[NFC] Improve time tracing data (#7146)

This is a bunch of small changes to improve the quality of the time
traces. This mostly adds new timers breakign down dxcompilerobj and the
always inliner code.

lib/DxilContainer/DxilContainerAssembler.cpp

@@ -37,6 +37,7 @@
 #include "llvm/IR/Module.h"
 #include "llvm/IR/Operator.h"
 #include "llvm/Support/MD5.h"
+#include "llvm/Support/TimeProfiler.h"
 #include "llvm/Transforms/Utils/Cloning.h"
 #include <algorithm>
 #include <assert.h> // Needed for DxilPipelineStateValidation.h
@@ -1895,6 +1896,7 @@ void hlsl::SerializeDxilContainerForModule(
     DxilShaderHash *pShaderHashOut, AbstractMemoryStream *pReflectionStreamOut,
     AbstractMemoryStream *pRootSigStreamOut, void *pPrivateData,
     size_t PrivateDataSize) {
+  llvm::TimeTraceScope TimeScope("SerializeDxilContainer", StringRef(""));
   // TODO: add a flag to update the module and remove information that is not
   // part of DXIL proper and is used only to assemble the container.
 

lib/Transforms/Utils/CloneFunction.cpp

@@ -13,7 +13,6 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "llvm/Transforms/Utils/Cloning.h"
 #include "llvm/ADT/SmallVector.h"
 #include "llvm/Analysis/ConstantFolding.h"
 #include "llvm/Analysis/InstructionSimplify.h"
@@ -29,7 +28,9 @@
 #include "llvm/IR/LLVMContext.h"
 #include "llvm/IR/Metadata.h"
 #include "llvm/IR/Module.h"
+#include "llvm/Support/TimeProfiler.h"
 #include "llvm/Transforms/Utils/BasicBlockUtils.h"
+#include "llvm/Transforms/Utils/Cloning.h"
 #include "llvm/Transforms/Utils/Local.h"
 #include "llvm/Transforms/Utils/ValueMapper.h"
 #include <map>
@@ -473,6 +474,9 @@ void llvm::CloneAndPruneIntoFromInst(Function *NewFunc, const Function *OldFunc,
                                      const char *NameSuffix, 
                                      ClonedCodeInfo *CodeInfo,
                                      CloningDirector *Director) {
+  TimeTraceScope TimeScope("CloneAndPruneIntoFromInst", [&] {
+    return (Twine(OldFunc->getName()) + "->" + NewFunc->getName()).str();
+  });
   assert(NameSuffix && "NameSuffix cannot be null!");
 
   ValueMapTypeRemapper *TypeMapper = nullptr;

lib/Transforms/Utils/InlineFunction.cpp

@@ -12,10 +12,9 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "llvm/Transforms/Utils/Cloning.h"
+#include "llvm/ADT/SetVector.h"
 #include "llvm/ADT/SmallSet.h"
 #include "llvm/ADT/SmallVector.h"
-#include "llvm/ADT/SetVector.h"
 #include "llvm/ADT/StringExtras.h"
 #include "llvm/Analysis/AliasAnalysis.h"
 #include "llvm/Analysis/AssumptionCache.h"
@@ -24,22 +23,24 @@
 #include "llvm/Analysis/InstructionSimplify.h"
 #include "llvm/Analysis/ValueTracking.h"
 #include "llvm/IR/Attributes.h"
-#include "llvm/IR/CallSite.h"
 #include "llvm/IR/CFG.h"
+#include "llvm/IR/CallSite.h"
 #include "llvm/IR/Constants.h"
+#include "llvm/IR/DIBuilder.h"
 #include "llvm/IR/DataLayout.h"
 #include "llvm/IR/DebugInfo.h"
 #include "llvm/IR/DerivedTypes.h"
-#include "llvm/IR/DIBuilder.h"
 #include "llvm/IR/Dominators.h"
 #include "llvm/IR/IRBuilder.h"
 #include "llvm/IR/Instructions.h"
 #include "llvm/IR/IntrinsicInst.h"
 #include "llvm/IR/Intrinsics.h"
 #include "llvm/IR/MDBuilder.h"
 #include "llvm/IR/Module.h"
-#include "llvm/Transforms/Utils/Local.h"
 #include "llvm/Support/CommandLine.h"
+#include "llvm/Support/TimeProfiler.h"
+#include "llvm/Transforms/Utils/Cloning.h"
+#include "llvm/Transforms/Utils/Local.h"
 #include <algorithm>
 using namespace llvm;
 
@@ -291,6 +292,8 @@ static void HandleInlinedInvoke(InvokeInst *II, BasicBlock *FirstNewBlock,
 /// non-aliasing property communicated by the metadata could have
 /// call-site-specific control dependencies).
 static void CloneAliasScopeMetadata(CallSite CS, ValueToValueMapTy &VMap) {
+  TimeTraceScope TimeScope("CloneAliasScopeMetadata",
+                           [&] { return CS.getCalledFunction()->getName(); });
   const Function *CalledFunc = CS.getCalledFunction();
   SetVector<const MDNode *> MD;
 
@@ -401,6 +404,8 @@ static void CloneAliasScopeMetadata(CallSite CS, ValueToValueMapTy &VMap) {
 /// non-derived loads, stores and memory intrinsics with the new alias scopes.
 static void AddAliasScopeMetadata(CallSite CS, ValueToValueMapTy &VMap,
                                   const DataLayout &DL, AliasAnalysis *AA) {
+  TimeTraceScope TimeScope("AddAliasScopeMetadata",
+                           [&] { return CS.getCalledFunction()->getName(); });
   if (!EnableNoAliasConversion)
     return;
 
@@ -872,6 +877,7 @@ updateInlinedAtInfo(DebugLoc DL, DILocation *InlinedAtNode, LLVMContext &Ctx,
 /// to encode location where these instructions are inlined.
 static void fixupLineNumbers(Function *Fn, Function::iterator FI,
                              Instruction *TheCall) {
+  TimeTraceScope TimeScope("fixupLineNumbers", [&] { return Fn->getName(); });
   DebugLoc TheCallDL = TheCall->getDebugLoc();
 #if 0 // HLSL Change
   if (!TheCallDL)

tools/clang/lib/CodeGen/CodeGenModule.cpp

@@ -3376,6 +3376,12 @@ void CodeGenModule::EmitLinkageSpec(const LinkageSpecDecl *LSD) {
 
 /// EmitTopLevelDecl - Emit code for a single top level declaration.
 void CodeGenModule::EmitTopLevelDecl(Decl *D) {
+  llvm::TimeTraceScope TimeScope("CGM::EmitTopLevelDecl", [&] {
+    if (const auto *ND = dyn_cast<NamedDecl>(D))
+      return ND->getName();
+    return StringRef("Unnamed decl");
+  });
+
   // Ignore dependent declarations.
   if (D->getDeclContext() && D->getDeclContext()->isDependentContext())
     return;

tools/clang/lib/Parse/ParseAST.cpp

@@ -100,8 +100,6 @@ void clang::ParseAST(Preprocessor &PP, ASTConsumer *Consumer,
 
 void clang::ParseAST(Sema &S, bool PrintStats, bool SkipFunctionBodies) {
 
-  // HLSL Change - Support hierarchial time tracing.
-  llvm::TimeTraceScope TimeScope("Frontend", StringRef(""));
   // Collect global stats on Decls/Stmts (until we have a module streamer).
   if (PrintStats) {
     Decl::EnableStatistics();
@@ -137,6 +135,8 @@ void clang::ParseAST(Sema &S, bool PrintStats, bool SkipFunctionBodies) {
     External->StartTranslationUnit(Consumer);
 
   if (!S.getDiagnostics().hasUnrecoverableErrorOccurred()) {  // HLSL Change: Skip if fatal error already occurred
+    // HLSL Change - Support hierarchial time tracing.
+    llvm::TimeTraceScope TimeScope("Frontend", StringRef(""));
     if (P.ParseTopLevelDecl(ADecl)) {
       if (!External && !S.getLangOpts().CPlusPlus)
         P.Diag(diag::ext_empty_translation_unit);
@@ -151,10 +151,14 @@ void clang::ParseAST(Sema &S, bool PrintStats, bool SkipFunctionBodies) {
     }
   } // HLSL Change: Skip if fatal error already occurred
 
-  // Process any TopLevelDecls generated by #pragma weak.
-  for (Decl *D : S.WeakTopLevelDecls())
-    Consumer->HandleTopLevelDecl(DeclGroupRef(D));
-
+  {
+    // HLSL Change - Support hierarchial time tracing.
+    llvm::TimeTraceScope TimeScope("Frontend - Consumer", StringRef(""));
+    // Process any TopLevelDecls generated by #pragma weak.
+    for (Decl *D : S.WeakTopLevelDecls())
+      Consumer->HandleTopLevelDecl(DeclGroupRef(D));
+  }
+
   // HLSL Change Starts
   // Provide the opportunity to generate translation-unit level validation
   // errors in the front-end, without relying on code generation being

tools/clang/tools/dxcompiler/dxcompilerobj.cpp

@@ -722,6 +722,7 @@ class DxcCompiler : public IDxcCompiler3,
       bool validateRootSigContainer = false;
 
       if (isPreprocessing) {
+        TimeTraceScope TimeScope("PreprocessAction", StringRef(""));
         // These settings are back-compatible with fxc.
         clang::PreprocessorOutputOptions &PPOutOpts =
             compiler.getPreprocessorOutputOpts();
@@ -867,6 +868,7 @@ class DxcCompiler : public IDxcCompiler3,
       compiler.getTarget().adjust(compiler.getLangOpts());
 
       if (opts.AstDump) {
+        TimeTraceScope TimeScope("DumpAST", StringRef(""));
         clang::ASTDumpAction dumpAction;
         // Consider - ASTDumpFilter, ASTDumpLookups
         compiler.getFrontendOpts().ASTDumpDecls = true;
@@ -876,6 +878,7 @@ class DxcCompiler : public IDxcCompiler3,
         dumpAction.EndSourceFile();
         outStream.flush();
       } else if (opts.DumpDependencies) {
+        TimeTraceScope TimeScope("DumpDependencies", StringRef(""));
         auto dependencyCollector = std::make_shared<DependencyCollector>();
         compiler.addDependencyCollector(dependencyCollector);
         compiler.createPreprocessor(clang::TranslationUnitKind::TU_Complete);
@@ -978,6 +981,7 @@ class DxcCompiler : public IDxcCompiler3,
         EmitBCAction action(&llvmContext);
         FrontendInputFile file(pUtf8SourceName, IK_HLSL);
         bool compileOK;
+        TimeTraceScope TimeScope("Compile Action", StringRef(""));
         if (action.BeginSourceFile(compiler, file)) {
           action.Execute();
           action.EndSourceFile();
@@ -1032,6 +1036,7 @@ class DxcCompiler : public IDxcCompiler3,
         // Do not create a container when there is only a a high-level
         // representation in the module.
         if (compileOK && !opts.CodeGenHighLevel) {
+          TimeTraceScope TimeScope("AssembleAndWriteContainer", StringRef(""));
           HRESULT valHR = S_OK;
           CComPtr<AbstractMemoryStream> pRootSigStream;
           IFT(CreateMemoryStream(DxcGetThreadMallocNoRef(),