +[[nodiscard]] constexpr Extent3D LevelTiles(const LevelInfo& info, u32 level) { + const Extent3D blocks = NumLevelBlocks(info, level); + const Extent3D tile_shift = TileShift(info, level); + const Extent2D gobs = NumGobs(info, level); + return Extent3D{ + .width = Common::DivCeilLog2(gobs.width, tile_shift.width), + .height = Common::DivCeilLog2(gobs.height, tile_shift.height), + .depth = Common::DivCeilLog2(blocks.depth, tile_shift.depth), + }; +} + +[[nodiscard]] constexpr u32 CalculateLevelSize(const LevelInfo& info, u32 level) { + const Extent3D tile_shift = TileShift(info, level); + const Extent3D tiles = LevelTiles(info, level); + const u32 num_tiles = tiles.width * tiles.height * tiles.depth; + const u32 shift = GOB_SIZE_SHIFT + tile_shift.width + tile_shift.height + tile_shift.depth; + return num_tiles << shift; +} + +[[nodiscard]] constexpr std::array<u32, MAX_MIP_LEVELS> CalculateLevelSizes(const LevelInfo& info, + u32 num_levels) { + ASSERT(num_levels <= MAX_MIP_LEVELS); + std::array<u32, MAX_MIP_LEVELS> sizes{}; + for (u32 level = 0; level < num_levels; ++level) { + sizes[level] = CalculateLevelSize(info, level); + } + return sizes; +} + +[[nodiscard]] constexpr LevelInfo MakeLevelInfo(PixelFormat format, Extent3D size, Extent3D block, + u32 num_samples, u32 tile_width_spacing) { + const auto [samples_x, samples_y] = Samples(num_samples); + const u32 bytes_per_block = BytesPerBlock(format); + return { + .size = + { + .width = size.width * samples_x, + .height = size.height * samples_y, + .depth = size.depth, + }, + .block = block, + .tile_size = DefaultBlockSize(format), + .bpp_log2 = BytesPerBlockLog2(bytes_per_block), + .tile_width_spacing = tile_width_spacing, + }; +} + +[[nodiscard]] constexpr LevelInfo MakeLevelInfo(const ImageInfo& info) { + return MakeLevelInfo(info.format, info.size, info.block, info.num_samples, + info.tile_width_spacing); +} + +[[nodiscard]] constexpr u32 CalculateLevelOffset(PixelFormat format, Extent3D size, Extent3D block, + u32 num_samples, u32 tile_width_spacing, + u32 level) { + const LevelInfo info = MakeLevelInfo(format, size, block, num_samples, tile_width_spacing); + u32 offset = 0; + for (u32 current_level = 0; current_level < level; ++current_level) { + offset += CalculateLevelSize(info, current_level); + } + return offset; +} + +[[nodiscard]] constexpr u32 AlignLayerSize(u32 size_bytes, Extent3D size, Extent3D block, + u32 tile_size_y, u32 tile_width_spacing) { + // https://github.com/Ryujinx/Ryujinx/blob/1c9aba6de1520aea5480c032e0ff5664ac1bb36f/Ryujinx.Graphics.Texture/SizeCalculator.cs#L134 + if (tile_width_spacing > 0) { + const u32 alignment_log2 = GOB_SIZE_SHIFT + tile_width_spacing + block.height + block.depth; + return Common::AlignBits(size_bytes, alignment_log2); + } + const u32 aligned_height = Common::AlignUp(size.height, tile_size_y); + while (block.height != 0 && aligned_height <= (1U << (block.height - 1)) * GOB_SIZE_Y) { + --block.height; + } + while (block.depth != 0 && size.depth <= (1U << (block.depth - 1))) { + --block.depth; + } + const u32 block_shift = GOB_SIZE_SHIFT + block.height + block.depth; + const u32 num_blocks = size_bytes >> block_shift; + if (size_bytes != num_blocks << block_shift) { + return (num_blocks + 1) << block_shift; + } + return size_bytes; +} + +[[nodiscard]] std::optional<SubresourceExtent> ResolveOverlapEqualAddress(const ImageInfo& new_info, + const ImageBase& overlap, + bool strict_size) { + const ImageInfo& info = overlap.info; + if (!IsBlockLinearSizeCompatible(new_info, info, 0, 0, strict_size)) { + return std::nullopt; + } + if (new_info.block != info.block) { + return std::nullopt; + } + const SubresourceExtent resources = new_info.resources; + return SubresourceExtent{ + .levels = std::max(resources.levels, info.resources.levels), + .layers = std::max(resources.layers, info.resources.layers), + }; +} + +[[nodiscard]] std::optional<SubresourceExtent> ResolveOverlapRightAddress3D( + const ImageInfo& new_info, GPUVAddr gpu_addr, const ImageBase& overlap, bool strict_size) { + const std::vector<u32> slice_offsets = CalculateSliceOffsets(new_info); + const u32 diff = static_cast<u32>(overlap.gpu_addr - gpu_addr); + const auto it = std::ranges::find(slice_offsets, diff); + if (it == slice_offsets.end()) { + return std::nullopt; + } + const std::vector subresources = CalculateSliceSubresources(new_info); + const SubresourceBase base = subresources[std::distance(slice_offsets.begin(), it)]; + const ImageInfo& info = overlap.info; + if (!IsBlockLinearSizeCompatible(new_info, info, base.level, 0, strict_size)) { + return std::nullopt; + } + const u32 mip_depth = std::max(1U, new_info.size.depth << base.level); + if (mip_depth < info.size.depth + base.layer) { + return std::nullopt; + } + if (MipBlockSize(new_info, base.level) != info.block) { + return std::nullopt; + } + return SubresourceExtent{ + .levels = std::max(new_info.resources.levels, info.resources.levels + base.level), + .layers = 1, + }; +} + +[[nodiscard]] std::optional<SubresourceExtent> ResolveOverlapRightAddress2D( + const ImageInfo& new_info, GPUVAddr gpu_addr, const ImageBase& overlap, bool strict_size) { + const u32 layer_stride = new_info.layer_stride; + const s32 new_size = layer_stride * new_info.resources.layers; + const s32 diff = static_cast<s32>(overlap.gpu_addr - gpu_addr); + if (diff > new_size) { + return std::nullopt; + } + const s32 base_layer = diff / layer_stride; + const s32 mip_offset = diff % layer_stride; + const std::array offsets = CalculateMipLevelOffsets(new_info); + const auto end = offsets.begin() + new_info.resources.levels; + const auto it = std::find(offsets.begin(), end, mip_offset); + if (it == end) { + // Mipmap is not aligned to any valid size + return std::nullopt; + } + const SubresourceBase base{ + .level = static_cast<s32>(std::distance(offsets.begin(), it)), + .layer = base_layer, + }; + const ImageInfo& info = overlap.info; + if (!IsBlockLinearSizeCompatible(new_info, info, base.level, 0, strict_size)) { + return std::nullopt; + } + if (MipBlockSize(new_info, base.level) != info.block) { + return std::nullopt; + } + return SubresourceExtent{ + .levels = std::max(new_info.resources.levels, info.resources.levels + base.level), + .layers = std::max(new_info.resources.layers, info.resources.layers + base.layer), + }; +} + +[[nodiscard]] std::optional<OverlapResult> ResolveOverlapRightAddress(const ImageInfo& new_info, + GPUVAddr gpu_addr, + VAddr cpu_addr, + const ImageBase& overlap, + bool strict_size) { + std::optional<SubresourceExtent> resources; + if (new_info.type != ImageType::e3D) { + resources = ResolveOverlapRightAddress2D(new_info, gpu_addr, overlap, strict_size); + } else { + resources = ResolveOverlapRightAddress3D(new_info, gpu_addr, overlap, strict_size); + } + if (!resources) { + return std::nullopt; + } + return OverlapResult{ + .gpu_addr = gpu_addr, + .cpu_addr = cpu_addr, + .resources = *resources, + }; +} + +[[nodiscard]] std::optional<OverlapResult> ResolveOverlapLeftAddress(const ImageInfo& new_info, + GPUVAddr gpu_addr, + VAddr cpu_addr, + const ImageBase& overlap, + bool strict_size) { + const std::optional<SubresourceBase> base = overlap.TryFindBase(gpu_addr); + if (!base) { + return std::nullopt; + } + const ImageInfo& info = overlap.info; + if (!IsBlockLinearSizeCompatible(new_info, info, base->level, 0, strict_size)) { + return std::nullopt; + } + if (new_info.block != MipBlockSize(info, base->level)) { + return std::nullopt; + } + const SubresourceExtent resources = new_info.resources; + s32 layers = 1; + if (info.type != ImageType::e3D) { + layers = std::max(resources.layers, info.resources.layers + base->layer); + } + return OverlapResult{ + .gpu_addr = overlap.gpu_addr, + .cpu_addr = overlap.cpu_addr, + .resources = + { + .levels = std::max(resources.levels + base->level, info.resources.levels), + .layers = layers, + }, + }; +} + +[[nodiscard]] Extent2D PitchLinearAlignedSize(const ImageInfo& info) { + // https://github.com/Ryujinx/Ryujinx/blob/1c9aba6de1520aea5480c032e0ff5664ac1bb36f/Ryujinx.Graphics.Texture/SizeCalculator.cs#L212 + static constexpr u32 STRIDE_ALIGNMENT = 32; + ASSERT(info.type == ImageType::Linear); + const Extent2D num_tiles{ + .width = Common::DivCeil(info.size.width, DefaultBlockWidth(info.format)), + .height = Common::DivCeil(info.size.height, DefaultBlockHeight(info.format)), + }; + const u32 width_alignment = STRIDE_ALIGNMENT / BytesPerBlock(info.format); + return Extent2D{ + .width = Common::AlignUp(num_tiles.width, width_alignment), + .height = num_tiles.height, + }; +} + +[[nodiscard]] Extent3D BlockLinearAlignedSize(const ImageInfo& info, u32 level) { + // https://github.com/Ryujinx/Ryujinx/blob/1c9aba6de1520aea5480c032e0ff5664ac1bb36f/Ryujinx.Graphics.Texture/SizeCalculator.cs#L176 + ASSERT(info.type != ImageType::Linear); + const Extent3D size = AdjustMipSize(info.size, level); + const Extent3D num_tiles{ + .width = Common::DivCeil(size.width, DefaultBlockWidth(info.format)), + .height = Common::DivCeil(size.height, DefaultBlockHeight(info.format)), + .depth = size.depth, + }; + const u32 bpp_log2 = BytesPerBlockLog2(info.format); + const u32 alignment = StrideAlignment(num_tiles, info.block, bpp_log2, info.tile_width_spacing); + const Extent3D mip_block = AdjustMipBlockSize(num_tiles, info.block, 0); + return Extent3D{ + .width = Common::AlignBits(num_tiles.width, alignment), + .height = Common::AlignBits(num_tiles.height, GOB_SIZE_Y_SHIFT + mip_block.height), + .depth = Common::AlignBits(num_tiles.depth, GOB_SIZE_Z_SHIFT + mip_block.depth), + }; +} + +[[nodiscard]] constexpr u32 NumBlocksPerLayer(const ImageInfo& info, Extent2D tile_size) noexcept { + u32 num_blocks = 0; + for (s32 level = 0; level < info.resources.levels; ++level) { + const Extent3D mip_size = AdjustMipSize(info.size, level); + num_blocks += NumBlocks(mip_size, tile_size); + } + return num_blocks; +} + +[[nodiscard]] u32 NumSlices(const ImageInfo& info) noexcept { + ASSERT(info.type == ImageType::e3D); + u32 num_slices = 0; + for (s32 level = 0; level < info.resources.levels; ++level) { + num_slices += AdjustMipSize(info.size.depth, level); + } + return num_slices; +} + +void SwizzlePitchLinearImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, + const ImageInfo& info, const BufferImageCopy& copy, + std::span<const u8> memory) { + ASSERT(copy.image_offset.z == 0); + ASSERT(copy.image_extent.depth == 1); + ASSERT(copy.image_subresource.base_level == 0); + ASSERT(copy.image_subresource.base_layer == 0); + ASSERT(copy.image_subresource.num_layers == 1); + + const u32 bytes_per_block = BytesPerBlock(info.format); + const u32 row_length = copy.image_extent.width * bytes_per_block; + const u32 guest_offset_x = copy.image_offset.x * bytes_per_block; + + for (u32 line = 0; line < copy.image_extent.height; ++line) { + const u32 host_offset_y = line * info.pitch; + const u32 guest_offset_y = (copy.image_offset.y + line) * info.pitch; + const u32 guest_offset = guest_offset_x + guest_offset_y; + gpu_memory.WriteBlockUnsafe(gpu_addr + guest_offset, memory.data() + host_offset_y, + row_length); + } +} + +void SwizzleBlockLinearImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, + const ImageInfo& info, const BufferImageCopy& copy, + std::span<const u8> input) { + const Extent3D size = info.size; + const LevelInfo level_info = MakeLevelInfo(info); + const Extent2D tile_size = DefaultBlockSize(info.format); + const u32 bytes_per_block = BytesPerBlock(info.format); + + const s32 level = copy.image_subresource.base_level; + const Extent3D level_size = AdjustMipSize(size, level); + const u32 num_blocks_per_layer = NumBlocks(level_size, tile_size); + const u32 host_bytes_per_layer = num_blocks_per_layer * bytes_per_block; + + UNIMPLEMENTED_IF(info.tile_width_spacing > 0); + + UNIMPLEMENTED_IF(copy.image_offset.x != 0); + UNIMPLEMENTED_IF(copy.image_offset.y != 0); + UNIMPLEMENTED_IF(copy.image_offset.z != 0); + UNIMPLEMENTED_IF(copy.image_extent != level_size); + + const Extent3D num_tiles = AdjustTileSize(level_size, tile_size); + const Extent3D block = AdjustMipBlockSize(num_tiles, level_info.block, level); + + size_t host_offset = copy.buffer_offset; + + const u32 num_levels = info.resources.levels; + const std::array sizes = CalculateLevelSizes(level_info, num_levels); + size_t guest_offset = std::reduce(sizes.begin(), sizes.begin() + level, 0); + const size_t layer_stride = + AlignLayerSize(std::reduce(sizes.begin(), sizes.begin() + num_levels, 0), size, + level_info.block, tile_size.height, info.tile_width_spacing); + const size_t subresource_size = sizes[level]; + + const auto dst_data = std::make_unique<u8[]>(subresource_size); + const std::span<u8> dst(dst_data.get(), subresource_size); + + for (s32 layer = 0; layer < info.resources.layers; ++layer) { + const std::span<const u8> src = input.subspan(host_offset); + SwizzleTexture(dst, src, bytes_per_block, num_tiles.width, num_tiles.height, + num_tiles.depth, block.height, block.depth); + + gpu_memory.WriteBlockUnsafe(gpu_addr + guest_offset, dst.data(), dst.size_bytes()); + + host_offset += host_bytes_per_layer; + guest_offset += layer_stride; + } + ASSERT(host_offset - copy.buffer_offset == copy.buffer_size); +} + +} // Anonymous namespace + +u32 CalculateGuestSizeInBytes(const ImageInfo& info) noexcept { + if (info.type == ImageType::Buffer) { + return info.size.width * BytesPerBlock(info.format); + } + if (info.type == ImageType::Linear) { + return info.pitch * Common::DivCeil(info.size.height, DefaultBlockHeight(info.format)); + } + if (info.resources.layers > 1) { + ASSERT(info.layer_stride != 0); + return info.layer_stride * info.resources.layers; + } else { + return CalculateLayerSize(info); + } +} + +u32 CalculateUnswizzledSizeBytes(const ImageInfo& info) noexcept { + if (info.type == ImageType::Buffer) { + return info.size.width * BytesPerBlock(info.format); + } + if (info.num_samples > 1) { + // Multisample images can't be uploaded or downloaded to the host + return 0; + } + if (info.type == ImageType::Linear) { + return info.pitch * Common::DivCeil(info.size.height, DefaultBlockHeight(info.format)); + } + const Extent2D tile_size = DefaultBlockSize(info.format); + return NumBlocksPerLayer(info, tile_size) * info.resources.layers * BytesPerBlock(info.format); +} + +u32 CalculateConvertedSizeBytes(const ImageInfo& info) noexcept { + if (info.type == ImageType::Buffer) { + return info.size.width * BytesPerBlock(info.format); + } + static constexpr Extent2D TILE_SIZE{1, 1}; + return NumBlocksPerLayer(info, TILE_SIZE) * info.resources.layers * CONVERTED_BYTES_PER_BLOCK; +} + +u32 CalculateLayerStride(const ImageInfo& info) noexcept { + ASSERT(info.type != ImageType::Linear); + const u32 layer_size = CalculateLayerSize(info); + const Extent3D size = info.size; + const Extent3D block = info.block; + const u32 tile_size_y = DefaultBlockHeight(info.format); + return AlignLayerSize(layer_size, size, block, tile_size_y, info.tile_width_spacing); +} + +u32 CalculateLayerSize(const ImageInfo& info) noexcept { + ASSERT(info.type != ImageType::Linear); + return CalculateLevelOffset(info.format, info.size, info.block, info.num_samples, + info.tile_width_spacing, info.resources.levels); +} + +std::array<u32, MAX_MIP_LEVELS> CalculateMipLevelOffsets(const ImageInfo& info) noexcept { + ASSERT(info.resources.levels <= MAX_MIP_LEVELS); + const LevelInfo level_info = MakeLevelInfo(info); + std::array<u32, MAX_MIP_LEVELS> offsets{}; + u32 offset = 0; + for (s32 level = 0; level < info.resources.levels; ++level) { + offsets[level] = offset; + offset += CalculateLevelSize(level_info, level); + } + return offsets; +} + +std::vector<u32> CalculateSliceOffsets(const ImageInfo& info) { + ASSERT(info.type == ImageType::e3D); + std::vector<u32> offsets; + offsets.reserve(NumSlices(info)); + + const LevelInfo level_info = MakeLevelInfo(info); + u32 mip_offset = 0; + for (s32 level = 0; level < info.resources.levels; ++level) { + const Extent3D tile_shift = TileShift(level_info, level); + const Extent3D tiles = LevelTiles(level_info, level); + const u32 gob_size_shift = tile_shift.height + GOB_SIZE_SHIFT; + const u32 slice_size = (tiles.width * tiles.height) << gob_size_shift; + const u32 z_mask = (1U << tile_shift.depth) - 1; + const u32 depth = AdjustMipSize(info.size.depth, level); + for (u32 slice = 0; slice < depth; ++slice) { + const u32 z_low = slice & z_mask; + const u32 z_high = slice & ~z_mask; + offsets.push_back(mip_offset + (z_low << gob_size_shift) + (z_high * slice_size)); + } + mip_offset += CalculateLevelSize(level_info, level); + } + return offsets; +} + +std::vector<SubresourceBase> CalculateSliceSubresources(const ImageInfo& info) { + ASSERT(info.type == ImageType::e3D); + std::vector<SubresourceBase> subresources; + subresources.reserve(NumSlices(info)); + for (s32 level = 0; level < info.resources.levels; ++level) { + const s32 depth = AdjustMipSize(info.size.depth, level); + for (s32 slice = 0; slice < depth; ++slice) { + subresources.emplace_back(SubresourceBase{ + .level = level, + .layer = slice, + }); + } + } + return subresources; +} + +u32 CalculateLevelStrideAlignment(const ImageInfo& info, u32 level) { + const Extent2D tile_size = DefaultBlockSize(info.format); + const Extent3D level_size = AdjustMipSize(info.size, level); + const Extent3D num_tiles = AdjustTileSize(level_size, tile_size); + const Extent3D block = AdjustMipBlockSize(num_tiles, info.block, level); + const u32 bpp_log2 = BytesPerBlockLog2(info.format); + return StrideAlignment(num_tiles, block, bpp_log2, info.tile_width_spacing); +} + +PixelFormat PixelFormatFromTIC(const TICEntry& config) noexcept { + return PixelFormatFromTextureInfo(config.format, config.r_type, config.g_type, config.b_type, + config.a_type, config.srgb_conversion); +} + +ImageViewType RenderTargetImageViewType(const ImageInfo& info) noexcept { + switch (info.type) { + case ImageType::e2D: + return info.resources.layers > 1 ? ImageViewType::e2DArray : ImageViewType::e2D; + case ImageType::e3D: + return ImageViewType::e2DArray; + case ImageType::Linear: + return ImageViewType::e2D; + default: + UNIMPLEMENTED_MSG("Unimplemented image type={}", static_cast<int>(info.type)); + return ImageViewType{}; + } +} + +std::vector<ImageCopy> MakeShrinkImageCopies(const ImageInfo& dst, const ImageInfo& src, + SubresourceBase base) { + ASSERT(dst.resources.levels >= src.resources.levels); + ASSERT(dst.num_samples == src.num_samples); + + const bool is_dst_3d = dst.type == ImageType::e3D; + if (is_dst_3d) { + ASSERT(src.type == ImageType::e3D); + ASSERT(src.resources.levels == 1); + } + + std::vector<ImageCopy> copies; + copies.reserve(src.resources.levels); + for (s32 level = 0; level < src.resources.levels; ++level) { + ImageCopy& copy = copies.emplace_back(); + copy.src_subresource = SubresourceLayers{ + .base_level = level, + .base_layer = 0, + .num_layers = src.resources.layers, + }; + copy.dst_subresource = SubresourceLayers{ + .base_level = base.level + level, + .base_layer = is_dst_3d ? 0 : base.layer, + .num_layers = is_dst_3d ? 1 : src.resources.layers, + }; + copy.src_offset = Offset3D{ + .x = 0, + .y = 0, + .z = 0, + }; + copy.dst_offset = Offset3D{ + .x = 0, + .y = 0, + .z = is_dst_3d ? base.layer : 0, + }; + const Extent3D mip_size = AdjustMipSize(dst.size, base.level + level); + copy.extent = AdjustSamplesSize(mip_size, dst.num_samples); + if (is_dst_3d) { + copy.extent.depth = src.size.depth; + } + } + return copies; +} + +bool IsValidAddress(const Tegra::MemoryManager& gpu_memory, const TICEntry& config) { + if (config.Address() == 0) { + return false; + } + if (config.Address() > (u64(1) << 48)) { + return false; + } + return gpu_memory.GpuToCpuAddress(config.Address()).has_value(); +} + +std::vector<BufferImageCopy> UnswizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, + const ImageInfo& info, std::span<u8> output) { + const size_t guest_size_bytes = CalculateGuestSizeInBytes(info); + const u32 bpp_log2 = BytesPerBlockLog2(info.format); + const Extent3D size = info.size; + + if (info.type == ImageType::Linear) { + gpu_memory.ReadBlockUnsafe(gpu_addr, output.data(), guest_size_bytes); + + ASSERT((info.pitch >> bpp_log2) << bpp_log2 == info.pitch); + return {{ + .buffer_offset = 0, + .buffer_size = guest_size_bytes, + .buffer_row_length = info.pitch >> bpp_log2, + .buffer_image_height = size.height, + .image_subresource = + { + .base_level = 0, + .base_layer = 0, + .num_layers = 1, + }, + .image_offset = {0, 0, 0}, + .image_extent = size, + }}; + } + const auto input_data = std::make_unique<u8[]>(guest_size_bytes); + gpu_memory.ReadBlockUnsafe(gpu_addr, input_data.get(), guest_size_bytes); + const std::span<const u8> input(input_data.get(), guest_size_bytes); + + const LevelInfo level_info = MakeLevelInfo(info); + const s32 num_layers = info.resources.layers; + const s32 num_levels = info.resources.levels; + const Extent2D tile_size = DefaultBlockSize(info.format); + const std::array level_sizes = CalculateLevelSizes(level_info, num_levels); + const Extent2D gob = GobSize(bpp_log2, info.block.height, info.tile_width_spacing); + const u32 layer_size = std::reduce(level_sizes.begin(), level_sizes.begin() + num_levels, 0); + const u32 layer_stride = AlignLayerSize(layer_size, size, level_info.block, tile_size.height, + info.tile_width_spacing); + size_t guest_offset = 0; + u32 host_offset = 0; + std::vector<BufferImageCopy> copies(num_levels); + + for (s32 level = 0; level < num_levels; ++level) { + const Extent3D level_size = AdjustMipSize(size, level); + const u32 num_blocks_per_layer = NumBlocks(level_size, tile_size); + const u32 host_bytes_per_layer = num_blocks_per_layer << bpp_log2; + copies[level] = BufferImageCopy{ + .buffer_offset = host_offset, + .buffer_size = static_cast<size_t>(host_bytes_per_layer) * num_layers, + .buffer_row_length = Common::AlignUp(level_size.width, tile_size.width), + .buffer_image_height = Common::AlignUp(level_size.height, tile_size.height), + .image_subresource = + { + .base_level = level, + .base_layer = 0, + .num_layers = info.resources.layers, + }, + .image_offset = {0, 0, 0}, + .image_extent = level_size, + }; + const Extent3D num_tiles = AdjustTileSize(level_size, tile_size); + const Extent3D block = AdjustMipBlockSize(num_tiles, level_info.block, level); + const u32 stride_alignment = StrideAlignment(num_tiles, info.block, gob, bpp_log2); + size_t guest_layer_offset = 0; + + for (s32 layer = 0; layer < info.resources.layers; ++layer) { + const std::span<u8> dst = output.subspan(host_offset); + const std::span<const u8> src = input.subspan(guest_offset + guest_layer_offset); + UnswizzleTexture(dst, src, 1U << bpp_log2, num_tiles.width, num_tiles.height, + num_tiles.depth, block.height, block.depth, stride_alignment); + guest_layer_offset += layer_stride; + host_offset += host_bytes_per_layer; + } + guest_offset += level_sizes[level]; + } + return copies; +} + +BufferCopy UploadBufferCopy(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, + const ImageBase& image, std::span<u8> output) { + gpu_memory.ReadBlockUnsafe(gpu_addr, output.data(), image.guest_size_bytes); + return BufferCopy{ + .src_offset = 0, + .dst_offset = 0, + .size = image.guest_size_bytes, + }; +} + +void ConvertImage(std::span<const u8> input, const ImageInfo& info, std::span<u8> output, + std::span<BufferImageCopy> copies) { + u32 output_offset = 0; + + const Extent2D tile_size = DefaultBlockSize(info.format); + for (BufferImageCopy& copy : copies) { + const u32 level = copy.image_subresource.base_level; + const Extent3D mip_size = AdjustMipSize(info.size, level); + ASSERT(copy.image_offset == Offset3D{}); + ASSERT(copy.image_subresource.base_layer == 0); + ASSERT(copy.image_extent == mip_size); + ASSERT(copy.buffer_row_length == Common::AlignUp(mip_size.width, tile_size.width)); + ASSERT(copy.buffer_image_height == Common::AlignUp(mip_size.height, tile_size.height)); + + if (IsPixelFormatASTC(info.format)) { + ASSERT(copy.image_extent.depth == 1); + Tegra::Texture::ASTC::Decompress(input.subspan(copy.buffer_offset), + copy.image_extent.width, copy.image_extent.height, + copy.image_subresource.num_layers, tile_size.width, + tile_size.height, output.subspan(output_offset)); + } else { + DecompressBC4(input.subspan(copy.buffer_offset), copy.image_extent, + output.subspan(output_offset)); + } + copy.buffer_offset = output_offset; + copy.buffer_row_length = mip_size.width; + copy.buffer_image_height = mip_size.height; + + output_offset += copy.image_extent.width * copy.image_extent.height * + copy.image_subresource.num_layers * CONVERTED_BYTES_PER_BLOCK; + } +} + +std::vector<BufferImageCopy> FullDownloadCopies(const ImageInfo& info) { + const Extent3D size = info.size; + const u32 bytes_per_block = BytesPerBlock(info.format); + if (info.type == ImageType::Linear) { + ASSERT(info.pitch % bytes_per_block == 0); + return {{ + .buffer_offset = 0, + .buffer_size = static_cast<size_t>(info.pitch) * size.height, + .buffer_row_length = info.pitch / bytes_per_block, + .buffer_image_height = size.height, + .image_subresource = + { + .base_level = 0, + .base_layer = 0, + .num_layers = 1, + }, + .image_offset = {0, 0, 0}, + .image_extent = size, + }}; + } + UNIMPLEMENTED_IF(info.tile_width_spacing > 0); + + const s32 num_layers = info.resources.layers; + const s32 num_levels = info.resources.levels; + const Extent2D tile_size = DefaultBlockSize(info.format); + + u32 host_offset = 0; + + std::vector<BufferImageCopy> copies(num_levels); + for (s32 level = 0; level < num_levels; ++level) { + const Extent3D level_size = AdjustMipSize(size, level); + const u32 num_blocks_per_layer = NumBlocks(level_size, tile_size); + const u32 host_bytes_per_level = num_blocks_per_layer * bytes_per_block * num_layers; + copies[level] = BufferImageCopy{ + .buffer_offset = host_offset, + .buffer_size = host_bytes_per_level, + .buffer_row_length = level_size.width, + .buffer_image_height = level_size.height, + .image_subresource = + { + .base_level = level, + .base_layer = 0, + .num_layers = info.resources.layers, + }, + .image_offset = {0, 0, 0}, + .image_extent = level_size, + }; + host_offset += host_bytes_per_level; + } + return copies; +} + +Extent3D MipSize(Extent3D size, u32 level) { + return AdjustMipSize(size, level); +} + +Extent3D MipBlockSize(const ImageInfo& info, u32 level) { + const LevelInfo level_info = MakeLevelInfo(info); + const Extent2D tile_size = DefaultBlockSize(info.format); + const Extent3D level_size = AdjustMipSize(info.size, level); + const Extent3D num_tiles = AdjustTileSize(level_size, tile_size); + return AdjustMipBlockSize(num_tiles, level_info.block, level); +} + +std::vector<SwizzleParameters> FullUploadSwizzles(const ImageInfo& info) { + const Extent2D tile_size = DefaultBlockSize(info.format); + if (info.type == ImageType::Linear) { + return std::vector{SwizzleParameters{ + .num_tiles = AdjustTileSize(info.size, tile_size), + .block = {}, + .buffer_offset = 0, + .level = 0, + }}; + } + const LevelInfo level_info = MakeLevelInfo(info); + const Extent3D size = info.size; + const s32 num_levels = info.resources.levels; + + u32 guest_offset = 0; + std::vector<SwizzleParameters> params(num_levels); + for (s32 level = 0; level < num_levels; ++level) { + const Extent3D level_size = AdjustMipSize(size, level); + const Extent3D num_tiles = AdjustTileSize(level_size, tile_size); + const Extent3D block = AdjustMipBlockSize(num_tiles, level_info.block, level); + params[level] = SwizzleParameters{ + .num_tiles = num_tiles, + .block = block, + .buffer_offset = guest_offset, + .level = level, + }; + guest_offset += CalculateLevelSize(level_info, level); + } + return params; +} + +void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const ImageInfo& info, + std::span<const BufferImageCopy> copies, std::span<const u8> memory) { + const bool is_pitch_linear = info.type == ImageType::Linear; + for (const BufferImageCopy& copy : copies) { + if (is_pitch_linear) { + SwizzlePitchLinearImage(gpu_memory, gpu_addr, info, copy, memory); + } else { + SwizzleBlockLinearImage(gpu_memory, gpu_addr, info, copy, memory); + } + } +} + +bool IsBlockLinearSizeCompatible(const ImageInfo& lhs, const ImageInfo& rhs, u32 lhs_level, + u32 rhs_level, bool strict_size) noexcept { + ASSERT(lhs.type != ImageType::Linear); + ASSERT(rhs.type != ImageType::Linear); + if (strict_size) { + const Extent3D lhs_size = AdjustMipSize(lhs.size, lhs_level); + const Extent3D rhs_size = AdjustMipSize(rhs.size, rhs_level); + return lhs_size.width == rhs_size.width && lhs_size.height == rhs_size.height; + } else { + const Extent3D lhs_size = BlockLinearAlignedSize(lhs, lhs_level); + const Extent3D rhs_size = BlockLinearAlignedSize(rhs, rhs_level); + return lhs_size.width == rhs_size.width && lhs_size.height == rhs_size.height; + } +} + +bool IsPitchLinearSameSize(const ImageInfo& lhs, const ImageInfo& rhs, bool strict_size) noexcept { + ASSERT(lhs.type == ImageType::Linear); + ASSERT(rhs.type == ImageType::Linear); + if (strict_size) { + return lhs.size.width == rhs.size.width && lhs.size.height == rhs.size.height; + } else { + const Extent2D lhs_size = PitchLinearAlignedSize(lhs); + const Extent2D rhs_size = PitchLinearAlignedSize(rhs); + return lhs_size == rhs_size; + } +} + +std::optional<OverlapResult> ResolveOverlap(const ImageInfo& new_info, GPUVAddr gpu_addr, + VAddr cpu_addr, const ImageBase& overlap, + bool strict_size, bool broken_views) { + ASSERT(new_info.type != ImageType::Linear); + ASSERT(overlap.info.type != ImageType::Linear); + if (!IsLayerStrideCompatible(new_info, overlap.info)) { + return std::nullopt; + } + if (!IsViewCompatible(overlap.info.format, new_info.format, broken_views)) { + return std::nullopt; + } + if (gpu_addr == overlap.gpu_addr) { + const std::optional solution = ResolveOverlapEqualAddress(new_info, overlap, strict_size); + if (!solution) { + return std::nullopt; + } + return OverlapResult{ + .gpu_addr = gpu_addr, + .cpu_addr = cpu_addr, + .resources = *solution, + }; + } + if (overlap.gpu_addr > gpu_addr) { + return ResolveOverlapRightAddress(new_info, gpu_addr, cpu_addr, overlap, strict_size); + } + // if overlap.gpu_addr < gpu_addr + return ResolveOverlapLeftAddress(new_info, gpu_addr, cpu_addr, overlap, strict_size); +} + +bool IsLayerStrideCompatible(const ImageInfo& lhs, const ImageInfo& rhs) { + // If either of the layer strides is zero, we can assume they are compatible + // These images generally come from rendertargets + if (lhs.layer_stride == 0) { + return true; + } + if (rhs.layer_stride == 0) { + return true; + } + // It's definitely compatible if the layer stride matches + if (lhs.layer_stride == rhs.layer_stride) { + return true; + } + // Although we also have to compare for cases where it can be unaligned + // This can happen if the image doesn't have layers, so the stride is not aligned + if (lhs.maybe_unaligned_layer_stride == rhs.maybe_unaligned_layer_stride) { + return true; + } + return false; +} + +std::optional<SubresourceBase> FindSubresource(const ImageInfo& candidate, const ImageBase& image, + GPUVAddr candidate_addr, RelaxedOptions options, + bool broken_views) { + const std::optional<SubresourceBase> base = image.TryFindBase(candidate_addr); + if (!base) { + return std::nullopt; + } + const ImageInfo& existing = image.info; + if (False(options & RelaxedOptions::Format)) { + if (!IsViewCompatible(existing.format, candidate.format, broken_views)) { + return std::nullopt; + } + } + if (!IsLayerStrideCompatible(existing, candidate)) { + return std::nullopt; + } + if (existing.type != candidate.type) { + return std::nullopt; + } + if (False(options & RelaxedOptions::Samples)) { + if (existing.num_samples != candidate.num_samples) { + return std::nullopt; + } + } + if (existing.resources.levels < candidate.resources.levels + base->level) { + return std::nullopt; + } + if (existing.type == ImageType::e3D) { + const u32 mip_depth = std::max(1U, existing.size.depth << base->level); + if (mip_depth < candidate.size.depth + base->layer) { + return std::nullopt; + } + } else { + if (existing.resources.layers < candidate.resources.layers + base->layer) { + return std::nullopt; + } + } + const bool strict_size = False(options & RelaxedOptions::Size); + if (!IsBlockLinearSizeCompatible(existing, candidate, base->level, 0, strict_size)) { + return std::nullopt; + } + // TODO: compare block sizes + return base; +} + +bool IsSubresource(const ImageInfo& candidate, const ImageBase& image, GPUVAddr candidate_addr, + RelaxedOptions options, bool broken_views) { + return FindSubresource(candidate, image, candidate_addr, options, broken_views).has_value(); +} + +void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst, + const ImageBase* src) { + if (src && GetFormatType(src->info.format) != SurfaceType::ColorTexture) { + src_info.format = src->info.format; + } + if (dst && GetFormatType(dst->info.format) != SurfaceType::ColorTexture) { + dst_info.format = dst->info.format; + } + if (!dst && src && GetFormatType(src->info.format) != SurfaceType::ColorTexture) { + dst_info.format = src->info.format; + } + if (!src && dst && GetFormatType(dst->info.format) != SurfaceType::ColorTexture) { + src_info.format = src->info.format; + } +} + +u32 MapSizeBytes(const ImageBase& image) { + if (True(image.flags & ImageFlagBits::AcceleratedUpload)) { + return image.guest_size_bytes; + } else if (True(image.flags & ImageFlagBits::Converted)) { + return image.converted_size_bytes; + } else { + return image.unswizzled_size_bytes; + } +} + +using P = PixelFormat; + +static_assert(CalculateLevelSize(LevelInfo{{1920, 1080}, {0, 2, 0}, {1, 1}, 2, 0}, 0) == 0x7f8000); +static_assert(CalculateLevelSize(LevelInfo{{32, 32}, {0, 0, 4}, {1, 1}, 4, 0}, 0) == 0x4000); + +static_assert(CalculateLevelOffset(P::R8_SINT, {1920, 1080}, {0, 2}, 1, 0, 7) == 0x2afc00); +static_assert(CalculateLevelOffset(P::ASTC_2D_12X12_UNORM, {8192, 4096}, {0, 2}, 1, 0, 12) == + 0x50d200); + +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 0) == 0); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 1) == 0x400000); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 2) == 0x500000); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 3) == 0x540000); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 4) == 0x550000); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 5) == 0x554000); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 6) == 0x555000); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 7) == 0x555400); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 8) == 0x555600); +static_assert(CalculateLevelOffset(P::A8B8G8R8_UNORM, {1024, 1024}, {0, 4}, 1, 0, 9) == 0x555800); + +constexpr u32 ValidateLayerSize(PixelFormat format, u32 width, u32 height, u32 block_height, + u32 tile_width_spacing, u32 level) { + const Extent3D size{width, height, 1}; + const Extent3D block{0, block_height, 0}; + const u32 offset = CalculateLevelOffset(format, size, block, 1, tile_width_spacing, level); + return AlignLayerSize(offset, size, block, DefaultBlockHeight(format), tile_width_spacing); +} + +static_assert(ValidateLayerSize(P::ASTC_2D_12X12_UNORM, 8192, 4096, 2, 0, 12) == 0x50d800); +static_assert(ValidateLayerSize(P::A8B8G8R8_UNORM, 1024, 1024, 2, 0, 10) == 0x556000); +static_assert(ValidateLayerSize(P::BC3_UNORM, 128, 128, 2, 0, 8) == 0x6000); + +static_assert(ValidateLayerSize(P::A8B8G8R8_UNORM, 518, 572, 4, 3, 1) == 0x190000, + "Tile width spacing is not working"); +static_assert(ValidateLayerSize(P::BC5_UNORM, 1024, 1024, 3, 4, 11) == 0x160000, + "Compressed tile width spacing is not working"); + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache/util.h b/src/video_core/texture_cache/util.h new file mode 100644 index 000000000..52a9207d6 --- /dev/null +++ b/src/video_core/texture_cache/util.h @@ -0,0 +1,109 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <optional> +#include <span> + +#include "common/common_types.h" + +#include "video_core/engines/maxwell_3d.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/image_base.h" +#include "video_core/texture_cache/image_view_base.h" +#include "video_core/texture_cache/types.h" +#include "video_core/textures/texture.h" + +namespace VideoCommon { + +using Tegra::Texture::TICEntry; + +struct OverlapResult { + GPUVAddr gpu_addr; + VAddr cpu_addr; + SubresourceExtent resources; +}; + +[[nodiscard]] u32 CalculateGuestSizeInBytes(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateUnswizzledSizeBytes(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateConvertedSizeBytes(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateLayerStride(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateLayerSize(const ImageInfo& info) noexcept; + +[[nodiscard]] std::array<u32, MAX_MIP_LEVELS> CalculateMipLevelOffsets( + const ImageInfo& info) noexcept; + +[[nodiscard]] std::vector<u32> CalculateSliceOffsets(const ImageInfo& info); + +[[nodiscard]] std::vector<SubresourceBase> CalculateSliceSubresources(const ImageInfo& info); + +[[nodiscard]] u32 CalculateLevelStrideAlignment(const ImageInfo& info, u32 level); + +[[nodiscard]] VideoCore::Surface::PixelFormat PixelFormatFromTIC( + const Tegra::Texture::TICEntry& config) noexcept; + +[[nodiscard]] ImageViewType RenderTargetImageViewType(const ImageInfo& info) noexcept; + +[[nodiscard]] std::vector<ImageCopy> MakeShrinkImageCopies(const ImageInfo& dst, + const ImageInfo& src, + SubresourceBase base); + +[[nodiscard]] bool IsValidAddress(const Tegra::MemoryManager& gpu_memory, const TICEntry& config); + +[[nodiscard]] std::vector<BufferImageCopy> UnswizzleImage(Tegra::MemoryManager& gpu_memory, + GPUVAddr gpu_addr, const ImageInfo& info, + std::span<u8> output); + +[[nodiscard]] BufferCopy UploadBufferCopy(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, + const ImageBase& image, std::span<u8> output); + +void ConvertImage(std::span<const u8> input, const ImageInfo& info, std::span<u8> output, + std::span<BufferImageCopy> copies); + +[[nodiscard]] std::vector<BufferImageCopy> FullDownloadCopies(const ImageInfo& info); + +[[nodiscard]] Extent3D MipSize(Extent3D size, u32 level); + +[[nodiscard]] Extent3D MipBlockSize(const ImageInfo& info, u32 level); + +[[nodiscard]] std::vector<SwizzleParameters> FullUploadSwizzles(const ImageInfo& info); + +void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const ImageInfo& info, + std::span<const BufferImageCopy> copies, std::span<const u8> memory); + +[[nodiscard]] bool IsBlockLinearSizeCompatible(const ImageInfo& new_info, + const ImageInfo& overlap_info, u32 new_level, + u32 overlap_level, bool strict_size) noexcept; + +[[nodiscard]] bool IsPitchLinearSameSize(const ImageInfo& lhs, const ImageInfo& rhs, + bool strict_size) noexcept; + +[[nodiscard]] std::optional<OverlapResult> ResolveOverlap(const ImageInfo& new_info, + GPUVAddr gpu_addr, VAddr cpu_addr, + const ImageBase& overlap, + bool strict_size, bool broken_views); + +[[nodiscard]] bool IsLayerStrideCompatible(const ImageInfo& lhs, const ImageInfo& rhs); + +[[nodiscard]] std::optional<SubresourceBase> FindSubresource(const ImageInfo& candidate, + const ImageBase& image, + GPUVAddr candidate_addr, + RelaxedOptions options, + bool broken_views); + +[[nodiscard]] bool IsSubresource(const ImageInfo& candidate, const ImageBase& image, + GPUVAddr candidate_addr, RelaxedOptions options, + bool broken_views); + +void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst, + const ImageBase* src); + +[[nodiscard]] u32 MapSizeBytes(const ImageBase& image); + +} // namespace VideoCommon diff --git a/src/video_core/textures/astc.cpp b/src/video_core/textures/astc.cpp index 365bde2f1..acd5bdd78 100644 --- a/src/video_core/textures/astc.cpp +++ b/src/video_core/textures/astc.cpp @@ -18,6 +18,7 @@ #include <algorithm> #include <cassert> #include <cstring> +#include <span> #include <vector> #include <boost/container/static_vector.hpp> @@ -600,7 +601,7 @@ static TexelWeightParams DecodeBlockInfo(InputBitStream& strm) { return params; } -static void FillVoidExtentLDR(InputBitStream& strm, u32* const outBuf, u32 blockWidth, +static void FillVoidExtentLDR(InputBitStream& strm, std::span<u32> outBuf, u32 blockWidth, u32 blockHeight) { // Don't actually care about the void extent, just read the bits... for (s32 i = 0; i < 4; ++i) { @@ -623,7 +624,7 @@ static void FillVoidExtentLDR(InputBitStream& strm, u32* const outBuf, u32 block } } -static void FillError(u32* outBuf, u32 blockWidth, u32 blockHeight) { +static void FillError(std::span<u32> outBuf, u32 blockWidth, u32 blockHeight) { for (u32 j = 0; j < blockHeight; j++) { for (u32 i = 0; i < blockWidth; i++) { outBuf[j * blockWidth + i] = 0xFFFF00FF; @@ -1438,9 +1439,9 @@ static void ComputeEndpos32s(Pixel& ep1, Pixel& ep2, const u32*& colorValues, #undef READ_INT_VALUES } -static void DecompressBlock(const u8 inBuf[16], const u32 blockWidth, const u32 blockHeight, - u32* outBuf) { - InputBitStream strm(inBuf); +static void DecompressBlock(std::span<const u8, 16> inBuf, const u32 blockWidth, + const u32 blockHeight, std::span<u32, 12 * 12> outBuf) { + InputBitStream strm(inBuf.data()); TexelWeightParams weightParams = DecodeBlockInfo(strm); // Was there an error? @@ -1601,8 +1602,8 @@ static void DecompressBlock(const u8 inBuf[16], const u32 blockWidth, const u32 } // Read the texel weight data.. - u8 texelWeightData[16]; - memcpy(texelWeightData, inBuf, sizeof(texelWeightData)); + std::array<u8, 16> texelWeightData; + std::ranges::copy(inBuf, texelWeightData.begin()); // Reverse everything for (u32 i = 0; i < 8; i++) { @@ -1618,14 +1619,15 @@ static void DecompressBlock(const u8 inBuf[16], const u32 blockWidth, const u32 // Make sure that higher non-texel bits are set to zero const u32 clearByteStart = (weightParams.GetPackedBitSize() >> 3) + 1; - texelWeightData[clearByteStart - 1] = - texelWeightData[clearByteStart - 1] & - static_cast<u8>((1 << (weightParams.GetPackedBitSize() % 8)) - 1); - memset(texelWeightData + clearByteStart, 0, 16 - clearByteStart); + if (clearByteStart > 0) { + texelWeightData[clearByteStart - 1] &= + static_cast<u8>((1 << (weightParams.GetPackedBitSize() % 8)) - 1); + } + std::memset(texelWeightData.data() + clearByteStart, 0, std::min(16U - clearByteStart, 16U)); IntegerEncodedVector texelWeightValues; - InputBitStream weightStream(texelWeightData); + InputBitStream weightStream(texelWeightData.data()); DecodeIntegerSequence(texelWeightValues, weightStream, weightParams.m_MaxWeight, weightParams.GetNumWeightValues()); @@ -1672,36 +1674,32 @@ static void DecompressBlock(const u8 inBuf[16], const u32 blockWidth, const u32 namespace Tegra::Texture::ASTC { -std::vector<u8> Decompress(const u8* data, u32 width, u32 height, u32 depth, u32 block_width, - u32 block_height) { - u32 blockIdx = 0; +void Decompress(std::span<const uint8_t> data, uint32_t width, uint32_t height, uint32_t depth, + uint32_t block_width, uint32_t block_height, std::span<uint8_t> output) { + u32 block_index = 0; std::size_t depth_offset = 0; - std::vector<u8> outData(height * width * depth * 4); - for (u32 k = 0; k < depth; k++) { - for (u32 j = 0; j < height; j += block_height) { - for (u32 i = 0; i < width; i += block_width) { - - const u8* blockPtr = data + blockIdx * 16; + for (u32 z = 0; z < depth; z++) { + for (u32 y = 0; y < height; y += block_height) { + for (u32 x = 0; x < width; x += block_width) { + const std::span<const u8, 16> blockPtr{data.subspan(block_index * 16, 16)}; // Blocks can be at most 12x12 - u32 uncompData[144]; + std::array<u32, 12 * 12> uncompData; ASTCC::DecompressBlock(blockPtr, block_width, block_height, uncompData); - u32 decompWidth = std::min(block_width, width - i); - u32 decompHeight = std::min(block_height, height - j); + u32 decompWidth = std::min(block_width, width - x); + u32 decompHeight = std::min(block_height, height - y); - u8* outRow = depth_offset + outData.data() + (j * width + i) * 4; + const std::span<u8> outRow = output.subspan(depth_offset + (y * width + x) * 4); for (u32 jj = 0; jj < decompHeight; jj++) { - memcpy(outRow + jj * width * 4, uncompData + jj * block_width, decompWidth * 4); + std::memcpy(outRow.data() + jj * width * 4, + uncompData.data() + jj * block_width, decompWidth * 4); } - - blockIdx++; + ++block_index; } } depth_offset += height * width * 4; } - - return outData; } } // namespace Tegra::Texture::ASTC diff --git a/src/video_core/textures/astc.h b/src/video_core/textures/astc.h index 991cdba72..9105119bc 100644 --- a/src/video_core/textures/astc.h +++ b/src/video_core/textures/astc.h @@ -5,11 +5,10 @@ #pragma once #include <cstdint> -#include <vector> namespace Tegra::Texture::ASTC { -std::vector<uint8_t> Decompress(const uint8_t* data, uint32_t width, uint32_t height, - uint32_t depth, uint32_t block_width, uint32_t block_height); +void Decompress(std::span<const uint8_t> data, uint32_t width, uint32_t height, uint32_t depth, + uint32_t block_width, uint32_t block_height, std::span<uint8_t> output); } // namespace Tegra::Texture::ASTC diff --git a/src/video_core/textures/convert.cpp b/src/video_core/textures/convert.cpp deleted file mode 100644 index 962921483..000000000 --- a/src/video_core/textures/convert.cpp +++ /dev/null @@ -1,93 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <algorithm> -#include <cstring> -#include <tuple> -#include <vector> - -#include "common/assert.h" -#include "common/common_types.h" -#include "common/logging/log.h" -#include "video_core/surface.h" -#include "video_core/textures/astc.h" -#include "video_core/textures/convert.h" - -namespace Tegra::Texture { - -using VideoCore::Surface::PixelFormat; - -template <bool reverse> -void SwapS8Z24ToZ24S8(u8* data, u32 width, u32 height) { - union S8Z24 { - BitField<0, 24, u32> z24; - BitField<24, 8, u32> s8; - }; - static_assert(sizeof(S8Z24) == 4, "S8Z24 is incorrect size"); - - union Z24S8 { - BitField<0, 8, u32> s8; - BitField<8, 24, u32> z24; - }; - static_assert(sizeof(Z24S8) == 4, "Z24S8 is incorrect size"); - - S8Z24 s8z24_pixel{}; - Z24S8 z24s8_pixel{}; - constexpr auto bpp{ - VideoCore::Surface::GetBytesPerPixel(VideoCore::Surface::PixelFormat::S8_UINT_D24_UNORM)}; - for (std::size_t y = 0; y < height; ++y) { - for (std::size_t x = 0; x < width; ++x) { - const std::size_t offset{bpp * (y * width + x)}; - if constexpr (reverse) { - std::memcpy(&z24s8_pixel, &data[offset], sizeof(Z24S8)); - s8z24_pixel.s8.Assign(z24s8_pixel.s8); - s8z24_pixel.z24.Assign(z24s8_pixel.z24); - std::memcpy(&data[offset], &s8z24_pixel, sizeof(S8Z24)); - } else { - std::memcpy(&s8z24_pixel, &data[offset], sizeof(S8Z24)); - z24s8_pixel.s8.Assign(s8z24_pixel.s8); - z24s8_pixel.z24.Assign(s8z24_pixel.z24); - std::memcpy(&data[offset], &z24s8_pixel, sizeof(Z24S8)); - } - } - } -} - -static void ConvertS8Z24ToZ24S8(u8* data, u32 width, u32 height) { - SwapS8Z24ToZ24S8<false>(data, width, height); -} - -static void ConvertZ24S8ToS8Z24(u8* data, u32 width, u32 height) { - SwapS8Z24ToZ24S8<true>(data, width, height); -} - -void ConvertFromGuestToHost(u8* in_data, u8* out_data, PixelFormat pixel_format, u32 width, - u32 height, u32 depth, bool convert_astc, bool convert_s8z24) { - if (convert_astc && IsPixelFormatASTC(pixel_format)) { - // Convert ASTC pixel formats to RGBA8, as most desktop GPUs do not support ASTC. - u32 block_width{}; - u32 block_height{}; - std::tie(block_width, block_height) = GetASTCBlockSize(pixel_format); - const std::vector<u8> rgba8_data = Tegra::Texture::ASTC::Decompress( - in_data, width, height, depth, block_width, block_height); - std::copy(rgba8_data.begin(), rgba8_data.end(), out_data); - - } else if (convert_s8z24 && pixel_format == PixelFormat::S8_UINT_D24_UNORM) { - Tegra::Texture::ConvertS8Z24ToZ24S8(in_data, width, height); - } -} - -void ConvertFromHostToGuest(u8* data, PixelFormat pixel_format, u32 width, u32 height, u32 depth, - bool convert_astc, bool convert_s8z24) { - if (convert_astc && IsPixelFormatASTC(pixel_format)) { - LOG_CRITICAL(HW_GPU, "Conversion of format {} after texture flushing is not implemented", - static_cast<u32>(pixel_format)); - UNREACHABLE(); - - } else if (convert_s8z24 && pixel_format == PixelFormat::S8_UINT_D24_UNORM) { - Tegra::Texture::ConvertZ24S8ToS8Z24(data, width, height); - } -} - -} // namespace Tegra::Texture diff --git a/src/video_core/textures/convert.h b/src/video_core/textures/convert.h deleted file mode 100644 index d5d6c77bb..000000000 --- a/src/video_core/textures/convert.h +++ /dev/null @@ -1,22 +0,0 @@ -// Copyright 2019 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include "common/common_types.h" - -namespace VideoCore::Surface { -enum class PixelFormat; -} - -namespace Tegra::Texture { - -void ConvertFromGuestToHost(u8* in_data, u8* out_data, VideoCore::Surface::PixelFormat pixel_format, - u32 width, u32 height, u32 depth, bool convert_astc, - bool convert_s8z24); - -void ConvertFromHostToGuest(u8* data, VideoCore::Surface::PixelFormat pixel_format, u32 width, - u32 height, u32 depth, bool convert_astc, bool convert_s8z24); - -} // namespace Tegra::Texture diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index 16d46a018..9f5181318 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -2,204 +2,111 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> #include <cmath> #include <cstring> +#include <span> +#include <utility> + #include "common/alignment.h" #include "common/assert.h" #include "common/bit_util.h" +#include "common/div_ceil.h" #include "video_core/gpu.h" #include "video_core/textures/decoders.h" #include "video_core/textures/texture.h" namespace Tegra::Texture { -namespace { +namespace { /** - * This table represents the internal swizzle of a gob, - * in format 16 bytes x 2 sector packing. + * This table represents the internal swizzle of a gob, in format 16 bytes x 2 sector packing. * Calculates the offset of an (x, y) position within a swizzled texture. * Taken from the Tegra X1 Technical Reference Manual. pages 1187-1188 */ -template <std::size_t N, std::size_t M, u32 Align> -struct alignas(64) SwizzleTable { - static_assert(M * Align == 64, "Swizzle Table does not align to GOB"); - constexpr SwizzleTable() { - for (u32 y = 0; y < N; ++y) { - for (u32 x = 0; x < M; ++x) { - const u32 x2 = x * Align; - values[y][x] = static_cast<u16>(((x2 % 64) / 32) * 256 + ((y % 8) / 2) * 64 + - ((x2 % 32) / 16) * 32 + (y % 2) * 16 + (x2 % 16)); - } +constexpr SwizzleTable MakeSwizzleTableConst() { + SwizzleTable table{}; + for (u32 y = 0; y < table.size(); ++y) { + for (u32 x = 0; x < table[0].size(); ++x) { + table[y][x] = ((x % 64) / 32) * 256 + ((y % 8) / 2) * 64 + ((x % 32) / 16) * 32 + + (y % 2) * 16 + (x % 16); } } - const std::array<u16, M>& operator[](std::size_t index) const { - return values[index]; - } - std::array<std::array<u16, M>, N> values{}; -}; + return table; +} -constexpr u32 FAST_SWIZZLE_ALIGN = 16; +constexpr SwizzleTable SWIZZLE_TABLE = MakeSwizzleTableConst(); -constexpr auto LEGACY_SWIZZLE_TABLE = SwizzleTable<GOB_SIZE_X, GOB_SIZE_X, GOB_SIZE_Z>(); -constexpr auto FAST_SWIZZLE_TABLE = SwizzleTable<GOB_SIZE_Y, 4, FAST_SWIZZLE_ALIGN>(); +template <bool TO_LINEAR> +void Swizzle(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, u32 width, + u32 height, u32 depth, u32 block_height, u32 block_depth, u32 stride_alignment) { + // The origin of the transformation can be configured here, leave it as zero as the current API + // doesn't expose it. + static constexpr u32 origin_x = 0; + static constexpr u32 origin_y = 0; + static constexpr u32 origin_z = 0; -/** - * This function manages ALL the GOBs(Group of Bytes) Inside a single block. - * Instead of going gob by gob, we map the coordinates inside a block and manage from - * those. Block_Width is assumed to be 1. - */ -void PreciseProcessBlock(u8* const swizzled_data, u8* const unswizzled_data, const bool unswizzle, - const u32 x_start, const u32 y_start, const u32 z_start, const u32 x_end, - const u32 y_end, const u32 z_end, const u32 tile_offset, - const u32 xy_block_size, const u32 layer_z, const u32 stride_x, - const u32 bytes_per_pixel, const u32 out_bytes_per_pixel) { - std::array<u8*, 2> data_ptrs; - u32 z_address = tile_offset; - - for (u32 z = z_start; z < z_end; z++) { - u32 y_address = z_address; - u32 pixel_base = layer_z * z + y_start * stride_x; - for (u32 y = y_start; y < y_end; y++) { - const auto& table = LEGACY_SWIZZLE_TABLE[y % GOB_SIZE_Y]; - for (u32 x = x_start; x < x_end; x++) { - const u32 swizzle_offset{y_address + table[x * bytes_per_pixel % GOB_SIZE_X]}; - const u32 pixel_index{x * out_bytes_per_pixel + pixel_base}; - data_ptrs[unswizzle] = swizzled_data + swizzle_offset; - data_ptrs[!unswizzle] = unswizzled_data + pixel_index; - std::memcpy(data_ptrs[0], data_ptrs[1], bytes_per_pixel); - } - pixel_base += stride_x; - if ((y + 1) % GOB_SIZE_Y == 0) - y_address += GOB_SIZE; - } - z_address += xy_block_size; - } -} + // We can configure here a custom pitch + // As it's not exposed 'width * bpp' will be the expected pitch. + const u32 pitch = width * bytes_per_pixel; + const u32 stride = Common::AlignBits(width, stride_alignment) * bytes_per_pixel; -/** - * This function manages ALL the GOBs(Group of Bytes) Inside a single block. - * Instead of going gob by gob, we map the coordinates inside a block and manage from - * those. Block_Width is assumed to be 1. - */ -void FastProcessBlock(u8* const swizzled_data, u8* const unswizzled_data, const bool unswizzle, - const u32 x_start, const u32 y_start, const u32 z_start, const u32 x_end, - const u32 y_end, const u32 z_end, const u32 tile_offset, - const u32 xy_block_size, const u32 layer_z, const u32 stride_x, - const u32 bytes_per_pixel, const u32 out_bytes_per_pixel) { - std::array<u8*, 2> data_ptrs; - u32 z_address = tile_offset; - const u32 x_startb = x_start * bytes_per_pixel; - const u32 x_endb = x_end * bytes_per_pixel; - - for (u32 z = z_start; z < z_end; z++) { - u32 y_address = z_address; - u32 pixel_base = layer_z * z + y_start * stride_x; - for (u32 y = y_start; y < y_end; y++) { - const auto& table = FAST_SWIZZLE_TABLE[y % GOB_SIZE_Y]; - for (u32 xb = x_startb; xb < x_endb; xb += FAST_SWIZZLE_ALIGN) { - const u32 swizzle_offset{y_address + table[(xb / FAST_SWIZZLE_ALIGN) % 4]}; - const u32 out_x = xb * out_bytes_per_pixel / bytes_per_pixel; - const u32 pixel_index{out_x + pixel_base}; - data_ptrs[unswizzle ? 1 : 0] = swizzled_data + swizzle_offset; - data_ptrs[unswizzle ? 0 : 1] = unswizzled_data + pixel_index; - std::memcpy(data_ptrs[0], data_ptrs[1], FAST_SWIZZLE_ALIGN); - } - pixel_base += stride_x; - if ((y + 1) % GOB_SIZE_Y == 0) - y_address += GOB_SIZE; - } - z_address += xy_block_size; - } -} + const u32 gobs_in_x = Common::DivCeilLog2(stride, GOB_SIZE_X_SHIFT); + const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height + block_depth); + const u32 slice_size = + Common::DivCeilLog2(height, block_height + GOB_SIZE_Y_SHIFT) * block_size; -/** - * This function unswizzles or swizzles a texture by mapping Linear to BlockLinear Textue. - * The body of this function takes care of splitting the swizzled texture into blocks, - * and managing the extents of it. Once all the parameters of a single block are obtained, - * the function calls 'ProcessBlock' to process that particular Block. - * - * Documentation for the memory layout and decoding can be found at: - * https://envytools.readthedocs.io/en/latest/hw/memory/g80-surface.html#blocklinear-surfaces - */ -template <bool fast> -void SwizzledData(u8* const swizzled_data, u8* const unswizzled_data, const bool unswizzle, - const u32 width, const u32 height, const u32 depth, const u32 bytes_per_pixel, - const u32 out_bytes_per_pixel, const u32 block_height, const u32 block_depth, - const u32 width_spacing) { - auto div_ceil = [](const u32 x, const u32 y) { return ((x + y - 1) / y); }; - const u32 stride_x = width * out_bytes_per_pixel; - const u32 layer_z = height * stride_x; - const u32 gob_elements_x = GOB_SIZE_X / bytes_per_pixel; - constexpr u32 gob_elements_y = GOB_SIZE_Y; - constexpr u32 gob_elements_z = GOB_SIZE_Z; - const u32 block_x_elements = gob_elements_x; - const u32 block_y_elements = gob_elements_y * block_height; - const u32 block_z_elements = gob_elements_z * block_depth; - const u32 aligned_width = Common::AlignUp(width, gob_elements_x * width_spacing); - const u32 blocks_on_x = div_ceil(aligned_width, block_x_elements); - const u32 blocks_on_y = div_ceil(height, block_y_elements); - const u32 blocks_on_z = div_ceil(depth, block_z_elements); - const u32 xy_block_size = GOB_SIZE * block_height; - const u32 block_size = xy_block_size * block_depth; - u32 tile_offset = 0; - for (u32 zb = 0; zb < blocks_on_z; zb++) { - const u32 z_start = zb * block_z_elements; - const u32 z_end = std::min(depth, z_start + block_z_elements); - for (u32 yb = 0; yb < blocks_on_y; yb++) { - const u32 y_start = yb * block_y_elements; - const u32 y_end = std::min(height, y_start + block_y_elements); - for (u32 xb = 0; xb < blocks_on_x; xb++) { - const u32 x_start = xb * block_x_elements; - const u32 x_end = std::min(width, x_start + block_x_elements); - if constexpr (fast) { - FastProcessBlock(swizzled_data, unswizzled_data, unswizzle, x_start, y_start, - z_start, x_end, y_end, z_end, tile_offset, xy_block_size, - layer_z, stride_x, bytes_per_pixel, out_bytes_per_pixel); - } else { - PreciseProcessBlock(swizzled_data, unswizzled_data, unswizzle, x_start, y_start, - z_start, x_end, y_end, z_end, tile_offset, xy_block_size, - layer_z, stride_x, bytes_per_pixel, out_bytes_per_pixel); - } - tile_offset += block_size; + const u32 block_height_mask = (1U << block_height) - 1; + const u32 block_depth_mask = (1U << block_depth) - 1; + const u32 x_shift = GOB_SIZE_SHIFT + block_height + block_depth; + + for (u32 slice = 0; slice < depth; ++slice) { + const u32 z = slice + origin_z; + const u32 offset_z = (z >> block_depth) * slice_size + + ((z & block_depth_mask) << (GOB_SIZE_SHIFT + block_height)); + for (u32 line = 0; line < height; ++line) { + const u32 y = line + origin_y; + const auto& table = SWIZZLE_TABLE[y % GOB_SIZE_Y]; + + const u32 block_y = y >> GOB_SIZE_Y_SHIFT; + const u32 offset_y = (block_y >> block_height) * block_size + + ((block_y & block_height_mask) << GOB_SIZE_SHIFT); + + for (u32 column = 0; column < width; ++column) { + const u32 x = (column + origin_x) * bytes_per_pixel; + const u32 offset_x = (x >> GOB_SIZE_X_SHIFT) << x_shift; + + const u32 base_swizzled_offset = offset_z + offset_y + offset_x; + const u32 swizzled_offset = base_swizzled_offset + table[x % GOB_SIZE_X]; + + const u32 unswizzled_offset = + slice * pitch * height + line * pitch + column * bytes_per_pixel; + + u8* const dst = &output[TO_LINEAR ? swizzled_offset : unswizzled_offset]; + const u8* const src = &input[TO_LINEAR ? unswizzled_offset : swizzled_offset]; + std::memcpy(dst, src, bytes_per_pixel); } } } } - } // Anonymous namespace -void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel, - u32 out_bytes_per_pixel, u8* const swizzled_data, u8* const unswizzled_data, - bool unswizzle, u32 block_height, u32 block_depth, u32 width_spacing) { - const u32 block_height_size{1U << block_height}; - const u32 block_depth_size{1U << block_depth}; - if (bytes_per_pixel % 3 != 0 && (width * bytes_per_pixel) % FAST_SWIZZLE_ALIGN == 0) { - SwizzledData<true>(swizzled_data, unswizzled_data, unswizzle, width, height, depth, - bytes_per_pixel, out_bytes_per_pixel, block_height_size, - block_depth_size, width_spacing); - } else { - SwizzledData<false>(swizzled_data, unswizzled_data, unswizzle, width, height, depth, - bytes_per_pixel, out_bytes_per_pixel, block_height_size, - block_depth_size, width_spacing); - } +SwizzleTable MakeSwizzleTable() { + return SWIZZLE_TABLE; } -void UnswizzleTexture(u8* const unswizzled_data, u8* address, u32 tile_size_x, u32 tile_size_y, - u32 bytes_per_pixel, u32 width, u32 height, u32 depth, u32 block_height, - u32 block_depth, u32 width_spacing) { - CopySwizzledData((width + tile_size_x - 1) / tile_size_x, - (height + tile_size_y - 1) / tile_size_y, depth, bytes_per_pixel, - bytes_per_pixel, address, unswizzled_data, true, block_height, block_depth, - width_spacing); +void UnswizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, + u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth, + u32 stride_alignment) { + Swizzle<false>(output, input, bytes_per_pixel, width, height, depth, block_height, block_depth, + stride_alignment); } -std::vector<u8> UnswizzleTexture(u8* address, u32 tile_size_x, u32 tile_size_y, u32 bytes_per_pixel, - u32 width, u32 height, u32 depth, u32 block_height, - u32 block_depth, u32 width_spacing) { - std::vector<u8> unswizzled_data(width * height * depth * bytes_per_pixel); - UnswizzleTexture(unswizzled_data.data(), address, tile_size_x, tile_size_y, bytes_per_pixel, - width, height, depth, block_height, block_depth, width_spacing); - return unswizzled_data; +void SwizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, u32 width, + u32 height, u32 depth, u32 block_height, u32 block_depth, + u32 stride_alignment) { + Swizzle<true>(output, input, bytes_per_pixel, width, height, depth, block_height, block_depth, + stride_alignment); } void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 swizzled_width, @@ -213,7 +120,7 @@ void SwizzleSubrect(u32 subrect_width, u32 subrect_height, u32 source_pitch, u32 const u32 gob_address_y = (dst_y / (GOB_SIZE_Y * block_height)) * GOB_SIZE * block_height * image_width_in_gobs + ((dst_y % (GOB_SIZE_Y * block_height)) / GOB_SIZE_Y) * GOB_SIZE; - const auto& table = LEGACY_SWIZZLE_TABLE[dst_y % GOB_SIZE_Y]; + const auto& table = SWIZZLE_TABLE[dst_y % GOB_SIZE_Y]; for (u32 x = 0; x < subrect_width; ++x) { const u32 dst_x = x + offset_x; const u32 gob_address = @@ -235,11 +142,11 @@ void UnswizzleSubrect(u32 line_length_in, u32 line_count, u32 pitch, u32 width, const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height); const u32 block_height_mask = (1U << block_height) - 1; - const u32 x_shift = static_cast<u32>(GOB_SIZE_SHIFT) + block_height; + const u32 x_shift = GOB_SIZE_SHIFT + block_height; for (u32 line = 0; line < line_count; ++line) { const u32 src_y = line + origin_y; - const auto& table = LEGACY_SWIZZLE_TABLE[src_y % GOB_SIZE_Y]; + const auto& table = SWIZZLE_TABLE[src_y % GOB_SIZE_Y]; const u32 block_y = src_y >> GOB_SIZE_Y_SHIFT; const u32 src_offset_y = (block_y >> block_height) * block_size + @@ -270,7 +177,7 @@ void SwizzleSliceToVoxel(u32 line_length_in, u32 line_count, u32 pitch, u32 widt const u32 x_shift = static_cast<u32>(GOB_SIZE_SHIFT) + block_height + block_depth; for (u32 line = 0; line < line_count; ++line) { - const auto& table = LEGACY_SWIZZLE_TABLE[line % GOB_SIZE_Y]; + const auto& table = SWIZZLE_TABLE[line % GOB_SIZE_Y]; const u32 block_y = line / GOB_SIZE_Y; const u32 dst_offset_y = (block_y >> block_height) * block_size + (block_y & block_height_mask) * GOB_SIZE; @@ -293,7 +200,7 @@ void SwizzleKepler(const u32 width, const u32 height, const u32 dst_x, const u32 const std::size_t gob_address_y = (y / (GOB_SIZE_Y * block_height)) * GOB_SIZE * block_height * image_width_in_gobs + ((y % (GOB_SIZE_Y * block_height)) / GOB_SIZE_Y) * GOB_SIZE; - const auto& table = LEGACY_SWIZZLE_TABLE[y % GOB_SIZE_Y]; + const auto& table = SWIZZLE_TABLE[y % GOB_SIZE_Y]; for (std::size_t x = dst_x; x < width && count < copy_size; ++x) { const std::size_t gob_address = gob_address_y + (x / GOB_SIZE_X) * GOB_SIZE * block_height; diff --git a/src/video_core/textures/decoders.h b/src/video_core/textures/decoders.h index 01e156bc8..d7cdc81e8 100644 --- a/src/video_core/textures/decoders.h +++ b/src/video_core/textures/decoders.h @@ -4,7 +4,8 @@ #pragma once -#include <vector> +#include <span> + #include "common/common_types.h" #include "video_core/textures/texture.h" @@ -15,28 +16,25 @@ constexpr u32 GOB_SIZE_Y = 8; constexpr u32 GOB_SIZE_Z = 1; constexpr u32 GOB_SIZE = GOB_SIZE_X * GOB_SIZE_Y * GOB_SIZE_Z; -constexpr std::size_t GOB_SIZE_X_SHIFT = 6; -constexpr std::size_t GOB_SIZE_Y_SHIFT = 3; -constexpr std::size_t GOB_SIZE_Z_SHIFT = 0; -constexpr std::size_t GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHIFT; - -/// Unswizzles a swizzled texture without changing its format. -void UnswizzleTexture(u8* unswizzled_data, u8* address, u32 tile_size_x, u32 tile_size_y, - u32 bytes_per_pixel, u32 width, u32 height, u32 depth, - u32 block_height = TICEntry::DefaultBlockHeight, - u32 block_depth = TICEntry::DefaultBlockHeight, u32 width_spacing = 0); - -/// Unswizzles a swizzled texture without changing its format. -std::vector<u8> UnswizzleTexture(u8* address, u32 tile_size_x, u32 tile_size_y, u32 bytes_per_pixel, - u32 width, u32 height, u32 depth, - u32 block_height = TICEntry::DefaultBlockHeight, - u32 block_depth = TICEntry::DefaultBlockHeight, - u32 width_spacing = 0); - -/// Copies texture data from a buffer and performs swizzling/unswizzling as necessary. -void CopySwizzledData(u32 width, u32 height, u32 depth, u32 bytes_per_pixel, - u32 out_bytes_per_pixel, u8* swizzled_data, u8* unswizzled_data, - bool unswizzle, u32 block_height, u32 block_depth, u32 width_spacing); +constexpr u32 GOB_SIZE_X_SHIFT = 6; +constexpr u32 GOB_SIZE_Y_SHIFT = 3; +constexpr u32 GOB_SIZE_Z_SHIFT = 0; +constexpr u32 GOB_SIZE_SHIFT = GOB_SIZE_X_SHIFT + GOB_SIZE_Y_SHIFT + GOB_SIZE_Z_SHIFT; + +using SwizzleTable = std::array<std::array<u32, GOB_SIZE_X>, GOB_SIZE_Y>; + +/// Returns a z-order swizzle table +SwizzleTable MakeSwizzleTable(); + +/// Unswizzles a block linear texture into linear memory. +void UnswizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, + u32 width, u32 height, u32 depth, u32 block_height, u32 block_depth, + u32 stride_alignment = 1); + +/// Swizzles linear memory into a block linear texture. +void SwizzleTexture(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixel, u32 width, + u32 height, u32 depth, u32 block_height, u32 block_depth, + u32 stride_alignment = 1); /// This function calculates the correct size of a texture depending if it's tiled or not. std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height, u32 depth, diff --git a/src/video_core/textures/texture.cpp b/src/video_core/textures/texture.cpp index 4171e3ef2..ae5621a7d 100644 --- a/src/video_core/textures/texture.cpp +++ b/src/video_core/textures/texture.cpp @@ -5,9 +5,13 @@ #include <algorithm> #include <array> +#include "common/cityhash.h" #include "core/settings.h" #include "video_core/textures/texture.h" +using Tegra::Texture::TICEntry; +using Tegra::Texture::TSCEntry; + namespace Tegra::Texture { namespace { @@ -65,7 +69,7 @@ unsigned SettingsMinimumAnisotropy() noexcept { } // Anonymous namespace -std::array<float, 4> TSCEntry::GetBorderColor() const noexcept { +std::array<float, 4> TSCEntry::BorderColor() const noexcept { if (!srgb_conversion) { return border_color; } @@ -73,8 +77,16 @@ std::array<float, 4> TSCEntry::GetBorderColor() const noexcept { SRGB_CONVERSION_LUT[srgb_border_color_b], border_color[3]}; } -float TSCEntry::GetMaxAnisotropy() const noexcept { +float TSCEntry::MaxAnisotropy() const noexcept { return static_cast<float>(std::max(1U << max_anisotropy, SettingsMinimumAnisotropy())); } } // namespace Tegra::Texture + +size_t std::hash<TICEntry>::operator()(const TICEntry& tic) const noexcept { + return Common::CityHash64(reinterpret_cast<const char*>(&tic), sizeof tic); +} + +size_t std::hash<TSCEntry>::operator()(const TSCEntry& tsc) const noexcept { + return Common::CityHash64(reinterpret_cast<const char*>(&tsc), sizeof tsc); +} diff --git a/src/video_core/textures/texture.h b/src/video_core/textures/texture.h index 0574fef12..c1d14335e 100644 --- a/src/video_core/textures/texture.h +++ b/src/video_core/textures/texture.h @@ -53,27 +53,27 @@ enum class TextureFormat : u32 { BC4 = 0x27, BC5 = 0x28, S8D24 = 0x29, - X8Z24 = 0x2a, + X8D24 = 0x2a, D24S8 = 0x2b, - X4V4Z24__COV4R4V = 0x2c, - X4V4Z24__COV8R8V = 0x2d, - V8Z24__COV4R12V = 0x2e, + X4V4D24__COV4R4V = 0x2c, + X4V4D24__COV8R8V = 0x2d, + V8D24__COV4R12V = 0x2e, D32 = 0x2f, D32S8 = 0x30, - X8Z24_X20V4S8__COV4R4V = 0x31, - X8Z24_X20V4S8__COV8R8V = 0x32, - ZF32_X20V4X8__COV4R4V = 0x33, - ZF32_X20V4X8__COV8R8V = 0x34, - ZF32_X20V4S8__COV4R4V = 0x35, - ZF32_X20V4S8__COV8R8V = 0x36, - X8Z24_X16V8S8__COV4R12V = 0x37, - ZF32_X16V8X8__COV4R12V = 0x38, - ZF32_X16V8S8__COV4R12V = 0x39, + X8D24_X20V4S8__COV4R4V = 0x31, + X8D24_X20V4S8__COV8R8V = 0x32, + D32_X20V4X8__COV4R4V = 0x33, + D32_X20V4X8__COV8R8V = 0x34, + D32_X20V4S8__COV4R4V = 0x35, + D32_X20V4S8__COV8R8V = 0x36, + X8D24_X16V8S8__COV4R12V = 0x37, + D32_X16V8X8__COV4R12V = 0x38, + D32_X16V8S8__COV4R12V = 0x39, D16 = 0x3a, - V8Z24__COV8R24V = 0x3b, - X8Z24_X16V8S8__COV8R24V = 0x3c, - ZF32_X16V8X8__COV8R24V = 0x3d, - ZF32_X16V8S8__COV8R24V = 0x3e, + V8D24__COV8R24V = 0x3b, + X8D24_X16V8S8__COV8R24V = 0x3c, + D32_X16V8X8__COV8R24V = 0x3d, + D32_X16V8S8__COV8R24V = 0x3e, ASTC_2D_4X4 = 0x40, ASTC_2D_5X5 = 0x41, ASTC_2D_6X6 = 0x42, @@ -146,7 +146,7 @@ enum class MsaaMode : u32 { }; union TextureHandle { - TextureHandle(u32 raw) : raw{raw} {} + /* implicit */ constexpr TextureHandle(u32 raw_) : raw{raw_} {} u32 raw; BitField<0, 20, u32> tic_id; @@ -155,124 +155,124 @@ union TextureHandle { static_assert(sizeof(TextureHandle) == 4, "TextureHandle has wrong size"); struct TICEntry { - static constexpr u32 DefaultBlockHeight = 16; - static constexpr u32 DefaultBlockDepth = 1; - - union { - u32 raw; - BitField<0, 7, TextureFormat> format; - BitField<7, 3, ComponentType> r_type; - BitField<10, 3, ComponentType> g_type; - BitField<13, 3, ComponentType> b_type; - BitField<16, 3, ComponentType> a_type; - - BitField<19, 3, SwizzleSource> x_source; - BitField<22, 3, SwizzleSource> y_source; - BitField<25, 3, SwizzleSource> z_source; - BitField<28, 3, SwizzleSource> w_source; - }; - u32 address_low; union { - BitField<0, 16, u32> address_high; - BitField<21, 3, TICHeaderVersion> header_version; - }; - union { - BitField<0, 3, u32> block_width; - BitField<3, 3, u32> block_height; - BitField<6, 3, u32> block_depth; + struct { + union { + BitField<0, 7, TextureFormat> format; + BitField<7, 3, ComponentType> r_type; + BitField<10, 3, ComponentType> g_type; + BitField<13, 3, ComponentType> b_type; + BitField<16, 3, ComponentType> a_type; + + BitField<19, 3, SwizzleSource> x_source; + BitField<22, 3, SwizzleSource> y_source; + BitField<25, 3, SwizzleSource> z_source; + BitField<28, 3, SwizzleSource> w_source; + }; + u32 address_low; + union { + BitField<0, 16, u32> address_high; + BitField<16, 5, u32> layer_base_3_7; + BitField<21, 3, TICHeaderVersion> header_version; + BitField<24, 1, u32> load_store_hint; + BitField<25, 4, u32> view_coherency_hash; + BitField<29, 3, u32> layer_base_8_10; + }; + union { + BitField<0, 3, u32> block_width; + BitField<3, 3, u32> block_height; + BitField<6, 3, u32> block_depth; - BitField<10, 3, u32> tile_width_spacing; + BitField<10, 3, u32> tile_width_spacing; - // High 16 bits of the pitch value - BitField<0, 16, u32> pitch_high; - BitField<26, 1, u32> use_header_opt_control; - BitField<27, 1, u32> depth_texture; - BitField<28, 4, u32> max_mip_level; + // High 16 bits of the pitch value + BitField<0, 16, u32> pitch_high; + BitField<26, 1, u32> use_header_opt_control; + BitField<27, 1, u32> depth_texture; + BitField<28, 4, u32> max_mip_level; - BitField<0, 16, u32> buffer_high_width_minus_one; - }; - union { - BitField<0, 16, u32> width_minus_1; - BitField<22, 1, u32> srgb_conversion; - BitField<23, 4, TextureType> texture_type; - BitField<29, 3, u32> border_size; + BitField<0, 16, u32> buffer_high_width_minus_one; + }; + union { + BitField<0, 16, u32> width_minus_one; + BitField<16, 3, u32> layer_base_0_2; + BitField<22, 1, u32> srgb_conversion; + BitField<23, 4, TextureType> texture_type; + BitField<29, 3, u32> border_size; - BitField<0, 16, u32> buffer_low_width_minus_one; - }; - union { - BitField<0, 16, u32> height_minus_1; - BitField<16, 14, u32> depth_minus_1; - }; - union { - BitField<6, 13, u32> mip_lod_bias; - BitField<27, 3, u32> max_anisotropy; + BitField<0, 16, u32> buffer_low_width_minus_one; + }; + union { + BitField<0, 16, u32> height_minus_1; + BitField<16, 14, u32> depth_minus_1; + BitField<30, 1, u32> is_sparse; + BitField<31, 1, u32> normalized_coords; + }; + union { + BitField<6, 13, u32> mip_lod_bias; + BitField<27, 3, u32> max_anisotropy; + }; + union { + BitField<0, 4, u32> res_min_mip_level; + BitField<4, 4, u32> res_max_mip_level; + BitField<8, 4, MsaaMode> msaa_mode; + BitField<12, 12, u32> min_lod_clamp; + }; + }; + std::array<u64, 4> raw; }; - union { - BitField<0, 4, u32> res_min_mip_level; - BitField<4, 4, u32> res_max_mip_level; - BitField<8, 4, MsaaMode> msaa_mode; - BitField<12, 12, u32> min_lod_clamp; - }; + constexpr bool operator==(const TICEntry& rhs) const noexcept { + return raw == rhs.raw; + } - GPUVAddr Address() const { + constexpr bool operator!=(const TICEntry& rhs) const noexcept { + return raw != rhs.raw; + } + + constexpr GPUVAddr Address() const { return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) | address_low); } - u32 Pitch() const { + constexpr u32 Pitch() const { ASSERT(header_version == TICHeaderVersion::Pitch || header_version == TICHeaderVersion::PitchColorKey); // The pitch value is 21 bits, and is 32B aligned. return pitch_high << 5; } - u32 Width() const { + constexpr u32 Width() const { if (header_version != TICHeaderVersion::OneDBuffer) { - return width_minus_1 + 1; + return width_minus_one + 1; } - return ((buffer_high_width_minus_one << 16) | buffer_low_width_minus_one) + 1; + return (buffer_high_width_minus_one << 16 | buffer_low_width_minus_one) + 1; } - u32 Height() const { + constexpr u32 Height() const { return height_minus_1 + 1; } - u32 Depth() const { + constexpr u32 Depth() const { return depth_minus_1 + 1; } - u32 BlockWidth() const { - ASSERT(IsTiled()); - return block_width; - } - - u32 BlockHeight() const { - ASSERT(IsTiled()); - return block_height; - } - - u32 BlockDepth() const { - ASSERT(IsTiled()); - return block_depth; + constexpr u32 BaseLayer() const { + return layer_base_0_2 | layer_base_3_7 << 3 | layer_base_8_10 << 8; } - bool IsTiled() const { + constexpr bool IsBlockLinear() const { return header_version == TICHeaderVersion::BlockLinear || header_version == TICHeaderVersion::BlockLinearColorKey; } - bool IsLineal() const { + constexpr bool IsPitchLinear() const { return header_version == TICHeaderVersion::Pitch || header_version == TICHeaderVersion::PitchColorKey; } - bool IsBuffer() const { + constexpr bool IsBuffer() const { return header_version == TICHeaderVersion::OneDBuffer; } - - bool IsSrgbConversionEnabled() const { - return srgb_conversion != 0; - } }; static_assert(sizeof(TICEntry) == 0x20, "TICEntry has wrong size"); @@ -309,6 +309,12 @@ enum class TextureMipmapFilter : u32 { Linear = 3, }; +enum class SamplerReduction : u32 { + WeightedAverage = 0, + Min = 1, + Max = 2, +}; + enum class Anisotropy { Default, Filter2x, @@ -333,8 +339,12 @@ struct TSCEntry { BitField<0, 2, TextureFilter> mag_filter; BitField<4, 2, TextureFilter> min_filter; BitField<6, 2, TextureMipmapFilter> mipmap_filter; + BitField<8, 1, u32> cubemap_anisotropy; BitField<9, 1, u32> cubemap_interface_filtering; + BitField<10, 2, SamplerReduction> reduction_filter; BitField<12, 13, u32> mip_lod_bias; + BitField<25, 1, u32> float_coord_normalization; + BitField<26, 5, u32> trilin_opt; }; union { BitField<0, 12, u32> min_lod_clamp; @@ -347,32 +357,45 @@ struct TSCEntry { }; std::array<f32, 4> border_color; }; - std::array<u8, 0x20> raw; + std::array<u64, 4> raw; }; - std::array<float, 4> GetBorderColor() const noexcept; + constexpr bool operator==(const TSCEntry& rhs) const noexcept { + return raw == rhs.raw; + } + + constexpr bool operator!=(const TSCEntry& rhs) const noexcept { + return raw != rhs.raw; + } + + std::array<float, 4> BorderColor() const noexcept; - float GetMaxAnisotropy() const noexcept; + float MaxAnisotropy() const noexcept; - float GetMinLod() const { + float MinLod() const { return static_cast<float>(min_lod_clamp) / 256.0f; } - float GetMaxLod() const { + float MaxLod() const { return static_cast<float>(max_lod_clamp) / 256.0f; } - float GetLodBias() const { + float LodBias() const { // Sign extend the 13-bit value. - constexpr u32 mask = 1U << (13 - 1); + static constexpr u32 mask = 1U << (13 - 1); return static_cast<float>(static_cast<s32>((mip_lod_bias ^ mask) - mask)) / 256.0f; } }; static_assert(sizeof(TSCEntry) == 0x20, "TSCEntry has wrong size"); -struct FullTextureInfo { - TICEntry tic; - TSCEntry tsc; +} // namespace Tegra::Texture + +template <> +struct std::hash<Tegra::Texture::TICEntry> { + size_t operator()(const Tegra::Texture::TICEntry& tic) const noexcept; }; -} // namespace Tegra::Texture +template <> +struct std::hash<Tegra::Texture::TSCEntry> { + size_t operator()(const Tegra::Texture::TSCEntry& tsc) const noexcept; +}; diff --git a/src/video_core/video_core.cpp b/src/video_core/video_core.cpp index a14df06a3..53444e945 100644 --- a/src/video_core/video_core.cpp +++ b/src/video_core/video_core.cpp @@ -7,13 +7,9 @@ #include "common/logging/log.h" #include "core/core.h" #include "core/settings.h" -#include "video_core/gpu_asynch.h" -#include "video_core/gpu_synch.h" #include "video_core/renderer_base.h" #include "video_core/renderer_opengl/renderer_opengl.h" -#ifdef HAS_VULKAN #include "video_core/renderer_vulkan/renderer_vulkan.h" -#endif #include "video_core/video_core.h" namespace { @@ -28,11 +24,9 @@ std::unique_ptr<VideoCore::RendererBase> CreateRenderer( case Settings::RendererBackend::OpenGL: return std::make_unique<OpenGL::RendererOpenGL>(telemetry_session, emu_window, cpu_memory, gpu, std::move(context)); -#ifdef HAS_VULKAN case Settings::RendererBackend::Vulkan: return std::make_unique<Vulkan::RendererVulkan>(telemetry_session, emu_window, cpu_memory, gpu, std::move(context)); -#endif default: return nullptr; } @@ -43,12 +37,9 @@ std::unique_ptr<VideoCore::RendererBase> CreateRenderer( namespace VideoCore { std::unique_ptr<Tegra::GPU> CreateGPU(Core::Frontend::EmuWindow& emu_window, Core::System& system) { - std::unique_ptr<Tegra::GPU> gpu; - if (Settings::values.use_asynchronous_gpu_emulation.GetValue()) { - gpu = std::make_unique<VideoCommon::GPUAsynch>(system); - } else { - gpu = std::make_unique<VideoCommon::GPUSynch>(system); - } + const bool use_nvdec = Settings::values.use_nvdec_emulation.GetValue(); + std::unique_ptr<Tegra::GPU> gpu = std::make_unique<Tegra::GPU>( + system, Settings::values.use_asynchronous_gpu_emulation.GetValue(), use_nvdec); auto context = emu_window.CreateSharedContext(); const auto scope = context->Acquire(); diff --git a/src/video_core/renderer_vulkan/nsight_aftermath_tracker.cpp b/src/video_core/vulkan_common/nsight_aftermath_tracker.cpp index 5b01020ec..8d10ac29e 100644 --- a/src/video_core/renderer_vulkan/nsight_aftermath_tracker.cpp +++ b/src/video_core/vulkan_common/nsight_aftermath_tracker.cpp @@ -32,20 +32,11 @@ namespace Vulkan { static constexpr char AFTERMATH_LIB_NAME[] = "GFSDK_Aftermath_Lib.x64.dll"; -NsightAftermathTracker::NsightAftermathTracker() = default; - -NsightAftermathTracker::~NsightAftermathTracker() { - if (initialized) { - (void)GFSDK_Aftermath_DisableGpuCrashDumps(); - } -} - -bool NsightAftermathTracker::Initialize() { +NsightAftermathTracker::NsightAftermathTracker() { if (!dl.Open(AFTERMATH_LIB_NAME)) { LOG_ERROR(Render_Vulkan, "Failed to load Nsight Aftermath DLL"); - return false; + return; } - if (!dl.GetSymbol("GFSDK_Aftermath_DisableGpuCrashDumps", &GFSDK_Aftermath_DisableGpuCrashDumps) || !dl.GetSymbol("GFSDK_Aftermath_EnableGpuCrashDumps", @@ -64,27 +55,28 @@ bool NsightAftermathTracker::Initialize() { LOG_ERROR(Render_Vulkan, "Failed to load Nsight Aftermath function pointers"); return false; } - dump_dir = Common::FS::GetUserPath(Common::FS::UserPath::LogDir) + "gpucrash"; - (void)Common::FS::DeleteDirRecursively(dump_dir); + void(Common::FS::DeleteDirRecursively(dump_dir)); if (!Common::FS::CreateDir(dump_dir)) { LOG_ERROR(Render_Vulkan, "Failed to create Nsight Aftermath dump directory"); - return false; + return; } - if (!GFSDK_Aftermath_SUCCEED(GFSDK_Aftermath_EnableGpuCrashDumps( GFSDK_Aftermath_Version_API, GFSDK_Aftermath_GpuCrashDumpWatchedApiFlags_Vulkan, GFSDK_Aftermath_GpuCrashDumpFeatureFlags_Default, GpuCrashDumpCallback, ShaderDebugInfoCallback, CrashDumpDescriptionCallback, this))) { LOG_ERROR(Render_Vulkan, "GFSDK_Aftermath_EnableGpuCrashDumps failed"); - return false; + return; } - LOG_INFO(Render_Vulkan, "Nsight Aftermath dump directory is \"{}\"", dump_dir); - initialized = true; - return true; +} + +NsightAftermathTracker::~NsightAftermathTracker() { + if (initialized) { + (void)GFSDK_Aftermath_DisableGpuCrashDumps(); + } } void NsightAftermathTracker::SaveShader(const std::vector<u32>& spirv) const { diff --git a/src/video_core/renderer_vulkan/nsight_aftermath_tracker.h b/src/video_core/vulkan_common/nsight_aftermath_tracker.h index afe7ae99e..cee3847fb 100644 --- a/src/video_core/renderer_vulkan/nsight_aftermath_tracker.h +++ b/src/video_core/vulkan_common/nsight_aftermath_tracker.h @@ -34,8 +34,6 @@ public: NsightAftermathTracker(NsightAftermathTracker&&) = delete; NsightAftermathTracker& operator=(NsightAftermathTracker&&) = delete; - bool Initialize(); - void SaveShader(const std::vector<u32>& spirv) const; private: @@ -78,9 +76,6 @@ private: #ifndef HAS_NSIGHT_AFTERMATH inline NsightAftermathTracker::NsightAftermathTracker() = default; inline NsightAftermathTracker::~NsightAftermathTracker() = default; -inline bool NsightAftermathTracker::Initialize() { - return false; -} inline void NsightAftermathTracker::SaveShader(const std::vector<u32>&) const {} #endif diff --git a/src/video_core/vulkan_common/vulkan_debug_callback.cpp b/src/video_core/vulkan_common/vulkan_debug_callback.cpp new file mode 100644 index 000000000..ea7af8ad4 --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_debug_callback.cpp @@ -0,0 +1,45 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <string_view> +#include "common/logging/log.h" +#include "video_core/vulkan_common/vulkan_debug_callback.h" + +namespace Vulkan { +namespace { +VkBool32 Callback(VkDebugUtilsMessageSeverityFlagBitsEXT severity, + VkDebugUtilsMessageTypeFlagsEXT type, + const VkDebugUtilsMessengerCallbackDataEXT* data, + [[maybe_unused]] void* user_data) { + const std::string_view message{data->pMessage}; + if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT) { + LOG_CRITICAL(Render_Vulkan, "{}", message); + } else if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT) { + LOG_WARNING(Render_Vulkan, "{}", message); + } else if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT) { + LOG_INFO(Render_Vulkan, "{}", message); + } else if (severity & VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT) { + LOG_DEBUG(Render_Vulkan, "{}", message); + } + return VK_FALSE; +} +} // Anonymous namespace + +vk::DebugUtilsMessenger CreateDebugCallback(const vk::Instance& instance) { + return instance.CreateDebugUtilsMessenger(VkDebugUtilsMessengerCreateInfoEXT{ + .sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_MESSENGER_CREATE_INFO_EXT, + .pNext = nullptr, + .flags = 0, + .messageSeverity = VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT, + .messageType = VK_DEBUG_UTILS_MESSAGE_TYPE_GENERAL_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_TYPE_VALIDATION_BIT_EXT | + VK_DEBUG_UTILS_MESSAGE_TYPE_PERFORMANCE_BIT_EXT, + .pfnUserCallback = Callback, + }); +} + +} // namespace Vulkan diff --git a/src/video_core/vulkan_common/vulkan_debug_callback.h b/src/video_core/vulkan_common/vulkan_debug_callback.h new file mode 100644 index 000000000..2efcd244c --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_debug_callback.h @@ -0,0 +1,11 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "video_core/vulkan_common/vulkan_wrapper.h" + +namespace Vulkan { + +vk::DebugUtilsMessenger CreateDebugCallback(const vk::Instance& instance); + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/vk_device.cpp b/src/video_core/vulkan_common/vulkan_device.cpp index 05e31f1de..75173324e 100644 --- a/src/video_core/renderer_vulkan/vk_device.cpp +++ b/src/video_core/vulkan_common/vulkan_device.cpp @@ -13,8 +13,9 @@ #include "common/assert.h" #include "core/settings.h" -#include "video_core/renderer_vulkan/vk_device.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/nsight_aftermath_tracker.h" +#include "video_core/vulkan_common/vulkan_device.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { @@ -38,11 +39,15 @@ constexpr std::array Depth16UnormS8_UINT{ constexpr std::array REQUIRED_EXTENSIONS{ VK_KHR_SWAPCHAIN_EXTENSION_NAME, + VK_KHR_MAINTENANCE1_EXTENSION_NAME, + VK_KHR_STORAGE_BUFFER_STORAGE_CLASS_EXTENSION_NAME, + VK_KHR_SHADER_DRAW_PARAMETERS_EXTENSION_NAME, VK_KHR_16BIT_STORAGE_EXTENSION_NAME, VK_KHR_8BIT_STORAGE_EXTENSION_NAME, VK_KHR_DRIVER_PROPERTIES_EXTENSION_NAME, VK_KHR_DESCRIPTOR_UPDATE_TEMPLATE_EXTENSION_NAME, VK_KHR_TIMELINE_SEMAPHORE_EXTENSION_NAME, + VK_KHR_SAMPLER_MIRROR_CLAMP_TO_EDGE_EXTENSION_NAME, VK_EXT_VERTEX_ATTRIBUTE_DIVISOR_EXTENSION_NAME, VK_EXT_SHADER_SUBGROUP_BALLOT_EXTENSION_NAME, VK_EXT_SHADER_SUBGROUP_VOTE_EXTENSION_NAME, @@ -79,6 +84,21 @@ VkFormatFeatureFlags GetFormatFeatures(VkFormatProperties properties, FormatType } } +[[nodiscard]] bool IsRDNA(std::string_view device_name, VkDriverIdKHR driver_id) { + static constexpr std::array RDNA_DEVICES{ + "5700", + "5600", + "5500", + "5300", + }; + if (driver_id != VK_DRIVER_ID_AMD_PROPRIETARY_KHR) { + return false; + } + return std::any_of(RDNA_DEVICES.begin(), RDNA_DEVICES.end(), [device_name](const char* name) { + return device_name.find(name) != std::string_view::npos; + }); +} + std::unordered_map<VkFormat, VkFormatProperties> GetFormatProperties( vk::PhysicalDevice physical, const vk::InstanceDispatch& dld) { static constexpr std::array formats{ @@ -104,6 +124,7 @@ std::unordered_map<VkFormat, VkFormatProperties> GetFormatProperties( VK_FORMAT_R16G16_UNORM, VK_FORMAT_R16G16_SNORM, VK_FORMAT_R16G16_SFLOAT, + VK_FORMAT_R16G16_SINT, VK_FORMAT_R16_UNORM, VK_FORMAT_R16_UINT, VK_FORMAT_R8G8B8A8_SRGB, @@ -143,18 +164,32 @@ std::unordered_map<VkFormat, VkFormatProperties> GetFormatProperties( VK_FORMAT_BC2_SRGB_BLOCK, VK_FORMAT_BC3_SRGB_BLOCK, VK_FORMAT_BC7_SRGB_BLOCK, + VK_FORMAT_ASTC_4x4_UNORM_BLOCK, VK_FORMAT_ASTC_4x4_SRGB_BLOCK, - VK_FORMAT_ASTC_8x8_SRGB_BLOCK, - VK_FORMAT_ASTC_8x5_SRGB_BLOCK, + VK_FORMAT_ASTC_5x4_UNORM_BLOCK, VK_FORMAT_ASTC_5x4_SRGB_BLOCK, VK_FORMAT_ASTC_5x5_UNORM_BLOCK, VK_FORMAT_ASTC_5x5_SRGB_BLOCK, - VK_FORMAT_ASTC_10x8_UNORM_BLOCK, - VK_FORMAT_ASTC_10x8_SRGB_BLOCK, + VK_FORMAT_ASTC_6x5_UNORM_BLOCK, + VK_FORMAT_ASTC_6x5_SRGB_BLOCK, VK_FORMAT_ASTC_6x6_UNORM_BLOCK, VK_FORMAT_ASTC_6x6_SRGB_BLOCK, + VK_FORMAT_ASTC_8x5_UNORM_BLOCK, + VK_FORMAT_ASTC_8x5_SRGB_BLOCK, + VK_FORMAT_ASTC_8x6_UNORM_BLOCK, + VK_FORMAT_ASTC_8x6_SRGB_BLOCK, + VK_FORMAT_ASTC_8x8_UNORM_BLOCK, + VK_FORMAT_ASTC_8x8_SRGB_BLOCK, + VK_FORMAT_ASTC_10x5_UNORM_BLOCK, + VK_FORMAT_ASTC_10x5_SRGB_BLOCK, + VK_FORMAT_ASTC_10x6_UNORM_BLOCK, + VK_FORMAT_ASTC_10x6_SRGB_BLOCK, + VK_FORMAT_ASTC_10x8_UNORM_BLOCK, + VK_FORMAT_ASTC_10x8_SRGB_BLOCK, VK_FORMAT_ASTC_10x10_UNORM_BLOCK, VK_FORMAT_ASTC_10x10_SRGB_BLOCK, + VK_FORMAT_ASTC_12x10_UNORM_BLOCK, + VK_FORMAT_ASTC_12x10_SRGB_BLOCK, VK_FORMAT_ASTC_12x12_UNORM_BLOCK, VK_FORMAT_ASTC_12x12_SRGB_BLOCK, VK_FORMAT_ASTC_8x6_UNORM_BLOCK, @@ -172,17 +207,14 @@ std::unordered_map<VkFormat, VkFormatProperties> GetFormatProperties( } // Anonymous namespace -VKDevice::VKDevice(VkInstance instance, vk::PhysicalDevice physical, VkSurfaceKHR surface, - const vk::InstanceDispatch& dld) - : dld{dld}, physical{physical}, properties{physical.GetProperties()}, +Device::Device(VkInstance instance_, vk::PhysicalDevice physical_, VkSurfaceKHR surface, + const vk::InstanceDispatch& dld_) + : instance{instance_}, dld{dld_}, physical{physical_}, properties{physical.GetProperties()}, format_properties{GetFormatProperties(physical, dld)} { + CheckSuitability(); SetupFamilies(surface); SetupFeatures(); -} - -VKDevice::~VKDevice() = default; -bool VKDevice::Create() { const auto queue_cis = GetDeviceQueueCreateInfos(); const std::vector extensions = LoadExtensions(); @@ -196,7 +228,7 @@ bool VKDevice::Create() { features2.features = { .robustBufferAccess = false, .fullDrawIndexUint32 = false, - .imageCubeArray = false, + .imageCubeArray = true, .independentBlend = true, .geometryShader = true, .tessellationShader = true, @@ -224,7 +256,7 @@ bool VKDevice::Create() { .shaderTessellationAndGeometryPointSize = false, .shaderImageGatherExtended = true, .shaderStorageImageExtendedFormats = false, - .shaderStorageImageMultisample = false, + .shaderStorageImageMultisample = true, .shaderStorageImageReadWithoutFormat = is_formatless_image_load_supported, .shaderStorageImageWriteWithoutFormat = true, .shaderUniformBufferArrayDynamicIndexing = false, @@ -250,7 +282,6 @@ bool VKDevice::Create() { .variableMultisampleRate = false, .inheritedQueries = false, }; - VkPhysicalDeviceTimelineSemaphoreFeaturesKHR timeline_semaphore{ .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TIMELINE_SEMAPHORE_FEATURES_KHR, .pNext = nullptr, @@ -362,13 +393,27 @@ bool VKDevice::Create() { LOG_INFO(Render_Vulkan, "Device doesn't support extended dynamic state"); } + VkPhysicalDeviceRobustness2FeaturesEXT robustness2; + if (ext_robustness2) { + robustness2 = { + .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_ROBUSTNESS_2_FEATURES_EXT, + .pNext = nullptr, + .robustBufferAccess2 = false, + .robustImageAccess2 = true, + .nullDescriptor = true, + }; + SetNext(next, robustness2); + } else { + LOG_INFO(Render_Vulkan, "Device doesn't support robustness2"); + } + if (!ext_depth_range_unrestricted) { LOG_INFO(Render_Vulkan, "Device doesn't support depth range unrestricted"); } VkDeviceDiagnosticsConfigCreateInfoNV diagnostics_nv; if (nv_device_diagnostics_config) { - nsight_aftermath_tracker.Initialize(); + nsight_aftermath_tracker = std::make_unique<NsightAftermathTracker>(); diagnostics_nv = { .sType = VK_STRUCTURE_TYPE_DEVICE_DIAGNOSTICS_CONFIG_CREATE_INFO_NV, @@ -379,20 +424,23 @@ bool VKDevice::Create() { }; first_next = &diagnostics_nv; } - logical = vk::Device::Create(physical, queue_cis, extensions, first_next, dld); - if (!logical) { - LOG_ERROR(Render_Vulkan, "Failed to create logical device"); - return false; - } CollectTelemetryParameters(); + CollectToolingInfo(); - if (ext_extended_dynamic_state && driver_id == VK_DRIVER_ID_AMD_PROPRIETARY_KHR) { - // AMD's proprietary driver supports VK_EXT_extended_dynamic_state but the <stride> field - // seems to be bugged. Blacklisting it for now. - LOG_WARNING(Render_Vulkan, - "Blacklisting AMD proprietary from VK_EXT_extended_dynamic_state"); + if (ext_extended_dynamic_state && driver_id == VK_DRIVER_ID_MESA_RADV) { + LOG_WARNING( + Render_Vulkan, + "Blacklisting RADV for VK_EXT_extended_dynamic state, likely due to a bug in yuzu"); + ext_extended_dynamic_state = false; + } + if (ext_extended_dynamic_state && IsRDNA(properties.deviceName, driver_id)) { + // AMD's proprietary driver supports VK_EXT_extended_dynamic_state but on RDNA devices it + // seems to cause stability issues + LOG_WARNING( + Render_Vulkan, + "Blacklisting AMD proprietary on RDNA devices from VK_EXT_extended_dynamic_state"); ext_extended_dynamic_state = false; } @@ -400,11 +448,12 @@ bool VKDevice::Create() { present_queue = logical.GetQueue(present_family); use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue(); - return true; } -VkFormat VKDevice::GetSupportedFormat(VkFormat wanted_format, VkFormatFeatureFlags wanted_usage, - FormatType format_type) const { +Device::~Device() = default; + +VkFormat Device::GetSupportedFormat(VkFormat wanted_format, VkFormatFeatureFlags wanted_usage, + FormatType format_type) const { if (IsFormatSupported(wanted_format, wanted_usage, format_type)) { return wanted_format; } @@ -435,18 +484,20 @@ VkFormat VKDevice::GetSupportedFormat(VkFormat wanted_format, VkFormatFeatureFla return wanted_format; } -void VKDevice::ReportLoss() const { +void Device::ReportLoss() const { LOG_CRITICAL(Render_Vulkan, "Device loss occured!"); // Wait for the log to flush and for Nsight Aftermath to dump the results - std::this_thread::sleep_for(std::chrono::seconds{3}); + std::this_thread::sleep_for(std::chrono::seconds{15}); } -void VKDevice::SaveShader(const std::vector<u32>& spirv) const { - nsight_aftermath_tracker.SaveShader(spirv); +void Device::SaveShader(const std::vector<u32>& spirv) const { + if (nsight_aftermath_tracker) { + nsight_aftermath_tracker->SaveShader(spirv); + } } -bool VKDevice::IsOptimalAstcSupported(const VkPhysicalDeviceFeatures& features) const { +bool Device::IsOptimalAstcSupported(const VkPhysicalDeviceFeatures& features) const { // Disable for now to avoid converting ASTC twice. static constexpr std::array astc_formats = { VK_FORMAT_ASTC_4x4_UNORM_BLOCK, VK_FORMAT_ASTC_4x4_SRGB_BLOCK, @@ -472,16 +523,26 @@ bool VKDevice::IsOptimalAstcSupported(const VkPhysicalDeviceFeatures& features) VK_FORMAT_FEATURE_BLIT_DST_BIT | VK_FORMAT_FEATURE_TRANSFER_SRC_BIT | VK_FORMAT_FEATURE_TRANSFER_DST_BIT}; for (const auto format : astc_formats) { - const auto format_properties{physical.GetFormatProperties(format)}; - if (!(format_properties.optimalTilingFeatures & format_feature_usage)) { + const auto physical_format_properties{physical.GetFormatProperties(format)}; + if ((physical_format_properties.optimalTilingFeatures & format_feature_usage) == 0) { return false; } } return true; } -bool VKDevice::IsFormatSupported(VkFormat wanted_format, VkFormatFeatureFlags wanted_usage, - FormatType format_type) const { +bool Device::TestDepthStencilBlits() const { + static constexpr VkFormatFeatureFlags required_features = + VK_FORMAT_FEATURE_BLIT_SRC_BIT | VK_FORMAT_FEATURE_BLIT_DST_BIT; + const auto test_features = [](VkFormatProperties props) { + return (props.optimalTilingFeatures & required_features) == required_features; + }; + return test_features(format_properties.at(VK_FORMAT_D32_SFLOAT_S8_UINT)) && + test_features(format_properties.at(VK_FORMAT_D24_UNORM_S8_UINT)); +} + +bool Device::IsFormatSupported(VkFormat wanted_format, VkFormatFeatureFlags wanted_usage, + FormatType format_type) const { const auto it = format_properties.find(wanted_format); if (it == format_properties.end()) { UNIMPLEMENTED_MSG("Unimplemented format query={}", wanted_format); @@ -491,65 +552,47 @@ bool VKDevice::IsFormatSupported(VkFormat wanted_format, VkFormatFeatureFlags wa return (supported_usage & wanted_usage) == wanted_usage; } -bool VKDevice::IsSuitable(vk::PhysicalDevice physical, VkSurfaceKHR surface) { - bool is_suitable = true; +void Device::CheckSuitability() const { std::bitset<REQUIRED_EXTENSIONS.size()> available_extensions; - - for (const auto& prop : physical.EnumerateDeviceExtensionProperties()) { + for (const VkExtensionProperties& property : physical.EnumerateDeviceExtensionProperties()) { for (std::size_t i = 0; i < REQUIRED_EXTENSIONS.size(); ++i) { if (available_extensions[i]) { continue; } - const std::string_view name{prop.extensionName}; + const std::string_view name{property.extensionName}; available_extensions[i] = name == REQUIRED_EXTENSIONS[i]; } } - if (!available_extensions.all()) { - for (std::size_t i = 0; i < REQUIRED_EXTENSIONS.size(); ++i) { - if (available_extensions[i]) { - continue; - } - LOG_ERROR(Render_Vulkan, "Missing required extension: {}", REQUIRED_EXTENSIONS[i]); - is_suitable = false; - } - } - - bool has_graphics{}, has_present{}; - const std::vector queue_family_properties = physical.GetQueueFamilyProperties(); - for (u32 i = 0; i < static_cast<u32>(queue_family_properties.size()); ++i) { - const auto& family = queue_family_properties[i]; - if (family.queueCount == 0) { + for (size_t i = 0; i < REQUIRED_EXTENSIONS.size(); ++i) { + if (available_extensions[i]) { continue; } - has_graphics |= family.queueFlags & VK_QUEUE_GRAPHICS_BIT; - has_present |= physical.GetSurfaceSupportKHR(i, surface); + LOG_ERROR(Render_Vulkan, "Missing required extension: {}", REQUIRED_EXTENSIONS[i]); + throw vk::Exception(VK_ERROR_EXTENSION_NOT_PRESENT); } - if (!has_graphics || !has_present) { - LOG_ERROR(Render_Vulkan, "Device lacks a graphics and present queue"); - is_suitable = false; - } - - // TODO(Rodrigo): Check if the device matches all requeriments. - const auto properties{physical.GetProperties()}; - const auto& limits{properties.limits}; - - constexpr u32 required_ubo_size = 65536; - if (limits.maxUniformBufferRange < required_ubo_size) { - LOG_ERROR(Render_Vulkan, "Device UBO size {} is too small, {} is required", - limits.maxUniformBufferRange, required_ubo_size); - is_suitable = false; - } - - constexpr u32 required_num_viewports = 16; - if (limits.maxViewports < required_num_viewports) { - LOG_INFO(Render_Vulkan, "Device number of viewports {} is too small, {} is required", - limits.maxViewports, required_num_viewports); - is_suitable = false; + struct LimitTuple { + u32 minimum; + u32 value; + const char* name; + }; + const VkPhysicalDeviceLimits& limits{properties.limits}; + const std::array limits_report{ + LimitTuple{65536, limits.maxUniformBufferRange, "maxUniformBufferRange"}, + LimitTuple{16, limits.maxViewports, "maxViewports"}, + LimitTuple{8, limits.maxColorAttachments, "maxColorAttachments"}, + LimitTuple{8, limits.maxClipDistances, "maxClipDistances"}, + }; + for (const auto& tuple : limits_report) { + if (tuple.value < tuple.minimum) { + LOG_ERROR(Render_Vulkan, "{} has to be {} or greater but it is {}", tuple.name, + tuple.minimum, tuple.value); + throw vk::Exception(VK_ERROR_FEATURE_NOT_PRESENT); + } } - - const auto features{physical.GetFeatures()}; - const std::array feature_report = { + const VkPhysicalDeviceFeatures features{physical.GetFeatures()}; + const std::array feature_report{ std::make_pair(features.vertexPipelineStoresAndAtomics, "vertexPipelineStoresAndAtomics"), + std::make_pair(features.imageCubeArray, "imageCubeArray"), std::make_pair(features.independentBlend, "independentBlend"), std::make_pair(features.depthClamp, "depthClamp"), std::make_pair(features.samplerAnisotropy, "samplerAnisotropy"), @@ -561,40 +604,21 @@ bool VKDevice::IsSuitable(vk::PhysicalDevice physical, VkSurfaceKHR surface) { std::make_pair(features.occlusionQueryPrecise, "occlusionQueryPrecise"), std::make_pair(features.fragmentStoresAndAtomics, "fragmentStoresAndAtomics"), std::make_pair(features.shaderImageGatherExtended, "shaderImageGatherExtended"), + std::make_pair(features.shaderStorageImageMultisample, "shaderStorageImageMultisample"), std::make_pair(features.shaderStorageImageWriteWithoutFormat, "shaderStorageImageWriteWithoutFormat"), }; - for (const auto& [supported, name] : feature_report) { - if (supported) { + for (const auto& [is_supported, name] : feature_report) { + if (is_supported) { continue; } LOG_ERROR(Render_Vulkan, "Missing required feature: {}", name); - is_suitable = false; - } - - if (!is_suitable) { - LOG_ERROR(Render_Vulkan, "{} is not suitable", properties.deviceName); + throw vk::Exception(VK_ERROR_FEATURE_NOT_PRESENT); } - - return is_suitable; } -std::vector<const char*> VKDevice::LoadExtensions() { +std::vector<const char*> Device::LoadExtensions() { std::vector<const char*> extensions; - const auto Test = [&](const VkExtensionProperties& extension, - std::optional<std::reference_wrapper<bool>> status, const char* name, - bool push) { - if (extension.extensionName != std::string_view(name)) { - return; - } - if (push) { - extensions.push_back(name); - } - if (status) { - status->get() = true; - } - }; - extensions.reserve(7 + REQUIRED_EXTENSIONS.size()); extensions.insert(extensions.begin(), REQUIRED_EXTENSIONS.begin(), REQUIRED_EXTENSIONS.end()); @@ -603,36 +627,47 @@ std::vector<const char*> VKDevice::LoadExtensions() { bool has_ext_transform_feedback{}; bool has_ext_custom_border_color{}; bool has_ext_extended_dynamic_state{}; - for (const auto& extension : physical.EnumerateDeviceExtensionProperties()) { - Test(extension, nv_viewport_swizzle, VK_NV_VIEWPORT_SWIZZLE_EXTENSION_NAME, true); - Test(extension, khr_uniform_buffer_standard_layout, + bool has_ext_robustness2{}; + for (const VkExtensionProperties& extension : physical.EnumerateDeviceExtensionProperties()) { + const auto test = [&](std::optional<std::reference_wrapper<bool>> status, const char* name, + bool push) { + if (extension.extensionName != std::string_view(name)) { + return; + } + if (push) { + extensions.push_back(name); + } + if (status) { + status->get() = true; + } + }; + test(nv_viewport_swizzle, VK_NV_VIEWPORT_SWIZZLE_EXTENSION_NAME, true); + test(khr_uniform_buffer_standard_layout, VK_KHR_UNIFORM_BUFFER_STANDARD_LAYOUT_EXTENSION_NAME, true); - Test(extension, has_khr_shader_float16_int8, VK_KHR_SHADER_FLOAT16_INT8_EXTENSION_NAME, - false); - Test(extension, ext_depth_range_unrestricted, - VK_EXT_DEPTH_RANGE_UNRESTRICTED_EXTENSION_NAME, true); - Test(extension, ext_index_type_uint8, VK_EXT_INDEX_TYPE_UINT8_EXTENSION_NAME, true); - Test(extension, ext_shader_viewport_index_layer, - VK_EXT_SHADER_VIEWPORT_INDEX_LAYER_EXTENSION_NAME, true); - Test(extension, has_ext_subgroup_size_control, VK_EXT_SUBGROUP_SIZE_CONTROL_EXTENSION_NAME, - false); - Test(extension, has_ext_transform_feedback, VK_EXT_TRANSFORM_FEEDBACK_EXTENSION_NAME, - false); - Test(extension, has_ext_custom_border_color, VK_EXT_CUSTOM_BORDER_COLOR_EXTENSION_NAME, - false); - Test(extension, has_ext_extended_dynamic_state, - VK_EXT_EXTENDED_DYNAMIC_STATE_EXTENSION_NAME, false); + test(has_khr_shader_float16_int8, VK_KHR_SHADER_FLOAT16_INT8_EXTENSION_NAME, false); + test(ext_depth_range_unrestricted, VK_EXT_DEPTH_RANGE_UNRESTRICTED_EXTENSION_NAME, true); + test(ext_index_type_uint8, VK_EXT_INDEX_TYPE_UINT8_EXTENSION_NAME, true); + test(ext_sampler_filter_minmax, VK_EXT_SAMPLER_FILTER_MINMAX_EXTENSION_NAME, true); + test(ext_shader_viewport_index_layer, VK_EXT_SHADER_VIEWPORT_INDEX_LAYER_EXTENSION_NAME, + true); + test(ext_tooling_info, VK_EXT_TOOLING_INFO_EXTENSION_NAME, true); + test(ext_shader_stencil_export, VK_EXT_SHADER_STENCIL_EXPORT_EXTENSION_NAME, true); + test(has_ext_transform_feedback, VK_EXT_TRANSFORM_FEEDBACK_EXTENSION_NAME, false); + test(has_ext_custom_border_color, VK_EXT_CUSTOM_BORDER_COLOR_EXTENSION_NAME, false); + test(has_ext_extended_dynamic_state, VK_EXT_EXTENDED_DYNAMIC_STATE_EXTENSION_NAME, false); + test(has_ext_robustness2, VK_EXT_ROBUSTNESS_2_EXTENSION_NAME, false); + test(has_ext_subgroup_size_control, VK_EXT_SUBGROUP_SIZE_CONTROL_EXTENSION_NAME, false); if (Settings::values.renderer_debug) { - Test(extension, nv_device_diagnostics_config, - VK_NV_DEVICE_DIAGNOSTICS_CONFIG_EXTENSION_NAME, true); + test(nv_device_diagnostics_config, VK_NV_DEVICE_DIAGNOSTICS_CONFIG_EXTENSION_NAME, + true); } } VkPhysicalDeviceFeatures2KHR features; features.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FEATURES_2_KHR; - VkPhysicalDeviceProperties2KHR properties; - properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR; + VkPhysicalDeviceProperties2KHR physical_properties; + physical_properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR; if (has_khr_shader_float16_int8) { VkPhysicalDeviceFloat16Int8FeaturesKHR float16_int8_features; @@ -657,8 +692,8 @@ std::vector<const char*> VKDevice::LoadExtensions() { subgroup_properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SUBGROUP_SIZE_CONTROL_PROPERTIES_EXT; subgroup_properties.pNext = nullptr; - properties.pNext = &subgroup_properties; - physical.GetProperties2KHR(properties); + physical_properties.pNext = &subgroup_properties; + physical.GetProperties2KHR(physical_properties); is_warp_potentially_bigger = subgroup_properties.maxSubgroupSize > GuestWarpSize; @@ -682,8 +717,8 @@ std::vector<const char*> VKDevice::LoadExtensions() { VkPhysicalDeviceTransformFeedbackPropertiesEXT tfb_properties; tfb_properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_TRANSFORM_FEEDBACK_PROPERTIES_EXT; tfb_properties.pNext = nullptr; - properties.pNext = &tfb_properties; - physical.GetProperties2KHR(properties); + physical_properties.pNext = &tfb_properties; + physical.GetProperties2KHR(physical_properties); if (tfb_features.transformFeedback && tfb_features.geometryStreams && tfb_properties.maxTransformFeedbackStreams >= 4 && @@ -720,51 +755,75 @@ std::vector<const char*> VKDevice::LoadExtensions() { } } + if (has_ext_robustness2) { + VkPhysicalDeviceRobustness2FeaturesEXT robustness2; + robustness2.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_ROBUSTNESS_2_FEATURES_EXT; + robustness2.pNext = nullptr; + features.pNext = &robustness2; + physical.GetFeatures2KHR(features); + if (robustness2.nullDescriptor && robustness2.robustImageAccess2) { + extensions.push_back(VK_EXT_ROBUSTNESS_2_EXTENSION_NAME); + ext_robustness2 = true; + } + } + return extensions; } -void VKDevice::SetupFamilies(VkSurfaceKHR surface) { - std::optional<u32> graphics_family_, present_family_; - +void Device::SetupFamilies(VkSurfaceKHR surface) { const std::vector queue_family_properties = physical.GetQueueFamilyProperties(); - for (u32 i = 0; i < static_cast<u32>(queue_family_properties.size()); ++i) { - if (graphics_family_ && present_family_) + std::optional<u32> graphics; + std::optional<u32> present; + for (u32 index = 0; index < static_cast<u32>(queue_family_properties.size()); ++index) { + if (graphics && (present || !surface)) { break; - - const auto& queue_family = queue_family_properties[i]; - if (queue_family.queueCount == 0) + } + const VkQueueFamilyProperties& queue_family = queue_family_properties[index]; + if (queue_family.queueCount == 0) { continue; - + } if (queue_family.queueFlags & VK_QUEUE_GRAPHICS_BIT) { - graphics_family_ = i; + graphics = index; } - if (physical.GetSurfaceSupportKHR(i, surface)) { - present_family_ = i; + if (surface && physical.GetSurfaceSupportKHR(index, surface)) { + present = index; } } - ASSERT(graphics_family_ && present_family_); - - graphics_family = *graphics_family_; - present_family = *present_family_; + if (!graphics) { + LOG_ERROR(Render_Vulkan, "Device lacks a graphics queue"); + throw vk::Exception(VK_ERROR_FEATURE_NOT_PRESENT); + } + if (surface && !present) { + LOG_ERROR(Render_Vulkan, "Device lacks a present queue"); + throw vk::Exception(VK_ERROR_FEATURE_NOT_PRESENT); + } + graphics_family = *graphics; + present_family = *present; } -void VKDevice::SetupFeatures() { +void Device::SetupFeatures() { const auto supported_features{physical.GetFeatures()}; is_formatless_image_load_supported = supported_features.shaderStorageImageReadWithoutFormat; + is_blit_depth_stencil_supported = TestDepthStencilBlits(); is_optimal_astc_supported = IsOptimalAstcSupported(supported_features); } -void VKDevice::CollectTelemetryParameters() { +void Device::CollectTelemetryParameters() { VkPhysicalDeviceDriverPropertiesKHR driver{ .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_DRIVER_PROPERTIES_KHR, .pNext = nullptr, + .driverID = {}, + .driverName = {}, + .driverInfo = {}, + .conformanceVersion = {}, }; - VkPhysicalDeviceProperties2KHR properties{ + VkPhysicalDeviceProperties2KHR device_properties{ .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR, .pNext = &driver, + .properties = {}, }; - physical.GetProperties2KHR(properties); + physical.GetProperties2KHR(device_properties); driver_id = driver.driverID; vendor_name = driver.driverName; @@ -776,7 +835,33 @@ void VKDevice::CollectTelemetryParameters() { } } -std::vector<VkDeviceQueueCreateInfo> VKDevice::GetDeviceQueueCreateInfos() const { +void Device::CollectToolingInfo() { + if (!ext_tooling_info) { + return; + } + const auto vkGetPhysicalDeviceToolPropertiesEXT = + reinterpret_cast<PFN_vkGetPhysicalDeviceToolPropertiesEXT>( + dld.vkGetInstanceProcAddr(instance, "vkGetPhysicalDeviceToolPropertiesEXT")); + if (!vkGetPhysicalDeviceToolPropertiesEXT) { + return; + } + u32 tool_count = 0; + if (vkGetPhysicalDeviceToolPropertiesEXT(physical, &tool_count, nullptr) != VK_SUCCESS) { + return; + } + std::vector<VkPhysicalDeviceToolPropertiesEXT> tools(tool_count); + if (vkGetPhysicalDeviceToolPropertiesEXT(physical, &tool_count, tools.data()) != VK_SUCCESS) { + return; + } + for (const VkPhysicalDeviceToolPropertiesEXT& tool : tools) { + const std::string_view name = tool.name; + LOG_INFO(Render_Vulkan, "{}", name); + has_renderdoc = has_renderdoc || name == "RenderDoc"; + has_nsight_graphics = has_nsight_graphics || name == "NVIDIA Nsight Graphics"; + } +} + +std::vector<VkDeviceQueueCreateInfo> Device::GetDeviceQueueCreateInfos() const { static constexpr float QUEUE_PRIORITY = 1.0f; std::unordered_set<u32> unique_queue_families{graphics_family, present_family}; diff --git a/src/video_core/renderer_vulkan/vk_device.h b/src/video_core/vulkan_common/vulkan_device.h index 26a233db1..a973c3ce4 100644 --- a/src/video_core/renderer_vulkan/vk_device.h +++ b/src/video_core/vulkan_common/vulkan_device.h @@ -10,11 +10,12 @@ #include <vector> #include "common/common_types.h" -#include "video_core/renderer_vulkan/nsight_aftermath_tracker.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan { +class NsightAftermathTracker; + /// Format usage descriptor. enum class FormatType { Linear, Optimal, Buffer }; @@ -22,14 +23,11 @@ enum class FormatType { Linear, Optimal, Buffer }; const u32 GuestWarpSize = 32; /// Handles data specific to a physical device. -class VKDevice final { +class Device final { public: - explicit VKDevice(VkInstance instance, vk::PhysicalDevice physical, VkSurfaceKHR surface, - const vk::InstanceDispatch& dld); - ~VKDevice(); - - /// Initializes the device. Returns true on success. - bool Create(); + explicit Device(VkInstance instance, vk::PhysicalDevice physical, VkSurfaceKHR surface, + const vk::InstanceDispatch& dld); + ~Device(); /** * Returns a format supported by the device for the passed requeriments. @@ -83,7 +81,7 @@ public: } /// Returns the current Vulkan API version provided in Vulkan-formatted version numbers. - u32 GetApiVersion() const { + u32 ApiVersion() const { return properties.apiVersion; } @@ -152,6 +150,11 @@ public: return is_formatless_image_load_supported; } + /// Returns true when blitting from and to depth stencil images is supported. + bool IsBlitDepthStencilSupported() const { + return is_blit_depth_stencil_supported; + } + /// Returns true if the device supports VK_NV_viewport_swizzle. bool IsNvViewportSwizzleSupported() const { return nv_viewport_swizzle; @@ -167,6 +170,11 @@ public: return ext_index_type_uint8; } + /// Returns true if the device supports VK_EXT_sampler_filter_minmax. + bool IsExtSamplerFilterMinmaxSupported() const { + return ext_sampler_filter_minmax; + } + /// Returns true if the device supports VK_EXT_depth_range_unrestricted. bool IsExtDepthRangeUnrestrictedSupported() const { return ext_depth_range_unrestricted; @@ -192,6 +200,16 @@ public: return ext_extended_dynamic_state; } + /// Returns true if the device supports VK_EXT_shader_stencil_export. + bool IsExtShaderStencilExportSupported() const { + return ext_shader_stencil_export; + } + + /// Returns true when a known debugging tool is attached. + bool HasDebuggingToolAttached() const { + return has_renderdoc || has_nsight_graphics; + } + /// Returns the vendor name reported from Vulkan. std::string_view GetVendorName() const { return vendor_name; @@ -207,10 +225,10 @@ public: return use_asynchronous_shaders; } +private: /// Checks if the physical device is suitable. - static bool IsSuitable(vk::PhysicalDevice physical, VkSurfaceKHR surface); + void CheckSuitability() const; -private: /// Loads extensions into a vector and stores available ones in this object. std::vector<const char*> LoadExtensions(); @@ -223,22 +241,30 @@ private: /// Collects telemetry information from the device. void CollectTelemetryParameters(); + /// Collects information about attached tools. + void CollectToolingInfo(); + /// Returns a list of queue initialization descriptors. std::vector<VkDeviceQueueCreateInfo> GetDeviceQueueCreateInfos() const; /// Returns true if ASTC textures are natively supported. bool IsOptimalAstcSupported(const VkPhysicalDeviceFeatures& features) const; + /// Returns true if the device natively supports blitting depth stencil images. + bool TestDepthStencilBlits() const; + /// Returns true if a format is supported. bool IsFormatSupported(VkFormat wanted_format, VkFormatFeatureFlags wanted_usage, FormatType format_type) const; + VkInstance instance; ///< Vulkan instance. vk::DeviceDispatch dld; ///< Device function pointers. vk::PhysicalDevice physical; ///< Physical device. VkPhysicalDeviceProperties properties; ///< Device properties. vk::Device logical; ///< Logical device. vk::Queue graphics_queue; ///< Main graphics queue. vk::Queue present_queue; ///< Main present queue. + u32 instance_version{}; ///< Vulkan onstance version. u32 graphics_family{}; ///< Main graphics queue family index. u32 present_family{}; ///< Main present queue family index. VkDriverIdKHR driver_id{}; ///< Driver ID. @@ -247,15 +273,22 @@ private: bool is_float16_supported{}; ///< Support for float16 arithmetics. bool is_warp_potentially_bigger{}; ///< Host warp size can be bigger than guest. bool is_formatless_image_load_supported{}; ///< Support for shader image read without format. + bool is_blit_depth_stencil_supported{}; ///< Support for blitting from and to depth stencil. bool nv_viewport_swizzle{}; ///< Support for VK_NV_viewport_swizzle. bool khr_uniform_buffer_standard_layout{}; ///< Support for std430 on UBOs. bool ext_index_type_uint8{}; ///< Support for VK_EXT_index_type_uint8. + bool ext_sampler_filter_minmax{}; ///< Support for VK_EXT_sampler_filter_minmax. bool ext_depth_range_unrestricted{}; ///< Support for VK_EXT_depth_range_unrestricted. bool ext_shader_viewport_index_layer{}; ///< Support for VK_EXT_shader_viewport_index_layer. + bool ext_tooling_info{}; ///< Support for VK_EXT_tooling_info. bool ext_transform_feedback{}; ///< Support for VK_EXT_transform_feedback. bool ext_custom_border_color{}; ///< Support for VK_EXT_custom_border_color. bool ext_extended_dynamic_state{}; ///< Support for VK_EXT_extended_dynamic_state. + bool ext_robustness2{}; ///< Support for VK_EXT_robustness2. + bool ext_shader_stencil_export{}; ///< Support for VK_EXT_shader_stencil_export. bool nv_device_diagnostics_config{}; ///< Support for VK_NV_device_diagnostics_config. + bool has_renderdoc{}; ///< Has RenderDoc attached + bool has_nsight_graphics{}; ///< Has Nsight Graphics attached // Asynchronous Graphics Pipeline setting bool use_asynchronous_shaders{}; ///< Setting to use asynchronous shaders/graphics pipeline @@ -268,7 +301,7 @@ private: std::unordered_map<VkFormat, VkFormatProperties> format_properties; /// Nsight Aftermath GPU crash tracker - NsightAftermathTracker nsight_aftermath_tracker; + std::unique_ptr<NsightAftermathTracker> nsight_aftermath_tracker; }; } // namespace Vulkan diff --git a/src/video_core/vulkan_common/vulkan_instance.cpp b/src/video_core/vulkan_common/vulkan_instance.cpp new file mode 100644 index 000000000..889ecda0c --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_instance.cpp @@ -0,0 +1,151 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <optional> +#include <span> +#include <utility> +#include <vector> + +#include "common/common_types.h" +#include "common/dynamic_library.h" +#include "common/logging/log.h" +#include "core/frontend/emu_window.h" +#include "video_core/vulkan_common/vulkan_instance.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" + +// Include these late to avoid polluting previous headers +#ifdef _WIN32 +#include <windows.h> +// ensure include order +#include <vulkan/vulkan_win32.h> +#endif + +#if !defined(_WIN32) && !defined(__APPLE__) +#include <X11/Xlib.h> +#include <vulkan/vulkan_wayland.h> +#include <vulkan/vulkan_xlib.h> +#endif + +namespace Vulkan { +namespace { +[[nodiscard]] std::vector<const char*> RequiredExtensions( + Core::Frontend::WindowSystemType window_type, bool enable_debug_utils) { + std::vector<const char*> extensions; + extensions.reserve(6); + switch (window_type) { + case Core::Frontend::WindowSystemType::Headless: + break; +#ifdef _WIN32 + case Core::Frontend::WindowSystemType::Windows: + extensions.push_back(VK_KHR_WIN32_SURFACE_EXTENSION_NAME); + break; +#endif +#if !defined(_WIN32) && !defined(__APPLE__) + case Core::Frontend::WindowSystemType::X11: + extensions.push_back(VK_KHR_XLIB_SURFACE_EXTENSION_NAME); + break; + case Core::Frontend::WindowSystemType::Wayland: + extensions.push_back(VK_KHR_WAYLAND_SURFACE_EXTENSION_NAME); + break; +#endif + default: + LOG_ERROR(Render_Vulkan, "Presentation not supported on this platform"); + break; + } + if (window_type != Core::Frontend::WindowSystemType::Headless) { + extensions.push_back(VK_KHR_SURFACE_EXTENSION_NAME); + } + if (enable_debug_utils) { + extensions.push_back(VK_EXT_DEBUG_UTILS_EXTENSION_NAME); + } + extensions.push_back(VK_KHR_GET_PHYSICAL_DEVICE_PROPERTIES_2_EXTENSION_NAME); + return extensions; +} + +[[nodiscard]] bool AreExtensionsSupported(const vk::InstanceDispatch& dld, + std::span<const char* const> extensions) { + const std::optional properties = vk::EnumerateInstanceExtensionProperties(dld); + if (!properties) { + LOG_ERROR(Render_Vulkan, "Failed to query extension properties"); + return false; + } + for (const char* extension : extensions) { + const auto it = std::ranges::find_if(*properties, [extension](const auto& prop) { + return std::strcmp(extension, prop.extensionName) == 0; + }); + if (it == properties->end()) { + LOG_ERROR(Render_Vulkan, "Required instance extension {} is not available", extension); + return false; + } + } + return true; +} + +[[nodiscard]] std::vector<const char*> Layers(bool enable_layers) { + std::vector<const char*> layers; + if (enable_layers) { + layers.push_back("VK_LAYER_KHRONOS_validation"); + } + return layers; +} + +void RemoveUnavailableLayers(const vk::InstanceDispatch& dld, std::vector<const char*>& layers) { + const std::optional layer_properties = vk::EnumerateInstanceLayerProperties(dld); + if (!layer_properties) { + LOG_ERROR(Render_Vulkan, "Failed to query layer properties, disabling layers"); + layers.clear(); + } + std::erase_if(layers, [&layer_properties](const char* layer) { + const auto comp = [layer](const VkLayerProperties& layer_property) { + return std::strcmp(layer, layer_property.layerName) == 0; + }; + const auto it = std::ranges::find_if(*layer_properties, comp); + if (it == layer_properties->end()) { + LOG_ERROR(Render_Vulkan, "Layer {} not available, removing it", layer); + return true; + } + return false; + }); +} +} // Anonymous namespace + +vk::Instance CreateInstance(const Common::DynamicLibrary& library, vk::InstanceDispatch& dld, + u32 required_version, Core::Frontend::WindowSystemType window_type, + bool enable_debug_utils, bool enable_layers) { + if (!library.IsOpen()) { + LOG_ERROR(Render_Vulkan, "Vulkan library not available"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + if (!library.GetSymbol("vkGetInstanceProcAddr", &dld.vkGetInstanceProcAddr)) { + LOG_ERROR(Render_Vulkan, "vkGetInstanceProcAddr not present in Vulkan"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + if (!vk::Load(dld)) { + LOG_ERROR(Render_Vulkan, "Failed to load Vulkan function pointers"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + const std::vector<const char*> extensions = RequiredExtensions(window_type, enable_debug_utils); + if (!AreExtensionsSupported(dld, extensions)) { + throw vk::Exception(VK_ERROR_EXTENSION_NOT_PRESENT); + } + std::vector<const char*> layers = Layers(enable_layers); + RemoveUnavailableLayers(dld, layers); + + const u32 available_version = vk::AvailableVersion(dld); + if (available_version < required_version) { + LOG_ERROR(Render_Vulkan, "Vulkan {}.{} is not supported, {}.{} is required", + VK_VERSION_MAJOR(available_version), VK_VERSION_MINOR(available_version), + VK_VERSION_MAJOR(required_version), VK_VERSION_MINOR(required_version)); + throw vk::Exception(VK_ERROR_INCOMPATIBLE_DRIVER); + } + vk::Instance instance = vk::Instance::Create(required_version, layers, extensions, dld); + if (!vk::Load(*instance, dld)) { + LOG_ERROR(Render_Vulkan, "Failed to load Vulkan instance function pointers"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + return instance; +} + +} // namespace Vulkan diff --git a/src/video_core/vulkan_common/vulkan_instance.h b/src/video_core/vulkan_common/vulkan_instance.h new file mode 100644 index 000000000..e5e3a7144 --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_instance.h @@ -0,0 +1,32 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/common_types.h" +#include "common/dynamic_library.h" +#include "core/frontend/emu_window.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" + +namespace Vulkan { + +/** + * Create a Vulkan instance + * + * @param library Dynamic library to load the Vulkan instance from + * @param dld Dispatch table to load function pointers into + * @param required_version Required Vulkan version (for example, VK_API_VERSION_1_1) + * @param window_type Window system type's enabled extension + * @param enable_debug_utils Whether to enable VK_EXT_debug_utils_extension_name or not + * @param enable_layers Whether to enable Vulkan validation layers or not + * + * @return A new Vulkan instance + * @throw vk::Exception on failure + */ +[[nodiscard]] vk::Instance CreateInstance( + const Common::DynamicLibrary& library, vk::InstanceDispatch& dld, u32 required_version, + Core::Frontend::WindowSystemType window_type = Core::Frontend::WindowSystemType::Headless, + bool enable_debug_utils = false, bool enable_layers = false); + +} // namespace Vulkan diff --git a/src/video_core/vulkan_common/vulkan_library.cpp b/src/video_core/vulkan_common/vulkan_library.cpp new file mode 100644 index 000000000..557871d81 --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_library.cpp @@ -0,0 +1,36 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <cstdlib> +#include <string> + +#include "common/dynamic_library.h" +#include "common/file_util.h" +#include "video_core/vulkan_common/vulkan_library.h" + +namespace Vulkan { + +Common::DynamicLibrary OpenLibrary() { + Common::DynamicLibrary library; +#ifdef __APPLE__ + // Check if a path to a specific Vulkan library has been specified. + char* const libvulkan_env = std::getenv("LIBVULKAN_PATH"); + if (!libvulkan_env || !library.Open(libvulkan_env)) { + // Use the libvulkan.dylib from the application bundle. + const std::string filename = + Common::FS::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib"; + void(library.Open(filename.c_str())); + } +#else + std::string filename = Common::DynamicLibrary::GetVersionedFilename("vulkan", 1); + if (!library.Open(filename.c_str())) { + // Android devices may not have libvulkan.so.1, only libvulkan.so. + filename = Common::DynamicLibrary::GetVersionedFilename("vulkan"); + void(library.Open(filename.c_str())); + } +#endif + return library; +} + +} // namespace Vulkan diff --git a/src/video_core/vulkan_common/vulkan_library.h b/src/video_core/vulkan_common/vulkan_library.h new file mode 100644 index 000000000..8b28b0e17 --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_library.h @@ -0,0 +1,13 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "common/dynamic_library.h" + +namespace Vulkan { + +Common::DynamicLibrary OpenLibrary(); + +} // namespace Vulkan diff --git a/src/video_core/vulkan_common/vulkan_surface.cpp b/src/video_core/vulkan_common/vulkan_surface.cpp new file mode 100644 index 000000000..3c3238f96 --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_surface.cpp @@ -0,0 +1,81 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/logging/log.h" +#include "core/frontend/emu_window.h" +#include "video_core/vulkan_common/vulkan_surface.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" + +// Include these late to avoid polluting previous headers +#ifdef _WIN32 +#include <windows.h> +// ensure include order +#include <vulkan/vulkan_win32.h> +#endif + +#if !defined(_WIN32) && !defined(__APPLE__) +#include <X11/Xlib.h> +#include <vulkan/vulkan_wayland.h> +#include <vulkan/vulkan_xlib.h> +#endif + +namespace Vulkan { + +vk::SurfaceKHR CreateSurface(const vk::Instance& instance, + const Core::Frontend::EmuWindow& emu_window) { + [[maybe_unused]] const vk::InstanceDispatch& dld = instance.Dispatch(); + [[maybe_unused]] const auto& window_info = emu_window.GetWindowInfo(); + VkSurfaceKHR unsafe_surface = nullptr; + +#ifdef _WIN32 + if (window_info.type == Core::Frontend::WindowSystemType::Windows) { + const HWND hWnd = static_cast<HWND>(window_info.render_surface); + const VkWin32SurfaceCreateInfoKHR win32_ci{VK_STRUCTURE_TYPE_WIN32_SURFACE_CREATE_INFO_KHR, + nullptr, 0, nullptr, hWnd}; + const auto vkCreateWin32SurfaceKHR = reinterpret_cast<PFN_vkCreateWin32SurfaceKHR>( + dld.vkGetInstanceProcAddr(*instance, "vkCreateWin32SurfaceKHR")); + if (!vkCreateWin32SurfaceKHR || + vkCreateWin32SurfaceKHR(*instance, &win32_ci, nullptr, &unsafe_surface) != VK_SUCCESS) { + LOG_ERROR(Render_Vulkan, "Failed to initialize Win32 surface"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + } +#endif +#if !defined(_WIN32) && !defined(__APPLE__) + if (window_info.type == Core::Frontend::WindowSystemType::X11) { + const VkXlibSurfaceCreateInfoKHR xlib_ci{ + VK_STRUCTURE_TYPE_XLIB_SURFACE_CREATE_INFO_KHR, nullptr, 0, + static_cast<Display*>(window_info.display_connection), + reinterpret_cast<Window>(window_info.render_surface)}; + const auto vkCreateXlibSurfaceKHR = reinterpret_cast<PFN_vkCreateXlibSurfaceKHR>( + dld.vkGetInstanceProcAddr(*instance, "vkCreateXlibSurfaceKHR")); + if (!vkCreateXlibSurfaceKHR || + vkCreateXlibSurfaceKHR(*instance, &xlib_ci, nullptr, &unsafe_surface) != VK_SUCCESS) { + LOG_ERROR(Render_Vulkan, "Failed to initialize Xlib surface"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + } + if (window_info.type == Core::Frontend::WindowSystemType::Wayland) { + const VkWaylandSurfaceCreateInfoKHR wayland_ci{ + VK_STRUCTURE_TYPE_WAYLAND_SURFACE_CREATE_INFO_KHR, nullptr, 0, + static_cast<wl_display*>(window_info.display_connection), + static_cast<wl_surface*>(window_info.render_surface)}; + const auto vkCreateWaylandSurfaceKHR = reinterpret_cast<PFN_vkCreateWaylandSurfaceKHR>( + dld.vkGetInstanceProcAddr(*instance, "vkCreateWaylandSurfaceKHR")); + if (!vkCreateWaylandSurfaceKHR || + vkCreateWaylandSurfaceKHR(*instance, &wayland_ci, nullptr, &unsafe_surface) != + VK_SUCCESS) { + LOG_ERROR(Render_Vulkan, "Failed to initialize Wayland surface"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + } +#endif + if (!unsafe_surface) { + LOG_ERROR(Render_Vulkan, "Presentation not supported on this platform"); + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); + } + return vk::SurfaceKHR(unsafe_surface, *instance, dld); +} + +} // namespace Vulkan diff --git a/src/video_core/vulkan_common/vulkan_surface.h b/src/video_core/vulkan_common/vulkan_surface.h new file mode 100644 index 000000000..05a169e32 --- /dev/null +++ b/src/video_core/vulkan_common/vulkan_surface.h @@ -0,0 +1,18 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include "video_core/vulkan_common/vulkan_wrapper.h" + +namespace Core::Frontend { +class EmuWindow; +} + +namespace Vulkan { + +[[nodiscard]] vk::SurfaceKHR CreateSurface(const vk::Instance& instance, + const Core::Frontend::EmuWindow& emu_window); + +} // namespace Vulkan diff --git a/src/video_core/renderer_vulkan/wrapper.cpp b/src/video_core/vulkan_common/vulkan_wrapper.cpp index 1fb14e190..5e15ad607 100644 --- a/src/video_core/renderer_vulkan/wrapper.cpp +++ b/src/video_core/vulkan_common/vulkan_wrapper.cpp @@ -6,32 +6,55 @@ #include <exception> #include <memory> #include <optional> +#include <string_view> #include <utility> #include <vector> #include "common/common_types.h" +#include "common/logging/log.h" -#include "video_core/renderer_vulkan/wrapper.h" +#include "video_core/vulkan_common/vulkan_wrapper.h" namespace Vulkan::vk { namespace { +template <typename Func> +void SortPhysicalDevices(std::vector<VkPhysicalDevice>& devices, const InstanceDispatch& dld, + Func&& func) { + // Calling GetProperties calls Vulkan more than needed. But they are supposed to be cheap + // functions. + std::stable_sort(devices.begin(), devices.end(), + [&dld, &func](VkPhysicalDevice lhs, VkPhysicalDevice rhs) { + return func(vk::PhysicalDevice(lhs, dld).GetProperties(), + vk::PhysicalDevice(rhs, dld).GetProperties()); + }); +} + +void SortPhysicalDevicesPerVendor(std::vector<VkPhysicalDevice>& devices, + const InstanceDispatch& dld, + std::initializer_list<u32> vendor_ids) { + for (auto it = vendor_ids.end(); it != vendor_ids.begin();) { + --it; + SortPhysicalDevices(devices, dld, [id = *it](const auto& lhs, const auto& rhs) { + return lhs.vendorID == id && rhs.vendorID != id; + }); + } +} + void SortPhysicalDevices(std::vector<VkPhysicalDevice>& devices, const InstanceDispatch& dld) { - std::stable_sort(devices.begin(), devices.end(), [&](auto lhs, auto rhs) { - // This will call Vulkan more than needed, but these calls are cheap. - const auto lhs_properties = vk::PhysicalDevice(lhs, dld).GetProperties(); - const auto rhs_properties = vk::PhysicalDevice(rhs, dld).GetProperties(); - - // Prefer discrete GPUs, Nvidia over AMD, AMD over Intel, Intel over the rest. - const bool preferred = - (lhs_properties.deviceType == VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU && - rhs_properties.deviceType != VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU) || - (lhs_properties.vendorID == 0x10DE && rhs_properties.vendorID != 0x10DE) || - (lhs_properties.vendorID == 0x1002 && rhs_properties.vendorID != 0x1002) || - (lhs_properties.vendorID == 0x8086 && rhs_properties.vendorID != 0x8086); - return !preferred; + // Sort by name, this will set a base and make GPUs with higher numbers appear first + // (e.g. GTX 1650 will intentionally be listed before a GTX 1080). + SortPhysicalDevices(devices, dld, [](const auto& lhs, const auto& rhs) { + return std::string_view{lhs.deviceName} > std::string_view{rhs.deviceName}; + }); + // Prefer discrete over non-discrete + SortPhysicalDevices(devices, dld, [](const auto& lhs, const auto& rhs) { + return lhs.deviceType == VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU && + rhs.deviceType != VK_PHYSICAL_DEVICE_TYPE_DISCRETE_GPU; }); + // Prefer Nvidia over AMD, AMD over Intel, Intel over the rest. + SortPhysicalDevicesPerVendor(devices, dld, {0x10DE, 0x1002, 0x8086}); } template <typename T> @@ -58,6 +81,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkCmdBeginQuery); X(vkCmdBeginRenderPass); X(vkCmdBeginTransformFeedbackEXT); + X(vkCmdBeginDebugUtilsLabelEXT); X(vkCmdBindDescriptorSets); X(vkCmdBindIndexBuffer); X(vkCmdBindPipeline); @@ -75,6 +99,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkCmdEndQuery); X(vkCmdEndRenderPass); X(vkCmdEndTransformFeedbackEXT); + X(vkCmdEndDebugUtilsLabelEXT); X(vkCmdFillBuffer); X(vkCmdPipelineBarrier); X(vkCmdPushConstants); @@ -98,6 +123,7 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkCmdSetPrimitiveTopologyEXT); X(vkCmdSetStencilOpEXT); X(vkCmdSetStencilTestEnableEXT); + X(vkCmdResolveImage); X(vkCreateBuffer); X(vkCreateBufferView); X(vkCreateCommandPool); @@ -153,6 +179,8 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { X(vkQueueSubmit); X(vkResetFences); X(vkResetQueryPoolEXT); + X(vkSetDebugUtilsObjectNameEXT); + X(vkSetDebugUtilsObjectTagEXT); X(vkUnmapMemory); X(vkUpdateDescriptorSetWithTemplateKHR); X(vkUpdateDescriptorSets); @@ -161,6 +189,19 @@ void Load(VkDevice device, DeviceDispatch& dld) noexcept { #undef X } +template <typename T> +void SetObjectName(const DeviceDispatch* dld, VkDevice device, T handle, VkObjectType type, + const char* name) { + const VkDebugUtilsObjectNameInfoEXT name_info{ + .sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_OBJECT_NAME_INFO_EXT, + .pNext = nullptr, + .objectType = VK_OBJECT_TYPE_IMAGE, + .objectHandle = reinterpret_cast<u64>(handle), + .pObjectName = name, + }; + Check(dld->vkSetDebugUtilsObjectNameEXT(device, &name_info)); +} + } // Anonymous namespace bool Load(InstanceDispatch& dld) noexcept { @@ -393,18 +434,17 @@ VkResult Free(VkDevice device, VkCommandPool handle, Span<VkCommandBuffer> buffe return VK_SUCCESS; } -Instance Instance::Create(Span<const char*> layers, Span<const char*> extensions, - InstanceDispatch& dld) noexcept { - static constexpr VkApplicationInfo application_info{ +Instance Instance::Create(u32 version, Span<const char*> layers, Span<const char*> extensions, + InstanceDispatch& dispatch) { + const VkApplicationInfo application_info{ .sType = VK_STRUCTURE_TYPE_APPLICATION_INFO, .pNext = nullptr, .pApplicationName = "yuzu Emulator", .applicationVersion = VK_MAKE_VERSION(0, 1, 0), .pEngineName = "yuzu Emulator", .engineVersion = VK_MAKE_VERSION(0, 1, 0), - .apiVersion = VK_API_VERSION_1_1, + .apiVersion = version, }; - const VkInstanceCreateInfo ci{ .sType = VK_STRUCTURE_TYPE_INSTANCE_CREATE_INFO, .pNext = nullptr, @@ -415,66 +455,68 @@ Instance Instance::Create(Span<const char*> layers, Span<const char*> extensions .enabledExtensionCount = extensions.size(), .ppEnabledExtensionNames = extensions.data(), }; - VkInstance instance; - if (dld.vkCreateInstance(&ci, nullptr, &instance) != VK_SUCCESS) { - // Failed to create the instance. - return {}; - } - if (!Proc(dld.vkDestroyInstance, dld, "vkDestroyInstance", instance)) { + Check(dispatch.vkCreateInstance(&ci, nullptr, &instance)); + if (!Proc(dispatch.vkDestroyInstance, dispatch, "vkDestroyInstance", instance)) { // We successfully created an instance but the destroy function couldn't be loaded. // This is a good moment to panic. - return {}; + throw vk::Exception(VK_ERROR_INITIALIZATION_FAILED); } - - return Instance(instance, dld); + return Instance(instance, dispatch); } -std::optional<std::vector<VkPhysicalDevice>> Instance::EnumeratePhysicalDevices() { +std::vector<VkPhysicalDevice> Instance::EnumeratePhysicalDevices() const { u32 num; - if (dld->vkEnumeratePhysicalDevices(handle, &num, nullptr) != VK_SUCCESS) { - return std::nullopt; - } + Check(dld->vkEnumeratePhysicalDevices(handle, &num, nullptr)); std::vector<VkPhysicalDevice> physical_devices(num); - if (dld->vkEnumeratePhysicalDevices(handle, &num, physical_devices.data()) != VK_SUCCESS) { - return std::nullopt; - } + Check(dld->vkEnumeratePhysicalDevices(handle, &num, physical_devices.data())); SortPhysicalDevices(physical_devices, *dld); - return std::make_optional(std::move(physical_devices)); + return physical_devices; } -DebugCallback Instance::TryCreateDebugCallback( - PFN_vkDebugUtilsMessengerCallbackEXT callback) noexcept { - const VkDebugUtilsMessengerCreateInfoEXT ci{ - .sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_MESSENGER_CREATE_INFO_EXT, - .pNext = nullptr, - .flags = 0, - .messageSeverity = VK_DEBUG_UTILS_MESSAGE_SEVERITY_ERROR_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_SEVERITY_WARNING_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_SEVERITY_INFO_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_SEVERITY_VERBOSE_BIT_EXT, - .messageType = VK_DEBUG_UTILS_MESSAGE_TYPE_GENERAL_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_TYPE_VALIDATION_BIT_EXT | - VK_DEBUG_UTILS_MESSAGE_TYPE_PERFORMANCE_BIT_EXT, - .pfnUserCallback = callback, - .pUserData = nullptr, - }; - - VkDebugUtilsMessengerEXT messenger; - if (dld->vkCreateDebugUtilsMessengerEXT(handle, &ci, nullptr, &messenger) != VK_SUCCESS) { - return {}; - } - return DebugCallback(messenger, handle, *dld); +DebugUtilsMessenger Instance::CreateDebugUtilsMessenger( + const VkDebugUtilsMessengerCreateInfoEXT& create_info) const { + VkDebugUtilsMessengerEXT object; + Check(dld->vkCreateDebugUtilsMessengerEXT(handle, &create_info, nullptr, &object)); + return DebugUtilsMessenger(object, handle, *dld); } void Buffer::BindMemory(VkDeviceMemory memory, VkDeviceSize offset) const { Check(dld->vkBindBufferMemory(owner, handle, memory, offset)); } +void Buffer::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_BUFFER, name); +} + +void BufferView::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_BUFFER_VIEW, name); +} + void Image::BindMemory(VkDeviceMemory memory, VkDeviceSize offset) const { Check(dld->vkBindImageMemory(owner, handle, memory, offset)); } +void Image::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_IMAGE, name); +} + +void ImageView::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_IMAGE_VIEW, name); +} + +void DeviceMemory::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_DEVICE_MEMORY, name); +} + +void Fence::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_FENCE, name); +} + +void Framebuffer::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_FRAMEBUFFER, name); +} + DescriptorSets DescriptorPool::Allocate(const VkDescriptorSetAllocateInfo& ai) const { const std::size_t num = ai.descriptorSetCount; std::unique_ptr sets = std::make_unique<VkDescriptorSet[]>(num); @@ -488,6 +530,10 @@ DescriptorSets DescriptorPool::Allocate(const VkDescriptorSetAllocateInfo& ai) c } } +void DescriptorPool::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_DESCRIPTOR_POOL, name); +} + CommandBuffers CommandPool::Allocate(std::size_t num_buffers, VkCommandBufferLevel level) const { const VkCommandBufferAllocateInfo ai{ .sType = VK_STRUCTURE_TYPE_COMMAND_BUFFER_ALLOCATE_INFO, @@ -508,6 +554,10 @@ CommandBuffers CommandPool::Allocate(std::size_t num_buffers, VkCommandBufferLev } } +void CommandPool::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_COMMAND_POOL, name); +} + std::vector<VkImage> SwapchainKHR::GetImages() const { u32 num; Check(dld->vkGetSwapchainImagesKHR(owner, handle, &num, nullptr)); @@ -516,9 +566,21 @@ std::vector<VkImage> SwapchainKHR::GetImages() const { return images; } +void Event::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_EVENT, name); +} + +void ShaderModule::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_SHADER_MODULE, name); +} + +void Semaphore::SetObjectNameEXT(const char* name) const { + SetObjectName(dld, owner, handle, VK_OBJECT_TYPE_SEMAPHORE, name); +} + Device Device::Create(VkPhysicalDevice physical_device, Span<VkDeviceQueueCreateInfo> queues_ci, Span<const char*> enabled_extensions, const void* next, - DeviceDispatch& dld) noexcept { + DeviceDispatch& dispatch) { const VkDeviceCreateInfo ci{ .sType = VK_STRUCTURE_TYPE_DEVICE_CREATE_INFO, .pNext = next, @@ -531,13 +593,10 @@ Device Device::Create(VkPhysicalDevice physical_device, Span<VkDeviceQueueCreate .ppEnabledExtensionNames = enabled_extensions.data(), .pEnabledFeatures = nullptr, }; - VkDevice device; - if (dld.vkCreateDevice(physical_device, &ci, nullptr, &device) != VK_SUCCESS) { - return {}; - } - Load(device, dld); - return Device(device, dld); + Check(dispatch.vkCreateDevice(physical_device, &ci, nullptr, &device)); + Load(device, dispatch); + return Device(device, dispatch); } Queue Device::GetQueue(u32 family_index) const noexcept { @@ -796,6 +855,21 @@ VkPhysicalDeviceMemoryProperties PhysicalDevice::GetMemoryProperties() const noe return properties; } +u32 AvailableVersion(const InstanceDispatch& dld) noexcept { + PFN_vkEnumerateInstanceVersion vkEnumerateInstanceVersion; + if (!Proc(vkEnumerateInstanceVersion, dld, "vkEnumerateInstanceVersion")) { + // If the procedure is not found, Vulkan 1.0 is assumed + return VK_API_VERSION_1_0; + } + u32 version; + if (const VkResult result = vkEnumerateInstanceVersion(&version); result != VK_SUCCESS) { + LOG_ERROR(Render_Vulkan, "vkEnumerateInstanceVersion returned {}, assuming Vulkan 1.1", + ToString(result)); + return VK_API_VERSION_1_1; + } + return version; +} + std::optional<std::vector<VkExtensionProperties>> EnumerateInstanceExtensionProperties( const InstanceDispatch& dld) { u32 num; @@ -807,7 +881,7 @@ std::optional<std::vector<VkExtensionProperties>> EnumerateInstanceExtensionProp VK_SUCCESS) { return std::nullopt; } - return std::move(properties); + return properties; } std::optional<std::vector<VkLayerProperties>> EnumerateInstanceLayerProperties( diff --git a/src/video_core/renderer_vulkan/wrapper.h b/src/video_core/vulkan_common/vulkan_wrapper.h index 234e01693..912cab46c 100644 --- a/src/video_core/renderer_vulkan/wrapper.h +++ b/src/video_core/vulkan_common/vulkan_wrapper.h @@ -9,6 +9,7 @@ #include <limits> #include <memory> #include <optional> +#include <span> #include <type_traits> #include <utility> #include <vector> @@ -18,6 +19,10 @@ #include "common/common_types.h" +#ifdef _MSC_VER +#pragma warning(disable : 26812) // Disable prefer enum class over enum +#endif + namespace Vulkan::vk { /** @@ -41,6 +46,9 @@ public: /// Construct an empty span. constexpr Span() noexcept = default; + /// Construct an empty span + constexpr Span(std::nullptr_t) noexcept {} + /// Construct a span from a single element. constexpr Span(const T& value) noexcept : ptr{&value}, num{1} {} @@ -52,7 +60,7 @@ public: /// Construct a span from a pointer and a size. /// This is inteded for subranges. - constexpr Span(const T* ptr, std::size_t num) noexcept : ptr{ptr}, num{num} {} + constexpr Span(const T* ptr_, std::size_t num_) noexcept : ptr{ptr_}, num{num_} {} /// Returns the data pointer by the span. constexpr const T* data() const noexcept { @@ -177,6 +185,7 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkCmdBeginQuery vkCmdBeginQuery; PFN_vkCmdBeginRenderPass vkCmdBeginRenderPass; PFN_vkCmdBeginTransformFeedbackEXT vkCmdBeginTransformFeedbackEXT; + PFN_vkCmdBeginDebugUtilsLabelEXT vkCmdBeginDebugUtilsLabelEXT; PFN_vkCmdBindDescriptorSets vkCmdBindDescriptorSets; PFN_vkCmdBindIndexBuffer vkCmdBindIndexBuffer; PFN_vkCmdBindPipeline vkCmdBindPipeline; @@ -194,6 +203,7 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkCmdEndQuery vkCmdEndQuery; PFN_vkCmdEndRenderPass vkCmdEndRenderPass; PFN_vkCmdEndTransformFeedbackEXT vkCmdEndTransformFeedbackEXT; + PFN_vkCmdEndDebugUtilsLabelEXT vkCmdEndDebugUtilsLabelEXT; PFN_vkCmdFillBuffer vkCmdFillBuffer; PFN_vkCmdPipelineBarrier vkCmdPipelineBarrier; PFN_vkCmdPushConstants vkCmdPushConstants; @@ -217,6 +227,7 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkCmdSetPrimitiveTopologyEXT vkCmdSetPrimitiveTopologyEXT; PFN_vkCmdSetStencilOpEXT vkCmdSetStencilOpEXT; PFN_vkCmdSetStencilTestEnableEXT vkCmdSetStencilTestEnableEXT; + PFN_vkCmdResolveImage vkCmdResolveImage; PFN_vkCreateBuffer vkCreateBuffer; PFN_vkCreateBufferView vkCreateBufferView; PFN_vkCreateCommandPool vkCreateCommandPool; @@ -272,6 +283,8 @@ struct DeviceDispatch : public InstanceDispatch { PFN_vkQueueSubmit vkQueueSubmit; PFN_vkResetFences vkResetFences; PFN_vkResetQueryPoolEXT vkResetQueryPoolEXT; + PFN_vkSetDebugUtilsObjectNameEXT vkSetDebugUtilsObjectNameEXT; + PFN_vkSetDebugUtilsObjectTagEXT vkSetDebugUtilsObjectTagEXT; PFN_vkUnmapMemory vkUnmapMemory; PFN_vkUpdateDescriptorSetWithTemplateKHR vkUpdateDescriptorSetWithTemplateKHR; PFN_vkUpdateDescriptorSets vkUpdateDescriptorSets; @@ -469,9 +482,10 @@ public: PoolAllocations() = default; /// Construct an allocation. Errors are reported through IsOutOfPoolMemory(). - explicit PoolAllocations(std::unique_ptr<AllocationType[]> allocations, std::size_t num, - VkDevice device, PoolType pool, const DeviceDispatch& dld) noexcept - : allocations{std::move(allocations)}, num{num}, device{device}, pool{pool}, dld{&dld} {} + explicit PoolAllocations(std::unique_ptr<AllocationType[]> allocations_, std::size_t num_, + VkDevice device_, PoolType pool_, const DeviceDispatch& dld_) noexcept + : allocations{std::move(allocations_)}, num{num_}, device{device_}, pool{pool_}, + dld{&dld_} {} /// Copying Vulkan allocations is not supported and will never be. PoolAllocations(const PoolAllocations&) = delete; @@ -541,18 +555,14 @@ private: const DeviceDispatch* dld = nullptr; }; -using BufferView = Handle<VkBufferView, VkDevice, DeviceDispatch>; -using DebugCallback = Handle<VkDebugUtilsMessengerEXT, VkInstance, InstanceDispatch>; +using DebugUtilsMessenger = Handle<VkDebugUtilsMessengerEXT, VkInstance, InstanceDispatch>; using DescriptorSetLayout = Handle<VkDescriptorSetLayout, VkDevice, DeviceDispatch>; using DescriptorUpdateTemplateKHR = Handle<VkDescriptorUpdateTemplateKHR, VkDevice, DeviceDispatch>; -using Framebuffer = Handle<VkFramebuffer, VkDevice, DeviceDispatch>; -using ImageView = Handle<VkImageView, VkDevice, DeviceDispatch>; using Pipeline = Handle<VkPipeline, VkDevice, DeviceDispatch>; using PipelineLayout = Handle<VkPipelineLayout, VkDevice, DeviceDispatch>; using QueryPool = Handle<VkQueryPool, VkDevice, DeviceDispatch>; using RenderPass = Handle<VkRenderPass, VkDevice, DeviceDispatch>; using Sampler = Handle<VkSampler, VkDevice, DeviceDispatch>; -using ShaderModule = Handle<VkShaderModule, VkDevice, DeviceDispatch>; using SurfaceKHR = Handle<VkSurfaceKHR, VkInstance, InstanceDispatch>; using DescriptorSets = PoolAllocations<VkDescriptorSet, VkDescriptorPool>; @@ -563,16 +573,25 @@ class Instance : public Handle<VkInstance, NoOwner, InstanceDispatch> { using Handle<VkInstance, NoOwner, InstanceDispatch>::Handle; public: - /// Creates a Vulkan instance. Use "operator bool" for error handling. - static Instance Create(Span<const char*> layers, Span<const char*> extensions, - InstanceDispatch& dld) noexcept; + /// Creates a Vulkan instance. + /// @throw Exception on initialization error. + static Instance Create(u32 version, Span<const char*> layers, Span<const char*> extensions, + InstanceDispatch& dispatch); /// Enumerates physical devices. /// @return Physical devices and an empty handle on failure. - std::optional<std::vector<VkPhysicalDevice>> EnumeratePhysicalDevices(); + /// @throw Exception on Vulkan error. + std::vector<VkPhysicalDevice> EnumeratePhysicalDevices() const; + + /// Creates a debug callback messenger. + /// @throw Exception on creation failure. + DebugUtilsMessenger CreateDebugUtilsMessenger( + const VkDebugUtilsMessengerCreateInfoEXT& create_info) const; - /// Tries to create a debug callback messenger. Returns an empty handle on failure. - DebugCallback TryCreateDebugCallback(PFN_vkDebugUtilsMessengerCallbackEXT callback) noexcept; + /// Returns dispatch table. + const InstanceDispatch& Dispatch() const noexcept { + return *dld; + } }; class Queue { @@ -581,7 +600,8 @@ public: constexpr Queue() noexcept = default; /// Construct a queue handle. - constexpr Queue(VkQueue queue, const DeviceDispatch& dld) noexcept : queue{queue}, dld{&dld} {} + constexpr Queue(VkQueue queue_, const DeviceDispatch& dld_) noexcept + : queue{queue_}, dld{&dld_} {} VkResult Submit(Span<VkSubmitInfo> submit_infos, VkFence fence = VK_NULL_HANDLE) const noexcept { @@ -603,6 +623,17 @@ class Buffer : public Handle<VkBuffer, VkDevice, DeviceDispatch> { public: /// Attaches a memory allocation. void BindMemory(VkDeviceMemory memory, VkDeviceSize offset) const; + + /// Set object name. + void SetObjectNameEXT(const char* name) const; +}; + +class BufferView : public Handle<VkBufferView, VkDevice, DeviceDispatch> { + using Handle<VkBufferView, VkDevice, DeviceDispatch>::Handle; + +public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; }; class Image : public Handle<VkImage, VkDevice, DeviceDispatch> { @@ -611,12 +642,26 @@ class Image : public Handle<VkImage, VkDevice, DeviceDispatch> { public: /// Attaches a memory allocation. void BindMemory(VkDeviceMemory memory, VkDeviceSize offset) const; + + /// Set object name. + void SetObjectNameEXT(const char* name) const; +}; + +class ImageView : public Handle<VkImageView, VkDevice, DeviceDispatch> { + using Handle<VkImageView, VkDevice, DeviceDispatch>::Handle; + +public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; }; class DeviceMemory : public Handle<VkDeviceMemory, VkDevice, DeviceDispatch> { using Handle<VkDeviceMemory, VkDevice, DeviceDispatch>::Handle; public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; + u8* Map(VkDeviceSize offset, VkDeviceSize size) const { void* data; Check(dld->vkMapMemory(owner, handle, offset, size, 0, &data)); @@ -632,6 +677,9 @@ class Fence : public Handle<VkFence, VkDevice, DeviceDispatch> { using Handle<VkFence, VkDevice, DeviceDispatch>::Handle; public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; + VkResult Wait(u64 timeout = std::numeric_limits<u64>::max()) const noexcept { return dld->vkWaitForFences(owner, 1, &handle, true, timeout); } @@ -645,11 +693,22 @@ public: } }; +class Framebuffer : public Handle<VkFramebuffer, VkDevice, DeviceDispatch> { + using Handle<VkFramebuffer, VkDevice, DeviceDispatch>::Handle; + +public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; +}; + class DescriptorPool : public Handle<VkDescriptorPool, VkDevice, DeviceDispatch> { using Handle<VkDescriptorPool, VkDevice, DeviceDispatch>::Handle; public: DescriptorSets Allocate(const VkDescriptorSetAllocateInfo& ai) const; + + /// Set object name. + void SetObjectNameEXT(const char* name) const; }; class CommandPool : public Handle<VkCommandPool, VkDevice, DeviceDispatch> { @@ -658,6 +717,9 @@ class CommandPool : public Handle<VkCommandPool, VkDevice, DeviceDispatch> { public: CommandBuffers Allocate(std::size_t num_buffers, VkCommandBufferLevel level = VK_COMMAND_BUFFER_LEVEL_PRIMARY) const; + + /// Set object name. + void SetObjectNameEXT(const char* name) const; }; class SwapchainKHR : public Handle<VkSwapchainKHR, VkDevice, DeviceDispatch> { @@ -671,15 +733,29 @@ class Event : public Handle<VkEvent, VkDevice, DeviceDispatch> { using Handle<VkEvent, VkDevice, DeviceDispatch>::Handle; public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; + VkResult GetStatus() const noexcept { return dld->vkGetEventStatus(owner, handle); } }; +class ShaderModule : public Handle<VkShaderModule, VkDevice, DeviceDispatch> { + using Handle<VkShaderModule, VkDevice, DeviceDispatch>::Handle; + +public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; +}; + class Semaphore : public Handle<VkSemaphore, VkDevice, DeviceDispatch> { using Handle<VkSemaphore, VkDevice, DeviceDispatch>::Handle; public: + /// Set object name. + void SetObjectNameEXT(const char* name) const; + [[nodiscard]] u64 GetCounter() const { u64 value; Check(dld->vkGetSemaphoreCounterValueKHR(owner, handle, &value)); @@ -720,7 +796,7 @@ class Device : public Handle<VkDevice, NoOwner, DeviceDispatch> { public: static Device Create(VkPhysicalDevice physical_device, Span<VkDeviceQueueCreateInfo> queues_ci, Span<const char*> enabled_extensions, const void* next, - DeviceDispatch& dld) noexcept; + DeviceDispatch& dispatch); Queue GetQueue(u32 family_index) const noexcept; @@ -809,8 +885,9 @@ class PhysicalDevice { public: constexpr PhysicalDevice() noexcept = default; - constexpr PhysicalDevice(VkPhysicalDevice physical_device, const InstanceDispatch& dld) noexcept - : physical_device{physical_device}, dld{&dld} {} + constexpr PhysicalDevice(VkPhysicalDevice physical_device_, + const InstanceDispatch& dld_) noexcept + : physical_device{physical_device_}, dld{&dld_} {} constexpr operator VkPhysicalDevice() const noexcept { return physical_device; @@ -849,8 +926,8 @@ class CommandBuffer { public: CommandBuffer() noexcept = default; - explicit CommandBuffer(VkCommandBuffer handle, const DeviceDispatch& dld) noexcept - : handle{handle}, dld{&dld} {} + explicit CommandBuffer(VkCommandBuffer handle_, const DeviceDispatch& dld_) noexcept + : handle{handle_}, dld{&dld_} {} const VkCommandBuffer* address() const noexcept { return &handle; @@ -929,6 +1006,12 @@ public: regions.data(), filter); } + void ResolveImage(VkImage src_image, VkImageLayout src_layout, VkImage dst_image, + VkImageLayout dst_layout, Span<VkImageResolve> regions) { + dld->vkCmdResolveImage(handle, src_image, src_layout, dst_image, dst_layout, regions.size(), + regions.data()); + } + void Dispatch(u32 x, u32 y, u32 z) const noexcept { dld->vkCmdDispatch(handle, x, y, z); } @@ -943,6 +1026,23 @@ public: image_barriers.size(), image_barriers.data()); } + void PipelineBarrier(VkPipelineStageFlags src_stage_mask, VkPipelineStageFlags dst_stage_mask, + VkDependencyFlags dependency_flags = 0) const noexcept { + PipelineBarrier(src_stage_mask, dst_stage_mask, dependency_flags, {}, {}, {}); + } + + void PipelineBarrier(VkPipelineStageFlags src_stage_mask, VkPipelineStageFlags dst_stage_mask, + VkDependencyFlags dependency_flags, + const VkBufferMemoryBarrier& buffer_barrier) const noexcept { + PipelineBarrier(src_stage_mask, dst_stage_mask, dependency_flags, {}, buffer_barrier, {}); + } + + void PipelineBarrier(VkPipelineStageFlags src_stage_mask, VkPipelineStageFlags dst_stage_mask, + VkDependencyFlags dependency_flags, + const VkImageMemoryBarrier& image_barrier) const noexcept { + PipelineBarrier(src_stage_mask, dst_stage_mask, dependency_flags, {}, {}, image_barrier); + } + void CopyBufferToImage(VkBuffer src_buffer, VkImage dst_image, VkImageLayout dst_image_layout, Span<VkBufferImageCopy> regions) const noexcept { dld->vkCmdCopyBufferToImage(handle, src_buffer, dst_image, dst_image_layout, regions.size(), @@ -976,6 +1076,13 @@ public: dld->vkCmdPushConstants(handle, layout, flags, offset, size, values); } + template <typename T> + void PushConstants(VkPipelineLayout layout, VkShaderStageFlags flags, + const T& data) const noexcept { + static_assert(std::is_trivially_copyable_v<T>, "<data> is not trivially copyable"); + dld->vkCmdPushConstants(handle, layout, flags, 0, static_cast<u32>(sizeof(T)), &data); + } + void SetViewport(u32 first, Span<VkViewport> viewports) const noexcept { dld->vkCmdSetViewport(handle, first, viewports.size(), viewports.data()); } @@ -1085,11 +1192,27 @@ public: counter_buffers, counter_buffer_offsets); } + void BeginDebugUtilsLabelEXT(const char* label, std::span<float, 4> color) const noexcept { + const VkDebugUtilsLabelEXT label_info{ + .sType = VK_STRUCTURE_TYPE_DEBUG_UTILS_LABEL_EXT, + .pNext = nullptr, + .pLabelName = label, + .color{color[0], color[1], color[2], color[3]}, + }; + dld->vkCmdBeginDebugUtilsLabelEXT(handle, &label_info); + } + + void EndDebugUtilsLabelEXT() const noexcept { + dld->vkCmdEndDebugUtilsLabelEXT(handle); + } + private: VkCommandBuffer handle; const DeviceDispatch* dld; }; +u32 AvailableVersion(const InstanceDispatch& dld) noexcept; + std::optional<std::vector<VkExtensionProperties>> EnumerateInstanceExtensionProperties( const InstanceDispatch& dld); diff --git a/src/web_service/CMakeLists.txt b/src/web_service/CMakeLists.txt index 7e484b906..ae85a72ea 100644 --- a/src/web_service/CMakeLists.txt +++ b/src/web_service/CMakeLists.txt @@ -9,4 +9,4 @@ add_library(web_service STATIC ) create_target_directory_groups(web_service) -target_link_libraries(web_service PRIVATE common nlohmann_json::nlohmann_json httplib lurlparser) +target_link_libraries(web_service PRIVATE common nlohmann_json::nlohmann_json httplib) diff --git a/src/web_service/web_backend.cpp b/src/web_service/web_backend.cpp index 74e287045..67183e64c 100644 --- a/src/web_service/web_backend.cpp +++ b/src/web_service/web_backend.cpp @@ -7,7 +7,6 @@ #include <mutex> #include <string> -#include <LUrlParser.h> #include <fmt/format.h> #include <httplib.h> @@ -19,9 +18,6 @@ namespace WebService { constexpr std::array<const char, 1> API_VERSION{'1'}; -constexpr int HTTP_PORT = 80; -constexpr int HTTPS_PORT = 443; - constexpr std::size_t TIMEOUT_SECONDS = 30; struct Client::Impl { @@ -67,28 +63,17 @@ struct Client::Impl { const std::string& jwt = "", const std::string& username = "", const std::string& token = "") { if (cli == nullptr) { - auto parsedUrl = LUrlParser::clParseURL::ParseURL(host); - int port; - if (parsedUrl.m_Scheme == "http") { - if (!parsedUrl.GetPort(&port)) { - port = HTTP_PORT; - } - cli = std::make_unique<httplib::Client>(parsedUrl.m_Host.c_str(), port); - } else if (parsedUrl.m_Scheme == "https") { - if (!parsedUrl.GetPort(&port)) { - port = HTTPS_PORT; - } - cli = std::make_unique<httplib::SSLClient>(parsedUrl.m_Host.c_str(), port); - } else { - LOG_ERROR(WebService, "Bad URL scheme {}", parsedUrl.m_Scheme); - return WebResult{WebResult::Code::InvalidURL, "Bad URL scheme", ""}; - } + cli = std::make_unique<httplib::Client>(host.c_str()); } - if (cli == nullptr) { - LOG_ERROR(WebService, "Invalid URL {}", host + path); - return WebResult{WebResult::Code::InvalidURL, "Invalid URL", ""}; + + if (!cli->is_valid()) { + LOG_ERROR(WebService, "Client is invalid, skipping request!"); + return {}; } - cli->set_timeout_sec(TIMEOUT_SECONDS); + + cli->set_connection_timeout(TIMEOUT_SECONDS); + cli->set_read_timeout(TIMEOUT_SECONDS); + cli->set_write_timeout(TIMEOUT_SECONDS); httplib::Headers params; if (!jwt.empty()) { diff --git a/src/yuzu/CMakeLists.txt b/src/yuzu/CMakeLists.txt index cc0291b15..e1bab2112 100644 --- a/src/yuzu/CMakeLists.txt +++ b/src/yuzu/CMakeLists.txt @@ -68,12 +68,12 @@ add_executable(yuzu configuration/configure_input_advanced.cpp configuration/configure_input_advanced.h configuration/configure_input_advanced.ui - configuration/configure_input_dialog.cpp - configuration/configure_input_dialog.h - configuration/configure_input_dialog.ui configuration/configure_input_player.cpp configuration/configure_input_player.h configuration/configure_input_player.ui + configuration/configure_input_profile_dialog.cpp + configuration/configure_input_profile_dialog.h + configuration/configure_input_profile_dialog.ui configuration/configure_motion_touch.cpp configuration/configure_motion_touch.h configuration/configure_motion_touch.ui @@ -105,9 +105,14 @@ add_executable(yuzu configuration/configure_ui.cpp configuration/configure_ui.h configuration/configure_ui.ui + configuration/configure_vibration.cpp + configuration/configure_vibration.h + configuration/configure_vibration.ui configuration/configure_web.cpp configuration/configure_web.h configuration/configure_web.ui + configuration/input_profiles.cpp + configuration/input_profiles.h debugger/console.cpp debugger/console.h debugger/profiler.cpp @@ -136,6 +141,8 @@ add_executable(yuzu util/limitable_input_dialog.h util/sequence_dialog/sequence_dialog.cpp util/sequence_dialog/sequence_dialog.h + util/url_request_interceptor.cpp + util/url_request_interceptor.h util/util.cpp util/util.h compatdb.cpp @@ -212,7 +219,8 @@ target_link_libraries(yuzu PRIVATE common core input_common video_core) target_link_libraries(yuzu PRIVATE Boost::boost glad Qt5::Widgets) target_link_libraries(yuzu PRIVATE ${PLATFORM_LIBRARIES} Threads::Threads) -if (ENABLE_VULKAN AND NOT WIN32) +target_include_directories(yuzu PRIVATE ../../externals/Vulkan-Headers/include) +if (NOT WIN32) target_include_directories(yuzu PRIVATE ${Qt5Gui_PRIVATE_INCLUDE_DIRS}) endif() @@ -264,17 +272,12 @@ endif() if (MSVC) include(CopyYuzuQt5Deps) include(CopyYuzuSDLDeps) - include(CopyYuzuUnicornDeps) + include(CopyYuzuFFmpegDeps) copy_yuzu_Qt5_deps(yuzu) copy_yuzu_SDL_deps(yuzu) - copy_yuzu_unicorn_deps(yuzu) + copy_yuzu_FFmpeg_deps(yuzu) endif() if (NOT APPLE) target_compile_definitions(yuzu PRIVATE HAS_OPENGL) endif() - -if (ENABLE_VULKAN) - target_include_directories(yuzu PRIVATE ../../externals/Vulkan-Headers/include) - target_compile_definitions(yuzu PRIVATE HAS_VULKAN) -endif() diff --git a/src/yuzu/aboutdialog.ui b/src/yuzu/aboutdialog.ui index f122ba39d..1b320630c 100644 --- a/src/yuzu/aboutdialog.ui +++ b/src/yuzu/aboutdialog.ui @@ -160,32 +160,12 @@ p, li { white-space: pre-wrap; } <signal>accepted()</signal> <receiver>AboutDialog</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>248</x> - <y>254</y> - </hint> - <hint type="destinationlabel"> - <x>157</x> - <y>274</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>AboutDialog</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>316</x> - <y>260</y> - </hint> - <hint type="destinationlabel"> - <x>286</x> - <y>274</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/applets/controller.cpp b/src/yuzu/applets/controller.cpp index 9d45f2a01..c680fd2c2 100644 --- a/src/yuzu/applets/controller.cpp +++ b/src/yuzu/applets/controller.cpp @@ -3,6 +3,7 @@ // Refer to the license.txt file included. #include <algorithm> +#include <thread> #include "common/assert.h" #include "common/string_util.h" @@ -13,20 +14,25 @@ #include "core/hle/service/sm/sm.h" #include "ui_controller.h" #include "yuzu/applets/controller.h" -#include "yuzu/configuration/configure_input_dialog.h" +#include "yuzu/configuration/configure_input.h" +#include "yuzu/configuration/configure_input_profile_dialog.h" +#include "yuzu/configuration/configure_vibration.h" +#include "yuzu/configuration/input_profiles.h" #include "yuzu/main.h" namespace { -constexpr std::array<std::array<bool, 4>, 8> led_patterns = {{ - {1, 0, 0, 0}, - {1, 1, 0, 0}, - {1, 1, 1, 0}, - {1, 1, 1, 1}, - {1, 0, 0, 1}, - {1, 0, 1, 0}, - {1, 0, 1, 1}, - {0, 1, 1, 0}, +constexpr std::size_t HANDHELD_INDEX = 8; + +constexpr std::array<std::array<bool, 4>, 8> led_patterns{{ + {true, false, false, false}, + {true, true, false, false}, + {true, true, true, false}, + {true, true, true, true}, + {true, false, false, true}, + {true, false, true, false}, + {true, false, true, true}, + {false, true, true, false}, }}; void UpdateController(Settings::ControllerType controller_type, std::size_t npad_index, @@ -66,47 +72,14 @@ bool IsControllerCompatible(Settings::ControllerType controller_type, } } -/// Maps the controller type combobox index to Controller Type enum -constexpr Settings::ControllerType GetControllerTypeFromIndex(int index) { - switch (index) { - case 0: - default: - return Settings::ControllerType::ProController; - case 1: - return Settings::ControllerType::DualJoyconDetached; - case 2: - return Settings::ControllerType::LeftJoycon; - case 3: - return Settings::ControllerType::RightJoycon; - case 4: - return Settings::ControllerType::Handheld; - } -} - -/// Maps the Controller Type enum to controller type combobox index -constexpr int GetIndexFromControllerType(Settings::ControllerType type) { - switch (type) { - case Settings::ControllerType::ProController: - default: - return 0; - case Settings::ControllerType::DualJoyconDetached: - return 1; - case Settings::ControllerType::LeftJoycon: - return 2; - case Settings::ControllerType::RightJoycon: - return 3; - case Settings::ControllerType::Handheld: - return 4; - } -} - } // namespace QtControllerSelectorDialog::QtControllerSelectorDialog( QWidget* parent, Core::Frontend::ControllerParameters parameters_, InputCommon::InputSubsystem* input_subsystem_) : QDialog(parent), ui(std::make_unique<Ui::QtControllerSelectorDialog>()), - parameters(std::move(parameters_)), input_subsystem(input_subsystem_) { + parameters(std::move(parameters_)), input_subsystem{input_subsystem_}, + input_profiles(std::make_unique<InputProfiles>()) { ui->setupUi(this); player_widgets = { @@ -177,6 +150,11 @@ QtControllerSelectorDialog::QtControllerSelectorDialog( // This avoids unintentionally changing the states of elements while loading them in. SetSupportedControllers(); DisableUnsupportedPlayers(); + + for (std::size_t player_index = 0; player_index < NUM_PLAYERS; ++player_index) { + SetEmulatedControllers(player_index); + } + LoadConfiguration(); for (std::size_t i = 0; i < NUM_PLAYERS; ++i) { @@ -216,19 +194,29 @@ QtControllerSelectorDialog::QtControllerSelectorDialog( if (i == 0) { connect(emulated_controllers[i], qOverload<int>(&QComboBox::currentIndexChanged), - [this](int index) { - UpdateDockedState(GetControllerTypeFromIndex(index) == + [this, i](int index) { + UpdateDockedState(GetControllerTypeFromIndex(index, i) == Settings::ControllerType::Handheld); }); } } + connect(ui->vibrationButton, &QPushButton::clicked, this, + &QtControllerSelectorDialog::CallConfigureVibrationDialog); + connect(ui->inputConfigButton, &QPushButton::clicked, this, - &QtControllerSelectorDialog::CallConfigureInputDialog); + &QtControllerSelectorDialog::CallConfigureInputProfileDialog); connect(ui->buttonBox, &QDialogButtonBox::accepted, this, &QtControllerSelectorDialog::ApplyConfiguration); + // Enhancement: Check if the parameters have already been met before disconnecting controllers. + // If all the parameters are met AND only allows a single player, + // stop the constructor here as we do not need to continue. + if (CheckIfParametersMet() && parameters.enable_single_mode) { + return; + } + // If keep_controllers_connected is false, forcefully disconnect all controllers if (!parameters.keep_controllers_connected) { for (auto player : player_groupboxes) { @@ -236,58 +224,66 @@ QtControllerSelectorDialog::QtControllerSelectorDialog( } } - CheckIfParametersMet(); - resize(0, 0); } QtControllerSelectorDialog::~QtControllerSelectorDialog() = default; -void QtControllerSelectorDialog::ApplyConfiguration() { - // Update the controller state once more, just to be sure they are properly applied. - for (std::size_t index = 0; index < NUM_PLAYERS; ++index) { - UpdateControllerState(index); +int QtControllerSelectorDialog::exec() { + if (parameters_met && parameters.enable_single_mode) { + return QDialog::Accepted; } + return QDialog::exec(); +} - const bool pre_docked_mode = Settings::values.use_docked_mode; - Settings::values.use_docked_mode = ui->radioDocked->isChecked(); - OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode); +void QtControllerSelectorDialog::ApplyConfiguration() { + const bool pre_docked_mode = Settings::values.use_docked_mode.GetValue(); + Settings::values.use_docked_mode.SetValue(ui->radioDocked->isChecked()); + OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode.GetValue()); - Settings::values.vibration_enabled = ui->vibrationGroup->isChecked(); + Settings::values.vibration_enabled.SetValue(ui->vibrationGroup->isChecked()); + Settings::values.motion_enabled.SetValue(ui->motionGroup->isChecked()); } void QtControllerSelectorDialog::LoadConfiguration() { for (std::size_t index = 0; index < NUM_PLAYERS; ++index) { - const auto connected = Settings::values.players[index].connected || - (index == 0 && Settings::values.players[8].connected); + const auto connected = + Settings::values.players.GetValue()[index].connected || + (index == 0 && Settings::values.players.GetValue()[HANDHELD_INDEX].connected); player_groupboxes[index]->setChecked(connected); connected_controller_checkboxes[index]->setChecked(connected); - emulated_controllers[index]->setCurrentIndex( - GetIndexFromControllerType(Settings::values.players[index].controller_type)); + emulated_controllers[index]->setCurrentIndex(GetIndexFromControllerType( + Settings::values.players.GetValue()[index].controller_type, index)); } - UpdateDockedState(Settings::values.players[8].connected); + UpdateDockedState(Settings::values.players.GetValue()[HANDHELD_INDEX].connected); - ui->vibrationGroup->setChecked(Settings::values.vibration_enabled); + ui->vibrationGroup->setChecked(Settings::values.vibration_enabled.GetValue()); + ui->motionGroup->setChecked(Settings::values.motion_enabled.GetValue()); } -void QtControllerSelectorDialog::CallConfigureInputDialog() { - const auto max_supported_players = parameters.enable_single_mode ? 1 : parameters.max_players; - - ConfigureInputDialog dialog(this, max_supported_players, input_subsystem); +void QtControllerSelectorDialog::CallConfigureVibrationDialog() { + ConfigureVibration dialog(this); dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | Qt::WindowSystemMenuHint); dialog.setWindowModality(Qt::WindowModal); - dialog.exec(); - dialog.ApplyConfiguration(); + if (dialog.exec() == QDialog::Accepted) { + dialog.ApplyConfiguration(); + } +} - LoadConfiguration(); - CheckIfParametersMet(); +void QtControllerSelectorDialog::CallConfigureInputProfileDialog() { + ConfigureInputProfileDialog dialog(this, input_subsystem, input_profiles.get()); + + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | + Qt::WindowSystemMenuHint); + dialog.setWindowModality(Qt::WindowModal); + dialog.exec(); } -void QtControllerSelectorDialog::CheckIfParametersMet() { +bool QtControllerSelectorDialog::CheckIfParametersMet() { // Here, we check and validate the current configuration against all applicable parameters. const auto num_connected_players = static_cast<int>( std::count_if(player_groupboxes.begin(), player_groupboxes.end(), @@ -301,7 +297,7 @@ void QtControllerSelectorDialog::CheckIfParametersMet() { num_connected_players > max_supported_players) { parameters_met = false; ui->buttonBox->setEnabled(parameters_met); - return; + return parameters_met; } // Next, check against all connected controllers. @@ -313,7 +309,7 @@ void QtControllerSelectorDialog::CheckIfParametersMet() { } const auto compatible = IsControllerCompatible( - GetControllerTypeFromIndex(emulated_controllers[index]->currentIndex()), + GetControllerTypeFromIndex(emulated_controllers[index]->currentIndex(), index), parameters); // If any controller is found to be incompatible, return false early. @@ -326,18 +322,13 @@ void QtControllerSelectorDialog::CheckIfParametersMet() { return true; }(); - if (!all_controllers_compatible) { - parameters_met = false; - ui->buttonBox->setEnabled(parameters_met); - return; - } - - parameters_met = true; + parameters_met = all_controllers_compatible; ui->buttonBox->setEnabled(parameters_met); + return parameters_met; } void QtControllerSelectorDialog::SetSupportedControllers() { - const QString theme = [this] { + const QString theme = [] { if (QIcon::themeName().contains(QStringLiteral("dark"))) { return QStringLiteral("_dark"); } else if (QIcon::themeName().contains(QStringLiteral("midnight"))) { @@ -402,6 +393,63 @@ void QtControllerSelectorDialog::SetSupportedControllers() { } } +void QtControllerSelectorDialog::SetEmulatedControllers(std::size_t player_index) { + auto& pairs = index_controller_type_pairs[player_index]; + + pairs.clear(); + emulated_controllers[player_index]->clear(); + + pairs.emplace_back(emulated_controllers[player_index]->count(), + Settings::ControllerType::ProController); + emulated_controllers[player_index]->addItem(tr("Pro Controller")); + + pairs.emplace_back(emulated_controllers[player_index]->count(), + Settings::ControllerType::DualJoyconDetached); + emulated_controllers[player_index]->addItem(tr("Dual Joycons")); + + pairs.emplace_back(emulated_controllers[player_index]->count(), + Settings::ControllerType::LeftJoycon); + emulated_controllers[player_index]->addItem(tr("Left Joycon")); + + pairs.emplace_back(emulated_controllers[player_index]->count(), + Settings::ControllerType::RightJoycon); + emulated_controllers[player_index]->addItem(tr("Right Joycon")); + + if (player_index == 0) { + pairs.emplace_back(emulated_controllers[player_index]->count(), + Settings::ControllerType::Handheld); + emulated_controllers[player_index]->addItem(tr("Handheld")); + } +} + +Settings::ControllerType QtControllerSelectorDialog::GetControllerTypeFromIndex( + int index, std::size_t player_index) const { + const auto& pairs = index_controller_type_pairs[player_index]; + + const auto it = std::find_if(pairs.begin(), pairs.end(), + [index](const auto& pair) { return pair.first == index; }); + + if (it == pairs.end()) { + return Settings::ControllerType::ProController; + } + + return it->second; +} + +int QtControllerSelectorDialog::GetIndexFromControllerType(Settings::ControllerType type, + std::size_t player_index) const { + const auto& pairs = index_controller_type_pairs[player_index]; + + const auto it = std::find_if(pairs.begin(), pairs.end(), + [type](const auto& pair) { return pair.second == type; }); + + if (it == pairs.end()) { + return 0; + } + + return it->first; +} + void QtControllerSelectorDialog::UpdateControllerIcon(std::size_t player_index) { if (!player_groupboxes[player_index]->isChecked()) { connected_controller_icons[player_index]->setStyleSheet(QString{}); @@ -410,7 +458,8 @@ void QtControllerSelectorDialog::UpdateControllerIcon(std::size_t player_index) } const QString stylesheet = [this, player_index] { - switch (GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex())) { + switch (GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex(), + player_index)) { case Settings::ControllerType::ProController: return QStringLiteral("image: url(:/controller/applet_pro_controller%0); "); case Settings::ControllerType::DualJoyconDetached: @@ -426,7 +475,13 @@ void QtControllerSelectorDialog::UpdateControllerIcon(std::size_t player_index) } }(); - const QString theme = [this] { + if (stylesheet.isEmpty()) { + connected_controller_icons[player_index]->setStyleSheet(QString{}); + player_labels[player_index]->show(); + return; + } + + const QString theme = [] { if (QIcon::themeName().contains(QStringLiteral("dark"))) { return QStringLiteral("_dark"); } else if (QIcon::themeName().contains(QStringLiteral("midnight"))) { @@ -441,38 +496,54 @@ void QtControllerSelectorDialog::UpdateControllerIcon(std::size_t player_index) } void QtControllerSelectorDialog::UpdateControllerState(std::size_t player_index) { - auto& player = Settings::values.players[player_index]; + auto& player = Settings::values.players.GetValue()[player_index]; - player.controller_type = - GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex()); - player.connected = player_groupboxes[player_index]->isChecked(); + const auto controller_type = GetControllerTypeFromIndex( + emulated_controllers[player_index]->currentIndex(), player_index); + const auto player_connected = player_groupboxes[player_index]->isChecked() && + controller_type != Settings::ControllerType::Handheld; - // Player 2-8 - if (player_index != 0) { - UpdateController(player.controller_type, player_index, player.connected); + if (player.controller_type == controller_type && player.connected == player_connected) { + // Set vibration devices in the event that the input device has changed. + ConfigureVibration::SetVibrationDevices(player_index); return; } - // Player 1 and Handheld - auto& handheld = Settings::values.players[8]; - // If Handheld is selected, copy all the settings from Player 1 to Handheld. - if (player.controller_type == Settings::ControllerType::Handheld) { - handheld = player; - handheld.connected = player_groupboxes[player_index]->isChecked(); - player.connected = false; // Disconnect Player 1 - } else { - player.connected = player_groupboxes[player_index]->isChecked(); - handheld.connected = false; // Disconnect Handheld + // Disconnect the controller first. + UpdateController(controller_type, player_index, false); + + player.controller_type = controller_type; + player.connected = player_connected; + + ConfigureVibration::SetVibrationDevices(player_index); + + // Handheld + if (player_index == 0) { + auto& handheld = Settings::values.players.GetValue()[HANDHELD_INDEX]; + if (controller_type == Settings::ControllerType::Handheld) { + handheld = player; + } + handheld.connected = player_groupboxes[player_index]->isChecked() && + controller_type == Settings::ControllerType::Handheld; + UpdateController(Settings::ControllerType::Handheld, 8, handheld.connected); } - UpdateController(player.controller_type, player_index, player.connected); - UpdateController(Settings::ControllerType::Handheld, 8, handheld.connected); + if (!player.connected) { + return; + } + + // This emulates a delay between disconnecting and reconnecting controllers as some games + // do not respond to a change in controller type if it was instantaneous. + using namespace std::chrono_literals; + std::this_thread::sleep_for(60ms); + + UpdateController(controller_type, player_index, player_connected); } void QtControllerSelectorDialog::UpdateLEDPattern(std::size_t player_index) { if (!player_groupboxes[player_index]->isChecked() || - GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex()) == - Settings::ControllerType::Handheld) { + GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex(), + player_index) == Settings::ControllerType::Handheld) { led_patterns_boxes[player_index][0]->setChecked(false); led_patterns_boxes[player_index][1]->setChecked(false); led_patterns_boxes[player_index][2]->setChecked(false); @@ -520,8 +591,8 @@ void QtControllerSelectorDialog::UpdateDockedState(bool is_handheld) { ui->radioDocked->setEnabled(!is_handheld); ui->radioUndocked->setEnabled(!is_handheld); - ui->radioDocked->setChecked(Settings::values.use_docked_mode); - ui->radioUndocked->setChecked(!Settings::values.use_docked_mode); + ui->radioDocked->setChecked(Settings::values.use_docked_mode.GetValue()); + ui->radioUndocked->setChecked(!Settings::values.use_docked_mode.GetValue()); // Also force into undocked mode if the controller type is handheld. if (is_handheld) { @@ -564,8 +635,8 @@ void QtControllerSelectorDialog::DisableUnsupportedPlayers() { for (std::size_t index = max_supported_players; index < NUM_PLAYERS; ++index) { // Disconnect any unsupported players here and disable or hide them if applicable. - Settings::values.players[index].connected = false; - UpdateController(Settings::values.players[index].controller_type, index, false); + Settings::values.players.GetValue()[index].connected = false; + UpdateController(Settings::values.players.GetValue()[index].controller_type, index, false); // Hide the player widgets when max_supported_controllers is less than or equal to 4. if (max_supported_players <= 4) { player_widgets[index]->hide(); @@ -589,13 +660,13 @@ QtControllerSelector::QtControllerSelector(GMainWindow& parent) { QtControllerSelector::~QtControllerSelector() = default; void QtControllerSelector::ReconfigureControllers( - std::function<void()> callback, Core::Frontend::ControllerParameters parameters) const { - this->callback = std::move(callback); + std::function<void()> callback_, const Core::Frontend::ControllerParameters& parameters) const { + callback = std::move(callback_); emit MainWindowReconfigureControllers(parameters); } void QtControllerSelector::MainWindowReconfigureFinished() { // Acquire the HLE mutex - std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock); + std::lock_guard lock(HLE::g_hle_lock); callback(); } diff --git a/src/yuzu/applets/controller.h b/src/yuzu/applets/controller.h index 2d6d588c6..3518eed56 100644 --- a/src/yuzu/applets/controller.h +++ b/src/yuzu/applets/controller.h @@ -16,10 +16,16 @@ class QDialogButtonBox; class QGroupBox; class QLabel; +class InputProfiles; + namespace InputCommon { class InputSubsystem; } +namespace Settings { +enum class ControllerType; +} + namespace Ui { class QtControllerSelectorDialog; } @@ -33,6 +39,8 @@ public: InputCommon::InputSubsystem* input_subsystem_); ~QtControllerSelectorDialog() override; + int exec() override; + private: // Applies the current configuration. void ApplyConfiguration(); @@ -40,16 +48,28 @@ private: // Loads the current input configuration into the frontend applet. void LoadConfiguration(); - // Initializes the "Configure Input" Dialog. - void CallConfigureInputDialog(); + // Initializes the "Configure Vibration" Dialog. + void CallConfigureVibrationDialog(); + + // Initializes the "Create Input Profile" Dialog. + void CallConfigureInputProfileDialog(); - // Checks the current configuration against the given parameters and - // sets the value of parameters_met. - void CheckIfParametersMet(); + // Checks the current configuration against the given parameters. + // This sets and returns the value of parameters_met. + bool CheckIfParametersMet(); // Sets the controller icons for "Supported Controller Types". void SetSupportedControllers(); + // Sets the emulated controllers per player. + void SetEmulatedControllers(std::size_t player_index); + + // Gets the Controller Type for a given controller combobox index per player. + Settings::ControllerType GetControllerTypeFromIndex(int index, std::size_t player_index) const; + + // Gets the controller combobox index for a given Controller Type per player. + int GetIndexFromControllerType(Settings::ControllerType type, std::size_t player_index) const; + // Updates the controller icons per player. void UpdateControllerIcon(std::size_t player_index); @@ -78,6 +98,8 @@ private: InputCommon::InputSubsystem* input_subsystem; + std::unique_ptr<InputProfiles> input_profiles; + // This is true if and only if all parameters are met. Otherwise, this is false. // This determines whether the "OK" button can be clicked to exit the applet. bool parameters_met{false}; @@ -105,6 +127,10 @@ private: // Comboboxes with a list of emulated controllers per player. std::array<QComboBox*, NUM_PLAYERS> emulated_controllers; + /// Pairs of emulated controller index and Controller Type enum per player. + std::array<std::vector<std::pair<int, Settings::ControllerType>>, NUM_PLAYERS> + index_controller_type_pairs; + // Labels representing the number of connected controllers // above the "Connected Controllers" checkboxes. std::array<QLabel*, NUM_PLAYERS> connected_controller_labels; @@ -120,11 +146,13 @@ public: explicit QtControllerSelector(GMainWindow& parent); ~QtControllerSelector() override; - void ReconfigureControllers(std::function<void()> callback, - Core::Frontend::ControllerParameters parameters) const override; + void ReconfigureControllers( + std::function<void()> callback_, + const Core::Frontend::ControllerParameters& parameters) const override; signals: - void MainWindowReconfigureControllers(Core::Frontend::ControllerParameters parameters) const; + void MainWindowReconfigureControllers( + const Core::Frontend::ControllerParameters& parameters) const; private: void MainWindowReconfigureFinished(); diff --git a/src/yuzu/applets/controller.ui b/src/yuzu/applets/controller.ui index c4108a979..c8cb6bcf3 100644 --- a/src/yuzu/applets/controller.ui +++ b/src/yuzu/applets/controller.ui @@ -1217,9 +1217,6 @@ </item> <item> <widget class="QComboBox" name="comboPlayer3Emulated"> - <property name="editable"> - <bool>false</bool> - </property> <item> <property name="text"> <string>Pro Controller</string> @@ -2279,7 +2276,7 @@ <number>6</number> </property> <property name="leftMargin"> - <number>6</number> + <number>8</number> </property> <property name="topMargin"> <number>6</number> @@ -2332,30 +2329,24 @@ <number>3</number> </property> <item> - <widget class="QSpinBox" name="vibrationSpin"> + <widget class="QPushButton" name="vibrationButton"> <property name="minimumSize"> <size> - <width>65</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>65</width> + <width>68</width> <height>16777215</height> </size> </property> - <property name="suffix"> - <string>%</string> - </property> - <property name="minimum"> - <number>1</number> - </property> - <property name="maximum"> - <number>200</number> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> </property> - <property name="value"> - <number>100</number> + <property name="text"> + <string>Configure</string> </property> </widget> </item> @@ -2387,18 +2378,18 @@ <widget class="QPushButton" name="motionButton"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Configure</string> @@ -2411,7 +2402,7 @@ <item> <widget class="QGroupBox" name="inputConfigGroup"> <property name="title"> - <string>Input Config</string> + <string>Profiles</string> </property> <layout class="QHBoxLayout" name="horizontalLayout_7"> <property name="leftMargin"> @@ -2430,15 +2421,15 @@ <widget class="QPushButton" name="inputConfigButton"> <property name="maximumSize"> <size> - <width>65</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> - <string>Open</string> + <string>Create</string> </property> </widget> </item> @@ -2657,16 +2648,6 @@ <signal>accepted()</signal> <receiver>QtControllerSelectorDialog</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>20</x> - <y>20</y> - </hint> - <hint type="destinationlabel"> - <x>20</x> - <y>20</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/applets/error.cpp b/src/yuzu/applets/error.cpp index 08ed57355..8ee03ddb3 100644 --- a/src/yuzu/applets/error.cpp +++ b/src/yuzu/applets/error.cpp @@ -17,9 +17,9 @@ QtErrorDisplay::QtErrorDisplay(GMainWindow& parent) { QtErrorDisplay::~QtErrorDisplay() = default; void QtErrorDisplay::ShowError(ResultCode error, std::function<void()> finished) const { - this->callback = std::move(finished); + callback = std::move(finished); emit MainWindowDisplayError( - tr("An error has occured.\nPlease try again or contact the developer of the " + tr("An error has occurred.\nPlease try again or contact the developer of the " "software.\n\nError Code: %1-%2 (0x%3)") .arg(static_cast<u32>(error.module.Value()) + 2000, 4, 10, QChar::fromLatin1('0')) .arg(error.description, 4, 10, QChar::fromLatin1('0')) @@ -28,11 +28,11 @@ void QtErrorDisplay::ShowError(ResultCode error, std::function<void()> finished) void QtErrorDisplay::ShowErrorWithTimestamp(ResultCode error, std::chrono::seconds time, std::function<void()> finished) const { - this->callback = std::move(finished); + callback = std::move(finished); const QDateTime date_time = QDateTime::fromSecsSinceEpoch(time.count()); emit MainWindowDisplayError( - tr("An error occured on %1 at %2.\nPlease try again or contact the " + tr("An error occurred on %1 at %2.\nPlease try again or contact the " "developer of the software.\n\nError Code: %3-%4 (0x%5)") .arg(date_time.toString(QStringLiteral("dddd, MMMM d, yyyy"))) .arg(date_time.toString(QStringLiteral("h:mm:ss A"))) @@ -44,9 +44,9 @@ void QtErrorDisplay::ShowErrorWithTimestamp(ResultCode error, std::chrono::secon void QtErrorDisplay::ShowCustomErrorText(ResultCode error, std::string dialog_text, std::string fullscreen_text, std::function<void()> finished) const { - this->callback = std::move(finished); + callback = std::move(finished); emit MainWindowDisplayError( - tr("An error has occured.\nError Code: %1-%2 (0x%3)\n\n%4\n\n%5") + tr("An error has occurred.\nError Code: %1-%2 (0x%3)\n\n%4\n\n%5") .arg(static_cast<u32>(error.module.Value()) + 2000, 4, 10, QChar::fromLatin1('0')) .arg(error.description, 4, 10, QChar::fromLatin1('0')) .arg(error.raw, 8, 16, QChar::fromLatin1('0')) diff --git a/src/yuzu/applets/profile_select.cpp b/src/yuzu/applets/profile_select.cpp index dca8835ed..4bf2bfd40 100644 --- a/src/yuzu/applets/profile_select.cpp +++ b/src/yuzu/applets/profile_select.cpp @@ -114,6 +114,15 @@ QtProfileSelectionDialog::QtProfileSelectionDialog(QWidget* parent) QtProfileSelectionDialog::~QtProfileSelectionDialog() = default; +int QtProfileSelectionDialog::exec() { + // Skip profile selection when there's only one. + if (profile_manager->GetUserCount() == 1) { + user_index = 0; + return QDialog::Accepted; + } + return QDialog::exec(); +} + void QtProfileSelectionDialog::accept() { QDialog::accept(); } @@ -141,8 +150,8 @@ QtProfileSelector::QtProfileSelector(GMainWindow& parent) { QtProfileSelector::~QtProfileSelector() = default; void QtProfileSelector::SelectProfile( - std::function<void(std::optional<Common::UUID>)> callback) const { - this->callback = std::move(callback); + std::function<void(std::optional<Common::UUID>)> callback_) const { + callback = std::move(callback_); emit MainWindowSelectProfile(); } diff --git a/src/yuzu/applets/profile_select.h b/src/yuzu/applets/profile_select.h index cee886a77..4e9037488 100644 --- a/src/yuzu/applets/profile_select.h +++ b/src/yuzu/applets/profile_select.h @@ -27,6 +27,7 @@ public: explicit QtProfileSelectionDialog(QWidget* parent); ~QtProfileSelectionDialog() override; + int exec() override; void accept() override; void reject() override; @@ -59,7 +60,7 @@ public: explicit QtProfileSelector(GMainWindow& parent); ~QtProfileSelector() override; - void SelectProfile(std::function<void(std::optional<Common::UUID>)> callback) const override; + void SelectProfile(std::function<void(std::optional<Common::UUID>)> callback_) const override; signals: void MainWindowSelectProfile() const; diff --git a/src/yuzu/applets/software_keyboard.cpp b/src/yuzu/applets/software_keyboard.cpp index af36f07c6..ab8cfd8ee 100644 --- a/src/yuzu/applets/software_keyboard.cpp +++ b/src/yuzu/applets/software_keyboard.cpp @@ -135,8 +135,8 @@ void QtSoftwareKeyboard::RequestText(std::function<void(std::optional<std::u16st } void QtSoftwareKeyboard::SendTextCheckDialog(std::u16string error_message, - std::function<void()> finished_check) const { - this->finished_check = std::move(finished_check); + std::function<void()> finished_check_) const { + finished_check = std::move(finished_check_); emit MainWindowTextCheckDialog(error_message); } diff --git a/src/yuzu/applets/software_keyboard.h b/src/yuzu/applets/software_keyboard.h index 44bcece75..9e1094cce 100644 --- a/src/yuzu/applets/software_keyboard.h +++ b/src/yuzu/applets/software_keyboard.h @@ -61,7 +61,7 @@ public: void RequestText(std::function<void(std::optional<std::u16string>)> out, Core::Frontend::SoftwareKeyboardParameters parameters) const override; void SendTextCheckDialog(std::u16string error_message, - std::function<void()> finished_check) const override; + std::function<void()> finished_check_) const override; signals: void MainWindowGetText(Core::Frontend::SoftwareKeyboardParameters parameters) const; diff --git a/src/yuzu/applets/web_browser.cpp b/src/yuzu/applets/web_browser.cpp index 33f1c385d..e482ba029 100644 --- a/src/yuzu/applets/web_browser.cpp +++ b/src/yuzu/applets/web_browser.cpp @@ -1,115 +1,414 @@ -// Copyright 2018 yuzu Emulator Project +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <mutex> - +#ifdef YUZU_USE_QT_WEB_ENGINE #include <QKeyEvent> -#include "core/hle/lock.h" +#include <QWebEngineProfile> +#include <QWebEngineScript> +#include <QWebEngineScriptCollection> +#include <QWebEngineSettings> +#include <QWebEngineUrlScheme> +#endif + +#include "common/file_util.h" +#include "core/core.h" +#include "core/frontend/input_interpreter.h" +#include "input_common/keyboard.h" +#include "input_common/main.h" #include "yuzu/applets/web_browser.h" +#include "yuzu/applets/web_browser_scripts.h" #include "yuzu/main.h" +#include "yuzu/util/url_request_interceptor.h" #ifdef YUZU_USE_QT_WEB_ENGINE -constexpr char NX_SHIM_INJECT_SCRIPT[] = R"( - window.nx = {}; - window.nx.playReport = {}; - window.nx.playReport.setCounterSetIdentifier = function () { - console.log("nx.playReport.setCounterSetIdentifier called - unimplemented"); - }; +namespace { - window.nx.playReport.incrementCounter = function () { - console.log("nx.playReport.incrementCounter called - unimplemented"); - }; +constexpr int HIDButtonToKey(HIDButton button) { + switch (button) { + case HIDButton::DLeft: + case HIDButton::LStickLeft: + return Qt::Key_Left; + case HIDButton::DUp: + case HIDButton::LStickUp: + return Qt::Key_Up; + case HIDButton::DRight: + case HIDButton::LStickRight: + return Qt::Key_Right; + case HIDButton::DDown: + case HIDButton::LStickDown: + return Qt::Key_Down; + default: + return 0; + } +} + +} // Anonymous namespace + +QtNXWebEngineView::QtNXWebEngineView(QWidget* parent, Core::System& system, + InputCommon::InputSubsystem* input_subsystem_) + : QWebEngineView(parent), input_subsystem{input_subsystem_}, + url_interceptor(std::make_unique<UrlRequestInterceptor>()), + input_interpreter(std::make_unique<InputInterpreter>(system)), + default_profile{QWebEngineProfile::defaultProfile()}, + global_settings{QWebEngineSettings::globalSettings()} { + QWebEngineScript gamepad; + QWebEngineScript window_nx; + + gamepad.setName(QStringLiteral("gamepad_script.js")); + window_nx.setName(QStringLiteral("window_nx_script.js")); + + gamepad.setSourceCode(QString::fromStdString(GAMEPAD_SCRIPT)); + window_nx.setSourceCode(QString::fromStdString(WINDOW_NX_SCRIPT)); + + gamepad.setInjectionPoint(QWebEngineScript::DocumentCreation); + window_nx.setInjectionPoint(QWebEngineScript::DocumentCreation); + + gamepad.setWorldId(QWebEngineScript::MainWorld); + window_nx.setWorldId(QWebEngineScript::MainWorld); + + gamepad.setRunsOnSubFrames(true); + window_nx.setRunsOnSubFrames(true); + + default_profile->scripts()->insert(gamepad); + default_profile->scripts()->insert(window_nx); + + default_profile->setRequestInterceptor(url_interceptor.get()); + + global_settings->setAttribute(QWebEngineSettings::LocalContentCanAccessRemoteUrls, true); + global_settings->setAttribute(QWebEngineSettings::FullScreenSupportEnabled, true); + global_settings->setAttribute(QWebEngineSettings::AllowRunningInsecureContent, true); + global_settings->setAttribute(QWebEngineSettings::FocusOnNavigationEnabled, true); + global_settings->setAttribute(QWebEngineSettings::AllowWindowActivationFromJavaScript, true); + global_settings->setAttribute(QWebEngineSettings::ShowScrollBars, false); + + global_settings->setFontFamily(QWebEngineSettings::StandardFont, QStringLiteral("Roboto")); + + connect( + page(), &QWebEnginePage::windowCloseRequested, page(), + [this] { + if (page()->url() == url_interceptor->GetRequestedURL()) { + SetFinished(true); + SetExitReason(Service::AM::Applets::WebExitReason::WindowClosed); + } + }, + Qt::QueuedConnection); +} + +QtNXWebEngineView::~QtNXWebEngineView() { + SetFinished(true); + StopInputThread(); +} + +void QtNXWebEngineView::LoadLocalWebPage(std::string_view main_url, + std::string_view additional_args) { + is_local = true; + + LoadExtractedFonts(); + SetUserAgent(UserAgent::WebApplet); + SetFinished(false); + SetExitReason(Service::AM::Applets::WebExitReason::EndButtonPressed); + SetLastURL("http://localhost/"); + StartInputThread(); + + load(QUrl(QUrl::fromLocalFile(QString::fromStdString(std::string(main_url))).toString() + + QString::fromStdString(std::string(additional_args)))); +} + +void QtNXWebEngineView::LoadExternalWebPage(std::string_view main_url, + std::string_view additional_args) { + is_local = false; + + SetUserAgent(UserAgent::WebApplet); + SetFinished(false); + SetExitReason(Service::AM::Applets::WebExitReason::EndButtonPressed); + SetLastURL("http://localhost/"); + StartInputThread(); + + load(QUrl(QString::fromStdString(std::string(main_url)) + + QString::fromStdString(std::string(additional_args)))); +} + +void QtNXWebEngineView::SetUserAgent(UserAgent user_agent) { + const QString user_agent_str = [user_agent] { + switch (user_agent) { + case UserAgent::WebApplet: + default: + return QStringLiteral("WebApplet"); + case UserAgent::ShopN: + return QStringLiteral("ShopN"); + case UserAgent::LoginApplet: + return QStringLiteral("LoginApplet"); + case UserAgent::ShareApplet: + return QStringLiteral("ShareApplet"); + case UserAgent::LobbyApplet: + return QStringLiteral("LobbyApplet"); + case UserAgent::WifiWebAuthApplet: + return QStringLiteral("WifiWebAuthApplet"); + } + }(); + + QWebEngineProfile::defaultProfile()->setHttpUserAgent( + QStringLiteral("Mozilla/5.0 (Nintendo Switch; %1) AppleWebKit/606.4 " + "(KHTML, like Gecko) NF/6.0.1.15.4 NintendoBrowser/5.1.0.20389") + .arg(user_agent_str)); +} + +bool QtNXWebEngineView::IsFinished() const { + return finished; +} + +void QtNXWebEngineView::SetFinished(bool finished_) { + finished = finished_; +} + +Service::AM::Applets::WebExitReason QtNXWebEngineView::GetExitReason() const { + return exit_reason; +} + +void QtNXWebEngineView::SetExitReason(Service::AM::Applets::WebExitReason exit_reason_) { + exit_reason = exit_reason_; +} + +const std::string& QtNXWebEngineView::GetLastURL() const { + return last_url; +} + +void QtNXWebEngineView::SetLastURL(std::string last_url_) { + last_url = std::move(last_url_); +} + +QString QtNXWebEngineView::GetCurrentURL() const { + return url_interceptor->GetRequestedURL().toString(); +} + +void QtNXWebEngineView::hide() { + SetFinished(true); + StopInputThread(); - window.nx.footer = {}; - window.nx.footer.unsetAssign = function () { - console.log("nx.footer.unsetAssign called - unimplemented"); + QWidget::hide(); +} + +void QtNXWebEngineView::keyPressEvent(QKeyEvent* event) { + if (is_local) { + input_subsystem->GetKeyboard()->PressKey(event->key()); + } +} + +void QtNXWebEngineView::keyReleaseEvent(QKeyEvent* event) { + if (is_local) { + input_subsystem->GetKeyboard()->ReleaseKey(event->key()); + } +} + +template <HIDButton... T> +void QtNXWebEngineView::HandleWindowFooterButtonPressedOnce() { + const auto f = [this](HIDButton button) { + if (input_interpreter->IsButtonPressedOnce(button)) { + page()->runJavaScript( + QStringLiteral("yuzu_key_callbacks[%1] == null;").arg(static_cast<u8>(button)), + [&](const QVariant& variant) { + if (variant.toBool()) { + switch (button) { + case HIDButton::A: + SendMultipleKeyPressEvents<Qt::Key_A, Qt::Key_Space, Qt::Key_Return>(); + break; + case HIDButton::B: + SendKeyPressEvent(Qt::Key_B); + break; + case HIDButton::X: + SendKeyPressEvent(Qt::Key_X); + break; + case HIDButton::Y: + SendKeyPressEvent(Qt::Key_Y); + break; + default: + break; + } + } + }); + + page()->runJavaScript( + QStringLiteral("if (yuzu_key_callbacks[%1] != null) { yuzu_key_callbacks[%1](); }") + .arg(static_cast<u8>(button))); + } }; - var yuzu_key_callbacks = []; - window.nx.footer.setAssign = function(key, discard1, func, discard2) { - switch (key) { - case 'A': - yuzu_key_callbacks[0] = func; - break; - case 'B': - yuzu_key_callbacks[1] = func; - break; - case 'X': - yuzu_key_callbacks[2] = func; - break; - case 'Y': - yuzu_key_callbacks[3] = func; - break; - case 'L': - yuzu_key_callbacks[6] = func; - break; - case 'R': - yuzu_key_callbacks[7] = func; - break; + (f(T), ...); +} + +template <HIDButton... T> +void QtNXWebEngineView::HandleWindowKeyButtonPressedOnce() { + const auto f = [this](HIDButton button) { + if (input_interpreter->IsButtonPressedOnce(button)) { + SendKeyPressEvent(HIDButtonToKey(button)); } }; - var applet_done = false; - window.nx.endApplet = function() { - applet_done = true; + (f(T), ...); +} + +template <HIDButton... T> +void QtNXWebEngineView::HandleWindowKeyButtonHold() { + const auto f = [this](HIDButton button) { + if (input_interpreter->IsButtonHeld(button)) { + SendKeyPressEvent(HIDButtonToKey(button)); + } }; - window.onkeypress = function(e) { if (e.keyCode === 13) { applet_done = true; } }; -)"; + (f(T), ...); +} + +void QtNXWebEngineView::SendKeyPressEvent(int key) { + if (key == 0) { + return; + } + + QCoreApplication::postEvent(focusProxy(), + new QKeyEvent(QKeyEvent::KeyPress, key, Qt::NoModifier)); + QCoreApplication::postEvent(focusProxy(), + new QKeyEvent(QKeyEvent::KeyRelease, key, Qt::NoModifier)); +} + +void QtNXWebEngineView::StartInputThread() { + if (input_thread_running) { + return; + } + + input_thread_running = true; + input_thread = std::thread(&QtNXWebEngineView::InputThread, this); +} + +void QtNXWebEngineView::StopInputThread() { + if (is_local) { + QWidget::releaseKeyboard(); + } -QString GetNXShimInjectionScript() { - return QString::fromStdString(NX_SHIM_INJECT_SCRIPT); + input_thread_running = false; + if (input_thread.joinable()) { + input_thread.join(); + } } -NXInputWebEngineView::NXInputWebEngineView(QWidget* parent) : QWebEngineView(parent) {} +void QtNXWebEngineView::InputThread() { + // Wait for 1 second before allowing any inputs to be processed. + std::this_thread::sleep_for(std::chrono::seconds(1)); + + if (is_local) { + QWidget::grabKeyboard(); + } + + while (input_thread_running) { + input_interpreter->PollInput(); + + HandleWindowFooterButtonPressedOnce<HIDButton::A, HIDButton::B, HIDButton::X, HIDButton::Y, + HIDButton::L, HIDButton::R>(); + + HandleWindowKeyButtonPressedOnce<HIDButton::DLeft, HIDButton::DUp, HIDButton::DRight, + HIDButton::DDown, HIDButton::LStickLeft, + HIDButton::LStickUp, HIDButton::LStickRight, + HIDButton::LStickDown>(); -void NXInputWebEngineView::keyPressEvent(QKeyEvent* event) { - parent()->event(event); + HandleWindowKeyButtonHold<HIDButton::DLeft, HIDButton::DUp, HIDButton::DRight, + HIDButton::DDown, HIDButton::LStickLeft, HIDButton::LStickUp, + HIDButton::LStickRight, HIDButton::LStickDown>(); + + std::this_thread::sleep_for(std::chrono::milliseconds(50)); + } } -void NXInputWebEngineView::keyReleaseEvent(QKeyEvent* event) { - parent()->event(event); +void QtNXWebEngineView::LoadExtractedFonts() { + QWebEngineScript nx_font_css; + QWebEngineScript load_nx_font; + + const QString fonts_dir = QString::fromStdString(Common::FS::SanitizePath( + fmt::format("{}/fonts", Common::FS::GetUserPath(Common::FS::UserPath::CacheDir)))); + + nx_font_css.setName(QStringLiteral("nx_font_css.js")); + load_nx_font.setName(QStringLiteral("load_nx_font.js")); + + nx_font_css.setSourceCode( + QString::fromStdString(NX_FONT_CSS) + .arg(fonts_dir + QStringLiteral("/FontStandard.ttf")) + .arg(fonts_dir + QStringLiteral("/FontChineseSimplified.ttf")) + .arg(fonts_dir + QStringLiteral("/FontExtendedChineseSimplified.ttf")) + .arg(fonts_dir + QStringLiteral("/FontChineseTraditional.ttf")) + .arg(fonts_dir + QStringLiteral("/FontKorean.ttf")) + .arg(fonts_dir + QStringLiteral("/FontNintendoExtended.ttf")) + .arg(fonts_dir + QStringLiteral("/FontNintendoExtended2.ttf"))); + load_nx_font.setSourceCode(QString::fromStdString(LOAD_NX_FONT)); + + nx_font_css.setInjectionPoint(QWebEngineScript::DocumentReady); + load_nx_font.setInjectionPoint(QWebEngineScript::Deferred); + + nx_font_css.setWorldId(QWebEngineScript::MainWorld); + load_nx_font.setWorldId(QWebEngineScript::MainWorld); + + nx_font_css.setRunsOnSubFrames(true); + load_nx_font.setRunsOnSubFrames(true); + + default_profile->scripts()->insert(nx_font_css); + default_profile->scripts()->insert(load_nx_font); + + connect( + url_interceptor.get(), &UrlRequestInterceptor::FrameChanged, url_interceptor.get(), + [this] { + std::this_thread::sleep_for(std::chrono::milliseconds(50)); + page()->runJavaScript(QString::fromStdString(LOAD_NX_FONT)); + }, + Qt::QueuedConnection); } #endif QtWebBrowser::QtWebBrowser(GMainWindow& main_window) { - connect(this, &QtWebBrowser::MainWindowOpenPage, &main_window, &GMainWindow::WebBrowserOpenPage, - Qt::QueuedConnection); - connect(&main_window, &GMainWindow::WebBrowserUnpackRomFS, this, - &QtWebBrowser::MainWindowUnpackRomFS, Qt::QueuedConnection); - connect(&main_window, &GMainWindow::WebBrowserFinishedBrowsing, this, - &QtWebBrowser::MainWindowFinishedBrowsing, Qt::QueuedConnection); + connect(this, &QtWebBrowser::MainWindowOpenWebPage, &main_window, + &GMainWindow::WebBrowserOpenWebPage, Qt::QueuedConnection); + connect(&main_window, &GMainWindow::WebBrowserExtractOfflineRomFS, this, + &QtWebBrowser::MainWindowExtractOfflineRomFS, Qt::QueuedConnection); + connect(&main_window, &GMainWindow::WebBrowserClosed, this, + &QtWebBrowser::MainWindowWebBrowserClosed, Qt::QueuedConnection); } QtWebBrowser::~QtWebBrowser() = default; -void QtWebBrowser::OpenPageLocal(std::string_view url, std::function<void()> unpack_romfs_callback, - std::function<void()> finished_callback) { - this->unpack_romfs_callback = std::move(unpack_romfs_callback); - this->finished_callback = std::move(finished_callback); +void QtWebBrowser::OpenLocalWebPage( + std::string_view local_url, std::function<void()> extract_romfs_callback_, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> callback_) const { + extract_romfs_callback = std::move(extract_romfs_callback_); + callback = std::move(callback_); + + const auto index = local_url.find('?'); + + if (index == std::string::npos) { + emit MainWindowOpenWebPage(local_url, "", true); + } else { + emit MainWindowOpenWebPage(local_url.substr(0, index), local_url.substr(index), true); + } +} + +void QtWebBrowser::OpenExternalWebPage( + std::string_view external_url, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> callback_) const { + callback = std::move(callback_); + + const auto index = external_url.find('?'); - const auto index = url.find('?'); if (index == std::string::npos) { - emit MainWindowOpenPage(url, ""); + emit MainWindowOpenWebPage(external_url, "", false); } else { - const auto front = url.substr(0, index); - const auto back = url.substr(index); - emit MainWindowOpenPage(front, back); + emit MainWindowOpenWebPage(external_url.substr(0, index), external_url.substr(index), + false); } } -void QtWebBrowser::MainWindowUnpackRomFS() { - // Acquire the HLE mutex - std::lock_guard lock{HLE::g_hle_lock}; - unpack_romfs_callback(); +void QtWebBrowser::MainWindowExtractOfflineRomFS() { + extract_romfs_callback(); } -void QtWebBrowser::MainWindowFinishedBrowsing() { - // Acquire the HLE mutex - std::lock_guard lock{HLE::g_hle_lock}; - finished_callback(); +void QtWebBrowser::MainWindowWebBrowserClosed(Service::AM::Applets::WebExitReason exit_reason, + std::string last_url) { + callback(exit_reason, last_url); } diff --git a/src/yuzu/applets/web_browser.h b/src/yuzu/applets/web_browser.h index b38437e46..47f960d69 100644 --- a/src/yuzu/applets/web_browser.h +++ b/src/yuzu/applets/web_browser.h @@ -1,10 +1,13 @@ -// Copyright 2018 yuzu Emulator Project +// Copyright 2020 yuzu Emulator Project // Licensed under GPLv2 or any later version // Refer to the license.txt file included. #pragma once -#include <functional> +#include <atomic> +#include <memory> +#include <thread> + #include <QObject> #ifdef YUZU_USE_QT_WEB_ENGINE @@ -13,19 +16,172 @@ #include "core/frontend/applets/web_browser.h" +enum class HIDButton : u8; + class GMainWindow; +class InputInterpreter; +class UrlRequestInterceptor; + +namespace Core { +class System; +} + +namespace InputCommon { +class InputSubsystem; +} #ifdef YUZU_USE_QT_WEB_ENGINE -QString GetNXShimInjectionScript(); +enum class UserAgent { + WebApplet, + ShopN, + LoginApplet, + ShareApplet, + LobbyApplet, + WifiWebAuthApplet, +}; + +class QWebEngineProfile; +class QWebEngineSettings; + +class QtNXWebEngineView : public QWebEngineView { + Q_OBJECT -class NXInputWebEngineView : public QWebEngineView { public: - explicit NXInputWebEngineView(QWidget* parent = nullptr); + explicit QtNXWebEngineView(QWidget* parent, Core::System& system, + InputCommon::InputSubsystem* input_subsystem_); + ~QtNXWebEngineView() override; + + /** + * Loads a HTML document that exists locally. Cannot be used to load external websites. + * + * @param main_url The url to the file. + * @param additional_args Additional arguments appended to the main url. + */ + void LoadLocalWebPage(std::string_view main_url, std::string_view additional_args); + + /** + * Loads an external website. Cannot be used to load local urls. + * + * @param main_url The url to the website. + * @param additional_args Additional arguments appended to the main url. + */ + void LoadExternalWebPage(std::string_view main_url, std::string_view additional_args); + + /** + * Sets the background color of the web page. + * + * @param color The color to set. + */ + void SetBackgroundColor(QColor color); + + /** + * Sets the user agent of the web browser. + * + * @param user_agent The user agent enum. + */ + void SetUserAgent(UserAgent user_agent); + + [[nodiscard]] bool IsFinished() const; + void SetFinished(bool finished_); + + [[nodiscard]] Service::AM::Applets::WebExitReason GetExitReason() const; + void SetExitReason(Service::AM::Applets::WebExitReason exit_reason_); + + [[nodiscard]] const std::string& GetLastURL() const; + void SetLastURL(std::string last_url_); + + /** + * This gets the current URL that has been requested by the webpage. + * This only applies to the main frame. Sub frames and other resources are ignored. + * + * @return Currently requested URL + */ + [[nodiscard]] QString GetCurrentURL() const; + +public slots: + void hide(); protected: void keyPressEvent(QKeyEvent* event) override; void keyReleaseEvent(QKeyEvent* event) override; + +private: + /** + * Handles button presses to execute functions assigned in yuzu_key_callbacks. + * yuzu_key_callbacks contains specialized functions for the buttons in the window footer + * that can be overriden by games to achieve desired functionality. + * + * @tparam HIDButton The list of buttons contained in yuzu_key_callbacks + */ + template <HIDButton... T> + void HandleWindowFooterButtonPressedOnce(); + + /** + * Handles button presses and converts them into keyboard input. + * This should only be used to convert D-Pad or Analog Stick input into arrow keys. + * + * @tparam HIDButton The list of buttons that can be converted into keyboard input. + */ + template <HIDButton... T> + void HandleWindowKeyButtonPressedOnce(); + + /** + * Handles button holds and converts them into keyboard input. + * This should only be used to convert D-Pad or Analog Stick input into arrow keys. + * + * @tparam HIDButton The list of buttons that can be converted into keyboard input. + */ + template <HIDButton... T> + void HandleWindowKeyButtonHold(); + + /** + * Sends a key press event to QWebEngineView. + * + * @param key Qt key code. + */ + void SendKeyPressEvent(int key); + + /** + * Sends multiple key press events to QWebEngineView. + * + * @tparam int Qt key code. + */ + template <int... T> + void SendMultipleKeyPressEvents() { + (SendKeyPressEvent(T), ...); + } + + void StartInputThread(); + void StopInputThread(); + + /// The thread where input is being polled and processed. + void InputThread(); + + /// Loads the extracted fonts using JavaScript. + void LoadExtractedFonts(); + + InputCommon::InputSubsystem* input_subsystem; + + std::unique_ptr<UrlRequestInterceptor> url_interceptor; + + std::unique_ptr<InputInterpreter> input_interpreter; + + std::thread input_thread; + + std::atomic<bool> input_thread_running{}; + + std::atomic<bool> finished{}; + + Service::AM::Applets::WebExitReason exit_reason{ + Service::AM::Applets::WebExitReason::EndButtonPressed}; + + std::string last_url{"http://localhost/"}; + + bool is_local{}; + + QWebEngineProfile* default_profile; + QWebEngineSettings* global_settings; }; #endif @@ -34,19 +190,28 @@ class QtWebBrowser final : public QObject, public Core::Frontend::WebBrowserAppl Q_OBJECT public: - explicit QtWebBrowser(GMainWindow& main_window); + explicit QtWebBrowser(GMainWindow& parent); ~QtWebBrowser() override; - void OpenPageLocal(std::string_view url, std::function<void()> unpack_romfs_callback, - std::function<void()> finished_callback) override; + void OpenLocalWebPage(std::string_view local_url, std::function<void()> extract_romfs_callback_, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> + callback_) const override; + + void OpenExternalWebPage(std::string_view external_url, + std::function<void(Service::AM::Applets::WebExitReason, std::string)> + callback_) const override; signals: - void MainWindowOpenPage(std::string_view filename, std::string_view additional_args) const; + void MainWindowOpenWebPage(std::string_view main_url, std::string_view additional_args, + bool is_local) const; private: - void MainWindowUnpackRomFS(); - void MainWindowFinishedBrowsing(); + void MainWindowExtractOfflineRomFS(); + + void MainWindowWebBrowserClosed(Service::AM::Applets::WebExitReason exit_reason, + std::string last_url); + + mutable std::function<void()> extract_romfs_callback; - std::function<void()> unpack_romfs_callback; - std::function<void()> finished_callback; + mutable std::function<void(Service::AM::Applets::WebExitReason, std::string)> callback; }; diff --git a/src/yuzu/applets/web_browser_scripts.h b/src/yuzu/applets/web_browser_scripts.h new file mode 100644 index 000000000..992837a85 --- /dev/null +++ b/src/yuzu/applets/web_browser_scripts.h @@ -0,0 +1,193 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +constexpr char NX_FONT_CSS[] = R"( +(function() { + css = document.createElement('style'); + css.type = 'text/css'; + css.id = 'nx_font'; + css.innerText = ` +/* FontStandard */ +@font-face { + font-family: 'FontStandard'; + src: url('%1') format('truetype'); +} + +/* FontChineseSimplified */ +@font-face { + font-family: 'FontChineseSimplified'; + src: url('%2') format('truetype'); +} + +/* FontExtendedChineseSimplified */ +@font-face { + font-family: 'FontExtendedChineseSimplified'; + src: url('%3') format('truetype'); +} + +/* FontChineseTraditional */ +@font-face { + font-family: 'FontChineseTraditional'; + src: url('%4') format('truetype'); +} + +/* FontKorean */ +@font-face { + font-family: 'FontKorean'; + src: url('%5') format('truetype'); +} + +/* FontNintendoExtended */ +@font-face { + font-family: 'NintendoExt003'; + src: url('%6') format('truetype'); +} + +/* FontNintendoExtended2 */ +@font-face { + font-family: 'NintendoExt003'; + src: url('%7') format('truetype'); +} +`; + + document.head.appendChild(css); +})(); +)"; + +constexpr char LOAD_NX_FONT[] = R"( +(function() { + var elements = document.querySelectorAll("*"); + + for (var i = 0; i < elements.length; i++) { + var style = window.getComputedStyle(elements[i], null); + if (style.fontFamily.includes("Arial") || style.fontFamily.includes("Calibri") || + style.fontFamily.includes("Century") || style.fontFamily.includes("Times New Roman")) { + elements[i].style.fontFamily = "FontStandard, FontChineseSimplified, FontExtendedChineseSimplified, FontChineseTraditional, FontKorean, NintendoExt003"; + } else { + elements[i].style.fontFamily = style.fontFamily + ", FontStandard, FontChineseSimplified, FontExtendedChineseSimplified, FontChineseTraditional, FontKorean, NintendoExt003"; + } + } +})(); +)"; + +constexpr char GAMEPAD_SCRIPT[] = R"( +window.addEventListener("gamepadconnected", function(e) { + console.log("Gamepad connected at index %d: %s. %d buttons, %d axes.", + e.gamepad.index, e.gamepad.id, e.gamepad.buttons.length, e.gamepad.axes.length); +}); + +window.addEventListener("gamepaddisconnected", function(e) { + console.log("Gamepad disconnected from index %d: %s", e.gamepad.index, e.gamepad.id); +}); +)"; + +constexpr char WINDOW_NX_SCRIPT[] = R"( +var end_applet = false; +var yuzu_key_callbacks = []; + +(function() { + class WindowNX { + constructor() { + yuzu_key_callbacks[1] = function() { window.history.back(); }; + yuzu_key_callbacks[2] = function() { window.nx.endApplet(); }; + } + + addEventListener(type, listener, options) { + console.log("nx.addEventListener called, type=%s", type); + + window.addEventListener(type, listener, options); + } + + endApplet() { + console.log("nx.endApplet called"); + + end_applet = true; + } + + playSystemSe(system_se) { + console.log("nx.playSystemSe is not implemented, system_se=%s", system_se); + } + + sendMessage(message) { + console.log("nx.sendMessage is not implemented, message=%s", message); + } + + setCursorScrollSpeed(scroll_speed) { + console.log("nx.setCursorScrollSpeed is not implemented, scroll_speed=%d", scroll_speed); + } + } + + class WindowNXFooter { + setAssign(key, label, func, option) { + console.log("nx.footer.setAssign called, key=%s", key); + + switch (key) { + case "A": + yuzu_key_callbacks[0] = func; + break; + case "B": + yuzu_key_callbacks[1] = func; + break; + case "X": + yuzu_key_callbacks[2] = func; + break; + case "Y": + yuzu_key_callbacks[3] = func; + break; + case "L": + yuzu_key_callbacks[6] = func; + break; + case "R": + yuzu_key_callbacks[7] = func; + break; + } + } + + setFixed(kind) { + console.log("nx.footer.setFixed is not implemented, kind=%s", kind); + } + + unsetAssign(key) { + console.log("nx.footer.unsetAssign called, key=%s", key); + + switch (key) { + case "A": + yuzu_key_callbacks[0] = function() {}; + break; + case "B": + yuzu_key_callbacks[1] = function() {}; + break; + case "X": + yuzu_key_callbacks[2] = function() {}; + break; + case "Y": + yuzu_key_callbacks[3] = function() {}; + break; + case "L": + yuzu_key_callbacks[6] = function() {}; + break; + case "R": + yuzu_key_callbacks[7] = function() {}; + break; + } + } + } + + class WindowNXPlayReport { + incrementCounter(counter_id) { + console.log("nx.playReport.incrementCounter is not implemented, counter_id=%d", counter_id); + } + + setCounterSetIdentifier(counter_id) { + console.log("nx.playReport.setCounterSetIdentifier is not implemented, counter_id=%d", counter_id); + } + } + + window.nx = new WindowNX(); + window.nx.footer = new WindowNXFooter(); + window.nx.playReport = new WindowNXPlayReport(); +})(); +)"; diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp index 408eac2b7..85ee2577d 100644 --- a/src/yuzu/bootmanager.cpp +++ b/src/yuzu/bootmanager.cpp @@ -10,6 +10,7 @@ #include <QMessageBox> #include <QPainter> #include <QScreen> +#include <QString> #include <QStringList> #include <QWindow> @@ -18,7 +19,7 @@ #include <QOpenGLContext> #endif -#if !defined(WIN32) && HAS_VULKAN +#if !defined(WIN32) #include <qpa/qplatformnativeinterface.h> #endif @@ -34,7 +35,7 @@ #include "core/settings.h" #include "input_common/keyboard.h" #include "input_common/main.h" -#include "input_common/motion_emu.h" +#include "input_common/mouse/mouse_input.h" #include "video_core/renderer_base.h" #include "video_core/video_core.h" #include "yuzu/bootmanager.h" @@ -240,14 +241,12 @@ private: std::unique_ptr<Core::Frontend::GraphicsContext> context; }; -#ifdef HAS_VULKAN class VulkanRenderWidget : public RenderWidget { public: explicit VulkanRenderWidget(GRenderWindow* parent) : RenderWidget(parent) { windowHandle()->setSurfaceType(QWindow::VulkanSurface); } }; -#endif static Core::Frontend::WindowSystemType GetWindowSystemType() { // Determine WSI type based on Qt platform. @@ -267,7 +266,6 @@ static Core::Frontend::EmuWindow::WindowSystemInfo GetWindowSystemInfo(QWindow* Core::Frontend::EmuWindow::WindowSystemInfo wsi; wsi.type = GetWindowSystemType(); -#ifdef HAS_VULKAN // Our Win32 Qt external doesn't have the private API. #if defined(WIN32) || defined(__APPLE__) wsi.render_surface = window ? reinterpret_cast<void*>(window->winId()) : nullptr; @@ -280,7 +278,6 @@ static Core::Frontend::EmuWindow::WindowSystemInfo GetWindowSystemInfo(QWindow* wsi.render_surface = window ? reinterpret_cast<void*>(window->winId()) : nullptr; #endif wsi.render_surface_scale = window ? static_cast<float>(window->devicePixelRatio()) : 1.0f; -#endif return wsi; } @@ -301,13 +298,19 @@ GRenderWindow::GRenderWindow(GMainWindow* parent, EmuThread* emu_thread_, this->setMouseTracking(true); connect(this, &GRenderWindow::FirstFrameDisplayed, parent, &GMainWindow::OnLoadComplete); + connect(this, &GRenderWindow::ExecuteProgramSignal, parent, &GMainWindow::OnExecuteProgram, + Qt::QueuedConnection); +} + +void GRenderWindow::ExecuteProgram(std::size_t program_index) { + emit ExecuteProgramSignal(program_index); } GRenderWindow::~GRenderWindow() { input_subsystem->Shutdown(); } -void GRenderWindow::PollEvents() { +void GRenderWindow::OnFrameDisplayed() { if (!first_frame) { first_frame = true; emit FirstFrameDisplayed(); @@ -381,44 +384,46 @@ void GRenderWindow::keyReleaseEvent(QKeyEvent* event) { } void GRenderWindow::mousePressEvent(QMouseEvent* event) { - // touch input is handled in TouchBeginEvent + // Touch input is handled in TouchBeginEvent if (event->source() == Qt::MouseEventSynthesizedBySystem) { return; } auto pos = event->pos(); + const auto [x, y] = ScaleTouch(pos); + input_subsystem->GetMouse()->PressButton(x, y, event->button()); + if (event->button() == Qt::LeftButton) { - const auto [x, y] = ScaleTouch(pos); this->TouchPressed(x, y); - } else if (event->button() == Qt::RightButton) { - input_subsystem->GetMotionEmu()->BeginTilt(pos.x(), pos.y()); } - QWidget::mousePressEvent(event); + + emit MouseActivity(); } void GRenderWindow::mouseMoveEvent(QMouseEvent* event) { - // touch input is handled in TouchUpdateEvent + // Touch input is handled in TouchUpdateEvent if (event->source() == Qt::MouseEventSynthesizedBySystem) { return; } auto pos = event->pos(); const auto [x, y] = ScaleTouch(pos); + input_subsystem->GetMouse()->MouseMove(x, y); this->TouchMoved(x, y); - input_subsystem->GetMotionEmu()->Tilt(pos.x(), pos.y()); - QWidget::mouseMoveEvent(event); + + emit MouseActivity(); } void GRenderWindow::mouseReleaseEvent(QMouseEvent* event) { - // touch input is handled in TouchEndEvent + // Touch input is handled in TouchEndEvent if (event->source() == Qt::MouseEventSynthesizedBySystem) { return; } + input_subsystem->GetMouse()->ReleaseButton(event->button()); + if (event->button() == Qt::LeftButton) { this->TouchReleased(); - } else if (event->button() == Qt::RightButton) { - input_subsystem->GetMotionEmu()->EndTilt(); } } @@ -560,6 +565,10 @@ void GRenderWindow::CaptureScreenshot(u32 res_scale, const QString& screenshot_p layout); } +bool GRenderWindow::IsLoadingComplete() const { + return first_frame; +} + void GRenderWindow::OnMinimalClientAreaChangeRequest(std::pair<u32, u32> minimal_size) { setMinimumSize(minimal_size.first, minimal_size.second); } @@ -585,37 +594,45 @@ bool GRenderWindow::InitializeOpenGL() { } bool GRenderWindow::InitializeVulkan() { -#ifdef HAS_VULKAN auto child = new VulkanRenderWidget(this); child_widget = child; child_widget->windowHandle()->create(); main_context = std::make_unique<DummyContext>(); return true; -#else - QMessageBox::critical(this, tr("Vulkan not available!"), - tr("yuzu has not been compiled with Vulkan support.")); - return false; -#endif } bool GRenderWindow::LoadOpenGL() { auto context = CreateSharedContext(); auto scope = context->Acquire(); if (!gladLoadGL()) { - QMessageBox::critical(this, tr("Error while initializing OpenGL 4.3!"), - tr("Your GPU may not support OpenGL 4.3, or you do not have the " - "latest graphics driver.")); + QMessageBox::warning( + this, tr("Error while initializing OpenGL!"), + tr("Your GPU may not support OpenGL, or you do not have the latest graphics driver.")); + return false; + } + + const QString renderer = + QString::fromUtf8(reinterpret_cast<const char*>(glGetString(GL_RENDERER))); + + if (!GLAD_GL_VERSION_4_3) { + LOG_ERROR(Frontend, "GPU does not support OpenGL 4.3: {}", renderer.toStdString()); + QMessageBox::warning(this, tr("Error while initializing OpenGL 4.3!"), + tr("Your GPU may not support OpenGL 4.3, or you do not have the " + "latest graphics driver.<br><br>GL Renderer:<br>%1") + .arg(renderer)); return false; } QStringList unsupported_gl_extensions = GetUnsupportedGLExtensions(); if (!unsupported_gl_extensions.empty()) { - QMessageBox::critical( + QMessageBox::warning( this, tr("Error while initializing OpenGL!"), tr("Your GPU may not support one or more required OpenGL extensions. Please ensure you " - "have the latest graphics driver.<br><br>Unsupported extensions:<br>") + - unsupported_gl_extensions.join(QStringLiteral("<br>"))); + "have the latest graphics driver.<br><br>GL Renderer:<br>%1<br><br>Unsupported " + "extensions:<br>%2") + .arg(renderer) + .arg(unsupported_gl_extensions.join(QStringLiteral("<br>")))); return false; } return true; @@ -645,8 +662,13 @@ QStringList GRenderWindow::GetUnsupportedGLExtensions() const { if (!GLAD_GL_ARB_depth_buffer_float) unsupported_ext.append(QStringLiteral("ARB_depth_buffer_float")); - for (const QString& ext : unsupported_ext) - LOG_CRITICAL(Frontend, "Unsupported GL extension: {}", ext.toStdString()); + if (!unsupported_ext.empty()) { + LOG_ERROR(Frontend, "GPU does not support all required extensions: {}", + glGetString(GL_RENDERER)); + } + for (const QString& ext : unsupported_ext) { + LOG_ERROR(Frontend, "Unsupported GL extension: {}", ext.toStdString()); + } return unsupported_ext; } @@ -666,3 +688,10 @@ void GRenderWindow::showEvent(QShowEvent* event) { connect(windowHandle(), &QWindow::screenChanged, this, &GRenderWindow::OnFramebufferSizeChanged, Qt::UniqueConnection); } + +bool GRenderWindow::eventFilter(QObject* object, QEvent* event) { + if (event->type() == QEvent::HoverMove) { + emit MouseActivity(); + } + return false; +} diff --git a/src/yuzu/bootmanager.h b/src/yuzu/bootmanager.h index ca35cf831..339095509 100644 --- a/src/yuzu/bootmanager.h +++ b/src/yuzu/bootmanager.h @@ -131,7 +131,7 @@ public: ~GRenderWindow() override; // EmuWindow implementation. - void PollEvents() override; + void OnFrameDisplayed() override; bool IsShown() const override; std::unique_ptr<Core::Frontend::GraphicsContext> CreateSharedContext() const override; @@ -162,10 +162,18 @@ public: /// Destroy the previous run's child_widget which should also destroy the child_window void ReleaseRenderTarget(); + bool IsLoadingComplete() const; + void CaptureScreenshot(u32 res_scale, const QString& screenshot_path); std::pair<u32, u32> ScaleTouch(const QPointF& pos) const; + /** + * Instructs the window to re-launch the application using the specified program_index. + * @param program_index Specifies the index within the application of the program to launch. + */ + void ExecuteProgram(std::size_t program_index); + public slots: void OnEmulationStarting(EmuThread* emu_thread); void OnEmulationStopping(); @@ -175,6 +183,8 @@ signals: /// Emitted when the window is closed void Closed(); void FirstFrameDisplayed(); + void ExecuteProgramSignal(std::size_t program_index); + void MouseActivity(); private: void TouchBeginEvent(const QTouchEvent* event); @@ -207,4 +217,5 @@ private: protected: void showEvent(QShowEvent* event) override; + bool eventFilter(QObject* object, QEvent* event) override; }; diff --git a/src/yuzu/compatdb.cpp b/src/yuzu/compatdb.cpp index 649912557..a470056ef 100644 --- a/src/yuzu/compatdb.cpp +++ b/src/yuzu/compatdb.cpp @@ -72,7 +72,7 @@ void CompatDB::Submit() { void CompatDB::OnTestcaseSubmitted() { if (!testcase_watcher.result()) { QMessageBox::critical(this, tr("Communication error"), - tr("An error occured while sending the Testcase")); + tr("An error occurred while sending the Testcase")); button(NextButton)->setEnabled(true); button(NextButton)->setText(tr("Next")); button(CancelButton)->setVisible(true); diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp index d2913d613..cda448718 100644 --- a/src/yuzu/configuration/config.cpp +++ b/src/yuzu/configuration/config.cpp @@ -5,7 +5,9 @@ #include <array> #include <QKeySequence> #include <QSettings> +#include "common/common_paths.h" #include "common/file_util.h" +#include "core/core.h" #include "core/hle/service/acc/profile_manager.h" #include "core/hle/service/hid/controllers/npad.h" #include "input_common/main.h" @@ -14,14 +16,10 @@ namespace FS = Common::FS; -Config::Config(const std::string& config_file, bool is_global) { - // TODO: Don't hardcode the path; let the frontend decide where to put the config files. - qt_config_loc = FS::GetUserPath(FS::UserPath::ConfigDir) + config_file; - FS::CreateFullPath(qt_config_loc); - qt_config = - std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), QSettings::IniFormat); - global = is_global; - Reload(); +Config::Config(const std::string& config_name, ConfigType config_type) : type(config_type) { + global = config_type == ConfigType::GlobalConfig; + + Initialize(config_name); } Config::~Config() { @@ -242,84 +240,152 @@ const std::array<UISettings::Shortcut, 16> Config::default_hotkeys{{ }}; // clang-format on -void Config::ReadPlayerValues() { - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { - auto& player = Settings::values.players[p]; +void Config::Initialize(const std::string& config_name) { + switch (type) { + case ConfigType::GlobalConfig: + qt_config_loc = fmt::format("{}" DIR_SEP "{}.ini", FS::GetUserPath(FS::UserPath::ConfigDir), + config_name); + FS::CreateFullPath(qt_config_loc); + qt_config = std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), + QSettings::IniFormat); + Reload(); + break; + case ConfigType::PerGameConfig: + qt_config_loc = fmt::format("{}custom" DIR_SEP "{}.ini", + FS::GetUserPath(FS::UserPath::ConfigDir), config_name); + FS::CreateFullPath(qt_config_loc); + qt_config = std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), + QSettings::IniFormat); + Reload(); + break; + case ConfigType::InputProfile: + qt_config_loc = fmt::format("{}input" DIR_SEP "{}.ini", + FS::GetUserPath(FS::UserPath::ConfigDir), config_name); + FS::CreateFullPath(qt_config_loc); + qt_config = std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), + QSettings::IniFormat); + break; + } +} + +void Config::ReadPlayerValue(std::size_t player_index) { + const QString player_prefix = [this, player_index] { + if (type == ConfigType::InputProfile) { + return QString{}; + } else { + return QStringLiteral("player_%1_").arg(player_index); + } + }(); + + auto& player = Settings::values.players.GetValue()[player_index]; + if (player_prefix.isEmpty()) { + const auto controller = static_cast<Settings::ControllerType>( + qt_config + ->value(QStringLiteral("%1type").arg(player_prefix), + static_cast<u8>(Settings::ControllerType::ProController)) + .toUInt()); + + if (controller == Settings::ControllerType::LeftJoycon || + controller == Settings::ControllerType::RightJoycon) { + player.controller_type = controller; + } + } else { player.connected = - ReadSetting(QStringLiteral("player_%1_connected").arg(p), false).toBool(); + ReadSetting(QStringLiteral("%1connected").arg(player_prefix), player_index == 0) + .toBool(); player.controller_type = static_cast<Settings::ControllerType>( qt_config - ->value(QStringLiteral("player_%1_type").arg(p), + ->value(QStringLiteral("%1type").arg(player_prefix), static_cast<u8>(Settings::ControllerType::ProController)) .toUInt()); + player.vibration_enabled = + qt_config->value(QStringLiteral("%1vibration_enabled").arg(player_prefix), true) + .toBool(); + + player.vibration_strength = + qt_config->value(QStringLiteral("%1vibration_strength").arg(player_prefix), 100) + .toInt(); + player.body_color_left = qt_config - ->value(QStringLiteral("player_%1_body_color_left").arg(p), + ->value(QStringLiteral("%1body_color_left").arg(player_prefix), Settings::JOYCON_BODY_NEON_BLUE) .toUInt(); - player.body_color_right = qt_config - ->value(QStringLiteral("player_%1_body_color_right").arg(p), - Settings::JOYCON_BODY_NEON_RED) - .toUInt(); - player.button_color_left = qt_config - ->value(QStringLiteral("player_%1_button_color_left").arg(p), - Settings::JOYCON_BUTTONS_NEON_BLUE) - .toUInt(); + player.body_color_right = + qt_config + ->value(QStringLiteral("%1body_color_right").arg(player_prefix), + Settings::JOYCON_BODY_NEON_RED) + .toUInt(); + player.button_color_left = + qt_config + ->value(QStringLiteral("%1button_color_left").arg(player_prefix), + Settings::JOYCON_BUTTONS_NEON_BLUE) + .toUInt(); player.button_color_right = qt_config - ->value(QStringLiteral("player_%1_button_color_right").arg(p), + ->value(QStringLiteral("%1button_color_right").arg(player_prefix), Settings::JOYCON_BUTTONS_NEON_RED) .toUInt(); + } - for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { - const std::string default_param = - InputCommon::GenerateKeyboardParam(default_buttons[i]); - auto& player_buttons = player.buttons[i]; - - player_buttons = qt_config - ->value(QStringLiteral("player_%1_").arg(p) + - QString::fromUtf8(Settings::NativeButton::mapping[i]), - QString::fromStdString(default_param)) - .toString() - .toStdString(); - if (player_buttons.empty()) { - player_buttons = default_param; - } + for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_buttons[i]); + auto& player_buttons = player.buttons[i]; + + player_buttons = qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeButton::mapping[i]), + QString::fromStdString(default_param)) + .toString() + .toStdString(); + if (player_buttons.empty()) { + player_buttons = default_param; } + } - for (int i = 0; i < Settings::NativeMotion::NumMotions; ++i) { - const std::string default_param = - InputCommon::GenerateKeyboardParam(default_motions[i]); - auto& player_motions = player.motions[i]; - - player_motions = qt_config - ->value(QStringLiteral("player_%1_").arg(p) + - QString::fromUtf8(Settings::NativeMotion::mapping[i]), - QString::fromStdString(default_param)) - .toString() - .toStdString(); - if (player_motions.empty()) { - player_motions = default_param; - } + for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { + const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( + default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], + default_analogs[i][3], default_stick_mod[i], 0.5f); + auto& player_analogs = player.analogs[i]; + + player_analogs = qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeAnalog::mapping[i]), + QString::fromStdString(default_param)) + .toString() + .toStdString(); + if (player_analogs.empty()) { + player_analogs = default_param; } + } - for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { - const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( - default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], - default_analogs[i][3], default_stick_mod[i], 0.5f); - auto& player_analogs = player.analogs[i]; - - player_analogs = qt_config - ->value(QStringLiteral("player_%1_").arg(p) + - QString::fromUtf8(Settings::NativeAnalog::mapping[i]), - QString::fromStdString(default_param)) - .toString() - .toStdString(); - if (player_analogs.empty()) { - player_analogs = default_param; - } + for (int i = 0; i < Settings::NativeVibration::NumVibrations; ++i) { + auto& player_vibrations = player.vibrations[i]; + + player_vibrations = + qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeVibration::mapping[i]), + QString{}) + .toString() + .toStdString(); + } + + for (int i = 0; i < Settings::NativeMotion::NumMotions; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_motions[i]); + auto& player_motions = player.motions[i]; + + player_motions = qt_config + ->value(QStringLiteral("%1").arg(player_prefix) + + QString::fromUtf8(Settings::NativeMotion::mapping[i]), + QString::fromStdString(default_param)) + .toString() + .toStdString(); + if (player_motions.empty()) { + player_motions = default_param; } } } @@ -436,18 +502,24 @@ void Config::ReadAudioValues() { void Config::ReadControlValues() { qt_config->beginGroup(QStringLiteral("Controls")); - ReadPlayerValues(); + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { + ReadPlayerValue(p); + } ReadDebugValues(); ReadKeyboardValues(); ReadMouseValues(); ReadTouchscreenValues(); ReadMotionTouchValues(); - Settings::values.vibration_enabled = - ReadSetting(QStringLiteral("vibration_enabled"), true).toBool(); - Settings::values.motion_enabled = ReadSetting(QStringLiteral("motion_enabled"), true).toBool(); - Settings::values.use_docked_mode = - ReadSetting(QStringLiteral("use_docked_mode"), false).toBool(); + Settings::values.emulate_analog_keyboard = + ReadSetting(QStringLiteral("emulate_analog_keyboard"), false).toBool(); + + ReadSettingGlobal(Settings::values.use_docked_mode, QStringLiteral("use_docked_mode"), true); + ReadSettingGlobal(Settings::values.vibration_enabled, QStringLiteral("vibration_enabled"), + true); + ReadSettingGlobal(Settings::values.enable_accurate_vibrations, + QStringLiteral("enable_accurate_vibrations"), false); + ReadSettingGlobal(Settings::values.motion_enabled, QStringLiteral("motion_enabled"), true); qt_config->endGroup(); } @@ -500,22 +572,17 @@ void Config::ReadMotionTouchValues() { ReadSetting(QStringLiteral("touch_from_button_map"), 0).toInt(); Settings::values.touch_from_button_map_index = std::clamp(Settings::values.touch_from_button_map_index, 0, num_touch_from_button_maps - 1); - Settings::values.udp_input_address = - ReadSetting(QStringLiteral("udp_input_address"), - QString::fromUtf8(InputCommon::CemuhookUDP::DEFAULT_ADDR)) + Settings::values.udp_input_servers = + ReadSetting(QStringLiteral("udp_input_servers"), + QString::fromUtf8(InputCommon::CemuhookUDP::DEFAULT_SRV)) .toString() .toStdString(); - Settings::values.udp_input_port = static_cast<u16>( - ReadSetting(QStringLiteral("udp_input_port"), InputCommon::CemuhookUDP::DEFAULT_PORT) - .toInt()); - Settings::values.udp_pad_index = - static_cast<u8>(ReadSetting(QStringLiteral("udp_pad_index"), 0).toUInt()); } void Config::ReadCoreValues() { qt_config->beginGroup(QStringLiteral("Core")); - ReadSettingGlobal(Settings::values.use_multi_core, QStringLiteral("use_multi_core"), false); + ReadSettingGlobal(Settings::values.use_multi_core, QStringLiteral("use_multi_core"), true); qt_config->endGroup(); } @@ -570,8 +637,6 @@ void Config::ReadDebuggingValues() { // Intentionally not using the QT default setting as this is intended to be changed in the ini Settings::values.record_frame_times = qt_config->value(QStringLiteral("record_frame_times"), false).toBool(); - Settings::values.use_gdbstub = ReadSetting(QStringLiteral("use_gdbstub"), false).toBool(); - Settings::values.gdbstub_port = ReadSetting(QStringLiteral("gdbstub_port"), 24689).toInt(); Settings::values.program_args = ReadSetting(QStringLiteral("program_args"), QString{}).toString().toStdString(); Settings::values.dump_exefs = ReadSetting(QStringLiteral("dump_exefs"), false).toBool(); @@ -581,6 +646,8 @@ void Config::ReadDebuggingValues() { Settings::values.quest_flag = ReadSetting(QStringLiteral("quest_flag"), false).toBool(); Settings::values.disable_macro_jit = ReadSetting(QStringLiteral("disable_macro_jit"), false).toBool(); + Settings::values.extended_logging = + ReadSetting(QStringLiteral("extended_logging"), false).toBool(); qt_config->endGroup(); } @@ -697,6 +764,8 @@ void Config::ReadCpuValues() { ReadSetting(QStringLiteral("cpuopt_unsafe_unfuse_fma"), true).toBool(); Settings::values.cpuopt_unsafe_reduce_fp_error = ReadSetting(QStringLiteral("cpuopt_unsafe_reduce_fp_error"), true).toBool(); + Settings::values.cpuopt_unsafe_inaccurate_nan = + ReadSetting(QStringLiteral("cpuopt_unsafe_inaccurate_nan"), true).toBool(); } qt_config->endGroup(); @@ -716,10 +785,12 @@ void Config::ReadRendererValues() { QStringLiteral("use_disk_shader_cache"), true); ReadSettingGlobal(Settings::values.gpu_accuracy, QStringLiteral("gpu_accuracy"), 0); ReadSettingGlobal(Settings::values.use_asynchronous_gpu_emulation, - QStringLiteral("use_asynchronous_gpu_emulation"), false); + QStringLiteral("use_asynchronous_gpu_emulation"), true); + ReadSettingGlobal(Settings::values.use_nvdec_emulation, QStringLiteral("use_nvdec_emulation"), + true); ReadSettingGlobal(Settings::values.use_vsync, QStringLiteral("use_vsync"), true); ReadSettingGlobal(Settings::values.use_assembly_shaders, QStringLiteral("use_assembly_shaders"), - false); + true); ReadSettingGlobal(Settings::values.use_asynchronous_shaders, QStringLiteral("use_asynchronous_shaders"), false); ReadSettingGlobal(Settings::values.use_fast_gpu_time, QStringLiteral("use_fast_gpu_time"), @@ -918,49 +989,64 @@ void Config::ReadValues() { ReadSystemValues(); } -void Config::SavePlayerValues() { - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { - const auto& player = Settings::values.players[p]; +void Config::SavePlayerValue(std::size_t player_index) { + const QString player_prefix = [this, player_index] { + if (type == ConfigType::InputProfile) { + return QString{}; + } else { + return QStringLiteral("player_%1_").arg(player_index); + } + }(); + + const auto& player = Settings::values.players.GetValue()[player_index]; - WriteSetting(QStringLiteral("player_%1_connected").arg(p), player.connected, false); - WriteSetting(QStringLiteral("player_%1_type").arg(p), - static_cast<u8>(player.controller_type), - static_cast<u8>(Settings::ControllerType::ProController)); + WriteSetting(QStringLiteral("%1type").arg(player_prefix), + static_cast<u8>(player.controller_type), + static_cast<u8>(Settings::ControllerType::ProController)); - WriteSetting(QStringLiteral("player_%1_body_color_left").arg(p), player.body_color_left, + if (!player_prefix.isEmpty()) { + WriteSetting(QStringLiteral("%1connected").arg(player_prefix), player.connected, false); + WriteSetting(QStringLiteral("%1vibration_enabled").arg(player_prefix), + player.vibration_enabled, true); + WriteSetting(QStringLiteral("%1vibration_strength").arg(player_prefix), + player.vibration_strength, 100); + WriteSetting(QStringLiteral("%1body_color_left").arg(player_prefix), player.body_color_left, Settings::JOYCON_BODY_NEON_BLUE); - WriteSetting(QStringLiteral("player_%1_body_color_right").arg(p), player.body_color_right, - Settings::JOYCON_BODY_NEON_RED); - WriteSetting(QStringLiteral("player_%1_button_color_left").arg(p), player.button_color_left, - Settings::JOYCON_BUTTONS_NEON_BLUE); - WriteSetting(QStringLiteral("player_%1_button_color_right").arg(p), + WriteSetting(QStringLiteral("%1body_color_right").arg(player_prefix), + player.body_color_right, Settings::JOYCON_BODY_NEON_RED); + WriteSetting(QStringLiteral("%1button_color_left").arg(player_prefix), + player.button_color_left, Settings::JOYCON_BUTTONS_NEON_BLUE); + WriteSetting(QStringLiteral("%1button_color_right").arg(player_prefix), player.button_color_right, Settings::JOYCON_BUTTONS_NEON_RED); + } - for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { - const std::string default_param = - InputCommon::GenerateKeyboardParam(default_buttons[i]); - WriteSetting(QStringLiteral("player_%1_").arg(p) + - QString::fromStdString(Settings::NativeButton::mapping[i]), - QString::fromStdString(player.buttons[i]), - QString::fromStdString(default_param)); - } - for (int i = 0; i < Settings::NativeMotion::NumMotions; ++i) { - const std::string default_param = - InputCommon::GenerateKeyboardParam(default_motions[i]); - WriteSetting(QStringLiteral("player_%1_").arg(p) + - QString::fromStdString(Settings::NativeMotion::mapping[i]), - QString::fromStdString(player.motions[i]), - QString::fromStdString(default_param)); - } - for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { - const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( - default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], - default_analogs[i][3], default_stick_mod[i], 0.5f); - WriteSetting(QStringLiteral("player_%1_").arg(p) + - QString::fromStdString(Settings::NativeAnalog::mapping[i]), - QString::fromStdString(player.analogs[i]), - QString::fromStdString(default_param)); - } + for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_buttons[i]); + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeButton::mapping[i]), + QString::fromStdString(player.buttons[i]), + QString::fromStdString(default_param)); + } + for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { + const std::string default_param = InputCommon::GenerateAnalogParamFromKeys( + default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], + default_analogs[i][3], default_stick_mod[i], 0.5f); + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeAnalog::mapping[i]), + QString::fromStdString(player.analogs[i]), + QString::fromStdString(default_param)); + } + for (int i = 0; i < Settings::NativeVibration::NumVibrations; ++i) { + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeVibration::mapping[i]), + QString::fromStdString(player.vibrations[i]), QString{}); + } + for (int i = 0; i < Settings::NativeMotion::NumMotions; ++i) { + const std::string default_param = InputCommon::GenerateKeyboardParam(default_motions[i]); + WriteSetting(QStringLiteral("%1").arg(player_prefix) + + QString::fromStdString(Settings::NativeMotion::mapping[i]), + QString::fromStdString(player.motions[i]), + QString::fromStdString(default_param)); } } @@ -1021,12 +1107,9 @@ void Config::SaveMotionTouchValues() { false); WriteSetting(QStringLiteral("touch_from_button_map"), Settings::values.touch_from_button_map_index, 0); - WriteSetting(QStringLiteral("udp_input_address"), - QString::fromStdString(Settings::values.udp_input_address), - QString::fromUtf8(InputCommon::CemuhookUDP::DEFAULT_ADDR)); - WriteSetting(QStringLiteral("udp_input_port"), Settings::values.udp_input_port, - InputCommon::CemuhookUDP::DEFAULT_PORT); - WriteSetting(QStringLiteral("udp_pad_index"), Settings::values.udp_pad_index, 0); + WriteSetting(QStringLiteral("udp_input_servers"), + QString::fromStdString(Settings::values.udp_input_servers), + QString::fromUtf8(InputCommon::CemuhookUDP::DEFAULT_SRV)); qt_config->beginWriteArray(QStringLiteral("touch_from_button_maps")); for (std::size_t p = 0; p < Settings::values.touch_from_button_maps.size(); ++p) { @@ -1085,14 +1168,20 @@ void Config::SaveAudioValues() { void Config::SaveControlValues() { qt_config->beginGroup(QStringLiteral("Controls")); - SavePlayerValues(); + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { + SavePlayerValue(p); + } SaveDebugValues(); SaveMouseValues(); SaveTouchscreenValues(); SaveMotionTouchValues(); - WriteSetting(QStringLiteral("vibration_enabled"), Settings::values.vibration_enabled, true); - WriteSetting(QStringLiteral("motion_enabled"), Settings::values.motion_enabled, true); + WriteSettingGlobal(QStringLiteral("use_docked_mode"), Settings::values.use_docked_mode, true); + WriteSettingGlobal(QStringLiteral("vibration_enabled"), Settings::values.vibration_enabled, + true); + WriteSettingGlobal(QStringLiteral("enable_accurate_vibrations"), + Settings::values.enable_accurate_vibrations, false); + WriteSettingGlobal(QStringLiteral("motion_enabled"), Settings::values.motion_enabled, true); WriteSetting(QStringLiteral("motion_device"), QString::fromStdString(Settings::values.motion_device), QStringLiteral("engine:motion_emu,update_period:100,sensitivity:0.01")); @@ -1100,7 +1189,8 @@ void Config::SaveControlValues() { QString::fromStdString(Settings::values.touch_device), QStringLiteral("engine:emu_window")); WriteSetting(QStringLiteral("keyboard_enabled"), Settings::values.keyboard_enabled, false); - WriteSetting(QStringLiteral("use_docked_mode"), Settings::values.use_docked_mode, false); + WriteSetting(QStringLiteral("emulate_analog_keyboard"), + Settings::values.emulate_analog_keyboard, false); qt_config->endGroup(); } @@ -1108,7 +1198,7 @@ void Config::SaveControlValues() { void Config::SaveCoreValues() { qt_config->beginGroup(QStringLiteral("Core")); - WriteSettingGlobal(QStringLiteral("use_multi_core"), Settings::values.use_multi_core, false); + WriteSettingGlobal(QStringLiteral("use_multi_core"), Settings::values.use_multi_core, true); qt_config->endGroup(); } @@ -1146,8 +1236,6 @@ void Config::SaveDebuggingValues() { // Intentionally not using the QT default setting as this is intended to be changed in the ini qt_config->setValue(QStringLiteral("record_frame_times"), Settings::values.record_frame_times); - WriteSetting(QStringLiteral("use_gdbstub"), Settings::values.use_gdbstub, false); - WriteSetting(QStringLiteral("gdbstub_port"), Settings::values.gdbstub_port, 24689); WriteSetting(QStringLiteral("program_args"), QString::fromStdString(Settings::values.program_args), QString{}); WriteSetting(QStringLiteral("dump_exefs"), Settings::values.dump_exefs, false); @@ -1241,6 +1329,8 @@ void Config::SaveCpuValues() { Settings::values.cpuopt_unsafe_unfuse_fma, true); WriteSetting(QStringLiteral("cpuopt_unsafe_reduce_fp_error"), Settings::values.cpuopt_unsafe_reduce_fp_error, true); + WriteSetting(QStringLiteral("cpuopt_unsafe_inaccurate_nan"), + Settings::values.cpuopt_unsafe_inaccurate_nan, true); } qt_config->endGroup(); @@ -1264,10 +1354,12 @@ void Config::SaveRendererValues() { static_cast<int>(Settings::values.gpu_accuracy.GetValue(global)), Settings::values.gpu_accuracy.UsingGlobal(), 0); WriteSettingGlobal(QStringLiteral("use_asynchronous_gpu_emulation"), - Settings::values.use_asynchronous_gpu_emulation, false); + Settings::values.use_asynchronous_gpu_emulation, true); + WriteSettingGlobal(QStringLiteral("use_nvdec_emulation"), Settings::values.use_nvdec_emulation, + true); WriteSettingGlobal(QStringLiteral("use_vsync"), Settings::values.use_vsync, true); WriteSettingGlobal(QStringLiteral("use_assembly_shaders"), - Settings::values.use_assembly_shaders, false); + Settings::values.use_assembly_shaders, true); WriteSettingGlobal(QStringLiteral("use_asynchronous_shaders"), Settings::values.use_asynchronous_shaders, false); WriteSettingGlobal(QStringLiteral("use_fast_gpu_time"), Settings::values.use_fast_gpu_time, @@ -1501,13 +1593,27 @@ void Config::WriteSettingGlobal(const QString& name, const QVariant& value, bool void Config::Reload() { ReadValues(); - Settings::Sanitize(); // To apply default value changes SaveValues(); - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); } void Config::Save() { - Settings::Sanitize(); SaveValues(); } + +void Config::ReadControlPlayerValue(std::size_t player_index) { + qt_config->beginGroup(QStringLiteral("Controls")); + ReadPlayerValue(player_index); + qt_config->endGroup(); +} + +void Config::SaveControlPlayerValue(std::size_t player_index) { + qt_config->beginGroup(QStringLiteral("Controls")); + SavePlayerValue(player_index); + qt_config->endGroup(); +} + +const std::string& Config::GetConfigFilePath() const { + return qt_config_loc; +} diff --git a/src/yuzu/configuration/config.h b/src/yuzu/configuration/config.h index 5d8e45d78..8a600e19d 100644 --- a/src/yuzu/configuration/config.h +++ b/src/yuzu/configuration/config.h @@ -16,12 +16,24 @@ class QSettings; class Config { public: - explicit Config(const std::string& config_loc = "qt-config.ini", bool is_global = true); + enum class ConfigType { + GlobalConfig, + PerGameConfig, + InputProfile, + }; + + explicit Config(const std::string& config_name = "qt-config", + ConfigType config_type = ConfigType::GlobalConfig); ~Config(); void Reload(); void Save(); + void ReadControlPlayerValue(std::size_t player_index); + void SaveControlPlayerValue(std::size_t player_index); + + const std::string& GetConfigFilePath() const; + static const std::array<int, Settings::NativeButton::NumButtons> default_buttons; static const std::array<int, Settings::NativeMotion::NumMotions> default_motions; static const std::array<std::array<int, 4>, Settings::NativeAnalog::NumAnalogs> default_analogs; @@ -33,8 +45,10 @@ public: static const std::array<UISettings::Shortcut, 16> default_hotkeys; private: + void Initialize(const std::string& config_name); + void ReadValues(); - void ReadPlayerValues(); + void ReadPlayerValue(std::size_t player_index); void ReadDebugValues(); void ReadKeyboardValues(); void ReadMouseValues(); @@ -62,7 +76,7 @@ private: void ReadWebServiceValues(); void SaveValues(); - void SavePlayerValues(); + void SavePlayerValue(std::size_t player_index); void SaveDebugValues(); void SaveMouseValues(); void SaveTouchscreenValues(); @@ -111,9 +125,9 @@ private: void WriteSettingGlobal(const QString& name, const QVariant& value, bool use_global, const QVariant& default_value); + ConfigType type; std::unique_ptr<QSettings> qt_config; std::string qt_config_loc; - bool global; }; diff --git a/src/yuzu/configuration/configure.ui b/src/yuzu/configuration/configure.ui index fcf42cdcb..f92c3aff3 100644 --- a/src/yuzu/configuration/configure.ui +++ b/src/yuzu/configuration/configure.ui @@ -275,32 +275,12 @@ <signal>accepted()</signal> <receiver>ConfigureDialog</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>220</x> - <y>380</y> - </hint> - <hint type="destinationlabel"> - <x>220</x> - <y>200</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigureDialog</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>220</x> - <y>380</y> - </hint> - <hint type="destinationlabel"> - <x>220</x> - <y>200</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_audio.cpp b/src/yuzu/configuration/configure_audio.cpp index fa9124ecf..db9518798 100644 --- a/src/yuzu/configuration/configure_audio.cpp +++ b/src/yuzu/configuration/configure_audio.cpp @@ -25,8 +25,8 @@ ConfigureAudio::ConfigureAudio(QWidget* parent) connect(ui->output_sink_combo_box, qOverload<int>(&QComboBox::currentIndexChanged), this, &ConfigureAudio::UpdateAudioDevices); - ui->volume_label->setVisible(Settings::configuring_global); - ui->volume_combo_box->setVisible(!Settings::configuring_global); + ui->volume_label->setVisible(Settings::IsConfiguringGlobal()); + ui->volume_combo_box->setVisible(!Settings::IsConfiguringGlobal()); SetupPerGameUI(); @@ -51,7 +51,7 @@ void ConfigureAudio::SetConfiguration() { ui->toggle_audio_stretching->setChecked(Settings::values.enable_audio_stretching.GetValue()); - if (!Settings::configuring_global) { + if (!Settings::IsConfiguringGlobal()) { if (Settings::values.volume.UsingGlobal()) { ui->volume_combo_box->setCurrentIndex(0); ui->volume_slider->setEnabled(false); @@ -99,7 +99,7 @@ void ConfigureAudio::SetVolumeIndicatorText(int percentage) { } void ConfigureAudio::ApplyConfiguration() { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { Settings::values.sink_id = ui->output_sink_combo_box->itemText(ui->output_sink_combo_box->currentIndex()) .toStdString(); @@ -165,7 +165,7 @@ void ConfigureAudio::RetranslateUI() { } void ConfigureAudio::SetupPerGameUI() { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->volume_slider->setEnabled(Settings::values.volume.UsingGlobal()); ui->toggle_audio_stretching->setEnabled( Settings::values.enable_audio_stretching.UsingGlobal()); diff --git a/src/yuzu/configuration/configure_cpu.cpp b/src/yuzu/configuration/configure_cpu.cpp index 37fcd6adc..d055cbd60 100644 --- a/src/yuzu/configuration/configure_cpu.cpp +++ b/src/yuzu/configuration/configure_cpu.cpp @@ -36,6 +36,8 @@ void ConfigureCpu::SetConfiguration() { ui->cpuopt_unsafe_unfuse_fma->setChecked(Settings::values.cpuopt_unsafe_unfuse_fma); ui->cpuopt_unsafe_reduce_fp_error->setEnabled(runtime_lock); ui->cpuopt_unsafe_reduce_fp_error->setChecked(Settings::values.cpuopt_unsafe_reduce_fp_error); + ui->cpuopt_unsafe_inaccurate_nan->setEnabled(runtime_lock); + ui->cpuopt_unsafe_inaccurate_nan->setChecked(Settings::values.cpuopt_unsafe_inaccurate_nan); } void ConfigureCpu::AccuracyUpdated(int index) { @@ -61,6 +63,7 @@ void ConfigureCpu::ApplyConfiguration() { static_cast<Settings::CPUAccuracy>(ui->accuracy->currentIndex()); Settings::values.cpuopt_unsafe_unfuse_fma = ui->cpuopt_unsafe_unfuse_fma->isChecked(); Settings::values.cpuopt_unsafe_reduce_fp_error = ui->cpuopt_unsafe_reduce_fp_error->isChecked(); + Settings::values.cpuopt_unsafe_inaccurate_nan = ui->cpuopt_unsafe_inaccurate_nan->isChecked(); } void ConfigureCpu::changeEvent(QEvent* event) { diff --git a/src/yuzu/configuration/configure_cpu.ui b/src/yuzu/configuration/configure_cpu.ui index ebdd2e6e9..bcd0962e9 100644 --- a/src/yuzu/configuration/configure_cpu.ui +++ b/src/yuzu/configuration/configure_cpu.ui @@ -109,6 +109,18 @@ </property> </widget> </item> + <item> + <widget class="QCheckBox" name="cpuopt_unsafe_inaccurate_nan"> + <property name="text"> + <string>Inaccurate NaN handling</string> + </property> + <property name="toolTip"> + <string> + <div>This option improves speed by removing NaN checking. Please note this also reduces accuracy of certain floating-point instructions.</div> + </string> + </property> + </widget> + </item> </layout> </widget> </item> diff --git a/src/yuzu/configuration/configure_debug.cpp b/src/yuzu/configuration/configure_debug.cpp index 2bfe2c306..121873f95 100644 --- a/src/yuzu/configuration/configure_debug.cpp +++ b/src/yuzu/configuration/configure_debug.cpp @@ -28,9 +28,6 @@ ConfigureDebug::ConfigureDebug(QWidget* parent) : QWidget(parent), ui(new Ui::Co ConfigureDebug::~ConfigureDebug() = default; void ConfigureDebug::SetConfiguration() { - ui->toggle_gdbstub->setChecked(Settings::values.use_gdbstub); - ui->gdbport_spinbox->setEnabled(Settings::values.use_gdbstub); - ui->gdbport_spinbox->setValue(Settings::values.gdbstub_port); ui->toggle_console->setEnabled(!Core::System::GetInstance().IsPoweredOn()); ui->toggle_console->setChecked(UISettings::values.show_console); ui->log_filter_edit->setText(QString::fromStdString(Settings::values.log_filter)); @@ -41,11 +38,10 @@ void ConfigureDebug::SetConfiguration() { ui->enable_graphics_debugging->setChecked(Settings::values.renderer_debug); ui->disable_macro_jit->setEnabled(!Core::System::GetInstance().IsPoweredOn()); ui->disable_macro_jit->setChecked(Settings::values.disable_macro_jit); + ui->extended_logging->setChecked(Settings::values.extended_logging); } void ConfigureDebug::ApplyConfiguration() { - Settings::values.use_gdbstub = ui->toggle_gdbstub->isChecked(); - Settings::values.gdbstub_port = ui->gdbport_spinbox->value(); UISettings::values.show_console = ui->toggle_console->isChecked(); Settings::values.log_filter = ui->log_filter_edit->text().toStdString(); Settings::values.program_args = ui->homebrew_args_edit->text().toStdString(); @@ -53,6 +49,7 @@ void ConfigureDebug::ApplyConfiguration() { Settings::values.quest_flag = ui->quest_flag->isChecked(); Settings::values.renderer_debug = ui->enable_graphics_debugging->isChecked(); Settings::values.disable_macro_jit = ui->disable_macro_jit->isChecked(); + Settings::values.extended_logging = ui->extended_logging->isChecked(); Debugger::ToggleConsole(); Log::Filter filter; filter.ParseFilterString(Settings::values.log_filter); diff --git a/src/yuzu/configuration/configure_debug.ui b/src/yuzu/configuration/configure_debug.ui index 9d6feb9f7..9186aa732 100644 --- a/src/yuzu/configuration/configure_debug.ui +++ b/src/yuzu/configuration/configure_debug.ui @@ -7,7 +7,7 @@ <x>0</x> <y>0</y> <width>400</width> - <height>467</height> + <height>486</height> </rect> </property> <property name="windowTitle"> @@ -15,57 +15,6 @@ </property> <layout class="QVBoxLayout" name="verticalLayout_1"> <item> - <layout class="QVBoxLayout" name="verticalLayout_2"> - <item> - <widget class="QGroupBox" name="groupBox"> - <property name="title"> - <string>GDB</string> - </property> - <layout class="QVBoxLayout" name="verticalLayout_3"> - <item> - <layout class="QHBoxLayout" name="horizontalLayout_1"> - <item> - <widget class="QCheckBox" name="toggle_gdbstub"> - <property name="text"> - <string>Enable GDB Stub</string> - </property> - </widget> - </item> - <item> - <spacer name="horizontalSpacer"> - <property name="orientation"> - <enum>Qt::Horizontal</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>40</width> - <height>20</height> - </size> - </property> - </spacer> - </item> - <item> - <widget class="QLabel" name="label_1"> - <property name="text"> - <string>Port:</string> - </property> - </widget> - </item> - <item> - <widget class="QSpinBox" name="gdbport_spinbox"> - <property name="maximum"> - <number>65536</number> - </property> - </widget> - </item> - </layout> - </item> - </layout> - </widget> - </item> - </layout> - </item> - <item> <widget class="QGroupBox" name="groupBox_2"> <property name="title"> <string>Logging</string> @@ -90,7 +39,7 @@ <item> <widget class="QCheckBox" name="toggle_console"> <property name="text"> - <string>Show Log Console (Windows Only)</string> + <string>Show Log in Console</string> </property> </widget> </item> @@ -103,6 +52,34 @@ </item> </layout> </item> + <item> + <widget class="QCheckBox" name="extended_logging"> + <property name="enabled"> + <bool>true</bool> + </property> + <property name="toolTip"> + <string>When checked, the max size of the log increases from 100 MB to 1 GB</string> + </property> + <property name="text"> + <string>Enable Extended Logging</string> + </property> + </widget> + </item> + <item> + <widget class="QLabel" name="label_3"> + <property name="font"> + <font> + <italic>true</italic> + </font> + </property> + <property name="text"> + <string>This will be reset automatically when yuzu closes.</string> + </property> + <property name="indent"> + <number>20</number> + </property> + </widget> + </item> </layout> </widget> </item> @@ -115,7 +92,7 @@ <item> <layout class="QHBoxLayout" name="horizontalLayout_4"> <item> - <widget class="QLabel" name="label_3"> + <widget class="QLabel" name="label_4"> <property name="text"> <string>Arguments String</string> </property> @@ -140,8 +117,8 @@ <property name="enabled"> <bool>true</bool> </property> - <property name="whatsThis"> - <string>When checked, the graphics API enters in a slower debugging mode</string> + <property name="toolTip"> + <string>When checked, the graphics API enters a slower debugging mode</string> </property> <property name="text"> <string>Enable Graphics Debugging</string> @@ -153,8 +130,8 @@ <property name="enabled"> <bool>true</bool> </property> - <property name="whatsThis"> - <string>When checked, it disables the macro Just In Time compiler. Enabled this makes games run slower</string> + <property name="toolTip"> + <string>When checked, it disables the macro Just In Time compiler. Enabling this makes games run slower</string> </property> <property name="text"> <string>Disable Macro JIT</string> @@ -169,7 +146,7 @@ <property name="title"> <string>Dump</string> </property> - <layout class="QVBoxLayout" name="verticalLayout_6"> + <layout class="QVBoxLayout" name="verticalLayout_7"> <item> <widget class="QCheckBox" name="reporting_services"> <property name="text"> @@ -178,7 +155,7 @@ </widget> </item> <item> - <widget class="QLabel" name="label"> + <widget class="QLabel" name="label_5"> <property name="font"> <font> <italic>true</italic> @@ -200,7 +177,7 @@ <property name="title"> <string>Advanced</string> </property> - <layout class="QVBoxLayout" name="verticalLayout_7"> + <layout class="QVBoxLayout" name="verticalLayout_8"> <item> <widget class="QCheckBox" name="quest_flag"> <property name="text"> @@ -230,8 +207,6 @@ </layout> </widget> <tabstops> - <tabstop>toggle_gdbstub</tabstop> - <tabstop>gdbport_spinbox</tabstop> <tabstop>log_filter_edit</tabstop> <tabstop>toggle_console</tabstop> <tabstop>open_log_button</tabstop> @@ -241,22 +216,5 @@ <tabstop>quest_flag</tabstop> </tabstops> <resources/> - <connections> - <connection> - <sender>toggle_gdbstub</sender> - <signal>toggled(bool)</signal> - <receiver>gdbport_spinbox</receiver> - <slot>setEnabled(bool)</slot> - <hints> - <hint type="sourcelabel"> - <x>84</x> - <y>157</y> - </hint> - <hint type="destinationlabel"> - <x>342</x> - <y>158</y> - </hint> - </hints> - </connection> - </connections> + <connections/> </ui> diff --git a/src/yuzu/configuration/configure_debug_controller.cpp b/src/yuzu/configuration/configure_debug_controller.cpp index 0097c9a29..a878ef9c6 100644 --- a/src/yuzu/configuration/configure_debug_controller.cpp +++ b/src/yuzu/configuration/configure_debug_controller.cpp @@ -4,11 +4,14 @@ #include "ui_configure_debug_controller.h" #include "yuzu/configuration/configure_debug_controller.h" +#include "yuzu/configuration/configure_input_player.h" ConfigureDebugController::ConfigureDebugController(QWidget* parent, - InputCommon::InputSubsystem* input_subsystem) + InputCommon::InputSubsystem* input_subsystem, + InputProfiles* profiles) : QDialog(parent), ui(std::make_unique<Ui::ConfigureDebugController>()), - debug_controller(new ConfigureInputPlayer(this, 9, nullptr, input_subsystem, true)) { + debug_controller( + new ConfigureInputPlayer(this, 9, nullptr, input_subsystem, profiles, true)) { ui->setupUi(this); ui->controllerLayout->addWidget(debug_controller); diff --git a/src/yuzu/configuration/configure_debug_controller.h b/src/yuzu/configuration/configure_debug_controller.h index 34dcf705f..b4f53fad5 100644 --- a/src/yuzu/configuration/configure_debug_controller.h +++ b/src/yuzu/configuration/configure_debug_controller.h @@ -6,10 +6,13 @@ #include <memory> #include <QDialog> -#include "yuzu/configuration/configure_input_player.h" class QPushButton; +class ConfigureInputPlayer; + +class InputProfiles; + namespace InputCommon { class InputSubsystem; } @@ -22,8 +25,8 @@ class ConfigureDebugController : public QDialog { Q_OBJECT public: - explicit ConfigureDebugController(QWidget* parent, - InputCommon::InputSubsystem* input_subsystem); + explicit ConfigureDebugController(QWidget* parent, InputCommon::InputSubsystem* input_subsystem, + InputProfiles* profiles); ~ConfigureDebugController() override; void ApplyConfiguration(); diff --git a/src/yuzu/configuration/configure_debug_controller.ui b/src/yuzu/configuration/configure_debug_controller.ui index a95ed50ff..7b7e6582c 100644 --- a/src/yuzu/configuration/configure_debug_controller.ui +++ b/src/yuzu/configuration/configure_debug_controller.ui @@ -66,32 +66,12 @@ <signal>accepted()</signal> <receiver>ConfigureDebugController</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>140</x> - <y>318</y> - </hint> - <hint type="destinationlabel"> - <x>140</x> - <y>169</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigureDebugController</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>140</x> - <y>318</y> - </hint> - <hint type="destinationlabel"> - <x>140</x> - <y>169</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_dialog.cpp b/src/yuzu/configuration/configure_dialog.cpp index 8186929a6..b33f8437a 100644 --- a/src/yuzu/configuration/configure_dialog.cpp +++ b/src/yuzu/configuration/configure_dialog.cpp @@ -5,6 +5,7 @@ #include <QHash> #include <QListWidgetItem> #include <QSignalBlocker> +#include "core/core.h" #include "core/settings.h" #include "ui_configure.h" #include "yuzu/configuration/config.h" @@ -15,7 +16,7 @@ ConfigureDialog::ConfigureDialog(QWidget* parent, HotkeyRegistry& registry, InputCommon::InputSubsystem* input_subsystem) : QDialog(parent), ui(new Ui::ConfigureDialog), registry(registry) { - Settings::configuring_global = true; + Settings::SetConfiguringGlobal(true); ui->setupUi(this); ui->hotkeysTab->Populate(registry); @@ -54,7 +55,7 @@ void ConfigureDialog::ApplyConfiguration() { ui->debugTab->ApplyConfiguration(); ui->webTab->ApplyConfiguration(); ui->serviceTab->ApplyConfiguration(); - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); Settings::LogSettings(); } diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp index 830096ea0..d4d29d422 100644 --- a/src/yuzu/configuration/configure_general.cpp +++ b/src/yuzu/configuration/configure_general.cpp @@ -19,7 +19,7 @@ ConfigureGeneral::ConfigureGeneral(QWidget* parent) SetConfiguration(); - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { connect(ui->toggle_frame_limit, &QCheckBox::clicked, ui->frame_limit, [this]() { ui->frame_limit->setEnabled(ui->toggle_frame_limit->isChecked()); }); } @@ -41,7 +41,7 @@ void ConfigureGeneral::SetConfiguration() { ui->toggle_frame_limit->setChecked(Settings::values.use_frame_limit.GetValue()); ui->frame_limit->setValue(Settings::values.frame_limit.GetValue()); - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->frame_limit->setEnabled(Settings::values.use_frame_limit.GetValue()); } else { ui->frame_limit->setEnabled(Settings::values.use_frame_limit.GetValue() && @@ -50,7 +50,7 @@ void ConfigureGeneral::SetConfiguration() { } void ConfigureGeneral::ApplyConfiguration() { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked(); UISettings::values.select_user_on_boot = ui->toggle_user_on_boot->isChecked(); UISettings::values.pause_when_in_background = ui->toggle_background_pause->isChecked(); @@ -93,7 +93,7 @@ void ConfigureGeneral::RetranslateUI() { } void ConfigureGeneral::SetupPerGameUI() { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->toggle_frame_limit->setEnabled(Settings::values.use_frame_limit.UsingGlobal()); ui->frame_limit->setEnabled(Settings::values.frame_limit.UsingGlobal()); diff --git a/src/yuzu/configuration/configure_graphics.cpp b/src/yuzu/configuration/configure_graphics.cpp index 07d818548..b78a5dff0 100644 --- a/src/yuzu/configuration/configure_graphics.cpp +++ b/src/yuzu/configuration/configure_graphics.cpp @@ -4,22 +4,17 @@ #include <QColorDialog> #include <QComboBox> -#ifdef HAS_VULKAN #include <QVulkanInstance> -#endif #include "common/common_types.h" #include "common/logging/log.h" #include "core/core.h" #include "core/settings.h" #include "ui_configure_graphics.h" +#include "video_core/renderer_vulkan/renderer_vulkan.h" #include "yuzu/configuration/configuration_shared.h" #include "yuzu/configuration/configure_graphics.h" -#ifdef HAS_VULKAN -#include "video_core/renderer_vulkan/renderer_vulkan.h" -#endif - ConfigureGraphics::ConfigureGraphics(QWidget* parent) : QWidget(parent), ui(new Ui::ConfigureGraphics) { vulkan_device = Settings::values.vulkan_device.GetValue(); @@ -33,7 +28,7 @@ ConfigureGraphics::ConfigureGraphics(QWidget* parent) connect(ui->api, qOverload<int>(&QComboBox::currentIndexChanged), this, [this] { UpdateDeviceComboBox(); - if (!Settings::configuring_global) { + if (!Settings::IsConfiguringGlobal()) { ConfigurationShared::SetHighlight( ui->api_layout, ui->api->currentIndex() != ConfigurationShared::USE_GLOBAL_INDEX); } @@ -49,8 +44,8 @@ ConfigureGraphics::ConfigureGraphics(QWidget* parent) UpdateBackgroundColorButton(new_bg_color); }); - ui->bg_label->setVisible(Settings::configuring_global); - ui->bg_combobox->setVisible(!Settings::configuring_global); + ui->bg_label->setVisible(Settings::IsConfiguringGlobal()); + ui->bg_combobox->setVisible(!Settings::IsConfiguringGlobal()); } void ConfigureGraphics::UpdateDeviceSelection(int device) { @@ -70,11 +65,13 @@ void ConfigureGraphics::SetConfiguration() { ui->api->setEnabled(runtime_lock); ui->use_asynchronous_gpu_emulation->setEnabled(runtime_lock); ui->use_disk_shader_cache->setEnabled(runtime_lock); + ui->use_nvdec_emulation->setEnabled(runtime_lock); ui->use_disk_shader_cache->setChecked(Settings::values.use_disk_shader_cache.GetValue()); ui->use_asynchronous_gpu_emulation->setChecked( Settings::values.use_asynchronous_gpu_emulation.GetValue()); + ui->use_nvdec_emulation->setChecked(Settings::values.use_nvdec_emulation.GetValue()); - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->api->setCurrentIndex(static_cast<int>(Settings::values.renderer_backend.GetValue())); ui->aspect_ratio_combobox->setCurrentIndex(Settings::values.aspect_ratio.GetValue()); } else { @@ -98,7 +95,7 @@ void ConfigureGraphics::SetConfiguration() { } void ConfigureGraphics::ApplyConfiguration() { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { // Guard if during game and set to game-specific value if (Settings::values.renderer_backend.UsingGlobal()) { Settings::values.renderer_backend.SetValue(GetCurrentGraphicsBackend()); @@ -116,6 +113,9 @@ void ConfigureGraphics::ApplyConfiguration() { Settings::values.use_asynchronous_gpu_emulation.SetValue( ui->use_asynchronous_gpu_emulation->isChecked()); } + if (Settings::values.use_nvdec_emulation.UsingGlobal()) { + Settings::values.use_nvdec_emulation.SetValue(ui->use_nvdec_emulation->isChecked()); + } if (Settings::values.bg_red.UsingGlobal()) { Settings::values.bg_red.SetValue(static_cast<float>(bg_color.redF())); Settings::values.bg_green.SetValue(static_cast<float>(bg_color.greenF())); @@ -144,6 +144,8 @@ void ConfigureGraphics::ApplyConfiguration() { ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_asynchronous_gpu_emulation, ui->use_asynchronous_gpu_emulation, use_asynchronous_gpu_emulation); + ConfigurationShared::ApplyPerGameSetting(&Settings::values.use_nvdec_emulation, + ui->use_nvdec_emulation, use_nvdec_emulation); if (ui->bg_combobox->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { Settings::values.bg_red.SetGlobal(true); @@ -187,7 +189,7 @@ void ConfigureGraphics::UpdateDeviceComboBox() { bool enabled = false; - if (!Settings::configuring_global && + if (!Settings::IsConfiguringGlobal() && ui->api->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX) { vulkan_device = Settings::values.vulkan_device.GetValue(); } @@ -205,22 +207,20 @@ void ConfigureGraphics::UpdateDeviceComboBox() { break; } // If in per-game config and use global is selected, don't enable. - enabled &= !(!Settings::configuring_global && + enabled &= !(!Settings::IsConfiguringGlobal() && ui->api->currentIndex() == ConfigurationShared::USE_GLOBAL_INDEX); ui->device->setEnabled(enabled && !Core::System::GetInstance().IsPoweredOn()); } void ConfigureGraphics::RetrieveVulkanDevices() { -#ifdef HAS_VULKAN vulkan_devices.clear(); - for (auto& name : Vulkan::RendererVulkan::EnumerateDevices()) { + for (const auto& name : Vulkan::RendererVulkan::EnumerateDevices()) { vulkan_devices.push_back(QString::fromStdString(name)); } -#endif } Settings::RendererBackend ConfigureGraphics::GetCurrentGraphicsBackend() const { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { return static_cast<Settings::RendererBackend>(ui->api->currentIndex()); } @@ -234,12 +234,13 @@ Settings::RendererBackend ConfigureGraphics::GetCurrentGraphicsBackend() const { } void ConfigureGraphics::SetupPerGameUI() { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->api->setEnabled(Settings::values.renderer_backend.UsingGlobal()); ui->device->setEnabled(Settings::values.renderer_backend.UsingGlobal()); ui->aspect_ratio_combobox->setEnabled(Settings::values.aspect_ratio.UsingGlobal()); ui->use_asynchronous_gpu_emulation->setEnabled( Settings::values.use_asynchronous_gpu_emulation.UsingGlobal()); + ui->use_nvdec_emulation->setEnabled(Settings::values.use_nvdec_emulation.UsingGlobal()); ui->use_disk_shader_cache->setEnabled(Settings::values.use_disk_shader_cache.UsingGlobal()); ui->bg_button->setEnabled(Settings::values.bg_red.UsingGlobal()); @@ -253,6 +254,8 @@ void ConfigureGraphics::SetupPerGameUI() { ConfigurationShared::SetColoredTristate( ui->use_disk_shader_cache, Settings::values.use_disk_shader_cache, use_disk_shader_cache); + ConfigurationShared::SetColoredTristate( + ui->use_nvdec_emulation, Settings::values.use_nvdec_emulation, use_nvdec_emulation); ConfigurationShared::SetColoredTristate(ui->use_asynchronous_gpu_emulation, Settings::values.use_asynchronous_gpu_emulation, use_asynchronous_gpu_emulation); diff --git a/src/yuzu/configuration/configure_graphics.h b/src/yuzu/configuration/configure_graphics.h index b4961f719..1fefc88eb 100644 --- a/src/yuzu/configuration/configure_graphics.h +++ b/src/yuzu/configuration/configure_graphics.h @@ -46,6 +46,7 @@ private: std::unique_ptr<Ui::ConfigureGraphics> ui; QColor bg_color; + ConfigurationShared::CheckState use_nvdec_emulation; ConfigurationShared::CheckState use_disk_shader_cache; ConfigurationShared::CheckState use_asynchronous_gpu_emulation; diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui index 62aa337e7..58486eb1e 100644 --- a/src/yuzu/configuration/configure_graphics.ui +++ b/src/yuzu/configuration/configure_graphics.ui @@ -98,6 +98,13 @@ </widget> </item> <item> + <widget class="QCheckBox" name="use_nvdec_emulation"> + <property name="text"> + <string>Use NVDEC emulation</string> + </property> + </widget> + </item> + <item> <widget class="QWidget" name="aspect_ratio_layout" native="true"> <layout class="QHBoxLayout" name="horizontalLayout_6"> <property name="leftMargin"> diff --git a/src/yuzu/configuration/configure_graphics_advanced.cpp b/src/yuzu/configuration/configure_graphics_advanced.cpp index 73f276949..383c7bac8 100644 --- a/src/yuzu/configuration/configure_graphics_advanced.cpp +++ b/src/yuzu/configuration/configure_graphics_advanced.cpp @@ -32,7 +32,7 @@ void ConfigureGraphicsAdvanced::SetConfiguration() { ui->use_asynchronous_shaders->setChecked(Settings::values.use_asynchronous_shaders.GetValue()); ui->use_fast_gpu_time->setChecked(Settings::values.use_fast_gpu_time.GetValue()); - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->gpu_accuracy->setCurrentIndex( static_cast<int>(Settings::values.gpu_accuracy.GetValue())); ui->anisotropic_filtering_combobox->setCurrentIndex( @@ -52,9 +52,9 @@ void ConfigureGraphicsAdvanced::ApplyConfiguration() { // Subtract 2 if configuring per-game (separator and "use global configuration" take 2 slots) const auto gpu_accuracy = static_cast<Settings::GPUAccuracy>( ui->gpu_accuracy->currentIndex() - - ((Settings::configuring_global) ? 0 : ConfigurationShared::USE_GLOBAL_OFFSET)); + ((Settings::IsConfiguringGlobal()) ? 0 : ConfigurationShared::USE_GLOBAL_OFFSET)); - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { // Must guard in case of a during-game configuration when set to be game-specific. if (Settings::values.gpu_accuracy.UsingGlobal()) { Settings::values.gpu_accuracy.SetValue(gpu_accuracy); @@ -118,7 +118,7 @@ void ConfigureGraphicsAdvanced::RetranslateUI() { void ConfigureGraphicsAdvanced::SetupPerGameUI() { // Disable if not global (only happens during game) - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->gpu_accuracy->setEnabled(Settings::values.gpu_accuracy.UsingGlobal()); ui->use_vsync->setEnabled(Settings::values.use_vsync.UsingGlobal()); ui->use_assembly_shaders->setEnabled(Settings::values.use_assembly_shaders.UsingGlobal()); diff --git a/src/yuzu/configuration/configure_input.cpp b/src/yuzu/configuration/configure_input.cpp index 2725fcb2b..567a36d9b 100644 --- a/src/yuzu/configuration/configure_input.cpp +++ b/src/yuzu/configuration/configure_input.cpp @@ -4,6 +4,7 @@ #include <algorithm> #include <memory> +#include <thread> #include <QSignalBlocker> #include <QTimer> @@ -23,6 +24,8 @@ #include "yuzu/configuration/configure_motion_touch.h" #include "yuzu/configuration/configure_mouse_advanced.h" #include "yuzu/configuration/configure_touchscreen_advanced.h" +#include "yuzu/configuration/configure_vibration.h" +#include "yuzu/configuration/input_profiles.h" namespace { template <typename Dialog, typename... Args> @@ -64,7 +67,8 @@ void OnDockedModeChanged(bool last_state, bool new_state) { } ConfigureInput::ConfigureInput(QWidget* parent) - : QWidget(parent), ui(std::make_unique<Ui::ConfigureInput>()) { + : QWidget(parent), ui(std::make_unique<Ui::ConfigureInput>()), + profiles(std::make_unique<InputProfiles>()) { ui->setupUi(this); } @@ -73,14 +77,22 @@ ConfigureInput::~ConfigureInput() = default; void ConfigureInput::Initialize(InputCommon::InputSubsystem* input_subsystem, std::size_t max_players) { player_controllers = { - new ConfigureInputPlayer(this, 0, ui->consoleInputSettings, input_subsystem), - new ConfigureInputPlayer(this, 1, ui->consoleInputSettings, input_subsystem), - new ConfigureInputPlayer(this, 2, ui->consoleInputSettings, input_subsystem), - new ConfigureInputPlayer(this, 3, ui->consoleInputSettings, input_subsystem), - new ConfigureInputPlayer(this, 4, ui->consoleInputSettings, input_subsystem), - new ConfigureInputPlayer(this, 5, ui->consoleInputSettings, input_subsystem), - new ConfigureInputPlayer(this, 6, ui->consoleInputSettings, input_subsystem), - new ConfigureInputPlayer(this, 7, ui->consoleInputSettings, input_subsystem), + new ConfigureInputPlayer(this, 0, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 1, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 2, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 3, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 4, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 5, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 6, ui->consoleInputSettings, input_subsystem, + profiles.get()), + new ConfigureInputPlayer(this, 7, ui->consoleInputSettings, input_subsystem, + profiles.get()), }; player_tabs = { @@ -113,8 +125,10 @@ void ConfigureInput::Initialize(InputCommon::InputSubsystem* input_subsystem, } } }); - connect(player_controllers[i], &ConfigureInputPlayer::RefreshInputDevices, - [this] { UpdateAllInputDevices(); }); + connect(player_controllers[i], &ConfigureInputPlayer::RefreshInputDevices, this, + &ConfigureInput::UpdateAllInputDevices); + connect(player_controllers[i], &ConfigureInputPlayer::RefreshInputProfiles, this, + &ConfigureInput::UpdateAllInputProfiles, Qt::QueuedConnection); connect(player_connected[i], &QCheckBox::stateChanged, [this, i](int state) { player_controllers[i]->ConnectPlayer(state == Qt::Checked); }); @@ -134,7 +148,7 @@ void ConfigureInput::Initialize(InputCommon::InputSubsystem* input_subsystem, ui->tabAdvanced->setLayout(new QHBoxLayout(ui->tabAdvanced)); ui->tabAdvanced->layout()->addWidget(advanced); connect(advanced, &ConfigureInputAdvanced::CallDebugControllerDialog, [this, input_subsystem] { - CallConfigureDialog<ConfigureDebugController>(*this, input_subsystem); + CallConfigureDialog<ConfigureDebugController>(*this, input_subsystem, profiles.get()); }); connect(advanced, &ConfigureInputAdvanced::CallMouseConfigDialog, [this, input_subsystem] { CallConfigureDialog<ConfigureMouseAdvanced>(*this, input_subsystem); @@ -146,6 +160,9 @@ void ConfigureInput::Initialize(InputCommon::InputSubsystem* input_subsystem, CallConfigureDialog<ConfigureMotionTouch>(*this, input_subsystem); }); + connect(ui->vibrationButton, &QPushButton::clicked, + [this] { CallConfigureDialog<ConfigureVibration>(*this); }); + connect(ui->motionButton, &QPushButton::clicked, [this, input_subsystem] { CallConfigureDialog<ConfigureMotionTouch>(*this, input_subsystem); }); @@ -165,18 +182,28 @@ QList<QWidget*> ConfigureInput::GetSubTabs() const { } void ConfigureInput::ApplyConfiguration() { - for (auto controller : player_controllers) { + for (auto* controller : player_controllers) { controller->ApplyConfiguration(); + controller->TryDisconnectSelectedController(); + } + + // This emulates a delay between disconnecting and reconnecting controllers as some games + // do not respond to a change in controller type if it was instantaneous. + using namespace std::chrono_literals; + std::this_thread::sleep_for(60ms); + + for (auto* controller : player_controllers) { + controller->TryConnectSelectedController(); } advanced->ApplyConfiguration(); - const bool pre_docked_mode = Settings::values.use_docked_mode; - Settings::values.use_docked_mode = ui->radioDocked->isChecked(); - OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode); + const bool pre_docked_mode = Settings::values.use_docked_mode.GetValue(); + Settings::values.use_docked_mode.SetValue(ui->radioDocked->isChecked()); + OnDockedModeChanged(pre_docked_mode, Settings::values.use_docked_mode.GetValue()); - Settings::values.vibration_enabled = ui->vibrationGroup->isChecked(); - Settings::values.motion_enabled = ui->motionGroup->isChecked(); + Settings::values.vibration_enabled.SetValue(ui->vibrationGroup->isChecked()); + Settings::values.motion_enabled.SetValue(ui->motionGroup->isChecked()); } void ConfigureInput::changeEvent(QEvent* event) { @@ -193,16 +220,16 @@ void ConfigureInput::RetranslateUI() { void ConfigureInput::LoadConfiguration() { LoadPlayerControllerIndices(); - UpdateDockedState(Settings::values.players[8].connected); + UpdateDockedState(Settings::values.players.GetValue()[8].connected); - ui->vibrationGroup->setChecked(Settings::values.vibration_enabled); - ui->motionGroup->setChecked(Settings::values.motion_enabled); + ui->vibrationGroup->setChecked(Settings::values.vibration_enabled.GetValue()); + ui->motionGroup->setChecked(Settings::values.motion_enabled.GetValue()); } void ConfigureInput::LoadPlayerControllerIndices() { for (std::size_t i = 0; i < player_connected.size(); ++i) { - const auto connected = Settings::values.players[i].connected || - (i == 0 && Settings::values.players[8].connected); + const auto connected = Settings::values.players.GetValue()[i].connected || + (i == 0 && Settings::values.players.GetValue()[8].connected); player_connected[i]->setChecked(connected); } } @@ -231,8 +258,8 @@ void ConfigureInput::UpdateDockedState(bool is_handheld) { ui->radioDocked->setEnabled(!is_handheld); ui->radioUndocked->setEnabled(!is_handheld); - ui->radioDocked->setChecked(Settings::values.use_docked_mode); - ui->radioUndocked->setChecked(!Settings::values.use_docked_mode); + ui->radioDocked->setChecked(Settings::values.use_docked_mode.GetValue()); + ui->radioUndocked->setChecked(!Settings::values.use_docked_mode.GetValue()); // Also force into undocked mode if the controller type is handheld. if (is_handheld) { @@ -242,6 +269,16 @@ void ConfigureInput::UpdateDockedState(bool is_handheld) { void ConfigureInput::UpdateAllInputDevices() { for (const auto& player : player_controllers) { - player->UpdateInputDevices(); + player->UpdateInputDeviceCombobox(); + } +} + +void ConfigureInput::UpdateAllInputProfiles(std::size_t player_index) { + for (std::size_t i = 0; i < player_controllers.size(); ++i) { + if (i == player_index) { + continue; + } + + player_controllers[i]->UpdateInputProfiles(); } } diff --git a/src/yuzu/configuration/configure_input.h b/src/yuzu/configuration/configure_input.h index 0e8b2fd4e..f4eb0d78b 100644 --- a/src/yuzu/configuration/configure_input.h +++ b/src/yuzu/configuration/configure_input.h @@ -8,17 +8,18 @@ #include <memory> #include <QKeyEvent> +#include <QList> #include <QWidget> -#include "yuzu/configuration/configure_input_advanced.h" -#include "yuzu/configuration/configure_input_player.h" - -#include "ui_configure_input.h" - class QCheckBox; class QString; class QTimer; +class ConfigureInputAdvanced; +class ConfigureInputPlayer; + +class InputProfiles; + namespace InputCommon { class InputSubsystem; } @@ -51,6 +52,7 @@ private: void UpdateDockedState(bool is_handheld); void UpdateAllInputDevices(); + void UpdateAllInputProfiles(std::size_t player_index); /// Load configuration settings. void LoadConfiguration(); @@ -61,6 +63,8 @@ private: std::unique_ptr<Ui::ConfigureInput> ui; + std::unique_ptr<InputProfiles> profiles; + std::array<ConfigureInputPlayer*, 8> player_controllers; std::array<QWidget*, 8> player_tabs; std::array<QCheckBox*, 8> player_connected; diff --git a/src/yuzu/configuration/configure_input.ui b/src/yuzu/configuration/configure_input.ui index 136955224..2707025e7 100644 --- a/src/yuzu/configuration/configure_input.ui +++ b/src/yuzu/configuration/configure_input.ui @@ -6,7 +6,7 @@ <rect> <x>0</x> <y>0</y> - <width>700</width> + <width>680</width> <height>540</height> </rect> </property> @@ -142,7 +142,7 @@ <number>6</number> </property> <property name="leftMargin"> - <number>3</number> + <number>8</number> </property> <property name="topMargin"> <number>6</number> @@ -195,30 +195,24 @@ <number>3</number> </property> <item> - <widget class="QSpinBox" name="vibrationSpin"> + <widget class="QPushButton" name="vibrationButton"> <property name="minimumSize"> <size> - <width>65</width> - <height>21</height> + <width>68</width> + <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>65</width> + <width>68</width> <height>16777215</height> </size> </property> - <property name="suffix"> - <string>%</string> - </property> - <property name="minimum"> - <number>1</number> - </property> - <property name="maximum"> - <number>200</number> + <property name="styleSheet"> + <string notr="true">min-width: 68px;</string> </property> - <property name="value"> - <number>100</number> + <property name="text"> + <string>Configure</string> </property> </widget> </item> @@ -250,18 +244,18 @@ <widget class="QPushButton" name="motionButton"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Configure</string> @@ -272,7 +266,7 @@ </widget> </item> <item alignment="Qt::AlignVCenter"> - <widget class="QWidget" name="widget" native="true"> + <widget class="QWidget" name="connectedControllers" native="true"> <layout class="QGridLayout" name="gridLayout_2"> <property name="leftMargin"> <number>5</number> @@ -468,13 +462,13 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -494,7 +488,7 @@ <enum>Qt::LeftToRight</enum> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Defaults</string> @@ -511,13 +505,13 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -537,7 +531,7 @@ <enum>Qt::LeftToRight</enum> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Clear</string> diff --git a/src/yuzu/configuration/configure_input_advanced.cpp b/src/yuzu/configuration/configure_input_advanced.cpp index 81f9dc16c..4e557bc6f 100644 --- a/src/yuzu/configuration/configure_input_advanced.cpp +++ b/src/yuzu/configuration/configure_input_advanced.cpp @@ -68,8 +68,7 @@ ConfigureInputAdvanced::ConfigureInputAdvanced(QWidget* parent) for (std::size_t button_idx = 0; button_idx < color_buttons.size(); ++button_idx) { connect(color_buttons[button_idx], &QPushButton::clicked, this, [this, player_idx, button_idx] { - OnControllerButtonClick(static_cast<int>(player_idx), - static_cast<int>(button_idx)); + OnControllerButtonClick(player_idx, button_idx); }); } } @@ -94,20 +93,21 @@ ConfigureInputAdvanced::ConfigureInputAdvanced(QWidget* parent) ConfigureInputAdvanced::~ConfigureInputAdvanced() = default; -void ConfigureInputAdvanced::OnControllerButtonClick(int player_idx, int button_idx) { +void ConfigureInputAdvanced::OnControllerButtonClick(std::size_t player_idx, + std::size_t button_idx) { const QColor new_bg_color = QColorDialog::getColor(controllers_colors[player_idx][button_idx]); if (!new_bg_color.isValid()) { return; } controllers_colors[player_idx][button_idx] = new_bg_color; controllers_color_buttons[player_idx][button_idx]->setStyleSheet( - QStringLiteral("background-color: %1; min-width: 55px;") + QStringLiteral("background-color: %1; min-width: 60px;") .arg(controllers_colors[player_idx][button_idx].name())); } void ConfigureInputAdvanced::ApplyConfiguration() { for (std::size_t player_idx = 0; player_idx < controllers_color_buttons.size(); ++player_idx) { - auto& player = Settings::values.players[player_idx]; + auto& player = Settings::values.players.GetValue()[player_idx]; std::array<u32, 4> colors{}; std::transform(controllers_colors[player_idx].begin(), controllers_colors[player_idx].end(), colors.begin(), [](QColor color) { return color.rgb(); }); @@ -121,12 +121,13 @@ void ConfigureInputAdvanced::ApplyConfiguration() { Settings::values.debug_pad_enabled = ui->debug_enabled->isChecked(); Settings::values.mouse_enabled = ui->mouse_enabled->isChecked(); Settings::values.keyboard_enabled = ui->keyboard_enabled->isChecked(); + Settings::values.emulate_analog_keyboard = ui->emulate_analog_keyboard->isChecked(); Settings::values.touchscreen.enabled = ui->touchscreen_enabled->isChecked(); } void ConfigureInputAdvanced::LoadConfiguration() { for (std::size_t player_idx = 0; player_idx < controllers_color_buttons.size(); ++player_idx) { - auto& player = Settings::values.players[player_idx]; + auto& player = Settings::values.players.GetValue()[player_idx]; std::array<u32, 4> colors = { player.body_color_left, player.button_color_left, @@ -139,7 +140,7 @@ void ConfigureInputAdvanced::LoadConfiguration() { for (std::size_t button_idx = 0; button_idx < colors.size(); ++button_idx) { controllers_color_buttons[player_idx][button_idx]->setStyleSheet( - QStringLiteral("background-color: %1; min-width: 55px;") + QStringLiteral("background-color: %1; min-width: 60px;") .arg(controllers_colors[player_idx][button_idx].name())); } } @@ -147,6 +148,7 @@ void ConfigureInputAdvanced::LoadConfiguration() { ui->debug_enabled->setChecked(Settings::values.debug_pad_enabled); ui->mouse_enabled->setChecked(Settings::values.mouse_enabled); ui->keyboard_enabled->setChecked(Settings::values.keyboard_enabled); + ui->emulate_analog_keyboard->setChecked(Settings::values.emulate_analog_keyboard); ui->touchscreen_enabled->setChecked(Settings::values.touchscreen.enabled); UpdateUIEnabled(); diff --git a/src/yuzu/configuration/configure_input_advanced.h b/src/yuzu/configuration/configure_input_advanced.h index 50bb87768..3083d55c1 100644 --- a/src/yuzu/configuration/configure_input_advanced.h +++ b/src/yuzu/configuration/configure_input_advanced.h @@ -35,7 +35,7 @@ private: void RetranslateUI(); void UpdateUIEnabled(); - void OnControllerButtonClick(int player_idx, int button_idx); + void OnControllerButtonClick(std::size_t player_idx, std::size_t button_idx); void LoadConfiguration(); diff --git a/src/yuzu/configuration/configure_input_advanced.ui b/src/yuzu/configuration/configure_input_advanced.ui index 5958435fc..f207e5d3b 100644 --- a/src/yuzu/configuration/configure_input_advanced.ui +++ b/src/yuzu/configuration/configure_input_advanced.ui @@ -192,18 +192,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -247,18 +247,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -323,18 +323,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -378,18 +378,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -478,18 +478,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -533,18 +533,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -609,18 +609,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -664,18 +664,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -782,18 +782,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -837,18 +837,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -913,18 +913,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -968,18 +968,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1068,18 +1068,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1123,18 +1123,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1199,18 +1199,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1254,18 +1254,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1393,18 +1393,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1448,18 +1448,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1524,18 +1524,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1579,18 +1579,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1679,18 +1679,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1734,18 +1734,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1810,18 +1810,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1865,18 +1865,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -1983,18 +1983,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2038,18 +2038,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2114,18 +2114,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2169,18 +2169,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2269,18 +2269,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2324,18 +2324,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2400,18 +2400,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2455,18 +2455,18 @@ </property> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string/> @@ -2546,14 +2546,27 @@ </property> </widget> </item> - <item row="4" column="2"> + <item row="1" column="0"> + <widget class="QCheckBox" name="emulate_analog_keyboard"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>23</height> + </size> + </property> + <property name="text"> + <string>Emulate Analog with Keyboard Input</string> + </property> + </widget> + </item> + <item row="5" column="2"> <widget class="QPushButton" name="touchscreen_advanced"> <property name="text"> <string>Advanced</string> </property> </widget> </item> - <item row="1" column="1"> + <item row="2" column="1"> <spacer name="horizontalSpacer_8"> <property name="orientation"> <enum>Qt::Horizontal</enum> @@ -2569,21 +2582,21 @@ </property> </spacer> </item> - <item row="1" column="2"> + <item row="2" column="2"> <widget class="QPushButton" name="mouse_advanced"> <property name="text"> <string>Advanced</string> </property> </widget> </item> - <item row="4" column="0"> + <item row="5" column="0"> <widget class="QCheckBox" name="touchscreen_enabled"> <property name="text"> <string>Touchscreen</string> </property> </widget> </item> - <item row="1" column="0"> + <item row="2" column="0"> <widget class="QCheckBox" name="mouse_enabled"> <property name="minimumSize"> <size> @@ -2596,28 +2609,28 @@ </property> </widget> </item> - <item row="6" column="0"> + <item row="7" column="0"> <widget class="QLabel" name="motion_touch"> <property name="text"> <string>Motion / Touch</string> </property> </widget> </item> - <item row="6" column="2"> + <item row="7" column="2"> <widget class="QPushButton" name="buttonMotionTouch"> <property name="text"> <string>Configure</string> </property> </widget> </item> - <item row="5" column="0"> + <item row="6" column="0"> <widget class="QCheckBox" name="debug_enabled"> <property name="text"> <string>Debug Controller</string> </property> </widget> </item> - <item row="5" column="2"> + <item row="6" column="2"> <widget class="QPushButton" name="debug_configure"> <property name="text"> <string>Configure</string> diff --git a/src/yuzu/configuration/configure_input_dialog.cpp b/src/yuzu/configuration/configure_input_dialog.cpp deleted file mode 100644 index 1866003c2..000000000 --- a/src/yuzu/configuration/configure_input_dialog.cpp +++ /dev/null @@ -1,37 +0,0 @@ -// Copyright 2020 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include "ui_configure_input_dialog.h" -#include "yuzu/configuration/configure_input_dialog.h" - -ConfigureInputDialog::ConfigureInputDialog(QWidget* parent, std::size_t max_players, - InputCommon::InputSubsystem* input_subsystem) - : QDialog(parent), ui(std::make_unique<Ui::ConfigureInputDialog>()), - input_widget(new ConfigureInput(this)) { - ui->setupUi(this); - - input_widget->Initialize(input_subsystem, max_players); - - ui->inputLayout->addWidget(input_widget); - - RetranslateUI(); -} - -ConfigureInputDialog::~ConfigureInputDialog() = default; - -void ConfigureInputDialog::ApplyConfiguration() { - input_widget->ApplyConfiguration(); -} - -void ConfigureInputDialog::changeEvent(QEvent* event) { - if (event->type() == QEvent::LanguageChange) { - RetranslateUI(); - } - - QDialog::changeEvent(event); -} - -void ConfigureInputDialog::RetranslateUI() { - ui->retranslateUi(this); -} diff --git a/src/yuzu/configuration/configure_input_dialog.h b/src/yuzu/configuration/configure_input_dialog.h deleted file mode 100644 index d1bd865f9..000000000 --- a/src/yuzu/configuration/configure_input_dialog.h +++ /dev/null @@ -1,38 +0,0 @@ -// Copyright 2020 yuzu Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <memory> -#include <QDialog> -#include "yuzu/configuration/configure_input.h" - -class QPushButton; - -namespace InputCommon { -class InputSubsystem; -} - -namespace Ui { -class ConfigureInputDialog; -} - -class ConfigureInputDialog : public QDialog { - Q_OBJECT - -public: - explicit ConfigureInputDialog(QWidget* parent, std::size_t max_players, - InputCommon::InputSubsystem* input_subsystem); - ~ConfigureInputDialog() override; - - void ApplyConfiguration(); - -private: - void changeEvent(QEvent* event) override; - void RetranslateUI(); - - std::unique_ptr<Ui::ConfigureInputDialog> ui; - - ConfigureInput* input_widget; -}; diff --git a/src/yuzu/configuration/configure_input_player.cpp b/src/yuzu/configuration/configure_input_player.cpp index 698cb1940..46ea026e4 100644 --- a/src/yuzu/configuration/configure_input_player.cpp +++ b/src/yuzu/configuration/configure_input_player.cpp @@ -18,12 +18,16 @@ #include "core/hle/service/sm/sm.h" #include "input_common/gcadapter/gc_poller.h" #include "input_common/main.h" +#include "input_common/mouse/mouse_poller.h" #include "input_common/udp/udp.h" #include "ui_configure_input_player.h" #include "yuzu/configuration/config.h" #include "yuzu/configuration/configure_input_player.h" +#include "yuzu/configuration/configure_vibration.h" +#include "yuzu/configuration/input_profiles.h" +#include "yuzu/util/limitable_input_dialog.h" -constexpr std::size_t HANDHELD_INDEX = 8; +using namespace Service::HID; const std::array<std::string, ConfigureInputPlayer::ANALOG_SUB_BUTTONS_NUM> ConfigureInputPlayer::analog_sub_buttons{{ @@ -35,6 +39,8 @@ const std::array<std::string, ConfigureInputPlayer::ANALOG_SUB_BUTTONS_NUM> namespace { +constexpr std::size_t HANDHELD_INDEX = 8; + void UpdateController(Settings::ControllerType controller_type, std::size_t npad_index, bool connected) { Core::System& system{Core::System::GetInstance()}; @@ -43,48 +49,12 @@ void UpdateController(Settings::ControllerType controller_type, std::size_t npad } Service::SM::ServiceManager& sm = system.ServiceManager(); - auto& npad = - sm.GetService<Service::HID::Hid>("hid") - ->GetAppletResource() - ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad); + auto& npad = sm.GetService<Hid>("hid")->GetAppletResource()->GetController<Controller_NPad>( + HidController::NPad); npad.UpdateControllerAt(npad.MapSettingsTypeToNPad(controller_type), npad_index, connected); } -/// Maps the controller type combobox index to Controller Type enum -constexpr Settings::ControllerType GetControllerTypeFromIndex(int index) { - switch (index) { - case 0: - default: - return Settings::ControllerType::ProController; - case 1: - return Settings::ControllerType::DualJoyconDetached; - case 2: - return Settings::ControllerType::LeftJoycon; - case 3: - return Settings::ControllerType::RightJoycon; - case 4: - return Settings::ControllerType::Handheld; - } -} - -/// Maps the Controller Type enum to controller type combobox index -constexpr int GetIndexFromControllerType(Settings::ControllerType type) { - switch (type) { - case Settings::ControllerType::ProController: - default: - return 0; - case Settings::ControllerType::DualJoyconDetached: - return 1; - case Settings::ControllerType::LeftJoycon: - return 2; - case Settings::ControllerType::RightJoycon: - return 3; - case Settings::ControllerType::Handheld: - return 4; - } -} - QString GetKeyName(int key_code) { switch (key_code) { case Qt::LeftButton: @@ -182,6 +152,14 @@ QString ButtonToText(const Common::ParamPackage& param) { return {}; } + if (param.Get("engine", "") == "mouse") { + if (param.Has("button")) { + const QString button_str = QString::number(int(param.Get("button", 0))); + return QObject::tr("Click %1").arg(button_str); + } + return GetKeyName(param.Get("code", 0)); + } + return QObject::tr("[unknown]"); } @@ -194,41 +172,31 @@ QString AnalogToText(const Common::ParamPackage& param, const std::string& dir) return ButtonToText(Common::ParamPackage{param.Get(dir, "")}); } - if (param.Get("engine", "") == "sdl") { + const auto engine_str = param.Get("engine", ""); + const QString axis_x_str = QString::fromStdString(param.Get("axis_x", "")); + const QString axis_y_str = QString::fromStdString(param.Get("axis_y", "")); + const bool invert_x = param.Get("invert_x", "+") == "-"; + const bool invert_y = param.Get("invert_y", "+") == "-"; + if (engine_str == "sdl" || engine_str == "gcpad" || engine_str == "mouse") { if (dir == "modifier") { return QObject::tr("[unused]"); } - if (dir == "left" || dir == "right") { - const QString axis_x_str = QString::fromStdString(param.Get("axis_x", "")); - - return QObject::tr("Axis %1").arg(axis_x_str); - } - - if (dir == "up" || dir == "down") { - const QString axis_y_str = QString::fromStdString(param.Get("axis_y", "")); - - return QObject::tr("Axis %1").arg(axis_y_str); + if (dir == "left") { + const QString invert_x_str = QString::fromStdString(invert_x ? "+" : "-"); + return QObject::tr("Axis %1%2").arg(axis_x_str, invert_x_str); } - - return {}; - } - - if (param.Get("engine", "") == "gcpad") { - if (dir == "modifier") { - return QObject::tr("[unused]"); + if (dir == "right") { + const QString invert_x_str = QString::fromStdString(invert_x ? "-" : "+"); + return QObject::tr("Axis %1%2").arg(axis_x_str, invert_x_str); } - - if (dir == "left" || dir == "right") { - const QString axis_x_str = QString::fromStdString(param.Get("axis_x", "")); - - return QObject::tr("GC Axis %1").arg(axis_x_str); + if (dir == "up") { + const QString invert_y_str = QString::fromStdString(invert_y ? "-" : "+"); + return QObject::tr("Axis %1%2").arg(axis_y_str, invert_y_str); } - - if (dir == "up" || dir == "down") { - const QString axis_y_str = QString::fromStdString(param.Get("axis_y", "")); - - return QObject::tr("GC Axis %1").arg(axis_y_str); + if (dir == "down") { + const QString invert_y_str = QString::fromStdString(invert_y ? "+" : "-"); + return QObject::tr("Axis %1%2").arg(axis_y_str, invert_y_str); } return {}; @@ -240,10 +208,11 @@ QString AnalogToText(const Common::ParamPackage& param, const std::string& dir) ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_index, QWidget* bottom_row, InputCommon::InputSubsystem* input_subsystem_, - bool debug) + InputProfiles* profiles_, bool debug) : QWidget(parent), ui(std::make_unique<Ui::ConfigureInputPlayer>()), player_index(player_index), - debug(debug), input_subsystem{input_subsystem_}, timeout_timer(std::make_unique<QTimer>()), - poll_timer(std::make_unique<QTimer>()), bottom_row(bottom_row) { + debug(debug), input_subsystem{input_subsystem_}, profiles(profiles_), + timeout_timer(std::make_unique<QTimer>()), poll_timer(std::make_unique<QTimer>()), + bottom_row(bottom_row) { ui->setupUi(this); setFocusPolicy(Qt::ClickFocus); @@ -256,11 +225,6 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i ui->buttonSL, ui->buttonSR, ui->buttonHome, ui->buttonScreenshot, }; - mod_buttons = { - ui->buttonLStickMod, - ui->buttonRStickMod, - }; - analog_map_buttons = {{ { ui->buttonLStickUp, @@ -284,6 +248,7 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i analog_map_deadzone_label = {ui->labelLStickDeadzone, ui->labelRStickDeadzone}; analog_map_deadzone_slider = {ui->sliderLStickDeadzone, ui->sliderRStickDeadzone}; analog_map_modifier_groupbox = {ui->buttonLStickModGroup, ui->buttonRStickModGroup}; + analog_map_modifier_button = {ui->buttonLStickMod, ui->buttonRStickMod}; analog_map_modifier_label = {ui->labelLStickModifierRange, ui->labelRStickModifierRange}; analog_map_modifier_slider = {ui->sliderLStickModifierRange, ui->sliderRStickModifierRange}; analog_map_range_groupbox = {ui->buttonLStickRangeGroup, ui->buttonRStickRangeGroup}; @@ -370,6 +335,18 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i } connect(analog_button, &QPushButton::clicked, [=, this] { + if (!map_analog_stick_accepted) { + map_analog_stick_accepted = + QMessageBox::information( + this, tr("Map Analog Stick"), + tr("After pressing OK, first move your joystick horizontally, and then " + "vertically.\nTo invert the axes, first move your joystick " + "vertically, and then horizontally."), + QMessageBox::Ok | QMessageBox::Cancel) == QMessageBox::Ok; + if (!map_analog_stick_accepted) { + return; + } + } HandleClick( analog_map_buttons[analog_id][sub_button_id], [=, this](const Common::ParamPackage& params) { @@ -388,26 +365,51 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i analogs_param[analog_id].Clear(); analog_map_buttons[analog_id][sub_button_id]->setText(tr("[not set]")); }); + context_menu.addAction(tr("Invert axis"), [&] { + if (sub_button_id == 2 || sub_button_id == 3) { + const bool invert_value = + analogs_param[analog_id].Get("invert_x", "+") == "-"; + const std::string invert_str = invert_value ? "+" : "-"; + analogs_param[analog_id].Set("invert_x", invert_str); + } + if (sub_button_id == 0 || sub_button_id == 1) { + const bool invert_value = + analogs_param[analog_id].Get("invert_y", "+") == "-"; + const std::string invert_str = invert_value ? "+" : "-"; + analogs_param[analog_id].Set("invert_y", invert_str); + } + for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; + ++sub_button_id) { + analog_map_buttons[analog_id][sub_button_id]->setText(AnalogToText( + analogs_param[analog_id], analog_sub_buttons[sub_button_id])); + } + }); context_menu.exec(analog_map_buttons[analog_id][sub_button_id]->mapToGlobal( menu_location)); }); } // Handle clicks for the modifier buttons as well. - ConfigureButtonClick(mod_buttons[analog_id], &stick_mod_param[analog_id], - Config::default_stick_mod[analog_id], - InputCommon::Polling::DeviceType::Button); + connect(analog_map_modifier_button[analog_id], &QPushButton::clicked, [=, this] { + HandleClick( + analog_map_modifier_button[analog_id], + [=, this](const Common::ParamPackage& params) { + analogs_param[analog_id].Set("modifier", params.Serialize()); + }, + InputCommon::Polling::DeviceType::Button); + }); - mod_buttons[analog_id]->setContextMenuPolicy(Qt::CustomContextMenu); + analog_map_modifier_button[analog_id]->setContextMenuPolicy(Qt::CustomContextMenu); - connect(mod_buttons[analog_id], &QPushButton::customContextMenuRequested, + connect(analog_map_modifier_button[analog_id], &QPushButton::customContextMenuRequested, [=, this](const QPoint& menu_location) { QMenu context_menu; context_menu.addAction(tr("Clear"), [&] { - stick_mod_param[analog_id].Clear(); - mod_buttons[analog_id]->setText(tr("[not set]")); + analogs_param[analog_id].Set("modifier", ""); + analog_map_modifier_button[analog_id]->setText(tr("[not set]")); }); - context_menu.exec(mod_buttons[analog_id]->mapToGlobal(menu_location)); + context_menu.exec( + analog_map_modifier_button[analog_id]->mapToGlobal(menu_location)); }); connect(analog_map_range_spinbox[analog_id], qOverload<int>(&QSpinBox::valueChanged), @@ -434,18 +436,7 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i connect(ui->groupConnectedController, &QGroupBox::toggled, [this](bool checked) { emit Connected(checked); }); - // Set up controller type. Only Player 1 can choose Handheld. - ui->comboControllerType->clear(); - - QStringList controller_types = { - tr("Pro Controller"), - tr("Dual Joycons"), - tr("Left Joycon"), - tr("Right Joycon"), - }; - if (player_index == 0) { - controller_types.append(tr("Handheld")); connect(ui->comboControllerType, qOverload<int>(&QComboBox::currentIndexChanged), [this](int index) { emit HandheldStateChanged(GetControllerTypeFromIndex(index) == @@ -453,17 +444,17 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i }); } + if (debug || player_index == 9) { + ui->groupConnectedController->setCheckable(false); + } + // The Debug Controller can only choose the Pro Controller. if (debug) { ui->buttonScreenshot->setEnabled(false); ui->buttonHome->setEnabled(false); - ui->groupConnectedController->setCheckable(false); - QStringList debug_controller_types = { - tr("Pro Controller"), - }; - ui->comboControllerType->addItems(debug_controller_types); + ui->comboControllerType->addItem(tr("Pro Controller")); } else { - ui->comboControllerType->addItems(controller_types); + SetConnectableControllers(); } UpdateControllerIcon(); @@ -475,11 +466,12 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i UpdateMotionButtons(); }); - connect(ui->comboDevices, qOverload<int>(&QComboBox::currentIndexChanged), this, + connect(ui->comboDevices, qOverload<int>(&QComboBox::activated), this, &ConfigureInputPlayer::UpdateMappingWithDefaults); + ui->comboDevices->setCurrentIndex(-1); + ui->buttonRefreshDevices->setIcon(QIcon::fromTheme(QStringLiteral("view-refresh"))); - UpdateInputDevices(); connect(ui->buttonRefreshDevices, &QPushButton::clicked, [this] { emit RefreshInputDevices(); }); @@ -490,14 +482,14 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i Common::ParamPackage params; if (input_subsystem->GetGCButtons()->IsPolling()) { params = input_subsystem->GetGCButtons()->GetNextInput(); - if (params.Has("engine")) { + if (params.Has("engine") && IsInputAcceptable(params)) { SetPollingResult(params, false); return; } } if (input_subsystem->GetGCAnalogs()->IsPolling()) { params = input_subsystem->GetGCAnalogs()->GetNextInput(); - if (params.Has("engine")) { + if (params.Has("engine") && IsInputAcceptable(params)) { SetPollingResult(params, false); return; } @@ -509,15 +501,54 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i return; } } + if (input_subsystem->GetMouseButtons()->IsPolling()) { + params = input_subsystem->GetMouseButtons()->GetNextInput(); + if (params.Has("engine") && IsInputAcceptable(params)) { + SetPollingResult(params, false); + return; + } + } + if (input_subsystem->GetMouseAnalogs()->IsPolling()) { + params = input_subsystem->GetMouseAnalogs()->GetNextInput(); + if (params.Has("engine") && IsInputAcceptable(params)) { + SetPollingResult(params, false); + return; + } + } + if (input_subsystem->GetMouseMotions()->IsPolling()) { + params = input_subsystem->GetMouseMotions()->GetNextInput(); + if (params.Has("engine") && IsInputAcceptable(params)) { + SetPollingResult(params, false); + return; + } + } + if (input_subsystem->GetMouseTouch()->IsPolling()) { + params = input_subsystem->GetMouseTouch()->GetNextInput(); + if (params.Has("engine") && IsInputAcceptable(params)) { + SetPollingResult(params, false); + return; + } + } for (auto& poller : device_pollers) { params = poller->GetNextInput(); - if (params.Has("engine")) { + if (params.Has("engine") && IsInputAcceptable(params)) { SetPollingResult(params, false); return; } } }); + UpdateInputProfiles(); + + connect(ui->buttonProfilesNew, &QPushButton::clicked, this, + &ConfigureInputPlayer::CreateProfile); + connect(ui->buttonProfilesDelete, &QPushButton::clicked, this, + &ConfigureInputPlayer::DeleteProfile); + connect(ui->comboProfiles, qOverload<int>(&QComboBox::activated), this, + &ConfigureInputPlayer::LoadProfile); + connect(ui->buttonProfilesSave, &QPushButton::clicked, this, + &ConfigureInputPlayer::SaveProfile); + LoadConfiguration(); // TODO(wwylele): enable this when we actually emulate it @@ -527,7 +558,7 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i ConfigureInputPlayer::~ConfigureInputPlayer() = default; void ConfigureInputPlayer::ApplyConfiguration() { - auto& player = Settings::values.players[player_index]; + auto& player = Settings::values.players.GetValue()[player_index]; auto& buttons = debug ? Settings::values.debug_pad_buttons : player.buttons; auto& analogs = debug ? Settings::values.debug_pad_analogs : player.analogs; @@ -541,33 +572,71 @@ void ConfigureInputPlayer::ApplyConfiguration() { } auto& motions = player.motions; + std::transform(motions_param.begin(), motions_param.end(), motions.begin(), [](const Common::ParamPackage& param) { return param.Serialize(); }); +} + +void ConfigureInputPlayer::TryConnectSelectedController() { + auto& player = Settings::values.players.GetValue()[player_index]; - player.controller_type = - static_cast<Settings::ControllerType>(ui->comboControllerType->currentIndex()); - player.connected = ui->groupConnectedController->isChecked(); + const auto controller_type = + GetControllerTypeFromIndex(ui->comboControllerType->currentIndex()); + const auto player_connected = ui->groupConnectedController->isChecked() && + controller_type != Settings::ControllerType::Handheld; - // Player 2-8 - if (player_index != 0) { - UpdateController(player.controller_type, player_index, player.connected); + if (player.controller_type == controller_type && player.connected == player_connected) { + // Set vibration devices in the event that the input device has changed. + ConfigureVibration::SetVibrationDevices(player_index); return; } - // Player 1 and Handheld - auto& handheld = Settings::values.players[HANDHELD_INDEX]; - // If Handheld is selected, copy all the settings from Player 1 to Handheld. - if (player.controller_type == Settings::ControllerType::Handheld) { - handheld = player; - handheld.connected = ui->groupConnectedController->isChecked(); - player.connected = false; // Disconnect Player 1 - } else { - player.connected = ui->groupConnectedController->isChecked(); - handheld.connected = false; // Disconnect Handheld + player.controller_type = controller_type; + player.connected = player_connected; + + ConfigureVibration::SetVibrationDevices(player_index); + + // Connect/Disconnect Handheld depending on Player 1's controller configuration. + if (player_index == 0) { + auto& handheld = Settings::values.players.GetValue()[HANDHELD_INDEX]; + if (controller_type == Settings::ControllerType::Handheld) { + handheld = player; + } + handheld.connected = ui->groupConnectedController->isChecked() && + controller_type == Settings::ControllerType::Handheld; + UpdateController(Settings::ControllerType::Handheld, HANDHELD_INDEX, handheld.connected); } - UpdateController(player.controller_type, player_index, player.connected); - UpdateController(Settings::ControllerType::Handheld, HANDHELD_INDEX, handheld.connected); + if (!player.connected) { + return; + } + + UpdateController(controller_type, player_index, player_connected); +} + +void ConfigureInputPlayer::TryDisconnectSelectedController() { + const auto& player = Settings::values.players.GetValue()[player_index]; + + const auto controller_type = + GetControllerTypeFromIndex(ui->comboControllerType->currentIndex()); + const auto player_connected = ui->groupConnectedController->isChecked() && + controller_type != Settings::ControllerType::Handheld; + + // Do not do anything if the controller configuration has not changed. + if (player.controller_type == controller_type && player.connected == player_connected) { + return; + } + + // Disconnect the controller first. + UpdateController(controller_type, player_index, false); +} + +void ConfigureInputPlayer::showEvent(QShowEvent* event) { + if (bottom_row == nullptr) { + return; + } + QWidget::showEvent(event); + ui->main->addWidget(bottom_row); } void ConfigureInputPlayer::changeEvent(QEvent* event) { @@ -584,7 +653,7 @@ void ConfigureInputPlayer::RetranslateUI() { } void ConfigureInputPlayer::LoadConfiguration() { - auto& player = Settings::values.players[player_index]; + auto& player = Settings::values.players.GetValue()[player_index]; if (debug) { std::transform(Settings::values.debug_pad_buttons.begin(), Settings::values.debug_pad_buttons.end(), buttons_param.begin(), @@ -602,52 +671,84 @@ void ConfigureInputPlayer::LoadConfiguration() { } UpdateUI(); + UpdateInputDeviceCombobox(); if (debug) { return; } - ui->comboControllerType->setCurrentIndex(static_cast<int>(player.controller_type)); + ui->comboControllerType->setCurrentIndex(GetIndexFromControllerType(player.controller_type)); ui->groupConnectedController->setChecked( player.connected || - (player_index == 0 && Settings::values.players[HANDHELD_INDEX].connected)); + (player_index == 0 && Settings::values.players.GetValue()[HANDHELD_INDEX].connected)); } -void ConfigureInputPlayer::UpdateInputDevices() { - input_devices = input_subsystem->GetInputDevices(); - ui->comboDevices->clear(); - for (auto device : input_devices) { - ui->comboDevices->addItem(QString::fromStdString(device.Get("display", "Unknown")), {}); - } +void ConfigureInputPlayer::ConnectPlayer(bool connected) { + ui->groupConnectedController->setChecked(connected); } -void ConfigureInputPlayer::RestoreDefaults() { - // Reset Buttons - for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; ++button_id) { - buttons_param[button_id] = Common::ParamPackage{ - InputCommon::GenerateKeyboardParam(Config::default_buttons[button_id])}; +void ConfigureInputPlayer::UpdateInputDeviceCombobox() { + // Skip input device persistence if "Input Devices" is set to "Any". + if (ui->comboDevices->currentIndex() == 0) { + UpdateInputDevices(); + return; } - // Reset Analogs and Modifier Buttons - for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; ++analog_id) { - for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; ++sub_button_id) { - Common::ParamPackage params{InputCommon::GenerateKeyboardParam( - Config::default_analogs[analog_id][sub_button_id])}; - SetAnalogParam(params, analogs_param[analog_id], analog_sub_buttons[sub_button_id]); - } + // Find the first button that isn't empty. + const auto button_param = + std::find_if(buttons_param.begin(), buttons_param.end(), + [](const Common::ParamPackage param) { return param.Has("engine"); }); + const bool buttons_empty = button_param == buttons_param.end(); + + const auto current_engine = button_param->Get("engine", ""); + const auto current_guid = button_param->Get("guid", ""); + const auto current_port = button_param->Get("port", ""); - stick_mod_param[analog_id] = Common::ParamPackage( - InputCommon::GenerateKeyboardParam(Config::default_stick_mod[analog_id])); + const bool is_keyboard_mouse = current_engine == "keyboard" || current_engine == "mouse"; + + UpdateInputDevices(); + + if (buttons_empty) { + return; } - for (int motion_id = 0; motion_id < Settings::NativeMotion::NumMotions; ++motion_id) { - motions_param[motion_id] = Common::ParamPackage{ - InputCommon::GenerateKeyboardParam(Config::default_motions[motion_id])}; + const bool all_one_device = + std::all_of(buttons_param.begin(), buttons_param.end(), + [current_engine, current_guid, current_port, + is_keyboard_mouse](const Common::ParamPackage param) { + if (is_keyboard_mouse) { + return !param.Has("engine") || param.Get("engine", "") == "keyboard" || + param.Get("engine", "") == "mouse"; + } + return !param.Has("engine") || (param.Get("engine", "") == current_engine && + param.Get("guid", "") == current_guid && + param.Get("port", "") == current_port); + }); + + if (all_one_device) { + if (is_keyboard_mouse) { + ui->comboDevices->setCurrentIndex(1); + return; + } + const auto devices_it = std::find_if( + input_devices.begin(), input_devices.end(), + [current_engine, current_guid, current_port](const Common::ParamPackage param) { + return param.Get("class", "") == current_engine && + param.Get("guid", "") == current_guid && + param.Get("port", "") == current_port; + }); + const int device_index = + devices_it != input_devices.end() + ? static_cast<int>(std::distance(input_devices.begin(), devices_it)) + : 0; + ui->comboDevices->setCurrentIndex(device_index); + } else { + ui->comboDevices->setCurrentIndex(0); } +} - UpdateUI(); - UpdateInputDevices(); - ui->comboControllerType->setCurrentIndex(0); +void ConfigureInputPlayer::RestoreDefaults() { + UpdateMappingWithDefaults(); } void ConfigureInputPlayer::ClearAll() { @@ -669,8 +770,6 @@ void ConfigureInputPlayer::ClearAll() { analogs_param[analog_id].Clear(); } - - stick_mod_param[analog_id].Clear(); } for (int motion_id = 0; motion_id < Settings::NativeMotion::NumMotions; ++motion_id) { @@ -707,7 +806,8 @@ void ConfigureInputPlayer::UpdateUI() { AnalogToText(analogs_param[analog_id], analog_sub_buttons[sub_button_id])); } - mod_buttons[analog_id]->setText(ButtonToText(stick_mod_param[analog_id])); + analog_map_modifier_button[analog_id]->setText( + ButtonToText(Common::ParamPackage{analogs_param[analog_id].Get("modifier", "")})); const auto deadzone_label = analog_map_deadzone_label[analog_id]; const auto deadzone_slider = analog_map_deadzone_slider[analog_id]; @@ -719,8 +819,9 @@ void ConfigureInputPlayer::UpdateUI() { int slider_value; auto& param = analogs_param[analog_id]; - const bool is_controller = - param.Get("engine", "") == "sdl" || param.Get("engine", "") == "gcpad"; + const bool is_controller = param.Get("engine", "") == "sdl" || + param.Get("engine", "") == "gcpad" || + param.Get("engine", "") == "mouse"; if (is_controller) { if (!param.Has("deadzone")) { @@ -751,117 +852,88 @@ void ConfigureInputPlayer::UpdateUI() { } } -void ConfigureInputPlayer::UpdateMappingWithDefaults() { - if (ui->comboDevices->currentIndex() < 2) { - return; - } - const auto& device = input_devices[ui->comboDevices->currentIndex()]; - auto button_mapping = input_subsystem->GetButtonMappingForDevice(device); - auto analog_mapping = input_subsystem->GetAnalogMappingForDevice(device); - for (std::size_t i = 0; i < buttons_param.size(); ++i) { - buttons_param[i] = button_mapping[static_cast<Settings::NativeButton::Values>(i)]; - } - for (std::size_t i = 0; i < analogs_param.size(); ++i) { - analogs_param[i] = analog_mapping[static_cast<Settings::NativeAnalog::Values>(i)]; - } - - UpdateUI(); -} - -void ConfigureInputPlayer::HandleClick( - QPushButton* button, std::function<void(const Common::ParamPackage&)> new_input_setter, - InputCommon::Polling::DeviceType type) { - if (button == ui->buttonMotionLeft || button == ui->buttonMotionRight) { - button->setText(tr("Shake!")); - } else { - button->setText(tr("[waiting]")); - } - button->setFocus(); +void ConfigureInputPlayer::SetConnectableControllers() { + const auto add_controllers = [this](bool enable_all, + Controller_NPad::NpadStyleSet npad_style_set = {}) { + index_controller_type_pairs.clear(); + ui->comboControllerType->clear(); - // The first two input devices are always Any and Keyboard/Mouse. If the user filtered to a - // controller, then they don't want keyboard/mouse input - want_keyboard_mouse = ui->comboDevices->currentIndex() < 2; + if (enable_all || npad_style_set.pro_controller == 1) { + index_controller_type_pairs.emplace_back(ui->comboControllerType->count(), + Settings::ControllerType::ProController); + ui->comboControllerType->addItem(tr("Pro Controller")); + } - input_setter = new_input_setter; + if (enable_all || npad_style_set.joycon_dual == 1) { + index_controller_type_pairs.emplace_back(ui->comboControllerType->count(), + Settings::ControllerType::DualJoyconDetached); + ui->comboControllerType->addItem(tr("Dual Joycons")); + } - device_pollers = input_subsystem->GetPollers(type); + if (enable_all || npad_style_set.joycon_left == 1) { + index_controller_type_pairs.emplace_back(ui->comboControllerType->count(), + Settings::ControllerType::LeftJoycon); + ui->comboControllerType->addItem(tr("Left Joycon")); + } - for (auto& poller : device_pollers) { - poller->Start(); - } + if (enable_all || npad_style_set.joycon_right == 1) { + index_controller_type_pairs.emplace_back(ui->comboControllerType->count(), + Settings::ControllerType::RightJoycon); + ui->comboControllerType->addItem(tr("Right Joycon")); + } - QWidget::grabMouse(); - QWidget::grabKeyboard(); + if (player_index == 0 && (enable_all || npad_style_set.handheld == 1)) { + index_controller_type_pairs.emplace_back(ui->comboControllerType->count(), + Settings::ControllerType::Handheld); + ui->comboControllerType->addItem(tr("Handheld")); + } + }; - if (type == InputCommon::Polling::DeviceType::Button) { - input_subsystem->GetGCButtons()->BeginConfiguration(); - } else { - input_subsystem->GetGCAnalogs()->BeginConfiguration(); - } + Core::System& system{Core::System::GetInstance()}; - if (type == InputCommon::Polling::DeviceType::Motion) { - input_subsystem->GetUDPMotions()->BeginConfiguration(); + if (!system.IsPoweredOn()) { + add_controllers(true); + return; } - timeout_timer->start(2500); // Cancel after 2.5 seconds - poll_timer->start(50); // Check for new inputs every 50ms -} - -void ConfigureInputPlayer::SetPollingResult(const Common::ParamPackage& params, bool abort) { - timeout_timer->stop(); - poll_timer->stop(); - for (auto& poller : device_pollers) { - poller->Stop(); - } + Service::SM::ServiceManager& sm = system.ServiceManager(); - QWidget::releaseMouse(); - QWidget::releaseKeyboard(); + auto& npad = sm.GetService<Hid>("hid")->GetAppletResource()->GetController<Controller_NPad>( + HidController::NPad); - input_subsystem->GetGCButtons()->EndConfiguration(); - input_subsystem->GetGCAnalogs()->EndConfiguration(); + add_controllers(false, npad.GetSupportedStyleSet()); +} - input_subsystem->GetUDPMotions()->EndConfiguration(); +Settings::ControllerType ConfigureInputPlayer::GetControllerTypeFromIndex(int index) const { + const auto it = + std::find_if(index_controller_type_pairs.begin(), index_controller_type_pairs.end(), + [index](const auto& pair) { return pair.first == index; }); - if (!abort) { - (*input_setter)(params); + if (it == index_controller_type_pairs.end()) { + return Settings::ControllerType::ProController; } - UpdateUI(); - input_setter = std::nullopt; + return it->second; } -void ConfigureInputPlayer::mousePressEvent(QMouseEvent* event) { - if (!input_setter || !event) { - return; - } +int ConfigureInputPlayer::GetIndexFromControllerType(Settings::ControllerType type) const { + const auto it = + std::find_if(index_controller_type_pairs.begin(), index_controller_type_pairs.end(), + [type](const auto& pair) { return pair.second == type; }); - if (want_keyboard_mouse) { - SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->button())}, - false); - } else { - // We don't want any mouse buttons, so don't stop polling - return; + if (it == index_controller_type_pairs.end()) { + return -1; } - SetPollingResult({}, true); + return it->first; } -void ConfigureInputPlayer::keyPressEvent(QKeyEvent* event) { - if (!input_setter || !event) { - return; - } - - if (event->key() != Qt::Key_Escape) { - if (want_keyboard_mouse) { - SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->key())}, - false); - } else { - // Escape key wasn't pressed and we don't want any keyboard keys, so don't stop polling - return; - } +void ConfigureInputPlayer::UpdateInputDevices() { + input_devices = input_subsystem->GetInputDevices(); + ui->comboDevices->clear(); + for (auto device : input_devices) { + ui->comboDevices->addItem(QString::fromStdString(device.Get("display", "Unknown")), {}); } - - SetPollingResult({}, true); } void ConfigureInputPlayer::UpdateControllerIcon() { @@ -884,7 +956,7 @@ void ConfigureInputPlayer::UpdateControllerIcon() { } }(); - const QString theme = [this] { + const QString theme = [] { if (QIcon::themeName().contains(QStringLiteral("dark"))) { return QStringLiteral("_dark"); } else if (QIcon::themeName().contains(QStringLiteral("midnight"))) { @@ -985,14 +1057,267 @@ void ConfigureInputPlayer::UpdateMotionButtons() { } } -void ConfigureInputPlayer::showEvent(QShowEvent* event) { - if (bottom_row == nullptr) { +void ConfigureInputPlayer::UpdateMappingWithDefaults() { + if (ui->comboDevices->currentIndex() == 0) { return; } - QWidget::showEvent(event); - ui->main->addWidget(bottom_row); + + if (ui->comboDevices->currentIndex() == 1) { + // Reset keyboard bindings + for (int button_id = 0; button_id < Settings::NativeButton::NumButtons; ++button_id) { + buttons_param[button_id] = Common::ParamPackage{ + InputCommon::GenerateKeyboardParam(Config::default_buttons[button_id])}; + } + for (int analog_id = 0; analog_id < Settings::NativeAnalog::NumAnalogs; ++analog_id) { + for (int sub_button_id = 0; sub_button_id < ANALOG_SUB_BUTTONS_NUM; ++sub_button_id) { + Common::ParamPackage params{InputCommon::GenerateKeyboardParam( + Config::default_analogs[analog_id][sub_button_id])}; + SetAnalogParam(params, analogs_param[analog_id], analog_sub_buttons[sub_button_id]); + } + + analogs_param[analog_id].Set("modifier", InputCommon::GenerateKeyboardParam( + Config::default_stick_mod[analog_id])); + } + + for (int motion_id = 0; motion_id < Settings::NativeMotion::NumMotions; ++motion_id) { + motions_param[motion_id] = Common::ParamPackage{ + InputCommon::GenerateKeyboardParam(Config::default_motions[motion_id])}; + } + + UpdateUI(); + return; + } + + // Reset controller bindings + const auto& device = input_devices[ui->comboDevices->currentIndex()]; + auto button_mapping = input_subsystem->GetButtonMappingForDevice(device); + auto analog_mapping = input_subsystem->GetAnalogMappingForDevice(device); + for (std::size_t i = 0; i < buttons_param.size(); ++i) { + buttons_param[i] = button_mapping[static_cast<Settings::NativeButton::Values>(i)]; + } + for (std::size_t i = 0; i < analogs_param.size(); ++i) { + analogs_param[i] = analog_mapping[static_cast<Settings::NativeAnalog::Values>(i)]; + } + + UpdateUI(); } -void ConfigureInputPlayer::ConnectPlayer(bool connected) { - ui->groupConnectedController->setChecked(connected); +void ConfigureInputPlayer::HandleClick( + QPushButton* button, std::function<void(const Common::ParamPackage&)> new_input_setter, + InputCommon::Polling::DeviceType type) { + if (button == ui->buttonMotionLeft || button == ui->buttonMotionRight) { + button->setText(tr("Shake!")); + } else { + button->setText(tr("[waiting]")); + } + button->setFocus(); + + // The first two input devices are always Any and Keyboard/Mouse. If the user filtered to a + // controller, then they don't want keyboard/mouse input + want_keyboard_mouse = ui->comboDevices->currentIndex() < 2; + + input_setter = new_input_setter; + + device_pollers = input_subsystem->GetPollers(type); + + for (auto& poller : device_pollers) { + poller->Start(); + } + + QWidget::grabMouse(); + QWidget::grabKeyboard(); + + if (type == InputCommon::Polling::DeviceType::Button) { + input_subsystem->GetGCButtons()->BeginConfiguration(); + } else { + input_subsystem->GetGCAnalogs()->BeginConfiguration(); + } + + if (type == InputCommon::Polling::DeviceType::Motion) { + input_subsystem->GetUDPMotions()->BeginConfiguration(); + } + + if (type == InputCommon::Polling::DeviceType::Button) { + input_subsystem->GetMouseButtons()->BeginConfiguration(); + } else if (type == InputCommon::Polling::DeviceType::AnalogPreferred) { + input_subsystem->GetMouseAnalogs()->BeginConfiguration(); + } else if (type == InputCommon::Polling::DeviceType::Motion) { + input_subsystem->GetMouseMotions()->BeginConfiguration(); + } else { + input_subsystem->GetMouseTouch()->BeginConfiguration(); + } + + timeout_timer->start(2500); // Cancel after 2.5 seconds + poll_timer->start(50); // Check for new inputs every 50ms +} + +void ConfigureInputPlayer::SetPollingResult(const Common::ParamPackage& params, bool abort) { + timeout_timer->stop(); + poll_timer->stop(); + for (auto& poller : device_pollers) { + poller->Stop(); + } + + QWidget::releaseMouse(); + QWidget::releaseKeyboard(); + + input_subsystem->GetGCButtons()->EndConfiguration(); + input_subsystem->GetGCAnalogs()->EndConfiguration(); + + input_subsystem->GetUDPMotions()->EndConfiguration(); + + input_subsystem->GetMouseButtons()->EndConfiguration(); + input_subsystem->GetMouseAnalogs()->EndConfiguration(); + input_subsystem->GetMouseMotions()->EndConfiguration(); + input_subsystem->GetMouseTouch()->EndConfiguration(); + + if (!abort) { + (*input_setter)(params); + } + + UpdateUI(); + UpdateInputDeviceCombobox(); + + input_setter = std::nullopt; +} + +bool ConfigureInputPlayer::IsInputAcceptable(const Common::ParamPackage& params) const { + if (ui->comboDevices->currentIndex() == 0) { + return true; + } + + // Keyboard/Mouse + if (ui->comboDevices->currentIndex() == 1) { + return params.Get("engine", "") == "keyboard" || params.Get("engine", "") == "mouse"; + } + + const auto current_input_device = input_devices[ui->comboDevices->currentIndex()]; + return params.Get("engine", "") == current_input_device.Get("class", "") && + params.Get("guid", "") == current_input_device.Get("guid", "") && + params.Get("port", "") == current_input_device.Get("port", ""); +} + +void ConfigureInputPlayer::mousePressEvent(QMouseEvent* event) { + if (!input_setter || !event) { + return; + } + + input_subsystem->GetMouse()->PressButton(0, 0, event->button()); +} + +void ConfigureInputPlayer::keyPressEvent(QKeyEvent* event) { + if (!input_setter || !event) { + return; + } + + if (event->key() != Qt::Key_Escape) { + if (want_keyboard_mouse) { + SetPollingResult(Common::ParamPackage{InputCommon::GenerateKeyboardParam(event->key())}, + false); + } else { + // Escape key wasn't pressed and we don't want any keyboard keys, so don't stop polling + return; + } + } + + SetPollingResult({}, true); +} + +void ConfigureInputPlayer::CreateProfile() { + const auto profile_name = + LimitableInputDialog::GetText(this, tr("New Profile"), tr("Enter a profile name:"), 1, 20); + + if (profile_name.isEmpty()) { + return; + } + + if (!InputProfiles::IsProfileNameValid(profile_name.toStdString())) { + QMessageBox::critical(this, tr("Create Input Profile"), + tr("The given profile name is not valid!")); + return; + } + + ApplyConfiguration(); + + if (!profiles->CreateProfile(profile_name.toStdString(), player_index)) { + QMessageBox::critical(this, tr("Create Input Profile"), + tr("Failed to create the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } + + emit RefreshInputProfiles(player_index); + + ui->comboProfiles->addItem(profile_name); + ui->comboProfiles->setCurrentIndex(ui->comboProfiles->count() - 1); +} + +void ConfigureInputPlayer::DeleteProfile() { + const QString profile_name = ui->comboProfiles->itemText(ui->comboProfiles->currentIndex()); + + if (profile_name.isEmpty()) { + return; + } + + if (!profiles->DeleteProfile(profile_name.toStdString())) { + QMessageBox::critical(this, tr("Delete Input Profile"), + tr("Failed to delete the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } + + emit RefreshInputProfiles(player_index); + + ui->comboProfiles->removeItem(ui->comboProfiles->currentIndex()); + ui->comboProfiles->setCurrentIndex(-1); +} + +void ConfigureInputPlayer::LoadProfile() { + const QString profile_name = ui->comboProfiles->itemText(ui->comboProfiles->currentIndex()); + + if (profile_name.isEmpty()) { + return; + } + + ApplyConfiguration(); + + if (!profiles->LoadProfile(profile_name.toStdString(), player_index)) { + QMessageBox::critical(this, tr("Load Input Profile"), + tr("Failed to load the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } + + LoadConfiguration(); +} + +void ConfigureInputPlayer::SaveProfile() { + const QString profile_name = ui->comboProfiles->itemText(ui->comboProfiles->currentIndex()); + + if (profile_name.isEmpty()) { + return; + } + + ApplyConfiguration(); + + if (!profiles->SaveProfile(profile_name.toStdString(), player_index)) { + QMessageBox::critical(this, tr("Save Input Profile"), + tr("Failed to save the input profile \"%1\"").arg(profile_name)); + UpdateInputProfiles(); + emit RefreshInputProfiles(player_index); + return; + } +} + +void ConfigureInputPlayer::UpdateInputProfiles() { + ui->comboProfiles->clear(); + + for (const auto& profile_name : profiles->GetInputProfileNames()) { + ui->comboProfiles->addItem(QString::fromStdString(profile_name)); + } + + ui->comboProfiles->setCurrentIndex(-1); } diff --git a/src/yuzu/configuration/configure_input_player.h b/src/yuzu/configuration/configure_input_player.h index ce443dec5..c4ae50de7 100644 --- a/src/yuzu/configuration/configure_input_player.h +++ b/src/yuzu/configuration/configure_input_player.h @@ -9,6 +9,7 @@ #include <memory> #include <optional> #include <string> +#include <vector> #include <QWidget> @@ -26,6 +27,8 @@ class QString; class QTimer; class QWidget; +class InputProfiles; + namespace InputCommon { class InputSubsystem; } @@ -45,14 +48,32 @@ class ConfigureInputPlayer : public QWidget { public: explicit ConfigureInputPlayer(QWidget* parent, std::size_t player_index, QWidget* bottom_row, InputCommon::InputSubsystem* input_subsystem_, - bool debug = false); + InputProfiles* profiles_, bool debug = false); ~ConfigureInputPlayer() override; /// Save all button configurations to settings file. void ApplyConfiguration(); + /** + * Attempts to connect the currently selected controller in the HID backend. + * This function will not do anything if it is not connected in the frontend. + */ + void TryConnectSelectedController(); + + /** + * Attempts to disconnect the currently selected controller in the HID backend. + * This function will not do anything if the configuration has not changed. + */ + void TryDisconnectSelectedController(); + + /// Set the connection state checkbox (used to sync state). + void ConnectPlayer(bool connected); + /// Update the input devices combobox. - void UpdateInputDevices(); + void UpdateInputDeviceCombobox(); + + /// Updates the list of controller profiles. + void UpdateInputProfiles(); /// Restore all buttons to their default values. void RestoreDefaults(); @@ -60,9 +81,6 @@ public: /// Clear all input configuration. void ClearAll(); - /// Set the connection state checkbox (used to sync state). - void ConnectPlayer(bool connected); - signals: /// Emitted when this controller is connected by the user. void Connected(bool connected); @@ -70,6 +88,12 @@ signals: void HandheldStateChanged(bool is_handheld); /// Emitted when the input devices combobox is being refreshed. void RefreshInputDevices(); + /** + * Emitted when the input profiles combobox is being refreshed. + * The player_index represents the current player's index, and the profile combobox + * will not be updated for this index as they are already updated by other mechanisms. + */ + void RefreshInputProfiles(std::size_t player_index); protected: void showEvent(QShowEvent* event) override; @@ -89,6 +113,9 @@ private: /// Finish polling and configure input using the input_setter. void SetPollingResult(const Common::ParamPackage& params, bool abort); + /// Checks whether a given input can be accepted. + bool IsInputAcceptable(const Common::ParamPackage& params) const; + /// Handle mouse button press events. void mousePressEvent(QMouseEvent* event) override; @@ -98,8 +125,17 @@ private: /// Update UI to reflect current configuration. void UpdateUI(); - /// Update the controller selection combobox - void UpdateControllerCombobox(); + /// Sets the available controllers. + void SetConnectableControllers(); + + /// Gets the Controller Type for a given controller combobox index. + Settings::ControllerType GetControllerTypeFromIndex(int index) const; + + /// Gets the controller combobox index for a given Controller Type. + int GetIndexFromControllerType(Settings::ControllerType type) const; + + /// Update the available input devices. + void UpdateInputDevices(); /// Update the current controller icon. void UpdateControllerIcon(); @@ -113,6 +149,18 @@ private: /// Gets the default controller mapping for this device and auto configures the input to match. void UpdateMappingWithDefaults(); + /// Creates a controller profile. + void CreateProfile(); + + /// Deletes the selected controller profile. + void DeleteProfile(); + + /// Loads the selected controller profile. + void LoadProfile(); + + /// Saves the current controller configuration into a selected controller profile. + void SaveProfile(); + std::unique_ptr<Ui::ConfigureInputPlayer> ui; std::size_t player_index; @@ -120,9 +168,14 @@ private: InputCommon::InputSubsystem* input_subsystem; + InputProfiles* profiles; + std::unique_ptr<QTimer> timeout_timer; std::unique_ptr<QTimer> poll_timer; + /// Stores a pair of "Connected Controllers" combobox index and Controller Type enum. + std::vector<std::pair<int, Settings::ControllerType>> index_controller_type_pairs; + static constexpr int PLAYER_COUNT = 8; std::array<QCheckBox*, PLAYER_COUNT> player_connected_checkbox; @@ -131,26 +184,25 @@ private: std::array<Common::ParamPackage, Settings::NativeButton::NumButtons> buttons_param; std::array<Common::ParamPackage, Settings::NativeAnalog::NumAnalogs> analogs_param; - std::array<Common::ParamPackage, Settings::NativeAnalog::NumAnalogs> stick_mod_param; std::array<Common::ParamPackage, Settings::NativeMotion::NumMotions> motions_param; static constexpr int ANALOG_SUB_BUTTONS_NUM = 4; /// Each button input is represented by a QPushButton. std::array<QPushButton*, Settings::NativeButton::NumButtons> button_map; - /// Each motion input is represented by a QPushButton. - std::array<QPushButton*, Settings::NativeMotion::NumMotions> motion_map; - /// Extra buttons for the modifiers. - std::array<QPushButton*, Settings::NativeAnalog::NumAnalogs> mod_buttons; /// A group of four QPushButtons represent one analog input. The buttons each represent up, /// down, left, right, respectively. std::array<std::array<QPushButton*, ANALOG_SUB_BUTTONS_NUM>, Settings::NativeAnalog::NumAnalogs> analog_map_buttons; + /// Each motion input is represented by a QPushButton. + std::array<QPushButton*, Settings::NativeMotion::NumMotions> motion_map; + std::array<QLabel*, Settings::NativeAnalog::NumAnalogs> analog_map_deadzone_label; std::array<QSlider*, Settings::NativeAnalog::NumAnalogs> analog_map_deadzone_slider; std::array<QGroupBox*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_groupbox; + std::array<QPushButton*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_button; std::array<QLabel*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_label; std::array<QSlider*, Settings::NativeAnalog::NumAnalogs> analog_map_modifier_slider; std::array<QGroupBox*, Settings::NativeAnalog::NumAnalogs> analog_map_range_groupbox; @@ -160,12 +212,15 @@ private: std::vector<std::unique_ptr<InputCommon::Polling::DevicePoller>> device_pollers; + /// A flag to indicate that the "Map Analog Stick" pop-up has been shown and accepted once. + bool map_analog_stick_accepted{}; + /// A flag to indicate if keyboard keys are okay when configuring an input. If this is false, /// keyboard events are ignored. - bool want_keyboard_mouse = false; + bool want_keyboard_mouse{}; /// List of physical devices users can map with. If a SDL backed device is selected, then you - /// can usue this device to get a default mapping. + /// can use this device to get a default mapping. std::vector<Common::ParamPackage> input_devices; /// Bottom row is where console wide settings are held, and its "owned" by the parent diff --git a/src/yuzu/configuration/configure_input_player.ui b/src/yuzu/configuration/configure_input_player.ui index e03461d9d..1e78b4c10 100644 --- a/src/yuzu/configuration/configure_input_player.ui +++ b/src/yuzu/configuration/configure_input_player.ui @@ -83,6 +83,12 @@ </property> <item> <widget class="QComboBox" name="comboControllerType"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>21</height> + </size> + </property> <item> <property name="text"> <string>Pro Controller</string> @@ -136,6 +142,12 @@ </property> <item> <widget class="QComboBox" name="comboDevices"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>21</height> + </size> + </property> <item> <property name="text"> <string>Any</string> @@ -152,14 +164,14 @@ <widget class="QPushButton" name="buttonRefreshDevices"> <property name="minimumSize"> <size> - <width>24</width> - <height>22</height> + <width>21</width> + <height>21</height> </size> </property> <property name="maximumSize"> <size> - <width>24</width> - <height>22</height> + <width>21</width> + <height>21</height> </size> </property> <property name="styleSheet"> @@ -198,18 +210,25 @@ <number>5</number> </property> <item> - <widget class="QComboBox" name="comboProfiles"/> + <widget class="QComboBox" name="comboProfiles"> + <property name="minimumSize"> + <size> + <width>0</width> + <height>21</height> + </size> + </property> + </widget> </item> <item> <widget class="QPushButton" name="buttonProfilesSave"> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Save</string> @@ -220,12 +239,12 @@ <widget class="QPushButton" name="buttonProfilesNew"> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>New</string> @@ -236,12 +255,12 @@ <widget class="QPushButton" name="buttonProfilesDelete"> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Delete</string> @@ -393,18 +412,18 @@ <widget class="QPushButton" name="buttonLStickUp"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Up</string> @@ -463,18 +482,18 @@ <widget class="QPushButton" name="buttonLStickLeft"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Left</string> @@ -512,18 +531,18 @@ <widget class="QPushButton" name="buttonLStickRight"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Right</string> @@ -594,18 +613,18 @@ <widget class="QPushButton" name="buttonLStickDown"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Down</string> @@ -664,18 +683,18 @@ <widget class="QPushButton" name="buttonLStick"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Pressed</string> @@ -713,18 +732,18 @@ <widget class="QPushButton" name="buttonLStickMod"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Modifier</string> @@ -759,13 +778,13 @@ <widget class="QSpinBox" name="spinboxLStickRange"> <property name="minimumSize"> <size> - <width>55</width> + <width>68</width> <height>21</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -966,18 +985,18 @@ <widget class="QPushButton" name="buttonDpadUp"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Up</string> @@ -1036,18 +1055,18 @@ <widget class="QPushButton" name="buttonDpadLeft"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Left</string> @@ -1085,18 +1104,18 @@ <widget class="QPushButton" name="buttonDpadRight"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Right</string> @@ -1167,18 +1186,18 @@ <widget class="QPushButton" name="buttonDpadDown"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Down</string> @@ -1292,18 +1311,18 @@ <widget class="QPushButton" name="buttonL"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>L</string> @@ -1341,18 +1360,18 @@ <widget class="QPushButton" name="buttonZL"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>ZL</string> @@ -1445,18 +1464,18 @@ <widget class="QPushButton" name="buttonMinus"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Minus</string> @@ -1494,18 +1513,18 @@ <widget class="QPushButton" name="buttonScreenshot"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Capture</string> @@ -1564,18 +1583,18 @@ <widget class="QPushButton" name="buttonPlus"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Plus</string> @@ -1613,18 +1632,18 @@ <widget class="QPushButton" name="buttonHome"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Home</string> @@ -1717,18 +1736,18 @@ <widget class="QPushButton" name="buttonR"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>R</string> @@ -1766,18 +1785,18 @@ <widget class="QPushButton" name="buttonZR"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>ZR</string> @@ -1870,18 +1889,18 @@ <widget class="QPushButton" name="buttonSL"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>SL</string> @@ -1919,18 +1938,18 @@ <widget class="QPushButton" name="buttonSR"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>SR</string> @@ -2027,18 +2046,18 @@ <widget class="QPushButton" name="buttonMotionLeft"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Left</string> @@ -2076,18 +2095,18 @@ <widget class="QPushButton" name="buttonMotionRight"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Right</string> @@ -2225,18 +2244,18 @@ <widget class="QPushButton" name="buttonX"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>X</string> @@ -2295,18 +2314,18 @@ <widget class="QPushButton" name="buttonY"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Y</string> @@ -2344,18 +2363,18 @@ <widget class="QPushButton" name="buttonA"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>A</string> @@ -2426,18 +2445,18 @@ <widget class="QPushButton" name="buttonB"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>B</string> @@ -2580,18 +2599,18 @@ <widget class="QPushButton" name="buttonRStickUp"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Up</string> @@ -2650,18 +2669,18 @@ <widget class="QPushButton" name="buttonRStickLeft"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Left</string> @@ -2699,18 +2718,18 @@ <widget class="QPushButton" name="buttonRStickRight"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Right</string> @@ -2781,18 +2800,18 @@ <widget class="QPushButton" name="buttonRStickDown"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Down</string> @@ -2851,18 +2870,18 @@ <widget class="QPushButton" name="buttonRStick"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Pressed</string> @@ -2900,18 +2919,18 @@ <widget class="QPushButton" name="buttonRStickMod"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> <property name="styleSheet"> - <string notr="true">min-width: 55px;</string> + <string notr="true">min-width: 68px;</string> </property> <property name="text"> <string>Modifier</string> @@ -2946,13 +2965,13 @@ <widget class="QSpinBox" name="spinboxRStickRange"> <property name="minimumSize"> <size> - <width>55</width> + <width>68</width> <height>21</height> </size> </property> <property name="maximumSize"> <size> - <width>55</width> + <width>68</width> <height>16777215</height> </size> </property> diff --git a/src/yuzu/configuration/configure_input_profile_dialog.cpp b/src/yuzu/configuration/configure_input_profile_dialog.cpp new file mode 100644 index 000000000..1f5cfa75b --- /dev/null +++ b/src/yuzu/configuration/configure_input_profile_dialog.cpp @@ -0,0 +1,37 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "ui_configure_input_profile_dialog.h" +#include "yuzu/configuration/configure_input_player.h" +#include "yuzu/configuration/configure_input_profile_dialog.h" + +ConfigureInputProfileDialog::ConfigureInputProfileDialog( + QWidget* parent, InputCommon::InputSubsystem* input_subsystem, InputProfiles* profiles) + : QDialog(parent), ui(std::make_unique<Ui::ConfigureInputProfileDialog>()), + profile_widget(new ConfigureInputPlayer(this, 9, nullptr, input_subsystem, profiles, false)) { + ui->setupUi(this); + + ui->controllerLayout->addWidget(profile_widget); + + connect(ui->clear_all_button, &QPushButton::clicked, this, + [this] { profile_widget->ClearAll(); }); + connect(ui->restore_defaults_button, &QPushButton::clicked, this, + [this] { profile_widget->RestoreDefaults(); }); + + RetranslateUI(); +} + +ConfigureInputProfileDialog::~ConfigureInputProfileDialog() = default; + +void ConfigureInputProfileDialog::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QDialog::changeEvent(event); +} + +void ConfigureInputProfileDialog::RetranslateUI() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_input_profile_dialog.h b/src/yuzu/configuration/configure_input_profile_dialog.h new file mode 100644 index 000000000..e6386bdbb --- /dev/null +++ b/src/yuzu/configuration/configure_input_profile_dialog.h @@ -0,0 +1,40 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <QDialog> + +class QPushButton; + +class ConfigureInputPlayer; + +class InputProfiles; + +namespace InputCommon { +class InputSubsystem; +} + +namespace Ui { +class ConfigureInputProfileDialog; +} + +class ConfigureInputProfileDialog : public QDialog { + Q_OBJECT + +public: + explicit ConfigureInputProfileDialog(QWidget* parent, + InputCommon::InputSubsystem* input_subsystem, + InputProfiles* profiles); + ~ConfigureInputProfileDialog() override; + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + std::unique_ptr<Ui::ConfigureInputProfileDialog> ui; + + ConfigureInputPlayer* profile_widget; +}; diff --git a/src/yuzu/configuration/configure_input_dialog.ui b/src/yuzu/configuration/configure_input_profile_dialog.ui index b92ddb200..726cf6905 100644 --- a/src/yuzu/configuration/configure_input_dialog.ui +++ b/src/yuzu/configuration/configure_input_profile_dialog.ui @@ -1,7 +1,7 @@ <?xml version="1.0" encoding="UTF-8"?> <ui version="4.0"> - <class>ConfigureInputDialog</class> - <widget class="QDialog" name="ConfigureInputDialog"> + <class>ConfigureInputProfileDialog</class> + <widget class="QDialog" name="ConfigureInputProfileDialog"> <property name="geometry"> <rect> <x>0</x> @@ -11,7 +11,7 @@ </rect> </property> <property name="windowTitle"> - <string>Configure Input</string> + <string>Create Input Profile</string> </property> <layout class="QVBoxLayout" name="verticalLayout"> <property name="spacing"> @@ -30,11 +30,25 @@ <number>9</number> </property> <item> - <layout class="QHBoxLayout" name="inputLayout"/> + <layout class="QHBoxLayout" name="controllerLayout"/> </item> <item> <layout class="QHBoxLayout" name="horizontalLayout"> <item> + <widget class="QPushButton" name="clear_all_button"> + <property name="text"> + <string>Clear</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="restore_defaults_button"> + <property name="text"> + <string>Defaults</string> + </property> + </widget> + </item> + <item> <widget class="QDialogButtonBox" name="buttonBox"> <property name="standardButtons"> <set>QDialogButtonBox::Ok</set> @@ -50,7 +64,7 @@ <connection> <sender>buttonBox</sender> <signal>accepted()</signal> - <receiver>ConfigureInputDialog</receiver> + <receiver>ConfigureInputProfileDialog</receiver> <slot>accept()</slot> </connection> </connections> diff --git a/src/yuzu/configuration/configure_motion_touch.cpp b/src/yuzu/configuration/configure_motion_touch.cpp index c7d085151..eb8eacbf9 100644 --- a/src/yuzu/configuration/configure_motion_touch.cpp +++ b/src/yuzu/configuration/configure_motion_touch.cpp @@ -3,10 +3,12 @@ // Refer to the license.txt file included. #include <array> +#include <sstream> #include <QCloseEvent> #include <QLabel> #include <QMessageBox> #include <QPushButton> +#include <QStringListModel> #include <QVBoxLayout> #include "common/logging/log.h" #include "core/settings.h" @@ -49,6 +51,8 @@ CalibrationConfigurationDialog::CalibrationConfigurationDialog(QWidget* parent, case CalibrationConfigurationJob::Status::Completed: text = tr("Configuration completed!"); break; + default: + break; } QMetaObject::invokeMethod(this, "UpdateLabelText", Q_ARG(QString, text)); if (status == CalibrationConfigurationJob::Status::Completed) { @@ -74,11 +78,6 @@ void CalibrationConfigurationDialog::UpdateButtonText(const QString& text) { cancel_button->setText(text); } -constexpr std::array<std::pair<const char*, const char*>, 2> MotionProviders = {{ - {"motion_emu", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "Mouse (Right Click)")}, - {"cemuhookudp", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "CemuhookUDP")}, -}}; - constexpr std::array<std::pair<const char*, const char*>, 2> TouchProviders = {{ {"emu_window", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "Emulator Window")}, {"cemuhookudp", QT_TRANSLATE_NOOP("ConfigureMotionTouch", "CemuhookUDP")}, @@ -89,9 +88,6 @@ ConfigureMotionTouch::ConfigureMotionTouch(QWidget* parent, : QDialog(parent), input_subsystem{input_subsystem_}, ui(std::make_unique<Ui::ConfigureMotionTouch>()) { ui->setupUi(this); - for (const auto& [provider, name] : MotionProviders) { - ui->motion_provider->addItem(tr(name), QString::fromUtf8(provider)); - } for (const auto& [provider, name] : TouchProviders) { ui->touch_provider->addItem(tr(name), QString::fromUtf8(provider)); } @@ -116,8 +112,6 @@ void ConfigureMotionTouch::SetConfiguration() { const std::string motion_engine = motion_param.Get("engine", "motion_emu"); const std::string touch_engine = touch_param.Get("engine", "emu_window"); - ui->motion_provider->setCurrentIndex( - ui->motion_provider->findData(QString::fromStdString(motion_engine))); ui->touch_provider->setCurrentIndex( ui->touch_provider->findData(QString::fromStdString(touch_engine))); ui->touch_from_button_checkbox->setChecked(Settings::values.use_touch_from_button); @@ -133,23 +127,30 @@ void ConfigureMotionTouch::SetConfiguration() { max_x = touch_param.Get("max_x", 1800); max_y = touch_param.Get("max_y", 850); - ui->udp_server->setText(QString::fromStdString(Settings::values.udp_input_address)); - ui->udp_port->setText(QString::number(Settings::values.udp_input_port)); - ui->udp_pad_index->setCurrentIndex(Settings::values.udp_pad_index); + ui->udp_server->setText(QString::fromStdString("127.0.0.1")); + ui->udp_port->setText(QString::number(26760)); + + udp_server_list_model = new QStringListModel(this); + udp_server_list_model->setStringList({}); + ui->udp_server_list->setModel(udp_server_list_model); + + std::stringstream ss(Settings::values.udp_input_servers); + std::string token; + + while (std::getline(ss, token, ',')) { + const int row = udp_server_list_model->rowCount(); + udp_server_list_model->insertRows(row, 1); + const QModelIndex index = udp_server_list_model->index(row); + udp_server_list_model->setData(index, QString::fromStdString(token)); + } } void ConfigureMotionTouch::UpdateUiDisplay() { - const QString motion_engine = ui->motion_provider->currentData().toString(); const QString touch_engine = ui->touch_provider->currentData().toString(); const QString cemuhook_udp = QStringLiteral("cemuhookudp"); - if (motion_engine == QStringLiteral("motion_emu")) { - ui->motion_sensitivity_label->setVisible(true); - ui->motion_sensitivity->setVisible(true); - } else { - ui->motion_sensitivity_label->setVisible(false); - ui->motion_sensitivity->setVisible(false); - } + ui->motion_sensitivity_label->setVisible(true); + ui->motion_sensitivity->setVisible(true); if (touch_engine == cemuhook_udp) { ui->touch_calibration->setVisible(true); @@ -163,19 +164,15 @@ void ConfigureMotionTouch::UpdateUiDisplay() { ui->touch_calibration_label->setVisible(false); } - if (motion_engine == cemuhook_udp || touch_engine == cemuhook_udp) { - ui->udp_config_group_box->setVisible(true); - } else { - ui->udp_config_group_box->setVisible(false); - } + ui->udp_config_group_box->setVisible(true); } void ConfigureMotionTouch::ConnectEvents() { - connect(ui->motion_provider, qOverload<int>(&QComboBox::currentIndexChanged), this, - [this](int index) { UpdateUiDisplay(); }); connect(ui->touch_provider, qOverload<int>(&QComboBox::currentIndexChanged), this, [this](int index) { UpdateUiDisplay(); }); connect(ui->udp_test, &QPushButton::clicked, this, &ConfigureMotionTouch::OnCemuhookUDPTest); + connect(ui->udp_add, &QPushButton::clicked, this, &ConfigureMotionTouch::OnUDPAddServer); + connect(ui->udp_remove, &QPushButton::clicked, this, &ConfigureMotionTouch::OnUDPDeleteServer); connect(ui->touch_calibration_config, &QPushButton::clicked, this, &ConfigureMotionTouch::OnConfigureTouchCalibration); connect(ui->touch_from_button_config_btn, &QPushButton::clicked, this, @@ -187,13 +184,58 @@ void ConfigureMotionTouch::ConnectEvents() { }); } +void ConfigureMotionTouch::OnUDPAddServer() { + QRegExp re(tr(R"re(^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4]" + "[0-9]|[01]?[0-9][0-9]?)$)re")); // a valid ip address + bool ok; + QString port_text = ui->udp_port->text(); + QString server_text = ui->udp_server->text(); + const QString server_string = tr("%1:%2").arg(server_text, port_text); + int port_number = port_text.toInt(&ok, 10); + int row = udp_server_list_model->rowCount(); + + if (!ok) { + QMessageBox::warning(this, tr("yuzu"), tr("Port number has invalid characters")); + return; + } + if (port_number < 0 || port_number > 65353) { + QMessageBox::warning(this, tr("yuzu"), tr("Port has to be in range 0 and 65353")); + return; + } + if (!re.exactMatch(server_text)) { + QMessageBox::warning(this, tr("yuzu"), tr("IP address is not valid")); + return; + } + // Search for duplicates + for (const auto& item : udp_server_list_model->stringList()) { + if (item == server_string) { + QMessageBox::warning(this, tr("yuzu"), tr("This UDP server already exists")); + return; + } + } + // Limit server count to 8 + if (row == 8) { + QMessageBox::warning(this, tr("yuzu"), tr("Unable to add more than 8 servers")); + return; + } + + udp_server_list_model->insertRows(row, 1); + QModelIndex index = udp_server_list_model->index(row); + udp_server_list_model->setData(index, server_string); + ui->udp_server_list->setCurrentIndex(index); +} + +void ConfigureMotionTouch::OnUDPDeleteServer() { + udp_server_list_model->removeRows(ui->udp_server_list->currentIndex().row(), 1); +} + void ConfigureMotionTouch::OnCemuhookUDPTest() { ui->udp_test->setEnabled(false); ui->udp_test->setText(tr("Testing")); udp_test_in_progress = true; InputCommon::CemuhookUDP::TestCommunication( - ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toInt()), - static_cast<u8>(ui->udp_pad_index->currentIndex()), 24872, + ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toInt()), 0, + 24872, [this] { LOG_INFO(Frontend, "UDP input test success"); QMetaObject::invokeMethod(this, "ShowUDPTestResult", Q_ARG(bool, true)); @@ -207,9 +249,9 @@ void ConfigureMotionTouch::OnCemuhookUDPTest() { void ConfigureMotionTouch::OnConfigureTouchCalibration() { ui->touch_calibration_config->setEnabled(false); ui->touch_calibration_config->setText(tr("Configuring")); - CalibrationConfigurationDialog dialog( - this, ui->udp_server->text().toStdString(), static_cast<u16>(ui->udp_port->text().toUInt()), - static_cast<u8>(ui->udp_pad_index->currentIndex()), 24872); + CalibrationConfigurationDialog dialog(this, ui->udp_server->text().toStdString(), + static_cast<u16>(ui->udp_port->text().toUInt()), 0, + 24872); dialog.exec(); if (dialog.completed) { min_x = dialog.min_x; @@ -269,7 +311,7 @@ void ConfigureMotionTouch::OnConfigureTouchFromButton() { bool ConfigureMotionTouch::CanCloseDialog() { if (udp_test_in_progress) { - QMessageBox::warning(this, tr("Citra"), + QMessageBox::warning(this, tr("yuzu"), tr("UDP Test or calibration configuration is in progress.<br>Please " "wait for them to finish.")); return false; @@ -282,17 +324,11 @@ void ConfigureMotionTouch::ApplyConfiguration() { return; } - std::string motion_engine = ui->motion_provider->currentData().toString().toStdString(); std::string touch_engine = ui->touch_provider->currentData().toString().toStdString(); - Common::ParamPackage motion_param{}, touch_param{}; - motion_param.Set("engine", std::move(motion_engine)); + Common::ParamPackage touch_param{}; touch_param.Set("engine", std::move(touch_engine)); - if (motion_engine == "motion_emu") { - motion_param.Set("sensitivity", static_cast<float>(ui->motion_sensitivity->value())); - } - if (touch_engine == "cemuhookudp") { touch_param.Set("min_x", min_x); touch_param.Set("min_y", min_y); @@ -300,15 +336,25 @@ void ConfigureMotionTouch::ApplyConfiguration() { touch_param.Set("max_y", max_y); } - Settings::values.motion_device = motion_param.Serialize(); Settings::values.touch_device = touch_param.Serialize(); Settings::values.use_touch_from_button = ui->touch_from_button_checkbox->isChecked(); Settings::values.touch_from_button_map_index = ui->touch_from_button_map->currentIndex(); Settings::values.touch_from_button_maps = touch_from_button_maps; - Settings::values.udp_input_address = ui->udp_server->text().toStdString(); - Settings::values.udp_input_port = static_cast<u16>(ui->udp_port->text().toInt()); - Settings::values.udp_pad_index = static_cast<u8>(ui->udp_pad_index->currentIndex()); + Settings::values.udp_input_servers = GetUDPServerString(); input_subsystem->ReloadInputDevices(); accept(); } + +std::string ConfigureMotionTouch::GetUDPServerString() const { + QString input_servers; + + for (const auto& item : udp_server_list_model->stringList()) { + input_servers += item; + input_servers += QLatin1Char{','}; + } + + // Remove last comma + input_servers.chop(1); + return input_servers.toStdString(); +} diff --git a/src/yuzu/configuration/configure_motion_touch.h b/src/yuzu/configuration/configure_motion_touch.h index 3d4b5d659..15d61e8ba 100644 --- a/src/yuzu/configuration/configure_motion_touch.h +++ b/src/yuzu/configuration/configure_motion_touch.h @@ -10,6 +10,7 @@ class QLabel; class QPushButton; +class QStringListModel; class QVBoxLayout; namespace InputCommon { @@ -62,6 +63,8 @@ public slots: void ApplyConfiguration(); private slots: + void OnUDPAddServer(); + void OnUDPDeleteServer(); void OnCemuhookUDPTest(); void OnConfigureTouchCalibration(); void OnConfigureTouchFromButton(); @@ -73,10 +76,12 @@ private: void UpdateUiDisplay(); void ConnectEvents(); bool CanCloseDialog(); + std::string GetUDPServerString() const; InputCommon::InputSubsystem* input_subsystem; std::unique_ptr<Ui::ConfigureMotionTouch> ui; + QStringListModel* udp_server_list_model; // Coordinate system of the CemuhookUDP touch provider int min_x{}; diff --git a/src/yuzu/configuration/configure_motion_touch.ui b/src/yuzu/configuration/configure_motion_touch.ui index 602cf8cd8..ebca835ac 100644 --- a/src/yuzu/configuration/configure_motion_touch.ui +++ b/src/yuzu/configuration/configure_motion_touch.ui @@ -2,41 +2,30 @@ <ui version="4.0"> <class>ConfigureMotionTouch</class> <widget class="QDialog" name="ConfigureMotionTouch"> - <property name="windowTitle"> - <string>Configure Motion / Touch</string> - </property> <property name="geometry"> <rect> <x>0</x> <y>0</y> <width>500</width> - <height>450</height> + <height>482</height> </rect> </property> + <property name="windowTitle"> + <string>Configure Motion / Touch</string> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> <layout class="QVBoxLayout"> <item> <widget class="QGroupBox" name="motion_group_box"> <property name="title"> - <string>Motion</string> + <string>Mouse Motion</string> </property> <layout class="QVBoxLayout"> <item> <layout class="QHBoxLayout"> <item> - <widget class="QLabel" name="motion_provider_label"> - <property name="text"> - <string>Motion Provider:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="motion_provider"/> - </item> - </layout> - </item> - <item> - <layout class="QHBoxLayout"> - <item> <widget class="QLabel" name="motion_sensitivity_label"> <property name="text"> <string>Sensitivity:</string> @@ -180,103 +169,171 @@ </widget> </item> <item> - <layout class="QHBoxLayout"> + <layout class="QHBoxLayout" name="horizontalLayout"> <item> - <widget class="QLabel" name="udp_server_label"> - <property name="text"> - <string>Server:</string> - </property> - </widget> + <widget class="QListView" name="udp_server_list"/> </item> <item> - <widget class="QLineEdit" name="udp_server"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> + <layout class="QVBoxLayout" name="verticalLayout"> + <property name="leftMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <layout class="QHBoxLayout"> - <item> - <widget class="QLabel" name="udp_port_label"> - <property name="text"> - <string>Port:</string> + <property name="topMargin"> + <number>0</number> </property> - </widget> - </item> - <item> - <widget class="QLineEdit" name="udp_port"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> + <property name="rightMargin"> + <number>0</number> </property> - </widget> - </item> - </layout> - </item> - <item> - <layout class="QHBoxLayout"> - <item> - <widget class="QLabel" name="udp_pad_index_label"> - <property name="text"> - <string>Pad:</string> + <property name="bottomMargin"> + <number>0</number> </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="udp_pad_index"> <item> - <property name="text"> - <string>Pad 1</string> - </property> + <layout class="QHBoxLayout"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="udp_server_label"> + <property name="text"> + <string>Server:</string> + </property> + </widget> + </item> + <item> + <widget class="QLineEdit" name="udp_server"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + </widget> + </item> + </layout> </item> <item> - <property name="text"> - <string>Pad 2</string> - </property> + <layout class="QHBoxLayout"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="udp_port_label"> + <property name="text"> + <string>Port:</string> + </property> + </widget> + </item> + <item> + <widget class="QLineEdit" name="udp_port"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Minimum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + </widget> + </item> + </layout> </item> <item> - <property name="text"> - <string>Pad 3</string> - </property> + <layout class="QHBoxLayout"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <item> + <widget class="QLabel" name="udp_learn_more"> + <property name="text"> + <string>Learn More</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="udp_test"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Test</string> + </property> + </widget> + </item> + <item> + <widget class="QPushButton" name="udp_add"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Add Server</string> + </property> + </widget> + </item> + </layout> </item> <item> - <property name="text"> - <string>Pad 4</string> - </property> + <spacer name="verticalSpacer_3"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> </item> - </widget> - </item> - </layout> - </item> - <item> - <layout class="QHBoxLayout"> - <item> - <widget class="QLabel" name="udp_learn_more"> - <property name="text"> - <string>Learn More</string> - </property> - </widget> - </item> - <item> - <widget class="QPushButton" name="udp_test"> - <property name="sizePolicy"> - <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> - <horstretch>0</horstretch> - <verstretch>0</verstretch> - </sizepolicy> - </property> - <property name="text"> - <string>Test</string> - </property> - </widget> + <item> + <layout class="QHBoxLayout" name="horizontalLayout_2"> + <property name="topMargin"> + <number>0</number> + </property> + <item> + <widget class="QPushButton" name="udp_remove"> + <property name="sizePolicy"> + <sizepolicy hsizetype="Maximum" vsizetype="Fixed"> + <horstretch>0</horstretch> + <verstretch>0</verstretch> + </sizepolicy> + </property> + <property name="text"> + <string>Remove Server</string> + </property> + </widget> + </item> + <item> + <spacer name="horizontalSpacer_2"> + <property name="orientation"> + <enum>Qt::Horizontal</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>40</width> + <height>20</height> + </size> + </property> + </spacer> + </item> + </layout> + </item> + </layout> </item> </layout> </item> @@ -314,12 +371,12 @@ <slot>ApplyConfiguration()</slot> <hints> <hint type="sourcelabel"> - <x>220</x> - <y>380</y> + <x>20</x> + <y>20</y> </hint> <hint type="destinationlabel"> - <x>220</x> - <y>200</y> + <x>20</x> + <y>20</y> </hint> </hints> </connection> diff --git a/src/yuzu/configuration/configure_mouse_advanced.ui b/src/yuzu/configuration/configure_mouse_advanced.ui index 74552fdbd..5b99e1c37 100644 --- a/src/yuzu/configuration/configure_mouse_advanced.ui +++ b/src/yuzu/configuration/configure_mouse_advanced.ui @@ -15,7 +15,7 @@ </property> <property name="styleSheet"> <string notr="true">QPushButton { - min-width: 55px; + min-width: 60px; }</string> </property> <layout class="QVBoxLayout" name="verticalLayout"> @@ -42,13 +42,13 @@ <widget class="QPushButton" name="forward_button"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>16777215</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -82,7 +82,7 @@ <widget class="QPushButton" name="back_button"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> @@ -110,7 +110,7 @@ <widget class="QPushButton" name="left_button"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> @@ -138,13 +138,13 @@ <widget class="QPushButton" name="middle_button"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>16777215</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -204,13 +204,13 @@ <widget class="QPushButton" name="right_button"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>16777215</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -256,13 +256,13 @@ <widget class="QPushButton" name="buttonClearAll"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>16777215</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -275,13 +275,13 @@ <widget class="QPushButton" name="buttonRestoreDefaults"> <property name="minimumSize"> <size> - <width>57</width> + <width>68</width> <height>0</height> </size> </property> <property name="maximumSize"> <size> - <width>16777215</width> + <width>68</width> <height>16777215</height> </size> </property> @@ -324,32 +324,12 @@ <signal>accepted()</signal> <receiver>ConfigureMouseAdvanced</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>124</x> - <y>266</y> - </hint> - <hint type="destinationlabel"> - <x>124</x> - <y>143</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigureMouseAdvanced</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>124</x> - <y>266</y> - </hint> - <hint type="destinationlabel"> - <x>124</x> - <y>143</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_per_game.cpp b/src/yuzu/configuration/configure_per_game.cpp index 1e49f0787..f598513df 100644 --- a/src/yuzu/configuration/configure_per_game.cpp +++ b/src/yuzu/configuration/configure_per_game.cpp @@ -16,6 +16,7 @@ #include "common/common_paths.h" #include "common/file_util.h" +#include "core/core.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" #include "core/file_sys/xts_archive.h" @@ -29,9 +30,10 @@ ConfigurePerGame::ConfigurePerGame(QWidget* parent, u64 title_id) : QDialog(parent), ui(std::make_unique<Ui::ConfigurePerGame>()), title_id(title_id) { - game_config = std::make_unique<Config>(fmt::format("{:016X}.ini", title_id), false); + game_config = std::make_unique<Config>(fmt::format("{:016X}", title_id), + Config::ConfigType::PerGameConfig); - Settings::configuring_global = false; + Settings::SetConfiguringGlobal(false); ui->setupUi(this); setFocusPolicy(Qt::ClickFocus); @@ -55,7 +57,7 @@ void ConfigurePerGame::ApplyConfiguration() { ui->graphicsAdvancedTab->ApplyConfiguration(); ui->audioTab->ApplyConfiguration(); - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); Settings::LogSettings(); game_config->Save(); @@ -88,9 +90,11 @@ void ConfigurePerGame::LoadConfiguration() { ui->display_title_id->setText( QStringLiteral("%1").arg(title_id, 16, 16, QLatin1Char{'0'}).toUpper()); - FileSys::PatchManager pm{title_id}; + auto& system = Core::System::GetInstance(); + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto control = pm.GetControlMetadata(); - const auto loader = Loader::GetLoader(file); + const auto loader = Loader::GetLoader(system, file); if (control.first != nullptr) { ui->display_version->setText(QString::fromStdString(control.first->GetVersionString())); diff --git a/src/yuzu/configuration/configure_per_game.ui b/src/yuzu/configuration/configure_per_game.ui index d2057c4ab..25975b3b9 100644 --- a/src/yuzu/configuration/configure_per_game.ui +++ b/src/yuzu/configuration/configure_per_game.ui @@ -319,32 +319,12 @@ <signal>accepted()</signal> <receiver>ConfigurePerGame</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>248</x> - <y>254</y> - </hint> - <hint type="destinationlabel"> - <x>157</x> - <y>274</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigurePerGame</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>316</x> - <y>260</y> - </hint> - <hint type="destinationlabel"> - <x>286</x> - <y>274</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_per_game_addons.cpp b/src/yuzu/configuration/configure_per_game_addons.cpp index 793fd8975..cdeeec01c 100644 --- a/src/yuzu/configuration/configure_per_game_addons.cpp +++ b/src/yuzu/configuration/configure_per_game_addons.cpp @@ -112,8 +112,10 @@ void ConfigurePerGameAddons::LoadConfiguration() { return; } - FileSys::PatchManager pm{title_id}; - const auto loader = Loader::GetLoader(file); + auto& system = Core::System::GetInstance(); + const FileSys::PatchManager pm{title_id, system.GetFileSystemController(), + system.GetContentProvider()}; + const auto loader = Loader::GetLoader(system, file); FileSys::VirtualFile update_raw; loader->ReadUpdateRaw(update_raw); diff --git a/src/yuzu/configuration/configure_profile_manager.cpp b/src/yuzu/configuration/configure_profile_manager.cpp index 6334c4c50..13d9a4757 100644 --- a/src/yuzu/configuration/configure_profile_manager.cpp +++ b/src/yuzu/configuration/configure_profile_manager.cpp @@ -180,7 +180,7 @@ void ConfigureProfileManager::ApplyConfiguration() { return; } - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); } void ConfigureProfileManager::SelectUser(const QModelIndex& index) { diff --git a/src/yuzu/configuration/configure_system.cpp b/src/yuzu/configuration/configure_system.cpp index 9ad43ed8f..6cf2032da 100644 --- a/src/yuzu/configuration/configure_system.cpp +++ b/src/yuzu/configuration/configure_system.cpp @@ -12,6 +12,7 @@ #include "common/assert.h" #include "common/file_util.h" #include "core/core.h" +#include "core/hle/service/time/time.h" #include "core/settings.h" #include "ui_configure_system.h" #include "yuzu/configuration/configuration_shared.h" @@ -36,8 +37,8 @@ ConfigureSystem::ConfigureSystem(QWidget* parent) : QWidget(parent), ui(new Ui:: } }); - ui->label_console_id->setVisible(Settings::configuring_global); - ui->button_regenerate_console_id->setVisible(Settings::configuring_global); + ui->label_console_id->setVisible(Settings::IsConfiguringGlobal()); + ui->button_regenerate_console_id->setVisible(Settings::IsConfiguringGlobal()); SetupPerGameUI(); @@ -77,7 +78,7 @@ void ConfigureSystem::SetConfiguration() { Settings::values.rng_seed.UsingGlobal()); ui->custom_rtc_edit->setDateTime(QDateTime::fromSecsSinceEpoch(rtc_time.count())); - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->combo_language->setCurrentIndex(Settings::values.language_index.GetValue()); ui->combo_region->setCurrentIndex(Settings::values.region_index.GetValue()); ui->combo_time_zone->setCurrentIndex(Settings::values.time_zone_index.GetValue()); @@ -104,11 +105,29 @@ void ConfigureSystem::SetConfiguration() { void ConfigureSystem::ReadSystemSettings() {} void ConfigureSystem::ApplyConfiguration() { + auto& system = Core::System::GetInstance(); + + // Allow setting custom RTC even if system is powered on, + // to allow in-game time to be fast forwarded + if (Settings::values.custom_rtc.UsingGlobal()) { + if (ui->custom_rtc_checkbox->isChecked()) { + Settings::values.custom_rtc.SetValue( + std::chrono::seconds(ui->custom_rtc_edit->dateTime().toSecsSinceEpoch())); + if (system.IsPoweredOn()) { + const s64 posix_time{Settings::values.custom_rtc.GetValue()->count() + + Service::Time::TimeManager::GetExternalTimeZoneOffset()}; + system.GetTimeManager().UpdateLocalSystemClockTime(posix_time); + } + } else { + Settings::values.custom_rtc.SetValue(std::nullopt); + } + } + if (!enabled) { return; } - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { // Guard if during game and set to game-specific value if (Settings::values.language_index.UsingGlobal()) { Settings::values.language_index.SetValue(ui->combo_language->currentIndex()); @@ -131,15 +150,6 @@ void ConfigureSystem::ApplyConfiguration() { Settings::values.rng_seed.SetValue(std::nullopt); } } - - if (Settings::values.custom_rtc.UsingGlobal()) { - if (ui->custom_rtc_checkbox->isChecked()) { - Settings::values.custom_rtc.SetValue( - std::chrono::seconds(ui->custom_rtc_edit->dateTime().toSecsSinceEpoch())); - } else { - Settings::values.custom_rtc.SetValue(std::nullopt); - } - } } else { ConfigurationShared::ApplyPerGameSetting(&Settings::values.language_index, ui->combo_language); @@ -189,7 +199,7 @@ void ConfigureSystem::ApplyConfiguration() { } } - Settings::Apply(); + Settings::Apply(system); } void ConfigureSystem::RefreshConsoleID() { @@ -210,7 +220,7 @@ void ConfigureSystem::RefreshConsoleID() { } void ConfigureSystem::SetupPerGameUI() { - if (Settings::configuring_global) { + if (Settings::IsConfiguringGlobal()) { ui->combo_language->setEnabled(Settings::values.language_index.UsingGlobal()); ui->combo_region->setEnabled(Settings::values.region_index.UsingGlobal()); ui->combo_time_zone->setEnabled(Settings::values.time_zone_index.UsingGlobal()); diff --git a/src/yuzu/configuration/configure_touch_from_button.ui b/src/yuzu/configuration/configure_touch_from_button.ui index f581e27e0..757219d54 100644 --- a/src/yuzu/configuration/configure_touch_from_button.ui +++ b/src/yuzu/configuration/configure_touch_from_button.ui @@ -216,16 +216,6 @@ Drag points to change position, or double-click table cells to edit values.</str <signal>rejected()</signal> <receiver>ConfigureTouchFromButton</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>249</x> - <y>428</y> - </hint> - <hint type="destinationlabel"> - <x>249</x> - <y>224</y> - </hint> - </hints> </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_touchscreen_advanced.ui b/src/yuzu/configuration/configure_touchscreen_advanced.ui index 1171c2dd1..30ceccddb 100644 --- a/src/yuzu/configuration/configure_touchscreen_advanced.ui +++ b/src/yuzu/configuration/configure_touchscreen_advanced.ui @@ -168,32 +168,12 @@ <signal>accepted()</signal> <receiver>ConfigureTouchscreenAdvanced</receiver> <slot>accept()</slot> - <hints> - <hint type="sourcelabel"> - <x>140</x> - <y>318</y> - </hint> - <hint type="destinationlabel"> - <x>140</x> - <y>169</y> - </hint> - </hints> </connection> <connection> <sender>buttonBox</sender> <signal>rejected()</signal> <receiver>ConfigureTouchscreenAdvanced</receiver> <slot>reject()</slot> - <hints> - <hint type="sourcelabel"> - <x>140</x> - <y>318</y> - </hint> - <hint type="destinationlabel"> - <x>140</x> - <y>169</y> - </hint> - </hints> - </connection> + </connection> </connections> </ui> diff --git a/src/yuzu/configuration/configure_ui.cpp b/src/yuzu/configuration/configure_ui.cpp index dbe3f78c8..aed876008 100644 --- a/src/yuzu/configuration/configure_ui.cpp +++ b/src/yuzu/configuration/configure_ui.cpp @@ -9,6 +9,7 @@ #include <QDirIterator> #include "common/common_types.h" #include "common/file_util.h" +#include "core/core.h" #include "core/settings.h" #include "ui_configure_ui.h" #include "yuzu/configuration/configure_ui.h" @@ -84,7 +85,7 @@ void ConfigureUi::ApplyConfiguration() { UISettings::values.enable_screenshot_save_as = ui->enable_screenshot_save_as->isChecked(); Common::FS::GetUserPath(Common::FS::UserPath::ScreenshotsDir, ui->screenshot_path_edit->text().toStdString()); - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); } void ConfigureUi::RequestGameListUpdate() { diff --git a/src/yuzu/configuration/configure_vibration.cpp b/src/yuzu/configuration/configure_vibration.cpp new file mode 100644 index 000000000..7dcb2c5b9 --- /dev/null +++ b/src/yuzu/configuration/configure_vibration.cpp @@ -0,0 +1,146 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <unordered_map> + +#include <fmt/format.h> + +#include "common/param_package.h" +#include "core/settings.h" +#include "ui_configure_vibration.h" +#include "yuzu/configuration/configure_vibration.h" + +ConfigureVibration::ConfigureVibration(QWidget* parent) + : QDialog(parent), ui(std::make_unique<Ui::ConfigureVibration>()) { + ui->setupUi(this); + + vibration_groupboxes = { + ui->vibrationGroupPlayer1, ui->vibrationGroupPlayer2, ui->vibrationGroupPlayer3, + ui->vibrationGroupPlayer4, ui->vibrationGroupPlayer5, ui->vibrationGroupPlayer6, + ui->vibrationGroupPlayer7, ui->vibrationGroupPlayer8, + }; + + vibration_spinboxes = { + ui->vibrationSpinPlayer1, ui->vibrationSpinPlayer2, ui->vibrationSpinPlayer3, + ui->vibrationSpinPlayer4, ui->vibrationSpinPlayer5, ui->vibrationSpinPlayer6, + ui->vibrationSpinPlayer7, ui->vibrationSpinPlayer8, + }; + + const auto& players = Settings::values.players.GetValue(); + + for (std::size_t i = 0; i < NUM_PLAYERS; ++i) { + vibration_groupboxes[i]->setChecked(players[i].vibration_enabled); + vibration_spinboxes[i]->setValue(players[i].vibration_strength); + } + + ui->checkBoxAccurateVibration->setChecked( + Settings::values.enable_accurate_vibrations.GetValue()); + + if (!Settings::IsConfiguringGlobal()) { + ui->checkBoxAccurateVibration->setDisabled(true); + } + + RetranslateUI(); +} + +ConfigureVibration::~ConfigureVibration() = default; + +void ConfigureVibration::ApplyConfiguration() { + auto& players = Settings::values.players.GetValue(); + + for (std::size_t i = 0; i < NUM_PLAYERS; ++i) { + players[i].vibration_enabled = vibration_groupboxes[i]->isChecked(); + players[i].vibration_strength = vibration_spinboxes[i]->value(); + } + + Settings::values.enable_accurate_vibrations.SetValue( + ui->checkBoxAccurateVibration->isChecked()); +} + +void ConfigureVibration::SetVibrationDevices(std::size_t player_index) { + using namespace Settings::NativeButton; + static constexpr std::array<std::array<Settings::NativeButton::Values, 6>, 2> buttons{{ + {DLeft, DUp, DRight, DDown, L, ZL}, // Left Buttons + {A, B, X, Y, R, ZR}, // Right Buttons + }}; + + auto& player = Settings::values.players.GetValue()[player_index]; + + for (std::size_t device_idx = 0; device_idx < buttons.size(); ++device_idx) { + std::unordered_map<std::string, int> params_count; + + for (const auto button_index : buttons[device_idx]) { + const auto& player_button = player.buttons[button_index]; + + if (params_count.find(player_button) != params_count.end()) { + ++params_count[player_button]; + continue; + } + + params_count.insert_or_assign(player_button, 1); + } + + const auto it = std::max_element( + params_count.begin(), params_count.end(), + [](const auto& lhs, const auto& rhs) { return lhs.second < rhs.second; }); + + auto& vibration_param_str = player.vibrations[device_idx]; + vibration_param_str.clear(); + + if (it->first.empty()) { + continue; + } + + const auto param = Common::ParamPackage(it->first); + + const auto engine = param.Get("engine", ""); + const auto guid = param.Get("guid", ""); + const auto port = param.Get("port", ""); + + if (engine.empty() || engine == "keyboard" || engine == "mouse") { + continue; + } + + vibration_param_str += fmt::format("engine:{}", engine); + + if (!port.empty()) { + vibration_param_str += fmt::format(",port:{}", port); + } + if (!guid.empty()) { + vibration_param_str += fmt::format(",guid:{}", guid); + } + } + + if (player.vibrations[0] != player.vibrations[1]) { + return; + } + + if (!player.vibrations[0].empty() && + player.controller_type != Settings::ControllerType::RightJoycon) { + player.vibrations[1].clear(); + } else if (!player.vibrations[1].empty() && + player.controller_type == Settings::ControllerType::RightJoycon) { + player.vibrations[0].clear(); + } +} + +void ConfigureVibration::SetAllVibrationDevices() { + // Set vibration devices for all player indices including handheld + for (std::size_t player_idx = 0; player_idx < NUM_PLAYERS + 1; ++player_idx) { + SetVibrationDevices(player_idx); + } +} + +void ConfigureVibration::changeEvent(QEvent* event) { + if (event->type() == QEvent::LanguageChange) { + RetranslateUI(); + } + + QDialog::changeEvent(event); +} + +void ConfigureVibration::RetranslateUI() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_vibration.h b/src/yuzu/configuration/configure_vibration.h new file mode 100644 index 000000000..07411a86f --- /dev/null +++ b/src/yuzu/configuration/configure_vibration.h @@ -0,0 +1,43 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <memory> +#include <QDialog> + +class QGroupBox; +class QSpinBox; + +namespace Ui { +class ConfigureVibration; +} + +class ConfigureVibration : public QDialog { + Q_OBJECT + +public: + explicit ConfigureVibration(QWidget* parent); + ~ConfigureVibration() override; + + void ApplyConfiguration(); + + static void SetVibrationDevices(std::size_t player_index); + static void SetAllVibrationDevices(); + +private: + void changeEvent(QEvent* event) override; + void RetranslateUI(); + + std::unique_ptr<Ui::ConfigureVibration> ui; + + static constexpr std::size_t NUM_PLAYERS = 8; + + // Groupboxes encapsulating the vibration strength spinbox. + std::array<QGroupBox*, NUM_PLAYERS> vibration_groupboxes; + + // Spinboxes representing the vibration strength percentage. + std::array<QSpinBox*, NUM_PLAYERS> vibration_spinboxes; +}; diff --git a/src/yuzu/configuration/configure_vibration.ui b/src/yuzu/configuration/configure_vibration.ui new file mode 100644 index 000000000..efdf317a9 --- /dev/null +++ b/src/yuzu/configuration/configure_vibration.ui @@ -0,0 +1,546 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureVibration</class> + <widget class="QDialog" name="ConfigureVibration"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>364</width> + <height>242</height> + </rect> + </property> + <property name="windowTitle"> + <string>Configure Vibration</string> + </property> + <property name="styleSheet"> + <string notr="true"/> + </property> + <layout class="QVBoxLayout"> + <item> + <widget class="QGroupBox" name="vibrationStrengthGroup"> + <property name="title"> + <string>Vibration</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout_3" stretch="0,0"> + <property name="leftMargin"> + <number>9</number> + </property> + <property name="topMargin"> + <number>9</number> + </property> + <property name="rightMargin"> + <number>9</number> + </property> + <property name="bottomMargin"> + <number>9</number> + </property> + <item> + <widget class="QWidget" name="player14Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_4"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer1"> + <property name="title"> + <string>Player 1</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_8"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer1"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer2"> + <property name="title"> + <string>Player 2</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_9"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer2"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer3"> + <property name="title"> + <string>Player 3</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_10"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer3"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer4"> + <property name="title"> + <string>Player 4</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_11"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer4"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QWidget" name="player58Widget" native="true"> + <layout class="QHBoxLayout" name="horizontalLayout_6"> + <property name="leftMargin"> + <number>0</number> + </property> + <property name="topMargin"> + <number>0</number> + </property> + <property name="rightMargin"> + <number>0</number> + </property> + <property name="bottomMargin"> + <number>0</number> + </property> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer7"> + <property name="title"> + <string>Player 5</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_14"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer7"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer8"> + <property name="title"> + <string>Player 6</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_15"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer8"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer5"> + <property name="title"> + <string>Player 7</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_12"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer5"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationGroupPlayer6"> + <property name="title"> + <string>Player 8</string> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QHBoxLayout" name="horizontalLayout_13"> + <property name="leftMargin"> + <number>3</number> + </property> + <property name="topMargin"> + <number>3</number> + </property> + <property name="rightMargin"> + <number>3</number> + </property> + <property name="bottomMargin"> + <number>3</number> + </property> + <item> + <widget class="QSpinBox" name="vibrationSpinPlayer6"> + <property name="minimumSize"> + <size> + <width>68</width> + <height>21</height> + </size> + </property> + <property name="maximumSize"> + <size> + <width>68</width> + <height>16777215</height> + </size> + </property> + <property name="suffix"> + <string>%</string> + </property> + <property name="minimum"> + <number>1</number> + </property> + <property name="maximum"> + <number>150</number> + </property> + <property name="value"> + <number>100</number> + </property> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <widget class="QGroupBox" name="vibrationSettingsGroup"> + <property name="title"> + <string>Settings</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <item> + <widget class="QCheckBox" name="checkBoxAccurateVibration"> + <property name="text"> + <string>Enable Accurate Vibration</string> + </property> + </widget> + </item> + </layout> + </widget> + </item> + <item> + <spacer name="spacerVibration"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>167</width> + <height>55</height> + </size> + </property> + </spacer> + </item> + <item> + <widget class="QDialogButtonBox" name="buttonBoxVibration"> + <property name="standardButtons"> + <set>QDialogButtonBox::Cancel|QDialogButtonBox::Ok</set> + </property> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections> + <connection> + <sender>buttonBoxVibration</sender> + <signal>accepted()</signal> + <receiver>ConfigureVibration</receiver> + <slot>accept()</slot> + </connection> + <connection> + <sender>buttonBoxVibration</sender> + <signal>rejected()</signal> + <receiver>ConfigureVibration</receiver> + <slot>reject()</slot> + </connection> + </connections> +</ui> diff --git a/src/yuzu/configuration/input_profiles.cpp b/src/yuzu/configuration/input_profiles.cpp new file mode 100644 index 000000000..e87aededb --- /dev/null +++ b/src/yuzu/configuration/input_profiles.cpp @@ -0,0 +1,131 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <fmt/format.h> + +#include "common/common_paths.h" +#include "common/file_util.h" +#include "yuzu/configuration/config.h" +#include "yuzu/configuration/input_profiles.h" + +namespace FS = Common::FS; + +namespace { + +bool ProfileExistsInFilesystem(std::string_view profile_name) { + return FS::Exists(fmt::format("{}input" DIR_SEP "{}.ini", + FS::GetUserPath(FS::UserPath::ConfigDir), profile_name)); +} + +bool IsINI(std::string_view filename) { + const std::size_t index = filename.rfind('.'); + + if (index == std::string::npos) { + return false; + } + + return filename.substr(index) == ".ini"; +} + +std::string GetNameWithoutExtension(const std::string& filename) { + const std::size_t index = filename.rfind('.'); + + if (index == std::string::npos) { + return filename; + } + + return filename.substr(0, index); +} + +} // namespace + +InputProfiles::InputProfiles() { + const std::string input_profile_loc = + fmt::format("{}input", FS::GetUserPath(FS::UserPath::ConfigDir)); + + FS::ForeachDirectoryEntry( + nullptr, input_profile_loc, + [this](u64* entries_out, const std::string& directory, const std::string& filename) { + if (IsINI(filename) && IsProfileNameValid(GetNameWithoutExtension(filename))) { + map_profiles.insert_or_assign( + GetNameWithoutExtension(filename), + std::make_unique<Config>(GetNameWithoutExtension(filename), + Config::ConfigType::InputProfile)); + } + return true; + }); +} + +InputProfiles::~InputProfiles() = default; + +std::vector<std::string> InputProfiles::GetInputProfileNames() { + std::vector<std::string> profile_names; + profile_names.reserve(map_profiles.size()); + + for (const auto& [profile_name, config] : map_profiles) { + if (!ProfileExistsInFilesystem(profile_name)) { + DeleteProfile(profile_name); + continue; + } + + profile_names.push_back(profile_name); + } + + return profile_names; +} + +bool InputProfiles::IsProfileNameValid(std::string_view profile_name) { + return profile_name.find_first_of("<>:;\"/\\|,.!?*") == std::string::npos; +} + +bool InputProfiles::CreateProfile(const std::string& profile_name, std::size_t player_index) { + if (ProfileExistsInMap(profile_name)) { + return false; + } + + map_profiles.insert_or_assign( + profile_name, std::make_unique<Config>(profile_name, Config::ConfigType::InputProfile)); + + return SaveProfile(profile_name, player_index); +} + +bool InputProfiles::DeleteProfile(const std::string& profile_name) { + if (!ProfileExistsInMap(profile_name)) { + return false; + } + + if (!ProfileExistsInFilesystem(profile_name) || + FS::Delete(map_profiles[profile_name]->GetConfigFilePath())) { + map_profiles.erase(profile_name); + } + + return !ProfileExistsInMap(profile_name) && !ProfileExistsInFilesystem(profile_name); +} + +bool InputProfiles::LoadProfile(const std::string& profile_name, std::size_t player_index) { + if (!ProfileExistsInMap(profile_name)) { + return false; + } + + if (!ProfileExistsInFilesystem(profile_name)) { + map_profiles.erase(profile_name); + return false; + } + + map_profiles[profile_name]->ReadControlPlayerValue(player_index); + return true; +} + +bool InputProfiles::SaveProfile(const std::string& profile_name, std::size_t player_index) { + if (!ProfileExistsInMap(profile_name)) { + return false; + } + + map_profiles[profile_name]->SaveControlPlayerValue(player_index); + return true; +} + +bool InputProfiles::ProfileExistsInMap(const std::string& profile_name) const { + return map_profiles.find(profile_name) != map_profiles.end(); +} diff --git a/src/yuzu/configuration/input_profiles.h b/src/yuzu/configuration/input_profiles.h new file mode 100644 index 000000000..cb41fd9be --- /dev/null +++ b/src/yuzu/configuration/input_profiles.h @@ -0,0 +1,32 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <string> +#include <string_view> +#include <unordered_map> + +class Config; + +class InputProfiles { + +public: + explicit InputProfiles(); + virtual ~InputProfiles(); + + std::vector<std::string> GetInputProfileNames(); + + static bool IsProfileNameValid(std::string_view profile_name); + + bool CreateProfile(const std::string& profile_name, std::size_t player_index); + bool DeleteProfile(const std::string& profile_name); + bool LoadProfile(const std::string& profile_name, std::size_t player_index); + bool SaveProfile(const std::string& profile_name, std::size_t player_index); + +private: + bool ProfileExistsInMap(const std::string& profile_name) const; + + std::unordered_map<std::string, std::unique_ptr<Config>> map_profiles; +}; diff --git a/src/yuzu/debugger/profiler.cpp b/src/yuzu/debugger/profiler.cpp index 0e26f765b..efdc6aa50 100644 --- a/src/yuzu/debugger/profiler.cpp +++ b/src/yuzu/debugger/profiler.cpp @@ -48,7 +48,7 @@ private: MicroProfileDialog::MicroProfileDialog(QWidget* parent) : QWidget(parent, Qt::Dialog) { setObjectName(QStringLiteral("MicroProfile")); - setWindowTitle(tr("MicroProfile")); + setWindowTitle(tr("&MicroProfile")); resize(1000, 600); // Remove the "?" button from the titlebar and enable the maximize button setWindowFlags((windowFlags() & ~Qt::WindowContextHelpButtonHint) | diff --git a/src/yuzu/debugger/wait_tree.cpp b/src/yuzu/debugger/wait_tree.cpp index 3439cb333..0925c10b4 100644 --- a/src/yuzu/debugger/wait_tree.cpp +++ b/src/yuzu/debugger/wait_tree.cpp @@ -13,10 +13,10 @@ #include "core/arm/arm_interface.h" #include "core/core.h" #include "core/hle/kernel/handle_table.h" +#include "core/hle/kernel/k_scheduler.h" #include "core/hle/kernel/mutex.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/readable_event.h" -#include "core/hle/kernel/scheduler.h" #include "core/hle/kernel/synchronization_object.h" #include "core/hle/kernel/thread.h" #include "core/memory.h" @@ -25,7 +25,6 @@ namespace { constexpr std::array<std::array<Qt::GlobalColor, 2>, 10> WaitTreeColors{{ {Qt::GlobalColor::darkGreen, Qt::GlobalColor::green}, - {Qt::GlobalColor::darkGreen, Qt::GlobalColor::green}, {Qt::GlobalColor::darkBlue, Qt::GlobalColor::cyan}, {Qt::GlobalColor::lightGray, Qt::GlobalColor::lightGray}, {Qt::GlobalColor::lightGray, Qt::GlobalColor::lightGray}, @@ -102,7 +101,7 @@ std::vector<std::unique_ptr<WaitTreeThread>> WaitTreeItem::MakeThreadItemList() }; const auto& system = Core::System::GetInstance(); - add_threads(system.GlobalScheduler().GetThreadList()); + add_threads(system.GlobalSchedulerContext().GetThreadList()); return item_list; } @@ -239,9 +238,6 @@ QString WaitTreeThread::GetText() const { const auto& thread = static_cast<const Kernel::Thread&>(object); QString status; switch (thread.GetStatus()) { - case Kernel::ThreadStatus::Running: - status = tr("running"); - break; case Kernel::ThreadStatus::Ready: if (!thread.IsPaused()) { if (thread.WasRunning()) { @@ -298,34 +294,32 @@ QColor WaitTreeThread::GetColor() const { const auto& thread = static_cast<const Kernel::Thread&>(object); switch (thread.GetStatus()) { - case Kernel::ThreadStatus::Running: - return QColor(WaitTreeColors[0][color_index]); case Kernel::ThreadStatus::Ready: if (!thread.IsPaused()) { if (thread.WasRunning()) { - return QColor(WaitTreeColors[1][color_index]); + return QColor(WaitTreeColors[0][color_index]); } else { - return QColor(WaitTreeColors[2][color_index]); + return QColor(WaitTreeColors[1][color_index]); } } else { - return QColor(WaitTreeColors[3][color_index]); + return QColor(WaitTreeColors[2][color_index]); } case Kernel::ThreadStatus::Paused: - return QColor(WaitTreeColors[4][color_index]); + return QColor(WaitTreeColors[3][color_index]); case Kernel::ThreadStatus::WaitHLEEvent: case Kernel::ThreadStatus::WaitIPC: - return QColor(WaitTreeColors[5][color_index]); + return QColor(WaitTreeColors[4][color_index]); case Kernel::ThreadStatus::WaitSleep: - return QColor(WaitTreeColors[6][color_index]); + return QColor(WaitTreeColors[5][color_index]); case Kernel::ThreadStatus::WaitSynch: case Kernel::ThreadStatus::WaitMutex: case Kernel::ThreadStatus::WaitCondVar: case Kernel::ThreadStatus::WaitArb: - return QColor(WaitTreeColors[7][color_index]); + return QColor(WaitTreeColors[6][color_index]); case Kernel::ThreadStatus::Dormant: - return QColor(WaitTreeColors[8][color_index]); + return QColor(WaitTreeColors[7][color_index]); case Kernel::ThreadStatus::Dead: - return QColor(WaitTreeColors[9][color_index]); + return QColor(WaitTreeColors[8][color_index]); default: return WaitTreeItem::GetColor(); } @@ -355,14 +349,14 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeThread::GetChildren() const { list.push_back(std::make_unique<WaitTreeText>(tr("processor = %1").arg(processor))); list.push_back( std::make_unique<WaitTreeText>(tr("ideal core = %1").arg(thread.GetIdealCore()))); - list.push_back( - std::make_unique<WaitTreeText>(tr("affinity mask = %1").arg(thread.GetAffinityMask()))); + list.push_back(std::make_unique<WaitTreeText>( + tr("affinity mask = %1").arg(thread.GetAffinityMask().GetAffinityMask()))); list.push_back(std::make_unique<WaitTreeText>(tr("thread id = %1").arg(thread.GetThreadID()))); list.push_back(std::make_unique<WaitTreeText>(tr("priority = %1(current) / %2(normal)") .arg(thread.GetPriority()) .arg(thread.GetNominalPriority()))); list.push_back(std::make_unique<WaitTreeText>( - tr("last running ticks = %1").arg(thread.GetLastRunningTicks()))); + tr("last running ticks = %1").arg(thread.GetLastScheduledTick()))); const VAddr mutex_wait_address = thread.GetMutexWaitAddress(); if (mutex_wait_address != 0) { @@ -463,7 +457,7 @@ void WaitTreeModel::InitItems() { thread_items = WaitTreeItem::MakeThreadItemList(); } -WaitTreeWidget::WaitTreeWidget(QWidget* parent) : QDockWidget(tr("Wait Tree"), parent) { +WaitTreeWidget::WaitTreeWidget(QWidget* parent) : QDockWidget(tr("&Wait Tree"), parent) { setObjectName(QStringLiteral("WaitTreeWidget")); view = new QTreeView(this); view->setHeaderHidden(true); diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp index a9738e298..70d865112 100644 --- a/src/yuzu/game_list.cpp +++ b/src/yuzu/game_list.cpp @@ -25,7 +25,8 @@ #include "yuzu/main.h" #include "yuzu/uisettings.h" -GameListSearchField::KeyReleaseEater::KeyReleaseEater(GameList* gamelist) : gamelist{gamelist} {} +GameListSearchField::KeyReleaseEater::KeyReleaseEater(GameList* gamelist, QObject* parent) + : QObject(parent), gamelist{gamelist} {} // EventFilter in order to process systemkeys while editing the searchfield bool GameListSearchField::KeyReleaseEater::eventFilter(QObject* obj, QEvent* event) { @@ -116,7 +117,7 @@ void GameListSearchField::setFocus() { } GameListSearchField::GameListSearchField(GameList* parent) : QWidget{parent} { - auto* const key_release_eater = new KeyReleaseEater(parent); + auto* const key_release_eater = new KeyReleaseEater(parent, this); layout_filter = new QHBoxLayout; layout_filter->setMargin(8); label_filter = new QLabel; diff --git a/src/yuzu/game_list_p.h b/src/yuzu/game_list_p.h index 92779a9c7..df935022d 100644 --- a/src/yuzu/game_list_p.h +++ b/src/yuzu/game_list_p.h @@ -174,7 +174,8 @@ public: } bool operator<(const QStandardItem& other) const override { - return data(CompatNumberRole) < other.data(CompatNumberRole); + return data(CompatNumberRole).value<QString>() < + other.data(CompatNumberRole).value<QString>(); } }; @@ -330,7 +331,7 @@ public: private: class KeyReleaseEater : public QObject { public: - explicit KeyReleaseEater(GameList* gamelist); + explicit KeyReleaseEater(GameList* gamelist, QObject* parent = nullptr); private: GameList* gamelist = nullptr; diff --git a/src/yuzu/game_list_worker.cpp b/src/yuzu/game_list_worker.cpp index e0ce45fd9..23643aea2 100644 --- a/src/yuzu/game_list_worker.cpp +++ b/src/yuzu/game_list_worker.cpp @@ -235,12 +235,11 @@ GameListWorker::~GameListWorker() = default; void GameListWorker::AddTitlesToGameList(GameListDir* parent_dir) { using namespace FileSys; - const auto& cache = - dynamic_cast<ContentProviderUnion&>(Core::System::GetInstance().GetContentProvider()); + auto& system = Core::System::GetInstance(); + const auto& cache = dynamic_cast<ContentProviderUnion&>(system.GetContentProvider()); - std::vector<std::pair<ContentProviderUnionSlot, ContentProviderEntry>> installed_games; - installed_games = cache.ListEntriesFilterOrigin(std::nullopt, TitleType::Application, - ContentRecordType::Program); + auto installed_games = cache.ListEntriesFilterOrigin(std::nullopt, TitleType::Application, + ContentRecordType::Program); if (parent_dir->type() == static_cast<int>(GameListItemType::SdmcDir)) { installed_games = cache.ListEntriesFilterOrigin( @@ -254,23 +253,27 @@ void GameListWorker::AddTitlesToGameList(GameListDir* parent_dir) { } for (const auto& [slot, game] : installed_games) { - if (slot == ContentProviderUnionSlot::FrontendManual) + if (slot == ContentProviderUnionSlot::FrontendManual) { continue; + } const auto file = cache.GetEntryUnparsed(game.title_id, game.type); - std::unique_ptr<Loader::AppLoader> loader = Loader::GetLoader(file); - if (!loader) + std::unique_ptr<Loader::AppLoader> loader = Loader::GetLoader(system, file); + if (!loader) { continue; + } std::vector<u8> icon; std::string name; u64 program_id = 0; loader->ReadProgramId(program_id); - const PatchManager patch{program_id}; + const PatchManager patch{program_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto control = cache.GetEntry(game.title_id, ContentRecordType::Control); - if (control != nullptr) + if (control != nullptr) { GetMetadataFromControlNCA(patch, *control, icon, name); + } emit EntryReady(MakeGameListEntry(file->GetFullPath(), name, icon, *loader, program_id, compatibility_list, patch), @@ -280,9 +283,11 @@ void GameListWorker::AddTitlesToGameList(GameListDir* parent_dir) { void GameListWorker::ScanFileSystem(ScanTarget target, const std::string& dir_path, unsigned int recursion, GameListDir* parent_dir) { - const auto callback = [this, target, recursion, - parent_dir](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { + auto& system = Core::System::GetInstance(); + + const auto callback = [this, target, recursion, parent_dir, + &system](u64* num_entries_out, const std::string& directory, + const std::string& virtual_name) -> bool { if (stop_processing) { // Breaks the callback loop. return false; @@ -293,7 +298,7 @@ void GameListWorker::ScanFileSystem(ScanTarget target, const std::string& dir_pa if (!is_dir && (HasSupportedFileExtension(physical_name) || IsExtractedNCAMain(physical_name))) { const auto file = vfs->OpenFile(physical_name, FileSys::Mode::Read); - auto loader = Loader::GetLoader(file); + auto loader = Loader::GetLoader(system, file); if (!loader) { return true; } @@ -331,7 +336,8 @@ void GameListWorker::ScanFileSystem(ScanTarget target, const std::string& dir_pa std::string name = " "; [[maybe_unused]] const auto res3 = loader->ReadTitle(name); - const FileSys::PatchManager patch{program_id}; + const FileSys::PatchManager patch{program_id, system.GetFileSystemController(), + system.GetContentProvider()}; emit EntryReady(MakeGameListEntry(physical_name, name, icon, *loader, program_id, compatibility_list, patch), diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index 6a2a88dd8..2e74037d1 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -18,6 +18,7 @@ #include "applets/web_browser.h" #include "configuration/configure_input.h" #include "configuration/configure_per_game.h" +#include "configuration/configure_vibration.h" #include "core/file_sys/vfs.h" #include "core/file_sys/vfs_real.h" #include "core/frontend/applets/controller.h" @@ -27,8 +28,6 @@ #include "core/hle/service/am/applet_ae.h" #include "core/hle/service/am/applet_oe.h" #include "core/hle/service/am/applets/applets.h" -#include "core/hle/service/hid/controllers/npad.h" -#include "core/hle/service/hid/hid.h" // These are wrappers to avoid the calls to CreateDirectory and CreateFile because of the Windows // defines. @@ -50,12 +49,14 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include <QDesktopServices> #include <QDesktopWidget> #include <QDialogButtonBox> +#include <QDir> #include <QFile> #include <QFileDialog> #include <QInputDialog> #include <QMessageBox> #include <QProgressBar> #include <QProgressDialog> +#include <QPushButton> #include <QShortcut> #include <QStatusBar> #include <QSysInfo> @@ -80,6 +81,7 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "core/core.h" #include "core/crypto/key_manager.h" #include "core/file_sys/card_image.h" +#include "core/file_sys/common_funcs.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/patch_manager.h" @@ -121,14 +123,6 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "yuzu/discord_impl.h" #endif -#ifdef YUZU_USE_QT_WEB_ENGINE -#include <QWebEngineProfile> -#include <QWebEngineScript> -#include <QWebEngineScriptCollection> -#include <QWebEngineSettings> -#include <QWebEngineView> -#endif - #ifdef QT_STATICPLUGIN Q_IMPORT_PLUGIN(QWindowsIntegrationPlugin); #endif @@ -145,12 +139,10 @@ __declspec(dllexport) int AmdPowerXpressRequestHighPerformance = 1; constexpr int default_mouse_timeout = 2500; -constexpr u64 DLC_BASE_TITLE_ID_MASK = 0xFFFFFFFFFFFFE000; - /** * "Callouts" are one-time instructional messages shown to the user. In the config settings, there * is a bitfield "callout_flags" options, used to track if a message has already been shown to the - * user. This is 32-bits - if we have more than 32 callouts, we should retire and recyle old ones. + * user. This is 32-bits - if we have more than 32 callouts, we should retire and recycle old ones. */ enum class CalloutFlag : uint32_t { Telemetry = 0x1, @@ -169,7 +161,7 @@ void GMainWindow::ShowTelemetryCallout() { "<br/><br/>Would you like to share your usage data with us?"); if (QMessageBox::question(this, tr("Telemetry"), telemetry_message) != QMessageBox::Yes) { Settings::values.enable_telemetry = false; - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); } } @@ -188,6 +180,30 @@ static void InitializeLogging() { #endif } +static void RemoveCachedContents() { + const auto offline_fonts = Common::FS::SanitizePath( + fmt::format("{}/fonts", Common::FS::GetUserPath(Common::FS::UserPath::CacheDir)), + Common::FS::DirectorySeparator::PlatformDefault); + + const auto offline_manual = Common::FS::SanitizePath( + fmt::format("{}/offline_web_applet_manual", + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir)), + Common::FS::DirectorySeparator::PlatformDefault); + const auto offline_legal_information = Common::FS::SanitizePath( + fmt::format("{}/offline_web_applet_legal_information", + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir)), + Common::FS::DirectorySeparator::PlatformDefault); + const auto offline_system_data = Common::FS::SanitizePath( + fmt::format("{}/offline_web_applet_system_data", + Common::FS::GetUserPath(Common::FS::UserPath::CacheDir)), + Common::FS::DirectorySeparator::PlatformDefault); + + Common::FS::DeleteDirRecursively(offline_fonts); + Common::FS::DeleteDirRecursively(offline_manual); + Common::FS::DeleteDirRecursively(offline_legal_information); + Common::FS::DeleteDirRecursively(offline_system_data); +} + GMainWindow::GMainWindow() : input_subsystem{std::make_shared<InputCommon::InputSubsystem>()}, config{std::make_unique<Config>()}, vfs{std::make_shared<FileSys::RealVfsFilesystem>()}, @@ -256,6 +272,9 @@ GMainWindow::GMainWindow() FileSys::ContentProviderUnionSlot::FrontendManual, provider.get()); Core::System::GetInstance().GetFileSystemController().CreateFactories(*vfs); + // Remove cached contents generated during the previous session + RemoveCachedContents(); + // Gen keys if necessary OnReinitializeKeys(ReinitializeKeyBehavior::NoWarning); @@ -273,9 +292,47 @@ GMainWindow::GMainWindow() connect(&mouse_hide_timer, &QTimer::timeout, this, &GMainWindow::HideMouseCursor); connect(ui.menubar, &QMenuBar::hovered, this, &GMainWindow::ShowMouseCursor); + MigrateConfigFiles(); + + ui.action_Fullscreen->setChecked(false); + QStringList args = QApplication::arguments(); - if (args.length() >= 2) { - BootGame(args[1]); + + if (args.size() < 2) { + return; + } + + QString game_path; + + for (int i = 1; i < args.size(); ++i) { + // Preserves drag/drop functionality + if (args.size() == 2 && !args[1].startsWith(QChar::fromLatin1('-'))) { + game_path = args[1]; + break; + } + + // Launch game in fullscreen mode + if (args[i] == QStringLiteral("-f")) { + ui.action_Fullscreen->setChecked(true); + continue; + } + + // Launch game at path + if (args[i] == QStringLiteral("-g")) { + if (i >= args.size() - 1) { + continue; + } + + if (args[i + 1].startsWith(QChar::fromLatin1('-'))) { + continue; + } + + game_path = args[++i]; + } + } + + if (!game_path.isEmpty()) { + BootGame(game_path); } } @@ -288,38 +345,34 @@ GMainWindow::~GMainWindow() { void GMainWindow::ControllerSelectorReconfigureControllers( const Core::Frontend::ControllerParameters& parameters) { QtControllerSelectorDialog dialog(this, parameters, input_subsystem.get()); - dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | - Qt::WindowSystemMenuHint); + + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowStaysOnTopHint | + Qt::WindowTitleHint | Qt::WindowSystemMenuHint); dialog.setWindowModality(Qt::WindowModal); dialog.exec(); emit ControllerSelectorReconfigureFinished(); // Don't forget to apply settings. - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); config->Save(); UpdateStatusButtons(); } void GMainWindow::ProfileSelectorSelectProfile() { - const Service::Account::ProfileManager manager; - int index = 0; - if (manager.GetUserCount() != 1) { - QtProfileSelectionDialog dialog(this); - dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | - Qt::WindowSystemMenuHint | Qt::WindowCloseButtonHint); - dialog.setWindowModality(Qt::WindowModal); - - if (dialog.exec() == QDialog::Rejected) { - emit ProfileSelectorFinishedSelection(std::nullopt); - return; - } - - index = dialog.GetIndex(); + QtProfileSelectionDialog dialog(this); + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowStaysOnTopHint | + Qt::WindowTitleHint | Qt::WindowSystemMenuHint | + Qt::WindowCloseButtonHint); + dialog.setWindowModality(Qt::WindowModal); + if (dialog.exec() == QDialog::Rejected) { + emit ProfileSelectorFinishedSelection(std::nullopt); + return; } - const auto uuid = manager.GetUser(static_cast<std::size_t>(index)); + const Service::Account::ProfileManager manager; + const auto uuid = manager.GetUser(static_cast<std::size_t>(dialog.GetIndex())); if (!uuid.has_value()) { emit ProfileSelectorFinishedSelection(std::nullopt); return; @@ -331,8 +384,9 @@ void GMainWindow::ProfileSelectorSelectProfile() { void GMainWindow::SoftwareKeyboardGetText( const Core::Frontend::SoftwareKeyboardParameters& parameters) { QtSoftwareKeyboardDialog dialog(this, parameters); - dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowTitleHint | - Qt::WindowSystemMenuHint | Qt::WindowCloseButtonHint); + dialog.setWindowFlags(Qt::Dialog | Qt::CustomizeWindowHint | Qt::WindowStaysOnTopHint | + Qt::WindowTitleHint | Qt::WindowSystemMenuHint | + Qt::WindowCloseButtonHint); dialog.setWindowModality(Qt::WindowModal); if (dialog.exec() == QDialog::Rejected) { @@ -348,148 +402,141 @@ void GMainWindow::SoftwareKeyboardInvokeCheckDialog(std::u16string error_message emit SoftwareKeyboardFinishedCheckDialog(); } +void GMainWindow::WebBrowserOpenWebPage(std::string_view main_url, std::string_view additional_args, + bool is_local) { #ifdef YUZU_USE_QT_WEB_ENGINE -void GMainWindow::WebBrowserOpenPage(std::string_view filename, std::string_view additional_args) { - NXInputWebEngineView web_browser_view(this); + if (disable_web_applet) { + emit WebBrowserClosed(Service::AM::Applets::WebExitReason::WindowClosed, + "http://localhost/"); + return; + } - // Scope to contain the QProgressDialog for initialization - { - QProgressDialog progress(this); - progress.setMinimumDuration(200); - progress.setLabelText(tr("Loading Web Applet...")); - progress.setRange(0, 4); - progress.setValue(0); - progress.show(); + QtNXWebEngineView web_browser_view(this, Core::System::GetInstance(), input_subsystem.get()); - auto future = QtConcurrent::run([this] { emit WebBrowserUnpackRomFS(); }); + ui.action_Pause->setEnabled(false); + ui.action_Restart->setEnabled(false); + ui.action_Stop->setEnabled(false); - while (!future.isFinished()) - QApplication::processEvents(); + { + QProgressDialog loading_progress(this); + loading_progress.setLabelText(tr("Loading Web Applet...")); + loading_progress.setRange(0, 3); + loading_progress.setValue(0); - progress.setValue(1); + if (is_local && !Common::FS::Exists(std::string(main_url))) { + loading_progress.show(); - // Load the special shim script to handle input and exit. - QWebEngineScript nx_shim; - nx_shim.setSourceCode(GetNXShimInjectionScript()); - nx_shim.setWorldId(QWebEngineScript::MainWorld); - nx_shim.setName(QStringLiteral("nx_inject.js")); - nx_shim.setInjectionPoint(QWebEngineScript::DocumentCreation); - nx_shim.setRunsOnSubFrames(true); - web_browser_view.page()->profile()->scripts()->insert(nx_shim); + auto future = QtConcurrent::run([this] { emit WebBrowserExtractOfflineRomFS(); }); - web_browser_view.load( - QUrl(QUrl::fromLocalFile(QString::fromStdString(std::string(filename))).toString() + - QString::fromStdString(std::string(additional_args)))); + while (!future.isFinished()) { + QCoreApplication::processEvents(); + } + } - progress.setValue(2); + loading_progress.setValue(1); - render_window->hide(); - web_browser_view.setFocus(); + if (is_local) { + web_browser_view.LoadLocalWebPage(main_url, additional_args); + } else { + web_browser_view.LoadExternalWebPage(main_url, additional_args); + } + + if (render_window->IsLoadingComplete()) { + render_window->hide(); + } const auto& layout = render_window->GetFramebufferLayout(); web_browser_view.resize(layout.screen.GetWidth(), layout.screen.GetHeight()); web_browser_view.move(layout.screen.left, layout.screen.top + menuBar()->height()); web_browser_view.setZoomFactor(static_cast<qreal>(layout.screen.GetWidth()) / - Layout::ScreenUndocked::Width); - web_browser_view.settings()->setAttribute( - QWebEngineSettings::LocalContentCanAccessRemoteUrls, true); + static_cast<qreal>(Layout::ScreenUndocked::Width)); + web_browser_view.setFocus(); web_browser_view.show(); - progress.setValue(3); + loading_progress.setValue(2); - QApplication::processEvents(); + QCoreApplication::processEvents(); - progress.setValue(4); + loading_progress.setValue(3); } - bool finished = false; - QAction* exit_action = new QAction(tr("Exit Web Applet"), this); - connect(exit_action, &QAction::triggered, this, [&finished] { finished = true; }); - ui.menubar->addAction(exit_action); + bool exit_check = false; - auto& npad = - Core::System::GetInstance() - .ServiceManager() - .GetService<Service::HID::Hid>("hid") - ->GetAppletResource() - ->GetController<Service::HID::Controller_NPad>(Service::HID::HidController::NPad); - - const auto fire_js_keypress = [&web_browser_view](u32 key_code) { - web_browser_view.page()->runJavaScript( - QStringLiteral("document.dispatchEvent(new KeyboardEvent('keydown', {'key': %1}));") - .arg(key_code)); - }; + // TODO (Morph): Remove this + QAction* exit_action = new QAction(tr("Disable Web Applet"), this); + connect(exit_action, &QAction::triggered, this, [this, &web_browser_view] { + const auto result = QMessageBox::warning( + this, tr("Disable Web Applet"), + tr("Disabling the web applet will cause it to not be shown again for the rest of the " + "emulated session. This can lead to undefined behavior and should only be used with " + "Super Mario 3D All-Stars. Are you sure you want to disable the web applet?"), + QMessageBox::Yes | QMessageBox::No); + if (result == QMessageBox::Yes) { + disable_web_applet = true; + web_browser_view.SetFinished(true); + } + }); + ui.menubar->addAction(exit_action); - QMessageBox::information( - this, tr("Exit"), - tr("To exit the web application, use the game provided controls to select exit, select the " - "'Exit Web Applet' option in the menu bar, or press the 'Enter' key.")); - - bool running_exit_check = false; - while (!finished) { - QApplication::processEvents(); - - if (!running_exit_check) { - web_browser_view.page()->runJavaScript(QStringLiteral("applet_done;"), - [&](const QVariant& res) { - running_exit_check = false; - if (res.toBool()) - finished = true; - }); - running_exit_check = true; + while (!web_browser_view.IsFinished()) { + QCoreApplication::processEvents(); + + if (!exit_check) { + web_browser_view.page()->runJavaScript( + QStringLiteral("end_applet;"), [&](const QVariant& variant) { + exit_check = false; + if (variant.toBool()) { + web_browser_view.SetFinished(true); + web_browser_view.SetExitReason( + Service::AM::Applets::WebExitReason::EndButtonPressed); + } + }); + + exit_check = true; } - const auto input = npad.GetAndResetPressState(); - for (std::size_t i = 0; i < Settings::NativeButton::NumButtons; ++i) { - if ((input & (1 << i)) != 0) { - LOG_DEBUG(Frontend, "firing input for button id={:02X}", i); - web_browser_view.page()->runJavaScript( - QStringLiteral("yuzu_key_callbacks[%1]();").arg(i)); + if (web_browser_view.GetCurrentURL().contains(QStringLiteral("localhost"))) { + if (!web_browser_view.IsFinished()) { + web_browser_view.SetFinished(true); + web_browser_view.SetExitReason(Service::AM::Applets::WebExitReason::CallbackURL); } + + web_browser_view.SetLastURL(web_browser_view.GetCurrentURL().toStdString()); } - if (input & 0x00888000) // RStick Down | LStick Down | DPad Down - fire_js_keypress(40); // Down Arrow Key - else if (input & 0x00444000) // RStick Right | LStick Right | DPad Right - fire_js_keypress(39); // Right Arrow Key - else if (input & 0x00222000) // RStick Up | LStick Up | DPad Up - fire_js_keypress(38); // Up Arrow Key - else if (input & 0x00111000) // RStick Left | LStick Left | DPad Left - fire_js_keypress(37); // Left Arrow Key - else if (input & 0x00000001) // A Button - fire_js_keypress(13); // Enter Key + std::this_thread::sleep_for(std::chrono::milliseconds(1)); } + const auto exit_reason = web_browser_view.GetExitReason(); + const auto last_url = web_browser_view.GetLastURL(); + web_browser_view.hide(); - render_window->show(); + render_window->setFocus(); - ui.menubar->removeAction(exit_action); - // Needed to update render window focus/show and remove menubar action - QApplication::processEvents(); - emit WebBrowserFinishedBrowsing(); -} + if (render_window->IsLoadingComplete()) { + render_window->show(); + } -#else + ui.action_Pause->setEnabled(true); + ui.action_Restart->setEnabled(true); + ui.action_Stop->setEnabled(true); -void GMainWindow::WebBrowserOpenPage(std::string_view filename, std::string_view additional_args) { - QMessageBox::warning( - this, tr("Web Applet"), - tr("This version of yuzu was built without QtWebEngine support, meaning that yuzu cannot " - "properly display the game manual or web page requested."), - QMessageBox::Ok, QMessageBox::Ok); + ui.menubar->removeAction(exit_action); - LOG_INFO(Frontend, - "(STUBBED) called - Missing QtWebEngine dependency needed to open website page at " - "'{}' with arguments '{}'!", - filename, additional_args); + QCoreApplication::processEvents(); - emit WebBrowserFinishedBrowsing(); -} + emit WebBrowserClosed(exit_reason, last_url); + +#else + + // Utilize the same fallback as the default web browser applet. + emit WebBrowserClosed(Service::AM::Applets::WebExitReason::WindowClosed, "http://localhost/"); #endif +} void GMainWindow::InitializeWidgets() { #ifdef YUZU_ENABLE_COMPATIBILITY_REPORTING @@ -551,13 +598,14 @@ void GMainWindow::InitializeWidgets() { dock_status_button->setObjectName(QStringLiteral("TogglableStatusBarButton")); dock_status_button->setFocusPolicy(Qt::NoFocus); connect(dock_status_button, &QPushButton::clicked, [&] { - Settings::values.use_docked_mode = !Settings::values.use_docked_mode; - dock_status_button->setChecked(Settings::values.use_docked_mode); - OnDockedModeChanged(!Settings::values.use_docked_mode, Settings::values.use_docked_mode); + Settings::values.use_docked_mode.SetValue(!Settings::values.use_docked_mode.GetValue()); + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); + OnDockedModeChanged(!Settings::values.use_docked_mode.GetValue(), + Settings::values.use_docked_mode.GetValue()); }); dock_status_button->setText(tr("DOCK")); dock_status_button->setCheckable(true); - dock_status_button->setChecked(Settings::values.use_docked_mode); + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); statusBar()->insertPermanentWidget(0, dock_status_button); // Setup ASync button @@ -568,11 +616,10 @@ void GMainWindow::InitializeWidgets() { if (emulation_running) { return; } - bool is_async = !Settings::values.use_asynchronous_gpu_emulation.GetValue() || - Settings::values.use_multi_core.GetValue(); - Settings::values.use_asynchronous_gpu_emulation.SetValue(is_async); + Settings::values.use_asynchronous_gpu_emulation.SetValue( + !Settings::values.use_asynchronous_gpu_emulation.GetValue()); async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation.GetValue()); - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); }); async_status_button->setText(tr("ASYNC")); async_status_button->setCheckable(true); @@ -587,16 +634,13 @@ void GMainWindow::InitializeWidgets() { return; } Settings::values.use_multi_core.SetValue(!Settings::values.use_multi_core.GetValue()); - bool is_async = Settings::values.use_asynchronous_gpu_emulation.GetValue() || - Settings::values.use_multi_core.GetValue(); - Settings::values.use_asynchronous_gpu_emulation.SetValue(is_async); - async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation.GetValue()); multicore_status_button->setChecked(Settings::values.use_multi_core.GetValue()); - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); }); multicore_status_button->setText(tr("MULTICORE")); multicore_status_button->setCheckable(true); multicore_status_button->setChecked(Settings::values.use_multi_core.GetValue()); + statusBar()->insertPermanentWidget(0, multicore_status_button); statusBar()->insertPermanentWidget(0, async_status_button); @@ -610,11 +654,6 @@ void GMainWindow::InitializeWidgets() { }); renderer_status_button->toggle(); -#ifndef HAS_VULKAN - renderer_status_button->setChecked(false); - renderer_status_button->setCheckable(false); - renderer_status_button->setDisabled(true); -#else renderer_status_button->setChecked(Settings::values.renderer_backend.GetValue() == Settings::RendererBackend::Vulkan); connect(renderer_status_button, &QPushButton::clicked, [this] { @@ -627,9 +666,8 @@ void GMainWindow::InitializeWidgets() { Settings::values.renderer_backend.SetValue(Settings::RendererBackend::OpenGL); } - Settings::Apply(); + Settings::Apply(Core::System::GetInstance()); }); -#endif // HAS_VULKAN statusBar()->insertPermanentWidget(0, renderer_status_button); statusBar()->setVisible(true); @@ -665,7 +703,7 @@ void GMainWindow::InitializeRecentFileMenuActions() { } ui.menu_recent_files->addSeparator(); QAction* action_clear_recent_files = new QAction(this); - action_clear_recent_files->setText(tr("Clear Recent Files")); + action_clear_recent_files->setText(tr("&Clear Recent Files")); connect(action_clear_recent_files, &QAction::triggered, this, [this] { UISettings::values.recent_files.clear(); UpdateRecentFiles(); @@ -796,10 +834,11 @@ void GMainWindow::InitializeHotkeys() { }); connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Change Docked Mode"), this), &QShortcut::activated, this, [&] { - Settings::values.use_docked_mode = !Settings::values.use_docked_mode; - OnDockedModeChanged(!Settings::values.use_docked_mode, - Settings::values.use_docked_mode); - dock_status_button->setChecked(Settings::values.use_docked_mode); + Settings::values.use_docked_mode.SetValue( + !Settings::values.use_docked_mode.GetValue()); + OnDockedModeChanged(!Settings::values.use_docked_mode.GetValue(), + Settings::values.use_docked_mode.GetValue()); + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); }); connect(hotkey_registry.GetHotkey(main_window, QStringLiteral("Mute Audio"), this), &QShortcut::activated, this, @@ -926,7 +965,10 @@ void GMainWindow::ConnectMenuEvents() { &GMainWindow::OnDisplayTitleBars); connect(ui.action_Show_Filter_Bar, &QAction::triggered, this, &GMainWindow::OnToggleFilterBar); connect(ui.action_Show_Status_Bar, &QAction::triggered, statusBar(), &QStatusBar::setVisible); - connect(ui.action_Reset_Window_Size, &QAction::triggered, this, &GMainWindow::ResetWindowSize); + connect(ui.action_Reset_Window_Size_720, &QAction::triggered, this, + &GMainWindow::ResetWindowSize720); + connect(ui.action_Reset_Window_Size_1080, &QAction::triggered, this, + &GMainWindow::ResetWindowSize1080); // Fullscreen connect(ui.action_Fullscreen, &QAction::triggered, this, &GMainWindow::ToggleFullscreen); @@ -974,7 +1016,7 @@ void GMainWindow::AllowOSSleep() { #endif } -bool GMainWindow::LoadROM(const QString& filename) { +bool GMainWindow::LoadROM(const QString& filename, std::size_t program_index) { // Shutdown previous session if the emu thread is still active... if (emu_thread != nullptr) ShutdownGame(); @@ -988,7 +1030,6 @@ bool GMainWindow::LoadROM(const QString& filename) { system.SetAppletFrontendSet({ std::make_unique<QtControllerSelector>(*this), // Controller Selector - nullptr, // E-Commerce std::make_unique<QtErrorDisplay>(*this), // Error Display nullptr, // Parental Controls nullptr, // Photo Viewer @@ -999,7 +1040,8 @@ bool GMainWindow::LoadROM(const QString& filename) { system.RegisterHostThread(); - const Core::System::ResultStatus result{system.Load(*render_window, filename.toStdString())}; + const Core::System::ResultStatus result{ + system.Load(*render_window, filename.toStdString(), program_index)}; const auto drd_callout = (UISettings::values.callout_flags & static_cast<u32>(CalloutFlag::DRDDeprecation)) == 0; @@ -1039,20 +1081,24 @@ bool GMainWindow::LoadROM(const QString& filename) { break; default: - if (static_cast<u32>(result) > - static_cast<u32>(Core::System::ResultStatus::ErrorLoader)) { + if (result > Core::System::ResultStatus::ErrorLoader) { const u16 loader_id = static_cast<u16>(Core::System::ResultStatus::ErrorLoader); const u16 error_id = static_cast<u16>(result) - loader_id; const std::string error_code = fmt::format("({:04X}-{:04X})", loader_id, error_id); LOG_CRITICAL(Frontend, "Failed to load ROM! {}", error_code); - QMessageBox::critical( - this, - tr("Error while loading ROM! ").append(QString::fromStdString(error_code)), - QString::fromStdString(fmt::format( - "{}<br>Please follow <a href='https://yuzu-emu.org/help/quickstart/'>the " - "yuzu quickstart guide</a> to redump your files.<br>You can refer " - "to the yuzu wiki</a> or the yuzu Discord</a> for help.", - static_cast<Loader::ResultStatus>(error_id)))); + + const auto title = + tr("Error while loading ROM! %1", "%1 signifies a numeric error code.") + .arg(QString::fromStdString(error_code)); + const auto description = + tr("%1<br>Please follow <a href='https://yuzu-emu.org/help/quickstart/'>the " + "yuzu quickstart guide</a> to redump your files.<br>You can refer " + "to the yuzu wiki</a> or the yuzu Discord</a> for help.", + "%1 signifies an error string.") + .arg(QString::fromStdString( + GetResultStatusString(static_cast<Loader::ResultStatus>(error_id)))); + + QMessageBox::critical(this, title, description); } else { QMessageBox::critical( this, tr("Error while loading ROM!"), @@ -1081,26 +1127,37 @@ void GMainWindow::SelectAndSetCurrentUser() { Settings::values.current_user = dialog.GetIndex(); } -void GMainWindow::BootGame(const QString& filename) { +void GMainWindow::BootGame(const QString& filename, std::size_t program_index) { LOG_INFO(Frontend, "yuzu starting..."); StoreRecentFile(filename); // Put the filename on top of the list u64 title_id{0}; + last_filename_booted = filename; + + auto& system = Core::System::GetInstance(); const auto v_file = Core::GetGameFileFromPath(vfs, filename.toUtf8().constData()); - const auto loader = Loader::GetLoader(v_file); + const auto loader = Loader::GetLoader(system, v_file, program_index); + if (!(loader == nullptr || loader->ReadProgramId(title_id) != Loader::ResultStatus::Success)) { // Load per game settings - Config per_game_config(fmt::format("{:016X}.ini", title_id), false); + Config per_game_config(fmt::format("{:016X}", title_id), Config::ConfigType::PerGameConfig); } + ConfigureVibration::SetAllVibrationDevices(); + + // Save configurations + UpdateUISettings(); + game_list->SaveInterfaceLayout(); + config->Save(); + Settings::LogSettings(); if (UISettings::values.select_user_on_boot) { SelectAndSetCurrentUser(); } - if (!LoadROM(filename)) + if (!LoadROM(filename, program_index)) return; // Create and start the emulation thread @@ -1108,7 +1165,12 @@ void GMainWindow::BootGame(const QString& filename) { emit EmulationStarting(emu_thread.get()); emu_thread->start(); + // Register an ExecuteProgram callback such that Core can execute a sub-program + system.RegisterExecuteProgramCallback( + [this](std::size_t program_index) { render_window->ExecuteProgram(program_index); }); + connect(render_window, &GRenderWindow::Closed, this, &GMainWindow::OnStopGame); + connect(render_window, &GRenderWindow::MouseActivity, this, &GMainWindow::OnMouseActivity); // BlockingQueuedConnection is important here, it makes sure we've finished refreshing our views // before the CPU continues connect(emu_thread.get(), &EmuThread::DebugModeEntered, waitTreeWidget, @@ -1132,15 +1194,19 @@ void GMainWindow::BootGame(const QString& filename) { if (UISettings::values.hide_mouse) { mouse_hide_timer.start(); - setMouseTracking(true); - ui.centralwidget->setMouseTracking(true); + render_window->installEventFilter(render_window); + render_window->setAttribute(Qt::WA_Hover, true); } std::string title_name; std::string title_version; - const auto res = Core::System::GetInstance().GetGameName(title_name); + const auto res = system.GetGameName(title_name); - const auto metadata = FileSys::PatchManager(title_id).GetControlMetadata(); + const auto metadata = [&system, title_id] { + const FileSys::PatchManager pm(title_id, system.GetFileSystemController(), + system.GetContentProvider()); + return pm.GetControlMetadata(); + }(); if (metadata.first != nullptr) { title_version = metadata.first->GetVersionString(); title_name = metadata.first->GetApplicationName(); @@ -1151,7 +1217,7 @@ void GMainWindow::BootGame(const QString& filename) { LOG_INFO(Frontend, "Booting game: {:016X} | {} | {}", title_id, title_name, title_version); UpdateWindowTitle(title_name, title_version); - loading_screen->Prepare(Core::System::GetInstance().GetAppLoader()); + loading_screen->Prepare(system.GetAppLoader()); loading_screen->show(); emulation_running = true; @@ -1206,8 +1272,8 @@ void GMainWindow::ShutdownGame() { } game_list->SetFilterFocus(); - setMouseTracking(false); - ui.centralwidget->setMouseTracking(false); + render_window->removeEventFilter(render_window); + render_window->setAttribute(Qt::WA_Hover, false); UpdateWindowTitle(); @@ -1219,9 +1285,7 @@ void GMainWindow::ShutdownGame() { emu_frametime_label->setVisible(false); async_status_button->setEnabled(true); multicore_status_button->setEnabled(true); -#ifdef HAS_VULKAN renderer_status_button->setEnabled(true); -#endif emulation_running = false; @@ -1270,16 +1334,18 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target const std::string& game_path) { std::string path; QString open_target; + auto& system = Core::System::GetInstance(); - const auto [user_save_size, device_save_size] = [this, &program_id, &game_path] { - FileSys::PatchManager pm{program_id}; + const auto [user_save_size, device_save_size] = [this, &game_path, &program_id, &system] { + const FileSys::PatchManager pm{program_id, system.GetFileSystemController(), + system.GetContentProvider()}; const auto control = pm.GetControlMetadata().first; if (control != nullptr) { return std::make_pair(control->GetDefaultNormalSaveSize(), control->GetDeviceSaveDataSize()); } else { const auto file = Core::GetGameFileFromPath(vfs, game_path); - const auto loader = Loader::GetLoader(file); + const auto loader = Loader::GetLoader(system, file); FileSys::NACP nacp{}; loader->ReadControlData(nacp); @@ -1321,12 +1387,12 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target const auto user_id = manager.GetUser(static_cast<std::size_t>(index)); ASSERT(user_id); path = nand_dir + FileSys::SaveDataFactory::GetFullPath( - FileSys::SaveDataSpaceId::NandUser, + system, FileSys::SaveDataSpaceId::NandUser, FileSys::SaveDataType::SaveData, program_id, user_id->uuid, 0); } else { // Device save data path = nand_dir + FileSys::SaveDataFactory::GetFullPath( - FileSys::SaveDataSpaceId::NandUser, + system, FileSys::SaveDataSpaceId::NandUser, FileSys::SaveDataType::SaveData, program_id, {}, 0); } @@ -1506,7 +1572,7 @@ void GMainWindow::RemoveAddOnContent(u64 program_id, const QString& entry_type) FileSys::TitleType::AOC, FileSys::ContentRecordType::Data); for (const auto& entry : dlc_entries) { - if ((entry.title_id & DLC_BASE_TITLE_ID_MASK) == program_id) { + if (FileSys::GetBaseTitleID(entry.title_id) == program_id) { const auto res = fs_controller.GetUserNANDContents()->RemoveExistingEntry(entry.title_id) || fs_controller.GetSDMCContents()->RemoveExistingEntry(entry.title_id); @@ -1581,7 +1647,8 @@ void GMainWindow::RemoveCustomConfiguration(u64 program_id) { const QString config_dir = QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::ConfigDir)); const QString custom_config_file_path = - config_dir + QString::fromStdString(fmt::format("{:016X}.ini", program_id)); + config_dir + QStringLiteral("custom") + QDir::separator() + + QString::fromStdString(fmt::format("{:016X}.ini", program_id)); if (!QFile::exists(custom_config_file_path)) { QMessageBox::warning(this, tr("Error Removing Custom Configuration"), @@ -1605,7 +1672,8 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa "cancelled the operation.")); }; - const auto loader = Loader::GetLoader(vfs->OpenFile(game_path, FileSys::Mode::Read)); + auto& system = Core::System::GetInstance(); + const auto loader = Loader::GetLoader(system, vfs->OpenFile(game_path, FileSys::Mode::Read)); if (loader == nullptr) { failed(); return; @@ -1617,7 +1685,7 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa return; } - const auto& installed = Core::System::GetInstance().GetContentProvider(); + const auto& installed = system.GetContentProvider(); const auto romfs_title_id = SelectRomFSDumpTarget(installed, program_id); if (!romfs_title_id) { @@ -1632,7 +1700,7 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa if (*romfs_title_id == program_id) { const u64 ivfc_offset = loader->ReadRomFSIVFCOffset(); - FileSys::PatchManager pm{program_id}; + const FileSys::PatchManager pm{program_id, system.GetFileSystemController(), installed}; romfs = pm.PatchRomFS(file, ivfc_offset, FileSys::ContentRecordType::Program); } else { romfs = installed.GetEntry(*romfs_title_id, FileSys::ContentRecordType::Data)->GetRomFS(); @@ -1749,7 +1817,8 @@ void GMainWindow::OnGameListShowList(bool show) { void GMainWindow::OnGameListOpenPerGameProperties(const std::string& file) { u64 title_id{}; const auto v_file = Core::GetGameFileFromPath(vfs, file); - const auto loader = Loader::GetLoader(v_file); + const auto loader = Loader::GetLoader(Core::System::GetInstance(), v_file); + if (loader == nullptr || loader->ReadProgramId(title_id) != Loader::ResultStatus::Success) { QMessageBox::information(this, tr("Properties"), tr("The game properties could not be loaded.")); @@ -2077,11 +2146,12 @@ void GMainWindow::OnStartGame() { qRegisterMetaType<std::string>("std::string"); qRegisterMetaType<std::optional<std::u16string>>("std::optional<std::u16string>"); qRegisterMetaType<std::string_view>("std::string_view"); + qRegisterMetaType<Service::AM::Applets::WebExitReason>("Service::AM::Applets::WebExitReason"); connect(emu_thread.get(), &EmuThread::ErrorThrown, this, &GMainWindow::OnCoreError); ui.action_Start->setEnabled(false); - ui.action_Start->setText(tr("Continue")); + ui.action_Start->setText(tr("&Continue")); ui.action_Pause->setEnabled(true); ui.action_Stop->setEnabled(true); @@ -2106,14 +2176,14 @@ void GMainWindow::OnPauseGame() { } void GMainWindow::OnStopGame() { - Core::System& system{Core::System::GetInstance()}; + auto& system{Core::System::GetInstance()}; if (system.GetExitLock() && !ConfirmForceLockedExit()) { return; } ShutdownGame(); - Settings::RestoreGlobalState(); + Settings::RestoreGlobalState(system.IsPoweredOn()); UpdateStatusButtons(); } @@ -2121,6 +2191,11 @@ void GMainWindow::OnLoadComplete() { loading_screen->OnLoadComplete(); } +void GMainWindow::OnExecuteProgram(std::size_t program_index) { + ShutdownGame(); + BootGame(last_filename_booted, program_index); +} + void GMainWindow::ErrorDisplayDisplayError(QString body) { QMessageBox::critical(this, tr("Error Display"), body); emit ErrorDisplayFinished(); @@ -2220,7 +2295,7 @@ void GMainWindow::ToggleWindowMode() { } } -void GMainWindow::ResetWindowSize() { +void GMainWindow::ResetWindowSize720() { const auto aspect_ratio = Layout::EmulationAspectRatio( static_cast<Layout::AspectRatio>(Settings::values.aspect_ratio.GetValue()), static_cast<float>(Layout::ScreenUndocked::Height) / Layout::ScreenUndocked::Width); @@ -2234,6 +2309,20 @@ void GMainWindow::ResetWindowSize() { } } +void GMainWindow::ResetWindowSize1080() { + const auto aspect_ratio = Layout::EmulationAspectRatio( + static_cast<Layout::AspectRatio>(Settings::values.aspect_ratio.GetValue()), + static_cast<float>(Layout::ScreenDocked::Height) / Layout::ScreenDocked::Width); + if (!ui.action_Single_Window_Mode->isChecked()) { + render_window->resize(Layout::ScreenDocked::Height / aspect_ratio, + Layout::ScreenDocked::Height); + } else { + resize(Layout::ScreenDocked::Height / aspect_ratio, + Layout::ScreenDocked::Height + menuBar()->height() + + (ui.action_Show_Status_Bar->isChecked() ? statusBar()->height() : 0)); + } +} + void GMainWindow::OnConfigure() { const auto old_theme = UISettings::values.theme; const bool old_discord_presence = UISettings::values.enable_discord_presence; @@ -2265,12 +2354,12 @@ void GMainWindow::OnConfigure() { config->Save(); if (UISettings::values.hide_mouse && emulation_running) { - setMouseTracking(true); - ui.centralwidget->setMouseTracking(true); + render_window->installEventFilter(render_window); + render_window->setAttribute(Qt::WA_Hover, true); mouse_hide_timer.start(); } else { - setMouseTracking(false); - ui.centralwidget->setMouseTracking(false); + render_window->removeEventFilter(render_window); + render_window->setAttribute(Qt::WA_Hover, false); } UpdateStatusButtons(); @@ -2283,10 +2372,11 @@ void GMainWindow::OnConfigurePerGame() { void GMainWindow::OpenPerGameConfiguration(u64 title_id, const std::string& file_name) { const auto v_file = Core::GetGameFileFromPath(vfs, file_name); + const auto& system = Core::System::GetInstance(); ConfigurePerGame dialog(this, title_id); dialog.LoadFromFile(v_file); - auto result = dialog.exec(); + const auto result = dialog.exec(); if (result == QDialog::Accepted) { dialog.ApplyConfiguration(); @@ -2296,13 +2386,14 @@ void GMainWindow::OpenPerGameConfiguration(u64 title_id, const std::string& file } // Do not cause the global config to write local settings into the config file - Settings::RestoreGlobalState(); + const bool is_powered_on = system.IsPoweredOn(); + Settings::RestoreGlobalState(is_powered_on); - if (!Core::System::GetInstance().IsPoweredOn()) { + if (!is_powered_on) { config->Save(); } } else { - Settings::RestoreGlobalState(); + Settings::RestoreGlobalState(system.IsPoweredOn()); } } @@ -2397,6 +2488,29 @@ void GMainWindow::OnCaptureScreenshot() { OnStartGame(); } +// TODO: Written 2020-10-01: Remove per-game config migration code when it is irrelevant +void GMainWindow::MigrateConfigFiles() { + const std::string& config_dir_str = Common::FS::GetUserPath(Common::FS::UserPath::ConfigDir); + const QDir config_dir = QDir(QString::fromStdString(config_dir_str)); + const QStringList config_dir_list = config_dir.entryList(QStringList(QStringLiteral("*.ini"))); + + Common::FS::CreateFullPath(fmt::format("{}custom" DIR_SEP, config_dir_str)); + for (QStringList::const_iterator it = config_dir_list.constBegin(); + it != config_dir_list.constEnd(); ++it) { + const auto filename = it->toStdString(); + if (filename.find_first_not_of("0123456789abcdefACBDEF", 0) < 16) { + continue; + } + const auto origin = fmt::format("{}{}", config_dir_str, filename); + const auto destination = fmt::format("{}custom" DIR_SEP "{}", config_dir_str, filename); + LOG_INFO(Frontend, "Migrating config file from {} to {}", origin, destination); + if (!Common::FS::Rename(origin, destination)) { + // Delete the old config file if one already exists in the new location. + Common::FS::Delete(origin); + } + } +} + void GMainWindow::UpdateWindowTitle(const std::string& title_name, const std::string& title_version) { const auto full_name = std::string(Common::g_build_fullname); @@ -2454,16 +2568,29 @@ void GMainWindow::UpdateStatusBar() { } void GMainWindow::UpdateStatusButtons() { - dock_status_button->setChecked(Settings::values.use_docked_mode); + dock_status_button->setChecked(Settings::values.use_docked_mode.GetValue()); multicore_status_button->setChecked(Settings::values.use_multi_core.GetValue()); - Settings::values.use_asynchronous_gpu_emulation.SetValue( - Settings::values.use_asynchronous_gpu_emulation.GetValue() || - Settings::values.use_multi_core.GetValue()); async_status_button->setChecked(Settings::values.use_asynchronous_gpu_emulation.GetValue()); -#ifdef HAS_VULKAN renderer_status_button->setChecked(Settings::values.renderer_backend.GetValue() == Settings::RendererBackend::Vulkan); +} + +void GMainWindow::UpdateUISettings() { + if (!ui.action_Fullscreen->isChecked()) { + UISettings::values.geometry = saveGeometry(); + UISettings::values.renderwindow_geometry = render_window->saveGeometry(); + } + UISettings::values.state = saveState(); +#if MICROPROFILE_ENABLED + UISettings::values.microprofile_geometry = microProfileDialog->saveGeometry(); + UISettings::values.microprofile_visible = microProfileDialog->isVisible(); #endif + UISettings::values.single_window_mode = ui.action_Single_Window_Mode->isChecked(); + UISettings::values.fullscreen = ui.action_Fullscreen->isChecked(); + UISettings::values.display_titlebar = ui.action_Display_Dock_Widget_Headers->isChecked(); + UISettings::values.show_filter_bar = ui.action_Show_Filter_Bar->isChecked(); + UISettings::values.show_status_bar = ui.action_Show_Status_Bar->isChecked(); + UISettings::values.first_start = false; } void GMainWindow::HideMouseCursor() { @@ -2472,21 +2599,17 @@ void GMainWindow::HideMouseCursor() { ShowMouseCursor(); return; } - setCursor(QCursor(Qt::BlankCursor)); + render_window->setCursor(QCursor(Qt::BlankCursor)); } void GMainWindow::ShowMouseCursor() { - unsetCursor(); + render_window->unsetCursor(); if (emu_thread != nullptr && UISettings::values.hide_mouse) { mouse_hide_timer.start(); } } -void GMainWindow::mouseMoveEvent(QMouseEvent* event) { - ShowMouseCursor(); -} - -void GMainWindow::mousePressEvent(QMouseEvent* event) { +void GMainWindow::OnMouseActivity() { ShowMouseCursor(); } @@ -2550,7 +2673,7 @@ void GMainWindow::OnCoreError(Core::System::ResultStatus result, std::string det if (emu_thread) { ShutdownGame(); - Settings::RestoreGlobalState(); + Settings::RestoreGlobalState(Core::System::GetInstance().IsPoweredOn()); UpdateStatusButtons(); } } else { @@ -2653,7 +2776,7 @@ std::optional<u64> GMainWindow::SelectRomFSDumpTarget(const FileSys::ContentProv dlc_match.reserve(dlc_entries.size()); std::copy_if(dlc_entries.begin(), dlc_entries.end(), std::back_inserter(dlc_match), [&program_id, &installed](const FileSys::ContentProviderEntry& entry) { - return (entry.title_id & DLC_BASE_TITLE_ID_MASK) == program_id && + return FileSys::GetBaseTitleID(entry.title_id) == program_id && installed.GetEntry(entry)->GetStatus() == Loader::ResultStatus::Success; }); @@ -2699,22 +2822,7 @@ void GMainWindow::closeEvent(QCloseEvent* event) { return; } - if (!ui.action_Fullscreen->isChecked()) { - UISettings::values.geometry = saveGeometry(); - UISettings::values.renderwindow_geometry = render_window->saveGeometry(); - } - UISettings::values.state = saveState(); -#if MICROPROFILE_ENABLED - UISettings::values.microprofile_geometry = microProfileDialog->saveGeometry(); - UISettings::values.microprofile_visible = microProfileDialog->isVisible(); -#endif - UISettings::values.single_window_mode = ui.action_Single_Window_Mode->isChecked(); - UISettings::values.fullscreen = ui.action_Fullscreen->isChecked(); - UISettings::values.display_titlebar = ui.action_Display_Dock_Widget_Headers->isChecked(); - UISettings::values.show_filter_bar = ui.action_Show_Filter_Bar->isChecked(); - UISettings::values.show_status_bar = ui.action_Show_Status_Bar->isChecked(); - UISettings::values.first_start = false; - + UpdateUISettings(); game_list->SaveInterfaceLayout(); hotkey_registry.SaveHotkeys(); @@ -2722,7 +2830,7 @@ void GMainWindow::closeEvent(QCloseEvent* event) { if (emu_thread != nullptr) { ShutdownGame(); - Settings::RestoreGlobalState(); + Settings::RestoreGlobalState(Core::System::GetInstance().IsPoweredOn()); UpdateStatusButtons(); } @@ -2890,7 +2998,7 @@ void GMainWindow::OnLanguageChanged(const QString& locale) { UpdateWindowTitle(); if (emulation_running) - ui.action_Start->setText(tr("Continue")); + ui.action_Start->setText(tr("&Continue")); } void GMainWindow::SetDiscordEnabled([[maybe_unused]] bool state) { diff --git a/src/yuzu/main.h b/src/yuzu/main.h index afcfa68a9..31788ea62 100644 --- a/src/yuzu/main.h +++ b/src/yuzu/main.h @@ -55,6 +55,10 @@ namespace InputCommon { class InputSubsystem; } +namespace Service::AM::Applets { +enum class WebExitReason : u32; +} + enum class EmulatedDirectoryTarget { NAND, SDMC, @@ -126,18 +130,20 @@ signals: void SoftwareKeyboardFinishedText(std::optional<std::u16string> text); void SoftwareKeyboardFinishedCheckDialog(); - void WebBrowserUnpackRomFS(); - void WebBrowserFinishedBrowsing(); + void WebBrowserExtractOfflineRomFS(); + void WebBrowserClosed(Service::AM::Applets::WebExitReason exit_reason, std::string last_url); public slots: void OnLoadComplete(); + void OnExecuteProgram(std::size_t program_index); void ControllerSelectorReconfigureControllers( const Core::Frontend::ControllerParameters& parameters); void ErrorDisplayDisplayError(QString body); void ProfileSelectorSelectProfile(); void SoftwareKeyboardGetText(const Core::Frontend::SoftwareKeyboardParameters& parameters); void SoftwareKeyboardInvokeCheckDialog(std::u16string error_message); - void WebBrowserOpenPage(std::string_view filename, std::string_view arguments); + void WebBrowserOpenWebPage(std::string_view main_url, std::string_view additional_args, + bool is_local); void OnAppFocusStateChanged(Qt::ApplicationState state); private: @@ -154,8 +160,8 @@ private: void PreventOSSleep(); void AllowOSSleep(); - bool LoadROM(const QString& filename); - void BootGame(const QString& filename); + bool LoadROM(const QString& filename, std::size_t program_index); + void BootGame(const QString& filename, std::size_t program_index = 0); void ShutdownGame(); void ShowTelemetryCallout(); @@ -236,11 +242,13 @@ private slots: void ShowFullscreen(); void HideFullscreen(); void ToggleWindowMode(); - void ResetWindowSize(); + void ResetWindowSize720(); + void ResetWindowSize1080(); void OnCaptureScreenshot(); void OnCoreError(Core::System::ResultStatus, std::string); void OnReinitializeKeys(ReinitializeKeyBehavior behavior); void OnLanguageChanged(const QString& locale); + void OnMouseActivity(); private: void RemoveBaseContent(u64 program_id, const QString& entry_type); @@ -251,10 +259,12 @@ private: std::optional<u64> SelectRomFSDumpTarget(const FileSys::ContentProvider&, u64 program_id); InstallResult InstallNSPXCI(const QString& filename); InstallResult InstallNCA(const QString& filename); + void MigrateConfigFiles(); void UpdateWindowTitle(const std::string& title_name = {}, const std::string& title_version = {}); void UpdateStatusBar(); void UpdateStatusButtons(); + void UpdateUISettings(); void HideMouseCursor(); void ShowMouseCursor(); void OpenURL(const QUrl& url); @@ -316,10 +326,14 @@ private: // Install progress dialog QProgressDialog* install_progress; + // Last game booted, used for multi-process apps + QString last_filename_booted; + + // Disables the web applet for the rest of the emulated session + bool disable_web_applet{}; + protected: void dropEvent(QDropEvent* event) override; void dragEnterEvent(QDragEnterEvent* event) override; void dragMoveEvent(QDragMoveEvent* event) override; - void mouseMoveEvent(QMouseEvent* event) override; - void mousePressEvent(QMouseEvent* event) override; }; diff --git a/src/yuzu/main.ui b/src/yuzu/main.ui index 2f3792247..e2ad5baf6 100644 --- a/src/yuzu/main.ui +++ b/src/yuzu/main.ui @@ -25,16 +25,7 @@ </property> <widget class="QWidget" name="centralwidget"> <layout class="QHBoxLayout" name="horizontalLayout"> - <property name="leftMargin"> - <number>0</number> - </property> - <property name="topMargin"> - <number>0</number> - </property> - <property name="rightMargin"> - <number>0</number> - </property> - <property name="bottomMargin"> + <property name="margin"> <number>0</number> </property> </layout> @@ -45,7 +36,7 @@ <x>0</x> <y>0</y> <width>1280</width> - <height>21</height> + <height>26</height> </rect> </property> <widget class="QMenu" name="menu_File"> @@ -54,7 +45,7 @@ </property> <widget class="QMenu" name="menu_recent_files"> <property name="title"> - <string>Recent Files</string> + <string>&Recent Files</string> </property> </widget> <addaction name="action_Install_File_NAND"/> @@ -89,7 +80,7 @@ </property> <widget class="QMenu" name="menu_View_Debugging"> <property name="title"> - <string>Debugging</string> + <string>&Debugging</string> </property> </widget> <addaction name="action_Fullscreen"/> @@ -97,13 +88,14 @@ <addaction name="action_Display_Dock_Widget_Headers"/> <addaction name="action_Show_Filter_Bar"/> <addaction name="action_Show_Status_Bar"/> - <addaction name="action_Reset_Window_Size"/> + <addaction name="action_Reset_Window_Size_720"/> + <addaction name="action_Reset_Window_Size_1080"/> <addaction name="separator"/> <addaction name="menu_View_Debugging"/> </widget> <widget class="QMenu" name="menu_Tools"> <property name="title"> - <string>Tools</string> + <string>&Tools</string> </property> <addaction name="action_Rederive"/> <addaction name="separator"/> @@ -131,17 +123,17 @@ <bool>true</bool> </property> <property name="text"> - <string>Install Files to NAND...</string> + <string>&Install Files to NAND...</string> </property> </action> <action name="action_Load_File"> <property name="text"> - <string>Load File...</string> + <string>L&oad File...</string> </property> </action> <action name="action_Load_Folder"> <property name="text"> - <string>Load Folder...</string> + <string>Load &Folder...</string> </property> </action> <action name="action_Exit"> @@ -175,12 +167,12 @@ </action> <action name="action_Rederive"> <property name="text"> - <string>Reinitialize keys...</string> + <string>&Reinitialize keys...</string> </property> </action> <action name="action_About"> <property name="text"> - <string>About yuzu</string> + <string>&About yuzu</string> </property> </action> <action name="action_Single_Window_Mode"> @@ -188,12 +180,12 @@ <bool>true</bool> </property> <property name="text"> - <string>Single Window Mode</string> + <string>Single &Window Mode</string> </property> </action> <action name="action_Configure"> <property name="text"> - <string>Configure...</string> + <string>Con&figure...</string> </property> </action> <action name="action_Display_Dock_Widget_Headers"> @@ -201,7 +193,7 @@ <bool>true</bool> </property> <property name="text"> - <string>Display Dock Widget Headers</string> + <string>Display D&ock Widget Headers</string> </property> </action> <action name="action_Show_Filter_Bar"> @@ -209,7 +201,7 @@ <bool>true</bool> </property> <property name="text"> - <string>Show Filter Bar</string> + <string>Show &Filter Bar</string> </property> </action> <action name="action_Show_Status_Bar"> @@ -217,12 +209,26 @@ <bool>true</bool> </property> <property name="text"> + <string>Show &Status Bar</string> + </property> + <property name="iconText"> <string>Show Status Bar</string> </property> </action> - <action name="action_Reset_Window_Size"> + <action name="action_Reset_Window_Size_720"> + <property name="text"> + <string>Reset Window Size to &720p</string> + </property> + <property name="iconText"> + <string>Reset Window Size to 720p</string> + </property> + </action> + <action name="action_Reset_Window_Size_1080"> <property name="text"> - <string>Reset Window Size</string> + <string>Reset Window Size to &1080p</string> + </property> + <property name="iconText"> + <string>Reset Window Size to 1080p</string> </property> </action> <action name="action_Fullscreen"> @@ -230,7 +236,7 @@ <bool>true</bool> </property> <property name="text"> - <string>Fullscreen</string> + <string>F&ullscreen</string> </property> </action> <action name="action_Restart"> @@ -238,7 +244,7 @@ <bool>false</bool> </property> <property name="text"> - <string>Restart</string> + <string>&Restart</string> </property> </action> <action name="action_Load_Amiibo"> @@ -246,7 +252,7 @@ <bool>false</bool> </property> <property name="text"> - <string>Load Amiibo...</string> + <string>Load &Amiibo...</string> </property> </action> <action name="action_Report_Compatibility"> @@ -254,7 +260,7 @@ <bool>false</bool> </property> <property name="text"> - <string>Report Compatibility</string> + <string>&Report Compatibility</string> </property> <property name="visible"> <bool>false</bool> @@ -262,22 +268,22 @@ </action> <action name="action_Open_Mods_Page"> <property name="text"> - <string>Open Mods Page</string> + <string>Open &Mods Page</string> </property> </action> <action name="action_Open_Quickstart_Guide"> <property name="text"> - <string>Open Quickstart Guide</string> + <string>Open &Quickstart Guide</string> </property> </action> <action name="action_Open_FAQ"> <property name="text"> - <string>FAQ</string> + <string>&FAQ</string> </property> </action> <action name="action_Open_yuzu_Folder"> <property name="text"> - <string>Open yuzu Folder</string> + <string>Open &yuzu Folder</string> </property> </action> <action name="action_Capture_Screenshot"> @@ -285,7 +291,7 @@ <bool>false</bool> </property> <property name="text"> - <string>Capture Screenshot</string> + <string>&Capture Screenshot</string> </property> </action> <action name="action_Configure_Current_Game"> @@ -293,7 +299,7 @@ <bool>false</bool> </property> <property name="text"> - <string>Configure Current Game...</string> + <string>Configure C&urrent Game...</string> </property> </action> </widget> diff --git a/src/yuzu/util/url_request_interceptor.cpp b/src/yuzu/util/url_request_interceptor.cpp new file mode 100644 index 000000000..b637e771e --- /dev/null +++ b/src/yuzu/util/url_request_interceptor.cpp @@ -0,0 +1,34 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#ifdef YUZU_USE_QT_WEB_ENGINE + +#include "yuzu/util/url_request_interceptor.h" + +UrlRequestInterceptor::UrlRequestInterceptor(QObject* p) : QWebEngineUrlRequestInterceptor(p) {} + +UrlRequestInterceptor::~UrlRequestInterceptor() = default; + +void UrlRequestInterceptor::interceptRequest(QWebEngineUrlRequestInfo& info) { + const auto resource_type = info.resourceType(); + + switch (resource_type) { + case QWebEngineUrlRequestInfo::ResourceTypeMainFrame: + requested_url = info.requestUrl(); + emit FrameChanged(); + break; + case QWebEngineUrlRequestInfo::ResourceTypeSubFrame: + case QWebEngineUrlRequestInfo::ResourceTypeXhr: + emit FrameChanged(); + break; + default: + break; + } +} + +QUrl UrlRequestInterceptor::GetRequestedURL() const { + return requested_url; +} + +#endif diff --git a/src/yuzu/util/url_request_interceptor.h b/src/yuzu/util/url_request_interceptor.h new file mode 100644 index 000000000..8a7f7499f --- /dev/null +++ b/src/yuzu/util/url_request_interceptor.h @@ -0,0 +1,30 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#ifdef YUZU_USE_QT_WEB_ENGINE + +#include <QObject> +#include <QWebEngineUrlRequestInterceptor> + +class UrlRequestInterceptor : public QWebEngineUrlRequestInterceptor { + Q_OBJECT + +public: + explicit UrlRequestInterceptor(QObject* p = nullptr); + ~UrlRequestInterceptor() override; + + void interceptRequest(QWebEngineUrlRequestInfo& info) override; + + QUrl GetRequestedURL() const; + +signals: + void FrameChanged(); + +private: + QUrl requested_url; +}; + +#endif diff --git a/src/yuzu_cmd/CMakeLists.txt b/src/yuzu_cmd/CMakeLists.txt index a15719a0f..0b3f2cb54 100644 --- a/src/yuzu_cmd/CMakeLists.txt +++ b/src/yuzu_cmd/CMakeLists.txt @@ -4,26 +4,17 @@ add_executable(yuzu-cmd config.cpp config.h default_ini.h - emu_window/emu_window_sdl2_gl.cpp - emu_window/emu_window_sdl2_gl.h emu_window/emu_window_sdl2.cpp emu_window/emu_window_sdl2.h emu_window/emu_window_sdl2_gl.cpp emu_window/emu_window_sdl2_gl.h + emu_window/emu_window_sdl2_vk.cpp + emu_window/emu_window_sdl2_vk.h resource.h yuzu.cpp yuzu.rc ) -if (ENABLE_VULKAN) - target_sources(yuzu-cmd PRIVATE - emu_window/emu_window_sdl2_vk.cpp - emu_window/emu_window_sdl2_vk.h) - - target_include_directories(yuzu-cmd PRIVATE ../../externals/Vulkan-Headers/include) - target_compile_definitions(yuzu-cmd PRIVATE HAS_VULKAN) -endif() - create_target_directory_groups(yuzu-cmd) target_link_libraries(yuzu-cmd PRIVATE common core input_common) @@ -33,13 +24,13 @@ if (MSVC) endif() target_link_libraries(yuzu-cmd PRIVATE ${PLATFORM_LIBRARIES} SDL2 Threads::Threads) +target_include_directories(yuzu-cmd PRIVATE ../../externals/Vulkan-Headers/include) + if(UNIX AND NOT APPLE) install(TARGETS yuzu-cmd RUNTIME DESTINATION "${CMAKE_INSTALL_PREFIX}/bin") endif() if (MSVC) include(CopyYuzuSDLDeps) - include(CopyYuzuUnicornDeps) copy_yuzu_SDL_deps(yuzu-cmd) - copy_yuzu_unicorn_deps(yuzu-cmd) endif() diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp index 23448e747..41ef6f6b8 100644 --- a/src/yuzu_cmd/config.cpp +++ b/src/yuzu_cmd/config.cpp @@ -228,24 +228,24 @@ static const std::array<int, 8> keyboard_mods{ void Config::ReadValues() { // Controls - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { const auto group = fmt::format("ControlsP{}", p); for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { std::string default_param = InputCommon::GenerateKeyboardParam(default_buttons[i]); - Settings::values.players[p].buttons[i] = + Settings::values.players.GetValue()[p].buttons[i] = sdl2_config->Get(group, Settings::NativeButton::mapping[i], default_param); - if (Settings::values.players[p].buttons[i].empty()) - Settings::values.players[p].buttons[i] = default_param; + if (Settings::values.players.GetValue()[p].buttons[i].empty()) + Settings::values.players.GetValue()[p].buttons[i] = default_param; } for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { std::string default_param = InputCommon::GenerateAnalogParamFromKeys( default_analogs[i][0], default_analogs[i][1], default_analogs[i][2], default_analogs[i][3], default_analogs[i][4], 0.5f); - Settings::values.players[p].analogs[i] = + Settings::values.players.GetValue()[p].analogs[i] = sdl2_config->Get(group, Settings::NativeAnalog::mapping[i], default_param); - if (Settings::values.players[p].analogs[i].empty()) - Settings::values.players[p].analogs[i] = default_param; + if (Settings::values.players.GetValue()[p].analogs[i].empty()) + Settings::values.players.GetValue()[p].analogs[i] = default_param; } } @@ -288,10 +288,12 @@ void Config::ReadValues() { Settings::values.debug_pad_analogs[i] = default_param; } - Settings::values.vibration_enabled = - sdl2_config->GetBoolean("ControlsGeneral", "vibration_enabled", true); - Settings::values.motion_enabled = - sdl2_config->GetBoolean("ControlsGeneral", "motion_enabled", true); + Settings::values.vibration_enabled.SetValue( + sdl2_config->GetBoolean("ControlsGeneral", "vibration_enabled", true)); + Settings::values.enable_accurate_vibrations.SetValue( + sdl2_config->GetBoolean("ControlsGeneral", "enable_accurate_vibrations", false)); + Settings::values.motion_enabled.SetValue( + sdl2_config->GetBoolean("ControlsGeneral", "motion_enabled", true)); Settings::values.touchscreen.enabled = sdl2_config->GetBoolean("ControlsGeneral", "touch_enabled", true); Settings::values.touchscreen.device = @@ -304,10 +306,8 @@ void Config::ReadValues() { sdl2_config->GetInteger("ControlsGeneral", "touch_diameter_x", 15); Settings::values.touchscreen.diameter_y = sdl2_config->GetInteger("ControlsGeneral", "touch_diameter_y", 15); - Settings::values.udp_input_address = - sdl2_config->Get("Controls", "udp_input_address", InputCommon::CemuhookUDP::DEFAULT_ADDR); - Settings::values.udp_input_port = static_cast<u16>(sdl2_config->GetInteger( - "Controls", "udp_input_port", InputCommon::CemuhookUDP::DEFAULT_PORT)); + Settings::values.udp_input_servers = + sdl2_config->Get("Controls", "udp_input_address", InputCommon::CemuhookUDP::DEFAULT_SRV); std::transform(keyboard_keys.begin(), keyboard_keys.end(), Settings::values.keyboard_keys.begin(), InputCommon::GenerateKeyboardParam); @@ -343,8 +343,8 @@ void Config::ReadValues() { Settings::values.gamecard_path = sdl2_config->Get("Data Storage", "gamecard_path", ""); // System - Settings::values.use_docked_mode = sdl2_config->GetBoolean("System", "use_docked_mode", false); - const auto size = sdl2_config->GetInteger("System", "users_size", 0); + Settings::values.use_docked_mode.SetValue( + sdl2_config->GetBoolean("System", "use_docked_mode", true)); Settings::values.current_user = std::clamp<int>( sdl2_config->GetInteger("System", "current_user", 0), 0, Service::Account::MAX_USERS - 1); @@ -371,7 +371,7 @@ void Config::ReadValues() { // Core Settings::values.use_multi_core.SetValue( - sdl2_config->GetBoolean("Core", "use_multi_core", false)); + sdl2_config->GetBoolean("Core", "use_multi_core", true)); // Renderer const int renderer_backend = sdl2_config->GetInteger( @@ -395,11 +395,11 @@ void Config::ReadValues() { const int gpu_accuracy_level = sdl2_config->GetInteger("Renderer", "gpu_accuracy", 0); Settings::values.gpu_accuracy.SetValue(static_cast<Settings::GPUAccuracy>(gpu_accuracy_level)); Settings::values.use_asynchronous_gpu_emulation.SetValue( - sdl2_config->GetBoolean("Renderer", "use_asynchronous_gpu_emulation", false)); + sdl2_config->GetBoolean("Renderer", "use_asynchronous_gpu_emulation", true)); Settings::values.use_vsync.SetValue( static_cast<u16>(sdl2_config->GetInteger("Renderer", "use_vsync", 1))); Settings::values.use_assembly_shaders.SetValue( - sdl2_config->GetBoolean("Renderer", "use_assembly_shaders", false)); + sdl2_config->GetBoolean("Renderer", "use_assembly_shaders", true)); Settings::values.use_asynchronous_shaders.SetValue( sdl2_config->GetBoolean("Renderer", "use_asynchronous_shaders", false)); Settings::values.use_asynchronous_shaders.SetValue( @@ -429,9 +429,6 @@ void Config::ReadValues() { // Debugging Settings::values.record_frame_times = sdl2_config->GetBoolean("Debugging", "record_frame_times", false); - Settings::values.use_gdbstub = sdl2_config->GetBoolean("Debugging", "use_gdbstub", false); - Settings::values.gdbstub_port = - static_cast<u16>(sdl2_config->GetInteger("Debugging", "gdbstub_port", 24689)); Settings::values.program_args = sdl2_config->Get("Debugging", "program_args", ""); Settings::values.dump_exefs = sdl2_config->GetBoolean("Debugging", "dump_exefs", false); Settings::values.dump_nso = sdl2_config->GetBoolean("Debugging", "dump_nso", false); diff --git a/src/yuzu_cmd/default_ini.h b/src/yuzu_cmd/default_ini.h index aa9e40380..3ee0e037d 100644 --- a/src/yuzu_cmd/default_ini.h +++ b/src/yuzu_cmd/default_ini.h @@ -65,6 +65,14 @@ button_screenshot= lstick= rstick= +# Whether to enable or disable vibration +# 0: Disabled, 1 (default): Enabled +vibration_enabled= + +# Whether to enable or disable accurate vibrations +# 0 (default): Disabled, 1: Enabled +enable_accurate_vibrations= + # for motion input, the following devices are available: # - "motion_emu" (default) for emulating motion input from mouse input. Required parameters: # - "update_period": update period in milliseconds (default to 100) @@ -94,7 +102,7 @@ udp_pad_index= [Core] # Whether to use multi-core for CPU emulation -# 0 (default): Disabled, 1: Enabled +# 0: Disabled, 1 (default): Enabled use_multi_core= [Cpu] @@ -163,7 +171,7 @@ max_anisotropy = use_vsync = # Whether to use OpenGL assembly shaders or not. NV_gpu_program5 is required. -# 0 (default): Off, 1: On +# 0: Off, 1 (default): On use_assembly_shaders = # Whether to allow asynchronous shader building. @@ -266,7 +274,7 @@ gamecard_path = [System] # Whether the system is docked -# 1: Yes, 0 (default): No +# 1 (default): Yes, 0: No use_docked_mode = # Allow the use of NFC in games @@ -310,9 +318,6 @@ log_filter = *:Trace [Debugging] # Record frame time data, can be found in the log directory. Boolean value record_frame_times = -# Port for listening to GDB connections. -use_gdbstub=false -gdbstub_port=24689 # Determines whether or not yuzu will dump the ExeFS of all games it attempts to load while loading them dump_exefs=false # Determines whether or not yuzu will dump all NSOs it attempts to load while loading them diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp index 521209622..e32bed5e6 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp @@ -9,7 +9,7 @@ #include "core/perf_stats.h" #include "input_common/keyboard.h" #include "input_common/main.h" -#include "input_common/motion_emu.h" +#include "input_common/mouse/mouse_input.h" #include "input_common/sdl/sdl.h" #include "yuzu_cmd/emu_window/emu_window_sdl2.h" @@ -30,7 +30,7 @@ EmuWindow_SDL2::~EmuWindow_SDL2() { void EmuWindow_SDL2::OnMouseMotion(s32 x, s32 y) { TouchMoved((unsigned)std::max(x, 0), (unsigned)std::max(y, 0)); - input_subsystem->GetMotionEmu()->Tilt(x, y); + input_subsystem->GetMouse()->MouseMove(x, y); } void EmuWindow_SDL2::OnMouseButton(u32 button, u8 state, s32 x, s32 y) { @@ -42,9 +42,9 @@ void EmuWindow_SDL2::OnMouseButton(u32 button, u8 state, s32 x, s32 y) { } } else if (button == SDL_BUTTON_RIGHT) { if (state == SDL_PRESSED) { - input_subsystem->GetMotionEmu()->BeginTilt(x, y); + input_subsystem->GetMouse()->PressButton(x, y, button); } else { - input_subsystem->GetMotionEmu()->EndTilt(); + input_subsystem->GetMouse()->ReleaseButton(button); } } } @@ -121,62 +121,64 @@ void EmuWindow_SDL2::Fullscreen() { SDL_MaximizeWindow(render_window); } -void EmuWindow_SDL2::PollEvents() { +void EmuWindow_SDL2::WaitEvent() { + // Called on main thread SDL_Event event; - // SDL_PollEvent returns 0 when there are no more events in the event queue - while (SDL_PollEvent(&event)) { - switch (event.type) { - case SDL_WINDOWEVENT: - switch (event.window.event) { - case SDL_WINDOWEVENT_SIZE_CHANGED: - case SDL_WINDOWEVENT_RESIZED: - case SDL_WINDOWEVENT_MAXIMIZED: - case SDL_WINDOWEVENT_RESTORED: - OnResize(); - break; - case SDL_WINDOWEVENT_MINIMIZED: - case SDL_WINDOWEVENT_EXPOSED: - is_shown = event.window.event == SDL_WINDOWEVENT_EXPOSED; - OnResize(); - break; - case SDL_WINDOWEVENT_CLOSE: - is_open = false; - break; - } - break; - case SDL_KEYDOWN: - case SDL_KEYUP: - OnKeyEvent(static_cast<int>(event.key.keysym.scancode), event.key.state); - break; - case SDL_MOUSEMOTION: - // ignore if it came from touch - if (event.button.which != SDL_TOUCH_MOUSEID) - OnMouseMotion(event.motion.x, event.motion.y); - break; - case SDL_MOUSEBUTTONDOWN: - case SDL_MOUSEBUTTONUP: - // ignore if it came from touch - if (event.button.which != SDL_TOUCH_MOUSEID) { - OnMouseButton(event.button.button, event.button.state, event.button.x, - event.button.y); - } - break; - case SDL_FINGERDOWN: - OnFingerDown(event.tfinger.x, event.tfinger.y); - break; - case SDL_FINGERMOTION: - OnFingerMotion(event.tfinger.x, event.tfinger.y); + if (!SDL_WaitEvent(&event)) { + LOG_CRITICAL(Frontend, "SDL_WaitEvent failed: {}", SDL_GetError()); + exit(1); + } + + switch (event.type) { + case SDL_WINDOWEVENT: + switch (event.window.event) { + case SDL_WINDOWEVENT_SIZE_CHANGED: + case SDL_WINDOWEVENT_RESIZED: + case SDL_WINDOWEVENT_MAXIMIZED: + case SDL_WINDOWEVENT_RESTORED: + OnResize(); break; - case SDL_FINGERUP: - OnFingerUp(); + case SDL_WINDOWEVENT_MINIMIZED: + case SDL_WINDOWEVENT_EXPOSED: + is_shown = event.window.event == SDL_WINDOWEVENT_EXPOSED; + OnResize(); break; - case SDL_QUIT: + case SDL_WINDOWEVENT_CLOSE: is_open = false; break; - default: - break; } + break; + case SDL_KEYDOWN: + case SDL_KEYUP: + OnKeyEvent(static_cast<int>(event.key.keysym.scancode), event.key.state); + break; + case SDL_MOUSEMOTION: + // ignore if it came from touch + if (event.button.which != SDL_TOUCH_MOUSEID) + OnMouseMotion(event.motion.x, event.motion.y); + break; + case SDL_MOUSEBUTTONDOWN: + case SDL_MOUSEBUTTONUP: + // ignore if it came from touch + if (event.button.which != SDL_TOUCH_MOUSEID) { + OnMouseButton(event.button.button, event.button.state, event.button.x, event.button.y); + } + break; + case SDL_FINGERDOWN: + OnFingerDown(event.tfinger.x, event.tfinger.y); + break; + case SDL_FINGERMOTION: + OnFingerMotion(event.tfinger.x, event.tfinger.y); + break; + case SDL_FINGERUP: + OnFingerUp(); + break; + case SDL_QUIT: + is_open = false; + break; + default: + break; } const u32 current_time = SDL_GetTicks(); diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.h b/src/yuzu_cmd/emu_window/emu_window_sdl2.h index 53d756c3c..a93141240 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.h +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.h @@ -23,38 +23,38 @@ public: explicit EmuWindow_SDL2(InputCommon::InputSubsystem* input_subsystem); ~EmuWindow_SDL2(); - /// Polls window events - void PollEvents() override; - /// Whether the window is still open, and a close request hasn't yet been sent bool IsOpen() const; /// Returns if window is shown (not minimized) bool IsShown() const override; + /// Wait for the next event on the main thread. + void WaitEvent(); + protected: - /// Called by PollEvents when a key is pressed or released. + /// Called by WaitEvent when a key is pressed or released. void OnKeyEvent(int key, u8 state); - /// Called by PollEvents when the mouse moves. + /// Called by WaitEvent when the mouse moves. void OnMouseMotion(s32 x, s32 y); - /// Called by PollEvents when a mouse button is pressed or released + /// Called by WaitEvent when a mouse button is pressed or released void OnMouseButton(u32 button, u8 state, s32 x, s32 y); /// Translates pixel position (0..1) to pixel positions std::pair<unsigned, unsigned> TouchToPixelPos(float touch_x, float touch_y) const; - /// Called by PollEvents when a finger starts touching the touchscreen + /// Called by WaitEvent when a finger starts touching the touchscreen void OnFingerDown(float x, float y); - /// Called by PollEvents when a finger moves while touching the touchscreen + /// Called by WaitEvent when a finger moves while touching the touchscreen void OnFingerMotion(float x, float y); - /// Called by PollEvents when a finger stops touching the touchscreen + /// Called by WaitEvent when a finger stops touching the touchscreen void OnFingerUp(); - /// Called by PollEvents when any event that may cause the window to be resized occurs + /// Called by WaitEvent when any event that may cause the window to be resized occurs void OnResize(); /// Called when user passes the fullscreen parameter flag diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp index 5f35233b5..a103b04bd 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2_gl.cpp @@ -17,7 +17,6 @@ #include "core/settings.h" #include "input_common/keyboard.h" #include "input_common/main.h" -#include "input_common/motion_emu.h" #include "video_core/renderer_base.h" #include "yuzu_cmd/emu_window/emu_window_sdl2_gl.h" diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp index 3a76c785f..4faf62ede 100644 --- a/src/yuzu_cmd/yuzu.cpp +++ b/src/yuzu_cmd/yuzu.cpp @@ -25,7 +25,6 @@ #include "core/crypto/key_manager.h" #include "core/file_sys/registered_cache.h" #include "core/file_sys/vfs_real.h" -#include "core/gdbstub/gdbstub.h" #include "core/hle/kernel/process.h" #include "core/hle/service/filesystem/filesystem.h" #include "core/loader/loader.h" @@ -36,9 +35,7 @@ #include "yuzu_cmd/config.h" #include "yuzu_cmd/emu_window/emu_window_sdl2.h" #include "yuzu_cmd/emu_window/emu_window_sdl2_gl.h" -#ifdef HAS_VULKAN #include "yuzu_cmd/emu_window/emu_window_sdl2_vk.h" -#endif #ifdef _WIN32 // windows.h needs to be included before shellapi.h @@ -65,7 +62,6 @@ __declspec(dllexport) int AmdPowerXpressRequestHighPerformance = 1; static void PrintHelp(const char* argv0) { std::cout << "Usage: " << argv0 << " [options] <filename>\n" - "-g, --gdbport=NUMBER Enable gdb stub on port NUMBER\n" "-f, --fullscreen Start in fullscreen mode\n" "-h, --help Display this help and exit\n" "-v, --version Output version information and exit\n" @@ -97,12 +93,8 @@ int main(int argc, char** argv) { Config config; int option_index = 0; - bool use_gdbstub = Settings::values.use_gdbstub; - u32 gdb_port = static_cast<u32>(Settings::values.gdbstub_port); InitializeLogging(); - - char* endarg; #ifdef _WIN32 int argc_w; auto argv_w = CommandLineToArgvW(GetCommandLineW(), &argc_w); @@ -117,26 +109,17 @@ int main(int argc, char** argv) { bool fullscreen = false; static struct option long_options[] = { - {"gdbport", required_argument, 0, 'g'}, {"fullscreen", no_argument, 0, 'f'}, - {"help", no_argument, 0, 'h'}, {"version", no_argument, 0, 'v'}, - {"program", optional_argument, 0, 'p'}, {0, 0, 0, 0}, + {"fullscreen", no_argument, 0, 'f'}, + {"help", no_argument, 0, 'h'}, + {"version", no_argument, 0, 'v'}, + {"program", optional_argument, 0, 'p'}, + {0, 0, 0, 0}, }; while (optind < argc) { int arg = getopt_long(argc, argv, "g:fhvp::", long_options, &option_index); if (arg != -1) { switch (static_cast<char>(arg)) { - case 'g': - errno = 0; - gdb_port = strtoul(optarg, &endarg, 0); - use_gdbstub = true; - if (endarg == optarg) - errno = EINVAL; - if (errno != 0) { - perror("--gdbport"); - exit(1); - } - break; case 'f': fullscreen = true; LOG_INFO(Frontend, "Starting in fullscreen mode..."); @@ -174,27 +157,20 @@ int main(int argc, char** argv) { return -1; } - // Apply the command line arguments - Settings::values.gdbstub_port = gdb_port; - Settings::values.use_gdbstub = use_gdbstub; - Settings::Apply(); - - Core::System& system{Core::System::GetInstance()}; + auto& system{Core::System::GetInstance()}; InputCommon::InputSubsystem input_subsystem; + // Apply the command line arguments + Settings::Apply(system); + std::unique_ptr<EmuWindow_SDL2> emu_window; switch (Settings::values.renderer_backend.GetValue()) { case Settings::RendererBackend::OpenGL: emu_window = std::make_unique<EmuWindow_SDL2_GL>(&input_subsystem, fullscreen); break; case Settings::RendererBackend::Vulkan: -#ifdef HAS_VULKAN emu_window = std::make_unique<EmuWindow_SDL2_VK>(&input_subsystem); break; -#else - LOG_CRITICAL(Frontend, "Vulkan backend has not been compiled!"); - return 1; -#endif } system.SetContentProvider(std::make_unique<FileSys::ContentProviderUnion>()); @@ -224,7 +200,7 @@ int main(int argc, char** argv) { const u16 loader_id = static_cast<u16>(Core::System::ResultStatus::ErrorLoader); const u16 error_id = static_cast<u16>(load_result) - loader_id; LOG_CRITICAL(Frontend, - "While attempting to load the ROM requested, an error occured. Please " + "While attempting to load the ROM requested, an error occurred. Please " "refer to the yuzu wiki for more information or the yuzu discord for " "additional help.\n\nError Code: {:04X}-{:04X}\nError Description: {}", loader_id, error_id, static_cast<Loader::ResultStatus>(error_id)); @@ -240,11 +216,11 @@ int main(int argc, char** argv) { system.CurrentProcess()->GetTitleID(), false, [](VideoCore::LoadCallbackStage, size_t value, size_t total) {}); - system.Run(); + void(system.Run()); while (emu_window->IsOpen()) { - std::this_thread::sleep_for(std::chrono::milliseconds(1)); + emu_window->WaitEvent(); } - system.Pause(); + void(system.Pause()); system.Shutdown(); detached_tasks.WaitForAllTasks(); diff --git a/src/yuzu_tester/CMakeLists.txt b/src/yuzu_tester/CMakeLists.txt index 06c2ee011..d8a2a1511 100644 --- a/src/yuzu_tester/CMakeLists.txt +++ b/src/yuzu_tester/CMakeLists.txt @@ -28,7 +28,5 @@ endif() if (MSVC) include(CopyYuzuSDLDeps) - include(CopyYuzuUnicornDeps) copy_yuzu_SDL_deps(yuzu-tester) - copy_yuzu_unicorn_deps(yuzu-tester) endif() diff --git a/src/yuzu_tester/config.cpp b/src/yuzu_tester/config.cpp index bc273fb51..0aa143e1f 100644 --- a/src/yuzu_tester/config.cpp +++ b/src/yuzu_tester/config.cpp @@ -47,13 +47,13 @@ bool Config::LoadINI(const std::string& default_contents, bool retry) { void Config::ReadValues() { // Controls - for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { + for (std::size_t p = 0; p < Settings::values.players.GetValue().size(); ++p) { for (int i = 0; i < Settings::NativeButton::NumButtons; ++i) { - Settings::values.players[p].buttons[i] = ""; + Settings::values.players.GetValue()[p].buttons[i] = ""; } for (int i = 0; i < Settings::NativeAnalog::NumAnalogs; ++i) { - Settings::values.players[p].analogs[i] = ""; + Settings::values.players.GetValue()[p].analogs[i] = ""; } } @@ -75,8 +75,9 @@ void Config::ReadValues() { Settings::values.debug_pad_analogs[i] = ""; } - Settings::values.vibration_enabled = true; - Settings::values.motion_enabled = true; + Settings::values.vibration_enabled.SetValue(true); + Settings::values.enable_accurate_vibrations.SetValue(false); + Settings::values.motion_enabled.SetValue(true); Settings::values.touchscreen.enabled = ""; Settings::values.touchscreen.device = ""; Settings::values.touchscreen.finger = 0; @@ -84,8 +85,8 @@ void Config::ReadValues() { Settings::values.touchscreen.diameter_x = 15; Settings::values.touchscreen.diameter_y = 15; - Settings::values.use_docked_mode = - sdl2_config->GetBoolean("Controls", "use_docked_mode", false); + Settings::values.use_docked_mode.SetValue( + sdl2_config->GetBoolean("Controls", "use_docked_mode", true)); // Data Storage Settings::values.use_virtual_sd = @@ -157,7 +158,6 @@ void Config::ReadValues() { Settings::values.use_dev_keys = sdl2_config->GetBoolean("Miscellaneous", "use_dev_keys", false); // Debugging - Settings::values.use_gdbstub = false; Settings::values.program_args = ""; Settings::values.dump_exefs = sdl2_config->GetBoolean("Debugging", "dump_exefs", false); Settings::values.dump_nso = sdl2_config->GetBoolean("Debugging", "dump_nso", false); diff --git a/src/yuzu_tester/default_ini.h b/src/yuzu_tester/default_ini.h index 3eb64e9d7..779c3791b 100644 --- a/src/yuzu_tester/default_ini.h +++ b/src/yuzu_tester/default_ini.h @@ -116,7 +116,7 @@ use_virtual_sd = [System] # Whether the system is docked -# 1: Yes, 0 (default): No +# 1 (default): Yes, 0: No use_docked_mode = # Allow the use of NFC in games diff --git a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp index 78f75fb38..358e03870 100644 --- a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp +++ b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.cpp @@ -109,8 +109,6 @@ EmuWindow_SDL2_Hide::~EmuWindow_SDL2_Hide() { SDL_Quit(); } -void EmuWindow_SDL2_Hide::PollEvents() {} - bool EmuWindow_SDL2_Hide::IsShown() const { return false; } diff --git a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h index a553b4b95..adccdf35e 100644 --- a/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h +++ b/src/yuzu_tester/emu_window/emu_window_sdl2_hide.h @@ -17,9 +17,6 @@ public: explicit EmuWindow_SDL2_Hide(); ~EmuWindow_SDL2_Hide(); - /// Polls window events - void PollEvents() override; - /// Whether the screen is being shown or not. bool IsShown() const override; diff --git a/src/yuzu_tester/service/yuzutest.cpp b/src/yuzu_tester/service/yuzutest.cpp index 2d3f6e3a7..e257fae25 100644 --- a/src/yuzu_tester/service/yuzutest.cpp +++ b/src/yuzu_tester/service/yuzutest.cpp @@ -4,6 +4,7 @@ #include <memory> #include "common/string_util.h" +#include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/service/service.h" #include "core/hle/service/sm/sm.h" @@ -15,10 +16,10 @@ constexpr u64 SERVICE_VERSION = 0x00000002; class YuzuTest final : public ServiceFramework<YuzuTest> { public: - explicit YuzuTest(std::string data, - std::function<void(std::vector<TestResult>)> finish_callback) - : ServiceFramework{"yuzutest"}, data(std::move(data)), - finish_callback(std::move(finish_callback)) { + explicit YuzuTest(Core::System& system_, std::string data_, + std::function<void(std::vector<TestResult>)> finish_callback_) + : ServiceFramework{system_, "yuzutest"}, data{std::move(data_)}, finish_callback{std::move( + finish_callback_)} { static const FunctionInfo functions[] = { {0, &YuzuTest::Initialize, "Initialize"}, {1, &YuzuTest::GetServiceVersion, "GetServiceVersion"}, @@ -104,9 +105,11 @@ private: std::function<void(std::vector<TestResult>)> finish_callback; }; -void InstallInterfaces(SM::ServiceManager& sm, std::string data, +void InstallInterfaces(Core::System& system, std::string data, std::function<void(std::vector<TestResult>)> finish_callback) { - std::make_shared<YuzuTest>(data, finish_callback)->InstallAsService(sm); + auto& sm = system.ServiceManager(); + std::make_shared<YuzuTest>(system, std::move(data), std::move(finish_callback)) + ->InstallAsService(sm); } } // namespace Service::Yuzu diff --git a/src/yuzu_tester/service/yuzutest.h b/src/yuzu_tester/service/yuzutest.h index eca129c8c..7794814fa 100644 --- a/src/yuzu_tester/service/yuzutest.h +++ b/src/yuzu_tester/service/yuzutest.h @@ -7,8 +7,8 @@ #include <functional> #include <string> -namespace Service::SM { -class ServiceManager; +namespace Core { +class System; } namespace Service::Yuzu { @@ -19,7 +19,7 @@ struct TestResult { std::string name; }; -void InstallInterfaces(SM::ServiceManager& sm, std::string data, +void InstallInterfaces(Core::System& system, std::string data, std::function<void(std::vector<TestResult>)> finish_callback); } // namespace Service::Yuzu diff --git a/src/yuzu_tester/yuzu.cpp b/src/yuzu_tester/yuzu.cpp index 5798ce43a..09cf2ad77 100644 --- a/src/yuzu_tester/yuzu.cpp +++ b/src/yuzu_tester/yuzu.cpp @@ -160,10 +160,11 @@ int main(int argc, char** argv) { return -1; } - Settings::values.use_gdbstub = false; - Settings::Apply(); + Core::System& system{Core::System::GetInstance()}; + + Settings::Apply(system); - std::unique_ptr<EmuWindow_SDL2_Hide> emu_window{std::make_unique<EmuWindow_SDL2_Hide>()}; + const auto emu_window{std::make_unique<EmuWindow_SDL2_Hide>()}; bool finished = false; int return_value = 0; @@ -212,7 +213,6 @@ int main(int argc, char** argv) { return_value = -1; }; - Core::System& system{Core::System::GetInstance()}; system.SetContentProvider(std::make_unique<FileSys::ContentProviderUnion>()); system.SetFilesystem(std::make_shared<FileSys::RealVfsFilesystem>()); system.GetFileSystemController().CreateFactories(*system.GetFilesystem()); @@ -242,25 +242,26 @@ int main(int argc, char** argv) { const u16 loader_id = static_cast<u16>(Core::System::ResultStatus::ErrorLoader); const u16 error_id = static_cast<u16>(load_result) - loader_id; LOG_CRITICAL(Frontend, - "While attempting to load the ROM requested, an error occured. Please " + "While attempting to load the ROM requested, an error occurred. Please " "refer to the yuzu wiki for more information or the yuzu discord for " "additional help.\n\nError Code: {:04X}-{:04X}\nError Description: {}", loader_id, error_id, static_cast<Loader::ResultStatus>(error_id)); } + break; } - Service::Yuzu::InstallInterfaces(system.ServiceManager(), datastring, callback); + Service::Yuzu::InstallInterfaces(system, datastring, callback); system.TelemetrySession().AddField(Common::Telemetry::FieldType::App, "Frontend", "SDLHideTester"); system.GPU().Start(); - system.Run(); + void(system.Run()); while (!finished) { std::this_thread::sleep_for(std::chrono::milliseconds(1)); } - system.Pause(); + void(system.Pause()); detached_tasks.WaitForAllTasks(); return return_value; |