[DirectX] Add DXIL_MODULE_FLAG for ShaderFlags. (#83217)

Add DXIL module flag bit offset for SHADER_FEATURE_FLAG.

Added DXIL_MODULE_FLAG for DXIL module flag which does not have feature flag.

Use DXILModuleFlags for ComputedShaderFlags instead of
ShaderFeatureFlags.

ComputedShaderFlags::getFeatureFlags() was added to get FeatureFlags.

Rename DXContainerGlobals::getShaderFlags to DXContainerGlobals::getFeatureFlags.

Fixes #57925

Author: python3kgae

llvm/include/llvm/BinaryFormat/DXContainer.h

@@ -141,7 +141,7 @@ enum class PartType {
 #include "DXContainerConstants.def"
 };
 
-#define SHADER_FEATURE_FLAG(Num, Val, Str) Val = 1ull << Num,
+#define SHADER_FEATURE_FLAG(Num, DxilModuleNum, Val, Str) Val = 1ull << Num,
 enum class FeatureFlags : uint64_t {
 #include "DXContainerConstants.def"
 };

llvm/include/llvm/BinaryFormat/DXContainerConstants.def

@@ -9,46 +9,63 @@ CONTAINER_PART(OSG1)
 CONTAINER_PART(PSG1)
 
 #undef CONTAINER_PART
-#endif 
+#endif // CONTAINER_PART
 
 #ifdef SHADER_FEATURE_FLAG
 
-SHADER_FEATURE_FLAG(0, Doubles, "Double-precision floating point")
-SHADER_FEATURE_FLAG(1, ComputeShadersPlusRawAndStructuredBuffers, "Raw and Structured buffers")
-SHADER_FEATURE_FLAG(2, UAVsAtEveryStage, "UAVs at every shader stage")
-SHADER_FEATURE_FLAG(3, Max64UAVs, "64 UAV slots")
-SHADER_FEATURE_FLAG(4, MinimumPrecision, "Minimum-precision data types")
-SHADER_FEATURE_FLAG(5, DX11_1_DoubleExtensions, "Double-precision extensions for 11.1")
-SHADER_FEATURE_FLAG(6, DX11_1_ShaderExtensions, "Shader extensions for 11.1")
-SHADER_FEATURE_FLAG(7, LEVEL9ComparisonFiltering, "Comparison filtering for feature level 9")
-SHADER_FEATURE_FLAG(8, TiledResources, "Tiled resources")
-SHADER_FEATURE_FLAG(9, StencilRef, "PS Output Stencil Ref")
-SHADER_FEATURE_FLAG(10, InnerCoverage, "PS Inner Coverage")
-SHADER_FEATURE_FLAG(11, TypedUAVLoadAdditionalFormats, "Typed UAV Load Additional Formats")
-SHADER_FEATURE_FLAG(12, ROVs, "Raster Ordered UAVs")
-SHADER_FEATURE_FLAG(13, ViewportAndRTArrayIndexFromAnyShaderFeedingRasterizer, "SV_RenderTargetArrayIndex or SV_ViewportArrayIndex from any shader feeding rasterizer")
-SHADER_FEATURE_FLAG(14, WaveOps, "Wave level operations")
-SHADER_FEATURE_FLAG(15, Int64Ops, "64-Bit integer")
-SHADER_FEATURE_FLAG(16, ViewID, "View Instancing")
-SHADER_FEATURE_FLAG(17, Barycentrics, "Barycentrics")
-SHADER_FEATURE_FLAG(18, NativeLowPrecision, "Use native low precision")
-SHADER_FEATURE_FLAG(19, ShadingRate, "Shading Rate")
-SHADER_FEATURE_FLAG(20, Raytracing_Tier_1_1, "Raytracing tier 1.1 features")
-SHADER_FEATURE_FLAG(21, SamplerFeedback, "Sampler feedback")
-SHADER_FEATURE_FLAG(22, AtomicInt64OnTypedResource, "64-bit Atomics on Typed Resources")
-SHADER_FEATURE_FLAG(23, AtomicInt64OnGroupShared, "64-bit Atomics on Group Shared")
-SHADER_FEATURE_FLAG(24, DerivativesInMeshAndAmpShaders, "Derivatives in mesh and amplification shaders")
-SHADER_FEATURE_FLAG(25, ResourceDescriptorHeapIndexing, "Resource descriptor heap indexing")
-SHADER_FEATURE_FLAG(26, SamplerDescriptorHeapIndexing, "Sampler descriptor heap indexing")
-SHADER_FEATURE_FLAG(27, RESERVED, "<RESERVED>")
-SHADER_FEATURE_FLAG(28, AtomicInt64OnHeapResource, "64-bit Atomics on Heap Resources")
-SHADER_FEATURE_FLAG(29, AdvancedTextureOps, "Advanced Texture Ops")
-SHADER_FEATURE_FLAG(30, WriteableMSAATextures, "Writeable MSAA Textures")
-
-SHADER_FEATURE_FLAG(31, NextUnusedBit, "Next reserved shader flag bit (not a flag)")
+// SHADER_FEATURE_FLAG(bit offset for the shader info flag, bit offset for DXIL module flag, name, description.
+
+SHADER_FEATURE_FLAG(0,   2, Doubles, "Double-precision floating point")
+SHADER_FEATURE_FLAG(1,  17, ComputeShadersPlusRawAndStructuredBuffers, "Raw and Structured buffers")
+SHADER_FEATURE_FLAG(2,  16, UAVsAtEveryStage, "UAVs at every shader stage")
+SHADER_FEATURE_FLAG(3,  15, Max64UAVs, "64 UAV slots")
+SHADER_FEATURE_FLAG(4,  -1, MinimumPrecision, "Minimum-precision data types")
+SHADER_FEATURE_FLAG(5,   6, DX11_1_DoubleExtensions, "Double-precision extensions for 11.1")
+SHADER_FEATURE_FLAG(6,   7, DX11_1_ShaderExtensions, "Shader extensions for 11.1")
+SHADER_FEATURE_FLAG(7,  14, LEVEL9ComparisonFiltering, "Comparison filtering for feature level 9")
+SHADER_FEATURE_FLAG(8,  12, TiledResources, "Tiled resources")
+SHADER_FEATURE_FLAG(9,  11, StencilRef, "PS Output Stencil Ref")
+SHADER_FEATURE_FLAG(10, 10, InnerCoverage, "PS Inner Coverage")
+SHADER_FEATURE_FLAG(11, 13, TypedUAVLoadAdditionalFormats, "Typed UAV Load Additional Formats")
+SHADER_FEATURE_FLAG(12, 18, ROVs, "Raster Ordered UAVs")
+SHADER_FEATURE_FLAG(13,  9, ViewportAndRTArrayIndexFromAnyShaderFeedingRasterizer, "SV_RenderTargetArrayIndex or SV_ViewportArrayIndex from any shader feeding rasterizer")
+SHADER_FEATURE_FLAG(14, 19, WaveOps, "Wave level operations")
+SHADER_FEATURE_FLAG(15, 20, Int64Ops, "64-Bit integer")
+SHADER_FEATURE_FLAG(16, 21, ViewID, "View Instancing")
+SHADER_FEATURE_FLAG(17, 22, Barycentrics, "Barycentrics")
+SHADER_FEATURE_FLAG(18, -1, NativeLowPrecision, "Use native low precision")
+SHADER_FEATURE_FLAG(19, 24, ShadingRate, "Shading Rate")
+SHADER_FEATURE_FLAG(20, 25, Raytracing_Tier_1_1, "Raytracing tier 1.1 features")
+SHADER_FEATURE_FLAG(21, 26, SamplerFeedback, "Sampler feedback")
+SHADER_FEATURE_FLAG(22, 27, AtomicInt64OnTypedResource, "64-bit Atomics on Typed Resources")
+SHADER_FEATURE_FLAG(23, 28, AtomicInt64OnGroupShared, "64-bit Atomics on Group Shared")
+SHADER_FEATURE_FLAG(24, 29, DerivativesInMeshAndAmpShaders, "Derivatives in mesh and amplification shaders")
+SHADER_FEATURE_FLAG(25, 30, ResourceDescriptorHeapIndexing, "Resource descriptor heap indexing")
+SHADER_FEATURE_FLAG(26, 31, SamplerDescriptorHeapIndexing, "Sampler descriptor heap indexing")
+SHADER_FEATURE_FLAG(27, 63, RESERVED, "<RESERVED>")
+SHADER_FEATURE_FLAG(28, 32, AtomicInt64OnHeapResource, "64-bit Atomics on Heap Resources")
+SHADER_FEATURE_FLAG(29, 34, AdvancedTextureOps, "Advanced Texture Ops")
+SHADER_FEATURE_FLAG(30, 35, WriteableMSAATextures, "Writeable MSAA Textures")
+
+SHADER_FEATURE_FLAG(31, 36, NextUnusedBit, "Next reserved shader flag bit (not a flag)")
 
 #undef SHADER_FEATURE_FLAG
-#endif
+#endif // SHADER_FEATURE_FLAG
+
+#ifdef DXIL_MODULE_FLAG
+
+// Only save DXIL module flags which not map to feature flags here.
+DXIL_MODULE_FLAG( 0,  DisableOptimizations,   "D3D11_1_SB_GLOBAL_FLAG_SKIP_OPTIMIZATION")
+DXIL_MODULE_FLAG( 1,  DisableMathRefactoring, "D3D10_SB_GLOBAL_FLAG_REFACTORING_ALLOWED")
+DXIL_MODULE_FLAG( 3,  ForceEarlyDepthStencil, "D3D11_SB_GLOBAL_FLAG_FORCE_EARLY_DEPTH_STENCIL")
+DXIL_MODULE_FLAG( 4,  EnableRawAndStructuredBuffers, "D3D11_SB_GLOBAL_FLAG_ENABLE_RAW_AND_STRUCTURED_BUFFERS")
+DXIL_MODULE_FLAG( 5,  LowPrecisionPresent, "D3D11_1_SB_GLOBAL_FLAG_ENABLE_MINIMUM_PRECISION")
+DXIL_MODULE_FLAG( 8,  AllResourcesBound, "D3D12_SB_GLOBAL_FLAG_ALL_RESOURCES_BOUND")
+DXIL_MODULE_FLAG(23,  UseNativeLowPrecision, "Native 16bit types enabled")
+DXIL_MODULE_FLAG(33,  ResMayNotAlias, "Any UAV may not alias any other UAV")
+
+#undef DXIL_MODULE_FLAG
+#endif // DXIL_MODULE_FLAG
 
 #ifdef SEMANTIC_KIND
 
@@ -86,7 +103,7 @@ SEMANTIC_KIND(30, CullPrimitive)
 SEMANTIC_KIND(30, Invalid)
 
 #undef SEMANTIC_KIND
-#endif
+#endif // SEMANTIC_KIND
 
 #ifdef COMPONENT_TYPE
 
@@ -102,7 +119,7 @@ COMPONENT_TYPE(8, SInt64)
 COMPONENT_TYPE(9, Float64)
 
 #undef COMPONENT_TYPE
-#endif
+#endif // COMPONENT_TYPE
 
 #ifdef COMPONENT_PRECISION
 
@@ -116,7 +133,7 @@ COMPONENT_PRECISION(0xf0, Any16)
 COMPONENT_PRECISION(0xf1, Any10)
 
 #undef COMPONENT_PRECISION
-#endif
+#endif // COMPONENT_PRECISION
 
 #ifdef INTERPOLATION_MODE
 
@@ -131,7 +148,7 @@ INTERPOLATION_MODE(7, LinearNoperspectiveSample)
 INTERPOLATION_MODE(8, Invalid)
 
 #undef INTERPOLATION_MODE
-#endif
+#endif // INTERPOLATION_MODE
 
 #ifdef D3D_SYSTEM_VALUE
 
@@ -165,4 +182,4 @@ D3D_SYSTEM_VALUE(70, InnerCoverage)
 
 #undef D3D_SYSTEM_VALUE
 
-#endif
+#endif // D3D_SYSTEM_VALUE

llvm/include/llvm/ObjectYAML/DXContainerYAML.h

@@ -56,7 +56,7 @@ struct DXILProgram {
   std::optional<std::vector<llvm::yaml::Hex8>> DXIL;
 };
 
-#define SHADER_FEATURE_FLAG(Num, Val, Str) bool Val = false;
+#define SHADER_FEATURE_FLAG(Num, DxilModuleNum, Val, Str) bool Val = false;
 struct ShaderFeatureFlags {
   ShaderFeatureFlags() = default;
   ShaderFeatureFlags(uint64_t FlagData);

llvm/lib/ObjectYAML/DXContainerYAML.cpp

@@ -24,14 +24,14 @@ static_assert((uint64_t)dxbc::FeatureFlags::NextUnusedBit <= 1ull << 63,
               "Shader flag bits exceed enum size.");
 
 DXContainerYAML::ShaderFeatureFlags::ShaderFeatureFlags(uint64_t FlagData) {
-#define SHADER_FEATURE_FLAG(Num, Val, Str)                                     \
+#define SHADER_FEATURE_FLAG(Num, DxilModuleNum, Val, Str)                      \
   Val = (FlagData & (uint64_t)dxbc::FeatureFlags::Val) > 0;
 #include "llvm/BinaryFormat/DXContainerConstants.def"
 }
 
 uint64_t DXContainerYAML::ShaderFeatureFlags::getEncodedFlags() {
   uint64_t Flag = 0;
-#define SHADER_FEATURE_FLAG(Num, Val, Str)                                     \
+#define SHADER_FEATURE_FLAG(Num, DxilModuleNum, Val, Str)                      \
   if (Val)                                                                     \
     Flag |= (uint64_t)dxbc::FeatureFlags::Val;
 #include "llvm/BinaryFormat/DXContainerConstants.def"
@@ -105,7 +105,8 @@ void MappingTraits<DXContainerYAML::DXILProgram>::mapping(
 
 void MappingTraits<DXContainerYAML::ShaderFeatureFlags>::mapping(
     IO &IO, DXContainerYAML::ShaderFeatureFlags &Flags) {
-#define SHADER_FEATURE_FLAG(Num, Val, Str) IO.mapRequired(#Val, Flags.Val);
+#define SHADER_FEATURE_FLAG(Num, DxilModuleNum, Val, Str)                      \
+  IO.mapRequired(#Val, Flags.Val);
 #include "llvm/BinaryFormat/DXContainerConstants.def"
 }
 

llvm/lib/Target/DirectX/DXContainerGlobals.cpp

@@ -28,7 +28,7 @@ using namespace llvm::dxil;
 namespace {
 class DXContainerGlobals : public llvm::ModulePass {
 
-  GlobalVariable *getShaderFlags(Module &M);
+  GlobalVariable *getFeatureFlags(Module &M);
   GlobalVariable *computeShaderHash(Module &M);
 
 public:
@@ -53,21 +53,24 @@ class DXContainerGlobals : public llvm::ModulePass {
 
 bool DXContainerGlobals::runOnModule(Module &M) {
   llvm::SmallVector<GlobalValue *> Globals;
-  Globals.push_back(getShaderFlags(M));
+  Globals.push_back(getFeatureFlags(M));
   Globals.push_back(computeShaderHash(M));
 
   appendToCompilerUsed(M, Globals);
   return true;
 }
 
-GlobalVariable *DXContainerGlobals::getShaderFlags(Module &M) {
-  const uint64_t Flags =
-      (uint64_t)(getAnalysis<ShaderFlagsAnalysisWrapper>().getShaderFlags());
+GlobalVariable *DXContainerGlobals::getFeatureFlags(Module &M) {
+  const uint64_t FeatureFlags =
+      static_cast<uint64_t>(getAnalysis<ShaderFlagsAnalysisWrapper>()
+                                .getShaderFlags()
+                                .getFeatureFlags());
 
-  Constant *FlagsConstant = ConstantInt::get(M.getContext(), APInt(64, Flags));
-  auto *GV = new llvm::GlobalVariable(M, FlagsConstant->getType(), true,
+  Constant *FeatureFlagsConstant =
+      ConstantInt::get(M.getContext(), APInt(64, FeatureFlags));
+  auto *GV = new llvm::GlobalVariable(M, FeatureFlagsConstant->getType(), true,
                                       GlobalValue::PrivateLinkage,
-                                      FlagsConstant, "dx.sfi0");
+                                      FeatureFlagsConstant, "dx.sfi0");
   GV->setSection("SFI0");
   GV->setAlignment(Align(4));
   return GV;

llvm/lib/Target/DirectX/DXILShaderFlags.cpp

@@ -51,9 +51,14 @@ void ComputedShaderFlags::print(raw_ostream &OS) const {
   if (FlagVal == 0)
     return;
   OS << "; Note: shader requires additional functionality:\n";
-#define SHADER_FEATURE_FLAG(bit, FlagName, Str)                                \
+#define SHADER_FEATURE_FLAG(FeatureBit, DxilModuleNum, FlagName, Str)          \
   if (FlagName)                                                                \
-    OS << ";       " Str "\n";
+    (OS << ";").indent(7) << Str << "\n";
+#include "llvm/BinaryFormat/DXContainerConstants.def"
+  OS << "; Note: extra DXIL module flags:\n";
+#define DXIL_MODULE_FLAG(DxilModuleBit, FlagName, Str)                         \
+  if (FlagName)                                                                \
+    (OS << ";").indent(7) << Str << "\n";
 #include "llvm/BinaryFormat/DXContainerConstants.def"
   OS << ";\n";
 }

llvm/lib/Target/DirectX/DXILShaderFlags.h

@@ -14,7 +14,6 @@
 #ifndef LLVM_TARGET_DIRECTX_DXILSHADERFLAGS_H
 #define LLVM_TARGET_DIRECTX_DXILSHADERFLAGS_H
 
-#include "llvm/BinaryFormat/DXContainer.h"
 #include "llvm/IR/PassManager.h"
 #include "llvm/Pass.h"
 #include "llvm/Support/Compiler.h"
@@ -29,22 +28,37 @@ class GlobalVariable;
 namespace dxil {
 
 struct ComputedShaderFlags {
-#define SHADER_FEATURE_FLAG(bit, FlagName, Str) bool FlagName : 1;
+#define SHADER_FEATURE_FLAG(FeatureBit, DxilModuleBit, FlagName, Str)          \
+  bool FlagName : 1;
+#define DXIL_MODULE_FLAG(DxilModuleBit, FlagName, Str) bool FlagName : 1;
 #include "llvm/BinaryFormat/DXContainerConstants.def"
 
-#define SHADER_FEATURE_FLAG(bit, FlagName, Str) FlagName = false;
+#define SHADER_FEATURE_FLAG(FeatureBit, DxilModuleBit, FlagName, Str)          \
+  FlagName = false;
+#define DXIL_MODULE_FLAG(DxilModuleBit, FlagName, Str) FlagName = false;
   ComputedShaderFlags() {
 #include "llvm/BinaryFormat/DXContainerConstants.def"
   }
 
+  constexpr uint64_t getMask(int Bit) const {
+    return Bit != -1 ? 1ull << Bit : 0;
+  }
   operator uint64_t() const {
     uint64_t FlagValue = 0;
-#define SHADER_FEATURE_FLAG(bit, FlagName, Str)                                \
-  FlagValue |=                                                                 \
-      FlagName ? static_cast<uint64_t>(dxbc::FeatureFlags::FlagName) : 0ull;
+#define SHADER_FEATURE_FLAG(FeatureBit, DxilModuleBit, FlagName, Str)          \
+  FlagValue |= FlagName ? getMask(DxilModuleBit) : 0ull;
+#define DXIL_MODULE_FLAG(DxilModuleBit, FlagName, Str)                         \
+  FlagValue |= FlagName ? getMask(DxilModuleBit) : 0ull;
 #include "llvm/BinaryFormat/DXContainerConstants.def"
     return FlagValue;
   }
+  uint64_t getFeatureFlags() const {
+    uint64_t FeatureFlags = 0;
+#define SHADER_FEATURE_FLAG(FeatureBit, DxilModuleBit, FlagName, Str)          \
+  FeatureFlags |= FlagName ? getMask(FeatureBit) : 0ull;
+#include "llvm/BinaryFormat/DXContainerConstants.def"
+    return FeatureFlags;
+  }
 
   static ComputedShaderFlags computeFlags(Module &M);
   void print(raw_ostream &OS = dbgs()) const;

llvm/lib/Target/DirectX/DXILTranslateMetadata.cpp

@@ -53,8 +53,8 @@ bool DXILTranslateMetadata::runOnModule(Module &M) {
       getAnalysis<DXILResourceWrapper>().getDXILResource();
   Res.write(M);
 
-  const uint64_t Flags =
-      (uint64_t)(getAnalysis<ShaderFlagsAnalysisWrapper>().getShaderFlags());
+  const uint64_t Flags = static_cast<uint64_t>(
+      getAnalysis<ShaderFlagsAnalysisWrapper>().getShaderFlags());
   dxil::createEntryMD(M, Flags);
 
   return false;

llvm/test/CodeGen/DirectX/ShaderFlags/double-extensions.ll

@@ -3,10 +3,11 @@
 
 target triple = "dxil-pc-shadermodel6.7-library"
 
-; CHECK: ; Shader Flags Value: 0x00000021
+; CHECK: ; Shader Flags Value: 0x00000044
 ; CHECK: ; Note: shader requires additional functionality:
 ; CHECK-NEXT: ;       Double-precision floating point
 ; CHECK-NEXT: ;       Double-precision extensions for 11.1
+; CHECK-NEXT: ; Note: extra DXIL module flags:
 ; CHECK-NEXT: {{^;$}}
 define double @div(double %a, double %b) {
   %res = fdiv double %a, %b

llvm/test/CodeGen/DirectX/ShaderFlags/doubles.ll

@@ -3,10 +3,12 @@
 
 target triple = "dxil-pc-shadermodel6.7-library"
 
-; CHECK: ; Shader Flags Value: 0x00000001
+; CHECK: ; Shader Flags Value: 0x00000004
 ; CHECK: ; Note: shader requires additional functionality:
 ; CHECK-NEXT: ;       Double-precision floating point
+; CHECK-NEXT: ; Note: extra DXIL module flags:
 ; CHECK-NEXT: {{^;$}}
+
 define double @add(double %a, double %b) {
   %sum = fadd double %a, %b
   ret double %sum

llvm/test/CodeGen/DirectX/lib_entry.ll

@@ -7,7 +7,7 @@ target triple = "dxil-unknown-shadermodel6.7-library"
 ; Make sure generate empty entry for lib profile.
 ;CHECK:![[empty_entry]] = !{null, !"", null, null, ![[shader_flags:[0-9]+]]}
 ; Make sure double is marked for shader flags.
-;CHECK:![[shader_flags]] = !{i32 0, i64 1}
+;CHECK:![[shader_flags]] = !{i32 0, i64 4}
 ;CHECK:![[entry]] = !{ptr @entry, !"entry", null, null, ![[extra:[0-9]+]]}
 ;CHECK:![[extra]] = !{i32 8, i32 5, i32 4, ![[numthreads:[0-9]+]]}
 ;CHECK:![[numthreads]] = !{i32 1, i32 2, i32 1}