summaryrefslogtreecommitdiffstats
path: root/src
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--src/audio_core/audio_renderer.h14
-rw-r--r--src/audio_core/cubeb_sink.cpp15
-rw-r--r--src/audio_core/cubeb_sink.h4
-rw-r--r--src/common/color.h40
-rw-r--r--src/common/logging/backend.cpp50
-rw-r--r--src/common/logging/backend.h4
-rw-r--r--src/common/math_util.h4
-rw-r--r--src/common/quaternion.h10
-rw-r--r--src/common/vector_math.h4
-rw-r--r--src/core/CMakeLists.txt1
-rw-r--r--src/core/core.cpp19
-rw-r--r--src/core/core.h4
-rw-r--r--src/core/frontend/emu_window.cpp2
-rw-r--r--src/core/frontend/emu_window.h2
-rw-r--r--src/core/frontend/framebuffer_layout.cpp12
-rw-r--r--src/core/frontend/framebuffer_layout.h2
-rw-r--r--src/core/frontend/input.h2
-rw-r--r--src/core/hle/ipc.h4
-rw-r--r--src/core/hle/kernel/address_arbiter.cpp147
-rw-r--r--src/core/hle/kernel/address_arbiter.h74
-rw-r--r--src/core/hle/kernel/errors.h1
-rw-r--r--src/core/hle/kernel/handle_table.cpp40
-rw-r--r--src/core/hle/kernel/handle_table.h25
-rw-r--r--src/core/hle/kernel/hle_ipc.h10
-rw-r--r--src/core/hle/kernel/kernel.cpp30
-rw-r--r--src/core/hle/kernel/kernel.h26
-rw-r--r--src/core/hle/kernel/process.cpp8
-rw-r--r--src/core/hle/kernel/process_capability.cpp4
-rw-r--r--src/core/hle/kernel/process_capability.h4
-rw-r--r--src/core/hle/kernel/svc.cpp40
-rw-r--r--src/core/hle/kernel/thread.cpp2
-rw-r--r--src/core/hle/kernel/vm_manager.cpp47
-rw-r--r--src/core/hle/kernel/vm_manager.h24
-rw-r--r--src/core/hle/result.h1
-rw-r--r--src/core/hle/service/am/applets/software_keyboard.cpp1
-rw-r--r--src/core/hle/service/am/applets/software_keyboard.h3
-rw-r--r--src/core/hle/service/audio/audout_u.cpp13
-rw-r--r--src/core/hle/service/audio/audren_u.cpp64
-rw-r--r--src/core/hle/service/audio/audren_u.h3
-rw-r--r--src/core/hle/service/audio/errors.h15
-rw-r--r--src/core/hle/service/hid/hid.h2
-rw-r--r--src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp2
-rw-r--r--src/core/hle/service/nvdrv/devices/nvdisp_disp0.h2
-rw-r--r--src/core/hle/service/nvflinger/buffer_queue.cpp2
-rw-r--r--src/core/hle/service/nvflinger/buffer_queue.h4
-rw-r--r--src/core/hle/service/vi/vi.cpp2
-rw-r--r--src/core/memory.cpp23
-rw-r--r--src/input_common/motion_emu.cpp28
-rw-r--r--src/tests/core/arm/arm_test_common.cpp6
-rw-r--r--src/video_core/CMakeLists.txt6
-rw-r--r--src/video_core/engines/fermi_2d.cpp15
-rw-r--r--src/video_core/engines/fermi_2d.h2
-rw-r--r--src/video_core/engines/kepler_compute.cpp3
-rw-r--r--src/video_core/engines/kepler_compute.h3
-rw-r--r--src/video_core/engines/kepler_memory.h1
-rw-r--r--src/video_core/engines/maxwell_3d.cpp66
-rw-r--r--src/video_core/engines/maxwell_3d.h17
-rw-r--r--src/video_core/engines/maxwell_dma.cpp1
-rw-r--r--src/video_core/engines/maxwell_dma.h1
-rw-r--r--src/video_core/engines/shader_bytecode.h1
-rw-r--r--src/video_core/gpu.h2
-rw-r--r--src/video_core/rasterizer_cache.h18
-rw-r--r--src/video_core/rasterizer_interface.h4
-rw-r--r--src/video_core/renderer_base.cpp1
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.cpp40
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer.h4
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.cpp138
-rw-r--r--src/video_core/renderer_opengl/gl_rasterizer_cache.h91
-rw-r--r--src/video_core/renderer_opengl/gl_shader_disk_cache.cpp2
-rw-r--r--src/video_core/renderer_opengl/gl_state.cpp4
-rw-r--r--src/video_core/renderer_opengl/renderer_opengl.cpp28
-rw-r--r--src/video_core/renderer_opengl/renderer_opengl.h5
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.cpp116
-rw-r--r--src/video_core/renderer_vulkan/vk_buffer_cache.h87
-rw-r--r--src/video_core/renderer_vulkan/vk_memory_manager.cpp2
-rw-r--r--src/video_core/renderer_vulkan/vk_resource_manager.cpp12
-rw-r--r--src/video_core/renderer_vulkan/vk_resource_manager.h2
-rw-r--r--src/video_core/renderer_vulkan/vk_stream_buffer.cpp90
-rw-r--r--src/video_core/renderer_vulkan/vk_stream_buffer.h72
-rw-r--r--src/video_core/surface.cpp2
-rw-r--r--src/web_service/verify_login.h2
-rw-r--r--src/web_service/web_backend.cpp1
-rw-r--r--src/yuzu/compatdb.cpp6
-rw-r--r--src/yuzu/debugger/graphics/graphics_surface.cpp2
-rw-r--r--src/yuzu/debugger/wait_tree.cpp8
-rw-r--r--src/yuzu/debugger/wait_tree.h3
-rw-r--r--src/yuzu/main.cpp7
-rw-r--r--src/yuzu_cmd/yuzu.cpp2
88 files changed, 1236 insertions, 477 deletions
diff --git a/src/audio_core/audio_renderer.h b/src/audio_core/audio_renderer.h
index 201ec7a3c..b2e5d336c 100644
--- a/src/audio_core/audio_renderer.h
+++ b/src/audio_core/audio_renderer.h
@@ -46,16 +46,18 @@ struct AudioRendererParameter {
u32_le sample_rate;
u32_le sample_count;
u32_le mix_buffer_count;
- u32_le unknown_c;
+ u32_le submix_count;
u32_le voice_count;
u32_le sink_count;
u32_le effect_count;
- u32_le unknown_1c;
- u8 unknown_20;
- INSERT_PADDING_BYTES(3);
+ u32_le performance_frame_count;
+ u8 is_voice_drop_enabled;
+ u8 unknown_21;
+ u8 unknown_22;
+ u8 execution_mode;
u32_le splitter_count;
- u32_le unknown_2c;
- INSERT_PADDING_WORDS(1);
+ u32_le num_splitter_send_channels;
+ u32_le unknown_30;
u32_le revision;
};
static_assert(sizeof(AudioRendererParameter) == 52, "AudioRendererParameter is an invalid size");
diff --git a/src/audio_core/cubeb_sink.cpp b/src/audio_core/cubeb_sink.cpp
index dc45dedd3..1da0b9f2a 100644
--- a/src/audio_core/cubeb_sink.cpp
+++ b/src/audio_core/cubeb_sink.cpp
@@ -12,6 +12,10 @@
#include "common/ring_buffer.h"
#include "core/settings.h"
+#ifdef _MSC_VER
+#include <objbase.h>
+#endif
+
namespace AudioCore {
class CubebSinkStream final : public SinkStream {
@@ -108,6 +112,11 @@ private:
};
CubebSink::CubebSink(std::string_view target_device_name) {
+ // Cubeb requires COM to be initialized on the thread calling cubeb_init on Windows
+#ifdef _MSC_VER
+ com_init_result = CoInitializeEx(nullptr, COINIT_MULTITHREADED);
+#endif
+
if (cubeb_init(&ctx, "yuzu", nullptr) != CUBEB_OK) {
LOG_CRITICAL(Audio_Sink, "cubeb_init failed");
return;
@@ -142,6 +151,12 @@ CubebSink::~CubebSink() {
}
cubeb_destroy(ctx);
+
+#ifdef _MSC_VER
+ if (SUCCEEDED(com_init_result)) {
+ CoUninitialize();
+ }
+#endif
}
SinkStream& CubebSink::AcquireSinkStream(u32 sample_rate, u32 num_channels,
diff --git a/src/audio_core/cubeb_sink.h b/src/audio_core/cubeb_sink.h
index efb9d1634..511df7bb1 100644
--- a/src/audio_core/cubeb_sink.h
+++ b/src/audio_core/cubeb_sink.h
@@ -25,6 +25,10 @@ private:
cubeb* ctx{};
cubeb_devid output_device{};
std::vector<SinkStreamPtr> sink_streams;
+
+#ifdef _MSC_VER
+ u32 com_init_result = 0;
+#endif
};
std::vector<std::string> ListCubebSinkDevices();
diff --git a/src/common/color.h b/src/common/color.h
index 0379040be..3a2222077 100644
--- a/src/common/color.h
+++ b/src/common/color.h
@@ -55,36 +55,36 @@ constexpr u8 Convert8To6(u8 value) {
/**
* Decode a color stored in RGBA8 format
* @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
*/
-inline Math::Vec4<u8> DecodeRGBA8(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGBA8(const u8* bytes) {
return {bytes[3], bytes[2], bytes[1], bytes[0]};
}
/**
* Decode a color stored in RGB8 format
* @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
*/
-inline Math::Vec4<u8> DecodeRGB8(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGB8(const u8* bytes) {
return {bytes[2], bytes[1], bytes[0], 255};
}
/**
* Decode a color stored in RG8 (aka HILO8) format
* @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
*/
-inline Math::Vec4<u8> DecodeRG8(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRG8(const u8* bytes) {
return {bytes[1], bytes[0], 0, 255};
}
/**
* Decode a color stored in RGB565 format
* @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
*/
-inline Math::Vec4<u8> DecodeRGB565(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGB565(const u8* bytes) {
u16_le pixel;
std::memcpy(&pixel, bytes, sizeof(pixel));
return {Convert5To8((pixel >> 11) & 0x1F), Convert6To8((pixel >> 5) & 0x3F),
@@ -94,9 +94,9 @@ inline Math::Vec4<u8> DecodeRGB565(const u8* bytes) {
/**
* Decode a color stored in RGB5A1 format
* @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
*/
-inline Math::Vec4<u8> DecodeRGB5A1(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGB5A1(const u8* bytes) {
u16_le pixel;
std::memcpy(&pixel, bytes, sizeof(pixel));
return {Convert5To8((pixel >> 11) & 0x1F), Convert5To8((pixel >> 6) & 0x1F),
@@ -106,9 +106,9 @@ inline Math::Vec4<u8> DecodeRGB5A1(const u8* bytes) {
/**
* Decode a color stored in RGBA4 format
* @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
*/
-inline Math::Vec4<u8> DecodeRGBA4(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGBA4(const u8* bytes) {
u16_le pixel;
std::memcpy(&pixel, bytes, sizeof(pixel));
return {Convert4To8((pixel >> 12) & 0xF), Convert4To8((pixel >> 8) & 0xF),
@@ -138,9 +138,9 @@ inline u32 DecodeD24(const u8* bytes) {
/**
* Decode a depth value and a stencil value stored in D24S8 format
* @param bytes Pointer to encoded source values
- * @return Resulting values stored as a Math::Vec2
+ * @return Resulting values stored as a Common::Vec2
*/
-inline Math::Vec2<u32> DecodeD24S8(const u8* bytes) {
+inline Common::Vec2<u32> DecodeD24S8(const u8* bytes) {
return {static_cast<u32>((bytes[2] << 16) | (bytes[1] << 8) | bytes[0]), bytes[3]};
}
@@ -149,7 +149,7 @@ inline Math::Vec2<u32> DecodeD24S8(const u8* bytes) {
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
-inline void EncodeRGBA8(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGBA8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[3] = color.r();
bytes[2] = color.g();
bytes[1] = color.b();
@@ -161,7 +161,7 @@ inline void EncodeRGBA8(const Math::Vec4<u8>& color, u8* bytes) {
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
-inline void EncodeRGB8(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGB8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[2] = color.r();
bytes[1] = color.g();
bytes[0] = color.b();
@@ -172,7 +172,7 @@ inline void EncodeRGB8(const Math::Vec4<u8>& color, u8* bytes) {
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
-inline void EncodeRG8(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRG8(const Common::Vec4<u8>& color, u8* bytes) {
bytes[1] = color.r();
bytes[0] = color.g();
}
@@ -181,7 +181,7 @@ inline void EncodeRG8(const Math::Vec4<u8>& color, u8* bytes) {
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
-inline void EncodeRGB565(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGB565(const Common::Vec4<u8>& color, u8* bytes) {
const u16_le data =
(Convert8To5(color.r()) << 11) | (Convert8To6(color.g()) << 5) | Convert8To5(color.b());
@@ -193,7 +193,7 @@ inline void EncodeRGB565(const Math::Vec4<u8>& color, u8* bytes) {
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
-inline void EncodeRGB5A1(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGB5A1(const Common::Vec4<u8>& color, u8* bytes) {
const u16_le data = (Convert8To5(color.r()) << 11) | (Convert8To5(color.g()) << 6) |
(Convert8To5(color.b()) << 1) | Convert8To1(color.a());
@@ -205,7 +205,7 @@ inline void EncodeRGB5A1(const Math::Vec4<u8>& color, u8* bytes) {
* @param color Source color to encode
* @param bytes Destination pointer to store encoded color
*/
-inline void EncodeRGBA4(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGBA4(const Common::Vec4<u8>& color, u8* bytes) {
const u16 data = (Convert8To4(color.r()) << 12) | (Convert8To4(color.g()) << 8) |
(Convert8To4(color.b()) << 4) | Convert8To4(color.a());
diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp
index b369f199f..4462ff3fb 100644
--- a/src/common/logging/backend.cpp
+++ b/src/common/logging/backend.cpp
@@ -39,8 +39,10 @@ public:
Impl(Impl const&) = delete;
const Impl& operator=(Impl const&) = delete;
- void PushEntry(Entry e) {
- message_queue.Push(std::move(e));
+ void PushEntry(Class log_class, Level log_level, const char* filename, unsigned int line_num,
+ const char* function, std::string message) {
+ message_queue.Push(
+ CreateEntry(log_class, log_level, filename, line_num, function, std::move(message)));
}
void AddBackend(std::unique_ptr<Backend> backend) {
@@ -108,11 +110,30 @@ private:
backend_thread.join();
}
+ Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr,
+ const char* function, std::string message) const {
+ using std::chrono::duration_cast;
+ using std::chrono::steady_clock;
+
+ Entry entry;
+ entry.timestamp =
+ duration_cast<std::chrono::microseconds>(steady_clock::now() - time_origin);
+ entry.log_class = log_class;
+ entry.log_level = log_level;
+ entry.filename = Common::TrimSourcePath(filename);
+ entry.line_num = line_nr;
+ entry.function = function;
+ entry.message = std::move(message);
+
+ return entry;
+ }
+
std::mutex writing_mutex;
std::thread backend_thread;
std::vector<std::unique_ptr<Backend>> backends;
Common::MPSCQueue<Log::Entry> message_queue;
Filter filter;
+ std::chrono::steady_clock::time_point time_origin{std::chrono::steady_clock::now()};
};
void ConsoleBackend::Write(const Entry& entry) {
@@ -271,25 +292,6 @@ const char* GetLevelName(Level log_level) {
#undef LVL
}
-Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr,
- const char* function, std::string message) {
- using std::chrono::duration_cast;
- using std::chrono::steady_clock;
-
- static steady_clock::time_point time_origin = steady_clock::now();
-
- Entry entry;
- entry.timestamp = duration_cast<std::chrono::microseconds>(steady_clock::now() - time_origin);
- entry.log_class = log_class;
- entry.log_level = log_level;
- entry.filename = Common::TrimSourcePath(filename);
- entry.line_num = line_nr;
- entry.function = function;
- entry.message = std::move(message);
-
- return entry;
-}
-
void SetGlobalFilter(const Filter& filter) {
Impl::Instance().SetGlobalFilter(filter);
}
@@ -314,9 +316,7 @@ void FmtLogMessageImpl(Class log_class, Level log_level, const char* filename,
if (!filter.CheckMessage(log_class, log_level))
return;
- Entry entry =
- CreateEntry(log_class, log_level, filename, line_num, function, fmt::vformat(format, args));
-
- instance.PushEntry(std::move(entry));
+ instance.PushEntry(log_class, log_level, filename, line_num, function,
+ fmt::vformat(format, args));
}
} // namespace Log
diff --git a/src/common/logging/backend.h b/src/common/logging/backend.h
index a31ee6968..fca0267a1 100644
--- a/src/common/logging/backend.h
+++ b/src/common/logging/backend.h
@@ -135,10 +135,6 @@ const char* GetLogClassName(Class log_class);
*/
const char* GetLevelName(Level log_level);
-/// Creates a log entry by formatting the given source location, and message.
-Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr,
- const char* function, std::string message);
-
/**
* The global filter will prevent any messages from even being processed if they are filtered. Each
* backend can have a filter, but if the level is lower than the global filter, the backend will
diff --git a/src/common/math_util.h b/src/common/math_util.h
index 94b4394c5..cff3d48c5 100644
--- a/src/common/math_util.h
+++ b/src/common/math_util.h
@@ -7,7 +7,7 @@
#include <cstdlib>
#include <type_traits>
-namespace MathUtil {
+namespace Common {
constexpr float PI = 3.14159265f;
@@ -41,4 +41,4 @@ struct Rectangle {
}
};
-} // namespace MathUtil
+} // namespace Common
diff --git a/src/common/quaternion.h b/src/common/quaternion.h
index c528c0b68..370198ae0 100644
--- a/src/common/quaternion.h
+++ b/src/common/quaternion.h
@@ -6,12 +6,12 @@
#include "common/vector_math.h"
-namespace Math {
+namespace Common {
template <typename T>
class Quaternion {
public:
- Math::Vec3<T> xyz;
+ Vec3<T> xyz;
T w{};
Quaternion<decltype(-T{})> Inverse() const {
@@ -38,12 +38,12 @@ public:
};
template <typename T>
-auto QuaternionRotate(const Quaternion<T>& q, const Math::Vec3<T>& v) {
+auto QuaternionRotate(const Quaternion<T>& q, const Vec3<T>& v) {
return v + 2 * Cross(q.xyz, Cross(q.xyz, v) + v * q.w);
}
-inline Quaternion<float> MakeQuaternion(const Math::Vec3<float>& axis, float angle) {
+inline Quaternion<float> MakeQuaternion(const Vec3<float>& axis, float angle) {
return {axis * std::sin(angle / 2), std::cos(angle / 2)};
}
-} // namespace Math
+} // namespace Common
diff --git a/src/common/vector_math.h b/src/common/vector_math.h
index 8feb49941..429485329 100644
--- a/src/common/vector_math.h
+++ b/src/common/vector_math.h
@@ -33,7 +33,7 @@
#include <cmath>
#include <type_traits>
-namespace Math {
+namespace Common {
template <typename T>
class Vec2;
@@ -690,4 +690,4 @@ constexpr Vec4<T> MakeVec(const T& x, const Vec3<T>& yzw) {
return MakeVec(x, yzw[0], yzw[1], yzw[2]);
}
-} // namespace Math
+} // namespace Common
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index 988356c65..8ccb2d5f0 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -217,6 +217,7 @@ add_library(core STATIC
hle/service/audio/audren_u.h
hle/service/audio/codecctl.cpp
hle/service/audio/codecctl.h
+ hle/service/audio/errors.h
hle/service/audio/hwopus.cpp
hle/service/audio/hwopus.h
hle/service/bcat/bcat.cpp
diff --git a/src/core/core.cpp b/src/core/core.cpp
index ab7181a05..d741ef90d 100644
--- a/src/core/core.cpp
+++ b/src/core/core.cpp
@@ -78,6 +78,7 @@ FileSys::VirtualFile GetGameFileFromPath(const FileSys::VirtualFilesystem& vfs,
return vfs->OpenFile(path, FileSys::Mode::Read);
}
struct System::Impl {
+ explicit Impl(System& system) : kernel{system} {}
Cpu& CurrentCpuCore() {
return cpu_core_manager.GetCurrentCore();
@@ -95,7 +96,7 @@ struct System::Impl {
LOG_DEBUG(HW_Memory, "initialized OK");
core_timing.Initialize();
- kernel.Initialize(core_timing);
+ kernel.Initialize();
const auto current_time = std::chrono::duration_cast<std::chrono::seconds>(
std::chrono::system_clock::now().time_since_epoch());
@@ -182,13 +183,13 @@ struct System::Impl {
void Shutdown() {
// Log last frame performance stats
- auto perf_results = GetAndResetPerfStats();
- Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_EmulationSpeed",
- perf_results.emulation_speed * 100.0);
- Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_Framerate",
- perf_results.game_fps);
- Telemetry().AddField(Telemetry::FieldType::Performance, "Shutdown_Frametime",
- perf_results.frametime * 1000.0);
+ const auto perf_results = GetAndResetPerfStats();
+ telemetry_session->AddField(Telemetry::FieldType::Performance, "Shutdown_EmulationSpeed",
+ perf_results.emulation_speed * 100.0);
+ telemetry_session->AddField(Telemetry::FieldType::Performance, "Shutdown_Framerate",
+ perf_results.game_fps);
+ telemetry_session->AddField(Telemetry::FieldType::Performance, "Shutdown_Frametime",
+ perf_results.frametime * 1000.0);
is_powered_on = false;
@@ -265,7 +266,7 @@ struct System::Impl {
Core::FrameLimiter frame_limiter;
};
-System::System() : impl{std::make_unique<Impl>()} {}
+System::System() : impl{std::make_unique<Impl>(*this)} {}
System::~System() = default;
Cpu& System::CurrentCpuCore() {
diff --git a/src/core/core.h b/src/core/core.h
index d720013f7..ba76a41d8 100644
--- a/src/core/core.h
+++ b/src/core/core.h
@@ -293,10 +293,6 @@ inline ARM_Interface& CurrentArmInterface() {
return System::GetInstance().CurrentArmInterface();
}
-inline TelemetrySession& Telemetry() {
- return System::GetInstance().TelemetrySession();
-}
-
inline Kernel::Process* CurrentProcess() {
return System::GetInstance().CurrentProcess();
}
diff --git a/src/core/frontend/emu_window.cpp b/src/core/frontend/emu_window.cpp
index 9dd493efb..e29afd630 100644
--- a/src/core/frontend/emu_window.cpp
+++ b/src/core/frontend/emu_window.cpp
@@ -67,7 +67,7 @@ static bool IsWithinTouchscreen(const Layout::FramebufferLayout& layout, unsigne
framebuffer_x >= layout.screen.left && framebuffer_x < layout.screen.right);
}
-std::tuple<unsigned, unsigned> EmuWindow::ClipToTouchScreen(unsigned new_x, unsigned new_y) {
+std::tuple<unsigned, unsigned> EmuWindow::ClipToTouchScreen(unsigned new_x, unsigned new_y) const {
new_x = std::max(new_x, framebuffer_layout.screen.left);
new_x = std::min(new_x, framebuffer_layout.screen.right - 1);
diff --git a/src/core/frontend/emu_window.h b/src/core/frontend/emu_window.h
index 7006a37b3..d0bcb4660 100644
--- a/src/core/frontend/emu_window.h
+++ b/src/core/frontend/emu_window.h
@@ -166,7 +166,7 @@ private:
/**
* Clip the provided coordinates to be inside the touchscreen area.
*/
- std::tuple<unsigned, unsigned> ClipToTouchScreen(unsigned new_x, unsigned new_y);
+ std::tuple<unsigned, unsigned> ClipToTouchScreen(unsigned new_x, unsigned new_y) const;
};
} // namespace Core::Frontend
diff --git a/src/core/frontend/framebuffer_layout.cpp b/src/core/frontend/framebuffer_layout.cpp
index f8662d193..a1357179f 100644
--- a/src/core/frontend/framebuffer_layout.cpp
+++ b/src/core/frontend/framebuffer_layout.cpp
@@ -12,12 +12,12 @@ namespace Layout {
// Finds the largest size subrectangle contained in window area that is confined to the aspect ratio
template <class T>
-static MathUtil::Rectangle<T> maxRectangle(MathUtil::Rectangle<T> window_area,
- float screen_aspect_ratio) {
+static Common::Rectangle<T> MaxRectangle(Common::Rectangle<T> window_area,
+ float screen_aspect_ratio) {
float scale = std::min(static_cast<float>(window_area.GetWidth()),
window_area.GetHeight() / screen_aspect_ratio);
- return MathUtil::Rectangle<T>{0, 0, static_cast<T>(std::round(scale)),
- static_cast<T>(std::round(scale * screen_aspect_ratio))};
+ return Common::Rectangle<T>{0, 0, static_cast<T>(std::round(scale)),
+ static_cast<T>(std::round(scale * screen_aspect_ratio))};
}
FramebufferLayout DefaultFrameLayout(unsigned width, unsigned height) {
@@ -29,8 +29,8 @@ FramebufferLayout DefaultFrameLayout(unsigned width, unsigned height) {
const float emulation_aspect_ratio{static_cast<float>(ScreenUndocked::Height) /
ScreenUndocked::Width};
- MathUtil::Rectangle<unsigned> screen_window_area{0, 0, width, height};
- MathUtil::Rectangle<unsigned> screen = maxRectangle(screen_window_area, emulation_aspect_ratio);
+ Common::Rectangle<unsigned> screen_window_area{0, 0, width, height};
+ Common::Rectangle<unsigned> screen = MaxRectangle(screen_window_area, emulation_aspect_ratio);
float window_aspect_ratio = static_cast<float>(height) / width;
diff --git a/src/core/frontend/framebuffer_layout.h b/src/core/frontend/framebuffer_layout.h
index e06647794..c2c63d08c 100644
--- a/src/core/frontend/framebuffer_layout.h
+++ b/src/core/frontend/framebuffer_layout.h
@@ -16,7 +16,7 @@ struct FramebufferLayout {
unsigned width{ScreenUndocked::Width};
unsigned height{ScreenUndocked::Height};
- MathUtil::Rectangle<unsigned> screen;
+ Common::Rectangle<unsigned> screen;
/**
* Returns the ration of pixel size of the screen, compared to the native size of the undocked
diff --git a/src/core/frontend/input.h b/src/core/frontend/input.h
index 16fdcd376..7c11d7546 100644
--- a/src/core/frontend/input.h
+++ b/src/core/frontend/input.h
@@ -124,7 +124,7 @@ using AnalogDevice = InputDevice<std::tuple<float, float>>;
* Orientation is determined by right-hand rule.
* Units: deg/sec
*/
-using MotionDevice = InputDevice<std::tuple<Math::Vec3<float>, Math::Vec3<float>>>;
+using MotionDevice = InputDevice<std::tuple<Common::Vec3<float>, Common::Vec3<float>>>;
/**
* A touch device is an input device that returns a tuple of two floats and a bool. The floats are
diff --git a/src/core/hle/ipc.h b/src/core/hle/ipc.h
index ed84197b3..455d1f346 100644
--- a/src/core/hle/ipc.h
+++ b/src/core/hle/ipc.h
@@ -4,10 +4,10 @@
#pragma once
+#include "common/bit_field.h"
+#include "common/common_funcs.h"
#include "common/common_types.h"
#include "common/swap.h"
-#include "core/hle/kernel/errors.h"
-#include "core/memory.h"
namespace IPC {
diff --git a/src/core/hle/kernel/address_arbiter.cpp b/src/core/hle/kernel/address_arbiter.cpp
index a250d088d..9780a7849 100644
--- a/src/core/hle/kernel/address_arbiter.cpp
+++ b/src/core/hle/kernel/address_arbiter.cpp
@@ -9,6 +9,7 @@
#include "common/common_types.h"
#include "core/core.h"
#include "core/core_cpu.h"
+#include "core/hle/kernel/address_arbiter.h"
#include "core/hle/kernel/errors.h"
#include "core/hle/kernel/object.h"
#include "core/hle/kernel/process.h"
@@ -17,58 +18,16 @@
#include "core/hle/result.h"
#include "core/memory.h"
-namespace Kernel::AddressArbiter {
-
-// Performs actual address waiting logic.
-static ResultCode WaitForAddress(VAddr address, s64 timeout) {
- SharedPtr<Thread> current_thread = GetCurrentThread();
- current_thread->SetArbiterWaitAddress(address);
- current_thread->SetStatus(ThreadStatus::WaitArb);
- current_thread->InvalidateWakeupCallback();
-
- current_thread->WakeAfterDelay(timeout);
-
- Core::System::GetInstance().CpuCore(current_thread->GetProcessorID()).PrepareReschedule();
- return RESULT_TIMEOUT;
-}
-
-// Gets the threads waiting on an address.
-static std::vector<SharedPtr<Thread>> GetThreadsWaitingOnAddress(VAddr address) {
- const auto RetrieveWaitingThreads = [](std::size_t core_index,
- std::vector<SharedPtr<Thread>>& waiting_threads,
- VAddr arb_addr) {
- const auto& scheduler = Core::System::GetInstance().Scheduler(core_index);
- const auto& thread_list = scheduler.GetThreadList();
-
- for (const auto& thread : thread_list) {
- if (thread->GetArbiterWaitAddress() == arb_addr)
- waiting_threads.push_back(thread);
- }
- };
-
- // Retrieve all threads that are waiting for this address.
- std::vector<SharedPtr<Thread>> threads;
- RetrieveWaitingThreads(0, threads, address);
- RetrieveWaitingThreads(1, threads, address);
- RetrieveWaitingThreads(2, threads, address);
- RetrieveWaitingThreads(3, threads, address);
-
- // Sort them by priority, such that the highest priority ones come first.
- std::sort(threads.begin(), threads.end(),
- [](const SharedPtr<Thread>& lhs, const SharedPtr<Thread>& rhs) {
- return lhs->GetPriority() < rhs->GetPriority();
- });
-
- return threads;
-}
-
+namespace Kernel {
+namespace {
// Wake up num_to_wake (or all) threads in a vector.
-static void WakeThreads(std::vector<SharedPtr<Thread>>& waiting_threads, s32 num_to_wake) {
+void WakeThreads(const std::vector<SharedPtr<Thread>>& waiting_threads, s32 num_to_wake) {
// Only process up to 'target' threads, unless 'target' is <= 0, in which case process
// them all.
std::size_t last = waiting_threads.size();
- if (num_to_wake > 0)
+ if (num_to_wake > 0) {
last = num_to_wake;
+ }
// Signal the waiting threads.
for (std::size_t i = 0; i < last; i++) {
@@ -78,42 +37,41 @@ static void WakeThreads(std::vector<SharedPtr<Thread>>& waiting_threads, s32 num
waiting_threads[i]->ResumeFromWait();
}
}
+} // Anonymous namespace
-// Signals an address being waited on.
-ResultCode SignalToAddress(VAddr address, s32 num_to_wake) {
- std::vector<SharedPtr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address);
+AddressArbiter::AddressArbiter(Core::System& system) : system{system} {}
+AddressArbiter::~AddressArbiter() = default;
+ResultCode AddressArbiter::SignalToAddress(VAddr address, s32 num_to_wake) {
+ const std::vector<SharedPtr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address);
WakeThreads(waiting_threads, num_to_wake);
return RESULT_SUCCESS;
}
-// Signals an address being waited on and increments its value if equal to the value argument.
-ResultCode IncrementAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake) {
+ResultCode AddressArbiter::IncrementAndSignalToAddressIfEqual(VAddr address, s32 value,
+ s32 num_to_wake) {
// Ensure that we can write to the address.
if (!Memory::IsValidVirtualAddress(address)) {
return ERR_INVALID_ADDRESS_STATE;
}
- if (static_cast<s32>(Memory::Read32(address)) == value) {
- Memory::Write32(address, static_cast<u32>(value + 1));
- } else {
+ if (static_cast<s32>(Memory::Read32(address)) != value) {
return ERR_INVALID_STATE;
}
+ Memory::Write32(address, static_cast<u32>(value + 1));
return SignalToAddress(address, num_to_wake);
}
-// Signals an address being waited on and modifies its value based on waiting thread count if equal
-// to the value argument.
-ResultCode ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value,
- s32 num_to_wake) {
+ResultCode AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value,
+ s32 num_to_wake) {
// Ensure that we can write to the address.
if (!Memory::IsValidVirtualAddress(address)) {
return ERR_INVALID_ADDRESS_STATE;
}
// Get threads waiting on the address.
- std::vector<SharedPtr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address);
+ const std::vector<SharedPtr<Thread>> waiting_threads = GetThreadsWaitingOnAddress(address);
// Determine the modified value depending on the waiting count.
s32 updated_value;
@@ -125,31 +83,31 @@ ResultCode ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 valu
updated_value = value;
}
- if (static_cast<s32>(Memory::Read32(address)) == value) {
- Memory::Write32(address, static_cast<u32>(updated_value));
- } else {
+ if (static_cast<s32>(Memory::Read32(address)) != value) {
return ERR_INVALID_STATE;
}
+ Memory::Write32(address, static_cast<u32>(updated_value));
WakeThreads(waiting_threads, num_to_wake);
return RESULT_SUCCESS;
}
-// Waits on an address if the value passed is less than the argument value, optionally decrementing.
-ResultCode WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout, bool should_decrement) {
+ResultCode AddressArbiter::WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout,
+ bool should_decrement) {
// Ensure that we can read the address.
if (!Memory::IsValidVirtualAddress(address)) {
return ERR_INVALID_ADDRESS_STATE;
}
- s32 cur_value = static_cast<s32>(Memory::Read32(address));
- if (cur_value < value) {
- if (should_decrement) {
- Memory::Write32(address, static_cast<u32>(cur_value - 1));
- }
- } else {
+ const s32 cur_value = static_cast<s32>(Memory::Read32(address));
+ if (cur_value >= value) {
return ERR_INVALID_STATE;
}
+
+ if (should_decrement) {
+ Memory::Write32(address, static_cast<u32>(cur_value - 1));
+ }
+
// Short-circuit without rescheduling, if timeout is zero.
if (timeout == 0) {
return RESULT_TIMEOUT;
@@ -158,8 +116,7 @@ ResultCode WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout, bool
return WaitForAddress(address, timeout);
}
-// Waits on an address if the value passed is equal to the argument value.
-ResultCode WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout) {
+ResultCode AddressArbiter::WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout) {
// Ensure that we can read the address.
if (!Memory::IsValidVirtualAddress(address)) {
return ERR_INVALID_ADDRESS_STATE;
@@ -175,4 +132,46 @@ ResultCode WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout) {
return WaitForAddress(address, timeout);
}
-} // namespace Kernel::AddressArbiter
+
+ResultCode AddressArbiter::WaitForAddress(VAddr address, s64 timeout) {
+ SharedPtr<Thread> current_thread = system.CurrentScheduler().GetCurrentThread();
+ current_thread->SetArbiterWaitAddress(address);
+ current_thread->SetStatus(ThreadStatus::WaitArb);
+ current_thread->InvalidateWakeupCallback();
+
+ current_thread->WakeAfterDelay(timeout);
+
+ system.CpuCore(current_thread->GetProcessorID()).PrepareReschedule();
+ return RESULT_TIMEOUT;
+}
+
+std::vector<SharedPtr<Thread>> AddressArbiter::GetThreadsWaitingOnAddress(VAddr address) const {
+ const auto RetrieveWaitingThreads = [this](std::size_t core_index,
+ std::vector<SharedPtr<Thread>>& waiting_threads,
+ VAddr arb_addr) {
+ const auto& scheduler = system.Scheduler(core_index);
+ const auto& thread_list = scheduler.GetThreadList();
+
+ for (const auto& thread : thread_list) {
+ if (thread->GetArbiterWaitAddress() == arb_addr) {
+ waiting_threads.push_back(thread);
+ }
+ }
+ };
+
+ // Retrieve all threads that are waiting for this address.
+ std::vector<SharedPtr<Thread>> threads;
+ RetrieveWaitingThreads(0, threads, address);
+ RetrieveWaitingThreads(1, threads, address);
+ RetrieveWaitingThreads(2, threads, address);
+ RetrieveWaitingThreads(3, threads, address);
+
+ // Sort them by priority, such that the highest priority ones come first.
+ std::sort(threads.begin(), threads.end(),
+ [](const SharedPtr<Thread>& lhs, const SharedPtr<Thread>& rhs) {
+ return lhs->GetPriority() < rhs->GetPriority();
+ });
+
+ return threads;
+}
+} // namespace Kernel
diff --git a/src/core/hle/kernel/address_arbiter.h b/src/core/hle/kernel/address_arbiter.h
index b58f21bec..e0c36f2e3 100644
--- a/src/core/hle/kernel/address_arbiter.h
+++ b/src/core/hle/kernel/address_arbiter.h
@@ -5,28 +5,68 @@
#pragma once
#include "common/common_types.h"
+#include "core/hle/kernel/address_arbiter.h"
union ResultCode;
-namespace Kernel::AddressArbiter {
+namespace Core {
+class System;
+}
-enum class ArbitrationType {
- WaitIfLessThan = 0,
- DecrementAndWaitIfLessThan = 1,
- WaitIfEqual = 2,
-};
+namespace Kernel {
-enum class SignalType {
- Signal = 0,
- IncrementAndSignalIfEqual = 1,
- ModifyByWaitingCountAndSignalIfEqual = 2,
-};
+class Thread;
+
+class AddressArbiter {
+public:
+ enum class ArbitrationType {
+ WaitIfLessThan = 0,
+ DecrementAndWaitIfLessThan = 1,
+ WaitIfEqual = 2,
+ };
+
+ enum class SignalType {
+ Signal = 0,
+ IncrementAndSignalIfEqual = 1,
+ ModifyByWaitingCountAndSignalIfEqual = 2,
+ };
+
+ explicit AddressArbiter(Core::System& system);
+ ~AddressArbiter();
+
+ AddressArbiter(const AddressArbiter&) = delete;
+ AddressArbiter& operator=(const AddressArbiter&) = delete;
+
+ AddressArbiter(AddressArbiter&&) = default;
+ AddressArbiter& operator=(AddressArbiter&&) = delete;
-ResultCode SignalToAddress(VAddr address, s32 num_to_wake);
-ResultCode IncrementAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake);
-ResultCode ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake);
+ /// Signals an address being waited on.
+ ResultCode SignalToAddress(VAddr address, s32 num_to_wake);
-ResultCode WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout, bool should_decrement);
-ResultCode WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout);
+ /// Signals an address being waited on and increments its value if equal to the value argument.
+ ResultCode IncrementAndSignalToAddressIfEqual(VAddr address, s32 value, s32 num_to_wake);
+
+ /// Signals an address being waited on and modifies its value based on waiting thread count if
+ /// equal to the value argument.
+ ResultCode ModifyByWaitingCountAndSignalToAddressIfEqual(VAddr address, s32 value,
+ s32 num_to_wake);
+
+ /// Waits on an address if the value passed is less than the argument value,
+ /// optionally decrementing.
+ ResultCode WaitForAddressIfLessThan(VAddr address, s32 value, s64 timeout,
+ bool should_decrement);
+
+ /// Waits on an address if the value passed is equal to the argument value.
+ ResultCode WaitForAddressIfEqual(VAddr address, s32 value, s64 timeout);
+
+private:
+ // Waits on the given address with a timeout in nanoseconds
+ ResultCode WaitForAddress(VAddr address, s64 timeout);
+
+ // Gets the threads waiting on an address.
+ std::vector<SharedPtr<Thread>> GetThreadsWaitingOnAddress(VAddr address) const;
+
+ Core::System& system;
+};
-} // namespace Kernel::AddressArbiter
+} // namespace Kernel
diff --git a/src/core/hle/kernel/errors.h b/src/core/hle/kernel/errors.h
index d17eb0cb6..8097b3863 100644
--- a/src/core/hle/kernel/errors.h
+++ b/src/core/hle/kernel/errors.h
@@ -14,6 +14,7 @@ constexpr ResultCode ERR_MAX_CONNECTIONS_REACHED{ErrorModule::Kernel, 7};
constexpr ResultCode ERR_INVALID_CAPABILITY_DESCRIPTOR{ErrorModule::Kernel, 14};
constexpr ResultCode ERR_INVALID_SIZE{ErrorModule::Kernel, 101};
constexpr ResultCode ERR_INVALID_ADDRESS{ErrorModule::Kernel, 102};
+constexpr ResultCode ERR_OUT_OF_MEMORY{ErrorModule::Kernel, 104};
constexpr ResultCode ERR_HANDLE_TABLE_FULL{ErrorModule::Kernel, 105};
constexpr ResultCode ERR_INVALID_ADDRESS_STATE{ErrorModule::Kernel, 106};
constexpr ResultCode ERR_INVALID_MEMORY_PERMISSIONS{ErrorModule::Kernel, 108};
diff --git a/src/core/hle/kernel/handle_table.cpp b/src/core/hle/kernel/handle_table.cpp
index c8acde5b1..bdfaa977f 100644
--- a/src/core/hle/kernel/handle_table.cpp
+++ b/src/core/hle/kernel/handle_table.cpp
@@ -14,32 +14,47 @@
namespace Kernel {
namespace {
constexpr u16 GetSlot(Handle handle) {
- return handle >> 15;
+ return static_cast<u16>(handle >> 15);
}
constexpr u16 GetGeneration(Handle handle) {
- return handle & 0x7FFF;
+ return static_cast<u16>(handle & 0x7FFF);
}
} // Anonymous namespace
HandleTable::HandleTable() {
- next_generation = 1;
Clear();
}
HandleTable::~HandleTable() = default;
+ResultCode HandleTable::SetSize(s32 handle_table_size) {
+ if (static_cast<u32>(handle_table_size) > MAX_COUNT) {
+ return ERR_OUT_OF_MEMORY;
+ }
+
+ // Values less than or equal to zero indicate to use the maximum allowable
+ // size for the handle table in the actual kernel, so we ignore the given
+ // value in that case, since we assume this by default unless this function
+ // is called.
+ if (handle_table_size > 0) {
+ table_size = static_cast<u16>(handle_table_size);
+ }
+
+ return RESULT_SUCCESS;
+}
+
ResultVal<Handle> HandleTable::Create(SharedPtr<Object> obj) {
DEBUG_ASSERT(obj != nullptr);
- u16 slot = next_free_slot;
- if (slot >= generations.size()) {
+ const u16 slot = next_free_slot;
+ if (slot >= table_size) {
LOG_ERROR(Kernel, "Unable to allocate Handle, too many slots in use.");
return ERR_HANDLE_TABLE_FULL;
}
next_free_slot = generations[slot];
- u16 generation = next_generation++;
+ const u16 generation = next_generation++;
// Overflow count so it fits in the 15 bits dedicated to the generation in the handle.
// Horizon OS uses zero to represent an invalid handle, so skip to 1.
@@ -64,10 +79,11 @@ ResultVal<Handle> HandleTable::Duplicate(Handle handle) {
}
ResultCode HandleTable::Close(Handle handle) {
- if (!IsValid(handle))
+ if (!IsValid(handle)) {
return ERR_INVALID_HANDLE;
+ }
- u16 slot = GetSlot(handle);
+ const u16 slot = GetSlot(handle);
objects[slot] = nullptr;
@@ -77,10 +93,10 @@ ResultCode HandleTable::Close(Handle handle) {
}
bool HandleTable::IsValid(Handle handle) const {
- std::size_t slot = GetSlot(handle);
- u16 generation = GetGeneration(handle);
+ const std::size_t slot = GetSlot(handle);
+ const u16 generation = GetGeneration(handle);
- return slot < MAX_COUNT && objects[slot] != nullptr && generations[slot] == generation;
+ return slot < table_size && objects[slot] != nullptr && generations[slot] == generation;
}
SharedPtr<Object> HandleTable::GetGeneric(Handle handle) const {
@@ -97,7 +113,7 @@ SharedPtr<Object> HandleTable::GetGeneric(Handle handle) const {
}
void HandleTable::Clear() {
- for (u16 i = 0; i < MAX_COUNT; ++i) {
+ for (u16 i = 0; i < table_size; ++i) {
generations[i] = i + 1;
objects[i] = nullptr;
}
diff --git a/src/core/hle/kernel/handle_table.h b/src/core/hle/kernel/handle_table.h
index 89a3bc740..44901391b 100644
--- a/src/core/hle/kernel/handle_table.h
+++ b/src/core/hle/kernel/handle_table.h
@@ -50,6 +50,20 @@ public:
~HandleTable();
/**
+ * Sets the number of handles that may be in use at one time
+ * for this handle table.
+ *
+ * @param handle_table_size The desired size to limit the handle table to.
+ *
+ * @returns an error code indicating if initialization was successful.
+ * If initialization was not successful, then ERR_OUT_OF_MEMORY
+ * will be returned.
+ *
+ * @pre handle_table_size must be within the range [0, 1024]
+ */
+ ResultCode SetSize(s32 handle_table_size);
+
+ /**
* Allocates a handle for the given object.
* @return The created Handle or one of the following errors:
* - `ERR_HANDLE_TABLE_FULL`: the maximum number of handles has been exceeded.
@@ -104,13 +118,20 @@ private:
std::array<u16, MAX_COUNT> generations;
/**
+ * The limited size of the handle table. This can be specified by process
+ * capabilities in order to restrict the overall number of handles that
+ * can be created in a process instance
+ */
+ u16 table_size = static_cast<u16>(MAX_COUNT);
+
+ /**
* Global counter of the number of created handles. Stored in `generations` when a handle is
* created, and wraps around to 1 when it hits 0x8000.
*/
- u16 next_generation;
+ u16 next_generation = 1;
/// Head of the free slots linked list.
- u16 next_free_slot;
+ u16 next_free_slot = 0;
};
} // namespace Kernel
diff --git a/src/core/hle/kernel/hle_ipc.h b/src/core/hle/kernel/hle_ipc.h
index cb1c5aff3..0107acea4 100644
--- a/src/core/hle/kernel/hle_ipc.h
+++ b/src/core/hle/kernel/hle_ipc.h
@@ -15,6 +15,8 @@
#include "core/hle/ipc.h"
#include "core/hle/kernel/object.h"
+union ResultCode;
+
namespace Service {
class ServiceFrameworkBase;
}
@@ -208,14 +210,12 @@ public:
template <typename T>
SharedPtr<T> GetCopyObject(std::size_t index) {
- ASSERT(index < copy_objects.size());
- return DynamicObjectCast<T>(copy_objects[index]);
+ return DynamicObjectCast<T>(copy_objects.at(index));
}
template <typename T>
SharedPtr<T> GetMoveObject(std::size_t index) {
- ASSERT(index < move_objects.size());
- return DynamicObjectCast<T>(move_objects[index]);
+ return DynamicObjectCast<T>(move_objects.at(index));
}
void AddMoveObject(SharedPtr<Object> object) {
@@ -232,7 +232,7 @@ public:
template <typename T>
std::shared_ptr<T> GetDomainRequestHandler(std::size_t index) const {
- return std::static_pointer_cast<T>(domain_request_handlers[index]);
+ return std::static_pointer_cast<T>(domain_request_handlers.at(index));
}
void SetDomainRequestHandlers(
diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp
index dd749eed4..04ea9349e 100644
--- a/src/core/hle/kernel/kernel.cpp
+++ b/src/core/hle/kernel/kernel.cpp
@@ -12,6 +12,7 @@
#include "core/core.h"
#include "core/core_timing.h"
+#include "core/hle/kernel/address_arbiter.h"
#include "core/hle/kernel/client_port.h"
#include "core/hle/kernel/handle_table.h"
#include "core/hle/kernel/kernel.h"
@@ -86,11 +87,13 @@ static void ThreadWakeupCallback(u64 thread_handle, [[maybe_unused]] int cycles_
}
struct KernelCore::Impl {
- void Initialize(KernelCore& kernel, Core::Timing::CoreTiming& core_timing) {
+ explicit Impl(Core::System& system) : address_arbiter{system}, system{system} {}
+
+ void Initialize(KernelCore& kernel) {
Shutdown();
InitializeSystemResourceLimit(kernel);
- InitializeThreads(core_timing);
+ InitializeThreads();
}
void Shutdown() {
@@ -122,9 +125,9 @@ struct KernelCore::Impl {
ASSERT(system_resource_limit->SetLimitValue(ResourceType::Sessions, 900).IsSuccess());
}
- void InitializeThreads(Core::Timing::CoreTiming& core_timing) {
+ void InitializeThreads() {
thread_wakeup_event_type =
- core_timing.RegisterEvent("ThreadWakeupCallback", ThreadWakeupCallback);
+ system.CoreTiming().RegisterEvent("ThreadWakeupCallback", ThreadWakeupCallback);
}
std::atomic<u32> next_object_id{0};
@@ -135,6 +138,8 @@ struct KernelCore::Impl {
std::vector<SharedPtr<Process>> process_list;
Process* current_process = nullptr;
+ Kernel::AddressArbiter address_arbiter;
+
SharedPtr<ResourceLimit> system_resource_limit;
Core::Timing::EventType* thread_wakeup_event_type = nullptr;
@@ -145,15 +150,18 @@ struct KernelCore::Impl {
/// Map of named ports managed by the kernel, which can be retrieved using
/// the ConnectToPort SVC.
NamedPortTable named_ports;
+
+ // System context
+ Core::System& system;
};
-KernelCore::KernelCore() : impl{std::make_unique<Impl>()} {}
+KernelCore::KernelCore(Core::System& system) : impl{std::make_unique<Impl>(system)} {}
KernelCore::~KernelCore() {
Shutdown();
}
-void KernelCore::Initialize(Core::Timing::CoreTiming& core_timing) {
- impl->Initialize(*this, core_timing);
+void KernelCore::Initialize() {
+ impl->Initialize(*this);
}
void KernelCore::Shutdown() {
@@ -184,6 +192,14 @@ const Process* KernelCore::CurrentProcess() const {
return impl->current_process;
}
+AddressArbiter& KernelCore::AddressArbiter() {
+ return impl->address_arbiter;
+}
+
+const AddressArbiter& KernelCore::AddressArbiter() const {
+ return impl->address_arbiter;
+}
+
void KernelCore::AddNamedPort(std::string name, SharedPtr<ClientPort> port) {
impl->named_ports.emplace(std::move(name), std::move(port));
}
diff --git a/src/core/hle/kernel/kernel.h b/src/core/hle/kernel/kernel.h
index 154bced42..4d292aca9 100644
--- a/src/core/hle/kernel/kernel.h
+++ b/src/core/hle/kernel/kernel.h
@@ -11,6 +11,10 @@
template <typename T>
class ResultVal;
+namespace Core {
+class System;
+}
+
namespace Core::Timing {
class CoreTiming;
struct EventType;
@@ -18,6 +22,7 @@ struct EventType;
namespace Kernel {
+class AddressArbiter;
class ClientPort;
class HandleTable;
class Process;
@@ -30,7 +35,14 @@ private:
using NamedPortTable = std::unordered_map<std::string, SharedPtr<ClientPort>>;
public:
- KernelCore();
+ /// Constructs an instance of the kernel using the given System
+ /// instance as a context for any necessary system-related state,
+ /// such as threads, CPU core state, etc.
+ ///
+ /// @post After execution of the constructor, the provided System
+ /// object *must* outlive the kernel instance itself.
+ ///
+ explicit KernelCore(Core::System& system);
~KernelCore();
KernelCore(const KernelCore&) = delete;
@@ -40,11 +52,7 @@ public:
KernelCore& operator=(KernelCore&&) = delete;
/// Resets the kernel to a clean slate for use.
- ///
- /// @param core_timing CoreTiming instance used to create any necessary
- /// kernel-specific callback events.
- ///
- void Initialize(Core::Timing::CoreTiming& core_timing);
+ void Initialize();
/// Clears all resources in use by the kernel instance.
void Shutdown();
@@ -67,6 +75,12 @@ public:
/// Retrieves a const pointer to the current process.
const Process* CurrentProcess() const;
+ /// Provides a reference to the kernel's address arbiter.
+ Kernel::AddressArbiter& AddressArbiter();
+
+ /// Provides a const reference to the kernel's address arbiter.
+ const Kernel::AddressArbiter& AddressArbiter() const;
+
/// Adds a port to the named port table
void AddNamedPort(std::string name, SharedPtr<ClientPort> port);
diff --git a/src/core/hle/kernel/process.cpp b/src/core/hle/kernel/process.cpp
index c5aa19afa..8009150e0 100644
--- a/src/core/hle/kernel/process.cpp
+++ b/src/core/hle/kernel/process.cpp
@@ -99,7 +99,13 @@ ResultCode Process::LoadFromMetadata(const FileSys::ProgramMetadata& metadata) {
vm_manager.Reset(metadata.GetAddressSpaceType());
const auto& caps = metadata.GetKernelCapabilities();
- return capabilities.InitializeForUserProcess(caps.data(), caps.size(), vm_manager);
+ const auto capability_init_result =
+ capabilities.InitializeForUserProcess(caps.data(), caps.size(), vm_manager);
+ if (capability_init_result.IsError()) {
+ return capability_init_result;
+ }
+
+ return handle_table.SetSize(capabilities.GetHandleTableSize());
}
void Process::Run(VAddr entry_point, s32 main_thread_priority, u32 stack_size) {
diff --git a/src/core/hle/kernel/process_capability.cpp b/src/core/hle/kernel/process_capability.cpp
index 3a2164b25..583e35b79 100644
--- a/src/core/hle/kernel/process_capability.cpp
+++ b/src/core/hle/kernel/process_capability.cpp
@@ -96,7 +96,7 @@ void ProcessCapabilities::InitializeForMetadatalessProcess() {
interrupt_capabilities.set();
// Allow using the maximum possible amount of handles
- handle_table_size = static_cast<u32>(HandleTable::MAX_COUNT);
+ handle_table_size = static_cast<s32>(HandleTable::MAX_COUNT);
// Allow all debugging capabilities.
is_debuggable = true;
@@ -337,7 +337,7 @@ ResultCode ProcessCapabilities::HandleHandleTableFlags(u32 flags) {
return ERR_RESERVED_VALUE;
}
- handle_table_size = (flags >> 16) & 0x3FF;
+ handle_table_size = static_cast<s32>((flags >> 16) & 0x3FF);
return RESULT_SUCCESS;
}
diff --git a/src/core/hle/kernel/process_capability.h b/src/core/hle/kernel/process_capability.h
index fbc8812a3..5cdd80747 100644
--- a/src/core/hle/kernel/process_capability.h
+++ b/src/core/hle/kernel/process_capability.h
@@ -156,7 +156,7 @@ public:
}
/// Gets the number of total allowable handles for the process' handle table.
- u32 GetHandleTableSize() const {
+ s32 GetHandleTableSize() const {
return handle_table_size;
}
@@ -252,7 +252,7 @@ private:
u64 core_mask = 0;
u64 priority_mask = 0;
- u32 handle_table_size = 0;
+ s32 handle_table_size = 0;
u32 kernel_version = 0;
ProgramType program_type = ProgramType::SysModule;
diff --git a/src/core/hle/kernel/svc.cpp b/src/core/hle/kernel/svc.cpp
index c5d399bab..7f5c0cc86 100644
--- a/src/core/hle/kernel/svc.cpp
+++ b/src/core/hle/kernel/svc.cpp
@@ -20,6 +20,7 @@
#include "core/hle/kernel/address_arbiter.h"
#include "core/hle/kernel/client_port.h"
#include "core/hle/kernel/client_session.h"
+#include "core/hle/kernel/errors.h"
#include "core/hle/kernel/handle_table.h"
#include "core/hle/kernel/kernel.h"
#include "core/hle/kernel/mutex.h"
@@ -47,23 +48,6 @@ constexpr bool IsValidAddressRange(VAddr address, u64 size) {
return address + size > address;
}
-// Checks if a given address range lies within a larger address range.
-constexpr bool IsInsideAddressRange(VAddr address, u64 size, VAddr address_range_begin,
- VAddr address_range_end) {
- const VAddr end_address = address + size - 1;
- return address_range_begin <= address && end_address <= address_range_end - 1;
-}
-
-bool IsInsideAddressSpace(const VMManager& vm, VAddr address, u64 size) {
- return IsInsideAddressRange(address, size, vm.GetAddressSpaceBaseAddress(),
- vm.GetAddressSpaceEndAddress());
-}
-
-bool IsInsideNewMapRegion(const VMManager& vm, VAddr address, u64 size) {
- return IsInsideAddressRange(address, size, vm.GetNewMapRegionBaseAddress(),
- vm.GetNewMapRegionEndAddress());
-}
-
// 8 GiB
constexpr u64 MAIN_MEMORY_SIZE = 0x200000000;
@@ -105,14 +89,14 @@ ResultCode MapUnmapMemorySanityChecks(const VMManager& vm_manager, VAddr dst_add
return ERR_INVALID_ADDRESS_STATE;
}
- if (!IsInsideAddressSpace(vm_manager, src_addr, size)) {
+ if (!vm_manager.IsWithinAddressSpace(src_addr, size)) {
LOG_ERROR(Kernel_SVC,
"Source is not within the address space, addr=0x{:016X}, size=0x{:016X}",
src_addr, size);
return ERR_INVALID_ADDRESS_STATE;
}
- if (!IsInsideNewMapRegion(vm_manager, dst_addr, size)) {
+ if (!vm_manager.IsWithinNewMapRegion(dst_addr, size)) {
LOG_ERROR(Kernel_SVC,
"Destination is not within the new map region, addr=0x{:016X}, size=0x{:016X}",
dst_addr, size);
@@ -238,7 +222,7 @@ static ResultCode SetMemoryPermission(VAddr addr, u64 size, u32 prot) {
auto* const current_process = Core::CurrentProcess();
auto& vm_manager = current_process->VMManager();
- if (!IsInsideAddressSpace(vm_manager, addr, size)) {
+ if (!vm_manager.IsWithinAddressSpace(addr, size)) {
LOG_ERROR(Kernel_SVC,
"Source is not within the address space, addr=0x{:016X}, size=0x{:016X}", addr,
size);
@@ -299,7 +283,7 @@ static ResultCode SetMemoryAttribute(VAddr address, u64 size, u32 mask, u32 attr
}
auto& vm_manager = Core::CurrentProcess()->VMManager();
- if (!IsInsideAddressSpace(vm_manager, address, size)) {
+ if (!vm_manager.IsWithinAddressSpace(address, size)) {
LOG_ERROR(Kernel_SVC,
"Given address (0x{:016X}) is outside the bounds of the address space.", address);
return ERR_INVALID_ADDRESS_STATE;
@@ -1495,13 +1479,14 @@ static ResultCode WaitForAddress(VAddr address, u32 type, s32 value, s64 timeout
return ERR_INVALID_ADDRESS;
}
+ auto& address_arbiter = Core::System::GetInstance().Kernel().AddressArbiter();
switch (static_cast<AddressArbiter::ArbitrationType>(type)) {
case AddressArbiter::ArbitrationType::WaitIfLessThan:
- return AddressArbiter::WaitForAddressIfLessThan(address, value, timeout, false);
+ return address_arbiter.WaitForAddressIfLessThan(address, value, timeout, false);
case AddressArbiter::ArbitrationType::DecrementAndWaitIfLessThan:
- return AddressArbiter::WaitForAddressIfLessThan(address, value, timeout, true);
+ return address_arbiter.WaitForAddressIfLessThan(address, value, timeout, true);
case AddressArbiter::ArbitrationType::WaitIfEqual:
- return AddressArbiter::WaitForAddressIfEqual(address, value, timeout);
+ return address_arbiter.WaitForAddressIfEqual(address, value, timeout);
default:
LOG_ERROR(Kernel_SVC,
"Invalid arbitration type, expected WaitIfLessThan, DecrementAndWaitIfLessThan "
@@ -1526,13 +1511,14 @@ static ResultCode SignalToAddress(VAddr address, u32 type, s32 value, s32 num_to
return ERR_INVALID_ADDRESS;
}
+ auto& address_arbiter = Core::System::GetInstance().Kernel().AddressArbiter();
switch (static_cast<AddressArbiter::SignalType>(type)) {
case AddressArbiter::SignalType::Signal:
- return AddressArbiter::SignalToAddress(address, num_to_wake);
+ return address_arbiter.SignalToAddress(address, num_to_wake);
case AddressArbiter::SignalType::IncrementAndSignalIfEqual:
- return AddressArbiter::IncrementAndSignalToAddressIfEqual(address, value, num_to_wake);
+ return address_arbiter.IncrementAndSignalToAddressIfEqual(address, value, num_to_wake);
case AddressArbiter::SignalType::ModifyByWaitingCountAndSignalIfEqual:
- return AddressArbiter::ModifyByWaitingCountAndSignalToAddressIfEqual(address, value,
+ return address_arbiter.ModifyByWaitingCountAndSignalToAddressIfEqual(address, value,
num_to_wake);
default:
LOG_ERROR(Kernel_SVC,
diff --git a/src/core/hle/kernel/thread.cpp b/src/core/hle/kernel/thread.cpp
index 6661e2130..eb54d6651 100644
--- a/src/core/hle/kernel/thread.cpp
+++ b/src/core/hle/kernel/thread.cpp
@@ -184,8 +184,6 @@ ResultVal<SharedPtr<Thread>> Thread::Create(KernelCore& kernel, std::string name
return ERR_INVALID_PROCESSOR_ID;
}
- // TODO(yuriks): Other checks, returning 0xD9001BEA
-
if (!Memory::IsValidVirtualAddress(owner_process, entry_point)) {
LOG_ERROR(Kernel_SVC, "(name={}): invalid entry {:016X}", name, entry_point);
// TODO (bunnei): Find the correct error code to use here
diff --git a/src/core/hle/kernel/vm_manager.cpp b/src/core/hle/kernel/vm_manager.cpp
index 10ad94aa6..05c59af34 100644
--- a/src/core/hle/kernel/vm_manager.cpp
+++ b/src/core/hle/kernel/vm_manager.cpp
@@ -17,8 +17,8 @@
#include "core/memory_setup.h"
namespace Kernel {
-
-static const char* GetMemoryStateName(MemoryState state) {
+namespace {
+const char* GetMemoryStateName(MemoryState state) {
static constexpr const char* names[] = {
"Unmapped", "Io",
"Normal", "CodeStatic",
@@ -35,6 +35,14 @@ static const char* GetMemoryStateName(MemoryState state) {
return names[ToSvcMemoryState(state)];
}
+// Checks if a given address range lies within a larger address range.
+constexpr bool IsInsideAddressRange(VAddr address, u64 size, VAddr address_range_begin,
+ VAddr address_range_end) {
+ const VAddr end_address = address + size - 1;
+ return address_range_begin <= address && end_address <= address_range_end - 1;
+}
+} // Anonymous namespace
+
bool VirtualMemoryArea::CanBeMergedWith(const VirtualMemoryArea& next) const {
ASSERT(base + size == next.base);
if (permissions != next.permissions || state != next.state || attribute != next.attribute ||
@@ -249,8 +257,7 @@ ResultCode VMManager::ReprotectRange(VAddr target, u64 size, VMAPermission new_p
}
ResultVal<VAddr> VMManager::HeapAllocate(VAddr target, u64 size, VMAPermission perms) {
- if (target < GetHeapRegionBaseAddress() || target + size > GetHeapRegionEndAddress() ||
- target + size < target) {
+ if (!IsWithinHeapRegion(target, size)) {
return ERR_INVALID_ADDRESS;
}
@@ -285,8 +292,7 @@ ResultVal<VAddr> VMManager::HeapAllocate(VAddr target, u64 size, VMAPermission p
}
ResultCode VMManager::HeapFree(VAddr target, u64 size) {
- if (target < GetHeapRegionBaseAddress() || target + size > GetHeapRegionEndAddress() ||
- target + size < target) {
+ if (!IsWithinHeapRegion(target, size)) {
return ERR_INVALID_ADDRESS;
}
@@ -706,6 +712,11 @@ u64 VMManager::GetAddressSpaceWidth() const {
return address_space_width;
}
+bool VMManager::IsWithinAddressSpace(VAddr address, u64 size) const {
+ return IsInsideAddressRange(address, size, GetAddressSpaceBaseAddress(),
+ GetAddressSpaceEndAddress());
+}
+
VAddr VMManager::GetASLRRegionBaseAddress() const {
return aslr_region_base;
}
@@ -750,6 +761,11 @@ u64 VMManager::GetCodeRegionSize() const {
return code_region_end - code_region_base;
}
+bool VMManager::IsWithinCodeRegion(VAddr address, u64 size) const {
+ return IsInsideAddressRange(address, size, GetCodeRegionBaseAddress(),
+ GetCodeRegionEndAddress());
+}
+
VAddr VMManager::GetHeapRegionBaseAddress() const {
return heap_region_base;
}
@@ -762,6 +778,11 @@ u64 VMManager::GetHeapRegionSize() const {
return heap_region_end - heap_region_base;
}
+bool VMManager::IsWithinHeapRegion(VAddr address, u64 size) const {
+ return IsInsideAddressRange(address, size, GetHeapRegionBaseAddress(),
+ GetHeapRegionEndAddress());
+}
+
VAddr VMManager::GetMapRegionBaseAddress() const {
return map_region_base;
}
@@ -774,6 +795,10 @@ u64 VMManager::GetMapRegionSize() const {
return map_region_end - map_region_base;
}
+bool VMManager::IsWithinMapRegion(VAddr address, u64 size) const {
+ return IsInsideAddressRange(address, size, GetMapRegionBaseAddress(), GetMapRegionEndAddress());
+}
+
VAddr VMManager::GetNewMapRegionBaseAddress() const {
return new_map_region_base;
}
@@ -786,6 +811,11 @@ u64 VMManager::GetNewMapRegionSize() const {
return new_map_region_end - new_map_region_base;
}
+bool VMManager::IsWithinNewMapRegion(VAddr address, u64 size) const {
+ return IsInsideAddressRange(address, size, GetNewMapRegionBaseAddress(),
+ GetNewMapRegionEndAddress());
+}
+
VAddr VMManager::GetTLSIORegionBaseAddress() const {
return tls_io_region_base;
}
@@ -798,4 +828,9 @@ u64 VMManager::GetTLSIORegionSize() const {
return tls_io_region_end - tls_io_region_base;
}
+bool VMManager::IsWithinTLSIORegion(VAddr address, u64 size) const {
+ return IsInsideAddressRange(address, size, GetTLSIORegionBaseAddress(),
+ GetTLSIORegionEndAddress());
+}
+
} // namespace Kernel
diff --git a/src/core/hle/kernel/vm_manager.h b/src/core/hle/kernel/vm_manager.h
index 6091533bc..88e0b3c02 100644
--- a/src/core/hle/kernel/vm_manager.h
+++ b/src/core/hle/kernel/vm_manager.h
@@ -432,18 +432,21 @@ public:
/// Gets the address space width in bits.
u64 GetAddressSpaceWidth() const;
+ /// Determines whether or not the given address range lies within the address space.
+ bool IsWithinAddressSpace(VAddr address, u64 size) const;
+
/// Gets the base address of the ASLR region.
VAddr GetASLRRegionBaseAddress() const;
/// Gets the end address of the ASLR region.
VAddr GetASLRRegionEndAddress() const;
- /// Determines whether or not the specified address range is within the ASLR region.
- bool IsWithinASLRRegion(VAddr address, u64 size) const;
-
/// Gets the size of the ASLR region
u64 GetASLRRegionSize() const;
+ /// Determines whether or not the specified address range is within the ASLR region.
+ bool IsWithinASLRRegion(VAddr address, u64 size) const;
+
/// Gets the base address of the code region.
VAddr GetCodeRegionBaseAddress() const;
@@ -453,6 +456,9 @@ public:
/// Gets the total size of the code region in bytes.
u64 GetCodeRegionSize() const;
+ /// Determines whether or not the specified range is within the code region.
+ bool IsWithinCodeRegion(VAddr address, u64 size) const;
+
/// Gets the base address of the heap region.
VAddr GetHeapRegionBaseAddress() const;
@@ -462,6 +468,9 @@ public:
/// Gets the total size of the heap region in bytes.
u64 GetHeapRegionSize() const;
+ /// Determines whether or not the specified range is within the heap region.
+ bool IsWithinHeapRegion(VAddr address, u64 size) const;
+
/// Gets the base address of the map region.
VAddr GetMapRegionBaseAddress() const;
@@ -471,6 +480,9 @@ public:
/// Gets the total size of the map region in bytes.
u64 GetMapRegionSize() const;
+ /// Determines whether or not the specified range is within the map region.
+ bool IsWithinMapRegion(VAddr address, u64 size) const;
+
/// Gets the base address of the new map region.
VAddr GetNewMapRegionBaseAddress() const;
@@ -480,6 +492,9 @@ public:
/// Gets the total size of the new map region in bytes.
u64 GetNewMapRegionSize() const;
+ /// Determines whether or not the given address range is within the new map region
+ bool IsWithinNewMapRegion(VAddr address, u64 size) const;
+
/// Gets the base address of the TLS IO region.
VAddr GetTLSIORegionBaseAddress() const;
@@ -489,6 +504,9 @@ public:
/// Gets the total size of the TLS IO region in bytes.
u64 GetTLSIORegionSize() const;
+ /// Determines if the given address range is within the TLS IO region.
+ bool IsWithinTLSIORegion(VAddr address, u64 size) const;
+
/// Each VMManager has its own page table, which is set as the main one when the owning process
/// is scheduled.
Memory::PageTable page_table;
diff --git a/src/core/hle/result.h b/src/core/hle/result.h
index bfb77cc31..1ed144481 100644
--- a/src/core/hle/result.h
+++ b/src/core/hle/result.h
@@ -8,7 +8,6 @@
#include <utility>
#include "common/assert.h"
#include "common/bit_field.h"
-#include "common/common_funcs.h"
#include "common/common_types.h"
// All the constants in this file come from http://switchbrew.org/index.php?title=Error_codes
diff --git a/src/core/hle/service/am/applets/software_keyboard.cpp b/src/core/hle/service/am/applets/software_keyboard.cpp
index f255f74b5..8c5bd6059 100644
--- a/src/core/hle/service/am/applets/software_keyboard.cpp
+++ b/src/core/hle/service/am/applets/software_keyboard.cpp
@@ -7,6 +7,7 @@
#include "common/string_util.h"
#include "core/core.h"
#include "core/frontend/applets/software_keyboard.h"
+#include "core/hle/result.h"
#include "core/hle/service/am/am.h"
#include "core/hle/service/am/applets/software_keyboard.h"
diff --git a/src/core/hle/service/am/applets/software_keyboard.h b/src/core/hle/service/am/applets/software_keyboard.h
index efd5753a1..b93a30d28 100644
--- a/src/core/hle/service/am/applets/software_keyboard.h
+++ b/src/core/hle/service/am/applets/software_keyboard.h
@@ -9,10 +9,13 @@
#include <vector>
#include "common/common_funcs.h"
+#include "common/common_types.h"
#include "common/swap.h"
#include "core/hle/service/am/am.h"
#include "core/hle/service/am/applets/applets.h"
+union ResultCode;
+
namespace Service::AM::Applets {
enum class KeysetDisable : u32 {
diff --git a/src/core/hle/service/audio/audout_u.cpp b/src/core/hle/service/audio/audout_u.cpp
index 6831c0735..bbe813490 100644
--- a/src/core/hle/service/audio/audout_u.cpp
+++ b/src/core/hle/service/audio/audout_u.cpp
@@ -18,17 +18,11 @@
#include "core/hle/kernel/readable_event.h"
#include "core/hle/kernel/writable_event.h"
#include "core/hle/service/audio/audout_u.h"
+#include "core/hle/service/audio/errors.h"
#include "core/memory.h"
namespace Service::Audio {
-namespace ErrCodes {
-enum {
- ErrorUnknown = 2,
- BufferCountExceeded = 8,
-};
-}
-
constexpr std::array<char, 10> DefaultDevice{{"DeviceOut"}};
constexpr int DefaultSampleRate{48000};
@@ -100,7 +94,7 @@ private:
if (stream->IsPlaying()) {
IPC::ResponseBuilder rb{ctx, 2};
- rb.Push(ResultCode(ErrorModule::Audio, ErrCodes::ErrorUnknown));
+ rb.Push(ERR_OPERATION_FAILED);
return;
}
@@ -143,7 +137,8 @@ private:
if (!audio_core.QueueBuffer(stream, tag, std::move(samples))) {
IPC::ResponseBuilder rb{ctx, 2};
- rb.Push(ResultCode(ErrorModule::Audio, ErrCodes::BufferCountExceeded));
+ rb.Push(ERR_BUFFER_COUNT_EXCEEDED);
+ return;
}
IPC::ResponseBuilder rb{ctx, 2};
diff --git a/src/core/hle/service/audio/audren_u.cpp b/src/core/hle/service/audio/audren_u.cpp
index 7e0cc64a8..c9de10a24 100644
--- a/src/core/hle/service/audio/audren_u.cpp
+++ b/src/core/hle/service/audio/audren_u.cpp
@@ -17,6 +17,7 @@
#include "core/hle/kernel/readable_event.h"
#include "core/hle/kernel/writable_event.h"
#include "core/hle/service/audio/audren_u.h"
+#include "core/hle/service/audio/errors.h"
namespace Service::Audio {
@@ -37,7 +38,7 @@ public:
{8, &IAudioRenderer::SetRenderingTimeLimit, "SetRenderingTimeLimit"},
{9, &IAudioRenderer::GetRenderingTimeLimit, "GetRenderingTimeLimit"},
{10, &IAudioRenderer::RequestUpdateImpl, "RequestUpdateAuto"},
- {11, nullptr, "ExecuteAudioRendererRendering"},
+ {11, &IAudioRenderer::ExecuteAudioRendererRendering, "ExecuteAudioRendererRendering"},
};
// clang-format on
RegisterHandlers(functions);
@@ -138,6 +139,17 @@ private:
rb.Push(rendering_time_limit_percent);
}
+ void ExecuteAudioRendererRendering(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_Audio, "called");
+
+ // This service command currently only reports an unsupported operation
+ // error code, or aborts. Given that, we just always return an error
+ // code in this case.
+
+ IPC::ResponseBuilder rb{ctx, 2};
+ rb.Push(ERR_NOT_SUPPORTED);
+ }
+
Kernel::EventPair system_event;
std::unique_ptr<AudioCore::AudioRenderer> renderer;
u32 rendering_time_limit_percent = 100;
@@ -235,7 +247,7 @@ AudRenU::AudRenU() : ServiceFramework("audren:u") {
{0, &AudRenU::OpenAudioRenderer, "OpenAudioRenderer"},
{1, &AudRenU::GetAudioRendererWorkBufferSize, "GetAudioRendererWorkBufferSize"},
{2, &AudRenU::GetAudioDeviceService, "GetAudioDeviceService"},
- {3, nullptr, "OpenAudioRendererAuto"},
+ {3, &AudRenU::OpenAudioRendererAuto, "OpenAudioRendererAuto"},
{4, &AudRenU::GetAudioDeviceServiceWithRevisionInfo, "GetAudioDeviceServiceWithRevisionInfo"},
};
// clang-format on
@@ -248,12 +260,7 @@ AudRenU::~AudRenU() = default;
void AudRenU::OpenAudioRenderer(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Service_Audio, "called");
- IPC::RequestParser rp{ctx};
- auto params = rp.PopRaw<AudioCore::AudioRendererParameter>();
- IPC::ResponseBuilder rb{ctx, 2, 0, 1};
-
- rb.Push(RESULT_SUCCESS);
- rb.PushIpcInterface<Audio::IAudioRenderer>(std::move(params));
+ OpenAudioRendererImpl(ctx);
}
void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
@@ -262,20 +269,20 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
LOG_DEBUG(Service_Audio, "called");
u64 buffer_sz = Common::AlignUp(4 * params.mix_buffer_count, 0x40);
- buffer_sz += params.unknown_c * 1024;
- buffer_sz += 0x940 * (params.unknown_c + 1);
+ buffer_sz += params.submix_count * 1024;
+ buffer_sz += 0x940 * (params.submix_count + 1);
buffer_sz += 0x3F0 * params.voice_count;
- buffer_sz += Common::AlignUp(8 * (params.unknown_c + 1), 0x10);
+ buffer_sz += Common::AlignUp(8 * (params.submix_count + 1), 0x10);
buffer_sz += Common::AlignUp(8 * params.voice_count, 0x10);
- buffer_sz +=
- Common::AlignUp((0x3C0 * (params.sink_count + params.unknown_c) + 4 * params.sample_count) *
- (params.mix_buffer_count + 6),
- 0x40);
+ buffer_sz += Common::AlignUp(
+ (0x3C0 * (params.sink_count + params.submix_count) + 4 * params.sample_count) *
+ (params.mix_buffer_count + 6),
+ 0x40);
if (IsFeatureSupported(AudioFeatures::Splitter, params.revision)) {
- u32 count = params.unknown_c + 1;
+ const u32 count = params.submix_count + 1;
u64 node_count = Common::AlignUp(count, 0x40);
- u64 node_state_buffer_sz =
+ const u64 node_state_buffer_sz =
4 * (node_count * node_count) + 0xC * node_count + 2 * (node_count / 8);
u64 edge_matrix_buffer_sz = 0;
node_count = Common::AlignUp(count * count, 0x40);
@@ -289,19 +296,19 @@ void AudRenU::GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx) {
buffer_sz += 0x20 * (params.effect_count + 4 * params.voice_count) + 0x50;
if (IsFeatureSupported(AudioFeatures::Splitter, params.revision)) {
- buffer_sz += 0xE0 * params.unknown_2c;
+ buffer_sz += 0xE0 * params.num_splitter_send_channels;
buffer_sz += 0x20 * params.splitter_count;
- buffer_sz += Common::AlignUp(4 * params.unknown_2c, 0x10);
+ buffer_sz += Common::AlignUp(4 * params.num_splitter_send_channels, 0x10);
}
buffer_sz = Common::AlignUp(buffer_sz, 0x40) + 0x170 * params.sink_count;
u64 output_sz = buffer_sz + 0x280 * params.sink_count + 0x4B0 * params.effect_count +
((params.voice_count * 256) | 0x40);
- if (params.unknown_1c >= 1) {
+ if (params.performance_frame_count >= 1) {
output_sz = Common::AlignUp(((16 * params.sink_count + 16 * params.effect_count +
16 * params.voice_count + 16) +
0x658) *
- (params.unknown_1c + 1) +
+ (params.performance_frame_count + 1) +
0xc0,
0x40) +
output_sz;
@@ -325,6 +332,12 @@ void AudRenU::GetAudioDeviceService(Kernel::HLERequestContext& ctx) {
rb.PushIpcInterface<Audio::IAudioDevice>();
}
+void AudRenU::OpenAudioRendererAuto(Kernel::HLERequestContext& ctx) {
+ LOG_DEBUG(Service_Audio, "called");
+
+ OpenAudioRendererImpl(ctx);
+}
+
void AudRenU::GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& ctx) {
LOG_WARNING(Service_Audio, "(STUBBED) called");
@@ -335,6 +348,15 @@ void AudRenU::GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& c
// based on the current revision
}
+void AudRenU::OpenAudioRendererImpl(Kernel::HLERequestContext& ctx) {
+ IPC::RequestParser rp{ctx};
+ const auto params = rp.PopRaw<AudioCore::AudioRendererParameter>();
+ IPC::ResponseBuilder rb{ctx, 2, 0, 1};
+
+ rb.Push(RESULT_SUCCESS);
+ rb.PushIpcInterface<IAudioRenderer>(params);
+}
+
bool AudRenU::IsFeatureSupported(AudioFeatures feature, u32_le revision) const {
u32_be version_num = (revision - Common::MakeMagic('R', 'E', 'V', '0')); // Byte swap
switch (feature) {
diff --git a/src/core/hle/service/audio/audren_u.h b/src/core/hle/service/audio/audren_u.h
index 3d63388fb..e55d25973 100644
--- a/src/core/hle/service/audio/audren_u.h
+++ b/src/core/hle/service/audio/audren_u.h
@@ -21,8 +21,11 @@ private:
void OpenAudioRenderer(Kernel::HLERequestContext& ctx);
void GetAudioRendererWorkBufferSize(Kernel::HLERequestContext& ctx);
void GetAudioDeviceService(Kernel::HLERequestContext& ctx);
+ void OpenAudioRendererAuto(Kernel::HLERequestContext& ctx);
void GetAudioDeviceServiceWithRevisionInfo(Kernel::HLERequestContext& ctx);
+ void OpenAudioRendererImpl(Kernel::HLERequestContext& ctx);
+
enum class AudioFeatures : u32 {
Splitter,
};
diff --git a/src/core/hle/service/audio/errors.h b/src/core/hle/service/audio/errors.h
new file mode 100644
index 000000000..6f8c09bcf
--- /dev/null
+++ b/src/core/hle/service/audio/errors.h
@@ -0,0 +1,15 @@
+// Copyright 2019 yuzu emulator team
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include "core/hle/result.h"
+
+namespace Service::Audio {
+
+constexpr ResultCode ERR_OPERATION_FAILED{ErrorModule::Audio, 2};
+constexpr ResultCode ERR_BUFFER_COUNT_EXCEEDED{ErrorModule::Audio, 8};
+constexpr ResultCode ERR_NOT_SUPPORTED{ErrorModule::Audio, 513};
+
+} // namespace Service::Audio
diff --git a/src/core/hle/service/hid/hid.h b/src/core/hle/service/hid/hid.h
index 6d897c842..7cc58db4c 100644
--- a/src/core/hle/service/hid/hid.h
+++ b/src/core/hle/service/hid/hid.h
@@ -15,7 +15,7 @@ namespace Kernel {
class SharedMemory;
}
-namespace SM {
+namespace Service::SM {
class ServiceManager;
}
diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp
index 21ccfe1f8..dbe7ee6e8 100644
--- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.cpp
@@ -23,7 +23,7 @@ u32 nvdisp_disp0::ioctl(Ioctl command, const std::vector<u8>& input, std::vector
void nvdisp_disp0::flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height,
u32 stride, NVFlinger::BufferQueue::BufferTransformFlags transform,
- const MathUtil::Rectangle<int>& crop_rect) {
+ const Common::Rectangle<int>& crop_rect) {
VAddr addr = nvmap_dev->GetObjectAddress(buffer_handle);
LOG_TRACE(Service,
"Drawing from address {:X} offset {:08X} Width {} Height {} Stride {} Format {}",
diff --git a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h
index a45086e45..ace71169f 100644
--- a/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h
+++ b/src/core/hle/service/nvdrv/devices/nvdisp_disp0.h
@@ -25,7 +25,7 @@ public:
/// Performs a screen flip, drawing the buffer pointed to by the handle.
void flip(u32 buffer_handle, u32 offset, u32 format, u32 width, u32 height, u32 stride,
NVFlinger::BufferQueue::BufferTransformFlags transform,
- const MathUtil::Rectangle<int>& crop_rect);
+ const Common::Rectangle<int>& crop_rect);
private:
std::shared_ptr<nvmap> nvmap_dev;
diff --git a/src/core/hle/service/nvflinger/buffer_queue.cpp b/src/core/hle/service/nvflinger/buffer_queue.cpp
index fc07d9bb8..4d150fc71 100644
--- a/src/core/hle/service/nvflinger/buffer_queue.cpp
+++ b/src/core/hle/service/nvflinger/buffer_queue.cpp
@@ -63,7 +63,7 @@ const IGBPBuffer& BufferQueue::RequestBuffer(u32 slot) const {
}
void BufferQueue::QueueBuffer(u32 slot, BufferTransformFlags transform,
- const MathUtil::Rectangle<int>& crop_rect) {
+ const Common::Rectangle<int>& crop_rect) {
auto itr = std::find_if(queue.begin(), queue.end(),
[&](const Buffer& buffer) { return buffer.slot == slot; });
ASSERT(itr != queue.end());
diff --git a/src/core/hle/service/nvflinger/buffer_queue.h b/src/core/hle/service/nvflinger/buffer_queue.h
index ab90d591e..e1ccb6171 100644
--- a/src/core/hle/service/nvflinger/buffer_queue.h
+++ b/src/core/hle/service/nvflinger/buffer_queue.h
@@ -67,14 +67,14 @@ public:
Status status = Status::Free;
IGBPBuffer igbp_buffer;
BufferTransformFlags transform;
- MathUtil::Rectangle<int> crop_rect;
+ Common::Rectangle<int> crop_rect;
};
void SetPreallocatedBuffer(u32 slot, const IGBPBuffer& igbp_buffer);
std::optional<u32> DequeueBuffer(u32 width, u32 height);
const IGBPBuffer& RequestBuffer(u32 slot) const;
void QueueBuffer(u32 slot, BufferTransformFlags transform,
- const MathUtil::Rectangle<int>& crop_rect);
+ const Common::Rectangle<int>& crop_rect);
std::optional<std::reference_wrapper<const Buffer>> AcquireBuffer();
void ReleaseBuffer(u32 slot);
u32 Query(QueryType type);
diff --git a/src/core/hle/service/vi/vi.cpp b/src/core/hle/service/vi/vi.cpp
index 74384a24d..a975767bb 100644
--- a/src/core/hle/service/vi/vi.cpp
+++ b/src/core/hle/service/vi/vi.cpp
@@ -420,7 +420,7 @@ public:
u32_le fence_is_valid;
std::array<Fence, 2> fences;
- MathUtil::Rectangle<int> GetCropRect() const {
+ Common::Rectangle<int> GetCropRect() const {
return {crop_left, crop_top, crop_right, crop_bottom};
}
};
diff --git a/src/core/memory.cpp b/src/core/memory.cpp
index e9166dbd9..ec279cef8 100644
--- a/src/core/memory.cpp
+++ b/src/core/memory.cpp
@@ -71,15 +71,20 @@ static void MapPages(PageTable& page_table, VAddr base, u64 size, u8* memory, Pa
FlushMode::FlushAndInvalidate);
VAddr end = base + size;
- while (base != end) {
- ASSERT_MSG(base < page_table.pointers.size(), "out of range mapping at {:016X}", base);
+ ASSERT_MSG(end <= page_table.pointers.size(), "out of range mapping at {:016X}",
+ base + page_table.pointers.size());
- page_table.attributes[base] = type;
- page_table.pointers[base] = memory;
+ std::fill(page_table.attributes.begin() + base, page_table.attributes.begin() + end, type);
- base += 1;
- if (memory != nullptr)
+ if (memory == nullptr) {
+ std::fill(page_table.pointers.begin() + base, page_table.pointers.begin() + end, memory);
+ } else {
+ while (base != end) {
+ page_table.pointers[base] = memory;
+
+ base += 1;
memory += PAGE_SIZE;
+ }
}
}
@@ -166,9 +171,6 @@ T Read(const VAddr vaddr) {
return value;
}
- // The memory access might do an MMIO or cached access, so we have to lock the HLE kernel state
- std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
-
PageType type = current_page_table->attributes[vaddr >> PAGE_BITS];
switch (type) {
case PageType::Unmapped:
@@ -199,9 +201,6 @@ void Write(const VAddr vaddr, const T data) {
return;
}
- // The memory access might do an MMIO or cached access, so we have to lock the HLE kernel state
- std::lock_guard<std::recursive_mutex> lock(HLE::g_hle_lock);
-
PageType type = current_page_table->attributes[vaddr >> PAGE_BITS];
switch (type) {
case PageType::Unmapped:
diff --git a/src/input_common/motion_emu.cpp b/src/input_common/motion_emu.cpp
index 9570c060e..6d96d4019 100644
--- a/src/input_common/motion_emu.cpp
+++ b/src/input_common/motion_emu.cpp
@@ -32,12 +32,12 @@ public:
}
void BeginTilt(int x, int y) {
- mouse_origin = Math::MakeVec(x, y);
+ mouse_origin = Common::MakeVec(x, y);
is_tilting = true;
}
void Tilt(int x, int y) {
- auto mouse_move = Math::MakeVec(x, y) - mouse_origin;
+ auto mouse_move = Common::MakeVec(x, y) - mouse_origin;
if (is_tilting) {
std::lock_guard<std::mutex> guard(tilt_mutex);
if (mouse_move.x == 0 && mouse_move.y == 0) {
@@ -45,7 +45,7 @@ public:
} else {
tilt_direction = mouse_move.Cast<float>();
tilt_angle =
- std::clamp(tilt_direction.Normalize() * sensitivity, 0.0f, MathUtil::PI * 0.5f);
+ std::clamp(tilt_direction.Normalize() * sensitivity, 0.0f, Common::PI * 0.5f);
}
}
}
@@ -56,7 +56,7 @@ public:
is_tilting = false;
}
- std::tuple<Math::Vec3<float>, Math::Vec3<float>> GetStatus() {
+ std::tuple<Common::Vec3<float>, Common::Vec3<float>> GetStatus() {
std::lock_guard<std::mutex> guard(status_mutex);
return status;
}
@@ -66,17 +66,17 @@ private:
const std::chrono::steady_clock::duration update_duration;
const float sensitivity;
- Math::Vec2<int> mouse_origin;
+ Common::Vec2<int> mouse_origin;
std::mutex tilt_mutex;
- Math::Vec2<float> tilt_direction;
+ Common::Vec2<float> tilt_direction;
float tilt_angle = 0;
bool is_tilting = false;
Common::Event shutdown_event;
- std::tuple<Math::Vec3<float>, Math::Vec3<float>> status;
+ std::tuple<Common::Vec3<float>, Common::Vec3<float>> status;
std::mutex status_mutex;
// Note: always keep the thread declaration at the end so that other objects are initialized
@@ -85,8 +85,8 @@ private:
void MotionEmuThread() {
auto update_time = std::chrono::steady_clock::now();
- Math::Quaternion<float> q = MakeQuaternion(Math::Vec3<float>(), 0);
- Math::Quaternion<float> old_q;
+ Common::Quaternion<float> q = Common::MakeQuaternion(Common::Vec3<float>(), 0);
+ Common::Quaternion<float> old_q;
while (!shutdown_event.WaitUntil(update_time)) {
update_time += update_duration;
@@ -96,18 +96,18 @@ private:
std::lock_guard<std::mutex> guard(tilt_mutex);
// Find the quaternion describing current 3DS tilting
- q = MakeQuaternion(Math::MakeVec(-tilt_direction.y, 0.0f, tilt_direction.x),
- tilt_angle);
+ q = Common::MakeQuaternion(
+ Common::MakeVec(-tilt_direction.y, 0.0f, tilt_direction.x), tilt_angle);
}
auto inv_q = q.Inverse();
// Set the gravity vector in world space
- auto gravity = Math::MakeVec(0.0f, -1.0f, 0.0f);
+ auto gravity = Common::MakeVec(0.0f, -1.0f, 0.0f);
// Find the angular rate vector in world space
auto angular_rate = ((q - old_q) * inv_q).xyz * 2;
- angular_rate *= 1000 / update_millisecond / MathUtil::PI * 180;
+ angular_rate *= 1000 / update_millisecond / Common::PI * 180;
// Transform the two vectors from world space to 3DS space
gravity = QuaternionRotate(inv_q, gravity);
@@ -131,7 +131,7 @@ public:
device = std::make_shared<MotionEmuDevice>(update_millisecond, sensitivity);
}
- std::tuple<Math::Vec3<float>, Math::Vec3<float>> GetStatus() const override {
+ std::tuple<Common::Vec3<float>, Common::Vec3<float>> GetStatus() const override {
return device->GetStatus();
}
diff --git a/src/tests/core/arm/arm_test_common.cpp b/src/tests/core/arm/arm_test_common.cpp
index 9b8a44fa1..ea27ef90d 100644
--- a/src/tests/core/arm/arm_test_common.cpp
+++ b/src/tests/core/arm/arm_test_common.cpp
@@ -13,11 +13,11 @@
namespace ArmTests {
TestEnvironment::TestEnvironment(bool mutable_memory_)
- : mutable_memory(mutable_memory_), test_memory(std::make_shared<TestMemory>(this)) {
-
+ : mutable_memory(mutable_memory_),
+ test_memory(std::make_shared<TestMemory>(this)), kernel{Core::System::GetInstance()} {
auto process = Kernel::Process::Create(kernel, "");
kernel.MakeCurrentProcess(process.get());
- page_table = &Core::CurrentProcess()->VMManager().page_table;
+ page_table = &process->VMManager().page_table;
std::fill(page_table->pointers.begin(), page_table->pointers.end(), nullptr);
page_table->special_regions.clear();
diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt
index b5a327936..c1ae83f4d 100644
--- a/src/video_core/CMakeLists.txt
+++ b/src/video_core/CMakeLists.txt
@@ -106,6 +106,8 @@ add_library(video_core STATIC
if (ENABLE_VULKAN)
target_sources(video_core PRIVATE
renderer_vulkan/declarations.h
+ renderer_vulkan/vk_buffer_cache.cpp
+ renderer_vulkan/vk_buffer_cache.h
renderer_vulkan/vk_device.cpp
renderer_vulkan/vk_device.h
renderer_vulkan/vk_memory_manager.cpp
@@ -113,7 +115,9 @@ if (ENABLE_VULKAN)
renderer_vulkan/vk_resource_manager.cpp
renderer_vulkan/vk_resource_manager.h
renderer_vulkan/vk_scheduler.cpp
- renderer_vulkan/vk_scheduler.h)
+ renderer_vulkan/vk_scheduler.h
+ renderer_vulkan/vk_stream_buffer.cpp
+ renderer_vulkan/vk_stream_buffer.h)
target_include_directories(video_core PRIVATE ../../externals/Vulkan-Headers/include)
target_compile_definitions(video_core PRIVATE HAS_VULKAN)
diff --git a/src/video_core/engines/fermi_2d.cpp b/src/video_core/engines/fermi_2d.cpp
index ec1a57226..03b7ee5d8 100644
--- a/src/video_core/engines/fermi_2d.cpp
+++ b/src/video_core/engines/fermi_2d.cpp
@@ -2,12 +2,11 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#include "core/core.h"
-#include "core/memory.h"
+#include "common/assert.h"
+#include "common/logging/log.h"
+#include "common/math_util.h"
#include "video_core/engines/fermi_2d.h"
-#include "video_core/engines/maxwell_3d.h"
#include "video_core/rasterizer_interface.h"
-#include "video_core/textures/decoders.h"
namespace Tegra::Engines {
@@ -44,10 +43,10 @@ void Fermi2D::HandleSurfaceCopy() {
const u32 src_blit_y2{
static_cast<u32>((regs.blit_src_y + (regs.blit_dst_height * regs.blit_dv_dy)) >> 32)};
- const MathUtil::Rectangle<u32> src_rect{src_blit_x1, src_blit_y1, src_blit_x2, src_blit_y2};
- const MathUtil::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y,
- regs.blit_dst_x + regs.blit_dst_width,
- regs.blit_dst_y + regs.blit_dst_height};
+ const Common::Rectangle<u32> src_rect{src_blit_x1, src_blit_y1, src_blit_x2, src_blit_y2};
+ const Common::Rectangle<u32> dst_rect{regs.blit_dst_x, regs.blit_dst_y,
+ regs.blit_dst_x + regs.blit_dst_width,
+ regs.blit_dst_y + regs.blit_dst_height};
if (!rasterizer.AccelerateSurfaceCopy(regs.src, regs.dst, src_rect, dst_rect)) {
UNIMPLEMENTED();
diff --git a/src/video_core/engines/fermi_2d.h b/src/video_core/engines/fermi_2d.h
index c69f74cc5..80523e320 100644
--- a/src/video_core/engines/fermi_2d.h
+++ b/src/video_core/engines/fermi_2d.h
@@ -5,7 +5,7 @@
#pragma once
#include <array>
-#include "common/assert.h"
+#include <cstddef>
#include "common/bit_field.h"
#include "common/common_funcs.h"
#include "common/common_types.h"
diff --git a/src/video_core/engines/kepler_compute.cpp b/src/video_core/engines/kepler_compute.cpp
index 4ca856b6b..b1d950460 100644
--- a/src/video_core/engines/kepler_compute.cpp
+++ b/src/video_core/engines/kepler_compute.cpp
@@ -2,9 +2,8 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
+#include "common/assert.h"
#include "common/logging/log.h"
-#include "core/core.h"
-#include "core/memory.h"
#include "video_core/engines/kepler_compute.h"
#include "video_core/memory_manager.h"
diff --git a/src/video_core/engines/kepler_compute.h b/src/video_core/engines/kepler_compute.h
index df0a32e0f..6575afd0f 100644
--- a/src/video_core/engines/kepler_compute.h
+++ b/src/video_core/engines/kepler_compute.h
@@ -5,8 +5,7 @@
#pragma once
#include <array>
-#include "common/assert.h"
-#include "common/bit_field.h"
+#include <cstddef>
#include "common/common_funcs.h"
#include "common/common_types.h"
#include "video_core/gpu.h"
diff --git a/src/video_core/engines/kepler_memory.h b/src/video_core/engines/kepler_memory.h
index f680c2ad9..9181e9d80 100644
--- a/src/video_core/engines/kepler_memory.h
+++ b/src/video_core/engines/kepler_memory.h
@@ -5,6 +5,7 @@
#pragma once
#include <array>
+#include <cstddef>
#include "common/bit_field.h"
#include "common/common_funcs.h"
#include "common/common_types.h"
diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp
index 2d2136067..144e7fa82 100644
--- a/src/video_core/engines/maxwell_3d.cpp
+++ b/src/video_core/engines/maxwell_3d.cpp
@@ -107,21 +107,23 @@ void Maxwell3D::CallMacroMethod(u32 method, std::vector<u32> parameters) {
void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
auto debug_context = system.GetGPUDebugContext();
+ const u32 method = method_call.method;
+
// It is an error to write to a register other than the current macro's ARG register before it
// has finished execution.
if (executing_macro != 0) {
- ASSERT(method_call.method == executing_macro + 1);
+ ASSERT(method == executing_macro + 1);
}
// Methods after 0xE00 are special, they're actually triggers for some microcode that was
// uploaded to the GPU during initialization.
- if (method_call.method >= MacroRegistersStart) {
+ if (method >= MacroRegistersStart) {
// We're trying to execute a macro
if (executing_macro == 0) {
// A macro call must begin by writing the macro method's register, not its argument.
- ASSERT_MSG((method_call.method % 2) == 0,
+ ASSERT_MSG((method % 2) == 0,
"Can't start macro execution by writing to the ARGS register");
- executing_macro = method_call.method;
+ executing_macro = method;
}
macro_params.push_back(method_call.argument);
@@ -133,66 +135,62 @@ void Maxwell3D::CallMethod(const GPU::MethodCall& method_call) {
return;
}
- ASSERT_MSG(method_call.method < Regs::NUM_REGS,
+ ASSERT_MSG(method < Regs::NUM_REGS,
"Invalid Maxwell3D register, increase the size of the Regs structure");
if (debug_context) {
debug_context->OnEvent(Tegra::DebugContext::Event::MaxwellCommandLoaded, nullptr);
}
- if (regs.reg_array[method_call.method] != method_call.argument) {
- regs.reg_array[method_call.method] = method_call.argument;
+ if (regs.reg_array[method] != method_call.argument) {
+ regs.reg_array[method] = method_call.argument;
// Color buffers
constexpr u32 first_rt_reg = MAXWELL3D_REG_INDEX(rt);
constexpr u32 registers_per_rt = sizeof(regs.rt[0]) / sizeof(u32);
- if (method_call.method >= first_rt_reg &&
- method_call.method < first_rt_reg + registers_per_rt * Regs::NumRenderTargets) {
- const std::size_t rt_index = (method_call.method - first_rt_reg) / registers_per_rt;
- dirty_flags.color_buffer |= 1u << static_cast<u32>(rt_index);
+ if (method >= first_rt_reg &&
+ method < first_rt_reg + registers_per_rt * Regs::NumRenderTargets) {
+ const std::size_t rt_index = (method - first_rt_reg) / registers_per_rt;
+ dirty_flags.color_buffer.set(rt_index);
}
// Zeta buffer
constexpr u32 registers_in_zeta = sizeof(regs.zeta) / sizeof(u32);
- if (method_call.method == MAXWELL3D_REG_INDEX(zeta_enable) ||
- method_call.method == MAXWELL3D_REG_INDEX(zeta_width) ||
- method_call.method == MAXWELL3D_REG_INDEX(zeta_height) ||
- (method_call.method >= MAXWELL3D_REG_INDEX(zeta) &&
- method_call.method < MAXWELL3D_REG_INDEX(zeta) + registers_in_zeta)) {
+ if (method == MAXWELL3D_REG_INDEX(zeta_enable) ||
+ method == MAXWELL3D_REG_INDEX(zeta_width) ||
+ method == MAXWELL3D_REG_INDEX(zeta_height) ||
+ (method >= MAXWELL3D_REG_INDEX(zeta) &&
+ method < MAXWELL3D_REG_INDEX(zeta) + registers_in_zeta)) {
dirty_flags.zeta_buffer = true;
}
// Shader
constexpr u32 shader_registers_count =
sizeof(regs.shader_config[0]) * Regs::MaxShaderProgram / sizeof(u32);
- if (method_call.method >= MAXWELL3D_REG_INDEX(shader_config[0]) &&
- method_call.method < MAXWELL3D_REG_INDEX(shader_config[0]) + shader_registers_count) {
+ if (method >= MAXWELL3D_REG_INDEX(shader_config[0]) &&
+ method < MAXWELL3D_REG_INDEX(shader_config[0]) + shader_registers_count) {
dirty_flags.shaders = true;
}
// Vertex format
- if (method_call.method >= MAXWELL3D_REG_INDEX(vertex_attrib_format) &&
- method_call.method <
- MAXWELL3D_REG_INDEX(vertex_attrib_format) + regs.vertex_attrib_format.size()) {
+ if (method >= MAXWELL3D_REG_INDEX(vertex_attrib_format) &&
+ method < MAXWELL3D_REG_INDEX(vertex_attrib_format) + regs.vertex_attrib_format.size()) {
dirty_flags.vertex_attrib_format = true;
}
// Vertex buffer
- if (method_call.method >= MAXWELL3D_REG_INDEX(vertex_array) &&
- method_call.method < MAXWELL3D_REG_INDEX(vertex_array) + 4 * 32) {
- dirty_flags.vertex_array |=
- 1u << ((method_call.method - MAXWELL3D_REG_INDEX(vertex_array)) >> 2);
- } else if (method_call.method >= MAXWELL3D_REG_INDEX(vertex_array_limit) &&
- method_call.method < MAXWELL3D_REG_INDEX(vertex_array_limit) + 2 * 32) {
- dirty_flags.vertex_array |=
- 1u << ((method_call.method - MAXWELL3D_REG_INDEX(vertex_array_limit)) >> 1);
- } else if (method_call.method >= MAXWELL3D_REG_INDEX(instanced_arrays) &&
- method_call.method < MAXWELL3D_REG_INDEX(instanced_arrays) + 32) {
- dirty_flags.vertex_array |=
- 1u << (method_call.method - MAXWELL3D_REG_INDEX(instanced_arrays));
+ if (method >= MAXWELL3D_REG_INDEX(vertex_array) &&
+ method < MAXWELL3D_REG_INDEX(vertex_array) + 4 * 32) {
+ dirty_flags.vertex_array.set((method - MAXWELL3D_REG_INDEX(vertex_array)) >> 2);
+ } else if (method >= MAXWELL3D_REG_INDEX(vertex_array_limit) &&
+ method < MAXWELL3D_REG_INDEX(vertex_array_limit) + 2 * 32) {
+ dirty_flags.vertex_array.set((method - MAXWELL3D_REG_INDEX(vertex_array_limit)) >> 1);
+ } else if (method >= MAXWELL3D_REG_INDEX(instanced_arrays) &&
+ method < MAXWELL3D_REG_INDEX(instanced_arrays) + 32) {
+ dirty_flags.vertex_array.set(method - MAXWELL3D_REG_INDEX(instanced_arrays));
}
}
- switch (method_call.method) {
+ switch (method) {
case MAXWELL3D_REG_INDEX(macros.data): {
ProcessMacroUpload(method_call.argument);
break;
diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h
index 0e3873ffd..7fbf1026e 100644
--- a/src/video_core/engines/maxwell_3d.h
+++ b/src/video_core/engines/maxwell_3d.h
@@ -5,8 +5,10 @@
#pragma once
#include <array>
+#include <bitset>
#include <unordered_map>
#include <vector>
+
#include "common/assert.h"
#include "common/bit_field.h"
#include "common/common_funcs.h"
@@ -503,7 +505,7 @@ public:
f32 translate_z;
INSERT_PADDING_WORDS(2);
- MathUtil::Rectangle<s32> GetRect() const {
+ Common::Rectangle<s32> GetRect() const {
return {
GetX(), // left
GetY() + GetHeight(), // top
@@ -1094,19 +1096,18 @@ public:
MemoryManager& memory_manager;
struct DirtyFlags {
- u8 color_buffer = 0xFF;
- bool zeta_buffer = true;
-
- bool shaders = true;
+ std::bitset<8> color_buffer{0xFF};
+ std::bitset<32> vertex_array{0xFFFFFFFF};
bool vertex_attrib_format = true;
- u32 vertex_array = 0xFFFFFFFF;
+ bool zeta_buffer = true;
+ bool shaders = true;
void OnMemoryWrite() {
- color_buffer = 0xFF;
zeta_buffer = true;
shaders = true;
- vertex_array = 0xFFFFFFFF;
+ color_buffer.set();
+ vertex_array.set();
}
};
diff --git a/src/video_core/engines/maxwell_dma.cpp b/src/video_core/engines/maxwell_dma.cpp
index 529a14ec7..0474c7ba3 100644
--- a/src/video_core/engines/maxwell_dma.cpp
+++ b/src/video_core/engines/maxwell_dma.cpp
@@ -3,6 +3,7 @@
// Refer to the license.txt file included.
#include "common/assert.h"
+#include "common/logging/log.h"
#include "core/core.h"
#include "core/memory.h"
#include "video_core/engines/maxwell_3d.h"
diff --git a/src/video_core/engines/maxwell_dma.h b/src/video_core/engines/maxwell_dma.h
index cf75aeb12..34c369320 100644
--- a/src/video_core/engines/maxwell_dma.h
+++ b/src/video_core/engines/maxwell_dma.h
@@ -5,6 +5,7 @@
#pragma once
#include <array>
+#include <cstddef>
#include "common/bit_field.h"
#include "common/common_funcs.h"
#include "common/common_types.h"
diff --git a/src/video_core/engines/shader_bytecode.h b/src/video_core/engines/shader_bytecode.h
index 252592edd..c7eb15b6a 100644
--- a/src/video_core/engines/shader_bytecode.h
+++ b/src/video_core/engines/shader_bytecode.h
@@ -6,7 +6,6 @@
#include <bitset>
#include <optional>
-#include <string>
#include <tuple>
#include <vector>
diff --git a/src/video_core/gpu.h b/src/video_core/gpu.h
index 0f5bfdcbf..6313702f2 100644
--- a/src/video_core/gpu.h
+++ b/src/video_core/gpu.h
@@ -100,7 +100,7 @@ struct FramebufferConfig {
using TransformFlags = Service::NVFlinger::BufferQueue::BufferTransformFlags;
TransformFlags transform_flags;
- MathUtil::Rectangle<int> crop_rect;
+ Common::Rectangle<int> crop_rect;
};
namespace Engines {
diff --git a/src/video_core/rasterizer_cache.h b/src/video_core/rasterizer_cache.h
index bcf0c15a4..a7bcf26fb 100644
--- a/src/video_core/rasterizer_cache.h
+++ b/src/video_core/rasterizer_cache.h
@@ -129,6 +129,15 @@ protected:
return ++modified_ticks;
}
+ /// Flushes the specified object, updating appropriate cache state as needed
+ void FlushObject(const T& object) {
+ if (!object->IsDirty()) {
+ return;
+ }
+ object->Flush();
+ object->MarkAsModified(false, *this);
+ }
+
private:
/// Returns a list of cached objects from the specified memory region, ordered by access time
std::vector<T> GetSortedObjectsFromRegion(VAddr addr, u64 size) {
@@ -154,15 +163,6 @@ private:
return objects;
}
- /// Flushes the specified object, updating appropriate cache state as needed
- void FlushObject(const T& object) {
- if (!object->IsDirty()) {
- return;
- }
- object->Flush();
- object->MarkAsModified(false, *this);
- }
-
using ObjectSet = std::set<T>;
using ObjectCache = std::unordered_map<VAddr, T>;
using IntervalCache = boost::icl::interval_map<VAddr, ObjectSet>;
diff --git a/src/video_core/rasterizer_interface.h b/src/video_core/rasterizer_interface.h
index b2a223705..6a1dc9cf6 100644
--- a/src/video_core/rasterizer_interface.h
+++ b/src/video_core/rasterizer_interface.h
@@ -47,8 +47,8 @@ public:
/// Attempt to use a faster method to perform a surface copy
virtual bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
const Tegra::Engines::Fermi2D::Regs::Surface& dst,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect) {
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect) {
return false;
}
diff --git a/src/video_core/renderer_base.cpp b/src/video_core/renderer_base.cpp
index 94223f45f..919d1f2d4 100644
--- a/src/video_core/renderer_base.cpp
+++ b/src/video_core/renderer_base.cpp
@@ -2,6 +2,7 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
+#include "common/logging/log.h"
#include "core/frontend/emu_window.h"
#include "core/settings.h"
#include "video_core/renderer_base.h"
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 12d876120..321d9dd3d 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -102,8 +102,8 @@ struct FramebufferCacheKey {
RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& window, Core::System& system,
ScreenInfo& info)
- : res_cache{*this}, shader_cache{*this, system}, emu_window{window}, screen_info{info},
- buffer_cache(*this, STREAM_BUFFER_SIZE), global_cache{*this} {
+ : res_cache{*this}, shader_cache{*this, system}, global_cache{*this}, emu_window{window},
+ screen_info{info}, buffer_cache(*this, STREAM_BUFFER_SIZE) {
// Create sampler objects
for (std::size_t i = 0; i < texture_samplers.size(); ++i) {
texture_samplers[i].Create();
@@ -200,7 +200,7 @@ GLuint RasterizerOpenGL::SetupVertexFormat() {
}
// Rebinding the VAO invalidates the vertex buffer bindings.
- gpu.dirty_flags.vertex_array = 0xFFFFFFFF;
+ gpu.dirty_flags.vertex_array.set();
state.draw.vertex_array = vao_entry.handle;
return vao_entry.handle;
@@ -210,14 +210,14 @@ void RasterizerOpenGL::SetupVertexBuffer(GLuint vao) {
auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
const auto& regs = gpu.regs;
- if (!gpu.dirty_flags.vertex_array)
+ if (gpu.dirty_flags.vertex_array.none())
return;
MICROPROFILE_SCOPE(OpenGL_VB);
// Upload all guest vertex arrays sequentially to our buffer
for (u32 index = 0; index < Maxwell::NumVertexArrays; ++index) {
- if (~gpu.dirty_flags.vertex_array & (1u << index))
+ if (!gpu.dirty_flags.vertex_array[index])
continue;
const auto& vertex_array = regs.vertex_array[index];
@@ -244,7 +244,7 @@ void RasterizerOpenGL::SetupVertexBuffer(GLuint vao) {
}
}
- gpu.dirty_flags.vertex_array = 0;
+ gpu.dirty_flags.vertex_array.reset();
}
DrawParameters RasterizerOpenGL::SetupDraw() {
@@ -488,13 +488,13 @@ std::pair<bool, bool> RasterizerOpenGL::ConfigureFramebuffers(
OpenGLState& current_state, bool using_color_fb, bool using_depth_fb, bool preserve_contents,
std::optional<std::size_t> single_color_target) {
MICROPROFILE_SCOPE(OpenGL_Framebuffer);
- const auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
+ auto& gpu = Core::System::GetInstance().GPU().Maxwell3D();
const auto& regs = gpu.regs;
const FramebufferConfigState fb_config_state{using_color_fb, using_depth_fb, preserve_contents,
single_color_target};
- if (fb_config_state == current_framebuffer_config_state && gpu.dirty_flags.color_buffer == 0 &&
- !gpu.dirty_flags.zeta_buffer) {
+ if (fb_config_state == current_framebuffer_config_state &&
+ gpu.dirty_flags.color_buffer.none() && !gpu.dirty_flags.zeta_buffer) {
// Only skip if the previous ConfigureFramebuffers call was from the same kind (multiple or
// single color targets). This is done because the guest registers may not change but the
// host framebuffer may contain different attachments
@@ -721,10 +721,10 @@ void RasterizerOpenGL::DrawArrays() {
// Add space for at least 18 constant buffers
buffer_size += Maxwell::MaxConstBuffers * (MaxConstbufferSize + uniform_buffer_alignment);
- bool invalidate = buffer_cache.Map(buffer_size);
+ const bool invalidate = buffer_cache.Map(buffer_size);
if (invalidate) {
// As all cached buffers are invalidated, we need to recheck their state.
- gpu.dirty_flags.vertex_array = 0xFFFFFFFF;
+ gpu.dirty_flags.vertex_array.set();
}
const GLuint vao = SetupVertexFormat();
@@ -738,19 +738,11 @@ void RasterizerOpenGL::DrawArrays() {
shader_program_manager->ApplyTo(state);
state.Apply();
- // Execute draw call
+ res_cache.SignalPreDrawCall();
params.DispatchDraw();
-
- // Disable scissor test
- state.viewports[0].scissor.enabled = false;
+ res_cache.SignalPostDrawCall();
accelerate_draw = AccelDraw::Disabled;
-
- // Unbind textures for potential future use as framebuffer attachments
- for (auto& texture_unit : state.texture_units) {
- texture_unit.Unbind();
- }
- state.Apply();
}
void RasterizerOpenGL::FlushAll() {}
@@ -779,8 +771,8 @@ void RasterizerOpenGL::FlushAndInvalidateRegion(VAddr addr, u64 size) {
bool RasterizerOpenGL::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
const Tegra::Engines::Fermi2D::Regs::Surface& dst,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect) {
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect) {
MICROPROFILE_SCOPE(OpenGL_Blits);
res_cache.FermiCopySurface(src, dst, src_rect, dst_rect);
return true;
@@ -1034,7 +1026,7 @@ void RasterizerOpenGL::SyncViewport(OpenGLState& current_state) {
for (std::size_t i = 0; i < viewport_count; i++) {
auto& viewport = current_state.viewports[i];
const auto& src = regs.viewports[i];
- const MathUtil::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
+ const Common::Rectangle<s32> viewport_rect{regs.viewport_transform[i].GetRect()};
viewport.x = viewport_rect.left;
viewport.y = viewport_rect.bottom;
viewport.width = viewport_rect.GetWidth();
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index 258d62259..2f0524f85 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -62,8 +62,8 @@ public:
void FlushAndInvalidateRegion(VAddr addr, u64 size) override;
bool AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Regs::Surface& src,
const Tegra::Engines::Fermi2D::Regs::Surface& dst,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect) override;
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect) override;
bool AccelerateDisplay(const Tegra::FramebufferConfig& config, VAddr framebuffer_addr,
u32 pixel_stride) override;
bool AccelerateDrawBatch(bool is_indexed) override;
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
index 642ccb269..876698b37 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.cpp
@@ -3,6 +3,7 @@
// Refer to the license.txt file included.
#include <algorithm>
+#include <optional>
#include <glad/glad.h>
#include "common/alignment.h"
@@ -399,7 +400,7 @@ static const FormatTuple& GetFormatTuple(PixelFormat pixel_format, ComponentType
return format;
}
-MathUtil::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
+Common::Rectangle<u32> SurfaceParams::GetRect(u32 mip_level) const {
u32 actual_height{std::max(1U, unaligned_height >> mip_level)};
if (IsPixelFormatASTC(pixel_format)) {
// ASTC formats must stop at the ATSC block size boundary
@@ -549,6 +550,8 @@ CachedSurface::CachedSurface(const SurfaceParams& params)
// alternatives. This signals a bug on those functions.
const auto width = static_cast<GLsizei>(params.MipWidth(0));
const auto height = static_cast<GLsizei>(params.MipHeight(0));
+ memory_size = params.MemorySize();
+ reinterpreted = false;
const auto& format_tuple = GetFormatTuple(params.pixel_format, params.component_type);
gl_internal_format = format_tuple.internal_format;
@@ -873,30 +876,31 @@ Surface RasterizerCacheOpenGL::GetColorBufferSurface(std::size_t index, bool pre
auto& gpu{Core::System::GetInstance().GPU().Maxwell3D()};
const auto& regs{gpu.regs};
- if ((gpu.dirty_flags.color_buffer & (1u << static_cast<u32>(index))) == 0) {
- return last_color_buffers[index];
+ if (!gpu.dirty_flags.color_buffer[index]) {
+ return current_color_buffers[index];
}
- gpu.dirty_flags.color_buffer &= ~(1u << static_cast<u32>(index));
+ gpu.dirty_flags.color_buffer.reset(index);
ASSERT(index < Tegra::Engines::Maxwell3D::Regs::NumRenderTargets);
if (index >= regs.rt_control.count) {
- return last_color_buffers[index] = {};
+ return current_color_buffers[index] = {};
}
if (regs.rt[index].Address() == 0 || regs.rt[index].format == Tegra::RenderTargetFormat::NONE) {
- return last_color_buffers[index] = {};
+ return current_color_buffers[index] = {};
}
const SurfaceParams color_params{SurfaceParams::CreateForFramebuffer(index)};
- return last_color_buffers[index] = GetSurface(color_params, preserve_contents);
+ return current_color_buffers[index] = GetSurface(color_params, preserve_contents);
}
void RasterizerCacheOpenGL::LoadSurface(const Surface& surface) {
surface->LoadGLBuffer();
surface->UploadGLTexture(read_framebuffer.handle, draw_framebuffer.handle);
surface->MarkAsModified(false, *this);
+ surface->MarkForReload(false);
}
Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool preserve_contents) {
@@ -908,18 +912,23 @@ Surface RasterizerCacheOpenGL::GetSurface(const SurfaceParams& params, bool pres
Surface surface{TryGet(params.addr)};
if (surface) {
if (surface->GetSurfaceParams().IsCompatibleSurface(params)) {
- // Use the cached surface as-is
+ // Use the cached surface as-is unless it's not synced with memory
+ if (surface->MustReload())
+ LoadSurface(surface);
return surface;
} else if (preserve_contents) {
// If surface parameters changed and we care about keeping the previous data, recreate
// the surface from the old one
Surface new_surface{RecreateSurface(surface, params)};
- Unregister(surface);
+ UnregisterSurface(surface);
Register(new_surface);
+ if (new_surface->IsUploaded()) {
+ RegisterReinterpretSurface(new_surface);
+ }
return new_surface;
} else {
// Delete the old surface before creating a new one to prevent collisions.
- Unregister(surface);
+ UnregisterSurface(surface);
}
}
@@ -973,8 +982,8 @@ void RasterizerCacheOpenGL::FastLayeredCopySurface(const Surface& src_surface,
}
static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect, GLuint read_fb_handle,
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect, GLuint read_fb_handle,
GLuint draw_fb_handle, GLenum src_attachment = 0, GLenum dst_attachment = 0,
std::size_t cubemap_face = 0) {
@@ -1104,7 +1113,7 @@ static bool BlitSurface(const Surface& src_surface, const Surface& dst_surface,
void RasterizerCacheOpenGL::FermiCopySurface(
const Tegra::Engines::Fermi2D::Regs::Surface& src_config,
const Tegra::Engines::Fermi2D::Regs::Surface& dst_config,
- const MathUtil::Rectangle<u32>& src_rect, const MathUtil::Rectangle<u32>& dst_rect) {
+ const Common::Rectangle<u32>& src_rect, const Common::Rectangle<u32>& dst_rect) {
const auto& src_params = SurfaceParams::CreateForFermiCopySurface(src_config);
const auto& dst_params = SurfaceParams::CreateForFermiCopySurface(dst_config);
@@ -1201,4 +1210,107 @@ Surface RasterizerCacheOpenGL::TryGetReservedSurface(const SurfaceParams& params
return {};
}
+static std::optional<u32> TryFindBestMipMap(std::size_t memory, const SurfaceParams params,
+ u32 height) {
+ for (u32 i = 0; i < params.max_mip_level; i++) {
+ if (memory == params.GetMipmapSingleSize(i) && params.MipHeight(i) == height) {
+ return {i};
+ }
+ }
+ return {};
+}
+
+static std::optional<u32> TryFindBestLayer(VAddr addr, const SurfaceParams params, u32 mipmap) {
+ const std::size_t size = params.LayerMemorySize();
+ VAddr start = params.addr + params.GetMipmapLevelOffset(mipmap);
+ for (u32 i = 0; i < params.depth; i++) {
+ if (start == addr) {
+ return {i};
+ }
+ start += size;
+ }
+ return {};
+}
+
+static bool LayerFitReinterpretSurface(RasterizerCacheOpenGL& cache, const Surface render_surface,
+ const Surface blitted_surface) {
+ const auto& dst_params = blitted_surface->GetSurfaceParams();
+ const auto& src_params = render_surface->GetSurfaceParams();
+ const std::size_t src_memory_size = src_params.size_in_bytes;
+ const std::optional<u32> level =
+ TryFindBestMipMap(src_memory_size, dst_params, src_params.height);
+ if (level.has_value()) {
+ if (src_params.width == dst_params.MipWidthGobAligned(*level) &&
+ src_params.height == dst_params.MipHeight(*level) &&
+ src_params.block_height >= dst_params.MipBlockHeight(*level)) {
+ const std::optional<u32> slot =
+ TryFindBestLayer(render_surface->GetAddr(), dst_params, *level);
+ if (slot.has_value()) {
+ glCopyImageSubData(render_surface->Texture().handle,
+ SurfaceTargetToGL(src_params.target), 0, 0, 0, 0,
+ blitted_surface->Texture().handle,
+ SurfaceTargetToGL(dst_params.target), *level, 0, 0, *slot,
+ dst_params.MipWidth(*level), dst_params.MipHeight(*level), 1);
+ blitted_surface->MarkAsModified(true, cache);
+ return true;
+ }
+ }
+ }
+ return false;
+}
+
+static bool IsReinterpretInvalid(const Surface render_surface, const Surface blitted_surface) {
+ const VAddr bound1 = blitted_surface->GetAddr() + blitted_surface->GetMemorySize();
+ const VAddr bound2 = render_surface->GetAddr() + render_surface->GetMemorySize();
+ if (bound2 > bound1)
+ return true;
+ const auto& dst_params = blitted_surface->GetSurfaceParams();
+ const auto& src_params = render_surface->GetSurfaceParams();
+ return (dst_params.component_type != src_params.component_type);
+}
+
+static bool IsReinterpretInvalidSecond(const Surface render_surface,
+ const Surface blitted_surface) {
+ const auto& dst_params = blitted_surface->GetSurfaceParams();
+ const auto& src_params = render_surface->GetSurfaceParams();
+ return (dst_params.height > src_params.height && dst_params.width > src_params.width);
+}
+
+bool RasterizerCacheOpenGL::PartialReinterpretSurface(Surface triggering_surface,
+ Surface intersect) {
+ if (IsReinterpretInvalid(triggering_surface, intersect)) {
+ UnregisterSurface(intersect);
+ return false;
+ }
+ if (!LayerFitReinterpretSurface(*this, triggering_surface, intersect)) {
+ if (IsReinterpretInvalidSecond(triggering_surface, intersect)) {
+ UnregisterSurface(intersect);
+ return false;
+ }
+ FlushObject(intersect);
+ FlushObject(triggering_surface);
+ intersect->MarkForReload(true);
+ }
+ return true;
+}
+
+void RasterizerCacheOpenGL::SignalPreDrawCall() {
+ if (texception && GLAD_GL_ARB_texture_barrier) {
+ glTextureBarrier();
+ }
+ texception = false;
+}
+
+void RasterizerCacheOpenGL::SignalPostDrawCall() {
+ for (u32 i = 0; i < Maxwell::NumRenderTargets; i++) {
+ if (current_color_buffers[i] != nullptr) {
+ Surface intersect = CollideOnReinterpretedSurface(current_color_buffers[i]->GetAddr());
+ if (intersect != nullptr) {
+ PartialReinterpretSurface(current_color_buffers[i], intersect);
+ texception = true;
+ }
+ }
+ }
+}
+
} // namespace OpenGL
diff --git a/src/video_core/renderer_opengl/gl_rasterizer_cache.h b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
index 89d733c50..797bbdc9c 100644
--- a/src/video_core/renderer_opengl/gl_rasterizer_cache.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer_cache.h
@@ -28,12 +28,13 @@ namespace OpenGL {
class CachedSurface;
using Surface = std::shared_ptr<CachedSurface>;
-using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, MathUtil::Rectangle<u32>>;
+using SurfaceSurfaceRect_Tuple = std::tuple<Surface, Surface, Common::Rectangle<u32>>;
using SurfaceTarget = VideoCore::Surface::SurfaceTarget;
using SurfaceType = VideoCore::Surface::SurfaceType;
using PixelFormat = VideoCore::Surface::PixelFormat;
using ComponentType = VideoCore::Surface::ComponentType;
+using Maxwell = Tegra::Engines::Maxwell3D::Regs;
struct SurfaceParams {
enum class SurfaceClass {
@@ -71,7 +72,7 @@ struct SurfaceParams {
}
/// Returns the rectangle corresponding to this surface
- MathUtil::Rectangle<u32> GetRect(u32 mip_level = 0) const;
+ Common::Rectangle<u32> GetRect(u32 mip_level = 0) const;
/// Returns the total size of this surface in bytes, adjusted for compression
std::size_t SizeInBytesRaw(bool ignore_tiled = false) const {
@@ -140,10 +141,18 @@ struct SurfaceParams {
return offset;
}
+ std::size_t GetMipmapSingleSize(u32 mip_level) const {
+ return InnerMipmapMemorySize(mip_level, false, is_layered);
+ }
+
u32 MipWidth(u32 mip_level) const {
return std::max(1U, width >> mip_level);
}
+ u32 MipWidthGobAligned(u32 mip_level) const {
+ return Common::AlignUp(std::max(1U, width >> mip_level), 64U * 8U / GetFormatBpp());
+ }
+
u32 MipHeight(u32 mip_level) const {
return std::max(1U, height >> mip_level);
}
@@ -346,6 +355,10 @@ public:
return cached_size_in_bytes;
}
+ std::size_t GetMemorySize() const {
+ return memory_size;
+ }
+
void Flush() override {
FlushGLBuffer();
}
@@ -395,6 +408,26 @@ public:
Tegra::Texture::SwizzleSource swizzle_z,
Tegra::Texture::SwizzleSource swizzle_w);
+ void MarkReinterpreted() {
+ reinterpreted = true;
+ }
+
+ bool IsReinterpreted() const {
+ return reinterpreted;
+ }
+
+ void MarkForReload(bool reload) {
+ must_reload = reload;
+ }
+
+ bool MustReload() const {
+ return must_reload;
+ }
+
+ bool IsUploaded() const {
+ return params.identity == SurfaceParams::SurfaceClass::Uploaded;
+ }
+
private:
void UploadGLMipmapTexture(u32 mip_map, GLuint read_fb_handle, GLuint draw_fb_handle);
@@ -408,6 +441,9 @@ private:
GLenum gl_internal_format{};
std::size_t cached_size_in_bytes{};
std::array<GLenum, 4> swizzle{GL_RED, GL_GREEN, GL_BLUE, GL_ALPHA};
+ std::size_t memory_size;
+ bool reinterpreted = false;
+ bool must_reload = false;
};
class RasterizerCacheOpenGL final : public RasterizerCache<Surface> {
@@ -430,8 +466,11 @@ public:
/// Copies the contents of one surface to another
void FermiCopySurface(const Tegra::Engines::Fermi2D::Regs::Surface& src_config,
const Tegra::Engines::Fermi2D::Regs::Surface& dst_config,
- const MathUtil::Rectangle<u32>& src_rect,
- const MathUtil::Rectangle<u32>& dst_rect);
+ const Common::Rectangle<u32>& src_rect,
+ const Common::Rectangle<u32>& dst_rect);
+
+ void SignalPreDrawCall();
+ void SignalPostDrawCall();
private:
void LoadSurface(const Surface& surface);
@@ -449,6 +488,10 @@ private:
/// Tries to get a reserved surface for the specified parameters
Surface TryGetReservedSurface(const SurfaceParams& params);
+ // Partialy reinterpret a surface based on a triggering_surface that collides with it.
+ // returns true if the reinterpret was successful, false in case it was not.
+ bool PartialReinterpretSurface(Surface triggering_surface, Surface intersect);
+
/// Performs a slow but accurate surface copy, flushing to RAM and reinterpreting the data
void AccurateCopySurface(const Surface& src_surface, const Surface& dst_surface);
void FastLayeredCopySurface(const Surface& src_surface, const Surface& dst_surface);
@@ -465,12 +508,50 @@ private:
OGLFramebuffer read_framebuffer;
OGLFramebuffer draw_framebuffer;
+ bool texception = false;
+
/// Use a Pixel Buffer Object to download the previous texture and then upload it to the new one
/// using the new format.
OGLBuffer copy_pbo;
- std::array<Surface, Tegra::Engines::Maxwell3D::Regs::NumRenderTargets> last_color_buffers;
+ std::array<Surface, Maxwell::NumRenderTargets> last_color_buffers;
+ std::array<Surface, Maxwell::NumRenderTargets> current_color_buffers;
Surface last_depth_buffer;
+
+ using SurfaceIntervalCache = boost::icl::interval_map<VAddr, Surface>;
+ using SurfaceInterval = typename SurfaceIntervalCache::interval_type;
+
+ static auto GetReinterpretInterval(const Surface& object) {
+ return SurfaceInterval::right_open(object->GetAddr() + 1,
+ object->GetAddr() + object->GetMemorySize() - 1);
+ }
+
+ // Reinterpreted surfaces are very fragil as the game may keep rendering into them.
+ SurfaceIntervalCache reinterpreted_surfaces;
+
+ void RegisterReinterpretSurface(Surface reinterpret_surface) {
+ auto interval = GetReinterpretInterval(reinterpret_surface);
+ reinterpreted_surfaces.insert({interval, reinterpret_surface});
+ reinterpret_surface->MarkReinterpreted();
+ }
+
+ Surface CollideOnReinterpretedSurface(VAddr addr) const {
+ const SurfaceInterval interval{addr};
+ for (auto& pair :
+ boost::make_iterator_range(reinterpreted_surfaces.equal_range(interval))) {
+ return pair.second;
+ }
+ return nullptr;
+ }
+
+ /// Unregisters an object from the cache
+ void UnregisterSurface(const Surface& object) {
+ if (object->IsReinterpreted()) {
+ auto interval = GetReinterpretInterval(object);
+ reinterpreted_surfaces.erase(interval);
+ }
+ Unregister(object);
+ }
};
} // namespace OpenGL
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 81882822b..82fc4d44b 100644
--- a/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_disk_cache.cpp
@@ -2,8 +2,6 @@
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
-#pragma once
-
#include <cstring>
#include <fmt/format.h>
#include <lz4.h>
diff --git a/src/video_core/renderer_opengl/gl_state.cpp b/src/video_core/renderer_opengl/gl_state.cpp
index 219f08053..9419326a3 100644
--- a/src/video_core/renderer_opengl/gl_state.cpp
+++ b/src/video_core/renderer_opengl/gl_state.cpp
@@ -461,7 +461,7 @@ void OpenGLState::ApplyTextures() const {
if (has_delta) {
glBindTextures(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
- textures.data());
+ textures.data() + first);
}
}
@@ -482,7 +482,7 @@ void OpenGLState::ApplySamplers() const {
}
if (has_delta) {
glBindSamplers(static_cast<GLuint>(first), static_cast<GLsizei>(last - first + 1),
- samplers.data());
+ samplers.data() + first);
}
}
diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp
index 272fc2e8e..8b510b6ae 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -244,6 +244,21 @@ void RendererOpenGL::InitOpenGLObjects() {
LoadColorToActiveGLTexture(0, 0, 0, 0, screen_info.texture);
}
+void RendererOpenGL::AddTelemetryFields() {
+ const char* const gl_version{reinterpret_cast<char const*>(glGetString(GL_VERSION))};
+ const char* const gpu_vendor{reinterpret_cast<char const*>(glGetString(GL_VENDOR))};
+ const char* const gpu_model{reinterpret_cast<char const*>(glGetString(GL_RENDERER))};
+
+ LOG_INFO(Render_OpenGL, "GL_VERSION: {}", gl_version);
+ LOG_INFO(Render_OpenGL, "GL_VENDOR: {}", gpu_vendor);
+ LOG_INFO(Render_OpenGL, "GL_RENDERER: {}", gpu_model);
+
+ auto& telemetry_session = system.TelemetrySession();
+ telemetry_session.AddField(Telemetry::FieldType::UserSystem, "GPU_Vendor", gpu_vendor);
+ telemetry_session.AddField(Telemetry::FieldType::UserSystem, "GPU_Model", gpu_model);
+ telemetry_session.AddField(Telemetry::FieldType::UserSystem, "GPU_OpenGL_Version", gl_version);
+}
+
void RendererOpenGL::CreateRasterizer() {
if (rasterizer) {
return;
@@ -257,6 +272,7 @@ void RendererOpenGL::ConfigureFramebufferTexture(TextureInfo& texture,
const Tegra::FramebufferConfig& framebuffer) {
texture.width = framebuffer.width;
texture.height = framebuffer.height;
+ texture.pixel_format = framebuffer.pixel_format;
GLint internal_format;
switch (framebuffer.pixel_format) {
@@ -465,17 +481,7 @@ bool RendererOpenGL::Init() {
glDebugMessageCallback(DebugHandler, nullptr);
}
- const char* gl_version{reinterpret_cast<char const*>(glGetString(GL_VERSION))};
- const char* gpu_vendor{reinterpret_cast<char const*>(glGetString(GL_VENDOR))};
- const char* gpu_model{reinterpret_cast<char const*>(glGetString(GL_RENDERER))};
-
- LOG_INFO(Render_OpenGL, "GL_VERSION: {}", gl_version);
- LOG_INFO(Render_OpenGL, "GL_VENDOR: {}", gpu_vendor);
- LOG_INFO(Render_OpenGL, "GL_RENDERER: {}", gpu_model);
-
- Core::Telemetry().AddField(Telemetry::FieldType::UserSystem, "GPU_Vendor", gpu_vendor);
- Core::Telemetry().AddField(Telemetry::FieldType::UserSystem, "GPU_Model", gpu_model);
- Core::Telemetry().AddField(Telemetry::FieldType::UserSystem, "GPU_OpenGL_Version", gl_version);
+ AddTelemetryFields();
if (!GLAD_GL_VERSION_4_3) {
return false;
diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h
index 7e13e566b..6cbf9d2cb 100644
--- a/src/video_core/renderer_opengl/renderer_opengl.h
+++ b/src/video_core/renderer_opengl/renderer_opengl.h
@@ -39,7 +39,7 @@ struct TextureInfo {
/// Structure used for storing information about the display target for the Switch screen
struct ScreenInfo {
GLuint display_texture;
- const MathUtil::Rectangle<float> display_texcoords{0.0f, 0.0f, 1.0f, 1.0f};
+ const Common::Rectangle<float> display_texcoords{0.0f, 0.0f, 1.0f, 1.0f};
TextureInfo texture;
};
@@ -60,6 +60,7 @@ public:
private:
void InitOpenGLObjects();
+ void AddTelemetryFields();
void CreateRasterizer();
void ConfigureFramebufferTexture(TextureInfo& texture,
@@ -102,7 +103,7 @@ private:
/// Used for transforming the framebuffer orientation
Tegra::FramebufferConfig::TransformFlags framebuffer_transform_flags;
- MathUtil::Rectangle<int> framebuffer_crop_rect;
+ Common::Rectangle<int> framebuffer_crop_rect;
};
} // namespace OpenGL
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
new file mode 100644
index 000000000..4a33a6c84
--- /dev/null
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
@@ -0,0 +1,116 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <cstring>
+#include <memory>
+#include <optional>
+#include <tuple>
+
+#include "common/alignment.h"
+#include "common/assert.h"
+#include "core/memory.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"
+#include "video_core/renderer_vulkan/vk_stream_buffer.h"
+
+namespace Vulkan {
+
+VKBufferCache::VKBufferCache(Tegra::MemoryManager& tegra_memory_manager,
+ VideoCore::RasterizerInterface& rasterizer, const VKDevice& device,
+ VKMemoryManager& memory_manager, VKScheduler& scheduler, u64 size)
+ : RasterizerCache{rasterizer}, tegra_memory_manager{tegra_memory_manager} {
+ const auto usage = vk::BufferUsageFlagBits::eVertexBuffer |
+ vk::BufferUsageFlagBits::eIndexBuffer |
+ vk::BufferUsageFlagBits::eUniformBuffer;
+ const auto access = vk::AccessFlagBits::eVertexAttributeRead | vk::AccessFlagBits::eIndexRead |
+ vk::AccessFlagBits::eUniformRead;
+ stream_buffer =
+ std::make_unique<VKStreamBuffer>(device, memory_manager, scheduler, size, usage, access,
+ vk::PipelineStageFlagBits::eAllCommands);
+ buffer_handle = stream_buffer->GetBuffer();
+}
+
+VKBufferCache::~VKBufferCache() = default;
+
+u64 VKBufferCache::UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size, u64 alignment,
+ bool cache) {
+ const auto cpu_addr{tegra_memory_manager.GpuToCpuAddress(gpu_addr)};
+ ASSERT(cpu_addr);
+
+ // Cache management is a big overhead, so only cache entries with a given size.
+ // TODO: Figure out which size is the best for given games.
+ cache &= size >= 2048;
+
+ if (cache) {
+ if (auto entry = TryGet(*cpu_addr); entry) {
+ if (entry->size >= size && entry->alignment == alignment) {
+ return entry->offset;
+ }
+ Unregister(entry);
+ }
+ }
+
+ AlignBuffer(alignment);
+ const u64 uploaded_offset = buffer_offset;
+
+ Memory::ReadBlock(*cpu_addr, buffer_ptr, size);
+
+ buffer_ptr += size;
+ buffer_offset += size;
+
+ if (cache) {
+ auto entry = std::make_shared<CachedBufferEntry>();
+ entry->offset = uploaded_offset;
+ entry->size = size;
+ entry->alignment = alignment;
+ entry->addr = *cpu_addr;
+ Register(entry);
+ }
+
+ return uploaded_offset;
+}
+
+u64 VKBufferCache::UploadHostMemory(const u8* raw_pointer, std::size_t size, u64 alignment) {
+ AlignBuffer(alignment);
+ std::memcpy(buffer_ptr, raw_pointer, size);
+ const u64 uploaded_offset = buffer_offset;
+
+ buffer_ptr += size;
+ buffer_offset += size;
+ return uploaded_offset;
+}
+
+std::tuple<u8*, u64> VKBufferCache::ReserveMemory(std::size_t size, u64 alignment) {
+ AlignBuffer(alignment);
+ u8* const uploaded_ptr = buffer_ptr;
+ const u64 uploaded_offset = buffer_offset;
+
+ buffer_ptr += size;
+ buffer_offset += size;
+ return {uploaded_ptr, uploaded_offset};
+}
+
+void VKBufferCache::Reserve(std::size_t max_size) {
+ bool invalidate;
+ std::tie(buffer_ptr, buffer_offset_base, invalidate) = stream_buffer->Reserve(max_size);
+ buffer_offset = buffer_offset_base;
+
+ if (invalidate) {
+ InvalidateAll();
+ }
+}
+
+VKExecutionContext VKBufferCache::Send(VKExecutionContext exctx) {
+ return stream_buffer->Send(exctx, buffer_offset - buffer_offset_base);
+}
+
+void VKBufferCache::AlignBuffer(std::size_t alignment) {
+ // Align the offset, not the mapped pointer
+ const u64 offset_aligned = Common::AlignUp(buffer_offset, alignment);
+ buffer_ptr += offset_aligned - buffer_offset;
+ buffer_offset = offset_aligned;
+}
+
+} // namespace Vulkan
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h
new file mode 100644
index 000000000..d8e916f31
--- /dev/null
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h
@@ -0,0 +1,87 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <memory>
+#include <tuple>
+
+#include "common/common_types.h"
+#include "video_core/gpu.h"
+#include "video_core/rasterizer_cache.h"
+#include "video_core/renderer_vulkan/declarations.h"
+#include "video_core/renderer_vulkan/vk_scheduler.h"
+
+namespace Tegra {
+class MemoryManager;
+}
+
+namespace Vulkan {
+
+class VKDevice;
+class VKFence;
+class VKMemoryManager;
+class VKStreamBuffer;
+
+struct CachedBufferEntry final : public RasterizerCacheObject {
+ VAddr GetAddr() const override {
+ return addr;
+ }
+
+ std::size_t GetSizeInBytes() const override {
+ return size;
+ }
+
+ // We do not have to flush this cache as things in it are never modified by us.
+ void Flush() override {}
+
+ VAddr addr;
+ std::size_t size;
+ u64 offset;
+ std::size_t alignment;
+};
+
+class VKBufferCache final : public RasterizerCache<std::shared_ptr<CachedBufferEntry>> {
+public:
+ explicit VKBufferCache(Tegra::MemoryManager& tegra_memory_manager,
+ VideoCore::RasterizerInterface& rasterizer, const VKDevice& device,
+ VKMemoryManager& memory_manager, VKScheduler& scheduler, u64 size);
+ ~VKBufferCache();
+
+ /// Uploads data from a guest GPU address. Returns host's buffer offset where it's been
+ /// allocated.
+ u64 UploadMemory(Tegra::GPUVAddr gpu_addr, std::size_t size, u64 alignment = 4,
+ bool cache = true);
+
+ /// Uploads from a host memory. Returns host's buffer offset where it's been allocated.
+ u64 UploadHostMemory(const u8* raw_pointer, std::size_t size, u64 alignment = 4);
+
+ /// Reserves memory to be used by host's CPU. Returns mapped address and offset.
+ std::tuple<u8*, u64> ReserveMemory(std::size_t size, u64 alignment = 4);
+
+ /// Reserves a region of memory to be used in subsequent upload/reserve operations.
+ void Reserve(std::size_t max_size);
+
+ /// Ensures that the set data is sent to the device.
+ [[nodiscard]] VKExecutionContext Send(VKExecutionContext exctx);
+
+ /// Returns the buffer cache handle.
+ vk::Buffer GetBuffer() const {
+ return buffer_handle;
+ }
+
+private:
+ void AlignBuffer(std::size_t alignment);
+
+ Tegra::MemoryManager& tegra_memory_manager;
+
+ std::unique_ptr<VKStreamBuffer> stream_buffer;
+ vk::Buffer buffer_handle;
+
+ u8* buffer_ptr = nullptr;
+ u64 buffer_offset = 0;
+ u64 buffer_offset_base = 0;
+};
+
+} // namespace Vulkan
diff --git a/src/video_core/renderer_vulkan/vk_memory_manager.cpp b/src/video_core/renderer_vulkan/vk_memory_manager.cpp
index 17ee93b91..0451babbf 100644
--- a/src/video_core/renderer_vulkan/vk_memory_manager.cpp
+++ b/src/video_core/renderer_vulkan/vk_memory_manager.cpp
@@ -238,7 +238,7 @@ bool VKMemoryManager::AllocMemory(vk::MemoryPropertyFlags wanted_properties, u32
VKMemoryCommitImpl::VKMemoryCommitImpl(VKMemoryAllocation* allocation, vk::DeviceMemory memory,
u8* data, u64 begin, u64 end)
- : allocation{allocation}, memory{memory}, data{data}, interval(std::make_pair(begin, end)) {}
+ : interval(std::make_pair(begin, end)), memory{memory}, allocation{allocation}, data{data} {}
VKMemoryCommitImpl::~VKMemoryCommitImpl() {
allocation->Free(this);
diff --git a/src/video_core/renderer_vulkan/vk_resource_manager.cpp b/src/video_core/renderer_vulkan/vk_resource_manager.cpp
index 1678463c7..a1e117443 100644
--- a/src/video_core/renderer_vulkan/vk_resource_manager.cpp
+++ b/src/video_core/renderer_vulkan/vk_resource_manager.cpp
@@ -125,11 +125,12 @@ void VKFence::Protect(VKResource* resource) {
protected_resources.push_back(resource);
}
-void VKFence::Unprotect(const VKResource* resource) {
+void VKFence::Unprotect(VKResource* resource) {
const auto it = std::find(protected_resources.begin(), protected_resources.end(), resource);
- if (it != protected_resources.end()) {
- protected_resources.erase(it);
- }
+ ASSERT(it != protected_resources.end());
+
+ resource->OnFenceRemoval(this);
+ protected_resources.erase(it);
}
VKFenceWatch::VKFenceWatch() = default;
@@ -141,12 +142,11 @@ VKFenceWatch::~VKFenceWatch() {
}
void VKFenceWatch::Wait() {
- if (!fence) {
+ if (fence == nullptr) {
return;
}
fence->Wait();
fence->Unprotect(this);
- fence = nullptr;
}
void VKFenceWatch::Watch(VKFence& new_fence) {
diff --git a/src/video_core/renderer_vulkan/vk_resource_manager.h b/src/video_core/renderer_vulkan/vk_resource_manager.h
index 5018dfa44..5bfe4cead 100644
--- a/src/video_core/renderer_vulkan/vk_resource_manager.h
+++ b/src/video_core/renderer_vulkan/vk_resource_manager.h
@@ -63,7 +63,7 @@ public:
void Protect(VKResource* resource);
/// Removes protection for a resource.
- void Unprotect(const VKResource* resource);
+ void Unprotect(VKResource* resource);
/// Retreives the fence.
operator vk::Fence() const {
diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.cpp b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp
new file mode 100644
index 000000000..58ffa42f2
--- /dev/null
+++ b/src/video_core/renderer_vulkan/vk_stream_buffer.cpp
@@ -0,0 +1,90 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <algorithm>
+#include <memory>
+#include <optional>
+#include <vector>
+
+#include "common/assert.h"
+#include "video_core/renderer_vulkan/declarations.h"
+#include "video_core/renderer_vulkan/vk_device.h"
+#include "video_core/renderer_vulkan/vk_memory_manager.h"
+#include "video_core/renderer_vulkan/vk_resource_manager.h"
+#include "video_core/renderer_vulkan/vk_scheduler.h"
+#include "video_core/renderer_vulkan/vk_stream_buffer.h"
+
+namespace Vulkan {
+
+constexpr u64 WATCHES_INITIAL_RESERVE = 0x4000;
+constexpr u64 WATCHES_RESERVE_CHUNK = 0x1000;
+
+VKStreamBuffer::VKStreamBuffer(const VKDevice& device, VKMemoryManager& memory_manager,
+ VKScheduler& scheduler, u64 size, vk::BufferUsageFlags usage,
+ vk::AccessFlags access, vk::PipelineStageFlags pipeline_stage)
+ : device{device}, scheduler{scheduler}, buffer_size{size}, access{access}, pipeline_stage{
+ pipeline_stage} {
+ CreateBuffers(memory_manager, usage);
+ ReserveWatches(WATCHES_INITIAL_RESERVE);
+}
+
+VKStreamBuffer::~VKStreamBuffer() = default;
+
+std::tuple<u8*, u64, bool> VKStreamBuffer::Reserve(u64 size) {
+ ASSERT(size <= buffer_size);
+ mapped_size = size;
+
+ if (offset + size > buffer_size) {
+ // The buffer would overflow, save the amount of used buffers, signal an invalidation and
+ // reset the state.
+ invalidation_mark = used_watches;
+ used_watches = 0;
+ offset = 0;
+ }
+
+ return {mapped_pointer + offset, offset, invalidation_mark.has_value()};
+}
+
+VKExecutionContext VKStreamBuffer::Send(VKExecutionContext exctx, u64 size) {
+ ASSERT_MSG(size <= mapped_size, "Reserved size is too small");
+
+ if (invalidation_mark) {
+ // TODO(Rodrigo): Find a better way to invalidate than waiting for all watches to finish.
+ exctx = scheduler.Flush();
+ std::for_each(watches.begin(), watches.begin() + *invalidation_mark,
+ [&](auto& resource) { resource->Wait(); });
+ invalidation_mark = std::nullopt;
+ }
+
+ if (used_watches + 1 >= watches.size()) {
+ // Ensure that there are enough watches.
+ ReserveWatches(WATCHES_RESERVE_CHUNK);
+ }
+ // Add a watch for this allocation.
+ watches[used_watches++]->Watch(exctx.GetFence());
+
+ offset += size;
+
+ return exctx;
+}
+
+void VKStreamBuffer::CreateBuffers(VKMemoryManager& memory_manager, vk::BufferUsageFlags usage) {
+ const vk::BufferCreateInfo buffer_ci({}, buffer_size, usage, vk::SharingMode::eExclusive, 0,
+ nullptr);
+
+ const auto dev = device.GetLogical();
+ const auto& dld = device.GetDispatchLoader();
+ buffer = dev.createBufferUnique(buffer_ci, nullptr, dld);
+ commit = memory_manager.Commit(*buffer, true);
+ mapped_pointer = commit->GetData();
+}
+
+void VKStreamBuffer::ReserveWatches(std::size_t grow_size) {
+ const std::size_t previous_size = watches.size();
+ watches.resize(previous_size + grow_size);
+ std::generate(watches.begin() + previous_size, watches.end(),
+ []() { return std::make_unique<VKFenceWatch>(); });
+}
+
+} // namespace Vulkan
diff --git a/src/video_core/renderer_vulkan/vk_stream_buffer.h b/src/video_core/renderer_vulkan/vk_stream_buffer.h
new file mode 100644
index 000000000..69d036ccd
--- /dev/null
+++ b/src/video_core/renderer_vulkan/vk_stream_buffer.h
@@ -0,0 +1,72 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <memory>
+#include <optional>
+#include <tuple>
+#include <vector>
+
+#include "common/common_types.h"
+#include "video_core/renderer_vulkan/declarations.h"
+#include "video_core/renderer_vulkan/vk_memory_manager.h"
+
+namespace Vulkan {
+
+class VKDevice;
+class VKFence;
+class VKFenceWatch;
+class VKResourceManager;
+class VKScheduler;
+
+class VKStreamBuffer {
+public:
+ explicit VKStreamBuffer(const VKDevice& device, VKMemoryManager& memory_manager,
+ VKScheduler& scheduler, u64 size, vk::BufferUsageFlags usage,
+ vk::AccessFlags access, vk::PipelineStageFlags pipeline_stage);
+ ~VKStreamBuffer();
+
+ /**
+ * Reserves a region of memory from the stream buffer.
+ * @param size Size to reserve.
+ * @returns A tuple in the following order: Raw memory pointer (with offset added), buffer
+ * offset and a boolean that's true when buffer has been invalidated.
+ */
+ std::tuple<u8*, u64, bool> Reserve(u64 size);
+
+ /// Ensures that "size" bytes of memory are available to the GPU, potentially recording a copy.
+ [[nodiscard]] VKExecutionContext Send(VKExecutionContext exctx, u64 size);
+
+ vk::Buffer GetBuffer() const {
+ return *buffer;
+ }
+
+private:
+ /// Creates Vulkan buffer handles committing the required the required memory.
+ void CreateBuffers(VKMemoryManager& memory_manager, vk::BufferUsageFlags usage);
+
+ /// Increases the amount of watches available.
+ void ReserveWatches(std::size_t grow_size);
+
+ const VKDevice& device; ///< Vulkan device manager.
+ VKScheduler& scheduler; ///< Command scheduler.
+ const u64 buffer_size; ///< Total size of the stream buffer.
+ const vk::AccessFlags access; ///< Access usage of this stream buffer.
+ const vk::PipelineStageFlags pipeline_stage; ///< Pipeline usage of this stream buffer.
+
+ UniqueBuffer buffer; ///< Mapped buffer.
+ VKMemoryCommit commit; ///< Memory commit.
+ u8* mapped_pointer{}; ///< Pointer to the host visible commit
+
+ u64 offset{}; ///< Buffer iterator.
+ u64 mapped_size{}; ///< Size reserved for the current copy.
+
+ std::vector<std::unique_ptr<VKFenceWatch>> watches; ///< Total watches
+ std::size_t used_watches{}; ///< Count of watches, reset on invalidation.
+ std::optional<std::size_t>
+ invalidation_mark{}; ///< Number of watches used in the current invalidation.
+};
+
+} // namespace Vulkan
diff --git a/src/video_core/surface.cpp b/src/video_core/surface.cpp
index 044ba116a..a7ac26d71 100644
--- a/src/video_core/surface.cpp
+++ b/src/video_core/surface.cpp
@@ -89,8 +89,6 @@ PixelFormat PixelFormatFromDepthFormat(Tegra::DepthFormat format) {
PixelFormat PixelFormatFromRenderTargetFormat(Tegra::RenderTargetFormat format) {
switch (format) {
- // TODO (Hexagon12): Converting SRGBA to RGBA is a hack and doesn't completely correct the
- // gamma.
case Tegra::RenderTargetFormat::RGBA8_SRGB:
return PixelFormat::RGBA8_SRGB;
case Tegra::RenderTargetFormat::RGBA8_UNORM:
diff --git a/src/web_service/verify_login.h b/src/web_service/verify_login.h
index 39db32dbb..821b345d7 100644
--- a/src/web_service/verify_login.h
+++ b/src/web_service/verify_login.h
@@ -4,8 +4,6 @@
#pragma once
-#include <functional>
-#include <future>
#include <string>
namespace WebService {
diff --git a/src/web_service/web_backend.cpp b/src/web_service/web_backend.cpp
index b7737b615..40da1a4e2 100644
--- a/src/web_service/web_backend.cpp
+++ b/src/web_service/web_backend.cpp
@@ -10,7 +10,6 @@
#include "common/common_types.h"
#include "common/logging/log.h"
#include "common/web_result.h"
-#include "core/settings.h"
#include "web_service/web_backend.h"
namespace WebService {
diff --git a/src/yuzu/compatdb.cpp b/src/yuzu/compatdb.cpp
index 5f0896f84..c8b0a5ec0 100644
--- a/src/yuzu/compatdb.cpp
+++ b/src/yuzu/compatdb.cpp
@@ -53,15 +53,15 @@ void CompatDB::Submit() {
case CompatDBPage::Final:
back();
LOG_DEBUG(Frontend, "Compatibility Rating: {}", compatibility->checkedId());
- Core::Telemetry().AddField(Telemetry::FieldType::UserFeedback, "Compatibility",
- compatibility->checkedId());
+ Core::System::GetInstance().TelemetrySession().AddField(
+ Telemetry::FieldType::UserFeedback, "Compatibility", compatibility->checkedId());
button(NextButton)->setEnabled(false);
button(NextButton)->setText(tr("Submitting"));
button(QWizard::CancelButton)->setVisible(false);
testcase_watcher.setFuture(QtConcurrent::run(
- [this]() { return Core::System::GetInstance().TelemetrySession().SubmitTestcase(); }));
+ [] { return Core::System::GetInstance().TelemetrySession().SubmitTestcase(); }));
break;
default:
LOG_ERROR(Frontend, "Unexpected page: {}", currentId());
diff --git a/src/yuzu/debugger/graphics/graphics_surface.cpp b/src/yuzu/debugger/graphics/graphics_surface.cpp
index 209798521..71683da8e 100644
--- a/src/yuzu/debugger/graphics/graphics_surface.cpp
+++ b/src/yuzu/debugger/graphics/graphics_surface.cpp
@@ -398,7 +398,7 @@ void GraphicsSurfaceWidget::OnUpdate() {
for (unsigned int y = 0; y < surface_height; ++y) {
for (unsigned int x = 0; x < surface_width; ++x) {
- Math::Vec4<u8> color;
+ 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];
diff --git a/src/yuzu/debugger/wait_tree.cpp b/src/yuzu/debugger/wait_tree.cpp
index f50225d5f..06ad74ffe 100644
--- a/src/yuzu/debugger/wait_tree.cpp
+++ b/src/yuzu/debugger/wait_tree.cpp
@@ -81,9 +81,8 @@ QString WaitTreeText::GetText() const {
return text;
}
-WaitTreeMutexInfo::WaitTreeMutexInfo(VAddr mutex_address) : mutex_address(mutex_address) {
- const auto& handle_table = Core::CurrentProcess()->GetHandleTable();
-
+WaitTreeMutexInfo::WaitTreeMutexInfo(VAddr mutex_address, const Kernel::HandleTable& handle_table)
+ : mutex_address(mutex_address) {
mutex_value = Memory::Read32(mutex_address);
owner_handle = static_cast<Kernel::Handle>(mutex_value & Kernel::Mutex::MutexOwnerMask);
owner = handle_table.Get<Kernel::Thread>(owner_handle);
@@ -316,7 +315,8 @@ std::vector<std::unique_ptr<WaitTreeItem>> WaitTreeThread::GetChildren() const {
const VAddr mutex_wait_address = thread.GetMutexWaitAddress();
if (mutex_wait_address != 0) {
- list.push_back(std::make_unique<WaitTreeMutexInfo>(mutex_wait_address));
+ const auto& handle_table = thread.GetOwnerProcess()->GetHandleTable();
+ list.push_back(std::make_unique<WaitTreeMutexInfo>(mutex_wait_address, handle_table));
} else {
list.push_back(std::make_unique<WaitTreeText>(tr("not waiting for mutex")));
}
diff --git a/src/yuzu/debugger/wait_tree.h b/src/yuzu/debugger/wait_tree.h
index 365c3dbfe..62886609d 100644
--- a/src/yuzu/debugger/wait_tree.h
+++ b/src/yuzu/debugger/wait_tree.h
@@ -17,6 +17,7 @@
class EmuThread;
namespace Kernel {
+class HandleTable;
class ReadableEvent;
class WaitObject;
class Thread;
@@ -72,7 +73,7 @@ public:
class WaitTreeMutexInfo : public WaitTreeExpandableItem {
Q_OBJECT
public:
- explicit WaitTreeMutexInfo(VAddr mutex_address);
+ explicit WaitTreeMutexInfo(VAddr mutex_address, const Kernel::HandleTable& handle_table);
~WaitTreeMutexInfo() override;
QString GetText() const override;
diff --git a/src/yuzu/main.cpp b/src/yuzu/main.cpp
index 1d460c189..0f5a14841 100644
--- a/src/yuzu/main.cpp
+++ b/src/yuzu/main.cpp
@@ -561,7 +561,10 @@ void GMainWindow::InitializeHotkeys() {
Settings::values.use_frame_limit = !Settings::values.use_frame_limit;
UpdateStatusBar();
});
- constexpr u16 SPEED_LIMIT_STEP = 5;
+ // TODO: Remove this comment/static whenever the next major release of
+ // MSVC occurs and we make it a requirement (see:
+ // https://developercommunity.visualstudio.com/content/problem/93922/constexprs-are-trying-to-be-captured-in-lambda-fun.html)
+ static constexpr u16 SPEED_LIMIT_STEP = 5;
connect(hotkey_registry.GetHotkey("Main Window", "Increase Speed Limit", this),
&QShortcut::activated, this, [&] {
if (Settings::values.frame_limit < 9999 - SPEED_LIMIT_STEP) {
@@ -846,7 +849,7 @@ bool GMainWindow::LoadROM(const QString& filename) {
}
game_path = filename;
- Core::Telemetry().AddField(Telemetry::FieldType::App, "Frontend", "Qt");
+ system.TelemetrySession().AddField(Telemetry::FieldType::App, "Frontend", "Qt");
return true;
}
diff --git a/src/yuzu_cmd/yuzu.cpp b/src/yuzu_cmd/yuzu.cpp
index c34b5467f..c6c66a787 100644
--- a/src/yuzu_cmd/yuzu.cpp
+++ b/src/yuzu_cmd/yuzu.cpp
@@ -216,7 +216,7 @@ int main(int argc, char** argv) {
}
}
- Core::Telemetry().AddField(Telemetry::FieldType::App, "Frontend", "SDL");
+ system.TelemetrySession().AddField(Telemetry::FieldType::App, "Frontend", "SDL");
system.Renderer().Rasterizer().LoadDiskResources();