blob: a51aa85a931c0e7e9325a11454f46e2561936906 [file] [log] [blame]
//===-- AArch64ISelDAGToDAG.cpp - A dag to dag inst selector for AArch64 --===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file defines an instruction selector for the AArch64 target.
//
//===----------------------------------------------------------------------===//
#include "AArch64TargetMachine.h"
#include "MCTargetDesc/AArch64AddressingModes.h"
#include "llvm/ADT/APSInt.h"
#include "llvm/CodeGen/SelectionDAGISel.h"
#include "llvm/IR/Function.h" // To access function attributes.
#include "llvm/IR/GlobalValue.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/IntrinsicsAArch64.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"
using namespace llvm;
#define DEBUG_TYPE "aarch64-isel"
//===--------------------------------------------------------------------===//
/// AArch64DAGToDAGISel - AArch64 specific code to select AArch64 machine
/// instructions for SelectionDAG operations.
///
namespace {
class AArch64DAGToDAGISel : public SelectionDAGISel {
/// Subtarget - Keep a pointer to the AArch64Subtarget around so that we can
/// make the right decision when generating code for different targets.
const AArch64Subtarget *Subtarget;
public:
explicit AArch64DAGToDAGISel(AArch64TargetMachine &tm,
CodeGenOpt::Level OptLevel)
: SelectionDAGISel(tm, OptLevel), Subtarget(nullptr) {}
StringRef getPassName() const override {
return "AArch64 Instruction Selection";
}
bool runOnMachineFunction(MachineFunction &MF) override {
Subtarget = &MF.getSubtarget<AArch64Subtarget>();
return SelectionDAGISel::runOnMachineFunction(MF);
}
void Select(SDNode *Node) override;
/// SelectInlineAsmMemoryOperand - Implement addressing mode selection for
/// inline asm expressions.
bool SelectInlineAsmMemoryOperand(const SDValue &Op,
unsigned ConstraintID,
std::vector<SDValue> &OutOps) override;
bool tryMLAV64LaneV128(SDNode *N);
bool tryMULLV64LaneV128(unsigned IntNo, SDNode *N);
bool SelectArithExtendedRegister(SDValue N, SDValue &Reg, SDValue &Shift);
bool SelectArithImmed(SDValue N, SDValue &Val, SDValue &Shift);
bool SelectNegArithImmed(SDValue N, SDValue &Val, SDValue &Shift);
bool SelectArithShiftedRegister(SDValue N, SDValue &Reg, SDValue &Shift) {
return SelectShiftedRegister(N, false, Reg, Shift);
}
bool SelectLogicalShiftedRegister(SDValue N, SDValue &Reg, SDValue &Shift) {
return SelectShiftedRegister(N, true, Reg, Shift);
}
bool SelectAddrModeIndexed7S8(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed7S(N, 1, Base, OffImm);
}
bool SelectAddrModeIndexed7S16(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed7S(N, 2, Base, OffImm);
}
bool SelectAddrModeIndexed7S32(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed7S(N, 4, Base, OffImm);
}
bool SelectAddrModeIndexed7S64(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed7S(N, 8, Base, OffImm);
}
bool SelectAddrModeIndexed7S128(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed7S(N, 16, Base, OffImm);
}
bool SelectAddrModeIndexedS9S128(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexedBitWidth(N, true, 9, 16, Base, OffImm);
}
bool SelectAddrModeIndexedU6S128(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexedBitWidth(N, false, 6, 16, Base, OffImm);
}
bool SelectAddrModeIndexed8(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed(N, 1, Base, OffImm);
}
bool SelectAddrModeIndexed16(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed(N, 2, Base, OffImm);
}
bool SelectAddrModeIndexed32(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed(N, 4, Base, OffImm);
}
bool SelectAddrModeIndexed64(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed(N, 8, Base, OffImm);
}
bool SelectAddrModeIndexed128(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeIndexed(N, 16, Base, OffImm);
}
bool SelectAddrModeUnscaled8(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeUnscaled(N, 1, Base, OffImm);
}
bool SelectAddrModeUnscaled16(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeUnscaled(N, 2, Base, OffImm);
}
bool SelectAddrModeUnscaled32(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeUnscaled(N, 4, Base, OffImm);
}
bool SelectAddrModeUnscaled64(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeUnscaled(N, 8, Base, OffImm);
}
bool SelectAddrModeUnscaled128(SDValue N, SDValue &Base, SDValue &OffImm) {
return SelectAddrModeUnscaled(N, 16, Base, OffImm);
}
template<int Width>
bool SelectAddrModeWRO(SDValue N, SDValue &Base, SDValue &Offset,
SDValue &SignExtend, SDValue &DoShift) {
return SelectAddrModeWRO(N, Width / 8, Base, Offset, SignExtend, DoShift);
}
template<int Width>
bool SelectAddrModeXRO(SDValue N, SDValue &Base, SDValue &Offset,
SDValue &SignExtend, SDValue &DoShift) {
return SelectAddrModeXRO(N, Width / 8, Base, Offset, SignExtend, DoShift);
}
bool SelectDupZeroOrUndef(SDValue N) {
switch(N->getOpcode()) {
case ISD::UNDEF:
return true;
case AArch64ISD::DUP:
case ISD::SPLAT_VECTOR: {
auto Opnd0 = N->getOperand(0);
if (auto CN = dyn_cast<ConstantSDNode>(Opnd0))
if (CN->isNullValue())
return true;
if (auto CN = dyn_cast<ConstantFPSDNode>(Opnd0))
if (CN->isZero())
return true;
break;
}
default:
break;
}
return false;
}
template<MVT::SimpleValueType VT>
bool SelectSVEAddSubImm(SDValue N, SDValue &Imm, SDValue &Shift) {
return SelectSVEAddSubImm(N, VT, Imm, Shift);
}
template<MVT::SimpleValueType VT>
bool SelectSVELogicalImm(SDValue N, SDValue &Imm) {
return SelectSVELogicalImm(N, VT, Imm);
}
// Returns a suitable CNT/INC/DEC/RDVL multiplier to calculate VSCALE*N.
template<signed Min, signed Max, signed Scale, bool Shift>
bool SelectCntImm(SDValue N, SDValue &Imm) {
if (!isa<ConstantSDNode>(N))
return false;
int64_t MulImm = cast<ConstantSDNode>(N)->getSExtValue();
if (Shift)
MulImm = 1LL << MulImm;
if ((MulImm % std::abs(Scale)) != 0)
return false;
MulImm /= Scale;
if ((MulImm >= Min) && (MulImm <= Max)) {
Imm = CurDAG->getTargetConstant(MulImm, SDLoc(N), MVT::i32);
return true;
}
return false;
}
/// Form sequences of consecutive 64/128-bit registers for use in NEON
/// instructions making use of a vector-list (e.g. ldN, tbl). Vecs must have
/// between 1 and 4 elements. If it contains a single element that is returned
/// unchanged; otherwise a REG_SEQUENCE value is returned.
SDValue createDTuple(ArrayRef<SDValue> Vecs);
SDValue createQTuple(ArrayRef<SDValue> Vecs);
/// Generic helper for the createDTuple/createQTuple
/// functions. Those should almost always be called instead.
SDValue createTuple(ArrayRef<SDValue> Vecs, const unsigned RegClassIDs[],
const unsigned SubRegs[]);
void SelectTable(SDNode *N, unsigned NumVecs, unsigned Opc, bool isExt);
bool tryIndexedLoad(SDNode *N);
bool trySelectStackSlotTagP(SDNode *N);
void SelectTagP(SDNode *N);
void SelectLoad(SDNode *N, unsigned NumVecs, unsigned Opc,
unsigned SubRegIdx);
void SelectPostLoad(SDNode *N, unsigned NumVecs, unsigned Opc,
unsigned SubRegIdx);
void SelectLoadLane(SDNode *N, unsigned NumVecs, unsigned Opc);
void SelectPostLoadLane(SDNode *N, unsigned NumVecs, unsigned Opc);
void SelectStore(SDNode *N, unsigned NumVecs, unsigned Opc);
void SelectPostStore(SDNode *N, unsigned NumVecs, unsigned Opc);
void SelectStoreLane(SDNode *N, unsigned NumVecs, unsigned Opc);
void SelectPostStoreLane(SDNode *N, unsigned NumVecs, unsigned Opc);
bool tryBitfieldExtractOp(SDNode *N);
bool tryBitfieldExtractOpFromSExt(SDNode *N);
bool tryBitfieldInsertOp(SDNode *N);
bool tryBitfieldInsertInZeroOp(SDNode *N);
bool tryShiftAmountMod(SDNode *N);
bool tryHighFPExt(SDNode *N);
bool tryReadRegister(SDNode *N);
bool tryWriteRegister(SDNode *N);
// Include the pieces autogenerated from the target description.
#include "AArch64GenDAGISel.inc"
private:
bool SelectShiftedRegister(SDValue N, bool AllowROR, SDValue &Reg,
SDValue &Shift);
bool SelectAddrModeIndexed7S(SDValue N, unsigned Size, SDValue &Base,
SDValue &OffImm) {
return SelectAddrModeIndexedBitWidth(N, true, 7, Size, Base, OffImm);
}
bool SelectAddrModeIndexedBitWidth(SDValue N, bool IsSignedImm, unsigned BW,
unsigned Size, SDValue &Base,
SDValue &OffImm);
bool SelectAddrModeIndexed(SDValue N, unsigned Size, SDValue &Base,
SDValue &OffImm);
bool SelectAddrModeUnscaled(SDValue N, unsigned Size, SDValue &Base,
SDValue &OffImm);
bool SelectAddrModeWRO(SDValue N, unsigned Size, SDValue &Base,
SDValue &Offset, SDValue &SignExtend,
SDValue &DoShift);
bool SelectAddrModeXRO(SDValue N, unsigned Size, SDValue &Base,
SDValue &Offset, SDValue &SignExtend,
SDValue &DoShift);
bool isWorthFolding(SDValue V) const;
bool SelectExtendedSHL(SDValue N, unsigned Size, bool WantExtend,
SDValue &Offset, SDValue &SignExtend);
template<unsigned RegWidth>
bool SelectCVTFixedPosOperand(SDValue N, SDValue &FixedPos) {
return SelectCVTFixedPosOperand(N, FixedPos, RegWidth);
}
bool SelectCVTFixedPosOperand(SDValue N, SDValue &FixedPos, unsigned Width);
bool SelectCMP_SWAP(SDNode *N);
bool SelectSVEAddSubImm(SDValue N, MVT VT, SDValue &Imm, SDValue &Shift);
bool SelectSVELogicalImm(SDValue N, MVT VT, SDValue &Imm);
bool SelectSVESignedArithImm(SDValue N, SDValue &Imm);
bool SelectSVEArithImm(SDValue N, SDValue &Imm);
};
} // end anonymous namespace
/// isIntImmediate - This method tests to see if the node is a constant
/// operand. If so Imm will receive the 32-bit value.
static bool isIntImmediate(const SDNode *N, uint64_t &Imm) {
if (const ConstantSDNode *C = dyn_cast<const ConstantSDNode>(N)) {
Imm = C->getZExtValue();
return true;
}
return false;
}
// isIntImmediate - This method tests to see if a constant operand.
// If so Imm will receive the value.
static bool isIntImmediate(SDValue N, uint64_t &Imm) {
return isIntImmediate(N.getNode(), Imm);
}
// isOpcWithIntImmediate - This method tests to see if the node is a specific
// opcode and that it has a immediate integer right operand.
// If so Imm will receive the 32 bit value.
static bool isOpcWithIntImmediate(const SDNode *N, unsigned Opc,
uint64_t &Imm) {
return N->getOpcode() == Opc &&
isIntImmediate(N->getOperand(1).getNode(), Imm);
}
bool AArch64DAGToDAGISel::SelectInlineAsmMemoryOperand(
const SDValue &Op, unsigned ConstraintID, std::vector<SDValue> &OutOps) {
switch(ConstraintID) {
default:
llvm_unreachable("Unexpected asm memory constraint");
case InlineAsm::Constraint_m:
case InlineAsm::Constraint_Q:
// We need to make sure that this one operand does not end up in XZR, thus
// require the address to be in a PointerRegClass register.
const TargetRegisterInfo *TRI = Subtarget->getRegisterInfo();
const TargetRegisterClass *TRC = TRI->getPointerRegClass(*MF);
SDLoc dl(Op);
SDValue RC = CurDAG->getTargetConstant(TRC->getID(), dl, MVT::i64);
SDValue NewOp =
SDValue(CurDAG->getMachineNode(TargetOpcode::COPY_TO_REGCLASS,
dl, Op.getValueType(),
Op, RC), 0);
OutOps.push_back(NewOp);
return false;
}
return true;
}
/// SelectArithImmed - Select an immediate value that can be represented as
/// a 12-bit value shifted left by either 0 or 12. If so, return true with
/// Val set to the 12-bit value and Shift set to the shifter operand.
bool AArch64DAGToDAGISel::SelectArithImmed(SDValue N, SDValue &Val,
SDValue &Shift) {
// This function is called from the addsub_shifted_imm ComplexPattern,
// which lists [imm] as the list of opcode it's interested in, however
// we still need to check whether the operand is actually an immediate
// here because the ComplexPattern opcode list is only used in
// root-level opcode matching.
if (!isa<ConstantSDNode>(N.getNode()))
return false;
uint64_t Immed = cast<ConstantSDNode>(N.getNode())->getZExtValue();
unsigned ShiftAmt;
if (Immed >> 12 == 0) {
ShiftAmt = 0;
} else if ((Immed & 0xfff) == 0 && Immed >> 24 == 0) {
ShiftAmt = 12;
Immed = Immed >> 12;
} else
return false;
unsigned ShVal = AArch64_AM::getShifterImm(AArch64_AM::LSL, ShiftAmt);
SDLoc dl(N);
Val = CurDAG->getTargetConstant(Immed, dl, MVT::i32);
Shift = CurDAG->getTargetConstant(ShVal, dl, MVT::i32);
return true;
}
/// SelectNegArithImmed - As above, but negates the value before trying to
/// select it.
bool AArch64DAGToDAGISel::SelectNegArithImmed(SDValue N, SDValue &Val,
SDValue &Shift) {
// This function is called from the addsub_shifted_imm ComplexPattern,
// which lists [imm] as the list of opcode it's interested in, however
// we still need to check whether the operand is actually an immediate
// here because the ComplexPattern opcode list is only used in
// root-level opcode matching.
if (!isa<ConstantSDNode>(N.getNode()))
return false;
// The immediate operand must be a 24-bit zero-extended immediate.
uint64_t Immed = cast<ConstantSDNode>(N.getNode())->getZExtValue();
// This negation is almost always valid, but "cmp wN, #0" and "cmn wN, #0"
// have the opposite effect on the C flag, so this pattern mustn't match under
// those circumstances.
if (Immed == 0)
return false;
if (N.getValueType() == MVT::i32)
Immed = ~((uint32_t)Immed) + 1;
else
Immed = ~Immed + 1ULL;
if (Immed & 0xFFFFFFFFFF000000ULL)
return false;
Immed &= 0xFFFFFFULL;
return SelectArithImmed(CurDAG->getConstant(Immed, SDLoc(N), MVT::i32), Val,
Shift);
}
/// getShiftTypeForNode - Translate a shift node to the corresponding
/// ShiftType value.
static AArch64_AM::ShiftExtendType getShiftTypeForNode(SDValue N) {
switch (N.getOpcode()) {
default:
return AArch64_AM::InvalidShiftExtend;
case ISD::SHL:
return AArch64_AM::LSL;
case ISD::SRL:
return AArch64_AM::LSR;
case ISD::SRA:
return AArch64_AM::ASR;
case ISD::ROTR:
return AArch64_AM::ROR;
}
}
/// Determine whether it is worth it to fold SHL into the addressing
/// mode.
static bool isWorthFoldingSHL(SDValue V) {
assert(V.getOpcode() == ISD::SHL && "invalid opcode");
// It is worth folding logical shift of up to three places.
auto *CSD = dyn_cast<ConstantSDNode>(V.getOperand(1));
if (!CSD)
return false;
unsigned ShiftVal = CSD->getZExtValue();
if (ShiftVal > 3)
return false;
// Check if this particular node is reused in any non-memory related
// operation. If yes, do not try to fold this node into the address
// computation, since the computation will be kept.
const SDNode *Node = V.getNode();
for (SDNode *UI : Node->uses())
if (!isa<MemSDNode>(*UI))
for (SDNode *UII : UI->uses())
if (!isa<MemSDNode>(*UII))
return false;
return true;
}
/// Determine whether it is worth to fold V into an extended register.
bool AArch64DAGToDAGISel::isWorthFolding(SDValue V) const {
// Trivial if we are optimizing for code size or if there is only
// one use of the value.
if (CurDAG->shouldOptForSize() || V.hasOneUse())
return true;
// If a subtarget has a fastpath LSL we can fold a logical shift into
// the addressing mode and save a cycle.
if (Subtarget->hasLSLFast() && V.getOpcode() == ISD::SHL &&
isWorthFoldingSHL(V))
return true;
if (Subtarget->hasLSLFast() && V.getOpcode() == ISD::ADD) {
const SDValue LHS = V.getOperand(0);
const SDValue RHS = V.getOperand(1);
if (LHS.getOpcode() == ISD::SHL && isWorthFoldingSHL(LHS))
return true;
if (RHS.getOpcode() == ISD::SHL && isWorthFoldingSHL(RHS))
return true;
}
// It hurts otherwise, since the value will be reused.
return false;
}
/// SelectShiftedRegister - Select a "shifted register" operand. If the value
/// is not shifted, set the Shift operand to default of "LSL 0". The logical
/// instructions allow the shifted register to be rotated, but the arithmetic
/// instructions do not. The AllowROR parameter specifies whether ROR is
/// supported.
bool AArch64DAGToDAGISel::SelectShiftedRegister(SDValue N, bool AllowROR,
SDValue &Reg, SDValue &Shift) {
AArch64_AM::ShiftExtendType ShType = getShiftTypeForNode(N);
if (ShType == AArch64_AM::InvalidShiftExtend)
return false;
if (!AllowROR && ShType == AArch64_AM::ROR)
return false;
if (ConstantSDNode *RHS = dyn_cast<ConstantSDNode>(N.getOperand(1))) {
unsigned BitSize = N.getValueSizeInBits();
unsigned Val = RHS->getZExtValue() & (BitSize - 1);
unsigned ShVal = AArch64_AM::getShifterImm(ShType, Val);
Reg = N.getOperand(0);
Shift = CurDAG->getTargetConstant(ShVal, SDLoc(N), MVT::i32);
return isWorthFolding(N);
}
return false;
}
/// getExtendTypeForNode - Translate an extend node to the corresponding
/// ExtendType value.
static AArch64_AM::ShiftExtendType
getExtendTypeForNode(SDValue N, bool IsLoadStore = false) {
if (N.getOpcode() == ISD::SIGN_EXTEND ||
N.getOpcode() == ISD::SIGN_EXTEND_INREG) {
EVT SrcVT;
if (N.getOpcode() == ISD::SIGN_EXTEND_INREG)
SrcVT = cast<VTSDNode>(N.getOperand(1))->getVT();
else
SrcVT = N.getOperand(0).getValueType();
if (!IsLoadStore && SrcVT == MVT::i8)
return AArch64_AM::SXTB;
else if (!IsLoadStore && SrcVT == MVT::i16)
return AArch64_AM::SXTH;
else if (SrcVT == MVT::i32)
return AArch64_AM::SXTW;
assert(SrcVT != MVT::i64 && "extend from 64-bits?");
return AArch64_AM::InvalidShiftExtend;
} else if (N.getOpcode() == ISD::ZERO_EXTEND ||
N.getOpcode() == ISD::ANY_EXTEND) {
EVT SrcVT = N.getOperand(0).getValueType();
if (!IsLoadStore && SrcVT == MVT::i8)
return AArch64_AM::UXTB;
else if (!IsLoadStore && SrcVT == MVT::i16)
return AArch64_AM::UXTH;
else if (SrcVT == MVT::i32)
return AArch64_AM::UXTW;
assert(SrcVT != MVT::i64 && "extend from 64-bits?");
return AArch64_AM::InvalidShiftExtend;
} else if (N.getOpcode() == ISD::AND) {
ConstantSDNode *CSD = dyn_cast<ConstantSDNode>(N.getOperand(1));
if (!CSD)
return AArch64_AM::InvalidShiftExtend;
uint64_t AndMask = CSD->getZExtValue();
switch (AndMask) {
default:
return AArch64_AM::InvalidShiftExtend;
case 0xFF:
return !IsLoadStore ? AArch64_AM::UXTB : AArch64_AM::InvalidShiftExtend;
case 0xFFFF:
return !IsLoadStore ? AArch64_AM::UXTH : AArch64_AM::InvalidShiftExtend;
case 0xFFFFFFFF:
return AArch64_AM::UXTW;
}
}
return AArch64_AM::InvalidShiftExtend;
}
// Helper for SelectMLAV64LaneV128 - Recognize high lane extracts.
static bool checkHighLaneIndex(SDNode *DL, SDValue &LaneOp, int &LaneIdx) {
if (DL->getOpcode() != AArch64ISD::DUPLANE16 &&
DL->getOpcode() != AArch64ISD::DUPLANE32)
return false;
SDValue SV = DL->getOperand(0);
if (SV.getOpcode() != ISD::INSERT_SUBVECTOR)
return false;
SDValue EV = SV.getOperand(1);
if (EV.getOpcode() != ISD::EXTRACT_SUBVECTOR)
return false;
ConstantSDNode *DLidx = cast<ConstantSDNode>(DL->getOperand(1).getNode());
ConstantSDNode *EVidx = cast<ConstantSDNode>(EV.getOperand(1).getNode());
LaneIdx = DLidx->getSExtValue() + EVidx->getSExtValue();
LaneOp = EV.getOperand(0);
return true;
}
// Helper for SelectOpcV64LaneV128 - Recognize operations where one operand is a
// high lane extract.
static bool checkV64LaneV128(SDValue Op0, SDValue Op1, SDValue &StdOp,
SDValue &LaneOp, int &LaneIdx) {
if (!checkHighLaneIndex(Op0.getNode(), LaneOp, LaneIdx)) {
std::swap(Op0, Op1);
if (!checkHighLaneIndex(Op0.getNode(), LaneOp, LaneIdx))
return false;
}
StdOp = Op1;
return true;
}
/// SelectMLAV64LaneV128 - AArch64 supports vector MLAs where one multiplicand
/// is a lane in the upper half of a 128-bit vector. Recognize and select this
/// so that we don't emit unnecessary lane extracts.
bool AArch64DAGToDAGISel::tryMLAV64LaneV128(SDNode *N) {
SDLoc dl(N);
SDValue Op0 = N->getOperand(0);
SDValue Op1 = N->getOperand(1);
SDValue MLAOp1; // Will hold ordinary multiplicand for MLA.
SDValue MLAOp2; // Will hold lane-accessed multiplicand for MLA.
int LaneIdx = -1; // Will hold the lane index.
if (Op1.getOpcode() != ISD::MUL ||
!checkV64LaneV128(Op1.getOperand(0), Op1.getOperand(1), MLAOp1, MLAOp2,
LaneIdx)) {
std::swap(Op0, Op1);
if (Op1.getOpcode() != ISD::MUL ||
!checkV64LaneV128(Op1.getOperand(0), Op1.getOperand(1), MLAOp1, MLAOp2,
LaneIdx))
return false;
}
SDValue LaneIdxVal = CurDAG->getTargetConstant(LaneIdx, dl, MVT::i64);
SDValue Ops[] = { Op0, MLAOp1, MLAOp2, LaneIdxVal };
unsigned MLAOpc = ~0U;
switch (N->getSimpleValueType(0).SimpleTy) {
default:
llvm_unreachable("Unrecognized MLA.");
case MVT::v4i16:
MLAOpc = AArch64::MLAv4i16_indexed;
break;
case MVT::v8i16:
MLAOpc = AArch64::MLAv8i16_indexed;
break;
case MVT::v2i32:
MLAOpc = AArch64::MLAv2i32_indexed;
break;
case MVT::v4i32:
MLAOpc = AArch64::MLAv4i32_indexed;
break;
}
ReplaceNode(N, CurDAG->getMachineNode(MLAOpc, dl, N->getValueType(0), Ops));
return true;
}
bool AArch64DAGToDAGISel::tryMULLV64LaneV128(unsigned IntNo, SDNode *N) {
SDLoc dl(N);
SDValue SMULLOp0;
SDValue SMULLOp1;
int LaneIdx;
if (!checkV64LaneV128(N->getOperand(1), N->getOperand(2), SMULLOp0, SMULLOp1,
LaneIdx))
return false;
SDValue LaneIdxVal = CurDAG->getTargetConstant(LaneIdx, dl, MVT::i64);
SDValue Ops[] = { SMULLOp0, SMULLOp1, LaneIdxVal };
unsigned SMULLOpc = ~0U;
if (IntNo == Intrinsic::aarch64_neon_smull) {
switch (N->getSimpleValueType(0).SimpleTy) {
default:
llvm_unreachable("Unrecognized SMULL.");
case MVT::v4i32:
SMULLOpc = AArch64::SMULLv4i16_indexed;
break;
case MVT::v2i64:
SMULLOpc = AArch64::SMULLv2i32_indexed;
break;
}
} else if (IntNo == Intrinsic::aarch64_neon_umull) {
switch (N->getSimpleValueType(0).SimpleTy) {
default:
llvm_unreachable("Unrecognized SMULL.");
case MVT::v4i32:
SMULLOpc = AArch64::UMULLv4i16_indexed;
break;
case MVT::v2i64:
SMULLOpc = AArch64::UMULLv2i32_indexed;
break;
}
} else
llvm_unreachable("Unrecognized intrinsic.");
ReplaceNode(N, CurDAG->getMachineNode(SMULLOpc, dl, N->getValueType(0), Ops));
return true;
}
/// Instructions that accept extend modifiers like UXTW expect the register
/// being extended to be a GPR32, but the incoming DAG might be acting on a
/// GPR64 (either via SEXT_INREG or AND). Extract the appropriate low bits if
/// this is the case.
static SDValue narrowIfNeeded(SelectionDAG *CurDAG, SDValue N) {
if (N.getValueType() == MVT::i32)
return N;
SDLoc dl(N);
SDValue SubReg = CurDAG->getTargetConstant(AArch64::sub_32, dl, MVT::i32);
MachineSDNode *Node = CurDAG->getMachineNode(TargetOpcode::EXTRACT_SUBREG,
dl, MVT::i32, N, SubReg);
return SDValue(Node, 0);
}
/// SelectArithExtendedRegister - Select a "extended register" operand. This
/// operand folds in an extend followed by an optional left shift.
bool AArch64DAGToDAGISel::SelectArithExtendedRegister(SDValue N, SDValue &Reg,
SDValue &Shift) {
unsigned ShiftVal = 0;
AArch64_AM::ShiftExtendType Ext;
if (N.getOpcode() == ISD::SHL) {
ConstantSDNode *CSD = dyn_cast<ConstantSDNode>(N.getOperand(1));
if (!CSD)
return false;
ShiftVal = CSD->getZExtValue();
if (ShiftVal > 4)
return false;
Ext = getExtendTypeForNode(N.getOperand(0));
if (Ext == AArch64_AM::InvalidShiftExtend)
return false;
Reg = N.getOperand(0).getOperand(0);
} else {
Ext = getExtendTypeForNode(N);
if (Ext == AArch64_AM::InvalidShiftExtend)
return false;
Reg = N.getOperand(0);
// Don't match if free 32-bit -> 64-bit zext can be used instead.
if (Ext == AArch64_AM::UXTW &&
Reg->getValueType(0).getSizeInBits() == 32 && isDef32(*Reg.getNode()))
return false;
}
// AArch64 mandates that the RHS of the operation must use the smallest
// register class that could contain the size being extended from. Thus,
// if we're folding a (sext i8), we need the RHS to be a GPR32, even though
// there might not be an actual 32-bit value in the program. We can
// (harmlessly) synthesize one by injected an EXTRACT_SUBREG here.
assert(Ext != AArch64_AM::UXTX && Ext != AArch64_AM::SXTX);
Reg = narrowIfNeeded(CurDAG, Reg);
Shift = CurDAG->getTargetConstant(getArithExtendImm(Ext, ShiftVal), SDLoc(N),
MVT::i32);
return isWorthFolding(N);
}
/// If there's a use of this ADDlow that's not itself a load/store then we'll
/// need to create a real ADD instruction from it anyway and there's no point in
/// folding it into the mem op. Theoretically, it shouldn't matter, but there's
/// a single pseudo-instruction for an ADRP/ADD pair so over-aggressive folding
/// leads to duplicated ADRP instructions.
static bool isWorthFoldingADDlow(SDValue N) {
for (auto Use : N->uses()) {
if (Use->getOpcode() != ISD::LOAD && Use->getOpcode() != ISD::STORE &&
Use->getOpcode() != ISD::ATOMIC_LOAD &&
Use->getOpcode() != ISD::ATOMIC_STORE)
return false;
// ldar and stlr have much more restrictive addressing modes (just a
// register).
if (isStrongerThanMonotonic(cast<MemSDNode>(Use)->getOrdering()))
return false;
}
return true;
}
/// SelectAddrModeIndexedBitWidth - Select a "register plus scaled (un)signed BW-bit
/// immediate" address. The "Size" argument is the size in bytes of the memory
/// reference, which determines the scale.
bool AArch64DAGToDAGISel::SelectAddrModeIndexedBitWidth(SDValue N, bool IsSignedImm,
unsigned BW, unsigned Size,
SDValue &Base,
SDValue &OffImm) {
SDLoc dl(N);
const DataLayout &DL = CurDAG->getDataLayout();
const TargetLowering *TLI = getTargetLowering();
if (N.getOpcode() == ISD::FrameIndex) {
int FI = cast<FrameIndexSDNode>(N)->getIndex();
Base = CurDAG->getTargetFrameIndex(FI, TLI->getPointerTy(DL));
OffImm = CurDAG->getTargetConstant(0, dl, MVT::i64);
return true;
}
// As opposed to the (12-bit) Indexed addressing mode below, the 7/9-bit signed
// selected here doesn't support labels/immediates, only base+offset.
if (CurDAG->isBaseWithConstantOffset(N)) {
if (ConstantSDNode *RHS = dyn_cast<ConstantSDNode>(N.getOperand(1))) {
if (IsSignedImm) {
int64_t RHSC = RHS->getSExtValue();
unsigned Scale = Log2_32(Size);
int64_t Range = 0x1LL << (BW - 1);
if ((RHSC & (Size - 1)) == 0 && RHSC >= -(Range << Scale) &&
RHSC < (Range << Scale)) {
Base = N.getOperand(0);
if (Base.getOpcode() == ISD::FrameIndex) {
int FI = cast<FrameIndexSDNode>(Base)->getIndex();
Base = CurDAG->getTargetFrameIndex(FI, TLI->getPointerTy(DL));
}
OffImm = CurDAG->getTargetConstant(RHSC >> Scale, dl, MVT::i64);
return true;
}
} else {
// unsigned Immediate
uint64_t RHSC = RHS->getZExtValue();
unsigned Scale = Log2_32(Size);
uint64_t Range = 0x1ULL << BW;
if ((RHSC & (Size - 1)) == 0 && RHSC < (Range << Scale)) {
Base = N.getOperand(0);
if (Base.getOpcode() == ISD::FrameIndex) {
int FI = cast<FrameIndexSDNode>(Base)->getIndex();
Base = CurDAG->getTargetFrameIndex(FI, TLI->getPointerTy(DL));
}
OffImm = CurDAG->getTargetConstant(RHSC >> Scale, dl, MVT::i64);
return true;
}
}
}
}
// Base only. The address will be materialized into a register before
// the memory is accessed.
// add x0, Xbase, #offset
// stp x1, x2, [x0]
Base = N;
OffImm = CurDAG->getTargetConstant(0, dl, MVT::i64);
return true;
}
/// SelectAddrModeIndexed - Select a "register plus scaled unsigned 12-bit
/// immediate" address. The "Size" argument is the size in bytes of the memory
/// reference, which determines the scale.
bool AArch64DAGToDAGISel::SelectAddrModeIndexed(SDValue N, unsigned Size,
SDValue &Base, SDValue &OffImm) {
SDLoc dl(N);
const DataLayout &DL = CurDAG->getDataLayout();
const TargetLowering *TLI = getTargetLowering();
if (N.getOpcode() == ISD::FrameIndex) {
int FI = cast<FrameIndexSDNode>(N)->getIndex();
Base = CurDAG->getTargetFrameIndex(FI, TLI->getPointerTy(DL));
OffImm = CurDAG->getTargetConstant(0, dl, MVT::i64);
return true;
}
if (N.getOpcode() == AArch64ISD::ADDlow && isWorthFoldingADDlow(N)) {
GlobalAddressSDNode *GAN =
dyn_cast<GlobalAddressSDNode>(N.getOperand(1).getNode());
Base = N.getOperand(0);
OffImm = N.getOperand(1);
if (!GAN)
return true;
if (GAN->getOffset() % Size == 0) {
const GlobalValue *GV = GAN->getGlobal();
unsigned Alignment = GV->getAlignment();
Type *Ty = GV->getValueType();
if (Alignment == 0 && Ty->isSized())
Alignment = DL.getABITypeAlignment(Ty);
if (Alignment >= Size)
return true;
}
}
if (CurDAG->isBaseWithConstantOffset(N)) {
if (ConstantSDNode *RHS = dyn_cast<ConstantSDNode>(N.getOperand(1))) {
int64_t RHSC = (int64_t)RHS->getZExtValue();
unsigned Scale = Log2_32(Size);
if ((RHSC & (Size - 1)) == 0 && RHSC >= 0 && RHSC < (0x1000 << Scale)) {
Base = N.getOperand(0);
if (Base.getOpcode() == ISD::FrameIndex) {
int FI = cast<FrameIndexSDNode>(Base)->getIndex();
Base = CurDAG->getTargetFrameIndex(FI, TLI->getPointerTy(DL));
}
OffImm = CurDAG->getTargetConstant(RHSC >> Scale, dl, MVT::i64);
return true;
}
}
}
// Before falling back to our general case, check if the unscaled
// instructions can handle this. If so, that's preferable.
if (SelectAddrModeUnscaled(N, Size, Base, OffImm))
return false;
// Base only. The address will be materialized into a register before
// the memory is accessed.
// add x0, Xbase, #offset
// ldr x0, [x0]
Base = N;
OffImm = CurDAG->getTargetConstant(0, dl, MVT::i64);
return true;
}
/// SelectAddrModeUnscaled - Select a "register plus unscaled signed 9-bit
/// immediate" address. This should only match when there is an offset that
/// is not valid for a scaled immediate addressing mode. The "Size" argument
/// is the size in bytes of the memory reference, which is needed here to know
/// what is valid for a scaled immediate.
bool AArch64DAGToDAGISel::SelectAddrModeUnscaled(SDValue N, unsigned Size,
SDValue &Base,
SDValue &OffImm) {
if (!CurDAG->isBaseWithConstantOffset(N))
return false;
if (ConstantSDNode *RHS = dyn_cast<ConstantSDNode>(N.getOperand(1))) {
int64_t RHSC = RHS->getSExtValue();
// If the offset is valid as a scaled immediate, don't match here.
if ((RHSC & (Size - 1)) == 0 && RHSC >= 0 &&
RHSC < (0x1000 << Log2_32(Size)))
return false;
if (RHSC >= -256 && RHSC < 256) {
Base = N.getOperand(0);
if (Base.getOpcode() == ISD::FrameIndex) {
int FI = cast<FrameIndexSDNode>(Base)->getIndex();
const TargetLowering *TLI = getTargetLowering();
Base = CurDAG->getTargetFrameIndex(
FI, TLI->getPointerTy(CurDAG->getDataLayout()));
}
OffImm = CurDAG->getTargetConstant(RHSC, SDLoc(N), MVT::i64);
return true;
}
}
return false;
}
static SDValue Widen(SelectionDAG *CurDAG, SDValue N) {
SDLoc dl(N);
SDValue SubReg = CurDAG->getTargetConstant(AArch64::sub_32, dl, MVT::i32);
SDValue ImpDef = SDValue(
CurDAG->getMachineNode(TargetOpcode::IMPLICIT_DEF, dl, MVT::i64), 0);
MachineSDNode *Node = CurDAG->getMachineNode(
TargetOpcode::INSERT_SUBREG, dl, MVT::i64, ImpDef, N, SubReg);
return SDValue(Node, 0);
}
/// Check if the given SHL node (\p N), can be used to form an
/// extended register for an addressing mode.
bool AArch64DAGToDAGISel::SelectExtendedSHL(SDValue N, unsigned Size,
bool WantExtend, SDValue &Offset,
SDValue &SignExtend) {
assert(N.getOpcode() == ISD::SHL && "Invalid opcode.");
ConstantSDNode *CSD = dyn_cast<ConstantSDNode>(N.getOperand(1));
if (!CSD || (CSD->getZExtValue() & 0x7) != CSD->getZExtValue())
return false;
SDLoc dl(N);
if (WantExtend) {
AArch64_AM::ShiftExtendType Ext =
getExtendTypeForNode(N.getOperand(0), true);
if (Ext == AArch64_AM::InvalidShiftExtend)
return false;
Offset = narrowIfNeeded(CurDAG, N.getOperand(0).getOperand(0));
SignExtend = CurDAG->getTargetConstant(Ext == AArch64_AM::SXTW, dl,
MVT::i32);
} else {
Offset = N.getOperand(0);
SignExtend = CurDAG->getTargetConstant(0, dl, MVT::i32);
}
unsigned LegalShiftVal = Log2_32(Size);
unsigned ShiftVal = CSD->getZExtValue();
if (ShiftVal != 0 && ShiftVal != LegalShiftVal)
return false;
return isWorthFolding(N);
}
bool AArch64DAGToDAGISel::SelectAddrModeWRO(SDValue N, unsigned Size,
SDValue &Base, SDValue &Offset,
SDValue &SignExtend,
SDValue &DoShift) {
if (N.getOpcode() != ISD::ADD)
return false;
SDValue LHS = N.getOperand(0);
SDValue RHS = N.getOperand(1);
SDLoc dl(N);
// We don't want to match immediate adds here, because they are better lowered
// to the register-immediate addressing modes.
if (isa<ConstantSDNode>(LHS) || isa<ConstantSDNode>(RHS))
return false;
// Check if this particular node is reused in any non-memory related
// operation. If yes, do not try to fold this node into the address
// computation, since the computation will be kept.
const SDNode *Node = N.getNode();
for (SDNode *UI : Node->uses()) {
if (!isa<MemSDNode>(*UI))
return false;
}
// Remember if it is worth folding N when it produces extended register.
bool IsExtendedRegisterWorthFolding = isWorthFolding(N);
// Try to match a shifted extend on the RHS.
if (IsExtendedRegisterWorthFolding && RHS.getOpcode() == ISD::SHL &&
SelectExtendedSHL(RHS, Size, true, Offset, SignExtend)) {
Base = LHS;
DoShift = CurDAG->getTargetConstant(true, dl, MVT::i32);
return true;
}
// Try to match a shifted extend on the LHS.
if (IsExtendedRegisterWorthFolding && LHS.getOpcode() == ISD::SHL &&
SelectExtendedSHL(LHS, Size, true, Offset, SignExtend)) {
Base = RHS;
DoShift = CurDAG->getTargetConstant(true, dl, MVT::i32);
return true;
}
// There was no shift, whatever else we find.
DoShift = CurDAG->getTargetConstant(false, dl, MVT::i32);
AArch64_AM::ShiftExtendType Ext = AArch64_AM::InvalidShiftExtend;
// Try to match an unshifted extend on the LHS.
if (IsExtendedRegisterWorthFolding &&
(Ext = getExtendTypeForNode(LHS, true)) !=
AArch64_AM::InvalidShiftExtend) {
Base = RHS;
Offset = narrowIfNeeded(CurDAG, LHS.getOperand(0));
SignExtend = CurDAG->getTargetConstant(Ext == AArch64_AM::SXTW, dl,
MVT::i32);
if (isWorthFolding(LHS))
return true;
}
// Try to match an unshifted extend on the RHS.
if (IsExtendedRegisterWorthFolding &&
(Ext = getExtendTypeForNode(RHS, true)) !=
AArch64_AM::InvalidShiftExtend) {
Base = LHS;
Offset = narrowIfNeeded(CurDAG, RHS.getOperand(0));
SignExtend = CurDAG->getTargetConstant(Ext == AArch64_AM::SXTW, dl,
MVT::i32);
if (isWorthFolding(RHS))
return true;
}
return false;
}
// Check if the given immediate is preferred by ADD. If an immediate can be
// encoded in an ADD, or it can be encoded in an "ADD LSL #12" and can not be
// encoded by one MOVZ, return true.
static bool isPreferredADD(int64_t ImmOff) {
// Constant in [0x0, 0xfff] can be encoded in ADD.
if ((ImmOff & 0xfffffffffffff000LL) == 0x0LL)
return true;
// Check if it can be encoded in an "ADD LSL #12".
if ((ImmOff & 0xffffffffff000fffLL) == 0x0LL)
// As a single MOVZ is faster than a "ADD of LSL #12", ignore such constant.
return (ImmOff & 0xffffffffff00ffffLL) != 0x0LL &&
(ImmOff & 0xffffffffffff0fffLL) != 0x0LL;
return false;
}
bool AArch64DAGToDAGISel::SelectAddrModeXRO(SDValue N, unsigned Size,
SDValue &Base, SDValue &Offset,
SDValue &SignExtend,
SDValue &DoShift) {
if (N.getOpcode() != ISD::ADD)
return false;
SDValue LHS = N.getOperand(0);
SDValue RHS = N.getOperand(1);
SDLoc DL(N);
// Check if this particular node is reused in any non-memory related
// operation. If yes, do not try to fold this node into the address
// computation, since the computation will be kept.
const SDNode *Node = N.getNode();
for (SDNode *UI : Node->uses()) {
if (!isa<MemSDNode>(*UI))
return false;
}
// Watch out if RHS is a wide immediate, it can not be selected into
// [BaseReg+Imm] addressing mode. Also it may not be able to be encoded into
// ADD/SUB. Instead it will use [BaseReg + 0] address mode and generate
// instructions like:
// MOV X0, WideImmediate
// ADD X1, BaseReg, X0
// LDR X2, [X1, 0]
// For such situation, using [BaseReg, XReg] addressing mode can save one
// ADD/SUB:
// MOV X0, WideImmediate
// LDR X2, [BaseReg, X0]
if (isa<ConstantSDNode>(RHS)) {
int64_t ImmOff = (int64_t)cast<ConstantSDNode>(RHS)->getZExtValue();
unsigned Scale = Log2_32(Size);
// Skip the immediate can be selected by load/store addressing mode.
// Also skip the immediate can be encoded by a single ADD (SUB is also
// checked by using -ImmOff).
if ((ImmOff % Size == 0 && ImmOff >= 0 && ImmOff < (0x1000 << Scale)) ||
isPreferredADD(ImmOff) || isPreferredADD(-ImmOff))
return false;
SDValue Ops[] = { RHS };
SDNode *MOVI =
CurDAG->getMachineNode(AArch64::MOVi64imm, DL, MVT::i64, Ops);
SDValue MOVIV = SDValue(MOVI, 0);
// This ADD of two X register will be selected into [Reg+Reg] mode.
N = CurDAG->getNode(ISD::ADD, DL, MVT::i64, LHS, MOVIV);
}
// Remember if it is worth folding N when it produces extended register.
bool IsExtendedRegisterWorthFolding = isWorthFolding(N);
// Try to match a shifted extend on the RHS.
if (IsExtendedRegisterWorthFolding && RHS.getOpcode() == ISD::SHL &&
SelectExtendedSHL(RHS, Size, false, Offset, SignExtend)) {
Base = LHS;
DoShift = CurDAG->getTargetConstant(true, DL, MVT::i32);
return true;
}
// Try to match a shifted extend on the LHS.
if (IsExtendedRegisterWorthFolding && LHS.getOpcode() == ISD::SHL &&
SelectExtendedSHL(LHS, Size, false, Offset, SignExtend)) {
Base = RHS;
DoShift = CurDAG->getTargetConstant(true, DL, MVT::i32);
return true;
}
// Match any non-shifted, non-extend, non-immediate add expression.
Base = LHS;
Offset = RHS;
SignExtend = CurDAG->getTargetConstant(false, DL, MVT::i32);
DoShift = CurDAG->getTargetConstant(false, DL, MVT::i32);
// Reg1 + Reg2 is free: no check needed.
return true;
}
SDValue AArch64DAGToDAGISel::createDTuple(ArrayRef<SDValue> Regs) {
static const unsigned RegClassIDs[] = {
AArch64::DDRegClassID, AArch64::DDDRegClassID, AArch64::DDDDRegClassID};
static const unsigned SubRegs[] = {AArch64::dsub0, AArch64::dsub1,
AArch64::dsub2, AArch64::dsub3};
return createTuple(Regs, RegClassIDs, SubRegs);
}
SDValue AArch64DAGToDAGISel::createQTuple(ArrayRef<SDValue> Regs) {
static const unsigned RegClassIDs[] = {
AArch64::QQRegClassID, AArch64::QQQRegClassID, AArch64::QQQQRegClassID};
static const unsigned SubRegs[] = {AArch64::qsub0, AArch64::qsub1,
AArch64::qsub2, AArch64::qsub3};
return createTuple(Regs, RegClassIDs, SubRegs);
}
SDValue AArch64DAGToDAGISel::createTuple(ArrayRef<SDValue> Regs,
const unsigned RegClassIDs[],
const unsigned SubRegs[]) {
// There's no special register-class for a vector-list of 1 element: it's just
// a vector.
if (Regs.size() == 1)
return Regs[0];
assert(Regs.size() >= 2 && Regs.size() <= 4);
SDLoc DL(Regs[0]);
SmallVector<SDValue, 4> Ops;
// First operand of REG_SEQUENCE is the desired RegClass.
Ops.push_back(
CurDAG->getTargetConstant(RegClassIDs[Regs.size() - 2], DL, MVT::i32));
// Then we get pairs of source & subregister-position for the components.
for (unsigned i = 0; i < Regs.size(); ++i) {
Ops.push_back(Regs[i]);
Ops.push_back(CurDAG->getTargetConstant(SubRegs[i], DL, MVT::i32));
}
SDNode *N =
CurDAG->getMachineNode(TargetOpcode::REG_SEQUENCE, DL, MVT::Untyped, Ops);
return SDValue(N, 0);
}
void AArch64DAGToDAGISel::SelectTable(SDNode *N, unsigned NumVecs, unsigned Opc,
bool isExt) {
SDLoc dl(N);
EVT VT = N->getValueType(0);
unsigned ExtOff = isExt;
// Form a REG_SEQUENCE to force register allocation.
unsigned Vec0Off = ExtOff + 1;
SmallVector<SDValue, 4> Regs(N->op_begin() + Vec0Off,
N->op_begin() + Vec0Off + NumVecs);
SDValue RegSeq = createQTuple(Regs);
SmallVector<SDValue, 6> Ops;
if (isExt)
Ops.push_back(N->getOperand(1));
Ops.push_back(RegSeq);
Ops.push_back(N->getOperand(NumVecs + ExtOff + 1));
ReplaceNode(N, CurDAG->getMachineNode(Opc, dl, VT, Ops));
}
bool AArch64DAGToDAGISel::tryIndexedLoad(SDNode *N) {
LoadSDNode *LD = cast<LoadSDNode>(N);
if (LD->isUnindexed())
return false;
EVT VT = LD->getMemoryVT();
EVT DstVT = N->getValueType(0);
ISD::MemIndexedMode AM = LD->getAddressingMode();
bool IsPre = AM == ISD::PRE_INC || AM == ISD::PRE_DEC;
// We're not doing validity checking here. That was done when checking
// if we should mark the load as indexed or not. We're just selecting
// the right instruction.
unsigned Opcode = 0;
ISD::LoadExtType ExtType = LD->getExtensionType();
bool InsertTo64 = false;
if (VT == MVT::i64)
Opcode = IsPre ? AArch64::LDRXpre : AArch64::LDRXpost;
else if (VT == MVT::i32) {
if (ExtType == ISD::NON_EXTLOAD)
Opcode = IsPre ? AArch64::LDRWpre : AArch64::LDRWpost;
else if (ExtType == ISD::SEXTLOAD)
Opcode = IsPre ? AArch64::LDRSWpre : AArch64::LDRSWpost;
else {
Opcode = IsPre ? AArch64::LDRWpre : AArch64::LDRWpost;
InsertTo64 = true;
// The result of the load is only i32. It's the subreg_to_reg that makes
// it into an i64.
DstVT = MVT::i32;
}
} else if (VT == MVT::i16) {
if (ExtType == ISD::SEXTLOAD) {
if (DstVT == MVT::i64)
Opcode = IsPre ? AArch64::LDRSHXpre : AArch64::LDRSHXpost;
else
Opcode = IsPre ? AArch64::LDRSHWpre : AArch64::LDRSHWpost;
} else {
Opcode = IsPre ? AArch64::LDRHHpre : AArch64::LDRHHpost;
InsertTo64 = DstVT == MVT::i64;
// The result of the load is only i32. It's the subreg_to_reg that makes
// it into an i64.
DstVT = MVT::i32;
}
} else if (VT == MVT::i8) {
if (ExtType == ISD::SEXTLOAD) {
if (DstVT == MVT::i64)
Opcode = IsPre ? AArch64::LDRSBXpre : AArch64::LDRSBXpost;
else
Opcode = IsPre ? AArch64::LDRSBWpre : AArch64::LDRSBWpost;
} else {
Opcode = IsPre ? AArch64::LDRBBpre : AArch64::LDRBBpost;
InsertTo64 = DstVT == MVT::i64;
// The result of the load is only i32. It's the subreg_to_reg that makes
// it into an i64.
DstVT = MVT::i32;
}
} else if (VT == MVT::f16) {
Opcode = IsPre ? AArch64::LDRHpre : AArch64::LDRHpost;
} else if (VT == MVT::f32) {
Opcode = IsPre ? AArch64::LDRSpre : AArch64::LDRSpost;
} else if (VT == MVT::f64 || VT.is64BitVector()) {
Opcode = IsPre ? AArch64::LDRDpre : AArch64::LDRDpost;
} else if (VT.is128BitVector()) {
Opcode = IsPre ? AArch64::LDRQpre : AArch64::LDRQpost;
} else
return false;
SDValue Chain = LD->getChain();
SDValue Base = LD->getBasePtr();
ConstantSDNode *OffsetOp = cast<ConstantSDNode>(LD->getOffset());
int OffsetVal = (int)OffsetOp->getZExtValue();
SDLoc dl(N);
SDValue Offset = CurDAG->getTargetConstant(OffsetVal, dl, MVT::i64);
SDValue Ops[] = { Base, Offset, Chain };
SDNode *Res = CurDAG->getMachineNode(Opcode, dl, MVT::i64, DstVT,
MVT::Other, Ops);
// Either way, we're replacing the node, so tell the caller that.
SDValue LoadedVal = SDValue(Res, 1);
if (InsertTo64) {
SDValue SubReg = CurDAG->getTargetConstant(AArch64::sub_32, dl, MVT::i32);
LoadedVal =
SDValue(CurDAG->getMachineNode(
AArch64::SUBREG_TO_REG, dl, MVT::i64,
CurDAG->getTargetConstant(0, dl, MVT::i64), LoadedVal,
SubReg),
0);
}
ReplaceUses(SDValue(N, 0), LoadedVal);
ReplaceUses(SDValue(N, 1), SDValue(Res, 0));
ReplaceUses(SDValue(N, 2), SDValue(Res, 2));
CurDAG->RemoveDeadNode(N);
return true;
}
void AArch64DAGToDAGISel::SelectLoad(SDNode *N, unsigned NumVecs, unsigned Opc,
unsigned SubRegIdx) {
SDLoc dl(N);
EVT VT = N->getValueType(0);
SDValue Chain = N->getOperand(0);
SDValue Ops[] = {N->getOperand(2), // Mem operand;
Chain};
const EVT ResTys[] = {MVT::Untyped, MVT::Other};
SDNode *Ld = CurDAG->getMachineNode(Opc, dl, ResTys, Ops);
SDValue SuperReg = SDValue(Ld, 0);
for (unsigned i = 0; i < NumVecs; ++i)
ReplaceUses(SDValue(N, i),
CurDAG->getTargetExtractSubreg(SubRegIdx + i, dl, VT, SuperReg));
ReplaceUses(SDValue(N, NumVecs), SDValue(Ld, 1));
// Transfer memoperands.
MachineMemOperand *MemOp = cast<MemIntrinsicSDNode>(N)->getMemOperand();
CurDAG->setNodeMemRefs(cast<MachineSDNode>(Ld), {MemOp});
CurDAG->RemoveDeadNode(N);
}
void AArch64DAGToDAGISel::SelectPostLoad(SDNode *N, unsigned NumVecs,
unsigned Opc, unsigned SubRegIdx) {
SDLoc dl(N);
EVT VT = N->getValueType(0);
SDValue Chain = N->getOperand(0);
SDValue Ops[] = {N->getOperand(1), // Mem operand
N->getOperand(2), // Incremental
Chain};
const EVT ResTys[] = {MVT::i64, // Type of the write back register
MVT::Untyped, MVT::Other};
SDNode *Ld = CurDAG->getMachineNode(Opc, dl, ResTys, Ops);
// Update uses of write back register
ReplaceUses(SDValue(N, NumVecs), SDValue(Ld, 0));
// Update uses of vector list
SDValue SuperReg = SDValue(Ld, 1);
if (NumVecs == 1)
ReplaceUses(SDValue(N, 0), SuperReg);
else
for (unsigned i = 0; i < NumVecs; ++i)
ReplaceUses(SDValue(N, i),
CurDAG->getTargetExtractSubreg(SubRegIdx + i, dl, VT, SuperReg));
// Update the chain
ReplaceUses(SDValue(N, NumVecs + 1), SDValue(Ld, 2));
CurDAG->RemoveDeadNode(N);
}
void AArch64DAGToDAGISel::SelectStore(SDNode *N, unsigned NumVecs,
unsigned Opc) {
SDLoc dl(N);
EVT VT = N->getOperand(2)->getValueType(0);
// Form a REG_SEQUENCE to force register allocation.
bool Is128Bit = VT.getSizeInBits() == 128;
SmallVector<SDValue, 4> Regs(N->op_begin() + 2, N->op_begin() + 2 + NumVecs);
SDValue RegSeq = Is128Bit ? createQTuple(Regs) : createDTuple(Regs);
SDValue Ops[] = {RegSeq, N->getOperand(NumVecs + 2), N->getOperand(0)};
SDNode *St = CurDAG->getMachineNode(Opc, dl, N->getValueType(0), Ops);
// Transfer memoperands.
MachineMemOperand *MemOp = cast<MemIntrinsicSDNode>(N)->getMemOperand();
CurDAG->setNodeMemRefs(cast<MachineSDNode>(St), {MemOp});
ReplaceNode(N, St);
}
void AArch64DAGToDAGISel::SelectPostStore(SDNode *N, unsigned NumVecs,
unsigned Opc) {
SDLoc dl(N);
EVT VT = N->getOperand(2)->getValueType(0);
const EVT ResTys[] = {MVT::i64, // Type of the write back register
MVT::Other}; // Type for the Chain
// Form a REG_SEQUENCE to force register allocation.
bool Is128Bit = VT.getSizeInBits() == 128;
SmallVector<SDValue, 4> Regs(N->op_begin() + 1, N->op_begin() + 1 + NumVecs);
SDValue RegSeq = Is128Bit ? createQTuple(Regs) : createDTuple(Regs);
SDValue Ops[] = {RegSeq,
N->getOperand(NumVecs + 1), // base register
N->getOperand(NumVecs + 2), // Incremental
N->getOperand(0)}; // Chain
SDNode *St = CurDAG->getMachineNode(Opc, dl, ResTys, Ops);
ReplaceNode(N, St);
}
namespace {
/// WidenVector - Given a value in the V64 register class, produce the
/// equivalent value in the V128 register class.
class WidenVector {
SelectionDAG &DAG;
public:
WidenVector(SelectionDAG &DAG) : DAG(DAG) {}
SDValue operator()(SDValue V64Reg) {
EVT VT = V64Reg.getValueType();
unsigned NarrowSize = VT.getVectorNumElements();
MVT EltTy = VT.getVectorElementType().getSimpleVT();
MVT WideTy = MVT::getVectorVT(EltTy, 2 * NarrowSize);
SDLoc DL(V64Reg);
SDValue Undef =
SDValue(DAG.getMachineNode(TargetOpcode::IMPLICIT_DEF, DL, WideTy), 0);
return DAG.getTargetInsertSubreg(AArch64::dsub, DL, WideTy, Undef, V64Reg);
}
};
} // namespace
/// NarrowVector - Given a value in the V128 register class, produce the
/// equivalent value in the V64 register class.
static SDValue NarrowVector(SDValue V128Reg, SelectionDAG &DAG) {
EVT VT = V128Reg.getValueType();
unsigned WideSize = VT.getVectorNumElements();
MVT EltTy = VT.getVectorElementType().getSimpleVT();
MVT NarrowTy = MVT::getVectorVT(EltTy, WideSize / 2);
return DAG.getTargetExtractSubreg(AArch64::dsub, SDLoc(V128Reg), NarrowTy,
V128Reg);
}
void AArch64DAGToDAGISel::SelectLoadLane(SDNode *N, unsigned NumVecs,
unsigned Opc) {
SDLoc dl(N);
EVT VT = N->getValueType(0);
bool Narrow = VT.getSizeInBits() == 64;
// Form a REG_SEQUENCE to force register allocation.
SmallVector<SDValue, 4> Regs(N->op_begin() + 2, N->op_begin() + 2 + NumVecs);
if (Narrow)
transform(Regs, Regs.begin(),
WidenVector(*CurDAG));
SDValue RegSeq = createQTuple(Regs);
const EVT ResTys[] = {MVT::Untyped, MVT::Other};
unsigned LaneNo =
cast<ConstantSDNode>(N->getOperand(NumVecs + 2))->getZExtValue();
SDValue Ops[] = {RegSeq, CurDAG->getTargetConstant(LaneNo, dl, MVT::i64),
N->getOperand(NumVecs + 3), N->getOperand(0)};
SDNode *Ld = CurDAG->getMachineNode(Opc, dl, ResTys, Ops);
SDValue SuperReg = SDValue(Ld, 0);
EVT WideVT = RegSeq.getOperand(1)->getValueType(0);
static const unsigned QSubs[] = { AArch64::qsub0, AArch64::qsub1,
AArch64::qsub2, AArch64::qsub3 };
for (unsigned i = 0; i < NumVecs; ++i) {
SDValue NV = CurDAG->getTargetExtractSubreg(QSubs[i], dl, WideVT, SuperReg);
if (Narrow)
NV = NarrowVector(NV, *CurDAG);
ReplaceUses(SDValue(N, i), NV);
}
ReplaceUses(SDValue(N, NumVecs), SDValue(Ld, 1));
CurDAG->RemoveDeadNode(N);
}
void AArch64DAGToDAGISel::SelectPostLoadLane(SDNode *N, unsigned NumVecs,
unsigned Opc) {
SDLoc dl(N);
EVT VT = N->getValueType(0);
bool Narrow = VT.getSizeInBits() == 64;
// Form a REG_SEQUENCE to force register allocation.
SmallVector<SDValue, 4> Regs(N->op_begin() + 1, N->op_begin() + 1 + NumVecs);
if (Narrow)
transform(Regs, Regs.begin(),
WidenVector(*CurDAG));
SDValue RegSeq = createQTuple(Regs);
const EVT ResTys[] = {MVT::i64, // Type of the write back register
RegSeq->getValueType(0), MVT::Other};
unsigned LaneNo =
cast<ConstantSDNode>(N->getOperand(NumVecs + 1))->getZExtValue();
SDValue Ops[] = {RegSeq,
CurDAG->getTargetConstant(LaneNo, dl,
MVT::i64), // Lane Number
N->getOperand(NumVecs + 2), // Base register
N->getOperand(NumVecs + 3), // Incremental
N->getOperand(0)};
SDNode *Ld = CurDAG->getMachineNode(Opc, dl, ResTys, Ops);
// Update uses of the write back register
ReplaceUses(SDValue(N, NumVecs), SDValue(Ld, 0));
// Update uses of the vector list
SDValue SuperReg = SDValue(Ld, 1);
if (NumVecs == 1) {
ReplaceUses(SDValue(N, 0),
Narrow ? NarrowVector(SuperReg, *CurDAG) : SuperReg);
} else {
EVT WideVT = RegSeq.getOperand(1)->getValueType(0);
static const unsigned QSubs[] = { AArch64::qsub0, AArch64::qsub1,
AArch64::qsub2, AArch64::qsub3 };
for (unsigned i = 0; i < NumVecs; ++i) {
SDValue NV = CurDAG->getTargetExtractSubreg(QSubs[i], dl, WideVT,
SuperReg);
if (Narrow)
NV = NarrowVector(NV, *CurDAG);
ReplaceUses(SDValue(N, i), NV);
}
}
// Update the Chain
ReplaceUses(SDValue(N, NumVecs + 1), SDValue(Ld, 2));
CurDAG->RemoveDeadNode(N);
}
void AArch64DAGToDAGISel::SelectStoreLane(SDNode *N, unsigned NumVecs,
unsigned Opc) {
SDLoc dl(N);
EVT VT = N->getOperand(2)->getValueType(0);
bool Narrow = VT.getSizeInBits() == 64;
// Form a REG_SEQUENCE to force register allocation.
SmallVector<SDValue, 4> Regs(N->op_begin() + 2, N->op_begin() + 2 + NumVecs);
if (Narrow)
transform(Regs, Regs.begin(),
WidenVector(*CurDAG));
SDValue RegSeq = createQTuple(Regs);
unsigned LaneNo =
cast<ConstantSDNode>(N->getOperand(NumVecs + 2))->getZExtValue();
SDValue Ops[] = {RegSeq, CurDAG->getTargetConstant(LaneNo, dl, MVT::i64),
N->getOperand(NumVecs + 3), N->getOperand(0)};
SDNode *St = CurDAG->getMachineNode(Opc, dl, MVT::Other, Ops);
// Transfer memoperands.
MachineMemOperand *MemOp = cast<MemIntrinsicSDNode>(N)->getMemOperand();
CurDAG->setNodeMemRefs(cast<MachineSDNode>(St), {MemOp});
ReplaceNode(N, St);
}
void AArch64DAGToDAGISel::SelectPostStoreLane(SDNode *N, unsigned NumVecs,
unsigned Opc) {
SDLoc dl(N);
EVT VT = N->getOperand(2)->getValueType(0);
bool Narrow = VT.getSizeInBits() == 64;
// Form a REG_SEQUENCE to force register allocation.
SmallVector<SDValue, 4> Regs(N->op_begin() + 1, N->op_begin() + 1 + NumVecs);
if (Narrow)
transform(Regs, Regs.begin(),
WidenVector(*CurDAG));
SDValue RegSeq = createQTuple(Regs);
const EVT ResTys[] = {MVT::i64, // Type of the write back register
MVT::Other};
unsigned LaneNo =
cast<ConstantSDNode>(N->getOperand(NumVecs + 1))->getZExtValue();
SDValue Ops[] = {RegSeq, CurDAG->getTargetConstant(LaneNo, dl, MVT::i64),
N->getOperand(NumVecs + 2), // Base Register
N->getOperand(NumVecs + 3), // Incremental
N->getOperand(0)};
SDNode *St = CurDAG->getMachineNode(Opc, dl, ResTys, Ops);
// Transfer memoperands.
MachineMemOperand *MemOp = cast<MemIntrinsicSDNode>(N)->getMemOperand();
CurDAG->setNodeMemRefs(cast<MachineSDNode>(St), {MemOp});
ReplaceNode(N, St);
}
static bool isBitfieldExtractOpFromAnd(SelectionDAG *CurDAG, SDNode *N,
unsigned &Opc, SDValue &Opd0,
unsigned &LSB, unsigned &MSB,
unsigned NumberOfIgnoredLowBits,
bool BiggerPattern) {
assert(N->getOpcode() == ISD::AND &&
"N must be a AND operation to call this function");
EVT VT = N->getValueType(0);
// Here we can test the type of VT and return false when the type does not
// match, but since it is done prior to that call in the current context
// we turned that into an assert to avoid redundant code.
assert((VT == MVT::i32 || VT == MVT::i64) &&
"Type checking must have been done before calling this function");
// FIXME: simplify-demanded-bits in DAGCombine will probably have
// changed the AND node to a 32-bit mask operation. We'll have to
// undo that as part of the transform here if we want to catch all
// the opportunities.
// Currently the NumberOfIgnoredLowBits argument helps to recover
// form these situations when matching bigger pattern (bitfield insert).
// For unsigned extracts, check for a shift right and mask
uint64_t AndImm = 0;
if (!isOpcWithIntImmediate(N, ISD::AND, AndImm))
return false;
const SDNode *Op0 = N->getOperand(0).getNode();
// Because of simplify-demanded-bits in DAGCombine, the mask may have been
// simplified. Try to undo that
AndImm |= maskTrailingOnes<uint64_t>(NumberOfIgnoredLowBits);
// The immediate is a mask of the low bits iff imm & (imm+1) == 0
if (AndImm & (AndImm + 1))
return false;
bool ClampMSB = false;
uint64_t SrlImm = 0;
// Handle the SRL + ANY_EXTEND case.
if (VT == MVT::i64 && Op0->getOpcode() == ISD::ANY_EXTEND &&
isOpcWithIntImmediate(Op0->getOperand(0).getNode(), ISD::SRL, SrlImm)) {
// Extend the incoming operand of the SRL to 64-bit.
Opd0 = Widen(CurDAG, Op0->getOperand(0).getOperand(0));
// Make sure to clamp the MSB so that we preserve the semantics of the
// original operations.
ClampMSB = true;
} else if (VT == MVT::i32 && Op0->getOpcode() == ISD::TRUNCATE &&
isOpcWithIntImmediate(Op0->getOperand(0).getNode(), ISD::SRL,
SrlImm)) {
// If the shift result was truncated, we can still combine them.
Opd0 = Op0->getOperand(0).getOperand(0);
// Use the type of SRL node.
VT = Opd0->getValueType(0);
} else if (isOpcWithIntImmediate(Op0, ISD::SRL, SrlImm)) {
Opd0 = Op0->getOperand(0);
} else if (BiggerPattern) {
// Let's pretend a 0 shift right has been performed.
// The resulting code will be at least as good as the original one
// plus it may expose more opportunities for bitfield insert pattern.
// FIXME: Currently we limit this to the bigger pattern, because
// some optimizations expect AND and not UBFM.
Opd0 = N->getOperand(0);
} else
return false;
// Bail out on large immediates. This happens when no proper
// combining/constant folding was performed.
if (!BiggerPattern && (SrlImm <= 0 || SrlImm >= VT.getSizeInBits())) {
LLVM_DEBUG(
(dbgs() << N
<< ": Found large shift immediate, this should not happen\n"));
return false;
}
LSB = SrlImm;
MSB = SrlImm + (VT == MVT::i32 ? countTrailingOnes<uint32_t>(AndImm)
: countTrailingOnes<uint64_t>(AndImm)) -
1;
if (ClampMSB)
// Since we're moving the extend before the right shift operation, we need
// to clamp the MSB to make sure we don't shift in undefined bits instead of
// the zeros which would get shifted in with the original right shift
// operation.
MSB = MSB > 31 ? 31 : MSB;
Opc = VT == MVT::i32 ? AArch64::UBFMWri : AArch64::UBFMXri;
return true;
}
static bool isBitfieldExtractOpFromSExtInReg(SDNode *N, unsigned &Opc,
SDValue &Opd0, unsigned &Immr,
unsigned &Imms) {
assert(N->getOpcode() == ISD::SIGN_EXTEND_INREG);
EVT VT = N->getValueType(0);
unsigned BitWidth = VT.getSizeInBits();
assert((VT == MVT::i32 || VT == MVT::i64) &&
"Type checking must have been done before calling this function");
SDValue Op = N->getOperand(0);
if (Op->getOpcode() == ISD::TRUNCATE) {
Op = Op->getOperand(0);
VT = Op->getValueType(0);
BitWidth = VT.getSizeInBits();
}
uint64_t ShiftImm;
if (!isOpcWithIntImmediate(Op.getNode(), ISD::SRL, ShiftImm) &&
!isOpcWithIntImmediate(Op.getNode(), ISD::SRA, ShiftImm))
return false;
unsigned Width = cast<VTSDNode>(N->getOperand(1))->getVT().getSizeInBits();
if (ShiftImm + Width > BitWidth)
return false;
Opc = (VT == MVT::i32) ? AArch64::SBFMWri : AArch64::SBFMXri;
Opd0 = Op.getOperand(0);
Immr = ShiftImm;
Imms = ShiftImm + Width - 1;
return true;
}
static bool isSeveralBitsExtractOpFromShr(SDNode *N, unsigned &Opc,
SDValue &Opd0, unsigned &LSB,
unsigned &MSB) {
// We are looking for the following pattern which basically extracts several
// continuous bits from the source value and places it from the LSB of the
// destination value, all other bits of the destination value or set to zero:
//
// Value2 = AND Value, MaskImm
// SRL Value2, ShiftImm
//
// with MaskImm >> ShiftImm to search for the bit width.
//
// This gets selected into a single UBFM:
//
// UBFM Value, ShiftImm, BitWide + SrlImm -1
//
if (N->getOpcode() != ISD::SRL)
return false;
uint64_t AndMask = 0;
if (!isOpcWithIntImmediate(N->getOperand(0).getNode(), ISD::AND, AndMask))
return false;
Opd0 = N->getOperand(0).getOperand(0);
uint64_t SrlImm = 0;
if (!isIntImmediate(N->getOperand(1), SrlImm))
return false;
// Check whether we really have several bits extract here.
unsigned BitWide = 64 - countLeadingOnes(~(AndMask >> SrlImm));
if (BitWide && isMask_64(AndMask >> SrlImm)) {
if (N->getValueType(0) == MVT::i32)
Opc = AArch64::UBFMWri;
else
Opc = AArch64::UBFMXri;
LSB = SrlImm;
MSB = BitWide + SrlImm - 1;
return true;
}
return false;
}
static bool isBitfieldExtractOpFromShr(SDNode *N, unsigned &Opc, SDValue &Opd0,
unsigned &Immr, unsigned &Imms,
bool BiggerPattern) {
assert((N->getOpcode() == ISD::SRA || N->getOpcode() == ISD::SRL) &&
"N must be a SHR/SRA operation to call this function");
EVT VT = N->getValueType(0);
// Here we can test the type of VT and return false when the type does not
// match, but since it is done prior to that call in the current context
// we turned that into an assert to avoid redundant code.
assert((VT == MVT::i32 || VT == MVT::i64) &&
"Type checking must have been done before calling this function");
// Check for AND + SRL doing several bits extract.
if (isSeveralBitsExtractOpFromShr(N, Opc, Opd0, Immr, Imms))
return true;
// We're looking for a shift of a shift.
uint64_t ShlImm = 0;
uint64_t TruncBits = 0;
if (isOpcWithIntImmediate(N->getOperand(0).getNode(), ISD::SHL, ShlImm)) {
Opd0 = N->getOperand(0).getOperand(0);
} else if (VT == MVT::i32 && N->getOpcode() == ISD::SRL &&
N->getOperand(0).getNode()->getOpcode() == ISD::TRUNCATE) {
// We are looking for a shift of truncate. Truncate from i64 to i32 could
// be considered as setting high 32 bits as zero. Our strategy here is to
// always generate 64bit UBFM. This consistency will help the CSE pass
// later find more redundancy.
Opd0 = N->getOperand(0).getOperand(0);
TruncBits = Opd0->getValueType(0).getSizeInBits() - VT.getSizeInBits();
VT = Opd0.getValueType();
assert(VT == MVT::i64 && "the promoted type should be i64");
} else if (BiggerPattern) {
// Let's pretend a 0 shift left has been performed.
// FIXME: Currently we limit this to the bigger pattern case,
// because some optimizations expect AND and not UBFM
Opd0 = N->getOperand(0);
} else
return false;
// Missing combines/constant folding may have left us with strange
// constants.
if (ShlImm >= VT.getSizeInBits()) {
LLVM_DEBUG(
(dbgs() << N
<< ": Found large shift immediate, this should not happen\n"));
return false;
}
uint64_t SrlImm = 0;
if (!isIntImmediate(N->getOperand(1), SrlImm))
return false;
assert(SrlImm > 0 && SrlImm < VT.getSizeInBits() &&
"bad amount in shift node!");
int immr = SrlImm - ShlImm;
Immr = immr < 0 ? immr + VT.getSizeInBits() : immr;
Imms = VT.getSizeInBits() - ShlImm - TruncBits - 1;
// SRA requires a signed extraction
if (VT == MVT::i32)
Opc = N->getOpcode() == ISD::SRA ? AArch64::SBFMWri : AArch64::UBFMWri;
else
Opc = N->getOpcode() == ISD::SRA ? AArch64::SBFMXri : AArch64::UBFMXri;
return true;
}
bool AArch64DAGToDAGISel::tryBitfieldExtractOpFromSExt(SDNode *N) {
assert(N->getOpcode() == ISD::SIGN_EXTEND);
EVT VT = N->getValueType(0);
EVT NarrowVT = N->getOperand(0)->getValueType(0);
if (VT != MVT::i64 || NarrowVT != MVT::i32)
return false;
uint64_t ShiftImm;
SDValue Op = N->getOperand(0);
if (!isOpcWithIntImmediate(Op.getNode(), ISD::SRA, ShiftImm))
return false;
SDLoc dl(N);
// Extend the incoming operand of the shift to 64-bits.
SDValue Opd0 = Widen(CurDAG, Op.getOperand(0));
unsigned Immr = ShiftImm;
unsigned Imms = NarrowVT.getSizeInBits() - 1;
SDValue Ops[] = {Opd0, CurDAG->getTargetConstant(Immr, dl, VT),
CurDAG->getTargetConstant(Imms, dl, VT)};
CurDAG->SelectNodeTo(N, AArch64::SBFMXri, VT, Ops);
return true;
}
/// Try to form fcvtl2 instructions from a floating-point extend of a high-half
/// extract of a subvector.
bool AArch64DAGToDAGISel::tryHighFPExt(SDNode *N) {
assert(N->getOpcode() == ISD::FP_EXTEND);
// There are 2 forms of fcvtl2 - extend to double or extend to float.
SDValue Extract = N->getOperand(0);
EVT VT = N->getValueType(0);
EVT NarrowVT = Extract.getValueType();
if ((VT != MVT::v2f64 || NarrowVT != MVT::v2f32) &&
(VT != MVT::v4f32 || NarrowVT != MVT::v4f16))
return false;
// Optionally look past a bitcast.
Extract = peekThroughBitcasts(Extract);
if (Extract.getOpcode() != ISD::EXTRACT_SUBVECTOR)
return false;
// Match extract from start of high half index.
// Example: v8i16 -> v4i16 means the extract must begin at index 4.
unsigned ExtractIndex = Extract.getConstantOperandVal(1);
if (ExtractIndex != Extract.getValueType().getVectorNumElements())
return false;
auto Opcode = VT == MVT::v2f64 ? AArch64::FCVTLv4i32 : AArch64::FCVTLv8i16;
CurDAG->SelectNodeTo(N, Opcode, VT, Extract.getOperand(0));
return true;
}
static bool isBitfieldExtractOp(SelectionDAG *CurDAG, SDNode *N, unsigned &Opc,
SDValue &Opd0, unsigned &Immr, unsigned &Imms,
unsigned NumberOfIgnoredLowBits = 0,
bool BiggerPattern = false) {
if (N->getValueType(0) != MVT::i32 && N->getValueType(0) != MVT::i64)
return false;
switch (N->getOpcode()) {
default:
if (!N->isMachineOpcode())
return false;
break;
case ISD::AND:
return isBitfieldExtractOpFromAnd(CurDAG, N, Opc, Opd0, Immr, Imms,
NumberOfIgnoredLowBits, BiggerPattern);
case ISD::SRL:
case ISD::SRA:
return isBitfieldExtractOpFromShr(N, Opc, Opd0, Immr, Imms, BiggerPattern);
case ISD::SIGN_EXTEND_INREG:
return isBitfieldExtractOpFromSExtInReg(N, Opc, Opd0, Immr, Imms);
}
unsigned NOpc = N->getMachineOpcode();
switch (NOpc) {
default:
return false;
case AArch64::SBFMWri:
case AArch64::UBFMWri:
case AArch64::SBFMXri:
case AArch64::UBFMXri:
Opc = NOpc;
Opd0 = N->getOperand(0);
Immr = cast<ConstantSDNode>(N->getOperand(1).getNode())->getZExtValue();
Imms = cast<ConstantSDNode>(N->getOperand(2).getNode())->getZExtValue();
return true;
}
// Unreachable
return false;
}
bool AArch64DAGToDAGISel::tryBitfieldExtractOp(SDNode *N) {
unsigned Opc, Immr, Imms;
SDValue Opd0;
if (!isBitfieldExtractOp(CurDAG, N, Opc, Opd0, Immr, Imms))
return false;
EVT VT = N->getValueType(0);
SDLoc dl(N);
// If the bit extract operation is 64bit but the original type is 32bit, we
// need to add one EXTRACT_SUBREG.
if ((Opc == AArch64::SBFMXri || Opc == AArch64::UBFMXri) && VT == MVT::i32) {
SDValue Ops64[] = {Opd0, CurDAG->getTargetConstant(Immr, dl, MVT::i64),
CurDAG->getTargetConstant(Imms, dl, MVT::i64)};
SDNode *BFM = CurDAG->getMachineNode(Opc, dl, MVT::i64, Ops64);
SDValue SubReg = CurDAG->getTargetConstant(AArch64::sub_32, dl, MVT::i32);
ReplaceNode(N, CurDAG->getMachineNode(TargetOpcode::EXTRACT_SUBREG, dl,
MVT::i32, SDValue(BFM, 0), SubReg));
return true;
}
SDValue Ops[] = {Opd0, CurDAG->getTargetConstant(Immr, dl, VT),
CurDAG->getTargetConstant(Imms, dl, VT)};
CurDAG->SelectNodeTo(N, Opc, VT, Ops);
return true;
}
/// Does DstMask form a complementary pair with the mask provided by
/// BitsToBeInserted, suitable for use in a BFI instruction. Roughly speaking,
/// this asks whether DstMask zeroes precisely those bits that will be set by
/// the other half.
static bool isBitfieldDstMask(uint64_t DstMask, const APInt &BitsToBeInserted,
unsigned NumberOfIgnoredHighBits, EVT VT) {
assert((VT == MVT::i32 || VT == MVT::i64) &&
"i32 or i64 mask type expected!");
unsigned BitWidth = VT.getSizeInBits() - NumberOfIgnoredHighBits;
APInt SignificantDstMask = APInt(BitWidth, DstMask);
APInt SignificantBitsToBeInserted = BitsToBeInserted.zextOrTrunc(BitWidth);
return (SignificantDstMask & SignificantBitsToBeInserted) == 0 &&
(SignificantDstMask | SignificantBitsToBeInserted).isAllOnesValue();
}
// Look for bits that will be useful for later uses.
// A bit is consider useless as soon as it is dropped and never used
// before it as been dropped.
// E.g., looking for useful bit of x
// 1. y = x & 0x7
// 2. z = y >> 2
// After #1, x useful bits are 0x7, then the useful bits of x, live through
// y.
// After #2, the useful bits of x are 0x4.
// However, if x is used on an unpredicatable instruction, then all its bits
// are useful.
// E.g.
// 1. y = x & 0x7
// 2. z = y >> 2
// 3. str x, [@x]
static void getUsefulBits(SDValue Op, APInt &UsefulBits, unsigned Depth = 0);
static void getUsefulBitsFromAndWithImmediate(SDValue Op, APInt &UsefulBits,
unsigned Depth) {
uint64_t Imm =
cast<const ConstantSDNode>(Op.getOperand(1).getNode())->getZExtValue();
Imm = AArch64_AM::decodeLogicalImmediate(Imm, UsefulBits.getBitWidth());
UsefulBits &= APInt(UsefulBits.getBitWidth(), Imm);
getUsefulBits(Op, UsefulBits, Depth + 1);
}
static void getUsefulBitsFromBitfieldMoveOpd(SDValue Op, APInt &UsefulBits,
uint64_t Imm, uint64_t MSB,
unsigned Depth) {
// inherit the bitwidth value
APInt OpUsefulBits(UsefulBits);
OpUsefulBits = 1;
if (MSB >= Imm) {
OpUsefulBits <<= MSB - Imm + 1;
--OpUsefulBits;
// The interesting part will be in the lower part of the result
getUsefulBits(Op, OpUsefulBits, Depth + 1);
// The interesting part was starting at Imm in the argument
OpUsefulBits <<= Imm;
} else {
OpUsefulBits <<= MSB + 1;
--OpUsefulBits;
// The interesting part will be shifted in the result
OpUsefulBits <<= OpUsefulBits.getBitWidth() - Imm;
getUsefulBits(Op, OpUsefulBits, Depth + 1);
// The interesting part was at zero in the argument
OpUsefulBits.lshrInPlace(OpUsefulBits.getBitWidth() - Imm);
}
UsefulBits &= OpUsefulBits;
}
static void getUsefulBitsFromUBFM(SDValue Op, APInt &UsefulBits,
unsigned Depth) {
uint64_t Imm =
cast<const ConstantSDNode>(Op.getOperand(1).getNode())->getZExtValue();
uint64_t MSB =
cast<const ConstantSDNode>(Op.getOperand(2).getNode())->getZExtValue();
getUsefulBitsFromBitfieldMoveOpd(Op, UsefulBits, Imm, MSB, Depth);
}
static void getUsefulBitsFromOrWithShiftedReg(SDValue Op, APInt &UsefulBits,
unsigned Depth) {
uint64_t ShiftTypeAndValue =
cast<const ConstantSDNode>(Op.getOperand(2).getNode())->getZExtValue();
APInt Mask(UsefulBits);
Mask.clearAllBits();
Mask.flipAllBits();
if (AArch64_AM::getShiftType(ShiftTypeAndValue) == AArch64_AM::LSL) {
// Shift Left
uint64_t ShiftAmt = AArch64_AM::getShiftValue(ShiftTypeAndValue);
Mask <<= ShiftAmt;
getUsefulBits(Op, Mask, Depth + 1);
Mask.lshrInPlace(ShiftAmt);
} else if (AArch64_AM::getShiftType(ShiftTypeAndValue) == AArch64_AM::LSR) {
// Shift Right
// We do not handle AArch64_AM::ASR, because the sign will change the
// number of useful bits
uint64_t ShiftAmt = AArch64_AM::getShiftValue(ShiftTypeAndValue);
Mask.lshrInPlace(ShiftAmt);
getUsefulBits(Op, Mask, Depth + 1);
Mask <<= ShiftAmt;
} else
return;
UsefulBits &= Mask;
}
static void getUsefulBitsFromBFM(SDValue Op, SDValue Orig, APInt &UsefulBits,
unsigned Depth) {
uint64_t Imm =
cast<const ConstantSDNode>(Op.getOperand(2).getNode())->getZExtValue();
uint64_t MSB =
cast<const ConstantSDNode>(Op.getOperand(3).getNode())->getZExtValue();
APInt OpUsefulBits(UsefulBits);
OpUsefulBits = 1;
APInt ResultUsefulBits(UsefulBits.getBitWidth(), 0);
ResultUsefulBits.flipAllBits();
APInt Mask(UsefulBits.getBitWidth(), 0);
getUsefulBits(Op, ResultUsefulBits, Depth + 1);
if (MSB >= Imm) {
// The instruction is a BFXIL.
uint64_t Width = MSB - Imm + 1;
uint64_t LSB = Imm;
OpUsefulBits <<= Width;
--OpUsefulBits;
if (Op.getOperand(1) == Orig) {
// Copy the low bits from the result to bits starting from LSB.
Mask = ResultUsefulBits & OpUsefulBits;
Mask <<= LSB;
}
if (Op.getOperand(0) == Orig)
// Bits starting from LSB in the input contribute to the result.
Mask |= (ResultUsefulBits & ~OpUsefulBits);
} else {
// The instruction is a BFI.
uint64_t Width = MSB + 1;
uint64_t LSB = UsefulBits.getBitWidth() - Imm;
OpUsefulBits <<= Width;
--OpUsefulBits;
OpUsefulBits <<= LSB;
if (Op.getOperand(1) == Orig) {
// Copy the bits from the result to the zero bits.
Mask = ResultUsefulBits & OpUsefulBits;
Mask.lshrInPlace(LSB);
}
if (Op.getOperand(0) == Orig)
Mask |= (ResultUsefulBits & ~OpUsefulBits);
}
UsefulBits &= Mask;
}
static void getUsefulBitsForUse(SDNode *UserNode, APInt &UsefulBits,
SDValue Orig, unsigned Depth) {
// Users of this node should have already been instruction selected
// FIXME: Can we turn that into an assert?
if (!UserNode->isMachineOpcode())
return;
switch (UserNode->getMachineOpcode()) {
default:
return;
case AArch64::ANDSWri:
case AArch64::ANDSXri:
case AArch64::ANDWri:
case AArch64::ANDXri:
// We increment Depth only when we call the getUsefulBits
return getUsefulBitsFromAndWithImmediate(SDValue(UserNode, 0), UsefulBits,
Depth);
case AArch64::UBFMWri:
case AArch64::UBFMXri:
return getUsefulBitsFromUBFM(SDValue(UserNode, 0), UsefulBits, Depth);
case AArch64::ORRWrs:
case AArch64::ORRXrs:
if (UserNode->getOperand(1) != Orig)
return;
return getUsefulBitsFromOrWithShiftedReg(SDValue(UserNode, 0), UsefulBits,
Depth);
case AArch64::BFMWri:
case AArch64::BFMXri:
return getUsefulBitsFromBFM(SDValue(UserNode, 0), Orig, UsefulBits, Depth);
case AArch64::STRBBui:
case AArch64::STURBBi:
if (UserNode->getOperand(0) != Orig)
return;
UsefulBits &= APInt(UsefulBits.getBitWidth(), 0xff);
return;
case AArch64::STRHHui:
case AArch64::STURHHi:
if (UserNode->getOperand(0) != Orig)
return;
UsefulBits &= APInt(UsefulBits.getBitWidth(), 0xffff);
return;
}
}
static void getUsefulBits(SDValue Op, APInt &UsefulBits, unsigned Depth) {
if (Depth >= SelectionDAG::MaxRecursionDepth)
return;
// Initialize UsefulBits
if (!Depth) {
unsigned Bitwidth = Op.getScalarValueSizeInBits();
// At the beginning, assume every produced bits is useful
UsefulBits = APInt(Bitwidth, 0);
UsefulBits.flipAllBits();
}
APInt UsersUsefulBits(UsefulBits.getBitWidth(), 0);
for (SDNode *Node : Op.getNode()->uses()) {
// A use cannot produce useful bits
APInt UsefulBitsForUse = APInt(UsefulBits);
getUsefulBitsForUse(Node, UsefulBitsForUse, Op, Depth);
UsersUsefulBits |= UsefulBitsForUse;
}
// UsefulBits contains the produced bits that are meaningful for the
// current definition, thus a user cannot make a bit meaningful at
// this point
UsefulBits &= UsersUsefulBits;
}
/// Create a machine node performing a notional SHL of Op by ShlAmount. If
/// ShlAmount is negative, do a (logical) right-shift instead. If ShlAmount is
/// 0, return Op unchanged.
static SDValue getLeftShift(SelectionDAG *CurDAG, SDValue Op, int ShlAmount) {
if (ShlAmount == 0)
return Op;
EVT VT = Op.getValueType();
SDLoc dl(Op);
unsigned BitWidth = VT.getSizeInBits();
unsigned UBFMOpc = BitWidth == 32 ? AArch64::UBFMWri : AArch64::UBFMXri;
SDNode *ShiftNode;
if (ShlAmount > 0) {
// LSL wD, wN, #Amt == UBFM wD, wN, #32-Amt, #31-Amt
ShiftNode = CurDAG->getMachineNode(
UBFMOpc, dl, VT, Op,
CurDAG->getTargetConstant(BitWidth - ShlAmount, dl, VT),
CurDAG->getTargetConstant(BitWidth - 1 - ShlAmount, dl, VT));
} else {
// LSR wD, wN, #Amt == UBFM wD, wN, #Amt, #32-1
assert(ShlAmount < 0 && "expected right shift");
int ShrAmount = -ShlAmount;
ShiftNode = CurDAG->getMachineNode(
UBFMOpc, dl, VT, Op, CurDAG->getTargetConstant(ShrAmount, dl, VT),
CurDAG->getTargetConstant(BitWidth - 1, dl, VT));
}
return SDValue(ShiftNode, 0);
}
/// Does this tree qualify as an attempt to move a bitfield into position,
/// essentially "(and (shl VAL, N), Mask)".
static bool isBitfieldPositioningOp(SelectionDAG *CurDAG, SDValue Op,
bool BiggerPattern,
SDValue &Src, int &ShiftAmount,
int &MaskWidth) {
EVT VT = Op.getValueType();
unsigned BitWidth = VT.getSizeInBits();
(void)BitWidth;
assert(BitWidth == 32 || BitWidth == 64);
KnownBits Known = CurDAG->computeKnownBits(Op);
// Non-zero in the sense that they're not provably zero, which is the key
// point if we want to use this value
uint64_t NonZeroBits = (~Known.Zero).getZExtValue();
// Discard a constant AND mask if present. It's safe because the node will
// already have been factored into the computeKnownBits calculation above.
uint64_t AndImm;
if (isOpcWithIntImmediate(Op.getNode(), ISD::AND, AndImm)) {
assert((~APInt(BitWidth, AndImm) & ~Known.Zero) == 0);
Op = Op.getOperand(0);
}
// Don't match if the SHL has more than one use, since then we'll end up
// generating SHL+UBFIZ instead of just keeping SHL+AND.
if (!BiggerPattern && !Op.hasOneUse())
return false;
uint64_t ShlImm;
if (!isOpcWithIntImmediate(Op.getNode(), ISD::SHL, ShlImm))
return false;
Op = Op.getOperand(0);
if (!isShiftedMask_64(NonZeroBits))
return false;
ShiftAmount = countTrailingZeros(NonZeroBits);
MaskWidth = countTrailingOnes(NonZeroBits >> ShiftAmount);
// BFI encompasses sufficiently many nodes that it's worth inserting an extra
// LSL/LSR if the mask in NonZeroBits doesn't quite match up with the ISD::SHL
// amount. BiggerPattern is true when this pattern is being matched for BFI,
// BiggerPattern is false when this pattern is being matched for UBFIZ, in
// which case it is not profitable to insert an extra shift.
if (ShlImm - ShiftAmount != 0 && !BiggerPattern)
return false;
Src = getLeftShift(CurDAG, Op, ShlImm - ShiftAmount);
return true;
}
static bool isShiftedMask(uint64_t Mask, EVT VT) {
assert(VT == MVT::i32 || VT == MVT::i64);
if (VT == MVT::i32)
return isShiftedMask_32(Mask);
return isShiftedMask_64(Mask);
}
// Generate a BFI/BFXIL from 'or (and X, MaskImm), OrImm' iff the value being
// inserted only sets known zero bits.
static bool tryBitfieldInsertOpFromOrAndImm(SDNode *N, SelectionDAG *CurDAG) {
assert(N->getOpcode() == ISD::OR && "Expect a OR operation");
EVT VT = N->getValueType(0);
if (VT != MVT::i32 && VT != MVT::i64)
return false;
unsigned BitWidth = VT.getSizeInBits();
uint64_t OrImm;
if (!isOpcWithIntImmediate(N, ISD::OR, OrImm))
return false;
// Skip this transformation if the ORR immediate can be encoded in the ORR.
// Otherwise, we'll trade an AND+ORR for ORR+BFI/BFXIL, which is most likely
// performance neutral.
if (AArch64_AM::isLogicalImmediate(OrImm, BitWidth))
return false;
uint64_t MaskImm;
SDValue And = N->getOperand(0);
// Must be a single use AND with an immediate operand.
if (!And.hasOneUse() ||
!isOpcWithIntImmediate(And.getNode(), ISD::AND, MaskImm))
return false;
// Compute the Known Zero for the AND as this allows us to catch more general
// cases than just looking for AND with imm.
KnownBits Known = CurDAG->computeKnownBits(And);
// Non-zero in the sense that they're not provably zero, which is the key
// point if we want to use this value.
uint64_t NotKnownZero = (~Known.Zero).getZExtValue();
// The KnownZero mask must be a shifted mask (e.g., 1110..011, 11100..00).
if (!isShiftedMask(Known.Zero.getZExtValue(), VT))