[CodeGen] emit CG profile for COFF object file

[zequanwu]

Summary

I forgot to add emission of CG profile for COFF object file, when adding the support(https://reviews.llvm.org/D81775)

[zequanwu]

test case.

[rnk]

lgtm!

I see the assembler support is already there, looks like it works.

[rnk]

llvm/lib/Target/TargetLoweringObjectFile.cpp

164 ↗

(On Diff #292846)

I had to revert this change in rG90242caca2074dab5a9b76e5bc36d9fafd2179a7. I think the fix might be to return null if the call graph edge target is marked dllimport here.

[zequanwu]

llvm/lib/Target/TargetLoweringObjectFile.cpp

164 ↗

(On Diff #292846)

I was thinking about a discussion in https://reviews.llvm.org/D81775?id=270540#inline-751699, which might be the cause. I was trying to do the same as ELF. Not sure if that's the cause.

[zequanwu]

Update. Don't emit CGProfileEntry if function has dll import storage class.

[rnk]

lgtm

llvm/lib/Target/TargetLoweringObjectFile.cpp

165 ↗

(On Diff #293899)

Thanks! You can see the diff here if you are curious:
https://reviews.llvm.org/D87811?vs=292846&id=293899#toc

[zequanwu]

For symbols haven't seen, just set them to external not weak external.

Previously I did it that way, because http://llvm.org/docs/Extensions.html#sht-llvm-call-graph-profile-section-call-graph-profile says

If either symbol is undefined, then that symbol is defined as if .weak symbol has been written at the end of the file.

[zequanwu]

llvm/lib/MC/MCWinCOFFStreamer.cpp

344 ↗

(On Diff #294817)

It's meaningless to set symbol to WeakExternal and External at same time here. At the end, the symbol just become WeakExternal.

[rnk]

Can you remove the weak external marking in a separate change from relanding? It changes existing functionality of the .cg_profile directive, so it can be committed independently, which is nice for making changes as small and independent as possible.

[zequanwu]

Split changes on .cg_profile to another patch D88456.

[probinson]

Can we have a separate NFC refactor commit, to move that lump of code into its own emitCGProfile method? It might make the functional bits easier to follow. Also we have a private change in that area, and the ping-pong for this change has been a real headache.

[zequanwu]

Will put the refactor part to a separate NFC commit.
Reland this as fix at D88456 seems work. (https://bugs.chromium.org/p/chromium/issues/detail?id=1130780#c18)

llvm/include/llvm/CodeGen/TargetLoweringObjectFileImpl.h

const MCExpr *getIndirectSymViaGOTPCRel(const GlobalValue *GV,
                                           MCStreamer &Streamer) const override;
 
   void getNameWithPrefix(SmallVectorImpl<char> &OutName, const GlobalValue *GV,
                          const TargetMachine &TM) const override;
 };
 
 class TargetLoweringObjectFileCOFF : public TargetLoweringObjectFile {
   mutable unsigned NextUniqueID = 0;
+  const TargetMachine *TM = nullptr;
 
 public:
   ~TargetLoweringObjectFileCOFF() override = default;
 
   void Initialize(MCContext &Ctx, const TargetMachine &TM) override;
   MCSection *getExplicitSectionGlobal(const GlobalObject *GO, SectionKind Kind,
                                       const TargetMachine &TM) const override;
 

llvm/lib/CodeGen/TargetLoweringObjectFileImpl.cpp

void TargetLoweringObjectFileCOFF::emitModuleMetadata(MCStreamer &Streamer,
                                                       Module &M) const {
   emitLinkerDirectives(Streamer, M);
 
   unsigned Version = 0;
   unsigned Flags = 0;
   StringRef Section;
 
   GetObjCImageInfo(M, Version, Flags, Section);
-  if (Section.empty())
-    return;
-
+  if (!Section.empty()) {
     auto &C = getContext();
-  auto *S = C.getCOFFSection(
-      Section, COFF::IMAGE_SCN_CNT_INITIALIZED_DATA | COFF::IMAGE_SCN_MEM_READ,
+    auto *S = C.getCOFFSection(Section,
+                               COFF::IMAGE_SCN_CNT_INITIALIZED_DATA |
+                                   COFF::IMAGE_SCN_MEM_READ,
                                SectionKind::getReadOnly());
     Streamer.SwitchSection(S);
     Streamer.emitLabel(C.getOrCreateSymbol(StringRef("OBJC_IMAGE_INFO")));
     Streamer.emitInt32(Version);
     Streamer.emitInt32(Flags);
     Streamer.AddBlankLine();
   }
 
+  auto &C = getContext();
+  SmallVector<Module::ModuleFlagEntry, 8> ModuleFlags;
+  M.getModuleFlagsMetadata(ModuleFlags);
+
+  MDNode *CFGProfile = nullptr;
+
+  for (const auto &MFE : ModuleFlags) {
+    StringRef Key = MFE.Key->getString();
+    if (Key == "CG Profile") {
+      CFGProfile = cast<MDNode>(MFE.Val);
+      break;
+    }
+  }
+
+  if (!CFGProfile)
+    return;
+
+  auto GetSym = [this](const MDOperand &MDO) -> MCSymbol * {
+    if (!MDO)
+      return nullptr;
+    auto V = cast<ValueAsMetadata>(MDO);
+    const Function *F = cast<Function>(V->getValue());
+    if (F->hasDLLImportStorageClass())
+      return nullptr;
+    return TM->getSymbol(F);
+  };
+
+  for (const auto &Edge : CFGProfile->operands()) {
+    MDNode *E = cast<MDNode>(Edge);
+    const MCSymbol *From = GetSym(E->getOperand(0));
+    const MCSymbol *To = GetSym(E->getOperand(1));
+    // Skip null functions. This can happen if functions are dead stripped after
+    // the CGProfile pass has been run.
+    if (!From || !To)
+      continue;
+    uint64_t Count = cast<ConstantAsMetadata>(E->getOperand(2))
+                         ->getValue()
+                         ->getUniqueInteger()
+                         .getZExtValue();
+    Streamer.emitCGProfileEntry(
+        MCSymbolRefExpr::create(From, MCSymbolRefExpr::VK_None, C),
+        MCSymbolRefExpr::create(To, MCSymbolRefExpr::VK_None, C), Count);
+  }
+}
+
 void TargetLoweringObjectFileCOFF::emitLinkerDirectives(
     MCStreamer &Streamer, Module &M) const {
   if (NamedMDNode *LinkerOptions = M.getNamedMetadata("llvm.linker.options")) {
     // Emit the linker options to the linker .drectve section.  According to the
     // spec, this section is a space-separated string containing flags for
     // linker.
     MCSection *Sec = getDrectveSection();
     Streamer.SwitchSection(Sec);
if (const auto *A = cast<ConstantArray>(LU->getInitializer())) {
       }
     }
   }
 }
 
 void TargetLoweringObjectFileCOFF::Initialize(MCContext &Ctx,
                                               const TargetMachine &TM) {
   TargetLoweringObjectFile::Initialize(Ctx, TM);
+  this->TM = &TM;
   const Triple &T = TM.getTargetTriple();
   if (T.isWindowsMSVCEnvironment() || T.isWindowsItaniumEnvironment()) {
     StaticCtorSection =
         Ctx.getCOFFSection(".CRT$XCU", COFF::IMAGE_SCN_CNT_INITIALIZED_DATA |
                                            COFF::IMAGE_SCN_MEM_READ,
                            SectionKind::getReadOnly());
     StaticDtorSection =
         Ctx.getCOFFSection(".CRT$XTX", COFF::IMAGE_SCN_CNT_INITIALIZED_DATA |

llvm/test/MC/COFF/cgprofile.ll

+; RUN: llc -filetype=asm %s -o - -mtriple x86_64-pc-windows-msvc | FileCheck %s
+; RUN: llc -filetype=obj %s -o %t -mtriple x86_64-pc-windows-msvc
+; RUN: llvm-readobj --cg-profile %t | FileCheck %s --check-prefix=OBJ
+
+declare void @b()
+
+define void @a() {
+  call void @b()
+  ret void
+}
+
+define void @freq(i1 %cond) {
+  br i1 %cond, label %A, label %B
+A:
+  call void @a();
+  ret void
+B:
+  call void @b();
+  ret void
+}
+
+!llvm.module.flags = !{!0}
+
+!0 = !{i32 5, !"CG Profile", !1}
+!1 = !{!2, !3, !4, !5}
+!2 = !{void ()* @a, void ()* @b, i64 32}
+!3 = !{void (i1)* @freq, void ()* @a, i64 11}
+!4 = !{void (i1)* @freq, void ()* @b, i64 20}
+!5 = !{void (i1)* @freq, null, i64 20}
+
+; CHECK: .cg_profile a, b, 32
+; CHECK: .cg_profile freq, a, 11
+; CHECK: .cg_profile freq, b, 20
+
+; OBJ: CGProfile [
+; OBJ:  CGProfileEntry {
+; OBJ:    From: a
+; OBJ:    To: b
+; OBJ:    Weight: 32
+; OBJ:  }
+; OBJ:  CGProfileEntry {
+; OBJ:    From: freq
+; OBJ:    To: a
+; OBJ:    Weight: 11
+; OBJ:  }
+; OBJ:  CGProfileEntry {
+; OBJ:    From: freq
+; OBJ:    To: b
+; OBJ:    Weight: 20
+; OBJ:  }
+; OBJ:]