Supporting -f(no)-reorder-functions flag, llvm side change

[twoh]

Summary

When profile data is given, .hot/.unlikely section prefix is
added to hot/cold functions for linker to improve code locality. GCC
controls this behavior with '-f(no)-reorder-functions' flag, while LLVM
uses opt tool option '-profile-guided-section-prefix=true/false'. This
patch is for LLVM to support the same driver flag with GCC. Clang side
patch with test cases will follow.

[twoh]

https://reviews.llvm.org/D34796 is clang side change.

[twoh]

Ping. Thanks!

[davidxl]

lib/CodeGen/CodeGenPrepare.cpp

411

Can TM be null?

[twoh]

@davidxl I think it is theoretically possible, if the if branch is not taken on line 294. Did I miss something? Thanks!

[xbolva00]

Ping, plesse rebase

[lebedev.ri]

Test?

[twoh]

Rebase. Tests are provided in the clang counterpart (D34796).

include/llvm/Target/TargetMachine.h

public:
   void setSupportsDefaultOutlining(bool Enable) {
     Options.SupportsDefaultOutlining = Enable;
   }
 
   bool shouldPrintMachineCode() const { return Options.PrintMachineCode; }
 
   bool getUniqueSectionNames() const { return Options.UniqueSectionNames; }
 
+  /// Return true if profile-guided section prefix that indicates hotness of
+  /// the function should be added, corresponds to -freorder-functions.
+  bool getReorderFunctions() const { return Options.ReorderFunctions; }
+
   /// Return true if data objects should be emitted into their own section,
   /// corresponds to -fdata-sections.
   bool getDataSections() const {
     return Options.DataSections;
   }
 
   /// Return true if functions should be emitted into their own section,
   /// corresponding to -ffunction-sections.

include/llvm/Target/TargetOptions.h

TargetOptions()
         : PrintMachineCode(false), UnsafeFPMath(false), NoInfsFPMath(false),
           NoNaNsFPMath(false), NoTrappingFPMath(false),
           NoSignedZerosFPMath(false),
           HonorSignDependentRoundingFPMathOption(false), NoZerosInBSS(false),
           GuaranteedTailCallOpt(false), StackSymbolOrdering(true),
           EnableFastISel(false), EnableGlobalISel(false), UseInitArray(false),
           DisableIntegratedAS(false), RelaxELFRelocations(false),
           FunctionSections(false), DataSections(false),
-          UniqueSectionNames(true), TrapUnreachable(false),
-          NoTrapAfterNoreturn(false), EmulatedTLS(false),
-          ExplicitEmulatedTLS(false), EnableIPRA(false),
+          UniqueSectionNames(true), ReorderFunctions(true),
+          TrapUnreachable(false), NoTrapAfterNoreturn(false),
+          EmulatedTLS(false), ExplicitEmulatedTLS(false), EnableIPRA(false),
           EmitStackSizeSection(false), EnableMachineOutliner(false),
           SupportsDefaultOutlining(false), EmitAddrsig(false) {}
 
     /// PrintMachineCode - This flag is enabled when the -print-machineinstrs
     /// option is specified on the command line, and should enable debugging
     /// output from the code generator.
     unsigned PrintMachineCode : 1;
 
public:
     /// Emit functions into separate sections.
     unsigned FunctionSections : 1;
 
     /// Emit data into separate sections.
     unsigned DataSections : 1;
 
     unsigned UniqueSectionNames : 1;
 
+    /// Add section prefix for hot/cold functions.
+    unsigned ReorderFunctions : 1;
+
     /// Emit target-specific trap instruction for 'unreachable' IR instructions.
     unsigned TrapUnreachable : 1;
 
     /// Do not emit a trap instruction for 'unreachable' IR instructions behind
     /// noreturn calls, even if TrapUnreachable is true.
     unsigned NoTrapAfterNoreturn : 1;
 
     /// EmulatedTLS - This flag enables emulated TLS model, using emutls

lib/CodeGen/CodeGenPrepare.cpp

bool CodeGenPrepare::runOnFunction(Function &F) {
   TTI = &getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
   LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
   BPI.reset(new BranchProbabilityInfo(F, *LI));
   BFI.reset(new BlockFrequencyInfo(F, *BPI, *LI));
   OptSize = F.optForSize();
 
   ProfileSummaryInfo *PSI =
       getAnalysis<ProfileSummaryInfoWrapperPass>().getPSI();
-  if (ProfileGuidedSectionPrefix) {
+  bool AddProfileGuidedSectionPrefix =
+      (!TM || ProfileGuidedSectionPrefix.getNumOccurrences())
+          ? ProfileGuidedSectionPrefix
+          : TM->getReorderFunctions();
+  if (AddProfileGuidedSectionPrefix) {
     if (PSI->isFunctionHotInCallGraph(&F, *BFI))
       F.setSectionPrefix(".hot");
     else if (PSI->isFunctionColdInCallGraph(&F, *BFI))
       F.setSectionPrefix(".unlikely");
   }
 
   /// This optimization identifies DIV instructions that can be
   /// profitably bypassed and carried out with a shorter, faster divide.
   if (!OptSize && !PSI->hasHugeWorkingSetSize() && TLI &&
       TLI->isSlowDivBypassed()) {
     const DenseMap<unsigned int, unsigned int> &BypassWidths =
        TLI->getBypassSlowDivWidths();
     BasicBlock* BB = &*F.begin();