diff options
Diffstat (limited to '')
211 files changed, 6511 insertions, 2815 deletions
diff --git a/.gitmodules b/.gitmodules index 2558a5ebc..3a49c4874 100644 --- a/.gitmodules +++ b/.gitmodules @@ -40,3 +40,9 @@ [submodule "Vulkan-Headers"] path = externals/Vulkan-Headers url = https://github.com/KhronosGroup/Vulkan-Headers.git +[submodule "externals/zstd"] + path = externals/zstd + url = https://github.com/facebook/zstd +[submodule "sirit"] + path = externals/sirit + url = https://github.com/ReinUsesLisp/sirit diff --git a/CMakeLists.txt b/CMakeLists.txt index ab18275d3..6a417017c 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -309,7 +309,7 @@ if (CLANG_FORMAT) set(CCOMMENT "Running clang format against all the .h and .cpp files in src/") if (WIN32) add_custom_target(clang-format - COMMAND powershell.exe -Command "Get-ChildItem ${SRCS}/* -Include *.cpp,*.h -Recurse | Foreach {${CLANG_FORMAT} -i $_.fullname}" + COMMAND powershell.exe -Command "Get-ChildItem '${SRCS}/*' -Include *.cpp,*.h -Recurse | Foreach {&'${CLANG_FORMAT}' -i $_.fullname}" COMMENT ${CCOMMENT}) elseif(MINGW) add_custom_target(clang-format diff --git a/externals/CMakeLists.txt b/externals/CMakeLists.txt index e156bbece..3f8b6cda8 100644 --- a/externals/CMakeLists.txt +++ b/externals/CMakeLists.txt @@ -49,6 +49,10 @@ add_subdirectory(open_source_archives EXCLUDE_FROM_ALL) add_library(unicorn-headers INTERFACE) target_include_directories(unicorn-headers INTERFACE ./unicorn/include) +# Zstandard +add_subdirectory(zstd/build/cmake EXCLUDE_FROM_ALL) +target_include_directories(libzstd_static INTERFACE ./zstd/lib) + # SoundTouch add_subdirectory(soundtouch) @@ -68,6 +72,11 @@ if (USE_DISCORD_PRESENCE) target_include_directories(discord-rpc INTERFACE ./discord-rpc/include) endif() +# Sirit +if (ENABLE_VULKAN) + add_subdirectory(sirit) +endif() + if (ENABLE_WEB_SERVICE) # LibreSSL set(LIBRESSL_SKIP_INSTALL ON CACHE BOOL "") diff --git a/externals/sirit b/externals/sirit new file mode 160000 +Subproject f7c4b07a7e14edb1dcd93bc9879c823423705c2 diff --git a/externals/zstd b/externals/zstd new file mode 160000 +Subproject 470344d33e1d52a2ada75d278466da8d4ee2faf diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt index 850ce8006..1e8e1b215 100644 --- a/src/common/CMakeLists.txt +++ b/src/common/CMakeLists.txt @@ -91,6 +91,8 @@ add_library(common STATIC logging/log.h logging/text_formatter.cpp logging/text_formatter.h + lz4_compression.cpp + lz4_compression.h math_util.h memory_hook.cpp memory_hook.h @@ -123,6 +125,8 @@ add_library(common STATIC uint128.h vector_math.h web_result.h + zstd_compression.cpp + zstd_compression.h ) if(ARCHITECTURE_x86_64) @@ -136,3 +140,4 @@ endif() create_target_directory_groups(common) target_link_libraries(common PUBLIC Boost::boost fmt microprofile) +target_link_libraries(common PRIVATE lz4_static libzstd_static) diff --git a/src/common/assert.h b/src/common/assert.h index 6002f7ab1..4b0e3f64e 100644 --- a/src/common/assert.h +++ b/src/common/assert.h @@ -57,3 +57,21 @@ __declspec(noinline, noreturn) #define UNIMPLEMENTED_IF(cond) ASSERT_MSG(!(cond), "Unimplemented code!") #define UNIMPLEMENTED_IF_MSG(cond, ...) ASSERT_MSG(!(cond), __VA_ARGS__) + +// If the assert is ignored, execute _b_ +#define ASSERT_OR_EXECUTE(_a_, _b_) \ + do { \ + ASSERT(_a_); \ + if (!(_a_)) { \ + _b_ \ + } \ + } while (0) + +// If the assert is ignored, execute _b_ +#define ASSERT_OR_EXECUTE_MSG(_a_, _b_, ...) \ + do { \ + ASSERT_MSG(_a_, __VA_ARGS__); \ + if (!(_a_)) { \ + _b_ \ + } \ + } while (0) diff --git a/src/common/bit_util.h b/src/common/bit_util.h index a4f9ed4aa..d032df413 100644 --- a/src/common/bit_util.h +++ b/src/common/bit_util.h @@ -32,7 +32,7 @@ inline u32 CountLeadingZeroes32(u32 value) { return 32; } -inline u64 CountLeadingZeroes64(u64 value) { +inline u32 CountLeadingZeroes64(u64 value) { unsigned long leading_zero = 0; if (_BitScanReverse64(&leading_zero, value) != 0) { @@ -47,15 +47,15 @@ inline u32 CountLeadingZeroes32(u32 value) { return 32; } - return __builtin_clz(value); + return static_cast<u32>(__builtin_clz(value)); } -inline u64 CountLeadingZeroes64(u64 value) { +inline u32 CountLeadingZeroes64(u64 value) { if (value == 0) { return 64; } - return __builtin_clzll(value); + return static_cast<u32>(__builtin_clzll(value)); } #endif @@ -70,7 +70,7 @@ inline u32 CountTrailingZeroes32(u32 value) { return 32; } -inline u64 CountTrailingZeroes64(u64 value) { +inline u32 CountTrailingZeroes64(u64 value) { unsigned long trailing_zero = 0; if (_BitScanForward64(&trailing_zero, value) != 0) { @@ -85,15 +85,15 @@ inline u32 CountTrailingZeroes32(u32 value) { return 32; } - return __builtin_ctz(value); + return static_cast<u32>(__builtin_ctz(value)); } -inline u64 CountTrailingZeroes64(u64 value) { +inline u32 CountTrailingZeroes64(u64 value) { if (value == 0) { return 64; } - return __builtin_ctzll(value); + return static_cast<u32>(__builtin_ctzll(value)); } #endif diff --git a/src/common/lz4_compression.cpp b/src/common/lz4_compression.cpp new file mode 100644 index 000000000..ade6759bb --- /dev/null +++ b/src/common/lz4_compression.cpp @@ -0,0 +1,76 @@ +// Copyright 2019 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <algorithm> +#include <lz4hc.h> + +#include "common/assert.h" +#include "common/lz4_compression.h" + +namespace Common::Compression { + +std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size) { + ASSERT_MSG(source_size <= LZ4_MAX_INPUT_SIZE, "Source size exceeds LZ4 maximum input size"); + + const auto source_size_int = static_cast<int>(source_size); + const int max_compressed_size = LZ4_compressBound(source_size_int); + std::vector<u8> compressed(max_compressed_size); + + const int compressed_size = LZ4_compress_default(reinterpret_cast<const char*>(source), + reinterpret_cast<char*>(compressed.data()), + source_size_int, max_compressed_size); + + if (compressed_size <= 0) { + // Compression failed + return {}; + } + + compressed.resize(compressed_size); + + return compressed; +} + +std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size, + s32 compression_level) { + ASSERT_MSG(source_size <= LZ4_MAX_INPUT_SIZE, "Source size exceeds LZ4 maximum input size"); + + compression_level = std::clamp(compression_level, LZ4HC_CLEVEL_MIN, LZ4HC_CLEVEL_MAX); + + const auto source_size_int = static_cast<int>(source_size); + const int max_compressed_size = LZ4_compressBound(source_size_int); + std::vector<u8> compressed(max_compressed_size); + + const int compressed_size = LZ4_compress_HC( + reinterpret_cast<const char*>(source), reinterpret_cast<char*>(compressed.data()), + source_size_int, max_compressed_size, compression_level); + + if (compressed_size <= 0) { + // Compression failed + return {}; + } + + compressed.resize(compressed_size); + + return compressed; +} + +std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size) { + return CompressDataLZ4HC(source, source_size, LZ4HC_CLEVEL_MAX); +} + +std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed, + std::size_t uncompressed_size) { + std::vector<u8> uncompressed(uncompressed_size); + const int size_check = LZ4_decompress_safe(reinterpret_cast<const char*>(compressed.data()), + reinterpret_cast<char*>(uncompressed.data()), + static_cast<int>(compressed.size()), + static_cast<int>(uncompressed.size())); + if (static_cast<int>(uncompressed_size) != size_check) { + // Decompression failed + return {}; + } + return uncompressed; +} + +} // namespace Common::Compression diff --git a/src/common/lz4_compression.h b/src/common/lz4_compression.h new file mode 100644 index 000000000..4c16f6e03 --- /dev/null +++ b/src/common/lz4_compression.h @@ -0,0 +1,57 @@ +// Copyright 2019 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::Compression { + +/** + * Compresses a source memory region with LZ4 and returns the compressed data in a vector. + * + * @param source the uncompressed source memory region. + * @param source_size the size in bytes of the uncompressed source memory region. + * + * @return the compressed data. + */ +std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size); + +/** + * Utilizes the LZ4 subalgorithm LZ4HC with the specified compression level. Higher compression + * levels result in a smaller compressed size, but require more CPU time for compression. The + * compression level has almost no impact on decompression speed. Data compressed with LZ4HC can + * also be decompressed with the default LZ4 decompression. + * + * @param source the uncompressed source memory region. + * @param source_size the size in bytes of the uncompressed source memory region. + * @param compression_level the used compression level. Should be between 3 and 12. + * + * @return the compressed data. + */ +std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size, s32 compression_level); + +/** + * Utilizes the LZ4 subalgorithm LZ4HC with the highest possible compression level. + * + * @param source the uncompressed source memory region. + * @param source_size the size in bytes of the uncompressed source memory region. + * + * @return the compressed data. + */ +std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size); + +/** + * Decompresses a source memory region with LZ4 and returns the uncompressed data in a vector. + * + * @param compressed the compressed source memory region. + * @param uncompressed_size the size in bytes of the uncompressed data. + * + * @return the decompressed data. + */ +std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed, std::size_t uncompressed_size); + +} // namespace Common::Compression
\ No newline at end of file diff --git a/src/common/multi_level_queue.h b/src/common/multi_level_queue.h index 2b61b91e0..9cb448f56 100644 --- a/src/common/multi_level_queue.h +++ b/src/common/multi_level_queue.h @@ -72,7 +72,7 @@ public: u64 prios = mlq.used_priorities; prios &= ~((1ULL << (current_priority + 1)) - 1); if (prios == 0) { - current_priority = mlq.depth(); + current_priority = static_cast<u32>(mlq.depth()); } else { current_priority = CountTrailingZeroes64(prios); it = GetBeginItForPrio(); diff --git a/src/common/scope_exit.h b/src/common/scope_exit.h index baf1f1c9e..1176a72b1 100644 --- a/src/common/scope_exit.h +++ b/src/common/scope_exit.h @@ -20,7 +20,7 @@ struct ScopeExitHelper { template <typename Func> ScopeExitHelper<Func> ScopeExit(Func&& func) { - return ScopeExitHelper<Func>(std::move(func)); + return ScopeExitHelper<Func>(std::forward<Func>(func)); } } // namespace detail diff --git a/src/common/swap.h b/src/common/swap.h index b3eab1324..71932c2bb 100644 --- a/src/common/swap.h +++ b/src/common/swap.h @@ -21,11 +21,6 @@ #if defined(_MSC_VER) #include <cstdlib> -#elif defined(__linux__) -#include <byteswap.h> -#elif defined(__Bitrig__) || defined(__DragonFly__) || defined(__FreeBSD__) || \ - defined(__NetBSD__) || defined(__OpenBSD__) -#include <sys/endian.h> #endif #include <cstring> #include "common/common_types.h" @@ -62,86 +57,49 @@ namespace Common { #ifdef _MSC_VER -inline u16 swap16(u16 _data) { - return _byteswap_ushort(_data); +[[nodiscard]] inline u16 swap16(u16 data) noexcept { + return _byteswap_ushort(data); } -inline u32 swap32(u32 _data) { - return _byteswap_ulong(_data); +[[nodiscard]] inline u32 swap32(u32 data) noexcept { + return _byteswap_ulong(data); } -inline u64 swap64(u64 _data) { - return _byteswap_uint64(_data); +[[nodiscard]] inline u64 swap64(u64 data) noexcept { + return _byteswap_uint64(data); } -#elif defined(ARCHITECTURE_ARM) && (__ARM_ARCH >= 6) -inline u16 swap16(u16 _data) { - u32 data = _data; - __asm__("rev16 %0, %1\n" : "=l"(data) : "l"(data)); - return (u16)data; -} -inline u32 swap32(u32 _data) { - __asm__("rev %0, %1\n" : "=l"(_data) : "l"(_data)); - return _data; -} -inline u64 swap64(u64 _data) { - return ((u64)swap32(_data) << 32) | swap32(_data >> 32); -} -#elif __linux__ -inline u16 swap16(u16 _data) { - return bswap_16(_data); -} -inline u32 swap32(u32 _data) { - return bswap_32(_data); -} -inline u64 swap64(u64 _data) { - return bswap_64(_data); -} -#elif __APPLE__ -inline __attribute__((always_inline)) u16 swap16(u16 _data) { - return (_data >> 8) | (_data << 8); -} -inline __attribute__((always_inline)) u32 swap32(u32 _data) { - return __builtin_bswap32(_data); -} -inline __attribute__((always_inline)) u64 swap64(u64 _data) { - return __builtin_bswap64(_data); -} -#elif defined(__Bitrig__) || defined(__OpenBSD__) +#elif defined(__clang__) || defined(__GNUC__) +#if defined(__Bitrig__) || defined(__OpenBSD__) // redefine swap16, swap32, swap64 as inline functions #undef swap16 #undef swap32 #undef swap64 -inline u16 swap16(u16 _data) { - return __swap16(_data); -} -inline u32 swap32(u32 _data) { - return __swap32(_data); -} -inline u64 swap64(u64 _data) { - return __swap64(_data); -} -#elif defined(__DragonFly__) || defined(__FreeBSD__) || defined(__NetBSD__) -inline u16 swap16(u16 _data) { - return bswap16(_data); +#endif +[[nodiscard]] inline u16 swap16(u16 data) noexcept { + return __builtin_bswap16(data); } -inline u32 swap32(u32 _data) { - return bswap32(_data); +[[nodiscard]] inline u32 swap32(u32 data) noexcept { + return __builtin_bswap32(data); } -inline u64 swap64(u64 _data) { - return bswap64(_data); +[[nodiscard]] inline u64 swap64(u64 data) noexcept { + return __builtin_bswap64(data); } #else -// Slow generic implementation. -inline u16 swap16(u16 data) { +// Generic implementation. +[[nodiscard]] inline u16 swap16(u16 data) noexcept { return (data >> 8) | (data << 8); } -inline u32 swap32(u32 data) { - return (swap16(data) << 16) | swap16(data >> 16); +[[nodiscard]] inline u32 swap32(u32 data) noexcept { + return ((data & 0xFF000000U) >> 24) | ((data & 0x00FF0000U) >> 8) | + ((data & 0x0000FF00U) << 8) | ((data & 0x000000FFU) << 24); } -inline u64 swap64(u64 data) { - return ((u64)swap32(data) << 32) | swap32(data >> 32); +[[nodiscard]] inline u64 swap64(u64 data) noexcept { + return ((data & 0xFF00000000000000ULL) >> 56) | ((data & 0x00FF000000000000ULL) >> 40) | + ((data & 0x0000FF0000000000ULL) >> 24) | ((data & 0x000000FF00000000ULL) >> 8) | + ((data & 0x00000000FF000000ULL) << 8) | ((data & 0x0000000000FF0000ULL) << 24) | + ((data & 0x000000000000FF00ULL) << 40) | ((data & 0x00000000000000FFULL) << 56); } #endif -inline float swapf(float f) { +[[nodiscard]] inline float swapf(float f) noexcept { static_assert(sizeof(u32) == sizeof(float), "float must be the same size as uint32_t."); u32 value; @@ -153,7 +111,7 @@ inline float swapf(float f) { return f; } -inline double swapd(double f) { +[[nodiscard]] inline double swapd(double f) noexcept { static_assert(sizeof(u64) == sizeof(double), "double must be the same size as uint64_t."); u64 value; diff --git a/src/common/zstd_compression.cpp b/src/common/zstd_compression.cpp new file mode 100644 index 000000000..60a35c67c --- /dev/null +++ b/src/common/zstd_compression.cpp @@ -0,0 +1,53 @@ +// Copyright 2019 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <algorithm> +#include <zstd.h> + +#include "common/assert.h" +#include "common/zstd_compression.h" + +namespace Common::Compression { + +std::vector<u8> CompressDataZSTD(const u8* source, std::size_t source_size, s32 compression_level) { + compression_level = std::clamp(compression_level, 1, ZSTD_maxCLevel()); + + const std::size_t max_compressed_size = ZSTD_compressBound(source_size); + std::vector<u8> compressed(max_compressed_size); + + const std::size_t compressed_size = + ZSTD_compress(compressed.data(), compressed.size(), source, source_size, compression_level); + + if (ZSTD_isError(compressed_size)) { + // Compression failed + return {}; + } + + compressed.resize(compressed_size); + + return compressed; +} + +std::vector<u8> CompressDataZSTDDefault(const u8* source, std::size_t source_size) { + return CompressDataZSTD(source, source_size, ZSTD_CLEVEL_DEFAULT); +} + +std::vector<u8> DecompressDataZSTD(const std::vector<u8>& compressed) { + const std::size_t decompressed_size = + ZSTD_getDecompressedSize(compressed.data(), compressed.size()); + std::vector<u8> decompressed(decompressed_size); + + const std::size_t uncompressed_result_size = ZSTD_decompress( + decompressed.data(), decompressed.size(), compressed.data(), compressed.size()); + + if (decompressed_size != uncompressed_result_size || ZSTD_isError(uncompressed_result_size)) { + // Decompression failed + return {}; + } + return decompressed; +} + +} // namespace Common::Compression diff --git a/src/common/zstd_compression.h b/src/common/zstd_compression.h new file mode 100644 index 000000000..e9de941c8 --- /dev/null +++ b/src/common/zstd_compression.h @@ -0,0 +1,44 @@ +// Copyright 2019 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::Compression { + +/** + * Compresses a source memory region with Zstandard and returns the compressed data in a vector. + * + * @param source the uncompressed source memory region. + * @param source_size the size in bytes of the uncompressed source memory region. + * @param compression_level the used compression level. Should be between 1 and 22. + * + * @return the compressed data. + */ +std::vector<u8> CompressDataZSTD(const u8* source, std::size_t source_size, s32 compression_level); + +/** + * Compresses a source memory region with Zstandard with the default compression level and returns + * the compressed data in a vector. + * + * @param source the uncompressed source memory region. + * @param source_size the size in bytes of the uncompressed source memory region. + * + * @return the compressed data. + */ +std::vector<u8> CompressDataZSTDDefault(const u8* source, std::size_t source_size); + +/** + * Decompresses a source memory region with Zstandard and returns the uncompressed data in a vector. + * + * @param compressed the compressed source memory region. + * + * @return the decompressed data. + */ +std::vector<u8> DecompressDataZSTD(const std::vector<u8>& compressed); + +} // namespace Common::Compression
\ No newline at end of file diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt index 9e23afe85..c59107102 100644 --- a/src/core/CMakeLists.txt +++ b/src/core/CMakeLists.txt @@ -458,7 +458,7 @@ add_library(core STATIC 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 lz4_static mbedtls opus unicorn open_source_archives) +target_link_libraries(core PUBLIC Boost::boost PRIVATE fmt mbedtls opus unicorn open_source_archives) if (ENABLE_WEB_SERVICE) target_compile_definitions(core PRIVATE -DENABLE_WEB_SERVICE) target_link_libraries(core PRIVATE web_service) diff --git a/src/core/arm/dynarmic/arm_dynarmic.cpp b/src/core/arm/dynarmic/arm_dynarmic.cpp index 4fdc12f11..dc96e35d5 100644 --- a/src/core/arm/dynarmic/arm_dynarmic.cpp +++ b/src/core/arm/dynarmic/arm_dynarmic.cpp @@ -14,6 +14,7 @@ #include "core/core_timing.h" #include "core/core_timing_util.h" #include "core/gdbstub/gdbstub.h" +#include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/svc.h" #include "core/hle/kernel/vm_manager.h" @@ -26,7 +27,6 @@ using Vector = Dynarmic::A64::Vector; class ARM_Dynarmic_Callbacks : public Dynarmic::A64::UserCallbacks { public: explicit ARM_Dynarmic_Callbacks(ARM_Dynarmic& parent) : parent(parent) {} - ~ARM_Dynarmic_Callbacks() = default; u8 MemoryRead8(u64 vaddr) override { return Memory::Read8(vaddr); @@ -100,7 +100,7 @@ public: } void CallSVC(u32 swi) override { - Kernel::CallSVC(swi); + Kernel::CallSVC(parent.system, swi); } void AddTicks(u64 ticks) override { @@ -113,14 +113,14 @@ public: // Always execute at least one tick. amortized_ticks = std::max<u64>(amortized_ticks, 1); - parent.core_timing.AddTicks(amortized_ticks); + parent.system.CoreTiming().AddTicks(amortized_ticks); num_interpreted_instructions = 0; } u64 GetTicksRemaining() override { - return std::max(parent.core_timing.GetDowncount(), 0); + return std::max(parent.system.CoreTiming().GetDowncount(), 0); } u64 GetCNTPCT() override { - return Timing::CpuCyclesToClockCycles(parent.core_timing.GetTicks()); + return Timing::CpuCyclesToClockCycles(parent.system.CoreTiming().GetTicks()); } ARM_Dynarmic& parent; @@ -130,7 +130,7 @@ public: }; std::unique_ptr<Dynarmic::A64::Jit> ARM_Dynarmic::MakeJit() const { - auto* current_process = Core::CurrentProcess(); + auto* current_process = system.Kernel().CurrentProcess(); auto** const page_table = current_process->VMManager().page_table.pointers.data(); Dynarmic::A64::UserConfig config; @@ -164,7 +164,6 @@ MICROPROFILE_DEFINE(ARM_Jit_Dynarmic, "ARM JIT", "Dynarmic", MP_RGB(255, 64, 64) void ARM_Dynarmic::Run() { MICROPROFILE_SCOPE(ARM_Jit_Dynarmic); - ASSERT(Memory::GetCurrentPageTable() == current_page_table); jit->Run(); } @@ -173,10 +172,10 @@ void ARM_Dynarmic::Step() { cb->InterpreterFallback(jit->GetPC(), 1); } -ARM_Dynarmic::ARM_Dynarmic(Timing::CoreTiming& core_timing, ExclusiveMonitor& exclusive_monitor, +ARM_Dynarmic::ARM_Dynarmic(System& system, ExclusiveMonitor& exclusive_monitor, std::size_t core_index) - : cb(std::make_unique<ARM_Dynarmic_Callbacks>(*this)), inner_unicorn{core_timing}, - core_index{core_index}, core_timing{core_timing}, + : cb(std::make_unique<ARM_Dynarmic_Callbacks>(*this)), inner_unicorn{system}, + core_index{core_index}, system{system}, exclusive_monitor{dynamic_cast<DynarmicExclusiveMonitor&>(exclusive_monitor)} { ThreadContext ctx{}; inner_unicorn.SaveContext(ctx); @@ -279,7 +278,6 @@ void ARM_Dynarmic::ClearExclusiveState() { void ARM_Dynarmic::PageTableChanged() { jit = MakeJit(); - current_page_table = Memory::GetCurrentPageTable(); } DynarmicExclusiveMonitor::DynarmicExclusiveMonitor(std::size_t core_count) : monitor(core_count) {} diff --git a/src/core/arm/dynarmic/arm_dynarmic.h b/src/core/arm/dynarmic/arm_dynarmic.h index aada1e862..c1db254e8 100644 --- a/src/core/arm/dynarmic/arm_dynarmic.h +++ b/src/core/arm/dynarmic/arm_dynarmic.h @@ -12,24 +12,16 @@ #include "core/arm/exclusive_monitor.h" #include "core/arm/unicorn/arm_unicorn.h" -namespace Common { -struct PageTable; -} - -namespace Core::Timing { -class CoreTiming; -} - namespace Core { class ARM_Dynarmic_Callbacks; class DynarmicExclusiveMonitor; +class System; class ARM_Dynarmic final : public ARM_Interface { public: - ARM_Dynarmic(Timing::CoreTiming& core_timing, ExclusiveMonitor& exclusive_monitor, - std::size_t core_index); - ~ARM_Dynarmic(); + ARM_Dynarmic(System& system, ExclusiveMonitor& exclusive_monitor, std::size_t core_index); + ~ARM_Dynarmic() override; void MapBackingMemory(VAddr address, std::size_t size, u8* memory, Kernel::VMAPermission perms) override; @@ -67,16 +59,14 @@ private: ARM_Unicorn inner_unicorn; std::size_t core_index; - Timing::CoreTiming& core_timing; + System& system; DynarmicExclusiveMonitor& exclusive_monitor; - - Common::PageTable* current_page_table = nullptr; }; class DynarmicExclusiveMonitor final : public ExclusiveMonitor { public: explicit DynarmicExclusiveMonitor(std::size_t core_count); - ~DynarmicExclusiveMonitor(); + ~DynarmicExclusiveMonitor() override; void SetExclusive(std::size_t core_index, VAddr addr) override; void ClearExclusive() override; diff --git a/src/core/arm/unicorn/arm_unicorn.cpp b/src/core/arm/unicorn/arm_unicorn.cpp index a542a098b..4e07fe8b5 100644 --- a/src/core/arm/unicorn/arm_unicorn.cpp +++ b/src/core/arm/unicorn/arm_unicorn.cpp @@ -10,7 +10,6 @@ #include "core/core.h" #include "core/core_timing.h" #include "core/hle/kernel/svc.h" -#include "core/memory.h" namespace Core { @@ -49,20 +48,6 @@ static void CodeHook(uc_engine* uc, uint64_t address, uint32_t size, void* user_ } } -static void InterruptHook(uc_engine* uc, u32 intNo, void* user_data) { - u32 esr{}; - CHECKED(uc_reg_read(uc, UC_ARM64_REG_ESR, &esr)); - - auto ec = esr >> 26; - auto iss = esr & 0xFFFFFF; - - switch (ec) { - case 0x15: // SVC - Kernel::CallSVC(iss); - break; - } -} - static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int size, u64 value, void* user_data) { ARM_Interface::ThreadContext ctx{}; @@ -72,7 +57,7 @@ static bool UnmappedMemoryHook(uc_engine* uc, uc_mem_type type, u64 addr, int si return {}; } -ARM_Unicorn::ARM_Unicorn(Timing::CoreTiming& core_timing) : core_timing{core_timing} { +ARM_Unicorn::ARM_Unicorn(System& system) : system{system} { CHECKED(uc_open(UC_ARCH_ARM64, UC_MODE_ARM, &uc)); auto fpv = 3 << 20; @@ -177,7 +162,7 @@ void ARM_Unicorn::Run() { if (GDBStub::IsServerEnabled()) { ExecuteInstructions(std::max(4000000, 0)); } else { - ExecuteInstructions(std::max(core_timing.GetDowncount(), 0)); + ExecuteInstructions(std::max(system.CoreTiming().GetDowncount(), 0)); } } @@ -190,14 +175,15 @@ MICROPROFILE_DEFINE(ARM_Jit_Unicorn, "ARM JIT", "Unicorn", MP_RGB(255, 64, 64)); void ARM_Unicorn::ExecuteInstructions(int num_instructions) { MICROPROFILE_SCOPE(ARM_Jit_Unicorn); CHECKED(uc_emu_start(uc, GetPC(), 1ULL << 63, 0, num_instructions)); - core_timing.AddTicks(num_instructions); + system.CoreTiming().AddTicks(num_instructions); if (GDBStub::IsServerEnabled()) { - if (last_bkpt_hit) { + if (last_bkpt_hit && last_bkpt.type == GDBStub::BreakpointType::Execute) { uc_reg_write(uc, UC_ARM64_REG_PC, &last_bkpt.address); } + Kernel::Thread* thread = Kernel::GetCurrentThread(); SaveContext(thread->GetContext()); - if (last_bkpt_hit || GDBStub::GetCpuStepFlag()) { + if (last_bkpt_hit || GDBStub::IsMemoryBreak() || GDBStub::GetCpuStepFlag()) { last_bkpt_hit = false; GDBStub::Break(); GDBStub::SendTrap(thread, 5); @@ -272,4 +258,20 @@ void ARM_Unicorn::RecordBreak(GDBStub::BreakpointAddress 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::CallSVC(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 index dbd6955ea..209fc16ad 100644 --- a/src/core/arm/unicorn/arm_unicorn.h +++ b/src/core/arm/unicorn/arm_unicorn.h @@ -9,16 +9,14 @@ #include "core/arm/arm_interface.h" #include "core/gdbstub/gdbstub.h" -namespace Core::Timing { -class CoreTiming; -} - namespace Core { +class System; + class ARM_Unicorn final : public ARM_Interface { public: - explicit ARM_Unicorn(Timing::CoreTiming& core_timing); - ~ARM_Unicorn(); + explicit ARM_Unicorn(System& system); + ~ARM_Unicorn() override; void MapBackingMemory(VAddr address, std::size_t size, u8* memory, Kernel::VMAPermission perms) override; @@ -47,10 +45,12 @@ public: void RecordBreak(GDBStub::BreakpointAddress bkpt); private: + static void InterruptHook(uc_engine* uc, u32 int_no, void* user_data); + uc_engine* uc{}; - Timing::CoreTiming& core_timing; + System& system; GDBStub::BreakpointAddress last_bkpt{}; - bool last_bkpt_hit; + bool last_bkpt_hit = false; }; } // namespace Core diff --git a/src/core/core.cpp b/src/core/core.cpp index 4fe77c25b..bc9e887b6 100644 --- a/src/core/core.cpp +++ b/src/core/core.cpp @@ -17,6 +17,7 @@ #include "core/core_timing.h" #include "core/cpu_core_manager.h" #include "core/file_sys/mode.h" +#include "core/file_sys/registered_cache.h" #include "core/file_sys/vfs_concat.h" #include "core/file_sys/vfs_real.h" #include "core/gdbstub/gdbstub.h" @@ -108,6 +109,8 @@ struct System::Impl { // Create a default fs if one doesn't already exist. if (virtual_filesystem == nullptr) virtual_filesystem = std::make_shared<FileSys::RealVfsFilesystem>(); + if (content_provider == nullptr) + content_provider = std::make_unique<FileSys::ContentProviderUnion>(); /// Create default implementations of applets if one is not provided. if (profile_selector == nullptr) @@ -249,6 +252,8 @@ struct System::Impl { Kernel::KernelCore kernel; /// RealVfsFilesystem instance FileSys::VirtualFilesystem virtual_filesystem; + /// ContentProviderUnion instance + std::unique_ptr<FileSys::ContentProviderUnion> content_provider; /// AppLoader used to load the current executing application std::unique_ptr<Loader::AppLoader> app_loader; std::unique_ptr<VideoCore::RendererBase> renderer; @@ -488,6 +493,27 @@ const Frontend::SoftwareKeyboardApplet& System::GetSoftwareKeyboard() const { return *impl->software_keyboard; } +void System::SetContentProvider(std::unique_ptr<FileSys::ContentProviderUnion> provider) { + impl->content_provider = std::move(provider); +} + +FileSys::ContentProvider& System::GetContentProvider() { + return *impl->content_provider; +} + +const FileSys::ContentProvider& System::GetContentProvider() const { + return *impl->content_provider; +} + +void System::RegisterContentProvider(FileSys::ContentProviderUnionSlot slot, + FileSys::ContentProvider* provider) { + impl->content_provider->SetSlot(slot, provider); +} + +void System::ClearContentProvider(FileSys::ContentProviderUnionSlot slot) { + impl->content_provider->ClearSlot(slot); +} + void System::SetWebBrowser(std::unique_ptr<Frontend::WebBrowserApplet> applet) { impl->web_browser = std::move(applet); } diff --git a/src/core/core.h b/src/core/core.h index 4d83b93cc..82b2e087e 100644 --- a/src/core/core.h +++ b/src/core/core.h @@ -21,6 +21,9 @@ class WebBrowserApplet; namespace FileSys { class CheatList; +class ContentProvider; +class ContentProviderUnion; +enum class ContentProviderUnionSlot; class VfsFilesystem; } // namespace FileSys @@ -270,6 +273,17 @@ public: Frontend::WebBrowserApplet& GetWebBrowser(); const Frontend::WebBrowserApplet& GetWebBrowser() const; + void SetContentProvider(std::unique_ptr<FileSys::ContentProviderUnion> provider); + + FileSys::ContentProvider& GetContentProvider(); + + const FileSys::ContentProvider& GetContentProvider() const; + + void RegisterContentProvider(FileSys::ContentProviderUnionSlot slot, + FileSys::ContentProvider* provider); + + void ClearContentProvider(FileSys::ContentProviderUnionSlot slot); + private: System(); diff --git a/src/core/core_cpu.cpp b/src/core/core_cpu.cpp index e75741db0..ba63c3e61 100644 --- a/src/core/core_cpu.cpp +++ b/src/core/core_cpu.cpp @@ -55,13 +55,13 @@ Cpu::Cpu(System& system, ExclusiveMonitor& exclusive_monitor, CpuBarrier& cpu_ba : cpu_barrier{cpu_barrier}, core_timing{system.CoreTiming()}, core_index{core_index} { if (Settings::values.use_cpu_jit) { #ifdef ARCHITECTURE_x86_64 - arm_interface = std::make_unique<ARM_Dynarmic>(core_timing, exclusive_monitor, core_index); + arm_interface = std::make_unique<ARM_Dynarmic>(system, exclusive_monitor, core_index); #else - arm_interface = std::make_unique<ARM_Unicorn>(); + arm_interface = std::make_unique<ARM_Unicorn>(system); LOG_WARNING(Core, "CPU JIT requested, but Dynarmic not available"); #endif } else { - arm_interface = std::make_unique<ARM_Unicorn>(core_timing); + arm_interface = std::make_unique<ARM_Unicorn>(system); } scheduler = std::make_unique<Kernel::Scheduler>(system, *arm_interface); diff --git a/src/core/crypto/key_manager.cpp b/src/core/crypto/key_manager.cpp index dfac9a4b3..dc006e2bb 100644 --- a/src/core/crypto/key_manager.cpp +++ b/src/core/crypto/key_manager.cpp @@ -22,6 +22,7 @@ #include "common/file_util.h" #include "common/hex_util.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/crypto/aes_util.h" #include "core/crypto/key_manager.h" #include "core/crypto/partition_data_manager.h" @@ -794,7 +795,7 @@ void KeyManager::DeriveBase() { void KeyManager::DeriveETicket(PartitionDataManager& data) { // ETicket keys - const auto es = Service::FileSystem::GetUnionContents().GetEntry( + const auto es = Core::System::GetInstance().GetContentProvider().GetEntry( 0x0100000000000033, FileSys::ContentRecordType::Program); if (es == nullptr) diff --git a/src/core/file_sys/control_metadata.cpp b/src/core/file_sys/control_metadata.cpp index 83c184750..60ea9ad12 100644 --- a/src/core/file_sys/control_metadata.cpp +++ b/src/core/file_sys/control_metadata.cpp @@ -67,7 +67,7 @@ std::string NACP::GetDeveloperName(Language language) const { } u64 NACP::GetTitleId() const { - return raw.title_id; + return raw.save_data_owner_id; } u64 NACP::GetDLCBaseTitleId() const { @@ -80,11 +80,11 @@ std::string NACP::GetVersionString() const { } u64 NACP::GetDefaultNormalSaveSize() const { - return raw.normal_save_data_size; + return raw.user_account_save_data_size; } u64 NACP::GetDefaultJournalSaveSize() const { - return raw.journal_sava_data_size; + return raw.user_account_save_data_journal_size; } std::vector<u8> NACP::GetRawBytes() const { diff --git a/src/core/file_sys/control_metadata.h b/src/core/file_sys/control_metadata.h index 7b9cdc910..280710ddf 100644 --- a/src/core/file_sys/control_metadata.h +++ b/src/core/file_sys/control_metadata.h @@ -38,23 +38,35 @@ struct RawNACP { u8 video_capture_mode; bool data_loss_confirmation; INSERT_PADDING_BYTES(1); - u64_le title_id; + u64_le presence_group_id; std::array<u8, 0x20> rating_age; std::array<char, 0x10> version_string; u64_le dlc_base_title_id; - u64_le title_id_2; - u64_le normal_save_data_size; - u64_le journal_sava_data_size; - INSERT_PADDING_BYTES(0x18); - u64_le product_code; + u64_le save_data_owner_id; + u64_le user_account_save_data_size; + u64_le user_account_save_data_journal_size; + u64_le device_save_data_size; + u64_le device_save_data_journal_size; + u64_le bcat_delivery_cache_storage_size; + char application_error_code_category[8]; std::array<u64_le, 0x8> local_communication; u8 logo_type; u8 logo_handling; bool runtime_add_on_content_install; INSERT_PADDING_BYTES(5); - u64_le title_id_update; - std::array<u8, 0x40> bcat_passphrase; - INSERT_PADDING_BYTES(0xEC0); + u64_le seed_for_pseudo_device_id; + std::array<u8, 0x41> bcat_passphrase; + INSERT_PADDING_BYTES(7); + u64_le user_account_save_data_max_size; + u64_le user_account_save_data_max_journal_size; + u64_le device_save_data_max_size; + u64_le device_save_data_max_journal_size; + u64_le temporary_storage_size; + u64_le cache_storage_size; + u64_le cache_storage_journal_size; + u64_le cache_storage_data_and_journal_max_size; + u64_le cache_storage_max_index; + INSERT_PADDING_BYTES(0xE70); }; static_assert(sizeof(RawNACP) == 0x4000, "RawNACP has incorrect size."); diff --git a/src/core/file_sys/fsmitm_romfsbuild.cpp b/src/core/file_sys/fsmitm_romfsbuild.cpp index 47b7526c7..d126ae8dd 100644 --- a/src/core/file_sys/fsmitm_romfsbuild.cpp +++ b/src/core/file_sys/fsmitm_romfsbuild.cpp @@ -23,6 +23,7 @@ */ #include <cstring> +#include <string_view> #include "common/alignment.h" #include "common/assert.h" #include "core/file_sys/fsmitm_romfsbuild.h" @@ -97,7 +98,8 @@ struct RomFSBuildFileContext { VirtualFile source; }; -static u32 romfs_calc_path_hash(u32 parent, std::string path, u32 start, std::size_t path_len) { +static u32 romfs_calc_path_hash(u32 parent, std::string_view path, u32 start, + std::size_t path_len) { u32 hash = parent ^ 123456789; for (u32 i = 0; i < path_len; i++) { hash = (hash >> 5) | (hash << 27); diff --git a/src/core/file_sys/nca_metadata.cpp b/src/core/file_sys/nca_metadata.cpp index 6f34b7836..93d0df6b9 100644 --- a/src/core/file_sys/nca_metadata.cpp +++ b/src/core/file_sys/nca_metadata.cpp @@ -10,14 +10,6 @@ namespace FileSys { -bool operator>=(TitleType lhs, TitleType rhs) { - return static_cast<std::size_t>(lhs) >= static_cast<std::size_t>(rhs); -} - -bool operator<=(TitleType lhs, TitleType rhs) { - return static_cast<std::size_t>(lhs) <= static_cast<std::size_t>(rhs); -} - CNMT::CNMT(VirtualFile file) { if (file->ReadObject(&header) != sizeof(CNMTHeader)) return; diff --git a/src/core/file_sys/nca_metadata.h b/src/core/file_sys/nca_metadata.h index a05d155f4..50bf38471 100644 --- a/src/core/file_sys/nca_metadata.h +++ b/src/core/file_sys/nca_metadata.h @@ -29,9 +29,6 @@ enum class TitleType : u8 { DeltaTitle = 0x83, }; -bool operator>=(TitleType lhs, TitleType rhs); -bool operator<=(TitleType lhs, TitleType rhs); - enum class ContentRecordType : u8 { Meta = 0, Program = 1, diff --git a/src/core/file_sys/patch_manager.cpp b/src/core/file_sys/patch_manager.cpp index e11217708..78dbadee3 100644 --- a/src/core/file_sys/patch_manager.cpp +++ b/src/core/file_sys/patch_manager.cpp @@ -10,6 +10,7 @@ #include "common/file_util.h" #include "common/hex_util.h" #include "common/logging/log.h" +#include "core/core.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" #include "core/file_sys/ips_layer.h" @@ -69,7 +70,7 @@ VirtualDir PatchManager::PatchExeFS(VirtualDir exefs) const { } } - const auto installed = Service::FileSystem::GetUnionContents(); + const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto& disabled = Settings::values.disabled_addons[title_id]; const auto update_disabled = @@ -155,7 +156,7 @@ std::vector<VirtualFile> PatchManager::CollectPatches(const std::vector<VirtualD return out; } -std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso) const { +std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso, const std::string& name) const { if (nso.size() < sizeof(Loader::NSOHeader)) { return nso; } @@ -171,18 +172,19 @@ std::vector<u8> PatchManager::PatchNSO(const std::vector<u8>& nso) const { const auto build_id = build_id_raw.substr(0, build_id_raw.find_last_not_of('0') + 1); if (Settings::values.dump_nso) { - LOG_INFO(Loader, "Dumping NSO for build_id={}, title_id={:016X}", build_id, title_id); + LOG_INFO(Loader, "Dumping NSO for name={}, build_id={}, title_id={:016X}", name, build_id, + title_id); const auto dump_dir = Service::FileSystem::GetModificationDumpRoot(title_id); if (dump_dir != nullptr) { const auto nso_dir = GetOrCreateDirectoryRelative(dump_dir, "/nso"); - const auto file = nso_dir->CreateFile(fmt::format("{}.nso", build_id)); + const auto file = nso_dir->CreateFile(fmt::format("{}-{}.nso", name, build_id)); file->Resize(nso.size()); file->WriteBytes(nso); } } - LOG_INFO(Loader, "Patching NSO for build_id={}", build_id); + LOG_INFO(Loader, "Patching NSO for name={}, build_id={}", name, build_id); const auto load_dir = Service::FileSystem::GetModificationLoadRoot(title_id); auto patch_dirs = load_dir->GetSubdirectories(); @@ -345,7 +347,7 @@ VirtualFile PatchManager::PatchRomFS(VirtualFile romfs, u64 ivfc_offset, Content if (romfs == nullptr) return romfs; - const auto installed = Service::FileSystem::GetUnionContents(); + const auto& installed = Core::System::GetInstance().GetContentProvider(); // Game Updates const auto update_tid = GetUpdateTitleID(title_id); @@ -392,7 +394,7 @@ static bool IsDirValidAndNonEmpty(const VirtualDir& dir) { std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNames( VirtualFile update_raw) const { std::map<std::string, std::string, std::less<>> out; - const auto installed = Service::FileSystem::GetUnionContents(); + const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto& disabled = Settings::values.disabled_addons[title_id]; // Game Updates @@ -466,10 +468,10 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam // DLC const auto dlc_entries = installed.ListEntriesFilter(TitleType::AOC, ContentRecordType::Data); - std::vector<RegisteredCacheEntry> dlc_match; + 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 RegisteredCacheEntry& entry) { + [this, &installed](const ContentProviderEntry& entry) { return (entry.title_id & DLC_BASE_TITLE_ID_MASK) == title_id && installed.GetEntry(entry)->GetStatus() == Loader::ResultStatus::Success; }); @@ -492,7 +494,7 @@ std::map<std::string, std::string, std::less<>> PatchManager::GetPatchVersionNam } std::pair<std::unique_ptr<NACP>, VirtualFile> PatchManager::GetControlMetadata() const { - const auto installed{Service::FileSystem::GetUnionContents()}; + const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto base_control_nca = installed.GetEntry(title_id, ContentRecordType::Control); if (base_control_nca == nullptr) diff --git a/src/core/file_sys/patch_manager.h b/src/core/file_sys/patch_manager.h index de2672c76..769f8c6f0 100644 --- a/src/core/file_sys/patch_manager.h +++ b/src/core/file_sys/patch_manager.h @@ -44,7 +44,7 @@ public: // Currently tracked NSO patches: // - IPS // - IPSwitch - std::vector<u8> PatchNSO(const std::vector<u8>& nso) const; + std::vector<u8> PatchNSO(const std::vector<u8>& nso, const std::string& name) const; // Checks to see if PatchNSO() will have any effect given the NSO's build ID. // Used to prevent expensive copies in NSO loader. diff --git a/src/core/file_sys/program_metadata.cpp b/src/core/file_sys/program_metadata.cpp index d3e00437f..d863253f8 100644 --- a/src/core/file_sys/program_metadata.cpp +++ b/src/core/file_sys/program_metadata.cpp @@ -3,7 +3,6 @@ // Refer to the license.txt file included. #include <cstddef> -#include <cstring> #include <vector> #include "common/logging/log.h" @@ -17,28 +16,30 @@ ProgramMetadata::ProgramMetadata() = default; ProgramMetadata::~ProgramMetadata() = default; Loader::ResultStatus ProgramMetadata::Load(VirtualFile file) { - std::size_t total_size = static_cast<std::size_t>(file->GetSize()); - if (total_size < sizeof(Header)) + const std::size_t total_size = file->GetSize(); + if (total_size < sizeof(Header)) { return Loader::ResultStatus::ErrorBadNPDMHeader; + } - // TODO(DarkLordZach): Use ReadObject when Header/AcidHeader becomes trivially copyable. - std::vector<u8> npdm_header_data = file->ReadBytes(sizeof(Header)); - if (sizeof(Header) != npdm_header_data.size()) + if (sizeof(Header) != file->ReadObject(&npdm_header)) { return Loader::ResultStatus::ErrorBadNPDMHeader; - std::memcpy(&npdm_header, npdm_header_data.data(), sizeof(Header)); + } - std::vector<u8> acid_header_data = file->ReadBytes(sizeof(AcidHeader), npdm_header.acid_offset); - if (sizeof(AcidHeader) != acid_header_data.size()) + if (sizeof(AcidHeader) != file->ReadObject(&acid_header, npdm_header.acid_offset)) { return Loader::ResultStatus::ErrorBadACIDHeader; - std::memcpy(&acid_header, acid_header_data.data(), sizeof(AcidHeader)); + } - if (sizeof(AciHeader) != file->ReadObject(&aci_header, npdm_header.aci_offset)) + if (sizeof(AciHeader) != file->ReadObject(&aci_header, npdm_header.aci_offset)) { return Loader::ResultStatus::ErrorBadACIHeader; + } - if (sizeof(FileAccessControl) != file->ReadObject(&acid_file_access, acid_header.fac_offset)) + if (sizeof(FileAccessControl) != file->ReadObject(&acid_file_access, acid_header.fac_offset)) { return Loader::ResultStatus::ErrorBadFileAccessControl; - if (sizeof(FileAccessHeader) != file->ReadObject(&aci_file_access, aci_header.fah_offset)) + } + + if (sizeof(FileAccessHeader) != file->ReadObject(&aci_file_access, aci_header.fah_offset)) { return Loader::ResultStatus::ErrorBadFileAccessHeader; + } aci_kernel_capabilities.resize(aci_header.kac_size / sizeof(u32)); const u64 read_size = aci_header.kac_size; diff --git a/src/core/file_sys/program_metadata.h b/src/core/file_sys/program_metadata.h index 0033ba347..7de5b9cf9 100644 --- a/src/core/file_sys/program_metadata.h +++ b/src/core/file_sys/program_metadata.h @@ -58,7 +58,6 @@ public: void Print() const; private: - // TODO(DarkLordZach): BitField is not trivially copyable. struct Header { std::array<char, 4> magic; std::array<u8, 8> reserved; @@ -85,7 +84,6 @@ private: static_assert(sizeof(Header) == 0x80, "NPDM header structure size is wrong"); - // TODO(DarkLordZach): BitField is not trivially copyable. struct AcidHeader { std::array<u8, 0x100> signature; std::array<u8, 0x100> nca_modulus; diff --git a/src/core/file_sys/registered_cache.cpp b/src/core/file_sys/registered_cache.cpp index 1c6bacace..3946ff871 100644 --- a/src/core/file_sys/registered_cache.cpp +++ b/src/core/file_sys/registered_cache.cpp @@ -23,19 +23,19 @@ namespace FileSys { // The size of blocks to use when vfs raw copying into nand. constexpr size_t VFS_RC_LARGE_COPY_BLOCK = 0x400000; -std::string RegisteredCacheEntry::DebugInfo() const { +std::string ContentProviderEntry::DebugInfo() const { return fmt::format("title_id={:016X}, content_type={:02X}", title_id, static_cast<u8>(type)); } -bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) { +bool operator<(const ContentProviderEntry& lhs, const ContentProviderEntry& rhs) { return (lhs.title_id < rhs.title_id) || (lhs.title_id == rhs.title_id && lhs.type < rhs.type); } -bool operator==(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) { +bool operator==(const ContentProviderEntry& lhs, const ContentProviderEntry& rhs) { return std::tie(lhs.title_id, lhs.type) == std::tie(rhs.title_id, rhs.type); } -bool operator!=(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs) { +bool operator!=(const ContentProviderEntry& lhs, const ContentProviderEntry& rhs) { return !operator==(lhs, rhs); } @@ -84,7 +84,7 @@ static std::string GetCNMTName(TitleType type, u64 title_id) { return fmt::format("{}_{:016x}.cnmt", TITLE_TYPE_NAMES[index], title_id); } -static ContentRecordType GetCRTypeFromNCAType(NCAContentType type) { +ContentRecordType GetCRTypeFromNCAType(NCAContentType type) { switch (type) { case NCAContentType::Program: // TODO(DarkLordZach): Differentiate between Program and Patch @@ -104,6 +104,28 @@ static ContentRecordType GetCRTypeFromNCAType(NCAContentType type) { } } +ContentProvider::~ContentProvider() = default; + +bool ContentProvider::HasEntry(ContentProviderEntry entry) const { + return HasEntry(entry.title_id, entry.type); +} + +VirtualFile ContentProvider::GetEntryUnparsed(ContentProviderEntry entry) const { + return GetEntryUnparsed(entry.title_id, entry.type); +} + +VirtualFile ContentProvider::GetEntryRaw(ContentProviderEntry entry) const { + return GetEntryRaw(entry.title_id, entry.type); +} + +std::unique_ptr<NCA> ContentProvider::GetEntry(ContentProviderEntry entry) const { + return GetEntry(entry.title_id, entry.type); +} + +std::vector<ContentProviderEntry> ContentProvider::ListEntries() const { + return ListEntriesFilter(std::nullopt, std::nullopt, std::nullopt); +} + VirtualFile RegisteredCache::OpenFileOrDirectoryConcat(const VirtualDir& dir, std::string_view path) const { const auto file = dir->GetFileRelative(path); @@ -161,8 +183,8 @@ VirtualFile RegisteredCache::GetFileAtID(NcaID id) const { return file; } -static std::optional<NcaID> CheckMapForContentRecord( - const boost::container::flat_map<u64, CNMT>& map, u64 title_id, ContentRecordType type) { +static std::optional<NcaID> CheckMapForContentRecord(const std::map<u64, CNMT>& map, u64 title_id, + ContentRecordType type) { if (map.find(title_id) == map.end()) return {}; @@ -268,7 +290,7 @@ void RegisteredCache::Refresh() { AccumulateYuzuMeta(); } -RegisteredCache::RegisteredCache(VirtualDir dir_, RegisteredCacheParsingFunction parsing_function) +RegisteredCache::RegisteredCache(VirtualDir dir_, ContentProviderParsingFunction parsing_function) : dir(std::move(dir_)), parser(std::move(parsing_function)) { Refresh(); } @@ -279,19 +301,11 @@ bool RegisteredCache::HasEntry(u64 title_id, ContentRecordType type) const { return GetEntryRaw(title_id, type) != nullptr; } -bool RegisteredCache::HasEntry(RegisteredCacheEntry entry) const { - return GetEntryRaw(entry) != nullptr; -} - VirtualFile RegisteredCache::GetEntryUnparsed(u64 title_id, ContentRecordType type) const { const auto id = GetNcaIDFromMetadata(title_id, type); return id ? GetFileAtID(*id) : nullptr; } -VirtualFile RegisteredCache::GetEntryUnparsed(RegisteredCacheEntry entry) const { - return GetEntryUnparsed(entry.title_id, entry.type); -} - std::optional<u32> RegisteredCache::GetEntryVersion(u64 title_id) const { const auto meta_iter = meta.find(title_id); if (meta_iter != meta.end()) @@ -309,10 +323,6 @@ VirtualFile RegisteredCache::GetEntryRaw(u64 title_id, ContentRecordType type) c return id ? parser(GetFileAtID(*id), *id) : nullptr; } -VirtualFile RegisteredCache::GetEntryRaw(RegisteredCacheEntry entry) const { - return GetEntryRaw(entry.title_id, entry.type); -} - std::unique_ptr<NCA> RegisteredCache::GetEntry(u64 title_id, ContentRecordType type) const { const auto raw = GetEntryRaw(title_id, type); if (raw == nullptr) @@ -320,10 +330,6 @@ std::unique_ptr<NCA> RegisteredCache::GetEntry(u64 title_id, ContentRecordType t return std::make_unique<NCA>(raw, nullptr, 0, keys); } -std::unique_ptr<NCA> RegisteredCache::GetEntry(RegisteredCacheEntry entry) const { - return GetEntry(entry.title_id, entry.type); -} - template <typename T> void RegisteredCache::IterateAllMetadata( std::vector<T>& out, std::function<T(const CNMT&, const ContentRecord&)> proc, @@ -348,25 +354,14 @@ void RegisteredCache::IterateAllMetadata( } } -std::vector<RegisteredCacheEntry> RegisteredCache::ListEntries() const { - std::vector<RegisteredCacheEntry> out; - IterateAllMetadata<RegisteredCacheEntry>( - out, - [](const CNMT& c, const ContentRecord& r) { - return RegisteredCacheEntry{c.GetTitleID(), r.type}; - }, - [](const CNMT& c, const ContentRecord& r) { return true; }); - return out; -} - -std::vector<RegisteredCacheEntry> RegisteredCache::ListEntriesFilter( +std::vector<ContentProviderEntry> RegisteredCache::ListEntriesFilter( std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type, std::optional<u64> title_id) const { - std::vector<RegisteredCacheEntry> out; - IterateAllMetadata<RegisteredCacheEntry>( + std::vector<ContentProviderEntry> out; + IterateAllMetadata<ContentProviderEntry>( out, [](const CNMT& c, const ContentRecord& r) { - return RegisteredCacheEntry{c.GetTitleID(), r.type}; + return ContentProviderEntry{c.GetTitleID(), r.type}; }, [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) { if (title_type && *title_type != c.GetType()) @@ -521,37 +516,56 @@ bool RegisteredCache::RawInstallYuzuMeta(const CNMT& cnmt) { }) != yuzu_meta.end(); } -RegisteredCacheUnion::RegisteredCacheUnion(std::vector<RegisteredCache*> caches) - : caches(std::move(caches)) {} +ContentProviderUnion::~ContentProviderUnion() = default; -void RegisteredCacheUnion::Refresh() { - for (const auto& c : caches) - c->Refresh(); +void ContentProviderUnion::SetSlot(ContentProviderUnionSlot slot, ContentProvider* provider) { + providers[slot] = provider; } -bool RegisteredCacheUnion::HasEntry(u64 title_id, ContentRecordType type) const { - return std::any_of(caches.begin(), caches.end(), [title_id, type](const auto& cache) { - return cache->HasEntry(title_id, type); - }); +void ContentProviderUnion::ClearSlot(ContentProviderUnionSlot slot) { + providers[slot] = nullptr; } -bool RegisteredCacheUnion::HasEntry(RegisteredCacheEntry entry) const { - return HasEntry(entry.title_id, entry.type); +void ContentProviderUnion::Refresh() { + for (auto& provider : providers) { + if (provider.second == nullptr) + continue; + + provider.second->Refresh(); + } } -std::optional<u32> RegisteredCacheUnion::GetEntryVersion(u64 title_id) const { - for (const auto& c : caches) { - const auto res = c->GetEntryVersion(title_id); - if (res) +bool ContentProviderUnion::HasEntry(u64 title_id, ContentRecordType type) const { + for (const auto& provider : providers) { + if (provider.second == nullptr) + continue; + + if (provider.second->HasEntry(title_id, type)) + return true; + } + + return false; +} + +std::optional<u32> ContentProviderUnion::GetEntryVersion(u64 title_id) const { + for (const auto& provider : providers) { + if (provider.second == nullptr) + continue; + + const auto res = provider.second->GetEntryVersion(title_id); + if (res != std::nullopt) return res; } - return {}; + return std::nullopt; } -VirtualFile RegisteredCacheUnion::GetEntryUnparsed(u64 title_id, ContentRecordType type) const { - for (const auto& c : caches) { - const auto res = c->GetEntryUnparsed(title_id, type); +VirtualFile ContentProviderUnion::GetEntryUnparsed(u64 title_id, ContentRecordType type) const { + for (const auto& provider : providers) { + if (provider.second == nullptr) + continue; + + const auto res = provider.second->GetEntryUnparsed(title_id, type); if (res != nullptr) return res; } @@ -559,13 +573,12 @@ VirtualFile RegisteredCacheUnion::GetEntryUnparsed(u64 title_id, ContentRecordTy return nullptr; } -VirtualFile RegisteredCacheUnion::GetEntryUnparsed(RegisteredCacheEntry entry) const { - return GetEntryUnparsed(entry.title_id, entry.type); -} +VirtualFile ContentProviderUnion::GetEntryRaw(u64 title_id, ContentRecordType type) const { + for (const auto& provider : providers) { + if (provider.second == nullptr) + continue; -VirtualFile RegisteredCacheUnion::GetEntryRaw(u64 title_id, ContentRecordType type) const { - for (const auto& c : caches) { - const auto res = c->GetEntryRaw(title_id, type); + const auto res = provider.second->GetEntryRaw(title_id, type); if (res != nullptr) return res; } @@ -573,30 +586,56 @@ VirtualFile RegisteredCacheUnion::GetEntryRaw(u64 title_id, ContentRecordType ty return nullptr; } -VirtualFile RegisteredCacheUnion::GetEntryRaw(RegisteredCacheEntry entry) const { - return GetEntryRaw(entry.title_id, entry.type); -} +std::unique_ptr<NCA> ContentProviderUnion::GetEntry(u64 title_id, ContentRecordType type) const { + for (const auto& provider : providers) { + if (provider.second == nullptr) + continue; -std::unique_ptr<NCA> RegisteredCacheUnion::GetEntry(u64 title_id, ContentRecordType type) const { - const auto raw = GetEntryRaw(title_id, type); - if (raw == nullptr) - return nullptr; - return std::make_unique<NCA>(raw); + auto res = provider.second->GetEntry(title_id, type); + if (res != nullptr) + return res; + } + + return nullptr; } -std::unique_ptr<NCA> RegisteredCacheUnion::GetEntry(RegisteredCacheEntry entry) const { - return GetEntry(entry.title_id, entry.type); +std::vector<ContentProviderEntry> ContentProviderUnion::ListEntriesFilter( + std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type, + std::optional<u64> title_id) const { + std::vector<ContentProviderEntry> out; + + for (const auto& provider : providers) { + if (provider.second == nullptr) + continue; + + const auto vec = provider.second->ListEntriesFilter(title_type, record_type, title_id); + std::copy(vec.begin(), vec.end(), std::back_inserter(out)); + } + + std::sort(out.begin(), out.end()); + out.erase(std::unique(out.begin(), out.end()), out.end()); + return out; } -std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntries() const { - std::vector<RegisteredCacheEntry> out; - for (const auto& c : caches) { - c->IterateAllMetadata<RegisteredCacheEntry>( - out, - [](const CNMT& c, const ContentRecord& r) { - return RegisteredCacheEntry{c.GetTitleID(), r.type}; - }, - [](const CNMT& c, const ContentRecord& r) { return true; }); +std::vector<std::pair<ContentProviderUnionSlot, ContentProviderEntry>> +ContentProviderUnion::ListEntriesFilterOrigin(std::optional<ContentProviderUnionSlot> origin, + std::optional<TitleType> title_type, + std::optional<ContentRecordType> record_type, + std::optional<u64> title_id) const { + std::vector<std::pair<ContentProviderUnionSlot, ContentProviderEntry>> out; + + for (const auto& provider : providers) { + if (provider.second == nullptr) + continue; + + if (origin.has_value() && *origin != provider.first) + continue; + + const auto vec = provider.second->ListEntriesFilter(title_type, record_type, title_id); + std::transform(vec.begin(), vec.end(), std::back_inserter(out), + [&provider](const ContentProviderEntry& entry) { + return std::make_pair(provider.first, entry); + }); } std::sort(out.begin(), out.end()); @@ -604,25 +643,61 @@ std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntries() const { return out; } -std::vector<RegisteredCacheEntry> RegisteredCacheUnion::ListEntriesFilter( +ManualContentProvider::~ManualContentProvider() = default; + +void ManualContentProvider::AddEntry(TitleType title_type, ContentRecordType content_type, + u64 title_id, VirtualFile file) { + entries.insert_or_assign({title_type, content_type, title_id}, file); +} + +void ManualContentProvider::ClearAllEntries() { + entries.clear(); +} + +void ManualContentProvider::Refresh() {} + +bool ManualContentProvider::HasEntry(u64 title_id, ContentRecordType type) const { + return GetEntryRaw(title_id, type) != nullptr; +} + +std::optional<u32> ManualContentProvider::GetEntryVersion(u64 title_id) const { + return std::nullopt; +} + +VirtualFile ManualContentProvider::GetEntryUnparsed(u64 title_id, ContentRecordType type) const { + return GetEntryRaw(title_id, type); +} + +VirtualFile ManualContentProvider::GetEntryRaw(u64 title_id, ContentRecordType type) const { + const auto iter = + std::find_if(entries.begin(), entries.end(), [title_id, type](const auto& entry) { + const auto [title_type, content_type, e_title_id] = entry.first; + return content_type == type && e_title_id == title_id; + }); + if (iter == entries.end()) + return nullptr; + return iter->second; +} + +std::unique_ptr<NCA> ManualContentProvider::GetEntry(u64 title_id, ContentRecordType type) const { + const auto res = GetEntryRaw(title_id, type); + if (res == nullptr) + return nullptr; + return std::make_unique<NCA>(res, nullptr, 0, keys); +} + +std::vector<ContentProviderEntry> ManualContentProvider::ListEntriesFilter( std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type, std::optional<u64> title_id) const { - std::vector<RegisteredCacheEntry> out; - for (const auto& c : caches) { - c->IterateAllMetadata<RegisteredCacheEntry>( - out, - [](const CNMT& c, const ContentRecord& r) { - return RegisteredCacheEntry{c.GetTitleID(), r.type}; - }, - [&title_type, &record_type, &title_id](const CNMT& c, const ContentRecord& r) { - if (title_type && *title_type != c.GetType()) - return false; - if (record_type && *record_type != r.type) - return false; - if (title_id && *title_id != c.GetTitleID()) - return false; - return true; - }); + std::vector<ContentProviderEntry> out; + + for (const auto& entry : entries) { + const auto [e_title_type, e_content_type, e_title_id] = entry.first; + if ((title_type == std::nullopt || e_title_type == *title_type) && + (record_type == std::nullopt || e_content_type == *record_type) && + (title_id == std::nullopt || e_title_id == *title_id)) { + out.emplace_back(ContentProviderEntry{e_title_id, e_content_type}); + } } std::sort(out.begin(), out.end()); diff --git a/src/core/file_sys/registered_cache.h b/src/core/file_sys/registered_cache.h index 3b77af4e0..ec9052653 100644 --- a/src/core/file_sys/registered_cache.h +++ b/src/core/file_sys/registered_cache.h @@ -21,12 +21,13 @@ class NSP; class XCI; enum class ContentRecordType : u8; +enum class NCAContentType : u8; enum class TitleType : u8; struct ContentRecord; using NcaID = std::array<u8, 0x10>; -using RegisteredCacheParsingFunction = std::function<VirtualFile(const VirtualFile&, const NcaID&)>; +using ContentProviderParsingFunction = std::function<VirtualFile(const VirtualFile&, const NcaID&)>; using VfsCopyFunction = std::function<bool(const VirtualFile&, const VirtualFile&, size_t)>; enum class InstallResult { @@ -36,7 +37,7 @@ enum class InstallResult { ErrorMetaFailed, }; -struct RegisteredCacheEntry { +struct ContentProviderEntry { u64 title_id; ContentRecordType type; @@ -47,12 +48,46 @@ constexpr u64 GetUpdateTitleID(u64 base_title_id) { return base_title_id | 0x800; } +ContentRecordType GetCRTypeFromNCAType(NCAContentType type); + // boost flat_map requires operator< for O(log(n)) lookups. -bool operator<(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs); +bool operator<(const ContentProviderEntry& lhs, const ContentProviderEntry& rhs); // std unique requires operator== to identify duplicates. -bool operator==(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs); -bool operator!=(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs); +bool operator==(const ContentProviderEntry& lhs, const ContentProviderEntry& rhs); +bool operator!=(const ContentProviderEntry& lhs, const ContentProviderEntry& rhs); + +class ContentProvider { +public: + virtual ~ContentProvider(); + + virtual void Refresh() = 0; + + virtual bool HasEntry(u64 title_id, ContentRecordType type) const = 0; + virtual 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; + + virtual VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const = 0; + virtual 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; + + virtual std::vector<ContentProviderEntry> ListEntries() const; + + // If a parameter is not std::nullopt, it will be filtered for from all entries. + virtual std::vector<ContentProviderEntry> ListEntriesFilter( + std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {}, + std::optional<u64> title_id = {}) const = 0; + +protected: + // A single instance of KeyManager to be used by GetEntry() + Core::Crypto::KeyManager keys; +}; /* * A class that catalogues NCAs in the registered directory structure. @@ -67,39 +102,32 @@ bool operator!=(const RegisteredCacheEntry& lhs, const RegisteredCacheEntry& rhs * (This impl also supports substituting the nca dir for an nca file, as that's more convenient * when 4GB splitting can be ignored.) */ -class RegisteredCache { - friend class RegisteredCacheUnion; - +class RegisteredCache : public ContentProvider { public: // Parsing function defines the conversion from raw file to NCA. If there are other steps // besides creating the NCA from the file (e.g. NAX0 on SD Card), that should go in a custom // parsing function. explicit RegisteredCache(VirtualDir dir, - RegisteredCacheParsingFunction parsing_function = + ContentProviderParsingFunction parsing_function = [](const VirtualFile& file, const NcaID& id) { return file; }); - ~RegisteredCache(); + ~RegisteredCache() override; - void Refresh(); + void Refresh() override; - bool HasEntry(u64 title_id, ContentRecordType type) const; - bool HasEntry(RegisteredCacheEntry entry) const; + bool HasEntry(u64 title_id, ContentRecordType type) const override; - std::optional<u32> GetEntryVersion(u64 title_id) const; + std::optional<u32> GetEntryVersion(u64 title_id) const override; - VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const; - VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const; + VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const override; - VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const; - VirtualFile GetEntryRaw(RegisteredCacheEntry entry) const; + VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const override; - std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const; - std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const; + std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const override; - std::vector<RegisteredCacheEntry> ListEntries() const; // If a parameter is not std::nullopt, it will be filtered for from all entries. - std::vector<RegisteredCacheEntry> ListEntriesFilter( + std::vector<ContentProviderEntry> ListEntriesFilter( std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {}, - std::optional<u64> title_id = {}) const; + std::optional<u64> title_id = {}) const override; // Raw copies all the ncas from the xci/nsp to the csache. Does some quick checks to make sure // there is a meta NCA and all of them are accessible. @@ -131,46 +159,70 @@ private: bool RawInstallYuzuMeta(const CNMT& cnmt); VirtualDir dir; - RegisteredCacheParsingFunction parser; - Core::Crypto::KeyManager keys; + ContentProviderParsingFunction parser; // maps tid -> NcaID of meta - boost::container::flat_map<u64, NcaID> meta_id; + std::map<u64, NcaID> meta_id; // maps tid -> meta - boost::container::flat_map<u64, CNMT> meta; + std::map<u64, CNMT> meta; // maps tid -> meta for CNMT in yuzu_meta - boost::container::flat_map<u64, CNMT> yuzu_meta; + std::map<u64, CNMT> yuzu_meta; }; -// Combines multiple RegisteredCaches (i.e. SysNAND, UserNAND, SDMC) into one interface. -class RegisteredCacheUnion { -public: - explicit RegisteredCacheUnion(std::vector<RegisteredCache*> caches); - - void Refresh(); - - bool HasEntry(u64 title_id, ContentRecordType type) const; - bool HasEntry(RegisteredCacheEntry entry) const; - - std::optional<u32> GetEntryVersion(u64 title_id) const; - - VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const; - VirtualFile GetEntryUnparsed(RegisteredCacheEntry entry) const; - - VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const; - VirtualFile GetEntryRaw(RegisteredCacheEntry entry) const; - - std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const; - std::unique_ptr<NCA> GetEntry(RegisteredCacheEntry entry) const; +enum class ContentProviderUnionSlot { + SysNAND, ///< System NAND + UserNAND, ///< User NAND + SDMC, ///< SD Card + FrontendManual, ///< Frontend-defined game list or similar +}; - std::vector<RegisteredCacheEntry> ListEntries() const; - // If a parameter is not std::nullopt, it will be filtered for from all entries. - std::vector<RegisteredCacheEntry> ListEntriesFilter( +// Combines multiple ContentProvider(s) (i.e. SysNAND, UserNAND, SDMC) into one interface. +class ContentProviderUnion : public ContentProvider { +public: + ~ContentProviderUnion() override; + + void SetSlot(ContentProviderUnionSlot slot, ContentProvider* provider); + void ClearSlot(ContentProviderUnionSlot slot); + + void Refresh() override; + bool HasEntry(u64 title_id, ContentRecordType type) const override; + std::optional<u32> GetEntryVersion(u64 title_id) const override; + VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const override; + VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const override; + std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const override; + std::vector<ContentProviderEntry> ListEntriesFilter( + std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type, + std::optional<u64> title_id) const override; + + std::vector<std::pair<ContentProviderUnionSlot, ContentProviderEntry>> ListEntriesFilterOrigin( + std::optional<ContentProviderUnionSlot> origin = {}, std::optional<TitleType> title_type = {}, std::optional<ContentRecordType> record_type = {}, std::optional<u64> title_id = {}) const; private: - std::vector<RegisteredCache*> caches; + std::map<ContentProviderUnionSlot, ContentProvider*> providers; +}; + +class ManualContentProvider : public ContentProvider { +public: + ~ManualContentProvider() override; + + void AddEntry(TitleType title_type, ContentRecordType content_type, u64 title_id, + VirtualFile file); + void ClearAllEntries(); + + void Refresh() override; + bool HasEntry(u64 title_id, ContentRecordType type) const override; + std::optional<u32> GetEntryVersion(u64 title_id) const override; + VirtualFile GetEntryUnparsed(u64 title_id, ContentRecordType type) const override; + VirtualFile GetEntryRaw(u64 title_id, ContentRecordType type) const override; + std::unique_ptr<NCA> GetEntry(u64 title_id, ContentRecordType type) const override; + std::vector<ContentProviderEntry> ListEntriesFilter( + std::optional<TitleType> title_type, std::optional<ContentRecordType> record_type, + std::optional<u64> title_id) const override; + +private: + std::map<std::tuple<TitleType, ContentRecordType, u64>, VirtualFile> entries; }; } // namespace FileSys diff --git a/src/core/file_sys/romfs_factory.cpp b/src/core/file_sys/romfs_factory.cpp index 6ad1e4f86..b2ccb2926 100644 --- a/src/core/file_sys/romfs_factory.cpp +++ b/src/core/file_sys/romfs_factory.cpp @@ -48,7 +48,7 @@ ResultVal<VirtualFile> RomFSFactory::Open(u64 title_id, StorageId storage, Conte switch (storage) { case StorageId::None: - res = Service::FileSystem::GetUnionContents().GetEntry(title_id, type); + res = Core::System::GetInstance().GetContentProvider().GetEntry(title_id, type); break; case StorageId::NandSystem: res = Service::FileSystem::GetSystemNANDContents()->GetEntry(title_id, type); diff --git a/src/core/file_sys/savedata_factory.cpp b/src/core/file_sys/savedata_factory.cpp index 1913dc956..7974b031d 100644 --- a/src/core/file_sys/savedata_factory.cpp +++ b/src/core/file_sys/savedata_factory.cpp @@ -16,8 +16,10 @@ namespace FileSys { constexpr char SAVE_DATA_SIZE_FILENAME[] = ".yuzu_save_size"; std::string SaveDataDescriptor::DebugInfo() const { - return fmt::format("[type={:02X}, title_id={:016X}, user_id={:016X}{:016X}, save_id={:016X}]", - static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id); + return fmt::format("[type={:02X}, title_id={:016X}, user_id={:016X}{:016X}, save_id={:016X}, " + "rank={}, index={}]", + static_cast<u8>(type), title_id, user_id[1], user_id[0], save_id, + static_cast<u8>(rank), index); } SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save_directory)) { @@ -28,7 +30,7 @@ SaveDataFactory::SaveDataFactory(VirtualDir save_directory) : dir(std::move(save SaveDataFactory::~SaveDataFactory() = default; -ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, SaveDataDescriptor meta) { +ResultVal<VirtualDir> SaveDataFactory::Open(SaveDataSpaceId space, const SaveDataDescriptor& meta) { if (meta.type == SaveDataType::SystemSaveData || meta.type == SaveDataType::SaveData) { if (meta.zero_1 != 0) { LOG_WARNING(Service_FS, diff --git a/src/core/file_sys/savedata_factory.h b/src/core/file_sys/savedata_factory.h index 3a1caf292..b73654571 100644 --- a/src/core/file_sys/savedata_factory.h +++ b/src/core/file_sys/savedata_factory.h @@ -32,12 +32,19 @@ enum class SaveDataType : u8 { CacheStorage = 5, }; +enum class SaveDataRank : u8 { + Primary, + Secondary, +}; + struct SaveDataDescriptor { u64_le title_id; u128 user_id; u64_le save_id; SaveDataType type; - INSERT_PADDING_BYTES(7); + SaveDataRank rank; + u16_le index; + INSERT_PADDING_BYTES(4); u64_le zero_1; u64_le zero_2; u64_le zero_3; @@ -57,7 +64,7 @@ public: explicit SaveDataFactory(VirtualDir dir); ~SaveDataFactory(); - ResultVal<VirtualDir> Open(SaveDataSpaceId space, SaveDataDescriptor meta); + ResultVal<VirtualDir> Open(SaveDataSpaceId space, const SaveDataDescriptor& meta); VirtualDir GetSaveDataSpaceDirectory(SaveDataSpaceId space) const; diff --git a/src/core/file_sys/submission_package.cpp b/src/core/file_sys/submission_package.cpp index e1a4210db..c69caae0f 100644 --- a/src/core/file_sys/submission_package.cpp +++ b/src/core/file_sys/submission_package.cpp @@ -143,11 +143,12 @@ std::multimap<u64, std::shared_ptr<NCA>> NSP::GetNCAsByTitleID() const { return out; } -std::map<u64, std::map<ContentRecordType, std::shared_ptr<NCA>>> NSP::GetNCAs() const { +std::map<u64, std::map<std::pair<TitleType, ContentRecordType>, std::shared_ptr<NCA>>> +NSP::GetNCAs() const { return ncas; } -std::shared_ptr<NCA> NSP::GetNCA(u64 title_id, ContentRecordType type) const { +std::shared_ptr<NCA> NSP::GetNCA(u64 title_id, ContentRecordType type, TitleType title_type) const { if (extracted) LOG_WARNING(Service_FS, "called on an NSP that is of type extracted."); @@ -155,14 +156,14 @@ std::shared_ptr<NCA> NSP::GetNCA(u64 title_id, ContentRecordType type) const { if (title_id_iter == ncas.end()) return nullptr; - const auto type_iter = title_id_iter->second.find(type); + const auto type_iter = title_id_iter->second.find({title_type, type}); if (type_iter == title_id_iter->second.end()) return nullptr; return type_iter->second; } -VirtualFile NSP::GetNCAFile(u64 title_id, ContentRecordType type) const { +VirtualFile NSP::GetNCAFile(u64 title_id, ContentRecordType type, TitleType title_type) const { if (extracted) LOG_WARNING(Service_FS, "called on an NSP that is of type extracted."); const auto nca = GetNCA(title_id, type); @@ -240,7 +241,7 @@ void NSP::ReadNCAs(const std::vector<VirtualFile>& files) { const CNMT cnmt(inner_file); auto& ncas_title = ncas[cnmt.GetTitleID()]; - ncas_title[ContentRecordType::Meta] = nca; + ncas_title[{cnmt.GetType(), ContentRecordType::Meta}] = nca; for (const auto& rec : cnmt.GetContentRecords()) { const auto id_string = Common::HexArrayToString(rec.nca_id, false); const auto next_file = pfs->GetFile(fmt::format("{}.nca", id_string)); @@ -258,7 +259,7 @@ void NSP::ReadNCAs(const std::vector<VirtualFile>& files) { if (next_nca->GetStatus() == Loader::ResultStatus::Success || (next_nca->GetStatus() == Loader::ResultStatus::ErrorMissingBKTRBaseRomFS && (cnmt.GetTitleID() & 0x800) != 0)) { - ncas_title[rec.type] = std::move(next_nca); + ncas_title[{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 9a28ed5bb..ee9b6ce17 100644 --- a/src/core/file_sys/submission_package.h +++ b/src/core/file_sys/submission_package.h @@ -42,9 +42,12 @@ public: // Type 0 Only (Collection of NCAs + Certificate + Ticket + Meta XML) std::vector<std::shared_ptr<NCA>> GetNCAsCollapsed() const; std::multimap<u64, std::shared_ptr<NCA>> GetNCAsByTitleID() const; - std::map<u64, std::map<ContentRecordType, std::shared_ptr<NCA>>> GetNCAs() const; - std::shared_ptr<NCA> GetNCA(u64 title_id, ContentRecordType type) const; - VirtualFile GetNCAFile(u64 title_id, ContentRecordType type) const; + std::map<u64, std::map<std::pair<TitleType, ContentRecordType>, std::shared_ptr<NCA>>> GetNCAs() + const; + std::shared_ptr<NCA> GetNCA(u64 title_id, ContentRecordType type, + TitleType title_type = TitleType::Application) const; + VirtualFile GetNCAFile(u64 title_id, ContentRecordType type, + TitleType title_type = TitleType::Application) const; std::vector<Core::Crypto::Key128> GetTitlekey() const; std::vector<VirtualFile> GetFiles() const override; @@ -67,7 +70,7 @@ private: std::shared_ptr<PartitionFilesystem> pfs; // Map title id -> {map type -> NCA} - std::map<u64, std::map<ContentRecordType, std::shared_ptr<NCA>>> ncas; + std::map<u64, std::map<std::pair<TitleType, ContentRecordType>, std::shared_ptr<NCA>>> ncas; std::vector<VirtualFile> ticket_files; Core::Crypto::KeyManager keys; diff --git a/src/core/frontend/emu_window.h b/src/core/frontend/emu_window.h index d0bcb4660..70a522556 100644 --- a/src/core/frontend/emu_window.h +++ b/src/core/frontend/emu_window.h @@ -13,6 +13,23 @@ namespace Core::Frontend { /** + * Represents a graphics context that can be used for background computation or drawing. If the + * graphics backend doesn't require the context, then the implementation of these methods can be + * stubs + */ +class GraphicsContext { +public: + /// Makes the graphics context current for the caller thread + virtual void MakeCurrent() = 0; + + /// Releases (dunno if this is the "right" word) the context from the caller thread + virtual void DoneCurrent() = 0; + + /// Swap buffers to display the next frame + virtual void SwapBuffers() = 0; +}; + +/** * Abstraction class used to provide an interface between emulation code and the frontend * (e.g. SDL, QGLWidget, GLFW, etc...). * @@ -30,7 +47,7 @@ namespace Core::Frontend { * - DO NOT TREAT THIS CLASS AS A GUI TOOLKIT ABSTRACTION LAYER. That's not what it is. Please * re-read the upper points again and think about it if you don't see this. */ -class EmuWindow { +class EmuWindow : public GraphicsContext { public: /// Data structure to store emuwindow configuration struct WindowConfig { @@ -40,17 +57,21 @@ public: std::pair<unsigned, unsigned> min_client_area_size; }; - /// Swap buffers to display the next frame - virtual void SwapBuffers() = 0; - /// Polls window events virtual void PollEvents() = 0; - /// Makes the graphics context current for the caller thread - virtual void MakeCurrent() = 0; - - /// Releases (dunno if this is the "right" word) the GLFW context from the caller thread - virtual void DoneCurrent() = 0; + /** + * Returns a GraphicsContext that the frontend provides that is shared with the emu window. This + * context can be used from other threads for background graphics computation. If the frontend + * is using a graphics backend that doesn't need anything specific to run on a different thread, + * then it can use a stubbed implemenation for GraphicsContext. + * + * If the return value is null, then the core should assume that the frontend cannot provide a + * Shared Context + */ + virtual std::unique_ptr<GraphicsContext> CreateSharedContext() const { + return nullptr; + } /** * Signal that a touch pressed event has occurred (e.g. mouse click pressed) diff --git a/src/core/gdbstub/gdbstub.cpp b/src/core/gdbstub/gdbstub.cpp index dafb32aae..afa812598 100644 --- a/src/core/gdbstub/gdbstub.cpp +++ b/src/core/gdbstub/gdbstub.cpp @@ -1030,7 +1030,7 @@ static void Step() { /// Tell the CPU if we hit a memory breakpoint. bool IsMemoryBreak() { - if (IsConnected()) { + if (!IsConnected()) { return false; } diff --git a/src/core/hle/ipc_helpers.h b/src/core/hle/ipc_helpers.h index 68406eb63..ac0e1d796 100644 --- a/src/core/hle/ipc_helpers.h +++ b/src/core/hle/ipc_helpers.h @@ -139,10 +139,8 @@ public: context->AddDomainObject(std::move(iface)); } else { auto& kernel = Core::System::GetInstance().Kernel(); - auto sessions = + auto [server, client] = Kernel::ServerSession::CreateSessionPair(kernel, iface->GetServiceName()); - auto server = std::get<Kernel::SharedPtr<Kernel::ServerSession>>(sessions); - auto client = std::get<Kernel::SharedPtr<Kernel::ClientSession>>(sessions); iface->ClientConnected(server); context->AddMoveObject(std::move(client)); } diff --git a/src/core/hle/kernel/client_port.cpp b/src/core/hle/kernel/client_port.cpp index aa432658e..744b1697d 100644 --- a/src/core/hle/kernel/client_port.cpp +++ b/src/core/hle/kernel/client_port.cpp @@ -2,8 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <tuple> - #include "core/hle/kernel/client_port.h" #include "core/hle/kernel/client_session.h" #include "core/hle/kernel/errors.h" @@ -31,18 +29,18 @@ ResultVal<SharedPtr<ClientSession>> ClientPort::Connect() { active_sessions++; // Create a new session pair, let the created sessions inherit the parent port's HLE handler. - auto sessions = ServerSession::CreateSessionPair(kernel, server_port->GetName(), this); + auto [server, client] = ServerSession::CreateSessionPair(kernel, server_port->GetName(), this); if (server_port->HasHLEHandler()) { - server_port->GetHLEHandler()->ClientConnected(std::get<SharedPtr<ServerSession>>(sessions)); + server_port->GetHLEHandler()->ClientConnected(server); } else { - server_port->AppendPendingSession(std::get<SharedPtr<ServerSession>>(sessions)); + server_port->AppendPendingSession(server); } // Wake the threads waiting on the ServerPort server_port->WakeupAllWaitingThreads(); - return MakeResult(std::get<SharedPtr<ClientSession>>(sessions)); + return MakeResult(client); } void ClientPort::ConnectionClosed() { diff --git a/src/core/hle/kernel/client_port.h b/src/core/hle/kernel/client_port.h index 6cd607206..4921ad4f0 100644 --- a/src/core/hle/kernel/client_port.h +++ b/src/core/hle/kernel/client_port.h @@ -25,7 +25,7 @@ public: return name; } - static const HandleType HANDLE_TYPE = HandleType::ClientPort; + static constexpr HandleType HANDLE_TYPE = HandleType::ClientPort; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/kernel/client_session.h b/src/core/hle/kernel/client_session.h index b1f39aad7..09cdff588 100644 --- a/src/core/hle/kernel/client_session.h +++ b/src/core/hle/kernel/client_session.h @@ -29,7 +29,7 @@ public: return name; } - static const HandleType HANDLE_TYPE = HandleType::ClientSession; + static constexpr HandleType HANDLE_TYPE = HandleType::ClientSession; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp index 3f14bfa86..4d58e7c69 100644 --- a/src/core/hle/kernel/kernel.cpp +++ b/src/core/hle/kernel/kernel.cpp @@ -21,6 +21,7 @@ #include "core/hle/kernel/thread.h" #include "core/hle/lock.h" #include "core/hle/result.h" +#include "core/memory.h" namespace Kernel { @@ -181,6 +182,7 @@ void KernelCore::AppendNewProcess(SharedPtr<Process> process) { void KernelCore::MakeCurrentProcess(Process* process) { impl->current_process = process; + Memory::SetCurrentPageTable(&process->VMManager().page_table); } Process* KernelCore::CurrentProcess() { diff --git a/src/core/hle/kernel/object.cpp b/src/core/hle/kernel/object.cpp index 217144efc..10431e94c 100644 --- a/src/core/hle/kernel/object.cpp +++ b/src/core/hle/kernel/object.cpp @@ -24,7 +24,6 @@ bool Object::IsWaitable() const { case HandleType::WritableEvent: case HandleType::SharedMemory: case HandleType::TransferMemory: - case HandleType::AddressArbiter: case HandleType::ResourceLimit: case HandleType::ClientPort: case HandleType::ClientSession: diff --git a/src/core/hle/kernel/object.h b/src/core/hle/kernel/object.h index 3f6baa094..332876c27 100644 --- a/src/core/hle/kernel/object.h +++ b/src/core/hle/kernel/object.h @@ -25,7 +25,6 @@ enum class HandleType : u32 { TransferMemory, Thread, Process, - AddressArbiter, ResourceLimit, ClientPort, ServerPort, diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp index 041267318..4e94048da 100644 --- a/src/core/hle/kernel/process.cpp +++ b/src/core/hle/kernel/process.cpp @@ -32,9 +32,6 @@ namespace { * @param priority The priority to give the main thread */ void SetupMainThread(Process& owner_process, KernelCore& kernel, VAddr entry_point, u32 priority) { - // Setup page table so we can write to memory - Memory::SetCurrentPageTable(&owner_process.VMManager().page_table); - // Initialize new "main" thread const VAddr stack_top = owner_process.VMManager().GetTLSIORegionEndAddress(); auto thread_res = Thread::Create(kernel, "main", entry_point, priority, 0, @@ -109,6 +106,8 @@ ResultCode Process::LoadFromMetadata(const FileSys::ProgramMetadata& metadata) { is_64bit_process = metadata.Is64BitProgram(); vm_manager.Reset(metadata.GetAddressSpaceType()); + // Ensure that the potentially resized page table is seen by CPU backends. + Memory::SetCurrentPageTable(&vm_manager.page_table); const auto& caps = metadata.GetKernelCapabilities(); const auto capability_init_result = diff --git a/src/core/hle/kernel/process.h b/src/core/hle/kernel/process.h index f060f2a3b..dda52f4c0 100644 --- a/src/core/hle/kernel/process.h +++ b/src/core/hle/kernel/process.h @@ -85,7 +85,7 @@ public: return name; } - static const HandleType HANDLE_TYPE = HandleType::Process; + static constexpr HandleType HANDLE_TYPE = HandleType::Process; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/kernel/readable_event.h b/src/core/hle/kernel/readable_event.h index 2eb9dcbb7..84215f572 100644 --- a/src/core/hle/kernel/readable_event.h +++ b/src/core/hle/kernel/readable_event.h @@ -31,7 +31,7 @@ public: return reset_type; } - static const HandleType HANDLE_TYPE = HandleType::ReadableEvent; + static constexpr HandleType HANDLE_TYPE = HandleType::ReadableEvent; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/kernel/resource_limit.h b/src/core/hle/kernel/resource_limit.h index 70e09858a..2613a6bb5 100644 --- a/src/core/hle/kernel/resource_limit.h +++ b/src/core/hle/kernel/resource_limit.h @@ -41,7 +41,7 @@ public: return GetTypeName(); } - static const HandleType HANDLE_TYPE = HandleType::ResourceLimit; + static constexpr HandleType HANDLE_TYPE = HandleType::ResourceLimit; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/kernel/scheduler.cpp b/src/core/hle/kernel/scheduler.cpp index ac501bf7f..e8447b69a 100644 --- a/src/core/hle/kernel/scheduler.cpp +++ b/src/core/hle/kernel/scheduler.cpp @@ -101,7 +101,6 @@ void Scheduler::SwitchContext(Thread* new_thread) { auto* const thread_owner_process = current_thread->GetOwnerProcess(); if (previous_process != thread_owner_process) { system.Kernel().MakeCurrentProcess(thread_owner_process); - Memory::SetCurrentPageTable(&thread_owner_process->VMManager().page_table); } cpu_core.LoadContext(new_thread->GetContext()); diff --git a/src/core/hle/kernel/server_port.cpp b/src/core/hle/kernel/server_port.cpp index 708fdf9e1..02e7c60e6 100644 --- a/src/core/hle/kernel/server_port.cpp +++ b/src/core/hle/kernel/server_port.cpp @@ -39,9 +39,8 @@ void ServerPort::Acquire(Thread* thread) { ASSERT_MSG(!ShouldWait(thread), "object unavailable!"); } -std::tuple<SharedPtr<ServerPort>, SharedPtr<ClientPort>> ServerPort::CreatePortPair( - KernelCore& kernel, u32 max_sessions, std::string name) { - +ServerPort::PortPair ServerPort::CreatePortPair(KernelCore& kernel, u32 max_sessions, + std::string name) { SharedPtr<ServerPort> server_port(new ServerPort(kernel)); SharedPtr<ClientPort> client_port(new ClientPort(kernel)); @@ -51,7 +50,7 @@ std::tuple<SharedPtr<ServerPort>, SharedPtr<ClientPort>> ServerPort::CreatePortP client_port->max_sessions = max_sessions; client_port->active_sessions = 0; - return std::make_tuple(std::move(server_port), std::move(client_port)); + return std::make_pair(std::move(server_port), std::move(client_port)); } } // namespace Kernel diff --git a/src/core/hle/kernel/server_port.h b/src/core/hle/kernel/server_port.h index 76293cb8b..dc88a1ebd 100644 --- a/src/core/hle/kernel/server_port.h +++ b/src/core/hle/kernel/server_port.h @@ -6,7 +6,7 @@ #include <memory> #include <string> -#include <tuple> +#include <utility> #include <vector> #include "common/common_types.h" #include "core/hle/kernel/object.h" @@ -23,6 +23,7 @@ class SessionRequestHandler; class ServerPort final : public WaitObject { public: using HLEHandler = std::shared_ptr<SessionRequestHandler>; + using PortPair = std::pair<SharedPtr<ServerPort>, SharedPtr<ClientPort>>; /** * Creates a pair of ServerPort and an associated ClientPort. @@ -32,8 +33,8 @@ public: * @param name Optional name of the ports * @return The created port tuple */ - static std::tuple<SharedPtr<ServerPort>, SharedPtr<ClientPort>> CreatePortPair( - KernelCore& kernel, u32 max_sessions, std::string name = "UnknownPort"); + static PortPair CreatePortPair(KernelCore& kernel, u32 max_sessions, + std::string name = "UnknownPort"); std::string GetTypeName() const override { return "ServerPort"; @@ -42,7 +43,7 @@ public: return name; } - static const HandleType HANDLE_TYPE = HandleType::ServerPort; + static constexpr HandleType HANDLE_TYPE = HandleType::ServerPort; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/kernel/server_session.cpp b/src/core/hle/kernel/server_session.cpp index 40cec143e..696a82cd9 100644 --- a/src/core/hle/kernel/server_session.cpp +++ b/src/core/hle/kernel/server_session.cpp @@ -28,11 +28,9 @@ ServerSession::~ServerSession() { // the emulated application. // Decrease the port's connection count. - if (parent->port) + if (parent->port) { parent->port->ConnectionClosed(); - - // TODO(Subv): Wake up all the ClientSession's waiting threads and set - // the SendSyncRequest result to 0xC920181A. + } parent->server = nullptr; } @@ -74,9 +72,6 @@ void ServerSession::ClientDisconnected() { handler->ClientDisconnected(this); } - // TODO(Subv): Force a wake up of all the ServerSession's waiting threads and set - // their WaitSynchronization result to 0xC920181A. - // Clean up the list of client threads with pending requests, they are unneeded now that the // client endpoint is closed. pending_requesting_threads.clear(); @@ -204,6 +199,6 @@ ServerSession::SessionPair ServerSession::CreateSessionPair(KernelCore& kernel, client_session->parent = parent; server_session->parent = parent; - return std::make_tuple(std::move(server_session), std::move(client_session)); + return std::make_pair(std::move(server_session), std::move(client_session)); } } // namespace Kernel diff --git a/src/core/hle/kernel/server_session.h b/src/core/hle/kernel/server_session.h index 79b84bade..738df30f8 100644 --- a/src/core/hle/kernel/server_session.h +++ b/src/core/hle/kernel/server_session.h @@ -6,6 +6,7 @@ #include <memory> #include <string> +#include <utility> #include <vector> #include "core/hle/kernel/object.h" @@ -41,7 +42,11 @@ public: return "ServerSession"; } - static const HandleType HANDLE_TYPE = HandleType::ServerSession; + std::string GetName() const override { + return name; + } + + static constexpr HandleType HANDLE_TYPE = HandleType::ServerSession; HandleType GetHandleType() const override { return HANDLE_TYPE; } @@ -54,7 +59,7 @@ public: return parent.get(); } - using SessionPair = std::tuple<SharedPtr<ServerSession>, SharedPtr<ClientSession>>; + using SessionPair = std::pair<SharedPtr<ServerSession>, SharedPtr<ClientSession>>; /** * Creates a pair of ServerSession and an associated ClientSession. diff --git a/src/core/hle/kernel/shared_memory.h b/src/core/hle/kernel/shared_memory.h index 37e18c443..c2b6155e1 100644 --- a/src/core/hle/kernel/shared_memory.h +++ b/src/core/hle/kernel/shared_memory.h @@ -76,7 +76,7 @@ public: return name; } - static const HandleType HANDLE_TYPE = HandleType::SharedMemory; + static constexpr HandleType HANDLE_TYPE = HandleType::SharedMemory; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp index ab10db3df..d48a2203a 100644 --- a/src/core/hle/kernel/svc.cpp +++ b/src/core/hle/kernel/svc.cpp @@ -131,16 +131,15 @@ enum class ResourceLimitValueType { LimitValue, }; -ResultVal<s64> RetrieveResourceLimitValue(Handle resource_limit, u32 resource_type, - ResourceLimitValueType value_type) { +ResultVal<s64> RetrieveResourceLimitValue(Core::System& system, Handle resource_limit, + u32 resource_type, ResourceLimitValueType value_type) { const auto type = static_cast<ResourceType>(resource_type); if (!IsValidResourceType(type)) { LOG_ERROR(Kernel_SVC, "Invalid resource limit type: '{}'", resource_type); return ERR_INVALID_ENUM_VALUE; } - const auto& kernel = Core::System::GetInstance().Kernel(); - const auto* const current_process = kernel.CurrentProcess(); + const auto* const current_process = system.Kernel().CurrentProcess(); ASSERT(current_process != nullptr); const auto resource_limit_object = @@ -160,7 +159,7 @@ ResultVal<s64> RetrieveResourceLimitValue(Handle resource_limit, u32 resource_ty } // Anonymous namespace /// Set the process heap to a given Size. It can both extend and shrink the heap. -static ResultCode SetHeapSize(VAddr* heap_addr, u64 heap_size) { +static ResultCode SetHeapSize(Core::System& system, VAddr* heap_addr, u64 heap_size) { LOG_TRACE(Kernel_SVC, "called, heap_size=0x{:X}", heap_size); // Size must be a multiple of 0x200000 (2MB) and be equal to or less than 8GB. @@ -175,7 +174,7 @@ static ResultCode SetHeapSize(VAddr* heap_addr, u64 heap_size) { return ERR_INVALID_SIZE; } - auto& vm_manager = Core::System::GetInstance().Kernel().CurrentProcess()->VMManager(); + auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); const auto alloc_result = vm_manager.SetHeapSize(heap_size); if (alloc_result.Failed()) { return alloc_result.Code(); @@ -185,7 +184,7 @@ static ResultCode SetHeapSize(VAddr* heap_addr, u64 heap_size) { return RESULT_SUCCESS; } -static ResultCode SetMemoryPermission(VAddr addr, u64 size, u32 prot) { +static ResultCode SetMemoryPermission(Core::System& system, VAddr addr, u64 size, u32 prot) { LOG_TRACE(Kernel_SVC, "called, addr=0x{:X}, size=0x{:X}, prot=0x{:X}", addr, size, prot); if (!Common::Is4KBAligned(addr)) { @@ -217,7 +216,7 @@ static ResultCode SetMemoryPermission(VAddr addr, u64 size, u32 prot) { return ERR_INVALID_MEMORY_PERMISSIONS; } - auto* const current_process = Core::CurrentProcess(); + auto* const current_process = system.Kernel().CurrentProcess(); auto& vm_manager = current_process->VMManager(); if (!vm_manager.IsWithinAddressSpace(addr, size)) { @@ -242,7 +241,8 @@ static ResultCode SetMemoryPermission(VAddr addr, u64 size, u32 prot) { return vm_manager.ReprotectRange(addr, size, converted_permissions); } -static ResultCode SetMemoryAttribute(VAddr address, u64 size, u32 mask, u32 attribute) { +static ResultCode SetMemoryAttribute(Core::System& system, VAddr address, u64 size, u32 mask, + u32 attribute) { LOG_DEBUG(Kernel_SVC, "called, address=0x{:016X}, size=0x{:X}, mask=0x{:08X}, attribute=0x{:08X}", address, size, mask, attribute); @@ -280,7 +280,7 @@ static ResultCode SetMemoryAttribute(VAddr address, u64 size, u32 mask, u32 attr return ERR_INVALID_COMBINATION; } - auto& vm_manager = Core::CurrentProcess()->VMManager(); + auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); if (!vm_manager.IsWithinAddressSpace(address, size)) { LOG_ERROR(Kernel_SVC, "Given address (0x{:016X}) is outside the bounds of the address space.", address); @@ -291,11 +291,11 @@ static ResultCode SetMemoryAttribute(VAddr address, u64 size, u32 mask, u32 attr } /// Maps a memory range into a different range. -static ResultCode MapMemory(VAddr dst_addr, VAddr src_addr, u64 size) { +static ResultCode MapMemory(Core::System& system, VAddr dst_addr, VAddr src_addr, u64 size) { LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr, src_addr, size); - auto& vm_manager = Core::CurrentProcess()->VMManager(); + auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size); if (result.IsError()) { @@ -306,11 +306,11 @@ static ResultCode MapMemory(VAddr dst_addr, VAddr src_addr, u64 size) { } /// Unmaps a region that was previously mapped with svcMapMemory -static ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size) { +static ResultCode UnmapMemory(Core::System& system, VAddr dst_addr, VAddr src_addr, u64 size) { LOG_TRACE(Kernel_SVC, "called, dst_addr=0x{:X}, src_addr=0x{:X}, size=0x{:X}", dst_addr, src_addr, size); - auto& vm_manager = Core::CurrentProcess()->VMManager(); + auto& vm_manager = system.Kernel().CurrentProcess()->VMManager(); const auto result = MapUnmapMemorySanityChecks(vm_manager, dst_addr, src_addr, size); if (result.IsError()) { @@ -321,7 +321,8 @@ static ResultCode UnmapMemory(VAddr dst_addr, VAddr src_addr, u64 size) { } /// Connect to an OS service given the port name, returns the handle to the port to out -static ResultCode ConnectToNamedPort(Handle* out_handle, VAddr port_name_address) { +static ResultCode ConnectToNamedPort(Core::System& system, Handle* out_handle, + VAddr port_name_address) { if (!Memory::IsValidVirtualAddress(port_name_address)) { LOG_ERROR(Kernel_SVC, "Port Name Address is not a valid virtual address, port_name_address=0x{:016X}", @@ -340,8 +341,8 @@ static ResultCode ConnectToNamedPort(Handle* out_handle, VAddr port_name_address LOG_TRACE(Kernel_SVC, "called port_name={}", port_name); - auto& kernel = Core::System::GetInstance().Kernel(); - auto it = kernel.FindNamedPort(port_name); + auto& kernel = system.Kernel(); + const auto it = kernel.FindNamedPort(port_name); if (!kernel.IsValidNamedPort(it)) { LOG_WARNING(Kernel_SVC, "tried to connect to unknown port: {}", port_name); return ERR_NOT_FOUND; @@ -353,14 +354,14 @@ static ResultCode ConnectToNamedPort(Handle* out_handle, VAddr port_name_address CASCADE_RESULT(client_session, client_port->Connect()); // Return the client session - auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + auto& handle_table = kernel.CurrentProcess()->GetHandleTable(); CASCADE_RESULT(*out_handle, handle_table.Create(client_session)); return RESULT_SUCCESS; } /// Makes a blocking IPC call to an OS service. -static ResultCode SendSyncRequest(Handle handle) { - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); +static ResultCode SendSyncRequest(Core::System& system, Handle handle) { + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); SharedPtr<ClientSession> session = handle_table.Get<ClientSession>(handle); if (!session) { LOG_ERROR(Kernel_SVC, "called with invalid handle=0x{:08X}", handle); @@ -369,18 +370,18 @@ static ResultCode SendSyncRequest(Handle handle) { LOG_TRACE(Kernel_SVC, "called handle=0x{:08X}({})", handle, session->GetName()); - Core::System::GetInstance().PrepareReschedule(); + system.PrepareReschedule(); // TODO(Subv): svcSendSyncRequest should put the caller thread to sleep while the server // responds and cause a reschedule. - return session->SendSyncRequest(GetCurrentThread()); + return session->SendSyncRequest(system.CurrentScheduler().GetCurrentThread()); } /// Get the ID for the specified thread. -static ResultCode GetThreadId(u64* thread_id, Handle thread_handle) { +static ResultCode GetThreadId(Core::System& system, u64* thread_id, Handle thread_handle) { LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", thread_handle); @@ -392,10 +393,10 @@ static ResultCode GetThreadId(u64* thread_id, Handle thread_handle) { } /// Gets the ID of the specified process or a specified thread's owning process. -static ResultCode GetProcessId(u64* process_id, Handle handle) { +static ResultCode GetProcessId(Core::System& system, u64* process_id, Handle handle) { LOG_DEBUG(Kernel_SVC, "called handle=0x{:08X}", handle); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const SharedPtr<Process> process = handle_table.Get<Process>(handle); if (process) { *process_id = process->GetProcessID(); @@ -437,8 +438,8 @@ static bool DefaultThreadWakeupCallback(ThreadWakeupReason reason, SharedPtr<Thr }; /// Wait for the given handles to synchronize, timeout after the specified nanoseconds -static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64 handle_count, - s64 nano_seconds) { +static ResultCode WaitSynchronization(Core::System& system, Handle* index, VAddr handles_address, + u64 handle_count, s64 nano_seconds) { LOG_TRACE(Kernel_SVC, "called handles_address=0x{:X}, handle_count={}, nano_seconds={}", handles_address, handle_count, nano_seconds); @@ -457,11 +458,11 @@ static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64 return ERR_OUT_OF_RANGE; } - auto* const thread = GetCurrentThread(); + auto* const thread = system.CurrentScheduler().GetCurrentThread(); using ObjectPtr = Thread::ThreadWaitObjects::value_type; Thread::ThreadWaitObjects objects(handle_count); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); for (u64 i = 0; i < handle_count; ++i) { const Handle handle = Memory::Read32(handles_address + i * sizeof(Handle)); @@ -507,16 +508,16 @@ static ResultCode WaitSynchronization(Handle* index, VAddr handles_address, u64 thread->WakeAfterDelay(nano_seconds); thread->SetWakeupCallback(DefaultThreadWakeupCallback); - Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule(); + system.CpuCore(thread->GetProcessorID()).PrepareReschedule(); return RESULT_TIMEOUT; } /// Resumes a thread waiting on WaitSynchronization -static ResultCode CancelSynchronization(Handle thread_handle) { +static ResultCode CancelSynchronization(Core::System& system, Handle thread_handle) { LOG_TRACE(Kernel_SVC, "called thread=0x{:X}", thread_handle); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}", @@ -531,8 +532,8 @@ static ResultCode CancelSynchronization(Handle thread_handle) { } /// Attempts to locks a mutex, creating it if it does not already exist -static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr, - Handle requesting_thread_handle) { +static ResultCode ArbitrateLock(Core::System& system, Handle holding_thread_handle, + VAddr mutex_addr, Handle requesting_thread_handle) { LOG_TRACE(Kernel_SVC, "called holding_thread_handle=0x{:08X}, mutex_addr=0x{:X}, " "requesting_current_thread_handle=0x{:08X}", @@ -549,13 +550,13 @@ static ResultCode ArbitrateLock(Handle holding_thread_handle, VAddr mutex_addr, return ERR_INVALID_ADDRESS; } - auto* const current_process = Core::System::GetInstance().Kernel().CurrentProcess(); + auto* const current_process = system.Kernel().CurrentProcess(); return current_process->GetMutex().TryAcquire(mutex_addr, holding_thread_handle, requesting_thread_handle); } /// Unlock a mutex -static ResultCode ArbitrateUnlock(VAddr mutex_addr) { +static ResultCode ArbitrateUnlock(Core::System& system, VAddr mutex_addr) { LOG_TRACE(Kernel_SVC, "called mutex_addr=0x{:X}", mutex_addr); if (Memory::IsKernelVirtualAddress(mutex_addr)) { @@ -569,7 +570,7 @@ static ResultCode ArbitrateUnlock(VAddr mutex_addr) { return ERR_INVALID_ADDRESS; } - auto* const current_process = Core::System::GetInstance().Kernel().CurrentProcess(); + auto* const current_process = system.Kernel().CurrentProcess(); return current_process->GetMutex().Release(mutex_addr); } @@ -592,7 +593,7 @@ struct BreakReason { }; /// Break program execution -static void Break(u32 reason, u64 info1, u64 info2) { +static void Break(Core::System& system, u32 reason, u64 info1, u64 info2) { BreakReason break_reason{reason}; bool has_dumped_buffer{}; @@ -670,22 +671,24 @@ static void Break(u32 reason, u64 info1, u64 info2) { Debug_Emulated, "Emulated program broke execution! reason=0x{:016X}, info1=0x{:016X}, info2=0x{:016X}", reason, info1, info2); + handle_debug_buffer(info1, info2); - Core::System::GetInstance() - .ArmInterface(static_cast<std::size_t>(GetCurrentThread()->GetProcessorID())) - .LogBacktrace(); + + auto* const current_thread = system.CurrentScheduler().GetCurrentThread(); + const auto thread_processor_id = current_thread->GetProcessorID(); + system.ArmInterface(static_cast<std::size_t>(thread_processor_id)).LogBacktrace(); ASSERT(false); - Core::CurrentProcess()->PrepareForTermination(); + system.Kernel().CurrentProcess()->PrepareForTermination(); // Kill the current thread - GetCurrentThread()->Stop(); - Core::System::GetInstance().PrepareReschedule(); + current_thread->Stop(); + system.PrepareReschedule(); } } /// Used to output a message on a debug hardware unit - does nothing on a retail unit -static void OutputDebugString(VAddr address, u64 len) { +static void OutputDebugString([[maybe_unused]] Core::System& system, VAddr address, u64 len) { if (len == 0) { return; } @@ -696,7 +699,8 @@ static void OutputDebugString(VAddr address, u64 len) { } /// Gets system/memory information for the current process -static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) { +static ResultCode GetInfo(Core::System& system, u64* result, u64 info_id, u64 handle, + u64 info_sub_id) { LOG_TRACE(Kernel_SVC, "called info_id=0x{:X}, info_sub_id=0x{:X}, handle=0x{:08X}", info_id, info_sub_id, handle); @@ -754,7 +758,8 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) return ERR_INVALID_ENUM_VALUE; } - const auto& current_process_handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& current_process_handle_table = + system.Kernel().CurrentProcess()->GetHandleTable(); const auto process = current_process_handle_table.Get<Process>(static_cast<Handle>(handle)); if (!process) { return ERR_INVALID_HANDLE; @@ -844,7 +849,7 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) return ERR_INVALID_COMBINATION; } - Process* const current_process = Core::CurrentProcess(); + Process* const current_process = system.Kernel().CurrentProcess(); HandleTable& handle_table = current_process->GetHandleTable(); const auto resource_limit = current_process->GetResourceLimit(); if (!resource_limit) { @@ -875,7 +880,7 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) return ERR_INVALID_COMBINATION; } - *result = Core::CurrentProcess()->GetRandomEntropy(info_sub_id); + *result = system.Kernel().CurrentProcess()->GetRandomEntropy(info_sub_id); return RESULT_SUCCESS; case GetInfoType::PrivilegedProcessId: @@ -892,15 +897,14 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) return ERR_INVALID_COMBINATION; } - const auto thread = - Core::CurrentProcess()->GetHandleTable().Get<Thread>(static_cast<Handle>(handle)); + const auto thread = system.Kernel().CurrentProcess()->GetHandleTable().Get<Thread>( + static_cast<Handle>(handle)); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", static_cast<Handle>(handle)); return ERR_INVALID_HANDLE; } - const auto& system = Core::System::GetInstance(); const auto& core_timing = system.CoreTiming(); const auto& scheduler = system.CurrentScheduler(); const auto* const current_thread = scheduler.GetCurrentThread(); @@ -927,13 +931,13 @@ static ResultCode GetInfo(u64* result, u64 info_id, u64 handle, u64 info_sub_id) } /// Sets the thread activity -static ResultCode SetThreadActivity(Handle handle, u32 activity) { +static ResultCode SetThreadActivity(Core::System& system, Handle handle, u32 activity) { LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, activity=0x{:08X}", handle, activity); if (activity > static_cast<u32>(ThreadActivity::Paused)) { return ERR_INVALID_ENUM_VALUE; } - const auto* current_process = Core::CurrentProcess(); + const auto* current_process = system.Kernel().CurrentProcess(); const SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle); @@ -950,7 +954,7 @@ static ResultCode SetThreadActivity(Handle handle, u32 activity) { return ERR_INVALID_HANDLE; } - if (thread == GetCurrentThread()) { + if (thread == system.CurrentScheduler().GetCurrentThread()) { LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread"); return ERR_BUSY; } @@ -960,10 +964,10 @@ static ResultCode SetThreadActivity(Handle handle, u32 activity) { } /// Gets the thread context -static ResultCode GetThreadContext(VAddr thread_context, Handle handle) { +static ResultCode GetThreadContext(Core::System& system, VAddr thread_context, Handle handle) { LOG_DEBUG(Kernel_SVC, "called, context=0x{:08X}, thread=0x{:X}", thread_context, handle); - const auto* current_process = Core::CurrentProcess(); + const auto* current_process = system.Kernel().CurrentProcess(); const SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle); @@ -980,7 +984,7 @@ static ResultCode GetThreadContext(VAddr thread_context, Handle handle) { return ERR_INVALID_HANDLE; } - if (thread == GetCurrentThread()) { + if (thread == system.CurrentScheduler().GetCurrentThread()) { LOG_ERROR(Kernel_SVC, "The thread handle specified is the current running thread"); return ERR_BUSY; } @@ -1001,10 +1005,10 @@ static ResultCode GetThreadContext(VAddr thread_context, Handle handle) { } /// Gets the priority for the specified thread -static ResultCode GetThreadPriority(u32* priority, Handle handle) { +static ResultCode GetThreadPriority(Core::System& system, u32* priority, Handle handle) { LOG_TRACE(Kernel_SVC, "called"); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const SharedPtr<Thread> thread = handle_table.Get<Thread>(handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, handle=0x{:08X}", handle); @@ -1016,7 +1020,7 @@ static ResultCode GetThreadPriority(u32* priority, Handle handle) { } /// Sets the priority for the specified thread -static ResultCode SetThreadPriority(Handle handle, u32 priority) { +static ResultCode SetThreadPriority(Core::System& system, Handle handle, u32 priority) { LOG_TRACE(Kernel_SVC, "called"); if (priority > THREADPRIO_LOWEST) { @@ -1027,7 +1031,7 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) { return ERR_INVALID_THREAD_PRIORITY; } - const auto* const current_process = Core::CurrentProcess(); + const auto* const current_process = system.Kernel().CurrentProcess(); SharedPtr<Thread> thread = current_process->GetHandleTable().Get<Thread>(handle); if (!thread) { @@ -1037,18 +1041,18 @@ static ResultCode SetThreadPriority(Handle handle, u32 priority) { thread->SetPriority(priority); - Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule(); + system.CpuCore(thread->GetProcessorID()).PrepareReschedule(); return RESULT_SUCCESS; } /// Get which CPU core is executing the current thread -static u32 GetCurrentProcessorNumber() { +static u32 GetCurrentProcessorNumber(Core::System& system) { LOG_TRACE(Kernel_SVC, "called"); - return GetCurrentThread()->GetProcessorID(); + return system.CurrentScheduler().GetCurrentThread()->GetProcessorID(); } -static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 size, - u32 permissions) { +static ResultCode MapSharedMemory(Core::System& system, Handle shared_memory_handle, VAddr addr, + u64 size, u32 permissions) { LOG_TRACE(Kernel_SVC, "called, shared_memory_handle=0x{:X}, addr=0x{:X}, size=0x{:X}, permissions=0x{:08X}", shared_memory_handle, addr, size, permissions); @@ -1082,7 +1086,7 @@ static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 s return ERR_INVALID_MEMORY_PERMISSIONS; } - auto* const current_process = Core::CurrentProcess(); + auto* const current_process = system.Kernel().CurrentProcess(); auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle); if (!shared_memory) { LOG_ERROR(Kernel_SVC, "Shared memory does not exist, shared_memory_handle=0x{:08X}", @@ -1100,7 +1104,8 @@ static ResultCode MapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 s return shared_memory->Map(*current_process, addr, permissions_type, MemoryPermission::DontCare); } -static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 size) { +static ResultCode UnmapSharedMemory(Core::System& system, Handle shared_memory_handle, VAddr addr, + u64 size) { LOG_WARNING(Kernel_SVC, "called, shared_memory_handle=0x{:08X}, addr=0x{:X}, size=0x{:X}", shared_memory_handle, addr, size); @@ -1125,7 +1130,7 @@ static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 return ERR_INVALID_ADDRESS_STATE; } - auto* const current_process = Core::CurrentProcess(); + auto* const current_process = system.Kernel().CurrentProcess(); auto shared_memory = current_process->GetHandleTable().Get<SharedMemory>(shared_memory_handle); if (!shared_memory) { LOG_ERROR(Kernel_SVC, "Shared memory does not exist, shared_memory_handle=0x{:08X}", @@ -1143,10 +1148,11 @@ static ResultCode UnmapSharedMemory(Handle shared_memory_handle, VAddr addr, u64 return shared_memory->Unmap(*current_process, addr, size); } -static ResultCode QueryProcessMemory(VAddr memory_info_address, VAddr page_info_address, - Handle process_handle, VAddr address) { +static ResultCode QueryProcessMemory(Core::System& system, VAddr memory_info_address, + VAddr page_info_address, Handle process_handle, + VAddr address) { LOG_TRACE(Kernel_SVC, "called process=0x{:08X} address={:X}", process_handle, address); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); SharedPtr<Process> process = handle_table.Get<Process>(process_handle); if (!process) { LOG_ERROR(Kernel_SVC, "Process handle does not exist, process_handle=0x{:08X}", @@ -1172,20 +1178,156 @@ static ResultCode QueryProcessMemory(VAddr memory_info_address, VAddr page_info_ return RESULT_SUCCESS; } -static ResultCode QueryMemory(VAddr memory_info_address, VAddr page_info_address, - VAddr query_address) { +static ResultCode QueryMemory(Core::System& system, VAddr memory_info_address, + VAddr page_info_address, VAddr query_address) { LOG_TRACE(Kernel_SVC, "called, memory_info_address=0x{:016X}, page_info_address=0x{:016X}, " "query_address=0x{:016X}", memory_info_address, page_info_address, query_address); - return QueryProcessMemory(memory_info_address, page_info_address, CurrentProcess, + return QueryProcessMemory(system, memory_info_address, page_info_address, CurrentProcess, query_address); } +static ResultCode MapProcessCodeMemory(Core::System& system, Handle process_handle, u64 dst_address, + u64 src_address, u64 size) { + LOG_DEBUG(Kernel_SVC, + "called. process_handle=0x{:08X}, dst_address=0x{:016X}, " + "src_address=0x{:016X}, size=0x{:016X}", + process_handle, dst_address, src_address, size); + + if (!Common::Is4KBAligned(src_address)) { + LOG_ERROR(Kernel_SVC, "src_address is not page-aligned (src_address=0x{:016X}).", + src_address); + return ERR_INVALID_ADDRESS; + } + + if (!Common::Is4KBAligned(dst_address)) { + LOG_ERROR(Kernel_SVC, "dst_address is not page-aligned (dst_address=0x{:016X}).", + dst_address); + return ERR_INVALID_ADDRESS; + } + + if (size == 0 || !Common::Is4KBAligned(size)) { + LOG_ERROR(Kernel_SVC, "Size is zero or not page-aligned (size=0x{:016X})", size); + return ERR_INVALID_SIZE; + } + + if (!IsValidAddressRange(dst_address, size)) { + LOG_ERROR(Kernel_SVC, + "Destination address range overflows the address space (dst_address=0x{:016X}, " + "size=0x{:016X}).", + dst_address, size); + return ERR_INVALID_ADDRESS_STATE; + } + + if (!IsValidAddressRange(src_address, size)) { + LOG_ERROR(Kernel_SVC, + "Source address range overflows the address space (src_address=0x{:016X}, " + "size=0x{:016X}).", + src_address, size); + return ERR_INVALID_ADDRESS_STATE; + } + + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); + auto process = handle_table.Get<Process>(process_handle); + if (!process) { + LOG_ERROR(Kernel_SVC, "Invalid process handle specified (handle=0x{:08X}).", + process_handle); + return ERR_INVALID_HANDLE; + } + + auto& vm_manager = process->VMManager(); + if (!vm_manager.IsWithinAddressSpace(src_address, size)) { + LOG_ERROR(Kernel_SVC, + "Source address range is not within the address space (src_address=0x{:016X}, " + "size=0x{:016X}).", + src_address, size); + return ERR_INVALID_ADDRESS_STATE; + } + + if (!vm_manager.IsWithinASLRRegion(dst_address, size)) { + LOG_ERROR(Kernel_SVC, + "Destination address range is not within the ASLR region (dst_address=0x{:016X}, " + "size=0x{:016X}).", + dst_address, size); + return ERR_INVALID_MEMORY_RANGE; + } + + return vm_manager.MapCodeMemory(dst_address, src_address, size); +} + +ResultCode UnmapProcessCodeMemory(Core::System& system, Handle process_handle, u64 dst_address, + u64 src_address, u64 size) { + LOG_DEBUG(Kernel_SVC, + "called. process_handle=0x{:08X}, dst_address=0x{:016X}, src_address=0x{:016X}, " + "size=0x{:016X}", + process_handle, dst_address, src_address, size); + + if (!Common::Is4KBAligned(dst_address)) { + LOG_ERROR(Kernel_SVC, "dst_address is not page-aligned (dst_address=0x{:016X}).", + dst_address); + return ERR_INVALID_ADDRESS; + } + + if (!Common::Is4KBAligned(src_address)) { + LOG_ERROR(Kernel_SVC, "src_address is not page-aligned (src_address=0x{:016X}).", + src_address); + return ERR_INVALID_ADDRESS; + } + + if (size == 0 || Common::Is4KBAligned(size)) { + LOG_ERROR(Kernel_SVC, "Size is zero or not page-aligned (size=0x{:016X}).", size); + return ERR_INVALID_SIZE; + } + + if (!IsValidAddressRange(dst_address, size)) { + LOG_ERROR(Kernel_SVC, + "Destination address range overflows the address space (dst_address=0x{:016X}, " + "size=0x{:016X}).", + dst_address, size); + return ERR_INVALID_ADDRESS_STATE; + } + + if (!IsValidAddressRange(src_address, size)) { + LOG_ERROR(Kernel_SVC, + "Source address range overflows the address space (src_address=0x{:016X}, " + "size=0x{:016X}).", + src_address, size); + return ERR_INVALID_ADDRESS_STATE; + } + + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); + auto process = handle_table.Get<Process>(process_handle); + if (!process) { + LOG_ERROR(Kernel_SVC, "Invalid process handle specified (handle=0x{:08X}).", + process_handle); + return ERR_INVALID_HANDLE; + } + + auto& vm_manager = process->VMManager(); + if (!vm_manager.IsWithinAddressSpace(src_address, size)) { + LOG_ERROR(Kernel_SVC, + "Source address range is not within the address space (src_address=0x{:016X}, " + "size=0x{:016X}).", + src_address, size); + return ERR_INVALID_ADDRESS_STATE; + } + + if (!vm_manager.IsWithinASLRRegion(dst_address, size)) { + LOG_ERROR(Kernel_SVC, + "Destination address range is not within the ASLR region (dst_address=0x{:016X}, " + "size=0x{:016X}).", + dst_address, size); + return ERR_INVALID_MEMORY_RANGE; + } + + return vm_manager.UnmapCodeMemory(dst_address, src_address, size); +} + /// Exits the current process -static void ExitProcess() { - auto* current_process = Core::CurrentProcess(); +static void ExitProcess(Core::System& system) { + auto* current_process = system.Kernel().CurrentProcess(); LOG_INFO(Kernel_SVC, "Process {} exiting", current_process->GetProcessID()); ASSERT_MSG(current_process->GetStatus() == ProcessStatus::Running, @@ -1194,20 +1336,20 @@ static void ExitProcess() { current_process->PrepareForTermination(); // Kill the current thread - GetCurrentThread()->Stop(); + system.CurrentScheduler().GetCurrentThread()->Stop(); - Core::System::GetInstance().PrepareReschedule(); + system.PrepareReschedule(); } /// Creates a new thread -static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, VAddr stack_top, - u32 priority, s32 processor_id) { +static ResultCode CreateThread(Core::System& system, Handle* out_handle, VAddr entry_point, u64 arg, + VAddr stack_top, u32 priority, s32 processor_id) { LOG_TRACE(Kernel_SVC, "called entrypoint=0x{:08X}, arg=0x{:08X}, stacktop=0x{:08X}, " "threadpriority=0x{:08X}, processorid=0x{:08X} : created handle=0x{:08X}", entry_point, arg, stack_top, priority, processor_id, *out_handle); - auto* const current_process = Core::CurrentProcess(); + auto* const current_process = system.Kernel().CurrentProcess(); if (processor_id == THREADPROCESSORID_IDEAL) { // Set the target CPU to the one specified by the process. @@ -1239,7 +1381,7 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V } const std::string name = fmt::format("thread-{:X}", entry_point); - auto& kernel = Core::System::GetInstance().Kernel(); + auto& kernel = system.Kernel(); CASCADE_RESULT(SharedPtr<Thread> thread, Thread::Create(kernel, name, entry_point, priority, arg, processor_id, stack_top, *current_process)); @@ -1253,16 +1395,16 @@ static ResultCode CreateThread(Handle* out_handle, VAddr entry_point, u64 arg, V thread->SetGuestHandle(*new_guest_handle); *out_handle = *new_guest_handle; - Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule(); + system.CpuCore(thread->GetProcessorID()).PrepareReschedule(); return RESULT_SUCCESS; } /// Starts the thread for the provided handle -static ResultCode StartThread(Handle thread_handle) { +static ResultCode StartThread(Core::System& system, Handle thread_handle) { LOG_TRACE(Kernel_SVC, "called thread=0x{:08X}", thread_handle); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}", @@ -1275,16 +1417,14 @@ static ResultCode StartThread(Handle thread_handle) { thread->ResumeFromWait(); if (thread->GetStatus() == ThreadStatus::Ready) { - Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule(); + system.CpuCore(thread->GetProcessorID()).PrepareReschedule(); } return RESULT_SUCCESS; } /// Called when a thread exits -static void ExitThread() { - auto& system = Core::System::GetInstance(); - +static void ExitThread(Core::System& system) { LOG_TRACE(Kernel_SVC, "called, pc=0x{:08X}", system.CurrentArmInterface().GetPC()); auto* const current_thread = system.CurrentScheduler().GetCurrentThread(); @@ -1294,7 +1434,7 @@ static void ExitThread() { } /// Sleep the current thread -static void SleepThread(s64 nanoseconds) { +static void SleepThread(Core::System& system, s64 nanoseconds) { LOG_TRACE(Kernel_SVC, "called nanoseconds={}", nanoseconds); enum class SleepType : s64 { @@ -1303,7 +1443,6 @@ static void SleepThread(s64 nanoseconds) { YieldAndWaitForLoadBalancing = -2, }; - auto& system = Core::System::GetInstance(); auto& scheduler = system.CurrentScheduler(); auto* const current_thread = scheduler.GetCurrentThread(); @@ -1332,14 +1471,29 @@ static void SleepThread(s64 nanoseconds) { } /// Wait process wide key atomic -static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_variable_addr, - Handle thread_handle, s64 nano_seconds) { +static ResultCode WaitProcessWideKeyAtomic(Core::System& system, VAddr mutex_addr, + VAddr condition_variable_addr, Handle thread_handle, + s64 nano_seconds) { LOG_TRACE( Kernel_SVC, "called mutex_addr={:X}, condition_variable_addr={:X}, thread_handle=0x{:08X}, timeout={}", mutex_addr, condition_variable_addr, thread_handle, nano_seconds); - auto* const current_process = Core::System::GetInstance().Kernel().CurrentProcess(); + if (Memory::IsKernelVirtualAddress(mutex_addr)) { + LOG_ERROR( + Kernel_SVC, + "Given mutex address must not be within the kernel address space. address=0x{:016X}", + mutex_addr); + return ERR_INVALID_ADDRESS_STATE; + } + + if (!Common::IsWordAligned(mutex_addr)) { + LOG_ERROR(Kernel_SVC, "Given mutex address must be word-aligned. address=0x{:016X}", + mutex_addr); + return ERR_INVALID_ADDRESS; + } + + auto* const current_process = system.Kernel().CurrentProcess(); const auto& handle_table = current_process->GetHandleTable(); SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle); ASSERT(thread); @@ -1349,7 +1503,7 @@ static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_var return release_result; } - SharedPtr<Thread> current_thread = GetCurrentThread(); + SharedPtr<Thread> current_thread = system.CurrentScheduler().GetCurrentThread(); current_thread->SetCondVarWaitAddress(condition_variable_addr); current_thread->SetMutexWaitAddress(mutex_addr); current_thread->SetWaitHandle(thread_handle); @@ -1360,19 +1514,20 @@ static ResultCode WaitProcessWideKeyAtomic(VAddr mutex_addr, VAddr condition_var // Note: Deliberately don't attempt to inherit the lock owner's priority. - Core::System::GetInstance().CpuCore(current_thread->GetProcessorID()).PrepareReschedule(); + system.CpuCore(current_thread->GetProcessorID()).PrepareReschedule(); return RESULT_SUCCESS; } /// Signal process wide key -static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target) { +static ResultCode SignalProcessWideKey(Core::System& system, VAddr condition_variable_addr, + s32 target) { LOG_TRACE(Kernel_SVC, "called, condition_variable_addr=0x{:X}, target=0x{:08X}", condition_variable_addr, target); - const auto RetrieveWaitingThreads = [](std::size_t core_index, - std::vector<SharedPtr<Thread>>& waiting_threads, - VAddr condvar_addr) { - const auto& scheduler = Core::System::GetInstance().Scheduler(core_index); + const auto RetrieveWaitingThreads = [&system](std::size_t core_index, + std::vector<SharedPtr<Thread>>& waiting_threads, + VAddr condvar_addr) { + const auto& scheduler = system.Scheduler(core_index); const auto& thread_list = scheduler.GetThreadList(); for (const auto& thread : thread_list) { @@ -1411,9 +1566,8 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target // liberate Cond Var Thread. thread->SetCondVarWaitAddress(0); - std::size_t current_core = Core::System::GetInstance().CurrentCoreIndex(); - - auto& monitor = Core::System::GetInstance().Monitor(); + const std::size_t current_core = system.CurrentCoreIndex(); + auto& monitor = system.Monitor(); // Atomically read the value of the mutex. u32 mutex_val = 0; @@ -1442,7 +1596,7 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target thread->SetLockOwner(nullptr); thread->SetMutexWaitAddress(0); thread->SetWaitHandle(0); - Core::System::GetInstance().CpuCore(thread->GetProcessorID()).PrepareReschedule(); + system.CpuCore(thread->GetProcessorID()).PrepareReschedule(); } else { // Atomically signal that the mutex now has a waiting thread. do { @@ -1458,7 +1612,7 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target // The mutex is already owned by some other thread, make this thread wait on it. const Handle owner_handle = static_cast<Handle>(mutex_val & Mutex::MutexOwnerMask); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); auto owner = handle_table.Get<Thread>(owner_handle); ASSERT(owner); ASSERT(thread->GetStatus() == ThreadStatus::WaitCondVar); @@ -1473,14 +1627,17 @@ static ResultCode SignalProcessWideKey(VAddr condition_variable_addr, s32 target } // Wait for an address (via Address Arbiter) -static ResultCode WaitForAddress(VAddr address, u32 type, s32 value, s64 timeout) { +static ResultCode WaitForAddress(Core::System& system, VAddr address, u32 type, s32 value, + s64 timeout) { LOG_WARNING(Kernel_SVC, "called, address=0x{:X}, type=0x{:X}, value=0x{:X}, timeout={}", address, type, value, timeout); + // If the passed address is a kernel virtual address, return invalid memory state. if (Memory::IsKernelVirtualAddress(address)) { LOG_ERROR(Kernel_SVC, "Address is a kernel virtual address, address={:016X}", address); return ERR_INVALID_ADDRESS_STATE; } + // If the address is not properly aligned to 4 bytes, return invalid address. if (!Common::IsWordAligned(address)) { LOG_ERROR(Kernel_SVC, "Address is not word aligned, address={:016X}", address); @@ -1488,20 +1645,22 @@ static ResultCode WaitForAddress(VAddr address, u32 type, s32 value, s64 timeout } const auto arbitration_type = static_cast<AddressArbiter::ArbitrationType>(type); - auto& address_arbiter = - Core::System::GetInstance().Kernel().CurrentProcess()->GetAddressArbiter(); + auto& address_arbiter = system.Kernel().CurrentProcess()->GetAddressArbiter(); return address_arbiter.WaitForAddress(address, arbitration_type, value, timeout); } // Signals to an address (via Address Arbiter) -static ResultCode SignalToAddress(VAddr address, u32 type, s32 value, s32 num_to_wake) { +static ResultCode SignalToAddress(Core::System& system, VAddr address, u32 type, s32 value, + s32 num_to_wake) { LOG_WARNING(Kernel_SVC, "called, address=0x{:X}, type=0x{:X}, value=0x{:X}, num_to_wake=0x{:X}", address, type, value, num_to_wake); + // If the passed address is a kernel virtual address, return invalid memory state. if (Memory::IsKernelVirtualAddress(address)) { LOG_ERROR(Kernel_SVC, "Address is a kernel virtual address, address={:016X}", address); return ERR_INVALID_ADDRESS_STATE; } + // If the address is not properly aligned to 4 bytes, return invalid address. if (!Common::IsWordAligned(address)) { LOG_ERROR(Kernel_SVC, "Address is not word aligned, address={:016X}", address); @@ -1509,16 +1668,15 @@ static ResultCode SignalToAddress(VAddr address, u32 type, s32 value, s32 num_to } const auto signal_type = static_cast<AddressArbiter::SignalType>(type); - auto& address_arbiter = - Core::System::GetInstance().Kernel().CurrentProcess()->GetAddressArbiter(); + auto& address_arbiter = system.Kernel().CurrentProcess()->GetAddressArbiter(); return address_arbiter.SignalToAddress(address, signal_type, value, num_to_wake); } /// This returns the total CPU ticks elapsed since the CPU was powered-on -static u64 GetSystemTick() { +static u64 GetSystemTick(Core::System& system) { LOG_TRACE(Kernel_SVC, "called"); - auto& core_timing = Core::System::GetInstance().CoreTiming(); + auto& core_timing = system.CoreTiming(); const u64 result{core_timing.GetTicks()}; // Advance time to defeat dumb games that busy-wait for the frame to end. @@ -1528,18 +1686,18 @@ static u64 GetSystemTick() { } /// Close a handle -static ResultCode CloseHandle(Handle handle) { +static ResultCode CloseHandle(Core::System& system, Handle handle) { LOG_TRACE(Kernel_SVC, "Closing handle 0x{:08X}", handle); - auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); return handle_table.Close(handle); } /// Clears the signaled state of an event or process. -static ResultCode ResetSignal(Handle handle) { +static ResultCode ResetSignal(Core::System& system, Handle handle) { LOG_DEBUG(Kernel_SVC, "called handle 0x{:08X}", handle); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); auto event = handle_table.Get<ReadableEvent>(handle); if (event) { @@ -1556,7 +1714,8 @@ static ResultCode ResetSignal(Handle handle) { } /// Creates a TransferMemory object -static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32 permissions) { +static ResultCode CreateTransferMemory(Core::System& system, Handle* handle, VAddr addr, u64 size, + u32 permissions) { LOG_DEBUG(Kernel_SVC, "called addr=0x{:X}, size=0x{:X}, perms=0x{:08X}", addr, size, permissions); @@ -1584,7 +1743,7 @@ static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32 return ERR_INVALID_MEMORY_PERMISSIONS; } - auto& kernel = Core::System::GetInstance().Kernel(); + auto& kernel = system.Kernel(); auto transfer_mem_handle = TransferMemory::Create(kernel, addr, size, perms); auto& handle_table = kernel.CurrentProcess()->GetHandleTable(); @@ -1597,7 +1756,8 @@ static ResultCode CreateTransferMemory(Handle* handle, VAddr addr, u64 size, u32 return RESULT_SUCCESS; } -static ResultCode MapTransferMemory(Handle handle, VAddr address, u64 size, u32 permission_raw) { +static ResultCode MapTransferMemory(Core::System& system, Handle handle, VAddr address, u64 size, + u32 permission_raw) { LOG_DEBUG(Kernel_SVC, "called. handle=0x{:08X}, address=0x{:016X}, size=0x{:016X}, permissions=0x{:08X}", handle, address, size, permission_raw); @@ -1631,7 +1791,7 @@ static ResultCode MapTransferMemory(Handle handle, VAddr address, u64 size, u32 return ERR_INVALID_STATE; } - const auto& kernel = Core::System::GetInstance().Kernel(); + const auto& kernel = system.Kernel(); const auto* const current_process = kernel.CurrentProcess(); const auto& handle_table = current_process->GetHandleTable(); @@ -1653,7 +1813,8 @@ static ResultCode MapTransferMemory(Handle handle, VAddr address, u64 size, u32 return transfer_memory->MapMemory(address, size, permissions); } -static ResultCode UnmapTransferMemory(Handle handle, VAddr address, u64 size) { +static ResultCode UnmapTransferMemory(Core::System& system, Handle handle, VAddr address, + u64 size) { LOG_DEBUG(Kernel_SVC, "called. handle=0x{:08X}, address=0x{:016X}, size=0x{:016X}", handle, address, size); @@ -1678,7 +1839,7 @@ static ResultCode UnmapTransferMemory(Handle handle, VAddr address, u64 size) { return ERR_INVALID_ADDRESS_STATE; } - const auto& kernel = Core::System::GetInstance().Kernel(); + const auto& kernel = system.Kernel(); const auto* const current_process = kernel.CurrentProcess(); const auto& handle_table = current_process->GetHandleTable(); @@ -1700,10 +1861,11 @@ static ResultCode UnmapTransferMemory(Handle handle, VAddr address, u64 size) { return transfer_memory->UnmapMemory(address, size); } -static ResultCode GetThreadCoreMask(Handle thread_handle, u32* core, u64* mask) { +static ResultCode GetThreadCoreMask(Core::System& system, Handle thread_handle, u32* core, + u64* mask) { LOG_TRACE(Kernel_SVC, "called, handle=0x{:08X}", thread_handle); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}", @@ -1717,11 +1879,12 @@ static ResultCode GetThreadCoreMask(Handle thread_handle, u32* core, u64* mask) return RESULT_SUCCESS; } -static ResultCode SetThreadCoreMask(Handle thread_handle, u32 core, u64 mask) { +static ResultCode SetThreadCoreMask(Core::System& system, Handle thread_handle, u32 core, + u64 mask) { LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, mask=0x{:016X}, core=0x{:X}", thread_handle, mask, core); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const SharedPtr<Thread> thread = handle_table.Get<Thread>(thread_handle); if (!thread) { LOG_ERROR(Kernel_SVC, "Thread handle does not exist, thread_handle=0x{:08X}", @@ -1766,8 +1929,8 @@ static ResultCode SetThreadCoreMask(Handle thread_handle, u32 core, u64 mask) { return RESULT_SUCCESS; } -static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permissions, - u32 remote_permissions) { +static ResultCode CreateSharedMemory(Core::System& system, Handle* handle, u64 size, + u32 local_permissions, u32 remote_permissions) { LOG_TRACE(Kernel_SVC, "called, size=0x{:X}, localPerms=0x{:08X}, remotePerms=0x{:08X}", size, local_permissions, remote_permissions); if (size == 0) { @@ -1803,7 +1966,7 @@ static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permiss return ERR_INVALID_MEMORY_PERMISSIONS; } - auto& kernel = Core::System::GetInstance().Kernel(); + auto& kernel = system.Kernel(); auto process = kernel.CurrentProcess(); auto& handle_table = process->GetHandleTable(); auto shared_mem_handle = SharedMemory::Create(kernel, process, size, local_perms, remote_perms); @@ -1812,10 +1975,10 @@ static ResultCode CreateSharedMemory(Handle* handle, u64 size, u32 local_permiss return RESULT_SUCCESS; } -static ResultCode CreateEvent(Handle* write_handle, Handle* read_handle) { +static ResultCode CreateEvent(Core::System& system, Handle* write_handle, Handle* read_handle) { LOG_DEBUG(Kernel_SVC, "called"); - auto& kernel = Core::System::GetInstance().Kernel(); + auto& kernel = system.Kernel(); const auto [readable_event, writable_event] = WritableEvent::CreateEventPair(kernel, ResetType::Sticky, "CreateEvent"); @@ -1840,10 +2003,10 @@ static ResultCode CreateEvent(Handle* write_handle, Handle* read_handle) { return RESULT_SUCCESS; } -static ResultCode ClearEvent(Handle handle) { +static ResultCode ClearEvent(Core::System& system, Handle handle) { LOG_TRACE(Kernel_SVC, "called, event=0x{:08X}", handle); - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); auto writable_event = handle_table.Get<WritableEvent>(handle); if (writable_event) { @@ -1861,10 +2024,10 @@ static ResultCode ClearEvent(Handle handle) { return ERR_INVALID_HANDLE; } -static ResultCode SignalEvent(Handle handle) { +static ResultCode SignalEvent(Core::System& system, Handle handle) { LOG_DEBUG(Kernel_SVC, "called. Handle=0x{:08X}", handle); - HandleTable& handle_table = Core::CurrentProcess()->GetHandleTable(); + HandleTable& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); auto writable_event = handle_table.Get<WritableEvent>(handle); if (!writable_event) { @@ -1876,7 +2039,7 @@ static ResultCode SignalEvent(Handle handle) { return RESULT_SUCCESS; } -static ResultCode GetProcessInfo(u64* out, Handle process_handle, u32 type) { +static ResultCode GetProcessInfo(Core::System& system, u64* out, Handle process_handle, u32 type) { LOG_DEBUG(Kernel_SVC, "called, handle=0x{:08X}, type=0x{:X}", process_handle, type); // This function currently only allows retrieving a process' status. @@ -1884,7 +2047,7 @@ static ResultCode GetProcessInfo(u64* out, Handle process_handle, u32 type) { Status, }; - const auto& handle_table = Core::CurrentProcess()->GetHandleTable(); + const auto& handle_table = system.Kernel().CurrentProcess()->GetHandleTable(); const auto process = handle_table.Get<Process>(process_handle); if (!process) { LOG_ERROR(Kernel_SVC, "Process handle does not exist, process_handle=0x{:08X}", @@ -1902,10 +2065,10 @@ static ResultCode GetProcessInfo(u64* out, Handle process_handle, u32 type) { return RESULT_SUCCESS; } -static ResultCode CreateResourceLimit(Handle* out_handle) { +static ResultCode CreateResourceLimit(Core::System& system, Handle* out_handle) { LOG_DEBUG(Kernel_SVC, "called"); - auto& kernel = Core::System::GetInstance().Kernel(); + auto& kernel = system.Kernel(); auto resource_limit = ResourceLimit::Create(kernel); auto* const current_process = kernel.CurrentProcess(); @@ -1920,11 +2083,11 @@ static ResultCode CreateResourceLimit(Handle* out_handle) { return RESULT_SUCCESS; } -static ResultCode GetResourceLimitLimitValue(u64* out_value, Handle resource_limit, - u32 resource_type) { +static ResultCode GetResourceLimitLimitValue(Core::System& system, u64* out_value, + Handle resource_limit, u32 resource_type) { LOG_DEBUG(Kernel_SVC, "called. Handle={:08X}, Resource type={}", resource_limit, resource_type); - const auto limit_value = RetrieveResourceLimitValue(resource_limit, resource_type, + const auto limit_value = RetrieveResourceLimitValue(system, resource_limit, resource_type, ResourceLimitValueType::LimitValue); if (limit_value.Failed()) { return limit_value.Code(); @@ -1934,11 +2097,11 @@ static ResultCode GetResourceLimitLimitValue(u64* out_value, Handle resource_lim return RESULT_SUCCESS; } -static ResultCode GetResourceLimitCurrentValue(u64* out_value, Handle resource_limit, - u32 resource_type) { +static ResultCode GetResourceLimitCurrentValue(Core::System& system, u64* out_value, + Handle resource_limit, u32 resource_type) { LOG_DEBUG(Kernel_SVC, "called. Handle={:08X}, Resource type={}", resource_limit, resource_type); - const auto current_value = RetrieveResourceLimitValue(resource_limit, resource_type, + const auto current_value = RetrieveResourceLimitValue(system, resource_limit, resource_type, ResourceLimitValueType::CurrentValue); if (current_value.Failed()) { return current_value.Code(); @@ -1948,7 +2111,8 @@ static ResultCode GetResourceLimitCurrentValue(u64* out_value, Handle resource_l return RESULT_SUCCESS; } -static ResultCode SetResourceLimitLimitValue(Handle resource_limit, u32 resource_type, u64 value) { +static ResultCode SetResourceLimitLimitValue(Core::System& system, Handle resource_limit, + u32 resource_type, u64 value) { LOG_DEBUG(Kernel_SVC, "called. Handle={:08X}, Resource type={}, Value={}", resource_limit, resource_type, value); @@ -1958,8 +2122,7 @@ static ResultCode SetResourceLimitLimitValue(Handle resource_limit, u32 resource return ERR_INVALID_ENUM_VALUE; } - auto& kernel = Core::System::GetInstance().Kernel(); - auto* const current_process = kernel.CurrentProcess(); + auto* const current_process = system.Kernel().CurrentProcess(); ASSERT(current_process != nullptr); auto resource_limit_object = @@ -1983,8 +2146,8 @@ static ResultCode SetResourceLimitLimitValue(Handle resource_limit, u32 resource return RESULT_SUCCESS; } -static ResultCode GetProcessList(u32* out_num_processes, VAddr out_process_ids, - u32 out_process_ids_size) { +static ResultCode GetProcessList(Core::System& system, u32* out_num_processes, + VAddr out_process_ids, u32 out_process_ids_size) { LOG_DEBUG(Kernel_SVC, "called. out_process_ids=0x{:016X}, out_process_ids_size={}", out_process_ids, out_process_ids_size); @@ -1996,7 +2159,7 @@ static ResultCode GetProcessList(u32* out_num_processes, VAddr out_process_ids, return ERR_OUT_OF_RANGE; } - const auto& kernel = Core::System::GetInstance().Kernel(); + const auto& kernel = system.Kernel(); const auto& vm_manager = kernel.CurrentProcess()->VMManager(); const auto total_copy_size = out_process_ids_size * sizeof(u64); @@ -2020,8 +2183,8 @@ static ResultCode GetProcessList(u32* out_num_processes, VAddr out_process_ids, return RESULT_SUCCESS; } -ResultCode GetThreadList(u32* out_num_threads, VAddr out_thread_ids, u32 out_thread_ids_size, - Handle debug_handle) { +ResultCode GetThreadList(Core::System& system, u32* out_num_threads, VAddr out_thread_ids, + u32 out_thread_ids_size, Handle debug_handle) { // TODO: Handle this case when debug events are supported. UNIMPLEMENTED_IF(debug_handle != InvalidHandle); @@ -2035,7 +2198,7 @@ ResultCode GetThreadList(u32* out_num_threads, VAddr out_thread_ids, u32 out_thr return ERR_OUT_OF_RANGE; } - const auto* const current_process = Core::System::GetInstance().Kernel().CurrentProcess(); + const auto* const current_process = system.Kernel().CurrentProcess(); const auto& vm_manager = current_process->VMManager(); const auto total_copy_size = out_thread_ids_size * sizeof(u64); @@ -2062,7 +2225,7 @@ ResultCode GetThreadList(u32* out_num_threads, VAddr out_thread_ids, u32 out_thr namespace { struct FunctionDef { - using Func = void(); + using Func = void(Core::System&); u32 id; Func* func; @@ -2190,8 +2353,8 @@ static const FunctionDef SVC_Table[] = { {0x74, nullptr, "MapProcessMemory"}, {0x75, nullptr, "UnmapProcessMemory"}, {0x76, SvcWrap<QueryProcessMemory>, "QueryProcessMemory"}, - {0x77, nullptr, "MapProcessCodeMemory"}, - {0x78, nullptr, "UnmapProcessCodeMemory"}, + {0x77, SvcWrap<MapProcessCodeMemory>, "MapProcessCodeMemory"}, + {0x78, SvcWrap<UnmapProcessCodeMemory>, "UnmapProcessCodeMemory"}, {0x79, nullptr, "CreateProcess"}, {0x7A, nullptr, "StartProcess"}, {0x7B, nullptr, "TerminateProcess"}, @@ -2211,7 +2374,7 @@ static const FunctionDef* GetSVCInfo(u32 func_num) { MICROPROFILE_DEFINE(Kernel_SVC, "Kernel", "SVC", MP_RGB(70, 200, 70)); -void CallSVC(u32 immediate) { +void CallSVC(Core::System& system, u32 immediate) { MICROPROFILE_SCOPE(Kernel_SVC); // Lock the global kernel mutex when we enter the kernel HLE. @@ -2220,7 +2383,7 @@ void CallSVC(u32 immediate) { const FunctionDef* info = GetSVCInfo(immediate); if (info) { if (info->func) { - info->func(); + info->func(system); } else { LOG_CRITICAL(Kernel_SVC, "Unimplemented SVC function {}(..)", info->name); } diff --git a/src/core/hle/kernel/svc.h b/src/core/hle/kernel/svc.h index c37ae0f98..c5539ac1c 100644 --- a/src/core/hle/kernel/svc.h +++ b/src/core/hle/kernel/svc.h @@ -6,8 +6,12 @@ #include "common/common_types.h" +namespace Core { +class System; +} + namespace Kernel { -void CallSVC(u32 immediate); +void CallSVC(Core::System& system, u32 immediate); } // namespace Kernel diff --git a/src/core/hle/kernel/svc_wrap.h b/src/core/hle/kernel/svc_wrap.h index b3733680f..865473c6f 100644 --- a/src/core/hle/kernel/svc_wrap.h +++ b/src/core/hle/kernel/svc_wrap.h @@ -11,278 +11,319 @@ namespace Kernel { -static inline u64 Param(int n) { - return Core::CurrentArmInterface().GetReg(n); +static inline u64 Param(const Core::System& system, int n) { + return system.CurrentArmInterface().GetReg(n); } /** * HLE a function return from the current ARM userland process - * @param res Result to return + * @param system System context + * @param result Result to return */ -static inline void FuncReturn(u64 res) { - Core::CurrentArmInterface().SetReg(0, res); +static inline void FuncReturn(Core::System& system, u64 result) { + system.CurrentArmInterface().SetReg(0, result); } //////////////////////////////////////////////////////////////////////////////////////////////////// // Function wrappers that return type ResultCode -template <ResultCode func(u64)> -void SvcWrap() { - FuncReturn(func(Param(0)).raw); +template <ResultCode func(Core::System&, u64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0)).raw); } -template <ResultCode func(u32)> -void SvcWrap() { - FuncReturn(func(static_cast<u32>(Param(0))).raw); +template <ResultCode func(Core::System&, u32)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, static_cast<u32>(Param(system, 0))).raw); } -template <ResultCode func(u32, u32)> -void SvcWrap() { - FuncReturn(func(static_cast<u32>(Param(0)), static_cast<u32>(Param(1))).raw); +template <ResultCode func(Core::System&, u32, u32)> +void SvcWrap(Core::System& system) { + FuncReturn( + system, + func(system, static_cast<u32>(Param(system, 0)), static_cast<u32>(Param(system, 1))).raw); +} + +template <ResultCode func(Core::System&, u32, u64, u64, u64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), + Param(system, 2), Param(system, 3)) + .raw); } -template <ResultCode func(u32*)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32*)> +void SvcWrap(Core::System& system) { u32 param = 0; - const u32 retval = func(¶m).raw; - Core::CurrentArmInterface().SetReg(1, param); - FuncReturn(retval); + const u32 retval = func(system, ¶m).raw; + system.CurrentArmInterface().SetReg(1, param); + FuncReturn(system, retval); } -template <ResultCode func(u32*, u32)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32*, u32)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; - u32 retval = func(¶m_1, static_cast<u32>(Param(1))).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, static_cast<u32>(Param(system, 1))).raw; + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u32*, u32*)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32*, u32*)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; u32 param_2 = 0; - const u32 retval = func(¶m_1, ¶m_2).raw; + const u32 retval = func(system, ¶m_1, ¶m_2).raw; - auto& arm_interface = Core::CurrentArmInterface(); + auto& arm_interface = system.CurrentArmInterface(); arm_interface.SetReg(1, param_1); arm_interface.SetReg(2, param_2); - FuncReturn(retval); + FuncReturn(system, retval); } -template <ResultCode func(u32*, u64)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32*, u64)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; - const u32 retval = func(¶m_1, Param(1)).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, Param(system, 1)).raw; + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u32*, u64, u32)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32*, u64, u32)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; - const u32 retval = func(¶m_1, Param(1), static_cast<u32>(Param(2))).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = + func(system, ¶m_1, Param(system, 1), static_cast<u32>(Param(system, 2))).raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u64*, u32)> -void SvcWrap() { +template <ResultCode func(Core::System&, u64*, u32)> +void SvcWrap(Core::System& system) { u64 param_1 = 0; - const u32 retval = func(¶m_1, static_cast<u32>(Param(1))).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, static_cast<u32>(Param(system, 1))).raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u64, s32)> -void SvcWrap() { - FuncReturn(func(Param(0), static_cast<s32>(Param(1))).raw); +template <ResultCode func(Core::System&, u64, s32)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), static_cast<s32>(Param(system, 1))).raw); } -template <ResultCode func(u64, u32)> -void SvcWrap() { - FuncReturn(func(Param(0), static_cast<u32>(Param(1))).raw); +template <ResultCode func(Core::System&, u64, u32)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), static_cast<u32>(Param(system, 1))).raw); } -template <ResultCode func(u64*, u64)> -void SvcWrap() { +template <ResultCode func(Core::System&, u64*, u64)> +void SvcWrap(Core::System& system) { u64 param_1 = 0; - u32 retval = func(¶m_1, Param(1)).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, Param(system, 1)).raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u64*, u32, u32)> -void SvcWrap() { +template <ResultCode func(Core::System&, u64*, u32, u32)> +void SvcWrap(Core::System& system) { u64 param_1 = 0; - u32 retval = func(¶m_1, static_cast<u32>(Param(1)), static_cast<u32>(Param(2))).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, static_cast<u32>(Param(system, 1)), + static_cast<u32>(Param(system, 2))) + .raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u32, u64)> -void SvcWrap() { - FuncReturn(func(static_cast<u32>(Param(0)), Param(1)).raw); +template <ResultCode func(Core::System&, u32, u64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1)).raw); } -template <ResultCode func(u32, u32, u64)> -void SvcWrap() { - FuncReturn(func(static_cast<u32>(Param(0)), static_cast<u32>(Param(1)), Param(2)).raw); +template <ResultCode func(Core::System&, u32, u32, u64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), + static_cast<u32>(Param(system, 1)), Param(system, 2)) + .raw); } -template <ResultCode func(u32, u32*, u64*)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32, u32*, u64*)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; u64 param_2 = 0; - ResultCode retval = func(static_cast<u32>(Param(2)), ¶m_1, ¶m_2); - Core::CurrentArmInterface().SetReg(1, param_1); - Core::CurrentArmInterface().SetReg(2, param_2); - FuncReturn(retval.raw); -} + const ResultCode retval = func(system, static_cast<u32>(Param(system, 2)), ¶m_1, ¶m_2); -template <ResultCode func(u64, u64, u32, u32)> -void SvcWrap() { - FuncReturn( - func(Param(0), Param(1), static_cast<u32>(Param(2)), static_cast<u32>(Param(3))).raw); + system.CurrentArmInterface().SetReg(1, param_1); + system.CurrentArmInterface().SetReg(2, param_2); + FuncReturn(system, retval.raw); } -template <ResultCode func(u64, u64, u32, u64)> -void SvcWrap() { - FuncReturn(func(Param(0), Param(1), static_cast<u32>(Param(2)), Param(3)).raw); +template <ResultCode func(Core::System&, u64, u64, u32, u32)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), Param(system, 1), + static_cast<u32>(Param(system, 2)), static_cast<u32>(Param(system, 3))) + .raw); } -template <ResultCode func(u32, u64, u32)> -void SvcWrap() { - FuncReturn(func(static_cast<u32>(Param(0)), Param(1), static_cast<u32>(Param(2))).raw); +template <ResultCode func(Core::System&, u64, u64, u32, u64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), Param(system, 1), + static_cast<u32>(Param(system, 2)), Param(system, 3)) + .raw); } -template <ResultCode func(u64, u64, u64)> -void SvcWrap() { - FuncReturn(func(Param(0), Param(1), Param(2)).raw); +template <ResultCode func(Core::System&, u32, u64, u32)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), + static_cast<u32>(Param(system, 2))) + .raw); } -template <ResultCode func(u64, u64, u32)> -void SvcWrap() { - FuncReturn(func(Param(0), Param(1), static_cast<u32>(Param(2))).raw); +template <ResultCode func(Core::System&, u64, u64, u64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), Param(system, 1), Param(system, 2)).raw); } -template <ResultCode func(u32, u64, u64, u32)> -void SvcWrap() { +template <ResultCode func(Core::System&, u64, u64, u32)> +void SvcWrap(Core::System& system) { FuncReturn( - func(static_cast<u32>(Param(0)), Param(1), Param(2), static_cast<u32>(Param(3))).raw); + system, + func(system, Param(system, 0), Param(system, 1), static_cast<u32>(Param(system, 2))).raw); } -template <ResultCode func(u32, u64, u64)> -void SvcWrap() { - FuncReturn(func(static_cast<u32>(Param(0)), Param(1), Param(2)).raw); +template <ResultCode func(Core::System&, u32, u64, u64, u32)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), + Param(system, 2), static_cast<u32>(Param(system, 3))) + .raw); } -template <ResultCode func(u32*, u64, u64, s64)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32, u64, u64)> +void SvcWrap(Core::System& system) { + FuncReturn( + system, + func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), Param(system, 2)).raw); +} + +template <ResultCode func(Core::System&, u32*, u64, u64, s64)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; - ResultCode retval = - func(¶m_1, Param(1), static_cast<u32>(Param(2)), static_cast<s64>(Param(3))); - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval.raw); + const u32 retval = func(system, ¶m_1, Param(system, 1), static_cast<u32>(Param(system, 2)), + static_cast<s64>(Param(system, 3))) + .raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u64, u64, u32, s64)> -void SvcWrap() { - FuncReturn( - func(Param(0), Param(1), static_cast<u32>(Param(2)), static_cast<s64>(Param(3))).raw); +template <ResultCode func(Core::System&, u64, u64, u32, s64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), Param(system, 1), + static_cast<u32>(Param(system, 2)), static_cast<s64>(Param(system, 3))) + .raw); } -template <ResultCode func(u64*, u64, u64, u64)> -void SvcWrap() { +template <ResultCode func(Core::System&, u64*, u64, u64, u64)> +void SvcWrap(Core::System& system) { u64 param_1 = 0; - u32 retval = func(¶m_1, Param(1), Param(2), Param(3)).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = + func(system, ¶m_1, Param(system, 1), Param(system, 2), Param(system, 3)).raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u32*, u64, u64, u64, u32, s32)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32*, u64, u64, u64, u32, s32)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; - u32 retval = func(¶m_1, Param(1), Param(2), Param(3), static_cast<u32>(Param(4)), - static_cast<s32>(Param(5))) - .raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, Param(system, 1), Param(system, 2), Param(system, 3), + static_cast<u32>(Param(system, 4)), static_cast<s32>(Param(system, 5))) + .raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u32*, u64, u64, u32)> -void SvcWrap() { +template <ResultCode func(Core::System&, u32*, u64, u64, u32)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; - u32 retval = func(¶m_1, Param(1), Param(2), static_cast<u32>(Param(3))).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, Param(system, 1), Param(system, 2), + static_cast<u32>(Param(system, 3))) + .raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(Handle*, u64, u32, u32)> -void SvcWrap() { +template <ResultCode func(Core::System&, Handle*, u64, u32, u32)> +void SvcWrap(Core::System& system) { u32 param_1 = 0; - u32 retval = - func(¶m_1, Param(1), static_cast<u32>(Param(2)), static_cast<u32>(Param(3))).raw; - Core::CurrentArmInterface().SetReg(1, param_1); - FuncReturn(retval); + const u32 retval = func(system, ¶m_1, Param(system, 1), static_cast<u32>(Param(system, 2)), + static_cast<u32>(Param(system, 3))) + .raw; + + system.CurrentArmInterface().SetReg(1, param_1); + FuncReturn(system, retval); } -template <ResultCode func(u64, u32, s32, s64)> -void SvcWrap() { - FuncReturn(func(Param(0), static_cast<u32>(Param(1)), static_cast<s32>(Param(2)), - static_cast<s64>(Param(3))) - .raw); +template <ResultCode func(Core::System&, u64, u32, s32, s64)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), static_cast<u32>(Param(system, 1)), + static_cast<s32>(Param(system, 2)), static_cast<s64>(Param(system, 3))) + .raw); } -template <ResultCode func(u64, u32, s32, s32)> -void SvcWrap() { - FuncReturn(func(Param(0), static_cast<u32>(Param(1)), static_cast<s32>(Param(2)), - static_cast<s32>(Param(3))) - .raw); +template <ResultCode func(Core::System&, u64, u32, s32, s32)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system, Param(system, 0), static_cast<u32>(Param(system, 1)), + static_cast<s32>(Param(system, 2)), static_cast<s32>(Param(system, 3))) + .raw); } //////////////////////////////////////////////////////////////////////////////////////////////////// // Function wrappers that return type u32 -template <u32 func()> -void SvcWrap() { - FuncReturn(func()); +template <u32 func(Core::System&)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system)); } //////////////////////////////////////////////////////////////////////////////////////////////////// // Function wrappers that return type u64 -template <u64 func()> -void SvcWrap() { - FuncReturn(func()); +template <u64 func(Core::System&)> +void SvcWrap(Core::System& system) { + FuncReturn(system, func(system)); } //////////////////////////////////////////////////////////////////////////////////////////////////// /// Function wrappers that return type void -template <void func()> -void SvcWrap() { - func(); +template <void func(Core::System&)> +void SvcWrap(Core::System& system) { + func(system); } -template <void func(s64)> -void SvcWrap() { - func(static_cast<s64>(Param(0))); +template <void func(Core::System&, s64)> +void SvcWrap(Core::System& system) { + func(system, static_cast<s64>(Param(system, 0))); } -template <void func(u64, u64 len)> -void SvcWrap() { - func(Param(0), Param(1)); +template <void func(Core::System&, u64, u64)> +void SvcWrap(Core::System& system) { + func(system, Param(system, 0), Param(system, 1)); } -template <void func(u64, u64, u64)> -void SvcWrap() { - func(Param(0), Param(1), Param(2)); +template <void func(Core::System&, u64, u64, u64)> +void SvcWrap(Core::System& system) { + func(system, Param(system, 0), Param(system, 1), Param(system, 2)); } -template <void func(u32, u64, u64)> -void SvcWrap() { - func(static_cast<u32>(Param(0)), Param(1), Param(2)); +template <void func(Core::System&, u32, u64, u64)> +void SvcWrap(Core::System& system) { + func(system, static_cast<u32>(Param(system, 0)), Param(system, 1), Param(system, 2)); } } // namespace Kernel diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp index 1b891f632..ca52267b2 100644 --- a/src/core/hle/kernel/thread.cpp +++ b/src/core/hle/kernel/thread.cpp @@ -220,11 +220,6 @@ void Thread::SetPriority(u32 priority) { UpdatePriority(); } -void Thread::BoostPriority(u32 priority) { - scheduler->SetThreadPriority(this, priority); - current_priority = priority; -} - void Thread::SetWaitSynchronizationResult(ResultCode result) { context.cpu_registers[0] = result.raw; } diff --git a/src/core/hle/kernel/thread.h b/src/core/hle/kernel/thread.h index 73e5d1bb4..32026d7f0 100644 --- a/src/core/hle/kernel/thread.h +++ b/src/core/hle/kernel/thread.h @@ -106,7 +106,7 @@ public: return "Thread"; } - static const HandleType HANDLE_TYPE = HandleType::Thread; + static constexpr HandleType HANDLE_TYPE = HandleType::Thread; HandleType GetHandleType() const override { return HANDLE_TYPE; } @@ -136,12 +136,6 @@ public: */ void SetPriority(u32 priority); - /** - * Temporarily boosts the thread's priority until the next time it is scheduled - * @param priority The new priority - */ - void BoostPriority(u32 priority); - /// Adds a thread to the list of threads that are waiting for a lock held by this thread. void AddMutexWaiter(SharedPtr<Thread> thread); diff --git a/src/core/hle/kernel/transfer_memory.cpp b/src/core/hle/kernel/transfer_memory.cpp index 23228e1b5..26c4e5e67 100644 --- a/src/core/hle/kernel/transfer_memory.cpp +++ b/src/core/hle/kernel/transfer_memory.cpp @@ -14,8 +14,8 @@ namespace Kernel { TransferMemory::TransferMemory(KernelCore& kernel) : Object{kernel} {} TransferMemory::~TransferMemory() = default; -SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_address, - size_t size, MemoryPermission permissions) { +SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_address, u64 size, + MemoryPermission permissions) { SharedPtr<TransferMemory> transfer_memory{new TransferMemory(kernel)}; transfer_memory->base_address = base_address; @@ -26,7 +26,15 @@ SharedPtr<TransferMemory> TransferMemory::Create(KernelCore& kernel, VAddr base_ return transfer_memory; } -ResultCode TransferMemory::MapMemory(VAddr address, size_t size, MemoryPermission permissions) { +const u8* TransferMemory::GetPointer() const { + return backing_block.get()->data(); +} + +u64 TransferMemory::GetSize() const { + return memory_size; +} + +ResultCode TransferMemory::MapMemory(VAddr address, u64 size, MemoryPermission permissions) { if (memory_size != size) { return ERR_INVALID_SIZE; } @@ -39,13 +47,13 @@ ResultCode TransferMemory::MapMemory(VAddr address, size_t size, MemoryPermissio return ERR_INVALID_STATE; } + backing_block = std::make_shared<std::vector<u8>>(size); + const auto map_state = owner_permissions == MemoryPermission::None ? MemoryState::TransferMemoryIsolated : MemoryState::TransferMemory; auto& vm_manager = owner_process->VMManager(); - const auto map_result = vm_manager.MapMemoryBlock( - address, std::make_shared<std::vector<u8>>(size), 0, size, map_state); - + const auto map_result = vm_manager.MapMemoryBlock(address, backing_block, 0, size, map_state); if (map_result.Failed()) { return map_result.Code(); } @@ -54,7 +62,7 @@ ResultCode TransferMemory::MapMemory(VAddr address, size_t size, MemoryPermissio return RESULT_SUCCESS; } -ResultCode TransferMemory::UnmapMemory(VAddr address, size_t size) { +ResultCode TransferMemory::UnmapMemory(VAddr address, u64 size) { if (memory_size != size) { return ERR_INVALID_SIZE; } diff --git a/src/core/hle/kernel/transfer_memory.h b/src/core/hle/kernel/transfer_memory.h index ec294951e..a140b1e2b 100644 --- a/src/core/hle/kernel/transfer_memory.h +++ b/src/core/hle/kernel/transfer_memory.h @@ -4,6 +4,9 @@ #pragma once +#include <memory> +#include <vector> + #include "core/hle/kernel/object.h" union ResultCode; @@ -25,7 +28,7 @@ class TransferMemory final : public Object { public: static constexpr HandleType HANDLE_TYPE = HandleType::TransferMemory; - static SharedPtr<TransferMemory> Create(KernelCore& kernel, VAddr base_address, size_t size, + static SharedPtr<TransferMemory> Create(KernelCore& kernel, VAddr base_address, u64 size, MemoryPermission permissions); TransferMemory(const TransferMemory&) = delete; @@ -46,6 +49,12 @@ public: return HANDLE_TYPE; } + /// Gets a pointer to the backing block of this instance. + const u8* GetPointer() const; + + /// Gets the size of the memory backing this instance in bytes. + u64 GetSize() const; + /// Attempts to map transfer memory with the given range and memory permissions. /// /// @param address The base address to being mapping memory at. @@ -56,7 +65,7 @@ public: /// the same values that were given when creating the transfer memory /// instance. /// - ResultCode MapMemory(VAddr address, size_t size, MemoryPermission permissions); + ResultCode MapMemory(VAddr address, u64 size, MemoryPermission permissions); /// Unmaps the transfer memory with the given range /// @@ -66,17 +75,20 @@ public: /// @pre The given address and size must be the same as the ones used /// to create the transfer memory instance. /// - ResultCode UnmapMemory(VAddr address, size_t size); + ResultCode UnmapMemory(VAddr address, u64 size); private: explicit TransferMemory(KernelCore& kernel); ~TransferMemory() override; + /// Memory block backing this instance. + std::shared_ptr<std::vector<u8>> backing_block; + /// The base address for the memory managed by this instance. VAddr base_address = 0; /// Size of the memory, in bytes, that this instance manages. - size_t memory_size = 0; + u64 memory_size = 0; /// The memory permissions that are applied to this instance. MemoryPermission owner_permissions{}; diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp index ec0a480ce..f0c0c12fc 100644 --- a/src/core/hle/kernel/vm_manager.cpp +++ b/src/core/hle/kernel/vm_manager.cpp @@ -302,6 +302,86 @@ ResultVal<VAddr> VMManager::SetHeapSize(u64 size) { return MakeResult<VAddr>(heap_region_base); } +ResultCode VMManager::MapCodeMemory(VAddr dst_address, VAddr src_address, u64 size) { + constexpr auto ignore_attribute = MemoryAttribute::LockedForIPC | MemoryAttribute::DeviceMapped; + const auto src_check_result = CheckRangeState( + src_address, size, MemoryState::All, MemoryState::Heap, VMAPermission::All, + VMAPermission::ReadWrite, MemoryAttribute::Mask, MemoryAttribute::None, ignore_attribute); + + if (src_check_result.Failed()) { + return src_check_result.Code(); + } + + const auto mirror_result = + MirrorMemory(dst_address, src_address, size, MemoryState::ModuleCode); + if (mirror_result.IsError()) { + return mirror_result; + } + + // Ensure we lock the source memory region. + const auto src_vma_result = CarveVMARange(src_address, size); + if (src_vma_result.Failed()) { + return src_vma_result.Code(); + } + auto src_vma_iter = *src_vma_result; + src_vma_iter->second.attribute = MemoryAttribute::Locked; + Reprotect(src_vma_iter, VMAPermission::Read); + + // The destination memory region is fine as is, however we need to make it read-only. + return ReprotectRange(dst_address, size, VMAPermission::Read); +} + +ResultCode VMManager::UnmapCodeMemory(VAddr dst_address, VAddr src_address, u64 size) { + constexpr auto ignore_attribute = MemoryAttribute::LockedForIPC | MemoryAttribute::DeviceMapped; + const auto src_check_result = CheckRangeState( + src_address, size, MemoryState::All, MemoryState::Heap, VMAPermission::None, + VMAPermission::None, MemoryAttribute::Mask, MemoryAttribute::Locked, ignore_attribute); + + if (src_check_result.Failed()) { + return src_check_result.Code(); + } + + // Yes, the kernel only checks the first page of the region. + const auto dst_check_result = + CheckRangeState(dst_address, Memory::PAGE_SIZE, MemoryState::FlagModule, + MemoryState::FlagModule, VMAPermission::None, VMAPermission::None, + MemoryAttribute::Mask, MemoryAttribute::None, ignore_attribute); + + if (dst_check_result.Failed()) { + return dst_check_result.Code(); + } + + const auto dst_memory_state = std::get<MemoryState>(*dst_check_result); + const auto dst_contiguous_check_result = CheckRangeState( + dst_address, size, MemoryState::All, dst_memory_state, VMAPermission::None, + VMAPermission::None, MemoryAttribute::Mask, MemoryAttribute::None, ignore_attribute); + + if (dst_contiguous_check_result.Failed()) { + return dst_contiguous_check_result.Code(); + } + + const auto unmap_result = UnmapRange(dst_address, size); + if (unmap_result.IsError()) { + return unmap_result; + } + + // With the mirrored portion unmapped, restore the original region's traits. + const auto src_vma_result = CarveVMARange(src_address, size); + if (src_vma_result.Failed()) { + return src_vma_result.Code(); + } + auto src_vma_iter = *src_vma_result; + src_vma_iter->second.state = MemoryState::Heap; + src_vma_iter->second.attribute = MemoryAttribute::None; + Reprotect(src_vma_iter, VMAPermission::ReadWrite); + + if (dst_memory_state == MemoryState::ModuleCode) { + Core::System::GetInstance().InvalidateCpuInstructionCaches(); + } + + return unmap_result; +} + MemoryInfo VMManager::QueryMemory(VAddr address) const { const auto vma = FindVMA(address); MemoryInfo memory_info{}; diff --git a/src/core/hle/kernel/vm_manager.h b/src/core/hle/kernel/vm_manager.h index 6f484b7bf..288eb9450 100644 --- a/src/core/hle/kernel/vm_manager.h +++ b/src/core/hle/kernel/vm_manager.h @@ -43,6 +43,9 @@ enum class VMAPermission : u8 { ReadExecute = Read | Execute, WriteExecute = Write | Execute, ReadWriteExecute = Read | Write | Execute, + + // Used as a wildcard when checking permissions across memory ranges + All = 0xFF, }; constexpr VMAPermission operator|(VMAPermission lhs, VMAPermission rhs) { @@ -152,6 +155,9 @@ enum class MemoryState : u32 { FlagUncached = 1U << 24, FlagCodeMemory = 1U << 25, + // Wildcard used in range checking to indicate all states. + All = 0xFFFFFFFF, + // Convenience flag sets to reduce repetition IPCFlags = FlagIPC0 | FlagIPC3 | FlagIPC1, @@ -415,6 +421,49 @@ public: /// ResultVal<VAddr> SetHeapSize(u64 size); + /// Maps a region of memory as code memory. + /// + /// @param dst_address The base address of the region to create the aliasing memory region. + /// @param src_address The base address of the region to be aliased. + /// @param size The total amount of memory to map in bytes. + /// + /// @pre Both memory regions lie within the actual addressable address space. + /// + /// @post After this function finishes execution, assuming success, then the address range + /// [dst_address, dst_address+size) will alias the memory region, + /// [src_address, src_address+size). + /// <p> + /// What this also entails is as follows: + /// 1. The aliased region gains the Locked memory attribute. + /// 2. The aliased region becomes read-only. + /// 3. The aliasing region becomes read-only. + /// 4. The aliasing region is created with a memory state of MemoryState::CodeModule. + /// + ResultCode MapCodeMemory(VAddr dst_address, VAddr src_address, u64 size); + + /// Unmaps a region of memory designated as code module memory. + /// + /// @param dst_address The base address of the memory region aliasing the source memory region. + /// @param src_address The base address of the memory region being aliased. + /// @param size The size of the memory region to unmap in bytes. + /// + /// @pre Both memory ranges lie within the actual addressable address space. + /// + /// @pre The memory region being unmapped has been previously been mapped + /// by a call to MapCodeMemory. + /// + /// @post After execution of the function, if successful. the aliasing memory region + /// will be unmapped and the aliased region will have various traits about it + /// restored to what they were prior to the original mapping call preceding + /// this function call. + /// <p> + /// What this also entails is as follows: + /// 1. The state of the memory region will now indicate a general heap region. + /// 2. All memory attributes for the memory region are cleared. + /// 3. Memory permissions for the region are restored to user read/write. + /// + ResultCode UnmapCodeMemory(VAddr dst_address, VAddr src_address, u64 size); + /// Queries the memory manager for information about the given address. /// /// @param address The address to query the memory manager about for information. diff --git a/src/core/hle/kernel/writable_event.h b/src/core/hle/kernel/writable_event.h index c9068dd3d..d00c92a6b 100644 --- a/src/core/hle/kernel/writable_event.h +++ b/src/core/hle/kernel/writable_event.h @@ -37,7 +37,7 @@ public: return name; } - static const HandleType HANDLE_TYPE = HandleType::WritableEvent; + static constexpr HandleType HANDLE_TYPE = HandleType::WritableEvent; HandleType GetHandleType() const override { return HANDLE_TYPE; } diff --git a/src/core/hle/result.h b/src/core/hle/result.h index ab84f5ddc..8a3701151 100644 --- a/src/core/hle/result.h +++ b/src/core/hle/result.h @@ -119,10 +119,6 @@ union ResultCode { BitField<0, 9, ErrorModule> module; BitField<9, 13, u32> description; - // The last bit of `level` is checked by apps and the kernel to determine if a result code is an - // error - BitField<31, 1, u32> is_error; - constexpr explicit ResultCode(u32 raw) : raw(raw) {} constexpr ResultCode(ErrorModule module_, u32 description_) diff --git a/src/core/hle/service/acc/acc.cpp b/src/core/hle/service/acc/acc.cpp index 1f8ed265e..ba7d7acbd 100644 --- a/src/core/hle/service/acc/acc.cpp +++ b/src/core/hle/service/acc/acc.cpp @@ -137,6 +137,7 @@ private: class IManagerForApplication final : public ServiceFramework<IManagerForApplication> { public: IManagerForApplication() : ServiceFramework("IManagerForApplication") { + // clang-format off static const FunctionInfo functions[] = { {0, &IManagerForApplication::CheckAvailability, "CheckAvailability"}, {1, &IManagerForApplication::GetAccountId, "GetAccountId"}, @@ -145,7 +146,10 @@ public: {130, nullptr, "GetNintendoAccountUserResourceCacheForApplication"}, {150, nullptr, "CreateAuthorizationRequest"}, {160, nullptr, "StoreOpenContext"}, + {170, nullptr, "LoadNetworkServiceLicenseKindAsync"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/acc/acc_su.cpp b/src/core/hle/service/acc/acc_su.cpp index 5e2030355..d66233cad 100644 --- a/src/core/hle/service/acc/acc_su.cpp +++ b/src/core/hle/service/acc/acc_su.cpp @@ -8,6 +8,7 @@ namespace Service::Account { ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager) : Module::Interface(std::move(module), std::move(profile_manager), "acc:su") { + // clang-format off static const FunctionInfo functions[] = { {0, &ACC_SU::GetUserCount, "GetUserCount"}, {1, &ACC_SU::GetUserExistence, "GetUserExistence"}, @@ -19,6 +20,7 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {50, &ACC_SU::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_SU::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, {60, nullptr, "ListOpenContextStoredUsers"}, + {99, nullptr, "DebugActivateOpenContextRetention"}, {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, {102, nullptr, "GetBaasAccountManagerForSystemService"}, @@ -29,6 +31,8 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {111, nullptr, "ClearSaveDataThumbnail"}, {112, nullptr, "LoadSaveDataThumbnail"}, {113, nullptr, "GetSaveDataThumbnailExistence"}, + {130, nullptr, "ActivateOpenContextRetention"}, + {140, nullptr, "ListQualifiedUsers"}, {190, nullptr, "GetUserLastOpenedApplication"}, {191, nullptr, "ActivateOpenContextHolder"}, {200, nullptr, "BeginUserRegistration"}, @@ -48,6 +52,8 @@ ACC_SU::ACC_SU(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {998, nullptr, "DebugSetUserStateClose"}, {999, nullptr, "DebugSetUserStateOpen"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/acc/acc_u0.cpp b/src/core/hle/service/acc/acc_u0.cpp index a4d705b45..182f7c7e5 100644 --- a/src/core/hle/service/acc/acc_u0.cpp +++ b/src/core/hle/service/acc/acc_u0.cpp @@ -8,6 +8,7 @@ namespace Service::Account { ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager) : Module::Interface(std::move(module), std::move(profile_manager), "acc:u0") { + // clang-format off static const FunctionInfo functions[] = { {0, &ACC_U0::GetUserCount, "GetUserCount"}, {1, &ACC_U0::GetUserExistence, "GetUserExistence"}, @@ -19,6 +20,7 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {50, &ACC_U0::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_U0::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, {60, nullptr, "ListOpenContextStoredUsers"}, + {99, nullptr, "DebugActivateOpenContextRetention"}, {100, &ACC_U0::InitializeApplicationInfo, "InitializeApplicationInfo"}, {101, &ACC_U0::GetBaasAccountManagerForApplication, "GetBaasAccountManagerForApplication"}, {102, nullptr, "AuthenticateApplicationAsync"}, @@ -27,7 +29,13 @@ ACC_U0::ACC_U0(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {111, nullptr, "ClearSaveDataThumbnail"}, {120, nullptr, "CreateGuestLoginRequest"}, {130, nullptr, "LoadOpenContext"}, + {131, nullptr, "ListOpenContextStoredUsers"}, + {140, nullptr, "InitializeApplicationInfo"}, + {141, nullptr, "ListQualifiedUsers"}, + {150, nullptr, "IsUserAccountSwitchLocked"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/acc/acc_u1.cpp b/src/core/hle/service/acc/acc_u1.cpp index 8fffc93b5..2dd17d935 100644 --- a/src/core/hle/service/acc/acc_u1.cpp +++ b/src/core/hle/service/acc/acc_u1.cpp @@ -8,6 +8,7 @@ namespace Service::Account { ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> profile_manager) : Module::Interface(std::move(module), std::move(profile_manager), "acc:u1") { + // clang-format off static const FunctionInfo functions[] = { {0, &ACC_U1::GetUserCount, "GetUserCount"}, {1, &ACC_U1::GetUserExistence, "GetUserExistence"}, @@ -19,6 +20,7 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {50, &ACC_U1::IsUserRegistrationRequestPermitted, "IsUserRegistrationRequestPermitted"}, {51, &ACC_U1::TrySelectUserWithoutInteraction, "TrySelectUserWithoutInteraction"}, {60, nullptr, "ListOpenContextStoredUsers"}, + {99, nullptr, "DebugActivateOpenContextRetention"}, {100, nullptr, "GetUserRegistrationNotifier"}, {101, nullptr, "GetUserStateChangeNotifier"}, {102, nullptr, "GetBaasAccountManagerForSystemService"}, @@ -29,12 +31,16 @@ ACC_U1::ACC_U1(std::shared_ptr<Module> module, std::shared_ptr<ProfileManager> p {111, nullptr, "ClearSaveDataThumbnail"}, {112, nullptr, "LoadSaveDataThumbnail"}, {113, nullptr, "GetSaveDataThumbnailExistence"}, + {130, nullptr, "ActivateOpenContextRetention"}, + {140, nullptr, "ListQualifiedUsers"}, {190, nullptr, "GetUserLastOpenedApplication"}, {191, nullptr, "ActivateOpenContextHolder"}, {997, nullptr, "DebugInvalidateTokenCacheForUser"}, {998, nullptr, "DebugSetUserStateClose"}, {999, nullptr, "DebugSetUserStateOpen"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/am/am.cpp b/src/core/hle/service/am/am.cpp index d31ab7970..1aa4ce1ac 100644 --- a/src/core/hle/service/am/am.cpp +++ b/src/core/hle/service/am/am.cpp @@ -13,7 +13,7 @@ #include "core/hle/kernel/kernel.h" #include "core/hle/kernel/process.h" #include "core/hle/kernel/readable_event.h" -#include "core/hle/kernel/shared_memory.h" +#include "core/hle/kernel/transfer_memory.h" #include "core/hle/kernel/writable_event.h" #include "core/hle/service/acc/profile_manager.h" #include "core/hle/service/am/am.h" @@ -224,6 +224,7 @@ IDebugFunctions::IDebugFunctions() : ServiceFramework{"IDebugFunctions"} { {20, nullptr, "InvalidateTransitionLayer"}, {30, nullptr, "RequestLaunchApplicationWithUserAndArgumentForDebug"}, {40, nullptr, "GetAppletResourceUsageInfo"}, + {41, nullptr, "SetCpuBoostModeForApplet"}, }; // clang-format on @@ -256,6 +257,7 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger {40, &ISelfController::CreateManagedDisplayLayer, "CreateManagedDisplayLayer"}, {41, nullptr, "IsSystemBufferSharingEnabled"}, {42, nullptr, "GetSystemSharedLayerHandle"}, + {43, nullptr, "GetSystemSharedBufferHandle"}, {50, &ISelfController::SetHandlesRequestToDisplay, "SetHandlesRequestToDisplay"}, {51, nullptr, "ApproveToDisplay"}, {60, nullptr, "OverrideAutoSleepTimeAndDimmingTime"}, @@ -269,9 +271,11 @@ ISelfController::ISelfController(std::shared_ptr<NVFlinger::NVFlinger> nvflinger {68, nullptr, "SetAutoSleepDisabled"}, {69, nullptr, "IsAutoSleepDisabled"}, {70, nullptr, "ReportMultimediaError"}, + {71, nullptr, "GetCurrentIlluminanceEx"}, {80, nullptr, "SetWirelessPriorityMode"}, {90, nullptr, "GetAccumulatedSuspendedTickValue"}, {91, nullptr, "GetAccumulatedSuspendedTickChangedEvent"}, + {100, nullptr, "SetAlbumImageTakenNotificationEnabled"}, {1000, nullptr, "GetDebugStorageChannel"}, }; // clang-format on @@ -363,8 +367,6 @@ void ISelfController::SetPerformanceModeChangedNotification(Kernel::HLERequestCo void ISelfController::SetFocusHandlingMode(Kernel::HLERequestContext& ctx) { // Takes 3 input u8s with each field located immediately after the previous // u8, these are bool flags. No output. - LOG_WARNING(Service_AM, "(STUBBED) called"); - IPC::RequestParser rp{ctx}; struct FocusHandlingModeParams { @@ -372,7 +374,10 @@ void ISelfController::SetFocusHandlingMode(Kernel::HLERequestContext& ctx) { u8 unknown1; u8 unknown2; }; - auto flags = rp.PopRaw<FocusHandlingModeParams>(); + const auto flags = rp.PopRaw<FocusHandlingModeParams>(); + + LOG_WARNING(Service_AM, "(STUBBED) called. unknown0={}, unknown1={}, unknown2={}", + flags.unknown0, flags.unknown1, flags.unknown2); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -515,11 +520,20 @@ ICommonStateGetter::ICommonStateGetter(std::shared_ptr<AppletMessageQueue> msg_q {50, nullptr, "IsVrModeEnabled"}, {51, nullptr, "SetVrModeEnabled"}, {52, nullptr, "SwitchLcdBacklight"}, + {53, nullptr, "BeginVrModeEx"}, + {54, nullptr, "EndVrModeEx"}, {55, nullptr, "IsInControllerFirmwareUpdateSection"}, {60, &ICommonStateGetter::GetDefaultDisplayResolution, "GetDefaultDisplayResolution"}, {61, &ICommonStateGetter::GetDefaultDisplayResolutionChangeEvent, "GetDefaultDisplayResolutionChangeEvent"}, {62, nullptr, "GetHdcpAuthenticationState"}, {63, nullptr, "GetHdcpAuthenticationStateChangeEvent"}, + {64, nullptr, "SetTvPowerStateMatchingMode"}, + {65, nullptr, "GetApplicationIdByContentActionName"}, + {66, nullptr, "SetCpuBoostMode"}, + {80, nullptr, "PerformSystemButtonPressingIfInFocus"}, + {90, nullptr, "SetPerformanceConfigurationChangedNotification"}, + {91, nullptr, "GetCurrentPerformanceConfiguration"}, + {200, nullptr, "GetOperationModeSystemInfo"}, }; // clang-format on @@ -931,19 +945,19 @@ void ILibraryAppletCreator::CreateTransferMemoryStorage(Kernel::HLERequestContex rp.SetCurrentOffset(3); const auto handle{rp.Pop<Kernel::Handle>()}; - const auto shared_mem = - Core::System::GetInstance().CurrentProcess()->GetHandleTable().Get<Kernel::SharedMemory>( + const auto transfer_mem = + Core::System::GetInstance().CurrentProcess()->GetHandleTable().Get<Kernel::TransferMemory>( handle); - if (shared_mem == nullptr) { + if (transfer_mem == nullptr) { LOG_ERROR(Service_AM, "shared_mem is a nullpr for handle={:08X}", handle); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ResultCode(-1)); return; } - const u8* mem_begin = shared_mem->GetPointer(); - const u8* mem_end = mem_begin + shared_mem->GetSize(); + const u8* const mem_begin = transfer_mem->GetPointer(); + const u8* const mem_end = mem_begin + transfer_mem->GetSize(); std::vector<u8> memory{mem_begin, mem_end}; IPC::ResponseBuilder rb{ctx, 2, 0, 1}; @@ -959,6 +973,8 @@ IApplicationFunctions::IApplicationFunctions() : ServiceFramework("IApplicationF {11, nullptr, "CreateApplicationAndPushAndRequestToStartForQuest"}, {12, nullptr, "CreateApplicationAndRequestToStart"}, {13, &IApplicationFunctions::CreateApplicationAndRequestToStartForQuest, "CreateApplicationAndRequestToStartForQuest"}, + {14, nullptr, "CreateApplicationWithAttributeAndPushAndRequestToStartForQuest"}, + {15, nullptr, "CreateApplicationWithAttributeAndRequestToStartForQuest"}, {20, &IApplicationFunctions::EnsureSaveData, "EnsureSaveData"}, {21, &IApplicationFunctions::GetDesiredLanguage, "GetDesiredLanguage"}, {22, &IApplicationFunctions::SetTerminateResult, "SetTerminateResult"}, @@ -1232,6 +1248,7 @@ IGlobalStateController::IGlobalStateController() : ServiceFramework("IGlobalStat {2, nullptr, "StartSleepSequence"}, {3, nullptr, "StartShutdownSequence"}, {4, nullptr, "StartRebootSequence"}, + {9, nullptr, "IsAutoPowerDownRequested"}, {10, nullptr, "LoadAndApplyIdlePolicySettings"}, {11, nullptr, "NotifyCecSettingsChanged"}, {12, nullptr, "SetDefaultHomeButtonLongPressTime"}, diff --git a/src/core/hle/service/am/applet_ae.cpp b/src/core/hle/service/am/applet_ae.cpp index b888f861d..488add8e7 100644 --- a/src/core/hle/service/am/applet_ae.cpp +++ b/src/core/hle/service/am/applet_ae.cpp @@ -16,6 +16,7 @@ public: std::shared_ptr<AppletMessageQueue> msg_queue) : ServiceFramework("ILibraryAppletProxy"), nvflinger(std::move(nvflinger)), msg_queue(std::move(msg_queue)) { + // clang-format off static const FunctionInfo functions[] = { {0, &ILibraryAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"}, {1, &ILibraryAppletProxy::GetSelfController, "GetSelfController"}, @@ -25,8 +26,11 @@ public: {10, &ILibraryAppletProxy::GetProcessWindingController, "GetProcessWindingController"}, {11, &ILibraryAppletProxy::GetLibraryAppletCreator, "GetLibraryAppletCreator"}, {20, &ILibraryAppletProxy::GetApplicationFunctions, "GetApplicationFunctions"}, + {21, nullptr, "GetAppletCommonFunctions"}, {1000, &ILibraryAppletProxy::GetDebugFunctions, "GetDebugFunctions"}, }; + // clang-format on + RegisterHandlers(functions); } @@ -113,6 +117,7 @@ public: std::shared_ptr<AppletMessageQueue> msg_queue) : ServiceFramework("ISystemAppletProxy"), nvflinger(std::move(nvflinger)), msg_queue(std::move(msg_queue)) { + // clang-format off static const FunctionInfo functions[] = { {0, &ISystemAppletProxy::GetCommonStateGetter, "GetCommonStateGetter"}, {1, &ISystemAppletProxy::GetSelfController, "GetSelfController"}, @@ -124,8 +129,11 @@ public: {20, &ISystemAppletProxy::GetHomeMenuFunctions, "GetHomeMenuFunctions"}, {21, &ISystemAppletProxy::GetGlobalStateController, "GetGlobalStateController"}, {22, &ISystemAppletProxy::GetApplicationCreator, "GetApplicationCreator"}, + {23, nullptr, "GetAppletCommonFunctions"}, {1000, &ISystemAppletProxy::GetDebugFunctions, "GetDebugFunctions"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/am/applets/web_browser.cpp b/src/core/hle/service/am/applets/web_browser.cpp index 9b0aa7f5f..7e17df98a 100644 --- a/src/core/hle/service/am/applets/web_browser.cpp +++ b/src/core/hle/service/am/applets/web_browser.cpp @@ -86,7 +86,7 @@ static FileSys::VirtualFile GetManualRomFS() { if (loader.ReadManualRomFS(out) == Loader::ResultStatus::Success) return out; - const auto& installed{FileSystem::GetUnionContents()}; + const auto& installed{Core::System::GetInstance().GetContentProvider()}; const auto res = installed.GetEntry(Core::System::GetInstance().CurrentProcess()->GetTitleID(), FileSys::ContentRecordType::Manual); diff --git a/src/core/hle/service/aoc/aoc_u.cpp b/src/core/hle/service/aoc/aoc_u.cpp index b506bc3dd..51d8c26b4 100644 --- a/src/core/hle/service/aoc/aoc_u.cpp +++ b/src/core/hle/service/aoc/aoc_u.cpp @@ -33,11 +33,11 @@ static bool CheckAOCTitleIDMatchesBase(u64 title_id, u64 base) { static std::vector<u64> AccumulateAOCTitleIDs() { std::vector<u64> add_on_content; - const auto rcu = FileSystem::GetUnionContents(); + const auto& rcu = Core::System::GetInstance().GetContentProvider(); const auto list = rcu.ListEntriesFilter(FileSys::TitleType::AOC, FileSys::ContentRecordType::Data); std::transform(list.begin(), list.end(), std::back_inserter(add_on_content), - [](const FileSys::RegisteredCacheEntry& rce) { return rce.title_id; }); + [](const FileSys::ContentProviderEntry& rce) { return rce.title_id; }); add_on_content.erase( std::remove_if( add_on_content.begin(), add_on_content.end(), @@ -50,6 +50,7 @@ static std::vector<u64> AccumulateAOCTitleIDs() { } AOC_U::AOC_U() : ServiceFramework("aoc:u"), add_on_content(AccumulateAOCTitleIDs()) { + // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "CountAddOnContentByApplicationId"}, {1, nullptr, "ListAddOnContentByApplicationId"}, @@ -60,7 +61,10 @@ AOC_U::AOC_U() : ServiceFramework("aoc:u"), add_on_content(AccumulateAOCTitleIDs {6, nullptr, "PrepareAddOnContentByApplicationId"}, {7, &AOC_U::PrepareAddOnContent, "PrepareAddOnContent"}, {8, &AOC_U::GetAddOnContentListChangedEvent, "GetAddOnContentListChangedEvent"}, + {100, nullptr, "CreateEcPurchasedEventManager"}, }; + // clang-format on + RegisterHandlers(functions); auto& kernel = Core::System::GetInstance().Kernel(); diff --git a/src/core/hle/service/apm/interface.cpp b/src/core/hle/service/apm/interface.cpp index fcacbab72..d058c0245 100644 --- a/src/core/hle/service/apm/interface.cpp +++ b/src/core/hle/service/apm/interface.cpp @@ -87,6 +87,8 @@ APM_Sys::APM_Sys() : ServiceFramework{"apm:sys"} { {3, nullptr, "GetLastThrottlingState"}, {4, nullptr, "ClearLastThrottlingState"}, {5, nullptr, "LoadAndApplySettings"}, + {6, nullptr, "SetCpuBoostMode"}, + {7, nullptr, "GetCurrentPerformanceConfiguration"}, }; // clang-format on diff --git a/src/core/hle/service/audio/audin_u.cpp b/src/core/hle/service/audio/audin_u.cpp index 088410564..d7f1d348d 100644 --- a/src/core/hle/service/audio/audin_u.cpp +++ b/src/core/hle/service/audio/audin_u.cpp @@ -2,9 +2,6 @@ // 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/audin_u.h" namespace Service::Audio { @@ -28,12 +25,12 @@ public: {11, nullptr, "GetAudioInBufferCount"}, {12, nullptr, "SetAudioInDeviceGain"}, {13, nullptr, "GetAudioInDeviceGain"}, + {14, nullptr, "FlushAudioInBuffers"}, }; // clang-format on RegisterHandlers(functions); } - ~IAudioIn() = default; }; AudInU::AudInU() : ServiceFramework("audin:u") { diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp index 21f5e64c7..12875fb42 100644 --- a/src/core/hle/service/audio/audout_u.cpp +++ b/src/core/hle/service/audio/audout_u.cpp @@ -44,7 +44,7 @@ public: std::string&& unique_name) : ServiceFramework("IAudioOut"), audio_core(audio_core), device_name(std::move(device_name)), audio_params(audio_params) { - + // clang-format off static const FunctionInfo functions[] = { {0, &IAudioOut::GetAudioOutState, "GetAudioOutState"}, {1, &IAudioOut::StartAudioOut, "StartAudioOut"}, @@ -58,7 +58,10 @@ public: {9, &IAudioOut::GetAudioOutBufferCount, "GetAudioOutBufferCount"}, {10, nullptr, "GetAudioOutPlayedSampleCount"}, {11, nullptr, "FlushAudioOutBuffers"}, + {12, nullptr, "SetAudioOutVolume"}, + {13, nullptr, "GetAudioOutVolume"}, }; + // clang-format on RegisterHandlers(functions); // This is the event handle used to check if the audio buffer was released @@ -150,7 +153,6 @@ private: void GetReleasedAudioOutBufferImpl(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Audio, "called {}", ctx.Description()); - IPC::RequestParser rp{ctx}; const u64 max_count{ctx.GetWriteBufferSize() / sizeof(u64)}; const auto released_buffers{audio_core.GetTagsAndReleaseBuffers(stream, max_count)}; @@ -194,12 +196,9 @@ private: void AudOutU::ListAudioOutsImpl(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_Audio, "called"); - IPC::RequestParser rp{ctx}; - ctx.WriteBuffer(DefaultDevice); IPC::ResponseBuilder rb{ctx, 3}; - rb.Push(RESULT_SUCCESS); rb.Push<u32>(1); // Amount of audio devices } diff --git a/src/core/hle/service/audio/audrec_u.cpp b/src/core/hle/service/audio/audrec_u.cpp index 6956a2e64..1a5aed9ed 100644 --- a/src/core/hle/service/audio/audrec_u.cpp +++ b/src/core/hle/service/audio/audrec_u.cpp @@ -2,9 +2,6 @@ // 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/audrec_u.h" namespace Service::Audio { @@ -30,7 +27,6 @@ public: RegisterHandlers(functions); } - ~IFinalOutputRecorder() = default; }; AudRecU::AudRecU() : ServiceFramework("audrec:u") { diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp index c9de10a24..1dde6edb7 100644 --- a/src/core/hle/service/audio/audren_u.cpp +++ b/src/core/hle/service/audio/audren_u.cpp @@ -10,6 +10,7 @@ #include "common/alignment.h" #include "common/common_funcs.h" #include "common/logging/log.h" +#include "common/string_util.h" #include "core/core.h" #include "core/hle/ipc_helpers.h" #include "core/hle/kernel/hle_ipc.h" @@ -184,7 +185,6 @@ public: private: void ListAudioDeviceName(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_Audio, "(STUBBED) called"); - IPC::RequestParser rp{ctx}; constexpr std::array<char, 15> audio_interface{{"AudioInterface"}}; ctx.WriteBuffer(audio_interface); @@ -195,13 +195,13 @@ private: } void SetAudioDeviceOutputVolume(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_Audio, "(STUBBED) called"); - IPC::RequestParser rp{ctx}; - f32 volume = static_cast<f32>(rp.Pop<u32>()); + const f32 volume = rp.Pop<f32>(); - auto file_buffer = ctx.ReadBuffer(); - auto end = std::find(file_buffer.begin(), file_buffer.end(), '\0'); + const auto device_name_buffer = ctx.ReadBuffer(); + const std::string name = Common::StringFromBuffer(device_name_buffer); + + LOG_WARNING(Service_Audio, "(STUBBED) called. name={}, volume={}", name, volume); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -209,7 +209,6 @@ private: void GetActiveAudioDeviceName(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_Audio, "(STUBBED) called"); - IPC::RequestParser rp{ctx}; constexpr std::array<char, 12> audio_interface{{"AudioDevice"}}; ctx.WriteBuffer(audio_interface); diff --git a/src/core/hle/service/btdrv/btdrv.cpp b/src/core/hle/service/btdrv/btdrv.cpp index 59ef603e1..974ff8e1a 100644 --- a/src/core/hle/service/btdrv/btdrv.cpp +++ b/src/core/hle/service/btdrv/btdrv.cpp @@ -154,7 +154,8 @@ public: {96, nullptr, "GetLeHidEventInfo"}, {97, nullptr, "RegisterBleHidEvent"}, {98, nullptr, "SetLeScanParameter"}, - {256, nullptr, "GetIsManufacturingMode"} + {256, nullptr, "GetIsManufacturingMode"}, + {257, nullptr, "EmulateBluetoothCrash"}, }; // clang-format on diff --git a/src/core/hle/service/caps/caps.cpp b/src/core/hle/service/caps/caps.cpp index ae7b0720b..907f464ab 100644 --- a/src/core/hle/service/caps/caps.cpp +++ b/src/core/hle/service/caps/caps.cpp @@ -15,32 +15,41 @@ public: explicit CAPS_A() : ServiceFramework{"caps:a"} { // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "Unknown1"}, - {1, nullptr, "Unknown2"}, - {2, nullptr, "Unknown3"}, - {3, nullptr, "Unknown4"}, - {4, nullptr, "Unknown5"}, - {5, nullptr, "Unknown6"}, - {6, nullptr, "Unknown7"}, - {7, nullptr, "Unknown8"}, - {8, nullptr, "Unknown9"}, - {9, nullptr, "Unknown10"}, - {10, nullptr, "Unknown11"}, - {11, nullptr, "Unknown12"}, - {12, nullptr, "Unknown13"}, - {13, nullptr, "Unknown14"}, - {14, nullptr, "Unknown15"}, - {301, nullptr, "Unknown16"}, - {401, nullptr, "Unknown17"}, - {501, nullptr, "Unknown18"}, - {1001, nullptr, "Unknown19"}, - {1002, nullptr, "Unknown20"}, - {8001, nullptr, "Unknown21"}, - {8002, nullptr, "Unknown22"}, - {8011, nullptr, "Unknown23"}, - {8012, nullptr, "Unknown24"}, - {8021, nullptr, "Unknown25"}, - {10011, nullptr, "Unknown26"}, + {0, nullptr, "GetAlbumFileCount"}, + {1, nullptr, "GetAlbumFileList"}, + {2, nullptr, "LoadAlbumFile"}, + {3, nullptr, "DeleteAlbumFile"}, + {4, nullptr, "StorageCopyAlbumFile"}, + {5, nullptr, "IsAlbumMounted"}, + {6, nullptr, "GetAlbumUsage"}, + {7, nullptr, "GetAlbumFileSize"}, + {8, nullptr, "LoadAlbumFileThumbnail"}, + {9, nullptr, "LoadAlbumScreenShotImage"}, + {10, nullptr, "LoadAlbumScreenShotThumbnailImage"}, + {11, nullptr, "GetAlbumEntryFromApplicationAlbumEntry"}, + {12, nullptr, "Unknown12"}, + {13, nullptr, "Unknown13"}, + {14, nullptr, "Unknown14"}, + {15, nullptr, "Unknown15"}, + {16, nullptr, "Unknown16"}, + {17, nullptr, "Unknown17"}, + {18, nullptr, "Unknown18"}, + {202, nullptr, "SaveEditedScreenShot"}, + {301, nullptr, "GetLastThumbnail"}, + {401, nullptr, "GetAutoSavingStorage"}, + {501, nullptr, "GetRequiredStorageSpaceSizeToCopyAll"}, + {1001, nullptr, "Unknown1001"}, + {1002, nullptr, "Unknown1002"}, + {1003, nullptr, "Unknown1003"}, + {8001, nullptr, "ForceAlbumUnmounted"}, + {8002, nullptr, "ResetAlbumMountStatus"}, + {8011, nullptr, "RefreshAlbumCache"}, + {8012, nullptr, "GetAlbumCache"}, + {8013, nullptr, "Unknown8013"}, + {8021, nullptr, "GetAlbumEntryFromApplicationAlbumEntryAruid"}, + {10011, nullptr, "SetInternalErrorConversionEnabled"}, + {50000, nullptr, "Unknown50000"}, + {60002, nullptr, "Unknown60002"}, }; // clang-format on @@ -53,16 +62,17 @@ public: explicit CAPS_C() : ServiceFramework{"caps:c"} { // clang-format off static const FunctionInfo functions[] = { - {2001, nullptr, "Unknown1"}, - {2002, nullptr, "Unknown2"}, - {2011, nullptr, "Unknown3"}, - {2012, nullptr, "Unknown4"}, - {2013, nullptr, "Unknown5"}, - {2014, nullptr, "Unknown6"}, - {2101, nullptr, "Unknown7"}, - {2102, nullptr, "Unknown8"}, - {2201, nullptr, "Unknown9"}, - {2301, nullptr, "Unknown10"}, + {33, nullptr, "Unknown33"}, + {2001, nullptr, "Unknown2001"}, + {2002, nullptr, "Unknown2002"}, + {2011, nullptr, "Unknown2011"}, + {2012, nullptr, "Unknown2012"}, + {2013, nullptr, "Unknown2013"}, + {2014, nullptr, "Unknown2014"}, + {2101, nullptr, "Unknown2101"}, + {2102, nullptr, "Unknown2102"}, + {2201, nullptr, "Unknown2201"}, + {2301, nullptr, "Unknown2301"}, }; // clang-format on @@ -127,11 +137,18 @@ public: explicit CAPS_U() : ServiceFramework{"caps:u"} { // clang-format off static const FunctionInfo functions[] = { + {32, nullptr, "SetShimLibraryVersion"}, {102, nullptr, "GetAlbumFileListByAruid"}, {103, nullptr, "DeleteAlbumFileByAruid"}, {104, nullptr, "GetAlbumFileSizeByAruid"}, + {105, nullptr, "DeleteAlbumFileByAruidForDebug"}, {110, nullptr, "LoadAlbumScreenShotImageByAruid"}, {120, nullptr, "LoadAlbumScreenShotThumbnailImageByAruid"}, + {130, nullptr, "PrecheckToCreateContentsByAruid"}, + {140, nullptr, "GetAlbumFileList1AafeAruidDeprecated"}, + {141, nullptr, "GetAlbumFileList2AafeUidAruidDeprecated"}, + {142, nullptr, "GetAlbumFileList3AaeAruid"}, + {143, nullptr, "GetAlbumFileList4AaeUidAruid"}, {60002, nullptr, "OpenAccessorSessionForApplication"}, }; // clang-format on diff --git a/src/core/hle/service/filesystem/filesystem.cpp b/src/core/hle/service/filesystem/filesystem.cpp index c6da2df43..1ebfeb4bf 100644 --- a/src/core/hle/service/filesystem/filesystem.cpp +++ b/src/core/hle/service/filesystem/filesystem.cpp @@ -197,13 +197,16 @@ ResultCode VfsDirectoryServiceWrapper::RenameDirectory(const std::string& src_pa ResultVal<FileSys::VirtualFile> VfsDirectoryServiceWrapper::OpenFile(const std::string& path_, FileSys::Mode mode) const { - std::string path(FileUtil::SanitizePath(path_)); - auto npath = path; - while (npath.size() > 0 && (npath[0] == '/' || npath[0] == '\\')) - npath = npath.substr(1); + const std::string path(FileUtil::SanitizePath(path_)); + std::string_view npath = path; + while (!npath.empty() && (npath[0] == '/' || npath[0] == '\\')) { + npath.remove_prefix(1); + } + auto file = backing->GetFileRelative(npath); - if (file == nullptr) + if (file == nullptr) { return FileSys::ERROR_PATH_NOT_FOUND; + } if (mode == FileSys::Mode::Append) { return MakeResult<FileSys::VirtualFile>( @@ -319,15 +322,15 @@ ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId stora } ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space, - FileSys::SaveDataDescriptor save_struct) { + const FileSys::SaveDataDescriptor& descriptor) { LOG_TRACE(Service_FS, "Opening Save Data for space_id={:01X}, save_struct={}", - static_cast<u8>(space), save_struct.DebugInfo()); + static_cast<u8>(space), descriptor.DebugInfo()); if (save_data_factory == nullptr) { return FileSys::ERROR_ENTITY_NOT_FOUND; } - return save_data_factory->Open(space, save_struct); + return save_data_factory->Open(space, descriptor); } ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space) { @@ -388,11 +391,6 @@ void WriteSaveDataSize(FileSys::SaveDataType type, u64 title_id, u128 user_id, save_data_factory->WriteSaveDataSize(type, title_id, user_id, new_value); } -FileSys::RegisteredCacheUnion GetUnionContents() { - return FileSys::RegisteredCacheUnion{ - {GetSystemNANDContents(), GetUserNANDContents(), GetSDMCContents()}}; -} - FileSys::RegisteredCache* GetSystemNANDContents() { LOG_TRACE(Service_FS, "Opening System NAND Contents"); @@ -457,6 +455,10 @@ void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) { if (bis_factory == nullptr) { bis_factory = std::make_unique<FileSys::BISFactory>(nand_directory, load_directory, dump_directory); + Core::System::GetInstance().RegisterContentProvider( + FileSys::ContentProviderUnionSlot::SysNAND, bis_factory->GetSystemNANDContents()); + Core::System::GetInstance().RegisterContentProvider( + FileSys::ContentProviderUnionSlot::UserNAND, bis_factory->GetUserNANDContents()); } if (save_data_factory == nullptr) { @@ -465,6 +467,8 @@ void CreateFactories(FileSys::VfsFilesystem& vfs, bool overwrite) { if (sdmc_factory == nullptr) { sdmc_factory = std::make_unique<FileSys::SDMCFactory>(std::move(sd_directory)); + Core::System::GetInstance().RegisterContentProvider(FileSys::ContentProviderUnionSlot::SDMC, + sdmc_factory->GetSDMCContents()); } } diff --git a/src/core/hle/service/filesystem/filesystem.h b/src/core/hle/service/filesystem/filesystem.h index 6fd5e7b23..6481f237c 100644 --- a/src/core/hle/service/filesystem/filesystem.h +++ b/src/core/hle/service/filesystem/filesystem.h @@ -46,7 +46,7 @@ ResultVal<FileSys::VirtualFile> OpenRomFSCurrentProcess(); ResultVal<FileSys::VirtualFile> OpenRomFS(u64 title_id, FileSys::StorageId storage_id, FileSys::ContentRecordType type); ResultVal<FileSys::VirtualDir> OpenSaveData(FileSys::SaveDataSpaceId space, - FileSys::SaveDataDescriptor save_struct); + const FileSys::SaveDataDescriptor& descriptor); ResultVal<FileSys::VirtualDir> OpenSaveDataSpace(FileSys::SaveDataSpaceId space); ResultVal<FileSys::VirtualDir> OpenSDMC(); @@ -54,8 +54,6 @@ FileSys::SaveDataSize ReadSaveDataSize(FileSys::SaveDataType type, u64 title_id, void WriteSaveDataSize(FileSys::SaveDataType type, u64 title_id, u128 user_id, FileSys::SaveDataSize new_value); -FileSys::RegisteredCacheUnion GetUnionContents(); - FileSys::RegisteredCache* GetSystemNANDContents(); FileSys::RegisteredCache* GetUserNANDContents(); FileSys::RegisteredCache* GetSDMCContents(); diff --git a/src/core/hle/service/filesystem/fsp_srv.cpp b/src/core/hle/service/filesystem/fsp_srv.cpp index f03fb629c..e7df8fd98 100644 --- a/src/core/hle/service/filesystem/fsp_srv.cpp +++ b/src/core/hle/service/filesystem/fsp_srv.cpp @@ -115,11 +115,12 @@ private: void Read(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const u64 unk = rp.Pop<u64>(); + const u64 option = rp.Pop<u64>(); const s64 offset = rp.Pop<s64>(); const s64 length = rp.Pop<s64>(); - LOG_DEBUG(Service_FS, "called, offset=0x{:X}, length={}", offset, length); + LOG_DEBUG(Service_FS, "called, option={}, offset=0x{:X}, length={}", option, offset, + length); // Error checking if (length < 0) { @@ -148,11 +149,12 @@ private: void Write(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - const u64 unk = rp.Pop<u64>(); + const u64 option = rp.Pop<u64>(); const s64 offset = rp.Pop<s64>(); const s64 length = rp.Pop<s64>(); - LOG_DEBUG(Service_FS, "called, offset=0x{:X}, length={}", offset, length); + LOG_DEBUG(Service_FS, "called, option={}, offset=0x{:X}, length={}", option, offset, + length); // Error checking if (length < 0) { @@ -250,10 +252,7 @@ private: u64 next_entry_index = 0; void Read(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - const u64 unk = rp.Pop<u64>(); - - LOG_DEBUG(Service_FS, "called, unk=0x{:X}", unk); + LOG_DEBUG(Service_FS, "called."); // Calculate how many entries we can fit in the output buffer const u64 count_entries = ctx.GetWriteBufferSize() / sizeof(FileSys::Entry); @@ -315,61 +314,53 @@ public: void CreateFile(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const auto file_buffer = ctx.ReadBuffer(); + const std::string name = Common::StringFromBuffer(file_buffer); - u64 mode = rp.Pop<u64>(); - u32 size = rp.Pop<u32>(); + const u64 mode = rp.Pop<u64>(); + const u32 size = rp.Pop<u32>(); - LOG_DEBUG(Service_FS, "called file {} mode 0x{:X} size 0x{:08X}", name, mode, size); + LOG_DEBUG(Service_FS, "called. file={}, mode=0x{:X}, size=0x{:08X}", name, mode, size); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(backend.CreateFile(name, size)); } void DeleteFile(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - - auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const auto file_buffer = ctx.ReadBuffer(); + const std::string name = Common::StringFromBuffer(file_buffer); - LOG_DEBUG(Service_FS, "called file {}", name); + LOG_DEBUG(Service_FS, "called. file={}", name); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(backend.DeleteFile(name)); } void CreateDirectory(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - - auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const auto file_buffer = ctx.ReadBuffer(); + const std::string name = Common::StringFromBuffer(file_buffer); - LOG_DEBUG(Service_FS, "called directory {}", name); + LOG_DEBUG(Service_FS, "called. directory={}", name); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(backend.CreateDirectory(name)); } void DeleteDirectory(Kernel::HLERequestContext& ctx) { - const IPC::RequestParser rp{ctx}; - const auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const std::string name = Common::StringFromBuffer(file_buffer); - LOG_DEBUG(Service_FS, "called directory {}", name); + LOG_DEBUG(Service_FS, "called. directory={}", name); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(backend.DeleteDirectory(name)); } void DeleteDirectoryRecursively(Kernel::HLERequestContext& ctx) { - const IPC::RequestParser rp{ctx}; - const auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const std::string name = Common::StringFromBuffer(file_buffer); - LOG_DEBUG(Service_FS, "called directory {}", name); + LOG_DEBUG(Service_FS, "called. directory={}", name); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(backend.DeleteDirectoryRecursively(name)); @@ -386,18 +377,16 @@ public: } void RenameFile(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - std::vector<u8> buffer; buffer.resize(ctx.BufferDescriptorX()[0].Size()); Memory::ReadBlock(ctx.BufferDescriptorX()[0].Address(), buffer.data(), buffer.size()); - std::string src_name = Common::StringFromBuffer(buffer); + const std::string src_name = Common::StringFromBuffer(buffer); buffer.resize(ctx.BufferDescriptorX()[1].Size()); Memory::ReadBlock(ctx.BufferDescriptorX()[1].Address(), buffer.data(), buffer.size()); - std::string dst_name = Common::StringFromBuffer(buffer); + const std::string dst_name = Common::StringFromBuffer(buffer); - LOG_DEBUG(Service_FS, "called file '{}' to file '{}'", src_name, dst_name); + LOG_DEBUG(Service_FS, "called. file '{}' to file '{}'", src_name, dst_name); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(backend.RenameFile(src_name, dst_name)); @@ -406,12 +395,12 @@ public: void OpenFile(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const auto file_buffer = ctx.ReadBuffer(); + const std::string name = Common::StringFromBuffer(file_buffer); - auto mode = static_cast<FileSys::Mode>(rp.Pop<u32>()); + 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, static_cast<u32>(mode)); auto result = backend.OpenFile(name, mode); if (result.Failed()) { @@ -430,13 +419,13 @@ public: void OpenDirectory(Kernel::HLERequestContext& ctx) { IPC::RequestParser rp{ctx}; - auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const auto file_buffer = ctx.ReadBuffer(); + const std::string name = Common::StringFromBuffer(file_buffer); // TODO(Subv): Implement this filter. - u32 filter_flags = rp.Pop<u32>(); + const u32 filter_flags = rp.Pop<u32>(); - LOG_DEBUG(Service_FS, "called directory {} filter {}", name, filter_flags); + LOG_DEBUG(Service_FS, "called. directory={}, filter={}", name, filter_flags); auto result = backend.OpenDirectory(name); if (result.Failed()) { @@ -453,12 +442,10 @@ public: } void GetEntryType(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - - auto file_buffer = ctx.ReadBuffer(); - std::string name = Common::StringFromBuffer(file_buffer); + const auto file_buffer = ctx.ReadBuffer(); + const std::string name = Common::StringFromBuffer(file_buffer); - LOG_DEBUG(Service_FS, "called file {}", name); + LOG_DEBUG(Service_FS, "called. file={}", name); auto result = backend.GetEntryType(name); if (result.Failed()) { @@ -616,7 +603,9 @@ private: u64_le save_id; u64_le title_id; u64_le save_image_size; - INSERT_PADDING_BYTES(0x28); + u16_le index; + FileSys::SaveDataRank rank; + INSERT_PADDING_BYTES(0x25); }; static_assert(sizeof(SaveDataInfo) == 0x60, "SaveDataInfo has incorrect size."); @@ -675,10 +664,13 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") { {100, nullptr, "OpenImageDirectoryFileSystem"}, {110, nullptr, "OpenContentStorageFileSystem"}, {120, nullptr, "OpenCloudBackupWorkStorageFileSystem"}, + {130, nullptr, "OpenCustomStorageFileSystem"}, {200, &FSP_SRV::OpenDataStorageByCurrentProcess, "OpenDataStorageByCurrentProcess"}, {201, nullptr, "OpenDataStorageByProgramId"}, {202, &FSP_SRV::OpenDataStorageByDataId, "OpenDataStorageByDataId"}, {203, &FSP_SRV::OpenPatchDataStorageByCurrentProcess, "OpenPatchDataStorageByCurrentProcess"}, + {204, nullptr, "OpenDataFileSystemByProgramIndex"}, + {205, nullptr, "OpenDataStorageByProgramIndex"}, {400, nullptr, "OpenDeviceOperator"}, {500, nullptr, "OpenSdCardDetectionEventNotifier"}, {501, nullptr, "OpenGameCardDetectionEventNotifier"}, @@ -702,6 +694,7 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") { {614, nullptr, "CorruptSaveDataFileSystemBySaveDataSpaceId"}, {615, nullptr, "QuerySaveDataInternalStorageTotalSize"}, {616, nullptr, "GetSaveDataCommitId"}, + {617, nullptr, "UnregisterExternalKey"}, {620, nullptr, "SetSdCardEncryptionSeed"}, {630, nullptr, "SetSdCardAccessibility"}, {631, nullptr, "IsSdCardAccessible"}, @@ -712,6 +705,7 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") { {710, nullptr, "ResolveAccessFailure"}, {720, nullptr, "AbandonAccessFailure"}, {800, nullptr, "GetAndClearFileSystemProxyErrorInfo"}, + {810, nullptr, "RegisterProgramIndexMapInfo"}, {1000, nullptr, "SetBisRootForHost"}, {1001, nullptr, "SetSaveDataSize"}, {1002, nullptr, "SetSaveDataRootPath"}, @@ -722,6 +716,8 @@ FSP_SRV::FSP_SRV() : ServiceFramework("fsp-srv") { {1007, nullptr, "RegisterUpdatePartition"}, {1008, nullptr, "OpenRegisteredUpdatePartition"}, {1009, nullptr, "GetAndClearMemoryReportInfo"}, + {1010, nullptr, "SetDataStorageRedirectTarget"}, + {1011, nullptr, "OutputAccessLogToSdCard2"}, {1100, nullptr, "OverrideSaveDataTransferTokenSignVerificationKey"}, {1110, nullptr, "CorruptSaveDataFileSystemBySaveDataSpaceId2"}, {1200, nullptr, "OpenMultiCommitManager"}, @@ -779,16 +775,17 @@ void FSP_SRV::CreateSaveDataFileSystem(Kernel::HLERequestContext& ctx) { } void FSP_SRV::OpenSaveDataFileSystem(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; + LOG_INFO(Service_FS, "called."); - auto space_id = rp.PopRaw<FileSys::SaveDataSpaceId>(); - auto unk = rp.Pop<u32>(); - LOG_INFO(Service_FS, "called with unknown={:08X}", unk); - - auto save_struct = rp.PopRaw<FileSys::SaveDataDescriptor>(); + struct Parameters { + FileSys::SaveDataSpaceId save_data_space_id; + FileSys::SaveDataDescriptor descriptor; + }; - auto dir = OpenSaveData(space_id, save_struct); + IPC::RequestParser rp{ctx}; + const auto parameters = rp.PopRaw<Parameters>(); + auto dir = OpenSaveData(parameters.save_data_space_id, parameters.descriptor); if (dir.Failed()) { IPC::ResponseBuilder rb{ctx, 2, 0, 0}; rb.Push(FileSys::ERROR_ENTITY_NOT_FOUND); diff --git a/src/core/hle/service/friend/friend.cpp b/src/core/hle/service/friend/friend.cpp index d9225d624..5100e376c 100644 --- a/src/core/hle/service/friend/friend.cpp +++ b/src/core/hle/service/friend/friend.cpp @@ -12,6 +12,7 @@ namespace Service::Friend { class IFriendService final : public ServiceFramework<IFriendService> { public: IFriendService() : ServiceFramework("IFriendService") { + // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetCompletionEvent"}, {1, nullptr, "Cancel"}, @@ -24,8 +25,7 @@ public: {10400, nullptr, "GetBlockedUserListIds"}, {10500, nullptr, "GetProfileList"}, {10600, nullptr, "DeclareOpenOnlinePlaySession"}, - {10601, &IFriendService::DeclareCloseOnlinePlaySession, - "DeclareCloseOnlinePlaySession"}, + {10601, &IFriendService::DeclareCloseOnlinePlaySession, "DeclareCloseOnlinePlaySession"}, {10610, &IFriendService::UpdateUserPresence, "UpdateUserPresence"}, {10700, nullptr, "GetPlayHistoryRegistrationKey"}, {10701, nullptr, "GetPlayHistoryRegistrationKeyWithNetworkServiceAccountId"}, @@ -88,6 +88,7 @@ public: {30830, nullptr, "ClearPlayLog"}, {49900, nullptr, "DeleteNetworkServiceAccountCache"}, }; + // clang-format on RegisterHandlers(functions); } diff --git a/src/core/hle/service/hid/hid.cpp b/src/core/hle/service/hid/hid.cpp index 63b55758b..a4ad95d96 100644 --- a/src/core/hle/service/hid/hid.cpp +++ b/src/core/hle/service/hid/hid.cpp @@ -210,6 +210,7 @@ Hid::Hid() : ServiceFramework("hid") { {131, nullptr, "IsUnintendedHomeButtonInputProtectionEnabled"}, {132, nullptr, "EnableUnintendedHomeButtonInputProtection"}, {133, nullptr, "SetNpadJoyAssignmentModeSingleWithDestination"}, + {134, nullptr, "SetNpadAnalogStickUseCenterClamp"}, {200, &Hid::GetVibrationDeviceInfo, "GetVibrationDeviceInfo"}, {201, &Hid::SendVibrationValue, "SendVibrationValue"}, {202, &Hid::GetActualVibrationValue, "GetActualVibrationValue"}, @@ -221,6 +222,7 @@ Hid::Hid() : ServiceFramework("hid") { {208, nullptr, "GetActualVibrationGcErmCommand"}, {209, &Hid::BeginPermitVibrationSession, "BeginPermitVibrationSession"}, {210, &Hid::EndPermitVibrationSession, "EndPermitVibrationSession"}, + {211, nullptr, "IsVibrationDeviceMounted"}, {300, &Hid::ActivateConsoleSixAxisSensor, "ActivateConsoleSixAxisSensor"}, {301, &Hid::StartConsoleSixAxisSensor, "StartConsoleSixAxisSensor"}, {302, nullptr, "StopConsoleSixAxisSensor"}, @@ -265,6 +267,7 @@ Hid::Hid() : ServiceFramework("hid") { {523, nullptr, "SetIsPalmaPairedConnectable"}, {524, nullptr, "PairPalma"}, {525, &Hid::SetPalmaBoostMode, "SetPalmaBoostMode"}, + {526, nullptr, "CancelWritePalmaWaveEntry"}, {1000, nullptr, "SetNpadCommunicationMode"}, {1001, nullptr, "GetNpadCommunicationMode"}, }; @@ -797,12 +800,22 @@ public: {232, nullptr, "EnableShipmentMode"}, {233, nullptr, "ClearPairingInfo"}, {234, nullptr, "GetUniquePadDeviceTypeSetInternal"}, + {235, nullptr, "EnableAnalogStickPower"}, {301, nullptr, "GetAbstractedPadHandles"}, {302, nullptr, "GetAbstractedPadState"}, {303, nullptr, "GetAbstractedPadsState"}, {321, nullptr, "SetAutoPilotVirtualPadState"}, {322, nullptr, "UnsetAutoPilotVirtualPadState"}, {323, nullptr, "UnsetAllAutoPilotVirtualPadState"}, + {324, nullptr, "AttachHdlsWorkBuffer"}, + {325, nullptr, "ReleaseHdlsWorkBuffer"}, + {326, nullptr, "DumpHdlsNpadAssignmentState"}, + {327, nullptr, "DumpHdlsStates"}, + {328, nullptr, "ApplyHdlsNpadAssignmentState"}, + {329, nullptr, "ApplyHdlsStateList"}, + {330, nullptr, "AttachHdlsVirtualDevice"}, + {331, nullptr, "DetachHdlsVirtualDevice"}, + {332, nullptr, "SetHdlsState"}, {350, nullptr, "AddRegisteredDevice"}, {400, nullptr, "DisableExternalMcuOnNxDevice"}, {401, nullptr, "DisableRailDeviceFiltering"}, @@ -825,6 +838,7 @@ public: {131, nullptr, "ActivateSleepButton"}, {141, nullptr, "AcquireCaptureButtonEventHandle"}, {151, nullptr, "ActivateCaptureButton"}, + {161, nullptr, "GetPlatformConfig"}, {210, nullptr, "AcquireNfcDeviceUpdateEventHandle"}, {211, nullptr, "GetNpadsWithNfc"}, {212, nullptr, "AcquireNfcActivateEventHandle"}, @@ -894,6 +908,7 @@ public: {827, nullptr, "IsAnalogStickButtonPressed"}, {828, nullptr, "IsAnalogStickInReleasePosition"}, {829, nullptr, "IsAnalogStickInCircumference"}, + {830, nullptr, "SetNotificationLedPattern"}, {850, nullptr, "IsUsbFullKeyControllerEnabled"}, {851, nullptr, "EnableUsbFullKeyController"}, {852, nullptr, "IsUsbConnected"}, diff --git a/src/core/hle/service/ldn/ldn.cpp b/src/core/hle/service/ldn/ldn.cpp index e250595e3..ed5059047 100644 --- a/src/core/hle/service/ldn/ldn.cpp +++ b/src/core/hle/service/ldn/ldn.cpp @@ -52,9 +52,11 @@ public: } }; -class ILocalCommunicationService final : public ServiceFramework<ILocalCommunicationService> { +class ISystemLocalCommunicationService final + : public ServiceFramework<ISystemLocalCommunicationService> { public: - explicit ILocalCommunicationService(const char* name) : ServiceFramework{name} { + explicit ISystemLocalCommunicationService() + : ServiceFramework{"ISystemLocalCommunicationService"} { // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "GetState"}, @@ -84,6 +86,50 @@ public: {304, nullptr, "Disconnect"}, {400, nullptr, "InitializeSystem"}, {401, nullptr, "FinalizeSystem"}, + {402, nullptr, "SetOperationMode"}, + {403, nullptr, "InitializeSystem2"}, + }; + // clang-format on + + RegisterHandlers(functions); + } +}; + +class IUserLocalCommunicationService final + : public ServiceFramework<IUserLocalCommunicationService> { +public: + explicit IUserLocalCommunicationService() : ServiceFramework{"IUserLocalCommunicationService"} { + // clang-format off + static const FunctionInfo functions[] = { + {0, nullptr, "GetState"}, + {1, nullptr, "GetNetworkInfo"}, + {2, nullptr, "GetIpv4Address"}, + {3, nullptr, "GetDisconnectReason"}, + {4, nullptr, "GetSecurityParameter"}, + {5, nullptr, "GetNetworkConfig"}, + {100, nullptr, "AttachStateChangeEvent"}, + {101, nullptr, "GetNetworkInfoLatestUpdate"}, + {102, nullptr, "Scan"}, + {103, nullptr, "ScanPrivate"}, + {104, nullptr, "SetWirelessControllerRestriction"}, + {200, nullptr, "OpenAccessPoint"}, + {201, nullptr, "CloseAccessPoint"}, + {202, nullptr, "CreateNetwork"}, + {203, nullptr, "CreateNetworkPrivate"}, + {204, nullptr, "DestroyNetwork"}, + {205, nullptr, "Reject"}, + {206, nullptr, "SetAdvertiseData"}, + {207, nullptr, "SetStationAcceptPolicy"}, + {208, nullptr, "AddAcceptFilterEntry"}, + {209, nullptr, "ClearAcceptFilter"}, + {300, nullptr, "OpenStation"}, + {301, nullptr, "CloseStation"}, + {302, nullptr, "Connect"}, + {303, nullptr, "ConnectPrivate"}, + {304, nullptr, "Disconnect"}, + {400, nullptr, "Initialize"}, + {401, nullptr, "Finalize"}, + {402, nullptr, "SetOperationMode"}, }; // clang-format on @@ -108,7 +154,7 @@ public: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ILocalCommunicationService>("ISystemLocalCommunicationService"); + rb.PushIpcInterface<ISystemLocalCommunicationService>(); } }; @@ -129,7 +175,7 @@ public: IPC::ResponseBuilder rb{ctx, 2, 0, 1}; rb.Push(RESULT_SUCCESS); - rb.PushIpcInterface<ILocalCommunicationService>("IUserLocalCommunicationService"); + rb.PushIpcInterface<IUserLocalCommunicationService>(); } }; diff --git a/src/core/hle/service/ldr/ldr.cpp b/src/core/hle/service/ldr/ldr.cpp index d65693fc7..5af925515 100644 --- a/src/core/hle/service/ldr/ldr.cpp +++ b/src/core/hle/service/ldr/ldr.cpp @@ -86,6 +86,7 @@ public: {2, &RelocatableObject::LoadNrr, "LoadNrr"}, {3, &RelocatableObject::UnloadNrr, "UnloadNrr"}, {4, &RelocatableObject::Initialize, "Initialize"}, + {10, nullptr, "LoadNrrEx"}, }; // clang-format on @@ -93,12 +94,18 @@ public: } void LoadNrr(Kernel::HLERequestContext& ctx) { + struct Parameters { + u64_le process_id; + u64_le nrr_address; + u64_le nrr_size; + }; + IPC::RequestParser rp{ctx}; - rp.Skip(2, false); - const VAddr nrr_addr{rp.Pop<VAddr>()}; - const u64 nrr_size{rp.Pop<u64>()}; - LOG_DEBUG(Service_LDR, "called with nrr_addr={:016X}, nrr_size={:016X}", nrr_addr, - nrr_size); + const auto [process_id, nrr_address, nrr_size] = rp.PopRaw<Parameters>(); + + LOG_DEBUG(Service_LDR, + "called with process_id={:016X}, nrr_address={:016X}, nrr_size={:016X}", + process_id, nrr_address, nrr_size); if (!initialized) { LOG_ERROR(Service_LDR, "LDR:RO not initialized before use!"); @@ -116,24 +123,26 @@ public: } // NRR Address does not fall on 0x1000 byte boundary - if (!Common::Is4KBAligned(nrr_addr)) { - LOG_ERROR(Service_LDR, "NRR Address has invalid alignment (actual {:016X})!", nrr_addr); + if (!Common::Is4KBAligned(nrr_address)) { + LOG_ERROR(Service_LDR, "NRR Address has invalid alignment (actual {:016X})!", + nrr_address); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_ALIGNMENT); return; } // NRR Size is zero or causes overflow - if (nrr_addr + nrr_size <= nrr_addr || nrr_size == 0 || !Common::Is4KBAligned(nrr_size)) { + if (nrr_address + nrr_size <= nrr_address || nrr_size == 0 || + !Common::Is4KBAligned(nrr_size)) { LOG_ERROR(Service_LDR, "NRR Size is invalid! (nrr_address={:016X}, nrr_size={:016X})", - nrr_addr, nrr_size); + nrr_address, nrr_size); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_SIZE); return; } // Read NRR data from memory std::vector<u8> nrr_data(nrr_size); - Memory::ReadBlock(nrr_addr, nrr_data.data(), nrr_size); + Memory::ReadBlock(nrr_address, nrr_data.data(), nrr_size); NRRHeader header; std::memcpy(&header, nrr_data.data(), sizeof(NRRHeader)); @@ -174,7 +183,7 @@ public: hashes.emplace_back(hash); } - nrr.insert_or_assign(nrr_addr, std::move(hashes)); + nrr.insert_or_assign(nrr_address, std::move(hashes)); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -188,23 +197,30 @@ public: return; } + struct Parameters { + u64_le process_id; + u64_le nrr_address; + }; + IPC::RequestParser rp{ctx}; - rp.Skip(2, false); - const auto nrr_addr{rp.Pop<VAddr>()}; - LOG_DEBUG(Service_LDR, "called with nrr_addr={:016X}", nrr_addr); + const auto [process_id, nrr_address] = rp.PopRaw<Parameters>(); + + LOG_DEBUG(Service_LDR, "called with process_id={:016X}, nrr_addr={:016X}", process_id, + nrr_address); - if (!Common::Is4KBAligned(nrr_addr)) { - LOG_ERROR(Service_LDR, "NRR Address has invalid alignment (actual {:016X})!", nrr_addr); + if (!Common::Is4KBAligned(nrr_address)) { + LOG_ERROR(Service_LDR, "NRR Address has invalid alignment (actual {:016X})!", + nrr_address); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_ALIGNMENT); return; } - const auto iter = nrr.find(nrr_addr); + const auto iter = nrr.find(nrr_address); if (iter == nrr.end()) { LOG_ERROR(Service_LDR, "Attempting to unload NRR which has not been loaded! (addr={:016X})", - nrr_addr); + nrr_address); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_NRR_ADDRESS); return; @@ -216,16 +232,22 @@ public: } void LoadNro(Kernel::HLERequestContext& ctx) { + struct Parameters { + u64_le process_id; + u64_le image_address; + u64_le image_size; + u64_le bss_address; + u64_le bss_size; + }; + IPC::RequestParser rp{ctx}; - rp.Skip(2, false); - const VAddr nro_addr{rp.Pop<VAddr>()}; - const u64 nro_size{rp.Pop<u64>()}; - const VAddr bss_addr{rp.Pop<VAddr>()}; - const u64 bss_size{rp.Pop<u64>()}; - LOG_DEBUG( - Service_LDR, - "called with nro_addr={:016X}, nro_size={:016X}, bss_addr={:016X}, bss_size={:016X}", - nro_addr, nro_size, bss_addr, bss_size); + const auto [process_id, nro_address, nro_size, bss_address, bss_size] = + rp.PopRaw<Parameters>(); + + LOG_DEBUG(Service_LDR, + "called with pid={:016X}, nro_addr={:016X}, nro_size={:016X}, bss_addr={:016X}, " + "bss_size={:016X}", + process_id, nro_address, nro_size, bss_address, bss_size); if (!initialized) { LOG_ERROR(Service_LDR, "LDR:RO not initialized before use!"); @@ -243,8 +265,9 @@ public: } // NRO Address does not fall on 0x1000 byte boundary - if (!Common::Is4KBAligned(nro_addr)) { - LOG_ERROR(Service_LDR, "NRO Address has invalid alignment (actual {:016X})!", nro_addr); + if (!Common::Is4KBAligned(nro_address)) { + LOG_ERROR(Service_LDR, "NRO Address has invalid alignment (actual {:016X})!", + nro_address); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_ALIGNMENT); return; @@ -252,15 +275,15 @@ public: // NRO Size or BSS Size is zero or causes overflow const auto nro_size_valid = - nro_size != 0 && nro_addr + nro_size > nro_addr && Common::Is4KBAligned(nro_size); - const auto bss_size_valid = - nro_size + bss_size >= nro_size && (bss_size == 0 || bss_addr + bss_size > bss_addr); + nro_size != 0 && nro_address + nro_size > nro_address && Common::Is4KBAligned(nro_size); + const auto bss_size_valid = nro_size + bss_size >= nro_size && + (bss_size == 0 || bss_address + bss_size > bss_address); if (!nro_size_valid || !bss_size_valid) { LOG_ERROR(Service_LDR, "NRO Size or BSS Size is invalid! (nro_address={:016X}, nro_size={:016X}, " "bss_address={:016X}, bss_size={:016X})", - nro_addr, nro_size, bss_addr, bss_size); + nro_address, nro_size, bss_address, bss_size); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_SIZE); return; @@ -268,7 +291,7 @@ public: // Read NRO data from memory std::vector<u8> nro_data(nro_size); - Memory::ReadBlock(nro_addr, nro_data.data(), nro_size); + Memory::ReadBlock(nro_address, nro_data.data(), nro_size); SHA256Hash hash{}; mbedtls_sha256(nro_data.data(), nro_data.size(), hash.data(), 0); @@ -318,17 +341,18 @@ public: return; } - ASSERT(vm_manager - .MirrorMemory(*map_address, nro_addr, nro_size, Kernel::MemoryState::ModuleCode) - .IsSuccess()); - ASSERT(vm_manager.UnmapRange(nro_addr, nro_size).IsSuccess()); + ASSERT( + vm_manager + .MirrorMemory(*map_address, nro_address, nro_size, Kernel::MemoryState::ModuleCode) + .IsSuccess()); + ASSERT(vm_manager.UnmapRange(nro_address, nro_size).IsSuccess()); if (bss_size > 0) { ASSERT(vm_manager - .MirrorMemory(*map_address + nro_size, bss_addr, bss_size, + .MirrorMemory(*map_address + nro_size, bss_address, bss_size, Kernel::MemoryState::ModuleCode) .IsSuccess()); - ASSERT(vm_manager.UnmapRange(bss_addr, bss_size).IsSuccess()); + ASSERT(vm_manager.UnmapRange(bss_address, bss_size).IsSuccess()); } vm_manager.ReprotectRange(*map_address, header.text_size, @@ -348,13 +372,6 @@ public: } void UnloadNro(Kernel::HLERequestContext& ctx) { - IPC::RequestParser rp{ctx}; - rp.Skip(2, false); - const VAddr mapped_addr{rp.PopRaw<VAddr>()}; - const VAddr heap_addr{rp.PopRaw<VAddr>()}; - LOG_DEBUG(Service_LDR, "called with mapped_addr={:016X}, heap_addr={:016X}", mapped_addr, - heap_addr); - if (!initialized) { LOG_ERROR(Service_LDR, "LDR:RO not initialized before use!"); IPC::ResponseBuilder rb{ctx, 2}; @@ -362,22 +379,30 @@ public: return; } - if (!Common::Is4KBAligned(mapped_addr) || !Common::Is4KBAligned(heap_addr)) { - LOG_ERROR(Service_LDR, - "NRO/BSS Address has invalid alignment (actual nro_addr={:016X}, " - "bss_addr={:016X})!", - mapped_addr, heap_addr); + struct Parameters { + u64_le process_id; + u64_le nro_address; + }; + + IPC::RequestParser rp{ctx}; + const auto [process_id, nro_address] = rp.PopRaw<Parameters>(); + LOG_DEBUG(Service_LDR, "called with process_id={:016X}, nro_address=0x{:016X}", process_id, + nro_address); + + if (!Common::Is4KBAligned(nro_address)) { + LOG_ERROR(Service_LDR, "NRO address has invalid alignment (nro_address=0x{:016X})", + nro_address); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_ALIGNMENT); return; } - const auto iter = nro.find(mapped_addr); + const auto iter = nro.find(nro_address); if (iter == nro.end()) { LOG_ERROR(Service_LDR, - "The NRO attempting to unmap was not mapped or has an invalid address " - "(actual {:016X})!", - mapped_addr); + "The NRO attempting to be unmapped was not mapped or has an invalid address " + "(nro_address=0x{:016X})!", + nro_address); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(ERROR_INVALID_NRO_ADDRESS); return; @@ -386,10 +411,7 @@ public: auto& vm_manager = Core::CurrentProcess()->VMManager(); const auto& nro_size = iter->second.size; - ASSERT(vm_manager - .MirrorMemory(heap_addr, mapped_addr, nro_size, Kernel::MemoryState::ModuleCode) - .IsSuccess()); - ASSERT(vm_manager.UnmapRange(mapped_addr, nro_size).IsSuccess()); + ASSERT(vm_manager.UnmapRange(nro_address, nro_size).IsSuccess()); Core::System::GetInstance().InvalidateCpuInstructionCaches(); @@ -459,11 +481,10 @@ private: std::map<VAddr, NROInfo> nro; std::map<VAddr, std::vector<SHA256Hash>> nrr; - bool IsValidNROHash(const SHA256Hash& hash) { - return std::any_of( - nrr.begin(), nrr.end(), [&hash](const std::pair<VAddr, std::vector<SHA256Hash>>& p) { - return std::find(p.second.begin(), p.second.end(), hash) != p.second.end(); - }); + bool IsValidNROHash(const SHA256Hash& hash) const { + return std::any_of(nrr.begin(), nrr.end(), [&hash](const auto& p) { + return std::find(p.second.begin(), p.second.end(), hash) != p.second.end(); + }); } static bool IsValidNRO(const NROHeader& header, u64 nro_size, u64 bss_size) { diff --git a/src/core/hle/service/nifm/nifm.cpp b/src/core/hle/service/nifm/nifm.cpp index 60479bb45..f92571008 100644 --- a/src/core/hle/service/nifm/nifm.cpp +++ b/src/core/hle/service/nifm/nifm.cpp @@ -15,12 +15,16 @@ namespace Service::NIFM { class IScanRequest final : public ServiceFramework<IScanRequest> { public: explicit IScanRequest() : ServiceFramework("IScanRequest") { + // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "Submit"}, {1, nullptr, "IsProcessing"}, {2, nullptr, "GetResult"}, {3, nullptr, "GetSystemEventReadableHandle"}, + {4, nullptr, "SetChannels"}, }; + // clang-format on + RegisterHandlers(functions); } }; diff --git a/src/core/hle/service/npns/npns.cpp b/src/core/hle/service/npns/npns.cpp index ccb6f9da9..8751522ca 100644 --- a/src/core/hle/service/npns/npns.cpp +++ b/src/core/hle/service/npns/npns.cpp @@ -45,7 +45,7 @@ public: {114, nullptr, "AttachJid"}, {115, nullptr, "DetachJid"}, {201, nullptr, "RequestChangeStateForceTimed"}, - {102, nullptr, "RequestChangeStateForceAsync"}, + {202, nullptr, "RequestChangeStateForceAsync"}, }; // clang-format on @@ -73,6 +73,7 @@ public: {103, nullptr, "GetState"}, {104, nullptr, "GetStatistics"}, {111, nullptr, "GetJid"}, + {120, nullptr, "CreateNotificationReceiver"}, }; // clang-format on diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h index ace71169f..12f3ef825 100644 --- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h +++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h @@ -18,7 +18,7 @@ class nvmap; class nvdisp_disp0 final : public nvdevice { public: explicit nvdisp_disp0(std::shared_ptr<nvmap> nvmap_dev); - ~nvdisp_disp0(); + ~nvdisp_disp0() override; u32 ioctl(Ioctl command, const std::vector<u8>& input, std::vector<u8>& output) override; diff --git a/src/core/hle/service/nvdrv/interface.h b/src/core/hle/service/nvdrv/interface.h index fe311b069..5b4889910 100644 --- a/src/core/hle/service/nvdrv/interface.h +++ b/src/core/hle/service/nvdrv/interface.h @@ -17,7 +17,7 @@ namespace Service::Nvidia { class NVDRV final : public ServiceFramework<NVDRV> { public: NVDRV(std::shared_ptr<Module> nvdrv, const char* name); - ~NVDRV(); + ~NVDRV() override; private: void Open(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/nvdrv/nvmemp.h b/src/core/hle/service/nvdrv/nvmemp.h index 5a4dfc1f9..6eafb1346 100644 --- a/src/core/hle/service/nvdrv/nvmemp.h +++ b/src/core/hle/service/nvdrv/nvmemp.h @@ -11,7 +11,7 @@ namespace Service::Nvidia { class NVMEMP final : public ServiceFramework<NVMEMP> { public: NVMEMP(); - ~NVMEMP(); + ~NVMEMP() override; private: void Cmd0(Kernel::HLERequestContext& ctx); diff --git a/src/core/hle/service/nvflinger/nvflinger.cpp b/src/core/hle/service/nvflinger/nvflinger.cpp index c7f5bbf28..3c5c53e24 100644 --- a/src/core/hle/service/nvflinger/nvflinger.cpp +++ b/src/core/hle/service/nvflinger/nvflinger.cpp @@ -21,12 +21,13 @@ #include "core/hle/service/vi/display/vi_display.h" #include "core/hle/service/vi/layer/vi_layer.h" #include "core/perf_stats.h" +#include "core/settings.h" #include "video_core/renderer_base.h" namespace Service::NVFlinger { -constexpr std::size_t SCREEN_REFRESH_RATE = 60; -constexpr s64 frame_ticks = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / SCREEN_REFRESH_RATE); +constexpr s64 frame_ticks = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / 60); +constexpr s64 frame_ticks_30fps = static_cast<s64>(Core::Timing::BASE_CLOCK_RATE / 30); NVFlinger::NVFlinger(Core::Timing::CoreTiming& core_timing) : core_timing{core_timing} { displays.emplace_back(0, "Default"); @@ -36,13 +37,15 @@ NVFlinger::NVFlinger(Core::Timing::CoreTiming& core_timing) : core_timing{core_t displays.emplace_back(4, "Null"); // Schedule the screen composition events - composition_event = - core_timing.RegisterEvent("ScreenComposition", [this](u64 userdata, s64 cycles_late) { + const auto ticks = Settings::values.force_30fps_mode ? frame_ticks_30fps : frame_ticks; + + composition_event = core_timing.RegisterEvent( + "ScreenComposition", [this, ticks](u64 userdata, s64 cycles_late) { Compose(); - this->core_timing.ScheduleEvent(frame_ticks - cycles_late, composition_event); + this->core_timing.ScheduleEvent(ticks - cycles_late, composition_event); }); - core_timing.ScheduleEvent(frame_ticks, composition_event); + core_timing.ScheduleEvent(ticks, composition_event); } NVFlinger::~NVFlinger() { @@ -62,6 +65,7 @@ std::optional<u64> NVFlinger::OpenDisplay(std::string_view name) { const auto itr = std::find_if(displays.begin(), displays.end(), [&](const VI::Display& display) { return display.GetName() == name; }); + if (itr == displays.end()) { return {}; } diff --git a/src/core/hle/service/pctl/module.cpp b/src/core/hle/service/pctl/module.cpp index 6081f41e1..c75b4ee34 100644 --- a/src/core/hle/service/pctl/module.cpp +++ b/src/core/hle/service/pctl/module.cpp @@ -12,10 +12,10 @@ namespace Service::PCTL { class IParentalControlService final : public ServiceFramework<IParentalControlService> { public: IParentalControlService() : ServiceFramework("IParentalControlService") { + // clang-format off static const FunctionInfo functions[] = { {1, &IParentalControlService::Initialize, "Initialize"}, - {1001, &IParentalControlService::CheckFreeCommunicationPermission, - "CheckFreeCommunicationPermission"}, + {1001, &IParentalControlService::CheckFreeCommunicationPermission, "CheckFreeCommunicationPermission"}, {1002, nullptr, "ConfirmLaunchApplicationPermission"}, {1003, nullptr, "ConfirmResumeApplicationPermission"}, {1004, nullptr, "ConfirmSnsPostPermission"}, @@ -30,6 +30,7 @@ public: {1013, nullptr, "ConfirmStereoVisionPermission"}, {1014, nullptr, "ConfirmPlayableApplicationVideoOld"}, {1015, nullptr, "ConfirmPlayableApplicationVideo"}, + {1016, nullptr, "ConfirmShowNewsPermission"}, {1031, nullptr, "IsRestrictionEnabled"}, {1032, nullptr, "GetSafetyLevel"}, {1033, nullptr, "SetSafetyLevel"}, @@ -45,6 +46,7 @@ public: {1045, nullptr, "UpdateFreeCommunicationApplicationList"}, {1046, nullptr, "DisableFeaturesForReset"}, {1047, nullptr, "NotifyApplicationDownloadStarted"}, + {1048, nullptr, "NotifyNetworkProfileCreated"}, {1061, nullptr, "ConfirmStereoVisionRestrictionConfigurable"}, {1062, nullptr, "GetStereoVisionRestriction"}, {1063, nullptr, "SetStereoVisionRestriction"}, @@ -63,6 +65,7 @@ public: {1411, nullptr, "GetPairingAccountInfo"}, {1421, nullptr, "GetAccountNickname"}, {1424, nullptr, "GetAccountState"}, + {1425, nullptr, "RequestPostEvents"}, {1432, nullptr, "GetSynchronizationEvent"}, {1451, nullptr, "StartPlayTimer"}, {1452, nullptr, "StopPlayTimer"}, diff --git a/src/core/hle/service/pm/pm.cpp b/src/core/hle/service/pm/pm.cpp index 6b27dc4a3..ebcc41a43 100644 --- a/src/core/hle/service/pm/pm.cpp +++ b/src/core/hle/service/pm/pm.cpp @@ -42,15 +42,18 @@ private: class DebugMonitor final : public ServiceFramework<DebugMonitor> { public: explicit DebugMonitor() : ServiceFramework{"pm:dmnt"} { + // clang-format off static const FunctionInfo functions[] = { - {0, nullptr, "IsDebugMode"}, - {1, nullptr, "GetDebugProcesses"}, - {2, nullptr, "StartDebugProcess"}, - {3, nullptr, "GetTitlePid"}, - {4, nullptr, "EnableDebugForTitleId"}, - {5, nullptr, "GetApplicationPid"}, - {6, nullptr, "EnableDebugForApplication"}, + {0, nullptr, "GetDebugProcesses"}, + {1, nullptr, "StartDebugProcess"}, + {2, nullptr, "GetTitlePid"}, + {3, nullptr, "EnableDebugForTitleId"}, + {4, nullptr, "GetApplicationPid"}, + {5, nullptr, "EnableDebugForApplication"}, + {6, nullptr, "DisableDebug"}, }; + // clang-format on + RegisterHandlers(functions); } }; @@ -68,6 +71,7 @@ public: class Shell final : public ServiceFramework<Shell> { public: explicit Shell() : ServiceFramework{"pm:shell"} { + // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "LaunchProcess"}, {1, nullptr, "TerminateProcessByPid"}, @@ -77,7 +81,10 @@ public: {5, nullptr, "NotifyBootFinished"}, {6, nullptr, "GetApplicationPid"}, {7, nullptr, "BoostSystemMemoryResourceLimit"}, + {8, nullptr, "EnableAdditionalSystemThreads"}, }; + // clang-format on + RegisterHandlers(functions); } }; diff --git a/src/core/hle/service/service.h b/src/core/hle/service/service.h index 830790269..abbfe5524 100644 --- a/src/core/hle/service/service.h +++ b/src/core/hle/service/service.h @@ -90,7 +90,7 @@ private: Kernel::HLERequestContext& ctx); ServiceFrameworkBase(const char* service_name, u32 max_sessions, InvokerFn* handler_invoker); - ~ServiceFrameworkBase(); + ~ServiceFrameworkBase() override; void RegisterHandlersBase(const FunctionInfoBase* functions, std::size_t n); void ReportUnimplementedFunction(Kernel::HLERequestContext& ctx, const FunctionInfoBase* info); diff --git a/src/core/hle/service/set/set.cpp b/src/core/hle/service/set/set.cpp index 1afc43f75..4ecb6bcef 100644 --- a/src/core/hle/service/set/set.cpp +++ b/src/core/hle/service/set/set.cpp @@ -116,6 +116,7 @@ void SET::GetLanguageCode(Kernel::HLERequestContext& ctx) { } SET::SET() : ServiceFramework("set") { + // clang-format off static const FunctionInfo functions[] = { {0, &SET::GetLanguageCode, "GetLanguageCode"}, {1, &SET::GetAvailableLanguageCodes, "GetAvailableLanguageCodes"}, @@ -126,7 +127,10 @@ SET::SET() : ServiceFramework("set") { {6, &SET::GetAvailableLanguageCodeCount2, "GetAvailableLanguageCodeCount2"}, {7, nullptr, "GetKeyCodeMap"}, {8, nullptr, "GetQuestFlag"}, + {9, nullptr, "GetKeyCodeMap2"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/set/set_cal.cpp b/src/core/hle/service/set/set_cal.cpp index 34654bb07..5981c575c 100644 --- a/src/core/hle/service/set/set_cal.cpp +++ b/src/core/hle/service/set/set_cal.cpp @@ -40,7 +40,7 @@ SET_CAL::SET_CAL() : ServiceFramework("set:cal") { {30, nullptr, "GetAmiiboEcqvBlsCertificate"}, {31, nullptr, "GetAmiiboEcqvBlsRootCertificate"}, {32, nullptr, "GetUsbTypeCPowerSourceCircuitVersion"}, - {33, nullptr, "GetBatteryVersion"}, + {41, nullptr, "GetBatteryVersion"}, }; RegisterHandlers(functions); } diff --git a/src/core/hle/service/set/set_cal.h b/src/core/hle/service/set/set_cal.h index 583036eac..a0677e815 100644 --- a/src/core/hle/service/set/set_cal.h +++ b/src/core/hle/service/set/set_cal.h @@ -11,7 +11,7 @@ namespace Service::Set { class SET_CAL final : public ServiceFramework<SET_CAL> { public: explicit SET_CAL(); - ~SET_CAL(); + ~SET_CAL() override; }; } // namespace Service::Set diff --git a/src/core/hle/service/set/set_sys.cpp b/src/core/hle/service/set/set_sys.cpp index ecee554bf..98d0cfdfd 100644 --- a/src/core/hle/service/set/set_sys.cpp +++ b/src/core/hle/service/set/set_sys.cpp @@ -104,6 +104,7 @@ void SET_SYS::SetColorSetId(Kernel::HLERequestContext& ctx) { } SET_SYS::SET_SYS() : ServiceFramework("set:sys") { + // clang-format off static const FunctionInfo functions[] = { {0, nullptr, "SetLanguageCode"}, {1, nullptr, "SetNetworkSettings"}, @@ -252,7 +253,33 @@ SET_SYS::SET_SYS() : ServiceFramework("set:sys") { {147, nullptr, "GetConsoleSixAxisSensorAngularAcceleration"}, {148, nullptr, "SetConsoleSixAxisSensorAngularAcceleration"}, {149, nullptr, "GetRebootlessSystemUpdateVersion"}, + {150, nullptr, "GetDeviceTimeZoneLocationUpdatedTime"}, + {151, nullptr, "SetDeviceTimeZoneLocationUpdatedTime"}, + {152, nullptr, "GetUserSystemClockAutomaticCorrectionUpdatedTime"}, + {153, nullptr, "SetUserSystemClockAutomaticCorrectionUpdatedTime"}, + {154, nullptr, "GetAccountOnlineStorageSettings"}, + {155, nullptr, "SetAccountOnlineStorageSettings"}, + {156, nullptr, "GetPctlReadyFlag"}, + {157, nullptr, "SetPctlReadyFlag"}, + {162, nullptr, "GetPtmBatteryVersion"}, + {163, nullptr, "SetPtmBatteryVersion"}, + {164, nullptr, "GetUsb30HostEnableFlag"}, + {165, nullptr, "SetUsb30HostEnableFlag"}, + {166, nullptr, "GetUsb30DeviceEnableFlag"}, + {167, nullptr, "SetUsb30DeviceEnableFlag"}, + {168, nullptr, "GetThemeId"}, + {169, nullptr, "SetThemeId"}, + {170, nullptr, "GetChineseTraditionalInputMethod"}, + {171, nullptr, "SetChineseTraditionalInputMethod"}, + {172, nullptr, "GetPtmCycleCountReliability"}, + {173, nullptr, "SetPtmCycleCountReliability"}, + {175, nullptr, "GetThemeSettings"}, + {176, nullptr, "SetThemeSettings"}, + {177, nullptr, "GetThemeKey"}, + {178, nullptr, "SetThemeKey"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/sockets/bsd.cpp b/src/core/hle/service/sockets/bsd.cpp index 4342f3b2d..884ad173b 100644 --- a/src/core/hle/service/sockets/bsd.cpp +++ b/src/core/hle/service/sockets/bsd.cpp @@ -73,6 +73,7 @@ void BSD::Close(Kernel::HLERequestContext& ctx) { } BSD::BSD(const char* name) : ServiceFramework(name) { + // clang-format off static const FunctionInfo functions[] = { {0, &BSD::RegisterClient, "RegisterClient"}, {1, &BSD::StartMonitoring, "StartMonitoring"}, @@ -105,7 +106,11 @@ BSD::BSD(const char* name) : ServiceFramework(name) { {28, nullptr, "GetResourceStatistics"}, {29, nullptr, "RecvMMsg"}, {30, nullptr, "SendMMsg"}, + {31, nullptr, "EventFd"}, + {32, nullptr, "RegisterResourceStatisticsName"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/sockets/sfdnsres.cpp b/src/core/hle/service/sockets/sfdnsres.cpp index 13ab1d31e..852e71e4b 100644 --- a/src/core/hle/service/sockets/sfdnsres.cpp +++ b/src/core/hle/service/sockets/sfdnsres.cpp @@ -8,12 +8,20 @@ namespace Service::Sockets { void SFDNSRES::GetAddrInfo(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"); + 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); } diff --git a/src/core/hle/service/spl/module.cpp b/src/core/hle/service/spl/module.cpp index 8db0c2f13..e724d4ab8 100644 --- a/src/core/hle/service/spl/module.cpp +++ b/src/core/hle/service/spl/module.cpp @@ -26,9 +26,7 @@ Module::Interface::~Interface() = default; void Module::Interface::GetRandomBytes(Kernel::HLERequestContext& ctx) { LOG_DEBUG(Service_SPL, "called"); - IPC::RequestParser rp{ctx}; - - std::size_t size = ctx.GetWriteBufferSize(); + const std::size_t size = ctx.GetWriteBufferSize(); std::uniform_int_distribution<u16> distribution(0, std::numeric_limits<u8>::max()); std::vector<u8> data(size); diff --git a/src/core/hle/service/ssl/ssl.cpp b/src/core/hle/service/ssl/ssl.cpp index af40a1815..65040c077 100644 --- a/src/core/hle/service/ssl/ssl.cpp +++ b/src/core/hle/service/ssl/ssl.cpp @@ -64,13 +64,19 @@ public: }; RegisterHandlers(functions); } - ~ISslContext() = default; private: void SetOption(Kernel::HLERequestContext& ctx) { - LOG_WARNING(Service_SSL, "(STUBBED) called"); + struct Parameters { + u8 enable; + u32 option; + }; IPC::RequestParser rp{ctx}; + const auto parameters = rp.PopRaw<Parameters>(); + + LOG_WARNING(Service_SSL, "(STUBBED) called. enable={}, option={}", parameters.enable, + parameters.option); IPC::ResponseBuilder rb{ctx, 2}; rb.Push(RESULT_SUCCESS); @@ -97,6 +103,8 @@ public: {4, nullptr, "DebugIoctl"}, {5, &SSL::SetInterfaceVersion, "SetInterfaceVersion"}, {6, nullptr, "FlushSessionCache"}, + {7, nullptr, "SetDebugOption"}, + {8, nullptr, "GetDebugOption"}, }; // clang-format on diff --git a/src/core/hle/service/time/interface.cpp b/src/core/hle/service/time/interface.cpp index b3a196f65..8d122ae33 100644 --- a/src/core/hle/service/time/interface.cpp +++ b/src/core/hle/service/time/interface.cpp @@ -8,6 +8,7 @@ namespace Service::Time { Time::Time(std::shared_ptr<Module> time, const char* name) : Module::Interface(std::move(time), name) { + // clang-format off static const FunctionInfo functions[] = { {0, &Time::GetStandardUserSystemClock, "GetStandardUserSystemClock"}, {1, &Time::GetStandardNetworkSystemClock, "GetStandardNetworkSystemClock"}, @@ -15,18 +16,23 @@ Time::Time(std::shared_ptr<Module> time, const char* name) {3, &Time::GetTimeZoneService, "GetTimeZoneService"}, {4, &Time::GetStandardLocalSystemClock, "GetStandardLocalSystemClock"}, {5, nullptr, "GetEphemeralNetworkSystemClock"}, + {20, nullptr, "GetSharedMemoryNativeHandle"}, + {30, nullptr, "GetStandardNetworkClockOperationEventReadableHandle"}, + {31, nullptr, "GetEphemeralNetworkClockOperationEventReadableHandle"}, {50, nullptr, "SetStandardSteadyClockInternalOffset"}, {100, nullptr, "IsStandardUserSystemClockAutomaticCorrectionEnabled"}, {101, nullptr, "SetStandardUserSystemClockAutomaticCorrectionEnabled"}, {102, nullptr, "GetStandardUserSystemClockInitialYear"}, {200, nullptr, "IsStandardNetworkSystemClockAccuracySufficient"}, + {201, nullptr, "GetStandardUserSystemClockAutomaticCorrectionUpdatedTime"}, {300, nullptr, "CalculateMonotonicSystemClockBaseTimePoint"}, {400, &Time::GetClockSnapshot, "GetClockSnapshot"}, {401, nullptr, "GetClockSnapshotFromSystemClockContext"}, - {500, &Time::CalculateStandardUserSystemClockDifferenceByUser, - "CalculateStandardUserSystemClockDifferenceByUser"}, + {500, &Time::CalculateStandardUserSystemClockDifferenceByUser, "CalculateStandardUserSystemClockDifferenceByUser"}, {501, nullptr, "CalculateSpanBetween"}, }; + // clang-format on + RegisterHandlers(functions); } diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp index 566cd6006..4e17249a9 100644 --- a/src/core/hle/service/vi/vi.cpp +++ b/src/core/hle/service/vi/vi.cpp @@ -498,7 +498,6 @@ public: }; RegisterHandlers(functions); } - ~IHOSBinderDriver() = default; private: enum class TransactionId { @@ -692,7 +691,6 @@ public: }; RegisterHandlers(functions); } - ~ISystemDisplayService() = default; private: void SetLayerZ(Kernel::HLERequestContext& ctx) { @@ -818,7 +816,6 @@ public: }; RegisterHandlers(functions); } - ~IManagerDisplayService() = default; private: void CloseDisplay(Kernel::HLERequestContext& ctx) { @@ -884,7 +881,6 @@ private: class IApplicationDisplayService final : public ServiceFramework<IApplicationDisplayService> { public: explicit IApplicationDisplayService(std::shared_ptr<NVFlinger::NVFlinger> nv_flinger); - ~IApplicationDisplayService() = default; private: enum class ConvertedScaleMode : u64 { @@ -1037,7 +1033,6 @@ private: void ListDisplays(Kernel::HLERequestContext& ctx) { LOG_WARNING(Service_VI, "(STUBBED) called"); - IPC::RequestParser rp{ctx}; DisplayInfo display_info; display_info.width *= static_cast<u64>(Settings::values.resolution_factor); display_info.height *= static_cast<u64>(Settings::values.resolution_factor); diff --git a/src/core/loader/nso.cpp b/src/core/loader/nso.cpp index babc7e646..d7c47c197 100644 --- a/src/core/loader/nso.cpp +++ b/src/core/loader/nso.cpp @@ -4,11 +4,12 @@ #include <cinttypes> #include <vector> -#include <lz4.h> + #include "common/common_funcs.h" #include "common/file_util.h" #include "common/hex_util.h" #include "common/logging/log.h" +#include "common/lz4_compression.h" #include "common/swap.h" #include "core/core.h" #include "core/file_sys/patch_manager.h" @@ -20,6 +21,8 @@ #include "core/memory.h" #include "core/settings.h" +#pragma optimize("", off) + namespace Loader { namespace { struct MODHeader { @@ -35,15 +38,11 @@ static_assert(sizeof(MODHeader) == 0x1c, "MODHeader has incorrect size."); std::vector<u8> DecompressSegment(const std::vector<u8>& compressed_data, const NSOSegmentHeader& header) { - std::vector<u8> uncompressed_data(header.size); - const int bytes_uncompressed = - LZ4_decompress_safe(reinterpret_cast<const char*>(compressed_data.data()), - reinterpret_cast<char*>(uncompressed_data.data()), - static_cast<int>(compressed_data.size()), header.size); + const std::vector<u8> uncompressed_data = + Common::Compression::DecompressDataLZ4(compressed_data, header.size); - ASSERT_MSG(bytes_uncompressed == static_cast<int>(header.size) && - bytes_uncompressed == static_cast<int>(uncompressed_data.size()), - "{} != {} != {}", bytes_uncompressed, header.size, uncompressed_data.size()); + ASSERT_MSG(uncompressed_data.size() == static_cast<int>(header.size), "{} != {}", header.size, + uncompressed_data.size()); return uncompressed_data; } @@ -139,13 +138,13 @@ std::optional<VAddr> AppLoader_NSO::LoadModule(Kernel::Process& process, // Apply patches if necessary if (pm && (pm->HasNSOPatch(nso_header.build_id) || Settings::values.dump_nso)) { - std::vector<u8> pi_header(sizeof(NSOHeader) + program_image.size()); + std::vector<u8> pi_header; pi_header.insert(pi_header.begin(), reinterpret_cast<u8*>(&nso_header), reinterpret_cast<u8*>(&nso_header) + sizeof(NSOHeader)); pi_header.insert(pi_header.begin() + sizeof(NSOHeader), program_image.begin(), program_image.end()); - pi_header = pm->PatchNSO(pi_header); + pi_header = pm->PatchNSO(pi_header, file.GetName()); std::copy(pi_header.begin() + sizeof(NSOHeader), pi_header.end(), program_image.begin()); } diff --git a/src/core/loader/xci.h b/src/core/loader/xci.h index d6995b61e..436f7387c 100644 --- a/src/core/loader/xci.h +++ b/src/core/loader/xci.h @@ -22,7 +22,7 @@ class AppLoader_NCA; class AppLoader_XCI final : public AppLoader { public: explicit AppLoader_XCI(FileSys::VirtualFile file); - ~AppLoader_XCI(); + ~AppLoader_XCI() override; /** * Returns the type of the file diff --git a/src/core/memory.cpp b/src/core/memory.cpp index 332c1037c..4e0538bc2 100644 --- a/src/core/memory.cpp +++ b/src/core/memory.cpp @@ -38,10 +38,6 @@ void SetCurrentPageTable(Common::PageTable* page_table) { } } -Common::PageTable* GetCurrentPageTable() { - return current_page_table; -} - static void MapPages(Common::PageTable& page_table, VAddr base, u64 size, u8* memory, Common::PageType type) { LOG_DEBUG(HW_Memory, "Mapping {} onto {:016X}-{:016X}", fmt::ptr(memory), base * PAGE_SIZE, diff --git a/src/core/memory.h b/src/core/memory.h index 1d38cdca8..6845f5fe1 100644 --- a/src/core/memory.h +++ b/src/core/memory.h @@ -28,16 +28,6 @@ constexpr u64 PAGE_MASK = PAGE_SIZE - 1; /// Virtual user-space memory regions enum : VAddr { - /// Read-only page containing kernel and system configuration values. - CONFIG_MEMORY_VADDR = 0x1FF80000, - CONFIG_MEMORY_SIZE = 0x00001000, - CONFIG_MEMORY_VADDR_END = CONFIG_MEMORY_VADDR + CONFIG_MEMORY_SIZE, - - /// Usually read-only page containing mostly values read from hardware. - SHARED_PAGE_VADDR = 0x1FF81000, - SHARED_PAGE_SIZE = 0x00001000, - SHARED_PAGE_VADDR_END = SHARED_PAGE_VADDR + SHARED_PAGE_SIZE, - /// TLS (Thread-Local Storage) related. TLS_ENTRY_SIZE = 0x200, @@ -50,9 +40,8 @@ enum : VAddr { KERNEL_REGION_END = KERNEL_REGION_VADDR + KERNEL_REGION_SIZE, }; -/// Currently active page table +/// Changes the currently active page table. void SetCurrentPageTable(Common::PageTable* page_table); -Common::PageTable* GetCurrentPageTable(); /// Determines if the given VAddr is valid for the specified process. bool IsValidVirtualAddress(const Kernel::Process& process, VAddr vaddr); diff --git a/src/core/settings.h b/src/core/settings.h index d543eb32f..b84390745 100644 --- a/src/core/settings.h +++ b/src/core/settings.h @@ -393,6 +393,7 @@ struct Values { bool use_disk_shader_cache; bool use_accurate_gpu_emulation; bool use_asynchronous_gpu_emulation; + bool force_30fps_mode; float bg_red; float bg_green; diff --git a/src/tests/core/arm/arm_test_common.cpp b/src/tests/core/arm/arm_test_common.cpp index 3e1a735c3..58af41f6e 100644 --- a/src/tests/core/arm/arm_test_common.cpp +++ b/src/tests/core/arm/arm_test_common.cpp @@ -17,7 +17,6 @@ TestEnvironment::TestEnvironment(bool mutable_memory_) : mutable_memory(mutable_memory_), test_memory(std::make_shared<TestMemory>(this)), kernel{Core::System::GetInstance()} { auto process = Kernel::Process::Create(Core::System::GetInstance(), ""); - kernel.MakeCurrentProcess(process.get()); page_table = &process->VMManager().page_table; std::fill(page_table->pointers.begin(), page_table->pointers.end(), nullptr); @@ -28,7 +27,7 @@ TestEnvironment::TestEnvironment(bool mutable_memory_) Memory::MapIoRegion(*page_table, 0x00000000, 0x80000000, test_memory); Memory::MapIoRegion(*page_table, 0x80000000, 0x80000000, test_memory); - Memory::SetCurrentPageTable(page_table); + kernel.MakeCurrentProcess(process.get()); } TestEnvironment::~TestEnvironment() { diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt index 44c761d3e..114bed20d 100644 --- a/src/video_core/CMakeLists.txt +++ b/src/video_core/CMakeLists.txt @@ -106,6 +106,8 @@ add_library(video_core STATIC textures/decoders.cpp textures/decoders.h textures/texture.h + texture_cache.cpp + texture_cache.h video_core.cpp video_core.h ) @@ -127,16 +129,21 @@ if (ENABLE_VULKAN) 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_stream_buffer.cpp renderer_vulkan/vk_stream_buffer.h renderer_vulkan/vk_swapchain.cpp renderer_vulkan/vk_swapchain.h) - target_include_directories(video_core PRIVATE ../../externals/Vulkan-Headers/include) + target_include_directories(video_core PRIVATE sirit ../../externals/Vulkan-Headers/include) target_compile_definitions(video_core PRIVATE HAS_VULKAN) endif() create_target_directory_groups(video_core) target_link_libraries(video_core PUBLIC common core) -target_link_libraries(video_core PRIVATE glad lz4_static) +target_link_libraries(video_core PRIVATE glad) +if (ENABLE_VULKAN) + target_link_libraries(video_core PRIVATE sirit) +endif() diff --git a/src/video_core/dma_pusher.cpp b/src/video_core/dma_pusher.cpp index 8b1bea1ae..046d047cb 100644 --- a/src/video_core/dma_pusher.cpp +++ b/src/video_core/dma_pusher.cpp @@ -8,6 +8,7 @@ #include "video_core/dma_pusher.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/gpu.h" +#include "video_core/memory_manager.h" namespace Tegra { diff --git a/src/video_core/engines/fermi_2d.cpp b/src/video_core/engines/fermi_2d.cpp index 03b7ee5d8..55966eef1 100644 --- a/src/video_core/engines/fermi_2d.cpp +++ b/src/video_core/engines/fermi_2d.cpp @@ -6,12 +6,13 @@ #include "common/logging/log.h" #include "common/math_util.h" #include "video_core/engines/fermi_2d.h" +#include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" namespace Tegra::Engines { Fermi2D::Fermi2D(VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager) - : memory_manager(memory_manager), rasterizer{rasterizer} {} + : rasterizer{rasterizer}, memory_manager{memory_manager} {} void Fermi2D::CallMethod(const GPU::MethodCall& method_call) { ASSERT_MSG(method_call.method < Regs::NUM_REGS, diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h index 80523e320..2e51b7f13 100644 --- a/src/video_core/engines/fermi_2d.h +++ b/src/video_core/engines/fermi_2d.h @@ -10,7 +10,10 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/gpu.h" -#include "video_core/memory_manager.h" + +namespace Tegra { +class MemoryManager; +} namespace VideoCore { class RasterizerInterface; @@ -115,10 +118,9 @@ public: }; } regs{}; - MemoryManager& memory_manager; - private: VideoCore::RasterizerInterface& rasterizer; + MemoryManager& memory_manager; /// Performs the copy from the source surface to the destination surface as configured in the /// registers. diff --git a/src/video_core/engines/kepler_compute.h b/src/video_core/engines/kepler_compute.h index 6575afd0f..fb6cdf432 100644 --- a/src/video_core/engines/kepler_compute.h +++ b/src/video_core/engines/kepler_compute.h @@ -9,7 +9,10 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/gpu.h" -#include "video_core/memory_manager.h" + +namespace Tegra { +class MemoryManager; +} namespace Tegra::Engines { @@ -40,10 +43,11 @@ public: static_assert(sizeof(Regs) == Regs::NUM_REGS * sizeof(u32), "KeplerCompute Regs has wrong size"); - MemoryManager& memory_manager; - /// Write the value to the register identified by method. void CallMethod(const GPU::MethodCall& method_call); + +private: + MemoryManager& memory_manager; }; #define ASSERT_REG_POSITION(field_name, position) \ diff --git a/src/video_core/engines/kepler_memory.cpp b/src/video_core/engines/kepler_memory.cpp index e259bf46b..cd51a31d7 100644 --- a/src/video_core/engines/kepler_memory.cpp +++ b/src/video_core/engines/kepler_memory.cpp @@ -5,9 +5,9 @@ #include "common/assert.h" #include "common/logging/log.h" #include "core/core.h" -#include "core/memory.h" #include "video_core/engines/kepler_memory.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_base.h" @@ -15,7 +15,7 @@ namespace Tegra::Engines { KeplerMemory::KeplerMemory(Core::System& system, VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager) - : system{system}, memory_manager(memory_manager), rasterizer{rasterizer} {} + : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager} {} KeplerMemory::~KeplerMemory() = default; diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h index 9181e9d80..78b6c3e45 100644 --- a/src/video_core/engines/kepler_memory.h +++ b/src/video_core/engines/kepler_memory.h @@ -10,12 +10,15 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/gpu.h" -#include "video_core/memory_manager.h" namespace Core { class System; } +namespace Tegra { +class MemoryManager; +} + namespace VideoCore { class RasterizerInterface; } @@ -82,8 +85,8 @@ public: private: Core::System& system; - MemoryManager& memory_manager; VideoCore::RasterizerInterface& rasterizer; + MemoryManager& memory_manager; void ProcessData(u32 data); }; diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index defcfbd3f..74403eed4 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -7,11 +7,10 @@ #include "common/assert.h" #include "core/core.h" #include "core/core_timing.h" -#include "core/memory.h" #include "video_core/debug_utils/debug_utils.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" -#include "video_core/renderer_base.h" #include "video_core/textures/texture.h" namespace Tegra::Engines { @@ -21,8 +20,8 @@ constexpr u32 MacroRegistersStart = 0xE00; Maxwell3D::Maxwell3D(Core::System& system, VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager) - : memory_manager(memory_manager), system{system}, rasterizer{rasterizer}, - macro_interpreter(*this) { + : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager}, macro_interpreter{ + *this} { InitializeRegisterDefaults(); } @@ -250,6 +249,10 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) { ProcessQueryGet(); break; } + case MAXWELL3D_REG_INDEX(sync_info): { + ProcessSyncPoint(); + break; + } default: break; } @@ -327,6 +330,14 @@ void Maxwell3D::ProcessQueryGet() { } } +void Maxwell3D::ProcessSyncPoint() { + const u32 sync_point = regs.sync_info.sync_point.Value(); + const u32 increment = regs.sync_info.increment.Value(); + const u32 cache_flush = regs.sync_info.unknown.Value(); + LOG_DEBUG(HW_GPU, "Syncpoint set {}, increment: {}, unk: {}", sync_point, increment, + cache_flush); +} + void Maxwell3D::DrawArrays() { LOG_DEBUG(HW_GPU, "called, topology={}, count={}", static_cast<u32>(regs.draw.topology.Value()), regs.vertex_buffer.count); diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 7fbf1026e..321af3297 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -16,13 +16,16 @@ #include "common/math_util.h" #include "video_core/gpu.h" #include "video_core/macro_interpreter.h" -#include "video_core/memory_manager.h" #include "video_core/textures/texture.h" namespace Core { class System; } +namespace Tegra { +class MemoryManager; +} + namespace VideoCore { class RasterizerInterface; } @@ -576,7 +579,17 @@ public: u32 bind; } macros; - INSERT_PADDING_WORDS(0x188); + INSERT_PADDING_WORDS(0x69); + + struct { + union { + BitField<0, 16, u32> sync_point; + BitField<16, 1, u32> unknown; + BitField<20, 1, u32> increment; + }; + } sync_info; + + INSERT_PADDING_WORDS(0x11E); u32 tfb_enabled; @@ -1093,7 +1106,6 @@ public: }; State state{}; - MemoryManager& memory_manager; struct DirtyFlags { std::bitset<8> color_buffer{0xFF}; @@ -1141,6 +1153,8 @@ private: VideoCore::RasterizerInterface& rasterizer; + MemoryManager& memory_manager; + /// Start offsets of each macro in macro_memory std::unordered_map<u32, u32> macro_offsets; @@ -1180,6 +1194,9 @@ private: /// Handles a write to the QUERY_GET register. void ProcessQueryGet(); + /// Handles writes to syncing register. + void ProcessSyncPoint(); + /// Handles a write to the CB_DATA[i] register. void ProcessCBData(u32 value); @@ -1195,6 +1212,7 @@ private: "Field " #field_name " has invalid position") ASSERT_REG_POSITION(macros, 0x45); +ASSERT_REG_POSITION(sync_info, 0xB2); ASSERT_REG_POSITION(tfb_enabled, 0x1D1); ASSERT_REG_POSITION(rt, 0x200); ASSERT_REG_POSITION(viewport_transform, 0x280); diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp index 5cca5c29a..2426d0067 100644 --- a/src/video_core/engines/maxwell_dma.cpp +++ b/src/video_core/engines/maxwell_dma.cpp @@ -5,9 +5,9 @@ #include "common/assert.h" #include "common/logging/log.h" #include "core/core.h" -#include "core/memory.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/engines/maxwell_dma.h" +#include "video_core/memory_manager.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_base.h" #include "video_core/textures/decoders.h" @@ -16,7 +16,7 @@ namespace Tegra::Engines { MaxwellDMA::MaxwellDMA(Core::System& system, VideoCore::RasterizerInterface& rasterizer, MemoryManager& memory_manager) - : memory_manager(memory_manager), system{system}, rasterizer{rasterizer} {} + : system{system}, rasterizer{rasterizer}, memory_manager{memory_manager} {} void MaxwellDMA::CallMethod(const GPU::MethodCall& method_call) { ASSERT_MSG(method_call.method < Regs::NUM_REGS, diff --git a/src/video_core/engines/maxwell_dma.h b/src/video_core/engines/maxwell_dma.h index 34c369320..c6b649842 100644 --- a/src/video_core/engines/maxwell_dma.h +++ b/src/video_core/engines/maxwell_dma.h @@ -10,12 +10,15 @@ #include "common/common_funcs.h" #include "common/common_types.h" #include "video_core/gpu.h" -#include "video_core/memory_manager.h" namespace Core { class System; } +namespace Tegra { +class MemoryManager; +} + namespace VideoCore { class RasterizerInterface; } @@ -139,13 +142,13 @@ public: }; } regs{}; - MemoryManager& memory_manager; - private: Core::System& system; VideoCore::RasterizerInterface& rasterizer; + MemoryManager& memory_manager; + /// Performs the copy from the source buffer to the destination buffer as configured in the /// registers. void HandleCopy(); diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h index c41e3f4f0..38db4addd 100644 --- a/src/video_core/engines/shader_bytecode.h +++ b/src/video_core/engines/shader_bytecode.h @@ -806,6 +806,12 @@ union Instruction { } ldg; union { + BitField<48, 3, UniformType> type; + BitField<46, 2, u64> cache_mode; + BitField<20, 24, s64> immediate_offset; + } stg; + + union { BitField<0, 3, u64> pred0; BitField<3, 3, u64> pred3; BitField<7, 1, u64> abs_a; @@ -1252,13 +1258,16 @@ union Instruction { union { BitField<20, 16, u64> imm20_16; + BitField<35, 1, u64> high_b_rr; // used on RR BitField<36, 1, u64> product_shift_left; BitField<37, 1, u64> merge_37; BitField<48, 1, u64> sign_a; BitField<49, 1, u64> sign_b; + BitField<50, 2, XmadMode> mode_cbf; // used by CR, RC BitField<50, 3, XmadMode> mode; BitField<52, 1, u64> high_b; BitField<53, 1, u64> high_a; + BitField<55, 1, u64> product_shift_left_second; // used on CR BitField<56, 1, u64> merge_56; } xmad; @@ -1676,7 +1685,7 @@ private: INST("0011011-11110---", Id::BFI_IMM_R, Type::Bfi, "BFI_IMM_R"), INST("0100110001000---", Id::LOP_C, Type::ArithmeticInteger, "LOP_C"), INST("0101110001000---", Id::LOP_R, Type::ArithmeticInteger, "LOP_R"), - INST("0011100001000---", Id::LOP_IMM, Type::ArithmeticInteger, "LOP_IMM"), + INST("0011100-01000---", Id::LOP_IMM, Type::ArithmeticInteger, "LOP_IMM"), INST("000001----------", Id::LOP32I, Type::ArithmeticIntegerImmediate, "LOP32I"), INST("0000001---------", Id::LOP3_C, Type::ArithmeticInteger, "LOP3_C"), INST("0101101111100---", Id::LOP3_R, Type::ArithmeticInteger, "LOP3_R"), diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp index 30b29e14d..4461083ff 100644 --- a/src/video_core/gpu.cpp +++ b/src/video_core/gpu.cpp @@ -31,7 +31,7 @@ u32 FramebufferConfig::BytesPerPixel(PixelFormat format) { GPU::GPU(Core::System& system, VideoCore::RendererBase& renderer) : renderer{renderer} { auto& rasterizer{renderer.Rasterizer()}; - memory_manager = std::make_unique<Tegra::MemoryManager>(); + memory_manager = std::make_unique<Tegra::MemoryManager>(rasterizer); dma_pusher = std::make_unique<Tegra::DmaPusher>(*this); maxwell_3d = std::make_unique<Engines::Maxwell3D>(system, rasterizer, *memory_manager); fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer, *memory_manager); diff --git a/src/video_core/gpu_asynch.cpp b/src/video_core/gpu_asynch.cpp index 8b355cf7b..db507cf04 100644 --- a/src/video_core/gpu_asynch.cpp +++ b/src/video_core/gpu_asynch.cpp @@ -9,7 +9,7 @@ namespace VideoCommon { GPUAsynch::GPUAsynch(Core::System& system, VideoCore::RendererBase& renderer) - : Tegra::GPU(system, renderer), gpu_thread{renderer, *dma_pusher} {} + : Tegra::GPU(system, renderer), gpu_thread{system, renderer, *dma_pusher} {} GPUAsynch::~GPUAsynch() = default; diff --git a/src/video_core/gpu_thread.cpp b/src/video_core/gpu_thread.cpp index c5dc199c5..cc56cf467 100644 --- a/src/video_core/gpu_thread.cpp +++ b/src/video_core/gpu_thread.cpp @@ -4,6 +4,9 @@ #include "common/assert.h" #include "common/microprofile.h" +#include "core/core.h" +#include "core/core_timing.h" +#include "core/core_timing_util.h" #include "core/frontend/scope_acquire_window_context.h" #include "video_core/dma_pusher.h" #include "video_core/gpu.h" @@ -36,7 +39,6 @@ static void RunThread(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_p dma_pusher.Push(std::move(submit_list->entries)); dma_pusher.DispatchCalls(); } else if (const auto data = std::get_if<SwapBuffersCommand>(&next.data)) { - state.DecrementFramesCounter(); renderer.SwapBuffers(std::move(data->framebuffer)); } else if (const auto data = std::get_if<FlushRegionCommand>(&next.data)) { renderer.Rasterizer().FlushRegion(data->addr, data->size); @@ -47,13 +49,18 @@ static void RunThread(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_p } else { UNREACHABLE(); } + state.signaled_fence = next.fence; + state.TrySynchronize(); } } } -ThreadManager::ThreadManager(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_pusher) - : renderer{renderer}, thread{RunThread, std::ref(renderer), std::ref(dma_pusher), - std::ref(state)} {} +ThreadManager::ThreadManager(Core::System& system, VideoCore::RendererBase& renderer, + Tegra::DmaPusher& dma_pusher) + : system{system}, thread{RunThread, std::ref(renderer), std::ref(dma_pusher), std::ref(state)} { + synchronization_event = system.CoreTiming().RegisterEvent( + "GPUThreadSynch", [this](u64 fence, s64) { state.WaitForSynchronization(fence); }); +} ThreadManager::~ThreadManager() { // Notify GPU thread that a shutdown is pending @@ -62,14 +69,14 @@ ThreadManager::~ThreadManager() { } void ThreadManager::SubmitList(Tegra::CommandList&& entries) { - PushCommand(SubmitListCommand(std::move(entries))); + const u64 fence{PushCommand(SubmitListCommand(std::move(entries)))}; + const s64 synchronization_ticks{Core::Timing::usToCycles(9000)}; + system.CoreTiming().ScheduleEvent(synchronization_ticks, synchronization_event, fence); } void ThreadManager::SwapBuffers( std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) { - state.IncrementFramesCounter(); PushCommand(SwapBuffersCommand(std::move(framebuffer))); - state.WaitForFrames(); } void ThreadManager::FlushRegion(CacheAddr addr, u64 size) { @@ -79,7 +86,7 @@ void ThreadManager::FlushRegion(CacheAddr addr, u64 size) { void ThreadManager::InvalidateRegion(CacheAddr addr, u64 size) { if (state.queue.Empty()) { // It's quicker to invalidate a single region on the CPU if the queue is already empty - renderer.Rasterizer().InvalidateRegion(addr, size); + system.Renderer().Rasterizer().InvalidateRegion(addr, size); } else { PushCommand(InvalidateRegionCommand(addr, size)); } @@ -90,9 +97,25 @@ void ThreadManager::FlushAndInvalidateRegion(CacheAddr addr, u64 size) { InvalidateRegion(addr, size); } -void ThreadManager::PushCommand(CommandData&& command_data) { - state.queue.Push(CommandDataContainer(std::move(command_data))); +u64 ThreadManager::PushCommand(CommandData&& command_data) { + const u64 fence{++state.last_fence}; + state.queue.Push(CommandDataContainer(std::move(command_data), fence)); state.SignalCommands(); + return fence; +} + +MICROPROFILE_DEFINE(GPU_wait, "GPU", "Wait for the GPU", MP_RGB(128, 128, 192)); +void SynchState::WaitForSynchronization(u64 fence) { + if (signaled_fence >= fence) { + return; + } + + // Wait for the GPU to be idle (all commands to be executed) + { + MICROPROFILE_SCOPE(GPU_wait); + std::unique_lock<std::mutex> lock{synchronization_mutex}; + synchronization_condition.wait(lock, [this, fence] { return signaled_fence >= fence; }); + } } } // namespace VideoCommon::GPUThread diff --git a/src/video_core/gpu_thread.h b/src/video_core/gpu_thread.h index 70acb2e79..62bcea5bb 100644 --- a/src/video_core/gpu_thread.h +++ b/src/video_core/gpu_thread.h @@ -19,9 +19,12 @@ struct FramebufferConfig; class DmaPusher; } // namespace Tegra -namespace VideoCore { -class RendererBase; -} // namespace VideoCore +namespace Core { +class System; +namespace Timing { +struct EventType; +} // namespace Timing +} // namespace Core namespace VideoCommon::GPUThread { @@ -75,63 +78,47 @@ using CommandData = struct CommandDataContainer { CommandDataContainer() = default; - CommandDataContainer(CommandData&& data) : data{std::move(data)} {} + CommandDataContainer(CommandData&& data, u64 next_fence) + : data{std::move(data)}, fence{next_fence} {} CommandDataContainer& operator=(const CommandDataContainer& t) { data = std::move(t.data); + fence = t.fence; return *this; } CommandData data; + u64 fence{}; }; /// Struct used to synchronize the GPU thread struct SynchState final { std::atomic_bool is_running{true}; std::atomic_int queued_frame_count{}; - std::mutex frames_mutex; + std::mutex synchronization_mutex; std::mutex commands_mutex; std::condition_variable commands_condition; - std::condition_variable frames_condition; + std::condition_variable synchronization_condition; - void IncrementFramesCounter() { - std::lock_guard lock{frames_mutex}; - ++queued_frame_count; + /// Returns true if the gap in GPU commands is small enough that we can consider the CPU and GPU + /// synchronized. This is entirely empirical. + bool IsSynchronized() const { + constexpr std::size_t max_queue_gap{5}; + return queue.Size() <= max_queue_gap; } - void DecrementFramesCounter() { - { - std::lock_guard lock{frames_mutex}; - --queued_frame_count; - - if (queued_frame_count) { - return; - } + void TrySynchronize() { + if (IsSynchronized()) { + std::lock_guard<std::mutex> lock{synchronization_mutex}; + synchronization_condition.notify_one(); } - frames_condition.notify_one(); } - void WaitForFrames() { - { - std::lock_guard lock{frames_mutex}; - if (!queued_frame_count) { - return; - } - } - - // Wait for the GPU to be idle (all commands to be executed) - { - std::unique_lock lock{frames_mutex}; - frames_condition.wait(lock, [this] { return !queued_frame_count; }); - } - } + void WaitForSynchronization(u64 fence); void SignalCommands() { - { - std::unique_lock lock{commands_mutex}; - if (queue.Empty()) { - return; - } + if (queue.Empty()) { + return; } commands_condition.notify_one(); @@ -144,12 +131,15 @@ struct SynchState final { using CommandQueue = Common::SPSCQueue<CommandDataContainer>; CommandQueue queue; + u64 last_fence{}; + std::atomic<u64> signaled_fence{}; }; /// Class used to manage the GPU thread class ThreadManager final { public: - explicit ThreadManager(VideoCore::RendererBase& renderer, Tegra::DmaPusher& dma_pusher); + explicit ThreadManager(Core::System& system, VideoCore::RendererBase& renderer, + Tegra::DmaPusher& dma_pusher); ~ThreadManager(); /// Push GPU command entries to be processed @@ -170,11 +160,12 @@ public: private: /// Pushes a command to be executed by the GPU thread - void PushCommand(CommandData&& command_data); + u64 PushCommand(CommandData&& command_data); private: SynchState state; - VideoCore::RendererBase& renderer; + Core::System& system; + Core::Timing::EventType* synchronization_event{}; std::thread thread; std::thread::id thread_id; }; diff --git a/src/video_core/macro_interpreter.cpp b/src/video_core/macro_interpreter.cpp index 64f75db43..524d9ea5a 100644 --- a/src/video_core/macro_interpreter.cpp +++ b/src/video_core/macro_interpreter.cpp @@ -223,27 +223,21 @@ void MacroInterpreter::ProcessResult(ResultOperation operation, u32 reg, u32 res } u32 MacroInterpreter::FetchParameter() { - ASSERT(next_parameter_index < parameters.size()); - return parameters[next_parameter_index++]; + return parameters.at(next_parameter_index++); } u32 MacroInterpreter::GetRegister(u32 register_id) const { - // Register 0 is supposed to always return 0. - if (register_id == 0) - return 0; - - ASSERT(register_id < registers.size()); - return registers[register_id]; + return registers.at(register_id); } void MacroInterpreter::SetRegister(u32 register_id, u32 value) { - // Register 0 is supposed to always return 0. NOP is implemented as a store to the zero - // register. - if (register_id == 0) + // Register 0 is hardwired as the zero register. + // Ensure no writes to it actually occur. + if (register_id == 0) { return; + } - ASSERT(register_id < registers.size()); - registers[register_id] = value; + registers.at(register_id) = value; } void MacroInterpreter::SetMethodAddress(u32 address) { diff --git a/src/video_core/memory_manager.cpp b/src/video_core/memory_manager.cpp index e76b59842..0f4e820aa 100644 --- a/src/video_core/memory_manager.cpp +++ b/src/video_core/memory_manager.cpp @@ -5,16 +5,13 @@ #include "common/alignment.h" #include "common/assert.h" #include "common/logging/log.h" -#include "core/core.h" #include "core/memory.h" -#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 { -MemoryManager::MemoryManager() { +MemoryManager::MemoryManager(VideoCore::RasterizerInterface& rasterizer) : rasterizer{rasterizer} { std::fill(page_table.pointers.begin(), page_table.pointers.end(), nullptr); std::fill(page_table.attributes.begin(), page_table.attributes.end(), Common::PageType::Unmapped); @@ -70,23 +67,23 @@ GPUVAddr MemoryManager::UnmapBuffer(GPUVAddr gpu_addr, u64 size) { const u64 aligned_size{Common::AlignUp(size, page_size)}; const CacheAddr cache_addr{ToCacheAddr(GetPointer(gpu_addr))}; - Core::System::GetInstance().Renderer().Rasterizer().FlushAndInvalidateRegion(cache_addr, - aligned_size); + rasterizer.FlushAndInvalidateRegion(cache_addr, aligned_size); UnmapRange(gpu_addr, aligned_size); return gpu_addr; } -GPUVAddr MemoryManager::FindFreeRegion(GPUVAddr region_start, u64 size) { +GPUVAddr MemoryManager::FindFreeRegion(GPUVAddr region_start, u64 size) const { // Find the first Free VMA. - const VMAHandle vma_handle{std::find_if(vma_map.begin(), vma_map.end(), [&](const auto& vma) { - if (vma.second.type != VirtualMemoryArea::Type::Unmapped) { - return false; - } + const VMAHandle vma_handle{ + std::find_if(vma_map.begin(), vma_map.end(), [region_start, size](const auto& vma) { + if (vma.second.type != VirtualMemoryArea::Type::Unmapped) { + return false; + } - const VAddr vma_end{vma.second.base + vma.second.size}; - return vma_end > region_start && vma_end >= region_start + size; - })}; + const VAddr vma_end{vma.second.base + vma.second.size}; + return vma_end > region_start && vma_end >= region_start + size; + })}; if (vma_handle == vma_map.end()) { return {}; @@ -99,12 +96,12 @@ bool MemoryManager::IsAddressValid(GPUVAddr addr) const { return (addr >> page_bits) < page_table.pointers.size(); } -std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) { +std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) const { if (!IsAddressValid(addr)) { return {}; } - VAddr cpu_addr{page_table.backing_addr[addr >> page_bits]}; + const VAddr cpu_addr{page_table.backing_addr[addr >> page_bits]}; if (cpu_addr) { return cpu_addr + (addr & page_mask); } @@ -113,7 +110,7 @@ std::optional<VAddr> MemoryManager::GpuToCpuAddress(GPUVAddr addr) { } template <typename T> -T MemoryManager::Read(GPUVAddr addr) { +T MemoryManager::Read(GPUVAddr addr) const { if (!IsAddressValid(addr)) { return {}; } @@ -165,10 +162,10 @@ void MemoryManager::Write(GPUVAddr addr, T data) { } } -template u8 MemoryManager::Read<u8>(GPUVAddr addr); -template u16 MemoryManager::Read<u16>(GPUVAddr addr); -template u32 MemoryManager::Read<u32>(GPUVAddr addr); -template u64 MemoryManager::Read<u64>(GPUVAddr addr); +template u8 MemoryManager::Read<u8>(GPUVAddr addr) const; +template u16 MemoryManager::Read<u16>(GPUVAddr addr) const; +template u32 MemoryManager::Read<u32>(GPUVAddr addr) const; +template u64 MemoryManager::Read<u64>(GPUVAddr addr) const; template void MemoryManager::Write<u8>(GPUVAddr addr, u8 data); template void MemoryManager::Write<u16>(GPUVAddr addr, u16 data); template void MemoryManager::Write<u32>(GPUVAddr addr, u32 data); @@ -179,8 +176,22 @@ u8* MemoryManager::GetPointer(GPUVAddr addr) { return {}; } - u8* page_pointer{page_table.pointers[addr >> page_bits]}; - if (page_pointer) { + u8* const page_pointer{page_table.pointers[addr >> page_bits]}; + if (page_pointer != nullptr) { + return page_pointer + (addr & page_mask); + } + + LOG_ERROR(HW_GPU, "Unknown GetPointer @ 0x{:016X}", addr); + return {}; +} + +const u8* MemoryManager::GetPointer(GPUVAddr addr) const { + if (!IsAddressValid(addr)) { + return {}; + } + + const u8* const page_pointer{page_table.pointers[addr >> page_bits]}; + if (page_pointer != nullptr) { return page_pointer + (addr & page_mask); } @@ -188,15 +199,86 @@ u8* MemoryManager::GetPointer(GPUVAddr addr) { return {}; } -void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) { - std::memcpy(dest_buffer, GetPointer(src_addr), size); +void MemoryManager::ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const { + std::size_t remaining_size{size}; + std::size_t page_index{src_addr >> page_bits}; + std::size_t page_offset{src_addr & page_mask}; + + while (remaining_size > 0) { + const std::size_t copy_amount{ + std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)}; + + switch (page_table.attributes[page_index]) { + case Common::PageType::Memory: { + const u8* src_ptr{page_table.pointers[page_index] + page_offset}; + rasterizer.FlushRegion(ToCacheAddr(src_ptr), copy_amount); + std::memcpy(dest_buffer, src_ptr, copy_amount); + break; + } + default: + UNREACHABLE(); + } + + page_index++; + page_offset = 0; + dest_buffer = static_cast<u8*>(dest_buffer) + copy_amount; + remaining_size -= copy_amount; + } } + void MemoryManager::WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size) { - std::memcpy(GetPointer(dest_addr), src_buffer, size); + std::size_t remaining_size{size}; + std::size_t page_index{dest_addr >> page_bits}; + std::size_t page_offset{dest_addr & page_mask}; + + while (remaining_size > 0) { + const std::size_t copy_amount{ + std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)}; + + switch (page_table.attributes[page_index]) { + case Common::PageType::Memory: { + u8* dest_ptr{page_table.pointers[page_index] + page_offset}; + rasterizer.InvalidateRegion(ToCacheAddr(dest_ptr), copy_amount); + std::memcpy(dest_ptr, src_buffer, copy_amount); + break; + } + default: + UNREACHABLE(); + } + + page_index++; + page_offset = 0; + src_buffer = static_cast<const u8*>(src_buffer) + copy_amount; + remaining_size -= copy_amount; + } } void MemoryManager::CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size) { - std::memcpy(GetPointer(dest_addr), GetPointer(src_addr), size); + std::size_t remaining_size{size}; + std::size_t page_index{src_addr >> page_bits}; + std::size_t page_offset{src_addr & page_mask}; + + while (remaining_size > 0) { + const std::size_t copy_amount{ + std::min(static_cast<std::size_t>(page_size) - page_offset, remaining_size)}; + + switch (page_table.attributes[page_index]) { + case Common::PageType::Memory: { + const u8* src_ptr{page_table.pointers[page_index] + page_offset}; + rasterizer.FlushRegion(ToCacheAddr(src_ptr), copy_amount); + WriteBlock(dest_addr, src_ptr, copy_amount); + break; + } + default: + UNREACHABLE(); + } + + page_index++; + page_offset = 0; + dest_addr += static_cast<VAddr>(copy_amount); + src_addr += static_cast<VAddr>(copy_amount); + remaining_size -= copy_amount; + } } void MemoryManager::MapPages(GPUVAddr base, u64 size, u8* memory, Common::PageType type, @@ -336,7 +418,7 @@ MemoryManager::VMAIter MemoryManager::CarveVMA(GPUVAddr base, u64 size) { const VirtualMemoryArea& vma{vma_handle->second}; if (vma.type == VirtualMemoryArea::Type::Mapped) { // Region is already allocated - return {}; + return vma_handle; } const VAddr start_in_vma{base - vma.base}; diff --git a/src/video_core/memory_manager.h b/src/video_core/memory_manager.h index 34744bb27..647cbf93a 100644 --- a/src/video_core/memory_manager.h +++ b/src/video_core/memory_manager.h @@ -10,6 +10,10 @@ #include "common/common_types.h" #include "common/page_table.h" +namespace VideoCore { +class RasterizerInterface; +} + namespace Tegra { /** @@ -43,24 +47,25 @@ struct VirtualMemoryArea { class MemoryManager final { public: - MemoryManager(); + MemoryManager(VideoCore::RasterizerInterface& rasterizer); GPUVAddr AllocateSpace(u64 size, u64 align); GPUVAddr AllocateSpace(GPUVAddr addr, u64 size, u64 align); GPUVAddr MapBufferEx(VAddr cpu_addr, u64 size); GPUVAddr MapBufferEx(VAddr cpu_addr, GPUVAddr addr, u64 size); GPUVAddr UnmapBuffer(GPUVAddr addr, u64 size); - std::optional<VAddr> GpuToCpuAddress(GPUVAddr addr); + std::optional<VAddr> GpuToCpuAddress(GPUVAddr addr) const; template <typename T> - T Read(GPUVAddr addr); + T Read(GPUVAddr addr) const; template <typename T> void Write(GPUVAddr addr, T data); u8* GetPointer(GPUVAddr addr); + const u8* GetPointer(GPUVAddr addr) const; - void ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size); + void ReadBlock(GPUVAddr src_addr, void* dest_buffer, std::size_t size) const; void WriteBlock(GPUVAddr dest_addr, const void* src_buffer, std::size_t size); void CopyBlock(GPUVAddr dest_addr, GPUVAddr src_addr, std::size_t size); @@ -127,7 +132,7 @@ private: void UpdatePageTableForVMA(const VirtualMemoryArea& vma); /// Finds a free (unmapped region) of the specified size starting at the specified address. - GPUVAddr FindFreeRegion(GPUVAddr region_start, u64 size); + GPUVAddr FindFreeRegion(GPUVAddr region_start, u64 size) const; private: static constexpr u64 page_bits{16}; @@ -143,6 +148,7 @@ private: Common::PageTable page_table{page_bits}; VMAMap vma_map; + VideoCore::RasterizerInterface& rasterizer; }; } // namespace Tegra diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp index fd091c84c..25652e794 100644 --- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp @@ -7,7 +7,7 @@ #include "common/alignment.h" #include "core/core.h" -#include "core/memory.h" +#include "video_core/memory_manager.h" #include "video_core/renderer_opengl/gl_buffer_cache.h" #include "video_core/renderer_opengl/gl_rasterizer.h" diff --git a/src/video_core/renderer_opengl/gl_global_cache.cpp b/src/video_core/renderer_opengl/gl_global_cache.cpp index da9326253..ea4a593af 100644 --- a/src/video_core/renderer_opengl/gl_global_cache.cpp +++ b/src/video_core/renderer_opengl/gl_global_cache.cpp @@ -4,9 +4,9 @@ #include <glad/glad.h> -#include "common/assert.h" #include "common/logging/log.h" #include "core/core.h" +#include "video_core/memory_manager.h" #include "video_core/renderer_opengl/gl_global_cache.h" #include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_shader_decompiler.h" @@ -14,28 +14,28 @@ namespace OpenGL { -CachedGlobalRegion::CachedGlobalRegion(VAddr cpu_addr, u32 size, u8* host_ptr) - : RasterizerCacheObject{host_ptr}, cpu_addr{cpu_addr}, size{size} { +CachedGlobalRegion::CachedGlobalRegion(VAddr cpu_addr, u8* host_ptr, u32 size, u32 max_size) + : RasterizerCacheObject{host_ptr}, cpu_addr{cpu_addr}, host_ptr{host_ptr}, size{size}, + max_size{max_size} { buffer.Create(); - // Bind and unbind the buffer so it gets allocated by the driver - glBindBuffer(GL_SHADER_STORAGE_BUFFER, buffer.handle); - glBindBuffer(GL_SHADER_STORAGE_BUFFER, 0); LabelGLObject(GL_BUFFER, buffer.handle, cpu_addr, "GlobalMemory"); } -void CachedGlobalRegion::Reload(u32 size_) { - constexpr auto max_size = static_cast<u32>(RasterizerOpenGL::MaxGlobalMemorySize); +CachedGlobalRegion::~CachedGlobalRegion() = default; +void CachedGlobalRegion::Reload(u32 size_) { size = size_; if (size > max_size) { size = max_size; - LOG_CRITICAL(HW_GPU, "Global region size {} exceeded the expected size {}!", size_, + LOG_CRITICAL(HW_GPU, "Global region size {} exceeded the supported size {}!", size_, max_size); } + glNamedBufferData(buffer.handle, size, host_ptr, GL_STREAM_DRAW); +} - // TODO(Rodrigo): Get rid of Memory::GetPointer with a staging buffer - glBindBuffer(GL_SHADER_STORAGE_BUFFER, buffer.handle); - glBufferData(GL_SHADER_STORAGE_BUFFER, size, GetHostPtr(), GL_DYNAMIC_DRAW); +void CachedGlobalRegion::Flush() { + LOG_DEBUG(Render_OpenGL, "Flushing {} bytes to CPU memory address 0x{:16}", size, cpu_addr); + glGetNamedBufferSubData(buffer.handle, 0, static_cast<GLsizeiptr>(size), host_ptr); } GlobalRegion GlobalRegionCacheOpenGL::TryGetReservedGlobalRegion(CacheAddr addr, u32 size) const { @@ -46,14 +46,16 @@ GlobalRegion GlobalRegionCacheOpenGL::TryGetReservedGlobalRegion(CacheAddr addr, return search->second; } -GlobalRegion GlobalRegionCacheOpenGL::GetUncachedGlobalRegion(GPUVAddr addr, u32 size, - u8* host_ptr) { +GlobalRegion GlobalRegionCacheOpenGL::GetUncachedGlobalRegion(GPUVAddr addr, u8* host_ptr, + u32 size) { GlobalRegion region{TryGetReservedGlobalRegion(ToCacheAddr(host_ptr), size)}; if (!region) { // No reserved surface available, create a new one and reserve it auto& memory_manager{Core::System::GetInstance().GPU().MemoryManager()}; - const auto cpu_addr = *memory_manager.GpuToCpuAddress(addr); - region = std::make_shared<CachedGlobalRegion>(cpu_addr, size, host_ptr); + const auto cpu_addr{memory_manager.GpuToCpuAddress(addr)}; + ASSERT(cpu_addr); + + region = std::make_shared<CachedGlobalRegion>(*cpu_addr, host_ptr, size, max_ssbo_size); ReserveGlobalRegion(region); } region->Reload(size); @@ -65,7 +67,11 @@ void GlobalRegionCacheOpenGL::ReserveGlobalRegion(GlobalRegion region) { } GlobalRegionCacheOpenGL::GlobalRegionCacheOpenGL(RasterizerOpenGL& rasterizer) - : RasterizerCache{rasterizer} {} + : RasterizerCache{rasterizer} { + GLint max_ssbo_size_; + glGetIntegerv(GL_MAX_SHADER_STORAGE_BLOCK_SIZE, &max_ssbo_size_); + max_ssbo_size = static_cast<u32>(max_ssbo_size_); +} GlobalRegion GlobalRegionCacheOpenGL::GetGlobalRegion( const GLShader::GlobalMemoryEntry& global_region, @@ -73,7 +79,7 @@ GlobalRegion GlobalRegionCacheOpenGL::GetGlobalRegion( auto& gpu{Core::System::GetInstance().GPU()}; auto& memory_manager{gpu.MemoryManager()}; - const auto cbufs{gpu.Maxwell3D().state.shader_stages[static_cast<u64>(stage)]}; + const auto cbufs{gpu.Maxwell3D().state.shader_stages[static_cast<std::size_t>(stage)]}; const auto addr{cbufs.const_buffers[global_region.GetCbufIndex()].address + global_region.GetCbufOffset()}; const auto actual_addr{memory_manager.Read<u64>(addr)}; @@ -85,7 +91,7 @@ GlobalRegion GlobalRegionCacheOpenGL::GetGlobalRegion( if (!region) { // No global region found - create a new one - region = GetUncachedGlobalRegion(actual_addr, size, host_ptr); + region = GetUncachedGlobalRegion(actual_addr, host_ptr, size); Register(region); } diff --git a/src/video_core/renderer_opengl/gl_global_cache.h b/src/video_core/renderer_opengl/gl_global_cache.h index 5a21ab66f..196e6e278 100644 --- a/src/video_core/renderer_opengl/gl_global_cache.h +++ b/src/video_core/renderer_opengl/gl_global_cache.h @@ -19,7 +19,7 @@ namespace OpenGL { namespace GLShader { class GlobalMemoryEntry; -} // namespace GLShader +} class RasterizerOpenGL; class CachedGlobalRegion; @@ -27,7 +27,8 @@ using GlobalRegion = std::shared_ptr<CachedGlobalRegion>; class CachedGlobalRegion final : public RasterizerCacheObject { public: - explicit CachedGlobalRegion(VAddr cpu_addr, u32 size, u8* host_ptr); + explicit CachedGlobalRegion(VAddr cpu_addr, u8* host_ptr, u32 size, u32 max_size); + ~CachedGlobalRegion(); VAddr GetCpuAddr() const override { return cpu_addr; @@ -45,14 +46,14 @@ public: /// Reloads the global region from guest memory void Reload(u32 size_); - // TODO(Rodrigo): When global memory is written (STG), implement flushing - void Flush() override { - UNIMPLEMENTED(); - } + void Flush() override; private: VAddr cpu_addr{}; + u8* host_ptr{}; u32 size{}; + u32 max_size{}; + OGLBuffer buffer; }; @@ -66,10 +67,11 @@ public: private: GlobalRegion TryGetReservedGlobalRegion(CacheAddr addr, u32 size) const; - GlobalRegion GetUncachedGlobalRegion(GPUVAddr addr, u32 size, u8* host_ptr); + GlobalRegion GetUncachedGlobalRegion(GPUVAddr addr, u8* host_ptr, u32 size); void ReserveGlobalRegion(GlobalRegion region); std::unordered_map<CacheAddr, GlobalRegion> reserve; + u32 max_ssbo_size{}; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp index 2bcbd3da2..c3e94d917 100644 --- a/src/video_core/renderer_opengl/gl_primitive_assembler.cpp +++ b/src/video_core/renderer_opengl/gl_primitive_assembler.cpp @@ -7,7 +7,7 @@ #include "common/assert.h" #include "common/common_types.h" #include "core/core.h" -#include "core/memory.h" +#include "video_core/memory_manager.h" #include "video_core/renderer_opengl/gl_buffer_cache.h" #include "video_core/renderer_opengl/gl_primitive_assembler.h" diff --git a/src/video_core/renderer_opengl/gl_primitive_assembler.h b/src/video_core/renderer_opengl/gl_primitive_assembler.h index 0e2e7dc36..4e87ce4d6 100644 --- a/src/video_core/renderer_opengl/gl_primitive_assembler.h +++ b/src/video_core/renderer_opengl/gl_primitive_assembler.h @@ -4,11 +4,9 @@ #pragma once -#include <vector> #include <glad/glad.h> #include "common/common_types.h" -#include "video_core/memory_manager.h" namespace OpenGL { diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp index 046fc935b..0b1fe3494 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp @@ -17,7 +17,6 @@ #include "common/microprofile.h" #include "common/scope_exit.h" #include "core/core.h" -#include "core/frontend/emu_window.h" #include "core/hle/kernel/process.h" #include "core/settings.h" #include "video_core/engines/maxwell_3d.h" @@ -26,7 +25,6 @@ #include "video_core/renderer_opengl/gl_shader_gen.h" #include "video_core/renderer_opengl/maxwell_to_gl.h" #include "video_core/renderer_opengl/renderer_opengl.h" -#include "video_core/video_core.h" namespace OpenGL { @@ -301,6 +299,10 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { BaseBindings base_bindings; std::array<bool, Maxwell::NumClipDistances> clip_distances{}; + // Prepare packed bindings + bind_ubo_pushbuffer.Setup(base_bindings.cbuf); + bind_ssbo_pushbuffer.Setup(base_bindings.gmem); + for (std::size_t index = 0; index < Maxwell::MaxShaderProgram; ++index) { const auto& shader_config = gpu.regs.shader_config[index]; const Maxwell::ShaderProgram program{static_cast<Maxwell::ShaderProgram>(index)}; @@ -318,13 +320,13 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { const std::size_t stage{index == 0 ? 0 : index - 1}; // Stage indices are 0 - 5 GLShader::MaxwellUniformData ubo{}; - ubo.SetFromRegs(gpu.state.shader_stages[stage]); + ubo.SetFromRegs(gpu, stage); const GLintptr offset = buffer_cache.UploadHostMemory( &ubo, sizeof(ubo), static_cast<std::size_t>(uniform_buffer_alignment)); // Bind the emulation info buffer - glBindBufferRange(GL_UNIFORM_BUFFER, base_bindings.cbuf, buffer_cache.GetHandle(), offset, - static_cast<GLsizeiptr>(sizeof(ubo))); + bind_ubo_pushbuffer.Push(buffer_cache.GetHandle(), offset, + static_cast<GLsizeiptr>(sizeof(ubo))); Shader shader{shader_cache.GetStageProgram(program)}; const auto [program_handle, next_bindings] = @@ -368,6 +370,9 @@ void RasterizerOpenGL::SetupShaders(GLenum primitive_mode) { base_bindings = next_bindings; } + bind_ubo_pushbuffer.Bind(); + bind_ssbo_pushbuffer.Bind(); + SyncClipEnabled(clip_distances); gpu.dirty_flags.shaders = false; @@ -577,9 +582,6 @@ std::pair<bool, bool> RasterizerOpenGL::ConfigureFramebuffers( } void RasterizerOpenGL::Clear() { - const auto prev_state{state}; - SCOPE_EXIT({ prev_state.Apply(); }); - const auto& regs = system.GPU().Maxwell3D().regs; bool use_color{}; bool use_depth{}; @@ -651,7 +653,10 @@ void RasterizerOpenGL::Clear() { clear_state.EmulateViewportWithScissor(); } - clear_state.Apply(); + clear_state.ApplyColorMask(); + clear_state.ApplyDepth(); + clear_state.ApplyStencilTest(); + clear_state.ApplyViewport(); if (use_color) { glClearBufferfv(GL_COLOR, regs.clear_buffers.RT, regs.clear_color); @@ -751,6 +756,7 @@ void RasterizerOpenGL::FlushRegion(CacheAddr addr, u64 size) { return; } res_cache.FlushRegion(addr, size); + global_cache.FlushRegion(addr, size); } void RasterizerOpenGL::InvalidateRegion(CacheAddr addr, u64 size) { @@ -902,23 +908,14 @@ void RasterizerOpenGL::SetupConstBuffers(Tegra::Engines::Maxwell3D::Regs::Shader const auto& shader_stage = maxwell3d.state.shader_stages[static_cast<std::size_t>(stage)]; const auto& entries = shader->GetShaderEntries().const_buffers; - constexpr u64 max_binds = Tegra::Engines::Maxwell3D::Regs::MaxConstBuffers; - std::array<GLuint, max_binds> bind_buffers; - std::array<GLintptr, max_binds> bind_offsets; - std::array<GLsizeiptr, max_binds> bind_sizes; - - ASSERT_MSG(entries.size() <= max_binds, "Exceeded expected number of binding points."); - // Upload only the enabled buffers from the 16 constbuffers of each shader stage for (u32 bindpoint = 0; bindpoint < entries.size(); ++bindpoint) { const auto& used_buffer = entries[bindpoint]; const auto& buffer = shader_stage.const_buffers[used_buffer.GetIndex()]; if (!buffer.enabled) { - // With disabled buffers set values as zero to unbind them - bind_buffers[bindpoint] = 0; - bind_offsets[bindpoint] = 0; - bind_sizes[bindpoint] = 0; + // Set values to zero to unbind buffers + bind_ubo_pushbuffer.Push(0, 0, 0); continue; } @@ -946,30 +943,22 @@ void RasterizerOpenGL::SetupConstBuffers(Tegra::Engines::Maxwell3D::Regs::Shader const GLintptr const_buffer_offset = buffer_cache.UploadMemory( buffer.address, size, static_cast<std::size_t>(uniform_buffer_alignment)); - // Prepare values for multibind - bind_buffers[bindpoint] = buffer_cache.GetHandle(); - bind_offsets[bindpoint] = const_buffer_offset; - bind_sizes[bindpoint] = size; + bind_ubo_pushbuffer.Push(buffer_cache.GetHandle(), const_buffer_offset, size); } - - // The first binding is reserved for emulation values - const GLuint ubo_base_binding = base_bindings.cbuf + 1; - glBindBuffersRange(GL_UNIFORM_BUFFER, ubo_base_binding, static_cast<GLsizei>(entries.size()), - bind_buffers.data(), bind_offsets.data(), bind_sizes.data()); } void RasterizerOpenGL::SetupGlobalRegions(Tegra::Engines::Maxwell3D::Regs::ShaderStage stage, const Shader& shader, GLenum primitive_mode, BaseBindings base_bindings) { - // TODO(Rodrigo): Use ARB_multi_bind here const auto& entries = shader->GetShaderEntries().global_memory_entries; - - for (u32 bindpoint = 0; bindpoint < static_cast<u32>(entries.size()); ++bindpoint) { - const auto& entry = entries[bindpoint]; - const u32 current_bindpoint = base_bindings.gmem + bindpoint; - const auto& region = global_cache.GetGlobalRegion(entry, stage); - - glBindBufferBase(GL_SHADER_STORAGE_BUFFER, current_bindpoint, region->GetBufferHandle()); + for (std::size_t bindpoint = 0; bindpoint < entries.size(); ++bindpoint) { + const auto& entry{entries[bindpoint]}; + const auto& region{global_cache.GetGlobalRegion(entry, stage)}; + if (entry.IsWritten()) { + region->MarkAsModified(true, global_cache); + } + bind_ssbo_pushbuffer.Push(region->GetBufferHandle(), 0, + static_cast<GLsizeiptr>(region->GetSizeInBytes())); } } diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h index 4de565321..d4c2cf80e 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer.h +++ b/src/video_core/renderer_opengl/gl_rasterizer.h @@ -12,15 +12,12 @@ #include <optional> #include <tuple> #include <utility> -#include <vector> #include <boost/icl/interval_map.hpp> -#include <boost/range/iterator_range.hpp> #include <glad/glad.h> #include "common/common_types.h" #include "video_core/engines/maxwell_3d.h" -#include "video_core/memory_manager.h" #include "video_core/rasterizer_cache.h" #include "video_core/rasterizer_interface.h" #include "video_core/renderer_opengl/gl_buffer_cache.h" @@ -29,10 +26,9 @@ #include "video_core/renderer_opengl/gl_rasterizer_cache.h" #include "video_core/renderer_opengl/gl_resource_manager.h" #include "video_core/renderer_opengl/gl_shader_cache.h" -#include "video_core/renderer_opengl/gl_shader_gen.h" #include "video_core/renderer_opengl/gl_shader_manager.h" #include "video_core/renderer_opengl/gl_state.h" -#include "video_core/renderer_opengl/gl_stream_buffer.h" +#include "video_core/renderer_opengl/utils.h" namespace Core { class System; @@ -75,10 +71,6 @@ public: static_assert(MaxConstbufferSize % sizeof(GLvec4) == 0, "The maximum size of a constbuffer must be a multiple of the size of GLvec4"); - static constexpr std::size_t MaxGlobalMemorySize = 0x10000; - static_assert(MaxGlobalMemorySize % sizeof(float) == 0, - "The maximum size of a global memory must be a multiple of the size of float"); - private: class SamplerInfo { public: @@ -234,6 +226,9 @@ private: PrimitiveAssembler primitive_assembler{buffer_cache}; GLint uniform_buffer_alignment; + BindBuffersRangePushBuffer bind_ubo_pushbuffer{GL_UNIFORM_BUFFER}; + BindBuffersRangePushBuffer bind_ssbo_pushbuffer{GL_SHADER_STORAGE_BUFFER}; + std::size_t CalculateVertexArraysSize() const; std::size_t CalculateIndexBufferSize() const; diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp index aba6ce731..7a68b8738 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp @@ -13,9 +13,9 @@ #include "common/scope_exit.h" #include "core/core.h" #include "core/hle/kernel/process.h" -#include "core/memory.h" #include "core/settings.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/memory_manager.h" #include "video_core/morton.h" #include "video_core/renderer_opengl/gl_rasterizer.h" #include "video_core/renderer_opengl/gl_rasterizer_cache.h" @@ -112,11 +112,26 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only, params.pixel_format = PixelFormatFromTextureFormat(config.tic.format, config.tic.r_type.Value(), params.srgb_conversion); - if (params.pixel_format == PixelFormat::R16U && config.tsc.depth_compare_enabled) { + if (config.tsc.depth_compare_enabled) { // Some titles create a 'R16U' (normalized 16-bit) texture with depth_compare enabled, // then attempt to sample from it via a shadow sampler. Convert format to Z16 (which also // causes GetFormatType to properly return 'Depth' below). - params.pixel_format = PixelFormat::Z16; + if (GetFormatType(params.pixel_format) == SurfaceType::ColorTexture) { + switch (params.pixel_format) { + case PixelFormat::R16S: + case PixelFormat::R16U: + case PixelFormat::R16F: + params.pixel_format = PixelFormat::Z16; + break; + case PixelFormat::R32F: + params.pixel_format = PixelFormat::Z32F; + break; + default: + LOG_WARNING(HW_GPU, "Color texture format being used with depth compare: {}", + static_cast<u32>(params.pixel_format)); + break; + } + } } params.component_type = ComponentTypeFromTexture(config.tic.r_type.Value()); @@ -266,6 +281,7 @@ std::size_t SurfaceParams::InnerMemorySize(bool force_gl, bool layer_only, params.component_type = ComponentTypeFromRenderTarget(config.format); params.type = GetFormatType(params.pixel_format); params.width = config.width; + params.pitch = config.pitch; params.height = config.height; params.unaligned_height = config.height; params.target = SurfaceTarget::Texture2D; @@ -662,8 +678,8 @@ void CachedSurface::FlushGLBuffer() { gl_buffer[0].resize(GetSizeInBytes()); const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type); - // Ensure no bad interactions with GL_UNPACK_ALIGNMENT - ASSERT(params.width * GetBytesPerPixel(params.pixel_format) % 4 == 0); + const u32 align = std::clamp(params.RowAlign(0), 1U, 8U); + glPixelStorei(GL_PACK_ALIGNMENT, align); glPixelStorei(GL_PACK_ROW_LENGTH, static_cast<GLint>(params.width)); ASSERT(!tuple.compressed); glBindBuffer(GL_PIXEL_PACK_BUFFER, 0); @@ -708,8 +724,8 @@ void CachedSurface::UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, const FormatTuple& tuple = GetFormatTuple(params.pixel_format, params.component_type); - // Ensure no bad interactions with GL_UNPACK_ALIGNMENT - ASSERT(params.MipWidth(mip_map) * GetBytesPerPixel(params.pixel_format) % 4 == 0); + const u32 align = std::clamp(params.RowAlign(mip_map), 1U, 8U); + glPixelStorei(GL_UNPACK_ALIGNMENT, align); glPixelStorei(GL_UNPACK_ROW_LENGTH, static_cast<GLint>(params.MipWidth(mip_map))); const auto image_size = static_cast<GLsizei>(params.GetMipmapSizeGL(mip_map, false)); @@ -1175,10 +1191,16 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface, return new_surface; } + const bool old_compressed = + GetFormatTuple(old_params.pixel_format, old_params.component_type).compressed; + const bool new_compressed = + GetFormatTuple(new_params.pixel_format, new_params.component_type).compressed; + const bool compatible_formats = + GetFormatBpp(old_params.pixel_format) == GetFormatBpp(new_params.pixel_format) && + !(old_compressed || new_compressed); // For compatible surfaces, we can just do fast glCopyImageSubData based copy - if (old_params.target == new_params.target && old_params.type == new_params.type && - old_params.depth == new_params.depth && old_params.depth == 1 && - GetFormatBpp(old_params.pixel_format) == GetFormatBpp(new_params.pixel_format)) { + if (old_params.target == new_params.target && old_params.depth == new_params.depth && + old_params.depth == 1 && compatible_formats) { FastCopySurface(old_surface, new_surface); return new_surface; } @@ -1193,7 +1215,7 @@ Surface RasterizerCacheOpenGL::RecreateSurface(const Surface& old_surface, case SurfaceTarget::TextureCubemap: case SurfaceTarget::Texture2DArray: case SurfaceTarget::TextureCubeArray: - if (old_params.pixel_format == new_params.pixel_format) + if (compatible_formats) FastLayeredCopySurface(old_surface, new_surface); else { AccurateCopySurface(old_surface, new_surface); diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h index e8073579f..db280dbb3 100644 --- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h +++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h @@ -5,13 +5,13 @@ #pragma once #include <array> -#include <map> #include <memory> #include <string> -#include <unordered_set> +#include <tuple> #include <vector> #include "common/alignment.h" +#include "common/bit_util.h" #include "common/common_types.h" #include "common/hash.h" #include "common/math_util.h" @@ -206,6 +206,13 @@ struct SurfaceParams { return bd; } + u32 RowAlign(u32 mip_level) const { + const u32 m_width = MipWidth(mip_level); + const u32 bytes_per_pixel = GetBytesPerPixel(pixel_format); + const u32 l2 = Common::CountTrailingZeroes32(m_width * bytes_per_pixel); + return (1U << l2); + } + /// Creates SurfaceParams from a texture configuration static SurfaceParams CreateForTexture(const Tegra::Texture::FullTextureInfo& config, const GLShader::SamplerEntry& entry); diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp index 290e654bc..99f67494c 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp @@ -6,13 +6,12 @@ #include "common/assert.h" #include "common/hash.h" #include "core/core.h" -#include "core/memory.h" #include "video_core/engines/maxwell_3d.h" +#include "video_core/memory_manager.h" #include "video_core/renderer_opengl/gl_rasterizer.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_disk_cache.h" -#include "video_core/renderer_opengl/gl_shader_manager.h" #include "video_core/renderer_opengl/utils.h" #include "video_core/shader/shader_ir.h" @@ -41,6 +40,10 @@ GPUVAddr GetShaderAddress(Maxwell::ShaderProgram program) { /// Gets the shader program code from memory for the specified address ProgramCode GetShaderCode(const u8* host_ptr) { ProgramCode program_code(VideoCommon::Shader::MAX_PROGRAM_LENGTH); + ASSERT_OR_EXECUTE(host_ptr != nullptr, { + std::fill(program_code.begin(), program_code.end(), 0); + return program_code; + }); std::memcpy(program_code.data(), host_ptr, program_code.size() * sizeof(u64)); return program_code; } diff --git a/src/video_core/renderer_opengl/gl_shader_cache.h b/src/video_core/renderer_opengl/gl_shader_cache.h index fd1c85115..0cf8e0b3d 100644 --- a/src/video_core/renderer_opengl/gl_shader_cache.h +++ b/src/video_core/renderer_opengl/gl_shader_cache.h @@ -5,21 +5,20 @@ #pragma once #include <array> +#include <atomic> #include <memory> #include <set> #include <tuple> #include <unordered_map> +#include <vector> #include <glad/glad.h> -#include "common/assert.h" #include "common/common_types.h" #include "video_core/rasterizer_cache.h" -#include "video_core/renderer_base.h" #include "video_core/renderer_opengl/gl_resource_manager.h" #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_shader_gen.h" namespace Core { class System; diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp index 11d1169f0..445048daf 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.cpp +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.cpp @@ -21,6 +21,8 @@ namespace OpenGL::GLShader { +namespace { + using Tegra::Shader::Attribute; using Tegra::Shader::AttributeUse; using Tegra::Shader::Header; @@ -34,13 +36,15 @@ using Maxwell = Tegra::Engines::Maxwell3D::Regs; using ShaderStage = Tegra::Engines::Maxwell3D::Regs::ShaderStage; using Operation = const OperationNode&; +enum class Type { Bool, Bool2, Float, Int, Uint, HalfFloat }; + +struct TextureAoffi {}; +using TextureArgument = std::pair<Type, Node>; +using TextureIR = std::variant<TextureAoffi, TextureArgument>; + enum : u32 { POSITION_VARYING_LOCATION = 0, GENERIC_VARYING_START_LOCATION = 1 }; constexpr u32 MAX_CONSTBUFFER_ELEMENTS = static_cast<u32>(RasterizerOpenGL::MaxConstbufferSize) / (4 * sizeof(float)); -constexpr u32 MAX_GLOBALMEMORY_ELEMENTS = - static_cast<u32>(RasterizerOpenGL::MaxGlobalMemorySize) / sizeof(float); - -enum class Type { Bool, Bool2, Float, Int, Uint, HalfFloat }; class ShaderWriter { public: @@ -69,10 +73,10 @@ public: shader_source += '\n'; } - std::string GenerateTemporal() { - std::string temporal = "tmp"; - temporal += std::to_string(temporal_index++); - return temporal; + std::string GenerateTemporary() { + std::string temporary = "tmp"; + temporary += std::to_string(temporary_index++); + return temporary; } std::string GetResult() { @@ -87,11 +91,11 @@ private: } std::string shader_source; - u32 temporal_index = 1; + u32 temporary_index = 1; }; /// Generates code to use for a swizzle operation. -static std::string GetSwizzle(u32 elem) { +std::string GetSwizzle(u32 elem) { ASSERT(elem <= 3); std::string swizzle = "."; swizzle += "xyzw"[elem]; @@ -99,7 +103,7 @@ static std::string GetSwizzle(u32 elem) { } /// Translate topology -static std::string GetTopologyName(Tegra::Shader::OutputTopology topology) { +std::string GetTopologyName(Tegra::Shader::OutputTopology topology) { switch (topology) { case Tegra::Shader::OutputTopology::PointList: return "points"; @@ -114,7 +118,7 @@ static std::string GetTopologyName(Tegra::Shader::OutputTopology topology) { } /// Returns true if an object has to be treated as precise -static bool IsPrecise(Operation operand) { +bool IsPrecise(Operation operand) { const auto& meta = operand.GetMeta(); if (const auto arithmetic = std::get_if<MetaArithmetic>(&meta)) { @@ -126,7 +130,7 @@ static bool IsPrecise(Operation operand) { return false; } -static bool IsPrecise(Node node) { +bool IsPrecise(Node node) { if (const auto operation = std::get_if<OperationNode>(node)) { return IsPrecise(*operation); } @@ -202,8 +206,10 @@ public: for (const auto& sampler : ir.GetSamplers()) { entries.samplers.emplace_back(sampler); } - for (const auto& gmem : ir.GetGlobalMemoryBases()) { - entries.global_memory_entries.emplace_back(gmem.cbuf_index, gmem.cbuf_offset); + for (const auto& gmem_pair : ir.GetGlobalMemory()) { + const auto& [base, usage] = gmem_pair; + entries.global_memory_entries.emplace_back(base.cbuf_index, base.cbuf_offset, + usage.is_read, usage.is_written); } entries.clip_distances = ir.GetClipDistances(); entries.shader_length = ir.GetLength(); @@ -374,12 +380,22 @@ private: } void DeclareGlobalMemory() { - for (const auto& entry : ir.GetGlobalMemoryBases()) { + for (const auto& gmem : ir.GetGlobalMemory()) { + const auto& [base, usage] = gmem; + + // Since we don't know how the shader will use the shader, hint the driver to disable as + // much optimizations as possible + std::string qualifier = "coherent volatile"; + if (usage.is_read && !usage.is_written) + qualifier += " readonly"; + else if (usage.is_written && !usage.is_read) + qualifier += " writeonly"; + const std::string binding = - fmt::format("GMEM_BINDING_{}_{}", entry.cbuf_index, entry.cbuf_offset); - code.AddLine("layout (std430, binding = " + binding + ") buffer " + - GetGlobalMemoryBlock(entry) + " {"); - code.AddLine(" float " + GetGlobalMemory(entry) + "[MAX_GLOBALMEMORY_ELEMENTS];"); + fmt::format("GMEM_BINDING_{}_{}", base.cbuf_index, base.cbuf_offset); + code.AddLine("layout (std430, binding = " + binding + ") " + qualifier + " buffer " + + GetGlobalMemoryBlock(base) + " {"); + code.AddLine(" float " + GetGlobalMemory(base) + "[];"); code.AddLine("};"); code.AddNewLine(); } @@ -426,9 +442,14 @@ private: std::string Visit(Node node) { if (const auto operation = std::get_if<OperationNode>(node)) { const auto operation_index = static_cast<std::size_t>(operation->GetCode()); + if (operation_index >= operation_decompilers.size()) { + UNREACHABLE_MSG("Out of bounds operation: {}", operation_index); + return {}; + } const auto decompiler = operation_decompilers[operation_index]; if (decompiler == nullptr) { - UNREACHABLE_MSG("Operation decompiler {} not defined", operation_index); + UNREACHABLE_MSG("Undefined operation: {}", operation_index); + return {}; } return (this->*decompiler)(*operation); @@ -540,9 +561,8 @@ private: } else if (std::holds_alternative<OperationNode>(*offset)) { // Indirect access - const std::string final_offset = code.GenerateTemporal(); - code.AddLine("uint " + final_offset + " = (ftou(" + Visit(offset) + ") / 4) & " + - std::to_string(MAX_CONSTBUFFER_ELEMENTS - 1) + ';'); + const std::string final_offset = code.GenerateTemporary(); + code.AddLine("uint " + final_offset + " = (ftou(" + Visit(offset) + ") / 4);"); return fmt::format("{}[{} / 4][{} % 4]", GetConstBuffer(cbuf->GetIndex()), final_offset, final_offset); @@ -587,9 +607,9 @@ private: // There's a bug in NVidia's proprietary drivers that makes precise fail on fragment shaders const std::string precise = stage != ShaderStage::Fragment ? "precise " : ""; - const std::string temporal = code.GenerateTemporal(); - code.AddLine(precise + "float " + temporal + " = " + value + ';'); - return temporal; + const std::string temporary = code.GenerateTemporary(); + code.AddLine(precise + "float " + temporary + " = " + value + ';'); + return temporary; } std::string VisitOperand(Operation operation, std::size_t operand_index) { @@ -601,9 +621,9 @@ private: return Visit(operand); } - const std::string temporal = code.GenerateTemporal(); - code.AddLine("float " + temporal + " = " + Visit(operand) + ';'); - return temporal; + const std::string temporary = code.GenerateTemporary(); + code.AddLine("float " + temporary + " = " + Visit(operand) + ';'); + return temporary; } std::string VisitOperand(Operation operation, std::size_t operand_index, Type type) { @@ -718,8 +738,8 @@ private: result_type)); } - std::string GenerateTexture(Operation operation, const std::string& func, - const std::vector<std::pair<Type, Node>>& extras) { + std::string GenerateTexture(Operation operation, const std::string& function_suffix, + const std::vector<TextureIR>& extras) { constexpr std::array<const char*, 4> coord_constructors = {"float", "vec2", "vec3", "vec4"}; const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); @@ -729,11 +749,11 @@ private: const bool has_array = meta->sampler.IsArray(); const bool has_shadow = meta->sampler.IsShadow(); - std::string expr = func; - expr += '('; - expr += GetSampler(meta->sampler); - expr += ", "; - + std::string expr = "texture" + function_suffix; + if (!meta->aoffi.empty()) { + expr += "Offset"; + } + expr += '(' + GetSampler(meta->sampler) + ", "; expr += coord_constructors.at(count + (has_array ? 1 : 0) + (has_shadow ? 1 : 0) - 1); expr += '('; for (std::size_t i = 0; i < count; ++i) { @@ -751,36 +771,74 @@ private: } expr += ')'; - for (const auto& extra_pair : extras) { - const auto [type, operand] = extra_pair; - if (operand == nullptr) { - continue; + for (const auto& variant : extras) { + if (const auto argument = std::get_if<TextureArgument>(&variant)) { + expr += GenerateTextureArgument(*argument); + } else if (std::get_if<TextureAoffi>(&variant)) { + expr += GenerateTextureAoffi(meta->aoffi); + } else { + UNREACHABLE(); } - expr += ", "; + } - switch (type) { - case Type::Int: - if (const auto immediate = std::get_if<ImmediateNode>(operand)) { - // Inline the string as an immediate integer in GLSL (some extra arguments are - // required to be constant) - expr += std::to_string(static_cast<s32>(immediate->GetValue())); - } else { - expr += "ftoi(" + Visit(operand) + ')'; - } - break; - case Type::Float: - expr += Visit(operand); - break; - default: { - const auto type_int = static_cast<u32>(type); - UNIMPLEMENTED_MSG("Unimplemented extra type={}", type_int); - expr += '0'; - break; + return expr + ')'; + } + + std::string GenerateTextureArgument(TextureArgument argument) { + const auto [type, operand] = argument; + if (operand == nullptr) { + return {}; + } + + std::string expr = ", "; + switch (type) { + case Type::Int: + if (const auto immediate = std::get_if<ImmediateNode>(operand)) { + // Inline the string as an immediate integer in GLSL (some extra arguments are + // required to be constant) + expr += std::to_string(static_cast<s32>(immediate->GetValue())); + } else { + expr += "ftoi(" + Visit(operand) + ')'; + } + break; + case Type::Float: + expr += Visit(operand); + break; + default: { + const auto type_int = static_cast<u32>(type); + UNIMPLEMENTED_MSG("Unimplemented extra type={}", type_int); + expr += '0'; + break; + } + } + return expr; + } + + std::string GenerateTextureAoffi(const std::vector<Node>& aoffi) { + if (aoffi.empty()) { + return {}; + } + constexpr std::array<const char*, 3> coord_constructors = {"int", "ivec2", "ivec3"}; + std::string expr = ", "; + expr += coord_constructors.at(aoffi.size() - 1); + expr += '('; + + for (std::size_t index = 0; index < aoffi.size(); ++index) { + const auto operand{aoffi.at(index)}; + if (const auto immediate = std::get_if<ImmediateNode>(operand)) { + // Inline the string as an immediate integer in GLSL (AOFFI arguments are required + // to be constant by the standard). + expr += std::to_string(static_cast<s32>(immediate->GetValue())); + } else { + expr += "ftoi(" + Visit(operand) + ')'; } + if (index + 1 < aoffi.size()) { + expr += ", "; } } + expr += ')'; - return expr + ')'; + return expr; } std::string Assign(Operation operation) { @@ -820,6 +878,12 @@ private: } else if (const auto lmem = std::get_if<LmemNode>(dest)) { target = GetLocalMemory() + "[ftou(" + Visit(lmem->GetAddress()) + ") / 4]"; + } else if (const auto gmem = std::get_if<GmemNode>(dest)) { + const std::string real = Visit(gmem->GetRealAddress()); + const std::string base = Visit(gmem->GetBaseAddress()); + const std::string final_offset = "(ftou(" + real + ") - ftou(" + base + ")) / 4"; + target = fmt::format("{}[{}]", GetGlobalMemory(gmem->GetDescriptor()), final_offset); + } else { UNREACHABLE_MSG("Assign called without a proper target"); } @@ -1159,7 +1223,8 @@ private: const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); ASSERT(meta); - std::string expr = GenerateTexture(operation, "texture", {{Type::Float, meta->bias}}); + std::string expr = GenerateTexture( + operation, "", {TextureAoffi{}, TextureArgument{Type::Float, meta->bias}}); if (meta->sampler.IsShadow()) { expr = "vec4(" + expr + ')'; } @@ -1170,7 +1235,8 @@ private: const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); ASSERT(meta); - std::string expr = GenerateTexture(operation, "textureLod", {{Type::Float, meta->lod}}); + std::string expr = GenerateTexture( + operation, "Lod", {TextureArgument{Type::Float, meta->lod}, TextureAoffi{}}); if (meta->sampler.IsShadow()) { expr = "vec4(" + expr + ')'; } @@ -1182,7 +1248,8 @@ private: ASSERT(meta); const auto type = meta->sampler.IsShadow() ? Type::Float : Type::Int; - return GenerateTexture(operation, "textureGather", {{type, meta->component}}) + + return GenerateTexture(operation, "Gather", + {TextureArgument{type, meta->component}, TextureAoffi{}}) + GetSwizzle(meta->element); } @@ -1196,11 +1263,12 @@ private: switch (meta->element) { case 0: case 1: - return "textureSize(" + sampler + ", " + lod + ')' + GetSwizzle(meta->element); + return "itof(int(textureSize(" + sampler + ", " + lod + ')' + + GetSwizzle(meta->element) + "))"; case 2: return "0"; case 3: - return "textureQueryLevels(" + sampler + ')'; + return "itof(textureQueryLevels(" + sampler + "))"; } UNREACHABLE(); return "0"; @@ -1211,8 +1279,8 @@ private: ASSERT(meta); if (meta->element < 2) { - return "itof(int((" + GenerateTexture(operation, "textureQueryLod", {}) + - " * vec2(256))" + GetSwizzle(meta->element) + "))"; + return "itof(int((" + GenerateTexture(operation, "QueryLod", {}) + " * vec2(256))" + + GetSwizzle(meta->element) + "))"; } return "0"; } @@ -1565,11 +1633,11 @@ private: ShaderWriter code; }; +} // Anonymous namespace + std::string GetCommonDeclarations() { const auto cbuf = std::to_string(MAX_CONSTBUFFER_ELEMENTS); - const auto gmem = std::to_string(MAX_GLOBALMEMORY_ELEMENTS); return "#define MAX_CONSTBUFFER_ELEMENTS " + cbuf + "\n" + - "#define MAX_GLOBALMEMORY_ELEMENTS " + gmem + "\n" + "#define ftoi floatBitsToInt\n" "#define ftou floatBitsToUint\n" "#define itof intBitsToFloat\n" diff --git a/src/video_core/renderer_opengl/gl_shader_decompiler.h b/src/video_core/renderer_opengl/gl_shader_decompiler.h index 72aca4938..55b3d4d7b 100644 --- a/src/video_core/renderer_opengl/gl_shader_decompiler.h +++ b/src/video_core/renderer_opengl/gl_shader_decompiler.h @@ -5,7 +5,6 @@ #pragma once #include <array> -#include <set> #include <string> #include <utility> #include <vector> @@ -40,8 +39,9 @@ private: class GlobalMemoryEntry { public: - explicit GlobalMemoryEntry(u32 cbuf_index, u32 cbuf_offset) - : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset} {} + 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 GetCbufIndex() const { return cbuf_index; @@ -51,9 +51,19 @@ public: return cbuf_offset; } + bool IsRead() const { + return is_read; + } + + bool IsWritten() const { + return is_written; + } + private: u32 cbuf_index{}; u32 cbuf_offset{}; + bool is_read{}; + bool is_written{}; }; struct ShaderEntries { 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 82fc4d44b..d5890a375 100644 --- a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp +++ b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp @@ -4,7 +4,6 @@ #include <cstring> #include <fmt/format.h> -#include <lz4.h> #include "common/assert.h" #include "common/common_paths.h" @@ -12,6 +11,7 @@ #include "common/file_util.h" #include "common/logging/log.h" #include "common/scm_rev.h" +#include "common/zstd_compression.h" #include "core/core.h" #include "core/hle/kernel/process.h" @@ -49,39 +49,6 @@ ShaderCacheVersionHash GetShaderCacheVersionHash() { return hash; } -template <typename T> -std::vector<u8> CompressData(const T* source, std::size_t source_size) { - if (source_size > LZ4_MAX_INPUT_SIZE) { - // Source size exceeds LZ4 maximum input size - return {}; - } - const auto source_size_int = static_cast<int>(source_size); - const int max_compressed_size = LZ4_compressBound(source_size_int); - std::vector<u8> compressed(max_compressed_size); - const int compressed_size = LZ4_compress_default(reinterpret_cast<const char*>(source), - reinterpret_cast<char*>(compressed.data()), - source_size_int, max_compressed_size); - if (compressed_size <= 0) { - // Compression failed - return {}; - } - compressed.resize(compressed_size); - return compressed; -} - -std::vector<u8> DecompressData(const std::vector<u8>& compressed, std::size_t uncompressed_size) { - std::vector<u8> uncompressed(uncompressed_size); - const int size_check = LZ4_decompress_safe(reinterpret_cast<const char*>(compressed.data()), - reinterpret_cast<char*>(uncompressed.data()), - static_cast<int>(compressed.size()), - static_cast<int>(uncompressed.size())); - if (static_cast<int>(uncompressed_size) != size_check) { - // Decompression failed - return {}; - } - return uncompressed; -} - } // namespace ShaderDiskCacheRaw::ShaderDiskCacheRaw(u64 unique_identifier, Maxwell::ShaderProgram program_type, @@ -292,7 +259,7 @@ ShaderDiskCacheOpenGL::LoadPrecompiledFile(FileUtil::IOFile& file) { return {}; } - dump.binary = DecompressData(compressed_binary, binary_length); + dump.binary = Common::Compression::DecompressDataZSTD(compressed_binary); if (dump.binary.empty()) { return {}; } @@ -321,7 +288,7 @@ std::optional<ShaderDiskCacheDecompiled> ShaderDiskCacheOpenGL::LoadDecompiledEn return {}; } - const std::vector<u8> code = DecompressData(compressed_code, code_size); + const std::vector<u8> code = Common::Compression::DecompressDataZSTD(compressed_code); if (code.empty()) { return {}; } @@ -370,11 +337,16 @@ std::optional<ShaderDiskCacheDecompiled> ShaderDiskCacheOpenGL::LoadDecompiledEn for (u32 i = 0; i < global_memory_count; ++i) { u32 cbuf_index{}; u32 cbuf_offset{}; + u8 is_read{}; + u8 is_written{}; if (file.ReadBytes(&cbuf_index, sizeof(u32)) != sizeof(u32) || - file.ReadBytes(&cbuf_offset, sizeof(u32)) != sizeof(u32)) { + file.ReadBytes(&cbuf_offset, sizeof(u32)) != sizeof(u32) || + file.ReadBytes(&is_read, sizeof(u8)) != sizeof(u8) || + file.ReadBytes(&is_written, sizeof(u8)) != sizeof(u8)) { return {}; } - entry.entries.global_memory_entries.emplace_back(cbuf_index, cbuf_offset); + entry.entries.global_memory_entries.emplace_back(cbuf_index, cbuf_offset, is_read != 0, + is_written != 0); } for (auto& clip_distance : entry.entries.clip_distances) { @@ -430,7 +402,9 @@ bool ShaderDiskCacheOpenGL::SaveDecompiledFile(FileUtil::IOFile& file, u64 uniqu return false; for (const auto& gmem : entries.global_memory_entries) { if (file.WriteObject(static_cast<u32>(gmem.GetCbufIndex())) != 1 || - file.WriteObject(static_cast<u32>(gmem.GetCbufOffset())) != 1) { + file.WriteObject(static_cast<u32>(gmem.GetCbufOffset())) != 1 || + file.WriteObject(static_cast<u8>(gmem.IsRead() ? 1 : 0)) != 1 || + file.WriteObject(static_cast<u8>(gmem.IsWritten() ? 1 : 0)) != 1) { return false; } } @@ -507,7 +481,8 @@ void ShaderDiskCacheOpenGL::SaveDecompiled(u64 unique_identifier, const std::str if (!IsUsable()) return; - const std::vector<u8> compressed_code{CompressData(code.data(), code.size())}; + const std::vector<u8> compressed_code{Common::Compression::CompressDataZSTDDefault( + reinterpret_cast<const u8*>(code.data()), code.size())}; if (compressed_code.empty()) { LOG_ERROR(Render_OpenGL, "Failed to compress GLSL code - skipping shader {:016x}", unique_identifier); @@ -537,7 +512,9 @@ void ShaderDiskCacheOpenGL::SaveDump(const ShaderDiskCacheUsage& usage, GLuint p std::vector<u8> binary(binary_length); glGetProgramBinary(program, binary_length, nullptr, &binary_format, binary.data()); - const std::vector<u8> compressed_binary = CompressData(binary.data(), binary.size()); + const std::vector<u8> compressed_binary = + Common::Compression::CompressDataZSTDDefault(binary.data(), binary.size()); + if (compressed_binary.empty()) { LOG_ERROR(Render_OpenGL, "Failed to compress binary program in shader={:016x}", usage.unique_identifier); diff --git a/src/video_core/renderer_opengl/gl_shader_gen.cpp b/src/video_core/renderer_opengl/gl_shader_gen.cpp index 7d96649af..8763d9c71 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.cpp +++ b/src/video_core/renderer_opengl/gl_shader_gen.cpp @@ -3,7 +3,6 @@ // Refer to the license.txt file included. #include <fmt/format.h> -#include "common/assert.h" #include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_opengl/gl_shader_decompiler.h" #include "video_core/renderer_opengl/gl_shader_gen.h" diff --git a/src/video_core/renderer_opengl/gl_shader_gen.h b/src/video_core/renderer_opengl/gl_shader_gen.h index fba8e681b..fad346b48 100644 --- a/src/video_core/renderer_opengl/gl_shader_gen.h +++ b/src/video_core/renderer_opengl/gl_shader_gen.h @@ -4,12 +4,9 @@ #pragma once -#include <array> -#include <string> #include <vector> #include "common/common_types.h" -#include "video_core/engines/shader_bytecode.h" #include "video_core/renderer_opengl/gl_shader_decompiler.h" #include "video_core/shader/shader_ir.h" diff --git a/src/video_core/renderer_opengl/gl_shader_manager.cpp b/src/video_core/renderer_opengl/gl_shader_manager.cpp index 6a30c28d2..05ab01dcb 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.cpp +++ b/src/video_core/renderer_opengl/gl_shader_manager.cpp @@ -2,23 +2,55 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include "core/core.h" +#include "common/common_types.h" +#include "video_core/engines/maxwell_3d.h" #include "video_core/renderer_opengl/gl_shader_manager.h" namespace OpenGL::GLShader { -void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage) { - const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D(); - const auto& regs = gpu.regs; - const auto& state = gpu.state; +using Tegra::Engines::Maxwell3D; + +ProgramManager::ProgramManager() { + pipeline.Create(); +} + +ProgramManager::~ProgramManager() = default; + +void ProgramManager::ApplyTo(OpenGLState& state) { + UpdatePipeline(); + state.draw.shader_program = 0; + state.draw.program_pipeline = pipeline.handle; +} + +void ProgramManager::UpdatePipeline() { + // Avoid updating the pipeline when values have no changed + if (old_state == current_state) { + return; + } + + // Workaround for AMD bug + constexpr GLenum all_used_stages{GL_VERTEX_SHADER_BIT | GL_GEOMETRY_SHADER_BIT | + GL_FRAGMENT_SHADER_BIT}; + glUseProgramStages(pipeline.handle, all_used_stages, 0); + + glUseProgramStages(pipeline.handle, GL_VERTEX_SHADER_BIT, current_state.vertex_shader); + glUseProgramStages(pipeline.handle, GL_GEOMETRY_SHADER_BIT, current_state.geometry_shader); + glUseProgramStages(pipeline.handle, GL_FRAGMENT_SHADER_BIT, current_state.fragment_shader); + + old_state = current_state; +} + +void MaxwellUniformData::SetFromRegs(const Maxwell3D& maxwell, std::size_t shader_stage) { + const auto& regs = maxwell.regs; + const auto& state = maxwell.state; // TODO(bunnei): Support more than one viewport viewport_flip[0] = regs.viewport_transform[0].scale_x < 0.0 ? -1.0f : 1.0f; viewport_flip[1] = regs.viewport_transform[0].scale_y < 0.0 ? -1.0f : 1.0f; - u32 func = static_cast<u32>(regs.alpha_test_func); + auto func{static_cast<u32>(regs.alpha_test_func)}; // Normalize the gl variants of opCompare to be the same as the normal variants - u32 op_gl_variant_base = static_cast<u32>(Tegra::Engines::Maxwell3D::Regs::ComparisonOp::Never); + const u32 op_gl_variant_base = static_cast<u32>(Maxwell3D::Regs::ComparisonOp::Never); if (func >= op_gl_variant_base) { func = func - op_gl_variant_base + 1U; } @@ -31,8 +63,9 @@ void MaxwellUniformData::SetFromRegs(const Maxwell3D::State::ShaderStageInfo& sh // Assign in which stage the position has to be flipped // (the last stage before the fragment shader). - if (gpu.regs.shader_config[static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry)].enable) { - flip_stage = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry); + constexpr u32 geometry_index = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::Geometry); + if (maxwell.regs.shader_config[geometry_index].enable) { + flip_stage = geometry_index; } else { flip_stage = static_cast<u32>(Maxwell3D::Regs::ShaderProgram::VertexB); } diff --git a/src/video_core/renderer_opengl/gl_shader_manager.h b/src/video_core/renderer_opengl/gl_shader_manager.h index 4970aafed..cec18a832 100644 --- a/src/video_core/renderer_opengl/gl_shader_manager.h +++ b/src/video_core/renderer_opengl/gl_shader_manager.h @@ -4,6 +4,8 @@ #pragma once +#include <cstddef> + #include <glad/glad.h> #include "video_core/renderer_opengl/gl_resource_manager.h" @@ -12,14 +14,13 @@ namespace OpenGL::GLShader { -using Tegra::Engines::Maxwell3D; - /// Uniform structure for the Uniform Buffer Object, all vectors must be 16-byte aligned -// NOTE: Always keep a vec4 at the end. The GL spec is not clear whether the alignment at -// the end of a uniform block is included in UNIFORM_BLOCK_DATA_SIZE or not. -// Not following that rule will cause problems on some AMD drivers. +/// @note Always keep a vec4 at the end. The GL spec is not clear whether the alignment at +/// the end of a uniform block is included in UNIFORM_BLOCK_DATA_SIZE or not. +/// Not following that rule will cause problems on some AMD drivers. struct MaxwellUniformData { - void SetFromRegs(const Maxwell3D::State::ShaderStageInfo& shader_stage); + void SetFromRegs(const Tegra::Engines::Maxwell3D& maxwell, std::size_t shader_stage); + alignas(16) GLvec4 viewport_flip; struct alignas(16) { GLuint instance_id; @@ -39,56 +40,48 @@ static_assert(sizeof(MaxwellUniformData) < 16384, class ProgramManager { public: - ProgramManager() { - pipeline.Create(); - } + explicit ProgramManager(); + ~ProgramManager(); + + void ApplyTo(OpenGLState& state); void UseProgrammableVertexShader(GLuint program) { - vs = program; + current_state.vertex_shader = program; } void UseProgrammableGeometryShader(GLuint program) { - gs = program; + current_state.geometry_shader = program; } void UseProgrammableFragmentShader(GLuint program) { - fs = program; + current_state.fragment_shader = program; } void UseTrivialGeometryShader() { - gs = 0; - } - - void ApplyTo(OpenGLState& state) { - UpdatePipeline(); - state.draw.shader_program = 0; - state.draw.program_pipeline = pipeline.handle; - state.geometry_shaders.enabled = (gs != 0); + current_state.geometry_shader = 0; } private: - void UpdatePipeline() { - // Avoid updating the pipeline when values have no changed - if (old_vs == vs && old_fs == fs && old_gs == gs) - return; - // Workaround for AMD bug - glUseProgramStages(pipeline.handle, - GL_VERTEX_SHADER_BIT | GL_GEOMETRY_SHADER_BIT | GL_FRAGMENT_SHADER_BIT, - 0); - - glUseProgramStages(pipeline.handle, GL_VERTEX_SHADER_BIT, vs); - glUseProgramStages(pipeline.handle, GL_GEOMETRY_SHADER_BIT, gs); - glUseProgramStages(pipeline.handle, GL_FRAGMENT_SHADER_BIT, fs); - - // Update the old values - old_vs = vs; - old_fs = fs; - old_gs = gs; - } + struct PipelineState { + bool operator==(const PipelineState& rhs) const { + return vertex_shader == rhs.vertex_shader && fragment_shader == rhs.fragment_shader && + geometry_shader == rhs.geometry_shader; + } + + bool operator!=(const PipelineState& rhs) const { + return !operator==(rhs); + } + + GLuint vertex_shader{}; + GLuint fragment_shader{}; + GLuint geometry_shader{}; + }; + + void UpdatePipeline(); OGLPipeline pipeline; - GLuint vs{}, fs{}, gs{}; - GLuint old_vs{}, old_fs{}, old_gs{}; + PipelineState current_state; + PipelineState old_state; }; } // namespace OpenGL::GLShader diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp index 9419326a3..52d569a1b 100644 --- a/src/video_core/renderer_opengl/gl_state.cpp +++ b/src/video_core/renderer_opengl/gl_state.cpp @@ -10,16 +10,62 @@ namespace OpenGL { -OpenGLState OpenGLState::cur_state; +using Maxwell = Tegra::Engines::Maxwell3D::Regs; +OpenGLState OpenGLState::cur_state; bool OpenGLState::s_rgb_used; +namespace { + +template <typename T> +bool UpdateValue(T& current_value, const T new_value) { + const bool changed = current_value != new_value; + current_value = new_value; + return changed; +} + +template <typename T1, typename T2> +bool UpdateTie(T1 current_value, const T2 new_value) { + const bool changed = current_value != new_value; + current_value = new_value; + return changed; +} + +void Enable(GLenum cap, bool enable) { + if (enable) { + glEnable(cap); + } else { + glDisable(cap); + } +} + +void Enable(GLenum cap, GLuint index, bool enable) { + if (enable) { + glEnablei(cap, index); + } else { + glDisablei(cap, index); + } +} + +void Enable(GLenum cap, bool& current_value, bool new_value) { + if (UpdateValue(current_value, new_value)) + Enable(cap, new_value); +} + +void Enable(GLenum cap, GLuint index, bool& current_value, bool new_value) { + if (UpdateValue(current_value, new_value)) + Enable(cap, index, new_value); +} + +} // namespace + OpenGLState::OpenGLState() { // These all match default OpenGL values - geometry_shaders.enabled = false; framebuffer_srgb.enabled = false; + multisample_control.alpha_to_coverage = false; multisample_control.alpha_to_one = false; + cull.enabled = false; cull.mode = GL_BACK; cull.front_face = GL_CCW; @@ -30,14 +76,15 @@ OpenGLState::OpenGLState() { primitive_restart.enabled = false; primitive_restart.index = 0; + for (auto& item : color_mask) { item.red_enabled = GL_TRUE; item.green_enabled = GL_TRUE; item.blue_enabled = GL_TRUE; item.alpha_enabled = GL_TRUE; } - stencil.test_enabled = false; - auto reset_stencil = [](auto& config) { + + const auto ResetStencil = [](auto& config) { config.test_func = GL_ALWAYS; config.test_ref = 0; config.test_mask = 0xFFFFFFFF; @@ -46,8 +93,10 @@ OpenGLState::OpenGLState() { config.action_depth_pass = GL_KEEP; config.action_stencil_fail = GL_KEEP; }; - reset_stencil(stencil.front); - reset_stencil(stencil.back); + stencil.test_enabled = false; + ResetStencil(stencil.front); + ResetStencil(stencil.back); + for (auto& item : viewports) { item.x = 0; item.y = 0; @@ -61,6 +110,7 @@ OpenGLState::OpenGLState() { item.scissor.width = 0; item.scissor.height = 0; } + for (auto& item : blend) { item.enabled = true; item.rgb_equation = GL_FUNC_ADD; @@ -70,11 +120,14 @@ OpenGLState::OpenGLState() { item.src_a_func = GL_ONE; item.dst_a_func = GL_ZERO; } + independant_blend.enabled = false; + blend_color.red = 0.0f; blend_color.green = 0.0f; blend_color.blue = 0.0f; blend_color.alpha = 0.0f; + logic_op.enabled = false; logic_op.operation = GL_COPY; @@ -91,9 +144,12 @@ OpenGLState::OpenGLState() { clip_distance = {}; point.size = 1; + fragment_color_clamp.enabled = false; + depth_clamp.far_plane = false; depth_clamp.near_plane = false; + polygon_offset.fill_enable = false; polygon_offset.line_enable = false; polygon_offset.point_enable = false; @@ -103,260 +159,255 @@ OpenGLState::OpenGLState() { } void OpenGLState::ApplyDefaultState() { + glEnable(GL_BLEND); glDisable(GL_FRAMEBUFFER_SRGB); glDisable(GL_CULL_FACE); glDisable(GL_DEPTH_TEST); glDisable(GL_PRIMITIVE_RESTART); glDisable(GL_STENCIL_TEST); - glEnable(GL_BLEND); glDisable(GL_COLOR_LOGIC_OP); glDisable(GL_SCISSOR_TEST); } +void OpenGLState::ApplyFramebufferState() const { + if (UpdateValue(cur_state.draw.read_framebuffer, draw.read_framebuffer)) { + glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer); + } + if (UpdateValue(cur_state.draw.draw_framebuffer, draw.draw_framebuffer)) { + glBindFramebuffer(GL_DRAW_FRAMEBUFFER, draw.draw_framebuffer); + } +} + +void OpenGLState::ApplyVertexArrayState() const { + if (UpdateValue(cur_state.draw.vertex_array, draw.vertex_array)) { + glBindVertexArray(draw.vertex_array); + } +} + +void OpenGLState::ApplyShaderProgram() const { + if (UpdateValue(cur_state.draw.shader_program, draw.shader_program)) { + glUseProgram(draw.shader_program); + } +} + +void OpenGLState::ApplyProgramPipeline() const { + if (UpdateValue(cur_state.draw.program_pipeline, draw.program_pipeline)) { + glBindProgramPipeline(draw.program_pipeline); + } +} + +void OpenGLState::ApplyClipDistances() const { + for (std::size_t i = 0; i < clip_distance.size(); ++i) { + Enable(GL_CLIP_DISTANCE0 + static_cast<GLenum>(i), cur_state.clip_distance[i], + clip_distance[i]); + } +} + +void OpenGLState::ApplyPointSize() const { + if (UpdateValue(cur_state.point.size, point.size)) { + glPointSize(point.size); + } +} + +void OpenGLState::ApplyFragmentColorClamp() const { + if (UpdateValue(cur_state.fragment_color_clamp.enabled, fragment_color_clamp.enabled)) { + glClampColor(GL_CLAMP_FRAGMENT_COLOR_ARB, + fragment_color_clamp.enabled ? GL_TRUE : GL_FALSE); + } +} + +void OpenGLState::ApplyMultisample() const { + Enable(GL_SAMPLE_ALPHA_TO_COVERAGE, cur_state.multisample_control.alpha_to_coverage, + multisample_control.alpha_to_coverage); + Enable(GL_SAMPLE_ALPHA_TO_ONE, cur_state.multisample_control.alpha_to_one, + multisample_control.alpha_to_one); +} + +void OpenGLState::ApplyDepthClamp() const { + if (depth_clamp.far_plane == cur_state.depth_clamp.far_plane && + depth_clamp.near_plane == cur_state.depth_clamp.near_plane) { + return; + } + cur_state.depth_clamp = depth_clamp; + + UNIMPLEMENTED_IF_MSG(depth_clamp.far_plane != depth_clamp.near_plane, + "Unimplemented Depth Clamp Separation!"); + + Enable(GL_DEPTH_CLAMP, depth_clamp.far_plane || depth_clamp.near_plane); +} + void OpenGLState::ApplySRgb() const { - if (framebuffer_srgb.enabled != cur_state.framebuffer_srgb.enabled) { - if (framebuffer_srgb.enabled) { - // Track if sRGB is used - s_rgb_used = true; - glEnable(GL_FRAMEBUFFER_SRGB); - } else { - glDisable(GL_FRAMEBUFFER_SRGB); - } + if (cur_state.framebuffer_srgb.enabled == framebuffer_srgb.enabled) + return; + cur_state.framebuffer_srgb.enabled = framebuffer_srgb.enabled; + if (framebuffer_srgb.enabled) { + // Track if sRGB is used + s_rgb_used = true; + glEnable(GL_FRAMEBUFFER_SRGB); + } else { + glDisable(GL_FRAMEBUFFER_SRGB); } } void OpenGLState::ApplyCulling() const { - if (cull.enabled != cur_state.cull.enabled) { - if (cull.enabled) { - glEnable(GL_CULL_FACE); - } else { - glDisable(GL_CULL_FACE); - } - } + Enable(GL_CULL_FACE, cur_state.cull.enabled, cull.enabled); - if (cull.mode != cur_state.cull.mode) { + if (UpdateValue(cur_state.cull.mode, cull.mode)) { glCullFace(cull.mode); } - if (cull.front_face != cur_state.cull.front_face) { + if (UpdateValue(cur_state.cull.front_face, cull.front_face)) { glFrontFace(cull.front_face); } } void OpenGLState::ApplyColorMask() const { - if (independant_blend.enabled) { - for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { - const auto& updated = color_mask[i]; - const auto& current = cur_state.color_mask[i]; - if (updated.red_enabled != current.red_enabled || - updated.green_enabled != current.green_enabled || - updated.blue_enabled != current.blue_enabled || - updated.alpha_enabled != current.alpha_enabled) { - glColorMaski(static_cast<GLuint>(i), updated.red_enabled, updated.green_enabled, - updated.blue_enabled, updated.alpha_enabled); - } - } - } else { - const auto& updated = color_mask[0]; - const auto& current = cur_state.color_mask[0]; + for (std::size_t i = 0; i < Maxwell::NumRenderTargets; ++i) { + const auto& updated = color_mask[i]; + auto& current = cur_state.color_mask[i]; if (updated.red_enabled != current.red_enabled || updated.green_enabled != current.green_enabled || updated.blue_enabled != current.blue_enabled || updated.alpha_enabled != current.alpha_enabled) { - glColorMask(updated.red_enabled, updated.green_enabled, updated.blue_enabled, - updated.alpha_enabled); + current = updated; + glColorMaski(static_cast<GLuint>(i), updated.red_enabled, updated.green_enabled, + updated.blue_enabled, updated.alpha_enabled); } } } void OpenGLState::ApplyDepth() const { - if (depth.test_enabled != cur_state.depth.test_enabled) { - if (depth.test_enabled) { - glEnable(GL_DEPTH_TEST); - } else { - glDisable(GL_DEPTH_TEST); - } - } + Enable(GL_DEPTH_TEST, cur_state.depth.test_enabled, depth.test_enabled); - if (depth.test_func != cur_state.depth.test_func) { + if (cur_state.depth.test_func != depth.test_func) { + cur_state.depth.test_func = depth.test_func; glDepthFunc(depth.test_func); } - if (depth.write_mask != cur_state.depth.write_mask) { + if (cur_state.depth.write_mask != depth.write_mask) { + cur_state.depth.write_mask = depth.write_mask; glDepthMask(depth.write_mask); } } void OpenGLState::ApplyPrimitiveRestart() const { - if (primitive_restart.enabled != cur_state.primitive_restart.enabled) { - if (primitive_restart.enabled) { - glEnable(GL_PRIMITIVE_RESTART); - } else { - glDisable(GL_PRIMITIVE_RESTART); - } - } + Enable(GL_PRIMITIVE_RESTART, cur_state.primitive_restart.enabled, primitive_restart.enabled); - if (primitive_restart.index != cur_state.primitive_restart.index) { + if (cur_state.primitive_restart.index != primitive_restart.index) { + cur_state.primitive_restart.index = primitive_restart.index; glPrimitiveRestartIndex(primitive_restart.index); } } void OpenGLState::ApplyStencilTest() const { - if (stencil.test_enabled != cur_state.stencil.test_enabled) { - if (stencil.test_enabled) { - glEnable(GL_STENCIL_TEST); - } else { - glDisable(GL_STENCIL_TEST); - } - } - - const auto ConfigStencil = [](GLenum face, const auto& config, const auto& prev_config) { - if (config.test_func != prev_config.test_func || config.test_ref != prev_config.test_ref || - config.test_mask != prev_config.test_mask) { + Enable(GL_STENCIL_TEST, cur_state.stencil.test_enabled, stencil.test_enabled); + + const auto ConfigStencil = [](GLenum face, const auto& config, auto& current) { + if (current.test_func != config.test_func || current.test_ref != config.test_ref || + current.test_mask != config.test_mask) { + current.test_func = config.test_func; + current.test_ref = config.test_ref; + current.test_mask = config.test_mask; glStencilFuncSeparate(face, config.test_func, config.test_ref, config.test_mask); } - if (config.action_depth_fail != prev_config.action_depth_fail || - config.action_depth_pass != prev_config.action_depth_pass || - config.action_stencil_fail != prev_config.action_stencil_fail) { + if (current.action_depth_fail != config.action_depth_fail || + current.action_depth_pass != config.action_depth_pass || + current.action_stencil_fail != config.action_stencil_fail) { + current.action_depth_fail = config.action_depth_fail; + current.action_depth_pass = config.action_depth_pass; + current.action_stencil_fail = config.action_stencil_fail; glStencilOpSeparate(face, config.action_stencil_fail, config.action_depth_fail, config.action_depth_pass); } - if (config.write_mask != prev_config.write_mask) { + if (current.write_mask != config.write_mask) { + current.write_mask = config.write_mask; glStencilMaskSeparate(face, config.write_mask); } }; ConfigStencil(GL_FRONT, stencil.front, cur_state.stencil.front); ConfigStencil(GL_BACK, stencil.back, cur_state.stencil.back); } -// Viewport does not affects glClearBuffer so emulate viewport using scissor test -void OpenGLState::EmulateViewportWithScissor() { - auto& current = viewports[0]; - if (current.scissor.enabled) { - const GLint left = std::max(current.x, current.scissor.x); - const GLint right = - std::max(current.x + current.width, current.scissor.x + current.scissor.width); - const GLint bottom = std::max(current.y, current.scissor.y); - const GLint top = - std::max(current.y + current.height, current.scissor.y + current.scissor.height); - current.scissor.x = std::max(left, 0); - current.scissor.y = std::max(bottom, 0); - current.scissor.width = std::max(right - left, 0); - current.scissor.height = std::max(top - bottom, 0); - } else { - current.scissor.enabled = true; - current.scissor.x = current.x; - current.scissor.y = current.y; - current.scissor.width = current.width; - current.scissor.height = current.height; - } -} void OpenGLState::ApplyViewport() const { - if (geometry_shaders.enabled) { - for (GLuint i = 0; i < static_cast<GLuint>(Tegra::Engines::Maxwell3D::Regs::NumViewports); - i++) { - const auto& current = cur_state.viewports[i]; - const auto& updated = viewports[i]; - if (updated.x != current.x || updated.y != current.y || - updated.width != current.width || updated.height != current.height) { - glViewportIndexedf( - i, static_cast<GLfloat>(updated.x), static_cast<GLfloat>(updated.y), - static_cast<GLfloat>(updated.width), static_cast<GLfloat>(updated.height)); - } - if (updated.depth_range_near != current.depth_range_near || - updated.depth_range_far != current.depth_range_far) { - glDepthRangeIndexed(i, updated.depth_range_near, updated.depth_range_far); - } - - if (updated.scissor.enabled != current.scissor.enabled) { - if (updated.scissor.enabled) { - glEnablei(GL_SCISSOR_TEST, i); - } else { - glDisablei(GL_SCISSOR_TEST, i); - } - } - - if (updated.scissor.x != current.scissor.x || updated.scissor.y != current.scissor.y || - updated.scissor.width != current.scissor.width || - updated.scissor.height != current.scissor.height) { - glScissorIndexed(i, updated.scissor.x, updated.scissor.y, updated.scissor.width, - updated.scissor.height); - } - } - } else { - const auto& current = cur_state.viewports[0]; - const auto& updated = viewports[0]; - if (updated.x != current.x || updated.y != current.y || updated.width != current.width || - updated.height != current.height) { - glViewport(updated.x, updated.y, updated.width, updated.height); - } - - if (updated.depth_range_near != current.depth_range_near || - updated.depth_range_far != current.depth_range_far) { - glDepthRange(updated.depth_range_near, updated.depth_range_far); + for (GLuint i = 0; i < static_cast<GLuint>(Maxwell::NumViewports); ++i) { + const auto& updated = viewports[i]; + auto& current = cur_state.viewports[i]; + + if (current.x != updated.x || current.y != updated.y || current.width != updated.width || + current.height != updated.height) { + current.x = updated.x; + current.y = updated.y; + current.width = updated.width; + current.height = updated.height; + glViewportIndexedf(i, static_cast<GLfloat>(updated.x), static_cast<GLfloat>(updated.y), + static_cast<GLfloat>(updated.width), + static_cast<GLfloat>(updated.height)); } - - if (updated.scissor.enabled != current.scissor.enabled) { - if (updated.scissor.enabled) { - glEnable(GL_SCISSOR_TEST); - } else { - glDisable(GL_SCISSOR_TEST); - } + if (current.depth_range_near != updated.depth_range_near || + current.depth_range_far != updated.depth_range_far) { + current.depth_range_near = updated.depth_range_near; + current.depth_range_far = updated.depth_range_far; + glDepthRangeIndexed(i, updated.depth_range_near, updated.depth_range_far); } - if (updated.scissor.x != current.scissor.x || updated.scissor.y != current.scissor.y || - updated.scissor.width != current.scissor.width || - updated.scissor.height != current.scissor.height) { - glScissor(updated.scissor.x, updated.scissor.y, updated.scissor.width, - updated.scissor.height); + Enable(GL_SCISSOR_TEST, i, current.scissor.enabled, updated.scissor.enabled); + + if (current.scissor.x != updated.scissor.x || current.scissor.y != updated.scissor.y || + current.scissor.width != updated.scissor.width || + current.scissor.height != updated.scissor.height) { + current.scissor.x = updated.scissor.x; + current.scissor.y = updated.scissor.y; + current.scissor.width = updated.scissor.width; + current.scissor.height = updated.scissor.height; + glScissorIndexed(i, updated.scissor.x, updated.scissor.y, updated.scissor.width, + updated.scissor.height); } } } void OpenGLState::ApplyGlobalBlending() const { - const Blend& current = cur_state.blend[0]; const Blend& updated = blend[0]; - if (updated.enabled != current.enabled) { - if (updated.enabled) { - glEnable(GL_BLEND); - } else { - glDisable(GL_BLEND); - } - } - if (!updated.enabled) { - return; - } - if (updated.src_rgb_func != current.src_rgb_func || - updated.dst_rgb_func != current.dst_rgb_func || updated.src_a_func != current.src_a_func || - updated.dst_a_func != current.dst_a_func) { + Blend& current = cur_state.blend[0]; + + Enable(GL_BLEND, current.enabled, updated.enabled); + + if (current.src_rgb_func != updated.src_rgb_func || + current.dst_rgb_func != updated.dst_rgb_func || current.src_a_func != updated.src_a_func || + current.dst_a_func != updated.dst_a_func) { + current.src_rgb_func = updated.src_rgb_func; + current.dst_rgb_func = updated.dst_rgb_func; + current.src_a_func = updated.src_a_func; + current.dst_a_func = updated.dst_a_func; glBlendFuncSeparate(updated.src_rgb_func, updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func); } - if (updated.rgb_equation != current.rgb_equation || updated.a_equation != current.a_equation) { + if (current.rgb_equation != updated.rgb_equation || current.a_equation != updated.a_equation) { + current.rgb_equation = updated.rgb_equation; + current.a_equation = updated.a_equation; glBlendEquationSeparate(updated.rgb_equation, updated.a_equation); } } void OpenGLState::ApplyTargetBlending(std::size_t target, bool force) const { const Blend& updated = blend[target]; - const Blend& current = cur_state.blend[target]; - if (updated.enabled != current.enabled || force) { - if (updated.enabled) { - glEnablei(GL_BLEND, static_cast<GLuint>(target)); - } else { - glDisablei(GL_BLEND, static_cast<GLuint>(target)); - } + Blend& current = cur_state.blend[target]; + + if (current.enabled != updated.enabled || force) { + current.enabled = updated.enabled; + Enable(GL_BLEND, static_cast<GLuint>(target), updated.enabled); } - if (updated.src_rgb_func != current.src_rgb_func || - updated.dst_rgb_func != current.dst_rgb_func || updated.src_a_func != current.src_a_func || - updated.dst_a_func != current.dst_a_func) { + if (UpdateTie(std::tie(current.src_rgb_func, current.dst_rgb_func, current.src_a_func, + current.dst_a_func), + std::tie(updated.src_rgb_func, updated.dst_rgb_func, updated.src_a_func, + updated.dst_a_func))) { glBlendFuncSeparatei(static_cast<GLuint>(target), updated.src_rgb_func, updated.dst_rgb_func, updated.src_a_func, updated.dst_a_func); } - if (updated.rgb_equation != current.rgb_equation || updated.a_equation != current.a_equation) { + if (UpdateTie(std::tie(current.rgb_equation, current.a_equation), + std::tie(updated.rgb_equation, updated.a_equation))) { glBlendEquationSeparatei(static_cast<GLuint>(target), updated.rgb_equation, updated.a_equation); } @@ -364,77 +415,48 @@ void OpenGLState::ApplyTargetBlending(std::size_t target, bool force) const { void OpenGLState::ApplyBlending() const { if (independant_blend.enabled) { - for (size_t i = 0; i < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets; i++) { - ApplyTargetBlending(i, - independant_blend.enabled != cur_state.independant_blend.enabled); + const bool force = independant_blend.enabled != cur_state.independant_blend.enabled; + for (std::size_t target = 0; target < Maxwell::NumRenderTargets; ++target) { + ApplyTargetBlending(target, force); } } else { ApplyGlobalBlending(); } - if (blend_color.red != cur_state.blend_color.red || - blend_color.green != cur_state.blend_color.green || - blend_color.blue != cur_state.blend_color.blue || - blend_color.alpha != cur_state.blend_color.alpha) { + cur_state.independant_blend.enabled = independant_blend.enabled; + + if (UpdateTie( + std::tie(cur_state.blend_color.red, cur_state.blend_color.green, + cur_state.blend_color.blue, cur_state.blend_color.alpha), + std::tie(blend_color.red, blend_color.green, blend_color.blue, blend_color.alpha))) { glBlendColor(blend_color.red, blend_color.green, blend_color.blue, blend_color.alpha); } } void OpenGLState::ApplyLogicOp() const { - if (logic_op.enabled != cur_state.logic_op.enabled) { - if (logic_op.enabled) { - glEnable(GL_COLOR_LOGIC_OP); - } else { - glDisable(GL_COLOR_LOGIC_OP); - } - } + Enable(GL_COLOR_LOGIC_OP, cur_state.logic_op.enabled, logic_op.enabled); - if (logic_op.operation != cur_state.logic_op.operation) { + if (UpdateValue(cur_state.logic_op.operation, logic_op.operation)) { glLogicOp(logic_op.operation); } } void OpenGLState::ApplyPolygonOffset() const { - const bool fill_enable_changed = - polygon_offset.fill_enable != cur_state.polygon_offset.fill_enable; - const bool line_enable_changed = - polygon_offset.line_enable != cur_state.polygon_offset.line_enable; - const bool point_enable_changed = - polygon_offset.point_enable != cur_state.polygon_offset.point_enable; - const bool factor_changed = polygon_offset.factor != cur_state.polygon_offset.factor; - const bool units_changed = polygon_offset.units != cur_state.polygon_offset.units; - const bool clamp_changed = polygon_offset.clamp != cur_state.polygon_offset.clamp; - - if (fill_enable_changed) { - if (polygon_offset.fill_enable) { - glEnable(GL_POLYGON_OFFSET_FILL); - } else { - glDisable(GL_POLYGON_OFFSET_FILL); - } - } - - if (line_enable_changed) { - if (polygon_offset.line_enable) { - glEnable(GL_POLYGON_OFFSET_LINE); - } else { - glDisable(GL_POLYGON_OFFSET_LINE); - } - } - - if (point_enable_changed) { - if (polygon_offset.point_enable) { - glEnable(GL_POLYGON_OFFSET_POINT); - } else { - glDisable(GL_POLYGON_OFFSET_POINT); - } - } - - if (factor_changed || units_changed || clamp_changed) { + Enable(GL_POLYGON_OFFSET_FILL, cur_state.polygon_offset.fill_enable, + polygon_offset.fill_enable); + Enable(GL_POLYGON_OFFSET_LINE, cur_state.polygon_offset.line_enable, + polygon_offset.line_enable); + Enable(GL_POLYGON_OFFSET_POINT, cur_state.polygon_offset.point_enable, + polygon_offset.point_enable); + + if (UpdateTie(std::tie(cur_state.polygon_offset.factor, cur_state.polygon_offset.units, + cur_state.polygon_offset.clamp), + std::tie(polygon_offset.factor, polygon_offset.units, polygon_offset.clamp))) { if (GLAD_GL_EXT_polygon_offset_clamp && polygon_offset.clamp != 0) { glPolygonOffsetClamp(polygon_offset.factor, polygon_offset.units, polygon_offset.clamp); } else { - glPolygonOffset(polygon_offset.factor, polygon_offset.units); UNIMPLEMENTED_IF_MSG(polygon_offset.clamp != 0, "Unimplemented Depth polygon offset clamp."); + glPolygonOffset(polygon_offset.factor, polygon_offset.units); } } } @@ -443,22 +465,21 @@ void OpenGLState::ApplyTextures() const { bool has_delta{}; std::size_t first{}; std::size_t last{}; - std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> textures; + std::array<GLuint, Maxwell::NumTextureSamplers> textures; for (std::size_t i = 0; i < std::size(texture_units); ++i) { const auto& texture_unit = texture_units[i]; - const auto& cur_state_texture_unit = cur_state.texture_units[i]; + auto& cur_state_texture_unit = cur_state.texture_units[i]; textures[i] = texture_unit.texture; - - if (textures[i] != cur_state_texture_unit.texture) { - if (!has_delta) { - first = i; - has_delta = true; - } - last = i; + if (cur_state_texture_unit.texture == textures[i]) + continue; + cur_state_texture_unit.texture = textures[i]; + if (!has_delta) { + first = i; + has_delta = true; } + last = i; } - if (has_delta) { glBindTextures(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1), textures.data() + first); @@ -469,16 +490,18 @@ void OpenGLState::ApplySamplers() const { bool has_delta{}; std::size_t first{}; std::size_t last{}; - std::array<GLuint, Tegra::Engines::Maxwell3D::Regs::NumTextureSamplers> samplers; + std::array<GLuint, Maxwell::NumTextureSamplers> samplers; + for (std::size_t i = 0; i < std::size(samplers); ++i) { + if (cur_state.texture_units[i].sampler == texture_units[i].sampler) + continue; + cur_state.texture_units[i].sampler = texture_units[i].sampler; samplers[i] = texture_units[i].sampler; - if (samplers[i] != cur_state.texture_units[i].sampler) { - if (!has_delta) { - first = i; - has_delta = true; - } - last = i; + if (!has_delta) { + first = i; + has_delta = true; } + last = i; } if (has_delta) { glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1), @@ -486,81 +509,15 @@ void OpenGLState::ApplySamplers() const { } } -void OpenGLState::ApplyFramebufferState() const { - if (draw.read_framebuffer != cur_state.draw.read_framebuffer) { - glBindFramebuffer(GL_READ_FRAMEBUFFER, draw.read_framebuffer); - } - if (draw.draw_framebuffer != cur_state.draw.draw_framebuffer) { - glBindFramebuffer(GL_DRAW_FRAMEBUFFER, draw.draw_framebuffer); - } -} - -void OpenGLState::ApplyVertexArrayState() const { - if (draw.vertex_array != cur_state.draw.vertex_array) { - glBindVertexArray(draw.vertex_array); - } -} - -void OpenGLState::ApplyDepthClamp() const { - if (depth_clamp.far_plane == cur_state.depth_clamp.far_plane && - depth_clamp.near_plane == cur_state.depth_clamp.near_plane) { - return; - } - UNIMPLEMENTED_IF_MSG(depth_clamp.far_plane != depth_clamp.near_plane, - "Unimplemented Depth Clamp Separation!"); - - if (depth_clamp.far_plane || depth_clamp.near_plane) { - glEnable(GL_DEPTH_CLAMP); - } else { - glDisable(GL_DEPTH_CLAMP); - } -} - void OpenGLState::Apply() const { ApplyFramebufferState(); ApplyVertexArrayState(); - - // Shader program - if (draw.shader_program != cur_state.draw.shader_program) { - glUseProgram(draw.shader_program); - } - - // Program pipeline - if (draw.program_pipeline != cur_state.draw.program_pipeline) { - glBindProgramPipeline(draw.program_pipeline); - } - // Clip distance - for (std::size_t i = 0; i < clip_distance.size(); ++i) { - if (clip_distance[i] != cur_state.clip_distance[i]) { - if (clip_distance[i]) { - glEnable(GL_CLIP_DISTANCE0 + static_cast<GLenum>(i)); - } else { - glDisable(GL_CLIP_DISTANCE0 + static_cast<GLenum>(i)); - } - } - } - // Point - if (point.size != cur_state.point.size) { - glPointSize(point.size); - } - if (fragment_color_clamp.enabled != cur_state.fragment_color_clamp.enabled) { - glClampColor(GL_CLAMP_FRAGMENT_COLOR_ARB, - fragment_color_clamp.enabled ? GL_TRUE : GL_FALSE); - } - if (multisample_control.alpha_to_coverage != cur_state.multisample_control.alpha_to_coverage) { - if (multisample_control.alpha_to_coverage) { - glEnable(GL_SAMPLE_ALPHA_TO_COVERAGE); - } else { - glDisable(GL_SAMPLE_ALPHA_TO_COVERAGE); - } - } - if (multisample_control.alpha_to_one != cur_state.multisample_control.alpha_to_one) { - if (multisample_control.alpha_to_one) { - glEnable(GL_SAMPLE_ALPHA_TO_ONE); - } else { - glDisable(GL_SAMPLE_ALPHA_TO_ONE); - } - } + ApplyShaderProgram(); + ApplyProgramPipeline(); + ApplyClipDistances(); + ApplyPointSize(); + ApplyFragmentColorClamp(); + ApplyMultisample(); ApplyDepthClamp(); ApplyColorMask(); ApplyViewport(); @@ -574,7 +531,28 @@ void OpenGLState::Apply() const { ApplyTextures(); ApplySamplers(); ApplyPolygonOffset(); - cur_state = *this; +} + +void OpenGLState::EmulateViewportWithScissor() { + auto& current = viewports[0]; + if (current.scissor.enabled) { + const GLint left = std::max(current.x, current.scissor.x); + const GLint right = + std::max(current.x + current.width, current.scissor.x + current.scissor.width); + const GLint bottom = std::max(current.y, current.scissor.y); + const GLint top = + std::max(current.y + current.height, current.scissor.y + current.scissor.height); + current.scissor.x = std::max(left, 0); + current.scissor.y = std::max(bottom, 0); + current.scissor.width = std::max(right - left, 0); + current.scissor.height = std::max(top - bottom, 0); + } else { + current.scissor.enabled = true; + current.scissor.x = current.x; + current.scissor.y = current.y; + current.scissor.width = current.width; + current.scissor.height = current.height; + } } OpenGLState& OpenGLState::UnbindTexture(GLuint handle) { diff --git a/src/video_core/renderer_opengl/gl_state.h b/src/video_core/renderer_opengl/gl_state.h index 9e1eda5b1..41418a7b8 100644 --- a/src/video_core/renderer_opengl/gl_state.h +++ b/src/video_core/renderer_opengl/gl_state.h @@ -54,10 +54,6 @@ public: } depth_clamp; // GL_DEPTH_CLAMP struct { - bool enabled; // viewports arrays are only supported when geometry shaders are enabled. - } geometry_shaders; - - struct { bool enabled; // GL_CULL_FACE GLenum mode; // GL_CULL_FACE_MODE GLenum front_face; // GL_FRONT_FACE @@ -184,34 +180,26 @@ public: static OpenGLState GetCurState() { return cur_state; } + static bool GetsRGBUsed() { return s_rgb_used; } + static void ClearsRGBUsed() { s_rgb_used = false; } + /// Apply this state as the current OpenGL state void Apply() const; - /// Apply only the state affecting the framebuffer + void ApplyFramebufferState() const; - /// Apply only the state affecting the vertex array void ApplyVertexArrayState() const; - /// Set the initial OpenGL state - static void ApplyDefaultState(); - /// Resets any references to the given resource - OpenGLState& UnbindTexture(GLuint handle); - OpenGLState& ResetSampler(GLuint handle); - OpenGLState& ResetProgram(GLuint handle); - OpenGLState& ResetPipeline(GLuint handle); - OpenGLState& ResetVertexArray(GLuint handle); - OpenGLState& ResetFramebuffer(GLuint handle); - void EmulateViewportWithScissor(); - -private: - static OpenGLState cur_state; - // Workaround for sRGB problems caused by - // QT not supporting srgb output - static bool s_rgb_used; + void ApplyShaderProgram() const; + void ApplyProgramPipeline() const; + void ApplyClipDistances() const; + void ApplyPointSize() const; + void ApplyFragmentColorClamp() const; + void ApplyMultisample() const; void ApplySRgb() const; void ApplyCulling() const; void ApplyColorMask() const; @@ -227,6 +215,26 @@ private: void ApplySamplers() const; void ApplyDepthClamp() const; void ApplyPolygonOffset() const; + + /// Set the initial OpenGL state + static void ApplyDefaultState(); + + /// Resets any references to the given resource + OpenGLState& UnbindTexture(GLuint handle); + OpenGLState& ResetSampler(GLuint handle); + OpenGLState& ResetProgram(GLuint handle); + OpenGLState& ResetPipeline(GLuint handle); + OpenGLState& ResetVertexArray(GLuint handle); + OpenGLState& ResetFramebuffer(GLuint handle); + + /// Viewport does not affects glClearBuffer so emulate viewport using scissor test + void EmulateViewportWithScissor(); + +private: + static OpenGLState cur_state; + + // Workaround for sRGB problems caused by QT not supporting srgb output + static bool s_rgb_used; }; } // namespace OpenGL diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp index a01efeb05..d69cba9c3 100644 --- a/src/video_core/renderer_opengl/renderer_opengl.cpp +++ b/src/video_core/renderer_opengl/renderer_opengl.cpp @@ -5,7 +5,6 @@ #include <algorithm> #include <cstddef> #include <cstdlib> -#include <cstring> #include <memory> #include <glad/glad.h> #include "common/assert.h" diff --git a/src/video_core/renderer_opengl/utils.cpp b/src/video_core/renderer_opengl/utils.cpp index d84634cb3..84a987371 100644 --- a/src/video_core/renderer_opengl/utils.cpp +++ b/src/video_core/renderer_opengl/utils.cpp @@ -5,11 +5,39 @@ #include <string> #include <fmt/format.h> #include <glad/glad.h> +#include "common/assert.h" #include "common/common_types.h" #include "video_core/renderer_opengl/utils.h" namespace OpenGL { +BindBuffersRangePushBuffer::BindBuffersRangePushBuffer(GLenum target) : target{target} {} + +BindBuffersRangePushBuffer::~BindBuffersRangePushBuffer() = default; + +void BindBuffersRangePushBuffer::Setup(GLuint first_) { + first = first_; + buffers.clear(); + offsets.clear(); + sizes.clear(); +} + +void BindBuffersRangePushBuffer::Push(GLuint buffer, GLintptr offset, GLsizeiptr size) { + buffers.push_back(buffer); + offsets.push_back(offset); + sizes.push_back(size); +} + +void BindBuffersRangePushBuffer::Bind() const { + const std::size_t count{buffers.size()}; + DEBUG_ASSERT(count == offsets.size() && count == sizes.size()); + if (count == 0) { + return; + } + glBindBuffersRange(target, first, static_cast<GLsizei>(count), buffers.data(), offsets.data(), + sizes.data()); +} + void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string extra_info) { if (!GLAD_GL_KHR_debug) { return; // We don't need to throw an error as this is just for debugging diff --git a/src/video_core/renderer_opengl/utils.h b/src/video_core/renderer_opengl/utils.h index 1fcb6fc11..aef45c9dc 100644 --- a/src/video_core/renderer_opengl/utils.h +++ b/src/video_core/renderer_opengl/utils.h @@ -5,11 +5,31 @@ #pragma once #include <string> +#include <vector> #include <glad/glad.h> #include "common/common_types.h" namespace OpenGL { +class BindBuffersRangePushBuffer { +public: + BindBuffersRangePushBuffer(GLenum target); + ~BindBuffersRangePushBuffer(); + + void Setup(GLuint first_); + + void Push(GLuint buffer, GLintptr offset, GLsizeiptr size); + + void Bind() const; + +private: + GLenum target; + GLuint first; + std::vector<GLuint> buffers; + std::vector<GLintptr> offsets; + std::vector<GLsizeiptr> sizes; +}; + void LabelGLObject(GLenum identifier, GLuint handle, VAddr addr, std::string extra_info = ""); } // namespace OpenGL
\ No newline at end of file diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp index 388b5ffd5..02a9f5ecb 100644 --- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp @@ -10,6 +10,7 @@ #include "common/alignment.h" #include "common/assert.h" #include "core/memory.h" +#include "video_core/memory_manager.h" #include "video_core/renderer_vulkan/declarations.h" #include "video_core/renderer_vulkan/vk_buffer_cache.h" #include "video_core/renderer_vulkan/vk_scheduler.h" diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp new file mode 100644 index 000000000..25500f9a3 --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp @@ -0,0 +1,1381 @@ +// Copyright 2019 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <functional> +#include <map> +#include <set> + +#include <fmt/format.h> + +#include <sirit/sirit.h> + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/common_types.h" +#include "common/logging/log.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/engines/shader_bytecode.h" +#include "video_core/engines/shader_header.h" +#include "video_core/renderer_vulkan/vk_shader_decompiler.h" +#include "video_core/shader/shader_ir.h" + +namespace Vulkan::VKShader { + +using Sirit::Id; +using Tegra::Shader::Attribute; +using Tegra::Shader::AttributeUse; +using Tegra::Shader::Register; +using namespace VideoCommon::Shader; + +using Maxwell = Tegra::Engines::Maxwell3D::Regs; +using ShaderStage = Tegra::Engines::Maxwell3D::Regs::ShaderStage; +using Operation = const OperationNode&; + +// TODO(Rodrigo): Use rasterizer's value +constexpr u32 MAX_CONSTBUFFER_ELEMENTS = 0x1000; +constexpr u32 STAGE_BINDING_STRIDE = 0x100; + +enum class Type { Bool, Bool2, Float, Int, Uint, HalfFloat }; + +struct SamplerImage { + Id image_type; + Id sampled_image_type; + Id sampler; +}; + +namespace { + +spv::Dim GetSamplerDim(const Sampler& sampler) { + switch (sampler.GetType()) { + case Tegra::Shader::TextureType::Texture1D: + return spv::Dim::Dim1D; + case Tegra::Shader::TextureType::Texture2D: + return spv::Dim::Dim2D; + case Tegra::Shader::TextureType::Texture3D: + return spv::Dim::Dim3D; + case Tegra::Shader::TextureType::TextureCube: + return spv::Dim::Cube; + default: + UNIMPLEMENTED_MSG("Unimplemented sampler type={}", static_cast<u32>(sampler.GetType())); + return spv::Dim::Dim2D; + } +} + +/// Returns true if an attribute index is one of the 32 generic attributes +constexpr bool IsGenericAttribute(Attribute::Index attribute) { + return attribute >= Attribute::Index::Attribute_0 && + attribute <= Attribute::Index::Attribute_31; +} + +/// Returns the location of a generic attribute +constexpr u32 GetGenericAttributeLocation(Attribute::Index attribute) { + ASSERT(IsGenericAttribute(attribute)); + return static_cast<u32>(attribute) - static_cast<u32>(Attribute::Index::Attribute_0); +} + +/// Returns true if an object has to be treated as precise +bool IsPrecise(Operation operand) { + const auto& meta = operand.GetMeta(); + + if (std::holds_alternative<MetaArithmetic>(meta)) { + return std::get<MetaArithmetic>(meta).precise; + } + if (std::holds_alternative<MetaHalfArithmetic>(meta)) { + return std::get<MetaHalfArithmetic>(meta).precise; + } + return false; +} + +} // namespace + +class SPIRVDecompiler : public Sirit::Module { +public: + explicit SPIRVDecompiler(const ShaderIR& ir, ShaderStage stage) + : Module(0x00010300), ir{ir}, stage{stage}, header{ir.GetHeader()} { + AddCapability(spv::Capability::Shader); + AddExtension("SPV_KHR_storage_buffer_storage_class"); + AddExtension("SPV_KHR_variable_pointers"); + } + + void Decompile() { + AllocateBindings(); + AllocateLabels(); + + DeclareVertex(); + DeclareGeometry(); + DeclareFragment(); + DeclareRegisters(); + DeclarePredicates(); + DeclareLocalMemory(); + DeclareInternalFlags(); + DeclareInputAttributes(); + DeclareOutputAttributes(); + DeclareConstantBuffers(); + DeclareGlobalBuffers(); + DeclareSamplers(); + + execute_function = + Emit(OpFunction(t_void, spv::FunctionControlMask::Inline, TypeFunction(t_void))); + Emit(OpLabel()); + + const u32 first_address = ir.GetBasicBlocks().begin()->first; + const Id loop_label = OpLabel("loop"); + const Id merge_label = OpLabel("merge"); + const Id dummy_label = OpLabel(); + const Id jump_label = OpLabel(); + continue_label = OpLabel("continue"); + + std::vector<Sirit::Literal> literals; + std::vector<Id> branch_labels; + for (const auto& pair : labels) { + const auto [literal, label] = pair; + literals.push_back(literal); + branch_labels.push_back(label); + } + + // TODO(Rodrigo): Figure out the actual depth of the flow stack, for now it seems unlikely + // that shaders will use 20 nested SSYs and PBKs. + constexpr u32 FLOW_STACK_SIZE = 20; + const Id flow_stack_type = TypeArray(t_uint, Constant(t_uint, FLOW_STACK_SIZE)); + jmp_to = Emit(OpVariable(TypePointer(spv::StorageClass::Function, t_uint), + spv::StorageClass::Function, Constant(t_uint, first_address))); + flow_stack = Emit(OpVariable(TypePointer(spv::StorageClass::Function, flow_stack_type), + spv::StorageClass::Function, ConstantNull(flow_stack_type))); + flow_stack_top = + Emit(OpVariable(t_func_uint, spv::StorageClass::Function, Constant(t_uint, 0))); + + Name(jmp_to, "jmp_to"); + Name(flow_stack, "flow_stack"); + Name(flow_stack_top, "flow_stack_top"); + + Emit(OpBranch(loop_label)); + Emit(loop_label); + Emit(OpLoopMerge(merge_label, continue_label, spv::LoopControlMask::Unroll)); + Emit(OpBranch(dummy_label)); + + Emit(dummy_label); + const Id default_branch = OpLabel(); + const Id jmp_to_load = Emit(OpLoad(t_uint, jmp_to)); + Emit(OpSelectionMerge(jump_label, spv::SelectionControlMask::MaskNone)); + Emit(OpSwitch(jmp_to_load, default_branch, literals, branch_labels)); + + Emit(default_branch); + Emit(OpReturn()); + + for (const auto& pair : ir.GetBasicBlocks()) { + const auto& [address, bb] = pair; + Emit(labels.at(address)); + + VisitBasicBlock(bb); + + const auto next_it = labels.lower_bound(address + 1); + const Id next_label = next_it != labels.end() ? next_it->second : default_branch; + Emit(OpBranch(next_label)); + } + + Emit(jump_label); + Emit(OpBranch(continue_label)); + Emit(continue_label); + Emit(OpBranch(loop_label)); + Emit(merge_label); + Emit(OpReturn()); + Emit(OpFunctionEnd()); + } + + ShaderEntries GetShaderEntries() const { + ShaderEntries entries; + entries.const_buffers_base_binding = const_buffers_base_binding; + entries.global_buffers_base_binding = global_buffers_base_binding; + entries.samplers_base_binding = samplers_base_binding; + for (const auto& cbuf : ir.GetConstantBuffers()) { + entries.const_buffers.emplace_back(cbuf.second, cbuf.first); + } + for (const auto& gmem_pair : ir.GetGlobalMemory()) { + const auto& [base, usage] = gmem_pair; + entries.global_buffers.emplace_back(base.cbuf_index, base.cbuf_offset); + } + for (const auto& sampler : ir.GetSamplers()) { + entries.samplers.emplace_back(sampler); + } + for (const auto& attr : ir.GetInputAttributes()) { + entries.attributes.insert(GetGenericAttributeLocation(attr.first)); + } + entries.clip_distances = ir.GetClipDistances(); + entries.shader_length = ir.GetLength(); + entries.entry_function = execute_function; + entries.interfaces = interfaces; + return entries; + } + +private: + using OperationDecompilerFn = Id (SPIRVDecompiler::*)(Operation); + using OperationDecompilersArray = + std::array<OperationDecompilerFn, static_cast<std::size_t>(OperationCode::Amount)>; + + static constexpr auto INTERNAL_FLAGS_COUNT = static_cast<std::size_t>(InternalFlag::Amount); + static constexpr u32 CBUF_STRIDE = 16; + + void AllocateBindings() { + const u32 binding_base = static_cast<u32>(stage) * STAGE_BINDING_STRIDE; + u32 binding_iterator = binding_base; + + const auto Allocate = [&binding_iterator](std::size_t count) { + const u32 current_binding = binding_iterator; + binding_iterator += static_cast<u32>(count); + return current_binding; + }; + const_buffers_base_binding = Allocate(ir.GetConstantBuffers().size()); + global_buffers_base_binding = Allocate(ir.GetGlobalMemory().size()); + samplers_base_binding = Allocate(ir.GetSamplers().size()); + + ASSERT_MSG(binding_iterator - binding_base < STAGE_BINDING_STRIDE, + "Stage binding stride is too small"); + } + + void AllocateLabels() { + for (const auto& pair : ir.GetBasicBlocks()) { + const u32 address = pair.first; + labels.emplace(address, OpLabel(fmt::format("label_0x{:x}", address))); + } + } + + void DeclareVertex() { + if (stage != ShaderStage::Vertex) + return; + + DeclareVertexRedeclarations(); + } + + void DeclareGeometry() { + if (stage != ShaderStage::Geometry) + return; + + UNIMPLEMENTED(); + } + + void DeclareFragment() { + if (stage != ShaderStage::Fragment) + return; + + for (u32 rt = 0; rt < static_cast<u32>(frag_colors.size()); ++rt) { + if (!IsRenderTargetUsed(rt)) { + continue; + } + + const Id id = AddGlobalVariable(OpVariable(t_out_float4, spv::StorageClass::Output)); + Name(id, fmt::format("frag_color{}", rt)); + Decorate(id, spv::Decoration::Location, rt); + + frag_colors[rt] = id; + interfaces.push_back(id); + } + + if (header.ps.omap.depth) { + frag_depth = AddGlobalVariable(OpVariable(t_out_float, spv::StorageClass::Output)); + Name(frag_depth, "frag_depth"); + Decorate(frag_depth, spv::Decoration::BuiltIn, + static_cast<u32>(spv::BuiltIn::FragDepth)); + + interfaces.push_back(frag_depth); + } + + frag_coord = DeclareBuiltIn(spv::BuiltIn::FragCoord, spv::StorageClass::Input, t_in_float4, + "frag_coord"); + front_facing = DeclareBuiltIn(spv::BuiltIn::FrontFacing, spv::StorageClass::Input, + t_in_bool, "front_facing"); + } + + void DeclareRegisters() { + for (const u32 gpr : ir.GetRegisters()) { + const Id id = OpVariable(t_prv_float, spv::StorageClass::Private, v_float_zero); + Name(id, fmt::format("gpr_{}", gpr)); + registers.emplace(gpr, AddGlobalVariable(id)); + } + } + + void DeclarePredicates() { + for (const auto pred : ir.GetPredicates()) { + const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false); + Name(id, fmt::format("pred_{}", static_cast<u32>(pred))); + predicates.emplace(pred, AddGlobalVariable(id)); + } + } + + void DeclareLocalMemory() { + if (const u64 local_memory_size = header.GetLocalMemorySize(); local_memory_size > 0) { + const auto element_count = static_cast<u32>(Common::AlignUp(local_memory_size, 4) / 4); + const Id type_array = TypeArray(t_float, Constant(t_uint, element_count)); + const Id type_pointer = TypePointer(spv::StorageClass::Private, type_array); + Name(type_pointer, "LocalMemory"); + + local_memory = + OpVariable(type_pointer, spv::StorageClass::Private, ConstantNull(type_array)); + AddGlobalVariable(Name(local_memory, "local_memory")); + } + } + + void DeclareInternalFlags() { + constexpr std::array<const char*, INTERNAL_FLAGS_COUNT> names = {"zero", "sign", "carry", + "overflow"}; + for (std::size_t flag = 0; flag < INTERNAL_FLAGS_COUNT; ++flag) { + const auto flag_code = static_cast<InternalFlag>(flag); + const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false); + internal_flags[flag] = AddGlobalVariable(Name(id, names[flag])); + } + } + + void DeclareInputAttributes() { + for (const auto element : ir.GetInputAttributes()) { + const Attribute::Index index = element.first; + if (!IsGenericAttribute(index)) { + continue; + } + + UNIMPLEMENTED_IF(stage == ShaderStage::Geometry); + + const u32 location = GetGenericAttributeLocation(index); + const Id id = OpVariable(t_in_float4, spv::StorageClass::Input); + Name(AddGlobalVariable(id), fmt::format("in_attr{}", location)); + input_attributes.emplace(index, id); + interfaces.push_back(id); + + Decorate(id, spv::Decoration::Location, location); + + if (stage != ShaderStage::Fragment) { + continue; + } + switch (header.ps.GetAttributeUse(location)) { + case AttributeUse::Constant: + Decorate(id, spv::Decoration::Flat); + break; + case AttributeUse::ScreenLinear: + Decorate(id, spv::Decoration::NoPerspective); + break; + case AttributeUse::Perspective: + // Default + break; + default: + UNREACHABLE_MSG("Unused attribute being fetched"); + } + } + } + + void DeclareOutputAttributes() { + for (const auto index : ir.GetOutputAttributes()) { + if (!IsGenericAttribute(index)) { + continue; + } + const auto location = GetGenericAttributeLocation(index); + const Id id = OpVariable(t_out_float4, spv::StorageClass::Output); + Name(AddGlobalVariable(id), fmt::format("out_attr{}", location)); + output_attributes.emplace(index, id); + interfaces.push_back(id); + + Decorate(id, spv::Decoration::Location, location); + } + } + + void DeclareConstantBuffers() { + u32 binding = const_buffers_base_binding; + for (const auto& entry : ir.GetConstantBuffers()) { + const auto [index, size] = entry; + const Id id = OpVariable(t_cbuf_ubo, spv::StorageClass::Uniform); + AddGlobalVariable(Name(id, fmt::format("cbuf_{}", index))); + + Decorate(id, spv::Decoration::Binding, binding++); + Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); + constant_buffers.emplace(index, id); + } + } + + void DeclareGlobalBuffers() { + u32 binding = global_buffers_base_binding; + for (const auto& entry : ir.GetGlobalMemory()) { + const auto [base, usage] = entry; + const Id id = OpVariable(t_gmem_ssbo, spv::StorageClass::StorageBuffer); + AddGlobalVariable( + Name(id, fmt::format("gmem_{}_{}", base.cbuf_index, base.cbuf_offset))); + + Decorate(id, spv::Decoration::Binding, binding++); + Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); + global_buffers.emplace(base, id); + } + } + + void DeclareSamplers() { + u32 binding = samplers_base_binding; + for (const auto& sampler : ir.GetSamplers()) { + const auto dim = GetSamplerDim(sampler); + const int depth = sampler.IsShadow() ? 1 : 0; + const int arrayed = sampler.IsArray() ? 1 : 0; + // TODO(Rodrigo): Sampled 1 indicates that the image will be used with a sampler. When + // SULD and SUST instructions are implemented, replace this value. + const int sampled = 1; + const Id image_type = + TypeImage(t_float, dim, depth, arrayed, false, sampled, spv::ImageFormat::Unknown); + const Id sampled_image_type = TypeSampledImage(image_type); + const Id pointer_type = + TypePointer(spv::StorageClass::UniformConstant, sampled_image_type); + const Id id = OpVariable(pointer_type, spv::StorageClass::UniformConstant); + AddGlobalVariable(Name(id, fmt::format("sampler_{}", sampler.GetIndex()))); + + sampler_images.insert( + {static_cast<u32>(sampler.GetIndex()), {image_type, sampled_image_type, id}}); + + Decorate(id, spv::Decoration::Binding, binding++); + Decorate(id, spv::Decoration::DescriptorSet, DESCRIPTOR_SET); + } + } + + void DeclareVertexRedeclarations() { + vertex_index = DeclareBuiltIn(spv::BuiltIn::VertexIndex, spv::StorageClass::Input, + t_in_uint, "vertex_index"); + instance_index = DeclareBuiltIn(spv::BuiltIn::InstanceIndex, spv::StorageClass::Input, + t_in_uint, "instance_index"); + + bool is_point_size_declared = false; + bool is_clip_distances_declared = false; + for (const auto index : ir.GetOutputAttributes()) { + if (index == Attribute::Index::PointSize) { + is_point_size_declared = true; + } else if (index == Attribute::Index::ClipDistances0123 || + index == Attribute::Index::ClipDistances4567) { + is_clip_distances_declared = true; + } + } + + std::vector<Id> members; + members.push_back(t_float4); + if (is_point_size_declared) { + members.push_back(t_float); + } + if (is_clip_distances_declared) { + members.push_back(TypeArray(t_float, Constant(t_uint, 8))); + } + + const Id gl_per_vertex_struct = Name(TypeStruct(members), "PerVertex"); + Decorate(gl_per_vertex_struct, spv::Decoration::Block); + + u32 declaration_index = 0; + const auto MemberDecorateBuiltIn = [&](spv::BuiltIn builtin, std::string name, + bool condition) { + if (!condition) + return u32{}; + MemberName(gl_per_vertex_struct, declaration_index, name); + MemberDecorate(gl_per_vertex_struct, declaration_index, spv::Decoration::BuiltIn, + static_cast<u32>(builtin)); + return declaration_index++; + }; + + position_index = MemberDecorateBuiltIn(spv::BuiltIn::Position, "position", true); + point_size_index = + MemberDecorateBuiltIn(spv::BuiltIn::PointSize, "point_size", is_point_size_declared); + clip_distances_index = MemberDecorateBuiltIn(spv::BuiltIn::ClipDistance, "clip_distances", + is_clip_distances_declared); + + const Id type_pointer = TypePointer(spv::StorageClass::Output, gl_per_vertex_struct); + per_vertex = OpVariable(type_pointer, spv::StorageClass::Output); + AddGlobalVariable(Name(per_vertex, "per_vertex")); + interfaces.push_back(per_vertex); + } + + void VisitBasicBlock(const NodeBlock& bb) { + for (const Node node : bb) { + static_cast<void>(Visit(node)); + } + } + + Id Visit(Node node) { + if (const auto operation = std::get_if<OperationNode>(node)) { + const auto operation_index = static_cast<std::size_t>(operation->GetCode()); + const auto decompiler = operation_decompilers[operation_index]; + if (decompiler == nullptr) { + UNREACHABLE_MSG("Operation decompiler {} not defined", operation_index); + } + return (this->*decompiler)(*operation); + + } else if (const auto gpr = std::get_if<GprNode>(node)) { + const u32 index = gpr->GetIndex(); + if (index == Register::ZeroIndex) { + return Constant(t_float, 0.0f); + } + return Emit(OpLoad(t_float, registers.at(index))); + + } else if (const auto immediate = std::get_if<ImmediateNode>(node)) { + return BitcastTo<Type::Float>(Constant(t_uint, immediate->GetValue())); + + } else if (const auto predicate = std::get_if<PredicateNode>(node)) { + const auto value = [&]() -> Id { + switch (const auto index = predicate->GetIndex(); index) { + case Tegra::Shader::Pred::UnusedIndex: + return v_true; + case Tegra::Shader::Pred::NeverExecute: + return v_false; + default: + return Emit(OpLoad(t_bool, predicates.at(index))); + } + }(); + if (predicate->IsNegated()) { + return Emit(OpLogicalNot(t_bool, value)); + } + return value; + + } else if (const auto abuf = std::get_if<AbufNode>(node)) { + const auto attribute = abuf->GetIndex(); + const auto element = abuf->GetElement(); + + switch (attribute) { + case Attribute::Index::Position: + if (stage != ShaderStage::Fragment) { + UNIMPLEMENTED(); + break; + } else { + if (element == 3) { + return Constant(t_float, 1.0f); + } + return Emit(OpLoad(t_float, AccessElement(t_in_float, frag_coord, element))); + } + case Attribute::Index::TessCoordInstanceIDVertexID: + // TODO(Subv): Find out what the values are for the first two elements when inside a + // vertex shader, and what's the value of the fourth element when inside a Tess Eval + // shader. + ASSERT(stage == ShaderStage::Vertex); + switch (element) { + case 2: + return BitcastFrom<Type::Uint>(Emit(OpLoad(t_uint, instance_index))); + case 3: + return BitcastFrom<Type::Uint>(Emit(OpLoad(t_uint, vertex_index))); + } + UNIMPLEMENTED_MSG("Unmanaged TessCoordInstanceIDVertexID element={}", element); + return Constant(t_float, 0); + case Attribute::Index::FrontFacing: + // TODO(Subv): Find out what the values are for the other elements. + ASSERT(stage == ShaderStage::Fragment); + if (element == 3) { + const Id is_front_facing = Emit(OpLoad(t_bool, front_facing)); + const Id true_value = + BitcastTo<Type::Float>(Constant(t_int, static_cast<s32>(-1))); + const Id false_value = BitcastTo<Type::Float>(Constant(t_int, 0)); + return Emit(OpSelect(t_float, is_front_facing, true_value, false_value)); + } + UNIMPLEMENTED_MSG("Unmanaged FrontFacing element={}", element); + return Constant(t_float, 0.0f); + default: + if (IsGenericAttribute(attribute)) { + const Id pointer = + AccessElement(t_in_float, input_attributes.at(attribute), element); + return Emit(OpLoad(t_float, pointer)); + } + break; + } + UNIMPLEMENTED_MSG("Unhandled input attribute: {}", static_cast<u32>(attribute)); + + } else if (const auto cbuf = std::get_if<CbufNode>(node)) { + const Node offset = cbuf->GetOffset(); + const Id buffer_id = constant_buffers.at(cbuf->GetIndex()); + + Id buffer_index{}; + Id buffer_element{}; + + if (const auto immediate = std::get_if<ImmediateNode>(offset)) { + // Direct access + const u32 offset_imm = immediate->GetValue(); + ASSERT(offset_imm % 4 == 0); + buffer_index = Constant(t_uint, offset_imm / 16); + buffer_element = Constant(t_uint, (offset_imm / 4) % 4); + + } else if (std::holds_alternative<OperationNode>(*offset)) { + // Indirect access + // TODO(Rodrigo): Use a uniform buffer stride of 4 and drop this slow math (which + // emits sub-optimal code on GLSL from my testing). + const Id offset_id = BitcastTo<Type::Uint>(Visit(offset)); + const Id unsafe_offset = Emit(OpUDiv(t_uint, offset_id, Constant(t_uint, 4))); + const Id final_offset = Emit( + OpUMod(t_uint, unsafe_offset, Constant(t_uint, MAX_CONSTBUFFER_ELEMENTS - 1))); + buffer_index = Emit(OpUDiv(t_uint, final_offset, Constant(t_uint, 4))); + buffer_element = Emit(OpUMod(t_uint, final_offset, Constant(t_uint, 4))); + + } else { + UNREACHABLE_MSG("Unmanaged offset node type"); + } + + const Id pointer = Emit(OpAccessChain(t_cbuf_float, buffer_id, Constant(t_uint, 0), + buffer_index, buffer_element)); + return Emit(OpLoad(t_float, pointer)); + + } else if (const auto gmem = std::get_if<GmemNode>(node)) { + const Id gmem_buffer = global_buffers.at(gmem->GetDescriptor()); + const Id real = BitcastTo<Type::Uint>(Visit(gmem->GetRealAddress())); + const Id base = BitcastTo<Type::Uint>(Visit(gmem->GetBaseAddress())); + + Id offset = Emit(OpISub(t_uint, real, base)); + offset = Emit(OpUDiv(t_uint, offset, Constant(t_uint, 4u))); + return Emit(OpLoad(t_float, Emit(OpAccessChain(t_gmem_float, gmem_buffer, + Constant(t_uint, 0u), offset)))); + + } else if (const auto conditional = std::get_if<ConditionalNode>(node)) { + // It's invalid to call conditional on nested nodes, use an operation instead + const Id true_label = OpLabel(); + const Id skip_label = OpLabel(); + Emit(OpBranchConditional(Visit(conditional->GetCondition()), true_label, skip_label)); + Emit(true_label); + + VisitBasicBlock(conditional->GetCode()); + + Emit(OpBranch(skip_label)); + Emit(skip_label); + return {}; + + } else if (const auto comment = std::get_if<CommentNode>(node)) { + Name(Emit(OpUndef(t_void)), comment->GetText()); + return {}; + } + + UNREACHABLE(); + return {}; + } + + template <Id (Module::*func)(Id, Id), Type result_type, Type type_a = result_type> + Id Unary(Operation operation) { + const Id type_def = GetTypeDefinition(result_type); + const Id op_a = VisitOperand<type_a>(operation, 0); + + const Id value = BitcastFrom<result_type>(Emit((this->*func)(type_def, op_a))); + if (IsPrecise(operation)) { + Decorate(value, spv::Decoration::NoContraction); + } + return value; + } + + template <Id (Module::*func)(Id, Id, Id), Type result_type, Type type_a = result_type, + Type type_b = type_a> + Id Binary(Operation operation) { + const Id type_def = GetTypeDefinition(result_type); + const Id op_a = VisitOperand<type_a>(operation, 0); + const Id op_b = VisitOperand<type_b>(operation, 1); + + const Id value = BitcastFrom<result_type>(Emit((this->*func)(type_def, op_a, op_b))); + if (IsPrecise(operation)) { + Decorate(value, spv::Decoration::NoContraction); + } + return value; + } + + template <Id (Module::*func)(Id, Id, Id, Id), Type result_type, Type type_a = result_type, + Type type_b = type_a, Type type_c = type_b> + Id Ternary(Operation operation) { + const Id type_def = GetTypeDefinition(result_type); + const Id op_a = VisitOperand<type_a>(operation, 0); + const Id op_b = VisitOperand<type_b>(operation, 1); + const Id op_c = VisitOperand<type_c>(operation, 2); + + const Id value = BitcastFrom<result_type>(Emit((this->*func)(type_def, op_a, op_b, op_c))); + if (IsPrecise(operation)) { + Decorate(value, spv::Decoration::NoContraction); + } + return value; + } + + template <Id (Module::*func)(Id, Id, Id, Id, Id), Type result_type, Type type_a = result_type, + Type type_b = type_a, Type type_c = type_b, Type type_d = type_c> + Id Quaternary(Operation operation) { + const Id type_def = GetTypeDefinition(result_type); + const Id op_a = VisitOperand<type_a>(operation, 0); + const Id op_b = VisitOperand<type_b>(operation, 1); + const Id op_c = VisitOperand<type_c>(operation, 2); + const Id op_d = VisitOperand<type_d>(operation, 3); + + const Id value = + BitcastFrom<result_type>(Emit((this->*func)(type_def, op_a, op_b, op_c, op_d))); + if (IsPrecise(operation)) { + Decorate(value, spv::Decoration::NoContraction); + } + return value; + } + + Id Assign(Operation operation) { + const Node dest = operation[0]; + const Node src = operation[1]; + + Id target{}; + if (const auto gpr = std::get_if<GprNode>(dest)) { + if (gpr->GetIndex() == Register::ZeroIndex) { + // Writing to Register::ZeroIndex is a no op + return {}; + } + target = registers.at(gpr->GetIndex()); + + } else if (const auto abuf = std::get_if<AbufNode>(dest)) { + target = [&]() -> Id { + switch (const auto attribute = abuf->GetIndex(); attribute) { + case Attribute::Index::Position: + return AccessElement(t_out_float, per_vertex, position_index, + abuf->GetElement()); + case Attribute::Index::PointSize: + return AccessElement(t_out_float, per_vertex, point_size_index); + case Attribute::Index::ClipDistances0123: + return AccessElement(t_out_float, per_vertex, clip_distances_index, + abuf->GetElement()); + case Attribute::Index::ClipDistances4567: + return AccessElement(t_out_float, per_vertex, clip_distances_index, + abuf->GetElement() + 4); + default: + if (IsGenericAttribute(attribute)) { + return AccessElement(t_out_float, output_attributes.at(attribute), + abuf->GetElement()); + } + UNIMPLEMENTED_MSG("Unhandled output attribute: {}", + static_cast<u32>(attribute)); + return {}; + } + }(); + + } else if (const auto lmem = std::get_if<LmemNode>(dest)) { + Id address = BitcastTo<Type::Uint>(Visit(lmem->GetAddress())); + address = Emit(OpUDiv(t_uint, address, Constant(t_uint, 4))); + target = Emit(OpAccessChain(t_prv_float, local_memory, {address})); + } + + Emit(OpStore(target, Visit(src))); + return {}; + } + + Id HNegate(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id HMergeF32(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id HMergeH0(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id HMergeH1(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id HPack2(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id LogicalAssign(Operation operation) { + const Node dest = operation[0]; + const Node src = operation[1]; + + Id target{}; + if (const auto pred = std::get_if<PredicateNode>(dest)) { + ASSERT_MSG(!pred->IsNegated(), "Negating logical assignment"); + + const auto index = pred->GetIndex(); + switch (index) { + case Tegra::Shader::Pred::NeverExecute: + case Tegra::Shader::Pred::UnusedIndex: + // Writing to these predicates is a no-op + return {}; + } + target = predicates.at(index); + + } else if (const auto flag = std::get_if<InternalFlagNode>(dest)) { + target = internal_flags.at(static_cast<u32>(flag->GetFlag())); + } + + Emit(OpStore(target, Visit(src))); + return {}; + } + + Id LogicalPick2(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id LogicalAll2(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id LogicalAny2(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id GetTextureSampler(Operation operation) { + const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); + const auto entry = sampler_images.at(static_cast<u32>(meta->sampler.GetIndex())); + return Emit(OpLoad(entry.sampled_image_type, entry.sampler)); + } + + Id GetTextureImage(Operation operation) { + const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); + const auto entry = sampler_images.at(static_cast<u32>(meta->sampler.GetIndex())); + return Emit(OpImage(entry.image_type, GetTextureSampler(operation))); + } + + Id GetTextureCoordinates(Operation operation) { + const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); + std::vector<Id> coords; + for (std::size_t i = 0; i < operation.GetOperandsCount(); ++i) { + coords.push_back(Visit(operation[i])); + } + if (meta->sampler.IsArray()) { + const Id array_integer = BitcastTo<Type::Int>(Visit(meta->array)); + coords.push_back(Emit(OpConvertSToF(t_float, array_integer))); + } + if (meta->sampler.IsShadow()) { + coords.push_back(Visit(meta->depth_compare)); + } + + const std::array<Id, 4> t_float_lut = {nullptr, t_float2, t_float3, t_float4}; + return coords.size() == 1 + ? coords[0] + : Emit(OpCompositeConstruct(t_float_lut.at(coords.size() - 1), coords)); + } + + Id GetTextureElement(Operation operation, Id sample_value) { + const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); + ASSERT(meta); + return Emit(OpCompositeExtract(t_float, sample_value, meta->element)); + } + + Id Texture(Operation operation) { + const Id texture = Emit(OpImageSampleImplicitLod(t_float4, GetTextureSampler(operation), + GetTextureCoordinates(operation))); + return GetTextureElement(operation, texture); + } + + Id TextureLod(Operation operation) { + const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); + const Id texture = Emit(OpImageSampleExplicitLod( + t_float4, GetTextureSampler(operation), GetTextureCoordinates(operation), + spv::ImageOperandsMask::Lod, Visit(meta->lod))); + return GetTextureElement(operation, texture); + } + + Id TextureGather(Operation operation) { + const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); + const auto coords = GetTextureCoordinates(operation); + + Id texture; + if (meta->sampler.IsShadow()) { + texture = Emit(OpImageDrefGather(t_float4, GetTextureSampler(operation), coords, + Visit(meta->component))); + } else { + u32 component_value = 0; + if (meta->component) { + const auto component = std::get_if<ImmediateNode>(meta->component); + ASSERT_MSG(component, "Component is not an immediate value"); + component_value = component->GetValue(); + } + texture = Emit(OpImageGather(t_float4, GetTextureSampler(operation), coords, + Constant(t_uint, component_value))); + } + + return GetTextureElement(operation, texture); + } + + Id TextureQueryDimensions(Operation operation) { + const auto meta = std::get_if<MetaTexture>(&operation.GetMeta()); + const auto image_id = GetTextureImage(operation); + AddCapability(spv::Capability::ImageQuery); + + if (meta->element == 3) { + return BitcastTo<Type::Float>(Emit(OpImageQueryLevels(t_int, image_id))); + } + + const Id lod = VisitOperand<Type::Uint>(operation, 0); + const std::size_t coords_count = [&]() { + switch (const auto type = meta->sampler.GetType(); type) { + case Tegra::Shader::TextureType::Texture1D: + return 1; + case Tegra::Shader::TextureType::Texture2D: + case Tegra::Shader::TextureType::TextureCube: + return 2; + case Tegra::Shader::TextureType::Texture3D: + return 3; + default: + UNREACHABLE_MSG("Invalid texture type={}", static_cast<u32>(type)); + return 2; + } + }(); + + if (meta->element >= coords_count) { + return Constant(t_float, 0.0f); + } + + const std::array<Id, 3> types = {t_int, t_int2, t_int3}; + const Id sizes = Emit(OpImageQuerySizeLod(types.at(coords_count - 1), image_id, lod)); + const Id size = Emit(OpCompositeExtract(t_int, sizes, meta->element)); + return BitcastTo<Type::Float>(size); + } + + Id TextureQueryLod(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id TexelFetch(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id Branch(Operation operation) { + const auto target = std::get_if<ImmediateNode>(operation[0]); + UNIMPLEMENTED_IF(!target); + + Emit(OpStore(jmp_to, Constant(t_uint, target->GetValue()))); + BranchingOp([&]() { Emit(OpBranch(continue_label)); }); + return {}; + } + + Id PushFlowStack(Operation operation) { + const auto target = std::get_if<ImmediateNode>(operation[0]); + ASSERT(target); + + const Id current = Emit(OpLoad(t_uint, flow_stack_top)); + const Id next = Emit(OpIAdd(t_uint, current, Constant(t_uint, 1))); + const Id access = Emit(OpAccessChain(t_func_uint, flow_stack, current)); + + Emit(OpStore(access, Constant(t_uint, target->GetValue()))); + Emit(OpStore(flow_stack_top, next)); + return {}; + } + + Id PopFlowStack(Operation operation) { + const Id current = Emit(OpLoad(t_uint, flow_stack_top)); + const Id previous = Emit(OpISub(t_uint, current, Constant(t_uint, 1))); + const Id access = Emit(OpAccessChain(t_func_uint, flow_stack, previous)); + const Id target = Emit(OpLoad(t_uint, access)); + + Emit(OpStore(flow_stack_top, previous)); + Emit(OpStore(jmp_to, target)); + BranchingOp([&]() { Emit(OpBranch(continue_label)); }); + return {}; + } + + Id Exit(Operation operation) { + switch (stage) { + case ShaderStage::Vertex: { + // TODO(Rodrigo): We should use VK_EXT_depth_range_unrestricted instead, but it doesn't + // seem to be working on Nvidia's drivers and Intel (mesa and blob) doesn't support it. + const Id position = AccessElement(t_float4, per_vertex, position_index); + Id depth = Emit(OpLoad(t_float, AccessElement(t_out_float, position, 2))); + depth = Emit(OpFAdd(t_float, depth, Constant(t_float, 1.0f))); + depth = Emit(OpFMul(t_float, depth, Constant(t_float, 0.5f))); + Emit(OpStore(AccessElement(t_out_float, position, 2), depth)); + break; + } + case ShaderStage::Fragment: { + const auto SafeGetRegister = [&](u32 reg) { + // TODO(Rodrigo): Replace with contains once C++20 releases + if (const auto it = registers.find(reg); it != registers.end()) { + return Emit(OpLoad(t_float, it->second)); + } + return Constant(t_float, 0.0f); + }; + + 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; + for (u32 rt = 0; rt < Maxwell::NumRenderTargets; ++rt) { + // TODO(Subv): Figure out how dual-source blending is configured in the Switch. + for (u32 component = 0; component < 4; ++component) { + if (header.ps.IsColorComponentOutputEnabled(rt, component)) { + Emit(OpStore(AccessElement(t_out_float, frag_colors.at(rt), component), + SafeGetRegister(current_reg))); + ++current_reg; + } + } + } + if (header.ps.omap.depth) { + // The depth output is always 2 registers after the last color output, and + // current_reg already contains one past the last color register. + Emit(OpStore(frag_depth, SafeGetRegister(current_reg + 1))); + } + break; + } + } + + BranchingOp([&]() { Emit(OpReturn()); }); + return {}; + } + + Id Discard(Operation operation) { + BranchingOp([&]() { Emit(OpKill()); }); + return {}; + } + + Id EmitVertex(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id EndPrimitive(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id YNegate(Operation operation) { + UNIMPLEMENTED(); + return {}; + } + + Id DeclareBuiltIn(spv::BuiltIn builtin, spv::StorageClass storage, Id type, + const std::string& name) { + const Id id = OpVariable(type, storage); + Decorate(id, spv::Decoration::BuiltIn, static_cast<u32>(builtin)); + AddGlobalVariable(Name(id, name)); + interfaces.push_back(id); + return id; + } + + bool IsRenderTargetUsed(u32 rt) const { + for (u32 component = 0; component < 4; ++component) { + if (header.ps.IsColorComponentOutputEnabled(rt, component)) { + return true; + } + } + return false; + } + + template <typename... Args> + Id AccessElement(Id pointer_type, Id composite, Args... elements_) { + std::vector<Id> members; + auto elements = {elements_...}; + for (const auto element : elements) { + members.push_back(Constant(t_uint, element)); + } + + return Emit(OpAccessChain(pointer_type, composite, members)); + } + + template <Type type> + Id VisitOperand(Operation operation, std::size_t operand_index) { + const Id value = Visit(operation[operand_index]); + + switch (type) { + case Type::Bool: + case Type::Bool2: + case Type::Float: + return value; + case Type::Int: + return Emit(OpBitcast(t_int, value)); + case Type::Uint: + return Emit(OpBitcast(t_uint, value)); + case Type::HalfFloat: + UNIMPLEMENTED(); + } + UNREACHABLE(); + return value; + } + + template <Type type> + Id BitcastFrom(Id value) { + switch (type) { + case Type::Bool: + case Type::Bool2: + case Type::Float: + return value; + case Type::Int: + case Type::Uint: + return Emit(OpBitcast(t_float, value)); + case Type::HalfFloat: + UNIMPLEMENTED(); + } + UNREACHABLE(); + return value; + } + + template <Type type> + Id BitcastTo(Id value) { + switch (type) { + case Type::Bool: + case Type::Bool2: + UNREACHABLE(); + case Type::Float: + return Emit(OpBitcast(t_float, value)); + case Type::Int: + return Emit(OpBitcast(t_int, value)); + case Type::Uint: + return Emit(OpBitcast(t_uint, value)); + case Type::HalfFloat: + UNIMPLEMENTED(); + } + UNREACHABLE(); + return value; + } + + Id GetTypeDefinition(Type type) { + switch (type) { + case Type::Bool: + return t_bool; + case Type::Bool2: + return t_bool2; + case Type::Float: + return t_float; + case Type::Int: + return t_int; + case Type::Uint: + return t_uint; + case Type::HalfFloat: + UNIMPLEMENTED(); + } + UNREACHABLE(); + return {}; + } + + void BranchingOp(std::function<void()> call) { + const Id true_label = OpLabel(); + const Id skip_label = OpLabel(); + Emit(OpSelectionMerge(skip_label, spv::SelectionControlMask::Flatten)); + Emit(OpBranchConditional(v_true, true_label, skip_label, 1, 0)); + Emit(true_label); + call(); + + Emit(skip_label); + } + + static constexpr OperationDecompilersArray operation_decompilers = { + &SPIRVDecompiler::Assign, + + &SPIRVDecompiler::Ternary<&Module::OpSelect, Type::Float, Type::Bool, Type::Float, + Type::Float>, + + &SPIRVDecompiler::Binary<&Module::OpFAdd, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFMul, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFDiv, Type::Float>, + &SPIRVDecompiler::Ternary<&Module::OpFma, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpFNegate, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpFAbs, Type::Float>, + &SPIRVDecompiler::Ternary<&Module::OpFClamp, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFMin, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFMax, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpCos, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpSin, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpExp2, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpLog2, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpInverseSqrt, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpSqrt, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpRoundEven, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpFloor, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpCeil, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpTrunc, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpConvertSToF, Type::Float, Type::Int>, + &SPIRVDecompiler::Unary<&Module::OpConvertUToF, Type::Float, Type::Uint>, + + &SPIRVDecompiler::Binary<&Module::OpIAdd, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpIMul, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpSDiv, Type::Int>, + &SPIRVDecompiler::Unary<&Module::OpSNegate, Type::Int>, + &SPIRVDecompiler::Unary<&Module::OpSAbs, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpSMin, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpSMax, Type::Int>, + + &SPIRVDecompiler::Unary<&Module::OpConvertFToS, Type::Int, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpBitcast, Type::Int, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpShiftLeftLogical, Type::Int, Type::Int, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpShiftRightLogical, Type::Int, Type::Int, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpShiftRightArithmetic, Type::Int, Type::Int, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpBitwiseAnd, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpBitwiseOr, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpBitwiseXor, Type::Int>, + &SPIRVDecompiler::Unary<&Module::OpNot, Type::Int>, + &SPIRVDecompiler::Quaternary<&Module::OpBitFieldInsert, Type::Int>, + &SPIRVDecompiler::Ternary<&Module::OpBitFieldSExtract, Type::Int>, + &SPIRVDecompiler::Unary<&Module::OpBitCount, Type::Int>, + + &SPIRVDecompiler::Binary<&Module::OpIAdd, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpIMul, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpUDiv, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpUMin, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpUMax, Type::Uint>, + &SPIRVDecompiler::Unary<&Module::OpConvertFToU, Type::Uint, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpBitcast, Type::Uint, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpShiftLeftLogical, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpShiftRightLogical, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpShiftRightArithmetic, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpBitwiseAnd, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpBitwiseOr, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpBitwiseXor, Type::Uint>, + &SPIRVDecompiler::Unary<&Module::OpNot, Type::Uint>, + &SPIRVDecompiler::Quaternary<&Module::OpBitFieldInsert, Type::Uint>, + &SPIRVDecompiler::Ternary<&Module::OpBitFieldUExtract, Type::Uint>, + &SPIRVDecompiler::Unary<&Module::OpBitCount, Type::Uint>, + + &SPIRVDecompiler::Binary<&Module::OpFAdd, Type::HalfFloat>, + &SPIRVDecompiler::Binary<&Module::OpFMul, Type::HalfFloat>, + &SPIRVDecompiler::Ternary<&Module::OpFma, Type::HalfFloat>, + &SPIRVDecompiler::Unary<&Module::OpFAbs, Type::HalfFloat>, + &SPIRVDecompiler::HNegate, + &SPIRVDecompiler::HMergeF32, + &SPIRVDecompiler::HMergeH0, + &SPIRVDecompiler::HMergeH1, + &SPIRVDecompiler::HPack2, + + &SPIRVDecompiler::LogicalAssign, + &SPIRVDecompiler::Binary<&Module::OpLogicalAnd, Type::Bool>, + &SPIRVDecompiler::Binary<&Module::OpLogicalOr, Type::Bool>, + &SPIRVDecompiler::Binary<&Module::OpLogicalNotEqual, Type::Bool>, + &SPIRVDecompiler::Unary<&Module::OpLogicalNot, Type::Bool>, + &SPIRVDecompiler::LogicalPick2, + &SPIRVDecompiler::LogicalAll2, + &SPIRVDecompiler::LogicalAny2, + + &SPIRVDecompiler::Binary<&Module::OpFOrdLessThan, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFOrdEqual, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFOrdLessThanEqual, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFOrdGreaterThan, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFOrdNotEqual, Type::Bool, Type::Float>, + &SPIRVDecompiler::Binary<&Module::OpFOrdGreaterThanEqual, Type::Bool, Type::Float>, + &SPIRVDecompiler::Unary<&Module::OpIsNan, Type::Bool>, + + &SPIRVDecompiler::Binary<&Module::OpSLessThan, Type::Bool, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpIEqual, Type::Bool, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpSLessThanEqual, Type::Bool, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpSGreaterThan, Type::Bool, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpINotEqual, Type::Bool, Type::Int>, + &SPIRVDecompiler::Binary<&Module::OpSGreaterThanEqual, Type::Bool, Type::Int>, + + &SPIRVDecompiler::Binary<&Module::OpULessThan, Type::Bool, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpIEqual, Type::Bool, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpULessThanEqual, Type::Bool, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpUGreaterThan, Type::Bool, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpINotEqual, Type::Bool, Type::Uint>, + &SPIRVDecompiler::Binary<&Module::OpUGreaterThanEqual, Type::Bool, Type::Uint>, + + &SPIRVDecompiler::Binary<&Module::OpFOrdLessThan, Type::Bool, Type::HalfFloat>, + &SPIRVDecompiler::Binary<&Module::OpFOrdEqual, Type::Bool, Type::HalfFloat>, + &SPIRVDecompiler::Binary<&Module::OpFOrdLessThanEqual, Type::Bool, Type::HalfFloat>, + &SPIRVDecompiler::Binary<&Module::OpFOrdGreaterThan, Type::Bool, Type::HalfFloat>, + &SPIRVDecompiler::Binary<&Module::OpFOrdNotEqual, Type::Bool, Type::HalfFloat>, + &SPIRVDecompiler::Binary<&Module::OpFOrdGreaterThanEqual, Type::Bool, Type::HalfFloat>, + + &SPIRVDecompiler::Texture, + &SPIRVDecompiler::TextureLod, + &SPIRVDecompiler::TextureGather, + &SPIRVDecompiler::TextureQueryDimensions, + &SPIRVDecompiler::TextureQueryLod, + &SPIRVDecompiler::TexelFetch, + + &SPIRVDecompiler::Branch, + &SPIRVDecompiler::PushFlowStack, + &SPIRVDecompiler::PopFlowStack, + &SPIRVDecompiler::Exit, + &SPIRVDecompiler::Discard, + + &SPIRVDecompiler::EmitVertex, + &SPIRVDecompiler::EndPrimitive, + + &SPIRVDecompiler::YNegate, + }; + + const ShaderIR& ir; + const ShaderStage stage; + const Tegra::Shader::Header header; + + const Id t_void = Name(TypeVoid(), "void"); + + const Id t_bool = Name(TypeBool(), "bool"); + const Id t_bool2 = Name(TypeVector(t_bool, 2), "bool2"); + + const Id t_int = Name(TypeInt(32, true), "int"); + const Id t_int2 = Name(TypeVector(t_int, 2), "int2"); + const Id t_int3 = Name(TypeVector(t_int, 3), "int3"); + const Id t_int4 = Name(TypeVector(t_int, 4), "int4"); + + const Id t_uint = Name(TypeInt(32, false), "uint"); + const Id t_uint2 = Name(TypeVector(t_uint, 2), "uint2"); + const Id t_uint3 = Name(TypeVector(t_uint, 3), "uint3"); + const Id t_uint4 = Name(TypeVector(t_uint, 4), "uint4"); + + const Id t_float = Name(TypeFloat(32), "float"); + const Id t_float2 = Name(TypeVector(t_float, 2), "float2"); + const Id t_float3 = Name(TypeVector(t_float, 3), "float3"); + const Id t_float4 = Name(TypeVector(t_float, 4), "float4"); + + const Id t_prv_bool = Name(TypePointer(spv::StorageClass::Private, t_bool), "prv_bool"); + const Id t_prv_float = Name(TypePointer(spv::StorageClass::Private, t_float), "prv_float"); + + const Id t_func_uint = Name(TypePointer(spv::StorageClass::Function, t_uint), "func_uint"); + + const Id t_in_bool = Name(TypePointer(spv::StorageClass::Input, t_bool), "in_bool"); + const Id t_in_uint = Name(TypePointer(spv::StorageClass::Input, t_uint), "in_uint"); + const Id t_in_float = Name(TypePointer(spv::StorageClass::Input, t_float), "in_float"); + const Id t_in_float4 = Name(TypePointer(spv::StorageClass::Input, t_float4), "in_float4"); + + const Id t_out_float = Name(TypePointer(spv::StorageClass::Output, t_float), "out_float"); + const Id t_out_float4 = Name(TypePointer(spv::StorageClass::Output, t_float4), "out_float4"); + + const Id t_cbuf_float = TypePointer(spv::StorageClass::Uniform, t_float); + const Id t_cbuf_array = + Decorate(Name(TypeArray(t_float4, Constant(t_uint, MAX_CONSTBUFFER_ELEMENTS)), "CbufArray"), + spv::Decoration::ArrayStride, CBUF_STRIDE); + const Id t_cbuf_struct = MemberDecorate( + Decorate(TypeStruct(t_cbuf_array), spv::Decoration::Block), 0, spv::Decoration::Offset, 0); + const Id t_cbuf_ubo = TypePointer(spv::StorageClass::Uniform, t_cbuf_struct); + + const Id t_gmem_float = TypePointer(spv::StorageClass::StorageBuffer, t_float); + const Id t_gmem_array = + Name(Decorate(TypeRuntimeArray(t_float), spv::Decoration::ArrayStride, 4u), "GmemArray"); + const Id t_gmem_struct = MemberDecorate( + Decorate(TypeStruct(t_gmem_array), spv::Decoration::Block), 0, spv::Decoration::Offset, 0); + const Id t_gmem_ssbo = TypePointer(spv::StorageClass::StorageBuffer, t_gmem_struct); + + const Id v_float_zero = Constant(t_float, 0.0f); + const Id v_true = ConstantTrue(t_bool); + const Id v_false = ConstantFalse(t_bool); + + Id per_vertex{}; + std::map<u32, Id> registers; + std::map<Tegra::Shader::Pred, Id> predicates; + Id local_memory{}; + std::array<Id, INTERNAL_FLAGS_COUNT> internal_flags{}; + std::map<Attribute::Index, Id> input_attributes; + std::map<Attribute::Index, Id> output_attributes; + std::map<u32, Id> constant_buffers; + std::map<GlobalMemoryBase, Id> global_buffers; + std::map<u32, SamplerImage> sampler_images; + + Id instance_index{}; + Id vertex_index{}; + std::array<Id, Maxwell::NumRenderTargets> frag_colors{}; + Id frag_depth{}; + Id frag_coord{}; + Id front_facing{}; + + u32 position_index{}; + u32 point_size_index{}; + u32 clip_distances_index{}; + + std::vector<Id> interfaces; + + u32 const_buffers_base_binding{}; + u32 global_buffers_base_binding{}; + u32 samplers_base_binding{}; + + Id execute_function{}; + Id jmp_to{}; + Id flow_stack_top{}; + Id flow_stack{}; + Id continue_label{}; + std::map<u32, Id> labels; +}; + +DecompilerResult Decompile(const VideoCommon::Shader::ShaderIR& ir, Maxwell::ShaderStage stage) { + auto decompiler = std::make_unique<SPIRVDecompiler>(ir, stage); + decompiler->Decompile(); + return {std::move(decompiler), decompiler->GetShaderEntries()}; +} + +} // namespace Vulkan::VKShader diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.h b/src/video_core/renderer_vulkan/vk_shader_decompiler.h new file mode 100644 index 000000000..329d8fa38 --- /dev/null +++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.h @@ -0,0 +1,80 @@ +// Copyright 2019 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <array> +#include <memory> +#include <set> +#include <utility> +#include <vector> + +#include <sirit/sirit.h> + +#include "common/common_types.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/shader/shader_ir.h" + +namespace VideoCommon::Shader { +class ShaderIR; +} + +namespace Vulkan::VKShader { + +using Maxwell = Tegra::Engines::Maxwell3D::Regs; + +using SamplerEntry = VideoCommon::Shader::Sampler; + +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} {} + + constexpr u32 GetIndex() const { + return index; + } + +private: + u32 index{}; +}; + +class GlobalBufferEntry { +public: + explicit GlobalBufferEntry(u32 cbuf_index, u32 cbuf_offset) + : cbuf_index{cbuf_index}, cbuf_offset{cbuf_offset} {} + + u32 GetCbufIndex() const { + return cbuf_index; + } + + u32 GetCbufOffset() const { + return cbuf_offset; + } + +private: + u32 cbuf_index{}; + u32 cbuf_offset{}; +}; + +struct ShaderEntries { + u32 const_buffers_base_binding{}; + u32 global_buffers_base_binding{}; + u32 samplers_base_binding{}; + std::vector<ConstBufferEntry> const_buffers; + std::vector<GlobalBufferEntry> global_buffers; + std::vector<SamplerEntry> samplers; + std::set<u32> attributes; + std::array<bool, Maxwell::NumClipDistances> clip_distances{}; + std::size_t shader_length{}; + Sirit::Id entry_function{}; + std::vector<Sirit::Id> interfaces; +}; + +using DecompilerResult = std::pair<std::unique_ptr<Sirit::Module>, ShaderEntries>; + +DecompilerResult Decompile(const VideoCommon::Shader::ShaderIR& ir, Maxwell::ShaderStage stage); + +} // namespace Vulkan::VKShader diff --git a/src/video_core/shader/decode/memory.cpp b/src/video_core/shader/decode/memory.cpp index b2b706cb8..ea1092db1 100644 --- a/src/video_core/shader/decode/memory.cpp +++ b/src/video_core/shader/decode/memory.cpp @@ -19,6 +19,23 @@ using Tegra::Shader::Instruction; using Tegra::Shader::OpCode; using Tegra::Shader::Register; +namespace { +u32 GetUniformTypeElementsCount(Tegra::Shader::UniformType uniform_type) { + switch (uniform_type) { + case Tegra::Shader::UniformType::Single: + return 1; + case Tegra::Shader::UniformType::Double: + return 2; + case Tegra::Shader::UniformType::Quad: + case Tegra::Shader::UniformType::UnsignedQuad: + return 4; + default: + UNIMPLEMENTED_MSG("Unimplemented size={}!", static_cast<u32>(uniform_type)); + return 1; + } +} +} // namespace + u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { const Instruction instr = {program_code[pc]}; const auto opcode = OpCode::Decode(instr); @@ -127,45 +144,15 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { break; } case OpCode::Id::LDG: { - const u32 count = [&]() { - switch (instr.ldg.type) { - case Tegra::Shader::UniformType::Single: - return 1; - case Tegra::Shader::UniformType::Double: - return 2; - case Tegra::Shader::UniformType::Quad: - case Tegra::Shader::UniformType::UnsignedQuad: - return 4; - default: - UNIMPLEMENTED_MSG("Unimplemented LDG size!"); - return 1; - } - }(); - - const Node addr_register = GetRegister(instr.gpr8); - const Node base_address = - TrackCbuf(addr_register, global_code, static_cast<s64>(global_code.size())); - const auto cbuf = std::get_if<CbufNode>(base_address); - ASSERT(cbuf != nullptr); - const auto cbuf_offset_imm = std::get_if<ImmediateNode>(cbuf->GetOffset()); - ASSERT(cbuf_offset_imm != nullptr); - const auto cbuf_offset = cbuf_offset_imm->GetValue(); - - bb.push_back(Comment( - fmt::format("Base address is c[0x{:x}][0x{:x}]", cbuf->GetIndex(), cbuf_offset))); - - const GlobalMemoryBase descriptor{cbuf->GetIndex(), cbuf_offset}; - used_global_memory_bases.insert(descriptor); - - const Node immediate_offset = - Immediate(static_cast<u32>(instr.ldg.immediate_offset.Value())); - const Node base_real_address = - Operation(OperationCode::UAdd, NO_PRECISE, immediate_offset, addr_register); + const auto [real_address_base, base_address, descriptor] = + TrackAndGetGlobalMemory(bb, GetRegister(instr.gpr8), + static_cast<u32>(instr.ldg.immediate_offset.Value()), false); + const u32 count = GetUniformTypeElementsCount(instr.ldg.type); for (u32 i = 0; i < count; ++i) { const Node it_offset = Immediate(i * 4); const Node real_address = - Operation(OperationCode::UAdd, NO_PRECISE, base_real_address, it_offset); + Operation(OperationCode::UAdd, NO_PRECISE, real_address_base, it_offset); const Node gmem = StoreNode(GmemNode(real_address, base_address, descriptor)); SetTemporal(bb, i, gmem); @@ -175,6 +162,28 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { } break; } + case OpCode::Id::STG: { + const auto [real_address_base, base_address, descriptor] = + TrackAndGetGlobalMemory(bb, GetRegister(instr.gpr8), + static_cast<u32>(instr.stg.immediate_offset.Value()), true); + + // Encode in temporary registers like this: real_base_address, {registers_to_be_written...} + SetTemporal(bb, 0, real_address_base); + + const u32 count = GetUniformTypeElementsCount(instr.stg.type); + for (u32 i = 0; i < count; ++i) { + SetTemporal(bb, i + 1, GetRegister(instr.gpr0.Value() + i)); + } + for (u32 i = 0; i < count; ++i) { + const Node it_offset = Immediate(i * 4); + const Node real_address = + Operation(OperationCode::UAdd, NO_PRECISE, real_address_base, it_offset); + const Node gmem = StoreNode(GmemNode(real_address, base_address, descriptor)); + + bb.push_back(Operation(OperationCode::Assign, gmem, GetTemporal(i + 1))); + } + break; + } case OpCode::Id::ST_A: { UNIMPLEMENTED_IF_MSG(instr.gpr8.Value() != Register::ZeroIndex, "Indirect attribute loads are not supported"); @@ -237,4 +246,34 @@ u32 ShaderIR::DecodeMemory(NodeBlock& bb, u32 pc) { return pc; } +std::tuple<Node, Node, GlobalMemoryBase> ShaderIR::TrackAndGetGlobalMemory(NodeBlock& bb, + Node addr_register, + u32 immediate_offset, + bool is_write) { + const Node base_address{ + TrackCbuf(addr_register, global_code, static_cast<s64>(global_code.size()))}; + const auto cbuf = std::get_if<CbufNode>(base_address); + ASSERT(cbuf != nullptr); + const auto cbuf_offset_imm = std::get_if<ImmediateNode>(cbuf->GetOffset()); + ASSERT(cbuf_offset_imm != nullptr); + const auto cbuf_offset = cbuf_offset_imm->GetValue(); + + bb.push_back( + Comment(fmt::format("Base address is c[0x{:x}][0x{:x}]", cbuf->GetIndex(), cbuf_offset))); + + const GlobalMemoryBase descriptor{cbuf->GetIndex(), cbuf_offset}; + const auto& [entry, is_new] = used_global_memory.try_emplace(descriptor); + auto& usage = entry->second; + if (is_write) { + usage.is_written = true; + } else { + usage.is_read = true; + } + + const auto real_address = + Operation(OperationCode::UAdd, NO_PRECISE, Immediate(immediate_offset), addr_register); + + return {real_address, base_address, descriptor}; +} + } // namespace VideoCommon::Shader diff --git a/src/video_core/shader/decode/texture.cpp b/src/video_core/shader/decode/texture.cpp index a99ae19bf..a775b402b 100644 --- a/src/video_core/shader/decode/texture.cpp +++ b/src/video_core/shader/decode/texture.cpp @@ -7,7 +7,9 @@ #include <fmt/format.h> #include "common/assert.h" +#include "common/bit_field.h" #include "common/common_types.h" +#include "common/logging/log.h" #include "video_core/engines/shader_bytecode.h" #include "video_core/shader/shader_ir.h" @@ -41,19 +43,18 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { switch (opcode->get().GetId()) { case OpCode::Id::TEX: { - UNIMPLEMENTED_IF_MSG(instr.tex.UsesMiscMode(TextureMiscMode::AOFFI), - "AOFFI is not implemented"); - if (instr.tex.UsesMiscMode(TextureMiscMode::NODEP)) { LOG_WARNING(HW_GPU, "TEX.NODEP implementation is incomplete"); } const TextureType texture_type{instr.tex.texture_type}; const bool is_array = instr.tex.array != 0; + const bool is_aoffi = instr.tex.UsesMiscMode(TextureMiscMode::AOFFI); const bool depth_compare = instr.tex.UsesMiscMode(TextureMiscMode::DC); const auto process_mode = instr.tex.GetTextureProcessMode(); WriteTexInstructionFloat( - bb, instr, GetTexCode(instr, texture_type, process_mode, depth_compare, is_array)); + bb, instr, + GetTexCode(instr, texture_type, process_mode, depth_compare, is_array, is_aoffi)); break; } case OpCode::Id::TEXS: { @@ -78,8 +79,6 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { } case OpCode::Id::TLD4: { ASSERT(instr.tld4.array == 0); - UNIMPLEMENTED_IF_MSG(instr.tld4.UsesMiscMode(TextureMiscMode::AOFFI), - "AOFFI is not implemented"); UNIMPLEMENTED_IF_MSG(instr.tld4.UsesMiscMode(TextureMiscMode::NDV), "NDV is not implemented"); UNIMPLEMENTED_IF_MSG(instr.tld4.UsesMiscMode(TextureMiscMode::PTP), @@ -92,8 +91,9 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { const auto texture_type = instr.tld4.texture_type.Value(); const bool depth_compare = instr.tld4.UsesMiscMode(TextureMiscMode::DC); const bool is_array = instr.tld4.array != 0; - WriteTexInstructionFloat(bb, instr, - GetTld4Code(instr, texture_type, depth_compare, is_array)); + const bool is_aoffi = instr.tld4.UsesMiscMode(TextureMiscMode::AOFFI); + WriteTexInstructionFloat( + bb, instr, GetTld4Code(instr, texture_type, depth_compare, is_array, is_aoffi)); break; } case OpCode::Id::TLD4S: { @@ -127,7 +127,7 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto coords_copy = coords; - MetaTexture meta{sampler, {}, {}, {}, {}, component, element}; + MetaTexture meta{sampler, {}, {}, {}, {}, {}, component, element}; values[element] = Operation(OperationCode::TextureGather, meta, std::move(coords_copy)); } @@ -152,7 +152,7 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { if (!instr.txq.IsComponentEnabled(element)) { continue; } - MetaTexture meta{sampler, {}, {}, {}, {}, {}, element}; + MetaTexture meta{sampler, {}, {}, {}, {}, {}, {}, element}; const Node value = Operation(OperationCode::TextureQueryDimensions, meta, GetRegister(instr.gpr8)); SetTemporal(bb, indexer++, value); @@ -202,7 +202,7 @@ u32 ShaderIR::DecodeTexture(NodeBlock& bb, u32 pc) { for (u32 element = 0; element < 2; ++element) { auto params = coords; - MetaTexture meta{sampler, {}, {}, {}, {}, {}, element}; + MetaTexture meta{sampler, {}, {}, {}, {}, {}, {}, element}; const Node value = Operation(OperationCode::TextureQueryLod, meta, std::move(params)); SetTemporal(bb, element, value); } @@ -325,7 +325,8 @@ void ShaderIR::WriteTexsInstructionHalfFloat(NodeBlock& bb, Instruction instr, Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, TextureProcessMode process_mode, std::vector<Node> coords, - Node array, Node depth_compare, u32 bias_offset) { + Node array, Node depth_compare, u32 bias_offset, + std::vector<Node> aoffi) { const bool is_array = array; const bool is_shadow = depth_compare; @@ -374,7 +375,7 @@ Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto copy_coords = coords; - MetaTexture meta{sampler, array, depth_compare, bias, lod, {}, element}; + MetaTexture meta{sampler, array, depth_compare, aoffi, bias, lod, {}, element}; values[element] = Operation(read_method, meta, std::move(copy_coords)); } @@ -382,9 +383,15 @@ Node4 ShaderIR::GetTextureCode(Instruction instr, TextureType texture_type, } Node4 ShaderIR::GetTexCode(Instruction instr, TextureType texture_type, - TextureProcessMode process_mode, bool depth_compare, bool is_array) { - const bool lod_bias_enabled = - (process_mode != TextureProcessMode::None && process_mode != TextureProcessMode::LZ); + TextureProcessMode process_mode, bool depth_compare, bool is_array, + bool is_aoffi) { + const bool lod_bias_enabled{ + (process_mode != TextureProcessMode::None && process_mode != TextureProcessMode::LZ)}; + + u64 parameter_register = instr.gpr20.Value(); + if (lod_bias_enabled) { + ++parameter_register; + } const auto [coord_count, total_coord_count] = ValidateAndGetCoordinateElement( texture_type, depth_compare, is_array, lod_bias_enabled, 4, 5); @@ -404,15 +411,19 @@ Node4 ShaderIR::GetTexCode(Instruction instr, TextureType texture_type, const Node array = is_array ? GetRegister(array_register) : nullptr; + std::vector<Node> aoffi; + if (is_aoffi) { + aoffi = GetAoffiCoordinates(GetRegister(parameter_register++), coord_count, false); + } + Node dc{}; if (depth_compare) { // Depth is always stored in the register signaled by gpr20 or in the next register if lod // or bias are used - const u64 depth_register = instr.gpr20.Value() + (lod_bias_enabled ? 1 : 0); - dc = GetRegister(depth_register); + dc = GetRegister(parameter_register++); } - return GetTextureCode(instr, texture_type, process_mode, coords, array, dc, 0); + return GetTextureCode(instr, texture_type, process_mode, coords, array, dc, 0, aoffi); } Node4 ShaderIR::GetTexsCode(Instruction instr, TextureType texture_type, @@ -448,11 +459,11 @@ Node4 ShaderIR::GetTexsCode(Instruction instr, TextureType texture_type, dc = GetRegister(depth_register); } - return GetTextureCode(instr, texture_type, process_mode, coords, array, dc, bias_offset); + return GetTextureCode(instr, texture_type, process_mode, coords, array, dc, bias_offset, {}); } Node4 ShaderIR::GetTld4Code(Instruction instr, TextureType texture_type, bool depth_compare, - bool is_array) { + bool is_array, bool is_aoffi) { const std::size_t coord_count = GetCoordCount(texture_type); const std::size_t total_coord_count = coord_count + (is_array ? 1 : 0); const std::size_t total_reg_count = total_coord_count + (depth_compare ? 1 : 0); @@ -463,15 +474,27 @@ Node4 ShaderIR::GetTld4Code(Instruction instr, TextureType texture_type, bool de const u64 coord_register = array_register + (is_array ? 1 : 0); std::vector<Node> coords; - for (size_t i = 0; i < coord_count; ++i) + for (std::size_t i = 0; i < coord_count; ++i) { coords.push_back(GetRegister(coord_register + i)); + } + + u64 parameter_register = instr.gpr20.Value(); + std::vector<Node> aoffi; + if (is_aoffi) { + aoffi = GetAoffiCoordinates(GetRegister(parameter_register++), coord_count, true); + } + + Node dc{}; + if (depth_compare) { + dc = GetRegister(parameter_register++); + } const auto& sampler = GetSampler(instr.sampler, texture_type, is_array, depth_compare); Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto coords_copy = coords; - MetaTexture meta{sampler, GetRegister(array_register), {}, {}, {}, {}, element}; + MetaTexture meta{sampler, GetRegister(array_register), dc, aoffi, {}, {}, {}, element}; values[element] = Operation(OperationCode::TextureGather, meta, std::move(coords_copy)); } @@ -507,7 +530,7 @@ Node4 ShaderIR::GetTldsCode(Instruction instr, TextureType texture_type, bool is Node4 values; for (u32 element = 0; element < values.size(); ++element) { auto coords_copy = coords; - MetaTexture meta{sampler, array, {}, {}, lod, {}, element}; + MetaTexture meta{sampler, array, {}, {}, {}, lod, {}, element}; values[element] = Operation(OperationCode::TexelFetch, meta, std::move(coords_copy)); } return values; @@ -531,4 +554,45 @@ std::tuple<std::size_t, std::size_t> ShaderIR::ValidateAndGetCoordinateElement( return {coord_count, total_coord_count}; } -} // namespace VideoCommon::Shader
\ No newline at end of file +std::vector<Node> ShaderIR::GetAoffiCoordinates(Node aoffi_reg, std::size_t coord_count, + bool is_tld4) { + const auto [coord_offsets, size, wrap_value, + diff_value] = [is_tld4]() -> std::tuple<std::array<u32, 3>, u32, s32, s32> { + if (is_tld4) { + return {{0, 8, 16}, 6, 32, 64}; + } else { + return {{0, 4, 8}, 4, 8, 16}; + } + }(); + const u32 mask = (1U << size) - 1; + + std::vector<Node> aoffi; + aoffi.reserve(coord_count); + + const auto aoffi_immediate{ + TrackImmediate(aoffi_reg, global_code, static_cast<s64>(global_code.size()))}; + if (!aoffi_immediate) { + // Variable access, not supported on AMD. + LOG_WARNING(HW_GPU, + "AOFFI constant folding failed, some hardware might have graphical issues"); + for (std::size_t coord = 0; coord < coord_count; ++coord) { + const Node value = BitfieldExtract(aoffi_reg, coord_offsets.at(coord), size); + const Node condition = + Operation(OperationCode::LogicalIGreaterEqual, value, Immediate(wrap_value)); + const Node negative = Operation(OperationCode::IAdd, value, Immediate(-diff_value)); + aoffi.push_back(Operation(OperationCode::Select, condition, negative, value)); + } + return aoffi; + } + + for (std::size_t coord = 0; coord < coord_count; ++coord) { + s32 value = (*aoffi_immediate >> coord_offsets.at(coord)) & mask; + if (value >= wrap_value) { + value -= diff_value; + } + aoffi.push_back(Immediate(value)); + } + return aoffi; +} + +} // namespace VideoCommon::Shader diff --git a/src/video_core/shader/decode/xmad.cpp b/src/video_core/shader/decode/xmad.cpp index c34843307..db15c0718 100644 --- a/src/video_core/shader/decode/xmad.cpp +++ b/src/video_core/shader/decode/xmad.cpp @@ -29,39 +29,55 @@ u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) { const bool is_signed_b = instr.xmad.sign_b == 1; const bool is_signed_c = is_signed_a; - auto [is_merge, op_b, op_c] = [&]() -> std::tuple<bool, Node, Node> { + auto [is_merge, is_psl, is_high_b, mode, op_b, + op_c] = [&]() -> std::tuple<bool, bool, bool, Tegra::Shader::XmadMode, Node, Node> { switch (opcode->get().GetId()) { case OpCode::Id::XMAD_CR: return {instr.xmad.merge_56, + instr.xmad.product_shift_left_second, + instr.xmad.high_b, + instr.xmad.mode_cbf, GetConstBuffer(instr.cbuf34.index, instr.cbuf34.GetOffset()), GetRegister(instr.gpr39)}; case OpCode::Id::XMAD_RR: - return {instr.xmad.merge_37, GetRegister(instr.gpr20), GetRegister(instr.gpr39)}; + return {instr.xmad.merge_37, instr.xmad.product_shift_left, instr.xmad.high_b_rr, + instr.xmad.mode, GetRegister(instr.gpr20), GetRegister(instr.gpr39)}; case OpCode::Id::XMAD_RC: - return {false, GetRegister(instr.gpr39), + return {false, + false, + instr.xmad.high_b, + instr.xmad.mode_cbf, + GetRegister(instr.gpr39), GetConstBuffer(instr.cbuf34.index, instr.cbuf34.GetOffset())}; case OpCode::Id::XMAD_IMM: - return {instr.xmad.merge_37, Immediate(static_cast<u32>(instr.xmad.imm20_16)), + return {instr.xmad.merge_37, + instr.xmad.product_shift_left, + false, + instr.xmad.mode, + Immediate(static_cast<u32>(instr.xmad.imm20_16)), GetRegister(instr.gpr39)}; } UNIMPLEMENTED_MSG("Unhandled XMAD instruction: {}", opcode->get().GetName()); - return {false, Immediate(0), Immediate(0)}; + return {false, false, false, Tegra::Shader::XmadMode::None, Immediate(0), Immediate(0)}; }(); op_a = BitfieldExtract(op_a, instr.xmad.high_a ? 16 : 0, 16); const Node original_b = op_b; - op_b = BitfieldExtract(op_b, instr.xmad.high_b ? 16 : 0, 16); + op_b = BitfieldExtract(op_b, is_high_b ? 16 : 0, 16); // TODO(Rodrigo): Use an appropiate sign for this operation Node product = Operation(OperationCode::IMul, NO_PRECISE, op_a, op_b); - if (instr.xmad.product_shift_left) { + if (is_psl) { product = Operation(OperationCode::ILogicalShiftLeft, NO_PRECISE, product, Immediate(16)); } + SetTemporal(bb, 0, product); + product = GetTemporal(0); const Node original_c = op_c; + const Tegra::Shader::XmadMode set_mode = mode; // Workaround to clang compile error op_c = [&]() { - switch (instr.xmad.mode) { + switch (set_mode) { case Tegra::Shader::XmadMode::None: return original_c; case Tegra::Shader::XmadMode::CLo: @@ -80,8 +96,13 @@ u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) { } }(); + SetTemporal(bb, 1, op_c); + op_c = GetTemporal(1); + // TODO(Rodrigo): Use an appropiate sign for this operation Node sum = Operation(OperationCode::IAdd, product, op_c); + SetTemporal(bb, 2, sum); + sum = GetTemporal(2); if (is_merge) { const Node a = BitfieldExtract(sum, 0, 16); const Node b = @@ -95,4 +116,4 @@ u32 ShaderIR::DecodeXmad(NodeBlock& bb, u32 pc) { return pc; } -} // namespace VideoCommon::Shader
\ No newline at end of file +} // namespace VideoCommon::Shader diff --git a/src/video_core/shader/shader_ir.h b/src/video_core/shader/shader_ir.h index 5bc3a3900..1afab08c0 100644 --- a/src/video_core/shader/shader_ir.h +++ b/src/video_core/shader/shader_ir.h @@ -7,6 +7,7 @@ #include <array> #include <cstring> #include <map> +#include <optional> #include <set> #include <string> #include <tuple> @@ -275,6 +276,11 @@ struct GlobalMemoryBase { } }; +struct GlobalMemoryUsage { + bool is_read{}; + bool is_written{}; +}; + struct MetaArithmetic { bool precise{}; }; @@ -290,6 +296,7 @@ struct MetaTexture { const Sampler& sampler; Node array{}; Node depth_compare{}; + std::vector<Node> aoffi; Node bias{}; Node lod{}; Node component{}; @@ -576,8 +583,8 @@ public: return used_clip_distances; } - const std::set<GlobalMemoryBase>& GetGlobalMemoryBases() const { - return used_global_memory_bases; + const std::map<GlobalMemoryBase, GlobalMemoryUsage>& GetGlobalMemory() const { + return used_global_memory; } std::size_t GetLength() const { @@ -741,14 +748,14 @@ private: Node4 GetTexCode(Tegra::Shader::Instruction instr, Tegra::Shader::TextureType texture_type, Tegra::Shader::TextureProcessMode process_mode, bool depth_compare, - bool is_array); + bool is_array, bool is_aoffi); Node4 GetTexsCode(Tegra::Shader::Instruction instr, Tegra::Shader::TextureType texture_type, Tegra::Shader::TextureProcessMode process_mode, bool depth_compare, bool is_array); Node4 GetTld4Code(Tegra::Shader::Instruction instr, Tegra::Shader::TextureType texture_type, - bool depth_compare, bool is_array); + bool depth_compare, bool is_array, bool is_aoffi); Node4 GetTldsCode(Tegra::Shader::Instruction instr, Tegra::Shader::TextureType texture_type, bool is_array); @@ -757,9 +764,11 @@ private: Tegra::Shader::TextureType texture_type, bool depth_compare, bool is_array, bool lod_bias_enabled, std::size_t max_coords, std::size_t max_inputs); + std::vector<Node> GetAoffiCoordinates(Node aoffi_reg, std::size_t coord_count, bool is_tld4); + Node4 GetTextureCode(Tegra::Shader::Instruction instr, Tegra::Shader::TextureType texture_type, Tegra::Shader::TextureProcessMode process_mode, std::vector<Node> coords, - Node array, Node depth_compare, u32 bias_offset); + Node array, Node depth_compare, u32 bias_offset, std::vector<Node> aoffi); Node GetVideoOperand(Node op, bool is_chunk, bool is_signed, Tegra::Shader::VideoType type, u64 byte_height); @@ -773,8 +782,15 @@ private: Node TrackCbuf(Node tracked, const NodeBlock& code, s64 cursor); + std::optional<u32> TrackImmediate(Node tracked, const NodeBlock& code, s64 cursor); + std::pair<Node, s64> TrackRegister(const GprNode* tracked, const NodeBlock& code, s64 cursor); + std::tuple<Node, Node, GlobalMemoryBase> TrackAndGetGlobalMemory(NodeBlock& bb, + Node addr_register, + u32 immediate_offset, + bool is_write); + template <typename... T> Node Operation(OperationCode code, const T*... operands) { return StoreNode(OperationNode(code, operands...)); @@ -828,7 +844,7 @@ private: std::map<u32, ConstBuffer> used_cbufs; std::set<Sampler> used_samplers; std::array<bool, Tegra::Engines::Maxwell3D::Regs::NumClipDistances> used_clip_distances{}; - std::set<GlobalMemoryBase> used_global_memory_bases; + std::map<GlobalMemoryBase, GlobalMemoryUsage> used_global_memory; Tegra::Shader::Header header; }; diff --git a/src/video_core/shader/track.cpp b/src/video_core/shader/track.cpp index 33b071747..4505667ff 100644 --- a/src/video_core/shader/track.cpp +++ b/src/video_core/shader/track.cpp @@ -6,6 +6,7 @@ #include <utility> #include <variant> +#include "common/common_types.h" #include "video_core/shader/shader_ir.h" namespace VideoCommon::Shader { @@ -14,7 +15,7 @@ namespace { std::pair<Node, s64> FindOperation(const NodeBlock& code, s64 cursor, OperationCode operation_code) { for (; cursor >= 0; --cursor) { - const Node node = code[cursor]; + const Node node = code.at(cursor); if (const auto operation = std::get_if<OperationNode>(node)) { if (operation->GetCode() == operation_code) return {node, cursor}; @@ -64,6 +65,20 @@ Node ShaderIR::TrackCbuf(Node tracked, const NodeBlock& code, s64 cursor) { return nullptr; } +std::optional<u32> ShaderIR::TrackImmediate(Node tracked, const NodeBlock& code, s64 cursor) { + // Reduce the cursor in one to avoid infinite loops when the instruction sets the same register + // that it uses as operand + const auto [found, found_cursor] = + TrackRegister(&std::get<GprNode>(*tracked), code, cursor - 1); + if (!found) { + return {}; + } + if (const auto immediate = std::get_if<ImmediateNode>(found)) { + return immediate->GetValue(); + } + return {}; +} + std::pair<Node, s64> ShaderIR::TrackRegister(const GprNode* tracked, const NodeBlock& code, s64 cursor) { for (; cursor >= 0; --cursor) { diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp index a7ac26d71..3b022a456 100644 --- a/src/video_core/surface.cpp +++ b/src/video_core/surface.cpp @@ -294,6 +294,8 @@ PixelFormat PixelFormatFromTextureFormat(Tegra::Texture::TextureFormat format, return PixelFormat::Z16; case Tegra::Texture::TextureFormat::Z24S8: return PixelFormat::Z24S8; + case Tegra::Texture::TextureFormat::ZF32_X24S8: + return PixelFormat::Z32FS8; case Tegra::Texture::TextureFormat::DXT1: return is_srgb ? PixelFormat::DXT1_SRGB : PixelFormat::DXT1; case Tegra::Texture::TextureFormat::DXT23: diff --git a/src/video_core/texture_cache.cpp b/src/video_core/texture_cache.cpp new file mode 100644 index 000000000..e96eba7cc --- /dev/null +++ b/src/video_core/texture_cache.cpp @@ -0,0 +1,386 @@ +// Copyright 2019 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include "common/alignment.h" +#include "common/assert.h" +#include "common/cityhash.h" +#include "common/common_types.h" +#include "core/core.h" +#include "video_core/surface.h" +#include "video_core/texture_cache.h" +#include "video_core/textures/decoders.h" +#include "video_core/textures/texture.h" + +namespace VideoCommon { + +using VideoCore::Surface::SurfaceTarget; + +using VideoCore::Surface::ComponentTypeFromDepthFormat; +using VideoCore::Surface::ComponentTypeFromRenderTarget; +using VideoCore::Surface::ComponentTypeFromTexture; +using VideoCore::Surface::PixelFormatFromDepthFormat; +using VideoCore::Surface::PixelFormatFromRenderTargetFormat; +using VideoCore::Surface::PixelFormatFromTextureFormat; +using VideoCore::Surface::SurfaceTargetFromTextureType; + +constexpr u32 GetMipmapSize(bool uncompressed, u32 mip_size, u32 tile) { + return uncompressed ? mip_size : std::max(1U, (mip_size + tile - 1) / tile); +} + +SurfaceParams SurfaceParams::CreateForTexture(Core::System& system, + const Tegra::Texture::FullTextureInfo& config) { + SurfaceParams params; + params.is_tiled = config.tic.IsTiled(); + params.block_width = params.is_tiled ? config.tic.BlockWidth() : 0, + params.block_height = params.is_tiled ? config.tic.BlockHeight() : 0, + params.block_depth = params.is_tiled ? config.tic.BlockDepth() : 0, + params.tile_width_spacing = params.is_tiled ? (1 << config.tic.tile_width_spacing.Value()) : 1; + params.pixel_format = + PixelFormatFromTextureFormat(config.tic.format, config.tic.r_type.Value(), false); + params.component_type = ComponentTypeFromTexture(config.tic.r_type.Value()); + params.type = GetFormatType(params.pixel_format); + params.target = SurfaceTargetFromTextureType(config.tic.texture_type); + params.width = Common::AlignUp(config.tic.Width(), GetCompressionFactor(params.pixel_format)); + params.height = Common::AlignUp(config.tic.Height(), GetCompressionFactor(params.pixel_format)); + params.depth = config.tic.Depth(); + if (params.target == SurfaceTarget::TextureCubemap || + params.target == SurfaceTarget::TextureCubeArray) { + params.depth *= 6; + } + params.pitch = params.is_tiled ? 0 : config.tic.Pitch(); + params.unaligned_height = config.tic.Height(); + params.num_levels = config.tic.max_mip_level + 1; + + params.CalculateCachedValues(); + return params; +} + +SurfaceParams SurfaceParams::CreateForDepthBuffer( + Core::System& system, u32 zeta_width, u32 zeta_height, Tegra::DepthFormat format, + u32 block_width, u32 block_height, u32 block_depth, + Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout type) { + SurfaceParams params; + params.is_tiled = type == Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear; + params.block_width = 1 << std::min(block_width, 5U); + params.block_height = 1 << std::min(block_height, 5U); + params.block_depth = 1 << std::min(block_depth, 5U); + params.tile_width_spacing = 1; + params.pixel_format = PixelFormatFromDepthFormat(format); + params.component_type = ComponentTypeFromDepthFormat(format); + params.type = GetFormatType(params.pixel_format); + params.width = zeta_width; + params.height = zeta_height; + params.unaligned_height = zeta_height; + params.target = SurfaceTarget::Texture2D; + params.depth = 1; + params.num_levels = 1; + + params.CalculateCachedValues(); + return params; +} + +SurfaceParams SurfaceParams::CreateForFramebuffer(Core::System& system, std::size_t index) { + const auto& config{system.GPU().Maxwell3D().regs.rt[index]}; + SurfaceParams params; + params.is_tiled = + config.memory_layout.type == Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout::BlockLinear; + params.block_width = 1 << config.memory_layout.block_width; + params.block_height = 1 << config.memory_layout.block_height; + params.block_depth = 1 << config.memory_layout.block_depth; + params.tile_width_spacing = 1; + params.pixel_format = PixelFormatFromRenderTargetFormat(config.format); + params.component_type = ComponentTypeFromRenderTarget(config.format); + params.type = GetFormatType(params.pixel_format); + if (params.is_tiled) { + 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.depth = 1; + params.unaligned_height = config.height; + params.target = SurfaceTarget::Texture2D; + params.num_levels = 1; + + params.CalculateCachedValues(); + return params; +} + +SurfaceParams SurfaceParams::CreateForFermiCopySurface( + const Tegra::Engines::Fermi2D::Regs::Surface& config) { + SurfaceParams params{}; + params.is_tiled = !config.linear; + params.block_width = params.is_tiled ? std::min(config.BlockWidth(), 32U) : 0, + params.block_height = params.is_tiled ? std::min(config.BlockHeight(), 32U) : 0, + params.block_depth = params.is_tiled ? std::min(config.BlockDepth(), 32U) : 0, + params.tile_width_spacing = 1; + params.pixel_format = PixelFormatFromRenderTargetFormat(config.format); + params.component_type = ComponentTypeFromRenderTarget(config.format); + params.type = GetFormatType(params.pixel_format); + params.width = config.width; + params.height = config.height; + params.unaligned_height = config.height; + // TODO(Rodrigo): Try to guess the surface target from depth and layer parameters + params.target = SurfaceTarget::Texture2D; + params.depth = 1; + params.num_levels = 1; + + params.CalculateCachedValues(); + return params; +} + +u32 SurfaceParams::GetMipWidth(u32 level) const { + return std::max(1U, width >> level); +} + +u32 SurfaceParams::GetMipHeight(u32 level) const { + return std::max(1U, height >> level); +} + +u32 SurfaceParams::GetMipDepth(u32 level) const { + return IsLayered() ? depth : std::max(1U, depth >> level); +} + +bool SurfaceParams::IsLayered() const { + switch (target) { + case SurfaceTarget::Texture1DArray: + case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubeArray: + case SurfaceTarget::TextureCubemap: + return true; + default: + return false; + } +} + +u32 SurfaceParams::GetMipBlockHeight(u32 level) const { + // Auto block resizing algorithm from: + // https://cgit.freedesktop.org/mesa/mesa/tree/src/gallium/drivers/nouveau/nv50/nv50_miptree.c + if (level == 0) { + return block_height; + } + const u32 height{GetMipHeight(level)}; + const u32 default_block_height{GetDefaultBlockHeight(pixel_format)}; + const u32 blocks_in_y{(height + default_block_height - 1) / default_block_height}; + u32 block_height = 16; + while (block_height > 1 && blocks_in_y <= block_height * 4) { + block_height >>= 1; + } + return block_height; +} + +u32 SurfaceParams::GetMipBlockDepth(u32 level) const { + if (level == 0) + return block_depth; + if (target != SurfaceTarget::Texture3D) + return 1; + + const u32 depth{GetMipDepth(level)}; + u32 block_depth = 32; + while (block_depth > 1 && depth * 2 <= block_depth) { + block_depth >>= 1; + } + if (block_depth == 32 && GetMipBlockHeight(level) >= 4) { + return 16; + } + return block_depth; +} + +std::size_t SurfaceParams::GetGuestMipmapLevelOffset(u32 level) const { + std::size_t offset = 0; + for (u32 i = 0; i < level; i++) { + offset += GetInnerMipmapMemorySize(i, false, IsLayered(), false); + } + return offset; +} + +std::size_t SurfaceParams::GetHostMipmapLevelOffset(u32 level) const { + std::size_t offset = 0; + for (u32 i = 0; i < level; i++) { + offset += GetInnerMipmapMemorySize(i, true, false, false); + } + return offset; +} + +std::size_t SurfaceParams::GetGuestLayerSize() const { + return GetInnerMemorySize(false, true, false); +} + +std::size_t SurfaceParams::GetHostLayerSize(u32 level) const { + return GetInnerMipmapMemorySize(level, true, IsLayered(), false); +} + +bool SurfaceParams::IsFamiliar(const SurfaceParams& view_params) const { + if (std::tie(is_tiled, tile_width_spacing, pixel_format, component_type, type) != + std::tie(view_params.is_tiled, view_params.tile_width_spacing, view_params.pixel_format, + view_params.component_type, view_params.type)) { + return false; + } + + const SurfaceTarget view_target{view_params.target}; + if (view_target == target) { + return true; + } + + switch (target) { + case SurfaceTarget::Texture1D: + case SurfaceTarget::Texture2D: + case SurfaceTarget::Texture3D: + return false; + case SurfaceTarget::Texture1DArray: + return view_target == SurfaceTarget::Texture1D; + case SurfaceTarget::Texture2DArray: + return view_target == SurfaceTarget::Texture2D; + case SurfaceTarget::TextureCubemap: + return view_target == SurfaceTarget::Texture2D || + view_target == SurfaceTarget::Texture2DArray; + case SurfaceTarget::TextureCubeArray: + return view_target == SurfaceTarget::Texture2D || + view_target == SurfaceTarget::Texture2DArray || + view_target == SurfaceTarget::TextureCubemap; + default: + UNIMPLEMENTED_MSG("Unimplemented texture family={}", static_cast<u32>(target)); + return false; + } +} + +bool SurfaceParams::IsPixelFormatZeta() const { + return pixel_format >= VideoCore::Surface::PixelFormat::MaxColorFormat && + pixel_format < VideoCore::Surface::PixelFormat::MaxDepthStencilFormat; +} + +void SurfaceParams::CalculateCachedValues() { + guest_size_in_bytes = GetInnerMemorySize(false, false, false); + + // ASTC is uncompressed in software, in emulated as RGBA8 + if (IsPixelFormatASTC(pixel_format)) { + host_size_in_bytes = width * height * depth * 4; + } else { + host_size_in_bytes = GetInnerMemorySize(true, false, false); + } + + switch (target) { + case SurfaceTarget::Texture1D: + case SurfaceTarget::Texture2D: + case SurfaceTarget::Texture3D: + num_layers = 1; + break; + case SurfaceTarget::Texture1DArray: + case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubemap: + case SurfaceTarget::TextureCubeArray: + num_layers = depth; + break; + default: + UNREACHABLE(); + } +} + +std::size_t SurfaceParams::GetInnerMipmapMemorySize(u32 level, bool as_host_size, bool layer_only, + bool uncompressed) const { + const bool tiled{as_host_size ? false : is_tiled}; + const u32 tile_x{GetDefaultBlockWidth(pixel_format)}; + const u32 tile_y{GetDefaultBlockHeight(pixel_format)}; + const u32 width{GetMipmapSize(uncompressed, GetMipWidth(level), tile_x)}; + const u32 height{GetMipmapSize(uncompressed, GetMipHeight(level), tile_y)}; + const u32 depth{layer_only ? 1U : GetMipDepth(level)}; + return Tegra::Texture::CalculateSize(tiled, GetBytesPerPixel(pixel_format), width, height, + depth, GetMipBlockHeight(level), GetMipBlockDepth(level)); +} + +std::size_t SurfaceParams::GetInnerMemorySize(bool as_host_size, bool layer_only, + bool uncompressed) const { + std::size_t size = 0; + for (u32 level = 0; level < num_levels; ++level) { + size += GetInnerMipmapMemorySize(level, as_host_size, layer_only, uncompressed); + } + if (!as_host_size && is_tiled) { + size = Common::AlignUp(size, Tegra::Texture::GetGOBSize() * block_height * block_depth); + } + return size; +} + +std::map<u64, std::pair<u32, u32>> SurfaceParams::CreateViewOffsetMap() const { + std::map<u64, std::pair<u32, u32>> view_offset_map; + switch (target) { + case SurfaceTarget::Texture1D: + case SurfaceTarget::Texture2D: + case SurfaceTarget::Texture3D: { + constexpr u32 layer = 0; + for (u32 level = 0; level < num_levels; ++level) { + const std::size_t offset{GetGuestMipmapLevelOffset(level)}; + view_offset_map.insert({offset, {layer, level}}); + } + break; + } + case SurfaceTarget::Texture1DArray: + case SurfaceTarget::Texture2DArray: + case SurfaceTarget::TextureCubemap: + case SurfaceTarget::TextureCubeArray: { + const std::size_t layer_size{GetGuestLayerSize()}; + for (u32 level = 0; level < num_levels; ++level) { + const std::size_t level_offset{GetGuestMipmapLevelOffset(level)}; + for (u32 layer = 0; layer < num_layers; ++layer) { + const auto layer_offset{static_cast<std::size_t>(layer_size * layer)}; + const std::size_t offset{level_offset + layer_offset}; + view_offset_map.insert({offset, {layer, level}}); + } + } + break; + } + default: + UNIMPLEMENTED_MSG("Unimplemented surface target {}", static_cast<u32>(target)); + } + return view_offset_map; +} + +bool SurfaceParams::IsViewValid(const SurfaceParams& view_params, u32 layer, u32 level) const { + return IsDimensionValid(view_params, level) && IsDepthValid(view_params, level) && + IsInBounds(view_params, layer, level); +} + +bool SurfaceParams::IsDimensionValid(const SurfaceParams& view_params, u32 level) const { + return view_params.width == GetMipWidth(level) && view_params.height == GetMipHeight(level); +} + +bool SurfaceParams::IsDepthValid(const SurfaceParams& view_params, u32 level) const { + if (view_params.target != SurfaceTarget::Texture3D) { + return true; + } + return view_params.depth == GetMipDepth(level); +} + +bool SurfaceParams::IsInBounds(const SurfaceParams& view_params, u32 layer, u32 level) const { + return layer + view_params.num_layers <= num_layers && + level + view_params.num_levels <= num_levels; +} + +std::size_t HasheableSurfaceParams::Hash() const { + return static_cast<std::size_t>( + Common::CityHash64(reinterpret_cast<const char*>(this), sizeof(*this))); +} + +bool HasheableSurfaceParams::operator==(const HasheableSurfaceParams& rhs) const { + return std::tie(is_tiled, block_width, block_height, block_depth, tile_width_spacing, width, + height, depth, pitch, unaligned_height, num_levels, pixel_format, + component_type, 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.unaligned_height, rhs.num_levels, rhs.pixel_format, rhs.component_type, + rhs.type, rhs.target); +} + +std::size_t ViewKey::Hash() const { + return static_cast<std::size_t>( + Common::CityHash64(reinterpret_cast<const char*>(this), sizeof(*this))); +} + +bool ViewKey::operator==(const ViewKey& rhs) const { + return std::tie(base_layer, num_layers, base_level, num_levels) == + std::tie(rhs.base_layer, rhs.num_layers, rhs.base_level, rhs.num_levels); +} + +} // namespace VideoCommon diff --git a/src/video_core/texture_cache.h b/src/video_core/texture_cache.h new file mode 100644 index 000000000..041551691 --- /dev/null +++ b/src/video_core/texture_cache.h @@ -0,0 +1,586 @@ +// Copyright 2019 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <list> +#include <memory> +#include <set> +#include <tuple> +#include <type_traits> +#include <unordered_map> + +#include <boost/icl/interval_map.hpp> +#include <boost/range/iterator_range.hpp> + +#include "common/assert.h" +#include "common/common_types.h" +#include "core/memory.h" +#include "video_core/engines/fermi_2d.h" +#include "video_core/engines/maxwell_3d.h" +#include "video_core/gpu.h" +#include "video_core/rasterizer_interface.h" +#include "video_core/surface.h" + +namespace Core { +class System; +} + +namespace Tegra::Texture { +struct FullTextureInfo; +} + +namespace VideoCore { +class RasterizerInterface; +} + +namespace VideoCommon { + +class HasheableSurfaceParams { +public: + std::size_t Hash() const; + + bool operator==(const HasheableSurfaceParams& rhs) const; + +protected: + // Avoid creation outside of a managed environment. + HasheableSurfaceParams() = default; + + bool is_tiled; + u32 block_width; + u32 block_height; + u32 block_depth; + u32 tile_width_spacing; + u32 width; + u32 height; + u32 depth; + u32 pitch; + u32 unaligned_height; + u32 num_levels; + VideoCore::Surface::PixelFormat pixel_format; + VideoCore::Surface::ComponentType component_type; + VideoCore::Surface::SurfaceType type; + VideoCore::Surface::SurfaceTarget target; +}; + +class SurfaceParams final : public HasheableSurfaceParams { +public: + /// Creates SurfaceCachedParams from a texture configuration. + static SurfaceParams CreateForTexture(Core::System& system, + const Tegra::Texture::FullTextureInfo& config); + + /// Creates SurfaceCachedParams for a depth buffer configuration. + static SurfaceParams CreateForDepthBuffer( + Core::System& system, u32 zeta_width, u32 zeta_height, Tegra::DepthFormat format, + u32 block_width, u32 block_height, u32 block_depth, + Tegra::Engines::Maxwell3D::Regs::InvMemoryLayout type); + + /// Creates SurfaceCachedParams from a framebuffer configuration. + static SurfaceParams CreateForFramebuffer(Core::System& system, std::size_t index); + + /// Creates SurfaceCachedParams from a Fermi2D surface configuration. + static SurfaceParams CreateForFermiCopySurface( + const Tegra::Engines::Fermi2D::Regs::Surface& config); + + bool IsTiled() const { + return is_tiled; + } + + u32 GetBlockWidth() const { + return block_width; + } + + u32 GetTileWidthSpacing() const { + return tile_width_spacing; + } + + u32 GetWidth() const { + return width; + } + + u32 GetHeight() const { + return height; + } + + u32 GetDepth() const { + return depth; + } + + u32 GetPitch() const { + return pitch; + } + + u32 GetNumLevels() const { + return num_levels; + } + + VideoCore::Surface::PixelFormat GetPixelFormat() const { + return pixel_format; + } + + VideoCore::Surface::ComponentType GetComponentType() const { + return component_type; + } + + VideoCore::Surface::SurfaceTarget GetTarget() const { + return target; + } + + VideoCore::Surface::SurfaceType GetType() const { + return type; + } + + std::size_t GetGuestSizeInBytes() const { + return guest_size_in_bytes; + } + + std::size_t GetHostSizeInBytes() const { + return host_size_in_bytes; + } + + u32 GetNumLayers() const { + return num_layers; + } + + /// Returns the width of a given mipmap level. + u32 GetMipWidth(u32 level) const; + + /// Returns the height of a given mipmap level. + u32 GetMipHeight(u32 level) const; + + /// Returns the depth of a given mipmap level. + u32 GetMipDepth(u32 level) const; + + /// Returns true if these parameters are from a layered surface. + bool IsLayered() const; + + /// 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 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) const; + + /// Returns the size of a layer in bytes in guest memory. + std::size_t GetGuestLayerSize() const; + + /// Returns the size of a layer in bytes in host memory for a given mipmap level. + std::size_t GetHostLayerSize(u32 level) const; + + /// Returns true if another surface can be familiar with this. This is a loosely defined term + /// that reflects the possibility of these two surface parameters potentially being part of a + /// bigger superset. + bool IsFamiliar(const SurfaceParams& view_params) const; + + /// Returns true if the pixel format is a depth and/or stencil format. + bool IsPixelFormatZeta() const; + + /// Creates a map that redirects an address difference to a layer and mipmap level. + std::map<u64, std::pair<u32, u32>> CreateViewOffsetMap() const; + + /// Returns true if the passed surface view parameters is equal or a valid subset of this. + bool IsViewValid(const SurfaceParams& view_params, u32 layer, u32 level) const; + +private: + /// Calculates values that can be deduced from HasheableSurfaceParams. + void CalculateCachedValues(); + + /// Returns the size of a given mipmap level. + std::size_t GetInnerMipmapMemorySize(u32 level, bool as_host_size, bool layer_only, + 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; + + /// Returns true if the passed view width and height match the size of this params in a given + /// mipmap level. + bool IsDimensionValid(const SurfaceParams& view_params, u32 level) const; + + /// Returns true if the passed view depth match the size of this params in a given mipmap level. + bool IsDepthValid(const SurfaceParams& view_params, u32 level) const; + + /// Returns true if the passed view layers and mipmap levels are in bounds. + bool IsInBounds(const SurfaceParams& view_params, u32 layer, u32 level) const; + + std::size_t guest_size_in_bytes; + std::size_t host_size_in_bytes; + u32 num_layers; +}; + +struct ViewKey { + std::size_t Hash() const; + + bool operator==(const ViewKey& rhs) const; + + u32 base_layer{}; + u32 num_layers{}; + u32 base_level{}; + u32 num_levels{}; +}; + +} // namespace VideoCommon + +namespace std { + +template <> +struct hash<VideoCommon::SurfaceParams> { + std::size_t operator()(const VideoCommon::SurfaceParams& k) const noexcept { + return k.Hash(); + } +}; + +template <> +struct hash<VideoCommon::ViewKey> { + std::size_t operator()(const VideoCommon::ViewKey& k) const noexcept { + return k.Hash(); + } +}; + +} // namespace std + +namespace VideoCommon { + +template <typename TView, typename TExecutionContext> +class SurfaceBase { + static_assert(std::is_trivially_copyable_v<TExecutionContext>); + +public: + virtual void LoadBuffer() = 0; + + virtual TExecutionContext FlushBuffer(TExecutionContext exctx) = 0; + + virtual TExecutionContext UploadTexture(TExecutionContext exctx) = 0; + + TView* TryGetView(VAddr view_addr, const SurfaceParams& view_params) { + if (view_addr < cpu_addr || !params.IsFamiliar(view_params)) { + // It can't be a view if it's in a prior address. + return {}; + } + + const auto relative_offset{static_cast<u64>(view_addr - cpu_addr)}; + const auto it{view_offset_map.find(relative_offset)}; + if (it == view_offset_map.end()) { + // Couldn't find an aligned view. + return {}; + } + const auto [layer, level] = it->second; + + if (!params.IsViewValid(view_params, layer, level)) { + return {}; + } + + return GetView(layer, view_params.GetNumLayers(), level, view_params.GetNumLevels()); + } + + VAddr GetCpuAddr() const { + ASSERT(is_registered); + return cpu_addr; + } + + u8* GetHostPtr() const { + ASSERT(is_registered); + return host_ptr; + } + + CacheAddr GetCacheAddr() const { + ASSERT(is_registered); + return cache_addr; + } + + std::size_t GetSizeInBytes() const { + return params.GetGuestSizeInBytes(); + } + + void MarkAsModified(bool is_modified_) { + is_modified = is_modified_; + } + + const SurfaceParams& GetSurfaceParams() const { + return params; + } + + TView* GetView(VAddr view_addr, const SurfaceParams& view_params) { + TView* view{TryGetView(view_addr, view_params)}; + ASSERT(view != nullptr); + return view; + } + + void Register(VAddr cpu_addr_, u8* host_ptr_) { + ASSERT(!is_registered); + is_registered = true; + cpu_addr = cpu_addr_; + host_ptr = host_ptr_; + cache_addr = ToCacheAddr(host_ptr_); + } + + void Register(VAddr cpu_addr_) { + Register(cpu_addr_, Memory::GetPointer(cpu_addr_)); + } + + void Unregister() { + ASSERT(is_registered); + is_registered = false; + } + + bool IsRegistered() const { + return is_registered; + } + +protected: + explicit SurfaceBase(const SurfaceParams& params) + : params{params}, view_offset_map{params.CreateViewOffsetMap()} {} + + ~SurfaceBase() = default; + + virtual std::unique_ptr<TView> CreateView(const ViewKey& view_key) = 0; + + bool IsModified() const { + return is_modified; + } + + const SurfaceParams params; + +private: + TView* GetView(u32 base_layer, u32 num_layers, u32 base_level, u32 num_levels) { + const ViewKey key{base_layer, num_layers, base_level, num_levels}; + const auto [entry, is_cache_miss] = views.try_emplace(key); + auto& view{entry->second}; + if (is_cache_miss) { + view = CreateView(key); + } + return view.get(); + } + + const std::map<u64, std::pair<u32, u32>> view_offset_map; + + VAddr cpu_addr{}; + u8* host_ptr{}; + CacheAddr cache_addr{}; + bool is_modified{}; + bool is_registered{}; + std::unordered_map<ViewKey, std::unique_ptr<TView>> views; +}; + +template <typename TSurface, typename TView, typename TExecutionContext> +class TextureCache { + static_assert(std::is_trivially_copyable_v<TExecutionContext>); + using ResultType = std::tuple<TView*, TExecutionContext>; + using IntervalMap = boost::icl::interval_map<CacheAddr, std::set<TSurface*>>; + using IntervalType = typename IntervalMap::interval_type; + +public: + void InvalidateRegion(CacheAddr addr, std::size_t size) { + for (TSurface* surface : GetSurfacesInRegion(addr, size)) { + if (!surface->IsRegistered()) { + // Skip duplicates + continue; + } + Unregister(surface); + } + } + + ResultType GetTextureSurface(TExecutionContext exctx, + const Tegra::Texture::FullTextureInfo& config) { + auto& memory_manager{system.GPU().MemoryManager()}; + const auto cpu_addr{memory_manager.GpuToCpuAddress(config.tic.Address())}; + if (!cpu_addr) { + return {{}, exctx}; + } + const auto params{SurfaceParams::CreateForTexture(system, config)}; + return GetSurfaceView(exctx, *cpu_addr, params, true); + } + + ResultType GetDepthBufferSurface(TExecutionContext exctx, bool preserve_contents) { + const auto& regs{system.GPU().Maxwell3D().regs}; + if (!regs.zeta.Address() || !regs.zeta_enable) { + return {{}, exctx}; + } + + auto& memory_manager{system.GPU().MemoryManager()}; + const auto cpu_addr{memory_manager.GpuToCpuAddress(regs.zeta.Address())}; + if (!cpu_addr) { + return {{}, exctx}; + } + + const auto depth_params{SurfaceParams::CreateForDepthBuffer( + system, regs.zeta_width, regs.zeta_height, regs.zeta.format, + regs.zeta.memory_layout.block_width, regs.zeta.memory_layout.block_height, + regs.zeta.memory_layout.block_depth, regs.zeta.memory_layout.type)}; + return GetSurfaceView(exctx, *cpu_addr, depth_params, preserve_contents); + } + + ResultType GetColorBufferSurface(TExecutionContext exctx, std::size_t index, + bool preserve_contents) { + ASSERT(index < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets); + + const auto& regs{system.GPU().Maxwell3D().regs}; + if (index >= regs.rt_control.count || regs.rt[index].Address() == 0 || + regs.rt[index].format == Tegra::RenderTargetFormat::NONE) { + return {{}, exctx}; + } + + auto& memory_manager{system.GPU().MemoryManager()}; + const auto& config{system.GPU().Maxwell3D().regs.rt[index]}; + const auto cpu_addr{memory_manager.GpuToCpuAddress( + config.Address() + config.base_layer * config.layer_stride * sizeof(u32))}; + if (!cpu_addr) { + return {{}, exctx}; + } + + return GetSurfaceView(exctx, *cpu_addr, SurfaceParams::CreateForFramebuffer(system, index), + preserve_contents); + } + + ResultType GetFermiSurface(TExecutionContext exctx, + const Tegra::Engines::Fermi2D::Regs::Surface& config) { + const auto cpu_addr{system.GPU().MemoryManager().GpuToCpuAddress(config.Address())}; + ASSERT(cpu_addr); + return GetSurfaceView(exctx, *cpu_addr, SurfaceParams::CreateForFermiCopySurface(config), + true); + } + + TSurface* TryFindFramebufferSurface(const u8* host_ptr) const { + const auto it{registered_surfaces.find(ToCacheAddr(host_ptr))}; + return it != registered_surfaces.end() ? *it->second.begin() : nullptr; + } + +protected: + TextureCache(Core::System& system, VideoCore::RasterizerInterface& rasterizer) + : system{system}, rasterizer{rasterizer} {} + + ~TextureCache() = default; + + virtual ResultType TryFastGetSurfaceView(TExecutionContext exctx, VAddr cpu_addr, u8* host_ptr, + const SurfaceParams& params, bool preserve_contents, + const std::vector<TSurface*>& overlaps) = 0; + + virtual std::unique_ptr<TSurface> CreateSurface(const SurfaceParams& params) = 0; + + void Register(TSurface* surface, VAddr cpu_addr, u8* host_ptr) { + surface->Register(cpu_addr, host_ptr); + registered_surfaces.add({GetSurfaceInterval(surface), {surface}}); + rasterizer.UpdatePagesCachedCount(surface->GetCpuAddr(), surface->GetSizeInBytes(), 1); + } + + void Unregister(TSurface* surface) { + registered_surfaces.subtract({GetSurfaceInterval(surface), {surface}}); + rasterizer.UpdatePagesCachedCount(surface->GetCpuAddr(), surface->GetSizeInBytes(), -1); + surface->Unregister(); + } + + TSurface* GetUncachedSurface(const SurfaceParams& params) { + if (TSurface* surface = TryGetReservedSurface(params); surface) + return surface; + // No reserved surface available, create a new one and reserve it + auto new_surface{CreateSurface(params)}; + TSurface* surface{new_surface.get()}; + ReserveSurface(params, std::move(new_surface)); + return surface; + } + + Core::System& system; + +private: + ResultType GetSurfaceView(TExecutionContext exctx, VAddr cpu_addr, const SurfaceParams& params, + bool preserve_contents) { + const auto host_ptr{Memory::GetPointer(cpu_addr)}; + const auto cache_addr{ToCacheAddr(host_ptr)}; + const auto overlaps{GetSurfacesInRegion(cache_addr, params.GetGuestSizeInBytes())}; + if (overlaps.empty()) { + return LoadSurfaceView(exctx, cpu_addr, host_ptr, params, preserve_contents); + } + + if (overlaps.size() == 1) { + if (TView* view = overlaps[0]->TryGetView(cpu_addr, params); view) + return {view, exctx}; + } + + TView* fast_view; + std::tie(fast_view, exctx) = + TryFastGetSurfaceView(exctx, cpu_addr, host_ptr, params, preserve_contents, overlaps); + + for (TSurface* surface : overlaps) { + if (!fast_view) { + // Flush even when we don't care about the contents, to preserve memory not written + // by the new surface. + exctx = surface->FlushBuffer(exctx); + } + Unregister(surface); + } + + if (fast_view) { + return {fast_view, exctx}; + } + + return LoadSurfaceView(exctx, cpu_addr, host_ptr, params, preserve_contents); + } + + ResultType LoadSurfaceView(TExecutionContext exctx, VAddr cpu_addr, u8* host_ptr, + const SurfaceParams& params, bool preserve_contents) { + TSurface* new_surface{GetUncachedSurface(params)}; + Register(new_surface, cpu_addr, host_ptr); + if (preserve_contents) { + exctx = LoadSurface(exctx, new_surface); + } + return {new_surface->GetView(cpu_addr, params), exctx}; + } + + TExecutionContext LoadSurface(TExecutionContext exctx, TSurface* surface) { + surface->LoadBuffer(); + exctx = surface->UploadTexture(exctx); + surface->MarkAsModified(false); + return exctx; + } + + std::vector<TSurface*> GetSurfacesInRegion(CacheAddr cache_addr, std::size_t size) const { + if (size == 0) { + return {}; + } + const IntervalType interval{cache_addr, cache_addr + size}; + + std::vector<TSurface*> surfaces; + for (auto& pair : boost::make_iterator_range(registered_surfaces.equal_range(interval))) { + surfaces.push_back(*pair.second.begin()); + } + return surfaces; + } + + void ReserveSurface(const SurfaceParams& params, std::unique_ptr<TSurface> surface) { + surface_reserve[params].push_back(std::move(surface)); + } + + TSurface* TryGetReservedSurface(const SurfaceParams& params) { + auto search{surface_reserve.find(params)}; + if (search == surface_reserve.end()) { + return {}; + } + for (auto& surface : search->second) { + if (!surface->IsRegistered()) { + return surface.get(); + } + } + return {}; + } + + IntervalType GetSurfaceInterval(TSurface* surface) const { + return IntervalType::right_open(surface->GetCacheAddr(), + surface->GetCacheAddr() + surface->GetSizeInBytes()); + } + + VideoCore::RasterizerInterface& rasterizer; + + IntervalMap registered_surfaces; + + /// 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::list<std::unique_ptr<TSurface>>> surface_reserve; +}; + +} // namespace VideoCommon diff --git a/src/video_core/textures/convert.cpp b/src/video_core/textures/convert.cpp index 5e439f036..82050bd51 100644 --- a/src/video_core/textures/convert.cpp +++ b/src/video_core/textures/convert.cpp @@ -10,6 +10,7 @@ #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" diff --git a/src/video_core/textures/convert.h b/src/video_core/textures/convert.h index 07cd8b5da..12542e71c 100644 --- a/src/video_core/textures/convert.h +++ b/src/video_core/textures/convert.h @@ -5,7 +5,10 @@ #pragma once #include "common/common_types.h" -#include "video_core/surface.h" + +namespace VideoCore::Surface { +enum class PixelFormat; +} namespace Tegra::Texture { diff --git a/src/video_core/textures/texture.h b/src/video_core/textures/texture.h index 93ecc6e31..bea0d5bc2 100644 --- a/src/video_core/textures/texture.h +++ b/src/video_core/textures/texture.h @@ -7,9 +7,7 @@ #include <array> #include "common/assert.h" #include "common/bit_field.h" -#include "common/common_funcs.h" #include "common/common_types.h" -#include "video_core/memory_manager.h" namespace Tegra::Texture { diff --git a/src/yuzu/CMakeLists.txt b/src/yuzu/CMakeLists.txt index 4cab599b4..31b65c04c 100644 --- a/src/yuzu/CMakeLists.txt +++ b/src/yuzu/CMakeLists.txt @@ -31,6 +31,8 @@ add_executable(yuzu configuration/configure_general.h configuration/configure_graphics.cpp configuration/configure_graphics.h + configuration/configure_hotkeys.cpp + configuration/configure_hotkeys.h configuration/configure_input.cpp configuration/configure_input.h configuration/configure_input_player.cpp @@ -54,8 +56,6 @@ add_executable(yuzu debugger/graphics/graphics_breakpoints.cpp debugger/graphics/graphics_breakpoints.h debugger/graphics/graphics_breakpoints_p.h - debugger/graphics/graphics_surface.cpp - debugger/graphics/graphics_surface.h debugger/console.cpp debugger/console.h debugger/profiler.cpp @@ -78,6 +78,8 @@ add_executable(yuzu ui_settings.h util/limitable_input_dialog.cpp util/limitable_input_dialog.h + util/sequence_dialog/sequence_dialog.cpp + util/sequence_dialog/sequence_dialog.h util/spinbox.cpp util/spinbox.h util/util.cpp @@ -95,6 +97,7 @@ set(UIS configuration/configure_gamelist.ui configuration/configure_general.ui configuration/configure_graphics.ui + configuration/configure_hotkeys.ui configuration/configure_input.ui configuration/configure_input_player.ui configuration/configure_input_simple.ui @@ -105,7 +108,6 @@ set(UIS configuration/configure_touchscreen_advanced.ui configuration/configure_web.ui compatdb.ui - hotkeys.ui loading_screen.ui main.ui ) @@ -149,6 +151,12 @@ target_link_libraries(yuzu PRIVATE common core input_common video_core) target_link_libraries(yuzu PRIVATE Boost::boost glad Qt5::OpenGL Qt5::Widgets) target_link_libraries(yuzu PRIVATE ${PLATFORM_LIBRARIES} Threads::Threads) +target_compile_definitions(yuzu PRIVATE + # Use QStringBuilder for string concatenation to reduce + # the overall number of temporary strings created. + -DQT_USE_QSTRINGBUILDER +) + if (YUZU_ENABLE_COMPATIBILITY_REPORTING) target_compile_definitions(yuzu PRIVATE -DYUZU_ENABLE_COMPATIBILITY_REPORTING) endif() diff --git a/src/yuzu/applets/profile_select.cpp b/src/yuzu/applets/profile_select.cpp index f95f7fe3c..743b24d76 100644 --- a/src/yuzu/applets/profile_select.cpp +++ b/src/yuzu/applets/profile_select.cpp @@ -4,6 +4,7 @@ #include <mutex> #include <QDialogButtonBox> +#include <QHeaderView> #include <QLabel> #include <QLineEdit> #include <QScrollArea> diff --git a/src/yuzu/applets/profile_select.h b/src/yuzu/applets/profile_select.h index 868573324..1c2922e54 100644 --- a/src/yuzu/applets/profile_select.h +++ b/src/yuzu/applets/profile_select.h @@ -7,6 +7,7 @@ #include <vector> #include <QDialog> #include <QList> +#include <QTreeView> #include "core/frontend/applets/profile_select.h" class GMainWindow; @@ -16,7 +17,6 @@ class QLabel; class QScrollArea; class QStandardItem; class QStandardItemModel; -class QTreeView; class QVBoxLayout; class QtProfileSelectionDialog final : public QDialog { diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp index 7438fbc0a..c29f2d2dc 100644 --- a/src/yuzu/bootmanager.cpp +++ b/src/yuzu/bootmanager.cpp @@ -1,6 +1,13 @@ +// Copyright 2014 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + #include <QApplication> #include <QHBoxLayout> #include <QKeyEvent> +#include <QOffscreenSurface> +#include <QOpenGLWindow> +#include <QPainter> #include <QScreen> #include <QWindow> #include <fmt/format.h> @@ -82,13 +89,36 @@ void EmuThread::run() { render_window->moveContext(); } +class GGLContext : public Core::Frontend::GraphicsContext { +public: + explicit GGLContext(QOpenGLContext* shared_context) : surface() { + context = std::make_unique<QOpenGLContext>(shared_context); + surface.setFormat(shared_context->format()); + surface.create(); + } + + void MakeCurrent() override { + context->makeCurrent(&surface); + } + + void DoneCurrent() override { + context->doneCurrent(); + } + + void SwapBuffers() override {} + +private: + std::unique_ptr<QOpenGLContext> context; + QOffscreenSurface surface; +}; + // This class overrides paintEvent and resizeEvent to prevent the GUI thread from stealing GL // context. // The corresponding functionality is handled in EmuThread instead -class GGLWidgetInternal : public QGLWidget { +class GGLWidgetInternal : public QOpenGLWindow { public: - GGLWidgetInternal(QGLFormat fmt, GRenderWindow* parent) - : QGLWidget(fmt, parent), parent(parent) {} + GGLWidgetInternal(GRenderWindow* parent, QOpenGLContext* shared_context) + : QOpenGLWindow(shared_context), parent(parent) {} void paintEvent(QPaintEvent* ev) override { if (do_painting) { @@ -101,9 +131,51 @@ public: parent->OnFramebufferSizeChanged(); } + void keyPressEvent(QKeyEvent* event) override { + InputCommon::GetKeyboard()->PressKey(event->key()); + } + + void keyReleaseEvent(QKeyEvent* event) override { + InputCommon::GetKeyboard()->ReleaseKey(event->key()); + } + + void mousePressEvent(QMouseEvent* event) override { + if (event->source() == Qt::MouseEventSynthesizedBySystem) + return; // touch input is handled in TouchBeginEvent + + const auto pos{event->pos()}; + if (event->button() == Qt::LeftButton) { + const auto [x, y] = parent->ScaleTouch(pos); + parent->TouchPressed(x, y); + } else if (event->button() == Qt::RightButton) { + InputCommon::GetMotionEmu()->BeginTilt(pos.x(), pos.y()); + } + } + + void mouseMoveEvent(QMouseEvent* event) override { + if (event->source() == Qt::MouseEventSynthesizedBySystem) + return; // touch input is handled in TouchUpdateEvent + + const auto pos{event->pos()}; + const auto [x, y] = parent->ScaleTouch(pos); + parent->TouchMoved(x, y); + InputCommon::GetMotionEmu()->Tilt(pos.x(), pos.y()); + } + + void mouseReleaseEvent(QMouseEvent* event) override { + if (event->source() == Qt::MouseEventSynthesizedBySystem) + return; // touch input is handled in TouchEndEvent + + if (event->button() == Qt::LeftButton) + parent->TouchReleased(); + else if (event->button() == Qt::RightButton) + InputCommon::GetMotionEmu()->EndTilt(); + } + void DisablePainting() { do_painting = false; } + void EnablePainting() { do_painting = true; } @@ -114,7 +186,7 @@ private: }; GRenderWindow::GRenderWindow(QWidget* parent, EmuThread* emu_thread) - : QWidget(parent), child(nullptr), emu_thread(emu_thread) { + : QWidget(parent), child(nullptr), context(nullptr), emu_thread(emu_thread) { setWindowTitle(QStringLiteral("yuzu %1 | %2-%3") .arg(Common::g_build_name, Common::g_scm_branch, Common::g_scm_desc)); @@ -137,19 +209,19 @@ void GRenderWindow::moveContext() { auto thread = (QThread::currentThread() == qApp->thread() && emu_thread != nullptr) ? emu_thread : qApp->thread(); - child->context()->moveToThread(thread); + context->moveToThread(thread); } void GRenderWindow::SwapBuffers() { - // In our multi-threaded QGLWidget use case we shouldn't need to call `makeCurrent`, + // In our multi-threaded QWidget use case we shouldn't need to call `makeCurrent`, // since we never call `doneCurrent` in this thread. // However: // - The Qt debug runtime prints a bogus warning on the console if `makeCurrent` wasn't called // since the last time `swapBuffers` was executed; // - On macOS, if `makeCurrent` isn't called explicitely, resizing the buffer breaks. - child->makeCurrent(); + context->makeCurrent(child); - child->swapBuffers(); + context->swapBuffers(child); if (!first_frame) { emit FirstFrameDisplayed(); first_frame = true; @@ -157,11 +229,11 @@ void GRenderWindow::SwapBuffers() { } void GRenderWindow::MakeCurrent() { - child->makeCurrent(); + context->makeCurrent(child); } void GRenderWindow::DoneCurrent() { - child->doneCurrent(); + context->doneCurrent(); } void GRenderWindow::PollEvents() {} @@ -174,14 +246,26 @@ void GRenderWindow::PollEvents() {} void GRenderWindow::OnFramebufferSizeChanged() { // Screen changes potentially incur a change in screen DPI, hence we should update the // framebuffer size - qreal pixelRatio = windowPixelRatio(); + qreal pixelRatio = GetWindowPixelRatio(); unsigned width = child->QPaintDevice::width() * pixelRatio; unsigned height = child->QPaintDevice::height() * pixelRatio; UpdateCurrentFramebufferLayout(width, height); } +void GRenderWindow::ForwardKeyPressEvent(QKeyEvent* event) { + if (child) { + child->keyPressEvent(event); + } +} + +void GRenderWindow::ForwardKeyReleaseEvent(QKeyEvent* event) { + if (child) { + child->keyReleaseEvent(event); + } +} + void GRenderWindow::BackupGeometry() { - geometry = ((QGLWidget*)this)->saveGeometry(); + geometry = ((QWidget*)this)->saveGeometry(); } void GRenderWindow::RestoreGeometry() { @@ -199,18 +283,18 @@ QByteArray GRenderWindow::saveGeometry() { // If we are a top-level widget, store the current geometry // otherwise, store the last backup if (parent() == nullptr) - return ((QGLWidget*)this)->saveGeometry(); + return ((QWidget*)this)->saveGeometry(); else return geometry; } -qreal GRenderWindow::windowPixelRatio() const { +qreal GRenderWindow::GetWindowPixelRatio() const { // windowHandle() might not be accessible until the window is displayed to screen. return windowHandle() ? windowHandle()->screen()->devicePixelRatio() : 1.0f; } std::pair<unsigned, unsigned> GRenderWindow::ScaleTouch(const QPointF pos) const { - const qreal pixel_ratio = windowPixelRatio(); + const qreal pixel_ratio = GetWindowPixelRatio(); return {static_cast<unsigned>(std::max(std::round(pos.x() * pixel_ratio), qreal{0.0})), static_cast<unsigned>(std::max(std::round(pos.y() * pixel_ratio), qreal{0.0}))}; } @@ -220,47 +304,6 @@ void GRenderWindow::closeEvent(QCloseEvent* event) { QWidget::closeEvent(event); } -void GRenderWindow::keyPressEvent(QKeyEvent* event) { - InputCommon::GetKeyboard()->PressKey(event->key()); -} - -void GRenderWindow::keyReleaseEvent(QKeyEvent* event) { - InputCommon::GetKeyboard()->ReleaseKey(event->key()); -} - -void GRenderWindow::mousePressEvent(QMouseEvent* event) { - if (event->source() == Qt::MouseEventSynthesizedBySystem) - return; // touch input is handled in TouchBeginEvent - - auto pos = event->pos(); - if (event->button() == Qt::LeftButton) { - const auto [x, y] = ScaleTouch(pos); - this->TouchPressed(x, y); - } else if (event->button() == Qt::RightButton) { - InputCommon::GetMotionEmu()->BeginTilt(pos.x(), pos.y()); - } -} - -void GRenderWindow::mouseMoveEvent(QMouseEvent* event) { - if (event->source() == Qt::MouseEventSynthesizedBySystem) - return; // touch input is handled in TouchUpdateEvent - - auto pos = event->pos(); - const auto [x, y] = ScaleTouch(pos); - this->TouchMoved(x, y); - InputCommon::GetMotionEmu()->Tilt(pos.x(), pos.y()); -} - -void GRenderWindow::mouseReleaseEvent(QMouseEvent* event) { - if (event->source() == Qt::MouseEventSynthesizedBySystem) - return; // touch input is handled in TouchEndEvent - - if (event->button() == Qt::LeftButton) - this->TouchReleased(); - else if (event->button() == Qt::RightButton) - InputCommon::GetMotionEmu()->EndTilt(); -} - void GRenderWindow::TouchBeginEvent(const QTouchEvent* event) { // TouchBegin always has exactly one touch point, so take the .first() const auto [x, y] = ScaleTouch(event->touchPoints().first().pos()); @@ -313,35 +356,60 @@ void GRenderWindow::OnClientAreaResized(unsigned width, unsigned height) { NotifyClientAreaSizeChanged(std::make_pair(width, height)); } +std::unique_ptr<Core::Frontend::GraphicsContext> GRenderWindow::CreateSharedContext() const { + return std::make_unique<GGLContext>(shared_context.get()); +} + void GRenderWindow::InitRenderTarget() { - if (child) { - delete child; - } + shared_context.reset(); + context.reset(); - if (layout()) { - delete layout(); - } + delete child; + child = nullptr; + + delete container; + container = nullptr; + + delete layout(); first_frame = false; // TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground, // WA_DontShowOnScreen, WA_DeleteOnClose - QGLFormat fmt; + QSurfaceFormat fmt; fmt.setVersion(4, 3); - fmt.setProfile(QGLFormat::CoreProfile); + fmt.setProfile(QSurfaceFormat::CoreProfile); + // TODO: expose a setting for buffer value (ie default/single/double/triple) + fmt.setSwapBehavior(QSurfaceFormat::DefaultSwapBehavior); + shared_context = std::make_unique<QOpenGLContext>(); + shared_context->setFormat(fmt); + shared_context->create(); + context = std::make_unique<QOpenGLContext>(); + context->setShareContext(shared_context.get()); + context->setFormat(fmt); + context->create(); fmt.setSwapInterval(false); - // Requests a forward-compatible context, which is required to get a 3.2+ context on OS X - fmt.setOption(QGL::NoDeprecatedFunctions); + child = new GGLWidgetInternal(this, shared_context.get()); + container = QWidget::createWindowContainer(child, this); - child = new GGLWidgetInternal(fmt, this); QBoxLayout* layout = new QHBoxLayout(this); - - resize(Layout::ScreenUndocked::Width, Layout::ScreenUndocked::Height); - layout->addWidget(child); + layout->addWidget(container); layout->setMargin(0); setLayout(layout); + // Reset minimum size to avoid unwanted resizes when this function is called for a second time. + setMinimumSize(1, 1); + + // Show causes the window to actually be created and the OpenGL context as well, but we don't + // want the widget to be shown yet, so immediately hide it. + show(); + hide(); + + resize(Layout::ScreenUndocked::Width, Layout::ScreenUndocked::Height); + child->resize(Layout::ScreenUndocked::Width, Layout::ScreenUndocked::Height); + container->resize(Layout::ScreenUndocked::Width, Layout::ScreenUndocked::Height); + OnMinimalClientAreaChangeRequest(GetActiveConfig().min_client_area_size); OnFramebufferSizeChanged(); diff --git a/src/yuzu/bootmanager.h b/src/yuzu/bootmanager.h index 3183621bc..9608b959f 100644 --- a/src/yuzu/bootmanager.h +++ b/src/yuzu/bootmanager.h @@ -7,9 +7,9 @@ #include <atomic> #include <condition_variable> #include <mutex> -#include <QGLWidget> #include <QImage> #include <QThread> +#include <QWidget> #include "common/thread.h" #include "core/core.h" #include "core/frontend/emu_window.h" @@ -21,6 +21,8 @@ class QTouchEvent; class GGLWidgetInternal; class GMainWindow; class GRenderWindow; +class QSurface; +class QOpenGLContext; namespace VideoCore { enum class LoadCallbackStage; @@ -121,25 +123,21 @@ public: void MakeCurrent() override; void DoneCurrent() override; void PollEvents() override; + std::unique_ptr<Core::Frontend::GraphicsContext> CreateSharedContext() const override; + + void ForwardKeyPressEvent(QKeyEvent* event); + void ForwardKeyReleaseEvent(QKeyEvent* event); void BackupGeometry(); void RestoreGeometry(); void restoreGeometry(const QByteArray& geometry); // overridden QByteArray saveGeometry(); // overridden - qreal windowPixelRatio() const; + qreal GetWindowPixelRatio() const; + std::pair<unsigned, unsigned> ScaleTouch(const QPointF pos) const; void closeEvent(QCloseEvent* event) override; - - void keyPressEvent(QKeyEvent* event) override; - void keyReleaseEvent(QKeyEvent* event) override; - - void mousePressEvent(QMouseEvent* event) override; - void mouseMoveEvent(QMouseEvent* event) override; - void mouseReleaseEvent(QMouseEvent* event) override; - bool event(QEvent* event) override; - void focusOutEvent(QFocusEvent* event) override; void OnClientAreaResized(unsigned width, unsigned height); @@ -161,7 +159,6 @@ signals: void FirstFrameDisplayed(); private: - std::pair<unsigned, unsigned> ScaleTouch(const QPointF pos) const; void TouchBeginEvent(const QTouchEvent* event); void TouchUpdateEvent(const QTouchEvent* event); void TouchEndEvent(); @@ -169,11 +166,17 @@ private: void OnMinimalClientAreaChangeRequest( const std::pair<unsigned, unsigned>& minimal_size) override; - GGLWidgetInternal* child; + QWidget* container = nullptr; + GGLWidgetInternal* child = nullptr; QByteArray geometry; EmuThread* emu_thread; + // Context that backs the GGLWidgetInternal (and will be used by core to render) + std::unique_ptr<QOpenGLContext> context; + // Context that will be shared between all newly created contexts. This should never be made + // current + std::unique_ptr<QOpenGLContext> shared_context; /// Temporary storage of the screenshot taken QImage screenshot_image; diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp index dead9f807..8725a78dc 100644 --- a/src/yuzu/configuration/config.cpp +++ b/src/yuzu/configuration/config.cpp @@ -2,6 +2,8 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. +#include <array> +#include <QKeySequence> #include <QSettings> #include "common/file_util.h" #include "configure_input_simple.h" @@ -9,7 +11,6 @@ #include "core/hle/service/hid/controllers/npad.h" #include "input_common/main.h" #include "yuzu/configuration/config.h" -#include "yuzu/ui_settings.h" Config::Config() { // TODO: Don't hardcode the path; let the frontend decide where to put the config files. @@ -17,7 +18,6 @@ Config::Config() { FileUtil::CreateFullPath(qt_config_loc); qt_config = std::make_unique<QSettings>(QString::fromStdString(qt_config_loc), QSettings::IniFormat); - Reload(); } @@ -205,6 +205,27 @@ const std::array<int, Settings::NativeKeyboard::NumKeyboardMods> Config::default Qt::Key_Control, Qt::Key_Shift, Qt::Key_AltGr, Qt::Key_ApplicationRight, }; +// This shouldn't have anything except static initializers (no functions). So +// QKeySequnce(...).toString() is NOT ALLOWED HERE. +// This must be in alphabetical order according to action name as it must have the same order as +// UISetting::values.shortcuts, which is alphabetically ordered. +const std::array<UISettings::Shortcut, 15> Config::default_hotkeys{ + {{"Capture Screenshot", "Main Window", {"Ctrl+P", Qt::ApplicationShortcut}}, + {"Continue/Pause Emulation", "Main Window", {"F4", Qt::WindowShortcut}}, + {"Decrease Speed Limit", "Main Window", {"-", Qt::ApplicationShortcut}}, + {"Exit yuzu", "Main Window", {"Ctrl+Q", Qt::WindowShortcut}}, + {"Exit Fullscreen", "Main Window", {"Esc", Qt::WindowShortcut}}, + {"Fullscreen", "Main Window", {"F11", Qt::WindowShortcut}}, + {"Increase Speed Limit", "Main Window", {"+", Qt::ApplicationShortcut}}, + {"Load Amiibo", "Main Window", {"F2", Qt::ApplicationShortcut}}, + {"Load File", "Main Window", {"Ctrl+O", Qt::WindowShortcut}}, + {"Restart Emulation", "Main Window", {"F6", Qt::WindowShortcut}}, + {"Stop Emulation", "Main Window", {"F5", Qt::WindowShortcut}}, + {"Toggle Filter Bar", "Main Window", {"Ctrl+F", Qt::WindowShortcut}}, + {"Toggle Speed Limit", "Main Window", {"Ctrl+Z", Qt::ApplicationShortcut}}, + {"Toggle Status Bar", "Main Window", {"Ctrl+S", Qt::WindowShortcut}}, + {"Change Docked Mode", "Main Window", {"F10", Qt::ApplicationShortcut}}}}; + void Config::ReadPlayerValues() { for (std::size_t p = 0; p < Settings::values.players.size(); ++p) { auto& player = Settings::values.players[p]; @@ -373,6 +394,7 @@ void Config::ReadValues() { ReadSetting("use_accurate_gpu_emulation", false).toBool(); Settings::values.use_asynchronous_gpu_emulation = ReadSetting("use_asynchronous_gpu_emulation", false).toBool(); + Settings::values.force_30fps_mode = ReadSetting("force_30fps_mode", false).toBool(); Settings::values.bg_red = ReadSetting("bg_red", 0.0).toFloat(); Settings::values.bg_green = ReadSetting("bg_green", 0.0).toFloat(); @@ -502,26 +524,21 @@ void Config::ReadValues() { qt_config->beginGroup("Paths"); UISettings::values.roms_path = ReadSetting("romsPath").toString(); UISettings::values.symbols_path = ReadSetting("symbolsPath").toString(); - UISettings::values.gamedir = ReadSetting("gameListRootDir", ".").toString(); - UISettings::values.gamedir_deepscan = ReadSetting("gameListDeepScan", false).toBool(); + UISettings::values.game_directory_path = ReadSetting("gameListRootDir", ".").toString(); + UISettings::values.game_directory_deepscan = ReadSetting("gameListDeepScan", false).toBool(); UISettings::values.recent_files = ReadSetting("recentFiles").toStringList(); qt_config->endGroup(); qt_config->beginGroup("Shortcuts"); - QStringList groups = qt_config->childGroups(); - for (auto group : groups) { + for (auto [name, group, shortcut] : default_hotkeys) { + auto [keyseq, context] = shortcut; qt_config->beginGroup(group); - - QStringList hotkeys = qt_config->childGroups(); - for (auto hotkey : hotkeys) { - qt_config->beginGroup(hotkey); - UISettings::values.shortcuts.emplace_back(UISettings::Shortcut( - group + "/" + hotkey, - UISettings::ContextualShortcut(ReadSetting("KeySeq").toString(), - ReadSetting("Context").toInt()))); - qt_config->endGroup(); - } - + qt_config->beginGroup(name); + UISettings::values.shortcuts.push_back( + {name, + group, + {ReadSetting("KeySeq", keyseq).toString(), ReadSetting("Context", context).toInt()}}); + qt_config->endGroup(); qt_config->endGroup(); } qt_config->endGroup(); @@ -648,6 +665,7 @@ void Config::SaveValues() { WriteSetting("use_accurate_gpu_emulation", Settings::values.use_accurate_gpu_emulation, false); WriteSetting("use_asynchronous_gpu_emulation", Settings::values.use_asynchronous_gpu_emulation, false); + WriteSetting("force_30fps_mode", Settings::values.force_30fps_mode, false); // Cast to double because Qt's written float values are not human-readable WriteSetting("bg_red", (double)Settings::values.bg_red, 0.0); @@ -752,15 +770,22 @@ void Config::SaveValues() { WriteSetting("romsPath", UISettings::values.roms_path); WriteSetting("symbolsPath", UISettings::values.symbols_path); WriteSetting("screenshotPath", UISettings::values.screenshot_path); - WriteSetting("gameListRootDir", UISettings::values.gamedir, "."); - WriteSetting("gameListDeepScan", UISettings::values.gamedir_deepscan, false); + WriteSetting("gameListRootDir", UISettings::values.game_directory_path, "."); + WriteSetting("gameListDeepScan", UISettings::values.game_directory_deepscan, false); WriteSetting("recentFiles", UISettings::values.recent_files); qt_config->endGroup(); qt_config->beginGroup("Shortcuts"); - for (auto shortcut : UISettings::values.shortcuts) { - WriteSetting(shortcut.first + "/KeySeq", shortcut.second.first); - WriteSetting(shortcut.first + "/Context", shortcut.second.second); + // Lengths of UISettings::values.shortcuts & default_hotkeys are same. + // However, their ordering must also be the same. + for (std::size_t i = 0; i < default_hotkeys.size(); i++) { + auto [name, group, shortcut] = UISettings::values.shortcuts[i]; + qt_config->beginGroup(group); + qt_config->beginGroup(name); + WriteSetting("KeySeq", shortcut.first, default_hotkeys[i].shortcut.first); + WriteSetting("Context", shortcut.second, default_hotkeys[i].shortcut.second); + qt_config->endGroup(); + qt_config->endGroup(); } qt_config->endGroup(); diff --git a/src/yuzu/configuration/config.h b/src/yuzu/configuration/config.h index f4185db18..221d2364c 100644 --- a/src/yuzu/configuration/config.h +++ b/src/yuzu/configuration/config.h @@ -9,6 +9,7 @@ #include <string> #include <QVariant> #include "core/settings.h" +#include "yuzu/ui_settings.h" class QSettings; @@ -47,6 +48,8 @@ private: void WriteSetting(const QString& name, const QVariant& value); void WriteSetting(const QString& name, const QVariant& value, const QVariant& default_value); + static const std::array<UISettings::Shortcut, 15> default_hotkeys; + std::unique_ptr<QSettings> qt_config; std::string qt_config_loc; }; diff --git a/src/yuzu/configuration/configure.ui b/src/yuzu/configuration/configure.ui index 3f03f0b77..267717bc9 100644 --- a/src/yuzu/configuration/configure.ui +++ b/src/yuzu/configuration/configure.ui @@ -7,9 +7,15 @@ <x>0</x> <y>0</y> <width>382</width> - <height>241</height> + <height>650</height> </rect> </property> + <property name="minimumSize"> + <size> + <width>0</width> + <height>650</height> + </size> + </property> <property name="windowTitle"> <string>yuzu Configuration</string> </property> @@ -62,6 +68,11 @@ <string>Input</string> </attribute> </widget> + <widget class="ConfigureHotkeys" name="hotkeysTab"> + <attribute name="title"> + <string>Hotkeys</string> + </attribute> + </widget> <widget class="ConfigureGraphics" name="graphicsTab"> <attribute name="title"> <string>Graphics</string> @@ -150,6 +161,12 @@ <header>configuration/configure_input_simple.h</header> <container>1</container> </customwidget> + <customwidget> + <class>ConfigureHotkeys</class> + <extends>QWidget</extends> + <header>configuration/configure_hotkeys.h</header> + <container>1</container> + </customwidget> </customwidgets> <resources/> <connections> diff --git a/src/yuzu/configuration/configure_dialog.cpp b/src/yuzu/configuration/configure_dialog.cpp index 777050405..a5218b051 100644 --- a/src/yuzu/configuration/configure_dialog.cpp +++ b/src/yuzu/configuration/configure_dialog.cpp @@ -8,20 +8,22 @@ #include "ui_configure.h" #include "yuzu/configuration/config.h" #include "yuzu/configuration/configure_dialog.h" +#include "yuzu/configuration/configure_input_player.h" #include "yuzu/hotkeys.h" -ConfigureDialog::ConfigureDialog(QWidget* parent, const HotkeyRegistry& registry) - : QDialog(parent), ui(new Ui::ConfigureDialog) { +ConfigureDialog::ConfigureDialog(QWidget* parent, HotkeyRegistry& registry) + : QDialog(parent), ui(new Ui::ConfigureDialog), registry(registry) { ui->setupUi(this); - ui->generalTab->PopulateHotkeyList(registry); + ui->hotkeysTab->Populate(registry); this->setConfiguration(); this->PopulateSelectionList(); connect(ui->selectorList, &QListWidget::itemSelectionChanged, this, &ConfigureDialog::UpdateVisibleTabs); - adjustSize(); - ui->selectorList->setCurrentRow(0); + + // Synchronise lists upon initialisation + ui->hotkeysTab->EmitHotkeysChanged(); } ConfigureDialog::~ConfigureDialog() = default; @@ -34,6 +36,7 @@ void ConfigureDialog::applyConfiguration() { ui->systemTab->applyConfiguration(); ui->profileManagerTab->applyConfiguration(); ui->inputTab->applyConfiguration(); + ui->hotkeysTab->applyConfiguration(registry); ui->graphicsTab->applyConfiguration(); ui->audioTab->applyConfiguration(); ui->debugTab->applyConfiguration(); @@ -47,7 +50,7 @@ void ConfigureDialog::PopulateSelectionList() { {{tr("General"), {tr("General"), tr("Web"), tr("Debug"), tr("Game List")}}, {tr("System"), {tr("System"), tr("Profiles"), tr("Audio")}}, {tr("Graphics"), {tr("Graphics")}}, - {tr("Controls"), {tr("Input")}}}}; + {tr("Controls"), {tr("Input"), tr("Hotkeys")}}}}; for (const auto& entry : items) { auto* const item = new QListWidgetItem(entry.first); @@ -66,6 +69,7 @@ void ConfigureDialog::UpdateVisibleTabs() { {tr("System"), ui->systemTab}, {tr("Profiles"), ui->profileManagerTab}, {tr("Input"), ui->inputTab}, + {tr("Hotkeys"), ui->hotkeysTab}, {tr("Graphics"), ui->graphicsTab}, {tr("Audio"), ui->audioTab}, {tr("Debug"), ui->debugTab}, diff --git a/src/yuzu/configuration/configure_dialog.h b/src/yuzu/configuration/configure_dialog.h index 243d9fa09..2363ba584 100644 --- a/src/yuzu/configuration/configure_dialog.h +++ b/src/yuzu/configuration/configure_dialog.h @@ -17,7 +17,7 @@ class ConfigureDialog : public QDialog { Q_OBJECT public: - explicit ConfigureDialog(QWidget* parent, const HotkeyRegistry& registry); + explicit ConfigureDialog(QWidget* parent, HotkeyRegistry& registry); ~ConfigureDialog() override; void applyConfiguration(); @@ -28,4 +28,5 @@ private: void PopulateSelectionList(); std::unique_ptr<Ui::ConfigureDialog> ui; + HotkeyRegistry& registry; }; diff --git a/src/yuzu/configuration/configure_general.cpp b/src/yuzu/configuration/configure_general.cpp index 389fcf667..e48f4f5a3 100644 --- a/src/yuzu/configuration/configure_general.cpp +++ b/src/yuzu/configuration/configure_general.cpp @@ -28,19 +28,15 @@ ConfigureGeneral::ConfigureGeneral(QWidget* parent) ConfigureGeneral::~ConfigureGeneral() = default; void ConfigureGeneral::setConfiguration() { - ui->toggle_deepscan->setChecked(UISettings::values.gamedir_deepscan); + ui->toggle_deepscan->setChecked(UISettings::values.game_directory_deepscan); ui->toggle_check_exit->setChecked(UISettings::values.confirm_before_closing); ui->toggle_user_on_boot->setChecked(UISettings::values.select_user_on_boot); ui->theme_combobox->setCurrentIndex(ui->theme_combobox->findData(UISettings::values.theme)); ui->use_cpu_jit->setChecked(Settings::values.use_cpu_jit); } -void ConfigureGeneral::PopulateHotkeyList(const HotkeyRegistry& registry) { - ui->widget->Populate(registry); -} - void ConfigureGeneral::applyConfiguration() { - UISettings::values.gamedir_deepscan = ui->toggle_deepscan->isChecked(); + UISettings::values.game_directory_deepscan = ui->toggle_deepscan->isChecked(); UISettings::values.confirm_before_closing = ui->toggle_check_exit->isChecked(); UISettings::values.select_user_on_boot = ui->toggle_user_on_boot->isChecked(); UISettings::values.theme = diff --git a/src/yuzu/configuration/configure_general.h b/src/yuzu/configuration/configure_general.h index 59738af40..df41d995b 100644 --- a/src/yuzu/configuration/configure_general.h +++ b/src/yuzu/configuration/configure_general.h @@ -20,7 +20,6 @@ public: explicit ConfigureGeneral(QWidget* parent = nullptr); ~ConfigureGeneral() override; - void PopulateHotkeyList(const HotkeyRegistry& registry); void applyConfiguration(); private: diff --git a/src/yuzu/configuration/configure_general.ui b/src/yuzu/configuration/configure_general.ui index 01d1c0b8e..1a5721fe7 100644 --- a/src/yuzu/configuration/configure_general.ui +++ b/src/yuzu/configuration/configure_general.ui @@ -98,22 +98,6 @@ </widget> </item> <item> - <widget class="QGroupBox" name="HotKeysGroupBox"> - <property name="title"> - <string>Hotkeys</string> - </property> - <layout class="QHBoxLayout" name="HotKeysHorizontalLayout"> - <item> - <layout class="QVBoxLayout" name="HotKeysVerticalLayout"> - <item> - <widget class="GHotkeysDialog" name="widget" native="true"/> - </item> - </layout> - </item> - </layout> - </widget> - </item> - <item> <spacer name="verticalSpacer"> <property name="orientation"> <enum>Qt::Vertical</enum> @@ -130,14 +114,6 @@ </item> </layout> </widget> - <customwidgets> - <customwidget> - <class>GHotkeysDialog</class> - <extends>QWidget</extends> - <header>hotkeys.h</header> - <container>1</container> - </customwidget> - </customwidgets> <resources/> <connections/> </ui> diff --git a/src/yuzu/configuration/configure_graphics.cpp b/src/yuzu/configuration/configure_graphics.cpp index dd1d67488..0a9883d37 100644 --- a/src/yuzu/configuration/configure_graphics.cpp +++ b/src/yuzu/configuration/configure_graphics.cpp @@ -77,6 +77,8 @@ void ConfigureGraphics::setConfiguration() { ui->use_accurate_gpu_emulation->setChecked(Settings::values.use_accurate_gpu_emulation); ui->use_asynchronous_gpu_emulation->setEnabled(!Core::System::GetInstance().IsPoweredOn()); ui->use_asynchronous_gpu_emulation->setChecked(Settings::values.use_asynchronous_gpu_emulation); + ui->force_30fps_mode->setEnabled(!Core::System::GetInstance().IsPoweredOn()); + ui->force_30fps_mode->setChecked(Settings::values.force_30fps_mode); UpdateBackgroundColorButton(QColor::fromRgbF(Settings::values.bg_red, Settings::values.bg_green, Settings::values.bg_blue)); } @@ -90,6 +92,7 @@ void ConfigureGraphics::applyConfiguration() { Settings::values.use_accurate_gpu_emulation = ui->use_accurate_gpu_emulation->isChecked(); Settings::values.use_asynchronous_gpu_emulation = ui->use_asynchronous_gpu_emulation->isChecked(); + Settings::values.force_30fps_mode = ui->force_30fps_mode->isChecked(); Settings::values.bg_red = static_cast<float>(bg_color.redF()); Settings::values.bg_green = static_cast<float>(bg_color.greenF()); Settings::values.bg_blue = static_cast<float>(bg_color.blueF()); diff --git a/src/yuzu/configuration/configure_graphics.ui b/src/yuzu/configuration/configure_graphics.ui index c6767e0ca..15ab18ecd 100644 --- a/src/yuzu/configuration/configure_graphics.ui +++ b/src/yuzu/configuration/configure_graphics.ui @@ -71,6 +71,13 @@ </widget> </item> <item> + <widget class="QCheckBox" name="force_30fps_mode"> + <property name="text"> + <string>Force 30 FPS mode</string> + </property> + </widget> + </item> + <item> <layout class="QHBoxLayout" name="horizontalLayout"> <item> <widget class="QLabel" name="label"> diff --git a/src/yuzu/configuration/configure_hotkeys.cpp b/src/yuzu/configuration/configure_hotkeys.cpp new file mode 100644 index 000000000..a7a8752e5 --- /dev/null +++ b/src/yuzu/configuration/configure_hotkeys.cpp @@ -0,0 +1,122 @@ +// Copyright 2017 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QMessageBox> +#include <QStandardItemModel> +#include "core/settings.h" +#include "ui_configure_hotkeys.h" +#include "yuzu/configuration/configure_hotkeys.h" +#include "yuzu/hotkeys.h" +#include "yuzu/util/sequence_dialog/sequence_dialog.h" + +ConfigureHotkeys::ConfigureHotkeys(QWidget* parent) + : QWidget(parent), ui(std::make_unique<Ui::ConfigureHotkeys>()) { + ui->setupUi(this); + setFocusPolicy(Qt::ClickFocus); + + model = new QStandardItemModel(this); + model->setColumnCount(3); + model->setHorizontalHeaderLabels({tr("Action"), tr("Hotkey"), tr("Context")}); + + connect(ui->hotkey_list, &QTreeView::doubleClicked, this, &ConfigureHotkeys::Configure); + ui->hotkey_list->setModel(model); + + // TODO(Kloen): Make context configurable as well (hiding the column for now) + ui->hotkey_list->hideColumn(2); + + ui->hotkey_list->setColumnWidth(0, 200); + ui->hotkey_list->resizeColumnToContents(1); +} + +ConfigureHotkeys::~ConfigureHotkeys() = default; + +void ConfigureHotkeys::EmitHotkeysChanged() { + emit HotkeysChanged(GetUsedKeyList()); +} + +QList<QKeySequence> ConfigureHotkeys::GetUsedKeyList() const { + QList<QKeySequence> list; + for (int r = 0; r < model->rowCount(); r++) { + const QStandardItem* parent = model->item(r, 0); + for (int r2 = 0; r2 < parent->rowCount(); r2++) { + const QStandardItem* keyseq = parent->child(r2, 1); + list << QKeySequence::fromString(keyseq->text(), QKeySequence::NativeText); + } + } + return list; +} + +void ConfigureHotkeys::Populate(const HotkeyRegistry& registry) { + for (const auto& group : registry.hotkey_groups) { + auto* parent_item = new QStandardItem(group.first); + parent_item->setEditable(false); + for (const auto& hotkey : group.second) { + auto* action = new QStandardItem(hotkey.first); + auto* keyseq = + new QStandardItem(hotkey.second.keyseq.toString(QKeySequence::NativeText)); + action->setEditable(false); + keyseq->setEditable(false); + parent_item->appendRow({action, keyseq}); + } + model->appendRow(parent_item); + } + + ui->hotkey_list->expandAll(); +} + +void ConfigureHotkeys::Configure(QModelIndex index) { + if (!index.parent().isValid()) { + return; + } + + index = index.sibling(index.row(), 1); + auto* const model = ui->hotkey_list->model(); + const auto previous_key = model->data(index); + + SequenceDialog hotkey_dialog{this}; + + const int return_code = hotkey_dialog.exec(); + const auto key_sequence = hotkey_dialog.GetSequence(); + if (return_code == QDialog::Rejected || key_sequence.isEmpty()) { + return; + } + + if (IsUsedKey(key_sequence) && key_sequence != QKeySequence(previous_key.toString())) { + QMessageBox::critical(this, tr("Error in inputted key"), + tr("You're using a key that's already bound.")); + } else { + model->setData(index, key_sequence.toString(QKeySequence::NativeText)); + EmitHotkeysChanged(); + } +} + +bool ConfigureHotkeys::IsUsedKey(QKeySequence key_sequence) const { + return GetUsedKeyList().contains(key_sequence); +} + +void ConfigureHotkeys::applyConfiguration(HotkeyRegistry& registry) { + for (int key_id = 0; key_id < model->rowCount(); key_id++) { + const QStandardItem* parent = model->item(key_id, 0); + for (int key_column_id = 0; key_column_id < parent->rowCount(); key_column_id++) { + const QStandardItem* action = parent->child(key_column_id, 0); + const QStandardItem* keyseq = parent->child(key_column_id, 1); + for (auto& [group, sub_actions] : registry.hotkey_groups) { + if (group != parent->text()) + continue; + for (auto& [action_name, hotkey] : sub_actions) { + if (action_name != action->text()) + continue; + hotkey.keyseq = QKeySequence(keyseq->text()); + } + } + } + } + + registry.SaveHotkeys(); + Settings::Apply(); +} + +void ConfigureHotkeys::retranslateUi() { + ui->retranslateUi(this); +} diff --git a/src/yuzu/configuration/configure_hotkeys.h b/src/yuzu/configuration/configure_hotkeys.h new file mode 100644 index 000000000..73fb8a175 --- /dev/null +++ b/src/yuzu/configuration/configure_hotkeys.h @@ -0,0 +1,47 @@ +// Copyright 2017 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <memory> +#include <QWidget> + +namespace Ui { +class ConfigureHotkeys; +} + +class HotkeyRegistry; +class QStandardItemModel; + +class ConfigureHotkeys : public QWidget { + Q_OBJECT + +public: + explicit ConfigureHotkeys(QWidget* parent = nullptr); + ~ConfigureHotkeys() override; + + void applyConfiguration(HotkeyRegistry& registry); + void retranslateUi(); + + void EmitHotkeysChanged(); + + /** + * Populates the hotkey list widget using data from the provided registry. + * Called everytime the Configure dialog is opened. + * @param registry The HotkeyRegistry whose data is used to populate the list. + */ + void Populate(const HotkeyRegistry& registry); + +signals: + void HotkeysChanged(QList<QKeySequence> new_key_list); + +private: + void Configure(QModelIndex index); + bool IsUsedKey(QKeySequence key_sequence) const; + QList<QKeySequence> GetUsedKeyList() const; + + std::unique_ptr<Ui::ConfigureHotkeys> ui; + + QStandardItemModel* model; +}; diff --git a/src/yuzu/configuration/configure_hotkeys.ui b/src/yuzu/configuration/configure_hotkeys.ui new file mode 100644 index 000000000..0d0b70f38 --- /dev/null +++ b/src/yuzu/configuration/configure_hotkeys.ui @@ -0,0 +1,42 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>ConfigureHotkeys</class> + <widget class="QWidget" name="ConfigureHotkeys"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>363</width> + <height>388</height> + </rect> + </property> + <property name="windowTitle"> + <string>Hotkey Settings</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <item> + <layout class="QVBoxLayout" name="verticalLayout_2"> + <item> + <widget class="QLabel" name="label_2"> + <property name="text"> + <string>Double-click on a binding to change it.</string> + </property> + </widget> + </item> + <item> + <widget class="QTreeView" name="hotkey_list"> + <property name="editTriggers"> + <set>QAbstractItemView::NoEditTriggers</set> + </property> + <property name="sortingEnabled"> + <bool>false</bool> + </property> + </widget> + </item> + </layout> + </item> + </layout> + </widget> + <resources/> + <connections/> +</ui>
\ No newline at end of file diff --git a/src/yuzu/debugger/graphics/graphics_surface.cpp b/src/yuzu/debugger/graphics/graphics_surface.cpp deleted file mode 100644 index 11023ed63..000000000 --- a/src/yuzu/debugger/graphics/graphics_surface.cpp +++ /dev/null @@ -1,460 +0,0 @@ -// Copyright 2014 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#include <QBoxLayout> -#include <QComboBox> -#include <QDebug> -#include <QFileDialog> -#include <QLabel> -#include <QMouseEvent> -#include <QPushButton> -#include <QScrollArea> -#include <QSpinBox> -#include "common/vector_math.h" -#include "core/core.h" -#include "core/memory.h" -#include "video_core/engines/maxwell_3d.h" -#include "video_core/gpu.h" -#include "video_core/textures/decoders.h" -#include "video_core/textures/texture.h" -#include "yuzu/debugger/graphics/graphics_surface.h" -#include "yuzu/util/spinbox.h" - -static Tegra::Texture::TextureFormat ConvertToTextureFormat( - Tegra::RenderTargetFormat render_target_format) { - switch (render_target_format) { - case Tegra::RenderTargetFormat::RGBA8_UNORM: - return Tegra::Texture::TextureFormat::A8R8G8B8; - case Tegra::RenderTargetFormat::RGB10_A2_UNORM: - return Tegra::Texture::TextureFormat::A2B10G10R10; - default: - UNIMPLEMENTED_MSG("Unimplemented RT format"); - return Tegra::Texture::TextureFormat::A8R8G8B8; - } -} - -SurfacePicture::SurfacePicture(QWidget* parent, GraphicsSurfaceWidget* surface_widget_) - : QLabel(parent), surface_widget(surface_widget_) {} - -SurfacePicture::~SurfacePicture() = default; - -void SurfacePicture::mousePressEvent(QMouseEvent* event) { - // Only do something while the left mouse button is held down - if (!(event->buttons() & Qt::LeftButton)) - return; - - if (pixmap() == nullptr) - return; - - if (surface_widget) - surface_widget->Pick(event->x() * pixmap()->width() / width(), - event->y() * pixmap()->height() / height()); -} - -void SurfacePicture::mouseMoveEvent(QMouseEvent* event) { - // We also want to handle the event if the user moves the mouse while holding down the LMB - mousePressEvent(event); -} - -GraphicsSurfaceWidget::GraphicsSurfaceWidget(std::shared_ptr<Tegra::DebugContext> debug_context, - QWidget* parent) - : BreakPointObserverDock(debug_context, tr("Maxwell Surface Viewer"), parent), - surface_source(Source::RenderTarget0) { - setObjectName("MaxwellSurface"); - - surface_source_list = new QComboBox; - surface_source_list->addItem(tr("Render Target 0")); - surface_source_list->addItem(tr("Render Target 1")); - surface_source_list->addItem(tr("Render Target 2")); - surface_source_list->addItem(tr("Render Target 3")); - surface_source_list->addItem(tr("Render Target 4")); - surface_source_list->addItem(tr("Render Target 5")); - surface_source_list->addItem(tr("Render Target 6")); - surface_source_list->addItem(tr("Render Target 7")); - surface_source_list->addItem(tr("Z Buffer")); - surface_source_list->addItem(tr("Custom")); - surface_source_list->setCurrentIndex(static_cast<int>(surface_source)); - - surface_address_control = new CSpinBox; - surface_address_control->SetBase(16); - surface_address_control->SetRange(0, 0x7FFFFFFFFFFFFFFF); - surface_address_control->SetPrefix("0x"); - - unsigned max_dimension = 16384; // TODO: Find actual maximum - - surface_width_control = new QSpinBox; - surface_width_control->setRange(0, max_dimension); - - surface_height_control = new QSpinBox; - surface_height_control->setRange(0, max_dimension); - - surface_picker_x_control = new QSpinBox; - surface_picker_x_control->setRange(0, max_dimension - 1); - - surface_picker_y_control = new QSpinBox; - surface_picker_y_control->setRange(0, max_dimension - 1); - - surface_format_control = new QComboBox; - - // Color formats sorted by Maxwell texture format index - surface_format_control->addItem(tr("None")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("A8R8G8B8")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("Unknown")); - surface_format_control->addItem(tr("DXT1")); - surface_format_control->addItem(tr("DXT23")); - surface_format_control->addItem(tr("DXT45")); - surface_format_control->addItem(tr("DXN1")); - surface_format_control->addItem(tr("DXN2")); - - surface_info_label = new QLabel(); - surface_info_label->setWordWrap(true); - - surface_picture_label = new SurfacePicture(0, this); - surface_picture_label->setSizePolicy(QSizePolicy::Fixed, QSizePolicy::Fixed); - surface_picture_label->setAlignment(Qt::AlignLeft | Qt::AlignTop); - surface_picture_label->setScaledContents(false); - - auto scroll_area = new QScrollArea(); - scroll_area->setBackgroundRole(QPalette::Dark); - scroll_area->setWidgetResizable(false); - scroll_area->setWidget(surface_picture_label); - - save_surface = new QPushButton(QIcon::fromTheme("document-save"), tr("Save")); - - // Connections - connect(this, &GraphicsSurfaceWidget::Update, this, &GraphicsSurfaceWidget::OnUpdate); - connect(surface_source_list, - static_cast<void (QComboBox::*)(int)>(&QComboBox::currentIndexChanged), this, - &GraphicsSurfaceWidget::OnSurfaceSourceChanged); - connect(surface_address_control, &CSpinBox::ValueChanged, this, - &GraphicsSurfaceWidget::OnSurfaceAddressChanged); - connect(surface_width_control, static_cast<void (QSpinBox::*)(int)>(&QSpinBox::valueChanged), - this, &GraphicsSurfaceWidget::OnSurfaceWidthChanged); - connect(surface_height_control, static_cast<void (QSpinBox::*)(int)>(&QSpinBox::valueChanged), - this, &GraphicsSurfaceWidget::OnSurfaceHeightChanged); - connect(surface_format_control, - static_cast<void (QComboBox::*)(int)>(&QComboBox::currentIndexChanged), this, - &GraphicsSurfaceWidget::OnSurfaceFormatChanged); - connect(surface_picker_x_control, static_cast<void (QSpinBox::*)(int)>(&QSpinBox::valueChanged), - this, &GraphicsSurfaceWidget::OnSurfacePickerXChanged); - connect(surface_picker_y_control, static_cast<void (QSpinBox::*)(int)>(&QSpinBox::valueChanged), - this, &GraphicsSurfaceWidget::OnSurfacePickerYChanged); - connect(save_surface, &QPushButton::clicked, this, &GraphicsSurfaceWidget::SaveSurface); - - auto main_widget = new QWidget; - auto main_layout = new QVBoxLayout; - { - auto sub_layout = new QHBoxLayout; - sub_layout->addWidget(new QLabel(tr("Source:"))); - sub_layout->addWidget(surface_source_list); - main_layout->addLayout(sub_layout); - } - { - auto sub_layout = new QHBoxLayout; - sub_layout->addWidget(new QLabel(tr("GPU Address:"))); - sub_layout->addWidget(surface_address_control); - main_layout->addLayout(sub_layout); - } - { - auto sub_layout = new QHBoxLayout; - sub_layout->addWidget(new QLabel(tr("Width:"))); - sub_layout->addWidget(surface_width_control); - main_layout->addLayout(sub_layout); - } - { - auto sub_layout = new QHBoxLayout; - sub_layout->addWidget(new QLabel(tr("Height:"))); - sub_layout->addWidget(surface_height_control); - main_layout->addLayout(sub_layout); - } - { - auto sub_layout = new QHBoxLayout; - sub_layout->addWidget(new QLabel(tr("Format:"))); - sub_layout->addWidget(surface_format_control); - main_layout->addLayout(sub_layout); - } - main_layout->addWidget(scroll_area); - - auto info_layout = new QHBoxLayout; - { - auto xy_layout = new QVBoxLayout; - { - { - auto sub_layout = new QHBoxLayout; - sub_layout->addWidget(new QLabel(tr("X:"))); - sub_layout->addWidget(surface_picker_x_control); - xy_layout->addLayout(sub_layout); - } - { - auto sub_layout = new QHBoxLayout; - sub_layout->addWidget(new QLabel(tr("Y:"))); - sub_layout->addWidget(surface_picker_y_control); - xy_layout->addLayout(sub_layout); - } - } - info_layout->addLayout(xy_layout); - surface_info_label->setSizePolicy(QSizePolicy::Expanding, QSizePolicy::Minimum); - info_layout->addWidget(surface_info_label); - } - main_layout->addLayout(info_layout); - - main_layout->addWidget(save_surface); - main_widget->setLayout(main_layout); - setWidget(main_widget); - - // Load current data - TODO: Make sure this works when emulation is not running - if (debug_context && debug_context->at_breakpoint) { - emit Update(); - widget()->setEnabled(debug_context->at_breakpoint); - } else { - widget()->setEnabled(false); - } -} - -void GraphicsSurfaceWidget::OnBreakPointHit(Tegra::DebugContext::Event event, void* data) { - emit Update(); - widget()->setEnabled(true); -} - -void GraphicsSurfaceWidget::OnResumed() { - widget()->setEnabled(false); -} - -void GraphicsSurfaceWidget::OnSurfaceSourceChanged(int new_value) { - surface_source = static_cast<Source>(new_value); - emit Update(); -} - -void GraphicsSurfaceWidget::OnSurfaceAddressChanged(qint64 new_value) { - if (surface_address != new_value) { - surface_address = static_cast<GPUVAddr>(new_value); - - surface_source_list->setCurrentIndex(static_cast<int>(Source::Custom)); - emit Update(); - } -} - -void GraphicsSurfaceWidget::OnSurfaceWidthChanged(int new_value) { - if (surface_width != static_cast<unsigned>(new_value)) { - surface_width = static_cast<unsigned>(new_value); - - surface_source_list->setCurrentIndex(static_cast<int>(Source::Custom)); - emit Update(); - } -} - -void GraphicsSurfaceWidget::OnSurfaceHeightChanged(int new_value) { - if (surface_height != static_cast<unsigned>(new_value)) { - surface_height = static_cast<unsigned>(new_value); - - surface_source_list->setCurrentIndex(static_cast<int>(Source::Custom)); - emit Update(); - } -} - -void GraphicsSurfaceWidget::OnSurfaceFormatChanged(int new_value) { - if (surface_format != static_cast<Tegra::Texture::TextureFormat>(new_value)) { - surface_format = static_cast<Tegra::Texture::TextureFormat>(new_value); - - surface_source_list->setCurrentIndex(static_cast<int>(Source::Custom)); - emit Update(); - } -} - -void GraphicsSurfaceWidget::OnSurfacePickerXChanged(int new_value) { - if (surface_picker_x != new_value) { - surface_picker_x = new_value; - Pick(surface_picker_x, surface_picker_y); - } -} - -void GraphicsSurfaceWidget::OnSurfacePickerYChanged(int new_value) { - if (surface_picker_y != new_value) { - surface_picker_y = new_value; - Pick(surface_picker_x, surface_picker_y); - } -} - -void GraphicsSurfaceWidget::Pick(int x, int y) { - surface_picker_x_control->setValue(x); - surface_picker_y_control->setValue(y); - - if (x < 0 || x >= static_cast<int>(surface_width) || y < 0 || - y >= static_cast<int>(surface_height)) { - surface_info_label->setText(tr("Pixel out of bounds")); - surface_info_label->setAlignment(Qt::AlignLeft | Qt::AlignVCenter); - return; - } - - surface_info_label->setText(QString("Raw: <Unimplemented>\n(%1)").arg("<Unimplemented>")); - surface_info_label->setAlignment(Qt::AlignLeft | Qt::AlignVCenter); -} - -void GraphicsSurfaceWidget::OnUpdate() { - auto& gpu = Core::System::GetInstance().GPU(); - - QPixmap pixmap; - - switch (surface_source) { - case Source::RenderTarget0: - case Source::RenderTarget1: - case Source::RenderTarget2: - case Source::RenderTarget3: - case Source::RenderTarget4: - case Source::RenderTarget5: - case Source::RenderTarget6: - case Source::RenderTarget7: { - // TODO: Store a reference to the registers in the debug context instead of accessing them - // directly... - - const auto& registers = gpu.Maxwell3D().regs; - const auto& rt = registers.rt[static_cast<std::size_t>(surface_source) - - static_cast<std::size_t>(Source::RenderTarget0)]; - - surface_address = rt.Address(); - surface_width = rt.width; - surface_height = rt.height; - if (rt.format != Tegra::RenderTargetFormat::NONE) { - surface_format = ConvertToTextureFormat(rt.format); - } - - break; - } - - case Source::Custom: { - // Keep user-specified values - break; - } - - default: - qDebug() << "Unknown surface source " << static_cast<int>(surface_source); - break; - } - - surface_address_control->SetValue(surface_address); - surface_width_control->setValue(surface_width); - surface_height_control->setValue(surface_height); - surface_format_control->setCurrentIndex(static_cast<int>(surface_format)); - - if (surface_address == 0) { - surface_picture_label->hide(); - surface_info_label->setText(tr("(invalid surface address)")); - surface_info_label->setAlignment(Qt::AlignCenter); - surface_picker_x_control->setEnabled(false); - surface_picker_y_control->setEnabled(false); - save_surface->setEnabled(false); - return; - } - - // TODO: Implement a good way to visualize alpha components! - - QImage decoded_image(surface_width, surface_height, QImage::Format_ARGB32); - - // TODO(bunnei): Will not work with BCn formats that swizzle 4x4 tiles. - // Needs to be fixed if we plan to use this feature more, otherwise we may remove it. - auto unswizzled_data = Tegra::Texture::UnswizzleTexture( - gpu.MemoryManager().GetPointer(surface_address), 1, 1, - Tegra::Texture::BytesPerPixel(surface_format), surface_width, surface_height, 1U); - - auto texture_data = Tegra::Texture::DecodeTexture(unswizzled_data, surface_format, - surface_width, surface_height); - - surface_picture_label->show(); - - for (unsigned int y = 0; y < surface_height; ++y) { - for (unsigned int x = 0; x < surface_width; ++x) { - Common::Vec4<u8> color; - color[0] = texture_data[x + y * surface_width + 0]; - color[1] = texture_data[x + y * surface_width + 1]; - color[2] = texture_data[x + y * surface_width + 2]; - color[3] = texture_data[x + y * surface_width + 3]; - decoded_image.setPixel(x, y, qRgba(color.r(), color.g(), color.b(), color.a())); - } - } - - pixmap = QPixmap::fromImage(decoded_image); - surface_picture_label->setPixmap(pixmap); - surface_picture_label->resize(pixmap.size()); - - // Update the info with pixel data - surface_picker_x_control->setEnabled(true); - surface_picker_y_control->setEnabled(true); - Pick(surface_picker_x, surface_picker_y); - - // Enable saving the converted pixmap to file - save_surface->setEnabled(true); -} - -void GraphicsSurfaceWidget::SaveSurface() { - QString png_filter = tr("Portable Network Graphic (*.png)"); - QString bin_filter = tr("Binary data (*.bin)"); - - QString selectedFilter; - QString filename = QFileDialog::getSaveFileName( - this, tr("Save Surface"), - QString("texture-0x%1.png").arg(QString::number(surface_address, 16)), - QString("%1;;%2").arg(png_filter, bin_filter), &selectedFilter); - - if (filename.isEmpty()) { - // If the user canceled the dialog, don't save anything. - return; - } - - if (selectedFilter == png_filter) { - const QPixmap* pixmap = surface_picture_label->pixmap(); - ASSERT_MSG(pixmap != nullptr, "No pixmap set"); - - QFile file(filename); - file.open(QIODevice::WriteOnly); - if (pixmap) - pixmap->save(&file, "PNG"); - } else if (selectedFilter == bin_filter) { - auto& gpu = Core::System::GetInstance().GPU(); - std::optional<VAddr> address = gpu.MemoryManager().GpuToCpuAddress(surface_address); - - const u8* buffer = Memory::GetPointer(*address); - ASSERT_MSG(buffer != nullptr, "Memory not accessible"); - - QFile file(filename); - file.open(QIODevice::WriteOnly); - int size = surface_width * surface_height * Tegra::Texture::BytesPerPixel(surface_format); - QByteArray data(reinterpret_cast<const char*>(buffer), size); - file.write(data); - } else { - UNREACHABLE_MSG("Unhandled filter selected"); - } -} diff --git a/src/yuzu/debugger/graphics/graphics_surface.h b/src/yuzu/debugger/graphics/graphics_surface.h deleted file mode 100644 index 89445b18f..000000000 --- a/src/yuzu/debugger/graphics/graphics_surface.h +++ /dev/null @@ -1,96 +0,0 @@ -// Copyright 2014 Citra Emulator Project -// Licensed under GPLv2 or any later version -// Refer to the license.txt file included. - -#pragma once - -#include <QLabel> -#include <QPushButton> -#include "video_core/memory_manager.h" -#include "video_core/textures/texture.h" -#include "yuzu/debugger/graphics/graphics_breakpoint_observer.h" - -class QComboBox; -class QSpinBox; -class CSpinBox; - -class GraphicsSurfaceWidget; - -class SurfacePicture : public QLabel { - Q_OBJECT - -public: - explicit SurfacePicture(QWidget* parent = nullptr, - GraphicsSurfaceWidget* surface_widget = nullptr); - ~SurfacePicture() override; - -protected slots: - void mouseMoveEvent(QMouseEvent* event) override; - void mousePressEvent(QMouseEvent* event) override; - -private: - GraphicsSurfaceWidget* surface_widget; -}; - -class GraphicsSurfaceWidget : public BreakPointObserverDock { - Q_OBJECT - - using Event = Tegra::DebugContext::Event; - - enum class Source { - RenderTarget0 = 0, - RenderTarget1 = 1, - RenderTarget2 = 2, - RenderTarget3 = 3, - RenderTarget4 = 4, - RenderTarget5 = 5, - RenderTarget6 = 6, - RenderTarget7 = 7, - ZBuffer = 8, - Custom = 9, - }; - -public: - explicit GraphicsSurfaceWidget(std::shared_ptr<Tegra::DebugContext> debug_context, - QWidget* parent = nullptr); - void Pick(int x, int y); - -public slots: - void OnSurfaceSourceChanged(int new_value); - void OnSurfaceAddressChanged(qint64 new_value); - void OnSurfaceWidthChanged(int new_value); - void OnSurfaceHeightChanged(int new_value); - void OnSurfaceFormatChanged(int new_value); - void OnSurfacePickerXChanged(int new_value); - void OnSurfacePickerYChanged(int new_value); - void OnUpdate(); - -signals: - void Update(); - -private: - void OnBreakPointHit(Tegra::DebugContext::Event event, void* data) override; - void OnResumed() override; - - void SaveSurface(); - - QComboBox* surface_source_list; - CSpinBox* surface_address_control; - QSpinBox* surface_width_control; - QSpinBox* surface_height_control; - QComboBox* surface_format_control; - - SurfacePicture* surface_picture_label; - QSpinBox* surface_picker_x_control; - QSpinBox* surface_picker_y_control; - QLabel* surface_info_label; - QPushButton* save_surface; - - Source surface_source; - GPUVAddr surface_address; - unsigned surface_width; - unsigned surface_height; - Tegra::Texture::TextureFormat surface_format; - int surface_picker_x = 0; - int surface_picker_y = 0; -}; diff --git a/src/yuzu/game_list.cpp b/src/yuzu/game_list.cpp index c0e3c5fa9..b0ca766ec 100644 --- a/src/yuzu/game_list.cpp +++ b/src/yuzu/game_list.cpp @@ -18,6 +18,7 @@ #include "common/common_types.h" #include "common/logging/log.h" #include "core/file_sys/patch_manager.h" +#include "core/file_sys/registered_cache.h" #include "yuzu/compatibility_list.h" #include "yuzu/game_list.h" #include "yuzu/game_list_p.h" @@ -193,8 +194,9 @@ void GameList::onFilterCloseClicked() { main_window->filterBarSetChecked(false); } -GameList::GameList(FileSys::VirtualFilesystem vfs, GMainWindow* parent) - : QWidget{parent}, vfs(std::move(vfs)) { +GameList::GameList(FileSys::VirtualFilesystem vfs, FileSys::ManualContentProvider* provider, + GMainWindow* parent) + : QWidget{parent}, vfs(std::move(vfs)), provider(provider) { watcher = new QFileSystemWatcher(this); connect(watcher, &QFileSystemWatcher::directoryChanged, this, &GameList::RefreshGameDirectory); @@ -329,6 +331,8 @@ void GameList::PopupContextMenu(const QPoint& menu_location) { QMenu context_menu; QAction* open_save_location = context_menu.addAction(tr("Open Save Data Location")); QAction* open_lfs_location = context_menu.addAction(tr("Open Mod Data Location")); + QAction* open_transferable_shader_cache = + context_menu.addAction(tr("Open Transferable Shader Cache")); context_menu.addSeparator(); QAction* dump_romfs = context_menu.addAction(tr("Dump RomFS")); QAction* copy_tid = context_menu.addAction(tr("Copy Title ID to Clipboard")); @@ -344,6 +348,8 @@ void GameList::PopupContextMenu(const QPoint& menu_location) { [&]() { emit OpenFolderRequested(program_id, GameListOpenTarget::SaveData); }); connect(open_lfs_location, &QAction::triggered, [&]() { emit OpenFolderRequested(program_id, GameListOpenTarget::ModData); }); + connect(open_transferable_shader_cache, &QAction::triggered, + [&]() { emit OpenTransferableShaderCacheRequested(program_id); }); connect(dump_romfs, &QAction::triggered, [&]() { emit DumpRomFSRequested(program_id, path); }); connect(copy_tid, &QAction::triggered, [&]() { emit CopyTIDRequested(program_id); }); connect(navigate_to_gamedb_entry, &QAction::triggered, @@ -428,7 +434,8 @@ void GameList::PopulateAsync(const QString& dir_path, bool deep_scan) { emit ShouldCancelWorker(); - GameListWorker* worker = new GameListWorker(vfs, dir_path, deep_scan, compatibility_list); + GameListWorker* worker = + new GameListWorker(vfs, provider, dir_path, deep_scan, compatibility_list); connect(worker, &GameListWorker::EntryReady, this, &GameList::AddEntry, Qt::QueuedConnection); connect(worker, &GameListWorker::Finished, this, &GameList::DonePopulating, @@ -460,9 +467,10 @@ void GameList::LoadInterfaceLayout() { const QStringList GameList::supported_file_extensions = {"nso", "nro", "nca", "xci", "nsp"}; void GameList::RefreshGameDirectory() { - if (!UISettings::values.gamedir.isEmpty() && current_worker != nullptr) { + if (!UISettings::values.game_directory_path.isEmpty() && current_worker != nullptr) { LOG_INFO(Frontend, "Change detected in the games directory. Reloading game list."); search_field->clear(); - PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan); + PopulateAsync(UISettings::values.game_directory_path, + UISettings::values.game_directory_deepscan); } } diff --git a/src/yuzu/game_list.h b/src/yuzu/game_list.h index b317eb2fc..56007eef8 100644 --- a/src/yuzu/game_list.h +++ b/src/yuzu/game_list.h @@ -26,8 +26,9 @@ class GameListSearchField; class GMainWindow; namespace FileSys { +class ManualContentProvider; class VfsFilesystem; -} +} // namespace FileSys enum class GameListOpenTarget { SaveData, @@ -47,7 +48,8 @@ public: COLUMN_COUNT, // Number of columns }; - explicit GameList(std::shared_ptr<FileSys::VfsFilesystem> vfs, GMainWindow* parent = nullptr); + explicit GameList(std::shared_ptr<FileSys::VfsFilesystem> vfs, + FileSys::ManualContentProvider* provider, GMainWindow* parent = nullptr); ~GameList() override; void clearFilter(); @@ -66,6 +68,7 @@ signals: void GameChosen(QString game_path); void ShouldCancelWorker(); void OpenFolderRequested(u64 program_id, GameListOpenTarget target); + void OpenTransferableShaderCacheRequested(u64 program_id); void DumpRomFSRequested(u64 program_id, const std::string& game_path); void CopyTIDRequested(u64 program_id); void NavigateToGamedbEntryRequested(u64 program_id, @@ -85,6 +88,7 @@ private: void RefreshGameDirectory(); std::shared_ptr<FileSys::VfsFilesystem> vfs; + FileSys::ManualContentProvider* provider; GameListSearchField* search_field; GMainWindow* main_window = nullptr; QVBoxLayout* layout = nullptr; diff --git a/src/yuzu/game_list_p.h b/src/yuzu/game_list_p.h index 3db0e90da..2cf5c58a0 100644 --- a/src/yuzu/game_list_p.h +++ b/src/yuzu/game_list_p.h @@ -95,7 +95,7 @@ public: if (row2.isEmpty()) return row1; - return row1 + "\n " + row2; + return QString(row1 + "\n " + row2); } return GameListItem::data(role); diff --git a/src/yuzu/game_list_worker.cpp b/src/yuzu/game_list_worker.cpp index b37710f59..8687e7c5a 100644 --- a/src/yuzu/game_list_worker.cpp +++ b/src/yuzu/game_list_worker.cpp @@ -12,12 +12,15 @@ #include "common/common_paths.h" #include "common/file_util.h" +#include "core/core.h" +#include "core/file_sys/card_image.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.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/submission_package.h" #include "core/hle/service/filesystem/filesystem.h" #include "core/loader/loader.h" #include "yuzu/compatibility_list.h" @@ -119,20 +122,25 @@ QList<QStandardItem*> MakeGameListEntry(const std::string& path, const std::stri } } // Anonymous namespace -GameListWorker::GameListWorker(FileSys::VirtualFilesystem vfs, QString dir_path, bool deep_scan, - const CompatibilityList& compatibility_list) - : vfs(std::move(vfs)), dir_path(std::move(dir_path)), deep_scan(deep_scan), +GameListWorker::GameListWorker(FileSys::VirtualFilesystem vfs, + FileSys::ManualContentProvider* provider, QString dir_path, + bool deep_scan, const CompatibilityList& compatibility_list) + : vfs(std::move(vfs)), provider(provider), dir_path(std::move(dir_path)), deep_scan(deep_scan), compatibility_list(compatibility_list) {} GameListWorker::~GameListWorker() = default; -void GameListWorker::AddInstalledTitlesToGameList() { - const auto cache = Service::FileSystem::GetUnionContents(); - const auto installed_games = cache.ListEntriesFilter(FileSys::TitleType::Application, - FileSys::ContentRecordType::Program); +void GameListWorker::AddTitlesToGameList() { + const auto& cache = dynamic_cast<FileSys::ContentProviderUnion&>( + Core::System::GetInstance().GetContentProvider()); + const auto installed_games = cache.ListEntriesFilterOrigin( + std::nullopt, FileSys::TitleType::Application, FileSys::ContentRecordType::Program); - for (const auto& game : installed_games) { - const auto file = cache.GetEntryUnparsed(game); + for (const auto& [slot, game] : installed_games) { + if (slot == FileSys::ContentProviderUnionSlot::FrontendManual) + continue; + + const auto file = cache.GetEntryUnparsed(game.title_id, game.type); std::unique_ptr<Loader::AppLoader> loader = Loader::GetLoader(file); if (!loader) continue; @@ -150,45 +158,13 @@ void GameListWorker::AddInstalledTitlesToGameList() { emit EntryReady(MakeGameListEntry(file->GetFullPath(), name, icon, *loader, program_id, compatibility_list, patch)); } - - const auto control_data = cache.ListEntriesFilter(FileSys::TitleType::Application, - FileSys::ContentRecordType::Control); - - for (const auto& entry : control_data) { - auto nca = cache.GetEntry(entry); - if (nca != nullptr) { - nca_control_map.insert_or_assign(entry.title_id, std::move(nca)); - } - } } -void GameListWorker::FillControlMap(const std::string& dir_path) { - const auto nca_control_callback = [this](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { - if (stop_processing) { - // Breaks the callback loop - return false; - } - - const std::string physical_name = directory + DIR_SEP + virtual_name; - const QFileInfo file_info(QString::fromStdString(physical_name)); - if (!file_info.isDir() && file_info.suffix() == QStringLiteral("nca")) { - auto nca = - std::make_unique<FileSys::NCA>(vfs->OpenFile(physical_name, FileSys::Mode::Read)); - if (nca->GetType() == FileSys::NCAContentType::Control) { - const u64 title_id = nca->GetTitleId(); - nca_control_map.insert_or_assign(title_id, std::move(nca)); - } - } - return true; - }; - - FileUtil::ForeachDirectoryEntry(nullptr, dir_path, nca_control_callback); -} - -void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsigned int recursion) { - const auto callback = [this, recursion](u64* num_entries_out, const std::string& directory, - const std::string& virtual_name) -> bool { +void GameListWorker::ScanFileSystem(ScanTarget target, const std::string& dir_path, + unsigned int recursion) { + const auto callback = [this, target, recursion](u64* num_entries_out, + const std::string& directory, + const std::string& virtual_name) -> bool { if (stop_processing) { // Breaks the callback loop. return false; @@ -198,7 +174,8 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign const bool is_dir = FileUtil::IsDirectory(physical_name); if (!is_dir && (HasSupportedFileExtension(physical_name) || IsExtractedNCAMain(physical_name))) { - auto loader = Loader::GetLoader(vfs->OpenFile(physical_name, FileSys::Mode::Read)); + const auto file = vfs->OpenFile(physical_name, FileSys::Mode::Read); + auto loader = Loader::GetLoader(file); if (!loader) { return true; } @@ -209,31 +186,42 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign return true; } - std::vector<u8> icon; - const auto res1 = loader->ReadIcon(icon); - u64 program_id = 0; const auto res2 = loader->ReadProgramId(program_id); - std::string name = " "; - const auto res3 = loader->ReadTitle(name); + if (target == ScanTarget::FillManualContentProvider) { + if (res2 == Loader::ResultStatus::Success && file_type == Loader::FileType::NCA) { + provider->AddEntry(FileSys::TitleType::Application, + FileSys::GetCRTypeFromNCAType(FileSys::NCA{file}.GetType()), + program_id, file); + } else if (res2 == Loader::ResultStatus::Success && + (file_type == Loader::FileType::XCI || + file_type == Loader::FileType::NSP)) { + const auto nsp = file_type == Loader::FileType::NSP + ? std::make_shared<FileSys::NSP>(file) + : FileSys::XCI{file}.GetSecurePartitionNSP(); + for (const auto& title : nsp->GetNCAs()) { + for (const auto& entry : title.second) { + provider->AddEntry(entry.first.first, entry.first.second, title.first, + entry.second->GetBaseFile()); + } + } + } + } else { + std::vector<u8> icon; + const auto res1 = loader->ReadIcon(icon); - const FileSys::PatchManager patch{program_id}; + std::string name = " "; + const auto res3 = loader->ReadTitle(name); - if (res1 != Loader::ResultStatus::Success && res3 != Loader::ResultStatus::Success && - res2 == Loader::ResultStatus::Success) { - // Use from metadata pool. - if (nca_control_map.find(program_id) != nca_control_map.end()) { - const auto& nca = nca_control_map[program_id]; - GetMetadataFromControlNCA(patch, *nca, icon, name); - } - } + const FileSys::PatchManager patch{program_id}; - emit EntryReady(MakeGameListEntry(physical_name, name, icon, *loader, program_id, - compatibility_list, patch)); + emit EntryReady(MakeGameListEntry(physical_name, name, icon, *loader, program_id, + compatibility_list, patch)); + } } else if (is_dir && recursion > 0) { watch_list.append(QString::fromStdString(physical_name)); - AddFstEntriesToGameList(physical_name, recursion - 1); + ScanFileSystem(target, physical_name, recursion - 1); } return true; @@ -245,10 +233,11 @@ void GameListWorker::AddFstEntriesToGameList(const std::string& dir_path, unsign void GameListWorker::run() { stop_processing = false; watch_list.append(dir_path); - FillControlMap(dir_path.toStdString()); - AddInstalledTitlesToGameList(); - AddFstEntriesToGameList(dir_path.toStdString(), deep_scan ? 256 : 0); - nca_control_map.clear(); + provider->ClearAllEntries(); + ScanFileSystem(ScanTarget::FillManualContentProvider, dir_path.toStdString(), + deep_scan ? 256 : 0); + AddTitlesToGameList(); + ScanFileSystem(ScanTarget::PopulateGameList, dir_path.toStdString(), deep_scan ? 256 : 0); emit Finished(watch_list); } diff --git a/src/yuzu/game_list_worker.h b/src/yuzu/game_list_worker.h index 0e42d0bde..7c3074af9 100644 --- a/src/yuzu/game_list_worker.h +++ b/src/yuzu/game_list_worker.h @@ -33,7 +33,8 @@ class GameListWorker : public QObject, public QRunnable { Q_OBJECT public: - GameListWorker(std::shared_ptr<FileSys::VfsFilesystem> vfs, QString dir_path, bool deep_scan, + GameListWorker(std::shared_ptr<FileSys::VfsFilesystem> vfs, + FileSys::ManualContentProvider* provider, QString dir_path, bool deep_scan, const CompatibilityList& compatibility_list); ~GameListWorker() override; @@ -58,12 +59,17 @@ signals: void Finished(QStringList watch_list); private: - void AddInstalledTitlesToGameList(); - void FillControlMap(const std::string& dir_path); - void AddFstEntriesToGameList(const std::string& dir_path, unsigned int recursion = 0); + void AddTitlesToGameList(); + + enum class ScanTarget { + FillManualContentProvider, + PopulateGameList, + }; + + void ScanFileSystem(ScanTarget target, const std::string& dir_path, unsigned int recursion = 0); std::shared_ptr<FileSys::VfsFilesystem> vfs; - std::map<u64, std::unique_ptr<FileSys::NCA>> nca_control_map; + FileSys::ManualContentProvider* provider; QStringList watch_list; QString dir_path; bool deep_scan; diff --git a/src/yuzu/hotkeys.cpp b/src/yuzu/hotkeys.cpp index dce399774..4582e7f21 100644 --- a/src/yuzu/hotkeys.cpp +++ b/src/yuzu/hotkeys.cpp @@ -2,7 +2,6 @@ // Licensed under GPLv2 or any later version // Refer to the license.txt file included. -#include <map> #include <QKeySequence> #include <QShortcut> #include <QTreeWidgetItem> @@ -13,47 +12,32 @@ HotkeyRegistry::HotkeyRegistry() = default; HotkeyRegistry::~HotkeyRegistry() = default; -void HotkeyRegistry::LoadHotkeys() { - // Make sure NOT to use a reference here because it would become invalid once we call - // beginGroup() - for (auto shortcut : UISettings::values.shortcuts) { - const QStringList cat = shortcut.first.split('/'); - Q_ASSERT(cat.size() >= 2); - - // RegisterHotkey assigns default keybindings, so use old values as default parameters - Hotkey& hk = hotkey_groups[cat[0]][cat[1]]; - if (!shortcut.second.first.isEmpty()) { - hk.keyseq = QKeySequence::fromString(shortcut.second.first); - hk.context = static_cast<Qt::ShortcutContext>(shortcut.second.second); - } - if (hk.shortcut) - hk.shortcut->setKey(hk.keyseq); - } -} - void HotkeyRegistry::SaveHotkeys() { UISettings::values.shortcuts.clear(); for (const auto& group : hotkey_groups) { for (const auto& hotkey : group.second) { - UISettings::values.shortcuts.emplace_back( - UISettings::Shortcut(group.first + '/' + hotkey.first, - UISettings::ContextualShortcut(hotkey.second.keyseq.toString(), - hotkey.second.context))); + UISettings::values.shortcuts.push_back( + {hotkey.first, group.first, + UISettings::ContextualShortcut(hotkey.second.keyseq.toString(), + hotkey.second.context)}); } } } -void HotkeyRegistry::RegisterHotkey(const QString& group, const QString& action, - const QKeySequence& default_keyseq, - Qt::ShortcutContext default_context) { - auto& hotkey_group = hotkey_groups[group]; - if (hotkey_group.find(action) != hotkey_group.end()) { - return; +void HotkeyRegistry::LoadHotkeys() { + // Make sure NOT to use a reference here because it would become invalid once we call + // beginGroup() + for (auto shortcut : UISettings::values.shortcuts) { + Hotkey& hk = hotkey_groups[shortcut.group][shortcut.name]; + if (!shortcut.shortcut.first.isEmpty()) { + hk.keyseq = QKeySequence::fromString(shortcut.shortcut.first, QKeySequence::NativeText); + hk.context = static_cast<Qt::ShortcutContext>(shortcut.shortcut.second); + } + if (hk.shortcut) { + hk.shortcut->disconnect(); + hk.shortcut->setKey(hk.keyseq); + } } - - auto& hotkey_action = hotkey_groups[group][action]; - hotkey_action.keyseq = default_keyseq; - hotkey_action.context = default_context; } QShortcut* HotkeyRegistry::GetHotkey(const QString& group, const QString& action, QWidget* widget) { @@ -65,24 +49,11 @@ QShortcut* HotkeyRegistry::GetHotkey(const QString& group, const QString& action return hk.shortcut; } -GHotkeysDialog::GHotkeysDialog(QWidget* parent) : QWidget(parent) { - ui.setupUi(this); +QKeySequence HotkeyRegistry::GetKeySequence(const QString& group, const QString& action) { + return hotkey_groups[group][action].keyseq; } -void GHotkeysDialog::Populate(const HotkeyRegistry& registry) { - for (const auto& group : registry.hotkey_groups) { - QTreeWidgetItem* toplevel_item = new QTreeWidgetItem(QStringList(group.first)); - for (const auto& hotkey : group.second) { - QStringList columns; - columns << hotkey.first << hotkey.second.keyseq.toString(); - QTreeWidgetItem* item = new QTreeWidgetItem(columns); - toplevel_item->addChild(item); - } - ui.treeWidget->addTopLevelItem(toplevel_item); - } - // TODO: Make context configurable as well (hiding the column for now) - ui.treeWidget->setColumnCount(2); - - ui.treeWidget->resizeColumnToContents(0); - ui.treeWidget->resizeColumnToContents(1); +Qt::ShortcutContext HotkeyRegistry::GetShortcutContext(const QString& group, + const QString& action) { + return hotkey_groups[group][action].context; } diff --git a/src/yuzu/hotkeys.h b/src/yuzu/hotkeys.h index f38e6c002..4f526dc7e 100644 --- a/src/yuzu/hotkeys.h +++ b/src/yuzu/hotkeys.h @@ -5,7 +5,6 @@ #pragma once #include <map> -#include "ui_hotkeys.h" class QDialog; class QKeySequence; @@ -14,7 +13,7 @@ class QShortcut; class HotkeyRegistry final { public: - friend class GHotkeysDialog; + friend class ConfigureHotkeys; explicit HotkeyRegistry(); ~HotkeyRegistry(); @@ -49,22 +48,27 @@ public: QShortcut* GetHotkey(const QString& group, const QString& action, QWidget* widget); /** - * Register a hotkey. + * Returns a QKeySequence object whose signal can be connected to QAction::setShortcut. * - * @param group General group this hotkey belongs to (e.g. "Main Window", "Debugger") - * @param action Name of the action (e.g. "Start Emulation", "Load Image") - * @param default_keyseq Default key sequence to assign if the hotkey wasn't present in the - * settings file before - * @param default_context Default context to assign if the hotkey wasn't present in the settings - * file before - * @warning Both the group and action strings will be displayed in the hotkey settings dialog + * @param group General group this hotkey belongs to (e.g. "Main Window", "Debugger"). + * @param action Name of the action (e.g. "Start Emulation", "Load Image"). + */ + QKeySequence GetKeySequence(const QString& group, const QString& action); + + /** + * Returns a Qt::ShortcutContext object who can be connected to other + * QAction::setShortcutContext. + * + * @param group General group this shortcut context belongs to (e.g. "Main Window", + * "Debugger"). + * @param action Name of the action (e.g. "Start Emulation", "Load Image"). */ - void RegisterHotkey(const QString& group, const QString& action, - const QKeySequence& default_keyseq = {}, - Qt::ShortcutContext default_context = Qt::WindowShortcut); + Qt::ShortcutContext GetShortcutContext(const QString& group, const QString& action); private: struct Hotkey { + Hotkey() : shortcut(nullptr), context(Qt::WindowShortcut) {} + QKeySequence keyseq; QShortcut* shortcut = nullptr; Qt::ShortcutContext context = Qt::WindowShortcut; @@ -75,15 +79,3 @@ private: HotkeyGroupMap hotkey_groups; }; - -class GHotkeysDialog : public QWidget { - Q_OBJECT - -public: - explicit GHotkeysDialog(QWidget* parent = nullptr); - - void Populate(const HotkeyRegistry& registry); - -private: - Ui::hotkeys ui; -}; diff --git a/src/yuzu/hotkeys.ui b/src/yuzu/hotkeys.ui deleted file mode 100644 index 050fe064e..000000000 --- a/src/yuzu/hotkeys.ui +++ /dev/null @@ -1,46 +0,0 @@ -<?xml version="1.0" encoding="UTF-8"?> -<ui version="4.0"> - <class>hotkeys</class> - <widget class="QWidget" name="hotkeys"> - <property name="geometry"> - <rect> - <x>0</x> - <y>0</y> - <width>363</width> - <height>388</height> - </rect> - </property> - <property name="windowTitle"> - <string>Hotkey Settings</string> - </property> - <layout class="QVBoxLayout" name="verticalLayout"> - <item> - <widget class="QTreeWidget" name="treeWidget"> - <property name="selectionBehavior"> - <enum>QAbstractItemView::SelectItems</enum> - </property> - <property name="headerHidden"> - <bool>false</bool> - </property> - <column> - <property name="text"> - <string>Action</string> - </property> - </column> - <column> - <property name="text"> - <string>Hotkey</string> - </property> - </column> - <column> - <property name="text"> - <string>Context</string> - </property> - </column> - </widget> - </item> - </layout> - </widget> - <resources/> - <connections/> -</ui> diff --git a/src/yuzu/loading_screen.cpp b/src/yuzu/loading_screen.cpp index 86f6d0165..4e2d988cd 100644 --- a/src/yuzu/loading_screen.cpp +++ b/src/yuzu/loading_screen.cpp @@ -192,7 +192,12 @@ void LoadingScreen::OnLoadProgress(VideoCore::LoadCallbackStage stage, std::size } // update labels and progress bar - ui->stage->setText(stage_translations[stage].arg(value).arg(total)); + if (stage == VideoCore::LoadCallbackStage::Decompile || + stage == VideoCore::LoadCallbackStage::Build) { + ui->stage->setText(stage_translations[stage].arg(value).arg(total)); + } else { + ui->stage->setText(stage_translations[stage]); + } ui->value->setText(estimate); ui->progress_bar->setValue(static_cast<int>(value)); previous_time = now; diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp index 41ba3c4c6..bdee44b04 100644 --- a/src/yuzu/main.cpp +++ b/src/yuzu/main.cpp @@ -37,14 +37,20 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include <glad/glad.h> #define QT_NO_OPENGL +#include <QClipboard> +#include <QDesktopServices> #include <QDesktopWidget> #include <QDialogButtonBox> #include <QFile> #include <QFileDialog> +#include <QInputDialog> #include <QMessageBox> +#include <QProgressBar> +#include <QProgressDialog> +#include <QShortcut> +#include <QStatusBar> #include <QtConcurrent/QtConcurrent> -#include <QtGui> -#include <QtWidgets> + #include <fmt/format.h> #include "common/common_paths.h" #include "common/detached_tasks.h" @@ -55,11 +61,9 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "common/microprofile.h" #include "common/scm_rev.h" #include "common/scope_exit.h" -#include "common/string_util.h" #include "common/telemetry.h" #include "core/core.h" #include "core/crypto/key_manager.h" -#include "core/file_sys/bis_factory.h" #include "core/file_sys/card_image.h" #include "core/file_sys/content_archive.h" #include "core/file_sys/control_metadata.h" @@ -71,7 +75,6 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "core/frontend/applets/software_keyboard.h" #include "core/hle/kernel/process.h" #include "core/hle/service/filesystem/filesystem.h" -#include "core/hle/service/filesystem/fsp_ldr.h" #include "core/hle/service/nfp/nfp.h" #include "core/hle/service/sm/sm.h" #include "core/loader/loader.h" @@ -87,7 +90,6 @@ static FileSys::VirtualFile VfsDirectoryCreateFileWrapper(const FileSys::Virtual #include "yuzu/configuration/configure_dialog.h" #include "yuzu/debugger/console.h" #include "yuzu/debugger/graphics/graphics_breakpoints.h" -#include "yuzu/debugger/graphics/graphics_surface.h" #include "yuzu/debugger/profiler.h" #include "yuzu/debugger/wait_tree.h" #include "yuzu/discord.h" @@ -168,7 +170,8 @@ static void InitializeLogging() { GMainWindow::GMainWindow() : config(new Config()), emu_thread(nullptr), - vfs(std::make_shared<FileSys::RealVfsFilesystem>()) { + vfs(std::make_shared<FileSys::RealVfsFilesystem>()), + provider(std::make_unique<FileSys::ManualContentProvider>()) { InitializeLogging(); debug_context = Tegra::DebugContext::Construct(); @@ -200,13 +203,18 @@ GMainWindow::GMainWindow() .arg(Common::g_build_fullname, Common::g_scm_branch, Common::g_scm_desc)); show(); + Core::System::GetInstance().SetContentProvider( + std::make_unique<FileSys::ContentProviderUnion>()); + Core::System::GetInstance().RegisterContentProvider( + FileSys::ContentProviderUnionSlot::FrontendManual, provider.get()); + Service::FileSystem::CreateFactories(*vfs); + // Gen keys if necessary OnReinitializeKeys(ReinitializeKeyBehavior::NoWarning); - // Necessary to load titles from nand in gamelist. - Service::FileSystem::CreateFactories(*vfs); game_list->LoadCompatibilityList(); - game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan); + game_list->PopulateAsync(UISettings::values.game_directory_path, + UISettings::values.game_directory_deepscan); // Show one-time "callout" messages to the user ShowTelemetryCallout(); @@ -416,7 +424,7 @@ void GMainWindow::InitializeWidgets() { render_window = new GRenderWindow(this, emu_thread.get()); render_window->hide(); - game_list = new GameList(vfs, this); + game_list = new GameList(vfs, provider.get(), this); ui.horizontalLayout->addWidget(game_list); loading_screen = new LoadingScreen(this); @@ -475,11 +483,6 @@ void GMainWindow::InitializeDebugWidgets() { graphicsBreakpointsWidget->hide(); debug_menu->addAction(graphicsBreakpointsWidget->toggleViewAction()); - graphicsSurfaceWidget = new GraphicsSurfaceWidget(debug_context, this); - addDockWidget(Qt::RightDockWidgetArea, graphicsSurfaceWidget); - graphicsSurfaceWidget->hide(); - debug_menu->addAction(graphicsSurfaceWidget->toggleViewAction()); - waitTreeWidget = new WaitTreeWidget(this); addDockWidget(Qt::LeftDockWidgetArea, waitTreeWidget); waitTreeWidget->hide(); @@ -511,33 +514,34 @@ void GMainWindow::InitializeRecentFileMenuActions() { } void GMainWindow::InitializeHotkeys() { - hotkey_registry.RegisterHotkey("Main Window", "Load File", QKeySequence::Open); - hotkey_registry.RegisterHotkey("Main Window", "Start Emulation"); - hotkey_registry.RegisterHotkey("Main Window", "Continue/Pause", QKeySequence(Qt::Key_F4)); - hotkey_registry.RegisterHotkey("Main Window", "Restart", QKeySequence(Qt::Key_F5)); - hotkey_registry.RegisterHotkey("Main Window", "Fullscreen", QKeySequence::FullScreen); - hotkey_registry.RegisterHotkey("Main Window", "Exit Fullscreen", QKeySequence(Qt::Key_Escape), - Qt::ApplicationShortcut); - hotkey_registry.RegisterHotkey("Main Window", "Toggle Speed Limit", QKeySequence("CTRL+Z"), - Qt::ApplicationShortcut); - hotkey_registry.RegisterHotkey("Main Window", "Increase Speed Limit", QKeySequence("+"), - Qt::ApplicationShortcut); - hotkey_registry.RegisterHotkey("Main Window", "Decrease Speed Limit", QKeySequence("-"), - Qt::ApplicationShortcut); - hotkey_registry.RegisterHotkey("Main Window", "Load Amiibo", QKeySequence(Qt::Key_F2), - Qt::ApplicationShortcut); - hotkey_registry.RegisterHotkey("Main Window", "Capture Screenshot", - QKeySequence(QKeySequence::Print)); - hotkey_registry.RegisterHotkey("Main Window", "Change Docked Mode", QKeySequence(Qt::Key_F10)); - hotkey_registry.LoadHotkeys(); + ui.action_Load_File->setShortcut(hotkey_registry.GetKeySequence("Main Window", "Load File")); + ui.action_Load_File->setShortcutContext( + hotkey_registry.GetShortcutContext("Main Window", "Load File")); + + ui.action_Exit->setShortcut(hotkey_registry.GetKeySequence("Main Window", "Exit yuzu")); + ui.action_Exit->setShortcutContext( + hotkey_registry.GetShortcutContext("Main Window", "Exit yuzu")); + + ui.action_Stop->setShortcut(hotkey_registry.GetKeySequence("Main Window", "Stop Emulation")); + ui.action_Stop->setShortcutContext( + hotkey_registry.GetShortcutContext("Main Window", "Stop Emulation")); + + ui.action_Show_Filter_Bar->setShortcut( + hotkey_registry.GetKeySequence("Main Window", "Toggle Filter Bar")); + ui.action_Show_Filter_Bar->setShortcutContext( + hotkey_registry.GetShortcutContext("Main Window", "Toggle Filter Bar")); + + ui.action_Show_Status_Bar->setShortcut( + hotkey_registry.GetKeySequence("Main Window", "Toggle Status Bar")); + ui.action_Show_Status_Bar->setShortcutContext( + hotkey_registry.GetShortcutContext("Main Window", "Toggle Status Bar")); + connect(hotkey_registry.GetHotkey("Main Window", "Load File", this), &QShortcut::activated, this, &GMainWindow::OnMenuLoadFile); - connect(hotkey_registry.GetHotkey("Main Window", "Start Emulation", this), - &QShortcut::activated, this, &GMainWindow::OnStartGame); - connect(hotkey_registry.GetHotkey("Main Window", "Continue/Pause", this), &QShortcut::activated, - this, [&] { + connect(hotkey_registry.GetHotkey("Main Window", "Continue/Pause Emulation", this), + &QShortcut::activated, this, [&] { if (emulation_running) { if (emu_thread->IsRunning()) { OnPauseGame(); @@ -546,8 +550,8 @@ void GMainWindow::InitializeHotkeys() { } } }); - connect(hotkey_registry.GetHotkey("Main Window", "Restart", this), &QShortcut::activated, this, - [this] { + connect(hotkey_registry.GetHotkey("Main Window", "Restart Emulation", this), + &QShortcut::activated, this, [this] { if (!Core::System::GetInstance().IsPoweredOn()) return; BootGame(QString(game_path)); @@ -648,6 +652,8 @@ void GMainWindow::RestoreUIState() { void GMainWindow::ConnectWidgetEvents() { connect(game_list, &GameList::GameChosen, this, &GMainWindow::OnGameListLoadFile); connect(game_list, &GameList::OpenFolderRequested, this, &GMainWindow::OnGameListOpenFolder); + connect(game_list, &GameList::OpenTransferableShaderCacheRequested, this, + &GMainWindow::OnTransferableShaderCacheOpenFile); connect(game_list, &GameList::DumpRomFSRequested, this, &GMainWindow::OnGameListDumpRomFS); connect(game_list, &GameList::CopyTIDRequested, this, &GMainWindow::OnGameListCopyTID); connect(game_list, &GameList::NavigateToGamedbEntryRequested, this, @@ -692,7 +698,6 @@ void GMainWindow::ConnectMenuEvents() { &GMainWindow::ToggleWindowMode); connect(ui.action_Display_Dock_Widget_Headers, &QAction::triggered, this, &GMainWindow::OnDisplayTitleBars); - ui.action_Show_Filter_Bar->setShortcut(tr("CTRL+F")); connect(ui.action_Show_Filter_Bar, &QAction::triggered, this, &GMainWindow::OnToggleFilterBar); connect(ui.action_Show_Status_Bar, &QAction::triggered, statusBar(), &QStatusBar::setVisible); @@ -1082,6 +1087,39 @@ void GMainWindow::OnGameListOpenFolder(u64 program_id, GameListOpenTarget target QDesktopServices::openUrl(QUrl::fromLocalFile(qpath)); } +void GMainWindow::OnTransferableShaderCacheOpenFile(u64 program_id) { + ASSERT(program_id != 0); + + const QString tranferable_shader_cache_folder_path = + QString::fromStdString(FileUtil::GetUserPath(FileUtil::UserPath::ShaderDir)) + "opengl" + + DIR_SEP + "transferable"; + + const QString transferable_shader_cache_file_path = + tranferable_shader_cache_folder_path + DIR_SEP + + QString::fromStdString(fmt::format("{:016X}.bin", program_id)); + + if (!QFile::exists(transferable_shader_cache_file_path)) { + QMessageBox::warning(this, tr("Error Opening Transferable Shader Cache"), + tr("A shader cache for this title does not exist.")); + return; + } + + // Windows supports opening a folder with selecting a specified file in explorer. On every other + // OS we just open the transferable shader cache folder without preselecting the transferable + // shader cache file for the selected game. +#if defined(Q_OS_WIN) + const QString explorer = QStringLiteral("explorer"); + QStringList param; + if (!QFileInfo(transferable_shader_cache_file_path).isDir()) { + param << QStringLiteral("/select,"); + } + param << QDir::toNativeSeparators(transferable_shader_cache_file_path); + QProcess::startDetached(explorer, param); +#else + QDesktopServices::openUrl(QUrl::fromLocalFile(tranferable_shader_cache_folder_path)); +#endif +} + static std::size_t CalculateRomFSEntrySize(const FileSys::VirtualDir& dir, bool full) { std::size_t out = 0; @@ -1141,7 +1179,7 @@ void GMainWindow::OnGameListDumpRomFS(u64 program_id, const std::string& game_pa return; } - const auto installed = Service::FileSystem::GetUnionContents(); + const auto& installed = Core::System::GetInstance().GetContentProvider(); const auto romfs_title_id = SelectRomFSDumpTarget(installed, program_id); if (!romfs_title_id) { @@ -1241,8 +1279,8 @@ void GMainWindow::OnGameListOpenPerGameProperties(const std::string& file) { const auto reload = UISettings::values.is_game_list_reload_pending.exchange(false); if (reload) { - game_list->PopulateAsync(UISettings::values.gamedir, - UISettings::values.gamedir_deepscan); + game_list->PopulateAsync(UISettings::values.game_directory_path, + UISettings::values.game_directory_deepscan); } config->Save(); @@ -1330,7 +1368,8 @@ void GMainWindow::OnMenuInstallToNAND() { const auto success = [this]() { QMessageBox::information(this, tr("Successfully Installed"), tr("The file was successfully installed.")); - game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan); + game_list->PopulateAsync(UISettings::values.game_directory_path, + UISettings::values.game_directory_deepscan); }; const auto failed = [this]() { @@ -1457,8 +1496,8 @@ void GMainWindow::OnMenuInstallToNAND() { void GMainWindow::OnMenuSelectGameListRoot() { QString dir_path = QFileDialog::getExistingDirectory(this, tr("Select Directory")); if (!dir_path.isEmpty()) { - UISettings::values.gamedir = dir_path; - game_list->PopulateAsync(dir_path, UISettings::values.gamedir_deepscan); + UISettings::values.game_directory_path = dir_path; + game_list->PopulateAsync(dir_path, UISettings::values.game_directory_deepscan); } } @@ -1480,7 +1519,8 @@ void GMainWindow::OnMenuSelectEmulatedDirectory(EmulatedDirectoryTarget target) : FileUtil::UserPath::NANDDir, dir_path.toStdString()); Service::FileSystem::CreateFactories(*vfs); - game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan); + game_list->PopulateAsync(UISettings::values.game_directory_path, + UISettings::values.game_directory_deepscan); } } @@ -1624,6 +1664,7 @@ void GMainWindow::OnConfigure() { auto result = configureDialog.exec(); if (result == QDialog::Accepted) { configureDialog.applyConfiguration(); + InitializeHotkeys(); if (UISettings::values.theme != old_theme) UpdateUITheme(); if (UISettings::values.enable_discord_presence != old_discord_presence) @@ -1631,8 +1672,8 @@ void GMainWindow::OnConfigure() { const auto reload = UISettings::values.is_game_list_reload_pending.exchange(false); if (reload) { - game_list->PopulateAsync(UISettings::values.gamedir, - UISettings::values.gamedir_deepscan); + game_list->PopulateAsync(UISettings::values.game_directory_path, + UISettings::values.game_directory_deepscan); } config->Save(); @@ -1882,18 +1923,19 @@ void GMainWindow::OnReinitializeKeys(ReinitializeKeyBehavior behavior) { Service::FileSystem::CreateFactories(*vfs); if (behavior == ReinitializeKeyBehavior::Warning) { - game_list->PopulateAsync(UISettings::values.gamedir, UISettings::values.gamedir_deepscan); + game_list->PopulateAsync(UISettings::values.game_directory_path, + UISettings::values.game_directory_deepscan); } } -std::optional<u64> GMainWindow::SelectRomFSDumpTarget( - const FileSys::RegisteredCacheUnion& installed, u64 program_id) { +std::optional<u64> GMainWindow::SelectRomFSDumpTarget(const FileSys::ContentProvider& installed, + u64 program_id) { const auto dlc_entries = installed.ListEntriesFilter(FileSys::TitleType::AOC, FileSys::ContentRecordType::Data); - std::vector<FileSys::RegisteredCacheEntry> dlc_match; + std::vector<FileSys::ContentProviderEntry> dlc_match; 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::RegisteredCacheEntry& entry) { + [&program_id, &installed](const FileSys::ContentProviderEntry& entry) { return (entry.title_id & DLC_BASE_TITLE_ID_MASK) == program_id && installed.GetEntry(entry)->GetStatus() == Loader::ResultStatus::Success; }); @@ -1989,6 +2031,18 @@ void GMainWindow::dragMoveEvent(QDragMoveEvent* event) { event->acceptProposedAction(); } +void GMainWindow::keyPressEvent(QKeyEvent* event) { + if (render_window) { + render_window->ForwardKeyPressEvent(event); + } +} + +void GMainWindow::keyReleaseEvent(QKeyEvent* event) { + if (render_window) { + render_window->ForwardKeyReleaseEvent(event); + } +} + bool GMainWindow::ConfirmChangeGame() { if (emu_thread == nullptr) return true; @@ -2056,7 +2110,8 @@ int main(int argc, char* argv[]) { QCoreApplication::setOrganizationName("yuzu team"); QCoreApplication::setApplicationName("yuzu"); - QApplication::setAttribute(Qt::AA_DontCheckOpenGLContextThreadAffinity); + // Enables the core to make the qt created contexts current on std::threads + QCoreApplication::setAttribute(Qt::AA_DontCheckOpenGLContextThreadAffinity); QApplication app(argc, argv); // Qt changes the locale and causes issues in float conversion using std::to_string() when diff --git a/src/yuzu/main.h b/src/yuzu/main.h index e07c892cf..ce5045819 100644 --- a/src/yuzu/main.h +++ b/src/yuzu/main.h @@ -23,7 +23,6 @@ class EmuThread; class GameList; class GImageInfo; class GraphicsBreakPointsWidget; -class GraphicsSurfaceWidget; class GRenderWindow; class LoadingScreen; class MicroProfileDialog; @@ -37,7 +36,8 @@ struct SoftwareKeyboardParameters; } // namespace Core::Frontend namespace FileSys { -class RegisteredCacheUnion; +class ContentProvider; +class ManualContentProvider; class VfsFilesystem; } // namespace FileSys @@ -120,7 +120,6 @@ private: void InitializeWidgets(); void InitializeDebugWidgets(); void InitializeRecentFileMenuActions(); - void InitializeHotkeys(); void SetDefaultUIGeometry(); void RestoreUIState(); @@ -176,6 +175,7 @@ private slots: /// Called whenever a user selects a game in the game list widget. void OnGameListLoadFile(QString game_path); void OnGameListOpenFolder(u64 program_id, GameListOpenTarget target); + void OnTransferableShaderCacheOpenFile(u64 program_id); void OnGameListDumpRomFS(u64 program_id, const std::string& game_path); void OnGameListCopyTID(u64 program_id); void OnGameListNavigateToGamedbEntry(u64 program_id, @@ -195,6 +195,7 @@ private slots: void OnAbout(); void OnToggleFilterBar(); void OnDisplayTitleBars(bool); + void InitializeHotkeys(); void ToggleFullscreen(); void ShowFullscreen(); void HideFullscreen(); @@ -204,7 +205,7 @@ private slots: void OnReinitializeKeys(ReinitializeKeyBehavior behavior); private: - std::optional<u64> SelectRomFSDumpTarget(const FileSys::RegisteredCacheUnion&, u64 program_id); + std::optional<u64> SelectRomFSDumpTarget(const FileSys::ContentProvider&, u64 program_id); void UpdateStatusBar(); Ui::MainWindow ui; @@ -232,12 +233,12 @@ private: // FS std::shared_ptr<FileSys::VfsFilesystem> vfs; + std::unique_ptr<FileSys::ManualContentProvider> provider; // Debugger panes ProfilerWidget* profilerWidget; MicroProfileDialog* microProfileDialog; GraphicsBreakPointsWidget* graphicsBreakpointsWidget; - GraphicsSurfaceWidget* graphicsSurfaceWidget; WaitTreeWidget* waitTreeWidget; QAction* actions_recent_files[max_recent_files_item]; @@ -251,4 +252,8 @@ protected: void dropEvent(QDropEvent* event) override; void dragEnterEvent(QDragEnterEvent* event) override; void dragMoveEvent(QDragMoveEvent* event) override; + + // Overrides used to forward signals to the render window when the focus moves out. + void keyPressEvent(QKeyEvent* event) override; + void keyReleaseEvent(QKeyEvent* event) override; }; diff --git a/src/yuzu/ui_settings.cpp b/src/yuzu/ui_settings.cpp index a314493fc..4bdc302e0 100644 --- a/src/yuzu/ui_settings.cpp +++ b/src/yuzu/ui_settings.cpp @@ -12,5 +12,4 @@ const Themes themes{{ }}; Values values = {}; - } // namespace UISettings diff --git a/src/yuzu/ui_settings.h b/src/yuzu/ui_settings.h index 82aaeedb0..dbd318e20 100644 --- a/src/yuzu/ui_settings.h +++ b/src/yuzu/ui_settings.h @@ -15,7 +15,12 @@ namespace UISettings { using ContextualShortcut = std::pair<QString, int>; -using Shortcut = std::pair<QString, ContextualShortcut>; + +struct Shortcut { + QString name; + QString group; + ContextualShortcut shortcut; +}; using Themes = std::array<std::pair<const char*, const char*>, 2>; extern const Themes themes; @@ -50,8 +55,8 @@ struct Values { QString roms_path; QString symbols_path; QString screenshot_path; - QString gamedir; - bool gamedir_deepscan; + QString game_directory_path; + bool game_directory_deepscan; QStringList recent_files; QString theme; diff --git a/src/yuzu/util/sequence_dialog/sequence_dialog.cpp b/src/yuzu/util/sequence_dialog/sequence_dialog.cpp new file mode 100644 index 000000000..d3edf6ec3 --- /dev/null +++ b/src/yuzu/util/sequence_dialog/sequence_dialog.cpp @@ -0,0 +1,37 @@ +// Copyright 2018 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#include <QDialogButtonBox> +#include <QKeySequenceEdit> +#include <QVBoxLayout> +#include "yuzu/util/sequence_dialog/sequence_dialog.h" + +SequenceDialog::SequenceDialog(QWidget* parent) : QDialog(parent) { + setWindowTitle(tr("Enter a hotkey")); + auto* layout = new QVBoxLayout(this); + key_sequence = new QKeySequenceEdit; + layout->addWidget(key_sequence); + auto* buttons = + new QDialogButtonBox(QDialogButtonBox::Ok | QDialogButtonBox::Cancel, Qt::Horizontal); + buttons->setCenterButtons(true); + layout->addWidget(buttons); + connect(buttons, &QDialogButtonBox::accepted, this, &QDialog::accept); + connect(buttons, &QDialogButtonBox::rejected, this, &QDialog::reject); + setWindowFlags(windowFlags() & ~Qt::WindowContextHelpButtonHint); +} + +SequenceDialog::~SequenceDialog() = default; + +QKeySequence SequenceDialog::GetSequence() const { + // Only the first key is returned. The other 3, if present, are ignored. + return QKeySequence(key_sequence->keySequence()[0]); +} + +bool SequenceDialog::focusNextPrevChild(bool next) { + return false; +} + +void SequenceDialog::closeEvent(QCloseEvent*) { + reject(); +} diff --git a/src/yuzu/util/sequence_dialog/sequence_dialog.h b/src/yuzu/util/sequence_dialog/sequence_dialog.h new file mode 100644 index 000000000..969c77740 --- /dev/null +++ b/src/yuzu/util/sequence_dialog/sequence_dialog.h @@ -0,0 +1,24 @@ +// Copyright 2018 Citra Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include <QDialog> + +class QKeySequenceEdit; + +class SequenceDialog : public QDialog { + Q_OBJECT + +public: + explicit SequenceDialog(QWidget* parent = nullptr); + ~SequenceDialog() override; + + QKeySequence GetSequence() const; + void closeEvent(QCloseEvent*) override; + +private: + QKeySequenceEdit* key_sequence; + bool focusNextPrevChild(bool next) override; +}; diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp index de7a26e14..68a176032 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.cpp @@ -19,6 +19,37 @@ #include "input_common/sdl/sdl.h" #include "yuzu_cmd/emu_window/emu_window_sdl2.h" +class SDLGLContext : public Core::Frontend::GraphicsContext { +public: + explicit SDLGLContext() { + // create a hidden window to make the shared context against + window = SDL_CreateWindow("", SDL_WINDOWPOS_UNDEFINED, // x position + SDL_WINDOWPOS_UNDEFINED, // y position + Layout::ScreenUndocked::Width, Layout::ScreenUndocked::Height, + SDL_WINDOW_OPENGL | SDL_WINDOW_HIDDEN); + context = SDL_GL_CreateContext(window); + } + + ~SDLGLContext() { + SDL_GL_DeleteContext(context); + SDL_DestroyWindow(window); + } + + void MakeCurrent() override { + SDL_GL_MakeCurrent(window, context); + } + + void DoneCurrent() override { + SDL_GL_MakeCurrent(window, nullptr); + } + + void SwapBuffers() override {} + +private: + SDL_Window* window; + SDL_GLContext context; +}; + void EmuWindow_SDL2::OnMouseMotion(s32 x, s32 y) { TouchMoved((unsigned)std::max(x, 0), (unsigned)std::max(y, 0)); InputCommon::GetMotionEmu()->Tilt(x, y); @@ -153,6 +184,7 @@ EmuWindow_SDL2::EmuWindow_SDL2(bool fullscreen) { SDL_GL_SetAttribute(SDL_GL_GREEN_SIZE, 8); SDL_GL_SetAttribute(SDL_GL_BLUE_SIZE, 8); SDL_GL_SetAttribute(SDL_GL_ALPHA_SIZE, 0); + SDL_GL_SetAttribute(SDL_GL_SHARE_WITH_CURRENT_CONTEXT, 1); std::string window_title = fmt::format("yuzu {} | {}-{}", Common::g_build_fullname, Common::g_scm_branch, Common::g_scm_desc); @@ -171,7 +203,6 @@ EmuWindow_SDL2::EmuWindow_SDL2(bool fullscreen) { if (fullscreen) { Fullscreen(); } - gl_context = SDL_GL_CreateContext(render_window); if (gl_context == nullptr) { @@ -278,3 +309,7 @@ void EmuWindow_SDL2::OnMinimalClientAreaChangeRequest( SDL_SetWindowMinimumSize(render_window, minimal_size.first, minimal_size.second); } + +std::unique_ptr<Core::Frontend::GraphicsContext> EmuWindow_SDL2::CreateSharedContext() const { + return std::make_unique<SDLGLContext>(); +} diff --git a/src/yuzu_cmd/emu_window/emu_window_sdl2.h b/src/yuzu_cmd/emu_window/emu_window_sdl2.h index b0d4116cc..17e98227f 100644 --- a/src/yuzu_cmd/emu_window/emu_window_sdl2.h +++ b/src/yuzu_cmd/emu_window/emu_window_sdl2.h @@ -27,6 +27,8 @@ public: /// Releases the GL context from the caller thread void DoneCurrent() override; + std::unique_ptr<Core::Frontend::GraphicsContext> CreateSharedContext() const override; + /// Whether the window is still open, and a close request hasn't yet been sent bool IsOpen() const; diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp index 245f25847..7ea4a1b18 100644 --- a/src/yuzu_cmd/yuzu.cpp +++ b/src/yuzu_cmd/yuzu.cpp @@ -33,6 +33,7 @@ #include "yuzu_cmd/emu_window/emu_window_sdl2.h" #include <getopt.h> +#include "core/file_sys/registered_cache.h" #ifndef _MSC_VER #include <unistd.h> #endif @@ -178,6 +179,7 @@ int main(int argc, char** argv) { } Core::System& system{Core::System::GetInstance()}; + system.SetContentProvider(std::make_unique<FileSys::ContentProviderUnion>()); system.SetFilesystem(std::make_shared<FileSys::RealVfsFilesystem>()); Service::FileSystem::CreateFactories(*system.GetFilesystem()); |