preserve-emulation/ppsspp
1
0
Fork 0
mirror of https://github.com/hrydgard/ppsspp.git synced 2026-07-11 01:25:07 +02:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #21741 from hrydgard/extracted-commits

Browse Source 
Various minor changes (logging, auto-load zipped frame dumps, etc)
This commit is contained in:
Henrik Rydgård
2026-05-28 11:36:19 +02:00
committed by GitHub
parent 9c4fdde20f 22b20cf2ea
commit 3b6ecc4001
9 changed files with 113 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Common/GPU/Vulkan/VulkanQueueRunner.cpp
+3 -1
View File
@@ -327,6 +327,8 @@ void VulkanQueueRunner::RunSteps(std::vector<VKRStep *> &steps, int curFrame, Fr
frameData.AcquireNextImage(vulkan_);
if (frameData.hasAcquired && frameData.curSwapchainImage != (uint32_t)-1) {
SetBackbuffer(framebuffers_[frameData.curSwapchainImage], frameDataShared.swapchainImages_[frameData.curSwapchainImage].image);
} else {
perform = false;
}
}
@@ -1888,4 +1890,4 @@ const char *VKRRenderCommandToString(VKRRenderCommand cmd) {
} else {
return "N/A";
}
}
}
Common/Math/CrossSIMD.h
+56
View File
@@ -189,6 +189,10 @@ inline bool AllCompareBitsSet(Vec4S32 value) {
return _mm_movemask_ps(_mm_castsi128_ps(value.v)) == 0xF;
}
inline bool AnyCompareBitsSet(Vec4S32 value) {
return _mm_movemask_ps(_mm_castsi128_ps(value.v)) != 0;
}
struct Vec4F32 {
__m128 v;
@@ -241,6 +245,12 @@ struct Vec4F32 {
return Vec4F32{ _mm_or_ps(_mm_and_ps(value, _mm_load_ps((const float *)mask)), _mm_load_ps(onelane3)) };
}
static Vec4F32 LoadF24x3_DontCare(const uint32_t *src) {
alignas(16) static const uint32_t mask[4] = {0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x0};
return Vec4F32{_mm_castsi128_ps(_mm_slli_epi32(_mm_loadu_si128((const __m128i *)src), 8))};
}
void Store(float *dst) { _mm_storeu_ps(dst, v); }
void Store2(float *dst) { _mm_storel_epi64((__m128i *)dst, _mm_castps_si128(v)); }
void StoreAligned(float *dst) { _mm_store_ps(dst, v); }
@@ -299,6 +309,12 @@ struct Vec4F32 {
return Vec4F32{ _mm_or_ps(_mm_and_ps(v, _mm_load_ps((const float *)mask)), _mm_load_ps((const float *)onelane3)) };
}
Vec4F32 WithLane3From(Vec4F32 other) const {
alignas(16) static const uint32_t mask[4] = {0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0x0};
__m128 maskVec = _mm_load_ps((const float *)mask);
return Vec4F32{_mm_or_ps(_mm_and_ps(maskVec, v), _mm_andnot_ps(maskVec, other.v))};
}
inline Vec4F32 AsVec3ByMatrix44(const Mat4F32 &m) {
return Vec4F32{ _mm_add_ps(
_mm_add_ps(
@@ -353,6 +369,9 @@ inline bool AnyZeroSignBit(Vec4F32 value) {
inline bool AllCompareBitsSet(Vec4F32 value) {
return _mm_movemask_ps(value.v) == 0xF;
}
inline bool AnyCompareBitsSet(Vec4F32 value) {
return _mm_movemask_ps(value.v) != 0;
}
// Make sure the W component of scale is 1.0f.
inline void ScaleInplace(Mat4F32 &m, Vec4F32 scale) {
@@ -600,6 +619,9 @@ struct Vec4F32 {
static Vec4F32 LoadF24x3_One(const uint32_t *src) {
return Vec4F32{ vsetq_lane_f32(1.0f, vreinterpretq_f32_u32(vshlq_n_u32(vld1q_u32(src), 8)), 3) };
}
static Vec4F32 LoadF24x3_DontCare(const uint32_t *src) {
return Vec4F32{vreinterpretq_f32_u32(vshlq_n_u32(vld1q_u32(src), 8))};
}
static Vec4F32 FromVec4S32(Vec4S32 other) {
return Vec4F32{ vcvtq_f32_s32(other.v) };
@@ -683,6 +705,10 @@ struct Vec4F32 {
return Vec4F32{ vsetq_lane_f32(1.0f, v, 3) };
}
Vec4F32 WithLane3From(Vec4F32 other) const {
return Vec4F32{vsetq_lane_f32(vgetq_lane_f32(other.v, 3), v, 3)};
}
Vec4F32 ShuffleXXYY() const {
float32x2_t low = vget_low_f32(v); // {X, Y} pair
// Combine them into {X, X, Y, Y}
@@ -796,6 +822,17 @@ inline bool AllCompareBitsSet(Vec4S32 value) {
#endif
}
inline bool AnyCompareBitsSet(Vec4S32 value) {
#if PPSSPP_ARCH(ARM64_NEON)
return vmaxvq_u32(vreinterpretq_u32_s32(value.v)) != 0;
#else
// Very suboptimal, let's optimize later.
int32x2_t prod = vand_s32(vget_low_s32(value.v), vget_high_s32(value.v));
int mask = vget_lane_s32(prod, 0) & vget_lane_s32(prod, 1);
return mask != 0;
#endif
}
inline bool AnyZeroSignBit(Vec4F32 value) {
int32x4_t ival = vreinterpretq_s32_f32(value.v);
#if PPSSPP_ARCH(ARM64_NEON)
@@ -1045,6 +1082,9 @@ struct Vec4F32 {
value = __lsx_vinsgr2vr_w(value, kOneF32Bits, 3);
return Vec4F32{ (__m128)value };
}
static Vec4F32 LoadF24x3_DontCare(const uint32_t *src) {
return Vec4F32{(__m128)__lsx_vslli_w(__lsx_vld(src, 0), 8)};
}
static Vec4F32 FromVec4S32(Vec4S32 other) {
return Vec4F32{ (__m128)__lsx_vffint_s_w(other.v) };
@@ -1115,6 +1155,13 @@ struct Vec4F32 {
return Vec4F32{ (__m128)__lsx_vinsgr2vr_w((__m128i)v, kOneF32Bits, 3) };
}
Vec4F32 WithLane3From(Vec4F32 other) const {
// Use vinsgr2vr_w to insert just the lane 3
// Extract lane 3 from other as an integer, then insert it into v
int lane3 = __lsx_vpickve2gr_w((__m128i)other.v, 3);
return Vec4F32{ (__m128)__lsx_vinsgr2vr_w((__m128i)v, lane3, 3) };
}
Vec4S32 CompareEq(Vec4F32 other) const { return Vec4S32{ (__m128i)__lsx_vfcmp_ceq_s(v, other.v) }; }
Vec4S32 CompareLt(Vec4F32 other) const { return Vec4S32{ (__m128i)__lsx_vfcmp_clt_s(v, other.v) }; }
Vec4S32 CompareGt(Vec4F32 other) const { return Vec4S32{ (__m128i)__lsx_vfcmp_clt_s(other.v, v) }; }
@@ -1223,6 +1270,11 @@ inline bool AllCompareBitsSet(Vec4S32 value) {
return mask == 0xF;
}
inline bool AnyCompareBitsSet(Vec4S32 value) {
int mask = __lsx_vpickve2gr_w(__lsx_vmskltz_w((__m128i)value.v), 0);
return mask != 0;
}
struct Vec4U16 {
__m128i v; // we only use the lower 64 bits.
@@ -1478,6 +1530,10 @@ struct Vec4F32 {
return temp;
}
static Vec4F32 LoadF24x3_DontCare(const uint32_t *src) {
return LoadR24x3_One(src);
}
static Vec4F32 FromVec4S32(Vec4S32 src) {
Vec4F32 temp;
for (int i = 0; i < 4; i++) {
Core/FileSystems/BlockDevices.cpp
+2 -2
View File
@@ -357,7 +357,7 @@ UDFFileBlockDevice::UDFFileBlockDevice(FileLoader *fileLoader)
return;
}
INFO_LOG(Log::Loader, "Detected PSP DVD-R wrapper: USER_L0=%u blocks at %u, USER_L1=%u blocks at %u",
DEBUG_LOG(Log::Loader, "Detected PSP DVD-R wrapper: USER_L0=%u blocks at %u, USER_L1=%u blocks at %u",
layer0_.numBlocks, layer0_.startBlock, layer1_.numBlocks, layer1_.startBlock);
}
@@ -436,7 +436,7 @@ ISOContainerFileBlockDevice::ISOContainerFileBlockDevice(FileLoader *fileLoader)
return;
}
INFO_LOG(Log::Loader, "Detected PSP ISO wrapper: USER_L0=%u blocks at %u, USER_L1=%u blocks at %u",
DEBUG_LOG(Log::Loader, "Detected PSP ISO wrapper: USER_L0=%u blocks at %u, USER_L1=%u blocks at %u",
layer0_.numBlocks, layer0_.startBlock, layer1_.numBlocks, layer1_.startBlock);
}
GPU/D3D11/D3D11Util.cpp
+32 -9
View File
@@ -22,7 +22,7 @@
using namespace Microsoft::WRL;
std::vector<uint8_t> CompileShaderToBytecodeD3D11(const char *code, size_t codeSize, const char *target, UINT flags) {
std::vector<uint8_t> CompileShaderToBytecodeD3D11(const char *code, size_t codeSize, const char *target, UINT flags, std::string *errorMessage) {
ComPtr<ID3DBlob> compiledCode;
ComPtr<ID3DBlob> errorMsgs;
HRESULT result = ptr_D3DCompile(code, codeSize, nullptr, nullptr, nullptr, "main", target, flags, 0, &compiledCode, &errorMsgs);
@@ -45,6 +45,9 @@ std::vector<uint8_t> CompileShaderToBytecodeD3D11(const char *code, size_t codeS
}
} else {
ERROR_LOG(Log::G3D, "%s: %s\n\n%s", "errors", errors.c_str(), numberedCode.c_str());
if (errorMessage) {
*errorMessage = errors;
}
OutputDebugStringA(errors.c_str());
OutputDebugStringA(numberedCode.c_str());
}
@@ -63,9 +66,14 @@ HRESULT CreateVertexShaderD3D11(ID3D11Device *device, const char *code, size_t c
if (ppVertexShader)
*ppVertexShader = nullptr;
const char *profile = featureLevel <= D3D_FEATURE_LEVEL_9_3 ? "vs_4_0_level_9_1" : "vs_4_0";
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, profile, flags);
if (byteCode.empty())
std::string errorMessage;
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, profile, flags, &errorMessage);
if (byteCode.empty()) {
if (!errorMessage.empty()) {
ERROR_LOG(Log::G3D, "%s", errorMessage.c_str());
}
return S_FALSE;
}
auto hr = device->CreateVertexShader(byteCode.data(), byteCode.size(), nullptr, ppVertexShader);
if (byteCodeOut)
@@ -77,9 +85,14 @@ HRESULT CreatePixelShaderD3D11(ID3D11Device *device, const char *code, size_t co
if (ppPixelShader)
*ppPixelShader = nullptr;
const char *profile = featureLevel <= D3D_FEATURE_LEVEL_9_3 ? "ps_4_0_level_9_1" : "ps_4_0";
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, profile, flags);
if (byteCode.empty())
std::string errorMessage;
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, profile, flags, &errorMessage);
if (byteCode.empty()) {
if (!errorMessage.empty()) {
ERROR_LOG(Log::G3D, "%s", errorMessage.c_str());
}
return S_FALSE;
}
return device->CreatePixelShader(byteCode.data(), byteCode.size(), nullptr, ppPixelShader);
}
@@ -89,9 +102,14 @@ HRESULT CreateComputeShaderD3D11(ID3D11Device *device, const char *code, size_t
*ppComputeShader = nullptr;
if (featureLevel <= D3D_FEATURE_LEVEL_9_3)
return S_FALSE;
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, "cs_4_0", flags);
if (byteCode.empty())
std::string errorMessage;
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, "cs_4_0", flags, &errorMessage);
if (byteCode.empty()) {
if (!errorMessage.empty()) {
ERROR_LOG(Log::G3D, "%s", errorMessage.c_str());
}
return S_FALSE;
}
return device->CreateComputeShader(byteCode.data(), byteCode.size(), nullptr, ppComputeShader);
}
@@ -101,9 +119,14 @@ HRESULT CreateGeometryShaderD3D11(ID3D11Device *device, const char *code, size_t
*ppGeometryShader = nullptr;
if (featureLevel <= D3D_FEATURE_LEVEL_9_3)
return S_FALSE;
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, "gs_5_0", flags);
if (byteCode.empty())
std::string errorMessage;
std::vector<uint8_t> byteCode = CompileShaderToBytecodeD3D11(code, codeSize, "gs_5_0", flags, &errorMessage);
if (byteCode.empty()) {
if (!errorMessage.empty()) {
ERROR_LOG(Log::G3D, "%s", errorMessage.c_str());
}
return S_FALSE;
}
return device->CreateGeometryShader(byteCode.data(), byteCode.size(), nullptr, ppGeometryShader);
}
GPU/D3D11/D3D11Util.h
+1 -1
View File
@@ -73,7 +73,7 @@ private:
bool nextMapDiscard_ = false;
};
std::vector<uint8_t> CompileShaderToBytecodeD3D11(const char *code, size_t codeSize, const char *target, UINT flags);
std::vector<uint8_t> CompileShaderToBytecodeD3D11(const char *code, size_t codeSize, const char *target, UINT flags, std::string *errorMessage = nullptr);
HRESULT CreateVertexShaderD3D11(ID3D11Device *device, const char *code, size_t codeSize, std::vector<uint8_t> *byteCodeOut, D3D_FEATURE_LEVEL featureLevel, UINT flags, ID3D11VertexShader **);
HRESULT CreatePixelShaderD3D11(ID3D11Device *device, const char *code, size_t codeSize, D3D_FEATURE_LEVEL featureLevel, UINT flags, ID3D11PixelShader **);
UI/InstallZipScreen.cpp
+5
View File
@@ -393,6 +393,11 @@ void InstallZipScreen::update() {
}
MainScreen::showHomebrewTab = returnToHomebrew_;
}
if (zipFileInfo_.contents == ZipFileContents::FRAME_DUMP) {
// For frame dumps, just launch them immediately.
screenManager()->switchScreen(new EmuScreen(zipPath_));
}
}
if (existingSaveView_) {
UI/InstallZipScreen.h
+3 -1
View File
@@ -42,7 +42,9 @@ protected:
void CreateSettingsViews(UI::ViewGroup *parent) override;
void CreateContentViews(UI::ViewGroup *parent) override;
std::string_view GetTitle() const override;
ViewLayoutMode LayoutMode() const override {
return ViewLayoutMode::ApplyInsets;
}
private:
void OnInstall(UI::EventParams &params);
void OnPlay(UI::EventParams &params);
UI/SimpleDialogScreen.cpp
+1 -1
View File
@@ -45,7 +45,7 @@ void UISimpleBaseDialogScreen::CreateViews() {
ViewLayoutMode UITwoPaneBaseDialogScreen::LayoutMode() const {
const bool portrait = GetDeviceOrientation() == DeviceOrientation::Portrait;
if (portrait) {
if ((flags_ & TwoPaneFlags::SettingsCanScroll) || (flags_ & TwoPaneFlags::ContentsCanScroll)) {
if (((flags_ & TwoPaneFlags::SettingsCanScroll) && !(flags_ & TwoPaneFlags::SettingsToTheRight)) || (flags_ & TwoPaneFlags::ContentsCanScroll)) {
return ViewLayoutMode::IgnoreBottomInset;
} else {
return ViewLayoutMode::ApplyInsets;
unittest/TestShaderGenerators.cpp
+10 -4
View File
@@ -139,9 +139,14 @@ bool TestCompileShader(const char *buffer, ShaderLanguage lang, ShaderStage stag
case ShaderStage::Vertex: programType = "vs_4_0"; break;
case ShaderStage::Fragment: programType = "ps_4_0"; break;
case ShaderStage::Geometry: programType = "gs_4_0"; break;
default: return false;
default:
*errorMessage = "Unknown shader stage";
return false;
}
auto output = CompileShaderToBytecodeD3D11(buffer, strlen(buffer), programType, 0, errorMessage);
if (output.empty() && errorMessage->empty()) {
*errorMessage = "Error compiling HLSL shader: bytecode empty";
}
auto output = CompileShaderToBytecodeD3D11(buffer, strlen(buffer), programType, 0);
return !output.empty();
}
#endif
@@ -153,6 +158,7 @@ bool TestCompileShader(const char *buffer, ShaderLanguage lang, ShaderStage stag
case ShaderLanguage::GLSL_3xx:
return GLSLtoSPV(StageToVulkan(stage), buffer, GLSLVariant::GLES300, spirv, errorMessage);
default:
*errorMessage = "Unknown shader language";
return false;
}
}
@@ -422,7 +428,7 @@ bool TestVertexShaders() {
if (generateSuccess[j]) {
std::string errorMessage;
if (!TestCompileShader(buffer[j], languages[j], ShaderStage::Vertex, &errorMessage)) {
printf("Error compiling vertex shader %d:\n\n%s\n\n%s\n", (int)j, LineNumberString(buffer[j]).c_str(), errorMessage.c_str());
printf("Error compiling vertex shader %d:\n\n%s\n\nERROR: %s\n\n(end of error)\n\n", (int)j, LineNumberString(buffer[j]).c_str(), errorMessage.c_str());
for (int i = 0; i < numLanguages; i++) {
delete[] buffer[i];
}
@@ -496,7 +502,7 @@ bool TestFragmentShaders() {
if (generateSuccess[j]) {
std::string errorMessage;
if (!TestCompileShader(buffer[j], languages[j], ShaderStage::Fragment, &errorMessage)) {
printf("Error compiling fragment shader:\n\n%s\n\n%s\n", LineNumberString(buffer[j]).c_str(), errorMessage.c_str());
printf("Error compiling fragment shader %d:\n\n%s\n\nERROR: %s\n\n(end of error)\n\n", (int)j, LineNumberString(buffer[j]).c_str(), errorMessage.c_str());
for (int i = 0; i < numLanguages; i++) {
delete[] buffer[i];
}