aotcompile.cpp

// This file is a part of Julia. License is MIT: https://julialang.org/license

#include "llvm-version.h"
#include "platform.h"
#include "options.h"

// target support
#include <llvm/ADT/Triple.h>
#include <llvm/Support/TargetRegistry.h>
#include <llvm/Target/TargetMachine.h>
#include <llvm/IR/DataLayout.h>
#include <llvm/Analysis/TargetTransformInfo.h>
#include <llvm/Analysis/TargetLibraryInfo.h>

// analysis passes
#include <llvm/Analysis/Passes.h>
#include <llvm/Analysis/BasicAliasAnalysis.h>
#include <llvm/Analysis/TypeBasedAliasAnalysis.h>
#include <llvm/Analysis/ScopedNoAliasAA.h>
#include <llvm/IR/Verifier.h>
#include <llvm/Transforms/IPO.h>
#include <llvm/Transforms/Scalar.h>
#include <llvm/Transforms/Vectorize.h>
#if defined(JL_ASAN_ENABLED)
#include <llvm/Transforms/Instrumentation/AddressSanitizer.h>
#endif
#include <llvm/Transforms/Scalar/GVN.h>
#include <llvm/Transforms/IPO/AlwaysInliner.h>
#include <llvm/Transforms/InstCombine/InstCombine.h>
#include <llvm/Transforms/Scalar/InstSimplifyPass.h>
#if defined(USE_POLLY)
#include <polly/RegisterPasses.h>
#include <polly/LinkAllPasses.h>
#include <polly/CodeGen/CodegenCleanup.h>
#if defined(USE_POLLY_ACC)
#include <polly/Support/LinkGPURuntime.h>
#endif
#endif

// for outputting assembly
#include <llvm/Bitcode/BitcodeWriter.h>
#include <llvm/Bitcode/BitcodeWriterPass.h>
#include "llvm/Object/ArchiveWriter.h"
#include <llvm/IR/IRPrintingPasses.h>
#include <llvm/CodeGen/AsmPrinter.h>
#include <llvm/CodeGen/MachineModuleInfo.h>
#include <llvm/CodeGen/TargetPassConfig.h>
#include <llvm/MC/MCAsmInfo.h>
#include <llvm/MC/MCStreamer.h>
#include <llvm/MC/MCAsmBackend.h>
#include <llvm/MC/MCCodeEmitter.h>
#if JL_LLVM_VERSION >= 100000
#include <llvm/Support/CodeGen.h>
#endif

#include <llvm/IR/LegacyPassManagers.h>
#include <llvm/Transforms/Utils/Cloning.h>


using namespace llvm;

// our passes
namespace llvm {
    extern Pass *createLowerSimdLoopPass();
}

#if JL_LLVM_VERSION < 100000
static const TargetMachine::CodeGenFileType CGFT_ObjectFile = TargetMachine::CGFT_ObjectFile;
#endif


#include "julia.h"
#include "julia_internal.h"
#include "jitlayers.h"
#include "julia_assert.h"

// MSVC's link.exe requires each function declaration to have a Comdat section
// So rather than litter the code with conditionals,
// all global values that get emitted call this function
// and it decides whether the definition needs a Comdat section and adds the appropriate declaration
template<class T> // for GlobalObject's
static T *addComdat(T *G)
{
#if defined(_OS_WINDOWS_)
    if (!G->isDeclaration()) {
        // Add comdat information to make MSVC link.exe happy
        // it's valid to emit this for ld.exe too,
        // but makes it very slow to link for no benefit
#if defined(_COMPILER_MICROSOFT_)
        Comdat *jl_Comdat = G->getParent()->getOrInsertComdat(G->getName());
        // ELF only supports Comdat::Any
        jl_Comdat->setSelectionKind(Comdat::NoDuplicates);
        G->setComdat(jl_Comdat);
#endif
        // add __declspec(dllexport) to everything marked for export
        if (G->getLinkage() == GlobalValue::ExternalLinkage)
            G->setDLLStorageClass(GlobalValue::DLLExportStorageClass);
        else
            G->setDLLStorageClass(GlobalValue::DefaultStorageClass);
    }
#endif
    return G;
}


typedef struct {
    std::unique_ptr<Module> M;
    std::vector<GlobalValue*> jl_sysimg_fvars;
    std::vector<GlobalValue*> jl_sysimg_gvars;
    std::map<jl_code_instance_t*, std::tuple<uint32_t, uint32_t>> jl_fvar_map;
    std::map<void*, int32_t> jl_value_to_llvm; // uses 1-based indexing
} jl_native_code_desc_t;

extern "C" JL_DLLEXPORT
void jl_get_function_id(void *native_code, jl_code_instance_t *codeinst,
        int32_t *func_idx, int32_t *specfunc_idx)
{
    jl_native_code_desc_t *data = (jl_native_code_desc_t*)native_code;
    if (data) {
        // get the function index in the fvar lookup table
        auto it = data->jl_fvar_map.find(codeinst);
        if (it != data->jl_fvar_map.end()) {
            std::tie(*func_idx, *specfunc_idx) = it->second;
        }
    }
}

extern "C"
int32_t jl_get_llvm_gv(void *native_code, jl_value_t *p)
{
    // map a jl_value_t memory location to a GlobalVariable
    jl_native_code_desc_t *data = (jl_native_code_desc_t*)native_code;
    if (data) {
        auto it = data->jl_value_to_llvm.find(p);
        if (it != data->jl_value_to_llvm.end()) {
            return it->second;
        }
    }
    return 0;
}

extern "C" JL_DLLEXPORT
Module* jl_get_llvm_module(void *native_code)
{
    jl_native_code_desc_t *data = (jl_native_code_desc_t*)native_code;
    if (data)
        return data->M.get();
    else
        return NULL;
}

extern "C" JL_DLLEXPORT
GlobalValue* jl_get_llvm_function(void *native_code, uint32_t idx)
{
    jl_native_code_desc_t *data = (jl_native_code_desc_t*)native_code;
    if (data)
        return data->jl_sysimg_fvars[idx];
    else
        return NULL;
}

extern "C" JL_DLLEXPORT
LLVMContext* jl_get_llvm_context(void *native_code)
{
    jl_native_code_desc_t *data = (jl_native_code_desc_t*)native_code;
    if (data)
        return &data->M->getContext();
    else
        return NULL;
}


static void emit_offset_table(Module &mod, const std::vector<GlobalValue*> &vars, StringRef name, Type *T_psize)
{
    // Emit a global variable with all the variable addresses.
    // The cloning pass will convert them into offsets.
    assert(!vars.empty());
    size_t nvars = vars.size();
    std::vector<Constant*> addrs(nvars);
    for (size_t i = 0; i < nvars; i++) {
        Constant *var = vars[i];
        addrs[i] = ConstantExpr::getBitCast(var, T_psize);
    }
    ArrayType *vars_type = ArrayType::get(T_psize, nvars);
    new GlobalVariable(mod, vars_type, true,
                       GlobalVariable::ExternalLinkage,
                       ConstantArray::get(vars_type, addrs),
                       name);
}

static bool is_safe_char(unsigned char c)
{
    return ('0' <= c && c <= '9') ||
           ('A' <= c && c <= 'Z') ||
           ('a' <= c && c <= 'z') ||
           (c == '_' || c == '$') ||
           (c >= 128 && c < 255);
}

static const char hexchars[16] = {
    '0', '1', '2', '3', '4', '5', '6', '7',
    '8', '9', 'A', 'B', 'C', 'D', 'E', 'F' };

static const char *const common_names[256] = {
//  0, 1, 2, 3, 4, 5, 6, 7, 8, 9, a, b, c, d, e, f
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x00
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x10
    "SP", "NOT", "DQT", "YY", 0, "REM", "AND", "SQT", // 0x20
      "LPR", "RPR", "MUL", "SUM", 0, "SUB", "DOT", "DIV", // 0x28
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, "COL", 0, "LT", "EQ", "GT", "QQ", // 0x30
    "AT", 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x40
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, "LBR", "RDV", "RBR", "POW", 0, // 0x50
    "TIC", 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 0x60
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, "LCR", "OR", "RCR", "TLD", "DEL", // 0x70
    0 }; // remainder is filled with zeros, though are also all safe characters

// reversibly removes special characters from the name of GlobalObjects,
// which might cause them to be treated special by LLVM or the system linker
// the only non-identifier characters we allow to appear are '.' and '$',
// and all of UTF-8 above code-point 128 (except 255)
// most are given "friendly" abbreviations
// the remaining few will print as hex
// e.g. mangles "llvm.a≠a$a!a##" as "llvmDOT.a≠a$aNOT.aYY.YY."
static void makeSafeName(GlobalObject &G)
{
    StringRef Name = G.getName();
    SmallVector<char, 32> SafeName;
    for (unsigned char c : Name.bytes()) {
        if (is_safe_char(c)) {
            SafeName.push_back(c);
        }
        else {
            if (common_names[c]) {
                SafeName.push_back(common_names[c][0]);
                SafeName.push_back(common_names[c][1]);
                if (common_names[c][2])
                    SafeName.push_back(common_names[c][2]);
            }
            else {
                SafeName.push_back(hexchars[(c >> 4) & 0xF]);
                SafeName.push_back(hexchars[c & 0xF]);
            }
            SafeName.push_back('.');
        }
    }
    if (SafeName.size() != Name.size())
        G.setName(StringRef(SafeName.data(), SafeName.size()));
}


// takes the running content that has collected in the shadow module and dump it to disk
// this builds the object file portion of the sysimage files for fast startup
extern "C" JL_DLLEXPORT
void *jl_create_native(jl_array_t *methods, const jl_cgparams_t cgparams)
{
    jl_native_code_desc_t *data = new jl_native_code_desc_t;
    jl_codegen_params_t params;
    params.params = &cgparams;
    std::map<jl_code_instance_t*, jl_compile_result_t> emitted;
    jl_method_instance_t *mi = NULL;
    jl_code_info_t *src = NULL;
    JL_GC_PUSH1(&src);
    JL_LOCK(&codegen_lock);

    // compile all methods for the current world and type-inference world
    size_t compile_for[] = { jl_typeinf_world, jl_world_counter };
    for (int worlds = 0; worlds < 2; worlds++) {
        params.world = compile_for[worlds];
        if (!params.world)
            continue;
        size_t i, l;
        for (i = 0, l = jl_array_len(methods); i < l; i++) {
            mi = (jl_method_instance_t*)jl_array_ptr_ref(methods, i);
            src = NULL;
            // if this method is generally visible to the current compilation world,
            // and this is either the primary world, or not applicable in the primary world
            // then we want to compile and emit this
            if (mi->def.method->primary_world <= params.world && params.world <= mi->def.method->deleted_world) {
                // find and prepare the source code to compile
                jl_value_t *ci = jl_rettype_inferred(mi, params.world, params.world);
                jl_code_instance_t *codeinst = NULL;
                if (ci != jl_nothing) {
                    codeinst = (jl_code_instance_t*)ci;
                    src = (jl_code_info_t*)codeinst->inferred;
                    jl_method_t *def = codeinst->def->def.method;
                    if ((jl_value_t*)src == jl_nothing)
                        src = NULL;
                    if (src && jl_is_method(def))
                        src = jl_uncompress_ir(def, codeinst, (jl_array_t*)src);
                }
                if (src == NULL || !jl_is_code_info(src)) {
                    src = jl_type_infer(mi, params.world, 0);
                    codeinst = jl_get_method_inferred(mi, src->rettype, src->min_world, src->max_world);
                    if (src->inferred && !codeinst->inferred)
                        codeinst->inferred = jl_nothing;
                }
                if (src && !emitted.count(codeinst)) {
                    // now add it to our compilation results
                    JL_GC_PROMISE_ROOTED(codeinst->rettype);
                    jl_compile_result_t result = jl_emit_code(mi, src, codeinst->rettype, params);
                    if (std::get<0>(result))
                        emitted[codeinst] = std::move(result);
                }
            }
        }
        // finally, make sure all referenced methods also get compiled or fixed up
        jl_compile_workqueue(emitted, params);
    }
    JL_GC_POP();

    // process the globals array, before jl_merge_module destroys them
    std::vector<std::string> gvars;

    for (auto &global : params.globals) {
        gvars.push_back(std::string(global.second->getName()));
        data->jl_value_to_llvm[global.first] = gvars.size();
    }

    // clones the contents of the module `m` to the shadow_output collector
    // while examining and recording what kind of function pointer we have
    ValueToValueMapTy VMap;
    std::unique_ptr<Module> clone(CloneModule(*shadow_output, VMap));
    for (auto &def : emitted) {
        jl_merge_module(clone.get(), std::move(std::get<0>(def.second)));
        jl_code_instance_t *this_code = def.first;
        jl_llvm_functions_t decls = std::get<1>(def.second);
        StringRef func = decls.functionObject;
        StringRef cfunc = decls.specFunctionObject;
        uint32_t func_id = 0;
        uint32_t cfunc_id = 0;
        if (func == "jl_fptr_args") {
            func_id = -1;
        }
        else if (func == "jl_fptr_sparam") {
            func_id = -2;
        }
        else {
            data->jl_sysimg_fvars.push_back(cast<Function>(clone->getNamedValue(func)));
            func_id = data->jl_sysimg_fvars.size();
        }
        if (!cfunc.empty()) {
            data->jl_sysimg_fvars.push_back(cast<Function>(clone->getNamedValue(cfunc)));
            cfunc_id = data->jl_sysimg_fvars.size();
        }
        data->jl_fvar_map[this_code] = std::make_tuple(func_id, cfunc_id);
    }

    // now get references to the globals in the merged module
    // and set them to be internalized and initialized at startup
    for (auto &global : gvars) {
        GlobalVariable *G = cast<GlobalVariable>(clone->getNamedValue(global));
        G->setInitializer(ConstantPointerNull::get(cast<PointerType>(G->getValueType())));
        G->setLinkage(GlobalVariable::InternalLinkage);
        data->jl_sysimg_gvars.push_back(G);
    }

#if defined(_OS_WINDOWS_) && defined(_CPU_X86_64_)
    // setting the function personality enables stack unwinding and catching exceptions
    // so make sure everything has something set
    Type *T_int32 = Type::getInt32Ty(clone->getContext());
    Function *juliapersonality_func =
       Function::Create(FunctionType::get(T_int32, true),
           Function::ExternalLinkage, "__julia_personality", clone.get());
    juliapersonality_func->setDLLStorageClass(GlobalValue::DLLImportStorageClass);
#endif

    // move everything inside, now that we've merged everything
    // (before adding the exported headers)
    for (GlobalObject &G : clone->global_objects()) {
        if (!G.isDeclaration()) {
            G.setLinkage(Function::InternalLinkage);
            makeSafeName(G);
            addComdat(&G);
#if defined(_OS_WINDOWS_) && defined(_CPU_X86_64_)
            // Add unwind exception personalities to functions to handle async exceptions
            if (Function *F = dyn_cast<Function>(&G))
                F->setPersonalityFn(juliapersonality_func);
#endif
        }
    }

    data->M = std::move(clone);

    JL_UNLOCK(&codegen_lock); // Might GC
    return (void*)data;
}


static void emit_result(std::vector<NewArchiveMember> &Archive, SmallVectorImpl<char> &OS,
        StringRef Name, std::vector<std::string> &outputs)
{
    outputs.push_back({ OS.data(), OS.size() });
    Archive.push_back(NewArchiveMember(MemoryBufferRef(outputs.back(), Name)));
    OS.clear();
}

static object::Archive::Kind getDefaultForHost(Triple &triple)
{
      if (triple.isOSDarwin())
          return object::Archive::K_DARWIN;
      return object::Archive::K_GNU;
}

typedef Error ArchiveWriterError;
static void reportWriterError(const ErrorInfoBase &E)
{
    std::string err = E.message();
    jl_safe_printf("ERROR: failed to emit output file %s\n", err.c_str());
}


// takes the running content that has collected in the shadow module and dump it to disk
// this builds the object file portion of the sysimage files for fast startup
extern "C"
void jl_dump_native(void *native_code,
        const char *bc_fname, const char *unopt_bc_fname, const char *obj_fname,
        const char *sysimg_data, size_t sysimg_len)
{
    JL_TIMING(NATIVE_DUMP);
    jl_native_code_desc_t *data = (jl_native_code_desc_t*)native_code;
    LLVMContext &Context = data->M->getContext();
    // We don't want to use MCJIT's target machine because
    // it uses the large code model and we may potentially
    // want less optimizations there.
    Triple TheTriple = Triple(jl_TargetMachine->getTargetTriple());
    // make sure to emit the native object format, even if FORCE_ELF was set in codegen
#if defined(_OS_WINDOWS_)
    TheTriple.setObjectFormat(Triple::COFF);
#elif defined(_OS_DARWIN_)
    TheTriple.setObjectFormat(Triple::MachO);
    TheTriple.setOS(llvm::Triple::MacOSX);
#endif
    std::unique_ptr<TargetMachine> TM(
        jl_TargetMachine->getTarget().createTargetMachine(
            TheTriple.getTriple(),
            jl_TargetMachine->getTargetCPU(),
            jl_TargetMachine->getTargetFeatureString(),
            jl_TargetMachine->Options,
#if defined(_OS_LINUX_) || defined(_OS_FREEBSD_)
            Reloc::PIC_,
#else
            Optional<Reloc::Model>(),
#endif
#if defined(_CPU_PPC_) || defined(_CPU_PPC64_)
            // On PPC the small model is limited to 16bit offsets
            CodeModel::Medium,
#else
            // Use small model so that we can use signed 32bits offset in the function and GV tables
            CodeModel::Small,
#endif
            CodeGenOpt::Aggressive // -O3 TODO: respect command -O0 flag?
            ));

    legacy::PassManager PM;
    addTargetPasses(&PM, TM.get());

    // set up optimization passes
    SmallVector<char, 128> bc_Buffer;
    SmallVector<char, 128> obj_Buffer;
    SmallVector<char, 128> unopt_bc_Buffer;
    raw_svector_ostream bc_OS(bc_Buffer);
    raw_svector_ostream obj_OS(obj_Buffer);
    raw_svector_ostream unopt_bc_OS(unopt_bc_Buffer);
    std::vector<NewArchiveMember> bc_Archive;
    std::vector<NewArchiveMember> obj_Archive;
    std::vector<NewArchiveMember> unopt_bc_Archive;
    std::vector<std::string> outputs;

    if (unopt_bc_fname)
        PM.add(createBitcodeWriterPass(unopt_bc_OS));
    if (bc_fname || obj_fname)
        addOptimizationPasses(&PM, jl_options.opt_level, true, true);
    if (bc_fname)
        PM.add(createBitcodeWriterPass(bc_OS));
    if (obj_fname)
        if (TM->addPassesToEmitFile(PM, obj_OS, nullptr, CGFT_ObjectFile, false))
            jl_safe_printf("ERROR: target does not support generation of object files\n");

    // Reset the target triple to make sure it matches the new target machine
    data->M->setTargetTriple(TM->getTargetTriple().str());
    DataLayout DL = TM->createDataLayout();
    DL.reset(DL.getStringRepresentation() + "-ni:10:11:12:13");
    data->M->setDataLayout(DL);
    Type *T_size;
    if (sizeof(size_t) == 8)
        T_size = Type::getInt64Ty(Context);
    else
        T_size = Type::getInt32Ty(Context);
    Type *T_psize = T_size->getPointerTo();

    // add metadata information
    if (imaging_mode) {
        emit_offset_table(*data->M, data->jl_sysimg_gvars, "jl_sysimg_gvars", T_psize);
        emit_offset_table(*data->M, data->jl_sysimg_fvars, "jl_sysimg_fvars", T_psize);

        // reflect the address of the jl_RTLD_DEFAULT_handle variable
        // back to the caller, so that we can check for consistency issues
        GlobalValue *jlRTLD_DEFAULT_var = data->M->getNamedValue("jl_RTLD_DEFAULT_handle");
        addComdat(new GlobalVariable(*data->M,
                                     jlRTLD_DEFAULT_var->getType(),
                                     true,
                                     GlobalVariable::ExternalLinkage,
                                     jlRTLD_DEFAULT_var,
                                     "jl_RTLD_DEFAULT_handle_pointer"));
    }

    // do the actual work
    auto add_output = [&] (Module &M, StringRef unopt_bc_Name, StringRef bc_Name, StringRef obj_Name) {
        PM.run(M);
        if (unopt_bc_fname)
            emit_result(unopt_bc_Archive, unopt_bc_Buffer, unopt_bc_Name, outputs);
        if (bc_fname)
            emit_result(bc_Archive, bc_Buffer, bc_Name, outputs);
        if (obj_fname)
            emit_result(obj_Archive, obj_Buffer, obj_Name, outputs);
    };

    add_output(*data->M, "unopt.bc", "text.bc", "text.o");

    std::unique_ptr<Module> sysimage(new Module("sysimage", Context));
    sysimage->setTargetTriple(data->M->getTargetTriple());
    sysimage->setDataLayout(data->M->getDataLayout());
    data->M.reset(); // free memory for data->M

    addComdat(new GlobalVariable(*sysimage,
                                 T_size,
                                 true,
                                 GlobalVariable::ExternalLinkage,
                                 ConstantInt::get(T_size, globalUnique + 1),
                                 "jl_globalUnique"));

    if (sysimg_data) {
        Constant *data = ConstantDataArray::get(Context,
            ArrayRef<uint8_t>((const unsigned char*)sysimg_data, sysimg_len));
        addComdat(new GlobalVariable(*sysimage, data->getType(), false,
                                     GlobalVariable::ExternalLinkage,
                                     data, "jl_system_image_data"))->setAlignment(Align(64));
        Constant *len = ConstantInt::get(T_size, sysimg_len);
        addComdat(new GlobalVariable(*sysimage, len->getType(), true,
                                     GlobalVariable::ExternalLinkage,
                                     len, "jl_system_image_size"));
    }
    add_output(*sysimage, "data.bc", "data.bc", "data.o");

    object::Archive::Kind Kind = getDefaultForHost(TheTriple);
    if (unopt_bc_fname)
        handleAllErrors(writeArchive(unopt_bc_fname, unopt_bc_Archive, true,
                    Kind, true, false), reportWriterError);
    if (bc_fname)
        handleAllErrors(writeArchive(bc_fname, bc_Archive, true,
                    Kind, true, false), reportWriterError);
    if (obj_fname)
        handleAllErrors(writeArchive(obj_fname, obj_Archive, true,
                    Kind, true, false), reportWriterError);

    delete data;
}

// clones the contents of the module `m` to the shadow_output collector
// TODO: this should be deprecated
void jl_add_to_shadow(Module *m)
{
    ValueToValueMapTy VMap;
    std::unique_ptr<Module> clone(CloneModule(*m, VMap));
    jl_merge_module(shadow_output, std::move(clone));
}


void addTargetPasses(legacy::PassManagerBase *PM, TargetMachine *TM)
{
    PM->add(new TargetLibraryInfoWrapperPass(Triple(TM->getTargetTriple())));
    PM->add(createTargetTransformInfoWrapperPass(TM->getTargetIRAnalysis()));
}

// this defines the set of optimization passes defined for Julia at various optimization levels.
// it assumes that the TLI and TTI wrapper passes have already been added.
void addOptimizationPasses(legacy::PassManagerBase *PM, int opt_level,
                           bool lower_intrinsics, bool dump_native)
{
#ifdef JL_DEBUG_BUILD
    PM->add(createGCInvariantVerifierPass(true));
    PM->add(createVerifierPass());
#endif

#if defined(JL_ASAN_ENABLED)
    PM->add(createAddressSanitizerFunctionPass());
#endif
#if defined(JL_MSAN_ENABLED)
    PM->add(llvm::createMemorySanitizerPass(true));
#endif
    if (opt_level < 2) {
        PM->add(createCFGSimplificationPass());
        if (opt_level == 1) {
            PM->add(createSROAPass());
            PM->add(createInstructionCombiningPass());
            PM->add(createEarlyCSEPass());
            // maybe add GVN?
            // also try GVNHoist and GVNSink
        }
        PM->add(createMemCpyOptPass());
        PM->add(createAlwaysInlinerLegacyPass()); // Respect always_inline
        if (lower_intrinsics) {
            PM->add(createBarrierNoopPass());
            PM->add(createLowerExcHandlersPass());
            PM->add(createGCInvariantVerifierPass(false));
            PM->add(createLateLowerGCFramePass());
            PM->add(createFinalLowerGCPass());
            PM->add(createLowerPTLSPass(dump_native));
        }
        PM->add(createLowerSimdLoopPass()); // Annotate loop marked with "loopinfo" as LLVM parallel loop
        if (dump_native)
            PM->add(createMultiVersioningPass());
        return;
    }
    PM->add(createPropagateJuliaAddrspaces());
    PM->add(createScopedNoAliasAAWrapperPass());
    PM->add(createTypeBasedAAWrapperPass());
    if (opt_level >= 3) {
        PM->add(createBasicAAWrapperPass());
    }

    PM->add(createCFGSimplificationPass());
    PM->add(createDeadCodeEliminationPass());
    PM->add(createSROAPass());

    //PM->add(createMemCpyOptPass());

    PM->add(createAlwaysInlinerLegacyPass()); // Respect always_inline

    // Running `memcpyopt` between this and `sroa` seems to give `sroa` a hard time
    // merging the `alloca` for the unboxed data and the `alloca` created by the `alloc_opt`
    // pass.
    PM->add(createAllocOptPass());
    // consider AggressiveInstCombinePass at optlevel > 2
    PM->add(createInstructionCombiningPass());
    PM->add(createCFGSimplificationPass());
    if (dump_native)
        PM->add(createMultiVersioningPass());
    PM->add(createSROAPass());
    PM->add(createInstSimplifyLegacyPass());
    PM->add(createJumpThreadingPass());

    PM->add(createReassociatePass());

    PM->add(createEarlyCSEPass());

    // Load forwarding above can expose allocations that aren't actually used
    // remove those before optimizing loops.
    PM->add(createAllocOptPass());
    PM->add(createLoopRotatePass());
    // moving IndVarSimplify here prevented removing the loop in perf_sumcartesian(10:-1:1)
    PM->add(createLoopIdiomPass());
#ifdef USE_POLLY
    // LCSSA (which has already run at this point due to the dependencies of the
    // above passes) introduces redundant phis that hinder Polly. Therefore we
    // run InstCombine here to remove them.
    PM->add(createInstructionCombiningPass());
    PM->add(polly::createCodePreparationPass());
    polly::registerPollyPasses(*PM);
    PM->add(polly::createCodegenCleanupPass());
#endif
    // LoopRotate strips metadata from terminator, so run LowerSIMD afterwards
    PM->add(createLowerSimdLoopPass()); // Annotate loop marked with "loopinfo" as LLVM parallel loop
    PM->add(createLICMPass());
    PM->add(createLoopUnswitchPass());
    // Subsequent passes not stripping metadata from terminator
    PM->add(createInstSimplifyLegacyPass());
    PM->add(createIndVarSimplifyPass());
    PM->add(createLoopDeletionPass());
    PM->add(createSimpleLoopUnrollPass());

    // Run our own SROA on heap objects before LLVM's
    PM->add(createAllocOptPass());
    // Re-run SROA after loop-unrolling (useful for small loops that operate,
    // over the structure of an aggregate)
    PM->add(createSROAPass());
    // might not be necessary:
    PM->add(createInstSimplifyLegacyPass());

    PM->add(createGVNPass());
    PM->add(createMemCpyOptPass());
    PM->add(createSCCPPass());

    // Run instcombine after redundancy elimination to exploit opportunities
    // opened up by them.
    // This needs to be InstCombine instead of InstSimplify to allow
    // loops over Union-typed arrays to vectorize.
    PM->add(createInstructionCombiningPass());
    PM->add(createJumpThreadingPass());
    PM->add(createDeadStoreEliminationPass());

    // More dead allocation (store) deletion before loop optimization
    // consider removing this:
    PM->add(createAllocOptPass());
    // see if all of the constant folding has exposed more loops
    // to simplification and deletion
    // this helps significantly with cleaning up iteration
    PM->add(createCFGSimplificationPass());
    PM->add(createLoopDeletionPass());
    PM->add(createInstructionCombiningPass());
    PM->add(createLoopVectorizePass());
    PM->add(createLoopLoadEliminationPass());
    PM->add(createCFGSimplificationPass());
    PM->add(createSLPVectorizerPass());
    // might need this after LLVM 11:
    //PM->add(createVectorCombinePass());

    PM->add(createAggressiveDCEPass());

    if (lower_intrinsics) {
        // LowerPTLS removes an indirect call. As a result, it is likely to trigger
        // LLVM's devirtualization heuristics, which would result in the entire
        // pass pipeline being re-exectuted. Prevent this by inserting a barrier.
        PM->add(createBarrierNoopPass());
        PM->add(createLowerExcHandlersPass());
        PM->add(createGCInvariantVerifierPass(false));
        PM->add(createLateLowerGCFramePass());
        PM->add(createFinalLowerGCPass());
        // Remove dead use of ptls
        PM->add(createDeadCodeEliminationPass());
        PM->add(createLowerPTLSPass(dump_native));
        // Clean up write barrier and ptls lowering
        PM->add(createCFGSimplificationPass());
    }
    PM->add(createCombineMulAddPass());
    PM->add(createDivRemPairsPass());
}

// An LLVM module pass that just runs all julia passes in order. Useful for
// debugging
template <int OptLevel>
class JuliaPipeline : public Pass {
public:
    static char ID;
    // A bit of a hack, but works
    struct TPMAdapter : public PassManagerBase {
        PMTopLevelManager *TPM;
        TPMAdapter(PMTopLevelManager *TPM) : TPM(TPM) {}
        void add(Pass *P) { TPM->schedulePass(P); }
    };
    void preparePassManager(PMStack &Stack) override {
        (void)jl_init_llvm();
        PMTopLevelManager *TPM = Stack.top()->getTopLevelManager();
        TPMAdapter Adapter(TPM);
        addTargetPasses(&Adapter, jl_TargetMachine);
        addOptimizationPasses(&Adapter, OptLevel);
    }
    JuliaPipeline() : Pass(PT_PassManager, ID) {}
    Pass *createPrinterPass(raw_ostream &O, const std::string &Banner) const override {
        return createPrintModulePass(O, Banner);
    }
};
template<> char JuliaPipeline<0>::ID = 0;
template<> char JuliaPipeline<2>::ID = 0;
template<> char JuliaPipeline<3>::ID = 0;
static RegisterPass<JuliaPipeline<0>> X("juliaO0", "Runs the entire julia pipeline (at -O0)", false, false);
static RegisterPass<JuliaPipeline<2>> Y("julia", "Runs the entire julia pipeline (at -O2)", false, false);
static RegisterPass<JuliaPipeline<3>> Z("juliaO3", "Runs the entire julia pipeline (at -O3)", false, false);

extern "C" JL_DLLEXPORT
void jl_add_optimization_passes(LLVMPassManagerRef PM, int opt_level, int lower_intrinsics) {
    addOptimizationPasses(unwrap(PM), opt_level, lower_intrinsics);
}

// --- native code info, and dump function to IR and ASM ---
// Get pointer to llvm::Function instance, compiling if necessary
// for use in reflection from Julia.
// this is paired with jl_dump_function_ir, jl_dump_method_asm, jl_dump_llvm_asm in particular ways:
// misuse will leak memory or cause read-after-free
extern "C" JL_DLLEXPORT
void *jl_get_llvmf_defn(jl_method_instance_t *mi, size_t world, char getwrapper, char optimize, const jl_cgparams_t params)
{
    if (jl_is_method(mi->def.method) && mi->def.method->source == NULL &&
            mi->def.method->generator == NULL) {
        // not a generic function
        return NULL;
    }

    static legacy::PassManager *PM;
    if (!PM) {
        PM = new legacy::PassManager();
        addTargetPasses(PM, jl_TargetMachine);
        addOptimizationPasses(PM, jl_options.opt_level);
    }

    // get the source code for this function
    jl_value_t *jlrettype = (jl_value_t*)jl_any_type;
    jl_code_info_t *src = NULL;
    JL_GC_PUSH2(&src, &jlrettype);
    jl_value_t *ci = jl_rettype_inferred(mi, world, world);
    if (ci != jl_nothing) {
        jl_code_instance_t *codeinst = (jl_code_instance_t*)ci;
        src = (jl_code_info_t*)codeinst->inferred;
        if ((jl_value_t*)src != jl_nothing && !jl_is_code_info(src) && jl_is_method(mi->def.method))
            src = jl_uncompress_ir(mi->def.method, codeinst, (jl_array_t*)src);
        jlrettype = codeinst->rettype;
    }
    if (!src || (jl_value_t*)src == jl_nothing) {
        src = jl_type_infer(mi, world, 0);
        if (src)
            jlrettype = src->rettype;
        else if (jl_is_method(mi->def.method)) {
            src = mi->def.method->generator ? jl_code_for_staged(mi) : (jl_code_info_t*)mi->def.method->source;
            if (src && !jl_is_code_info(src) && jl_is_method(mi->def.method))
                src = jl_uncompress_ir(mi->def.method, NULL, (jl_array_t*)src);
        }
        // TODO: use mi->uninferred
    }

    // emit this function into a new llvm module
    if (src && jl_is_code_info(src)) {
        jl_codegen_params_t output;
        output.world = world;
        output.params = &params;
        std::unique_ptr<Module> m;
        jl_llvm_functions_t decls;
        JL_LOCK(&codegen_lock);
        std::tie(m, decls) = jl_emit_code(mi, src, jlrettype, output);

        Function *F = NULL;
        if (m) {
            // if compilation succeeded, prepare to return the result
            if (optimize)
                PM->run(*m.get());
            const std::string *fname;
            if (decls.functionObject == "jl_fptr_args" || decls.functionObject == "jl_fptr_sparam")
                getwrapper = false;
            if (!getwrapper)
                fname = &decls.specFunctionObject;
            else
                fname = &decls.functionObject;
            F = cast<Function>(m->getNamedValue(*fname));
            m.release(); // the return object `llvmf` will be the owning pointer
        }
        JL_GC_POP();
        JL_UNLOCK(&codegen_lock); // Might GC
        if (F)
            return F;
    }

    const char *mname = name_from_method_instance(mi);
    jl_errorf("unable to compile source for function %s", mname);
}

/// addPassesToX helper drives creation and initialization of TargetPassConfig.
static MCContext *
addPassesToGenerateCode(LLVMTargetMachine *TM, PassManagerBase &PM) {
    TargetPassConfig *PassConfig = TM->createPassConfig(PM);
    PassConfig->setDisableVerify(false);
    PM.add(PassConfig);
#if JL_LLVM_VERSION >= 100000
    MachineModuleInfoWrapperPass *MMIWP =
        new MachineModuleInfoWrapperPass(TM);
    PM.add(MMIWP);
#else
    MachineModuleInfo *MMI = new MachineModuleInfo(TM);
    PM.add(MMI);
#endif
    if (PassConfig->addISelPasses())
        return NULL;
    PassConfig->addMachinePasses();
    PassConfig->setInitialized();
#if JL_LLVM_VERSION >= 100000
    return &MMIWP->getMMI().getContext();
#else
    return &MMI->getContext();
#endif
}

void jl_strip_llvm_debug(Module *m);


// get a native assembly for llvm::Function
// TODO: implement debuginfo handling
extern "C" JL_DLLEXPORT
jl_value_t *jl_dump_llvm_asm(void *F, const char* asm_variant, const char *debuginfo)
{
    // precise printing via IR assembler
    SmallVector<char, 4096> ObjBufferSV;
    { // scope block
        Function *f = (Function*)F;
        llvm::raw_svector_ostream asmfile(ObjBufferSV);
        assert(!f->isDeclaration());
        std::unique_ptr<Module> m(f->getParent());
        for (auto &f2 : m->functions()) {
            if (f != &f2 && !f->isDeclaration())
                f2.deleteBody();
        }
        jl_strip_llvm_debug(m.get());
        legacy::PassManager PM;
        LLVMTargetMachine *TM = static_cast<LLVMTargetMachine*>(jl_TargetMachine);
        MCContext *Context = addPassesToGenerateCode(TM, PM);
        if (Context) {
            const MCSubtargetInfo &STI = *TM->getMCSubtargetInfo();
            const MCAsmInfo &MAI = *TM->getMCAsmInfo();
            const MCRegisterInfo &MRI = *TM->getMCRegisterInfo();
            const MCInstrInfo &MII = *TM->getMCInstrInfo();
            unsigned OutputAsmDialect = MAI.getAssemblerDialect();
            if (!strcmp(asm_variant, "att"))
                OutputAsmDialect = 0;
            if (!strcmp(asm_variant, "intel"))
                OutputAsmDialect = 1;
            MCInstPrinter *InstPrinter = TM->getTarget().createMCInstPrinter(
                TM->getTargetTriple(), OutputAsmDialect, MAI, MII, MRI);
             std::unique_ptr<MCAsmBackend> MAB(TM->getTarget().createMCAsmBackend(
                STI, MRI, TM->Options.MCOptions));
            std::unique_ptr<MCCodeEmitter> MCE;
#if JL_LLVM_VERSION >= 100000
            auto FOut = std::make_unique<formatted_raw_ostream>(asmfile);
#else
            auto FOut = llvm::make_unique<formatted_raw_ostream>(asmfile);
#endif
            std::unique_ptr<MCStreamer> S(TM->getTarget().createAsmStreamer(
                *Context, std::move(FOut), true,
                true, InstPrinter,
                std::move(MCE), std::move(MAB),
                false));
            std::unique_ptr<AsmPrinter> Printer(
                TM->getTarget().createAsmPrinter(*TM, std::move(S)));
            if (Printer) {
                PM.add(Printer.release());
                PM.run(*m);
            }
        }
    }
    return jl_pchar_to_string(ObjBufferSV.data(), ObjBufferSV.size());
}