https://github.com/JuliaLang/julia
Tip revision: 5d1ca6a19fdc8f961c0ce8a3745e3a5d4cb5f90b authored by Rafael Fourquet on 11 July 2023, 16:00:54 UTC
speed-up `randperm` by using our current `rand(1:n)`
speed-up `randperm` by using our current `rand(1:n)`
Tip revision: 5d1ca6a
llvm-final-gc-lowering.cpp
// This file is a part of Julia. License is MIT: https://julialang.org/license
#include "llvm-version.h"
#include "passes.h"
#include <llvm/ADT/Statistic.h>
#include <llvm/IR/LegacyPassManager.h>
#include <llvm/IR/Function.h>
#include <llvm/IR/IntrinsicInst.h>
#include <llvm/IR/Module.h>
#include <llvm/IR/IRBuilder.h>
#include <llvm/IR/Verifier.h>
#include <llvm/Pass.h>
#include <llvm/Support/Debug.h>
#include <llvm/Transforms/Utils/ModuleUtils.h>
#include "llvm-codegen-shared.h"
#include "julia.h"
#include "julia_internal.h"
#include "llvm-pass-helpers.h"
#define DEBUG_TYPE "final_gc_lowering"
STATISTIC(NewGCFrameCount, "Number of lowered newGCFrameFunc intrinsics");
STATISTIC(PushGCFrameCount, "Number of lowered pushGCFrameFunc intrinsics");
STATISTIC(PopGCFrameCount, "Number of lowered popGCFrameFunc intrinsics");
STATISTIC(GetGCFrameSlotCount, "Number of lowered getGCFrameSlotFunc intrinsics");
STATISTIC(GCAllocBytesCount, "Number of lowered GCAllocBytesFunc intrinsics");
STATISTIC(QueueGCRootCount, "Number of lowered queueGCRootFunc intrinsics");
STATISTIC(SafepointCount, "Number of lowered safepoint intrinsics");
using namespace llvm;
// The final GC lowering pass. This pass lowers platform-agnostic GC
// intrinsics to platform-dependent instruction sequences. The
// intrinsics it targets are those produced by the late GC frame
// lowering pass.
//
// This pass targets typical back-ends for which the standard Julia
// runtime library is available. Atypical back-ends should supply
// their own lowering pass.
struct FinalLowerGC: private JuliaPassContext {
bool runOnFunction(Function &F);
bool doInitialization(Module &M);
bool doFinalization(Module &M);
private:
Function *queueRootFunc;
Function *poolAllocFunc;
Function *bigAllocFunc;
Function *allocTypedFunc;
Instruction *pgcstack;
Type *T_size;
// Lowers a `julia.new_gc_frame` intrinsic.
Value *lowerNewGCFrame(CallInst *target, Function &F);
// Lowers a `julia.push_gc_frame` intrinsic.
void lowerPushGCFrame(CallInst *target, Function &F);
// Lowers a `julia.pop_gc_frame` intrinsic.
void lowerPopGCFrame(CallInst *target, Function &F);
// Lowers a `julia.get_gc_frame_slot` intrinsic.
Value *lowerGetGCFrameSlot(CallInst *target, Function &F);
// Lowers a `julia.gc_alloc_bytes` intrinsic.
Value *lowerGCAllocBytes(CallInst *target, Function &F);
// Lowers a `julia.queue_gc_root` intrinsic.
Value *lowerQueueGCRoot(CallInst *target, Function &F);
// Lowers a `julia.safepoint` intrinsic.
Value *lowerSafepoint(CallInst *target, Function &F);
};
Value *FinalLowerGC::lowerNewGCFrame(CallInst *target, Function &F)
{
++NewGCFrameCount;
assert(target->arg_size() == 1);
unsigned nRoots = cast<ConstantInt>(target->getArgOperand(0))->getLimitedValue(INT_MAX);
// Create the GC frame.
IRBuilder<> builder(target->getNextNode());
auto gcframe_alloca = builder.CreateAlloca(T_prjlvalue, ConstantInt::get(Type::getInt32Ty(F.getContext()), nRoots + 2));
gcframe_alloca->setAlignment(Align(16));
// addrspacecast as needed for non-0 alloca addrspace
auto gcframe = cast<Instruction>(builder.CreateAddrSpaceCast(gcframe_alloca, T_prjlvalue->getPointerTo(0)));
gcframe->takeName(target);
// Zero out the GC frame.
auto ptrsize = F.getParent()->getDataLayout().getPointerSize();
builder.CreateMemSet(gcframe, Constant::getNullValue(Type::getInt8Ty(F.getContext())), ptrsize * (nRoots + 2), Align(16), tbaa_gcframe);
return gcframe;
}
void FinalLowerGC::lowerPushGCFrame(CallInst *target, Function &F)
{
++PushGCFrameCount;
assert(target->arg_size() == 2);
auto gcframe = target->getArgOperand(0);
unsigned nRoots = cast<ConstantInt>(target->getArgOperand(1))->getLimitedValue(INT_MAX);
IRBuilder<> builder(target->getContext());
builder.SetInsertPoint(&*(++BasicBlock::iterator(target)));
StoreInst *inst = builder.CreateAlignedStore(
ConstantInt::get(T_size, JL_GC_ENCODE_PUSHARGS(nRoots)),
builder.CreateBitCast(
builder.CreateConstInBoundsGEP1_32(T_prjlvalue, gcframe, 0),
T_size->getPointerTo()),
Align(sizeof(void*)));
inst->setMetadata(LLVMContext::MD_tbaa, tbaa_gcframe);
auto T_ppjlvalue = JuliaType::get_ppjlvalue_ty(F.getContext());
inst = builder.CreateAlignedStore(
builder.CreateAlignedLoad(T_ppjlvalue, pgcstack, Align(sizeof(void*))),
builder.CreatePointerCast(
builder.CreateConstInBoundsGEP1_32(T_prjlvalue, gcframe, 1),
PointerType::get(T_ppjlvalue, 0)),
Align(sizeof(void*)));
inst->setMetadata(LLVMContext::MD_tbaa, tbaa_gcframe);
inst = builder.CreateAlignedStore(
gcframe,
builder.CreateBitCast(pgcstack, PointerType::get(PointerType::get(T_prjlvalue, 0), 0)),
Align(sizeof(void*)));
}
void FinalLowerGC::lowerPopGCFrame(CallInst *target, Function &F)
{
++PopGCFrameCount;
assert(target->arg_size() == 1);
auto gcframe = target->getArgOperand(0);
IRBuilder<> builder(target->getContext());
builder.SetInsertPoint(target);
Instruction *gcpop =
cast<Instruction>(builder.CreateConstInBoundsGEP1_32(T_prjlvalue, gcframe, 1));
Instruction *inst = builder.CreateAlignedLoad(T_prjlvalue, gcpop, Align(sizeof(void*)));
inst->setMetadata(LLVMContext::MD_tbaa, tbaa_gcframe);
inst = builder.CreateAlignedStore(
inst,
builder.CreateBitCast(pgcstack,
PointerType::get(T_prjlvalue, 0)),
Align(sizeof(void*)));
inst->setMetadata(LLVMContext::MD_tbaa, tbaa_gcframe);
}
Value *FinalLowerGC::lowerGetGCFrameSlot(CallInst *target, Function &F)
{
++GetGCFrameSlotCount;
assert(target->arg_size() == 2);
auto gcframe = target->getArgOperand(0);
auto index = target->getArgOperand(1);
// Initialize an IR builder.
IRBuilder<> builder(target->getContext());
builder.SetInsertPoint(target);
// The first two slots are reserved, so we'll add two to the index.
index = builder.CreateAdd(index, ConstantInt::get(Type::getInt32Ty(F.getContext()), 2));
// Lower the intrinsic as a GEP.
auto gep = builder.CreateInBoundsGEP(T_prjlvalue, gcframe, index);
gep->takeName(target);
return gep;
}
Value *FinalLowerGC::lowerQueueGCRoot(CallInst *target, Function &F)
{
++QueueGCRootCount;
assert(target->arg_size() == 1);
target->setCalledFunction(queueRootFunc);
return target;
}
Value *FinalLowerGC::lowerSafepoint(CallInst *target, Function &F)
{
++SafepointCount;
assert(target->arg_size() == 1);
IRBuilder<> builder(target->getContext());
builder.SetInsertPoint(target);
Value* signal_page = target->getOperand(0);
Value* load = builder.CreateLoad(T_size, signal_page, true);
return load;
}
Value *FinalLowerGC::lowerGCAllocBytes(CallInst *target, Function &F)
{
++GCAllocBytesCount;
assert(target->arg_size() == 2);
CallInst *newI;
IRBuilder<> builder(target);
builder.SetCurrentDebugLocation(target->getDebugLoc());
auto ptls = target->getArgOperand(0);
Attribute derefAttr;
if (auto CI = dyn_cast<ConstantInt>(target->getArgOperand(1))) {
size_t sz = (size_t)CI->getZExtValue();
// This is strongly architecture and OS dependent
int osize;
int offset = jl_gc_classify_pools(sz, &osize);
if (offset < 0) {
newI = builder.CreateCall(
bigAllocFunc,
{ ptls, ConstantInt::get(T_size, sz + sizeof(void*)) });
derefAttr = Attribute::getWithDereferenceableBytes(F.getContext(), sz + sizeof(void*));
}
else {
auto pool_offs = ConstantInt::get(Type::getInt32Ty(F.getContext()), offset);
auto pool_osize = ConstantInt::get(Type::getInt32Ty(F.getContext()), osize);
newI = builder.CreateCall(poolAllocFunc, { ptls, pool_offs, pool_osize });
derefAttr = Attribute::getWithDereferenceableBytes(F.getContext(), osize);
}
} else {
auto size = builder.CreateZExtOrTrunc(target->getArgOperand(1), T_size);
size = builder.CreateAdd(size, ConstantInt::get(T_size, sizeof(void*)));
newI = builder.CreateCall(allocTypedFunc, { ptls, size, ConstantPointerNull::get(Type::getInt8PtrTy(F.getContext())) });
derefAttr = Attribute::getWithDereferenceableBytes(F.getContext(), sizeof(void*));
}
newI->setAttributes(newI->getCalledFunction()->getAttributes());
newI->addRetAttr(derefAttr);
newI->takeName(target);
return newI;
}
bool FinalLowerGC::doInitialization(Module &M) {
// Initialize platform-agnostic references.
initAll(M);
// Initialize platform-specific references.
queueRootFunc = getOrDeclare(jl_well_known::GCQueueRoot);
poolAllocFunc = getOrDeclare(jl_well_known::GCPoolAlloc);
bigAllocFunc = getOrDeclare(jl_well_known::GCBigAlloc);
allocTypedFunc = getOrDeclare(jl_well_known::GCAllocTyped);
T_size = M.getDataLayout().getIntPtrType(M.getContext());
GlobalValue *functionList[] = {queueRootFunc, poolAllocFunc, bigAllocFunc, allocTypedFunc};
unsigned j = 0;
for (unsigned i = 0; i < sizeof(functionList) / sizeof(void*); i++) {
if (!functionList[i])
continue;
if (i != j)
functionList[j] = functionList[i];
j++;
}
if (j != 0)
appendToCompilerUsed(M, ArrayRef<GlobalValue*>(functionList, j));
return true;
}
bool FinalLowerGC::doFinalization(Module &M)
{
GlobalValue *functionList[] = {queueRootFunc, poolAllocFunc, bigAllocFunc, allocTypedFunc};
queueRootFunc = poolAllocFunc = bigAllocFunc = allocTypedFunc = nullptr;
auto used = M.getGlobalVariable("llvm.compiler.used");
if (!used)
return false;
SmallPtrSet<Constant*, 16> InitAsSet(
functionList,
functionList + sizeof(functionList) / sizeof(void*));
bool changed = false;
SmallVector<Constant*, 16> init;
ConstantArray *CA = cast<ConstantArray>(used->getInitializer());
for (auto &Op : CA->operands()) {
Constant *C = cast_or_null<Constant>(Op);
if (InitAsSet.count(C->stripPointerCasts())) {
changed = true;
continue;
}
init.push_back(C);
}
if (!changed)
return false;
used->eraseFromParent();
if (init.empty())
return true;
ArrayType *ATy = ArrayType::get(Type::getInt8PtrTy(M.getContext()), init.size());
used = new GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
ConstantArray::get(ATy, init), "llvm.compiler.used");
used->setSection("llvm.metadata");
return true;
}
template<typename TIterator>
static void replaceInstruction(
Instruction *oldInstruction,
Value *newInstruction,
TIterator &it)
{
if (newInstruction != oldInstruction) {
oldInstruction->replaceAllUsesWith(newInstruction);
it = oldInstruction->eraseFromParent();
}
else {
++it;
}
}
bool FinalLowerGC::runOnFunction(Function &F)
{
// Check availability of functions again since they might have been deleted.
initFunctions(*F.getParent());
if (!pgcstack_getter && !adoptthread_func) {
LLVM_DEBUG(dbgs() << "FINAL GC LOWERING: Skipping function " << F.getName() << "\n");
return false;
}
// Look for a call to 'julia.get_pgcstack'.
pgcstack = getPGCstack(F);
if (!pgcstack) {
LLVM_DEBUG(dbgs() << "FINAL GC LOWERING: Skipping function " << F.getName() << " no pgcstack\n");
return false;
}
LLVM_DEBUG(dbgs() << "FINAL GC LOWERING: Processing function " << F.getName() << "\n");
// Acquire intrinsic functions.
auto newGCFrameFunc = getOrNull(jl_intrinsics::newGCFrame);
auto pushGCFrameFunc = getOrNull(jl_intrinsics::pushGCFrame);
auto popGCFrameFunc = getOrNull(jl_intrinsics::popGCFrame);
auto getGCFrameSlotFunc = getOrNull(jl_intrinsics::getGCFrameSlot);
auto GCAllocBytesFunc = getOrNull(jl_intrinsics::GCAllocBytes);
auto queueGCRootFunc = getOrNull(jl_intrinsics::queueGCRoot);
auto safepointFunc = getOrNull(jl_intrinsics::safepoint);
// Lower all calls to supported intrinsics.
for (BasicBlock &BB : F) {
for (auto it = BB.begin(); it != BB.end();) {
auto *CI = dyn_cast<CallInst>(&*it);
if (!CI) {
++it;
continue;
}
Value *callee = CI->getCalledOperand();
assert(callee);
if (callee == newGCFrameFunc) {
replaceInstruction(CI, lowerNewGCFrame(CI, F), it);
}
else if (callee == pushGCFrameFunc) {
lowerPushGCFrame(CI, F);
it = CI->eraseFromParent();
}
else if (callee == popGCFrameFunc) {
lowerPopGCFrame(CI, F);
it = CI->eraseFromParent();
}
else if (callee == getGCFrameSlotFunc) {
replaceInstruction(CI, lowerGetGCFrameSlot(CI, F), it);
}
else if (callee == GCAllocBytesFunc) {
replaceInstruction(CI, lowerGCAllocBytes(CI, F), it);
}
else if (callee == queueGCRootFunc) {
replaceInstruction(CI, lowerQueueGCRoot(CI, F), it);
}
else if (callee == safepointFunc) {
lowerSafepoint(CI, F);
it = CI->eraseFromParent();
}
else {
++it;
}
}
}
return true;
}
struct FinalLowerGCLegacy: public FunctionPass {
static char ID;
FinalLowerGCLegacy() : FunctionPass(ID), finalLowerGC(FinalLowerGC()) {}
protected:
void getAnalysisUsage(AnalysisUsage &AU) const override {
FunctionPass::getAnalysisUsage(AU);
}
private:
bool runOnFunction(Function &F) override;
bool doInitialization(Module &M) override;
bool doFinalization(Module &M) override;
FinalLowerGC finalLowerGC;
};
bool FinalLowerGCLegacy::runOnFunction(Function &F) {
return finalLowerGC.runOnFunction(F);
}
bool FinalLowerGCLegacy::doInitialization(Module &M) {
return finalLowerGC.doInitialization(M);
}
bool FinalLowerGCLegacy::doFinalization(Module &M) {
auto ret = finalLowerGC.doFinalization(M);
#ifdef JL_VERIFY_PASSES
assert(!verifyModule(M, &errs()));
#endif
return ret;
}
PreservedAnalyses FinalLowerGCPass::run(Module &M, ModuleAnalysisManager &AM)
{
auto finalLowerGC = FinalLowerGC();
bool modified = false;
modified |= finalLowerGC.doInitialization(M);
for (auto &F : M.functions()) {
if (F.isDeclaration())
continue;
modified |= finalLowerGC.runOnFunction(F);
}
modified |= finalLowerGC.doFinalization(M);
#ifdef JL_VERIFY_PASSES
assert(!verifyModule(M, &errs()));
#endif
if (modified) {
return PreservedAnalyses::allInSet<CFGAnalyses>();
}
return PreservedAnalyses::all();
}
char FinalLowerGCLegacy::ID = 0;
static RegisterPass<FinalLowerGCLegacy> X("FinalLowerGC", "Final GC intrinsic lowering pass", false, false);
Pass *createFinalLowerGCPass()
{
return new FinalLowerGCLegacy();
}
extern "C" JL_DLLEXPORT_CODEGEN
void LLVMExtraAddFinalLowerGCPass_impl(LLVMPassManagerRef PM)
{
unwrap(PM)->add(createFinalLowerGCPass());
}