blob: 6f9475da8c0ae8029f51312884ebb3ec875dcbd3 [file] [log] [blame]
//===- CodeGenPrepare.cpp - Prepare a function for code generation --------===//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
// This pass munges the code in the input function to better prepare it for
// SelectionDAG-based code generation. This works around limitations in it's
// basic-block-at-a-time approach. It should eventually be removed.
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/DenseMap.h"
#include "llvm/ADT/MapVector.h"
#include "llvm/ADT/PointerIntPair.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/BlockFrequencyInfo.h"
#include "llvm/Analysis/BranchProbabilityInfo.h"
#include "llvm/Analysis/ConstantFolding.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/MemoryBuiltins.h"
#include "llvm/Analysis/ProfileSummaryInfo.h"
#include "llvm/Analysis/TargetLibraryInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/Analysis/VectorUtils.h"
#include "llvm/CodeGen/Analysis.h"
#include "llvm/CodeGen/ISDOpcodes.h"
#include "llvm/CodeGen/SelectionDAGNodes.h"
#include "llvm/CodeGen/TargetLowering.h"
#include "llvm/CodeGen/TargetPassConfig.h"
#include "llvm/CodeGen/TargetSubtargetInfo.h"
#include "llvm/CodeGen/ValueTypes.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/IR/Argument.h"
#include "llvm/IR/Attributes.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/Constant.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalValue.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/InlineAsm.h"
#include "llvm/IR/InstrTypes.h"
#include "llvm/IR/Instruction.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/IntrinsicsAArch64.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/MDBuilder.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/Operator.h"
#include "llvm/IR/PatternMatch.h"
#include "llvm/IR/Statepoint.h"
#include "llvm/IR/Type.h"
#include "llvm/IR/Use.h"
#include "llvm/IR/User.h"
#include "llvm/IR/Value.h"
#include "llvm/IR/ValueHandle.h"
#include "llvm/IR/ValueMap.h"
#include "llvm/InitializePasses.h"
#include "llvm/Pass.h"
#include "llvm/Support/BlockFrequency.h"
#include "llvm/Support/BranchProbability.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Compiler.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/MachineValueType.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetOptions.h"
#include "llvm/Transforms/Utils/BasicBlockUtils.h"
#include "llvm/Transforms/Utils/BypassSlowDivision.h"
#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Transforms/Utils/SimplifyLibCalls.h"
#include "llvm/Transforms/Utils/SizeOpts.h"
#include <algorithm>
#include <cassert>
#include <cstdint>
#include <iterator>
#include <limits>
#include <memory>
#include <utility>
#include <vector>
using namespace llvm;
using namespace llvm::PatternMatch;
#define DEBUG_TYPE "codegenprepare"
STATISTIC(NumBlocksElim, "Number of blocks eliminated");
STATISTIC(NumPHIsElim, "Number of trivial PHIs eliminated");
STATISTIC(NumGEPsElim, "Number of GEPs converted to casts");
STATISTIC(NumCmpUses, "Number of uses of Cmp expressions replaced with uses of "
"sunken Cmps");
STATISTIC(NumCastUses, "Number of uses of Cast expressions replaced with uses "
"of sunken Casts");
STATISTIC(NumMemoryInsts, "Number of memory instructions whose address "
"computations were sunk");
"Number of phis created when address "
"computations were sunk to memory instructions");
"Number of select created when address "
"computations were sunk to memory instructions");
STATISTIC(NumExtsMoved, "Number of [s|z]ext instructions combined with loads");
STATISTIC(NumExtUses, "Number of uses of [s|z]ext instructions optimized");
"Number of and mask instructions added to form ext loads");
STATISTIC(NumAndUses, "Number of uses of and mask instructions optimized");
STATISTIC(NumRetsDup, "Number of return instructions duplicated");
STATISTIC(NumDbgValueMoved, "Number of debug value instructions moved");
STATISTIC(NumSelectsExpanded, "Number of selects turned into branches");
STATISTIC(NumStoreExtractExposed, "Number of store(extractelement) exposed");
static cl::opt<bool> DisableBranchOpts(
"disable-cgp-branch-opts", cl::Hidden, cl::init(false),
cl::desc("Disable branch optimizations in CodeGenPrepare"));
static cl::opt<bool>
DisableGCOpts("disable-cgp-gc-opts", cl::Hidden, cl::init(false),
cl::desc("Disable GC optimizations in CodeGenPrepare"));
static cl::opt<bool> DisableSelectToBranch(
"disable-cgp-select2branch", cl::Hidden, cl::init(false),
cl::desc("Disable select to branch conversion."));
static cl::opt<bool> AddrSinkUsingGEPs(
"addr-sink-using-gep", cl::Hidden, cl::init(true),
cl::desc("Address sinking in CGP using GEPs."));
static cl::opt<bool> EnableAndCmpSinking(
"enable-andcmp-sinking", cl::Hidden, cl::init(true),
cl::desc("Enable sinkinig and/cmp into branches."));
static cl::opt<bool> DisableStoreExtract(
"disable-cgp-store-extract", cl::Hidden, cl::init(false),
cl::desc("Disable store(extract) optimizations in CodeGenPrepare"));
static cl::opt<bool> StressStoreExtract(
"stress-cgp-store-extract", cl::Hidden, cl::init(false),
cl::desc("Stress test store(extract) optimizations in CodeGenPrepare"));
static cl::opt<bool> DisableExtLdPromotion(
"disable-cgp-ext-ld-promotion", cl::Hidden, cl::init(false),
cl::desc("Disable ext(promotable(ld)) -> promoted(ext(ld)) optimization in "
static cl::opt<bool> StressExtLdPromotion(
"stress-cgp-ext-ld-promotion", cl::Hidden, cl::init(false),
cl::desc("Stress test ext(promotable(ld)) -> promoted(ext(ld)) "
"optimization in CodeGenPrepare"));
static cl::opt<bool> DisablePreheaderProtect(
"disable-preheader-prot", cl::Hidden, cl::init(false),
cl::desc("Disable protection against removing loop preheaders"));
static cl::opt<bool> ProfileGuidedSectionPrefix(
"profile-guided-section-prefix", cl::Hidden, cl::init(true), cl::ZeroOrMore,
cl::desc("Use profile info to add section prefix for hot/cold functions"));
static cl::opt<bool> ProfileUnknownInSpecialSection(
"profile-unknown-in-special-section", cl::Hidden, cl::init(false),
cl::desc("In profiling mode like sampleFDO, if a function doesn't have "
"profile, we cannot tell the function is cold for sure because "
"it may be a function newly added without ever being sampled. "
"With the flag enabled, compiler can put such profile unknown "
"functions into a special section, so runtime system can choose "
"to handle it in a different way than .text section, to save "
"RAM for example. "));
static cl::opt<unsigned> FreqRatioToSkipMerge(
"cgp-freq-ratio-to-skip-merge", cl::Hidden, cl::init(2),
cl::desc("Skip merging empty blocks if (frequency of empty block) / "
"(frequency of destination block) is greater than this ratio"));
static cl::opt<bool> ForceSplitStore(
"force-split-store", cl::Hidden, cl::init(false),
cl::desc("Force store splitting no matter what the target query says."));
static cl::opt<bool>
EnableTypePromotionMerge("cgp-type-promotion-merge", cl::Hidden,
cl::desc("Enable merging of redundant sexts when one is dominating"
" the other."), cl::init(true));
static cl::opt<bool> DisableComplexAddrModes(
"disable-complex-addr-modes", cl::Hidden, cl::init(false),
cl::desc("Disables combining addressing modes with different parts "
"in optimizeMemoryInst."));
static cl::opt<bool>
AddrSinkNewPhis("addr-sink-new-phis", cl::Hidden, cl::init(false),
cl::desc("Allow creation of Phis in Address sinking."));
static cl::opt<bool>
AddrSinkNewSelects("addr-sink-new-select", cl::Hidden, cl::init(true),
cl::desc("Allow creation of selects in Address sinking."));
static cl::opt<bool> AddrSinkCombineBaseReg(
"addr-sink-combine-base-reg", cl::Hidden, cl::init(true),
cl::desc("Allow combining of BaseReg field in Address sinking."));
static cl::opt<bool> AddrSinkCombineBaseGV(
"addr-sink-combine-base-gv", cl::Hidden, cl::init(true),
cl::desc("Allow combining of BaseGV field in Address sinking."));
static cl::opt<bool> AddrSinkCombineBaseOffs(
"addr-sink-combine-base-offs", cl::Hidden, cl::init(true),
cl::desc("Allow combining of BaseOffs field in Address sinking."));
static cl::opt<bool> AddrSinkCombineScaledReg(
"addr-sink-combine-scaled-reg", cl::Hidden, cl::init(true),
cl::desc("Allow combining of ScaledReg field in Address sinking."));
static cl::opt<bool>
EnableGEPOffsetSplit("cgp-split-large-offset-gep", cl::Hidden,
cl::desc("Enable splitting large offset of GEP."));
static cl::opt<bool> EnableICMP_EQToICMP_ST(
"cgp-icmp-eq2icmp-st", cl::Hidden, cl::init(false),
cl::desc("Enable ICMP_EQ to ICMP_S(L|G)T conversion."));
static cl::opt<bool>
VerifyBFIUpdates("cgp-verify-bfi-updates", cl::Hidden, cl::init(false),
cl::desc("Enable BFI update verification for "
static cl::opt<bool> OptimizePhiTypes(
"cgp-optimize-phi-types", cl::Hidden, cl::init(false),
cl::desc("Enable converting phi types in CodeGenPrepare"));
namespace {
enum ExtType {
ZeroExtension, // Zero extension has been seen.
SignExtension, // Sign extension has been seen.
BothExtension // This extension type is used if we saw sext after
// ZeroExtension had been set, or if we saw zext after
// SignExtension had been set. It makes the type
// information of a promoted instruction invalid.
using SetOfInstrs = SmallPtrSet<Instruction *, 16>;
using TypeIsSExt = PointerIntPair<Type *, 2, ExtType>;
using InstrToOrigTy = DenseMap<Instruction *, TypeIsSExt>;
using SExts = SmallVector<Instruction *, 16>;
using ValueToSExts = DenseMap<Value *, SExts>;
class TypePromotionTransaction;
class CodeGenPrepare : public FunctionPass {
const TargetMachine *TM = nullptr;
const TargetSubtargetInfo *SubtargetInfo;
const TargetLowering *TLI = nullptr;
const TargetRegisterInfo *TRI;
const TargetTransformInfo *TTI = nullptr;
const TargetLibraryInfo *TLInfo;
const LoopInfo *LI;
std::unique_ptr<BlockFrequencyInfo> BFI;
std::unique_ptr<BranchProbabilityInfo> BPI;
ProfileSummaryInfo *PSI;
/// As we scan instructions optimizing them, this is the next instruction
/// to optimize. Transforms that can invalidate this should update it.
BasicBlock::iterator CurInstIterator;
/// Keeps track of non-local addresses that have been sunk into a block.
/// This allows us to avoid inserting duplicate code for blocks with
/// multiple load/stores of the same address. The usage of WeakTrackingVH
/// enables SunkAddrs to be treated as a cache whose entries can be
/// invalidated if a sunken address computation has been erased.
ValueMap<Value*, WeakTrackingVH> SunkAddrs;
/// Keeps track of all instructions inserted for the current function.
SetOfInstrs InsertedInsts;
/// Keeps track of the type of the related instruction before their
/// promotion for the current function.
InstrToOrigTy PromotedInsts;
/// Keep track of instructions removed during promotion.
SetOfInstrs RemovedInsts;
/// Keep track of sext chains based on their initial value.
DenseMap<Value *, Instruction *> SeenChainsForSExt;
/// Keep track of GEPs accessing the same data structures such as structs or
/// arrays that are candidates to be split later because of their large
/// size.
SmallVector<std::pair<AssertingVH<GetElementPtrInst>, int64_t>, 32>>
/// Keep track of new GEP base after splitting the GEPs having large offset.
SmallSet<AssertingVH<Value>, 2> NewGEPBases;
/// Map serial numbers to Large offset GEPs.
DenseMap<AssertingVH<GetElementPtrInst>, int> LargeOffsetGEPID;
/// Keep track of SExt promoted.
ValueToSExts ValToSExtendedUses;
/// True if the function has the OptSize attribute.
bool OptSize;
/// DataLayout for the Function being processed.
const DataLayout *DL = nullptr;
/// Building the dominator tree can be expensive, so we only build it
/// lazily and update it when required.
std::unique_ptr<DominatorTree> DT;
static char ID; // Pass identification, replacement for typeid
CodeGenPrepare() : FunctionPass(ID) {
bool runOnFunction(Function &F) override;
StringRef getPassName() const override { return "CodeGen Prepare"; }
void getAnalysisUsage(AnalysisUsage &AU) const override {
// FIXME: When we can selectively preserve passes, preserve the domtree.
template <typename F>
void resetIteratorIfInvalidatedWhileCalling(BasicBlock *BB, F f) {
// Substituting can cause recursive simplifications, which can invalidate
// our iterator. Use a WeakTrackingVH to hold onto it in case this
// happens.
Value *CurValue = &*CurInstIterator;
WeakTrackingVH IterHandle(CurValue);
// If the iterator instruction was recursively deleted, start over at the
// start of the block.
if (IterHandle != CurValue) {
CurInstIterator = BB->begin();
// Get the DominatorTree, building if necessary.
DominatorTree &getDT(Function &F) {
if (!DT)
DT = std::make_unique<DominatorTree>(F);
return *DT;
void removeAllAssertingVHReferences(Value *V);
bool eliminateFallThrough(Function &F);
bool eliminateMostlyEmptyBlocks(Function &F);
BasicBlock *findDestBlockOfMergeableEmptyBlock(BasicBlock *BB);
bool canMergeBlocks(const BasicBlock *BB, const BasicBlock *DestBB) const;
void eliminateMostlyEmptyBlock(BasicBlock *BB);
bool isMergingEmptyBlockProfitable(BasicBlock *BB, BasicBlock *DestBB,
bool isPreheader);
bool makeBitReverse(Instruction &I);
bool optimizeBlock(BasicBlock &BB, bool &ModifiedDT);
bool optimizeInst(Instruction *I, bool &ModifiedDT);
bool optimizeMemoryInst(Instruction *MemoryInst, Value *Addr,
Type *AccessTy, unsigned AddrSpace);
bool optimizeGatherScatterInst(Instruction *MemoryInst, Value *Ptr);
bool optimizeInlineAsmInst(CallInst *CS);
bool optimizeCallInst(CallInst *CI, bool &ModifiedDT);
bool optimizeExt(Instruction *&I);
bool optimizeExtUses(Instruction *I);
bool optimizeLoadExt(LoadInst *Load);
bool optimizeShiftInst(BinaryOperator *BO);
bool optimizeFunnelShift(IntrinsicInst *Fsh);
bool optimizeSelectInst(SelectInst *SI);
bool optimizeShuffleVectorInst(ShuffleVectorInst *SVI);
bool optimizeSwitchInst(SwitchInst *SI);
bool optimizeExtractElementInst(Instruction *Inst);
bool dupRetToEnableTailCallOpts(BasicBlock *BB, bool &ModifiedDT);
bool fixupDbgValue(Instruction *I);
bool placeDbgValues(Function &F);
bool canFormExtLd(const SmallVectorImpl<Instruction *> &MovedExts,
LoadInst *&LI, Instruction *&Inst, bool HasPromoted);
bool tryToPromoteExts(TypePromotionTransaction &TPT,
const SmallVectorImpl<Instruction *> &Exts,
SmallVectorImpl<Instruction *> &ProfitablyMovedExts,
unsigned CreatedInstsCost = 0);
bool mergeSExts(Function &F);
bool splitLargeGEPOffsets();
bool optimizePhiType(PHINode *Inst, SmallPtrSetImpl<PHINode *> &Visited,
SmallPtrSetImpl<Instruction *> &DeletedInstrs);
bool optimizePhiTypes(Function &F);
bool performAddressTypePromotion(
Instruction *&Inst,
bool AllowPromotionWithoutCommonHeader,
bool HasPromoted, TypePromotionTransaction &TPT,
SmallVectorImpl<Instruction *> &SpeculativelyMovedExts);
bool splitBranchCondition(Function &F, bool &ModifiedDT);
bool simplifyOffsetableRelocate(GCStatepointInst &I);
bool tryToSinkFreeOperands(Instruction *I);
bool replaceMathCmpWithIntrinsic(BinaryOperator *BO, Value *Arg0,
Value *Arg1, CmpInst *Cmp,
Intrinsic::ID IID);
bool optimizeCmp(CmpInst *Cmp, bool &ModifiedDT);
bool combineToUSubWithOverflow(CmpInst *Cmp, bool &ModifiedDT);
bool combineToUAddWithOverflow(CmpInst *Cmp, bool &ModifiedDT);
void verifyBFIUpdates(Function &F);
} // end anonymous namespace
char CodeGenPrepare::ID = 0;
"Optimize for code generation", false, false)
"Optimize for code generation", false, false)
FunctionPass *llvm::createCodeGenPreparePass() { return new CodeGenPrepare(); }
bool CodeGenPrepare::runOnFunction(Function &F) {
if (skipFunction(F))
return false;
DL = &F.getParent()->getDataLayout();
bool EverMadeChange = false;
// Clear per function information.
TM = &getAnalysis<TargetPassConfig>().getTM<TargetMachine>();
SubtargetInfo = TM->getSubtargetImpl(F);
TLI = SubtargetInfo->getTargetLowering();
TRI = SubtargetInfo->getRegisterInfo();
TLInfo = &getAnalysis<TargetLibraryInfoWrapperPass>().getTLI(F);
TTI = &getAnalysis<TargetTransformInfoWrapperPass>().getTTI(F);
LI = &getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
BPI.reset(new BranchProbabilityInfo(F, *LI));
BFI.reset(new BlockFrequencyInfo(F, *BPI, *LI));
PSI = &getAnalysis<ProfileSummaryInfoWrapperPass>().getPSI();
OptSize = F.hasOptSize();
if (ProfileGuidedSectionPrefix) {
// The hot attribute overwrites profile count based hotness while profile
// counts based hotness overwrite the cold attribute.
// This is a conservative behabvior.
if (F.hasFnAttribute(Attribute::Hot) ||
PSI->isFunctionHotInCallGraph(&F, *BFI))
// If PSI shows this function is not hot, we will placed the function
// into unlikely section if (1) PSI shows this is a cold function, or
// (2) the function has a attribute of cold.
else if (PSI->isFunctionColdInCallGraph(&F, *BFI) ||
else if (ProfileUnknownInSpecialSection && PSI->hasPartialSampleProfile() &&
/// This optimization identifies DIV instructions that can be
/// profitably bypassed and carried out with a shorter, faster divide.
if (!OptSize && !PSI->hasHugeWorkingSetSize() && TLI->isSlowDivBypassed()) {
const DenseMap<unsigned int, unsigned int> &BypassWidths =
BasicBlock* BB = &*F.begin();
while (BB != nullptr) {
// bypassSlowDivision may create new BBs, but we don't want to reapply the
// optimization to those blocks.
BasicBlock* Next = BB->getNextNode();
// F.hasOptSize is already checked in the outer if statement.
if (!llvm::shouldOptimizeForSize(BB, PSI, BFI.get()))
EverMadeChange |= bypassSlowDivision(BB, BypassWidths);
BB = Next;
// Eliminate blocks that contain only PHI nodes and an
// unconditional branch.
EverMadeChange |= eliminateMostlyEmptyBlocks(F);
bool ModifiedDT = false;
if (!DisableBranchOpts)
EverMadeChange |= splitBranchCondition(F, ModifiedDT);
// Split some critical edges where one of the sources is an indirect branch,
// to help generate sane code for PHIs involving such edges.
EverMadeChange |= SplitIndirectBrCriticalEdges(F);
bool MadeChange = true;
while (MadeChange) {
MadeChange = false;
for (Function::iterator I = F.begin(); I != F.end(); ) {
BasicBlock *BB = &*I++;
bool ModifiedDTOnIteration = false;
MadeChange |= optimizeBlock(*BB, ModifiedDTOnIteration);
// Restart BB iteration if the dominator tree of the Function was changed
if (ModifiedDTOnIteration)
if (EnableTypePromotionMerge && !ValToSExtendedUses.empty())
MadeChange |= mergeSExts(F);
if (!LargeOffsetGEPMap.empty())
MadeChange |= splitLargeGEPOffsets();
MadeChange |= optimizePhiTypes(F);
if (MadeChange)
// Really free removed instructions during promotion.
for (Instruction *I : RemovedInsts)
EverMadeChange |= MadeChange;
if (!DisableBranchOpts) {
MadeChange = false;
// Use a set vector to get deterministic iteration order. The order the
// blocks are removed may affect whether or not PHI nodes in successors
// are removed.
SmallSetVector<BasicBlock*, 8> WorkList;
for (BasicBlock &BB : F) {
SmallVector<BasicBlock *, 2> Successors(successors(&BB));
MadeChange |= ConstantFoldTerminator(&BB, true);
if (!MadeChange) continue;
for (BasicBlock *Succ : Successors)
if (pred_empty(Succ))
// Delete the dead blocks and any of their dead successors.
MadeChange |= !WorkList.empty();
while (!WorkList.empty()) {
BasicBlock *BB = WorkList.pop_back_val();
SmallVector<BasicBlock*, 2> Successors(successors(BB));
for (BasicBlock *Succ : Successors)
if (pred_empty(Succ))
// Merge pairs of basic blocks with unconditional branches, connected by
// a single edge.
if (EverMadeChange || MadeChange)
MadeChange |= eliminateFallThrough(F);
EverMadeChange |= MadeChange;
if (!DisableGCOpts) {
SmallVector<GCStatepointInst *, 2> Statepoints;
for (BasicBlock &BB : F)
for (Instruction &I : BB)
if (auto *SP = dyn_cast<GCStatepointInst>(&I))
for (auto &I : Statepoints)
EverMadeChange |= simplifyOffsetableRelocate(*I);
// Do this last to clean up use-before-def scenarios introduced by other
// preparatory transforms.
EverMadeChange |= placeDbgValues(F);
#ifndef NDEBUG
if (VerifyBFIUpdates)
return EverMadeChange;
/// An instruction is about to be deleted, so remove all references to it in our
/// GEP-tracking data strcutures.
void CodeGenPrepare::removeAllAssertingVHReferences(Value *V) {
auto GEP = dyn_cast<GetElementPtrInst>(V);
if (!GEP)
auto VecI = LargeOffsetGEPMap.find(GEP->getPointerOperand());
if (VecI == LargeOffsetGEPMap.end())
auto &GEPVector = VecI->second;
const auto &I =
llvm::find_if(GEPVector, [=](auto &Elt) { return Elt.first == GEP; });
if (I == GEPVector.end())
if (GEPVector.empty())
// Verify BFI has been updated correctly by recomputing BFI and comparing them.
void LLVM_ATTRIBUTE_UNUSED CodeGenPrepare::verifyBFIUpdates(Function &F) {
DominatorTree NewDT(F);
LoopInfo NewLI(NewDT);
BranchProbabilityInfo NewBPI(F, NewLI, TLInfo);
BlockFrequencyInfo NewBFI(F, NewBPI, NewLI);
/// Merge basic blocks which are connected by a single edge, where one of the
/// basic blocks has a single successor pointing to the other basic block,
/// which has a single predecessor.
bool CodeGenPrepare::eliminateFallThrough(Function &F) {
bool Changed = false;
// Scan all of the blocks in the function, except for the entry block.
// Use a temporary array to avoid iterator being invalidated when
// deleting blocks.
SmallVector<WeakTrackingVH, 16> Blocks;
for (auto &Block : llvm::drop_begin(F))
SmallSet<WeakTrackingVH, 16> Preds;
for (auto &Block : Blocks) {
auto *BB = cast_or_null<BasicBlock>(Block);
if (!BB)
// If the destination block has a single pred, then this is a trivial
// edge, just collapse it.
BasicBlock *SinglePred = BB->getSinglePredecessor();
// Don't merge if BB's address is taken.
if (!SinglePred || SinglePred == BB || BB->hasAddressTaken()) continue;
BranchInst *Term = dyn_cast<BranchInst>(SinglePred->getTerminator());
if (Term && !Term->isConditional()) {
Changed = true;
LLVM_DEBUG(dbgs() << "To merge:\n" << *BB << "\n\n\n");
// Merge BB into SinglePred and delete it.
// (Repeatedly) merging blocks into their predecessors can create redundant
// debug intrinsics.
for (auto &Pred : Preds)
if (auto *BB = cast_or_null<BasicBlock>(Pred))
return Changed;
/// Find a destination block from BB if BB is mergeable empty block.
BasicBlock *CodeGenPrepare::findDestBlockOfMergeableEmptyBlock(BasicBlock *BB) {
// If this block doesn't end with an uncond branch, ignore it.
BranchInst *BI = dyn_cast<BranchInst>(BB->getTerminator());
if (!BI || !BI->isUnconditional())
return nullptr;
// If the instruction before the branch (skipping debug info) isn't a phi
// node, then other stuff is happening here.
BasicBlock::iterator BBI = BI->getIterator();
if (BBI != BB->begin()) {
while (isa<DbgInfoIntrinsic>(BBI)) {
if (BBI == BB->begin())
if (!isa<DbgInfoIntrinsic>(BBI) && !isa<PHINode>(BBI))
return nullptr;
// Do not break infinite loops.
BasicBlock *DestBB = BI->getSuccessor(0);
if (DestBB == BB)
return nullptr;
if (!canMergeBlocks(BB, DestBB))
DestBB = nullptr;
return DestBB;
/// Eliminate blocks that contain only PHI nodes, debug info directives, and an
/// unconditional branch. Passes before isel (e.g. LSR/loopsimplify) often split
/// edges in ways that are non-optimal for isel. Start by eliminating these
/// blocks so we can split them the way we want them.
bool CodeGenPrepare::eliminateMostlyEmptyBlocks(Function &F) {
SmallPtrSet<BasicBlock *, 16> Preheaders;
SmallVector<Loop *, 16> LoopList(LI->begin(), LI->end());
while (!LoopList.empty()) {
Loop *L = LoopList.pop_back_val();
llvm::append_range(LoopList, *L);
if (BasicBlock *Preheader = L->getLoopPreheader())
bool MadeChange = false;
// Copy blocks into a temporary array to avoid iterator invalidation issues
// as we remove them.
// Note that this intentionally skips the entry block.
SmallVector<WeakTrackingVH, 16> Blocks;
for (auto &Block : llvm::drop_begin(F))
for (auto &Block : Blocks) {
BasicBlock *BB = cast_or_null<BasicBlock>(Block);
if (!BB)
BasicBlock *DestBB = findDestBlockOfMergeableEmptyBlock(BB);
if (!DestBB ||
!isMergingEmptyBlockProfitable(BB, DestBB, Preheaders.count(BB)))
MadeChange = true;
return MadeChange;
bool CodeGenPrepare::isMergingEmptyBlockProfitable(BasicBlock *BB,
BasicBlock *DestBB,
bool isPreheader) {
// Do not delete loop preheaders if doing so would create a critical edge.
// Loop preheaders can be good locations to spill registers. If the
// preheader is deleted and we create a critical edge, registers may be
// spilled in the loop body instead.
if (!DisablePreheaderProtect && isPreheader &&
!(BB->getSinglePredecessor() &&
return false;
// Skip merging if the block's successor is also a successor to any callbr
// that leads to this block.
// FIXME: Is this really needed? Is this a correctness issue?
for (BasicBlock *Pred : predecessors(BB)) {
if (auto *CBI = dyn_cast<CallBrInst>((Pred)->getTerminator()))
for (unsigned i = 0, e = CBI->getNumSuccessors(); i != e; ++i)
if (DestBB == CBI->getSuccessor(i))
return false;
// Try to skip merging if the unique predecessor of BB is terminated by a
// switch or indirect branch instruction, and BB is used as an incoming block
// of PHIs in DestBB. In such case, merging BB and DestBB would cause ISel to
// add COPY instructions in the predecessor of BB instead of BB (if it is not
// merged). Note that the critical edge created by merging such blocks wont be
// split in MachineSink because the jump table is not analyzable. By keeping
// such empty block (BB), ISel will place COPY instructions in BB, not in the
// predecessor of BB.
BasicBlock *Pred = BB->getUniquePredecessor();
if (!Pred ||
!(isa<SwitchInst>(Pred->getTerminator()) ||
return true;
if (BB->getTerminator() != BB->getFirstNonPHIOrDbg())
return true;
// We use a simple cost heuristic which determine skipping merging is
// profitable if the cost of skipping merging is less than the cost of
// merging : Cost(skipping merging) < Cost(merging BB), where the
// Cost(skipping merging) is Freq(BB) * (Cost(Copy) + Cost(Branch)), and
// the Cost(merging BB) is Freq(Pred) * Cost(Copy).
// Assuming Cost(Copy) == Cost(Branch), we could simplify it to :
// Freq(Pred) / Freq(BB) > 2.
// Note that if there are multiple empty blocks sharing the same incoming
// value for the PHIs in the DestBB, we consider them together. In such
// case, Cost(merging BB) will be the sum of their frequencies.
if (!isa<PHINode>(DestBB->begin()))
return true;
SmallPtrSet<BasicBlock *, 16> SameIncomingValueBBs;
// Find all other incoming blocks from which incoming values of all PHIs in
// DestBB are the same as the ones from BB.
for (BasicBlock *DestBBPred : predecessors(DestBB)) {
if (DestBBPred == BB)
if (llvm::all_of(DestBB->phis(), [&](const PHINode &DestPN) {
return DestPN.getIncomingValueForBlock(BB) ==
// See if all BB's incoming values are same as the value from Pred. In this
// case, no reason to skip merging because COPYs are expected to be place in
// Pred already.
if (SameIncomingValueBBs.count(Pred))
return true;
BlockFrequency PredFreq = BFI->getBlockFreq(Pred);
BlockFrequency BBFreq = BFI->getBlockFreq(BB);
for (auto *SameValueBB : SameIncomingValueBBs)
if (SameValueBB->getUniquePredecessor() == Pred &&
DestBB == findDestBlockOfMergeableEmptyBlock(SameValueBB))
BBFreq += BFI->getBlockFreq(SameValueBB);
return PredFreq.getFrequency() <=
BBFreq.getFrequency() * FreqRatioToSkipMerge;
/// Return true if we can merge BB into DestBB if there is a single
/// unconditional branch between them, and BB contains no other non-phi
/// instructions.
bool CodeGenPrepare::canMergeBlocks(const BasicBlock *BB,
const BasicBlock *DestBB) const {
// We only want to eliminate blocks whose phi nodes are used by phi nodes in
// the successor. If there are more complex condition (e.g. preheaders),
// don't mess around with them.
for (const PHINode &PN : BB->phis()) {
for (const User *U : PN.users()) {
const Instruction *UI = cast<Instruction>(U);
if (UI->getParent() != DestBB || !isa<PHINode>(UI))
return false;
// If User is inside DestBB block and it is a PHINode then check
// incoming value. If incoming value is not from BB then this is
// a complex condition (e.g. preheaders) we want to avoid here.
if (UI->getParent() == DestBB) {
if (const PHINode *UPN = dyn_cast<PHINode>(UI))
for (unsigned I = 0, E = UPN->getNumIncomingValues(); I != E; ++I) {
Instruction *Insn = dyn_cast<Instruction>(UPN->getIncomingValue(I));
if (Insn && Insn->getParent() == BB &&
Insn->getParent() != UPN->getIncomingBlock(I))
return false;
// If BB and DestBB contain any common predecessors, then the phi nodes in BB
// and DestBB may have conflicting incoming values for the block. If so, we
// can't merge the block.
const PHINode *DestBBPN = dyn_cast<PHINode>(DestBB->begin());
if (!DestBBPN) return true; // no conflict.
// Collect the preds of BB.
SmallPtrSet<const BasicBlock*, 16> BBPreds;
if (const PHINode *BBPN = dyn_cast<PHINode>(BB->begin())) {
// It is faster to get preds from a PHI than with pred_iterator.
for (unsigned i = 0, e = BBPN->getNumIncomingValues(); i != e; ++i)
} else {
BBPreds.insert(pred_begin(BB), pred_end(BB));
// Walk the preds of DestBB.
for (unsigned i = 0, e = DestBBPN->getNumIncomingValues(); i != e; ++i) {
BasicBlock *Pred = DestBBPN->getIncomingBlock(i);
if (BBPreds.count(Pred)) { // Common predecessor?
for (const PHINode &PN : DestBB->phis()) {
const Value *V1 = PN.getIncomingValueForBlock(Pred);
const Value *V2 = PN.getIncomingValueForBlock(BB);
// If V2 is a phi node in BB, look up what the mapped value will be.
if (const PHINode *V2PN = dyn_cast<PHINode>(V2))
if (V2PN->getParent() == BB)
V2 = V2PN->getIncomingValueForBlock(Pred);
// If there is a conflict, bail out.
if (V1 != V2) return false;
return true;
/// Eliminate a basic block that has only phi's and an unconditional branch in
/// it.
void CodeGenPrepare::eliminateMostlyEmptyBlock(BasicBlock *BB) {
BranchInst *BI = cast<BranchInst>(BB->getTerminator());
BasicBlock *DestBB = BI->getSuccessor(0);
<< *BB << *DestBB);
// If the destination block has a single pred, then this is a trivial edge,
// just collapse it.
if (BasicBlock *SinglePred = DestBB->getSinglePredecessor()) {
if (SinglePred != DestBB) {
assert(SinglePred == BB &&
"Single predecessor not the same as predecessor");
// Merge DestBB into SinglePred/BB and delete it.
// Note: BB(=SinglePred) will not be deleted on this path.
// DestBB(=its single successor) is the one that was deleted.
LLVM_DEBUG(dbgs() << "AFTER:\n" << *SinglePred << "\n\n\n");
// Otherwise, we have multiple predecessors of BB. Update the PHIs in DestBB
// to handle the new incoming edges it is about to have.
for (PHINode &PN : DestBB->phis()) {
// Remove the incoming value for BB, and remember it.
Value *InVal = PN.removeIncomingValue(BB, false);
// Two options: either the InVal is a phi node defined in BB or it is some
// value that dominates BB.
PHINode *InValPhi = dyn_cast<PHINode>(InVal);
if (InValPhi && InValPhi->getParent() == BB) {
// Add all of the input values of the input PHI as inputs of this phi.
for (unsigned i = 0, e = InValPhi->getNumIncomingValues(); i != e; ++i)
} else {
// Otherwise, add one instance of the dominating value for each edge that
// we will be adding.
if (PHINode *BBPN = dyn_cast<PHINode>(BB->begin())) {
for (unsigned i = 0, e = BBPN->getNumIncomingValues(); i != e; ++i)
PN.addIncoming(InVal, BBPN->getIncomingBlock(i));
} else {
for (BasicBlock *Pred : predecessors(BB))
PN.addIncoming(InVal, Pred);
// The PHIs are now updated, change everything that refers to BB to use
// DestBB and remove BB.
LLVM_DEBUG(dbgs() << "AFTER:\n" << *DestBB << "\n\n\n");
// Computes a map of base pointer relocation instructions to corresponding
// derived pointer relocation instructions given a vector of all relocate calls
static void computeBaseDerivedRelocateMap(
const SmallVectorImpl<GCRelocateInst *> &AllRelocateCalls,
DenseMap<GCRelocateInst *, SmallVector<GCRelocateInst *, 2>>
&RelocateInstMap) {
// Collect information in two maps: one primarily for locating the base object
// while filling the second map; the second map is the final structure holding
// a mapping between Base and corresponding Derived relocate calls
DenseMap<std::pair<unsigned, unsigned>, GCRelocateInst *> RelocateIdxMap;
for (auto *ThisRelocate : AllRelocateCalls) {
auto K = std::make_pair(ThisRelocate->getBasePtrIndex(),
RelocateIdxMap.insert(std::make_pair(K, ThisRelocate));
for (auto &Item : RelocateIdxMap) {
std::pair<unsigned, unsigned> Key = Item.first;
if (Key.first == Key.second)
// Base relocation: nothing to insert
GCRelocateInst *I = Item.second;
auto BaseKey = std::make_pair(Key.first, Key.first);
// We're iterating over RelocateIdxMap so we cannot modify it.
auto MaybeBase = RelocateIdxMap.find(BaseKey);
if (MaybeBase == RelocateIdxMap.end())
// TODO: We might want to insert a new base object relocate and gep off
// that, if there are enough derived object relocates.
// Accepts a GEP and extracts the operands into a vector provided they're all
// small integer constants
static bool getGEPSmallConstantIntOffsetV(GetElementPtrInst *GEP,
SmallVectorImpl<Value *> &OffsetV) {
for (unsigned i = 1; i < GEP->getNumOperands(); i++) {
// Only accept small constant integer operands
auto *Op = dyn_cast<ConstantInt>(GEP->getOperand(i));
if (!Op || Op->getZExtValue() > 20)
return false;
for (unsigned i = 1; i < GEP->getNumOperands(); i++)
return true;
// Takes a RelocatedBase (base pointer relocation instruction) and Targets to
// replace, computes a replacement, and affects it.
static bool
simplifyRelocatesOffABase(GCRelocateInst *RelocatedBase,
const SmallVectorImpl<GCRelocateInst *> &Targets) {
bool MadeChange = false;
// We must ensure the relocation of derived pointer is defined after
// relocation of base pointer. If we find a relocation corresponding to base
// defined earlier than relocation of base then we move relocation of base
// right before found relocation. We consider only relocation in the same
// basic block as relocation of base. Relocations from other basic block will
// be skipped by optimization and we do not care about them.
for (auto R = RelocatedBase->getParent()->getFirstInsertionPt();
&*R != RelocatedBase; ++R)
if (auto *RI = dyn_cast<GCRelocateInst>(R))
if (RI->getStatepoint() == RelocatedBase->getStatepoint())
if (RI->getBasePtrIndex() == RelocatedBase->getBasePtrIndex()) {
for (GCRelocateInst *ToReplace : Targets) {
assert(ToReplace->getBasePtrIndex() == RelocatedBase->getBasePtrIndex() &&
"Not relocating a derived object of the original base object");
if (ToReplace->getBasePtrIndex() == ToReplace->getDerivedPtrIndex()) {
// A duplicate relocate call. TODO: coalesce duplicates.
if (RelocatedBase->getParent() != ToReplace->getParent()) {
// Base and derived relocates are in different basic blocks.
// In this case transform is only valid when base dominates derived
// relocate. However it would be too expensive to check dominance
// for each such relocate, so we skip the whole transformation.
Value *Base = ToReplace->getBasePtr();
auto *Derived = dyn_cast<GetElementPtrInst>(ToReplace->getDerivedPtr());
if (!Derived || Derived->getPointerOperand() != Base)
SmallVector<Value *, 2> OffsetV;
if (!getGEPSmallConstantIntOffsetV(Derived, OffsetV))
// Create a Builder and replace the target callsite with a gep
assert(RelocatedBase->getNextNode() &&
"Should always have one since it's not a terminator");
// Insert after RelocatedBase
IRBuilder<> Builder(RelocatedBase->getNextNode());
// If gc_relocate does not match the actual type, cast it to the right type.
// In theory, there must be a bitcast after gc_relocate if the type does not
// match, and we should reuse it to get the derived pointer. But it could be
// cases like this:
// bb1:
// ...
// %g1 = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(...)
// br label %merge
// bb2:
// ...
// %g2 = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(...)
// br label %merge
// merge:
// %p1 = phi i8 addrspace(1)* [ %g1, %bb1 ], [ %g2, %bb2 ]
// %cast = bitcast i8 addrspace(1)* %p1 in to i32 addrspace(1)*
// In this case, we can not find the bitcast any more. So we insert a new bitcast
// no matter there is already one or not. In this way, we can handle all cases, and
// the extra bitcast should be optimized away in later passes.
Value *ActualRelocatedBase = RelocatedBase;
if (RelocatedBase->getType() != Base->getType()) {
ActualRelocatedBase =
Builder.CreateBitCast(RelocatedBase, Base->getType());
Value *Replacement = Builder.CreateGEP(
Derived->getSourceElementType(), ActualRelocatedBase, makeArrayRef(OffsetV));
// If the newly generated derived pointer's type does not match the original derived
// pointer's type, cast the new derived pointer to match it. Same reasoning as above.
Value *ActualReplacement = Replacement;
if (Replacement->getType() != ToReplace->getType()) {
ActualReplacement =
Builder.CreateBitCast(Replacement, ToReplace->getType());
MadeChange = true;
return MadeChange;
// Turns this:
// %base = ...
// %ptr = gep %base + 15
// %tok = statepoint (%fun, i32 0, i32 0, i32 0, %base, %ptr)
// %base' = relocate(%tok, i32 4, i32 4)
// %ptr' = relocate(%tok, i32 4, i32 5)
// %val = load %ptr'
// into this:
// %base = ...
// %ptr = gep %base + 15
// %tok = statepoint (%fun, i32 0, i32 0, i32 0, %base, %ptr)
// %base' = gc.relocate(%tok, i32 4, i32 4)
// %ptr' = gep %base' + 15
// %val = load %ptr'
bool CodeGenPrepare::simplifyOffsetableRelocate(GCStatepointInst &I) {
bool MadeChange = false;
SmallVector<GCRelocateInst *, 2> AllRelocateCalls;
for (auto *U : I.users())
if (GCRelocateInst *Relocate = dyn_cast<GCRelocateInst>(U))
// Collect all the relocate calls associated with a statepoint
// We need at least one base pointer relocation + one derived pointer
// relocation to mangle
if (AllRelocateCalls.size() < 2)
return false;
// RelocateInstMap is a mapping from the base relocate instruction to the
// corresponding derived relocate instructions
DenseMap<GCRelocateInst *, SmallVector<GCRelocateInst *, 2>> RelocateInstMap;
computeBaseDerivedRelocateMap(AllRelocateCalls, RelocateInstMap);
if (RelocateInstMap.empty())
return false;
for (auto &Item : RelocateInstMap)
// Item.first is the RelocatedBase to offset against
// Item.second is the vector of Targets to replace
MadeChange = simplifyRelocatesOffABase(Item.first, Item.second);
return MadeChange;
/// Sink the specified cast instruction into its user blocks.
static bool SinkCast(CastInst *CI) {
BasicBlock *DefBB = CI->getParent();
/// InsertedCasts - Only insert a cast in each block once.
DenseMap<BasicBlock*, CastInst*> InsertedCasts;
bool MadeChange = false;
for (Value::user_iterator UI = CI->user_begin(), E = CI->user_end();
UI != E; ) {
Use &TheUse = UI.getUse();
Instruction *User = cast<Instruction>(*UI);
// Figure out which BB this cast is used in. For PHI's this is the
// appropriate predecessor block.
BasicBlock *UserBB = User->getParent();
if (PHINode *PN = dyn_cast<PHINode>(User)) {
UserBB = PN->getIncomingBlock(TheUse);
// Preincrement use iterator so we don't invalidate it.
// The first insertion point of a block containing an EH pad is after the
// pad. If the pad is the user, we cannot sink the cast past the pad.
if (User->isEHPad())
// If the block selected to receive the cast is an EH pad that does not
// allow non-PHI instructions before the terminator, we can't sink the
// cast.
if (UserBB->getTerminator()->isEHPad())
// If this user is in the same block as the cast, don't change the cast.
if (UserBB == DefBB) continue;
// If we have already inserted a cast into this block, use it.
CastInst *&InsertedCast = InsertedCasts[UserBB];
if (!InsertedCast) {
BasicBlock::iterator InsertPt = UserBB->getFirstInsertionPt();
assert(InsertPt != UserBB->end());
InsertedCast = CastInst::Create(CI->getOpcode(), CI->getOperand(0),
CI->getType(), "", &*InsertPt);
// Replace a use of the cast with a use of the new cast.
TheUse = InsertedCast;
MadeChange = true;
// If we removed all uses, nuke the cast.
if (CI->use_empty()) {
MadeChange = true;
return MadeChange;
/// If the specified cast instruction is a noop copy (e.g. it's casting from
/// one pointer type to another, i32->i8 on PPC), sink it into user blocks to
/// reduce the number of virtual registers that must be created and coalesced.
/// Return true if any changes are made.
static bool OptimizeNoopCopyExpression(CastInst *CI, const TargetLowering &TLI,
const DataLayout &DL) {
// Sink only "cheap" (or nop) address-space casts. This is a weaker condition
// than sinking only nop casts, but is helpful on some platforms.
if (auto *ASC = dyn_cast<AddrSpaceCastInst>(CI)) {
if (!TLI.isFreeAddrSpaceCast(ASC->getSrcAddressSpace(),
return false;
// If this is a noop copy,
EVT SrcVT = TLI.getValueType(DL, CI->getOperand(0)->getType());
EVT DstVT = TLI.getValueType(DL, CI->getType());
// This is an fp<->int conversion?
if (SrcVT.isInteger() != DstVT.isInteger())
return false;
// If this is an extension, it will be a zero or sign extension, which
// isn't a noop.
if (SrcVT.bitsLT(DstVT)) return false;
// If these values will be promoted, find out what they will be promoted
// to. This helps us consider truncates on PPC as noop copies when they
// are.
if (TLI.getTypeAction(CI->getContext(), SrcVT) ==
SrcVT = TLI.getTypeToTransformTo(CI->getContext(), SrcVT);
if (TLI.getTypeAction(CI->getContext(), DstVT) ==
DstVT = TLI.getTypeToTransformTo(CI->getContext(), DstVT);
// If, after promotion, these are the same types, this is a noop copy.
if (SrcVT != DstVT)
return false;
return SinkCast(CI);
bool CodeGenPrepare::replaceMathCmpWithIntrinsic(BinaryOperator *BO,
Value *Arg0, Value *Arg1,
CmpInst *Cmp,
Intrinsic::ID IID) {
auto isIVIncrement = [this, &Cmp](BinaryOperator *BO) {
auto *PN = dyn_cast<PHINode>(BO->getOperand(0));
if (!PN)
return false;
const Loop *L = LI->getLoopFor(BO->getParent());
if (!L || L->getHeader() != PN->getParent() || !L->getLoopLatch())
return false;
const BasicBlock *Latch = L->getLoopLatch();
if (PN->getIncomingValueForBlock(Latch) != BO)
return false;
if (auto *Step = dyn_cast<Instruction>(BO->getOperand(1)))
if (L->contains(Step->getParent()))
return false;
// IV increment may have other users than the IV. We do not want to make
// dominance queries to analyze the legality of moving it towards the cmp,
// so just check that there is no other users.
if (!BO->hasOneUse())
return false;
// Do not risk on moving increment into a child loop.
if (LI->getLoopFor(Cmp->getParent()) != L)
return false;
// Ultimately, the insertion point must dominate latch. This should be a
// cheap check because no CFG changes & dom tree recomputation happens
// during the transform.
Function *F = BO->getParent()->getParent();
return getDT(*F).dominates(Cmp->getParent(), Latch);
if (BO->getParent() != Cmp->getParent() && !isIVIncrement(BO)) {
// We used to use a dominator tree here to allow multi-block optimization.
// But that was problematic because:
// 1. It could cause a perf regression by hoisting the math op into the
// critical path.
// 2. It could cause a perf regression by creating a value that was live
// across multiple blocks and increasing register pressure.
// 3. Use of a dominator tree could cause large compile-time regression.
// This is because we recompute the DT on every change in the main CGP
// run-loop. The recomputing is probably unnecessary in many cases, so if
// that was fixed, using a DT here would be ok.
// There is one important particular case we still want to handle: if BO is
// the IV increment. Important properties that make it profitable:
// - We can speculate IV increment anywhere in the loop (as long as the
// indvar Phi is its only user);
// - Upon computing Cmp, we effectively compute something equivalent to the
// IV increment (despite it loops differently in the IR). So moving it up
// to the cmp point does not really increase register pressure.
return false;
// We allow matching the canonical IR (add X, C) back to (usubo X, -C).
if (BO->getOpcode() == Instruction::Add &&
IID == Intrinsic::usub_with_overflow) {
assert(isa<Constant>(Arg1) && "Unexpected input for usubo");
Arg1 = ConstantExpr::getNeg(cast<Constant>(Arg1));
// Insert at the first instruction of the pair.
Instruction *InsertPt = nullptr;
for (Instruction &Iter : *Cmp->getParent()) {
// If BO is an XOR, it is not guaranteed that it comes after both inputs to
// the overflow intrinsic are defined.
if ((BO->getOpcode() != Instruction::Xor && &Iter == BO) || &Iter == Cmp) {
InsertPt = &Iter;
assert(InsertPt != nullptr && "Parent block did not contain cmp or binop");
IRBuilder<> Builder(InsertPt);
Value *MathOV = Builder.CreateBinaryIntrinsic(IID, Arg0, Arg1);
if (BO->getOpcode() != Instruction::Xor) {
Value *Math = Builder.CreateExtractValue(MathOV, 0, "math");
} else
assert(BO->hasOneUse() &&
"Patterns with XOr should use the BO only in the compare");
Value *OV = Builder.CreateExtractValue(MathOV, 1, "ov");
return true;
/// Match special-case patterns that check for unsigned add overflow.
static bool matchUAddWithOverflowConstantEdgeCases(CmpInst *Cmp,
BinaryOperator *&Add) {
// Add = add A, 1; Cmp = icmp eq A,-1 (overflow if A is max val)
// Add = add A,-1; Cmp = icmp ne A, 0 (overflow if A is non-zero)
Value *A = Cmp->getOperand(0), *B = Cmp->getOperand(1);
// We are not expecting non-canonical/degenerate code. Just bail out.
if (isa<Constant>(A))
return false;
ICmpInst::Predicate Pred = Cmp->getPredicate();
if (Pred == ICmpInst::ICMP_EQ && match(B, m_AllOnes()))
B = ConstantInt::get(B->getType(), 1);
else if (Pred == ICmpInst::ICMP_NE && match(B, m_ZeroInt()))
B = ConstantInt::get(B->getType(), -1);
return false;
// Check the users of the variable operand of the compare looking for an add
// with the adjusted constant.
for (User *U : A->users()) {
if (match(U, m_Add(m_Specific(A), m_Specific(B)))) {
Add = cast<BinaryOperator>(U);
return true;
return false;
/// Try to combine the compare into a call to the llvm.uadd.with.overflow
/// intrinsic. Return true if any changes were made.
bool CodeGenPrepare::combineToUAddWithOverflow(CmpInst *Cmp,
bool &ModifiedDT) {
Value *A, *B;
BinaryOperator *Add;
if (!match(Cmp, m_UAddWithOverflow(m_Value(A), m_Value(B), m_BinOp(Add)))) {
if (!matchUAddWithOverflowConstantEdgeCases(Cmp, Add))
return false;
// Set A and B in case we match matchUAddWithOverflowConstantEdgeCases.
A = Add->getOperand(0);
B = Add->getOperand(1);
if (!TLI->shouldFormOverflowOp(ISD::UADDO,
TLI->getValueType(*DL, Add->getType()),
return false;
// We don't want to move around uses of condition values this late, so we
// check if it is legal to create the call to the intrinsic in the basic
// block containing the icmp.
if (Add->getParent() != Cmp->getParent() && !Add->hasOneUse())
return false;
if (!replaceMathCmpWithIntrinsic(Add, A, B, Cmp,
return false;
// Reset callers - do not crash by iterating over a dead instruction.
ModifiedDT = true;
return true;
bool CodeGenPrepare::combineToUSubWithOverflow(CmpInst *Cmp,
bool &ModifiedDT) {
// We are not expecting non-canonical/degenerate code. Just bail out.
Value *A = Cmp->getOperand(0), *B = Cmp->getOperand(1);
if (isa<Constant>(A) && isa<Constant>(B))
return false;
// Convert (A u> B) to (A u< B) to simplify pattern matching.
ICmpInst::Predicate Pred = Cmp->getPredicate();
if (Pred == ICmpInst::ICMP_UGT) {
std::swap(A, B);
Pred = ICmpInst::ICMP_ULT;
// Convert special-case: (A == 0) is the same as (A u< 1).
if (Pred == ICmpInst::ICMP_EQ && match(B, m_ZeroInt())) {
B = ConstantInt::get(B->getType(), 1);
Pred = ICmpInst::ICMP_ULT;
// Convert special-case: (A != 0) is the same as (0 u< A).
if (Pred == ICmpInst::ICMP_NE && match(B, m_ZeroInt())) {
std::swap(A, B);
Pred = ICmpInst::ICMP_ULT;
if (Pred != ICmpInst::ICMP_ULT)
return false;
// Walk the users of a variable operand of a compare looking for a subtract or
// add with that same operand. Also match the 2nd operand of the compare to
// the add/sub, but that may be a negated constant operand of an add.
Value *CmpVariableOperand = isa<Constant>(A) ? B : A;
BinaryOperator *Sub = nullptr;
for (User *U : CmpVariableOperand->users()) {
// A - B, A u< B --> usubo(A, B)
if (match(U, m_Sub(m_Specific(A), m_Specific(B)))) {
Sub = cast<BinaryOperator>(U);
// A + (-C), A u< C (canonicalized form of (sub A, C))
const APInt *CmpC, *AddC;
if (match(U, m_Add(m_Specific(A), m_APInt(AddC))) &&
match(B, m_APInt(CmpC)) && *AddC == -(*CmpC)) {
Sub = cast<BinaryOperator>(U);
if (!Sub)
return false;
if (!TLI->shouldFormOverflowOp(ISD::USUBO,
TLI->getValueType(*DL, Sub->getType()),
return false;
if (!replaceMathCmpWithIntrinsic(Sub, Sub->getOperand(0), Sub->getOperand(1),
Cmp, Intrinsic::usub_with_overflow))
return false;
// Reset callers - do not crash by iterating over a dead instruction.
ModifiedDT = true;
return true;
/// Sink the given CmpInst into user blocks to reduce the number of virtual
/// registers that must be created and coalesced. This is a clear win except on
/// targets with multiple condition code registers (PowerPC), where it might
/// lose; some adjustment may be wanted there.
/// Return true if any changes are made.
static bool sinkCmpExpression(CmpInst *Cmp, const TargetLowering &TLI) {
if (TLI.hasMultipleConditionRegisters())
return false;
// Avoid sinking soft-FP comparisons, since this can move them into a loop.
if (TLI.useSoftFloat() && isa<FCmpInst>(Cmp))
return false;
// Only insert a cmp in each block once.
DenseMap<BasicBlock*, CmpInst*> InsertedCmps;
bool MadeChange = false;
for (Value::user_iterator UI = Cmp->user_begin(), E = Cmp->user_end();
UI != E; ) {
Use &TheUse = UI.getUse();
Instruction *User = cast<Instruction>(*UI);
// Preincrement use iterator so we don't invalidate it.
// Don't bother for PHI nodes.
if (isa<PHINode>(User))
// Figure out which BB this cmp is used in.
BasicBlock *UserBB = User->getParent();
BasicBlock *DefBB = Cmp->getParent();
// If this user is in the same block as the cmp, don't change the cmp.
if (UserBB == DefBB) continue;
// If we have already inserted a cmp into this block, use it.
CmpInst *&InsertedCmp = InsertedCmps[UserBB];
if (!InsertedCmp) {
BasicBlock::iterator InsertPt = UserBB->getFirstInsertionPt();
assert(InsertPt != UserBB->end());
InsertedCmp =
CmpInst::Create(Cmp->getOpcode(), Cmp->getPredicate(),
Cmp->getOperand(0), Cmp->getOperand(1), "",
// Propagate the debug info.
// Replace a use of the cmp with a use of the new cmp.
TheUse = InsertedCmp;
MadeChange = true;
// If we removed all uses, nuke the cmp.
if (Cmp->use_empty()) {
MadeChange = true;
return MadeChange;
/// For pattern like:
/// DomCond = icmp sgt/slt CmpOp0, CmpOp1 (might not be in DomBB)
/// ...
/// DomBB:
/// ...
/// br DomCond, TrueBB, CmpBB
/// CmpBB: (with DomBB being the single predecessor)
/// ...
/// Cmp = icmp eq CmpOp0, CmpOp1
/// ...
/// It would use two comparison on targets that lowering of icmp sgt/slt is
/// different from lowering of icmp eq (PowerPC). This function try to convert
/// 'Cmp = icmp eq CmpOp0, CmpOp1' to ' Cmp = icmp slt/sgt CmpOp0, CmpOp1'.
/// After that, DomCond and Cmp can use the same comparison so reduce one
/// comparison.
/// Return true if any changes are made.
static bool foldICmpWithDominatingICmp(CmpInst *Cmp,
const TargetLowering &TLI) {
if (!EnableICMP_EQToICMP_ST && TLI.isEqualityCmpFoldedWithSignedCmp())
return false;
ICmpInst::Predicate Pred = Cmp->getPredicate();
if (Pred != ICmpInst::ICMP_EQ)
return false;
// If icmp eq has users other than BranchInst and SelectInst, converting it to
// icmp slt/sgt would introduce more redundant LLVM IR.
for (User *U : Cmp->users()) {
if (isa<BranchInst>(U))
if (isa<SelectInst>(U) && cast<SelectInst>(U)->getCondition() == Cmp)
return false;
// This is a cheap/incomplete check for dominance - just match a single
// predecessor with a conditional branch.
BasicBlock *CmpBB = Cmp->getParent();
BasicBlock *DomBB = CmpBB->getSinglePredecessor();
if (!DomBB)
return false;
// We want to ensure that the only way control gets to the comparison of
// interest is that a less/greater than comparison on the same operands is
// false.
Value *DomCond;
BasicBlock *TrueBB, *FalseBB;
if (!match(DomBB->getTerminator(), m_Br(m_Value(DomCond), TrueBB, FalseBB)))
return false;
if (CmpBB != FalseBB)
return false;
Value *CmpOp0 = Cmp->getOperand(0), *CmpOp1 = Cmp->getOperand(1);
ICmpInst::Predicate DomPred;
if (!match(DomCond, m_ICmp(DomPred, m_Specific(CmpOp0), m_Specific(CmpOp1))))
return false;
if (DomPred != ICmpInst::ICMP_SGT && DomPred != ICmpInst::ICMP_SLT)
return false;
// Convert the equality comparison to the opposite of the dominating
// comparison and swap the direction for all branch/select users.
// We have conceptually converted:
// Res = (a < b) ? <LT_RES> : (a == b) ? <EQ_RES> : <GT_RES>;
// to
// Res = (a < b) ? <LT_RES> : (a > b) ? <GT_RES> : <EQ_RES>;
// And similarly for branches.
for (User *U : Cmp->users()) {
if (auto *BI = dyn_cast<BranchInst>(U)) {
assert(BI->isConditional() && "Must be conditional");
if (auto *SI = dyn_cast<SelectInst>(U)) {
// Swap operands
llvm_unreachable("Must be a branch or a select");
return true;
bool CodeGenPrepare::optimizeCmp(CmpInst *Cmp, bool &ModifiedDT) {
if (sinkCmpExpression(Cmp, *TLI))
return true;
if (combineToUAddWithOverflow(Cmp, ModifiedDT))
return true;
if (combineToUSubWithOverflow(Cmp, ModifiedDT))
return true;
if (foldICmpWithDominatingICmp(Cmp, *TLI))
return true;
return false;
/// Duplicate and sink the given 'and' instruction into user blocks where it is
/// used in a compare to allow isel to generate better code for targets where
/// this operation can be combined.
/// Return true if any changes are made.
static bool sinkAndCmp0Expression(Instruction *AndI,
const TargetLowering &TLI,
SetOfInstrs &InsertedInsts) {
// Double-check that we're not trying to optimize an instruction that was
// already optimized by some other part of this pass.
assert(!InsertedInsts.count(AndI) &&
"Attempting to optimize already optimized and instruction");
(void) InsertedInsts;
// Nothing to do for single use in same basic block.
if (AndI->hasOneUse() &&
AndI->getParent() == cast<Instruction>(*AndI->user_begin())->getParent())
return false;
// Try to avoid cases where sinking/duplicating is likely to increase register
// pressure.
if (!isa<ConstantInt>(AndI->getOperand(0)) &&
!isa<ConstantInt>(AndI->getOperand(1)) &&
AndI->getOperand(0)->hasOneUse() && AndI->getOperand(1)->hasOneUse())
return false;
for (auto *U : AndI->users()) {
Instruction *User = cast<Instruction>(U);
// Only sink 'and' feeding icmp with 0.
if (!isa<ICmpInst>(User))
return false;
auto *CmpC = dyn_cast<ConstantInt>(User->getOperand(1));
if (!CmpC || !CmpC->isZero())
return false;
if (!TLI.isMaskAndCmp0FoldingBeneficial(*AndI))
return false;
LLVM_DEBUG(dbgs() << "found 'and' feeding only icmp 0;\n");
// Push the 'and' into the same block as the icmp 0. There should only be
// one (icmp (and, 0)) in each block, since CSE/GVN should have removed any
// others, so we don't need to keep track of which BBs we insert into.
for (Value::user_iterator UI = AndI->user_begin(), E = AndI->user_end();
UI != E; ) {
Use &TheUse = UI.getUse();
Instruction *User = cast<Instruction>(*UI);
// Preincrement use iterator so we don't invalidate it.
LLVM_DEBUG(dbgs() << "sinking 'and' use: " << *User << "\n");
// Keep the 'and' in the same place if the use is already in the same block.
Instruction *InsertPt =
User->getParent() == AndI->getParent() ? AndI : User;
Instruction *InsertedAnd =
BinaryOperator::Create(Instruction::And, AndI->getOperand(0),
AndI->getOperand(1), "", InsertPt);
// Propagate the debug info.
// Replace a use of the 'and' with a use of the new 'and'.
TheUse = InsertedAnd;
// We removed all uses, nuke the and.
return true;
/// Check if the candidates could be combined with a shift instruction, which
/// includes:
/// 1. Truncate instruction
/// 2. And instruction and the imm is a mask of the low bits:
/// imm & (imm+1) == 0
static bool isExtractBitsCandidateUse(Instruction *User) {
if (!isa<TruncInst>(User)) {
if (User->getOpcode() != Instruction::And ||
return false;
const APInt &Cimm = cast<ConstantInt>(User->getOperand(1))->getValue();
if ((Cimm & (Cimm + 1)).getBoolValue())
return false;
return true;
/// Sink both shift and truncate instruction to the use of truncate's BB.
static bool
SinkShiftAndTruncate(BinaryOperator *ShiftI, Instruction *User, ConstantInt *CI,
DenseMap<BasicBlock *, BinaryOperator *> &InsertedShifts,
const TargetLowering &TLI, const DataLayout &DL) {
BasicBlock *UserBB = User->getParent();
DenseMap<BasicBlock *, CastInst *> InsertedTruncs;
auto *TruncI = cast<TruncInst>(User);
bool MadeChange = false;
for (Value::user_iterator TruncUI = TruncI->user_begin(),
TruncE = TruncI->user_end();
TruncUI != TruncE;) {
Use &TruncTheUse = TruncUI.getUse();
Instruction *TruncUser = cast<Instruction>(*TruncUI);
// Preincrement use iterator so we don't invalidate it.
int ISDOpcode = TLI.InstructionOpcodeToISD(TruncUser->getOpcode());
if (!ISDOpcode)
// If the use is actually a legal node, there will not be an
// implicit truncate.
// FIXME: always querying the result type is just an
// approximation; some nodes' legality is determined by the
// operand or other means. There's no good way to find out though.
if (TLI.isOperationLegalOrCustom(
ISDOpcode, TLI.getValueType(DL, TruncUser->getType(), true)))
// Don't bother for PHI nodes.
if (isa<PHINode>(TruncUser))
BasicBlock *TruncUserBB = TruncUser->getParent();
if (UserBB == TruncUserBB)
BinaryOperator *&InsertedShift = InsertedShifts[TruncUserBB];
CastInst *&InsertedTrunc = InsertedTruncs[TruncUserBB];
if (!InsertedShift && !InsertedTrunc) {
BasicBlock::iterator InsertPt = TruncUserBB->getFirstInsertionPt();
assert(InsertPt != TruncUserBB->end());
// Sink the shift
if (ShiftI->getOpcode() == Instruction::AShr)
InsertedShift = BinaryOperator::CreateAShr(ShiftI->getOperand(0), CI,
"", &*InsertPt);
InsertedShift = BinaryOperator::CreateLShr(ShiftI->getOperand(0), CI,
"", &*InsertPt);
// Sink the trunc
BasicBlock::iterator TruncInsertPt = TruncUserBB->getFirstInsertionPt();
assert(TruncInsertPt != TruncUserBB->end());
InsertedTrunc = CastInst::Create(TruncI->getOpcode(), InsertedShift,
TruncI->getType(), "", &*TruncInsertPt);
MadeChange = true;
TruncTheUse = InsertedTrunc;
return MadeChange;
/// Sink the shift *right* instruction into user blocks if the uses could
/// potentially be combined with this shift instruction and generate BitExtract
/// instruction. It will only be applied if the architecture supports BitExtract
/// instruction. Here is an example:
/// BB1:
/// %x.extract.shift = lshr i64 %arg1, 32
/// BB2:
/// %x.extract.trunc = trunc i64 %x.extract.shift to i16
/// ==>
/// BB2:
/// %x.extract.shift.1 = lshr i64 %arg1, 32
/// %x.extract.trunc = trunc i64 %x.extract.shift.1 to i16
/// CodeGen will recognize the pattern in BB2 and generate BitExtract
/// instruction.
/// Return true if any changes are made.
static bool OptimizeExtractBits(BinaryOperator *ShiftI, ConstantInt *CI,
const TargetLowering &TLI,
const DataLayout &DL) {
BasicBlock *DefBB = ShiftI->getParent();
/// Only insert instructions in each block once.
DenseMap<BasicBlock *, BinaryOperator *> InsertedShifts;
bool shiftIsLegal = TLI.isTypeLegal(TLI.getValueType(DL, ShiftI->getType()));
bool MadeChange = false;
for (Value::user_iterator UI = ShiftI->user_begin(), E = ShiftI->user_end();
UI != E;) {
Use &TheUse = UI.getUse();
Instruction *User = cast<Instruction>(*UI);
// Preincrement use iterator so we don't invalidate it.
// Don't bother for PHI nodes.
if (isa<PHINode>(User))
if (!isExtractBitsCandidateUse(User))
BasicBlock *UserBB = User->getParent();
if (UserBB == DefBB) {
// If the shift and truncate instruction are in the same BB. The use of
// the truncate(TruncUse) may still introduce another truncate if not
// legal. In this case, we would like to sink both shift and truncate
// instruction to the BB of TruncUse.
// for example:
// BB1:
// i64 shift.result = lshr i64 opnd, imm
// trunc.result = trunc shift.result to i16
// BB2:
// ----> We will have an implicit truncate here if the architecture does
// not have i16 compare.
// cmp i16 trunc.result, opnd2
if (isa<TruncInst>(User) && shiftIsLegal
// If the type of the truncate is legal, no truncate will be
// introduced in other basic blocks.
(!TLI.isTypeLegal(TLI.getValueType(DL, User->getType()))))
MadeChange =
SinkShiftAndTruncate(ShiftI, User, CI, InsertedShifts, TLI, DL);
// If we have already inserted a shift into this block, use it.
BinaryOperator *&InsertedShift = InsertedShifts[UserBB];
if (!InsertedShift) {
BasicBlock::iterator InsertPt = UserBB->getFirstInsertionPt();
assert(InsertPt != UserBB->end());
if (ShiftI->getOpcode() == Instruction::AShr)
InsertedShift = BinaryOperator::CreateAShr(ShiftI->getOperand(0), CI,
"", &*InsertPt);
InsertedShift = BinaryOperator::CreateLShr(ShiftI->getOperand(0), CI,
"", &*InsertPt);
MadeChange = true;
// Replace a use of the shift with a use of the new shift.
TheUse = InsertedShift;
// If we removed all uses, or there are none, nuke the shift.
if (ShiftI->use_empty()) {
MadeChange = true;
return MadeChange;
/// If counting leading or trailing zeros is an expensive operation and a zero
/// input is defined, add a check for zero to avoid calling the intrinsic.
/// We want to transform:
/// %z = call i64 @llvm.cttz.i64(i64 %A, i1 false)
/// into:
/// entry:
/// %cmpz = icmp eq i64 %A, 0
/// br i1 %cmpz, label %cond.end, label %cond.false
/// cond.false:
/// %z = call i64 @llvm.cttz.i64(i64 %A, i1 true)
/// br label %cond.end
/// cond.end:
/// %ctz = phi i64 [ 64, %entry ], [ %z, %cond.false ]
/// If the transform is performed, return true and set ModifiedDT to true.
static bool despeculateCountZeros(IntrinsicInst *CountZeros,
const TargetLowering *TLI,
const DataLayout *DL,
bool &ModifiedDT) {
// If a zero input is undefined, it doesn't make sense to despeculate that.
if (match(CountZeros->getOperand(1), m_One()))
return false;
// If it's cheap to speculate, there's nothing to do.
auto IntrinsicID = CountZeros->getIntrinsicID();
if ((IntrinsicID == Intrinsic::cttz && TLI->isCheapToSpeculateCttz()) ||
(IntrinsicID == Intrinsic::ctlz && TLI->isCheapToSpeculateCtlz()))
return false;
// Only handle legal scalar cases. Anything else requires too much work.
Type *Ty = CountZeros->getType();
unsigned SizeInBits = Ty->getPrimitiveSizeInBits();
if (Ty->isVectorTy() || SizeInBits > DL->getLargestLegalIntTypeSizeInBits())
return false;
// The intrinsic will be sunk behind a compare against zero and branch.
BasicBlock *StartBlock = CountZeros->getParent();
BasicBlock *CallBlock = StartBlock->splitBasicBlock(CountZeros, "cond.false");
// Create another block after the count zero intrinsic. A PHI will be added
// in this block to select the result of the intrinsic or the bit-width
// constant if the input to the intrinsic is zero.
BasicBlock::iterator SplitPt = ++(BasicBlock::iterator(CountZeros));
BasicBlock *EndBlock = CallBlock->splitBasicBlock(SplitPt, "cond.end");
// Set up a builder to create a compare, conditional branch, and PHI.
IRBuilder<> Builder(CountZeros->getContext());
// Replace the unconditional branch that was created by the first split with
// a compare against zero and a conditional branch.
Value *Zero = Constant::getNullValue(Ty);
Value *Cmp = Builder.CreateICmpEQ(CountZeros->getOperand(0), Zero, "cmpz");
Builder.CreateCondBr(Cmp, EndBlock, CallBlock);
// Create a PHI in the end block to select either the output of the intrinsic
// or the bit width of the operand.
PHINode *PN = Builder.CreatePHI(Ty, 2, "ctz");
Value *BitWidth = Builder.getInt(APInt(SizeInBits, SizeInBits));
PN->addIncoming(BitWidth, StartBlock);
PN->addIncoming(CountZeros, CallBlock);
// We are explicitly handling the zero case, so we can set the intrinsic's
// undefined zero argument to 'true'. This will also prevent reprocessing the
// intrinsic; we only despeculate when a zero input is defined.
CountZeros->setArgOperand(1, Builder.getTrue());
ModifiedDT = true;
return true;
bool CodeGenPrepare::optimizeCallInst(CallInst *CI, bool &ModifiedDT) {
BasicBlock *BB = CI->getParent();
// Lower inline assembly if we can.
// If we found an inline asm expession, and if the target knows how to
// lower it to normal LLVM code, do so now.
if (CI->isInlineAsm()) {
if (TLI->ExpandInlineAsm(CI)) {
// Avoid invalidating the iterator.
CurInstIterator = BB->begin();
// Avoid processing instructions out of order, which could cause
// reuse before a value is defined.
return true;
// Sink address computing for memory operands into the block.
if (optimizeInlineAsmInst(CI))
return true;
// Align the pointer arguments to this call if the target thinks it's a good
// idea
unsigned MinSize, PrefAlign;
if (TLI->shouldAlignPointerArgs(CI, MinSize, PrefAlign)) {
for (auto &Arg : CI->arg_operands()) {
// We want to align both objects whose address is used directly and
// objects whose address is used in casts and GEPs, though it only makes
// sense for GEPs if the offset is a multiple of the desired alignment and
// if size - offset meets the size threshold.
if (!Arg->getType()->isPointerTy())
APInt Offset(DL->getIndexSizeInBits(
Value *Val = Arg->stripAndAccumulateInBoundsConstantOffsets(*DL, Offset);
uint64_t Offset2 = Offset.getLimitedValue();
if ((Offset2 & (PrefAlign-1)) != 0)
AllocaInst *AI;
if ((AI = dyn_cast<AllocaInst>(Val)) && AI->getAlignment() < PrefAlign &&
DL->getTypeAllocSize(AI->getAllocatedType()) >= MinSize + Offset2)
// Global variables can only be aligned if they are defined in this
// object (i.e. they are uniquely initialized in this object), and
// over-aligning global variables that have an explicit section is
// forbidden.
GlobalVariable *GV;
if ((GV = dyn_cast<GlobalVariable>(Val)) && GV->canIncreaseAlignment() &&
GV->getPointerAlignment(*DL) < PrefAlign &&
DL->getTypeAllocSize(GV->getValueType()) >=
MinSize + Offset2)
// If this is a memcpy (or similar) then we may be able to improve the
// alignment
if (MemIntrinsic *MI = dyn_cast<MemIntrinsic>(CI)) {
Align DestAlign = getKnownAlignment(MI->getDest(), *DL);
MaybeAlign MIDestAlign = MI->getDestAlign();
if (!MIDestAlign || DestAlign > *MIDestAlign)
if (MemTransferInst *MTI = dyn_cast<MemTransferInst>(MI)) {
MaybeAlign MTISrcAlign = MTI->getSourceAlign();
Align SrcAlign = getKnownAlignment(MTI->getSource(), *DL);
if (!MTISrcAlign || SrcAlign > *MTISrcAlign)
// If we have a cold call site, try to sink addressing computation into the
// cold block. This interacts with our handling for loads and stores to
// ensure that we can fold all uses of a potential addressing computation
// into their uses. TODO: generalize this to work over profiling data
if (CI->hasFnAttr(Attribute::Cold) &&
!OptSize && !llvm::shouldOptimizeForSize(BB, PSI, BFI.get()))
for (auto &Arg : CI->arg_operands()) {
if (!Arg->getType()->isPointerTy())
unsigned AS = Arg->getType()->getPointerAddressSpace();
return optimizeMemoryInst(CI, Arg, Arg->getType(), AS);
IntrinsicInst *II = dyn_cast<IntrinsicInst>(CI);
if (II) {
switch (II->getIntrinsicID()) {
default: break;
case Intrinsic::assume: {
Value *Operand = II->getOperand(0);
// Prune the operand, it's most likely dead.
resetIteratorIfInvalidatedWhileCalling(BB, [&]() {
Operand, TLInfo, nullptr,
[&](Value *V) { removeAllAssertingVHReferences(V); });
return true;
case Intrinsic::experimental_widenable_condition: {
// Give up on future widening oppurtunties so that we can fold away dead
// paths and merge blocks before going into block-local instruction
// selection.
if (II->use_empty()) {
return true;
Constant *RetVal = ConstantInt::getTrue(II->getContext());
resetIteratorIfInvalidatedWhileCalling(BB, [&]() {
replaceAndRecursivelySimplify(CI, RetVal, TLInfo, nullptr);
return true;
case Intrinsic::objectsize:
llvm_unreachable("llvm.objectsize.* should have been lowered already");
case Intrinsic::is_constant:
llvm_unreachable("* should have been lowered already");
case Intrinsic::aarch64_stlxr:
case Intrinsic::aarch64_stxr: {
ZExtInst *ExtVal = dyn_cast<ZExtInst>(CI->getArgOperand(0));
if (!ExtVal || !ExtVal->hasOneUse() ||
ExtVal->getParent() == CI->getParent())
return false;
// Sink a zext feeding stlxr/stxr before it, so it can be folded into it.
// Mark this instruction as "inserted by CGP", so that other
// optimizations don't touch it.
return true;
case Intrinsic::launder_invariant_group:
case Intrinsic::strip_invariant_group: {
Value *ArgVal = II->getArgOperand(0);
auto it = LargeOffsetGEPMap.find(II);
if (it != LargeOffsetGEPMap.end()) {
// Merge entries in LargeOffsetGEPMap to reflect the RAUW.
// Make sure not to have to deal with iterator invalidation
// after possibly adding ArgVal to LargeOffsetGEPMap.
auto GEPs = std::move(it->second);
LargeOffsetGEPMap[ArgVal].append(GEPs.begin(), GEPs.end());
return true;
case Intrinsic::cttz:
case Intrinsic::ctlz:
// If counting zeros is expensive, try to avoid it.
return despeculateCountZeros(II, TLI, DL, ModifiedDT);
case Intrinsic::fshl:
case Intrinsic::fshr:
return optimizeFunnelShift(II);
case Intrinsic::dbg_value:
return fixupDbgValue(II);
case Intrinsic::vscale: {
// If datalayout has no special restrictions on vector data layout,
// replace `llvm.vscale` by an equivalent constant expression
// to benefit from cheap constant propagation.
Type *ScalableVectorTy =
VectorType::get(Type::getInt8Ty(II->getContext()), 1, true);
if (DL->getTypeAllocSize(ScalableVectorTy).getKnownMinSize() == 8) {
auto *Null = Constant::getNullValue(ScalableVectorTy->getPointerTo());
auto *One = ConstantInt::getSigned(II->getType(), 1);
auto *CGep =
ConstantExpr::getGetElementPtr(ScalableVectorTy, Null, One);
II->replaceAllUsesWith(ConstantExpr::getPtrToInt(CGep, II->getType()));
return true;
case Intrinsic::masked_gather:
return optimizeGatherScatterInst(II, II->getArgOperand(0));
case Intrinsic::masked_scatter:
return optimizeGatherScatterInst(II, II->getArgOperand(1));
SmallVector<Value *, 2> PtrOps;
Type *AccessTy;
if (TLI->getAddrModeArguments(II, PtrOps, AccessTy))
while (!PtrOps.empty()) {
Value *PtrVal = PtrOps.pop_back_val();
unsigned AS = PtrVal->getType()->getPointerAddressSpace();
if (optimizeMemoryInst(II, PtrVal, AccessTy, AS))
return true;
// From here on out we're working with named functions.
if (!CI->getCalledFunction()) return false;
// Lower all default uses of _chk calls. This is very similar
// to what InstCombineCalls does, but here we are only lowering calls
// to fortified library functions (e.g. __memcpy_chk) that have the default
// "don't know" as the objectsize. Anything else should be left alone.
FortifiedLibCallSimplifier Simplifier(TLInfo, true);
IRBuilder<> Builder(CI);
if (Value *V = Simplifier.optimizeCall(CI, Builder)) {
return true;
return false;
/// Look for opportunities to duplicate return instructions to the predecessor
/// to enable tail call optimizations. The case it is currently looking for is:
/// @code
/// bb0:
/// %tmp0 = tail call i32 @f0()
/// br label %return
/// bb1:
/// %tmp1 = tail call i32 @f1()
/// br label %return
/// bb2:
/// %tmp2 = tail call i32 @f2()
/// br label %return
/// return:
/// %retval = phi i32 [ %tmp0, %bb0 ], [ %tmp1, %bb1 ], [ %tmp2, %bb2 ]
/// ret i32 %retval
/// @endcode
/// =>
/// @code
/// bb0:
/// %tmp0 = tail call i32 @f0()
/// ret i32 %tmp0
/// bb1:
/// %tmp1 = tail call i32 @f1()
/// ret i32 %tmp1
/// bb2:
/// %tmp2 = tail call i32 @f2()
/// ret i32 %tmp2
/// @endcode
bool CodeGenPrepare::dupRetToEnableTailCallOpts(BasicBlock *BB, bool &ModifiedDT) {
ReturnInst *RetI = dyn_cast<ReturnInst>(BB->getTerminator());
if (!RetI)
return false;
PHINode *PN = nullptr;
ExtractValueInst *EVI = nullptr;
BitCastInst *BCI = nullptr;
Value *V = RetI->getReturnValue();
if (V) {
BCI = dyn_cast<BitCastInst>(V);
if (BCI)
V = BCI->getOperand(0);
EVI = dyn_cast<ExtractValueInst>(V);
if (EVI) {
V = EVI->getOperand(0);
if (!llvm::all_of(EVI->indices(), [](unsigned idx) { return idx == 0; }))
return false;
PN = dyn_cast<PHINode>(V);
if (!PN)
return false;
if (PN && PN->getParent() != BB)
return false;
auto isLifetimeEndOrBitCastFor = [](const Instruction *Inst) {
const BitCastInst *BC = dyn_cast<BitCastInst>(Inst);
if (BC && BC->hasOneUse())
Inst = BC->user_back();
if (const IntrinsicInst *II = dyn_cast<IntrinsicInst>(Inst))
return II->getIntrinsicID() == Intrinsic::lifetime_end;
return false;
// Make sure there are no instructions between the first instruction
// and return.
const Instruction *BI = BB->getFirstNonPHI();
// Skip over debug and the bitcast.
while (isa<DbgInfoIntrinsic>(BI) || BI == BCI || BI == EVI ||
isa<PseudoProbeInst>(BI) || isLifetimeEndOrBitCastFor(BI))
BI = BI->getNextNode();
if (BI != RetI)
return false;
/// Only dup the ReturnInst if the CallInst is likely to be emitted as a tail
/// call.
const Function *F = BB->getParent();
SmallVector<BasicBlock*, 4> TailCallBBs;
if (PN) {
for (unsigned I = 0, E = PN->getNumIncomingValues(); I != E; ++I) {
// Look through bitcasts.
Value *IncomingVal = PN->getIncomingValue(I)->stripPointerCasts();
CallInst *CI = dyn_cast<CallInst>(IncomingVal);
BasicBlock *PredBB = PN->getIncomingBlock(I);
// Make sure the phi value is indeed produced by the tail call.
if (CI && CI->hasOneUse() && CI->getParent() == PredBB &&
TLI->mayBeEmittedAsTailCall(CI) &&
attributesPermitTailCall(F, CI, RetI, *TLI))
} else {
SmallPtrSet<BasicBlock*, 4> VisitedBBs;
for (BasicBlock *Pred : predecessors(BB)) {
if (!VisitedBBs.insert(Pred).second)
if (Instruction *I = Pred->rbegin()->getPrevNonDebugInstruction(true)) {
CallInst *CI = dyn_cast<CallInst>(I);
if (CI && CI->use_empty() && TLI->mayBeEmittedAsTailCall(CI) &&
attributesPermitTailCall(F, CI, RetI, *TLI))
bool Changed = false;
for (auto const &TailCallBB : TailCallBBs) {
// Make sure the call instruction is followed by an unconditional branch to
// the return block.
BranchInst *BI = dyn_cast<BranchInst>(TailCallBB->getTerminator());
if (!BI || !BI->isUnconditional() || BI->getSuccessor(0) != BB)
// Duplicate the return into TailCallBB.
(void)FoldReturnIntoUncondBranch(RetI, BB, TailCallBB);
assert(!VerifyBFIUpdates ||
BFI->getBlockFreq(BB) >= BFI->getBlockFreq(TailCallBB));
(BFI->getBlockFreq(BB) - BFI->getBlockFreq(TailCallBB)).getFrequency());
ModifiedDT = Changed = true;
// If we eliminated all predecessors of the block, delete the block now.
if (Changed && !BB->hasAddressTaken() && pred_empty(BB))
return Changed;
// Memory Optimization
namespace {
/// This is an extended version of TargetLowering::AddrMode
/// which holds actual Value*'s for register values.
struct ExtAddrMode : public TargetLowering::AddrMode {
Value *BaseReg = nullptr;
Value *ScaledReg = nullptr;
Value *OriginalValue = nullptr;
bool InBounds = true;
enum FieldName {
NoField = 0x00,
BaseRegField = 0x01,
BaseGVField = 0x02,
BaseOffsField = 0x04,
ScaledRegField = 0x08,
ScaleField = 0x10,
MultipleFields = 0xff
ExtAddrMode() = default;
void print(raw_ostream &OS) const;
void dump() const;
FieldName compare(const ExtAddrMode &other) {
// First check that the types are the same on each field, as differing types
// is something we can't cope with later on.
if (BaseReg && other.BaseReg &&
BaseReg->getType() != other.BaseReg->getType())
return MultipleFields;
if (BaseGV && other.BaseGV &&
BaseGV->getType() != other.BaseGV->getType())
return MultipleFields;
if (ScaledReg && other.ScaledReg &&
ScaledReg->getType() != other.ScaledReg->getType())
return MultipleFields;
// Conservatively reject 'inbounds' mismatches.
if (InBounds != other.InBounds)
return MultipleFields;
// Check each field to see if it differs.
unsigned Result = NoField;
if (BaseReg != other.BaseReg)
Result |= BaseRegField;
if (BaseGV != other.BaseGV)
Result |= BaseGVField;
if (BaseOffs != other.BaseOffs)
Result |= BaseOffsField;
if (ScaledReg != other.ScaledReg)
Result |= ScaledRegField;
// Don't count 0 as being a different scale, because that actually means
// unscaled (which will already be counted by having no ScaledReg).
if (Scale && other.Scale && Scale != other.Scale)
Result |= ScaleField;
if (countPopulation(Result) > 1)
return MultipleFields;
return static_cast<FieldName>(Result);
// An AddrMode is trivial if it involves no calculation i.e. it is just a base
// with no offset.
bool isTrivial() {
// An AddrMode is (BaseGV + BaseReg + BaseOffs + ScaleReg * Scale) so it is
// trivial if at most one of these terms is nonzero, except that BaseGV and
// BaseReg both being zero actually means a null pointer value, which we
// consider to be 'non-zero' here.
return !BaseOffs && !Scale && !(BaseGV && BaseReg);
Value *GetFieldAsValue(FieldName Field, Type *IntPtrTy) {
switch (Field) {
return nullptr;
case BaseRegField:
return BaseReg;
case BaseGVField:
return BaseGV;
case ScaledRegField:
return ScaledReg;
case BaseOffsField:
return ConstantInt::get(IntPtrTy, BaseOffs);
void SetCombinedField(FieldName Field, Value *V,
const SmallVectorImpl<ExtAddrMode> &AddrModes) {
switch (Field) {
llvm_unreachable("Unhandled fields are expected to be rejected earlier");
case ExtAddrMode::BaseRegField:
BaseReg = V;
case ExtAddrMode::BaseGVField:
// A combined BaseGV is an Instruction, not a GlobalValue, so it goes
// in the BaseReg field.
assert(BaseReg == nullptr);
BaseReg = V;
BaseGV = nullptr;
case ExtAddrMode::ScaledRegField:
ScaledReg = V;
// If we have a mix of scaled and unscaled addrmodes then we want scale
// to be the scale and not zero.
if (!Scale)
for (const ExtAddrMode &AM : AddrModes)
if (AM.Scale) {
Scale = AM.Scale;
case ExtAddrMode::BaseOffsField:
// The offset is no longer a constant, so it goes in ScaledReg with a
// scale of 1.
assert(ScaledReg == nullptr);
ScaledReg = V;
Scale = 1;
BaseOffs = 0;
} // end anonymous namespace
#ifndef NDEBUG
static inline raw_ostream &operator<<(raw_ostream &OS, const ExtAddrMode &AM) {
return OS;
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
void ExtAddrMode::print(raw_ostream &OS) const {
bool NeedPlus = false;
OS << "[";
if (InBounds)
OS << "inbounds ";
if (BaseGV) {
OS << (NeedPlus ? " + " : "")
<< "GV:";
BaseGV->printAsOperand(OS, /*PrintType=*/false);
NeedPlus = true;
if (BaseOffs) {
OS << (NeedPlus ? " + " : "")
<< BaseOffs;
NeedPlus = true;
if (BaseReg) {
OS << (NeedPlus ? " + " : "")
<< "Base:";
BaseReg->printAsOperand(OS, /*PrintType=*/false);
NeedPlus = true;
if (Scale) {