maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/lib/Target/X86/X86ISelDAGToDAG.cpp

3908 lines
145 KiB
C++
Raw Blame History

//===- X86ISelDAGToDAG.cpp - A DAG pattern matching inst selector for X86 -===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file defines a DAG pattern matching instruction selector for X86,
// converting from a legalized dag to a X86 dag.
//
//===----------------------------------------------------------------------===//
#include "X86.h"
#include "X86MachineFunctionInfo.h"
#include "X86RegisterInfo.h"
#include "X86Subtarget.h"
#include "X86TargetMachine.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineFunction.h"
#include "llvm/CodeGen/SelectionDAGISel.h"
#include "llvm/Config/llvm-config.h"
#include "llvm/IR/ConstantRange.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/Type.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/KnownBits.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetOptions.h"
#include <stdint.h>
using namespace llvm;
#define DEBUG_TYPE "x86-isel"
STATISTIC(NumLoadMoved, "Number of loads moved below TokenFactor");
static cl::opt<bool> AndImmShrink("x86-and-imm-shrink", cl::init(true),
cl::desc("Enable setting constant bits to reduce size of mask immediates"),
cl::Hidden);
//===----------------------------------------------------------------------===//
// Pattern Matcher Implementation
//===----------------------------------------------------------------------===//
namespace {
/// This corresponds to X86AddressMode, but uses SDValue's instead of register
/// numbers for the leaves of the matched tree.
struct X86ISelAddressMode {
enum {
RegBase,
FrameIndexBase
} BaseType;
// This is really a union, discriminated by BaseType!
SDValue Base_Reg;
int Base_FrameIndex;
unsigned Scale;
SDValue IndexReg;
int32_t Disp;
SDValue Segment;
const GlobalValue *GV;
const Constant *CP;
const BlockAddress *BlockAddr;
const char *ES;
MCSymbol *MCSym;
int JT;
unsigned Align; // CP alignment.
unsigned char SymbolFlags; // X86II::MO_*
X86ISelAddressMode()
: BaseType(RegBase), Base_FrameIndex(0), Scale(1), IndexReg(), Disp(0),
Segment(), GV(nullptr), CP(nullptr), BlockAddr(nullptr), ES(nullptr),
MCSym(nullptr), JT(-1), Align(0), SymbolFlags(X86II::MO_NO_FLAG) {}
bool hasSymbolicDisplacement() const {
return GV != nullptr || CP != nullptr || ES != nullptr ||
MCSym != nullptr || JT != -1 || BlockAddr != nullptr;
}
bool hasBaseOrIndexReg() const {
return BaseType == FrameIndexBase ||
IndexReg.getNode() != nullptr || Base_Reg.getNode() != nullptr;
}
/// Return true if this addressing mode is already RIP-relative.
bool isRIPRelative() const {
if (BaseType != RegBase) return false;
if (RegisterSDNode *RegNode =
dyn_cast_or_null<RegisterSDNode>(Base_Reg.getNode()))
return RegNode->getReg() == X86::RIP;
return false;
}
void setBaseReg(SDValue Reg) {
BaseType = RegBase;
Base_Reg = Reg;
}
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
void dump(SelectionDAG *DAG = nullptr) {
dbgs() << "X86ISelAddressMode " << this << '\n';
dbgs() << "Base_Reg ";
if (Base_Reg.getNode())
Base_Reg.getNode()->dump(DAG);
else
dbgs() << "nul\n";
if (BaseType == FrameIndexBase)
dbgs() << " Base.FrameIndex " << Base_FrameIndex << '\n';
dbgs() << " Scale " << Scale << '\n'
<< "IndexReg ";
if (IndexReg.getNode())
IndexReg.getNode()->dump(DAG);
else
dbgs() << "nul\n";
dbgs() << " Disp " << Disp << '\n'
<< "GV ";
if (GV)
GV->dump();
else
dbgs() << "nul";
dbgs() << " CP ";
if (CP)
CP->dump();
else
dbgs() << "nul";
dbgs() << '\n'
<< "ES ";
if (ES)
dbgs() << ES;
else
dbgs() << "nul";
dbgs() << " MCSym ";
if (MCSym)
dbgs() << MCSym;
else
dbgs() << "nul";
dbgs() << " JT" << JT << " Align" << Align << '\n';
}
#endif
};
}
namespace {
//===--------------------------------------------------------------------===//
/// ISel - X86-specific code to select X86 machine instructions for
/// SelectionDAG operations.
///
class X86DAGToDAGISel final : public SelectionDAGISel {
/// Keep a pointer to the X86Subtarget around so that we can
/// make the right decision when generating code for different targets.
const X86Subtarget *Subtarget;
/// If true, selector should try to optimize for code size instead of
/// performance.
bool OptForSize;
/// If true, selector should try to optimize for minimum code size.
bool OptForMinSize;
/// Disable direct TLS access through segment registers.
bool IndirectTlsSegRefs;
public:
explicit X86DAGToDAGISel(X86TargetMachine &tm, CodeGenOpt::Level OptLevel)
: SelectionDAGISel(tm, OptLevel), OptForSize(false),
OptForMinSize(false) {}
StringRef getPassName() const override {
return "X86 DAG->DAG Instruction Selection";
}
bool runOnMachineFunction(MachineFunction &MF) override {
// Reset the subtarget each time through.
Subtarget = &MF.getSubtarget<X86Subtarget>();
IndirectTlsSegRefs = MF.getFunction().hasFnAttribute(
"indirect-tls-seg-refs");
SelectionDAGISel::runOnMachineFunction(MF);
return true;
}
void EmitFunctionEntryCode() override;
bool IsProfitableToFold(SDValue N, SDNode *U, SDNode *Root) const override;
void PreprocessISelDAG() override;
void PostprocessISelDAG() override;
// Include the pieces autogenerated from the target description.
#include "X86GenDAGISel.inc"
private:
void Select(SDNode *N) override;
bool foldOffsetIntoAddress(uint64_t Offset, X86ISelAddressMode &AM);
bool matchLoadInAddress(LoadSDNode *N, X86ISelAddressMode &AM);
bool matchWrapper(SDValue N, X86ISelAddressMode &AM);
bool matchAddress(SDValue N, X86ISelAddressMode &AM);
bool matchVectorAddress(SDValue N, X86ISelAddressMode &AM);
bool matchAdd(SDValue N, X86ISelAddressMode &AM, unsigned Depth);
bool matchAddressRecursively(SDValue N, X86ISelAddressMode &AM,
unsigned Depth);
bool matchAddressBase(SDValue N, X86ISelAddressMode &AM);
bool selectAddr(SDNode *Parent, SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index, SDValue &Disp,
SDValue &Segment);
bool selectVectorAddr(SDNode *Parent, SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index, SDValue &Disp,
SDValue &Segment);
bool selectMOV64Imm32(SDValue N, SDValue &Imm);
bool selectLEAAddr(SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index, SDValue &Disp,
SDValue &Segment);
bool selectLEA64_32Addr(SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index, SDValue &Disp,
SDValue &Segment);
bool selectTLSADDRAddr(SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index, SDValue &Disp,
SDValue &Segment);
bool selectScalarSSELoad(SDNode *Root, SDNode *Parent, SDValue N,
SDValue &Base, SDValue &Scale,
SDValue &Index, SDValue &Disp,
SDValue &Segment,
SDValue &NodeWithChain);
bool selectRelocImm(SDValue N, SDValue &Op);
bool tryFoldLoad(SDNode *Root, SDNode *P, SDValue N,
SDValue &Base, SDValue &Scale,
SDValue &Index, SDValue &Disp,
SDValue &Segment);
// Convenience method where P is also root.
bool tryFoldLoad(SDNode *P, SDValue N,
SDValue &Base, SDValue &Scale,
SDValue &Index, SDValue &Disp,
SDValue &Segment) {
return tryFoldLoad(P, P, N, Base, Scale, Index, Disp, Segment);
}
/// Implement addressing mode selection for inline asm expressions.
bool SelectInlineAsmMemoryOperand(const SDValue &Op,
unsigned ConstraintID,
std::vector<SDValue> &OutOps) override;
void emitSpecialCodeForMain();
inline void getAddressOperands(X86ISelAddressMode &AM, const SDLoc &DL,
SDValue &Base, SDValue &Scale,
SDValue &Index, SDValue &Disp,
SDValue &Segment) {
Base = (AM.BaseType == X86ISelAddressMode::FrameIndexBase)
? CurDAG->getTargetFrameIndex(
AM.Base_FrameIndex,
TLI->getPointerTy(CurDAG->getDataLayout()))
: AM.Base_Reg;
Scale = getI8Imm(AM.Scale, DL);
Index = AM.IndexReg;
// These are 32-bit even in 64-bit mode since RIP-relative offset
// is 32-bit.
if (AM.GV)
Disp = CurDAG->getTargetGlobalAddress(AM.GV, SDLoc(),
MVT::i32, AM.Disp,
AM.SymbolFlags);
else if (AM.CP)
Disp = CurDAG->getTargetConstantPool(AM.CP, MVT::i32,
AM.Align, AM.Disp, AM.SymbolFlags);
else if (AM.ES) {
assert(!AM.Disp && "Non-zero displacement is ignored with ES.");
Disp = CurDAG->getTargetExternalSymbol(AM.ES, MVT::i32, AM.SymbolFlags);
} else if (AM.MCSym) {
assert(!AM.Disp && "Non-zero displacement is ignored with MCSym.");
assert(AM.SymbolFlags == 0 && "oo");
Disp = CurDAG->getMCSymbol(AM.MCSym, MVT::i32);
} else if (AM.JT != -1) {
assert(!AM.Disp && "Non-zero displacement is ignored with JT.");
Disp = CurDAG->getTargetJumpTable(AM.JT, MVT::i32, AM.SymbolFlags);
} else if (AM.BlockAddr)
Disp = CurDAG->getTargetBlockAddress(AM.BlockAddr, MVT::i32, AM.Disp,
AM.SymbolFlags);
else
Disp = CurDAG->getTargetConstant(AM.Disp, DL, MVT::i32);
if (AM.Segment.getNode())
Segment = AM.Segment;
else
Segment = CurDAG->getRegister(0, MVT::i32);
}
// Utility function to determine whether we should avoid selecting
// immediate forms of instructions for better code size or not.
// At a high level, we'd like to avoid such instructions when
// we have similar constants used within the same basic block
// that can be kept in a register.
//
bool shouldAvoidImmediateInstFormsForSize(SDNode *N) const {
uint32_t UseCount = 0;
// Do not want to hoist if we're not optimizing for size.
// TODO: We'd like to remove this restriction.
// See the comment in X86InstrInfo.td for more info.
if (!OptForSize)
return false;
// Walk all the users of the immediate.
for (SDNode::use_iterator UI = N->use_begin(),
UE = N->use_end(); (UI != UE) && (UseCount < 2); ++UI) {
SDNode *User = *UI;
// This user is already selected. Count it as a legitimate use and
// move on.
if (User->isMachineOpcode()) {
UseCount++;
continue;
}
// We want to count stores of immediates as real uses.
if (User->getOpcode() == ISD::STORE &&
User->getOperand(1).getNode() == N) {
UseCount++;
continue;
}
// We don't currently match users that have > 2 operands (except
// for stores, which are handled above)
// Those instruction won't match in ISEL, for now, and would
// be counted incorrectly.
// This may change in the future as we add additional instruction
// types.
if (User->getNumOperands() != 2)
continue;
// Immediates that are used for offsets as part of stack
// manipulation should be left alone. These are typically
// used to indicate SP offsets for argument passing and
// will get pulled into stores/pushes (implicitly).
if (User->getOpcode() == X86ISD::ADD ||
User->getOpcode() == ISD::ADD ||
User->getOpcode() == X86ISD::SUB ||
User->getOpcode() == ISD::SUB) {
// Find the other operand of the add/sub.
SDValue OtherOp = User->getOperand(0);
if (OtherOp.getNode() == N)
OtherOp = User->getOperand(1);
// Don't count if the other operand is SP.
RegisterSDNode *RegNode;
if (OtherOp->getOpcode() == ISD::CopyFromReg &&
(RegNode = dyn_cast_or_null<RegisterSDNode>(
OtherOp->getOperand(1).getNode())))
if ((RegNode->getReg() == X86::ESP) ||
(RegNode->getReg() == X86::RSP))
continue;
}
// ... otherwise, count this and move on.
UseCount++;
}
// If we have more than 1 use, then recommend for hoisting.
return (UseCount > 1);
}
/// Return a target constant with the specified value of type i8.
inline SDValue getI8Imm(unsigned Imm, const SDLoc &DL) {
return CurDAG->getTargetConstant(Imm, DL, MVT::i8);
}
/// Return a target constant with the specified value, of type i32.
inline SDValue getI32Imm(unsigned Imm, const SDLoc &DL) {
return CurDAG->getTargetConstant(Imm, DL, MVT::i32);
}
/// Return a target constant with the specified value, of type i64.
inline SDValue getI64Imm(uint64_t Imm, const SDLoc &DL) {
return CurDAG->getTargetConstant(Imm, DL, MVT::i64);
}
SDValue getExtractVEXTRACTImmediate(SDNode *N, unsigned VecWidth,
const SDLoc &DL) {
assert((VecWidth == 128 || VecWidth == 256) && "Unexpected vector width");
uint64_t Index = N->getConstantOperandVal(1);
MVT VecVT = N->getOperand(0).getSimpleValueType();
return getI8Imm((Index * VecVT.getScalarSizeInBits()) / VecWidth, DL);
}
SDValue getInsertVINSERTImmediate(SDNode *N, unsigned VecWidth,
const SDLoc &DL) {
assert((VecWidth == 128 || VecWidth == 256) && "Unexpected vector width");
uint64_t Index = N->getConstantOperandVal(2);
MVT VecVT = N->getSimpleValueType(0);
return getI8Imm((Index * VecVT.getScalarSizeInBits()) / VecWidth, DL);
}
/// Return an SDNode that returns the value of the global base register.
/// Output instructions required to initialize the global base register,
/// if necessary.
SDNode *getGlobalBaseReg();
/// Return a reference to the TargetMachine, casted to the target-specific
/// type.
const X86TargetMachine &getTargetMachine() const {
return static_cast<const X86TargetMachine &>(TM);
}
/// Return a reference to the TargetInstrInfo, casted to the target-specific
/// type.
const X86InstrInfo *getInstrInfo() const {
return Subtarget->getInstrInfo();
}
/// Address-mode matching performs shift-of-and to and-of-shift
/// reassociation in order to expose more scaled addressing
/// opportunities.
bool ComplexPatternFuncMutatesDAG() const override {
return true;
}
bool isSExtAbsoluteSymbolRef(unsigned Width, SDNode *N) const;
/// Returns whether this is a relocatable immediate in the range
/// [-2^Width .. 2^Width-1].
template <unsigned Width> bool isSExtRelocImm(SDNode *N) const {
if (auto *CN = dyn_cast<ConstantSDNode>(N))
return isInt<Width>(CN->getSExtValue());
return isSExtAbsoluteSymbolRef(Width, N);
}
// Indicates we should prefer to use a non-temporal load for this load.
bool useNonTemporalLoad(LoadSDNode *N) const {
if (!N->isNonTemporal())
return false;
unsigned StoreSize = N->getMemoryVT().getStoreSize();
if (N->getAlignment() < StoreSize)
return false;
switch (StoreSize) {
default: llvm_unreachable("Unsupported store size");
case 4:
case 8:
return false;
case 16:
return Subtarget->hasSSE41();
case 32:
return Subtarget->hasAVX2();
case 64:
return Subtarget->hasAVX512();
}
}
bool foldLoadStoreIntoMemOperand(SDNode *Node);
MachineSDNode *matchBEXTRFromAndImm(SDNode *Node);
bool matchBitExtract(SDNode *Node);
bool shrinkAndImmediate(SDNode *N);
bool isMaskZeroExtended(SDNode *N) const;
bool tryShiftAmountMod(SDNode *N);
MachineSDNode *emitPCMPISTR(unsigned ROpc, unsigned MOpc, bool MayFoldLoad,
const SDLoc &dl, MVT VT, SDNode *Node);
MachineSDNode *emitPCMPESTR(unsigned ROpc, unsigned MOpc, bool MayFoldLoad,
const SDLoc &dl, MVT VT, SDNode *Node,
SDValue &InFlag);
bool tryOptimizeRem8Extend(SDNode *N);
};
}
// Returns true if this masked compare can be implemented legally with this
// type.
static bool isLegalMaskCompare(SDNode *N, const X86Subtarget *Subtarget) {
unsigned Opcode = N->getOpcode();
if (Opcode == X86ISD::CMPM || Opcode == ISD::SETCC ||
Opcode == X86ISD::CMPM_RND || Opcode == X86ISD::VFPCLASS) {
// We can get 256-bit 8 element types here without VLX being enabled. When
// this happens we will use 512-bit operations and the mask will not be
// zero extended.
EVT OpVT = N->getOperand(0).getValueType();
if (OpVT.is256BitVector() || OpVT.is128BitVector())
return Subtarget->hasVLX();
return true;
}
// Scalar opcodes use 128 bit registers, but aren't subject to the VLX check.
if (Opcode == X86ISD::VFPCLASSS || Opcode == X86ISD::FSETCCM ||
Opcode == X86ISD::FSETCCM_RND)
return true;
return false;
}
// Returns true if we can assume the writer of the mask has zero extended it
// for us.
bool X86DAGToDAGISel::isMaskZeroExtended(SDNode *N) const {
// If this is an AND, check if we have a compare on either side. As long as
// one side guarantees the mask is zero extended, the AND will preserve those
// zeros.
if (N->getOpcode() == ISD::AND)
return isLegalMaskCompare(N->getOperand(0).getNode(), Subtarget) ||
isLegalMaskCompare(N->getOperand(1).getNode(), Subtarget);
return isLegalMaskCompare(N, Subtarget);
}
bool
X86DAGToDAGISel::IsProfitableToFold(SDValue N, SDNode *U, SDNode *Root) const {
if (OptLevel == CodeGenOpt::None) return false;
if (!N.hasOneUse())
return false;
if (N.getOpcode() != ISD::LOAD)
return true;
// Don't fold non-temporal loads if we have an instruction for them.
if (useNonTemporalLoad(cast<LoadSDNode>(N)))
return false;
// If N is a load, do additional profitability checks.
if (U == Root) {
switch (U->getOpcode()) {
default: break;
case X86ISD::ADD:
case X86ISD::ADC:
case X86ISD::SUB:
case X86ISD::SBB:
case X86ISD::AND:
case X86ISD::XOR:
case X86ISD::OR:
case ISD::ADD:
case ISD::ADDCARRY:
case ISD::AND:
case ISD::OR:
case ISD::XOR: {
SDValue Op1 = U->getOperand(1);
// If the other operand is a 8-bit immediate we should fold the immediate
// instead. This reduces code size.
// e.g.
// movl 4(%esp), %eax
// addl $4, %eax
// vs.
// movl $4, %eax
// addl 4(%esp), %eax
// The former is 2 bytes shorter. In case where the increment is 1, then
// the saving can be 4 bytes (by using incl %eax).
if (ConstantSDNode *Imm = dyn_cast<ConstantSDNode>(Op1)) {
if (Imm->getAPIntValue().isSignedIntN(8))
return false;
// If this is a 64-bit AND with an immediate that fits in 32-bits,
// prefer using the smaller and over folding the load. This is needed to
// make sure immediates created by shrinkAndImmediate are always folded.
// Ideally we would narrow the load during DAG combine and get the
// best of both worlds.
if (U->getOpcode() == ISD::AND &&
Imm->getAPIntValue().getBitWidth() == 64 &&
Imm->getAPIntValue().isIntN(32))
return false;
}
// If the other operand is a TLS address, we should fold it instead.
// This produces
// movl %gs:0, %eax
// leal i@NTPOFF(%eax), %eax
// instead of
// movl $i@NTPOFF, %eax
// addl %gs:0, %eax
// if the block also has an access to a second TLS address this will save
// a load.
// FIXME: This is probably also true for non-TLS addresses.
if (Op1.getOpcode() == X86ISD::Wrapper) {
SDValue Val = Op1.getOperand(0);
if (Val.getOpcode() == ISD::TargetGlobalTLSAddress)
return false;
}
// Don't fold load if this matches the BTS/BTR/BTC patterns.
// BTS: (or X, (shl 1, n))
// BTR: (and X, (rotl -2, n))
// BTC: (xor X, (shl 1, n))
if (U->getOpcode() == ISD::OR || U->getOpcode() == ISD::XOR) {
if (U->getOperand(0).getOpcode() == ISD::SHL &&
isOneConstant(U->getOperand(0).getOperand(0)))
return false;
if (U->getOperand(1).getOpcode() == ISD::SHL &&
isOneConstant(U->getOperand(1).getOperand(0)))
return false;
}
if (U->getOpcode() == ISD::AND) {
SDValue U0 = U->getOperand(0);
SDValue U1 = U->getOperand(1);
if (U0.getOpcode() == ISD::ROTL) {
auto *C = dyn_cast<ConstantSDNode>(U0.getOperand(0));
if (C && C->getSExtValue() == -2)
return false;
}
if (U1.getOpcode() == ISD::ROTL) {
auto *C = dyn_cast<ConstantSDNode>(U1.getOperand(0));
if (C && C->getSExtValue() == -2)
return false;
}
}
break;
}
case ISD::SHL:
case ISD::SRA:
case ISD::SRL:
// Don't fold a load into a shift by immediate. The BMI2 instructions
// support folding a load, but not an immediate. The legacy instructions
// support folding an immediate, but can't fold a load. Folding an
// immediate is preferable to folding a load.
if (isa<ConstantSDNode>(U->getOperand(1)))
return false;
break;
}
}
// Prevent folding a load if this can implemented with an insert_subreg or
// a move that implicitly zeroes.
if (Root->getOpcode() == ISD::INSERT_SUBVECTOR &&
isNullConstant(Root->getOperand(2)) &&
(Root->getOperand(0).isUndef() ||
ISD::isBuildVectorAllZeros(Root->getOperand(0).getNode())))
return false;
return true;
}
/// Replace the original chain operand of the call with
/// load's chain operand and move load below the call's chain operand.
static void moveBelowOrigChain(SelectionDAG *CurDAG, SDValue Load,
SDValue Call, SDValue OrigChain) {
SmallVector<SDValue, 8> Ops;
SDValue Chain = OrigChain.getOperand(0);
if (Chain.getNode() == Load.getNode())
Ops.push_back(Load.getOperand(0));
else {
assert(Chain.getOpcode() == ISD::TokenFactor &&
"Unexpected chain operand");
for (unsigned i = 0, e = Chain.getNumOperands(); i != e; ++i)
if (Chain.getOperand(i).getNode() == Load.getNode())
Ops.push_back(Load.getOperand(0));
else
Ops.push_back(Chain.getOperand(i));
SDValue NewChain =
CurDAG->getNode(ISD::TokenFactor, SDLoc(Load), MVT::Other, Ops);
Ops.clear();
Ops.push_back(NewChain);
}
Ops.append(OrigChain->op_begin() + 1, OrigChain->op_end());
CurDAG->UpdateNodeOperands(OrigChain.getNode(), Ops);
CurDAG->UpdateNodeOperands(Load.getNode(), Call.getOperand(0),
Load.getOperand(1), Load.getOperand(2));
Ops.clear();
Ops.push_back(SDValue(Load.getNode(), 1));
Ops.append(Call->op_begin() + 1, Call->op_end());
CurDAG->UpdateNodeOperands(Call.getNode(), Ops);
}
/// Return true if call address is a load and it can be
/// moved below CALLSEQ_START and the chains leading up to the call.
/// Return the CALLSEQ_START by reference as a second output.
/// In the case of a tail call, there isn't a callseq node between the call
/// chain and the load.
static bool isCalleeLoad(SDValue Callee, SDValue &Chain, bool HasCallSeq) {
// The transformation is somewhat dangerous if the call's chain was glued to
// the call. After MoveBelowOrigChain the load is moved between the call and
// the chain, this can create a cycle if the load is not folded. So it is
// *really* important that we are sure the load will be folded.
if (Callee.getNode() == Chain.getNode() || !Callee.hasOneUse())
return false;
LoadSDNode *LD = dyn_cast<LoadSDNode>(Callee.getNode());
if (!LD ||
LD->isVolatile() ||
LD->getAddressingMode() != ISD::UNINDEXED ||
LD->getExtensionType() != ISD::NON_EXTLOAD)
return false;
// Now let's find the callseq_start.
while (HasCallSeq && Chain.getOpcode() != ISD::CALLSEQ_START) {
if (!Chain.hasOneUse())
return false;
Chain = Chain.getOperand(0);
}
if (!Chain.getNumOperands())
return false;
// Since we are not checking for AA here, conservatively abort if the chain
// writes to memory. It's not safe to move the callee (a load) across a store.
if (isa<MemSDNode>(Chain.getNode()) &&
cast<MemSDNode>(Chain.getNode())->writeMem())
return false;
if (Chain.getOperand(0).getNode() == Callee.getNode())
return true;
if (Chain.getOperand(0).getOpcode() == ISD::TokenFactor &&
Callee.getValue(1).isOperandOf(Chain.getOperand(0).getNode()) &&
Callee.getValue(1).hasOneUse())
return true;
return false;
}
void X86DAGToDAGISel::PreprocessISelDAG() {
// OptFor[Min]Size are used in pattern predicates that isel is matching.
OptForSize = MF->getFunction().optForSize();
OptForMinSize = MF->getFunction().optForMinSize();
assert((!OptForMinSize || OptForSize) && "OptForMinSize implies OptForSize");
for (SelectionDAG::allnodes_iterator I = CurDAG->allnodes_begin(),
E = CurDAG->allnodes_end(); I != E; ) {
SDNode *N = &*I++; // Preincrement iterator to avoid invalidation issues.
// If this is a target specific AND node with no flag usages, turn it back
// into ISD::AND to enable test instruction matching.
if (N->getOpcode() == X86ISD::AND && !N->hasAnyUseOfValue(1)) {
SDValue Res = CurDAG->getNode(ISD::AND, SDLoc(N), N->getValueType(0),
N->getOperand(0), N->getOperand(1));
--I;
CurDAG->ReplaceAllUsesOfValueWith(SDValue(N, 0), Res);
++I;
CurDAG->DeleteNode(N);
continue;
}
if (OptLevel != CodeGenOpt::None &&
// Only do this when the target can fold the load into the call or
// jmp.
!Subtarget->useRetpolineIndirectCalls() &&
((N->getOpcode() == X86ISD::CALL && !Subtarget->slowTwoMemOps()) ||
(N->getOpcode() == X86ISD::TC_RETURN &&
(Subtarget->is64Bit() ||
!getTargetMachine().isPositionIndependent())))) {
/// Also try moving call address load from outside callseq_start to just
/// before the call to allow it to be folded.
///
/// [Load chain]
/// ^
/// |
/// [Load]
/// ^ ^
/// | |
/// / \--
/// / |
///[CALLSEQ_START] |
/// ^ |
/// | |
/// [LOAD/C2Reg] |
/// | |
/// \ /
/// \ /
/// [CALL]
bool HasCallSeq = N->getOpcode() == X86ISD::CALL;
SDValue Chain = N->getOperand(0);
SDValue Load = N->getOperand(1);
if (!isCalleeLoad(Load, Chain, HasCallSeq))
continue;
moveBelowOrigChain(CurDAG, Load, SDValue(N, 0), Chain);
++NumLoadMoved;
continue;
}
// Lower fpround and fpextend nodes that target the FP stack to be store and
// load to the stack. This is a gross hack. We would like to simply mark
// these as being illegal, but when we do that, legalize produces these when
// it expands calls, then expands these in the same legalize pass. We would
// like dag combine to be able to hack on these between the call expansion
// and the node legalization. As such this pass basically does "really
// late" legalization of these inline with the X86 isel pass.
// FIXME: This should only happen when not compiled with -O0.
if (N->getOpcode() != ISD::FP_ROUND && N->getOpcode() != ISD::FP_EXTEND)
continue;
MVT SrcVT = N->getOperand(0).getSimpleValueType();
MVT DstVT = N->getSimpleValueType(0);
// If any of the sources are vectors, no fp stack involved.
if (SrcVT.isVector() || DstVT.isVector())
continue;
// If the source and destination are SSE registers, then this is a legal
// conversion that should not be lowered.
const X86TargetLowering *X86Lowering =
static_cast<const X86TargetLowering *>(TLI);
bool SrcIsSSE = X86Lowering->isScalarFPTypeInSSEReg(SrcVT);
bool DstIsSSE = X86Lowering->isScalarFPTypeInSSEReg(DstVT);
if (SrcIsSSE && DstIsSSE)
continue;
if (!SrcIsSSE && !DstIsSSE) {
// If this is an FPStack extension, it is a noop.
if (N->getOpcode() == ISD::FP_EXTEND)
continue;
// If this is a value-preserving FPStack truncation, it is a noop.
if (N->getConstantOperandVal(1))
continue;
}
// Here we could have an FP stack truncation or an FPStack <-> SSE convert.
// FPStack has extload and truncstore. SSE can fold direct loads into other
// operations. Based on this, decide what we want to do.
MVT MemVT;
if (N->getOpcode() == ISD::FP_ROUND)
MemVT = DstVT; // FP_ROUND must use DstVT, we can't do a 'trunc load'.
else
MemVT = SrcIsSSE ? SrcVT : DstVT;
SDValue MemTmp = CurDAG->CreateStackTemporary(MemVT);
SDLoc dl(N);
// FIXME: optimize the case where the src/dest is a load or store?
SDValue Store =
CurDAG->getTruncStore(CurDAG->getEntryNode(), dl, N->getOperand(0),
MemTmp, MachinePointerInfo(), MemVT);
SDValue Result = CurDAG->getExtLoad(ISD::EXTLOAD, dl, DstVT, Store, MemTmp,
MachinePointerInfo(), MemVT);
// We're about to replace all uses of the FP_ROUND/FP_EXTEND with the
// extload we created. This will cause general havok on the dag because
// anything below the conversion could be folded into other existing nodes.
// To avoid invalidating 'I', back it up to the convert node.
--I;
CurDAG->ReplaceAllUsesOfValueWith(SDValue(N, 0), Result);
// Now that we did that, the node is dead. Increment the iterator to the
// next node to process, then delete N.
++I;
CurDAG->DeleteNode(N);
}
}
// Look for a redundant movzx/movsx that can occur after an 8-bit divrem.
bool X86DAGToDAGISel::tryOptimizeRem8Extend(SDNode *N) {
unsigned Opc = N->getMachineOpcode();
if (Opc != X86::MOVZX32rr8 && Opc != X86::MOVSX32rr8 &&
Opc != X86::MOVSX64rr8)
return false;
SDValue N0 = N->getOperand(0);
// We need to be extracting the lower bit of an extend.
if (!N0.isMachineOpcode() ||
N0.getMachineOpcode() != TargetOpcode::EXTRACT_SUBREG ||
N0.getConstantOperandVal(1) != X86::sub_8bit)
return false;
// We're looking for either a movsx or movzx to match the original opcode.
unsigned ExpectedOpc = Opc == X86::MOVZX32rr8 ? X86::MOVZX32rr8_NOREX
: X86::MOVSX32rr8_NOREX;
SDValue N00 = N0.getOperand(0);
if (!N00.isMachineOpcode() || N00.getMachineOpcode() != ExpectedOpc)
return false;
if (Opc == X86::MOVSX64rr8) {
// If we had a sign extend from 8 to 64 bits. We still need to go from 32
// to 64.
MachineSDNode *Extend = CurDAG->getMachineNode(X86::MOVSX64rr32, SDLoc(N),
MVT::i64, N00);
ReplaceUses(N, Extend);
} else {
// Ok we can drop this extend and just use the original extend.
ReplaceUses(N, N00.getNode());
}
return true;
}
void X86DAGToDAGISel::PostprocessISelDAG() {
// Skip peepholes at -O0.
if (TM.getOptLevel() == CodeGenOpt::None)
return;
SelectionDAG::allnodes_iterator Position = CurDAG->allnodes_end();
bool MadeChange = false;
while (Position != CurDAG->allnodes_begin()) {
SDNode *N = &*--Position;
// Skip dead nodes and any non-machine opcodes.
if (N->use_empty() || !N->isMachineOpcode())
continue;
if (tryOptimizeRem8Extend(N)) {
MadeChange = true;
continue;
}
// Attempt to remove vectors moves that were inserted to zero upper bits.
if (N->getMachineOpcode() != TargetOpcode::SUBREG_TO_REG)
continue;
unsigned SubRegIdx = N->getConstantOperandVal(2);
if (SubRegIdx != X86::sub_xmm && SubRegIdx != X86::sub_ymm)
continue;
SDValue Move = N->getOperand(1);
if (!Move.isMachineOpcode())
continue;
// Make sure its one of the move opcodes we recognize.
switch (Move.getMachineOpcode()) {
default:
continue;
case X86::VMOVAPDrr: case X86::VMOVUPDrr:
case X86::VMOVAPSrr: case X86::VMOVUPSrr:
case X86::VMOVDQArr: case X86::VMOVDQUrr:
case X86::VMOVAPDYrr: case X86::VMOVUPDYrr:
case X86::VMOVAPSYrr: case X86::VMOVUPSYrr:
case X86::VMOVDQAYrr: case X86::VMOVDQUYrr:
case X86::VMOVAPDZ128rr: case X86::VMOVUPDZ128rr:
case X86::VMOVAPSZ128rr: case X86::VMOVUPSZ128rr:
case X86::VMOVDQA32Z128rr: case X86::VMOVDQU32Z128rr:
case X86::VMOVDQA64Z128rr: case X86::VMOVDQU64Z128rr:
case X86::VMOVAPDZ256rr: case X86::VMOVUPDZ256rr:
case X86::VMOVAPSZ256rr: case X86::VMOVUPSZ256rr:
case X86::VMOVDQA32Z256rr: case X86::VMOVDQU32Z256rr:
case X86::VMOVDQA64Z256rr: case X86::VMOVDQU64Z256rr:
break;
}
SDValue In = Move.getOperand(0);
if (!In.isMachineOpcode() ||
In.getMachineOpcode() <= TargetOpcode::GENERIC_OP_END)
continue;
// Make sure the instruction has a VEX, XOP, or EVEX prefix. This covers
// the SHA instructions which use a legacy encoding.
uint64_t TSFlags = getInstrInfo()->get(In.getMachineOpcode()).TSFlags;
if ((TSFlags & X86II::EncodingMask) != X86II::VEX &&
(TSFlags & X86II::EncodingMask) != X86II::EVEX &&
(TSFlags & X86II::EncodingMask) != X86II::XOP)
continue;
// Producing instruction is another vector instruction. We can drop the
// move.
CurDAG->UpdateNodeOperands(N, N->getOperand(0), In, N->getOperand(2));
MadeChange = true;
}
if (MadeChange)
CurDAG->RemoveDeadNodes();
}
/// Emit any code that needs to be executed only in the main function.
void X86DAGToDAGISel::emitSpecialCodeForMain() {
if (Subtarget->isTargetCygMing()) {
TargetLowering::ArgListTy Args;
auto &DL = CurDAG->getDataLayout();
TargetLowering::CallLoweringInfo CLI(*CurDAG);
CLI.setChain(CurDAG->getRoot())
.setCallee(CallingConv::C, Type::getVoidTy(*CurDAG->getContext()),
CurDAG->getExternalSymbol("__main", TLI->getPointerTy(DL)),
std::move(Args));
const TargetLowering &TLI = CurDAG->getTargetLoweringInfo();
std::pair<SDValue, SDValue> Result = TLI.LowerCallTo(CLI);
CurDAG->setRoot(Result.second);
}
}
void X86DAGToDAGISel::EmitFunctionEntryCode() {
// If this is main, emit special code for main.
const Function &F = MF->getFunction();
if (F.hasExternalLinkage() && F.getName() == "main")
emitSpecialCodeForMain();
}
static bool isDispSafeForFrameIndex(int64_t Val) {
// On 64-bit platforms, we can run into an issue where a frame index
// includes a displacement that, when added to the explicit displacement,
// will overflow the displacement field. Assuming that the frame index
// displacement fits into a 31-bit integer (which is only slightly more
// aggressive than the current fundamental assumption that it fits into
// a 32-bit integer), a 31-bit disp should always be safe.
return isInt<31>(Val);
}
bool X86DAGToDAGISel::foldOffsetIntoAddress(uint64_t Offset,
X86ISelAddressMode &AM) {
// If there's no offset to fold, we don't need to do any work.
if (Offset == 0)
return false;
// Cannot combine ExternalSymbol displacements with integer offsets.
if (AM.ES || AM.MCSym)
return true;
int64_t Val = AM.Disp + Offset;
CodeModel::Model M = TM.getCodeModel();
if (Subtarget->is64Bit()) {
if (!X86::isOffsetSuitableForCodeModel(Val, M,
AM.hasSymbolicDisplacement()))
return true;
// In addition to the checks required for a register base, check that
// we do not try to use an unsafe Disp with a frame index.
if (AM.BaseType == X86ISelAddressMode::FrameIndexBase &&
!isDispSafeForFrameIndex(Val))
return true;
}
AM.Disp = Val;
return false;
}
bool X86DAGToDAGISel::matchLoadInAddress(LoadSDNode *N, X86ISelAddressMode &AM){
SDValue Address = N->getOperand(1);
// load gs:0 -> GS segment register.
// load fs:0 -> FS segment register.
//
// This optimization is valid because the GNU TLS model defines that
// gs:0 (or fs:0 on X86-64) contains its own address.
// For more information see http://people.redhat.com/drepper/tls.pdf
if (ConstantSDNode *C = dyn_cast<ConstantSDNode>(Address))
if (C->getSExtValue() == 0 && AM.Segment.getNode() == nullptr &&
!IndirectTlsSegRefs &&
(Subtarget->isTargetGlibc() || Subtarget->isTargetAndroid() ||
Subtarget->isTargetFuchsia()))
switch (N->getPointerInfo().getAddrSpace()) {
case 256:
AM.Segment = CurDAG->getRegister(X86::GS, MVT::i16);
return false;
case 257:
AM.Segment = CurDAG->getRegister(X86::FS, MVT::i16);
return false;
// Address space 258 is not handled here, because it is not used to
// address TLS areas.
}
return true;
}
/// Try to match X86ISD::Wrapper and X86ISD::WrapperRIP nodes into an addressing
/// mode. These wrap things that will resolve down into a symbol reference.
/// If no match is possible, this returns true, otherwise it returns false.
bool X86DAGToDAGISel::matchWrapper(SDValue N, X86ISelAddressMode &AM) {
// If the addressing mode already has a symbol as the displacement, we can
// never match another symbol.
if (AM.hasSymbolicDisplacement())
return true;
bool IsRIPRel = N.getOpcode() == X86ISD::WrapperRIP;
// We can't use an addressing mode in the 64-bit large code model. In the
// medium code model, we use can use an mode when RIP wrappers are present.
// That signifies access to globals that are known to be "near", such as the
// GOT itself.
CodeModel::Model M = TM.getCodeModel();
if (Subtarget->is64Bit() &&
(M == CodeModel::Large || (M == CodeModel::Medium && !IsRIPRel)))
return true;
// Base and index reg must be 0 in order to use %rip as base.
if (IsRIPRel && AM.hasBaseOrIndexReg())
return true;
// Make a local copy in case we can't do this fold.
X86ISelAddressMode Backup = AM;
int64_t Offset = 0;
SDValue N0 = N.getOperand(0);
if (GlobalAddressSDNode *G = dyn_cast<GlobalAddressSDNode>(N0)) {
AM.GV = G->getGlobal();
AM.SymbolFlags = G->getTargetFlags();
Offset = G->getOffset();
} else if (ConstantPoolSDNode *CP = dyn_cast<ConstantPoolSDNode>(N0)) {
AM.CP = CP->getConstVal();
AM.Align = CP->getAlignment();
AM.SymbolFlags = CP->getTargetFlags();
Offset = CP->getOffset();
} else if (ExternalSymbolSDNode *S = dyn_cast<ExternalSymbolSDNode>(N0)) {
AM.ES = S->getSymbol();
AM.SymbolFlags = S->getTargetFlags();
} else if (auto *S = dyn_cast<MCSymbolSDNode>(N0)) {
AM.MCSym = S->getMCSymbol();
} else if (JumpTableSDNode *J = dyn_cast<JumpTableSDNode>(N0)) {
AM.JT = J->getIndex();
AM.SymbolFlags = J->getTargetFlags();
} else if (BlockAddressSDNode *BA = dyn_cast<BlockAddressSDNode>(N0)) {
AM.BlockAddr = BA->getBlockAddress();
AM.SymbolFlags = BA->getTargetFlags();
Offset = BA->getOffset();
} else
llvm_unreachable("Unhandled symbol reference node.");
if (foldOffsetIntoAddress(Offset, AM)) {
AM = Backup;
return true;
}
if (IsRIPRel)
AM.setBaseReg(CurDAG->getRegister(X86::RIP, MVT::i64));
// Commit the changes now that we know this fold is safe.
return false;
}
/// Add the specified node to the specified addressing mode, returning true if
/// it cannot be done. This just pattern matches for the addressing mode.
bool X86DAGToDAGISel::matchAddress(SDValue N, X86ISelAddressMode &AM) {
if (matchAddressRecursively(N, AM, 0))
return true;
// Post-processing: Convert lea(,%reg,2) to lea(%reg,%reg), which has
// a smaller encoding and avoids a scaled-index.
if (AM.Scale == 2 &&
AM.BaseType == X86ISelAddressMode::RegBase &&
AM.Base_Reg.getNode() == nullptr) {
AM.Base_Reg = AM.IndexReg;
AM.Scale = 1;
}
// Post-processing: Convert foo to foo(%rip), even in non-PIC mode,
// because it has a smaller encoding.
// TODO: Which other code models can use this?
if (TM.getCodeModel() == CodeModel::Small &&
Subtarget->is64Bit() &&
AM.Scale == 1 &&
AM.BaseType == X86ISelAddressMode::RegBase &&
AM.Base_Reg.getNode() == nullptr &&
AM.IndexReg.getNode() == nullptr &&
AM.SymbolFlags == X86II::MO_NO_FLAG &&
AM.hasSymbolicDisplacement())
AM.Base_Reg = CurDAG->getRegister(X86::RIP, MVT::i64);
return false;
}
bool X86DAGToDAGISel::matchAdd(SDValue N, X86ISelAddressMode &AM,
unsigned Depth) {
// Add an artificial use to this node so that we can keep track of
// it if it gets CSE'd with a different node.
HandleSDNode Handle(N);
X86ISelAddressMode Backup = AM;
if (!matchAddressRecursively(N.getOperand(0), AM, Depth+1) &&
!matchAddressRecursively(Handle.getValue().getOperand(1), AM, Depth+1))
return false;
AM = Backup;
// Try again after commuting the operands.
if (!matchAddressRecursively(Handle.getValue().getOperand(1), AM, Depth+1) &&
!matchAddressRecursively(Handle.getValue().getOperand(0), AM, Depth+1))
return false;
AM = Backup;
// If we couldn't fold both operands into the address at the same time,
// see if we can just put each operand into a register and fold at least
// the add.
if (AM.BaseType == X86ISelAddressMode::RegBase &&
!AM.Base_Reg.getNode() &&
!AM.IndexReg.getNode()) {
N = Handle.getValue();
AM.Base_Reg = N.getOperand(0);
AM.IndexReg = N.getOperand(1);
AM.Scale = 1;
return false;
}
N = Handle.getValue();
return true;
}
// Insert a node into the DAG at least before the Pos node's position. This
// will reposition the node as needed, and will assign it a node ID that is <=
// the Pos node's ID. Note that this does *not* preserve the uniqueness of node
// IDs! The selection DAG must no longer depend on their uniqueness when this
// is used.
static void insertDAGNode(SelectionDAG &DAG, SDValue Pos, SDValue N) {
if (N->getNodeId() == -1 ||
(SelectionDAGISel::getUninvalidatedNodeId(N.getNode()) >
SelectionDAGISel::getUninvalidatedNodeId(Pos.getNode()))) {
DAG.RepositionNode(Pos->getIterator(), N.getNode());
// Mark Node as invalid for pruning as after this it may be a successor to a
// selected node but otherwise be in the same position of Pos.
// Conservatively mark it with the same -abs(Id) to assure node id
// invariant is preserved.
N->setNodeId(Pos->getNodeId());
SelectionDAGISel::InvalidateNodeId(N.getNode());
}
}
// Transform "(X >> (8-C1)) & (0xff << C1)" to "((X >> 8) & 0xff) << C1" if
// safe. This allows us to convert the shift and and into an h-register
// extract and a scaled index. Returns false if the simplification is
// performed.
static bool foldMaskAndShiftToExtract(SelectionDAG &DAG, SDValue N,
uint64_t Mask,
SDValue Shift, SDValue X,
X86ISelAddressMode &AM) {
if (Shift.getOpcode() != ISD::SRL ||
!isa<ConstantSDNode>(Shift.getOperand(1)) ||
!Shift.hasOneUse())
return true;
int ScaleLog = 8 - Shift.getConstantOperandVal(1);
if (ScaleLog <= 0 || ScaleLog >= 4 ||
Mask != (0xffu << ScaleLog))
return true;
MVT VT = N.getSimpleValueType();
SDLoc DL(N);
SDValue Eight = DAG.getConstant(8, DL, MVT::i8);
SDValue NewMask = DAG.getConstant(0xff, DL, VT);
SDValue Srl = DAG.getNode(ISD::SRL, DL, VT, X, Eight);
SDValue And = DAG.getNode(ISD::AND, DL, VT, Srl, NewMask);
SDValue ShlCount = DAG.getConstant(ScaleLog, DL, MVT::i8);
SDValue Shl = DAG.getNode(ISD::SHL, DL, VT, And, ShlCount);
// Insert the new nodes into the topological ordering. We must do this in
// a valid topological ordering as nothing is going to go back and re-sort
// these nodes. We continually insert before 'N' in sequence as this is
// essentially a pre-flattened and pre-sorted sequence of nodes. There is no
// hierarchy left to express.
insertDAGNode(DAG, N, Eight);
insertDAGNode(DAG, N, Srl);
insertDAGNode(DAG, N, NewMask);
insertDAGNode(DAG, N, And);
insertDAGNode(DAG, N, ShlCount);
insertDAGNode(DAG, N, Shl);
DAG.ReplaceAllUsesWith(N, Shl);
AM.IndexReg = And;
AM.Scale = (1 << ScaleLog);
return false;
}
// Transforms "(X << C1) & C2" to "(X & (C2>>C1)) << C1" if safe and if this
// allows us to fold the shift into this addressing mode. Returns false if the
// transform succeeded.
static bool foldMaskedShiftToScaledMask(SelectionDAG &DAG, SDValue N,
uint64_t Mask,
SDValue Shift, SDValue X,
X86ISelAddressMode &AM) {
if (Shift.getOpcode() != ISD::SHL ||
!isa<ConstantSDNode>(Shift.getOperand(1)))
return true;
// Not likely to be profitable if either the AND or SHIFT node has more
// than one use (unless all uses are for address computation). Besides,
// isel mechanism requires their node ids to be reused.
if (!N.hasOneUse() || !Shift.hasOneUse())
return true;
// Verify that the shift amount is something we can fold.
unsigned ShiftAmt = Shift.getConstantOperandVal(1);
if (ShiftAmt != 1 && ShiftAmt != 2 && ShiftAmt != 3)
return true;
MVT VT = N.getSimpleValueType();
SDLoc DL(N);
SDValue NewMask = DAG.getConstant(Mask >> ShiftAmt, DL, VT);
SDValue NewAnd = DAG.getNode(ISD::AND, DL, VT, X, NewMask);
SDValue NewShift = DAG.getNode(ISD::SHL, DL, VT, NewAnd, Shift.getOperand(1));
// Insert the new nodes into the topological ordering. We must do this in
// a valid topological ordering as nothing is going to go back and re-sort
// these nodes. We continually insert before 'N' in sequence as this is
// essentially a pre-flattened and pre-sorted sequence of nodes. There is no
// hierarchy left to express.
insertDAGNode(DAG, N, NewMask);
insertDAGNode(DAG, N, NewAnd);
insertDAGNode(DAG, N, NewShift);
DAG.ReplaceAllUsesWith(N, NewShift);
AM.Scale = 1 << ShiftAmt;
AM.IndexReg = NewAnd;
return false;
}
// Implement some heroics to detect shifts of masked values where the mask can
// be replaced by extending the shift and undoing that in the addressing mode
// scale. Patterns such as (shl (srl x, c1), c2) are canonicalized into (and
// (srl x, SHIFT), MASK) by DAGCombines that don't know the shl can be done in
// the addressing mode. This results in code such as:
//
// int f(short *y, int *lookup_table) {
// ...
// return *y + lookup_table[*y >> 11];
// }
//
// Turning into:
// movzwl (%rdi), %eax
// movl %eax, %ecx
// shrl $11, %ecx
// addl (%rsi,%rcx,4), %eax
//
// Instead of:
// movzwl (%rdi), %eax
// movl %eax, %ecx
// shrl $9, %ecx
// andl $124, %rcx
// addl (%rsi,%rcx), %eax
//
// Note that this function assumes the mask is provided as a mask *after* the
// value is shifted. The input chain may or may not match that, but computing
// such a mask is trivial.
static bool foldMaskAndShiftToScale(SelectionDAG &DAG, SDValue N,
uint64_t Mask,
SDValue Shift, SDValue X,
X86ISelAddressMode &AM) {
if (Shift.getOpcode() != ISD::SRL || !Shift.hasOneUse() ||
!isa<ConstantSDNode>(Shift.getOperand(1)))
return true;
unsigned ShiftAmt = Shift.getConstantOperandVal(1);
unsigned MaskLZ = countLeadingZeros(Mask);
unsigned MaskTZ = countTrailingZeros(Mask);
// The amount of shift we're trying to fit into the addressing mode is taken
// from the trailing zeros of the mask.
unsigned AMShiftAmt = MaskTZ;
// There is nothing we can do here unless the mask is removing some bits.
// Also, the addressing mode can only represent shifts of 1, 2, or 3 bits.
if (AMShiftAmt <= 0 || AMShiftAmt > 3) return true;
// We also need to ensure that mask is a continuous run of bits.
if (countTrailingOnes(Mask >> MaskTZ) + MaskTZ + MaskLZ != 64) return true;
// Scale the leading zero count down based on the actual size of the value.
// Also scale it down based on the size of the shift.
unsigned ScaleDown = (64 - X.getSimpleValueType().getSizeInBits()) + ShiftAmt;
if (MaskLZ < ScaleDown)
return true;
MaskLZ -= ScaleDown;
// The final check is to ensure that any masked out high bits of X are
// already known to be zero. Otherwise, the mask has a semantic impact
// other than masking out a couple of low bits. Unfortunately, because of
// the mask, zero extensions will be removed from operands in some cases.
// This code works extra hard to look through extensions because we can
// replace them with zero extensions cheaply if necessary.
bool ReplacingAnyExtend = false;
if (X.getOpcode() == ISD::ANY_EXTEND) {
unsigned ExtendBits = X.getSimpleValueType().getSizeInBits() -
X.getOperand(0).getSimpleValueType().getSizeInBits();
// Assume that we'll replace the any-extend with a zero-extend, and
// narrow the search to the extended value.
X = X.getOperand(0);
MaskLZ = ExtendBits > MaskLZ ? 0 : MaskLZ - ExtendBits;
ReplacingAnyExtend = true;
}
APInt MaskedHighBits =
APInt::getHighBitsSet(X.getSimpleValueType().getSizeInBits(), MaskLZ);
KnownBits Known;
DAG.computeKnownBits(X, Known);
if (MaskedHighBits != Known.Zero) return true;
// We've identified a pattern that can be transformed into a single shift
// and an addressing mode. Make it so.
MVT VT = N.getSimpleValueType();
if (ReplacingAnyExtend) {
assert(X.getValueType() != VT);
// We looked through an ANY_EXTEND node, insert a ZERO_EXTEND.
SDValue NewX = DAG.getNode(ISD::ZERO_EXTEND, SDLoc(X), VT, X);
insertDAGNode(DAG, N, NewX);
X = NewX;
}
SDLoc DL(N);
SDValue NewSRLAmt = DAG.getConstant(ShiftAmt + AMShiftAmt, DL, MVT::i8);
SDValue NewSRL = DAG.getNode(ISD::SRL, DL, VT, X, NewSRLAmt);
SDValue NewSHLAmt = DAG.getConstant(AMShiftAmt, DL, MVT::i8);
SDValue NewSHL = DAG.getNode(ISD::SHL, DL, VT, NewSRL, NewSHLAmt);
// Insert the new nodes into the topological ordering. We must do this in
// a valid topological ordering as nothing is going to go back and re-sort
// these nodes. We continually insert before 'N' in sequence as this is
// essentially a pre-flattened and pre-sorted sequence of nodes. There is no
// hierarchy left to express.
insertDAGNode(DAG, N, NewSRLAmt);
insertDAGNode(DAG, N, NewSRL);
insertDAGNode(DAG, N, NewSHLAmt);
insertDAGNode(DAG, N, NewSHL);
DAG.ReplaceAllUsesWith(N, NewSHL);
AM.Scale = 1 << AMShiftAmt;
AM.IndexReg = NewSRL;
return false;
}
// Transform "(X >> SHIFT) & (MASK << C1)" to
// "((X >> (SHIFT + C1)) & (MASK)) << C1". Everything before the SHL will be
// matched to a BEXTR later. Returns false if the simplification is performed.
static bool foldMaskedShiftToBEXTR(SelectionDAG &DAG, SDValue N,
uint64_t Mask,
SDValue Shift, SDValue X,
X86ISelAddressMode &AM,
const X86Subtarget &Subtarget) {
if (Shift.getOpcode() != ISD::SRL ||
!isa<ConstantSDNode>(Shift.getOperand(1)) ||
!Shift.hasOneUse() || !N.hasOneUse())
return true;
// Only do this if BEXTR will be matched by matchBEXTRFromAndImm.
if (!Subtarget.hasTBM() &&
!(Subtarget.hasBMI() && Subtarget.hasFastBEXTR()))
return true;
// We need to ensure that mask is a continuous run of bits.
if (!isShiftedMask_64(Mask)) return true;
unsigned ShiftAmt = Shift.getConstantOperandVal(1);
// The amount of shift we're trying to fit into the addressing mode is taken
// from the trailing zeros of the mask.
unsigned AMShiftAmt = countTrailingZeros(Mask);
// There is nothing we can do here unless the mask is removing some bits.
// Also, the addressing mode can only represent shifts of 1, 2, or 3 bits.
if (AMShiftAmt <= 0 || AMShiftAmt > 3) return true;
MVT VT = N.getSimpleValueType();
SDLoc DL(N);
SDValue NewSRLAmt = DAG.getConstant(ShiftAmt + AMShiftAmt, DL, MVT::i8);
SDValue NewSRL = DAG.getNode(ISD::SRL, DL, VT, X, NewSRLAmt);
SDValue NewMask = DAG.getConstant(Mask >> AMShiftAmt, DL, VT);
SDValue NewAnd = DAG.getNode(ISD::AND, DL, VT, NewSRL, NewMask);
SDValue NewSHLAmt = DAG.getConstant(AMShiftAmt, DL, MVT::i8);
SDValue NewSHL = DAG.getNode(ISD::SHL, DL, VT, NewAnd, NewSHLAmt);
// Insert the new nodes into the topological ordering. We must do this in
// a valid topological ordering as nothing is going to go back and re-sort
// these nodes. We continually insert before 'N' in sequence as this is
// essentially a pre-flattened and pre-sorted sequence of nodes. There is no
// hierarchy left to express.
insertDAGNode(DAG, N, NewSRLAmt);
insertDAGNode(DAG, N, NewSRL);
insertDAGNode(DAG, N, NewMask);
insertDAGNode(DAG, N, NewAnd);
insertDAGNode(DAG, N, NewSHLAmt);
insertDAGNode(DAG, N, NewSHL);
DAG.ReplaceAllUsesWith(N, NewSHL);
AM.Scale = 1 << AMShiftAmt;
AM.IndexReg = NewAnd;
return false;
}
bool X86DAGToDAGISel::matchAddressRecursively(SDValue N, X86ISelAddressMode &AM,
unsigned Depth) {
SDLoc dl(N);
LLVM_DEBUG({
dbgs() << "MatchAddress: ";
AM.dump(CurDAG);
});
// Limit recursion.
if (Depth > 5)
return matchAddressBase(N, AM);
// If this is already a %rip relative address, we can only merge immediates
// into it. Instead of handling this in every case, we handle it here.
// RIP relative addressing: %rip + 32-bit displacement!
if (AM.isRIPRelative()) {
// FIXME: JumpTable and ExternalSymbol address currently don't like
// displacements. It isn't very important, but this should be fixed for
// consistency.
if (!(AM.ES || AM.MCSym) && AM.JT != -1)
return true;
if (ConstantSDNode *Cst = dyn_cast<ConstantSDNode>(N))
if (!foldOffsetIntoAddress(Cst->getSExtValue(), AM))
return false;
return true;
}
switch (N.getOpcode()) {
default: break;
case ISD::LOCAL_RECOVER: {
if (!AM.hasSymbolicDisplacement() && AM.Disp == 0)
if (const auto *ESNode = dyn_cast<MCSymbolSDNode>(N.getOperand(0))) {
// Use the symbol and don't prefix it.
AM.MCSym = ESNode->getMCSymbol();
return false;
}
break;
}
case ISD::Constant: {
uint64_t Val = cast<ConstantSDNode>(N)->getSExtValue();
if (!foldOffsetIntoAddress(Val, AM))
return false;
break;
}
case X86ISD::Wrapper:
case X86ISD::WrapperRIP:
if (!matchWrapper(N, AM))
return false;
break;
case ISD::LOAD:
if (!matchLoadInAddress(cast<LoadSDNode>(N), AM))
return false;
break;
case ISD::FrameIndex:
if (AM.BaseType == X86ISelAddressMode::RegBase &&
AM.Base_Reg.getNode() == nullptr &&
(!Subtarget->is64Bit() || isDispSafeForFrameIndex(AM.Disp))) {
AM.BaseType = X86ISelAddressMode::FrameIndexBase;
AM.Base_FrameIndex = cast<FrameIndexSDNode>(N)->getIndex();
return false;
}
break;
case ISD::SHL:
if (AM.IndexReg.getNode() != nullptr || AM.Scale != 1)
break;
if (ConstantSDNode *CN = dyn_cast<ConstantSDNode>(N.getOperand(1))) {
unsigned Val = CN->getZExtValue();
// Note that we handle x<<1 as (,x,2) rather than (x,x) here so
// that the base operand remains free for further matching. If
// the base doesn't end up getting used, a post-processing step
// in MatchAddress turns (,x,2) into (x,x), which is cheaper.
if (Val == 1 || Val == 2 || Val == 3) {
AM.Scale = 1 << Val;
SDValue ShVal = N.getOperand(0);
// Okay, we know that we have a scale by now. However, if the scaled
// value is an add of something and a constant, we can fold the
// constant into the disp field here.
if (CurDAG->isBaseWithConstantOffset(ShVal)) {
AM.IndexReg = ShVal.getOperand(0);
ConstantSDNode *AddVal = cast<ConstantSDNode>(ShVal.getOperand(1));
uint64_t Disp = (uint64_t)AddVal->getSExtValue() << Val;
if (!foldOffsetIntoAddress(Disp, AM))
return false;
}
AM.IndexReg = ShVal;
return false;
}
}
break;
case ISD::SRL: {
// Scale must not be used already.
if (AM.IndexReg.getNode() != nullptr || AM.Scale != 1) break;
SDValue And = N.getOperand(0);
if (And.getOpcode() != ISD::AND) break;
SDValue X = And.getOperand(0);
// We only handle up to 64-bit values here as those are what matter for
// addressing mode optimizations.
if (X.getSimpleValueType().getSizeInBits() > 64) break;
// The mask used for the transform is expected to be post-shift, but we
// found the shift first so just apply the shift to the mask before passing
// it down.
if (!isa<ConstantSDNode>(N.getOperand(1)) ||
!isa<ConstantSDNode>(And.getOperand(1)))
break;
uint64_t Mask = And.getConstantOperandVal(1) >> N.getConstantOperandVal(1);
// Try to fold the mask and shift into the scale, and return false if we
// succeed.
if (!foldMaskAndShiftToScale(*CurDAG, N, Mask, N, X, AM))
return false;
break;
}
case ISD::SMUL_LOHI:
case ISD::UMUL_LOHI:
// A mul_lohi where we need the low part can be folded as a plain multiply.
if (N.getResNo() != 0) break;
LLVM_FALLTHROUGH;
case ISD::MUL:
case X86ISD::MUL_IMM:
// X*[3,5,9] -> X+X*[2,4,8]
if (AM.BaseType == X86ISelAddressMode::RegBase &&
AM.Base_Reg.getNode() == nullptr &&
AM.IndexReg.getNode() == nullptr) {
if (ConstantSDNode *CN = dyn_cast<ConstantSDNode>(N.getOperand(1)))
if (CN->getZExtValue() == 3 || CN->getZExtValue() == 5 ||
CN->getZExtValue() == 9) {
AM.Scale = unsigned(CN->getZExtValue())-1;
SDValue MulVal = N.getOperand(0);
SDValue Reg;
// Okay, we know that we have a scale by now. However, if the scaled
// value is an add of something and a constant, we can fold the
// constant into the disp field here.
if (MulVal.getNode()->getOpcode() == ISD::ADD && MulVal.hasOneUse() &&
isa<ConstantSDNode>(MulVal.getOperand(1))) {
Reg = MulVal.getOperand(0);
ConstantSDNode *AddVal =
cast<ConstantSDNode>(MulVal.getOperand(1));
uint64_t Disp = AddVal->getSExtValue() * CN->getZExtValue();
if (foldOffsetIntoAddress(Disp, AM))
Reg = N.getOperand(0);
} else {
Reg = N.getOperand(0);
}
AM.IndexReg = AM.Base_Reg = Reg;
return false;
}
}
break;
case ISD::SUB: {
// Given A-B, if A can be completely folded into the address and
// the index field with the index field unused, use -B as the index.
// This is a win if a has multiple parts that can be folded into
// the address. Also, this saves a mov if the base register has
// other uses, since it avoids a two-address sub instruction, however
// it costs an additional mov if the index register has other uses.
// Add an artificial use to this node so that we can keep track of
// it if it gets CSE'd with a different node.
HandleSDNode Handle(N);
// Test if the LHS of the sub can be folded.
X86ISelAddressMode Backup = AM;
if (matchAddressRecursively(N.getOperand(0), AM, Depth+1)) {
AM = Backup;
break;
}
// Test if the index field is free for use.
if (AM.IndexReg.getNode() || AM.isRIPRelative()) {
AM = Backup;
break;
}
int Cost = 0;
SDValue RHS = Handle.getValue().getOperand(1);
// If the RHS involves a register with multiple uses, this
// transformation incurs an extra mov, due to the neg instruction
// clobbering its operand.
if (!RHS.getNode()->hasOneUse() ||
RHS.getNode()->getOpcode() == ISD::CopyFromReg ||
RHS.getNode()->getOpcode() == ISD::TRUNCATE ||
RHS.getNode()->getOpcode() == ISD::ANY_EXTEND ||
(RHS.getNode()->getOpcode() == ISD::ZERO_EXTEND &&
RHS.getOperand(0).getValueType() == MVT::i32))
++Cost;
// If the base is a register with multiple uses, this
// transformation may save a mov.
// FIXME: Don't rely on DELETED_NODEs.
if ((AM.BaseType == X86ISelAddressMode::RegBase && AM.Base_Reg.getNode() &&
AM.Base_Reg->getOpcode() != ISD::DELETED_NODE &&
!AM.Base_Reg.getNode()->hasOneUse()) ||
AM.BaseType == X86ISelAddressMode::FrameIndexBase)
--Cost;
// If the folded LHS was interesting, this transformation saves
// address arithmetic.
if ((AM.hasSymbolicDisplacement() && !Backup.hasSymbolicDisplacement()) +
((AM.Disp != 0) && (Backup.Disp == 0)) +
(AM.Segment.getNode() && !Backup.Segment.getNode()) >= 2)
--Cost;
// If it doesn't look like it may be an overall win, don't do it.
if (Cost >= 0) {
AM = Backup;
break;
}
// Ok, the transformation is legal and appears profitable. Go for it.
SDValue Zero = CurDAG->getConstant(0, dl, N.getValueType());
SDValue Neg = CurDAG->getNode(ISD::SUB, dl, N.getValueType(), Zero, RHS);
AM.IndexReg = Neg;
AM.Scale = 1;
// Insert the new nodes into the topological ordering.
insertDAGNode(*CurDAG, Handle.getValue(), Zero);
insertDAGNode(*CurDAG, Handle.getValue(), Neg);
return false;
}
case ISD::ADD:
if (!matchAdd(N, AM, Depth))
return false;
break;
case ISD::OR:
// We want to look through a transform in InstCombine and DAGCombiner that
// turns 'add' into 'or', so we can treat this 'or' exactly like an 'add'.
// Example: (or (and x, 1), (shl y, 3)) --> (add (and x, 1), (shl y, 3))
// An 'lea' can then be used to match the shift (multiply) and add:
// and $1, %esi
// lea (%rsi, %rdi, 8), %rax
if (CurDAG->haveNoCommonBitsSet(N.getOperand(0), N.getOperand(1)) &&
!matchAdd(N, AM, Depth))
return false;
break;
case ISD::AND: {
// Perform some heroic transforms on an and of a constant-count shift
// with a constant to enable use of the scaled offset field.
// Scale must not be used already.
if (AM.IndexReg.getNode() != nullptr || AM.Scale != 1) break;
SDValue Shift = N.getOperand(0);
if (Shift.getOpcode() != ISD::SRL && Shift.getOpcode() != ISD::SHL) break;
SDValue X = Shift.getOperand(0);
// We only handle up to 64-bit values here as those are what matter for
// addressing mode optimizations.
if (X.getSimpleValueType().getSizeInBits() > 64) break;
if (!isa<ConstantSDNode>(N.getOperand(1)))
break;
uint64_t Mask = N.getConstantOperandVal(1);
// Try to fold the mask and shift into an extract and scale.
if (!foldMaskAndShiftToExtract(*CurDAG, N, Mask, Shift, X, AM))
return false;
// Try to fold the mask and shift directly into the scale.
if (!foldMaskAndShiftToScale(*CurDAG, N, Mask, Shift, X, AM))
return false;
// Try to swap the mask and shift to place shifts which can be done as
// a scale on the outside of the mask.
if (!foldMaskedShiftToScaledMask(*CurDAG, N, Mask, Shift, X, AM))
return false;
// Try to fold the mask and shift into BEXTR and scale.
if (!foldMaskedShiftToBEXTR(*CurDAG, N, Mask, Shift, X, AM, *Subtarget))
return false;
break;
}
}
return matchAddressBase(N, AM);
}
/// Helper for MatchAddress. Add the specified node to the
/// specified addressing mode without any further recursion.
bool X86DAGToDAGISel::matchAddressBase(SDValue N, X86ISelAddressMode &AM) {
// Is the base register already occupied?
if (AM.BaseType != X86ISelAddressMode::RegBase || AM.Base_Reg.getNode()) {
// If so, check to see if the scale index register is set.
if (!AM.IndexReg.getNode()) {
AM.IndexReg = N;
AM.Scale = 1;
return false;
}
// Otherwise, we cannot select it.
return true;
}
// Default, generate it as a register.
AM.BaseType = X86ISelAddressMode::RegBase;
AM.Base_Reg = N;
return false;
}
/// Helper for selectVectorAddr. Handles things that can be folded into a
/// gather scatter address. The index register and scale should have already
/// been handled.
bool X86DAGToDAGISel::matchVectorAddress(SDValue N, X86ISelAddressMode &AM) {
// TODO: Support other operations.
switch (N.getOpcode()) {
case ISD::Constant: {
uint64_t Val = cast<ConstantSDNode>(N)->getSExtValue();
if (!foldOffsetIntoAddress(Val, AM))
return false;
break;
}
case X86ISD::Wrapper:
if (!matchWrapper(N, AM))
return false;
break;
}
return matchAddressBase(N, AM);
}
bool X86DAGToDAGISel::selectVectorAddr(SDNode *Parent, SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index,
SDValue &Disp, SDValue &Segment) {
X86ISelAddressMode AM;
auto *Mgs = cast<X86MaskedGatherScatterSDNode>(Parent);
AM.IndexReg = Mgs->getIndex();
AM.Scale = cast<ConstantSDNode>(Mgs->getScale())->getZExtValue();
unsigned AddrSpace = cast<MemSDNode>(Parent)->getPointerInfo().getAddrSpace();
// AddrSpace 256 -> GS, 257 -> FS, 258 -> SS.
if (AddrSpace == 256)
AM.Segment = CurDAG->getRegister(X86::GS, MVT::i16);
if (AddrSpace == 257)
AM.Segment = CurDAG->getRegister(X86::FS, MVT::i16);
if (AddrSpace == 258)
AM.Segment = CurDAG->getRegister(X86::SS, MVT::i16);
// Try to match into the base and displacement fields.
if (matchVectorAddress(N, AM))
return false;
MVT VT = N.getSimpleValueType();
if (AM.BaseType == X86ISelAddressMode::RegBase) {
if (!AM.Base_Reg.getNode())
AM.Base_Reg = CurDAG->getRegister(0, VT);
}
getAddressOperands(AM, SDLoc(N), Base, Scale, Index, Disp, Segment);
return true;
}
/// Returns true if it is able to pattern match an addressing mode.
/// It returns the operands which make up the maximal addressing mode it can
/// match by reference.
///
/// Parent is the parent node of the addr operand that is being matched. It
/// is always a load, store, atomic node, or null. It is only null when
/// checking memory operands for inline asm nodes.
bool X86DAGToDAGISel::selectAddr(SDNode *Parent, SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index,
SDValue &Disp, SDValue &Segment) {
X86ISelAddressMode AM;
if (Parent &&
// This list of opcodes are all the nodes that have an "addr:$ptr" operand
// that are not a MemSDNode, and thus don't have proper addrspace info.
Parent->getOpcode() != ISD::INTRINSIC_W_CHAIN && // unaligned loads, fixme
Parent->getOpcode() != ISD::INTRINSIC_VOID && // nontemporal stores
Parent->getOpcode() != X86ISD::TLSCALL && // Fixme
Parent->getOpcode() != X86ISD::EH_SJLJ_SETJMP && // setjmp
Parent->getOpcode() != X86ISD::EH_SJLJ_LONGJMP) { // longjmp
unsigned AddrSpace =
cast<MemSDNode>(Parent)->getPointerInfo().getAddrSpace();
// AddrSpace 256 -> GS, 257 -> FS, 258 -> SS.
if (AddrSpace == 256)
AM.Segment = CurDAG->getRegister(X86::GS, MVT::i16);
if (AddrSpace == 257)
AM.Segment = CurDAG->getRegister(X86::FS, MVT::i16);
if (AddrSpace == 258)
AM.Segment = CurDAG->getRegister(X86::SS, MVT::i16);
}
if (matchAddress(N, AM))
return false;
MVT VT = N.getSimpleValueType();
if (AM.BaseType == X86ISelAddressMode::RegBase) {
if (!AM.Base_Reg.getNode())
AM.Base_Reg = CurDAG->getRegister(0, VT);
}
if (!AM.IndexReg.getNode())
AM.IndexReg = CurDAG->getRegister(0, VT);
getAddressOperands(AM, SDLoc(N), Base, Scale, Index, Disp, Segment);
return true;
}
// We can only fold a load if all nodes between it and the root node have a
// single use. If there are additional uses, we could end up duplicating the
// load.
static bool hasSingleUsesFromRoot(SDNode *Root, SDNode *User) {
while (User != Root) {
if (!User->hasOneUse())
return false;
User = *User->use_begin();
}
return true;
}
/// Match a scalar SSE load. In particular, we want to match a load whose top
/// elements are either undef or zeros. The load flavor is derived from the
/// type of N, which is either v4f32 or v2f64.
///
/// We also return:
/// PatternChainNode: this is the matched node that has a chain input and
/// output.
bool X86DAGToDAGISel::selectScalarSSELoad(SDNode *Root, SDNode *Parent,
SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index,
SDValue &Disp, SDValue &Segment,
SDValue &PatternNodeWithChain) {
if (!hasSingleUsesFromRoot(Root, Parent))
return false;
// We can allow a full vector load here since narrowing a load is ok.
if (ISD::isNON_EXTLoad(N.getNode())) {
PatternNodeWithChain = N;
if (IsProfitableToFold(PatternNodeWithChain, N.getNode(), Root) &&
IsLegalToFold(PatternNodeWithChain, Parent, Root, OptLevel)) {
LoadSDNode *LD = cast<LoadSDNode>(PatternNodeWithChain);
return selectAddr(LD, LD->getBasePtr(), Base, Scale, Index, Disp,
Segment);
}
}
// We can also match the special zero extended load opcode.
if (N.getOpcode() == X86ISD::VZEXT_LOAD) {
PatternNodeWithChain = N;
if (IsProfitableToFold(PatternNodeWithChain, N.getNode(), Root) &&
IsLegalToFold(PatternNodeWithChain, Parent, Root, OptLevel)) {
auto *MI = cast<MemIntrinsicSDNode>(PatternNodeWithChain);
return selectAddr(MI, MI->getBasePtr(), Base, Scale, Index, Disp,
Segment);
}
}
// Need to make sure that the SCALAR_TO_VECTOR and load are both only used
// once. Otherwise the load might get duplicated and the chain output of the
// duplicate load will not be observed by all dependencies.
if (N.getOpcode() == ISD::SCALAR_TO_VECTOR && N.getNode()->hasOneUse()) {
PatternNodeWithChain = N.getOperand(0);
if (ISD::isNON_EXTLoad(PatternNodeWithChain.getNode()) &&
IsProfitableToFold(PatternNodeWithChain, N.getNode(), Root) &&
IsLegalToFold(PatternNodeWithChain, N.getNode(), Root, OptLevel)) {
LoadSDNode *LD = cast<LoadSDNode>(PatternNodeWithChain);
return selectAddr(LD, LD->getBasePtr(), Base, Scale, Index, Disp,
Segment);
}
}
// Also handle the case where we explicitly require zeros in the top
// elements. This is a vector shuffle from the zero vector.
if (N.getOpcode() == X86ISD::VZEXT_MOVL && N.getNode()->hasOneUse() &&
// Check to see if the top elements are all zeros (or bitcast of zeros).
N.getOperand(0).getOpcode() == ISD::SCALAR_TO_VECTOR &&
N.getOperand(0).getNode()->hasOneUse()) {
PatternNodeWithChain = N.getOperand(0).getOperand(0);
if (ISD::isNON_EXTLoad(PatternNodeWithChain.getNode()) &&
IsProfitableToFold(PatternNodeWithChain, N.getNode(), Root) &&
IsLegalToFold(PatternNodeWithChain, N.getNode(), Root, OptLevel)) {
// Okay, this is a zero extending load. Fold it.
LoadSDNode *LD = cast<LoadSDNode>(PatternNodeWithChain);
return selectAddr(LD, LD->getBasePtr(), Base, Scale, Index, Disp,
Segment);
}
}
return false;
}
bool X86DAGToDAGISel::selectMOV64Imm32(SDValue N, SDValue &Imm) {
if (const ConstantSDNode *CN = dyn_cast<ConstantSDNode>(N)) {
uint64_t ImmVal = CN->getZExtValue();
if (!isUInt<32>(ImmVal))
return false;
Imm = CurDAG->getTargetConstant(ImmVal, SDLoc(N), MVT::i64);
return true;
}
// In static codegen with small code model, we can get the address of a label
// into a register with 'movl'
if (N->getOpcode() != X86ISD::Wrapper)
return false;
N = N.getOperand(0);
// At least GNU as does not accept 'movl' for TPOFF relocations.
// FIXME: We could use 'movl' when we know we are targeting MC.
if (N->getOpcode() == ISD::TargetGlobalTLSAddress)
return false;
Imm = N;
if (N->getOpcode() != ISD::TargetGlobalAddress)
return TM.getCodeModel() == CodeModel::Small;
Optional<ConstantRange> CR =
cast<GlobalAddressSDNode>(N)->getGlobal()->getAbsoluteSymbolRange();
if (!CR)
return TM.getCodeModel() == CodeModel::Small;
return CR->getUnsignedMax().ult(1ull << 32);
}
bool X86DAGToDAGISel::selectLEA64_32Addr(SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index,
SDValue &Disp, SDValue &Segment) {
// Save the debug loc before calling selectLEAAddr, in case it invalidates N.
SDLoc DL(N);
if (!selectLEAAddr(N, Base, Scale, Index, Disp, Segment))
return false;
RegisterSDNode *RN = dyn_cast<RegisterSDNode>(Base);
if (RN && RN->getReg() == 0)
Base = CurDAG->getRegister(0, MVT::i64);
else if (Base.getValueType() == MVT::i32 && !dyn_cast<FrameIndexSDNode>(Base)) {
// Base could already be %rip, particularly in the x32 ABI.
Base = SDValue(CurDAG->getMachineNode(
TargetOpcode::SUBREG_TO_REG, DL, MVT::i64,
CurDAG->getTargetConstant(0, DL, MVT::i64),
Base,
CurDAG->getTargetConstant(X86::sub_32bit, DL, MVT::i32)),
0);
}
RN = dyn_cast<RegisterSDNode>(Index);
if (RN && RN->getReg() == 0)
Index = CurDAG->getRegister(0, MVT::i64);
else {
assert(Index.getValueType() == MVT::i32 &&
"Expect to be extending 32-bit registers for use in LEA");
Index = SDValue(CurDAG->getMachineNode(
TargetOpcode::SUBREG_TO_REG, DL, MVT::i64,
CurDAG->getTargetConstant(0, DL, MVT::i64),
Index,
CurDAG->getTargetConstant(X86::sub_32bit, DL,
MVT::i32)),
0);
}
return true;
}
/// Calls SelectAddr and determines if the maximal addressing
/// mode it matches can be cost effectively emitted as an LEA instruction.
bool X86DAGToDAGISel::selectLEAAddr(SDValue N,
SDValue &Base, SDValue &Scale,
SDValue &Index, SDValue &Disp,
SDValue &Segment) {
X86ISelAddressMode AM;
// Save the DL and VT before calling matchAddress, it can invalidate N.
SDLoc DL(N);
MVT VT = N.getSimpleValueType();
// Set AM.Segment to prevent MatchAddress from using one. LEA doesn't support
// segments.
SDValue Copy = AM.Segment;
SDValue T = CurDAG->getRegister(0, MVT::i32);
AM.Segment = T;
if (matchAddress(N, AM))
return false;
assert (T == AM.Segment);
AM.Segment = Copy;
unsigned Complexity = 0;
if (AM.BaseType == X86ISelAddressMode::RegBase)
if (AM.Base_Reg.getNode())
Complexity = 1;
else
AM.Base_Reg = CurDAG->getRegister(0, VT);
else if (AM.BaseType == X86ISelAddressMode::FrameIndexBase)
Complexity = 4;
if (AM.IndexReg.getNode())
Complexity++;
else
AM.IndexReg = CurDAG->getRegister(0, VT);
// Don't match just leal(,%reg,2). It's cheaper to do addl %reg, %reg, or with
// a simple shift.
if (AM.Scale > 1)
Complexity++;
// FIXME: We are artificially lowering the criteria to turn ADD %reg, $GA
// to a LEA. This is determined with some experimentation but is by no means
// optimal (especially for code size consideration). LEA is nice because of
// its three-address nature. Tweak the cost function again when we can run
// convertToThreeAddress() at register allocation time.
if (AM.hasSymbolicDisplacement()) {
// For X86-64, always use LEA to materialize RIP-relative addresses.
if (Subtarget->is64Bit())
Complexity = 4;
else
Complexity += 2;
}
if (AM.Disp && (AM.Base_Reg.getNode() || AM.IndexReg.getNode()))
Complexity++;
// If it isn't worth using an LEA, reject it.
if (Complexity <= 2)
return false;
getAddressOperands(AM, DL, Base, Scale, Index, Disp, Segment);
return true;
}
/// This is only run on TargetGlobalTLSAddress nodes.
bool X86DAGToDAGISel::selectTLSADDRAddr(SDValue N, SDValue &Base,
SDValue &Scale, SDValue &Index,
SDValue &Disp, SDValue &Segment) {
assert(N.getOpcode() == ISD::TargetGlobalTLSAddress);
const GlobalAddressSDNode *GA = cast<GlobalAddressSDNode>(N);
X86ISelAddressMode AM;
AM.GV = GA->getGlobal();
AM.Disp += GA->getOffset();
AM.Base_Reg = CurDAG->getRegister(0, N.getValueType());
AM.SymbolFlags = GA->getTargetFlags();
if (N.getValueType() == MVT::i32) {
AM.Scale = 1;
AM.IndexReg = CurDAG->getRegister(X86::EBX, MVT::i32);
} else {
AM.IndexReg = CurDAG->getRegister(0, MVT::i64);
}
getAddressOperands(AM, SDLoc(N), Base, Scale, Index, Disp, Segment);
return true;
}
bool X86DAGToDAGISel::selectRelocImm(SDValue N, SDValue &Op) {
if (auto *CN = dyn_cast<ConstantSDNode>(N)) {
Op = CurDAG->getTargetConstant(CN->getAPIntValue(), SDLoc(CN),
N.getValueType());
return true;
}
// Keep track of the original value type and whether this value was
// truncated. If we see a truncation from pointer type to VT that truncates
// bits that are known to be zero, we can use a narrow reference.
EVT VT = N.getValueType();
bool WasTruncated = false;
if (N.getOpcode() == ISD::TRUNCATE) {
WasTruncated = true;
N = N.getOperand(0);
}
if (N.getOpcode() != X86ISD::Wrapper)
return false;
// We can only use non-GlobalValues as immediates if they were not truncated,
// as we do not have any range information. If we have a GlobalValue and the
// address was not truncated, we can select it as an operand directly.
unsigned Opc = N.getOperand(0)->getOpcode();
if (Opc != ISD::TargetGlobalAddress || !WasTruncated) {
Op = N.getOperand(0);
// We can only select the operand directly if we didn't have to look past a
// truncate.
return !WasTruncated;
}
// Check that the global's range fits into VT.
auto *GA = cast<GlobalAddressSDNode>(N.getOperand(0));
Optional<ConstantRange> CR = GA->getGlobal()->getAbsoluteSymbolRange();
if (!CR || CR->getUnsignedMax().uge(1ull << VT.getSizeInBits()))
return false;
// Okay, we can use a narrow reference.
Op = CurDAG->getTargetGlobalAddress(GA->getGlobal(), SDLoc(N), VT,
GA->getOffset(), GA->getTargetFlags());
return true;
}
bool X86DAGToDAGISel::tryFoldLoad(SDNode *Root, SDNode *P, SDValue N,
SDValue &Base, SDValue &Scale,
SDValue &Index, SDValue &Disp,
SDValue &Segment) {
if (!ISD::isNON_EXTLoad(N.getNode()) ||
!IsProfitableToFold(N, P, Root) ||
!IsLegalToFold(N, P, Root, OptLevel))
return false;
return selectAddr(N.getNode(),
N.getOperand(1), Base, Scale, Index, Disp, Segment);
}
/// Return an SDNode that returns the value of the global base register.
/// Output instructions required to initialize the global base register,
/// if necessary.
SDNode *X86DAGToDAGISel::getGlobalBaseReg() {
unsigned GlobalBaseReg = getInstrInfo()->getGlobalBaseReg(MF);
auto &DL = MF->getDataLayout();
return CurDAG->getRegister(GlobalBaseReg, TLI->getPointerTy(DL)).getNode();
}
bool X86DAGToDAGISel::isSExtAbsoluteSymbolRef(unsigned Width, SDNode *N) const {
if (N->getOpcode() == ISD::TRUNCATE)
N = N->getOperand(0).getNode();
if (N->getOpcode() != X86ISD::Wrapper)
return false;
auto *GA = dyn_cast<GlobalAddressSDNode>(N->getOperand(0));
if (!GA)
return false;
Optional<ConstantRange> CR = GA->getGlobal()->getAbsoluteSymbolRange();
return CR && CR->getSignedMin().sge(-1ull << Width) &&
CR->getSignedMax().slt(1ull << Width);
}
/// Test whether the given X86ISD::CMP node has any uses which require the SF
/// or OF bits to be accurate.
static bool hasNoSignedComparisonUses(SDNode *N) {
// Examine each user of the node.
for (SDNode::use_iterator UI = N->use_begin(),
UE = N->use_end(); UI != UE; ++UI) {
// Only examine CopyToReg uses.
if (UI->getOpcode() != ISD::CopyToReg)
return false;
// Only examine CopyToReg uses that copy to EFLAGS.
if (cast<RegisterSDNode>(UI->getOperand(1))->getReg() !=
X86::EFLAGS)
return false;
// Examine each user of the CopyToReg use.
for (SDNode::use_iterator FlagUI = UI->use_begin(),
FlagUE = UI->use_end(); FlagUI != FlagUE; ++FlagUI) {
// Only examine the Flag result.
if (FlagUI.getUse().getResNo() != 1) continue;
// Anything unusual: assume conservatively.
if (!FlagUI->isMachineOpcode()) return false;
// Examine the opcode of the user.
switch (FlagUI->getMachineOpcode()) {
// These comparisons don't treat the most significant bit specially.
case X86::SETAr: case X86::SETAEr: case X86::SETBr: case X86::SETBEr:
case X86::SETEr: case X86::SETNEr: case X86::SETPr: case X86::SETNPr:
case X86::SETAm: case X86::SETAEm: case X86::SETBm: case X86::SETBEm:
case X86::SETEm: case X86::SETNEm: case X86::SETPm: case X86::SETNPm:
case X86::JA_1: case X86::JAE_1: case X86::JB_1: case X86::JBE_1:
case X86::JE_1: case X86::JNE_1: case X86::JP_1: case X86::JNP_1:
case X86::CMOVA16rr: case X86::CMOVA16rm:
case X86::CMOVA32rr: case X86::CMOVA32rm:
case X86::CMOVA64rr: case X86::CMOVA64rm:
case X86::CMOVAE16rr: case X86::CMOVAE16rm:
case X86::CMOVAE32rr: case X86::CMOVAE32rm:
case X86::CMOVAE64rr: case X86::CMOVAE64rm:
case X86::CMOVB16rr: case X86::CMOVB16rm:
case X86::CMOVB32rr: case X86::CMOVB32rm:
case X86::CMOVB64rr: case X86::CMOVB64rm:
case X86::CMOVBE16rr: case X86::CMOVBE16rm:
case X86::CMOVBE32rr: case X86::CMOVBE32rm:
case X86::CMOVBE64rr: case X86::CMOVBE64rm:
case X86::CMOVE16rr: case X86::CMOVE16rm:
case X86::CMOVE32rr: case X86::CMOVE32rm:
case X86::CMOVE64rr: case X86::CMOVE64rm:
case X86::CMOVNE16rr: case X86::CMOVNE16rm:
case X86::CMOVNE32rr: case X86::CMOVNE32rm:
case X86::CMOVNE64rr: case X86::CMOVNE64rm:
case X86::CMOVNP16rr: case X86::CMOVNP16rm:
case X86::CMOVNP32rr: case X86::CMOVNP32rm:
case X86::CMOVNP64rr: case X86::CMOVNP64rm:
case X86::CMOVP16rr: case X86::CMOVP16rm:
case X86::CMOVP32rr: case X86::CMOVP32rm:
case X86::CMOVP64rr: case X86::CMOVP64rm:
continue;
// Anything else: assume conservatively.
default: return false;
}
}
}
return true;
}
/// Test whether the given node which sets flags has any uses which require the
/// CF flag to be accurate.
static bool hasNoCarryFlagUses(SDNode *N) {
// Examine each user of the node.
for (SDNode::use_iterator UI = N->use_begin(), UE = N->use_end(); UI != UE;
++UI) {
// Only check things that use the flags.
if (UI.getUse().getResNo() != 1)
continue;
// Only examine CopyToReg uses.
if (UI->getOpcode() != ISD::CopyToReg)
return false;
// Only examine CopyToReg uses that copy to EFLAGS.
if (cast<RegisterSDNode>(UI->getOperand(1))->getReg() != X86::EFLAGS)
return false;
// Examine each user of the CopyToReg use.
for (SDNode::use_iterator FlagUI = UI->use_begin(), FlagUE = UI->use_end();
FlagUI != FlagUE; ++FlagUI) {
// Only examine the Flag result.
if (FlagUI.getUse().getResNo() != 1)
continue;
// Anything unusual: assume conservatively.
if (!FlagUI->isMachineOpcode())
return false;
// Examine the opcode of the user.
switch (FlagUI->getMachineOpcode()) {
// Comparisons which don't examine the CF flag.
case X86::SETOr: case X86::SETNOr: case X86::SETEr: case X86::SETNEr:
case X86::SETSr: case X86::SETNSr: case X86::SETPr: case X86::SETNPr:
case X86::SETLr: case X86::SETGEr: case X86::SETLEr: case X86::SETGr:
case X86::JO_1: case X86::JNO_1: case X86::JE_1: case X86::JNE_1:
case X86::JS_1: case X86::JNS_1: case X86::JP_1: case X86::JNP_1:
case X86::JL_1: case X86::JGE_1: case X86::JLE_1: case X86::JG_1:
case X86::CMOVO16rr: case X86::CMOVO32rr: case X86::CMOVO64rr:
case X86::CMOVO16rm: case X86::CMOVO32rm: case X86::CMOVO64rm:
case X86::CMOVNO16rr: case X86::CMOVNO32rr: case X86::CMOVNO64rr:
case X86::CMOVNO16rm: case X86::CMOVNO32rm: case X86::CMOVNO64rm:
case X86::CMOVE16rr: case X86::CMOVE32rr: case X86::CMOVE64rr:
case X86::CMOVE16rm: case X86::CMOVE32rm: case X86::CMOVE64rm:
case X86::CMOVNE16rr: case X86::CMOVNE32rr: case X86::CMOVNE64rr:
case X86::CMOVNE16rm: case X86::CMOVNE32rm: case X86::CMOVNE64rm:
case X86::CMOVS16rr: case X86::CMOVS32rr: case X86::CMOVS64rr:
case X86::CMOVS16rm: case X86::CMOVS32rm: case X86::CMOVS64rm:
case X86::CMOVNS16rr: case X86::CMOVNS32rr: case X86::CMOVNS64rr:
case X86::CMOVNS16rm: case X86::CMOVNS32rm: case X86::CMOVNS64rm:
case X86::CMOVP16rr: case X86::CMOVP32rr: case X86::CMOVP64rr:
case X86::CMOVP16rm: case X86::CMOVP32rm: case X86::CMOVP64rm:
case X86::CMOVNP16rr: case X86::CMOVNP32rr: case X86::CMOVNP64rr:
case X86::CMOVNP16rm: case X86::CMOVNP32rm: case X86::CMOVNP64rm:
case X86::CMOVL16rr: case X86::CMOVL32rr: case X86::CMOVL64rr:
case X86::CMOVL16rm: case X86::CMOVL32rm: case X86::CMOVL64rm:
case X86::CMOVGE16rr: case X86::CMOVGE32rr: case X86::CMOVGE64rr:
case X86::CMOVGE16rm: case X86::CMOVGE32rm: case X86::CMOVGE64rm:
case X86::CMOVLE16rr: case X86::CMOVLE32rr: case X86::CMOVLE64rr:
case X86::CMOVLE16rm: case X86::CMOVLE32rm: case X86::CMOVLE64rm:
case X86::CMOVG16rr: case X86::CMOVG32rr: case X86::CMOVG64rr:
case X86::CMOVG16rm: case X86::CMOVG32rm: case X86::CMOVG64rm:
continue;
// Anything else: assume conservatively.
default:
return false;
}
}
}
return true;
}
/// Check whether or not the chain ending in StoreNode is suitable for doing
/// the {load; op; store} to modify transformation.
static bool isFusableLoadOpStorePattern(StoreSDNode *StoreNode,
SDValue StoredVal, SelectionDAG *CurDAG,
unsigned LoadOpNo,
LoadSDNode *&LoadNode,
SDValue &InputChain) {
// Is the stored value result 0 of the operation?
if (StoredVal.getResNo() != 0) return false;
// Are there other uses of the operation other than the store?
if (!StoredVal.getNode()->hasNUsesOfValue(1, 0)) return false;
// Is the store non-extending and non-indexed?
if (!ISD::isNormalStore(StoreNode) || StoreNode->isNonTemporal())
return false;
SDValue Load = StoredVal->getOperand(LoadOpNo);
// Is the stored value a non-extending and non-indexed load?
if (!ISD::isNormalLoad(Load.getNode())) return false;
// Return LoadNode by reference.
LoadNode = cast<LoadSDNode>(Load);
// Is store the only read of the loaded value?
if (!Load.hasOneUse())
return false;
// Is the address of the store the same as the load?
if (LoadNode->getBasePtr() != StoreNode->getBasePtr() ||
LoadNode->getOffset() != StoreNode->getOffset())
return false;
bool FoundLoad = false;
SmallVector<SDValue, 4> ChainOps;
SmallVector<const SDNode *, 4> LoopWorklist;
SmallPtrSet<const SDNode *, 16> Visited;
const unsigned int Max = 1024;
// Visualization of Load-Op-Store fusion:
// -------------------------
// Legend:
// *-lines = Chain operand dependencies.
// |-lines = Normal operand dependencies.
// Dependencies flow down and right. n-suffix references multiple nodes.
//
// C Xn C
// * * *
// * * *
// Xn A-LD Yn TF Yn
// * * \ | * |
// * * \ | * |
// * * \ | => A--LD_OP_ST
// * * \| \
// TF OP \
// * | \ Zn
// * | \
// A-ST Zn
//
// This merge induced dependences from: #1: Xn -> LD, OP, Zn
// #2: Yn -> LD
// #3: ST -> Zn
// Ensure the transform is safe by checking for the dual
// dependencies to make sure we do not induce a loop.
// As LD is a predecessor to both OP and ST we can do this by checking:
// a). if LD is a predecessor to a member of Xn or Yn.
// b). if a Zn is a predecessor to ST.
// However, (b) can only occur through being a chain predecessor to
// ST, which is the same as Zn being a member or predecessor of Xn,
// which is a subset of LD being a predecessor of Xn. So it's
// subsumed by check (a).
SDValue Chain = StoreNode->getChain();
// Gather X elements in ChainOps.
if (Chain == Load.getValue(1)) {
FoundLoad = true;
ChainOps.push_back(Load.getOperand(0));
} else if (Chain.getOpcode() == ISD::TokenFactor) {
for (unsigned i = 0, e = Chain.getNumOperands(); i != e; ++i) {
SDValue Op = Chain.getOperand(i);
if (Op == Load.getValue(1)) {
FoundLoad = true;
// Drop Load, but keep its chain. No cycle check necessary.
ChainOps.push_back(Load.getOperand(0));
continue;
}
LoopWorklist.push_back(Op.getNode());
ChainOps.push_back(Op);
}
}
if (!FoundLoad)
return false;
// Worklist is currently Xn. Add Yn to worklist.
for (SDValue Op : StoredVal->ops())
if (Op.getNode() != LoadNode)
LoopWorklist.push_back(Op.getNode());
// Check (a) if Load is a predecessor to Xn + Yn
if (SDNode::hasPredecessorHelper(Load.getNode(), Visited, LoopWorklist, Max,
true))
return false;
InputChain =
CurDAG->getNode(ISD::TokenFactor, SDLoc(Chain), MVT::Other, ChainOps);
return true;
}
// Change a chain of {load; op; store} of the same value into a simple op
// through memory of that value, if the uses of the modified value and its
// address are suitable.
//
// The tablegen pattern memory operand pattern is currently not able to match
// the case where the EFLAGS on the original operation are used.
//
// To move this to tablegen, we'll need to improve tablegen to allow flags to
// be transferred from a node in the pattern to the result node, probably with
// a new keyword. For example, we have this
// def DEC64m : RI<0xFF, MRM1m, (outs), (ins i64mem:$dst), "dec{q}\t$dst",
// [(store (add (loadi64 addr:$dst), -1), addr:$dst),
// (implicit EFLAGS)]>;
// but maybe need something like this
// def DEC64m : RI<0xFF, MRM1m, (outs), (ins i64mem:$dst), "dec{q}\t$dst",
// [(store (add (loadi64 addr:$dst), -1), addr:$dst),
// (transferrable EFLAGS)]>;
//
// Until then, we manually fold these and instruction select the operation
// here.
bool X86DAGToDAGISel::foldLoadStoreIntoMemOperand(SDNode *Node) {
StoreSDNode *StoreNode = cast<StoreSDNode>(Node);
SDValue StoredVal = StoreNode->getOperand(1);
unsigned Opc = StoredVal->getOpcode();
// Before we try to select anything, make sure this is memory operand size
// and opcode we can handle. Note that this must match the code below that
// actually lowers the opcodes.
EVT MemVT = StoreNode->getMemoryVT();
if (MemVT != MVT::i64 && MemVT != MVT::i32 && MemVT != MVT::i16 &&
MemVT != MVT::i8)
return false;
bool IsCommutable = false;
switch (Opc) {
default:
return false;
case X86ISD::INC:
case X86ISD::DEC:
case X86ISD::SUB:
case X86ISD::SBB:
break;
case X86ISD::ADD:
case X86ISD::ADC:
case X86ISD::AND:
case X86ISD::OR:
case X86ISD::XOR:
IsCommutable = true;
break;
}
unsigned LoadOpNo = 0;
LoadSDNode *LoadNode = nullptr;
SDValue InputChain;
if (!isFusableLoadOpStorePattern(StoreNode, StoredVal, CurDAG, LoadOpNo,
LoadNode, InputChain)) {
if (!IsCommutable)
return false;
// This operation is commutable, try the other operand.
LoadOpNo = 1;
if (!isFusableLoadOpStorePattern(StoreNode, StoredVal, CurDAG, LoadOpNo,
LoadNode, InputChain))
return false;
}
SDValue Base, Scale, Index, Disp, Segment;
if (!selectAddr(LoadNode, LoadNode->getBasePtr(), Base, Scale, Index, Disp,
Segment))
return false;
auto SelectOpcode = [&](unsigned Opc64, unsigned Opc32, unsigned Opc16,
unsigned Opc8) {
switch (MemVT.getSimpleVT().SimpleTy) {
case MVT::i64:
return Opc64;
case MVT::i32:
return Opc32;
case MVT::i16:
return Opc16;
case MVT::i8:
return Opc8;
default:
llvm_unreachable("Invalid size!");
}
};
MachineSDNode *Result;
switch (Opc) {
case X86ISD::INC:
case X86ISD::DEC: {
unsigned NewOpc =
Opc == X86ISD::INC
? SelectOpcode(X86::INC64m, X86::INC32m, X86::INC16m, X86::INC8m)
: SelectOpcode(X86::DEC64m, X86::DEC32m, X86::DEC16m, X86::DEC8m);
const SDValue Ops[] = {Base, Scale, Index, Disp, Segment, InputChain};
Result =
CurDAG->getMachineNode(NewOpc, SDLoc(Node), MVT::i32, MVT::Other, Ops);
break;
}
case X86ISD::ADD:
case X86ISD::ADC:
case X86ISD::SUB:
case X86ISD::SBB:
case X86ISD::AND:
case X86ISD::OR:
case X86ISD::XOR: {
auto SelectRegOpcode = [SelectOpcode](unsigned Opc) {
switch (Opc) {
case X86ISD::ADD:
return SelectOpcode(X86::ADD64mr, X86::ADD32mr, X86::ADD16mr,
X86::ADD8mr);
case X86ISD::ADC:
return SelectOpcode(X86::ADC64mr, X86::ADC32mr, X86::ADC16mr,
X86::ADC8mr);
case X86ISD::SUB:
return SelectOpcode(X86::SUB64mr, X86::SUB32mr, X86::SUB16mr,
X86::SUB8mr);
case X86ISD::SBB:
return SelectOpcode(X86::SBB64mr, X86::SBB32mr, X86::SBB16mr,
X86::SBB8mr);
case X86ISD::AND:
return SelectOpcode(X86::AND64mr, X86::AND32mr, X86::AND16mr,
X86::AND8mr);
case X86ISD::OR:
return SelectOpcode(X86::OR64mr, X86::OR32mr, X86::OR16mr, X86::OR8mr);
case X86ISD::XOR:
return SelectOpcode(X86::XOR64mr, X86::XOR32mr, X86::XOR16mr,
X86::XOR8mr);
default:
llvm_unreachable("Invalid opcode!");
}
};
auto SelectImm8Opcode = [SelectOpcode](unsigned Opc) {
switch (Opc) {
case X86ISD::ADD:
return SelectOpcode(X86::ADD64mi8, X86::ADD32mi8, X86::ADD16mi8, 0);
case X86ISD::ADC:
return SelectOpcode(X86::ADC64mi8, X86::ADC32mi8, X86::ADC16mi8, 0);
case X86ISD::SUB:
return SelectOpcode(X86::SUB64mi8, X86::SUB32mi8, X86::SUB16mi8, 0);
case X86ISD::SBB:
return SelectOpcode(X86::SBB64mi8, X86::SBB32mi8, X86::SBB16mi8, 0);
case X86ISD::AND:
return SelectOpcode(X86::AND64mi8, X86::AND32mi8, X86::AND16mi8, 0);
case X86ISD::OR:
return SelectOpcode(X86::OR64mi8, X86::OR32mi8, X86::OR16mi8, 0);
case X86ISD::XOR:
return SelectOpcode(X86::XOR64mi8, X86::XOR32mi8, X86::XOR16mi8, 0);
default:
llvm_unreachable("Invalid opcode!");
}
};
auto SelectImmOpcode = [SelectOpcode](unsigned Opc) {
switch (Opc) {
case X86ISD::ADD:
return SelectOpcode(X86::ADD64mi32, X86::ADD32mi, X86::ADD16mi,
X86::ADD8mi);
case X86ISD::ADC:
return SelectOpcode(X86::ADC64mi32, X86::ADC32mi, X86::ADC16mi,
X86::ADC8mi);
case X86ISD::SUB:
return SelectOpcode(X86::SUB64mi32, X86::SUB32mi, X86::SUB16mi,
X86::SUB8mi);
case X86ISD::SBB:
return SelectOpcode(X86::SBB64mi32, X86::SBB32mi, X86::SBB16mi,
X86::SBB8mi);
case X86ISD::AND:
return SelectOpcode(X86::AND64mi32, X86::AND32mi, X86::AND16mi,
X86::AND8mi);
case X86ISD::OR:
return SelectOpcode(X86::OR64mi32, X86::OR32mi, X86::OR16mi,
X86::OR8mi);
case X86ISD::XOR:
return SelectOpcode(X86::XOR64mi32, X86::XOR32mi, X86::XOR16mi,
X86::XOR8mi);
default:
llvm_unreachable("Invalid opcode!");
}
};
unsigned NewOpc = SelectRegOpcode(Opc);
SDValue Operand = StoredVal->getOperand(1-LoadOpNo);
// See if the operand is a constant that we can fold into an immediate
// operand.
if (auto *OperandC = dyn_cast<ConstantSDNode>(Operand)) {
auto OperandV = OperandC->getAPIntValue();
// Check if we can shrink the operand enough to fit in an immediate (or
// fit into a smaller immediate) by negating it and switching the
// operation.
if ((Opc == X86ISD::ADD || Opc == X86ISD::SUB) &&
((MemVT != MVT::i8 && OperandV.getMinSignedBits() > 8 &&
(-OperandV).getMinSignedBits() <= 8) ||
(MemVT == MVT::i64 && OperandV.getMinSignedBits() > 32 &&
(-OperandV).getMinSignedBits() <= 32)) &&
hasNoCarryFlagUses(StoredVal.getNode())) {
OperandV = -OperandV;
Opc = Opc == X86ISD::ADD ? X86ISD::SUB : X86ISD::ADD;
}
// First try to fit this into an Imm8 operand. If it doesn't fit, then try
// the larger immediate operand.
if (MemVT != MVT::i8 && OperandV.getMinSignedBits() <= 8) {
Operand = CurDAG->getTargetConstant(OperandV, SDLoc(Node), MemVT);
NewOpc = SelectImm8Opcode(Opc);
} else if (OperandV.getActiveBits() <= MemVT.getSizeInBits() &&
(MemVT != MVT::i64 || OperandV.getMinSignedBits() <= 32)) {
Operand = CurDAG->getTargetConstant(OperandV, SDLoc(Node), MemVT);
NewOpc = SelectImmOpcode(Opc);
}
}
if (Opc == X86ISD::ADC || Opc == X86ISD::SBB) {
SDValue CopyTo =
CurDAG->getCopyToReg(InputChain, SDLoc(Node), X86::EFLAGS,
StoredVal.getOperand(2), SDValue());
const SDValue Ops[] = {Base, Scale, Index, Disp,
Segment, Operand, CopyTo, CopyTo.getValue(1)};
Result = CurDAG->getMachineNode(NewOpc, SDLoc(Node), MVT::i32, MVT::Other,
Ops);
} else {
const SDValue Ops[] = {Base, Scale, Index, Disp,
Segment, Operand, InputChain};
Result = CurDAG->getMachineNode(NewOpc, SDLoc(Node), MVT::i32, MVT::Other,
Ops);
}
break;
}
default:
llvm_unreachable("Invalid opcode!");
}
MachineMemOperand *MemOps[] = {StoreNode->getMemOperand(),
LoadNode->getMemOperand()};
CurDAG->setNodeMemRefs(Result, MemOps);
// Update Load Chain uses as well.
ReplaceUses(SDValue(LoadNode, 1), SDValue(Result, 1));
ReplaceUses(SDValue(StoreNode, 0), SDValue(Result, 1));
ReplaceUses(SDValue(StoredVal.getNode(), 1), SDValue(Result, 0));
CurDAG->RemoveDeadNode(Node);
return true;
}
// See if this is an X & Mask that we can match to BEXTR/BZHI.
// Where Mask is one of the following patterns:
// a) x & (1 << nbits) - 1
// b) x & ~(-1 << nbits)
// c) x & (-1 >> (32 - y))
// d) x << (32 - y) >> (32 - y)
bool X86DAGToDAGISel::matchBitExtract(SDNode *Node) {
assert(
(Node->getOpcode() == ISD::AND || Node->getOpcode() == ISD::SRL) &&
"Should be either an and-mask, or right-shift after clearing high bits.");
// BEXTR is BMI instruction, BZHI is BMI2 instruction. We need at least one.
if (!Subtarget->hasBMI() && !Subtarget->hasBMI2())
return false;
MVT NVT = Node->getSimpleValueType(0);
// Only supported for 32 and 64 bits.
if (NVT != MVT::i32 && NVT != MVT::i64)
return false;
unsigned Size = NVT.getSizeInBits();
SDValue NBits;
// If we have BMI2's BZHI, we are ok with muti-use patterns.
// Else, if we only have BMI1's BEXTR, we require one-use.
const bool CanHaveExtraUses = Subtarget->hasBMI2();
auto checkUses = [CanHaveExtraUses](SDValue Op, unsigned NUses) {
return CanHaveExtraUses ||
Op.getNode()->hasNUsesOfValue(NUses, Op.getResNo());
};
auto checkOneUse = [checkUses](SDValue Op) { return checkUses(Op, 1); };
auto checkTwoUse = [checkUses](SDValue Op) { return checkUses(Op, 2); };
// a) x & ((1 << nbits) + (-1))
auto matchPatternA = [&checkOneUse, &NBits](SDValue Mask) -> bool {
// Match `add`. Must only have one use!
if (Mask->getOpcode() != ISD::ADD || !checkOneUse(Mask))
return false;
// We should be adding all-ones constant (i.e. subtracting one.)
if (!isAllOnesConstant(Mask->getOperand(1)))
return false;
// Match `1 << nbits`. Must only have one use!
SDValue M0 = Mask->getOperand(0);
if (M0->getOpcode() != ISD::SHL || !checkOneUse(M0))
return false;
if (!isOneConstant(M0->getOperand(0)))
return false;
NBits = M0->getOperand(1);
return true;
};
// b) x & ~(-1 << nbits)
auto matchPatternB = [&checkOneUse, &NBits](SDValue Mask) -> bool {
// Match `~()`. Must only have one use!
if (!isBitwiseNot(Mask) || !checkOneUse(Mask))
return false;
// Match `-1 << nbits`. Must only have one use!
SDValue M0 = Mask->getOperand(0);
if (M0->getOpcode() != ISD::SHL || !checkOneUse(M0))
return false;
if (!isAllOnesConstant(M0->getOperand(0)))
return false;
NBits = M0->getOperand(1);
return true;
};
// Match potentially-truncated (bitwidth - y)
auto matchShiftAmt = [checkOneUse, Size, &NBits](SDValue ShiftAmt) {
// Skip over a truncate of the shift amount.
if (ShiftAmt.getOpcode() == ISD::TRUNCATE) {
ShiftAmt = ShiftAmt.getOperand(0);
// The trunc should have been the only user of the real shift amount.
if (!checkOneUse(ShiftAmt))
return false;
}
// Match the shift amount as: (bitwidth - y). It should go away, too.
if (ShiftAmt.getOpcode() != ISD::SUB)
return false;
auto V0 = dyn_cast<ConstantSDNode>(ShiftAmt.getOperand(0));
if (!V0 || V0->getZExtValue() != Size)
return false;
NBits = ShiftAmt.getOperand(1);
return true;
};
// c) x & (-1 >> (32 - y))
auto matchPatternC = [&checkOneUse, matchShiftAmt](SDValue Mask) -> bool {
// Match `l>>`. Must only have one use!
if (Mask.getOpcode() != ISD::SRL || !checkOneUse(Mask))
return false;
// We should be shifting all-ones constant.
if (!isAllOnesConstant(Mask.getOperand(0)))
return false;
SDValue M1 = Mask.getOperand(1);
// The shift amount should not be used externally.
if (!checkOneUse(M1))
return false;
return matchShiftAmt(M1);
};
SDValue X;
// d) x << (32 - y) >> (32 - y)
auto matchPatternD = [&checkOneUse, &checkTwoUse, matchShiftAmt,
&X](SDNode *Node) -> bool {
if (Node->getOpcode() != ISD::SRL)
return false;
SDValue N0 = Node->getOperand(0);
if (N0->getOpcode() != ISD::SHL || !checkOneUse(N0))
return false;
SDValue N1 = Node->getOperand(1);
SDValue N01 = N0->getOperand(1);
// Both of the shifts must be by the exact same value.
// There should not be any uses of the shift amount outside of the pattern.
if (N1 != N01 || !checkTwoUse(N1))
return false;
if (!matchShiftAmt(N1))
return false;
X = N0->getOperand(0);
return true;
};
auto matchLowBitMask = [&matchPatternA, &matchPatternB,
&matchPatternC](SDValue Mask) -> bool {
// FIXME: pattern c.
return matchPatternA(Mask) || matchPatternB(Mask) || matchPatternC(Mask);
};
if (Node->getOpcode() == ISD::AND) {
X = Node->getOperand(0);
SDValue Mask = Node->getOperand(1);
if (matchLowBitMask(Mask)) {
// Great.
} else {
std::swap(X, Mask);
if (!matchLowBitMask(Mask))
return false;
}
} else if (!matchPatternD(Node))
return false;
SDLoc DL(Node);
SDValue OrigNBits = NBits;
if (NBits.getValueType() != NVT) {
// Truncate the shift amount.
NBits = CurDAG->getNode(ISD::TRUNCATE, DL, MVT::i8, NBits);
insertDAGNode(*CurDAG, OrigNBits, NBits);
// Insert 8-bit NBits into lowest 8 bits of NVT-sized (32 or 64-bit)
// register. All the other bits are undefined, we do not care about them.
SDValue ImplDef =
SDValue(CurDAG->getMachineNode(TargetOpcode::IMPLICIT_DEF, DL, NVT), 0);
insertDAGNode(*CurDAG, OrigNBits, ImplDef);
NBits =
CurDAG->getTargetInsertSubreg(X86::sub_8bit, DL, NVT, ImplDef, NBits);
insertDAGNode(*CurDAG, OrigNBits, NBits);
}
if (Subtarget->hasBMI2()) {
// Great, just emit the the BZHI..
SDValue Extract = CurDAG->getNode(X86ISD::BZHI, DL, NVT, X, NBits);
ReplaceNode(Node, Extract.getNode());
SelectCode(Extract.getNode());
return true;
}
// Else, emitting BEXTR requires one more step.
// The 'control' of BEXTR has the pattern of:
// [15...8 bit][ 7...0 bit] location
// [ bit count][ shift] name
// I.e. 0b000000011'00000001 means (x >> 0b1) & 0b11
// Shift NBits left by 8 bits, thus producing 'control'.
SDValue C8 = CurDAG->getConstant(8, DL, MVT::i8);
SDValue Control = CurDAG->getNode(ISD::SHL, DL, NVT, NBits, C8);
insertDAGNode(*CurDAG, OrigNBits, Control);
// NOTE: could also try to extract start from (x >> start)
// And finally, form the BEXTR itself.
SDValue Extract = CurDAG->getNode(X86ISD::BEXTR, DL, NVT, X, Control);
ReplaceNode(Node, Extract.getNode());
SelectCode(Extract.getNode());
return true;
}
// See if this is an (X >> C1) & C2 that we can match to BEXTR/BEXTRI.
MachineSDNode *X86DAGToDAGISel::matchBEXTRFromAndImm(SDNode *Node) {
MVT NVT = Node->getSimpleValueType(0);
SDLoc dl(Node);
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
// If we have TBM we can use an immediate for the control. If we have BMI
// we should only do this if the BEXTR instruction is implemented well.
// Otherwise moving the control into a register makes this more costly.
// TODO: Maybe load folding, greater than 32-bit masks, or a guarantee of LICM
// hoisting the move immediate would make it worthwhile with a less optimal
// BEXTR?
if (!Subtarget->hasTBM() &&
!(Subtarget->hasBMI() && Subtarget->hasFastBEXTR()))
return nullptr;
// Must have a shift right.
if (N0->getOpcode() != ISD::SRL && N0->getOpcode() != ISD::SRA)
return nullptr;
// Shift can't have additional users.
if (!N0->hasOneUse())
return nullptr;
// Only supported for 32 and 64 bits.
if (NVT != MVT::i32 && NVT != MVT::i64)
return nullptr;
// Shift amount and RHS of and must be constant.
ConstantSDNode *MaskCst = dyn_cast<ConstantSDNode>(N1);
ConstantSDNode *ShiftCst = dyn_cast<ConstantSDNode>(N0->getOperand(1));
if (!MaskCst || !ShiftCst)
return nullptr;
// And RHS must be a mask.
uint64_t Mask = MaskCst->getZExtValue();
if (!isMask_64(Mask))
return nullptr;
uint64_t Shift = ShiftCst->getZExtValue();
uint64_t MaskSize = countPopulation(Mask);
// Don't interfere with something that can be handled by extracting AH.
// TODO: If we are able to fold a load, BEXTR might still be better than AH.
if (Shift == 8 && MaskSize == 8)
return nullptr;
// Make sure we are only using bits that were in the original value, not
// shifted in.
if (Shift + MaskSize > NVT.getSizeInBits())
return nullptr;
SDValue New = CurDAG->getTargetConstant(Shift | (MaskSize << 8), dl, NVT);
unsigned ROpc = NVT == MVT::i64 ? X86::BEXTRI64ri : X86::BEXTRI32ri;
unsigned MOpc = NVT == MVT::i64 ? X86::BEXTRI64mi : X86::BEXTRI32mi;
// BMI requires the immediate to placed in a register.
if (!Subtarget->hasTBM()) {
ROpc = NVT == MVT::i64 ? X86::BEXTR64rr : X86::BEXTR32rr;
MOpc = NVT == MVT::i64 ? X86::BEXTR64rm : X86::BEXTR32rm;
unsigned NewOpc = NVT == MVT::i64 ? X86::MOV32ri64 : X86::MOV32ri;
New = SDValue(CurDAG->getMachineNode(NewOpc, dl, NVT, New), 0);
}
MachineSDNode *NewNode;
SDValue Input = N0->getOperand(0);
SDValue Tmp0, Tmp1, Tmp2, Tmp3, Tmp4;
if (tryFoldLoad(Node, N0.getNode(), Input, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4)) {
SDValue Ops[] = { Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, New, Input.getOperand(0) };
SDVTList VTs = CurDAG->getVTList(NVT, MVT::Other);
NewNode = CurDAG->getMachineNode(MOpc, dl, VTs, Ops);
// Update the chain.
ReplaceUses(Input.getValue(1), SDValue(NewNode, 1));
// Record the mem-refs
CurDAG->setNodeMemRefs(NewNode, {cast<LoadSDNode>(Input)->getMemOperand()});
} else {
NewNode = CurDAG->getMachineNode(ROpc, dl, NVT, Input, New);
}
return NewNode;
}
// Emit a PCMISTR(I/M) instruction.
MachineSDNode *X86DAGToDAGISel::emitPCMPISTR(unsigned ROpc, unsigned MOpc,
bool MayFoldLoad, const SDLoc &dl,
MVT VT, SDNode *Node) {
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
SDValue Imm = Node->getOperand(2);
const ConstantInt *Val = cast<ConstantSDNode>(Imm)->getConstantIntValue();
Imm = CurDAG->getTargetConstant(*Val, SDLoc(Node), Imm.getValueType());
// Try to fold a load. No need to check alignment.
SDValue Tmp0, Tmp1, Tmp2, Tmp3, Tmp4;
if (MayFoldLoad && tryFoldLoad(Node, N1, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4)) {
SDValue Ops[] = { N0, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, Imm,
N1.getOperand(0) };
SDVTList VTs = CurDAG->getVTList(VT, MVT::i32, MVT::Other);
MachineSDNode *CNode = CurDAG->getMachineNode(MOpc, dl, VTs, Ops);
// Update the chain.
ReplaceUses(N1.getValue(1), SDValue(CNode, 2));
// Record the mem-refs
CurDAG->setNodeMemRefs(CNode, {cast<LoadSDNode>(N1)->getMemOperand()});
return CNode;
}
SDValue Ops[] = { N0, N1, Imm };
SDVTList VTs = CurDAG->getVTList(VT, MVT::i32);
MachineSDNode *CNode = CurDAG->getMachineNode(ROpc, dl, VTs, Ops);
return CNode;
}
// Emit a PCMESTR(I/M) instruction. Also return the Glue result in case we need
// to emit a second instruction after this one. This is needed since we have two
// copyToReg nodes glued before this and we need to continue that glue through.
MachineSDNode *X86DAGToDAGISel::emitPCMPESTR(unsigned ROpc, unsigned MOpc,
bool MayFoldLoad, const SDLoc &dl,
MVT VT, SDNode *Node,
SDValue &InFlag) {
SDValue N0 = Node->getOperand(0);
SDValue N2 = Node->getOperand(2);
SDValue Imm = Node->getOperand(4);
const ConstantInt *Val = cast<ConstantSDNode>(Imm)->getConstantIntValue();
Imm = CurDAG->getTargetConstant(*Val, SDLoc(Node), Imm.getValueType());
// Try to fold a load. No need to check alignment.
SDValue Tmp0, Tmp1, Tmp2, Tmp3, Tmp4;
if (MayFoldLoad && tryFoldLoad(Node, N2, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4)) {
SDValue Ops[] = { N0, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, Imm,
N2.getOperand(0), InFlag };
SDVTList VTs = CurDAG->getVTList(VT, MVT::i32, MVT::Other, MVT::Glue);
MachineSDNode *CNode = CurDAG->getMachineNode(MOpc, dl, VTs, Ops);
InFlag = SDValue(CNode, 3);
// Update the chain.
ReplaceUses(N2.getValue(1), SDValue(CNode, 2));
// Record the mem-refs
CurDAG->setNodeMemRefs(CNode, {cast<LoadSDNode>(N2)->getMemOperand()});
return CNode;
}
SDValue Ops[] = { N0, N2, Imm, InFlag };
SDVTList VTs = CurDAG->getVTList(VT, MVT::i32, MVT::Glue);
MachineSDNode *CNode = CurDAG->getMachineNode(ROpc, dl, VTs, Ops);
InFlag = SDValue(CNode, 2);
return CNode;
}
bool X86DAGToDAGISel::tryShiftAmountMod(SDNode *N) {
EVT VT = N->getValueType(0);
// Only handle scalar shifts.
if (VT.isVector())
return false;
// Narrower shifts only mask to 5 bits in hardware.
unsigned Size = VT == MVT::i64 ? 64 : 32;
SDValue OrigShiftAmt = N->getOperand(1);
SDValue ShiftAmt = OrigShiftAmt;
SDLoc DL(N);
// Skip over a truncate of the shift amount.
if (ShiftAmt->getOpcode() == ISD::TRUNCATE)
ShiftAmt = ShiftAmt->getOperand(0);
// This function is called after X86DAGToDAGISel::matchBitExtract(),
// so we are not afraid that we might mess up BZHI/BEXTR pattern.
SDValue NewShiftAmt;
if (ShiftAmt->getOpcode() == ISD::ADD || ShiftAmt->getOpcode() == ISD::SUB) {
SDValue Add0 = ShiftAmt->getOperand(0);
SDValue Add1 = ShiftAmt->getOperand(1);
// If we are shifting by X+/-N where N == 0 mod Size, then just shift by X
// to avoid the ADD/SUB.
if (isa<ConstantSDNode>(Add1) &&
cast<ConstantSDNode>(Add1)->getZExtValue() % Size == 0) {
NewShiftAmt = Add0;
// If we are shifting by N-X where N == 0 mod Size, then just shift by -X to
// generate a NEG instead of a SUB of a constant.
} else if (ShiftAmt->getOpcode() == ISD::SUB &&
isa<ConstantSDNode>(Add0) &&
cast<ConstantSDNode>(Add0)->getZExtValue() != 0 &&
cast<ConstantSDNode>(Add0)->getZExtValue() % Size == 0) {
// Insert a negate op.
// TODO: This isn't guaranteed to replace the sub if there is a logic cone
// that uses it that's not a shift.
EVT SubVT = ShiftAmt.getValueType();
SDValue Zero = CurDAG->getConstant(0, DL, SubVT);
SDValue Neg = CurDAG->getNode(ISD::SUB, DL, SubVT, Zero, Add1);
NewShiftAmt = Neg;
// Insert these operands into a valid topological order so they can
// get selected independently.
insertDAGNode(*CurDAG, OrigShiftAmt, Zero);
insertDAGNode(*CurDAG, OrigShiftAmt, Neg);
} else
return false;
} else
return false;
if (NewShiftAmt.getValueType() != MVT::i8) {
// Need to truncate the shift amount.
NewShiftAmt = CurDAG->getNode(ISD::TRUNCATE, DL, MVT::i8, NewShiftAmt);
// Add to a correct topological ordering.
insertDAGNode(*CurDAG, OrigShiftAmt, NewShiftAmt);
}
// Insert a new mask to keep the shift amount legal. This should be removed
// by isel patterns.
NewShiftAmt = CurDAG->getNode(ISD::AND, DL, MVT::i8, NewShiftAmt,
CurDAG->getConstant(Size - 1, DL, MVT::i8));
// Place in a correct topological ordering.
insertDAGNode(*CurDAG, OrigShiftAmt, NewShiftAmt);
SDNode *UpdatedNode = CurDAG->UpdateNodeOperands(N, N->getOperand(0),
NewShiftAmt);
if (UpdatedNode != N) {
// If we found an existing node, we should replace ourselves with that node
// and wait for it to be selected after its other users.
ReplaceNode(N, UpdatedNode);
return true;
}
// If the original shift amount is now dead, delete it so that we don't run
// it through isel.
if (OrigShiftAmt.getNode()->use_empty())
CurDAG->RemoveDeadNode(OrigShiftAmt.getNode());
// Now that we've optimized the shift amount, defer to normal isel to get
// load folding and legacy vs BMI2 selection without repeating it here.
SelectCode(N);
return true;
}
/// If the high bits of an 'and' operand are known zero, try setting the
/// high bits of an 'and' constant operand to produce a smaller encoding by
/// creating a small, sign-extended negative immediate rather than a large
/// positive one. This reverses a transform in SimplifyDemandedBits that
/// shrinks mask constants by clearing bits. There is also a possibility that
/// the 'and' mask can be made -1, so the 'and' itself is unnecessary. In that
/// case, just replace the 'and'. Return 'true' if the node is replaced.
bool X86DAGToDAGISel::shrinkAndImmediate(SDNode *And) {
// i8 is unshrinkable, i16 should be promoted to i32, and vector ops don't
// have immediate operands.
MVT VT = And->getSimpleValueType(0);
if (VT != MVT::i32 && VT != MVT::i64)
return false;
auto *And1C = dyn_cast<ConstantSDNode>(And->getOperand(1));
if (!And1C)
return false;
// Bail out if the mask constant is already negative. It's can't shrink more.
// If the upper 32 bits of a 64 bit mask are all zeros, we have special isel
// patterns to use a 32-bit and instead of a 64-bit and by relying on the
// implicit zeroing of 32 bit ops. So we should check if the lower 32 bits
// are negative too.
APInt MaskVal = And1C->getAPIntValue();
unsigned MaskLZ = MaskVal.countLeadingZeros();
if (!MaskLZ || (VT == MVT::i64 && MaskLZ == 32))
return false;
// Don't extend into the upper 32 bits of a 64 bit mask.
if (VT == MVT::i64 && MaskLZ >= 32) {
MaskLZ -= 32;
MaskVal = MaskVal.trunc(32);
}
SDValue And0 = And->getOperand(0);
APInt HighZeros = APInt::getHighBitsSet(MaskVal.getBitWidth(), MaskLZ);
APInt NegMaskVal = MaskVal | HighZeros;
// If a negative constant would not allow a smaller encoding, there's no need
// to continue. Only change the constant when we know it's a win.
unsigned MinWidth = NegMaskVal.getMinSignedBits();
if (MinWidth > 32 || (MinWidth > 8 && MaskVal.getMinSignedBits() <= 32))
return false;
// Extend masks if we truncated above.
if (VT == MVT::i64 && MaskVal.getBitWidth() < 64) {
NegMaskVal = NegMaskVal.zext(64);
HighZeros = HighZeros.zext(64);
}
// The variable operand must be all zeros in the top bits to allow using the
// new, negative constant as the mask.
if (!CurDAG->MaskedValueIsZero(And0, HighZeros))
return false;
// Check if the mask is -1. In that case, this is an unnecessary instruction
// that escaped earlier analysis.
if (NegMaskVal.isAllOnesValue()) {
ReplaceNode(And, And0.getNode());
return true;
}
// A negative mask allows a smaller encoding. Create a new 'and' node.
SDValue NewMask = CurDAG->getConstant(NegMaskVal, SDLoc(And), VT);
SDValue NewAnd = CurDAG->getNode(ISD::AND, SDLoc(And), VT, And0, NewMask);
ReplaceNode(And, NewAnd.getNode());
SelectCode(NewAnd.getNode());
return true;
}
void X86DAGToDAGISel::Select(SDNode *Node) {
MVT NVT = Node->getSimpleValueType(0);
unsigned Opcode = Node->getOpcode();
SDLoc dl(Node);
if (Node->isMachineOpcode()) {
LLVM_DEBUG(dbgs() << "== "; Node->dump(CurDAG); dbgs() << '\n');
Node->setNodeId(-1);
return; // Already selected.
}
switch (Opcode) {
default: break;
case ISD::BRIND: {
if (Subtarget->isTargetNaCl())
// NaCl has its own pass where jmp %r32 are converted to jmp %r64. We
// leave the instruction alone.
break;
if (Subtarget->isTarget64BitILP32()) {
// Converts a 32-bit register to a 64-bit, zero-extended version of
// it. This is needed because x86-64 can do many things, but jmp %r32
// ain't one of them.
const SDValue &Target = Node->getOperand(1);
assert(Target.getSimpleValueType() == llvm::MVT::i32);
SDValue ZextTarget = CurDAG->getZExtOrTrunc(Target, dl, EVT(MVT::i64));
SDValue Brind = CurDAG->getNode(ISD::BRIND, dl, MVT::Other,
Node->getOperand(0), ZextTarget);
ReplaceNode(Node, Brind.getNode());
SelectCode(ZextTarget.getNode());
SelectCode(Brind.getNode());
return;
}
break;
}
case X86ISD::GlobalBaseReg:
ReplaceNode(Node, getGlobalBaseReg());
return;
case ISD::BITCAST:
// Just drop all 128/256/512-bit bitcasts.
if (NVT.is512BitVector() || NVT.is256BitVector() || NVT.is128BitVector() ||
NVT == MVT::f128) {
ReplaceUses(SDValue(Node, 0), Node->getOperand(0));
CurDAG->RemoveDeadNode(Node);
return;
}
break;
case X86ISD::SELECT:
case X86ISD::SHRUNKBLEND: {
// SHRUNKBLEND selects like a regular VSELECT. Same with X86ISD::SELECT.
SDValue VSelect = CurDAG->getNode(
ISD::VSELECT, SDLoc(Node), Node->getValueType(0), Node->getOperand(0),
Node->getOperand(1), Node->getOperand(2));
ReplaceNode(Node, VSelect.getNode());
SelectCode(VSelect.getNode());
// We already called ReplaceUses.
return;
}
case ISD::SRL:
if (matchBitExtract(Node))
return;
LLVM_FALLTHROUGH;
case ISD::SRA:
case ISD::SHL:
if (tryShiftAmountMod(Node))
return;
break;
case ISD::AND:
if (MachineSDNode *NewNode = matchBEXTRFromAndImm(Node)) {
ReplaceUses(SDValue(Node, 0), SDValue(NewNode, 0));
CurDAG->RemoveDeadNode(Node);
return;
}
if (matchBitExtract(Node))
return;
if (AndImmShrink && shrinkAndImmediate(Node))
return;
LLVM_FALLTHROUGH;
case ISD::OR:
case ISD::XOR: {
// For operations of the form (x << C1) op C2, check if we can use a smaller
// encoding for C2 by transforming it into (x op (C2>>C1)) << C1.
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
if (N0->getOpcode() != ISD::SHL || !N0->hasOneUse())
break;
// i8 is unshrinkable, i16 should be promoted to i32.
if (NVT != MVT::i32 && NVT != MVT::i64)
break;
ConstantSDNode *Cst = dyn_cast<ConstantSDNode>(N1);
ConstantSDNode *ShlCst = dyn_cast<ConstantSDNode>(N0->getOperand(1));
if (!Cst || !ShlCst)
break;
int64_t Val = Cst->getSExtValue();
uint64_t ShlVal = ShlCst->getZExtValue();
// Make sure that we don't change the operation by removing bits.
// This only matters for OR and XOR, AND is unaffected.
uint64_t RemovedBitsMask = (1ULL << ShlVal) - 1;
if (Opcode != ISD::AND && (Val & RemovedBitsMask) != 0)
break;
unsigned ShlOp, AddOp, Op;
MVT CstVT = NVT;
// Check the minimum bitwidth for the new constant.
// TODO: AND32ri is the same as AND64ri32 with zext imm.
// TODO: MOV32ri+OR64r is cheaper than MOV64ri64+OR64rr
// TODO: Using 16 and 8 bit operations is also possible for or32 & xor32.
if (!isInt<8>(Val) && isInt<8>(Val >> ShlVal))
CstVT = MVT::i8;
else if (!isInt<32>(Val) && isInt<32>(Val >> ShlVal))
CstVT = MVT::i32;
// Bail if there is no smaller encoding.
if (NVT == CstVT)
break;
switch (NVT.SimpleTy) {
default: llvm_unreachable("Unsupported VT!");
case MVT::i32:
assert(CstVT == MVT::i8);
ShlOp = X86::SHL32ri;
AddOp = X86::ADD32rr;
switch (Opcode) {
default: llvm_unreachable("Impossible opcode");
case ISD::AND: Op = X86::AND32ri8; break;
case ISD::OR: Op = X86::OR32ri8; break;
case ISD::XOR: Op = X86::XOR32ri8; break;
}
break;
case MVT::i64:
assert(CstVT == MVT::i8 || CstVT == MVT::i32);
ShlOp = X86::SHL64ri;
AddOp = X86::ADD64rr;
switch (Opcode) {
default: llvm_unreachable("Impossible opcode");
case ISD::AND: Op = CstVT==MVT::i8? X86::AND64ri8 : X86::AND64ri32; break;
case ISD::OR: Op = CstVT==MVT::i8? X86::OR64ri8 : X86::OR64ri32; break;
case ISD::XOR: Op = CstVT==MVT::i8? X86::XOR64ri8 : X86::XOR64ri32; break;
}
break;
}
// Emit the smaller op and the shift.
SDValue NewCst = CurDAG->getTargetConstant(Val >> ShlVal, dl, CstVT);
SDNode *New = CurDAG->getMachineNode(Op, dl, NVT, N0->getOperand(0),NewCst);
if (ShlVal == 1)
CurDAG->SelectNodeTo(Node, AddOp, NVT, SDValue(New, 0),
SDValue(New, 0));
else
CurDAG->SelectNodeTo(Node, ShlOp, NVT, SDValue(New, 0),
getI8Imm(ShlVal, dl));
return;
}
case X86ISD::UMUL8:
case X86ISD::SMUL8: {
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
unsigned Opc = (Opcode == X86ISD::SMUL8 ? X86::IMUL8r : X86::MUL8r);
SDValue InFlag = CurDAG->getCopyToReg(CurDAG->getEntryNode(), dl, X86::AL,
N0, SDValue()).getValue(1);
SDVTList VTs = CurDAG->getVTList(NVT, MVT::i32);
SDValue Ops[] = {N1, InFlag};
SDNode *CNode = CurDAG->getMachineNode(Opc, dl, VTs, Ops);
ReplaceNode(Node, CNode);
return;
}
case X86ISD::UMUL: {
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
unsigned LoReg, Opc;
switch (NVT.SimpleTy) {
default: llvm_unreachable("Unsupported VT!");
// MVT::i8 is handled by X86ISD::UMUL8.
case MVT::i16: LoReg = X86::AX; Opc = X86::MUL16r; break;
case MVT::i32: LoReg = X86::EAX; Opc = X86::MUL32r; break;
case MVT::i64: LoReg = X86::RAX; Opc = X86::MUL64r; break;
}
SDValue InFlag = CurDAG->getCopyToReg(CurDAG->getEntryNode(), dl, LoReg,
N0, SDValue()).getValue(1);
SDVTList VTs = CurDAG->getVTList(NVT, NVT, MVT::i32);
SDValue Ops[] = {N1, InFlag};
SDNode *CNode = CurDAG->getMachineNode(Opc, dl, VTs, Ops);
ReplaceNode(Node, CNode);
return;
}
case ISD::SMUL_LOHI:
case ISD::UMUL_LOHI: {
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
unsigned Opc, MOpc;
bool isSigned = Opcode == ISD::SMUL_LOHI;
bool hasBMI2 = Subtarget->hasBMI2();
if (!isSigned) {
switch (NVT.SimpleTy) {
default: llvm_unreachable("Unsupported VT!");
case MVT::i32: Opc = hasBMI2 ? X86::MULX32rr : X86::MUL32r;
MOpc = hasBMI2 ? X86::MULX32rm : X86::MUL32m; break;
case MVT::i64: Opc = hasBMI2 ? X86::MULX64rr : X86::MUL64r;
MOpc = hasBMI2 ? X86::MULX64rm : X86::MUL64m; break;
}
} else {
switch (NVT.SimpleTy) {
default: llvm_unreachable("Unsupported VT!");
case MVT::i32: Opc = X86::IMUL32r; MOpc = X86::IMUL32m; break;
case MVT::i64: Opc = X86::IMUL64r; MOpc = X86::IMUL64m; break;
}
}
unsigned SrcReg, LoReg, HiReg;
switch (Opc) {
default: llvm_unreachable("Unknown MUL opcode!");
case X86::IMUL32r:
case X86::MUL32r:
SrcReg = LoReg = X86::EAX; HiReg = X86::EDX;
break;
case X86::IMUL64r:
case X86::MUL64r:
SrcReg = LoReg = X86::RAX; HiReg = X86::RDX;
break;
case X86::MULX32rr:
SrcReg = X86::EDX; LoReg = HiReg = 0;
break;
case X86::MULX64rr:
SrcReg = X86::RDX; LoReg = HiReg = 0;
break;
}
SDValue Tmp0, Tmp1, Tmp2, Tmp3, Tmp4;
bool foldedLoad = tryFoldLoad(Node, N1, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4);
// Multiply is commmutative.
if (!foldedLoad) {
foldedLoad = tryFoldLoad(Node, N0, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4);
if (foldedLoad)
std::swap(N0, N1);
}
SDValue InFlag = CurDAG->getCopyToReg(CurDAG->getEntryNode(), dl, SrcReg,
N0, SDValue()).getValue(1);
SDValue ResHi, ResLo;
if (foldedLoad) {
SDValue Chain;
MachineSDNode *CNode = nullptr;
SDValue Ops[] = { Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, N1.getOperand(0),
InFlag };
if (MOpc == X86::MULX32rm || MOpc == X86::MULX64rm) {
SDVTList VTs = CurDAG->getVTList(NVT, NVT, MVT::Other, MVT::Glue);
CNode = CurDAG->getMachineNode(MOpc, dl, VTs, Ops);
ResHi = SDValue(CNode, 0);
ResLo = SDValue(CNode, 1);
Chain = SDValue(CNode, 2);
InFlag = SDValue(CNode, 3);
} else {
SDVTList VTs = CurDAG->getVTList(MVT::Other, MVT::Glue);
CNode = CurDAG->getMachineNode(MOpc, dl, VTs, Ops);
Chain = SDValue(CNode, 0);
InFlag = SDValue(CNode, 1);
}
// Update the chain.
ReplaceUses(N1.getValue(1), Chain);
// Record the mem-refs
CurDAG->setNodeMemRefs(CNode, {cast<LoadSDNode>(N1)->getMemOperand()});
} else {
SDValue Ops[] = { N1, InFlag };
if (Opc == X86::MULX32rr || Opc == X86::MULX64rr) {
SDVTList VTs = CurDAG->getVTList(NVT, NVT, MVT::Glue);
SDNode *CNode = CurDAG->getMachineNode(Opc, dl, VTs, Ops);
ResHi = SDValue(CNode, 0);
ResLo = SDValue(CNode, 1);
InFlag = SDValue(CNode, 2);
} else {
SDVTList VTs = CurDAG->getVTList(MVT::Glue);
SDNode *CNode = CurDAG->getMachineNode(Opc, dl, VTs, Ops);
InFlag = SDValue(CNode, 0);
}
}
// Copy the low half of the result, if it is needed.
if (!SDValue(Node, 0).use_empty()) {
if (!ResLo.getNode()) {
assert(LoReg && "Register for low half is not defined!");
ResLo = CurDAG->getCopyFromReg(CurDAG->getEntryNode(), dl, LoReg, NVT,
InFlag);
InFlag = ResLo.getValue(2);
}
ReplaceUses(SDValue(Node, 0), ResLo);
LLVM_DEBUG(dbgs() << "=> "; ResLo.getNode()->dump(CurDAG);
dbgs() << '\n');
}
// Copy the high half of the result, if it is needed.
if (!SDValue(Node, 1).use_empty()) {
if (!ResHi.getNode()) {
assert(HiReg && "Register for high half is not defined!");
ResHi = CurDAG->getCopyFromReg(CurDAG->getEntryNode(), dl, HiReg, NVT,
InFlag);
InFlag = ResHi.getValue(2);
}
ReplaceUses(SDValue(Node, 1), ResHi);
LLVM_DEBUG(dbgs() << "=> "; ResHi.getNode()->dump(CurDAG);
dbgs() << '\n');
}
CurDAG->RemoveDeadNode(Node);
return;
}
case ISD::SDIVREM:
case ISD::UDIVREM: {
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
unsigned Opc, MOpc;
bool isSigned = Opcode == ISD::SDIVREM;
if (!isSigned) {
switch (NVT.SimpleTy) {
default: llvm_unreachable("Unsupported VT!");
case MVT::i8: Opc = X86::DIV8r; MOpc = X86::DIV8m; break;
case MVT::i16: Opc = X86::DIV16r; MOpc = X86::DIV16m; break;
case MVT::i32: Opc = X86::DIV32r; MOpc = X86::DIV32m; break;
case MVT::i64: Opc = X86::DIV64r; MOpc = X86::DIV64m; break;
}
} else {
switch (NVT.SimpleTy) {
default: llvm_unreachable("Unsupported VT!");
case MVT::i8: Opc = X86::IDIV8r; MOpc = X86::IDIV8m; break;
case MVT::i16: Opc = X86::IDIV16r; MOpc = X86::IDIV16m; break;
case MVT::i32: Opc = X86::IDIV32r; MOpc = X86::IDIV32m; break;
case MVT::i64: Opc = X86::IDIV64r; MOpc = X86::IDIV64m; break;
}
}
unsigned LoReg, HiReg, ClrReg;
unsigned SExtOpcode;
switch (NVT.SimpleTy) {
default: llvm_unreachable("Unsupported VT!");
case MVT::i8:
LoReg = X86::AL; ClrReg = HiReg = X86::AH;
SExtOpcode = X86::CBW;
break;
case MVT::i16:
LoReg = X86::AX; HiReg = X86::DX;
ClrReg = X86::DX;
SExtOpcode = X86::CWD;
break;
case MVT::i32:
LoReg = X86::EAX; ClrReg = HiReg = X86::EDX;
SExtOpcode = X86::CDQ;
break;
case MVT::i64:
LoReg = X86::RAX; ClrReg = HiReg = X86::RDX;
SExtOpcode = X86::CQO;
break;
}
SDValue Tmp0, Tmp1, Tmp2, Tmp3, Tmp4;
bool foldedLoad = tryFoldLoad(Node, N1, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4);
bool signBitIsZero = CurDAG->SignBitIsZero(N0);
SDValue InFlag;
if (NVT == MVT::i8 && (!isSigned || signBitIsZero)) {
// Special case for div8, just use a move with zero extension to AX to
// clear the upper 8 bits (AH).
SDValue Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, Chain;
MachineSDNode *Move;
if (tryFoldLoad(Node, N0, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4)) {
SDValue Ops[] = { Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, N0.getOperand(0) };
Move = CurDAG->getMachineNode(X86::MOVZX32rm8, dl, MVT::i32,
MVT::Other, Ops);
Chain = SDValue(Move, 1);
ReplaceUses(N0.getValue(1), Chain);
// Record the mem-refs
CurDAG->setNodeMemRefs(Move, {cast<LoadSDNode>(N0)->getMemOperand()});
} else {
Move = CurDAG->getMachineNode(X86::MOVZX32rr8, dl, MVT::i32, N0);
Chain = CurDAG->getEntryNode();
}
Chain = CurDAG->getCopyToReg(Chain, dl, X86::EAX, SDValue(Move, 0),
SDValue());
InFlag = Chain.getValue(1);
} else {
InFlag =
CurDAG->getCopyToReg(CurDAG->getEntryNode(), dl,
LoReg, N0, SDValue()).getValue(1);
if (isSigned && !signBitIsZero) {
// Sign extend the low part into the high part.
InFlag =
SDValue(CurDAG->getMachineNode(SExtOpcode, dl, MVT::Glue, InFlag),0);
} else {
// Zero out the high part, effectively zero extending the input.
SDValue ClrNode = SDValue(CurDAG->getMachineNode(X86::MOV32r0, dl, NVT), 0);
switch (NVT.SimpleTy) {
case MVT::i16:
ClrNode =
SDValue(CurDAG->getMachineNode(
TargetOpcode::EXTRACT_SUBREG, dl, MVT::i16, ClrNode,
CurDAG->getTargetConstant(X86::sub_16bit, dl,
MVT::i32)),
0);
break;
case MVT::i32:
break;
case MVT::i64:
ClrNode =
SDValue(CurDAG->getMachineNode(
TargetOpcode::SUBREG_TO_REG, dl, MVT::i64,
CurDAG->getTargetConstant(0, dl, MVT::i64), ClrNode,
CurDAG->getTargetConstant(X86::sub_32bit, dl,
MVT::i32)),
0);
break;
default:
llvm_unreachable("Unexpected division source");
}
InFlag = CurDAG->getCopyToReg(CurDAG->getEntryNode(), dl, ClrReg,
ClrNode, InFlag).getValue(1);
}
}
if (foldedLoad) {
SDValue Ops[] = { Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, N1.getOperand(0),
InFlag };
MachineSDNode *CNode =
CurDAG->getMachineNode(MOpc, dl, MVT::Other, MVT::Glue, Ops);
InFlag = SDValue(CNode, 1);
// Update the chain.
ReplaceUses(N1.getValue(1), SDValue(CNode, 0));
// Record the mem-refs
CurDAG->setNodeMemRefs(CNode, {cast<LoadSDNode>(N1)->getMemOperand()});
} else {
InFlag =
SDValue(CurDAG->getMachineNode(Opc, dl, MVT::Glue, N1, InFlag), 0);
}
// Prevent use of AH in a REX instruction by explicitly copying it to
// an ABCD_L register.
//
// The current assumption of the register allocator is that isel
// won't generate explicit references to the GR8_ABCD_H registers. If
// the allocator and/or the backend get enhanced to be more robust in
// that regard, this can be, and should be, removed.
if (HiReg == X86::AH && !SDValue(Node, 1).use_empty()) {
SDValue AHCopy = CurDAG->getRegister(X86::AH, MVT::i8);
unsigned AHExtOpcode =
isSigned ? X86::MOVSX32rr8_NOREX : X86::MOVZX32rr8_NOREX;
SDNode *RNode = CurDAG->getMachineNode(AHExtOpcode, dl, MVT::i32,
MVT::Glue, AHCopy, InFlag);
SDValue Result(RNode, 0);
InFlag = SDValue(RNode, 1);
Result =
CurDAG->getTargetExtractSubreg(X86::sub_8bit, dl, MVT::i8, Result);
ReplaceUses(SDValue(Node, 1), Result);
LLVM_DEBUG(dbgs() << "=> "; Result.getNode()->dump(CurDAG);
dbgs() << '\n');
}
// Copy the division (low) result, if it is needed.
if (!SDValue(Node, 0).use_empty()) {
SDValue Result = CurDAG->getCopyFromReg(CurDAG->getEntryNode(), dl,
LoReg, NVT, InFlag);
InFlag = Result.getValue(2);
ReplaceUses(SDValue(Node, 0), Result);
LLVM_DEBUG(dbgs() << "=> "; Result.getNode()->dump(CurDAG);
dbgs() << '\n');
}
// Copy the remainder (high) result, if it is needed.
if (!SDValue(Node, 1).use_empty()) {
SDValue Result = CurDAG->getCopyFromReg(CurDAG->getEntryNode(), dl,
HiReg, NVT, InFlag);
InFlag = Result.getValue(2);
ReplaceUses(SDValue(Node, 1), Result);
LLVM_DEBUG(dbgs() << "=> "; Result.getNode()->dump(CurDAG);
dbgs() << '\n');
}
CurDAG->RemoveDeadNode(Node);
return;
}
case X86ISD::CMP: {
SDValue N0 = Node->getOperand(0);
SDValue N1 = Node->getOperand(1);
// Save the original VT of the compare.
MVT CmpVT = N0.getSimpleValueType();
// If we are comparing (and (shr X, C, Mask) with 0, emit a BEXTR followed
// by a test instruction. The test should be removed later by
// analyzeCompare if we are using only the zero flag.
// TODO: Should we check the users and use the BEXTR flags directly?
if (isNullConstant(N1) && N0.getOpcode() == ISD::AND && N0.hasOneUse()) {
if (MachineSDNode *NewNode = matchBEXTRFromAndImm(N0.getNode())) {
unsigned TestOpc = CmpVT == MVT::i64 ? X86::TEST64rr
: X86::TEST32rr;
SDValue BEXTR = SDValue(NewNode, 0);
NewNode = CurDAG->getMachineNode(TestOpc, dl, MVT::i32, BEXTR, BEXTR);
ReplaceUses(SDValue(Node, 0), SDValue(NewNode, 0));
CurDAG->RemoveDeadNode(Node);
return;
}
}
// We can peek through truncates, but we need to be careful below.
if (N0.getOpcode() == ISD::TRUNCATE && N0.hasOneUse())
N0 = N0.getOperand(0);
// Look for (X86cmp (and $op, $imm), 0) and see if we can convert it to
// use a smaller encoding.
// Look past the truncate if CMP is the only use of it.
if (N0.getOpcode() == ISD::AND &&
N0.getNode()->hasOneUse() &&
N0.getValueType() != MVT::i8 &&
isNullConstant(N1)) {
ConstantSDNode *C = dyn_cast<ConstantSDNode>(N0.getOperand(1));
if (!C) break;
uint64_t Mask = C->getZExtValue();
MVT VT;
int SubRegOp;
unsigned ROpc, MOpc;
// For each of these checks we need to be careful if the sign flag is
// being used. It is only safe to use the sign flag in two conditions,
// either the sign bit in the shrunken mask is zero or the final test
// size is equal to the original compare size.
if (isUInt<8>(Mask) &&
(!(Mask & 0x80) || CmpVT == MVT::i8 ||
hasNoSignedComparisonUses(Node))) {
// For example, convert "testl %eax, $8" to "testb %al, $8"
VT = MVT::i8;
SubRegOp = X86::sub_8bit;
ROpc = X86::TEST8ri;
MOpc = X86::TEST8mi;
} else if (OptForMinSize && isUInt<16>(Mask) &&
(!(Mask & 0x8000) || CmpVT == MVT::i16 ||
hasNoSignedComparisonUses(Node))) {
// For example, "testl %eax, $32776" to "testw %ax, $32776".
// NOTE: We only want to form TESTW instructions if optimizing for
// min size. Otherwise we only save one byte and possibly get a length
// changing prefix penalty in the decoders.
VT = MVT::i16;
SubRegOp = X86::sub_16bit;
ROpc = X86::TEST16ri;
MOpc = X86::TEST16mi;
} else if (isUInt<32>(Mask) && N0.getValueType() != MVT::i16 &&
((!(Mask & 0x80000000) &&
// Without minsize 16-bit Cmps can get here so we need to
// be sure we calculate the correct sign flag if needed.
(CmpVT != MVT::i16 || !(Mask & 0x8000))) ||
CmpVT == MVT::i32 || hasNoSignedComparisonUses(Node))) {
// For example, "testq %rax, $268468232" to "testl %eax, $268468232".
// NOTE: We only want to run that transform if N0 is 32 or 64 bits.
// Otherwize, we find ourselves in a position where we have to do
// promotion. If previous passes did not promote the and, we assume
// they had a good reason not to and do not promote here.
VT = MVT::i32;
SubRegOp = X86::sub_32bit;
ROpc = X86::TEST32ri;
MOpc = X86::TEST32mi;
} else {
// No eligible transformation was found.
break;
}
// FIXME: We should be able to fold loads here.
SDValue Imm = CurDAG->getTargetConstant(Mask, dl, VT);
SDValue Reg = N0.getOperand(0);
// Emit a testl or testw.
MachineSDNode *NewNode;
SDValue Tmp0, Tmp1, Tmp2, Tmp3, Tmp4;
if (tryFoldLoad(Node, N0.getNode(), Reg, Tmp0, Tmp1, Tmp2, Tmp3, Tmp4)) {
SDValue Ops[] = { Tmp0, Tmp1, Tmp2, Tmp3, Tmp4, Imm,
Reg.getOperand(0) };
NewNode = CurDAG->getMachineNode(MOpc, dl, MVT::i32, MVT::Other, Ops);
// Update the chain.
ReplaceUses(Reg.getValue(1), SDValue(NewNode, 1));
// Record the mem-refs
CurDAG->setNodeMemRefs(NewNode,
{cast<LoadSDNode>(Reg)->getMemOperand()});
} else {
// Extract the subregister if necessary.
if (N0.getValueType() != VT)
Reg = CurDAG->getTargetExtractSubreg(SubRegOp, dl, VT, Reg);
NewNode = CurDAG->getMachineNode(ROpc, dl, MVT::i32, Reg, Imm);
}
// Replace CMP with TEST.
ReplaceNode(Node, NewNode);
return;
}
break;
}
case X86ISD::PCMPISTR: {
if (!Subtarget->hasSSE42())
break;
bool NeedIndex = !SDValue(Node, 0).use_empty();
bool NeedMask = !SDValue(Node, 1).use_empty();
// We can't fold a load if we are going to make two instructions.
bool MayFoldLoad = !NeedIndex || !NeedMask;
MachineSDNode *CNode;
if (NeedMask) {
unsigned ROpc = Subtarget->hasAVX() ? X86::VPCMPISTRMrr : X86::PCMPISTRMrr;
unsigned MOpc = Subtarget->hasAVX() ? X86::VPCMPISTRMrm : X86::PCMPISTRMrm;
CNode = emitPCMPISTR(ROpc, MOpc, MayFoldLoad, dl, MVT::v16i8, Node);
ReplaceUses(SDValue(Node, 1), SDValue(CNode, 0));
}
if (NeedIndex || !NeedMask) {
unsigned ROpc = Subtarget->hasAVX() ? X86::VPCMPISTRIrr : X86::PCMPISTRIrr;
unsigned MOpc = Subtarget->hasAVX() ? X86::VPCMPISTRIrm : X86::PCMPISTRIrm;
CNode = emitPCMPISTR(ROpc, MOpc, MayFoldLoad, dl, MVT::i32, Node);
ReplaceUses(SDValue(Node, 0), SDValue(CNode, 0));
}
// Connect the flag usage to the last instruction created.
ReplaceUses(SDValue(Node, 2), SDValue(CNode, 1));
CurDAG->RemoveDeadNode(Node);
return;
}
case X86ISD::PCMPESTR: {
if (!Subtarget->hasSSE42())
break;
// Copy the two implicit register inputs.
SDValue InFlag = CurDAG->getCopyToReg(CurDAG->getEntryNode(), dl, X86::EAX,
Node->getOperand(1),
SDValue()).getValue(1);
InFlag = CurDAG->getCopyToReg(CurDAG->getEntryNode(), dl, X86::EDX,
Node->getOperand(3), InFlag).getValue(1);
bool NeedIndex = !SDValue(Node, 0).use_empty();
bool NeedMask = !SDValue(Node, 1).use_empty();
// We can't fold a load if we are going to make two instructions.
bool MayFoldLoad = !NeedIndex || !NeedMask;
MachineSDNode *CNode;
if (NeedMask) {
unsigned ROpc = Subtarget->hasAVX() ? X86::VPCMPESTRMrr : X86::PCMPESTRMrr;
unsigned MOpc = Subtarget->hasAVX() ? X86::VPCMPESTRMrm : X86::PCMPESTRMrm;
CNode = emitPCMPESTR(ROpc, MOpc, MayFoldLoad, dl, MVT::v16i8, Node,
InFlag);
ReplaceUses(SDValue(Node, 1), SDValue(CNode, 0));
}
if (NeedIndex || !NeedMask) {
unsigned ROpc = Subtarget->hasAVX() ? X86::VPCMPESTRIrr : X86::PCMPESTRIrr;
unsigned MOpc = Subtarget->hasAVX() ? X86::VPCMPESTRIrm : X86::PCMPESTRIrm;
CNode = emitPCMPESTR(ROpc, MOpc, MayFoldLoad, dl, MVT::i32, Node, InFlag);
ReplaceUses(SDValue(Node, 0), SDValue(CNode, 0));
}
// Connect the flag usage to the last instruction created.
ReplaceUses(SDValue(Node, 2), SDValue(CNode, 1));
CurDAG->RemoveDeadNode(Node);
return;
}
case ISD::STORE:
if (foldLoadStoreIntoMemOperand(Node))
return;
break;
}
SelectCode(Node);
}
bool X86DAGToDAGISel::
SelectInlineAsmMemoryOperand(const SDValue &Op, unsigned ConstraintID,
std::vector<SDValue> &OutOps) {
SDValue Op0, Op1, Op2, Op3, Op4;
switch (ConstraintID) {
default:
llvm_unreachable("Unexpected asm memory constraint");
case InlineAsm::Constraint_i:
// FIXME: It seems strange that 'i' is needed here since it's supposed to
// be an immediate and not a memory constraint.
LLVM_FALLTHROUGH;
case InlineAsm::Constraint_o: // offsetable ??
case InlineAsm::Constraint_v: // not offsetable ??
case InlineAsm::Constraint_m: // memory
case InlineAsm::Constraint_X:
if (!selectAddr(nullptr, Op, Op0, Op1, Op2, Op3, Op4))
return true;
break;
}
OutOps.push_back(Op0);
OutOps.push_back(Op1);
OutOps.push_back(Op2);
OutOps.push_back(Op3);
OutOps.push_back(Op4);
return false;
}
/// This pass converts a legalized DAG into a X86-specific DAG,
/// ready for instruction scheduling.
FunctionPass *llvm::createX86ISelDag(X86TargetMachine &TM,
CodeGenOpt::Level OptLevel) {
return new X86DAGToDAGISel(TM, OptLevel);
}
Reference in New Issue View Git Blame Copy Permalink