[DirectX] Infrastructure to collect shader flags for each function

llvm/lib/Target/DirectX/DXContainerGlobals.cpp

@@ -78,13 +78,18 @@ bool DXContainerGlobals::runOnModule(Module &M) {
 }
 
 GlobalVariable *DXContainerGlobals::getFeatureFlags(Module &M) {
-  const uint64_t FeatureFlags =
-      static_cast<uint64_t>(getAnalysis<ShaderFlagsAnalysisWrapper>()
-                                .getShaderFlags()
-                                .getFeatureFlags());
+  const DXILModuleShaderFlagsInfo &MSFI =
+      getAnalysis<ShaderFlagsAnalysisWrapper>().getShaderFlags();
+  // TODO: Feature flags mask is obtained as a collection of feature flags
+  // of the shader flags of all functions in the module. Need to verify
+  // and modify the computation of feature flags to be used.
+  uint64_t ConsolidatedFeatureFlags = 0;
+  for (const auto &FuncFlags : MSFI.FuncShaderFlagsVec) {
+    ConsolidatedFeatureFlags |= FuncFlags.second.getFeatureFlags();
+  }
 
   Constant *FeatureFlagsConstant =
-      ConstantInt::get(M.getContext(), APInt(64, FeatureFlags));
+      ConstantInt::get(M.getContext(), APInt(64, ConsolidatedFeatureFlags));
   return buildContainerGlobal(M, FeatureFlagsConstant, "dx.sfi0", "SFI0");
 }
 

llvm/lib/Target/DirectX/DXILShaderFlags.cpp

@@ -13,36 +13,88 @@
 
 #include "DXILShaderFlags.h"
 #include "DirectX.h"
+#include "llvm/ADT/STLExtras.h"
+#include "llvm/IR/DiagnosticInfo.h"
+#include "llvm/IR/DiagnosticPrinter.h"
 #include "llvm/IR/Instruction.h"
 #include "llvm/IR/Module.h"
 #include "llvm/Support/FormatVariadic.h"
+#include "llvm/Support/raw_ostream.h"
 
 using namespace llvm;
 using namespace llvm::dxil;
 
-static void updateFlags(ComputedShaderFlags &Flags, const Instruction &I) {
+namespace {
+/// A simple Wrapper DiagnosticInfo that generates Module-level diagnostic
+/// for ShaderFlagsAnalysis pass
+class DiagnosticInfoShaderFlags : public DiagnosticInfo {
+private:
+  const Twine &Msg;
+  const Module &Mod;
+
+public:
+  /// \p M is the module for which the diagnostic is being emitted. \p Msg is
+  /// the message to show. Note that this class does not copy this message, so
+  /// this reference must be valid for the whole life time of the diagnostic.
+  DiagnosticInfoShaderFlags(const Module &M, const Twine &Msg,
+                            DiagnosticSeverity Severity = DS_Error)
+      : DiagnosticInfo(DK_Unsupported, Severity), Msg(Msg), Mod(M) {}
+
+  void print(DiagnosticPrinter &DP) const override {
+    DP << Mod.getName() << ": " << Msg << '\n';
+  }
+};
+} // namespace
+
+static void updateFlags(ComputedShaderFlags &CSF, const Instruction &I) {
   Type *Ty = I.getType();
-  if (Ty->isDoubleTy()) {
-    Flags.Doubles = true;
+  bool DoubleTyInUse = Ty->isDoubleTy();
+  for (Value *Op : I.operands()) {
+    DoubleTyInUse |= Op->getType()->isDoubleTy();
+  }
+
+  if (DoubleTyInUse) {
+    CSF.Doubles = true;
     switch (I.getOpcode()) {
     case Instruction::FDiv:
     case Instruction::UIToFP:
     case Instruction::SIToFP:
     case Instruction::FPToUI:
     case Instruction::FPToSI:
-      Flags.DX11_1_DoubleExtensions = true;
+      // TODO: To be set if I is a call to DXIL intrinsic DXIL::Opcode::Fma
+      CSF.DX11_1_DoubleExtensions = true;
       break;
     }
   }
 }
 
-ComputedShaderFlags ComputedShaderFlags::computeFlags(Module &M) {
-  ComputedShaderFlags Flags;
-  for (const auto &F : M)
+static bool compareFuncSFPairs(const FuncShaderFlagsMask &First,
+                               const FuncShaderFlagsMask &Second) {
+  return (First.first->getName().compare(Second.first->getName()) < 0);
+}
+
+static DXILModuleShaderFlagsInfo computeFlags(Module &M) {
+  DXILModuleShaderFlagsInfo MSFI;
+  for (auto &F : M) {
+    if (F.isDeclaration())
+      continue;
+    // Each of the functions in a module are unique. Hence no prior shader flags
+    // mask of the function should be present.
+    if (MSFI.hasShaderFlagsMask(&F)) {
+      M.getContext().diagnose(
+          DiagnosticInfoShaderFlags(M, "Shader Flags mask for Function '" +
+                                           F.getName() + "' already exists"));
+    }
+    ComputedShaderFlags CSF{};
     for (const auto &BB : F)
       for (const auto &I : BB)
-        updateFlags(Flags, I);
-  return Flags;
+        updateFlags(CSF, I);
+    // Insert shader flag mask for function F
+    MSFI.FuncShaderFlagsVec.push_back({&F, CSF});
+  }
+  // Sort MSFI.FuncShaderFlagsVec for later lookup that uses binary search
+  llvm::sort(MSFI.FuncShaderFlagsVec, compareFuncSFPairs);
+  return MSFI;
 }
 
 void ComputedShaderFlags::print(raw_ostream &OS) const {
@@ -63,16 +115,53 @@ void ComputedShaderFlags::print(raw_ostream &OS) const {
   OS << ";\n";
 }
 
+void DXILModuleShaderFlagsInfo::print(raw_ostream &OS) const {
+  OS << "; Shader Flags mask for Module:\n";
+  ModuleFlags.print(OS);
+  for (auto SF : FuncShaderFlagsVec) {
+    OS << "; Shader Flags mask for Function: " << SF.first->getName() << "\n";
+    SF.second.print(OS);
+  }
+}
+
+const ComputedShaderFlags
+DXILModuleShaderFlagsInfo::getShaderFlagsMask(const Function *Func) const {
+  FuncShaderFlagsMask V{Func, {}};
+  auto Iter = llvm::lower_bound(FuncShaderFlagsVec, V, compareFuncSFPairs);
+  if (Iter == FuncShaderFlagsVec.end()) {
+    Func->getContext().diagnose(DiagnosticInfoShaderFlags(
+        *(Func->getParent()), "Shader Flags information of Function '" +
+                                  Twine(Func->getName()) + "' not found"));
+  }
+  if (Iter->first != Func) {
+    Func->getContext().diagnose(DiagnosticInfoShaderFlags(
+        *(Func->getParent()),
+        "Inconsistent Shader Flags information of Function '" +
+            Twine(Func->getName()) + "' retrieved"));
+  }
+  return Iter->second;
+}
+
+bool DXILModuleShaderFlagsInfo::hasShaderFlagsMask(const Function *Func) const {
+  FuncShaderFlagsMask V{Func, {}};
+  return llvm::binary_search(FuncShaderFlagsVec, V);
+}
+
 AnalysisKey ShaderFlagsAnalysis::Key;
 
-ComputedShaderFlags ShaderFlagsAnalysis::run(Module &M,
-                                             ModuleAnalysisManager &AM) {
-  return ComputedShaderFlags::computeFlags(M);
+DXILModuleShaderFlagsInfo ShaderFlagsAnalysis::run(Module &M,
+                                                   ModuleAnalysisManager &AM) {
+  return computeFlags(M);
+}
+
+bool ShaderFlagsAnalysisWrapper::runOnModule(Module &M) {
+  MSFI = computeFlags(M);
+  return false;
 }
 
 PreservedAnalyses ShaderFlagsAnalysisPrinter::run(Module &M,
                                                   ModuleAnalysisManager &AM) {
-  ComputedShaderFlags Flags = AM.getResult<ShaderFlagsAnalysis>(M);
+  DXILModuleShaderFlagsInfo Flags = AM.getResult<ShaderFlagsAnalysis>(M);
   Flags.print(OS);
   return PreservedAnalyses::all();
 }

llvm/lib/Target/DirectX/DXILShaderFlags.h

@@ -14,6 +14,7 @@
 #ifndef LLVM_TARGET_DIRECTX_DXILSHADERFLAGS_H
 #define LLVM_TARGET_DIRECTX_DXILSHADERFLAGS_H
 
+#include "llvm/IR/Function.h"
 #include "llvm/IR/PassManager.h"
 #include "llvm/Pass.h"
 #include "llvm/Support/Compiler.h"
@@ -60,21 +61,34 @@ struct ComputedShaderFlags {
     return FeatureFlags;
   }
 
-  static ComputedShaderFlags computeFlags(Module &M);
   void print(raw_ostream &OS = dbgs()) const;
   LLVM_DUMP_METHOD void dump() const { print(); }
 };
 
+using FuncShaderFlagsMask = std::pair<Function const *, ComputedShaderFlags>;
+using FunctionShaderFlagsVec = SmallVector<FuncShaderFlagsMask>;
+struct DXILModuleShaderFlagsInfo {
+  // Shader Flag mask representing module-level properties
+  ComputedShaderFlags ModuleFlags;
+  // Vector of Function-Shader Flag mask pairs representing properties of each
+  // of the functions in the module
+  FunctionShaderFlagsVec FuncShaderFlagsVec;
+
+  const ComputedShaderFlags getShaderFlagsMask(const Function *Func) const;
+  bool hasShaderFlagsMask(const Function *Func) const;
+  void print(raw_ostream &OS = dbgs()) const;
+};
+
 class ShaderFlagsAnalysis : public AnalysisInfoMixin<ShaderFlagsAnalysis> {
   friend AnalysisInfoMixin<ShaderFlagsAnalysis>;
   static AnalysisKey Key;
 
 public:
   ShaderFlagsAnalysis() = default;
 
-  using Result = ComputedShaderFlags;
+  using Result = DXILModuleShaderFlagsInfo;
 
-  ComputedShaderFlags run(Module &M, ModuleAnalysisManager &AM);
+  DXILModuleShaderFlagsInfo run(Module &M, ModuleAnalysisManager &AM);
 };
 
 /// Printer pass for ShaderFlagsAnalysis results.
@@ -92,19 +106,16 @@ class ShaderFlagsAnalysisPrinter
 /// This is required because the passes that will depend on this are codegen
 /// passes which run through the legacy pass manager.
 class ShaderFlagsAnalysisWrapper : public ModulePass {
-  ComputedShaderFlags Flags;
+  DXILModuleShaderFlagsInfo MSFI;
 
 public:
   static char ID;
 
   ShaderFlagsAnalysisWrapper() : ModulePass(ID) {}
 
-  const ComputedShaderFlags &getShaderFlags() { return Flags; }
+  const DXILModuleShaderFlagsInfo &getShaderFlags() { return MSFI; }
 
-  bool runOnModule(Module &M) override {
-    Flags = ComputedShaderFlags::computeFlags(M);
-    return false;
-  }
+  bool runOnModule(Module &M) override;
 
   void getAnalysisUsage(AnalysisUsage &AU) const override {
     AU.setPreservesAll();

llvm/lib/Target/DirectX/DXILTranslateMetadata.cpp

@@ -286,11 +286,6 @@ static MDTuple *emitTopLevelLibraryNode(Module &M, MDNode *RMD,
   MDTuple *Properties = nullptr;
   if (ShaderFlags != 0) {
     SmallVector<Metadata *> MDVals;
-    // FIXME: ShaderFlagsAnalysis pass needs to collect and provide
-    // ShaderFlags for each entry function. Currently, ShaderFlags value
-    // provided by ShaderFlagsAnalysis pass is created by walking *all* the
-    // function instructions of the module. Is it is correct to use this value
-    // for metadata of the empty library entry?
     MDVals.append(
         getTagValueAsMetadata(EntryPropsTag::ShaderFlags, ShaderFlags, Ctx));
     Properties = MDNode::get(Ctx, MDVals);
@@ -302,7 +297,7 @@ static MDTuple *emitTopLevelLibraryNode(Module &M, MDNode *RMD,
 
 static void translateMetadata(Module &M, const DXILResourceMap &DRM,
                               const Resources &MDResources,
-                              const ComputedShaderFlags &ShaderFlags,
+                              const DXILModuleShaderFlagsInfo &ShaderFlags,
                               const ModuleMetadataInfo &MMDI) {
   LLVMContext &Ctx = M.getContext();
   IRBuilder<> IRB(Ctx);
@@ -318,22 +313,32 @@ static void translateMetadata(Module &M, const DXILResourceMap &DRM,
   // See https://github.com/llvm/llvm-project/issues/57928
   MDTuple *Signatures = nullptr;
 
-  if (MMDI.ShaderProfile == Triple::EnvironmentType::Library)
+  if (MMDI.ShaderProfile == Triple::EnvironmentType::Library) {
+    // Create a consolidated shader flag mask of all functions in the library
+    // to be used as shader flags mask value associated with top-level library
+    // entry metadata.
+    uint64_t ConsolidatedMask = ShaderFlags.ModuleFlags;
+    for (const auto &FunFlags : ShaderFlags.FuncShaderFlagsVec) {
+      ConsolidatedMask |= FunFlags.second;
+    }
     EntryFnMDNodes.emplace_back(
-        emitTopLevelLibraryNode(M, ResourceMD, ShaderFlags));
-  else if (MMDI.EntryPropertyVec.size() > 1) {
+        emitTopLevelLibraryNode(M, ResourceMD, ConsolidatedMask));
+  } else if (MMDI.EntryPropertyVec.size() > 1) {
     M.getContext().diagnose(DiagnosticInfoTranslateMD(
         M, "Non-library shader: One and only one entry expected"));
   }
 
   for (const EntryProperties &EntryProp : MMDI.EntryPropertyVec) {
-    // FIXME: ShaderFlagsAnalysis pass needs to collect and provide
-    // ShaderFlags for each entry function. For now, assume shader flags value
-    // of entry functions being compiled for lib_* shader profile viz.,
-    // EntryPro.Entry is 0.
-    uint64_t EntryShaderFlags =
-        (MMDI.ShaderProfile == Triple::EnvironmentType::Library) ? 0
-                                                                 : ShaderFlags;
+    ComputedShaderFlags ECSF = ShaderFlags.getShaderFlagsMask(EntryProp.Entry);
+    // If ShaderProfile is Library, mask is already consolidated in the
+    // top-level library node. Hence it is not emitted.
+    uint64_t EntryShaderFlags = 0;
+    if (MMDI.ShaderProfile != Triple::EnvironmentType::Library) {
+      // TODO: Create a consolidated shader flag mask of all the entry
+      // functions and its callees. The following is correct only if
+      // EntryProp.Entry has no call instructions.
+      EntryShaderFlags = ECSF | ShaderFlags.ModuleFlags;
+    }
     if (MMDI.ShaderProfile != Triple::EnvironmentType::Library) {
       if (EntryProp.ShaderStage != MMDI.ShaderProfile) {
         M.getContext().diagnose(DiagnosticInfoTranslateMD(
@@ -361,7 +366,7 @@ PreservedAnalyses DXILTranslateMetadata::run(Module &M,
                                              ModuleAnalysisManager &MAM) {
   const DXILResourceMap &DRM = MAM.getResult<DXILResourceAnalysis>(M);
   const dxil::Resources &MDResources = MAM.getResult<DXILResourceMDAnalysis>(M);
-  const ComputedShaderFlags &ShaderFlags =
+  const DXILModuleShaderFlagsInfo &ShaderFlags =
       MAM.getResult<ShaderFlagsAnalysis>(M);
   const dxil::ModuleMetadataInfo MMDI = MAM.getResult<DXILMetadataAnalysis>(M);
 
@@ -393,7 +398,7 @@ class DXILTranslateMetadataLegacy : public ModulePass {
         getAnalysis<DXILResourceWrapperPass>().getResourceMap();
     const dxil::Resources &MDResources =
         getAnalysis<DXILResourceMDWrapper>().getDXILResource();
-    const ComputedShaderFlags &ShaderFlags =
+    const DXILModuleShaderFlagsInfo &ShaderFlags =
         getAnalysis<ShaderFlagsAnalysisWrapper>().getShaderFlags();
     dxil::ModuleMetadataInfo MMDI =
         getAnalysis<DXILMetadataAnalysisWrapperPass>().getModuleMetadata();

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

@@ -0,0 +1,19 @@
+; RUN: llc %s --filetype=obj -o - | obj2yaml | FileCheck %s --check-prefix=DXC
+
+target triple = "dxil-pc-shadermodel6.7-library"
+define double @div(double %a, double %b) #0 {
+  %res = fdiv double %a, %b
+  ret double %res
+}
+
+attributes #0 = { convergent norecurse nounwind "hlsl.export"}
+
+; DXC: - Name:            SFI0
+; DXC-NEXT:     Size:            8
+; DXC-NEXT:     Flags:
+; DXC-NEXT:       Doubles:         true
+; DXC-NOT:   {{[A-Za-z]+: +true}}
+; DXC:            DX11_1_DoubleExtensions:         true
+; DXC-NOT:   {{[A-Za-z]+: +true}}
+; DXC:       NextUnusedBit:   false
+; DXC: ...