renderer: emulate clip-disabled screen-space positions

This reverts commit 6758d67942.

An EXEC pair loses its mask provenance and falls into the raw high-word path. The saved all-ones high word then keeps waterfall branches true and wedges GPU work.

.github/workflows/build.yml

@@ -208,34 +208,66 @@ jobs:
         run: |
           archive="$RUNNER_TEMP/MoltenVK-macos.tar"
           package="$RUNNER_TEMP/MoltenVK"
+          install_dir="_Build/macos/install"
+          app_dir="$install_dir/KytyPS5.app"
+          moltenvk="$install_dir/libMoltenVK.dylib"
           curl --fail --location --retry 3 \
             --output "$archive" \
             "https://github.com/KhronosGroup/MoltenVK/releases/download/$MOLTENVK_VERSION/MoltenVK-macos.tar"
           echo "$MOLTENVK_SHA256  $archive" | shasum -a 256 --check
           tar -xf "$archive" -C "$RUNNER_TEMP"
+          test -d "$app_dir"
           install -m 755 \
             "$package/MoltenVK/dynamic/dylib/macOS/libMoltenVK.dylib" \
-            _Build/macos/install/libMoltenVK.dylib
-          install -m 644 "$package/LICENSE" _Build/macos/install/LICENSE.MoltenVK
-          codesign --force --sign - --timestamp=none _Build/macos/install/libMoltenVK.dylib
+            "$moltenvk"
+          install -m 644 "$package/LICENSE" "$install_dir/LICENSE.MoltenVK"
+          codesign --force --sign - --timestamp=none "$moltenvk"
+          install -d -m 755 "$app_dir/Contents/Frameworks" "$app_dir/Contents/Resources"
+          install -m 755 "$moltenvk" "$app_dir/Contents/Frameworks/libMoltenVK.dylib"
+          install -m 644 "$package/LICENSE" "$app_dir/Contents/Resources/LICENSE.MoltenVK"
+          codesign --force --sign - --timestamp=none "$app_dir"
 
       - name: Verify artifacts
         shell: bash
         run: |
-          file _Build/macos/install/launcher
-          file _Build/macos/install/kyty_emulator
-          file _Build/macos/install/libMoltenVK.dylib
-          lipo _Build/macos/install/launcher -verify_arch x86_64
-          lipo _Build/macos/install/kyty_emulator -verify_arch x86_64
-          lipo _Build/macos/install/libMoltenVK.dylib -verify_arch x86_64
-          codesign --verify --strict _Build/macos/install/kyty_emulator
-          codesign --verify --strict _Build/macos/install/libMoltenVK.dylib
+          install_dir="_Build/macos/install"
+          app_dir="$install_dir/KytyPS5.app"
+          test -d "$app_dir"
+          for binary in \
+            "$install_dir/kyty_emulator" \
+            "$install_dir/libMoltenVK.dylib" \
+            "$app_dir/Contents/MacOS/KytyPS5" \
+            "$app_dir/Contents/MacOS/kyty_emulator" \
+            "$app_dir/Contents/Frameworks/libMoltenVK.dylib"; do
+            file "$binary"
+            lipo "$binary" -verify_arch x86_64
+          done
+          for binary in \
+            "$install_dir/kyty_emulator" \
+            "$install_dir/libMoltenVK.dylib" \
+            "$app_dir/Contents/MacOS/KytyPS5" \
+            "$app_dir/Contents/MacOS/kyty_emulator" \
+            "$app_dir/Contents/Frameworks/libMoltenVK.dylib"; do
+            codesign --verify --strict "$binary"
+          done
+          codesign --verify --deep --strict --verbose=2 "$app_dir"
+
+      - name: Package macOS artifacts
+        shell: bash
+        run: |
+          install_dir="_Build/macos/install"
+          archive="$RUNNER_TEMP/KytyPS5-macOS-x86_64.zip"
+          (
+            cd "$install_dir"
+            zip -qry "$archive" .
+          )
+          unzip -t "$archive"
 
       - name: Upload macOS artifacts
         uses: actions/upload-artifact@v4
         with:
           name: KytyPS5-macOS-x86_64
-          path: _Build/macos/install/**
+          path: ${{ runner.temp }}/KytyPS5-macOS-x86_64.zip
           if-no-files-found: error
 
   linux:
@@ -424,16 +456,9 @@ jobs:
         if: needs.macos.result == 'success'
         shell: bash
         run: |
-          macos_dir="artifacts/KytyPS5-macOS-x86_64"
-          test -d "$macos_dir"
-          chmod a+x \
-            "$macos_dir/launcher" \
-            "$macos_dir/kyty_emulator" \
-            "$macos_dir/libMoltenVK.dylib"
-          (
-            cd "$macos_dir"
-            zip -r "$GITHUB_WORKSPACE/$RELEASE_NAME-macOS-x86_64.zip" .
-          )
+          macos_archive="artifacts/KytyPS5-macOS-x86_64/KytyPS5-macOS-x86_64.zip"
+          test -f "$macos_archive"
+          cp "$macos_archive" "$RELEASE_NAME-macOS-x86_64.zip"
 
       - name: Package Linux build
         if: needs.linux.result == 'success'

.gitignore

@@ -5,3 +5,4 @@ build/
 _Build/vscode-clang/
 _Build/
 cmake-build-debug/
+_PipelineCache/

CMakeLists.txt

@@ -574,6 +574,8 @@ if(BUILD_TESTING)
 		COMMAND $<TARGET_FILE:shader_recompiler_compute_tests> --storage-sampled-only)
 	add_test(NAME texture_cache_depth_readback
 		COMMAND $<TARGET_FILE:shader_recompiler_compute_tests> --depth-readback-only)
+	add_test(NAME compute_meta_clear_classification
+		COMMAND $<TARGET_FILE:shader_recompiler_compute_tests> --compute-meta-clear-only)
 	add_test(NAME buffer_cache_dirty_gc
 		COMMAND $<TARGET_FILE:shader_recompiler_compute_tests> --buffer-cache-gc-only)
 

README.md

@@ -220,17 +220,23 @@ cmake --install _Build/macos --prefix _Build/macos/install
 ```
 
 The build re-signs `kyty_emulator` with the JIT entitlements it needs to execute translated
-guest code; no manual signing step is required.
+guest code; no manual signing step is required. When the launcher is built, the install
+also produces `_Build/macos/install/KytyPS5.app` — double-click to launch the GUI.
+A flat `kyty_emulator` is kept for CLI usage.
 
 Vulkan comes from MoltenVK. Download `MoltenVK-macos.tar` from the
 [MoltenVK releases](https://github.com/KhronosGroup/MoltenVK/releases), then copy
-`MoltenVK/dynamic/dylib/macOS/libMoltenVK.dylib` next to `kyty_emulator` and ad-hoc sign it:
+`MoltenVK/dynamic/dylib/macOS/libMoltenVK.dylib` next to the flat `kyty_emulator`
+(and, for the bundle, into `KytyPS5.app/Contents/Frameworks/`) and ad-hoc sign it:
 
 ```bash
 codesign --force --sign - _Build/macos/install/libMoltenVK.dylib
+# For the bundle (if present):
+codesign --force --sign - _Build/macos/install/KytyPS5.app/Contents/Frameworks/libMoltenVK.dylib
+codesign --force --sign - _Build/macos/install/KytyPS5.app
 ```
 
-Release archives already include a signed `libMoltenVK.dylib`.
+Release archives already include a signed `libMoltenVK.dylib` (both flat and inside the bundle).
 
 ### Regression tests
 
@@ -274,6 +280,10 @@ To use the graphical launcher:
 ./_Build/linux/install/launcher
 ```
 
+```bash
+open _Build/macos/install/KytyPS5.app  # or double-click in Finder
+```
+
 On first launch, add one or more game folders in the global settings. The launcher searches those
 folders recursively for game directories containing `eboot.bin`. Select a detected game and run it
 from the game list.
@@ -288,12 +298,17 @@ The emulator can also be started directly with a legally obtained game directory
 ./_Build/linux/install/kyty_emulator --game "/games/ExampleGame"
 ```
 
-On macOS, point SDL at the MoltenVK library explicitly; the hardened runtime prevents it from
-being picked up from the executable's directory:
+On macOS, the adjacent flat or app-bundled `libMoltenVK.dylib` is found automatically; no
+environment variable is required:
 
 ```bash
-cd _Build/macos/install
-SDL_VULKAN_LIBRARY="$PWD/libMoltenVK.dylib" ./kyty_emulator --game "/games/ExampleGame"
+./_Build/macos/install/kyty_emulator --game "/games/ExampleGame"
+```
+
+To override the Vulkan loader, set `SDL_VULKAN_LIBRARY`:
+
+```bash
+SDL_VULKAN_LIBRARY=/path/to/libMoltenVK.dylib ./kyty_emulator --game "/games/ExampleGame"
 ```
 
 Run `kyty_emulator --help` to see the available graphics, logging, validation, profiling, and

src/common/emulatorConfig.cpp

@@ -47,6 +47,10 @@ PresentMode GetPresentMode() {
 	return g_config->present_mode;
 }
 
+int32_t GetGpuIndex() {
+	return g_config->gpu_index;
+}
+
 bool FullscreenEnabled() {
 	return g_config->fullscreen_enabled;
 }

src/common/emulatorConfig.h

@@ -48,6 +48,7 @@ struct ConfigOptions {
 	std::string            user_name                   = "Kyty";
 	int32_t                user_id                     = DEFAULT_USER_ID;
 	PresentMode            present_mode                = PresentMode::Fifo;
+	int32_t                gpu_index                   = -1;
 	bool                   fullscreen_enabled          = false;
 	uint32_t               vblank_frequency            = 60;
 	uint32_t               console_language            = DEFAULT_CONSOLE_LANGUAGE;
@@ -59,7 +60,7 @@ struct ConfigOptions {
 	bool                   command_buffer_dump_enabled = false;
 	std::filesystem::path  command_buffer_dump_folder  = "_Buffers";
 	bool                   graphics_debug_dump_enabled = false;
-	OutputDirection        printf_direction            = OutputDirection::Console;
+	OutputDirection        printf_direction            = OutputDirection::Silent;
 	std::filesystem::path  printf_output_file          = "_kyty.txt";
 	ProfilerDirection      profiler_direction          = ProfilerDirection::None;
 	bool                   spirv_debug_printf_enabled  = false;
@@ -80,6 +81,7 @@ uint32_t GetScreenHeight();
 const std::string& GetUserName();
 int32_t  GetUserId();
 PresentMode GetPresentMode();
+int32_t GetGpuIndex();
 bool     FullscreenEnabled();
 uint32_t GetVblankFrequency();
 uint32_t GetConsoleLanguage();

src/common/file.cpp

@@ -25,9 +25,6 @@
 #ifdef CopyFile
 #undef CopyFile
 #endif
-#ifdef MoveFile
-#undef MoveFile
-#endif
 #endif
 
 namespace Common {
@@ -544,14 +541,14 @@ bool File::CopyFile(const std::filesystem::path& src,
 	return SysFileCopyFile(src, dst);
 }
 
-bool File::MoveFile(const std::filesystem::path& src,
-                    const std::filesystem::path& dst) // @suppress("Member declaration not found")
+bool File::RenameFile(const std::filesystem::path& src,
+                      const std::filesystem::path& dst) // @suppress("Member declaration not found")
 {
 	if (IsFileExisting(dst)) {
 		DeleteFile(dst); // @suppress("Invalid arguments")
 	}
 
-	return SysFileMoveFile(src, dst);
+	return SysFileRenameFile(src, dst);
 }
 
 void File::RemoveReadonly(const std::filesystem::path& name) {

src/common/file.h

@@ -20,9 +20,6 @@
 #ifdef CopyFile
 #undef CopyFile
 #endif
-#ifdef MoveFile
-#undef MoveFile
-#endif
 #endif
 
 namespace Common {
@@ -107,7 +104,7 @@ public:
 	static std::vector<FindInfo> FindFiles(const std::filesystem::path& path);
 	static std::vector<DirEntry> GetDirEntries(const std::filesystem::path& path);
 	static bool CopyFile(const std::filesystem::path& src, const std::filesystem::path& dst);
-	static bool MoveFile(const std::filesystem::path& src, const std::filesystem::path& dst);
+	static bool RenameFile(const std::filesystem::path& src, const std::filesystem::path& dst);
 	static void RemoveReadonly(const std::filesystem::path& name);
 
 	KYTY_CLASS_NO_COPY(File);

src/common/platform/sysFileIO.h

@@ -80,7 +80,7 @@ void SysFileFindFiles(const std::filesystem::path& path, std::vector<sys_file_fi
 // NOLINTNEXTLINE(google-runtime-references)
 void SysFileGetDents(const std::filesystem::path& path, std::vector<sys_dir_entry_t>& out);
 bool SysFileCopyFile(const std::filesystem::path& src, const std::filesystem::path& dst);
-bool SysFileMoveFile(const std::filesystem::path& src, const std::filesystem::path& dst);
+bool SysFileRenameFile(const std::filesystem::path& src, const std::filesystem::path& dst);
 void SysFileRemoveReadonly(const std::filesystem::path& name);
 
 #endif /* KYTY_COMMON_PLATFORM_SYSFILEIO_H_ */

src/common/platform/sysLinuxFileIO.cpp

@@ -693,7 +693,7 @@ bool SysFileCopyFile(const std::filesystem::path& src, const std::filesystem::pa
 	       !error;
 }
 
-bool SysFileMoveFile(const std::filesystem::path& src, const std::filesystem::path& dst) {
+bool SysFileRenameFile(const std::filesystem::path& src, const std::filesystem::path& dst) {
 	auto real_src = get_internal_name(src);
 	auto real_dst = get_internal_name(dst);
 

src/common/platform/sysWindowsFileIO.cpp

@@ -565,7 +565,7 @@ bool SysFileCopyFile(const std::filesystem::path& src, const std::filesystem::pa
 	return CopyFileW(src_wide.c_str(), dst_wide.c_str(), FALSE) != 0;
 }
 
-bool SysFileMoveFile(const std::filesystem::path& src, const std::filesystem::path& dst) {
+bool SysFileRenameFile(const std::filesystem::path& src, const std::filesystem::path& dst) {
 	auto src_wide = src.wstring();
 	auto dst_wide = dst.wstring();
 	return MoveFileW(src_wide.c_str(), dst_wide.c_str()) != 0;

src/generate_version.cmake

@@ -1,5 +1,6 @@
 set(KYTY_GIT_VERSION "unknown")
 set(KYTY_GIT_HASH "unknown")
+set(KYTY_GIT_REVISION "unknown")
 if(GIT_EXECUTABLE)
 	execute_process(
 		COMMAND "${GIT_EXECUTABLE}" describe --tags --always --dirty
@@ -14,16 +15,18 @@ if(GIT_EXECUTABLE)
 	endif()
 
 	execute_process(
-		COMMAND "${GIT_EXECUTABLE}" rev-parse --short=7 HEAD
+		COMMAND "${GIT_EXECUTABLE}" rev-parse HEAD
 		WORKING_DIRECTORY "${GIT_WORKING_DIRECTORY}"
-		OUTPUT_VARIABLE KYTY_GIT_HASH
+		OUTPUT_VARIABLE KYTY_GIT_REVISION
 		OUTPUT_STRIP_TRAILING_WHITESPACE
 		RESULT_VARIABLE GIT_HASH_RESULT
 		ERROR_QUIET
 	)
 	if(NOT GIT_HASH_RESULT EQUAL 0)
 		set(KYTY_GIT_HASH "unknown")
+		set(KYTY_GIT_REVISION "unknown")
 	else()
+		string(SUBSTRING "${KYTY_GIT_REVISION}" 0 7 KYTY_GIT_HASH)
 		execute_process(
 			COMMAND "${GIT_EXECUTABLE}" diff-index --quiet HEAD --
 			WORKING_DIRECTORY "${GIT_WORKING_DIRECTORY}"

src/graphics/guest_gpu/command_processor/pm4Dispatch.cpp

@@ -23,9 +23,8 @@ constexpr auto MakeContextDispatchTable() {
 	g_hw_ctx_func[Pm4::DB_STENCIL_CLEAR]             = HwCtxSetStencilClear;
 	g_hw_ctx_func[Pm4::DB_DEPTH_CLEAR]               = HwCtxSetDepthClear;
 	g_hw_ctx_func[Pm4::PA_SC_SCREEN_SCISSOR_TL]      = HwCtxSetScreenScissor;
-	g_hw_ctx_func[Pm4::DB_Z_INFO]                    = HwCtxSetDepthRenderTarget;
 	g_hw_ctx_func[Pm4::DB_STENCIL_INFO]              = HwCtxSetStencilInfo;
-	g_hw_ctx_func[Pm4::DB_HTILE_SURFACE]             = HwCtxSetDepthHtileSurface;
+	g_hw_ctx_func[Pm4::DB_SHADING_RATE_ENCODING]     = HwCtxSetDepthShadingRateEncoding;
 	g_hw_ctx_func[Pm4::PA_SU_HARDWARE_SCREEN_OFFSET] = HwCtxSetHardwareScreenOffset;
 	g_hw_ctx_func[Pm4::PA_SC_WINDOW_OFFSET]          = HwCtxSetWindowOffset;
 	g_hw_ctx_func[Pm4::PA_SC_WINDOW_SCISSOR_TL]      = HwCtxSetWindowScissor;
@@ -75,7 +74,6 @@ constexpr auto MakeContextDispatchTable() {
 	g_hw_ctx_func[Pm4::PA_SC_BINNER_CNTL_0]                   = HwCtxSetPaScExtendedControl;
 	g_hw_ctx_func[Pm4::PA_SC_BINNER_CNTL_1]                   = HwCtxSetPaScExtendedControl;
 	g_hw_ctx_func[Pm4::PA_SC_CONSERVATIVE_RASTERIZATION_CNTL] = HwCtxSetPaScExtendedControl;
-	g_hw_ctx_func[Pm4::PA_SC_NGG_MODE_CNTL]                   = HwCtxSetPaScExtendedControl;
 	g_hw_ctx_func[Pm4::DB_ALPHA_TO_MASK]                      = HwCtxSetAlphaToMask;
 	g_hw_ctx_func[Pm4::VGT_DRAW_PAYLOAD_CNTL]                 = HwCtxSetDrawPayloadControl;
 	g_hw_ctx_func[Pm4::VGT_PRIMITIVEID_RESET]                 = HwCtxSetPrimitiveIdReset;
@@ -83,7 +81,6 @@ constexpr auto MakeContextDispatchTable() {
 	g_hw_ctx_func[Pm4::VGT_SHADER_STAGES_EN]                  = HwCtxSetShaderStages;
 
 	for (uint32_t slot = 0; slot < 8; slot++) {
-		g_hw_ctx_func[Pm4::CB_COLOR0_BASE + slot * 15] = HwCtxSetRenderTarget;
 		g_hw_ctx_func[Pm4::CB_COLOR0_INFO + slot * 15] = HwCtxSetColorInfo;
 
 		g_hw_ctx_func[Pm4::CB_BLEND0_CONTROL + slot * 1] = HwCtxSetBlendControl;
@@ -116,25 +113,21 @@ constexpr auto MakeShaderDispatchTable() {
 	}
 
 	for (uint32_t slot = 0; slot < 4; slot++) {
-		g_hw_sh_func[Pm4::SPI_SHADER_USER_ACCUM_PS_0 + slot] = HwShSetPsUserAccumSgpr;
+		g_hw_sh_func[Pm4::SPI_SHADER_USER_ACCUM_PS_0 + slot] = HwShIgnoreUserAccumulator;
 	}
 
 	for (uint32_t slot = 0; slot < 4; slot++) {
-		g_hw_sh_func[Pm4::SPI_SHADER_USER_ACCUM_ESGS_0 + slot] = HwShSetGsUserAccumSgpr;
+		g_hw_sh_func[Pm4::SPI_SHADER_USER_ACCUM_ESGS_0 + slot] = HwShIgnoreUserAccumulator;
 	}
 	for (uint32_t slot = 0; slot < 4; slot++) {
-		g_hw_sh_func[Pm4::SPI_SHADER_USER_ACCUM_LSHS_0 + slot] = HwShSetHsUserAccumSgpr;
+		g_hw_sh_func[Pm4::SPI_SHADER_USER_ACCUM_LSHS_0 + slot] = HwShIgnoreUserAccumulator;
 	}
 
-	for (uint32_t slot = 0; Pm4::COMPUTE_USER_ACCUM_0 + slot < Pm4::COMPUTE_USER_DATA_0; slot++) {
-		g_hw_sh_func[Pm4::COMPUTE_USER_ACCUM_0 + slot] = HwShSetCsUserAccumSgpr;
+	for (uint32_t slot = 0; slot < 4; slot++) {
+		g_hw_sh_func[Pm4::COMPUTE_USER_ACCUM_0 + slot] = HwShIgnoreUserAccumulator;
 	}
 	g_hw_sh_func[Pm4::COMPUTE_PGM_LO]                  = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_PGM_HI]                  = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_TBA_LO]                  = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_TBA_HI]                  = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_TMA_LO]                  = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_TMA_HI]                  = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_PGM_RSRC1]               = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_PGM_RSRC2]               = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_START_X]                 = HwShSetCsRegisters;
@@ -144,30 +137,16 @@ constexpr auto MakeShaderDispatchTable() {
 	g_hw_sh_func[Pm4::COMPUTE_NUM_THREAD_Y]            = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_NUM_THREAD_Z]            = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_RESOURCE_LIMITS]         = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_DESTINATION_EN_SE0]      = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_DESTINATION_EN_SE1]      = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_TMPRING_SIZE]            = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_DESTINATION_EN_SE2]      = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_DESTINATION_EN_SE3]      = HwShSetCsRegisters;
 	g_hw_sh_func[Pm4::COMPUTE_PGM_RSRC3]               = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_SHADER_CHKSUM]           = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::COMPUTE_DISPATCH_TUNNEL]         = HwShSetCsRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_TBA_LO_PS]            = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_PGM_RSRC3_PS]         = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_PGM_RSRC4_PS]         = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_TMA_LO_PS]            = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_TMA_HI_PS]            = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_REQ_CTRL_PS]          = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_PGM_RSRC3_GS]         = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_PGM_RSRC4_GS]         = HwShIgnoreRegisters;
+	g_hw_sh_func[Pm4::COMPUTE_PACE_ID]                 = HwShSetCsRegisters;
+	g_hw_sh_func[Pm4::SPI_GRAPHICS_SHADER_CONTROL_PS]  = HwShIgnoreRegisters;
+	g_hw_sh_func[Pm4::SPI_GRAPHICS_SHADER_CONTROL_GS]  = HwShIgnoreRegisters;
 	g_hw_sh_func[Pm4::SPI_SHADER_USER_DATA_ADDR_LO_GS] = HwShIgnoreRegisters;
 	g_hw_sh_func[Pm4::SPI_SHADER_USER_DATA_ADDR_HI_GS] = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_REQ_CTRL_ESGS]        = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_PGM_RSRC3_HS]         = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_PGM_RSRC4_HS]         = HwShIgnoreRegisters;
+	g_hw_sh_func[Pm4::SPI_GRAPHICS_SHADER_CONTROL_HS]  = HwShIgnoreRegisters;
 	g_hw_sh_func[Pm4::SPI_SHADER_USER_DATA_ADDR_LO_HS] = HwShIgnoreRegisters;
 	g_hw_sh_func[Pm4::SPI_SHADER_USER_DATA_ADDR_HI_HS] = HwShIgnoreRegisters;
-	g_hw_sh_func[Pm4::SPI_SHADER_REQ_CTRL_LSHS]        = HwShIgnoreRegisters;
 
 	return g_hw_sh_func;
 }
@@ -179,6 +158,7 @@ constexpr auto MakeUconfigDispatchTable() {
 	g_hw_uc_func[Pm4::VGT_INDEX_TYPE]            = HwUcSetIndexType;
 	g_hw_uc_func[Pm4::VGT_OBJECT_ID]             = HwUcSetObjectId;
 	g_hw_uc_func[Pm4::TEXTURE_GRADIENT_FACTORS]  = HwUcSetTextureGradientFactors;
+	g_hw_uc_func[Pm4::TEXTURE_GRADIENT_CONTROL]  = HwUcSetTextureGradientFactors;
 	g_hw_uc_func[Pm4::IA_MULTI_VGT_PARAM]        = HwUcSetIaMultiVgtParam;
 	g_hw_uc_func[Pm4::GE_MULTI_PRIM_IB_RESET_EN] = HwUcSetMultiPrimIbReset;
 	g_hw_uc_func[Pm4::TA_CS_BC_BASE_ADDR]        = HwUcSetBorderColorTableAddr;

src/graphics/guest_gpu/command_processor/pm4Dispatch.h

@@ -44,11 +44,9 @@ uint32_t HwCtxSetBorderColorTableAddr(CommandProcessor&, uint32_t, uint32_t, con
 uint32_t HwCtxSetStencilClear(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwCtxSetDepthClear(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwCtxSetScreenScissor(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
-uint32_t HwCtxSetDepthRenderTarget(CommandProcessor&, uint32_t, uint32_t, const uint32_t*,
-                                   uint32_t);
 uint32_t HwCtxSetStencilInfo(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
-uint32_t HwCtxSetDepthHtileSurface(CommandProcessor&, uint32_t, uint32_t, const uint32_t*,
-                                   uint32_t);
+uint32_t HwCtxSetDepthShadingRateEncoding(CommandProcessor&, uint32_t, uint32_t,
+                                          const uint32_t*, uint32_t);
 uint32_t HwCtxSetHardwareScreenOffset(CommandProcessor&, uint32_t, uint32_t, const uint32_t*,
                                       uint32_t);
 uint32_t HwCtxSetWindowOffset(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
@@ -89,7 +87,6 @@ uint32_t HwCtxSetDrawPayloadControl(CommandProcessor&, uint32_t, uint32_t, const
 uint32_t HwCtxSetPrimitiveIdReset(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwCtxSetObjprimIdControl(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwCtxSetShaderStages(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
-uint32_t HwCtxSetRenderTarget(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwCtxSetColorInfo(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwCtxSetBlendControl(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwCtxSetViewportScissor(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
@@ -100,10 +97,8 @@ uint32_t HwShSetPsUserSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t
 uint32_t HwShSetGsUserSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwShSetHsUserSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwShSetCsUserSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
-uint32_t HwShSetPsUserAccumSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
-uint32_t HwShSetGsUserAccumSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
-uint32_t HwShSetHsUserAccumSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
-uint32_t HwShSetCsUserAccumSgpr(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
+uint32_t HwShIgnoreUserAccumulator(CommandProcessor&, uint32_t, uint32_t, const uint32_t*,
+                                   uint32_t);
 uint32_t HwShSetCsRegisters(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwShIgnoreRegisters(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);
 uint32_t HwUcSetPrimitiveType(CommandProcessor&, uint32_t, uint32_t, const uint32_t*, uint32_t);

src/graphics/guest_gpu/gpu_defs.h

@@ -407,7 +407,19 @@ enum class BufferFormat : uint32_t {
 	k5_5_5_1UNorm       = 134,
 	k1_5_5_5UNorm       = 135,
 	k4_4_4_4UNorm       = 136,
+	kFmask8_S2_F1       = 156,
+	kFmask8_S4_F1       = 157,
+	kFmask8_S8_F1       = 158,
+	kFmask8_S2_F2       = 159,
+	kFmask8_S4_F2       = 160,
 	kFmask8_S4_F4       = 161,
+	kFmask16_S16_F1     = 162,
+	kFmask16_S8_F2      = 163,
+	kFmask32_S16_F2     = 164,
+	kFmask32_S8_F4      = 165,
+	kFmask32_S8_F8      = 166,
+	kFmask64_S16_F4     = 167,
+	kFmask64_S16_F8     = 168,
 	kBc1UNorm           = 169,
 	kBc1Srgb            = 170,
 	kBc2UNorm           = 171,

src/graphics/guest_gpu/gpu_format.cpp

@@ -68,7 +68,19 @@ constexpr FormatInfo kFormatInfo[] = {
 	{BufferFormat::k5_5_5_1UNorm, 2, 0, 2, true, false},
 	{BufferFormat::k1_5_5_5UNorm, 0, 0, 2, false, false},
 	{BufferFormat::k4_4_4_4UNorm, 2, 0, 2, true, false},
+	{BufferFormat::kFmask8_S2_F1, 1, 0, 1, true, false},
+	{BufferFormat::kFmask8_S4_F1, 1, 0, 1, true, false},
+	{BufferFormat::kFmask8_S8_F1, 1, 0, 1, true, false},
+	{BufferFormat::kFmask8_S2_F2, 1, 0, 1, true, false},
+	{BufferFormat::kFmask8_S4_F2, 1, 0, 1, true, false},
 	{BufferFormat::kFmask8_S4_F4, 1, 0, 1, true, false},
+	{BufferFormat::kFmask16_S16_F1, 2, 0, 2, true, false},
+	{BufferFormat::kFmask16_S8_F2, 2, 0, 2, true, false},
+	{BufferFormat::kFmask32_S16_F2, 4, 0, 4, true, false},
+	{BufferFormat::kFmask32_S8_F4, 4, 0, 4, true, false},
+	{BufferFormat::kFmask32_S8_F8, 4, 0, 4, true, false},
+	{BufferFormat::kFmask64_S16_F4, 8, 0, 8, true, false},
+	{BufferFormat::kFmask64_S16_F8, 8, 0, 8, true, false},
 	{BufferFormat::kBc1UNorm, 0, 8, 0, true, false},
 	{BufferFormat::kBc1Srgb, 0, 8, 0, true, false},
 	{BufferFormat::kBc2UNorm, 0, 16, 0, true, false},
@@ -203,7 +215,7 @@ uint32_t RenderTargetBytesPerElement(BufferFormat format) {
 }
 
 bool IsFmaskTextureFormat(BufferFormat format) {
-	return format == BufferFormat::kFmask8_S4_F4;
+	return format >= BufferFormat::kFmask8_S2_F1 && format <= BufferFormat::kFmask64_S16_F8;
 }
 
 bool IsSampledTextureFormat(BufferFormat format) {

src/graphics/guest_gpu/hardwareContext.h

@@ -11,15 +11,6 @@ struct ColorBase {
 	uint64_t addr = 0;
 };
 
-struct ColorPitch {
-	uint32_t pitch_div8_minus1       = 0;
-	uint32_t fmask_pitch_div8_minus1 = 0;
-};
-
-struct ColorSlice {
-	uint32_t slice_div64_minus1 = 0;
-};
-
 struct ColorView {
 	uint32_t base_array_slice_index = 0;
 	uint32_t last_array_slice_index = 0;
@@ -36,20 +27,15 @@ struct ColorInfo {
 	bool                    blend_clamp                    = false;
 	bool                    blend_bypass                   = false;
 	bool                    round_mode                     = false;
-	uint32_t                cmask_is_linear                = 0;
-	uint32_t                cmask_addr_type                = 0;
-	bool                    alt_tile_mode                  = false;
 	Prospero::ChannelLayout format                         = Prospero::ChannelLayout::kInvalid;
 	Prospero::ChannelType   channel_type                   = Prospero::ChannelType::kUNorm;
 	Prospero::ChannelOrder  channel_order                  = Prospero::ChannelOrder::kStandard;
 };
 
 struct ColorAttrib {
-	bool               force_dest_alpha_to_one = false;
-	Prospero::TileMode tile_mode               = Prospero::TileMode::kLinear;
-	Prospero::TileMode fmask_tile_mode         = Prospero::TileMode::kLinear;
-	uint32_t           num_samples             = 0;
-	uint32_t           num_fragments           = 0;
+	bool     force_dest_alpha_to_one = false;
+	uint32_t num_samples             = 0;
+	uint32_t num_fragments           = 0;
 };
 
 struct ColorAttrib2 {
@@ -59,41 +45,33 @@ struct ColorAttrib2 {
 };
 
 struct ColorAttrib3 {
-	uint32_t           depth              = 0;
-	Prospero::TileMode tile_mode          = Prospero::TileMode::kLinear;
-	uint32_t           dimension          = 0;
-	bool               cmask_pipe_aligned = false;
-	bool               dcc_pipe_aligned   = false;
+	uint32_t           depth                        = 0;
+	Prospero::TileMode tile_mode                    = Prospero::TileMode::kRenderTarget;
+	uint32_t           dimension                    = 0;
+	bool               metadata_pipe_aligned        = false;
+	bool               write_vrs_rate_hint_to_cmask = false;
 };
 
 struct ColorDccControl {
-	uint32_t max_uncompressed_block_size    = 0;
-	uint32_t max_compressed_block_size      = 0;
-	uint32_t min_compressed_block_size      = 0;
-	uint32_t color_transform                = 0;
-	bool     dcc_clear_key_enable           = false;
-	bool     overwrite_combiner_disable     = false;
-	bool     independent_64b_blocks         = false;
-	bool     independent_128b_blocks        = false;
-	bool     data_write_on_dcc_clear_to_reg = false;
+	enum class IndependentBlockSize : uint8_t { Disabled, Bytes64, Bytes128 };
+
+	uint32_t             max_uncompressed_block_size    = 2;
+	uint32_t             max_compressed_block_size      = 2;
+	uint32_t             color_transform                = 0;
+	bool                 dcc_clear_key_enable           = false;
+	bool                 overwrite_combiner_disable     = false;
+	IndependentBlockSize independent_block_size         = IndependentBlockSize::Disabled;
+	bool                 data_write_on_dcc_clear_to_reg = false;
 };
 
 struct ColorCmask {
 	uint64_t addr = 0;
 };
 
-struct ColorCmaskSlice {
-	uint32_t slice_minus1 = 0;
-};
-
 struct ColorFmask {
 	uint64_t addr = 0;
 };
 
-struct ColorFmaskSlice {
-	uint32_t slice_minus1 = 0;
-};
-
 struct ColorClearWord0 {
 	uint32_t word0 = 0;
 };
@@ -106,15 +84,8 @@ struct ColorDccAddr {
 	uint64_t addr = 0;
 };
 
-struct ColorSize {
-	uint32_t width  = 0;
-	uint32_t height = 0;
-};
-
 struct RenderTarget {
 	ColorBase       base;
-	ColorPitch      pitch;
-	ColorSlice      slice;
 	ColorView       view;
 	ColorInfo       info;
 	ColorAttrib     attrib;
@@ -122,13 +93,10 @@ struct RenderTarget {
 	ColorAttrib3    attrib3;
 	ColorDccControl dcc;
 	ColorCmask      cmask;
-	ColorCmaskSlice cmask_slice;
 	ColorFmask      fmask;
-	ColorFmaskSlice fmask_slice;
 	ColorClearWord0 clear_word0;
 	ColorClearWord1 clear_word1;
 	ColorDccAddr    dcc_addr;
-	ColorSize       size;
 };
 
 struct DepthZInfo {
@@ -136,7 +104,7 @@ struct DepthZInfo {
 	uint32_t                                    num_samples = 0;
 	Prospero::TextureCompatiblePlaneCompression texture_compatibility =
 	    Prospero::TextureCompatiblePlaneCompression::kDisable;
-	Prospero::ZCompareBase z_compare_base     = Prospero::ZCompareBase::kZMin;
+	Prospero::ZCompareBase z_compare_base     = Prospero::ZCompareBase::kZMax;
 	bool                   htile_acceleration = false;
 	bool                   expclear_enabled   = false;
 	bool                   partially_resident = false;
@@ -171,7 +139,7 @@ struct DepthStencilInfo {
 	Prospero::TextureCompatibleStencil texture_compatibility =
 	    Prospero::TextureCompatibleStencil::kDisable;
 	bool expclear_enabled       = false;
-	bool htile_stencil_disabled = false;
+	bool htile_stencil_disabled = true;
 	bool partially_resident     = false;
 
 	[[nodiscard]] static DepthStencilInfo Decode(uint32_t value) {
@@ -194,16 +162,6 @@ struct DepthStencilInfo {
 	}
 };
 
-struct DepthRenderTargetDepthInfo {
-	uint32_t addr5_swizzle_mask = 0;
-	uint32_t array_mode         = 0;
-	uint32_t pipe_config        = 0;
-	uint32_t bank_width         = 0;
-	uint32_t bank_height        = 0;
-	uint32_t macro_tile_aspect  = 0;
-	uint32_t num_banks          = 0;
-};
-
 struct DepthDepthView {
 	uint32_t slice_start           = 0;
 	uint32_t slice_max             = 0;
@@ -218,36 +176,18 @@ struct DepthDepthSizeXY {
 	bool     valid = false;
 };
 
-struct DepthRenderTargetHTileSurface {
-	uint32_t linear                  = 0;
-	uint32_t full_cache              = 0;
-	uint32_t htile_uses_preload_win  = 0;
-	uint32_t preload                 = 0;
-	uint32_t prefetch_width          = 0;
-	uint32_t prefetch_height         = 0;
-	uint32_t dst_outside_zero_to_one = 0;
-};
-
 struct DepthRenderTarget {
-	DepthZInfo                    z_info;
-	DepthStencilInfo              stencil_info;
-	DepthDepthView                depth_view;
-	DepthDepthSizeXY              size;
-	DepthRenderTargetDepthInfo    depth_info;
-	DepthRenderTargetHTileSurface htile_surface;
+	DepthZInfo       z_info;
+	DepthStencilInfo stencil_info;
+	DepthDepthView   depth_view;
+	DepthDepthSizeXY size;
 
 	uint64_t z_read_base_addr        = 0;
 	uint64_t stencil_read_base_addr  = 0;
 	uint64_t z_write_base_addr       = 0;
 	uint64_t stencil_write_base_addr = 0;
-	uint32_t pitch_div8_minus1       = 0;
-	uint32_t height_div8_minus1      = 0;
-	bool     pitch_height_valid      = false;
-	uint32_t slice_div64_minus1      = 0;
 	uint64_t htile_data_base_addr    = 0;
-	uint32_t width                   = 0;
-	uint32_t height                  = 0;
-	bool     width_height_valid      = false;
+	uint8_t  shading_rate_encoding   = 0;
 };
 
 struct RenderControl {
@@ -256,6 +196,8 @@ struct RenderControl {
 	bool    resummarize_enable       = false;
 	bool    stencil_compress_disable = false;
 	bool    depth_compress_disable   = false;
+	bool    copy_depth_to_color      = false;
+	bool    copy_stencil_to_color    = false;
 	bool    copy_centroid            = false;
 	uint8_t copy_sample              = 0;
 };
@@ -299,16 +241,14 @@ struct ClipControl {
 };
 
 struct DepthControl {
-	bool    stencil_enable                     = false;
-	bool    z_enable                           = false;
-	bool    z_write_enable                     = false;
-	bool    depth_bounds_enable                = false;
-	uint8_t zfunc                              = 0;
-	bool    backface_enable                    = false;
-	uint8_t stencilfunc                        = 0;
-	uint8_t stencilfunc_bf                     = 0;
-	bool    color_writes_on_depth_fail_enable  = false;
-	bool    color_writes_on_depth_pass_disable = false;
+	bool    stencil_enable      = false;
+	bool    z_enable            = false;
+	bool    z_write_enable      = false;
+	bool    depth_bounds_enable = false;
+	uint8_t zfunc               = 0;
+	bool    backface_enable     = false;
+	uint8_t stencilfunc         = 0;
+	uint8_t stencilfunc_bf      = 0;
 };
 
 struct StencilControl {
@@ -468,27 +408,6 @@ struct ScreenViewport {
 	bool     clip_rect_window_offset_enable[4]    = {};
 };
 
-struct VsShaderResource1 {
-	uint8_t vgprs                    = 0;
-	uint8_t sgprs                    = 0;
-	uint8_t priority                 = 0;
-	uint8_t float_mode               = 0;
-	bool    dx10_clamp               = false;
-	bool    ieee_mode                = false;
-	uint8_t vgpr_component_count     = 0;
-	bool    cu_group_enable          = false;
-	bool    require_forward_progress = false;
-	bool    fp16_overflow            = false;
-};
-
-struct VsShaderResource2 {
-	bool    scratch_en        = false;
-	uint8_t user_sgpr         = 0;
-	bool    offchip_lds       = false;
-	bool    streamout_enabled = false;
-	uint8_t shared_vgprs      = 0;
-};
-
 struct HsShaderResource1 {
 	uint8_t vgprs                     = 0;
 	uint8_t priority                  = 0;
@@ -511,7 +430,6 @@ struct HsShaderResource2 {
 
 struct PsShaderResource1 {
 	uint8_t vgprs                    = 0;
-	uint8_t sgprs                    = 0;
 	uint8_t priority                 = 0;
 	uint8_t float_mode               = 0;
 	bool    dx10_clamp               = false;
@@ -535,7 +453,6 @@ struct PsStageRegisters {
 	uint64_t          data_addr = 0;
 	PsShaderResource1 rsrc1;
 	PsShaderResource2 rsrc2;
-	uint64_t          chksum = 0;
 };
 
 struct CsStageRegisters {
@@ -545,12 +462,13 @@ struct CsStageRegisters {
 	uint32_t num_thread_y              = 0;
 	uint32_t num_thread_z              = 0;
 	uint8_t  vgprs                     = 0;
-	uint8_t  sgprs                     = 0;
+	uint8_t  priority                  = 0;
 	uint8_t  float_mode                = 0;
 	bool     dx10_clamp                = false;
+	bool     debug_mode                = false;
 	bool     ieee_mode                 = false;
+	bool     require_forward_progress  = false;
 	bool     fp16_overflow             = false;
-	bool     bulky                     = false;
 	bool     threadgroup_configuration = false;
 	uint8_t  wave_size                 = 64;
 	bool     scratch_en                = false;
@@ -561,6 +479,7 @@ struct CsStageRegisters {
 	bool     tg_size_en                = false;
 	uint8_t  tidig_comp_cnt            = 0;
 	uint16_t lds_size                  = 0;
+	uint8_t  shared_vgprs              = 0;
 };
 
 struct EsStageRegisters {
@@ -568,21 +487,17 @@ struct EsStageRegisters {
 };
 
 struct LsStageRegisters {
-	uint64_t          data_addr = 0;
-	VsShaderResource1 rsrc1;
-	VsShaderResource2 rsrc2;
+	uint64_t data_addr = 0;
 };
 
 struct HsStageRegisters {
 	uint64_t          data_addr = 0;
 	HsShaderResource1 rsrc1;
 	HsShaderResource2 rsrc2;
-	uint64_t          chksum = 0;
 };
 
 struct GsShaderResource1 {
 	uint8_t vgprs                     = 0;
-	uint8_t sgprs                     = 0;
 	uint8_t priority                  = 0;
 	uint8_t float_mode                = 0;
 	bool    dx10_clamp                = false;
@@ -608,7 +523,6 @@ struct GsStageRegisters {
 	uint64_t          data_addr = 0;
 	GsShaderResource1 rsrc1;
 	GsShaderResource2 rsrc2;
-	uint64_t          chksum = 0;
 };
 
 struct ShaderRegisters {
@@ -664,8 +578,7 @@ struct ShaderRegisters {
 enum class UserSgprType { Unknown, Region, Vsharp };
 
 struct UserSgprInfo {
-	// Changed from 16 to 64
-	static constexpr int SGPRS_MAX = 64;
+	static constexpr int SGPRS_MAX = 32;
 
 	uint32_t     value[SGPRS_MAX] = {0};
 	UserSgprType type[SGPRS_MAX]  = {};
@@ -712,12 +625,6 @@ public:
 	void Reset() { *this = Context(); }
 
 	void SetColorBase(uint32_t slot, const ColorBase& base) { m_render_targets[slot].base = base; }
-	void SetColorPitch(uint32_t slot, const ColorPitch& pitch) {
-		m_render_targets[slot].pitch = pitch;
-	}
-	void SetColorSlice(uint32_t slot, const ColorSlice& slice) {
-		m_render_targets[slot].slice = slice;
-	}
 	void SetColorView(uint32_t slot, const ColorView& view) { m_render_targets[slot].view = view; }
 	void SetColorInfo(uint32_t slot, const ColorInfo& info) { m_render_targets[slot].info = info; }
 	void SetColorAttrib(uint32_t slot, const ColorAttrib& attrib) {
@@ -735,15 +642,9 @@ public:
 	void SetColorCmask(uint32_t slot, const ColorCmask& cmask) {
 		m_render_targets[slot].cmask = cmask;
 	}
-	void SetColorCmaskSlice(uint32_t slot, const ColorCmaskSlice& cmask_slice) {
-		m_render_targets[slot].cmask_slice = cmask_slice;
-	}
 	void SetColorFmask(uint32_t slot, const ColorFmask& fmask) {
 		m_render_targets[slot].fmask = fmask;
 	}
-	void SetColorFmaskSlice(uint32_t slot, const ColorFmaskSlice& fmask_slice) {
-		m_render_targets[slot].fmask_slice = fmask_slice;
-	}
 	void SetColorClearWord0(uint32_t slot, const ColorClearWord0& clear_word0) {
 		m_render_targets[slot].clear_word0 = clear_word0;
 	}
@@ -753,7 +654,6 @@ public:
 	void SetColorDccAddr(uint32_t slot, const ColorDccAddr& dcc_addr) {
 		m_render_targets[slot].dcc_addr = dcc_addr;
 	}
-	void SetColorSize(uint32_t slot, const ColorSize& size) { m_render_targets[slot].size = size; }
 	[[nodiscard]] const RenderTarget& GetRenderTarget(uint32_t slot) const {
 		return m_render_targets[slot];
 	}
@@ -1072,21 +972,16 @@ public:
 
 	void SetEsShaderBase(uint64_t addr) { m_vs.es_regs.data_addr = addr; }
 	void SetLsShaderBase(uint64_t addr) { m_vs.ls_regs.data_addr = addr; }
-	void SetLsShaderResource1(const VsShaderResource1& rsrc1) { m_vs.ls_regs.rsrc1 = rsrc1; }
-	void SetLsShaderResource2(const VsShaderResource2& rsrc2) { m_vs.ls_regs.rsrc2 = rsrc2; }
 	void SetHsShaderBase(uint64_t addr) { m_vs.hs_regs.data_addr = addr; }
 	void SetHsShaderResource1(const HsShaderResource1& rsrc1) { m_vs.hs_regs.rsrc1 = rsrc1; }
 	void SetHsShaderResource2(const HsShaderResource2& rsrc2) { m_vs.hs_regs.rsrc2 = rsrc2; }
-	void SetHsShaderChksum(uint32_t value) { m_vs.hs_regs.chksum = value; }
 	void SetGsShaderBase(uint64_t addr) { m_vs.gs_regs.data_addr = addr; }
 	void SetGsShaderResource1(const GsShaderResource1& rsrc1) { m_vs.gs_regs.rsrc1 = rsrc1; }
 	void SetGsShaderResource2(const GsShaderResource2& rsrc2) { m_vs.gs_regs.rsrc2 = rsrc2; }
-	void SetGsShaderChksum(uint32_t value) { m_vs.gs_regs.chksum = value; }
 
 	void SetPsShaderBase(uint64_t addr) { m_ps.ps_regs.data_addr = addr; }
 	void SetPsShaderResource1(const PsShaderResource1& rsrc1) { m_ps.ps_regs.rsrc1 = rsrc1; }
 	void SetPsShaderResource2(const PsShaderResource2& rsrc2) { m_ps.ps_regs.rsrc2 = rsrc2; }
-	void SetPsShaderChksum(uint32_t value) { m_ps.ps_regs.chksum = value; }
 
 	void SetCsShader(const CsStageRegisters& cs_regs) { m_cs.cs_regs = cs_regs; }
 	void SetCsWaveSize(uint8_t wave_size) { m_cs.cs_regs.wave_size = wave_size; }

src/graphics/guest_gpu/pm4.h

@@ -98,6 +98,10 @@ constexpr uint32_t DB_RENDER_CONTROL_DEPTH_CLEAR_ENABLE_SHIFT       = 0;
 constexpr uint32_t DB_RENDER_CONTROL_DEPTH_CLEAR_ENABLE_MASK        = 0x1;
 constexpr uint32_t DB_RENDER_CONTROL_STENCIL_CLEAR_ENABLE_SHIFT     = 1;
 constexpr uint32_t DB_RENDER_CONTROL_STENCIL_CLEAR_ENABLE_MASK      = 0x1;
+constexpr uint32_t DB_RENDER_CONTROL_COPY_DEPTH_TO_COLOR_SHIFT      = 2;
+constexpr uint32_t DB_RENDER_CONTROL_COPY_DEPTH_TO_COLOR_MASK       = 0x1;
+constexpr uint32_t DB_RENDER_CONTROL_COPY_STENCIL_TO_COLOR_SHIFT    = 3;
+constexpr uint32_t DB_RENDER_CONTROL_COPY_STENCIL_TO_COLOR_MASK     = 0x1;
 constexpr uint32_t DB_RENDER_CONTROL_RESUMMARIZE_ENABLE_SHIFT       = 4;
 constexpr uint32_t DB_RENDER_CONTROL_RESUMMARIZE_ENABLE_MASK        = 0x1;
 constexpr uint32_t DB_RENDER_CONTROL_STENCIL_COMPRESS_DISABLE_SHIFT = 5;
@@ -163,22 +167,6 @@ constexpr uint32_t PA_SC_SCREEN_SCISSOR_BR_BR_Y_MASK  = 0xFFFF;
 
 constexpr uint32_t DB_DFSM_CONTROL = 0xE;
 
-constexpr uint32_t DB_DEPTH_INFO                          = 0xF;
-constexpr uint32_t DB_DEPTH_INFO_ADDR5_SWIZZLE_MASK_SHIFT = 0;
-constexpr uint32_t DB_DEPTH_INFO_ADDR5_SWIZZLE_MASK_MASK  = 0xF;
-constexpr uint32_t DB_DEPTH_INFO_ARRAY_MODE_SHIFT         = 4;
-constexpr uint32_t DB_DEPTH_INFO_ARRAY_MODE_MASK          = 0xF;
-constexpr uint32_t DB_DEPTH_INFO_PIPE_CONFIG_SHIFT        = 8;
-constexpr uint32_t DB_DEPTH_INFO_PIPE_CONFIG_MASK         = 0x1F;
-constexpr uint32_t DB_DEPTH_INFO_BANK_WIDTH_SHIFT         = 13;
-constexpr uint32_t DB_DEPTH_INFO_BANK_WIDTH_MASK          = 0x3;
-constexpr uint32_t DB_DEPTH_INFO_BANK_HEIGHT_SHIFT        = 15;
-constexpr uint32_t DB_DEPTH_INFO_BANK_HEIGHT_MASK         = 0x3;
-constexpr uint32_t DB_DEPTH_INFO_MACRO_TILE_ASPECT_SHIFT  = 17;
-constexpr uint32_t DB_DEPTH_INFO_MACRO_TILE_ASPECT_MASK   = 0x3;
-constexpr uint32_t DB_DEPTH_INFO_NUM_BANKS_SHIFT          = 19;
-constexpr uint32_t DB_DEPTH_INFO_NUM_BANKS_MASK           = 0x3;
-
 constexpr uint32_t DB_Z_INFO                               = 0x10;
 constexpr uint32_t DB_Z_INFO_FORMAT_SHIFT                  = 0;
 constexpr uint32_t DB_Z_INFO_FORMAT_MASK                   = 0x3;
@@ -222,16 +210,6 @@ constexpr uint32_t DB_STENCIL_READ_BASE  = 0x13;
 constexpr uint32_t DB_Z_WRITE_BASE       = 0x14;
 constexpr uint32_t DB_STENCIL_WRITE_BASE = 0x15;
 
-constexpr uint32_t DB_DEPTH_SIZE                       = 0x16;
-constexpr uint32_t DB_DEPTH_SIZE_PITCH_TILE_MAX_SHIFT  = 0;
-constexpr uint32_t DB_DEPTH_SIZE_PITCH_TILE_MAX_MASK   = 0x7FF;
-constexpr uint32_t DB_DEPTH_SIZE_HEIGHT_TILE_MAX_SHIFT = 11;
-constexpr uint32_t DB_DEPTH_SIZE_HEIGHT_TILE_MAX_MASK  = 0x7FF;
-
-constexpr uint32_t DB_DEPTH_SLICE                      = 0x17;
-constexpr uint32_t DB_DEPTH_SLICE_SLICE_TILE_MAX_SHIFT = 0;
-constexpr uint32_t DB_DEPTH_SLICE_SLICE_TILE_MAX_MASK  = 0x3FFFFF;
-
 constexpr uint32_t DB_Z_READ_BASE_HI                  = 0x1A;
 constexpr uint32_t DB_STENCIL_READ_BASE_HI            = 0x1B;
 constexpr uint32_t DB_Z_WRITE_BASE_HI                 = 0x1C;
@@ -424,11 +402,6 @@ constexpr uint32_t DB_DEPTH_CONTROL_STENCILFUNC_SHIFT                        = 8
 constexpr uint32_t DB_DEPTH_CONTROL_STENCILFUNC_MASK                         = 0x7;
 constexpr uint32_t DB_DEPTH_CONTROL_STENCILFUNC_BF_SHIFT                     = 20;
 constexpr uint32_t DB_DEPTH_CONTROL_STENCILFUNC_BF_MASK                      = 0x7;
-constexpr uint32_t DB_DEPTH_CONTROL_ENABLE_COLOR_WRITES_ON_DEPTH_FAIL_SHIFT  = 30;
-constexpr uint32_t DB_DEPTH_CONTROL_ENABLE_COLOR_WRITES_ON_DEPTH_FAIL_MASK   = 0x1;
-constexpr uint32_t DB_DEPTH_CONTROL_DISABLE_COLOR_WRITES_ON_DEPTH_PASS_SHIFT = 31;
-constexpr uint32_t DB_DEPTH_CONTROL_DISABLE_COLOR_WRITES_ON_DEPTH_PASS_MASK  = 0x1;
-
 constexpr uint32_t DB_EQAA                                  = 0x201;
 constexpr uint32_t DB_EQAA_MAX_ANCHOR_SAMPLES_SHIFT         = 0;
 constexpr uint32_t DB_EQAA_MAX_ANCHOR_SAMPLES_MASK          = 0x7;
@@ -560,21 +533,9 @@ constexpr uint32_t VGT_DRAW_PAYLOAD_CNTL  = 0x2A6;
 constexpr uint32_t VGT_ESGS_RING_ITEMSIZE = 0x2AB;
 constexpr uint32_t VGT_REUSE_OFF          = 0x2AD;
 
-constexpr uint32_t DB_HTILE_SURFACE                               = 0x2AF;
-constexpr uint32_t DB_HTILE_SURFACE_LINEAR_SHIFT                  = 0;
-constexpr uint32_t DB_HTILE_SURFACE_LINEAR_MASK                   = 0x1;
-constexpr uint32_t DB_HTILE_SURFACE_FULL_CACHE_SHIFT              = 1;
-constexpr uint32_t DB_HTILE_SURFACE_FULL_CACHE_MASK               = 0x1;
-constexpr uint32_t DB_HTILE_SURFACE_HTILE_USES_PRELOAD_WIN_SHIFT  = 2;
-constexpr uint32_t DB_HTILE_SURFACE_HTILE_USES_PRELOAD_WIN_MASK   = 0x1;
-constexpr uint32_t DB_HTILE_SURFACE_PRELOAD_SHIFT                 = 3;
-constexpr uint32_t DB_HTILE_SURFACE_PRELOAD_MASK                  = 0x1;
-constexpr uint32_t DB_HTILE_SURFACE_PREFETCH_WIDTH_SHIFT          = 4;
-constexpr uint32_t DB_HTILE_SURFACE_PREFETCH_WIDTH_MASK           = 0x3F;
-constexpr uint32_t DB_HTILE_SURFACE_PREFETCH_HEIGHT_SHIFT         = 10;
-constexpr uint32_t DB_HTILE_SURFACE_PREFETCH_HEIGHT_MASK          = 0x3F;
-constexpr uint32_t DB_HTILE_SURFACE_DST_OUTSIDE_ZERO_TO_ONE_SHIFT = 16;
-constexpr uint32_t DB_HTILE_SURFACE_DST_OUTSIDE_ZERO_TO_ONE_MASK  = 0x1;
+constexpr uint32_t DB_SHADING_RATE_ENCODING                             = 0x2AF;
+constexpr uint32_t DB_SHADING_RATE_ENCODING_SHADING_RATE_ENCODING_SHIFT = 19;
+constexpr uint32_t DB_SHADING_RATE_ENCODING_SHADING_RATE_ENCODING_MASK  = 0x3;
 
 constexpr uint32_t DB_SRESULTS_COMPARE_STATE0                          = 0x2B0;
 constexpr uint32_t DB_SRESULTS_COMPARE_STATE1                          = 0x2B1;
@@ -621,7 +582,6 @@ constexpr uint32_t PA_SC_SHADER_CONTROL                  = 0x310;
 constexpr uint32_t PA_SC_BINNER_CNTL_0                   = 0x311;
 constexpr uint32_t PA_SC_BINNER_CNTL_1                   = 0x312;
 constexpr uint32_t PA_SC_CONSERVATIVE_RASTERIZATION_CNTL = 0x313;
-constexpr uint32_t PA_SC_NGG_MODE_CNTL                   = 0x314;
 constexpr uint32_t CB_COLOR0_BASE                        = 0x318;
 
 constexpr uint32_t CB_COLOR0_VIEW                   = 0x31B;
@@ -649,24 +609,13 @@ constexpr uint32_t CB_COLOR0_INFO_BLEND_BYPASS_SHIFT              = 16;
 constexpr uint32_t CB_COLOR0_INFO_BLEND_BYPASS_MASK               = 0x1;
 constexpr uint32_t CB_COLOR0_INFO_ROUND_MODE_SHIFT                = 18;
 constexpr uint32_t CB_COLOR0_INFO_ROUND_MODE_MASK                 = 0x1;
-constexpr uint32_t CB_COLOR0_INFO_CMASK_IS_LINEAR_SHIFT           = 19;
-constexpr uint32_t CB_COLOR0_INFO_CMASK_IS_LINEAR_MASK            = 0x1;
 constexpr uint32_t CB_COLOR0_INFO_FMASK_COMPRESSION_DISABLE_SHIFT = 26;
 constexpr uint32_t CB_COLOR0_INFO_FMASK_COMPRESSION_DISABLE_MASK  = 0x1;
 constexpr uint32_t CB_COLOR0_INFO_FMASK_COMPRESS_1FRAG_ONLY_SHIFT = 27;
 constexpr uint32_t CB_COLOR0_INFO_FMASK_COMPRESS_1FRAG_ONLY_MASK  = 0x1;
 constexpr uint32_t CB_COLOR0_INFO_DCC_ENABLE_SHIFT                = 28;
 constexpr uint32_t CB_COLOR0_INFO_DCC_ENABLE_MASK                 = 0x1;
-constexpr uint32_t CB_COLOR0_INFO_CMASK_ADDR_TYPE_SHIFT           = 29;
-constexpr uint32_t CB_COLOR0_INFO_CMASK_ADDR_TYPE_MASK            = 0x3;
-constexpr uint32_t CB_COLOR0_INFO_ALT_TILE_MODE_SHIFT             = 31;
-constexpr uint32_t CB_COLOR0_INFO_ALT_TILE_MODE_MASK              = 0x1;
-
 constexpr uint32_t CB_COLOR0_ATTRIB                             = 0x31D;
-constexpr uint32_t CB_COLOR0_ATTRIB_TILE_MODE_INDEX_SHIFT       = 0;
-constexpr uint32_t CB_COLOR0_ATTRIB_TILE_MODE_INDEX_MASK        = 0x1F;
-constexpr uint32_t CB_COLOR0_ATTRIB_FMASK_TILE_MODE_INDEX_SHIFT = 5;
-constexpr uint32_t CB_COLOR0_ATTRIB_FMASK_TILE_MODE_INDEX_MASK  = 0x1F;
 constexpr uint32_t CB_COLOR0_ATTRIB_NUM_SAMPLES_SHIFT           = 12;
 constexpr uint32_t CB_COLOR0_ATTRIB_NUM_SAMPLES_MASK            = 0x7;
 constexpr uint32_t CB_COLOR0_ATTRIB_NUM_FRAGMENTS_SHIFT         = 15;
@@ -681,8 +630,6 @@ constexpr uint32_t CB_COLOR0_DCC_CONTROL_KEY_CLEAR_ENABLE_SHIFT
 constexpr uint32_t CB_COLOR0_DCC_CONTROL_KEY_CLEAR_ENABLE_MASK                  = 0x1;
 constexpr uint32_t CB_COLOR0_DCC_CONTROL_MAX_UNCOMPRESSED_BLOCK_SIZE_SHIFT      = 2;
 constexpr uint32_t CB_COLOR0_DCC_CONTROL_MAX_UNCOMPRESSED_BLOCK_SIZE_MASK       = 0x3;
-constexpr uint32_t CB_COLOR0_DCC_CONTROL_MIN_COMPRESSED_BLOCK_SIZE_SHIFT        = 4;
-constexpr uint32_t CB_COLOR0_DCC_CONTROL_MIN_COMPRESSED_BLOCK_SIZE_MASK         = 0x1;
 constexpr uint32_t CB_COLOR0_DCC_CONTROL_MAX_COMPRESSED_BLOCK_SIZE_SHIFT        = 5;
 constexpr uint32_t CB_COLOR0_DCC_CONTROL_MAX_COMPRESSED_BLOCK_SIZE_MASK         = 0x3;
 constexpr uint32_t CB_COLOR0_DCC_CONTROL_COLOR_TRANSFORM_SHIFT                  = 7;
@@ -739,31 +686,25 @@ constexpr uint32_t CB_COLOR0_ATTRIB3_CMASK_PIPE_ALIGNED_SHIFT = 26;
 constexpr uint32_t CB_COLOR0_ATTRIB3_CMASK_PIPE_ALIGNED_MASK  = 0x1;
 constexpr uint32_t CB_COLOR0_ATTRIB3_DCC_PIPE_ALIGNED_SHIFT   = 30;
 constexpr uint32_t CB_COLOR0_ATTRIB3_DCC_PIPE_ALIGNED_MASK    = 0x1;
+constexpr uint32_t CB_COLOR0_ATTRIB3_WRITE_VRS_RATE_HINT_TO_CMASK_SHIFT = 31;
+constexpr uint32_t CB_COLOR0_ATTRIB3_WRITE_VRS_RATE_HINT_TO_CMASK_MASK  = 0x1;
 
 constexpr uint32_t CB_COLOR7_ATTRIB3 = 0x3BF;
 
-/* Fake codes. Don't exist on real HW */
-
-constexpr uint32_t CX_NOP = 0x800003FF;
+constexpr uint32_t CX_NOP = 0xDB;
 
 constexpr uint32_t CX_NUM = 0x3FF + 1;
 
 /* Shader registers */
 
-constexpr uint32_t SPI_SHADER_TBA_LO_PS     = 0x0;
-constexpr uint32_t SPI_SHADER_PGM_RSRC4_PS  = 0x1;
-constexpr uint32_t SPI_SHADER_TMA_LO_PS     = 0x2;
-constexpr uint32_t SPI_SHADER_TMA_HI_PS     = 0x3;
-constexpr uint32_t SPI_SHADER_PGM_CHKSUM_PS = 0x6;
-constexpr uint32_t SPI_SHADER_PGM_RSRC3_PS  = 0x7;
-constexpr uint32_t SPI_SHADER_PGM_LO_PS     = 0x8;
-constexpr uint32_t SPI_SHADER_PGM_HI_PS     = 0x9;
+constexpr uint32_t SPI_SHADER_PACE_ID_PS            = 0x6;
+constexpr uint32_t SPI_GRAPHICS_SHADER_CONTROL_PS    = 0x7;
+constexpr uint32_t SPI_SHADER_PGM_LO_PS              = 0x8;
+constexpr uint32_t SPI_SHADER_PGM_HI_PS              = 0x9;
 
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS                        = 0xA;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS_VGPRS_SHIFT            = 0;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS_VGPRS_MASK             = 0x3F;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS_SGPRS_SHIFT            = 6;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS_SGPRS_MASK             = 0xF;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS_PRIORITY_SHIFT         = 10;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS_PRIORITY_MASK          = 0x3;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_PS_FLOAT_MODE_SHIFT       = 12;
@@ -800,14 +741,11 @@ constexpr uint32_t SPI_SHADER_PGM_RSRC2_PS_SHARED_VGPR_CNT_MASK         = 0xF;
 constexpr uint32_t SPI_SHADER_USER_DATA_PS_0  = 0xC;
 constexpr uint32_t SPI_SHADER_USER_DATA_PS_15 = 0x1B;
 constexpr uint32_t SPI_SHADER_USER_DATA_PS_31 = 0x2B;
-constexpr uint32_t SPI_SHADER_REQ_CTRL_PS     = 0x30;
 constexpr uint32_t SPI_SHADER_USER_ACCUM_PS_0 = 0x32;
 constexpr uint32_t SPI_SHADER_USER_ACCUM_PS_3 = 0x35;
 
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_VS_VGPRS_SHIFT           = 0;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_VS_VGPRS_MASK            = 0x3F;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_VS_SGPRS_SHIFT           = 6;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_VS_SGPRS_MASK            = 0xF;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_VS_PRIORITY_SHIFT        = 10;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_VS_PRIORITY_MASK         = 0x3;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_VS_FLOAT_MODE_SHIFT      = 12;
@@ -838,19 +776,20 @@ constexpr uint32_t SPI_SHADER_PGM_RSRC2_VS_USER_SGPR_MSB_MASK    = 0x1;
 constexpr uint32_t SPI_SHADER_PGM_RSRC2_VS_SHARED_VGPR_CNT_SHIFT = 28;
 constexpr uint32_t SPI_SHADER_PGM_RSRC2_VS_SHARED_VGPR_CNT_MASK  = 0xF;
 
-constexpr uint32_t SPI_SHADER_PGM_CHKSUM_GS        = 0x80;
+constexpr uint32_t SPI_SHADER_PACE_ID_GS           = 0x80;
+// AGC compiler records use this numeric tag as opaque metadata. It is not a shader checksum
+// register and intentionally has no checksum-state consumer in HardwareContext.
+constexpr uint32_t SPI_SHADER_PGM_CHKSUM_GS        = SPI_SHADER_PACE_ID_GS;
 constexpr uint32_t SPI_SHADER_PGM_RSRC4_GS         = 0x81;
 constexpr uint32_t SPI_SHADER_USER_DATA_ADDR_LO_GS = 0x82;
 constexpr uint32_t SPI_SHADER_USER_DATA_ADDR_HI_GS = 0x83;
-constexpr uint32_t SPI_SHADER_PGM_RSRC3_GS         = 0x87;
+constexpr uint32_t SPI_GRAPHICS_SHADER_CONTROL_GS  = 0x87;
 constexpr uint32_t SPI_SHADER_PGM_LO_GS            = 0x88;
 constexpr uint32_t SPI_SHADER_PGM_HI_GS            = 0x89;
 
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS                        = 0x8A;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS_VGPRS_SHIFT            = 0;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS_VGPRS_MASK             = 0x3F;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS_SGPRS_SHIFT            = 6;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS_SGPRS_MASK             = 0xF;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS_PRIORITY_SHIFT         = 10;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS_PRIORITY_MASK          = 0x3;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_GS_FLOAT_MODE_SHIFT       = 12;
@@ -891,22 +830,16 @@ constexpr uint32_t SPI_SHADER_PGM_RSRC2_GS_SHARED_VGPR_CNT_MASK   = 0xF;
 constexpr uint32_t SPI_SHADER_USER_DATA_GS_0                      = 0x8C;
 constexpr uint32_t SPI_SHADER_USER_DATA_GS_15                     = 0x9B;
 constexpr uint32_t SPI_SHADER_USER_DATA_GS_31                     = 0xAB;
-constexpr uint32_t SPI_SHADER_REQ_CTRL_ESGS                       = 0xB0;
 constexpr uint32_t SPI_SHADER_USER_ACCUM_ESGS_0                   = 0xB2;
 constexpr uint32_t SPI_SHADER_USER_ACCUM_ESGS_3                   = 0xB5;
-constexpr uint32_t SPI_SHADER_PGM_RSRC2_ES_GS                     = 0xBC;
 constexpr uint32_t SPI_SHADER_PGM_LO_ES                           = 0xC8;
 constexpr uint32_t SPI_SHADER_PGM_HI_ES                           = 0xC9;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_ES                        = 0xCA;
-constexpr uint32_t SPI_SHADER_PGM_RSRC2_ES                        = 0xCB;
-constexpr uint32_t SPI_SHADER_USER_DATA_ES_0                      = 0xCC;
-constexpr uint32_t SPI_SHADER_USER_DATA_ES_15                     = 0xDB;
-constexpr uint32_t SPI_SHADER_USER_DATA_ES_31                     = 0xEB;
+// Private AGC registers used by shader binaries submitted through SET_SH_REG_INDIRECT.
 constexpr uint32_t SPI_SHADER_PGM_CHKSUM_HS                       = 0x100;
 constexpr uint32_t SPI_SHADER_PGM_RSRC4_HS                        = 0x101;
 constexpr uint32_t SPI_SHADER_USER_DATA_ADDR_LO_HS                = 0x102;
 constexpr uint32_t SPI_SHADER_USER_DATA_ADDR_HI_HS                = 0x103;
-constexpr uint32_t SPI_SHADER_PGM_RSRC3_HS                        = 0x107;
+constexpr uint32_t SPI_GRAPHICS_SHADER_CONTROL_HS                 = 0x107;
 constexpr uint32_t SPI_SHADER_PGM_LO_HS                           = 0x108;
 constexpr uint32_t SPI_SHADER_PGM_HI_HS                           = 0x109;
 constexpr uint32_t SPI_SHADER_PGM_RSRC1_HS                        = 0x10A;
@@ -944,13 +877,10 @@ constexpr uint32_t SPI_SHADER_PGM_RSRC2_HS_SHARED_VGPR_CNT_MASK   = 0xF;
 constexpr uint32_t SPI_SHADER_USER_DATA_HS_0                      = 0x10C;
 constexpr uint32_t SPI_SHADER_USER_DATA_HS_15                     = 0x11B;
 constexpr uint32_t SPI_SHADER_USER_DATA_HS_31                     = 0x12B;
-constexpr uint32_t SPI_SHADER_REQ_CTRL_LSHS                       = 0x130;
 constexpr uint32_t SPI_SHADER_USER_ACCUM_LSHS_0                   = 0x132;
 constexpr uint32_t SPI_SHADER_USER_ACCUM_LSHS_3                   = 0x135;
 constexpr uint32_t SPI_SHADER_PGM_LO_LS                           = 0x148;
 constexpr uint32_t SPI_SHADER_PGM_HI_LS                           = 0x149;
-constexpr uint32_t SPI_SHADER_PGM_RSRC1_LS                        = 0x14A;
-constexpr uint32_t SPI_SHADER_PGM_RSRC2_LS                        = 0x14B;
 constexpr uint32_t COMPUTE_START_X                                = 0x204;
 constexpr uint32_t COMPUTE_START_Y                                = 0x205;
 constexpr uint32_t COMPUTE_START_Z                                = 0x206;
@@ -959,28 +889,26 @@ constexpr uint32_t COMPUTE_NUM_THREAD_Y                           = 0x208;
 constexpr uint32_t COMPUTE_NUM_THREAD_Z                           = 0x209;
 constexpr uint32_t COMPUTE_PGM_LO                                 = 0x20C;
 constexpr uint32_t COMPUTE_PGM_HI                                 = 0x20D;
-constexpr uint32_t COMPUTE_TBA_LO                                 = 0x20E;
-constexpr uint32_t COMPUTE_TBA_HI                                 = 0x20F;
-constexpr uint32_t COMPUTE_TMA_LO                                 = 0x210;
-constexpr uint32_t COMPUTE_TMA_HI                                 = 0x211;
 
 constexpr uint32_t COMPUTE_PGM_RSRC1                  = 0x212;
 constexpr uint32_t COMPUTE_PGM_RSRC1_VGPRS_SHIFT      = 0;
 constexpr uint32_t COMPUTE_PGM_RSRC1_VGPRS_MASK       = 0x3F;
-constexpr uint32_t COMPUTE_PGM_RSRC1_SGPRS_SHIFT      = 6;
-constexpr uint32_t COMPUTE_PGM_RSRC1_SGPRS_MASK       = 0xF;
+constexpr uint32_t COMPUTE_PGM_RSRC1_PRIORITY_SHIFT   = 10;
+constexpr uint32_t COMPUTE_PGM_RSRC1_PRIORITY_MASK    = 0x3;
 constexpr uint32_t COMPUTE_PGM_RSRC1_FLOAT_MODE_SHIFT = 12;
 constexpr uint32_t COMPUTE_PGM_RSRC1_FLOAT_MODE_MASK  = 0xFF;
 constexpr uint32_t COMPUTE_PGM_RSRC1_DX10_CLAMP_SHIFT = 21;
 constexpr uint32_t COMPUTE_PGM_RSRC1_DX10_CLAMP_MASK  = 0x1;
+constexpr uint32_t COMPUTE_PGM_RSRC1_DEBUG_MODE_SHIFT = 22;
+constexpr uint32_t COMPUTE_PGM_RSRC1_DEBUG_MODE_MASK  = 0x1;
 constexpr uint32_t COMPUTE_PGM_RSRC1_IEEE_MODE_SHIFT  = 23;
 constexpr uint32_t COMPUTE_PGM_RSRC1_IEEE_MODE_MASK   = 0x1;
-constexpr uint32_t COMPUTE_PGM_RSRC1_BULKY_SHIFT      = 24;
-constexpr uint32_t COMPUTE_PGM_RSRC1_BULKY_MASK       = 0x1;
 constexpr uint32_t COMPUTE_PGM_RSRC1_FP16_OVFL_SHIFT  = 26;
 constexpr uint32_t COMPUTE_PGM_RSRC1_FP16_OVFL_MASK   = 0x1;
 constexpr uint32_t COMPUTE_PGM_RSRC1_WGP_MODE_SHIFT   = 29;
 constexpr uint32_t COMPUTE_PGM_RSRC1_WGP_MODE_MASK    = 0x1;
+constexpr uint32_t COMPUTE_PGM_RSRC1_FWD_PROGRESS_SHIFT = 31;
+constexpr uint32_t COMPUTE_PGM_RSRC1_FWD_PROGRESS_MASK  = 0x1;
 
 constexpr uint32_t COMPUTE_DISPATCH_INITIATOR_CS_W32_EN_SHIFT = 15;
 constexpr uint32_t COMPUTE_DISPATCH_INITIATOR_CS_W32_EN_MASK  = 0x1;
@@ -1011,21 +939,16 @@ constexpr uint32_t COMPUTE_PGM_RSRC2_LDS_SIZE_SHIFT       = 15;
 constexpr uint32_t COMPUTE_PGM_RSRC2_LDS_SIZE_MASK        = 0x1FF;
 
 constexpr uint32_t COMPUTE_RESOURCE_LIMITS    = 0x215;
-constexpr uint32_t COMPUTE_DESTINATION_EN_SE0 = 0x216;
-constexpr uint32_t COMPUTE_DESTINATION_EN_SE1 = 0x217;
 constexpr uint32_t COMPUTE_TMPRING_SIZE       = 0x218;
-constexpr uint32_t COMPUTE_DESTINATION_EN_SE2 = 0x219;
-constexpr uint32_t COMPUTE_DESTINATION_EN_SE3 = 0x21A;
 constexpr uint32_t COMPUTE_USER_ACCUM_0       = 0x224;
 constexpr uint32_t COMPUTE_PGM_RSRC3          = 0x228;
-constexpr uint32_t COMPUTE_SHADER_CHKSUM      = 0x22A;
+constexpr uint32_t COMPUTE_PGM_RSRC3_SHARED_VGPRS_SHIFT = 0;
+constexpr uint32_t COMPUTE_PGM_RSRC3_SHARED_VGPRS_MASK  = 0xF;
+constexpr uint32_t COMPUTE_PACE_ID             = 0x22A;
 constexpr uint32_t COMPUTE_USER_DATA_0        = 0x240;
 constexpr uint32_t COMPUTE_USER_DATA_15       = 0x24F;
-constexpr uint32_t COMPUTE_DISPATCH_TUNNEL    = 0x27D;
 
-/* Fake codes. Don't exist on real HW */
-
-constexpr uint32_t SH_NOP = 0x800002FF;
+constexpr uint32_t SH_NOP = 0x280;
 
 constexpr uint32_t SH_NUM = 0x2FF + 1;
 
@@ -1042,6 +965,7 @@ constexpr uint32_t GE_INDX_OFFSET            = 0x24A;
 constexpr uint32_t GE_MULTI_PRIM_IB_RESET_EN = 0x24B;
 constexpr uint32_t VGT_HS_OFFCHIP_PARAM      = 0x24F;
 constexpr uint32_t VGT_TF_MEMORY_BASE        = 0x250;
+constexpr uint32_t IA_MULTI_VGT_PARAM        = 0x258;
 
 constexpr uint32_t GE_CNTL                     = 0x25B;
 constexpr uint32_t GE_CNTL_PRIM_GRP_SIZE_SHIFT = 0;
@@ -1053,7 +977,7 @@ constexpr uint32_t GE_USER_VGPR1  = 0x25C;
 constexpr uint32_t GE_USER_VGPR2  = 0x25D;
 constexpr uint32_t GE_USER_VGPR3  = 0x25E;
 constexpr uint32_t GE_STEREO_CNTL = 0x25F;
-constexpr uint32_t GE_PC_ALLOC    = 0x260;
+constexpr uint32_t UC_PARAMETER_OVERSUBSCRIPTION = 0x260;
 
 constexpr uint32_t GE_USER_VGPR_EN                     = 0x262;
 constexpr uint32_t GE_USER_VGPR_EN_EN_USER_VGPR1_SHIFT = 0;
@@ -1063,28 +987,15 @@ constexpr uint32_t GE_USER_VGPR_EN_EN_USER_VGPR2_MASK  = 0x1;
 constexpr uint32_t GE_USER_VGPR_EN_EN_USER_VGPR3_SHIFT = 2;
 constexpr uint32_t GE_USER_VGPR_EN_EN_USER_VGPR3_MASK  = 0x1;
 
-constexpr uint32_t IA_MULTI_VGT_PARAM       = 0x258;
 constexpr uint32_t TA_CS_BC_BASE_ADDR       = 0x380;
 constexpr uint32_t TA_CS_BC_BASE_ADDR_HI    = 0x381;
-constexpr uint32_t TEXTURE_GRADIENT_FACTORS = 0x382;
+constexpr uint32_t TEXTURE_GRADIENT_FACTORS = 0x383;
+constexpr uint32_t TEXTURE_GRADIENT_CONTROL = 0x384;
 constexpr uint32_t GDS_OA_CNTL              = 0x41D;
 constexpr uint32_t GDS_OA_COUNTER           = 0x41E;
 constexpr uint32_t GDS_OA_ADDRESS           = 0x41F;
 
-/* Fake codes. Don't exist on real HW */
-
-constexpr uint32_t FSR_EXTEND_SUBPIXEL_ROUNDING = 0x80003FF4;
-constexpr uint32_t FSR_ALPHA_VALUE0             = 0x80003FF5;
-constexpr uint32_t FSR_ALPHA_VALUE1             = 0x80003FF6;
-constexpr uint32_t FSR_CONTROL_POINT0           = 0x80003FF7;
-constexpr uint32_t FSR_CONTROL_POINT1           = 0x80003FF8;
-constexpr uint32_t FSR_CONTROL_POINT2           = 0x80003FF9;
-constexpr uint32_t FSR_CONTROL_POINT3           = 0x80003FFA;
-constexpr uint32_t FSR_WINDOW0                  = 0x80003FFB;
-constexpr uint32_t FSR_WINDOW1                  = 0x80003FFC;
-constexpr uint32_t TEXTURE_GRADIENT_CONTROL     = 0x80003FFD;
-constexpr uint32_t MEMORY_MAPPING_MASK          = 0x80003FFE;
-constexpr uint32_t UC_NOP                       = 0x80003FFF;
+constexpr uint32_t UC_NOP = 0xA2;
 
 constexpr uint32_t UC_NUM = 0x3FFF + 1;
 

src/graphics/host_gpu/memoryTracker.cpp

@@ -18,8 +18,8 @@ MemoryTracker::~MemoryTracker() = default;
 #if KYTY_BUILD == KYTY_BUILD_DEBUG
 void MemoryTracker::ValidateGpuDirtyPages(const RangeSet& dirty, uint64_t vaddr, uint64_t size,
                                           const char* operation) const noexcept {
-	if (vaddr == 0 || size == 0 || size > UINT64_MAX - vaddr ||
-	    (vaddr & (TRACKER_PAGE_SIZE - 1)) != 0 || (size & (TRACKER_PAGE_SIZE - 1)) != 0) {
+	if (!GuestRange {vaddr, size}.Valid() || (vaddr & (TRACKER_PAGE_SIZE - 1)) != 0 ||
+	    (size & (TRACKER_PAGE_SIZE - 1)) != 0) {
 		EXIT("MemoryTracker: invalid dirty-page validation range\n");
 	}
 	for (auto page = vaddr; page < vaddr + size; page += TRACKER_PAGE_SIZE) {
@@ -37,9 +37,6 @@ void MemoryTracker::ValidateGpuDirtyPages(const RangeSet& dirty, uint64_t vaddr,
 void MemoryTracker::ValidateGpuDirtyOwnership(const RangeSet& dirty, uint64_t vaddr, uint64_t size,
                                               const char* operation) {
 	ValidateRange(vaddr, size);
-	if (vaddr + size > UINT64_MAX - (TRACKER_PAGE_SIZE - 1)) {
-		EXIT("MemoryTracker: dirty ownership range alignment overflow\n");
-	}
 	const auto begin = vaddr & ~(TRACKER_PAGE_SIZE - 1);
 	const auto end   = (vaddr + size + TRACKER_PAGE_SIZE - 1) & ~(TRACKER_PAGE_SIZE - 1);
 	for (auto page = begin; page < end; page += TRACKER_PAGE_SIZE) {
@@ -56,8 +53,7 @@ void MemoryTracker::ValidateGpuDirtyOwnership(const RangeSet& dirty, uint64_t va
 #endif
 
 void MemoryTracker::ValidateRange(uint64_t vaddr, uint64_t size) {
-	if (vaddr == 0 || size == 0 || vaddr >= TRACKER_ADDRESS_SIZE ||
-	    size > TRACKER_ADDRESS_SIZE - vaddr) {
+	if (!GuestRange {vaddr, size}.Valid()) {
 		EXIT("invalid memory tracker range\n");
 	}
 }

src/graphics/host_gpu/pageManager.cpp

@@ -110,7 +110,7 @@ private:
 };
 
 void ValidateRange(uint64_t vaddr, uint64_t size) {
-	if (vaddr == 0 || size == 0 || vaddr >= ADDRESS_SIZE || size > ADDRESS_SIZE - vaddr) {
+	if (!GuestRange {vaddr, size}.Valid()) {
 		Fatal("invalid range vaddr=0x%016" PRIx64 ", size=0x%016" PRIx64, vaddr, size);
 	}
 }

src/graphics/host_gpu/regionDefinitions.h

@@ -4,6 +4,8 @@
 #include "common/bitArray.h"
 #include "common/common.h"
 
+#include <compare>
+
 namespace Libs::Graphics {
 
 constexpr uint64_t TRACKER_PAGE_SIZE    = 4ull * 1024ull;
@@ -11,6 +13,20 @@ constexpr uint64_t TRACKER_REGION_SIZE  = 4ull * 1024ull * 1024ull;
 constexpr uint64_t TRACKER_ADDRESS_SIZE = 1ull << 40u;
 constexpr size_t   TRACKER_REGION_PAGES = TRACKER_REGION_SIZE / TRACKER_PAGE_SIZE;
 
+struct GuestRange {
+	uint64_t address = 0;
+	uint64_t size    = 0;
+
+	[[nodiscard]] constexpr bool Empty() const noexcept { return address == 0 && size == 0; }
+	[[nodiscard]] constexpr bool Valid() const noexcept {
+		return address != 0 && size != 0 && address < TRACKER_ADDRESS_SIZE &&
+		       size <= TRACKER_ADDRESS_SIZE - address;
+	}
+	[[nodiscard]] constexpr bool     ValidOrEmpty() const noexcept { return Empty() || Valid(); }
+	[[nodiscard]] constexpr uint64_t End() const noexcept { return address + size; }
+	auto                             operator<=>(const GuestRange&) const = default;
+};
+
 enum class DirtySource { Cpu, Gpu };
 using RegionBits = Common::BitArray<TRACKER_REGION_PAGES>;
 static_assert(sizeof(RegionBits) == TRACKER_REGION_PAGES / 8);

src/graphics/host_gpu/renderer/cache/bufferCache.cpp

@@ -241,8 +241,7 @@ BufferCache::~BufferCache() {
 }
 
 void BufferCache::InvalidateMemory(uint64_t vaddr, uint64_t size) {
-	if (vaddr == 0 || size == 0 || vaddr >= TRACKER_ADDRESS_SIZE ||
-	    size > TRACKER_ADDRESS_SIZE - vaddr) {
+	if (!GuestRange {vaddr, size}.Valid()) {
 		EXIT("BufferCache: invalid memory-invalidation range\n");
 	}
 	m_memory_tracker.InvalidateRegion(vaddr, size,
@@ -314,7 +313,7 @@ BufferId BufferCache::FindBuffer(uint64_t vaddr, uint64_t size) {
 	if (vaddr == 0) {
 		return NULL_BUFFER_ID;
 	}
-	if (size == 0 || vaddr >= TRACKER_ADDRESS_SIZE || size > TRACKER_ADDRESS_SIZE - vaddr) {
+	if (!GuestRange {vaddr, size}.Valid()) {
 		EXIT("BufferCache: invalid buffer discovery request\n");
 	}
 	const auto* owner = m_page_table.Find(vaddr >> PageTable::kPageBits);
@@ -443,8 +442,7 @@ std::pair<Buffer*, uint64_t> BufferCache::ObtainBuffer(uint64_t vaddr, uint64_t
                                                        bool is_written, bool is_texel_buffer,
                                                        BufferId id) {
 	auto& command = m_scheduler.Current();
-	if (command.IsInvalid() || vaddr == 0 || size == 0 ||
-	    vaddr >= TRACKER_ADDRESS_SIZE || size > TRACKER_ADDRESS_SIZE - vaddr) {
+	if (command.IsInvalid() || !GuestRange {vaddr, size}.Valid()) {
 		EXIT("BufferCache: buffer request requires a recording command buffer\n");
 	}
 
@@ -474,8 +472,7 @@ std::pair<Buffer*, uint64_t> BufferCache::ObtainBuffer(uint64_t vaddr, uint64_t
 }
 
 std::pair<Buffer*, uint64_t> BufferCache::ObtainBufferForImage(uint64_t vaddr, uint64_t size) {
-	if (vaddr == 0 || size == 0 || vaddr >= TRACKER_ADDRESS_SIZE ||
-	    size > TRACKER_ADDRESS_SIZE - vaddr) {
+	if (!GuestRange {vaddr, size}.Valid()) {
 		EXIT("BufferCache: invalid image source\n");
 	}
 	auto find_owner = [&]() -> Buffer* {
@@ -661,8 +658,7 @@ void BufferCache::CopyBuffer(uint64_t dst_vaddr, uint64_t src_vaddr, uint64_t si
 }
 
 bool BufferCache::IsRegionRegistered(uint64_t vaddr, uint64_t size) {
-	if (vaddr == 0 || size == 0 || vaddr >= TRACKER_ADDRESS_SIZE ||
-	    size > TRACKER_ADDRESS_SIZE - vaddr) {
+	if (!GuestRange {vaddr, size}.Valid()) {
 		EXIT("BufferCache: invalid registered-region query\n");
 	}
 	// Cached buffers are ordered and non-overlapping. The last buffer beginning before the query

src/graphics/host_gpu/renderer/cache/gpuResourceManager.cpp

@@ -35,8 +35,7 @@ bool GpuResourceManager::InvalidateMemory(uint64_t vaddr, uint64_t size) {
 }
 
 bool GpuResourceManager::IsMapped(uint64_t vaddr, uint64_t size) const noexcept {
-	if (vaddr == 0 || size == 0 || vaddr >= TRACKER_ADDRESS_SIZE ||
-	    size > TRACKER_ADDRESS_SIZE - vaddr) {
+	if (!GuestRange {vaddr, size}.Valid()) {
 		return false;
 	}
 	std::shared_lock lock(m_mapped_ranges_mutex);

src/graphics/host_gpu/renderer/colorRenderTarget.cpp

@@ -204,13 +204,11 @@ void RenderExecutor::ResolveRenderColorTarget(uint64_t submit_id, CommandBuffer&
 	if (texture_tile &&
 	    (!TileGetTextureBlockLayout(transfer_format, rt.attrib3.tile_mode, volume,
 	                                texture_tile_layout) ||
-	     rt.pitch.pitch_div8_minus1 != 0 ||
 	     (rt.base.addr & (texture_tile_layout.block.block_size - 1u)) != 0 ||
 	     rt.info.fmask_compression_enable || rt.info.fmask_data_compression_disable ||
 	     rt.info.fmask_one_frag_mode || rt.info.cmask_fast_clear_enable ||
-	     rt.info.dcc_compression_enable || rt.info.cmask_is_linear != 0 ||
-	     rt.info.cmask_addr_type != 0 || rt.info.alt_tile_mode || rt.cmask.addr != 0 ||
-	     rt.fmask.addr != 0 || rt.dcc_addr.addr != 0 || rt.dcc.data_write_on_dcc_clear_to_reg)) {
+	     rt.info.dcc_compression_enable || rt.cmask.addr != 0 || rt.fmask.addr != 0 ||
+	     rt.dcc_addr.addr != 0 || rt.dcc.data_write_on_dcc_clear_to_reg)) {
 		EXIT("unsupported texture-tiled render target: addr=0x%016" PRIx64 " tile=%u"
 		     " dimension=%u depth=%u levels=%u layer=%u/%u samples=%u fragments=%u bpe=%u"
 		     " cmask=0x%016" PRIx64 " fmask=0x%016" PRIx64 " dcc=0x%016" PRIx64 "\n",
@@ -219,15 +217,14 @@ void RenderExecutor::ResolveRenderColorTarget(uint64_t submit_id, CommandBuffer&
 		     rt.attrib.num_fragments, bytes_per_element, rt.cmask.addr, rt.fmask.addr,
 		     rt.dcc_addr.addr);
 	}
-	if ((standard64 || depth_tile) && (rt.attrib3.dimension != 1 || rt.attrib3.depth != 0 ||
-	                                   view.base_layer != 0 || view.image_layers != 1)) {
+	if ((standard64 || depth_tile) &&
+	    (rt.attrib3.dimension != 1 || rt.attrib3.depth != 0 ||
+	     (depth_tile && (view.base_layer != 0 || view.image_layers != 1)))) {
 		EXIT("unsupported 64KB texture-tiled render-target view: dimension=%u depth=%u"
 		     " layer=%u/%u\n",
 		     rt.attrib3.dimension, rt.attrib3.depth, view.base_layer, view.image_layers);
 	}
-	if (rt.pitch.pitch_div8_minus1 != 0) {
-		pitch = (rt.pitch.pitch_div8_minus1 + 1u) << 3u;
-	} else if (tile) {
+	if (tile) {
 		if (volume) {
 			pitch = TileGetTexturePitch(transfer_format, width, rt.attrib3.tile_mode);
 		} else if (texture_tile) {
@@ -293,12 +290,6 @@ void RenderExecutor::ResolveRenderColorTarget(uint64_t submit_id, CommandBuffer&
 		mip_sizes[0]  = {static_cast<uint32_t>(size), 0, 0, 0, 0, 0};
 		mip_padded[0] = {pitch, height};
 	}
-	if (rt.slice.slice_div64_minus1 != 0 &&
-	    (static_cast<uint64_t>(rt.slice.slice_div64_minus1) + 1u) * 64u != size) {
-		EXIT("render-target slice span mismatch: encoded=0x%016" PRIx64 " derived=0x%016" PRIx64
-		     "\n",
-		     (static_cast<uint64_t>(rt.slice.slice_div64_minus1) + 1u) * 64u, size);
-	}
 	if (size == 0 || (!volume && size > UINT64_MAX / view.image_layers)) {
 		EXIT("render-target memory footprint is invalid\n");
 	}
@@ -308,7 +299,7 @@ void RenderExecutor::ResolveRenderColorTarget(uint64_t submit_id, CommandBuffer&
 	if (backing_size == 0) {
 		EXIT("render-target backing is empty\n");
 	}
-	if (backing_size > TRACKER_ADDRESS_SIZE - rt.base.addr) {
+	if (!GuestRange {rt.base.addr, backing_size}.Valid()) {
 		EXIT("render-target backing range is invalid\n");
 	}
 

src/graphics/host_gpu/renderer/debug.cpp

@@ -97,17 +97,11 @@ void sh_print(const char* func, const HW::Shader& /*uc*/) {
 
 std::vector<std::string> rt_print(const char* func, const HW::RenderTarget& rt) {
 	std::vector<std::string> dst;
-	dst.reserve(53);
+	dst.reserve(40);
 
 	dst.push_back(fmt::format("{}\n", func));
 
 	dst.push_back(fmt::format("\t base.addr                       = 0x{:016x}\n", rt.base.addr));
-	dst.push_back(
-	    fmt::format("\t pitch.pitch_div8_minus1         = 0x{:08x}\n", rt.pitch.pitch_div8_minus1));
-	dst.push_back(fmt::format("\t pitch.fmask_pitch_div8_minus1   = 0x{:08x}\n",
-	                          rt.pitch.fmask_pitch_div8_minus1));
-	dst.push_back(fmt::format("\t slice.slice_div64_minus1        = 0x{:08x}\n",
-	                          rt.slice.slice_div64_minus1));
 	dst.push_back(fmt::format("\t view.base_array_slice_index     = 0x{:08x}\n",
 	                          rt.view.base_array_slice_index));
 	dst.push_back(fmt::format("\t view.last_array_slice_index     = 0x{:08x}\n",
@@ -128,12 +122,6 @@ std::vector<std::string> rt_print(const char* func, const HW::RenderTarget& rt)
 	                          rt.info.cmask_fast_clear_enable ? "true" : "false"));
 	dst.push_back(fmt::format("\t info.dcc_compression_enable     = {}\n",
 	                          rt.info.dcc_compression_enable ? "true" : "false"));
-	dst.push_back(
-	    fmt::format("\t info.cmask_is_linear            = 0x{:08x}\n", rt.info.cmask_is_linear));
-	dst.push_back(
-	    fmt::format("\t info.cmask_addr_type            = 0x{:08x}\n", rt.info.cmask_addr_type));
-	dst.push_back(fmt::format("\t info.alt_tile_mode              = {}\n",
-	                          rt.info.alt_tile_mode ? "true" : "false"));
 	dst.push_back(fmt::format("\t info.format                     = 0x{:08x}\n",
 	                          static_cast<uint32_t>(rt.info.format)));
 	dst.push_back(fmt::format("\t info.channel_type               = 0x{:08x}\n",
@@ -148,10 +136,6 @@ std::vector<std::string> rt_print(const char* func, const HW::RenderTarget& rt)
 	                          rt.info.round_mode ? "true" : "false"));
 	dst.push_back(fmt::format("\t attrib.force_dest_alpha_to_one  = {}\n",
 	                          rt.attrib.force_dest_alpha_to_one ? "true" : "false"));
-	dst.push_back(fmt::format("\t attrib.tile_mode                = 0x{:08x}\n",
-	                          static_cast<uint32_t>(rt.attrib.tile_mode)));
-	dst.push_back(fmt::format("\t attrib.fmask_tile_mode          = 0x{:08x}\n",
-	                          static_cast<uint32_t>(rt.attrib.fmask_tile_mode)));
 	dst.push_back(
 	    fmt::format("\t attrib.num_samples              = 0x{:08x}\n", rt.attrib.num_samples));
 	dst.push_back(
@@ -166,42 +150,32 @@ std::vector<std::string> rt_print(const char* func, const HW::RenderTarget& rt)
 	                          static_cast<uint32_t>(rt.attrib3.tile_mode)));
 	dst.push_back(
 	    fmt::format("\t attrib3.dimension               = 0x{:08x}\n", rt.attrib3.dimension));
-	dst.push_back(fmt::format("\t attrib3.cmask_pipe_aligned      = {}\n",
-	                          rt.attrib3.cmask_pipe_aligned ? "true" : "false"));
-	dst.push_back(fmt::format("\t attrib3.dcc_pipe_aligned        = {}\n",
-	                          rt.attrib3.dcc_pipe_aligned ? "true" : "false"));
+	dst.push_back(fmt::format("\t attrib3.metadata_pipe_aligned   = {}\n",
+	                          rt.attrib3.metadata_pipe_aligned ? "true" : "false"));
+	dst.push_back(fmt::format("\t attrib3.write_vrs_rate_hint_to_cmask = {}\n",
+	                          rt.attrib3.write_vrs_rate_hint_to_cmask ? "true" : "false"));
 	dst.push_back(fmt::format("\t dcc.max_uncompressed_block_size = 0x{:08x}\n",
 	                          rt.dcc.max_uncompressed_block_size));
 	dst.push_back(fmt::format("\t dcc.max_compressed_block_size   = 0x{:08x}\n",
 	                          rt.dcc.max_compressed_block_size));
-	dst.push_back(fmt::format("\t dcc.min_compressed_block_size   = 0x{:08x}\n",
-	                          rt.dcc.min_compressed_block_size));
 	dst.push_back(
 	    fmt::format("\t dcc.color_transform             = 0x{:08x}\n", rt.dcc.color_transform));
 	dst.push_back(fmt::format("\t dcc.overwrite_combiner_disable  = {}\n",
 	                          rt.dcc.overwrite_combiner_disable ? "true" : "false"));
-	dst.push_back(fmt::format("\t dcc.independent_64b_blocks      = {}\n",
-	                          rt.dcc.independent_64b_blocks ? "true" : "false"));
-	dst.push_back(fmt::format("\t dcc.independent_128b_blocks     = {}\n",
-	                          rt.dcc.independent_128b_blocks ? "true" : "false"));
+	dst.push_back(fmt::format("\t dcc.independent_block_size      = 0x{:02x}\n",
+	                          static_cast<uint8_t>(rt.dcc.independent_block_size)));
 	dst.push_back(fmt::format("\t data_write_on_dcc_clear_to_reg  = {}\n",
 	                          rt.dcc.data_write_on_dcc_clear_to_reg ? "true" : "false"));
 	dst.push_back(fmt::format("\t dcc.dcc_clear_key_enable        = {}\n",
 	                          rt.dcc.dcc_clear_key_enable ? "true" : "false"));
 	dst.push_back(fmt::format("\t cmask.addr                      = 0x{:016x}\n", rt.cmask.addr));
-	dst.push_back(fmt::format("\t cmask_slice.slice_minus1        = 0x{:08x}\n",
-	                          rt.cmask_slice.slice_minus1));
 	dst.push_back(fmt::format("\t fmask.addr                      = 0x{:016x}\n", rt.fmask.addr));
-	dst.push_back(fmt::format("\t fmask_slice.slice_minus1        = 0x{:08x}\n",
-	                          rt.fmask_slice.slice_minus1));
 	dst.push_back(
 	    fmt::format("\t clear_word0.word0               = 0x{:08x}\n", rt.clear_word0.word0));
 	dst.push_back(
 	    fmt::format("\t clear_word1.word1               = 0x{:08x}\n", rt.clear_word1.word1));
 	dst.push_back(
 	    fmt::format("\t dcc_addr.addr                   = 0x{:016x}\n", rt.dcc_addr.addr));
-	dst.push_back(fmt::format("\t size.width                      = 0x{:08x}\n", rt.size.width));
-	dst.push_back(fmt::format("\t size.height                     = 0x{:08x}\n", rt.size.height));
 
 	return dst;
 }
@@ -231,13 +205,8 @@ static bool RenderIsColorDimension(uint32_t dimension) {
 // NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void RtCheck(const HW::RenderTarget& rt) {
 	if (rt.base.addr != 0) {
-		// bool render_to_texture = (rt.attrib.tile_mode == 0x0d);
 		//  EXIT_NOT_IMPLEMENTED(rt.base_addr == 0);
 
-		EXIT_NOT_IMPLEMENTED(rt.pitch.pitch_div8_minus1 != 0);
-		EXIT_NOT_IMPLEMENTED(rt.pitch.fmask_pitch_div8_minus1 != 0);
-		EXIT_NOT_IMPLEMENTED(rt.slice.slice_div64_minus1 != 0);
-
 		EXIT_NOT_IMPLEMENTED(rt.view.base_array_slice_index > rt.view.last_array_slice_index);
 		if (rt.view.base_array_slice_index != 0x00000000 ||
 		    rt.view.last_array_slice_index != 0x00000000) {
@@ -257,11 +226,6 @@ static void RtCheck(const HW::RenderTarget& rt) {
 				logged = true;
 			}
 		}
-		const bool native_2x_one_frag =
-		    rt.info.fmask_compression_enable && !rt.info.fmask_data_compression_disable &&
-		    rt.info.cmask_fast_clear_enable && rt.fmask.addr != 0 && rt.cmask.addr != 0 &&
-		    rt.attrib.num_samples == 1 && rt.attrib.num_fragments == 1;
-		EXIT_NOT_IMPLEMENTED(rt.info.fmask_one_frag_mode && !native_2x_one_frag);
 		if (rt.info.fmask_compression_enable) {
 			EXIT_NOT_IMPLEMENTED(rt.attrib.num_samples == 0 && rt.attrib.num_fragments == 0);
 			static bool logged = false;
@@ -286,22 +250,6 @@ static void RtCheck(const HW::RenderTarget& rt) {
 				logged = true;
 			}
 		}
-		if (rt.info.cmask_is_linear != 0x00000000 || rt.info.cmask_addr_type != 0x00000000) {
-			static bool logged = false;
-			if (!logged) {
-				LOGF("RenderTarget: temporary: ignoring PS5 CMASK metadata fields "
-				     "is_linear=%u addr_type=%u\n",
-				     rt.info.cmask_is_linear, rt.info.cmask_addr_type);
-				logged = true;
-			}
-		}
-		if (rt.info.alt_tile_mode) {
-			static bool logged = false;
-			if (!logged) {
-				LOGF("RenderTarget: temporary: ignoring alternate tile mode flag\n");
-				logged = true;
-			}
-		}
 		if (rt.info.blend_bypass) {
 			static bool logged = false;
 			if (!logged) {
@@ -328,8 +276,6 @@ static void RtCheck(const HW::RenderTarget& rt) {
 				logged = true;
 			}
 		}
-		// EXIT_NOT_IMPLEMENTED(rt.tile_mode != 0x0000000a);
-		// EXIT_NOT_IMPLEMENTED(rt.fmask_tile_mode != 0x0000000a);
 		if (rt.attrib.num_samples != 0x00000000 || rt.attrib.num_fragments != 0x00000000) {
 			static bool logged = false;
 			if (!logged) {
@@ -364,34 +310,22 @@ static void RtCheck(const HW::RenderTarget& rt) {
 		if (!RenderIsColorDimension(rt.attrib3.dimension)) {
 			EXIT("unknown PS5 render-target dimension: 0x%08" PRIx32 "\n", rt.attrib3.dimension);
 		}
-		if (!rt.attrib3.cmask_pipe_aligned) {
+		if (!rt.attrib3.metadata_pipe_aligned) {
 			static bool logged = false;
 			if (!logged) {
-				LOGF("RenderTarget: temporary: accepting unaligned PS5 CMASK pipe flag\n");
-				logged = true;
-			}
-		}
-		if (!rt.attrib3.dcc_pipe_aligned) {
-			static bool logged = false;
-			if (!logged) {
-				LOGF("RenderTarget: temporary: accepting unaligned PS5 DCC pipe flag\n");
+				LOGF("RenderTarget: temporary: accepting unaligned PS5 metadata pipe flag\n");
 				logged = true;
 			}
 		}
+		EXIT_NOT_IMPLEMENTED(rt.attrib3.write_vrs_rate_hint_to_cmask);
 
 		// EXIT_NOT_IMPLEMENTED(rt.dcc_max_uncompressed_block_size != 0x00000002);
 		// EXIT_NOT_IMPLEMENTED(rt.dcc.max_compressed_block_size != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(rt.dcc.min_compressed_block_size != 0x00000000);
 		// EXIT_NOT_IMPLEMENTED(rt.dcc.color_transform != 0x00000000);
 		EXIT_NOT_IMPLEMENTED(rt.dcc.overwrite_combiner_disable != false);
-		// EXIT_NOT_IMPLEMENTED(rt.dcc.force_independent_blocks != false);
-		// EXIT_NOT_IMPLEMENTED(rt.dcc.independent_128b_blocks != false);
 		// EXIT_NOT_IMPLEMENTED(rt.dcc.data_write_on_dcc_clear_to_reg != false);
 		EXIT_NOT_IMPLEMENTED(rt.dcc.dcc_clear_key_enable != false);
-		if (rt.cmask.addr != 0x0000000000000000 || rt.cmask_slice.slice_minus1 != 0x00000000 ||
-		    rt.fmask.addr != 0x0000000000000000 ||
-		    (rt.fmask_slice.slice_minus1 != 0x00000000 &&
-		     rt.fmask_slice.slice_minus1 != rt.slice.slice_div64_minus1) ||
+		if (rt.cmask.addr != 0x0000000000000000 || rt.fmask.addr != 0x0000000000000000 ||
 		    rt.dcc_addr.addr != 0x0000000000000000) {
 			static bool logged = false;
 			if (!logged) {
@@ -401,9 +335,6 @@ static void RtCheck(const HW::RenderTarget& rt) {
 				logged = true;
 			}
 		}
-
-		EXIT_NOT_IMPLEMENTED(rt.size.width != 0);
-		EXIT_NOT_IMPLEMENTED(rt.size.height != 0);
 	}
 }
 
@@ -423,37 +354,20 @@ static void ZPrint(const char* func, const HW::DepthRenderTarget& z) {
 	     "\t stencil_info.htile_stencil_disabled   = %s\n"
 	     "\t stencil_info.expclear_enabled         = %s\n"
 	     "\t stencil_info.partially_resident       = %s\n"
-	     "\t depth_info.addr5_swizzle_mask         = 0x%08" PRIx32 "\n"
-	     "\t depth_info.array_mode                 = 0x%08" PRIx32 "\n"
-	     "\t depth_info.pipe_config                = 0x%08" PRIx32 "\n"
-	     "\t depth_info.bank_width                 = 0x%08" PRIx32 "\n"
-	     "\t depth_info.bank_height                = 0x%08" PRIx32 "\n"
-	     "\t depth_info.macro_tile_aspect          = 0x%08" PRIx32 "\n"
-	     "\t depth_info.num_banks                  = 0x%08" PRIx32 "\n"
 	     "\t depth_view.slice_start                = 0x%08" PRIx32 "\n"
 	     "\t depth_view.slice_max                  = 0x%08" PRIx32 "\n"
 	     "\t depth_view.current_mip_level          = 0x%02" PRIx8 "\n"
 	     "\t depth_view.depth_write_disable        = %s\n"
 	     "\t depth_view.stencil_write_disable      = %s\n"
-	     "\t htile_surface.linear                  = 0x%08" PRIx32 "\n"
-	     "\t htile_surface.full_cache              = 0x%08" PRIx32 "\n"
-	     "\t htile_surface.htile_uses_preload_win  = 0x%08" PRIx32 "\n"
-	     "\t htile_surface.preload                 = 0x%08" PRIx32 "\n"
-	     "\t htile_surface.prefetch_width          = 0x%08" PRIx32 "\n"
-	     "\t htile_surface.prefetch_height         = 0x%08" PRIx32 "\n"
-	     "\t htile_surface.dst_outside_zero_to_one = 0x%08" PRIx32 "\n"
 	     "\t z_read_base_addr                      = 0x%016" PRIx64 "\n"
 	     "\t stencil_read_base_addr                = 0x%016" PRIx64 "\n"
 	     "\t z_write_base_addr                     = 0x%016" PRIx64 "\n"
 	     "\t stencil_write_base_addr               = 0x%016" PRIx64 "\n"
-	     "\t pitch_div8_minus1                     = 0x%08" PRIx32 "\n"
-	     "\t height_div8_minus1                    = 0x%08" PRIx32 "\n"
-	     "\t slice_div64_minus1                    = 0x%08" PRIx32 "\n"
 	     "\t htile_data_base_addr                  = 0x%016" PRIx64 "\n"
-	     "\t width                                 = 0x%08" PRIx32 "\n"
-	     "\t height                                = 0x%08" PRIx32 "\n"
+	     "\t shading_rate_encoding                 = 0x%02" PRIx8 "\n"
 	     "\t size.x_max                            = 0x%04" PRIx16 "\n"
-	     "\t size.y_max                            = 0x%04" PRIx16 "\n",
+	     "\t size.y_max                            = 0x%04" PRIx16 "\n"
+	     "\t size.valid                            = %s\n",
 	     static_cast<uint32_t>(z.z_info.format), z.z_info.num_samples,
 	     static_cast<uint32_t>(z.z_info.texture_compatibility),
 	     z.z_info.htile_acceleration ? "true" : "false",
@@ -464,30 +378,27 @@ static void ZPrint(const char* func, const HW::DepthRenderTarget& z) {
 	     static_cast<uint32_t>(z.stencil_info.texture_compatibility),
 	     z.stencil_info.htile_stencil_disabled ? "true" : "false",
 	     z.stencil_info.expclear_enabled ? "true" : "false",
-	     z.stencil_info.partially_resident ? "true" : "false", z.depth_info.addr5_swizzle_mask,
-	     z.depth_info.array_mode, z.depth_info.pipe_config, z.depth_info.bank_width,
-	     z.depth_info.bank_height, z.depth_info.macro_tile_aspect, z.depth_info.num_banks,
-	     z.depth_view.slice_start, z.depth_view.slice_max, z.depth_view.current_mip_level,
+	     z.stencil_info.partially_resident ? "true" : "false", z.depth_view.slice_start,
+	     z.depth_view.slice_max, z.depth_view.current_mip_level,
 	     z.depth_view.depth_write_disable ? "true" : "false",
-	     z.depth_view.stencil_write_disable ? "true" : "false", z.htile_surface.linear,
-	     z.htile_surface.full_cache, z.htile_surface.htile_uses_preload_win,
-	     z.htile_surface.preload, z.htile_surface.prefetch_width, z.htile_surface.prefetch_height,
-	     z.htile_surface.dst_outside_zero_to_one, z.z_read_base_addr, z.stencil_read_base_addr,
-	     z.z_write_base_addr, z.stencil_write_base_addr, z.pitch_div8_minus1, z.height_div8_minus1,
-	     z.slice_div64_minus1, z.htile_data_base_addr, z.width, z.height, z.size.x_max,
-	     z.size.y_max);
+	     z.depth_view.stencil_write_disable ? "true" : "false", z.z_read_base_addr,
+	     z.stencil_read_base_addr, z.z_write_base_addr, z.stencil_write_base_addr,
+	     z.htile_data_base_addr, z.shading_rate_encoding, z.size.x_max, z.size.y_max,
+	     z.size.valid ? "true" : "false");
 }
 
 // NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void ZCheck(const HW::DepthRenderTarget& z, const HW::DepthControl& dc,
                    const HW::RenderControl& rc) {
-	const bool depth_active =
-	    dc.z_enable || dc.z_write_enable || dc.depth_bounds_enable || rc.depth_clear_enable;
-	const bool stencil_active = dc.stencil_enable || rc.stencil_clear_enable;
+	const bool depth_active = dc.z_enable || dc.z_write_enable || dc.depth_bounds_enable ||
+	                          rc.depth_clear_enable || rc.copy_depth_to_color;
+	const bool stencil_active =
+	    dc.stencil_enable || rc.stencil_clear_enable || rc.copy_stencil_to_color;
 	if (!depth_active && !stencil_active) {
 		return;
 	}
 
+	EXIT_NOT_IMPLEMENTED(rc.copy_depth_to_color || rc.copy_stencil_to_color);
 	EXIT_NOT_IMPLEMENTED(!z.z_info.HasValidTextureCompatibility());
 	EXIT_NOT_IMPLEMENTED(!z.stencil_info.HasValidTextureCompatibility());
 	if (z.z_info.format == Prospero::DepthFormat::kInvalid) {
@@ -524,21 +435,7 @@ static void ZCheck(const HW::DepthRenderTarget& z, const HW::DepthControl& dc,
 
 	if (z.z_info.format != Prospero::DepthFormat::kInvalid ||
 	    z.stencil_info.format != Prospero::StencilFormat::kInvalid) {
-
-		EXIT_NOT_IMPLEMENTED(z.depth_info.addr5_swizzle_mask != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.depth_info.array_mode != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.depth_info.pipe_config != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.depth_info.bank_width != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.depth_info.bank_height != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.depth_info.macro_tile_aspect != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.depth_info.num_banks != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.htile_surface.linear != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.htile_surface.full_cache != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.htile_surface.htile_uses_preload_win != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.htile_surface.preload != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.htile_surface.prefetch_width != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.htile_surface.prefetch_height != 0x00000000);
-		EXIT_NOT_IMPLEMENTED(z.htile_surface.dst_outside_zero_to_one != 0x00000000);
+		EXIT_NOT_IMPLEMENTED(z.shading_rate_encoding != 0);
 
 		if (z.depth_view.current_mip_level != 0x00000000) {
 			static std::atomic<uint32_t> log_count {0};
@@ -560,23 +457,8 @@ static void ZCheck(const HW::DepthRenderTarget& z, const HW::DepthControl& dc,
 		EXIT_NOT_IMPLEMENTED(!z.depth_view.stencil_write_disable &&
 		                     z.stencil_read_base_addr != z.stencil_write_base_addr);
 		EXIT_NOT_IMPLEMENTED(!z.depth_view.depth_write_disable && z.z_write_base_addr == 0);
-		// EXIT_NOT_IMPLEMENTED(z.pitch_div8_minus1 != 0x000000ff);
-		// EXIT_NOT_IMPLEMENTED(z.height_div8_minus1 != 0x0000008f);
-		// EXIT_NOT_IMPLEMENTED(z.slice_div64_minus1 != 0x00008fff);
 		// EXIT_NOT_IMPLEMENTED(z.htile_data_base_addr == 0);
-		// EXIT_NOT_IMPLEMENTED(z.width != 0x00000780);
-		// EXIT_NOT_IMPLEMENTED(z.height != 0x00000438);
-
-		EXIT_NOT_IMPLEMENTED(z.width != 0);
-		EXIT_NOT_IMPLEMENTED(z.height != 0);
-		if (z.size.x_max == 0 || z.size.y_max == 0) {
-			static std::atomic<uint32_t> log_count {0};
-			if (log_count.fetch_add(1) < 16) {
-				LOGF("DepthTarget: temporary: inferring PS5 default depth extent x_max=0x%04" PRIx16
-				     ", y_max=0x%04" PRIx16 "\n",
-				     z.size.x_max, z.size.y_max);
-			}
-		}
+		EXIT_NOT_IMPLEMENTED(!z.size.valid);
 	}
 }
 
@@ -608,7 +490,7 @@ static void ClipCheck(const HW::ClipControl& c) {
 	// dx_linear_attr_clip_enable preserves linear (noperspective) attributes at clip-generated
 	// vertices, which Vulkan provides as part of clipping and interpolation.
 	EXIT_NOT_IMPLEMENTED(c.user_clip_planes != 0 || c.user_clip_plane_mode != 0 ||
-	                     c.vertex_kill_any || c.user_clip_plane_negate_y || c.clip_disable ||
+	                     c.vertex_kill_any || c.user_clip_plane_negate_y ||
 	                     c.user_clip_plane_cull_only || c.cull_on_clipping_error_disable ||
 	                     c.force_viewport_index_from_vs_enable);
 }
@@ -759,8 +641,6 @@ static void DPrint(const char* func, const HW::DepthControl& c, const HW::Stenci
 	     "\t backface_enable      = %s\n"
 	     "\t stencilfunc          = %" PRIu8 "\n"
 	     "\t stencilfunc_bf       = %" PRIu8 "\n"
-	     "\t color_writes_on_depth_fail_enable  = %s\n"
-	     "\t color_writes_on_depth_pass_disable = %s\n"
 	     "\t stencil_fail         = %" PRIu8 "\n"
 	     "\t stencil_zpass        = %" PRIu8 "\n"
 	     "\t stencil_zfail        = %" PRIu8 "\n"
@@ -777,12 +657,11 @@ static void DPrint(const char* func, const HW::DepthControl& c, const HW::Stenci
 	     "\t stencil_opval_bf     = %" PRIu8 "\n",
 	     c.stencil_enable ? "true" : "false", c.z_enable ? "true" : "false",
 	     c.z_write_enable ? "true" : "false", c.depth_bounds_enable ? "true" : "false", c.zfunc,
-	     c.backface_enable ? "true" : "false", c.stencilfunc, c.stencilfunc_bf,
-	     c.color_writes_on_depth_fail_enable ? "true" : "false",
-	     c.color_writes_on_depth_pass_disable ? "true" : "false", s.stencil_fail, s.stencil_zpass,
-	     s.stencil_zfail, s.stencil_fail_bf, s.stencil_zpass_bf, s.stencil_zfail_bf,
-	     sm.stencil_testval, sm.stencil_mask, sm.stencil_writemask, sm.stencil_opval,
-	     sm.stencil_testval_bf, sm.stencil_mask_bf, sm.stencil_writemask_bf, sm.stencil_opval_bf);
+	     c.backface_enable ? "true" : "false", c.stencilfunc, c.stencilfunc_bf, s.stencil_fail,
+	     s.stencil_zpass, s.stencil_zfail, s.stencil_fail_bf, s.stencil_zpass_bf,
+	     s.stencil_zfail_bf, sm.stencil_testval, sm.stencil_mask, sm.stencil_writemask,
+	     sm.stencil_opval, sm.stencil_testval_bf, sm.stencil_mask_bf, sm.stencil_writemask_bf,
+	     sm.stencil_opval_bf);
 }
 
 static void DCheck(const HW::DepthControl& c, const HW::StencilControl& s,
@@ -794,8 +673,6 @@ static void DCheck(const HW::DepthControl& c, const HW::StencilControl& s,
 	// Back-face stencil state is handled separately when enabled.
 	// EXIT_NOT_IMPLEMENTED(c.stencilfunc != 0);
 	// EXIT_NOT_IMPLEMENTED(c.stencilfunc_bf != 0);
-	EXIT_NOT_IMPLEMENTED(c.color_writes_on_depth_fail_enable != false);
-	EXIT_NOT_IMPLEMENTED(c.color_writes_on_depth_pass_disable != false);
 	// EXIT_NOT_IMPLEMENTED(s.stencil_fail != 0);
 	// EXIT_NOT_IMPLEMENTED(s.stencil_zpass != 0);
 	// EXIT_NOT_IMPLEMENTED(s.stencil_zfail != 0);
@@ -827,7 +704,12 @@ static void EqaaPrint(const char* func, const HW::EqaaControl& c) {
 	     c.static_anchor_associations ? "true" : "false");
 }
 
-static void EqaaCheck(const HW::EqaaControl& c) {
+static void EqaaCheck(const HW::EqaaControl& c, const HW::AaConfig& cf) {
+	// The EQAA controls only take effect while multisampling is on, so a single-sample
+	// target leaves them inert and has nothing to warn about.
+	if (cf.msaa_num_samples == 0) {
+		return;
+	}
 	if (c.max_anchor_samples != 0 || c.ps_iter_samples != 0 || c.mask_export_num_samples != 0 ||
 	    c.alpha_to_mask_num_samples != 0 || c.high_quality_intersections ||
 	    c.incoherent_eqaa_reads || c.interpolate_comp_z || c.static_anchor_associations) {
@@ -1192,12 +1074,12 @@ void hw_check(const CommandBuffer& buffer) {
 	log_phase("blend");
 	BcCheck(bc, bclr, cc);
 	log_phase("eqaa");
-	EqaaCheck(eqaa);
+	EqaaCheck(eqaa, ac);
 	log_phase("aa");
 	AaCheck(aa, ac);
 	log_phase("done");
 
-	if (RenderTargetMaskHasBoundMrt(buffer)) {
+	if (graphics_debug_dump_enabled() && RenderTargetMaskHasBoundMrt(buffer)) {
 		LOGF("MRT render target mask: 0x%08" PRIx32 "\n", hw.GetRenderTargetMask());
 		for (uint32_t i = 0; i < 8; i++) {
 			const auto& mrt = hw.GetRenderTarget(i);

src/graphics/host_gpu/renderer/depthRenderTarget.cpp

@@ -108,9 +108,10 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 	const auto& sc          = hw.GetStencilControl();
 	const auto& sm          = hw.GetStencilMask();
 	const bool  has_stencil = z.stencil_info.format != Prospero::StencilFormat::kInvalid;
-	const bool  depth_active =
-	    dc.z_enable || dc.z_write_enable || dc.depth_bounds_enable || rc.depth_clear_enable;
-	const bool stencil_active = has_stencil && (dc.stencil_enable || rc.stencil_clear_enable);
+	const bool depth_active = dc.z_enable || dc.z_write_enable || dc.depth_bounds_enable ||
+	                          rc.depth_clear_enable || rc.copy_depth_to_color;
+	const bool stencil_active =
+	    has_stencil && (dc.stencil_enable || rc.stencil_clear_enable || rc.copy_stencil_to_color);
 	if (!depth_active && !stencil_active) {
 		return;
 	}
@@ -127,22 +128,14 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 	    !z.stencil_info.expclear_enabled && !z.stencil_info.partially_resident &&
 	    z.depth_view.slice_start == 0 && z.depth_view.slice_max == 0 &&
 	    z.depth_view.current_mip_level == 0 && !z.depth_view.depth_write_disable &&
-	    !z.depth_view.stencil_write_disable && z.depth_info.addr5_swizzle_mask == 0 &&
-	    z.depth_info.array_mode == 0 && z.depth_info.pipe_config == 0 &&
-	    z.depth_info.bank_width == 0 && z.depth_info.bank_height == 0 &&
-	    z.depth_info.macro_tile_aspect == 0 && z.depth_info.num_banks == 0 &&
-	    z.htile_surface.linear == 0 && z.htile_surface.full_cache == 0 &&
-	    z.htile_surface.htile_uses_preload_win == 0 && z.htile_surface.preload == 0 &&
-	    z.htile_surface.prefetch_width == 0 && z.htile_surface.prefetch_height == 0 &&
-	    z.htile_surface.dst_outside_zero_to_one == 0 && z.z_read_base_addr == 0 &&
-	    z.z_write_base_addr == 0 && z.stencil_read_base_addr == 0 &&
-	    z.stencil_write_base_addr == 0 && z.htile_data_base_addr == 0 &&
+	    !z.depth_view.stencil_write_disable && z.z_read_base_addr == 0 && z.z_write_base_addr == 0 &&
+	    z.stencil_read_base_addr == 0 && z.stencil_write_base_addr == 0 &&
+	    z.htile_data_base_addr == 0 &&
 	    // DB_DEPTH_SIZE_XY is independent state and may remain programmed after the attachment
 	    // formats and addresses are unbound. A zero encoding is the valid 1x1 value, so its
 	    // presence alone must not manufacture a depth attachment.
-	    !z.z_info.htile_acceleration && !z.width_height_valid && !z.pitch_height_valid &&
-	    z.size.x_max == 0 && z.size.y_max == 0 && z.pitch_div8_minus1 == 0 &&
-	    z.height_div8_minus1 == 0 && z.slice_div64_minus1 == 0 && z.width == 0 && z.height == 0;
+	    !z.z_info.htile_acceleration && z.shading_rate_encoding == 0 && z.size.x_max == 0 &&
+	    z.size.y_max == 0;
 	if (attachment_unbound) {
 		static std::atomic_bool logged = false;
 		if (!logged.exchange(true, std::memory_order_relaxed)) {
@@ -165,17 +158,12 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 			DepthFatal("invalid depth view: base=%u last=%u", z.depth_view.slice_start,
 			           z.depth_view.slice_max);
 	}
-	if (rc.copy_centroid || rc.copy_sample != 0 || z.z_info.expclear_enabled ||
+	if (rc.copy_depth_to_color || rc.copy_stencil_to_color || rc.copy_centroid ||
+	    rc.copy_sample != 0 || z.z_info.expclear_enabled ||
 	    z.stencil_info.expclear_enabled || z.z_info.partially_resident ||
 	    z.stencil_info.partially_resident || z.z_info.max_mip_level != 0 ||
-	    z.depth_view.current_mip_level != 0 || z.depth_info.addr5_swizzle_mask != 0 ||
-	    z.depth_info.array_mode != 0 || z.depth_info.pipe_config != 0 ||
-	    z.depth_info.bank_width != 0 || z.depth_info.bank_height != 0 ||
-	    z.depth_info.macro_tile_aspect != 0 || z.depth_info.num_banks != 0 ||
-	    z.htile_surface.linear != 0 || z.htile_surface.full_cache != 0 ||
-	    z.htile_surface.htile_uses_preload_win != 0 || z.htile_surface.preload != 0 ||
-	    z.htile_surface.prefetch_width != 0 || z.htile_surface.prefetch_height != 0 ||
-	    z.htile_surface.dst_outside_zero_to_one != 0 || z.z_read_base_addr == 0 ||
+	    z.depth_view.current_mip_level != 0 || z.shading_rate_encoding != 0 ||
+	    z.z_read_base_addr == 0 ||
 	    (!z.depth_view.depth_write_disable && z.z_write_base_addr != z.z_read_base_addr) ||
 	    (z.z_read_base_addr & 0xffffu) != 0 ||
 	    dc.zfunc > static_cast<uint8_t>(vk::CompareOp::eAlways)) {
@@ -189,10 +177,7 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 		    (z.stencil_read_base_addr & 0xffffu) != 0) {
 			DepthFatal("unsupported stencil attachment state");
 		}
-	} else if (z.stencil_read_base_addr != 0 || z.stencil_write_base_addr != 0 ||
-	           !htile_stencil_compat ||
-	           z.stencil_info.texture_compatibility !=
-	               Prospero::TextureCompatibleStencil::kDisable) {
+	} else if (z.stencil_read_base_addr != 0 || z.stencil_write_base_addr != 0) {
 		DepthFatal("stencil state without an active stencil attachment");
 	}
 	if (has_htile) {
@@ -203,18 +188,13 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 			DepthFatal("HTile clear tracking supports at most 32 slices");
 		}
 	}
-	const bool size_xy_valid = z.size.valid;
-	const bool wh_valid      = z.width_height_valid && z.width != 0 && z.height != 0;
-	if (!size_xy_valid && !wh_valid) {
+	if (!z.size.valid) {
 		DepthFatal("missing depth extent");
 	}
-	const uint32_t width  = size_xy_valid ? static_cast<uint32_t>(z.size.x_max) + 1u : z.width;
-	const uint32_t height = size_xy_valid ? static_cast<uint32_t>(z.size.y_max) + 1u : z.height;
-	if (width > 16384 || height > 16384 ||
-	    (size_xy_valid && wh_valid && (width != z.width || height != z.height)) ||
-	    (!z.pitch_height_valid &&
-	     (z.pitch_div8_minus1 != 0 || z.height_div8_minus1 != 0 || z.slice_div64_minus1 != 0))) {
-		DepthFatal("inconsistent depth extent or encoded layout");
+	const uint32_t width  = static_cast<uint32_t>(z.size.x_max) + 1u;
+	const uint32_t height = static_cast<uint32_t>(z.size.y_max) + 1u;
+	if (width > 16384 || height > 16384) {
+		DepthFatal("invalid depth extent");
 	}
 	const auto* policy = FindDepthFormatPolicy(z.z_info.format);
 	if (policy == nullptr) {
@@ -229,11 +209,6 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 	const auto     guest_format = policy->guest_format;
 	const uint32_t bytes        = policy->bytes_per_element;
 	const auto     pitch        = TileGetDepthPitch(width, bytes, z.z_info.num_samples);
-	if (z.pitch_height_valid && ((static_cast<uint64_t>(z.pitch_div8_minus1) + 1u) * 8u != pitch ||
-	                             (static_cast<uint64_t>(z.height_div8_minus1) + 1u) * 8u !=
-	                                 ((static_cast<uint64_t>(height) + 7u) & ~7ull))) {
-		DepthFatal("encoded depth pitch or height mismatch");
-	}
 	TileSizeAlign depth_size {};
 	TileSizeAlign stencil_size {};
 	TileSizeAlign htile_size {};
@@ -244,14 +219,6 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 	    (has_htile != (htile_size.align == 32768 && htile_size.size != 0))) {
 		DepthFatal("unsupported depth/stencil/HTile footprint");
 	}
-	if (z.pitch_height_valid &&
-	    (static_cast<uint64_t>(z.slice_div64_minus1) + 1u) * 64u != depth_size.size) {
-		DepthFatal("depth footprint mismatch: extent=%ux%u pitch=%u expected=0x%016" PRIx64
-		           " align=0x%016" PRIx64 " encoded_valid=%u encoded=0x%016" PRIx64,
-		           width, height, pitch, depth_size.size, depth_size.align,
-		           z.pitch_height_valid ? 1u : 0u,
-		           (static_cast<uint64_t>(z.slice_div64_minus1) + 1u) * 64u);
-	}
 	if (depth_size.size > UINT64_MAX / view.image_layers ||
 	    stencil_size.size > UINT64_MAX / view.image_layers ||
 	    htile_size.size > UINT64_MAX / view.image_layers) {
@@ -260,9 +227,9 @@ void RenderExecutor::ResolveRenderDepthTarget(uint64_t submit_id, CommandBuffer&
 	const auto depth_backing_size   = depth_size.size * view.image_layers;
 	const auto stencil_backing_size = stencil_size.size * view.image_layers;
 	const auto htile_backing_size   = htile_size.size * view.image_layers;
-	if (depth_backing_size > TRACKER_ADDRESS_SIZE - z.z_read_base_addr ||
-	    (has_stencil && stencil_backing_size > TRACKER_ADDRESS_SIZE - z.stencil_read_base_addr) ||
-	    (has_htile && htile_backing_size > TRACKER_ADDRESS_SIZE - z.htile_data_base_addr)) {
+	if (!GuestRange {z.z_read_base_addr, depth_backing_size}.Valid() ||
+	    (has_stencil && !GuestRange {z.stencil_read_base_addr, stencil_backing_size}.Valid()) ||
+	    (has_htile && !GuestRange {z.htile_data_base_addr, htile_backing_size}.Valid())) {
 		DepthFatal("layered depth backing range is invalid");
 	}
 	r.htile                   = has_htile;

src/graphics/host_gpu/renderer/depthRenderTarget.h

@@ -11,9 +11,11 @@
 
 namespace Libs::Graphics {
 
+// Without a stencil plane, Hi-Stencil fields are inactive. An active plane is compatible when
+// Hi-Stencil is disabled or HTile backing is present.
 inline constexpr bool depth_htile_stencil_acceleration_compatible(bool has_stencil, bool has_htile,
                                                                   bool htile_stencil_disabled) {
-	return htile_stencil_disabled || (has_stencil && has_htile);
+	return !has_stencil || htile_stencil_disabled || has_htile;
 }
 
 struct RenderDepthInfo {

src/graphics/host_gpu/renderer/image/image.cpp

@@ -73,10 +73,8 @@ namespace {
 	return usage;
 }
 
-void ValidateRange(GuestRange range, const char* name) {
-	if ((range.address == 0) != (range.size == 0) ||
-	    (range.address != 0 && (range.address >= TRACKER_ADDRESS_SIZE ||
-	                            range.size > TRACKER_ADDRESS_SIZE - range.address))) {
+void ValidateOptionalRange(GuestRange range, const char* name) {
+	if (!range.ValidOrEmpty()) {
 		EXIT("invalid %s image range: address=0x%016llx size=0x%016llx\n", name,
 		     static_cast<unsigned long long>(range.address),
 		     static_cast<unsigned long long>(range.size));
@@ -567,13 +565,13 @@ void Image::CopyMip(Image& source, uint32_t mip, uint32_t layer) {
 namespace ImageOps {
 
 void Validate(const ImageInfo& info) {
-	ValidateRange(info.data, "data");
-	ValidateRange(info.stencil, "stencil");
+	ValidateOptionalRange(info.data, "data");
+	ValidateOptionalRange(info.stencil, "stencil");
 
 	if (info.pixel_format == vk::Format::eUndefined) {
 		const bool metadata_empty =
-		    info.metadata.range.address == 0 && info.metadata.range.size == 0 &&
-		    info.metadata.kind == ImageMetadataKind::None && info.metadata.control == 0 &&
+		    info.metadata.range.Empty() && info.metadata.kind == ImageMetadataKind::None &&
+		    info.metadata.control == 0 &&
 		    info.metadata.compression == VideoOutCompression::Uncompressed &&
 		    !info.metadata.stencil_compressed;
 		if (info.data.Empty() || info.HasStencil() || !metadata_empty || info.extent.width == 0 ||
@@ -620,16 +618,14 @@ void Validate(const ImageInfo& info) {
 	}
 	switch (info.metadata.kind) {
 		case ImageMetadataKind::None:
-			if (info.metadata.range.address != 0 || info.metadata.range.size != 0 ||
-			    info.metadata.control != 0 ||
+			if (!info.metadata.range.Empty() || info.metadata.control != 0 ||
 			    info.metadata.compression != VideoOutCompression::Uncompressed ||
 			    info.metadata.stencil_compressed) {
 				EXIT("metadata-free image has metadata state\n");
 			}
 			break;
 		case ImageMetadataKind::Htile:
-			ValidateRange(info.metadata.range, "HTILE");
-			if (info.metadata.range.Empty() ||
+			if (!info.metadata.range.Valid() ||
 			    info.metadata.compression != VideoOutCompression::Uncompressed) {
 				EXIT("invalid HTILE metadata\n");
 			}

src/graphics/host_gpu/renderer/image/imageInfo.h

@@ -15,18 +15,6 @@
 
 namespace Libs::Graphics {
 
-struct GuestRange {
-	uint64_t address = 0;
-	uint64_t size    = 0;
-
-	[[nodiscard]] constexpr bool Empty() const noexcept { return address == 0 || size == 0; }
-	[[nodiscard]] constexpr bool Valid() const noexcept {
-		return !Empty() && address < TRACKER_ADDRESS_SIZE && size <= TRACKER_ADDRESS_SIZE - address;
-	}
-	[[nodiscard]] constexpr uint64_t End() const noexcept { return address + size; }
-	auto                             operator<=>(const GuestRange&) const = default;
-};
-
 enum class VideoOutCompression : uint8_t { Uncompressed, Dcc256_256_0, Dcc256_64_64, Unsupported };
 
 enum class ImageMetadataKind : uint8_t { None, Htile, Dcc };
@@ -443,6 +431,10 @@ IsSupportedDisplayRenderTargetTileMode(Prospero::TileMode tile_mode) noexcept {
 		return encoded <= 0.04045f ? encoded / 12.92f : std::pow((encoded + 0.055f) / 1.055f, 2.4f);
 	};
 	switch (format) {
+		// A single-plane float target carries its clear as raw float bits, the same encoding the
+		// depth decoder below uses. Without this the clear is discarded and the target keeps stale
+		// contents.
+		case vk::Format::eR32Sfloat: next.float32[0] = std::bit_cast<float>(packed); break;
 		case vk::Format::eR32Uint: next.uint32[0] = packed; break;
 		case vk::Format::eR32Sint: next.int32[0] = static_cast<int32_t>(packed); break;
 		case vk::Format::eR8G8B8A8Srgb:

src/graphics/host_gpu/renderer/pipeline/descriptors.cpp

@@ -895,9 +895,9 @@ PreparedBindings RenderExecutor::PrepareBindings(const ShaderStageRuntime& runti
 	EXIT_IF(!runtime);
 	const auto& program  = *runtime.program;
 	const auto& snapshot = *runtime.resources;
-	std::string error;
-	if (!ShaderRecompiler::IR::ValidateResourceSpecialization(program, snapshot, &error)) {
-		EXIT("invalid native shader runtime snapshot: %s\n", error.c_str());
+	if (!ShaderRecompiler::IR::ValidateResourceSpecialization(program, snapshot)) {
+		EXIT("invalid native shader runtime snapshot: hash=0x%016" PRIx64 " stage=%u\n",
+		     program.shader_hash, static_cast<unsigned>(program.stage));
 	}
 
 	PreparedBindings prepared;

src/graphics/host_gpu/renderer/pipeline/pipelineCache.cpp

@@ -1,6 +1,7 @@
 #include "graphics/host_gpu/renderer/pipeline/pipelineCache.h"
 
 #include "common/assert.h"
+#include "common/file.h"
 #include "common/logging/log.h"
 #include "common/profiler.h"
 #include "graphics/guest_gpu/hardwareContext.h"
@@ -11,18 +12,64 @@
 #include "graphics/host_gpu/renderer/render.h"
 #include "graphics/host_gpu/renderer/renderContext.h"
 #include "graphics/shader/shaderCompiler.h"
+#include "kytyGitVersion.h"
+#include "loader/systemContent.h"
 
+#include <algorithm>
 #include <atomic>
+#include <cctype>
 #include <cstdio>
 #include <cstring>
+#include <limits>
 #include <span>
+#include <string_view>
 #include <utility>
 #include <vector>
 
+#include <fmt/format.h>
+#include <xxhash.h>
+
 namespace Libs::Graphics {
 
 namespace {
 
+std::string DriverCacheSignature(const vk::PhysicalDeviceProperties& properties) {
+	constexpr char hex[] = "0123456789abcdef";
+	std::string    uuid(VK_UUID_SIZE * 2, '0');
+	for (size_t i = 0; i < VK_UUID_SIZE; i++) {
+		uuid[i * 2]     = hex[properties.pipelineCacheUUID[i] >> 4u];
+		uuid[i * 2 + 1] = hex[properties.pipelineCacheUUID[i] & 0xfu];
+	}
+	return fmt::format("KytyPC1:{}:{:08x}:{:08x}:{:08x}:{}\n", KYTY_GIT_REVISION,
+	                   properties.vendorID, properties.deviceID, properties.driverVersion, uuid);
+}
+
+std::string PipelineCacheTitleId() {
+	std::string title_id;
+	if ((!Loader::SystemContentParamSfoGetString("TITLE_ID", &title_id) || title_id.empty()) &&
+	    (!Loader::SystemContentParamSfoGetString("CONTENT_ID", &title_id) || title_id.empty())) {
+		return {};
+	}
+	if (!std::ranges::all_of(title_id, [](unsigned char c) {
+		    return std::isalnum(c) != 0 || c == '-' || c == '_';
+	    })) {
+		return {};
+	}
+	return title_id;
+}
+
+template <typename... Args>
+void PipelineCacheLog(fmt::format_string<Args...> format, Args&&... args) {
+	auto message = fmt::format(format, std::forward<Args>(args)...);
+	message += '\n';
+	if (Log::GetDirection() != Log::Direction::Console) {
+		std::fwrite(message.data(), 1, message.size(), stdout);
+		std::fflush(stdout);
+	}
+	Log::Write(message);
+	Log::Flush();
+}
+
 void NormalizeStaticParamsForDynamicState(PipelineStaticParameters& static_params) {
 	static_params.viewport_scale[0]  = 0.5f;
 	static_params.viewport_scale[1]  = 0.5f;
@@ -69,7 +116,7 @@ struct PipelineCache::ProgramCache {
 	}
 
 	static constexpr std::size_t MaxStaticKeyWords =
-	    5 + ShaderVertexInputInfo::RES_MAX * 17;
+	    12 + ShaderVertexInputInfo::RES_MAX * 17;
 
 	template <typename InputInfo>
 	ShaderProgram Get(const ShaderParams& params, InputInfo& input_info) {
@@ -119,9 +166,11 @@ struct PipelineCache::ProgramCache {
 PipelineCache::PipelineCache(GraphicContext& graphics)
     : m_graphics(graphics), m_program_cache(std::make_unique<ProgramCache>(graphics.device)) {
 	EXIT_NOT_IMPLEMENTED(!Common::Thread::IsMainThread());
+	InitializeDriverCache();
 }
 
 PipelineCache::~PipelineCache() {
+	Save();
 	auto destroy = [this](const auto& pipelines) {
 		for (const auto& [key, pipeline]: pipelines) {
 			(void)key;
@@ -138,12 +187,176 @@ PipelineCache::~PipelineCache() {
 			m_graphics.device.destroyShaderModule(permutation.handle.module, nullptr);
 		}
 	}
+	if (m_driver_cache != nullptr) {
+		m_graphics.device.destroyPipelineCache(m_driver_cache, nullptr);
+	}
+}
+
+void PipelineCache::InitializeDriverCache() {
+	const auto title_id = PipelineCacheTitleId();
+	if (title_id.empty()) {
+		return;
+	}
+	if (KYTY_BUILD != KYTY_BUILD_RELEASE) {
+		PipelineCacheLog("Vulkan pipeline cache: disabled (non-Release build)");
+		return;
+	}
+	const std::string_view git_hash     = KYTY_GIT_HASH;
+	const std::string_view git_revision = KYTY_GIT_REVISION;
+	if (git_hash == "unknown" || git_revision == "unknown") {
+		PipelineCacheLog("Vulkan pipeline cache: disabled (unknown git revision)");
+		return;
+	}
+	if (git_hash.ends_with("-dirty")) {
+		PipelineCacheLog("Vulkan pipeline cache: disabled (dirty build)");
+		return;
+	}
+
+	m_driver_cache_path = std::filesystem::path("_PipelineCache") / (title_id + ".bin");
+	const auto path         = Common::PathToString(m_driver_cache_path);
+	const bool cache_exists = Common::File::IsFileExisting(m_driver_cache_path);
+	if (cache_exists) {
+		PipelineCacheLog("Vulkan pipeline cache: loading {}", path);
+	} else {
+		PipelineCacheLog("Vulkan pipeline cache: initializing {}", path);
+	}
+	std::vector<uint8_t> initial_data;
+	if (cache_exists) {
+		Common::File file(m_driver_cache_path, Common::File::Mode::Read);
+		const auto   file_size = file.IsInvalid() ? 0 : file.Size();
+		const auto signature = DriverCacheSignature(m_graphics.GetPhysicalDeviceProperties());
+		if (file_size >= signature.size() + sizeof(uint64_t) &&
+		    file_size <= std::numeric_limits<uint32_t>::max()) {
+			std::string cached_signature(signature.size(), '\0');
+			uint64_t    payload_hash = 0;
+			initial_data.resize(file_size - signature.size() - sizeof(payload_hash));
+			uint32_t signature_read = 0;
+			uint32_t hash_read      = 0;
+			uint32_t payload_read   = 0;
+			file.Read(cached_signature.data(), static_cast<uint32_t>(cached_signature.size()),
+			          &signature_read);
+			file.Read(&payload_hash, sizeof(payload_hash), &hash_read);
+			file.Read(initial_data.data(), static_cast<uint32_t>(initial_data.size()), &payload_read);
+			file.Close();
+			if (signature_read != cached_signature.size() || hash_read != sizeof(payload_hash) ||
+			    payload_read != initial_data.size() || cached_signature != signature ||
+			    XXH3_64bits(initial_data.data(), initial_data.size()) != payload_hash) {
+				initial_data.clear();
+				PipelineCacheLog(
+				    "Vulkan pipeline cache: invalidating {} (driver, emulator, or data mismatch)",
+				    path);
+			}
+		} else {
+			file.Close();
+			PipelineCacheLog("Vulkan pipeline cache: invalidating {} (invalid file size)",
+			                 path);
+		}
+	}
+
+	vk::PipelineCacheCreateInfo create {};
+	create.sType           = vk::StructureType::ePipelineCacheCreateInfo;
+	create.initialDataSize = initial_data.size();
+	create.pInitialData    = initial_data.empty() ? nullptr : initial_data.data();
+	auto result = m_graphics.device.createPipelineCache(&create, nullptr, &m_driver_cache);
+	if (result != vk::Result::eSuccess && !initial_data.empty()) {
+		PipelineCacheLog("Vulkan pipeline cache: driver rejected {} ({}); starting empty", path,
+		                 VulkanToString(result));
+		initial_data.clear();
+		create.initialDataSize = 0;
+		create.pInitialData    = nullptr;
+		result = m_graphics.device.createPipelineCache(&create, nullptr, &m_driver_cache);
+	}
+	if (result != vk::Result::eSuccess) {
+		PipelineCacheLog("Vulkan pipeline cache: disabled ({})", VulkanToString(result));
+		m_driver_cache = nullptr;
+		return;
+	}
+	if (!initial_data.empty()) {
+		PipelineCacheLog("Vulkan pipeline cache: loaded {} bytes from {}", initial_data.size(),
+		                 path);
+	} else {
+		PipelineCacheLog("Vulkan pipeline cache: initialized empty");
+	}
+}
+
+void PipelineCache::Save() {
+	Common::LockGuard lock(m_mutex);
+	if (m_driver_cache == nullptr) {
+		return;
+	}
+
+	size_t               size = 0;
+	vk::Result           result;
+	std::vector<uint8_t> payload;
+	for (uint32_t attempt = 0; attempt < 3; attempt++) {
+		size   = 0;
+		result = m_graphics.device.getPipelineCacheData(m_driver_cache, &size, nullptr);
+		if (result != vk::Result::eSuccess || size == 0 ||
+		    size > std::numeric_limits<uint32_t>::max()) {
+			break;
+		}
+		payload.resize(size);
+		result = m_graphics.device.getPipelineCacheData(m_driver_cache, &size, payload.data());
+		if (result != vk::Result::eIncomplete) {
+			break;
+		}
+	}
+	if (result != vk::Result::eSuccess || size == 0 ||
+	    size > std::numeric_limits<uint32_t>::max()) {
+		PipelineCacheLog("Vulkan pipeline cache: save failed ({}, {} bytes)",
+		                 VulkanToString(result), size);
+		return;
+	}
+	payload.resize(size);
+	auto prefix = DriverCacheSignature(m_graphics.GetPhysicalDeviceProperties());
+	const auto payload_hash = XXH3_64bits(payload.data(), payload.size());
+	prefix.append(reinterpret_cast<const char*>(&payload_hash), sizeof(payload_hash));
+	if (!Common::File::CreateDirectories(m_driver_cache_path.parent_path())) {
+		PipelineCacheLog("Vulkan pipeline cache: failed to create cache directory");
+		return;
+	}
+	auto temp_path = m_driver_cache_path;
+	temp_path += ".tmp";
+	Common::File file;
+	uint32_t     prefix_written  = 0;
+	uint32_t     payload_written = 0;
+	if (file.Create(temp_path)) {
+		file.Write(prefix.data(), static_cast<uint32_t>(prefix.size()), &prefix_written);
+		file.Write(payload.data(), static_cast<uint32_t>(payload.size()), &payload_written);
+	}
+	const bool flushed = !file.IsInvalid() && file.Flush();
+	file.Close();
+	if (prefix_written != prefix.size() || payload_written != payload.size() || !flushed ||
+	    !Common::File::RenameFile(temp_path, m_driver_cache_path)) {
+		PipelineCacheLog("Vulkan pipeline cache: failed to write {}",
+		                 Common::PathToString(m_driver_cache_path));
+		return;
+	}
+	PipelineCacheLog("Vulkan pipeline cache: saved {} bytes to {}", payload.size(),
+	                 Common::PathToString(m_driver_cache_path));
+	m_graphics.device.destroyPipelineCache(m_driver_cache, nullptr);
+	m_driver_cache = nullptr;
 }
 
 ShaderProgram PipelineCache::GetVertexProgram(const HW::VertexShaderInfo& regs,
                                               const HW::ShaderRegisters&  sh,
+                                              const HW::Context&          context,
                                               ShaderVertexInputInfo&      input_info) {
 	const auto params = PrepareProgram(regs, sh, input_info);
+	if (context.GetClipControl().clip_disable) {
+		const auto& viewport = context.GetScreenViewport().viewports[0];
+		const auto& limits   = m_graphics.GetPhysicalDeviceProperties().limits;
+		auto&       clip     = input_info.clip_space;
+		clip.scale[0]        = viewport.xscale;
+		clip.scale[1]        = viewport.yscale;
+		clip.offset[0]       = viewport.xoffset;
+		clip.offset[1]       = viewport.yoffset;
+		clip.half_extent[0] =
+		    static_cast<float>(std::min(limits.maxViewportDimensions[0], 16384u)) * 0.5f;
+		clip.half_extent[1] =
+		    static_cast<float>(std::min(limits.maxViewportDimensions[1], 16384u)) * 0.5f;
+		clip.enabled = true;
+	}
 	Common::LockGuard lock(m_mutex);
 	return m_program_cache->Get(params, input_info);
 }
@@ -324,7 +537,7 @@ PipelineCache::GraphicsPipeline& PipelineCache::CreateGraphicsPipeline(
 	auto cached = std::make_unique<GraphicsPipeline>(p);
 	LogPipelineTrace("CreatePipelineInternal begin", vs_id, ps_id);
 	CreatePipelineInternal(m_graphics, *cached, rendering, vs_input_info, vertex_program.module,
-	                       ps_input_info, pixel_program.module, static_params);
+	                       ps_input_info, pixel_program.module, static_params, m_driver_cache);
 	LogPipelineTrace("CreatePipelineInternal done", vs_id, ps_id);
 
 	EXIT_NOT_IMPLEMENTED(cached->pipeline == nullptr);
@@ -360,7 +573,7 @@ PipelineCache::CreateComputePipeline(ShaderComputeInputInfo& input_info,
 	}
 
 	auto cached = std::make_unique<ComputePipeline>(p);
-	CreatePipelineInternal(m_graphics, *cached, input_info, compute_program.module);
+	CreatePipelineInternal(m_graphics, *cached, input_info, compute_program.module, m_driver_cache);
 
 	EXIT_NOT_IMPLEMENTED(cached->pipeline == nullptr);
 	EXIT_NOT_IMPLEMENTED(cached->pipeline_layout == nullptr);

src/graphics/host_gpu/renderer/pipeline/pipelineCache.h

@@ -10,6 +10,7 @@
 #include "graphics/shader/shader.h"
 
 #include <cstddef>
+#include <filesystem>
 #include <memory>
 #include <span>
 #include <type_traits>
@@ -109,6 +110,7 @@ public:
 	explicit PipelineCache(GraphicContext& graphics);
 	~PipelineCache();
 	KYTY_CLASS_NO_COPY(PipelineCache);
+	void Save();
 
 	struct Pipeline {
 		vk::PipelineLayout      pipeline_layout       = nullptr;
@@ -128,6 +130,7 @@ public:
 
 	ShaderProgram GetVertexProgram(const HW::VertexShaderInfo& regs,
 	                               const HW::ShaderRegisters&  sh,
+	                               const HW::Context&          context,
 	                               ShaderVertexInputInfo&      input_info);
 	ShaderProgram
 	GetPixelProgram(const HW::PixelShaderInfo& regs, const HW::ShaderRegisters& sh,
@@ -214,12 +217,16 @@ private:
 
 	GraphicContext&                m_graphics;
 	std::unique_ptr<ProgramCache> m_program_cache;
+	vk::PipelineCache             m_driver_cache = nullptr;
+	std::filesystem::path         m_driver_cache_path;
 	std::unordered_map<GraphicsPipelineKey, std::unique_ptr<GraphicsPipeline>,
 	                   GraphicsPipelineKeyHash>
 	    m_graphics_pipelines;
 	std::unordered_map<ComputePipelineKey, std::unique_ptr<ComputePipeline>, ComputePipelineKeyHash>
 	              m_compute_pipelines;
 	Common::Mutex m_mutex;
+
+	void InitializeDriverCache();
 };
 
 void LogPipelineTrace(const char* phase, uint64_t vertex_program_id, uint64_t pixel_program_id);
@@ -227,10 +234,11 @@ void CreatePipelineInternal(
     GraphicContext& graphics, PipelineCache::GraphicsPipeline& pipeline,
     const PipelineRenderingState& rendering, const ShaderVertexInputInfo& vs_input_info,
     vk::ShaderModule vertex_module, const ShaderPixelInputInfo* ps_input_info,
-	vk::ShaderModule pixel_module, const PipelineStaticParameters& static_params);
+	vk::ShaderModule pixel_module, const PipelineStaticParameters& static_params,
+	vk::PipelineCache driver_cache);
 void CreatePipelineInternal(GraphicContext& graphics, PipelineCache::ComputePipeline& pipeline,
                             const ShaderComputeInputInfo& input_info,
-                            vk::ShaderModule              compute_module);
+	vk::ShaderModule compute_module, vk::PipelineCache driver_cache);
 
 } // namespace Libs::Graphics
 

src/graphics/host_gpu/renderer/pipeline/shaders.cpp

@@ -423,7 +423,8 @@ void CreatePipelineInternal(
     GraphicContext& graphics, PipelineCache::GraphicsPipeline& pipeline,
     const PipelineRenderingState& rendering, const ShaderVertexInputInfo& vs_input_info,
     vk::ShaderModule vertex_module, const ShaderPixelInputInfo* ps_input_info,
-	vk::ShaderModule pixel_module, const PipelineStaticParameters& static_params) {
+	vk::ShaderModule pixel_module, const PipelineStaticParameters& static_params,
+	vk::PipelineCache driver_cache) {
 	const bool ps_active = ps_input_info != nullptr;
 	EXIT_IF(vertex_module == nullptr || (ps_active && pixel_module == nullptr));
 
@@ -940,7 +941,7 @@ void CreatePipelineInternal(
 		     viewport.y, viewport.width, viewport.height, scissor.offset.x, scissor.offset.y,
 		     scissor.extent.width, scissor.extent.height);
 	}
-	result = graphics.device.createGraphicsPipelines(nullptr, 1, &pipeline_info, nullptr,
+	result = graphics.device.createGraphicsPipelines(driver_cache, 1, &pipeline_info, nullptr,
 	                                                 &pipeline.pipeline);
 	if (graphics_debug_dump_enabled()) {
 		LOGF("PipelineTrace: vkCreateGraphicsPipelines done result=%s pipeline=%p\n",
@@ -961,7 +962,7 @@ void CreatePipelineInternal(
 // NOLINTNEXTLINE(readability-function-cognitive-complexity)
 void CreatePipelineInternal(GraphicContext& graphics, PipelineCache::ComputePipeline& pipeline,
                             const ShaderComputeInputInfo& input_info,
-                            vk::ShaderModule              compute_module) {
+	vk::ShaderModule compute_module, vk::PipelineCache driver_cache) {
 	EXIT_IF(compute_module == nullptr);
 
 	vk::PipelineShaderStageCreateInfo                     comp_shader_stage_info {};
@@ -1024,7 +1025,8 @@ void CreatePipelineInternal(GraphicContext& graphics, PipelineCache::ComputePipe
 
 	LOGF("PipelineTrace: vkCreateComputePipelines begin layout=%p\n",
 	     static_cast<void*>(pipeline.pipeline_layout));
-	result = graphics.device.createComputePipelines(nullptr, 1, &info, nullptr, &pipeline.pipeline);
+	result = graphics.device.createComputePipelines(driver_cache, 1, &info, nullptr,
+	                                                &pipeline.pipeline);
 	LOGF("PipelineTrace: vkCreateComputePipelines done result=%s pipeline=%p\n",
 	     VulkanToString(result).c_str(), static_cast<void*>(pipeline.pipeline));
 	EXIT_NOT_IMPLEMENTED(result != vk::Result::eSuccess);

src/graphics/host_gpu/renderer/renderCompute.cpp

@@ -54,7 +54,9 @@ bool RenderExecutor::TryConsumeComputeMetaClear(const ShaderComputeInputInfo& in
 	for (uint32_t i = 0; i < program.info.buffers.size(); i++) {
 		const auto& resource   = program.info.buffers[i];
 		const auto  descriptor = DecodeNativeDescriptor<ShaderBufferResource>(resources.buffers[i]);
-		if (!resource.written && cache.IsMeta(descriptor.Base48())) {
+		// A metadata resource that is also read is not proven to be a full overwrite. Execute it
+		// conservatively instead of replacing the dispatch with a coarse full-surface clear.
+		if (cache.IsMeta(descriptor.Base48()) && (!resource.written || resource.read)) {
 			return false;
 		}
 	}

src/graphics/host_gpu/renderer/renderDraw.cpp

@@ -331,10 +331,18 @@ static void SetGraphicsDynamicParams(const CommandBuffer& buffer, vk::CommandBuf
 	const auto final_scissor = calc_final_scissor(vp, ctx.GetScanModeControl(), framebuffer_extent);
 
 	vk::Viewport viewport {};
-	viewport.x        = vp.viewports[0].xoffset - vp.viewports[0].xscale;
-	viewport.y        = vp.viewports[0].yoffset - vp.viewports[0].yscale;
-	viewport.width    = vp.viewports[0].xscale * 2.0f;
-	viewport.height   = vp.viewports[0].yscale * 2.0f;
+	if (ctx.GetClipControl().clip_disable) {
+		const auto& limits = buffer.GetGraphics().GetPhysicalDeviceProperties().limits;
+		viewport.x         = 0.0f;
+		viewport.y         = 0.0f;
+		viewport.width     = static_cast<float>(std::min(limits.maxViewportDimensions[0], 16384u));
+		viewport.height    = static_cast<float>(std::min(limits.maxViewportDimensions[1], 16384u));
+	} else {
+		viewport.x      = vp.viewports[0].xoffset - vp.viewports[0].xscale;
+		viewport.y      = vp.viewports[0].yoffset - vp.viewports[0].yscale;
+		viewport.width  = vp.viewports[0].xscale * 2.0f;
+		viewport.height = vp.viewports[0].yscale * 2.0f;
+	}
 	viewport.minDepth = vp.viewports[0].zoffset;
 	viewport.maxDepth = vp.viewports[0].zscale + vp.viewports[0].zoffset;
 	vk_buffer.setViewport(0, 1, &viewport);
@@ -411,7 +419,7 @@ static void SetGraphicsDynamicParams(const CommandBuffer& buffer, vk::CommandBuf
 static bool DrawHasValidVertexShader(const HW::Shader& sh_ctx) {
 
 	const auto& vs = sh_ctx.GetVs();
-	return vs.gs_regs.chksum != 0 && ShaderAddressValid(vs.es_regs.data_addr);
+	return ShaderAddressValid(vs.es_regs.data_addr);
 }
 
 static bool PixelShaderHasDepthOrCoverageSideEffects(const HW::ShaderRegisters& sh_regs) {
@@ -443,7 +451,6 @@ static bool ShouldSkipGeShader(const CommandBuffer& buffer) {
 	};
 
 	const bool ps5_ngg_vertex_path = stages == 0x02002000 && vertex_info.es_regs.data_addr != 0 &&
-	                                 vertex_info.gs_regs.chksum != 0 &&
 	                                 sh_regs.m_vgtGsMaxVertOut == 0x00000000 &&
 	                                 is_known_gs_out_prim_type(sh_regs.m_vgtGsOutPrimType);
 
@@ -1014,8 +1021,8 @@ static void RefreshShaders(CommandBuffer& buffer, const DrawCallInfo& draw, bool
 		LogDrawPhase(draw.name, "GetVertexProgram");
 	}
 	auto& pipeline_cache = buffer.GetContext().GetPipelineCache();
-	state.vertex_program =
-	    pipeline_cache.GetVertexProgram(vertex_shader_info, shader_regs, state.vs_input_info);
+	state.vertex_program = pipeline_cache.GetVertexProgram(vertex_shader_info, shader_regs, ctx,
+	                                                         state.vs_input_info);
 
 	if (!state.ps_active) {
 		return;
@@ -1450,7 +1457,7 @@ void RenderExecutor::DrawAuto(uint64_t submit_id, CommandBuffer& buffer, uint32_
 		if (graphics_debug_dump_enabled()) {
 			LOGF("DrawIndexAuto: skipping rect-list draw with no VS param exports and PS inputs: "
 			     "ps_inputs=%u ps=0x%016" PRIx64 " es=0x%016" PRIx64 " gs=0x%016" PRIx64 "\n",
-			     state.ps_input_info.input_num, sh_ctx.GetPs().ps_regs.chksum,
+			     state.ps_input_info.input_num, sh_ctx.GetPs().ps_regs.data_addr,
 			     sh_ctx.GetVs().es_regs.data_addr, sh_ctx.GetVs().gs_regs.data_addr);
 		}
 		ResetBindings();

src/graphics/presentation/window/hostInput.cpp

@@ -1,18 +1,24 @@
 #include "graphics/presentation/window/hostInput.h"
 
+#include "SDL_error.h"
+#include "SDL_events.h"
 #include "SDL_keyboard.h"
 #include "SDL_keycode.h"
 #include "SDL_mouse.h"
+#include "SDL_timer.h"
+#include "SDL_video.h"
 #include "common/assert.h"
 #include "common/emulatorConfig.h"
+#include "common/logging/log.h"
 #include "libs/controller.h"
 
 #include <algorithm>
 #include <array>
+#include <cmath>
 #include <cstddef>
+#include <cstdlib>
 #include <string>
 #include <string_view>
-#include <utility>
 
 namespace Libs::Graphics {
 
@@ -23,6 +29,7 @@ struct ControlInfo {
 	uint32_t         button   = 0;
 	Controller::Axis axis     = Controller::Axis::AxisMax;
 	bool             positive = false;
+	float            touch_x  = 0.0f;
 };
 
 static constexpr std::array CONTROL_INFO = {
@@ -41,7 +48,10 @@ static constexpr std::array CONTROL_INFO = {
     ControlInfo {"Circle", Controller::PAD_BUTTON_CIRCLE},
     ControlInfo {"Cross", Controller::PAD_BUTTON_CROSS},
     ControlInfo {"Square", Controller::PAD_BUTTON_SQUARE},
-    ControlInfo {"TouchPad", Controller::PAD_BUTTON_TOUCH_PAD},
+    ControlInfo {"TouchPad", Controller::PAD_BUTTON_TOUCH_PAD, Controller::Axis::AxisMax, false,
+                 0.25f},
+    ControlInfo {"TouchPadRight", Controller::PAD_BUTTON_TOUCH_PAD, Controller::Axis::AxisMax,
+                 false, 0.75f},
     ControlInfo {"LeftStickLeft", 0, Controller::Axis::LeftX, false},
     ControlInfo {"LeftStickRight", 0, Controller::Axis::LeftX, true},
     ControlInfo {"LeftStickUp", 0, Controller::Axis::LeftY, false},
@@ -60,17 +70,23 @@ struct Binding {
 	std::size_t control      = INVALID_CONTROL;
 };
 
+constexpr int              MOUSE_POLL_INTERVAL_MS = 33;
+constexpr std::string_view MOUSE_SENSITIVITY      = "MouseSensitivity=";
+
+struct MouseJoystickState {
+	bool     enabled   = false;
+	bool     output    = false;
+	uint64_t next_poll = 0;
+};
+
+MouseJoystickState g_mouse;
+
 std::size_t ControlFromName(std::string_view name) {
 	const auto info = std::find_if(CONTROL_INFO.begin(), CONTROL_INFO.end(),
 	                               [name](const auto& item) { return item.name == name; });
 	return static_cast<std::size_t>(std::distance(CONTROL_INFO.begin(), info));
 }
 
-const ControlInfo& Info(std::size_t control) {
-	EXIT_IF(control >= CONTROL_INFO.size());
-	return CONTROL_INFO[control];
-}
-
 SDL_Keycode NormalizeKey(SDL_Keycode key) {
 	switch (key) {
 		case SDLK_RSHIFT: return SDLK_LSHIFT;
@@ -103,10 +119,17 @@ bool Conflicts(const Binding& first, const Binding& second) {
 
 class InputMap {
 public:
-	InputMap(): m_custom(!Config::GetKeymap().empty()) {
+	InputMap() {
 		for (const auto& value: Config::GetKeymap()) {
 			const std::string_view entry = value;
-			const auto             split = entry.find('=');
+			if (entry.starts_with(MOUSE_SENSITIVITY)) {
+				const float sensitivity =
+				    std::strtof(value.c_str() + MOUSE_SENSITIVITY.size(), nullptr);
+				m_mouse_sensitivity =
+				    std::clamp(std::isfinite(sensitivity) ? sensitivity : 1.0f, 0.1f, 5.0f);
+				continue;
+			}
+			const auto split = entry.find('=');
 
 			Binding binding;
 			if (split != std::string_view::npos) {
@@ -118,9 +141,9 @@ public:
 				}
 			}
 
-			const bool reserved =
-			    binding.key == SDLK_ESCAPE || binding.key == SDLK_SPACE || binding.key == SDLK_F1 ||
-			    binding.key == SDLK_F11;
+			const bool reserved = binding.key == SDLK_ESCAPE || binding.key == SDLK_SPACE ||
+			                      binding.key == SDLK_F1 || binding.key == SDLK_F7 ||
+			                      binding.key == SDLK_F11;
 			if (binding.control == INVALID_CONTROL || reserved ||
 			    (binding.key == SDLK_UNKNOWN && binding.mouse_button == 0)) {
 				EXIT("Invalid input mapping: %s\n", value.c_str());
@@ -128,7 +151,8 @@ public:
 			Add(binding);
 		}
 	}
-	[[nodiscard]] bool Custom() const { return m_custom; }
+	[[nodiscard]] bool  Custom() const { return m_size != 0; }
+	[[nodiscard]] float MouseSensitivity() const { return m_mouse_sensitivity; }
 
 	[[nodiscard]] std::size_t FindKey(int key_code) const {
 		key_code = NormalizeKey(static_cast<SDL_Keycode>(key_code));
@@ -159,8 +183,8 @@ private:
 	}
 
 	std::array<Binding, CONTROL_INFO.size()> m_bindings {};
-	std::size_t                              m_size = 0;
-	bool                                     m_custom;
+	std::size_t                              m_size              = 0;
+	float                                    m_mouse_sensitivity = 1.0f;
 };
 
 const InputMap& GetInputMap() {
@@ -180,6 +204,10 @@ void SetButton(uint32_t button, bool down) {
 	}
 }
 
+void SetTouchPad(float x, bool down) {
+	Controller::ControllerTouchPad(Controller::HOST_INPUT_CONTROLLER_ID, 0, down, x, 0.5f);
+}
+
 uint32_t DefaultKeyboardButton(int key_code) {
 	switch (NormalizeKey(static_cast<SDL_Keycode>(key_code))) {
 		case SDLK_UP: return Controller::PAD_BUTTON_UP;
@@ -196,8 +224,6 @@ uint32_t DefaultKeyboardButton(int key_code) {
 		case SDLK_LCTRL: return Controller::PAD_BUTTON_R3;
 		case SDLK_RETURN:
 		case SDLK_RETURN2: return Controller::PAD_BUTTON_OPTIONS;
-		case SDLK_BACKSPACE:
-		case SDLK_TAB: return Controller::PAD_BUTTON_TOUCH_PAD;
 		default: return 0;
 	}
 }
@@ -210,7 +236,12 @@ struct StickKeys {
 };
 
 void SetStickAxis(Controller::Axis axis, bool negative, bool positive) {
-	const int value = negative == positive ? 128 : negative ? 0 : 255;
+	int value = 128;
+	if (negative && !positive) {
+		value = 0;
+	} else if (positive && !negative) {
+		value = 255;
+	}
 	Controller::ControllerAxis(Controller::HOST_INPUT_CONTROLLER_ID, axis, value);
 }
 
@@ -219,7 +250,11 @@ void SetControl(std::size_t control, bool down) {
 		return;
 	}
 
-	const auto& info = Info(control);
+	const auto& info = CONTROL_INFO[control];
+	if (info.button == Controller::PAD_BUTTON_TOUCH_PAD) {
+		SetTouchPad(info.touch_x, down);
+		return;
+	}
 	if (info.button != 0) {
 		SetButton(info.button, down);
 		return;
@@ -246,6 +281,8 @@ void DefaultKeyboardInput(int key_code, bool down) {
 	static StickKeys right;
 
 	switch (NormalizeKey(static_cast<SDL_Keycode>(key_code))) {
+		case SDLK_BACKSPACE: SetTouchPad(0.25f, down); return;
+		case SDLK_TAB: SetTouchPad(0.75f, down); return;
 		case SDLK_a:
 			left.left = down;
 			SetStickAxis(Controller::Axis::LeftX, left.left, left.right);
@@ -281,6 +318,33 @@ void DefaultKeyboardInput(int key_code, bool down) {
 		default: SetButton(DefaultKeyboardButton(key_code), down); return;
 	}
 }
+
+void MouseToJoystick(int delta_x, int delta_y) {
+	const double distance = std::hypot(delta_x, delta_y);
+	const double scale =
+	    std::clamp(distance * GetInputMap().MouseSensitivity() + 16.0, 64.0, 128.0) / distance;
+	const auto map_axis = [scale](int delta) {
+		return std::clamp(128 + static_cast<int>(std::lround(delta * scale)), 0, 255);
+	};
+	Controller::ControllerRightStick(Controller::HOST_INPUT_CONTROLLER_ID, map_axis(delta_x),
+	                                 map_axis(delta_y));
+}
+
+void CenterMouseStick() {
+	if (!g_mouse.output) {
+		return;
+	}
+	Controller::ControllerRightStick(Controller::HOST_INPUT_CONTROLLER_ID, 128, 128);
+	g_mouse.output = false;
+}
+
+bool SetRelativeMouseMode(bool enabled) {
+	if (SDL_SetRelativeMouseMode(enabled ? SDL_TRUE : SDL_FALSE) == 0) {
+		return true;
+	}
+	LOGF("Mouse-to-joystick relative mode failed: %s\n", SDL_GetError());
+	return false;
+}
 } // namespace
 
 void HostInputInit() {
@@ -303,4 +367,63 @@ void HostInputMouseButton(uint8_t mouse_button, bool down) {
 	}
 }
 
+void HostInputToggleMouseToJoystick() {
+	if (g_mouse.enabled) {
+		SetRelativeMouseMode(false);
+		CenterMouseStick();
+		g_mouse = {};
+		LOGF("Mouse to right stick: disabled\n");
+		return;
+	}
+
+	if (!SetRelativeMouseMode(true)) {
+		return;
+	}
+	int ignored_x = 0;
+	int ignored_y = 0;
+	SDL_GetRelativeMouseState(&ignored_x, &ignored_y);
+	g_mouse.enabled   = true;
+	g_mouse.next_poll = SDL_GetTicks64() + MOUSE_POLL_INTERVAL_MS;
+	LOGF("Mouse to right stick: enabled (F7 to release)\n");
+}
+
+int PollMouse(uint64_t now_ms) {
+	if (now_ms < g_mouse.next_poll) {
+		return static_cast<int>(g_mouse.next_poll - now_ms);
+	}
+	g_mouse.next_poll = now_ms + MOUSE_POLL_INTERVAL_MS;
+
+	int delta_x = 0;
+	int delta_y = 0;
+	SDL_GetRelativeMouseState(&delta_x, &delta_y);
+	if (delta_x == 0 && delta_y == 0) {
+		CenterMouseStick();
+		return MOUSE_POLL_INTERVAL_MS;
+	}
+
+	MouseToJoystick(delta_x, delta_y);
+	g_mouse.output = true;
+	return MOUSE_POLL_INTERVAL_MS;
+}
+
+bool HostInputWaitEvent(SDL_Event* event) {
+	if (!g_mouse.enabled || SDL_GetKeyboardFocus() == nullptr) {
+		CenterMouseStick();
+		if (SDL_WaitEvent(event) == 0) {
+			EXIT("%s\n", SDL_GetError());
+		}
+		return true;
+	}
+
+	const int timeout_ms = PollMouse(SDL_GetTicks64());
+	SDL_ClearError();
+	if (SDL_WaitEventTimeout(event, timeout_ms) != 0) {
+		return true;
+	}
+	if (SDL_GetError()[0] != '\0') {
+		EXIT("%s\n", SDL_GetError());
+	}
+	return false;
+}
+
 } // namespace Libs::Graphics

src/graphics/presentation/window/hostInput.h

@@ -3,11 +3,15 @@
 
 #include <cstdint>
 
+union SDL_Event;
+
 namespace Libs::Graphics {
 
-void HostInputInit();
-void HostInputKey(int key_code, bool down);
-void HostInputMouseButton(uint8_t mouse_button, bool down);
+void               HostInputInit();
+void               HostInputKey(int key_code, bool down);
+void               HostInputMouseButton(uint8_t mouse_button, bool down);
+void               HostInputToggleMouseToJoystick();
+[[nodiscard]] bool HostInputWaitEvent(SDL_Event* event);
 
 } // namespace Libs::Graphics
 

src/graphics/presentation/window/vulkanWindow.cpp

@@ -157,6 +157,10 @@ static void VulkanFindPhysicalDevice(vk::Instance instance, vk::SurfaceKHR surfa
 	    });
 	EXIT_NOT_IMPLEMENTED(devices.empty());
 
+	if (Config::GetGpuIndex() >= 0) {
+		devices = {devices[Config::GetGpuIndex()]};
+	}
+
 	vk::PhysicalDevice  best_device       = nullptr;
 	uint32_t            best_queue_family = static_cast<uint32_t>(-1);
 	SurfaceCapabilities best_capabilities;
@@ -942,6 +946,18 @@ void WindowContext::CreateVulkan() {
 	inst_info.sType                   = vk::StructureType::eInstanceCreateInfo;
 	inst_info.pNext                   = (r.enable_validation_layers ? &dbg_create_info : nullptr);
 	inst_info.flags                   = {};
+#if defined(__APPLE__)
+	// MoltenVK requires VK_KHR_portability_enumeration + flag to surface
+	// portability devices. Without this, enumeratePhysicalDevices hides the
+	// MoltenVK adapter on some driver versions.
+	if (HasExtension(r.available_extensions, VK_KHR_PORTABILITY_ENUMERATION_EXTENSION_NAME)) {
+		if (!HasExtension(r.required_extensions, VK_KHR_PORTABILITY_ENUMERATION_EXTENSION_NAME)) {
+			r.required_extensions.push_back(VK_KHR_PORTABILITY_ENUMERATION_EXTENSION_NAME);
+		}
+		inst_info.flags |= vk::InstanceCreateFlagBits::eEnumeratePortabilityKHR;
+		LOGF("Vulkan instance: enabled %s\n", VK_KHR_PORTABILITY_ENUMERATION_EXTENSION_NAME);
+	}
+#endif
 	inst_info.pApplicationInfo        = &app_info;
 	inst_info.enabledExtensionCount   = static_cast<uint32_t>(r.required_extensions.size());
 	inst_info.ppEnabledExtensionNames = r.required_extensions.data();

src/graphics/presentation/window/window.cpp

@@ -3,6 +3,7 @@
 #include "SDL.h"
 #include "SDL_error.h"
 #include "SDL_events.h"
+#include "SDL_filesystem.h"
 #include "SDL_gamecontroller.h"
 #include "SDL_hints.h"
 #include "SDL_joystick.h"
@@ -28,6 +29,7 @@
 #include "common/timer.h"
 #include "graphics/host_gpu/graphicContext.h"
 #include "graphics/host_gpu/renderer/render.h"
+#include "graphics/host_gpu/renderer/renderContext.h"
 #include "graphics/host_gpu/vma.h"
 #include "graphics/host_gpu/vulkanCommon.h"
 #include "graphics/presentation/imeOverlay.h"
@@ -39,13 +41,12 @@
 #include "loader/systemContent.h"
 
 #include <cstdlib>
+#include <fmt/format.h>
 #include <memory>
 #include <string>
 #include <vector>
 #include <vulkan/vk_platform.h>
 
-#include <fmt/format.h>
-
 #define STB_IMAGE_IMPLEMENTATION
 #define STBI_NO_SIMD
 #include "stb_image.h"
@@ -75,7 +76,6 @@ static uint32_t ControllerButtonToPadButton(int button) {
 		case SDL_CONTROLLER_BUTTON_B: return Controller::PAD_BUTTON_CIRCLE;
 		case SDL_CONTROLLER_BUTTON_X: return Controller::PAD_BUTTON_SQUARE;
 		case SDL_CONTROLLER_BUTTON_Y: return Controller::PAD_BUTTON_TRIANGLE;
-		case SDL_CONTROLLER_BUTTON_BACK: return Controller::PAD_BUTTON_TOUCH_PAD;
 		case SDL_CONTROLLER_BUTTON_START: return Controller::PAD_BUTTON_OPTIONS;
 		case SDL_CONTROLLER_BUTTON_LEFTSTICK: return Controller::PAD_BUTTON_L3;
 		case SDL_CONTROLLER_BUTTON_RIGHTSTICK: return Controller::PAD_BUTTON_R3;
@@ -85,6 +85,7 @@ static uint32_t ControllerButtonToPadButton(int button) {
 		case SDL_CONTROLLER_BUTTON_DPAD_DOWN: return Controller::PAD_BUTTON_DOWN;
 		case SDL_CONTROLLER_BUTTON_DPAD_LEFT: return Controller::PAD_BUTTON_LEFT;
 		case SDL_CONTROLLER_BUTTON_DPAD_RIGHT: return Controller::PAD_BUTTON_RIGHT;
+		case SDL_CONTROLLER_BUTTON_TOUCHPAD: return Controller::PAD_BUTTON_TOUCH_PAD;
 		default: return 0;
 	}
 }
@@ -241,7 +242,6 @@ static void GameEventKeyboard(WindowLoopState& game, const EventKeyboard& key) {
 #if KYTY_PLATFORM == KYTY_PLATFORM_WINDOWS || KYTY_PLATFORM == KYTY_PLATFORM_LINUX
 	if (key.down) {
 		switch (key.key_code) {
-			case SDLK_ESCAPE: game.need_exit = true; break;
 			case SDLK_SPACE: SetPause(game, !game.paused.load(std::memory_order_acquire)); break;
 			case SDLK_F1:
 				if (!key.repeat) {
@@ -506,6 +506,13 @@ void WindowContext::ProcessEvent(double time_s) {
 	auto& game  = loop;
 	auto* event = &game.event;
 	EXIT_IF(SDL_GetEventState(SDL_DISPLAYEVENT) != SDL_ENABLE);
+	if ((event->type == SDL_KEYDOWN || event->type == SDL_KEYUP) &&
+	    event->key.keysym.sym == SDLK_F7) {
+		if (event->type == SDL_KEYDOWN && event->key.repeat == 0) {
+			HostInputToggleMouseToJoystick();
+		}
+		return;
+	}
 	if (ProcessImeInput(*event)) {
 		return;
 	}
@@ -697,6 +704,16 @@ void WindowContext::ProcessEvent(double time_s) {
 			break;
 		}
 
+		case SDL_CONTROLLERTOUCHPADDOWN:
+		case SDL_CONTROLLERTOUCHPADMOTION:
+		case SDL_CONTROLLERTOUCHPADUP:
+			if (event->ctouchpad.touchpad == 0) {
+				Controller::ControllerTouchPad(event->ctouchpad.which, event->ctouchpad.finger,
+				                               event->ctouchpad.type != SDL_CONTROLLERTOUCHPADUP,
+				                               event->ctouchpad.x, event->ctouchpad.y);
+			}
+			break;
+
 		case SDL_CONTROLLERDEVICEADDED:
 		case SDL_CONTROLLERDEVICEREMOVED:
 		case SDL_CONTROLLERDEVICEREMAPPED: {
@@ -782,8 +799,8 @@ void WindowContext::Run() {
 			timer.Resume();
 		}
 
-		if (SDL_WaitEvent(&loop.event) == 0) {
-			EXIT("%s\n", SDL_GetError());
+		if (!HostInputWaitEvent(&loop.event)) {
+			continue;
 		}
 		ProcessEvent(timer.GetTimeS());
 	}
@@ -811,6 +828,24 @@ static void WindowCreate(WindowContext& context) {
 
 	uint32_t window_flags = WindowContext::InitialWindowFlags(Config::FullscreenEnabled());
 #if defined(__APPLE__)
+	// SDL loads Vulkan while creating a Vulkan window, so select the bundled
+	// MoltenVK loader before calling SDL_CreateWindow. Keep an explicit user
+	// override, and fall back to SDL's normal loader search when no bundle is present.
+	if (std::getenv("SDL_VULKAN_LIBRARY") == nullptr) {
+		if (char* base_path = SDL_GetBasePath(); base_path != nullptr) {
+			const std::string base_path_str = base_path;
+			SDL_free(base_path);
+			std::string moltenvk_path = base_path_str + "libMoltenVK.dylib";
+			if (!Common::File::IsFileExisting(moltenvk_path)) {
+				moltenvk_path = base_path_str + "../Frameworks/libMoltenVK.dylib";
+			}
+			if (Common::File::IsFileExisting(moltenvk_path) &&
+			    SDL_setenv("SDL_VULKAN_LIBRARY", moltenvk_path.c_str(), 0) == 0) {
+				LOGF("Vulkan loader: %s\n", moltenvk_path.c_str());
+			}
+		}
+	}
+
 	// macOS 26 window chrome (CoreUI asset decode, SwiftUI titlebar) has been observed
 	// throwing NSExceptions under Rosetta during the first CATransaction commit. A
 	// borderless window skips that machinery entirely.
@@ -858,6 +893,7 @@ void WindowRun() {
 	EXIT_IF(g_window == nullptr);
 
 	g_window->Run();
+	g_window->render_context->GetPipelineCache().Save();
 }
 
 void WindowShutdown() {

src/graphics/shader/recompiler/ShaderRecompiler.cpp

@@ -30,14 +30,6 @@ namespace Libs::Graphics::ShaderRecompiler {
 
 namespace {
 
-bool ReadZeroMemory(void*, uint64_t, uint32_t* value) {
-	if (value == nullptr) {
-		return false;
-	}
-	*value = 0;
-	return true;
-}
-
 const char* GetDumpLabel(const CompileOptions& options) {
 	return options.dump_label != nullptr ? options.dump_label : "ShaderRecompiler";
 }
@@ -472,20 +464,14 @@ EmbeddedFetchData DetectEmbeddedVertexFetch(const Decoder::Program&      decoded
 
 } // namespace
 
-bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
-                  CompileResult& result, std::string* error) {
+CompileResult Recompile(std::span<const uint32_t> code, const CompileOptions& options) {
 	if (code.empty()) {
-		if (error != nullptr) {
-			*error = "invalid shader recompiler input";
-		}
-		return false;
+		EXIT("shader recompiler input is empty\n");
 	}
 	if (options.stage != ShaderType::Compute && options.stage != ShaderType::Vertex &&
 	    options.stage != ShaderType::Pixel) {
-		if (error != nullptr) {
-			*error = "shader recompiler supports compute, vertex, and pixel stages";
-		}
-		return false;
+		EXIT("shader recompiler received unsupported stage %u\n",
+		     static_cast<unsigned>(options.stage));
 	}
 
 	const auto compile_begin = std::chrono::steady_clock::now();
@@ -500,9 +486,7 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 	     static_cast<uint64_t>(code.size()));
 
 	Decoder::Program decoded;
-	if (!Decoder::DecodeProgram(code, decoded, error)) {
-		return false;
-	}
+	Decoder::DecodeProgram(code, decoded);
 	LOGF("%s phase end: stage=%s hash=0x%016" PRIx64 " decode instructions=%" PRIu64
 	     " elapsed_ms=%" PRIu64 "\n",
 	     GetDumpLabel(options), StageName(options.stage), options.shader_hash,
@@ -516,12 +500,9 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 		}
 	}
 
-	CFG::Graph cfg;
 	LOGF("%s phase begin: stage=%s hash=0x%016" PRIx64 " CFG BuildGraph\n", GetDumpLabel(options),
 	     StageName(options.stage), options.shader_hash);
-	if (!CFG::BuildGraph(decoded, cfg, error)) {
-		return false;
-	}
+	auto cfg = CFG::BuildGraph(decoded);
 	LOGF("%s phase end: stage=%s hash=0x%016" PRIx64 " CFG BuildGraph blocks=%" PRIu64
 	     " loops=%" PRIu64 " back_edges=%" PRIu64 " elapsed_ms=%" PRIu64 "\n",
 	     GetDumpLabel(options), StageName(options.stage), options.shader_hash,
@@ -534,13 +515,12 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 		dispatcher_reason   = cfg.unsupported_reason;
 		LogDispatcherFallback(options, cfg, "build", dispatcher_reason);
 	} else {
-		std::string structure_error;
-		const auto  unstructured_cfg = cfg;
+		const auto unstructured_cfg = cfg;
 		LOGF("%s phase begin: stage=%s hash=0x%016" PRIx64 " CFG Structurize\n",
 		     GetDumpLabel(options), StageName(options.stage), options.shader_hash);
-		if (!CFG::Structurize(cfg, &structure_error)) {
+		if (!CFG::Structurize(cfg)) {
 			dispatcher_fallback      = true;
-			dispatcher_reason        = structure_error;
+			dispatcher_reason        = cfg.unsupported_reason;
 			const auto failure_kind  = cfg.failure_kind;
 			const auto failure_block = cfg.failure_block;
 			LogDispatcherFallback(options, cfg, "structurize", dispatcher_reason);
@@ -548,7 +528,7 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 			cfg.unsupported        = true;
 			cfg.failure_kind       = failure_kind;
 			cfg.failure_block      = failure_block;
-			cfg.unsupported_reason = structure_error;
+			cfg.unsupported_reason = dispatcher_reason;
 		} else {
 			LOGF("%s structured CFG success: blocks=%" PRIu64 "\n", GetDumpLabel(options),
 			     static_cast<uint64_t>(cfg.blocks.size()));
@@ -560,20 +540,26 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 		     static_cast<uint64_t>(cfg.natural_loops.size()), phase_ms());
 	}
 
-	IR::Program ir;
 	const ShaderVertexInputInfo*  vertex  = nullptr;
 	const ShaderPixelInputInfo*   pixel   = nullptr;
 	const ShaderComputeInputInfo* compute = nullptr;
 	switch (options.stage) {
-		case ShaderType::Vertex: vertex = options.input_info.vertex; break;
-		case ShaderType::Pixel: pixel = options.input_info.pixel; break;
-		case ShaderType::Compute: compute = options.input_info.compute; break;
+		case ShaderType::Vertex:
+			vertex = options.input_info.vertex;
+			break;
+		case ShaderType::Pixel:
+			pixel = options.input_info.pixel;
+			break;
+		case ShaderType::Compute:
+			compute = options.input_info.compute;
+			break;
 		default: break;
 	}
 	EmbeddedFetchData embedded_fetch;
 	if (options.stage == ShaderType::Vertex && vertex->fetch_embedded) {
 		embedded_fetch = DetectEmbeddedVertexFetch(decoded, vertex, options.user_data_base,
-		                                           options.user_data_count, options.wave_size);
+		                                           static_cast<uint32_t>(options.user_data.size()),
+		                                           options.wave_size);
 		if (!embedded_fetch.loads.empty()) {
 			LOGF("%s embedded vertex fetch plan: detected=%" PRIu64 "\n", GetDumpLabel(options),
 			     static_cast<uint64_t>(embedded_fetch.loads.size()));
@@ -584,7 +570,7 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 	    .wave_size           = options.wave_size,
 	    .shader_hash         = options.shader_hash,
 	    .user_data_base      = options.user_data_base,
-	    .user_data_count     = options.user_data_count,
+	    .user_data_count     = static_cast<uint32_t>(options.user_data.size()),
 	    .scratch_dwords      = options.scratch_dwords,
 	    .dispatcher_fallback = dispatcher_fallback,
 	    .cfg_failure_kind    = cfg.failure_kind,
@@ -597,9 +583,7 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 	};
 	LOGF("%s phase begin: stage=%s hash=0x%016" PRIx64 " IR TranslateProgram\n",
 	     GetDumpLabel(options), StageName(options.stage), options.shader_hash);
-	if (!Frontend::TranslateProgram(decoded, cfg, translate_options, ir, error)) {
-		return false;
-	}
+	auto ir = Frontend::TranslateProgram(decoded, cfg, translate_options);
 	LOGF("%s phase end: stage=%s hash=0x%016" PRIx64 " IR TranslateProgram blocks=%" PRIu64
 	     " elapsed_ms=%" PRIu64 "\n",
 	     GetDumpLabel(options), StageName(options.stage), options.shader_hash,
@@ -626,66 +610,35 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 			     lds_barriers.inserted_barriers);
 		}
 	}
-	std::string srt_error;
-	if (!IR::BuildSrtPlan(ir, &srt_error)) {
-		LOGF("%s SRT planning failed: %s\n", GetDumpLabel(options), srt_error.c_str());
-		if (error != nullptr) {
-			*error = std::move(srt_error);
-		}
-		return false;
-	}
+	IR::BuildSrtPlan(ir);
 	IR::EliminateDeadCode(ir.blocks);
-	if (!IR::TrackResources(ir, error)) {
-		return false;
-	}
+	IR::TrackResources(ir);
 	IR::EliminateDeadCode(ir.blocks);
 	if (options.stage == ShaderType::Vertex) {
 		ir.info.vertex_offset_sgpr = embedded_fetch.vertex_offset_sgpr;
 	}
 
+	const IR::SrtRuntime runtime {
+	    .user_data                  = options.user_data,
+	    .shader_base                = reinterpret_cast<uint64_t>(code.data()),
+	    .read_memory                = options.read_memory,
+	    .userdata                   = options.read_memory_data,
+	    .read_specialization_memory = options.read_specialization_memory,
+	};
 	IR::ResourceSnapshot resources;
-	if (options.resource_snapshot != nullptr) {
-		resources = *options.resource_snapshot;
-	} else {
-		IR::SrtRuntime           runtime;
-		std::array<uint32_t, 64> zero_user_data {};
-		if (options.user_data == nullptr) {
-			for (uint32_t i = 2; i < zero_user_data.size(); i += 4) {
-				zero_user_data[i] = UINT32_MAX;
-			}
-		}
-		const auto* user_data =
-		    options.user_data != nullptr ? options.user_data : zero_user_data.data();
-		runtime.user_data   = std::span<const uint32_t>(user_data, options.user_data_count);
-		runtime.shader_base = options.shader_base != 0 ? options.shader_base
-		                                               : reinterpret_cast<uint64_t>(code.data());
-		runtime.read_memory = options.read_memory;
-		runtime.read_specialization_memory = options.read_specialization_memory;
-		if (runtime.read_memory == nullptr && options.user_data == nullptr) {
-			runtime.read_memory = ReadZeroMemory;
-		}
-		runtime.userdata         = options.read_memory_data;
-		if (!IR::MaterializeResources(ir, runtime, resources, error)) {
-			return false;
-		}
-	}
-	if (!IR::SpecializeResources(ir, resources, error)) {
-		return false;
+	if (!IR::MaterializeResources(ir, runtime, resources)) {
+		EXIT("shader resource materialization failed: stage=%s hash=0x%016" PRIx64 "\n",
+		     StageName(options.stage), options.shader_hash);
 	}
+	IR::SpecializeResources(ir, resources);
 
 	IR::ShaderInfoOptions info_options;
 	info_options.vertex  = vertex;
 	info_options.pixel   = pixel;
 	info_options.compute = compute;
-	if (!IR::CollectShaderInfo(ir, info_options, error)) {
-		return false;
-	}
-	if (!IR::AllocateBindings(ir, options.push_constant_offset, error)) {
-		return false;
-	}
-	if (!Spirv::AnalyzeProgramRequirements(ir, error)) {
-		return false;
-	}
+	IR::CollectShaderInfo(ir, info_options);
+	IR::AllocateBindings(ir, options.push_constant_offset);
+	Spirv::AnalyzeProgramRequirements(ir);
 	std::string ir_dump;
 	if (options.dump_ir) {
 		ir_dump = MakeIrDump(cfg, ir);
@@ -694,36 +647,22 @@ bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
 		}
 	}
 
-	std::vector<uint32_t> spirv;
-	std::string           emit_error;
 	LOGF("%s phase begin: stage=%s hash=0x%016" PRIx64 " SPIR-V EmitProgram\n",
 	     GetDumpLabel(options), StageName(options.stage), options.shader_hash);
-	if (!Spirv::EmitProgram(ir, resources, options.input_info, spirv, &emit_error)) {
-		LOGF("%s typed SPIR-V emit failed: %s\n", GetDumpLabel(options), emit_error.c_str());
-		if (dispatcher_fallback && error != nullptr) {
-			*error = fmt::format("dispatcher fallback failed after {}: {}",
-			                     dispatcher_reason.c_str(), emit_error.c_str());
-			LOGF("%s dispatcher fallback emit failed: %s\n", GetDumpLabel(options), error->c_str());
-		} else if (error != nullptr) {
-			*error = emit_error;
-		}
-		return false;
-	}
+	auto spirv = Spirv::EmitProgram(ir, resources, options.input_info);
 	LOGF("%s phase end: stage=%s hash=0x%016" PRIx64 " SPIR-V EmitProgram words=%" PRIu64
 	     " elapsed_ms=%" PRIu64 "\n",
 	     GetDumpLabel(options), StageName(options.stage), options.shader_hash,
 	     static_cast<uint64_t>(spirv.size()), phase_ms());
+	CompileResult result;
 	result.spirv     = std::move(spirv);
 	result.program   = std::move(ir);
 	result.resources = std::move(resources);
 	if (options.dump_ir) {
 		result.decoded_dump = std::move(decoded_dump);
 		result.ir_dump      = std::move(ir_dump);
-	} else {
-		result.decoded_dump.clear();
-		result.ir_dump.clear();
 	}
-	return true;
+	return result;
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler

src/graphics/shader/recompiler/ShaderRecompiler.h

@@ -15,19 +15,16 @@ struct CompileOptions {
 	ShaderType                  stage           = ShaderType::Compute;
 	uint32_t                    wave_size       = 64;
 	uint32_t                    user_data_base  = 0;
-	uint32_t                    user_data_count = 64;
 	uint32_t                    scratch_dwords  = 0;
 	uint64_t                    shader_hash     = 0;
-	uint64_t                    shader_base     = 0;
 	uint32_t                    push_constant_offset      = 0;
 	bool                        dump_ir                    = true;
 	bool                        early_dump                 = false;
 	const char*                 dump_label                 = nullptr;
-	const uint32_t*             user_data                  = nullptr;
+	std::span<const uint32_t>   user_data;
 	IR::SrtMemoryReader         read_memory                = nullptr;
 	IR::SrtMemoryReader         read_specialization_memory = nullptr;
 	void*                       read_memory_data           = nullptr;
-	const IR::ResourceSnapshot* resource_snapshot          = nullptr;
 	ShaderStageInputInfo        input_info;
 };
 
@@ -39,8 +36,8 @@ struct CompileResult {
 	IR::ResourceSnapshot  resources;
 };
 
-bool TryRecompile(std::span<const uint32_t> code, const CompileOptions& options,
-                  CompileResult& result, std::string* error);
+[[nodiscard]] CompileResult Recompile(std::span<const uint32_t> code,
+                                      const CompileOptions& options);
 
 } // namespace Libs::Graphics::ShaderRecompiler
 

src/graphics/shader/recompiler/backend/spirv/SpirvEmitter.cpp

@@ -1,19 +1,23 @@
 #include "graphics/shader/recompiler/backend/spirv/SpirvEmitter.h"
 
+#include "common/assert.h"
 #include "graphics/shader/recompiler/backend/spirv/spirvEmitterInternal.h"
 #include "graphics/shader/recompiler/ir/ShaderIR.h"
-#include "graphics/shader/shader.h"
 
 #include <algorithm>
 #include <array>
 
 namespace Libs::Graphics::ShaderRecompiler::Spirv {
 
-using ShaderError::Fail;
-
 namespace {
 
-bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
+[[noreturn]] void Fail(const IR::Program& program, const char* reason) {
+	EXIT("SPIR-V validation failed: hash=0x%016" PRIx64 " stage=%u reason=%s\n",
+	     program.shader_hash, static_cast<unsigned>(program.stage), reason);
+	std::abort();
+}
+
+void ValidateNativeProgram(const IR::Program& program) {
 	using Kind                                             = IR::DescriptorBindingKind;
 	constexpr auto                               KindCount = static_cast<size_t>(Kind::Count);
 	std::array<std::vector<uint32_t>, KindCount> expected;
@@ -36,13 +40,13 @@ bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
 	for (uint32_t i = 0; i < program.info.images.size(); i++) {
 		const auto kind = IR::DescriptorBindingForImage(program.info.images[i]);
 		if (!kind.has_value()) {
-			return Fail(error, "native shader plan has an invalid image class");
+			Fail(program, "native shader plan has an invalid image class");
 		}
 		present[static_cast<size_t>(*kind)] = true;
 		const auto dynamic = program.info.images[i].mip_mode == IR::ImageMipMode::DynamicStorage;
 		const auto count   = dynamic ? program.info.images[i].mip_count : 1u;
 		if (count == 0u || (!dynamic && program.info.images[i].mip_count != 1u)) {
-			return Fail(error, "native shader plan has an invalid image mip descriptor count");
+			Fail(program, "native shader plan has an invalid image mip descriptor count");
 		}
 		expected[static_cast<size_t>(*kind)].insert(expected[static_cast<size_t>(*kind)].end(),
 		                                            count, i);
@@ -58,11 +62,11 @@ bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
 			}
 			const auto index = inst.Flags<IR::MemoryFlags>().index;
 			if (index >= program.memory_info.size()) {
-				return Fail(error, "shared operation has invalid memory metadata");
+				Fail(program, "shared operation has invalid memory metadata");
 			}
 			const auto kind = program.memory_info[index].kind;
 			if (kind != IR::ResourceKind::Lds && kind != IR::ResourceKind::Gds) {
-				return Fail(error, "shared operation has invalid resource kind");
+				Fail(program, "shared operation has invalid resource kind");
 			}
 			uses_gds |= kind == IR::ResourceKind::Gds;
 		}
@@ -91,13 +95,13 @@ bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
 		const auto kind = static_cast<size_t>(binding.kind);
 		if (kind >= KindCount || seen[kind] || !present[kind] ||
 		    binding.resources != expected[kind]) {
-			return Fail(error, "native descriptor groups do not match shader topology");
+			Fail(program, "native descriptor groups do not match shader topology");
 		}
 		seen[kind] = true;
 	}
 	for (size_t i = 0; i < KindCount; i++) {
 		if (present[i] != seen[i]) {
-			return Fail(error, "native shader plan is missing a required descriptor group");
+			Fail(program, "native shader plan is missing a required descriptor group");
 		}
 	}
 	const auto has_user_storage = present[static_cast<size_t>(Kind::UserData)];
@@ -114,7 +118,7 @@ bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
 	    std::adjacent_find(program.bindings.user_data_registers.begin(),
 	                       program.bindings.user_data_registers.end()) !=
 	        program.bindings.user_data_registers.end()) {
-		return Fail(error, "native user-data layout is inconsistent");
+		Fail(program, "native user-data layout is inconsistent");
 	}
 
 	const auto planning_only_handle = [&](const IR::Inst& handle) {
@@ -139,7 +143,7 @@ bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
 						break;
 					}
 					if (dense >= program.info.buffers.size()) {
-						return Fail(error, "typed buffer handle has an invalid dense resource");
+						Fail(program, "typed buffer handle has an invalid dense resource");
 					}
 					break;
 				case IR::ValueOpcode::GetAddressResource:
@@ -147,29 +151,29 @@ bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
 						break;
 					}
 					if (inst.NumArgs() != 2 || !program.info.uses_dma) {
-						return Fail(error, "typed address handle has invalid DMA metadata");
+						Fail(program, "typed address handle has invalid DMA metadata");
 					}
 					break;
 				case IR::ValueOpcode::GetScratchResource:
 					if (inst.NumArgs() != 0 || program.scratch_dwords == 0) {
-						return Fail(error, "typed scratch handle has invalid shader metadata");
+						Fail(program, "typed scratch handle has invalid shader metadata");
 					}
 					break;
 				case IR::ValueOpcode::GetImageResource:
 					if (dense >= program.info.images.size()) {
-						return Fail(error, "typed image handle has an invalid dense resource");
+						Fail(program, "typed image handle has an invalid dense resource");
 					}
 					break;
 				case IR::ValueOpcode::GetSamplerResource:
 					if (dense >= program.info.samplers.size()) {
-						return Fail(error, "typed sampler handle has an invalid dense resource");
+						Fail(program, "typed sampler handle has an invalid dense resource");
 					}
 					break;
 				case IR::ValueOpcode::ReadConst: {
 					const auto slot = inst.Arg(1).Resolve();
 					if (!slot.IsImmediate() || slot.GetType() != IR::Type::U32 ||
 					    slot.U32() >= program.srt_reads.size()) {
-						return Fail(error, "flattened SRT read has an invalid dense slot");
+						Fail(program, "flattened SRT read has an invalid dense slot");
 					}
 					break;
 				}
@@ -177,12 +181,11 @@ bool ValidateNativeProgram(const IR::Program& program, std::string* error) {
 			}
 		}
 	}
-	return true;
 }
 
 } // namespace
 
-bool AnalyzeProgramRequirements(IR::Program& program, std::string* error) {
+void AnalyzeProgramRequirements(IR::Program& program) {
 	program.spirv_requirements.reset();
 	IR::SpirvRequirements requirements {};
 	const auto MarkBallot = [&] { requirements.subgroup_ballot = true; };
@@ -192,30 +195,30 @@ bool AnalyzeProgramRequirements(IR::Program& program, std::string* error) {
 			if (address_access != IR::AddressAccess::None) {
 				const auto memory_index = inst.Flags<IR::MemoryFlags>().index;
 				if (memory_index >= program.memory_info.size()) {
-					return Fail(error, "address operation has invalid memory metadata");
+					Fail(program, "address operation has invalid memory metadata");
 				}
 				if (program.memory_info[memory_index].kind == IR::ResourceKind::Scratch) {
 					if (program.scratch_dwords == 0) {
-						return Fail(error, "scratch operation has no per-thread storage");
+						Fail(program, "scratch operation has no per-thread storage");
 					}
 					requirements.function_scratch = true;
 				} else if (address_access == IR::AddressAccess::Write) {
-					return Fail(error, "writable FLAT/GLOBAL addresses require GPU ownership tracking");
+					Fail(program, "writable FLAT/GLOBAL addresses require GPU ownership tracking");
 				}
 			}
 			if (IR::BufferAccessOf(inst.GetOpcode()) != IR::BufferAccess::None) {
 				const auto memory_index = inst.Flags<IR::MemoryFlags>().index;
 				if (memory_index >= program.memory_info.size()) {
-					return Fail(error, "buffer operation has invalid memory metadata");
+					Fail(program, "buffer operation has invalid memory metadata");
 				}
 				const auto& memory = program.memory_info[memory_index];
 				if (memory.kind == IR::ResourceKind::Buffer) {
 					if (memory.resource >= program.info.buffers.size()) {
-						return Fail(error, "buffer operation has invalid resource metadata");
+						Fail(program, "buffer operation has invalid resource metadata");
 					}
 					if ((program.info.buffers[memory.resource].packed_stride & (1u << 20u)) != 0u) {
 						if (program.stage != ShaderType::Compute) {
-							return Fail(error, "buffer ADD_TID is only valid for compute shaders");
+							Fail(program, "buffer ADD_TID is only valid for compute shaders");
 						}
 						requirements.subgroup_local_invocation_id = true;
 					}
@@ -225,11 +228,11 @@ bool AnalyzeProgramRequirements(IR::Program& program, std::string* error) {
 			if (shared_access != IR::SharedAccess::None) {
 				const auto index = inst.Flags<IR::MemoryFlags>().index;
 				if (index >= program.memory_info.size()) {
-					return Fail(error, "shared operation has invalid memory metadata");
+					Fail(program, "shared operation has invalid memory metadata");
 				}
 				const auto kind = program.memory_info[index].kind;
 				if (kind != IR::ResourceKind::Lds && kind != IR::ResourceKind::Gds) {
-					return Fail(error, "shared operation has invalid resource kind");
+					Fail(program, "shared operation has invalid resource kind");
 				}
 				if (program.stage != ShaderType::Compute && kind == IR::ResourceKind::Lds) {
 					requirements.function_lds = true;
@@ -266,7 +269,8 @@ bool AnalyzeProgramRequirements(IR::Program& program, std::string* error) {
 					requirements.subgroup_local_invocation_id = true;
 					break;
 				}
-				case IR::ValueOpcode::SwizzleU32: {
+				case IR::ValueOpcode::SwizzleU32:
+				case IR::ValueOpcode::BpermuteU32: {
 					MarkBallot();
 					requirements.subgroup_shuffle             = true;
 					requirements.subgroup_local_invocation_id = true;
@@ -282,7 +286,7 @@ bool AnalyzeProgramRequirements(IR::Program& program, std::string* error) {
 				case IR::ValueOpcode::SetAttribute: {
 					const auto index = inst.Flags<IR::ExportFlags>().index;
 					if (index >= program.export_info.size()) {
-						return Fail(error, "attribute export has invalid metadata");
+						Fail(program, "attribute export has invalid metadata");
 					}
 					if (program.stage == ShaderType::Pixel &&
 					    program.export_info[index].vm) {
@@ -295,37 +299,30 @@ bool AnalyzeProgramRequirements(IR::Program& program, std::string* error) {
 		}
 	}
 	program.spirv_requirements.emplace(requirements);
-	return true;
 }
 
-bool EmitProgram(const IR::Program& program, const IR::ResourceSnapshot& resources,
-                 ShaderStageInputInfo input_info, std::vector<uint32_t>& spirv,
-                 std::string* error) {
+std::vector<uint32_t> EmitProgram(const IR::Program& program,
+                                  const IR::ResourceSnapshot& resources,
+                                  ShaderStageInputInfo input_info) {
 	using namespace Emitter;
 
 	if (program.stage != ShaderType::Compute && program.stage != ShaderType::Vertex &&
 	    program.stage != ShaderType::Pixel) {
-		SetError(error, "binary SPIR-V emitter supports compute, vertex, and pixel shaders");
-		return false;
+		Fail(program, "binary SPIR-V emitter supports compute, vertex, and pixel shaders");
 	}
 	if (!program.srt_plan_complete || !program.resource_tracking_complete ||
 	    !program.shader_info_complete || !program.binding_layout_complete ||
 	    !program.spirv_requirements.has_value()) {
-		SetError(error, "SPIR-V emitter requires a fully planned native shader program");
-		return false;
+		Fail(program, "SPIR-V emitter requires a fully planned native shader program");
 	}
-	if (!IR::ValidateResourceSnapshot(program, resources, error)) {
-		return false;
+	if (!IR::ValidateResourceSnapshot(program, resources)) {
+		Fail(program, "resource snapshot is inconsistent with the native shader plan");
 	}
-	if (!IR::ValidateResourceSpecialization(program, resources, error)) {
-		return false;
-	}
-	if (!ValidateNativeProgram(program, error)) {
-		return false;
-	}
-	if (!IR::ValidateProgram(program, true, error)) {
-		return false;
+	if (!IR::ValidateResourceSpecialization(program, resources)) {
+		Fail(program, "resource specialization is inconsistent with the native shader plan");
 	}
+	ValidateNativeProgram(program);
+	IR::ValidateProgram(program, true);
 	EmitterState state(program, resources, input_info);
 	state.stage     = program.stage;
 	state.wave_size = program.wave_size;
@@ -336,17 +333,13 @@ bool EmitProgram(const IR::Program& program, const IR::ResourceSnapshot& resourc
 	AllocateInputVariables(state);
 	AllocateOutputVariables(state);
 	DefineModule(state);
-	if (!EmitProgram(state, program, error)) {
-		return false;
-	}
+	EmitProgram(state, program);
 
 	auto binary = state.builder.Build();
 	if (binary.empty()) {
-		SetError(error, "SPIR-V builder returned an empty module");
-		return false;
+		Fail(program, "SPIR-V builder returned an empty module");
 	}
-	spirv = std::move(binary);
-	return true;
+	return binary;
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::Spirv

src/graphics/shader/recompiler/backend/spirv/SpirvEmitter.h

@@ -9,10 +9,11 @@
 
 namespace Libs::Graphics::ShaderRecompiler::Spirv {
 
-bool AnalyzeProgramRequirements(IR::Program& program, std::string* error);
+void AnalyzeProgramRequirements(IR::Program& program);
 
-bool EmitProgram(const IR::Program& program, const IR::ResourceSnapshot& resources,
-                 ShaderStageInputInfo input_info, std::vector<uint32_t>& spirv, std::string* error);
+std::vector<uint32_t> EmitProgram(const IR::Program& program,
+                                  const IR::ResourceSnapshot& resources,
+                                  ShaderStageInputInfo input_info);
 
 } // namespace Libs::Graphics::ShaderRecompiler::Spirv
 

src/graphics/shader/recompiler/backend/spirv/spirvEmitterAnalysis.cpp

@@ -37,12 +37,6 @@ bool PixelParameterIsCustom(const EmitterState& state, uint32_t attr) {
 	       ShaderPixelParameterIsCustom(*state.input_info.pixel, attr);
 }
 
-void SetError(std::string* error, const char* message) {
-	if (error != nullptr) {
-		*error = message;
-	}
-}
-
 bool HasOutput(const std::vector<OutputBinding>& outputs, IR::StageOutputKind kind,
                uint32_t index) {
 	return std::any_of(outputs.begin(), outputs.end(), [kind, index](const OutputBinding& binding) {

src/graphics/shader/recompiler/backend/spirv/spirvEmitterFlow.cpp

@@ -408,6 +408,41 @@ void EmitAuxPositionExport(ValueEmitContext& ctx, uint32_t data, const IR::Expor
 	}
 }
 
+uint32_t ConvertClipCoordinate(EmitterState& state, uint32_t coordinate, float scale,
+                               float offset, float half_extent) {
+	const auto window  = state.builder.AllocateId();
+	const auto biased  = state.builder.AllocateId();
+	const auto divided = state.builder.AllocateId();
+	const auto ndc     = state.builder.AllocateId();
+	state.builder.AddFunction(
+	    {OpFMul, TypeF32(state), window, coordinate, ConstantF32Value(state, scale)});
+	state.builder.AddFunction(
+	    {OpFAdd, TypeF32(state), biased, window, ConstantF32Value(state, offset)});
+	state.builder.AddFunction(
+	    {OpFDiv, TypeF32(state), divided, biased, ConstantF32Value(state, half_extent)});
+	state.builder.AddFunction(
+	    {OpFSub, TypeF32(state), ndc, divided, ConstantF32Value(state, 1.0f)});
+	return ndc;
+}
+
+uint32_t ConvertPositionToClipSpace(EmitterState& state, uint32_t position) {
+	const auto& transform = state.input_info.vertex->clip_space;
+	uint32_t    components[4] {};
+	for (uint32_t i = 0; i < 4; i++) {
+		components[i] = state.builder.AllocateId();
+		state.builder.AddFunction(
+		    {OpCompositeExtract, TypeF32(state), components[i], position, i});
+	}
+	components[0] = ConvertClipCoordinate(state, components[0], transform.scale[0],
+	                                      transform.offset[0], transform.half_extent[0]);
+	components[1] = ConvertClipCoordinate(state, components[1], transform.scale[1],
+	                                      transform.offset[1], transform.half_extent[1]);
+	const auto converted = state.builder.AllocateId();
+	state.builder.AddFunction({OpCompositeConstruct, TypeF32Vector(state, 4), converted,
+	                           components[0], components[1], components[2], components[3]});
+	return converted;
+}
+
 void EmitExport(ValueEmitContext& ctx, const IR::Inst& inst) {
 	auto&       state = ctx.state;
 	const auto& exp   = ctx.Export(inst);
@@ -467,6 +502,9 @@ void EmitExport(ValueEmitContext& ctx, const IR::Inst& inst) {
 			}
 		}
 		if (exp.kind == IR::ExportTargetKind::Position) {
+			if (state.stage == ShaderType::Vertex && state.input_info.vertex->clip_space.enabled) {
+				value = ConvertPositionToClipSpace(state, value);
+			}
 			const auto pointer = state.builder.AllocateId();
 			state.builder.AddFunction(
 			    {OpAccessChain, TypePointer(state, StorageClassOutput, TypeF32Vector(state, 4)),

src/graphics/shader/recompiler/backend/spirv/spirvEmitterHelpers.cpp

@@ -110,6 +110,12 @@ DppTargetLane EmitDppTargetLane(EmitterState& state, uint32_t control) {
 	if (control == 0x141u) {
 		return EmitDppMirrorTargetLane(state, subid, true);
 	}
+	if (control >= 0x160u && control <= 0x16fu) {
+		const auto target = state.builder.AllocateId();
+		state.builder.AddFunction(
+		    {OpBitwiseXor, TypeU32(state), target, subid, ConstantU32(state, control & 0xfu)});
+		return {target, EmitTrueBool(state)};
+	}
 	return {subid, EmitTrueBool(state)};
 }
 

src/graphics/shader/recompiler/backend/spirv/spirvEmitterImage.cpp

@@ -439,6 +439,49 @@ uint32_t UnpackImageGather(ValueEmitContext& ctx, const IR::MemoryInfo& mem, uin
 	return result;
 }
 
+uint32_t EmitOneDimensionalGatherLz(ValueEmitContext& ctx, const IR::MemoryInfo& mem, uint32_t pc,
+	                                 uint32_t coord, bool integer) {
+	auto& state = ctx.state;
+	state.builder.RequireCapability(CapabilityImageQuery);
+	const auto image = LoadSampledImageDescriptor(state, mem, pc, ImageViewKind::Dim1D);
+	const auto width = state.builder.AllocateId();
+	state.builder.AddFunction(
+	    {OpImageQuerySizeLod, TypeU32(state), width, image, ConstantU32(state, 0)});
+	const auto width_f32 = state.builder.AllocateId();
+	state.builder.AddFunction({OpConvertUToF, TypeF32(state), width_f32, width});
+	const auto left = state.builder.AllocateId();
+	state.builder.AddFunction(
+	    {OpExtInst, TypeF32(state), left, GlslStd450(state), GlslFloor,
+	     Binary(state, OpFSub, TypeF32(state),
+	            Binary(state, OpFMul, TypeF32(state), coord, width_f32),
+	            ConstantF32(state, 0x3f000000u))});
+
+	const auto sampled = MakeSampledImage(state, mem, pc, ImageViewKind::Dim1D);
+	const auto vector_type = integer ? TypeU32Vector(state, 4) : TypeF32Vector(state, 4);
+	const auto scalar_type = integer ? TypeU32(state) : TypeF32(state);
+	const auto component = ImageConversionFormat(state, mem).format == Prospero::BufferFormat::kInvalid
+	                           ? ImageGatherComponent(mem.dmask)
+	                           : 0u;
+	uint32_t values[2] {};
+	for (uint32_t index = 0; index < 2u; index++) {
+		const auto sample_coord = Binary(
+		    state, OpFDiv, TypeF32(state),
+		    Binary(state, OpFAdd, TypeF32(state), left,
+		           ConstantF32(state, index == 0u ? 0x3f000000u : 0x3fc00000u)),
+		    width_f32);
+		const auto texel = state.builder.AllocateId();
+		state.builder.AddFunction({OpImageSampleExplicitLod, vector_type, texel, sampled, sample_coord,
+		                           ImageOperandsLodMask, ZeroF32(state)});
+		values[index] = state.builder.AllocateId();
+		state.builder.AddFunction(
+		    {OpCompositeExtract, scalar_type, values[index], texel, component});
+	}
+	const auto result = state.builder.AllocateId();
+	state.builder.AddFunction({OpCompositeConstruct, vector_type, result, values[0], values[1],
+	                           values[1], values[0]});
+	return result;
+}
+
 uint32_t PackImageTexel(ValueEmitContext& ctx, const IR::MemoryInfo& mem, uint32_t texel) {
 	const auto info = ImageConversionFormat(ctx.state, mem);
 	if (info.format == Prospero::BufferFormat::kInvalid) return texel;
@@ -613,6 +656,22 @@ bool EmitValueImage(ValueEmitContext& ctx, const IR::Inst& inst) {
 		const auto coord =
 		    CoordF32(ctx, mem, *address, layout.coord, ImageViewCoordinateComponents(view));
 		if (op == IR::ValueOpcode::ImageGatherRaw) {
+			if (view == ImageViewKind::Dim1D) {
+				if (dref || !HasFlag(mem, Decoder::ImageSampleFlagLevelZero) ||
+				    HasFlag(mem, Decoder::ImageSampleFlagOffset) ||
+				    HasFlag(mem, Decoder::ImageSampleFlagGatherHorizontal)) {
+					ctx.Fail(inst, "has an unsupported 1D gather variant");
+					return true;
+				}
+				const auto sample = EmitOneDimensionalGatherLz(ctx, mem, pc, coord, integer);
+				ctx.Define(inst, ResultVector(ctx, UnpackImageGather(ctx, mem, sample), integer,
+				                              false, mem, true));
+				return true;
+			}
+			if (view == ImageViewKind::Dim1DArray) {
+				ctx.Fail(inst, "has an unsupported 1D-array gather");
+				return true;
+			}
 			const auto            sampled = MakeSampledImage(state, mem, pc, view);
 			const auto            sample  = state.builder.AllocateId();
 			std::vector<uint32_t> words =

src/graphics/shader/recompiler/backend/spirv/spirvEmitterInternal.h

@@ -439,7 +439,8 @@ struct ValueEmitContext {
 	const IR::MemoryInfo& Memory(const IR::Inst& inst) const;
 	const IR::ExportInfo& Export(const IR::Inst& inst) const;
 	uint32_t              Label(const IR::Block* block) const;
-	void                  Fail(const IR::Inst& inst, const char* reason);
+	[[noreturn]] void     Fail(const char* reason) const;
+	[[noreturn]] void     Fail(const IR::Inst& inst, const char* reason) const;
 
 	EmitterState&                                                      state;
 	const IR::Program&                                            program;
@@ -449,8 +450,6 @@ struct ValueEmitContext {
 	std::unordered_map<const IR::Inst*, std::pair<uint32_t, uint32_t>> dispatcher_block_loads;
 	const IR::Block*                                                   current_block = nullptr;
 	uint32_t                                                           scratch_u32_variable = 0;
-	bool                                                               failed               = false;
-	std::string                                                        error;
 };
 
 enum class VertexInputScalarKind { Float, Sint, Uint };
@@ -613,8 +612,6 @@ uint32_t VertexParameterVectorOrScalarType(EmitterState& state, VertexInputScala
 uint32_t VertexParameterInputPointerType(EmitterState& state, VertexInputScalarKind kind,
                                          uint32_t components);
 
-void SetError(std::string* error, const char* message);
-
 bool HasOutput(const std::vector<OutputBinding>& outputs, IR::StageOutputKind kind, uint32_t index);
 
 void CopyProgramInputsAndOutputs(EmitterState& state, const IR::Program& program);
@@ -864,7 +861,7 @@ bool EmitValueMemory(ValueEmitContext& ctx, const IR::Inst& inst);
 
 bool EmitValueImage(ValueEmitContext& ctx, const IR::Inst& inst);
 
-bool EmitProgram(EmitterState& state, const IR::Program& program, std::string* error);
+void EmitProgram(EmitterState& state, const IR::Program& program);
 
 void DefineGetBdaPointer(EmitterState& state);
 

src/graphics/shader/recompiler/backend/spirv/spirvEmitterMemory.cpp

@@ -30,6 +30,19 @@ uint32_t AndCondition(EmitterState& state, uint32_t lhs, uint32_t rhs) {
 	return Binary(state, OpLogicalAnd, TypeBool(state), lhs, rhs);
 }
 
+uint32_t EmitDsMaskedLaneRead(EmitterState& state, uint32_t source, uint32_t target,
+                              uint32_t exec) {
+	const auto shuffled = state.builder.AllocateId();
+	state.builder.AddFunction({OpGroupNonUniformShuffle, TypeU32(state), shuffled,
+	                           ConstantU32(state, ScopeSubgroup), source, target});
+	const auto source_exec = state.builder.AllocateId();
+	state.builder.AddFunction({OpGroupNonUniformShuffle, TypeBool(state), source_exec,
+	                           ConstantU32(state, ScopeSubgroup), exec, target});
+	const auto source_active =
+	    AndCondition(state, source_exec, EmitSubgroupLaneActiveBool(state, target));
+	return Select(state, TypeU32(state), source_active, shuffled, ConstantU32(state, 0));
+}
+
 uint32_t BufferByteAddress(ValueEmitContext& ctx, const IR::Inst& inst, const IR::MemoryInfo& mem,
                            uint32_t index, uint32_t offset, uint32_t soffset) {
 	auto&          state   = ctx.state;
@@ -1181,24 +1194,27 @@ bool EmitValueMemory(ValueEmitContext& ctx, const IR::Inst& inst) {
 		ctx.Define(inst, AppendConsume(ctx, inst, op == IR::ValueOpcode::DataAppend));
 		return true;
 	}
-	if (op == IR::ValueOpcode::SwizzleU32) {
-		ctx.state.builder.AddFunction({OpStore, ctx.scratch_u32_variable, ctx.Arg(inst, 0)});
-		const auto source = ctx.state.builder.AllocateId();
-		ctx.state.builder.AddFunction(
-		    {OpLoad, TypeU32(ctx.state), source, ctx.scratch_u32_variable});
-		const auto target =
-		    EmitDsSwizzleTargetLane(state, EmitSubgroupLocalInvocationId(state),
-		                            inst.Arg(1).IsImmediate() ? inst.Arg(1).U32() : 0);
-		const auto shuffled = state.builder.AllocateId();
-		state.builder.AddFunction({OpGroupNonUniformShuffle, TypeU32(state), shuffled,
-		                           ConstantU32(state, ScopeSubgroup), source, target});
-		const auto source_exec = state.builder.AllocateId();
-		state.builder.AddFunction({OpGroupNonUniformShuffle, TypeBool(state), source_exec,
-		                           ConstantU32(state, ScopeSubgroup), ctx.Arg(inst, 2), target});
-		const auto source_active =
-		    AndCondition(state, source_exec, EmitSubgroupLaneActiveBool(state, target));
-		ctx.Define(inst,
-		           Select(state, TypeU32(state), source_active, shuffled, ConstantU32(state, 0)));
+	if (op == IR::ValueOpcode::SwizzleU32 || op == IR::ValueOpcode::BpermuteU32) {
+		uint32_t source = ctx.Arg(inst, 0);
+		uint32_t target = 0;
+		if (op == IR::ValueOpcode::SwizzleU32) {
+			state.builder.AddFunction({OpStore, ctx.scratch_u32_variable, source});
+			source = state.builder.AllocateId();
+			state.builder.AddFunction({OpLoad, TypeU32(state), source, ctx.scratch_u32_variable});
+			target = EmitDsSwizzleTargetLane(state, EmitSubgroupLocalInvocationId(state),
+			                                 inst.Arg(1).IsImmediate() ? inst.Arg(1).U32() : 0);
+		} else {
+			const auto index = Binary(
+			    state, OpBitwiseAnd, TypeU32(state),
+			    Binary(state, OpShiftRightLogical, TypeU32(state), ctx.Arg(inst, 1),
+			           ConstantU32(state, 2)),
+			    ConstantU32(state, 31));
+			const auto base =
+			    Binary(state, OpBitwiseAnd, TypeU32(state), EmitSubgroupLocalInvocationId(state),
+			           ConstantU32(state, ~31u));
+			target = Binary(state, OpBitwiseOr, TypeU32(state), base, index);
+		}
+		ctx.Define(inst, EmitDsMaskedLaneRead(state, source, target, ctx.Arg(inst, 2)));
 		return true;
 	}
 	return false;

src/graphics/shader/recompiler/backend/spirv/spirvEmitterProgram.cpp

@@ -1,8 +1,9 @@
 #include "graphics/shader/recompiler/backend/spirv/spirvEmitterInternal.h"
 
+#include "common/assert.h"
+
 #include <algorithm>
 #include <bit>
-#include <fmt/format.h>
 #include <functional>
 #include <optional>
 #include <unordered_set>
@@ -196,57 +197,50 @@ uint32_t EmitDispatcherNextPc(ValueEmitContext& ctx, const DispatcherFunctionSta
 	}
 }
 
-bool EmitDirectInstruction(ValueEmitContext& ctx, const IR::Inst& inst) {
+void EmitDirectInstruction(ValueEmitContext& ctx, const IR::Inst& inst) {
 	if (EmitValueFlow(ctx, inst) || EmitValueAlu(ctx, inst) || EmitValueMemory(ctx, inst) ||
 	    EmitValueImage(ctx, inst)) {
-		return true;
+		return;
 	}
 	ctx.Fail(inst, "has no direct SPIR-V emitter");
-	return false;
 }
 
-bool EmitStructuredInstruction(ValueEmitContext& ctx, StructuredFunctionState& structured,
+void EmitStructuredInstruction(ValueEmitContext& ctx, StructuredFunctionState& structured,
                                const IR::Inst& inst) {
 	if (inst.GetOpcode() == IR::ValueOpcode::Phi) {
 		const auto type = ctx.TypeId(inst.GetType());
 		if (type == 0 || inst.NumArgs() == 0) {
 			ctx.Fail(inst, "has no native SPIR-V representation");
-			return false;
 		}
 		for (size_t index = 0; index < inst.NumArgs(); index++) {
 			const auto* predecessor = inst.PhiBlock(index);
 			if (predecessor == nullptr || !ctx.labels.contains(predecessor)) {
 				ctx.Fail(inst, "has a predecessor outside the structured function");
-				return false;
 			}
 		}
 		structured.deferred_phis.push_back(
 		    {ctx.state.builder.AddDeferredPhi(type, ctx.Result(inst), inst.NumArgs()), &inst});
-		return true;
+		return;
 	}
-	return EmitDirectInstruction(ctx, inst);
+	EmitDirectInstruction(ctx, inst);
 }
 
-bool EmitDispatcherInstruction(ValueEmitContext& ctx, const DispatcherFunctionState& dispatcher,
+void EmitDispatcherInstruction(ValueEmitContext& ctx, const DispatcherFunctionState& dispatcher,
                                const IR::Inst& inst) {
 	if (inst.GetOpcode() == IR::ValueOpcode::Phi) {
 		const auto type = ctx.TypeId(inst.GetType());
 		if (type == 0) {
 			ctx.Fail(inst, "cannot be loaded by the dispatcher");
-			return false;
 		}
 		ctx.state.builder.AddFunction(
 		    {OpLoad, type, ctx.Result(inst), dispatcher.spills.at(&inst)});
-		return true;
-	}
-	if (!EmitDirectInstruction(ctx, inst)) {
-		return false;
+		return;
 	}
+	EmitDirectInstruction(ctx, inst);
 	if (const auto found = dispatcher.spills.find(&inst); found != dispatcher.spills.end()) {
 		ctx.state.builder.AddFunction(
 		    {OpStore, found->second, ctx.Def(IR::Value(const_cast<IR::Inst*>(&inst)))});
 	}
-	return true;
 }
 
 template <typename EmitInstruction>
@@ -258,14 +252,11 @@ void EmitBlock(ValueEmitContext& ctx, const IR::Block* block, EmitInstruction&&
 		if (inst.GetOpcode() == IR::ValueOpcode::Phi) {
 			if (emitted_non_phi) {
 				ctx.Fail(inst, "appears after a non-Phi instruction");
-				return;
 			}
 		} else {
 			emitted_non_phi = true;
 		}
-		if (!emit_instruction(inst)) {
-			return;
-		}
+		emit_instruction(inst);
 	}
 }
 
@@ -275,15 +266,10 @@ void PatchStructuredPhis(ValueEmitContext& ctx, StructuredFunctionState& structu
 			const auto* predecessor = deferred.instruction->PhiBlock(index);
 			const auto  found       = structured.block_exit_labels.find(predecessor);
 			if (found == structured.block_exit_labels.end()) {
-				ctx.failed = true;
-				ctx.error  = "typed Phi predecessor was not emitted";
-				return;
+				ctx.Fail(*deferred.instruction, "has a predecessor that was not emitted");
 			}
 			ctx.state.builder.PatchDeferredPhi(
 			    deferred.phi, index, ctx.Def(deferred.instruction->Arg(index)), found->second);
-			if (ctx.failed) {
-				return;
-			}
 		}
 	}
 }
@@ -291,20 +277,15 @@ void PatchStructuredPhis(ValueEmitContext& ctx, StructuredFunctionState& structu
 void EmitStructuredFunction(ValueEmitContext& ctx) {
 	StructuredFunctionState structured;
 	ctx.state.builder.AddFunction({OpBranch, ctx.Label(ctx.program.blocks.front())});
-	for (size_t index = 0; index < ctx.program.blocks.size() && !ctx.failed; index++) {
+	for (size_t index = 0; index < ctx.program.blocks.size(); index++) {
 		const auto* block = ctx.program.blocks[index];
 		EmitBlock(ctx, block, [&](const IR::Inst& inst) {
-			return EmitStructuredInstruction(ctx, structured, inst);
+			EmitStructuredInstruction(ctx, structured, inst);
 		});
-		if (ctx.failed) {
-			break;
-		}
 		structured.block_exit_labels.emplace(block, ctx.state.current_label);
 		EmitStructuredTerminator(ctx, block, ctx.program.block_info[index]);
 	}
-	if (!ctx.failed) {
-		PatchStructuredPhis(ctx, structured);
-	}
+	PatchStructuredPhis(ctx, structured);
 }
 
 void EmitDispatcherFunction(ValueEmitContext& ctx, const DispatcherFunctionState& dispatcher) {
@@ -312,11 +293,8 @@ void EmitDispatcherFunction(ValueEmitContext& ctx, const DispatcherFunctionState
 	const auto* entry = ctx.program.blocks.front();
 	state.builder.AddFunction({OpBranch, ctx.Label(entry)});
 	EmitBlock(ctx, entry, [&](const IR::Inst& inst) {
-		return EmitDispatcherInstruction(ctx, dispatcher, inst);
+		EmitDispatcherInstruction(ctx, dispatcher, inst);
 	});
-	if (ctx.failed) {
-		return;
-	}
 	const auto initial_pc =
 	    EmitDispatcherNextPc(ctx, dispatcher, entry, ctx.program.block_info.front());
 	const auto initial_parent = state.current_label;
@@ -347,22 +325,16 @@ void EmitDispatcherFunction(ValueEmitContext& ctx, const DispatcherFunctionState
 	std::vector<uint32_t> next_pc_words {OpPhi, TypeU32(state), next_pc,
 	                                     ConstantU32(state, UINT32_MAX), dispatcher.select_label};
 
-	for (size_t index = 1; index < ctx.program.blocks.size() && !ctx.failed; index++) {
+	for (size_t index = 1; index < ctx.program.blocks.size(); index++) {
 		EmitBlock(ctx, ctx.program.blocks[index], [&](const IR::Inst& inst) {
-			return EmitDispatcherInstruction(ctx, dispatcher, inst);
+			EmitDispatcherInstruction(ctx, dispatcher, inst);
 		});
-		if (ctx.failed) {
-			break;
-		}
 		const auto selected = EmitDispatcherNextPc(ctx, dispatcher, ctx.program.blocks[index],
 		                                           ctx.program.block_info[index]);
 		next_pc_words.push_back(selected);
 		next_pc_words.push_back(state.current_label);
 		state.builder.AddFunction({OpBranch, dispatcher.after_switch_label});
 	}
-	if (ctx.failed) {
-		return;
-	}
 	EmitLabel(state, dispatcher.after_switch_label);
 	state.builder.AddFunction(next_pc_words);
 	state.builder.AddFunction({OpBranch, dispatcher.continue_label});
@@ -408,9 +380,7 @@ uint32_t ValueEmitContext::Def(IR::Value value) {
 	}
 	const auto* inst = value.ResolveInstruction();
 	if (inst == nullptr) {
-		failed = true;
-		error  = "direct SPIR-V emitter received a non-value argument";
-		return ConstantU32(state, 0);
+		Fail("direct SPIR-V emitter received a non-value argument");
 	}
 	if (dispatcher_spills != nullptr && current_block != nullptr &&
 	    inst->Parent() != current_block) {
@@ -469,9 +439,7 @@ uint32_t ValueEmitContext::Define(const IR::Inst& inst, uint32_t value) {
 uint32_t ValueEmitContext::ResourceIndex(IR::Value value, IR::ValueOpcode opcode) {
 	const auto* inst = value.ResolveInstruction();
 	if (inst == nullptr || inst->GetOpcode() != opcode) {
-		failed = true;
-		error  = "typed resource handle has the wrong producer";
-		return 0;
+		Fail("typed resource handle has the wrong producer");
 	}
 	return inst->Flags<uint32_t>();
 }
@@ -479,9 +447,7 @@ uint32_t ValueEmitContext::ResourceIndex(IR::Value value, IR::ValueOpcode opcode
 const IR::Inst* ValueEmitContext::ImageAddress(IR::Value value) {
 	const auto* inst = value.ResolveInstruction();
 	if (inst == nullptr || inst->GetOpcode() != IR::ValueOpcode::MakeImageAddress) {
-		failed = true;
-		error  = "typed image address was not constructed by MakeImageAddress";
-		return nullptr;
+		Fail("typed image address was not constructed by MakeImageAddress");
 	}
 	return inst;
 }
@@ -498,12 +464,20 @@ uint32_t ValueEmitContext::Label(const IR::Block* block) const {
 	return labels.at(block);
 }
 
-void ValueEmitContext::Fail(const IR::Inst& inst, const char* reason) {
-	failed = true;
-	error  = fmt::format("typed opcode {} {}", IR::ValueOpcodeName(inst.GetOpcode()), reason);
+[[noreturn]] void ValueEmitContext::Fail(const char* reason) const {
+	EXIT("SPIR-V emission failed: hash=0x%016" PRIx64 " stage=%u reason=%s\n",
+	     program.shader_hash, static_cast<unsigned>(program.stage), reason);
+	std::abort();
 }
 
-bool EmitProgram(EmitterState& state, const IR::Program& program, std::string* error) {
+[[noreturn]] void ValueEmitContext::Fail(const IR::Inst& inst, const char* reason) const {
+	EXIT("SPIR-V emission failed: hash=0x%016" PRIx64 " stage=%u opcode=%s reason=%s\n",
+	     program.shader_hash, static_cast<unsigned>(program.stage),
+	     IR::ValueOpcodeName(inst.GetOpcode()), reason);
+	std::abort();
+}
+
+void EmitProgram(EmitterState& state, const IR::Program& program) {
 	ValueEmitContext                       ctx(state, program);
 	std::optional<DispatcherFunctionState> dispatcher;
 	if (state.stage == ShaderType::Pixel && state.requirements.pixel_valid_mask) {
@@ -562,10 +536,6 @@ bool EmitProgram(EmitterState& state, const IR::Program& program, std::string* e
 		dispatch.merge_label        = state.builder.AllocateId();
 		ctx.dispatcher_spills       = &dispatch.spills;
 	}
-	if (ctx.failed) {
-		SetError(error, ctx.error.c_str());
-		return false;
-	}
 	DefineGetBdaPointer(state);
 	for (const auto* block: program.blocks) {
 		if (std::ranges::any_of(*block, [](const IR::Inst& inst) {
@@ -630,11 +600,6 @@ bool EmitProgram(EmitterState& state, const IR::Program& program, std::string* e
 		EmitStructuredFunction(ctx);
 	}
 	state.builder.AddFunction({OpFunctionEnd});
-	if (ctx.failed) {
-		SetError(error, ctx.error.c_str());
-		return false;
-	}
-	return true;
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::Spirv::Emitter

src/graphics/shader/recompiler/frontend/cfg/ShaderCFG.cpp

@@ -1,5 +1,7 @@
 #include "graphics/shader/recompiler/frontend/cfg/ShaderCFG.h"
 
+#include "common/assert.h"
+
 #include <algorithm>
 #include <fmt/format.h>
 #include <iterator>
@@ -24,19 +26,18 @@ struct SetpcTargetInfo {
 	std::vector<uint32_t> selector_target_pcs;
 };
 
-void SetError(std::string* error, const std::string& message) {
-	if (error != nullptr) {
-		*error = message;
-	}
-}
-
-void SetFailure(Graph& graph, FailureKind kind, uint32_t block_id, const std::string& message,
-                std::string* error) {
+void SetFailure(Graph& graph, FailureKind kind, uint32_t block_id, const std::string& message) {
 	graph.unsupported        = true;
 	graph.failure_kind       = kind;
 	graph.failure_block      = block_id;
 	graph.unsupported_reason = message;
-	SetError(error, message);
+}
+
+[[noreturn]] void ExitBuildFailure(Graph& graph, FailureKind kind, uint32_t block_id,
+                                   const std::string& message) {
+	SetFailure(graph, kind, block_id, message);
+	EXIT("shader CFG build failed: %s", message.c_str());
+	std::abort();
 }
 
 uint32_t InstructionEndPc(const Instruction& inst) {
@@ -1267,7 +1268,7 @@ bool MergeLeavesContainingLoop(const Graph& graph, uint32_t header, uint32_t mer
 	return false;
 }
 
-bool CanonicalizeNaturalLoops(Graph& graph, std::string* error) {
+bool CanonicalizeNaturalLoops(Graph& graph) {
 	const auto rewrite_budget = graph.blocks.size() * 2u + 16u;
 	for (size_t rewrite = 0; rewrite < rewrite_budget; rewrite++) {
 		bool changed = false;
@@ -1322,19 +1323,18 @@ bool CanonicalizeNaturalLoops(Graph& graph, std::string* error) {
 	}
 
 	SetFailure(graph, FailureKind::StructuredControlFlow, graph.entry_block,
-	           "CFG loop canonicalization exceeded rewrite budget", error);
+	           "CFG loop canonicalization exceeded rewrite budget");
 	return false;
 }
 
-bool IsolateSemanticLoopHeaders(Graph& graph, std::string* error) {
+bool IsolateSemanticLoopHeaders(Graph& graph) {
 	const auto isolation_budget = graph.natural_loops.size() + 1u;
 	for (size_t isolation = 0; isolation < isolation_budget; isolation++) {
-		const auto loop = std::find_if(graph.natural_loops.begin(), graph.natural_loops.end(),
-		                               [&](const auto& value) {
-			                               const auto* header = graph.FindBlock(value.header);
-			                               return header != nullptr &&
-			                                      header->inst_begin != header->inst_end;
-		                               });
+		const auto loop = std::find_if(
+		    graph.natural_loops.begin(), graph.natural_loops.end(), [&](const auto& value) {
+			    const auto* header = graph.FindBlock(value.header);
+			    return header != nullptr && header->inst_begin != header->inst_end;
+		    });
 		if (loop == graph.natural_loops.end()) {
 			return true;
 		}
@@ -1346,14 +1346,13 @@ bool IsolateSemanticLoopHeaders(Graph& graph, std::string* error) {
 		// displacing OpLoopMerge into a helper-created block.
 		if (!IsolateSemanticLoopHeader(graph, loop->header)) {
 			SetFailure(graph, FailureKind::StructuredControlFlow, loop->header,
-			           fmt::format("failed to isolate semantic loop header {}", loop->header),
-			           error);
+			           fmt::format("failed to isolate semantic loop header {}", loop->header));
 			return false;
 		}
 	}
 
 	SetFailure(graph, FailureKind::StructuredControlFlow, graph.entry_block,
-	           "CFG semantic loop-header isolation exceeded rewrite budget", error);
+	           "CFG semantic loop-header isolation exceeded rewrite budget");
 	return false;
 }
 
@@ -1453,7 +1452,7 @@ std::vector<uint32_t> SelectionRegion(const Graph& graph, const BasicBlock& head
 	return region;
 }
 
-bool SplitOneSelectionMerge(Graph& graph, std::string* error) {
+bool SplitOneSelectionMerge(Graph& graph) {
 	std::vector<uint32_t> loop_headers;
 	loop_headers.reserve(graph.natural_loops.size());
 	for (const auto& loop: graph.natural_loops) {
@@ -1501,8 +1500,7 @@ bool SplitOneSelectionMerge(Graph& graph, std::string* error) {
 			    graph, FailureKind::StructuredControlFlow, block_id,
 			    fmt::format("selection header block {} has externally entered region block {}; "
 			                "semantic block cloning is disabled",
-			                block_id, *external),
-			    error);
+			                block_id, *external));
 			return false;
 		}
 		const auto construct_blocks = DominatedBlocks(graph, block_id, merge);
@@ -1514,11 +1512,11 @@ bool SplitOneSelectionMerge(Graph& graph, std::string* error) {
 	return false;
 }
 
-bool SplitSharedMergeBlocks(Graph& graph, std::string* error) {
+bool SplitSharedMergeBlocks(Graph& graph) {
 	const auto original_block_count = static_cast<uint32_t>(graph.blocks.size());
 	const auto split_budget         = std::max<uint32_t>(16u, original_block_count * 4u);
 	for (uint32_t splits = 0; splits < split_budget; splits++) {
-		if (!SplitOneLoopMerge(graph) && !SplitOneSelectionMerge(graph, error)) {
+		if (!SplitOneLoopMerge(graph) && !SplitOneSelectionMerge(graph)) {
 			return !graph.unsupported;
 		}
 		RebuildPredecessors(graph);
@@ -1528,8 +1526,7 @@ bool SplitSharedMergeBlocks(Graph& graph, std::string* error) {
 	           fmt::format("CFG shared merge splitting exceeded budget: original_blocks={} "
 	                       "current_blocks={} split_budget={}",
 	                       original_block_count, static_cast<uint64_t>(graph.blocks.size()),
-	                       split_budget),
-	           error);
+	                       split_budget));
 	return false;
 }
 
@@ -1890,12 +1887,11 @@ bool RouteSharedSelectionArm(Graph& graph, uint32_t route_variable) {
 
 } // namespace
 
-bool BuildGraph(const Decoder::Program& program, Graph& graph, std::string* error) {
-	graph = {};
-
+Graph BuildGraph(const Decoder::Program& program) {
+	Graph graph;
 	if (program.instructions.empty()) {
-		SetError(error, "cannot build CFG for empty shader");
-		return false;
+		ExitBuildFailure(graph, FailureKind::InvalidLabel, UINT32_MAX,
+		                 "cannot build CFG for empty shader");
 	}
 
 	const auto first_pc = program.instructions.front().pc;
@@ -1905,11 +1901,10 @@ bool BuildGraph(const Decoder::Program& program, Graph& graph, std::string* erro
 	for (const auto& inst: program.instructions) {
 		instruction_pcs.insert(inst.pc);
 		if (inst.opcode == Opcode::UNSUPPORTED) {
-			SetFailure(graph, FailureKind::UnsupportedInstruction, UINT32_MAX,
-			           fmt::format("unsupported decoded instruction in CFG at pc 0x{:08x}: {}",
-			                       inst.pc, Decoder::InstructionToString(inst).c_str()),
-			           error);
-			return false;
+			ExitBuildFailure(
+			    graph, FailureKind::UnsupportedInstruction, UINT32_MAX,
+			    fmt::format("unsupported decoded instruction in CFG at pc 0x{:08x}: {}", inst.pc,
+			                Decoder::InstructionToString(inst).c_str()));
 		}
 	}
 
@@ -1923,11 +1918,9 @@ bool BuildGraph(const Decoder::Program& program, Graph& graph, std::string* erro
 		const auto  next_pc = InstructionEndPc(inst);
 		if (IsBranch(inst.opcode)) {
 			if (!IsValidTarget(inst.branch_target, instruction_pcs, first_pc, end_pc)) {
-				SetFailure(graph, FailureKind::InvalidBranchTarget, UINT32_MAX,
-				           fmt::format("branch at pc 0x{:08x} targets invalid pc 0x{:08x}", inst.pc,
-				                       inst.branch_target),
-				           error);
-				return false;
+				ExitBuildFailure(graph, FailureKind::InvalidBranchTarget, UINT32_MAX,
+				                 fmt::format("branch at pc 0x{:08x} targets invalid pc 0x{:08x}",
+				                             inst.pc, inst.branch_target));
 			}
 			labels.insert(inst.branch_target);
 			if (next_pc <= end_pc) {
@@ -1936,21 +1929,19 @@ bool BuildGraph(const Decoder::Program& program, Graph& graph, std::string* erro
 		} else if (inst.opcode == Opcode::S_SETPC_B64) {
 			SetpcTargetInfo target_info;
 			if (!ResolveSetpcTargets(program, i, target_info)) {
-				SetFailure(graph, FailureKind::InvalidBranchTarget, UINT32_MAX,
-				           fmt::format("unsupported dynamic S_SETPC_B64 at pc 0x{:08x}", inst.pc),
-				           error);
-				return false;
+				ExitBuildFailure(
+				    graph, FailureKind::InvalidBranchTarget, UINT32_MAX,
+				    fmt::format("unsupported dynamic S_SETPC_B64 at pc 0x{:08x}", inst.pc));
 			}
 			const auto& target_pcs = target_info.indirect
 			                             ? target_info.target_pcs
 			                             : std::vector<uint32_t> {target_info.target};
 			for (const auto target: target_pcs) {
 				if (!IsValidTarget(target, instruction_pcs, first_pc, end_pc)) {
-					SetFailure(graph, FailureKind::InvalidBranchTarget, UINT32_MAX,
-					           fmt::format("S_SETPC_B64 at pc 0x{:08x} targets invalid pc 0x{:08x}",
-					                       inst.pc, target),
-					           error);
-					return false;
+					ExitBuildFailure(
+					    graph, FailureKind::InvalidBranchTarget, UINT32_MAX,
+					    fmt::format("S_SETPC_B64 at pc 0x{:08x} targets invalid pc 0x{:08x}",
+					                inst.pc, target));
 				}
 				labels.insert(target);
 			}
@@ -1971,10 +1962,9 @@ bool BuildGraph(const Decoder::Program& program, Graph& graph, std::string* erro
 			continue;
 		}
 		if (start != end_pc && !instruction_pcs.contains(start)) {
-			SetFailure(graph, FailureKind::InvalidLabel, UINT32_MAX,
-			           fmt::format("CFG label does not start on an instruction: 0x{:08x}", start),
-			           error);
-			return false;
+			ExitBuildFailure(
+			    graph, FailureKind::InvalidLabel, UINT32_MAX,
+			    fmt::format("CFG label does not start on an instruction: 0x{:08x}", start));
 		}
 
 		BasicBlock block;
@@ -2046,11 +2036,10 @@ bool BuildGraph(const Decoder::Program& program, Graph& graph, std::string* erro
 			block.terminator.true_block = pc_to_block.at(last.branch_target);
 			const auto fallthrough      = pc_to_block.find(next_pc);
 			if (fallthrough == pc_to_block.end()) {
-				SetFailure(graph, FailureKind::MissingFallthrough, block.id,
-				           fmt::format("conditional branch at pc 0x{:08x} has no fallthrough block",
-				                       last.pc),
-				           error);
-				return false;
+				ExitBuildFailure(
+				    graph, FailureKind::MissingFallthrough, block.id,
+				    fmt::format("conditional branch at pc 0x{:08x} has no fallthrough block",
+				                last.pc));
 			}
 			block.terminator.false_block = fallthrough->second;
 		} else {
@@ -2131,27 +2120,26 @@ bool BuildGraph(const Decoder::Program& program, Graph& graph, std::string* erro
 		graph.unsupported = false;
 	}
 
-	return true;
+	return graph;
 }
 
 namespace {
 
-bool StructurizeImpl(Graph& graph, std::string* error) {
+bool StructurizeImpl(Graph& graph) {
 	if (graph.unsupported || graph.irreducible) {
 		if (graph.unsupported_reason.empty()) {
 			graph.unsupported_reason = "unsupported CFG";
 		}
-		SetError(error, graph.unsupported_reason);
 		return false;
 	}
 
-	if (!CanonicalizeNaturalLoops(graph, error)) {
+	if (!CanonicalizeNaturalLoops(graph)) {
 		return false;
 	}
-	if (!SplitSharedMergeBlocks(graph, error)) {
+	if (!SplitSharedMergeBlocks(graph)) {
 		return false;
 	}
-	if (!IsolateSemanticLoopHeaders(graph, error)) {
+	if (!IsolateSemanticLoopHeaders(graph)) {
 		return false;
 	}
 	ClearStructuredTerminators(graph);
@@ -2164,8 +2152,7 @@ bool StructurizeImpl(Graph& graph, std::string* error) {
 			    graph, FailureKind::StructuredControlFlow, header,
 			    fmt::format(
 			        "duplicate structured merge block {} for header {} (already used by header {})",
-			        merge, header, it->second),
-			    error);
+			        merge, header, it->second));
 			return false;
 		}
 		return true;
@@ -2176,16 +2163,14 @@ bool StructurizeImpl(Graph& graph, std::string* error) {
 		if (header == nullptr || loop.merge == UINT32_MAX || loop.continue_block == UINT32_MAX) {
 			SetFailure(
 			    graph, FailureKind::StructuredControlFlow, loop.header,
-			    fmt::format("loop at block {} has no structured merge/continue", loop.header),
-			    error);
+			    fmt::format("loop at block {} has no structured merge/continue", loop.header));
 			return false;
 		}
 		if (header->inst_begin != header->inst_end ||
 		    header->terminator.kind != TerminatorKind::Branch) {
-			SetFailure(graph, FailureKind::StructuredControlFlow, loop.header,
-			           fmt::format("loop header {} is not a dedicated empty control block",
-			                       loop.header),
-			           error);
+			SetFailure(
+			    graph, FailureKind::StructuredControlFlow, loop.header,
+			    fmt::format("loop header {} is not a dedicated empty control block", loop.header));
 			return false;
 		}
 		if (!reserve_merge_block(loop.header, loop.merge)) {
@@ -2208,8 +2193,7 @@ bool StructurizeImpl(Graph& graph, std::string* error) {
 		const auto merge = FindSelectionMerge(graph, block);
 		if (merge == UINT32_MAX) {
 			SetFailure(graph, FailureKind::StructuredControlFlow, block.id,
-			           fmt::format("conditional block {} has no structured merge", block.id),
-			           error);
+			           fmt::format("conditional block {} has no structured merge", block.id));
 			return false;
 		}
 
@@ -2224,16 +2208,15 @@ bool StructurizeImpl(Graph& graph, std::string* error) {
 
 } // namespace
 
-bool Structurize(Graph& graph, std::string* error) {
+bool Structurize(Graph& graph) {
 	Graph original = graph;
-	if (StructurizeImpl(graph, error)) {
+	if (StructurizeImpl(graph)) {
 		return true;
 	}
 
-	Graph       failed_graph = std::move(graph);
-	std::string failed_error = error != nullptr ? *error : std::string {};
-	graph                    = std::move(original);
-	const auto route_budget  = static_cast<uint32_t>(graph.blocks.size());
+	Graph failed_graph      = std::move(graph);
+	graph                   = std::move(original);
+	const auto route_budget = static_cast<uint32_t>(graph.blocks.size());
 	// Apply one route at a time and retry. Eagerly routing every matching diamond can
 	// rewrite unrelated selections that were already structurally valid.
 	for (uint32_t route_variable = 0; route_variable < route_budget; route_variable++) {
@@ -2241,18 +2224,12 @@ bool Structurize(Graph& graph, std::string* error) {
 			break;
 		}
 		Graph routed = graph;
-		if (error != nullptr) {
-			error->clear();
-		}
-		if (StructurizeImpl(routed, error)) {
+		if (StructurizeImpl(routed)) {
 			graph = std::move(routed);
 			return true;
 		}
 	}
 	graph = std::move(failed_graph);
-	if (error != nullptr) {
-		*error = std::move(failed_error);
-	}
 	return false;
 }
 

src/graphics/shader/recompiler/frontend/cfg/ShaderCFG.h

@@ -108,8 +108,8 @@ struct Graph {
 	uint32_t          FindNearestCommonPostDominator(uint32_t block_a, uint32_t block_b) const;
 };
 
-bool        BuildGraph(const Decoder::Program& program, Graph& graph, std::string* error);
-bool        Structurize(Graph& graph, std::string* error);
+Graph       BuildGraph(const Decoder::Program& program);
+bool        Structurize(Graph& graph);
 std::string BranchConditionToString(BranchCondition condition);
 std::string FailureKindToString(FailureKind kind);
 std::string GraphToString(const Graph& graph);

src/graphics/shader/recompiler/frontend/decode/ImageOps.cpp

@@ -357,8 +357,8 @@ void DecodeMimg(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 
 	DecodeVectorGpr(vdata, inst.dst);
 	DecodeVectorGpr(vaddr, inst.src0);
-	DecodeScalarSource(srsrc * 4u, pc, inst.src1, nullptr);
-	DecodeScalarSource(ssamp * 4u, pc, inst.src2, nullptr);
+	DecodeScalarSource(srsrc * 4u, pc, inst.src1);
+	DecodeScalarSource(ssamp * 4u, pc, inst.src2);
 	inst.src_count = 3;
 }
 

src/graphics/shader/recompiler/frontend/decode/MemoryOps.cpp

@@ -103,8 +103,11 @@ constexpr MemoryOpcodeInfo DS_OPCODE_LIST[] = {
     {0x4du, Opcode::DS_WRITE_B64, 2, 32},        {0x4eu, Opcode::DS_WRITE2_B64, 4, 32},
     {0x4fu, Opcode::DS_WRITE2ST64_B64, 4, 32},   {0x76u, Opcode::DS_READ_B64, 2, 32},
     {0x77u, Opcode::DS_READ2_B64, 4, 32},        {0x78u, Opcode::DS_READ2ST64_B64, 4, 32},
+    {0xa1u, Opcode::DS_WRITE_B16_D16_HI, 1, 16},
     {0xa6u, Opcode::DS_READ_U16_D16, 1, 16},
+    {0xa7u, Opcode::DS_READ_U16_D16_HI, 1, 16},
     {0xb0u, Opcode::DS_WRITE_ADDTID_B32, 1, 32}, {0xb1u, Opcode::DS_READ_ADDTID_B32, 1, 32},
+    {0xb3u, Opcode::DS_BPERMUTE_B32, 1, 32},
     {0xdeu, Opcode::DS_WRITE_B96, 3, 32},        {0xdfu, Opcode::DS_WRITE_B128, 4, 32},
     {0xfeu, Opcode::DS_READ_B96, 3, 32},         {0xffu, Opcode::DS_READ_B128, 4, 32},
 };
@@ -147,6 +150,7 @@ bool IsDsWriteOpcode(Opcode opcode) {
 	switch (opcode) {
 		case Opcode::DS_WRITE_B8:
 		case Opcode::DS_WRITE_B16:
+		case Opcode::DS_WRITE_B16_D16_HI:
 		case Opcode::DS_WRITE2_B32:
 		case Opcode::DS_WRITE2ST64_B32:
 		case Opcode::DS_WRITE2_B64:
@@ -192,6 +196,7 @@ uint32_t DsSourceCount(Opcode opcode) {
 		case Opcode::DS_WRITE2ST64_B64:
 		case Opcode::DS_MIN_F32:
 		case Opcode::DS_MAX_F32: return 3u;
+		case Opcode::DS_BPERMUTE_B32: return 2u;
 		case Opcode::DS_READ_ADDTID_B32:
 		case Opcode::DS_CONSUME:
 		case Opcode::DS_APPEND: return 0u;
@@ -236,11 +241,11 @@ void DecodeSmem(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 		MarkMemoryUnsupported(inst, Family::SMEM, opcode, "SMEM opcode is not implemented");
 	}
 
-	DecodeScalarDestination(sdst, pc, inst.dst, nullptr);
+	DecodeScalarDestination(sdst, pc, inst.dst);
 	// SMEM encodes SBASE in SGPR pairs. Scalar-buffer loads still use the same
 	// pair index; their descriptor operand consumes four SGPRs from that base.
-	DecodeScalarSource(sbase * 2u, pc, inst.src0, nullptr);
-	DecodeScalarSource(soffset, pc, inst.src1, nullptr);
+	DecodeScalarSource(sbase * 2u, pc, inst.src0);
+	DecodeScalarSource(soffset, pc, inst.src1);
 	inst.src_count = 2;
 }
 
@@ -273,8 +278,8 @@ void DecodeMubuf(uint32_t pc, std::span<const uint32_t> code, uint32_t word_inde
 
 	DecodeVectorGpr(vdata, inst.dst);
 	DecodeVectorGpr(vaddr, inst.src0);
-	DecodeScalarSource(srsrc * 4u, pc, inst.src1, nullptr);
-	DecodeScalarSource(soffset, pc, inst.src2, nullptr);
+	DecodeScalarSource(srsrc * 4u, pc, inst.src1);
+	DecodeScalarSource(soffset, pc, inst.src2);
 	inst.src_count = 3;
 }
 
@@ -311,8 +316,8 @@ void DecodeMtbuf(uint32_t pc, std::span<const uint32_t> code, uint32_t word_inde
 
 	DecodeVectorGpr(vdata, inst.dst);
 	DecodeVectorGpr(vaddr, inst.src0);
-	DecodeScalarSource(srsrc * 4u, pc, inst.src1, nullptr);
-	DecodeScalarSource(soffset, pc, inst.src2, nullptr);
+	DecodeScalarSource(srsrc * 4u, pc, inst.src1);
+	DecodeScalarSource(soffset, pc, inst.src2);
 	inst.src_count = 3;
 }
 
@@ -359,7 +364,7 @@ void DecodeFlat(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 		DecodeVectorGpr(addr + 1u, inst.src1);
 		inst.src_count = 2;
 	} else {
-		DecodeScalarSource(saddr, pc, inst.src1, nullptr);
+		DecodeScalarSource(saddr, pc, inst.src1);
 		inst.src_count = 2;
 	}
 }
@@ -392,9 +397,10 @@ void DecodeDs(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
 		SetUnsupported(inst, Family::DS, opcode, "DS swizzle FFT mode is not implemented");
 	}
 	if (inst.gds &&
-	    (inst.opcode == Opcode::DS_SWIZZLE_B32 || inst.opcode == Opcode::DS_WRITE_ADDTID_B32 ||
+	    (inst.opcode == Opcode::DS_SWIZZLE_B32 || inst.opcode == Opcode::DS_BPERMUTE_B32 ||
+	     inst.opcode == Opcode::DS_WRITE_ADDTID_B32 ||
 	     inst.opcode == Opcode::DS_READ_ADDTID_B32)) {
-		SetUnsupported(inst, Family::DS, opcode, "DS swizzle/addtid is available only for LDS");
+		SetUnsupported(inst, Family::DS, opcode, "DS lane operation is available only for LDS");
 	}
 	if (inst.opcode == Opcode::DS_WRITE_ADDTID_B32 && data1 != 0u) {
 		SetUnsupported(inst, Family::DS, opcode,
@@ -421,13 +427,17 @@ void DecodeDs(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
 	}
 
 	DecodeVectorGpr(vdst, inst.dst);
-	if (inst.opcode == Opcode::DS_READ_U16_D16) {
+	if (inst.opcode == Opcode::DS_READ_U16_D16 ||
+	    inst.opcode == Opcode::DS_READ_U16_D16_HI) {
 		// Native D16 reads update only the selected destination half. Reuse the partial
-		// destination representation so typed IR translation preserves the upper word.
-		inst.dst.sdwa_sel = 4u;
+		// destination representation so typed IR translation preserves the other half.
+		inst.dst.sdwa_sel = inst.opcode == Opcode::DS_READ_U16_D16 ? 4u : 5u;
 	}
 	DecodeVectorGpr(addr, inst.src0);
 	DecodeVectorGpr(data0, inst.src1);
+	if (inst.opcode == Opcode::DS_WRITE_B16_D16_HI) {
+		inst.src1.sdwa_sel = 5u;
+	}
 	DecodeVectorGpr(data1, inst.src2);
 	inst.src_count = DsSourceCount(inst.opcode);
 }

src/graphics/shader/recompiler/frontend/decode/ScalarAluOps.cpp

@@ -45,6 +45,7 @@ constexpr OpcodeMap SOP1_OPCODE_LIST[] = {
     {0x10u, Opcode::S_BCNT1_I32_B64},
     {0x13u, Opcode::S_FF1_I32_B32},
     {0x14u, Opcode::S_FF1_I32_B64},
+    {0x15u, Opcode::S_FLBIT_I32_B32},
     {0x16u, Opcode::S_FLBIT_I32_B64},
     {0x1bu, Opcode::S_BITSET0_B32},
     {0x1du, Opcode::S_BITSET1_B32},
@@ -105,21 +106,18 @@ constexpr auto SOPK_OPS = Detail::MakeOpcodeTable<0x20>(SOPK_OPCODE_LIST);
 constexpr auto SOPC_OPS = Detail::MakeOpcodeTable<0x80>(SOPC_OPCODE_LIST);
 constexpr auto SOPP_OPS = Detail::MakeOpcodeTable<0x80>(SOPP_OPCODE_LIST);
 
-bool DecodeBinarySources(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
-                         Instruction& inst, uint32_t ssrc0, uint32_t ssrc1, std::string* error) {
-	if (!DecodeScalarSource(ssrc0, pc, inst.src0, error) ||
-	    !DecodeScalarSource(ssrc1, pc, inst.src1, error)) {
-		return false;
-	}
+void DecodeBinarySources(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                         Instruction& inst, uint32_t ssrc0, uint32_t ssrc1) {
+	DecodeScalarSource(ssrc0, pc, inst.src0);
+	DecodeScalarSource(ssrc1, pc, inst.src1);
 	inst.src_count = 2;
 	ReadLiteralOperands(code, word_index, inst);
-	return true;
 }
 
 } // namespace
 
-bool DecodeSop1(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error) {
+void DecodeSop1(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst) {
 	const uint32_t word   = code[word_index];
 	const uint32_t opcode = (word >> 8u) & 0xffu;
 	const uint32_t ssrc0  = word & 0xffu;
@@ -134,35 +132,31 @@ bool DecodeSop1(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 
 	if (inst.opcode == Opcode::UNSUPPORTED) {
 		SetUnsupported(inst, Family::SOP1, opcode, "SOP1 opcode is not implemented");
-		return true;
+		return;
 	}
 
 	switch (inst.opcode) {
 		case Opcode::S_GETPC_B64:
 			inst.src_count = 0;
-			return DecodeScalarDestination(sdst, pc, inst.dst, error);
+			DecodeScalarDestination(sdst, pc, inst.dst);
+			return;
 		case Opcode::S_SETPC_B64:
 			inst.src_count = 1;
 			inst.dst.kind  = OperandKind::Null;
-			if (!DecodeScalarSource(ssrc0, pc, inst.src0, error)) {
-				return false;
-			}
+			DecodeScalarSource(ssrc0, pc, inst.src0);
 			ReadLiteralOperands(code, word_index, inst);
-			return true;
+			return;
 		default: break;
 	}
 
-	if (!DecodeScalarSource(ssrc0, pc, inst.src0, error) ||
-	    !DecodeScalarDestination(sdst, pc, inst.dst, error)) {
-		return false;
-	}
+	DecodeScalarSource(ssrc0, pc, inst.src0);
+	DecodeScalarDestination(sdst, pc, inst.dst);
 	inst.src_count = 1;
 	ReadLiteralOperands(code, word_index, inst);
-	return true;
 }
 
-bool DecodeSop2(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error) {
+void DecodeSop2(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst) {
 	const uint32_t word   = code[word_index];
 	const uint32_t opcode = (word >> 23u) & 0x7fu;
 	const uint32_t ssrc1  = (word >> 8u) & 0xffu;
@@ -178,17 +172,15 @@ bool DecodeSop2(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 
 	if (inst.opcode == Opcode::UNSUPPORTED) {
 		SetUnsupported(inst, Family::SOP2, opcode, "SOP2 opcode is not implemented");
-		return true;
+		return;
 	}
 
-	if (!DecodeScalarDestination(sdst, pc, inst.dst, error)) {
-		return false;
-	}
-	return DecodeBinarySources(pc, code, word_index, inst, ssrc0, ssrc1, error);
+	DecodeScalarDestination(sdst, pc, inst.dst);
+	DecodeBinarySources(pc, code, word_index, inst, ssrc0, ssrc1);
 }
 
-bool DecodeSopk(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error) {
+void DecodeSopk(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst) {
 	const uint32_t word   = code[word_index];
 	const uint32_t opcode = (word >> 23u) & 0x1fu;
 	const uint32_t sdst   = (word >> 16u) & 0x7fu;
@@ -207,18 +199,18 @@ bool DecodeSopk(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 
 	if (inst.opcode == Opcode::UNSUPPORTED) {
 		SetUnsupported(inst, Family::SOPK, opcode, "SOPK opcode is not implemented");
-		return true;
+		return;
 	}
 
 	switch (inst.opcode) {
-		case Opcode::S_MOVK_I32: return DecodeScalarDestination(sdst, pc, inst.dst, error);
+		case Opcode::S_MOVK_I32: DecodeScalarDestination(sdst, pc, inst.dst); return;
 		case Opcode::S_WAITCNT: {
 			const uint32_t waitcnt = word & 0xffffu;
 			inst.dst.kind          = OperandKind::Null;
 			inst.src0.signed_val   = static_cast<int32_t>(waitcnt);
 			inst.src0.value        = waitcnt;
 			inst.src_count         = 1;
-			return true;
+			return;
 		}
 		case Opcode::S_SETREG_B32:
 			inst.dst.kind        = OperandKind::Null;
@@ -226,26 +218,25 @@ bool DecodeSopk(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 			inst.src1.value      = word & 0xffffu;
 			inst.src1.signed_val = static_cast<int32_t>(imm);
 			inst.src_count       = 2;
-			return DecodeScalarSource(sdst, pc, inst.src0, error);
+			DecodeScalarSource(sdst, pc, inst.src0);
+			return;
 		default: break;
 	}
 
 	inst.src1 = inst.src0;
-	if (!DecodeScalarSource(sdst, pc, inst.src0, error)) {
-		return false;
-	}
+	DecodeScalarSource(sdst, pc, inst.src0);
 	if (inst.opcode == Opcode::S_ADD_I32 || inst.opcode == Opcode::S_MULK_I32) {
 		inst.src_count = 2;
-		return DecodeScalarDestination(sdst, pc, inst.dst, error);
+		DecodeScalarDestination(sdst, pc, inst.dst);
+		return;
 	}
 
 	inst.dst.kind  = OperandKind::Scc;
 	inst.src_count = 2;
-	return true;
 }
 
-bool DecodeSopc(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error) {
+void DecodeSopc(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst) {
 	const uint32_t word   = code[word_index];
 	const uint32_t ssrc1  = (word >> 8u) & 0xffu;
 	const uint32_t ssrc0  = word & 0xffu;
@@ -261,10 +252,10 @@ bool DecodeSopc(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index
 
 	if (inst.opcode == Opcode::UNSUPPORTED) {
 		SetUnsupported(inst, Family::SOPC, opcode, "SOPC opcode is not implemented");
-		return true;
+		return;
 	}
 
-	return DecodeBinarySources(pc, code, word_index, inst, ssrc0, ssrc1, error);
+	DecodeBinarySources(pc, code, word_index, inst, ssrc0, ssrc1);
 }
 
 void DecodeSopp(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,

src/graphics/shader/recompiler/frontend/decode/ScalarAluOps.h

@@ -5,14 +5,14 @@
 
 namespace Libs::Graphics::ShaderRecompiler::Decoder {
 
-bool DecodeSop1(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
-bool DecodeSop2(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
-bool DecodeSopk(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
-bool DecodeSopc(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
+void DecodeSop1(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
+void DecodeSop2(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
+void DecodeSopk(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
+void DecodeSopc(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
 void DecodeSopp(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
                 Instruction& inst);
 

src/graphics/shader/recompiler/frontend/decode/ShaderDecoder.cpp

@@ -1,5 +1,6 @@
 #include "graphics/shader/recompiler/frontend/decode/ShaderDecoder.h"
 
+#include "common/assert.h"
 #include "graphics/shader/recompiler/frontend/decode/ExportOps.h"
 #include "graphics/shader/recompiler/frontend/decode/ImageOps.h"
 #include "graphics/shader/recompiler/frontend/decode/MemoryOps.h"
@@ -17,12 +18,6 @@ uint32_t FloatBits(float value) {
 	return std::bit_cast<uint32_t>(value);
 }
 
-void SetError(std::string* error, const char* message) {
-	if (error != nullptr) {
-		*error = message;
-	}
-}
-
 bool HasLiteral(const Instruction& inst) {
 	return inst.src0.kind == OperandKind::LiteralConstant ||
 	       inst.src1.kind == OperandKind::LiteralConstant ||
@@ -216,86 +211,76 @@ const char* ImageDimensionToString(ImageDimension dimension) {
 	}
 }
 
-bool DecodeScalarSource(uint32_t code, uint32_t pc, Operand& operand, std::string* error) {
+void DecodeScalarSource(uint32_t code, uint32_t pc, Operand& operand) {
 	operand = {};
 
 	if (code <= 105u) {
 		operand.kind = OperandKind::Sgpr;
 		operand.reg  = code;
-		return true;
+		return;
 	}
 	if (code >= 128u && code <= 192u) {
 		operand.kind       = OperandKind::IntegerInlineConstant;
 		operand.signed_val = static_cast<int32_t>(code - 128u);
 		operand.value      = static_cast<uint32_t>(operand.signed_val);
-		return true;
+		return;
 	}
 	if (code >= 193u && code <= 208u) {
 		operand.kind       = OperandKind::IntegerInlineConstant;
 		operand.signed_val = 192 - static_cast<int32_t>(code);
 		operand.value      = static_cast<uint32_t>(operand.signed_val);
-		return true;
+		return;
 	}
 	if (code >= 240u && code <= 247u) {
 		constexpr float values[] = {0.5f, -0.5f, 1.0f, -1.0f, 2.0f, -2.0f, 4.0f, -4.0f};
 		operand.kind             = OperandKind::FloatInlineConstant;
 		operand.float_val        = values[code - 240u];
 		operand.value            = FloatBits(operand.float_val);
-		return true;
+		return;
 	}
 	if (code >= 256u && code <= 511u) {
 		DecodeVectorGpr(code - 256u, operand);
-		return true;
+		return;
 	}
 
 	switch (code) {
-		case 106u: operand.kind = OperandKind::VccLo; return true;
-		case 107u: operand.kind = OperandKind::VccHi; return true;
-		case 124u: operand.kind = OperandKind::M0; return true;
-		case 125u: operand.kind = OperandKind::Null; return true;
-		case 126u: operand.kind = OperandKind::ExecLo; return true;
-		case 127u: operand.kind = OperandKind::ExecHi; return true;
-		case 239u: operand.kind = OperandKind::PopsExitingWaveId; return true;
+		case 106u: operand.kind = OperandKind::VccLo; return;
+		case 107u: operand.kind = OperandKind::VccHi; return;
+		case 124u: operand.kind = OperandKind::M0; return;
+		case 125u: operand.kind = OperandKind::Null; return;
+		case 126u: operand.kind = OperandKind::ExecLo; return;
+		case 127u: operand.kind = OperandKind::ExecHi; return;
+		case 239u: operand.kind = OperandKind::PopsExitingWaveId; return;
 		case 248u:
 			operand.kind      = OperandKind::FloatInlineConstant;
 			operand.float_val = 0.15915494309189535f;
 			operand.value     = FloatBits(operand.float_val);
-			return true;
-		case 251u: operand.kind = OperandKind::VccZ; return true;
-		case 252u: operand.kind = OperandKind::ExecZ; return true;
-		case 253u: operand.kind = OperandKind::Scc; return true;
-		case 255u: operand.kind = OperandKind::LiteralConstant; return true;
-		default:
-			if (error != nullptr) {
-				*error = fmt::format("unsupported scalar source operand 0x{:08x} at pc 0x{:08x}",
-				                     code, pc);
-			}
-			return false;
+			return;
+		case 251u: operand.kind = OperandKind::VccZ; return;
+		case 252u: operand.kind = OperandKind::ExecZ; return;
+		case 253u: operand.kind = OperandKind::Scc; return;
+		case 255u: operand.kind = OperandKind::LiteralConstant; return;
+		default: EXIT("unsupported scalar source operand 0x%08x at pc 0x%08x", code, pc);
 	}
 }
 
-bool DecodeScalarDestination(uint32_t code, uint32_t pc, Operand& operand, std::string* error) {
+void DecodeScalarDestination(uint32_t code, uint32_t pc, Operand& operand) {
 	operand = {};
 
 	if (code <= 105u) {
 		operand.kind = OperandKind::Sgpr;
 		operand.reg  = code;
-		return true;
+		return;
 	}
 
 	switch (code) {
-		case 106u: operand.kind = OperandKind::VccLo; return true;
-		case 107u: operand.kind = OperandKind::VccHi; return true;
-		case 124u: operand.kind = OperandKind::M0; return true;
-		case 125u: operand.kind = OperandKind::Null; return true;
-		case 126u: operand.kind = OperandKind::ExecLo; return true;
-		case 127u: operand.kind = OperandKind::ExecHi; return true;
-		default:
-			if (error != nullptr) {
-				*error = fmt::format(
-				    "unsupported scalar destination operand 0x{:08x} at pc 0x{:08x}", code, pc);
-			}
-			return false;
+		case 106u: operand.kind = OperandKind::VccLo; return;
+		case 107u: operand.kind = OperandKind::VccHi; return;
+		case 124u: operand.kind = OperandKind::M0; return;
+		case 125u: operand.kind = OperandKind::Null; return;
+		case 126u: operand.kind = OperandKind::ExecLo; return;
+		case 127u: operand.kind = OperandKind::ExecHi; return;
+		default: EXIT("unsupported scalar destination operand 0x%08x at pc 0x%08x", code, pc);
 	}
 }
 
@@ -368,39 +353,33 @@ Family GetInstructionFamily(uint32_t word) {
 	}
 }
 
-bool DecodeInstruction(std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                       std::string* error) {
+void DecodeInstruction(std::span<const uint32_t> code, uint32_t word_index, Instruction& inst) {
 	const uint32_t pc = word_index * sizeof(uint32_t);
 	switch (GetInstructionFamily(code[word_index])) {
-		case Family::SOP1: return DecodeSop1(pc, code, word_index, inst, error);
-		case Family::SOP2: return DecodeSop2(pc, code, word_index, inst, error);
-		case Family::SOPK: return DecodeSopk(pc, code, word_index, inst, error);
-		case Family::SOPC: return DecodeSopc(pc, code, word_index, inst, error);
-		case Family::SOPP: DecodeSopp(pc, code, word_index, inst); return true;
-		case Family::VOP1: return DecodeVop1(pc, code, word_index, inst, error);
-		case Family::VOP2: return DecodeVop2(pc, code, word_index, inst, error);
-		case Family::VOP3: return DecodeVop3(pc, code, word_index, inst, error);
-		case Family::VOP3P: return DecodeVop3p(pc, code, word_index, inst, error);
-		case Family::VOPC: return DecodeVopc(pc, code, word_index, inst, error);
-		case Family::VINTRP: DecodeVintrp(pc, code, word_index, inst); return true;
-		case Family::SMEM: DecodeSmem(pc, code, word_index, inst); return true;
-		case Family::MUBUF: DecodeMubuf(pc, code, word_index, inst); return true;
-		case Family::MTBUF: DecodeMtbuf(pc, code, word_index, inst); return true;
-		case Family::FLAT: DecodeFlat(pc, code, word_index, inst); return true;
-		case Family::DS: DecodeDs(pc, code, word_index, inst); return true;
-		case Family::MIMG: DecodeMimg(pc, code, word_index, inst); return true;
-		case Family::EXP: DecodeExp(pc, code, word_index, inst); return true;
+		case Family::SOP1: DecodeSop1(pc, code, word_index, inst); return;
+		case Family::SOP2: DecodeSop2(pc, code, word_index, inst); return;
+		case Family::SOPK: DecodeSopk(pc, code, word_index, inst); return;
+		case Family::SOPC: DecodeSopc(pc, code, word_index, inst); return;
+		case Family::SOPP: DecodeSopp(pc, code, word_index, inst); return;
+		case Family::VOP1: DecodeVop1(pc, code, word_index, inst); return;
+		case Family::VOP2: DecodeVop2(pc, code, word_index, inst); return;
+		case Family::VOP3: DecodeVop3(pc, code, word_index, inst); return;
+		case Family::VOP3P: DecodeVop3p(pc, code, word_index, inst); return;
+		case Family::VOPC: DecodeVopc(pc, code, word_index, inst); return;
+		case Family::VINTRP: DecodeVintrp(pc, code, word_index, inst); return;
+		case Family::SMEM: DecodeSmem(pc, code, word_index, inst); return;
+		case Family::MUBUF: DecodeMubuf(pc, code, word_index, inst); return;
+		case Family::MTBUF: DecodeMtbuf(pc, code, word_index, inst); return;
+		case Family::FLAT: DecodeFlat(pc, code, word_index, inst); return;
+		case Family::DS: DecodeDs(pc, code, word_index, inst); return;
+		case Family::MIMG: DecodeMimg(pc, code, word_index, inst); return;
+		case Family::EXP: DecodeExp(pc, code, word_index, inst); return;
 		default:
-			if (error != nullptr) {
-				*error =
-				    fmt::format("unknown RDNA2 instruction family at pc 0x{:08x}, raw=0x{:08x}", pc,
-				                code[word_index]);
-			}
-			return false;
+			EXIT("unknown RDNA2 instruction family at pc 0x%08x, raw=0x%08x", pc, code[word_index]);
 	}
 }
 
-bool DecodeProgram(std::span<const uint32_t> code, Program& program, std::string* error) {
+void DecodeProgram(std::span<const uint32_t> code, Program& program) {
 	program.instructions.clear();
 	program.instructions.reserve(code.size());
 	program.code = code;
@@ -408,10 +387,7 @@ bool DecodeProgram(std::span<const uint32_t> code, Program& program, std::string
 	std::vector<bool> branch_targets;
 	for (uint32_t word_index = 0; word_index < code.size();) {
 		program.instructions.emplace_back();
-		if (!DecodeInstruction(code, word_index, program.instructions.back(), error)) {
-			program.instructions.pop_back();
-			return false;
-		}
+		DecodeInstruction(code, word_index, program.instructions.back());
 
 		const auto& inst = program.instructions.back();
 		word_index += inst.word_count;
@@ -425,12 +401,11 @@ bool DecodeProgram(std::span<const uint32_t> code, Program& program, std::string
 		}
 		if (inst.opcode == Opcode::S_ENDPGM &&
 		    (word_index >= code.size() || branch_targets.empty() || !branch_targets[word_index])) {
-			return true;
+			return;
 		}
 	}
 
-	SetError(error, "shader decode reached the code boundary before S_ENDPGM");
-	return false;
+	EXIT("shader decode reached the code boundary before S_ENDPGM");
 }
 
 std::string OperandToString(const Operand& operand) {
@@ -501,6 +476,7 @@ std::string InstructionToString(const Instruction& inst) {
 		case Opcode::S_ABS_I32:
 		case Opcode::S_BREV_B32:
 		case Opcode::S_BCNT1_I32_B32:
+		case Opcode::S_FLBIT_I32_B32:
 		case Opcode::S_FF1_I32_B32:
 		case Opcode::S_FF1_I32_B64:
 		case Opcode::S_NOT_B64:
@@ -661,6 +637,7 @@ std::string InstructionToString(const Instruction& inst) {
 		case Opcode::DS_MIN_F32:
 		case Opcode::DS_MAX_F32:
 		case Opcode::DS_SWIZZLE_B32:
+		case Opcode::DS_BPERMUTE_B32:
 		case Opcode::DS_READ_I8:
 		case Opcode::DS_READ_U8:
 		case Opcode::DS_READ_I16:
@@ -675,6 +652,7 @@ std::string InstructionToString(const Instruction& inst) {
 		case Opcode::DS_READ_B128:
 		case Opcode::DS_WRITE_B8:
 		case Opcode::DS_WRITE_B16:
+		case Opcode::DS_WRITE_B16_D16_HI:
 		case Opcode::DS_WRITE2_B32:
 		case Opcode::DS_WRITE2ST64_B32:
 		case Opcode::DS_WRITE2_B64:

src/graphics/shader/recompiler/frontend/decode/ShaderDecoder.h

@@ -45,6 +45,7 @@ enum class Opcode {
 	S_BCNT1_I32_B64,
 	S_FF1_I32_B32,
 	S_FF1_I32_B64,
+	S_FLBIT_I32_B32,
 	S_FLBIT_I32_B64,
 	S_BITREPLICATE_B64_B32,
 	S_QUADMASK_B64,
@@ -179,12 +180,15 @@ enum class Opcode {
 	V_CEIL_F16,
 	V_TRUNC_F16,
 	V_RNDNE_F16,
+	V_SIN_F16,
+	V_COS_F16,
 	V_SIN_F32,
 	V_COS_F32,
 	V_NOT_B32,
 	V_BFREV_B32,
 	V_FFBH_U32,
 	V_FFBL_B32,
+	V_FFBH_I32,
 	V_ADD_F32,
 	V_SUB_F32,
 	V_SUBREV_F32,
@@ -348,6 +352,7 @@ enum class Opcode {
 	V_CMP_GE_I32,
 	V_CMP_T_I32,
 	V_CMP_CLASS_F32,
+	V_CMPX_CLASS_F32,
 	V_CMP_LT_I16,
 	V_CMP_EQ_I16,
 	V_CMP_LE_I16,
@@ -391,6 +396,8 @@ enum class Opcode {
 	V_CMP_GE_U32,
 	V_CMP_T_U32,
 	V_CMP_EQ_I64,
+	V_CMP_LT_U64,
+	V_CMP_EQ_U64,
 	V_CMP_GT_U64,
 	V_CMP_NE_U64,
 	V_CMPX_NE_I64,
@@ -490,6 +497,7 @@ enum class Opcode {
 	DS_MIN_F32,
 	DS_MAX_F32,
 	DS_SWIZZLE_B32,
+	DS_BPERMUTE_B32,
 	DS_CONSUME,
 	DS_APPEND,
 	DS_READ_I8,
@@ -497,6 +505,7 @@ enum class Opcode {
 	DS_READ_I16,
 	DS_READ_U16,
 	DS_READ_U16_D16,
+	DS_READ_U16_D16_HI,
 	DS_READ2_B32,
 	DS_READ2ST64_B32,
 	DS_READ_B32,
@@ -507,6 +516,7 @@ enum class Opcode {
 	DS_READ_B128,
 	DS_WRITE_B8,
 	DS_WRITE_B16,
+	DS_WRITE_B16_D16_HI,
 	DS_WRITE2_B32,
 	DS_WRITE2ST64_B32,
 	DS_WRITE2_B64,
@@ -624,6 +634,7 @@ struct Operand {
 	uint32_t dpp_ctrl           = 0;
 	uint32_t dpp_row_mask       = 0xf;
 	uint32_t dpp_bank_mask      = 0xf;
+	bool     explicit_sdwa_dst  = false;
 	bool     sdwa_sext          = false;
 	bool     dpp_fetch_inactive = false;
 	bool     dpp_bound_ctrl     = false;
@@ -695,12 +706,11 @@ struct Program {
 // Code spans are trusted to contain complete instructions, valid branch targets, and 32-bit PCs.
 Family GetInstructionFamily(uint32_t word);
 // The output object must be freshly initialized.
-bool DecodeInstruction(std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                       std::string* error);
-bool DecodeProgram(std::span<const uint32_t> code, Program& program, std::string* error);
+void DecodeInstruction(std::span<const uint32_t> code, uint32_t word_index, Instruction& inst);
+void DecodeProgram(std::span<const uint32_t> code, Program& program);
 
-bool DecodeScalarSource(uint32_t code, uint32_t pc, Operand& operand, std::string* error);
-bool DecodeScalarDestination(uint32_t code, uint32_t pc, Operand& operand, std::string* error);
+void DecodeScalarSource(uint32_t code, uint32_t pc, Operand& operand);
+void DecodeScalarDestination(uint32_t code, uint32_t pc, Operand& operand);
 void DecodeVectorGpr(uint32_t reg, Operand& operand);
 void ReadLiteralOperands(std::span<const uint32_t> code, uint32_t word_index, Instruction& inst);
 void SetRawWords(Instruction& inst, std::span<const uint32_t> code, uint32_t word_index,

src/graphics/shader/recompiler/frontend/decode/VectorAluOps.h

@@ -5,16 +5,16 @@
 
 namespace Libs::Graphics::ShaderRecompiler::Decoder {
 
-bool DecodeVop2(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
-bool DecodeVop1(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
-bool DecodeVopc(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
-bool DecodeVop3(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index, Instruction& inst,
-                std::string* error);
-bool DecodeVop3p(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
-                 Instruction& inst, std::string* error);
+void DecodeVop2(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
+void DecodeVop1(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
+void DecodeVopc(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
+void DecodeVop3(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                Instruction& inst);
+void DecodeVop3p(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
+                 Instruction& inst);
 void DecodeVintrp(uint32_t pc, std::span<const uint32_t> code, uint32_t word_index,
                   Instruction& inst);
 

src/graphics/shader/recompiler/frontend/translate/Attribute.cpp

@@ -1,9 +1,8 @@
+#include "common/assert.h"
 #include "graphics/shader/recompiler/frontend/translate/Translator.h"
-#include "graphics/shader/shader.h"
 
 #include <algorithm>
 #include <array>
-#include <fmt/format.h>
 
 namespace Libs::Graphics::ShaderRecompiler::Frontend {
 
@@ -47,14 +46,13 @@ IR::ExportFlags Translator::AddExportInfo(const Decoder::Instruction& inst) {
 	return {.index = index, .pc = inst.pc};
 }
 
-bool Translator::TranslateEmbeddedFetch(const Decoder::Instruction& inst, uint32_t attribute,
+void Translator::TranslateEmbeddedFetch(const Decoder::Instruction& inst, uint32_t attribute,
                                         uint32_t component_count) {
 	for (uint32_t component = 0; component < component_count; component++) {
 		const auto value =
 		    ir.Emit(IR::ValueOpcode::GetAttribute, {IR::Value(attribute), IR::Value(component)});
 		WriteOperand(OffsetOperand(inst.dst, component), value);
 	}
-	return true;
 }
 
 void Translator::V_INTERP_P1_F32() {}
@@ -65,25 +63,20 @@ void Translator::V_INTERP_P2_F32(const Decoder::Instruction& inst) {
 	WriteOperand(inst.dst, value);
 }
 
-bool Translator::V_INTERP_MOV_F32(const Decoder::Instruction& inst, std::string* error) {
+void Translator::V_INTERP_MOV_F32(const Decoder::Instruction& inst) {
 	if (inst.src0.value >= 3u) {
-		return ShaderError::Fail(error,
-		                         fmt::format("v_interp_mov_f32 mode {} is reserved at pc 0x{:08x}",
-		                                     inst.src0.value, inst.pc));
+		EXIT("v_interp_mov_f32 mode %u is reserved at pc 0x%08x", inst.src0.value, inst.pc);
 	}
 	const auto value = ir.Emit(
 	    IR::ValueOpcode::GetInterpolationParameter,
 	    {IR::Value(inst.src1.value), IR::Value(inst.src2.value), IR::Value(inst.src0.value)});
 	WriteOperand(inst.dst, value);
-	return true;
 }
 
-bool Translator::EXP(const Decoder::Instruction& inst, std::string* error) {
+void Translator::EXP(const Decoder::Instruction& inst) {
 	uint32_t index = 0;
 	if (ExportTargetKindFromTarget(inst.exp.target, index) == IR::ExportTargetKind::Unknown) {
-		return ShaderError::Fail(error,
-		                         fmt::format("unsupported EXP target 0x{:02x} at pc 0x{:08x}",
-		                                     inst.exp.target, inst.pc));
+		EXIT("unsupported EXP target 0x%02x at pc 0x%08x", inst.exp.target, inst.pc);
 	}
 	std::array<IR::Value, 4> components {IR::Value(0u), IR::Value(0u), IR::Value(0u),
 	                                     IR::Value(0u)};
@@ -93,14 +86,13 @@ bool Translator::EXP(const Decoder::Instruction& inst, std::string* error) {
 	const auto data = ir.Emit(IR::ValueOpcode::CompositeConstructU32x4,
 	                          {components[0], components[1], components[2], components[3]});
 	ir.Emit(IR::ValueOpcode::SetAttribute, {data, ir.GetExec()}, AddExportInfo(inst));
-	return true;
 }
 
-bool Translator::EmitInterpolation(const Decoder::Instruction& inst, std::string* error) {
+bool Translator::EmitInterpolation(const Decoder::Instruction& inst) {
 	switch (inst.opcode) {
 		case Decoder::Opcode::V_INTERP_P1_F32: V_INTERP_P1_F32(); return true;
 		case Decoder::Opcode::V_INTERP_P2_F32: V_INTERP_P2_F32(inst); return true;
-		case Decoder::Opcode::V_INTERP_MOV_F32: return V_INTERP_MOV_F32(inst, error);
+		case Decoder::Opcode::V_INTERP_MOV_F32: V_INTERP_MOV_F32(inst); return true;
 		default: return false;
 	}
 }

src/graphics/shader/recompiler/frontend/translate/Compare.cpp

@@ -55,11 +55,11 @@ void Translator::EmitFloatOrderedCompare(const Decoder::Instruction& inst, bool
 	EmitCompareResult(inst, ordered ? ir.LogicalNot(unordered) : unordered, false, false);
 }
 
-void Translator::V_CMP_CLASS_F32(const Decoder::Instruction& inst) {
+void Translator::EmitFloatClassCompare(const Decoder::Instruction& inst, bool cmpx) {
 	const auto value = ReadOperand(inst.src0, IR::Type::F32);
 	const auto mask  = ReadOperand(inst.src1, IR::Type::U32);
 	EmitCompareResult(inst, IR::U1(ir.Emit(IR::ValueOpcode::FPCmpClass32, {value, mask})), false,
-	                  false);
+	                  cmpx);
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::Frontend

src/graphics/shader/recompiler/frontend/translate/Control.cpp

@@ -1,3 +1,4 @@
+#include "common/assert.h"
 #include "graphics/shader/recompiler/frontend/translate/Translator.h"
 
 namespace Libs::Graphics::ShaderRecompiler::Frontend {
@@ -153,12 +154,12 @@ void Translator::S_INST_PREFETCH() {
 	ir.Emit(IR::ValueOpcode::InstPrefetch);
 }
 
-bool Translator::S_GETPC_B64(const Decoder::Instruction& inst, std::string* error) {
+void Translator::S_GETPC_B64(const Decoder::Instruction& inst) {
 	const auto base    = IR::U64(ir.Emit(IR::ValueOpcode::GetShaderBase));
 	const auto address = IR::U64(
 	    ir.Emit(IR::ValueOpcode::IAdd64, {base, IR::Value(static_cast<uint64_t>(inst.pc) + 4u)}));
 	if (inst.dst.kind == Decoder::OperandKind::Null) {
-		return true;
+		return;
 	}
 	auto high = PlainOperand(inst.dst);
 	switch (inst.dst.kind) {
@@ -169,25 +170,19 @@ bool Translator::S_GETPC_B64(const Decoder::Instruction& inst, std::string* erro
 				high.kind = Decoder::OperandKind::VccLo;
 				high.reg  = 0u;
 			} else {
-				if (error != nullptr) {
-					*error = "S_GETPC_B64 destination does not name a valid scalar pair";
-				}
-				return false;
+				EXIT("S_GETPC_B64 destination does not name a valid scalar pair at pc 0x%08x",
+				     inst.pc);
 			}
 			break;
 		case Decoder::OperandKind::VccLo: high.kind = Decoder::OperandKind::VccHi; break;
 		case Decoder::OperandKind::M0: high.kind = Decoder::OperandKind::Null; break;
 		case Decoder::OperandKind::ExecLo: high.kind = Decoder::OperandKind::ExecHi; break;
 		default:
-			if (error != nullptr) {
-				*error = "S_GETPC_B64 destination does not name a valid scalar pair";
-			}
-			return false;
+			EXIT("S_GETPC_B64 destination does not name a valid scalar pair at pc 0x%08x", inst.pc);
 	}
 	const auto words = ExtractU64(address);
 	WriteOperand(inst.dst, words[0]);
 	WriteOperand(high, words[1]);
-	return true;
 }
 
 void Translator::S_CSELECT_B32(const Decoder::Instruction& inst) {
@@ -274,21 +269,15 @@ void Translator::S_WQM_B64(const Decoder::Instruction& inst) {
 	ir.SetScc(result);
 }
 
-bool Translator::V_MOVRELS_B32(const Decoder::Instruction& inst, std::string* error) {
+void Translator::V_MOVRELS_B32(const Decoder::Instruction& inst) {
 	if (inst.dst.kind != Decoder::OperandKind::Vgpr ||
 	    inst.src0.kind != Decoder::OperandKind::Vgpr) {
-		if (error != nullptr) {
-			*error = "V_MOVRELS_B32 requires VGPR source and destination";
-		}
-		return false;
+		EXIT("V_MOVRELS_B32 requires VGPR source and destination at pc 0x%08x", inst.pc);
 	}
 	if (inst.dst.sdwa_sel != 6u || inst.dst.omod != 0u || inst.dst.clamp ||
 	    inst.src0.sdwa_sel != 6u || inst.src0.sdwa_sext || inst.src0.negate || inst.src0.absolute ||
 	    inst.src0.dpp) {
-		if (error != nullptr) {
-			*error = "V_MOVRELS_B32 modifiers are not implemented";
-		}
-		return false;
+		EXIT("V_MOVRELS_B32 modifiers are not implemented at pc 0x%08x", inst.pc);
 	}
 	const auto base     = inst.src0.reg;
 	const auto m0       = ir.BitwiseAnd(ReadU32(ConditionOperand(Decoder::OperandKind::M0)),
@@ -299,23 +288,16 @@ bool Translator::V_MOVRELS_B32(const Decoder::Instruction& inst, std::string* er
 		selected = ir.Select(match, ir.GetVectorReg(static_cast<IR::VectorReg>(index)), selected);
 	}
 	WriteOperand(DestinationOperand(inst), selected);
-	return true;
 }
 
-bool Translator::V_MOVRELD_B32(const Decoder::Instruction& inst, std::string* error) {
+void Translator::V_MOVRELD_B32(const Decoder::Instruction& inst) {
 	if (inst.dst.kind != Decoder::OperandKind::Vgpr) {
-		if (error != nullptr) {
-			*error = "V_MOVRELD_B32 requires VGPR destination";
-		}
-		return false;
+		EXIT("V_MOVRELD_B32 requires VGPR destination at pc 0x%08x", inst.pc);
 	}
 	if (inst.dst.sdwa_sel != 6u || inst.dst.omod != 0u || inst.dst.clamp ||
 	    inst.src0.sdwa_sel != 6u || inst.src0.sdwa_sext || inst.src0.negate || inst.src0.absolute ||
 	    inst.src0.dpp) {
-		if (error != nullptr) {
-			*error = "V_MOVRELD_B32 modifiers are not implemented";
-		}
-		return false;
+		EXIT("V_MOVRELD_B32 modifiers are not implemented at pc 0x%08x", inst.pc);
 	}
 	const auto base  = inst.dst.reg;
 	const auto value = ReadU32(inst.src0);
@@ -327,7 +309,6 @@ bool Translator::V_MOVRELD_B32(const Decoder::Instruction& inst, std::string* er
 		const auto write = ir.LogicalAnd(ir.GetExec(), match);
 		ir.SetVectorReg(reg, ir.Select(write, value, ir.GetVectorReg(reg)));
 	}
-	return true;
 }
 
 void Translator::V_READFIRSTLANE_B32(const Decoder::Instruction& inst) {

src/graphics/shader/recompiler/frontend/translate/Convert.cpp

@@ -141,8 +141,10 @@ void Translator::V_CVT_OFF_F32_I4(const Decoder::Instruction& inst) {
 }
 
 void Translator::V_CVT_PKRTZ_F16_F32(const Decoder::Instruction& inst) {
-	const auto lhs = ReadOperand(SourceAt(inst, 0), IR::Type::F32);
-	const auto rhs = ReadOperand(SourceAt(inst, 1), IR::Type::F32);
+	const auto lhs = ApplyF32ResultModifiers(
+	    inst.dst, IR::F32(ReadOperand(SourceAt(inst, 0), IR::Type::F32)));
+	const auto rhs = ApplyF32ResultModifiers(
+	    inst.dst, IR::F32(ReadOperand(SourceAt(inst, 1), IR::Type::F32)));
 	WriteOperand(DestinationOperand(inst), ir.Emit(IR::ValueOpcode::PackFloat2x16Rtz, {lhs, rhs}));
 }
 

src/graphics/shader/recompiler/frontend/translate/Dispatch.cpp

@@ -1,23 +1,18 @@
+#include "common/assert.h"
 #include "graphics/shader/recompiler/frontend/translate/Translator.h"
 
 namespace Libs::Graphics::ShaderRecompiler::Frontend {
 
-bool Translator::TranslateInstruction(const Decoder::Instruction& inst, std::string* error) {
+void Translator::TranslateInstruction(const Decoder::Instruction& inst) {
 	current_opcode = inst.opcode;
 	current_pc     = inst.pc;
 
 	switch (inst.opcode) {
 		case Decoder::Opcode::UNKNOWN:
 		case Decoder::Opcode::COUNT:
-			if (error != nullptr) {
-				*error = "decoded opcode has no IR translation";
-			}
-			return false;
+			EXIT("decoded opcode has no IR translation at pc 0x%08x", inst.pc);
 		case Decoder::Opcode::UNSUPPORTED:
-			if (error != nullptr) {
-				*error = "unsupported decoded instruction: " + Decoder::InstructionToString(inst);
-			}
-			return false;
+			EXIT("unsupported decoded instruction: %s", Decoder::InstructionToString(inst).c_str());
 		default: break;
 	}
 
@@ -27,27 +22,30 @@ bool Translator::TranslateInstruction(const Decoder::Instruction& inst, std::str
 		case Decoder::Family::SOP2:
 		case Decoder::Family::SOPK:
 		case Decoder::Family::SOPC:
-		case Decoder::Family::SOPP: translated = EmitScalar(inst, error); break;
+		case Decoder::Family::SOPP: translated = EmitScalar(inst); break;
 		case Decoder::Family::VOP1:
 		case Decoder::Family::VOP2:
 		case Decoder::Family::VOP3:
 		case Decoder::Family::VOP3P:
-		case Decoder::Family::VOPC: translated = EmitVector(inst, error); break;
+		case Decoder::Family::VOPC: translated = EmitVector(inst); break;
 		case Decoder::Family::SMEM:
 		case Decoder::Family::MUBUF:
 		case Decoder::Family::MTBUF:
 		case Decoder::Family::FLAT:
 		case Decoder::Family::DS:
-		case Decoder::Family::MIMG: translated = EmitMemory(inst, error); break;
-		case Decoder::Family::VINTRP: translated = EmitInterpolation(inst, error); break;
-		case Decoder::Family::EXP: translated = EXP(inst, error); break;
+		case Decoder::Family::MIMG: translated = EmitMemory(inst); break;
+		case Decoder::Family::VINTRP: translated = EmitInterpolation(inst); break;
+		case Decoder::Family::EXP:
+			EXP(inst);
+			translated = true;
+			break;
 		default: break;
 	}
 
-	if (!translated && error != nullptr && error->empty()) {
-		*error = "decoded opcode has no IR translation";
+	if (!translated) {
+		EXIT("opcode %s at pc 0x%08x has no IR translation",
+		     Decoder::InstructionToString(inst).c_str(), inst.pc);
 	}
-	return translated;
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::Frontend

src/graphics/shader/recompiler/frontend/translate/Float.cpp

@@ -63,6 +63,42 @@ bool Translator::Float16Unary(const Decoder::Instruction& inst, IR::ValueOpcode
 	return true;
 }
 
+bool Translator::Float16Trig(const Decoder::Instruction& inst, IR::ValueOpcode opcode) {
+	const auto argument = ReadF16AsF32(inst.src0);
+	const auto magnitude =
+	    ir.BitwiseAnd(ir.BitCastU32(argument), IR::U32(IR::Value(0x7fffffffu)));
+	auto result = IR::F32(ir.Emit(opcode, {argument}));
+	if (opcode == IR::ValueOpcode::FPSin) {
+		const auto fraction = IR::F32(ir.Emit(IR::ValueOpcode::FPFract32, {argument}));
+		const auto whole = IR::U1(
+		    ir.Emit(IR::ValueOpcode::FPOrdEqual32, {fraction, IR::Value::F32(0.0f)}));
+		const auto half = IR::U1(
+		    ir.Emit(IR::ValueOpcode::FPOrdEqual32, {fraction, IR::Value::F32(0.5f)}));
+		const auto nonzero = ir.INotEqual(magnitude, IR::U32(IR::Value(0u)));
+		const auto cardinal = ir.LogicalOr(ir.LogicalAnd(whole, nonzero), half);
+		result = IR::F32(ir.Emit(
+		    IR::ValueOpcode::SelectF32, {cardinal, IR::Value::F32(0.0f), result}));
+	} else {
+		// The architectural quarter-cycle result is exactly zero. Evaluating cos(pi/2) with a
+		// rounded F32 PI can otherwise produce a half-precision subnormal instead.
+		const auto fraction = IR::F32(ir.Emit(IR::ValueOpcode::FPFract32, {argument}));
+		const auto quarter = IR::U1(
+		    ir.Emit(IR::ValueOpcode::FPOrdEqual32, {fraction, IR::Value::F32(0.25f)}));
+		const auto three_quarters = IR::U1(
+		    ir.Emit(IR::ValueOpcode::FPOrdEqual32, {fraction, IR::Value::F32(0.75f)}));
+		result = IR::F32(
+		    ir.Emit(IR::ValueOpcode::SelectF32,
+		            {ir.LogicalOr(quarter, three_quarters), IR::Value::F32(0.0f), result}));
+	}
+
+	const auto infinite = ir.IEqual(magnitude, IR::U32(IR::Value(0x7f800000u)));
+	result              = ApplyF32ResultModifiers(inst.dst, result);
+	const auto bits    = PackHalf2x16(result, IR::F32(IR::Value::F32(0.0f)));
+	const auto invalid = IR::U32(IR::Value(inst.dst.clamp ? 0u : 0xfe00u));
+	WriteU16(DestinationOperand(inst), ir.Select(infinite, invalid, bits));
+	return true;
+}
+
 bool Translator::Float16Binary(const Decoder::Instruction& inst, IR::ValueOpcode opcode,
                                bool reverse) {
 	const auto lhs = ReadF16AsF32(reverse ? inst.src1 : inst.src0);

src/graphics/shader/recompiler/frontend/translate/Integer.cpp

@@ -224,11 +224,14 @@ bool Translator::S_FF1_I32_B64(const Decoder::Instruction& inst) {
 	return true;
 }
 
-bool Translator::V_FFBH_U32(const Decoder::Instruction& inst) {
-	const auto source   = ReadU32(inst.src0);
-	const auto msb      = IR::U32(ir.Emit(IR::ValueOpcode::FindUMsb32, {source}));
+bool Translator::V_FFBH_32(const Decoder::Instruction& inst, bool sign) {
+	const auto source = ReadU32(inst.src0);
+	const auto value  = sign ? ir.BitwiseXor(
+	                              source, ir.ShiftRightArithmetic(source, IR::U32(IR::Value(31u))))
+	                         : source;
+	const auto msb      = IR::U32(ir.Emit(IR::ValueOpcode::FindUMsb32, {value}));
 	const auto position = ir.ISub(IR::U32(IR::Value(31u)), msb);
-	const auto result   = ir.Select(ir.INotEqual(source, IR::U32(IR::Value(0u))), position,
+	const auto result   = ir.Select(ir.INotEqual(value, IR::U32(IR::Value(0u))), position,
 	                                IR::U32(IR::Value(0xffffffffu)));
 	WriteOperand(DestinationOperand(inst), result);
 	return true;

src/graphics/shader/recompiler/frontend/translate/Memory.cpp

@@ -1,9 +1,7 @@
-#include "common/magicEnum.h"
 #include "graphics/shader/recompiler/frontend/translate/Translator.h"
 
 #include <algorithm>
 #include <array>
-#include <fmt/format.h>
 
 namespace Libs::Graphics::ShaderRecompiler::Frontend {
 
@@ -37,6 +35,7 @@ Decoder::Operand OffsetDecodedRegister(const Decoder::Operand& operand, uint32_t
 	result.dpp_ctrl           = 0;
 	result.dpp_row_mask       = 0xf;
 	result.dpp_bank_mask      = 0xf;
+	result.explicit_sdwa_dst  = false;
 	result.dpp_fetch_inactive = false;
 	result.dpp_bound_ctrl     = false;
 	result.dpp                = false;
@@ -257,6 +256,7 @@ Decoder::Operand MemorySourceAt(const Decoder::Instruction& decoded, uint32_t in
 				return index == 0u ? decoded.src1 : index == 1u ? decoded.src0 : decoded.src2;
 			case Decoder::Opcode::DS_WRITE_B8:
 			case Decoder::Opcode::DS_WRITE_B16:
+			case Decoder::Opcode::DS_WRITE_B16_D16_HI:
 			case Decoder::Opcode::DS_WRITE_B32:
 			case Decoder::Opcode::DS_WRITE_B64:
 			case Decoder::Opcode::DS_WRITE_B96:
@@ -877,7 +877,14 @@ bool Translator::DS_SWIZZLE_B32(const Decoder::Instruction& inst) {
 	return true;
 }
 
-bool Translator::EmitMemory(const Decoder::Instruction& inst, std::string* error) {
+bool Translator::DS_BPERMUTE_B32(const Decoder::Instruction& inst) {
+	const auto address = ir.IAdd(ReadU32(inst.src0), IR::U32(IR::Value(inst.offset)));
+	WriteOperand(inst.dst, ir.Emit(IR::ValueOpcode::BpermuteU32,
+	                               {ReadU32(inst.src1), address, ir.GetExec()}));
+	return true;
+}
+
+bool Translator::EmitMemory(const Decoder::Instruction& inst) {
 	switch (inst.opcode) {
 		case Decoder::Opcode::S_LOAD_DWORD:
 		case Decoder::Opcode::S_LOAD_DWORDX2:
@@ -1035,6 +1042,7 @@ bool Translator::EmitMemory(const Decoder::Instruction& inst, std::string* error
 		case Decoder::Opcode::DS_MAX_F32:
 			return DS_MINMAX_F32(inst, IR::ValueOpcode::SharedAtomicFMax32);
 		case Decoder::Opcode::DS_SWIZZLE_B32: return DS_SWIZZLE_B32(inst);
+		case Decoder::Opcode::DS_BPERMUTE_B32: return DS_BPERMUTE_B32(inst);
 		case Decoder::Opcode::DS_CONSUME:
 			return DS_APPEND_CONSUME(inst, IR::ValueOpcode::DataConsume);
 		case Decoder::Opcode::DS_APPEND:
@@ -1050,6 +1058,7 @@ bool Translator::EmitMemory(const Decoder::Instruction& inst, std::string* error
 		case Decoder::Opcode::DS_READ_I16:
 		case Decoder::Opcode::DS_READ_U16:
 		case Decoder::Opcode::DS_READ_U16_D16:
+		case Decoder::Opcode::DS_READ_U16_D16_HI:
 		case Decoder::Opcode::DS_READ_B32:
 		case Decoder::Opcode::DS_READ_B64:
 		case Decoder::Opcode::DS_READ_B96:
@@ -1060,16 +1069,12 @@ bool Translator::EmitMemory(const Decoder::Instruction& inst, std::string* error
 		case Decoder::Opcode::DS_WRITE2ST64_B64: return DS_WRITE2(inst);
 		case Decoder::Opcode::DS_WRITE_B8:
 		case Decoder::Opcode::DS_WRITE_B16:
+		case Decoder::Opcode::DS_WRITE_B16_D16_HI:
 		case Decoder::Opcode::DS_WRITE_B32:
 		case Decoder::Opcode::DS_WRITE_B64:
 		case Decoder::Opcode::DS_WRITE_B96:
 		case Decoder::Opcode::DS_WRITE_B128: return DS_WRITE(inst);
-		default:
-			if (error != nullptr) {
-				*error = fmt::format("memory-family opcode has no specialized IR translation: {}",
-				                     magic_enum::enum_name(inst.opcode));
-			}
-			return false;
+		default: return false;
 	}
 }
 

src/graphics/shader/recompiler/frontend/translate/Scalar.cpp

@@ -2,14 +2,14 @@
 
 namespace Libs::Graphics::ShaderRecompiler::Frontend {
 
-bool Translator::EmitScalar(const Decoder::Instruction& inst, std::string* error) {
+bool Translator::EmitScalar(const Decoder::Instruction& inst) {
 	using O = Decoder::Opcode;
 	switch (inst.opcode) {
 		case O::S_MOV_B32:
 		case O::S_MOVK_I32: MOV_B32(inst, false); return true;
 		case O::S_MOV_B64: S_MOV_B64(inst); return true;
 		case O::S_WQM_B64: S_WQM_B64(inst); return true;
-		case O::S_GETPC_B64: return S_GETPC_B64(inst, error);
+		case O::S_GETPC_B64: S_GETPC_B64(inst); return true;
 		case O::S_SETPC_B64: return true;
 		case O::S_CSELECT_B32: S_CSELECT_B32(inst); return true;
 		case O::S_CSELECT_B64: S_CSELECT_B64(inst); return true;
@@ -181,6 +181,7 @@ bool Translator::EmitScalar(const Decoder::Instruction& inst, std::string* error
 		case O::S_XNOR_B32:
 			return ComposedIntegerBinary(inst, IR::ValueOpcode::BitwiseXor32, false, true, true);
 		case O::S_FF1_I32_B64: return S_FF1_I32_B64(inst);
+		case O::S_FLBIT_I32_B32: return V_FFBH_32(inst, false);
 		case O::S_FLBIT_I32_B64: return S_FLBIT_I32_B64(inst);
 
 		case O::S_BITSET0_B32: return S_BITSET_B32(inst, false);

src/graphics/shader/recompiler/frontend/translate/Translate.cpp

@@ -1,16 +1,14 @@
+#include "common/assert.h"
 #include "graphics/shader/recompiler/frontend/translate/Translator.h"
 #include "graphics/shader/shader.h"
 
 #include <algorithm>
 #include <array>
-#include <fmt/format.h>
 #include <unordered_map>
 #include <utility>
 
 namespace Libs::Graphics::ShaderRecompiler::Frontend {
 
-using ShaderError::Fail;
-
 Decoder::Operand Translator::SourceAt(const Decoder::Instruction& inst, uint32_t index) {
 	switch (index) {
 		case 0: return inst.src0;
@@ -104,6 +102,7 @@ Decoder::Operand Translator::PlainOperand(const Decoder::Operand& operand) {
 	result.dpp_ctrl           = 0;
 	result.dpp_row_mask       = 0xf;
 	result.dpp_bank_mask      = 0xf;
+	result.explicit_sdwa_dst  = false;
 	result.dpp_fetch_inactive = false;
 	result.dpp_bound_ctrl     = false;
 	result.dpp                = false;
@@ -392,12 +391,12 @@ void Translator::WriteOperand(const Decoder::Operand& operand, IR::Value value)
 		}
 	}
 	if (type == IR::Type::U16) {
-		WriteRawU32(operand, IR::U32(ir.Emit(IR::ValueOpcode::ConvertU32U16, {IR::U16(value)})));
+		Write16Bits(operand, IR::U32(ir.Emit(IR::ValueOpcode::ConvertU32U16, {IR::U16(value)})));
 		return;
 	}
 	if (type == IR::Type::F16) {
 		const auto bits = IR::U16(ir.Emit(IR::ValueOpcode::BitCastU16F16, {value}));
-		WriteRawU32(operand, IR::U32(ir.Emit(IR::ValueOpcode::ConvertU32U16, {bits})));
+		Write16Bits(operand, IR::U32(ir.Emit(IR::ValueOpcode::ConvertU32U16, {bits})));
 		return;
 	}
 	if (type == IR::Type::U64) {
@@ -417,38 +416,32 @@ IR::U32 Translator::PackHalf2x16(IR::F32 low, IR::F32 high) {
 	return IR::U32(ir.Emit(IR::ValueOpcode::PackHalf2x16, {pair}));
 }
 
+void Translator::Write16Bits(const Decoder::Operand& operand, IR::U32 value) {
+	auto destination = operand;
+	// Native GFX10 16-bit results preserve the unselected half. Plain/DPP destinations use the
+	// low half; native VOP3 may already select either half. Explicit SDWA retains encoded DST_U.
+	if (!destination.explicit_sdwa_dst) {
+		if (destination.sdwa_sel == 6u) {
+			destination.sdwa_sel = 4u;
+		}
+		destination.sdwa_dst_unused = 2u;
+	}
+	destination.omod   = 0u;
+	destination.op_sel = false;
+	destination.clamp  = false;
+	WriteRawU32(destination, ir.BitwiseAnd(value, IR::U32(IR::Value(0xffffu))));
+}
+
 void Translator::WriteF16(const Decoder::Operand& operand, IR::F32 value) {
 	value           = ApplyF32ResultModifiers(operand, value);
 	const auto half = IR::F16(ir.Emit(IR::ValueOpcode::ConvertF16F32, {value}));
 	const auto bits = IR::U32(ir.Emit(IR::ValueOpcode::ConvertU32U16,
 	                                  {IR::U16(ir.Emit(IR::ValueOpcode::BitCastU16F16, {half}))}));
-	auto       merged = bits;
-	if (operand.sdwa_sel == 4u || operand.sdwa_sel == 5u) {
-		merged = IR::U32(ir.Emit(IR::ValueOpcode::BitFieldInsert,
-		                         {ReadRawU32(PlainOperand(operand)), bits,
-		                          IR::Value(operand.sdwa_sel == 5u ? 16u : 0u), IR::Value(16u)}));
-	}
-	auto raw     = operand;
-	raw.sdwa_sel = 6u;
-	raw.omod     = 0u;
-	raw.op_sel   = false;
-	raw.clamp    = false;
-	WriteRawU32(raw, merged);
+	Write16Bits(operand, bits);
 }
 
 void Translator::WriteU16(const Decoder::Operand& operand, IR::U32 value) {
-	auto merged = ir.BitwiseAnd(value, IR::U32(IR::Value(0xffffu)));
-	if (operand.sdwa_sel == 4u || operand.sdwa_sel == 5u) {
-		merged = IR::U32(ir.Emit(IR::ValueOpcode::BitFieldInsert,
-		                         {ReadRawU32(PlainOperand(operand)), merged,
-		                          IR::Value(operand.sdwa_sel == 5u ? 16u : 0u), IR::Value(16u)}));
-	}
-	auto raw     = operand;
-	raw.sdwa_sel = 6u;
-	raw.omod     = 0u;
-	raw.op_sel   = false;
-	raw.clamp    = false;
-	WriteRawU32(raw, merged);
+	Write16Bits(operand, value);
 }
 
 IR::U32 Translator::ReadU32(const Decoder::Operand& operand) {
@@ -745,11 +738,10 @@ void Translator::WriteCompareResult(const Decoder::Operand& operand, IR::U1 valu
 	WriteMask(operand, ir.LogicalAnd(ir.GetExec(), value));
 }
 
-bool Translator::AddBranchCondition(const CFG::BasicBlock& source, IR::BlockInfo& info,
-                                    std::string* error) {
+void Translator::AddBranchCondition(const CFG::BasicBlock& source, IR::BlockInfo& info) {
 	if (source.terminator.goto_value >= 0) {
 		if (source.terminator.goto_variable == UINT32_MAX) {
-			return Fail(error, fmt::format("block {} sets an invalid goto variable", source.id));
+			EXIT("block %u sets an invalid goto variable", source.id);
 		}
 		ir.SetGotoVariable(source.terminator.goto_variable,
 		                   IR::U1(IR::Value(source.terminator.goto_value != 0)));
@@ -761,13 +753,13 @@ bool Translator::AddBranchCondition(const CFG::BasicBlock& source, IR::BlockInfo
 			info.indirect_target =
 			    ir.GetScalarReg(static_cast<IR::ScalarReg>(source.terminator.indirect_pc_sgpr));
 		} else {
-			return Fail(error, fmt::format("block {} has no indirect branch selector", source.id));
+			EXIT("block %u has no indirect branch selector", source.id);
 		}
 		ir.Emit(IR::ValueOpcode::ReferenceU32, {info.indirect_target});
-		return true;
+		return;
 	}
 	if (source.terminator.kind != CFG::TerminatorKind::ConditionalBranch) {
-		return true;
+		return;
 	}
 	// EXEC and VCC are invocation-local Boolean masks. Branching on that Boolean lets inactive
 	// invocations leave the region without reconstructing a host-subgroup mask.
@@ -782,17 +774,15 @@ bool Translator::AddBranchCondition(const CFG::BasicBlock& source, IR::BlockInfo
 		case CFG::BranchCondition::ExecNonZero: condition = ir.GetExec(); break;
 		case CFG::BranchCondition::GotoVariable:
 			if (source.terminator.goto_variable == UINT32_MAX) {
-				return Fail(error,
-				            fmt::format("block {} reads an invalid goto variable", source.id));
+				EXIT("block %u reads an invalid goto variable", source.id);
 			}
 			condition = ir.GetGotoVariable(source.terminator.goto_variable);
 			break;
 		case CFG::BranchCondition::Unknown:
-			return Fail(error, fmt::format("block {} has an unknown branch condition", source.id));
+			EXIT("block %u has an unknown branch condition", source.id);
 	}
 	info.condition = condition;
 	ir.Emit(IR::ValueOpcode::Reference, {condition});
-	return true;
 }
 
 namespace {
@@ -914,47 +904,42 @@ void IncludeInstructionVectorRegisters(const Decoder::Instruction& inst, uint32_
 	}
 }
 
-bool ValidateTranslateOptions(const TranslateOptions& options, std::string* error) {
+void ValidateTranslateOptions(const TranslateOptions& options) {
 	if (options.wave_size != 32u && options.wave_size != 64u) {
-		return Fail(error, "shader translation requires wave32 or wave64");
+		EXIT("shader translation requires wave32 or wave64, got %u", options.wave_size);
 	}
 	switch (options.stage) {
 		case ShaderType::Vertex:
-			return options.vertex != nullptr ||
-			       Fail(error, "vertex shader translation has no vertex input metadata");
+			if (options.vertex == nullptr) {
+				EXIT("vertex shader translation has no vertex input metadata");
+			}
+			return;
 		case ShaderType::Pixel:
-			return options.pixel != nullptr ||
-			       Fail(error, "pixel shader translation has no pixel input metadata");
+			if (options.pixel == nullptr) {
+				EXIT("pixel shader translation has no pixel input metadata");
+			}
+			return;
 		case ShaderType::Compute:
-			return options.compute != nullptr ||
-			       Fail(error, "compute shader translation has no compute input metadata");
-		default: return Fail(error, "shader translation has an unsupported stage");
+			if (options.compute == nullptr) {
+				EXIT("compute shader translation has no compute input metadata");
+			}
+			return;
+		default:
+			EXIT("shader translation has unsupported stage %u",
+			     static_cast<uint32_t>(options.stage));
 	}
 }
 
 } // namespace
 
-bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
-                      const TranslateOptions& options, IR::Program& result, std::string* error) {
-	result = {};
-	struct ResetOnFailure {
-		IR::Program& program;
-		bool         complete = false;
-
-		~ResetOnFailure() {
-			if (!complete) {
-				program = {};
-			}
-		}
-	} reset_on_failure {result};
-
-	if (!ValidateTranslateOptions(options, error)) {
-		return false;
-	}
+IR::Program TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
+                             const TranslateOptions& options) {
+	ValidateTranslateOptions(options);
 	if (cfg.blocks.empty()) {
-		return Fail(error, "cannot translate an empty CFG");
+		EXIT("cannot translate an empty CFG");
 	}
 
+	IR::Program result;
 	result.stage               = options.stage;
 	result.wave_size           = options.wave_size;
 	result.shader_hash         = options.shader_hash;
@@ -972,8 +957,8 @@ bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
 	for (const auto& cfg_block: cfg.blocks) {
 		for (uint32_t index = cfg_block.inst_begin; index < cfg_block.inst_end; index++) {
 			if (index >= decoded.instructions.size()) {
-				return Fail(error,
-				            "CFG block references an instruction outside the decoded program");
+				EXIT("CFG block %u references instruction %u outside decoded program of size %zu",
+				     cfg_block.id, index, decoded.instructions.size());
 			}
 			const auto& instruction = decoded.instructions[index];
 			if (IsCodeTableLoad(cfg, instruction.pc)) {
@@ -990,7 +975,7 @@ bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
 		                    return block.id;
 	                    })->id;
 	if (max_id == UINT32_MAX) {
-		return Fail(error, "cannot allocate a typed entry block id");
+		EXIT("cannot allocate a typed entry block id");
 	}
 	CFG::Terminator terminator;
 	terminator.kind       = CFG::TerminatorKind::Branch;
@@ -1004,7 +989,7 @@ bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
 	block_indices.reserve(cfg.blocks.size());
 	for (const auto& source_block: cfg.blocks) {
 		if (!block_indices.emplace(source_block.id, result.blocks.size()).second) {
-			return Fail(error, fmt::format("CFG contains duplicate block id {}", source_block.id));
+			EXIT("CFG contains duplicate block id %u", source_block.id);
 		}
 		result.block_storage.push_back(std::make_unique<IR::Block>());
 		result.blocks.push_back(result.block_storage.back().get());
@@ -1016,8 +1001,7 @@ bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
 		for (const auto successor: source_block.successors) {
 			const auto target = block_indices.find(successor);
 			if (target == block_indices.end()) {
-				return Fail(error, fmt::format("CFG block {} has unknown successor {}",
-				                               source_block.id, successor));
+				EXIT("CFG block %u has unknown successor %u", source_block.id, successor);
 			}
 			result.blocks[source_index]->AddBranch(result.blocks[target->second]);
 		}
@@ -1115,7 +1099,7 @@ bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
 		}
 	}
 	for (const auto& cfg_block: cfg.blocks) {
-		const auto         typed_index = block_indices.at(cfg_block.id);
+		const auto typed_index = block_indices.at(cfg_block.id);
 		Translator translator(result, result.blocks[typed_index], vector_limit, options.wave_size);
 		for (uint32_t index = cfg_block.inst_begin; index < cfg_block.inst_end; index++) {
 			const auto& instruction = decoded.instructions[index];
@@ -1131,44 +1115,27 @@ bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
 			    instruction.data_dwords == embedded->components &&
 			    instruction.dst.kind == Decoder::OperandKind::Vgpr) {
 				if (options.vertex == nullptr) {
-					return Fail(error, "embedded vertex fetch plan has no vertex input metadata");
+					EXIT("embedded vertex fetch plan has no vertex input metadata");
 				}
 				const auto resource = ResolveEmbeddedFetchResource(*options.vertex, *embedded);
 				if (resource < 0 || resource >= options.vertex->resources_num) {
-					return Fail(
-					    error,
-					    fmt::format(
-					        "embedded vertex fetch at 0x{:08x} has no resource for attribute {}",
-					        instruction.pc, embedded->attrib_id));
+					EXIT("embedded vertex fetch at 0x%08x has no resource for attribute %d",
+					     instruction.pc, embedded->attrib_id);
 				}
 				result.info.vertex_fetch_components[static_cast<size_t>(resource)] =
 				    static_cast<uint8_t>(std::max<uint32_t>(
 				        result.info.vertex_fetch_components[static_cast<size_t>(resource)],
 				        embedded->components));
-				if (!translator.TranslateEmbeddedFetch(instruction, static_cast<uint32_t>(resource),
-				                                       embedded->components)) {
-					return false;
-				}
+				translator.TranslateEmbeddedFetch(instruction, static_cast<uint32_t>(resource),
+				                                  embedded->components);
 				continue;
 			}
-			if (!translator.TranslateInstruction(instruction, error)) {
-				if (error == nullptr || error->empty()) {
-					return Fail(error, fmt::format("opcode {} at 0x{:08x} has no IR translation",
-					                               magic_enum::enum_name(instruction.opcode),
-					                               instruction.pc));
-				}
-				return false;
-			}
-		}
-		if (!translator.AddBranchCondition(cfg_block, result.block_info[typed_index], error)) {
-			return false;
+			translator.TranslateInstruction(instruction);
 		}
+		translator.AddBranchCondition(cfg_block, result.block_info[typed_index]);
 	}
-	if (!IR::ValidateProgram(result, false, error)) {
-		return false;
-	}
-	reset_on_failure.complete = true;
-	return true;
+	IR::ValidateProgram(result, false);
+	return result;
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::Frontend

src/graphics/shader/recompiler/frontend/translate/Translate.h

@@ -38,7 +38,7 @@ struct TranslateOptions {
 	const EmbeddedFetchPlan*      embedded_fetch = nullptr;
 };
 
-bool TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
-                      const TranslateOptions& options, IR::Program& result, std::string* error);
+IR::Program TranslateProgram(const Decoder::Program& decoded, const CFG::Graph& cfg,
+                             const TranslateOptions& options);
 
 } // namespace Libs::Graphics::ShaderRecompiler::Frontend

src/graphics/shader/recompiler/frontend/translate/Translator.h

@@ -13,10 +13,10 @@ public:
 	    : program(program), ir(block), current_vector_limit(vector_limit),
 	      current_wave_size(wave_size) {}
 
-	bool TranslateInstruction(const Decoder::Instruction& inst, std::string* error);
-	bool TranslateEmbeddedFetch(const Decoder::Instruction& inst, uint32_t attribute,
+	void TranslateInstruction(const Decoder::Instruction& inst);
+	void TranslateEmbeddedFetch(const Decoder::Instruction& inst, uint32_t attribute,
 	                            uint32_t component_count);
-	bool AddBranchCondition(const CFG::BasicBlock& source, IR::BlockInfo& info, std::string* error);
+	void AddBranchCondition(const CFG::BasicBlock& source, IR::BlockInfo& info);
 
 private:
 	Decoder::Operand SourceAt(const Decoder::Instruction& inst, uint32_t index);
@@ -34,6 +34,7 @@ private:
 	IR::F32                ApplyF32ResultModifiers(const Decoder::Operand& operand, IR::F32 value);
 	void                   WriteOperand(const Decoder::Operand& operand, IR::Value value);
 	IR::U32                PackHalf2x16(IR::F32 low, IR::F32 high);
+	void                   Write16Bits(const Decoder::Operand& operand, IR::U32 value);
 	void                   WriteF16(const Decoder::Operand& operand, IR::F32 value);
 	void                   WriteU16(const Decoder::Operand& operand, IR::U32 value);
 	IR::U32                ReadU32(const Decoder::Operand& operand);
@@ -109,6 +110,7 @@ private:
 	bool DS_APPEND_CONSUME(const Decoder::Instruction& inst, IR::ValueOpcode opcode);
 	bool DS_ADDTID(const Decoder::Instruction& inst, bool write);
 	bool DS_SWIZZLE_B32(const Decoder::Instruction& inst);
+	bool DS_BPERMUTE_B32(const Decoder::Instruction& inst);
 
 	IR::F32 SelectF32(IR::U1 condition, IR::F32 true_value, IR::F32 false_value);
 	IR::U32 ConvertF32ToU32Saturated(IR::F32 value, float upper_bound, float safe_upper,
@@ -126,7 +128,7 @@ private:
 	void EmitFloatCompare(const Decoder::Instruction& inst, IR::ValueOpcode opcode, bool half,
 	                      bool cmpx);
 	void EmitFloatOrderedCompare(const Decoder::Instruction& inst, bool ordered);
-	void V_CMP_CLASS_F32(const Decoder::Instruction& inst);
+	void EmitFloatClassCompare(const Decoder::Instruction& inst, bool cmpx);
 	void V_CVT_F32_UBYTE(const Decoder::Instruction& inst, uint32_t byte_index);
 	void V_CVT_F32_U32(const Decoder::Instruction& inst);
 	void V_CVT_F32_I32(const Decoder::Instruction& inst);
@@ -148,6 +150,7 @@ private:
 	                   bool quiet_snan);
 	bool Float16Unary(const Decoder::Instruction& inst, IR::ValueOpcode opcode,
 	                  bool invalid_negative);
+	bool Float16Trig(const Decoder::Instruction& inst, IR::ValueOpcode opcode);
 	bool Float16Binary(const Decoder::Instruction& inst, IR::ValueOpcode opcode, bool reverse);
 	bool Float16Ternary(const Decoder::Instruction& inst, IR::ValueOpcode opcode, bool accumulator,
 	                    bool mix);
@@ -172,28 +175,28 @@ private:
 	bool PackedInteger16MinMax(const Decoder::Instruction& inst, IR::ValueOpcode opcode, bool sign);
 	bool S_U64_MASK(const Decoder::Instruction& inst, IR::ValueOpcode logical_opcode,
 	                IR::ValueOpcode bit_opcode, bool negate_rhs, bool negate_result, bool unary);
-	IR::U1  U64MaskBinary(const Decoder::Instruction& inst, IR::ValueOpcode opcode, bool negate_rhs,
-	                      bool negate_result);
-	bool    SimpleInteger(const Decoder::Instruction& inst, IR::ValueOpcode opcode, IR::Type type,
-	                      bool reverse, bool mask_shift_count, bool update_scc);
-	bool    ComposedIntegerBinary(const Decoder::Instruction& inst, IR::ValueOpcode opcode,
-	                              bool negate_rhs, bool negate_result, bool update_scc);
-	bool    V_AND_OR_B32(const Decoder::Instruction& inst);
-	bool    V_OR3_B32(const Decoder::Instruction& inst);
-	bool    V_XOR3_B32(const Decoder::Instruction& inst);
-	bool    S_FF1_I32_B64(const Decoder::Instruction& inst);
-	bool    V_FFBH_U32(const Decoder::Instruction& inst);
-	bool    S_FLBIT_I32_B64(const Decoder::Instruction& inst);
-	bool    Integer24(const Decoder::Instruction& inst, bool sign, bool addend);
-	bool    V_MAD_U64_U32(const Decoder::Instruction& inst);
-	bool    V_SAD_U32(const Decoder::Instruction& inst);
-	bool    V_ADD3_U32(const Decoder::Instruction& inst);
-	bool    S_BITSET_B32(const Decoder::Instruction& inst, bool set);
-	bool    V_BCNT_U32_B32(const Decoder::Instruction& inst);
-	bool    V_MBCNT_U32_B32(const Decoder::Instruction& inst, bool low);
-	bool    S_BITREPLICATE_B64_B32(const Decoder::Instruction& inst);
-	bool    S_QUADMASK_B64(const Decoder::Instruction& inst);
-	bool    BFM_B32(const Decoder::Instruction& inst);
+	IR::U1 U64MaskBinary(const Decoder::Instruction& inst, IR::ValueOpcode opcode, bool negate_rhs,
+	                     bool negate_result);
+	bool SimpleInteger(const Decoder::Instruction& inst, IR::ValueOpcode opcode, IR::Type type,
+	                   bool reverse, bool mask_shift_count, bool update_scc);
+	bool ComposedIntegerBinary(const Decoder::Instruction& inst, IR::ValueOpcode opcode,
+	                           bool negate_rhs, bool negate_result, bool update_scc);
+	bool V_AND_OR_B32(const Decoder::Instruction& inst);
+	bool V_OR3_B32(const Decoder::Instruction& inst);
+	bool V_XOR3_B32(const Decoder::Instruction& inst);
+	bool S_FF1_I32_B64(const Decoder::Instruction& inst);
+	bool V_FFBH_32(const Decoder::Instruction& inst, bool sign);
+	bool S_FLBIT_I32_B64(const Decoder::Instruction& inst);
+	bool Integer24(const Decoder::Instruction& inst, bool sign, bool addend);
+	bool V_MAD_U64_U32(const Decoder::Instruction& inst);
+	bool V_SAD_U32(const Decoder::Instruction& inst);
+	bool V_ADD3_U32(const Decoder::Instruction& inst);
+	bool S_BITSET_B32(const Decoder::Instruction& inst, bool set);
+	bool V_BCNT_U32_B32(const Decoder::Instruction& inst);
+	bool V_MBCNT_U32_B32(const Decoder::Instruction& inst, bool low);
+	bool S_BITREPLICATE_B64_B32(const Decoder::Instruction& inst);
+	bool S_QUADMASK_B64(const Decoder::Instruction& inst);
+	bool BFM_B32(const Decoder::Instruction& inst);
 	IR::U32 RightMask32(IR::U32 count);
 	IR::U64 RightMask64(IR::U32 count);
 	bool    S_BFM_B64(const Decoder::Instruction& inst);
@@ -226,27 +229,27 @@ private:
 	void S_SENDMSG();
 	void S_TTRACEDATA();
 	void S_INST_PREFETCH();
-	bool S_GETPC_B64(const Decoder::Instruction& inst, std::string* error);
+	void S_GETPC_B64(const Decoder::Instruction& inst);
 	void S_CSELECT_B32(const Decoder::Instruction& inst);
 	void S_CSELECT_B64(const Decoder::Instruction& inst);
 	void MOV_B32(const Decoder::Instruction& inst, bool apply_float_modifiers);
 	void S_MOV_B64(const Decoder::Instruction& inst);
 	void S_WQM_B64(const Decoder::Instruction& inst);
-	bool V_MOVRELS_B32(const Decoder::Instruction& inst, std::string* error);
-	bool V_MOVRELD_B32(const Decoder::Instruction& inst, std::string* error);
+	void V_MOVRELS_B32(const Decoder::Instruction& inst);
+	void V_MOVRELD_B32(const Decoder::Instruction& inst);
 	void V_READFIRSTLANE_B32(const Decoder::Instruction& inst);
 	void V_READLANE_B32(const Decoder::Instruction& inst);
 	void V_WRITELANE_B32(const Decoder::Instruction& inst);
 	void V_PERMLANE16_B32(const Decoder::Instruction& inst, bool x16);
 	void V_INTERP_P1_F32();
 	void V_INTERP_P2_F32(const Decoder::Instruction& inst);
-	bool V_INTERP_MOV_F32(const Decoder::Instruction& inst, std::string* error);
-	bool EXP(const Decoder::Instruction& inst, std::string* error);
+	void V_INTERP_MOV_F32(const Decoder::Instruction& inst);
+	void EXP(const Decoder::Instruction& inst);
 
-	bool EmitScalar(const Decoder::Instruction& inst, std::string* error);
-	bool EmitVector(const Decoder::Instruction& inst, std::string* error);
-	bool EmitInterpolation(const Decoder::Instruction& inst, std::string* error);
-	bool EmitMemory(const Decoder::Instruction& inst, std::string* error);
+	bool EmitScalar(const Decoder::Instruction& inst);
+	bool EmitVector(const Decoder::Instruction& inst);
+	bool EmitInterpolation(const Decoder::Instruction& inst);
+	bool EmitMemory(const Decoder::Instruction& inst);
 
 	IR::Program&    program;
 	IR::IREmitter   ir;

src/graphics/shader/recompiler/frontend/translate/Vector.cpp

@@ -2,7 +2,7 @@
 
 namespace Libs::Graphics::ShaderRecompiler::Frontend {
 
-bool Translator::EmitVector(const Decoder::Instruction& inst, std::string* error) {
+bool Translator::EmitVector(const Decoder::Instruction& inst) {
 	using O = Decoder::Opcode;
 	switch (inst.opcode) {
 		case O::V_NOP: return true;
@@ -13,8 +13,8 @@ bool Translator::EmitVector(const Decoder::Instruction& inst, std::string* error
 		case O::V_SUBREV_CO_CI_U32: SUBB_U32(inst, true); return true;
 
 		case O::V_MOV_B32: MOV_B32(inst, true); return true;
-		case O::V_MOVRELS_B32: return V_MOVRELS_B32(inst, error);
-		case O::V_MOVRELD_B32: return V_MOVRELD_B32(inst, error);
+		case O::V_MOVRELS_B32: V_MOVRELS_B32(inst); return true;
+		case O::V_MOVRELD_B32: V_MOVRELD_B32(inst); return true;
 		case O::V_READFIRSTLANE_B32: V_READFIRSTLANE_B32(inst); return true;
 		case O::V_READLANE_B32: V_READLANE_B32(inst); return true;
 		case O::V_WRITELANE_B32: V_WRITELANE_B32(inst); return true;
@@ -96,8 +96,12 @@ bool Translator::EmitVector(const Decoder::Instruction& inst, std::string* error
 			EmitIntegerCompare(inst, IR::ValueOpcode::SLessThanEqual32, IR::Type::U32, false, true);
 			return true;
 		case O::V_CMP_EQ_I64:
+		case O::V_CMP_EQ_U64:
 			EmitIntegerCompare(inst, IR::ValueOpcode::IEqual64, IR::Type::U64, false, false);
 			return true;
+		case O::V_CMP_LT_U64:
+			EmitIntegerCompare(inst, IR::ValueOpcode::ULessThan64, IR::Type::U64, false, false);
+			return true;
 		case O::V_CMP_GT_U64:
 			EmitIntegerCompare(inst, IR::ValueOpcode::UGreaterThan64, IR::Type::U64, false, false);
 			return true;
@@ -270,7 +274,8 @@ bool Translator::EmitVector(const Decoder::Instruction& inst, std::string* error
 			return true;
 		case O::V_CMP_O_F32: EmitFloatOrderedCompare(inst, true); return true;
 		case O::V_CMP_U_F32: EmitFloatOrderedCompare(inst, false); return true;
-		case O::V_CMP_CLASS_F32: V_CMP_CLASS_F32(inst); return true;
+		case O::V_CMP_CLASS_F32: EmitFloatClassCompare(inst, false); return true;
+		case O::V_CMPX_CLASS_F32: EmitFloatClassCompare(inst, true); return true;
 
 		case O::V_CVT_F32_UBYTE0: V_CVT_F32_UBYTE(inst, 0); return true;
 		case O::V_CVT_F32_UBYTE1: V_CVT_F32_UBYTE(inst, 1); return true;
@@ -364,6 +369,8 @@ bool Translator::EmitVector(const Decoder::Instruction& inst, std::string* error
 		case O::V_CEIL_F16: return Float16Unary(inst, IR::ValueOpcode::FPCeil32, false);
 		case O::V_TRUNC_F16: return Float16Unary(inst, IR::ValueOpcode::FPTrunc32, false);
 		case O::V_RNDNE_F16: return Float16Unary(inst, IR::ValueOpcode::FPRoundEven32, false);
+		case O::V_SIN_F16: return Float16Trig(inst, IR::ValueOpcode::FPSin);
+		case O::V_COS_F16: return Float16Trig(inst, IR::ValueOpcode::FPCos);
 		case O::V_MIN3_F16: return Float16Ternary(inst, IR::ValueOpcode::FPMinTri32, false, false);
 		case O::V_MAX3_F16: return Float16Ternary(inst, IR::ValueOpcode::FPMaxTri32, false, false);
 		case O::V_MED3_F16: return Float16Ternary(inst, IR::ValueOpcode::FPMedTri32, false, false);
@@ -490,7 +497,8 @@ bool Translator::EmitVector(const Decoder::Instruction& inst, std::string* error
 		case O::V_AND_OR_B32: return V_AND_OR_B32(inst);
 		case O::V_OR3_B32: return V_OR3_B32(inst);
 		case O::V_XOR3_B32: return V_XOR3_B32(inst);
-		case O::V_FFBH_U32: return V_FFBH_U32(inst);
+		case O::V_FFBH_U32: return V_FFBH_32(inst, false);
+		case O::V_FFBH_I32: return V_FFBH_32(inst, true);
 
 		case O::V_MAD_I32_I24: return Integer24(inst, true, true);
 		case O::V_MAD_U32_U24: return Integer24(inst, false, true);

src/graphics/shader/recompiler/ir/Program.cpp

@@ -1,7 +1,6 @@
+#include "common/assert.h"
 #include "graphics/shader/recompiler/ir/ShaderIR.h"
 
-#include "graphics/shader/shader.h"
-
 #include <fmt/format.h>
 #include <map>
 #include <new>
@@ -11,10 +10,13 @@
 
 namespace Libs::Graphics::ShaderRecompiler::IR {
 
-using ShaderError::Fail;
-
 namespace {
 
+[[noreturn]] void Fail(std::string_view message) {
+	EXIT("shader IR validation failed: %s", std::string(message).c_str());
+	std::abort();
+}
+
 bool IsRegisterStatePseudo(ValueOpcode opcode) {
 	switch (opcode) {
 		case ValueOpcode::GetThreadBitScalarRegister:
@@ -149,10 +151,10 @@ Value ResolveInvariantPhi(const Program& program, Value value) {
 	return invariant;
 }
 
-bool ValidateProgram(const Program& program, bool require_ssa, std::string* error) {
+void ValidateProgram(const Program& program, bool require_ssa) {
 	if (program.blocks.size() != program.block_info.size() ||
 	    program.blocks.size() != program.block_storage.size()) {
-		return Fail(error, "value IR block storage is inconsistent");
+		return Fail("value IR block storage is inconsistent");
 	}
 	std::unordered_set<const Block*>           blocks;
 	std::unordered_map<const Block*, size_t>   block_indices;
@@ -162,26 +164,26 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 		const auto* block = program.blocks[block_index];
 		if (block == nullptr || program.block_storage[block_index] == nullptr ||
 		    block != program.block_storage[block_index].get()) {
-			return Fail(error, "value IR block pointer is inconsistent");
+			return Fail("value IR block pointer is inconsistent");
 		}
 		if (!blocks.insert(block).second) {
-			return Fail(error, "value IR block pointer is duplicated");
+			return Fail("value IR block pointer is duplicated");
 		}
 		block_indices.emplace(block, block_index);
 		if (program.block_info[block_index].id == UINT32_MAX) {
-			return Fail(error, "value IR block uses the reserved exit id");
+			return Fail("value IR block uses the reserved exit id");
 		}
 		if (!blocks_by_id.emplace(program.block_info[block_index].id, block).second) {
-			return Fail(error, "value IR block id is duplicated");
+			return Fail("value IR block id is duplicated");
 		}
 		for (const auto& inst: *block) {
 			if (!instructions.insert(&inst).second) {
-				return Fail(error, "value IR instruction is duplicated");
+				return Fail("value IR instruction is duplicated");
 			}
 		}
 	}
 	if (!program.blocks.empty() && !program.blocks.front()->ImmPredecessors().empty()) {
-		return Fail(error, "value IR entry block has a predecessor");
+		return Fail("value IR entry block has a predecessor");
 	}
 
 	for (size_t block_index = 0; block_index < program.blocks.size(); block_index++) {
@@ -189,28 +191,28 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 		std::unordered_set<const Block*> predecessors;
 		for (const auto* predecessor: block->ImmPredecessors()) {
 			if (predecessor == nullptr || !blocks.contains(predecessor)) {
-				return Fail(error, "value IR block has a foreign predecessor");
+				return Fail("value IR block has a foreign predecessor");
 			}
 			if (!predecessors.insert(predecessor).second) {
-				return Fail(error, "value IR block predecessor is duplicated");
+				return Fail("value IR block predecessor is duplicated");
 			}
 			if (std::ranges::find(predecessor->ImmSuccessors(), block) ==
 			    predecessor->ImmSuccessors().end()) {
-				return Fail(error, "value IR predecessor edge is not reciprocal");
+				return Fail("value IR predecessor edge is not reciprocal");
 			}
 		}
 
 		std::unordered_set<const Block*> successors;
 		for (const auto* successor: block->ImmSuccessors()) {
 			if (successor == nullptr || !blocks.contains(successor)) {
-				return Fail(error, "value IR block has a foreign successor");
+				return Fail("value IR block has a foreign successor");
 			}
 			if (!successors.insert(successor).second) {
-				return Fail(error, "value IR block successor is duplicated");
+				return Fail("value IR block successor is duplicated");
 			}
 			if (std::ranges::find(successor->ImmPredecessors(), block) ==
 			    successor->ImmPredecessors().end()) {
-				return Fail(error, "value IR successor edge is not reciprocal");
+				return Fail("value IR successor edge is not reciprocal");
 			}
 		}
 
@@ -234,41 +236,40 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 		switch (terminator.kind) {
 			case CFG::TerminatorKind::Branch:
 				if (!add_target(terminator.true_block)) {
-					return Fail(error, "value IR branch target is missing");
+					return Fail("value IR branch target is missing");
 				}
 				break;
 			case CFG::TerminatorKind::ConditionalBranch:
 				if (!add_target(terminator.true_block) || !add_target(terminator.false_block)) {
-					return Fail(error, "value IR conditional branch target is missing");
+					return Fail("value IR conditional branch target is missing");
 				}
 				if (!validate_control_value(program.block_info[block_index].condition, Type::U1)) {
-					return Fail(error, "value IR conditional branch condition is invalid");
+					return Fail("value IR conditional branch condition is invalid");
 				}
 				break;
 			case CFG::TerminatorKind::IndirectBranch: {
 				if (!validate_control_value(program.block_info[block_index].indirect_target,
 				                            Type::U32)) {
-					return Fail(error, "value IR indirect branch selector is invalid");
+					return Fail("value IR indirect branch selector is invalid");
 				}
 				std::unordered_set<uint32_t> indirect_targets;
 				for (const auto target: terminator.indirect_targets) {
 					if (!indirect_targets.insert(target).second) {
-						return Fail(error, "value IR indirect branch target is duplicated");
+						return Fail("value IR indirect branch target is duplicated");
 					}
 					if (!add_target(target)) {
-						return Fail(error, "value IR indirect branch target is missing");
+						return Fail("value IR indirect branch target is missing");
 					}
 				}
 				if (terminator.indirect_selector_values.size() !=
 				    terminator.indirect_selector_targets.size()) {
-					return Fail(error, "value IR indirect selector table is inconsistent");
+					return Fail("value IR indirect selector table is inconsistent");
 				}
 				for (const auto target: terminator.indirect_selector_targets) {
 					const auto found = blocks_by_id.find(target);
 					if (found == blocks_by_id.end() ||
 					    !expected_successors.contains(found->second)) {
-						return Fail(error,
-						            "value IR indirect selector target is not a CFG successor");
+						return Fail("value IR indirect selector target is not a CFG successor");
 					}
 				}
 				break;
@@ -280,73 +281,73 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 		     !blocks_by_id.contains(terminator.merge_block)) ||
 		    (terminator.continue_block != UINT32_MAX &&
 		     !blocks_by_id.contains(terminator.continue_block))) {
-			return Fail(error, "value IR structured control target is missing");
+			return Fail("value IR structured control target is missing");
 		}
 		if (successors != expected_successors) {
-			return Fail(error, "value IR terminator and successor graph disagree");
+			return Fail("value IR terminator and successor graph disagree");
 		}
 
 		bool saw_non_phi = false;
 		for (const auto& inst: *block) {
 			if (inst.Parent() != block) {
-				return Fail(error, "value IR instruction has the wrong parent block");
+				return Fail("value IR instruction has the wrong parent block");
 			}
 			if (inst.GetOpcode() == ValueOpcode::Phi) {
 				if (saw_non_phi) {
-					return Fail(error, "value IR Phi appears after a non-Phi instruction");
+					return Fail("value IR Phi appears after a non-Phi instruction");
 				}
 				if (inst.NumPhiBlocks() != inst.NumArgs()) {
-					return Fail(error, "value IR Phi parent table is inconsistent");
+					return Fail("value IR Phi parent table is inconsistent");
 				}
 				if (inst.NumArgs() == 0) {
-					return Fail(error, "value IR Phi has no incoming values");
+					return Fail("value IR Phi has no incoming values");
 				}
 				std::unordered_set<const Block*> incoming_blocks;
 				for (size_t arg_index = 0; arg_index < inst.NumArgs(); arg_index++) {
 					const auto* predecessor = inst.PhiBlock(arg_index);
 					if (predecessor == nullptr || !blocks.contains(predecessor) ||
 					    !predecessors.contains(predecessor)) {
-						return Fail(error, "value IR Phi has a foreign or non-predecessor parent");
+						return Fail("value IR Phi has a foreign or non-predecessor parent");
 					}
 					if (!incoming_blocks.insert(predecessor).second) {
-						return Fail(error, "value IR Phi parent is duplicated");
+						return Fail("value IR Phi parent is duplicated");
 					}
 					if (inst.Arg(arg_index).GetType() != inst.GetType()) {
-						return Fail(error, "value IR Phi incoming type does not match its result");
+						return Fail("value IR Phi incoming type does not match its result");
 					}
 				}
 				if (incoming_blocks != predecessors) {
-					return Fail(error, "value IR Phi does not cover every predecessor");
+					return Fail("value IR Phi does not cover every predecessor");
 				}
 			} else {
 				saw_non_phi = true;
 			}
 			if (require_ssa && IsRegisterStatePseudo(inst.GetOpcode())) {
-				return Fail(error, fmt::format("register-state pseudo {} survived SSA rewrite",
-				                               ValueOpcodeName(inst.GetOpcode())));
+				return Fail(fmt::format("register-state pseudo {} survived SSA rewrite",
+				                        ValueOpcodeName(inst.GetOpcode())));
 			}
 			if (inst.GetOpcode() != ValueOpcode::Phi && inst.GetOpcode() != ValueOpcode::Identity &&
 			    inst.GetType() == Type::Opaque) {
-				return Fail(error, fmt::format("untyped opcode {} survived translation",
-				                               ValueOpcodeName(inst.GetOpcode())));
+				return Fail(fmt::format("untyped opcode {} survived translation",
+				                        ValueOpcodeName(inst.GetOpcode())));
 			}
 			const bool fixed_signature =
 			    inst.GetOpcode() != ValueOpcode::Phi && inst.GetOpcode() != ValueOpcode::Identity;
 			if (fixed_signature && inst.NumArgs() != NumArgsOf(inst.GetOpcode())) {
-				return Fail(error, fmt::format("{} has {} arguments, expected {}",
-				                               ValueOpcodeName(inst.GetOpcode()), inst.NumArgs(),
-				                               NumArgsOf(inst.GetOpcode())));
+				return Fail(fmt::format("{} has {} arguments, expected {}",
+				                        ValueOpcodeName(inst.GetOpcode()), inst.NumArgs(),
+				                        NumArgsOf(inst.GetOpcode())));
 			}
 			if (inst.GetOpcode() == ValueOpcode::ReadConstBuffer) {
 				const auto memory_index = inst.Flags<MemoryFlags>().index;
 				if (memory_index >= program.memory_info.size()) {
-					return Fail(error, fmt::format("{} has an invalid memory-info index",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has an invalid memory-info index",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				const auto& memory = program.memory_info[memory_index];
 				if (memory.kind != ResourceKind::ScalarBuffer) {
-					return Fail(error, fmt::format("{} has an invalid scalar-memory resource kind",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has an invalid scalar-memory resource kind",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				const bool valid_group_width =
 				    memory.component_count == 1u || memory.component_count == 2u ||
@@ -354,23 +355,23 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 				    memory.component_count == 16u;
 				if (memory.data_bits != 32u || memory.data_dwords != 1u || !valid_group_width ||
 				    memory.component_index >= memory.component_count) {
-					return Fail(error, fmt::format("{} has inconsistent scalar-memory metadata",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has inconsistent scalar-memory metadata",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 			}
 			const auto address_info = AddressOpcodeInfoOf(inst.GetOpcode());
 			if (address_info.access != AddressAccess::None) {
 				const auto memory_index = inst.Flags<MemoryFlags>().index;
 				if (memory_index >= program.memory_info.size()) {
-					return Fail(error, fmt::format("{} has an invalid memory-info index",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has an invalid memory-info index",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				const auto& memory = program.memory_info[memory_index];
 				if (!IsAddressResourceKind(memory.kind) ||
 				    (memory.kind == ResourceKind::ScalarAddress &&
 				     inst.GetOpcode() != ValueOpcode::LoadAddressU32)) {
-					return Fail(error, fmt::format("{} has an invalid address resource kind",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has an invalid address resource kind",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				const bool scalar_address = memory.kind == ResourceKind::ScalarAddress;
 				const bool valid_group_width =
@@ -382,50 +383,50 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 				if (memory.data_bits != address_info.data_bits || memory.data_dwords != 1u ||
 				    !valid_group_width || memory.component_index >= memory.component_count ||
 				    memory.sampler != 0u) {
-					return Fail(error, fmt::format("{} has inconsistent address-memory metadata",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has inconsistent address-memory metadata",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 			}
 			const auto buffer_components = BufferComponentCount(inst.GetOpcode());
 			if (buffer_components != 0u) {
 				const auto memory_index = inst.Flags<MemoryFlags>().index;
 				if (memory_index >= program.memory_info.size()) {
-					return Fail(error, fmt::format("{} has an invalid memory-info index",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has an invalid memory-info index",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				const auto& memory = program.memory_info[memory_index];
 				if (memory.kind != ResourceKind::Buffer &&
 				    memory.kind != ResourceKind::ScalarBuffer) {
-					return Fail(error, fmt::format("{} has a non-buffer resource kind",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has a non-buffer resource kind",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				if (buffer_components > 1u &&
 				    (memory.kind != ResourceKind::Buffer || memory.data_bits != 32u ||
 				     memory.data_dwords != buffer_components || memory.component_index != 0u)) {
-					return Fail(error, fmt::format("{} has inconsistent native-wide metadata",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has inconsistent native-wide metadata",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				if (buffer_components == 1u &&
 				    (inst.GetOpcode() == ValueOpcode::LoadBufferU32 ||
 				     inst.GetOpcode() == ValueOpcode::StoreBufferU32) &&
 				    memory.data_dwords != 1u) {
-					return Fail(error, fmt::format("{} retains scalar-sibling width metadata",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} retains scalar-sibling width metadata",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 			}
 			const auto shared_components = SharedComponentCount(inst.GetOpcode());
 			if (shared_components != 0u) {
 				const auto memory_index = inst.Flags<MemoryFlags>().index;
 				if (memory_index >= program.memory_info.size()) {
-					return Fail(error, fmt::format("{} has an invalid memory-info index",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has an invalid memory-info index",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				const auto& memory = program.memory_info[memory_index];
 				if ((memory.kind != ResourceKind::Lds && memory.kind != ResourceKind::Gds) ||
 				    memory.resource != 0u || memory.sampler != 0u || memory.component_count == 0u ||
 				    memory.component_index >= memory.component_count) {
-					return Fail(error, fmt::format("{} has invalid shared-memory metadata",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has invalid shared-memory metadata",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				uint32_t expected_bits = 32u;
 				if (inst.GetOpcode() == ValueOpcode::LoadSharedU8 ||
@@ -437,26 +438,26 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 				}
 				if (memory.data_bits != expected_bits || memory.data_dwords != shared_components ||
 				    (shared_components > 1u && memory.component_index != 0u)) {
-					return Fail(error, fmt::format("{} has inconsistent shared-memory width",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has inconsistent shared-memory width",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 			}
 			const auto image_info = ImageOpcodeInfoOf(inst.GetOpcode());
 			if (image_info.access != ImageAccess::None) {
 				const auto memory_index = inst.Flags<MemoryFlags>().index;
 				if (memory_index >= program.memory_info.size()) {
-					return Fail(error, fmt::format("{} has an invalid memory-info index",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has an invalid memory-info index",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 				const auto& memory = program.memory_info[memory_index];
 				if (!ImageResourceKindMatches(memory.kind, image_info.resource_class)) {
-					return Fail(error, fmt::format("{} has invalid image-memory metadata",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(fmt::format("{} has invalid image-memory metadata",
+					                        ValueOpcodeName(inst.GetOpcode())));
 				}
 			}
 			if (inst.GetOpcode() == ValueOpcode::SetAttribute &&
 			    inst.Flags<ExportFlags>().index >= program.export_info.size()) {
-				return Fail(error, "SetAttribute has an invalid export-info index");
+				return Fail("SetAttribute has an invalid export-info index");
 			}
 			uint32_t composite_components = 0u;
 			switch (inst.GetOpcode()) {
@@ -469,25 +470,24 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 			if (composite_components != 0u &&
 			    (!inst.Arg(1).IsImmediate() || inst.Arg(1).GetType() != Type::U32 ||
 			     inst.Arg(1).U32() >= composite_components)) {
-				return Fail(error, fmt::format("{} has an invalid component index",
-				                               ValueOpcodeName(inst.GetOpcode())));
+				return Fail(fmt::format("{} has an invalid component index",
+				                        ValueOpcodeName(inst.GetOpcode())));
 			}
 			for (size_t arg_index = 0; arg_index < inst.NumArgs(); arg_index++) {
 				const auto arg = inst.Arg(arg_index);
 				if (arg.IsEmpty()) {
-					return Fail(error, fmt::format("{} has an empty argument",
-					                               ValueOpcodeName(inst.GetOpcode())));
+					return Fail(
+					    fmt::format("{} has an empty argument", ValueOpcodeName(inst.GetOpcode())));
 				}
 				if (fixed_signature && arg.GetType() != ArgTypeOf(inst.GetOpcode(), arg_index)) {
-					return Fail(error,
-					            fmt::format("{} argument {} has type {}, expected {}",
+					return Fail(fmt::format("{} argument {} has type {}, expected {}",
 					                        ValueOpcodeName(inst.GetOpcode()), arg_index,
 					                        TypeName(arg.GetType()),
 					                        TypeName(ArgTypeOf(inst.GetOpcode(), arg_index))));
 				}
 				if (const auto* definition = arg.TryInstruction();
 				    definition != nullptr && !instructions.contains(definition)) {
-					return Fail(error, "value IR argument has a foreign definition");
+					return Fail("value IR argument has a foreign definition");
 				}
 				if (const auto* definition = arg.TryInstruction(); definition != nullptr) {
 					const auto& uses = definition->Uses();
@@ -495,8 +495,7 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 						return candidate.user == &inst && candidate.operand == arg_index;
 					});
 					if (use == uses.end()) {
-						return Fail(error,
-						            fmt::format("{} argument {} is absent from {} reverse uses",
+						return Fail(fmt::format("{} argument {} is absent from {} reverse uses",
 						                        ValueOpcodeName(inst.GetOpcode()), arg_index,
 						                        ValueOpcodeName(definition->GetOpcode())));
 					}
@@ -506,7 +505,7 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 	}
 
 	if (program.blocks.empty()) {
-		return true;
+		return;
 	}
 
 	std::vector<bool>   reachable(program.blocks.size(), false);
@@ -523,7 +522,7 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 		}
 	}
 	if (!std::ranges::all_of(reachable, [](bool value) { return value; })) {
-		return Fail(error, "value IR contains an unreachable block");
+		return Fail("value IR contains an unreachable block");
 	}
 
 	std::vector<std::vector<bool>> dominators(program.blocks.size(),
@@ -577,28 +576,25 @@ bool ValidateProgram(const Program& program, bool require_ssa, std::string* erro
 					const auto* predecessor = inst.PhiBlock(arg_index);
 					if (definition->Parent() != predecessor &&
 					    !dominates(definition->Parent(), predecessor)) {
-						return Fail(error,
-						            "value IR Phi incoming definition does not dominate its edge");
+						return Fail("value IR Phi incoming definition does not dominate its edge");
 					}
 				} else if (definition->Parent() == block) {
 					if (instruction_positions.at(definition) >= instruction_positions.at(&inst)) {
-						return Fail(error,
-						            "value IR instruction uses a same-block definition before it");
+						return Fail("value IR instruction uses a same-block definition before it");
 					}
 				} else if (!dominates(definition->Parent(), block)) {
-					return Fail(error, "value IR instruction definition does not dominate its use");
+					return Fail("value IR instruction definition does not dominate its use");
 				}
 			}
 		}
 		const auto& info = program.block_info[block_index];
 		if (!info.condition.IsEmpty() && !control_dominates(info.condition, block)) {
-			return Fail(error, "value IR branch condition definition does not dominate its use");
+			return Fail("value IR branch condition definition does not dominate its use");
 		}
 		if (!info.indirect_target.IsEmpty() && !control_dominates(info.indirect_target, block)) {
-			return Fail(error, "value IR indirect selector definition does not dominate its use");
+			return Fail("value IR indirect selector definition does not dominate its use");
 		}
 	}
-	return true;
 }
 
 void ResolveControlFlowIdentities(Program& program) {

src/graphics/shader/recompiler/ir/ShaderIR.h

@@ -477,7 +477,6 @@ struct DescriptorSource {
 struct SrtRead {
 	Value    value;
 	uint32_t flat_offset = 0;
-	uint32_t use_pc      = 0;
 
 	bool operator==(const SrtRead& other) const = default;
 };
@@ -523,7 +522,7 @@ struct Program {
 
 std::string ProgramToString(const Program& program);
 
-bool  ValidateProgram(const Program& program, bool require_ssa, std::string* error);
+void  ValidateProgram(const Program& program, bool require_ssa);
 void  ResolveControlFlowIdentities(Program& program);
 bool  EquivalentValue(const Program& program, Value left, Value right);
 Value ResolveInvariantPhi(const Program& program, Value value);

src/graphics/shader/recompiler/ir/opcodes/ValueOpcodes.inc

@@ -188,6 +188,7 @@ VALUE_OPCODE(ReadFirstLane, U32, U32, U1)
 VALUE_OPCODE(ReadLane, U32, U32, U32)
 VALUE_OPCODE(WriteLane, U32, U32, U32, U32)
 VALUE_OPCODE(Permlane16U32, U32, U32, U32, U32, U1)
+VALUE_OPCODE(BpermuteU32, U32, U32, U32, U1)
 
 // Typed shader resources and memory operations. Descriptor state is represented by SSA values;
 // dense bindings are assigned by resource tracking.

src/graphics/shader/recompiler/ir/passes/BindingLayout.cpp

@@ -1,5 +1,6 @@
 #include "graphics/shader/recompiler/ir/passes/BindingLayout.h"
 
+#include "common/assert.h"
 #include "graphics/shader/recompiler/ir/ShaderIR.h"
 
 #include <algorithm>
@@ -9,12 +10,17 @@
 namespace Libs::Graphics::ShaderRecompiler::IR {
 namespace {
 
-constexpr auto     FirstImageBinding     = DescriptorBindingKind::Sampled1D;
+constexpr auto     FirstImageBinding = DescriptorBindingKind::Sampled1D;
 constexpr uint32_t ImageBindingCount =
     static_cast<uint32_t>(DescriptorBindingKind::StorageAtomic3D) -
     static_cast<uint32_t>(FirstImageBinding) + 1u;
 
-bool CollectUserData(const Program& program, std::vector<uint32_t>& result) {
+[[noreturn]] void BindingFail(const char* message) {
+	EXIT("shader binding layout failed: %s", message);
+	std::abort();
+}
+
+std::vector<uint32_t> CollectUserData(const Program& program) {
 	std::array<bool, NumScalarRegs> registers {};
 	for (const auto* block: program.blocks) {
 		for (const auto& inst: *block) {
@@ -22,22 +28,22 @@ bool CollectUserData(const Program& program, std::vector<uint32_t>& result) {
 				continue;
 			}
 			if (inst.Arg(0).GetType() != Type::ScalarReg) {
-				return false;
+				BindingFail("typed shader contains an invalid user-data register");
 			}
 			const auto index = RegIndex(inst.Arg(0).ScalarRegister());
 			if (index >= NumScalarRegs) {
-				return false;
+				BindingFail("typed shader contains an invalid user-data register");
 			}
 			registers[index] = true;
 		}
 	}
-	result.clear();
+	std::vector<uint32_t> result;
 	for (uint32_t index = 0; index < registers.size(); index++) {
 		if (registers[index]) {
 			result.push_back(index);
 		}
 	}
-	return true;
+	return result;
 }
 
 void AddBinding(BindingLayout& layout, DescriptorBindingKind kind,
@@ -45,8 +51,8 @@ void AddBinding(BindingLayout& layout, DescriptorBindingKind kind,
 	layout.descriptors.push_back({kind, std::move(resources)});
 }
 
-bool UsesGds(const Program& program, bool& uses_gds) {
-	uses_gds = false;
+bool UsesGds(const Program& program) {
+	bool uses_gds = false;
 	for (const auto* block: program.blocks) {
 		for (const auto& inst: *block) {
 			if (SharedAccessOf(inst.GetOpcode()) == SharedAccess::None) {
@@ -54,44 +60,36 @@ bool UsesGds(const Program& program, bool& uses_gds) {
 			}
 			const auto index = inst.Flags<MemoryFlags>().index;
 			if (index >= program.memory_info.size()) {
-				return false;
+				BindingFail("typed shader contains invalid shared-memory metadata");
 			}
 			const auto kind = program.memory_info[index].kind;
 			if (kind != ResourceKind::Lds && kind != ResourceKind::Gds) {
-				return false;
+				BindingFail("typed shader contains invalid shared-memory metadata");
 			}
 			uses_gds |= kind == ResourceKind::Gds;
 		}
 	}
-	return true;
+	return uses_gds;
 }
 
 } // namespace
 
-bool AllocateBindings(Program& program, uint32_t push_constant_offset, std::string* error) {
+void AllocateBindings(Program& program, uint32_t push_constant_offset) {
 	if (!program.shader_info_complete || program.binding_layout_complete) {
-		if (error != nullptr) {
-			*error = !program.shader_info_complete ? "shader info is not ready"
-			                                       : "binding layout already allocated";
-		}
-		return false;
+		EXIT("shader binding layout failed: %s", !program.shader_info_complete
+		                                             ? "shader info is not ready"
+		                                             : "binding layout already allocated");
 	}
 	if (push_constant_offset > NativePushConstantSize ||
 	    push_constant_offset % sizeof(uint32_t) != 0) {
-		if (error != nullptr) {
-			*error = "push-constant offset exceeds the Vulkan minimum guarantee or is unaligned";
-		}
-		return false;
+		EXIT("shader binding layout failed: push-constant offset %u exceeds the Vulkan minimum "
+		     "guarantee or is unaligned",
+		     push_constant_offset);
 	}
 
 	BindingLayout next;
 	next.push_constant_offset = push_constant_offset;
-	if (!CollectUserData(program, next.user_data_registers)) {
-		if (error != nullptr) {
-			*error = "typed shader contains an invalid user-data register";
-		}
-		return false;
-	}
+	next.user_data_registers = CollectUserData(program);
 	next.memory_offset_dword = static_cast<uint32_t>(next.user_data_registers.size());
 	next.memory_offset_count = static_cast<uint32_t>(program.info.buffers.size());
 
@@ -107,26 +105,18 @@ bool AllocateBindings(Program& program, uint32_t push_constant_offset, std::stri
 	for (uint32_t i = 0; i < program.info.images.size(); i++) {
 		const auto kind = DescriptorBindingForImage(program.info.images[i]);
 		if (!kind.has_value()) {
-			if (error != nullptr) {
-				*error = "shader info contains an invalid image binding class";
-			}
-			return false;
+			EXIT("shader binding layout failed: image %u has an invalid binding class", i);
 		}
 		const auto group = static_cast<uint32_t>(*kind) - static_cast<uint32_t>(FirstImageBinding);
 		if (group >= image_groups.size()) {
-			if (error != nullptr) {
-				*error = "shader info contains an unmapped image binding class";
-			}
-			return false;
+			EXIT("shader binding layout failed: image %u has an unmapped binding class", i);
 		}
 		auto&      resources = image_groups[group];
 		const auto dynamic   = program.info.images[i].mip_mode == ImageMipMode::DynamicStorage;
 		const auto count     = dynamic ? program.info.images[i].mip_count : 1u;
 		if (count == 0u || (!dynamic && program.info.images[i].mip_count != 1u)) {
-			if (error != nullptr) {
-				*error = "image has an invalid specialized mip descriptor count";
-			}
-			return false;
+			EXIT("shader binding layout failed: image %u has invalid specialized mip count %u", i,
+			     program.info.images[i].mip_count);
 		}
 		resources.insert(resources.end(), count, i);
 	}
@@ -146,14 +136,7 @@ bool AllocateBindings(Program& program, uint32_t push_constant_offset, std::stri
 		}
 		AddBinding(next, DescriptorBindingKind::Samplers, std::move(resources));
 	}
-	bool uses_gds = false;
-	if (!UsesGds(program, uses_gds)) {
-		if (error != nullptr) {
-			*error = "typed shader contains invalid shared-memory metadata";
-		}
-		return false;
-	}
-	if (uses_gds) {
+	if (UsesGds(program)) {
 		AddBinding(next, DescriptorBindingKind::Gds);
 	}
 	if (program.info.uses_dma) {
@@ -178,7 +161,6 @@ bool AllocateBindings(Program& program, uint32_t push_constant_offset, std::stri
 
 	program.bindings                = std::move(next);
 	program.binding_layout_complete = true;
-	return true;
 }
 
 const DescriptorBinding* FindBinding(const BindingLayout& layout, DescriptorBindingKind kind) {

src/graphics/shader/recompiler/ir/passes/BindingLayout.h

@@ -5,7 +5,7 @@
 
 namespace Libs::Graphics::ShaderRecompiler::IR {
 
-bool AllocateBindings(Program& program, uint32_t push_constant_offset, std::string* error);
+void AllocateBindings(Program& program, uint32_t push_constant_offset);
 
 const DescriptorBinding* FindBinding(const BindingLayout& layout, DescriptorBindingKind kind);
 

src/graphics/shader/recompiler/ir/passes/ResourceMaterialization.cpp

@@ -1,5 +1,6 @@
 #include "graphics/shader/recompiler/ir/passes/ResourceMaterialization.h"
 
+#include "common/assert.h"
 #include "graphics/guest_gpu/gpu_format.h"
 #include "graphics/shader/recompiler/ir/ShaderIR.h"
 #include "graphics/shader/shaderBindings.h"
@@ -18,6 +19,11 @@ namespace {
 constexpr uint64_t AddressMask            = 0x0000ffffffffffffull;
 constexpr uint64_t MaxIndirectImageProbes = 65536u;
 
+[[noreturn]] void SpecializationFail(std::string_view message) {
+	EXIT("shader resource specialization failed: %s", std::string(message).c_str());
+	std::abort();
+}
+
 Decoder::ImageDimension DescriptorDimension(const DescriptorValue&  descriptor,
                                             Decoder::ImageDimension requested) {
 	const bool is_array = requested == Decoder::ImageDimension::Dim1DArray ||
@@ -55,7 +61,8 @@ bool NullImageDescriptor(const DescriptorValue& descriptor) {
 bool ValidImageDescriptor(const DescriptorValue& descriptor, bool r128 = false) {
 	const auto type   = static_cast<Prospero::ImageType>((descriptor.dwords[3] >> 28u) & 0xfu);
 	const auto format = static_cast<Prospero::BufferFormat>((descriptor.dwords[1] >> 20u) & 0x1ffu);
-	if (type < Prospero::ImageType::kColor1D || format == Prospero::BufferFormat::kInvalid) {
+	if (type < Prospero::ImageType::kColor1D || format == Prospero::BufferFormat::kInvalid ||
+	    format > Prospero::BufferFormat::kBc7Srgb) {
 		return false;
 	}
 	if (r128 && type != Prospero::ImageType::kColor1D && type != Prospero::ImageType::kColor2D &&
@@ -79,7 +86,7 @@ uint32_t DescriptorImageSwizzle(const DescriptorValue& descriptor) {
 
 Prospero::BufferFormat ImageConversionFormat(Prospero::BufferFormat format) {
 	return Prospero::RemapTextureFormat(format) != format ? format
-	                                                     : Prospero::BufferFormat::kInvalid;
+	                                                      : Prospero::BufferFormat::kInvalid;
 }
 
 bool DescriptorIsCube(const DescriptorValue& descriptor) {
@@ -153,8 +160,7 @@ bool ReadSpecializationWord(const SrtRuntime& runtime, uint64_t address, uint32_
 }
 
 bool ReadScalarBufferWord(const ShaderBufferResource& descriptor, uint32_t dynamic_offset,
-                          uint32_t immediate_offset, const SrtRuntime& runtime, uint32_t& word,
-                          std::string* error, uint32_t use_pc) {
+                          uint32_t immediate_offset, const SrtRuntime& runtime, uint32_t& word) {
 	const auto byte_offset = static_cast<uint64_t>(dynamic_offset) + immediate_offset;
 	const auto aligned     = byte_offset & ~uint64_t {3};
 	const auto size        = ScalarBufferSize(descriptor);
@@ -164,20 +170,10 @@ bool ReadScalarBufferWord(const ShaderBufferResource& descriptor, uint32_t dynam
 	}
 	const auto base = descriptor.Base48() & ~uint64_t {3};
 	if (aligned > AddressMask - base) {
-		if (error != nullptr) {
-			*error = fmt::format("indirect image scalar read at pc 0x{:08x} exceeds the "
-			                     "48-bit address space",
-			                     use_pc);
-		}
 		return false;
 	}
 	const auto address = base + aligned;
 	if (!ReadSpecializationWord(runtime, address, word)) {
-		if (error != nullptr) {
-			*error = fmt::format("indirect image scalar read at pc 0x{:08x} failed at "
-			                     "0x{:016x}",
-			                     use_pc, address);
-		}
 		return false;
 	}
 	return true;
@@ -186,23 +182,14 @@ bool ReadScalarBufferWord(const ShaderBufferResource& descriptor, uint32_t dynam
 bool MaterializeIndirectImage(const DescriptorSource::IndirectImage& indirect,
                               const DescriptorValue&                 material_value,
                               const DescriptorValue& heap_value, const SrtRuntime& runtime,
-                              ResourceSnapshot::IndirectImage& result, std::string* error,
-                              uint32_t use_pc) {
+                              ResourceSnapshot::IndirectImage& result) {
 	ShaderBufferResource material;
 	ShaderBufferResource heap;
 	if (!DecodeBufferDescriptor(material_value, material) ||
 	    !DecodeBufferDescriptor(heap_value, heap)) {
-		if (error != nullptr) {
-			*error = fmt::format("indirect image tables at pc 0x{:08x} have invalid descriptors",
-			                     use_pc);
-		}
 		return false;
 	}
 	if (material.Stride() != indirect.selector_stride) {
-		if (error != nullptr) {
-			*error =
-			    fmt::format("indirect image material table at pc 0x{:08x} changed stride", use_pc);
-		}
 		return false;
 	}
 
@@ -215,11 +202,6 @@ bool MaterializeIndirectImage(const DescriptorSource::IndirectImage& indirect,
 	const auto limit       = std::min<uint64_t>(UINT32_MAX, size + 3u);
 	const auto probe_count = residue <= limit ? (limit - residue) / step + 1u : 0u;
 	if (probe_count > MaxIndirectImageProbes) {
-		if (error != nullptr) {
-			*error = fmt::format("indirect image material table at pc 0x{:08x} requires {} "
-			                     "bounded probes",
-			                     use_pc, probe_count);
-		}
 		return false;
 	}
 
@@ -227,8 +209,7 @@ bool MaterializeIndirectImage(const DescriptorSource::IndirectImage& indirect,
 	std::unordered_set<uint32_t> seen {0u};
 	for (uint64_t offset = residue; offset <= limit && probe_count != 0u; offset += step) {
 		uint32_t key = 0;
-		if (!ReadScalarBufferWord(material, static_cast<uint32_t>(offset), 0u, runtime, key, error,
-		                          use_pc)) {
+		if (!ReadScalarBufferWord(material, static_cast<uint32_t>(offset), 0u, runtime, key)) {
 			return false;
 		}
 		if (seen.insert(key).second) {
@@ -248,7 +229,7 @@ bool MaterializeIndirectImage(const DescriptorSource::IndirectImage& indirect,
 		const auto heap_offset = key << 5u;
 		for (uint32_t dword = 0; dword < candidate.dword_count; dword++) {
 			if (!ReadScalarBufferWord(heap, heap_offset, dword * sizeof(uint32_t), runtime,
-			                          candidate.dwords[dword], error, use_pc)) {
+			                          candidate.dwords[dword])) {
 				return false;
 			}
 		}
@@ -281,35 +262,22 @@ size_t FlattenedRuntimeDwords(const Program& program) {
 
 } // namespace
 
-bool ValidateResourceSnapshot(const Program& program, const ResourceSnapshot& snapshot,
-                              std::string* error) {
+bool ValidateResourceSnapshot(const Program& program, const ResourceSnapshot& snapshot) {
 	if (!program.resource_tracking_complete) {
-		if (error != nullptr) {
-			*error = "shader resources were not tracked";
-		}
 		return false;
 	}
 	if (snapshot.buffers.size() != program.info.buffers.size() ||
 	    snapshot.images.size() != program.info.images.size() ||
 	    snapshot.samplers.size() != program.info.samplers.size()) {
-		if (error != nullptr) {
-			*error = "resource snapshot does not match dense shader topology";
-		}
 		return false;
 	}
 	if (snapshot.flattened_srt.size() != FlattenedRuntimeDwords(program)) {
-		if (error != nullptr) {
-			*error = "flattened SRT snapshot does not match the shader plan";
-		}
 		return false;
 	}
 	if (program.binding_layout_complete) {
 		for (const auto reg: program.bindings.user_data_registers) {
 			if (reg < program.user_data_base ||
 			    reg - program.user_data_base >= snapshot.user_data.size()) {
-				if (error != nullptr) {
-					*error = fmt::format("runtime snapshot is missing user SGPR {}", reg);
-				}
 				return false;
 			}
 		}
@@ -317,19 +285,12 @@ bool ValidateResourceSnapshot(const Program& program, const ResourceSnapshot& sn
 	for (uint32_t i = 0; i < program.info.buffers.size(); i++) {
 		const auto alias = program.info.buffers[i].image_alias;
 		if (alias != BufferResource::NoImageAlias && alias >= program.info.images.size()) {
-			if (error != nullptr) {
-				*error = fmt::format("buffer resource {} has invalid image alias {}", i, alias);
-			}
 			return false;
 		}
 	}
-	const auto CheckWidth = [&](const auto& values, uint32_t width, const char* kind) {
+	const auto CheckWidth = [&](const auto& values, uint32_t width) {
 		for (uint32_t i = 0; i < values.size(); i++) {
 			if (values[i].dword_count != width) {
-				if (error != nullptr) {
-					*error = fmt::format("{} descriptor {} has {} dwords", kind, i,
-					                     values[i].dword_count);
-				}
 				return false;
 			}
 		}
@@ -345,56 +306,37 @@ bool ValidateResourceSnapshot(const Program& program, const ResourceSnapshot& sn
 		    table.keys.empty() || table.keys.size() != table.candidates.size() ||
 		    table.capacity < table.keys.size() || table.descriptors.empty() ||
 		    !indirect_resources.insert(table.resource).second) {
-			if (error != nullptr) {
-				*error = "indirect image snapshot does not match the shader plan";
-			}
 			return false;
 		}
 		for (const auto candidate: table.candidates) {
 			if (candidate >= table.descriptors.size()) {
-				if (error != nullptr) {
-					*error = "indirect image snapshot has an invalid candidate";
-				}
 				return false;
 			}
 		}
 		if (snapshot.images[table.resource] != table.descriptors[table.candidates[0]]) {
-			if (error != nullptr) {
-				*error = "indirect image snapshot has an inconsistent root descriptor";
-			}
 			return false;
 		}
 	}
-	return CheckWidth(snapshot.buffers, 4, "buffer") && CheckWidth(snapshot.images, 8, "image") &&
-	       CheckWidth(snapshot.samplers, 4, "sampler");
+	return CheckWidth(snapshot.buffers, 4) && CheckWidth(snapshot.images, 8) &&
+	       CheckWidth(snapshot.samplers, 4);
 }
 
-bool ValidateResourceSpecialization(const Program& program, const ResourceSnapshot& snapshot,
-                                    std::string* error) {
-	if (!ValidateResourceSnapshot(program, snapshot, error)) {
+bool ValidateResourceSpecialization(const Program& program, const ResourceSnapshot& snapshot) {
+	if (!ValidateResourceSnapshot(program, snapshot)) {
 		return false;
 	}
 	if (!snapshot.indirect_images.empty()) {
-		if (error != nullptr) {
-			*error = "indirect image snapshot was not specialized";
-		}
 		return false;
 	}
 	for (uint32_t i = 0; i < program.info.buffers.size(); i++) {
 		const auto&          buffer = program.info.buffers[i];
 		ShaderBufferResource descriptor;
 		if (!DecodeBufferDescriptor(snapshot.buffers[i], descriptor)) {
-			if (error != nullptr) {
-				*error = fmt::format("buffer descriptor {} has invalid width", i);
-			}
 			return false;
 		}
 		if (buffer.packed_stride != descriptor.PackedStride() ||
 		    buffer.descriptor_format != descriptor.Format() ||
 		    buffer.descriptor_swizzle != descriptor.DstSelXYZW()) {
-			if (error != nullptr) {
-				*error = fmt::format("buffer descriptor {} no longer matches specialization", i);
-			}
 			return false;
 		}
 	}
@@ -403,9 +345,6 @@ bool ValidateResourceSpecialization(const Program& program, const ResourceSnapsh
 		const auto& descriptor = snapshot.images[i];
 		const auto  mip_count  = StorageMipCount(image, descriptor);
 		if (mip_count == 0u || mip_count != image.mip_count) {
-			if (error != nullptr) {
-				*error = fmt::format("storage image descriptor {} changed mip range", i);
-			}
 			return false;
 		}
 		if (NullImageDescriptor(descriptor)) {
@@ -419,10 +358,6 @@ bool ValidateResourceSpecialization(const Program& program, const ResourceSnapsh
 				    program.info.images[root].conversion_format != image.conversion_format ||
 				    program.info.images[root].shader_swizzle != image.shader_swizzle ||
 				    program.info.images[root].cube != image.cube) {
-					if (error != nullptr) {
-						*error = fmt::format(
-						    "null indirect image descriptor {} changed binding class", i);
-					}
 					return false;
 				}
 				continue;
@@ -434,10 +369,6 @@ bool ValidateResourceSpecialization(const Program& program, const ResourceSnapsh
 			}
 			if (image.dimension != Decoder::ImageDimension::Dim2D || image.cube ||
 			    !canonical_kind) {
-				if (error != nullptr) {
-					*error = fmt::format(
-					    "image descriptor {} no longer matches canonical null specialization", i);
-				}
 				return false;
 			}
 			continue;
@@ -445,14 +376,6 @@ bool ValidateResourceSpecialization(const Program& program, const ResourceSnapsh
 		const auto dimension = DescriptorDimension(descriptor, image.dimension);
 		if (dimension == Decoder::ImageDimension::Unknown || dimension != image.dimension ||
 		    DescriptorIsCube(descriptor) != image.cube) {
-			if (error != nullptr) {
-				*error =
-				    fmt::format("image descriptor {} no longer matches specialized dimension: "
-				                "{:08x},{:08x},{:08x},{:08x},{:08x},{:08x},{:08x},{:08x}",
-				                i, descriptor.dwords[0], descriptor.dwords[1], descriptor.dwords[2],
-				                descriptor.dwords[3], descriptor.dwords[4], descriptor.dwords[5],
-				                descriptor.dwords[6], descriptor.dwords[7]);
-			}
 			return false;
 		}
 		if (image.kind == ResourceKind::Image || image.kind == ResourceKind::ImageUint ||
@@ -463,49 +386,25 @@ bool ValidateResourceSpecialization(const Program& program, const ResourceSnapsh
 			const auto format =
 			    static_cast<Prospero::BufferFormat>((descriptor.dwords[1] >> 20u) & 0x1ffu);
 			if (image.atomic && format != Prospero::BufferFormat::k32UInt) {
-				if (error != nullptr) {
-					*error = fmt::format(
-					    "atomic image descriptor {} changed to unsupported format {}", i,
-					    static_cast<uint32_t>(format));
-				}
 				return false;
 			}
 			const auto conversion_format = ImageConversionFormat(format);
 			if (image.conversion_format != conversion_format) {
-				if (error != nullptr) {
-					*error = fmt::format("image descriptor {} changed format conversion", i);
-				}
 				return false;
 			}
 			if ((storage || conversion_format != Prospero::BufferFormat::kInvalid) &&
 			    image.shader_swizzle != DescriptorImageSwizzle(descriptor)) {
-				if (error != nullptr) {
-					*error = fmt::format("image descriptor {} changed swizzle", i);
-				}
 				return false;
 			}
-			const bool raw_sint_storage =
-			    storage && format == Prospero::BufferFormat::k32SInt && image.written &&
-			    !image.read && !image.atomic;
-			const bool uint_descriptor =
-			    Prospero::IsUintTextureFormat(format) || raw_sint_storage;
-			const auto uint_program = image.kind == ResourceKind::ImageUint ||
-			                          image.kind == ResourceKind::StorageImageUint;
+			const bool raw_sint_storage = storage && format == Prospero::BufferFormat::k32SInt &&
+			                              image.written && !image.read && !image.atomic;
+			const bool uint_descriptor  = Prospero::IsUintTextureFormat(format) || raw_sint_storage;
+			const auto uint_program     = image.kind == ResourceKind::ImageUint ||
+			                              image.kind == ResourceKind::StorageImageUint;
 			if (uint_descriptor != uint_program && !(image.atomic && uint_program)) {
-				if (error != nullptr) {
-					*error = fmt::format(
-					    "image descriptor {} no longer matches specialized format: "
-					    "{:08x},{:08x},{:08x},{:08x},{:08x},{:08x},{:08x},{:08x}",
-					    i, descriptor.dwords[0], descriptor.dwords[1], descriptor.dwords[2],
-					    descriptor.dwords[3], descriptor.dwords[4], descriptor.dwords[5],
-					    descriptor.dwords[6], descriptor.dwords[7]);
-				}
 				return false;
 			}
 			if (image.depth_compare && uint_program) {
-				if (error != nullptr) {
-					*error = fmt::format("integer image descriptor {} uses depth comparison", i);
-				}
 				return false;
 			}
 		}
@@ -514,11 +413,8 @@ bool ValidateResourceSpecialization(const Program& program, const ResourceSnapsh
 }
 
 bool MaterializeResources(const Program& program, const SrtRuntime& runtime,
-                          ResourceSnapshot& snapshot, std::string* error) {
+                          ResourceSnapshot& snapshot) {
 	if (!program.resource_tracking_complete) {
-		if (error != nullptr) {
-			*error = "shader resources were not tracked";
-		}
 		return false;
 	}
 
@@ -527,17 +423,12 @@ bool MaterializeResources(const Program& program, const SrtRuntime& runtime,
 	requests.reserve(program.info.buffers.size() + program.info.images.size() * 2u +
 	                 program.info.samplers.size());
 	for (const auto& buffer: program.info.buffers) {
-		requests.push_back({buffer.source, buffer.first_use_pc});
+		requests.push_back({buffer.source});
 	}
 	for (const auto& image: program.info.images) {
 		const auto* source = Source(program, image.source);
 		if (source != nullptr && source->indirect_image.has_value()) {
 			if (runtime.read_specialization_memory == nullptr) {
-				if (error != nullptr) {
-					*error = fmt::format("indirect image table at pc 0x{:08x} has no clean "
-					                     "memory reader",
-					                     image.first_use_pc);
-				}
 				return false;
 			}
 			MarkCleanFlatSlots(program, Source(program, source->indirect_image->material_source),
@@ -545,16 +436,16 @@ bool MaterializeResources(const Program& program, const SrtRuntime& runtime,
 			MarkCleanFlatSlots(program, Source(program, source->indirect_image->heap_source),
 			                   clean_flat_slots);
 		} else {
-			requests.push_back({image.source, image.first_use_pc});
+			requests.push_back({image.source});
 		}
 	}
 	for (const auto& sampler: program.info.samplers) {
-		requests.push_back({sampler.source, sampler.first_use_pc});
+		requests.push_back({sampler.source});
 	}
 	std::vector<DescriptorValue> values;
 	std::vector<uint32_t>        flattened_srt;
-	if (!EvaluateRuntimeSources(program, requests, runtime, values, flattened_srt, clean_flat_slots,
-	                            error)) {
+	if (!EvaluateRuntimeSources(program, requests, runtime, values, flattened_srt,
+	                            clean_flat_slots)) {
 		return false;
 	}
 
@@ -581,20 +472,19 @@ bool MaterializeResources(const Program& program, const SrtRuntime& runtime,
 				continue;
 			}
 			const std::array requests {
-			    DescriptorSourceRequest {source->indirect_image->material_source,
-			                             image.first_use_pc},
-			    DescriptorSourceRequest {source->indirect_image->heap_source, image.first_use_pc}};
+			    DescriptorSourceRequest {source->indirect_image->material_source},
+			    DescriptorSourceRequest {source->indirect_image->heap_source}};
 			SrtRuntime clean_runtime  = runtime;
 			clean_runtime.read_memory = runtime.read_specialization_memory;
 			std::vector<DescriptorValue> tables;
-			if (!EvaluateDescriptorSources(program, requests, clean_runtime, tables, error)) {
+			if (!EvaluateDescriptorSources(program, requests, clean_runtime, tables)) {
 				return false;
 			}
 			const auto&                     material = tables[0];
 			const auto&                     heap     = tables[1];
 			ResourceSnapshot::IndirectImage table;
-			if (!MaterializeIndirectImage(*source->indirect_image, material, heap, runtime, table,
-			                              error, image.first_use_pc)) {
+			if (!MaterializeIndirectImage(*source->indirect_image, material, heap, runtime,
+			                              table)) {
 				return false;
 			}
 			if (image.indirect_root == ImageResource::NoIndirectImage) {
@@ -608,20 +498,12 @@ bool MaterializeResources(const Program& program, const SrtRuntime& runtime,
 			}
 			if (table.keys.size() > image.indirect_mapping_capacity ||
 			    table.descriptors.size() > image.indirect_resources.size()) {
-				if (error != nullptr) {
-					*error = fmt::format(
-					    "indirect image table at pc 0x{:08x} changed candidate topology",
-					    image.first_use_pc);
-				}
 				return false;
 			}
 			for (uint32_t candidate = 0; candidate < image.indirect_resources.size(); candidate++) {
 				const auto resource = image.indirect_resources[candidate];
 				if (resource >= program.info.images.size() ||
 				    program.info.images[resource].indirect_root != image_index) {
-					if (error != nullptr) {
-						*error = "indirect image specialization has an invalid candidate resource";
-					}
 					return false;
 				}
 				next.images[resource] =
@@ -649,31 +531,26 @@ bool MaterializeResources(const Program& program, const SrtRuntime& runtime,
 		}
 	}
 	if (std::ranges::find(image_written, uint8_t {0}) != image_written.end()) {
-		if (error != nullptr) {
-			*error = "indirect image specialization left an unmaterialized candidate";
-		}
 		return false;
 	}
 	next.samplers.assign(cursor, cursor + program.info.samplers.size());
 	next.user_data.assign(runtime.user_data.begin(), runtime.user_data.end());
-	if (!ValidateResourceSnapshot(program, next, error)) {
+	if (!ValidateResourceSnapshot(program, next)) {
 		return false;
 	}
 	snapshot = std::move(next);
 	return true;
 }
 
-bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::string* error) {
+void SpecializeResources(Program& program, ResourceSnapshot& snapshot) {
 	if (!program.resource_tracking_complete || program.shader_info_complete ||
 	    program.binding_layout_complete) {
-		if (error != nullptr) {
-			*error = !program.resource_tracking_complete ? "shader resources were not tracked"
-			                                             : "resource specialization is too late";
-		}
-		return false;
+		return SpecializationFail(!program.resource_tracking_complete
+		                              ? "shader resources were not tracked"
+		                              : "resource specialization is too late");
 	}
-	if (!ValidateResourceSnapshot(program, snapshot, error)) {
-		return false;
+	if (!ValidateResourceSnapshot(program, snapshot)) {
+		return SpecializationFail("invalid resource snapshot");
 	}
 
 	auto next          = program.info;
@@ -681,10 +558,8 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 	for (const auto& table: snapshot.indirect_images) {
 		if (table.resource >= next.images.size() || table.descriptors.size() < 2u ||
 		    next.images.size() + table.descriptors.size() - 1u > ShaderInfo::MaxImages) {
-			if (error != nullptr) {
-				*error = "indirect image candidates exceed the dense image resource limit";
-			}
-			return false;
+			return SpecializationFail(
+			    "indirect image candidates exceed the dense image resource limit");
 		}
 		const auto            root_image = next.images[table.resource];
 		std::vector<uint32_t> resources(table.descriptors.size());
@@ -720,10 +595,7 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 	for (uint32_t i = 0; i < next.buffers.size(); i++) {
 		ShaderBufferResource descriptor;
 		if (!DecodeBufferDescriptor(next_snapshot.buffers[i], descriptor)) {
-			if (error != nullptr) {
-				*error = fmt::format("buffer descriptor {} has invalid width", i);
-			}
-			return false;
+			return SpecializationFail(fmt::format("buffer descriptor {} has invalid width", i));
 		}
 		next.buffers[i].packed_stride      = descriptor.PackedStride();
 		next.buffers[i].descriptor_format  = descriptor.Format();
@@ -734,10 +606,8 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 		auto&       image      = next.images[i];
 		image.mip_count        = StorageMipCount(image, descriptor);
 		if (image.mip_count == 0u) {
-			if (error != nullptr) {
-				*error = fmt::format("storage image descriptor {} has an invalid mip range", i);
-			}
-			return false;
+			return SpecializationFail(
+			    fmt::format("storage image descriptor {} has an invalid mip range", i));
 		}
 		if (NullImageDescriptor(descriptor)) {
 			image.dimension = Decoder::ImageDimension::Dim2D;
@@ -755,37 +625,31 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 		}
 		const auto descriptor_dimension = DescriptorDimension(descriptor, image.dimension);
 		if (descriptor_dimension == Decoder::ImageDimension::Unknown) {
-			if (error != nullptr) {
-				*error = fmt::format(
-				    "image descriptor {} has unsupported type {}: {:08x},{:08x},{:08x},{:08x},"
-				    "{:08x},{:08x},{:08x},{:08x}",
-				    i, (descriptor.dwords[3] >> 28u) & 0xfu, descriptor.dwords[0],
-				    descriptor.dwords[1], descriptor.dwords[2], descriptor.dwords[3],
-				    descriptor.dwords[4], descriptor.dwords[5], descriptor.dwords[6],
-				    descriptor.dwords[7]);
-			}
-			return false;
+			return SpecializationFail(fmt::format(
+			    "image descriptor {} has unsupported type {}: {:08x},{:08x},{:08x},{:08x},"
+			    "{:08x},{:08x},{:08x},{:08x}",
+			    i, (descriptor.dwords[3] >> 28u) & 0xfu, descriptor.dwords[0], descriptor.dwords[1],
+			    descriptor.dwords[2], descriptor.dwords[3], descriptor.dwords[4],
+			    descriptor.dwords[5], descriptor.dwords[6], descriptor.dwords[7]));
 		}
 		image.dimension = descriptor_dimension;
 		image.cube      = DescriptorIsCube(descriptor);
 		const auto format =
 		    static_cast<Prospero::BufferFormat>((descriptor.dwords[1] >> 20u) & 0x1ffu);
 		if (image.atomic && format != Prospero::BufferFormat::k32UInt) {
-			if (error != nullptr) {
-				*error = fmt::format("atomic image descriptor {} uses unsupported format {}", i,
-				                     static_cast<uint32_t>(format));
-			}
-			return false;
+			return SpecializationFail(
+			    fmt::format("atomic image descriptor {} uses unsupported format {}", i,
+			                static_cast<uint32_t>(format)));
 		}
-		const bool storage = image.kind == ResourceKind::StorageImage ||
-		                     image.kind == ResourceKind::StorageImageUint;
+		const bool storage      = image.kind == ResourceKind::StorageImage ||
+		                          image.kind == ResourceKind::StorageImageUint;
 		image.conversion_format = ImageConversionFormat(format);
 		if (storage || image.conversion_format != Prospero::BufferFormat::kInvalid) {
 			image.shader_swizzle = DescriptorImageSwizzle(descriptor);
 		}
 		const bool raw_sint_storage = storage && format == Prospero::BufferFormat::k32SInt &&
 		                              image.written && !image.read && !image.atomic;
-		const bool uint_image = Prospero::IsUintTextureFormat(format) || raw_sint_storage;
+		const bool uint_image       = Prospero::IsUintTextureFormat(format) || raw_sint_storage;
 		if (uint_image) {
 			switch (image.kind) {
 				case ResourceKind::Image: image.kind = ResourceKind::ImageUint; break;
@@ -803,19 +667,13 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 		    static_cast<size_t>(root.indirect_mapping_offset) + 1u +
 		            static_cast<size_t>(root.indirect_mapping_capacity) * 2u >
 		        next_snapshot.flattened_srt.size()) {
-			if (error != nullptr) {
-				*error = "indirect image specialization has an invalid key mapping";
-			}
-			return false;
+			return SpecializationFail("indirect image specialization has an invalid key mapping");
 		}
 		uint32_t exemplar = ImageResource::NoIndirectImage;
 		for (const auto resource: root.indirect_resources) {
 			if (resource >= next.images.size() ||
 			    next.images[resource].indirect_root != root_index) {
-				if (error != nullptr) {
-					*error = "indirect image specialization has an invalid candidate";
-				}
-				return false;
+				return SpecializationFail("indirect image specialization has an invalid candidate");
 			}
 			if (!NullImageDescriptor(next_snapshot.images[resource])) {
 				exemplar = resource;
@@ -823,10 +681,7 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 			}
 		}
 		if (exemplar == ImageResource::NoIndirectImage) {
-			if (error != nullptr) {
-				*error = "indirect image specialization has no typed candidate";
-			}
-			return false;
+			return SpecializationFail("indirect image specialization has no typed candidate");
 		}
 		const auto& image_class = next.images[exemplar];
 		for (uint32_t candidate = 0; candidate < next.images.size(); candidate++) {
@@ -849,12 +704,9 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 			    image.shader_swizzle != image_class.shader_swizzle ||
 			    image.depth_compare != image_class.depth_compare ||
 			    image.cube != image_class.cube) {
-				if (error != nullptr) {
-					*error = fmt::format(
-					    "indirect image table at pc 0x{:08x} has incompatible candidates",
-					    root.first_use_pc);
-				}
-				return false;
+				return SpecializationFail(
+				    fmt::format("indirect image table at pc 0x{:08x} has incompatible candidates",
+				                root.first_use_pc));
 			}
 		}
 	}
@@ -867,11 +719,8 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 	std::vector<SamplerUsage> sampler_usage(next.samplers.size());
 	for (const auto& pair: next.sampled_pairs) {
 		if (pair.image >= next.images.size() || pair.sampler >= next.samplers.size()) {
-			if (error != nullptr) {
-				*error = fmt::format("sampled resource pair at pc 0x{:08x} is invalid",
-				                     pair.first_use_pc);
-			}
-			return false;
+			return SpecializationFail(
+			    fmt::format("sampled resource pair at pc 0x{:08x} is invalid", pair.first_use_pc));
 		}
 		auto& usage = sampler_usage[pair.sampler];
 		if (next.images[pair.image].conversion_format != Prospero::BufferFormat::kInvalid) {
@@ -891,10 +740,8 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 			continue;
 		}
 		if (next.samplers.size() >= ShaderInfo::MaxSamplers) {
-			if (error != nullptr) {
-				*error = "point-filter sampler variants exceed the dense sampler resource limit";
-			}
-			return false;
+			return SpecializationFail(
+			    "point-filter sampler variants exceed the dense sampler resource limit");
 		}
 		point_sampler[i]              = static_cast<uint32_t>(next.samplers.size());
 		auto sampler                  = next.samplers[i];
@@ -915,19 +762,13 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 			}
 			const auto index = inst.Flags<MemoryFlags>().index;
 			if (index >= memory_info.size()) {
-				if (error != nullptr) {
-					*error = fmt::format("typed image instruction has invalid memory metadata {}",
-					                     index);
-				}
-				return false;
+				return SpecializationFail(
+				    fmt::format("typed image instruction has invalid memory metadata {}", index));
 			}
 			auto& memory = memory_info[index];
 			if (memory.resource >= next.images.size()) {
-				if (error != nullptr) {
-					*error = fmt::format("typed image instruction has invalid resource {}",
-					                     memory.resource);
-				}
-				return false;
+				return SpecializationFail(fmt::format(
+				    "typed image instruction has invalid resource {}", memory.resource));
 			}
 			const auto& image = next.images[memory.resource];
 			if (image_opcode.needs_sampler &&
@@ -937,22 +778,18 @@ bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::stri
 			}
 			if (image.indirect_root == memory.resource &&
 			    inst.GetOpcode() != ValueOpcode::ImageSampleRaw) {
-				if (error != nullptr) {
-					*error = fmt::format(
-					    "indirect image table at pc 0x{:08x} is used by unsupported {}",
-					    inst.Flags<MemoryFlags>().pc, ValueOpcodeName(inst.GetOpcode()));
-				}
-				return false;
+				return SpecializationFail(
+				    fmt::format("indirect image table at pc 0x{:08x} is used by unsupported {}",
+				                inst.Flags<MemoryFlags>().pc, ValueOpcodeName(inst.GetOpcode())));
 			}
 			memory.kind            = image.kind;
 			memory.image_dimension = image.dimension;
 			memory.image_cube      = image.cube;
 		}
 	}
-	program.info                = std::move(next);
+	program.info        = std::move(next);
 	program.memory_info = std::move(memory_info);
-	snapshot                    = std::move(next_snapshot);
-	return true;
+	snapshot            = std::move(next_snapshot);
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR

src/graphics/shader/recompiler/ir/passes/ResourceMaterialization.h

@@ -22,19 +22,17 @@ struct ResourceSnapshot {
 	std::vector<IndirectImage>   indirect_images;
 };
 
-bool ValidateResourceSnapshot(const Program& program, const ResourceSnapshot& snapshot,
-                              std::string* error);
-bool ValidateResourceSpecialization(const Program& program, const ResourceSnapshot& snapshot,
-                                    std::string* error);
+bool ValidateResourceSnapshot(const Program& program, const ResourceSnapshot& snapshot);
+bool ValidateResourceSpecialization(const Program& program, const ResourceSnapshot& snapshot);
 
 // Resolves the immutable dense resource topology against one runtime user-data/SRT snapshot.
 // On failure the destination is unchanged.
 bool MaterializeResources(const Program& program, const SrtRuntime& runtime,
-                          ResourceSnapshot& snapshot, std::string* error);
+                          ResourceSnapshot& snapshot);
 
 // Applies runtime descriptor shape/format facts to a copied dense topology before layout and
-// emission. On failure the program is unchanged.
-bool SpecializeResources(Program& program, ResourceSnapshot& snapshot, std::string* error);
+// emission. Invalid specializations are fatal.
+void SpecializeResources(Program& program, ResourceSnapshot& snapshot);
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR
 

src/graphics/shader/recompiler/ir/passes/ResourceTracking.cpp

@@ -1,8 +1,8 @@
 #include "graphics/shader/recompiler/ir/passes/ResourceTracking.h"
 
+#include "common/assert.h"
 #include "graphics/shader/recompiler/ir/ShaderIR.h"
 #include "graphics/shader/recompiler/ir/passes/SrtWalker.h"
-#include "graphics/shader/shader.h"
 
 #include <algorithm>
 #include <fmt/format.h>
@@ -12,7 +12,7 @@
 namespace Libs::Graphics::ShaderRecompiler::IR {
 namespace {
 
-constexpr uint32_t SamplerBorderClampMask = (1u << 2u) | (1u << 5u) | (1u << 8u);
+constexpr uint32_t SamplerBorderClampMask    = (1u << 2u) | (1u << 5u) | (1u << 8u);
 constexpr uint32_t SamplerDword3ReservedMask = 0x3ffff000u;
 
 uint32_t PossibleU32Bits(Value value) {
@@ -101,19 +101,17 @@ public:
 		m_info.uses_dma = false;
 	}
 
-	bool Run(std::string* error) {
+	void Run() {
 		if (m_program.resource_tracking_complete) {
-			return Fail(0, error, "resources already tracked");
+			Fail(0, "resources already tracked");
 		}
 		if (!m_program.srt_plan_complete) {
-			return Fail(0, error, "SRT plan is not ready");
+			Fail(0, "SRT plan is not ready");
 		}
 		PlanIndirectImages();
 		for (auto* block: m_program.blocks) {
 			for (auto& inst: *block) {
-				if (!Collect(inst, error)) {
-					return false;
-				}
+				Collect(inst);
 			}
 		}
 		LinkImageAliases();
@@ -141,14 +139,14 @@ public:
 		}
 		std::erase_if(m_program.dynamic_reads, [&](Value value) {
 			const auto* inst = value.Resolve().TryInstruction();
-			return std::ranges::any_of(m_indirect_images, [&](const IndirectImagePlan& plan) {
+			return std::any_of(m_indirect_images.begin(), m_indirect_images.end(),
+			                   [&](const IndirectImagePlan& plan) {
 				return std::ranges::find(plan.reads, inst) != plan.reads.end();
 			});
 		});
-		m_program.descriptor_sources          = std::move(m_sources);
+		m_program.descriptor_sources         = std::move(m_sources);
 		m_program.info                       = std::move(m_info);
 		m_program.resource_tracking_complete = true;
-		return true;
 	}
 
 private:
@@ -173,18 +171,19 @@ private:
 		std::array<const Inst*, 8> reads {};
 	};
 
-	bool Fail(uint32_t pc, std::string* error, const std::string& reason) const {
-		return ShaderError::Fail(
-		    error, fmt::format("shader resource tracking: hash=0x{:016x} stage={} pc=0x{:08x} {}",
-		                       m_program.shader_hash, StageName(m_program.stage), pc, reason));
+	[[noreturn]] void Fail(uint32_t pc, const std::string& reason) const {
+		const auto message =
+		    fmt::format("shader resource tracking: hash=0x{:016x} stage={} pc=0x{:08x} {}",
+		                m_program.shader_hash, StageName(m_program.stage), pc, reason);
+		EXIT("%s", message.c_str());
+		std::abort();
 	}
 
-	bool MakeSource(const Inst& handle, uint32_t width, bool sampler, bool sample_adjust,
-	                DescriptorSource& descriptor, uint32_t pc, std::string* error) const {
+	void MakeSource(const Inst& handle, uint32_t width, bool sampler, bool sample_adjust,
+	                DescriptorSource& descriptor, uint32_t pc) const {
 		if (handle.NumArgs() != width) {
-			return Fail(pc, error,
-			            fmt::format("{} has {} descriptor dwords, expected {}",
-			                        ValueOpcodeName(handle.GetOpcode()), handle.NumArgs(), width));
+			Fail(pc, fmt::format("{} has {} descriptor dwords, expected {}",
+			                     ValueOpcodeName(handle.GetOpcode()), handle.NumArgs(), width));
 		}
 		descriptor.dword_count = width;
 		for (uint32_t i = 0; i < width; i++) {
@@ -199,18 +198,15 @@ private:
 			// Border color and its table index are unused unless a clamp axis selects border mode.
 			descriptor.dwords[3] = Value(0u);
 		}
-		return true;
 	}
 
-	bool ValidateSource(const DescriptorSource& descriptor, std::string& reason,
-	                    uint32_t& bad_dword) const {
+	bool ValidateSource(const DescriptorSource& descriptor, uint32_t& bad_dword) const {
 		for (uint32_t i = 0; i < descriptor.dword_count; i++) {
 			bad_dword = i;
 			if (descriptor.dwords[i].Resolve().GetType() != Type::U32) {
-				reason = "is not U32";
 				return false;
 			}
-			if (!ValidateRuntimeValue(m_program, descriptor.dwords[i], reason)) {
+			if (!ValidateRuntimeValue(m_program, descriptor.dwords[i])) {
 				return false;
 			}
 		}
@@ -286,13 +282,12 @@ private:
 
 	bool MakeRuntimeBufferSource(const Inst& handle, uint32_t pc, uint32_t& source,
 	                             DescriptorSource& descriptor) {
-		if (handle.GetOpcode() != ValueOpcode::GetBufferResource ||
-		    !MakeSource(handle, 4u, false, false, descriptor, pc, nullptr)) {
+		if (handle.GetOpcode() != ValueOpcode::GetBufferResource) {
 			return false;
 		}
-		std::string reason;
-		uint32_t    bad_dword = 0;
-		if (!ValidateSource(descriptor, reason, bad_dword)) {
+		MakeSource(handle, 4u, false, false, descriptor, pc);
+		uint32_t bad_dword = 0;
+		if (!ValidateSource(descriptor, bad_dword)) {
 			return false;
 		}
 		source = InternSource(descriptor);
@@ -438,14 +433,16 @@ private:
 
 	const IndirectImagePlan* FindIndirectImage(const Inst& handle) const {
 		const auto found =
-		    std::ranges::find_if(m_indirect_images, [&](const IndirectImagePlan& plan) {
+		    std::find_if(m_indirect_images.begin(), m_indirect_images.end(),
+		                 [&](const IndirectImagePlan& plan) {
 			    return plan.handle == &handle;
 		    });
 		return found == m_indirect_images.end() ? nullptr : &*found;
 	}
 
 	bool IsIndirectPlanningMemory(uint32_t index) const {
-		return std::ranges::any_of(m_indirect_images, [&](const IndirectImagePlan& plan) {
+		return std::any_of(m_indirect_images.begin(), m_indirect_images.end(),
+		                   [&](const IndirectImagePlan& plan) {
 			return std::ranges::find(plan.memory, index) != plan.memory.end();
 		});
 	}
@@ -469,38 +466,30 @@ private:
 		}
 	}
 
-	bool GetHandle(Value value, ValueOpcode expected, uint32_t width, uint32_t pc, Inst*& handle,
-	               uint32_t& source, std::string* error, bool sampler = false,
-	               bool sample_adjust = false) {
+	void GetHandle(Value value, ValueOpcode expected, uint32_t width, uint32_t pc, Inst*& handle,
+	               uint32_t& source, bool sampler = false, bool sample_adjust = false) {
 		handle = value.Resolve().TryInstruction();
 		if (handle == nullptr || handle->GetOpcode() != expected) {
-			return Fail(pc, error,
-			            fmt::format("memory operation requires {}", ValueOpcodeName(expected)));
+			Fail(pc, fmt::format("memory operation requires {}", ValueOpcodeName(expected)));
 		}
 		DescriptorSource descriptor;
-		if (!MakeSource(*handle, width, sampler, sample_adjust, descriptor, pc, error)) {
-			return false;
-		}
-		std::string reason;
-		uint32_t    bad_dword = 0;
-		if (!ValidateSource(descriptor, reason, bad_dword)) {
-			return Fail(
-			    pc, error,
-			    fmt::format("{} dword {} {}", ValueOpcodeName(expected), bad_dword, reason));
+		MakeSource(*handle, width, sampler, sample_adjust, descriptor, pc);
+		uint32_t bad_dword = 0;
+		if (!ValidateSource(descriptor, bad_dword)) {
+			Fail(pc, fmt::format("{} dword {} is not a valid runtime value",
+			                     ValueOpcodeName(expected), bad_dword));
 		}
 		source = InternSource(descriptor);
-		return true;
 	}
 
-	bool ValidateAddressHandle(Value value, uint32_t pc, std::string* error) const {
+	void ValidateAddressHandle(Value value, uint32_t pc) const {
 		const auto* handle = value.Resolve().TryInstruction();
 		if (handle == nullptr || handle->GetOpcode() != ValueOpcode::GetAddressResource) {
-			return Fail(pc, error, "address operation requires GetAddressResource");
+			Fail(pc, "address operation requires GetAddressResource");
 		}
 		if (handle->NumArgs() != 2) {
-			return Fail(pc, error, "GetAddressResource must have two address dwords");
+			Fail(pc, "GetAddressResource must have two address dwords");
 		}
-		return true;
 	}
 
 	uint32_t AddBuffer(uint32_t source, const MemoryInfo& memory, ValueOpcode op, uint32_t pc) {
@@ -540,7 +529,7 @@ private:
 		const auto mip   = MipMode(memory);
 		const bool depth = (memory.image_sample_flags & Decoder::ImageSampleFlagCompare) != 0;
 		for (uint32_t i = 0; i < m_info.images.size(); i++) {
-			auto& image = m_info.images[i];
+			auto&      image           = m_info.images[i];
 			const bool compatible_kind = image.kind == memory.kind || (IsStorageImage(image.kind) &&
 			                                                           IsStorageImage(memory.kind));
 			if (image.source == source && compatible_kind &&
@@ -593,155 +582,139 @@ private:
 		return static_cast<uint32_t>(m_info.samplers.size() - 1);
 	}
 
-	bool AddSampledPair(uint32_t image, uint32_t sampler, uint32_t pc, std::string* error) {
+	void AddSampledPair(uint32_t image, uint32_t sampler, uint32_t pc) {
 		for (auto& pair: m_info.sampled_pairs) {
 			if (pair.image == image && pair.sampler == sampler) {
 				pair.first_use_pc = std::min(pair.first_use_pc, pc);
-				return true;
+				return;
 			}
 		}
 		if (m_info.sampled_pairs.size() >= ShaderInfo::MaxSampledPairs) {
-			return Fail(pc, error, "sampled image/sampler pair limit exceeded");
+			Fail(pc, "sampled image/sampler pair limit exceeded");
 		}
 		m_info.sampled_pairs.push_back({image, sampler, pc});
-		return true;
 	}
 
-	bool AddHandlePatch(Inst* handle, uint32_t resource, uint32_t pc, std::string* error) {
+	void AddHandlePatch(Inst* handle, uint32_t resource, uint32_t pc) {
 		for (const auto& patch: m_handle_patches) {
 			if (patch.handle == handle) {
-				return patch.resource == resource ||
-				       Fail(pc, error,
-				            fmt::format("{} is reused with incompatible resource classes",
-				                        ValueOpcodeName(handle->GetOpcode())));
+				if (patch.resource != resource) {
+					Fail(pc, fmt::format("{} is reused with incompatible resource classes",
+					                     ValueOpcodeName(handle->GetOpcode())));
+				}
+				return;
 			}
 		}
 		m_handle_patches.push_back({handle, resource});
-		return true;
 	}
 
-	bool AddMemoryPatch(uint32_t index, uint32_t resource, uint32_t sampler, bool has_sampler,
-	                    uint32_t pc, std::string* error) {
+	void AddMemoryPatch(uint32_t index, uint32_t resource, uint32_t sampler, bool has_sampler,
+	                    uint32_t pc) {
 		for (auto& patch: m_memory_patches) {
 			if (patch.index != index) {
 				continue;
 			}
 			if (patch.resource != resource ||
 			    (has_sampler && patch.has_sampler && patch.sampler != sampler)) {
-				return Fail(pc, error, "memory metadata is reused with incompatible resources");
+				Fail(pc, "memory metadata is reused with incompatible resources");
 			}
 			if (has_sampler) {
 				patch.sampler     = sampler;
 				patch.has_sampler = true;
 			}
-			return true;
+			return;
 		}
 		m_memory_patches.push_back({index, resource, sampler, has_sampler});
-		return true;
 	}
 
-	bool Collect(Inst& inst, std::string* error) {
+	void Collect(Inst& inst) {
 		const auto op           = inst.GetOpcode();
 		const auto buffer       = BufferAccessOf(op);
 		const auto address_info = AddressOpcodeInfoOf(op);
 		const auto image_info   = ImageOpcodeInfoOf(op);
 		if (buffer == BufferAccess::None && address_info.access == AddressAccess::None &&
 		    image_info.access == ImageAccess::None) {
-			return true;
+			return;
 		}
 		const auto flags = inst.Flags<MemoryFlags>();
 		if (flags.index >= m_program.memory_info.size()) {
-			return Fail(flags.pc, error,
-			            fmt::format("memory metadata index {} is out of range", flags.index));
+			Fail(flags.pc, fmt::format("memory metadata index {} is out of range", flags.index));
 		}
 		if (inst.NumArgs() == 0) {
-			return Fail(flags.pc, error, "memory operation has no resource handle");
+			Fail(flags.pc, "memory operation has no resource handle");
 		}
 		const auto& memory = m_program.memory_info[flags.index];
 		if (memory.planning_only || IsIndirectPlanningMemory(flags.index)) {
-			return true;
+			return;
 		}
 		Inst*    handle   = nullptr;
 		uint32_t source   = 0;
 		uint32_t resource = 0;
 
 		if (buffer != BufferAccess::None) {
-			if (!GetHandle(inst.Arg(0), ValueOpcode::GetBufferResource, 4, flags.pc, handle, source,
-			               error)) {
-				return false;
-			}
+			GetHandle(inst.Arg(0), ValueOpcode::GetBufferResource, 4, flags.pc, handle, source);
 			resource = AddBuffer(source, memory, op, flags.pc);
 			if (resource == UINT32_MAX) {
-				return Fail(flags.pc, error, "buffer resource limit exceeded");
+				Fail(flags.pc, "buffer resource limit exceeded");
 			}
-			return AddHandlePatch(handle, resource, flags.pc, error) &&
-			       AddMemoryPatch(flags.index, resource, 0, false, flags.pc, error);
+			AddHandlePatch(handle, resource, flags.pc);
+			AddMemoryPatch(flags.index, resource, 0, false, flags.pc);
+			return;
 		}
 		if (address_info.access != AddressAccess::None) {
 			if (!IsAddressResourceKind(memory.kind)) {
-				return Fail(flags.pc, error, "address operation has invalid resource kind");
+				Fail(flags.pc, "address operation has invalid resource kind");
 			}
 			if (memory.kind == ResourceKind::Scratch) {
 				handle = inst.Arg(0).Resolve().TryInstruction();
 				if (handle == nullptr || handle->GetOpcode() != ValueOpcode::GetScratchResource ||
 				    handle->NumArgs() != 0) {
-					return Fail(flags.pc, error, "scratch operation requires GetScratchResource");
+					Fail(flags.pc, "scratch operation requires GetScratchResource");
 				}
 				if (m_program.scratch_dwords == 0) {
-					return Fail(flags.pc, error,
-					            "scratch operation requires a nonzero AGC per-thread size");
+					Fail(flags.pc, "scratch operation requires a nonzero AGC per-thread size");
 				}
-				return true;
-			}
-			if (!ValidateAddressHandle(inst.Arg(0), flags.pc, error)) {
-				return false;
+				return;
 			}
+			ValidateAddressHandle(inst.Arg(0), flags.pc);
 			m_info.uses_dma = true;
-			return true;
+			return;
 		}
 
 		if (!ImageResourceKindMatches(memory.kind, image_info.resource_class)) {
-			return Fail(flags.pc, error, "image operation has invalid resource kind");
+			Fail(flags.pc, "image operation has invalid resource kind");
 		}
 		handle               = inst.Arg(0).Resolve().TryInstruction();
 		const auto* indirect = handle != nullptr ? FindIndirectImage(*handle) : nullptr;
 		if (indirect != nullptr) {
 			source = indirect->source;
-		} else if (!GetHandle(inst.Arg(0), ValueOpcode::GetImageResource, 8, flags.pc, handle,
-		                      source, error)) {
-			return false;
+		} else {
+			GetHandle(inst.Arg(0), ValueOpcode::GetImageResource, 8, flags.pc, handle, source);
 		}
 		resource = AddImage(source, memory, op, flags.pc);
 		if (resource == UINT32_MAX) {
-			return Fail(flags.pc, error, "image resource limit exceeded");
-		}
-		if (!AddHandlePatch(handle, resource, flags.pc, error)) {
-			return false;
+			Fail(flags.pc, "image resource limit exceeded");
 		}
+		AddHandlePatch(handle, resource, flags.pc);
 		uint32_t sampler = 0;
 		if (image_info.needs_sampler) {
 			if (inst.NumArgs() < 2) {
-				return Fail(flags.pc, error, "sampled image operation has no sampler handle");
+				Fail(flags.pc, "sampled image operation has no sampler handle");
 			}
-			Inst*    sampler_handle = nullptr;
-			uint32_t sampler_source = 0;
+			Inst*      sampler_handle = nullptr;
+			uint32_t   sampler_source = 0;
 			const bool sample_adjust =
 			    (memory.image_sample_flags & Decoder::ImageSampleFlagAdjust) != 0;
-			if (!GetHandle(inst.Arg(1), ValueOpcode::GetSamplerResource, 4, flags.pc,
-			               sampler_handle, sampler_source, error, true, sample_adjust)) {
-				return false;
-			}
+			GetHandle(inst.Arg(1), ValueOpcode::GetSamplerResource, 4, flags.pc, sampler_handle,
+			          sampler_source, true, sample_adjust);
 			sampler = AddSampler(sampler_source, flags.pc);
 			if (sampler == UINT32_MAX) {
-				return Fail(flags.pc, error, "sampler resource limit exceeded");
-			}
-			if (!AddHandlePatch(sampler_handle, sampler, flags.pc, error) ||
-			    !AddSampledPair(resource, sampler, flags.pc, error)) {
-				return false;
+				Fail(flags.pc, "sampler resource limit exceeded");
 			}
+			AddHandlePatch(sampler_handle, sampler, flags.pc);
+			AddSampledPair(resource, sampler, flags.pc);
 		}
-		return AddMemoryPatch(flags.index, resource, sampler, image_info.needs_sampler, flags.pc,
-		                      error);
+		AddMemoryPatch(flags.index, resource, sampler, image_info.needs_sampler, flags.pc);
 	}
 
 	const DescriptorSource* Source(uint32_t source) const {
@@ -783,8 +756,8 @@ private:
 
 } // namespace
 
-bool TrackResources(Program& program, std::string* error) {
-	return Tracker(program).Run(error);
+void TrackResources(Program& program) {
+	Tracker(program).Run();
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR

src/graphics/shader/recompiler/ir/passes/ResourceTracking.h

@@ -7,7 +7,7 @@ namespace Libs::Graphics::ShaderRecompiler::IR {
 
 // Collects immutable resource topology from typed SSA handles, interns their resolved dwords in
 // descriptor_sources, then writes dense indices to handle flags and MemoryInfo.
-bool TrackResources(Program& program, std::string* error);
+void TrackResources(Program& program);
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR
 

src/graphics/shader/recompiler/ir/passes/ShaderInfoCollection.cpp

@@ -1,5 +1,6 @@
 #include "graphics/shader/recompiler/ir/passes/ShaderInfoCollection.h"
 
+#include "common/assert.h"
 #include "graphics/shader/recompiler/ir/ShaderIR.h"
 
 #include <algorithm>
@@ -8,6 +9,11 @@
 namespace Libs::Graphics::ShaderRecompiler::IR {
 namespace {
 
+[[noreturn]] void Fail(std::string_view message) {
+	EXIT("shader info collection failed: %s", std::string(message).c_str());
+	std::abort();
+}
+
 void AddInput(ShaderInfo& info, StageInputKind kind, uint32_t location, uint32_t components,
               std::string name, bool per_vertex = false) {
 	const auto input = std::find_if(info.inputs.begin(), info.inputs.end(), [=](const auto& value) {
@@ -34,38 +40,38 @@ void AddOutput(ShaderInfo& info, StageOutputKind kind, uint32_t index, uint32_t
 	}
 }
 
-bool ValidateOptions(const Program& program, const ShaderInfoOptions& options, std::string* error) {
-	auto Fail = [&](const char* message) {
-		if (error != nullptr) {
-			*error = message;
-		}
-		return false;
-	};
+void ValidateOptions(const Program& program, const ShaderInfoOptions& options) {
 	switch (program.stage) {
 		case ShaderType::Vertex:
+			if (options.vertex == nullptr) {
+				return Fail("vertex shader has no input metadata");
+			}
 			if (options.vertex->resources_num < 0 ||
 			    options.vertex->resources_num > ShaderVertexInputInfo::RES_MAX) {
 				return Fail("vertex resource count is out of range");
 			}
-			return true;
+			return;
 		case ShaderType::Pixel:
-			return options.pixel->input_num <= std::size(options.pixel->interpolator_settings) ||
-			       Fail("pixel input count is out of range");
+			if (options.pixel == nullptr) {
+				return Fail("pixel shader has no input metadata");
+			}
+			if (options.pixel->input_num > std::size(options.pixel->interpolator_settings)) {
+				return Fail("pixel input count is out of range");
+			}
+			return;
 		case ShaderType::Compute:
-			return (options.compute->thread_ids_num >= 0 && options.compute->thread_ids_num <= 3) ||
-			       Fail("compute thread ID count is out of range");
+			if (options.compute == nullptr) {
+				return Fail("compute shader has no input metadata");
+			}
+			if (options.compute->thread_ids_num < 0 || options.compute->thread_ids_num > 3) {
+				return Fail("compute thread ID count is out of range");
+			}
+			return;
 		default: return Fail("unsupported shader stage for info collection");
 	}
 }
 
-bool ValidateValueReferences(const Program& program, const ShaderInfoOptions& options,
-                             std::string* error) {
-	const auto Fail = [&](const char* message) {
-		if (error != nullptr) {
-			*error = message;
-		}
-		return false;
-	};
+void ValidateValueReferences(const Program& program, const ShaderInfoOptions& options) {
 	for (const auto* block: program.blocks) {
 		for (const auto& inst: *block) {
 			switch (inst.GetOpcode()) {
@@ -148,7 +154,6 @@ bool ValidateValueReferences(const Program& program, const ShaderInfoOptions& op
 			}
 		}
 	}
-	return true;
 }
 
 void CollectVertexInputs(const Program& program, const ShaderVertexInputInfo* vertex,
@@ -269,14 +274,8 @@ void CollectBuiltinInputs(const Program& program, ShaderInfo& info) {
 	}
 }
 
-bool CollectOutputs(const Program& program, const ShaderVertexInputInfo* vertex,
-	                const ShaderPixelInputInfo* pixel, ShaderInfo& info, std::string* error) {
-	auto Fail = [&](std::string message) {
-		if (error != nullptr) {
-			*error = std::move(message);
-		}
-		return false;
-	};
+void CollectOutputs(const Program& program, const ShaderVertexInputInfo* vertex,
+                    const ShaderPixelInputInfo* pixel, ShaderInfo& info) {
 	for (const auto* block: program.blocks) {
 		for (const auto& inst: *block) {
 			if (inst.GetOpcode() != ValueOpcode::SetAttribute) {
@@ -347,31 +346,23 @@ bool CollectOutputs(const Program& program, const ShaderVertexInputInfo* vertex,
 			}
 		}
 	}
-	return true;
 }
 
 } // namespace
 
-bool CollectShaderInfo(Program& program, const ShaderInfoOptions& options, std::string* error) {
+void CollectShaderInfo(Program& program, const ShaderInfoOptions& options) {
 	if (!program.resource_tracking_complete || program.shader_info_complete) {
-		if (error != nullptr) {
-			*error = !program.resource_tracking_complete ? "shader resources were not tracked"
-			                                             : "shader info already collected";
-		}
-		return false;
-	}
-	if (!ValidateOptions(program, options, error)) {
-		return false;
-	}
-	if (!ValidateValueReferences(program, options, error)) {
-		return false;
+		return Fail(!program.resource_tracking_complete ? "shader resources were not tracked"
+		                                                : "shader info already collected");
 	}
+	ValidateOptions(program, options);
+	ValidateValueReferences(program, options);
 
 	auto next = program.info;
 	next.inputs.clear();
 	next.outputs.clear();
-	next.has_bitwise_xor = std::any_of(
-	    program.blocks.begin(), program.blocks.end(), [](const auto* block) {
+	next.has_bitwise_xor =
+	    std::any_of(program.blocks.begin(), program.blocks.end(), [](const auto* block) {
 		    return std::any_of(block->begin(), block->end(), [](const auto& inst) {
 			    return inst.GetOpcode() == ValueOpcode::BitwiseXor32;
 		    });
@@ -380,15 +371,12 @@ bool CollectShaderInfo(Program& program, const ShaderInfoOptions& options, std::
 		case ShaderType::Vertex: CollectVertexInputs(program, options.vertex, next); break;
 		case ShaderType::Pixel: CollectPixelInputs(program, options.pixel, next); break;
 		case ShaderType::Compute: CollectComputeInputs(options.compute, next); break;
-		default: return false;
+		default: return Fail("unsupported shader stage for info collection");
 	}
 	CollectBuiltinInputs(program, next);
-	if (!CollectOutputs(program, options.vertex, options.pixel, next, error)) {
-		return false;
-	}
+	CollectOutputs(program, options.vertex, options.pixel, next);
 	program.info                 = std::move(next);
 	program.shader_info_complete = true;
-	return true;
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR

src/graphics/shader/recompiler/ir/passes/ShaderInfoCollection.h

@@ -13,7 +13,7 @@ struct ShaderInfoOptions {
 
 // Completes the immutable shader interface after resource tracking. On failure Program::info and
 // all completion state remain unchanged.
-bool CollectShaderInfo(Program& program, const ShaderInfoOptions& options, std::string* error);
+void CollectShaderInfo(Program& program, const ShaderInfoOptions& options);
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR
 

src/graphics/shader/recompiler/ir/passes/SrtWalker.cpp

@@ -1,7 +1,7 @@
 #include "graphics/shader/recompiler/ir/passes/SrtWalker.h"
 
+#include "common/assert.h"
 #include "graphics/shader/recompiler/ir/ShaderIR.h"
-#include "graphics/shader/shader.h"
 
 #include <algorithm>
 #include <bit>
@@ -120,10 +120,10 @@ class RuntimeValidator {
 public:
 	explicit RuntimeValidator(const Program& program): m_program(program) {}
 
-	bool Run(Value value, std::string& reason) { return Validate(value, reason); }
+	bool Run(Value value) { return Validate(value); }
 
 private:
-	bool Validate(Value value, std::string& reason) {
+	bool Validate(Value value) {
 		value            = value.Resolve();
 		const auto* inst = value.TryInstruction();
 		if (inst == nullptr) {
@@ -133,14 +133,10 @@ private:
 				case Type::U16:
 				case Type::U32:
 				case Type::U64: return true;
-				default:
-					reason = fmt::format("contains a non-integer immediate of type {}",
-					                     TypeName(value.GetType()));
-					return false;
+				default: return false;
 			}
 		}
 		if (!m_visiting.insert(inst).second) {
-			reason = fmt::format("contains a cyclic {} value", ValueOpcodeName(inst->GetOpcode()));
 			return false;
 		}
 		const auto finish = [&](bool valid) {
@@ -151,25 +147,21 @@ private:
 		if (op == ValueOpcode::UndefU1 || op == ValueOpcode::UndefU8 ||
 		    op == ValueOpcode::UndefU16 || op == ValueOpcode::UndefU32 ||
 		    op == ValueOpcode::UndefU64 || op == ValueOpcode::Void) {
-			reason = fmt::format("contains {}", ValueOpcodeName(op));
 			return finish(false);
 		}
 		if (op == ValueOpcode::GetUserData) {
 			if (inst->NumArgs() != 1 || inst->Arg(0).GetType() != Type::ScalarReg) {
-				reason = "contains a malformed GetUserData";
 				return finish(false);
 			}
 			const auto reg = RegIndex(inst->Arg(0).ScalarRegister());
 			if (reg < m_program.user_data_base ||
 			    reg - m_program.user_data_base >= m_program.user_data_count) {
-				reason = fmt::format("references unavailable user SGPR {}", reg);
 				return finish(false);
 			}
 			return finish(true);
 		}
 		if (op == ValueOpcode::GetShaderBase) {
 			if (inst->NumArgs() != 0) {
-				reason = "contains a malformed GetShaderBase";
 				return finish(false);
 			}
 			return finish(true);
@@ -177,14 +169,12 @@ private:
 		if (op == ValueOpcode::Phi) {
 			const auto invariant = ResolveInvariantPhi(m_program, value);
 			if (invariant.IsEmpty()) {
-				reason = "contains a control-dependent phi";
 				return finish(false);
 			}
-			return finish(Validate(invariant, reason));
+			return finish(Validate(invariant));
 		}
 		if (op == ValueOpcode::GetSrtResource) {
 			if (inst->NumArgs() != 0) {
-				reason = "contains a malformed GetSrtResource";
 				return finish(false);
 			}
 			return finish(true);
@@ -196,7 +186,6 @@ private:
 			        ValueOpcode::GetSrtResource ||
 			    !slot.IsImmediate() || slot.GetType() != Type::U32 ||
 			    slot.U32() >= m_program.srt_reads.size()) {
-				reason = "contains a malformed flattened SRT read";
 				return finish(false);
 			}
 		} else if (op == ValueOpcode::LoadAddressU32 || op == ValueOpcode::ReadConstBuffer) {
@@ -206,13 +195,11 @@ private:
 			const auto* handle = inst->NumArgs() != 0 ? inst->Arg(0).ResolveInstruction() : nullptr;
 			if (!IsRawRead(m_program, *inst) || handle == nullptr ||
 			    handle->GetOpcode() != expected) {
-				reason = fmt::format("contains a non-scalar {}", ValueOpcodeName(op));
 				return finish(false);
 			}
 		} else if (op == ValueOpcode::CompositeExtractU64) {
 			const auto index = inst->NumArgs() == 2 ? inst->Arg(1).Resolve() : Value {};
 			if (!index.IsImmediate() || index.GetType() != Type::U32 || index.U32() >= 2u) {
-				reason = "contains an invalid U64 component index";
 				return finish(false);
 			}
 		} else if (op == ValueOpcode::CompositeExtractU32x2) {
@@ -222,7 +209,6 @@ private:
 			    index.U32() >= 2u ||
 			    (source->GetOpcode() != ValueOpcode::CompositeConstructU32x2 &&
 			     source->GetOpcode() != ValueOpcode::IAddCarry32)) {
-				reason = "contains an unsupported composite runtime source";
 				return finish(false);
 			}
 		}
@@ -233,17 +219,14 @@ private:
 			                        : op == ValueOpcode::GetSamplerResource ? 4u
 			                                                                : 2u;
 			if (inst->NumArgs() != expected) {
-				reason = fmt::format("contains a malformed {}", ValueOpcodeName(op));
 				return finish(false);
 			}
 		} else if (op != ValueOpcode::ReadConst && op != ValueOpcode::ReadConstBuffer &&
 		           op != ValueOpcode::LoadAddressU32 && !IsRuntimeIntegerOp(op)) {
-			reason =
-			    fmt::format("contains unsupported or control-dependent {}", ValueOpcodeName(op));
 			return finish(false);
 		}
 		for (size_t index = 0; index < inst->NumArgs(); index++) {
-			if (!Validate(inst->Arg(index), reason)) {
+			if (!Validate(inst->Arg(index))) {
 				return finish(false);
 			}
 		}
@@ -258,7 +241,7 @@ class PlanBuilder {
 public:
 	explicit PlanBuilder(Program& program): m_program(program) {}
 
-	bool Run(std::string* error) {
+	void Run() {
 		m_program.srt_reads.clear();
 		m_program.dynamic_reads.clear();
 		for (auto* block: m_program.blocks) {
@@ -273,34 +256,29 @@ public:
 						                        (op == ValueOpcode::ReadConstBuffer &&
 						                         kind == ResourceKind::ScalarAddress);
 						if (crosswired) {
-							return Fail(flags.pc, error,
-							            fmt::format("{} has incompatible scalar memory metadata",
-							                        ValueOpcodeName(op)));
+							Fail(flags.pc,
+							     fmt::format("{} has incompatible scalar memory metadata",
+							                 ValueOpcodeName(op)));
 						}
 					}
 				}
 				if (IsDescriptorHandle(inst.GetOpcode())) {
 					for (size_t index = 0; index < inst.NumArgs(); index++) {
-						if (!Collect(inst.Arg(index), 0, error)) {
-							return false;
-						}
+						Collect(inst.Arg(index), 0);
 					}
 				}
 			}
 		}
 		for (auto* block: m_program.blocks) {
 			for (auto& inst: *block) {
-				std::string reason;
 				if (inst.GetOpcode() == ValueOpcode::LoadAddressU32 && IsRawRead(m_program, inst) &&
 				    inst.Arg(1).Resolve().IsImmediate() &&
-				    ValidateRuntimeValue(m_program, Value(&inst), reason) &&
-				    !Collect(Value(&inst), inst.Flags<MemoryFlags>().pc, error)) {
-					return false;
+				    ValidateRuntimeValue(m_program, Value(&inst))) {
+					Collect(Value(&inst), inst.Flags<MemoryFlags>().pc);
 				}
 			}
 		}
 		PatchReads();
-		return true;
 	}
 
 private:
@@ -310,18 +288,20 @@ private:
 		bool     keep = false;
 	};
 
-	bool Fail(uint32_t pc, std::string* error, const std::string& message) const {
-		return ShaderError::Fail(error, Diagnostic(m_program, pc, message));
+	[[noreturn]] void Fail(uint32_t pc, const std::string& message) const {
+		const auto diagnostic = Diagnostic(m_program, pc, message);
+		EXIT("shader SRT planning failed: %s", diagnostic.c_str());
+		std::abort();
 	}
 
-	bool Collect(Value value, uint32_t use_pc, std::string* error) {
+	void Collect(Value value, uint32_t use_pc) {
 		value = value.Resolve();
 		if (value.IsImmediate()) {
-			return true;
+			return;
 		}
 		auto* inst = value.TryInstruction();
 		if (inst == nullptr) {
-			return Fail(use_pc, error, "invalid typed planning value");
+			Fail(use_pc, "invalid typed planning value");
 		}
 		const auto cycle = std::ranges::find(m_visiting, inst);
 		if (cycle != m_visiting.end()) {
@@ -329,44 +309,40 @@ private:
 				return value->GetOpcode() == ValueOpcode::Phi;
 			});
 			if (contains_phi) {
-				return true;
+				return;
 			}
-			return Fail(use_pc, error,
-			            fmt::format("cyclic typed planning value {} without a phi",
-			                        ValueOpcodeName(inst->GetOpcode())));
+			Fail(use_pc, fmt::format("cyclic typed planning value {} without a phi",
+			                         ValueOpcodeName(inst->GetOpcode())));
 		}
 		if (std::ranges::find(m_visited, inst) != m_visited.end()) {
-			return true;
+			return;
 		}
 		m_visiting.push_back(inst);
 		for (size_t index = 0; index < inst->NumArgs(); index++) {
-			if (!Collect(inst->Arg(index), use_pc, error)) {
-				return false;
-			}
+			Collect(inst->Arg(index), use_pc);
 		}
 		m_visiting.pop_back();
 		m_visited.push_back(inst);
 		if (!IsRawRead(m_program, *inst)) {
-			return true;
+			return;
 		}
 		const auto offset = inst->Arg(1).Resolve();
 		if (!offset.IsImmediate() || offset.GetType() != Type::U32) {
-			if (std::ranges::find(m_program.dynamic_reads, value) == m_program.dynamic_reads.end()) {
+			if (std::ranges::find(m_program.dynamic_reads, value) ==
+			    m_program.dynamic_reads.end()) {
 				m_program.dynamic_reads.push_back(value);
 			}
-			return true;
+			return;
 		}
-		use_pc = inst->Flags<MemoryFlags>().pc;
 		for (uint32_t slot = 0; slot < m_program.srt_reads.size(); slot++) {
 			if (EquivalentValue(m_program, value, m_program.srt_reads[slot].value)) {
 				m_patches.push_back({inst, slot, false});
-				return true;
+				return;
 			}
 		}
 		const auto slot = static_cast<uint32_t>(m_program.srt_reads.size());
-		m_program.srt_reads.push_back({value, slot, use_pc});
+		m_program.srt_reads.push_back({value, slot});
 		m_patches.push_back({inst, slot, true});
-		return true;
 	}
 
 	void PatchReads() {
@@ -411,14 +387,12 @@ class Evaluator {
 public:
 	Evaluator(const Program& program, const SrtRuntime& runtime,
 	          std::span<const uint8_t> clean_flat_slots = {}, Evaluator* clean_evaluator = nullptr)
-	    : m_program(program), m_runtime(runtime),
-	      m_clean_flat_slots(clean_flat_slots), m_clean_evaluator(clean_evaluator) {}
+	    : m_program(program), m_runtime(runtime), m_clean_flat_slots(clean_flat_slots),
+	      m_clean_evaluator(clean_evaluator) {}
 
-	void SetUsePc(uint32_t pc) { m_use_pc = pc; }
-
-	bool Evaluate(Value value, uint32_t& result, std::string* error) {
+	bool Evaluate(Value value, uint32_t& result) {
 		uint64_t wide = 0;
-		if (!EvaluateWide(value, wide, error)) {
+		if (!EvaluateWide(value, wide)) {
 			return false;
 		}
 		result = static_cast<uint32_t>(wide);
@@ -426,11 +400,7 @@ public:
 	}
 
 private:
-	bool Fail(std::string* error, const std::string& message) const {
-		return ShaderError::Fail(error, Diagnostic(m_program, m_use_pc, message));
-	}
-
-	bool EvaluateWide(Value value, uint64_t& result, std::string* error) {
+	bool EvaluateWide(Value value, uint64_t& result) {
 		value = value.Resolve();
 		if (value.IsImmediate()) {
 			switch (value.GetType()) {
@@ -439,23 +409,23 @@ private:
 				case Type::U16: result = value.U16(); return true;
 				case Type::U32: result = value.U32(); return true;
 				case Type::U64: result = value.U64(); return true;
-				default: return Fail(error, "non-integer immediate in runtime expression");
+				default: return false;
 			}
 		}
 		auto* inst = value.TryInstruction();
 		if (inst == nullptr) {
-			return Fail(error, "invalid typed runtime value");
+			return false;
 		}
 		if (const auto found = m_cache.find(inst); found != m_cache.end()) {
 			result = found->second;
 			return true;
 		}
 		if (std::ranges::find(m_visiting, inst) != m_visiting.end()) {
-			return Fail(error, "cyclic typed runtime value");
+			return false;
 		}
 		m_visiting.push_back(inst);
 		uint64_t out = 0;
-		if (!EvaluateInst(*inst, out, error)) {
+		if (!EvaluateInst(*inst, out)) {
 			return false;
 		}
 		m_visiting.pop_back();
@@ -464,28 +434,27 @@ private:
 		return true;
 	}
 
-	bool Arg(const Inst& inst, size_t index, uint64_t& result, std::string* error) {
-		return EvaluateWide(inst.Arg(index), result, error);
+	bool Arg(const Inst& inst, size_t index, uint64_t& result) {
+		return EvaluateWide(inst.Arg(index), result);
 	}
 
-	bool EvaluatePhi(const Inst& inst, uint64_t& result, std::string* error) {
+	bool EvaluatePhi(const Inst& inst, uint64_t& result) {
 		const auto value = ResolveInvariantPhi(m_program, Value(const_cast<Inst*>(&inst)));
-		return !value.IsEmpty() ? EvaluateWide(value, result, error)
-		                        : Fail(error, "typed phi has runtime-dependent values");
+		return !value.IsEmpty() && EvaluateWide(value, result);
 	}
 
-	bool EvaluateExtract(const Inst& inst, uint64_t& result, std::string* error) {
+	bool EvaluateExtract(const Inst& inst, uint64_t& result) {
 		const auto index = inst.Arg(1).Resolve();
 		if (!index.IsImmediate() || index.GetType() != Type::U32) {
-			return Fail(error, "dynamic composite extract in runtime expression");
+			return false;
 		}
 		const auto component = index.U32();
 		if (component >= 2u) {
-			return Fail(error, "unsupported composite runtime source");
+			return false;
 		}
 		if (inst.GetOpcode() == ValueOpcode::CompositeExtractU64) {
 			uint64_t packed = 0;
-			if (!Arg(inst, 0, packed, error)) {
+			if (!Arg(inst, 0, packed)) {
 				return false;
 			}
 			result = static_cast<uint32_t>(packed >> (component * 32u));
@@ -493,15 +462,15 @@ private:
 		}
 		const auto* source = inst.Arg(0).ResolveInstruction();
 		if (source == nullptr) {
-			return Fail(error, "unsupported composite runtime source");
+			return false;
 		}
 		if (source->GetOpcode() == ValueOpcode::CompositeConstructU32x2) {
-			return EvaluateWide(source->Arg(component), result, error);
+			return EvaluateWide(source->Arg(component), result);
 		}
 		if (source->GetOpcode() == ValueOpcode::IAddCarry32) {
 			uint64_t lhs = 0;
 			uint64_t rhs = 0;
-			if (!Arg(*source, 0, lhs, error) || !Arg(*source, 1, rhs, error)) {
+			if (!Arg(*source, 0, lhs) || !Arg(*source, 1, rhs)) {
 				return false;
 			}
 			const auto sum =
@@ -510,24 +479,23 @@ private:
 			    component == 0u ? static_cast<uint32_t>(sum) : static_cast<uint32_t>(sum >> 32u);
 			return true;
 		}
-		return Fail(error, "unsupported composite runtime source");
+		return false;
 	}
 
-	bool EvaluateRawRead(const Inst& inst, uint64_t& result, std::string* error) {
+	bool EvaluateRawRead(const Inst& inst, uint64_t& result) {
 		const auto flags = inst.Flags<MemoryFlags>();
 		if (flags.index >= m_program.memory_info.size()) {
-			return Fail(error, "raw scalar read has invalid metadata");
+			return false;
 		}
 		const auto& mem    = m_program.memory_info[flags.index];
 		const auto* handle = inst.Arg(0).ResolveInstruction();
 		if (handle == nullptr) {
-			return Fail(error, "raw scalar read has no descriptor handle");
+			return false;
 		}
 		uint64_t low    = 0;
 		uint64_t high   = 0;
 		uint64_t offset = 0;
-		if (!Arg(*handle, 0, low, error) || !Arg(*handle, 1, high, error) ||
-		    !Arg(inst, 1, offset, error)) {
+		if (!Arg(*handle, 0, low) || !Arg(*handle, 1, high) || !Arg(inst, 1, offset)) {
 			return false;
 		}
 		const auto base      = ((high << 32u) | static_cast<uint32_t>(low)) & AddressMask;
@@ -536,12 +504,11 @@ private:
 		if (inst.GetOpcode() == ValueOpcode::ReadConstBuffer) {
 			uint64_t records = 0;
 			uint64_t word3   = 0;
-			if (handle->NumArgs() != 4u || !Arg(*handle, 2, records, error) ||
-			    !Arg(*handle, 3, word3, error)) {
-				return Fail(error, "constant-buffer descriptor has invalid width");
+			if (handle->NumArgs() != 4u || !Arg(*handle, 2, records) || !Arg(*handle, 3, word3)) {
+				return false;
 			}
 			if (immediate < 0) {
-				return Fail(error, "constant-buffer read has a negative immediate offset");
+				return false;
 			}
 			const auto byte_offset =
 			    static_cast<uint64_t>(immediate) + static_cast<uint32_t>(offset);
@@ -551,21 +518,20 @@ private:
 			                      ? static_cast<uint64_t>(static_cast<uint32_t>(records))
 			                      : static_cast<uint64_t>(stride) * static_cast<uint32_t>(records);
 			if (aligned > size || size - aligned < sizeof(uint32_t)) {
-				return Fail(
-				    error, fmt::format("constant-buffer offset {} exceeds size {}", aligned, size));
+				return false;
 			}
 			address = ((base & ~uint64_t {3}) + byte_offset) & ~uint64_t {3};
 		} else {
 			const auto relative = (immediate & ~int64_t {3}) +
 			                      static_cast<int64_t>(static_cast<uint32_t>(offset) & ~3u);
 			if (!AddSignedAddress(base & ~uint64_t {3}, relative, address)) {
-				return Fail(error, "raw scalar read is outside the 48-bit address space");
+				return false;
 			}
 		}
 		uint32_t word = 0;
 		if (m_runtime.read_memory != nullptr) {
 			if (!m_runtime.read_memory(m_runtime.userdata, address, &word)) {
-				return Fail(error, fmt::format("constant read failed at 0x{:016x}", address));
+				return false;
 			}
 		} else {
 			std::memcpy(&word, reinterpret_cast<const void*>(address), sizeof(word));
@@ -574,28 +540,28 @@ private:
 		return true;
 	}
 
-	bool EvaluateInst(const Inst& inst, uint64_t& result, std::string* error) {
+	bool EvaluateInst(const Inst& inst, uint64_t& result) {
 		uint64_t   a       = 0;
 		uint64_t   b       = 0;
 		uint64_t   c       = 0;
-		const auto binary  = [&]() { return Arg(inst, 0, a, error) && Arg(inst, 1, b, error); };
+		const auto binary  = [&]() { return Arg(inst, 0, a) && Arg(inst, 1, b); };
 		const auto ternary = [&]() {
-			return Arg(inst, 0, a, error) && Arg(inst, 1, b, error) && Arg(inst, 2, c, error);
+			return Arg(inst, 0, a) && Arg(inst, 1, b) && Arg(inst, 2, c);
 		};
 		switch (inst.GetOpcode()) {
 			case ValueOpcode::GetUserData: {
 				const auto reg = RegIndex(inst.Arg(0).ScalarRegister());
 				if (reg < m_program.user_data_base ||
 				    reg - m_program.user_data_base >= m_runtime.user_data.size()) {
-					return Fail(error, fmt::format("user SGPR {} is unavailable", reg));
+					return false;
 				}
 				result = m_runtime.user_data[reg - m_program.user_data_base];
 				return true;
 			}
 			case ValueOpcode::GetShaderBase: result = m_runtime.shader_base; return true;
-			case ValueOpcode::Phi: return EvaluatePhi(inst, result, error);
+			case ValueOpcode::Phi: return EvaluatePhi(inst, result);
 			case ValueOpcode::CompositeExtractU64:
-			case ValueOpcode::CompositeExtractU32x2: return EvaluateExtract(inst, result, error);
+			case ValueOpcode::CompositeExtractU32x2: return EvaluateExtract(inst, result);
 			case ValueOpcode::CompositeConstructU64:
 				if (!binary()) {
 					return false;
@@ -607,20 +573,19 @@ private:
 				const auto slot = inst.Arg(1).Resolve();
 				if (!slot.IsImmediate() || slot.GetType() != Type::U32 ||
 				    slot.U32() >= m_program.srt_reads.size()) {
-					return Fail(error, "invalid flattened SRT index");
+					return false;
 				}
 				if (slot.U32() < m_clean_flat_slots.size() &&
 				    m_clean_flat_slots[slot.U32()] != 0u && m_clean_evaluator != nullptr) {
-					m_clean_evaluator->SetUsePc(m_use_pc);
 					return m_clean_evaluator->EvaluateWide(m_program.srt_reads[slot.U32()].value,
-					                                       result, error);
+					                                       result);
 				}
-				return EvaluateWide(m_program.srt_reads[slot.U32()].value, result, error);
+				return EvaluateWide(m_program.srt_reads[slot.U32()].value, result);
 			}
 			case ValueOpcode::LoadAddressU32:
 			case ValueOpcode::ReadConstBuffer:
 				if (IsRawRead(m_program, inst)) {
-					return EvaluateRawRead(inst, result, error);
+					return EvaluateRawRead(inst, result);
 				}
 				break;
 			case ValueOpcode::IAdd32:
@@ -690,7 +655,7 @@ private:
 				}
 				return false;
 			case ValueOpcode::BitwiseNot32:
-				if (Arg(inst, 0, a, error)) {
+				if (Arg(inst, 0, a)) {
 					result = ~static_cast<uint32_t>(a);
 					return true;
 				}
@@ -737,7 +702,7 @@ private:
 					const auto offset = static_cast<uint32_t>(b);
 					const auto width  = static_cast<uint32_t>(c);
 					if (offset > 32u || width > 32u - offset) {
-						return Fail(error, "invalid unsigned bit-field range");
+						return false;
 					}
 					const auto mask = width == 32u  ? UINT32_MAX
 					                  : width == 0u ? 0u
@@ -751,7 +716,7 @@ private:
 					const auto offset = static_cast<uint32_t>(b);
 					const auto width  = static_cast<uint32_t>(c);
 					if (offset > 32u || width > 32u - offset) {
-						return Fail(error, "invalid signed bit-field range");
+						return false;
 					}
 					if (width == 0u) {
 						result = 0;
@@ -768,13 +733,13 @@ private:
 				return false;
 			case ValueOpcode::BitFieldInsert: {
 				uint64_t d = 0;
-				if (!ternary() || !Arg(inst, 3, d, error)) {
+				if (!ternary() || !Arg(inst, 3, d)) {
 					return false;
 				}
 				const auto offset = static_cast<uint32_t>(c);
 				const auto width  = static_cast<uint32_t>(d);
 				if (offset > 32u || width > 32u - offset) {
-					return Fail(error, "invalid inserted bit-field range");
+					return false;
 				}
 				if (width == 0u) {
 					result = static_cast<uint32_t>(a);
@@ -836,7 +801,7 @@ private:
 				}
 				return false;
 			case ValueOpcode::LogicalNot:
-				if (Arg(inst, 0, a, error)) {
+				if (Arg(inst, 0, a)) {
 					result = a == 0u;
 					return true;
 				}
@@ -845,18 +810,16 @@ private:
 			case ValueOpcode::UndefU8:
 			case ValueOpcode::UndefU16:
 			case ValueOpcode::UndefU32:
-			case ValueOpcode::UndefU64: return Fail(error, "undefined typed runtime value");
+			case ValueOpcode::UndefU64: return false;
 			default: break;
 		}
-		return Fail(error, fmt::format("unsupported typed runtime opcode {}",
-		                               ValueOpcodeName(inst.GetOpcode())));
+		return false;
 	}
 
 	const Program&                            m_program;
 	const SrtRuntime&                         m_runtime;
 	std::span<const uint8_t>                  m_clean_flat_slots;
 	Evaluator*                                m_clean_evaluator = nullptr;
-	uint32_t                                  m_use_pc          = 0;
 	std::unordered_map<const Inst*, uint64_t> m_cache;
 	std::vector<const Inst*>                  m_visiting;
 };
@@ -872,18 +835,12 @@ bool EvaluateRuntimeSourcesImpl(const Program&                           program
                                 std::span<const DescriptorSourceRequest> requests,
                                 const SrtRuntime& runtime, std::vector<DescriptorValue>& results,
                                 std::vector<uint32_t>& flat, bool evaluate_flat,
-                                std::span<const uint8_t> clean_flat_slots, std::string* error) {
+                                std::span<const uint8_t> clean_flat_slots) {
 	if (!program.srt_plan_complete) {
-		if (error != nullptr) {
-			*error = Diagnostic(program, 0, "typed SRT plan is not ready");
-		}
 		return false;
 	}
 	if (std::ranges::any_of(clean_flat_slots, [](uint8_t clean) { return clean != 0u; }) &&
 	    runtime.read_specialization_memory == nullptr) {
-		if (error != nullptr) {
-			*error = Diagnostic(program, 0, "clean flattened SRT read has no memory reader");
-		}
 		return false;
 	}
 	SrtRuntime clean_runtime  = runtime;
@@ -895,17 +852,12 @@ bool EvaluateRuntimeSourcesImpl(const Program&                           program
 	for (const auto& request: requests) {
 		const auto* source = Source(program, request.source);
 		if (source == nullptr) {
-			if (error != nullptr) {
-				*error = Diagnostic(program, request.use_pc,
-				                    fmt::format("invalid descriptor source {}", request.source));
-			}
 			return false;
 		}
-		evaluator.SetUsePc(request.use_pc);
 		DescriptorValue value;
 		value.dword_count = source->dword_count;
 		for (uint32_t index = 0; index < source->dword_count; index++) {
-			if (!evaluator.Evaluate(source->dwords[index], value.dwords[index], error)) {
+			if (!evaluator.Evaluate(source->dwords[index], value.dwords[index])) {
 				return false;
 			}
 		}
@@ -918,9 +870,8 @@ bool EvaluateRuntimeSourcesImpl(const Program&                           program
 			const bool clean    = read.flat_offset < clean_flat_slots.size() &&
 			                      clean_flat_slots[read.flat_offset] != 0u;
 			auto&      selected = clean ? clean_evaluator : evaluator;
-			selected.SetUsePc(read.use_pc);
 			if (read.flat_offset >= flattened.size() ||
-			    !selected.Evaluate(read.value, flattened[read.flat_offset], error)) {
+			    !selected.Evaluate(read.value, flattened[read.flat_offset])) {
 				return false;
 			}
 		}
@@ -934,31 +885,24 @@ bool EvaluateRuntimeSourcesImpl(const Program&                           program
 
 } // namespace
 
-bool ValidateRuntimeValue(const Program& program, Value value, std::string& reason) {
-	return RuntimeValidator(program).Run(value, reason);
+bool ValidateRuntimeValue(const Program& program, Value value) {
+	return RuntimeValidator(program).Run(value);
 }
 
-bool BuildSrtPlan(Program& program, std::string* error) {
+void BuildSrtPlan(Program& program) {
 	if (program.resource_tracking_complete) {
-		if (error != nullptr) {
-			*error = "cannot rebuild SRT after resource tracking";
-		}
-		return false;
+		EXIT("shader SRT planning failed: cannot rebuild SRT after resource tracking");
 	}
 	program.srt_plan_complete = false;
-	if (!PlanBuilder(program).Run(error)) {
-		return false;
-	}
+	PlanBuilder(program).Run();
 	program.srt_plan_complete = true;
-	return true;
 }
 
-bool EvaluateDescriptorSource(const Program& program, uint32_t source, uint32_t use_pc,
-                              const SrtRuntime& runtime, DescriptorValue& result,
-                              std::string* error) {
-	const DescriptorSourceRequest request {source, use_pc};
+bool EvaluateDescriptorSource(const Program& program, uint32_t source, const SrtRuntime& runtime,
+                              DescriptorValue& result) {
+	const DescriptorSourceRequest request {source};
 	std::vector<DescriptorValue>  results;
-	if (!EvaluateDescriptorSources(program, std::span {&request, 1}, runtime, results, error)) {
+	if (!EvaluateDescriptorSources(program, std::span {&request, 1}, runtime, results)) {
 		return false;
 	}
 	result = results.front();
@@ -967,26 +911,23 @@ bool EvaluateDescriptorSource(const Program& program, uint32_t source, uint32_t
 
 bool EvaluateDescriptorSources(const Program&                           program,
                                std::span<const DescriptorSourceRequest> requests,
-                               const SrtRuntime& runtime, std::vector<DescriptorValue>& results,
-                               std::string* error) {
+                               const SrtRuntime& runtime, std::vector<DescriptorValue>& results) {
 	std::vector<uint32_t> ignored;
-	return EvaluateRuntimeSourcesImpl(program, requests, runtime, results, ignored, false, {},
-	                                  error);
+	return EvaluateRuntimeSourcesImpl(program, requests, runtime, results, ignored, false, {});
 }
 
 bool EvaluateRuntimeSources(const Program&                           program,
                             std::span<const DescriptorSourceRequest> requests,
                             const SrtRuntime& runtime, std::vector<DescriptorValue>& results,
-                            std::vector<uint32_t>& flat, std::span<const uint8_t> clean_flat_slots,
-                            std::string* error) {
+                            std::vector<uint32_t>&   flat,
+                            std::span<const uint8_t> clean_flat_slots) {
 	return EvaluateRuntimeSourcesImpl(program, requests, runtime, results, flat, true,
-	                                  clean_flat_slots, error);
+	                                  clean_flat_slots);
 }
 
-bool WalkSrt(const Program& program, const SrtRuntime& runtime, std::vector<uint32_t>& flat,
-             std::string* error) {
+bool WalkSrt(const Program& program, const SrtRuntime& runtime, std::vector<uint32_t>& flat) {
 	std::vector<DescriptorValue> ignored;
-	return EvaluateRuntimeSources(program, {}, runtime, ignored, flat, {}, error);
+	return EvaluateRuntimeSources(program, {}, runtime, ignored, flat, {});
 }
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR

src/graphics/shader/recompiler/ir/passes/SrtWalker.h

@@ -13,43 +13,38 @@ using SrtMemoryReader = bool (*)(void* userdata, uint64_t address, uint32_t* val
 
 struct SrtRuntime {
 	std::span<const uint32_t> user_data;
-	uint64_t                  shader_base = 0;
-	SrtMemoryReader           read_memory = nullptr;
-	void*                     userdata    = nullptr;
+	uint64_t                  shader_base                = 0;
+	SrtMemoryReader           read_memory                = nullptr;
+	void*                     userdata                   = nullptr;
 	SrtMemoryReader           read_specialization_memory = nullptr;
 };
 
 struct DescriptorSourceRequest {
 	uint32_t source = 0;
-	uint32_t use_pc = 0;
 };
 
 // Collects reachable ReadConst values. Immediate offsets receive compact flat-buffer slots;
 // dynamic offsets remain explicit and are never assigned a fake slot.
-bool BuildSrtPlan(Program& program, std::string* error);
-bool ValidateRuntimeValue(const Program& program, Value value, std::string& reason);
+void BuildSrtPlan(Program& program);
+bool ValidateRuntimeValue(const Program& program, Value value);
 
-bool EvaluateDescriptorSource(const Program& program, uint32_t source, uint32_t use_pc,
-                              const SrtRuntime& runtime, DescriptorValue& result,
-                              std::string* error);
+bool EvaluateDescriptorSource(const Program& program, uint32_t source, const SrtRuntime& runtime,
+                              DescriptorValue& result);
 
 // Evaluates one runtime snapshot transactionally. Scalar values and ReadConst results shared by
 // several descriptors are memoized once across the batch.
 bool EvaluateDescriptorSources(const Program&                           program,
                                std::span<const DescriptorSourceRequest> requests,
-                               const SrtRuntime& runtime, std::vector<DescriptorValue>& results,
-                               std::string* error);
+                               const SrtRuntime& runtime, std::vector<DescriptorValue>& results);
 
 // Evaluates descriptor sources and the flattened immediate SRT with one memoized scalar walk.
 // On failure neither destination is changed.
 bool EvaluateRuntimeSources(const Program&                           program,
                             std::span<const DescriptorSourceRequest> requests,
                             const SrtRuntime& runtime, std::vector<DescriptorValue>& results,
-                            std::vector<uint32_t>& flat, std::span<const uint8_t> clean_flat_slots,
-                            std::string* error);
+                            std::vector<uint32_t>& flat, std::span<const uint8_t> clean_flat_slots);
 
-bool WalkSrt(const Program& program, const SrtRuntime& runtime, std::vector<uint32_t>& flat,
-             std::string* error);
+bool WalkSrt(const Program& program, const SrtRuntime& runtime, std::vector<uint32_t>& flat);
 
 } // namespace Libs::Graphics::ShaderRecompiler::IR
 

src/graphics/shader/shader.cpp

@@ -25,6 +25,7 @@
 
 #include <algorithm>
 #include <atomic>
+#include <bit>
 #include <cstdio>
 #include <filesystem>
 #include <fmt/format.h>
@@ -173,6 +174,11 @@ static const ShaderBinaryInfo* GetBinaryInfo(const uint32_t* code) {
 	return nullptr;
 }
 
+static uint64_t GetDeclaredShaderHash(uint64_t shader_addr) {
+	const auto* header = GetBinaryInfo(reinterpret_cast<const uint32_t*>(shader_addr));
+	return header != nullptr ? (static_cast<uint64_t>(header->hash1) << 32u) | header->hash0 : 0;
+}
+
 static ShaderParams GetShaderParams(uint64_t shader_addr, const char* label, uint64_t declared_hash,
 	                                std::span<const uint32_t> user_data,
 	                                const ShaderMappedData& data) {
@@ -210,7 +216,6 @@ static void vs_check(const HW::VertexShaderInfo& vs, const HW::ShaderRegisters&
 	};
 	const bool ps5_ngg_passthrough_triangle_path =
 	    vs.es_regs.data_addr != 0 && vs.gs_regs.data_addr == vs.es_regs.data_addr &&
-	    vs.gs_regs.chksum != 0 &&
 	    sh.m_geNggSubgrpCntl == 0x00000001 && sh.m_vgtGsMaxVertOut == 0x00000003 &&
 	    sh.m_vgtGsOutPrimType == 0x00000002 && sh.m_geMaxOutputPerSubgroup <= 0x000000c0;
 
@@ -275,7 +280,6 @@ static void ps_check(const HW::PsStageRegisters& ps, const HW::ShaderRegisters&
 	// EXIT_NOT_IMPLEMENTED(ps.m_spiShaderPgmRsrc1Ps != 0x002c0000);
 	// EXIT_NOT_IMPLEMENTED(ps.m_spiShaderPgmRsrc2Ps != 0x00000000);
 	// EXIT_NOT_IMPLEMENTED(ps.vgprs != 0x00 && ps.vgprs != 0x01);
-	// EXIT_NOT_IMPLEMENTED(ps.sgprs != 0x00 && ps.sgprs != 0x01);
 	EXIT_NOT_IMPLEMENTED(ps.rsrc1.priority != 0);
 	EXIT_NOT_IMPLEMENTED(ps.rsrc1.float_mode != 192);
 	EXIT_NOT_IMPLEMENTED(ps.rsrc1.dx10_clamp != true);
@@ -365,8 +369,10 @@ static void cs_check(const HW::CsStageRegisters& cs, const HW::ShaderRegisters&
 	// EXIT_NOT_IMPLEMENTED(cs.num_thread_y != 0x00000001);
 	// EXIT_NOT_IMPLEMENTED(cs.num_thread_z != 0x00000001);
 	// EXIT_NOT_IMPLEMENTED(cs.vgprs != 0x00 && cs.vgprs != 0x01);
-	// EXIT_NOT_IMPLEMENTED(cs.sgprs != 0x01 && cs.sgprs != 0x02);
-	EXIT_NOT_IMPLEMENTED(cs.bulky != 0x00);
+	EXIT_NOT_IMPLEMENTED(cs.priority != 0x00);
+	EXIT_NOT_IMPLEMENTED(cs.debug_mode != false);
+	EXIT_NOT_IMPLEMENTED(cs.require_forward_progress != false);
+	EXIT_NOT_IMPLEMENTED(cs.shared_vgprs != 0x00);
 	EXIT_NOT_IMPLEMENTED(cs.scratch_en != 0x00);
 	// EXIT_NOT_IMPLEMENTED(cs.user_sgpr != 0x0c);
 	if (cs.tgid_x_en == 0x00) {
@@ -566,7 +572,10 @@ static void ShaderApplyAttribSemantics(ShaderVertexInputInfo& info,
 		uint32_t reg  = in.hardware_mapping;
 		uint32_t size = in.size_in_elements;
 
-		LOGF("reg = %u, size = %u, va[%u] = 0x%08" PRIx32 "\n", reg, size, i, attrib[in.semantic]);
+		if (Config::GraphicsDebugDumpEnabled()) {
+			LOGF("reg = %u, size = %u, va[%u] = 0x%08" PRIx32 "\n", reg, size, i,
+			     attrib[in.semantic]);
+		}
 
 		size_t index = attrib[in.semantic] & 0x1fu;
 		auto   format =
@@ -689,7 +698,7 @@ static bool ShaderGetStaticInputInfoVS(const HW::VertexShaderInfo& regs,
 	info.export_count      = static_cast<int>(sh.GetExportCount());
 	info.pa_cl_vs_out_cntl = sh.m_paClVsOutCntl;
 
-	EXIT_NOT_IMPLEMENTED(regs.es_regs.data_addr == 0 || regs.gs_regs.chksum == 0);
+	EXIT_NOT_IMPLEMENTED(regs.es_regs.data_addr == 0);
 
 	uint64_t                shader_addr   = regs.es_regs.data_addr;
 	const HW::UserSgprInfo& user_sgpr     = regs.gs_user_sgpr;
@@ -698,8 +707,8 @@ static bool ShaderGetStaticInputInfoVS(const HW::VertexShaderInfo& regs,
 
 	if (data.user_data == nullptr) {
 		LOGF("ShaderGetInputInfoVS(): no AGC user data for shader=0x%016" PRIx64 " es=0x%016" PRIx64
-		     " gs=0x%016" PRIx64 " chksum=0x%016" PRIx64 " user_sgpr_num=%u\n",
-		     shader_addr, regs.es_regs.data_addr, regs.gs_regs.data_addr, regs.gs_regs.chksum,
+		     " gs=0x%016" PRIx64 " user_sgpr_num=%u\n",
+		     shader_addr, regs.es_regs.data_addr, regs.gs_regs.data_addr,
 		     static_cast<uint32_t>(user_sgpr_num));
 	}
 	ShaderVertexMetadata metadata;
@@ -841,23 +850,22 @@ static void ApplyPixelOutputs(ShaderPixelInputInfo&                info,
 	}
 }
 
-static bool LogPermutationMismatch(const ShaderRecompiler::IR::Program& program,
-                                   const char* stage, const std::string& error) {
+static bool LogPermutationMismatch(const ShaderRecompiler::IR::Program* program,
+                                   const char* stage) {
 	static std::atomic<uint32_t> log_count {0};
 	if (Config::GraphicsDebugDumpEnabled() &&
 	    log_count.fetch_add(1, std::memory_order_relaxed) < 64) {
-		LOGF("ShaderProgramCache native runtime mismatch %s shader=0x%016" PRIx64 ": %s\n",
-		     stage, program.shader_hash, error.c_str());
+		LOGF("ShaderProgramCache native runtime mismatch %s shader=0x%016" PRIx64 "\n",
+		     stage, program != nullptr ? program->shader_hash : 0);
 	}
 	return false;
 }
 
 bool MaterializeProgram(const std::shared_ptr<const ShaderRecompiler::IR::Program>& program,
                         const ShaderParams& params, ShaderVertexInputInfo& info) {
-	std::string error;
-	if (!ShaderMaterializeStageRuntime(program, params.user_data, params.Base(), info.stage, &error,
+	if (!ShaderMaterializeStageRuntime(program, params.user_data, params.Base(), info.stage,
 	                                   ReadShaderGuestMemory)) {
-		return LogPermutationMismatch(*program, "VS", error);
+		return LogPermutationMismatch(program.get(), "VS");
 	}
 	ApplyVertexOutputs(info, *program);
 	return true;
@@ -865,10 +873,9 @@ bool MaterializeProgram(const std::shared_ptr<const ShaderRecompiler::IR::Progra
 
 bool MaterializeProgram(const std::shared_ptr<const ShaderRecompiler::IR::Program>& program,
                         const ShaderParams& params, ShaderPixelInputInfo& info) {
-	std::string error;
-	if (!ShaderMaterializeStageRuntime(program, params.user_data, params.Base(), info.stage, &error,
+	if (!ShaderMaterializeStageRuntime(program, params.user_data, params.Base(), info.stage,
 	                                   ReadShaderGuestMemory)) {
-		return LogPermutationMismatch(*program, "PS", error);
+		return LogPermutationMismatch(program.get(), "PS");
 	}
 	ApplyPixelOutputs(info, *program);
 	return true;
@@ -876,10 +883,9 @@ bool MaterializeProgram(const std::shared_ptr<const ShaderRecompiler::IR::Progra
 
 bool MaterializeProgram(const std::shared_ptr<const ShaderRecompiler::IR::Program>& program,
                         const ShaderParams& params, ShaderComputeInputInfo& info) {
-	std::string error;
-	if (!ShaderMaterializeStageRuntime(program, params.user_data, params.Base(), info.stage, &error,
+	if (!ShaderMaterializeStageRuntime(program, params.user_data, params.Base(), info.stage,
 	                                   ReadShaderGuestMemory)) {
-		return LogPermutationMismatch(*program, "CS", error);
+		return LogPermutationMismatch(program.get(), "CS");
 	}
 	return true;
 }
@@ -892,6 +898,18 @@ void BuildStageStaticKey(const ShaderVertexInputInfo& info, std::vector<uint32_t
 	key.push_back(info.resources_num);
 	key.push_back(info.scratch_size_dwords);
 	key.push_back(info.pa_cl_vs_out_cntl);
+	key.push_back(static_cast<uint32_t>(info.clip_space.enabled));
+	if (info.clip_space.enabled) {
+		for (const float value: info.clip_space.scale) {
+			key.push_back(std::bit_cast<uint32_t>(value));
+		}
+		for (const float value: info.clip_space.offset) {
+			key.push_back(std::bit_cast<uint32_t>(value));
+		}
+		for (const float value: info.clip_space.half_extent) {
+			key.push_back(std::bit_cast<uint32_t>(value));
+		}
+	}
 
 	for (int i = 0; i < info.resources_num; i++) {
 		const auto& resource    = info.resources[i];
@@ -982,7 +1000,7 @@ ShaderParams PrepareProgram(const HW::VertexShaderInfo& regs, const HW::ShaderRe
 		EXIT("failed to prepare vertex shader program\n");
 	}
 	return GetShaderParams(
-	    regs.es_regs.data_addr, "ShaderRecompiler VS", regs.gs_regs.chksum,
+	    regs.es_regs.data_addr, "ShaderRecompiler VS", GetDeclaredShaderHash(regs.es_regs.data_addr),
 	    std::span<const uint32_t>(regs.gs_user_sgpr.value, regs.gs_regs.rsrc2.user_sgpr), data);
 }
 
@@ -994,7 +1012,7 @@ ShaderParams PrepareProgram(
 	const auto data = ShaderGetMappedData(regs.ps_regs.data_addr, "ShaderGetInputInfoPS():");
 	ShaderGetStaticInputInfoPS(regs, sh, vs_info, target_export_mapping, data, ps_info);
 	return GetShaderParams(
-	    regs.ps_regs.data_addr, "ShaderRecompiler PS", regs.ps_regs.chksum,
+	    regs.ps_regs.data_addr, "ShaderRecompiler PS", GetDeclaredShaderHash(regs.ps_regs.data_addr),
 	    std::span<const uint32_t>(regs.ps_user_sgpr.value, regs.ps_regs.rsrc2.user_sgpr), data);
 }
 
@@ -1002,11 +1020,8 @@ ShaderParams PrepareProgram(const HW::ComputeShaderInfo& regs, const HW::ShaderR
                             ShaderComputeInputInfo& info) {
 	const auto data = ShaderGetMappedData(regs.cs_regs.data_addr, "ShaderGetInputInfoCS():");
 	ShaderGetStaticInputInfoCS(regs, sh, data, info);
-	const auto* header = GetBinaryInfo(reinterpret_cast<const uint32_t*>(regs.cs_regs.data_addr));
-	const auto declared_hash =
-	    header != nullptr ? (static_cast<uint64_t>(header->hash1) << 32u) | header->hash0 : 0;
 	return GetShaderParams(
-	    regs.cs_regs.data_addr, "ShaderRecompiler CS", declared_hash,
+	    regs.cs_regs.data_addr, "ShaderRecompiler CS", GetDeclaredShaderHash(regs.cs_regs.data_addr),
 	    std::span<const uint32_t>(regs.cs_user_sgpr.value, regs.cs_regs.user_sgpr), data);
 }
 
@@ -1237,9 +1252,7 @@ static ShaderRecompiler::CompileOptions MakeCompileOptions(const ShaderParams& p
 	ShaderRecompiler::CompileOptions options;
 	options.stage                      = stage;
 	options.shader_hash                = params.hash;
-	options.shader_base                = params.Base();
-	options.user_data_count            = static_cast<uint32_t>(params.user_data.size());
-	options.user_data                  = params.user_data.data();
+	options.user_data                  = params.user_data;
 	options.read_specialization_memory = ReadShaderGuestMemory;
 	options.dump_ir                    = ShaderRecompilerTextDumpEnabled();
 	options.early_dump                 = options.dump_ir;
@@ -1252,11 +1265,7 @@ static vk::ShaderModule CompileModule(vk::Device device, const ShaderParams& par
                                       ShaderStageRuntime& stage) {
 	const auto* stage_name = ShaderStageName(options.stage);
 	const auto* label      = ShaderStageLabel(options.stage);
-	ShaderRecompiler::CompileResult result;
-	std::string                     error;
-	if (!ShaderRecompiler::TryRecompile(params.code, options, result, &error)) {
-		ExitShaderRecompilerFailure(label, options.shader_hash, error.c_str());
-	}
+	auto result = ShaderRecompiler::Recompile(params.code, options);
 
 	DumpShaderRecompilerOriginal(stage_name, options.shader_hash, params.code, result.decoded_dump);
 	if (!SpirvValidateBinary(label, options.shader_hash, result.spirv)) {

src/graphics/shader/shader.h

@@ -57,7 +57,7 @@ using ShaderSpecializationMemoryReader = bool (*)(void* userdata, uint64_t addre
 bool ShaderMaterializeStageRuntime(
     std::shared_ptr<const ShaderRecompiler::IR::Program> program,
     std::span<const uint32_t> user_data, uint64_t shader_base, ShaderStageRuntime& stage,
-    std::string* error, ShaderSpecializationMemoryReader read_specialization_memory = nullptr,
+    ShaderSpecializationMemoryReader read_specialization_memory = nullptr,
     void* read_memory_data = nullptr);
 
 constexpr uint32_t DstSel(uint32_t x, uint32_t y = 0, uint32_t z = 0, uint32_t w = 0) {
@@ -68,6 +68,13 @@ inline uint8_t GetDstSel(uint32_t swizzle, uint32_t channel) {
 	return (swizzle >> (channel * 3u)) & 0x7u;
 }
 
+struct ShaderClipSpaceTransform {
+	float scale[2]       = {};
+	float offset[2]      = {};
+	float half_extent[2] = {};
+	bool  enabled        = false;
+};
+
 struct ShaderVertexInputInfo {
 	static constexpr int RES_MAX = 32;
 
@@ -84,6 +91,7 @@ struct ShaderVertexInputInfo {
 	uint32_t                scratch_size_dwords = 0;
 	uint32_t                param_export_mask   = 0;
 	uint32_t                pa_cl_vs_out_cntl    = 0;
+	ShaderClipSpaceTransform clip_space;
 	bool                    fetch_external      = false;
 	bool                    fetch_embedded      = false;
 };

src/graphics/shader/shaderStageRuntime.cpp

@@ -7,13 +7,10 @@ namespace Libs::Graphics {
 
 bool ShaderMaterializeStageRuntime(std::shared_ptr<const ShaderRecompiler::IR::Program> program,
                                    std::span<const uint32_t> user_data, uint64_t shader_base,
-                                   ShaderStageRuntime& stage, std::string* error,
+                                   ShaderStageRuntime& stage,
                                    ShaderSpecializationMemoryReader read_specialization_memory,
                                    void*                            read_memory_data) {
 	if (program == nullptr) {
-		if (error != nullptr) {
-			*error = "missing native shader plan";
-		}
 		return false;
 	}
 	ShaderRecompiler::IR::SrtRuntime runtime;
@@ -22,8 +19,8 @@ bool ShaderMaterializeStageRuntime(std::shared_ptr<const ShaderRecompiler::IR::P
 	runtime.read_specialization_memory = read_specialization_memory;
 	runtime.userdata                   = read_memory_data;
 	ShaderRecompiler::IR::ResourceSnapshot snapshot;
-	if (!ShaderRecompiler::IR::MaterializeResources(*program, runtime, snapshot, error) ||
-	    !ShaderRecompiler::IR::ValidateResourceSpecialization(*program, snapshot, error)) {
+	if (!ShaderRecompiler::IR::MaterializeResources(*program, runtime, snapshot) ||
+	    !ShaderRecompiler::IR::ValidateResourceSpecialization(*program, snapshot)) {
 		return false;
 	}
 	auto resources =

src/kernel/fileSystem.cpp

@@ -1066,7 +1066,7 @@ int KYTY_SYSV_ABI KernelRename(const char* from, const char* to) {
 		Common::File::DeleteFile(real_to);
 	}
 
-	if (!Common::File::MoveFile(real_from, real_to)) {
+	if (!Common::File::RenameFile(real_from, real_to)) {
 		return KERNEL_ERROR_EIO;
 	}
 

src/kernel/memory.cpp

@@ -3804,8 +3804,8 @@ int KYTY_SYSV_ABI KernelBatchMap2(KernelBatchMapEntry* entries, int num_entries,
 		LOGF("\t [%d] start = %p, offset = 0x%016" PRIx64 ", length = 0x%016" PRIx64
 		     ", prot = 0x%02" PRIx32 ", type = 0x%02" PRIx32 ", op = %d\n",
 		     i, entry->start, entry->offset, entry->length,
-		     static_cast<uint32_t>(static_cast<unsigned char>(entry->protection)),
-		     static_cast<uint32_t>(static_cast<unsigned char>(entry->type)), entry->operation);
+		     static_cast<uint32_t>(entry->protection), static_cast<uint32_t>(entry->type),
+		     entry->operation);
 
 		if (entry->length == 0 || entry->operation < MAP_OP_MAP_DIRECT ||
 		    entry->operation > MAP_OP_TYPE_PROTECT) {

src/kernel/memory.h

@@ -47,13 +47,13 @@ struct VirtualQueryInfo {
 static_assert(sizeof(VirtualQueryInfo) == 72, "VirtualQueryInfo struct size is incorrect");
 
 struct KernelBatchMapEntry {
-	void*    start;
-	uint64_t offset;
-	uint64_t length;
-	char     protection;
-	char     type;
-	int16_t  reserved;
-	int32_t  operation;
+	void*         start;
+	uint64_t      offset;
+	uint64_t      length;
+	unsigned char protection;
+	unsigned char type;
+	int16_t       reserved;
+	int32_t       operation;
 };
 
 static_assert(sizeof(KernelBatchMapEntry) == 32, "KernelBatchMapEntry struct size is incorrect");

src/kernel/memoryAddressSpace.inc

@@ -44,9 +44,9 @@ public:
 
 	explicit GuestBackingStore(uint64_t size): m_size(size) {
 #if KYTY_PLATFORM == KYTY_PLATFORM_WINDOWS
-		m_handle = CreateFileMappingA(INVALID_HANDLE_VALUE, nullptr, PAGE_EXECUTE_READWRITE,
-		                              static_cast<DWORD>(m_size >> 32u),
-		                              static_cast<DWORD>(m_size & 0xffffffffu), nullptr);
+		m_handle = CreateFileMappingA(
+		    INVALID_HANDLE_VALUE, nullptr, PAGE_EXECUTE_READWRITE | SEC_COMMIT,
+		    static_cast<DWORD>(m_size >> 32u), static_cast<DWORD>(m_size & 0xffffffffu), nullptr);
 		if (m_handle == nullptr) {
 			LOGF_COLOR(Log::Color::Red,
 			           "\t direct-memory backing: CreateFileMapping failed: 0x%08" PRIx32 "\n",
@@ -659,7 +659,16 @@ public:
 #endif
 		ReserveGuestRegions();
 		m_backing = new GuestBackingStore(backing_size);
-		EXIT_IF(!m_backing->IsAvailable());
+		if (!m_backing->IsAvailable()) {
+#if KYTY_PLATFORM == KYTY_PLATFORM_WINDOWS
+			EXIT("could not reserve %" PRIu64 " MB for guest direct memory. Windows commits\n"
+			     "this up front, so the paging file is usually what needs to be larger.\n",
+			     backing_size / (1024U * 1024U));
+#else
+			EXIT("could not reserve %" PRIu64 " MB for guest direct memory\n",
+			     backing_size / (1024U * 1024U));
+#endif
+		}
 	}
 
 	~GuestAddressSpace() {

src/kytyGitVersion.h.in

@@ -1,7 +1,10 @@
 #include "cmake_config.h"
 
 #define KYTY_GIT_VERSION "@KYTY_GIT_VERSION@"
+// Short, human-facing commit name used in build labels.
 #define KYTY_GIT_HASH "@KYTY_GIT_HASH@"
+// Full commit identity used for compatibility checks.
+#define KYTY_GIT_REVISION "@KYTY_GIT_REVISION@"
 
 /* Build provenance, shown in the window title and in crash reports. */
 #if defined(KYTY_OFFICIAL_BUILD)

src/launcher/CMakeLists.txt

@@ -66,10 +66,46 @@ if(WIN32)
 endif()
 
 set(launcher_name "launcher")
+if(APPLE)
+	set(launcher_name "KytyPS5")
+endif()
 
 set_target_properties(launcher PROPERTIES OUTPUT_NAME ${launcher_name})
 
-install(TARGETS launcher DESTINATION .)
+if(APPLE)
+	# macOS bundle: KytyPS5.app with launcher as main executable and
+	# kyty_emulator as helper inside Contents/MacOS. MoltenVK is found
+	# next to a flat executable or in the bundle's Contents/Frameworks;
+	# no rpath fixup is needed because SDL loads it by absolute path.
+	# Qt deployment (Frameworks + plugins) is handled by macdeployqt.
+	set_target_properties(launcher PROPERTIES
+		MACOSX_BUNDLE TRUE
+		MACOSX_BUNDLE_INFO_PLIST "${CMAKE_CURRENT_SOURCE_DIR}/Info.plist.in"
+		INSTALL_RPATH "@executable_path/../Frameworks"
+		BUILD_WITH_INSTALL_RPATH TRUE
+	)
+endif()
+
+install(TARGETS launcher
+	BUNDLE DESTINATION .
+	RUNTIME DESTINATION .)
+
+if(APPLE)
+	install(TARGETS kyty_emulator DESTINATION "KytyPS5.app/Contents/MacOS")
+	# Deploy and ad-hoc sign the Qt frameworks and plugins in the bundle.
+	find_program(KYTY_MACDEPLOYQT NAMES macdeployqt PATHS "${Qt6_DIR}/../../../bin" NO_DEFAULT_PATH)
+	if(NOT KYTY_MACDEPLOYQT)
+		message(FATAL_ERROR "Could not find macdeployqt")
+	endif()
+	install(CODE "
+		execute_process(COMMAND \"${KYTY_MACDEPLOYQT}\" \"\${CMAKE_INSTALL_PREFIX}/KytyPS5.app\"
+			-codesign=-
+			RESULT_VARIABLE _kyty_deploy_result)
+		if(NOT _kyty_deploy_result EQUAL 0)
+			message(FATAL_ERROR \"macdeployqt failed with \${_kyty_deploy_result}\")
+		endif()
+	")
+endif()
 
 if(WIN32)
     install(CODE "

src/launcher/Info.plist.in

@@ -0,0 +1,17 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+	<key>CFBundleName</key><string>KytyPS5</string>
+	<key>CFBundleDisplayName</key><string>KytyPS5</string>
+	<key>CFBundleIdentifier</key><string>org.kytyps5.kytyps5</string>
+	<key>CFBundleVersion</key><string>@PROJECT_VERSION@</string>
+	<key>CFBundleShortVersionString</key><string>@PROJECT_VERSION@</string>
+	<key>CFBundlePackageType</key><string>APPL</string>
+	<key>CFBundleSignature</key><string>????</string>
+	<key>CFBundleExecutable</key><string>KytyPS5</string>
+	<key>LSMinimumSystemVersion</key><string>13.0</string>
+	<key>NSHighResolutionCapable</key><true/>
+	<key>NSRequiresAquaSystemAppearance</key><false/>
+</dict>
+</plist>

src/launcher/include/inputMappingDialog.h

@@ -5,6 +5,7 @@
 #include <QStringList>
 
 class QPushButton;
+class QDoubleSpinBox;
 class QTreeWidget;
 class QTreeWidgetItem;
 class QWidget;
@@ -22,10 +23,11 @@ private:
 	void SetBinding(QTreeWidgetItem* item, const QString& binding);
 	void UpdateButtons();
 
-	QTreeWidget* m_bindings      = nullptr;
-	QPushButton* m_change_button = nullptr;
-	QPushButton* m_clear_button  = nullptr;
-	bool         m_customized    = false;
+	QTreeWidget*    m_bindings        = nullptr;
+	QPushButton*    m_change_button   = nullptr;
+	QPushButton*    m_clear_button    = nullptr;
+	QDoubleSpinBox* m_sensitivity     = nullptr;
+	bool            m_custom_bindings = false;
 };
 
 #endif /* LAUNCHER_INCLUDE_INPUT_MAPPING_DIALOG_H_ */

src/launcher/src/configurationListWidget.cpp

@@ -889,7 +889,7 @@ void ConfigurationListWidget::show_context_menu(const QPoint& pos) {
 	    style()->standardIcon(QStyle::SP_FileDialogContentsView), tr("View trophies..."));
 	connect(action_view_trophies, &QAction::triggered, this,
 	        &ConfigurationListWidget::ViewTrophies);
-	QAction* action_patches = menu.addAction(tr("Patches (experimental)..."));
+	QAction* action_patches = menu.addAction(tr("Cheats (experimental)..."));
 	connect(action_patches, &QAction::triggered, this, [this, item]() {
 		if (item != nullptr) {
 			auto* dialog = new PatchesDialog(item->GetInfo(), this);

src/launcher/src/inputMappingDialog.cpp

@@ -2,6 +2,7 @@
 
 #include <QAbstractItemView>
 #include <QDialogButtonBox>
+#include <QDoubleSpinBox>
 #include <QHBoxLayout>
 #include <QHash>
 #include <QHeaderView>
@@ -16,7 +17,9 @@
 
 namespace {
 
-constexpr int BINDING_COLUMN = 1;
+constexpr int  BINDING_COLUMN            = 1;
+constexpr auto DEFAULT_MOUSE_SENSITIVITY = 1.0;
+constexpr char MOUSE_SENSITIVITY[]       = "MouseSensitivity=";
 
 struct PadControl {
 	const char* id;
@@ -48,7 +51,8 @@ constexpr PadControl PAD_CONTROLS[] = {
     {"L3", "L3", "Left Shift"},
     {"R3", "R3", "Left Ctrl"},
     {"Options", "Options", "Return"},
-    {"TouchPad", "Touch pad", "Backspace"},
+    {"TouchPad", "Touch pad left (SELECT)", "Backspace"},
+    {"TouchPadRight", "Touch pad right (START)", "Tab"},
 };
 
 QString KeypadName(int key) {
@@ -128,7 +132,8 @@ public:
 
 		auto* layout = new QVBoxLayout(this);
 		m_label      = new QLabel(
-		    tr("Press a key or mouse button.\nSpace and F1 are reserved; Esc cancels."), this);
+		    tr("Press a key or mouse button.\nSpace, F1, F7, and F11 are reserved; Esc cancels."),
+		    this);
 		m_label->setAlignment(Qt::AlignCenter);
 		layout->addWidget(m_label);
 	}
@@ -144,7 +149,8 @@ protected:
 			reject();
 			return;
 		}
-		if (event->key() == Qt::Key_Space || event->key() == Qt::Key_F1) {
+		if (event->key() == Qt::Key_Space || event->key() == Qt::Key_F1 ||
+		    event->key() == Qt::Key_F7 || event->key() == Qt::Key_F11) {
 			m_label->setText(tr("That key is reserved by the emulator."));
 			return;
 		}
@@ -177,6 +183,9 @@ private:
 QHash<QString, QString> ParseMapping(const QStringList& mapping) {
 	QHash<QString, QString> result;
 	for (const auto& entry: mapping) {
+		if (entry.startsWith(QLatin1String(MOUSE_SENSITIVITY))) {
+			continue;
+		}
 		const auto separator = entry.indexOf(QLatin1Char('='));
 		if (separator > 0 && separator + 1 < entry.size()) {
 			const auto binding = entry.mid(separator + 1);
@@ -193,15 +202,42 @@ QHash<QString, QString> ParseMapping(const QStringList& mapping) {
 	return result;
 }
 
+double ParseMouseSensitivity(const QStringList& mapping) {
+	for (const auto& entry: mapping) {
+		if (entry.startsWith(QLatin1String(MOUSE_SENSITIVITY))) {
+			return entry.mid(sizeof(MOUSE_SENSITIVITY) - 1).toDouble();
+		}
+	}
+	return DEFAULT_MOUSE_SENSITIVITY;
+}
+
 } // namespace
 
 InputMappingDialog::InputMappingDialog(const QStringList& mapping, QWidget* parent)
-    : QDialog(parent), m_customized(!mapping.isEmpty()) {
+    : QDialog(parent) {
 	setWindowTitle(tr("Input Mapping"));
 	resize(460, 650);
 
 	auto* layout = new QVBoxLayout(this);
-	layout->addWidget(new QLabel(tr("Map keyboard or mouse input to DualSense controls."), this));
+	layout->addWidget(
+	    new QLabel(tr("Map keyboard or mouse buttons to DualSense controls.\n"
+	                  "Press F7 in-game to toggle mouse movement on the right stick."),
+	               this));
+
+	const auto parsed = ParseMapping(mapping);
+	m_custom_bindings = !parsed.isEmpty();
+
+	auto* sensitivity_layout = new QHBoxLayout;
+	sensitivity_layout->addWidget(new QLabel(tr("Mouse sensitivity"), this));
+	m_sensitivity = new QDoubleSpinBox(this);
+	m_sensitivity->setRange(0.1, 5.0);
+	m_sensitivity->setSingleStep(0.1);
+	m_sensitivity->setDecimals(1);
+	m_sensitivity->setSuffix(QStringLiteral("x"));
+	m_sensitivity->setValue(ParseMouseSensitivity(mapping));
+	sensitivity_layout->addWidget(m_sensitivity);
+	sensitivity_layout->addStretch();
+	layout->addLayout(sensitivity_layout);
 
 	m_bindings = new QTreeWidget(this);
 	m_bindings->setColumnCount(2);
@@ -212,13 +248,12 @@ InputMappingDialog::InputMappingDialog(const QStringList& mapping, QWidget* pare
 	m_bindings->header()->setSectionResizeMode(1, QHeaderView::Stretch);
 	layout->addWidget(m_bindings);
 
-	const auto parsed = ParseMapping(mapping);
 	for (const auto& control: PAD_CONTROLS) {
 		auto* item = new QTreeWidgetItem(m_bindings);
 		item->setText(0, tr(control.label));
 		item->setData(0, Qt::UserRole, QString::fromLatin1(control.id));
-		SetBinding(item, m_customized ? parsed.value(QString::fromLatin1(control.id))
-		                              : QString::fromLatin1(control.default_binding));
+		SetBinding(item, m_custom_bindings ? parsed.value(QString::fromLatin1(control.id))
+		                                   : QString::fromLatin1(control.default_binding));
 	}
 	m_bindings->setCurrentItem(m_bindings->topLevelItem(0));
 
@@ -248,16 +283,19 @@ InputMappingDialog::InputMappingDialog(const QStringList& mapping, QWidget* pare
 
 QStringList InputMappingDialog::Mapping() const {
 	QStringList result;
-	if (!m_customized) {
-		return result;
-	}
-	for (int index = 0; index < m_bindings->topLevelItemCount(); index++) {
-		const auto* item    = m_bindings->topLevelItem(index);
-		const auto  binding = item->data(BINDING_COLUMN, Qt::UserRole).toString();
-		if (!binding.isEmpty()) {
-			result.append(item->data(0, Qt::UserRole).toString() + QLatin1Char('=') + binding);
+	if (m_custom_bindings) {
+		for (int index = 0; index < m_bindings->topLevelItemCount(); index++) {
+			const auto* item    = m_bindings->topLevelItem(index);
+			const auto  binding = item->data(BINDING_COLUMN, Qt::UserRole).toString();
+			if (!binding.isEmpty()) {
+				result.append(item->data(0, Qt::UserRole).toString() + QLatin1Char('=') + binding);
+			}
 		}
 	}
+	if (m_sensitivity->value() != DEFAULT_MOUSE_SENSITIVITY) {
+		result.append(QLatin1String(MOUSE_SENSITIVITY) +
+		              QString::number(m_sensitivity->value(), 'f', 1));
+	}
 	return result;
 }
 
@@ -281,12 +319,12 @@ void InputMappingDialog::ChangeBinding() {
 		}
 	}
 	SetBinding(item, dialog.Binding());
-	m_customized = true;
+	m_custom_bindings = true;
 }
 
 void InputMappingDialog::ClearBinding() {
 	SetBinding(m_bindings->currentItem(), {});
-	m_customized = true;
+	m_custom_bindings = true;
 }
 
 void InputMappingDialog::RestoreDefaults() {
@@ -294,7 +332,8 @@ void InputMappingDialog::RestoreDefaults() {
 		SetBinding(m_bindings->topLevelItem(index),
 		           QString::fromLatin1(PAD_CONTROLS[index].default_binding));
 	}
-	m_customized = false;
+	m_sensitivity->setValue(DEFAULT_MOUSE_SENSITIVITY);
+	m_custom_bindings = false;
 }
 
 void InputMappingDialog::SetBinding(QTreeWidgetItem* item, const QString& binding) {