Google Git Google Git
Sign in
llvm/llvm-project/6b26457a0ef2994de54606eb5c356cb897cadda2/./llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp
blob: daa4ed4cf4faf9316c73bc15b6edd1967740665c [file]
//===- InstCombineSelect.cpp ----------------------------------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file implements the visitSelect function.
//
//===----------------------------------------------------------------------===//
#include "InstCombineInternal.h"
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/CmpInstAnalysis.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/Loads.h"
#include "llvm/Analysis/OverflowInstAnalysis.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/Analysis/VectorUtils.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/Constant.h"
#include "llvm/IR/ConstantRange.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/FMF.h"
#include "llvm/IR/IRBuilder.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/Operator.h"
#include "llvm/IR/PatternMatch.h"
#include "llvm/IR/ProfDataUtils.h"
#include "llvm/IR/Type.h"
#include "llvm/IR/User.h"
#include "llvm/IR/Value.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/KnownBits.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Transforms/InstCombine/InstCombiner.h"
#include <cassert>
#include <optional>
#include <utility>
#define DEBUG_TYPE "instcombine"
#include "llvm/Transforms/Utils/InstructionWorklist.h"
using namespace llvm;
using namespace PatternMatch;
namespace llvm {
extern cl::opt<bool> ProfcheckDisableMetadataFixes;
}
/// Replace a select operand based on an equality comparison with the identity
/// constant of a binop.
static Instruction *foldSelectBinOpIdentity(SelectInst &Sel,
const TargetLibraryInfo &TLI,
InstCombinerImpl &IC) {
// The select condition must be an equality compare with a constant operand.
Value *X;
Constant *C;
CmpPredicate Pred;
if (!match(Sel.getCondition(), m_Cmp(Pred, m_Value(X), m_Constant(C))))
return nullptr;
bool IsEq;
if (ICmpInst::isEquality(Pred))
IsEq = Pred == ICmpInst::ICMP_EQ;
else if (Pred == FCmpInst::FCMP_OEQ)
IsEq = true;
else if (Pred == FCmpInst::FCMP_UNE)
IsEq = false;
else
return nullptr;
// A select operand must be a binop.
BinaryOperator *BO;
if (!match(Sel.getOperand(IsEq ? 1 : 2), m_BinOp(BO)))
return nullptr;
// For absorbing values, we can fold to the compared value.
bool IsAbsorbingValue = false;
// Last, match the compare variable operand with a binop operand.
Value *Y;
if (BO->isCommutative()) {
// Recognized 0 as an absorbing value for fmul, but we need to be careful
// about the sign. This could be more aggressive, by handling arbitrary sign
// bit operations as long as we know the fmul sign matches (and handling
// arbitrary opcodes).
if (match(BO, m_c_FMul(m_FAbs(m_Specific(X)), m_Value(Y))) &&
match(C, m_AnyZeroFP()) &&
IC.fmulByZeroIsZero(Y, BO->getFastMathFlags(), &Sel))
IsAbsorbingValue = true;
else if (!match(BO, m_c_BinOp(m_Value(Y), m_Specific(X))))
return nullptr;
} else {
if (!match(BO, m_BinOp(m_Value(Y), m_Specific(X))))
return nullptr;
}
// The compare constant must be the identity constant for that binop.
// If this a floating-point compare with 0.0, any zero constant will do.
Type *Ty = BO->getType();
Value *FoldedVal;
if (IsAbsorbingValue) {
FoldedVal = C;
} else {
Constant *IdC = ConstantExpr::getBinOpIdentity(BO->getOpcode(), Ty, true);
if (IdC != C) {
if (!IdC || !CmpInst::isFPPredicate(Pred))
return nullptr;
if (!match(IdC, m_AnyZeroFP()) || !match(C, m_AnyZeroFP()))
return nullptr;
}
// +0.0 compares equal to -0.0, and so it does not behave as required for
// this transform. Bail out if we can not exclude that possibility.
if (const auto *FPO = dyn_cast<FPMathOperator>(BO))
if (!FPO->hasNoSignedZeros() &&
!cannotBeNegativeZero(Y,
IC.getSimplifyQuery().getWithInstruction(&Sel)))
return nullptr;
FoldedVal = Y;
}
// BO = binop Y, X
// S = { select (cmp eq X, C), BO, ? } or { select (cmp ne X, C), ?, BO }
// =>
// S = { select (cmp eq X, C), Y, ? } or { select (cmp ne X, C), ?, Y }
return IC.replaceOperand(Sel, IsEq ? 1 : 2, FoldedVal);
}
/// This folds:
/// select (icmp eq (and X, C1)), TC, FC
/// iff C1 is a power 2 and the difference between TC and FC is a power-of-2.
/// To something like:
/// (shr (and (X, C1)), (log2(C1) - log2(TC-FC))) + FC
/// Or:
/// (shl (and (X, C1)), (log2(TC-FC) - log2(C1))) + FC
/// With some variations depending if FC is larger than TC, or the shift
/// isn't needed, or the bit widths don't match.
static Value *foldSelectICmpAnd(SelectInst &Sel, Value *CondVal, Value *TrueVal,
Value *FalseVal, Value *V, const APInt &AndMask,
bool CreateAnd,
InstCombiner::BuilderTy &Builder) {
const APInt *SelTC, *SelFC;
if (!match(TrueVal, m_APInt(SelTC)) || !match(FalseVal, m_APInt(SelFC)))
return nullptr;
Type *SelType = Sel.getType();
// In general, when both constants are non-zero, we would need an offset to
// replace the select. This would require more instructions than we started
// with. But there's one special-case that we handle here because it can
// simplify/reduce the instructions.
const APInt &TC = *SelTC;
const APInt &FC = *SelFC;
if (!TC.isZero() && !FC.isZero()) {
if (TC.getBitWidth() != AndMask.getBitWidth())
return nullptr;
// If we have to create an 'and', then we must kill the cmp to not
// increase the instruction count.
if (CreateAnd && !CondVal->hasOneUse())
return nullptr;
// (V & AndMaskC) == 0 ? TC : FC --> TC | (V & AndMaskC)
// (V & AndMaskC) == 0 ? TC : FC --> TC ^ (V & AndMaskC)
// (V & AndMaskC) == 0 ? TC : FC --> TC + (V & AndMaskC)
// (V & AndMaskC) == 0 ? TC : FC --> TC - (V & AndMaskC)
Constant *TCC = ConstantInt::get(SelType, TC);
Constant *FCC = ConstantInt::get(SelType, FC);
Constant *MaskC = ConstantInt::get(SelType, AndMask);
for (auto Opc : {Instruction::Or, Instruction::Xor, Instruction::Add,
Instruction::Sub}) {
if (ConstantFoldBinaryOpOperands(Opc, TCC, MaskC, Sel.getDataLayout()) ==
FCC) {
if (CreateAnd)
V = Builder.CreateAnd(V, MaskC);
return Builder.CreateBinOp(Opc, TCC, V);
}
}
return nullptr;
}
// Make sure one of the select arms is a power-of-2.
if (!TC.isPowerOf2() && !FC.isPowerOf2())
return nullptr;
// Determine which shift is needed to transform result of the 'and' into the
// desired result.
const APInt &ValC = !TC.isZero() ? TC : FC;
unsigned ValZeros = ValC.logBase2();
unsigned AndZeros = AndMask.logBase2();
bool ShouldNotVal = !TC.isZero();
bool NeedShift = ValZeros != AndZeros;
bool NeedZExtTrunc =
SelType->getScalarSizeInBits() != V->getType()->getScalarSizeInBits();
// If we would need to create an 'and' + 'shift' + 'xor' + cast to replace
// a 'select' + 'icmp', then this transformation would result in more
// instructions and potentially interfere with other folding.
if (CreateAnd + ShouldNotVal + NeedShift + NeedZExtTrunc >
1 + CondVal->hasOneUse())
return nullptr;
// Insert the 'and' instruction on the input to the truncate.
if (CreateAnd)
V = Builder.CreateAnd(V, ConstantInt::get(V->getType(), AndMask));
// If types don't match, we can still convert the select by introducing a zext
// or a trunc of the 'and'.
if (ValZeros > AndZeros) {
V = Builder.CreateZExtOrTrunc(V, SelType);
V = Builder.CreateShl(V, ValZeros - AndZeros);
} else if (ValZeros < AndZeros) {
V = Builder.CreateLShr(V, AndZeros - ValZeros);
V = Builder.CreateZExtOrTrunc(V, SelType);
} else {
V = Builder.CreateZExtOrTrunc(V, SelType);
}
// Okay, now we know that everything is set up, we just don't know whether we
// have a icmp_ne or icmp_eq and whether the true or false val is the zero.
if (ShouldNotVal)
V = Builder.CreateXor(V, ValC);
return V;
}
/// We want to turn code that looks like this:
/// %C = or %A, %B
/// %D = select %cond, %C, %A
/// into:
/// %C = select %cond, %B, 0
/// %D = or %A, %C
///
/// Assuming that the specified instruction is an operand to the select, return
/// a bitmask indicating which operands of this instruction are foldable if they
/// equal the other incoming value of the select.
static unsigned getSelectFoldableOperands(BinaryOperator *I) {
switch (I->getOpcode()) {
case Instruction::Add:
case Instruction::FAdd:
case Instruction::Mul:
case Instruction::FMul:
case Instruction::And:
case Instruction::Or:
case Instruction::Xor:
return 3; // Can fold through either operand.
case Instruction::Sub: // Can only fold on the amount subtracted.
case Instruction::FSub:
case Instruction::FDiv: // Can only fold on the divisor amount.
case Instruction::Shl: // Can only fold on the shift amount.
case Instruction::LShr:
case Instruction::AShr:
return 1;
default:
return 0; // Cannot fold
}
}
/// We have (select c, TI, FI), and we know that TI and FI have the same opcode.
Instruction *InstCombinerImpl::foldSelectOpOp(SelectInst &SI, Instruction *TI,
Instruction *FI) {
// If this is a cast from the same type, merge.
Value *Cond = SI.getCondition();
Type *CondTy = Cond->getType();
if (TI->getNumOperands() == 1 && TI->isCast()) {
Type *FIOpndTy = FI->getOperand(0)->getType();
if (TI->getOperand(0)->getType() != FIOpndTy)
return nullptr;
// The select condition may be a vector. We may only change the operand
// type if the vector width remains the same (and matches the condition).
if (auto *CondVTy = dyn_cast<VectorType>(CondTy)) {
if (!FIOpndTy->isVectorTy() ||
CondVTy->getElementCount() !=
cast<VectorType>(FIOpndTy)->getElementCount())
return nullptr;
// TODO: If the backend knew how to deal with casts better, we could
// remove this limitation. For now, there's too much potential to create
// worse codegen by promoting the select ahead of size-altering casts
// (PR28160).
//
// Note that ValueTracking's matchSelectPattern() looks through casts
// without checking 'hasOneUse' when it matches min/max patterns, so this
// transform may end up happening anyway.
if (TI->getOpcode() != Instruction::BitCast &&
(!TI->hasOneUse() || !FI->hasOneUse()))
return nullptr;
} else if (!TI->hasOneUse() || !FI->hasOneUse()) {
// TODO: The one-use restrictions for a scalar select could be eased if
// the fold of a select in visitLoadInst() was enhanced to match a pattern
// that includes a cast.
return nullptr;
}
// Fold this by inserting a select from the input values.
Value *NewSI =
Builder.CreateSelect(Cond, TI->getOperand(0), FI->getOperand(0),
SI.getName() + ".v", &SI);
return CastInst::Create(Instruction::CastOps(TI->getOpcode()), NewSI,
TI->getType());
}
Value *OtherOpT, *OtherOpF;
bool MatchIsOpZero;
auto getCommonOp = [&](Instruction *TI, Instruction *FI, bool Commute,
bool Swapped = false) -> Value * {
assert(!(Commute && Swapped) &&
"Commute and Swapped can't set at the same time");
if (!Swapped) {
if (TI->getOperand(0) == FI->getOperand(0)) {
OtherOpT = TI->getOperand(1);
OtherOpF = FI->getOperand(1);
MatchIsOpZero = true;
return TI->getOperand(0);
} else if (TI->getOperand(1) == FI->getOperand(1)) {
OtherOpT = TI->getOperand(0);
OtherOpF = FI->getOperand(0);
MatchIsOpZero = false;
return TI->getOperand(1);
}
}
if (!Commute && !Swapped)
return nullptr;
// If we are allowing commute or swap of operands, then
// allow a cross-operand match. In that case, MatchIsOpZero
// means that TI's operand 0 (FI's operand 1) is the common op.
if (TI->getOperand(0) == FI->getOperand(1)) {
OtherOpT = TI->getOperand(1);
OtherOpF = FI->getOperand(0);
MatchIsOpZero = true;
return TI->getOperand(0);
} else if (TI->getOperand(1) == FI->getOperand(0)) {
OtherOpT = TI->getOperand(0);
OtherOpF = FI->getOperand(1);
MatchIsOpZero = false;
return TI->getOperand(1);
}
return nullptr;
};
if (TI->hasOneUse() || FI->hasOneUse()) {
// Cond ? -X : -Y --> -(Cond ? X : Y)
Value *X, *Y;
if (match(TI, m_FNeg(m_Value(X))) && match(FI, m_FNeg(m_Value(Y)))) {
// Intersect FMF from the fneg instructions and union those with the
// select.
FastMathFlags FMF = TI->getFastMathFlags();
FMF &= FI->getFastMathFlags();
FMF |= SI.getFastMathFlags();
Value *NewSel =
Builder.CreateSelect(Cond, X, Y, SI.getName() + ".v", &SI);
if (auto *NewSelI = dyn_cast<Instruction>(NewSel))
NewSelI->setFastMathFlags(FMF);
Instruction *NewFNeg = UnaryOperator::CreateFNeg(NewSel);
NewFNeg->setFastMathFlags(FMF);
return NewFNeg;
}
// Min/max intrinsic with a common operand can have the common operand
// pulled after the select. This is the same transform as below for binops,
// but specialized for intrinsic matching and without the restrictive uses
// clause.
auto *TII = dyn_cast<IntrinsicInst>(TI);
auto *FII = dyn_cast<IntrinsicInst>(FI);
if (TII && FII && TII->getIntrinsicID() == FII->getIntrinsicID()) {
if (match(TII, m_MaxOrMin(m_Value(), m_Value()))) {
if (Value *MatchOp = getCommonOp(TI, FI, true)) {
Value *NewSel =
Builder.CreateSelect(Cond, OtherOpT, OtherOpF, "minmaxop", &SI);
return CallInst::Create(TII->getCalledFunction(), {NewSel, MatchOp});
}
}
// select c, (ldexp v, e0), (ldexp v, e1) -> ldexp v, (select c, e0, e1)
// select c, (ldexp v0, e), (ldexp v1, e) -> ldexp (select c, v0, v1), e
//
// select c, (ldexp v0, e0), (ldexp v1, e1) ->
// ldexp (select c, v0, v1), (select c, e0, e1)
if (TII->getIntrinsicID() == Intrinsic::ldexp) {
Value *LdexpVal0 = TII->getArgOperand(0);
Value *LdexpExp0 = TII->getArgOperand(1);
Value *LdexpVal1 = FII->getArgOperand(0);
Value *LdexpExp1 = FII->getArgOperand(1);
if (LdexpExp0->getType() == LdexpExp1->getType()) {
FPMathOperator *SelectFPOp = cast<FPMathOperator>(&SI);
FastMathFlags FMF = cast<FPMathOperator>(TII)->getFastMathFlags();
FMF &= cast<FPMathOperator>(FII)->getFastMathFlags();
FMF |= SelectFPOp->getFastMathFlags();
Value *SelectVal = Builder.CreateSelect(Cond, LdexpVal0, LdexpVal1);
Value *SelectExp = Builder.CreateSelect(Cond, LdexpExp0, LdexpExp1);
Value *NewLdexp = Builder.CreateIntrinsic(
TII->getType(), Intrinsic::ldexp, {SelectVal, SelectExp}, FMF);
return replaceInstUsesWith(SI, NewLdexp);
}
}
}
auto CreateCmpSel = [&](std::optional<CmpPredicate> P,
bool Swapped) -> CmpInst * {
if (!P)
return nullptr;
auto *MatchOp = getCommonOp(TI, FI, ICmpInst::isEquality(*P),
ICmpInst::isRelational(*P) && Swapped);
if (!MatchOp)
return nullptr;
Value *NewSel = Builder.CreateSelect(Cond, OtherOpT, OtherOpF,
SI.getName() + ".v", &SI);
return new ICmpInst(MatchIsOpZero ? *P
: ICmpInst::getSwappedCmpPredicate(*P),
MatchOp, NewSel);
};
// icmp with a common operand also can have the common operand
// pulled after the select.
CmpPredicate TPred, FPred;
if (match(TI, m_ICmp(TPred, m_Value(), m_Value())) &&
match(FI, m_ICmp(FPred, m_Value(), m_Value()))) {
if (auto *R =
CreateCmpSel(CmpPredicate::getMatching(TPred, FPred), false))
return R;
if (auto *R =
CreateCmpSel(CmpPredicate::getMatching(
TPred, ICmpInst::getSwappedCmpPredicate(FPred)),
true))
return R;
}
}
// Only handle binary operators (including two-operand getelementptr) with
// one-use here. As with the cast case above, it may be possible to relax the
// one-use constraint, but that needs be examined carefully since it may not
// reduce the total number of instructions.
if (TI->getNumOperands() != 2 || FI->getNumOperands() != 2 ||
!TI->isSameOperationAs(FI) ||
(!isa<BinaryOperator>(TI) && !isa<GetElementPtrInst>(TI)) ||
!TI->hasOneUse() || !FI->hasOneUse())
return nullptr;
// Figure out if the operations have any operands in common.
Value *MatchOp = getCommonOp(TI, FI, TI->isCommutative());
if (!MatchOp)
return nullptr;
// If the select condition is a vector, the operands of the original select's
// operands also must be vectors. This may not be the case for getelementptr
// for example.
if (CondTy->isVectorTy() && (!OtherOpT->getType()->isVectorTy() ||
!OtherOpF->getType()->isVectorTy()))
return nullptr;
// If we are sinking div/rem after a select, we may need to freeze the
// condition because div/rem may induce immediate UB with a poison operand.
// For example, the following transform is not safe if Cond can ever be poison
// because we can replace poison with zero and then we have div-by-zero that
// didn't exist in the original code:
// Cond ? x/y : x/z --> x / (Cond ? y : z)
auto *BO = dyn_cast<BinaryOperator>(TI);
if (BO && BO->isIntDivRem() && !isGuaranteedNotToBePoison(Cond)) {
// A udiv/urem with a common divisor is safe because UB can only occur with
// div-by-zero, and that would be present in the original code.
if (BO->getOpcode() == Instruction::SDiv ||
BO->getOpcode() == Instruction::SRem || MatchIsOpZero)
Cond = Builder.CreateFreeze(Cond);
}
// If we reach here, they do have operations in common.
Value *NewSI = Builder.CreateSelect(Cond, OtherOpT, OtherOpF,
SI.getName() + ".v", &SI);
Value *Op0 = MatchIsOpZero ? MatchOp : NewSI;
Value *Op1 = MatchIsOpZero ? NewSI : MatchOp;
if (auto *BO = dyn_cast<BinaryOperator>(TI)) {
BinaryOperator *NewBO = BinaryOperator::Create(BO->getOpcode(), Op0, Op1);
NewBO->copyIRFlags(TI);
NewBO->andIRFlags(FI);
return NewBO;
}
if (auto *TGEP = dyn_cast<GetElementPtrInst>(TI)) {
auto *FGEP = cast<GetElementPtrInst>(FI);
Type *ElementType = TGEP->getSourceElementType();
return GetElementPtrInst::Create(
ElementType, Op0, Op1, TGEP->getNoWrapFlags() & FGEP->getNoWrapFlags());
}
llvm_unreachable("Expected BinaryOperator or GEP");
return nullptr;
}
/// This transforms patterns of the form:
/// select cond, intrinsic(x, ...), intrinsic(y, ...)
/// into:
/// intrinsic(select cond, x, y, ...)
Instruction *InstCombinerImpl::foldSelectIntrinsic(SelectInst &SI) {
auto *LHSIntrinsic = dyn_cast<IntrinsicInst>(SI.getTrueValue());
if (!LHSIntrinsic)
return nullptr;
auto *RHSIntrinsic = dyn_cast<IntrinsicInst>(SI.getFalseValue());
if (!RHSIntrinsic ||
LHSIntrinsic->getIntrinsicID() != RHSIntrinsic->getIntrinsicID() ||
!LHSIntrinsic->hasOneUse() || !RHSIntrinsic->hasOneUse())
return nullptr;
const Intrinsic::ID IID = LHSIntrinsic->getIntrinsicID();
switch (IID) {
case Intrinsic::abs:
case Intrinsic::cttz:
case Intrinsic::ctlz: {
auto *TZ = cast<ConstantInt>(LHSIntrinsic->getArgOperand(1));
auto *FZ = cast<ConstantInt>(RHSIntrinsic->getArgOperand(1));
Value *TV = LHSIntrinsic->getArgOperand(0);
Value *FV = RHSIntrinsic->getArgOperand(0);
Value *NewSel = Builder.CreateSelect(SI.getCondition(), TV, FV, "", &SI);
Value *NewPoisonFlag = Builder.CreateAnd(TZ, FZ);
Value *NewCall = Builder.CreateBinaryIntrinsic(IID, NewSel, NewPoisonFlag);
return replaceInstUsesWith(SI, NewCall);
}
case Intrinsic::ctpop: {
Value *TV = LHSIntrinsic->getArgOperand(0);
Value *FV = RHSIntrinsic->getArgOperand(0);
Value *NewSel = Builder.CreateSelect(SI.getCondition(), TV, FV, "", &SI);
Value *NewCall = Builder.CreateUnaryIntrinsic(IID, NewSel);
return replaceInstUsesWith(SI, NewCall);
}
default:
return nullptr;
}
}
static bool isSelect01(const APInt &C1I, const APInt &C2I) {
if (!C1I.isZero() && !C2I.isZero()) // One side must be zero.
return false;
return C1I.isOne() || C1I.isAllOnes() || C2I.isOne() || C2I.isAllOnes();
}
/// Try to fold the select into one of the operands to allow further
/// optimization.
Instruction *InstCombinerImpl::foldSelectIntoOp(SelectInst &SI, Value *TrueVal,
Value *FalseVal) {
// See the comment above getSelectFoldableOperands for a description of the
// transformation we are doing here.
auto TryFoldSelectIntoOp = [&](SelectInst &SI, Value *TrueVal,
Value *FalseVal,
bool Swapped) -> Instruction * {
auto *TVI = dyn_cast<BinaryOperator>(TrueVal);
if (!TVI || !TVI->hasOneUse() || isa<Constant>(FalseVal))
return nullptr;
unsigned SFO = getSelectFoldableOperands(TVI);
unsigned OpToFold = 0;
if ((SFO & 1) && FalseVal == TVI->getOperand(0))
OpToFold = 1;
else if ((SFO & 2) && FalseVal == TVI->getOperand(1))
OpToFold = 2;
if (!OpToFold)
return nullptr;
FastMathFlags FMF;
if (const auto *FPO = dyn_cast<FPMathOperator>(&SI))
FMF = FPO->getFastMathFlags();
Constant *C = ConstantExpr::getBinOpIdentity(
TVI->getOpcode(), TVI->getType(), true, FMF.noSignedZeros());
Value *OOp = TVI->getOperand(2 - OpToFold);
// Avoid creating select between 2 constants unless it's selecting
// between 0, 1 and -1.
const APInt *OOpC;
bool OOpIsAPInt = match(OOp, m_APInt(OOpC));
if (isa<Constant>(OOp) &&
(!OOpIsAPInt || !isSelect01(C->getUniqueInteger(), *OOpC)))
return nullptr;
// If the false value is a NaN then we have that the floating point math
// operation in the transformed code may not preserve the exact NaN
// bit-pattern -- e.g. `fadd sNaN, 0.0 -> qNaN`.
// This makes the transformation incorrect since the original program would
// have preserved the exact NaN bit-pattern.
// Avoid the folding if the false value might be a NaN.
if (isa<FPMathOperator>(&SI) &&
!computeKnownFPClass(FalseVal, FMF, fcNan, SQ.getWithInstruction(&SI))
.isKnownNeverNaN())
return nullptr;
Value *NewSel = Builder.CreateSelect(SI.getCondition(), Swapped ? C : OOp,
Swapped ? OOp : C, "", &SI);
if (isa<FPMathOperator>(&SI)) {
FastMathFlags NewSelFMF = FMF;
// We cannot propagate ninf from the original select, because OOp may be
// inf and the flag only guarantees that FalseVal (op OOp) is never
// infinity.
// Examples: -inf + +inf = NaN, -inf - -inf = NaN, 0 * inf = NaN
// Specifically, if the original select has both ninf and nnan, we can
// safely propagate the flag.
// Note: This property holds for fadd, fsub, and fmul, but does not
// hold for fdiv (e.g. A / Inf == 0.0).
bool CanInferFiniteOperandsFromResult =
TVI->getOpcode() == Instruction::FAdd ||
TVI->getOpcode() == Instruction::FSub ||
TVI->getOpcode() == Instruction::FMul;
NewSelFMF.setNoInfs(TVI->hasNoInfs() ||
(CanInferFiniteOperandsFromResult &&
NewSelFMF.noInfs() && NewSelFMF.noNaNs()));
cast<Instruction>(NewSel)->setFastMathFlags(NewSelFMF);
}
NewSel->takeName(TVI);
BinaryOperator *BO =
BinaryOperator::Create(TVI->getOpcode(), FalseVal, NewSel);
BO->copyIRFlags(TVI);
if (isa<FPMathOperator>(&SI)) {
// Merge poison generating flags from the select.
BO->setHasNoNaNs(BO->hasNoNaNs() && FMF.noNaNs());
BO->setHasNoInfs(BO->hasNoInfs() && FMF.noInfs());
// Merge no-signed-zeros flag from the select.
// Otherwise we may produce zeros with different sign.
BO->setHasNoSignedZeros(BO->hasNoSignedZeros() && FMF.noSignedZeros());
}
return BO;
};
if (Instruction *R = TryFoldSelectIntoOp(SI, TrueVal, FalseVal, false))
return R;
if (Instruction *R = TryFoldSelectIntoOp(SI, FalseVal, TrueVal, true))
return R;
return nullptr;
}
static Value *canoncalizeSelectICmpMinMax(const ICmpInst *Cmp, Value *TVal,
Value *FVal,
InstCombiner::BuilderTy &Builder,
const SimplifyQuery &SQ) {
Value *CmpLHS = Cmp->getOperand(0);
Value *CmpRHS = Cmp->getOperand(1);
ICmpInst::Predicate Pred = Cmp->getPredicate();
if (match(FVal, m_Zero())) {
std::swap(TVal, FVal);
Pred = ICmpInst::getInversePredicate(Pred);
}
if (!match(TVal, m_Zero()))
return nullptr;
if (Pred == CmpInst::ICMP_SGT || Pred == CmpInst::ICMP_SGE) {
std::swap(CmpLHS, CmpRHS);
Pred = ICmpInst::getSwappedPredicate(Pred);
}
// Handles:
// (X <= Y) ? 0 : (X - Y)
// (X <= Y) ? (Y - X) : 0
// (X >= Y) ? 0 : (Y - X)
// (X >= Y) ? (X - Y) : 0
if ((Pred == CmpInst::ICMP_SLT || Pred == CmpInst::ICMP_SLE) &&
match(FVal, m_NSWSub(m_Specific(CmpLHS), m_Specific(CmpRHS))) &&
isGuaranteedNotToBeUndef(CmpLHS, SQ.AC, SQ.CtxI, SQ.DT)) {
Value *SMin =
Builder.CreateBinaryIntrinsic(Intrinsic::smin, CmpRHS, CmpLHS);
return Builder.CreateNSWSub(CmpLHS, SMin);
}
return nullptr;
}
/// Try to fold a select to a min/max intrinsic. Many cases are already handled
/// by matchDecomposedSelectPattern but here we handle the cases where more
/// extensive modification of the IR is required.
static Value *foldSelectICmpMinMax(const ICmpInst *Cmp, Value *TVal,
Value *FVal,
InstCombiner::BuilderTy &Builder,
const SimplifyQuery &SQ) {
Value *CmpLHS = Cmp->getOperand(0);
Value *CmpRHS = Cmp->getOperand(1);
ICmpInst::Predicate Pred = Cmp->getPredicate();
if (Value *V = canoncalizeSelectICmpMinMax(Cmp, TVal, FVal, Builder, SQ))
return V;
// (X > Y) ? X : (Y - 1) ==> MIN(X, Y - 1)
// (X < Y) ? X : (Y + 1) ==> MAX(X, Y + 1)
// This transformation is valid when overflow corresponding to the sign of
// the comparison is poison and we must drop the non-matching overflow flag.
if (CmpRHS == TVal) {
std::swap(CmpLHS, CmpRHS);
Pred = CmpInst::getSwappedPredicate(Pred);
}
// TODO: consider handling 'or disjoint' as well, though these would need to
// be converted to 'add' instructions.
if (!(CmpLHS == TVal && isa<Instruction>(FVal)))
return nullptr;
if (Pred == CmpInst::ICMP_SGT &&
match(FVal, m_NSWAdd(m_Specific(CmpRHS), m_One()))) {
cast<Instruction>(FVal)->setHasNoUnsignedWrap(false);
return Builder.CreateBinaryIntrinsic(Intrinsic::smax, TVal, FVal);
}
if (Pred == CmpInst::ICMP_SLT &&
match(FVal, m_NSWAdd(m_Specific(CmpRHS), m_AllOnes()))) {
cast<Instruction>(FVal)->setHasNoUnsignedWrap(false);
return Builder.CreateBinaryIntrinsic(Intrinsic::smin, TVal, FVal);
}
if (Pred == CmpInst::ICMP_UGT &&
match(FVal, m_NUWAdd(m_Specific(CmpRHS), m_One()))) {
cast<Instruction>(FVal)->setHasNoSignedWrap(false);
return Builder.CreateBinaryIntrinsic(Intrinsic::umax, TVal, FVal);
}
// Note: We must use isKnownNonZero here because "sub nuw %x, 1" will be
// canonicalized to "add %x, -1" discarding the nuw flag.
if (Pred == CmpInst::ICMP_ULT &&
match(FVal, m_Add(m_Specific(CmpRHS), m_AllOnes())) &&
isKnownNonZero(CmpRHS, SQ)) {
cast<Instruction>(FVal)->setHasNoSignedWrap(false);
cast<Instruction>(FVal)->setHasNoUnsignedWrap(false);
return Builder.CreateBinaryIntrinsic(Intrinsic::umin, TVal, FVal);
}
return nullptr;
}
/// We want to turn:
/// (select (icmp eq (and X, Y), 0), (and (lshr X, Z), 1), 1)
/// into:
/// zext (icmp ne i32 (and X, (or Y, (shl 1, Z))), 0)
/// Note:
/// Z may be 0 if lshr is missing.
/// Worst-case scenario is that we will replace 5 instructions with 5 different
/// instructions, but we got rid of select.
static Instruction *foldSelectICmpAndAnd(Type *SelType, const Value *Cond,
Value *TVal, Value *FVal,
InstCombiner::BuilderTy &Builder) {
Value *A, *X, *Y, *Z;
CmpPredicate Pred;
unsigned NumReplaced = 1 + Cond->hasOneUse();
if (match(Cond, m_Trunc(m_Value(X)))) {
Y = ConstantInt::get(X->getType(), 1);
Pred = ICmpInst::ICMP_NE;
} else if (match(Cond,
m_ICmp(Pred, m_And(m_Value(X), m_Value(Y)), m_Zero())) &&
ICmpInst::isEquality(Pred)) {
NumReplaced +=
Cond->hasOneUse() && cast<ICmpInst>(Cond)->getOperand(0)->hasOneUse();
} else
return nullptr;
if (Pred == ICmpInst::ICMP_NE)
std::swap(TVal, FVal);
if (!match(FVal, m_One()))
return nullptr;
// The TrueVal has general form of: and %B, 1
if (!match(TVal, m_And(m_Value(A), m_One())))
return nullptr;
APInt BitWidth(SelType->getScalarSizeInBits(),
SelType->getScalarSizeInBits());
auto TValPattern = m_CombineOr(
m_Deferred(X),
m_LShr(m_Deferred(X), m_Value(Z, m_SpecificInt_ICMP_ForbidPoison(
CmpInst::ICMP_ULT, BitWidth))));
if (!match(A, TValPattern)) {
std::swap(X, Y);
if (!match(A, TValPattern))
return nullptr;
}
bool HasShift = A != X;
if (TVal->hasOneUse())
NumReplaced += 1 + (HasShift && A->hasOneUse());
if (NumReplaced < (4u - isa<Constant>(Y)))
return nullptr;
// ((X & Y) == 0) ? ((X >> Z) & 1) : 1 --> (X & (Y | (1 << Z))) != 0
// ((X & Y) == 0) ? (X & 1) : 1 --> (X & (Y | 1)) != 0
Constant *One = ConstantInt::get(SelType, 1);
Value *MaskB = HasShift ? Builder.CreateShl(One, Z) : One;
Value *FullMask = Builder.CreateOr(Y, MaskB);
Value *MaskedX = Builder.CreateAnd(X, FullMask);
Value *ICmpNeZero = Builder.CreateIsNotNull(MaskedX);
return new ZExtInst(ICmpNeZero, SelType);
}
/// We want to turn:
/// (select (icmp eq (and X, C1), 0), 0, (shl [nsw/nuw] X, C2));
/// iff C1 is a mask and the number of its leading zeros is equal to C2
/// into:
/// shl X, C2
static Value *foldSelectICmpAndZeroShl(const ICmpInst *Cmp, Value *TVal,
Value *FVal,
InstCombiner::BuilderTy &Builder) {
CmpPredicate Pred;
Value *AndVal;
if (!match(Cmp, m_ICmp(Pred, m_Value(AndVal), m_Zero())))
return nullptr;
if (Pred == ICmpInst::ICMP_NE) {
Pred = ICmpInst::ICMP_EQ;
std::swap(TVal, FVal);
}
Value *X;
const APInt *C2, *C1;
if (Pred != ICmpInst::ICMP_EQ ||
!match(AndVal, m_And(m_Value(X), m_APInt(C1))) ||
!match(TVal, m_Zero()) || !match(FVal, m_Shl(m_Specific(X), m_APInt(C2))))
return nullptr;
if (!C1->isMask() ||
C1->countLeadingZeros() != static_cast<unsigned>(C2->getZExtValue()))
return nullptr;
auto *FI = dyn_cast<Instruction>(FVal);
if (!FI)
return nullptr;
FI->setHasNoSignedWrap(false);
FI->setHasNoUnsignedWrap(false);
return FVal;
}
/// We want to turn:
/// (select (icmp sgt x, C), lshr (X, Y), ashr (X, Y)); iff C s>= -1
/// (select (icmp slt x, C), ashr (X, Y), lshr (X, Y)); iff C s>= 0
/// into:
/// ashr (X, Y)
static Value *foldSelectICmpLshrAshr(const ICmpInst *IC, Value *TrueVal,
Value *FalseVal,
InstCombiner::BuilderTy &Builder) {
ICmpInst::Predicate Pred = IC->getPredicate();
Value *CmpLHS = IC->getOperand(0);
Value *CmpRHS = IC->getOperand(1);
if (!CmpRHS->getType()->isIntOrIntVectorTy())
return nullptr;
Value *X, *Y;
unsigned Bitwidth = CmpRHS->getType()->getScalarSizeInBits();
if ((Pred != ICmpInst::ICMP_SGT ||
!match(CmpRHS, m_SpecificInt_ICMP(ICmpInst::ICMP_SGE,
APInt::getAllOnes(Bitwidth)))) &&
(Pred != ICmpInst::ICMP_SLT ||
!match(CmpRHS, m_SpecificInt_ICMP(ICmpInst::ICMP_SGE,
APInt::getZero(Bitwidth)))))
return nullptr;
// Canonicalize so that ashr is in FalseVal.
if (Pred == ICmpInst::ICMP_SLT)
std::swap(TrueVal, FalseVal);
if (match(TrueVal, m_LShr(m_Value(X), m_Value(Y))) &&
match(FalseVal, m_AShr(m_Specific(X), m_Specific(Y))) &&
match(CmpLHS, m_Specific(X))) {
const auto *Ashr = cast<Instruction>(FalseVal);
// if lshr is not exact and ashr is, this new ashr must not be exact.
bool IsExact = Ashr->isExact() && cast<Instruction>(TrueVal)->isExact();
return Builder.CreateAShr(X, Y, IC->getName(), IsExact);
}
return nullptr;
}
/// We want to turn:
/// (select (icmp eq (and X, C1), 0), Y, (BinOp Y, C2))
/// into:
/// IF C2 u>= C1
/// (BinOp Y, (shl (and X, C1), C3))
/// ELSE
/// (BinOp Y, (lshr (and X, C1), C3))
/// iff:
/// 0 on the RHS is the identity value (i.e add, xor, shl, etc...)
/// C1 and C2 are both powers of 2
/// where:
/// IF C2 u>= C1
/// C3 = Log(C2) - Log(C1)
/// ELSE
/// C3 = Log(C1) - Log(C2)
///
/// This transform handles cases where:
/// 1. The icmp predicate is inverted
/// 2. The select operands are reversed
/// 3. The magnitude of C2 and C1 are flipped
static Value *foldSelectICmpAndBinOp(Value *CondVal, Value *TrueVal,
Value *FalseVal, Value *V,
const APInt &AndMask, bool CreateAnd,
InstCombiner::BuilderTy &Builder) {
// Only handle integer compares.
if (!TrueVal->getType()->isIntOrIntVectorTy())
return nullptr;
unsigned C1Log = AndMask.logBase2();
Value *Y;
BinaryOperator *BinOp;
const APInt *C2;
bool NeedXor;
if (match(FalseVal, m_BinOp(m_Specific(TrueVal), m_Power2(C2)))) {
Y = TrueVal;
BinOp = cast<BinaryOperator>(FalseVal);
NeedXor = false;
} else if (match(TrueVal, m_BinOp(m_Specific(FalseVal), m_Power2(C2)))) {
Y = FalseVal;
BinOp = cast<BinaryOperator>(TrueVal);
NeedXor = true;
} else {
return nullptr;
}
// Check that 0 on RHS is identity value for this binop.
auto *IdentityC =
ConstantExpr::getBinOpIdentity(BinOp->getOpcode(), BinOp->getType(),
/*AllowRHSConstant*/ true);
if (IdentityC == nullptr || !IdentityC->isNullValue())
return nullptr;
unsigned C2Log = C2->logBase2();
bool NeedShift = C1Log != C2Log;
bool NeedZExtTrunc = Y->getType()->getScalarSizeInBits() !=
V->getType()->getScalarSizeInBits();
// the demanded bits for the created shl make the and redundant
if (AndMask.isOne() && C2->isSignBitSet())
CreateAnd = false;
// Make sure we don't create more instructions than we save.
if ((NeedShift + NeedXor + NeedZExtTrunc + CreateAnd) >
(CondVal->hasOneUse() + BinOp->hasOneUse()))
return nullptr;
if (CreateAnd) {
// Insert the AND instruction on the input to the truncate.
V = Builder.CreateAnd(V, ConstantInt::get(V->getType(), AndMask));
}
if (C2Log > C1Log) {
V = Builder.CreateZExtOrTrunc(V, Y->getType());
V = Builder.CreateShl(V, C2Log - C1Log);
} else if (C1Log > C2Log) {
V = Builder.CreateLShr(V, C1Log - C2Log);
V = Builder.CreateZExtOrTrunc(V, Y->getType());
} else
V = Builder.CreateZExtOrTrunc(V, Y->getType());
if (NeedXor)
V = Builder.CreateXor(V, *C2);
auto *Res = Builder.CreateBinOp(BinOp->getOpcode(), Y, V);
if (auto *BO = dyn_cast<BinaryOperator>(Res))
BO->copyIRFlags(BinOp);
return Res;
}
/// Canonicalize a set or clear of a masked set of constant bits to
/// select-of-constants form.
static Instruction *foldSetClearBits(SelectInst &Sel,
InstCombiner::BuilderTy &Builder) {
Value *Cond = Sel.getCondition();
Value *T = Sel.getTrueValue();
Value *F = Sel.getFalseValue();
Type *Ty = Sel.getType();
Value *X;
const APInt *NotC, *C;
// Cond ? (X & ~C) : (X | C) --> (X & ~C) | (Cond ? 0 : C)
if (match(T, m_And(m_Value(X), m_APInt(NotC))) &&
match(F, m_OneUse(m_Or(m_Specific(X), m_APInt(C)))) && *NotC == ~(*C)) {
Constant *Zero = ConstantInt::getNullValue(Ty);
Constant *OrC = ConstantInt::get(Ty, *C);
Value *NewSel = Builder.CreateSelect(Cond, Zero, OrC, "masksel", &Sel);
return BinaryOperator::CreateOr(T, NewSel);
}
// Cond ? (X | C) : (X & ~C) --> (X & ~C) | (Cond ? C : 0)
if (match(F, m_And(m_Value(X), m_APInt(NotC))) &&
match(T, m_OneUse(m_Or(m_Specific(X), m_APInt(C)))) && *NotC == ~(*C)) {
Constant *Zero = ConstantInt::getNullValue(Ty);
Constant *OrC = ConstantInt::get(Ty, *C);
Value *NewSel = Builder.CreateSelect(Cond, OrC, Zero, "masksel", &Sel);
return BinaryOperator::CreateOr(F, NewSel);
}
return nullptr;
}
// select (x == 0), 0, x * y --> freeze(y) * x
// select (y == 0), 0, x * y --> freeze(x) * y
// select (x == 0), undef, x * y --> freeze(y) * x
// select (x == undef), 0, x * y --> freeze(y) * x
// Usage of mul instead of 0 will make the result more poisonous,
// so the operand that was not checked in the condition should be frozen.
// The latter folding is applied only when a constant compared with x is
// is a vector consisting of 0 and undefs. If a constant compared with x
// is a scalar undefined value or undefined vector then an expression
// should be already folded into a constant.
//
// This also holds all operations such that Op(0) == 0
// e.g. Shl, Umin, etc
static Instruction *foldSelectZeroOrFixedOp(SelectInst &SI,
InstCombinerImpl &IC) {
auto *CondVal = SI.getCondition();
auto *TrueVal = SI.getTrueValue();
auto *FalseVal = SI.getFalseValue();
Value *X, *Y;
CmpPredicate Predicate;
// Assuming that constant compared with zero is not undef (but it may be
// a vector with some undef elements). Otherwise (when a constant is undef)
// the select expression should be already simplified.
if (!match(CondVal, m_ICmp(Predicate, m_Value(X), m_Zero())) ||
!ICmpInst::isEquality(Predicate))
return nullptr;
if (Predicate == ICmpInst::ICMP_NE)
std::swap(TrueVal, FalseVal);
// Check that TrueVal is a constant instead of matching it with m_Zero()
// to handle the case when it is a scalar undef value or a vector containing
// non-zero elements that are masked by undef elements in the compare
// constant.
auto *TrueValC = dyn_cast<Constant>(TrueVal);
if (TrueValC == nullptr || !isa<Instruction>(FalseVal))
return nullptr;
bool FreezeY;
if (match(FalseVal, m_c_Mul(m_Specific(X), m_Value(Y))) ||
match(FalseVal, m_c_And(m_Specific(X), m_Value(Y))) ||
match(FalseVal, m_FShl(m_Specific(X), m_Specific(X), m_Value(Y))) ||
match(FalseVal, m_FShr(m_Specific(X), m_Specific(X), m_Value(Y))) ||
match(FalseVal,
m_c_Intrinsic<Intrinsic::umin>(m_Specific(X), m_Value(Y)))) {
FreezeY = true;
} else if (match(FalseVal, m_IDiv(m_Specific(X), m_Value(Y))) ||
match(FalseVal, m_IRem(m_Specific(X), m_Value(Y)))) {
FreezeY = false;
} else {
return nullptr;
}
auto *ZeroC = cast<Constant>(cast<Instruction>(CondVal)->getOperand(1));
auto *MergedC = Constant::mergeUndefsWith(TrueValC, ZeroC);
// If X is compared with 0 then TrueVal could be either zero or undef.
// m_Zero match vectors containing some undef elements, but for scalars
// m_Undef should be used explicitly.
if (!match(MergedC, m_Zero()) && !match(MergedC, m_Undef()))
return nullptr;
auto *FalseValI = cast<Instruction>(FalseVal);
if (FreezeY) {
auto *FrY = IC.InsertNewInstBefore(new FreezeInst(Y, Y->getName() + ".fr"),
FalseValI->getIterator());
IC.replaceOperand(*FalseValI,
FalseValI->getOperand(0) == Y
? 0
: (FalseValI->getOperand(1) == Y ? 1 : 2),
FrY);
}
return IC.replaceInstUsesWith(SI, FalseValI);
}
/// Transform patterns such as (a > b) ? a - b : 0 into usub.sat(a, b).
/// There are 8 commuted/swapped variants of this pattern.
static Value *
canonicalizeSaturatedSubtractUnsigned(const ICmpInst *ICI, const Value *TrueVal,
const Value *FalseVal,
InstCombiner::BuilderTy &Builder) {
ICmpInst::Predicate Pred = ICI->getPredicate();
Value *A = ICI->getOperand(0);
Value *B = ICI->getOperand(1);
// (b > a) ? 0 : a - b -> (b <= a) ? a - b : 0
// (a == 0) ? 0 : a - 1 -> (a != 0) ? a - 1 : 0
if (match(TrueVal, m_Zero())) {
Pred = ICmpInst::getInversePredicate(Pred);
std::swap(TrueVal, FalseVal);
}
if (!match(FalseVal, m_Zero()))
return nullptr;
// ugt 0 is canonicalized to ne 0 and requires special handling
// (a != 0) ? a + -1 : 0 -> usub.sat(a, 1)
if (Pred == ICmpInst::ICMP_NE) {
if (match(B, m_Zero()) && match(TrueVal, m_Add(m_Specific(A), m_AllOnes())))
return Builder.CreateBinaryIntrinsic(Intrinsic::usub_sat, A,
ConstantInt::get(A->getType(), 1));
return nullptr;
}
if (!ICmpInst::isUnsigned(Pred))
return nullptr;
if (Pred == ICmpInst::ICMP_ULE || Pred == ICmpInst::ICMP_ULT) {
// (b < a) ? a - b : 0 -> (a > b) ? a - b : 0
std::swap(A, B);
Pred = ICmpInst::getSwappedPredicate(Pred);
}
assert((Pred == ICmpInst::ICMP_UGE || Pred == ICmpInst::ICMP_UGT) &&
"Unexpected isUnsigned predicate!");
// Ensure the sub is of the form:
// (a > b) ? a - b : 0 -> usub.sat(a, b)
// (a > b) ? b - a : 0 -> -usub.sat(a, b)
// Checking for both a-b and a+(-b) as a constant.
bool IsNegative = false;
const APInt *C;
if (match(TrueVal, m_Sub(m_Specific(B), m_Specific(A))) ||
(match(A, m_APInt(C)) &&
match(TrueVal, m_Add(m_Specific(B), m_SpecificInt(-*C)))))
IsNegative = true;
else if (!match(TrueVal, m_Sub(m_Specific(A), m_Specific(B))) &&
!(match(B, m_APInt(C)) &&
match(TrueVal, m_Add(m_Specific(A), m_SpecificInt(-*C)))))
return nullptr;
// If we are adding a negate and the sub and icmp are used anywhere else, we
// would end up with more instructions.
if (IsNegative && !TrueVal->hasOneUse() && !ICI->hasOneUse())
return nullptr;
// (a > b) ? a - b : 0 -> usub.sat(a, b)
// (a > b) ? b - a : 0 -> -usub.sat(a, b)
Value *Result = Builder.CreateBinaryIntrinsic(Intrinsic::usub_sat, A, B);
if (IsNegative)
Result = Builder.CreateNeg(Result);
return Result;
}
static Value *
canonicalizeSaturatedSubtractSigned(const ICmpInst *ICI, const Value *TrueVal,
const Value *FalseVal,
InstCombiner::BuilderTy &Builder) {
ICmpInst::Predicate Pred = ICI->getPredicate();
Value *CmpLHS = ICI->getOperand(0);
Value *CmpRHS = ICI->getOperand(1);
// `A != B ? X : Y` --> `A == B ? Y : X`
// This canonicalization allows us to handle more patterns with fewer checks.
if (Pred == ICmpInst::ICMP_NE) {
Pred = ICmpInst::ICMP_EQ;
std::swap(TrueVal, FalseVal);
}
// `A == MIN_INT ? MAX_INT : 0 - A` --> `ssub_sat 0, A`
if (Pred == ICmpInst::ICMP_EQ && match(CmpRHS, m_SignMask()) &&
match(TrueVal, m_MaxSignedValue()) &&
match(FalseVal, m_Neg(m_Specific(CmpLHS)))) {
return Builder.CreateBinaryIntrinsic(
Intrinsic::ssub_sat, ConstantInt::getNullValue(CmpLHS->getType()),
CmpLHS);
}
return nullptr;
}
static Value *canonicalizeSaturatedSubtract(const ICmpInst *ICI,
const Value *TrueVal,
const Value *FalseVal,
InstCombiner::BuilderTy &Builder) {
if (Value *V = canonicalizeSaturatedSubtractUnsigned(ICI, TrueVal, FalseVal,
Builder))
return V;
if (Value *V =
canonicalizeSaturatedSubtractSigned(ICI, TrueVal, FalseVal, Builder))
return V;
return nullptr;
}
static Value *
canonicalizeSaturatedAddUnsigned(ICmpInst *Cmp, Value *TVal, Value *FVal,
InstCombiner::BuilderTy &Builder) {
// Match unsigned saturated add with constant.
Value *Cmp0 = Cmp->getOperand(0);
Value *Cmp1 = Cmp->getOperand(1);
ICmpInst::Predicate Pred = Cmp->getPredicate();
Value *X;
const APInt *C;
// Match unsigned saturated add of 2 variables with an unnecessary 'not'.
// There are 8 commuted variants.
// Canonicalize -1 (saturated result) to true value of the select.
if (match(FVal, m_AllOnes())) {
std::swap(TVal, FVal);
Pred = CmpInst::getInversePredicate(Pred);
}
if (!match(TVal, m_AllOnes()))
return nullptr;
// uge -1 is canonicalized to eq -1 and requires special handling
// (a == -1) ? -1 : a + 1 -> uadd.sat(a, 1)
if (Pred == ICmpInst::ICMP_EQ) {
if (match(FVal, m_Add(m_Specific(Cmp0), m_One())) &&
match(Cmp1, m_AllOnes())) {
return Builder.CreateBinaryIntrinsic(
Intrinsic::uadd_sat, Cmp0, ConstantInt::get(Cmp0->getType(), 1));
}
return nullptr;
}
if ((Pred == ICmpInst::ICMP_UGE || Pred == ICmpInst::ICMP_UGT) &&
match(FVal, m_Add(m_Specific(Cmp0), m_APIntAllowPoison(C))) &&
match(Cmp1, m_SpecificIntAllowPoison(~*C))) {
// (X u> ~C) ? -1 : (X + C) --> uadd.sat(X, C)
// (X u>= ~C)? -1 : (X + C) --> uadd.sat(X, C)
return Builder.CreateBinaryIntrinsic(Intrinsic::uadd_sat, Cmp0,
ConstantInt::get(Cmp0->getType(), *C));
}
// Negative one does not work here because X u> -1 ? -1, X + -1 is not a
// saturated add.
if (Pred == ICmpInst::ICMP_UGT &&
match(FVal, m_Add(m_Specific(Cmp0), m_APIntAllowPoison(C))) &&
match(Cmp1, m_SpecificIntAllowPoison(~*C - 1)) && !C->isAllOnes()) {
// (X u> ~C - 1) ? -1 : (X + C) --> uadd.sat(X, C)
return Builder.CreateBinaryIntrinsic(Intrinsic::uadd_sat, Cmp0,
ConstantInt::get(Cmp0->getType(), *C));
}
// Zero does not work here because X u>= 0 ? -1 : X -> is always -1, which is
// not a saturated add.
if (Pred == ICmpInst::ICMP_UGE &&
match(FVal, m_Add(m_Specific(Cmp0), m_APIntAllowPoison(C))) &&
match(Cmp1, m_SpecificIntAllowPoison(-*C)) && !C->isZero()) {
// (X u >= -C) ? -1 : (X + C) --> uadd.sat(X, C)
return Builder.CreateBinaryIntrinsic(Intrinsic::uadd_sat, Cmp0,
ConstantInt::get(Cmp0->getType(), *C));
}
// Canonicalize predicate to less-than or less-or-equal-than.
if (Pred == ICmpInst::ICMP_UGT || Pred == ICmpInst::ICMP_UGE) {
std::swap(Cmp0, Cmp1);
Pred = CmpInst::getSwappedPredicate(Pred);
}
if (Pred != ICmpInst::ICMP_ULT && Pred != ICmpInst::ICMP_ULE)
return nullptr;
// Match unsigned saturated add of 2 variables with an unnecessary 'not'.
// Strictness of the comparison is irrelevant.
Value *Y;
if (match(Cmp0, m_Not(m_Value(X))) &&
match(FVal, m_c_Add(m_Specific(X), m_Value(Y))) && Y == Cmp1) {
// (~X u< Y) ? -1 : (X + Y) --> uadd.sat(X, Y)
// (~X u< Y) ? -1 : (Y + X) --> uadd.sat(X, Y)
return Builder.CreateBinaryIntrinsic(Intrinsic::uadd_sat, X, Y);
}
// The 'not' op may be included in the sum but not the compare.
// Strictness of the comparison is irrelevant.
X = Cmp0;
Y = Cmp1;
if (match(FVal, m_c_Add(m_NotForbidPoison(m_Specific(X)), m_Specific(Y)))) {
// (X u< Y) ? -1 : (~X + Y) --> uadd.sat(~X, Y)
// (X u< Y) ? -1 : (Y + ~X) --> uadd.sat(Y, ~X)
BinaryOperator *BO = cast<BinaryOperator>(FVal);
return Builder.CreateBinaryIntrinsic(
Intrinsic::uadd_sat, BO->getOperand(0), BO->getOperand(1));
}
// The overflow may be detected via the add wrapping round.
// This is only valid for strict comparison!
if (Pred == ICmpInst::ICMP_ULT &&
match(Cmp0, m_c_Add(m_Specific(Cmp1), m_Value(Y))) &&
match(FVal, m_c_Add(m_Specific(Cmp1), m_Specific(Y)))) {
// ((X + Y) u< X) ? -1 : (X + Y) --> uadd.sat(X, Y)
// ((X + Y) u< Y) ? -1 : (X + Y) --> uadd.sat(X, Y)
return Builder.CreateBinaryIntrinsic(Intrinsic::uadd_sat, Cmp1, Y);
}
return nullptr;
}
static Value *canonicalizeSaturatedAddSigned(ICmpInst *Cmp, Value *TVal,
Value *FVal,
InstCombiner::BuilderTy &Builder) {
// Match saturated add with constant.
Value *Cmp0 = Cmp->getOperand(0);
Value *Cmp1 = Cmp->getOperand(1);
ICmpInst::Predicate Pred = Cmp->getPredicate();
// Canonicalize TVal to be the saturation constant.
if (match(FVal, m_MaxSignedValue()) || match(FVal, m_SignMask())) {
std::swap(TVal, FVal);
Pred = CmpInst::getInversePredicate(Pred);
}
const APInt *SatC;
if (!match(TVal, m_APInt(SatC)) ||
!(SatC->isMaxSignedValue() || SatC->isSignMask()))
return nullptr;
bool IsMax = SatC->isMaxSignedValue();
// sge maximum signed value is canonicalized to eq maximum signed value and
// requires special handling. sle minimum signed value is similarly
// canonicalized to eq minimum signed value.
if (Pred == ICmpInst::ICMP_EQ && Cmp1 == TVal) {
// (a == INT_MAX) ? INT_MAX : a + 1 -> sadd.sat(a, 1)
if (IsMax && match(FVal, m_Add(m_Specific(Cmp0), m_One()))) {
return Builder.CreateBinaryIntrinsic(
Intrinsic::sadd_sat, Cmp0, ConstantInt::get(Cmp0->getType(), 1));
}
// (a == INT_MIN) ? INT_MIN : a + -1 -> sadd.sat(a, -1)
if (!IsMax && match(FVal, m_Add(m_Specific(Cmp0), m_AllOnes()))) {
return Builder.CreateBinaryIntrinsic(
Intrinsic::sadd_sat, Cmp0,
ConstantInt::getAllOnesValue(Cmp0->getType()));
}
return nullptr;
}
const APInt *C;
// (X > Y) ? INT_MAX : (X + C) --> sadd.sat(X, C)
// (X >= Y) ? INT_MAX : (X + C) --> sadd.sat(X, C)
// where C > 0 and Y is INT_MAX - C or INT_MAX - C - 1
if (IsMax && (Pred == ICmpInst::ICMP_SGT || Pred == ICmpInst::ICMP_SGE) &&
isa<Constant>(Cmp1) &&
match(FVal, m_Add(m_Specific(Cmp0), m_StrictlyPositive(C)))) {
// Normalize SGE to SGT for threshold comparison.
if (Pred == ICmpInst::ICMP_SGE) {
if (auto Flipped = getFlippedStrictnessPredicateAndConstant(
Pred, cast<Constant>(Cmp1))) {
Pred = Flipped->first;
Cmp1 = Flipped->second;
}
}
// Check: X > INT_MAX - C or X > INT_MAX - C - 1
APInt Threshold = *SatC - *C;
if (Pred == ICmpInst::ICMP_SGT &&
(match(Cmp1, m_SpecificIntAllowPoison(Threshold)) ||
match(Cmp1, m_SpecificIntAllowPoison(Threshold - 1))))
return Builder.CreateBinaryIntrinsic(
Intrinsic::sadd_sat, Cmp0, ConstantInt::get(Cmp0->getType(), *C));
}
// (X < Y) ? INT_MIN : (X + C) --> sadd.sat(X, C)
// (X <= Y) ? INT_MIN : (X + C) --> sadd.sat(X, C)
// where C < 0 and Y is INT_MIN - C or INT_MIN - C + 1
if (!IsMax && (Pred == ICmpInst::ICMP_SLT || Pred == ICmpInst::ICMP_SLE) &&
isa<Constant>(Cmp1) &&
match(FVal, m_Add(m_Specific(Cmp0), m_Negative(C)))) {
// Normalize SLE to SLT for threshold comparison.
if (Pred == ICmpInst::ICMP_SLE) {
if (auto Flipped = getFlippedStrictnessPredicateAndConstant(
Pred, cast<Constant>(Cmp1))) {
Pred = Flipped->first;
Cmp1 = Flipped->second;
}
}
// Check: X < INT_MIN - C or X < INT_MIN - C + 1
// INT_MIN - C for negative C is like INT_MIN + |C|
APInt Threshold = *SatC - *C;
if (Pred == ICmpInst::ICMP_SLT &&
(match(Cmp1, m_SpecificIntAllowPoison(Threshold)) ||
match(Cmp1, m_SpecificIntAllowPoison(Threshold + 1))))
return Builder.CreateBinaryIntrinsic(
Intrinsic::sadd_sat, Cmp0, ConstantInt::get(Cmp0->getType(), *C));
}
// Canonicalize predicate to less-than or less-or-equal-than.
if (Pred == ICmpInst::ICMP_SGT || Pred == ICmpInst::ICMP_SGE) {
std::swap(Cmp0, Cmp1);
Pred = CmpInst::getSwappedPredicate(Pred);
}
if (Pred != ICmpInst::ICMP_SLT && Pred != ICmpInst::ICMP_SLE)
return nullptr;
Value *X;
// (INT_MAX - X s< Y) ? INT_MAX : (X + Y) --> sadd.sat(X, Y)
// (INT_MAX - X s< Y) ? INT_MAX : (Y + X) --> sadd.sat(X, Y)
if (IsMax && match(Cmp0, m_NSWSub(m_SpecificInt(*SatC), m_Value(X))) &&
match(FVal, m_c_Add(m_Specific(X), m_Specific(Cmp1)))) {
return Builder.CreateBinaryIntrinsic(Intrinsic::sadd_sat, X, Cmp1);
}
// (INT_MIN - X s> Y) ? INT_MIN : (X + Y) --> sadd.sat(X, Y)
// (INT_MIN - X s> Y) ? INT_MIN : (Y + X) --> sadd.sat(X, Y)
// After swapping operands from the SGT/SGE canonicalization above,
// this becomes (Y s< INT_MIN - X).
if (!IsMax && match(Cmp1, m_NSWSub(m_SpecificInt(*SatC), m_Value(X))) &&
match(FVal, m_c_Add(m_Specific(X), m_Specific(Cmp0)))) {
return Builder.CreateBinaryIntrinsic(Intrinsic::sadd_sat, X, Cmp0);
}
return nullptr;
}
static Value *canonicalizeSaturatedAdd(ICmpInst *Cmp, Value *TVal, Value *FVal,
InstCombiner::BuilderTy &Builder) {
if (!Cmp->hasOneUse())
return nullptr;
if (Value *V = canonicalizeSaturatedAddUnsigned(Cmp, TVal, FVal, Builder))
return V;
if (Value *V = canonicalizeSaturatedAddSigned(Cmp, TVal, FVal, Builder))
return V;
return nullptr;
}
/// Try to match patterns with select and subtract as absolute difference.
static Value *foldAbsDiff(ICmpInst *Cmp, Value *TVal, Value *FVal,
InstCombiner::BuilderTy &Builder) {
auto *TI = dyn_cast<Instruction>(TVal);
auto *FI = dyn_cast<Instruction>(FVal);
if (!TI || !FI)
return nullptr;
// Normalize predicate to gt/lt rather than ge/le.
ICmpInst::Predicate Pred = Cmp->getStrictPredicate();
Value *A = Cmp->getOperand(0);
Value *B = Cmp->getOperand(1);
// Normalize "A - B" as the true value of the select.
if (match(FI, m_Sub(m_Specific(A), m_Specific(B)))) {
std::swap(FI, TI);
Pred = ICmpInst::getSwappedPredicate(Pred);
}
// With any pair of no-wrap subtracts:
// (A > B) ? (A - B) : (B - A) --> abs(A - B)
if (Pred == CmpInst::ICMP_SGT &&
match(TI, m_Sub(m_Specific(A), m_Specific(B))) &&
match(FI, m_Sub(m_Specific(B), m_Specific(A))) &&
(TI->hasNoSignedWrap() || TI->hasNoUnsignedWrap()) &&
(FI->hasNoSignedWrap() || FI->hasNoUnsignedWrap())) {
// The remaining subtract is not "nuw" any more.
// If there's one use of the subtract (no other use than the use we are
// about to replace), then we know that the sub is "nsw" in this context
// even if it was only "nuw" before. If there's another use, then we can't
// add "nsw" to the existing instruction because it may not be safe in the
// other user's context.
TI->setHasNoUnsignedWrap(false);
if (!TI->hasNoSignedWrap())
TI->setHasNoSignedWrap(TI->hasOneUse());
return Builder.CreateBinaryIntrinsic(Intrinsic::abs, TI, Builder.getTrue());
}
// Match: (A > B) ? (A - B) : (0 - (A - B)) --> abs(A - B)
if (Pred == CmpInst::ICMP_SGT &&
match(TI, m_NSWSub(m_Specific(A), m_Specific(B))) &&
match(FI, m_Neg(m_Specific(TI)))) {
return Builder.CreateBinaryIntrinsic(Intrinsic::abs, TI,
Builder.getFalse());
}
// Match: (A < B) ? (0 - (A - B)) : (A - B) --> abs(A - B)
if (Pred == CmpInst::ICMP_SLT &&
match(FI, m_NSWSub(m_Specific(A), m_Specific(B))) &&
match(TI, m_Neg(m_Specific(FI)))) {
return Builder.CreateBinaryIntrinsic(Intrinsic::abs, FI,
Builder.getFalse());
}
// Match: (A > B) ? (0 - (B - A)) : (B - A) --> abs(B - A)
if (Pred == CmpInst::ICMP_SGT &&
match(FI, m_NSWSub(m_Specific(B), m_Specific(A))) &&
match(TI, m_Neg(m_Specific(FI)))) {
return Builder.CreateBinaryIntrinsic(Intrinsic::abs, FI,
Builder.getFalse());
}
// Match: (A < B) ? (B - A) : (0 - (B - A)) --> abs(B - A)
if (Pred == CmpInst::ICMP_SLT &&
match(TI, m_NSWSub(m_Specific(B), m_Specific(A))) &&
match(FI, m_Neg(m_Specific(TI)))) {
return Builder.CreateBinaryIntrinsic(Intrinsic::abs, TI,
Builder.getFalse());
}
return nullptr;
}
/// Fold the following code sequence:
/// \code
/// int a = ctlz(x & -x);
// x ? 31 - a : 32;
/// \code
///
/// into:
/// cttz(x)
static Instruction *foldSelectCtlzToCttz(ICmpInst *ICI, Value *TrueVal,
Value *FalseVal,
InstCombiner::BuilderTy &Builder) {
unsigned BitWidth = TrueVal->getType()->getScalarSizeInBits();
if (!ICI->isEquality() || !match(ICI->getOperand(1), m_Zero()))
return nullptr;
if (ICI->getPredicate() == ICmpInst::ICMP_NE)
std::swap(TrueVal, FalseVal);
Value *Ctlz;
if (match(FalseVal,
m_Xor(m_Value(Ctlz), m_SpecificIntAllowPoison(BitWidth - 1)))) {
if (!isPowerOf2_32(BitWidth))
return nullptr;
} else if (!match(FalseVal, m_Sub(m_SpecificIntAllowPoison(BitWidth - 1),
m_Value(Ctlz)))) {
return nullptr;
}
if (!match(Ctlz, m_Ctlz(m_Value(), m_Value())))
return nullptr;
if (!match(TrueVal, m_SpecificInt(BitWidth)))
return nullptr;
Value *X = ICI->getOperand(0);
auto *II = cast<IntrinsicInst>(Ctlz);
if (!match(II->getOperand(0), m_c_And(m_Specific(X), m_Neg(m_Specific(X)))))
return nullptr;
// The original select returns the constant bitwidth when x == 0, so the
// result is defined there; the cttz must use is_zero_poison = false.
Function *F = Intrinsic::getOrInsertDeclaration(
II->getModule(), Intrinsic::cttz, II->getType());
return CallInst::Create(F, {X, Builder.getFalse()});
}
/// Attempt to fold a cttz/ctlz followed by a icmp plus select into a single
/// call to cttz/ctlz with flag 'is_zero_poison' cleared.
///
/// For example, we can fold the following code sequence:
/// \code
/// %0 = tail call i32 @llvm.cttz.i32(i32 %x, i1 true)
/// %1 = icmp ne i32 %x, 0
/// %2 = select i1 %1, i32 %0, i32 32
/// \code
///
/// into:
/// %0 = tail call i32 @llvm.cttz.i32(i32 %x, i1 false)
static Value *foldSelectCttzCtlz(ICmpInst *ICI, Value *TrueVal, Value *FalseVal,
InstCombinerImpl &IC) {
ICmpInst::Predicate Pred = ICI->getPredicate();
Value *CmpLHS = ICI->getOperand(0);
Value *CmpRHS = ICI->getOperand(1);
// Check if the select condition compares a value for equality.
if (!ICI->isEquality())
return nullptr;
Value *SelectArg = FalseVal;
Value *ValueOnZero = TrueVal;
if (Pred == ICmpInst::ICMP_NE)
std::swap(SelectArg, ValueOnZero);
// Skip zero extend/truncate.
Value *Count = nullptr;
if (!match(SelectArg, m_ZExt(m_Value(Count))) &&
!match(SelectArg, m_Trunc(m_Value(Count))))
Count = SelectArg;
// Check that 'Count' is a call to intrinsic cttz/ctlz. Also check that the
// input to the cttz/ctlz is used as LHS for the compare instruction.
Value *X;
if (!match(Count, m_Cttz(m_Value(X), m_Value())) &&
!match(Count, m_Ctlz(m_Value(X), m_Value())))
return nullptr;
// (X == 0) ? BitWidth : ctz(X)
// (X == -1) ? BitWidth : ctz(~X)
// (X == Y) ? BitWidth : ctz(X ^ Y)
if ((X != CmpLHS || !match(CmpRHS, m_Zero())) &&
(!match(X, m_Not(m_Specific(CmpLHS))) || !match(CmpRHS, m_AllOnes())) &&
!match(X, m_c_Xor(m_Specific(CmpLHS), m_Specific(CmpRHS))))
return nullptr;
IntrinsicInst *II = cast<IntrinsicInst>(Count);
// Check if the value propagated on zero is a constant number equal to the
// sizeof in bits of 'Count'.
unsigned SizeOfInBits = Count->getType()->getScalarSizeInBits();
if (match(ValueOnZero, m_SpecificInt(SizeOfInBits))) {
// A range annotation on the intrinsic may no longer be valid.
II->dropPoisonGeneratingAnnotations();
IC.addToWorklist(II);
return SelectArg;
}
// The ValueOnZero is not the bitwidth. But if the cttz/ctlz (and optional
// zext/trunc) have one use (ending at the select), the cttz/ctlz result will
// not be used if the input is zero. Relax to 'zero is poison' for that case.
if (II->hasOneUse() && SelectArg->hasOneUse() &&
!match(II->getArgOperand(1), m_One())) {
II->setArgOperand(1, ConstantInt::getTrue(II->getContext()));
// noundef attribute on the intrinsic may no longer be valid.
II->dropUBImplyingAttrsAndMetadata();
IC.addToWorklist(II);
}
return nullptr;
}
static Value *canonicalizeSPF(ICmpInst &Cmp, Value *TrueVal, Value *FalseVal,
InstCombinerImpl &IC) {
Value *LHS, *RHS;
// TODO: What to do with pointer min/max patterns?
if (!TrueVal->getType()->isIntOrIntVectorTy())
return nullptr;
SelectPatternFlavor SPF =
matchDecomposedSelectPattern(&Cmp, TrueVal, FalseVal, LHS, RHS).Flavor;
if (SPF == SelectPatternFlavor::SPF_ABS ||
SPF == SelectPatternFlavor::SPF_NABS) {
if (!Cmp.hasOneUse() && !RHS->hasOneUse())
return nullptr; // TODO: Relax this restriction.
// Note that NSW flag can only be propagated for normal, non-negated abs!
bool IntMinIsPoison = SPF == SelectPatternFlavor::SPF_ABS &&
match(RHS, m_NSWNeg(m_Specific(LHS)));
Constant *IntMinIsPoisonC =
ConstantInt::get(Type::getInt1Ty(Cmp.getContext()), IntMinIsPoison);
Value *Abs =
IC.Builder.CreateBinaryIntrinsic(Intrinsic::abs, LHS, IntMinIsPoisonC);
if (SPF == SelectPatternFlavor::SPF_NABS)
return IC.Builder.CreateNeg(Abs); // Always without NSW flag!
return Abs;
}
if (SelectPatternResult::isMinOrMax(SPF)) {
Intrinsic::ID IntrinsicID = getMinMaxIntrinsic(SPF);
return IC.Builder.CreateBinaryIntrinsic(IntrinsicID, LHS, RHS);
}
return nullptr;
}
bool InstCombinerImpl::replaceInInstruction(Value *V, Value *Old, Value *New,
unsigned Depth) {
// Conservatively limit replacement to two instructions upwards.
if (Depth == 2)
return false;
assert(!isa<Constant>(Old) && "Only replace non-constant values");
auto *I = dyn_cast<Instruction>(V);
if (!I || !I->hasOneUse() ||
!isSafeToSpeculativelyExecuteWithVariableReplaced(I))
return false;
// Forbid potentially lane-crossing instructions.
if (Old->getType()->isVectorTy() && !isNotCrossLaneOperation(I))
return false;
bool Changed = false;
for (Use &U : I->operands()) {
if (U == Old) {
replaceUse(U, New);
Worklist.add(I);
Changed = true;
} else {
Changed |= replaceInInstruction(U, Old, New, Depth + 1);
}
}
return Changed;
}
/// If we have a select with an equality comparison, then we know the value in
/// one of the arms of the select. See if substituting this value into an arm
/// and simplifying the result yields the same value as the other arm.
///
/// To make this transform safe, we must drop poison-generating flags
/// (nsw, etc) if we simplified to a binop because the select may be guarding
/// that poison from propagating. If the existing binop already had no
/// poison-generating flags, then this transform can be done by instsimplify.
///
/// Consider:
/// %cmp = icmp eq i32 %x, 2147483647
/// %add = add nsw i32 %x, 1
/// %sel = select i1 %cmp, i32 -2147483648, i32 %add
///
/// We can't replace %sel with %add unless we strip away the flags.
/// TODO: Wrapping flags could be preserved in some cases with better analysis.
Instruction *InstCombinerImpl::foldSelectValueEquivalence(SelectInst &Sel,
CmpInst &Cmp) {
// Canonicalize the pattern to an equivalence on the predicate by swapping the
// select operands.
Value *TrueVal = Sel.getTrueValue(), *FalseVal = Sel.getFalseValue();
bool Swapped = false;
if (Cmp.isEquivalence(/*Invert=*/true)) {
std::swap(TrueVal, FalseVal);
Swapped = true;
} else if (!Cmp.isEquivalence()) {
return nullptr;
}
Value *CmpLHS = Cmp.getOperand(0), *CmpRHS = Cmp.getOperand(1);
auto ReplaceOldOpWithNewOp = [&](Value *OldOp,
Value *NewOp) -> Instruction * {
// In X == Y ? f(X) : Z, try to evaluate f(Y) and replace the operand.
// Take care to avoid replacing X == Y ? X : Z with X == Y ? Y : Z, as that
// would lead to an infinite replacement cycle.
// If we will be able to evaluate f(Y) to a constant, we can allow undef,
// otherwise Y cannot be undef as we might pick different values for undef
// in the cmp and in f(Y).
if (TrueVal == OldOp && (isa<Constant>(OldOp) || !isa<Constant>(NewOp)))
return nullptr;
if (Value *V = simplifyWithOpReplaced(TrueVal, OldOp, NewOp, SQ,
/* AllowRefinement=*/true)) {
// Need some guarantees about the new simplified op to ensure we don't inf
// loop.
// If we simplify to a constant, replace if we aren't creating new undef.
if (match(V, m_ImmConstant()) &&
isGuaranteedNotToBeUndef(V, SQ.AC, &Sel, &DT))
return replaceOperand(Sel, Swapped ? 2 : 1, V);
// If NewOp is a constant and OldOp is not replace iff NewOp doesn't
// contain and undef elements.
// Make sure that V is always simpler than TrueVal, otherwise we might
// end up in an infinite loop.
if (match(NewOp, m_ImmConstant()) ||
(isa<Instruction>(TrueVal) &&
is_contained(cast<Instruction>(TrueVal)->operands(), V))) {
if (isGuaranteedNotToBeUndef(NewOp, SQ.AC, &Sel, &DT))
return replaceOperand(Sel, Swapped ? 2 : 1, V);
return nullptr;
}
}
// Even if TrueVal does not simplify, we can directly replace a use of
// CmpLHS with CmpRHS, as long as the instruction is not used anywhere
// else and is safe to speculatively execute (we may end up executing it
// with different operands, which should not cause side-effects or trigger
// undefined behavior). Only do this if CmpRHS is a constant, as
// profitability is not clear for other cases.
if (OldOp == CmpLHS && match(NewOp, m_ImmConstant()) &&
!match(OldOp, m_Constant()) &&
isGuaranteedNotToBeUndef(NewOp, SQ.AC, &Sel, &DT))
if (replaceInInstruction(TrueVal, OldOp, NewOp))
return &Sel;
return nullptr;
};
bool CanReplaceCmpLHSWithRHS = canReplacePointersIfEqual(CmpLHS, CmpRHS, DL);
if (CanReplaceCmpLHSWithRHS) {
if (Instruction *R = ReplaceOldOpWithNewOp(CmpLHS, CmpRHS))
return R;
}
bool CanReplaceCmpRHSWithLHS = canReplacePointersIfEqual(CmpRHS, CmpLHS, DL);
if (CanReplaceCmpRHSWithLHS) {
if (Instruction *R = ReplaceOldOpWithNewOp(CmpRHS, CmpLHS))
return R;
}
auto *FalseInst = dyn_cast<Instruction>(FalseVal);
if (!FalseInst)
return nullptr;
// InstSimplify already performed this fold if it was possible subject to
// current poison-generating flags. Check whether dropping poison-generating
// flags enables the transform.
// Try each equivalence substitution possibility.
// We have an 'EQ' comparison, so the select's false value will propagate.
// Example:
// (X == 42) ? 43 : (X + 1) --> (X == 42) ? (X + 1) : (X + 1) --> X + 1
SmallVector<Instruction *> DropFlags;
if ((CanReplaceCmpLHSWithRHS &&
simplifyWithOpReplaced(FalseVal, CmpLHS, CmpRHS, SQ,
/* AllowRefinement */ false,
&DropFlags) == TrueVal) ||
(CanReplaceCmpRHSWithLHS &&
simplifyWithOpReplaced(FalseVal, CmpRHS, CmpLHS, SQ,
/* AllowRefinement */ false,
&DropFlags) == TrueVal)) {
for (Instruction *I : DropFlags) {
I->dropPoisonGeneratingAnnotations();
Worklist.add(I);
}
return replaceInstUsesWith(Sel, FalseVal);
}
Constant *CmpC;
if (FalseVal->getType()->isIntOrIntVectorTy(1) &&
match(FalseVal, m_NUWTrunc(m_Specific(CmpLHS))) &&
match(CmpRHS, m_ImmConstant(CmpC)) &&
ConstantFoldCompareInstOperands(
ICmpInst::Predicate::ICMP_NE, CmpC,
ConstantInt::getNullValue(CmpLHS->getType()), DL) == TrueVal) {
return new ICmpInst(CmpInst::Predicate::ICMP_NE, CmpLHS,
ConstantInt::getNullValue(CmpLHS->getType()));
}
return nullptr;
}
/// Fold the following code sequence:
/// \code
/// %XeqZ = icmp eq i64 %X, %Z
/// %YeqZ = icmp eq i64 %Y, %Z
/// %XeqY = icmp eq i64 %X, %Y
/// %not.YeqZ = xor i1 %YeqZ, true
/// %and = select i1 %not.YeqZ, i1 %XeqY, i1 false
/// %equal = select i1 %XeqZ, i1 %YeqZ, i1 %and
/// \code
///
/// into:
/// %equal = icmp eq i64 %X, %Y
Instruction *InstCombinerImpl::foldSelectEqualityTest(SelectInst &Sel) {
Value *X, *Y, *Z;
Value *XeqY, *XeqZ = Sel.getCondition(), *YeqZ = Sel.getTrueValue();
if (!match(XeqZ, m_SpecificICmp(ICmpInst::ICMP_EQ, m_Value(X), m_Value(Z))))
return nullptr;
if (!match(YeqZ,
m_c_SpecificICmp(ICmpInst::ICMP_EQ, m_Value(Y), m_Specific(Z))))
std::swap(X, Z);
if (!match(YeqZ,
m_c_SpecificICmp(ICmpInst::ICMP_EQ, m_Value(Y), m_Specific(Z))))
return nullptr;
if (!match(Sel.getFalseValue(),
m_c_LogicalAnd(m_Not(m_Specific(YeqZ)), m_Value(XeqY))))
return nullptr;
if (!match(XeqY,
m_c_SpecificICmp(ICmpInst::ICMP_EQ, m_Specific(X), m_Specific(Y))))
return nullptr;
cast<ICmpInst>(XeqY)->setSameSign(false);
return replaceInstUsesWith(Sel, XeqY);
}
// See if this is a pattern like:
// %old_cmp1 = icmp slt i32 %x, C2
// %old_replacement = select i1 %old_cmp1, i32 %target_low, i32 %target_high
// %old_x_offseted = add i32 %x, C1
// %old_cmp0 = icmp ult i32 %old_x_offseted, C0
// %r = select i1 %old_cmp0, i32 %x, i32 %old_replacement
// This can be rewritten as more canonical pattern:
// %new_cmp1 = icmp slt i32 %x, -C1
// %new_cmp2 = icmp sge i32 %x, C0-C1
// %new_clamped_low = select i1 %new_cmp1, i32 %target_low, i32 %x
// %r = select i1 %new_cmp2, i32 %target_high, i32 %new_clamped_low
// Iff -C1 s<= C2 s<= C0-C1
// Also ULT predicate can also be UGT iff C0 != -1 (+invert result)
// SLT predicate can also be SGT iff C2 != INT_MAX (+invert res.)
static Value *canonicalizeClampLike(SelectInst &Sel0, ICmpInst &Cmp0,
InstCombiner::BuilderTy &Builder,
InstCombiner &IC) {
Value *X = Sel0.getTrueValue();
Value *Sel1 = Sel0.getFalseValue();
// First match the condition of the outermost select.
// Said condition must be one-use.
if (!Cmp0.hasOneUse())
return nullptr;
ICmpInst::Predicate Pred0 = Cmp0.getPredicate();
Value *Cmp00 = Cmp0.getOperand(0);
Constant *C0;
if (!match(Cmp0.getOperand(1),
m_CombineAnd(m_AnyIntegralConstant(), m_Constant(C0))))
return nullptr;
if (!match(Sel1, m_SelectLike(m_Value(), m_Value(), m_Value()))) {
Pred0 = ICmpInst::getInversePredicate(Pred0);
std::swap(X, Sel1);
}
// Canonicalize Cmp0 into ult or uge.
// FIXME: we shouldn't care about lanes that are 'undef' in the end?
switch (Pred0) {
case ICmpInst::Predicate::ICMP_ULT:
case ICmpInst::Predicate::ICMP_UGE:
// Although icmp ult %x, 0 is an unusual thing to try and should generally
// have been simplified, it does not verify with undef inputs so ensure we
// are not in a strange state.
if (!match(C0, m_SpecificInt_ICMP(
ICmpInst::Predicate::ICMP_NE,
APInt::getZero(C0->getType()->getScalarSizeInBits()))))
return nullptr;
break; // Great!
case ICmpInst::Predicate::ICMP_ULE:
case ICmpInst::Predicate::ICMP_UGT:
// We want to canonicalize it to 'ult' or 'uge', so we'll need to increment
// C0, which again means it must not have any all-ones elements.
if (!match(C0,
m_SpecificInt_ICMP(
ICmpInst::Predicate::ICMP_NE,
APInt::getAllOnes(C0->getType()->getScalarSizeInBits()))))
return nullptr; // Can't do, have all-ones element[s].
Pred0 = ICmpInst::getFlippedStrictnessPredicate(Pred0);
C0 = InstCombiner::AddOne(C0);
break;
default:
return nullptr; // Unknown predicate.
}
// Now that we've canonicalized the ICmp, we know the X we expect;
// the select in other hand should be one-use.
if (!Sel1->hasOneUse())
return nullptr;
// If the types do not match, look through any truncs to the underlying
// instruction.
if (Cmp00->getType() != X->getType() && X->hasOneUse())
match(X, m_TruncOrSelf(m_Value(X)));
// We now can finish matching the condition of the outermost select:
// it should either be the X itself, or an addition of some constant to X.
Constant *C1;
if (Cmp00 == X)
C1 = ConstantInt::getNullValue(X->getType());
else if (!match(Cmp00,
m_Add(m_Specific(X),
m_CombineAnd(m_AnyIntegralConstant(), m_Constant(C1)))))
return nullptr;
Value *Cmp1;
CmpPredicate Pred1;
Constant *C2;
Value *ReplacementLow, *ReplacementHigh;
if (!match(Sel1, m_SelectLike(m_Value(Cmp1), m_Value(ReplacementLow),
m_Value(ReplacementHigh))) ||
!match(Cmp1,
m_ICmp(Pred1, m_Specific(X),
m_CombineAnd(m_AnyIntegralConstant(), m_Constant(C2)))))
return nullptr;
if (!Cmp1->hasOneUse() && (Cmp00 == X || !Cmp00->hasOneUse()))
return nullptr; // Not enough one-use instructions for the fold.
// FIXME: this restriction could be relaxed if Cmp1 can be reused as one of
// two comparisons we'll need to build.
// Canonicalize Cmp1 into the form we expect.
// FIXME: we shouldn't care about lanes that are 'undef' in the end?
switch (Pred1) {
case ICmpInst::Predicate::ICMP_SLT:
break;
case ICmpInst::Predicate::ICMP_SLE:
// We'd have to increment C2 by one, and for that it must not have signed
// max element, but then it would have been canonicalized to 'slt' before
// we get here. So we can't do anything useful with 'sle'.
return nullptr;
case ICmpInst::Predicate::ICMP_SGT:
// We want to canonicalize it to 'slt', so we'll need to increment C2,
// which again means it must not have any signed max elements.
if (!match(C2,
m_SpecificInt_ICMP(ICmpInst::Predicate::ICMP_NE,
APInt::getSignedMaxValue(
C2->getType()->getScalarSizeInBits()))))
return nullptr; // Can't do, have signed max element[s].
C2 = InstCombiner::AddOne(C2);
[[fallthrough]];
case ICmpInst::Predicate::ICMP_SGE:
// Also non-canonical, but here we don't need to change C2,
// so we don't have any restrictions on C2, so we can just handle it.
Pred1 = ICmpInst::Predicate::ICMP_SLT;
std::swap(ReplacementLow, ReplacementHigh);
break;
default:
return nullptr; // Unknown predicate.
}
assert(Pred1 == ICmpInst::Predicate::ICMP_SLT &&
"Unexpected predicate type.");
// The thresholds of this clamp-like pattern.
auto *ThresholdLowIncl = ConstantExpr::getNeg(C1);
auto *ThresholdHighExcl = ConstantExpr::getSub(C0, C1);
assert((Pred0 == ICmpInst::Predicate::ICMP_ULT ||
Pred0 == ICmpInst::Predicate::ICMP_UGE) &&
"Unexpected predicate type.");
if (Pred0 == ICmpInst::Predicate::ICMP_UGE)
std::swap(ThresholdLowIncl, ThresholdHighExcl);
// The fold has a precondition 1: C2 s>= ThresholdLow
auto *Precond1 = ConstantFoldCompareInstOperands(
ICmpInst::Predicate::ICMP_SGE, C2, ThresholdLowIncl, IC.getDataLayout());
if (!Precond1 || !match(Precond1, m_One()))
return nullptr;
// The fold has a precondition 2: C2 s<= ThresholdHigh
auto *Precond2 = ConstantFoldCompareInstOperands(
ICmpInst::Predicate::ICMP_SLE, C2, ThresholdHighExcl, IC.getDataLayout());
if (!Precond2 || !match(Precond2, m_One()))
return nullptr;
// If we are matching from a truncated input, we need to sext the
// ReplacementLow and ReplacementHigh values. Only do the transform if they
// are free to extend due to being constants.
if (X->getType() != Sel0.getType()) {
Constant *LowC, *HighC;
if (!match(ReplacementLow, m_ImmConstant(LowC)) ||
!match(ReplacementHigh, m_ImmConstant(HighC)))
return nullptr;
const DataLayout &DL = Sel0.getDataLayout();
ReplacementLow =
ConstantFoldCastOperand(Instruction::SExt, LowC, X->getType(), DL);
ReplacementHigh =
ConstantFoldCastOperand(Instruction::SExt, HighC, X->getType(), DL);
assert(ReplacementLow && ReplacementHigh &&
"Constant folding of ImmConstant cannot fail");
}
// All good, finally emit the new pattern.
Value *ShouldReplaceLow = Builder.CreateICmpSLT(X, ThresholdLowIncl);
Value *ShouldReplaceHigh = Builder.CreateICmpSGE(X, ThresholdHighExcl);
Value *MaybeReplacedLow =
Builder.CreateSelect(ShouldReplaceLow, ReplacementLow, X);
// Create the final select. If we looked through a truncate above, we will
// need to retruncate the result.
Value *MaybeReplacedHigh = Builder.CreateSelect(
ShouldReplaceHigh, ReplacementHigh, MaybeReplacedLow);
return Builder.CreateTrunc(MaybeReplacedHigh, Sel0.getType());
}
// If we have
// %cmp = icmp [canonical predicate] i32 %x, C0
// %r = select i1 %cmp, i32 %y, i32 C1
// Where C0 != C1 and %x may be different from %y, see if the constant that we
// will have if we flip the strictness of the predicate (i.e. without changing
// the result) is identical to the C1 in select. If it matches we can change
// original comparison to one with swapped predicate, reuse the constant,
// and swap the hands of select.
static Instruction *
tryToReuseConstantFromSelectInComparison(SelectInst &Sel, ICmpInst &Cmp,
InstCombinerImpl &IC) {
CmpPredicate Pred;
Value *X;
Constant *C0;
if (!match(&Cmp, m_OneUse(m_ICmp(
Pred, m_Value(X),
m_CombineAnd(m_AnyIntegralConstant(), m_Constant(C0))))))
return nullptr;
// If comparison predicate is non-relational, we won't be able to do anything.
if (ICmpInst::isEquality(Pred))
return nullptr;
// If comparison predicate is non-canonical, then we certainly won't be able
// to make it canonical; canonicalizeCmpWithConstant() already tried.
if (!InstCombiner::isCanonicalPredicate(Pred))
return nullptr;
// If the [input] type of comparison and select type are different, lets abort
// for now. We could try to compare constants with trunc/[zs]ext though.
if (C0->getType() != Sel.getType())
return nullptr;
// ULT with 'add' of a constant is canonical. See foldICmpAddConstant().
// FIXME: Are there more magic icmp predicate+constant pairs we must avoid?
// Or should we just abandon this transform entirely?
if (Pred == CmpInst::ICMP_ULT && match(X, m_Add(m_Value(), m_Constant())))
return nullptr;
Value *SelVal0, *SelVal1; // We do not care which one is from where.
match(&Sel, m_Select(m_Value(), m_Value(SelVal0), m_Value(SelVal1)));
// At least one of these values we are selecting between must be a constant
// else we'll never succeed.
if (!match(SelVal0, m_AnyIntegralConstant()) &&
!match(SelVal1, m_AnyIntegralConstant()))
return nullptr;
// Does this constant C match any of the `select` values?
auto MatchesSelectValue = [SelVal0, SelVal1](Constant *C) {
return C->isElementWiseEqual(SelVal0) || C->isElementWiseEqual(SelVal1);
};
// If C0 *already* matches true/false value of select, we are done.
if (MatchesSelectValue(C0))
return nullptr;
// Check the constant we'd have with flipped-strictness predicate.
auto FlippedStrictness = getFlippedStrictnessPredicateAndConstant(Pred, C0);
if (!FlippedStrictness)
return nullptr;
// If said constant doesn't match either, then there is no hope,
if (!MatchesSelectValue(FlippedStrictness->second))
return nullptr;
// It matched! Lets insert the new comparison just before select.
InstCombiner::BuilderTy::InsertPointGuard Guard(IC.Builder);
IC.Builder.SetInsertPoint(&Sel);
Pred = ICmpInst::getSwappedPredicate(Pred); // Yes, swapped.
Value *NewCmp = IC.Builder.CreateICmp(Pred, X, FlippedStrictness->second,
Cmp.getName() + ".inv");
IC.replaceOperand(Sel, 0, NewCmp);
Sel.swapValues();
Sel.swapProfMetadata();
return &Sel;
}
static Instruction *foldSelectZeroOrOnes(ICmpInst *Cmp, Value *TVal,
Value *FVal,
InstCombiner::BuilderTy &Builder) {
if (!Cmp->hasOneUse())
return nullptr;
const APInt *CmpC;
if (!match(Cmp->getOperand(1), m_APIntAllowPoison(CmpC)))
return nullptr;
// (X u< 2) ? -X : -1 --> sext (X != 0)
Value *X = Cmp->getOperand(0);
if (Cmp->getPredicate() == ICmpInst::ICMP_ULT && *CmpC == 2 &&
match(TVal, m_Neg(m_Specific(X))) && match(FVal, m_AllOnes()))
return new SExtInst(Builder.CreateIsNotNull(X), TVal->getType());
// (X u> 1) ? -1 : -X --> sext (X != 0)
if (Cmp->getPredicate() == ICmpInst::ICMP_UGT && *CmpC == 1 &&
match(FVal, m_Neg(m_Specific(X))) && match(TVal, m_AllOnes()))
return new SExtInst(Builder.CreateIsNotNull(X), TVal->getType());
return nullptr;
}
static Value *foldSelectInstWithICmpConst(SelectInst &SI, ICmpInst *ICI,
InstCombiner::BuilderTy &Builder) {
const APInt *CmpC;
Value *V;
CmpPredicate Pred;
if (!match(ICI, m_ICmp(Pred, m_Value(V), m_APInt(CmpC))))
return nullptr;
// Match clamp away from min/max value as a max/min operation.
Value *TVal = SI.getTrueValue();
Value *FVal = SI.getFalseValue();
if (Pred == ICmpInst::ICMP_EQ && V == FVal) {
// (V == UMIN) ? UMIN+1 : V --> umax(V, UMIN+1)
if (CmpC->isMinValue() && match(TVal, m_SpecificInt(*CmpC + 1)))
return Builder.CreateBinaryIntrinsic(Intrinsic::umax, V, TVal);
// (V == UMAX) ? UMAX-1 : V --> umin(V, UMAX-1)
if (CmpC->isMaxValue() && match(TVal, m_SpecificInt(*CmpC - 1)))
return Builder.CreateBinaryIntrinsic(Intrinsic::umin, V, TVal);
// (V == SMIN) ? SMIN+1 : V --> smax(V, SMIN+1)
if (CmpC->isMinSignedValue() && match(TVal, m_SpecificInt(*CmpC + 1)))
return Builder.CreateBinaryIntrinsic(Intrinsic::smax, V, TVal);
// (V == SMAX) ? SMAX-1 : V --> smin(V, SMAX-1)
if (CmpC->isMaxSignedValue() && match(TVal, m_SpecificInt(*CmpC - 1)))
return Builder.CreateBinaryIntrinsic(Intrinsic::smin, V, TVal);
}
// Fold icmp(X) ? f(X) : C to f(X) when f(X) is guaranteed to be equal to C
// for all X in the exact range of the inverse predicate.
Instruction *Op;
const APInt *C;
CmpInst::Predicate CPred;
if (match(&SI, m_Select(m_Specific(ICI), m_APInt(C), m_Instruction(Op))))
CPred = ICI->getPredicate();
else if (match(&SI, m_Select(m_Specific(ICI), m_Instruction(Op), m_APInt(C))))
CPred = ICI->getInversePredicate();
else
return nullptr;
ConstantRange InvDomCR = ConstantRange::makeExactICmpRegion(CPred, *CmpC);
const APInt *OpC;
if (match(Op, m_BinOp(m_Specific(V), m_APInt(OpC)))) {
ConstantRange R = InvDomCR.binaryOp(
static_cast<Instruction::BinaryOps>(Op->getOpcode()), *OpC);
if (R == *C) {
Op->dropPoisonGeneratingFlags();
return Op;
}
}
if (auto *MMI = dyn_cast<MinMaxIntrinsic>(Op);
MMI && MMI->getLHS() == V && match(MMI->getRHS(), m_APInt(OpC))) {
ConstantRange R = ConstantRange::intrinsic(MMI->getIntrinsicID(),
{InvDomCR, ConstantRange(*OpC)});
if (R == *C) {
MMI->dropPoisonGeneratingAnnotations();
return MMI;
}
}
return nullptr;
}
/// `A == MIN_INT ? B != MIN_INT : A < B` --> `A < B`
/// `A == MAX_INT ? B != MAX_INT : A > B` --> `A > B`
static Instruction *foldSelectWithExtremeEqCond(Value *CmpLHS, Value *CmpRHS,
Value *TrueVal,
Value *FalseVal) {
Type *Ty = CmpLHS->getType();
if (Ty->isPtrOrPtrVectorTy())
return nullptr;
CmpPredicate Pred;
Value *B;
if (!match(FalseVal, m_c_ICmp(Pred, m_Specific(CmpLHS), m_Value(B))))
return nullptr;
Value *TValRHS;
if (!match(TrueVal, m_SpecificICmp(ICmpInst::ICMP_NE, m_Specific(B),
m_Value(TValRHS))))
return nullptr;
APInt C;
unsigned BitWidth = Ty->getScalarSizeInBits();
if (ICmpInst::isLT(Pred)) {
C = CmpInst::isSigned(Pred) ? APInt::getSignedMinValue(BitWidth)
: APInt::getMinValue(BitWidth);
} else if (ICmpInst::isGT(Pred)) {
C = CmpInst::isSigned(Pred) ? APInt::getSignedMaxValue(BitWidth)
: APInt::getMaxValue(BitWidth);
} else {
return nullptr;
}
if (!match(CmpRHS, m_SpecificInt(C)) || !match(TValRHS, m_SpecificInt(C)))
return nullptr;
return new ICmpInst(Pred, CmpLHS, B);
}
static Instruction *foldSelectICmpEq(SelectInst &SI, ICmpInst *ICI,
InstCombinerImpl &IC) {
ICmpInst::Predicate Pred = ICI->getPredicate();
if (!ICmpInst::isEquality(Pred))
return nullptr;
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
Value *CmpLHS = ICI->getOperand(0);
Value *CmpRHS = ICI->getOperand(1);
if (Pred == ICmpInst::ICMP_NE)
std::swap(TrueVal, FalseVal);
if (Instruction *Res =
foldSelectWithExtremeEqCond(CmpLHS, CmpRHS, TrueVal, FalseVal))
return Res;
return nullptr;
}
/// Fold `X Pred C1 ? X BOp C2 : C1 BOp C2` to `min/max(X, C1) BOp C2`.
/// This allows for better canonicalization.
Value *InstCombinerImpl::foldSelectWithConstOpToBinOp(ICmpInst *Cmp,
Value *TrueVal,
Value *FalseVal) {
Constant *C1, *C2, *C3;
Value *X;
CmpPredicate Predicate;
if (!match(Cmp, m_ICmp(Predicate, m_Value(X), m_Constant(C1))))
return nullptr;
if (!ICmpInst::isRelational(Predicate))
return nullptr;
if (match(TrueVal, m_Constant())) {
std::swap(FalseVal, TrueVal);
Predicate = ICmpInst::getInversePredicate(Predicate);
}
if (!match(FalseVal, m_Constant(C3)) || !TrueVal->hasOneUse())
return nullptr;
bool IsIntrinsic;
unsigned Opcode;
if (BinaryOperator *BOp = dyn_cast<BinaryOperator>(TrueVal)) {
Opcode = BOp->getOpcode();
IsIntrinsic = false;
// This fold causes some regressions and is primarily intended for
// add and sub. So we early exit for div and rem to minimize the
// regressions.
if (Instruction::isIntDivRem(Opcode))
return nullptr;
if (!match(BOp, m_BinOp(m_Specific(X), m_Constant(C2))))
return nullptr;
} else if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(TrueVal)) {
if (!match(II, m_MaxOrMin(m_Specific(X), m_Constant(C2))))
return nullptr;
Opcode = II->getIntrinsicID();
IsIntrinsic = true;
} else {
return nullptr;
}
Value *RHS;
SelectPatternFlavor SPF;
const DataLayout &DL = Cmp->getDataLayout();
auto Flipped = getFlippedStrictnessPredicateAndConstant(Predicate, C1);
auto FoldBinaryOpOrIntrinsic = [&](Constant *LHS, Constant *RHS) {
return IsIntrinsic
? ConstantFoldIntrinsic(Opcode, {LHS, RHS}, LHS->getType(), DL)
: ConstantFoldBinaryOpOperands(Opcode, LHS, RHS, DL);
};
if (C3 == FoldBinaryOpOrIntrinsic(C1, C2)) {
SPF = getSelectPattern(Predicate).Flavor;
RHS = C1;
} else if (Flipped && C3 == FoldBinaryOpOrIntrinsic(Flipped->second, C2)) {
SPF = getSelectPattern(Flipped->first).Flavor;
RHS = Flipped->second;
} else {
return nullptr;
}
Intrinsic::ID MinMaxID = getMinMaxIntrinsic(SPF);
Value *MinMax = Builder.CreateBinaryIntrinsic(MinMaxID, X, RHS);
if (IsIntrinsic)
return Builder.CreateBinaryIntrinsic(Opcode, MinMax, C2);
const auto BinOpc = Instruction::BinaryOps(Opcode);
Value *BinOp = Builder.CreateBinOp(BinOpc, MinMax, C2);
// If we can attach no-wrap flags to the new instruction, do so if the
// old instruction had them and C1 BinOp C2 does not overflow.
if (Instruction *BinOpInst = dyn_cast<Instruction>(BinOp)) {
if (BinOpc == Instruction::Add || BinOpc == Instruction::Sub ||
BinOpc == Instruction::Mul) {
Instruction *OldBinOp = cast<BinaryOperator>(TrueVal);
if (OldBinOp->hasNoSignedWrap() &&
willNotOverflow(BinOpc, RHS, C2, *BinOpInst, /*IsSigned=*/true))
BinOpInst->setHasNoSignedWrap();
if (OldBinOp->hasNoUnsignedWrap() &&
willNotOverflow(BinOpc, RHS, C2, *BinOpInst, /*IsSigned=*/false))
BinOpInst->setHasNoUnsignedWrap();
}
}
return BinOp;
}
/// Folds:
/// %a_sub = call @llvm.usub.sat(x, IntConst1)
/// %b_sub = call @llvm.usub.sat(y, IntConst2)
/// %or = or %a_sub, %b_sub
/// %cmp = icmp eq %or, 0
/// %sel = select %cmp, 0, MostSignificantBit
/// into:
/// %a_sub' = usub.sat(x, IntConst1 - MostSignificantBit)
/// %b_sub' = usub.sat(y, IntConst2 - MostSignificantBit)
/// %or = or %a_sub', %b_sub'
/// %and = and %or, MostSignificantBit
/// Likewise, for vector arguments as well.
static Instruction *foldICmpUSubSatWithAndForMostSignificantBitCmp(
SelectInst &SI, ICmpInst *ICI, InstCombiner::BuilderTy &Builder) {
if (!SI.hasOneUse() || !ICI->hasOneUse())
return nullptr;
CmpPredicate Pred;
Value *A, *B;
const APInt *Constant1, *Constant2;
if (!match(SI.getCondition(),
m_ICmp(Pred,
m_OneUse(m_Or(m_OneUse(m_Intrinsic<Intrinsic::usub_sat>(
m_Value(A), m_APInt(Constant1))),
m_OneUse(m_Intrinsic<Intrinsic::usub_sat>(
m_Value(B), m_APInt(Constant2))))),
m_Zero())))
return nullptr;
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
if (!((Pred == ICmpInst::ICMP_EQ && match(TrueVal, m_Zero()) &&
match(FalseVal, m_SignMask())) ||
(Pred == ICmpInst::ICMP_NE && match(TrueVal, m_SignMask()) &&
match(FalseVal, m_Zero()))))
return nullptr;
auto *Ty = A->getType();
unsigned BW = Constant1->getBitWidth();
APInt MostSignificantBit = APInt::getSignMask(BW);
// Anything over MSB is negative
if (Constant1->isNonNegative() || Constant2->isNonNegative())
return nullptr;
APInt AdjAP1 = *Constant1 - MostSignificantBit + 1;
APInt AdjAP2 = *Constant2 - MostSignificantBit + 1;
auto *Adj1 = ConstantInt::get(Ty, AdjAP1);
auto *Adj2 = ConstantInt::get(Ty, AdjAP2);
Value *NewA = Builder.CreateBinaryIntrinsic(Intrinsic::usub_sat, A, Adj1);
Value *NewB = Builder.CreateBinaryIntrinsic(Intrinsic::usub_sat, B, Adj2);
Value *Or = Builder.CreateOr(NewA, NewB);
Constant *MSBConst = ConstantInt::get(Ty, MostSignificantBit);
return BinaryOperator::CreateAnd(Or, MSBConst);
}
/// Visit a SelectInst that has an ICmpInst as its first operand.
Instruction *InstCombinerImpl::foldSelectInstWithICmp(SelectInst &SI,
ICmpInst *ICI) {
if (Value *V =
canonicalizeSPF(*ICI, SI.getTrueValue(), SI.getFalseValue(), *this))
return replaceInstUsesWith(SI, V);
if (Value *V = foldSelectInstWithICmpConst(SI, ICI, Builder))
return replaceInstUsesWith(SI, V);
if (Value *V = canonicalizeClampLike(SI, *ICI, Builder, *this))
return replaceInstUsesWith(SI, V);
if (Instruction *NewSel =
tryToReuseConstantFromSelectInComparison(SI, *ICI, *this))
return NewSel;
if (Instruction *Folded =
foldICmpUSubSatWithAndForMostSignificantBitCmp(SI, ICI, Builder))
return Folded;
// NOTE: if we wanted to, this is where to detect integer MIN/MAX
bool Changed = false;
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
ICmpInst::Predicate Pred = ICI->getPredicate();
Value *CmpLHS = ICI->getOperand(0);
Value *CmpRHS = ICI->getOperand(1);
if (Instruction *NewSel = foldSelectICmpEq(SI, ICI, *this))
return NewSel;
// Canonicalize a signbit condition to use zero constant by swapping:
// (CmpLHS > -1) ? TV : FV --> (CmpLHS < 0) ? FV : TV
// To avoid conflicts (infinite loops) with other canonicalizations, this is
// not applied with any constant select arm.
if (Pred == ICmpInst::ICMP_SGT && match(CmpRHS, m_AllOnes()) &&
!match(TrueVal, m_Constant()) && !match(FalseVal, m_Constant()) &&
ICI->hasOneUse()) {
InstCombiner::BuilderTy::InsertPointGuard Guard(Builder);
Builder.SetInsertPoint(&SI);
Value *IsNeg = Builder.CreateIsNeg(CmpLHS, ICI->getName());
replaceOperand(SI, 0, IsNeg);
SI.swapValues();
SI.swapProfMetadata();
return &SI;
}
if (Value *V = foldSelectICmpMinMax(ICI, TrueVal, FalseVal, Builder,
SQ.getWithInstruction(&SI)))
return replaceInstUsesWith(SI, V);
if (Value *V = foldSelectICmpAndZeroShl(ICI, TrueVal, FalseVal, Builder))
return replaceInstUsesWith(SI, V);
if (Instruction *V = foldSelectCtlzToCttz(ICI, TrueVal, FalseVal, Builder))
return V;
if (Instruction *V = foldSelectZeroOrOnes(ICI, TrueVal, FalseVal, Builder))
return V;
if (Value *V = foldSelectICmpLshrAshr(ICI, TrueVal, FalseVal, Builder))
return replaceInstUsesWith(SI, V);
if (Value *V = foldSelectCttzCtlz(ICI, TrueVal, FalseVal, *this))
return replaceInstUsesWith(SI, V);
if (Value *V = canonicalizeSaturatedSubtract(ICI, TrueVal, FalseVal, Builder))
return replaceInstUsesWith(SI, V);
if (Value *V = canonicalizeSaturatedAdd(ICI, TrueVal, FalseVal, Builder))
return replaceInstUsesWith(SI, V);
if (Value *V = foldAbsDiff(ICI, TrueVal, FalseVal, Builder))
return replaceInstUsesWith(SI, V);
if (Value *V = foldSelectWithConstOpToBinOp(ICI, TrueVal, FalseVal))
return replaceInstUsesWith(SI, V);
return Changed ? &SI : nullptr;
}
/// We have an SPF (e.g. a min or max) of an SPF of the form:
/// SPF2(SPF1(A, B), C)
Instruction *InstCombinerImpl::foldSPFofSPF(Instruction *Inner,
SelectPatternFlavor SPF1, Value *A,
Value *B, Instruction &Outer,
SelectPatternFlavor SPF2,
Value *C) {
if (Outer.getType() != Inner->getType())
return nullptr;
if (C == A || C == B) {
// MAX(MAX(A, B), B) -> MAX(A, B)
// MIN(MIN(a, b), a) -> MIN(a, b)
// TODO: This could be done in instsimplify.
if (SPF1 == SPF2 && SelectPatternResult::isMinOrMax(SPF1))
return replaceInstUsesWith(Outer, Inner);
}
return nullptr;
}
/// Turn select C, (X + Y), (X - Y) --> (X + (select C, Y, (-Y))).
/// This is even legal for FP.
static Instruction *foldAddSubSelect(SelectInst &SI,
InstCombiner::BuilderTy &Builder) {
Value *CondVal = SI.getCondition();
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
auto *TI = dyn_cast<Instruction>(TrueVal);
auto *FI = dyn_cast<Instruction>(FalseVal);
if (!TI || !FI || !TI->hasOneUse() || !FI->hasOneUse())
return nullptr;
Instruction *AddOp = nullptr, *SubOp = nullptr;
if ((TI->getOpcode() == Instruction::Sub &&
FI->getOpcode() == Instruction::Add) ||
(TI->getOpcode() == Instruction::FSub &&
FI->getOpcode() == Instruction::FAdd)) {
AddOp = FI;
SubOp = TI;
} else if ((FI->getOpcode() == Instruction::Sub &&
TI->getOpcode() == Instruction::Add) ||
(FI->getOpcode() == Instruction::FSub &&
TI->getOpcode() == Instruction::FAdd)) {
AddOp = TI;
SubOp = FI;
}
if (AddOp) {
Value *OtherAddOp = nullptr;
if (SubOp->getOperand(0) == AddOp->getOperand(0)) {
OtherAddOp = AddOp->getOperand(1);
} else if (SubOp->getOperand(0) == AddOp->getOperand(1)) {
OtherAddOp = AddOp->getOperand(0);
}
if (OtherAddOp) {
// So at this point we know we have (Y -> OtherAddOp):
// select C, (add X, Y), (sub X, Z)
Value *NegVal; // Compute -Z
if (SI.getType()->isFPOrFPVectorTy()) {
NegVal = Builder.CreateFNeg(SubOp->getOperand(1));
if (Instruction *NegInst = dyn_cast<Instruction>(NegVal)) {
FastMathFlags Flags = AddOp->getFastMathFlags();
Flags &= SubOp->getFastMathFlags();
NegInst->setFastMathFlags(Flags);
}
} else {
NegVal = Builder.CreateNeg(SubOp->getOperand(1));
}
Value *NewTrueOp = OtherAddOp;
Value *NewFalseOp = NegVal;
if (AddOp != TI)
std::swap(NewTrueOp, NewFalseOp);
Value *NewSel = Builder.CreateSelect(CondVal, NewTrueOp, NewFalseOp,
SI.getName() + ".p", &SI);
if (SI.getType()->isFPOrFPVectorTy()) {
Instruction *RI =
BinaryOperator::CreateFAdd(SubOp->getOperand(0), NewSel);
FastMathFlags Flags = AddOp->getFastMathFlags();
Flags &= SubOp->getFastMathFlags();
RI->setFastMathFlags(Flags);
return RI;
} else
return BinaryOperator::CreateAdd(SubOp->getOperand(0), NewSel);
}
}
return nullptr;
}
/// Turn X + Y overflows ? -1 : X + Y -> uadd_sat X, Y
/// And X - Y overflows ? 0 : X - Y -> usub_sat X, Y
/// Along with a number of patterns similar to:
/// X + Y overflows ? (X < 0 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
/// X - Y overflows ? (X > 0 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
static Instruction *
foldOverflowingAddSubSelect(SelectInst &SI, InstCombiner::BuilderTy &Builder) {
Value *CondVal = SI.getCondition();
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
WithOverflowInst *II;
if (!match(CondVal, m_ExtractValue<1>(m_WithOverflowInst(II))) ||
!match(FalseVal, m_ExtractValue<0>(m_Specific(II))))
return nullptr;
Value *X = II->getLHS();
Value *Y = II->getRHS();
auto IsSignedSaturateLimit = [&](Value *Limit, bool IsAdd) {
Type *Ty = Limit->getType();
CmpPredicate Pred;
Value *TrueVal, *FalseVal, *Op;
const APInt *C;
if (!match(Limit, m_Select(m_ICmp(Pred, m_Value(Op), m_APInt(C)),
m_Value(TrueVal), m_Value(FalseVal))))
return false;
auto IsZeroOrOne = [](const APInt &C) { return C.isZero() || C.isOne(); };
auto IsMinMax = [&](Value *Min, Value *Max) {
APInt MinVal = APInt::getSignedMinValue(Ty->getScalarSizeInBits());
APInt MaxVal = APInt::getSignedMaxValue(Ty->getScalarSizeInBits());
return match(Min, m_SpecificInt(MinVal)) &&
match(Max, m_SpecificInt(MaxVal));
};
if (Op != X && Op != Y)
return false;
if (IsAdd) {
// X + Y overflows ? (X <s 0 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (X <s 1 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y <s 0 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y <s 1 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
if (Pred == ICmpInst::ICMP_SLT && IsZeroOrOne(*C) &&
IsMinMax(TrueVal, FalseVal))
return true;
// X + Y overflows ? (X >s 0 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (X >s -1 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y >s 0 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y >s -1 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
if (Pred == ICmpInst::ICMP_SGT && IsZeroOrOne(*C + 1) &&
IsMinMax(FalseVal, TrueVal))
return true;
} else {
// X - Y overflows ? (X <s 0 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (X <s -1 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
if (Op == X && Pred == ICmpInst::ICMP_SLT && IsZeroOrOne(*C + 1) &&
IsMinMax(TrueVal, FalseVal))
return true;
// X - Y overflows ? (X >s -1 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (X >s -2 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
if (Op == X && Pred == ICmpInst::ICMP_SGT && IsZeroOrOne(*C + 2) &&
IsMinMax(FalseVal, TrueVal))
return true;
// X - Y overflows ? (Y <s 0 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (Y <s 1 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
if (Op == Y && Pred == ICmpInst::ICMP_SLT && IsZeroOrOne(*C) &&
IsMinMax(FalseVal, TrueVal))
return true;
// X - Y overflows ? (Y >s 0 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (Y >s -1 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
if (Op == Y && Pred == ICmpInst::ICMP_SGT && IsZeroOrOne(*C + 1) &&
IsMinMax(TrueVal, FalseVal))
return true;
}
return false;
};
Intrinsic::ID NewIntrinsicID;
if (II->getIntrinsicID() == Intrinsic::uadd_with_overflow &&
match(TrueVal, m_AllOnes()))
// X + Y overflows ? -1 : X + Y -> uadd_sat X, Y
NewIntrinsicID = Intrinsic::uadd_sat;
else if (II->getIntrinsicID() == Intrinsic::usub_with_overflow &&
match(TrueVal, m_Zero()))
// X - Y overflows ? 0 : X - Y -> usub_sat X, Y
NewIntrinsicID = Intrinsic::usub_sat;
else if (II->getIntrinsicID() == Intrinsic::sadd_with_overflow &&
IsSignedSaturateLimit(TrueVal, /*IsAdd=*/true))
// X + Y overflows ? (X <s 0 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (X <s 1 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (X >s 0 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (X >s -1 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y <s 0 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y <s 1 ? INTMIN : INTMAX) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y >s 0 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
// X + Y overflows ? (Y >s -1 ? INTMAX : INTMIN) : X + Y --> sadd_sat X, Y
NewIntrinsicID = Intrinsic::sadd_sat;
else if (II->getIntrinsicID() == Intrinsic::ssub_with_overflow &&
IsSignedSaturateLimit(TrueVal, /*IsAdd=*/false))
// X - Y overflows ? (X <s 0 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (X <s -1 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (X >s -1 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (X >s -2 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (Y <s 0 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (Y <s 1 ? INTMAX : INTMIN) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (Y >s 0 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
// X - Y overflows ? (Y >s -1 ? INTMIN : INTMAX) : X - Y --> ssub_sat X, Y
NewIntrinsicID = Intrinsic::ssub_sat;
else
return nullptr;
Function *F = Intrinsic::getOrInsertDeclaration(SI.getModule(),
NewIntrinsicID, SI.getType());
return CallInst::Create(F, {X, Y});
}
Instruction *InstCombinerImpl::foldSelectExtConst(SelectInst &Sel) {
Constant *C;
if (!match(Sel.getTrueValue(), m_Constant(C)) &&
!match(Sel.getFalseValue(), m_Constant(C)))
return nullptr;
Instruction *ExtInst;
if (!match(Sel.getTrueValue(), m_Instruction(ExtInst)) &&
!match(Sel.getFalseValue(), m_Instruction(ExtInst)))
return nullptr;
auto ExtOpcode = ExtInst->getOpcode();
if (ExtOpcode != Instruction::ZExt && ExtOpcode != Instruction::SExt)
return nullptr;
// If we are extending from a boolean type or if we can create a select that
// has the same size operands as its condition, try to narrow the select.
Value *X = ExtInst->getOperand(0);
Type *SmallType = X->getType();
Value *Cond = Sel.getCondition();
if (!SmallType->isIntOrIntVectorTy(1) &&
(!isa<CmpInst, TruncInst>(Cond) ||
cast<Instruction>(Cond)->getOperand(0)->getType() != SmallType))
return nullptr;
// If the constant is the same after truncation to the smaller type and
// extension to the original type, we can narrow the select.
Type *SelType = Sel.getType();
Constant *TruncC = getLosslessInvCast(C, SmallType, ExtOpcode, DL);
if (TruncC && ExtInst->hasOneUse()) {
Value *TruncCVal = cast<Value>(TruncC);
if (ExtInst == Sel.getFalseValue())
std::swap(X, TruncCVal);
// select Cond, (ext X), C --> ext(select Cond, X, C')
// select Cond, C, (ext X) --> ext(select Cond, C', X)
Value *NewSel = Builder.CreateSelect(Cond, X, TruncCVal, "narrow", &Sel);
return CastInst::Create(Instruction::CastOps(ExtOpcode), NewSel, SelType);
}
return nullptr;
}
/// Try to transform a vector select with a constant condition vector into a
/// shuffle for easier combining with other shuffles and insert/extract.
static Instruction *canonicalizeSelectToShuffle(SelectInst &SI) {
Value *CondVal = SI.getCondition();
Constant *CondC;
auto *CondValTy = dyn_cast<FixedVectorType>(CondVal->getType());
if (!CondValTy || !match(CondVal, m_Constant(CondC)))
return nullptr;
unsigned NumElts = CondValTy->getNumElements();
SmallVector<int, 16> Mask;
Mask.reserve(NumElts);
for (unsigned i = 0; i != NumElts; ++i) {
Constant *Elt = CondC->getAggregateElement(i);
if (!Elt)
return nullptr;
if (Elt->isOneValue()) {
// If the select condition element is true, choose from the 1st vector.
Mask.push_back(i);
} else if (Elt->isNullValue()) {
// If the select condition element is false, choose from the 2nd vector.
Mask.push_back(i + NumElts);
} else if (isa<UndefValue>(Elt)) {
// Undef in a select condition (choose one of the operands) does not mean
// the same thing as undef in a shuffle mask (any value is acceptable), so
// give up.
return nullptr;
} else {
// Bail out on a constant expression.
return nullptr;
}
}
return new ShuffleVectorInst(SI.getTrueValue(), SI.getFalseValue(), Mask);
}
/// If we have a select of vectors with a scalar condition, try to convert that
/// to a vector select by splatting the condition. A splat may get folded with
/// other operations in IR and having all operands of a select be vector types
/// is likely better for vector codegen.
static Instruction *canonicalizeScalarSelectOfVecs(SelectInst &Sel,
InstCombinerImpl &IC) {
auto *Ty = dyn_cast<VectorType>(Sel.getType());
if (!Ty)
return nullptr;
// We can replace a single-use extract with constant index.
Value *Cond = Sel.getCondition();
if (!match(Cond, m_OneUse(m_ExtractElt(m_Value(), m_ConstantInt()))))
return nullptr;
// select (extelt V, Index), T, F --> select (splat V, Index), T, F
// Splatting the extracted condition reduces code (we could directly create a
// splat shuffle of the source vector to eliminate the intermediate step).
return IC.replaceOperand(
Sel, 0, IC.Builder.CreateVectorSplat(Ty->getElementCount(), Cond));
}
/// Reuse bitcasted operands between a compare and select:
/// select (cmp (bitcast C), (bitcast D)), (bitcast' C), (bitcast' D) -->
/// bitcast (select (cmp (bitcast C), (bitcast D)), (bitcast C), (bitcast D))
static Instruction *foldSelectCmpBitcasts(SelectInst &Sel,
InstCombiner::BuilderTy &Builder) {
Value *Cond = Sel.getCondition();
Value *TVal = Sel.getTrueValue();
Value *FVal = Sel.getFalseValue();
CmpPredicate Pred;
Value *A, *B;
if (!match(Cond, m_Cmp(Pred, m_Value(A), m_Value(B))))
return nullptr;
// The select condition is a compare instruction. If the select's true/false
// values are already the same as the compare operands, there's nothing to do.
if (TVal == A || TVal == B || FVal == A || FVal == B)
return nullptr;
Value *C, *D;
if (!match(A, m_BitCast(m_Value(C))) || !match(B, m_BitCast(m_Value(D))))
return nullptr;
// select (cmp (bitcast C), (bitcast D)), (bitcast TSrc), (bitcast FSrc)
Value *TSrc, *FSrc;
if (!match(TVal, m_BitCast(m_Value(TSrc))) ||
!match(FVal, m_BitCast(m_Value(FSrc))))
return nullptr;
// If the select true/false values are *different bitcasts* of the same source
// operands, make the select operands the same as the compare operands and
// cast the result. This is the canonical select form for min/max.
Value *NewSel;
if (TSrc == C && FSrc == D) {
// select (cmp (bitcast C), (bitcast D)), (bitcast' C), (bitcast' D) -->
// bitcast (select (cmp A, B), A, B)
NewSel = Builder.CreateSelect(Cond, A, B, "", &Sel);
} else if (TSrc == D && FSrc == C) {
// select (cmp (bitcast C), (bitcast D)), (bitcast' D), (bitcast' C) -->
// bitcast (select (cmp A, B), B, A)
NewSel = Builder.CreateSelect(Cond, B, A, "", &Sel);
} else {
return nullptr;
}
return new BitCastInst(NewSel, Sel.getType());
}
/// Try to eliminate select instructions that test the returned flag of cmpxchg
/// instructions.
///
/// If a select instruction tests the returned flag of a cmpxchg instruction and
/// selects between the returned value of the cmpxchg instruction its compare
/// operand, the result of the select will always be equal to its false value.
/// For example:
///
/// %cmpxchg = cmpxchg ptr %ptr, i64 %compare, i64 %new_value seq_cst seq_cst
/// %val = extractvalue { i64, i1 } %cmpxchg, 0
/// %success = extractvalue { i64, i1 } %cmpxchg, 1
/// %sel = select i1 %success, i64 %compare, i64 %val
/// ret i64 %sel
///
/// The returned value of the cmpxchg instruction (%val) is the original value
/// located at %ptr prior to any update. If the cmpxchg operation succeeds, %val
/// must have been equal to %compare. Thus, the result of the select is always
/// equal to %val, and the code can be simplified to:
///
/// %cmpxchg = cmpxchg ptr %ptr, i64 %compare, i64 %new_value seq_cst seq_cst
/// %val = extractvalue { i64, i1 } %cmpxchg, 0
/// ret i64 %val
///
static Value *foldSelectCmpXchg(SelectInst &SI) {
// A helper that determines if V is an extractvalue instruction whose
// aggregate operand is a cmpxchg instruction and whose single index is equal
// to I. If such conditions are true, the helper returns the cmpxchg
// instruction; otherwise, a nullptr is returned.
auto isExtractFromCmpXchg = [](Value *V, unsigned I) -> AtomicCmpXchgInst * {
// When extracting the value loaded by a cmpxchg, allow peeking through a
// bitcast. These are inserted for floating-point cmpxchg, for example:
// %bc = bitcast float %compare to i32
// %cmpxchg = cmpxchg ptr %ptr, i32 %bc, i32 %new_value seq_cst seq_cst
// %val = extractvalue { i32, i1 } %cmpxchg, 0
// %success = extractvalue { i32, i1 } %cmpxchg, 1
// %val.bc = bitcast i32 %val to float
// %sel = select i1 %success, float %compare, float %val.bc
if (auto *BI = dyn_cast<BitCastInst>(V); BI && I == 0)
V = BI->getOperand(0);
auto *Extract = dyn_cast<ExtractValueInst>(V);
if (!Extract)
return nullptr;
if (Extract->getIndices()[0] != I)
return nullptr;
return dyn_cast<AtomicCmpXchgInst>(Extract->getAggregateOperand());
};
// Check if the compare value of a cmpxchg matches another value.
auto isCompareSameAsValue = [](Value *CmpVal, Value *SelVal) {
// The values match if they are the same or %CmpVal = bitcast %SelVal (see
// above).
if (CmpVal == SelVal || match(CmpVal, m_BitCast(m_Specific(SelVal))))
return true;
// For FP constants, the value may have been bitcast to Int directly.
auto *IntC = dyn_cast<ConstantInt>(CmpVal);
auto *FpC = dyn_cast<ConstantFP>(SelVal);
return IntC && FpC && IntC->getValue() == FpC->getValue().bitcastToAPInt();
};
// If the select has a single user, and this user is a select instruction that
// we can simplify, skip the cmpxchg simplification for now.
if (SI.hasOneUse())
if (auto *Select = dyn_cast<SelectInst>(SI.user_back()))
if (Select->getCondition() == SI.getCondition())
if (Select->getFalseValue() == SI.getTrueValue() ||
Select->getTrueValue() == SI.getFalseValue())
return nullptr;
// Ensure the select condition is the returned flag of a cmpxchg instruction.
auto *CmpXchg = isExtractFromCmpXchg(SI.getCondition(), 1);
if (!CmpXchg)
return nullptr;
// Check the true value case: The true value of the select is the returned
// value of the same cmpxchg used by the condition, and the false value is the
// cmpxchg instruction's compare operand.
if (auto *X = isExtractFromCmpXchg(SI.getTrueValue(), 0))
if (X == CmpXchg &&
isCompareSameAsValue(X->getCompareOperand(), SI.getFalseValue()))
return SI.getFalseValue();
// Check the false value case: The false value of the select is the returned
// value of the same cmpxchg used by the condition, and the true value is the
// cmpxchg instruction's compare operand.
if (auto *X = isExtractFromCmpXchg(SI.getFalseValue(), 0))
if (X == CmpXchg &&
isCompareSameAsValue(X->getCompareOperand(), SI.getTrueValue()))
return SI.getFalseValue();
return nullptr;
}
/// Try to reduce a funnel/rotate pattern that includes a compare and select
/// into a funnel shift intrinsic. Example:
/// rotl32(a, b) --> (b == 0 ? a : ((a >> (32 - b)) | (a << b)))
/// --> call llvm.fshl.i32(a, a, b)
/// fshl32(a, b, c) --> (c == 0 ? a : ((b >> (32 - c)) | (a << c)))
/// --> call llvm.fshl.i32(a, b, c)
/// fshr32(a, b, c) --> (c == 0 ? b : ((a >> (32 - c)) | (b << c)))
/// --> call llvm.fshr.i32(a, b, c)
static Instruction *foldSelectFunnelShift(SelectInst &Sel,
InstCombiner::BuilderTy &Builder) {
// This must be a power-of-2 type for a bitmasking transform to be valid.
unsigned Width = Sel.getType()->getScalarSizeInBits();
if (!isPowerOf2_32(Width))
return nullptr;
BinaryOperator *Or0, *Or1;
if (!match(Sel.getFalseValue(), m_OneUse(m_Or(m_BinOp(Or0), m_BinOp(Or1)))))
return nullptr;
Value *SV0, *SV1, *SA0, *SA1;
if (!match(Or0, m_OneUse(m_LogicalShift(m_Value(SV0),
m_ZExtOrSelf(m_Value(SA0))))) ||
!match(Or1, m_OneUse(m_LogicalShift(m_Value(SV1),
m_ZExtOrSelf(m_Value(SA1))))) ||
Or0->getOpcode() == Or1->getOpcode())
return nullptr;
// Canonicalize to or(shl(SV0, SA0), lshr(SV1, SA1)).
if (Or0->getOpcode() == BinaryOperator::LShr) {
std::swap(Or0, Or1);
std::swap(SV0, SV1);
std::swap(SA0, SA1);
}
assert(Or0->getOpcode() == BinaryOperator::Shl &&
Or1->getOpcode() == BinaryOperator::LShr &&
"Illegal or(shift,shift) pair");
// Check the shift amounts to see if they are an opposite pair.
Value *ShAmt;
if (match(SA1, m_OneUse(m_Sub(m_SpecificInt(Width), m_Specific(SA0)))))
ShAmt = SA0;
else if (match(SA0, m_OneUse(m_Sub(m_SpecificInt(Width), m_Specific(SA1)))))
ShAmt = SA1;
else
return nullptr;
// We should now have this pattern:
// select ?, TVal, (or (shl SV0, SA0), (lshr SV1, SA1))
// The false value of the select must be a funnel-shift of the true value:
// IsFShl -> TVal must be SV0 else TVal must be SV1.
bool IsFshl = (ShAmt == SA0);
Value *TVal = Sel.getTrueValue();
if ((IsFshl && TVal != SV0) || (!IsFshl && TVal != SV1))
return nullptr;
// Finally, see if the select is filtering out a shift-by-zero.
Value *Cond = Sel.getCondition();
if (!match(Cond, m_OneUse(m_SpecificICmp(ICmpInst::ICMP_EQ, m_Specific(ShAmt),
m_ZeroInt()))))
return nullptr;
// If this is not a rotate then the select was blocking poison from the
// 'shift-by-zero' non-TVal, but a funnel shift won't - so freeze it.
if (SV0 != SV1) {
if (IsFshl && !llvm::isGuaranteedNotToBePoison(SV1))
SV1 = Builder.CreateFreeze(SV1);
else if (!IsFshl && !llvm::isGuaranteedNotToBePoison(SV0))
SV0 = Builder.CreateFreeze(SV0);
}
// This is a funnel/rotate that avoids shift-by-bitwidth UB in a suboptimal way.
// Convert to funnel shift intrinsic.
Intrinsic::ID IID = IsFshl ? Intrinsic::fshl : Intrinsic::fshr;
Function *F =
Intrinsic::getOrInsertDeclaration(Sel.getModule(), IID, Sel.getType());
ShAmt = Builder.CreateZExt(ShAmt, Sel.getType());
return CallInst::Create(F, { SV0, SV1, ShAmt });
}
static Instruction *foldSelectToCopysign(SelectInst &Sel,
InstCombiner::BuilderTy &Builder) {
Value *Cond = Sel.getCondition();
Value *TVal = Sel.getTrueValue();
Value *FVal = Sel.getFalseValue();
Type *SelType = Sel.getType();
// Match select ?, TC, FC where the constants are equal but negated.
// TODO: Generalize to handle a negated variable operand?
const APFloat *TC, *FC;
if (!match(TVal, m_APFloatAllowPoison(TC)) ||
!match(FVal, m_APFloatAllowPoison(FC)) ||
!abs(*TC).bitwiseIsEqual(abs(*FC)))
return nullptr;
assert(TC != FC && "Expected equal select arms to simplify");
Value *X;
const APInt *C;
bool IsTrueIfSignSet;
CmpPredicate Pred;
if (!match(Cond, m_OneUse(m_ICmp(Pred, m_ElementWiseBitCast(m_Value(X)),
m_APInt(C)))) ||
!isSignBitCheck(Pred, *C, IsTrueIfSignSet) || X->getType() != SelType)
return nullptr;
// If needed, negate the value that will be the sign argument of the copysign:
// (bitcast X) < 0 ? -TC : TC --> copysign(TC, X)
// (bitcast X) < 0 ? TC : -TC --> copysign(TC, -X)
// (bitcast X) >= 0 ? -TC : TC --> copysign(TC, -X)
// (bitcast X) >= 0 ? TC : -TC --> copysign(TC, X)
// Note: FMF from the select can not be propagated to the new instructions.
if (IsTrueIfSignSet ^ TC->isNegative())
X = Builder.CreateFNeg(X);
// Canonicalize the magnitude argument as the positive constant since we do
// not care about its sign.
Value *MagArg = ConstantFP::get(SelType, abs(*TC));
Function *F = Intrinsic::getOrInsertDeclaration(
Sel.getModule(), Intrinsic::copysign, Sel.getType());
return CallInst::Create(F, { MagArg, X });
}
Instruction *InstCombinerImpl::foldVectorSelect(SelectInst &Sel) {
if (!isa<VectorType>(Sel.getType()))
return nullptr;
Value *Cond = Sel.getCondition();
Value *TVal = Sel.getTrueValue();
Value *FVal = Sel.getFalseValue();
Value *C, *X, *Y;
if (match(Cond, m_VecReverse(m_Value(C)))) {
auto createSelReverse = [&](Value *C, Value *X, Value *Y) {
Value *V = Builder.CreateSelect(C, X, Y, Sel.getName(), &Sel);
if (auto *I = dyn_cast<Instruction>(V))
I->copyIRFlags(&Sel);
Module *M = Sel.getModule();
Function *F = Intrinsic::getOrInsertDeclaration(
M, Intrinsic::vector_reverse, V->getType());
return CallInst::Create(F, V);
};
if (match(TVal, m_VecReverse(m_Value(X)))) {
// select rev(C), rev(X), rev(Y) --> rev(select C, X, Y)
if (match(FVal, m_VecReverse(m_Value(Y))) &&
(Cond->hasOneUse() || TVal->hasOneUse() || FVal->hasOneUse()))
return createSelReverse(C, X, Y);
// select rev(C), rev(X), FValSplat --> rev(select C, X, FValSplat)
if ((Cond->hasOneUse() || TVal->hasOneUse()) && isSplatValue(FVal))
return createSelReverse(C, X, FVal);
}
// select rev(C), TValSplat, rev(Y) --> rev(select C, TValSplat, Y)
else if (isSplatValue(TVal) && match(FVal, m_VecReverse(m_Value(Y))) &&
(Cond->hasOneUse() || FVal->hasOneUse()))
return createSelReverse(C, TVal, Y);
}
auto *VecTy = dyn_cast<FixedVectorType>(Sel.getType());
if (!VecTy)
return nullptr;
unsigned NumElts = VecTy->getNumElements();
APInt PoisonElts(NumElts, 0);
APInt AllOnesEltMask(APInt::getAllOnes(NumElts));
if (Value *V = SimplifyDemandedVectorElts(&Sel, AllOnesEltMask, PoisonElts)) {
if (V != &Sel)
return replaceInstUsesWith(Sel, V);
return &Sel;
}
// A select of a "select shuffle" with a common operand can be rearranged
// to select followed by "select shuffle". Because of poison, this only works
// in the case of a shuffle with no undefined mask elements.
ArrayRef<int> Mask;
if (match(TVal, m_OneUse(m_Shuffle(m_Value(X), m_Value(Y), m_Mask(Mask)))) &&
!is_contained(Mask, PoisonMaskElem) &&
cast<ShuffleVectorInst>(TVal)->isSelect()) {
if (X == FVal) {
// select Cond, (shuf_sel X, Y), X --> shuf_sel X, (select Cond, Y, X)
Value *NewSel = Builder.CreateSelect(Cond, Y, X, "sel", &Sel);
return new ShuffleVectorInst(X, NewSel, Mask);
}
if (Y == FVal) {
// select Cond, (shuf_sel X, Y), Y --> shuf_sel (select Cond, X, Y), Y
Value *NewSel = Builder.CreateSelect(Cond, X, Y, "sel", &Sel);
return new ShuffleVectorInst(NewSel, Y, Mask);
}
}
if (match(FVal, m_OneUse(m_Shuffle(m_Value(X), m_Value(Y), m_Mask(Mask)))) &&
!is_contained(Mask, PoisonMaskElem) &&
cast<ShuffleVectorInst>(FVal)->isSelect()) {
if (X == TVal) {
// select Cond, X, (shuf_sel X, Y) --> shuf_sel X, (select Cond, X, Y)
Value *NewSel = Builder.CreateSelect(Cond, X, Y, "sel", &Sel);
return new ShuffleVectorInst(X, NewSel, Mask);
}
if (Y == TVal) {
// select Cond, Y, (shuf_sel X, Y) --> shuf_sel (select Cond, Y, X), Y
Value *NewSel = Builder.CreateSelect(Cond, Y, X, "sel", &Sel);
return new ShuffleVectorInst(NewSel, Y, Mask);
}
}
return nullptr;
}
static Instruction *foldSelectToPhiImpl(SelectInst &Sel, BasicBlock *BB,
const DominatorTree &DT,
InstCombiner::BuilderTy &Builder) {
// Find the block's immediate dominator that ends with a conditional branch
// that matches select's condition (maybe inverted).
auto *IDomNode = DT[BB]->getIDom();
if (!IDomNode)
return nullptr;
BasicBlock *IDom = IDomNode->getBlock();
Value *Cond = Sel.getCondition();
Value *IfTrue, *IfFalse;
BasicBlock *TrueSucc, *FalseSucc;
if (match(IDom->getTerminator(),
m_Br(m_Specific(Cond), m_BasicBlock(TrueSucc),
m_BasicBlock(FalseSucc)))) {
IfTrue = Sel.getTrueValue();
IfFalse = Sel.getFalseValue();
} else if (match(IDom->getTerminator(),
m_Br(m_Not(m_Specific(Cond)), m_BasicBlock(TrueSucc),
m_BasicBlock(FalseSucc)))) {
IfTrue = Sel.getFalseValue();
IfFalse = Sel.getTrueValue();
} else
return nullptr;
// Make sure the branches are actually different.
if (TrueSucc == FalseSucc)
return nullptr;
// We want to replace select %cond, %a, %b with a phi that takes value %a
// for all incoming edges that are dominated by condition `%cond == true`,
// and value %b for edges dominated by condition `%cond == false`. If %a
// or %b are also phis from the same basic block, we can go further and take
// their incoming values from the corresponding blocks.
BasicBlockEdge TrueEdge(IDom, TrueSucc);
BasicBlockEdge FalseEdge(IDom, FalseSucc);
DenseMap<BasicBlock *, Value *> Inputs;
for (auto *Pred : predecessors(BB)) {
// Check implication.
BasicBlockEdge Incoming(Pred, BB);
if (DT.dominates(TrueEdge, Incoming))
Inputs[Pred] = IfTrue->DoPHITranslation(BB, Pred);
else if (DT.dominates(FalseEdge, Incoming))
Inputs[Pred] = IfFalse->DoPHITranslation(BB, Pred);
else
return nullptr;
// Check availability.
if (auto *Insn = dyn_cast<Instruction>(Inputs[Pred]))
if (!DT.dominates(Insn, Pred->getTerminator()))
return nullptr;
}
Builder.SetInsertPoint(BB, BB->begin());
auto *PN = Builder.CreatePHI(Sel.getType(), Inputs.size());
for (auto *Pred : predecessors(BB))
PN->addIncoming(Inputs[Pred], Pred);
PN->takeName(&Sel);
return PN;
}
static Instruction *foldSelectToPhi(SelectInst &Sel, const DominatorTree &DT,
InstCombiner::BuilderTy &Builder) {
// Try to replace this select with Phi in one of these blocks.
SmallSetVector<BasicBlock *, 4> CandidateBlocks;
CandidateBlocks.insert(Sel.getParent());
for (Value *V : Sel.operands())
if (auto *I = dyn_cast<Instruction>(V))
CandidateBlocks.insert(I->getParent());
for (BasicBlock *BB : CandidateBlocks)
if (auto *PN = foldSelectToPhiImpl(Sel, BB, DT, Builder))
return PN;
return nullptr;
}
/// Tries to reduce a pattern that arises when calculating the remainder of the
/// Euclidean division. When the divisor is a power of two and is guaranteed not
/// to be negative, a signed remainder can be folded with a bitwise and.
///
/// (x % n) < 0 ? (x % n) + n : (x % n)
/// -> x & (n - 1)
static Instruction *foldSelectWithSRem(SelectInst &SI, InstCombinerImpl &IC,
IRBuilderBase &Builder) {
Value *CondVal = SI.getCondition();
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
CmpPredicate Pred;
Value *Op, *RemRes, *Remainder;
const APInt *C;
bool TrueIfSigned = false;
if (!(match(CondVal, m_ICmp(Pred, m_Value(RemRes), m_APInt(C))) &&
isSignBitCheck(Pred, *C, TrueIfSigned)))
return nullptr;
// If the sign bit is not set, we have a SGE/SGT comparison, and the operands
// of the select are inverted.
if (!TrueIfSigned)
std::swap(TrueVal, FalseVal);
auto FoldToBitwiseAnd = [&](Value *Remainder) -> Instruction * {
Value *Add = Builder.CreateAdd(
Remainder, Constant::getAllOnesValue(RemRes->getType()));
return BinaryOperator::CreateAnd(Op, Add);
};
// Match the general case:
// %rem = srem i32 %x, %n
// %cnd = icmp slt i32 %rem, 0
// %add = add i32 %rem, %n
// %sel = select i1 %cnd, i32 %add, i32 %rem
if (match(TrueVal, m_c_Add(m_Specific(RemRes), m_Value(Remainder))) &&
match(RemRes, m_SRem(m_Value(Op), m_Specific(Remainder))) &&
IC.isKnownToBeAPowerOfTwo(Remainder, /*OrZero=*/true) &&
FalseVal == RemRes)
return FoldToBitwiseAnd(Remainder);
// Match the case where the one arm has been replaced by constant 1:
// %rem = srem i32 %n, 2
// %cnd = icmp slt i32 %rem, 0
// %sel = select i1 %cnd, i32 1, i32 %rem
if (match(TrueVal, m_One()) &&
match(RemRes, m_SRem(m_Value(Op), m_SpecificInt(2))) &&
FalseVal == RemRes)
return FoldToBitwiseAnd(ConstantInt::get(RemRes->getType(), 2));
return nullptr;
}
/// Given that \p CondVal is known to be \p CondIsTrue, try to simplify \p SI.
static Value *simplifyNestedSelectsUsingImpliedCond(SelectInst &SI,
Value *CondVal,
bool CondIsTrue,
const DataLayout &DL) {
Value *InnerCondVal = SI.getCondition();
Value *InnerTrueVal = SI.getTrueValue();
Value *InnerFalseVal = SI.getFalseValue();
assert(CondVal->getType() == InnerCondVal->getType() &&
"The type of inner condition must match with the outer.");
if (auto Implied = isImpliedCondition(CondVal, InnerCondVal, DL, CondIsTrue))
return *Implied ? InnerTrueVal : InnerFalseVal;
return nullptr;
}
Instruction *InstCombinerImpl::foldAndOrOfSelectUsingImpliedCond(Value *Op,
SelectInst &SI,
bool IsAnd) {
assert(Op->getType()->isIntOrIntVectorTy(1) &&
"Op must be either i1 or vector of i1.");
if (SI.getCondition()->getType() != Op->getType())
return nullptr;
if (Value *V = simplifyNestedSelectsUsingImpliedCond(SI, Op, IsAnd, DL))
return createSelectInstWithUnknownProfile(
Op, IsAnd ? V : ConstantInt::getTrue(Op->getType()),
IsAnd ? ConstantInt::getFalse(Op->getType()) : V);
return nullptr;
}
// Canonicalize select with fcmp to fabs(). -0.0 makes this tricky. We need
// fast-math-flags (nsz) or fsub with +0.0 (not fneg) for this to work.
static Instruction *foldSelectWithFCmpToFabs(SelectInst &SI,
InstCombinerImpl &IC) {
Value *CondVal = SI.getCondition();
bool ChangedFMF = false;
for (bool Swap : {false, true}) {
Value *TrueVal = SI.getTrueValue();
Value *X = SI.getFalseValue();
CmpPredicate Pred;
if (Swap)
std::swap(TrueVal, X);
if (!match(CondVal, m_FCmp(Pred, m_Specific(X), m_AnyZeroFP())))
continue;
// fold (X <= +/-0.0) ? (0.0 - X) : X to fabs(X), when 'Swap' is false
// fold (X > +/-0.0) ? X : (0.0 - X) to fabs(X), when 'Swap' is true
// Note: We require "nnan" for this fold because fcmp ignores the signbit
// of NAN, but IEEE-754 specifies the signbit of NAN values with
// fneg/fabs operations.
if (match(TrueVal, m_FSub(m_PosZeroFP(), m_Specific(X))) &&
(cast<FPMathOperator>(CondVal)->hasNoNaNs() || SI.hasNoNaNs() ||
(SI.hasOneUse() && canIgnoreSignBitOfNaN(*SI.use_begin())) ||
isKnownNeverNaN(X, IC.getSimplifyQuery().getWithInstruction(
cast<Instruction>(CondVal))))) {
if (!Swap && (Pred == FCmpInst::FCMP_OLE || Pred == FCmpInst::FCMP_ULE)) {
Value *Fabs = IC.Builder.CreateFAbs(X, &SI);
return IC.replaceInstUsesWith(SI, Fabs);
}
if (Swap && (Pred == FCmpInst::FCMP_OGT || Pred == FCmpInst::FCMP_UGT)) {
Value *Fabs = IC.Builder.CreateFAbs(X, &SI);
return IC.replaceInstUsesWith(SI, Fabs);
}
}
if (!match(TrueVal, m_FNeg(m_Specific(X))))
return nullptr;
// Forward-propagate nnan and ninf from the fcmp to the select.
// If all inputs are not those values, then the select is not either.
// Note: nsz is defined differently, so it may not be correct to propagate.
FastMathFlags FMF = cast<FPMathOperator>(CondVal)->getFastMathFlags();
if (FMF.noNaNs() && !SI.hasNoNaNs()) {
SI.setHasNoNaNs(true);
ChangedFMF = true;
}
if (FMF.noInfs() && !SI.hasNoInfs()) {
SI.setHasNoInfs(true);
ChangedFMF = true;
}
// Forward-propagate nnan from the fneg to the select.
// The nnan flag can be propagated iff fneg is selected when X is NaN.
if (!SI.hasNoNaNs() && cast<FPMathOperator>(TrueVal)->hasNoNaNs() &&
(Swap ? FCmpInst::isOrdered(Pred) : FCmpInst::isUnordered(Pred))) {
SI.setHasNoNaNs(true);
ChangedFMF = true;
}
// With nsz, when 'Swap' is false:
// fold (X < +/-0.0) ? -X : X or (X <= +/-0.0) ? -X : X to fabs(X)
// fold (X > +/-0.0) ? -X : X or (X >= +/-0.0) ? -X : X to -fabs(x)
// when 'Swap' is true:
// fold (X > +/-0.0) ? X : -X or (X >= +/-0.0) ? X : -X to fabs(X)
// fold (X < +/-0.0) ? X : -X or (X <= +/-0.0) ? X : -X to -fabs(X)
//
// Note: We require "nnan" for this fold because fcmp ignores the signbit
// of NAN, but IEEE-754 specifies the signbit of NAN values with
// fneg/fabs operations.
if (!SI.hasNoSignedZeros() &&
(!SI.hasOneUse() || !canIgnoreSignBitOfZero(*SI.use_begin())))
return nullptr;
if (!SI.hasNoNaNs() &&
(!SI.hasOneUse() || !canIgnoreSignBitOfNaN(*SI.use_begin())))
return nullptr;
if (Swap)
Pred = FCmpInst::getSwappedPredicate(Pred);
bool IsLTOrLE = Pred == FCmpInst::FCMP_OLT || Pred == FCmpInst::FCMP_OLE ||
Pred == FCmpInst::FCMP_ULT || Pred == FCmpInst::FCMP_ULE;
bool IsGTOrGE = Pred == FCmpInst::FCMP_OGT || Pred == FCmpInst::FCMP_OGE ||
Pred == FCmpInst::FCMP_UGT || Pred == FCmpInst::FCMP_UGE;
if (IsLTOrLE) {
Value *Fabs = IC.Builder.CreateFAbs(X, &SI);
return IC.replaceInstUsesWith(SI, Fabs);
}
if (IsGTOrGE) {
Value *Fabs = IC.Builder.CreateFAbs(X, &SI);
Instruction *NewFNeg = UnaryOperator::CreateFNeg(Fabs);
NewFNeg->setFastMathFlags(SI.getFastMathFlags());
return NewFNeg;
}
}
// Match select with (icmp slt (bitcast X to int), 0)
// or (icmp sgt (bitcast X to int), -1)
for (bool Swap : {false, true}) {
Value *TrueVal = SI.getTrueValue();
Value *X = SI.getFalseValue();
if (Swap)
std::swap(TrueVal, X);
CmpPredicate Pred;
const APInt *C;
bool TrueIfSigned;
if (!match(CondVal,
m_ICmp(Pred, m_ElementWiseBitCast(m_Specific(X)), m_APInt(C))) ||
!isSignBitCheck(Pred, *C, TrueIfSigned))
continue;
if (!match(TrueVal, m_FNeg(m_Specific(X))))
return nullptr;
if (Swap == TrueIfSigned && !CondVal->hasOneUse() && !TrueVal->hasOneUse())
return nullptr;
// Fold (IsNeg ? -X : X) or (!IsNeg ? X : -X) to fabs(X)
// Fold (IsNeg ? X : -X) or (!IsNeg ? -X : X) to -fabs(X)
Value *Fabs = IC.Builder.CreateFAbs(X, &SI);
if (Swap != TrueIfSigned)
return IC.replaceInstUsesWith(SI, Fabs);
return UnaryOperator::CreateFNegFMF(Fabs, &SI);
}
return ChangedFMF ? &SI : nullptr;
}
// Fold a select of an ordered fcmp using fabs of a NaN-scrubbed value:
// %s = select i1 (isnotnan T %x), T %x, T %y
// %a = call T @llvm.fabs.T(T %s)
// %c = fcmp <ordered-pred> T %a, %k
// %r = select i1 %c, T %s, T %y
// =>
// %a2 = call T @llvm.fabs.T(T %x)
// %c2 = fcmp <ordered-pred> T %a2, %k
// %r2 = select i1 %c2, T %x, T %y
static Instruction *
foldSelectOfOrderedFAbsCmpOfNaNScrubbedValue(SelectInst &SI,
InstCombinerImpl &IC) {
Instruction *OuterCmpI;
Value *Cmp0, *Cmp1;
if (!match(SI.getCondition(),
m_OneUse(m_Instruction(OuterCmpI,
m_FCmp(m_Value(Cmp0), m_Value(Cmp1))))))
return nullptr;
auto *OuterCmp = cast<FCmpInst>(OuterCmpI);
CmpInst::Predicate Pred = OuterCmp->getPredicate();
if (!FCmpInst::isOrdered(Pred))
return nullptr;
Value *Y = SI.getFalseValue();
Value *InnerSel = SI.getTrueValue();
// Match a select that returns X when X is not NaN, and Y otherwise:
// select (fcmp ord X, 0.0), X, Y
Value *X;
if (!match(InnerSel,
m_Select(m_OneUse(m_SpecificFCmp(FCmpInst::FCMP_ORD, m_Value(X),
m_AnyZeroFP())),
m_Deferred(X), m_Specific(Y))))
return nullptr;
Instruction *FAbsI;
auto MatchFAbsOfInnerSel = [&](Value *V) {
return match(V,
m_OneUse(m_Instruction(FAbsI, m_FAbs(m_Specific(InnerSel)))));
};
if (!MatchFAbsOfInnerSel(Cmp0)) {
if (!MatchFAbsOfInnerSel(Cmp1))
return nullptr;
std::swap(Cmp0, Cmp1);
Pred = CmpInst::getSwappedPredicate(Pred);
}
FastMathFlags FAbsFMF = FAbsI->getFastMathFlags();
FastMathFlags CmpFMF = OuterCmp->getFastMathFlags();
FastMathFlags CommonRewriteFMF =
FastMathFlags::intersectRewrite(FAbsFMF, CmpFMF);
// unionValue with FastMathFlags() drops all rewriter based flags
FastMathFlags NewFAbsFMF =
CommonRewriteFMF | FastMathFlags::unionValue(FAbsFMF, FastMathFlags());
FastMathFlags NewCmpFMF =
CommonRewriteFMF | FastMathFlags::unionValue(CmpFMF, FastMathFlags());
// When X is NaN, the old code evaluated fabs(Y), while the new code evaluates
// fabs(X). Do not preserve nnan on either newly-created instruction.
NewFAbsFMF.setNoNaNs(false);
NewCmpFMF.setNoNaNs(false);
Value *NewAbs = IC.Builder.CreateFAbs(X, FMFSource(NewFAbsFMF));
Value *NewCmp =
IC.Builder.CreateFCmpFMF(Pred, NewAbs, Cmp1, FMFSource(NewCmpFMF));
Value *NewSel = IC.Builder.CreateSelectFMF(NewCmp, X, Y, &SI);
return IC.replaceInstUsesWith(SI, NewSel);
}
// Match the following IR pattern:
// %x.lowbits = and i8 %x, %lowbitmask
// %x.lowbits.are.zero = icmp eq i8 %x.lowbits, 0
// %x.biased = add i8 %x, %bias
// %x.biased.highbits = and i8 %x.biased, %highbitmask
// %x.roundedup = select i1 %x.lowbits.are.zero, i8 %x, i8 %x.biased.highbits
// Define:
// %alignment = add i8 %lowbitmask, 1
// Iff 1. an %alignment is a power-of-two (aka, %lowbitmask is a low bit mask)
// and 2. %bias is equal to either %lowbitmask or %alignment,
// and 3. %highbitmask is equal to ~%lowbitmask (aka, to -%alignment)
// then this pattern can be transformed into:
// %x.offset = add i8 %x, %lowbitmask
// %x.roundedup = and i8 %x.offset, %highbitmask
static Value *
foldRoundUpIntegerWithPow2Alignment(SelectInst &SI,
InstCombiner::BuilderTy &Builder) {
Value *Cond = SI.getCondition();
Value *X = SI.getTrueValue();
Value *XBiasedHighBits = SI.getFalseValue();
CmpPredicate Pred;
Value *XLowBits;
if (!match(Cond, m_ICmp(Pred, m_Value(XLowBits), m_ZeroInt())) ||
!ICmpInst::isEquality(Pred))
return nullptr;
if (Pred == ICmpInst::Predicate::ICMP_NE)
std::swap(X, XBiasedHighBits);
// FIXME: we could support non non-splats here.
const APInt *LowBitMaskCst;
if (!match(XLowBits, m_And(m_Specific(X), m_APIntAllowPoison(LowBitMaskCst))))
return nullptr;
// Match even if the AND and ADD are swapped.
const APInt *BiasCst, *HighBitMaskCst;
if (!match(XBiasedHighBits,
m_And(m_Add(m_Specific(X), m_APIntAllowPoison(BiasCst)),
m_APIntAllowPoison(HighBitMaskCst))) &&
!match(XBiasedHighBits,
m_Add(m_And(m_Specific(X), m_APIntAllowPoison(HighBitMaskCst)),
m_APIntAllowPoison(BiasCst))))
return nullptr;
if (!LowBitMaskCst->isMask())
return nullptr;
APInt InvertedLowBitMaskCst = ~*LowBitMaskCst;
if (InvertedLowBitMaskCst != *HighBitMaskCst)
return nullptr;
APInt AlignmentCst = *LowBitMaskCst + 1;
if (*BiasCst != AlignmentCst && *BiasCst != *LowBitMaskCst)
return nullptr;
if (!XBiasedHighBits->hasOneUse()) {
// We can't directly return XBiasedHighBits if it is more poisonous.
if (*BiasCst == *LowBitMaskCst && impliesPoison(XBiasedHighBits, X))
return XBiasedHighBits;
return nullptr;
}
// FIXME: could we preserve undef's here?
Type *Ty = X->getType();
Value *XOffset = Builder.CreateAdd(X, ConstantInt::get(Ty, *LowBitMaskCst),
X->getName() + ".biased");
Value *R = Builder.CreateAnd(XOffset, ConstantInt::get(Ty, *HighBitMaskCst));
R->takeName(&SI);
return R;
}
namespace {
struct DecomposedSelect {
Value *Cond = nullptr;
Value *TrueVal = nullptr;
Value *FalseVal = nullptr;
};
} // namespace
/// Folds patterns like:
/// select c2 (select c1 a b) (select c1 b a)
/// into:
/// select (xor c1 c2) b a
static Instruction *
foldSelectOfSymmetricSelect(SelectInst &OuterSelVal,
InstCombiner::BuilderTy &Builder) {
Value *OuterCond, *InnerCond, *InnerTrueVal, *InnerFalseVal;
if (!match(
&OuterSelVal,
m_Select(m_Value(OuterCond),
m_OneUse(m_Select(m_Value(InnerCond), m_Value(InnerTrueVal),
m_Value(InnerFalseVal))),
m_OneUse(m_Select(m_Deferred(InnerCond),
m_Deferred(InnerFalseVal),
m_Deferred(InnerTrueVal))))))
return nullptr;
if (OuterCond->getType() != InnerCond->getType())
return nullptr;
Value *Xor = Builder.CreateXor(InnerCond, OuterCond);
return SelectInst::Create(Xor, InnerFalseVal, InnerTrueVal);
}
/// Look for patterns like
/// %outer.cond = select i1 %inner.cond, i1 %alt.cond, i1 false
/// %inner.sel = select i1 %inner.cond, i8 %inner.sel.t, i8 %inner.sel.f
/// %outer.sel = select i1 %outer.cond, i8 %outer.sel.t, i8 %inner.sel
/// and rewrite it as
/// %inner.sel = select i1 %cond.alternative, i8 %sel.outer.t, i8 %sel.inner.t
/// %sel.outer = select i1 %cond.inner, i8 %inner.sel, i8 %sel.inner.f
static Instruction *foldNestedSelects(SelectInst &OuterSelVal,
InstCombiner::BuilderTy &Builder) {
// We must start with a `select`.
DecomposedSelect OuterSel;
match(&OuterSelVal,
m_Select(m_Value(OuterSel.Cond), m_Value(OuterSel.TrueVal),
m_Value(OuterSel.FalseVal)));
// Canonicalize inversion of the outermost `select`'s condition.
if (match(OuterSel.Cond, m_Not(m_Value(OuterSel.Cond))))
std::swap(OuterSel.TrueVal, OuterSel.FalseVal);
// The condition of the outermost select must be an `and`/`or`.
if (!match(OuterSel.Cond, m_c_LogicalOp(m_Value(), m_Value())))
return nullptr;
// Depending on the logical op, inner select might be in different hand.
bool IsAndVariant = match(OuterSel.Cond, m_LogicalAnd());
Value *InnerSelVal = IsAndVariant ? OuterSel.FalseVal : OuterSel.TrueVal;
// Profitability check - avoid increasing instruction count.
if (none_of(ArrayRef<Value *>({OuterSelVal.getCondition(), InnerSelVal}),
match_fn(m_OneUse(m_Value()))))
return nullptr;
// The appropriate hand of the outermost `select` must be a select itself.
DecomposedSelect InnerSel;
if (!match(InnerSelVal,
m_Select(m_Value(InnerSel.Cond), m_Value(InnerSel.TrueVal),
m_Value(InnerSel.FalseVal))))
return nullptr;
// Canonicalize inversion of the innermost `select`'s condition.
if (match(InnerSel.Cond, m_Not(m_Value(InnerSel.Cond))))
std::swap(InnerSel.TrueVal, InnerSel.FalseVal);
Value *AltCond = nullptr;
auto matchOuterCond = [OuterSel, IsAndVariant, &AltCond](auto m_InnerCond) {
// An unsimplified select condition can match both LogicalAnd and LogicalOr
// (select true, true, false). Since below we assume that LogicalAnd implies
// InnerSel match the FVal and vice versa for LogicalOr, we can't match the
// alternative pattern here.
return IsAndVariant ? match(OuterSel.Cond,
m_c_LogicalAnd(m_InnerCond, m_Value(AltCond)))
: match(OuterSel.Cond,
m_c_LogicalOr(m_InnerCond, m_Value(AltCond)));
};
// Finally, match the condition that was driving the outermost `select`,
// it should be a logical operation between the condition that was driving
// the innermost `select` (after accounting for the possible inversions
// of the condition), and some other condition.
if (matchOuterCond(m_Specific(InnerSel.Cond))) {
// Done!
} else if (Value * NotInnerCond; matchOuterCond(m_CombineAnd(
m_Not(m_Specific(InnerSel.Cond)), m_Value(NotInnerCond)))) {
// Done!
std::swap(InnerSel.TrueVal, InnerSel.FalseVal);
InnerSel.Cond = NotInnerCond;
} else // Not the pattern we were looking for.
return nullptr;
Value *SelInner = Builder.CreateSelect(
AltCond, IsAndVariant ? OuterSel.TrueVal : InnerSel.FalseVal,
IsAndVariant ? InnerSel.TrueVal : OuterSel.FalseVal);
SelInner->takeName(InnerSelVal);
return SelectInst::Create(InnerSel.Cond,
IsAndVariant ? SelInner : InnerSel.TrueVal,
!IsAndVariant ? SelInner : InnerSel.FalseVal);
}
/// Return true if V is poison or \p Expected given that ValAssumedPoison is
/// already poison. For example, if ValAssumedPoison is `icmp samesign X, 10`
/// and V is `icmp ne X, 5`, impliesPoisonOrCond returns true.
static bool impliesPoisonOrCond(const Value *ValAssumedPoison, const Value *V,
bool Expected, const SimplifyQuery &SQ) {
if (impliesPoison(ValAssumedPoison, V))
return true;
// Handle the case that ValAssumedPoison is `icmp samesign pred X, C1` and V
// is `icmp pred X, C2`, where C1 is well-defined.
if (auto *ICmp = dyn_cast<ICmpInst>(ValAssumedPoison)) {
Value *LHS = ICmp->getOperand(0);
const APInt *RHSC1;
const APInt *RHSC2;
CmpPredicate Pred;
if (ICmp->hasSameSign() &&
match(ICmp->getOperand(1), m_APIntForbidPoison(RHSC1)) &&
match(V, m_ICmp(Pred, m_Specific(LHS), m_APIntAllowPoison(RHSC2)))) {
unsigned BitWidth = RHSC1->getBitWidth();
ConstantRange CRX =
RHSC1->isNonNegative()
? ConstantRange(APInt::getSignedMinValue(BitWidth),
APInt::getZero(BitWidth))
: ConstantRange(APInt::getZero(BitWidth),
APInt::getSignedMinValue(BitWidth));
return CRX.icmp(Expected ? Pred : ICmpInst::getInverseCmpPredicate(Pred),
*RHSC2);
}
}
// For non-poison X in [0, 1], `trunc nuw X to i1` is not poison, but an
// additional `nsw` flag makes it poison for X == 1.
Value *A;
if (match(ValAssumedPoison, m_NUWTrunc(m_Value(A))) &&
!cast<TruncInst>(ValAssumedPoison)->hasNoSignedWrap() &&
isGuaranteedNotToBePoison(A)) {
assert(ValAssumedPoison->getType()->isIntOrIntVectorTy(1));
return computeKnownBits(
A, SQ.getWithInstruction(cast<Instruction>(ValAssumedPoison)))
.getMaxValue() == 1;
}
return false;
}
Instruction *InstCombinerImpl::foldSelectOfBools(SelectInst &SI) {
Value *CondVal = SI.getCondition();
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
Type *SelType = SI.getType();
// Avoid potential infinite loops by checking for non-constant condition.
// TODO: Can we assert instead by improving canonicalizeSelectToShuffle()?
// Scalar select must have simplified?
if (!SelType->isIntOrIntVectorTy(1) || isa<Constant>(CondVal) ||
TrueVal->getType() != CondVal->getType())
return nullptr;
auto *One = ConstantInt::getTrue(SelType);
auto *Zero = ConstantInt::getFalse(SelType);
Value *A, *B, *C, *D;
// Folding select to and/or i1 isn't poison safe in general. impliesPoison
// checks whether folding it does not convert a well-defined value into
// poison.
if (match(TrueVal, m_One())) {
if (impliesPoisonOrCond(FalseVal, CondVal, /*Expected=*/false, SQ)) {
// Change: A = select B, true, C --> A = or B, C
return BinaryOperator::CreateOr(CondVal, FalseVal);
}
if (match(CondVal, m_OneUse(m_Select(m_Value(A), m_One(), m_Value(B)))) &&
impliesPoisonOrCond(FalseVal, B, /*Expected=*/false, SQ)) {
// (A || B) || C --> A || (B | C)
Value *LOr = Builder.CreateLogicalOr(A, Builder.CreateOr(B, FalseVal));
if (auto *I = dyn_cast<Instruction>(LOr)) {
setExplicitlyUnknownBranchWeightsIfProfiled(*I, DEBUG_TYPE);
}
return replaceInstUsesWith(SI, LOr);
}
// (A && B) || (C && B) --> (A || C) && B
if (match(CondVal, m_LogicalAnd(m_Value(A), m_Value(B))) &&
match(FalseVal, m_LogicalAnd(m_Value(C), m_Value(D))) &&
(CondVal->hasOneUse() || FalseVal->hasOneUse())) {
bool CondLogicAnd = isa<SelectInst>(CondVal);
bool FalseLogicAnd = isa<SelectInst>(FalseVal);
auto AndFactorization = [&](Value *Common, Value *InnerCond,
Value *InnerVal,
bool SelFirst = false) -> Instruction * {
Value *InnerSel = Builder.CreateSelectWithUnknownProfile(
InnerCond, One, InnerVal, DEBUG_TYPE);
if (SelFirst)
std::swap(Common, InnerSel);
if (FalseLogicAnd || (CondLogicAnd && Common == A))
return createSelectInstWithUnknownProfile(Common, InnerSel, Zero);
else
return BinaryOperator::CreateAnd(Common, InnerSel);
};
if (A == C)
return AndFactorization(A, B, D);
if (A == D)
return AndFactorization(A, B, C);
if (B == C)
return AndFactorization(B, A, D);
if (B == D)
return AndFactorization(B, A, C, CondLogicAnd && FalseLogicAnd);
}
}
if (match(FalseVal, m_Zero())) {
if (impliesPoisonOrCond(TrueVal, CondVal, /*Expected=*/true, SQ)) {
// Change: A = select B, C, false --> A = and B, C
return BinaryOperator::CreateAnd(CondVal, TrueVal);
}
if (match(CondVal, m_OneUse(m_Select(m_Value(A), m_Value(B), m_Zero()))) &&
impliesPoisonOrCond(TrueVal, B, /*Expected=*/true, SQ)) {
// (A && B) && C --> A && (B & C)
Value *LAnd = Builder.CreateLogicalAnd(A, Builder.CreateAnd(B, TrueVal));
if (auto *I = dyn_cast<Instruction>(LAnd)) {
setExplicitlyUnknownBranchWeightsIfProfiled(*I, DEBUG_TYPE);
}
return replaceInstUsesWith(SI, LAnd);
}
// (A || B) && (C || B) --> (A && C) || B
if (match(CondVal, m_LogicalOr(m_Value(A), m_Value(B))) &&
match(TrueVal, m_LogicalOr(m_Value(C), m_Value(D))) &&
(CondVal->hasOneUse() || TrueVal->hasOneUse())) {
bool CondLogicOr = isa<SelectInst>(CondVal);
bool TrueLogicOr = isa<SelectInst>(TrueVal);
auto OrFactorization = [&](Value *Common, Value *InnerCond,
Value *InnerVal,
bool SelFirst = false) -> Instruction * {
Value *InnerSel = Builder.CreateSelectWithUnknownProfile(
InnerCond, InnerVal, Zero, DEBUG_TYPE);
if (SelFirst)
std::swap(Common, InnerSel);
if (TrueLogicOr || (CondLogicOr && Common == A))
return createSelectInstWithUnknownProfile(Common, One, InnerSel);
else
return BinaryOperator::CreateOr(Common, InnerSel);
};
if (A == C)
return OrFactorization(A, B, D);
if (A == D)
return OrFactorization(A, B, C);
if (B == C)
return OrFactorization(B, A, D);
if (B == D)
return OrFactorization(B, A, C, CondLogicOr && TrueLogicOr);
}
}
// We match the "full" 0 or 1 constant here to avoid a potential infinite
// loop with vectors that may have undefined/poison elements.
// select a, false, b -> select !a, b, false
if (match(TrueVal, m_Specific(Zero))) {
Value *NotCond = Builder.CreateNot(CondVal, "not." + CondVal->getName());
SelectInst *NewSI = SelectInst::Create(NotCond, FalseVal, Zero, "", nullptr,
/*MDFrom=*/&SI);
NewSI->swapProfMetadata();
return NewSI;
}
// select a, b, true -> select !a, true, b
if (match(FalseVal, m_Specific(One))) {
Value *NotCond = Builder.CreateNot(CondVal, "not." + CondVal->getName());
SelectInst *NewSI =
SelectInst::Create(NotCond, One, TrueVal, "", nullptr, /*MDFrom=*/&SI);
NewSI->swapProfMetadata();
return NewSI;
}
// DeMorgan in select form: !a && !b --> !(a || b)
// select !a, !b, false --> not (select a, true, b)
if (match(&SI, m_LogicalAnd(m_Not(m_Value(A)), m_Not(m_Value(B)))) &&
(CondVal->hasOneUse() || TrueVal->hasOneUse()) &&
!match(A, m_ConstantExpr()) && !match(B, m_ConstantExpr())) {
SelectInst *NewSI =
cast<SelectInst>(Builder.CreateSelect(A, One, B, "", /*MDFrom=*/&SI));
NewSI->swapProfMetadata();
return BinaryOperator::CreateNot(NewSI);
}
// DeMorgan in select form: !a || !b --> !(a && b)
// select !a, true, !b --> not (select a, b, false)
if (match(&SI, m_LogicalOr(m_Not(m_Value(A)), m_Not(m_Value(B)))) &&
(CondVal->hasOneUse() || FalseVal->hasOneUse()) &&
!match(A, m_ConstantExpr()) && !match(B, m_ConstantExpr())) {
SelectInst *NewSI =
cast<SelectInst>(Builder.CreateSelect(A, B, Zero, "", /*MDFrom=*/&SI));
NewSI->swapProfMetadata();
return BinaryOperator::CreateNot(NewSI);
}
// select (select a, true, b), true, b -> select a, true, b
if (match(CondVal, m_Select(m_Value(A), m_One(), m_Value(B))) &&
match(TrueVal, m_One()) && match(FalseVal, m_Specific(B)))
return replaceOperand(SI, 0, A);
// select (select a, b, false), b, false -> select a, b, false
if (match(CondVal, m_Select(m_Value(A), m_Value(B), m_Zero())) &&
match(TrueVal, m_Specific(B)) && match(FalseVal, m_Zero()))
return replaceOperand(SI, 0, A);
// ~(A & B) & (A | B) --> A ^ B
if (match(&SI, m_c_LogicalAnd(m_Not(m_LogicalAnd(m_Value(A), m_Value(B))),
m_c_LogicalOr(m_Deferred(A), m_Deferred(B)))))
return BinaryOperator::CreateXor(A, B);
// select (~a | c), a, b -> select a, (select c, true, b), false
if (match(CondVal,
m_OneUse(m_c_Or(m_Not(m_Specific(TrueVal)), m_Value(C))))) {
// TODO(#183864): We could improve the profile if P(~a | c) < 0.5, which
// implies strong bounds on both operands (P(a) is high, P(c) is low).
Value *OrV =
Builder.CreateSelectWithUnknownProfile(C, One, FalseVal, DEBUG_TYPE);
return createSelectInstWithUnknownProfile(TrueVal, OrV, Zero);
}
// select (c & b), a, b -> select b, (select ~c, true, a), false
if (match(CondVal, m_OneUse(m_c_And(m_Value(C), m_Specific(FalseVal))))) {
if (Value *NotC = getFreelyInverted(C, C->hasOneUse(), &Builder)) {
Value *OrV = Builder.CreateSelectWithUnknownProfile(NotC, One, TrueVal,
DEBUG_TYPE);
return createSelectInstWithUnknownProfile(FalseVal, OrV, Zero);
}
}
// select (a | c), a, b -> select a, true, (select ~c, b, false)
if (match(CondVal, m_OneUse(m_c_Or(m_Specific(TrueVal), m_Value(C))))) {
if (Value *NotC = getFreelyInverted(C, C->hasOneUse(), &Builder)) {
// TODO(#183864): We could improve the profile if P(a | c) < 0.5, which
// implies strong bounds on both operands (both P(a) and P(c) are low).
Value *AndV = Builder.CreateSelectWithUnknownProfile(NotC, FalseVal, Zero,
DEBUG_TYPE);
return createSelectInstWithUnknownProfile(TrueVal, One, AndV);
}
}
// select (c & ~b), a, b -> select b, true, (select c, a, false)
if (match(CondVal,
m_OneUse(m_c_And(m_Value(C), m_Not(m_Specific(FalseVal)))))) {
Value *AndV =
Builder.CreateSelectWithUnknownProfile(C, TrueVal, Zero, DEBUG_TYPE);
return createSelectInstWithUnknownProfile(FalseVal, One, AndV);
}
if (match(FalseVal, m_Zero()) || match(TrueVal, m_One())) {
Use *Y = nullptr;
bool IsAnd = match(FalseVal, m_Zero()) ? true : false;
Value *Op1 = IsAnd ? TrueVal : FalseVal;
if (isCheckForZeroAndMulWithOverflow(CondVal, Op1, IsAnd, Y)) {
auto *FI = new FreezeInst(*Y, (*Y)->getName() + ".fr");
InsertNewInstBefore(FI, cast<Instruction>(Y->getUser())->getIterator());
replaceUse(*Y, FI);
return replaceInstUsesWith(SI, Op1);
}
if (auto *V = foldBooleanAndOr(CondVal, Op1, SI, IsAnd,
/*IsLogical=*/true))
return replaceInstUsesWith(SI, V);
}
// select (a || b), c, false -> select a, c, false
// select c, (a || b), false -> select c, a, false
// if c implies that b is false.
if (match(CondVal, m_LogicalOr(m_Value(A), m_Value(B))) &&
match(FalseVal, m_Zero())) {
std::optional<bool> Res = isImpliedCondition(TrueVal, B, DL);
if (Res && *Res == false)
return replaceOperand(SI, 0, A);
}
if (match(TrueVal, m_LogicalOr(m_Value(A), m_Value(B))) &&
match(FalseVal, m_Zero())) {
std::optional<bool> Res = isImpliedCondition(CondVal, B, DL);
if (Res && *Res == false)
return replaceOperand(SI, 1, A);
}
// select c, true, (a && b) -> select c, true, a
// select (a && b), true, c -> select a, true, c
// if c = false implies that b = true
if (match(TrueVal, m_One()) &&
match(FalseVal, m_LogicalAnd(m_Value(A), m_Value(B)))) {
std::optional<bool> Res = isImpliedCondition(CondVal, B, DL, false);
if (Res && *Res == true)
return replaceOperand(SI, 2, A);
}
if (match(CondVal, m_LogicalAnd(m_Value(A), m_Value(B))) &&
match(TrueVal, m_One())) {
std::optional<bool> Res = isImpliedCondition(FalseVal, B, DL, false);
if (Res && *Res == true)
return replaceOperand(SI, 0, A);
}
if (match(TrueVal, m_One())) {
// (C && A) || (!C && B) --> select C, A, B (and similar cases)
if (auto *V = FoldOrOfLogicalAnds(CondVal, FalseVal)) {
return V;
}
}
return nullptr;
}
// Return true if we can safely remove the select instruction for std::bit_ceil
// pattern.
static bool isSafeToRemoveBitCeilSelect(ICmpInst::Predicate Pred, Value *Cond0,
const APInt *Cond1, Value *CtlzOp,
unsigned BitWidth,
bool &ShouldDropNoWrap) {
// The challenge in recognizing std::bit_ceil(X) is that the operand is used
// for the CTLZ proper and select condition, each possibly with some
// operation like add and sub.
//
// Our aim is to make sure that -ctlz & (BitWidth - 1) == 0 even when the
// select instruction would select 1, which allows us to get rid of the select
// instruction.
//
// To see if we can do so, we do some symbolic execution with ConstantRange.
// Specifically, we compute the range of values that Cond0 could take when
// Cond == false. Then we successively transform the range until we obtain
// the range of values that CtlzOp could take.
//
// Conceptually, we follow the def-use chain backward from Cond0 while
// transforming the range for Cond0 until we meet the common ancestor of Cond0
// and CtlzOp. Then we follow the def-use chain forward until we obtain the
// range for CtlzOp. That said, we only follow at most one ancestor from
// Cond0. Likewise, we only follow at most one ancestor from CtrlOp.
ConstantRange CR = ConstantRange::makeExactICmpRegion(
CmpInst::getInversePredicate(Pred), *Cond1);
ShouldDropNoWrap = false;
// Match the operation that's used to compute CtlzOp from CommonAncestor. If
// CtlzOp == CommonAncestor, return true as no operation is needed. If a
// match is found, execute the operation on CR, update CR, and return true.
// Otherwise, return false.
auto MatchForward = [&](Value *CommonAncestor) {
const APInt *C = nullptr;
if (CtlzOp == CommonAncestor)
return true;
if (match(CtlzOp, m_Add(m_Specific(CommonAncestor), m_APInt(C)))) {
ShouldDropNoWrap = true;
CR = CR.add(*C);
return true;
}
if (match(CtlzOp, m_Sub(m_APInt(C), m_Specific(CommonAncestor)))) {
ShouldDropNoWrap = true;
CR = ConstantRange(*C).sub(CR);
return true;
}
if (match(CtlzOp, m_Not(m_Specific(CommonAncestor)))) {
CR = CR.binaryNot();
return true;
}
return false;
};
const APInt *C = nullptr;
Value *CommonAncestor;
if (MatchForward(Cond0)) {
// Cond0 is either CtlzOp or CtlzOp's parent. CR has been updated.
} else if (match(Cond0, m_Add(m_Value(CommonAncestor), m_APInt(C)))) {
CR = CR.sub(*C);
if (!MatchForward(CommonAncestor))
return false;
// Cond0's parent is either CtlzOp or CtlzOp's parent. CR has been updated.
} else {
return false;
}
// Return true if all the values in the range are either 0 or negative (if
// treated as signed). We do so by evaluating:
//
// CR - 1 u>= (1 << BitWidth) - 1.
APInt IntMax = APInt::getSignMask(BitWidth) - 1;
CR = CR.sub(APInt(BitWidth, 1));
return CR.icmp(ICmpInst::ICMP_UGE, IntMax);
}
// Transform the std::bit_ceil(X) pattern like:
//
// %dec = add i32 %x, -1
// %ctlz = tail call i32 @llvm.ctlz.i32(i32 %dec, i1 false)
// %sub = sub i32 32, %ctlz
// %shl = shl i32 1, %sub
// %ugt = icmp ugt i32 %x, 1
// %sel = select i1 %ugt, i32 %shl, i32 1
//
// into:
//
// %dec = add i32 %x, -1
// %ctlz = tail call i32 @llvm.ctlz.i32(i32 %dec, i1 false)
// %neg = sub i32 0, %ctlz
// %masked = and i32 %ctlz, 31
// %shl = shl i32 1, %sub
//
// Note that the select is optimized away while the shift count is masked with
// 31. We handle some variations of the input operand like std::bit_ceil(X +
// 1).
static Instruction *foldBitCeil(SelectInst &SI, IRBuilderBase &Builder,
InstCombinerImpl &IC) {
Type *SelType = SI.getType();
unsigned BitWidth = SelType->getScalarSizeInBits();
if (!isPowerOf2_32(BitWidth))
return nullptr;
Value *FalseVal = SI.getFalseValue();
Value *TrueVal = SI.getTrueValue();
CmpPredicate Pred;
const APInt *Cond1;
Value *Cond0, *Ctlz, *CtlzOp;
if (!match(SI.getCondition(), m_ICmp(Pred, m_Value(Cond0), m_APInt(Cond1))))
return nullptr;
if (match(TrueVal, m_One())) {
std::swap(FalseVal, TrueVal);
Pred = CmpInst::getInversePredicate(Pred);
}
bool ShouldDropNoWrap;
if (!match(FalseVal, m_One()) ||
!match(TrueVal,
m_OneUse(m_Shl(m_One(), m_OneUse(m_Sub(m_SpecificInt(BitWidth),
m_Value(Ctlz)))))) ||
!match(Ctlz, m_Ctlz(m_Value(CtlzOp), m_Value())) ||
!isSafeToRemoveBitCeilSelect(Pred, Cond0, Cond1, CtlzOp, BitWidth,
ShouldDropNoWrap))
return nullptr;
if (ShouldDropNoWrap) {
cast<Instruction>(CtlzOp)->setHasNoUnsignedWrap(false);
cast<Instruction>(CtlzOp)->setHasNoSignedWrap(false);
}
// Build 1 << (-CTLZ & (BitWidth-1)). The negation likely corresponds to a
// single hardware instruction as opposed to BitWidth - CTLZ, where BitWidth
// is an integer constant. Masking with BitWidth-1 comes free on some
// hardware as part of the shift instruction.
// Drop range attributes and re-infer them in the next iteration.
cast<Instruction>(Ctlz)->dropPoisonGeneratingAnnotations();
IC.addToWorklist(cast<Instruction>(Ctlz));
Value *Neg = Builder.CreateNeg(Ctlz);
Value *Masked =
Builder.CreateAnd(Neg, ConstantInt::get(SelType, BitWidth - 1));
return BinaryOperator::Create(Instruction::Shl, ConstantInt::get(SelType, 1),
Masked);
}
// This function tries to fold the following operations:
// (x < y) ? -1 : zext(x != y)
// (x < y) ? -1 : zext(x > y)
// (x > y) ? 1 : sext(x != y)
// (x > y) ? 1 : sext(x < y)
// (x == y) ? 0 : (x > y ? 1 : -1)
// (x == y) ? 0 : (x < y ? -1 : 1)
// Special case: x == C ? 0 : (x > C - 1 ? 1 : -1)
// Special case: x == C ? 0 : (x < C + 1 ? -1 : 1)
// Into ucmp/scmp(x, y), where signedness is determined by the signedness
// of the comparison in the original sequence.
Instruction *InstCombinerImpl::foldSelectToCmp(SelectInst &SI) {
Value *TV = SI.getTrueValue();
Value *FV = SI.getFalseValue();
CmpPredicate Pred;
Value *LHS, *RHS;
if (!match(SI.getCondition(), m_ICmp(Pred, m_Value(LHS), m_Value(RHS))))
return nullptr;
if (!LHS->getType()->isIntOrIntVectorTy())
return nullptr;
// If there is no -1, 0 or 1 at TV, then invert the select statement and try
// to canonicalize to one of the forms above
if (!isa<Constant>(TV)) {
if (!isa<Constant>(FV))
return nullptr;
Pred = ICmpInst::getInverseCmpPredicate(Pred);
std::swap(TV, FV);
}
if (ICmpInst::isNonStrictPredicate(Pred)) {
if (Constant *C = dyn_cast<Constant>(RHS)) {
auto FlippedPredAndConst =
getFlippedStrictnessPredicateAndConstant(Pred, C);
if (!FlippedPredAndConst)
return nullptr;
Pred = FlippedPredAndConst->first;
RHS = FlippedPredAndConst->second;
} else {
return nullptr;
}
}
// Try to swap operands and the predicate. We need to be careful when doing
// so because two of the patterns have opposite predicates, so use the
// constant inside select to determine if swapping operands would be
// beneficial to us.
if ((ICmpInst::isGT(Pred) && match(TV, m_AllOnes())) ||
(ICmpInst::isLT(Pred) && match(TV, m_One()))) {
Pred = ICmpInst::getSwappedPredicate(Pred);
std::swap(LHS, RHS);
}
bool IsSigned = ICmpInst::isSigned(Pred);
bool Replace = false;
CmpPredicate ExtendedCmpPredicate;
// (x < y) ? -1 : zext(x != y)
// (x < y) ? -1 : zext(x > y)
if (ICmpInst::isLT(Pred) && match(TV, m_AllOnes()) &&
match(FV, m_ZExt(m_c_ICmp(ExtendedCmpPredicate, m_Specific(LHS),
m_Specific(RHS)))) &&
(ExtendedCmpPredicate == ICmpInst::ICMP_NE ||
ICmpInst::getSwappedPredicate(ExtendedCmpPredicate) == Pred))
Replace = true;
// (x > y) ? 1 : sext(x != y)
// (x > y) ? 1 : sext(x < y)
if (ICmpInst::isGT(Pred) && match(TV, m_One()) &&
match(FV, m_SExt(m_c_ICmp(ExtendedCmpPredicate, m_Specific(LHS),
m_Specific(RHS)))) &&
(ExtendedCmpPredicate == ICmpInst::ICMP_NE ||
ICmpInst::getSwappedPredicate(ExtendedCmpPredicate) == Pred))
Replace = true;
// (x == y) ? 0 : (x > y ? 1 : -1)
CmpPredicate FalseBranchSelectPredicate;
const APInt *InnerTV, *InnerFV;
if (Pred == ICmpInst::ICMP_EQ && match(TV, m_Zero()) &&
match(FV, m_Select(m_c_ICmp(FalseBranchSelectPredicate, m_Specific(LHS),
m_Specific(RHS)),
m_APInt(InnerTV), m_APInt(InnerFV)))) {
if (!ICmpInst::isGT(FalseBranchSelectPredicate)) {
FalseBranchSelectPredicate =
ICmpInst::getSwappedPredicate(FalseBranchSelectPredicate);
std::swap(LHS, RHS);
}
if (!InnerTV->isOne()) {
std::swap(InnerTV, InnerFV);
std::swap(LHS, RHS);
}
if (ICmpInst::isGT(FalseBranchSelectPredicate) && InnerTV->isOne() &&
InnerFV->isAllOnes()) {
IsSigned = ICmpInst::isSigned(FalseBranchSelectPredicate);
Replace = true;
}
}
// Special cases with constants: x == C ? 0 : (x > C-1 ? 1 : -1)
if (Pred == ICmpInst::ICMP_EQ && match(TV, m_Zero())) {
const APInt *C;
if (match(RHS, m_APInt(C))) {
CmpPredicate InnerPred;
Value *InnerRHS;
const APInt *InnerTV, *InnerFV;
if (match(FV,
m_Select(m_ICmp(InnerPred, m_Specific(LHS), m_Value(InnerRHS)),
m_APInt(InnerTV), m_APInt(InnerFV)))) {
// x == C ? 0 : (x > C-1 ? 1 : -1)
if (ICmpInst::isGT(InnerPred) && InnerTV->isOne() &&
InnerFV->isAllOnes()) {
IsSigned = ICmpInst::isSigned(InnerPred);
bool CanSubOne = IsSigned ? !C->isMinSignedValue() : !C->isMinValue();
if (CanSubOne) {
APInt Cminus1 = *C - 1;
if (match(InnerRHS, m_SpecificInt(Cminus1)))
Replace = true;
}
}
// x == C ? 0 : (x < C+1 ? -1 : 1)
if (ICmpInst::isLT(InnerPred) && InnerTV->isAllOnes() &&
InnerFV->isOne()) {
IsSigned = ICmpInst::isSigned(InnerPred);
bool CanAddOne = IsSigned ? !C->isMaxSignedValue() : !C->isMaxValue();
if (CanAddOne) {
APInt Cplus1 = *C + 1;
if (match(InnerRHS, m_SpecificInt(Cplus1)))
Replace = true;
}
}
}
}
}
Intrinsic::ID IID = IsSigned ? Intrinsic::scmp : Intrinsic::ucmp;
if (Replace)
return replaceInstUsesWith(
SI, Builder.CreateIntrinsic(SI.getType(), IID, {LHS, RHS}));
return nullptr;
}
bool InstCombinerImpl::fmulByZeroIsZero(Value *MulVal, FastMathFlags FMF,
const Instruction *CtxI) const {
KnownFPClass Known =
computeKnownFPClass(MulVal, FMF, fcNegative, SQ.getWithInstruction(CtxI));
return Known.isKnownNeverNaN() && Known.isKnownNeverInfinity() &&
(FMF.noSignedZeros() || Known.signBitIsZeroOrNaN());
}
static bool matchFMulByZeroIfResultEqZero(InstCombinerImpl &IC, Value *Cmp0,
Value *Cmp1, Value *TrueVal,
Value *FalseVal, Instruction &CtxI,
bool SelectIsNSZ) {
Value *MulRHS;
if (match(Cmp1, m_PosZeroFP()) &&
match(TrueVal, m_c_FMul(m_Specific(Cmp0), m_Value(MulRHS)))) {
FastMathFlags FMF = cast<FPMathOperator>(TrueVal)->getFastMathFlags();
// nsz must be on the select, it must be ignored on the multiply. We
// need nnan and ninf on the multiply for the other value.
FMF.setNoSignedZeros(SelectIsNSZ);
return IC.fmulByZeroIsZero(MulRHS, FMF, &CtxI);
}
return false;
}
/// Check whether the KnownBits of a select arm may be affected by the
/// select condition.
static bool hasAffectedValue(Value *V, SmallPtrSetImpl<Value *> &Affected,
unsigned Depth) {
if (Depth == MaxAnalysisRecursionDepth)
return false;
// Ignore the case where the select arm itself is affected. These cases
// are handled more efficiently by other optimizations.
if (Depth != 0 && Affected.contains(V))
return true;
if (auto *I = dyn_cast<Instruction>(V)) {
if (isa<PHINode>(I)) {
if (Depth == MaxAnalysisRecursionDepth - 1)
return false;
Depth = MaxAnalysisRecursionDepth - 2;
}
return any_of(I->operands(), [&](Value *Op) {
return Op->getType()->isIntOrIntVectorTy() &&
hasAffectedValue(Op, Affected, Depth + 1);
});
}
return false;
}
// This transformation enables the possibility of transforming fcmp + sel into
// a fmaxnum/fminnum intrinsic.
static Value *foldSelectIntoAddConstant(SelectInst &SI,
InstCombiner::BuilderTy &Builder) {
// Do this transformation only when select instruction gives NaN and NSZ
// guarantee.
auto *SIFOp = dyn_cast<FPMathOperator>(&SI);
if (!SIFOp || !SIFOp->hasNoSignedZeros() || !SIFOp->hasNoNaNs())
return nullptr;
auto TryFoldIntoAddConstant =
[&Builder, &SI](CmpInst::Predicate Pred, Value *X, Value *Z,
Instruction *FAdd, Constant *C, bool Swapped) -> Value * {
// Only these relational predicates can be transformed into maxnum/minnum
// intrinsic.
if (!CmpInst::isRelational(Pred) || !match(Z, m_AnyZeroFP()))
return nullptr;
if (!match(FAdd, m_FAdd(m_Specific(X), m_Specific(C))))
return nullptr;
Value *NewSelect = Builder.CreateSelect(SI.getCondition(), Swapped ? Z : X,
Swapped ? X : Z, "", &SI);
NewSelect->takeName(&SI);
Value *NewFAdd = Builder.CreateFAdd(NewSelect, C);
NewFAdd->takeName(FAdd);
// Propagate FastMath flags
FastMathFlags SelectFMF = SI.getFastMathFlags();
FastMathFlags FAddFMF = FAdd->getFastMathFlags();
FastMathFlags NewFMF = FastMathFlags::intersectRewrite(SelectFMF, FAddFMF) |
FastMathFlags::unionValue(SelectFMF, FAddFMF);
cast<Instruction>(NewFAdd)->setFastMathFlags(NewFMF);
cast<Instruction>(NewSelect)->setFastMathFlags(NewFMF);
return NewFAdd;
};
// select((fcmp Pred, X, 0), (fadd X, C), C)
// => fadd((select (fcmp Pred, X, 0), X, 0), C)
//
// Pred := OGT, OGE, OLT, OLE, UGT, UGE, ULT, and ULE
Instruction *FAdd;
Constant *C;
Value *X, *Z;
CmpPredicate Pred;
// Note: OneUse check for `Cmp` is necessary because it makes sure that other
// InstCombine folds don't undo this transformation and cause an infinite
// loop. Furthermore, it could also increase the operation count.
if (match(&SI, m_Select(m_OneUse(m_FCmp(Pred, m_Value(X), m_Value(Z))),
m_OneUse(m_Instruction(FAdd)), m_Constant(C))))
return TryFoldIntoAddConstant(Pred, X, Z, FAdd, C, /*Swapped=*/false);
if (match(&SI, m_Select(m_OneUse(m_FCmp(Pred, m_Value(X), m_Value(Z))),
m_Constant(C), m_OneUse(m_Instruction(FAdd)))))
return TryFoldIntoAddConstant(Pred, X, Z, FAdd, C, /*Swapped=*/true);
return nullptr;
}
static Value *foldSelectBitTest(SelectInst &Sel, Value *CondVal, Value *TrueVal,
Value *FalseVal,
InstCombiner::BuilderTy &Builder,
const SimplifyQuery &SQ) {
// If this is a vector select, we need a vector compare.
Type *SelType = Sel.getType();
if (SelType->isVectorTy() != CondVal->getType()->isVectorTy())
return nullptr;
Value *V;
APInt AndMask;
bool CreateAnd = false;
CmpPredicate Pred;
Value *CmpLHS, *CmpRHS;
if (match(CondVal, m_ICmp(Pred, m_Value(CmpLHS), m_Value(CmpRHS)))) {
if (ICmpInst::isEquality(Pred)) {
if (!match(CmpRHS, m_Zero()))
return nullptr;
V = CmpLHS;
const APInt *AndRHS;
if (!match(CmpLHS, m_And(m_Value(), m_Power2(AndRHS))))
return nullptr;
AndMask = *AndRHS;
} else if (auto Res = decomposeBitTestICmp(CmpLHS, CmpRHS, Pred)) {
assert(ICmpInst::isEquality(Res->Pred) && "Not equality test?");
AndMask = Res->Mask;
V = Res->X;
KnownBits Known = computeKnownBits(V, SQ.getWithInstruction(&Sel));
AndMask &= Known.getMaxValue();
if (!AndMask.isPowerOf2())
return nullptr;
Pred = Res->Pred;
CreateAnd = true;
} else {
return nullptr;
}
} else if (auto *Trunc = dyn_cast<TruncInst>(CondVal)) {
V = Trunc->getOperand(0);
AndMask = APInt(V->getType()->getScalarSizeInBits(), 1);
Pred = ICmpInst::ICMP_NE;
CreateAnd = !Trunc->hasNoUnsignedWrap();
} else {
return nullptr;
}
if (Pred == ICmpInst::ICMP_NE)
std::swap(TrueVal, FalseVal);
if (Value *X = foldSelectICmpAnd(Sel, CondVal, TrueVal, FalseVal, V, AndMask,
CreateAnd, Builder))
return X;
if (Value *X = foldSelectICmpAndBinOp(CondVal, TrueVal, FalseVal, V, AndMask,
CreateAnd, Builder))
return X;
return nullptr;
}
/// This function makes the following folds:
/// select C, (sub 0, X), (xor X, -1)
/// -> sub (sext !C), X
/// select C, (xor X, -1), (sub 0, X)
/// -> sub (sext C), X
static Instruction *foldSelectNegNot(SelectInst &SI,
InstCombiner::BuilderTy &Builder) {
auto *CondVal = SI.getCondition();
auto *TrueVal = SI.getTrueValue();
auto *FalseVal = SI.getFalseValue();
auto *SelTy = SI.getType();
if (!SelTy->isIntOrIntVectorTy() || SelTy->isIntOrIntVectorTy(1))
return nullptr;
if (CondVal->getType()->isVectorTy() != SelTy->isVectorTy())
return nullptr;
auto matchNegNot = [&](Value *Neg, Value *Not, Value *&X) -> bool {
return match(Neg, m_OneUse(m_Neg(m_Value(X)))) &&
match(Not, m_OneUse(m_Not(m_Specific(X))));
};
Value *X;
Value *Mask;
// select C, (sub 0, X), (xor X, -1) -> sub (sext !C), X
if (matchNegNot(TrueVal, FalseVal, X)) {
Value *NotCond = Builder.CreateNot(CondVal, "not." + CondVal->getName());
Mask = Builder.CreateSExt(NotCond, SelTy);
return BinaryOperator::CreateSub(Mask, X);
}
// select C, (xor X, -1), (sub 0, X) -> sub (sext C), X
if (matchNegNot(FalseVal, TrueVal, X)) {
Mask = Builder.CreateSExt(CondVal, SelTy);
return BinaryOperator::CreateSub(Mask, X);
}
return nullptr;
}
/// Fold select (A & Shift == 0 | B & Shift == 0), 0, Shift -> Shift & A & B
/// where Shift is known to be a power of two.
static Instruction *foldSelectAndOrPowerOfTwo(SelectInst &SI,
InstCombiner::BuilderTy &Builder,
const SimplifyQuery &SQ) {
Value *Cond = SI.getCondition();
if (!Cond->hasOneUse())
return nullptr;
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
Value *A, *B, *Shift;
bool Case1 =
match(TrueVal, m_Zero()) && match(FalseVal, m_Value(Shift)) &&
match(Cond, m_Or(m_SpecificICmp(ICmpInst::ICMP_EQ,
m_c_And(m_Specific(Shift), m_Value(A)),
m_Zero()),
m_SpecificICmp(ICmpInst::ICMP_EQ,
m_c_And(m_Specific(Shift), m_Value(B)),
m_Zero())));
bool Case2 =
match(FalseVal, m_Zero()) && match(TrueVal, m_Value(Shift)) &&
match(Cond, m_And(m_SpecificICmp(ICmpInst::ICMP_NE,
m_c_And(m_Specific(Shift), m_Value(A)),
m_Zero()),
m_SpecificICmp(ICmpInst::ICMP_NE,
m_c_And(m_Specific(Shift), m_Value(B)),
m_Zero())));
if ((Case1 || Case2) && isKnownToBeAPowerOfTwo(Shift, /*OrZero=*/true,
SQ.getWithInstruction(&SI))) {
Value *And1 = Builder.CreateAnd(Shift, A);
return BinaryOperator::CreateAnd(And1, B);
}
return nullptr;
}
// Return true if no use can observe the sign of zero of the select result,
// looking through phis, selects and the loop back edge to the select itself.
static bool isSelectZeroSignInsignificant(SelectInst &SI) {
// Bound the number of uses to look through to keep the compile time in
// check.
constexpr unsigned MaxUsesToLookThrough = 16;
unsigned NumUses = 0;
SmallPtrSet<Instruction *, 4> Visited;
SmallVector<Instruction *> Worklist(1, &SI);
while (!Worklist.empty()) {
for (Use &U : Worklist.pop_back_val()->uses()) {
if (++NumUses > MaxUsesToLookThrough)
return false;
auto *User = cast<Instruction>(U.getUser());
if (User == &SI)
continue;
if (canIgnoreSignBitOfZero(U))
continue;
if (isa<PHINode, SelectInst>(User)) {
if (Visited.insert(User).second)
Worklist.push_back(User);
continue;
}
return false;
}
}
return true;
}
Instruction *InstCombinerImpl::visitSelectInst(SelectInst &SI) {
Value *CondVal = SI.getCondition();
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
Type *SelType = SI.getType();
FastMathFlags FMF;
if (auto *FPMO = dyn_cast_if_present<FPMathOperator>(&SI))
FMF = FPMO->getFastMathFlags();
if (Value *V = simplifySelectInst(CondVal, TrueVal, FalseVal, FMF,
SQ.getWithInstruction(&SI)))
return replaceInstUsesWith(SI, V);
if (Instruction *I = canonicalizeSelectToShuffle(SI))
return I;
if (Instruction *I = canonicalizeScalarSelectOfVecs(SI, *this))
return I;
// Fold: select (icmp ult X, 2), X, ctpop(X) --> ctpop(X)
// ctpop(0)==0 and ctpop(1)==1, so the guard is always redundant.
if (match(FalseVal, m_Ctpop(m_Specific(TrueVal))) &&
match(CondVal, m_SpecificICmp(ICmpInst::ICMP_ULT, m_Specific(TrueVal),
m_SpecificInt(2)))) {
cast<Instruction>(FalseVal)->dropPoisonGeneratingAnnotations();
addToWorklist(cast<Instruction>(FalseVal));
return replaceInstUsesWith(SI, FalseVal);
}
// If the type of select is not an integer type or if the condition and
// the selection type are not both scalar nor both vector types, there is no
// point in attempting to match these patterns.
Type *CondType = CondVal->getType();
if (!isa<Constant>(CondVal) && SelType->isIntOrIntVectorTy() &&
CondType->isVectorTy() == SelType->isVectorTy()) {
if (Value *S = simplifyWithOpReplaced(TrueVal, CondVal,
ConstantInt::getTrue(CondType), SQ,
/* AllowRefinement */ true))
return replaceOperand(SI, 1, S);
if (Value *S = simplifyWithOpReplaced(FalseVal, CondVal,
ConstantInt::getFalse(CondType), SQ,
/* AllowRefinement */ true))
return replaceOperand(SI, 2, S);
if (replaceInInstruction(TrueVal, CondVal,
ConstantInt::getTrue(CondType)) ||
replaceInInstruction(FalseVal, CondVal,
ConstantInt::getFalse(CondType)))
return &SI;
}
if (Instruction *R = foldSelectOfBools(SI))
return R;
// Selecting between two integer or vector splat integer constants?
//
// Note that we don't handle a scalar select of vectors:
// select i1 %c, <2 x i8> <1, 1>, <2 x i8> <0, 0>
// because that may need 3 instructions to splat the condition value:
// extend, insertelement, shufflevector.
//
// Do not handle i1 TrueVal and FalseVal otherwise would result in
// zext/sext i1 to i1.
if (SelType->isIntOrIntVectorTy() && !SelType->isIntOrIntVectorTy(1) &&
CondVal->getType()->isVectorTy() == SelType->isVectorTy()) {
// select C, 1, 0 -> zext C to int
if (match(TrueVal, m_One()) && match(FalseVal, m_Zero()))
return new ZExtInst(CondVal, SelType);
// select C, -1, 0 -> sext C to int
if (match(TrueVal, m_AllOnes()) && match(FalseVal, m_Zero()))
return new SExtInst(CondVal, SelType);
// select C, 0, 1 -> zext !C to int
if (match(TrueVal, m_Zero()) && match(FalseVal, m_One())) {
Value *NotCond = Builder.CreateNot(CondVal, "not." + CondVal->getName());
return new ZExtInst(NotCond, SelType);
}
// select C, 0, -1 -> sext !C to int
if (match(TrueVal, m_Zero()) && match(FalseVal, m_AllOnes())) {
Value *NotCond = Builder.CreateNot(CondVal, "not." + CondVal->getName());
return new SExtInst(NotCond, SelType);
}
}
if (Instruction *I = foldSelectNegNot(SI, Builder))
return I;
if (Instruction *I = foldSelectAndOrPowerOfTwo(SI, Builder, SQ))
return I;
auto *SIFPOp = dyn_cast<FPMathOperator>(&SI);
if (auto *FCmp = dyn_cast<FCmpInst>(CondVal)) {
FCmpInst::Predicate Pred = FCmp->getPredicate();
Value *Cmp0 = FCmp->getOperand(0), *Cmp1 = FCmp->getOperand(1);
// Are we selecting a value based on a comparison of the two values?
if ((Cmp0 == TrueVal && Cmp1 == FalseVal) ||
(Cmp0 == FalseVal && Cmp1 == TrueVal)) {
// Canonicalize to use ordered comparisons by swapping the select
// operands.
//
// e.g.
// (X ugt Y) ? X : Y -> (X ole Y) ? Y : X
if (FCmp->hasOneUse() && FCmpInst::isUnordered(Pred)) {
FCmpInst::Predicate InvPred = FCmp->getInversePredicate();
Value *NewCond = Builder.CreateFCmpFMF(InvPred, Cmp0, Cmp1, FCmp,
FCmp->getName() + ".inv");
// Propagate ninf/nnan from fcmp to select.
FastMathFlags FMF = SI.getFastMathFlags();
if (FCmp->hasNoNaNs())
FMF.setNoNaNs(true);
if (FCmp->hasNoInfs())
FMF.setNoInfs(true);
Value *NewSel =
Builder.CreateSelectFMF(NewCond, FalseVal, TrueVal, FMF);
return replaceInstUsesWith(SI, NewSel);
}
}
if (SIFPOp) {
// Fold out scale-if-equals-zero pattern.
//
// This pattern appears in code with denormal range checks after it's
// assumed denormals are treated as zero. This drops a canonicalization.
// TODO: Could relax the signed zero logic. We just need to know the sign
// of the result matches (fmul x, y has the same sign as x).
//
// TODO: Handle always-canonicalizing variant that selects some value or 1
// scaling factor in the fmul visitor.
// TODO: Handle ldexp too
Value *MatchCmp0 = nullptr;
Value *MatchCmp1 = nullptr;
// (select (fcmp [ou]eq x, 0.0), (fmul x, K), x => x
// (select (fcmp [ou]ne x, 0.0), x, (fmul x, K) => x
if (Pred == CmpInst::FCMP_OEQ || Pred == CmpInst::FCMP_UEQ) {
MatchCmp0 = FalseVal;
MatchCmp1 = TrueVal;
} else if (Pred == CmpInst::FCMP_ONE || Pred == CmpInst::FCMP_UNE) {
MatchCmp0 = TrueVal;
MatchCmp1 = FalseVal;
}
if (Cmp0 == MatchCmp0 &&
matchFMulByZeroIfResultEqZero(*this, Cmp0, Cmp1, MatchCmp1, MatchCmp0,
SI, SIFPOp->hasNoSignedZeros()))
return replaceInstUsesWith(SI, Cmp0);
Type *EltTy = SelType->getScalarType();
// TODO: Generalize to any ordered / unordered compare.
if ((Pred == CmpInst::FCMP_ORD || Pred == CmpInst::FCMP_UNO) &&
match(Cmp1, m_PosZeroFP()) && EltTy->isIEEELikeFPTy()) {
// Fold out only-canonicalize-non-nans pattern. This implements a
// wrapper around llvm.canonicalize which is not required to quiet
// signaling nans or preserve nan payload bits.
//
// %hard.canonical = call @llvm.canonicalize(%x)
// %soft.canonical = fdiv 1.0, %x
// %ord = fcmp ord %x, 0.0
// %x.canon = select i1 %ord, %hard.canonical, %soft.canonical
//
// With known IEEE handling:
// => %x
//
// With other denormal behaviors:
// => llvm.canonicalize(%x)
//
// Note the fdiv could be any value preserving, potentially
// canonicalizing floating-point operation such as fmul by 1.0. However,
// since in the llvm model canonicalization is not mandatory, the fmul
// would have been dropped by the time we reached here. The trick here
// is to use a reciprocal fdiv. It's not a droppable no-op, as it could
// return an infinity if %x were sufficiently small, but in this pattern
// we're only using the output for nan values.
if (Pred == CmpInst::FCMP_ORD) {
MatchCmp0 = TrueVal;
MatchCmp1 = FalseVal;
} else {
MatchCmp0 = FalseVal;
MatchCmp1 = TrueVal;
}
bool RcpIfNan = match(MatchCmp1, m_FDiv(m_FPOne(), m_Specific(Cmp0)));
bool CanonicalizeIfNotNan =
match(MatchCmp0, m_FCanonicalize(m_Specific(Cmp0)));
if (RcpIfNan || CanonicalizeIfNotNan) {
const fltSemantics &FPSem = EltTy->getFltSemantics();
DenormalMode Mode = F.getDenormalMode(FPSem);
if (RcpIfNan) {
if (Mode == DenormalMode::getIEEE()) {
// Special case for the other select operand. Otherwise, we may
// need to insert freeze on Cmp0 in the compare and select.
if (CanonicalizeIfNotNan)
return replaceInstUsesWith(SI, Cmp0);
if (isGuaranteedNotToBeUndef(Cmp0, &AC, &SI, &DT)) {
// select (fcmp ord x, 0), y, (fdiv 1, x)
// => select (fcmp ord x, 0), y, x
//
// select (fcmp uno x, 0), (fdiv 1, x), y
// => select (fcmp uno x, 0), x, y
replaceOperand(SI, Pred == CmpInst::FCMP_ORD ? 2 : 1, Cmp0);
return &SI;
}
auto *FrCmp0 = InsertNewInstBefore(
new FreezeInst(Cmp0, Cmp0->getName() + ".fr"),
FCmp->getIterator());
replaceOperand(*FCmp, 0, FrCmp0);
return replaceOperand(SI, Pred == CmpInst::FCMP_ORD ? 2 : 1,
FrCmp0);
}
}
if (CanonicalizeIfNotNan) {
// IEEE handling does not have non-canonical values, so the
// canonicalize can be dropped for direct replacement without
// looking for the intermediate maybe-canonicalizing operation.
if (Mode == DenormalMode::getIEEE()) {
// select (fcmp ord x, 0), canonicalize(x), y
// => select (fcmp ord x, 0), x, y
replaceOperand(SI, Pred == CmpInst::FCMP_ORD ? 1 : 2, Cmp0);
return &SI;
}
// If denormals may be flushed, we need to retain the canonicalize
// call. This introduces a canonicalization on the nan path, which
// we are not free to do as that could change the sign bit or
// payload bits. We can only do this if there were a no-op like
// floating-point instruction which may have changed the nan bits
// anyway.
// Leave the dynamic mode case alone. This would introduce new
// constraints if the mode may be refined later.
if (RcpIfNan && (Mode.inputsAreZero() || Mode.outputsAreZero()))
return replaceInstUsesWith(SI, MatchCmp0);
assert(RcpIfNan || Mode != DenormalMode::getIEEE());
}
}
}
}
}
if (SIFPOp) {
// TODO: Try to forward-propagate FMF from select arms to the select.
auto *FCmp = dyn_cast<FCmpInst>(CondVal);
// Canonicalize select of FP values where NaN and -0.0 are not valid as
// minnum/maxnum intrinsics.
//
// Note that the `nnan` flag is propagated from the comparison, not from the
// select. While it's technically possible to transform a `fcmp` + `select
// nnan` to a `minnum`/`maxnum` call *without* an `nnan`, that would be a
// pessimization in practice. Many targets can't map `minnum`/`maxnum` to a
// single instruction, and if they cannot prove the absence of NaN, must
// lower it to a routine or a libcall. There are additional reasons besides
// performance to avoid introducing libcalls where none existed before
// (https://github.com/llvm/llvm-project/issues/54554).
//
// As such, we want to ensure that the generated `minnum`/`maxnum` intrinsic
// has the `nnan nsz` flags, which allow it to be lowered *back* to a
// fcmp+select if that's the best way to express it on the target.
if (FCmp && FCmp->hasNoNaNs() &&
(SIFPOp->hasNoSignedZeros() || isSelectZeroSignInsignificant(SI))) {
Value *X, *Y;
if (match(&SI, m_OrdOrUnordFMax(m_Value(X), m_Value(Y)))) {
Value *BinIntr =
Builder.CreateBinaryIntrinsic(Intrinsic::maxnum, X, Y, &SI);
if (auto *BinIntrInst = dyn_cast<Instruction>(BinIntr)) {
// `ninf` must be propagated from the comparison too, rather than the
// select: https://github.com/llvm/llvm-project/pull/136433
BinIntrInst->setHasNoInfs(FCmp->hasNoInfs());
// The `nsz` flag is a precondition, so let's ensure it's always added
// to the min/max operation, even if it wasn't on the select. This
// could happen if the select doesn't have `nsz`, but no use of the
// result can observe the sign of zero.
BinIntrInst->setHasNoSignedZeros(true);
// As mentioned above, `nnan` is also a precondition, so we always set
// the flag.
BinIntrInst->setHasNoNaNs(true);
}
return replaceInstUsesWith(SI, BinIntr);
}
if (match(&SI, m_OrdOrUnordFMin(m_Value(X), m_Value(Y)))) {
Value *BinIntr =
Builder.CreateBinaryIntrinsic(Intrinsic::minnum, X, Y, &SI);
if (auto *BinIntrInst = dyn_cast<Instruction>(BinIntr)) {
BinIntrInst->setHasNoInfs(FCmp->hasNoInfs());
BinIntrInst->setHasNoSignedZeros(true);
BinIntrInst->setHasNoNaNs(true);
}
return replaceInstUsesWith(SI, BinIntr);
}
}
}
// Fold selecting to fabs.
if (Instruction *Fabs = foldSelectWithFCmpToFabs(SI, *this))
return Fabs;
if (Instruction *I = foldSelectOfOrderedFAbsCmpOfNaNScrubbedValue(SI, *this))
return I;
// See if we are selecting two values based on a comparison of the two values.
if (CmpInst *CI = dyn_cast<CmpInst>(CondVal))
if (Instruction *NewSel = foldSelectValueEquivalence(SI, *CI))
return NewSel;
if (ICmpInst *ICI = dyn_cast<ICmpInst>(CondVal))
if (Instruction *Result = foldSelectInstWithICmp(SI, ICI))
return Result;
if (Instruction *V =
foldSelectICmpAndAnd(SelType, CondVal, TrueVal, FalseVal, Builder))
return V;
if (Value *V = foldSelectBitTest(SI, CondVal, TrueVal, FalseVal, Builder, SQ))
return replaceInstUsesWith(SI, V);
if (Instruction *Add = foldAddSubSelect(SI, Builder))
return Add;
if (Instruction *Add = foldOverflowingAddSubSelect(SI, Builder))
return Add;
if (Instruction *Or = foldSetClearBits(SI, Builder))
return Or;
if (Instruction *Mul = foldSelectZeroOrFixedOp(SI, *this))
return Mul;
// Turn (select C, (op X, Y), (op X, Z)) -> (op X, (select C, Y, Z))
auto *TI = dyn_cast<Instruction>(TrueVal);
auto *FI = dyn_cast<Instruction>(FalseVal);
if (TI && FI && TI->getOpcode() == FI->getOpcode())
if (Instruction *IV = foldSelectOpOp(SI, TI, FI))
return IV;
if (Instruction *I = foldSelectIntrinsic(SI))
return I;
if (Instruction *I = foldSelectExtConst(SI))
return I;
if (Instruction *I = foldSelectWithSRem(SI, *this, Builder))
return I;
// Fold (select C, (gep Ptr, Idx), Ptr) -> (gep Ptr, (select C, Idx, 0))
// Fold (select C, Ptr, (gep Ptr, Idx)) -> (gep Ptr, (select C, 0, Idx))
auto SelectGepWithBase = [&](GetElementPtrInst *Gep, Value *Base,
bool Swap) -> GetElementPtrInst * {
Value *Ptr = Gep->getPointerOperand();
if (Gep->getNumOperands() != 2 || Gep->getPointerOperand() != Base ||
!Gep->hasOneUse())
return nullptr;
Value *Idx = Gep->getOperand(1);
if (isa<VectorType>(CondVal->getType()) && !isa<VectorType>(Idx->getType()))
return nullptr;
Type *ElementType = Gep->getSourceElementType();
Value *NewT = Idx;
Value *NewF = Constant::getNullValue(Idx->getType());
if (Swap)
std::swap(NewT, NewF);
Value *NewSI =
Builder.CreateSelect(CondVal, NewT, NewF, SI.getName() + ".idx", &SI);
return GetElementPtrInst::Create(ElementType, Ptr, NewSI,
Gep->getNoWrapFlags());
};
if (auto *TrueGep = dyn_cast<GetElementPtrInst>(TrueVal))
if (auto *NewGep = SelectGepWithBase(TrueGep, FalseVal, false))
return NewGep;
if (auto *FalseGep = dyn_cast<GetElementPtrInst>(FalseVal))
if (auto *NewGep = SelectGepWithBase(FalseGep, TrueVal, true))
return NewGep;
// See if we can fold the select into one of our operands.
if (SelType->isIntOrIntVectorTy() || SelType->isFPOrFPVectorTy()) {
if (Instruction *FoldI = foldSelectIntoOp(SI, TrueVal, FalseVal))
return FoldI;
Value *LHS, *RHS;
Instruction::CastOps CastOp;
SelectPatternResult SPR = matchSelectPattern(&SI, LHS, RHS, &CastOp);
auto SPF = SPR.Flavor;
if (SPF) {
Value *LHS2, *RHS2;
if (SelectPatternFlavor SPF2 = matchSelectPattern(LHS, LHS2, RHS2).Flavor)
if (Instruction *R = foldSPFofSPF(cast<Instruction>(LHS), SPF2, LHS2,
RHS2, SI, SPF, RHS))
return R;
if (SelectPatternFlavor SPF2 = matchSelectPattern(RHS, LHS2, RHS2).Flavor)
if (Instruction *R = foldSPFofSPF(cast<Instruction>(RHS), SPF2, LHS2,
RHS2, SI, SPF, LHS))
return R;
}
if (SelectPatternResult::isMinOrMax(SPF)) {
// Canonicalize so that
// - type casts are outside select patterns.
// - float clamp is transformed to min/max pattern
bool IsCastNeeded = LHS->getType() != SelType;
Value *CmpLHS = cast<CmpInst>(CondVal)->getOperand(0);
Value *CmpRHS = cast<CmpInst>(CondVal)->getOperand(1);
if (IsCastNeeded ||
(LHS->getType()->isFPOrFPVectorTy() &&
((CmpLHS != LHS && CmpLHS != RHS) ||
(CmpRHS != LHS && CmpRHS != RHS)))) {
CmpInst::Predicate MinMaxPred = getMinMaxPred(SPF, SPR.Ordered);
Value *Cmp;
if (CmpInst::isIntPredicate(MinMaxPred))
Cmp = Builder.CreateICmp(MinMaxPred, LHS, RHS);
else
Cmp = Builder.CreateFCmpFMF(MinMaxPred, LHS, RHS,
cast<Instruction>(SI.getCondition()));
Value *NewSI = Builder.CreateSelect(Cmp, LHS, RHS, SI.getName(), &SI);
if (!IsCastNeeded)
return replaceInstUsesWith(SI, NewSI);
Value *NewCast = Builder.CreateCast(CastOp, NewSI, SelType);
return replaceInstUsesWith(SI, NewCast);
}
}
}
// See if we can fold the select into a phi node if the condition is a select.
if (auto *PN = dyn_cast<PHINode>(SI.getCondition()))
if (Instruction *NV = foldOpIntoPhi(SI, PN))
return NV;
if (SelectInst *TrueSI = dyn_cast<SelectInst>(TrueVal)) {
if (TrueSI->getCondition()->getType() == CondVal->getType()) {
// Fold nested selects if the inner condition can be implied by the outer
// condition.
if (Value *V = simplifyNestedSelectsUsingImpliedCond(
*TrueSI, CondVal, /*CondIsTrue=*/true, DL))
return replaceOperand(SI, 1, V);
// We choose this as normal form to enable folding on the And and
// shortening paths for the values (this helps getUnderlyingObjects() for
// example).
if (TrueSI->hasOneUse()) {
Value *And = nullptr, *OtherVal = nullptr;
// select(C0, select(C1, a, b), b) -> select(C0&&C1, a, b)
if (TrueSI->getFalseValue() == FalseVal) {
And = Builder.CreateLogicalAnd(CondVal, TrueSI->getCondition(), "",
&SI);
OtherVal = TrueSI->getTrueValue();
}
// select(C0, select(C1, b, a), b) -> select(C0&&!C1, a, b)
else if (TrueSI->getTrueValue() == FalseVal) {
Value *InvertedCond = Builder.CreateNot(TrueSI->getCondition());
And = Builder.CreateLogicalAnd(CondVal, InvertedCond, "", &SI);
OtherVal = TrueSI->getFalseValue();
}
if (And && OtherVal) {
replaceOperand(SI, 0, And);
replaceOperand(SI, 1, OtherVal);
setExplicitlyUnknownBranchWeightsIfProfiled(SI, DEBUG_TYPE);
return &SI;
}
}
}
}
if (SelectInst *FalseSI = dyn_cast<SelectInst>(FalseVal)) {
if (FalseSI->getCondition()->getType() == CondVal->getType()) {
// Fold nested selects if the inner condition can be implied by the outer
// condition.
if (Value *V = simplifyNestedSelectsUsingImpliedCond(
*FalseSI, CondVal, /*CondIsTrue=*/false, DL))
return replaceOperand(SI, 2, V);
if (FalseSI->hasOneUse()) {
Value *Or = nullptr, *OtherVal = nullptr;
// select(C0, a, select(C1, a, b)) -> select(C0||C1, a, b)
if (FalseSI->getTrueValue() == TrueVal) {
Or = Builder.CreateLogicalOr(CondVal, FalseSI->getCondition(), "",
&SI);
OtherVal = FalseSI->getFalseValue();
}
// select(C0, a, select(C1, b, a)) -> select(C0||!C1, a, b)
else if (FalseSI->getFalseValue() == TrueVal) {
Value *InvertedCond = Builder.CreateNot(FalseSI->getCondition());
Or = Builder.CreateLogicalOr(CondVal, InvertedCond, "", &SI);
OtherVal = FalseSI->getTrueValue();
}
if (Or && OtherVal) {
replaceOperand(SI, 0, Or);
replaceOperand(SI, 2, OtherVal);
setExplicitlyUnknownBranchWeightsIfProfiled(SI, DEBUG_TYPE);
return &SI;
}
}
}
}
// Try to simplify a binop sandwiched between 2 selects with the same
// condition. This is not valid for div/rem because the select might be
// preventing a division-by-zero.
// TODO: A div/rem restriction is conservative; use something like
// isSafeToSpeculativelyExecute().
// select(C, binop(select(C, X, Y), W), Z) -> select(C, binop(X, W), Z)
BinaryOperator *TrueBO;
if (match(TrueVal, m_OneUse(m_BinOp(TrueBO))) && !TrueBO->isIntDivRem()) {
if (auto *TrueBOSI = dyn_cast<SelectInst>(TrueBO->getOperand(0))) {
if (TrueBOSI->getCondition() == CondVal) {
replaceOperand(*TrueBO, 0, TrueBOSI->getTrueValue());
Worklist.push(TrueBO);
return &SI;
}
}
if (auto *TrueBOSI = dyn_cast<SelectInst>(TrueBO->getOperand(1))) {
if (TrueBOSI->getCondition() == CondVal) {
replaceOperand(*TrueBO, 1, TrueBOSI->getTrueValue());
Worklist.push(TrueBO);
return &SI;
}
}
}
// select(C, Z, binop(select(C, X, Y), W)) -> select(C, Z, binop(Y, W))
BinaryOperator *FalseBO;
if (match(FalseVal, m_OneUse(m_BinOp(FalseBO))) && !FalseBO->isIntDivRem()) {
if (auto *FalseBOSI = dyn_cast<SelectInst>(FalseBO->getOperand(0))) {
if (FalseBOSI->getCondition() == CondVal) {
replaceOperand(*FalseBO, 0, FalseBOSI->getFalseValue());
Worklist.push(FalseBO);
return &SI;
}
}
if (auto *FalseBOSI = dyn_cast<SelectInst>(FalseBO->getOperand(1))) {
if (FalseBOSI->getCondition() == CondVal) {
replaceOperand(*FalseBO, 1, FalseBOSI->getFalseValue());
Worklist.push(FalseBO);
return &SI;
}
}
}
Value *NotCond;
if (match(CondVal, m_Not(m_Value(NotCond))) &&
!InstCombiner::shouldAvoidAbsorbingNotIntoSelect(SI)) {
replaceOperand(SI, 0, NotCond);
SI.swapValues();
SI.swapProfMetadata();
return &SI;
}
if (Instruction *I = foldVectorSelect(SI))
return I;
// If we can compute the condition, there's no need for a select.
// Like the above fold, we are attempting to reduce compile-time cost by
// putting this fold here with limitations rather than in InstSimplify.
// The motivation for this call into value tracking is to take advantage of
// the assumption cache, so make sure that is populated.
if (!CondVal->getType()->isVectorTy() && !AC.assumptions().empty()) {
KnownBits Known(1);
computeKnownBits(CondVal, Known, &SI);
if (Known.One.isOne())
return replaceInstUsesWith(SI, TrueVal);
if (Known.Zero.isOne())
return replaceInstUsesWith(SI, FalseVal);
}
if (Instruction *BitCastSel = foldSelectCmpBitcasts(SI, Builder))
return BitCastSel;
// Simplify selects that test the returned flag of cmpxchg instructions.
if (Value *V = foldSelectCmpXchg(SI))
return replaceInstUsesWith(SI, V);
if (Instruction *Select = foldSelectBinOpIdentity(SI, TLI, *this))
return Select;
if (Instruction *Funnel = foldSelectFunnelShift(SI, Builder))
return Funnel;
if (Instruction *Copysign = foldSelectToCopysign(SI, Builder))
return Copysign;
if (Instruction *PN = foldSelectToPhi(SI, DT, Builder))
return replaceInstUsesWith(SI, PN);
if (Value *V = foldRoundUpIntegerWithPow2Alignment(SI, Builder))
return replaceInstUsesWith(SI, V);
if (Value *V = foldSelectIntoAddConstant(SI, Builder))
return replaceInstUsesWith(SI, V);
// select(mask, mload(ptr,mask,0), 0) -> mload(ptr,mask,0)
// Load inst is intentionally not checked for hasOneUse()
if (match(FalseVal, m_Zero()) &&
(match(TrueVal, m_MaskedLoad(m_Value(), m_Specific(CondVal),
m_CombineOr(m_Undef(), m_Zero()))) ||
match(TrueVal, m_MaskedGather(m_Value(), m_Specific(CondVal),
m_CombineOr(m_Undef(), m_Zero()))))) {
auto *MaskedInst = cast<IntrinsicInst>(TrueVal);
if (isa<UndefValue>(MaskedInst->getArgOperand(2)))
MaskedInst->setArgOperand(2, FalseVal /* Zero */);
return replaceInstUsesWith(SI, MaskedInst);
}
Value *Mask;
if (match(TrueVal, m_Zero()) &&
(match(FalseVal, m_MaskedLoad(m_Value(), m_Value(Mask),
m_CombineOr(m_Undef(), m_Zero()))) ||
match(FalseVal, m_MaskedGather(m_Value(), m_Value(Mask),
m_CombineOr(m_Undef(), m_Zero())))) &&
(CondVal->getType() == Mask->getType())) {
// We can remove the select by ensuring the load zeros all lanes the
// select would have. We determine this by proving there is no overlap
// between the load and select masks.
// (i.e (load_mask & select_mask) == 0 == no overlap)
bool CanMergeSelectIntoLoad = false;
if (Value *V = simplifyAndInst(CondVal, Mask, SQ.getWithInstruction(&SI)))
CanMergeSelectIntoLoad = match(V, m_Zero());
if (CanMergeSelectIntoLoad) {
auto *MaskedInst = cast<IntrinsicInst>(FalseVal);
if (isa<UndefValue>(MaskedInst->getArgOperand(2)))
MaskedInst->setArgOperand(2, TrueVal /* Zero */);
return replaceInstUsesWith(SI, MaskedInst);
}
}
if (Instruction *I = foldSelectOfSymmetricSelect(SI, Builder))
return I;
if (Instruction *I = foldNestedSelects(SI, Builder))
return I;
// Match logical variants of the pattern,
// and transform them iff that gets rid of inversions.
// (~x) | y --> ~(x & (~y))
// (~x) & y --> ~(x | (~y))
if (sinkNotIntoOtherHandOfLogicalOp(SI))
return &SI;
if (Instruction *I = foldBitCeil(SI, Builder, *this))
return I;
if (Instruction *I = foldSelectToCmp(SI))
return I;
if (Instruction *I = foldSelectEqualityTest(SI))
return I;
// Fold:
// (select A && B, T, F) -> (select A, (select B, T, F), F)
// (select A || B, T, F) -> (select A, T, (select B, T, F))
// if (select B, T, F) is foldable.
// TODO: preserve FMF flags
auto FoldSelectWithAndOrCond = [&](bool IsAnd, Value *A,
Value *B) -> Instruction * {
if (Value *V = simplifySelectInst(B, TrueVal, FalseVal, FMF,
SQ.getWithInstruction(&SI))) {
Value *NewTrueVal = IsAnd ? V : TrueVal;
Value *NewFalseVal = IsAnd ? FalseVal : V;
// If the True and False values don't change, then preserve the branch
// metadata of the original select as the net effect of this change is to
// simplify the conditional.
Instruction *MDFrom = nullptr;
if (NewTrueVal == TrueVal && NewFalseVal == FalseVal) {
MDFrom = &SI;
}
return SelectInst::Create(A, NewTrueVal, NewFalseVal, "", nullptr,
MDFrom);
}
// Is (select B, T, F) a SPF?
if (CondVal->hasOneUse() && SelType->isIntOrIntVectorTy()) {
if (ICmpInst *Cmp = dyn_cast<ICmpInst>(B))
if (Value *V = canonicalizeSPF(*Cmp, TrueVal, FalseVal, *this)) {
return SelectInst::Create(
A, IsAnd ? V : TrueVal, IsAnd ? FalseVal : V, "", nullptr,
ProfcheckDisableMetadataFixes ? nullptr : &SI);
}
}
return nullptr;
};
Value *LHS, *RHS;
if (match(CondVal, m_And(m_Value(LHS), m_Value(RHS)))) {
if (Instruction *I = FoldSelectWithAndOrCond(/*IsAnd*/ true, LHS, RHS))
return I;
if (Instruction *I = FoldSelectWithAndOrCond(/*IsAnd*/ true, RHS, LHS))
return I;
} else if (match(CondVal, m_Or(m_Value(LHS), m_Value(RHS)))) {
if (Instruction *I = FoldSelectWithAndOrCond(/*IsAnd*/ false, LHS, RHS))
return I;
if (Instruction *I = FoldSelectWithAndOrCond(/*IsAnd*/ false, RHS, LHS))
return I;
} else {
// We cannot swap the operands of logical and/or.
// TODO: Can we swap the operands by inserting a freeze?
if (match(CondVal, m_LogicalAnd(m_Value(LHS), m_Value(RHS)))) {
if (Instruction *I = FoldSelectWithAndOrCond(/*IsAnd*/ true, LHS, RHS))
return I;
} else if (match(CondVal, m_LogicalOr(m_Value(LHS), m_Value(RHS)))) {
if (Instruction *I = FoldSelectWithAndOrCond(/*IsAnd*/ false, LHS, RHS))
return I;
}
}
// select Cond, !X, X -> xor Cond, X
if (CondVal->getType() == SI.getType() && isKnownInversion(FalseVal, TrueVal))
return BinaryOperator::CreateXor(CondVal, FalseVal);
// For vectors, this transform is only safe if the simplification does not
// look through any lane-crossing operations. For now, limit to scalars only.
if (SelType->isIntegerTy() &&
(!isa<Constant>(TrueVal) || !isa<Constant>(FalseVal))) {
// Try to simplify select arms based on KnownBits implied by the condition.
CondContext CC(CondVal);
findValuesAffectedByCondition(CondVal, /*IsAssume=*/false, [&](Value *V) {
CC.AffectedValues.insert(V);
});
SimplifyQuery Q = SQ.getWithInstruction(&SI).getWithCondContext(CC);
if (!CC.AffectedValues.empty()) {
if (!isa<Constant>(TrueVal) &&
hasAffectedValue(TrueVal, CC.AffectedValues, /*Depth=*/0)) {
KnownBits Known = llvm::computeKnownBits(TrueVal, Q);
if (Known.isConstant())
return replaceOperand(SI, 1,
ConstantInt::get(SelType, Known.getConstant()));
}
CC.Invert = true;
if (!isa<Constant>(FalseVal) &&
hasAffectedValue(FalseVal, CC.AffectedValues, /*Depth=*/0)) {
KnownBits Known = llvm::computeKnownBits(FalseVal, Q);
if (Known.isConstant())
return replaceOperand(SI, 2,
ConstantInt::get(SelType, Known.getConstant()));
}
}
}
// select (trunc nuw X to i1), X, Y --> select (trunc nuw X to i1), 1, Y
// select (trunc nuw X to i1), Y, X --> select (trunc nuw X to i1), Y, 0
// select (trunc nsw X to i1), X, Y --> select (trunc nsw X to i1), -1, Y
// select (trunc nsw X to i1), Y, X --> select (trunc nsw X to i1), Y, 0
Value *Trunc;
if (match(CondVal, m_NUWTrunc(m_Value(Trunc))) && !isa<Constant>(Trunc)) {
if (TrueVal == Trunc)
return replaceOperand(SI, 1, ConstantInt::get(TrueVal->getType(), 1));
if (FalseVal == Trunc)
return replaceOperand(SI, 2, ConstantInt::get(FalseVal->getType(), 0));
}
if (match(CondVal, m_NSWTrunc(m_Value(Trunc))) && !isa<Constant>(Trunc)) {
if (TrueVal == Trunc)
return replaceOperand(SI, 1,
Constant::getAllOnesValue(TrueVal->getType()));
if (FalseVal == Trunc)
return replaceOperand(SI, 2, ConstantInt::get(FalseVal->getType(), 0));
}
if (match(CondVal, m_Trunc(m_Value(Trunc))) && Trunc->getType() == SelType) {
if (match(FalseVal, m_Zero()) && impliesPoison(TrueVal, CondVal) &&
llvm::computeKnownBits(TrueVal, SQ.getWithInstruction(&SI))
.countMaxActiveBits() == 1)
return BinaryOperator::CreateAnd(Trunc, TrueVal);
if (cast<TruncInst>(CondVal)->hasNoUnsignedWrap() &&
match(TrueVal, m_One()) && impliesPoison(FalseVal, CondVal) &&
llvm::computeKnownBits(FalseVal, SQ.getWithInstruction(&SI))
.countMaxActiveBits() == 1) {
return BinaryOperator::CreateOr(Trunc, FalseVal);
}
}
Value *MaskedLoadPtr;
if (match(TrueVal, m_OneUse(m_MaskedLoad(m_Value(MaskedLoadPtr),
m_Specific(CondVal), m_Value())))) {
auto *LoadInst = cast<IntrinsicInst>(TrueVal);
// Keep the load at its original position to avoid crossing writes. The new
// passthrough must therefore be available there.
if (DT.dominates(FalseVal, LoadInst)) {
Builder.SetInsertPoint(LoadInst);
Instruction *In = Builder.CreateMaskedLoad(
TrueVal->getType(), MaskedLoadPtr,
LoadInst->getParamAlign(0).valueOrOne(), CondVal, FalseVal);
In->setAAMetadata(LoadInst->getAAMetadata());
return replaceInstUsesWith(SI, In);
}
}
// Canonicalize sign function ashr pattern: select (icmp slt X, 1), ashr X,
// bitwidth-1, 1 -> scmp(X, 0)
// Also handles: select (icmp sgt X, 0), 1, ashr X, bitwidth-1 -> scmp(X, 0)
unsigned BitWidth = SI.getType()->getScalarSizeInBits();
CmpPredicate Pred;
Value *CmpLHS, *CmpRHS;
// Canonicalize sign function ashr patterns:
// select (icmp slt X, 1), ashr X, bitwidth-1, 1 -> scmp(X, 0)
// select (icmp sgt X, 0), 1, ashr X, bitwidth-1 -> scmp(X, 0)
if (match(&SI, m_Select(m_ICmp(Pred, m_Value(CmpLHS), m_Value(CmpRHS)),
m_Value(TrueVal), m_Value(FalseVal))) &&
((Pred == ICmpInst::ICMP_SLT && match(CmpRHS, m_One()) &&
match(TrueVal,
m_AShr(m_Specific(CmpLHS), m_SpecificInt(BitWidth - 1))) &&
match(FalseVal, m_One())) ||
(Pred == ICmpInst::ICMP_SGT && match(CmpRHS, m_Zero()) &&
match(TrueVal, m_One()) &&
match(FalseVal,
m_AShr(m_Specific(CmpLHS), m_SpecificInt(BitWidth - 1)))))) {
Function *Scmp = Intrinsic::getOrInsertDeclaration(
SI.getModule(), Intrinsic::scmp, {SI.getType(), SI.getType()});
return CallInst::Create(Scmp, {CmpLHS, ConstantInt::get(SI.getType(), 0)});
}
return nullptr;
}