ScalarEvolutionExpander.cpp [plain text]
#include "llvm/Analysis/ScalarEvolutionExpander.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/PatternMatch.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
using namespace llvm;
using namespace PatternMatch;
Value *SCEVExpander::ReuseOrCreateCast(Value *V, Type *Ty,
Instruction::CastOps Op,
BasicBlock::iterator IP) {
BasicBlock::iterator BIP = Builder.GetInsertPoint();
Instruction *Ret = nullptr;
for (User *U : V->users())
if (U->getType() == Ty)
if (CastInst *CI = dyn_cast<CastInst>(U))
if (CI->getOpcode() == Op) {
if (BasicBlock::iterator(CI) != IP || BIP == IP) {
Ret = CastInst::Create(Op, V, Ty, "", &*IP);
Ret->takeName(CI);
CI->replaceAllUsesWith(Ret);
CI->setOperand(0, UndefValue::get(V->getType()));
break;
}
Ret = CI;
break;
}
if (!Ret)
Ret = CastInst::Create(Op, V, Ty, V->getName(), &*IP);
assert(SE.DT.dominates(Ret, &*BIP));
rememberInstruction(Ret);
return Ret;
}
static BasicBlock::iterator findInsertPointAfter(Instruction *I,
BasicBlock *MustDominate) {
BasicBlock::iterator IP = ++I->getIterator();
if (auto *II = dyn_cast<InvokeInst>(I))
IP = II->getNormalDest()->begin();
while (isa<PHINode>(IP))
++IP;
while (IP->isEHPad()) {
if (isa<FuncletPadInst>(IP) || isa<LandingPadInst>(IP)) {
++IP;
} else if (isa<CatchSwitchInst>(IP)) {
IP = MustDominate->getFirstInsertionPt();
} else {
llvm_unreachable("unexpected eh pad!");
}
}
return IP;
}
Value *SCEVExpander::InsertNoopCastOfTo(Value *V, Type *Ty) {
Instruction::CastOps Op = CastInst::getCastOpcode(V, false, Ty, false);
assert((Op == Instruction::BitCast ||
Op == Instruction::PtrToInt ||
Op == Instruction::IntToPtr) &&
"InsertNoopCastOfTo cannot perform non-noop casts!");
assert(SE.getTypeSizeInBits(V->getType()) == SE.getTypeSizeInBits(Ty) &&
"InsertNoopCastOfTo cannot change sizes!");
if (Op == Instruction::BitCast) {
if (V->getType() == Ty)
return V;
if (CastInst *CI = dyn_cast<CastInst>(V)) {
if (CI->getOperand(0)->getType() == Ty)
return CI->getOperand(0);
}
}
if ((Op == Instruction::PtrToInt || Op == Instruction::IntToPtr) &&
SE.getTypeSizeInBits(Ty) == SE.getTypeSizeInBits(V->getType())) {
if (CastInst *CI = dyn_cast<CastInst>(V))
if ((CI->getOpcode() == Instruction::PtrToInt ||
CI->getOpcode() == Instruction::IntToPtr) &&
SE.getTypeSizeInBits(CI->getType()) ==
SE.getTypeSizeInBits(CI->getOperand(0)->getType()))
return CI->getOperand(0);
if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V))
if ((CE->getOpcode() == Instruction::PtrToInt ||
CE->getOpcode() == Instruction::IntToPtr) &&
SE.getTypeSizeInBits(CE->getType()) ==
SE.getTypeSizeInBits(CE->getOperand(0)->getType()))
return CE->getOperand(0);
}
if (Constant *C = dyn_cast<Constant>(V))
return ConstantExpr::getCast(Op, C, Ty);
if (Argument *A = dyn_cast<Argument>(V)) {
BasicBlock::iterator IP = A->getParent()->getEntryBlock().begin();
while ((isa<BitCastInst>(IP) &&
isa<Argument>(cast<BitCastInst>(IP)->getOperand(0)) &&
cast<BitCastInst>(IP)->getOperand(0) != A) ||
isa<DbgInfoIntrinsic>(IP))
++IP;
return ReuseOrCreateCast(A, Ty, Op, IP);
}
Instruction *I = cast<Instruction>(V);
BasicBlock::iterator IP = findInsertPointAfter(I, Builder.GetInsertBlock());
return ReuseOrCreateCast(I, Ty, Op, IP);
}
Value *SCEVExpander::InsertBinop(Instruction::BinaryOps Opcode,
Value *LHS, Value *RHS) {
if (Constant *CLHS = dyn_cast<Constant>(LHS))
if (Constant *CRHS = dyn_cast<Constant>(RHS))
return ConstantExpr::get(Opcode, CLHS, CRHS);
unsigned ScanLimit = 6;
BasicBlock::iterator BlockBegin = Builder.GetInsertBlock()->begin();
BasicBlock::iterator IP = Builder.GetInsertPoint();
if (IP != BlockBegin) {
--IP;
for (; ScanLimit; --IP, --ScanLimit) {
if (isa<DbgInfoIntrinsic>(IP))
ScanLimit++;
if (IP->getOpcode() == (unsigned)Opcode && IP->getOperand(0) == LHS &&
IP->getOperand(1) == RHS)
return &*IP;
if (IP == BlockBegin) break;
}
}
DebugLoc Loc = Builder.GetInsertPoint()->getDebugLoc();
BuilderType::InsertPointGuard Guard(Builder);
while (const Loop *L = SE.LI.getLoopFor(Builder.GetInsertBlock())) {
if (!L->isLoopInvariant(LHS) || !L->isLoopInvariant(RHS)) break;
BasicBlock *Preheader = L->getLoopPreheader();
if (!Preheader) break;
Builder.SetInsertPoint(Preheader->getTerminator());
}
Instruction *BO = cast<Instruction>(Builder.CreateBinOp(Opcode, LHS, RHS));
BO->setDebugLoc(Loc);
rememberInstruction(BO);
return BO;
}
static bool FactorOutConstant(const SCEV *&S, const SCEV *&Remainder,
const SCEV *Factor, ScalarEvolution &SE,
const DataLayout &DL) {
if (Factor->isOne())
return true;
if (S == Factor) {
S = SE.getConstant(S->getType(), 1);
return true;
}
if (const SCEVConstant *C = dyn_cast<SCEVConstant>(S)) {
if (C->isZero())
return true;
if (const SCEVConstant *FC = dyn_cast<SCEVConstant>(Factor)) {
ConstantInt *CI =
ConstantInt::get(SE.getContext(), C->getAPInt().sdiv(FC->getAPInt()));
if (!CI->isZero()) {
const SCEV *Div = SE.getConstant(CI);
S = Div;
Remainder = SE.getAddExpr(
Remainder, SE.getConstant(C->getAPInt().srem(FC->getAPInt())));
return true;
}
}
}
if (const SCEVMulExpr *M = dyn_cast<SCEVMulExpr>(S)) {
const SCEVConstant *FC = cast<SCEVConstant>(Factor);
if (const SCEVConstant *C = dyn_cast<SCEVConstant>(M->getOperand(0)))
if (!C->getAPInt().srem(FC->getAPInt())) {
SmallVector<const SCEV *, 4> NewMulOps(M->op_begin(), M->op_end());
NewMulOps[0] = SE.getConstant(C->getAPInt().sdiv(FC->getAPInt()));
S = SE.getMulExpr(NewMulOps);
return true;
}
}
if (const SCEVAddRecExpr *A = dyn_cast<SCEVAddRecExpr>(S)) {
const SCEV *Step = A->getStepRecurrence(SE);
const SCEV *StepRem = SE.getConstant(Step->getType(), 0);
if (!FactorOutConstant(Step, StepRem, Factor, SE, DL))
return false;
if (!StepRem->isZero())
return false;
const SCEV *Start = A->getStart();
if (!FactorOutConstant(Start, Remainder, Factor, SE, DL))
return false;
S = SE.getAddRecExpr(Start, Step, A->getLoop(),
A->getNoWrapFlags(SCEV::FlagNW));
return true;
}
return false;
}
static void SimplifyAddOperands(SmallVectorImpl<const SCEV *> &Ops,
Type *Ty,
ScalarEvolution &SE) {
unsigned NumAddRecs = 0;
for (unsigned i = Ops.size(); i > 0 && isa<SCEVAddRecExpr>(Ops[i-1]); --i)
++NumAddRecs;
SmallVector<const SCEV *, 8> NoAddRecs(Ops.begin(), Ops.end() - NumAddRecs);
SmallVector<const SCEV *, 8> AddRecs(Ops.end() - NumAddRecs, Ops.end());
const SCEV *Sum = NoAddRecs.empty() ?
SE.getConstant(Ty, 0) :
SE.getAddExpr(NoAddRecs);
Ops.clear();
if (const SCEVAddExpr *Add = dyn_cast<SCEVAddExpr>(Sum))
Ops.append(Add->op_begin(), Add->op_end());
else if (!Sum->isZero())
Ops.push_back(Sum);
Ops.append(AddRecs.begin(), AddRecs.end());
}
static void SplitAddRecs(SmallVectorImpl<const SCEV *> &Ops,
Type *Ty,
ScalarEvolution &SE) {
SmallVector<const SCEV *, 8> AddRecs;
for (unsigned i = 0, e = Ops.size(); i != e; ++i)
while (const SCEVAddRecExpr *A = dyn_cast<SCEVAddRecExpr>(Ops[i])) {
const SCEV *Start = A->getStart();
if (Start->isZero()) break;
const SCEV *Zero = SE.getConstant(Ty, 0);
AddRecs.push_back(SE.getAddRecExpr(Zero,
A->getStepRecurrence(SE),
A->getLoop(),
A->getNoWrapFlags(SCEV::FlagNW)));
if (const SCEVAddExpr *Add = dyn_cast<SCEVAddExpr>(Start)) {
Ops[i] = Zero;
Ops.append(Add->op_begin(), Add->op_end());
e += Add->getNumOperands();
} else {
Ops[i] = Start;
}
}
if (!AddRecs.empty()) {
Ops.append(AddRecs.begin(), AddRecs.end());
SimplifyAddOperands(Ops, Ty, SE);
}
}
Value *SCEVExpander::expandAddToGEP(const SCEV *const *op_begin,
const SCEV *const *op_end,
PointerType *PTy,
Type *Ty,
Value *V) {
Type *OriginalElTy = PTy->getElementType();
Type *ElTy = OriginalElTy;
SmallVector<Value *, 4> GepIndices;
SmallVector<const SCEV *, 8> Ops(op_begin, op_end);
bool AnyNonZeroIndices = false;
SplitAddRecs(Ops, Ty, SE);
Type *IntPtrTy = DL.getIntPtrType(PTy);
for (;;) {
SmallVector<const SCEV *, 8> ScaledOps;
if (ElTy->isSized()) {
const SCEV *ElSize = SE.getSizeOfExpr(IntPtrTy, ElTy);
if (!ElSize->isZero()) {
SmallVector<const SCEV *, 8> NewOps;
for (const SCEV *Op : Ops) {
const SCEV *Remainder = SE.getConstant(Ty, 0);
if (FactorOutConstant(Op, Remainder, ElSize, SE, DL)) {
ScaledOps.push_back(Op);
if (!Remainder->isZero())
NewOps.push_back(Remainder);
AnyNonZeroIndices = true;
} else {
NewOps.push_back(Op);
}
}
if (!ScaledOps.empty()) {
Ops = NewOps;
SimplifyAddOperands(Ops, Ty, SE);
}
}
}
Value *Scaled = ScaledOps.empty() ?
Constant::getNullValue(Ty) :
expandCodeFor(SE.getAddExpr(ScaledOps), Ty);
GepIndices.push_back(Scaled);
while (StructType *STy = dyn_cast<StructType>(ElTy)) {
bool FoundFieldNo = false;
if (STy->getNumElements() == 0) break;
if (Ops.empty())
break;
if (const SCEVConstant *C = dyn_cast<SCEVConstant>(Ops[0]))
if (SE.getTypeSizeInBits(C->getType()) <= 64) {
const StructLayout &SL = *DL.getStructLayout(STy);
uint64_t FullOffset = C->getValue()->getZExtValue();
if (FullOffset < SL.getSizeInBytes()) {
unsigned ElIdx = SL.getElementContainingOffset(FullOffset);
GepIndices.push_back(
ConstantInt::get(Type::getInt32Ty(Ty->getContext()), ElIdx));
ElTy = STy->getTypeAtIndex(ElIdx);
Ops[0] =
SE.getConstant(Ty, FullOffset - SL.getElementOffset(ElIdx));
AnyNonZeroIndices = true;
FoundFieldNo = true;
}
}
if (!FoundFieldNo) {
ElTy = STy->getTypeAtIndex(0u);
GepIndices.push_back(
Constant::getNullValue(Type::getInt32Ty(Ty->getContext())));
}
}
if (ArrayType *ATy = dyn_cast<ArrayType>(ElTy))
ElTy = ATy->getElementType();
else
break;
}
if (!AnyNonZeroIndices) {
V = InsertNoopCastOfTo(V,
Type::getInt8PtrTy(Ty->getContext(), PTy->getAddressSpace()));
assert(!isa<Instruction>(V) ||
SE.DT.dominates(cast<Instruction>(V), &*Builder.GetInsertPoint()));
Value *Idx = expandCodeFor(SE.getAddExpr(Ops), Ty);
if (Constant *CLHS = dyn_cast<Constant>(V))
if (Constant *CRHS = dyn_cast<Constant>(Idx))
return ConstantExpr::getGetElementPtr(Type::getInt8Ty(Ty->getContext()),
CLHS, CRHS);
unsigned ScanLimit = 6;
BasicBlock::iterator BlockBegin = Builder.GetInsertBlock()->begin();
BasicBlock::iterator IP = Builder.GetInsertPoint();
if (IP != BlockBegin) {
--IP;
for (; ScanLimit; --IP, --ScanLimit) {
if (isa<DbgInfoIntrinsic>(IP))
ScanLimit++;
if (IP->getOpcode() == Instruction::GetElementPtr &&
IP->getOperand(0) == V && IP->getOperand(1) == Idx)
return &*IP;
if (IP == BlockBegin) break;
}
}
BuilderType::InsertPointGuard Guard(Builder);
while (const Loop *L = SE.LI.getLoopFor(Builder.GetInsertBlock())) {
if (!L->isLoopInvariant(V) || !L->isLoopInvariant(Idx)) break;
BasicBlock *Preheader = L->getLoopPreheader();
if (!Preheader) break;
Builder.SetInsertPoint(Preheader->getTerminator());
}
Value *GEP = Builder.CreateGEP(Builder.getInt8Ty(), V, Idx, "uglygep");
rememberInstruction(GEP);
return GEP;
}
BuilderType::InsertPoint SaveInsertPt = Builder.saveIP();
while (const Loop *L = SE.LI.getLoopFor(Builder.GetInsertBlock())) {
if (!L->isLoopInvariant(V)) break;
bool AnyIndexNotLoopInvariant =
std::any_of(GepIndices.begin(), GepIndices.end(),
[L](Value *Op) { return !L->isLoopInvariant(Op); });
if (AnyIndexNotLoopInvariant)
break;
BasicBlock *Preheader = L->getLoopPreheader();
if (!Preheader) break;
Builder.SetInsertPoint(Preheader->getTerminator());
}
Value *Casted = V;
if (V->getType() != PTy)
Casted = InsertNoopCastOfTo(Casted, PTy);
Value *GEP = Builder.CreateGEP(OriginalElTy, Casted, GepIndices, "scevgep");
Ops.push_back(SE.getUnknown(GEP));
rememberInstruction(GEP);
Builder.restoreIP(SaveInsertPt);
return expand(SE.getAddExpr(Ops));
}
static const Loop *PickMostRelevantLoop(const Loop *A, const Loop *B,
DominatorTree &DT) {
if (!A) return B;
if (!B) return A;
if (A->contains(B)) return B;
if (B->contains(A)) return A;
if (DT.dominates(A->getHeader(), B->getHeader())) return B;
if (DT.dominates(B->getHeader(), A->getHeader())) return A;
return A; }
const Loop *SCEVExpander::getRelevantLoop(const SCEV *S) {
auto Pair = RelevantLoops.insert(std::make_pair(S, nullptr));
if (!Pair.second)
return Pair.first->second;
if (isa<SCEVConstant>(S))
return nullptr;
if (const SCEVUnknown *U = dyn_cast<SCEVUnknown>(S)) {
if (const Instruction *I = dyn_cast<Instruction>(U->getValue()))
return Pair.first->second = SE.LI.getLoopFor(I->getParent());
return nullptr;
}
if (const SCEVNAryExpr *N = dyn_cast<SCEVNAryExpr>(S)) {
const Loop *L = nullptr;
if (const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(S))
L = AR->getLoop();
for (const SCEV *Op : N->operands())
L = PickMostRelevantLoop(L, getRelevantLoop(Op), SE.DT);
return RelevantLoops[N] = L;
}
if (const SCEVCastExpr *C = dyn_cast<SCEVCastExpr>(S)) {
const Loop *Result = getRelevantLoop(C->getOperand());
return RelevantLoops[C] = Result;
}
if (const SCEVUDivExpr *D = dyn_cast<SCEVUDivExpr>(S)) {
const Loop *Result = PickMostRelevantLoop(
getRelevantLoop(D->getLHS()), getRelevantLoop(D->getRHS()), SE.DT);
return RelevantLoops[D] = Result;
}
llvm_unreachable("Unexpected SCEV type!");
}
namespace {
class LoopCompare {
DominatorTree &DT;
public:
explicit LoopCompare(DominatorTree &dt) : DT(dt) {}
bool operator()(std::pair<const Loop *, const SCEV *> LHS,
std::pair<const Loop *, const SCEV *> RHS) const {
if (LHS.second->getType()->isPointerTy() !=
RHS.second->getType()->isPointerTy())
return LHS.second->getType()->isPointerTy();
if (LHS.first != RHS.first)
return PickMostRelevantLoop(LHS.first, RHS.first, DT) != LHS.first;
if (LHS.second->isNonConstantNegative()) {
if (!RHS.second->isNonConstantNegative())
return false;
} else if (RHS.second->isNonConstantNegative())
return true;
return false;
}
};
}
Value *SCEVExpander::visitAddExpr(const SCEVAddExpr *S) {
Type *Ty = SE.getEffectiveSCEVType(S->getType());
SmallVector<std::pair<const Loop *, const SCEV *>, 8> OpsAndLoops;
for (std::reverse_iterator<SCEVAddExpr::op_iterator> I(S->op_end()),
E(S->op_begin()); I != E; ++I)
OpsAndLoops.push_back(std::make_pair(getRelevantLoop(*I), *I));
std::stable_sort(OpsAndLoops.begin(), OpsAndLoops.end(), LoopCompare(SE.DT));
Value *Sum = nullptr;
for (auto I = OpsAndLoops.begin(), E = OpsAndLoops.end(); I != E;) {
const Loop *CurLoop = I->first;
const SCEV *Op = I->second;
if (!Sum) {
Sum = expand(Op);
++I;
} else if (PointerType *PTy = dyn_cast<PointerType>(Sum->getType())) {
SmallVector<const SCEV *, 4> NewOps;
for (; I != E && I->first == CurLoop; ++I) {
const SCEV *X = I->second;
if (const SCEVUnknown *U = dyn_cast<SCEVUnknown>(X))
if (!isa<Instruction>(U->getValue()))
X = SE.getSCEV(U->getValue());
NewOps.push_back(X);
}
Sum = expandAddToGEP(NewOps.begin(), NewOps.end(), PTy, Ty, Sum);
} else if (PointerType *PTy = dyn_cast<PointerType>(Op->getType())) {
SmallVector<const SCEV *, 4> NewOps;
NewOps.push_back(isa<Instruction>(Sum) ? SE.getUnknown(Sum) :
SE.getSCEV(Sum));
for (++I; I != E && I->first == CurLoop; ++I)
NewOps.push_back(I->second);
Sum = expandAddToGEP(NewOps.begin(), NewOps.end(), PTy, Ty, expand(Op));
} else if (Op->isNonConstantNegative()) {
Value *W = expandCodeFor(SE.getNegativeSCEV(Op), Ty);
Sum = InsertNoopCastOfTo(Sum, Ty);
Sum = InsertBinop(Instruction::Sub, Sum, W);
++I;
} else {
Value *W = expandCodeFor(Op, Ty);
Sum = InsertNoopCastOfTo(Sum, Ty);
if (isa<Constant>(Sum)) std::swap(Sum, W);
Sum = InsertBinop(Instruction::Add, Sum, W);
++I;
}
}
return Sum;
}
Value *SCEVExpander::visitMulExpr(const SCEVMulExpr *S) {
Type *Ty = SE.getEffectiveSCEVType(S->getType());
SmallVector<std::pair<const Loop *, const SCEV *>, 8> OpsAndLoops;
for (std::reverse_iterator<SCEVMulExpr::op_iterator> I(S->op_end()),
E(S->op_begin()); I != E; ++I)
OpsAndLoops.push_back(std::make_pair(getRelevantLoop(*I), *I));
std::stable_sort(OpsAndLoops.begin(), OpsAndLoops.end(), LoopCompare(SE.DT));
Value *Prod = nullptr;
for (const auto &I : OpsAndLoops) {
const SCEV *Op = I.second;
if (!Prod) {
Prod = expand(Op);
} else if (Op->isAllOnesValue()) {
Prod = InsertNoopCastOfTo(Prod, Ty);
Prod = InsertBinop(Instruction::Sub, Constant::getNullValue(Ty), Prod);
} else {
Value *W = expandCodeFor(Op, Ty);
Prod = InsertNoopCastOfTo(Prod, Ty);
if (isa<Constant>(Prod)) std::swap(Prod, W);
const APInt *RHS;
if (match(W, m_Power2(RHS))) {
assert(!Ty->isVectorTy() && "vector types are not SCEVable");
Prod = InsertBinop(Instruction::Shl, Prod,
ConstantInt::get(Ty, RHS->logBase2()));
} else {
Prod = InsertBinop(Instruction::Mul, Prod, W);
}
}
}
return Prod;
}
Value *SCEVExpander::visitUDivExpr(const SCEVUDivExpr *S) {
Type *Ty = SE.getEffectiveSCEVType(S->getType());
Value *LHS = expandCodeFor(S->getLHS(), Ty);
if (const SCEVConstant *SC = dyn_cast<SCEVConstant>(S->getRHS())) {
const APInt &RHS = SC->getAPInt();
if (RHS.isPowerOf2())
return InsertBinop(Instruction::LShr, LHS,
ConstantInt::get(Ty, RHS.logBase2()));
}
Value *RHS = expandCodeFor(S->getRHS(), Ty);
return InsertBinop(Instruction::UDiv, LHS, RHS);
}
static void ExposePointerBase(const SCEV *&Base, const SCEV *&Rest,
ScalarEvolution &SE) {
while (const SCEVAddRecExpr *A = dyn_cast<SCEVAddRecExpr>(Base)) {
Base = A->getStart();
Rest = SE.getAddExpr(Rest,
SE.getAddRecExpr(SE.getConstant(A->getType(), 0),
A->getStepRecurrence(SE),
A->getLoop(),
A->getNoWrapFlags(SCEV::FlagNW)));
}
if (const SCEVAddExpr *A = dyn_cast<SCEVAddExpr>(Base)) {
Base = A->getOperand(A->getNumOperands()-1);
SmallVector<const SCEV *, 8> NewAddOps(A->op_begin(), A->op_end());
NewAddOps.back() = Rest;
Rest = SE.getAddExpr(NewAddOps);
ExposePointerBase(Base, Rest, SE);
}
}
bool SCEVExpander::isNormalAddRecExprPHI(PHINode *PN, Instruction *IncV,
const Loop *L) {
if (IncV->getNumOperands() == 0 || isa<PHINode>(IncV) ||
(isa<CastInst>(IncV) && !isa<BitCastInst>(IncV)))
return false;
if (L == IVIncInsertLoop) {
for (User::op_iterator OI = IncV->op_begin()+1,
OE = IncV->op_end(); OI != OE; ++OI)
if (Instruction *OInst = dyn_cast<Instruction>(OI))
if (!SE.DT.dominates(OInst, IVIncInsertPos))
return false;
}
IncV = dyn_cast<Instruction>(IncV->getOperand(0));
if (!IncV)
return false;
if (IncV->mayHaveSideEffects())
return false;
if (IncV != PN)
return true;
return isNormalAddRecExprPHI(PN, IncV, L);
}
Instruction *SCEVExpander::getIVIncOperand(Instruction *IncV,
Instruction *InsertPos,
bool allowScale) {
if (IncV == InsertPos)
return nullptr;
switch (IncV->getOpcode()) {
default:
return nullptr;
case Instruction::Add:
case Instruction::Sub: {
Instruction *OInst = dyn_cast<Instruction>(IncV->getOperand(1));
if (!OInst || SE.DT.dominates(OInst, InsertPos))
return dyn_cast<Instruction>(IncV->getOperand(0));
return nullptr;
}
case Instruction::BitCast:
return dyn_cast<Instruction>(IncV->getOperand(0));
case Instruction::GetElementPtr:
for (auto I = IncV->op_begin() + 1, E = IncV->op_end(); I != E; ++I) {
if (isa<Constant>(*I))
continue;
if (Instruction *OInst = dyn_cast<Instruction>(*I)) {
if (!SE.DT.dominates(OInst, InsertPos))
return nullptr;
}
if (allowScale) {
continue;
}
if (IncV->getNumOperands() != 2)
return nullptr;
unsigned AS = cast<PointerType>(IncV->getType())->getAddressSpace();
if (IncV->getType() != Type::getInt1PtrTy(SE.getContext(), AS)
&& IncV->getType() != Type::getInt8PtrTy(SE.getContext(), AS))
return nullptr;
break;
}
return dyn_cast<Instruction>(IncV->getOperand(0));
}
}
bool SCEVExpander::hoistIVInc(Instruction *IncV, Instruction *InsertPos) {
if (SE.DT.dominates(IncV, InsertPos))
return true;
if (isa<PHINode>(InsertPos) ||
!SE.DT.dominates(InsertPos->getParent(), IncV->getParent()))
return false;
if (!SE.LI.movementPreservesLCSSAForm(IncV, InsertPos))
return false;
SmallVector<Instruction*, 4> IVIncs;
for(;;) {
Instruction *Oper = getIVIncOperand(IncV, InsertPos, true);
if (!Oper)
return false;
IVIncs.push_back(IncV);
IncV = Oper;
if (SE.DT.dominates(IncV, InsertPos))
break;
}
for (auto I = IVIncs.rbegin(), E = IVIncs.rend(); I != E; ++I) {
(*I)->moveBefore(InsertPos);
}
return true;
}
bool SCEVExpander::isExpandedAddRecExprPHI(PHINode *PN, Instruction *IncV,
const Loop *L) {
for(Instruction *IVOper = IncV;
(IVOper = getIVIncOperand(IVOper, L->getLoopPreheader()->getTerminator(),
false));) {
if (IVOper == PN)
return true;
}
return false;
}
Value *SCEVExpander::expandIVInc(PHINode *PN, Value *StepV, const Loop *L,
Type *ExpandTy, Type *IntTy,
bool useSubtract) {
Value *IncV;
if (ExpandTy->isPointerTy()) {
PointerType *GEPPtrTy = cast<PointerType>(ExpandTy);
if (!isa<ConstantInt>(StepV))
GEPPtrTy = PointerType::get(Type::getInt1Ty(SE.getContext()),
GEPPtrTy->getAddressSpace());
const SCEV *const StepArray[1] = { SE.getSCEV(StepV) };
IncV = expandAddToGEP(StepArray, StepArray+1, GEPPtrTy, IntTy, PN);
if (IncV->getType() != PN->getType()) {
IncV = Builder.CreateBitCast(IncV, PN->getType());
rememberInstruction(IncV);
}
} else {
IncV = useSubtract ?
Builder.CreateSub(PN, StepV, Twine(IVName) + ".iv.next") :
Builder.CreateAdd(PN, StepV, Twine(IVName) + ".iv.next");
rememberInstruction(IncV);
}
return IncV;
}
static void hoistBeforePos(DominatorTree *DT, Instruction *InstToHoist,
Instruction *Pos, PHINode *LoopPhi) {
do {
if (DT->dominates(InstToHoist, Pos))
break;
InstToHoist->moveBefore(Pos);
Pos = InstToHoist;
InstToHoist = cast<Instruction>(InstToHoist->getOperand(0));
} while (InstToHoist != LoopPhi);
}
static bool canBeCheaplyTransformed(ScalarEvolution &SE,
const SCEVAddRecExpr *Phi,
const SCEVAddRecExpr *Requested,
bool &InvertStep) {
Type *PhiTy = SE.getEffectiveSCEVType(Phi->getType());
Type *RequestedTy = SE.getEffectiveSCEVType(Requested->getType());
if (RequestedTy->getIntegerBitWidth() > PhiTy->getIntegerBitWidth())
return false;
Phi = dyn_cast<SCEVAddRecExpr>(SE.getTruncateOrNoop(Phi, RequestedTy));
if (!Phi)
return false;
if (Phi == Requested) {
InvertStep = false;
return true;
}
if (SE.getAddExpr(Requested->getStart(),
SE.getNegativeSCEV(Requested)) == Phi) {
InvertStep = true;
return true;
}
return false;
}
static bool IsIncrementNSW(ScalarEvolution &SE, const SCEVAddRecExpr *AR) {
if (!isa<IntegerType>(AR->getType()))
return false;
unsigned BitWidth = cast<IntegerType>(AR->getType())->getBitWidth();
Type *WideTy = IntegerType::get(AR->getType()->getContext(), BitWidth * 2);
const SCEV *Step = AR->getStepRecurrence(SE);
const SCEV *OpAfterExtend = SE.getAddExpr(SE.getSignExtendExpr(Step, WideTy),
SE.getSignExtendExpr(AR, WideTy));
const SCEV *ExtendAfterOp =
SE.getSignExtendExpr(SE.getAddExpr(AR, Step), WideTy);
return ExtendAfterOp == OpAfterExtend;
}
static bool IsIncrementNUW(ScalarEvolution &SE, const SCEVAddRecExpr *AR) {
if (!isa<IntegerType>(AR->getType()))
return false;
unsigned BitWidth = cast<IntegerType>(AR->getType())->getBitWidth();
Type *WideTy = IntegerType::get(AR->getType()->getContext(), BitWidth * 2);
const SCEV *Step = AR->getStepRecurrence(SE);
const SCEV *OpAfterExtend = SE.getAddExpr(SE.getZeroExtendExpr(Step, WideTy),
SE.getZeroExtendExpr(AR, WideTy));
const SCEV *ExtendAfterOp =
SE.getZeroExtendExpr(SE.getAddExpr(AR, Step), WideTy);
return ExtendAfterOp == OpAfterExtend;
}
PHINode *
SCEVExpander::getAddRecExprPHILiterally(const SCEVAddRecExpr *Normalized,
const Loop *L,
Type *ExpandTy,
Type *IntTy,
Type *&TruncTy,
bool &InvertStep) {
assert((!IVIncInsertLoop||IVIncInsertPos) && "Uninitialized insert position");
BasicBlock *LatchBlock = L->getLoopLatch();
if (LatchBlock) {
PHINode *AddRecPhiMatch = nullptr;
Instruction *IncV = nullptr;
TruncTy = nullptr;
InvertStep = false;
bool TryNonMatchingSCEV =
IVIncInsertLoop &&
SE.DT.properlyDominates(LatchBlock, IVIncInsertLoop->getHeader());
for (auto &I : *L->getHeader()) {
auto *PN = dyn_cast<PHINode>(&I);
if (!PN || !SE.isSCEVable(PN->getType()))
continue;
const SCEVAddRecExpr *PhiSCEV = dyn_cast<SCEVAddRecExpr>(SE.getSCEV(PN));
if (!PhiSCEV)
continue;
bool IsMatchingSCEV = PhiSCEV == Normalized;
if (!IsMatchingSCEV && !TryNonMatchingSCEV)
continue;
Instruction *TempIncV =
cast<Instruction>(PN->getIncomingValueForBlock(LatchBlock));
if (LSRMode) {
if (!isExpandedAddRecExprPHI(PN, TempIncV, L))
continue;
if (L == IVIncInsertLoop && !hoistIVInc(TempIncV, IVIncInsertPos))
continue;
} else {
if (!isNormalAddRecExprPHI(PN, TempIncV, L))
continue;
}
if (IsMatchingSCEV) {
IncV = TempIncV;
TruncTy = nullptr;
InvertStep = false;
AddRecPhiMatch = PN;
break;
}
if ((!TruncTy || InvertStep) &&
canBeCheaplyTransformed(SE, PhiSCEV, Normalized, InvertStep)) {
AddRecPhiMatch = PN;
IncV = TempIncV;
TruncTy = SE.getEffectiveSCEVType(Normalized->getType());
}
}
if (AddRecPhiMatch) {
if (L == IVIncInsertLoop)
hoistBeforePos(&SE.DT, IncV, IVIncInsertPos, AddRecPhiMatch);
InsertedValues.insert(AddRecPhiMatch);
rememberInstruction(IncV);
return AddRecPhiMatch;
}
}
BuilderType::InsertPointGuard Guard(Builder);
PostIncLoopSet SavedPostIncLoops = PostIncLoops;
PostIncLoops.clear();
Value *StartV =
expandCodeFor(Normalized->getStart(), ExpandTy, &L->getHeader()->front());
assert(!isa<Instruction>(StartV) ||
SE.DT.properlyDominates(cast<Instruction>(StartV)->getParent(),
L->getHeader()));
const SCEV *Step = Normalized->getStepRecurrence(SE);
bool useSubtract = !ExpandTy->isPointerTy() && Step->isNonConstantNegative();
if (useSubtract)
Step = SE.getNegativeSCEV(Step);
Value *StepV = expandCodeFor(Step, IntTy, &L->getHeader()->front());
bool IncrementIsNUW = !useSubtract && IsIncrementNUW(SE, Normalized);
bool IncrementIsNSW = !useSubtract && IsIncrementNSW(SE, Normalized);
BasicBlock *Header = L->getHeader();
Builder.SetInsertPoint(Header, Header->begin());
pred_iterator HPB = pred_begin(Header), HPE = pred_end(Header);
PHINode *PN = Builder.CreatePHI(ExpandTy, std::distance(HPB, HPE),
Twine(IVName) + ".iv");
rememberInstruction(PN);
for (pred_iterator HPI = HPB; HPI != HPE; ++HPI) {
BasicBlock *Pred = *HPI;
if (!L->contains(Pred)) {
PN->addIncoming(StartV, Pred);
continue;
}
Instruction *InsertPos = L == IVIncInsertLoop ?
IVIncInsertPos : Pred->getTerminator();
Builder.SetInsertPoint(InsertPos);
Value *IncV = expandIVInc(PN, StepV, L, ExpandTy, IntTy, useSubtract);
if (isa<OverflowingBinaryOperator>(IncV)) {
if (IncrementIsNUW)
cast<BinaryOperator>(IncV)->setHasNoUnsignedWrap();
if (IncrementIsNSW)
cast<BinaryOperator>(IncV)->setHasNoSignedWrap();
}
PN->addIncoming(IncV, Pred);
}
PostIncLoops = SavedPostIncLoops;
InsertedValues.insert(PN);
return PN;
}
Value *SCEVExpander::expandAddRecExprLiterally(const SCEVAddRecExpr *S) {
Type *STy = S->getType();
Type *IntTy = SE.getEffectiveSCEVType(STy);
const Loop *L = S->getLoop();
const SCEVAddRecExpr *Normalized = S;
if (PostIncLoops.count(L)) {
PostIncLoopSet Loops;
Loops.insert(L);
Normalized = cast<SCEVAddRecExpr>(TransformForPostIncUse(
Normalize, S, nullptr, nullptr, Loops, SE, SE.DT));
}
const SCEV *Start = Normalized->getStart();
const SCEV *PostLoopOffset = nullptr;
if (!SE.properlyDominates(Start, L->getHeader())) {
PostLoopOffset = Start;
Start = SE.getConstant(Normalized->getType(), 0);
Normalized = cast<SCEVAddRecExpr>(
SE.getAddRecExpr(Start, Normalized->getStepRecurrence(SE),
Normalized->getLoop(),
Normalized->getNoWrapFlags(SCEV::FlagNW)));
}
const SCEV *Step = Normalized->getStepRecurrence(SE);
const SCEV *PostLoopScale = nullptr;
if (!SE.dominates(Step, L->getHeader())) {
PostLoopScale = Step;
Step = SE.getConstant(Normalized->getType(), 1);
Normalized =
cast<SCEVAddRecExpr>(SE.getAddRecExpr(
Start, Step, Normalized->getLoop(),
Normalized->getNoWrapFlags(SCEV::FlagNW)));
}
Type *ExpandTy = PostLoopScale ? IntTy : STy;
Type *TruncTy = nullptr;
bool InvertStep = false;
PHINode *PN = getAddRecExprPHILiterally(Normalized, L, ExpandTy, IntTy,
TruncTy, InvertStep);
Value *Result;
if (!PostIncLoops.count(L))
Result = PN;
else {
BasicBlock *LatchBlock = L->getLoopLatch();
assert(LatchBlock && "PostInc mode requires a unique loop latch!");
Result = PN->getIncomingValueForBlock(LatchBlock);
if (isa<Instruction>(Result) &&
!SE.DT.dominates(cast<Instruction>(Result),
&*Builder.GetInsertPoint())) {
bool useSubtract =
!ExpandTy->isPointerTy() && Step->isNonConstantNegative();
if (useSubtract)
Step = SE.getNegativeSCEV(Step);
Value *StepV;
{
BuilderType::InsertPointGuard Guard(Builder);
StepV = expandCodeFor(Step, IntTy, &L->getHeader()->front());
}
Result = expandIVInc(PN, StepV, L, ExpandTy, IntTy, useSubtract);
}
}
if (TruncTy) {
Type *ResTy = Result->getType();
if (ResTy != SE.getEffectiveSCEVType(ResTy))
Result = InsertNoopCastOfTo(Result, SE.getEffectiveSCEVType(ResTy));
if (TruncTy != Result->getType()) {
Result = Builder.CreateTrunc(Result, TruncTy);
rememberInstruction(Result);
}
if (InvertStep) {
Result = Builder.CreateSub(expandCodeFor(Normalized->getStart(), TruncTy),
Result);
rememberInstruction(Result);
}
}
if (PostLoopScale) {
assert(S->isAffine() && "Can't linearly scale non-affine recurrences.");
Result = InsertNoopCastOfTo(Result, IntTy);
Result = Builder.CreateMul(Result,
expandCodeFor(PostLoopScale, IntTy));
rememberInstruction(Result);
}
if (PostLoopOffset) {
if (PointerType *PTy = dyn_cast<PointerType>(ExpandTy)) {
const SCEV *const OffsetArray[1] = { PostLoopOffset };
Result = expandAddToGEP(OffsetArray, OffsetArray+1, PTy, IntTy, Result);
} else {
Result = InsertNoopCastOfTo(Result, IntTy);
Result = Builder.CreateAdd(Result,
expandCodeFor(PostLoopOffset, IntTy));
rememberInstruction(Result);
}
}
return Result;
}
Value *SCEVExpander::visitAddRecExpr(const SCEVAddRecExpr *S) {
if (!CanonicalMode) return expandAddRecExprLiterally(S);
Type *Ty = SE.getEffectiveSCEVType(S->getType());
const Loop *L = S->getLoop();
PHINode *CanonicalIV = nullptr;
if (PHINode *PN = L->getCanonicalInductionVariable())
if (SE.getTypeSizeInBits(PN->getType()) >= SE.getTypeSizeInBits(Ty))
CanonicalIV = PN;
if (CanonicalIV &&
SE.getTypeSizeInBits(CanonicalIV->getType()) >
SE.getTypeSizeInBits(Ty)) {
SmallVector<const SCEV *, 4> NewOps(S->getNumOperands());
for (unsigned i = 0, e = S->getNumOperands(); i != e; ++i)
NewOps[i] = SE.getAnyExtendExpr(S->op_begin()[i], CanonicalIV->getType());
Value *V = expand(SE.getAddRecExpr(NewOps, S->getLoop(),
S->getNoWrapFlags(SCEV::FlagNW)));
BasicBlock::iterator NewInsertPt =
findInsertPointAfter(cast<Instruction>(V), Builder.GetInsertBlock());
V = expandCodeFor(SE.getTruncateExpr(SE.getUnknown(V), Ty), nullptr,
&*NewInsertPt);
return V;
}
if (!S->getStart()->isZero()) {
SmallVector<const SCEV *, 4> NewOps(S->op_begin(), S->op_end());
NewOps[0] = SE.getConstant(Ty, 0);
const SCEV *Rest = SE.getAddRecExpr(NewOps, L,
S->getNoWrapFlags(SCEV::FlagNW));
const SCEV *Base = S->getStart();
const SCEV *RestArray[1] = { Rest };
ExposePointerBase(Base, RestArray[0], SE);
if (PointerType *PTy = dyn_cast<PointerType>(Base->getType())) {
if (!isa<SCEVMulExpr>(Base) && !isa<SCEVUDivExpr>(Base)) {
Value *StartV = expand(Base);
assert(StartV->getType() == PTy && "Pointer type mismatch for GEP!");
return expandAddToGEP(RestArray, RestArray+1, PTy, Ty, StartV);
}
}
return expand(SE.getAddExpr(SE.getUnknown(expand(S->getStart())),
SE.getUnknown(expand(Rest))));
}
if (!CanonicalIV) {
BasicBlock *Header = L->getHeader();
pred_iterator HPB = pred_begin(Header), HPE = pred_end(Header);
CanonicalIV = PHINode::Create(Ty, std::distance(HPB, HPE), "indvar",
&Header->front());
rememberInstruction(CanonicalIV);
SmallSet<BasicBlock *, 4> PredSeen;
Constant *One = ConstantInt::get(Ty, 1);
for (pred_iterator HPI = HPB; HPI != HPE; ++HPI) {
BasicBlock *HP = *HPI;
if (!PredSeen.insert(HP).second) {
CanonicalIV->addIncoming(CanonicalIV->getIncomingValueForBlock(HP), HP);
continue;
}
if (L->contains(HP)) {
Instruction *Add = BinaryOperator::CreateAdd(CanonicalIV, One,
"indvar.next",
HP->getTerminator());
Add->setDebugLoc(HP->getTerminator()->getDebugLoc());
rememberInstruction(Add);
CanonicalIV->addIncoming(Add, HP);
} else {
CanonicalIV->addIncoming(Constant::getNullValue(Ty), HP);
}
}
}
if (S->isAffine() && S->getOperand(1)->isOne()) {
assert(Ty == SE.getEffectiveSCEVType(CanonicalIV->getType()) &&
"IVs with types different from the canonical IV should "
"already have been handled!");
return CanonicalIV;
}
if (S->isAffine()) return
expand(SE.getTruncateOrNoop(
SE.getMulExpr(SE.getUnknown(CanonicalIV),
SE.getNoopOrAnyExtend(S->getOperand(1),
CanonicalIV->getType())),
Ty));
const SCEV *IH = SE.getUnknown(CanonicalIV);
const SCEV *NewS = S;
const SCEV *Ext = SE.getNoopOrAnyExtend(S, CanonicalIV->getType());
if (isa<SCEVAddRecExpr>(Ext))
NewS = Ext;
const SCEV *V = cast<SCEVAddRecExpr>(NewS)->evaluateAtIteration(IH, SE);
const SCEV *T = SE.getTruncateOrNoop(V, Ty);
return expand(T);
}
Value *SCEVExpander::visitTruncateExpr(const SCEVTruncateExpr *S) {
Type *Ty = SE.getEffectiveSCEVType(S->getType());
Value *V = expandCodeFor(S->getOperand(),
SE.getEffectiveSCEVType(S->getOperand()->getType()));
Value *I = Builder.CreateTrunc(V, Ty);
rememberInstruction(I);
return I;
}
Value *SCEVExpander::visitZeroExtendExpr(const SCEVZeroExtendExpr *S) {
Type *Ty = SE.getEffectiveSCEVType(S->getType());
Value *V = expandCodeFor(S->getOperand(),
SE.getEffectiveSCEVType(S->getOperand()->getType()));
Value *I = Builder.CreateZExt(V, Ty);
rememberInstruction(I);
return I;
}
Value *SCEVExpander::visitSignExtendExpr(const SCEVSignExtendExpr *S) {
Type *Ty = SE.getEffectiveSCEVType(S->getType());
Value *V = expandCodeFor(S->getOperand(),
SE.getEffectiveSCEVType(S->getOperand()->getType()));
Value *I = Builder.CreateSExt(V, Ty);
rememberInstruction(I);
return I;
}
Value *SCEVExpander::visitSMaxExpr(const SCEVSMaxExpr *S) {
Value *LHS = expand(S->getOperand(S->getNumOperands()-1));
Type *Ty = LHS->getType();
for (int i = S->getNumOperands()-2; i >= 0; --i) {
if (S->getOperand(i)->getType() != Ty) {
Ty = SE.getEffectiveSCEVType(Ty);
LHS = InsertNoopCastOfTo(LHS, Ty);
}
Value *RHS = expandCodeFor(S->getOperand(i), Ty);
Value *ICmp = Builder.CreateICmpSGT(LHS, RHS);
rememberInstruction(ICmp);
Value *Sel = Builder.CreateSelect(ICmp, LHS, RHS, "smax");
rememberInstruction(Sel);
LHS = Sel;
}
if (LHS->getType() != S->getType())
LHS = InsertNoopCastOfTo(LHS, S->getType());
return LHS;
}
Value *SCEVExpander::visitUMaxExpr(const SCEVUMaxExpr *S) {
Value *LHS = expand(S->getOperand(S->getNumOperands()-1));
Type *Ty = LHS->getType();
for (int i = S->getNumOperands()-2; i >= 0; --i) {
if (S->getOperand(i)->getType() != Ty) {
Ty = SE.getEffectiveSCEVType(Ty);
LHS = InsertNoopCastOfTo(LHS, Ty);
}
Value *RHS = expandCodeFor(S->getOperand(i), Ty);
Value *ICmp = Builder.CreateICmpUGT(LHS, RHS);
rememberInstruction(ICmp);
Value *Sel = Builder.CreateSelect(ICmp, LHS, RHS, "umax");
rememberInstruction(Sel);
LHS = Sel;
}
if (LHS->getType() != S->getType())
LHS = InsertNoopCastOfTo(LHS, S->getType());
return LHS;
}
Value *SCEVExpander::expandCodeFor(const SCEV *SH, Type *Ty,
Instruction *IP) {
assert(IP);
Builder.SetInsertPoint(IP);
return expandCodeFor(SH, Ty);
}
Value *SCEVExpander::expandCodeFor(const SCEV *SH, Type *Ty) {
Value *V = expand(SH);
if (Ty) {
assert(SE.getTypeSizeInBits(Ty) == SE.getTypeSizeInBits(SH->getType()) &&
"non-trivial casts should be done with the SCEVs directly!");
V = InsertNoopCastOfTo(V, Ty);
}
return V;
}
Value *SCEVExpander::expand(const SCEV *S) {
Instruction *InsertPt = &*Builder.GetInsertPoint();
for (Loop *L = SE.LI.getLoopFor(Builder.GetInsertBlock());;
L = L->getParentLoop())
if (SE.isLoopInvariant(S, L)) {
if (!L) break;
if (BasicBlock *Preheader = L->getLoopPreheader())
InsertPt = Preheader->getTerminator();
else {
InsertPt = &*L->getHeader()->getFirstInsertionPt();
}
} else {
if (L && SE.hasComputableLoopEvolution(S, L) && !PostIncLoops.count(L))
InsertPt = &*L->getHeader()->getFirstInsertionPt();
while (InsertPt != Builder.GetInsertPoint()
&& (isInsertedInstruction(InsertPt)
|| isa<DbgInfoIntrinsic>(InsertPt))) {
InsertPt = &*std::next(InsertPt->getIterator());
}
break;
}
auto I = InsertedExpressions.find(std::make_pair(S, InsertPt));
if (I != InsertedExpressions.end())
return I->second;
BuilderType::InsertPointGuard Guard(Builder);
Builder.SetInsertPoint(InsertPt);
Value *V = visit(S);
InsertedExpressions[std::make_pair(S, InsertPt)] = V;
return V;
}
void SCEVExpander::rememberInstruction(Value *I) {
if (!PostIncLoops.empty())
InsertedPostIncValues.insert(I);
else
InsertedValues.insert(I);
}
PHINode *
SCEVExpander::getOrInsertCanonicalInductionVariable(const Loop *L,
Type *Ty) {
assert(Ty->isIntegerTy() && "Can only insert integer induction variables!");
const SCEV *H = SE.getAddRecExpr(SE.getConstant(Ty, 0),
SE.getConstant(Ty, 1), L, SCEV::FlagAnyWrap);
BuilderType::InsertPointGuard Guard(Builder);
PHINode *V =
cast<PHINode>(expandCodeFor(H, nullptr, &L->getHeader()->front()));
return V;
}
unsigned SCEVExpander::replaceCongruentIVs(Loop *L, const DominatorTree *DT,
SmallVectorImpl<WeakVH> &DeadInsts,
const TargetTransformInfo *TTI) {
SmallVector<PHINode*, 8> Phis;
for (auto &I : *L->getHeader()) {
if (auto *PN = dyn_cast<PHINode>(&I))
Phis.push_back(PN);
else
break;
}
if (TTI)
std::sort(Phis.begin(), Phis.end(), [](Value *LHS, Value *RHS) {
if (!LHS->getType()->isIntegerTy() || !RHS->getType()->isIntegerTy())
return RHS->getType()->isIntegerTy() && !LHS->getType()->isIntegerTy();
return RHS->getType()->getPrimitiveSizeInBits() <
LHS->getType()->getPrimitiveSizeInBits();
});
unsigned NumElim = 0;
DenseMap<const SCEV *, PHINode *> ExprToIVMap;
for (PHINode *Phi : Phis) {
auto SimplifyPHINode = [&](PHINode *PN) -> Value * {
if (Value *V = SimplifyInstruction(PN, DL, &SE.TLI, &SE.DT, &SE.AC))
return V;
if (!SE.isSCEVable(PN->getType()))
return nullptr;
auto *Const = dyn_cast<SCEVConstant>(SE.getSCEV(PN));
if (!Const)
return nullptr;
return Const->getValue();
};
if (Value *V = SimplifyPHINode(Phi)) {
if (V->getType() != Phi->getType())
continue;
Phi->replaceAllUsesWith(V);
DeadInsts.emplace_back(Phi);
++NumElim;
DEBUG_WITH_TYPE(DebugType, dbgs()
<< "INDVARS: Eliminated constant iv: " << *Phi << '\n');
continue;
}
if (!SE.isSCEVable(Phi->getType()))
continue;
PHINode *&OrigPhiRef = ExprToIVMap[SE.getSCEV(Phi)];
if (!OrigPhiRef) {
OrigPhiRef = Phi;
if (Phi->getType()->isIntegerTy() && TTI
&& TTI->isTruncateFree(Phi->getType(), Phis.back()->getType())) {
const SCEV *TruncExpr =
SE.getTruncateExpr(SE.getSCEV(Phi), Phis.back()->getType());
ExprToIVMap[TruncExpr] = Phi;
}
continue;
}
if (OrigPhiRef->getType()->isPointerTy() != Phi->getType()->isPointerTy())
continue;
if (BasicBlock *LatchBlock = L->getLoopLatch()) {
Instruction *OrigInc =
cast<Instruction>(OrigPhiRef->getIncomingValueForBlock(LatchBlock));
Instruction *IsomorphicInc =
cast<Instruction>(Phi->getIncomingValueForBlock(LatchBlock));
if (OrigPhiRef->getType() == Phi->getType()
&& !(ChainedPhis.count(Phi)
|| isExpandedAddRecExprPHI(OrigPhiRef, OrigInc, L))
&& (ChainedPhis.count(Phi)
|| isExpandedAddRecExprPHI(Phi, IsomorphicInc, L))) {
std::swap(OrigPhiRef, Phi);
std::swap(OrigInc, IsomorphicInc);
}
const SCEV *TruncExpr = SE.getTruncateOrNoop(SE.getSCEV(OrigInc),
IsomorphicInc->getType());
if (OrigInc != IsomorphicInc && TruncExpr == SE.getSCEV(IsomorphicInc) &&
SE.LI.replacementPreservesLCSSAForm(IsomorphicInc, OrigInc) &&
hoistIVInc(OrigInc, IsomorphicInc)) {
DEBUG_WITH_TYPE(DebugType,
dbgs() << "INDVARS: Eliminated congruent iv.inc: "
<< *IsomorphicInc << '\n');
Value *NewInc = OrigInc;
if (OrigInc->getType() != IsomorphicInc->getType()) {
Instruction *IP = nullptr;
if (PHINode *PN = dyn_cast<PHINode>(OrigInc))
IP = &*PN->getParent()->getFirstInsertionPt();
else
IP = OrigInc->getNextNode();
IRBuilder<> Builder(IP);
Builder.SetCurrentDebugLocation(IsomorphicInc->getDebugLoc());
NewInc = Builder.
CreateTruncOrBitCast(OrigInc, IsomorphicInc->getType(), IVName);
}
IsomorphicInc->replaceAllUsesWith(NewInc);
DeadInsts.emplace_back(IsomorphicInc);
}
}
DEBUG_WITH_TYPE(DebugType, dbgs()
<< "INDVARS: Eliminated congruent iv: " << *Phi << '\n');
++NumElim;
Value *NewIV = OrigPhiRef;
if (OrigPhiRef->getType() != Phi->getType()) {
IRBuilder<> Builder(&*L->getHeader()->getFirstInsertionPt());
Builder.SetCurrentDebugLocation(Phi->getDebugLoc());
NewIV = Builder.CreateTruncOrBitCast(OrigPhiRef, Phi->getType(), IVName);
}
Phi->replaceAllUsesWith(NewIV);
DeadInsts.emplace_back(Phi);
}
return NumElim;
}
Value *SCEVExpander::findExistingExpansion(const SCEV *S,
const Instruction *At, Loop *L) {
using namespace llvm::PatternMatch;
SmallVector<BasicBlock *, 4> ExitingBlocks;
L->getExitingBlocks(ExitingBlocks);
for (BasicBlock *BB : ExitingBlocks) {
ICmpInst::Predicate Pred;
Instruction *LHS, *RHS;
BasicBlock *TrueBB, *FalseBB;
if (!match(BB->getTerminator(),
m_Br(m_ICmp(Pred, m_Instruction(LHS), m_Instruction(RHS)),
TrueBB, FalseBB)))
continue;
if (SE.getSCEV(LHS) == S && SE.DT.dominates(LHS, At))
return LHS;
if (SE.getSCEV(RHS) == S && SE.DT.dominates(RHS, At))
return RHS;
}
return nullptr;
}
bool SCEVExpander::isHighCostExpansionHelper(
const SCEV *S, Loop *L, const Instruction *At,
SmallPtrSetImpl<const SCEV *> &Processed) {
if (At && findExistingExpansion(S, At, L) != nullptr)
return false;
switch (S->getSCEVType()) {
case scUnknown:
case scConstant:
return false;
case scTruncate:
return isHighCostExpansionHelper(cast<SCEVTruncateExpr>(S)->getOperand(),
L, At, Processed);
case scZeroExtend:
return isHighCostExpansionHelper(cast<SCEVZeroExtendExpr>(S)->getOperand(),
L, At, Processed);
case scSignExtend:
return isHighCostExpansionHelper(cast<SCEVSignExtendExpr>(S)->getOperand(),
L, At, Processed);
}
if (!Processed.insert(S).second)
return false;
if (auto *UDivExpr = dyn_cast<SCEVUDivExpr>(S)) {
if (auto *SC = dyn_cast<SCEVConstant>(UDivExpr->getRHS()))
if (SC->getAPInt().isPowerOf2()) {
const DataLayout &DL =
L->getHeader()->getParent()->getParent()->getDataLayout();
unsigned Width = cast<IntegerType>(UDivExpr->getType())->getBitWidth();
return DL.isIllegalInteger(Width);
}
BasicBlock *ExitingBB = L->getExitingBlock();
if (!ExitingBB)
return true;
if (!At)
At = &ExitingBB->back();
if (!findExistingExpansion(
SE.getAddExpr(S, SE.getConstant(S->getType(), 1)), At, L))
return true;
}
if (isa<SCEVSMaxExpr>(S) || isa<SCEVUMaxExpr>(S))
return true;
if (const SCEVNAryExpr *NAry = dyn_cast<SCEVNAryExpr>(S)) {
for (auto *Op : NAry->operands())
if (isHighCostExpansionHelper(Op, L, At, Processed))
return true;
}
return false;
}
Value *SCEVExpander::expandCodeForPredicate(const SCEVPredicate *Pred,
Instruction *IP) {
assert(IP);
switch (Pred->getKind()) {
case SCEVPredicate::P_Union:
return expandUnionPredicate(cast<SCEVUnionPredicate>(Pred), IP);
case SCEVPredicate::P_Equal:
return expandEqualPredicate(cast<SCEVEqualPredicate>(Pred), IP);
}
llvm_unreachable("Unknown SCEV predicate type");
}
Value *SCEVExpander::expandEqualPredicate(const SCEVEqualPredicate *Pred,
Instruction *IP) {
Value *Expr0 = expandCodeFor(Pred->getLHS(), Pred->getLHS()->getType(), IP);
Value *Expr1 = expandCodeFor(Pred->getRHS(), Pred->getRHS()->getType(), IP);
Builder.SetInsertPoint(IP);
auto *I = Builder.CreateICmpNE(Expr0, Expr1, "ident.check");
return I;
}
Value *SCEVExpander::expandUnionPredicate(const SCEVUnionPredicate *Union,
Instruction *IP) {
auto *BoolType = IntegerType::get(IP->getContext(), 1);
Value *Check = ConstantInt::getNullValue(BoolType);
for (auto Pred : Union->getPredicates()) {
auto *NextCheck = expandCodeForPredicate(Pred, IP);
Builder.SetInsertPoint(IP);
Check = Builder.CreateOr(Check, NextCheck);
}
return Check;
}
namespace {
struct SCEVFindUnsafe {
ScalarEvolution &SE;
bool IsUnsafe;
SCEVFindUnsafe(ScalarEvolution &se): SE(se), IsUnsafe(false) {}
bool follow(const SCEV *S) {
if (const SCEVUDivExpr *D = dyn_cast<SCEVUDivExpr>(S)) {
const SCEVConstant *SC = dyn_cast<SCEVConstant>(D->getRHS());
if (!SC || SC->getValue()->isZero()) {
IsUnsafe = true;
return false;
}
}
if (const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(S)) {
const SCEV *Step = AR->getStepRecurrence(SE);
if (!AR->isAffine() && !SE.dominates(Step, AR->getLoop()->getHeader())) {
IsUnsafe = true;
return false;
}
}
return true;
}
bool isDone() const { return IsUnsafe; }
};
}
namespace llvm {
bool isSafeToExpand(const SCEV *S, ScalarEvolution &SE) {
SCEVFindUnsafe Search(SE);
visitAll(S, Search);
return !Search.IsUnsafe;
}
}