//===--- BackendUtil.cpp - LLVM Backend Utilities -------------------------===//
//
// 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
//
//===----------------------------------------------------------------------===//

#include "clang/CodeGen/BackendUtil.h"
#include "BackendConsumer.h"
#include "LinkInModulesPass.h"
#include "clang/Basic/CodeGenOptions.h"
#include "clang/Basic/Diagnostic.h"
#include "clang/Basic/DiagnosticFrontend.h"
#include "clang/Basic/LangOptions.h"
#include "clang/Basic/TargetOptions.h"
#include "clang/Frontend/Utils.h"
#include "clang/Lex/HeaderSearchOptions.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/ADT/StringSwitch.h"
#include "llvm/Analysis/GlobalsModRef.h"
#include "llvm/Analysis/RuntimeLibcallInfo.h"
#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/BinaryFormat/ELF.h"
#include "llvm/Bitcode/BitcodeReader.h"
#include "llvm/Bitcode/BitcodeWriter.h"
#include "llvm/Bitcode/BitcodeWriterPass.h"
#include "llvm/CodeGen/MachineModuleInfo.h"
#include "llvm/CodeGen/TargetSubtargetInfo.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/Frontend/Driver/CodeGenOptions.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DebugInfo.h"
#include "llvm/IR/LLVMRemarkStreamer.h"
#include "llvm/IR/LegacyPassManager.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/ModuleSummaryIndex.h"
#include "llvm/IR/PassManager.h"
#include "llvm/IR/Verifier.h"
#include "llvm/IRPrinter/IRPrintingPasses.h"
#include "llvm/LTO/LTOBackend.h"
#include "llvm/MC/MCTargetOptions.h"
#include "llvm/MC/TargetRegistry.h"
#include "llvm/Object/OffloadBinary.h"
#include "llvm/Passes/PassBuilder.h"
#include "llvm/Passes/RunCodeGen.h"
#include "llvm/Passes/StandardInstrumentations.h"
#include "llvm/Plugins/PassPlugin.h"
#include "llvm/ProfileData/InstrProfCorrelator.h"
#include "llvm/Support/BuryPointer.h"
#include "llvm/Support/CodeGen.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Compiler.h"
#include "llvm/Support/IOSandbox.h"
#include "llvm/Support/MemoryBuffer.h"
#include "llvm/Support/PrettyStackTrace.h"
#include "llvm/Support/Program.h"
#include "llvm/Support/TimeProfiler.h"
#include "llvm/Support/Timer.h"
#include "llvm/Support/ToolOutputFile.h"
#include "llvm/Support/VirtualFileSystem.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetOptions.h"
#include "llvm/TargetParser/SubtargetFeature.h"
#include "llvm/TargetParser/Triple.h"
#include "llvm/Transforms/HipStdPar/HipStdPar.h"
#include "llvm/Transforms/IPO/EmbedBitcodePass.h"
#include "llvm/Transforms/IPO/InferFunctionAttrs.h"
#include "llvm/Transforms/IPO/LowerTypeTests.h"
#include "llvm/Transforms/IPO/ThinLTOBitcodeWriter.h"
#include "llvm/Transforms/InstCombine/InstCombine.h"
#include "llvm/Transforms/Instrumentation/AddressSanitizer.h"
#include "llvm/Transforms/Instrumentation/AddressSanitizerOptions.h"
#include "llvm/Transforms/Instrumentation/BoundsChecking.h"
#include "llvm/Transforms/Instrumentation/CopyProf.h"
#include "llvm/Transforms/Instrumentation/DataFlowSanitizer.h"
#include "llvm/Transforms/Instrumentation/GCOVProfiler.h"
#include "llvm/Transforms/Instrumentation/HWAddressSanitizer.h"
#include "llvm/Transforms/Instrumentation/InstrProfiling.h"
#include "llvm/Transforms/Instrumentation/KCFI.h"
#include "llvm/Transforms/Instrumentation/LowerAllowCheckPass.h"
#include "llvm/Transforms/Instrumentation/MemProfInstrumentation.h"
#include "llvm/Transforms/Instrumentation/MemProfUse.h"
#include "llvm/Transforms/Instrumentation/MemorySanitizer.h"
#include "llvm/Transforms/Instrumentation/NumericalStabilitySanitizer.h"
#include "llvm/Transforms/Instrumentation/PGOInstrumentation.h"
#include "llvm/Transforms/Instrumentation/RealtimeSanitizer.h"
#include "llvm/Transforms/Instrumentation/SanitizerBinaryMetadata.h"
#include "llvm/Transforms/Instrumentation/SanitizerCoverage.h"
#include "llvm/Transforms/Instrumentation/ThreadSanitizer.h"
#include "llvm/Transforms/Instrumentation/TypeSanitizer.h"
#include "llvm/Transforms/ObjCARC.h"
#include "llvm/Transforms/Scalar/EarlyCSE.h"
#include "llvm/Transforms/Scalar/GVN.h"
#include "llvm/Transforms/Scalar/JumpThreading.h"
#include "llvm/Transforms/Utils/AssignGUID.h"
#include "llvm/Transforms/Utils/Debugify.h"
#include "llvm/Transforms/Utils/DynamicDebugging.h"
#include "llvm/Transforms/Utils/ModuleUtils.h"
#include <limits>
#include <memory>
#include <optional>
using namespace clang;
using namespace llvm;

#define HANDLE_EXTENSION(Ext)                                                  \
  llvm::PassPluginLibraryInfo get##Ext##PluginInfo();
#include "llvm/Support/Extension.def"

namespace llvm {
// Experiment to move sanitizers earlier.
static cl::opt<bool> ClSanitizeOnOptimizerEarlyEP(
    "sanitizer-early-opt-ep",
    cl::desc("Insert sanitizers on OptimizerEarlyEP."));

// Experiment to mark cold functions as optsize/minsize/optnone.
// TODO: remove once this is exposed as a proper driver flag.
static cl::opt<PGOOptions::ColdFuncOpt> ClPGOColdFuncAttr(
    "pgo-cold-func-opt", cl::init(PGOOptions::ColdFuncOpt::Default), cl::Hidden,
    cl::desc(
        "Function attribute to apply to cold functions as determined by PGO"),
    cl::values(clEnumValN(PGOOptions::ColdFuncOpt::Default, "default",
                          "Default (no attribute)"),
               clEnumValN(PGOOptions::ColdFuncOpt::OptSize, "optsize",
                          "Mark cold functions with optsize."),
               clEnumValN(PGOOptions::ColdFuncOpt::MinSize, "minsize",
                          "Mark cold functions with minsize."),
               clEnumValN(PGOOptions::ColdFuncOpt::OptNone, "optnone",
                          "Mark cold functions with optnone.")));

LLVM_ABI extern cl::opt<InstrProfCorrelator::ProfCorrelatorKind>
    ProfileCorrelate;
} // namespace llvm
namespace clang {
extern llvm::cl::opt<bool> ClSanitizeGuardChecks;
}

// Path and name of file used for profile generation
static std::string getProfileGenName(const CodeGenOptions &CodeGenOpts) {
  std::string FileName = CodeGenOpts.InstrProfileOutput.empty()
                             ? llvm::driver::getDefaultProfileGenName()
                             : CodeGenOpts.InstrProfileOutput;
  if (CodeGenOpts.ContinuousProfileSync)
    FileName = "%c" + FileName;
  return FileName;
}

namespace {

class EmitAssemblyHelper {
  CompilerInstance &CI;
  DiagnosticsEngine &Diags;
  const CodeGenOptions &CodeGenOpts;
  const clang::TargetOptions &TargetOpts;
  const LangOptions &LangOpts;
  llvm::Module *TheModule;
  IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS;

  std::unique_ptr<raw_pwrite_stream> OS;

  Triple TargetTriple;

  TargetIRAnalysis getTargetIRAnalysis() const {
    if (TM)
      return TM->getTargetIRAnalysis();

    return TargetIRAnalysis();
  }

  /// Generates the TargetMachine.
  /// Leaves TM unchanged if it is unable to create the target machine.
  /// Some of our clang tests specify triples which are not built
  /// into clang. This is okay because these tests check the generated
  /// IR, and they require DataLayout which depends on the triple.
  /// In this case, we allow this method to fail and not report an error.
  /// When MustCreateTM is used, we print an error if we are unable to load
  /// the requested target.
  void CreateTargetMachine(bool MustCreateTM);

  std::unique_ptr<llvm::ToolOutputFile> openOutputFile(StringRef Path) {
    std::error_code EC;
    auto F = std::make_unique<llvm::ToolOutputFile>(Path, EC,
                                                     llvm::sys::fs::OF_None);
    if (EC) {
      Diags.Report(diag::err_fe_unable_to_open_output) << Path << EC.message();
      F.reset();
    }
    return F;
  }

  void RunOptimizationPipeline(
      BackendAction Action, std::unique_ptr<raw_pwrite_stream> &OS,
      std::unique_ptr<llvm::ToolOutputFile> &ThinLinkOS, BackendConsumer *BC);
  void RunCodegenPipeline(BackendAction Action,
                          std::unique_ptr<raw_pwrite_stream> &OS,
                          std::unique_ptr<llvm::ToolOutputFile> &DwoOS);
  void TimeCodegenPasses(llvm::function_ref<void()> RunPasses);

  /// Check whether we should emit a module summary for regular LTO.
  /// The module summary should be emitted by default for regular LTO
  /// except for ld64 targets.
  ///
  /// \return True if the module summary should be emitted.
  bool shouldEmitRegularLTOSummary() const {
    return CodeGenOpts.PrepareForLTO && !CodeGenOpts.DisableLLVMPasses &&
           TargetTriple.getVendor() != llvm::Triple::Apple;
  }

  /// Check whether we should emit a flag for UnifiedLTO.
  /// The UnifiedLTO module flag should be set when UnifiedLTO is enabled for
  /// ThinLTO or Full LTO with module summaries.
  bool shouldEmitUnifiedLTOModueFlag() const {
    return CodeGenOpts.UnifiedLTO &&
           (CodeGenOpts.PrepareForThinLTO || shouldEmitRegularLTOSummary());
  }

public:
  EmitAssemblyHelper(CompilerInstance &CI, CodeGenOptions &CGOpts,
                     llvm::Module *M,
                     IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS)
      : CI(CI), Diags(CI.getDiagnostics()), CodeGenOpts(CGOpts),
        TargetOpts(CI.getTargetOpts()), LangOpts(CI.getLangOpts()),
        TheModule(M), VFS(std::move(VFS)),
        TargetTriple(TheModule->getTargetTriple()) {}

  ~EmitAssemblyHelper() {
    if (CodeGenOpts.DisableFree)
      BuryPointer(std::move(TM));
  }

  std::unique_ptr<TargetMachine> TM;

  // Emit output using the new pass manager for the optimization pipeline.
  void emitAssembly(BackendAction Action, std::unique_ptr<raw_pwrite_stream> OS,
                    BackendConsumer *BC);
};
} // namespace

static SanitizerCoverageOptions
getSancovOptsFromCGOpts(const CodeGenOptions &CGOpts) {
  SanitizerCoverageOptions Opts;
  Opts.CoverageType =
      static_cast<SanitizerCoverageOptions::Type>(CGOpts.SanitizeCoverageType);
  Opts.IndirectCalls = CGOpts.SanitizeCoverageIndirectCalls;
  Opts.TraceBB = CGOpts.SanitizeCoverageTraceBB;
  Opts.TraceCmp = CGOpts.SanitizeCoverageTraceCmp;
  Opts.TraceDiv = CGOpts.SanitizeCoverageTraceDiv;
  Opts.TraceGep = CGOpts.SanitizeCoverageTraceGep;
  Opts.Use8bitCounters = CGOpts.SanitizeCoverage8bitCounters;
  Opts.TracePC = CGOpts.SanitizeCoverageTracePC;
  Opts.TracePCEntryExit = CGOpts.SanitizeCoverageTracePCEntryExit;
  Opts.TracePCGuard = CGOpts.SanitizeCoverageTracePCGuard;
  Opts.NoPrune = CGOpts.SanitizeCoverageNoPrune;
  Opts.Inline8bitCounters = CGOpts.SanitizeCoverageInline8bitCounters;
  Opts.InlineBoolFlag = CGOpts.SanitizeCoverageInlineBoolFlag;
  Opts.PCTable = CGOpts.SanitizeCoveragePCTable;
  Opts.StackDepth = CGOpts.SanitizeCoverageStackDepth;
  Opts.StackDepthCallbackMin = CGOpts.SanitizeCoverageStackDepthCallbackMin;
  Opts.TraceLoads = CGOpts.SanitizeCoverageTraceLoads;
  Opts.TraceStores = CGOpts.SanitizeCoverageTraceStores;
  Opts.CollectControlFlow = CGOpts.SanitizeCoverageControlFlow;
  return Opts;
}

static SanitizerBinaryMetadataOptions
getSanitizerBinaryMetadataOptions(const CodeGenOptions &CGOpts) {
  SanitizerBinaryMetadataOptions Opts;
  Opts.Covered = CGOpts.SanitizeBinaryMetadataCovered;
  Opts.Atomics = CGOpts.SanitizeBinaryMetadataAtomics;
  Opts.UAR = CGOpts.SanitizeBinaryMetadataUAR;
  return Opts;
}

// Check if ASan should use GC-friendly instrumentation for globals.
// First of all, there is no point if -fdata-sections is off (expect for MachO,
// where this is not a factor). Also, on ELF this feature requires an assembler
// extension that only works with -integrated-as at the moment.
static bool asanUseGlobalsGC(const Triple &T, const CodeGenOptions &CGOpts) {
  if (!CGOpts.SanitizeAddressGlobalsDeadStripping)
    return false;
  switch (T.getObjectFormat()) {
  case Triple::MachO:
  case Triple::COFF:
    return true;
  case Triple::ELF:
    return !CGOpts.DisableIntegratedAS;
  case Triple::GOFF:
    llvm::report_fatal_error("ASan not implemented for GOFF");
  case Triple::XCOFF:
    llvm::report_fatal_error("ASan not implemented for XCOFF.");
  case Triple::Wasm:
  case Triple::DXContainer:
  case Triple::SPIRV:
  case Triple::UnknownObjectFormat:
    break;
  }
  return false;
}

static std::optional<llvm::CodeModel::Model>
getCodeModel(const CodeGenOptions &CodeGenOpts) {
  unsigned CodeModel = llvm::StringSwitch<unsigned>(CodeGenOpts.CodeModel)
                           .Case("tiny", llvm::CodeModel::Tiny)
                           .Case("small", llvm::CodeModel::Small)
                           .Case("kernel", llvm::CodeModel::Kernel)
                           .Case("medium", llvm::CodeModel::Medium)
                           .Case("large", llvm::CodeModel::Large)
                           .Cases({"default", ""}, ~1u)
                           .Default(~0u);
  assert(CodeModel != ~0u && "invalid code model!");
  if (CodeModel == ~1u)
    return std::nullopt;
  return static_cast<llvm::CodeModel::Model>(CodeModel);
}

static CodeGenFileType getCodeGenFileType(BackendAction Action) {
  if (Action == Backend_EmitObj)
    return CodeGenFileType::ObjectFile;
  else if (Action == Backend_EmitMCNull)
    return CodeGenFileType::Null;
  else {
    assert(Action == Backend_EmitAssembly && "Invalid action!");
    return CodeGenFileType::AssemblyFile;
  }
}

static bool actionRequiresCodeGen(BackendAction Action) {
  return Action != Backend_EmitNothing && Action != Backend_EmitBC &&
         Action != Backend_EmitLL;
}

static std::string flattenClangCommandLine(ArrayRef<std::string> Args,
                                           StringRef MainFilename,
                                           ArrayRef<StringRef> InputFiles) {
  if (Args.empty())
    return std::string{};

  std::string FlatCmdLine;
  raw_string_ostream OS(FlatCmdLine);
  bool PrintedOneArg = false;
  if (!StringRef(Args[0]).contains("-cc1")) {
    llvm::sys::printArg(OS, "-cc1", /*Quote=*/true);
    PrintedOneArg = true;
  }
  for (unsigned i = 0; i < Args.size(); i++) {
    StringRef Arg = Args[i];
    if (Arg.empty())
      continue;
    if (Arg == "-main-file-name" || Arg == "-o") {
      i++; // Skip this argument and next one.
      continue;
    }
    if (Arg.starts_with("-object-file-name"))
      continue;
    // Strip the source positional, matching either MainFilename (the
    // -main-file-name basename) or one of the resolved frontend input paths
    // (which is what the cc1 positional looks like for an absolute driver
    // input). Avoid a generic basename match: it would also strip values of
    // args like `-include <path>` whose trailing component happens to equal
    // the source basename.
    if (Arg == MainFilename || llvm::is_contained(InputFiles, Arg))
      continue;
    // Skip fmessage-length for reproducibility.
    if (Arg.starts_with("-fmessage-length"))
      continue;
    if (PrintedOneArg)
      OS << " ";
    llvm::sys::printArg(OS, Arg, /*Quote=*/true);
    PrintedOneArg = true;
  }
  return FlatCmdLine;
}

static bool initTargetOptions(const CompilerInstance &CI,
                              DiagnosticsEngine &Diags,
                              llvm::TargetOptions &Options) {
  const auto &CodeGenOpts = CI.getCodeGenOpts();
  const auto &TargetOpts = CI.getTargetOpts();
  const auto &LangOpts = CI.getLangOpts();
  const auto &HSOpts = CI.getHeaderSearchOpts();

  Options.MCOptions.BinutilsVersion =
      llvm::MCTargetOptions::parseBinutilsVersion(CodeGenOpts.BinutilsVersion);
  Options.UseInitArray = CodeGenOpts.UseInitArray;
  Options.MCOptions.DisableIntegratedAS = CodeGenOpts.DisableIntegratedAS;

  if (CodeGenOpts.hasSjLjExceptions())
    Options.ExceptionModel = llvm::ExceptionHandling::SjLj;
  if (CodeGenOpts.hasSEHExceptions())
    Options.ExceptionModel = llvm::ExceptionHandling::WinEH;
  if (CodeGenOpts.hasDWARFExceptions())
    Options.ExceptionModel = llvm::ExceptionHandling::DwarfCFI;
  if (CodeGenOpts.hasWasmExceptions())
    Options.ExceptionModel = llvm::ExceptionHandling::Wasm;
  if (CodeGenOpts.hasEmscriptenExceptions())
    Options.ExceptionModel = llvm::ExceptionHandling::Emscripten;

  Options.NoZerosInBSS = CodeGenOpts.NoZeroInitializedInBSS;

  Options.BBAddrMap = CodeGenOpts.BBAddrMap;
  Options.BBSections =
      llvm::StringSwitch<llvm::BasicBlockSection>(CodeGenOpts.BBSections)
          .Case("all", llvm::BasicBlockSection::All)
          .StartsWith("list=", llvm::BasicBlockSection::List)
          .Case("none", llvm::BasicBlockSection::None)
          .Default(llvm::BasicBlockSection::None);

  if (Options.BBSections == llvm::BasicBlockSection::List) {
    ErrorOr<std::unique_ptr<MemoryBuffer>> MBOrErr =
        CI.getVirtualFileSystem().getBufferForFile(
            CodeGenOpts.BBSections.substr(5));
    if (!MBOrErr) {
      Diags.Report(diag::err_fe_unable_to_load_basic_block_sections_file)
          << MBOrErr.getError().message();
      return false;
    }
    Options.BBSectionsFuncListBuf = std::move(*MBOrErr);
  }

  Options.EnableMachineFunctionSplitter = CodeGenOpts.SplitMachineFunctions;
  Options.EnableStaticDataPartitioning =
      CodeGenOpts.PartitionStaticDataSections;
  Options.FunctionSections = CodeGenOpts.FunctionSections;
  Options.DataSections = CodeGenOpts.DataSections;
  Options.IgnoreXCOFFVisibility = LangOpts.IgnoreXCOFFVisibility;
  Options.UniqueSectionNames = CodeGenOpts.UniqueSectionNames;
  Options.UniqueBasicBlockSectionNames =
      CodeGenOpts.UniqueBasicBlockSectionNames;
  Options.SeparateNamedSections = CodeGenOpts.SeparateNamedSections;
  Options.TLSSize = CodeGenOpts.TLSSize;
  Options.EnableTLSDESC = CodeGenOpts.EnableTLSDESC;
  Options.EmulatedTLS = CodeGenOpts.EmulatedTLS;
  Options.DebuggerTuning = CodeGenOpts.getDebuggerTuning();
  Options.EmitStackSizeSection = CodeGenOpts.StackSizeSection;
  Options.StackUsageFile = CodeGenOpts.StackUsageFile;
  Options.EmitAddrsig = CodeGenOpts.Addrsig;
  Options.ForceDwarfFrameSection = CodeGenOpts.ForceDwarfFrameSection;
  Options.EmitCallGraphSection = CodeGenOpts.CallGraphSection;
  Options.EmitCallSiteInfo = CodeGenOpts.EmitCallSiteInfo;
  Options.XRayFunctionIndex = CodeGenOpts.XRayFunctionIndex;
  Options.LoopAlignment = CodeGenOpts.LoopAlignment;
  Options.DebugStrictDwarf = CodeGenOpts.DebugStrictDwarf;
  Options.ObjectFilenameForDebug =
      CodeGenOpts.remapDebugPathPrefix(CodeGenOpts.ObjectFilenameForDebug);
  Options.Hotpatch = CodeGenOpts.HotPatch;
  Options.JMCInstrument = CodeGenOpts.JMCInstrument;
  Options.XCOFFReadOnlyPointers = CodeGenOpts.XCOFFReadOnlyPointers;
  Options.VecLib =
      convertDriverVectorLibraryToVectorLibrary(CodeGenOpts.getVecLib());

  switch (CodeGenOpts.getSwiftAsyncFramePointer()) {
  case CodeGenOptions::SwiftAsyncFramePointerKind::Auto:
    Options.SwiftAsyncFramePointer =
        SwiftAsyncFramePointerMode::DeploymentBased;
    break;

  case CodeGenOptions::SwiftAsyncFramePointerKind::Always:
    Options.SwiftAsyncFramePointer = SwiftAsyncFramePointerMode::Always;
    break;

  case CodeGenOptions::SwiftAsyncFramePointerKind::Never:
    Options.SwiftAsyncFramePointer = SwiftAsyncFramePointerMode::Never;
    break;
  }

  Options.MCOptions.SplitDwarfFile = CodeGenOpts.SplitDwarfFile;
  Options.MCOptions.EmitDwarfUnwind = CodeGenOpts.getEmitDwarfUnwind();
  Options.MCOptions.EmitCompactUnwindNonCanonical =
      CodeGenOpts.EmitCompactUnwindNonCanonical;
  Options.MCOptions.EmitSFrameUnwind = CodeGenOpts.EmitSFrameUnwind;
  Options.MCOptions.MCRelaxAll = CodeGenOpts.RelaxAll;
  Options.MCOptions.MCSaveTempLabels = CodeGenOpts.SaveTempLabels;
  Options.MCOptions.MCUseDwarfDirectory =
      CodeGenOpts.NoDwarfDirectoryAsm
          ? llvm::MCTargetOptions::DisableDwarfDirectory
          : llvm::MCTargetOptions::EnableDwarfDirectory;
  Options.MCOptions.MCNoExecStack = CodeGenOpts.NoExecStack;
  Options.MCOptions.MCIncrementalLinkerCompatible =
      CodeGenOpts.IncrementalLinkerCompatible;
  Options.MCOptions.MCFatalWarnings = CodeGenOpts.FatalWarnings;
  Options.MCOptions.MCNoWarn = CodeGenOpts.NoWarn;
  Options.MCOptions.AsmVerbose = CodeGenOpts.AsmVerbose;
  Options.MCOptions.Dwarf64 = CodeGenOpts.Dwarf64;
  Options.MCOptions.PreserveAsmComments = CodeGenOpts.PreserveAsmComments;
  Options.MCOptions.Crel = CodeGenOpts.Crel;
  Options.MCOptions.RelocSectionSym = CodeGenOpts.getRelocSectionSym();
  Options.MCOptions.ImplicitMapSyms = CodeGenOpts.ImplicitMapSyms;
  Options.MCOptions.X86RelaxRelocations = CodeGenOpts.X86RelaxRelocations;
  Options.MCOptions.CompressDebugSections =
      CodeGenOpts.getCompressDebugSections();
  if (CodeGenOpts.OutputAsmVariant != 3) // 3 (default): not specified
    Options.MCOptions.OutputAsmVariant = CodeGenOpts.OutputAsmVariant;
  Options.MCOptions.ABIName = TargetOpts.ABI;
  for (const auto &Entry : HSOpts.UserEntries)
    if (!Entry.IsFramework &&
        (Entry.Group == frontend::IncludeDirGroup::Quoted ||
         Entry.Group == frontend::IncludeDirGroup::Angled ||
         Entry.Group == frontend::IncludeDirGroup::System))
      Options.MCOptions.IASSearchPaths.push_back(
          Entry.IgnoreSysRoot ? Entry.Path : HSOpts.Sysroot + Entry.Path);
  Options.MCOptions.Argv0 = CodeGenOpts.Argv0 ? CodeGenOpts.Argv0 : "";
  // Pass the resolved frontend inputs so flattenClangCommandLine can strip
  // the cc1 source positional even when the driver received an absolute path
  // (which won't match CodeGenOpts.MainFileName, that's just the basename).
  SmallVector<StringRef, 1> InputFiles;
  for (const auto &Input : CI.getFrontendOpts().Inputs)
    if (Input.isFile())
      InputFiles.push_back(Input.getFile());
  Options.MCOptions.CommandlineArgs = flattenClangCommandLine(
      CodeGenOpts.CommandLineArgs, CodeGenOpts.MainFileName, InputFiles);
  Options.MCOptions.AsSecureLogFile = CodeGenOpts.AsSecureLogFile;
  Options.MCOptions.PPCUseFullRegisterNames =
      CodeGenOpts.PPCUseFullRegisterNames;
  Options.MisExpect = CodeGenOpts.MisExpect;

  return true;
}

static std::optional<GCOVOptions>
getGCOVOptions(const CodeGenOptions &CodeGenOpts, const LangOptions &LangOpts) {
  if (CodeGenOpts.CoverageNotesFile.empty() &&
      CodeGenOpts.CoverageDataFile.empty())
    return std::nullopt;
  // Not using 'GCOVOptions::getDefault' allows us to avoid exiting if
  // LLVM's -default-gcov-version flag is set to something invalid.
  GCOVOptions Options;
  Options.EmitNotes = !CodeGenOpts.CoverageNotesFile.empty();
  Options.EmitData = !CodeGenOpts.CoverageDataFile.empty();
  llvm::copy(CodeGenOpts.CoverageVersion, std::begin(Options.Version));
  Options.NoRedZone = CodeGenOpts.DisableRedZone;
  Options.Filter = CodeGenOpts.ProfileFilterFiles;
  Options.Exclude = CodeGenOpts.ProfileExcludeFiles;
  Options.Atomic = CodeGenOpts.AtomicProfileUpdate;
  return Options;
}

static std::optional<InstrProfOptions>
getInstrProfOptions(const CodeGenOptions &CodeGenOpts,
                    const LangOptions &LangOpts) {
  if (!CodeGenOpts.hasProfileClangInstr())
    return std::nullopt;
  InstrProfOptions Options;
  Options.NoRedZone = CodeGenOpts.DisableRedZone;
  Options.InstrProfileOutput = CodeGenOpts.ContinuousProfileSync
                                   ? ("%c" + CodeGenOpts.InstrProfileOutput)
                                   : CodeGenOpts.InstrProfileOutput;
  Options.Atomic = CodeGenOpts.AtomicProfileUpdate;
  return Options;
}

static void setCommandLineOpts(const CodeGenOptions &CodeGenOpts,
                               vfs::FileSystem &VFS) {
  SmallVector<const char *, 16> BackendArgs;
  BackendArgs.push_back("clang"); // Fake program name.
  if (!CodeGenOpts.DebugPass.empty()) {
    BackendArgs.push_back("-debug-pass");
    BackendArgs.push_back(CodeGenOpts.DebugPass.c_str());
  }
  if (!CodeGenOpts.LimitFloatPrecision.empty()) {
    BackendArgs.push_back("-limit-float-precision");
    BackendArgs.push_back(CodeGenOpts.LimitFloatPrecision.c_str());
  }

  // Check for the default "clang" invocation that won't set any cl::opt values.
  // Skip trying to parse the command line invocation to avoid the issues
  // described below.
  if (BackendArgs.size() == 1)
    return;
  BackendArgs.push_back(nullptr);
  // FIXME: The command line parser below is not thread-safe and shares a global
  // state, so this call might crash or overwrite the options of another Clang
  // instance in the same process.
  llvm::cl::ParseCommandLineOptions(BackendArgs.size() - 1, BackendArgs.data(),
                                    /*Overview=*/"", /*Errs=*/nullptr,
                                    /*VFS=*/&VFS);
}

void EmitAssemblyHelper::CreateTargetMachine(bool MustCreateTM) {
  // Create the TargetMachine for generating code.
  std::string Error;
  const llvm::Triple &Triple = TheModule->getTargetTriple();
  const llvm::Target *TheTarget = TargetRegistry::lookupTarget(Triple, Error);
  if (!TheTarget) {
    if (MustCreateTM)
      Diags.Report(diag::err_fe_unable_to_create_target) << Error;
    return;
  }

  std::optional<llvm::CodeModel::Model> CM = getCodeModel(CodeGenOpts);
  std::string FeaturesStr =
      llvm::join(TargetOpts.Features.begin(), TargetOpts.Features.end(), ",");
  llvm::Reloc::Model RM = CodeGenOpts.RelocationModel;
  std::optional<CodeGenOptLevel> OptLevelOrNone =
      CodeGenOpt::getLevel(CodeGenOpts.OptimizationLevel);
  assert(OptLevelOrNone && "Invalid optimization level!");
  CodeGenOptLevel OptLevel = *OptLevelOrNone;

  llvm::TargetOptions Options;
  if (!initTargetOptions(CI, Diags, Options))
    return;
  TM.reset(TheTarget->createTargetMachine(Triple, TargetOpts.CPU, FeaturesStr,
                                          Options, RM, CM, OptLevel));
  if (TM)
    TM->setLargeDataThreshold(CodeGenOpts.LargeDataThreshold);
}

static OptimizationLevel mapToLevel(const CodeGenOptions &Opts) {
  switch (Opts.OptimizationLevel) {
  default:
    llvm_unreachable("Invalid optimization level!");

  case 0:
    return OptimizationLevel::O0;

  case 1:
    return OptimizationLevel::O1;

  case 2:
    return OptimizationLevel::O2;

  case 3:
    return OptimizationLevel::O3;
  }
}

static void addKCFIPass(const Triple &TargetTriple, const LangOptions &LangOpts,
                        PassBuilder &PB) {
  // If the back-end supports KCFI operand bundle lowering, skip KCFIPass.
  if (TargetTriple.getArch() == llvm::Triple::x86_64 ||
      TargetTriple.isAArch64(64) || TargetTriple.isRISCV() ||
      TargetTriple.isARM() || TargetTriple.isThumb() ||
      TargetTriple.getArch() == llvm::Triple::hexagon)
    return;

  // Ensure we lower KCFI operand bundles with -O0.
  PB.registerOptimizerLastEPCallback(
      [&](ModulePassManager &MPM, OptimizationLevel Level, ThinOrFullLTOPhase) {
        if (Level == OptimizationLevel::O0 &&
            LangOpts.Sanitize.has(SanitizerKind::KCFI))
          MPM.addPass(createModuleToFunctionPassAdaptor(KCFIPass()));
      });

  // When optimizations are requested, run KCIFPass after InstCombine to
  // avoid unnecessary checks.
  PB.registerPeepholeEPCallback(
      [&](FunctionPassManager &FPM, OptimizationLevel Level) {
        if (Level != OptimizationLevel::O0 &&
            LangOpts.Sanitize.has(SanitizerKind::KCFI))
          FPM.addPass(KCFIPass());
      });
}

static void addSanitizers(const Triple &TargetTriple,
                          const CodeGenOptions &CodeGenOpts,
                          const LangOptions &LangOpts, PassBuilder &PB) {
  auto SanitizersCallback = [&](ModulePassManager &MPM, OptimizationLevel Level,
                                ThinOrFullLTOPhase) {
    if (CodeGenOpts.hasSanitizeCoverage()) {
      auto SancovOpts = getSancovOptsFromCGOpts(CodeGenOpts);
      MPM.addPass(
          SanitizerCoveragePass(SancovOpts, PB.getVirtualFileSystemPtr(),
                                CodeGenOpts.SanitizeCoverageAllowlistFiles,
                                CodeGenOpts.SanitizeCoverageIgnorelistFiles));
    }

    if (CodeGenOpts.hasSanitizeBinaryMetadata()) {
      MPM.addPass(SanitizerBinaryMetadataPass(
          getSanitizerBinaryMetadataOptions(CodeGenOpts),
          PB.getVirtualFileSystemPtr(),
          CodeGenOpts.SanitizeMetadataIgnorelistFiles));
    }

    auto MSanPass = [&](SanitizerMask Mask, bool CompileKernel) {
      if (LangOpts.Sanitize.has(Mask)) {
        int TrackOrigins = CodeGenOpts.SanitizeMemoryTrackOrigins;
        bool Recover = CodeGenOpts.SanitizeRecover.has(Mask);

        MemorySanitizerOptions options(TrackOrigins, Recover, CompileKernel,
                                       CodeGenOpts.SanitizeMemoryParamRetval);
        MPM.addPass(MemorySanitizerPass(options));
        if (Level != OptimizationLevel::O0) {
          // MemorySanitizer inserts complex instrumentation that mostly follows
          // the logic of the original code, but operates on "shadow" values. It
          // can benefit from re-running some general purpose optimization
          // passes.
          MPM.addPass(RequireAnalysisPass<GlobalsAA, llvm::Module>());
          FunctionPassManager FPM;
          FPM.addPass(EarlyCSEPass(true /* Enable mem-ssa. */));
          FPM.addPass(InstCombinePass());
          FPM.addPass(JumpThreadingPass());
          FPM.addPass(GVNPass());
          FPM.addPass(InstCombinePass());
          MPM.addPass(createModuleToFunctionPassAdaptor(std::move(FPM)));
        }
      }
    };
    MSanPass(SanitizerKind::Memory, false);
    MSanPass(SanitizerKind::KernelMemory, true);

    if (LangOpts.Sanitize.has(SanitizerKind::Thread)) {
      MPM.addPass(ModuleThreadSanitizerPass());
      MPM.addPass(createModuleToFunctionPassAdaptor(ThreadSanitizerPass()));
    }

    if (LangOpts.Sanitize.has(SanitizerKind::Type))
      MPM.addPass(TypeSanitizerPass());

    if (LangOpts.Sanitize.has(SanitizerKind::NumericalStability))
      MPM.addPass(NumericalStabilitySanitizerPass());

    if (LangOpts.Sanitize.has(SanitizerKind::Realtime))
      MPM.addPass(RealtimeSanitizerPass());

    auto ASanPass = [&](SanitizerMask Mask, bool CompileKernel) {
      if (LangOpts.Sanitize.has(Mask)) {
        bool UseGlobalGC = asanUseGlobalsGC(TargetTriple, CodeGenOpts);
        bool UseOdrIndicator = CodeGenOpts.SanitizeAddressUseOdrIndicator;
        llvm::AsanDtorKind DestructorKind =
            CodeGenOpts.getSanitizeAddressDtor();
        AddressSanitizerOptions Opts;
        Opts.CompileKernel = CompileKernel;
        Opts.Recover = CodeGenOpts.SanitizeRecover.has(Mask);
        Opts.UseAfterScope = CodeGenOpts.SanitizeAddressUseAfterScope;
        Opts.UseAfterReturn = CodeGenOpts.getSanitizeAddressUseAfterReturn();
        MPM.addPass(AddressSanitizerPass(Opts, UseGlobalGC, UseOdrIndicator,
                                         DestructorKind));
      }
    };
    ASanPass(SanitizerKind::Address, false);
    ASanPass(SanitizerKind::KernelAddress, true);

    auto HWASanPass = [&](SanitizerMask Mask, bool CompileKernel) {
      if (LangOpts.Sanitize.has(Mask)) {
        bool Recover = CodeGenOpts.SanitizeRecover.has(Mask);
        MPM.addPass(HWAddressSanitizerPass(
            {CompileKernel, Recover,
             /*DisableOptimization=*/CodeGenOpts.OptimizationLevel == 0}));
      }
    };
    HWASanPass(SanitizerKind::HWAddress, false);
    HWASanPass(SanitizerKind::KernelHWAddress, true);

    if (LangOpts.Sanitize.has(SanitizerKind::DataFlow)) {
      MPM.addPass(DataFlowSanitizerPass(LangOpts.NoSanitizeFiles,
                                        PB.getVirtualFileSystemPtr()));
    }
  };
  if (ClSanitizeOnOptimizerEarlyEP) {
    PB.registerOptimizerEarlyEPCallback(
        [SanitizersCallback](ModulePassManager &MPM, OptimizationLevel Level,
                             ThinOrFullLTOPhase Phase) {
          ModulePassManager NewMPM;
          SanitizersCallback(NewMPM, Level, Phase);
          if (!NewMPM.isEmpty()) {
            // Sanitizers can abandon<GlobalsAA>.
            NewMPM.addPass(RequireAnalysisPass<GlobalsAA, llvm::Module>());
            MPM.addPass(std::move(NewMPM));
          }
        });
  } else {
    // LastEP does not need GlobalsAA.
    PB.registerOptimizerLastEPCallback(SanitizersCallback);
  }
}

void addLowerAllowCheckPass(const CodeGenOptions &CodeGenOpts,
                            const LangOptions &LangOpts, PassBuilder &PB) {
  // SanitizeSkipHotCutoffs: doubles with range [0, 1]
  // Opts.cutoffs: unsigned ints with range [0, 1000000]
  auto ScaledCutoffs = CodeGenOpts.SanitizeSkipHotCutoffs.getAllScaled(1000000);
  uint64_t AllowRuntimeCheckSkipHotCutoff =
      CodeGenOpts.AllowRuntimeCheckSkipHotCutoff.value_or(0.0) * 1000000;
  // Only register the pass if one of the relevant sanitizers is enabled.
  // This avoids pipeline overhead for builds that do not use these sanitizers.
  bool LowerAllowSanitize = LangOpts.Sanitize.hasOneOf(
      SanitizerKind::Address | SanitizerKind::KernelAddress |
      SanitizerKind::Thread | SanitizerKind::Memory |
      SanitizerKind::KernelMemory | SanitizerKind::HWAddress |
      SanitizerKind::KernelHWAddress);

  // TODO: remove IsRequested()
  if (LowerAllowCheckPass::IsRequested() || ScaledCutoffs.has_value() ||
      CodeGenOpts.AllowRuntimeCheckSkipHotCutoff.has_value() ||
      LowerAllowSanitize) {
    // We want to call it after inline, which is about OptimizerEarlyEPCallback.
    PB.registerOptimizerEarlyEPCallback(
        [ScaledCutoffs, AllowRuntimeCheckSkipHotCutoff](
            ModulePassManager &MPM, OptimizationLevel Level,
            ThinOrFullLTOPhase Phase) {
          LowerAllowCheckPass::Options Opts;
          // TODO: after removing IsRequested(), make this unconditional
          if (ScaledCutoffs.has_value())
            Opts.cutoffs = ScaledCutoffs.value();
          Opts.runtime_check = AllowRuntimeCheckSkipHotCutoff;
          MPM.addPass(
              createModuleToFunctionPassAdaptor(LowerAllowCheckPass(Opts)));
        });
  }
}

void EmitAssemblyHelper::RunOptimizationPipeline(
    BackendAction Action, std::unique_ptr<raw_pwrite_stream> &OS,
    std::unique_ptr<llvm::ToolOutputFile> &ThinLinkOS, BackendConsumer *BC) {
  std::optional<PGOOptions> PGOOpt;

  if (CodeGenOpts.hasProfileIRInstr())
    // -fprofile-generate.
    PGOOpt = PGOOptions(getProfileGenName(CodeGenOpts), "", "",
                        CodeGenOpts.MemoryProfileUsePath, PGOOptions::IRInstr,
                        PGOOptions::NoCSAction, ClPGOColdFuncAttr,
                        CodeGenOpts.DebugInfoForProfiling,
                        /*PseudoProbeForProfiling=*/false,
                        CodeGenOpts.AtomicProfileUpdate);
  else if (CodeGenOpts.hasProfileIRUse()) {
    // -fprofile-use.
    auto CSAction = CodeGenOpts.hasProfileCSIRUse() ? PGOOptions::CSIRUse
                                                    : PGOOptions::NoCSAction;
    PGOOpt = PGOOptions(CodeGenOpts.ProfileInstrumentUsePath, "",
                        CodeGenOpts.ProfileRemappingFile,
                        CodeGenOpts.MemoryProfileUsePath, PGOOptions::IRUse,
                        CSAction, ClPGOColdFuncAttr,
                        CodeGenOpts.DebugInfoForProfiling);
  } else if (!CodeGenOpts.SampleProfileFile.empty())
    // -fprofile-sample-use
    PGOOpt = PGOOptions(
        CodeGenOpts.SampleProfileFile, "", CodeGenOpts.ProfileRemappingFile,
        CodeGenOpts.MemoryProfileUsePath, PGOOptions::SampleUse,
        PGOOptions::NoCSAction, ClPGOColdFuncAttr,
        CodeGenOpts.DebugInfoForProfiling, CodeGenOpts.PseudoProbeForProfiling);
  else if (!CodeGenOpts.MemoryProfileUsePath.empty())
    // -fmemory-profile-use (without any of the above options)
    PGOOpt = PGOOptions("", "", "", CodeGenOpts.MemoryProfileUsePath,
                        PGOOptions::NoAction, PGOOptions::NoCSAction,
                        ClPGOColdFuncAttr, CodeGenOpts.DebugInfoForProfiling);
  else if (CodeGenOpts.PseudoProbeForProfiling)
    // -fpseudo-probe-for-profiling
    PGOOpt = PGOOptions("", "", "", /*MemoryProfile=*/"", PGOOptions::NoAction,
                        PGOOptions::NoCSAction, ClPGOColdFuncAttr,
                        CodeGenOpts.DebugInfoForProfiling, true);
  else if (CodeGenOpts.DebugInfoForProfiling)
    // -fdebug-info-for-profiling
    PGOOpt = PGOOptions("", "", "", /*MemoryProfile=*/"", PGOOptions::NoAction,
                        PGOOptions::NoCSAction, ClPGOColdFuncAttr, true);

  // Check to see if we want to generate a CS profile.
  if (CodeGenOpts.hasProfileCSIRInstr()) {
    assert(!CodeGenOpts.hasProfileCSIRUse() &&
           "Cannot have both CSProfileUse pass and CSProfileGen pass at "
           "the same time");
    if (PGOOpt) {
      assert(PGOOpt->Action != PGOOptions::IRInstr &&
             PGOOpt->Action != PGOOptions::SampleUse &&
             "Cannot run CSProfileGen pass with ProfileGen or SampleUse "
             " pass");
      PGOOpt->CSProfileGenFile = getProfileGenName(CodeGenOpts);
      PGOOpt->CSAction = PGOOptions::CSIRInstr;
    } else
      PGOOpt = PGOOptions("", getProfileGenName(CodeGenOpts), "",
                          /*MemoryProfile=*/"", PGOOptions::NoAction,
                          PGOOptions::CSIRInstr, ClPGOColdFuncAttr,
                          CodeGenOpts.DebugInfoForProfiling);
  }
  if (TM)
    TM->setPGOOption(PGOOpt);

  PipelineTuningOptions PTO;
  PTO.LoopUnrolling = CodeGenOpts.UnrollLoops;
  PTO.LoopInterchange = CodeGenOpts.InterchangeLoops;
  PTO.LoopFusion = CodeGenOpts.FuseLoops;
  // For historical reasons, loop interleaving is set to mirror setting for loop
  // unrolling.
  PTO.LoopInterleaving = CodeGenOpts.UnrollLoops;
  PTO.LoopVectorization = CodeGenOpts.VectorizeLoop;
  PTO.SLPVectorization = CodeGenOpts.VectorizeSLP;
  PTO.MergeFunctions = CodeGenOpts.MergeFunctions;
  // Only enable CGProfilePass when using integrated assembler, since
  // non-integrated assemblers don't recognize .cgprofile section.
  PTO.CallGraphProfile = !CodeGenOpts.DisableIntegratedAS;
  PTO.UnifiedLTO = CodeGenOpts.UnifiedLTO;
  PTO.DevirtualizeSpeculatively = CodeGenOpts.DevirtualizeSpeculatively;

  LoopAnalysisManager LAM;
  FunctionAnalysisManager FAM;
  CGSCCAnalysisManager CGAM;
  ModuleAnalysisManager MAM;

  bool DebugPassStructure = CodeGenOpts.DebugPass == "Structure";
  PassInstrumentationCallbacks PIC;
  PrintPassOptions PrintPassOpts;
  PrintPassOpts.Indent = DebugPassStructure;
  PrintPassOpts.SkipAnalyses = DebugPassStructure;
  StandardInstrumentations SI(
      TheModule->getContext(),
      (CodeGenOpts.DebugPassManager || DebugPassStructure),
      CodeGenOpts.VerifyEach, PrintPassOpts);
  SI.registerCallbacks(PIC, &MAM);
  PassBuilder PB(TM.get(), PTO, PGOOpt, &PIC, CI.getVirtualFileSystemPtr());

  // Handle the assignment tracking feature options.
  switch (CodeGenOpts.getAssignmentTrackingMode()) {
  case CodeGenOptions::AssignmentTrackingOpts::Forced:
    PB.registerPipelineStartEPCallback(
        [&](ModulePassManager &MPM, OptimizationLevel Level) {
          MPM.addPass(AssignmentTrackingPass());
        });
    break;
  case CodeGenOptions::AssignmentTrackingOpts::Enabled:
    // Disable assignment tracking in LTO builds for now as the performance
    // cost is too high. Disable for LLDB tuning due to llvm.org/PR43126.
    if (!CodeGenOpts.PrepareForThinLTO && !CodeGenOpts.PrepareForLTO &&
        CodeGenOpts.getDebuggerTuning() != llvm::DebuggerKind::LLDB) {
      PB.registerPipelineStartEPCallback(
          [&](ModulePassManager &MPM, OptimizationLevel Level) {
            // Only use assignment tracking if optimisations are enabled.
            if (Level != OptimizationLevel::O0)
              MPM.addPass(AssignmentTrackingPass());
          });
    }
    break;
  case CodeGenOptions::AssignmentTrackingOpts::Disabled:
    break;
  }

  // Enable verify-debuginfo-preserve-each for new PM.
  DebugifyEachInstrumentation Debugify;
  DebugInfoPerPass DebugInfoBeforePass;
  if (CodeGenOpts.EnableDIPreservationVerify) {
    Debugify.setDebugifyMode(DebugifyMode::OriginalDebugInfo);
    Debugify.setDebugInfoBeforePass(DebugInfoBeforePass);

    if (!CodeGenOpts.DIBugsReportFilePath.empty())
      Debugify.setOrigDIVerifyBugsReportFilePath(
          CodeGenOpts.DIBugsReportFilePath);
    Debugify.registerCallbacks(PIC, MAM);

#if LLVM_ENABLE_DEBUGLOC_TRACKING_COVERAGE
    // If we're using debug location coverage tracking, mark all the
    // instructions coming out of the frontend without a DebugLoc as being
    // compiler-generated, to prevent both those instructions and new
    // instructions that inherit their location from being treated as
    // incorrectly empty locations.
    for (Function &F : *TheModule) {
      if (!F.getSubprogram())
        continue;
      for (BasicBlock &BB : F)
        for (Instruction &I : BB)
          if (!I.getDebugLoc())
            I.setDebugLoc(DebugLoc::getCompilerGenerated());
    }
#endif
  }
  // Register plugin callbacks with PB.
  for (const std::unique_ptr<PassPlugin> &Plugin : CI.getPassPlugins())
    Plugin->registerPassBuilderCallbacks(PB);
  for (const auto &PassCallback : CodeGenOpts.PassBuilderCallbacks)
    PassCallback(PB);
#define HANDLE_EXTENSION(Ext)                                                  \
  get##Ext##PluginInfo().RegisterPassBuilderCallbacks(PB);
#include "llvm/Support/Extension.def"

  // Register the target library analysis directly and give it a customized
  // preset TLI.
  std::unique_ptr<TargetLibraryInfoImpl> TLII(
      llvm::driver::createTLII(TargetTriple, CodeGenOpts.getVecLib()));
  FAM.registerPass([&] { return TargetLibraryAnalysis(*TLII); });

  // Register all the basic analyses with the managers.
  PB.registerModuleAnalyses(MAM);
  PB.registerCGSCCAnalyses(CGAM);
  PB.registerFunctionAnalyses(FAM);
  PB.registerLoopAnalyses(LAM);
  PB.crossRegisterProxies(LAM, FAM, CGAM, MAM);

  ModulePassManager MPM;
  // Add a verifier pass, before any other passes, to catch CodeGen issues.
  if (CodeGenOpts.VerifyModule)
    MPM.addPass(VerifierPass());

  if (!CodeGenOpts.DisableLLVMPasses) {
    // Map our optimization levels into one of the distinct levels used to
    // configure the pipeline.
    OptimizationLevel Level = mapToLevel(CodeGenOpts);

    const bool PrepareForThinLTO = CodeGenOpts.PrepareForThinLTO;
    const bool PrepareForLTO = CodeGenOpts.PrepareForLTO;

    if (LangOpts.ObjCAutoRefCount) {
      PB.registerPipelineStartEPCallback(
          [](ModulePassManager &MPM, OptimizationLevel Level) {
            if (Level != OptimizationLevel::O0)
              MPM.addPass(
                  createModuleToFunctionPassAdaptor(ObjCARCExpandPass()));
          });
      PB.registerScalarOptimizerLateEPCallback(
          [](FunctionPassManager &FPM, OptimizationLevel Level) {
            if (Level != OptimizationLevel::O0)
              FPM.addPass(ObjCARCOptPass());
          });
    }

    // If we reached here with a non-empty index file name, then the index
    // file was empty and we are not performing ThinLTO backend compilation
    // (used in testing in a distributed build environment).
    bool IsThinLTOPostLink = !CodeGenOpts.ThinLTOIndexFile.empty();
    // If so drop any the type test assume sequences inserted for whole program
    // vtables so that codegen doesn't complain.
    if (IsThinLTOPostLink)
      PB.registerPipelineStartEPCallback(
          [](ModulePassManager &MPM, OptimizationLevel Level) {
            MPM.addPass(DropTypeTestsPass());
          });

    // Register callbacks to schedule sanitizer passes at the appropriate part
    // of the pipeline.
    if (LangOpts.Sanitize.has(SanitizerKind::LocalBounds))
      PB.registerScalarOptimizerLateEPCallback([this](FunctionPassManager &FPM,
                                                      OptimizationLevel Level) {
        BoundsCheckingPass::Options Options;
        if (CodeGenOpts.SanitizeSkipHotCutoffs[SanitizerKind::SO_LocalBounds] ||
            ClSanitizeGuardChecks) {
          static_assert(SanitizerKind::SO_LocalBounds <=
                            std::numeric_limits<
                                decltype(Options.GuardKind)::value_type>::max(),
                        "Update type of llvm.allow.ubsan.check to represent "
                        "SanitizerKind::SO_LocalBounds.");
          Options.GuardKind = SanitizerKind::SO_LocalBounds;
        }
        Options.Merge =
            CodeGenOpts.SanitizeMergeHandlers.has(SanitizerKind::LocalBounds);
        if (!CodeGenOpts.SanitizeTrap.has(SanitizerKind::LocalBounds)) {
          Options.Rt = {
              /*MinRuntime=*/static_cast<bool>(
                  CodeGenOpts.SanitizeMinimalRuntime),
              /*MayReturn=*/
              CodeGenOpts.SanitizeRecover.has(SanitizerKind::LocalBounds),
              /*HandlerPreserveAllRegs=*/
              static_cast<bool>(CodeGenOpts.SanitizeHandlerPreserveAllRegs),
          };
        }
        FPM.addPass(BoundsCheckingPass(Options));
      });

    if (!IsThinLTOPostLink) {
      // Most sanitizers only run during PreLink stage.
      addSanitizers(TargetTriple, CodeGenOpts, LangOpts, PB);
      addKCFIPass(TargetTriple, LangOpts, PB);
      addLowerAllowCheckPass(CodeGenOpts, LangOpts, PB);

      PB.registerPipelineStartEPCallback(
          [&](ModulePassManager &MPM, OptimizationLevel Level) {
            if (Level == OptimizationLevel::O0 &&
                LangOpts.Sanitize.has(SanitizerKind::AllocToken)) {
              // With the default O0 pipeline, LibFunc attrs are not inferred,
              // so we insert it here because we need it for accurate memory
              // allocation function detection with -fsanitize=alloc-token.
              // Note: This could also be added to the default O0 pipeline, but
              // has a non-trivial effect on generated IR size (attributes).
              MPM.addPass(InferFunctionAttrsPass());
            }
          });

      if (CodeGenOpts.CopyProf) {
        // Early pass: insert callbacks into special member functions before the
        // inliner removes function boundaries.
        PB.registerPipelineEarlySimplificationEPCallback(
            [](ModulePassManager &MPM, OptimizationLevel, ThinOrFullLTOPhase) {
              MPM.addPass(createModuleToFunctionPassAdaptor(CopyProfPass()));
              MPM.addPass(ModuleCopyProfPass());
            });
        // Late pass: to reduce runtime overhead, instrument stores only after
        // optimizations have been run so only useful stores are instrumented.
        PB.registerOptimizerLastEPCallback([](ModulePassManager &MPM,
                                              OptimizationLevel,
                                              ThinOrFullLTOPhase) {
          MPM.addPass(createModuleToFunctionPassAdaptor(CopyProfStoresPass()));
        });
      }
    }

    if (std::optional<GCOVOptions> Options =
            getGCOVOptions(CodeGenOpts, LangOpts))
      PB.registerPipelineStartEPCallback(
          [this, Options](ModulePassManager &MPM, OptimizationLevel Level) {
            MPM.addPass(
                GCOVProfilerPass(*Options, CI.getVirtualFileSystemPtr()));
          });
    if (std::optional<InstrProfOptions> Options =
            getInstrProfOptions(CodeGenOpts, LangOpts))
      PB.registerPipelineStartEPCallback(
          [Options](ModulePassManager &MPM, OptimizationLevel Level) {
            MPM.addPass(InstrProfilingLoweringPass(*Options, false));
          });

    // TODO: Consider passing the MemoryProfileOutput to the pass builder via
    // the PGOOptions, and set this up there.
    if (!CodeGenOpts.MemoryProfileOutput.empty()) {
      PB.registerOptimizerLastEPCallback([](ModulePassManager &MPM,
                                            OptimizationLevel Level,
                                            ThinOrFullLTOPhase) {
        MPM.addPass(createModuleToFunctionPassAdaptor(MemProfilerPass()));
        MPM.addPass(ModuleMemProfilerPass());
      });
    }

    if (CodeGenOpts.FatLTO) {
      MPM.addPass(PB.buildFatLTODefaultPipeline(
          Level, PrepareForThinLTO,
          PrepareForThinLTO || shouldEmitRegularLTOSummary(),
          CodeGenOpts.VerifyModule));
    } else if (PrepareForThinLTO) {
      MPM.addPass(PB.buildThinLTOPreLinkDefaultPipeline(Level));
    } else if (PrepareForLTO) {
      MPM.addPass(PB.buildLTOPreLinkDefaultPipeline(Level));
    } else {
      MPM.addPass(PB.buildPerModuleDefaultPipeline(Level));
    }
  }

  // Link against bitcodes supplied via the -mlink-builtin-bitcode option
  if (CodeGenOpts.LinkBitcodePostopt) {
    MPM.addPass(LinkInModulesPass(BC));
    MPM.addPass(AssignGUIDPass());
  }

  if (LangOpts.HIPStdPar && !LangOpts.CUDAIsDevice &&
      LangOpts.HIPStdParInterposeAlloc)
    MPM.addPass(HipStdParAllocationInterpositionPass());

  // Add a verifier pass if requested. We don't have to do this if the action
  // requires code generation because there will already be a verifier pass in
  // the code-generation pipeline.
  // Since we already added a verifier pass above, this
  // might even not run the analysis, if previous passes caused no changes.
  if (!actionRequiresCodeGen(Action) && CodeGenOpts.VerifyModule)
    MPM.addPass(VerifierPass());

  if (Action == Backend_EmitBC || Action == Backend_EmitLL ||
      CodeGenOpts.FatLTO) {
    if (CodeGenOpts.PrepareForThinLTO && !CodeGenOpts.DisableLLVMPasses) {
      if (!TheModule->getModuleFlag("EnableSplitLTOUnit"))
        TheModule->addModuleFlag(llvm::Module::Error, "EnableSplitLTOUnit",
                                 CodeGenOpts.EnableSplitLTOUnit);
      if (Action == Backend_EmitBC) {
        if (!CodeGenOpts.ThinLinkBitcodeFile.empty()) {
          ThinLinkOS = openOutputFile(CodeGenOpts.ThinLinkBitcodeFile);
          if (!ThinLinkOS)
            return;
        }
        MPM.addPass(ThinLTOBitcodeWriterPass(
            *OS, ThinLinkOS ? &ThinLinkOS->os() : nullptr));
      } else if (Action == Backend_EmitLL) {
        MPM.addPass(PrintModulePass(*OS, "", CodeGenOpts.EmitLLVMUseLists,
                                    /*EmitLTOSummary=*/true,
                                    /*ShouldRenumberMetadata=*/true));
      }
    } else {
      // Emit a module summary by default for Regular LTO except for ld64
      // targets
      bool EmitLTOSummary = shouldEmitRegularLTOSummary();
      if (EmitLTOSummary) {
        if (!TheModule->getModuleFlag("ThinLTO") && !CodeGenOpts.UnifiedLTO)
          TheModule->addModuleFlag(llvm::Module::Error, "ThinLTO", uint32_t(0));
        if (!TheModule->getModuleFlag("EnableSplitLTOUnit"))
          TheModule->addModuleFlag(llvm::Module::Error, "EnableSplitLTOUnit",
                                   uint32_t(1));
      }
      if (Action == Backend_EmitBC) {
        MPM.addPass(BitcodeWriterPass(*OS, CodeGenOpts.EmitLLVMUseLists,
                                      EmitLTOSummary));
      } else if (Action == Backend_EmitLL) {
        MPM.addPass(PrintModulePass(*OS, "", CodeGenOpts.EmitLLVMUseLists,
                                    EmitLTOSummary,
                                    /*ShouldRenumberMetadata=*/true));
      }
    }

    if (shouldEmitUnifiedLTOModueFlag() &&
        !TheModule->getModuleFlag("UnifiedLTO"))
      TheModule->addModuleFlag(llvm::Module::Error, "UnifiedLTO", uint32_t(1));
  }

  // FIXME: This should eventually be replaced by a first-class driver option.
  // This should be done for both clang and flang simultaneously.
  // Print a textual, '-passes=' compatible, representation of pipeline if
  // requested.
  if (PrintPipelinePasses) {
    MPM.printPipeline(outs(), [&PIC](StringRef ClassName) {
      auto PassName = PIC.getPassNameForClassName(ClassName);
      return PassName.empty() ? ClassName : PassName;
    });
    outs() << "\n";
    return;
  }

  // Now that we have all of the passes ready, run them.
  {
    PrettyStackTraceString CrashInfo("Optimizer");
    llvm::TimeTraceScope TimeScope("Optimizer");
    Timer timer;
    if (CI.getCodeGenOpts().TimePasses) {
      timer.init("optimizer", "Optimizer", CI.getTimerGroup());
      CI.getFrontendTimer().yieldTo(timer);
    }
    MPM.run(*TheModule, MAM);
    if (CI.getCodeGenOpts().TimePasses)
      timer.yieldTo(CI.getFrontendTimer());
  }
}

void EmitAssemblyHelper::RunCodegenPipeline(
    BackendAction Action, std::unique_ptr<raw_pwrite_stream> &OS,
    std::unique_ptr<llvm::ToolOutputFile> &DwoOS) {
  if (!actionRequiresCodeGen(Action))
    return;

  // Normal mode, emit a .s or .o file by running the code generator. Note,
  // this also adds codegenerator level optimization passes.
  CodeGenFileType CGFT = getCodeGenFileType(Action);

  // Invoke pre-codegen callback from plugin, which might want to take over the
  // entire code generation itself.
  for (const std::unique_ptr<llvm::PassPlugin> &Plugin : CI.getPassPlugins()) {
    if (Plugin->invokePreCodeGenCallback(*TheModule, *TM, CGFT, *OS))
      return;
  }

  if (!CodeGenOpts.SplitDwarfOutput.empty()) {
    DwoOS = openOutputFile(CodeGenOpts.SplitDwarfOutput);
    if (!DwoOS)
      return;
  }

  TimeCodegenPasses([&]() {
    Error CodeGenError = runCodeGenPipeline(
        *TM, *TheModule, *OS, DwoOS, CGFT, PrintPipelinePasses.has_value(),
        !CodeGenOpts.VerifyModule, /*DisableSimplifyLibCalls=*/false,
        CI.getVirtualFileSystemPtr());
    if (CodeGenError)
      Diags.Report(diag::err_fe_unable_to_interface_with_target);
  });
}

void EmitAssemblyHelper::TimeCodegenPasses(
    llvm::function_ref<void()> RunPasses) {
  PrettyStackTraceString CrashInfo("Code generation");
  llvm::TimeTraceScope TimeScope("CodeGenPasses");
  Timer timer;
  if (CI.getCodeGenOpts().TimePasses) {
    timer.init("codegen", "Machine code generation", CI.getTimerGroup());
    CI.getFrontendTimer().yieldTo(timer);
  }
  RunPasses();
  if (CI.getCodeGenOpts().TimePasses)
    timer.yieldTo(CI.getFrontendTimer());
}

void EmitAssemblyHelper::emitAssembly(BackendAction Action,
                                      std::unique_ptr<raw_pwrite_stream> OS,
                                      BackendConsumer *BC) {
  setCommandLineOpts(CodeGenOpts, CI.getVirtualFileSystem());

  bool RequiresCodeGen = actionRequiresCodeGen(Action);
  CreateTargetMachine(RequiresCodeGen);

  if (RequiresCodeGen && !TM)
    return;
  if (TM && TheModule->getDataLayout().isDefault())
    TheModule->setDataLayout(TheModule->getTargetTriple().computeDataLayout(
        TM->getTargetABIName(*TheModule)));

  // Before executing passes, print the final values of the LLVM options.
  cl::PrintOptionValues();

  std::unique_ptr<llvm::ToolOutputFile> ThinLinkOS, DwoOS;
  RunOptimizationPipeline(Action, OS, ThinLinkOS, BC);
  RunCodegenPipeline(Action, OS, DwoOS);

  if (ThinLinkOS)
    ThinLinkOS->keep();
  if (DwoOS)
    DwoOS->keep();
}

static void
runThinLTOBackend(CompilerInstance &CI, ModuleSummaryIndex *CombinedIndex,
                  llvm::Module *M, std::unique_ptr<raw_pwrite_stream> OS,
                  std::string SampleProfile, std::string ProfileRemapping,
                  BackendAction Action) {
  DiagnosticsEngine &Diags = CI.getDiagnostics();
  const auto &CGOpts = CI.getCodeGenOpts();
  const auto &TOpts = CI.getTargetOpts();
  DenseMap<StringRef, DenseMap<GlobalValue::GUID, GlobalValueSummary *>>
      ModuleToDefinedGVSummaries;
  CombinedIndex->collectDefinedGVSummariesPerModule(ModuleToDefinedGVSummaries);

  setCommandLineOpts(CGOpts, CI.getVirtualFileSystem());

  // We can simply import the values mentioned in the combined index, since
  // we should only invoke this using the individual indexes written out
  // via a WriteIndexesThinBackend.
  FunctionImporter::ImportIDTable ImportIDs;
  FunctionImporter::ImportMapTy ImportList(ImportIDs);
  if (!lto::initImportList(*M, *CombinedIndex, ImportList))
    return;

  auto AddStream = [&](size_t Task, const Twine &ModuleName) {
    return std::make_unique<CachedFileStream>(std::move(OS),
                                              CGOpts.ObjectFilenameForDebug);
  };
  lto::Config Conf;
  if (CGOpts.SaveTempsFilePrefix != "") {
    if (Error E = Conf.addSaveTemps(CGOpts.SaveTempsFilePrefix + ".",
                                    /* UseInputModulePath */ false)) {
      handleAllErrors(std::move(E), [&](ErrorInfoBase &EIB) {
        errs() << "Error setting up ThinLTO save-temps: " << EIB.message()
               << '\n';
      });
    }
  }
  Conf.CPU = TOpts.CPU;
  Conf.CodeModel = getCodeModel(CGOpts);
  Conf.MAttrs = TOpts.Features;
  Conf.RelocModel = CGOpts.RelocationModel;
  std::optional<CodeGenOptLevel> OptLevelOrNone =
      CodeGenOpt::getLevel(CGOpts.OptimizationLevel);
  assert(OptLevelOrNone && "Invalid optimization level!");
  Conf.CGOptLevel = *OptLevelOrNone;
  Conf.OptLevel = CGOpts.OptimizationLevel;
  initTargetOptions(CI, Diags, Conf.Options);
  Conf.SampleProfile = std::move(SampleProfile);
  Conf.PTO.LoopUnrolling = CGOpts.UnrollLoops;
  Conf.PTO.LoopInterchange = CGOpts.InterchangeLoops;
  Conf.PTO.LoopFusion = CGOpts.FuseLoops;
  // For historical reasons, loop interleaving is set to mirror setting for loop
  // unrolling.
  Conf.PTO.LoopInterleaving = CGOpts.UnrollLoops;
  Conf.PTO.LoopVectorization = CGOpts.VectorizeLoop;
  Conf.PTO.SLPVectorization = CGOpts.VectorizeSLP;
  // Only enable CGProfilePass when using integrated assembler, since
  // non-integrated assemblers don't recognize .cgprofile section.
  Conf.PTO.CallGraphProfile = !CGOpts.DisableIntegratedAS;

  // Context sensitive profile.
  if (CGOpts.hasProfileCSIRInstr()) {
    Conf.RunCSIRInstr = true;
    Conf.CSIRProfile = getProfileGenName(CGOpts);
  } else if (CGOpts.hasProfileCSIRUse()) {
    Conf.RunCSIRInstr = false;
    Conf.CSIRProfile = std::move(CGOpts.ProfileInstrumentUsePath);
  }

  Conf.ProfileRemapping = std::move(ProfileRemapping);
  Conf.DebugPassManager = CGOpts.DebugPassManager;
  Conf.VerifyEach = CGOpts.VerifyEach;
  Conf.RemarksWithHotness = CGOpts.DiagnosticsWithHotness;
  Conf.RemarksFilename = CGOpts.OptRecordFile;
  Conf.RemarksPasses = CGOpts.OptRecordPasses;
  Conf.RemarksFormat = CGOpts.OptRecordFormat;
  Conf.SplitDwarfFile = CGOpts.SplitDwarfFile;
  Conf.SplitDwarfOutput = CGOpts.SplitDwarfOutput;
  for (auto &Plugin : CI.getPassPlugins())
    Conf.LoadedPassPlugins.push_back(Plugin.get());
  switch (Action) {
  case Backend_EmitNothing:
    Conf.PreCodeGenModuleHook = [](size_t Task, const llvm::Module &Mod) {
      return false;
    };
    break;
  case Backend_EmitLL:
    Conf.PreCodeGenModuleHook = [&](size_t Task, const llvm::Module &Mod) {
      M->renumberMetadataForAssembly();
      M->print(*OS, nullptr, CGOpts.EmitLLVMUseLists);
      return false;
    };
    break;
  case Backend_EmitBC:
    Conf.PreCodeGenModuleHook = [&](size_t Task, const llvm::Module &Mod) {
      WriteBitcodeToFile(*M, *OS, CGOpts.EmitLLVMUseLists);
      return false;
    };
    break;
  default:
    Conf.CGFileType = getCodeGenFileType(Action);
    break;
  }

  // FIXME: Both ExecuteAction and thinBackend set up optimization remarks for
  // the same context.
  // FIXME: This does not yet set the list of bitcode libfuncs that it isn't
  // safe to call. This precludes bitcode libc in distributed ThinLTO.
  finalizeLLVMOptimizationRemarks(M->getContext());
  if (Error E = thinBackend(
          Conf, -1, AddStream, *M, *CombinedIndex, ImportList,
          ModuleToDefinedGVSummaries[M->getModuleIdentifier()],
          /*ModuleMap=*/nullptr, Conf.CodeGenOnly, /*BitcodeLibFuncs=*/{},
          /*IRAddStream=*/nullptr, CGOpts.CmdArgs)) {
    handleAllErrors(std::move(E), [&](ErrorInfoBase &EIB) {
      errs() << "Error running ThinLTO backend: " << EIB.message() << '\n';
    });
  }
}

static void createAndEmbedModuleForDynamicDebugging(
    CompilerInstance &CI, CodeGenOptions &CGOpts, llvm::Module *M,
    IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS, BackendConsumer *BC) {
  /// Helper for saving the module(s) at various dyndbg stages.
  auto SaveModule = [&](StringRef Name, llvm::Module &M) {
    if (CGOpts.SaveDynDbgTempsFilePrefix == "")
      return;
    std::error_code EC;
    std::string Path =
        Twine(CGOpts.SaveDynDbgTempsFilePrefix + "." + Name + ".ll").str();
    raw_fd_ostream OS(Path, EC, sys::fs::OpenFlags::OF_None);
    if (EC) {
      // Copy -save-temps behaviour: this is a debugging option so we simply
      // exit if there's an issue.
      errs() << "failed to open " << Path << ": " << EC.message() << '\n';
      errs().flush();
      exit(1);
    }
    M.print(OS, nullptr);
  };

  // Compute a hash suffix for promoting static globals (once per TU).
  std::string PromotionSuffix;
  {
    // LLVM's hash/hash_combine is not guaranteed to be stable.
    MD5 Hash;
    // Include args in the hash else preprocessor definitions used to alter
    // the same source file compiled twice won't generate unique hashes.
    Hash.update(CGOpts.CmdArgs);
    for (auto *CU : M->debug_compile_units()) {
      if (CU->getDirectory().size() > 0)
        Hash.update(CU->getDirectory());

      Hash.update(CU->getFilename());
    }

    MD5::MD5Result Result;
    Hash.final(Result);
    PromotionSuffix = ".dyndbg." + utohexstr(Result.low());
  }

  SaveModule("dyndbg.0.input", *M);
  // Modify M as needed and create an "unoptimized" clone.
  auto UnoptM = prepareForDynamicDebugging(M, PromotionSuffix);
  SaveModule("dyndbg.1.inner", *UnoptM);

  if (!CGOpts.DiscardDynamicDebuggingDebugModule) {
    CodeGenOptions UnoptOpts = CGOpts;
    UnoptOpts.OptimizationLevel = 0;
    UnoptOpts.OptimizeSize = 0;
    EmitAssemblyHelper AsmHelper(CI, UnoptOpts, UnoptM.get(), VFS);

    // Create a buffer and ostream for the inner ELF.
    SmallVector<char, 0> UnoptBuf;
    std::unique_ptr<llvm::raw_pwrite_stream> UnoptOS =
        std::make_unique<llvm::raw_svector_ostream>(UnoptBuf);

    // Always run the full codegen pipeline (Backend_EmitObj). This causes
    // assertion failures if there's no registered backend which is why we
    // disable the feature if that's the case (see
    // warn_dyndbg_unable_to_create_target above).
    AsmHelper.emitAssembly(Backend_EmitObj, std::move(UnoptOS), BC);
    assert(!UnoptBuf.empty() && "Expected emitAssembly to fill UnoptBuf");

    // Inject the inner ELF into the outer module.
    StringRef SR(UnoptBuf.data(), UnoptBuf.size());
    std::unique_ptr<MemoryBuffer> Buf =
        MemoryBuffer::getMemBuffer(SR, "", false);

    GlobalVariable *EmbeddedGV =
        llvm::embedBufferInModule(*M, *Buf, ".debug_llvm_dyndbg", Align(8),
                                  /*SectionExclude*/ false);
    // Add ELF section properties metadata.
    auto &C = M->getContext();
    auto getU32Metadata = [&C](unsigned Val) {
      return ConstantAsMetadata::get(ConstantInt::get(C, APInt(32, Val)));
    };
    EmbeddedGV->addMetadata(
        LLVMContext::MD_elf_section_properties,
        *MDTuple::get(C, {/*sh_type*/ getU32Metadata(ELF::SHT_LLVM_DYNDBG_ELF),
                          /*sh_entsize*/ getU32Metadata(0)}));
  }
  SaveModule("dyndbg.2.outer", *M);
}

void clang::emitBackendOutput(CompilerInstance &CI, CodeGenOptions &CGOpts,
                              llvm::Module *M, BackendAction Action,
                              IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS,
                              std::unique_ptr<raw_pwrite_stream> OS,
                              BackendConsumer *BC) {
  llvm::TimeTraceScope TimeScope("Backend");
  DiagnosticsEngine &Diags = CI.getDiagnostics();

  std::unique_ptr<llvm::Module> EmptyModule;
  if (!CGOpts.ThinLTOIndexFile.empty()) {
    // FIXME(sandboxing): Figure out how to support distributed indexing.
    auto BypassSandbox = sys::sandbox::scopedDisable();
    // If we are performing a ThinLTO importing compile, load the function index
    // into memory and pass it into runThinLTOBackend, which will run the
    // function importer and invoke LTO passes.
    std::unique_ptr<ModuleSummaryIndex> CombinedIndex;
    if (Error E = llvm::getModuleSummaryIndexForFile(
                      CGOpts.ThinLTOIndexFile,
                      /*IgnoreEmptyThinLTOIndexFile*/ true)
                      .moveInto(CombinedIndex)) {
      logAllUnhandledErrors(std::move(E), errs(),
                            "Error loading index file '" +
                                CGOpts.ThinLTOIndexFile + "': ");
      return;
    }

    // A null CombinedIndex means we should skip ThinLTO compilation
    // (LLVM will optionally ignore empty index files, returning null instead
    // of an error).
    if (CombinedIndex) {
      if (!CombinedIndex->skipModuleByDistributedBackend()) {
        runThinLTOBackend(CI, CombinedIndex.get(), M, std::move(OS),
                          CGOpts.SampleProfileFile, CGOpts.ProfileRemappingFile,
                          Action);
        return;
      }
      // Distributed indexing detected that nothing from the module is needed
      // for the final linking. So we can skip the compilation. We sill need to
      // output an empty object file to make sure that a linker does not fail
      // trying to read it. Also for some features, like CFI, we must skip
      // the compilation as CombinedIndex does not contain all required
      // information.
      EmptyModule = std::make_unique<llvm::Module>("empty", M->getContext());
      EmptyModule->setTargetTriple(M->getTargetTriple());
      M = EmptyModule.get();
    }
  }

  bool EnableDynamicDebugging = CGOpts.DynamicDebugging;
  if (EnableDynamicDebugging) {
    // Disable dyndbg if the target isn't available as we're compiling to the
    // inner module (unless we're discarding it for debugging/testing).
    std::string Error;
    const llvm::Target *TheTarget =
        TargetRegistry::lookupTarget(M->getTargetTriple(), Error);
    if (!TheTarget && !CGOpts.DiscardDynamicDebuggingDebugModule) {
      Diags.Report(diag::warn_dyndbg_unable_to_create_target) << Error;
      EnableDynamicDebugging = false;
    }

    // Instrumentation causes issues (parts of LLVM expect certain globals to
    // have initializers). Intrinsics may already have been added to IR by now,
    // so we can't just turn it off for the inner module (we'd have to strip
    // them out / not clone them). TODO: Support instrumentation.
    if (CGOpts.getProfileInstr() != driver::ProfileInstrKind::ProfileNone) {
      Diags.Report(diag::err_dyndbg_no_instrumentation);
      EnableDynamicDebugging = false;
    }
  }
  if (EnableDynamicDebugging)
    createAndEmbedModuleForDynamicDebugging(CI, CGOpts, M, VFS, BC);

  EmitAssemblyHelper AsmHelper(CI, CGOpts, M, VFS);
  AsmHelper.emitAssembly(Action, std::move(OS), BC);

  // Verify the module's DataLayout against the one the target computes for the
  // module's ABI. This respects the target-abi module flag rather than assuming
  // the DataLayout is a fixed property of the target options.
  if (AsmHelper.TM) {
    std::string DLDesc = M->getDataLayout().getStringRepresentation();
    std::string TDesc = M->getTargetTriple().computeDataLayout(
        AsmHelper.TM->getTargetABIName(*M));
    if (DLDesc != TDesc)
      Diags.Report(diag::err_data_layout_mismatch) << DLDesc << TDesc;
  }
}

// With -fembed-bitcode, save a copy of the llvm IR as data in the
// __LLVM,__bitcode section.
void clang::EmbedBitcode(llvm::Module *M, const CodeGenOptions &CGOpts,
                         llvm::MemoryBufferRef Buf) {
  if (CGOpts.getEmbedBitcode() == CodeGenOptions::Embed_Off)
    return;
  llvm::embedBitcodeInModule(
      *M, Buf, CGOpts.getEmbedBitcode() != CodeGenOptions::Embed_Marker,
      CGOpts.getEmbedBitcode() != CodeGenOptions::Embed_Bitcode,
      CGOpts.CmdArgs);
}

void clang::EmbedObject(llvm::Module *M, const CodeGenOptions &CGOpts,
                        llvm::vfs::FileSystem &VFS, DiagnosticsEngine &Diags) {
  if (CGOpts.OffloadObjects.empty())
    return;

  for (StringRef OffloadObject : CGOpts.OffloadObjects) {
    llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> ObjectOrErr =
        VFS.getBufferForFile(OffloadObject);
    if (ObjectOrErr.getError()) {
      Diags.Report(diag::err_failed_to_open_for_embedding) << OffloadObject;
      return;
    }

    llvm::embedBufferInModule(*M, **ObjectOrErr, ".llvm.offloading",
                              Align(object::OffloadBinary::getAlignment()));
  }
}