OSchip
/
llvm-project
mirror of https://github.com/THU-DSP-LAB/llvm-project.git
13
0
Fork 6
Code Issues Packages Projects Releases Wiki Activity
llvm-project/llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp

2166 lines
81 KiB
C++
Raw Blame History

//===- 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/Optional.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/ValueTracking.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/Constant.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DerivedTypes.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/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/Transforms/InstCombine/InstCombineWorklist.h"
#include <cassert>
#include <utility>
using namespace llvm;
using namespace PatternMatch;
#define DEBUG_TYPE "instcombine"
static Value *createMinMax(InstCombiner::BuilderTy &Builder,
SelectPatternFlavor SPF, Value *A, Value *B) {
CmpInst::Predicate Pred = getMinMaxPred(SPF);
assert(CmpInst::isIntPredicate(Pred) && "Expected integer predicate");
return Builder.CreateSelect(Builder.CreateICmp(Pred, A, B), A, B);
}
/// Replace a select operand based on an equality comparison with the identity
/// constant of a binop.
static Instruction *foldSelectBinOpIdentity(SelectInst &Sel,
const TargetLibraryInfo &TLI) {
// The select condition must be an equality compare with a constant operand.
Value *X;
Constant *C;
CmpInst::Predicate 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;
// 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();
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;
}
// Last, match the compare variable operand with a binop operand.
Value *Y;
if (!BO->isCommutative() && !match(BO, m_BinOp(m_Value(Y), m_Specific(X))))
return nullptr;
if (!match(BO, m_c_BinOp(m_Value(Y), m_Specific(X))))
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 (isa<FPMathOperator>(BO))
if (!BO->hasNoSignedZeros() && !CannotBeNegativeZero(Y, &TLI))
return nullptr;
// 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 }
Sel.setOperand(IsEq ? 1 : 2, Y);
return &Sel;
}
/// 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, ICmpInst *Cmp,
InstCombiner::BuilderTy &Builder) {
const APInt *SelTC, *SelFC;
if (!match(Sel.getTrueValue(), m_APInt(SelTC)) ||
!match(Sel.getFalseValue(), m_APInt(SelFC)))
return nullptr;
// If this is a vector select, we need a vector compare.
Type *SelType = Sel.getType();
if (SelType->isVectorTy() != Cmp->getType()->isVectorTy())
return nullptr;
Value *V;
APInt AndMask;
bool CreateAnd = false;
ICmpInst::Predicate Pred = Cmp->getPredicate();
if (ICmpInst::isEquality(Pred)) {
if (!match(Cmp->getOperand(1), m_Zero()))
return nullptr;
V = Cmp->getOperand(0);
const APInt *AndRHS;
if (!match(V, m_And(m_Value(), m_Power2(AndRHS))))
return nullptr;
AndMask = *AndRHS;
} else if (decomposeBitTestICmp(Cmp->getOperand(0), Cmp->getOperand(1),
Pred, V, AndMask)) {
assert(ICmpInst::isEquality(Pred) && "Not equality test?");
if (!AndMask.isPowerOf2())
return nullptr;
CreateAnd = true;
} else {
return nullptr;
}
// 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.
APInt TC = *SelTC;
APInt FC = *SelFC;
if (!TC.isNullValue() && !FC.isNullValue()) {
// If the select constants differ by exactly one bit and that's the same
// bit that is masked and checked by the select condition, the select can
// be replaced by bitwise logic to set/clear one bit of the constant result.
if (TC.getBitWidth() != AndMask.getBitWidth() || (TC ^ FC) != AndMask)
return nullptr;
if (CreateAnd) {
// If we have to create an 'and', then we must kill the cmp to not
// increase the instruction count.
if (!Cmp->hasOneUse())
return nullptr;
V = Builder.CreateAnd(V, ConstantInt::get(SelType, AndMask));
}
bool ExtraBitInTC = TC.ugt(FC);
if (Pred == ICmpInst::ICMP_EQ) {
// If the masked bit in V is clear, clear or set the bit in the result:
// (V & AndMaskC) == 0 ? TC : FC --> (V & AndMaskC) ^ TC
// (V & AndMaskC) == 0 ? TC : FC --> (V & AndMaskC) | TC
Constant *C = ConstantInt::get(SelType, TC);
return ExtraBitInTC ? Builder.CreateXor(V, C) : Builder.CreateOr(V, C);
}
if (Pred == ICmpInst::ICMP_NE) {
// If the masked bit in V is set, set or clear the bit in the result:
// (V & AndMaskC) != 0 ? TC : FC --> (V & AndMaskC) | FC
// (V & AndMaskC) != 0 ? TC : FC --> (V & AndMaskC) ^ FC
Constant *C = ConstantInt::get(SelType, FC);
return ExtraBitInTC ? Builder.CreateOr(V, C) : Builder.CreateXor(V, C);
}
llvm_unreachable("Only expecting equality predicates");
}
// 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.isNullValue() ? TC : FC;
unsigned ValZeros = ValC.logBase2();
unsigned AndZeros = AndMask.logBase2();
// 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.
bool ShouldNotVal = !TC.isNullValue();
ShouldNotVal ^= Pred == ICmpInst::ICMP_NE;
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::Mul:
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::Shl: // Can only fold on the shift amount.
case Instruction::LShr:
case Instruction::AShr:
return 1;
default:
return 0; // Cannot fold
}
}
/// For the same transformation as the previous function, return the identity
/// constant that goes into the select.
static APInt getSelectFoldableConstant(BinaryOperator *I) {
switch (I->getOpcode()) {
default: llvm_unreachable("This cannot happen!");
case Instruction::Add:
case Instruction::Sub:
case Instruction::Or:
case Instruction::Xor:
case Instruction::Shl:
case Instruction::LShr:
case Instruction::AShr:
return APInt::getNullValue(I->getType()->getScalarSizeInBits());
case Instruction::And:
return APInt::getAllOnesValue(I->getType()->getScalarSizeInBits());
case Instruction::Mul:
return APInt(I->getType()->getScalarSizeInBits(), 1);
}
}
/// We have (select c, TI, FI), and we know that TI and FI have the same opcode.
Instruction *InstCombiner::foldSelectOpOp(SelectInst &SI, Instruction *TI,
Instruction *FI) {
// Don't break up min/max patterns. The hasOneUse checks below prevent that
// for most cases, but vector min/max with bitcasts can be transformed. If the
// one-use restrictions are eased for other patterns, we still don't want to
// obfuscate min/max.
if ((match(&SI, m_SMin(m_Value(), m_Value())) ||
match(&SI, m_SMax(m_Value(), m_Value())) ||
match(&SI, m_UMin(m_Value(), m_Value())) ||
match(&SI, m_UMax(m_Value(), m_Value()))))
return nullptr;
// 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 (CondTy->isVectorTy()) {
if (!FIOpndTy->isVectorTy())
return nullptr;
if (CondTy->getVectorNumElements() != FIOpndTy->getVectorNumElements())
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());
}
// Cond ? -X : -Y --> -(Cond ? X : Y)
Value *X, *Y;
if (match(TI, m_FNeg(m_Value(X))) && match(FI, m_FNeg(m_Value(Y))) &&
(TI->hasOneUse() || FI->hasOneUse())) {
Value *NewSel = Builder.CreateSelect(Cond, X, Y, SI.getName() + ".v", &SI);
// TODO: Remove the hack for the binop form when the unary op is optimized
// properly with all IR passes.
if (TI->getOpcode() != Instruction::FNeg)
return BinaryOperator::CreateFNegFMF(NewSel, cast<BinaryOperator>(TI));
return UnaryOperator::CreateFNeg(NewSel);
}
// 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 ||
(!isa<BinaryOperator>(TI) && !isa<GetElementPtrInst>(TI)) ||
!TI->hasOneUse() || !FI->hasOneUse())
return nullptr;
// Figure out if the operations have any operands in common.
Value *MatchOp, *OtherOpT, *OtherOpF;
bool MatchIsOpZero;
if (TI->getOperand(0) == FI->getOperand(0)) {
MatchOp = TI->getOperand(0);
OtherOpT = TI->getOperand(1);
OtherOpF = FI->getOperand(1);
MatchIsOpZero = true;
} else if (TI->getOperand(1) == FI->getOperand(1)) {
MatchOp = TI->getOperand(1);
OtherOpT = TI->getOperand(0);
OtherOpF = FI->getOperand(0);
MatchIsOpZero = false;
} else if (!TI->isCommutative()) {
return nullptr;
} else if (TI->getOperand(0) == FI->getOperand(1)) {
MatchOp = TI->getOperand(0);
OtherOpT = TI->getOperand(1);
OtherOpF = FI->getOperand(0);
MatchIsOpZero = true;
} else if (TI->getOperand(1) == FI->getOperand(0)) {
MatchOp = TI->getOperand(1);
OtherOpT = TI->getOperand(0);
OtherOpF = FI->getOperand(1);
MatchIsOpZero = true;
} else {
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 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->getResultElementType();
return TGEP->isInBounds() && FGEP->isInBounds()
? GetElementPtrInst::CreateInBounds(ElementType, Op0, {Op1})
: GetElementPtrInst::Create(ElementType, Op0, {Op1});
}
llvm_unreachable("Expected BinaryOperator or GEP");
return nullptr;
}
static bool isSelect01(const APInt &C1I, const APInt &C2I) {
if (!C1I.isNullValue() && !C2I.isNullValue()) // One side must be zero.
return false;
return C1I.isOneValue() || C1I.isAllOnesValue() ||
C2I.isOneValue() || C2I.isAllOnesValue();
}
/// Try to fold the select into one of the operands to allow further
/// optimization.
Instruction *InstCombiner::foldSelectIntoOp(SelectInst &SI, Value *TrueVal,
Value *FalseVal) {
// See the comment above GetSelectFoldableOperands for a description of the
// transformation we are doing here.
if (auto *TVI = dyn_cast<BinaryOperator>(TrueVal)) {
if (TVI->hasOneUse() && !isa<Constant>(FalseVal)) {
if (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) {
APInt CI = getSelectFoldableConstant(TVI);
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(CI, *OOpC))) {
Value *C = ConstantInt::get(OOp->getType(), CI);
Value *NewSel = Builder.CreateSelect(SI.getCondition(), OOp, C);
NewSel->takeName(TVI);
BinaryOperator *BO = BinaryOperator::Create(TVI->getOpcode(),
FalseVal, NewSel);
BO->copyIRFlags(TVI);
return BO;
}
}
}
}
}
if (auto *FVI = dyn_cast<BinaryOperator>(FalseVal)) {
if (FVI->hasOneUse() && !isa<Constant>(TrueVal)) {
if (unsigned SFO = getSelectFoldableOperands(FVI)) {
unsigned OpToFold = 0;
if ((SFO & 1) && TrueVal == FVI->getOperand(0)) {
OpToFold = 1;
} else if ((SFO & 2) && TrueVal == FVI->getOperand(1)) {
OpToFold = 2;
}
if (OpToFold) {
APInt CI = getSelectFoldableConstant(FVI);
Value *OOp = FVI->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(CI, *OOpC))) {
Value *C = ConstantInt::get(OOp->getType(), CI);
Value *NewSel = Builder.CreateSelect(SI.getCondition(), C, OOp);
NewSel->takeName(FVI);
BinaryOperator *BO = BinaryOperator::Create(FVI->getOpcode(),
TrueVal, NewSel);
BO->copyIRFlags(FVI);
return BO;
}
}
}
}
}
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 ICmpInst *Cmp,
Value *TVal, Value *FVal,
InstCombiner::BuilderTy &Builder) {
if (!(Cmp->hasOneUse() && Cmp->getOperand(0)->hasOneUse() &&
Cmp->getPredicate() == ICmpInst::ICMP_EQ &&
match(Cmp->getOperand(1), m_Zero()) && match(FVal, m_One())))
return nullptr;
// The TrueVal has general form of: and %B, 1
Value *B;
if (!match(TVal, m_OneUse(m_And(m_Value(B), m_One()))))
return nullptr;
// Where %B may be optionally shifted: lshr %X, %Z.
Value *X, *Z;
const bool HasShift = match(B, m_OneUse(m_LShr(m_Value(X), m_Value(Z))));
if (!HasShift)
X = B;
Value *Y;
if (!match(Cmp->getOperand(0), m_c_And(m_Specific(X), m_Value(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), Y, (or Y, C2))
/// into:
/// (or (shl (and X, C1), C3), Y)
/// iff:
/// C1 and C2 are both powers of 2
/// where:
/// C3 = Log(C2) - Log(C1)
///
/// 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 *foldSelectICmpAndOr(const ICmpInst *IC, Value *TrueVal,
Value *FalseVal,
InstCombiner::BuilderTy &Builder) {
// Only handle integer compares. Also, if this is a vector select, we need a
// vector compare.
if (!TrueVal->getType()->isIntOrIntVectorTy() ||
TrueVal->getType()->isVectorTy() != IC->getType()->isVectorTy())
return nullptr;
Value *CmpLHS = IC->getOperand(0);
Value *CmpRHS = IC->getOperand(1);
Value *V;
unsigned C1Log;
bool IsEqualZero;
bool NeedAnd = false;
if (IC->isEquality()) {
if (!match(CmpRHS, m_Zero()))
return nullptr;
const APInt *C1;
if (!match(CmpLHS, m_And(m_Value(), m_Power2(C1))))
return nullptr;
V = CmpLHS;
C1Log = C1->logBase2();
IsEqualZero = IC->getPredicate() == ICmpInst::ICMP_EQ;
} else if (IC->getPredicate() == ICmpInst::ICMP_SLT ||
IC->getPredicate() == ICmpInst::ICMP_SGT) {
// We also need to recognize (icmp slt (trunc (X)), 0) and
// (icmp sgt (trunc (X)), -1).
IsEqualZero = IC->getPredicate() == ICmpInst::ICMP_SGT;
if ((IsEqualZero && !match(CmpRHS, m_AllOnes())) ||
(!IsEqualZero && !match(CmpRHS, m_Zero())))
return nullptr;
if (!match(CmpLHS, m_OneUse(m_Trunc(m_Value(V)))))
return nullptr;
C1Log = CmpLHS->getType()->getScalarSizeInBits() - 1;
NeedAnd = true;
} else {
return nullptr;
}
const APInt *C2;
bool OrOnTrueVal = false;
bool OrOnFalseVal = match(FalseVal, m_Or(m_Specific(TrueVal), m_Power2(C2)));
if (!OrOnFalseVal)
OrOnTrueVal = match(TrueVal, m_Or(m_Specific(FalseVal), m_Power2(C2)));
if (!OrOnFalseVal && !OrOnTrueVal)
return nullptr;
Value *Y = OrOnFalseVal ? TrueVal : FalseVal;
unsigned C2Log = C2->logBase2();
bool NeedXor = (!IsEqualZero && OrOnFalseVal) || (IsEqualZero && OrOnTrueVal);
bool NeedShift = C1Log != C2Log;
bool NeedZExtTrunc = Y->getType()->getScalarSizeInBits() !=
V->getType()->getScalarSizeInBits();
// Make sure we don't create more instructions than we save.
Value *Or = OrOnFalseVal ? FalseVal : TrueVal;
if ((NeedShift + NeedXor + NeedZExtTrunc) >
(IC->hasOneUse() + Or->hasOneUse()))
return nullptr;
if (NeedAnd) {
// Insert the AND instruction on the input to the truncate.
APInt C1 = APInt::getOneBitSet(V->getType()->getScalarSizeInBits(), C1Log);
V = Builder.CreateAnd(V, ConstantInt::get(V->getType(), C1));
}
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);
return Builder.CreateOr(V, Y);
}
/// Transform patterns such as (a > b) ? a - b : 0 into usub.sat(a, b).
/// There are 8 commuted/swapped variants of this pattern.
/// TODO: Also support a - UMIN(a,b) patterns.
static Value *canonicalizeSaturatedSubtract(const ICmpInst *ICI,
const Value *TrueVal,
const Value *FalseVal,
InstCombiner::BuilderTy &Builder) {
ICmpInst::Predicate Pred = ICI->getPredicate();
if (!ICmpInst::isUnsigned(Pred))
return nullptr;
// (b > a) ? 0 : a - b -> (b <= a) ? a - b : 0
if (match(TrueVal, m_Zero())) {
Pred = ICmpInst::getInversePredicate(Pred);
std::swap(TrueVal, FalseVal);
}
if (!match(FalseVal, m_Zero()))
return nullptr;
Value *A = ICI->getOperand(0);
Value *B = ICI->getOperand(1);
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!");
// Account for swapped form of subtraction: ((a > b) ? b - a : 0).
bool IsNegative = false;
if (match(TrueVal, m_Sub(m_Specific(B), m_Specific(A))))
IsNegative = true;
else if (!match(TrueVal, m_Sub(m_Specific(A), m_Specific(B))))
return nullptr;
// If sub is used anywhere else, we wouldn't be able to eliminate it
// afterwards.
if (!TrueVal->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 *canonicalizeSaturatedAdd(ICmpInst *Cmp, Value *TVal, Value *FVal,
InstCombiner::BuilderTy &Builder) {
if (!Cmp->hasOneUse())
return nullptr;
// 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, *CmpC;
if (Pred == ICmpInst::ICMP_ULT &&
match(TVal, m_Add(m_Value(X), m_APInt(C))) && X == Cmp0 &&
match(FVal, m_AllOnes()) && match(Cmp1, m_APInt(CmpC)) && *CmpC == ~*C) {
// (X u< ~C) ? (X + C) : -1 --> uadd.sat(X, C)
return Builder.CreateBinaryIntrinsic(
Intrinsic::uadd_sat, X, ConstantInt::get(X->getType(), *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. Just
// swapping the compare operands is legal, because the selected value is the
// same in case of equality, so we can interchange u< and u<=.
if (match(FVal, m_AllOnes())) {
std::swap(TVal, FVal);
std::swap(Cmp0, Cmp1);
}
if (!match(TVal, m_AllOnes()))
return nullptr;
// Canonicalize predicate to 'ULT'.
if (Pred == ICmpInst::ICMP_UGT) {
Pred = ICmpInst::ICMP_ULT;
std::swap(Cmp0, Cmp1);
}
if (Pred != ICmpInst::ICMP_ULT)
return nullptr;
// Match unsigned saturated add of 2 variables with an unnecessary 'not'.
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.
X = Cmp0;
Y = Cmp1;
if (match(FVal, m_c_Add(m_Not(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));
}
return nullptr;
}
/// Attempt to fold a cttz/ctlz followed by a icmp plus select into a single
/// call to cttz/ctlz with flag 'is_zero_undef' 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,
InstCombiner::BuilderTy &Builder) {
ICmpInst::Predicate Pred = ICI->getPredicate();
Value *CmpLHS = ICI->getOperand(0);
Value *CmpRHS = ICI->getOperand(1);
// Check if the condition value compares a value for equality against zero.
if (!ICI->isEquality() || !match(CmpRHS, m_Zero()))
return nullptr;
Value *Count = FalseVal;
Value *ValueOnZero = TrueVal;
if (Pred == ICmpInst::ICMP_NE)
std::swap(Count, ValueOnZero);
// Skip zero extend/truncate.
Value *V = nullptr;
if (match(Count, m_ZExt(m_Value(V))) ||
match(Count, m_Trunc(m_Value(V))))
Count = V;
// 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.
if (!match(Count, m_Intrinsic<Intrinsic::cttz>(m_Specific(CmpLHS))) &&
!match(Count, m_Intrinsic<Intrinsic::ctlz>(m_Specific(CmpLHS))))
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))) {
// Explicitly clear the 'undef_on_zero' flag.
IntrinsicInst *NewI = cast<IntrinsicInst>(II->clone());
NewI->setArgOperand(1, ConstantInt::getFalse(NewI->getContext()));
Builder.Insert(NewI);
return Builder.CreateZExtOrTrunc(NewI, ValueOnZero->getType());
}
// If the ValueOnZero is not the bitwidth, we can at least make use of the
// fact that the cttz/ctlz result will not be used if the input is zero, so
// it's okay to relax it to undef for that case.
if (II->hasOneUse() && !match(II->getArgOperand(1), m_One()))
II->setArgOperand(1, ConstantInt::getTrue(II->getContext()));
return nullptr;
}
/// Return true if we find and adjust an icmp+select pattern where the compare
/// is with a constant that can be incremented or decremented to match the
/// minimum or maximum idiom.
static bool adjustMinMax(SelectInst &Sel, ICmpInst &Cmp) {
ICmpInst::Predicate Pred = Cmp.getPredicate();
Value *CmpLHS = Cmp.getOperand(0);
Value *CmpRHS = Cmp.getOperand(1);
Value *TrueVal = Sel.getTrueValue();
Value *FalseVal = Sel.getFalseValue();
// We may move or edit the compare, so make sure the select is the only user.
const APInt *CmpC;
if (!Cmp.hasOneUse() || !match(CmpRHS, m_APInt(CmpC)))
return false;
// These transforms only work for selects of integers or vector selects of
// integer vectors.
Type *SelTy = Sel.getType();
auto *SelEltTy = dyn_cast<IntegerType>(SelTy->getScalarType());
if (!SelEltTy || SelTy->isVectorTy() != Cmp.getType()->isVectorTy())
return false;
Constant *AdjustedRHS;
if (Pred == ICmpInst::ICMP_UGT || Pred == ICmpInst::ICMP_SGT)
AdjustedRHS = ConstantInt::get(CmpRHS->getType(), *CmpC + 1);
else if (Pred == ICmpInst::ICMP_ULT || Pred == ICmpInst::ICMP_SLT)
AdjustedRHS = ConstantInt::get(CmpRHS->getType(), *CmpC - 1);
else
return false;
// X > C ? X : C+1 --> X < C+1 ? C+1 : X
// X < C ? X : C-1 --> X > C-1 ? C-1 : X
if ((CmpLHS == TrueVal && AdjustedRHS == FalseVal) ||
(CmpLHS == FalseVal && AdjustedRHS == TrueVal)) {
; // Nothing to do here. Values match without any sign/zero extension.
}
// Types do not match. Instead of calculating this with mixed types, promote
// all to the larger type. This enables scalar evolution to analyze this
// expression.
else if (CmpRHS->getType()->getScalarSizeInBits() < SelEltTy->getBitWidth()) {
Constant *SextRHS = ConstantExpr::getSExt(AdjustedRHS, SelTy);
// X = sext x; x >s c ? X : C+1 --> X = sext x; X <s C+1 ? C+1 : X
// X = sext x; x <s c ? X : C-1 --> X = sext x; X >s C-1 ? C-1 : X
// X = sext x; x >u c ? X : C+1 --> X = sext x; X <u C+1 ? C+1 : X
// X = sext x; x <u c ? X : C-1 --> X = sext x; X >u C-1 ? C-1 : X
if (match(TrueVal, m_SExt(m_Specific(CmpLHS))) && SextRHS == FalseVal) {
CmpLHS = TrueVal;
AdjustedRHS = SextRHS;
} else if (match(FalseVal, m_SExt(m_Specific(CmpLHS))) &&
SextRHS == TrueVal) {
CmpLHS = FalseVal;
AdjustedRHS = SextRHS;
} else if (Cmp.isUnsigned()) {
Constant *ZextRHS = ConstantExpr::getZExt(AdjustedRHS, SelTy);
// X = zext x; x >u c ? X : C+1 --> X = zext x; X <u C+1 ? C+1 : X
// X = zext x; x <u c ? X : C-1 --> X = zext x; X >u C-1 ? C-1 : X
// zext + signed compare cannot be changed:
// 0xff <s 0x00, but 0x00ff >s 0x0000
if (match(TrueVal, m_ZExt(m_Specific(CmpLHS))) && ZextRHS == FalseVal) {
CmpLHS = TrueVal;
AdjustedRHS = ZextRHS;
} else if (match(FalseVal, m_ZExt(m_Specific(CmpLHS))) &&
ZextRHS == TrueVal) {
CmpLHS = FalseVal;
AdjustedRHS = ZextRHS;
} else {
return false;
}
} else {
return false;
}
} else {
return false;
}
Pred = ICmpInst::getSwappedPredicate(Pred);
CmpRHS = AdjustedRHS;
std::swap(FalseVal, TrueVal);
Cmp.setPredicate(Pred);
Cmp.setOperand(0, CmpLHS);
Cmp.setOperand(1, CmpRHS);
Sel.setOperand(1, TrueVal);
Sel.setOperand(2, FalseVal);
Sel.swapProfMetadata();
// Move the compare instruction right before the select instruction. Otherwise
// the sext/zext value may be defined after the compare instruction uses it.
Cmp.moveBefore(&Sel);
return true;
}
/// If this is an integer min/max (icmp + select) with a constant operand,
/// create the canonical icmp for the min/max operation and canonicalize the
/// constant to the 'false' operand of the select:
/// select (icmp Pred X, C1), C2, X --> select (icmp Pred' X, C2), X, C2
/// Note: if C1 != C2, this will change the icmp constant to the existing
/// constant operand of the select.
static Instruction *
canonicalizeMinMaxWithConstant(SelectInst &Sel, ICmpInst &Cmp,
InstCombiner::BuilderTy &Builder) {
if (!Cmp.hasOneUse() || !isa<Constant>(Cmp.getOperand(1)))
return nullptr;
// Canonicalize the compare predicate based on whether we have min or max.
Value *LHS, *RHS;
SelectPatternResult SPR = matchSelectPattern(&Sel, LHS, RHS);
if (!SelectPatternResult::isMinOrMax(SPR.Flavor))
return nullptr;
// Is this already canonical?
ICmpInst::Predicate CanonicalPred = getMinMaxPred(SPR.Flavor);
if (Cmp.getOperand(0) == LHS && Cmp.getOperand(1) == RHS &&
Cmp.getPredicate() == CanonicalPred)
return nullptr;
// Create the canonical compare and plug it into the select.
Sel.setCondition(Builder.CreateICmp(CanonicalPred, LHS, RHS));
// If the select operands did not change, we're done.
if (Sel.getTrueValue() == LHS && Sel.getFalseValue() == RHS)
return &Sel;
// If we are swapping the select operands, swap the metadata too.
assert(Sel.getTrueValue() == RHS && Sel.getFalseValue() == LHS &&
"Unexpected results from matchSelectPattern");
Sel.setTrueValue(LHS);
Sel.setFalseValue(RHS);
Sel.swapProfMetadata();
return &Sel;
}
/// There are many select variants for each of ABS/NABS.
/// In matchSelectPattern(), there are different compare constants, compare
/// predicates/operands and select operands.
/// In isKnownNegation(), there are different formats of negated operands.
/// Canonicalize all these variants to 1 pattern.
/// This makes CSE more likely.
static Instruction *canonicalizeAbsNabs(SelectInst &Sel, ICmpInst &Cmp,
InstCombiner::BuilderTy &Builder) {
if (!Cmp.hasOneUse() || !isa<Constant>(Cmp.getOperand(1)))
return nullptr;
// Choose a sign-bit check for the compare (likely simpler for codegen).
// ABS: (X <s 0) ? -X : X
// NABS: (X <s 0) ? X : -X
Value *LHS, *RHS;
SelectPatternFlavor SPF = matchSelectPattern(&Sel, LHS, RHS).Flavor;
if (SPF != SelectPatternFlavor::SPF_ABS &&
SPF != SelectPatternFlavor::SPF_NABS)
return nullptr;
Value *TVal = Sel.getTrueValue();
Value *FVal = Sel.getFalseValue();
assert(isKnownNegation(TVal, FVal) &&
"Unexpected result from matchSelectPattern");
// The compare may use the negated abs()/nabs() operand, or it may use
// negation in non-canonical form such as: sub A, B.
bool CmpUsesNegatedOp = match(Cmp.getOperand(0), m_Neg(m_Specific(TVal))) ||
match(Cmp.getOperand(0), m_Neg(m_Specific(FVal)));
bool CmpCanonicalized = !CmpUsesNegatedOp &&
match(Cmp.getOperand(1), m_ZeroInt()) &&
Cmp.getPredicate() == ICmpInst::ICMP_SLT;
bool RHSCanonicalized = match(RHS, m_Neg(m_Specific(LHS)));
// Is this already canonical?
if (CmpCanonicalized && RHSCanonicalized)
return nullptr;
// If RHS is used by other instructions except compare and select, don't
// canonicalize it to not increase the instruction count.
if (!(RHS->hasOneUse() || (RHS->hasNUses(2) && CmpUsesNegatedOp)))
return nullptr;
// Create the canonical compare: icmp slt LHS 0.
if (!CmpCanonicalized) {
Cmp.setPredicate(ICmpInst::ICMP_SLT);
Cmp.setOperand(1, ConstantInt::getNullValue(Cmp.getOperand(0)->getType()));
if (CmpUsesNegatedOp)
Cmp.setOperand(0, LHS);
}
// Create the canonical RHS: RHS = sub (0, LHS).
if (!RHSCanonicalized) {
assert(RHS->hasOneUse() && "RHS use number is not right");
RHS = Builder.CreateNeg(LHS);
if (TVal == LHS) {
Sel.setFalseValue(RHS);
FVal = RHS;
} else {
Sel.setTrueValue(RHS);
TVal = RHS;
}
}
// If the select operands do not change, we're done.
if (SPF == SelectPatternFlavor::SPF_NABS) {
if (TVal == LHS)
return &Sel;
assert(FVal == LHS && "Unexpected results from matchSelectPattern");
} else {
if (FVal == LHS)
return &Sel;
assert(TVal == LHS && "Unexpected results from matchSelectPattern");
}
// We are swapping the select operands, so swap the metadata too.
Sel.setTrueValue(FVal);
Sel.setFalseValue(TVal);
Sel.swapProfMetadata();
return &Sel;
}
/// Visit a SelectInst that has an ICmpInst as its first operand.
Instruction *InstCombiner::foldSelectInstWithICmp(SelectInst &SI,
ICmpInst *ICI) {
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
if (Instruction *NewSel = canonicalizeMinMaxWithConstant(SI, *ICI, Builder))
return NewSel;
if (Instruction *NewAbs = canonicalizeAbsNabs(SI, *ICI, Builder))
return NewAbs;
bool Changed = adjustMinMax(SI, *ICI);
if (Value *V = foldSelectICmpAnd(SI, ICI, Builder))
return replaceInstUsesWith(SI, V);
// NOTE: if we wanted to, this is where to detect integer MIN/MAX
ICmpInst::Predicate Pred = ICI->getPredicate();
Value *CmpLHS = ICI->getOperand(0);
Value *CmpRHS = ICI->getOperand(1);
if (CmpRHS != CmpLHS && isa<Constant>(CmpRHS)) {
if (CmpLHS == TrueVal && Pred == ICmpInst::ICMP_EQ) {
// Transform (X == C) ? X : Y -> (X == C) ? C : Y
SI.setOperand(1, CmpRHS);
Changed = true;
} else if (CmpLHS == FalseVal && Pred == ICmpInst::ICMP_NE) {
// Transform (X != C) ? Y : X -> (X != C) ? Y : C
SI.setOperand(2, CmpRHS);
Changed = true;
}
}
// FIXME: This code is nearly duplicated in InstSimplify. Using/refactoring
// decomposeBitTestICmp() might help.
{
unsigned BitWidth =
DL.getTypeSizeInBits(TrueVal->getType()->getScalarType());
APInt MinSignedValue = APInt::getSignedMinValue(BitWidth);
Value *X;
const APInt *Y, *C;
bool TrueWhenUnset;
bool IsBitTest = false;
if (ICmpInst::isEquality(Pred) &&
match(CmpLHS, m_And(m_Value(X), m_Power2(Y))) &&
match(CmpRHS, m_Zero())) {
IsBitTest = true;
TrueWhenUnset = Pred == ICmpInst::ICMP_EQ;
} else if (Pred == ICmpInst::ICMP_SLT && match(CmpRHS, m_Zero())) {
X = CmpLHS;
Y = &MinSignedValue;
IsBitTest = true;
TrueWhenUnset = false;
} else if (Pred == ICmpInst::ICMP_SGT && match(CmpRHS, m_AllOnes())) {
X = CmpLHS;
Y = &MinSignedValue;
IsBitTest = true;
TrueWhenUnset = true;
}
if (IsBitTest) {
Value *V = nullptr;
// (X & Y) == 0 ? X : X ^ Y --> X & ~Y
if (TrueWhenUnset && TrueVal == X &&
match(FalseVal, m_Xor(m_Specific(X), m_APInt(C))) && *Y == *C)
V = Builder.CreateAnd(X, ~(*Y));
// (X & Y) != 0 ? X ^ Y : X --> X & ~Y
else if (!TrueWhenUnset && FalseVal == X &&
match(TrueVal, m_Xor(m_Specific(X), m_APInt(C))) && *Y == *C)
V = Builder.CreateAnd(X, ~(*Y));
// (X & Y) == 0 ? X ^ Y : X --> X | Y
else if (TrueWhenUnset && FalseVal == X &&
match(TrueVal, m_Xor(m_Specific(X), m_APInt(C))) && *Y == *C)
V = Builder.CreateOr(X, *Y);
// (X & Y) != 0 ? X : X ^ Y --> X | Y
else if (!TrueWhenUnset && TrueVal == X &&
match(FalseVal, m_Xor(m_Specific(X), m_APInt(C))) && *Y == *C)
V = Builder.CreateOr(X, *Y);
if (V)
return replaceInstUsesWith(SI, V);
}
}
if (Instruction *V =
foldSelectICmpAndAnd(SI.getType(), ICI, TrueVal, FalseVal, Builder))
return V;
if (Value *V = foldSelectICmpAndOr(ICI, TrueVal, FalseVal, Builder))
return replaceInstUsesWith(SI, V);
if (Value *V = foldSelectCttzCtlz(ICI, TrueVal, FalseVal, Builder))
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);
return Changed ? &SI : nullptr;
}
/// SI is a select whose condition is a PHI node (but the two may be in
/// different blocks). See if the true/false values (V) are live in all of the
/// predecessor blocks of the PHI. For example, cases like this can't be mapped:
///
/// X = phi [ C1, BB1], [C2, BB2]
/// Y = add
/// Z = select X, Y, 0
///
/// because Y is not live in BB1/BB2.
static bool canSelectOperandBeMappingIntoPredBlock(const Value *V,
const SelectInst &SI) {
// If the value is a non-instruction value like a constant or argument, it
// can always be mapped.
const Instruction *I = dyn_cast<Instruction>(V);
if (!I) return true;
// If V is a PHI node defined in the same block as the condition PHI, we can
// map the arguments.
const PHINode *CondPHI = cast<PHINode>(SI.getCondition());
if (const PHINode *VP = dyn_cast<PHINode>(I))
if (VP->getParent() == CondPHI->getParent())
return true;
// Otherwise, if the PHI and select are defined in the same block and if V is
// defined in a different block, then we can transform it.
if (SI.getParent() == CondPHI->getParent() &&
I->getParent() != CondPHI->getParent())
return true;
// Otherwise we have a 'hard' case and we can't tell without doing more
// detailed dominator based analysis, punt.
return false;
}
/// We have an SPF (e.g. a min or max) of an SPF of the form:
/// SPF2(SPF1(A, B), C)
Instruction *InstCombiner::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);
// MAX(MIN(a, b), a) -> a
// MIN(MAX(a, b), a) -> a
// TODO: This could be done in instsimplify.
if ((SPF1 == SPF_SMIN && SPF2 == SPF_SMAX) ||
(SPF1 == SPF_SMAX && SPF2 == SPF_SMIN) ||
(SPF1 == SPF_UMIN && SPF2 == SPF_UMAX) ||
(SPF1 == SPF_UMAX && SPF2 == SPF_UMIN))
return replaceInstUsesWith(Outer, C);
}
if (SPF1 == SPF2) {
const APInt *CB, *CC;
if (match(B, m_APInt(CB)) && match(C, m_APInt(CC))) {
// MIN(MIN(A, 23), 97) -> MIN(A, 23)
// MAX(MAX(A, 97), 23) -> MAX(A, 97)
// TODO: This could be done in instsimplify.
if ((SPF1 == SPF_UMIN && CB->ule(*CC)) ||
(SPF1 == SPF_SMIN && CB->sle(*CC)) ||
(SPF1 == SPF_UMAX && CB->uge(*CC)) ||
(SPF1 == SPF_SMAX && CB->sge(*CC)))
return replaceInstUsesWith(Outer, Inner);
// MIN(MIN(A, 97), 23) -> MIN(A, 23)
// MAX(MAX(A, 23), 97) -> MAX(A, 97)
if ((SPF1 == SPF_UMIN && CB->ugt(*CC)) ||
(SPF1 == SPF_SMIN && CB->sgt(*CC)) ||
(SPF1 == SPF_UMAX && CB->ult(*CC)) ||
(SPF1 == SPF_SMAX && CB->slt(*CC))) {
Outer.replaceUsesOfWith(Inner, A);
return &Outer;
}
}
}
// ABS(ABS(X)) -> ABS(X)
// NABS(NABS(X)) -> NABS(X)
// TODO: This could be done in instsimplify.
if (SPF1 == SPF2 && (SPF1 == SPF_ABS || SPF1 == SPF_NABS)) {
return replaceInstUsesWith(Outer, Inner);
}
// ABS(NABS(X)) -> ABS(X)
// NABS(ABS(X)) -> NABS(X)
if ((SPF1 == SPF_ABS && SPF2 == SPF_NABS) ||
(SPF1 == SPF_NABS && SPF2 == SPF_ABS)) {
SelectInst *SI = cast<SelectInst>(Inner);
Value *NewSI =
Builder.CreateSelect(SI->getCondition(), SI->getFalseValue(),
SI->getTrueValue(), SI->getName(), SI);
return replaceInstUsesWith(Outer, NewSI);
}
auto IsFreeOrProfitableToInvert =
[&](Value *V, Value *&NotV, bool &ElidesXor) {
if (match(V, m_Not(m_Value(NotV)))) {
// If V has at most 2 uses then we can get rid of the xor operation
// entirely.
ElidesXor |= !V->hasNUsesOrMore(3);
return true;
}
if (IsFreeToInvert(V, !V->hasNUsesOrMore(3))) {
NotV = nullptr;
return true;
}
return false;
};
Value *NotA, *NotB, *NotC;
bool ElidesXor = false;
// MIN(MIN(~A, ~B), ~C) == ~MAX(MAX(A, B), C)
// MIN(MAX(~A, ~B), ~C) == ~MAX(MIN(A, B), C)
// MAX(MIN(~A, ~B), ~C) == ~MIN(MAX(A, B), C)
// MAX(MAX(~A, ~B), ~C) == ~MIN(MIN(A, B), C)
//
// This transform is performance neutral if we can elide at least one xor from
// the set of three operands, since we'll be tacking on an xor at the very
// end.
if (SelectPatternResult::isMinOrMax(SPF1) &&
SelectPatternResult::isMinOrMax(SPF2) &&
IsFreeOrProfitableToInvert(A, NotA, ElidesXor) &&
IsFreeOrProfitableToInvert(B, NotB, ElidesXor) &&
IsFreeOrProfitableToInvert(C, NotC, ElidesXor) && ElidesXor) {
if (!NotA)
NotA = Builder.CreateNot(A);
if (!NotB)
NotB = Builder.CreateNot(B);
if (!NotC)
NotC = Builder.CreateNot(C);
Value *NewInner = createMinMax(Builder, getInverseMinMaxFlavor(SPF1), NotA,
NotB);
Value *NewOuter = Builder.CreateNot(
createMinMax(Builder, getInverseMinMaxFlavor(SPF2), NewInner, NotC));
return replaceInstUsesWith(Outer, NewOuter);
}
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;
}
Instruction *InstCombiner::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();
auto *Cmp = dyn_cast<CmpInst>(Cond);
if (!SmallType->isIntOrIntVectorTy(1) &&
(!Cmp || Cmp->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 = ConstantExpr::getTrunc(C, SmallType);
Constant *ExtC = ConstantExpr::getCast(ExtOpcode, TruncC, SelType);
if (ExtC == C) {
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);
}
// If one arm of the select is the extend of the condition, replace that arm
// with the extension of the appropriate known bool value.
if (Cond == X) {
if (ExtInst == Sel.getTrueValue()) {
// select X, (sext X), C --> select X, -1, C
// select X, (zext X), C --> select X, 1, C
Constant *One = ConstantInt::getTrue(SmallType);
Constant *AllOnesOrOne = ConstantExpr::getCast(ExtOpcode, One, SelType);
return SelectInst::Create(Cond, AllOnesOrOne, C, "", nullptr, &Sel);
} else {
// select X, C, (sext X) --> select X, C, 0
// select X, C, (zext X) --> select X, C, 0
Constant *Zero = ConstantInt::getNullValue(SelType);
return SelectInst::Create(Cond, C, Zero, "", nullptr, &Sel);
}
}
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;
if (!CondVal->getType()->isVectorTy() || !match(CondVal, m_Constant(CondC)))
return nullptr;
unsigned NumElts = CondVal->getType()->getVectorNumElements();
SmallVector<Constant *, 16> Mask;
Mask.reserve(NumElts);
Type *Int32Ty = Type::getInt32Ty(CondVal->getContext());
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(ConstantInt::get(Int32Ty, i));
} else if (Elt->isNullValue()) {
// If the select condition element is false, choose from the 2nd vector.
Mask.push_back(ConstantInt::get(Int32Ty, 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(),
ConstantVector::get(Mask));
}
/// 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();
CmpInst::Predicate 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 CastInst::CreateBitOrPointerCast(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:
///
/// %0 = cmpxchg i64* %ptr, i64 %compare, i64 %new_value seq_cst seq_cst
/// %1 = extractvalue { i64, i1 } %0, 1
/// %2 = extractvalue { i64, i1 } %0, 0
/// %3 = select i1 %1, i64 %compare, i64 %2
/// ret i64 %3
///
/// The returned value of the cmpxchg instruction (%2) is the original value
/// located at %ptr prior to any update. If the cmpxchg operation succeeds, %2
/// must have been equal to %compare. Thus, the result of the select is always
/// equal to %2, and the code can be simplified to:
///
/// %0 = cmpxchg i64* %ptr, i64 %compare, i64 %new_value seq_cst seq_cst
/// %1 = extractvalue { i64, i1 } %0, 0
/// ret i64 %1
///
static Instruction *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 * {
auto *Extract = dyn_cast<ExtractValueInst>(V);
if (!Extract)
return nullptr;
if (Extract->getIndices()[0] != I)
return nullptr;
return dyn_cast<AtomicCmpXchgInst>(Extract->getAggregateOperand());
};
// 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 && X->getCompareOperand() == SI.getFalseValue()) {
SI.setTrueValue(SI.getFalseValue());
return &SI;
}
// 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 && X->getCompareOperand() == SI.getTrueValue()) {
SI.setTrueValue(SI.getFalseValue());
return &SI;
}
return nullptr;
}
static Instruction *moveAddAfterMinMax(SelectPatternFlavor SPF, Value *X,
Value *Y,
InstCombiner::BuilderTy &Builder) {
assert(SelectPatternResult::isMinOrMax(SPF) && "Expected min/max pattern");
bool IsUnsigned = SPF == SelectPatternFlavor::SPF_UMIN ||
SPF == SelectPatternFlavor::SPF_UMAX;
// TODO: If InstSimplify could fold all cases where C2 <= C1, we could change
// the constant value check to an assert.
Value *A;
const APInt *C1, *C2;
if (IsUnsigned && match(X, m_NUWAdd(m_Value(A), m_APInt(C1))) &&
match(Y, m_APInt(C2)) && C2->uge(*C1) && X->hasNUses(2)) {
// umin (add nuw A, C1), C2 --> add nuw (umin A, C2 - C1), C1
// umax (add nuw A, C1), C2 --> add nuw (umax A, C2 - C1), C1
Value *NewMinMax = createMinMax(Builder, SPF, A,
ConstantInt::get(X->getType(), *C2 - *C1));
return BinaryOperator::CreateNUW(BinaryOperator::Add, NewMinMax,
ConstantInt::get(X->getType(), *C1));
}
if (!IsUnsigned && match(X, m_NSWAdd(m_Value(A), m_APInt(C1))) &&
match(Y, m_APInt(C2)) && X->hasNUses(2)) {
bool Overflow;
APInt Diff = C2->ssub_ov(*C1, Overflow);
if (!Overflow) {
// smin (add nsw A, C1), C2 --> add nsw (smin A, C2 - C1), C1
// smax (add nsw A, C1), C2 --> add nsw (smax A, C2 - C1), C1
Value *NewMinMax = createMinMax(Builder, SPF, A,
ConstantInt::get(X->getType(), Diff));
return BinaryOperator::CreateNSW(BinaryOperator::Add, NewMinMax,
ConstantInt::get(X->getType(), *C1));
}
}
return nullptr;
}
/// Reduce a sequence of min/max with a common operand.
static Instruction *factorizeMinMaxTree(SelectPatternFlavor SPF, Value *LHS,
Value *RHS,
InstCombiner::BuilderTy &Builder) {
assert(SelectPatternResult::isMinOrMax(SPF) && "Expected a min/max");
// TODO: Allow FP min/max with nnan/nsz.
if (!LHS->getType()->isIntOrIntVectorTy())
return nullptr;
// Match 3 of the same min/max ops. Example: umin(umin(), umin()).
Value *A, *B, *C, *D;
SelectPatternResult L = matchSelectPattern(LHS, A, B);
SelectPatternResult R = matchSelectPattern(RHS, C, D);
if (SPF != L.Flavor || L.Flavor != R.Flavor)
return nullptr;
// Look for a common operand. The use checks are different than usual because
// a min/max pattern typically has 2 uses of each op: 1 by the cmp and 1 by
// the select.
Value *MinMaxOp = nullptr;
Value *ThirdOp = nullptr;
if (!LHS->hasNUsesOrMore(3) && RHS->hasNUsesOrMore(3)) {
// If the LHS is only used in this chain and the RHS is used outside of it,
// reuse the RHS min/max because that will eliminate the LHS.
if (D == A || C == A) {
// min(min(a, b), min(c, a)) --> min(min(c, a), b)
// min(min(a, b), min(a, d)) --> min(min(a, d), b)
MinMaxOp = RHS;
ThirdOp = B;
} else if (D == B || C == B) {
// min(min(a, b), min(c, b)) --> min(min(c, b), a)
// min(min(a, b), min(b, d)) --> min(min(b, d), a)
MinMaxOp = RHS;
ThirdOp = A;
}
} else if (!RHS->hasNUsesOrMore(3)) {
// Reuse the LHS. This will eliminate the RHS.
if (D == A || D == B) {
// min(min(a, b), min(c, a)) --> min(min(a, b), c)
// min(min(a, b), min(c, b)) --> min(min(a, b), c)
MinMaxOp = LHS;
ThirdOp = C;
} else if (C == A || C == B) {
// min(min(a, b), min(b, d)) --> min(min(a, b), d)
// min(min(a, b), min(c, b)) --> min(min(a, b), d)
MinMaxOp = LHS;
ThirdOp = D;
}
}
if (!MinMaxOp || !ThirdOp)
return nullptr;
CmpInst::Predicate P = getMinMaxPred(SPF);
Value *CmpABC = Builder.CreateICmp(P, MinMaxOp, ThirdOp);
return SelectInst::Create(CmpABC, MinMaxOp, ThirdOp);
}
/// Try to reduce a 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)
static Instruction *foldSelectRotate(SelectInst &Sel) {
// The false value of the select must be a rotate of the true value.
Value *Or0, *Or1;
if (!match(Sel.getFalseValue(), m_OneUse(m_Or(m_Value(Or0), m_Value(Or1)))))
return nullptr;
Value *TVal = Sel.getTrueValue();
Value *SA0, *SA1;
if (!match(Or0, m_OneUse(m_LogicalShift(m_Specific(TVal), m_Value(SA0)))) ||
!match(Or1, m_OneUse(m_LogicalShift(m_Specific(TVal), m_Value(SA1)))))
return nullptr;
auto ShiftOpcode0 = cast<BinaryOperator>(Or0)->getOpcode();
auto ShiftOpcode1 = cast<BinaryOperator>(Or1)->getOpcode();
if (ShiftOpcode0 == ShiftOpcode1)
return nullptr;
// We have one of these patterns so far:
// select ?, TVal, (or (lshr TVal, SA0), (shl TVal, SA1))
// select ?, TVal, (or (shl TVal, SA0), (lshr TVal, SA1))
// This must be a power-of-2 rotate for a bitmasking transform to be valid.
unsigned Width = Sel.getType()->getScalarSizeInBits();
if (!isPowerOf2_32(Width))
return nullptr;
// 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;
// Finally, see if the select is filtering out a shift-by-zero.
Value *Cond = Sel.getCondition();
ICmpInst::Predicate Pred;
if (!match(Cond, m_OneUse(m_ICmp(Pred, m_Specific(ShAmt), m_ZeroInt()))) ||
Pred != ICmpInst::ICMP_EQ)
return nullptr;
// This is a rotate that avoids shift-by-bitwidth UB in a suboptimal way.
// Convert to funnel shift intrinsic.
bool IsFshl = (ShAmt == SA0 && ShiftOpcode0 == BinaryOperator::Shl) ||
(ShAmt == SA1 && ShiftOpcode1 == BinaryOperator::Shl);
Intrinsic::ID IID = IsFshl ? Intrinsic::fshl : Intrinsic::fshr;
Function *F = Intrinsic::getDeclaration(Sel.getModule(), IID, Sel.getType());
return IntrinsicInst::Create(F, { TVal, TVal, ShAmt });
}
Instruction *InstCombiner::visitSelectInst(SelectInst &SI) {
Value *CondVal = SI.getCondition();
Value *TrueVal = SI.getTrueValue();
Value *FalseVal = SI.getFalseValue();
Type *SelType = SI.getType();
// FIXME: Remove this workaround when freeze related patches are done.
// For select with undef operand which feeds into an equality comparison,
// don't simplify it so loop unswitch can know the equality comparison
// may have an undef operand. This is a workaround for PR31652 caused by
// descrepancy about branch on undef between LoopUnswitch and GVN.
if (isa<UndefValue>(TrueVal) || isa<UndefValue>(FalseVal)) {
if (llvm::any_of(SI.users(), [&](User *U) {
ICmpInst *CI = dyn_cast<ICmpInst>(U);
if (CI && CI->isEquality())
return true;
return false;
})) {
return nullptr;
}
}
if (Value *V = SimplifySelectInst(CondVal, TrueVal, FalseVal,
SQ.getWithInstruction(&SI)))
return replaceInstUsesWith(SI, V);
if (Instruction *I = canonicalizeSelectToShuffle(SI))
return I;
// Canonicalize a one-use integer compare with a non-canonical predicate by
// inverting the predicate and swapping the select operands. This matches a
// compare canonicalization for conditional branches.
// TODO: Should we do the same for FP compares?
CmpInst::Predicate Pred;
if (match(CondVal, m_OneUse(m_ICmp(Pred, m_Value(), m_Value()))) &&
!isCanonicalPredicate(Pred)) {
// Swap true/false values and condition.
CmpInst *Cond = cast<CmpInst>(CondVal);
Cond->setPredicate(CmpInst::getInversePredicate(Pred));
SI.setOperand(1, FalseVal);
SI.setOperand(2, TrueVal);
SI.swapProfMetadata();
Worklist.Add(Cond);
return &SI;
}
if (SelType->isIntOrIntVectorTy(1) &&
TrueVal->getType() == CondVal->getType()) {
if (match(TrueVal, m_One())) {
// Change: A = select B, true, C --> A = or B, C
return BinaryOperator::CreateOr(CondVal, FalseVal);
}
if (match(TrueVal, m_Zero())) {
// Change: A = select B, false, C --> A = and !B, C
Value *NotCond = Builder.CreateNot(CondVal, "not." + CondVal->getName());
return BinaryOperator::CreateAnd(NotCond, FalseVal);
}
if (match(FalseVal, m_Zero())) {
// Change: A = select B, C, false --> A = and B, C
return BinaryOperator::CreateAnd(CondVal, TrueVal);
}
if (match(FalseVal, m_One())) {
// Change: A = select B, C, true --> A = or !B, C
Value *NotCond = Builder.CreateNot(CondVal, "not." + CondVal->getName());
return BinaryOperator::CreateOr(NotCond, TrueVal);
}
// select a, a, b -> a | b
// select a, b, a -> a & b
if (CondVal == TrueVal)
return BinaryOperator::CreateOr(CondVal, FalseVal);
if (CondVal == FalseVal)
return BinaryOperator::CreateAnd(CondVal, TrueVal);
// select a, ~a, b -> (~a) & b
// select a, b, ~a -> (~a) | b
if (match(TrueVal, m_Not(m_Specific(CondVal))))
return BinaryOperator::CreateAnd(TrueVal, FalseVal);
if (match(FalseVal, m_Not(m_Specific(CondVal))))
return BinaryOperator::CreateOr(TrueVal, FalseVal);
}
// 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.
if (SelType->isIntOrIntVectorTy() &&
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);
}
}
// See if we are selecting two values based on a comparison of the two values.
if (FCmpInst *FCI = dyn_cast<FCmpInst>(CondVal)) {
if (FCI->getOperand(0) == TrueVal && FCI->getOperand(1) == FalseVal) {
// Canonicalize to use ordered comparisons by swapping the select
// operands.
//
// e.g.
// (X ugt Y) ? X : Y -> (X ole Y) ? Y : X
if (FCI->hasOneUse() && FCmpInst::isUnordered(FCI->getPredicate())) {
FCmpInst::Predicate InvPred = FCI->getInversePredicate();
IRBuilder<>::FastMathFlagGuard FMFG(Builder);
Builder.setFastMathFlags(FCI->getFastMathFlags());
Value *NewCond = Builder.CreateFCmp(InvPred, TrueVal, FalseVal,
FCI->getName() + ".inv");
return SelectInst::Create(NewCond, FalseVal, TrueVal,
SI.getName() + ".p");
}
// NOTE: if we wanted to, this is where to detect MIN/MAX
} else if (FCI->getOperand(0) == FalseVal && FCI->getOperand(1) == TrueVal){
// Canonicalize to use ordered comparisons by swapping the select
// operands.
//
// e.g.
// (X ugt Y) ? X : Y -> (X ole Y) ? X : Y
if (FCI->hasOneUse() && FCmpInst::isUnordered(FCI->getPredicate())) {
FCmpInst::Predicate InvPred = FCI->getInversePredicate();
IRBuilder<>::FastMathFlagGuard FMFG(Builder);
Builder.setFastMathFlags(FCI->getFastMathFlags());
Value *NewCond = Builder.CreateFCmp(InvPred, FalseVal, TrueVal,
FCI->getName() + ".inv");
return SelectInst::Create(NewCond, FalseVal, TrueVal,
SI.getName() + ".p");
}
// NOTE: if we wanted to, this is where to detect MIN/MAX
}
// 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. We
// also require nnan because we do not want to unintentionally change the
// sign of a NaN value.
Value *X = FCI->getOperand(0);
FCmpInst::Predicate Pred = FCI->getPredicate();
if (match(FCI->getOperand(1), m_AnyZeroFP()) && FCI->hasNoNaNs()) {
// (X <= +/-0.0) ? (0.0 - X) : X --> fabs(X)
// (X > +/-0.0) ? X : (0.0 - X) --> fabs(X)
if ((X == FalseVal && Pred == FCmpInst::FCMP_OLE &&
match(TrueVal, m_FSub(m_PosZeroFP(), m_Specific(X)))) ||
(X == TrueVal && Pred == FCmpInst::FCMP_OGT &&
match(FalseVal, m_FSub(m_PosZeroFP(), m_Specific(X))))) {
Value *Fabs = Builder.CreateUnaryIntrinsic(Intrinsic::fabs, X, FCI);
return replaceInstUsesWith(SI, Fabs);
}
// With nsz:
// (X < +/-0.0) ? -X : X --> fabs(X)
// (X <= +/-0.0) ? -X : X --> fabs(X)
// (X > +/-0.0) ? X : -X --> fabs(X)
// (X >= +/-0.0) ? X : -X --> fabs(X)
if (FCI->hasNoSignedZeros() &&
((X == FalseVal && match(TrueVal, m_FNeg(m_Specific(X))) &&
(Pred == FCmpInst::FCMP_OLT || Pred == FCmpInst::FCMP_OLE)) ||
(X == TrueVal && match(FalseVal, m_FNeg(m_Specific(X))) &&
(Pred == FCmpInst::FCMP_OGT || Pred == FCmpInst::FCMP_OGE)))) {
Value *Fabs = Builder.CreateUnaryIntrinsic(Intrinsic::fabs, X, FCI);
return replaceInstUsesWith(SI, Fabs);
}
}
}
// See if we are selecting two values based on a comparison of the two values.
if (ICmpInst *ICI = dyn_cast<ICmpInst>(CondVal))
if (Instruction *Result = foldSelectInstWithICmp(SI, ICI))
return Result;
if (Instruction *Add = foldAddSubSelect(SI, Builder))
return Add;
// 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 = foldSelectExtConst(SI))
return I;
// 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;
// TODO.
// ABS(-X) -> ABS(X)
}
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 Pred = getMinMaxPred(SPF, SPR.Ordered);
Value *Cmp;
if (CmpInst::isIntPredicate(Pred)) {
Cmp = Builder.CreateICmp(Pred, LHS, RHS);
} else {
IRBuilder<>::FastMathFlagGuard FMFG(Builder);
auto FMF = cast<FPMathOperator>(SI.getCondition())->getFastMathFlags();
Builder.setFastMathFlags(FMF);
Cmp = Builder.CreateFCmp(Pred, LHS, RHS);
}
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);
}
// MAX(~a, ~b) -> ~MIN(a, b)
// MAX(~a, C) -> ~MIN(a, ~C)
// MIN(~a, ~b) -> ~MAX(a, b)
// MIN(~a, C) -> ~MAX(a, ~C)
auto moveNotAfterMinMax = [&](Value *X, Value *Y) -> Instruction * {
Value *A;
if (match(X, m_Not(m_Value(A))) && !X->hasNUsesOrMore(3) &&
!IsFreeToInvert(A, A->hasOneUse()) &&
// Passing false to only consider m_Not and constants.
IsFreeToInvert(Y, false)) {
Value *B = Builder.CreateNot(Y);
Value *NewMinMax = createMinMax(Builder, getInverseMinMaxFlavor(SPF),
A, B);
// Copy the profile metadata.
if (MDNode *MD = SI.getMetadata(LLVMContext::MD_prof)) {
cast<SelectInst>(NewMinMax)->setMetadata(LLVMContext::MD_prof, MD);
// Swap the metadata if the operands are swapped.
if (X == SI.getFalseValue() && Y == SI.getTrueValue())
cast<SelectInst>(NewMinMax)->swapProfMetadata();
}
return BinaryOperator::CreateNot(NewMinMax);
}
return nullptr;
};
if (Instruction *I = moveNotAfterMinMax(LHS, RHS))
return I;
if (Instruction *I = moveNotAfterMinMax(RHS, LHS))
return I;
if (Instruction *I = moveAddAfterMinMax(SPF, LHS, RHS, Builder))
return I;
if (Instruction *I = factorizeMinMaxTree(SPF, LHS, RHS, Builder))
return I;
}
}
// 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()))
// The true/false values have to be live in the PHI predecessor's blocks.
if (canSelectOperandBeMappingIntoPredBlock(TrueVal, SI) &&
canSelectOperandBeMappingIntoPredBlock(FalseVal, SI))
if (Instruction *NV = foldOpIntoPhi(SI, PN))
return NV;
if (SelectInst *TrueSI = dyn_cast<SelectInst>(TrueVal)) {
if (TrueSI->getCondition()->getType() == CondVal->getType()) {
// select(C, select(C, a, b), c) -> select(C, a, c)
if (TrueSI->getCondition() == CondVal) {
if (SI.getTrueValue() == TrueSI->getTrueValue())
return nullptr;
SI.setOperand(1, TrueSI->getTrueValue());
return &SI;
}
// select(C0, select(C1, a, b), b) -> select(C0&C1, a, b)
// 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->getFalseValue() == FalseVal && TrueSI->hasOneUse()) {
Value *And = Builder.CreateAnd(CondVal, TrueSI->getCondition());
SI.setOperand(0, And);
SI.setOperand(1, TrueSI->getTrueValue());
return &SI;
}
}
}
if (SelectInst *FalseSI = dyn_cast<SelectInst>(FalseVal)) {
if (FalseSI->getCondition()->getType() == CondVal->getType()) {
// select(C, a, select(C, b, c)) -> select(C, a, c)
if (FalseSI->getCondition() == CondVal) {
if (SI.getFalseValue() == FalseSI->getFalseValue())
return nullptr;
SI.setOperand(2, FalseSI->getFalseValue());
return &SI;
}
// select(C0, a, select(C1, a, b)) -> select(C0|C1, a, b)
if (FalseSI->getTrueValue() == TrueVal && FalseSI->hasOneUse()) {
Value *Or = Builder.CreateOr(CondVal, FalseSI->getCondition());
SI.setOperand(0, Or);
SI.setOperand(2, FalseSI->getFalseValue());
return &SI;
}
}
}
auto canMergeSelectThroughBinop = [](BinaryOperator *BO) {
// The select might be preventing a division by 0.
switch (BO->getOpcode()) {
default:
return true;
case Instruction::SRem:
case Instruction::URem:
case Instruction::SDiv:
case Instruction::UDiv:
return false;
}
};
// Try to simplify a binop sandwiched between 2 selects with the same
// condition.
// 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))) &&
canMergeSelectThroughBinop(TrueBO)) {
if (auto *TrueBOSI = dyn_cast<SelectInst>(TrueBO->getOperand(0))) {
if (TrueBOSI->getCondition() == CondVal) {
TrueBO->setOperand(0, TrueBOSI->getTrueValue());
Worklist.Add(TrueBO);
return &SI;
}
}
if (auto *TrueBOSI = dyn_cast<SelectInst>(TrueBO->getOperand(1))) {
if (TrueBOSI->getCondition() == CondVal) {
TrueBO->setOperand(1, TrueBOSI->getTrueValue());
Worklist.Add(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))) &&
canMergeSelectThroughBinop(FalseBO)) {
if (auto *FalseBOSI = dyn_cast<SelectInst>(FalseBO->getOperand(0))) {
if (FalseBOSI->getCondition() == CondVal) {
FalseBO->setOperand(0, FalseBOSI->getFalseValue());
Worklist.Add(FalseBO);
return &SI;
}
}
if (auto *FalseBOSI = dyn_cast<SelectInst>(FalseBO->getOperand(1))) {
if (FalseBOSI->getCondition() == CondVal) {
FalseBO->setOperand(1, FalseBOSI->getFalseValue());
Worklist.Add(FalseBO);
return &SI;
}
}
}
Value *NotCond;
if (match(CondVal, m_Not(m_Value(NotCond)))) {
SI.setOperand(0, NotCond);
SI.setOperand(1, FalseVal);
SI.setOperand(2, TrueVal);
SI.swapProfMetadata();
return &SI;
}
if (VectorType *VecTy = dyn_cast<VectorType>(SelType)) {
unsigned VWidth = VecTy->getNumElements();
APInt UndefElts(VWidth, 0);
APInt AllOnesEltMask(APInt::getAllOnesValue(VWidth));
if (Value *V = SimplifyDemandedVectorElts(&SI, AllOnesEltMask, UndefElts)) {
if (V != &SI)
return replaceInstUsesWith(SI, V);
return &SI;
}
}
// 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, 0, &SI);
if (Known.One.isOneValue())
return replaceInstUsesWith(SI, TrueVal);
if (Known.Zero.isOneValue())
return replaceInstUsesWith(SI, FalseVal);
}
if (Instruction *BitCastSel = foldSelectCmpBitcasts(SI, Builder))
return BitCastSel;
// Simplify selects that test the returned flag of cmpxchg instructions.
if (Instruction *Select = foldSelectCmpXchg(SI))
return Select;
if (Instruction *Select = foldSelectBinOpIdentity(SI, TLI))
return Select;
if (Instruction *Rot = foldSelectRotate(SI))
return Rot;
return nullptr;
}
Reference in New Issue View Git Blame Copy Permalink