llvm/lib/Transforms/IPO/SampleProfile.cpp

//===- SampleProfile.cpp - Incorporate sample profiles into the IR --------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file implements the SampleProfileLoader transformation. This pass
// reads a profile file generated by a sampling profiler (e.g. Linux Perf -
// http://perf.wiki.kernel.org/) and generates IR metadata to reflect the
// profile information in the given profile.
//
// This pass generates branch weight annotations on the IR:
//
// - prof: Represents branch weights. This annotation is added to branches
//      to indicate the weights of each edge coming out of the branch.
//      The weight of each edge is the weight of the target block for
//      that edge. The weight of a block B is computed as the maximum
//      number of samples found in B.
//
//===----------------------------------------------------------------------===//

#include "llvm/Transforms/IPO/SampleProfile.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/DenseSet.h"
#include "llvm/ADT/None.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/StringMap.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/ADT/Twine.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/InlineCost.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/OptimizationRemarkEmitter.h"
#include "llvm/Analysis/PostDominators.h"
#include "llvm/Analysis/ProfileSummaryInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/CFG.h"
#include "llvm/IR/CallSite.h"
#include "llvm/IR/DebugInfoMetadata.h"
#include "llvm/IR/DebugLoc.h"
#include "llvm/IR/DiagnosticInfo.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/GlobalValue.h"
#include "llvm/IR/InstrTypes.h"
#include "llvm/IR/Instruction.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/MDBuilder.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/PassManager.h"
#include "llvm/IR/ValueSymbolTable.h"
#include "llvm/Pass.h"
#include "llvm/ProfileData/InstrProf.h"
#include "llvm/ProfileData/SampleProf.h"
#include "llvm/ProfileData/SampleProfReader.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/ErrorOr.h"
#include "llvm/Support/GenericDomTree.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/IPO.h"
#include "llvm/Transforms/Instrumentation.h"
#include "llvm/Transforms/Utils/CallPromotionUtils.h"
#include "llvm/Transforms/Utils/Cloning.h"
#include <algorithm>
#include <cassert>
#include <cstdint>
#include <functional>
#include <limits>
#include <map>
#include <memory>
#include <string>
#include <system_error>
#include <utility>
#include <vector>

using namespace llvm;
using namespace sampleprof;
using ProfileCount = Function::ProfileCount;
#define DEBUG_TYPE "sample-profile"

// Command line option to specify the file to read samples from. This is
// mainly used for debugging.
static cl::opt<std::string> SampleProfileFile(
    "sample-profile-file", cl::init(""), cl::value_desc("filename"),
    cl::desc("Profile file loaded by -sample-profile"), cl::Hidden);

// The named file contains a set of transformations that may have been applied
// to the symbol names between the program from which the sample data was
// collected and the current program's symbols.
static cl::opt<std::string> SampleProfileRemappingFile(
    "sample-profile-remapping-file", cl::init(""), cl::value_desc("filename"),
    cl::desc("Profile remapping file loaded by -sample-profile"), cl::Hidden);

static cl::opt<unsigned> SampleProfileMaxPropagateIterations(
    "sample-profile-max-propagate-iterations", cl::init(100),
    cl::desc("Maximum number of iterations to go through when propagating "
             "sample block/edge weights through the CFG."));

static cl::opt<unsigned> SampleProfileRecordCoverage(
    "sample-profile-check-record-coverage", cl::init(0), cl::value_desc("N"),
    cl::desc("Emit a warning if less than N% of records in the input profile "
             "are matched to the IR."));

static cl::opt<unsigned> SampleProfileSampleCoverage(
    "sample-profile-check-sample-coverage", cl::init(0), cl::value_desc("N"),
    cl::desc("Emit a warning if less than N% of samples in the input profile "
             "are matched to the IR."));

static cl::opt<bool> NoWarnSampleUnused(
    "no-warn-sample-unused", cl::init(false), cl::Hidden,
    cl::desc("Use this option to turn off/on warnings about function with "
             "samples but without debug information to use those samples. "));

static cl::opt<bool> ProfileSampleAccurate(
    "profile-sample-accurate", cl::Hidden, cl::init(false),
    cl::desc("If the sample profile is accurate, we will mark all un-sampled "
             "callsite and function as having 0 samples. Otherwise, treat "
             "un-sampled callsites and functions conservatively as unknown. "));

namespace {

using BlockWeightMap = DenseMap<const BasicBlock *, uint64_t>;
using EquivalenceClassMap = DenseMap<const BasicBlock *, const BasicBlock *>;
using Edge = std::pair<const BasicBlock *, const BasicBlock *>;
using EdgeWeightMap = DenseMap<Edge, uint64_t>;
using BlockEdgeMap =
    DenseMap<const BasicBlock *, SmallVector<const BasicBlock *, 8>>;

class SampleCoverageTracker {
public:
  SampleCoverageTracker() = default;

  bool markSamplesUsed(const FunctionSamples *FS, uint32_t LineOffset,
                       uint32_t Discriminator, uint64_t Samples);
  unsigned computeCoverage(unsigned Used, unsigned Total) const;
  unsigned countUsedRecords(const FunctionSamples *FS,
                            ProfileSummaryInfo *PSI) const;
  unsigned countBodyRecords(const FunctionSamples *FS,
                            ProfileSummaryInfo *PSI) const;
  uint64_t getTotalUsedSamples() const { return TotalUsedSamples; }
  uint64_t countBodySamples(const FunctionSamples *FS,
                            ProfileSummaryInfo *PSI) const;

  void clear() {
    SampleCoverage.clear();
    TotalUsedSamples = 0;
  }

private:
  using BodySampleCoverageMap = std::map<LineLocation, unsigned>;
  using FunctionSamplesCoverageMap =
      DenseMap<const FunctionSamples *, BodySampleCoverageMap>;

  /// Coverage map for sampling records.
  ///
  /// This map keeps a record of sampling records that have been matched to
  /// an IR instruction. This is used to detect some form of staleness in
  /// profiles (see flag -sample-profile-check-coverage).
  ///
  /// Each entry in the map corresponds to a FunctionSamples instance.  This is
  /// another map that counts how many times the sample record at the
  /// given location has been used.
  FunctionSamplesCoverageMap SampleCoverage;

  /// Number of samples used from the profile.
  ///
  /// When a sampling record is used for the first time, the samples from
  /// that record are added to this accumulator.  Coverage is later computed
  /// based on the total number of samples available in this function and
  /// its callsites.
  ///
  /// Note that this accumulator tracks samples used from a single function
  /// and all the inlined callsites. Strictly, we should have a map of counters
  /// keyed by FunctionSamples pointers, but these stats are cleared after
  /// every function, so we just need to keep a single counter.
  uint64_t TotalUsedSamples = 0;
};

/// Sample profile pass.
///
/// This pass reads profile data from the file specified by
/// -sample-profile-file and annotates every affected function with the
/// profile information found in that file.
class SampleProfileLoader {
public:
  SampleProfileLoader(
      StringRef Name, StringRef RemapName, bool IsThinLTOPreLink,
      std::function<AssumptionCache &(Function &)> GetAssumptionCache,
      std::function<TargetTransformInfo &(Function &)> GetTargetTransformInfo)
      : GetAC(std::move(GetAssumptionCache)),
        GetTTI(std::move(GetTargetTransformInfo)), Filename(Name),
        RemappingFilename(RemapName), IsThinLTOPreLink(IsThinLTOPreLink) {}

  bool doInitialization(Module &M);
  bool runOnModule(Module &M, ModuleAnalysisManager *AM,
                   ProfileSummaryInfo *_PSI);

  void dump() { Reader->dump(); }

protected:
  bool runOnFunction(Function &F, ModuleAnalysisManager *AM);
  unsigned getFunctionLoc(Function &F);
  bool emitAnnotations(Function &F);
  ErrorOr<uint64_t> getInstWeight(const Instruction &I);
  ErrorOr<uint64_t> getBlockWeight(const BasicBlock *BB);
  const FunctionSamples *findCalleeFunctionSamples(const Instruction &I) const;
  std::vector<const FunctionSamples *>
  findIndirectCallFunctionSamples(const Instruction &I, uint64_t &Sum) const;
  mutable DenseMap<const DILocation *, const FunctionSamples *> DILocation2SampleMap;
  const FunctionSamples *findFunctionSamples(const Instruction &I) const;
  bool inlineCallInstruction(Instruction *I);
  bool inlineHotFunctions(Function &F,
                          DenseSet<GlobalValue::GUID> &InlinedGUIDs);
  void printEdgeWeight(raw_ostream &OS, Edge E);
  void printBlockWeight(raw_ostream &OS, const BasicBlock *BB) const;
  void printBlockEquivalence(raw_ostream &OS, const BasicBlock *BB);
  bool computeBlockWeights(Function &F);
  void findEquivalenceClasses(Function &F);
  template <bool IsPostDom>
  void findEquivalencesFor(BasicBlock *BB1, ArrayRef<BasicBlock *> Descendants,
                           DominatorTreeBase<BasicBlock, IsPostDom> *DomTree);

  void propagateWeights(Function &F);
  uint64_t visitEdge(Edge E, unsigned *NumUnknownEdges, Edge *UnknownEdge);
  void buildEdges(Function &F);
  bool propagateThroughEdges(Function &F, bool UpdateBlockCount);
  void computeDominanceAndLoopInfo(Function &F);
  void clearFunctionData();

  /// Map basic blocks to their computed weights.
  ///
  /// The weight of a basic block is defined to be the maximum
  /// of all the instruction weights in that block.
  BlockWeightMap BlockWeights;

  /// Map edges to their computed weights.
  ///
  /// Edge weights are computed by propagating basic block weights in
  /// SampleProfile::propagateWeights.
  EdgeWeightMap EdgeWeights;

  /// Set of visited blocks during propagation.
  SmallPtrSet<const BasicBlock *, 32> VisitedBlocks;

  /// Set of visited edges during propagation.
  SmallSet<Edge, 32> VisitedEdges;

  /// Equivalence classes for block weights.
  ///
  /// Two blocks BB1 and BB2 are in the same equivalence class if they
  /// dominate and post-dominate each other, and they are in the same loop
  /// nest. When this happens, the two blocks are guaranteed to execute
  /// the same number of times.
  EquivalenceClassMap EquivalenceClass;

  /// Map from function name to Function *. Used to find the function from
  /// the function name. If the function name contains suffix, additional
  /// entry is added to map from the stripped name to the function if there
  /// is one-to-one mapping.
  StringMap<Function *> SymbolMap;

  /// Dominance, post-dominance and loop information.
  std::unique_ptr<DominatorTree> DT;
  std::unique_ptr<PostDominatorTree> PDT;
  std::unique_ptr<LoopInfo> LI;

  std::function<AssumptionCache &(Function &)> GetAC;
  std::function<TargetTransformInfo &(Function &)> GetTTI;

  /// Predecessors for each basic block in the CFG.
  BlockEdgeMap Predecessors;

  /// Successors for each basic block in the CFG.
  BlockEdgeMap Successors;

  SampleCoverageTracker CoverageTracker;

  /// Profile reader object.
  std::unique_ptr<SampleProfileReader> Reader;

  /// Samples collected for the body of this function.
  FunctionSamples *Samples = nullptr;

  /// Name of the profile file to load.
  std::string Filename;

  /// Name of the profile remapping file to load.
  std::string RemappingFilename;

  /// Flag indicating whether the profile input loaded successfully.
  bool ProfileIsValid = false;

  /// Flag indicating if the pass is invoked in ThinLTO compile phase.
  ///
  /// In this phase, in annotation, we should not promote indirect calls.
  /// Instead, we will mark GUIDs that needs to be annotated to the function.
  bool IsThinLTOPreLink;

  /// Profile Summary Info computed from sample profile.
  ProfileSummaryInfo *PSI = nullptr;

  /// Total number of samples collected in this profile.
  ///
  /// This is the sum of all the samples collected in all the functions executed
  /// at runtime.
  uint64_t TotalCollectedSamples = 0;

  /// Optimization Remark Emitter used to emit diagnostic remarks.
  OptimizationRemarkEmitter *ORE = nullptr;

  // Information recorded when we declined to inline a call site
  // because we have determined it is too cold is accumulated for
  // each callee function. Initially this is just the entry count.
  struct NotInlinedProfileInfo {
    uint64_t entryCount;
  };
  DenseMap<Function *, NotInlinedProfileInfo> notInlinedCallInfo;
};

class SampleProfileLoaderLegacyPass : public ModulePass {
public:
  // Class identification, replacement for typeinfo
  static char ID;

  SampleProfileLoaderLegacyPass(StringRef Name = SampleProfileFile,
                                bool IsThinLTOPreLink = false)
      : ModulePass(ID),
        SampleLoader(Name, SampleProfileRemappingFile, IsThinLTOPreLink,
                     [&](Function &F) -> AssumptionCache & {
                       return ACT->getAssumptionCache(F);
                     },
                     [&](Function &F) -> TargetTransformInfo & {
                       return TTIWP->getTTI(F);
                     }) {
    initializeSampleProfileLoaderLegacyPassPass(
        *PassRegistry::getPassRegistry());
  }

  void dump() { SampleLoader.dump(); }

  bool doInitialization(Module &M) override {
    return SampleLoader.doInitialization(M);
  }

  StringRef getPassName() const override { return "Sample profile pass"; }
  bool runOnModule(Module &M) override;

  void getAnalysisUsage(AnalysisUsage &AU) const override {
    AU.addRequired<AssumptionCacheTracker>();
    AU.addRequired<TargetTransformInfoWrapperPass>();
    AU.addRequired<ProfileSummaryInfoWrapperPass>();
  }

private:
  SampleProfileLoader SampleLoader;
  AssumptionCacheTracker *ACT = nullptr;
  TargetTransformInfoWrapperPass *TTIWP = nullptr;
};

} // end anonymous namespace

/// Return true if the given callsite is hot wrt to hot cutoff threshold.
///
/// Functions that were inlined in the original binary will be represented
/// in the inline stack in the sample profile. If the profile shows that
/// the original inline decision was "good" (i.e., the callsite is executed
/// frequently), then we will recreate the inline decision and apply the
/// profile from the inlined callsite.
///
/// To decide whether an inlined callsite is hot, we compare the callsite
/// sample count with the hot cutoff computed by ProfileSummaryInfo, it is
/// regarded as hot if the count is above the cutoff value.
static bool callsiteIsHot(const FunctionSamples *CallsiteFS,
                          ProfileSummaryInfo *PSI) {
  if (!CallsiteFS)
    return false; // The callsite was not inlined in the original binary.

  assert(PSI && "PSI is expected to be non null");
  uint64_t CallsiteTotalSamples = CallsiteFS->getTotalSamples();
  return PSI->isHotCount(CallsiteTotalSamples);
}

/// Mark as used the sample record for the given function samples at
/// (LineOffset, Discriminator).
///
/// \returns true if this is the first time we mark the given record.
bool SampleCoverageTracker::markSamplesUsed(const FunctionSamples *FS,
                                            uint32_t LineOffset,
                                            uint32_t Discriminator,
                                            uint64_t Samples) {
  LineLocation Loc(LineOffset, Discriminator);
  unsigned &Count = SampleCoverage[FS][Loc];
  bool FirstTime = (++Count == 1);
  if (FirstTime)
    TotalUsedSamples += Samples;
  return FirstTime;
}

/// Return the number of sample records that were applied from this profile.
///
/// This count does not include records from cold inlined callsites.
unsigned
SampleCoverageTracker::countUsedRecords(const FunctionSamples *FS,
                                        ProfileSummaryInfo *PSI) const {
  auto I = SampleCoverage.find(FS);

  // The size of the coverage map for FS represents the number of records
  // that were marked used at least once.
  unsigned Count = (I != SampleCoverage.end()) ? I->second.size() : 0;

  // If there are inlined callsites in this function, count the samples found
  // in the respective bodies. However, do not bother counting callees with 0
  // total samples, these are callees that were never invoked at runtime.
  for (const auto &I : FS->getCallsiteSamples())
    for (const auto &J : I.second) {
      const FunctionSamples *CalleeSamples = &J.second;
      if (callsiteIsHot(CalleeSamples, PSI))
        Count += countUsedRecords(CalleeSamples, PSI);
    }

  return Count;
}

/// Return the number of sample records in the body of this profile.
///
/// This count does not include records from cold inlined callsites.
unsigned
SampleCoverageTracker::countBodyRecords(const FunctionSamples *FS,
                                        ProfileSummaryInfo *PSI) const {
  unsigned Count = FS->getBodySamples().size();

  // Only count records in hot callsites.
  for (const auto &I : FS->getCallsiteSamples())
    for (const auto &J : I.second) {
      const FunctionSamples *CalleeSamples = &J.second;
      if (callsiteIsHot(CalleeSamples, PSI))
        Count += countBodyRecords(CalleeSamples, PSI);
    }

  return Count;
}

/// Return the number of samples collected in the body of this profile.
///
/// This count does not include samples from cold inlined callsites.
uint64_t
SampleCoverageTracker::countBodySamples(const FunctionSamples *FS,
                                        ProfileSummaryInfo *PSI) const {
  uint64_t Total = 0;
  for (const auto &I : FS->getBodySamples())
    Total += I.second.getSamples();

  // Only count samples in hot callsites.
  for (const auto &I : FS->getCallsiteSamples())
    for (const auto &J : I.second) {
      const FunctionSamples *CalleeSamples = &J.second;
      if (callsiteIsHot(CalleeSamples, PSI))
        Total += countBodySamples(CalleeSamples, PSI);
    }

  return Total;
}

/// Return the fraction of sample records used in this profile.
///
/// The returned value is an unsigned integer in the range 0-100 indicating
/// the percentage of sample records that were used while applying this
/// profile to the associated function.
unsigned SampleCoverageTracker::computeCoverage(unsigned Used,
                                                unsigned Total) const {
  assert(Used <= Total &&
         "number of used records cannot exceed the total number of records");
  return Total > 0 ? Used * 100 / Total : 100;
}

/// Clear all the per-function data used to load samples and propagate weights.
void SampleProfileLoader::clearFunctionData() {
  BlockWeights.clear();
  EdgeWeights.clear();
  VisitedBlocks.clear();
  VisitedEdges.clear();
  EquivalenceClass.clear();
  DT = nullptr;
  PDT = nullptr;
  LI = nullptr;
  Predecessors.clear();
  Successors.clear();
  CoverageTracker.clear();
}

#ifndef NDEBUG
/// Print the weight of edge \p E on stream \p OS.
///
/// \param OS  Stream to emit the output to.
/// \param E  Edge to print.
void SampleProfileLoader::printEdgeWeight(raw_ostream &OS, Edge E) {
  OS << "weight[" << E.first->getName() << "->" << E.second->getName()
     << "]: " << EdgeWeights[E] << "\n";
}

/// Print the equivalence class of block \p BB on stream \p OS.
///
/// \param OS  Stream to emit the output to.
/// \param BB  Block to print.
void SampleProfileLoader::printBlockEquivalence(raw_ostream &OS,
                                                const BasicBlock *BB) {
  const BasicBlock *Equiv = EquivalenceClass[BB];
  OS << "equivalence[" << BB->getName()
     << "]: " << ((Equiv) ? EquivalenceClass[BB]->getName() : "NONE") << "\n";
}

/// Print the weight of block \p BB on stream \p OS.
///
/// \param OS  Stream to emit the output to.
/// \param BB  Block to print.
void SampleProfileLoader::printBlockWeight(raw_ostream &OS,
                                           const BasicBlock *BB) const {
  const auto &I = BlockWeights.find(BB);
  uint64_t W = (I == BlockWeights.end() ? 0 : I->second);
  OS << "weight[" << BB->getName() << "]: " << W << "\n";
}
#endif

/// Get the weight for an instruction.
///
/// The "weight" of an instruction \p Inst is the number of samples
/// collected on that instruction at runtime. To retrieve it, we
/// need to compute the line number of \p Inst relative to the start of its
/// function. We use HeaderLineno to compute the offset. We then
/// look up the samples collected for \p Inst using BodySamples.
///
/// \param Inst Instruction to query.
///
/// \returns the weight of \p Inst.
ErrorOr<uint64_t> SampleProfileLoader::getInstWeight(const Instruction &Inst) {
  const DebugLoc &DLoc = Inst.getDebugLoc();
  if (!DLoc)
    return std::error_code();

  const FunctionSamples *FS = findFunctionSamples(Inst);
  if (!FS)
    return std::error_code();

  // Ignore all intrinsics, phinodes and branch instructions.
  // Branch and phinodes instruction usually contains debug info from sources outside of
  // the residing basic block, thus we ignore them during annotation.
  if (isa<BranchInst>(Inst) || isa<IntrinsicInst>(Inst) || isa<PHINode>(Inst))
    return std::error_code();

  // If a direct call/invoke instruction is inlined in profile
  // (findCalleeFunctionSamples returns non-empty result), but not inlined here,
  // it means that the inlined callsite has no sample, thus the call
  // instruction should have 0 count.
  if ((isa<CallInst>(Inst) || isa<InvokeInst>(Inst)) &&
      !ImmutableCallSite(&Inst).isIndirectCall() &&
      findCalleeFunctionSamples(Inst))
    return 0;

  const DILocation *DIL = DLoc;
  uint32_t LineOffset = FunctionSamples::getOffset(DIL);
  uint32_t Discriminator = DIL->getBaseDiscriminator();
  ErrorOr<uint64_t> R = FS->findSamplesAt(LineOffset, Discriminator);
  if (R) {
    bool FirstMark =
        CoverageTracker.markSamplesUsed(FS, LineOffset, Discriminator, R.get());
    if (FirstMark) {
      ORE->emit([&]() {
        OptimizationRemarkAnalysis Remark(DEBUG_TYPE, "AppliedSamples", &Inst);
        Remark << "Applied " << ore::NV("NumSamples", *R);
        Remark << " samples from profile (offset: ";
        Remark << ore::NV("LineOffset", LineOffset);
        if (Discriminator) {
          Remark << ".";
          Remark << ore::NV("Discriminator", Discriminator);
        }
        Remark << ")";
        return Remark;
      });
    }
    LLVM_DEBUG(dbgs() << "    " << DLoc.getLine() << "."
                      << DIL->getBaseDiscriminator() << ":" << Inst
                      << " (line offset: " << LineOffset << "."
                      << DIL->getBaseDiscriminator() << " - weight: " << R.get()
                      << ")\n");
  }
  return R;
}

/// Compute the weight of a basic block.
///
/// The weight of basic block \p BB is the maximum weight of all the
/// instructions in BB.
///
/// \param BB The basic block to query.
///
/// \returns the weight for \p BB.
ErrorOr<uint64_t> SampleProfileLoader::getBlockWeight(const BasicBlock *BB) {
  uint64_t Max = 0;
  bool HasWeight = false;
  for (auto &I : BB->getInstList()) {
    const ErrorOr<uint64_t> &R = getInstWeight(I);
    if (R) {
      Max = std::max(Max, R.get());
      HasWeight = true;
    }
  }
  return HasWeight ? ErrorOr<uint64_t>(Max) : std::error_code();
}

/// Compute and store the weights of every basic block.
///
/// This populates the BlockWeights map by computing
/// the weights of every basic block in the CFG.
///
/// \param F The function to query.
bool SampleProfileLoader::computeBlockWeights(Function &F) {
  bool Changed = false;
  LLVM_DEBUG(dbgs() << "Block weights\n");
  for (const auto &BB : F) {
    ErrorOr<uint64_t> Weight = getBlockWeight(&BB);
    if (Weight) {
      BlockWeights[&BB] = Weight.get();
      VisitedBlocks.insert(&BB);
      Changed = true;
    }
    LLVM_DEBUG(printBlockWeight(dbgs(), &BB));
  }

  return Changed;
}

/// Get the FunctionSamples for a call instruction.
///
/// The FunctionSamples of a call/invoke instruction \p Inst is the inlined
/// instance in which that call instruction is calling to. It contains
/// all samples that resides in the inlined instance. We first find the
/// inlined instance in which the call instruction is from, then we
/// traverse its children to find the callsite with the matching
/// location.
///
/// \param Inst Call/Invoke instruction to query.
///
/// \returns The FunctionSamples pointer to the inlined instance.
const FunctionSamples *
SampleProfileLoader::findCalleeFunctionSamples(const Instruction &Inst) const {
  const DILocation *DIL = Inst.getDebugLoc();
  if (!DIL) {
    return nullptr;
  }

  StringRef CalleeName;
  if (const CallInst *CI = dyn_cast<CallInst>(&Inst))
    if (Function *Callee = CI->getCalledFunction())
      CalleeName = Callee->getName();

  const FunctionSamples *FS = findFunctionSamples(Inst);
  if (FS == nullptr)
    return nullptr;

  return FS->findFunctionSamplesAt(LineLocation(FunctionSamples::getOffset(DIL),
                                                DIL->getBaseDiscriminator()),
                                   CalleeName);
}

/// Returns a vector of FunctionSamples that are the indirect call targets
/// of \p Inst. The vector is sorted by the total number of samples. Stores
/// the total call count of the indirect call in \p Sum.
std::vector<const FunctionSamples *>
SampleProfileLoader::findIndirectCallFunctionSamples(
    const Instruction &Inst, uint64_t &Sum) const {
  const DILocation *DIL = Inst.getDebugLoc();
  std::vector<const FunctionSamples *> R;

  if (!DIL) {
    return R;
  }

  const FunctionSamples *FS = findFunctionSamples(Inst);
  if (FS == nullptr)
    return R;

  uint32_t LineOffset = FunctionSamples::getOffset(DIL);
  uint32_t Discriminator = DIL->getBaseDiscriminator();

  auto T = FS->findCallTargetMapAt(LineOffset, Discriminator);
  Sum = 0;
  if (T)
    for (const auto &T_C : T.get())
      Sum += T_C.second;
  if (const FunctionSamplesMap *M = FS->findFunctionSamplesMapAt(LineLocation(
          FunctionSamples::getOffset(DIL), DIL->getBaseDiscriminator()))) {
    if (M->empty())
      return R;
    for (const auto &NameFS : *M) {
      Sum += NameFS.second.getEntrySamples();
      R.push_back(&NameFS.second);
    }
    llvm::sort(R, [](const FunctionSamples *L, const FunctionSamples *R) {
      if (L->getEntrySamples() != R->getEntrySamples())
        return L->getEntrySamples() > R->getEntrySamples();
      return FunctionSamples::getGUID(L->getName()) <
             FunctionSamples::getGUID(R->getName());
    });
  }
  return R;
}

/// Get the FunctionSamples for an instruction.
///
/// The FunctionSamples of an instruction \p Inst is the inlined instance
/// in which that instruction is coming from. We traverse the inline stack
/// of that instruction, and match it with the tree nodes in the profile.
///
/// \param Inst Instruction to query.
///
/// \returns the FunctionSamples pointer to the inlined instance.
const FunctionSamples *
SampleProfileLoader::findFunctionSamples(const Instruction &Inst) const {
  const DILocation *DIL = Inst.getDebugLoc();
  if (!DIL)
    return Samples;

  auto it = DILocation2SampleMap.try_emplace(DIL,nullptr);
  if (it.second)
    it.first->second = Samples->findFunctionSamples(DIL);
  return it.first->second;
}

bool SampleProfileLoader::inlineCallInstruction(Instruction *I) {
  assert(isa<CallInst>(I) || isa<InvokeInst>(I));
  CallSite CS(I);
  Function *CalledFunction = CS.getCalledFunction();
  assert(CalledFunction);
  DebugLoc DLoc = I->getDebugLoc();
  BasicBlock *BB = I->getParent();
  InlineParams Params = getInlineParams();
  Params.ComputeFullInlineCost = true;
  // Checks if there is anything in the reachable portion of the callee at
  // this callsite that makes this inlining potentially illegal. Need to
  // set ComputeFullInlineCost, otherwise getInlineCost may return early
  // when cost exceeds threshold without checking all IRs in the callee.
  // The acutal cost does not matter because we only checks isNever() to
  // see if it is legal to inline the callsite.
  InlineCost Cost = getInlineCost(CS, Params, GetTTI(*CalledFunction), GetAC,
                                  None, nullptr, nullptr);
  if (Cost.isNever()) {
    ORE->emit(OptimizationRemark(DEBUG_TYPE, "Not inline", DLoc, BB)
              << "incompatible inlining");
    return false;
  }
  InlineFunctionInfo IFI(nullptr, &GetAC);
  if (InlineFunction(CS, IFI)) {
    // The call to InlineFunction erases I, so we can't pass it here.
    ORE->emit(OptimizationRemark(DEBUG_TYPE, "HotInline", DLoc, BB)
              << "inlined hot callee '" << ore::NV("Callee", CalledFunction)
              << "' into '" << ore::NV("Caller", BB->getParent()) << "'");
    return true;
  }
  return false;
}

/// Iteratively inline hot callsites of a function.
///
/// Iteratively traverse all callsites of the function \p F, and find if
/// the corresponding inlined instance exists and is hot in profile. If
/// it is hot enough, inline the callsites and adds new callsites of the
/// callee into the caller. If the call is an indirect call, first promote
/// it to direct call. Each indirect call is limited with a single target.
///
/// \param F function to perform iterative inlining.
/// \param InlinedGUIDs a set to be updated to include all GUIDs that are
///     inlined in the profiled binary.
///
/// \returns True if there is any inline happened.
bool SampleProfileLoader::inlineHotFunctions(
    Function &F, DenseSet<GlobalValue::GUID> &InlinedGUIDs) {
  DenseSet<Instruction *> PromotedInsns;

  DenseMap<Instruction *, const FunctionSamples *> localNotInlinedCallSites;
  bool Changed = false;
  while (true) {
    bool LocalChanged = false;
    SmallVector<Instruction *, 10> CIS;
    for (auto &BB : F) {
      bool Hot = false;
      SmallVector<Instruction *, 10> Candidates;
      for (auto &I : BB.getInstList()) {
        const FunctionSamples *FS = nullptr;
        if ((isa<CallInst>(I) || isa<InvokeInst>(I)) &&
            !isa<IntrinsicInst>(I) && (FS = findCalleeFunctionSamples(I))) {
          Candidates.push_back(&I);
          if (FS->getEntrySamples() > 0)
            localNotInlinedCallSites.try_emplace(&I, FS);
          if (callsiteIsHot(FS, PSI))
            Hot = true;
        }
      }
      if (Hot) {
        CIS.insert(CIS.begin(), Candidates.begin(), Candidates.end());
      }
    }
    for (auto I : CIS) {
      Function *CalledFunction = CallSite(I).getCalledFunction();
      // Do not inline recursive calls.
      if (CalledFunction == &F)
        continue;
      if (CallSite(I).isIndirectCall()) {
        if (PromotedInsns.count(I))
          continue;
        uint64_t Sum;
        for (const auto *FS : findIndirectCallFunctionSamples(*I, Sum)) {
          if (IsThinLTOPreLink) {
            FS->findInlinedFunctions(InlinedGUIDs, F.getParent(),
                                     PSI->getOrCompHotCountThreshold());
            continue;
          }
          auto CalleeFunctionName = FS->getFuncNameInModule(F.getParent());
          // If it is a recursive call, we do not inline it as it could bloat
          // the code exponentially. There is way to better handle this, e.g.
          // clone the caller first, and inline the cloned caller if it is
          // recursive. As llvm does not inline recursive calls, we will
          // simply ignore it instead of handling it explicitly.
          if (CalleeFunctionName == F.getName())
            continue;

          const char *Reason = "Callee function not available";
          auto R = SymbolMap.find(CalleeFunctionName);
          if (R != SymbolMap.end() && R->getValue() &&
              !R->getValue()->isDeclaration() &&
              R->getValue()->getSubprogram() &&
              isLegalToPromote(CallSite(I), R->getValue(), &Reason)) {
            uint64_t C = FS->getEntrySamples();
            Instruction *DI =
                pgo::promoteIndirectCall(I, R->getValue(), C, Sum, false, ORE);
            Sum -= C;
            PromotedInsns.insert(I);
            // If profile mismatches, we should not attempt to inline DI.
            if ((isa<CallInst>(DI) || isa<InvokeInst>(DI)) &&
                inlineCallInstruction(DI)) {
              localNotInlinedCallSites.erase(I);
              LocalChanged = true;
            }
          } else {
            LLVM_DEBUG(dbgs()
                       << "\nFailed to promote indirect call to "
                       << CalleeFunctionName << " because " << Reason << "\n");
          }
        }
      } else if (CalledFunction && CalledFunction->getSubprogram() &&
                 !CalledFunction->isDeclaration()) {
        if (inlineCallInstruction(I)) {
          localNotInlinedCallSites.erase(I);
          LocalChanged = true;
        }
      } else if (IsThinLTOPreLink) {
        findCalleeFunctionSamples(*I)->findInlinedFunctions(
            InlinedGUIDs, F.getParent(), PSI->getOrCompHotCountThreshold());
      }
    }
    if (LocalChanged) {
      Changed = true;
    } else {
      break;
    }
  }

  // Accumulate not inlined callsite information into notInlinedSamples
  for (const auto &Pair : localNotInlinedCallSites) {
    Instruction *I = Pair.getFirst();
    Function *Callee = CallSite(I).getCalledFunction();
    if (!Callee || Callee->isDeclaration())
      continue;
    const FunctionSamples *FS = Pair.getSecond();
    auto pair =
        notInlinedCallInfo.try_emplace(Callee, NotInlinedProfileInfo{0});
    pair.first->second.entryCount += FS->getEntrySamples();
  }
  return Changed;
}

/// Find equivalence classes for the given block.
///
/// This finds all the blocks that are guaranteed to execute the same
/// number of times as \p BB1. To do this, it traverses all the
/// descendants of \p BB1 in the dominator or post-dominator tree.
///
/// A block BB2 will be in the same equivalence class as \p BB1 if
/// the following holds:
///
/// 1- \p BB1 is a descendant of BB2 in the opposite tree. So, if BB2
///    is a descendant of \p BB1 in the dominator tree, then BB2 should
///    dominate BB1 in the post-dominator tree.
///
/// 2- Both BB2 and \p BB1 must be in the same loop.
///
/// For every block BB2 that meets those two requirements, we set BB2's
/// equivalence class to \p BB1.
///
/// \param BB1  Block to check.
/// \param Descendants  Descendants of \p BB1 in either the dom or pdom tree.
/// \param DomTree  Opposite dominator tree. If \p Descendants is filled
///                 with blocks from \p BB1's dominator tree, then
///                 this is the post-dominator tree, and vice versa.
template <bool IsPostDom>
void SampleProfileLoader::findEquivalencesFor(
    BasicBlock *BB1, ArrayRef<BasicBlock *> Descendants,
    DominatorTreeBase<BasicBlock, IsPostDom> *DomTree) {
  const BasicBlock *EC = EquivalenceClass[BB1];
  uint64_t Weight = BlockWeights[EC];
  for (const auto *BB2 : Descendants) {
    bool IsDomParent = DomTree->dominates(BB2, BB1);
    bool IsInSameLoop = LI->getLoopFor(BB1) == LI->getLoopFor(BB2);
    if (BB1 != BB2 && IsDomParent && IsInSameLoop) {
      EquivalenceClass[BB2] = EC;
      // If BB2 is visited, then the entire EC should be marked as visited.
      if (VisitedBlocks.count(BB2)) {
        VisitedBlocks.insert(EC);
      }

      // If BB2 is heavier than BB1, make BB2 have the same weight
      // as BB1.
      //
      // Note that we don't worry about the opposite situation here
      // (when BB2 is lighter than BB1). We will deal with this
      // during the propagation phase. Right now, we just want to
      // make sure that BB1 has the largest weight of all the
      // members of its equivalence set.
      Weight = std::max(Weight, BlockWeights[BB2]);
    }
  }
  if (EC == &EC->getParent()->getEntryBlock()) {
    BlockWeights[EC] = Samples->getHeadSamples() + 1;
  } else {
    BlockWeights[EC] = Weight;
  }
}

/// Find equivalence classes.
///
/// Since samples may be missing from blocks, we can fill in the gaps by setting
/// the weights of all the blocks in the same equivalence class to the same
/// weight. To compute the concept of equivalence, we use dominance and loop
/// information. Two blocks B1 and B2 are in the same equivalence class if B1
/// dominates B2, B2 post-dominates B1 and both are in the same loop.
///
/// \param F The function to query.
void SampleProfileLoader::findEquivalenceClasses(Function &F) {
  SmallVector<BasicBlock *, 8> DominatedBBs;
  LLVM_DEBUG(dbgs() << "\nBlock equivalence classes\n");
  // Find equivalence sets based on dominance and post-dominance information.
  for (auto &BB : F) {
    BasicBlock *BB1 = &BB;

    // Compute BB1's equivalence class once.
    if (EquivalenceClass.count(BB1)) {
      LLVM_DEBUG(printBlockEquivalence(dbgs(), BB1));
      continue;
    }

    // By default, blocks are in their own equivalence class.
    EquivalenceClass[BB1] = BB1;

    // Traverse all the blocks dominated by BB1. We are looking for
    // every basic block BB2 such that:
    //
    // 1- BB1 dominates BB2.
    // 2- BB2 post-dominates BB1.
    // 3- BB1 and BB2 are in the same loop nest.
    //
    // If all those conditions hold, it means that BB2 is executed
    // as many times as BB1, so they are placed in the same equivalence
    // class by making BB2's equivalence class be BB1.
    DominatedBBs.clear();
    DT->getDescendants(BB1, DominatedBBs);
    findEquivalencesFor(BB1, DominatedBBs, PDT.get());

    LLVM_DEBUG(printBlockEquivalence(dbgs(), BB1));
  }

  // Assign weights to equivalence classes.
  //
  // All the basic blocks in the same equivalence class will execute
  // the same number of times. Since we know that the head block in
  // each equivalence class has the largest weight, assign that weight
  // to all the blocks in that equivalence class.
  LLVM_DEBUG(
      dbgs() << "\nAssign the same weight to all blocks in the same class\n");
  for (auto &BI : F) {
    const BasicBlock *BB = &BI;
    const BasicBlock *EquivBB = EquivalenceClass[BB];
    if (BB != EquivBB)
      BlockWeights[BB] = BlockWeights[EquivBB];
    LLVM_DEBUG(printBlockWeight(dbgs(), BB));
  }
}

/// Visit the given edge to decide if it has a valid weight.
///
/// If \p E has not been visited before, we copy to \p UnknownEdge
/// and increment the count of unknown edges.
///
/// \param E  Edge to visit.
/// \param NumUnknownEdges  Current number of unknown edges.
/// \param UnknownEdge  Set if E has not been visited before.
///
/// \returns E's weight, if known. Otherwise, return 0.
uint64_t SampleProfileLoader::visitEdge(Edge E, unsigned *NumUnknownEdges,
                                        Edge *UnknownEdge) {
  if (!VisitedEdges.count(E)) {
    (*NumUnknownEdges)++;
    *UnknownEdge = E;
    return 0;
  }

  return EdgeWeights[E];
}

/// Propagate weights through incoming/outgoing edges.
///
/// If the weight of a basic block is known, and there is only one edge
/// with an unknown weight, we can calculate the weight of that edge.
///
/// Similarly, if all the edges have a known count, we can calculate the
/// count of the basic block, if needed.
///
/// \param F  Function to process.
/// \param UpdateBlockCount  Whether we should update basic block counts that
///                          has already been annotated.
///
/// \returns  True if new weights were assigned to edges or blocks.
bool SampleProfileLoader::propagateThroughEdges(Function &F,
                                                bool UpdateBlockCount) {
  bool Changed = false;
  LLVM_DEBUG(dbgs() << "\nPropagation through edges\n");
  for (const auto &BI : F) {
    const BasicBlock *BB = &BI;
    const BasicBlock *EC = EquivalenceClass[BB];

    // Visit all the predecessor and successor edges to determine
    // which ones have a weight assigned already. Note that it doesn't
    // matter that we only keep track of a single unknown edge. The
    // only case we are interested in handling is when only a single
    // edge is unknown (see setEdgeOrBlockWeight).
    for (unsigned i = 0; i < 2; i++) {
      uint64_t TotalWeight = 0;
      unsigned NumUnknownEdges = 0, NumTotalEdges = 0;
      Edge UnknownEdge, SelfReferentialEdge, SingleEdge;

      if (i == 0) {
        // First, visit all predecessor edges.
        NumTotalEdges = Predecessors[BB].size();
        for (auto *Pred : Predecessors[BB]) {
          Edge E = std::make_pair(Pred, BB);
          TotalWeight += visitEdge(E, &NumUnknownEdges, &UnknownEdge);
          if (E.first == E.second)
            SelfReferentialEdge = E;
        }
        if (NumTotalEdges == 1) {
          SingleEdge = std::make_pair(Predecessors[BB][0], BB);
        }
      } else {
        // On the second round, visit all successor edges.
        NumTotalEdges = Successors[BB].size();
        for (auto *Succ : Successors[BB]) {
          Edge E = std::make_pair(BB, Succ);
          TotalWeight += visitEdge(E, &NumUnknownEdges, &UnknownEdge);
        }
        if (NumTotalEdges == 1) {
          SingleEdge = std::make_pair(BB, Successors[BB][0]);
        }
      }

      // After visiting all the edges, there are three cases that we
      // can handle immediately:
      //
      // - All the edge weights are known (i.e., NumUnknownEdges == 0).
      //   In this case, we simply check that the sum of all the edges
      //   is the same as BB's weight. If not, we change BB's weight
      //   to match. Additionally, if BB had not been visited before,
      //   we mark it visited.
      //
      // - Only one edge is unknown and BB has already been visited.
      //   In this case, we can compute the weight of the edge by
      //   subtracting the total block weight from all the known
      //   edge weights. If the edges weight more than BB, then the
      //   edge of the last remaining edge is set to zero.
      //
      // - There exists a self-referential edge and the weight of BB is
      //   known. In this case, this edge can be based on BB's weight.
      //   We add up all the other known edges and set the weight on
      //   the self-referential edge as we did in the previous case.
      //
      // In any other case, we must continue iterating. Eventually,
      // all edges will get a weight, or iteration will stop when
      // it reaches SampleProfileMaxPropagateIterations.
      if (NumUnknownEdges <= 1) {
        uint64_t &BBWeight = BlockWeights[EC];
        if (NumUnknownEdges == 0) {
          if (!VisitedBlocks.count(EC)) {
            // If we already know the weight of all edges, the weight of the
            // basic block can be computed. It should be no larger than the sum
            // of all edge weights.
            if (TotalWeight > BBWeight) {
              BBWeight = TotalWeight;
              Changed = true;
              LLVM_DEBUG(dbgs() << "All edge weights for " << BB->getName()
                                << " known. Set weight for block: ";
                         printBlockWeight(dbgs(), BB););
            }
          } else if (NumTotalEdges == 1 &&
                     EdgeWeights[SingleEdge] < BlockWeights[EC]) {
            // If there is only one edge for the visited basic block, use the
            // block weight to adjust edge weight if edge weight is smaller.
            EdgeWeights[SingleEdge] = BlockWeights[EC];
            Changed = true;
          }
        } else if (NumUnknownEdges == 1 && VisitedBlocks.count(EC)) {
          // If there is a single unknown edge and the block has been
          // visited, then we can compute E's weight.
          if (BBWeight >= TotalWeight)
            EdgeWeights[UnknownEdge] = BBWeight - TotalWeight;
          else
            EdgeWeights[UnknownEdge] = 0;
          const BasicBlock *OtherEC;
          if (i == 0)
            OtherEC = EquivalenceClass[UnknownEdge.first];
          else
            OtherEC = EquivalenceClass[UnknownEdge.second];
          // Edge weights should never exceed the BB weights it connects.
          if (VisitedBlocks.count(OtherEC) &&
              EdgeWeights[UnknownEdge] > BlockWeights[OtherEC])
            EdgeWeights[UnknownEdge] = BlockWeights[OtherEC];
          VisitedEdges.insert(UnknownEdge);
          Changed = true;
          LLVM_DEBUG(dbgs() << "Set weight for edge: ";
                     printEdgeWeight(dbgs(), UnknownEdge));
        }
      } else if (VisitedBlocks.count(EC) && BlockWeights[EC] == 0) {
        // If a block Weights 0, all its in/out edges should weight 0.
        if (i == 0) {
          for (auto *Pred : Predecessors[BB]) {
            Edge E = std::make_pair(Pred, BB);
            EdgeWeights[E] = 0;
            VisitedEdges.insert(E);
          }
        } else {
          for (auto *Succ : Successors[BB]) {
            Edge E = std::make_pair(BB, Succ);
            EdgeWeights[E] = 0;
            VisitedEdges.insert(E);
          }
        }
      } else if (SelfReferentialEdge.first && VisitedBlocks.count(EC)) {
        uint64_t &BBWeight = BlockWeights[BB];
        // We have a self-referential edge and the weight of BB is known.
        if (BBWeight >= TotalWeight)
          EdgeWeights[SelfReferentialEdge] = BBWeight - TotalWeight;
        else
          EdgeWeights[SelfReferentialEdge] = 0;
        VisitedEdges.insert(SelfReferentialEdge);
        Changed = true;
        LLVM_DEBUG(dbgs() << "Set self-referential edge weight to: ";
                   printEdgeWeight(dbgs(), SelfReferentialEdge));
      }
      if (UpdateBlockCount && !VisitedBlocks.count(EC) && TotalWeight > 0) {
        BlockWeights[EC] = TotalWeight;
        VisitedBlocks.insert(EC);
        Changed = true;
      }
    }
  }

  return Changed;
}

/// Build in/out edge lists for each basic block in the CFG.
///
/// We are interested in unique edges. If a block B1 has multiple
/// edges to another block B2, we only add a single B1->B2 edge.
void SampleProfileLoader::buildEdges(Function &F) {
  for (auto &BI : F) {
    BasicBlock *B1 = &BI;

    // Add predecessors for B1.
    SmallPtrSet<BasicBlock *, 16> Visited;
    if (!Predecessors[B1].empty())
      llvm_unreachable("Found a stale predecessors list in a basic block.");
    for (pred_iterator PI = pred_begin(B1), PE = pred_end(B1); PI != PE; ++PI) {
      BasicBlock *B2 = *PI;
      if (Visited.insert(B2).second)
        Predecessors[B1].push_back(B2);
    }

    // Add successors for B1.
    Visited.clear();
    if (!Successors[B1].empty())
      llvm_unreachable("Found a stale successors list in a basic block.");
    for (succ_iterator SI = succ_begin(B1), SE = succ_end(B1); SI != SE; ++SI) {
      BasicBlock *B2 = *SI;
      if (Visited.insert(B2).second)
        Successors[B1].push_back(B2);
    }
  }
}

/// Returns the sorted CallTargetMap \p M by count in descending order.
static SmallVector<InstrProfValueData, 2> SortCallTargets(
    const SampleRecord::CallTargetMap &M) {
  SmallVector<InstrProfValueData, 2> R;
  for (auto I = M.begin(); I != M.end(); ++I)
    R.push_back({FunctionSamples::getGUID(I->getKey()), I->getValue()});
  llvm::sort(R, [](const InstrProfValueData &L, const InstrProfValueData &R) {
    if (L.Count == R.Count)
      return L.Value > R.Value;
    else
      return L.Count > R.Count;
  });
  return R;
}

/// Propagate weights into edges
///
/// The following rules are applied to every block BB in the CFG:
///
/// - If BB has a single predecessor/successor, then the weight
///   of that edge is the weight of the block.
///
/// - If all incoming or outgoing edges are known except one, and the
///   weight of the block is already known, the weight of the unknown
///   edge will be the weight of the block minus the sum of all the known
///   edges. If the sum of all the known edges is larger than BB's weight,
///   we set the unknown edge weight to zero.
///
/// - If there is a self-referential edge, and the weight of the block is
///   known, the weight for that edge is set to the weight of the block
///   minus the weight of the other incoming edges to that block (if
///   known).
void SampleProfileLoader::propagateWeights(Function &F) {
  bool Changed = true;
  unsigned I = 0;

  // If BB weight is larger than its corresponding loop's header BB weight,
  // use the BB weight to replace the loop header BB weight.
  for (auto &BI : F) {
    BasicBlock *BB = &BI;
    Loop *L = LI->getLoopFor(BB);
    if (!L) {
      continue;
    }
    BasicBlock *Header = L->getHeader();
    if (Header && BlockWeights[BB] > BlockWeights[Header]) {
      BlockWeights[Header] = BlockWeights[BB];
    }
  }

  // Before propagation starts, build, for each block, a list of
  // unique predecessors and successors. This is necessary to handle
  // identical edges in multiway branches. Since we visit all blocks and all
  // edges of the CFG, it is cleaner to build these lists once at the start
  // of the pass.
  buildEdges(F);

  // Propagate until we converge or we go past the iteration limit.
  while (Changed && I++ < SampleProfileMaxPropagateIterations) {
    Changed = propagateThroughEdges(F, false);
  }

  // The first propagation propagates BB counts from annotated BBs to unknown
  // BBs. The 2nd propagation pass resets edges weights, and use all BB weights
  // to propagate edge weights.
  VisitedEdges.clear();
  Changed = true;
  while (Changed && I++ < SampleProfileMaxPropagateIterations) {
    Changed = propagateThroughEdges(F, false);
  }

  // The 3rd propagation pass allows adjust annotated BB weights that are
  // obviously wrong.
  Changed = true;
  while (Changed && I++ < SampleProfileMaxPropagateIterations) {
    Changed = propagateThroughEdges(F, true);
  }

  // Generate MD_prof metadata for every branch instruction using the
  // edge weights computed during propagation.
  LLVM_DEBUG(dbgs() << "\nPropagation complete. Setting branch weights\n");
  LLVMContext &Ctx = F.getContext();
  MDBuilder MDB(Ctx);
  for (auto &BI : F) {
    BasicBlock *BB = &BI;

    if (BlockWeights[BB]) {
      for (auto &I : BB->getInstList()) {
        if (!isa<CallInst>(I) && !isa<InvokeInst>(I))
          continue;
        CallSite CS(&I);
        if (!CS.getCalledFunction()) {
          const DebugLoc &DLoc = I.getDebugLoc();
          if (!DLoc)
            continue;
          const DILocation *DIL = DLoc;
          uint32_t LineOffset = FunctionSamples::getOffset(DIL);
          uint32_t Discriminator = DIL->getBaseDiscriminator();

          const FunctionSamples *FS = findFunctionSamples(I);
          if (!FS)
            continue;
          auto T = FS->findCallTargetMapAt(LineOffset, Discriminator);
          if (!T || T.get().empty())
            continue;
          SmallVector<InstrProfValueData, 2> SortedCallTargets =
              SortCallTargets(T.get());
          uint64_t Sum;
          findIndirectCallFunctionSamples(I, Sum);
          annotateValueSite(*I.getParent()->getParent()->getParent(), I,
                            SortedCallTargets, Sum, IPVK_IndirectCallTarget,
                            SortedCallTargets.size());
        } else if (!dyn_cast<IntrinsicInst>(&I)) {
          SmallVector<uint32_t, 1> Weights;
          Weights.push_back(BlockWeights[BB]);
          I.setMetadata(LLVMContext::MD_prof, MDB.createBranchWeights(Weights));
        }
      }
    }
    Instruction *TI = BB->getTerminator();
    if (TI->getNumSuccessors() == 1)
      continue;
    if (!isa<BranchInst>(TI) && !isa<SwitchInst>(TI))
      continue;

    DebugLoc BranchLoc = TI->getDebugLoc();
    LLVM_DEBUG(dbgs() << "\nGetting weights for branch at line "
                      << ((BranchLoc) ? Twine(BranchLoc.getLine())
                                      : Twine("<UNKNOWN LOCATION>"))
                      << ".\n");
    SmallVector<uint32_t, 4> Weights;
    uint32_t MaxWeight = 0;
    Instruction *MaxDestInst;
    for (unsigned I = 0; I < TI->getNumSuccessors(); ++I) {
      BasicBlock *Succ = TI->getSuccessor(I);
      Edge E = std::make_pair(BB, Succ);
      uint64_t Weight = EdgeWeights[E];
      LLVM_DEBUG(dbgs() << "\t"; printEdgeWeight(dbgs(), E));
      // Use uint32_t saturated arithmetic to adjust the incoming weights,
      // if needed. Sample counts in profiles are 64-bit unsigned values,
      // but internally branch weights are expressed as 32-bit values.
      if (Weight > std::numeric_limits<uint32_t>::max()) {
        LLVM_DEBUG(dbgs() << " (saturated due to uint32_t overflow)");
        Weight = std::numeric_limits<uint32_t>::max();
      }
      // Weight is added by one to avoid propagation errors introduced by
      // 0 weights.
      Weights.push_back(static_cast<uint32_t>(Weight + 1));
      if (Weight != 0) {
        if (Weight > MaxWeight) {
          MaxWeight = Weight;
          MaxDestInst = Succ->getFirstNonPHIOrDbgOrLifetime();
        }
      }
    }

    uint64_t TempWeight;
    // Only set weights if there is at least one non-zero weight.
    // In any other case, let the analyzer set weights.
    // Do not set weights if the weights are present. In ThinLTO, the profile
    // annotation is done twice. If the first annotation already set the
    // weights, the second pass does not need to set it.
    if (MaxWeight > 0 && !TI->extractProfTotalWeight(TempWeight)) {
      LLVM_DEBUG(dbgs() << "SUCCESS. Found non-zero weights.\n");
      TI->setMetadata(LLVMContext::MD_prof,
                      MDB.createBranchWeights(Weights));
      ORE->emit([&]() {
        return OptimizationRemark(DEBUG_TYPE, "PopularDest", MaxDestInst)
               << "most popular destination for conditional branches at "
               << ore::NV("CondBranchesLoc", BranchLoc);
      });
    } else {
      LLVM_DEBUG(dbgs() << "SKIPPED. All branch weights are zero.\n");
    }
  }
}

/// Get the line number for the function header.
///
/// This looks up function \p F in the current compilation unit and
/// retrieves the line number where the function is defined. This is
/// line 0 for all the samples read from the profile file. Every line
/// number is relative to this line.
///
/// \param F  Function object to query.
///
/// \returns the line number where \p F is defined. If it returns 0,
///          it means that there is no debug information available for \p F.
unsigned SampleProfileLoader::getFunctionLoc(Function &F) {
  if (DISubprogram *S = F.getSubprogram())
    return S->getLine();

  if (NoWarnSampleUnused)
    return 0;

  // If the start of \p F is missing, emit a diagnostic to inform the user
  // about the missed opportunity.
  F.getContext().diagnose(DiagnosticInfoSampleProfile(
      "No debug information found in function " + F.getName() +
          ": Function profile not used",
      DS_Warning));
  return 0;
}

void SampleProfileLoader::computeDominanceAndLoopInfo(Function &F) {
  DT.reset(new DominatorTree);
  DT->recalculate(F);

  PDT.reset(new PostDominatorTree(F));

  LI.reset(new LoopInfo);
  LI->analyze(*DT);
}

/// Generate branch weight metadata for all branches in \p F.
///
/// Branch weights are computed out of instruction samples using a
/// propagation heuristic. Propagation proceeds in 3 phases:
///
/// 1- Assignment of block weights. All the basic blocks in the function
///    are initial assigned the same weight as their most frequently
///    executed instruction.
///
/// 2- Creation of equivalence classes. Since samples may be missing from
///    blocks, we can fill in the gaps by setting the weights of all the
///    blocks in the same equivalence class to the same weight. To compute
///    the concept of equivalence, we use dominance and loop information.
///    Two blocks B1 and B2 are in the same equivalence class if B1
///    dominates B2, B2 post-dominates B1 and both are in the same loop.
///
/// 3- Propagation of block weights into edges. This uses a simple
///    propagation heuristic. The following rules are applied to every
///    block BB in the CFG:
///
///    - If BB has a single predecessor/successor, then the weight
///      of that edge is the weight of the block.
///
///    - If all the edges are known except one, and the weight of the
///      block is already known, the weight of the unknown edge will
///      be the weight of the block minus the sum of all the known
///      edges. If the sum of all the known edges is larger than BB's weight,
///      we set the unknown edge weight to zero.
///
///    - If there is a self-referential edge, and the weight of the block is
///      known, the weight for that edge is set to the weight of the block
///      minus the weight of the other incoming edges to that block (if
///      known).
///
/// Since this propagation is not guaranteed to finalize for every CFG, we
/// only allow it to proceed for a limited number of iterations (controlled
/// by -sample-profile-max-propagate-iterations).
///
/// FIXME: Try to replace this propagation heuristic with a scheme
/// that is guaranteed to finalize. A work-list approach similar to
/// the standard value propagation algorithm used by SSA-CCP might
/// work here.
///
/// Once all the branch weights are computed, we emit the MD_prof
/// metadata on BB using the computed values for each of its branches.
///
/// \param F The function to query.
///
/// \returns true if \p F was modified. Returns false, otherwise.
bool SampleProfileLoader::emitAnnotations(Function &F) {
  bool Changed = false;

  if (getFunctionLoc(F) == 0)
    return false;

  LLVM_DEBUG(dbgs() << "Line number for the first instruction in "
                    << F.getName() << ": " << getFunctionLoc(F) << "\n");

  DenseSet<GlobalValue::GUID> InlinedGUIDs;
  Changed |= inlineHotFunctions(F, InlinedGUIDs);

  // Compute basic block weights.
  Changed |= computeBlockWeights(F);

  if (Changed) {
    // Add an entry count to the function using the samples gathered at the
    // function entry.
    // Sets the GUIDs that are inlined in the profiled binary. This is used
    // for ThinLink to make correct liveness analysis, and also make the IR
    // match the profiled binary before annotation.
    F.setEntryCount(
        ProfileCount(Samples->getHeadSamples() + 1, Function::PCT_Real),
        &InlinedGUIDs);

    // Compute dominance and loop info needed for propagation.
    computeDominanceAndLoopInfo(F);

    // Find equivalence classes.
    findEquivalenceClasses(F);

    // Propagate weights to all edges.
    propagateWeights(F);
  }

  // If coverage checking was requested, compute it now.
  if (SampleProfileRecordCoverage) {
    unsigned Used = CoverageTracker.countUsedRecords(Samples, PSI);
    unsigned Total = CoverageTracker.countBodyRecords(Samples, PSI);
    unsigned Coverage = CoverageTracker.computeCoverage(Used, Total);
    if (Coverage < SampleProfileRecordCoverage) {
      F.getContext().diagnose(DiagnosticInfoSampleProfile(
          F.getSubprogram()->getFilename(), getFunctionLoc(F),
          Twine(Used) + " of " + Twine(Total) + " available profile records (" +
              Twine(Coverage) + "%) were applied",
          DS_Warning));
    }
  }

  if (SampleProfileSampleCoverage) {
    uint64_t Used = CoverageTracker.getTotalUsedSamples();
    uint64_t Total = CoverageTracker.countBodySamples(Samples, PSI);
    unsigned Coverage = CoverageTracker.computeCoverage(Used, Total);
    if (Coverage < SampleProfileSampleCoverage) {
      F.getContext().diagnose(DiagnosticInfoSampleProfile(
          F.getSubprogram()->getFilename(), getFunctionLoc(F),
          Twine(Used) + " of " + Twine(Total) + " available profile samples (" +
              Twine(Coverage) + "%) were applied",
          DS_Warning));
    }
  }
  return Changed;
}

char SampleProfileLoaderLegacyPass::ID = 0;

INITIALIZE_PASS_BEGIN(SampleProfileLoaderLegacyPass, "sample-profile",
                      "Sample Profile loader", false, false)
INITIALIZE_PASS_DEPENDENCY(AssumptionCacheTracker)
INITIALIZE_PASS_DEPENDENCY(TargetTransformInfoWrapperPass)
INITIALIZE_PASS_DEPENDENCY(ProfileSummaryInfoWrapperPass)
INITIALIZE_PASS_END(SampleProfileLoaderLegacyPass, "sample-profile",
                    "Sample Profile loader", false, false)

bool SampleProfileLoader::doInitialization(Module &M) {
  auto &Ctx = M.getContext();
  auto ReaderOrErr = SampleProfileReader::create(Filename, Ctx);
  if (std::error_code EC = ReaderOrErr.getError()) {
    std::string Msg = "Could not open profile: " + EC.message();
    Ctx.diagnose(DiagnosticInfoSampleProfile(Filename, Msg));
    return false;
  }
  Reader = std::move(ReaderOrErr.get());
  Reader->collectFuncsToUse(M);
  ProfileIsValid = (Reader->read() == sampleprof_error::success);

  if (!RemappingFilename.empty()) {
    // Apply profile remappings to the loaded profile data if requested.
    // For now, we only support remapping symbols encoded using the Itanium
    // C++ ABI's name mangling scheme.
    ReaderOrErr = SampleProfileReaderItaniumRemapper::create(
        RemappingFilename, Ctx, std::move(Reader));
    if (std::error_code EC = ReaderOrErr.getError()) {
      std::string Msg = "Could not open profile remapping file: " + EC.message();
      Ctx.diagnose(DiagnosticInfoSampleProfile(Filename, Msg));
      return false;
    }
    Reader = std::move(ReaderOrErr.get());
    ProfileIsValid = (Reader->read() == sampleprof_error::success);
  }
  return true;
}

ModulePass *llvm::createSampleProfileLoaderPass() {
  return new SampleProfileLoaderLegacyPass();
}

ModulePass *llvm::createSampleProfileLoaderPass(StringRef Name) {
  return new SampleProfileLoaderLegacyPass(Name);
}

bool SampleProfileLoader::runOnModule(Module &M, ModuleAnalysisManager *AM,
                                      ProfileSummaryInfo *_PSI) {
  FunctionSamples::GUIDToFuncNameMapper Mapper(M);
  if (!ProfileIsValid)
    return false;

  PSI = _PSI;
  if (M.getProfileSummary() == nullptr)
    M.setProfileSummary(Reader->getSummary().getMD(M.getContext()));

  // Compute the total number of samples collected in this profile.
  for (const auto &I : Reader->getProfiles())
    TotalCollectedSamples += I.second.getTotalSamples();

  // Populate the symbol map.
  for (const auto &N_F : M.getValueSymbolTable()) {
    StringRef OrigName = N_F.getKey();
    Function *F = dyn_cast<Function>(N_F.getValue());
    if (F == nullptr)
      continue;
    SymbolMap[OrigName] = F;
    auto pos = OrigName.find('.');
    if (pos != StringRef::npos) {
      StringRef NewName = OrigName.substr(0, pos);
      auto r = SymbolMap.insert(std::make_pair(NewName, F));
      // Failiing to insert means there is already an entry in SymbolMap,
      // thus there are multiple functions that are mapped to the same
      // stripped name. In this case of name conflicting, set the value
      // to nullptr to avoid confusion.
      if (!r.second)
        r.first->second = nullptr;
    }
  }

  bool retval = false;
  for (auto &F : M)
    if (!F.isDeclaration()) {
      clearFunctionData();
      retval |= runOnFunction(F, AM);
    }

  // Account for cold calls not inlined....
  for (const std::pair<Function *, NotInlinedProfileInfo> &pair :
       notInlinedCallInfo)
    updateProfileCallee(pair.first, pair.second.entryCount);

  return retval;
}

bool SampleProfileLoaderLegacyPass::runOnModule(Module &M) {
  ACT = &getAnalysis<AssumptionCacheTracker>();
  TTIWP = &getAnalysis<TargetTransformInfoWrapperPass>();
  ProfileSummaryInfo *PSI =
      &getAnalysis<ProfileSummaryInfoWrapperPass>().getPSI();
  return SampleLoader.runOnModule(M, nullptr, PSI);
}

bool SampleProfileLoader::runOnFunction(Function &F, ModuleAnalysisManager *AM) {
  
  DILocation2SampleMap.clear();
  // By default the entry count is initialized to -1, which will be treated
  // conservatively by getEntryCount as the same as unknown (None). This is
  // to avoid newly added code to be treated as cold. If we have samples
  // this will be overwritten in emitAnnotations.
  // If ProfileSampleAccurate is true or F has profile-sample-accurate
  // attribute, initialize the entry count to 0 so callsites or functions
  // unsampled will be treated as cold.
  uint64_t initialEntryCount =
      (ProfileSampleAccurate || F.hasFnAttribute("profile-sample-accurate"))
          ? 0
          : -1;
  F.setEntryCount(ProfileCount(initialEntryCount, Function::PCT_Real));
  std::unique_ptr<OptimizationRemarkEmitter> OwnedORE;
  if (AM) {
    auto &FAM =
        AM->getResult<FunctionAnalysisManagerModuleProxy>(*F.getParent())
            .getManager();
    ORE = &FAM.getResult<OptimizationRemarkEmitterAnalysis>(F);
  } else {
    OwnedORE = make_unique<OptimizationRemarkEmitter>(&F);
    ORE = OwnedORE.get();
  }
  Samples = Reader->getSamplesFor(F);
  if (Samples && !Samples->empty())
    return emitAnnotations(F);
  return false;
}

PreservedAnalyses SampleProfileLoaderPass::run(Module &M,
                                               ModuleAnalysisManager &AM) {
  FunctionAnalysisManager &FAM =
      AM.getResult<FunctionAnalysisManagerModuleProxy>(M).getManager();

  auto GetAssumptionCache = [&](Function &F) -> AssumptionCache & {
    return FAM.getResult<AssumptionAnalysis>(F);
  };
  auto GetTTI = [&](Function &F) -> TargetTransformInfo & {
    return FAM.getResult<TargetIRAnalysis>(F);
  };

  SampleProfileLoader SampleLoader(
      ProfileFileName.empty() ? SampleProfileFile : ProfileFileName,
      ProfileRemappingFileName.empty() ? SampleProfileRemappingFile
                                       : ProfileRemappingFileName,
      IsThinLTOPreLink, GetAssumptionCache, GetTTI);

  SampleLoader.doInitialization(M);

  ProfileSummaryInfo *PSI = &AM.getResult<ProfileSummaryAnalysis>(M);
  if (!SampleLoader.runOnModule(M, &AM, PSI))
    return PreservedAnalyses::all();

  return PreservedAnalyses::none();
}