Google Git
Sign in
llvm / llvm-project / b9ca4c5fab4023e01d14d08f29b13e0c9b4dcad1 / . / llvm / lib / Target / AArch64 / AArch64ISelLowering.cpp
blob: c8b1eafd35495aa0c73bfe193d62977575be5949 [file] [log] [blame]
//===-- AArch64ISelLowering.cpp - AArch64 DAG Lowering Implementation ----===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file implements the AArch64TargetLowering class.
//
//===----------------------------------------------------------------------===//
#include "AArch64ISelLowering.h"
#include "AArch64CallingConvention.h"
#include "AArch64ExpandImm.h"
#include "AArch64MachineFunctionInfo.h"
#include "AArch64PerfectShuffle.h"
#include "AArch64RegisterInfo.h"
#include "AArch64Subtarget.h"
#include "MCTargetDesc/AArch64AddressingModes.h"
#include "Utils/AArch64BaseInfo.h"
#include "Utils/AArch64SMEAttributes.h"
#include "llvm/ADT/APFloat.h"
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/ArrayRef.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/SmallVectorExtras.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/ADT/Twine.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/MemoryLocation.h"
#include "llvm/Analysis/ObjCARCUtil.h"
#include "llvm/Analysis/OptimizationRemarkEmitter.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/Analysis/VectorUtils.h"
#include "llvm/CodeGen/Analysis.h"
#include "llvm/CodeGen/CallingConvLower.h"
#include "llvm/CodeGen/ComplexDeinterleavingPass.h"
#include "llvm/CodeGen/GlobalISel/Utils.h"
#include "llvm/CodeGen/ISDOpcodes.h"
#include "llvm/CodeGen/MachineBasicBlock.h"
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineFunction.h"
#include "llvm/CodeGen/MachineInstr.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineMemOperand.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/SelectionDAG.h"
#include "llvm/CodeGen/SelectionDAGNodes.h"
#include "llvm/CodeGen/TargetCallingConv.h"
#include "llvm/CodeGen/TargetInstrInfo.h"
#include "llvm/CodeGen/TargetOpcodes.h"
#include "llvm/CodeGen/ValueTypes.h"
#include "llvm/CodeGenTypes/MachineValueType.h"
#include "llvm/IR/Attributes.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DebugLoc.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalValue.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Instruction.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/IntrinsicsAArch64.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/PatternMatch.h"
#include "llvm/IR/Type.h"
#include "llvm/IR/Use.h"
#include "llvm/IR/Value.h"
#include "llvm/Support/AtomicOrdering.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/CodeGen.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/InstructionCost.h"
#include "llvm/Support/KnownBits.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Support/SipHash.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Target/TargetOptions.h"
#include "llvm/TargetParser/Triple.h"
#include <algorithm>
#include <bitset>
#include <cassert>
#include <cctype>
#include <cstdint>
#include <cstdlib>
#include <iterator>
#include <limits>
#include <optional>
#include <tuple>
#include <utility>
#include <vector>
using namespace llvm;
using namespace llvm::PatternMatch;
#define DEBUG_TYPE "aarch64-lower"
STATISTIC(NumTailCalls, "Number of tail calls");
STATISTIC(NumOptimizedImms, "Number of times immediates were optimized");
// FIXME: The necessary dtprel relocations don't seem to be supported
// well in the GNU bfd and gold linkers at the moment. Therefore, by
// default, for now, fall back to GeneralDynamic code generation.
cl::opt<bool> EnableAArch64ELFLocalDynamicTLSGeneration(
"aarch64-elf-ldtls-generation", cl::Hidden,
cl::desc("Allow AArch64 Local Dynamic TLS code generation"),
cl::init(false));
static cl::opt<bool>
EnableOptimizeLogicalImm("aarch64-enable-logical-imm", cl::Hidden,
cl::desc("Enable AArch64 logical imm instruction "
"optimization"),
cl::init(true));
// Temporary option added for the purpose of testing functionality added
// to DAGCombiner.cpp in D92230. It is expected that this can be removed
// in future when both implementations will be based off MGATHER rather
// than the GLD1 nodes added for the SVE gather load intrinsics.
static cl::opt<bool>
EnableCombineMGatherIntrinsics("aarch64-enable-mgather-combine", cl::Hidden,
cl::desc("Combine extends of AArch64 masked "
"gather intrinsics"),
cl::init(true));
static cl::opt<bool> EnableExtToTBL("aarch64-enable-ext-to-tbl", cl::Hidden,
cl::desc("Combine ext and trunc to TBL"),
cl::init(true));
// All of the XOR, OR and CMP use ALU ports, and data dependency will become the
// bottleneck after this transform on high end CPU. So this max leaf node
// limitation is guard cmp+ccmp will be profitable.
static cl::opt<unsigned> MaxXors("aarch64-max-xors", cl::init(16), cl::Hidden,
cl::desc("Maximum of xors"));
// By turning this on, we will not fallback to DAG ISel when encountering
// scalable vector types for all instruction, even if SVE is not yet supported
// with some instructions.
// See [AArch64TargetLowering::fallbackToDAGISel] for implementation details.
cl::opt<bool> EnableSVEGISel(
"aarch64-enable-gisel-sve", cl::Hidden,
cl::desc("Enable / disable SVE scalable vectors in Global ISel"),
cl::init(false));
// TODO: This option should be removed once we switch to always using PTRADD in
// the SelectionDAG.
static cl::opt<bool> UseFEATCPACodegen(
"aarch64-use-featcpa-codegen", cl::Hidden,
cl::desc("Generate ISD::PTRADD nodes for pointer arithmetic in "
"SelectionDAG for FEAT_CPA"),
cl::init(false));
/// Value type used for condition codes.
static const MVT MVT_CC = MVT::i32;
static const MCPhysReg GPRArgRegs[] = {AArch64::X0, AArch64::X1, AArch64::X2,
AArch64::X3, AArch64::X4, AArch64::X5,
AArch64::X6, AArch64::X7};
static const MCPhysReg FPRArgRegs[] = {AArch64::Q0, AArch64::Q1, AArch64::Q2,
AArch64::Q3, AArch64::Q4, AArch64::Q5,
AArch64::Q6, AArch64::Q7};
ArrayRef<MCPhysReg> llvm::AArch64::getGPRArgRegs() { return GPRArgRegs; }
ArrayRef<MCPhysReg> llvm::AArch64::getFPRArgRegs() { return FPRArgRegs; }
static inline EVT getPackedSVEVectorVT(EVT VT) {
switch (VT.getSimpleVT().SimpleTy) {
default:
llvm_unreachable("unexpected element type for vector");
case MVT::i8:
return MVT::nxv16i8;
case MVT::i16:
return MVT::nxv8i16;
case MVT::i32:
return MVT::nxv4i32;
case MVT::i64:
return MVT::nxv2i64;
case MVT::f16:
return MVT::nxv8f16;
case MVT::f32:
return MVT::nxv4f32;
case MVT::f64:
return MVT::nxv2f64;
case MVT::bf16:
return MVT::nxv8bf16;
}
}
// NOTE: Currently there's only a need to return integer vector types. If this
// changes then just add an extra "type" parameter.
static inline EVT getPackedSVEVectorVT(ElementCount EC) {
switch (EC.getKnownMinValue()) {
default:
llvm_unreachable("unexpected element count for vector");
case 16:
return MVT::nxv16i8;
case 8:
return MVT::nxv8i16;
case 4:
return MVT::nxv4i32;
case 2:
return MVT::nxv2i64;
}
}
static inline EVT getPromotedVTForPredicate(EVT VT) {
assert(VT.isScalableVector() && (VT.getVectorElementType() == MVT::i1) &&
"Expected scalable predicate vector type!");
switch (VT.getVectorMinNumElements()) {
default:
llvm_unreachable("unexpected element count for vector");
case 2:
return MVT::nxv2i64;
case 4:
return MVT::nxv4i32;
case 8:
return MVT::nxv8i16;
case 16:
return MVT::nxv16i8;
}
}
/// Returns true if VT's elements occupy the lowest bit positions of its
/// associated register class without any intervening space.
///
/// For example, nxv2f16, nxv4f16 and nxv8f16 are legal types that belong to the
/// same register class, but only nxv8f16 can be treated as a packed vector.
static inline bool isPackedVectorType(EVT VT, SelectionDAG &DAG) {
assert(VT.isVector() && DAG.getTargetLoweringInfo().isTypeLegal(VT) &&
"Expected legal vector type!");
return VT.isFixedLengthVector() ||
VT.getSizeInBits().getKnownMinValue() == AArch64::SVEBitsPerBlock;
}
// Returns true for ####_MERGE_PASSTHRU opcodes, whose operands have a leading
// predicate and end with a passthru value matching the result type.
static bool isMergePassthruOpcode(unsigned Opc) {
switch (Opc) {
default:
return false;
case AArch64ISD::BITREVERSE_MERGE_PASSTHRU:
case AArch64ISD::BSWAP_MERGE_PASSTHRU:
case AArch64ISD::REVH_MERGE_PASSTHRU:
case AArch64ISD::REVW_MERGE_PASSTHRU:
case AArch64ISD::REVD_MERGE_PASSTHRU:
case AArch64ISD::CTLZ_MERGE_PASSTHRU:
case AArch64ISD::CTPOP_MERGE_PASSTHRU:
case AArch64ISD::DUP_MERGE_PASSTHRU:
case AArch64ISD::ABS_MERGE_PASSTHRU:
case AArch64ISD::NEG_MERGE_PASSTHRU:
case AArch64ISD::FNEG_MERGE_PASSTHRU:
case AArch64ISD::SIGN_EXTEND_INREG_MERGE_PASSTHRU:
case AArch64ISD::ZERO_EXTEND_INREG_MERGE_PASSTHRU:
case AArch64ISD::FCEIL_MERGE_PASSTHRU:
case AArch64ISD::FFLOOR_MERGE_PASSTHRU:
case AArch64ISD::FNEARBYINT_MERGE_PASSTHRU:
case AArch64ISD::FRINT_MERGE_PASSTHRU:
case AArch64ISD::FROUND_MERGE_PASSTHRU:
case AArch64ISD::FROUNDEVEN_MERGE_PASSTHRU:
case AArch64ISD::FTRUNC_MERGE_PASSTHRU:
case AArch64ISD::FP_ROUND_MERGE_PASSTHRU:
case AArch64ISD::FP_EXTEND_MERGE_PASSTHRU:
case AArch64ISD::SINT_TO_FP_MERGE_PASSTHRU:
case AArch64ISD::UINT_TO_FP_MERGE_PASSTHRU:
case AArch64ISD::FCVTX_MERGE_PASSTHRU:
case AArch64ISD::FCVTZU_MERGE_PASSTHRU:
case AArch64ISD::FCVTZS_MERGE_PASSTHRU:
case AArch64ISD::FSQRT_MERGE_PASSTHRU:
case AArch64ISD::FRECPX_MERGE_PASSTHRU:
case AArch64ISD::FABS_MERGE_PASSTHRU:
return true;
}
}
// Returns true if inactive lanes are known to be zeroed by construction.
static bool isZeroingInactiveLanes(SDValue Op) {
switch (Op.getOpcode()) {
default:
return false;
// We guarantee i1 splat_vectors to zero the other lanes
case ISD::SPLAT_VECTOR:
case ISD::GET_ACTIVE_LANE_MASK:
case AArch64ISD::PTRUE:
case AArch64ISD::SETCC_MERGE_ZERO:
return true;
case ISD::INTRINSIC_WO_CHAIN:
switch (Op.getConstantOperandVal(0)) {
default:
return false;
case Intrinsic::aarch64_sve_ptrue:
case Intrinsic::aarch64_sve_pnext:
case Intrinsic::aarch64_sve_cmpeq:
case Intrinsic::aarch64_sve_cmpne:
case Intrinsic::aarch64_sve_cmpge:
case Intrinsic::aarch64_sve_cmpgt:
case Intrinsic::aarch64_sve_cmphs:
case Intrinsic::aarch64_sve_cmphi:
case Intrinsic::aarch64_sve_cmpeq_wide:
case Intrinsic::aarch64_sve_cmpne_wide:
case Intrinsic::aarch64_sve_cmpge_wide:
case Intrinsic::aarch64_sve_cmpgt_wide:
case Intrinsic::aarch64_sve_cmplt_wide:
case Intrinsic::aarch64_sve_cmple_wide:
case Intrinsic::aarch64_sve_cmphs_wide:
case Intrinsic::aarch64_sve_cmphi_wide:
case Intrinsic::aarch64_sve_cmplo_wide:
case Intrinsic::aarch64_sve_cmpls_wide:
case Intrinsic::aarch64_sve_fcmpeq:
case Intrinsic::aarch64_sve_fcmpne:
case Intrinsic::aarch64_sve_fcmpge:
case Intrinsic::aarch64_sve_fcmpgt:
case Intrinsic::aarch64_sve_fcmpuo:
case Intrinsic::aarch64_sve_facgt:
case Intrinsic::aarch64_sve_facge:
case Intrinsic::aarch64_sve_whilege:
case Intrinsic::aarch64_sve_whilegt:
case Intrinsic::aarch64_sve_whilehi:
case Intrinsic::aarch64_sve_whilehs:
case Intrinsic::aarch64_sve_whilele:
case Intrinsic::aarch64_sve_whilelo:
case Intrinsic::aarch64_sve_whilels:
case Intrinsic::aarch64_sve_whilelt:
case Intrinsic::aarch64_sve_match:
case Intrinsic::aarch64_sve_nmatch:
case Intrinsic::aarch64_sve_whilege_x2:
case Intrinsic::aarch64_sve_whilegt_x2:
case Intrinsic::aarch64_sve_whilehi_x2:
case Intrinsic::aarch64_sve_whilehs_x2:
case Intrinsic::aarch64_sve_whilele_x2:
case Intrinsic::aarch64_sve_whilelo_x2:
case Intrinsic::aarch64_sve_whilels_x2:
case Intrinsic::aarch64_sve_whilelt_x2:
return true;
}
}
}
static std::tuple<SDValue, SDValue>
extractPtrauthBlendDiscriminators(SDValue Disc, SelectionDAG *DAG) {
SDLoc DL(Disc);
SDValue AddrDisc;
SDValue ConstDisc;
// If this is a blend, remember the constant and address discriminators.
// Otherwise, it's either a constant discriminator, or a non-blended
// address discriminator.
if (Disc->getOpcode() == ISD::INTRINSIC_WO_CHAIN &&
Disc->getConstantOperandVal(0) == Intrinsic::ptrauth_blend) {
AddrDisc = Disc->getOperand(1);
ConstDisc = Disc->getOperand(2);
} else {
ConstDisc = Disc;
}
// If the constant discriminator (either the blend RHS, or the entire
// discriminator value) isn't a 16-bit constant, bail out, and let the
// discriminator be computed separately.
const auto *ConstDiscN = dyn_cast<ConstantSDNode>(ConstDisc);
if (!ConstDiscN || !isUInt<16>(ConstDiscN->getZExtValue()))
return std::make_tuple(DAG->getTargetConstant(0, DL, MVT::i64), Disc);
// If there's no address discriminator, use NoRegister, which we'll later
// replace with XZR, or directly use a Z variant of the inst. when available.
if (!AddrDisc)
AddrDisc = DAG->getRegister(AArch64::NoRegister, MVT::i64);
return std::make_tuple(
DAG->getTargetConstant(ConstDiscN->getZExtValue(), DL, MVT::i64),
AddrDisc);
}
AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
const AArch64Subtarget &STI)
: TargetLowering(TM), Subtarget(&STI) {
// AArch64 doesn't have comparisons which set GPRs or setcc instructions, so
// we have to make something up. Arbitrarily, choose ZeroOrOne.
setBooleanContents(ZeroOrOneBooleanContent);
// When comparing vectors the result sets the different elements in the
// vector to all-one or all-zero.
setBooleanVectorContents(ZeroOrNegativeOneBooleanContent);
// Set up the register classes.
addRegisterClass(MVT::i32, &AArch64::GPR32allRegClass);
addRegisterClass(MVT::i64, &AArch64::GPR64allRegClass);
if (Subtarget->hasLS64()) {
addRegisterClass(MVT::i64x8, &AArch64::GPR64x8ClassRegClass);
setOperationAction(ISD::LOAD, MVT::i64x8, Custom);
setOperationAction(ISD::STORE, MVT::i64x8, Custom);
}
if (Subtarget->hasFPARMv8()) {
addRegisterClass(MVT::aarch64mfp8, &AArch64::FPR8RegClass);
addRegisterClass(MVT::f16, &AArch64::FPR16RegClass);
addRegisterClass(MVT::bf16, &AArch64::FPR16RegClass);
addRegisterClass(MVT::f32, &AArch64::FPR32RegClass);
addRegisterClass(MVT::f64, &AArch64::FPR64RegClass);
addRegisterClass(MVT::f128, &AArch64::FPR128RegClass);
}
if (Subtarget->hasNEON()) {
addRegisterClass(MVT::v16i8, &AArch64::FPR8RegClass);
addRegisterClass(MVT::v8i16, &AArch64::FPR16RegClass);
addDRType(MVT::v2f32);
addDRType(MVT::v8i8);
addDRType(MVT::v4i16);
addDRType(MVT::v2i32);
addDRType(MVT::v1i64);
addDRType(MVT::v1f64);
addDRType(MVT::v4f16);
addDRType(MVT::v4bf16);
addQRType(MVT::v4f32);
addQRType(MVT::v2f64);
addQRType(MVT::v16i8);
addQRType(MVT::v8i16);
addQRType(MVT::v4i32);
addQRType(MVT::v2i64);
addQRType(MVT::v8f16);
addQRType(MVT::v8bf16);
}
if (Subtarget->isSVEorStreamingSVEAvailable()) {
// Add legal sve predicate types
addRegisterClass(MVT::nxv1i1, &AArch64::PPRRegClass);
addRegisterClass(MVT::nxv2i1, &AArch64::PPRRegClass);
addRegisterClass(MVT::nxv4i1, &AArch64::PPRRegClass);
addRegisterClass(MVT::nxv8i1, &AArch64::PPRRegClass);
addRegisterClass(MVT::nxv16i1, &AArch64::PPRRegClass);
// Add legal sve data types
addRegisterClass(MVT::nxv16i8, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv8i16, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv4i32, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv2i64, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv2f16, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv4f16, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv8f16, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv2f32, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv4f32, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv2f64, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv2bf16, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv4bf16, &AArch64::ZPRRegClass);
addRegisterClass(MVT::nxv8bf16, &AArch64::ZPRRegClass);
if (Subtarget->useSVEForFixedLengthVectors()) {
for (MVT VT : MVT::integer_fixedlen_vector_valuetypes())
if (useSVEForFixedLengthVectorVT(VT))
addRegisterClass(VT, &AArch64::ZPRRegClass);
for (MVT VT : MVT::fp_fixedlen_vector_valuetypes())
if (useSVEForFixedLengthVectorVT(VT))
addRegisterClass(VT, &AArch64::ZPRRegClass);
}
}
if (Subtarget->hasSVE2p1() || Subtarget->hasSME2()) {
addRegisterClass(MVT::aarch64svcount, &AArch64::PPRRegClass);
setOperationPromotedToType(ISD::LOAD, MVT::aarch64svcount, MVT::nxv16i1);
setOperationPromotedToType(ISD::STORE, MVT::aarch64svcount, MVT::nxv16i1);
setOperationAction(ISD::SELECT, MVT::aarch64svcount, Custom);
setOperationAction(ISD::SELECT_CC, MVT::aarch64svcount, Expand);
}
// Compute derived properties from the register classes
computeRegisterProperties(Subtarget->getRegisterInfo());
// Provide all sorts of operation actions
setOperationAction(ISD::GlobalAddress, MVT::i64, Custom);
setOperationAction(ISD::GlobalTLSAddress, MVT::i64, Custom);
setOperationAction(ISD::SETCC, MVT::i32, Custom);
setOperationAction(ISD::SETCC, MVT::i64, Custom);
setOperationAction(ISD::SETCC, MVT::bf16, Custom);
setOperationAction(ISD::SETCC, MVT::f16, Custom);
setOperationAction(ISD::SETCC, MVT::f32, Custom);
setOperationAction(ISD::SETCC, MVT::f64, Custom);
setOperationAction(ISD::STRICT_FSETCC, MVT::bf16, Custom);
setOperationAction(ISD::STRICT_FSETCC, MVT::f16, Custom);
setOperationAction(ISD::STRICT_FSETCC, MVT::f32, Custom);
setOperationAction(ISD::STRICT_FSETCC, MVT::f64, Custom);
setOperationAction(ISD::STRICT_FSETCCS, MVT::f16, Custom);
setOperationAction(ISD::STRICT_FSETCCS, MVT::f32, Custom);
setOperationAction(ISD::STRICT_FSETCCS, MVT::f64, Custom);
setOperationAction(ISD::BITREVERSE, MVT::i32, Legal);
setOperationAction(ISD::BITREVERSE, MVT::i64, Legal);
setOperationAction(ISD::BRCOND, MVT::Other, Custom);
setOperationAction(ISD::BR_CC, MVT::i32, Custom);
setOperationAction(ISD::BR_CC, MVT::i64, Custom);
setOperationAction(ISD::BR_CC, MVT::f16, Custom);
setOperationAction(ISD::BR_CC, MVT::f32, Custom);
setOperationAction(ISD::BR_CC, MVT::f64, Custom);
setOperationAction(ISD::SELECT, MVT::i32, Custom);
setOperationAction(ISD::SELECT, MVT::i64, Custom);
if (Subtarget->hasFPARMv8()) {
setOperationAction(ISD::SELECT, MVT::f16, Custom);
setOperationAction(ISD::SELECT, MVT::bf16, Custom);
}
setOperationAction(ISD::SELECT, MVT::f32, Custom);
setOperationAction(ISD::SELECT, MVT::f64, Custom);
setOperationAction(ISD::SELECT_CC, MVT::i32, Custom);
setOperationAction(ISD::SELECT_CC, MVT::i64, Custom);
setOperationAction(ISD::SELECT_CC, MVT::f16, Custom);
setOperationAction(ISD::SELECT_CC, MVT::bf16, Custom);
setOperationAction(ISD::SELECT_CC, MVT::f32, Custom);
setOperationAction(ISD::SELECT_CC, MVT::f64, Custom);
setOperationAction(ISD::BR_JT, MVT::Other, Custom);
setOperationAction(ISD::JumpTable, MVT::i64, Custom);
setOperationAction(ISD::BRIND, MVT::Other, Custom);
setOperationAction(ISD::SETCCCARRY, MVT::i64, Custom);
setOperationAction(ISD::PtrAuthGlobalAddress, MVT::i64, Custom);
setOperationAction(ISD::SHL_PARTS, MVT::i64, Custom);
setOperationAction(ISD::SRA_PARTS, MVT::i64, Custom);
setOperationAction(ISD::SRL_PARTS, MVT::i64, Custom);
setOperationAction(ISD::FREM, MVT::f32, Expand);
setOperationAction(ISD::FREM, MVT::f64, Expand);
setOperationAction(ISD::FREM, MVT::f80, Expand);
setOperationAction(ISD::BUILD_PAIR, MVT::i64, Expand);
// Custom lowering hooks are needed for XOR
// to fold it into CSINC/CSINV.
setOperationAction(ISD::XOR, MVT::i32, Custom);
setOperationAction(ISD::XOR, MVT::i64, Custom);
setOperationAction(ISD::ADDRSPACECAST, MVT::i32, Custom);
setOperationAction(ISD::ADDRSPACECAST, MVT::i64, Custom);
// Virtually no operation on f128 is legal, but LLVM can't expand them when
// there's a valid register class, so we need custom operations in most cases.
setOperationAction(ISD::FABS, MVT::f128, Expand);
setOperationAction(ISD::FADD, MVT::f128, LibCall);
setOperationAction(ISD::FCOPYSIGN, MVT::f128, Expand);
setOperationAction(ISD::FCOS, MVT::f128, Expand);
setOperationAction(ISD::FDIV, MVT::f128, LibCall);
setOperationAction(ISD::FMA, MVT::f128, Expand);
setOperationAction(ISD::FMUL, MVT::f128, LibCall);
setOperationAction(ISD::FNEG, MVT::f128, Expand);
setOperationAction(ISD::FPOW, MVT::f128, Expand);
setOperationAction(ISD::FREM, MVT::f128, Expand);
setOperationAction(ISD::FRINT, MVT::f128, Expand);
setOperationAction(ISD::FSIN, MVT::f128, Expand);
setOperationAction(ISD::FSINCOS, MVT::f128, Expand);
setOperationAction(ISD::FSQRT, MVT::f128, Expand);
setOperationAction(ISD::FSUB, MVT::f128, LibCall);
setOperationAction(ISD::FTAN, MVT::f128, Expand);
setOperationAction(ISD::FTRUNC, MVT::f128, Expand);
setOperationAction(ISD::SETCC, MVT::f128, Custom);
setOperationAction(ISD::STRICT_FSETCC, MVT::f128, Custom);
setOperationAction(ISD::STRICT_FSETCCS, MVT::f128, Custom);
setOperationAction(ISD::BR_CC, MVT::f128, Custom);
setOperationAction(ISD::SELECT, MVT::f128, Custom);
setOperationAction(ISD::SELECT_CC, MVT::f128, Custom);
setOperationAction(ISD::FP_EXTEND, MVT::f128, Custom);
// FIXME: f128 FMINIMUM and FMAXIMUM (including STRICT versions) currently
// aren't handled.
// Lowering for many of the conversions is actually specified by the non-f128
// type. The LowerXXX function will be trivial when f128 isn't involved.
setOperationAction(ISD::FP_TO_SINT, MVT::i32, Custom);
setOperationAction(ISD::FP_TO_SINT, MVT::i64, Custom);
setOperationAction(ISD::FP_TO_SINT, MVT::i128, Custom);
setOperationAction(ISD::STRICT_FP_TO_SINT, MVT::i32, Custom);
setOperationAction(ISD::STRICT_FP_TO_SINT, MVT::i64, Custom);
setOperationAction(ISD::STRICT_FP_TO_SINT, MVT::i128, Custom);
setOperationAction(ISD::FP_TO_UINT, MVT::i32, Custom);
setOperationAction(ISD::FP_TO_UINT, MVT::i64, Custom);
setOperationAction(ISD::FP_TO_UINT, MVT::i128, Custom);
setOperationAction(ISD::STRICT_FP_TO_UINT, MVT::i32, Custom);
setOperationAction(ISD::STRICT_FP_TO_UINT, MVT::i64, Custom);
setOperationAction(ISD::STRICT_FP_TO_UINT, MVT::i128, Custom);
setOperationAction(ISD::SINT_TO_FP, MVT::i32, Custom);
setOperationAction(ISD::SINT_TO_FP, MVT::i64, Custom);
setOperationAction(ISD::SINT_TO_FP, MVT::i128, Custom);
setOperationAction(ISD::STRICT_SINT_TO_FP, MVT::i32, Custom);
setOperationAction(ISD::STRICT_SINT_TO_FP, MVT::i64, Custom);
setOperationAction(ISD::STRICT_SINT_TO_FP, MVT::i128, Custom);
setOperationAction(ISD::UINT_TO_FP, MVT::i32, Custom);
setOperationAction(ISD::UINT_TO_FP, MVT::i64, Custom);
setOperationAction(ISD::UINT_TO_FP, MVT::i128, Custom);
setOperationAction(ISD::STRICT_UINT_TO_FP, MVT::i32, Custom);
setOperationAction(ISD::STRICT_UINT_TO_FP, MVT::i64, Custom);
setOperationAction(ISD::STRICT_UINT_TO_FP, MVT::i128, Custom);
if (Subtarget->hasFPARMv8()) {
setOperationAction(ISD::FP_ROUND, MVT::f16, Custom);
setOperationAction(ISD::FP_ROUND, MVT::bf16, Custom);
}
setOperationAction(ISD::FP_ROUND, MVT::f32, Custom);
setOperationAction(ISD::FP_ROUND, MVT::f64, Custom);
if (Subtarget->hasFPARMv8()) {
setOperationAction(ISD::STRICT_FP_ROUND, MVT::f16, Custom);
setOperationAction(ISD::STRICT_FP_ROUND, MVT::bf16, Custom);
}
setOperationAction(ISD::STRICT_FP_ROUND, MVT::f32, Custom);
setOperationAction(ISD::STRICT_FP_ROUND, MVT::f64, Custom);
setOperationAction(ISD::FP_TO_UINT_SAT, MVT::i32, Custom);
setOperationAction(ISD::FP_TO_UINT_SAT, MVT::i64, Custom);
setOperationAction(ISD::FP_TO_SINT_SAT, MVT::i32, Custom);
setOperationAction(ISD::FP_TO_SINT_SAT, MVT::i64, Custom);
// Variable arguments.
setOperationAction(ISD::VASTART, MVT::Other, Custom);
setOperationAction(ISD::VAARG, MVT::Other, Custom);
setOperationAction(ISD::VACOPY, MVT::Other, Custom);
setOperationAction(ISD::VAEND, MVT::Other, Expand);
// Variable-sized objects.
setOperationAction(ISD::STACKSAVE, MVT::Other, Expand);
setOperationAction(ISD::STACKRESTORE, MVT::Other, Expand);
// Lowering Funnel Shifts to EXTR
setOperationAction(ISD::FSHR, MVT::i32, Custom);
setOperationAction(ISD::FSHR, MVT::i64, Custom);
setOperationAction(ISD::FSHL, MVT::i32, Custom);
setOperationAction(ISD::FSHL, MVT::i64, Custom);
setOperationAction(ISD::DYNAMIC_STACKALLOC, MVT::i64, Custom);
// Constant pool entries
setOperationAction(ISD::ConstantPool, MVT::i64, Custom);
// BlockAddress
setOperationAction(ISD::BlockAddress, MVT::i64, Custom);
// AArch64 lacks both left-rotate and popcount instructions.
setOperationAction(ISD::ROTL, MVT::i32, Expand);
setOperationAction(ISD::ROTL, MVT::i64, Expand);
for (MVT VT : MVT::fixedlen_vector_valuetypes()) {
setOperationAction(ISD::ROTL, VT, Expand);
setOperationAction(ISD::ROTR, VT, Expand);
}
// AArch64 doesn't have i32 MULH{S|U}.
setOperationAction(ISD::MULHU, MVT::i32, Expand);
setOperationAction(ISD::MULHS, MVT::i32, Expand);
// AArch64 doesn't have {U|S}MUL_LOHI.
setOperationAction(ISD::UMUL_LOHI, MVT::i32, Expand);
setOperationAction(ISD::SMUL_LOHI, MVT::i32, Expand);
setOperationAction(ISD::UMUL_LOHI, MVT::i64, Expand);
setOperationAction(ISD::SMUL_LOHI, MVT::i64, Expand);
if (Subtarget->hasCSSC()) {
setOperationAction(ISD::CTPOP, MVT::i32, Legal);
setOperationAction(ISD::CTPOP, MVT::i64, Legal);
setOperationAction(ISD::CTPOP, MVT::i128, Expand);
setOperationAction(ISD::PARITY, MVT::i128, Expand);
setOperationAction(ISD::CTTZ, MVT::i32, Legal);
setOperationAction(ISD::CTTZ, MVT::i64, Legal);
setOperationAction(ISD::CTTZ, MVT::i128, Expand);
setOperationAction(ISD::ABS, MVT::i32, Legal);
setOperationAction(ISD::ABS, MVT::i64, Legal);
setOperationAction(ISD::SMAX, MVT::i32, Legal);
setOperationAction(ISD::SMAX, MVT::i64, Legal);
setOperationAction(ISD::UMAX, MVT::i32, Legal);
setOperationAction(ISD::UMAX, MVT::i64, Legal);
setOperationAction(ISD::SMIN, MVT::i32, Legal);
setOperationAction(ISD::SMIN, MVT::i64, Legal);
setOperationAction(ISD::UMIN, MVT::i32, Legal);
setOperationAction(ISD::UMIN, MVT::i64, Legal);
} else {
setOperationAction(ISD::CTPOP, MVT::i32, Custom);
setOperationAction(ISD::CTPOP, MVT::i64, Custom);
setOperationAction(ISD::CTPOP, MVT::i128, Custom);
setOperationAction(ISD::PARITY, MVT::i64, Custom);
setOperationAction(ISD::PARITY, MVT::i128, Custom);
setOperationAction(ISD::ABS, MVT::i32, Custom);
setOperationAction(ISD::ABS, MVT::i64, Custom);
}
setOperationAction(ISD::SDIVREM, MVT::i32, Expand);
setOperationAction(ISD::SDIVREM, MVT::i64, Expand);
for (MVT VT : MVT::fixedlen_vector_valuetypes()) {
setOperationAction(ISD::SDIVREM, VT, Expand);
setOperationAction(ISD::UDIVREM, VT, Expand);
}
setOperationAction(ISD::SREM, MVT::i32, Expand);
setOperationAction(ISD::SREM, MVT::i64, Expand);
setOperationAction(ISD::UDIVREM, MVT::i32, Expand);
setOperationAction(ISD::UDIVREM, MVT::i64, Expand);
setOperationAction(ISD::UREM, MVT::i32, Expand);
setOperationAction(ISD::UREM, MVT::i64, Expand);
// Custom lower Add/Sub/Mul with overflow.
setOperationAction(ISD::SADDO, MVT::i32, Custom);
setOperationAction(ISD::SADDO, MVT::i64, Custom);
setOperationAction(ISD::UADDO, MVT::i32, Custom);
setOperationAction(ISD::UADDO, MVT::i64, Custom);
setOperationAction(ISD::SSUBO, MVT::i32, Custom);
setOperationAction(ISD::SSUBO, MVT::i64, Custom);
setOperationAction(ISD::USUBO, MVT::i32, Custom);
setOperationAction(ISD::USUBO, MVT::i64, Custom);
setOperationAction(ISD::SMULO, MVT::i32, Custom);
setOperationAction(ISD::SMULO, MVT::i64, Custom);
setOperationAction(ISD::UMULO, MVT::i32, Custom);
setOperationAction(ISD::UMULO, MVT::i64, Custom);
setOperationAction(ISD::UADDO_CARRY, MVT::i32, Custom);
setOperationAction(ISD::UADDO_CARRY, MVT::i64, Custom);
setOperationAction(ISD::USUBO_CARRY, MVT::i32, Custom);
setOperationAction(ISD::USUBO_CARRY, MVT::i64, Custom);
setOperationAction(ISD::SADDO_CARRY, MVT::i32, Custom);
setOperationAction(ISD::SADDO_CARRY, MVT::i64, Custom);
setOperationAction(ISD::SSUBO_CARRY, MVT::i32, Custom);
setOperationAction(ISD::SSUBO_CARRY, MVT::i64, Custom);
setOperationAction(ISD::FSIN, MVT::f32, Expand);
setOperationAction(ISD::FSIN, MVT::f64, Expand);
setOperationAction(ISD::FCOS, MVT::f32, Expand);
setOperationAction(ISD::FCOS, MVT::f64, Expand);
setOperationAction(ISD::FPOW, MVT::f32, Expand);
setOperationAction(ISD::FPOW, MVT::f64, Expand);
setOperationAction(ISD::FCOPYSIGN, MVT::f64, Custom);
setOperationAction(ISD::FCOPYSIGN, MVT::f32, Custom);
if (Subtarget->hasFullFP16()) {
setOperationAction(ISD::FCOPYSIGN, MVT::f16, Custom);
setOperationAction(ISD::FCOPYSIGN, MVT::bf16, Custom);
} else {
setOperationAction(ISD::FCOPYSIGN, MVT::f16, Promote);
setOperationAction(ISD::FCOPYSIGN, MVT::bf16, Promote);
}
for (auto Op : {ISD::FREM, ISD::FPOW, ISD::FPOWI,
ISD::FCOS, ISD::FSIN, ISD::FSINCOS,
ISD::FSINCOSPI, ISD::FMODF, ISD::FACOS,
ISD::FASIN, ISD::FATAN, ISD::FATAN2,
ISD::FCOSH, ISD::FSINH, ISD::FTANH,
ISD::FTAN, ISD::FEXP, ISD::FEXP2,
ISD::FEXP10, ISD::FLOG, ISD::FLOG2,
ISD::FLOG10, ISD::STRICT_FREM, ISD::STRICT_FPOW,
ISD::STRICT_FPOWI, ISD::STRICT_FCOS, ISD::STRICT_FSIN,
ISD::STRICT_FACOS, ISD::STRICT_FASIN, ISD::STRICT_FATAN,
ISD::STRICT_FATAN2, ISD::STRICT_FCOSH, ISD::STRICT_FSINH,
ISD::STRICT_FTANH, ISD::STRICT_FEXP, ISD::STRICT_FEXP2,
ISD::STRICT_FLOG, ISD::STRICT_FLOG2, ISD::STRICT_FLOG10,
ISD::STRICT_FTAN}) {
setOperationAction(Op, MVT::f16, Promote);
setOperationAction(Op, MVT::v4f16, Expand);
setOperationAction(Op, MVT::v8f16, Expand);
setOperationAction(Op, MVT::bf16, Promote);
setOperationAction(Op, MVT::v4bf16, Expand);
setOperationAction(Op, MVT::v8bf16, Expand);
}
// fpextend from f16 or bf16 to f32 is legal
setOperationAction(ISD::FP_EXTEND, MVT::f32, Legal);
setOperationAction(ISD::FP_EXTEND, MVT::v4f32, Legal);
setOperationAction(ISD::STRICT_FP_EXTEND, MVT::f32, Legal);
setOperationAction(ISD::STRICT_FP_EXTEND, MVT::v4f32, Legal);
// fpextend from bf16 to f64 needs to be split into two fpextends
setOperationAction(ISD::FP_EXTEND, MVT::f64, Custom);
setOperationAction(ISD::STRICT_FP_EXTEND, MVT::f64, Custom);
auto LegalizeNarrowFP = [this](MVT ScalarVT) {
for (auto Op : {
ISD::SETCC,
ISD::SELECT_CC,
ISD::BR_CC,
ISD::FADD,
ISD::FSUB,
ISD::FMUL,
ISD::FDIV,
ISD::FMA,
ISD::FCEIL,
ISD::FSQRT,
ISD::FFLOOR,
ISD::FNEARBYINT,
ISD::FRINT,
ISD::FROUND,
ISD::FROUNDEVEN,
ISD::FTRUNC,
ISD::FMINNUM,
ISD::FMAXNUM,
ISD::FMINIMUM,
ISD::FMAXIMUM,
ISD::FMINIMUMNUM,
ISD::FMAXIMUMNUM,
ISD::FCANONICALIZE,
ISD::STRICT_FADD,
ISD::STRICT_FSUB,
ISD::STRICT_FMUL,
ISD::STRICT_FDIV,
ISD::STRICT_FMA,
ISD::STRICT_FCEIL,
ISD::STRICT_FFLOOR,
ISD::STRICT_FSQRT,
ISD::STRICT_FRINT,
ISD::STRICT_FNEARBYINT,
ISD::STRICT_FROUND,
ISD::STRICT_FTRUNC,
ISD::STRICT_FROUNDEVEN,
ISD::STRICT_FMINNUM,
ISD::STRICT_FMAXNUM,
ISD::STRICT_FMINIMUM,
ISD::STRICT_FMAXIMUM,
})
setOperationAction(Op, ScalarVT, Promote);
for (auto Op : {ISD::FNEG, ISD::FABS})
setOperationAction(Op, ScalarVT, Legal);
// Round-to-integer need custom lowering for fp16, as Promote doesn't work
// because the result type is integer.
for (auto Op : {ISD::LROUND, ISD::LLROUND, ISD::LRINT, ISD::LLRINT,
ISD::STRICT_LROUND, ISD::STRICT_LLROUND, ISD::STRICT_LRINT,
ISD::STRICT_LLRINT})
setOperationAction(Op, ScalarVT, Custom);
// promote v4f16 to v4f32 when that is known to be safe.
auto V4Narrow = MVT::getVectorVT(ScalarVT, 4);
setOperationPromotedToType(ISD::FADD, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FSUB, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FMUL, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FDIV, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FCEIL, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FFLOOR, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FROUND, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FTRUNC, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FROUNDEVEN, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FRINT, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FNEARBYINT, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::FCANONICALIZE, V4Narrow, MVT::v4f32);
setOperationPromotedToType(ISD::SETCC, V4Narrow, MVT::v4f32);
setOperationAction(ISD::FABS, V4Narrow, Legal);
setOperationAction(ISD::FNEG, V4Narrow, Legal);
setOperationAction(ISD::FMA, V4Narrow, Expand);
setOperationAction(ISD::BR_CC, V4Narrow, Expand);
setOperationAction(ISD::SELECT, V4Narrow, Expand);
setOperationAction(ISD::SELECT_CC, V4Narrow, Expand);
setOperationAction(ISD::FCOPYSIGN, V4Narrow, Custom);
setOperationAction(ISD::FSQRT, V4Narrow, Expand);
auto V8Narrow = MVT::getVectorVT(ScalarVT, 8);
setOperationPromotedToType(ISD::FCANONICALIZE, V8Narrow, MVT::v8f32);
setOperationPromotedToType(ISD::SETCC, V8Narrow, MVT::v8f32);
setOperationAction(ISD::FABS, V8Narrow, Legal);
setOperationAction(ISD::FADD, V8Narrow, Legal);
setOperationAction(ISD::FCEIL, V8Narrow, Legal);
setOperationAction(ISD::FCOPYSIGN, V8Narrow, Custom);
setOperationAction(ISD::FDIV, V8Narrow, Legal);
setOperationAction(ISD::FFLOOR, V8Narrow, Legal);
setOperationAction(ISD::FMA, V8Narrow, Expand);
setOperationAction(ISD::FMUL, V8Narrow, Legal);
setOperationAction(ISD::FNEARBYINT, V8Narrow, Legal);
setOperationAction(ISD::FNEG, V8Narrow, Legal);
setOperationAction(ISD::FROUND, V8Narrow, Legal);
setOperationAction(ISD::FROUNDEVEN, V8Narrow, Legal);
setOperationAction(ISD::FRINT, V8Narrow, Legal);
setOperationAction(ISD::FSQRT, V8Narrow, Expand);
setOperationAction(ISD::FSUB, V8Narrow, Legal);
setOperationAction(ISD::FTRUNC, V8Narrow, Legal);
setOperationAction(ISD::BR_CC, V8Narrow, Expand);
setOperationAction(ISD::SELECT, V8Narrow, Expand);
setOperationAction(ISD::SELECT_CC, V8Narrow, Expand);
setOperationAction(ISD::FP_EXTEND, V8Narrow, Expand);
};
if (!Subtarget->hasFullFP16()) {
LegalizeNarrowFP(MVT::f16);
}
LegalizeNarrowFP(MVT::bf16);
setOperationAction(ISD::FP_ROUND, MVT::v4f32, Custom);
setOperationAction(ISD::FP_ROUND, MVT::v4bf16, Custom);
// AArch64 has implementations of a lot of rounding-like FP operations.
// clang-format off
for (auto Op :
{ISD::FFLOOR, ISD::FNEARBYINT, ISD::FCEIL,
ISD::FRINT, ISD::FTRUNC, ISD::FROUND,
ISD::FROUNDEVEN, ISD::FMINNUM, ISD::FMAXNUM,
ISD::FMINIMUM, ISD::FMAXIMUM, ISD::LROUND,
ISD::LLROUND, ISD::LRINT, ISD::LLRINT,
ISD::FMINNUM_IEEE, ISD::FMAXNUM_IEEE,
ISD::STRICT_FFLOOR, ISD::STRICT_FCEIL, ISD::STRICT_FNEARBYINT,
ISD::STRICT_FRINT, ISD::STRICT_FTRUNC, ISD::STRICT_FROUNDEVEN,
ISD::STRICT_FROUND, ISD::STRICT_FMINNUM, ISD::STRICT_FMAXNUM,
ISD::STRICT_FMINIMUM, ISD::STRICT_FMAXIMUM, ISD::STRICT_LROUND,
ISD::STRICT_LLROUND, ISD::STRICT_LRINT, ISD::STRICT_LLRINT}) {
for (MVT Ty : {MVT::f32, MVT::f64})
setOperationAction(Op, Ty, Legal);
if (Subtarget->hasFullFP16())
setOperationAction(Op, MVT::f16, Legal);
}
// clang-format on
// Basic strict FP operations are legal
for (auto Op : {ISD::STRICT_FADD, ISD::STRICT_FSUB, ISD::STRICT_FMUL,
ISD::STRICT_FDIV, ISD::STRICT_FMA, ISD::STRICT_FSQRT}) {
for (MVT Ty : {MVT::f32, MVT::f64})
setOperationAction(Op, Ty, Legal);
if (Subtarget->hasFullFP16())
setOperationAction(Op, MVT::f16, Legal);
}
setOperationAction(ISD::PREFETCH, MVT::Other, Custom);
setOperationAction(ISD::GET_ROUNDING, MVT::i32, Custom);
setOperationAction(ISD::SET_ROUNDING, MVT::Other, Custom);
setOperationAction(ISD::GET_FPMODE, MVT::i32, Custom);
setOperationAction(ISD::SET_FPMODE, MVT::i32, Custom);
setOperationAction(ISD::RESET_FPMODE, MVT::Other, Custom);
setOperationAction(ISD::ATOMIC_CMP_SWAP, MVT::i128, Custom);
if (!Subtarget->hasLSE() && !Subtarget->outlineAtomics()) {
setOperationAction(ISD::ATOMIC_LOAD_SUB, MVT::i32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_SUB, MVT::i64, LibCall);
} else {
setOperationAction(ISD::ATOMIC_LOAD_SUB, MVT::i32, Expand);
setOperationAction(ISD::ATOMIC_LOAD_SUB, MVT::i64, Expand);
}
setOperationAction(ISD::ATOMIC_LOAD_AND, MVT::i32, Custom);
setOperationAction(ISD::ATOMIC_LOAD_AND, MVT::i64, Custom);
// Generate outline atomics library calls only if LSE was not specified for
// subtarget
if (Subtarget->outlineAtomics() && !Subtarget->hasLSE()) {
setOperationAction(ISD::ATOMIC_CMP_SWAP, MVT::i8, LibCall);
setOperationAction(ISD::ATOMIC_CMP_SWAP, MVT::i16, LibCall);
setOperationAction(ISD::ATOMIC_CMP_SWAP, MVT::i32, LibCall);
setOperationAction(ISD::ATOMIC_CMP_SWAP, MVT::i64, LibCall);
setOperationAction(ISD::ATOMIC_CMP_SWAP, MVT::i128, LibCall);
setOperationAction(ISD::ATOMIC_SWAP, MVT::i8, LibCall);
setOperationAction(ISD::ATOMIC_SWAP, MVT::i16, LibCall);
setOperationAction(ISD::ATOMIC_SWAP, MVT::i32, LibCall);
setOperationAction(ISD::ATOMIC_SWAP, MVT::i64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_ADD, MVT::i8, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_ADD, MVT::i16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_ADD, MVT::i32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_ADD, MVT::i64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_OR, MVT::i8, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_OR, MVT::i16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_OR, MVT::i32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_OR, MVT::i64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_CLR, MVT::i8, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_CLR, MVT::i16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_CLR, MVT::i32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_CLR, MVT::i64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_XOR, MVT::i8, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_XOR, MVT::i16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_XOR, MVT::i32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_XOR, MVT::i64, LibCall);
}
if (Subtarget->outlineAtomics() && !Subtarget->hasLSFE()) {
setOperationAction(ISD::ATOMIC_LOAD_FADD, MVT::f16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FADD, MVT::f32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FADD, MVT::f64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FADD, MVT::bf16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAX, MVT::f16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAX, MVT::f32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAX, MVT::f64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAX, MVT::bf16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMIN, MVT::f16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMIN, MVT::f32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMIN, MVT::f64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMIN, MVT::bf16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAXIMUM, MVT::f16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAXIMUM, MVT::f32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAXIMUM, MVT::f64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMAXIMUM, MVT::bf16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMINIMUM, MVT::f16, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMINIMUM, MVT::f32, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMINIMUM, MVT::f64, LibCall);
setOperationAction(ISD::ATOMIC_LOAD_FMINIMUM, MVT::bf16, LibCall);
}
if (Subtarget->hasLSE128()) {
// Custom lowering because i128 is not legal. Must be replaced by 2x64
// values. ATOMIC_LOAD_AND also needs op legalisation to emit LDCLRP.
setOperationAction(ISD::ATOMIC_LOAD_AND, MVT::i128, Custom);
setOperationAction(ISD::ATOMIC_LOAD_OR, MVT::i128, Custom);
setOperationAction(ISD::ATOMIC_SWAP, MVT::i128, Custom);
}
// 128-bit loads and stores can be done without expanding
setOperationAction(ISD::LOAD, MVT::i128, Custom);
setOperationAction(ISD::STORE, MVT::i128, Custom);
// Aligned 128-bit loads and stores are single-copy atomic according to the
// v8.4a spec. LRCPC3 introduces 128-bit STILP/LDIAPP but still requires LSE2.
if (Subtarget->hasLSE2()) {
setOperationAction(ISD::ATOMIC_LOAD, MVT::i128, Custom);
setOperationAction(ISD::ATOMIC_STORE, MVT::i128, Custom);
}
// 256 bit non-temporal stores can be lowered to STNP. Do this as part of the
// custom lowering, as there are no un-paired non-temporal stores and
// legalization will break up 256 bit inputs.
setOperationAction(ISD::STORE, MVT::v32i8, Custom);
setOperationAction(ISD::STORE, MVT::v16i16, Custom);
setOperationAction(ISD::STORE, MVT::v16f16, Custom);
setOperationAction(ISD::STORE, MVT::v16bf16, Custom);
setOperationAction(ISD::STORE, MVT::v8i32, Custom);
setOperationAction(ISD::STORE, MVT::v8f32, Custom);
setOperationAction(ISD::STORE, MVT::v4f64, Custom);
setOperationAction(ISD::STORE, MVT::v4i64, Custom);
// 256 bit non-temporal loads can be lowered to LDNP. This is done using
// custom lowering, as there are no un-paired non-temporal loads legalization
// will break up 256 bit inputs.
setOperationAction(ISD::LOAD, MVT::v32i8, Custom);
setOperationAction(ISD::LOAD, MVT::v16i16, Custom);
setOperationAction(ISD::LOAD, MVT::v16f16, Custom);
setOperationAction(ISD::LOAD, MVT::v16bf16, Custom);
setOperationAction(ISD::LOAD, MVT::v8i32, Custom);
setOperationAction(ISD::LOAD, MVT::v8f32, Custom);
setOperationAction(ISD::LOAD, MVT::v4f64, Custom);
setOperationAction(ISD::LOAD, MVT::v4i64, Custom);
// Lower READCYCLECOUNTER using an mrs from CNTVCT_EL0.
setOperationAction(ISD::READCYCLECOUNTER, MVT::i64, Legal);
if (getLibcallName(RTLIB::SINCOS_STRET_F32) != nullptr &&
getLibcallName(RTLIB::SINCOS_STRET_F64) != nullptr) {
// Issue __sincos_stret if available.
setOperationAction(ISD::FSINCOS, MVT::f64, Custom);
setOperationAction(ISD::FSINCOS, MVT::f32, Custom);
} else {
setOperationAction(ISD::FSINCOS, MVT::f64, Expand);
setOperationAction(ISD::FSINCOS, MVT::f32, Expand);
}
// Make floating-point constants legal for the large code model, so they don't
// become loads from the constant pool.
if (Subtarget->isTargetMachO() && TM.getCodeModel() == CodeModel::Large) {
setOperationAction(ISD::ConstantFP, MVT::f32, Legal);
setOperationAction(ISD::ConstantFP, MVT::f64, Legal);
}
// AArch64 does not have floating-point extending loads, i1 sign-extending
// load, floating-point truncating stores, or v2i32->v2i16 truncating store.
for (MVT VT : MVT::fp_valuetypes()) {
setLoadExtAction(ISD::EXTLOAD, VT, MVT::bf16, Expand);
setLoadExtAction(ISD::EXTLOAD, VT, MVT::f16, Expand);
setLoadExtAction(ISD::EXTLOAD, VT, MVT::f32, Expand);
setLoadExtAction(ISD::EXTLOAD, VT, MVT::f64, Expand);
setLoadExtAction(ISD::EXTLOAD, VT, MVT::f80, Expand);
}
for (MVT VT : MVT::integer_valuetypes())
setLoadExtAction(ISD::SEXTLOAD, VT, MVT::i1, Expand);
for (MVT WideVT : MVT::fp_valuetypes()) {
for (MVT NarrowVT : MVT::fp_valuetypes()) {
if (WideVT.getScalarSizeInBits() > NarrowVT.getScalarSizeInBits()) {
setTruncStoreAction(WideVT, NarrowVT, Expand);
}
}
}
if (Subtarget->hasFPARMv8()) {
setOperationAction(ISD::BITCAST, MVT::i16, Custom);
setOperationAction(ISD::BITCAST, MVT::f16, Custom);
setOperationAction(ISD::BITCAST, MVT::bf16, Custom);
}
// Indexed loads and stores are supported.
for (unsigned im = (unsigned)ISD::PRE_INC;
im != (unsigned)ISD::LAST_INDEXED_MODE; ++im) {
setIndexedLoadAction(im, MVT::i8, Legal);
setIndexedLoadAction(im, MVT::i16, Legal);
setIndexedLoadAction(im, MVT::i32, Legal);
setIndexedLoadAction(im, MVT::i64, Legal);
setIndexedLoadAction(im, MVT::f64, Legal);
setIndexedLoadAction(im, MVT::f32, Legal);
setIndexedLoadAction(im, MVT::f16, Legal);
setIndexedLoadAction(im, MVT::bf16, Legal);
setIndexedStoreAction(im, MVT::i8, Legal);
setIndexedStoreAction(im, MVT::i16, Legal);
setIndexedStoreAction(im, MVT::i32, Legal);
setIndexedStoreAction(im, MVT::i64, Legal);
setIndexedStoreAction(im, MVT::f64, Legal);
setIndexedStoreAction(im, MVT::f32, Legal);
setIndexedStoreAction(im, MVT::f16, Legal);
setIndexedStoreAction(im, MVT::bf16, Legal);
}
// Trap.
setOperationAction(ISD::TRAP, MVT::Other, Legal);
setOperationAction(ISD::DEBUGTRAP, MVT::Other, Legal);
setOperationAction(ISD::UBSANTRAP, MVT::Other, Legal);
// We combine OR nodes for bitfield operations.
setTargetDAGCombine(ISD::OR);
// Try to create BICs for vector ANDs.
setTargetDAGCombine(ISD::AND);
// llvm.init.trampoline and llvm.adjust.trampoline
setOperationAction(ISD::INIT_TRAMPOLINE, MVT::Other, Custom);
setOperationAction(ISD::ADJUST_TRAMPOLINE, MVT::Other, Custom);
// Vector add and sub nodes may conceal a high-half opportunity.
// Also, try to fold ADD into CSINC/CSINV..
setTargetDAGCombine({ISD::ADD, ISD::ABS, ISD::SUB, ISD::XOR, ISD::SINT_TO_FP,
ISD::UINT_TO_FP});
setTargetDAGCombine({ISD::FP_TO_SINT, ISD::FP_TO_UINT, ISD::FP_TO_SINT_SAT,
ISD::FP_TO_UINT_SAT, ISD::FADD});
// Try and combine setcc with csel
setTargetDAGCombine(ISD::SETCC);
setTargetDAGCombine(ISD::INTRINSIC_WO_CHAIN);
setTargetDAGCombine({ISD::ANY_EXTEND, ISD::ZERO_EXTEND, ISD::SIGN_EXTEND,
ISD::SIGN_EXTEND_INREG, ISD::CONCAT_VECTORS,
ISD::EXTRACT_SUBVECTOR, ISD::INSERT_SUBVECTOR,
ISD::STORE, ISD::BUILD_VECTOR});
setTargetDAGCombine(ISD::TRUNCATE);
setTargetDAGCombine(ISD::LOAD);
setTargetDAGCombine(ISD::MSTORE);
setTargetDAGCombine(ISD::MUL);
setTargetDAGCombine({ISD::SELECT, ISD::VSELECT});
setTargetDAGCombine({ISD::INTRINSIC_VOID, ISD::INTRINSIC_W_CHAIN,
ISD::INSERT_VECTOR_ELT, ISD::EXTRACT_VECTOR_ELT,
ISD::VECREDUCE_ADD, ISD::STEP_VECTOR});
setTargetDAGCombine(
{ISD::MGATHER, ISD::MSCATTER, ISD::EXPERIMENTAL_VECTOR_HISTOGRAM});
setTargetDAGCombine(ISD::FP_EXTEND);
setTargetDAGCombine(ISD::GlobalAddress);
setTargetDAGCombine(ISD::CTLZ);
setTargetDAGCombine(ISD::GET_ACTIVE_LANE_MASK);
setTargetDAGCombine(ISD::VECREDUCE_AND);
setTargetDAGCombine(ISD::VECREDUCE_OR);
setTargetDAGCombine(ISD::VECREDUCE_XOR);
setTargetDAGCombine(ISD::SCALAR_TO_VECTOR);
setTargetDAGCombine(ISD::SHL);
// In case of strict alignment, avoid an excessive number of byte wide stores.
MaxStoresPerMemsetOptSize = 8;
MaxStoresPerMemset =
Subtarget->requiresStrictAlign() ? MaxStoresPerMemsetOptSize : 32;
MaxGluedStoresPerMemcpy = 4;
MaxStoresPerMemcpyOptSize = 4;
MaxStoresPerMemcpy =
Subtarget->requiresStrictAlign() ? MaxStoresPerMemcpyOptSize : 16;
MaxStoresPerMemmoveOptSize = 4;
MaxStoresPerMemmove =
Subtarget->requiresStrictAlign() ? MaxStoresPerMemmoveOptSize : 16;
MaxLoadsPerMemcmpOptSize = 4;
MaxLoadsPerMemcmp =
Subtarget->requiresStrictAlign() ? MaxLoadsPerMemcmpOptSize : 8;
setStackPointerRegisterToSaveRestore(AArch64::SP);
setSchedulingPreference(Sched::Hybrid);
EnableExtLdPromotion = true;
// Set required alignment.
setMinFunctionAlignment(Align(4));
// Set preferred alignments.
// Don't align loops on Windows. The SEH unwind info generation needs to
// know the exact length of functions before the alignments have been
// expanded.
if (!Subtarget->isTargetWindows())
setPrefLoopAlignment(STI.getPrefLoopAlignment());
setMaxBytesForAlignment(STI.getMaxBytesForLoopAlignment());
setPrefFunctionAlignment(STI.getPrefFunctionAlignment());
// Only change the limit for entries in a jump table if specified by
// the sub target, but not at the command line.
unsigned MaxJT = STI.getMaximumJumpTableSize();
if (MaxJT && getMaximumJumpTableSize() == UINT_MAX)
setMaximumJumpTableSize(MaxJT);
setHasExtractBitsInsn(true);
setMaxDivRemBitWidthSupported(128);
setOperationAction(ISD::INTRINSIC_WO_CHAIN, MVT::Other, Custom);
if (Subtarget->hasSME())
setOperationAction(ISD::INTRINSIC_WO_CHAIN, MVT::i1, Custom);
if (Subtarget->isNeonAvailable()) {
// FIXME: v1f64 shouldn't be legal if we can avoid it, because it leads to
// silliness like this:
// clang-format off
for (auto Op :
{ISD::SELECT, ISD::SELECT_CC, ISD::FATAN2,
ISD::BR_CC, ISD::FADD, ISD::FSUB,
ISD::FMUL, ISD::FDIV, ISD::FMA,
ISD::FNEG, ISD::FABS, ISD::FCEIL,
ISD::FSQRT, ISD::FFLOOR, ISD::FNEARBYINT,
ISD::FSIN, ISD::FCOS, ISD::FTAN,
ISD::FASIN, ISD::FACOS, ISD::FATAN,
ISD::FSINH, ISD::FCOSH, ISD::FTANH,
ISD::FPOW, ISD::FLOG, ISD::FLOG2,
ISD::FLOG10, ISD::FEXP, ISD::FEXP2,
ISD::FEXP10, ISD::FRINT, ISD::FROUND,
ISD::FROUNDEVEN, ISD::FTRUNC, ISD::FMINNUM,
ISD::FMAXNUM, ISD::FMINIMUM, ISD::FMAXIMUM,
ISD::FMAXNUM_IEEE, ISD::FMINNUM_IEEE,
ISD::STRICT_FADD, ISD::STRICT_FSUB, ISD::STRICT_FMUL,
ISD::STRICT_FDIV, ISD::STRICT_FMA, ISD::STRICT_FCEIL,
ISD::STRICT_FFLOOR, ISD::STRICT_FSQRT, ISD::STRICT_FRINT,
ISD::STRICT_FNEARBYINT, ISD::STRICT_FROUND, ISD::STRICT_FTRUNC,
ISD::STRICT_FROUNDEVEN, ISD::STRICT_FMINNUM, ISD::STRICT_FMAXNUM,
ISD::STRICT_FMINIMUM, ISD::STRICT_FMAXIMUM})
setOperationAction(Op, MVT::v1f64, Expand);
// clang-format on
for (auto Op :
{ISD::FP_TO_SINT, ISD::FP_TO_UINT, ISD::SINT_TO_FP, ISD::UINT_TO_FP,
ISD::FP_ROUND, ISD::FP_TO_SINT_SAT, ISD::FP_TO_UINT_SAT, ISD::MUL,
ISD::STRICT_FP_TO_SINT, ISD::STRICT_FP_TO_UINT,
ISD::STRICT_SINT_TO_FP, ISD::STRICT_UINT_TO_FP, ISD::STRICT_FP_ROUND})
setOperationAction(Op, MVT::v1i64, Expand);
// AArch64 doesn't have a direct vector ->f32 conversion instructions for
// elements smaller than i32, so promote the input to i32 first.
setOperationPromotedToType(ISD::UINT_TO_FP, MVT::v4i8, MVT::v4i32);
setOperationPromotedToType(ISD::SINT_TO_FP, MVT::v4i8, MVT::v4i32);
// Similarly, there is no direct i32 -> f64 vector conversion instruction.
// Or, direct i32 -> f16 vector conversion. Set it so custom, so the
// conversion happens in two steps: v4i32 -> v4f32 -> v4f16
for (auto Op : {ISD::SINT_TO_FP, ISD::UINT_TO_FP, ISD::STRICT_SINT_TO_FP,
ISD::STRICT_UINT_TO_FP})
for (auto VT : {MVT::v2i32, MVT::v2i64, MVT::v4i32})
setOperationAction(Op, VT, Custom);
if (Subtarget->hasFullFP16()) {
setOperationAction(ISD::ConstantFP, MVT::f16, Legal);
setOperationAction(ISD::ConstantFP, MVT::bf16, Legal);
setOperationAction(ISD::SINT_TO_FP, MVT::v8i8, Custom);
setOperationAction(ISD::UINT_TO_FP, MVT::v8i8, Custom);
setOperationAction(ISD::SINT_TO_FP, MVT::v16i8, Custom);
setOperationAction(ISD::UINT_TO_FP, MVT::v16i8, Custom);
setOperationAction(ISD::SINT_TO_FP, MVT::v4i16, Custom);
setOperationAction(ISD::UINT_TO_FP, MVT::v4i16, Custom);
setOperationAction(ISD::SINT_TO_FP, MVT::v8i16, Custom);
setOperationAction(ISD::UINT_TO_FP, MVT::v8i16, Custom);
} else {
// when AArch64 doesn't have fullfp16 support, promote the input
// to i32 first.
setOperationPromotedToType(ISD::SINT_TO_FP, MVT::v8i8, MVT::v8i32);
setOperationPromotedToType(ISD::UINT_TO_FP, MVT::v8i8, MVT::v8i32);
setOperationPromotedToType(ISD::UINT_TO_FP, MVT::v16i8, MVT::v16i32);
setOperationPromotedToType(ISD::SINT_TO_FP, MVT::v16i8, MVT::v16i32);
setOperationPromotedToType(ISD::UINT_TO_FP, MVT::v4i16, MVT::v4i32);
setOperationPromotedToType(ISD::SINT_TO_FP, MVT::v4i16, MVT::v4i32);
setOperationPromotedToType(ISD::SINT_TO_FP, MVT::v8i16, MVT::v8i32);
setOperationPromotedToType(ISD::UINT_TO_FP, MVT::v8i16, MVT::v8i32);
}
setOperationAction(ISD::CTLZ, MVT::v1i64, Expand);
setOperationAction(ISD::CTLZ, MVT::v2i64, Expand);
setOperationAction(ISD::BITREVERSE, MVT::v8i8, Legal);
setOperationAction(ISD::BITREVERSE, MVT::v16i8, Legal);
setOperationAction(ISD::BITREVERSE, MVT::v2i32, Custom);
setOperationAction(ISD::BITREVERSE, MVT::v4i32, Custom);
setOperationAction(ISD::BITREVERSE, MVT::v1i64, Custom);
setOperationAction(ISD::BITREVERSE, MVT::v2i64, Custom);
for (auto VT : {MVT::v1i64, MVT::v2i64}) {
setOperationAction(ISD::UMAX, VT, Custom);
setOperationAction(ISD::SMAX, VT, Custom);
setOperationAction(ISD::UMIN, VT, Custom);
setOperationAction(ISD::SMIN, VT, Custom);
}
// Custom handling for some quad-vector types to detect MULL.
setOperationAction(ISD::MUL, MVT::v8i16, Custom);
setOperationAction(ISD::MUL, MVT::v4i32, Custom);
setOperationAction(ISD::MUL, MVT::v2i64, Custom);
setOperationAction(ISD::MUL, MVT::v4i16, Custom);
setOperationAction(ISD::MUL, MVT::v2i32, Custom);
setOperationAction(ISD::MUL, MVT::v1i64, Custom);
// Saturates
for (MVT VT : { MVT::v8i8, MVT::v4i16, MVT::v2i32, MVT::v1i64,
MVT::v16i8, MVT::v8i16, MVT::v4i32, MVT::v2i64 }) {
setOperationAction(ISD::SADDSAT, VT, Legal);
setOperationAction(ISD::UADDSAT, VT, Legal);
setOperationAction(ISD::SSUBSAT, VT, Legal);
setOperationAction(ISD::USUBSAT, VT, Legal);
}
for (MVT VT : {MVT::v8i8, MVT::v4i16, MVT::v2i32, MVT::v16i8, MVT::v8i16,
MVT::v4i32}) {
setOperationAction(ISD::AVGFLOORS, VT, Legal);
setOperationAction(ISD::AVGFLOORU, VT, Legal);
setOperationAction(ISD::AVGCEILS, VT, Legal);
setOperationAction(ISD::AVGCEILU, VT, Legal);
setOperationAction(ISD::ABDS, VT, Legal);
setOperationAction(ISD::ABDU, VT, Legal);
}
// Vector reductions
for (MVT VT : { MVT::v4f16, MVT::v2f32,
MVT::v8f16, MVT::v4f32, MVT::v2f64 }) {
if (VT.getVectorElementType() != MVT::f16 || Subtarget->hasFullFP16()) {
setOperationAction(ISD::VECREDUCE_FMAX, VT, Legal);
setOperationAction(ISD::VECREDUCE_FMIN, VT, Legal);
setOperationAction(ISD::VECREDUCE_FMAXIMUM, VT, Legal);
setOperationAction(ISD::VECREDUCE_FMINIMUM, VT, Legal);
setOperationAction(ISD::VECREDUCE_FADD, VT, Legal);
}
}
for (MVT VT : { MVT::v8i8, MVT::v4i16, MVT::v2i32,
MVT::v16i8, MVT::v8i16, MVT::v4i32 }) {
setOperationAction(ISD::VECREDUCE_ADD, VT, Custom);
setOperationAction(ISD::VECREDUCE_SMAX, VT, Custom);
setOperationAction(ISD::VECREDUCE_SMIN, VT, Custom);
setOperationAction(ISD::VECREDUCE_UMAX, VT, Custom);
setOperationAction(ISD::VECREDUCE_UMIN, VT, Custom);
setOperationAction(ISD::VECREDUCE_AND, VT, Custom);
setOperationAction(ISD::VECREDUCE_OR, VT, Custom);
setOperationAction(ISD::VECREDUCE_XOR, VT, Custom);
}
setOperationAction(ISD::VECREDUCE_ADD, MVT::v2i64, Custom);
setOperationAction(ISD::VECREDUCE_AND, MVT::v2i64, Custom);
setOperationAction(ISD::VECREDUCE_OR, MVT::v2i64, Custom);
setOperationAction(ISD::VECREDUCE_XOR, MVT::v2i64, Custom);
setOperationAction(ISD::ANY_EXTEND, MVT::v4i32, Legal);
setTruncStoreAction(MVT::v2i32, MVT::v2i16, Expand);
// Likewise, narrowing and extending vector loads/stores aren't handled
// directly.
for (MVT VT : MVT::fixedlen_vector_valuetypes()) {
setOperationAction(ISD::SIGN_EXTEND_INREG, VT, Expand);
if (VT == MVT::v16i8 || VT == MVT::v8i16 || VT == MVT::v4i32) {
setOperationAction(ISD::MULHS, VT, Legal);
setOperationAction(ISD::MULHU, VT, Legal);
} else {
setOperationAction(ISD::MULHS, VT, Expand);
setOperationAction(ISD::MULHU, VT, Expand);
}
setOperationAction(ISD::SMUL_LOHI, VT, Expand);
setOperationAction(ISD::UMUL_LOHI, VT, Expand);
setOperationAction(ISD::BSWAP, VT, Expand);
setOperationAction(ISD::CTTZ, VT, Expand);
for (MVT InnerVT : MVT::fixedlen_vector_valuetypes()) {
setTruncStoreAction(VT, InnerVT, Expand);
setLoadExtAction(ISD::SEXTLOAD, VT, InnerVT, Expand);
setLoadExtAction(ISD::ZEXTLOAD, VT, InnerVT, Expand);
setLoadExtAction(ISD::EXTLOAD, VT, InnerVT, Expand);
}
}
for (auto Op :
{ISD::FFLOOR, ISD::FNEARBYINT, ISD::FCEIL, ISD::FRINT, ISD::FTRUNC,
ISD::FROUND, ISD::FROUNDEVEN, ISD::FMAXNUM_IEEE, ISD::FMINNUM_IEEE,
ISD::STRICT_FFLOOR, ISD::STRICT_FNEARBYINT, ISD::STRICT_FCEIL,
ISD::STRICT_FRINT, ISD::STRICT_FTRUNC, ISD::STRICT_FROUND,
ISD::STRICT_FROUNDEVEN}) {
for (MVT Ty : {MVT::v2f32, MVT::v4f32, MVT::v2f64})
setOperationAction(Op, Ty, Legal);
if (Subtarget->hasFullFP16())
for (MVT Ty : {MVT::v4f16, MVT::v8f16})
setOperationAction(Op, Ty, Legal);
}
// LRINT and LLRINT.
for (auto Op : {ISD::LRINT, ISD::LLRINT}) {
for (MVT Ty : {MVT::v2f32, MVT::v4f32, MVT::v2f64})
setOperationAction(Op, Ty, Custom);
if (Subtarget->hasFullFP16())
for (MVT Ty : {MVT::v4f16, MVT::v8f16})
setOperationAction(Op, Ty, Custom);
}
setTruncStoreAction(MVT::v4i16, MVT::v4i8, Custom);
setOperationAction(ISD::BITCAST, MVT::i2, Custom);
setOperationAction(ISD::BITCAST, MVT::i4, Custom);
setOperationAction(ISD::BITCAST, MVT::i8, Custom);
setOperationAction(ISD::BITCAST, MVT::i16, Custom);
setOperationAction(ISD::BITCAST, MVT::v2i8, Custom);
setOperationAction(ISD::BITCAST, MVT::v2i16, Custom);
setOperationAction(ISD::BITCAST, MVT::v4i8, Custom);
setLoadExtAction(ISD::EXTLOAD, MVT::v4i16, MVT::v4i8, Custom);
setLoadExtAction(ISD::SEXTLOAD, MVT::v4i16, MVT::v4i8, Custom);
setLoadExtAction(ISD::ZEXTLOAD, MVT::v4i16, MVT::v4i8, Custom);
setLoadExtAction(ISD::EXTLOAD, MVT::v4i32, MVT::v4i8, Custom);
setLoadExtAction(ISD::SEXTLOAD, MVT::v4i32, MVT::v4i8, Custom);
setLoadExtAction(ISD::ZEXTLOAD, MVT::v4i32, MVT::v4i8, Custom);
// ADDP custom lowering
for (MVT VT : { MVT::v32i8, MVT::v16i16, MVT::v8i32, MVT::v4i64 })
setOperationAction(ISD::ADD, VT, Custom);
// FADDP custom lowering
for (MVT VT : { MVT::v16f16, MVT::v8f32, MVT::v4f64 })
setOperationAction(ISD::FADD, VT, Custom);
if (Subtarget->hasDotProd()) {
static const unsigned MLAOps[] = {ISD::PARTIAL_REDUCE_SMLA,
ISD::PARTIAL_REDUCE_UMLA};
setPartialReduceMLAAction(MLAOps, MVT::v4i32, MVT::v16i8, Legal);
setPartialReduceMLAAction(MLAOps, MVT::v2i32, MVT::v8i8, Legal);
setPartialReduceMLAAction(MLAOps, MVT::v2i64, MVT::v16i8, Custom);
if (Subtarget->hasMatMulInt8()) {
setPartialReduceMLAAction(ISD::PARTIAL_REDUCE_SUMLA, MVT::v4i32,
MVT::v16i8, Legal);
setPartialReduceMLAAction(ISD::PARTIAL_REDUCE_SUMLA, MVT::v2i64,
MVT::v16i8, Custom);
setPartialReduceMLAAction(ISD::PARTIAL_REDUCE_SUMLA, MVT::v2i32,
MVT::v8i8, Legal);
}
}
} else /* !isNeonAvailable */ {
for (MVT VT : MVT::fixedlen_vector_valuetypes()) {
for (unsigned Op = 0; Op < ISD::BUILTIN_OP_END; ++Op)
setOperationAction(Op, VT, Expand);
if (VT.is128BitVector() || VT.is64BitVector()) {
setOperationAction(ISD::LOAD, VT, Legal);
setOperationAction(ISD::STORE, VT, Legal);
setOperationAction(ISD::BITCAST, VT,
Subtarget->isLittleEndian() ? Legal : Expand);
}
for (MVT InnerVT : MVT::fixedlen_vector_valuetypes()) {
setTruncStoreAction(VT, InnerVT, Expand);
setLoadExtAction(ISD::SEXTLOAD, VT, InnerVT, Expand);
setLoadExtAction(ISD::ZEXTLOAD, VT, InnerVT, Expand);
setLoadExtAction(ISD::EXTLOAD, VT, InnerVT, Expand);
}
}
}
for (MVT VT : {MVT::v8i16, MVT::v4i32, MVT::v2i64}) {
setOperationAction(ISD::TRUNCATE_SSAT_S, VT, Legal);
setOperationAction(ISD::TRUNCATE_SSAT_U, VT, Legal);
setOperationAction(ISD::TRUNCATE_USAT_U, VT, Legal);
}
if (Subtarget->hasSME()) {
setOperationAction(ISD::INTRINSIC_W_CHAIN, MVT::Other, Custom);
}
// FIXME: Move lowering for more nodes here if those are common between
// SVE and SME.
if (Subtarget->isSVEorStreamingSVEAvailable()) {
for (auto VT :
{MVT::nxv16i1, MVT::nxv8i1, MVT::nxv4i1, MVT::nxv2i1, MVT::nxv1i1}) {
setOperationAction(ISD::SPLAT_VECTOR, VT, Custom);
setOperationAction(ISD::EXTRACT_VECTOR_ELT, VT, Custom);
setOperationAction(ISD::VECTOR_DEINTERLEAVE, VT, Custom);
setOperationAction(ISD::VECTOR_INTERLEAVE, VT, Custom);
}
for (auto VT : {MVT::nxv16i1, MVT::nxv8i1, MVT::nxv4i1, MVT::nxv2i1}) {
setOperationAction(ISD::VECTOR_FIND_LAST_ACTIVE, VT, Legal);
setOperationAction(ISD::GET_ACTIVE_LANE_MASK, VT, Legal);
}
if (Subtarget->hasSVE2p1() ||
(Subtarget->hasSME2() && Subtarget->isStreaming()))
setOperationAction(ISD::GET_ACTIVE_LANE_MASK, MVT::nxv32i1, Custom);
for (auto VT : {MVT::v16i8, MVT::v8i8, MVT::v4i16, MVT::v2i32})
setOperationAction(ISD::GET_ACTIVE_LANE_MASK, VT, Custom);
}
if (Subtarget->isSVEorStreamingSVEAvailable()) {
for (auto VT : {MVT::nxv16i8, MVT::nxv8i16, MVT::nxv4i32, MVT::nxv2i64}) {
setOperationAction(ISD::BITREVERSE, VT, Custom);
setOperationAction(ISD::BSWAP, VT, Custom);
setOperationAction(ISD::CTLZ, VT, Custom);
setOperationAction(ISD::CTPOP, VT, Custom);
setOperationAction(ISD::CTTZ, VT, Custom);
setOperationAction(ISD::INSERT_SUBVECTOR, VT, Custom);
setOperationAction(ISD::UINT_TO_FP, VT, Custom);
setOperationAction(ISD::SINT_TO_FP, VT, Custom);
setOperationAction(ISD::FP_TO_UINT, VT, Custom);
setOperationAction(ISD::FP_TO_SINT, VT, Custom);
setOperationAction(ISD::MLOAD, VT, Custom);
setOperationAction(ISD::MUL, VT, Custom);
setOperationAction(ISD::MULHS, VT, Custom);
setOperationAction(ISD::MULHU, VT, Custom);
setOperationAction(ISD::SPLAT_VECTOR, VT, Legal);
setOperationAction(ISD::VECTOR_SPLICE, VT, Custom);
setOperationAction(ISD::SELECT, VT, Custom);
setOperationAction(ISD::SETCC, VT, Custom);
setOperationAction(ISD::SDIV, VT, Custom);
setOperationAction(ISD::UDIV, VT, Custom);
setOperationAction(ISD::SMIN, VT, Custom);
setOperationAction(ISD::UMIN, VT, Custom);
setOperationAction(ISD::SMAX, VT, Custom);
setOperationAction(ISD::UMAX, VT, Custom);
setOperationAction(ISD::SHL, VT, Custom);
setOperationAction(ISD::SRL, VT, Custom);
setOperationAction(ISD::SRA, VT, Custom);
setOperationAction(ISD::ABS, VT, Custom);
setOperationAction(ISD::ABDS, VT, Custom);
setOperationAction(ISD::ABDU, VT, Custom);
setOperationAction(ISD::VECREDUCE_ADD, VT, Custom);
setOperationAction(ISD::VECREDUCE_AND, VT, Custom);
setOperationAction(ISD::VECREDUCE_OR, VT, Custom);
setOperationAction(ISD::VECREDUCE_XOR, VT, Custom);
setOperationAction(ISD::VECREDUCE_UMIN, VT, Custom);
setOperationAction(ISD::VECREDUCE_UMAX, VT, Custom);
setOperationAction(ISD::VECREDUCE_SMIN, VT, Custom);
setOperationAction(ISD::VECREDUCE_SMAX, VT, Custom);
setOperationAction(ISD::VECTOR_DEINTERLEAVE, VT, Custom);
setOperationAction(ISD::VECTOR_INTERLEAVE, VT, Custom);
setOperationAction(ISD::UMUL_LOHI, VT, Expand);
setOperationAction(ISD::SMUL_LOHI, VT, Expand);
setOperationAction(ISD::SELECT_CC, VT, Expand);
setOperationAction(ISD::ROTL, VT, Expand);
setOperationAction(ISD::ROTR, VT, Expand);
setOperationAction(ISD::SADDSAT, VT, Legal);
setOperationAction(ISD::UADDSAT, VT, Legal);
setOperationAction(ISD::SSUBSAT, VT, Legal);
setOperationAction(ISD::USUBSAT, VT, Legal);
setOperationAction(ISD::UREM, VT, Expand);
setOperationAction(ISD::SREM, VT, Expand);
setOperationAction(ISD::SDIVREM, VT, Expand);
setOperationAction(ISD::UDIVREM, VT, Expand);
setOperationAction(ISD::AVGFLOORS, VT, Custom);
setOperationAction(ISD::AVGFLOORU, VT, Custom);
setOperationAction(ISD::AVGCEILS, VT, Custom);
setOperationAction(ISD::AVGCEILU, VT, Custom);
if (!Subtarget->isLittleEndian())
setOperationAction(ISD::BITCAST, VT, Custom);
if (Subtarget->hasSVE2() ||
(Subtarget->hasSME() && Subtarget->isStreaming()))
// For SLI/SRI.
setOperationAction(ISD::OR, VT, Custom);
}
// Illegal unpacked integer vector types.
for (auto VT : {MVT::nxv8i8, MVT::nxv4i16, MVT::nxv2i32}) {
setOperationAction(ISD::EXTRACT_SUBVECTOR, VT, Custom);
setOperationAction(ISD::INSERT_SUBVECTOR, VT, Custom);
}
// Type legalize unpacked bitcasts.
for (auto VT : {MVT::nxv2i16, MVT::nxv4i16, MVT::nxv2i32})
setOperationAction(ISD::BITCAST, VT, Custom);
for (auto VT :
{ MVT::nxv2i8, MVT::nxv2i16, MVT::nxv2i32, MVT::nxv2i64, MVT::nxv4i8,
MVT::nxv4i16, MVT::nxv4i32, MVT::nxv8i8, MVT::nxv8i16 })
setOperationAction(ISD::SIGN_EXTEND_INREG, VT, Legal);
for (auto VT :
{MVT::nxv16i1, MVT::nxv8i1, MVT::nxv4i1, MVT::nxv2i1, MVT::nxv1i1}) {
setOperationAction(ISD::CONCAT_VECTORS, VT, Custom);
setOperationAction(ISD::SELECT, VT, Custom);
setOperationAction(ISD::SETCC, VT, Custom);
setOperationAction(ISD::TRUNCATE, VT, Custom);
setOperationAction(ISD::VECREDUCE_AND, VT, Custom);
setOperationAction(ISD::VECREDUCE_OR, VT, Custom);
setOperationAction(ISD::VECREDUCE_XOR, VT, Custom);
setOperationAction(ISD::SELECT_CC, VT, Expand);
setOperationAction(ISD::INSERT_VECTOR_ELT, VT, Custom);
setOperationAction(ISD::INSERT_SUBVECTOR, VT, Custom);
// There are no legal MVT::nxv16f## based types.
if (VT != MVT::nxv16i1) {
setOperationAction(ISD::FP_TO_SINT, VT, Custom);
setOperationAction(ISD::FP_TO_UINT, VT, Custom);
setOperationAction(ISD::SINT_TO_FP, VT, Custom);
setOperationAction(ISD::UINT_TO_FP, VT, Custom);
}
}
// NEON doesn't support masked loads/stores, but SME and SVE do.
for (auto VT :
{MVT::v4f16, MVT::v8f16, MVT::v2f32, MVT::v4f32, MVT::v1f64,
MVT::v2f64, MVT::v8i8, MVT::v16i8, MVT::v4i16, MVT::v8i16,
MVT::v2i32, MVT::v4i32, MVT::v1i64, MVT::v2i64}) {
setOperationAction(ISD::MLOAD, VT, Custom);
setOperationAction(ISD::MSTORE, VT, Custom);
}
// Firstly, exclude all scalable vector extending loads/truncating stores,
// include both integer and floating scalable vector.
for (MVT VT : MVT::scalable_vector_valuetypes()) {
for (MVT InnerVT : MVT::scalable_vector_valuetypes()) {
setTruncStoreAction(VT, InnerVT, Expand);
setLoadExtAction(ISD::SEXTLOAD, VT, InnerVT, Expand);
setLoadExtAction(ISD::ZEXTLOAD, VT, InnerVT, Expand);
setLoadExtAction(ISD::EXTLOAD, VT, InnerVT, Expand);
}
}
// Then, selectively enable those which we directly support.
setTruncStoreAction(MVT::nxv2i64, MVT::nxv2i8, Legal);
setTruncStoreAction(MVT::nxv2i64, MVT::nxv2i16, Legal);
setTruncStoreAction(MVT::nxv2i64, MVT::nxv2i32, Legal);
setTruncStoreAction(MVT::nxv4i32, MVT::nxv4i8, Legal);
setTruncStoreAction(MVT::nxv4i32, MVT::nxv4i16, Legal);
setTruncStoreAction(MVT::nxv8i16, MVT::nxv8i8, Legal);
for (auto Op : {ISD::ZEXTLOAD, ISD::SEXTLOAD, ISD::EXTLOAD}) {
setLoadExtAction(Op, MVT::nxv2i64, MVT::nxv2i8, Legal);
setLoadExtAction(Op, MVT::nxv2i64, MVT::nxv2i16, Legal);
setLoadExtAction(Op, MVT::nxv2i64, MVT::nxv2i32, Legal);
setLoadExtAction(Op, MVT::nxv4i32, MVT::nxv4i8, Legal);
setLoadExtAction(Op, MVT::nxv4i32, MVT::nxv4i16, Legal);
setLoadExtAction(Op, MVT::nxv8i16, MVT::nxv8i8, Legal);
}
// SVE supports truncating stores of 64 and 128-bit vectors
setTruncStoreAction(MVT::v2i64, MVT::v2i8, Custom);
setTruncStoreAction(MVT::v2i64, MVT::v2i16, Custom);
setTruncStoreAction(MVT::v2i64, MVT::v2i32, Custom);
setTruncStoreAction(MVT::v2i32, MVT::v2i8, Custom);
setTruncStoreAction(MVT::v2i32, MVT::v2i16, Custom);
for (auto VT : {MVT::nxv2f16, MVT::nxv4f16, MVT::nxv8f16, MVT::nxv2f32,
MVT::nxv4f32, MVT::nxv2f64}) {
setOperationAction(ISD::BITCAST, VT, Custom);
setOperationAction(ISD::CONCAT_VECTORS, VT, Custom);
setOperationAction(ISD::INSERT_SUBVECTOR, VT, Custom);
setOperationAction(ISD::MLOAD, VT, Custom);
setOperationAction(ISD::SPLAT_VECTOR, VT, Legal);
setOperationAction(ISD::SELECT, VT, Custom);
setOperationAction(ISD::SETCC, VT, Custom);
setOperationAction(ISD::FADD, VT, Custom);
setOperationAction(ISD::FCOPYSIGN, VT, Custom);
setOperationAction(ISD::FDIV, VT, Custom);
setOperationAction(ISD::FMA, VT, Custom);
setOperationAction(ISD::FMAXIMUM, VT, Custom);
setOperationAction(ISD::FMAXNUM, VT, Custom);
setOperationAction(ISD::FMINIMUM, VT, Custom);
setOperationAction(ISD::FMINNUM, VT, Custom);
setOperationAction(ISD::FMUL, VT, Custom);
setOperationAction(ISD::FNEG, VT, Custom);
setOperationAction(ISD::FSUB, VT, Custom);
setOperationAction(ISD::FCEIL, VT, Custom);
setOperationAction(ISD::FFLOOR, VT, Custom);
setOperationAction(ISD::FNEARBYINT, VT, Custom);
setOperationAction(ISD::FRINT, VT, Custom);
setOperationAction(ISD::LRINT, VT, Custom);
setOperationAction(ISD::LLRINT, VT, Custom);
setOperationAction(ISD::FROUND, VT, Custom);
setOperationAction(ISD::FROUNDEVEN, VT, Custom);
setOperationAction(ISD::FTRUNC, VT, Custom);
setOperationAction(ISD::FSQRT, VT, Custom);
setOperationAction(ISD::FABS, VT, Custom);
setOperationAction(ISD::FP_EXTEND, VT, Custom);
setOperationAction(ISD::FP_ROUND, VT, Custom);
setOperationAction(ISD::VECREDUCE_FADD, VT, Custom);
setOperationAction(ISD::VECREDUCE_FMAX, VT, Custom);
setOperationAction(ISD::VECREDUCE_FMIN, VT, Custom);
setOperationAction(ISD::VECREDUCE_FMAXIMUM, VT, Custom);
setOperationAction(ISD::VECREDUCE_FMINIMUM, VT, Custom);
setOperationAction(ISD::VECTOR_SPLICE, VT, Custom);
setOperationAction(ISD::VECTOR_DEINTERLEAVE, VT, Custom);
setOperationAction(ISD::VECTOR_INTERLEAVE, VT, Custom);
setOperationAction(ISD::SELECT_CC, VT, Expand);
setOperationAction(ISD::FREM, VT, Expand);
setOperationAction(ISD::FPOW, VT, Expand);
setOperationAction(ISD::FPOWI, VT, Expand);
setOperationAction(ISD::FCOS, VT, Expand);
setOperationAction(ISD::FSIN, VT, Expand);
setOperationAction(ISD::FSINCOS, VT, Expand);
setOperationAction(ISD::FTAN, VT, Expand);
setOperationAction(ISD::FACOS, VT, Expand);
setOperationAction(ISD::FASIN, VT, Expand);
setOperationAction(ISD::FATAN, VT, Expand);
setOperationAction(ISD::FATAN2, VT, Expand);
setOperationAction(ISD::FCOSH, VT, Expand);
setOperationAction(ISD::FSINH, VT, Expand);
setOperationAction(ISD::FTANH, VT, Expand);
setOperationAction(ISD::FEXP, VT, Expand);
setOperationAction(ISD::FEXP2, VT, Expand);
setOperationAction(ISD::FEXP10, VT, Expand);
setOperationAction(ISD::FLOG, VT, Expand);
setOperationAction(ISD::FLOG2, VT, Expand);
setOperationAction(ISD::FLOG10, VT, Expand);
setCondCodeAction(ISD::SETO, VT, Expand);
setCondCodeAction(ISD::SETOLT, VT, Expand);
setCondCodeAction(ISD::SETLT, VT, Expand);
setCondCodeAction(ISD::SETOLE, VT, Expand);
setCondCodeAction(ISD::SETLE, VT, Expand);
setCondCodeAction(ISD::SETULT, VT, Expand);
setCondCodeAction(ISD::SETULE, VT, Expand);
setCondCodeAction(ISD::SETUGE, VT, Expand);
setCondCodeAction(ISD::SETUGT, VT, Expand);
setCondCodeAction(ISD::SETUEQ, VT, Expand);
setCondCodeAction(ISD::SETONE, VT, Expand);
}
for (auto VT : {MVT::nxv2bf16, MVT::nxv4bf16, MVT::nxv8bf16}) {
setOperationAction(ISD::BITCAST, VT, Custom);
setOperationAction(ISD::CONCAT_VECTORS, VT, Custom);
setOperationAction(ISD::FABS, VT, Legal);
setOperationAction(ISD::FCOPYSIGN, VT, Custom);
setOperationAction(ISD::FNEG, VT, Legal);
setOperationAction(ISD::FP_EXTEND, VT, Custom);
setOperationAction(ISD::FP_ROUND, VT, Custom);
setOperationAction(ISD::MLOAD, VT, Custom);
setOperationAction(ISD::INSERT_SUBVECTOR, VT, Custom);
setOperationAction(ISD::SELECT, VT, Custom);
setOperationAction(ISD::SELECT_CC, VT, Expand);
setOperationAction(ISD::SPLAT_VECTOR, VT, Legal);
setOperationAction(ISD::VECTOR_DEINTERLEAVE, VT, Custom);
setOperationAction(ISD::VECTOR_INTERLEAVE, VT, Custom);
setOperationAction(ISD::VECTOR_SPLICE, VT, Custom);
if (Subtarget->hasSVEB16B16()) {
setOperationAction(ISD::FADD, VT, Legal);
setOperationAction(ISD::FMA, VT, Custom);
setOperationAction(ISD::FMAXIMUM, VT, Custom);
setOperationAction(ISD::FMAXNUM, VT, Custom);
setOperationAction(ISD::FMINIMUM, VT, Custom);
setOperationAction(ISD::FMINNUM, VT, Custom);
setOperationAction(ISD::FMUL, VT, Legal);
setOperationAction(ISD::FSUB, VT, Legal);
}
}
for (auto Opcode :
{ISD::FCEIL, ISD::FDIV, ISD::FFLOOR, ISD::FNEARBYINT, ISD::FRINT,
ISD::FROUND, ISD::FROUNDEVEN, ISD::FSQRT, ISD::FTRUNC, ISD::SETCC,
ISD::VECREDUCE_FADD, ISD::VECREDUCE_FMAX, ISD::VECREDUCE_FMAXIMUM,
ISD::VECREDUCE_FMIN, ISD::VECREDUCE_FMINIMUM}) {
setOperationPromotedToType(Opcode, MVT::nxv2bf16, MVT::nxv2f32);
setOperationPromotedToType(Opcode, MVT::nxv4bf16, MVT::nxv4f32);
setOperationPromotedToType(Opcode, MVT::nxv8bf16, MVT::nxv8f32);
}
if (!Subtarget->hasSVEB16B16()) {
for (auto Opcode : {ISD::FADD, ISD::FMA, ISD::FMAXIMUM, ISD::FMAXNUM,
ISD::FMINIMUM, ISD::FMINNUM, ISD::FMUL, ISD::FSUB}) {
setOperationPromotedToType(Opcode, MVT::nxv2bf16, MVT::nxv2f32);
setOperationPromotedToType(Opcode, MVT::nxv4bf16, MVT::nxv4f32);
setOperationPromotedToType(Opcode, MVT::nxv8bf16, MVT::nxv8f32);
}
}
setOperationAction(ISD::INTRINSIC_WO_CHAIN, MVT::i8, Custom);
setOperationAction(ISD::INTRINSIC_WO_CHAIN, MVT::i16, Custom);
// NEON doesn't support integer divides, but SVE does
for (auto VT : {MVT::v8i8, MVT::v16i8, MVT::v4i16, MVT::v8i16, MVT::v2i32,
MVT::v4i32, MVT::v1i64, MVT::v2i64}) {
setOperationAction(ISD::SDIV, VT, Custom);
setOperationAction(ISD::UDIV, VT, Custom);
}
// NEON doesn't support 64-bit vector integer muls, but SVE does.
setOperationAction(ISD::MUL, MVT::v1i64, Custom);
setOperationAction(ISD::MUL, MVT::v2i64, Custom);
// NOTE: Currently this has to happen after computeRegisterProperties rather
// than the preferred option of combining it with the addRegisterClass call.
if (Subtarget->useSVEForFixedLengthVectors()) {
for (MVT VT : MVT::integer_fixedlen_vector_valuetypes()) {
if (useSVEForFixedLengthVectorVT(
VT, /*OverrideNEON=*/!Subtarget->isNeonAvailable()))
addTypeForFixedLengthSVE(VT);
}
for (MVT VT : MVT::fp_fixedlen_vector_valuetypes()) {
if (useSVEForFixedLengthVectorVT(
VT, /*OverrideNEON=*/!Subtarget->isNeonAvailable()))
addTypeForFixedLengthSVE(VT);
}
// 64bit results can mean a bigger than NEON input.
for (auto VT : {MVT::v8i8, MVT::v4i16})
setOperationAction(ISD::TRUNCATE, VT, Custom);
setOperationAction(ISD::FP_ROUND, MVT::v4f16, Custom);
// 128bit results imply a bigger than NEON input.
for (auto VT : {MVT::v16i8, MVT::v8i16, MVT::v4i32})
setOperationAction(ISD::TRUNCATE, VT, Custom);
for (auto VT : {MVT::v8f16, MVT::v4f32})
setOperationAction(ISD::FP_ROUND, VT, Custom);
// These operations are not supported on NEON but SVE can do them.
setOperationAction(ISD::BITREVERSE, MVT::v1i64, Custom);
setOperationAction(ISD::CTLZ, MVT::v1i64, Custom);
setOperationAction(ISD::CTLZ, MVT::v2i64, Custom);
setOperationAction(ISD::CTTZ, MVT::v1i64, Custom);
setOperationAction(ISD::MULHS, MVT::v1i64, Custom);
setOperationAction(ISD::MULHS, MVT::v2i64, Custom);
setOperationAction(ISD::MULHU, MVT::v1i64, Custom);
setOperationAction(ISD::MULHU, MVT::v2i64, Custom);
setOperationAction(ISD::SMAX, MVT::v1i64, Custom);
setOperationAction(ISD::SMAX, MVT::v2i64, Custom);
setOperationAction(ISD::SMIN, MVT::v1i64, Custom);
setOperationAction(ISD::SMIN, MVT::v2i64, Custom);
setOperationAction(ISD::UMAX, MVT::v1i64, Custom);
setOperationAction(ISD::UMAX, MVT::v2i64, Custom);
setOperationAction(ISD::UMIN, MVT::v1i64, Custom);
setOperationAction(ISD::UMIN, MVT::v2i64, Custom);
setOperationAction(ISD::VECREDUCE_SMAX, MVT::v2i64, Custom);
setOperationAction(ISD::VECREDUCE_SMIN, MVT::v2i64, Custom);
setOperationAction(ISD::VECREDUCE_UMAX, MVT::v2i64, Custom);
setOperationAction(ISD::VECREDUCE_UMIN, MVT::v2i64, Custom);
// Int operations with no NEON support.
for (auto VT : {MVT::v8i8, MVT::v16i8, MVT::v4i16, MVT::v8i16,
MVT::v2i32, MVT::v4i32, MVT::v2i64}) {
setOperationAction(ISD::BITREVERSE, VT, Custom);
setOperationAction(ISD::CTTZ, VT, Custom);
setOperationAction(ISD::VECREDUCE_AND, VT, Custom);
setOperationAction(ISD::VECREDUCE_OR, VT, Custom);
setOperationAction(ISD::VECREDUCE_XOR, VT, Custom);
setOperationAction(ISD::MULHS, VT, Custom);
setOperationAction(ISD::MULHU, VT, Custom);
}
// Use SVE for vectors with more than 2 elements.
for (auto VT : {MVT::v4f16, MVT::v8f16, MVT::v4f32})
setOperationAction(ISD::VECREDUCE_FADD, VT, Custom);
}
setOperationPromotedToType(ISD::VECTOR_SPLICE, MVT::nxv2i1, MVT::nxv2i64);
setOperationPromotedToType(ISD::VECTOR_SPLICE, MVT::nxv4i1, MVT::nxv4i32);
setOperationPromotedToType(ISD::VECTOR_SPLICE, MVT::nxv8i1, MVT::nxv8i16);
setOperationPromotedToType(ISD::VECTOR_SPLICE, MVT::nxv16i1, MVT::nxv16i8);
setOperationAction(ISD::VSCALE, MVT::i32, Custom);
for (auto VT : {MVT::v16i1, MVT::v8i1, MVT::v4i1, MVT::v2i1})
setOperationAction(ISD::INTRINSIC_WO_CHAIN, VT, Custom);
}
// Handle partial reduction operations
if (Subtarget->isSVEorStreamingSVEAvailable()) {
// Mark known legal pairs as 'Legal' (these will expand to UDOT or SDOT).
// Other pairs will default to 'Expand'.
static const unsigned MLAOps[] = {ISD::PARTIAL_REDUCE_SMLA,
ISD::PARTIAL_REDUCE_UMLA};
setPartialReduceMLAAction(MLAOps, MVT::nxv2i64, MVT::nxv8i16, Legal);
setPartialReduceMLAAction(MLAOps, MVT::nxv4i32, MVT::nxv16i8, Legal);
setPartialReduceMLAAction(MLAOps, MVT::nxv2i64, MVT::nxv16i8, Custom);
if (Subtarget->hasMatMulInt8()) {
setPartialReduceMLAAction(ISD::PARTIAL_REDUCE_SUMLA, MVT::nxv4i32,
MVT::nxv16i8, Legal);
setPartialReduceMLAAction(ISD::PARTIAL_REDUCE_SUMLA, MVT::nxv2i64,
MVT::nxv16i8, Custom);
}
// Wide add types
if (Subtarget->hasSVE2() || Subtarget->hasSME()) {
setPartialReduceMLAAction(MLAOps, MVT::nxv2i64, MVT::nxv4i32, Legal);
setPartialReduceMLAAction(MLAOps, MVT::nxv4i32, MVT::nxv8i16, Legal);
setPartialReduceMLAAction(MLAOps, MVT::nxv8i16, MVT::nxv16i8, Legal);
}
}
// Handle operations that are only available in non-streaming SVE mode.
if (Subtarget->isSVEAvailable()) {
for (auto VT : {MVT::nxv16i8, MVT::nxv8i16, MVT::nxv4i32, MVT::nxv2i64,
MVT::nxv2f16, MVT::nxv4f16, MVT::nxv8f16, MVT::nxv2f32,
MVT::nxv4f32, MVT::nxv2f64, MVT::nxv2bf16, MVT::nxv4bf16,
MVT::nxv8bf16, MVT::v4f16, MVT::v8f16, MVT::v2f32,
MVT::v4f32, MVT::v1f64, MVT::v2f64, MVT::v8i8,
MVT::v16i8, MVT::v4i16, MVT::v8i16, MVT::v2i32,
MVT::v4i32, MVT::v1i64, MVT::v2i64}) {
setOperationAction(ISD::MGATHER, VT, Custom);
setOperationAction(ISD::MSCATTER, VT, Custom);
}
for (auto VT : {MVT::nxv2f16, MVT::nxv4f16, MVT::nxv8f16, MVT::nxv2f32,
MVT::nxv4f32, MVT::nxv2f64, MVT::v4f16, MVT::v8f16,
MVT::v2f32, MVT::v4f32, MVT::v2f64})
setOperationAction(ISD::VECREDUCE_SEQ_FADD, VT, Custom);
// We can lower types that have <vscale x {2|4}> elements to compact.
for (auto VT :
{MVT::nxv2i8, MVT::nxv2i16, MVT::nxv2i32, MVT::nxv2i64, MVT::nxv2f32,
MVT::nxv2f64, MVT::nxv4i8, MVT::nxv4i16, MVT::nxv4i32, MVT::nxv4f32})
setOperationAction(ISD::VECTOR_COMPRESS, VT, Custom);
// If we have SVE, we can use SVE logic for legal (or smaller than legal)
// NEON vectors in the lowest bits of the SVE register.
for (auto VT : {MVT::v2i8, MVT::v2i16, MVT::v2i32, MVT::v2i64, MVT::v2f32,
MVT::v2f64, MVT::v4i8, MVT::v4i16, MVT::v4i32, MVT::v4f32})
setOperationAction(ISD::VECTOR_COMPRESS, VT, Custom);
// Histcnt is SVE2 only
if (Subtarget->hasSVE2()) {
setOperationAction(ISD::EXPERIMENTAL_VECTOR_HISTOGRAM, MVT::nxv4i32,
Custom);
setOperationAction(ISD::EXPERIMENTAL_VECTOR_HISTOGRAM, MVT::nxv2i64,
Custom);
static const unsigned MLAOps[] = {ISD::PARTIAL_REDUCE_SMLA,
ISD::PARTIAL_REDUCE_UMLA};
// Must be lowered to SVE instructions.
setPartialReduceMLAAction(MLAOps, MVT::v2i64, MVT::v4i32, Custom);
setPartialReduceMLAAction(MLAOps, MVT::v2i64, MVT::v8i16, Custom);
setPartialReduceMLAAction(MLAOps, MVT::v2i64, MVT::v16i8, Custom);
setPartialReduceMLAAction(MLAOps, MVT::v4i32, MVT::v8i16, Custom);
setPartialReduceMLAAction(MLAOps, MVT::v4i32, MVT::v16i8, Custom);
setPartialReduceMLAAction(MLAOps, MVT::v8i16, MVT::v16i8, Custom);
}
}
if (Subtarget->hasMOPS() && Subtarget->hasMTE()) {
// Only required for llvm.aarch64.mops.memset.tag
setOperationAction(ISD::INTRINSIC_W_CHAIN, MVT::i8, Custom);
}
setOperationAction(ISD::INTRINSIC_VOID, MVT::Other, Custom);
if (Subtarget->hasSVE()) {
setOperationAction(ISD::FLDEXP, MVT::f64, Custom);
setOperationAction(ISD::FLDEXP, MVT::f32, Custom);
setOperationAction(ISD::FLDEXP, MVT::f16, Custom);
setOperationAction(ISD::FLDEXP, MVT::bf16, Custom);
}
PredictableSelectIsExpensive = Subtarget->predictableSelectIsExpensive();
IsStrictFPEnabled = true;
setMaxAtomicSizeInBitsSupported(128);
// On MSVC, both 32-bit and 64-bit, ldexpf(f32) is not defined. MinGW has
// it, but it's just a wrapper around ldexp.
if (Subtarget->isTargetWindows()) {
for (ISD::NodeType Op : {ISD::FLDEXP, ISD::STRICT_FLDEXP, ISD::FFREXP})
if (isOperationExpand(Op, MVT::f32))
setOperationAction(Op, MVT::f32, Promote);
}
// LegalizeDAG currently can't expand fp16 LDEXP/FREXP on targets where i16
// isn't legal.
for (ISD::NodeType Op : {ISD::FLDEXP, ISD::STRICT_FLDEXP, ISD::FFREXP})
if (isOperationExpand(Op, MVT::f16))
setOperationAction(Op, MVT::f16, Promote);
}
void AArch64TargetLowering::addTypeForNEON(MVT VT) {
assert(VT.isVector() && "VT should be a vector type");
if (VT.isFloatingPoint()) {
MVT PromoteTo = EVT(VT).changeVectorElementTypeToInteger().getSimpleVT();
setOperationPromotedToType(ISD::LOAD, VT, PromoteTo);
setOperationPromotedToType(ISD::STORE, VT, PromoteTo);
}
// Mark vector float intrinsics as expand.
if (VT == MVT::v2f32 || VT == MVT::v4f32 || VT == MVT::v2f64) {
setOperationAction(ISD::FSIN, VT, Expand);
setOperationAction(ISD::FCOS, VT, Expand);
setOperationAction(ISD::FTAN, VT, Expand);
setOperationAction(ISD::FASIN, VT, Expand);
setOperationAction(ISD::FACOS, VT, Expand);
setOperationAction(ISD::FATAN, VT, Expand);
setOperationAction(ISD::FATAN2, VT, Expand);
setOperationAction(ISD::FSINH, VT, Expand);
setOperationAction(ISD::FCOSH, VT, Expand);
setOperationAction(ISD::FTANH, VT, Expand);
setOperationAction(ISD::FPOW, VT, Expand);
setOperationAction(ISD::FLOG, VT, Expand);
setOperationAction(ISD::FLOG2, VT, Expand);
setOperationAction(ISD::FLOG10, VT, Expand);
setOperationAction(ISD::FEXP, VT, Expand);
setOperationAction(ISD::FEXP2, VT, Expand);
setOperationAction(ISD::FEXP10, VT, Expand);
}
// But we do support custom-lowering for FCOPYSIGN.
if (VT == MVT::v2f32 || VT == MVT::v4f32 || VT == MVT::v2f64 ||
((VT == MVT::v4bf16 || VT == MVT::v8bf16 || VT == MVT::v4f16 ||
VT == MVT::v8f16) &&
Subtarget->hasFullFP16()))
setOperationAction(ISD::FCOPYSIGN, VT, Custom);
setOperationAction(ISD::EXTRACT_VECTOR_ELT, VT, Custom);
setOperationAction(ISD::INSERT_VECTOR_ELT, VT, Custom);
setOperationAction(ISD::BUILD_VECTOR, VT, Custom);
setOperationAction(ISD::ZERO_EXTEND_VECTOR_INREG, VT, Custom);
setOperationAction(ISD::VECTOR_SHUFFLE, VT, Custom);
setOperationAction(ISD::EXTRACT_SUBVECTOR, VT, Custom);
setOperationAction(ISD::SRA, VT, Custom);
setOperationAction(ISD::SRL, VT, Custom);
setOperationAction(ISD::SHL, VT, Custom);
setOperationAction(ISD::OR, VT, Custom);
setOperationAction(ISD::SETCC, VT, Custom);
setOperationAction(ISD::CONCAT_VECTORS, VT, Legal);
setOperationAction(ISD::SELECT, VT, Expand);
setOperationAction(ISD::SELECT_CC, VT, Expand);
setOperationAction(ISD::VSELECT, VT, Expand);
for (MVT InnerVT : MVT::all_valuetypes())
setLoadExtAction(ISD::EXTLOAD, InnerVT, VT, Expand);
// CNT supports only B element sizes, then use UADDLP to widen.
if (VT != MVT::v8i8 && VT != MVT::v16i8)
setOperationAction(ISD::CTPOP, VT, Custom);
setOperationAction(ISD::UDIV, VT, Expand);
setOperationAction(ISD::SDIV, VT, Expand);
setOperationAction(ISD::UREM, VT, Expand);
setOperationAction(ISD::SREM, VT, Expand);
setOperationAction(ISD::FREM, VT, Expand);
for (unsigned Opcode :
{ISD::FP_TO_SINT, ISD::FP_TO_UINT, ISD::FP_TO_SINT_SAT,
ISD::FP_TO_UINT_SAT, ISD::STRICT_FP_TO_SINT, ISD::STRICT_FP_TO_UINT})
setOperationAction(Opcode, VT, Custom);
if (!VT.isFloatingPoint())
setOperationAction(ISD::ABS, VT, Legal);
// [SU][MIN|MAX] are available for all NEON types apart from i64.
if (!VT.isFloatingPoint() && VT != MVT::v2i64 && VT != MVT::v1i64)
for (unsigned Opcode : {ISD::SMIN, ISD::SMAX, ISD::UMIN, ISD::UMAX})
setOperationAction(Opcode, VT, Legal);
// F[MIN|MAX][NUM|NAN] and simple strict operations are available for all FP
// NEON types.
if (VT.isFloatingPoint() &&
VT.getVectorElementType() != MVT::bf16 &&
(VT.getVectorElementType() != MVT::f16 || Subtarget->hasFullFP16()))
for (unsigned Opcode :
{ISD::FMINIMUM, ISD::FMAXIMUM, ISD::FMINNUM, ISD::FMAXNUM,
ISD::FMINNUM_IEEE, ISD::FMAXNUM_IEEE, ISD::STRICT_FMINIMUM,
ISD::STRICT_FMAXIMUM, ISD::STRICT_FMINNUM, ISD::STRICT_FMAXNUM,
ISD::STRICT_FADD, ISD::STRICT_FSUB, ISD::STRICT_FMUL,
ISD::STRICT_FDIV, ISD::STRICT_FMA, ISD::STRICT_FSQRT})
setOperationAction(Opcode, VT, Legal);
// Strict fp extend and trunc are legal
if (VT.isFloatingPoint() && VT.getScalarSizeInBits() != 16)
setOperationAction(ISD::STRICT_FP_EXTEND, VT, Legal);
if (VT.isFloatingPoint() && VT.getScalarSizeInBits() != 64)
setOperationAction(ISD::STRICT_FP_ROUND, VT, Legal);
// FIXME: We could potentially make use of the vector comparison instructions
// for STRICT_FSETCC and STRICT_FSETCSS, but there's a number of
// complications:
// * FCMPEQ/NE are quiet comparisons, the rest are signalling comparisons,
// so we would need to expand when the condition code doesn't match the
// kind of comparison.
// * Some kinds of comparison require more than one FCMXY instruction so
// would need to be expanded instead.
// * The lowering of the non-strict versions involves target-specific ISD
// nodes so we would likely need to add strict versions of all of them and
// handle them appropriately.
setOperationAction(ISD::STRICT_FSETCC, VT, Expand);
setOperationAction(ISD::STRICT_FSETCCS, VT, Expand);
// When little-endian we can use ordinary d and q register loads/stores for
// vector types, but when big-endian we need to use structure load/store which
// only allow post-index addressing.
if (Subtarget->isLittleEndian()) {
for (unsigned im = (unsigned)ISD::PRE_INC;
im != (unsigned)ISD::LAST_INDEXED_MODE; ++im) {
setIndexedLoadAction(im, VT, Legal);
setIndexedStoreAction(im, VT, Legal);
}
} else {
setIndexedLoadAction(ISD::POST_INC, VT, Legal);
setIndexedStoreAction(ISD::POST_INC, VT, Legal);
}
if (Subtarget->hasD128()) {
setOperationAction(ISD::READ_REGISTER, MVT::i128, Custom);
setOperationAction(ISD::WRITE_REGISTER, MVT::i128, Custom);
}
if (VT.isInteger()) {
// Let common code emit inverted variants of compares we do support.
setCondCodeAction(ISD::SETNE, VT, Expand);
setCondCodeAction(ISD::SETLE, VT, Expand);
setCondCodeAction(ISD::SETLT, VT, Expand);
setCondCodeAction(ISD::SETULE, VT, Expand);
setCondCodeAction(ISD::SETULT, VT, Expand);
}
}
bool AArch64TargetLowering::shouldExpandGetActiveLaneMask(EVT ResVT,
EVT OpVT) const {
// Only SVE has a 1:1 mapping from intrinsic -> instruction (whilelo).
if (!Subtarget->isSVEorStreamingSVEAvailable() ||
ResVT.getVectorElementType() != MVT::i1)
return true;
// Only support illegal types if the result is scalable and min elements > 1.
if (ResVT.getVectorMinNumElements() == 1 ||
(ResVT.isFixedLengthVector() && (ResVT.getVectorNumElements() > 16 ||
(OpVT != MVT::i32 && OpVT != MVT::i64))))
return true;
// 32 & 64 bit operands are supported. We can promote anything < 64 bits,
// but anything larger should be expanded.
if (OpVT.getFixedSizeInBits() > 64)
return true;
return false;
}
bool AArch64TargetLowering::shouldExpandPartialReductionIntrinsic(
const IntrinsicInst *I) const {
assert(I->getIntrinsicID() ==
Intrinsic::experimental_vector_partial_reduce_add &&
"Unexpected intrinsic!");
return true;
}
bool AArch64TargetLowering::shouldExpandCttzElements(EVT VT) const {
if (!Subtarget->isSVEorStreamingSVEAvailable())
return true;
// We can only use the BRKB + CNTP sequence with legal predicate types. We can
// also support fixed-width predicates.
return VT != MVT::nxv16i1 && VT != MVT::nxv8i1 && VT != MVT::nxv4i1 &&
VT != MVT::nxv2i1 && VT != MVT::v16i1 && VT != MVT::v8i1 &&
VT != MVT::v4i1 && VT != MVT::v2i1;
}
bool AArch64TargetLowering::shouldExpandVectorMatch(EVT VT,
unsigned SearchSize) const {
// MATCH is SVE2 and only available in non-streaming mode.
if (!Subtarget->hasSVE2() || !Subtarget->isSVEAvailable())
return true;
// Furthermore, we can only use it for 8-bit or 16-bit elements.
if (VT == MVT::nxv8i16 || VT == MVT::v8i16)
return SearchSize != 8;
if (VT == MVT::nxv16i8 || VT == MVT::v16i8 || VT == MVT::v8i8)
return SearchSize != 8 && SearchSize != 16;
return true;
}
void AArch64TargetLowering::addTypeForFixedLengthSVE(MVT VT) {
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
// By default everything must be expanded.
for (unsigned Op = 0; Op < ISD::BUILTIN_OP_END; ++Op)
setOperationAction(Op, VT, Expand);
if (VT.isFloatingPoint()) {
setCondCodeAction(ISD::SETO, VT, Expand);
setCondCodeAction(ISD::SETOLT, VT, Expand);
setCondCodeAction(ISD::SETOLE, VT, Expand);
setCondCodeAction(ISD::SETULT, VT, Expand);
setCondCodeAction(ISD::SETULE, VT, Expand);
setCondCodeAction(ISD::SETUGE, VT, Expand);
setCondCodeAction(ISD::SETUGT, VT, Expand);
setCondCodeAction(ISD::SETUEQ, VT, Expand);
setCondCodeAction(ISD::SETONE, VT, Expand);
}
TargetLoweringBase::LegalizeAction Default =
VT == MVT::v1f64 ? Expand : Custom;
// Mark integer truncating stores/extending loads as having custom lowering
if (VT.isInteger()) {
MVT InnerVT = VT.changeVectorElementType(MVT::i8);
while (InnerVT != VT) {
setTruncStoreAction(VT, InnerVT, Default);
setLoadExtAction(ISD::ZEXTLOAD, VT, InnerVT, Default);
setLoadExtAction(ISD::SEXTLOAD, VT, InnerVT, Default);
setLoadExtAction(ISD::EXTLOAD, VT, InnerVT, Default);
InnerVT = InnerVT.changeVectorElementType(
MVT::getIntegerVT(2 * InnerVT.getScalarSizeInBits()));
}
}
// Mark floating-point truncating stores/extending loads as having custom
// lowering
if (VT.isFloatingPoint()) {
MVT InnerVT = VT.changeVectorElementType(MVT::f16);
while (InnerVT != VT) {
setTruncStoreAction(VT, InnerVT, Custom);
setLoadExtAction(ISD::EXTLOAD, VT, InnerVT, Default);
InnerVT = InnerVT.changeVectorElementType(
MVT::getFloatingPointVT(2 * InnerVT.getScalarSizeInBits()));
}
}
bool PreferNEON = VT.is64BitVector() || VT.is128BitVector();
bool PreferSVE = !PreferNEON && Subtarget->isSVEAvailable();
static const unsigned MLAOps[] = {ISD::PARTIAL_REDUCE_SMLA,
ISD::PARTIAL_REDUCE_UMLA};
unsigned NumElts = VT.getVectorNumElements();
if (VT.getVectorElementType() == MVT::i64) {
setPartialReduceMLAAction(MLAOps, VT,
MVT::getVectorVT(MVT::i8, NumElts * 8), Custom);
setPartialReduceMLAAction(MLAOps, VT,
MVT::getVectorVT(MVT::i16, NumElts * 4), Custom);
setPartialReduceMLAAction(MLAOps, VT,
MVT::getVectorVT(MVT::i32, NumElts * 2), Custom);
} else if (VT.getVectorElementType() == MVT::i32) {
setPartialReduceMLAAction(MLAOps, VT,
MVT::getVectorVT(MVT::i8, NumElts * 4), Custom);
setPartialReduceMLAAction(MLAOps, VT,
MVT::getVectorVT(MVT::i16, NumElts * 2), Custom);
} else if (VT.getVectorElementType() == MVT::i16) {
setPartialReduceMLAAction(MLAOps, VT,
MVT::getVectorVT(MVT::i8, NumElts * 2), Custom);
}
if (Subtarget->hasMatMulInt8()) {
if (VT.getVectorElementType() == MVT::i32)
setPartialReduceMLAAction(ISD::PARTIAL_REDUCE_SUMLA, VT,
MVT::getVectorVT(MVT::i8, NumElts * 4), Custom);
else if (VT.getVectorElementType() == MVT::i64)
setPartialReduceMLAAction(ISD::PARTIAL_REDUCE_SUMLA, VT,
MVT::getVectorVT(MVT::i8, NumElts * 8), Custom);
}
// Lower fixed length vector operations to scalable equivalents.
setOperationAction(ISD::ABDS, VT, Default);
setOperationAction(ISD::ABDU, VT, Default);
setOperationAction(ISD::ABS, VT, Default);
setOperationAction(ISD::ADD, VT, Default);
setOperationAction(ISD::AND, VT, Default);
setOperationAction(ISD::ANY_EXTEND, VT, Default);
setOperationAction(ISD::BITCAST, VT, PreferNEON ? Legal : Default);
setOperationAction(ISD::BITREVERSE, VT, Default);
setOperationAction(ISD::BSWAP, VT, Default);
setOperationAction(ISD::BUILD_VECTOR, VT, Custom);
setOperationAction(ISD::CONCAT_VECTORS, VT, Default);
setOperationAction(ISD::CTLZ, VT, Default);
setOperationAction(ISD::CTPOP, VT, Default);
setOperationAction(ISD::CTTZ, VT, Default);
setOperationAction(ISD::EXTRACT_SUBVECTOR, VT, Default);
setOperationAction(ISD::EXTRACT_VECTOR_ELT, VT, Default);
setOperationAction(ISD::FABS, VT, Default);
setOperationAction(ISD::FADD, VT, Default);
setOperationAction(ISD::FCEIL, VT, Default);
setOperationAction(ISD::FCOPYSIGN, VT, Default);
setOperationAction(ISD::FDIV, VT, Default);
setOperationAction(ISD::FFLOOR, VT, Default);
setOperationAction(ISD::FMA, VT, Default);
setOperationAction(ISD::FMAXIMUM, VT, Default);
setOperationAction(ISD::FMAXNUM, VT, Default);
setOperationAction(ISD::FMINIMUM, VT, Default);
setOperationAction(ISD::FMINNUM, VT, Default);
setOperationAction(ISD::FMUL, VT, Default);
setOperationAction(ISD::FNEARBYINT, VT, Default);
setOperationAction(ISD::FNEG, VT, Default);
setOperationAction(ISD::FP_EXTEND, VT, Default);
setOperationAction(ISD::FP_ROUND, VT, Default);
setOperationAction(ISD::FP_TO_SINT, VT, Default);
setOperationAction(ISD::FP_TO_UINT, VT, Default);
setOperationAction(ISD::FRINT, VT, Default);
setOperationAction(ISD::LRINT, VT, Default);
setOperationAction(ISD::LLRINT, VT, Default);
setOperationAction(ISD::FROUND, VT, Default);
setOperationAction(ISD::FROUNDEVEN, VT, Default);
setOperationAction(ISD::FSQRT, VT, Default);
setOperationAction(ISD::FSUB, VT, Default);
setOperationAction(ISD::FTRUNC, VT, Default);
setOperationAction(ISD::GET_ACTIVE_LANE_MASK, VT, Default);
setOperationAction(ISD::INSERT_VECTOR_ELT, VT, Default);
setOperationAction(ISD::LOAD, VT, PreferNEON ? Legal : Default);
setOperationAction(ISD::MGATHER, VT, PreferSVE ? Default : Expand);
setOperationAction(ISD::MLOAD, VT, Default);
setOperationAction(ISD::MSCATTER, VT, PreferSVE ? Default : Expand);
setOperationAction(ISD::MSTORE, VT, Default);
setOperationAction(ISD::MUL, VT, Default);
setOperationAction(ISD::MULHS, VT, Default);
setOperationAction(ISD::MULHU, VT, Default);
setOperationAction(ISD::OR, VT, Default);
setOperationAction(ISD::SCALAR_TO_VECTOR, VT, PreferNEON ? Legal : Expand);
setOperationAction(ISD::SDIV, VT, Default);
setOperationAction(ISD::SELECT, VT, Default);
setOperationAction(ISD::SETCC, VT, Default);
setOperationAction(ISD::SHL, VT, Default);
setOperationAction(ISD::SIGN_EXTEND, VT, Default);
setOperationAction(ISD::SIGN_EXTEND_INREG, VT, Default);
setOperationAction(ISD::SINT_TO_FP, VT, Default);
setOperationAction(ISD::SMAX, VT, Default);
setOperationAction(ISD::SMIN, VT, Default);
setOperationAction(ISD::SPLAT_VECTOR, VT, Default);
setOperationAction(ISD::SRA, VT, Default);
setOperationAction(ISD::SRL, VT, Default);
setOperationAction(ISD::STORE, VT, PreferNEON ? Legal : Default);
setOperationAction(ISD::SUB, VT, Default);
setOperationAction(ISD::TRUNCATE, VT, Default);
setOperationAction(ISD::UDIV, VT, Default);
setOperationAction(ISD::UINT_TO_FP, VT, Default);
setOperationAction(ISD::UMAX, VT, Default);
setOperationAction(ISD::UMIN, VT, Default);
setOperationAction(ISD::VECREDUCE_ADD, VT, Default);
setOperationAction(ISD::VECREDUCE_AND, VT, Default);
setOperationAction(ISD::VECREDUCE_FADD, VT, Default);
setOperationAction(ISD::VECREDUCE_FMAX, VT, Default);
setOperationAction(ISD::VECREDUCE_FMIN, VT, Default);
setOperationAction(ISD::VECREDUCE_FMAXIMUM, VT, Default);
setOperationAction(ISD::VECREDUCE_FMINIMUM, VT, Default);
setOperationAction(ISD::VECREDUCE_OR, VT, Default);
setOperationAction(ISD::VECREDUCE_SEQ_FADD, VT, PreferSVE ? Default : Expand);
setOperationAction(ISD::VECREDUCE_SMAX, VT, Default);
setOperationAction(ISD::VECREDUCE_SMIN, VT, Default);
setOperationAction(ISD::VECREDUCE_UMAX, VT, Default);
setOperationAction(ISD::VECREDUCE_UMIN, VT, Default);
setOperationAction(ISD::VECREDUCE_XOR, VT, Default);
setOperationAction(ISD::VECTOR_SHUFFLE, VT, Default);
setOperationAction(ISD::VECTOR_SPLICE, VT, Default);
setOperationAction(ISD::VSELECT, VT, Default);
setOperationAction(ISD::XOR, VT, Default);
setOperationAction(ISD::ZERO_EXTEND, VT, Default);
}
void AArch64TargetLowering::addDRType(MVT VT) {
addRegisterClass(VT, &AArch64::FPR64RegClass);
if (Subtarget->isNeonAvailable())
addTypeForNEON(VT);
}
void AArch64TargetLowering::addQRType(MVT VT) {
addRegisterClass(VT, &AArch64::FPR128RegClass);
if (Subtarget->isNeonAvailable())
addTypeForNEON(VT);
}
EVT AArch64TargetLowering::getSetCCResultType(const DataLayout &,
LLVMContext &C, EVT VT) const {
if (!VT.isVector())
return MVT::i32;
if (VT.isScalableVector())
return EVT::getVectorVT(C, MVT::i1, VT.getVectorElementCount());
return VT.changeVectorElementTypeToInteger();
}
// isIntImmediate - This method tests to see if the node is a constant
// operand. If so Imm will receive the 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;
}
// 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 value.
static bool isOpcWithIntImmediate(const SDNode *N, unsigned Opc,
uint64_t &Imm) {
return N->getOpcode() == Opc &&
isIntImmediate(N->getOperand(1).getNode(), Imm);
}
static bool optimizeLogicalImm(SDValue Op, unsigned Size, uint64_t Imm,
const APInt &Demanded,
TargetLowering::TargetLoweringOpt &TLO,
unsigned NewOpc) {
uint64_t OldImm = Imm, NewImm, Enc;
uint64_t Mask = ((uint64_t)(-1LL) >> (64 - Size)), OrigMask = Mask;
// Return if the immediate is already all zeros, all ones, a bimm32 or a
// bimm64.
if (Imm == 0 || Imm == Mask ||
AArch64_AM::isLogicalImmediate(Imm & Mask, Size))
return false;
unsigned EltSize = Size;
uint64_t DemandedBits = Demanded.getZExtValue();
// Clear bits that are not demanded.
Imm &= DemandedBits;
while (true) {
// The goal here is to set the non-demanded bits in a way that minimizes
// the number of switching between 0 and 1. In order to achieve this goal,
// we set the non-demanded bits to the value of the preceding demanded bits.
// For example, if we have an immediate 0bx10xx0x1 ('x' indicates a
// non-demanded bit), we copy bit0 (1) to the least significant 'x',
// bit2 (0) to 'xx', and bit6 (1) to the most significant 'x'.
// The final result is 0b11000011.
uint64_t NonDemandedBits = ~DemandedBits;
uint64_t InvertedImm = ~Imm & DemandedBits;
uint64_t RotatedImm =
((InvertedImm << 1) | (InvertedImm >> (EltSize - 1) & 1)) &
NonDemandedBits;
uint64_t Sum = RotatedImm + NonDemandedBits;
bool Carry = NonDemandedBits & ~Sum & (1ULL << (EltSize - 1));
uint64_t Ones = (Sum + Carry) & NonDemandedBits;
NewImm = (Imm | Ones) & Mask;
// If NewImm or its bitwise NOT is a shifted mask, it is a bitmask immediate
// or all-ones or all-zeros, in which case we can stop searching. Otherwise,
// we halve the element size and continue the search.
if (isShiftedMask_64(NewImm) || isShiftedMask_64(~(NewImm | ~Mask)))
break;
// We cannot shrink the element size any further if it is 2-bits.
if (EltSize == 2)
return false;
EltSize /= 2;
Mask >>= EltSize;
uint64_t Hi = Imm >> EltSize, DemandedBitsHi = DemandedBits >> EltSize;
// Return if there is mismatch in any of the demanded bits of Imm and Hi.
if (((Imm ^ Hi) & (DemandedBits & DemandedBitsHi) & Mask) != 0)
return false;
// Merge the upper and lower halves of Imm and DemandedBits.
Imm |= Hi;
DemandedBits |= DemandedBitsHi;
}
++NumOptimizedImms;
// Replicate the element across the register width.
while (EltSize < Size) {
NewImm |= NewImm << EltSize;
EltSize *= 2;
}
(void)OldImm;
assert(((OldImm ^ NewImm) & Demanded.getZExtValue()) == 0 &&
"demanded bits should never be altered");
assert(OldImm != NewImm && "the new imm shouldn't be equal to the old imm");
// Create the new constant immediate node.
EVT VT = Op.getValueType();
SDLoc DL(Op);
SDValue New;
// If the new constant immediate is all-zeros or all-ones, let the target
// independent DAG combine optimize this node.
if (NewImm == 0 || NewImm == OrigMask) {
New = TLO.DAG.getNode(Op.getOpcode(), DL, VT, Op.getOperand(0),
TLO.DAG.getConstant(NewImm, DL, VT));
// Otherwise, create a machine node so that target independent DAG combine
// doesn't undo this optimization.
} else {
Enc = AArch64_AM::encodeLogicalImmediate(NewImm, Size);
SDValue EncConst = TLO.DAG.getTargetConstant(Enc, DL, VT);
New = SDValue(
TLO.DAG.getMachineNode(NewOpc, DL, VT, Op.getOperand(0), EncConst), 0);
}
return TLO.CombineTo(Op, New);
}
bool AArch64TargetLowering::targetShrinkDemandedConstant(
SDValue Op, const APInt &DemandedBits, const APInt &DemandedElts,
TargetLoweringOpt &TLO) const {
// Delay this optimization to as late as possible.
if (!TLO.LegalOps)
return false;
if (!EnableOptimizeLogicalImm)
return false;
EVT VT = Op.getValueType();
if (VT.isVector())
return false;
unsigned Size = VT.getSizeInBits();
if (Size != 32 && Size != 64)
return false;
// Exit early if we demand all bits.
if (DemandedBits.popcount() == Size)
return false;
unsigned NewOpc;
switch (Op.getOpcode()) {
default:
return false;
case ISD::AND:
NewOpc = Size == 32 ? AArch64::ANDWri : AArch64::ANDXri;
break;
case ISD::OR:
NewOpc = Size == 32 ? AArch64::ORRWri : AArch64::ORRXri;
break;
case ISD::XOR:
NewOpc = Size == 32 ? AArch64::EORWri : AArch64::EORXri;
break;
}
ConstantSDNode *C = dyn_cast<ConstantSDNode>(Op.getOperand(1));
if (!C)
return false;
uint64_t Imm = C->getZExtValue();
return optimizeLogicalImm(Op, Size, Imm, DemandedBits, TLO, NewOpc);
}
/// computeKnownBitsForTargetNode - Determine which of the bits specified in
/// Mask are known to be either zero or one and return them Known.
void AArch64TargetLowering::computeKnownBitsForTargetNode(
const SDValue Op, KnownBits &Known, const APInt &DemandedElts,
const SelectionDAG &DAG, unsigned Depth) const {
switch (Op.getOpcode()) {
default:
break;
case AArch64ISD::DUP: {
SDValue SrcOp = Op.getOperand(0);
Known = DAG.computeKnownBits(SrcOp, Depth + 1);
if (SrcOp.getValueSizeInBits() != Op.getScalarValueSizeInBits()) {
assert(SrcOp.getValueSizeInBits() > Op.getScalarValueSizeInBits() &&
"Expected DUP implicit truncation");
Known = Known.trunc(Op.getScalarValueSizeInBits());
}
break;
}
case AArch64ISD::CSEL: {
KnownBits Known2;
Known = DAG.computeKnownBits(Op->getOperand(0), Depth + 1);
Known2 = DAG.computeKnownBits(Op->getOperand(1), Depth + 1);
Known = Known.intersectWith(Known2);
break;
}
case AArch64ISD::BICi: {
// Compute the bit cleared value.
APInt Mask =
~(Op->getConstantOperandAPInt(1) << Op->getConstantOperandAPInt(2))
.trunc(Known.getBitWidth());
Known = DAG.computeKnownBits(Op->getOperand(0), Depth + 1);
Known &= KnownBits::makeConstant(Mask);
break;
}
case AArch64ISD::VLSHR: {
KnownBits Known2;
Known = DAG.computeKnownBits(Op->getOperand(0), Depth + 1);
Known2 = DAG.computeKnownBits(Op->getOperand(1), Depth + 1);
Known = KnownBits::lshr(Known, Known2);
break;
}
case AArch64ISD::VASHR: {
KnownBits Known2;
Known = DAG.computeKnownBits(Op->getOperand(0), Depth + 1);
Known2 = DAG.computeKnownBits(Op->getOperand(1), Depth + 1);
Known = KnownBits::ashr(Known, Known2);
break;
}
case AArch64ISD::VSHL: {
KnownBits Known2;
Known = DAG.computeKnownBits(Op->getOperand(0), Depth + 1);
Known2 = DAG.computeKnownBits(Op->getOperand(1), Depth + 1);
Known = KnownBits::shl(Known, Known2);
break;
}
case AArch64ISD::MOVI: {
Known = KnownBits::makeConstant(
APInt(Known.getBitWidth(), Op->getConstantOperandVal(0)));
break;
}
case AArch64ISD::LOADgot:
case AArch64ISD::ADDlow: {
if (!Subtarget->isTargetILP32())
break;
// In ILP32 mode all valid pointers are in the low 4GB of the address-space.
Known.Zero = APInt::getHighBitsSet(64, 32);
break;
}
case AArch64ISD::ASSERT_ZEXT_BOOL: {
Known = DAG.computeKnownBits(Op->getOperand(0), Depth + 1);
Known.Zero |= APInt(Known.getBitWidth(), 0xFE);
break;
}
case ISD::INTRINSIC_W_CHAIN: {
Intrinsic::ID IntID =
static_cast<Intrinsic::ID>(Op->getConstantOperandVal(1));
switch (IntID) {
default: return;
case Intrinsic::aarch64_ldaxr:
case Intrinsic::aarch64_ldxr: {
unsigned BitWidth = Known.getBitWidth();
EVT VT = cast<MemIntrinsicSDNode>(Op)->getMemoryVT();
unsigned MemBits = VT.getScalarSizeInBits();
Known.Zero |= APInt::getHighBitsSet(BitWidth, BitWidth - MemBits);
return;
}
}
break;
}
case ISD::INTRINSIC_WO_CHAIN:
case ISD::INTRINSIC_VOID: {
unsigned IntNo = Op.getConstantOperandVal(0);
switch (IntNo) {
default:
break;
case Intrinsic::aarch64_neon_uaddlv: {
MVT VT = Op.getOperand(1).getValueType().getSimpleVT();
unsigned BitWidth = Known.getBitWidth();
if (VT == MVT::v8i8 || VT == MVT::v16i8) {
unsigned Bound = (VT == MVT::v8i8) ? 11 : 12;
assert(BitWidth >= Bound && "Unexpected width!");
APInt Mask = APInt::getHighBitsSet(BitWidth, BitWidth - Bound);
Known.Zero |= Mask;
}
break;
}
case Intrinsic::aarch64_neon_umaxv:
case Intrinsic::aarch64_neon_uminv: {
// Figure out the datatype of the vector operand. The UMINV instruction
// will zero extend the result, so we can mark as known zero all the
// bits larger than the element datatype. 32-bit or larget doesn't need
// this as those are legal types and will be handled by isel directly.
MVT VT = Op.getOperand(1).getValueType().getSimpleVT();
unsigned BitWidth = Known.getBitWidth();
if (VT == MVT::v8i8 || VT == MVT::v16i8) {
assert(BitWidth >= 8 && "Unexpected width!");
APInt Mask = APInt::getHighBitsSet(BitWidth, BitWidth - 8);
Known.Zero |= Mask;
} else if (VT == MVT::v4i16 || VT == MVT::v8i16) {
assert(BitWidth >= 16 && "Unexpected width!");
APInt Mask = APInt::getHighBitsSet(BitWidth, BitWidth - 16);
Known.Zero |= Mask;
}
break;
} break;
}
}
}
}
unsigned AArch64TargetLowering::ComputeNumSignBitsForTargetNode(
SDValue Op, const APInt &DemandedElts, const SelectionDAG &DAG,
unsigned Depth) const {
EVT VT = Op.getValueType();
unsigned VTBits = VT.getScalarSizeInBits();
unsigned Opcode = Op.getOpcode();
switch (Opcode) {
case AArch64ISD::FCMEQ:
case AArch64ISD::FCMGE:
case AArch64ISD::FCMGT:
// Compares return either 0 or all-ones
return VTBits;
case AArch64ISD::VASHR: {
unsigned Tmp =
DAG.ComputeNumSignBits(Op.getOperand(0), DemandedElts, Depth + 1);
return std::min<uint64_t>(Tmp + Op.getConstantOperandVal(1), VTBits);
}
}
return 1;
}
MVT AArch64TargetLowering::getScalarShiftAmountTy(const DataLayout &DL,
EVT) const {
return MVT::i64;
}
bool AArch64TargetLowering::allowsMisalignedMemoryAccesses(
EVT VT, unsigned AddrSpace, Align Alignment, MachineMemOperand::Flags Flags,
unsigned *Fast) const {
// Allow SVE loads/stores where the alignment >= the size of the element type,
// even with +strict-align. Predicated SVE loads/stores (e.g. ld1/st1), used
// for stores that come from IR, only require element-size alignment (even if
// unaligned accesses are disabled). Without this, these will be forced to
// have 16-byte alignment with +strict-align (and fail to lower as we don't
// yet support TLI.expandUnalignedLoad() and TLI.expandUnalignedStore()).
if (VT.isScalableVector()) {
unsigned ElementSizeBits = VT.getScalarSizeInBits();
if (ElementSizeBits % 8 == 0 && Alignment >= Align(ElementSizeBits / 8))
return true;
}
if (Subtarget->requiresStrictAlign())
return false;
if (Fast) {
// Some CPUs are fine with unaligned stores except for 128-bit ones.
*Fast = !Subtarget->isMisaligned128StoreSlow() || VT.getStoreSize() != 16 ||
// See comments in performSTORECombine() for more details about
// these conditions.
// Code that uses clang vector extensions can mark that it
// wants unaligned accesses to be treated as fast by
// underspecifying alignment to be 1 or 2.
Alignment <= 2 ||
// Disregard v2i64. Memcpy lowering produces those and splitting
// them regresses performance on micro-benchmarks and olden/bh.
VT == MVT::v2i64;
}
return true;
}
// Same as above but handling LLTs instead.
bool AArch64TargetLowering::allowsMisalignedMemoryAccesses(
LLT Ty, unsigned AddrSpace, Align Alignment, MachineMemOperand::Flags Flags,
unsigned *Fast) const {
if (Subtarget->requiresStrictAlign())
return false;
if (Fast) {
// Some CPUs are fine with unaligned stores except for 128-bit ones.
*Fast = !Subtarget->isMisaligned128StoreSlow() ||
Ty.getSizeInBytes() != 16 ||
// See comments in performSTORECombine() for more details about
// these conditions.
// Code that uses clang vector extensions can mark that it
// wants unaligned accesses to be treated as fast by
// underspecifying alignment to be 1 or 2.
Alignment <= 2 ||
// Disregard v2i64. Memcpy lowering produces those and splitting
// them regresses performance on micro-benchmarks and olden/bh.
Ty == LLT::fixed_vector(2, 64);
}
return true;
}
FastISel *
AArch64TargetLowering::createFastISel(FunctionLoweringInfo &funcInfo,
const TargetLibraryInfo *libInfo) const {
return AArch64::createFastISel(funcInfo, libInfo);
}
MachineBasicBlock *
AArch64TargetLowering::EmitF128CSEL(MachineInstr &MI,
MachineBasicBlock *MBB) const {
// We materialise the F128CSEL pseudo-instruction as some control flow and a
// phi node:
// OrigBB:
// [... previous instrs leading to comparison ...]
// b.ne TrueBB
// b EndBB
// TrueBB:
// ; Fallthrough
// EndBB:
// Dest = PHI [IfTrue, TrueBB], [IfFalse, OrigBB]
MachineFunction *MF = MBB->getParent();
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
const BasicBlock *LLVM_BB = MBB->getBasicBlock();
DebugLoc DL = MI.getDebugLoc();
MachineFunction::iterator It = ++MBB->getIterator();
Register DestReg = MI.getOperand(0).getReg();
Register IfTrueReg = MI.getOperand(1).getReg();
Register IfFalseReg = MI.getOperand(2).getReg();
unsigned CondCode = MI.getOperand(3).getImm();
bool NZCVKilled = MI.getOperand(4).isKill();
MachineBasicBlock *TrueBB = MF->CreateMachineBasicBlock(LLVM_BB);
MachineBasicBlock *EndBB = MF->CreateMachineBasicBlock(LLVM_BB);
MF->insert(It, TrueBB);
MF->insert(It, EndBB);
// Transfer rest of current basic-block to EndBB
EndBB->splice(EndBB->begin(), MBB, std::next(MachineBasicBlock::iterator(MI)),
MBB->end());
EndBB->transferSuccessorsAndUpdatePHIs(MBB);
BuildMI(MBB, DL, TII->get(AArch64::Bcc)).addImm(CondCode).addMBB(TrueBB);
BuildMI(MBB, DL, TII->get(AArch64::B)).addMBB(EndBB);
MBB->addSuccessor(TrueBB);
MBB->addSuccessor(EndBB);
// TrueBB falls through to the end.
TrueBB->addSuccessor(EndBB);
if (!NZCVKilled) {
TrueBB->addLiveIn(AArch64::NZCV);
EndBB->addLiveIn(AArch64::NZCV);
}
BuildMI(*EndBB, EndBB->begin(), DL, TII->get(AArch64::PHI), DestReg)
.addReg(IfTrueReg)
.addMBB(TrueBB)
.addReg(IfFalseReg)
.addMBB(MBB);
MI.eraseFromParent();
return EndBB;
}
MachineBasicBlock *AArch64TargetLowering::EmitLoweredCatchRet(
MachineInstr &MI, MachineBasicBlock *BB) const {
assert(!isAsynchronousEHPersonality(classifyEHPersonality(
BB->getParent()->getFunction().getPersonalityFn())) &&
"SEH does not use catchret!");
return BB;
}
MachineBasicBlock *
AArch64TargetLowering::EmitDynamicProbedAlloc(MachineInstr &MI,
MachineBasicBlock *MBB) const {
MachineFunction &MF = *MBB->getParent();
MachineBasicBlock::iterator MBBI = MI.getIterator();
const AArch64InstrInfo &TII =
*MF.getSubtarget<AArch64Subtarget>().getInstrInfo();
Register TargetReg = MI.getOperand(0).getReg();
MachineBasicBlock::iterator NextInst =
TII.probedStackAlloc(MBBI, TargetReg, false);
MI.eraseFromParent();
return NextInst->getParent();
}
MachineBasicBlock *
AArch64TargetLowering::EmitTileLoad(unsigned Opc, unsigned BaseReg,
MachineInstr &MI,
MachineBasicBlock *BB) const {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
MachineInstrBuilder MIB = BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(Opc));
MIB.addReg(BaseReg + MI.getOperand(0).getImm(), RegState::Define);
MIB.add(MI.getOperand(1)); // slice index register
MIB.add(MI.getOperand(2)); // slice index offset
MIB.add(MI.getOperand(3)); // pg
MIB.add(MI.getOperand(4)); // base
MIB.add(MI.getOperand(5)); // offset
MI.eraseFromParent(); // The pseudo is gone now.
return BB;
}
MachineBasicBlock *
AArch64TargetLowering::EmitFill(MachineInstr &MI, MachineBasicBlock *BB) const {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
MachineInstrBuilder MIB =
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::LDR_ZA));
MIB.addReg(AArch64::ZA, RegState::Define);
MIB.add(MI.getOperand(0)); // Vector select register
MIB.add(MI.getOperand(1)); // Vector select offset
MIB.add(MI.getOperand(2)); // Base
MIB.add(MI.getOperand(1)); // Offset, same as vector select offset
MI.eraseFromParent(); // The pseudo is gone now.
return BB;
}
MachineBasicBlock *AArch64TargetLowering::EmitZTInstr(MachineInstr &MI,
MachineBasicBlock *BB,
unsigned Opcode,
bool Op0IsDef) const {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
MachineInstrBuilder MIB;
MIB = BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(Opcode))
.addReg(MI.getOperand(0).getReg(), Op0IsDef ? RegState::Define : 0);
for (unsigned I = 1; I < MI.getNumOperands(); ++I)
MIB.add(MI.getOperand(I));
MI.eraseFromParent(); // The pseudo is gone now.
return BB;
}
MachineBasicBlock *
AArch64TargetLowering::EmitZAInstr(unsigned Opc, unsigned BaseReg,
MachineInstr &MI,
MachineBasicBlock *BB) const {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
MachineInstrBuilder MIB = BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(Opc));
unsigned StartIdx = 0;
bool HasTile = BaseReg != AArch64::ZA;
bool HasZPROut = HasTile && MI.getOperand(0).isReg();
if (HasZPROut) {
MIB.add(MI.getOperand(StartIdx)); // Output ZPR
++StartIdx;
}
if (HasTile) {
MIB.addReg(BaseReg + MI.getOperand(StartIdx).getImm(),
RegState::Define); // Output ZA Tile
MIB.addReg(BaseReg + MI.getOperand(StartIdx).getImm()); // Input Za Tile
StartIdx++;
} else {
// Avoids all instructions with mnemonic za.<sz>[Reg, Imm,
if (MI.getOperand(0).isReg() && !MI.getOperand(1).isImm()) {
MIB.add(MI.getOperand(StartIdx)); // Output ZPR
++StartIdx;
}
MIB.addReg(BaseReg, RegState::Define).addReg(BaseReg);
}
for (unsigned I = StartIdx; I < MI.getNumOperands(); ++I)
MIB.add(MI.getOperand(I));
MI.eraseFromParent(); // The pseudo is gone now.
return BB;
}
MachineBasicBlock *
AArch64TargetLowering::EmitZero(MachineInstr &MI, MachineBasicBlock *BB) const {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
MachineInstrBuilder MIB =
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::ZERO_M));
MIB.add(MI.getOperand(0)); // Mask
unsigned Mask = MI.getOperand(0).getImm();
for (unsigned I = 0; I < 8; I++) {
if (Mask & (1 << I))
MIB.addDef(AArch64::ZAD0 + I, RegState::ImplicitDefine);
}
MI.eraseFromParent(); // The pseudo is gone now.
return BB;
}
MachineBasicBlock *
AArch64TargetLowering::EmitInitTPIDR2Object(MachineInstr &MI,
MachineBasicBlock *BB) const {
MachineFunction *MF = BB->getParent();
MachineFrameInfo &MFI = MF->getFrameInfo();
AArch64FunctionInfo *FuncInfo = MF->getInfo<AArch64FunctionInfo>();
TPIDR2Object &TPIDR2 = FuncInfo->getTPIDR2Obj();
if (TPIDR2.Uses > 0) {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
// Store the buffer pointer to the TPIDR2 stack object.
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::STRXui))
.addReg(MI.getOperand(0).getReg())
.addFrameIndex(TPIDR2.FrameIndex)
.addImm(0);
// Set the reserved bytes (10-15) to zero
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::STRHHui))
.addReg(AArch64::WZR)
.addFrameIndex(TPIDR2.FrameIndex)
.addImm(5);
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::STRWui))
.addReg(AArch64::WZR)
.addFrameIndex(TPIDR2.FrameIndex)
.addImm(3);
} else
MFI.RemoveStackObject(TPIDR2.FrameIndex);
BB->remove_instr(&MI);
return BB;
}
MachineBasicBlock *
AArch64TargetLowering::EmitAllocateZABuffer(MachineInstr &MI,
MachineBasicBlock *BB) const {
MachineFunction *MF = BB->getParent();
MachineFrameInfo &MFI = MF->getFrameInfo();
AArch64FunctionInfo *FuncInfo = MF->getInfo<AArch64FunctionInfo>();
// TODO This function grows the stack with a subtraction, which doesn't work
// on Windows. Some refactoring to share the functionality in
// LowerWindowsDYNAMIC_STACKALLOC will be required once the Windows ABI
// supports SME
assert(!MF->getSubtarget<AArch64Subtarget>().isTargetWindows() &&
"Lazy ZA save is not yet supported on Windows");
TPIDR2Object &TPIDR2 = FuncInfo->getTPIDR2Obj();
if (TPIDR2.Uses > 0) {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
MachineRegisterInfo &MRI = MF->getRegInfo();
// The SUBXrs below won't always be emitted in a form that accepts SP
// directly
Register SP = MRI.createVirtualRegister(&AArch64::GPR64RegClass);
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(TargetOpcode::COPY), SP)
.addReg(AArch64::SP);
// Allocate a lazy-save buffer object of the size given, normally SVL * SVL
auto Size = MI.getOperand(1).getReg();
auto Dest = MI.getOperand(0).getReg();
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::MSUBXrrr), Dest)
.addReg(Size)
.addReg(Size)
.addReg(SP);
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(TargetOpcode::COPY),
AArch64::SP)
.addReg(Dest);
// We have just allocated a variable sized object, tell this to PEI.
MFI.CreateVariableSizedObject(Align(16), nullptr);
}
BB->remove_instr(&MI);
return BB;
}
// TODO: Find a way to merge this with EmitAllocateZABuffer.
MachineBasicBlock *
AArch64TargetLowering::EmitAllocateSMESaveBuffer(MachineInstr &MI,
MachineBasicBlock *BB) const {
MachineFunction *MF = BB->getParent();
MachineFrameInfo &MFI = MF->getFrameInfo();
AArch64FunctionInfo *FuncInfo = MF->getInfo<AArch64FunctionInfo>();
assert(!MF->getSubtarget<AArch64Subtarget>().isTargetWindows() &&
"Lazy ZA save is not yet supported on Windows");
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
if (FuncInfo->isSMESaveBufferUsed()) {
// Allocate a buffer object of the size given by MI.getOperand(1).
auto Size = MI.getOperand(1).getReg();
auto Dest = MI.getOperand(0).getReg();
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::SUBXrx64), AArch64::SP)
.addReg(AArch64::SP)
.addReg(Size)
.addImm(AArch64_AM::getArithExtendImm(AArch64_AM::UXTX, 0));
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(TargetOpcode::COPY), Dest)
.addReg(AArch64::SP);
// We have just allocated a variable sized object, tell this to PEI.
MFI.CreateVariableSizedObject(Align(16), nullptr);
} else
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(TargetOpcode::IMPLICIT_DEF),
MI.getOperand(0).getReg());
BB->remove_instr(&MI);
return BB;
}
MachineBasicBlock *
AArch64TargetLowering::EmitGetSMESaveSize(MachineInstr &MI,
MachineBasicBlock *BB) const {
// If the buffer is used, emit a call to __arm_sme_state_size()
MachineFunction *MF = BB->getParent();
AArch64FunctionInfo *FuncInfo = MF->getInfo<AArch64FunctionInfo>();
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
if (FuncInfo->isSMESaveBufferUsed()) {
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(AArch64::BL))
.addExternalSymbol("__arm_sme_state_size")
.addReg(AArch64::X0, RegState::ImplicitDefine)
.addRegMask(TRI->getCallPreservedMask(
*MF, CallingConv::
AArch64_SME_ABI_Support_Routines_PreserveMost_From_X1));
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(TargetOpcode::COPY),
MI.getOperand(0).getReg())
.addReg(AArch64::X0);
} else
BuildMI(*BB, MI, MI.getDebugLoc(), TII->get(TargetOpcode::COPY),
MI.getOperand(0).getReg())
.addReg(AArch64::XZR);
BB->remove_instr(&MI);
return BB;
}
MachineBasicBlock *AArch64TargetLowering::EmitInstrWithCustomInserter(
MachineInstr &MI, MachineBasicBlock *BB) const {
int SMEOrigInstr = AArch64::getSMEPseudoMap(MI.getOpcode());
if (SMEOrigInstr != -1) {
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
uint64_t SMEMatrixType =
TII->get(MI.getOpcode()).TSFlags & AArch64::SMEMatrixTypeMask;
switch (SMEMatrixType) {
case (AArch64::SMEMatrixArray):
return EmitZAInstr(SMEOrigInstr, AArch64::ZA, MI, BB);
case (AArch64::SMEMatrixTileB):
return EmitZAInstr(SMEOrigInstr, AArch64::ZAB0, MI, BB);
case (AArch64::SMEMatrixTileH):
return EmitZAInstr(SMEOrigInstr, AArch64::ZAH0, MI, BB);
case (AArch64::SMEMatrixTileS):
return EmitZAInstr(SMEOrigInstr, AArch64::ZAS0, MI, BB);
case (AArch64::SMEMatrixTileD):
return EmitZAInstr(SMEOrigInstr, AArch64::ZAD0, MI, BB);
case (AArch64::SMEMatrixTileQ):
return EmitZAInstr(SMEOrigInstr, AArch64::ZAQ0, MI, BB);
}
}
switch (MI.getOpcode()) {
default:
#ifndef NDEBUG
MI.dump();
#endif
llvm_unreachable("Unexpected instruction for custom inserter!");
case AArch64::InitTPIDR2Obj:
return EmitInitTPIDR2Object(MI, BB);
case AArch64::AllocateZABuffer:
return EmitAllocateZABuffer(MI, BB);
case AArch64::AllocateSMESaveBuffer:
return EmitAllocateSMESaveBuffer(MI, BB);
case AArch64::GetSMESaveSize:
return EmitGetSMESaveSize(MI, BB);
case AArch64::F128CSEL:
return EmitF128CSEL(MI, BB);
case TargetOpcode::STATEPOINT:
// STATEPOINT is a pseudo instruction which has no implicit defs/uses
// while bl call instruction (where statepoint will be lowered at the end)
// has implicit def. This def is early-clobber as it will be set at
// the moment of the call and earlier than any use is read.
// Add this implicit dead def here as a workaround.
MI.addOperand(*MI.getMF(),
MachineOperand::CreateReg(
AArch64::LR, /*isDef*/ true,
/*isImp*/ true, /*isKill*/ false, /*isDead*/ true,
/*isUndef*/ false, /*isEarlyClobber*/ true));
[[fallthrough]];
case TargetOpcode::STACKMAP:
case TargetOpcode::PATCHPOINT:
return emitPatchPoint(MI, BB);
case TargetOpcode::PATCHABLE_EVENT_CALL:
case TargetOpcode::PATCHABLE_TYPED_EVENT_CALL:
return BB;
case AArch64::CATCHRET:
return EmitLoweredCatchRet(MI, BB);
case AArch64::PROBED_STACKALLOC_DYN:
return EmitDynamicProbedAlloc(MI, BB);
case AArch64::LD1_MXIPXX_H_PSEUDO_B:
return EmitTileLoad(AArch64::LD1_MXIPXX_H_B, AArch64::ZAB0, MI, BB);
case AArch64::LD1_MXIPXX_H_PSEUDO_H:
return EmitTileLoad(AArch64::LD1_MXIPXX_H_H, AArch64::ZAH0, MI, BB);
case AArch64::LD1_MXIPXX_H_PSEUDO_S:
return EmitTileLoad(AArch64::LD1_MXIPXX_H_S, AArch64::ZAS0, MI, BB);
case AArch64::LD1_MXIPXX_H_PSEUDO_D:
return EmitTileLoad(AArch64::LD1_MXIPXX_H_D, AArch64::ZAD0, MI, BB);
case AArch64::LD1_MXIPXX_H_PSEUDO_Q:
return EmitTileLoad(AArch64::LD1_MXIPXX_H_Q, AArch64::ZAQ0, MI, BB);
case AArch64::LD1_MXIPXX_V_PSEUDO_B:
return EmitTileLoad(AArch64::LD1_MXIPXX_V_B, AArch64::ZAB0, MI, BB);
case AArch64::LD1_MXIPXX_V_PSEUDO_H:
return EmitTileLoad(AArch64::LD1_MXIPXX_V_H, AArch64::ZAH0, MI, BB);
case AArch64::LD1_MXIPXX_V_PSEUDO_S:
return EmitTileLoad(AArch64::LD1_MXIPXX_V_S, AArch64::ZAS0, MI, BB);
case AArch64::LD1_MXIPXX_V_PSEUDO_D:
return EmitTileLoad(AArch64::LD1_MXIPXX_V_D, AArch64::ZAD0, MI, BB);
case AArch64::LD1_MXIPXX_V_PSEUDO_Q:
return EmitTileLoad(AArch64::LD1_MXIPXX_V_Q, AArch64::ZAQ0, MI, BB);
case AArch64::LDR_ZA_PSEUDO:
return EmitFill(MI, BB);
case AArch64::LDR_TX_PSEUDO:
return EmitZTInstr(MI, BB, AArch64::LDR_TX, /*Op0IsDef=*/true);
case AArch64::STR_TX_PSEUDO:
return EmitZTInstr(MI, BB, AArch64::STR_TX, /*Op0IsDef=*/false);
case AArch64::ZERO_M_PSEUDO:
return EmitZero(MI, BB);
case AArch64::ZERO_T_PSEUDO:
return EmitZTInstr(MI, BB, AArch64::ZERO_T, /*Op0IsDef=*/true);
case AArch64::MOVT_TIZ_PSEUDO:
return EmitZTInstr(MI, BB, AArch64::MOVT_TIZ, /*Op0IsDef=*/true);
}
}
//===----------------------------------------------------------------------===//
// AArch64 Lowering private implementation.
//===----------------------------------------------------------------------===//
//===----------------------------------------------------------------------===//
// Lowering Code
//===----------------------------------------------------------------------===//
// Forward declarations of SVE fixed length lowering helpers
static EVT getContainerForFixedLengthVector(SelectionDAG &DAG, EVT VT);
static SDValue convertToScalableVector(SelectionDAG &DAG, EVT VT, SDValue V);
static SDValue convertFromScalableVector(SelectionDAG &DAG, EVT VT, SDValue V);
static SDValue convertFixedMaskToScalableVector(SDValue Mask,
SelectionDAG &DAG);
static SDValue getPredicateForVector(SelectionDAG &DAG, SDLoc &DL, EVT VT);
static SDValue getPredicateForScalableVector(SelectionDAG &DAG, SDLoc &DL,
EVT VT);
/// isZerosVector - Check whether SDNode N is a zero-filled vector.
static bool isZerosVector(const SDNode *N) {
// Look through a bit convert.
while (N->getOpcode() == ISD::BITCAST)
N = N->getOperand(0).getNode();
if (ISD::isConstantSplatVectorAllZeros(N))
return true;
if (N->getOpcode() != AArch64ISD::DUP)
return false;
auto Opnd0 = N->getOperand(0);
return isNullConstant(Opnd0) || isNullFPConstant(Opnd0);
}
/// changeIntCCToAArch64CC - Convert a DAG integer condition code to an AArch64
/// CC
static AArch64CC::CondCode changeIntCCToAArch64CC(ISD::CondCode CC) {
switch (CC) {
default:
llvm_unreachable("Unknown condition code!");
case ISD::SETNE:
return AArch64CC::NE;
case ISD::SETEQ:
return AArch64CC::EQ;
case ISD::SETGT:
return AArch64CC::GT;
case ISD::SETGE:
return AArch64CC::GE;
case ISD::SETLT:
return AArch64CC::LT;
case ISD::SETLE:
return AArch64CC::LE;
case ISD::SETUGT:
return AArch64CC::HI;
case ISD::SETUGE:
return AArch64CC::HS;
case ISD::SETULT:
return AArch64CC::LO;
case ISD::SETULE:
return AArch64CC::LS;
}
}
/// changeFPCCToAArch64CC - Convert a DAG fp condition code to an AArch64 CC.
static void changeFPCCToAArch64CC(ISD::CondCode CC,
AArch64CC::CondCode &CondCode,
AArch64CC::CondCode &CondCode2) {
CondCode2 = AArch64CC::AL;
switch (CC) {
default:
llvm_unreachable("Unknown FP condition!");
case ISD::SETEQ:
case ISD::SETOEQ:
CondCode = AArch64CC::EQ;
break;
case ISD::SETGT:
case ISD::SETOGT:
CondCode = AArch64CC::GT;
break;
case ISD::SETGE:
case ISD::SETOGE:
CondCode = AArch64CC::GE;
break;
case ISD::SETOLT:
CondCode = AArch64CC::MI;
break;
case ISD::SETOLE:
CondCode = AArch64CC::LS;
break;
case ISD::SETONE:
CondCode = AArch64CC::MI;
CondCode2 = AArch64CC::GT;
break;
case ISD::SETO:
CondCode = AArch64CC::VC;
break;
case ISD::SETUO:
CondCode = AArch64CC::VS;
break;
case ISD::SETUEQ:
CondCode = AArch64CC::EQ;
CondCode2 = AArch64CC::VS;
break;
case ISD::SETUGT:
CondCode = AArch64CC::HI;
break;
case ISD::SETUGE:
CondCode = AArch64CC::PL;
break;
case ISD::SETLT:
case ISD::SETULT:
CondCode = AArch64CC::LT;
break;
case ISD::SETLE:
case ISD::SETULE:
CondCode = AArch64CC::LE;
break;
case ISD::SETNE:
case ISD::SETUNE:
CondCode = AArch64CC::NE;
break;
}
}
/// Convert a DAG fp condition code to an AArch64 CC.
/// This differs from changeFPCCToAArch64CC in that it returns cond codes that
/// should be AND'ed instead of OR'ed.
static void changeFPCCToANDAArch64CC(ISD::CondCode CC,
AArch64CC::CondCode &CondCode,
AArch64CC::CondCode &CondCode2) {
CondCode2 = AArch64CC::AL;
switch (CC) {
default:
changeFPCCToAArch64CC(CC, CondCode, CondCode2);
assert(CondCode2 == AArch64CC::AL);
break;
case ISD::SETONE:
// (a one b)
// == ((a olt b) || (a ogt b))
// == ((a ord b) && (a une b))
CondCode = AArch64CC::VC;
CondCode2 = AArch64CC::NE;
break;
case ISD::SETUEQ:
// (a ueq b)
// == ((a uno b) || (a oeq b))
// == ((a ule b) && (a uge b))
CondCode = AArch64CC::PL;
CondCode2 = AArch64CC::LE;
break;
}
}
/// changeVectorFPCCToAArch64CC - Convert a DAG fp condition code to an AArch64
/// CC usable with the vector instructions. Fewer operations are available
/// without a real NZCV register, so we have to use less efficient combinations
/// to get the same effect.
static void changeVectorFPCCToAArch64CC(ISD::CondCode CC,
AArch64CC::CondCode &CondCode,
AArch64CC::CondCode &CondCode2,
bool &Invert) {
Invert = false;
switch (CC) {
default:
// Mostly the scalar mappings work fine.
changeFPCCToAArch64CC(CC, CondCode, CondCode2);
break;
case ISD::SETUO:
Invert = true;
[[fallthrough]];
case ISD::SETO:
CondCode = AArch64CC::MI;
CondCode2 = AArch64CC::GE;
break;
case ISD::SETUEQ:
case ISD::SETULT:
case ISD::SETULE:
case ISD::SETUGT:
case ISD::SETUGE:
// All of the compare-mask comparisons are ordered, but we can switch
// between the two by a double inversion. E.g. ULE == !OGT.
Invert = true;
changeFPCCToAArch64CC(getSetCCInverse(CC, /* FP inverse */ MVT::f32),
CondCode, CondCode2);
break;
}
}
static bool isLegalArithImmed(uint64_t C) {
// Matches AArch64DAGToDAGISel::SelectArithImmed().
bool IsLegal = (C >> 12 == 0) || ((C & 0xFFFULL) == 0 && C >> 24 == 0);
LLVM_DEBUG(dbgs() << "Is imm " << C
<< " legal: " << (IsLegal ? "yes\n" : "no\n"));
return IsLegal;
}
bool isLegalCmpImmed(APInt C) {
// Works for negative immediates too, as it can be written as an ADDS
// instruction with a negated immediate.
return isLegalArithImmed(C.abs().getZExtValue());
}
static bool isSafeSignedCMN(SDValue Op, SelectionDAG &DAG) {
// 0 - INT_MIN sign wraps, so no signed wrap means cmn is safe.
if (Op->getFlags().hasNoSignedWrap())
return true;
// We can still figure out if the second operand is safe to use
// in a CMN instruction by checking if it is known to be not the minimum
// signed value. If it is not, then we can safely use CMN.
// Note: We can eventually remove this check and simply rely on
// Op->getFlags().hasNoSignedWrap() once SelectionDAG/ISelLowering
// consistently sets them appropriately when making said nodes.
KnownBits KnownSrc = DAG.computeKnownBits(Op.getOperand(1));
return !KnownSrc.getSignedMinValue().isMinSignedValue();
}
// Can a (CMP op1, (sub 0, op2) be turned into a CMN instruction on
// the grounds that "op1 - (-op2) == op1 + op2" ? Not always, the C and V flags
// can be set differently by this operation. It comes down to whether
// "SInt(~op2)+1 == SInt(~op2+1)" (and the same for UInt). If they are then
// everything is fine. If not then the optimization is wrong. Thus general
// comparisons are only valid if op2 != 0 and op2 != INT_MIN.
//
// So, finally, the only LLVM-native comparisons that don't mention C or V
// are the ones that aren't unsigned comparisons. They're the only ones we can
// safely use CMN for in the absence of information about op2.
static bool isCMN(SDValue Op, ISD::CondCode CC, SelectionDAG &DAG) {
return Op.getOpcode() == ISD::SUB && isNullConstant(Op.getOperand(0)) &&
(isIntEqualitySetCC(CC) ||
(isUnsignedIntSetCC(CC) && DAG.isKnownNeverZero(Op.getOperand(1))) ||
(isSignedIntSetCC(CC) && isSafeSignedCMN(Op, DAG)));
}
static SDValue emitStrictFPComparison(SDValue LHS, SDValue RHS, const SDLoc &DL,
SelectionDAG &DAG, SDValue Chain,
bool IsSignaling) {
EVT VT = LHS.getValueType();
assert(VT != MVT::f128);
const bool FullFP16 = DAG.getSubtarget<AArch64Subtarget>().hasFullFP16();
if ((VT == MVT::f16 && !FullFP16) || VT == MVT::bf16) {
LHS = DAG.getNode(ISD::STRICT_FP_EXTEND, DL, {MVT::f32, MVT::Other},
{Chain, LHS});
RHS = DAG.getNode(ISD::STRICT_FP_EXTEND, DL, {MVT::f32, MVT::Other},
{LHS.getValue(1), RHS});
Chain = RHS.getValue(1);
}
unsigned Opcode =
IsSignaling ? AArch64ISD::STRICT_FCMPE : AArch64ISD::STRICT_FCMP;
return DAG.getNode(Opcode, DL, {MVT::i32, MVT::Other}, {Chain, LHS, RHS});
}
static SDValue emitComparison(SDValue LHS, SDValue RHS, ISD::CondCode CC,
const SDLoc &DL, SelectionDAG &DAG) {
EVT VT = LHS.getValueType();
const bool FullFP16 = DAG.getSubtarget<AArch64Subtarget>().hasFullFP16();
if (VT.isFloatingPoint()) {
assert(VT != MVT::f128);
if ((VT == MVT::f16 && !FullFP16) || VT == MVT::bf16) {
LHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, LHS);
RHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, RHS);
}
return DAG.getNode(AArch64ISD::FCMP, DL, MVT::i32, LHS, RHS);
}
// The CMP instruction is just an alias for SUBS, and representing it as
// SUBS means that it's possible to get CSE with subtract operations.
// A later phase can perform the optimization of setting the destination
// register to WZR/XZR if it ends up being unused.
unsigned Opcode = AArch64ISD::SUBS;
if (isCMN(RHS, CC, DAG)) {
// Can we combine a (CMP op1, (sub 0, op2) into a CMN instruction ?
Opcode = AArch64ISD::ADDS;
RHS = RHS.getOperand(1);
} else if (LHS.getOpcode() == ISD::SUB && isNullConstant(LHS.getOperand(0)) &&
isIntEqualitySetCC(CC)) {
// As we are looking for EQ/NE compares, the operands can be commuted ; can
// we combine a (CMP (sub 0, op1), op2) into a CMN instruction ?
Opcode = AArch64ISD::ADDS;
LHS = LHS.getOperand(1);
} else if (isNullConstant(RHS) && !isUnsignedIntSetCC(CC)) {
if (LHS.getOpcode() == ISD::AND) {
// Similarly, (CMP (and X, Y), 0) can be implemented with a TST
// (a.k.a. ANDS) except that the flags are only guaranteed to work for one
// of the signed comparisons.
const SDValue ANDSNode =
DAG.getNode(AArch64ISD::ANDS, DL, DAG.getVTList(VT, MVT_CC),
LHS.getOperand(0), LHS.getOperand(1));
// Replace all users of (and X, Y) with newly generated (ands X, Y)
DAG.ReplaceAllUsesWith(LHS, ANDSNode);
return ANDSNode.getValue(1);
} else if (LHS.getOpcode() == AArch64ISD::ANDS) {
// Use result of ANDS
return LHS.getValue(1);
}
}
return DAG.getNode(Opcode, DL, DAG.getVTList(VT, MVT_CC), LHS, RHS)
.getValue(1);
}
/// \defgroup AArch64CCMP CMP;CCMP matching
///
/// These functions deal with the formation of CMP;CCMP;... sequences.
/// The CCMP/CCMN/FCCMP/FCCMPE instructions allow the conditional execution of
/// a comparison. They set the NZCV flags to a predefined value if their
/// predicate is false. This allows to express arbitrary conjunctions, for
/// example "cmp 0 (and (setCA (cmp A)) (setCB (cmp B)))"
/// expressed as:
/// cmp A
/// ccmp B, inv(CB), CA
/// check for CB flags
///
/// This naturally lets us implement chains of AND operations with SETCC
/// operands. And we can even implement some other situations by transforming
/// them:
/// - We can implement (NEG SETCC) i.e. negating a single comparison by
/// negating the flags used in a CCMP/FCCMP operations.
/// - We can negate the result of a whole chain of CMP/CCMP/FCCMP operations
/// by negating the flags we test for afterwards. i.e.
/// NEG (CMP CCMP CCCMP ...) can be implemented.
/// - Note that we can only ever negate all previously processed results.
/// What we can not implement by flipping the flags to test is a negation
/// of two sub-trees (because the negation affects all sub-trees emitted so
/// far, so the 2nd sub-tree we emit would also affect the first).
/// With those tools we can implement some OR operations:
/// - (OR (SETCC A) (SETCC B)) can be implemented via:
/// NEG (AND (NEG (SETCC A)) (NEG (SETCC B)))
/// - After transforming OR to NEG/AND combinations we may be able to use NEG
/// elimination rules from earlier to implement the whole thing as a
/// CCMP/FCCMP chain.
///
/// As complete example:
/// or (or (setCA (cmp A)) (setCB (cmp B)))
/// (and (setCC (cmp C)) (setCD (cmp D)))"
/// can be reassociated to:
/// or (and (setCC (cmp C)) setCD (cmp D))
// (or (setCA (cmp A)) (setCB (cmp B)))
/// can be transformed to:
/// not (and (not (and (setCC (cmp C)) (setCD (cmp D))))
/// (and (not (setCA (cmp A)) (not (setCB (cmp B))))))"
/// which can be implemented as:
/// cmp C
/// ccmp D, inv(CD), CC
/// ccmp A, CA, inv(CD)
/// ccmp B, CB, inv(CA)
/// check for CB flags
///
/// A counterexample is "or (and A B) (and C D)" which translates to
/// not (and (not (and (not A) (not B))) (not (and (not C) (not D)))), we
/// can only implement 1 of the inner (not) operations, but not both!
/// @{
/// Create a conditional comparison; Use CCMP, CCMN or FCCMP as appropriate.
static SDValue emitConditionalComparison(SDValue LHS, SDValue RHS,
ISD::CondCode CC, SDValue CCOp,
AArch64CC::CondCode Predicate,
AArch64CC::CondCode OutCC,
const SDLoc &DL, SelectionDAG &DAG) {
unsigned Opcode = 0;
const bool FullFP16 = DAG.getSubtarget<AArch64Subtarget>().hasFullFP16();
if (LHS.getValueType().isFloatingPoint()) {
assert(LHS.getValueType() != MVT::f128);
if ((LHS.getValueType() == MVT::f16 && !FullFP16) ||
LHS.getValueType() == MVT::bf16) {
LHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, LHS);
RHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, RHS);
}
Opcode = AArch64ISD::FCCMP;
} else if (ConstantSDNode *Const = dyn_cast<ConstantSDNode>(RHS)) {
APInt Imm = Const->getAPIntValue();
if (Imm.isNegative() && Imm.sgt(-32)) {
Opcode = AArch64ISD::CCMN;
RHS = DAG.getConstant(Imm.abs(), DL, Const->getValueType(0));
}
} else if (isCMN(RHS, CC, DAG)) {
Opcode = AArch64ISD::CCMN;
RHS = RHS.getOperand(1);
} else if (LHS.getOpcode() == ISD::SUB && isNullConstant(LHS.getOperand(0)) &&
isIntEqualitySetCC(CC)) {
// As we are looking for EQ/NE compares, the operands can be commuted ; can
// we combine a (CCMP (sub 0, op1), op2) into a CCMN instruction ?
Opcode = AArch64ISD::CCMN;
LHS = LHS.getOperand(1);
}
if (Opcode == 0)
Opcode = AArch64ISD::CCMP;
SDValue Condition = DAG.getConstant(Predicate, DL, MVT_CC);
AArch64CC::CondCode InvOutCC = AArch64CC::getInvertedCondCode(OutCC);
unsigned NZCV = AArch64CC::getNZCVToSatisfyCondCode(InvOutCC);
SDValue NZCVOp = DAG.getConstant(NZCV, DL, MVT::i32);
return DAG.getNode(Opcode, DL, MVT_CC, LHS, RHS, NZCVOp, Condition, CCOp);
}
/// Returns true if @p Val is a tree of AND/OR/SETCC operations that can be
/// expressed as a conjunction. See \ref AArch64CCMP.
/// \param CanNegate Set to true if we can negate the whole sub-tree just by
/// changing the conditions on the SETCC tests.
/// (this means we can call emitConjunctionRec() with
/// Negate==true on this sub-tree)
/// \param MustBeFirst Set to true if this subtree needs to be negated and we
/// cannot do the negation naturally. We are required to
/// emit the subtree first in this case.
/// \param WillNegate Is true if are called when the result of this
/// subexpression must be negated. This happens when the
/// outer expression is an OR. We can use this fact to know
/// that we have a double negation (or (or ...) ...) that
/// can be implemented for free.
static bool canEmitConjunction(const SDValue Val, bool &CanNegate,
bool &MustBeFirst, bool WillNegate,
unsigned Depth = 0) {
if (!Val.hasOneUse())
return false;
unsigned Opcode = Val->getOpcode();
if (Opcode == ISD::SETCC) {
if (Val->getOperand(0).getValueType() == MVT::f128)
return false;
CanNegate = true;
MustBeFirst = false;
return true;
}
// Protect against exponential runtime and stack overflow.
if (Depth > 6)
return false;
if (Opcode == ISD::AND || Opcode == ISD::OR) {
bool IsOR = Opcode == ISD::OR;
SDValue O0 = Val->getOperand(0);
SDValue O1 = Val->getOperand(1);
bool CanNegateL;
bool MustBeFirstL;
if (!canEmitConjunction(O0, CanNegateL, MustBeFirstL, IsOR, Depth+1))
return false;
bool CanNegateR;
bool MustBeFirstR;
if (!canEmitConjunction(O1, CanNegateR, MustBeFirstR, IsOR, Depth+1))
return false;
if (MustBeFirstL && MustBeFirstR)
return false;
if (IsOR) {
// For an OR expression we need to be able to naturally negate at least
// one side or we cannot do the transformation at all.
if (!CanNegateL && !CanNegateR)
return false;
// If we the result of the OR will be negated and we can naturally negate
// the leafs, then this sub-tree as a whole negates naturally.
CanNegate = WillNegate && CanNegateL && CanNegateR;
// If we cannot naturally negate the whole sub-tree, then this must be
// emitted first.
MustBeFirst = !CanNegate;
} else {
assert(Opcode == ISD::AND && "Must be OR or AND");
// We cannot naturally negate an AND operation.
CanNegate = false;
MustBeFirst = MustBeFirstL || MustBeFirstR;
}
return true;
}
return false;
}
/// Emit conjunction or disjunction tree with the CMP/FCMP followed by a chain
/// of CCMP/CFCMP ops. See @ref AArch64CCMP.
/// Tries to transform the given i1 producing node @p Val to a series compare
/// and conditional compare operations. @returns an NZCV flags producing node
/// and sets @p OutCC to the flags that should be tested or returns SDValue() if
/// transformation was not possible.
/// \p Negate is true if we want this sub-tree being negated just by changing
/// SETCC conditions.
static SDValue emitConjunctionRec(SelectionDAG &DAG, SDValue Val,
AArch64CC::CondCode &OutCC, bool Negate, SDValue CCOp,
AArch64CC::CondCode Predicate) {
// We're at a tree leaf, produce a conditional comparison operation.
unsigned Opcode = Val->getOpcode();
if (Opcode == ISD::SETCC) {
SDValue LHS = Val->getOperand(0);
SDValue RHS = Val->getOperand(1);
ISD::CondCode CC = cast<CondCodeSDNode>(Val->getOperand(2))->get();
bool isInteger = LHS.getValueType().isInteger();
if (Negate)
CC = getSetCCInverse(CC, LHS.getValueType());
SDLoc DL(Val);
// Determine OutCC and handle FP special case.
if (isInteger) {
OutCC = changeIntCCToAArch64CC(CC);
} else {
assert(LHS.getValueType().isFloatingPoint());
AArch64CC::CondCode ExtraCC;
changeFPCCToANDAArch64CC(CC, OutCC, ExtraCC);
// Some floating point conditions can't be tested with a single condition
// code. Construct an additional comparison in this case.
if (ExtraCC != AArch64CC::AL) {
SDValue ExtraCmp;
if (!CCOp.getNode())
ExtraCmp = emitComparison(LHS, RHS, CC, DL, DAG);
else
ExtraCmp = emitConditionalComparison(LHS, RHS, CC, CCOp, Predicate,
ExtraCC, DL, DAG);
CCOp = ExtraCmp;
Predicate = ExtraCC;
}
}
// Produce a normal comparison if we are first in the chain
if (!CCOp)
return emitComparison(LHS, RHS, CC, DL, DAG);
// Otherwise produce a ccmp.
return emitConditionalComparison(LHS, RHS, CC, CCOp, Predicate, OutCC, DL,
DAG);
}
assert(Val->hasOneUse() && "Valid conjunction/disjunction tree");
bool IsOR = Opcode == ISD::OR;
SDValue LHS = Val->getOperand(0);
bool CanNegateL;
bool MustBeFirstL;
bool ValidL = canEmitConjunction(LHS, CanNegateL, MustBeFirstL, IsOR);
assert(ValidL && "Valid conjunction/disjunction tree");
(void)ValidL;
SDValue RHS = Val->getOperand(1);
bool CanNegateR;
bool MustBeFirstR;
bool ValidR = canEmitConjunction(RHS, CanNegateR, MustBeFirstR, IsOR);
assert(ValidR && "Valid conjunction/disjunction tree");
(void)ValidR;
// Swap sub-tree that must come first to the right side.
if (MustBeFirstL) {
assert(!MustBeFirstR && "Valid conjunction/disjunction tree");
std::swap(LHS, RHS);
std::swap(CanNegateL, CanNegateR);
std::swap(MustBeFirstL, MustBeFirstR);
}
bool NegateR;
bool NegateAfterR;
bool NegateL;
bool NegateAfterAll;
if (Opcode == ISD::OR) {
// Swap the sub-tree that we can negate naturally to the left.
if (!CanNegateL) {
assert(CanNegateR && "at least one side must be negatable");
assert(!MustBeFirstR && "invalid conjunction/disjunction tree");
assert(!Negate);
std::swap(LHS, RHS);
NegateR = false;
NegateAfterR = true;
} else {
// Negate the left sub-tree if possible, otherwise negate the result.
NegateR = CanNegateR;
NegateAfterR = !CanNegateR;
}
NegateL = true;
NegateAfterAll = !Negate;
} else {
assert(Opcode == ISD::AND && "Valid conjunction/disjunction tree");
assert(!Negate && "Valid conjunction/disjunction tree");
NegateL = false;
NegateR = false;
NegateAfterR = false;
NegateAfterAll = false;
}
// Emit sub-trees.
AArch64CC::CondCode RHSCC;
SDValue CmpR = emitConjunctionRec(DAG, RHS, RHSCC, NegateR, CCOp, Predicate);
if (NegateAfterR)
RHSCC = AArch64CC::getInvertedCondCode(RHSCC);
SDValue CmpL = emitConjunctionRec(DAG, LHS, OutCC, NegateL, CmpR, RHSCC);
if (NegateAfterAll)
OutCC = AArch64CC::getInvertedCondCode(OutCC);
return CmpL;
}
/// Emit expression as a conjunction (a series of CCMP/CFCMP ops).
/// In some cases this is even possible with OR operations in the expression.
/// See \ref AArch64CCMP.
/// \see emitConjunctionRec().
static SDValue emitConjunction(SelectionDAG &DAG, SDValue Val,
AArch64CC::CondCode &OutCC) {
bool DummyCanNegate;
bool DummyMustBeFirst;
if (!canEmitConjunction(Val, DummyCanNegate, DummyMustBeFirst, false))
return SDValue();
return emitConjunctionRec(DAG, Val, OutCC, false, SDValue(), AArch64CC::AL);
}
/// @}
/// Returns how profitable it is to fold a comparison's operand's shift and/or
/// extension operations.
static unsigned getCmpOperandFoldingProfit(SDValue Op) {
auto isSupportedExtend = [&](SDValue V) {
if (V.getOpcode() == ISD::SIGN_EXTEND_INREG)
return true;
if (V.getOpcode() == ISD::AND)
if (ConstantSDNode *MaskCst = dyn_cast<ConstantSDNode>(V.getOperand(1))) {
uint64_t Mask = MaskCst->getZExtValue();
return (Mask == 0xFF || Mask == 0xFFFF || Mask == 0xFFFFFFFF);
}
return false;
};
if (!Op.hasOneUse())
return 0;
if (isSupportedExtend(Op))
return 1;
unsigned Opc = Op.getOpcode();
if (Opc == ISD::SHL || Opc == ISD::SRL || Opc == ISD::SRA)
if (ConstantSDNode *ShiftCst = dyn_cast<ConstantSDNode>(Op.getOperand(1))) {
uint64_t Shift = ShiftCst->getZExtValue();
if (isSupportedExtend(Op.getOperand(0)))
return (Shift <= 4) ? 2 : 1;
EVT VT = Op.getValueType();
if ((VT == MVT::i32 && Shift <= 31) || (VT == MVT::i64 && Shift <= 63))
return 1;
}
return 0;
}
// emitComparison() converts comparison with one or negative one to comparison
// with 0. Note that this only works for signed comparisons because of how ANDS
// works.
static bool shouldBeAdjustedToZero(SDValue LHS, APInt C, ISD::CondCode &CC) {
// Only works for ANDS and AND.
if (LHS.getOpcode() != ISD::AND && LHS.getOpcode() != AArch64ISD::ANDS)
return false;
if (C.isOne() && (CC == ISD::SETLT || CC == ISD::SETGE)) {
CC = (CC == ISD::SETLT) ? ISD::SETLE : ISD::SETGT;
return true;
}
if (C.isAllOnes() && (CC == ISD::SETLE || CC == ISD::SETGT)) {
CC = (CC == ISD::SETLE) ? ISD::SETLT : ISD::SETGE;
return true;
}
return false;
}
static SDValue getAArch64Cmp(SDValue LHS, SDValue RHS, ISD::CondCode CC,
SDValue &AArch64cc, SelectionDAG &DAG,
const SDLoc &DL) {
if (ConstantSDNode *RHSC = dyn_cast<ConstantSDNode>(RHS.getNode())) {
EVT VT = RHS.getValueType();
APInt C = RHSC->getAPIntValue();
// shouldBeAdjustedToZero is a special case to better fold with
// emitComparison().
if (shouldBeAdjustedToZero(LHS, C, CC)) {
// Adjust the constant to zero.
// CC has already been adjusted.
RHS = DAG.getConstant(0, DL, VT);
} else if (!isLegalCmpImmed(C)) {
// Constant does not fit, try adjusting it by one?
switch (CC) {
default:
break;
case ISD::SETLT:
case ISD::SETGE:
if (!C.isMinSignedValue()) {
APInt CMinusOne = C - 1;
if (isLegalCmpImmed(CMinusOne)) {
CC = (CC == ISD::SETLT) ? ISD::SETLE : ISD::SETGT;
RHS = DAG.getConstant(CMinusOne, DL, VT);
}
}
break;
case ISD::SETULT:
case ISD::SETUGE:
if (!C.isZero()) {
APInt CMinusOne = C - 1;
if (isLegalCmpImmed(CMinusOne)) {
CC = (CC == ISD::SETULT) ? ISD::SETULE : ISD::SETUGT;
RHS = DAG.getConstant(CMinusOne, DL, VT);
}
}
break;
case ISD::SETLE:
case ISD::SETGT:
if (!C.isMaxSignedValue()) {
APInt CPlusOne = C + 1;
if (isLegalCmpImmed(CPlusOne)) {
CC = (CC == ISD::SETLE) ? ISD::SETLT : ISD::SETGE;
RHS = DAG.getConstant(CPlusOne, DL, VT);
}
}
break;
case ISD::SETULE:
case ISD::SETUGT:
if (!C.isAllOnes()) {
APInt CPlusOne = C + 1;
if (isLegalCmpImmed(CPlusOne)) {
CC = (CC == ISD::SETULE) ? ISD::SETULT : ISD::SETUGE;
RHS = DAG.getConstant(CPlusOne, DL, VT);
}
}
break;
}
}
}
// Comparisons are canonicalized so that the RHS operand is simpler than the
// LHS one, the extreme case being when RHS is an immediate. However, AArch64
// can fold some shift+extend operations on the RHS operand, so swap the
// operands if that can be done.
//
// For example:
// lsl w13, w11, #1
// cmp w13, w12
// can be turned into:
// cmp w12, w11, lsl #1
if (!isa<ConstantSDNode>(RHS) || !isLegalCmpImmed(RHS->getAsAPIntVal())) {
bool LHSIsCMN = isCMN(LHS, CC, DAG);
bool RHSIsCMN = isCMN(RHS, CC, DAG);
SDValue TheLHS = LHSIsCMN ? LHS.getOperand(1) : LHS;
SDValue TheRHS = RHSIsCMN ? RHS.getOperand(1) : RHS;
if (getCmpOperandFoldingProfit(TheLHS) + (LHSIsCMN ? 1 : 0) >
getCmpOperandFoldingProfit(TheRHS) + (RHSIsCMN ? 1 : 0)) {
std::swap(LHS, RHS);
CC = ISD::getSetCCSwappedOperands(CC);
}
}
SDValue Cmp;
AArch64CC::CondCode AArch64CC;
if (isIntEqualitySetCC(CC) && isa<ConstantSDNode>(RHS)) {
const ConstantSDNode *RHSC = cast<ConstantSDNode>(RHS);
// The imm operand of ADDS is an unsigned immediate, in the range 0 to 4095.
// For the i8 operand, the largest immediate is 255, so this can be easily
// encoded in the compare instruction. For the i16 operand, however, the
// largest immediate cannot be encoded in the compare.
// Therefore, use a sign extending load and cmn to avoid materializing the
// -1 constant. For example,
// movz w1, #65535
// ldrh w0, [x0, #0]
// cmp w0, w1
// >
// ldrsh w0, [x0, #0]
// cmn w0, #1
// Fundamental, we're relying on the property that (zext LHS) == (zext RHS)
// if and only if (sext LHS) == (sext RHS). The checks are in place to
// ensure both the LHS and RHS are truly zero extended and to make sure the
// transformation is profitable.
if ((RHSC->getZExtValue() >> 16 == 0) && isa<LoadSDNode>(LHS) &&
cast<LoadSDNode>(LHS)->getExtensionType() == ISD::ZEXTLOAD &&
cast<LoadSDNode>(LHS)->getMemoryVT() == MVT::i16 &&
LHS.getNode()->hasNUsesOfValue(1, 0)) {
int16_t ValueofRHS = RHS->getAsZExtVal();
if (ValueofRHS < 0 && isLegalArithImmed(-ValueofRHS)) {
SDValue SExt =
DAG.getNode(ISD::SIGN_EXTEND_INREG, DL, LHS.getValueType(), LHS,
DAG.getValueType(MVT::i16));
Cmp = emitComparison(
SExt, DAG.getSignedConstant(ValueofRHS, DL, RHS.getValueType()), CC,
DL, DAG);
AArch64CC = changeIntCCToAArch64CC(CC);
}
}
if (!Cmp && (RHSC->isZero() || RHSC->isOne())) {
if ((Cmp = emitConjunction(DAG, LHS, AArch64CC))) {
if ((CC == ISD::SETNE) ^ RHSC->isZero())
AArch64CC = AArch64CC::getInvertedCondCode(AArch64CC);
}
}
}
if (!Cmp) {
Cmp = emitComparison(LHS, RHS, CC, DL, DAG);
AArch64CC = changeIntCCToAArch64CC(CC);
}
AArch64cc = DAG.getConstant(AArch64CC, DL, MVT_CC);
return Cmp;
}
static std::pair<SDValue, SDValue>
getAArch64XALUOOp(AArch64CC::CondCode &CC, SDValue Op, SelectionDAG &DAG) {
assert((Op.getValueType() == MVT::i32 || Op.getValueType() == MVT::i64) &&
"Unsupported value type");
SDValue Value, Overflow;
SDLoc DL(Op);
SDValue LHS = Op.getOperand(0);
SDValue RHS = Op.getOperand(1);
unsigned Opc = 0;
switch (Op.getOpcode()) {
default:
llvm_unreachable("Unknown overflow instruction!");
case ISD::SADDO:
Opc = AArch64ISD::ADDS;
CC = AArch64CC::VS;
break;
case ISD::UADDO:
Opc = AArch64ISD::ADDS;
CC = AArch64CC::HS;
break;
case ISD::SSUBO:
Opc = AArch64ISD::SUBS;
CC = AArch64CC::VS;
break;
case ISD::USUBO:
Opc = AArch64ISD::SUBS;
CC = AArch64CC::LO;
break;
// Multiply needs a little bit extra work.
case ISD::SMULO:
case ISD::UMULO: {
CC = AArch64CC::NE;
bool IsSigned = Op.getOpcode() == ISD::SMULO;
if (Op.getValueType() == MVT::i32) {
// Extend to 64-bits, then perform a 64-bit multiply.
unsigned ExtendOpc = IsSigned ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND;
LHS = DAG.getNode(ExtendOpc, DL, MVT::i64, LHS);
RHS = DAG.getNode(ExtendOpc, DL, MVT::i64, RHS);
SDValue Mul = DAG.getNode(ISD::MUL, DL, MVT::i64, LHS, RHS);
Value = DAG.getNode(ISD::TRUNCATE, DL, MVT::i32, Mul);
// Check that the result fits into a 32-bit integer.
SDVTList VTs = DAG.getVTList(MVT::i64, MVT_CC);
if (IsSigned) {
// cmp xreg, wreg, sxtw
SDValue SExtMul = DAG.getNode(ISD::SIGN_EXTEND, DL, MVT::i64, Value);
Overflow =
DAG.getNode(AArch64ISD::SUBS, DL, VTs, Mul, SExtMul).getValue(1);
} else {
// tst xreg, #0xffffffff00000000
SDValue UpperBits = DAG.getConstant(0xFFFFFFFF00000000, DL, MVT::i64);
Overflow =
DAG.getNode(AArch64ISD::ANDS, DL, VTs, Mul, UpperBits).getValue(1);
}
break;
}
assert(Op.getValueType() == MVT::i64 && "Expected an i64 value type");
// For the 64 bit multiply
Value = DAG.getNode(ISD::MUL, DL, MVT::i64, LHS, RHS);
if (IsSigned) {
SDValue UpperBits = DAG.getNode(ISD::MULHS, DL, MVT::i64, LHS, RHS);
SDValue LowerBits = DAG.getNode(ISD::SRA, DL, MVT::i64, Value,
DAG.getConstant(63, DL, MVT::i64));
// It is important that LowerBits is last, otherwise the arithmetic
// shift will not be folded into the compare (SUBS).
SDVTList VTs = DAG.getVTList(MVT::i64, MVT::i32);
Overflow = DAG.getNode(AArch64ISD::SUBS, DL, VTs, UpperBits, LowerBits)
.getValue(1);
} else {
SDValue UpperBits = DAG.getNode(ISD::MULHU, DL, MVT::i64, LHS, RHS);
SDVTList VTs = DAG.getVTList(MVT::i64, MVT::i32);
Overflow =
DAG.getNode(AArch64ISD::SUBS, DL, VTs,
DAG.getConstant(0, DL, MVT::i64),
UpperBits).getValue(1);
}
break;
}
} // switch (...)
if (Opc) {
SDVTList VTs = DAG.getVTList(Op->getValueType(0), MVT::i32);
// Emit the AArch64 operation with overflow check.
Value = DAG.getNode(Opc, DL, VTs, LHS, RHS);
Overflow = Value.getValue(1);
}
return std::make_pair(Value, Overflow);
}
SDValue AArch64TargetLowering::LowerXOR(SDValue Op, SelectionDAG &DAG) const {
if (useSVEForFixedLengthVectorVT(Op.getValueType(),
!Subtarget->isNeonAvailable()))
return LowerToScalableOp(Op, DAG);
SDValue Sel = Op.getOperand(0);
SDValue Other = Op.getOperand(1);
SDLoc DL(Sel);
// If the operand is an overflow checking operation, invert the condition
// code and kill the Not operation. I.e., transform:
// (xor (overflow_op_bool, 1))
// -->
// (csel 1, 0, invert(cc), overflow_op_bool)
// ... which later gets transformed to just a cset instruction with an
// inverted condition code, rather than a cset + eor sequence.
if (isOneConstant(Other) && ISD::isOverflowIntrOpRes(Sel)) {
// Only lower legal XALUO ops.
if (!DAG.getTargetLoweringInfo().isTypeLegal(Sel->getValueType(0)))
return SDValue();
SDValue TVal = DAG.getConstant(1, DL, MVT::i32);
SDValue FVal = DAG.getConstant(0, DL, MVT::i32);
AArch64CC::CondCode CC;
SDValue Value, Overflow;
std::tie(Value, Overflow) = getAArch64XALUOOp(CC, Sel.getValue(0), DAG);
SDValue CCVal = DAG.getConstant(getInvertedCondCode(CC), DL, MVT::i32);
return DAG.getNode(AArch64ISD::CSEL, DL, Op.getValueType(), TVal, FVal,
CCVal, Overflow);
}
// If neither operand is a SELECT_CC, give up.
if (Sel.getOpcode() != ISD::SELECT_CC)
std::swap(Sel, Other);
if (Sel.getOpcode() != ISD::SELECT_CC)
return Op;
// The folding we want to perform is:
// (xor x, (select_cc a, b, cc, 0, -1) )
// -->
// (csel x, (xor x, -1), cc ...)
//
// The latter will get matched to a CSINV instruction.
ISD::CondCode CC = cast<CondCodeSDNode>(Sel.getOperand(4))->get();
SDValue LHS = Sel.getOperand(0);
SDValue RHS = Sel.getOperand(1);
SDValue TVal = Sel.getOperand(2);
SDValue FVal = Sel.getOperand(3);
// FIXME: This could be generalized to non-integer comparisons.
if (LHS.getValueType() != MVT::i32 && LHS.getValueType() != MVT::i64)
return Op;
ConstantSDNode *CFVal = dyn_cast<ConstantSDNode>(FVal);
ConstantSDNode *CTVal = dyn_cast<ConstantSDNode>(TVal);
// The values aren't constants, this isn't the pattern we're looking for.
if (!CFVal || !CTVal)
return Op;
// We can commute the SELECT_CC by inverting the condition. This
// might be needed to make this fit into a CSINV pattern.
if (CTVal->isAllOnes() && CFVal->isZero()) {
std::swap(TVal, FVal);
std::swap(CTVal, CFVal);
CC = ISD::getSetCCInverse(CC, LHS.getValueType());
}
// If the constants line up, perform the transform!
if (CTVal->isZero() && CFVal->isAllOnes()) {
SDValue CCVal;
SDValue Cmp = getAArch64Cmp(LHS, RHS, CC, CCVal, DAG, DL);
FVal = Other;
TVal = DAG.getNode(ISD::XOR, DL, Other.getValueType(), Other,
DAG.getAllOnesConstant(DL, Other.getValueType()));
return DAG.getNode(AArch64ISD::CSEL, DL, Sel.getValueType(), FVal, TVal,
CCVal, Cmp);
}
return Op;
}
// If Invert is false, sets 'C' bit of NZCV to 0 if value is 0, else sets 'C'
// bit to 1. If Invert is true, sets 'C' bit of NZCV to 1 if value is 0, else
// sets 'C' bit to 0.
static SDValue valueToCarryFlag(SDValue Value, SelectionDAG &DAG, bool Invert) {
SDLoc DL(Value);
EVT VT = Value.getValueType();
SDValue Op0 = Invert ? DAG.getConstant(0, DL, VT) : Value;
SDValue Op1 = Invert ? Value : DAG.getConstant(1, DL, VT);
SDValue Cmp =
DAG.getNode(AArch64ISD::SUBS, DL, DAG.getVTList(VT, MVT::Glue), Op0, Op1);
return Cmp.getValue(1);
}
// If Invert is false, value is 1 if 'C' bit of NZCV is 1, else 0.
// If Invert is true, value is 0 if 'C' bit of NZCV is 1, else 1.
static SDValue carryFlagToValue(SDValue Glue, EVT VT, SelectionDAG &DAG,
bool Invert) {
assert(Glue.getResNo() == 1);
SDLoc DL(Glue);
SDValue Zero = DAG.getConstant(0, DL, VT);
SDValue One = DAG.getConstant(1, DL, VT);
unsigned Cond = Invert ? AArch64CC::LO : AArch64CC::HS;
SDValue CC = DAG.getConstant(Cond, DL, MVT::i32);
return DAG.getNode(AArch64ISD::CSEL, DL, VT, One, Zero, CC, Glue);
}
// Value is 1 if 'V' bit of NZCV is 1, else 0
static SDValue overflowFlagToValue(SDValue Glue, EVT VT, SelectionDAG &DAG) {
assert(Glue.getResNo() == 1);
SDLoc DL(Glue);
SDValue Zero = DAG.getConstant(0, DL, VT);
SDValue One = DAG.getConstant(1, DL, VT);
SDValue CC = DAG.getConstant(AArch64CC::VS, DL, MVT::i32);
return DAG.getNode(AArch64ISD::CSEL, DL, VT, One, Zero, CC, Glue);
}
// This lowering is inefficient, but it will get cleaned up by
// `foldOverflowCheck`
static SDValue lowerADDSUBO_CARRY(SDValue Op, SelectionDAG &DAG,
unsigned Opcode, bool IsSigned) {
EVT VT0 = Op.getValue(0).getValueType();
EVT VT1 = Op.getValue(1).getValueType();
if (VT0 != MVT::i32 && VT0 != MVT::i64)
return SDValue();
bool InvertCarry = Opcode == AArch64ISD::SBCS;
SDValue OpLHS = Op.getOperand(0);
SDValue OpRHS = Op.getOperand(1);
SDValue OpCarryIn = valueToCarryFlag(Op.getOperand(2), DAG, InvertCarry);
SDLoc DL(Op);
SDVTList VTs = DAG.getVTList(VT0, VT1);
SDValue Sum = DAG.getNode(Opcode, DL, DAG.getVTList(VT0, MVT::Glue), OpLHS,
OpRHS, OpCarryIn);
SDValue OutFlag =
IsSigned ? overflowFlagToValue(Sum.getValue(1), VT1, DAG)
: carryFlagToValue(Sum.getValue(1), VT1, DAG, InvertCarry);
return DAG.getNode(ISD::MERGE_VALUES, DL, VTs, Sum, OutFlag);
}
static SDValue LowerXALUO(SDValue Op, SelectionDAG &DAG) {
// Let legalize expand this if it isn't a legal type yet.
if (!DAG.getTargetLoweringInfo().isTypeLegal(Op.getValueType()))
return SDValue();
SDLoc DL(Op);
AArch64CC::CondCode CC;
// The actual operation that sets the overflow or carry flag.
SDValue Value, Overflow;
std::tie(Value, Overflow) = getAArch64XALUOOp(CC, Op, DAG);
// We use 0 and 1 as false and true values.
SDValue TVal = DAG.getConstant(1, DL, MVT::i32);
SDValue FVal = DAG.getConstant(0, DL, MVT::i32);
// We use an inverted condition, because the conditional select is inverted
// too. This will allow it to be selected to a single instruction:
// CSINC Wd, WZR, WZR, invert(cond).
SDValue CCVal = DAG.getConstant(getInvertedCondCode(CC), DL, MVT::i32);
Overflow =
DAG.getNode(AArch64ISD::CSEL, DL, MVT::i32, FVal, TVal, CCVal, Overflow);
SDVTList VTs = DAG.getVTList(Op.getValueType(), MVT::i32);
return DAG.getNode(ISD::MERGE_VALUES, DL, VTs, Value, Overflow);
}
// Prefetch operands are:
// 1: Address to prefetch
// 2: bool isWrite
// 3: int locality (0 = no locality ... 3 = extreme locality)
// 4: bool isDataCache
static SDValue LowerPREFETCH(SDValue Op, SelectionDAG &DAG) {
SDLoc DL(Op);
unsigned IsWrite = Op.getConstantOperandVal(2);
unsigned Locality = Op.getConstantOperandVal(3);
unsigned IsData = Op.getConstantOperandVal(4);
bool IsStream = !Locality;
// When the locality number is set
if (Locality) {
// The front-end should have filtered out the out-of-range values
assert(Locality <= 3 && "Prefetch locality out-of-range");
// The locality degree is the opposite of the cache speed.
// Put the number the other way around.
// The encoding starts at 0 for level 1
Locality = 3 - Locality;
}
// built the mask value encoding the expected behavior.
unsigned PrfOp = (IsWrite << 4) | // Load/Store bit
(!IsData << 3) | // IsDataCache bit
(Locality << 1) | // Cache level bits
(unsigned)IsStream; // Stream bit
return DAG.getNode(AArch64ISD::PREFETCH, DL, MVT::Other, Op.getOperand(0),
DAG.getTargetConstant(PrfOp, DL, MVT::i32),
Op.getOperand(1));
}
// Converts SETCC (AND X Y) Z ULT -> SETCC (AND X (Y & ~(Z - 1)) 0 EQ when Y is
// a power of 2. This is then lowered to ANDS X (Y & ~(Z - 1)) instead of SUBS
// (AND X Y) Z which produces a better opt with EmitComparison
static void simplifySetCCIntoEq(ISD::CondCode &CC, SDValue &LHS, SDValue &RHS,
SelectionDAG &DAG, const SDLoc DL) {
if (CC == ISD::SETULT && LHS.getOpcode() == ISD::AND && LHS->hasOneUse()) {
ConstantSDNode *LHSConstOp = dyn_cast<ConstantSDNode>(LHS.getOperand(1));
ConstantSDNode *RHSConst = dyn_cast<ConstantSDNode>(RHS);
if (LHSConstOp && RHSConst) {
uint64_t LHSConstValue = LHSConstOp->getZExtValue();
uint64_t RHSConstant = RHSConst->getZExtValue();
if (isPowerOf2_64(RHSConstant)) {
uint64_t NewMaskValue = LHSConstValue & ~(RHSConstant - 1);
LHS =
DAG.getNode(ISD::AND, DL, LHS.getValueType(), LHS.getOperand(0),
DAG.getConstant(NewMaskValue, DL, LHS.getValueType()));
RHS = DAG.getConstant(0, DL, RHS.getValueType());
CC = ISD::SETEQ;
}
}
}
}
SDValue AArch64TargetLowering::LowerFP_EXTEND(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
if (VT.isScalableVector()) {
SDValue SrcVal = Op.getOperand(0);
if (VT == MVT::nxv2f64 && SrcVal.getValueType() == MVT::nxv2bf16) {
// Break conversion in two with the first part converting to f32 and the
// second using native f32->VT instructions.
SDLoc DL(Op);
return DAG.getNode(ISD::FP_EXTEND, DL, VT,
DAG.getNode(ISD::FP_EXTEND, DL, MVT::nxv2f32, SrcVal));
}
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FP_EXTEND_MERGE_PASSTHRU);
}
if (useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()))
return LowerFixedLengthFPExtendToSVE(Op, DAG);
bool IsStrict = Op->isStrictFPOpcode();
SDValue Op0 = Op.getOperand(IsStrict ? 1 : 0);
EVT Op0VT = Op0.getValueType();
if (VT == MVT::f64) {
// FP16->FP32 extends are legal for v32 and v4f32.
if (Op0VT == MVT::f32 || Op0VT == MVT::f16)
return Op;
// Split bf16->f64 extends into two fpextends.
if (Op0VT == MVT::bf16 && IsStrict) {
SDValue Ext1 =
DAG.getNode(ISD::STRICT_FP_EXTEND, SDLoc(Op), {MVT::f32, MVT::Other},
{Op0, Op.getOperand(0)});
return DAG.getNode(ISD::STRICT_FP_EXTEND, SDLoc(Op), {VT, MVT::Other},
{Ext1, Ext1.getValue(1)});
}
if (Op0VT == MVT::bf16)
return DAG.getNode(ISD::FP_EXTEND, SDLoc(Op), VT,
DAG.getNode(ISD::FP_EXTEND, SDLoc(Op), MVT::f32, Op0));
return SDValue();
}
assert(Op.getValueType() == MVT::f128 && "Unexpected lowering");
return SDValue();
}
SDValue AArch64TargetLowering::LowerFP_ROUND(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
bool IsStrict = Op->isStrictFPOpcode();
SDValue SrcVal = Op.getOperand(IsStrict ? 1 : 0);
EVT SrcVT = SrcVal.getValueType();
bool Trunc = Op.getConstantOperandVal(IsStrict ? 2 : 1) == 1;
if (VT.isScalableVector()) {
// Let common code split the operation.
if (SrcVT == MVT::nxv8f32)
return Op;
if (VT.getScalarType() != MVT::bf16)
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FP_ROUND_MERGE_PASSTHRU);
SDLoc DL(Op);
constexpr EVT I32 = MVT::nxv4i32;
auto ImmV = [&](int I) -> SDValue { return DAG.getConstant(I, DL, I32); };
SDValue NaN;
SDValue Narrow;
if (SrcVT == MVT::nxv2f32 || SrcVT == MVT::nxv4f32) {
if (Subtarget->hasBF16())
return LowerToPredicatedOp(Op, DAG,
AArch64ISD::FP_ROUND_MERGE_PASSTHRU);
Narrow = getSVESafeBitCast(I32, SrcVal, DAG);
// Set the quiet bit.
if (!DAG.isKnownNeverSNaN(SrcVal))
NaN = DAG.getNode(ISD::OR, DL, I32, Narrow, ImmV(0x400000));
} else if (SrcVT == MVT::nxv2f64 &&
(Subtarget->hasSVE2() || Subtarget->isStreamingSVEAvailable())) {
// Round to float without introducing rounding errors and try again.
SDValue Pg = getPredicateForVector(DAG, DL, MVT::nxv2f32);
Narrow = DAG.getNode(AArch64ISD::FCVTX_MERGE_PASSTHRU, DL, MVT::nxv2f32,
Pg, SrcVal, DAG.getUNDEF(MVT::nxv2f32));
SmallVector<SDValue, 3> NewOps;
if (IsStrict)
NewOps.push_back(Op.getOperand(0));
NewOps.push_back(Narrow);
NewOps.push_back(Op.getOperand(IsStrict ? 2 : 1));
return DAG.getNode(Op.getOpcode(), DL, VT, NewOps, Op->getFlags());
} else
return SDValue();
if (!Trunc) {
SDValue Lsb = DAG.getNode(ISD::SRL, DL, I32, Narrow, ImmV(16));
Lsb = DAG.getNode(ISD::AND, DL, I32, Lsb, ImmV(1));
SDValue RoundingBias = DAG.getNode(ISD::ADD, DL, I32, Lsb, ImmV(0x7fff));
Narrow = DAG.getNode(ISD::ADD, DL, I32, Narrow, RoundingBias);
}
// Don't round if we had a NaN, we don't want to turn 0x7fffffff into
// 0x80000000.
if (NaN) {
EVT I1 = I32.changeElementType(MVT::i1);
EVT CondVT = VT.changeElementType(MVT::i1);
SDValue IsNaN = DAG.getSetCC(DL, CondVT, SrcVal, SrcVal, ISD::SETUO);
IsNaN = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, I1, IsNaN);
Narrow = DAG.getSelect(DL, I32, IsNaN, NaN, Narrow);
}
// Now that we have rounded, shift the bits into position.
Narrow = DAG.getNode(ISD::SRL, DL, I32, Narrow, ImmV(16));
return getSVESafeBitCast(VT, Narrow, DAG);
}
if (useSVEForFixedLengthVectorVT(SrcVT, !Subtarget->isNeonAvailable()))
return LowerFixedLengthFPRoundToSVE(Op, DAG);
// Expand cases where the result type is BF16 but we don't have hardware
// instructions to lower it.
if (VT.getScalarType() == MVT::bf16 &&
!((Subtarget->hasNEON() || Subtarget->hasSME()) &&
Subtarget->hasBF16())) {
SDLoc DL(Op);
SDValue Narrow = SrcVal;
SDValue NaN;
EVT I32 = SrcVT.changeElementType(MVT::i32);
EVT F32 = SrcVT.changeElementType(MVT::f32);
if (SrcVT.getScalarType() == MVT::f32) {
bool NeverSNaN = DAG.isKnownNeverSNaN(Narrow);
Narrow = DAG.getNode(ISD::BITCAST, DL, I32, Narrow);
if (!NeverSNaN) {
// Set the quiet bit.
NaN = DAG.getNode(ISD::OR, DL, I32, Narrow,
DAG.getConstant(0x400000, DL, I32));
}
} else if (SrcVT.getScalarType() == MVT::f64) {
Narrow = DAG.getNode(AArch64ISD::FCVTXN, DL, F32, Narrow);
Narrow = DAG.getNode(ISD::BITCAST, DL, I32, Narrow);
} else {
return SDValue();
}
if (!Trunc) {
SDValue One = DAG.getConstant(1, DL, I32);
SDValue Lsb = DAG.getNode(ISD::SRL, DL, I32, Narrow,
DAG.getShiftAmountConstant(16, I32, DL));
Lsb = DAG.getNode(ISD::AND, DL, I32, Lsb, One);
SDValue RoundingBias =
DAG.getNode(ISD::ADD, DL, I32, DAG.getConstant(0x7fff, DL, I32), Lsb);
Narrow = DAG.getNode(ISD::ADD, DL, I32, Narrow, RoundingBias);
}
// Don't round if we had a NaN, we don't want to turn 0x7fffffff into
// 0x80000000.
if (NaN) {
SDValue IsNaN = DAG.getSetCC(
DL, getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(), SrcVT),
SrcVal, SrcVal, ISD::SETUO);
Narrow = DAG.getSelect(DL, I32, IsNaN, NaN, Narrow);
}
// Now that we have rounded, shift the bits into position.
Narrow = DAG.getNode(ISD::SRL, DL, I32, Narrow,
DAG.getShiftAmountConstant(16, I32, DL));
if (VT.isVector()) {
EVT I16 = I32.changeVectorElementType(MVT::i16);
Narrow = DAG.getNode(ISD::TRUNCATE, DL, I16, Narrow);
return DAG.getNode(ISD::BITCAST, DL, VT, Narrow);
}
Narrow = DAG.getNode(ISD::BITCAST, DL, F32, Narrow);
SDValue Result = DAG.getTargetExtractSubreg(AArch64::hsub, DL, VT, Narrow);
return IsStrict ? DAG.getMergeValues({Result, Op.getOperand(0)}, DL)
: Result;
}
if (SrcVT != MVT::f128) {
// Expand cases where the input is a vector bigger than NEON.
if (useSVEForFixedLengthVectorVT(SrcVT))
return SDValue();
// It's legal except when f128 is involved
return Op;
}
return SDValue();
}
SDValue AArch64TargetLowering::LowerVectorFP_TO_INT(SDValue Op,
SelectionDAG &DAG) const {
// Warning: We maintain cost tables in AArch64TargetTransformInfo.cpp.
// Any additional optimization in this function should be recorded
// in the cost tables.
bool IsStrict = Op->isStrictFPOpcode();
EVT InVT = Op.getOperand(IsStrict ? 1 : 0).getValueType();
EVT VT = Op.getValueType();
assert(!(IsStrict && VT.isScalableVector()) &&
"Unimplemented SVE support for STRICT_FP_to_INT!");
// f16 conversions are promoted to f32 when full fp16 is not supported.
if ((InVT.getVectorElementType() == MVT::f16 && !Subtarget->hasFullFP16()) ||
InVT.getVectorElementType() == MVT::bf16) {
EVT NewVT = VT.changeElementType(MVT::f32);
SDLoc DL(Op);
if (IsStrict) {
SDValue Ext = DAG.getNode(ISD::STRICT_FP_EXTEND, DL, {NewVT, MVT::Other},
{Op.getOperand(0), Op.getOperand(1)});
return DAG.getNode(Op.getOpcode(), DL, {VT, MVT::Other},
{Ext.getValue(1), Ext.getValue(0)});
}
return DAG.getNode(
Op.getOpcode(), DL, Op.getValueType(),
DAG.getNode(ISD::FP_EXTEND, DL, NewVT, Op.getOperand(0)));
}
if (VT.isScalableVector()) {
if (VT.getVectorElementType() == MVT::i1) {
SDLoc DL(Op);
EVT CvtVT = getPromotedVTForPredicate(VT);
SDValue Cvt = DAG.getNode(Op.getOpcode(), DL, CvtVT, Op.getOperand(0));
SDValue Zero = DAG.getConstant(0, DL, CvtVT);
return DAG.getSetCC(DL, VT, Cvt, Zero, ISD::SETNE);
}
// Let common code split the operation.
if (InVT == MVT::nxv8f32)
return Op;
unsigned Opcode = Op.getOpcode() == ISD::FP_TO_UINT
? AArch64ISD::FCVTZU_MERGE_PASSTHRU
: AArch64ISD::FCVTZS_MERGE_PASSTHRU;
return LowerToPredicatedOp(Op, DAG, Opcode);
}
if (useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()) ||
useSVEForFixedLengthVectorVT(InVT, !Subtarget->isNeonAvailable()))
return LowerFixedLengthFPToIntToSVE(Op, DAG);
uint64_t VTSize = VT.getFixedSizeInBits();
uint64_t InVTSize = InVT.getFixedSizeInBits();
if (VTSize < InVTSize) {
SDLoc DL(Op);
if (IsStrict) {
InVT = InVT.changeVectorElementTypeToInteger();
SDValue Cv = DAG.getNode(Op.getOpcode(), DL, {InVT, MVT::Other},
{Op.getOperand(0), Op.getOperand(1)});
SDValue Trunc = DAG.getNode(ISD::TRUNCATE, DL, VT, Cv);
return DAG.getMergeValues({Trunc, Cv.getValue(1)}, DL);
}
SDValue Cv =
DAG.getNode(Op.getOpcode(), DL, InVT.changeVectorElementTypeToInteger(),
Op.getOperand(0));
return DAG.getNode(ISD::TRUNCATE, DL, VT, Cv);
}
if (VTSize > InVTSize) {
SDLoc DL(Op);
MVT ExtVT =
MVT::getVectorVT(MVT::getFloatingPointVT(VT.getScalarSizeInBits()),
VT.getVectorNumElements());
if (IsStrict) {
SDValue Ext = DAG.getNode(ISD::STRICT_FP_EXTEND, DL, {ExtVT, MVT::Other},
{Op.getOperand(0), Op.getOperand(1)});
return DAG.getNode(Op.getOpcode(), DL, {VT, MVT::Other},
{Ext.getValue(1), Ext.getValue(0)});
}
SDValue Ext = DAG.getNode(ISD::FP_EXTEND, DL, ExtVT, Op.getOperand(0));
return DAG.getNode(Op.getOpcode(), DL, VT, Ext);
}
// Use a scalar operation for conversions between single-element vectors of
// the same size.
if (InVT.getVectorNumElements() == 1) {
SDLoc DL(Op);
SDValue Extract = DAG.getNode(
ISD::EXTRACT_VECTOR_ELT, DL, InVT.getScalarType(),
Op.getOperand(IsStrict ? 1 : 0), DAG.getConstant(0, DL, MVT::i64));
EVT ScalarVT = VT.getScalarType();
if (IsStrict)
return DAG.getNode(Op.getOpcode(), DL, {ScalarVT, MVT::Other},
{Op.getOperand(0), Extract});
return DAG.getNode(Op.getOpcode(), DL, ScalarVT, Extract);
}
// Type changing conversions are illegal.
return Op;
}
SDValue AArch64TargetLowering::LowerFP_TO_INT(SDValue Op,
SelectionDAG &DAG) const {
bool IsStrict = Op->isStrictFPOpcode();
SDValue SrcVal = Op.getOperand(IsStrict ? 1 : 0);
if (SrcVal.getValueType().isVector())
return LowerVectorFP_TO_INT(Op, DAG);
// f16 conversions are promoted to f32 when full fp16 is not supported.
if ((SrcVal.getValueType() == MVT::f16 && !Subtarget->hasFullFP16()) ||
SrcVal.getValueType() == MVT::bf16) {
SDLoc DL(Op);
if (IsStrict) {
SDValue Ext =
DAG.getNode(ISD::STRICT_FP_EXTEND, DL, {MVT::f32, MVT::Other},
{Op.getOperand(0), SrcVal});
return DAG.getNode(Op.getOpcode(), DL, {Op.getValueType(), MVT::Other},
{Ext.getValue(1), Ext.getValue(0)});
}
return DAG.getNode(Op.getOpcode(), DL, Op.getValueType(),
DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, SrcVal));
}
if (SrcVal.getValueType() != MVT::f128) {
// It's legal except when f128 is involved
return Op;
}
return SDValue();
}
SDValue
AArch64TargetLowering::LowerVectorFP_TO_INT_SAT(SDValue Op,
SelectionDAG &DAG) const {
// AArch64 FP-to-int conversions saturate to the destination element size, so
// we can lower common saturating conversions to simple instructions.
SDValue SrcVal = Op.getOperand(0);
EVT SrcVT = SrcVal.getValueType();
EVT DstVT = Op.getValueType();
EVT SatVT = cast<VTSDNode>(Op.getOperand(1))->getVT();
uint64_t SrcElementWidth = SrcVT.getScalarSizeInBits();
uint64_t DstElementWidth = DstVT.getScalarSizeInBits();
uint64_t SatWidth = SatVT.getScalarSizeInBits();
assert(SatWidth <= DstElementWidth &&
"Saturation width cannot exceed result width");
// TODO: Consider lowering to SVE operations, as in LowerVectorFP_TO_INT.
// Currently, the `llvm.fpto[su]i.sat.*` intrinsics don't accept scalable
// types, so this is hard to reach.
if (DstVT.isScalableVector())
return SDValue();
EVT SrcElementVT = SrcVT.getVectorElementType();
// In the absence of FP16 support, promote f16 to f32 and saturate the result.
SDLoc DL(Op);
SDValue SrcVal2;
if ((SrcElementVT == MVT::f16 &&
(!Subtarget->hasFullFP16() || DstElementWidth > 16)) ||
SrcElementVT == MVT::bf16) {
MVT F32VT = MVT::getVectorVT(MVT::f32, SrcVT.getVectorNumElements());
SrcVal = DAG.getNode(ISD::FP_EXTEND, DL, F32VT, SrcVal);
// If we are extending to a v8f32, split into two v4f32 to produce legal
// types.
if (F32VT.getSizeInBits() > 128) {
std::tie(SrcVal, SrcVal2) = DAG.SplitVector(SrcVal, DL);
F32VT = F32VT.getHalfNumVectorElementsVT();
}
SrcVT = F32VT;
SrcElementVT = MVT::f32;
SrcElementWidth = 32;
} else if (SrcElementVT != MVT::f64 && SrcElementVT != MVT::f32 &&
SrcElementVT != MVT::f16 && SrcElementVT != MVT::bf16)
return SDValue();
// Expand to f64 if we are saturating to i64, to help keep the lanes the same
// width and produce a fcvtzu.
if (SatWidth == 64 && SrcElementWidth < 64) {
MVT F64VT = MVT::getVectorVT(MVT::f64, SrcVT.getVectorNumElements());
SrcVal = DAG.getNode(ISD::FP_EXTEND, DL, F64VT, SrcVal);
SrcVT = F64VT;
SrcElementVT = MVT::f64;
SrcElementWidth = 64;
}
// Cases that we can emit directly.
if (SrcElementWidth == DstElementWidth && SrcElementWidth == SatWidth) {
SDValue Res = DAG.getNode(Op.getOpcode(), DL, DstVT, SrcVal,
DAG.getValueType(DstVT.getScalarType()));
if (SrcVal2) {
SDValue Res2 = DAG.getNode(Op.getOpcode(), DL, DstVT, SrcVal2,
DAG.getValueType(DstVT.getScalarType()));
return DAG.getNode(ISD::CONCAT_VECTORS, DL, DstVT, Res, Res2);
}
return Res;
}
// Otherwise we emit a cvt that saturates to a higher BW, and saturate the
// result. This is only valid if the legal cvt is larger than the saturate
// width. For double, as we don't have MIN/MAX, it can be simpler to scalarize
// (at least until sqxtn is selected).
if (SrcElementWidth < SatWidth || SrcElementVT == MVT::f64)
return SDValue();
EVT IntVT = SrcVT.changeVectorElementTypeToInteger();
SDValue NativeCvt = DAG.getNode(Op.getOpcode(), DL, IntVT, SrcVal,
DAG.getValueType(IntVT.getScalarType()));
SDValue NativeCvt2 =
SrcVal2 ? DAG.getNode(Op.getOpcode(), DL, IntVT, SrcVal2,
DAG.getValueType(IntVT.getScalarType()))
: SDValue();
SDValue Sat, Sat2;
if (Op.getOpcode() == ISD::FP_TO_SINT_SAT) {
SDValue MinC = DAG.getConstant(
APInt::getSignedMaxValue(SatWidth).sext(SrcElementWidth), DL, IntVT);
SDValue Min = DAG.getNode(ISD::SMIN, DL, IntVT, NativeCvt, MinC);
SDValue Min2 = SrcVal2 ? DAG.getNode(ISD::SMIN, DL, IntVT, NativeCvt2, MinC) : SDValue();
SDValue MaxC = DAG.getConstant(
APInt::getSignedMinValue(SatWidth).sext(SrcElementWidth), DL, IntVT);
Sat = DAG.getNode(ISD::SMAX, DL, IntVT, Min, MaxC);
Sat2 = SrcVal2 ? DAG.getNode(ISD::SMAX, DL, IntVT, Min2, MaxC) : SDValue();
} else {
SDValue MinC = DAG.getConstant(
APInt::getAllOnes(SatWidth).zext(SrcElementWidth), DL, IntVT);
Sat = DAG.getNode(ISD::UMIN, DL, IntVT, NativeCvt, MinC);
Sat2 = SrcVal2 ? DAG.getNode(ISD::UMIN, DL, IntVT, NativeCvt2, MinC) : SDValue();
}
if (SrcVal2)
Sat = DAG.getNode(ISD::CONCAT_VECTORS, DL,
IntVT.getDoubleNumVectorElementsVT(*DAG.getContext()),
Sat, Sat2);
return DAG.getNode(ISD::TRUNCATE, DL, DstVT, Sat);
}
SDValue AArch64TargetLowering::LowerFP_TO_INT_SAT(SDValue Op,
SelectionDAG &DAG) const {
// AArch64 FP-to-int conversions saturate to the destination register size, so
// we can lower common saturating conversions to simple instructions.
SDValue SrcVal = Op.getOperand(0);
EVT SrcVT = SrcVal.getValueType();
if (SrcVT.isVector())
return LowerVectorFP_TO_INT_SAT(Op, DAG);
EVT DstVT = Op.getValueType();
EVT SatVT = cast<VTSDNode>(Op.getOperand(1))->getVT();
uint64_t SatWidth = SatVT.getScalarSizeInBits();
uint64_t DstWidth = DstVT.getScalarSizeInBits();
assert(SatWidth <= DstWidth && "Saturation width cannot exceed result width");
// In the absence of FP16 support, promote f16 to f32 and saturate the result.
if ((SrcVT == MVT::f16 && !Subtarget->hasFullFP16()) || SrcVT == MVT::bf16) {
SrcVal = DAG.getNode(ISD::FP_EXTEND, SDLoc(Op), MVT::f32, SrcVal);
SrcVT = MVT::f32;
} else if (SrcVT != MVT::f64 && SrcVT != MVT::f32 && SrcVT != MVT::f16 &&
SrcVT != MVT::bf16)
return SDValue();
SDLoc DL(Op);
// Cases that we can emit directly.
if ((SrcVT == MVT::f64 || SrcVT == MVT::f32 ||
(SrcVT == MVT::f16 && Subtarget->hasFullFP16())) &&
DstVT == SatVT && (DstVT == MVT::i64 || DstVT == MVT::i32))
return DAG.getNode(Op.getOpcode(), DL, DstVT, SrcVal,
DAG.getValueType(DstVT));
// Otherwise we emit a cvt that saturates to a higher BW, and saturate the
// result. This is only valid if the legal cvt is larger than the saturate
// width.
if (DstWidth < SatWidth)
return SDValue();
SDValue NativeCvt =
DAG.getNode(Op.getOpcode(), DL, DstVT, SrcVal, DAG.getValueType(DstVT));
SDValue Sat;
if (Op.getOpcode() == ISD::FP_TO_SINT_SAT) {
SDValue MinC = DAG.getConstant(
APInt::getSignedMaxValue(SatWidth).sext(DstWidth), DL, DstVT);
SDValue Min = DAG.getNode(ISD::SMIN, DL, DstVT, NativeCvt, MinC);
SDValue MaxC = DAG.getConstant(
APInt::getSignedMinValue(SatWidth).sext(DstWidth), DL, DstVT);
Sat = DAG.getNode(ISD::SMAX, DL, DstVT, Min, MaxC);
} else {
SDValue MinC = DAG.getConstant(
APInt::getAllOnes(SatWidth).zext(DstWidth), DL, DstVT);
Sat = DAG.getNode(ISD::UMIN, DL, DstVT, NativeCvt, MinC);
}
return DAG.getNode(ISD::TRUNCATE, DL, DstVT, Sat);
}
SDValue AArch64TargetLowering::LowerVectorXRINT(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
SDValue Src = Op.getOperand(0);
SDLoc DL(Op);
assert(VT.isVector() && "Expected vector type");
EVT CastVT =
VT.changeVectorElementType(Src.getValueType().getVectorElementType());
// Round the floating-point value into a floating-point register with the
// current rounding mode.
SDValue FOp = DAG.getNode(ISD::FRINT, DL, CastVT, Src);
// Truncate the rounded floating point to an integer.
return DAG.getNode(ISD::FP_TO_SINT_SAT, DL, VT, FOp,
DAG.getValueType(VT.getVectorElementType()));
}
SDValue AArch64TargetLowering::LowerVectorINT_TO_FP(SDValue Op,
SelectionDAG &DAG) const {
// Warning: We maintain cost tables in AArch64TargetTransformInfo.cpp.
// Any additional optimization in this function should be recorded
// in the cost tables.
bool IsStrict = Op->isStrictFPOpcode();
EVT VT = Op.getValueType();
SDLoc DL(Op);
SDValue In = Op.getOperand(IsStrict ? 1 : 0);
EVT InVT = In.getValueType();
unsigned Opc = Op.getOpcode();
bool IsSigned = Opc == ISD::SINT_TO_FP || Opc == ISD::STRICT_SINT_TO_FP;
assert(!(IsStrict && VT.isScalableVector()) &&
"Unimplemented SVE support for ISD:::STRICT_INT_TO_FP!");
// NOTE: i1->bf16 does not require promotion to f32.
if (VT.isScalableVector() && InVT.getVectorElementType() == MVT::i1) {
SDValue FalseVal = DAG.getConstantFP(0.0, DL, VT);
SDValue TrueVal = IsSigned ? DAG.getConstantFP(-1.0, DL, VT)
: DAG.getConstantFP(1.0, DL, VT);
return DAG.getNode(ISD::VSELECT, DL, VT, In, TrueVal, FalseVal);
}
// Promote bf16 conversions to f32.
if (VT.getVectorElementType() == MVT::bf16) {
EVT F32 = VT.changeElementType(MVT::f32);
if (IsStrict) {
SDValue Val = DAG.getNode(Op.getOpcode(), DL, {F32, MVT::Other},
{Op.getOperand(0), In});
return DAG.getNode(ISD::STRICT_FP_ROUND, DL,
{Op.getValueType(), MVT::Other},
{Val.getValue(1), Val.getValue(0),
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true)});
}
return DAG.getNode(ISD::FP_ROUND, DL, Op.getValueType(),
DAG.getNode(Op.getOpcode(), DL, F32, In),
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true));
}
if (VT.isScalableVector()) {
// Let common code split the operation.
if (VT == MVT::nxv8f32)
return Op;
unsigned Opcode = IsSigned ? AArch64ISD::SINT_TO_FP_MERGE_PASSTHRU
: AArch64ISD::UINT_TO_FP_MERGE_PASSTHRU;
return LowerToPredicatedOp(Op, DAG, Opcode);
}
if (useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()) ||
useSVEForFixedLengthVectorVT(InVT, !Subtarget->isNeonAvailable()))
return LowerFixedLengthIntToFPToSVE(Op, DAG);
uint64_t VTSize = VT.getFixedSizeInBits();
uint64_t InVTSize = InVT.getFixedSizeInBits();
if (VTSize < InVTSize) {
// AArch64 doesn't have a direct vector instruction to convert
// fixed point to floating point AND narrow it at the same time.
// Additional rounding when the target is f32/f64 causes double
// rounding issues. Conversion to f16 is fine due to narrow width.
bool IsTargetf32 = VT.getVectorElementType() == MVT::f32;
bool IsTargetf16 = false;
if (Op.hasOneUse() &&
Op->user_begin()->getOpcode() == ISD::CONCAT_VECTORS) {
// Some vector types are split during legalization into half, followed by
// concatenation, followed by rounding to the original vector type. If we
// end up resolving to f16 type, we shouldn't worry about rounding errors.
SDNode *U = *Op->user_begin();
if (U->hasOneUse() && U->user_begin()->getOpcode() == ISD::FP_ROUND) {
EVT TmpVT = U->user_begin()->getValueType(0);
if (TmpVT.getScalarType() == MVT::f16)
IsTargetf16 = true;
}
}
if (IsTargetf32 && !IsTargetf16) {
return !IsStrict ? DAG.UnrollVectorOp(Op.getNode()) : SDValue();
}
MVT CastVT =
MVT::getVectorVT(MVT::getFloatingPointVT(InVT.getScalarSizeInBits()),
InVT.getVectorNumElements());
if (IsStrict) {
In = DAG.getNode(Opc, DL, {CastVT, MVT::Other}, {Op.getOperand(0), In});
return DAG.getNode(ISD::STRICT_FP_ROUND, DL, {VT, MVT::Other},
{In.getValue(1), In.getValue(0),
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true)});
}
In = DAG.getNode(Opc, DL, CastVT, In);
return DAG.getNode(ISD::FP_ROUND, DL, VT, In,
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true));
}
if (VTSize > InVTSize) {
unsigned CastOpc = IsSigned ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND;
EVT CastVT = VT.changeVectorElementTypeToInteger();
In = DAG.getNode(CastOpc, DL, CastVT, In);
if (IsStrict)
return DAG.getNode(Opc, DL, {VT, MVT::Other}, {Op.getOperand(0), In});
return DAG.getNode(Opc, DL, VT, In);
}
// Use a scalar operation for conversions between single-element vectors of
// the same size.
if (VT.getVectorNumElements() == 1) {
SDValue Extract =
DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, InVT.getScalarType(), In,
DAG.getConstant(0, DL, MVT::i64));
EVT ScalarVT = VT.getScalarType();
if (IsStrict)
return DAG.getNode(Op.getOpcode(), DL, {ScalarVT, MVT::Other},
{Op.getOperand(0), Extract});
return DAG.getNode(Op.getOpcode(), DL, ScalarVT, Extract);
}
return Op;
}
SDValue AArch64TargetLowering::LowerINT_TO_FP(SDValue Op,
SelectionDAG &DAG) const {
if (Op.getValueType().isVector())
return LowerVectorINT_TO_FP(Op, DAG);
bool IsStrict = Op->isStrictFPOpcode();
SDValue SrcVal = Op.getOperand(IsStrict ? 1 : 0);
bool IsSigned = Op->getOpcode() == ISD::STRICT_SINT_TO_FP ||
Op->getOpcode() == ISD::SINT_TO_FP;
auto IntToFpViaPromotion = [&](EVT PromoteVT) {
SDLoc DL(Op);
if (IsStrict) {
SDValue Val = DAG.getNode(Op.getOpcode(), DL, {PromoteVT, MVT::Other},
{Op.getOperand(0), SrcVal});
return DAG.getNode(ISD::STRICT_FP_ROUND, DL,
{Op.getValueType(), MVT::Other},
{Val.getValue(1), Val.getValue(0),
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true)});
}
return DAG.getNode(ISD::FP_ROUND, DL, Op.getValueType(),
DAG.getNode(Op.getOpcode(), DL, PromoteVT, SrcVal),
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true));
};
if (Op.getValueType() == MVT::bf16) {
unsigned MaxWidth = IsSigned
? DAG.ComputeMaxSignificantBits(SrcVal)
: DAG.computeKnownBits(SrcVal).countMaxActiveBits();
// bf16 conversions are promoted to f32 when converting from i16.
if (MaxWidth <= 24) {
return IntToFpViaPromotion(MVT::f32);
}
// bf16 conversions are promoted to f64 when converting from i32.
if (MaxWidth <= 53) {
return IntToFpViaPromotion(MVT::f64);
}
// We need to be careful about i64 -> bf16.
// Consider an i32 22216703.
// This number cannot be represented exactly as an f32 and so a itofp will
// turn it into 22216704.0 fptrunc to bf16 will turn this into 22282240.0
// However, the correct bf16 was supposed to be 22151168.0
// We need to use sticky rounding to get this correct.
if (SrcVal.getValueType() == MVT::i64) {
SDLoc DL(Op);
// This algorithm is equivalent to the following:
// uint64_t SrcHi = SrcVal & ~0xfffull;
// uint64_t SrcLo = SrcVal & 0xfffull;
// uint64_t Highest = SrcVal >> 53;
// bool HasHighest = Highest != 0;
// uint64_t ToRound = HasHighest ? SrcHi : SrcVal;
// double Rounded = static_cast<double>(ToRound);
// uint64_t RoundedBits = std::bit_cast<uint64_t>(Rounded);
// uint64_t HasLo = SrcLo != 0;
// bool NeedsAdjustment = HasHighest & HasLo;
// uint64_t AdjustedBits = RoundedBits | uint64_t{NeedsAdjustment};
// double Adjusted = std::bit_cast<double>(AdjustedBits);
// return static_cast<__bf16>(Adjusted);
//
// Essentially, what happens is that SrcVal either fits perfectly in a
// double-precision value or it is too big. If it is sufficiently small,
// we should just go u64 -> double -> bf16 in a naive way. Otherwise, we
// ensure that u64 -> double has no rounding error by only using the 52
// MSB of the input. The low order bits will get merged into a sticky bit
// which will avoid issues incurred by double rounding.
// Signed conversion is more or less like so:
// copysign((__bf16)abs(SrcVal), SrcVal)
SDValue SignBit;
if (IsSigned) {
SignBit = DAG.getNode(ISD::AND, DL, MVT::i64, SrcVal,
DAG.getConstant(1ull << 63, DL, MVT::i64));
SrcVal = DAG.getNode(ISD::ABS, DL, MVT::i64, SrcVal);
}
SDValue SrcHi = DAG.getNode(ISD::AND, DL, MVT::i64, SrcVal,
DAG.getConstant(~0xfffull, DL, MVT::i64));
SDValue SrcLo = DAG.getNode(ISD::AND, DL, MVT::i64, SrcVal,
DAG.getConstant(0xfffull, DL, MVT::i64));
SDValue Highest =
DAG.getNode(ISD::SRL, DL, MVT::i64, SrcVal,
DAG.getShiftAmountConstant(53, MVT::i64, DL));
SDValue Zero64 = DAG.getConstant(0, DL, MVT::i64);
SDValue ToRound =
DAG.getSelectCC(DL, Highest, Zero64, SrcHi, SrcVal, ISD::SETNE);
SDValue Rounded =
IsStrict ? DAG.getNode(Op.getOpcode(), DL, {MVT::f64, MVT::Other},
{Op.getOperand(0), ToRound})
: DAG.getNode(Op.getOpcode(), DL, MVT::f64, ToRound);
SDValue RoundedBits = DAG.getNode(ISD::BITCAST, DL, MVT::i64, Rounded);
if (SignBit) {
RoundedBits = DAG.getNode(ISD::OR, DL, MVT::i64, RoundedBits, SignBit);
}
SDValue HasHighest = DAG.getSetCC(
DL,
getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(), MVT::i64),
Highest, Zero64, ISD::SETNE);
SDValue HasLo = DAG.getSetCC(
DL,
getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(), MVT::i64),
SrcLo, Zero64, ISD::SETNE);
SDValue NeedsAdjustment =
DAG.getNode(ISD::AND, DL, HasLo.getValueType(), HasHighest, HasLo);
NeedsAdjustment = DAG.getZExtOrTrunc(NeedsAdjustment, DL, MVT::i64);
SDValue AdjustedBits =
DAG.getNode(ISD::OR, DL, MVT::i64, RoundedBits, NeedsAdjustment);
SDValue Adjusted = DAG.getNode(ISD::BITCAST, DL, MVT::f64, AdjustedBits);
return IsStrict
? DAG.getNode(
ISD::STRICT_FP_ROUND, DL,
{Op.getValueType(), MVT::Other},
{Rounded.getValue(1), Adjusted,
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true)})
: DAG.getNode(ISD::FP_ROUND, DL, Op.getValueType(), Adjusted,
DAG.getIntPtrConstant(0, DL, /*isTarget=*/true));
}
}
// f16 conversions are promoted to f32 when full fp16 is not supported.
if (Op.getValueType() == MVT::f16 && !Subtarget->hasFullFP16()) {
return IntToFpViaPromotion(MVT::f32);
}
// i128 conversions are libcalls.
if (SrcVal.getValueType() == MVT::i128)
return SDValue();
// Other conversions are legal, unless it's to the completely software-based
// fp128.
if (Op.getValueType() != MVT::f128)
return Op;
return SDValue();
}
SDValue AArch64TargetLowering::LowerFSINCOS(SDValue Op,
SelectionDAG &DAG) const {
// For iOS, we want to call an alternative entry point: __sincos_stret,
// which returns the values in two S / D registers.
SDLoc DL(Op);
SDValue Arg = Op.getOperand(0);
EVT ArgVT = Arg.getValueType();
Type *ArgTy = ArgVT.getTypeForEVT(*DAG.getContext());
ArgListTy Args;
ArgListEntry Entry;
Entry.Node = Arg;
Entry.Ty = ArgTy;
Entry.IsSExt = false;
Entry.IsZExt = false;
Args.push_back(Entry);
RTLIB::Libcall LC = ArgVT == MVT::f64 ? RTLIB::SINCOS_STRET_F64
: RTLIB::SINCOS_STRET_F32;
const char *LibcallName = getLibcallName(LC);
SDValue Callee =
DAG.getExternalSymbol(LibcallName, getPointerTy(DAG.getDataLayout()));
StructType *RetTy = StructType::get(ArgTy, ArgTy);
TargetLowering::CallLoweringInfo CLI(DAG);
CallingConv::ID CC = getLibcallCallingConv(LC);
CLI.setDebugLoc(DL)
.setChain(DAG.getEntryNode())
.setLibCallee(CC, RetTy, Callee, std::move(Args));
std::pair<SDValue, SDValue> CallResult = LowerCallTo(CLI);
return CallResult.first;
}
static MVT getSVEContainerType(EVT ContentTy);
SDValue AArch64TargetLowering::LowerBITCAST(SDValue Op,
SelectionDAG &DAG) const {
EVT OpVT = Op.getValueType();
EVT ArgVT = Op.getOperand(0).getValueType();
if (useSVEForFixedLengthVectorVT(OpVT))
return LowerFixedLengthBitcastToSVE(Op, DAG);
if (OpVT.isScalableVector()) {
assert(isTypeLegal(OpVT) && "Unexpected result type!");
// Handle type legalisation first.
if (!isTypeLegal(ArgVT)) {
assert(OpVT.isFloatingPoint() && !ArgVT.isFloatingPoint() &&
"Expected int->fp bitcast!");
// Bitcasting between unpacked vector types of different element counts is
// not a NOP because the live elements are laid out differently.
// 01234567
// e.g. nxv2i32 = XX??XX??
// nxv4f16 = X?X?X?X?
if (OpVT.getVectorElementCount() != ArgVT.getVectorElementCount())
return SDValue();
SDValue ExtResult =
DAG.getNode(ISD::ANY_EXTEND, SDLoc(Op), getSVEContainerType(ArgVT),
Op.getOperand(0));
return getSVESafeBitCast(OpVT, ExtResult, DAG);
}
// Bitcasts between legal types with the same element count are legal.
if (OpVT.getVectorElementCount() == ArgVT.getVectorElementCount())
return Op;
// getSVESafeBitCast does not support casting between unpacked types.
if (!isPackedVectorType(OpVT, DAG))
return SDValue();
return getSVESafeBitCast(OpVT, Op.getOperand(0), DAG);
}
if (OpVT != MVT::f16 && OpVT != MVT::bf16)
return SDValue();
// Bitcasts between f16 and bf16 are legal.
if (ArgVT == MVT::f16 || ArgVT == MVT::bf16)
return Op;
assert(ArgVT == MVT::i16);
SDLoc DL(Op);
Op = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, Op.getOperand(0));
Op = DAG.getNode(ISD::BITCAST, DL, MVT::f32, Op);
return DAG.getTargetExtractSubreg(AArch64::hsub, DL, OpVT, Op);
}
// Returns lane if Op extracts from a two-element vector and lane is constant
// (i.e., extractelt(<2 x Ty> %v, ConstantLane)), and std::nullopt otherwise.
static std::optional<uint64_t>
getConstantLaneNumOfExtractHalfOperand(SDValue &Op) {
SDNode *OpNode = Op.getNode();
if (OpNode->getOpcode() != ISD::EXTRACT_VECTOR_ELT)
return std::nullopt;
EVT VT = OpNode->getOperand(0).getValueType();
ConstantSDNode *C = dyn_cast<ConstantSDNode>(OpNode->getOperand(1));
if (!VT.isFixedLengthVector() || VT.getVectorNumElements() != 2 || !C)
return std::nullopt;
return C->getZExtValue();
}
static bool isExtendedBUILD_VECTOR(SDValue N, SelectionDAG &DAG,
bool isSigned) {
EVT VT = N.getValueType();
if (N.getOpcode() != ISD::BUILD_VECTOR)
return false;
for (const SDValue &Elt : N->op_values()) {
if (ConstantSDNode *C = dyn_cast<ConstantSDNode>(Elt)) {
unsigned EltSize = VT.getScalarSizeInBits();
unsigned HalfSize = EltSize / 2;
if (isSigned) {
if (!isIntN(HalfSize, C->getSExtValue()))
return false;
} else {
if (!isUIntN(HalfSize, C->getZExtValue()))
return false;
}
continue;
}
return false;
}
return true;
}
static SDValue skipExtensionForVectorMULL(SDValue N, SelectionDAG &DAG) {
EVT VT = N.getValueType();
assert(VT.is128BitVector() && "Unexpected vector MULL size");
EVT HalfVT = EVT::getVectorVT(
*DAG.getContext(),
VT.getScalarType().getHalfSizedIntegerVT(*DAG.getContext()),
VT.getVectorElementCount());
return DAG.getNode(ISD::TRUNCATE, SDLoc(N), HalfVT, N);
}
static bool isSignExtended(SDValue N, SelectionDAG &DAG) {
return N.getOpcode() == ISD::SIGN_EXTEND ||
N.getOpcode() == ISD::ANY_EXTEND ||
isExtendedBUILD_VECTOR(N, DAG, true);
}
static bool isZeroExtended(SDValue N, SelectionDAG &DAG) {
return N.getOpcode() == ISD::ZERO_EXTEND ||
N.getOpcode() == ISD::ANY_EXTEND ||
isExtendedBUILD_VECTOR(N, DAG, false);
}
static bool isAddSubSExt(SDValue N, SelectionDAG &DAG) {
unsigned Opcode = N.getOpcode();
if (Opcode == ISD::ADD || Opcode == ISD::SUB) {
SDValue N0 = N.getOperand(0);
SDValue N1 = N.getOperand(1);
return N0->hasOneUse() && N1->hasOneUse() &&
isSignExtended(N0, DAG) && isSignExtended(N1, DAG);
}
return false;
}
static bool isAddSubZExt(SDValue N, SelectionDAG &DAG) {
unsigned Opcode = N.getOpcode();
if (Opcode == ISD::ADD || Opcode == ISD::SUB) {
SDValue N0 = N.getOperand(0);
SDValue N1 = N.getOperand(1);
return N0->hasOneUse() && N1->hasOneUse() &&
isZeroExtended(N0, DAG) && isZeroExtended(N1, DAG);
}
return false;
}
SDValue AArch64TargetLowering::LowerGET_ROUNDING(SDValue Op,
SelectionDAG &DAG) const {
// The rounding mode is in bits 23:22 of the FPSCR.
// The ARM rounding mode value to FLT_ROUNDS mapping is 0->1, 1->2, 2->3, 3->0
// The formula we use to implement this is (((FPSCR + 1 << 22) >> 22) & 3)
// so that the shift + and get folded into a bitfield extract.
SDLoc DL(Op);
SDValue Chain = Op.getOperand(0);
SDValue FPCR_64 = DAG.getNode(
ISD::INTRINSIC_W_CHAIN, DL, {MVT::i64, MVT::Other},
{Chain, DAG.getConstant(Intrinsic::aarch64_get_fpcr, DL, MVT::i64)});
Chain = FPCR_64.getValue(1);
SDValue FPCR_32 = DAG.getNode(ISD::TRUNCATE, DL, MVT::i32, FPCR_64);
SDValue FltRounds = DAG.getNode(ISD::ADD, DL, MVT::i32, FPCR_32,
DAG.getConstant(1U << 22, DL, MVT::i32));
SDValue RMODE = DAG.getNode(ISD::SRL, DL, MVT::i32, FltRounds,
DAG.getConstant(22, DL, MVT::i32));
SDValue AND = DAG.getNode(ISD::AND, DL, MVT::i32, RMODE,
DAG.getConstant(3, DL, MVT::i32));
return DAG.getMergeValues({AND, Chain}, DL);
}
SDValue AArch64TargetLowering::LowerSET_ROUNDING(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
SDValue Chain = Op->getOperand(0);
SDValue RMValue = Op->getOperand(1);
// The rounding mode is in bits 23:22 of the FPCR.
// The llvm.set.rounding argument value to the rounding mode in FPCR mapping
// is 0->3, 1->0, 2->1, 3->2. The formula we use to implement this is
// ((arg - 1) & 3) << 22).
//
// The argument of llvm.set.rounding must be within the segment [0, 3], so
// NearestTiesToAway (4) is not handled here. It is responsibility of the code
// generated llvm.set.rounding to ensure this condition.
// Calculate new value of FPCR[23:22].
RMValue = DAG.getNode(ISD::SUB, DL, MVT::i32, RMValue,
DAG.getConstant(1, DL, MVT::i32));
RMValue = DAG.getNode(ISD::AND, DL, MVT::i32, RMValue,
DAG.getConstant(0x3, DL, MVT::i32));
RMValue =
DAG.getNode(ISD::SHL, DL, MVT::i32, RMValue,
DAG.getConstant(AArch64::RoundingBitsPos, DL, MVT::i32));
RMValue = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i64, RMValue);
// Get current value of FPCR.
SDValue Ops[] = {
Chain, DAG.getTargetConstant(Intrinsic::aarch64_get_fpcr, DL, MVT::i64)};
SDValue FPCR =
DAG.getNode(ISD::INTRINSIC_W_CHAIN, DL, {MVT::i64, MVT::Other}, Ops);
Chain = FPCR.getValue(1);
FPCR = FPCR.getValue(0);
// Put new rounding mode into FPSCR[23:22].
const int RMMask = ~(AArch64::Rounding::rmMask << AArch64::RoundingBitsPos);
FPCR = DAG.getNode(ISD::AND, DL, MVT::i64, FPCR,
DAG.getConstant(RMMask, DL, MVT::i64));
FPCR = DAG.getNode(ISD::OR, DL, MVT::i64, FPCR, RMValue);
SDValue Ops2[] = {
Chain, DAG.getTargetConstant(Intrinsic::aarch64_set_fpcr, DL, MVT::i64),
FPCR};
return DAG.getNode(ISD::INTRINSIC_VOID, DL, MVT::Other, Ops2);
}
SDValue AArch64TargetLowering::LowerGET_FPMODE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
SDValue Chain = Op->getOperand(0);
// Get current value of FPCR.
SDValue Ops[] = {
Chain, DAG.getTargetConstant(Intrinsic::aarch64_get_fpcr, DL, MVT::i64)};
SDValue FPCR =
DAG.getNode(ISD::INTRINSIC_W_CHAIN, DL, {MVT::i64, MVT::Other}, Ops);
Chain = FPCR.getValue(1);
FPCR = FPCR.getValue(0);
// Truncate FPCR to 32 bits.
SDValue Result = DAG.getNode(ISD::TRUNCATE, DL, MVT::i32, FPCR);
return DAG.getMergeValues({Result, Chain}, DL);
}
SDValue AArch64TargetLowering::LowerSET_FPMODE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
SDValue Chain = Op->getOperand(0);
SDValue Mode = Op->getOperand(1);
// Extend the specified value to 64 bits.
SDValue FPCR = DAG.getZExtOrTrunc(Mode, DL, MVT::i64);
// Set new value of FPCR.
SDValue Ops2[] = {
Chain, DAG.getConstant(Intrinsic::aarch64_set_fpcr, DL, MVT::i64), FPCR};
return DAG.getNode(ISD::INTRINSIC_VOID, DL, MVT::Other, Ops2);
}
SDValue AArch64TargetLowering::LowerRESET_FPMODE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
SDValue Chain = Op->getOperand(0);
// Get current value of FPCR.
SDValue Ops[] = {
Chain, DAG.getTargetConstant(Intrinsic::aarch64_get_fpcr, DL, MVT::i64)};
SDValue FPCR =
DAG.getNode(ISD::INTRINSIC_W_CHAIN, DL, {MVT::i64, MVT::Other}, Ops);
Chain = FPCR.getValue(1);
FPCR = FPCR.getValue(0);
// Clear bits that are not reserved.
SDValue FPSCRMasked = DAG.getNode(
ISD::AND, DL, MVT::i64, FPCR,
DAG.getConstant(AArch64::ReservedFPControlBits, DL, MVT::i64));
// Set new value of FPCR.
SDValue Ops2[] = {Chain,
DAG.getConstant(Intrinsic::aarch64_set_fpcr, DL, MVT::i64),
FPSCRMasked};
return DAG.getNode(ISD::INTRINSIC_VOID, DL, MVT::Other, Ops2);
}
static unsigned selectUmullSmull(SDValue &N0, SDValue &N1, SelectionDAG &DAG,
SDLoc DL, bool &IsMLA) {
bool IsN0SExt = isSignExtended(N0, DAG);
bool IsN1SExt = isSignExtended(N1, DAG);
if (IsN0SExt && IsN1SExt)
return AArch64ISD::SMULL;
bool IsN0ZExt = isZeroExtended(N0, DAG);
bool IsN1ZExt = isZeroExtended(N1, DAG);
if (IsN0ZExt && IsN1ZExt)
return AArch64ISD::UMULL;
// Select UMULL if we can replace the other operand with an extend.
EVT VT = N0.getValueType();
unsigned EltSize = VT.getScalarSizeInBits();
APInt Mask = APInt::getHighBitsSet(EltSize, EltSize / 2);
if (IsN0ZExt || IsN1ZExt) {
if (DAG.MaskedValueIsZero(IsN0ZExt ? N1 : N0, Mask))
return AArch64ISD::UMULL;
} else if (VT == MVT::v2i64 && DAG.MaskedValueIsZero(N0, Mask) &&
DAG.MaskedValueIsZero(N1, Mask)) {
// For v2i64 we look more aggressively at both operands being zero, to avoid
// scalarization.
return AArch64ISD::UMULL;
}
if (IsN0SExt || IsN1SExt) {
if (DAG.ComputeNumSignBits(IsN0SExt ? N1 : N0) > EltSize / 2)
return AArch64ISD::SMULL;
} else if (VT == MVT::v2i64 && DAG.ComputeNumSignBits(N0) > EltSize / 2 &&
DAG.ComputeNumSignBits(N1) > EltSize / 2) {
return AArch64ISD::SMULL;
}
if (!IsN1SExt && !IsN1ZExt)
return 0;
// Look for (s/zext A + s/zext B) * (s/zext C). We want to turn these
// into (s/zext A * s/zext C) + (s/zext B * s/zext C)
if (IsN1SExt && isAddSubSExt(N0, DAG)) {
IsMLA = true;
return AArch64ISD::SMULL;
}
if (IsN1ZExt && isAddSubZExt(N0, DAG)) {
IsMLA = true;
return AArch64ISD::UMULL;
}
if (IsN0ZExt && isAddSubZExt(N1, DAG)) {
std::swap(N0, N1);
IsMLA = true;
return AArch64ISD::UMULL;
}
return 0;
}
SDValue AArch64TargetLowering::LowerMUL(SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
bool OverrideNEON = !Subtarget->isNeonAvailable();
if (VT.isScalableVector() || useSVEForFixedLengthVectorVT(VT, OverrideNEON))
return LowerToPredicatedOp(Op, DAG, AArch64ISD::MUL_PRED);
// Multiplications are only custom-lowered for 128-bit and 64-bit vectors so
// that VMULL can be detected. Otherwise v2i64 multiplications are not legal.
assert((VT.is128BitVector() || VT.is64BitVector()) && VT.isInteger() &&
"unexpected type for custom-lowering ISD::MUL");
SDValue N0 = Op.getOperand(0);
SDValue N1 = Op.getOperand(1);
bool isMLA = false;
EVT OVT = VT;
if (VT.is64BitVector()) {
if (N0.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
isNullConstant(N0.getOperand(1)) &&
N1.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
isNullConstant(N1.getOperand(1))) {
N0 = N0.getOperand(0);
N1 = N1.getOperand(0);
VT = N0.getValueType();
} else {
if (VT == MVT::v1i64) {
if (Subtarget->hasSVE())
return LowerToPredicatedOp(Op, DAG, AArch64ISD::MUL_PRED);
// Fall through to expand this. It is not legal.
return SDValue();
} else
// Other vector multiplications are legal.
return Op;
}
}
SDLoc DL(Op);
unsigned NewOpc = selectUmullSmull(N0, N1, DAG, DL, isMLA);
if (!NewOpc) {
if (VT.getVectorElementType() == MVT::i64) {
// If SVE is available then i64 vector multiplications can also be made
// legal.
if (Subtarget->hasSVE())
return LowerToPredicatedOp(Op, DAG, AArch64ISD::MUL_PRED);
// Fall through to expand this. It is not legal.
return SDValue();
} else
// Other vector multiplications are legal.
return Op;
}
// Legalize to a S/UMULL instruction
SDValue Op0;
SDValue Op1 = skipExtensionForVectorMULL(N1, DAG);
if (!isMLA) {
Op0 = skipExtensionForVectorMULL(N0, DAG);
assert(Op0.getValueType().is64BitVector() &&
Op1.getValueType().is64BitVector() &&
"unexpected types for extended operands to VMULL");
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, OVT,
DAG.getNode(NewOpc, DL, VT, Op0, Op1),
DAG.getConstant(0, DL, MVT::i64));
}
// Optimizing (zext A + zext B) * C, to (S/UMULL A, C) + (S/UMULL B, C) during
// isel lowering to take advantage of no-stall back to back s/umul + s/umla.
// This is true for CPUs with accumulate forwarding such as Cortex-A53/A57
SDValue N00 = skipExtensionForVectorMULL(N0.getOperand(0), DAG);
SDValue N01 = skipExtensionForVectorMULL(N0.getOperand(1), DAG);
EVT Op1VT = Op1.getValueType();
return DAG.getNode(
ISD::EXTRACT_SUBVECTOR, DL, OVT,
DAG.getNode(N0.getOpcode(), DL, VT,
DAG.getNode(NewOpc, DL, VT,
DAG.getNode(ISD::BITCAST, DL, Op1VT, N00), Op1),
DAG.getNode(NewOpc, DL, VT,
DAG.getNode(ISD::BITCAST, DL, Op1VT, N01), Op1)),
DAG.getConstant(0, DL, MVT::i64));
}
static inline SDValue getPTrue(SelectionDAG &DAG, SDLoc DL, EVT VT,
int Pattern) {
if (Pattern == AArch64SVEPredPattern::all)
return DAG.getConstant(1, DL, VT);
return DAG.getNode(AArch64ISD::PTRUE, DL, VT,
DAG.getTargetConstant(Pattern, DL, MVT::i32));
}
static SDValue optimizeIncrementingWhile(SDNode *N, SelectionDAG &DAG,
bool IsSigned, bool IsEqual) {
unsigned Op0 = N->getOpcode() == ISD::INTRINSIC_WO_CHAIN ? 1 : 0;
unsigned Op1 = N->getOpcode() == ISD::INTRINSIC_WO_CHAIN ? 2 : 1;
if (!isa<ConstantSDNode>(N->getOperand(Op1)))
return SDValue();
SDLoc DL(N);
APInt Y = N->getConstantOperandAPInt(Op1);
// When the second operand is the maximum value, comparisons that include
// equality can never fail and thus we can return an all active predicate.
if (IsEqual)
if (IsSigned ? Y.isMaxSignedValue() : Y.isMaxValue())
return DAG.getConstant(1, DL, N->getValueType(0));
if (!isa<ConstantSDNode>(N->getOperand(Op0)))
return SDValue();
APInt X = N->getConstantOperandAPInt(Op0);
bool Overflow;
APInt NumActiveElems =
IsSigned ? Y.ssub_ov(X, Overflow) : Y.usub_ov(X, Overflow);
if (Overflow)
return SDValue();
if (IsEqual) {
APInt One(NumActiveElems.getBitWidth(), 1, IsSigned);
NumActiveElems = IsSigned ? NumActiveElems.sadd_ov(One, Overflow)
: NumActiveElems.uadd_ov(One, Overflow);
if (Overflow)
return SDValue();
}
std::optional<unsigned> PredPattern =
getSVEPredPatternFromNumElements(NumActiveElems.getZExtValue());
unsigned MinSVEVectorSize = std::max(
DAG.getSubtarget<AArch64Subtarget>().getMinSVEVectorSizeInBits(), 128u);
unsigned ElementSize = 128 / N->getValueType(0).getVectorMinNumElements();
if (PredPattern != std::nullopt &&
NumActiveElems.getZExtValue() <= (MinSVEVectorSize / ElementSize))
return getPTrue(DAG, DL, N->getValueType(0), *PredPattern);
return SDValue();
}
// Returns a safe bitcast between two scalable vector predicates, where
// any newly created lanes from a widening bitcast are defined as zero.
static SDValue getSVEPredicateBitCast(EVT VT, SDValue Op, SelectionDAG &DAG) {
SDLoc DL(Op);
EVT InVT = Op.getValueType();
assert(InVT.getVectorElementType() == MVT::i1 &&
VT.getVectorElementType() == MVT::i1 &&
"Expected a predicate-to-predicate bitcast");
assert(VT.isScalableVector() && DAG.getTargetLoweringInfo().isTypeLegal(VT) &&
InVT.isScalableVector() &&
DAG.getTargetLoweringInfo().isTypeLegal(InVT) &&
"Only expect to cast between legal scalable predicate types!");
// Return the operand if the cast isn't changing type,
if (InVT == VT)
return Op;
// Look through casts to <vscale x 16 x i1> when their input has more lanes
// than VT. This will increase the chances of removing casts that introduce
// new lanes, which have to be explicitly zero'd.
if (Op.getOpcode() == ISD::INTRINSIC_WO_CHAIN &&
Op.getConstantOperandVal(0) == Intrinsic::aarch64_sve_convert_to_svbool &&
Op.getOperand(1).getValueType().bitsGT(VT))
Op = Op.getOperand(1);
SDValue Reinterpret = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, VT, Op);
// We only have to zero the lanes if new lanes are being defined, e.g. when
// casting from <vscale x 2 x i1> to <vscale x 16 x i1>. If this is not the
// case (e.g. when casting from <vscale x 16 x i1> -> <vscale x 2 x i1>) then
// we can return here.
if (InVT.bitsGT(VT))
return Reinterpret;
// Check if the other lanes are already known to be zeroed by
// construction.
if (isZeroingInactiveLanes(Op))
return Reinterpret;
// Zero the newly introduced lanes.
SDValue Mask = DAG.getConstant(1, DL, InVT);
Mask = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, VT, Mask);
return DAG.getNode(ISD::AND, DL, VT, Reinterpret, Mask);
}
SDValue AArch64TargetLowering::getRuntimePStateSM(SelectionDAG &DAG,
SDValue Chain, SDLoc DL,
EVT VT) const {
SDValue Callee = DAG.getExternalSymbol("__arm_sme_state",
getPointerTy(DAG.getDataLayout()));
Type *Int64Ty = Type::getInt64Ty(*DAG.getContext());
Type *RetTy = StructType::get(Int64Ty, Int64Ty);
TargetLowering::CallLoweringInfo CLI(DAG);
ArgListTy Args;
CLI.setDebugLoc(DL).setChain(Chain).setLibCallee(
CallingConv::AArch64_SME_ABI_Support_Routines_PreserveMost_From_X2,
RetTy, Callee, std::move(Args));
std::pair<SDValue, SDValue> CallResult = LowerCallTo(CLI);
SDValue Mask = DAG.getConstant(/*PSTATE.SM*/ 1, DL, MVT::i64);
return DAG.getNode(ISD::AND, DL, MVT::i64, CallResult.first.getOperand(0),
Mask);
}
// Lower an SME LDR/STR ZA intrinsic
// Case 1: If the vector number (vecnum) is an immediate in range, it gets
// folded into the instruction
// ldr(%tileslice, %ptr, 11) -> ldr [%tileslice, 11], [%ptr, 11]
// Case 2: If the vecnum is not an immediate, then it is used to modify the base
// and tile slice registers
// ldr(%tileslice, %ptr, %vecnum)
// ->
// %svl = rdsvl
// %ptr2 = %ptr + %svl * %vecnum
// %tileslice2 = %tileslice + %vecnum
// ldr [%tileslice2, 0], [%ptr2, 0]
// Case 3: If the vecnum is an immediate out of range, then the same is done as
// case 2, but the base and slice registers are modified by the greatest
// multiple of 15 lower than the vecnum and the remainder is folded into the
// instruction. This means that successive loads and stores that are offset from
// each other can share the same base and slice register updates.
// ldr(%tileslice, %ptr, 22)
// ldr(%tileslice, %ptr, 23)
// ->
// %svl = rdsvl
// %ptr2 = %ptr + %svl * 15
// %tileslice2 = %tileslice + 15
// ldr [%tileslice2, 7], [%ptr2, 7]
// ldr [%tileslice2, 8], [%ptr2, 8]
// Case 4: If the vecnum is an add of an immediate, then the non-immediate
// operand and the immediate can be folded into the instruction, like case 2.
// ldr(%tileslice, %ptr, %vecnum + 7)
// ldr(%tileslice, %ptr, %vecnum + 8)
// ->
// %svl = rdsvl
// %ptr2 = %ptr + %svl * %vecnum
// %tileslice2 = %tileslice + %vecnum
// ldr [%tileslice2, 7], [%ptr2, 7]
// ldr [%tileslice2, 8], [%ptr2, 8]
// Case 5: The vecnum being an add of an immediate out of range is also handled,
// in which case the same remainder logic as case 3 is used.
SDValue LowerSMELdrStr(SDValue N, SelectionDAG &DAG, bool IsLoad) {
SDLoc DL(N);
SDValue TileSlice = N->getOperand(2);
SDValue Base = N->getOperand(3);
SDValue VecNum = N->getOperand(4);
int32_t ConstAddend = 0;
SDValue VarAddend = VecNum;
// If the vnum is an add of an immediate, we can fold it into the instruction
if (VecNum.getOpcode() == ISD::ADD &&
isa<ConstantSDNode>(VecNum.getOperand(1))) {
ConstAddend = cast<ConstantSDNode>(VecNum.getOperand(1))->getSExtValue();
VarAddend = VecNum.getOperand(0);
} else if (auto ImmNode = dyn_cast<ConstantSDNode>(VecNum)) {
ConstAddend = ImmNode->getSExtValue();
VarAddend = SDValue();
}
int32_t ImmAddend = ConstAddend % 16;
if (int32_t C = (ConstAddend - ImmAddend)) {
SDValue CVal = DAG.getTargetConstant(C, DL, MVT::i32);
VarAddend = VarAddend
? DAG.getNode(ISD::ADD, DL, MVT::i32, {VarAddend, CVal})
: CVal;
}
if (VarAddend) {
// Get the vector length that will be multiplied by vnum
auto SVL = DAG.getNode(AArch64ISD::RDSVL, DL, MVT::i64,
DAG.getConstant(1, DL, MVT::i32));
// Multiply SVL and vnum then add it to the base
SDValue Mul = DAG.getNode(
ISD::MUL, DL, MVT::i64,
{SVL, DAG.getNode(ISD::SIGN_EXTEND, DL, MVT::i64, VarAddend)});
Base = DAG.getNode(ISD::ADD, DL, MVT::i64, {Base, Mul});
// Just add vnum to the tileslice
TileSlice = DAG.getNode(ISD::ADD, DL, MVT::i32, {TileSlice, VarAddend});
}
return DAG.getNode(IsLoad ? AArch64ISD::SME_ZA_LDR : AArch64ISD::SME_ZA_STR,
DL, MVT::Other,
{/*Chain=*/N.getOperand(0), TileSlice, Base,
DAG.getTargetConstant(ImmAddend, DL, MVT::i32)});
}
SDValue LowerVectorMatch(SDValue Op, SelectionDAG &DAG) {
SDLoc DL(Op);
SDValue ID =
DAG.getTargetConstant(Intrinsic::aarch64_sve_match, DL, MVT::i64);
auto Op1 = Op.getOperand(1);
auto Op2 = Op.getOperand(2);
auto Mask = Op.getOperand(3);
EVT Op1VT = Op1.getValueType();
EVT Op2VT = Op2.getValueType();
EVT ResVT = Op.getValueType();
assert((Op1VT.getVectorElementType() == MVT::i8 ||
Op1VT.getVectorElementType() == MVT::i16) &&
"Expected 8-bit or 16-bit characters.");
// Scalable vector type used to wrap operands.
// A single container is enough for both operands because ultimately the
// operands will have to be wrapped to the same type (nxv16i8 or nxv8i16).
EVT OpContainerVT = Op1VT.isScalableVector()
? Op1VT
: getContainerForFixedLengthVector(DAG, Op1VT);
if (Op2VT.is128BitVector()) {
// If Op2 is a full 128-bit vector, wrap it trivially in a scalable vector.
Op2 = convertToScalableVector(DAG, OpContainerVT, Op2);
// Further, if the result is scalable, broadcast Op2 to a full SVE register.
if (ResVT.isScalableVector())
Op2 = DAG.getNode(AArch64ISD::DUPLANE128, DL, OpContainerVT, Op2,
DAG.getTargetConstant(0, DL, MVT::i64));
} else {
// If Op2 is not a full 128-bit vector, we always need to broadcast it.
unsigned Op2BitWidth = Op2VT.getFixedSizeInBits();
MVT Op2IntVT = MVT::getIntegerVT(Op2BitWidth);
EVT Op2PromotedVT = getPackedSVEVectorVT(Op2IntVT);
Op2 = DAG.getBitcast(MVT::getVectorVT(Op2IntVT, 1), Op2);
Op2 = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, Op2IntVT, Op2,
DAG.getConstant(0, DL, MVT::i64));
Op2 = DAG.getSplatVector(Op2PromotedVT, DL, Op2);
Op2 = DAG.getBitcast(OpContainerVT, Op2);
}
// If the result is scalable, we just need to carry out the MATCH.
if (ResVT.isScalableVector())
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, ResVT, ID, Mask, Op1, Op2);
// If the result is fixed, we can still use MATCH but we need to wrap the
// first operand and the mask in scalable vectors before doing so.
// Wrap the operands.
Op1 = convertToScalableVector(DAG, OpContainerVT, Op1);
Mask = DAG.getNode(ISD::SIGN_EXTEND, DL, Op1VT, Mask);
Mask = convertFixedMaskToScalableVector(Mask, DAG);
// Carry out the match.
SDValue Match = DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, Mask.getValueType(),
ID, Mask, Op1, Op2);
// Extract and promote the match result (nxv16i1/nxv8i1) to ResVT
// (v16i8/v8i8).
Match = DAG.getNode(ISD::SIGN_EXTEND, DL, OpContainerVT, Match);
Match = convertFromScalableVector(DAG, Op1VT, Match);
return DAG.getNode(ISD::TRUNCATE, DL, ResVT, Match);
}
SDValue AArch64TargetLowering::LowerINTRINSIC_VOID(SDValue Op,
SelectionDAG &DAG) const {
unsigned IntNo = Op.getConstantOperandVal(1);
SDLoc DL(Op);
switch (IntNo) {
default:
return SDValue(); // Don't custom lower most intrinsics.
case Intrinsic::aarch64_prefetch: {
SDValue Chain = Op.getOperand(0);
SDValue Addr = Op.getOperand(2);
unsigned IsWrite = Op.getConstantOperandVal(3);
unsigned Locality = Op.getConstantOperandVal(4);
unsigned IsStream = Op.getConstantOperandVal(5);
unsigned IsData = Op.getConstantOperandVal(6);
unsigned PrfOp = (IsWrite << 4) | // Load/Store bit
(!IsData << 3) | // IsDataCache bit
(Locality << 1) | // Cache level bits
(unsigned)IsStream; // Stream bit
return DAG.getNode(AArch64ISD::PREFETCH, DL, MVT::Other, Chain,
DAG.getTargetConstant(PrfOp, DL, MVT::i32), Addr);
}
case Intrinsic::aarch64_sme_str:
case Intrinsic::aarch64_sme_ldr: {
return LowerSMELdrStr(Op, DAG, IntNo == Intrinsic::aarch64_sme_ldr);
}
case Intrinsic::aarch64_sme_za_enable:
return DAG.getNode(
AArch64ISD::SMSTART, DL, DAG.getVTList(MVT::Other, MVT::Glue),
Op->getOperand(0), // Chain
DAG.getTargetConstant((int32_t)(AArch64SVCR::SVCRZA), DL, MVT::i32));
case Intrinsic::aarch64_sme_za_disable:
return DAG.getNode(
AArch64ISD::SMSTOP, DL, DAG.getVTList(MVT::Other, MVT::Glue),
Op->getOperand(0), // Chain
DAG.getTargetConstant((int32_t)(AArch64SVCR::SVCRZA), DL, MVT::i32));
}
}
SDValue AArch64TargetLowering::LowerINTRINSIC_W_CHAIN(SDValue Op,
SelectionDAG &DAG) const {
unsigned IntNo = Op.getConstantOperandVal(1);
SDLoc DL(Op);
switch (IntNo) {
default:
return SDValue(); // Don't custom lower most intrinsics.
case Intrinsic::aarch64_mops_memset_tag: {
auto Node = cast<MemIntrinsicSDNode>(Op.getNode());
SDValue Chain = Node->getChain();
SDValue Dst = Op.getOperand(2);
SDValue Val = Op.getOperand(3);
Val = DAG.getAnyExtOrTrunc(Val, DL, MVT::i64);
SDValue Size = Op.getOperand(4);
auto Alignment = Node->getMemOperand()->getAlign();
bool IsVol = Node->isVolatile();
auto DstPtrInfo = Node->getPointerInfo();
const auto &SDI =
static_cast<const AArch64SelectionDAGInfo &>(DAG.getSelectionDAGInfo());
SDValue MS = SDI.EmitMOPS(AArch64::MOPSMemorySetTaggingPseudo, DAG, DL,
Chain, Dst, Val, Size, Alignment, IsVol,
DstPtrInfo, MachinePointerInfo{});
// MOPS_MEMSET_TAGGING has 3 results (DstWb, SizeWb, Chain) whereas the
// intrinsic has 2. So hide SizeWb using MERGE_VALUES. Otherwise
// LowerOperationWrapper will complain that the number of results has
// changed.
return DAG.getMergeValues({MS.getValue(0), MS.getValue(2)}, DL);
}
}
}
SDValue AArch64TargetLowering::LowerINTRINSIC_WO_CHAIN(SDValue Op,
SelectionDAG &DAG) const {
unsigned IntNo = Op.getConstantOperandVal(0);
SDLoc DL(Op);
switch (IntNo) {
default: return SDValue(); // Don't custom lower most intrinsics.
case Intrinsic::thread_pointer: {
EVT PtrVT = getPointerTy(DAG.getDataLayout());
return DAG.getNode(AArch64ISD::THREAD_POINTER, DL, PtrVT);
}
case Intrinsic::aarch64_neon_abs: {
EVT Ty = Op.getValueType();
if (Ty == MVT::i64) {
SDValue Result =
DAG.getNode(ISD::BITCAST, DL, MVT::v1i64, Op.getOperand(1));
Result = DAG.getNode(ISD::ABS, DL, MVT::v1i64, Result);
return DAG.getNode(ISD::BITCAST, DL, MVT::i64, Result);
} else if (Ty.isVector() && Ty.isInteger() && isTypeLegal(Ty)) {
return DAG.getNode(ISD::ABS, DL, Ty, Op.getOperand(1));
} else {
report_fatal_error("Unexpected type for AArch64 NEON intrinsic");
}
}
case Intrinsic::aarch64_neon_pmull64: {
SDValue LHS = Op.getOperand(1);
SDValue RHS = Op.getOperand(2);
std::optional<uint64_t> LHSLane =
getConstantLaneNumOfExtractHalfOperand(LHS);
std::optional<uint64_t> RHSLane =
getConstantLaneNumOfExtractHalfOperand(RHS);
assert((!LHSLane || *LHSLane < 2) && "Expect lane to be None or 0 or 1");
assert((!RHSLane || *RHSLane < 2) && "Expect lane to be None or 0 or 1");
// 'aarch64_neon_pmull64' takes i64 parameters; while pmull/pmull2
// instructions execute on SIMD registers. So canonicalize i64 to v1i64,
// which ISel recognizes better. For example, generate a ldr into d*
// registers as opposed to a GPR load followed by a fmov.
auto TryVectorizeOperand = [](SDValue N, std::optional<uint64_t> NLane,
std::optional<uint64_t> OtherLane,
const SDLoc &DL,
SelectionDAG &DAG) -> SDValue {
// If the operand is an higher half itself, rewrite it to
// extract_high_v2i64; this way aarch64_neon_pmull64 could
// re-use the dag-combiner function with aarch64_neon_{pmull,smull,umull}.
if (NLane == 1)
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v1i64,
N.getOperand(0), DAG.getConstant(1, DL, MVT::i64));
// Operand N is not a higher half but the other operand is.
if (OtherLane == 1) {
// If this operand is a lower half, rewrite it to
// extract_high_v2i64(duplane(<2 x Ty>, 0)). This saves a roundtrip to
// align lanes of two operands. A roundtrip sequence (to move from lane
// 1 to lane 0) is like this:
// mov x8, v0.d[1]
// fmov d0, x8
if (NLane == 0)
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v1i64,
DAG.getNode(AArch64ISD::DUPLANE64, DL, MVT::v2i64,
N.getOperand(0),
DAG.getConstant(0, DL, MVT::i64)),
DAG.getConstant(1, DL, MVT::i64));
// Otherwise just dup from main to all lanes.
return DAG.getNode(AArch64ISD::DUP, DL, MVT::v1i64, N);
}
// Neither operand is an extract of higher half, so codegen may just use
// the non-high version of PMULL instruction. Use v1i64 to represent i64.
assert(N.getValueType() == MVT::i64 &&
"Intrinsic aarch64_neon_pmull64 requires i64 parameters");
return DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, MVT::v1i64, N);
};
LHS = TryVectorizeOperand(LHS, LHSLane, RHSLane, DL, DAG);
RHS = TryVectorizeOperand(RHS, RHSLane, LHSLane, DL, DAG);
return DAG.getNode(AArch64ISD::PMULL, DL, Op.getValueType(), LHS, RHS);
}
case Intrinsic::aarch64_neon_smax:
return DAG.getNode(ISD::SMAX, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
case Intrinsic::aarch64_neon_umax:
return DAG.getNode(ISD::UMAX, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
case Intrinsic::aarch64_neon_smin:
return DAG.getNode(ISD::SMIN, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
case Intrinsic::aarch64_neon_umin:
return DAG.getNode(ISD::UMIN, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
case Intrinsic::aarch64_neon_scalar_sqxtn:
case Intrinsic::aarch64_neon_scalar_sqxtun:
case Intrinsic::aarch64_neon_scalar_uqxtn: {
assert(Op.getValueType() == MVT::i32 || Op.getValueType() == MVT::f32);
if (Op.getValueType() == MVT::i32)
return DAG.getNode(ISD::BITCAST, DL, MVT::i32,
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, MVT::f32,
Op.getOperand(0),
DAG.getNode(ISD::BITCAST, DL, MVT::f64,
Op.getOperand(1))));
return SDValue();
}
case Intrinsic::aarch64_neon_sqxtn:
return DAG.getNode(ISD::TRUNCATE_SSAT_S, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_neon_sqxtun:
return DAG.getNode(ISD::TRUNCATE_SSAT_U, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_neon_uqxtn:
return DAG.getNode(ISD::TRUNCATE_USAT_U, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_neon_sqshrn:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::TRUNCATE_SSAT_S, DL, Op.getValueType(),
DAG.getNode(AArch64ISD::VASHR, DL,
Op.getOperand(1).getValueType(),
Op.getOperand(1), Op.getOperand(2)));
return SDValue();
case Intrinsic::aarch64_neon_sqshrun:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::TRUNCATE_SSAT_U, DL, Op.getValueType(),
DAG.getNode(AArch64ISD::VASHR, DL,
Op.getOperand(1).getValueType(),
Op.getOperand(1), Op.getOperand(2)));
return SDValue();
case Intrinsic::aarch64_neon_uqshrn:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::TRUNCATE_USAT_U, DL, Op.getValueType(),
DAG.getNode(AArch64ISD::VLSHR, DL,
Op.getOperand(1).getValueType(),
Op.getOperand(1), Op.getOperand(2)));
return SDValue();
case Intrinsic::aarch64_neon_sqrshrn:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::TRUNCATE_SSAT_S, DL, Op.getValueType(),
DAG.getNode(AArch64ISD::SRSHR_I, DL,
Op.getOperand(1).getValueType(),
Op.getOperand(1), Op.getOperand(2)));
return SDValue();
case Intrinsic::aarch64_neon_sqrshrun:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::TRUNCATE_SSAT_U, DL, Op.getValueType(),
DAG.getNode(AArch64ISD::SRSHR_I, DL,
Op.getOperand(1).getValueType(),
Op.getOperand(1), Op.getOperand(2)));
return SDValue();
case Intrinsic::aarch64_neon_uqrshrn:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::TRUNCATE_USAT_U, DL, Op.getValueType(),
DAG.getNode(AArch64ISD::URSHR_I, DL,
Op.getOperand(1).getValueType(),
Op.getOperand(1), Op.getOperand(2)));
return SDValue();
case Intrinsic::aarch64_neon_sqadd:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::SADDSAT, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
return SDValue();
case Intrinsic::aarch64_neon_sqsub:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::SSUBSAT, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
return SDValue();
case Intrinsic::aarch64_neon_uqadd:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::UADDSAT, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
return SDValue();
case Intrinsic::aarch64_neon_uqsub:
if (Op.getValueType().isVector())
return DAG.getNode(ISD::USUBSAT, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
return SDValue();
case Intrinsic::aarch64_sve_whilelt:
return optimizeIncrementingWhile(Op.getNode(), DAG, /*IsSigned=*/true,
/*IsEqual=*/false);
case Intrinsic::aarch64_sve_whilels:
return optimizeIncrementingWhile(Op.getNode(), DAG, /*IsSigned=*/false,
/*IsEqual=*/true);
case Intrinsic::aarch64_sve_whilele:
return optimizeIncrementingWhile(Op.getNode(), DAG, /*IsSigned=*/true,
/*IsEqual=*/true);
case Intrinsic::aarch64_sve_sunpkhi:
return DAG.getNode(AArch64ISD::SUNPKHI, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_sve_sunpklo:
return DAG.getNode(AArch64ISD::SUNPKLO, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_sve_uunpkhi:
return DAG.getNode(AArch64ISD::UUNPKHI, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_sve_uunpklo:
return DAG.getNode(AArch64ISD::UUNPKLO, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_sve_clasta_n:
return DAG.getNode(AArch64ISD::CLASTA_N, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2), Op.getOperand(3));
case Intrinsic::aarch64_sve_clastb_n:
return DAG.getNode(AArch64ISD::CLASTB_N, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2), Op.getOperand(3));
case Intrinsic::aarch64_sve_lasta:
return DAG.getNode(AArch64ISD::LASTA, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_lastb:
return DAG.getNode(AArch64ISD::LASTB, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_rev:
return DAG.getNode(ISD::VECTOR_REVERSE, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_sve_tbl:
return DAG.getNode(AArch64ISD::TBL, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
case Intrinsic::aarch64_sve_trn1:
return DAG.getNode(AArch64ISD::TRN1, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_trn2:
return DAG.getNode(AArch64ISD::TRN2, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_uzp1:
return DAG.getNode(AArch64ISD::UZP1, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_uzp2:
return DAG.getNode(AArch64ISD::UZP2, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_zip1:
return DAG.getNode(AArch64ISD::ZIP1, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_zip2:
return DAG.getNode(AArch64ISD::ZIP2, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_splice:
return DAG.getNode(AArch64ISD::SPLICE, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2), Op.getOperand(3));
case Intrinsic::aarch64_sve_ptrue:
return getPTrue(DAG, DL, Op.getValueType(), Op.getConstantOperandVal(1));
case Intrinsic::aarch64_sve_clz:
return DAG.getNode(AArch64ISD::CTLZ_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sme_cntsb:
return DAG.getNode(AArch64ISD::RDSVL, DL, Op.getValueType(),
DAG.getConstant(1, DL, MVT::i32));
case Intrinsic::aarch64_sme_cntsh: {
SDValue One = DAG.getConstant(1, DL, MVT::i32);
SDValue Bytes = DAG.getNode(AArch64ISD::RDSVL, DL, Op.getValueType(), One);
return DAG.getNode(ISD::SRL, DL, Op.getValueType(), Bytes, One);
}
case Intrinsic::aarch64_sme_cntsw: {
SDValue Bytes = DAG.getNode(AArch64ISD::RDSVL, DL, Op.getValueType(),
DAG.getConstant(1, DL, MVT::i32));
return DAG.getNode(ISD::SRL, DL, Op.getValueType(), Bytes,
DAG.getConstant(2, DL, MVT::i32));
}
case Intrinsic::aarch64_sme_cntsd: {
SDValue Bytes = DAG.getNode(AArch64ISD::RDSVL, DL, Op.getValueType(),
DAG.getConstant(1, DL, MVT::i32));
return DAG.getNode(ISD::SRL, DL, Op.getValueType(), Bytes,
DAG.getConstant(3, DL, MVT::i32));
}
case Intrinsic::aarch64_sve_cnt: {
SDValue Data = Op.getOperand(3);
// CTPOP only supports integer operands.
if (Data.getValueType().isFloatingPoint())
Data = DAG.getNode(ISD::BITCAST, DL, Op.getValueType(), Data);
return DAG.getNode(AArch64ISD::CTPOP_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Data, Op.getOperand(1));
}
case Intrinsic::aarch64_sve_dupq_lane:
return LowerDUPQLane(Op, DAG);
case Intrinsic::aarch64_sve_convert_from_svbool:
if (Op.getValueType() == MVT::aarch64svcount)
return DAG.getNode(ISD::BITCAST, DL, Op.getValueType(), Op.getOperand(1));
return getSVEPredicateBitCast(Op.getValueType(), Op.getOperand(1), DAG);
case Intrinsic::aarch64_sve_convert_to_svbool:
if (Op.getOperand(1).getValueType() == MVT::aarch64svcount)
return DAG.getNode(ISD::BITCAST, DL, MVT::nxv16i1, Op.getOperand(1));
return getSVEPredicateBitCast(MVT::nxv16i1, Op.getOperand(1), DAG);
case Intrinsic::aarch64_sve_fneg:
return DAG.getNode(AArch64ISD::FNEG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_frintp:
return DAG.getNode(AArch64ISD::FCEIL_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_frintm:
return DAG.getNode(AArch64ISD::FFLOOR_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_frinti:
return DAG.getNode(AArch64ISD::FNEARBYINT_MERGE_PASSTHRU, DL,
Op.getValueType(), Op.getOperand(2), Op.getOperand(3),
Op.getOperand(1));
case Intrinsic::aarch64_sve_frintx:
return DAG.getNode(AArch64ISD::FRINT_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_frinta:
return DAG.getNode(AArch64ISD::FROUND_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_frintn:
return DAG.getNode(AArch64ISD::FROUNDEVEN_MERGE_PASSTHRU, DL,
Op.getValueType(), Op.getOperand(2), Op.getOperand(3),
Op.getOperand(1));
case Intrinsic::aarch64_sve_frintz:
return DAG.getNode(AArch64ISD::FTRUNC_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_ucvtf:
return DAG.getNode(AArch64ISD::UINT_TO_FP_MERGE_PASSTHRU, DL,
Op.getValueType(), Op.getOperand(2), Op.getOperand(3),
Op.getOperand(1));
case Intrinsic::aarch64_sve_scvtf:
return DAG.getNode(AArch64ISD::SINT_TO_FP_MERGE_PASSTHRU, DL,
Op.getValueType(), Op.getOperand(2), Op.getOperand(3),
Op.getOperand(1));
case Intrinsic::aarch64_sve_fcvtzu:
return DAG.getNode(AArch64ISD::FCVTZU_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_fcvtzs:
return DAG.getNode(AArch64ISD::FCVTZS_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_fsqrt:
return DAG.getNode(AArch64ISD::FSQRT_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_frecpx:
return DAG.getNode(AArch64ISD::FRECPX_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_frecpe_x:
return DAG.getNode(AArch64ISD::FRECPE, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_sve_frecps_x:
return DAG.getNode(AArch64ISD::FRECPS, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_frsqrte_x:
return DAG.getNode(AArch64ISD::FRSQRTE, DL, Op.getValueType(),
Op.getOperand(1));
case Intrinsic::aarch64_sve_frsqrts_x:
return DAG.getNode(AArch64ISD::FRSQRTS, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2));
case Intrinsic::aarch64_sve_fabs:
return DAG.getNode(AArch64ISD::FABS_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_abs:
return DAG.getNode(AArch64ISD::ABS_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_neg:
return DAG.getNode(AArch64ISD::NEG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_insr: {
SDValue Scalar = Op.getOperand(2);
EVT ScalarTy = Scalar.getValueType();
if ((ScalarTy == MVT::i8) || (ScalarTy == MVT::i16))
Scalar = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, Scalar);
return DAG.getNode(AArch64ISD::INSR, DL, Op.getValueType(),
Op.getOperand(1), Scalar);
}
case Intrinsic::aarch64_sve_rbit:
return DAG.getNode(AArch64ISD::BITREVERSE_MERGE_PASSTHRU, DL,
Op.getValueType(), Op.getOperand(2), Op.getOperand(3),
Op.getOperand(1));
case Intrinsic::aarch64_sve_revb:
return DAG.getNode(AArch64ISD::BSWAP_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_revh:
return DAG.getNode(AArch64ISD::REVH_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_revw:
return DAG.getNode(AArch64ISD::REVW_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_revd:
return DAG.getNode(AArch64ISD::REVD_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3), Op.getOperand(1));
case Intrinsic::aarch64_sve_sxtb:
return DAG.getNode(
AArch64ISD::SIGN_EXTEND_INREG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3),
DAG.getValueType(Op.getValueType().changeVectorElementType(MVT::i8)),
Op.getOperand(1));
case Intrinsic::aarch64_sve_sxth:
return DAG.getNode(
AArch64ISD::SIGN_EXTEND_INREG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3),
DAG.getValueType(Op.getValueType().changeVectorElementType(MVT::i16)),
Op.getOperand(1));
case Intrinsic::aarch64_sve_sxtw:
return DAG.getNode(
AArch64ISD::SIGN_EXTEND_INREG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3),
DAG.getValueType(Op.getValueType().changeVectorElementType(MVT::i32)),
Op.getOperand(1));
case Intrinsic::aarch64_sve_uxtb:
return DAG.getNode(
AArch64ISD::ZERO_EXTEND_INREG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3),
DAG.getValueType(Op.getValueType().changeVectorElementType(MVT::i8)),
Op.getOperand(1));
case Intrinsic::aarch64_sve_uxth:
return DAG.getNode(
AArch64ISD::ZERO_EXTEND_INREG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3),
DAG.getValueType(Op.getValueType().changeVectorElementType(MVT::i16)),
Op.getOperand(1));
case Intrinsic::aarch64_sve_uxtw:
return DAG.getNode(
AArch64ISD::ZERO_EXTEND_INREG_MERGE_PASSTHRU, DL, Op.getValueType(),
Op.getOperand(2), Op.getOperand(3),
DAG.getValueType(Op.getValueType().changeVectorElementType(MVT::i32)),
Op.getOperand(1));
case Intrinsic::localaddress: {
const auto &MF = DAG.getMachineFunction();
const auto *RegInfo = Subtarget->getRegisterInfo();
unsigned Reg = RegInfo->getLocalAddressRegister(MF);
return DAG.getCopyFromReg(DAG.getEntryNode(), DL, Reg,
Op.getSimpleValueType());
}
case Intrinsic::eh_recoverfp: {
// FIXME: This needs to be implemented to correctly handle highly aligned
// stack objects. For now we simply return the incoming FP. Refer D53541
// for more details.
SDValue FnOp = Op.getOperand(1);
SDValue IncomingFPOp = Op.getOperand(2);
GlobalAddressSDNode *GSD = dyn_cast<GlobalAddressSDNode>(FnOp);
auto *Fn = dyn_cast_or_null<Function>(GSD ? GSD->getGlobal() : nullptr);
if (!Fn)
report_fatal_error(
"llvm.eh.recoverfp must take a function as the first argument");
return IncomingFPOp;
}
case Intrinsic::aarch64_neon_vsri:
case Intrinsic::aarch64_neon_vsli:
case Intrinsic::aarch64_sve_sri:
case Intrinsic::aarch64_sve_sli: {
EVT Ty = Op.getValueType();
if (!Ty.isVector())
report_fatal_error("Unexpected type for aarch64_neon_vsli");
assert(Op.getConstantOperandVal(3) <= Ty.getScalarSizeInBits());
bool IsShiftRight = IntNo == Intrinsic::aarch64_neon_vsri ||
IntNo == Intrinsic::aarch64_sve_sri;
unsigned Opcode = IsShiftRight ? AArch64ISD::VSRI : AArch64ISD::VSLI;
return DAG.getNode(Opcode, DL, Ty, Op.getOperand(1), Op.getOperand(2),
Op.getOperand(3));
}
case Intrinsic::aarch64_neon_srhadd:
case Intrinsic::aarch64_neon_urhadd:
case Intrinsic::aarch64_neon_shadd:
case Intrinsic::aarch64_neon_uhadd: {
bool IsSignedAdd = (IntNo == Intrinsic::aarch64_neon_srhadd ||
IntNo == Intrinsic::aarch64_neon_shadd);
bool IsRoundingAdd = (IntNo == Intrinsic::aarch64_neon_srhadd ||
IntNo == Intrinsic::aarch64_neon_urhadd);
unsigned Opcode = IsSignedAdd
? (IsRoundingAdd ? ISD::AVGCEILS : ISD::AVGFLOORS)
: (IsRoundingAdd ? ISD::AVGCEILU : ISD::AVGFLOORU);
return DAG.getNode(Opcode, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2));
}
case Intrinsic::aarch64_neon_saddlp:
case Intrinsic::aarch64_neon_uaddlp: {
unsigned Opcode = IntNo == Intrinsic::aarch64_neon_uaddlp
? AArch64ISD::UADDLP
: AArch64ISD::SADDLP;
return DAG.getNode(Opcode, DL, Op.getValueType(), Op.getOperand(1));
}
case Intrinsic::aarch64_neon_sdot:
case Intrinsic::aarch64_neon_udot:
case Intrinsic::aarch64_sve_sdot:
case Intrinsic::aarch64_sve_udot: {
unsigned Opcode = (IntNo == Intrinsic::aarch64_neon_udot ||
IntNo == Intrinsic::aarch64_sve_udot)
? AArch64ISD::UDOT
: AArch64ISD::SDOT;
return DAG.getNode(Opcode, DL, Op.getValueType(), Op.getOperand(1),
Op.getOperand(2), Op.getOperand(3));
}
case Intrinsic::aarch64_neon_usdot:
case Intrinsic::aarch64_sve_usdot: {
return DAG.getNode(AArch64ISD::USDOT, DL, Op.getValueType(),
Op.getOperand(1), Op.getOperand(2), Op.getOperand(3));
}
case Intrinsic::aarch64_neon_saddlv:
case Intrinsic::aarch64_neon_uaddlv: {
EVT OpVT = Op.getOperand(1).getValueType();
EVT ResVT = Op.getValueType();
assert(
((ResVT == MVT::i32 && (OpVT == MVT::v8i8 || OpVT == MVT::v16i8 ||
OpVT == MVT::v8i16 || OpVT == MVT::v4i16)) ||
(ResVT == MVT::i64 && (OpVT == MVT::v4i32 || OpVT == MVT::v2i32))) &&
"Unexpected aarch64_neon_u/saddlv type");
(void)OpVT;
// In order to avoid insert_subvector, use v4i32 rather than v2i32.
SDValue ADDLV = DAG.getNode(
IntNo == Intrinsic::aarch64_neon_uaddlv ? AArch64ISD::UADDLV
: AArch64ISD::SADDLV,
DL, ResVT == MVT::i32 ? MVT::v4i32 : MVT::v2i64, Op.getOperand(1));
SDValue EXTRACT_VEC_ELT = DAG.getNode(
ISD::EXTRACT_VECTOR_ELT, DL, ResVT == MVT::i32 ? MVT::i32 : MVT::i64,
ADDLV, DAG.getConstant(0, DL, MVT::i64));
return EXTRACT_VEC_ELT;
}
case Intrinsic::experimental_cttz_elts: {
SDValue CttzOp = Op.getOperand(1);
EVT VT = CttzOp.getValueType();
assert(VT.getVectorElementType() == MVT::i1 && "Expected MVT::i1");
if (VT.isFixedLengthVector()) {
// We can use SVE instructions to lower this intrinsic by first creating
// an SVE predicate register mask from the fixed-width vector.
EVT NewVT = getTypeToTransformTo(*DAG.getContext(), VT);
SDValue Mask = DAG.getNode(ISD::SIGN_EXTEND, DL, NewVT, CttzOp);
CttzOp = convertFixedMaskToScalableVector(Mask, DAG);
}
SDValue NewCttzElts =
DAG.getNode(AArch64ISD::CTTZ_ELTS, DL, MVT::i64, CttzOp);
return DAG.getZExtOrTrunc(NewCttzElts, DL, Op.getValueType());
}
case Intrinsic::experimental_vector_match: {
return LowerVectorMatch(Op, DAG);
}
}
}
bool AArch64TargetLowering::shouldExtendGSIndex(EVT VT, EVT &EltTy) const {
if (VT.getVectorElementType() == MVT::i8 ||
VT.getVectorElementType() == MVT::i16) {
EltTy = MVT::i32;
return true;
}
return false;
}
bool AArch64TargetLowering::shouldRemoveExtendFromGSIndex(SDValue Extend,
EVT DataVT) const {
const EVT IndexVT = Extend.getOperand(0).getValueType();
// SVE only supports implicit extension of 32-bit indices.
if (!Subtarget->hasSVE() || IndexVT.getVectorElementType() != MVT::i32)
return false;
// Indices cannot be smaller than the main data type.
if (IndexVT.getScalarSizeInBits() < DataVT.getScalarSizeInBits())
return false;
// Scalable vectors with "vscale * 2" or fewer elements sit within a 64-bit
// element container type, which would violate the previous clause.
return DataVT.isFixedLengthVector() || DataVT.getVectorMinNumElements() > 2;
}
bool AArch64TargetLowering::isVectorLoadExtDesirable(SDValue ExtVal) const {
EVT ExtVT = ExtVal.getValueType();
if (!ExtVT.isScalableVector() && !Subtarget->useSVEForFixedLengthVectors())
return false;
// It may be worth creating extending masked loads if there are multiple
// masked loads using the same predicate. That way we'll end up creating
// extending masked loads that may then get split by the legaliser. This
// results in just one set of predicate unpacks at the start, instead of
// multiple sets of vector unpacks after each load.
if (auto *Ld = dyn_cast<MaskedLoadSDNode>(ExtVal->getOperand(0))) {
if (!isLoadExtLegalOrCustom(ISD::ZEXTLOAD, ExtVT, Ld->getValueType(0))) {
// Disable extending masked loads for fixed-width for now, since the code
// quality doesn't look great.
if (!ExtVT.isScalableVector())
return false;
unsigned NumExtMaskedLoads = 0;
for (auto *U : Ld->getMask()->users())
if (isa<MaskedLoadSDNode>(U))
NumExtMaskedLoads++;
if (NumExtMaskedLoads <= 1)
return false;
}
}
return true;
}
unsigned getGatherVecOpcode(bool IsScaled, bool IsSigned, bool NeedsExtend) {
std::map<std::tuple<bool, bool, bool>, unsigned> AddrModes = {
{std::make_tuple(/*Scaled*/ false, /*Signed*/ false, /*Extend*/ false),
AArch64ISD::GLD1_MERGE_ZERO},
{std::make_tuple(/*Scaled*/ false, /*Signed*/ false, /*Extend*/ true),
AArch64ISD::GLD1_UXTW_MERGE_ZERO},
{std::make_tuple(/*Scaled*/ false, /*Signed*/ true, /*Extend*/ false),
AArch64ISD::GLD1_MERGE_ZERO},
{std::make_tuple(/*Scaled*/ false, /*Signed*/ true, /*Extend*/ true),
AArch64ISD::GLD1_SXTW_MERGE_ZERO},
{std::make_tuple(/*Scaled*/ true, /*Signed*/ false, /*Extend*/ false),
AArch64ISD::GLD1_SCALED_MERGE_ZERO},
{std::make_tuple(/*Scaled*/ true, /*Signed*/ false, /*Extend*/ true),
AArch64ISD::GLD1_UXTW_SCALED_MERGE_ZERO},
{std::make_tuple(/*Scaled*/ true, /*Signed*/ true, /*Extend*/ false),
AArch64ISD::GLD1_SCALED_MERGE_ZERO},
{std::make_tuple(/*Scaled*/ true, /*Signed*/ true, /*Extend*/ true),
AArch64ISD::GLD1_SXTW_SCALED_MERGE_ZERO},
};
auto Key = std::make_tuple(IsScaled, IsSigned, NeedsExtend);
return AddrModes.find(Key)->second;
}
unsigned getSignExtendedGatherOpcode(unsigned Opcode) {
switch (Opcode) {
default:
llvm_unreachable("unimplemented opcode");
return Opcode;
case AArch64ISD::GLD1_MERGE_ZERO:
return AArch64ISD::GLD1S_MERGE_ZERO;
case AArch64ISD::GLD1_IMM_MERGE_ZERO:
return AArch64ISD::GLD1S_IMM_MERGE_ZERO;
case AArch64ISD::GLD1_UXTW_MERGE_ZERO:
return AArch64ISD::GLD1S_UXTW_MERGE_ZERO;
case AArch64ISD::GLD1_SXTW_MERGE_ZERO:
return AArch64ISD::GLD1S_SXTW_MERGE_ZERO;
case AArch64ISD::GLD1_SCALED_MERGE_ZERO:
return AArch64ISD::GLD1S_SCALED_MERGE_ZERO;
case AArch64ISD::GLD1_UXTW_SCALED_MERGE_ZERO:
return AArch64ISD::GLD1S_UXTW_SCALED_MERGE_ZERO;
case AArch64ISD::GLD1_SXTW_SCALED_MERGE_ZERO:
return AArch64ISD::GLD1S_SXTW_SCALED_MERGE_ZERO;
}
}
SDValue AArch64TargetLowering::LowerMGATHER(SDValue Op,
SelectionDAG &DAG) const {
MaskedGatherSDNode *MGT = cast<MaskedGatherSDNode>(Op);
SDLoc DL(Op);
SDValue Chain = MGT->getChain();
SDValue PassThru = MGT->getPassThru();
SDValue Mask = MGT->getMask();
SDValue BasePtr = MGT->getBasePtr();
SDValue Index = MGT->getIndex();
SDValue Scale = MGT->getScale();
EVT VT = Op.getValueType();
EVT MemVT = MGT->getMemoryVT();
ISD::LoadExtType ExtType = MGT->getExtensionType();
ISD::MemIndexType IndexType = MGT->getIndexType();
// SVE supports zero (and so undef) passthrough values only, everything else
// must be handled manually by an explicit select on the load's output.
if (!PassThru->isUndef() && !isZerosVector(PassThru.getNode())) {
SDValue Ops[] = {Chain, DAG.getUNDEF(VT), Mask, BasePtr, Index, Scale};
SDValue Load =
DAG.getMaskedGather(MGT->getVTList(), MemVT, DL, Ops,
MGT->getMemOperand(), IndexType, ExtType);
SDValue Select = DAG.getSelect(DL, VT, Mask, Load, PassThru);
return DAG.getMergeValues({Select, Load.getValue(1)}, DL);
}
bool IsScaled = MGT->isIndexScaled();
bool IsSigned = MGT->isIndexSigned();
// SVE supports an index scaled by sizeof(MemVT.elt) only, everything else
// must be calculated before hand.
uint64_t ScaleVal = Scale->getAsZExtVal();
if (IsScaled && ScaleVal != MemVT.getScalarStoreSize()) {
assert(isPowerOf2_64(ScaleVal) && "Expecting power-of-two types");
EVT IndexVT = Index.getValueType();
Index = DAG.getNode(ISD::SHL, DL, IndexVT, Index,
DAG.getConstant(Log2_32(ScaleVal), DL, IndexVT));
Scale = DAG.getTargetConstant(1, DL, Scale.getValueType());
SDValue Ops[] = {Chain, PassThru, Mask, BasePtr, Index, Scale};
return DAG.getMaskedGather(MGT->getVTList(), MemVT, DL, Ops,
MGT->getMemOperand(), IndexType, ExtType);
}
// Lower fixed length gather to a scalable equivalent.
if (VT.isFixedLengthVector()) {
assert(Subtarget->useSVEForFixedLengthVectors() &&
"Cannot lower when not using SVE for fixed vectors!");
// NOTE: Handle floating-point as if integer then bitcast the result.
EVT DataVT = VT.changeVectorElementTypeToInteger();
MemVT = MemVT.changeVectorElementTypeToInteger();
// Find the smallest integer fixed length vector we can use for the gather.
EVT PromotedVT = VT.changeVectorElementType(MVT::i32);
if (DataVT.getVectorElementType() == MVT::i64 ||
Index.getValueType().getVectorElementType() == MVT::i64 ||
Mask.getValueType().getVectorElementType() == MVT::i64)
PromotedVT = VT.changeVectorElementType(MVT::i64);
// Promote vector operands except for passthrough, which we know is either
// undef or zero, and thus best constructed directly.
unsigned ExtOpcode = IsSigned ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND;
Index = DAG.getNode(ExtOpcode, DL, PromotedVT, Index);
Mask = DAG.getNode(ISD::SIGN_EXTEND, DL, PromotedVT, Mask);
// A promoted result type forces the need for an extending load.
if (PromotedVT != DataVT && ExtType == ISD::NON_EXTLOAD)
ExtType = ISD::EXTLOAD;
EVT ContainerVT = getContainerForFixedLengthVector(DAG, PromotedVT);
// Convert fixed length vector operands to scalable.
MemVT = ContainerVT.changeVectorElementType(MemVT.getVectorElementType());
Index = convertToScalableVector(DAG, ContainerVT, Index);
Mask = convertFixedMaskToScalableVector(Mask, DAG);
PassThru = PassThru->isUndef() ? DAG.getUNDEF(ContainerVT)
: DAG.getConstant(0, DL, ContainerVT);
// Emit equivalent scalable vector gather.
SDValue Ops[] = {Chain, PassThru, Mask, BasePtr, Index, Scale};
SDValue Load =
DAG.getMaskedGather(DAG.getVTList(ContainerVT, MVT::Other), MemVT, DL,
Ops, MGT->getMemOperand(), IndexType, ExtType);
// Extract fixed length data then convert to the required result type.
SDValue Result = convertFromScalableVector(DAG, PromotedVT, Load);
Result = DAG.getNode(ISD::TRUNCATE, DL, DataVT, Result);
if (VT.isFloatingPoint())
Result = DAG.getNode(ISD::BITCAST, DL, VT, Result);
return DAG.getMergeValues({Result, Load.getValue(1)}, DL);
}
// Everything else is legal.
return Op;
}
SDValue AArch64TargetLowering::LowerMSCATTER(SDValue Op,
SelectionDAG &DAG) const {
MaskedScatterSDNode *MSC = cast<MaskedScatterSDNode>(Op);
SDLoc DL(Op);
SDValue Chain = MSC->getChain();
SDValue StoreVal = MSC->getValue();
SDValue Mask = MSC->getMask();
SDValue BasePtr = MSC->getBasePtr();
SDValue Index = MSC->getIndex();
SDValue Scale = MSC->getScale();
EVT VT = StoreVal.getValueType();
EVT MemVT = MSC->getMemoryVT();
ISD::MemIndexType IndexType = MSC->getIndexType();
bool Truncating = MSC->isTruncatingStore();
bool IsScaled = MSC->isIndexScaled();
bool IsSigned = MSC->isIndexSigned();
// SVE supports an index scaled by sizeof(MemVT.elt) only, everything else
// must be calculated before hand.
uint64_t ScaleVal = Scale->getAsZExtVal();
if (IsScaled && ScaleVal != MemVT.getScalarStoreSize()) {
assert(isPowerOf2_64(ScaleVal) && "Expecting power-of-two types");
EVT IndexVT = Index.getValueType();
Index = DAG.getNode(ISD::SHL, DL, IndexVT, Index,
DAG.getConstant(Log2_32(ScaleVal), DL, IndexVT));
Scale = DAG.getTargetConstant(1, DL, Scale.getValueType());
SDValue Ops[] = {Chain, StoreVal, Mask, BasePtr, Index, Scale};
return DAG.getMaskedScatter(MSC->getVTList(), MemVT, DL, Ops,
MSC->getMemOperand(), IndexType, Truncating);
}
// Lower fixed length scatter to a scalable equivalent.
if (VT.isFixedLengthVector()) {
assert(Subtarget->useSVEForFixedLengthVectors() &&
"Cannot lower when not using SVE for fixed vectors!");
// Once bitcast we treat floating-point scatters as if integer.
if (VT.isFloatingPoint()) {
VT = VT.changeVectorElementTypeToInteger();
MemVT = MemVT.changeVectorElementTypeToInteger();
StoreVal = DAG.getNode(ISD::BITCAST, DL, VT, StoreVal);
}
// Find the smallest integer fixed length vector we can use for the scatter.
EVT PromotedVT = VT.changeVectorElementType(MVT::i32);
if (VT.getVectorElementType() == MVT::i64 ||
Index.getValueType().getVectorElementType() == MVT::i64 ||
Mask.getValueType().getVectorElementType() == MVT::i64)
PromotedVT = VT.changeVectorElementType(MVT::i64);
// Promote vector operands.
unsigned ExtOpcode = IsSigned ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND;
Index = DAG.getNode(ExtOpcode, DL, PromotedVT, Index);
Mask = DAG.getNode(ISD::SIGN_EXTEND, DL, PromotedVT, Mask);
StoreVal = DAG.getNode(ISD::ANY_EXTEND, DL, PromotedVT, StoreVal);
// A promoted value type forces the need for a truncating store.
if (PromotedVT != VT)
Truncating = true;
EVT ContainerVT = getContainerForFixedLengthVector(DAG, PromotedVT);
// Convert fixed length vector operands to scalable.
MemVT = ContainerVT.changeVectorElementType(MemVT.getVectorElementType());
Index = convertToScalableVector(DAG, ContainerVT, Index);
Mask = convertFixedMaskToScalableVector(Mask, DAG);
StoreVal = convertToScalableVector(DAG, ContainerVT, StoreVal);
// Emit equivalent scalable vector scatter.
SDValue Ops[] = {Chain, StoreVal, Mask, BasePtr, Index, Scale};
return DAG.getMaskedScatter(MSC->getVTList(), MemVT, DL, Ops,
MSC->getMemOperand(), IndexType, Truncating);
}
// Everything else is legal.
return Op;
}
SDValue AArch64TargetLowering::LowerMLOAD(SDValue Op, SelectionDAG &DAG) const {
SDLoc DL(Op);
MaskedLoadSDNode *LoadNode = cast<MaskedLoadSDNode>(Op);
assert(LoadNode && "Expected custom lowering of a masked load node");
EVT VT = Op->getValueType(0);
if (useSVEForFixedLengthVectorVT(VT, /*OverrideNEON=*/true))
return LowerFixedLengthVectorMLoadToSVE(Op, DAG);
SDValue PassThru = LoadNode->getPassThru();
SDValue Mask = LoadNode->getMask();
if (PassThru->isUndef() || isZerosVector(PassThru.getNode()))
return Op;
SDValue Load = DAG.getMaskedLoad(
VT, DL, LoadNode->getChain(), LoadNode->getBasePtr(),
LoadNode->getOffset(), Mask, DAG.getUNDEF(VT), LoadNode->getMemoryVT(),
LoadNode->getMemOperand(), LoadNode->getAddressingMode(),
LoadNode->getExtensionType());
SDValue Result = DAG.getSelect(DL, VT, Mask, Load, PassThru);
return DAG.getMergeValues({Result, Load.getValue(1)}, DL);
}
// Custom lower trunc store for v4i8 vectors, since it is promoted to v4i16.
static SDValue LowerTruncateVectorStore(SDLoc DL, StoreSDNode *ST,
EVT VT, EVT MemVT,
SelectionDAG &DAG) {
assert(VT.isVector() && "VT should be a vector type");
assert(MemVT == MVT::v4i8 && VT == MVT::v4i16);
SDValue Value = ST->getValue();
// It first extend the promoted v4i16 to v8i16, truncate to v8i8, and extract
// the word lane which represent the v4i8 subvector. It optimizes the store
// to:
//
// xtn v0.8b, v0.8h
// str s0, [x0]
SDValue Undef = DAG.getUNDEF(MVT::i16);
SDValue UndefVec = DAG.getBuildVector(MVT::v4i16, DL,
{Undef, Undef, Undef, Undef});
SDValue TruncExt = DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v8i16,
Value, UndefVec);
SDValue Trunc = DAG.getNode(ISD::TRUNCATE, DL, MVT::v8i8, TruncExt);
Trunc = DAG.getNode(ISD::BITCAST, DL, MVT::v2i32, Trunc);
SDValue ExtractTrunc = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i32,
Trunc, DAG.getConstant(0, DL, MVT::i64));
return DAG.getStore(ST->getChain(), DL, ExtractTrunc,
ST->getBasePtr(), ST->getMemOperand());
}
static SDValue LowerADDRSPACECAST(SDValue Op, SelectionDAG &DAG) {
SDLoc DL(Op);
SDValue Src = Op.getOperand(0);
MVT DestVT = Op.getSimpleValueType();
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
AddrSpaceCastSDNode *N = cast<AddrSpaceCastSDNode>(Op.getNode());
unsigned SrcAS = N->getSrcAddressSpace();
unsigned DestAS = N->getDestAddressSpace();
assert(SrcAS != DestAS &&
"addrspacecast must be between different address spaces");
assert(TLI.getTargetMachine().getPointerSize(SrcAS) !=
TLI.getTargetMachine().getPointerSize(DestAS) &&
"addrspacecast must be between different ptr sizes");
(void)TLI;
if (SrcAS == ARM64AS::PTR32_SPTR) {
return DAG.getNode(ISD::SIGN_EXTEND, DL, DestVT, Src,
DAG.getTargetConstant(0, DL, DestVT));
} else if (SrcAS == ARM64AS::PTR32_UPTR) {
return DAG.getNode(ISD::ZERO_EXTEND, DL, DestVT, Src,
DAG.getTargetConstant(0, DL, DestVT));
} else if ((DestAS == ARM64AS::PTR32_SPTR) ||
(DestAS == ARM64AS::PTR32_UPTR)) {
SDValue Ext = DAG.getAnyExtOrTrunc(Src, DL, DestVT);
SDValue Trunc = DAG.getZeroExtendInReg(Ext, DL, DestVT);
return Trunc;
} else {
return Src;
}
}
// Custom lowering for any store, vector or scalar and/or default or with
// a truncate operations. Currently only custom lower truncate operation
// from vector v4i16 to v4i8 or volatile stores of i128.
SDValue AArch64TargetLowering::LowerSTORE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc Dl(Op);
StoreSDNode *StoreNode = cast<StoreSDNode>(Op);
assert (StoreNode && "Can only custom lower store nodes");
SDValue Value = StoreNode->getValue();
EVT VT = Value.getValueType();
EVT MemVT = StoreNode->getMemoryVT();
if (VT.isVector()) {
if (useSVEForFixedLengthVectorVT(
VT,
/*OverrideNEON=*/Subtarget->useSVEForFixedLengthVectors()))
return LowerFixedLengthVectorStoreToSVE(Op, DAG);
unsigned AS = StoreNode->getAddressSpace();
Align Alignment = StoreNode->getAlign();
if (Alignment < MemVT.getStoreSize() &&
!allowsMisalignedMemoryAccesses(MemVT, AS, Alignment,
StoreNode->getMemOperand()->getFlags(),
nullptr)) {
return scalarizeVectorStore(StoreNode, DAG);
}
if (StoreNode->isTruncatingStore() && VT == MVT::v4i16 &&
MemVT == MVT::v4i8) {
return LowerTruncateVectorStore(Dl, StoreNode, VT, MemVT, DAG);
}
// 256 bit non-temporal stores can be lowered to STNP. Do this as part of
// the custom lowering, as there are no un-paired non-temporal stores and
// legalization will break up 256 bit inputs.
ElementCount EC = MemVT.getVectorElementCount();
if (StoreNode->isNonTemporal() && MemVT.getSizeInBits() == 256u &&
EC.isKnownEven() && DAG.getDataLayout().isLittleEndian() &&
(MemVT.getScalarSizeInBits() == 8u ||
MemVT.getScalarSizeInBits() == 16u ||
MemVT.getScalarSizeInBits() == 32u ||
MemVT.getScalarSizeInBits() == 64u)) {
SDValue Lo =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, Dl,
MemVT.getHalfNumVectorElementsVT(*DAG.getContext()),
StoreNode->getValue(), DAG.getConstant(0, Dl, MVT::i64));
SDValue Hi =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, Dl,
MemVT.getHalfNumVectorElementsVT(*DAG.getContext()),
StoreNode->getValue(),
DAG.getConstant(EC.getKnownMinValue() / 2, Dl, MVT::i64));
SDValue Result = DAG.getMemIntrinsicNode(
AArch64ISD::STNP, Dl, DAG.getVTList(MVT::Other),
{StoreNode->getChain(), Lo, Hi, StoreNode->getBasePtr()},
StoreNode->getMemoryVT(), StoreNode->getMemOperand());
return Result;
}
} else if (MemVT == MVT::i128 && StoreNode->isVolatile()) {
return LowerStore128(Op, DAG);
} else if (MemVT == MVT::i64x8) {
SDValue Value = StoreNode->getValue();
assert(Value->getValueType(0) == MVT::i64x8);
SDValue Chain = StoreNode->getChain();
SDValue Base = StoreNode->getBasePtr();
EVT PtrVT = Base.getValueType();
for (unsigned i = 0; i < 8; i++) {
SDValue Part = DAG.getNode(AArch64ISD::LS64_EXTRACT, Dl, MVT::i64,
Value, DAG.getConstant(i, Dl, MVT::i32));
SDValue Ptr = DAG.getNode(ISD::ADD, Dl, PtrVT, Base,
DAG.getConstant(i * 8, Dl, PtrVT));
Chain = DAG.getStore(Chain, Dl, Part, Ptr, StoreNode->getPointerInfo(),
StoreNode->getBaseAlign());
}
return Chain;
}
return SDValue();
}
/// Lower atomic or volatile 128-bit stores to a single STP instruction.
SDValue AArch64TargetLowering::LowerStore128(SDValue Op,
SelectionDAG &DAG) const {
MemSDNode *StoreNode = cast<MemSDNode>(Op);
assert(StoreNode->getMemoryVT() == MVT::i128);
assert(StoreNode->isVolatile() || StoreNode->isAtomic());
bool IsStoreRelease =
StoreNode->getMergedOrdering() == AtomicOrdering::Release;
if (StoreNode->isAtomic())
assert((Subtarget->hasFeature(AArch64::FeatureLSE2) &&
Subtarget->hasFeature(AArch64::FeatureRCPC3) && IsStoreRelease) ||
StoreNode->getMergedOrdering() == AtomicOrdering::Unordered ||
StoreNode->getMergedOrdering() == AtomicOrdering::Monotonic);
SDValue Value = (StoreNode->getOpcode() == ISD::STORE ||
StoreNode->getOpcode() == ISD::ATOMIC_STORE)
? StoreNode->getOperand(1)
: StoreNode->getOperand(2);
SDLoc DL(Op);
auto StoreValue = DAG.SplitScalar(Value, DL, MVT::i64, MVT::i64);
unsigned Opcode = IsStoreRelease ? AArch64ISD::STILP : AArch64ISD::STP;
if (DAG.getDataLayout().isBigEndian())
std::swap(StoreValue.first, StoreValue.second);
SDValue Result = DAG.getMemIntrinsicNode(
Opcode, DL, DAG.getVTList(MVT::Other),
{StoreNode->getChain(), StoreValue.first, StoreValue.second,
StoreNode->getBasePtr()},
StoreNode->getMemoryVT(), StoreNode->getMemOperand());
return Result;
}
SDValue AArch64TargetLowering::LowerLOAD(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
LoadSDNode *LoadNode = cast<LoadSDNode>(Op);
assert(LoadNode && "Expected custom lowering of a load node");
if (LoadNode->getMemoryVT() == MVT::i64x8) {
SmallVector<SDValue, 8> Ops;
SDValue Base = LoadNode->getBasePtr();
SDValue Chain = LoadNode->getChain();
EVT PtrVT = Base.getValueType();
for (unsigned i = 0; i < 8; i++) {
SDValue Ptr = DAG.getNode(ISD::ADD, DL, PtrVT, Base,
DAG.getConstant(i * 8, DL, PtrVT));
SDValue Part =
DAG.getLoad(MVT::i64, DL, Chain, Ptr, LoadNode->getPointerInfo(),
LoadNode->getBaseAlign());
Ops.push_back(Part);
Chain = SDValue(Part.getNode(), 1);
}
SDValue Loaded = DAG.getNode(AArch64ISD::LS64_BUILD, DL, MVT::i64x8, Ops);
return DAG.getMergeValues({Loaded, Chain}, DL);
}
// Custom lowering for extending v4i8 vector loads.
EVT VT = Op->getValueType(0);
assert((VT == MVT::v4i16 || VT == MVT::v4i32) && "Expected v4i16 or v4i32");
if (LoadNode->getMemoryVT() != MVT::v4i8)
return SDValue();
// Avoid generating unaligned loads.
if (Subtarget->requiresStrictAlign() && LoadNode->getAlign() < Align(4))
return SDValue();
unsigned ExtType;
if (LoadNode->getExtensionType() == ISD::SEXTLOAD)
ExtType = ISD::SIGN_EXTEND;
else if (LoadNode->getExtensionType() == ISD::ZEXTLOAD ||
LoadNode->getExtensionType() == ISD::EXTLOAD)
ExtType = ISD::ZERO_EXTEND;
else
return SDValue();
SDValue Load = DAG.getLoad(MVT::f32, DL, LoadNode->getChain(),
LoadNode->getBasePtr(), MachinePointerInfo());
SDValue Chain = Load.getValue(1);
SDValue Vec = DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, MVT::v2f32, Load);
SDValue BC = DAG.getNode(ISD::BITCAST, DL, MVT::v8i8, Vec);
SDValue Ext = DAG.getNode(ExtType, DL, MVT::v8i16, BC);
Ext = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v4i16, Ext,
DAG.getConstant(0, DL, MVT::i64));
if (VT == MVT::v4i32)
Ext = DAG.getNode(ExtType, DL, MVT::v4i32, Ext);
return DAG.getMergeValues({Ext, Chain}, DL);
}
SDValue AArch64TargetLowering::LowerVECTOR_COMPRESS(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
SDValue Vec = Op.getOperand(0);
SDValue Mask = Op.getOperand(1);
SDValue Passthru = Op.getOperand(2);
EVT VecVT = Vec.getValueType();
EVT MaskVT = Mask.getValueType();
EVT ElmtVT = VecVT.getVectorElementType();
const bool IsFixedLength = VecVT.isFixedLengthVector();
const bool HasPassthru = !Passthru.isUndef();
unsigned MinElmts = VecVT.getVectorElementCount().getKnownMinValue();
EVT FixedVecVT = MVT::getVectorVT(ElmtVT.getSimpleVT(), MinElmts);
assert(VecVT.isVector() && "Input to VECTOR_COMPRESS must be vector.");
if (!Subtarget->isSVEAvailable())
return SDValue();
if (IsFixedLength && VecVT.getSizeInBits().getFixedValue() > 128)
return SDValue();
// Only <vscale x {4|2} x {i32|i64}> supported for compact.
if (MinElmts != 2 && MinElmts != 4)
return SDValue();
// We can use the SVE register containing the NEON vector in its lowest bits.
if (IsFixedLength) {
EVT ScalableVecVT =
MVT::getScalableVectorVT(ElmtVT.getSimpleVT(), MinElmts);
EVT ScalableMaskVT = MVT::getScalableVectorVT(
MaskVT.getVectorElementType().getSimpleVT(), MinElmts);
Vec = DAG.getNode(ISD::INSERT_SUBVECTOR, DL, ScalableVecVT,
DAG.getUNDEF(ScalableVecVT), Vec,
DAG.getConstant(0, DL, MVT::i64));
Mask = DAG.getNode(ISD::INSERT_SUBVECTOR, DL, ScalableMaskVT,
DAG.getUNDEF(ScalableMaskVT), Mask,
DAG.getConstant(0, DL, MVT::i64));
Mask = DAG.getNode(ISD::TRUNCATE, DL,
ScalableMaskVT.changeVectorElementType(MVT::i1), Mask);
Passthru = DAG.getNode(ISD::INSERT_SUBVECTOR, DL, ScalableVecVT,
DAG.getUNDEF(ScalableVecVT), Passthru,
DAG.getConstant(0, DL, MVT::i64));
VecVT = Vec.getValueType();
MaskVT = Mask.getValueType();
}
// Get legal type for compact instruction
EVT ContainerVT = getSVEContainerType(VecVT);
EVT CastVT = VecVT.changeVectorElementTypeToInteger();
// Convert to i32 or i64 for smaller types, as these are the only supported
// sizes for compact.
if (ContainerVT != VecVT) {
Vec = DAG.getBitcast(CastVT, Vec);
Vec = DAG.getNode(ISD::ANY_EXTEND, DL, ContainerVT, Vec);
}
SDValue Compressed = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, Vec.getValueType(),
DAG.getConstant(Intrinsic::aarch64_sve_compact, DL, MVT::i64), Mask, Vec);
// compact fills with 0s, so if our passthru is all 0s, do nothing here.
if (HasPassthru && !ISD::isConstantSplatVectorAllZeros(Passthru.getNode())) {
SDValue Offset = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, MVT::i64,
DAG.getConstant(Intrinsic::aarch64_sve_cntp, DL, MVT::i64), Mask, Mask);
SDValue IndexMask = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, MaskVT,
DAG.getConstant(Intrinsic::aarch64_sve_whilelo, DL, MVT::i64),
DAG.getConstant(0, DL, MVT::i64), Offset);
Compressed =
DAG.getNode(ISD::VSELECT, DL, VecVT, IndexMask, Compressed, Passthru);
}
// Extracting from a legal SVE type before truncating produces better code.
if (IsFixedLength) {
Compressed = DAG.getNode(
ISD::EXTRACT_SUBVECTOR, DL,
FixedVecVT.changeVectorElementType(ContainerVT.getVectorElementType()),
Compressed, DAG.getConstant(0, DL, MVT::i64));
CastVT = FixedVecVT.changeVectorElementTypeToInteger();
VecVT = FixedVecVT;
}
// If we changed the element type before, we need to convert it back.
if (ContainerVT != VecVT) {
Compressed = DAG.getNode(ISD::TRUNCATE, DL, CastVT, Compressed);
Compressed = DAG.getBitcast(VecVT, Compressed);
}
return Compressed;
}
// Generate SUBS and CSEL for integer abs.
SDValue AArch64TargetLowering::LowerABS(SDValue Op, SelectionDAG &DAG) const {
MVT VT = Op.getSimpleValueType();
if (VT.isVector())
return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABS_MERGE_PASSTHRU);
SDLoc DL(Op);
SDValue Neg = DAG.getNode(ISD::SUB, DL, VT, DAG.getConstant(0, DL, VT),
Op.getOperand(0));
// Generate SUBS & CSEL.
SDValue Cmp =
DAG.getNode(AArch64ISD::SUBS, DL, DAG.getVTList(VT, MVT::i32),
Op.getOperand(0), DAG.getConstant(0, DL, VT));
return DAG.getNode(AArch64ISD::CSEL, DL, VT, Op.getOperand(0), Neg,
DAG.getConstant(AArch64CC::PL, DL, MVT::i32),
Cmp.getValue(1));
}
static SDValue LowerBRCOND(SDValue Op, SelectionDAG &DAG) {
SDValue Chain = Op.getOperand(0);
SDValue Cond = Op.getOperand(1);
SDValue Dest = Op.getOperand(2);
AArch64CC::CondCode CC;
if (SDValue Cmp = emitConjunction(DAG, Cond, CC)) {
SDLoc DL(Op);
SDValue CCVal = DAG.getConstant(CC, DL, MVT::i32);
return DAG.getNode(AArch64ISD::BRCOND, DL, MVT::Other, Chain, Dest, CCVal,
Cmp);
}
return SDValue();
}
// Treat FSHR with constant shifts as legal operation, otherwise it is expanded
// FSHL is converted to FSHR before deciding what to do with it
static SDValue LowerFunnelShift(SDValue Op, SelectionDAG &DAG) {
SDValue Shifts = Op.getOperand(2);
// Check if the shift amount is a constant and normalise to [0, SrcBitLen)
// If opcode is FSHL, convert it to FSHR
if (auto *ShiftNo = dyn_cast<ConstantSDNode>(Shifts)) {
SDLoc DL(Op);
MVT VT = Op.getSimpleValueType();
unsigned int NewShiftNo = ShiftNo->getZExtValue() % VT.getFixedSizeInBits();
if (Op.getOpcode() == ISD::FSHL) {
if (NewShiftNo == 0)
return Op.getOperand(0);
NewShiftNo = VT.getFixedSizeInBits() - NewShiftNo;
return DAG.getNode(
ISD::FSHR, DL, VT, Op.getOperand(0), Op.getOperand(1),
DAG.getConstant(NewShiftNo, DL, Shifts.getValueType()));
}
if (Op.getOpcode() == ISD::FSHR) {
if (NewShiftNo == 0)
return Op.getOperand(1);
if (ShiftNo->getZExtValue() == NewShiftNo)
return Op;
// Rewrite using the normalised shift amount.
return DAG.getNode(
ISD::FSHR, DL, VT, Op.getOperand(0), Op.getOperand(1),
DAG.getConstant(NewShiftNo, DL, Shifts.getValueType()));
}
}
return SDValue();
}
static SDValue LowerFLDEXP(SDValue Op, SelectionDAG &DAG) {
SDValue X = Op.getOperand(0);
EVT XScalarTy = X.getValueType();
SDValue Exp = Op.getOperand(1);
SDLoc DL(Op);
EVT XVT, ExpVT;
switch (Op.getSimpleValueType().SimpleTy) {
default:
return SDValue();
case MVT::bf16:
case MVT::f16:
X = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, X);
[[fallthrough]];
case MVT::f32:
XVT = MVT::nxv4f32;
ExpVT = MVT::nxv4i32;
break;
case MVT::f64:
XVT = MVT::nxv2f64;
ExpVT = MVT::nxv2i64;
Exp = DAG.getNode(ISD::SIGN_EXTEND, DL, MVT::i64, Exp);
break;
}
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
SDValue VX =
DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, XVT, DAG.getUNDEF(XVT), X, Zero);
SDValue VExp = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, ExpVT,
DAG.getUNDEF(ExpVT), Exp, Zero);
SDValue VPg = getPTrue(DAG, DL, XVT.changeVectorElementType(MVT::i1),
AArch64SVEPredPattern::all);
SDValue FScale =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, XVT,
DAG.getConstant(Intrinsic::aarch64_sve_fscale, DL, MVT::i64),
VPg, VX, VExp);
SDValue Final =
DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, X.getValueType(), FScale, Zero);
if (X.getValueType() != XScalarTy)
Final = DAG.getNode(ISD::FP_ROUND, DL, XScalarTy, Final,
DAG.getIntPtrConstant(1, SDLoc(Op), /*isTarget=*/true));
return Final;
}
SDValue AArch64TargetLowering::LowerADJUST_TRAMPOLINE(SDValue Op,
SelectionDAG &DAG) const {
return Op.getOperand(0);
}
SDValue AArch64TargetLowering::LowerINIT_TRAMPOLINE(SDValue Op,
SelectionDAG &DAG) const {
SDValue Chain = Op.getOperand(0);
SDValue Trmp = Op.getOperand(1); // trampoline, >=32 bytes
SDValue FPtr = Op.getOperand(2); // nested function
SDValue Nest = Op.getOperand(3); // 'nest' parameter value
const Value *TrmpAddr = cast<SrcValueSDNode>(Op.getOperand(4))->getValue();
// ldr NestReg, .+16
// ldr x17, .+20
// br x17
// .word 0
// .nest: .qword nest
// .fptr: .qword fptr
SDValue OutChains[5];
const Function *Func =
cast<Function>(cast<SrcValueSDNode>(Op.getOperand(5))->getValue());
CallingConv::ID CC = Func->getCallingConv();
unsigned NestReg;
switch (CC) {
default:
NestReg = 0x0f; // X15
break;
case CallingConv::ARM64EC_Thunk_X64:
// Must be kept in sync with AArch64CallingConv.td
NestReg = 0x04; // X4
break;
}
const char FptrReg = 0x11; // X17
SDValue Addr = Trmp;
SDLoc DL(Op);
OutChains[0] = DAG.getStore(
Chain, DL, DAG.getConstant(0x58000080u | NestReg, DL, MVT::i32), Addr,
MachinePointerInfo(TrmpAddr));
Addr = DAG.getNode(ISD::ADD, DL, MVT::i64, Trmp,
DAG.getConstant(4, DL, MVT::i64));
OutChains[1] = DAG.getStore(
Chain, DL, DAG.getConstant(0x580000b0u | FptrReg, DL, MVT::i32), Addr,
MachinePointerInfo(TrmpAddr, 4));
Addr = DAG.getNode(ISD::ADD, DL, MVT::i64, Trmp,
DAG.getConstant(8, DL, MVT::i64));
OutChains[2] =
DAG.getStore(Chain, DL, DAG.getConstant(0xd61f0220u, DL, MVT::i32), Addr,
MachinePointerInfo(TrmpAddr, 8));
Addr = DAG.getNode(ISD::ADD, DL, MVT::i64, Trmp,
DAG.getConstant(16, DL, MVT::i64));
OutChains[3] =
DAG.getStore(Chain, DL, Nest, Addr, MachinePointerInfo(TrmpAddr, 16));
Addr = DAG.getNode(ISD::ADD, DL, MVT::i64, Trmp,
DAG.getConstant(24, DL, MVT::i64));
OutChains[4] =
DAG.getStore(Chain, DL, FPtr, Addr, MachinePointerInfo(TrmpAddr, 24));
SDValue StoreToken = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, OutChains);
SDValue EndOfTrmp = DAG.getNode(ISD::ADD, DL, MVT::i64, Trmp,
DAG.getConstant(12, DL, MVT::i64));
// Call clear cache on the trampoline instructions.
return DAG.getNode(ISD::CLEAR_CACHE, DL, MVT::Other, StoreToken, Trmp,
EndOfTrmp);
}
SDValue AArch64TargetLowering::LowerOperation(SDValue Op,
SelectionDAG &DAG) const {
LLVM_DEBUG(dbgs() << "Custom lowering: ");
LLVM_DEBUG(Op.dump());
switch (Op.getOpcode()) {
default:
llvm_unreachable("unimplemented operand");
return SDValue();
case ISD::BITCAST:
return LowerBITCAST(Op, DAG);
case ISD::GlobalAddress:
return LowerGlobalAddress(Op, DAG);
case ISD::GlobalTLSAddress:
return LowerGlobalTLSAddress(Op, DAG);
case ISD::PtrAuthGlobalAddress:
return LowerPtrAuthGlobalAddress(Op, DAG);
case ISD::ADJUST_TRAMPOLINE:
return LowerADJUST_TRAMPOLINE(Op, DAG);
case ISD::INIT_TRAMPOLINE:
return LowerINIT_TRAMPOLINE(Op, DAG);
case ISD::SETCC:
case ISD::STRICT_FSETCC:
case ISD::STRICT_FSETCCS:
return LowerSETCC(Op, DAG);
case ISD::SETCCCARRY:
return LowerSETCCCARRY(Op, DAG);
case ISD::BRCOND:
return LowerBRCOND(Op, DAG);
case ISD::BR_CC:
return LowerBR_CC(Op, DAG);
case ISD::SELECT:
return LowerSELECT(Op, DAG);
case ISD::SELECT_CC:
return LowerSELECT_CC(Op, DAG);
case ISD::JumpTable:
return LowerJumpTable(Op, DAG);
case ISD::BR_JT:
return LowerBR_JT(Op, DAG);
case ISD::BRIND:
return LowerBRIND(Op, DAG);
case ISD::ConstantPool:
return LowerConstantPool(Op, DAG);
case ISD::BlockAddress:
return LowerBlockAddress(Op, DAG);
case ISD::VASTART:
return LowerVASTART(Op, DAG);
case ISD::VACOPY:
return LowerVACOPY(Op, DAG);
case ISD::VAARG:
return LowerVAARG(Op, DAG);
case ISD::UADDO_CARRY:
return lowerADDSUBO_CARRY(Op, DAG, AArch64ISD::ADCS, false /*unsigned*/);
case ISD::USUBO_CARRY:
return lowerADDSUBO_CARRY(Op, DAG, AArch64ISD::SBCS, false /*unsigned*/);
case ISD::SADDO_CARRY:
return lowerADDSUBO_CARRY(Op, DAG, AArch64ISD::ADCS, true /*signed*/);
case ISD::SSUBO_CARRY:
return lowerADDSUBO_CARRY(Op, DAG, AArch64ISD::SBCS, true /*signed*/);
case ISD::SADDO:
case ISD::UADDO:
case ISD::SSUBO:
case ISD::USUBO:
case ISD::SMULO:
case ISD::UMULO:
return LowerXALUO(Op, DAG);
case ISD::FADD:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FADD_PRED);
case ISD::FSUB:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FSUB_PRED);
case ISD::FMUL:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FMUL_PRED);
case ISD::FMA:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FMA_PRED);
case ISD::FDIV:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FDIV_PRED);
case ISD::FNEG:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FNEG_MERGE_PASSTHRU);
case ISD::FCEIL:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FCEIL_MERGE_PASSTHRU);
case ISD::FFLOOR:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FFLOOR_MERGE_PASSTHRU);
case ISD::FNEARBYINT:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FNEARBYINT_MERGE_PASSTHRU);
case ISD::FRINT:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FRINT_MERGE_PASSTHRU);
case ISD::FROUND:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FROUND_MERGE_PASSTHRU);
case ISD::FROUNDEVEN:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FROUNDEVEN_MERGE_PASSTHRU);
case ISD::FTRUNC:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FTRUNC_MERGE_PASSTHRU);
case ISD::FSQRT:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FSQRT_MERGE_PASSTHRU);
case ISD::FABS:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FABS_MERGE_PASSTHRU);
case ISD::FP_ROUND:
case ISD::STRICT_FP_ROUND:
return LowerFP_ROUND(Op, DAG);
case ISD::FP_EXTEND:
case ISD::STRICT_FP_EXTEND:
return LowerFP_EXTEND(Op, DAG);
case ISD::FRAMEADDR:
return LowerFRAMEADDR(Op, DAG);
case ISD::SPONENTRY:
return LowerSPONENTRY(Op, DAG);
case ISD::RETURNADDR:
return LowerRETURNADDR(Op, DAG);
case ISD::ADDROFRETURNADDR:
return LowerADDROFRETURNADDR(Op, DAG);
case ISD::CONCAT_VECTORS:
return LowerCONCAT_VECTORS(Op, DAG);
case ISD::INSERT_VECTOR_ELT:
return LowerINSERT_VECTOR_ELT(Op, DAG);
case ISD::EXTRACT_VECTOR_ELT:
return LowerEXTRACT_VECTOR_ELT(Op, DAG);
case ISD::BUILD_VECTOR:
return LowerBUILD_VECTOR(Op, DAG);
case ISD::ZERO_EXTEND_VECTOR_INREG:
return LowerZERO_EXTEND_VECTOR_INREG(Op, DAG);
case ISD::VECTOR_SHUFFLE:
return LowerVECTOR_SHUFFLE(Op, DAG);
case ISD::SPLAT_VECTOR:
return LowerSPLAT_VECTOR(Op, DAG);
case ISD::EXTRACT_SUBVECTOR:
return LowerEXTRACT_SUBVECTOR(Op, DAG);
case ISD::INSERT_SUBVECTOR:
return LowerINSERT_SUBVECTOR(Op, DAG);
case ISD::SDIV:
case ISD::UDIV:
return LowerDIV(Op, DAG);
case ISD::SMIN:
case ISD::UMIN:
case ISD::SMAX:
case ISD::UMAX:
return LowerMinMax(Op, DAG);
case ISD::SRA:
case ISD::SRL:
case ISD::SHL:
return LowerVectorSRA_SRL_SHL(Op, DAG);
case ISD::SHL_PARTS:
case ISD::SRL_PARTS:
case ISD::SRA_PARTS:
return LowerShiftParts(Op, DAG);
case ISD::CTPOP:
case ISD::PARITY:
return LowerCTPOP_PARITY(Op, DAG);
case ISD::FCOPYSIGN:
return LowerFCOPYSIGN(Op, DAG);
case ISD::OR:
return LowerVectorOR(Op, DAG);
case ISD::XOR:
return LowerXOR(Op, DAG);
case ISD::PREFETCH:
return LowerPREFETCH(Op, DAG);
case ISD::SINT_TO_FP:
case ISD::UINT_TO_FP:
case ISD::STRICT_SINT_TO_FP:
case ISD::STRICT_UINT_TO_FP:
return LowerINT_TO_FP(Op, DAG);
case ISD::FP_TO_SINT:
case ISD::FP_TO_UINT:
case ISD::STRICT_FP_TO_SINT:
case ISD::STRICT_FP_TO_UINT:
return LowerFP_TO_INT(Op, DAG);
case ISD::FP_TO_SINT_SAT:
case ISD::FP_TO_UINT_SAT:
return LowerFP_TO_INT_SAT(Op, DAG);
case ISD::FSINCOS:
return LowerFSINCOS(Op, DAG);
case ISD::GET_ROUNDING:
return LowerGET_ROUNDING(Op, DAG);
case ISD::SET_ROUNDING:
return LowerSET_ROUNDING(Op, DAG);
case ISD::GET_FPMODE:
return LowerGET_FPMODE(Op, DAG);
case ISD::SET_FPMODE:
return LowerSET_FPMODE(Op, DAG);
case ISD::RESET_FPMODE:
return LowerRESET_FPMODE(Op, DAG);
case ISD::MUL:
return LowerMUL(Op, DAG);
case ISD::MULHS:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::MULHS_PRED);
case ISD::MULHU:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::MULHU_PRED);
case ISD::INTRINSIC_W_CHAIN:
return LowerINTRINSIC_W_CHAIN(Op, DAG);
case ISD::INTRINSIC_WO_CHAIN:
return LowerINTRINSIC_WO_CHAIN(Op, DAG);
case ISD::INTRINSIC_VOID:
return LowerINTRINSIC_VOID(Op, DAG);
case ISD::ATOMIC_STORE:
if (cast<MemSDNode>(Op)->getMemoryVT() == MVT::i128) {
assert(Subtarget->hasLSE2() || Subtarget->hasRCPC3());
return LowerStore128(Op, DAG);
}
return SDValue();
case ISD::STORE:
return LowerSTORE(Op, DAG);
case ISD::MSTORE:
return LowerFixedLengthVectorMStoreToSVE(Op, DAG);
case ISD::MGATHER:
return LowerMGATHER(Op, DAG);
case ISD::MSCATTER:
return LowerMSCATTER(Op, DAG);
case ISD::VECREDUCE_SEQ_FADD:
return LowerVECREDUCE_SEQ_FADD(Op, DAG);
case ISD::VECREDUCE_ADD:
case ISD::VECREDUCE_AND:
case ISD::VECREDUCE_OR:
case ISD::VECREDUCE_XOR:
case ISD::VECREDUCE_SMAX:
case ISD::VECREDUCE_SMIN:
case ISD::VECREDUCE_UMAX:
case ISD::VECREDUCE_UMIN:
case ISD::VECREDUCE_FADD:
case ISD::VECREDUCE_FMAX:
case ISD::VECREDUCE_FMIN:
case ISD::VECREDUCE_FMAXIMUM:
case ISD::VECREDUCE_FMINIMUM:
return LowerVECREDUCE(Op, DAG);
case ISD::ATOMIC_LOAD_AND:
return LowerATOMIC_LOAD_AND(Op, DAG);
case ISD::DYNAMIC_STACKALLOC:
return LowerDYNAMIC_STACKALLOC(Op, DAG);
case ISD::VSCALE:
return LowerVSCALE(Op, DAG);
case ISD::VECTOR_COMPRESS:
return LowerVECTOR_COMPRESS(Op, DAG);
case ISD::ANY_EXTEND:
case ISD::SIGN_EXTEND:
case ISD::ZERO_EXTEND:
return LowerFixedLengthVectorIntExtendToSVE(Op, DAG);
case ISD::ADDRSPACECAST:
return LowerADDRSPACECAST(Op, DAG);
case ISD::SIGN_EXTEND_INREG: {
// Only custom lower when ExtraVT has a legal byte based element type.
EVT ExtraVT = cast<VTSDNode>(Op.getOperand(1))->getVT();
EVT ExtraEltVT = ExtraVT.getVectorElementType();
if ((ExtraEltVT != MVT::i8) && (ExtraEltVT != MVT::i16) &&
(ExtraEltVT != MVT::i32) && (ExtraEltVT != MVT::i64))
return SDValue();
return LowerToPredicatedOp(Op, DAG,
AArch64ISD::SIGN_EXTEND_INREG_MERGE_PASSTHRU);
}
case ISD::TRUNCATE:
return LowerTRUNCATE(Op, DAG);
case ISD::MLOAD:
return LowerMLOAD(Op, DAG);
case ISD::LOAD:
if (useSVEForFixedLengthVectorVT(Op.getValueType(),
!Subtarget->isNeonAvailable()))
return LowerFixedLengthVectorLoadToSVE(Op, DAG);
return LowerLOAD(Op, DAG);
case ISD::ADD:
case ISD::AND:
case ISD::SUB:
return LowerToScalableOp(Op, DAG);
case ISD::FMAXIMUM:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FMAX_PRED);
case ISD::FMAXNUM:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FMAXNM_PRED);
case ISD::FMINIMUM:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FMIN_PRED);
case ISD::FMINNUM:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::FMINNM_PRED);
case ISD::VSELECT:
return LowerFixedLengthVectorSelectToSVE(Op, DAG);
case ISD::ABS:
return LowerABS(Op, DAG);
case ISD::ABDS:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABDS_PRED);
case ISD::ABDU:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::ABDU_PRED);
case ISD::AVGFLOORS:
return LowerAVG(Op, DAG, AArch64ISD::HADDS_PRED);
case ISD::AVGFLOORU:
return LowerAVG(Op, DAG, AArch64ISD::HADDU_PRED);
case ISD::AVGCEILS:
return LowerAVG(Op, DAG, AArch64ISD::RHADDS_PRED);
case ISD::AVGCEILU:
return LowerAVG(Op, DAG, AArch64ISD::RHADDU_PRED);
case ISD::BITREVERSE:
return LowerBitreverse(Op, DAG);
case ISD::BSWAP:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::BSWAP_MERGE_PASSTHRU);
case ISD::CTLZ:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::CTLZ_MERGE_PASSTHRU);
case ISD::CTTZ:
return LowerCTTZ(Op, DAG);
case ISD::VECTOR_SPLICE:
return LowerVECTOR_SPLICE(Op, DAG);
case ISD::VECTOR_DEINTERLEAVE:
return LowerVECTOR_DEINTERLEAVE(Op, DAG);
case ISD::VECTOR_INTERLEAVE:
return LowerVECTOR_INTERLEAVE(Op, DAG);
case ISD::GET_ACTIVE_LANE_MASK:
return LowerGET_ACTIVE_LANE_MASK(Op, DAG);
case ISD::LRINT:
case ISD::LLRINT:
if (Op.getValueType().isVector())
return LowerVectorXRINT(Op, DAG);
[[fallthrough]];
case ISD::LROUND:
case ISD::LLROUND: {
assert((Op.getOperand(0).getValueType() == MVT::f16 ||
Op.getOperand(0).getValueType() == MVT::bf16) &&
"Expected custom lowering of rounding operations only for f16");
SDLoc DL(Op);
SDValue Ext = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, Op.getOperand(0));
return DAG.getNode(Op.getOpcode(), DL, Op.getValueType(), Ext);
}
case ISD::STRICT_LROUND:
case ISD::STRICT_LLROUND:
case ISD::STRICT_LRINT:
case ISD::STRICT_LLRINT: {
assert((Op.getOperand(1).getValueType() == MVT::f16 ||
Op.getOperand(1).getValueType() == MVT::bf16) &&
"Expected custom lowering of rounding operations only for f16");
SDLoc DL(Op);
SDValue Ext = DAG.getNode(ISD::STRICT_FP_EXTEND, DL, {MVT::f32, MVT::Other},
{Op.getOperand(0), Op.getOperand(1)});
return DAG.getNode(Op.getOpcode(), DL, {Op.getValueType(), MVT::Other},
{Ext.getValue(1), Ext.getValue(0)});
}
case ISD::WRITE_REGISTER: {
assert(Op.getOperand(2).getValueType() == MVT::i128 &&
"WRITE_REGISTER custom lowering is only for 128-bit sysregs");
SDLoc DL(Op);
SDValue Chain = Op.getOperand(0);
SDValue SysRegName = Op.getOperand(1);
std::pair<SDValue, SDValue> Pair =
DAG.SplitScalar(Op.getOperand(2), DL, MVT::i64, MVT::i64);
// chain = MSRR(chain, sysregname, lo, hi)
SDValue Result = DAG.getNode(AArch64ISD::MSRR, DL, MVT::Other, Chain,
SysRegName, Pair.first, Pair.second);
return Result;
}
case ISD::FSHL:
case ISD::FSHR:
return LowerFunnelShift(Op, DAG);
case ISD::FLDEXP:
return LowerFLDEXP(Op, DAG);
case ISD::EXPERIMENTAL_VECTOR_HISTOGRAM:
return LowerVECTOR_HISTOGRAM(Op, DAG);
case ISD::PARTIAL_REDUCE_SMLA:
case ISD::PARTIAL_REDUCE_UMLA:
case ISD::PARTIAL_REDUCE_SUMLA:
return LowerPARTIAL_REDUCE_MLA(Op, DAG);
}
}
bool AArch64TargetLowering::mergeStoresAfterLegalization(EVT VT) const {
return !Subtarget->useSVEForFixedLengthVectors();
}
bool AArch64TargetLowering::useSVEForFixedLengthVectorVT(
EVT VT, bool OverrideNEON) const {
if (!VT.isFixedLengthVector() || !VT.isSimple())
return false;
// Don't use SVE for vectors we cannot scalarize if required.
switch (VT.getVectorElementType().getSimpleVT().SimpleTy) {
// Fixed length predicates should be promoted to i8.
// NOTE: This is consistent with how NEON (and thus 64/128bit vectors) work.
case MVT::i1:
default:
return false;
case MVT::i8:
case MVT::i16:
case MVT::i32:
case MVT::i64:
case MVT::f16:
case MVT::f32:
case MVT::f64:
break;
}
// NEON-sized vectors can be emulated using SVE instructions.
if (OverrideNEON && (VT.is128BitVector() || VT.is64BitVector()))
return Subtarget->isSVEorStreamingSVEAvailable();
// Ensure NEON MVTs only belong to a single register class.
if (VT.getFixedSizeInBits() <= 128)
return false;
// Ensure wider than NEON code generation is enabled.
if (!Subtarget->useSVEForFixedLengthVectors())
return false;
// Don't use SVE for types that don't fit.
if (VT.getFixedSizeInBits() > Subtarget->getMinSVEVectorSizeInBits())
return false;
// TODO: Perhaps an artificial restriction, but worth having whilst getting
// the base fixed length SVE support in place.
if (!VT.isPow2VectorType())
return false;
return true;
}
//===----------------------------------------------------------------------===//
// Calling Convention Implementation
//===----------------------------------------------------------------------===//
static unsigned getIntrinsicID(const SDNode *N) {
unsigned Opcode = N->getOpcode();
switch (Opcode) {
default:
return Intrinsic::not_intrinsic;
case ISD::INTRINSIC_WO_CHAIN: {
unsigned IID = N->getConstantOperandVal(0);
if (IID < Intrinsic::num_intrinsics)
return IID;
return Intrinsic::not_intrinsic;
}
}
}
bool AArch64TargetLowering::isReassocProfitable(SelectionDAG &DAG, SDValue N0,
SDValue N1) const {
if (!N0.hasOneUse())
return false;
unsigned IID = getIntrinsicID(N1.getNode());
// Avoid reassociating expressions that can be lowered to smlal/umlal.
if (IID == Intrinsic::aarch64_neon_umull ||
N1.getOpcode() == AArch64ISD::UMULL ||
IID == Intrinsic::aarch64_neon_smull ||
N1.getOpcode() == AArch64ISD::SMULL)
return N0.getOpcode() != ISD::ADD;
return true;
}
/// Selects the correct CCAssignFn for a given CallingConvention value.
CCAssignFn *AArch64TargetLowering::CCAssignFnForCall(CallingConv::ID CC,
bool IsVarArg) const {
switch (CC) {
default:
reportFatalUsageError("unsupported calling convention");
case CallingConv::GHC:
return CC_AArch64_GHC;
case CallingConv::PreserveNone:
// The VarArg implementation makes assumptions about register
// argument passing that do not hold for preserve_none, so we
// instead fall back to C argument passing.
// The non-vararg case is handled in the CC function itself.
if (!IsVarArg)
return CC_AArch64_Preserve_None;
[[fallthrough]];
case CallingConv::C:
case CallingConv::Fast:
case CallingConv::PreserveMost:
case CallingConv::PreserveAll:
case CallingConv::CXX_FAST_TLS:
case CallingConv::Swift:
case CallingConv::SwiftTail:
case CallingConv::Tail:
case CallingConv::GRAAL:
if (Subtarget->isTargetWindows()) {
if (IsVarArg) {
if (Subtarget->isWindowsArm64EC())
return CC_AArch64_Arm64EC_VarArg;
return CC_AArch64_Win64_VarArg;
}
return CC_AArch64_Win64PCS;
}
if (!Subtarget->isTargetDarwin())
return CC_AArch64_AAPCS;
if (!IsVarArg)
return CC_AArch64_DarwinPCS;
return Subtarget->isTargetILP32() ? CC_AArch64_DarwinPCS_ILP32_VarArg
: CC_AArch64_DarwinPCS_VarArg;
case CallingConv::Win64:
if (IsVarArg) {
if (Subtarget->isWindowsArm64EC())
return CC_AArch64_Arm64EC_VarArg;
return CC_AArch64_Win64_VarArg;
}
return CC_AArch64_Win64PCS;
case CallingConv::CFGuard_Check:
if (Subtarget->isWindowsArm64EC())
return CC_AArch64_Arm64EC_CFGuard_Check;
return CC_AArch64_Win64_CFGuard_Check;
case CallingConv::AArch64_VectorCall:
case CallingConv::AArch64_SVE_VectorCall:
case CallingConv::AArch64_SME_ABI_Support_Routines_PreserveMost_From_X0:
case CallingConv::AArch64_SME_ABI_Support_Routines_PreserveMost_From_X1:
case CallingConv::AArch64_SME_ABI_Support_Routines_PreserveMost_From_X2:
return CC_AArch64_AAPCS;
case CallingConv::ARM64EC_Thunk_X64:
return CC_AArch64_Arm64EC_Thunk;
case CallingConv::ARM64EC_Thunk_Native:
return CC_AArch64_Arm64EC_Thunk_Native;
}
}
CCAssignFn *
AArch64TargetLowering::CCAssignFnForReturn(CallingConv::ID CC) const {
switch (CC) {
default:
return RetCC_AArch64_AAPCS;
case CallingConv::ARM64EC_Thunk_X64:
return RetCC_AArch64_Arm64EC_Thunk;
case CallingConv::CFGuard_Check:
if (Subtarget->isWindowsArm64EC())
return RetCC_AArch64_Arm64EC_CFGuard_Check;
return RetCC_AArch64_AAPCS;
}
}
static bool isPassedInFPR(EVT VT) {
return VT.isFixedLengthVector() ||
(VT.isFloatingPoint() && !VT.isScalableVector());
}
SDValue AArch64TargetLowering::LowerFormalArguments(
SDValue Chain, CallingConv::ID CallConv, bool isVarArg,
const SmallVectorImpl<ISD::InputArg> &Ins, const SDLoc &DL,
SelectionDAG &DAG, SmallVectorImpl<SDValue> &InVals) const {
MachineFunction &MF = DAG.getMachineFunction();
const Function &F = MF.getFunction();
MachineFrameInfo &MFI = MF.getFrameInfo();
bool IsWin64 =
Subtarget->isCallingConvWin64(F.getCallingConv(), F.isVarArg());
bool StackViaX4 = CallConv == CallingConv::ARM64EC_Thunk_X64 ||
(isVarArg && Subtarget->isWindowsArm64EC());
AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
SmallVector<ISD::OutputArg, 4> Outs;
GetReturnInfo(CallConv, F.getReturnType(), F.getAttributes(), Outs,
DAG.getTargetLoweringInfo(), MF.getDataLayout());
if (any_of(Outs, [](ISD::OutputArg &Out){ return Out.VT.isScalableVector(); }))
FuncInfo->setIsSVECC(true);
// Assign locations to all of the incoming arguments.
SmallVector<CCValAssign, 16> ArgLocs;
CCState CCInfo(CallConv, isVarArg, MF, ArgLocs, *DAG.getContext());
// At this point, Ins[].VT may already be promoted to i32. To correctly
// handle passing i8 as i8 instead of i32 on stack, we pass in both i32 and
// i8 to CC_AArch64_AAPCS with i32 being ValVT and i8 being LocVT.
// Since AnalyzeFormalArguments uses Ins[].VT for both ValVT and LocVT, here
// we use a special version of AnalyzeFormalArguments to pass in ValVT and
// LocVT.
unsigned NumArgs = Ins.size();
Function::const_arg_iterator CurOrigArg = F.arg_begin();
unsigned CurArgIdx = 0;
bool UseVarArgCC = false;
if (IsWin64)
UseVarArgCC = isVarArg;
CCAssignFn *AssignFn = CCAssignFnForCall(CallConv, UseVarArgCC);
for (unsigned i = 0; i != NumArgs; ++i) {
MVT ValVT = Ins[i].VT;
if (Ins[i].isOrigArg()) {
std::advance(CurOrigArg, Ins[i].getOrigArgIndex() - CurArgIdx);
CurArgIdx = Ins[i].getOrigArgIndex();
// Get type of the original argument.
EVT ActualVT = getValueType(DAG.getDataLayout(), CurOrigArg->getType(),
/*AllowUnknown*/ true);
MVT ActualMVT = ActualVT.isSimple() ? ActualVT.getSimpleVT() : MVT::Other;
// If ActualMVT is i1/i8/i16, we should set LocVT to i8/i8/i16.
if (ActualMVT == MVT::i1 || ActualMVT == MVT::i8)
ValVT = MVT::i8;
else if (ActualMVT == MVT::i16)
ValVT = MVT::i16;
}
bool Res =
AssignFn(i, ValVT, ValVT, CCValAssign::Full, Ins[i].Flags, CCInfo);
assert(!Res && "Call operand has unhandled type");
(void)Res;
}
SMEAttrs Attrs = FuncInfo->getSMEFnAttrs();
bool IsLocallyStreaming =
!Attrs.hasStreamingInterface() && Attrs.hasStreamingBody();
assert(Chain.getOpcode() == ISD::EntryToken && "Unexpected Chain value");
SDValue Glue = Chain.getValue(1);
unsigned ExtraArgLocs = 0;
for (unsigned i = 0, e = Ins.size(); i != e; ++i) {
CCValAssign &VA = ArgLocs[i - ExtraArgLocs];
if (Ins[i].Flags.isByVal()) {
// Byval is used for HFAs in the PCS, but the system should work in a
// non-compliant manner for larger structs.
EVT PtrVT = getPointerTy(DAG.getDataLayout());
int Size = Ins[i].Flags.getByValSize();
unsigned NumRegs = (Size + 7) / 8;
// FIXME: This works on big-endian for composite byvals, which are the common
// case. It should also work for fundamental types too.
unsigned FrameIdx =
MFI.CreateFixedObject(8 * NumRegs, VA.getLocMemOffset(), false);
SDValue FrameIdxN = DAG.getFrameIndex(FrameIdx, PtrVT);
InVals.push_back(FrameIdxN);
continue;
}
if (Ins[i].Flags.isSwiftAsync())
MF.getInfo<AArch64FunctionInfo>()->setHasSwiftAsyncContext(true);
SDValue ArgValue;
if (VA.isRegLoc()) {
// Arguments stored in registers.
EVT RegVT = VA.getLocVT();
const TargetRegisterClass *RC;
if (RegVT == MVT::i32)
RC = &AArch64::GPR32RegClass;
else if (RegVT == MVT::i64)
RC = &AArch64::GPR64RegClass;
else if (RegVT == MVT::f16 || RegVT == MVT::bf16)
RC = &AArch64::FPR16RegClass;
else if (RegVT == MVT::f32)
RC = &AArch64::FPR32RegClass;
else if (RegVT == MVT::f64 || RegVT.is64BitVector())
RC = &AArch64::FPR64RegClass;
else if (RegVT == MVT::f128 || RegVT.is128BitVector())
RC = &AArch64::FPR128RegClass;
else if (RegVT.isScalableVector() &&
RegVT.getVectorElementType() == MVT::i1) {
FuncInfo->setIsSVECC(true);
RC = &AArch64::PPRRegClass;
} else if (RegVT == MVT::aarch64svcount) {
FuncInfo->setIsSVECC(true);
RC = &AArch64::PPRRegClass;
} else if (RegVT.isScalableVector()) {
FuncInfo->setIsSVECC(true);
RC = &AArch64::ZPRRegClass;
} else
llvm_unreachable("RegVT not supported by FORMAL_ARGUMENTS Lowering");
// Transform the arguments in physical registers into virtual ones.
Register Reg = MF.addLiveIn(VA.getLocReg(), RC);
if (IsLocallyStreaming) {
// LocallyStreamingFunctions must insert the SMSTART in the correct
// position, so we use Glue to ensure no instructions can be scheduled
// between the chain of:
// t0: ch,glue = EntryNode
// t1: res,ch,glue = CopyFromReg
// ...
// tn: res,ch,glue = CopyFromReg t(n-1), ..
// t(n+1): ch, glue = SMSTART t0:0, ...., tn:2
// ^^^^^^
// This will be the new Chain/Root node.
ArgValue = DAG.getCopyFromReg(Chain, DL, Reg, RegVT, Glue);
Glue = ArgValue.getValue(2);
if (isPassedInFPR(ArgValue.getValueType())) {
ArgValue =
DAG.getNode(AArch64ISD::COALESCER_BARRIER, DL,
DAG.getVTList(ArgValue.getValueType(), MVT::Glue),
{ArgValue, Glue});
Glue = ArgValue.getValue(1);
}
} else
ArgValue = DAG.getCopyFromReg(Chain, DL, Reg, RegVT);
// If this is an 8, 16 or 32-bit value, it is really passed promoted
// to 64 bits. Insert an assert[sz]ext to capture this, then
// truncate to the right size.
switch (VA.getLocInfo()) {
default:
llvm_unreachable("Unknown loc info!");
case CCValAssign::Full:
break;
case CCValAssign::Indirect:
assert(
(VA.getValVT().isScalableVT() || Subtarget->isWindowsArm64EC()) &&
"Indirect arguments should be scalable on most subtargets");
break;
case CCValAssign::BCvt:
ArgValue = DAG.getNode(ISD::BITCAST, DL, VA.getValVT(), ArgValue);
break;
case CCValAssign::AExt:
case CCValAssign::SExt:
case CCValAssign::ZExt:
break;
case CCValAssign::AExtUpper:
ArgValue = DAG.getNode(ISD::SRL, DL, RegVT, ArgValue,
DAG.getConstant(32, DL, RegVT));
ArgValue = DAG.getZExtOrTrunc(ArgValue, DL, VA.getValVT());
break;
}
} else { // VA.isRegLoc()
assert(VA.isMemLoc() && "CCValAssign is neither reg nor mem");
unsigned ArgOffset = VA.getLocMemOffset();
unsigned ArgSize = (VA.getLocInfo() == CCValAssign::Indirect
? VA.getLocVT().getSizeInBits()
: VA.getValVT().getSizeInBits()) / 8;
uint32_t BEAlign = 0;
if (!Subtarget->isLittleEndian() && ArgSize < 8 &&
!Ins[i].Flags.isInConsecutiveRegs())
BEAlign = 8 - ArgSize;
SDValue FIN;
MachinePointerInfo PtrInfo;
if (StackViaX4) {
// In both the ARM64EC varargs convention and the thunk convention,
// arguments on the stack are accessed relative to x4, not sp. In
// the thunk convention, there's an additional offset of 32 bytes
// to account for the shadow store.
unsigned ObjOffset = ArgOffset + BEAlign;
if (CallConv == CallingConv::ARM64EC_Thunk_X64)
ObjOffset += 32;
Register VReg = MF.addLiveIn(AArch64::X4, &AArch64::GPR64RegClass);
SDValue Val = DAG.getCopyFromReg(Chain, DL, VReg, MVT::i64);
FIN = DAG.getNode(ISD::ADD, DL, MVT::i64, Val,
DAG.getConstant(ObjOffset, DL, MVT::i64));
PtrInfo = MachinePointerInfo::getUnknownStack(MF);
} else {
int FI = MFI.CreateFixedObject(ArgSize, ArgOffset + BEAlign, true);
// Create load nodes to retrieve arguments from the stack.
FIN = DAG.getFrameIndex(FI, getPointerTy(DAG.getDataLayout()));
PtrInfo = MachinePointerInfo::getFixedStack(MF, FI);
}
// For NON_EXTLOAD, generic code in getLoad assert(ValVT == MemVT)
ISD::LoadExtType ExtType = ISD::NON_EXTLOAD;
MVT MemVT = VA.getValVT();
switch (VA.getLocInfo()) {
default:
break;
case CCValAssign::Trunc:
case CCValAssign::BCvt:
MemVT = VA.getLocVT();
break;
case CCValAssign::Indirect:
assert((VA.getValVT().isScalableVector() ||
Subtarget->isWindowsArm64EC()) &&
"Indirect arguments should be scalable on most subtargets");
MemVT = VA.getLocVT();
break;
case CCValAssign::SExt:
ExtType = ISD::SEXTLOAD;
break;
case CCValAssign::ZExt:
ExtType = ISD::ZEXTLOAD;
break;
case CCValAssign::AExt:
ExtType = ISD::EXTLOAD;
break;
}
ArgValue = DAG.getExtLoad(ExtType, DL, VA.getLocVT(), Chain, FIN, PtrInfo,
MemVT);
}
if (VA.getLocInfo() == CCValAssign::Indirect) {
assert((VA.getValVT().isScalableVT() ||
Subtarget->isWindowsArm64EC()) &&
"Indirect arguments should be scalable on most subtargets");
uint64_t PartSize = VA.getValVT().getStoreSize().getKnownMinValue();
unsigned NumParts = 1;
if (Ins[i].Flags.isInConsecutiveRegs()) {
while (!Ins[i + NumParts - 1].Flags.isInConsecutiveRegsLast())
++NumParts;
}
MVT PartLoad = VA.getValVT();
SDValue Ptr = ArgValue;
// Ensure we generate all loads for each tuple part, whilst updating the
// pointer after each load correctly using vscale.
while (NumParts > 0) {
ArgValue = DAG.getLoad(PartLoad, DL, Chain, Ptr, MachinePointerInfo());
InVals.push_back(ArgValue);
NumParts--;
if (NumParts > 0) {
SDValue BytesIncrement;
if (PartLoad.isScalableVector()) {
BytesIncrement = DAG.getVScale(
DL, Ptr.getValueType(),
APInt(Ptr.getValueSizeInBits().getFixedValue(), PartSize));
} else {
BytesIncrement = DAG.getConstant(
APInt(Ptr.getValueSizeInBits().getFixedValue(), PartSize), DL,
Ptr.getValueType());
}
Ptr = DAG.getNode(ISD::ADD, DL, Ptr.getValueType(), Ptr,
BytesIncrement, SDNodeFlags::NoUnsignedWrap);
ExtraArgLocs++;
i++;
}
}
} else {
if (Subtarget->isTargetILP32() && Ins[i].Flags.isPointer())
ArgValue = DAG.getNode(ISD::AssertZext, DL, ArgValue.getValueType(),
ArgValue, DAG.getValueType(MVT::i32));
// i1 arguments are zero-extended to i8 by the caller. Emit a
// hint to reflect this.
if (Ins[i].isOrigArg()) {
Argument *OrigArg = F.getArg(Ins[i].getOrigArgIndex());
if (OrigArg->getType()->isIntegerTy(1)) {
if (!Ins[i].Flags.isZExt()) {
ArgValue = DAG.getNode(AArch64ISD::ASSERT_ZEXT_BOOL, DL,
ArgValue.getValueType(), ArgValue);
}
}
}
InVals.push_back(ArgValue);
}
}
assert((ArgLocs.size() + ExtraArgLocs) == Ins.size());
// Insert the SMSTART if this is a locally streaming function and
// make sure it is Glued to the last CopyFromReg value.
if (IsLocallyStreaming) {
SDValue PStateSM;
if (Attrs.hasStreamingCompatibleInterface()) {
PStateSM = getRuntimePStateSM(DAG, Chain, DL, MVT::i64);
Register Reg = MF.getRegInfo().createVirtualRegister(
getRegClassFor(PStateSM.getValueType().getSimpleVT()));
FuncInfo->setPStateSMReg(Reg);
Chain = DAG.getCopyToReg(Chain, DL, Reg, PStateSM);
Chain = changeStreamingMode(DAG, DL, /*Enable*/ true, Chain, Glue,
AArch64SME::IfCallerIsNonStreaming, PStateSM);
} else
Chain = changeStreamingMode(DAG, DL, /*Enable*/ true, Chain, Glue,
AArch64SME::Always);
// Ensure that the SMSTART happens after the CopyWithChain such that its
// chain result is used.
for (unsigned I=0; I<InVals.size(); ++I) {
Register Reg = MF.getRegInfo().createVirtualRegister(
getRegClassFor(InVals[I].getValueType().getSimpleVT()));
Chain = DAG.getCopyToReg(Chain, DL, Reg, InVals[I]);
InVals[I] = DAG.getCopyFromReg(Chain, DL, Reg,
InVals[I].getValueType());
}
}
// varargs
if (isVarArg) {
if (DAG.getMachineFunction().getFrameInfo().hasVAStart()) {
if (!Subtarget->isTargetDarwin() || IsWin64) {
// The AAPCS variadic function ABI is identical to the non-variadic
// one. As a result there may be more arguments in registers and we
// should save them for future reference.
// Win64 variadic functions also pass arguments in registers, but all
// float arguments are passed in integer registers.
saveVarArgRegisters(CCInfo, DAG, DL, Chain);
}
// This will point to the next argument passed via stack.
unsigned VarArgsOffset = CCInfo.getStackSize();
// We currently pass all varargs at 8-byte alignment, or 4 for ILP32
VarArgsOffset =
alignTo(VarArgsOffset, Subtarget->isTargetILP32() ? 4 : 8);
FuncInfo->setVarArgsStackOffset(VarArgsOffset);
FuncInfo->setVarArgsStackIndex(
MFI.CreateFixedObject(4, VarArgsOffset, true));
}
if (MFI.hasMustTailInVarArgFunc()) {
SmallVector<MVT, 2> RegParmTypes;
RegParmTypes.push_back(MVT::i64);
RegParmTypes.push_back(MVT::f128);
// Compute the set of forwarded registers. The rest are scratch.
SmallVectorImpl<ForwardedRegister> &Forwards =
FuncInfo->getForwardedMustTailRegParms();
CCInfo.analyzeMustTailForwardedRegisters(Forwards, RegParmTypes,
CC_AArch64_AAPCS);
// Conservatively forward X8, since it might be used for aggregate return.
if (!CCInfo.isAllocated(AArch64::X8)) {
Register X8VReg = MF.addLiveIn(AArch64::X8, &AArch64::GPR64RegClass);
Forwards.push_back(ForwardedRegister(X8VReg, AArch64::X8, MVT::i64));
}
}
}
// On Windows, InReg pointers must be returned, so record the pointer in a
// virtual register at the start of the function so it can be returned in the
// epilogue.
if (IsWin64 || F.getCallingConv() == CallingConv::ARM64EC_Thunk_X64) {
for (unsigned I = 0, E = Ins.size(); I != E; ++I) {
if ((F.getCallingConv() == CallingConv::ARM64EC_Thunk_X64 ||
Ins[I].Flags.isInReg()) &&
Ins[I].Flags.isSRet()) {
assert(!FuncInfo->getSRetReturnReg());
MVT PtrTy = getPointerTy(DAG.getDataLayout());
Register Reg =
MF.getRegInfo().createVirtualRegister(getRegClassFor(PtrTy));
FuncInfo->setSRetReturnReg(Reg);
SDValue Copy = DAG.getCopyToReg(DAG.getEntryNode(), DL, Reg, InVals[I]);
Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, Copy, Chain);
break;
}
}
}
unsigned StackArgSize = CCInfo.getStackSize();
bool TailCallOpt = MF.getTarget().Options.GuaranteedTailCallOpt;
if (DoesCalleeRestoreStack(CallConv, TailCallOpt)) {
// This is a non-standard ABI so by fiat I say we're allowed to make full
// use of the stack area to be popped, which must be aligned to 16 bytes in
// any case:
StackArgSize = alignTo(StackArgSize, 16);
// If we're expected to restore the stack (e.g. fastcc) then we'll be adding
// a multiple of 16.
FuncInfo->setArgumentStackToRestore(StackArgSize);
// This realignment carries over to the available bytes below. Our own
// callers will guarantee the space is free by giving an aligned value to
// CALLSEQ_START.
}
// Even if we're not expected to free up the space, it's useful to know how
// much is there while considering tail calls (because we can reuse it).
FuncInfo->setBytesInStackArgArea(StackArgSize);
if (Subtarget->hasCustomCallingConv())
Subtarget->getRegisterInfo()->UpdateCustomCalleeSavedRegs(MF);
// Create a 16 Byte TPIDR2 object. The dynamic buffer
// will be expanded and stored in the static object later using a pseudonode.
if (Attrs.hasZAState()) {
TPIDR2Object &TPIDR2 = FuncInfo->getTPIDR2Obj();
TPIDR2.FrameIndex = MFI.CreateStackObject(16, Align(16), false);
SDValue SVL = DAG.getNode(AArch64ISD::RDSVL, DL, MVT::i64,
DAG.getConstant(1, DL, MVT::i32));
SDValue Buffer;
if (!Subtarget->isTargetWindows() && !hasInlineStackProbe(MF)) {
Buffer = DAG.getNode(AArch64ISD::ALLOCATE_ZA_BUFFER, DL,
DAG.getVTList(MVT::i64, MVT::Other), {Chain, SVL});
} else {
SDValue Size = DAG.getNode(ISD::MUL, DL, MVT::i64, SVL, SVL);
Buffer = DAG.getNode(ISD::DYNAMIC_STACKALLOC, DL,
DAG.getVTList(MVT::i64, MVT::Other),
{Chain, Size, DAG.getConstant(1, DL, MVT::i64)});
MFI.CreateVariableSizedObject(Align(16), nullptr);
}
Chain = DAG.getNode(
AArch64ISD::INIT_TPIDR2OBJ, DL, DAG.getVTList(MVT::Other),
{/*Chain*/ Buffer.getValue(1), /*Buffer ptr*/ Buffer.getValue(0)});
} else if (Attrs.hasAgnosticZAInterface()) {
// Call __arm_sme_state_size().
SDValue BufferSize =
DAG.getNode(AArch64ISD::GET_SME_SAVE_SIZE, DL,
DAG.getVTList(MVT::i64, MVT::Other), Chain);
Chain = BufferSize.getValue(1);
SDValue Buffer;
if (!Subtarget->isTargetWindows() && !hasInlineStackProbe(MF)) {
Buffer =
DAG.getNode(AArch64ISD::ALLOC_SME_SAVE_BUFFER, DL,
DAG.getVTList(MVT::i64, MVT::Other), {Chain, BufferSize});
} else {
// Allocate space dynamically.
Buffer = DAG.getNode(
ISD::DYNAMIC_STACKALLOC, DL, DAG.getVTList(MVT::i64, MVT::Other),
{Chain, BufferSize, DAG.getConstant(1, DL, MVT::i64)});
MFI.CreateVariableSizedObject(Align(16), nullptr);
}
// Copy the value to a virtual register, and save that in FuncInfo.
Register BufferPtr =
MF.getRegInfo().createVirtualRegister(&AArch64::GPR64RegClass);
FuncInfo->setSMESaveBufferAddr(BufferPtr);
Chain = DAG.getCopyToReg(Chain, DL, BufferPtr, Buffer);
}
if (CallConv == CallingConv::PreserveNone) {
for (const ISD::InputArg &I : Ins) {
if (I.Flags.isSwiftSelf() || I.Flags.isSwiftError() ||
I.Flags.isSwiftAsync()) {
MachineFunction &MF = DAG.getMachineFunction();
DAG.getContext()->diagnose(DiagnosticInfoUnsupported(
MF.getFunction(),
"Swift attributes can't be used with preserve_none",
DL.getDebugLoc()));
break;
}
}
}
return Chain;
}
void AArch64TargetLowering::saveVarArgRegisters(CCState &CCInfo,
SelectionDAG &DAG,
const SDLoc &DL,
SDValue &Chain) const {
MachineFunction &MF = DAG.getMachineFunction();
MachineFrameInfo &MFI = MF.getFrameInfo();
AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
auto PtrVT = getPointerTy(DAG.getDataLayout());
Function &F = MF.getFunction();
bool IsWin64 =
Subtarget->isCallingConvWin64(F.getCallingConv(), F.isVarArg());
SmallVector<SDValue, 8> MemOps;
auto GPRArgRegs = AArch64::getGPRArgRegs();
unsigned NumGPRArgRegs = GPRArgRegs.size();
if (Subtarget->isWindowsArm64EC()) {
// In the ARM64EC ABI, only x0-x3 are used to pass arguments to varargs
// functions.
NumGPRArgRegs = 4;
}
unsigned FirstVariadicGPR = CCInfo.getFirstUnallocated(GPRArgRegs);
unsigned GPRSaveSize = 8 * (NumGPRArgRegs - FirstVariadicGPR);
int GPRIdx = 0;
if (GPRSaveSize != 0) {
if (IsWin64) {
GPRIdx = MFI.CreateFixedObject(GPRSaveSize, -(int)GPRSaveSize, false);
if (GPRSaveSize & 15)
// The extra size here, if triggered, will always be 8.
MFI.CreateFixedObject(16 - (GPRSaveSize & 15), -(int)alignTo(GPRSaveSize, 16), false);
} else
GPRIdx = MFI.CreateStackObject(GPRSaveSize, Align(8), false);
SDValue FIN;
if (Subtarget->isWindowsArm64EC()) {
// With the Arm64EC ABI, we reserve the save area as usual, but we
// compute its address relative to x4. For a normal AArch64->AArch64
// call, x4 == sp on entry, but calls from an entry thunk can pass in a
// different address.
Register VReg = MF.addLiveIn(AArch64::X4, &AArch64::GPR64RegClass);
SDValue Val = DAG.getCopyFromReg(Chain, DL, VReg, MVT::i64);
FIN = DAG.getNode(ISD::SUB, DL, MVT::i64, Val,
DAG.getConstant(GPRSaveSize, DL, MVT::i64));
} else {
FIN = DAG.getFrameIndex(GPRIdx, PtrVT);
}
for (unsigned i = FirstVariadicGPR; i < NumGPRArgRegs; ++i) {
Register VReg = MF.addLiveIn(GPRArgRegs[i], &AArch64::GPR64RegClass);
SDValue Val = DAG.getCopyFromReg(Chain, DL, VReg, MVT::i64);
SDValue Store =
DAG.getStore(Val.getValue(1), DL, Val, FIN,
IsWin64 ? MachinePointerInfo::getFixedStack(
MF, GPRIdx, (i - FirstVariadicGPR) * 8)
: MachinePointerInfo::getStack(MF, i * 8));
MemOps.push_back(Store);
FIN =
DAG.getNode(ISD::ADD, DL, PtrVT, FIN, DAG.getConstant(8, DL, PtrVT));
}
}
FuncInfo->setVarArgsGPRIndex(GPRIdx);
FuncInfo->setVarArgsGPRSize(GPRSaveSize);
if (Subtarget->hasFPARMv8() && !IsWin64) {
auto FPRArgRegs = AArch64::getFPRArgRegs();
const unsigned NumFPRArgRegs = FPRArgRegs.size();
unsigned FirstVariadicFPR = CCInfo.getFirstUnallocated(FPRArgRegs);
unsigned FPRSaveSize = 16 * (NumFPRArgRegs - FirstVariadicFPR);
int FPRIdx = 0;
if (FPRSaveSize != 0) {
FPRIdx = MFI.CreateStackObject(FPRSaveSize, Align(16), false);
SDValue FIN = DAG.getFrameIndex(FPRIdx, PtrVT);
for (unsigned i = FirstVariadicFPR; i < NumFPRArgRegs; ++i) {
Register VReg = MF.addLiveIn(FPRArgRegs[i], &AArch64::FPR128RegClass);
SDValue Val = DAG.getCopyFromReg(Chain, DL, VReg, MVT::f128);
SDValue Store = DAG.getStore(Val.getValue(1), DL, Val, FIN,
MachinePointerInfo::getStack(MF, i * 16));
MemOps.push_back(Store);
FIN = DAG.getNode(ISD::ADD, DL, PtrVT, FIN,
DAG.getConstant(16, DL, PtrVT));
}
}
FuncInfo->setVarArgsFPRIndex(FPRIdx);
FuncInfo->setVarArgsFPRSize(FPRSaveSize);
}
if (!MemOps.empty()) {
Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, MemOps);
}
}
/// LowerCallResult - Lower the result values of a call into the
/// appropriate copies out of appropriate physical registers.
SDValue AArch64TargetLowering::LowerCallResult(
SDValue Chain, SDValue InGlue, CallingConv::ID CallConv, bool isVarArg,
const SmallVectorImpl<CCValAssign> &RVLocs, const SDLoc &DL,
SelectionDAG &DAG, SmallVectorImpl<SDValue> &InVals, bool isThisReturn,
SDValue ThisVal, bool RequiresSMChange) const {
DenseMap<unsigned, SDValue> CopiedRegs;
// Copy all of the result registers out of their specified physreg.
for (unsigned i = 0; i != RVLocs.size(); ++i) {
CCValAssign VA = RVLocs[i];
// Pass 'this' value directly from the argument to return value, to avoid
// reg unit interference
if (i == 0 && isThisReturn) {
assert(!VA.needsCustom() && VA.getLocVT() == MVT::i64 &&
"unexpected return calling convention register assignment");
InVals.push_back(ThisVal);
continue;
}
// Avoid copying a physreg twice since RegAllocFast is incompetent and only
// allows one use of a physreg per block.
SDValue Val = CopiedRegs.lookup(VA.getLocReg());
if (!Val) {
Val =
DAG.getCopyFromReg(Chain, DL, VA.getLocReg(), VA.getLocVT(), InGlue);
Chain = Val.getValue(1);
InGlue = Val.getValue(2);
CopiedRegs[VA.getLocReg()] = Val;
}
switch (VA.getLocInfo()) {
default:
llvm_unreachable("Unknown loc info!");
case CCValAssign::Full:
break;
case CCValAssign::BCvt:
Val = DAG.getNode(ISD::BITCAST, DL, VA.getValVT(), Val);
break;
case CCValAssign::AExtUpper:
Val = DAG.getNode(ISD::SRL, DL, VA.getLocVT(), Val,
DAG.getConstant(32, DL, VA.getLocVT()));
[[fallthrough]];
case CCValAssign::AExt:
[[fallthrough]];
case CCValAssign::ZExt:
Val = DAG.getZExtOrTrunc(Val, DL, VA.getValVT());
break;
}
if (RequiresSMChange && isPassedInFPR(VA.getValVT()))
Val = DAG.getNode(AArch64ISD::COALESCER_BARRIER, DL,
DAG.getVTList(Val.getValueType(), MVT::Glue), Val);
InVals.push_back(Val);
}
return Chain;
}
/// Return true if the calling convention is one that we can guarantee TCO for.
static bool canGuaranteeTCO(CallingConv::ID CC, bool GuaranteeTailCalls) {
return (CC == CallingConv::Fast && GuaranteeTailCalls) ||
CC == CallingConv::Tail || CC == CallingConv::SwiftTail;
}
/// Return true if we might ever do TCO for calls with this calling convention.
static bool mayTailCallThisCC(CallingConv::ID CC) {
switch (CC) {
case CallingConv::C:
case CallingConv::AArch64_SVE_VectorCall:
case CallingConv::PreserveMost:
case CallingConv::PreserveAll:
case CallingConv::PreserveNone:
case CallingConv::Swift:
case CallingConv::SwiftTail:
case CallingConv::Tail:
case CallingConv::Fast:
return true;
default:
return false;
}
}
/// Return true if the call convention supports varargs
/// Currently only those that pass varargs like the C
/// calling convention does are eligible
/// Calling conventions listed in this function must also
/// be properly handled in AArch64Subtarget::isCallingConvWin64
static bool callConvSupportsVarArgs(CallingConv::ID CC) {
switch (CC) {
case CallingConv::C:
case CallingConv::PreserveNone:
// SVE vector call is only partially supported, but it should
// support named arguments being passed. Any arguments being passed
// as varargs, are still unsupported.
case CallingConv::AArch64_SVE_VectorCall:
return true;
default:
return false;
}
}
static void analyzeCallOperands(const AArch64TargetLowering &TLI,
const AArch64Subtarget *Subtarget,
const TargetLowering::CallLoweringInfo &CLI,
CCState &CCInfo) {
const SelectionDAG &DAG = CLI.DAG;
CallingConv::ID CalleeCC = CLI.CallConv;
bool IsVarArg = CLI.IsVarArg;
const SmallVector<ISD::OutputArg, 32> &Outs = CLI.Outs;
bool IsCalleeWin64 = Subtarget->isCallingConvWin64(CalleeCC, IsVarArg);
// For Arm64EC thunks, allocate 32 extra bytes at the bottom of the stack
// for the shadow store.
if (CalleeCC == CallingConv::ARM64EC_Thunk_X64)
CCInfo.AllocateStack(32, Align(16));
unsigned NumArgs = Outs.size();
for (unsigned i = 0; i != NumArgs; ++i) {
MVT ArgVT = Outs[i].VT;
ISD::ArgFlagsTy ArgFlags = Outs[i].Flags;
bool UseVarArgCC = false;
if (IsVarArg) {
// On Windows, the fixed arguments in a vararg call are passed in GPRs
// too, so use the vararg CC to force them to integer registers.
if (IsCalleeWin64) {
UseVarArgCC = true;
} else {
UseVarArgCC = !Outs[i].IsFixed;
}
}
if (!UseVarArgCC) {
// Get type of the original argument.
EVT ActualVT =
TLI.getValueType(DAG.getDataLayout(), CLI.Args[Outs[i].OrigArgIndex].Ty,
/*AllowUnknown*/ true);
MVT ActualMVT = ActualVT.isSimple() ? ActualVT.getSimpleVT() : ArgVT;
// If ActualMVT is i1/i8/i16, we should set LocVT to i8/i8/i16.
if (ActualMVT == MVT::i1 || ActualMVT == MVT::i8)
ArgVT = MVT::i8;
else if (ActualMVT == MVT::i16)
ArgVT = MVT::i16;
}
// FIXME: CCAssignFnForCall should be called once, for the call and not per
// argument. This logic should exactly mirror LowerFormalArguments.
CCAssignFn *AssignFn = TLI.CCAssignFnForCall(CalleeCC, UseVarArgCC);
bool Res = AssignFn(i, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, CCInfo);
assert(!Res && "Call operand has unhandled type");
(void)Res;
}
}
static SMECallAttrs
getSMECallAttrs(const Function &Caller,
const TargetLowering::CallLoweringInfo &CLI) {
if (CLI.CB)
return SMECallAttrs(*CLI.CB);
if (auto *ES = dyn_cast<ExternalSymbolSDNode>(CLI.Callee))
return SMECallAttrs(SMEAttrs(Caller), SMEAttrs(ES->getSymbol()));
return SMECallAttrs(SMEAttrs(Caller), SMEAttrs(SMEAttrs::Normal));
}
bool AArch64TargetLowering::isEligibleForTailCallOptimization(
const CallLoweringInfo &CLI) const {
CallingConv::ID CalleeCC = CLI.CallConv;
if (!mayTailCallThisCC(CalleeCC))
return false;
SDValue Callee = CLI.Callee;
bool IsVarArg = CLI.IsVarArg;
const SmallVector<ISD::OutputArg, 32> &Outs = CLI.Outs;
const SmallVector<SDValue, 32> &OutVals = CLI.OutVals;
const SmallVector<ISD::InputArg, 32> &Ins = CLI.Ins;
const SelectionDAG &DAG = CLI.DAG;
MachineFunction &MF = DAG.getMachineFunction();
const Function &CallerF = MF.getFunction();
CallingConv::ID CallerCC = CallerF.getCallingConv();
// SME Streaming functions are not eligible for TCO as they may require
// the streaming mode or ZA to be restored after returning from the call.
SMECallAttrs CallAttrs = getSMECallAttrs(CallerF, CLI);
if (CallAttrs.requiresSMChange() || CallAttrs.requiresLazySave() ||
CallAttrs.requiresPreservingAllZAState() ||
CallAttrs.caller().hasStreamingBody())
return false;
// Functions using the C or Fast calling convention that have an SVE signature
// preserve more registers and should assume the SVE_VectorCall CC.
// The check for matching callee-saved regs will determine whether it is
// eligible for TCO.
if ((CallerCC == CallingConv::C || CallerCC == CallingConv::Fast) &&
MF.getInfo<AArch64FunctionInfo>()->isSVECC())
CallerCC = CallingConv::AArch64_SVE_VectorCall;
bool CCMatch = CallerCC == CalleeCC;
// When using the Windows calling convention on a non-windows OS, we want
// to back up and restore X18 in such functions; we can't do a tail call
// from those functions.
if (CallerCC == CallingConv::Win64 && !Subtarget->isTargetWindows() &&
CalleeCC != CallingConv::Win64)
return false;
// Byval parameters hand the function a pointer directly into the stack area
// we want to reuse during a tail call. Working around this *is* possible (see
// X86) but less efficient and uglier in LowerCall.
for (Function::const_arg_iterator i = CallerF.arg_begin(),
e = CallerF.arg_end();
i != e; ++i) {
if (i->hasByValAttr())
return false;
// On Windows, "inreg" attributes signify non-aggregate indirect returns.
// In this case, it is necessary to save X0/X1 in the callee and return it
// in X0. Tail call opt may interfere with this, so we disable tail call
// opt when the caller has an "inreg" attribute -- except if the callee
// also has that attribute on the same argument, and the same value is
// passed.
if (i->hasInRegAttr()) {
unsigned ArgIdx = i - CallerF.arg_begin();
if (!CLI.CB || CLI.CB->arg_size() <= ArgIdx)
return false;
AttributeSet Attrs = CLI.CB->getParamAttributes(ArgIdx);
if (!Attrs.hasAttribute(Attribute::InReg) ||
!Attrs.hasAttribute(Attribute::StructRet) || !i->hasStructRetAttr() ||
CLI.CB->getArgOperand(ArgIdx) != i) {
return false;
}
}
}
if (canGuaranteeTCO(CalleeCC, getTargetMachine().Options.GuaranteedTailCallOpt))
return CCMatch;
// Externally-defined functions with weak linkage should not be
// tail-called on AArch64 when the OS does not support dynamic
// pre-emption of symbols, as the AAELF spec requires normal calls
// to undefined weak functions to be replaced with a NOP or jump to the
// next instruction. The behaviour of branch instructions in this
// situation (as used for tail calls) is implementation-defined, so we
// cannot rely on the linker replacing the tail call with a return.
if (GlobalAddressSDNode *G = dyn_cast<GlobalAddressSDNode>(Callee)) {
const GlobalValue *GV = G->getGlobal();
const Triple &TT = getTargetMachine().getTargetTriple();
if (GV->hasExternalWeakLinkage() &&
(!TT.isOSWindows() || TT.isOSBinFormatELF() || TT.isOSBinFormatMachO()))
return false;
}
// Now we search for cases where we can use a tail call without changing the
// ABI. Sibcall is used in some places (particularly gcc) to refer to this
// concept.
// I want anyone implementing a new calling convention to think long and hard
// about this assert.
if (IsVarArg && !callConvSupportsVarArgs(CalleeCC))
report_fatal_error("Unsupported variadic calling convention");
LLVMContext &C = *DAG.getContext();
// Check that the call results are passed in the same way.
if (!CCState::resultsCompatible(CalleeCC, CallerCC, MF, C, Ins,
CCAssignFnForCall(CalleeCC, IsVarArg),
CCAssignFnForCall(CallerCC, IsVarArg)))
return false;
// The callee has to preserve all registers the caller needs to preserve.
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
const uint32_t *CallerPreserved = TRI->getCallPreservedMask(MF, CallerCC);
if (!CCMatch) {
const uint32_t *CalleePreserved = TRI->getCallPreservedMask(MF, CalleeCC);
if (Subtarget->hasCustomCallingConv()) {
TRI->UpdateCustomCallPreservedMask(MF, &CallerPreserved);
TRI->UpdateCustomCallPreservedMask(MF, &CalleePreserved);
}
if (!TRI->regmaskSubsetEqual(CallerPreserved, CalleePreserved))
return false;
}
// Nothing more to check if the callee is taking no arguments
if (Outs.empty())
return true;
SmallVector<CCValAssign, 16> ArgLocs;
CCState CCInfo(CalleeCC, IsVarArg, MF, ArgLocs, C);
analyzeCallOperands(*this, Subtarget, CLI, CCInfo);
if (IsVarArg && !(CLI.CB && CLI.CB->isMustTailCall())) {
// When we are musttail, additional checks have been done and we can safely ignore this check
// At least two cases here: if caller is fastcc then we can't have any
// memory arguments (we'd be expected to clean up the stack afterwards). If
// caller is C then we could potentially use its argument area.
// FIXME: for now we take the most conservative of these in both cases:
// disallow all variadic memory operands.
for (const CCValAssign &ArgLoc : ArgLocs)
if (!ArgLoc.isRegLoc())
return false;
}
const AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
// If any of the arguments is passed indirectly, it must be SVE, so the
// 'getBytesInStackArgArea' is not sufficient to determine whether we need to
// allocate space on the stack. That is why we determine this explicitly here
// the call cannot be a tailcall.
if (llvm::any_of(ArgLocs, [&](CCValAssign &A) {
assert((A.getLocInfo() != CCValAssign::Indirect ||
A.getValVT().isScalableVector() ||
Subtarget->isWindowsArm64EC()) &&
"Expected value to be scalable");
return A.getLocInfo() == CCValAssign::Indirect;
}))
return false;
// If the stack arguments for this call do not fit into our own save area then
// the call cannot be made tail.
if (CCInfo.getStackSize() > FuncInfo->getBytesInStackArgArea())
return false;
const MachineRegisterInfo &MRI = MF.getRegInfo();
if (!parametersInCSRMatch(MRI, CallerPreserved, ArgLocs, OutVals))
return false;
return true;
}
SDValue AArch64TargetLowering::addTokenForArgument(SDValue Chain,
SelectionDAG &DAG,
MachineFrameInfo &MFI,
int ClobberedFI) const {
SmallVector<SDValue, 8> ArgChains;
int64_t FirstByte = MFI.getObjectOffset(ClobberedFI);
int64_t LastByte = FirstByte + MFI.getObjectSize(ClobberedFI) - 1;
// Include the original chain at the beginning of the list. When this is
// used by target LowerCall hooks, this helps legalize find the
// CALLSEQ_BEGIN node.
ArgChains.push_back(Chain);
// Add a chain value for each stack argument corresponding
for (SDNode *U : DAG.getEntryNode().getNode()->users())
if (LoadSDNode *L = dyn_cast<LoadSDNode>(U))
if (FrameIndexSDNode *FI = dyn_cast<FrameIndexSDNode>(L->getBasePtr()))
if (FI->getIndex() < 0) {
int64_t InFirstByte = MFI.getObjectOffset(FI->getIndex());
int64_t InLastByte = InFirstByte;
InLastByte += MFI.getObjectSize(FI->getIndex()) - 1;
if ((InFirstByte <= FirstByte && FirstByte <= InLastByte) ||
(FirstByte <= InFirstByte && InFirstByte <= LastByte))
ArgChains.push_back(SDValue(L, 1));
}
// Build a tokenfactor for all the chains.
return DAG.getNode(ISD::TokenFactor, SDLoc(Chain), MVT::Other, ArgChains);
}
bool AArch64TargetLowering::DoesCalleeRestoreStack(CallingConv::ID CallCC,
bool TailCallOpt) const {
return (CallCC == CallingConv::Fast && TailCallOpt) ||
CallCC == CallingConv::Tail || CallCC == CallingConv::SwiftTail;
}
// Check if the value is zero-extended from i1 to i8
static bool checkZExtBool(SDValue Arg, const SelectionDAG &DAG) {
unsigned SizeInBits = Arg.getValueType().getSizeInBits();
if (SizeInBits < 8)
return false;
APInt RequiredZero(SizeInBits, 0xFE);
KnownBits Bits = DAG.computeKnownBits(Arg, 4);
bool ZExtBool = (Bits.Zero & RequiredZero) == RequiredZero;
return ZExtBool;
}
void AArch64TargetLowering::AdjustInstrPostInstrSelection(MachineInstr &MI,
SDNode *Node) const {
// Live-in physreg copies that are glued to SMSTART are applied as
// implicit-def's in the InstrEmitter. Here we remove them, allowing the
// register allocator to pass call args in callee saved regs, without extra
// copies to avoid these fake clobbers of actually-preserved GPRs.
if (MI.getOpcode() == AArch64::MSRpstatesvcrImm1 ||
MI.getOpcode() == AArch64::MSRpstatePseudo) {
for (unsigned I = MI.getNumOperands() - 1; I > 0; --I)
if (MachineOperand &MO = MI.getOperand(I);
MO.isReg() && MO.isImplicit() && MO.isDef() &&
(AArch64::GPR32RegClass.contains(MO.getReg()) ||
AArch64::GPR64RegClass.contains(MO.getReg())))
MI.removeOperand(I);
// The SVE vector length can change when entering/leaving streaming mode.
// FPMR is set to 0 when entering/leaving streaming mode.
if (MI.getOperand(0).getImm() == AArch64SVCR::SVCRSM ||
MI.getOperand(0).getImm() == AArch64SVCR::SVCRSMZA) {
MI.addOperand(MachineOperand::CreateReg(AArch64::VG, /*IsDef=*/false,
/*IsImplicit=*/true));
MI.addOperand(MachineOperand::CreateReg(AArch64::VG, /*IsDef=*/true,
/*IsImplicit=*/true));
MI.addOperand(MachineOperand::CreateReg(AArch64::FPMR, /*IsDef=*/true,
/*IsImplicit=*/true));
}
}
// Add an implicit use of 'VG' for ADDXri/SUBXri, which are instructions that
// have nothing to do with VG, were it not that they are used to materialise a
// frame-address. If they contain a frame-index to a scalable vector, this
// will likely require an ADDVL instruction to materialise the address, thus
// reading VG.
const MachineFunction &MF = *MI.getMF();
if (MF.getInfo<AArch64FunctionInfo>()->hasStreamingModeChanges() &&
(MI.getOpcode() == AArch64::ADDXri ||
MI.getOpcode() == AArch64::SUBXri)) {
const MachineOperand &MO = MI.getOperand(1);
if (MO.isFI() && MF.getFrameInfo().getStackID(MO.getIndex()) ==
TargetStackID::ScalableVector)
MI.addOperand(MachineOperand::CreateReg(AArch64::VG, /*IsDef=*/false,
/*IsImplicit=*/true));
}
}
SDValue AArch64TargetLowering::changeStreamingMode(SelectionDAG &DAG, SDLoc DL,
bool Enable, SDValue Chain,
SDValue InGlue,
unsigned Condition,
SDValue PStateSM) const {
MachineFunction &MF = DAG.getMachineFunction();
AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
FuncInfo->setHasStreamingModeChanges(true);
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
SDValue RegMask = DAG.getRegisterMask(TRI->getSMStartStopCallPreservedMask());
SDValue MSROp =
DAG.getTargetConstant((int32_t)AArch64SVCR::SVCRSM, DL, MVT::i32);
SmallVector<SDValue> Ops = {Chain, MSROp};
unsigned Opcode;
if (Condition != AArch64SME::Always) {
SDValue ConditionOp = DAG.getTargetConstant(Condition, DL, MVT::i64);
Opcode = Enable ? AArch64ISD::COND_SMSTART : AArch64ISD::COND_SMSTOP;
assert(PStateSM && "PStateSM should be defined");
Ops.push_back(ConditionOp);
Ops.push_back(PStateSM);
} else {
Opcode = Enable ? AArch64ISD::SMSTART : AArch64ISD::SMSTOP;
}
Ops.push_back(RegMask);
if (InGlue)
Ops.push_back(InGlue);
return DAG.getNode(Opcode, DL, DAG.getVTList(MVT::Other, MVT::Glue), Ops);
}
// Emit a call to __arm_sme_save or __arm_sme_restore.
static SDValue emitSMEStateSaveRestore(const AArch64TargetLowering &TLI,
SelectionDAG &DAG,
AArch64FunctionInfo *Info, SDLoc DL,
SDValue Chain, bool IsSave) {
MachineFunction &MF = DAG.getMachineFunction();
AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
FuncInfo->setSMESaveBufferUsed();
TargetLowering::ArgListTy Args;
TargetLowering::ArgListEntry Entry;
Entry.Ty = PointerType::getUnqual(*DAG.getContext());
Entry.Node =
DAG.getCopyFromReg(Chain, DL, Info->getSMESaveBufferAddr(), MVT::i64);
Args.push_back(Entry);
SDValue Callee =
DAG.getExternalSymbol(IsSave ? "__arm_sme_save" : "__arm_sme_restore",
TLI.getPointerTy(DAG.getDataLayout()));
auto *RetTy = Type::getVoidTy(*DAG.getContext());
TargetLowering::CallLoweringInfo CLI(DAG);
CLI.setDebugLoc(DL).setChain(Chain).setLibCallee(
CallingConv::AArch64_SME_ABI_Support_Routines_PreserveMost_From_X1, RetTy,
Callee, std::move(Args));
return TLI.LowerCallTo(CLI).second;
}
static AArch64SME::ToggleCondition
getSMToggleCondition(const SMECallAttrs &CallAttrs) {
if (!CallAttrs.caller().hasStreamingCompatibleInterface() ||
CallAttrs.caller().hasStreamingBody())
return AArch64SME::Always;
if (CallAttrs.callee().hasNonStreamingInterface())
return AArch64SME::IfCallerIsStreaming;
if (CallAttrs.callee().hasStreamingInterface())
return AArch64SME::IfCallerIsNonStreaming;
llvm_unreachable("Unsupported attributes");
}
/// Check whether a stack argument requires lowering in a tail call.
static bool shouldLowerTailCallStackArg(const MachineFunction &MF,
const CCValAssign &VA, SDValue Arg,
ISD::ArgFlagsTy Flags, int CallOffset) {
// FIXME: We should be able to handle this case, but it's not clear how to.
if (Flags.isZExt() || Flags.isSExt())
return true;
for (;;) {
// Look through nodes that don't alter the bits of the incoming value.
unsigned Op = Arg.getOpcode();
if (Op == ISD::ZERO_EXTEND || Op == ISD::ANY_EXTEND || Op == ISD::BITCAST ||
Arg->isAssert() || Op == AArch64ISD::ASSERT_ZEXT_BOOL) {
Arg = Arg.getOperand(0);
continue;
}
break;
}
// If the argument is a load from the same immutable stack slot, we can reuse
// it.
if (auto *LoadNode = dyn_cast<LoadSDNode>(Arg)) {
if (auto *FINode = dyn_cast<FrameIndexSDNode>(LoadNode->getBasePtr())) {
const MachineFrameInfo &MFI = MF.getFrameInfo();
int FI = FINode->getIndex();
if (!MFI.isImmutableObjectIndex(FI))
return true;
if (CallOffset != MFI.getObjectOffset(FI))
return true;
uint64_t SizeInBits = LoadNode->getMemoryVT().getFixedSizeInBits();
if (SizeInBits / 8 != MFI.getObjectSize(FI))
return true;
return false;
}
}
return true;
}
/// LowerCall - Lower a call to a callseq_start + CALL + callseq_end chain,
/// and add input and output parameter nodes.
SDValue
AArch64TargetLowering::LowerCall(CallLoweringInfo &CLI,
SmallVectorImpl<SDValue> &InVals) const {
SelectionDAG &DAG = CLI.DAG;
SDLoc &DL = CLI.DL;
SmallVector<ISD::OutputArg, 32> &Outs = CLI.Outs;
SmallVector<SDValue, 32> &OutVals = CLI.OutVals;
SmallVector<ISD::InputArg, 32> &Ins = CLI.Ins;
SDValue Chain = CLI.Chain;
SDValue Callee = CLI.Callee;
bool &IsTailCall = CLI.IsTailCall;
CallingConv::ID &CallConv = CLI.CallConv;
bool IsVarArg = CLI.IsVarArg;
MachineFunction &MF = DAG.getMachineFunction();
MachineFunction::CallSiteInfo CSInfo;
bool IsThisReturn = false;
AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
bool TailCallOpt = MF.getTarget().Options.GuaranteedTailCallOpt;
bool IsCFICall = CLI.CB && CLI.CB->isIndirectCall() && CLI.CFIType;
bool IsSibCall = false;
bool GuardWithBTI = false;
if (CLI.CB && CLI.CB->hasFnAttr(Attribute::ReturnsTwice) &&
!Subtarget->noBTIAtReturnTwice()) {
GuardWithBTI = FuncInfo->branchTargetEnforcement();
}
// Analyze operands of the call, assigning locations to each operand.
SmallVector<CCValAssign, 16> ArgLocs;
CCState CCInfo(CallConv, IsVarArg, MF, ArgLocs, *DAG.getContext());
if (IsVarArg) {
unsigned NumArgs = Outs.size();
for (unsigned i = 0; i != NumArgs; ++i) {
if (!Outs[i].IsFixed && Outs[i].VT.isScalableVector())
report_fatal_error("Passing SVE types to variadic functions is "
"currently not supported");
}
}
analyzeCallOperands(*this, Subtarget, CLI, CCInfo);
CCAssignFn *RetCC = CCAssignFnForReturn(CallConv);
// Assign locations to each value returned by this call.
SmallVector<CCValAssign, 16> RVLocs;
CCState RetCCInfo(CallConv, IsVarArg, DAG.getMachineFunction(), RVLocs,
*DAG.getContext());
RetCCInfo.AnalyzeCallResult(Ins, RetCC);
// Check callee args/returns for SVE registers and set calling convention
// accordingly.
if (CallConv == CallingConv::C || CallConv == CallingConv::Fast) {
auto HasSVERegLoc = [](CCValAssign &Loc) {
if (!Loc.isRegLoc())
return false;
return AArch64::ZPRRegClass.contains(Loc.getLocReg()) ||
AArch64::PPRRegClass.contains(Loc.getLocReg());
};
if (any_of(RVLocs, HasSVERegLoc) || any_of(ArgLocs, HasSVERegLoc))
CallConv = CallingConv::AArch64_SVE_VectorCall;
}
if (IsTailCall) {
// Check if it's really possible to do a tail call.
IsTailCall = isEligibleForTailCallOptimization(CLI);
// A sibling call is one where we're under the usual C ABI and not planning
// to change that but can still do a tail call:
if (!TailCallOpt && IsTailCall && CallConv != CallingConv::Tail &&
CallConv != CallingConv::SwiftTail)
IsSibCall = true;
if (IsTailCall)
++NumTailCalls;
}
if (!IsTailCall && CLI.CB && CLI.CB->isMustTailCall())
report_fatal_error("failed to perform tail call elimination on a call "
"site marked musttail");
// Get a count of how many bytes are to be pushed on the stack.
unsigned NumBytes = CCInfo.getStackSize();
if (IsSibCall) {
// Since we're not changing the ABI to make this a tail call, the memory
// operands are already available in the caller's incoming argument space.
NumBytes = 0;
}
// FPDiff is the byte offset of the call's argument area from the callee's.
// Stores to callee stack arguments will be placed in FixedStackSlots offset
// by this amount for a tail call. In a sibling call it must be 0 because the
// caller will deallocate the entire stack and the callee still expects its
// arguments to begin at SP+0. Completely unused for non-tail calls.
int FPDiff = 0;
if (IsTailCall && !IsSibCall) {
unsigned NumReusableBytes = FuncInfo->getBytesInStackArgArea();
// Since callee will pop argument stack as a tail call, we must keep the
// popped size 16-byte aligned.
NumBytes = alignTo(NumBytes, 16);
// FPDiff will be negative if this tail call requires more space than we
// would automatically have in our incoming argument space. Positive if we
// can actually shrink the stack.
FPDiff = NumReusableBytes - NumBytes;
// Update the required reserved area if this is the tail call requiring the
// most argument stack space.
if (FPDiff < 0 && FuncInfo->getTailCallReservedStack() < (unsigned)-FPDiff)
FuncInfo->setTailCallReservedStack(-FPDiff);
// The stack pointer must be 16-byte aligned at all times it's used for a
// memory operation, which in practice means at *all* times and in
// particular across call boundaries. Therefore our own arguments started at
// a 16-byte aligned SP and the delta applied for the tail call should
// satisfy the same constraint.
assert(FPDiff % 16 == 0 && "unaligned stack on tail call");
}
// Determine whether we need any streaming mode changes.
SMECallAttrs CallAttrs = getSMECallAttrs(MF.getFunction(), CLI);
auto DescribeCallsite =
[&](OptimizationRemarkAnalysis &R) -> OptimizationRemarkAnalysis & {
R << "call from '" << ore::NV("Caller", MF.getName()) << "' to '";
if (auto *ES = dyn_cast<ExternalSymbolSDNode>(CLI.Callee))
R << ore::NV("Callee", ES->getSymbol());
else if (CLI.CB && CLI.CB->getCalledFunction())
R << ore::NV("Callee", CLI.CB->getCalledFunction()->getName());
else
R << "unknown callee";
R << "'";
return R;
};
bool RequiresLazySave = CallAttrs.requiresLazySave();
bool RequiresSaveAllZA = CallAttrs.requiresPreservingAllZAState();
if (RequiresLazySave) {
const TPIDR2Object &TPIDR2 = FuncInfo->getTPIDR2Obj();
MachinePointerInfo MPI =
MachinePointerInfo::getStack(MF, TPIDR2.FrameIndex);
SDValue TPIDR2ObjAddr = DAG.getFrameIndex(
TPIDR2.FrameIndex,
DAG.getTargetLoweringInfo().getFrameIndexTy(DAG.getDataLayout()));
SDValue NumZaSaveSlicesAddr =
DAG.getNode(ISD::ADD, DL, TPIDR2ObjAddr.getValueType(), TPIDR2ObjAddr,
DAG.getConstant(8, DL, TPIDR2ObjAddr.getValueType()));
SDValue NumZaSaveSlices = DAG.getNode(AArch64ISD::RDSVL, DL, MVT::i64,
DAG.getConstant(1, DL, MVT::i32));
Chain = DAG.getTruncStore(Chain, DL, NumZaSaveSlices, NumZaSaveSlicesAddr,
MPI, MVT::i16);
Chain = DAG.getNode(
ISD::INTRINSIC_VOID, DL, MVT::Other, Chain,
DAG.getConstant(Intrinsic::aarch64_sme_set_tpidr2, DL, MVT::i32),
TPIDR2ObjAddr);
OptimizationRemarkEmitter ORE(&MF.getFunction());
ORE.emit([&]() {
auto R = CLI.CB ? OptimizationRemarkAnalysis("sme", "SMELazySaveZA",
CLI.CB)
: OptimizationRemarkAnalysis("sme", "SMELazySaveZA",
&MF.getFunction());
return DescribeCallsite(R) << " sets up a lazy save for ZA";
});
} else if (RequiresSaveAllZA) {
assert(!CallAttrs.callee().hasSharedZAInterface() &&
"Cannot share state that may not exist");
Chain = emitSMEStateSaveRestore(*this, DAG, FuncInfo, DL, Chain,
/*IsSave=*/true);
}
SDValue PStateSM;
bool RequiresSMChange = CallAttrs.requiresSMChange();
if (RequiresSMChange) {
if (CallAttrs.caller().hasStreamingInterfaceOrBody())
PStateSM = DAG.getConstant(1, DL, MVT::i64);
else if (CallAttrs.caller().hasNonStreamingInterface())
PStateSM = DAG.getConstant(0, DL, MVT::i64);
else
PStateSM = getRuntimePStateSM(DAG, Chain, DL, MVT::i64);
OptimizationRemarkEmitter ORE(&MF.getFunction());
ORE.emit([&]() {
auto R = CLI.CB ? OptimizationRemarkAnalysis("sme", "SMETransition",
CLI.CB)
: OptimizationRemarkAnalysis("sme", "SMETransition",
&MF.getFunction());
DescribeCallsite(R) << " requires a streaming mode transition";
return R;
});
}
SDValue ZTFrameIdx;
MachineFrameInfo &MFI = MF.getFrameInfo();
bool ShouldPreserveZT0 = CallAttrs.requiresPreservingZT0();
// If the caller has ZT0 state which will not be preserved by the callee,
// spill ZT0 before the call.
if (ShouldPreserveZT0) {
unsigned ZTObj = MFI.CreateSpillStackObject(64, Align(16));
ZTFrameIdx = DAG.getFrameIndex(
ZTObj,
DAG.getTargetLoweringInfo().getFrameIndexTy(DAG.getDataLayout()));
Chain = DAG.getNode(AArch64ISD::SAVE_ZT, DL, DAG.getVTList(MVT::Other),
{Chain, DAG.getConstant(0, DL, MVT::i32), ZTFrameIdx});
}
// If caller shares ZT0 but the callee is not shared ZA, we need to stop
// PSTATE.ZA before the call if there is no lazy-save active.
bool DisableZA = CallAttrs.requiresDisablingZABeforeCall();
assert((!DisableZA || !RequiresLazySave) &&
"Lazy-save should have PSTATE.SM=1 on entry to the function");
if (DisableZA)
Chain = DAG.getNode(
AArch64ISD::SMSTOP, DL, DAG.getVTList(MVT::Other, MVT::Glue), Chain,
DAG.getTargetConstant((int32_t)(AArch64SVCR::SVCRZA), DL, MVT::i32));
// Adjust the stack pointer for the new arguments...
// These operations are automatically eliminated by the prolog/epilog pass
if (!IsSibCall)
Chain = DAG.getCALLSEQ_START(Chain, IsTailCall ? 0 : NumBytes, 0, DL);
SDValue StackPtr = DAG.getCopyFromReg(Chain, DL, AArch64::SP,
getPointerTy(DAG.getDataLayout()));
SmallVector<std::pair<unsigned, SDValue>, 8> RegsToPass;
SmallSet<unsigned, 8> RegsUsed;
SmallVector<SDValue, 8> MemOpChains;
auto PtrVT = getPointerTy(DAG.getDataLayout());
if (IsVarArg && CLI.CB && CLI.CB->isMustTailCall()) {
const auto &Forwards = FuncInfo->getForwardedMustTailRegParms();
for (const auto &F : Forwards) {
SDValue Val = DAG.getCopyFromReg(Chain, DL, F.VReg, F.VT);
RegsToPass.emplace_back(F.PReg, Val);
}
}
// Walk the register/memloc assignments, inserting copies/loads.
unsigned ExtraArgLocs = 0;
for (unsigned i = 0, e = Outs.size(); i != e; ++i) {
CCValAssign &VA = ArgLocs[i - ExtraArgLocs];
SDValue Arg = OutVals[i];
ISD::ArgFlagsTy Flags = Outs[i].Flags;
// Promote the value if needed.
switch (VA.getLocInfo()) {
default:
llvm_unreachable("Unknown loc info!");
case CCValAssign::Full:
break;
case CCValAssign::SExt:
Arg = DAG.getNode(ISD::SIGN_EXTEND, DL, VA.getLocVT(), Arg);
break;
case CCValAssign::ZExt:
Arg = DAG.getNode(ISD::ZERO_EXTEND, DL, VA.getLocVT(), Arg);
break;
case CCValAssign::AExt:
if (Outs[i].ArgVT == MVT::i1) {
// AAPCS requires i1 to be zero-extended to 8-bits by the caller.
//
// Check if we actually have to do this, because the value may
// already be zero-extended.
//
// We cannot just emit a (zext i8 (trunc (assert-zext i8)))
// and rely on DAGCombiner to fold this, because the following
// (anyext i32) is combined with (zext i8) in DAG.getNode:
//
// (ext (zext x)) -> (zext x)
//
// This will give us (zext i32), which we cannot remove, so
// try to check this beforehand.
if (!checkZExtBool(Arg, DAG)) {
Arg = DAG.getNode(ISD::TRUNCATE, DL, MVT::i1, Arg);
Arg = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i8, Arg);
}
}
Arg = DAG.getNode(ISD::ANY_EXTEND, DL, VA.getLocVT(), Arg);
break;
case CCValAssign::AExtUpper:
assert(VA.getValVT() == MVT::i32 && "only expect 32 -> 64 upper bits");
Arg = DAG.getNode(ISD::ANY_EXTEND, DL, VA.getLocVT(), Arg);
Arg = DAG.getNode(ISD::SHL, DL, VA.getLocVT(), Arg,
DAG.getConstant(32, DL, VA.getLocVT()));
break;
case CCValAssign::BCvt:
Arg = DAG.getBitcast(VA.getLocVT(), Arg);
break;
case CCValAssign::Trunc:
Arg = DAG.getZExtOrTrunc(Arg, DL, VA.getLocVT());
break;
case CCValAssign::FPExt:
Arg = DAG.getNode(ISD::FP_EXTEND, DL, VA.getLocVT(), Arg);
break;
case CCValAssign::Indirect:
bool isScalable = VA.getValVT().isScalableVT();
assert((isScalable || Subtarget->isWindowsArm64EC()) &&
"Indirect arguments should be scalable on most subtargets");
uint64_t StoreSize = VA.getValVT().getStoreSize().getKnownMinValue();
uint64_t PartSize = StoreSize;
unsigned NumParts = 1;
if (Outs[i].Flags.isInConsecutiveRegs()) {
while (!Outs[i + NumParts - 1].Flags.isInConsecutiveRegsLast())
++NumParts;
StoreSize *= NumParts;
}
Type *Ty = EVT(VA.getValVT()).getTypeForEVT(*DAG.getContext());
Align Alignment = DAG.getDataLayout().getPrefTypeAlign(Ty);
MachineFrameInfo &MFI = MF.getFrameInfo();
int FI = MFI.CreateStackObject(StoreSize, Alignment, false);
if (isScalable)
MFI.setStackID(FI, TargetStackID::ScalableVector);
MachinePointerInfo MPI = MachinePointerInfo::getFixedStack(MF, FI);
SDValue Ptr = DAG.getFrameIndex(
FI, DAG.getTargetLoweringInfo().getFrameIndexTy(DAG.getDataLayout()));
SDValue SpillSlot = Ptr;
// Ensure we generate all stores for each tuple part, whilst updating the
// pointer after each store correctly using vscale.
while (NumParts) {
SDValue Store = DAG.getStore(Chain, DL, OutVals[i], Ptr, MPI);
MemOpChains.push_back(Store);
NumParts--;
if (NumParts > 0) {
SDValue BytesIncrement;
if (isScalable) {
BytesIncrement = DAG.getVScale(
DL, Ptr.getValueType(),
APInt(Ptr.getValueSizeInBits().getFixedValue(), PartSize));
} else {
BytesIncrement = DAG.getConstant(
APInt(Ptr.getValueSizeInBits().getFixedValue(), PartSize), DL,
Ptr.getValueType());
}
MPI = MachinePointerInfo(MPI.getAddrSpace());
Ptr = DAG.getNode(ISD::ADD, DL, Ptr.getValueType(), Ptr,
BytesIncrement, SDNodeFlags::NoUnsignedWrap);
ExtraArgLocs++;
i++;
}
}
Arg = SpillSlot;
break;
}
if (VA.isRegLoc()) {
if (i == 0 && Flags.isReturned() && !Flags.isSwiftSelf() &&
Outs[0].VT == MVT::i64) {
assert(VA.getLocVT() == MVT::i64 &&
"unexpected calling convention register assignment");
assert(!Ins.empty() && Ins[0].VT == MVT::i64 &&
"unexpected use of 'returned'");
IsThisReturn = true;
}
if (RegsUsed.count(VA.getLocReg())) {
// If this register has already been used then we're trying to pack
// parts of an [N x i32] into an X-register. The extension type will
// take care of putting the two halves in the right place but we have to
// combine them.
SDValue &Bits =
llvm::find_if(RegsToPass,
[=](const std::pair<unsigned, SDValue> &Elt) {
return Elt.first == VA.getLocReg();
})
->second;
Bits = DAG.getNode(ISD::OR, DL, Bits.getValueType(), Bits, Arg);
// Call site info is used for function's parameter entry value
// tracking. For now we track only simple cases when parameter
// is transferred through whole register.
llvm::erase_if(CSInfo.ArgRegPairs,
[&VA](MachineFunction::ArgRegPair ArgReg) {
return ArgReg.Reg == VA.getLocReg();
});
} else {
// Add an extra level of indirection for streaming mode changes by
// using a pseudo copy node that cannot be rematerialised between a
// smstart/smstop and the call by the simple register coalescer.
if (RequiresSMChange && isPassedInFPR(Arg.getValueType()))
Arg = DAG.getNode(AArch64ISD::COALESCER_BARRIER, DL,
DAG.getVTList(Arg.getValueType(), MVT::Glue), Arg);
RegsToPass.emplace_back(VA.getLocReg(), Arg);
RegsUsed.insert(VA.getLocReg());
const TargetOptions &Options = DAG.getTarget().Options;
if (Options.EmitCallSiteInfo)
CSInfo.ArgRegPairs.emplace_back(VA.getLocReg(), i);
}
} else {
assert(VA.isMemLoc());
SDValue DstAddr;
MachinePointerInfo DstInfo;
// FIXME: This works on big-endian for composite byvals, which are the
// common case. It should also work for fundamental types too.
uint32_t BEAlign = 0;
unsigned OpSize;
if (VA.getLocInfo() == CCValAssign::Indirect ||
VA.getLocInfo() == CCValAssign::Trunc)
OpSize = VA.getLocVT().getFixedSizeInBits();
else
OpSize = Flags.isByVal() ? Flags.getByValSize() * 8
: VA.getValVT().getSizeInBits();
OpSize = (OpSize + 7) / 8;
if (!Subtarget->isLittleEndian() && !Flags.isByVal() &&
!Flags.isInConsecutiveRegs()) {
if (OpSize < 8)
BEAlign = 8 - OpSize;
}
unsigned LocMemOffset = VA.getLocMemOffset();
int32_t Offset = LocMemOffset + BEAlign;
if (IsTailCall) {
// When the frame pointer is perfectly aligned for the tail call and the
// same stack argument is passed down intact, we can reuse it.
if (!FPDiff && !shouldLowerTailCallStackArg(MF, VA, Arg, Flags, Offset))
continue;
Offset = Offset + FPDiff;
int FI = MF.getFrameInfo().CreateFixedObject(OpSize, Offset, true);
DstAddr = DAG.getFrameIndex(FI, PtrVT);
DstInfo = MachinePointerInfo::getFixedStack(MF, FI);
// Make sure any stack arguments overlapping with where we're storing
// are loaded before this eventual operation. Otherwise they'll be
// clobbered.
Chain = addTokenForArgument(Chain, DAG, MF.getFrameInfo(), FI);
} else {
SDValue PtrOff = DAG.getIntPtrConstant(Offset, DL);
DstAddr = DAG.getNode(ISD::ADD, DL, PtrVT, StackPtr, PtrOff);
DstInfo = MachinePointerInfo::getStack(MF, LocMemOffset);
}
if (Outs[i].Flags.isByVal()) {
SDValue SizeNode =
DAG.getConstant(Outs[i].Flags.getByValSize(), DL, MVT::i64);
SDValue Cpy = DAG.getMemcpy(
Chain, DL, DstAddr, Arg, SizeNode,
Outs[i].Flags.getNonZeroByValAlign(),
/*isVol = */ false, /*AlwaysInline = */ false,
/*CI=*/nullptr, std::nullopt, DstInfo, MachinePointerInfo());
MemOpChains.push_back(Cpy);
} else {
// Since we pass i1/i8/i16 as i1/i8/i16 on stack and Arg is already
// promoted to a legal register type i32, we should truncate Arg back to
// i1/i8/i16.
if (VA.getValVT() == MVT::i1 || VA.getValVT() == MVT::i8 ||
VA.getValVT() == MVT::i16)
Arg = DAG.getNode(ISD::TRUNCATE, DL, VA.getValVT(), Arg);
SDValue Store = DAG.getStore(Chain, DL, Arg, DstAddr, DstInfo);
MemOpChains.push_back(Store);
}
}
}
if (IsVarArg && Subtarget->isWindowsArm64EC() &&
!(CLI.CB && CLI.CB->isMustTailCall())) {
SDValue ParamPtr = StackPtr;
if (IsTailCall) {
// Create a dummy object at the top of the stack that can be used to get
// the SP after the epilogue
int FI = MF.getFrameInfo().CreateFixedObject(1, FPDiff, true);
ParamPtr = DAG.getFrameIndex(FI, PtrVT);
}
// For vararg calls, the Arm64EC ABI requires values in x4 and x5
// describing the argument list. x4 contains the address of the
// first stack parameter. x5 contains the size in bytes of all parameters
// passed on the stack.
RegsToPass.emplace_back(AArch64::X4, ParamPtr);
RegsToPass.emplace_back(AArch64::X5,
DAG.getConstant(NumBytes, DL, MVT::i64));
}
if (!MemOpChains.empty())
Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, MemOpChains);
SDValue InGlue;
if (RequiresSMChange) {
if (!Subtarget->isTargetDarwin() || Subtarget->hasSVE()) {
Chain = DAG.getNode(AArch64ISD::VG_SAVE, DL,
DAG.getVTList(MVT::Other, MVT::Glue), Chain);
InGlue = Chain.getValue(1);
}
SDValue NewChain = changeStreamingMode(
DAG, DL, CallAttrs.callee().hasStreamingInterface(), Chain, InGlue,
getSMToggleCondition(CallAttrs), PStateSM);
Chain = NewChain.getValue(0);
InGlue = NewChain.getValue(1);
}
// Build a sequence of copy-to-reg nodes chained together with token chain
// and flag operands which copy the outgoing args into the appropriate regs.
for (auto &RegToPass : RegsToPass) {
Chain = DAG.getCopyToReg(Chain, DL, RegToPass.first,
RegToPass.second, InGlue);
InGlue = Chain.getValue(1);
}
// If the callee is a GlobalAddress/ExternalSymbol node (quite common, every
// direct call is) turn it into a TargetGlobalAddress/TargetExternalSymbol
// node so that legalize doesn't hack it.
const GlobalValue *CalledGlobal = nullptr;
unsigned OpFlags = 0;
if (auto *G = dyn_cast<GlobalAddressSDNode>(Callee)) {
CalledGlobal = G->getGlobal();
OpFlags = Subtarget->classifyGlobalFunctionReference(CalledGlobal,
getTargetMachine());
if (OpFlags & AArch64II::MO_GOT) {
Callee = DAG.getTargetGlobalAddress(CalledGlobal, DL, PtrVT, 0, OpFlags);
Callee = DAG.getNode(AArch64ISD::LOADgot, DL, PtrVT, Callee);
} else {
const GlobalValue *GV = G->getGlobal();
Callee = DAG.getTargetGlobalAddress(GV, DL, PtrVT, 0, OpFlags);
}
} else if (auto *S = dyn_cast<ExternalSymbolSDNode>(Callee)) {
bool UseGot = (getTargetMachine().getCodeModel() == CodeModel::Large &&
Subtarget->isTargetMachO()) ||
MF.getFunction().getParent()->getRtLibUseGOT();
const char *Sym = S->getSymbol();
if (UseGot) {
Callee = DAG.getTargetExternalSymbol(Sym, PtrVT, AArch64II::MO_GOT);
Callee = DAG.getNode(AArch64ISD::LOADgot, DL, PtrVT, Callee);
} else {
Callee = DAG.getTargetExternalSymbol(Sym, PtrVT, 0);
}
}
// We don't usually want to end the call-sequence here because we would tidy
// the frame up *after* the call, however in the ABI-changing tail-call case
// we've carefully laid out the parameters so that when sp is reset they'll be
// in the correct location.
if (IsTailCall && !IsSibCall) {
Chain = DAG.getCALLSEQ_END(Chain, 0, 0, InGlue, DL);
InGlue = Chain.getValue(1);
}
unsigned Opc = IsTailCall ? AArch64ISD::TC_RETURN : AArch64ISD::CALL;
std::vector<SDValue> Ops;
Ops.push_back(Chain);
Ops.push_back(Callee);
// Calls with operand bundle "clang.arc.attachedcall" are special. They should
// be expanded to the call, directly followed by a special marker sequence and
// a call to an ObjC library function. Use CALL_RVMARKER to do that.
if (CLI.CB && objcarc::hasAttachedCallOpBundle(CLI.CB)) {
assert(!IsTailCall &&
"tail calls cannot be marked with clang.arc.attachedcall");
Opc = AArch64ISD::CALL_RVMARKER;
// Add a target global address for the retainRV/claimRV runtime function
// just before the call target.
Function *ARCFn = *objcarc::getAttachedARCFunction(CLI.CB);
auto GA = DAG.getTargetGlobalAddress(ARCFn, DL, PtrVT);
Ops.insert(Ops.begin() + 1, GA);
// We may or may not need to emit both the marker and the retain/claim call.
// Tell the pseudo expansion using an additional boolean op.
bool ShouldEmitMarker = objcarc::attachedCallOpBundleNeedsMarker(CLI.CB);
SDValue DoEmitMarker =
DAG.getTargetConstant(ShouldEmitMarker, DL, MVT::i32);
Ops.insert(Ops.begin() + 2, DoEmitMarker);
} else if (CallConv == CallingConv::ARM64EC_Thunk_X64) {
Opc = AArch64ISD::CALL_ARM64EC_TO_X64;
} else if (GuardWithBTI) {
Opc = AArch64ISD::CALL_BTI;
}
if (IsTailCall) {
// Each tail call may have to adjust the stack by a different amount, so
// this information must travel along with the operation for eventual
// consumption by emitEpilogue.
Ops.push_back(DAG.getSignedTargetConstant(FPDiff, DL, MVT::i32));
}
if (CLI.PAI) {
const uint64_t Key = CLI.PAI->Key;
assert((Key == AArch64PACKey::IA || Key == AArch64PACKey::IB) &&
"Invalid auth call key");
// Split the discriminator into address/integer components.
SDValue AddrDisc, IntDisc;
std::tie(IntDisc, AddrDisc) =
extractPtrauthBlendDiscriminators(CLI.PAI->Discriminator, &DAG);
if (Opc == AArch64ISD::CALL_RVMARKER)
Opc = AArch64ISD::AUTH_CALL_RVMARKER;
else
Opc = IsTailCall ? AArch64ISD::AUTH_TC_RETURN : AArch64ISD::AUTH_CALL;
Ops.push_back(DAG.getTargetConstant(Key, DL, MVT::i32));
Ops.push_back(IntDisc);
Ops.push_back(AddrDisc);
}
// Add argument registers to the end of the list so that they are known live
// into the call.
for (auto &RegToPass : RegsToPass)
Ops.push_back(DAG.getRegister(RegToPass.first,
RegToPass.second.getValueType()));
// Add a register mask operand representing the call-preserved registers.
const uint32_t *Mask;
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
if (IsThisReturn) {
// For 'this' returns, use the X0-preserving mask if applicable
Mask = TRI->getThisReturnPreservedMask(MF, CallConv);
if (!Mask) {
IsThisReturn = false;
Mask = TRI->getCallPreservedMask(MF, CallConv);
}
} else
Mask = TRI->getCallPreservedMask(MF, CallConv);
if (Subtarget->hasCustomCallingConv())
TRI->UpdateCustomCallPreservedMask(MF, &Mask);
if (TRI->isAnyArgRegReserved(MF))
TRI->emitReservedArgRegCallError(MF);
assert(Mask && "Missing call preserved mask for calling convention");
Ops.push_back(DAG.getRegisterMask(Mask));
if (InGlue.getNode())
Ops.push_back(InGlue);
// If we're doing a tall call, use a TC_RETURN here rather than an
// actual call instruction.
if (IsTailCall) {
MF.getFrameInfo().setHasTailCall();
SDValue Ret = DAG.getNode(Opc, DL, MVT::Other, Ops);
if (IsCFICall)
Ret.getNode()->setCFIType(CLI.CFIType->getZExtValue());
DAG.addNoMergeSiteInfo(Ret.getNode(), CLI.NoMerge);
DAG.addCallSiteInfo(Ret.getNode(), std::move(CSInfo));
if (CalledGlobal &&
MF.getFunction().getParent()->getModuleFlag("import-call-optimization"))
DAG.addCalledGlobal(Ret.getNode(), CalledGlobal, OpFlags);
return Ret;
}
// Returns a chain and a flag for retval copy to use.
Chain = DAG.getNode(Opc, DL, {MVT::Other, MVT::Glue}, Ops);
if (IsCFICall)
Chain.getNode()->setCFIType(CLI.CFIType->getZExtValue());
DAG.addNoMergeSiteInfo(Chain.getNode(), CLI.NoMerge);
InGlue = Chain.getValue(1);
DAG.addCallSiteInfo(Chain.getNode(), std::move(CSInfo));
if (CalledGlobal &&
MF.getFunction().getParent()->getModuleFlag("import-call-optimization"))
DAG.addCalledGlobal(Chain.getNode(), CalledGlobal, OpFlags);
uint64_t CalleePopBytes =
DoesCalleeRestoreStack(CallConv, TailCallOpt) ? alignTo(NumBytes, 16) : 0;
Chain = DAG.getCALLSEQ_END(Chain, NumBytes, CalleePopBytes, InGlue, DL);
InGlue = Chain.getValue(1);
// Handle result values, copying them out of physregs into vregs that we
// return.
SDValue Result = LowerCallResult(
Chain, InGlue, CallConv, IsVarArg, RVLocs, DL, DAG, InVals, IsThisReturn,
IsThisReturn ? OutVals[0] : SDValue(), RequiresSMChange);
if (!Ins.empty())
InGlue = Result.getValue(Result->getNumValues() - 1);
if (RequiresSMChange) {
assert(PStateSM && "Expected a PStateSM to be set");
Result = changeStreamingMode(
DAG, DL, !CallAttrs.callee().hasStreamingInterface(), Result, InGlue,
getSMToggleCondition(CallAttrs), PStateSM);
if (!Subtarget->isTargetDarwin() || Subtarget->hasSVE()) {
InGlue = Result.getValue(1);
Result =
DAG.getNode(AArch64ISD::VG_RESTORE, DL,
DAG.getVTList(MVT::Other, MVT::Glue), {Result, InGlue});
}
}
if (CallAttrs.requiresEnablingZAAfterCall())
// Unconditionally resume ZA.
Result = DAG.getNode(
AArch64ISD::SMSTART, DL, DAG.getVTList(MVT::Other, MVT::Glue), Result,
DAG.getTargetConstant((int32_t)(AArch64SVCR::SVCRZA), DL, MVT::i32));
if (ShouldPreserveZT0)
Result =
DAG.getNode(AArch64ISD::RESTORE_ZT, DL, DAG.getVTList(MVT::Other),
{Result, DAG.getConstant(0, DL, MVT::i32), ZTFrameIdx});
if (RequiresLazySave) {
// Conditionally restore the lazy save using a pseudo node.
TPIDR2Object &TPIDR2 = FuncInfo->getTPIDR2Obj();
SDValue RegMask = DAG.getRegisterMask(
TRI->SMEABISupportRoutinesCallPreservedMaskFromX0());
SDValue RestoreRoutine = DAG.getTargetExternalSymbol(
"__arm_tpidr2_restore", getPointerTy(DAG.getDataLayout()));
SDValue TPIDR2_EL0 = DAG.getNode(
ISD::INTRINSIC_W_CHAIN, DL, MVT::i64, Result,
DAG.getConstant(Intrinsic::aarch64_sme_get_tpidr2, DL, MVT::i32));
// Copy the address of the TPIDR2 block into X0 before 'calling' the
// RESTORE_ZA pseudo.
SDValue Glue;
SDValue TPIDR2Block = DAG.getFrameIndex(
TPIDR2.FrameIndex,
DAG.getTargetLoweringInfo().getFrameIndexTy(DAG.getDataLayout()));
Result = DAG.getCopyToReg(Result, DL, AArch64::X0, TPIDR2Block, Glue);
Result =
DAG.getNode(AArch64ISD::RESTORE_ZA, DL, MVT::Other,
{Result, TPIDR2_EL0, DAG.getRegister(AArch64::X0, MVT::i64),
RestoreRoutine, RegMask, Result.getValue(1)});
// Finally reset the TPIDR2_EL0 register to 0.
Result = DAG.getNode(
ISD::INTRINSIC_VOID, DL, MVT::Other, Result,
DAG.getConstant(Intrinsic::aarch64_sme_set_tpidr2, DL, MVT::i32),
DAG.getConstant(0, DL, MVT::i64));
TPIDR2.Uses++;
} else if (RequiresSaveAllZA) {
Result = emitSMEStateSaveRestore(*this, DAG, FuncInfo, DL, Result,
/*IsSave=*/false);
}
if (RequiresSMChange || RequiresLazySave || ShouldPreserveZT0 ||
RequiresSaveAllZA) {
for (unsigned I = 0; I < InVals.size(); ++I) {
// The smstart/smstop is chained as part of the call, but when the
// resulting chain is discarded (which happens when the call is not part
// of a chain, e.g. a call to @llvm.cos()), we need to ensure the
// smstart/smstop is chained to the result value. We can do that by doing
// a vreg -> vreg copy.
Register Reg = MF.getRegInfo().createVirtualRegister(
getRegClassFor(InVals[I].getValueType().getSimpleVT()));
SDValue X = DAG.getCopyToReg(Result, DL, Reg, InVals[I]);
InVals[I] = DAG.getCopyFromReg(X, DL, Reg,
InVals[I].getValueType());
}
}
if (CallConv == CallingConv::PreserveNone) {
for (const ISD::OutputArg &O : Outs) {
if (O.Flags.isSwiftSelf() || O.Flags.isSwiftError() ||
O.Flags.isSwiftAsync()) {
MachineFunction &MF = DAG.getMachineFunction();
DAG.getContext()->diagnose(DiagnosticInfoUnsupported(
MF.getFunction(),
"Swift attributes can't be used with preserve_none",
DL.getDebugLoc()));
break;
}
}
}
return Result;
}
bool AArch64TargetLowering::CanLowerReturn(
CallingConv::ID CallConv, MachineFunction &MF, bool isVarArg,
const SmallVectorImpl<ISD::OutputArg> &Outs, LLVMContext &Context,
const Type *RetTy) const {
CCAssignFn *RetCC = CCAssignFnForReturn(CallConv);
SmallVector<CCValAssign, 16> RVLocs;
CCState CCInfo(CallConv, isVarArg, MF, RVLocs, Context);
return CCInfo.CheckReturn(Outs, RetCC);
}
SDValue
AArch64TargetLowering::LowerReturn(SDValue Chain, CallingConv::ID CallConv,
bool isVarArg,
const SmallVectorImpl<ISD::OutputArg> &Outs,
const SmallVectorImpl<SDValue> &OutVals,
const SDLoc &DL, SelectionDAG &DAG) const {
auto &MF = DAG.getMachineFunction();
auto *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
CCAssignFn *RetCC = CCAssignFnForReturn(CallConv);
SmallVector<CCValAssign, 16> RVLocs;
CCState CCInfo(CallConv, isVarArg, MF, RVLocs, *DAG.getContext());
CCInfo.AnalyzeReturn(Outs, RetCC);
// Copy the result values into the output registers.
SDValue Glue;
SmallVector<std::pair<unsigned, SDValue>, 4> RetVals;
SmallSet<unsigned, 4> RegsUsed;
for (unsigned i = 0, realRVLocIdx = 0; i != RVLocs.size();
++i, ++realRVLocIdx) {
CCValAssign &VA = RVLocs[i];
assert(VA.isRegLoc() && "Can only return in registers!");
SDValue Arg = OutVals[realRVLocIdx];
switch (VA.getLocInfo()) {
default:
llvm_unreachable("Unknown loc info!");
case CCValAssign::Full:
if (Outs[i].ArgVT == MVT::i1) {
// AAPCS requires i1 to be zero-extended to i8 by the producer of the
// value. This is strictly redundant on Darwin (which uses "zeroext
// i1"), but will be optimised out before ISel.
Arg = DAG.getNode(ISD::TRUNCATE, DL, MVT::i1, Arg);
Arg = DAG.getNode(ISD::ZERO_EXTEND, DL, VA.getLocVT(), Arg);
}
break;
case CCValAssign::BCvt:
Arg = DAG.getNode(ISD::BITCAST, DL, VA.getLocVT(), Arg);
break;
case CCValAssign::AExt:
case CCValAssign::ZExt:
Arg = DAG.getZExtOrTrunc(Arg, DL, VA.getLocVT());
break;
case CCValAssign::AExtUpper:
assert(VA.getValVT() == MVT::i32 && "only expect 32 -> 64 upper bits");
Arg = DAG.getZExtOrTrunc(Arg, DL, VA.getLocVT());
Arg = DAG.getNode(ISD::SHL, DL, VA.getLocVT(), Arg,
DAG.getConstant(32, DL, VA.getLocVT()));
break;
}
if (RegsUsed.count(VA.getLocReg())) {
SDValue &Bits =
llvm::find_if(RetVals, [=](const std::pair<unsigned, SDValue> &Elt) {
return Elt.first == VA.getLocReg();
})->second;
Bits = DAG.getNode(ISD::OR, DL, Bits.getValueType(), Bits, Arg);
} else {
RetVals.emplace_back(VA.getLocReg(), Arg);
RegsUsed.insert(VA.getLocReg());
}
}
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
// Emit SMSTOP before returning from a locally streaming function
SMEAttrs FuncAttrs = FuncInfo->getSMEFnAttrs();
if (FuncAttrs.hasStreamingBody() && !FuncAttrs.hasStreamingInterface()) {
if (FuncAttrs.hasStreamingCompatibleInterface()) {
Register Reg = FuncInfo->getPStateSMReg();
assert(Reg.isValid() && "PStateSM Register is invalid");
SDValue PStateSM = DAG.getCopyFromReg(Chain, DL, Reg, MVT::i64);
Chain = changeStreamingMode(DAG, DL, /*Enable*/ false, Chain,
/*Glue*/ SDValue(),
AArch64SME::IfCallerIsNonStreaming, PStateSM);
} else
Chain = changeStreamingMode(DAG, DL, /*Enable*/ false, Chain,
/*Glue*/ SDValue(), AArch64SME::Always);
Glue = Chain.getValue(1);
}
SmallVector<SDValue, 4> RetOps(1, Chain);
for (auto &RetVal : RetVals) {
if (FuncAttrs.hasStreamingBody() && !FuncAttrs.hasStreamingInterface() &&
isPassedInFPR(RetVal.second.getValueType()))
RetVal.second =
DAG.getNode(AArch64ISD::COALESCER_BARRIER, DL,
DAG.getVTList(RetVal.second.getValueType(), MVT::Glue),
RetVal.second);
Chain = DAG.getCopyToReg(Chain, DL, RetVal.first, RetVal.second, Glue);
Glue = Chain.getValue(1);
RetOps.push_back(
DAG.getRegister(RetVal.first, RetVal.second.getValueType()));
}
// Windows AArch64 ABIs require that for returning structs by value we copy
// the sret argument into X0 for the return.
// We saved the argument into a virtual register in the entry block,
// so now we copy the value out and into X0.
if (unsigned SRetReg = FuncInfo->getSRetReturnReg()) {
SDValue Val = DAG.getCopyFromReg(RetOps[0], DL, SRetReg,
getPointerTy(MF.getDataLayout()));
unsigned RetValReg = AArch64::X0;
if (CallConv == CallingConv::ARM64EC_Thunk_X64)
RetValReg = AArch64::X8;
Chain = DAG.getCopyToReg(Chain, DL, RetValReg, Val, Glue);
Glue = Chain.getValue(1);
RetOps.push_back(
DAG.getRegister(RetValReg, getPointerTy(DAG.getDataLayout())));
}
const MCPhysReg *I = TRI->getCalleeSavedRegsViaCopy(&MF);
if (I) {
for (; *I; ++I) {
if (AArch64::GPR64RegClass.contains(*I))
RetOps.push_back(DAG.getRegister(*I, MVT::i64));
else if (AArch64::FPR64RegClass.contains(*I))
RetOps.push_back(DAG.getRegister(*I, MVT::getFloatingPointVT(64)));
else
llvm_unreachable("Unexpected register class in CSRsViaCopy!");
}
}
RetOps[0] = Chain; // Update chain.
// Add the glue if we have it.
if (Glue.getNode())
RetOps.push_back(Glue);
if (CallConv == CallingConv::ARM64EC_Thunk_X64) {
// ARM64EC entry thunks use a special return sequence: instead of a regular
// "ret" instruction, they need to explicitly call the emulator.
EVT PtrVT = getPointerTy(DAG.getDataLayout());
SDValue Arm64ECRetDest =
DAG.getExternalSymbol("__os_arm64x_dispatch_ret", PtrVT);
Arm64ECRetDest =
getAddr(cast<ExternalSymbolSDNode>(Arm64ECRetDest), DAG, 0);
Arm64ECRetDest = DAG.getLoad(PtrVT, DL, DAG.getEntryNode(), Arm64ECRetDest,
MachinePointerInfo());
RetOps.insert(RetOps.begin() + 1, Arm64ECRetDest);
RetOps.insert(RetOps.begin() + 2, DAG.getTargetConstant(0, DL, MVT::i32));
return DAG.getNode(AArch64ISD::TC_RETURN, DL, MVT::Other, RetOps);
}
return DAG.getNode(AArch64ISD::RET_GLUE, DL, MVT::Other, RetOps);
}
//===----------------------------------------------------------------------===//
// Other Lowering Code
//===----------------------------------------------------------------------===//
SDValue AArch64TargetLowering::getTargetNode(GlobalAddressSDNode *N, EVT Ty,
SelectionDAG &DAG,
unsigned Flag) const {
return DAG.getTargetGlobalAddress(N->getGlobal(), SDLoc(N), Ty,
N->getOffset(), Flag);
}
SDValue AArch64TargetLowering::getTargetNode(JumpTableSDNode *N, EVT Ty,
SelectionDAG &DAG,
unsigned Flag) const {
return DAG.getTargetJumpTable(N->getIndex(), Ty, Flag);
}
SDValue AArch64TargetLowering::getTargetNode(ConstantPoolSDNode *N, EVT Ty,
SelectionDAG &DAG,
unsigned Flag) const {
return DAG.getTargetConstantPool(N->getConstVal(), Ty, N->getAlign(),
N->getOffset(), Flag);
}
SDValue AArch64TargetLowering::getTargetNode(BlockAddressSDNode* N, EVT Ty,
SelectionDAG &DAG,
unsigned Flag) const {
return DAG.getTargetBlockAddress(N->getBlockAddress(), Ty, 0, Flag);
}
SDValue AArch64TargetLowering::getTargetNode(ExternalSymbolSDNode *N, EVT Ty,
SelectionDAG &DAG,
unsigned Flag) const {
return DAG.getTargetExternalSymbol(N->getSymbol(), Ty, Flag);
}
// (loadGOT sym)
template <class NodeTy>
SDValue AArch64TargetLowering::getGOT(NodeTy *N, SelectionDAG &DAG,
unsigned Flags) const {
LLVM_DEBUG(dbgs() << "AArch64TargetLowering::getGOT\n");
SDLoc DL(N);
EVT Ty = getPointerTy(DAG.getDataLayout());
SDValue GotAddr = getTargetNode(N, Ty, DAG, AArch64II::MO_GOT | Flags);
// FIXME: Once remat is capable of dealing with instructions with register
// operands, expand this into two nodes instead of using a wrapper node.
if (DAG.getMachineFunction()
.getInfo<AArch64FunctionInfo>()
->hasELFSignedGOT())
return SDValue(DAG.getMachineNode(AArch64::LOADgotAUTH, DL, Ty, GotAddr),
0);
return DAG.getNode(AArch64ISD::LOADgot, DL, Ty, GotAddr);
}
// (wrapper %highest(sym), %higher(sym), %hi(sym), %lo(sym))
template <class NodeTy>
SDValue AArch64TargetLowering::getAddrLarge(NodeTy *N, SelectionDAG &DAG,
unsigned Flags) const {
LLVM_DEBUG(dbgs() << "AArch64TargetLowering::getAddrLarge\n");
SDLoc DL(N);
EVT Ty = getPointerTy(DAG.getDataLayout());
const unsigned char MO_NC = AArch64II::MO_NC;
return DAG.getNode(
AArch64ISD::WrapperLarge, DL, Ty,
getTargetNode(N, Ty, DAG, AArch64II::MO_G3 | Flags),
getTargetNode(N, Ty, DAG, AArch64II::MO_G2 | MO_NC | Flags),
getTargetNode(N, Ty, DAG, AArch64II::MO_G1 | MO_NC | Flags),
getTargetNode(N, Ty, DAG, AArch64II::MO_G0 | MO_NC | Flags));
}
// (addlow (adrp %hi(sym)) %lo(sym))
template <class NodeTy>
SDValue AArch64TargetLowering::getAddr(NodeTy *N, SelectionDAG &DAG,
unsigned Flags) const {
LLVM_DEBUG(dbgs() << "AArch64TargetLowering::getAddr\n");
SDLoc DL(N);
EVT Ty = getPointerTy(DAG.getDataLayout());
SDValue Hi = getTargetNode(N, Ty, DAG, AArch64II::MO_PAGE | Flags);
SDValue Lo = getTargetNode(N, Ty, DAG,
AArch64II::MO_PAGEOFF | AArch64II::MO_NC | Flags);
SDValue ADRP = DAG.getNode(AArch64ISD::ADRP, DL, Ty, Hi);
return DAG.getNode(AArch64ISD::ADDlow, DL, Ty, ADRP, Lo);
}
// (adr sym)
template <class NodeTy>
SDValue AArch64TargetLowering::getAddrTiny(NodeTy *N, SelectionDAG &DAG,
unsigned Flags) const {
LLVM_DEBUG(dbgs() << "AArch64TargetLowering::getAddrTiny\n");
SDLoc DL(N);
EVT Ty = getPointerTy(DAG.getDataLayout());
SDValue Sym = getTargetNode(N, Ty, DAG, Flags);
return DAG.getNode(AArch64ISD::ADR, DL, Ty, Sym);
}
SDValue AArch64TargetLowering::LowerGlobalAddress(SDValue Op,
SelectionDAG &DAG) const {
GlobalAddressSDNode *GN = cast<GlobalAddressSDNode>(Op);
const GlobalValue *GV = GN->getGlobal();
unsigned OpFlags = Subtarget->ClassifyGlobalReference(GV, getTargetMachine());
if (OpFlags != AArch64II::MO_NO_FLAG)
assert(cast<GlobalAddressSDNode>(Op)->getOffset() == 0 &&
"unexpected offset in global node");
// This also catches the large code model case for Darwin, and tiny code
// model with got relocations.
if ((OpFlags & AArch64II::MO_GOT) != 0) {
return getGOT(GN, DAG, OpFlags);
}
SDValue Result;
if (getTargetMachine().getCodeModel() == CodeModel::Large &&
!getTargetMachine().isPositionIndependent()) {
Result = getAddrLarge(GN, DAG, OpFlags);
} else if (getTargetMachine().getCodeModel() == CodeModel::Tiny) {
Result = getAddrTiny(GN, DAG, OpFlags);
} else {
Result = getAddr(GN, DAG, OpFlags);
}
EVT PtrVT = getPointerTy(DAG.getDataLayout());
SDLoc DL(GN);
if (OpFlags & (AArch64II::MO_DLLIMPORT | AArch64II::MO_COFFSTUB))
Result = DAG.getLoad(PtrVT, DL, DAG.getEntryNode(), Result,
MachinePointerInfo::getGOT(DAG.getMachineFunction()));
return Result;
}
/// Convert a TLS address reference into the correct sequence of loads
/// and calls to compute the variable's address (for Darwin, currently) and
/// return an SDValue containing the final node.
/// Darwin only has one TLS scheme which must be capable of dealing with the
/// fully general situation, in the worst case. This means:
/// + "extern __thread" declaration.
/// + Defined in a possibly unknown dynamic library.
///
/// The general system is that each __thread variable has a [3 x i64] descriptor
/// which contains information used by the runtime to calculate the address. The
/// only part of this the compiler needs to know about is the first xword, which
/// contains a function pointer that must be called with the address of the
/// entire descriptor in "x0".
///
/// Since this descriptor may be in a different unit, in general even the
/// descriptor must be accessed via an indirect load. The "ideal" code sequence
/// is:
/// adrp x0, _var@TLVPPAGE
/// ldr x0, [x0, _var@TLVPPAGEOFF] ; x0 now contains address of descriptor
/// ldr x1, [x0] ; x1 contains 1st entry of descriptor,
/// ; the function pointer
/// blr x1 ; Uses descriptor address in x0
/// ; Address of _var is now in x0.
///
/// If the address of _var's descriptor *is* known to the linker, then it can
/// change the first "ldr" instruction to an appropriate "add x0, x0, #imm" for
/// a slight efficiency gain.
SDValue
AArch64TargetLowering::LowerDarwinGlobalTLSAddress(SDValue Op,
SelectionDAG &DAG) const {
assert(Subtarget->isTargetDarwin() &&
"This function expects a Darwin target");
SDLoc DL(Op);
MVT PtrVT = getPointerTy(DAG.getDataLayout());
MVT PtrMemVT = getPointerMemTy(DAG.getDataLayout());
const GlobalValue *GV = cast<GlobalAddressSDNode>(Op)->getGlobal();
SDValue TLVPAddr =
DAG.getTargetGlobalAddress(GV, DL, PtrVT, 0, AArch64II::MO_TLS);
SDValue DescAddr = DAG.getNode(AArch64ISD::LOADgot, DL, PtrVT, TLVPAddr);
// The first entry in the descriptor is a function pointer that we must call
// to obtain the address of the variable.
SDValue Chain = DAG.getEntryNode();
SDValue FuncTLVGet = DAG.getLoad(
PtrMemVT, DL, Chain, DescAddr,
MachinePointerInfo::getGOT(DAG.getMachineFunction()),
Align(PtrMemVT.getSizeInBits() / 8),
MachineMemOperand::MOInvariant | MachineMemOperand::MODereferenceable);
Chain = FuncTLVGet.getValue(1);
// Extend loaded pointer if necessary (i.e. if ILP32) to DAG pointer.
FuncTLVGet = DAG.getZExtOrTrunc(FuncTLVGet, DL, PtrVT);
MachineFrameInfo &MFI = DAG.getMachineFunction().getFrameInfo();
MFI.setAdjustsStack(true);
// TLS calls preserve all registers except those that absolutely must be
// trashed: X0 (it takes an argument), LR (it's a call) and NZCV (let's not be
// silly).
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
const uint32_t *Mask = TRI->getTLSCallPreservedMask();
if (Subtarget->hasCustomCallingConv())
TRI->UpdateCustomCallPreservedMask(DAG.getMachineFunction(), &Mask);
// Finally, we can make the call. This is just a degenerate version of a
// normal AArch64 call node: x0 takes the address of the descriptor, and
// returns the address of the variable in this thread.
Chain = DAG.getCopyToReg(Chain, DL, AArch64::X0, DescAddr, SDValue());
unsigned Opcode = AArch64ISD::CALL;
SmallVector<SDValue, 8> Ops;
Ops.push_back(Chain);
Ops.push_back(FuncTLVGet);
// With ptrauth-calls, the tlv access thunk pointer is authenticated (IA, 0).
if (DAG.getMachineFunction().getFunction().hasFnAttribute("ptrauth-calls")) {
Opcode = AArch64ISD::AUTH_CALL;
Ops.push_back(DAG.getTargetConstant(AArch64PACKey::IA, DL, MVT::i32));
Ops.push_back(DAG.getTargetConstant(0, DL, MVT::i64)); // Integer Disc.
Ops.push_back(DAG.getRegister(AArch64::NoRegister, MVT::i64)); // Addr Disc.
}
Ops.push_back(DAG.getRegister(AArch64::X0, MVT::i64));
Ops.push_back(DAG.getRegisterMask(Mask));
Ops.push_back(Chain.getValue(1));
Chain = DAG.getNode(Opcode, DL, DAG.getVTList(MVT::Other, MVT::Glue), Ops);
return DAG.getCopyFromReg(Chain, DL, AArch64::X0, PtrVT, Chain.getValue(1));
}
/// Convert a thread-local variable reference into a sequence of instructions to
/// compute the variable's address for the local exec TLS model of ELF targets.
/// The sequence depends on the maximum TLS area size.
SDValue AArch64TargetLowering::LowerELFTLSLocalExec(const GlobalValue *GV,
SDValue ThreadBase,
const SDLoc &DL,
SelectionDAG &DAG) const {
EVT PtrVT = getPointerTy(DAG.getDataLayout());
SDValue TPOff, Addr;
switch (DAG.getTarget().Options.TLSSize) {
default:
llvm_unreachable("Unexpected TLS size");
case 12: {
// mrs x0, TPIDR_EL0
// add x0, x0, :tprel_lo12:a
SDValue Var = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0, AArch64II::MO_TLS | AArch64II::MO_PAGEOFF);
return SDValue(DAG.getMachineNode(AArch64::ADDXri, DL, PtrVT, ThreadBase,
Var,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
}
case 24: {
// mrs x0, TPIDR_EL0
// add x0, x0, :tprel_hi12:a
// add x0, x0, :tprel_lo12_nc:a
SDValue HiVar = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0, AArch64II::MO_TLS | AArch64II::MO_HI12);
SDValue LoVar = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0,
AArch64II::MO_TLS | AArch64II::MO_PAGEOFF | AArch64II::MO_NC);
Addr = SDValue(DAG.getMachineNode(AArch64::ADDXri, DL, PtrVT, ThreadBase,
HiVar,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
return SDValue(DAG.getMachineNode(AArch64::ADDXri, DL, PtrVT, Addr,
LoVar,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
}
case 32: {
// mrs x1, TPIDR_EL0
// movz x0, #:tprel_g1:a
// movk x0, #:tprel_g0_nc:a
// add x0, x1, x0
SDValue HiVar = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0, AArch64II::MO_TLS | AArch64II::MO_G1);
SDValue LoVar = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0,
AArch64II::MO_TLS | AArch64II::MO_G0 | AArch64II::MO_NC);
TPOff = SDValue(DAG.getMachineNode(AArch64::MOVZXi, DL, PtrVT, HiVar,
DAG.getTargetConstant(16, DL, MVT::i32)),
0);
TPOff = SDValue(DAG.getMachineNode(AArch64::MOVKXi, DL, PtrVT, TPOff, LoVar,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
return DAG.getNode(ISD::ADD, DL, PtrVT, ThreadBase, TPOff);
}
case 48: {
// mrs x1, TPIDR_EL0
// movz x0, #:tprel_g2:a
// movk x0, #:tprel_g1_nc:a
// movk x0, #:tprel_g0_nc:a
// add x0, x1, x0
SDValue HiVar = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0, AArch64II::MO_TLS | AArch64II::MO_G2);
SDValue MiVar = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0,
AArch64II::MO_TLS | AArch64II::MO_G1 | AArch64II::MO_NC);
SDValue LoVar = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0,
AArch64II::MO_TLS | AArch64II::MO_G0 | AArch64II::MO_NC);
TPOff = SDValue(DAG.getMachineNode(AArch64::MOVZXi, DL, PtrVT, HiVar,
DAG.getTargetConstant(32, DL, MVT::i32)),
0);
TPOff = SDValue(DAG.getMachineNode(AArch64::MOVKXi, DL, PtrVT, TPOff, MiVar,
DAG.getTargetConstant(16, DL, MVT::i32)),
0);
TPOff = SDValue(DAG.getMachineNode(AArch64::MOVKXi, DL, PtrVT, TPOff, LoVar,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
return DAG.getNode(ISD::ADD, DL, PtrVT, ThreadBase, TPOff);
}
}
}
/// When accessing thread-local variables under either the general-dynamic or
/// local-dynamic system, we make a "TLS-descriptor" call. The variable will
/// have a descriptor, accessible via a PC-relative ADRP, and whose first entry
/// is a function pointer to carry out the resolution.
///
/// The sequence is:
/// adrp x0, :tlsdesc:var
/// ldr x1, [x0, #:tlsdesc_lo12:var]
/// add x0, x0, #:tlsdesc_lo12:var
/// .tlsdesccall var
/// blr x1
/// (TPIDR_EL0 offset now in x0)
///
/// The above sequence must be produced unscheduled, to enable the linker to
/// optimize/relax this sequence.
/// Therefore, a pseudo-instruction (TLSDESC_CALLSEQ) is used to represent the
/// above sequence, and expanded really late in the compilation flow, to ensure
/// the sequence is produced as per above.
SDValue AArch64TargetLowering::LowerELFTLSDescCallSeq(SDValue SymAddr,
const SDLoc &DL,
SelectionDAG &DAG) const {
EVT PtrVT = getPointerTy(DAG.getDataLayout());
SDValue Chain = DAG.getEntryNode();
SDVTList NodeTys = DAG.getVTList(MVT::Other, MVT::Glue);
unsigned Opcode =
DAG.getMachineFunction().getInfo<AArch64FunctionInfo>()->hasELFSignedGOT()
? AArch64ISD::TLSDESC_AUTH_CALLSEQ
: AArch64ISD::TLSDESC_CALLSEQ;
Chain = DAG.getNode(Opcode, DL, NodeTys, {Chain, SymAddr});
SDValue Glue = Chain.getValue(1);
return DAG.getCopyFromReg(Chain, DL, AArch64::X0, PtrVT, Glue);
}
SDValue
AArch64TargetLowering::LowerELFGlobalTLSAddress(SDValue Op,
SelectionDAG &DAG) const {
assert(Subtarget->isTargetELF() && "This function expects an ELF target");
const GlobalAddressSDNode *GA = cast<GlobalAddressSDNode>(Op);
AArch64FunctionInfo *MFI =
DAG.getMachineFunction().getInfo<AArch64FunctionInfo>();
TLSModel::Model Model = MFI->hasELFSignedGOT()
? TLSModel::GeneralDynamic
: getTargetMachine().getTLSModel(GA->getGlobal());
if (!EnableAArch64ELFLocalDynamicTLSGeneration) {
if (Model == TLSModel::LocalDynamic)
Model = TLSModel::GeneralDynamic;
}
if (getTargetMachine().getCodeModel() == CodeModel::Large &&
Model != TLSModel::LocalExec)
report_fatal_error("ELF TLS only supported in small memory model or "
"in local exec TLS model");
// Different choices can be made for the maximum size of the TLS area for a
// module. For the small address model, the default TLS size is 16MiB and the
// maximum TLS size is 4GiB.
// FIXME: add tiny and large code model support for TLS access models other
// than local exec. We currently generate the same code as small for tiny,
// which may be larger than needed.
SDValue TPOff;
EVT PtrVT = getPointerTy(DAG.getDataLayout());
SDLoc DL(Op);
const GlobalValue *GV = GA->getGlobal();
SDValue ThreadBase = DAG.getNode(AArch64ISD::THREAD_POINTER, DL, PtrVT);
if (Model == TLSModel::LocalExec) {
return LowerELFTLSLocalExec(GV, ThreadBase, DL, DAG);
} else if (Model == TLSModel::InitialExec) {
TPOff = DAG.getTargetGlobalAddress(GV, DL, PtrVT, 0, AArch64II::MO_TLS);
TPOff = DAG.getNode(AArch64ISD::LOADgot, DL, PtrVT, TPOff);
} else if (Model == TLSModel::LocalDynamic) {
// Local-dynamic accesses proceed in two phases. A general-dynamic TLS
// descriptor call against the special symbol _TLS_MODULE_BASE_ to calculate
// the beginning of the module's TLS region, followed by a DTPREL offset
// calculation.
// These accesses will need deduplicating if there's more than one.
MFI->incNumLocalDynamicTLSAccesses();
// The call needs a relocation too for linker relaxation. It doesn't make
// sense to call it MO_PAGE or MO_PAGEOFF though so we need another copy of
// the address.
SDValue SymAddr = DAG.getTargetExternalSymbol("_TLS_MODULE_BASE_", PtrVT,
AArch64II::MO_TLS);
// Now we can calculate the offset from TPIDR_EL0 to this module's
// thread-local area.
TPOff = LowerELFTLSDescCallSeq(SymAddr, DL, DAG);
// Now use :dtprel_whatever: operations to calculate this variable's offset
// in its thread-storage area.
SDValue HiVar = DAG.getTargetGlobalAddress(
GV, DL, MVT::i64, 0, AArch64II::MO_TLS | AArch64II::MO_HI12);
SDValue LoVar = DAG.getTargetGlobalAddress(
GV, DL, MVT::i64, 0,
AArch64II::MO_TLS | AArch64II::MO_PAGEOFF | AArch64II::MO_NC);
TPOff = SDValue(DAG.getMachineNode(AArch64::ADDXri, DL, PtrVT, TPOff, HiVar,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
TPOff = SDValue(DAG.getMachineNode(AArch64::ADDXri, DL, PtrVT, TPOff, LoVar,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
} else if (Model == TLSModel::GeneralDynamic) {
// The call needs a relocation too for linker relaxation. It doesn't make
// sense to call it MO_PAGE or MO_PAGEOFF though so we need another copy of
// the address.
SDValue SymAddr =
DAG.getTargetGlobalAddress(GV, DL, PtrVT, 0, AArch64II::MO_TLS);
// Finally we can make a call to calculate the offset from tpidr_el0.
TPOff = LowerELFTLSDescCallSeq(SymAddr, DL, DAG);
} else
llvm_unreachable("Unsupported ELF TLS access model");
return DAG.getNode(ISD::ADD, DL, PtrVT, ThreadBase, TPOff);
}
SDValue
AArch64TargetLowering::LowerWindowsGlobalTLSAddress(SDValue Op,
SelectionDAG &DAG) const {
assert(Subtarget->isTargetWindows() && "Windows specific TLS lowering");
SDValue Chain = DAG.getEntryNode();
EVT PtrVT = getPointerTy(DAG.getDataLayout());
SDLoc DL(Op);
SDValue TEB = DAG.getRegister(AArch64::X18, MVT::i64);
// Load the ThreadLocalStoragePointer from the TEB
// A pointer to the TLS array is located at offset 0x58 from the TEB.
SDValue TLSArray =
DAG.getNode(ISD::ADD, DL, PtrVT, TEB, DAG.getIntPtrConstant(0x58, DL));
TLSArray = DAG.getLoad(PtrVT, DL, Chain, TLSArray, MachinePointerInfo());
Chain = TLSArray.getValue(1);
// Load the TLS index from the C runtime;
// This does the same as getAddr(), but without having a GlobalAddressSDNode.
// This also does the same as LOADgot, but using a generic i32 load,
// while LOADgot only loads i64.
SDValue TLSIndexHi =
DAG.getTargetExternalSymbol("_tls_index", PtrVT, AArch64II::MO_PAGE);
SDValue TLSIndexLo = DAG.getTargetExternalSymbol(
"_tls_index", PtrVT, AArch64II::MO_PAGEOFF | AArch64II::MO_NC);
SDValue ADRP = DAG.getNode(AArch64ISD::ADRP, DL, PtrVT, TLSIndexHi);
SDValue TLSIndex =
DAG.getNode(AArch64ISD::ADDlow, DL, PtrVT, ADRP, TLSIndexLo);
TLSIndex = DAG.getLoad(MVT::i32, DL, Chain, TLSIndex, MachinePointerInfo());
Chain = TLSIndex.getValue(1);
// The pointer to the thread's TLS data area is at the TLS Index scaled by 8
// offset into the TLSArray.
TLSIndex = DAG.getNode(ISD::ZERO_EXTEND, DL, PtrVT, TLSIndex);
SDValue Slot = DAG.getNode(ISD::SHL, DL, PtrVT, TLSIndex,
DAG.getConstant(3, DL, PtrVT));
SDValue TLS = DAG.getLoad(PtrVT, DL, Chain,
DAG.getNode(ISD::ADD, DL, PtrVT, TLSArray, Slot),
MachinePointerInfo());
Chain = TLS.getValue(1);
const GlobalAddressSDNode *GA = cast<GlobalAddressSDNode>(Op);
const GlobalValue *GV = GA->getGlobal();
SDValue TGAHi = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0, AArch64II::MO_TLS | AArch64II::MO_HI12);
SDValue TGALo = DAG.getTargetGlobalAddress(
GV, DL, PtrVT, 0,
AArch64II::MO_TLS | AArch64II::MO_PAGEOFF | AArch64II::MO_NC);
// Add the offset from the start of the .tls section (section base).
SDValue Addr =
SDValue(DAG.getMachineNode(AArch64::ADDXri, DL, PtrVT, TLS, TGAHi,
DAG.getTargetConstant(0, DL, MVT::i32)),
0);
Addr = DAG.getNode(AArch64ISD::ADDlow, DL, PtrVT, Addr, TGALo);
return Addr;
}
SDValue AArch64TargetLowering::LowerGlobalTLSAddress(SDValue Op,
SelectionDAG &DAG) const {
const GlobalAddressSDNode *GA = cast<GlobalAddressSDNode>(Op);
if (DAG.getTarget().useEmulatedTLS())
return LowerToTLSEmulatedModel(GA, DAG);
if (Subtarget->isTargetDarwin())
return LowerDarwinGlobalTLSAddress(Op, DAG);
if (Subtarget->isTargetELF())
return LowerELFGlobalTLSAddress(Op, DAG);
if (Subtarget->isTargetWindows())
return LowerWindowsGlobalTLSAddress(Op, DAG);
llvm_unreachable("Unexpected platform trying to use TLS");
}
//===----------------------------------------------------------------------===//
// PtrAuthGlobalAddress lowering
//
// We have 3 lowering alternatives to choose from:
// - MOVaddrPAC: similar to MOVaddr, with added PAC.
// If the GV doesn't need a GOT load (i.e., is locally defined)
// materialize the pointer using adrp+add+pac. See LowerMOVaddrPAC.
//
// - LOADgotPAC: similar to LOADgot, with added PAC.
// If the GV needs a GOT load, materialize the pointer using the usual
// GOT adrp+ldr, +pac. Pointers in GOT are assumed to be not signed, the GOT
// section is assumed to be read-only (for example, via relro mechanism). See
// LowerMOVaddrPAC.
//
// - LOADauthptrstatic: similar to LOADgot, but use a
// special stub slot instead of a GOT slot.
// Load a signed pointer for symbol 'sym' from a stub slot named
// 'sym$auth_ptr$key$disc' filled by dynamic linker during relocation
// resolving. This usually lowers to adrp+ldr, but also emits an entry into
// .data with an @AUTH relocation. See LowerLOADauthptrstatic.
//
// All 3 are pseudos that are expand late to longer sequences: this lets us
// provide integrity guarantees on the to-be-signed intermediate values.
//
// LOADauthptrstatic is undesirable because it requires a large section filled
// with often similarly-signed pointers, making it a good harvesting target.
// Thus, it's only used for ptrauth references to extern_weak to avoid null
// checks.
static SDValue LowerPtrAuthGlobalAddressStatically(
SDValue TGA, SDLoc DL, EVT VT, AArch64PACKey::ID KeyC,
SDValue Discriminator, SDValue AddrDiscriminator, SelectionDAG &DAG) {
const auto *TGN = cast<GlobalAddressSDNode>(TGA.getNode());
assert(TGN->getGlobal()->hasExternalWeakLinkage());
// Offsets and extern_weak don't mix well: ptrauth aside, you'd get the
// offset alone as a pointer if the symbol wasn't available, which would
// probably break null checks in users. Ptrauth complicates things further:
// error out.
if (TGN->getOffset() != 0)
report_fatal_error(
"unsupported non-zero offset in weak ptrauth global reference");
if (!isNullConstant(AddrDiscriminator))
report_fatal_error("unsupported weak addr-div ptrauth global");
SDValue Key = DAG.getTargetConstant(KeyC, DL, MVT::i32);
return SDValue(DAG.getMachineNode(AArch64::LOADauthptrstatic, DL, MVT::i64,
{TGA, Key, Discriminator}),
0);
}
SDValue
AArch64TargetLowering::LowerPtrAuthGlobalAddress(SDValue Op,
SelectionDAG &DAG) const {
SDValue Ptr = Op.getOperand(0);
uint64_t KeyC = Op.getConstantOperandVal(1);
SDValue AddrDiscriminator = Op.getOperand(2);
uint64_t DiscriminatorC = Op.getConstantOperandVal(3);
EVT VT = Op.getValueType();
SDLoc DL(Op);
if (KeyC > AArch64PACKey::LAST)
report_fatal_error("key in ptrauth global out of range [0, " +
Twine((int)AArch64PACKey::LAST) + "]");
// Blend only works if the integer discriminator is 16-bit wide.
if (!isUInt<16>(DiscriminatorC))
report_fatal_error(
"constant discriminator in ptrauth global out of range [0, 0xffff]");
// Choosing between 3 lowering alternatives is target-specific.
if (!Subtarget->isTargetELF() && !Subtarget->isTargetMachO())
report_fatal_error("ptrauth global lowering only supported on MachO/ELF");
int64_t PtrOffsetC = 0;
if (Ptr.getOpcode() == ISD::ADD) {
PtrOffsetC = Ptr.getConstantOperandVal(1);
Ptr = Ptr.getOperand(0);
}
const auto *PtrN = cast<GlobalAddressSDNode>(Ptr.getNode());
const GlobalValue *PtrGV = PtrN->getGlobal();
// Classify the reference to determine whether it needs a GOT load.
const unsigned OpFlags =
Subtarget->ClassifyGlobalReference(PtrGV, getTargetMachine());
const bool NeedsGOTLoad = ((OpFlags & AArch64II::MO_GOT) != 0);
assert(((OpFlags & (~AArch64II::MO_GOT)) == 0) &&
"unsupported non-GOT op flags on ptrauth global reference");
// Fold any offset into the GV; our pseudos expect it there.
PtrOffsetC += PtrN->getOffset();
SDValue TPtr = DAG.getTargetGlobalAddress(PtrGV, DL, VT, PtrOffsetC,
/*TargetFlags=*/0);
assert(PtrN->getTargetFlags() == 0 &&
"unsupported target flags on ptrauth global");
SDValue Key = DAG.getTargetConstant(KeyC, DL, MVT::i32);
SDValue Discriminator = DAG.getTargetConstant(DiscriminatorC, DL, MVT::i64);
SDValue TAddrDiscriminator = !isNullConstant(AddrDiscriminator)
? AddrDiscriminator
: DAG.getRegister(AArch64::XZR, MVT::i64);
// No GOT load needed -> MOVaddrPAC
if (!NeedsGOTLoad) {
assert(!PtrGV->hasExternalWeakLinkage() && "extern_weak should use GOT");
return SDValue(
DAG.getMachineNode(AArch64::MOVaddrPAC, DL, MVT::i64,
{TPtr, Key, TAddrDiscriminator, Discriminator}),
0);
}
// GOT load -> LOADgotPAC
// Note that we disallow extern_weak refs to avoid null checks later.
if (!PtrGV->hasExternalWeakLinkage())
return SDValue(
DAG.getMachineNode(AArch64::LOADgotPAC, DL, MVT::i64,
{TPtr, Key, TAddrDiscriminator, Discriminator}),
0);
// extern_weak ref -> LOADauthptrstatic
return LowerPtrAuthGlobalAddressStatically(
TPtr, DL, VT, (AArch64PACKey::ID)KeyC, Discriminator, AddrDiscriminator,
DAG);
}
// Looks through \param Val to determine the bit that can be used to
// check the sign of the value. It returns the unextended value and
// the sign bit position.
std::pair<SDValue, uint64_t> lookThroughSignExtension(SDValue Val) {
if (Val.getOpcode() == ISD::SIGN_EXTEND_INREG)
return {Val.getOperand(0),
cast<VTSDNode>(Val.getOperand(1))->getVT().getFixedSizeInBits() -
1};
if (Val.getOpcode() == ISD::SIGN_EXTEND)
return {Val.getOperand(0),
Val.getOperand(0)->getValueType(0).getFixedSizeInBits() - 1};
return {Val, Val.getValueSizeInBits() - 1};
}
SDValue AArch64TargetLowering::LowerBR_CC(SDValue Op, SelectionDAG &DAG) const {
SDValue Chain = Op.getOperand(0);
ISD::CondCode CC = cast<CondCodeSDNode>(Op.getOperand(1))->get();
SDValue LHS = Op.getOperand(2);
SDValue RHS = Op.getOperand(3);
SDValue Dest = Op.getOperand(4);
SDLoc DL(Op);
MachineFunction &MF = DAG.getMachineFunction();
// Speculation tracking/SLH assumes that optimized TB(N)Z/CB(N)Z instructions
// will not be produced, as they are conditional branch instructions that do
// not set flags.
bool ProduceNonFlagSettingCondBr =
!MF.getFunction().hasFnAttribute(Attribute::SpeculativeLoadHardening);
// Handle f128 first, since lowering it will result in comparing the return
// value of a libcall against zero, which is just what the rest of LowerBR_CC
// is expecting to deal with.
if (LHS.getValueType() == MVT::f128) {
softenSetCCOperands(DAG, MVT::f128, LHS, RHS, CC, DL, LHS, RHS);
// If softenSetCCOperands returned a scalar, we need to compare the result
// against zero to select between true and false values.
if (!RHS.getNode()) {
RHS = DAG.getConstant(0, DL, LHS.getValueType());
CC = ISD::SETNE;
}
}
// Optimize {s|u}{add|sub|mul}.with.overflow feeding into a branch
// instruction.
if (ISD::isOverflowIntrOpRes(LHS) && isOneConstant(RHS) &&
(CC == ISD::SETEQ || CC == ISD::SETNE)) {
// Only lower legal XALUO ops.
if (!DAG.getTargetLoweringInfo().isTypeLegal(LHS->getValueType(0)))
return SDValue();
// The actual operation with overflow check.
AArch64CC::CondCode OFCC;
SDValue Value, Overflow;
std::tie(Value, Overflow) = getAArch64XALUOOp(OFCC, LHS.getValue(0), DAG);
if (CC == ISD::SETNE)
OFCC = getInvertedCondCode(OFCC);
SDValue CCVal = DAG.getConstant(OFCC, DL, MVT::i32);
return DAG.getNode(AArch64ISD::BRCOND, DL, MVT::Other, Chain, Dest, CCVal,
Overflow);
}
if (LHS.getValueType().isInteger()) {
assert((LHS.getValueType() == RHS.getValueType()) &&
(LHS.getValueType() == MVT::i32 || LHS.getValueType() == MVT::i64));
// If the RHS of the comparison is zero, we can potentially fold this
// to a specialized branch.
const ConstantSDNode *RHSC = dyn_cast<ConstantSDNode>(RHS);
if (RHSC && RHSC->getZExtValue() == 0 && ProduceNonFlagSettingCondBr) {
if (CC == ISD::SETEQ) {
// See if we can use a TBZ to fold in an AND as well.
// TBZ has a smaller branch displacement than CBZ. If the offset is
// out of bounds, a late MI-layer pass rewrites branches.
// 403.gcc is an example that hits this case.
if (LHS.getOpcode() == ISD::AND &&
isa<ConstantSDNode>(LHS.getOperand(1)) &&
isPowerOf2_64(LHS.getConstantOperandVal(1))) {
SDValue Test = LHS.getOperand(0);
uint64_t Mask = LHS.getConstantOperandVal(1);
return DAG.getNode(AArch64ISD::TBZ, DL, MVT::Other, Chain, Test,
DAG.getConstant(Log2_64(Mask), DL, MVT::i64),
Dest);
}
return DAG.getNode(AArch64ISD::CBZ, DL, MVT::Other, Chain, LHS, Dest);
} else if (CC == ISD::SETNE) {
// See if we can use a TBZ to fold in an AND as well.
// TBZ has a smaller branch displacement than CBZ. If the offset is
// out of bounds, a late MI-layer pass rewrites branches.
// 403.gcc is an example that hits this case.
if (LHS.getOpcode() == ISD::AND &&
isa<ConstantSDNode>(LHS.getOperand(1)) &&
isPowerOf2_64(LHS.getConstantOperandVal(1))) {
SDValue Test = LHS.getOperand(0);
uint64_t Mask = LHS.getConstantOperandVal(1);
return DAG.getNode(AArch64ISD::TBNZ, DL, MVT::Other, Chain, Test,
DAG.getConstant(Log2_64(Mask), DL, MVT::i64),
Dest);
}
return DAG.getNode(AArch64ISD::CBNZ, DL, MVT::Other, Chain, LHS, Dest);
} else if (CC == ISD::SETLT && LHS.getOpcode() != ISD::AND) {
// Don't combine AND since emitComparison converts the AND to an ANDS
// (a.k.a. TST) and the test in the test bit and branch instruction
// becomes redundant. This would also increase register pressure.
uint64_t SignBitPos;
std::tie(LHS, SignBitPos) = lookThroughSignExtension(LHS);
return DAG.getNode(AArch64ISD::TBNZ, DL, MVT::Other, Chain, LHS,
DAG.getConstant(SignBitPos, DL, MVT::i64), Dest);
}
}
if (RHSC && RHSC->getSExtValue() == -1 && CC == ISD::SETGT &&
LHS.getOpcode() != ISD::AND && ProduceNonFlagSettingCondBr) {
// Don't combine AND since emitComparison converts the AND to an ANDS
// (a.k.a. TST) and the test in the test bit and branch instruction
// becomes redundant. This would also increase register pressure.
uint64_t SignBitPos;
std::tie(LHS, SignBitPos) = lookThroughSignExtension(LHS);
return DAG.getNode(AArch64ISD::TBZ, DL, MVT::Other, Chain, LHS,
DAG.getConstant(SignBitPos, DL, MVT::i64), Dest);
}
// Try to emit Armv9.6 CB instructions. We prefer tb{n}z/cb{n}z due to their
// larger branch displacement but do prefer CB over cmp + br.
if (Subtarget->hasCMPBR() &&
AArch64CC::isValidCBCond(changeIntCCToAArch64CC(CC)) &&
ProduceNonFlagSettingCondBr) {
SDValue Cond =
DAG.getTargetConstant(changeIntCCToAArch64CC(CC), DL, MVT::i32);
return DAG.getNode(AArch64ISD::CB, DL, MVT::Other, Chain, Cond, LHS, RHS,
Dest);
}
SDValue CCVal;
SDValue Cmp = getAArch64Cmp(LHS, RHS, CC, CCVal, DAG, DL);
return DAG.getNode(AArch64ISD::BRCOND, DL, MVT::Other, Chain, Dest, CCVal,
Cmp);
}
assert(LHS.getValueType() == MVT::f16 || LHS.getValueType() == MVT::bf16 ||
LHS.getValueType() == MVT::f32 || LHS.getValueType() == MVT::f64);
// Unfortunately, the mapping of LLVM FP CC's onto AArch64 CC's isn't totally
// clean. Some of them require two branches to implement.
SDValue Cmp = emitComparison(LHS, RHS, CC, DL, DAG);
AArch64CC::CondCode CC1, CC2;
changeFPCCToAArch64CC(CC, CC1, CC2);
SDValue CC1Val = DAG.getConstant(CC1, DL, MVT::i32);
SDValue BR1 =
DAG.getNode(AArch64ISD::BRCOND, DL, MVT::Other, Chain, Dest, CC1Val, Cmp);
if (CC2 != AArch64CC::AL) {
SDValue CC2Val = DAG.getConstant(CC2, DL, MVT::i32);
return DAG.getNode(AArch64ISD::BRCOND, DL, MVT::Other, BR1, Dest, CC2Val,
Cmp);
}
return BR1;
}
SDValue AArch64TargetLowering::LowerFCOPYSIGN(SDValue Op,
SelectionDAG &DAG) const {
if (!Subtarget->isNeonAvailable() &&
!Subtarget->useSVEForFixedLengthVectors())
return SDValue();
EVT VT = Op.getValueType();
EVT IntVT = VT.changeTypeToInteger();
SDLoc DL(Op);
SDValue In1 = Op.getOperand(0);
SDValue In2 = Op.getOperand(1);
EVT SrcVT = In2.getValueType();
if (!SrcVT.bitsEq(VT))
In2 = DAG.getFPExtendOrRound(In2, DL, VT);
if (VT.isScalableVector())
IntVT =
getPackedSVEVectorVT(VT.getVectorElementType().changeTypeToInteger());
if (VT.isFixedLengthVector() &&
useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable())) {
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
In1 = convertToScalableVector(DAG, ContainerVT, In1);
In2 = convertToScalableVector(DAG, ContainerVT, In2);
SDValue Res = DAG.getNode(ISD::FCOPYSIGN, DL, ContainerVT, In1, In2);
return convertFromScalableVector(DAG, VT, Res);
}
// With SVE, but without Neon, extend the scalars to scalable vectors and use
// a SVE FCOPYSIGN.
if (!VT.isVector() && !Subtarget->isNeonAvailable() &&
Subtarget->isSVEorStreamingSVEAvailable()) {
if (VT != MVT::f16 && VT != MVT::f32 && VT != MVT::f64 && VT != MVT::bf16)
return SDValue();
EVT SVT = getPackedSVEVectorVT(VT);
SDValue Ins1 =
DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, SVT, DAG.getUNDEF(SVT), In1,
DAG.getConstant(0, DL, MVT::i64));
SDValue Ins2 =
DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, SVT, DAG.getUNDEF(SVT), In2,
DAG.getConstant(0, DL, MVT::i64));
SDValue FCS = DAG.getNode(ISD::FCOPYSIGN, DL, SVT, Ins1, Ins2);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, FCS,
DAG.getConstant(0, DL, MVT::i64));
}
auto BitCast = [this](EVT VT, SDValue Op, SelectionDAG &DAG) {
if (VT.isScalableVector())
return getSVESafeBitCast(VT, Op, DAG);
return DAG.getBitcast(VT, Op);
};
SDValue VecVal1, VecVal2;
EVT VecVT;
auto SetVecVal = [&](int Idx = -1) {
if (!VT.isVector()) {
VecVal1 =
DAG.getTargetInsertSubreg(Idx, DL, VecVT, DAG.getUNDEF(VecVT), In1);
VecVal2 =
DAG.getTargetInsertSubreg(Idx, DL, VecVT, DAG.getUNDEF(VecVT), In2);
} else {
VecVal1 = BitCast(VecVT, In1, DAG);
VecVal2 = BitCast(VecVT, In2, DAG);
}
};
if (VT.isVector()) {
VecVT = IntVT;
SetVecVal();
} else if (VT == MVT::f64) {
VecVT = MVT::v2i64;
SetVecVal(AArch64::dsub);
} else if (VT == MVT::f32) {
VecVT = MVT::v4i32;
SetVecVal(AArch64::ssub);
} else if (VT == MVT::f16 || VT == MVT::bf16) {
VecVT = MVT::v8i16;
SetVecVal(AArch64::hsub);
} else {
llvm_unreachable("Invalid type for copysign!");
}
unsigned BitWidth = In1.getScalarValueSizeInBits();
SDValue SignMaskV = DAG.getConstant(~APInt::getSignMask(BitWidth), DL, VecVT);
// We want to materialize a mask with every bit but the high bit set, but the
// AdvSIMD immediate moves cannot materialize that in a single instruction for
// 64-bit elements. Instead, materialize all bits set and then negate that.
if (VT == MVT::f64 || VT == MVT::v2f64) {
SignMaskV = DAG.getConstant(APInt::getAllOnes(BitWidth), DL, VecVT);
SignMaskV = DAG.getNode(ISD::BITCAST, DL, MVT::v2f64, SignMaskV);
SignMaskV = DAG.getNode(ISD::FNEG, DL, MVT::v2f64, SignMaskV);
SignMaskV = DAG.getNode(ISD::BITCAST, DL, MVT::v2i64, SignMaskV);
}
SDValue BSP =
DAG.getNode(AArch64ISD::BSP, DL, VecVT, SignMaskV, VecVal1, VecVal2);
if (VT == MVT::f16 || VT == MVT::bf16)
return DAG.getTargetExtractSubreg(AArch64::hsub, DL, VT, BSP);
if (VT == MVT::f32)
return DAG.getTargetExtractSubreg(AArch64::ssub, DL, VT, BSP);
if (VT == MVT::f64)
return DAG.getTargetExtractSubreg(AArch64::dsub, DL, VT, BSP);
return BitCast(VT, BSP, DAG);
}
SDValue AArch64TargetLowering::LowerCTPOP_PARITY(SDValue Op,
SelectionDAG &DAG) const {
if (DAG.getMachineFunction().getFunction().hasFnAttribute(
Attribute::NoImplicitFloat))
return SDValue();
EVT VT = Op.getValueType();
if (VT.isScalableVector() ||
useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()))
return LowerToPredicatedOp(Op, DAG, AArch64ISD::CTPOP_MERGE_PASSTHRU);
if (!Subtarget->isNeonAvailable())
return SDValue();
bool IsParity = Op.getOpcode() == ISD::PARITY;
SDValue Val = Op.getOperand(0);
SDLoc DL(Op);
// for i32, general parity function using EORs is more efficient compared to
// using floating point
if (VT == MVT::i32 && IsParity)
return SDValue();
// If there is no CNT instruction available, GPR popcount can
// be more efficiently lowered to the following sequence that uses
// AdvSIMD registers/instructions as long as the copies to/from
// the AdvSIMD registers are cheap.
// FMOV D0, X0 // copy 64-bit int to vector, high bits zero'd
// CNT V0.8B, V0.8B // 8xbyte pop-counts
// ADDV B0, V0.8B // sum 8xbyte pop-counts
// FMOV X0, D0 // copy result back to integer reg
if (VT == MVT::i32 || VT == MVT::i64) {
if (VT == MVT::i32)
Val = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i64, Val);
Val = DAG.getNode(ISD::BITCAST, DL, MVT::v8i8, Val);
SDValue CtPop = DAG.getNode(ISD::CTPOP, DL, MVT::v8i8, Val);
SDValue AddV = DAG.getNode(AArch64ISD::UADDV, DL, MVT::v8i8, CtPop);
AddV = DAG.getNode(AArch64ISD::NVCAST, DL,
VT == MVT::i32 ? MVT::v2i32 : MVT::v1i64, AddV);
AddV = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, AddV,
DAG.getConstant(0, DL, MVT::i64));
if (IsParity)
AddV = DAG.getNode(ISD::AND, DL, VT, AddV, DAG.getConstant(1, DL, VT));
return AddV;
} else if (VT == MVT::i128) {
Val = DAG.getNode(ISD::BITCAST, DL, MVT::v16i8, Val);
SDValue CtPop = DAG.getNode(ISD::CTPOP, DL, MVT::v16i8, Val);
SDValue AddV = DAG.getNode(AArch64ISD::UADDV, DL, MVT::v16i8, CtPop);
AddV = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i64,
DAG.getNode(AArch64ISD::NVCAST, DL, MVT::v2i64, AddV),
DAG.getConstant(0, DL, MVT::i64));
AddV = DAG.getZExtOrTrunc(AddV, DL, VT);
if (IsParity)
AddV = DAG.getNode(ISD::AND, DL, VT, AddV, DAG.getConstant(1, DL, VT));
return AddV;
}
assert(!IsParity && "ISD::PARITY of vector types not supported");
assert((VT == MVT::v1i64 || VT == MVT::v2i64 || VT == MVT::v2i32 ||
VT == MVT::v4i32 || VT == MVT::v4i16 || VT == MVT::v8i16) &&
"Unexpected type for custom ctpop lowering");
EVT VT8Bit = VT.is64BitVector() ? MVT::v8i8 : MVT::v16i8;
Val = DAG.getBitcast(VT8Bit, Val);
Val = DAG.getNode(ISD::CTPOP, DL, VT8Bit, Val);
if (Subtarget->hasDotProd() && VT.getScalarSizeInBits() != 16 &&
VT.getVectorNumElements() >= 2) {
EVT DT = VT == MVT::v2i64 ? MVT::v4i32 : VT;
SDValue Zeros = DAG.getConstant(0, DL, DT);
SDValue Ones = DAG.getConstant(1, DL, VT8Bit);
if (VT == MVT::v2i64) {
Val = DAG.getNode(AArch64ISD::UDOT, DL, DT, Zeros, Ones, Val);
Val = DAG.getNode(AArch64ISD::UADDLP, DL, VT, Val);
} else if (VT == MVT::v2i32) {
Val = DAG.getNode(AArch64ISD::UDOT, DL, DT, Zeros, Ones, Val);
} else if (VT == MVT::v4i32) {
Val = DAG.getNode(AArch64ISD::UDOT, DL, DT, Zeros, Ones, Val);
} else {
llvm_unreachable("Unexpected type for custom ctpop lowering");
}
return Val;
}
// Widen v8i8/v16i8 CTPOP result to VT by repeatedly widening pairwise adds.
unsigned EltSize = 8;
unsigned NumElts = VT.is64BitVector() ? 8 : 16;
while (EltSize != VT.getScalarSizeInBits()) {
EltSize *= 2;
NumElts /= 2;
MVT WidenVT = MVT::getVectorVT(MVT::getIntegerVT(EltSize), NumElts);
Val = DAG.getNode(AArch64ISD::UADDLP, DL, WidenVT, Val);
}
return Val;
}
SDValue AArch64TargetLowering::LowerCTTZ(SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isScalableVector() ||
useSVEForFixedLengthVectorVT(
VT, /*OverrideNEON=*/Subtarget->useSVEForFixedLengthVectors()));
SDLoc DL(Op);
SDValue RBIT = DAG.getNode(ISD::BITREVERSE, DL, VT, Op.getOperand(0));
return DAG.getNode(ISD::CTLZ, DL, VT, RBIT);
}
SDValue AArch64TargetLowering::LowerMinMax(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
SDLoc DL(Op);
unsigned Opcode = Op.getOpcode();
ISD::CondCode CC;
switch (Opcode) {
default:
llvm_unreachable("Wrong instruction");
case ISD::SMAX:
CC = ISD::SETGT;
break;
case ISD::SMIN:
CC = ISD::SETLT;
break;
case ISD::UMAX:
CC = ISD::SETUGT;
break;
case ISD::UMIN:
CC = ISD::SETULT;
break;
}
if (VT.isScalableVector() ||
useSVEForFixedLengthVectorVT(
VT, /*OverrideNEON=*/Subtarget->useSVEForFixedLengthVectors())) {
switch (Opcode) {
default:
llvm_unreachable("Wrong instruction");
case ISD::SMAX:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::SMAX_PRED);
case ISD::SMIN:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::SMIN_PRED);
case ISD::UMAX:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::UMAX_PRED);
case ISD::UMIN:
return LowerToPredicatedOp(Op, DAG, AArch64ISD::UMIN_PRED);
}
}
SDValue Op0 = Op.getOperand(0);
SDValue Op1 = Op.getOperand(1);
SDValue Cond = DAG.getSetCC(DL, VT, Op0, Op1, CC);
return DAG.getSelect(DL, VT, Cond, Op0, Op1);
}
SDValue AArch64TargetLowering::LowerBitreverse(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
if (VT.isScalableVector() ||
useSVEForFixedLengthVectorVT(
VT, /*OverrideNEON=*/Subtarget->useSVEForFixedLengthVectors()))
return LowerToPredicatedOp(Op, DAG, AArch64ISD::BITREVERSE_MERGE_PASSTHRU);
SDLoc DL(Op);
SDValue REVB;
MVT VST;
switch (VT.getSimpleVT().SimpleTy) {
default:
llvm_unreachable("Invalid type for bitreverse!");
case MVT::v2i32: {
VST = MVT::v8i8;
REVB = DAG.getNode(AArch64ISD::REV32, DL, VST, Op.getOperand(0));
break;
}
case MVT::v4i32: {
VST = MVT::v16i8;
REVB = DAG.getNode(AArch64ISD::REV32, DL, VST, Op.getOperand(0));
break;
}
case MVT::v1i64: {
VST = MVT::v8i8;
REVB = DAG.getNode(AArch64ISD::REV64, DL, VST, Op.getOperand(0));
break;
}
case MVT::v2i64: {
VST = MVT::v16i8;
REVB = DAG.getNode(AArch64ISD::REV64, DL, VST, Op.getOperand(0));
break;
}
}
return DAG.getNode(AArch64ISD::NVCAST, DL, VT,
DAG.getNode(ISD::BITREVERSE, DL, VST, REVB));
}
// Check whether the continuous comparison sequence.
static bool
isOrXorChain(SDValue N, unsigned &Num,
SmallVector<std::pair<SDValue, SDValue>, 16> &WorkList) {
if (Num == MaxXors)
return false;
// Skip the one-use zext
if (N->getOpcode() == ISD::ZERO_EXTEND && N->hasOneUse())
N = N->getOperand(0);
// The leaf node must be XOR
if (N->getOpcode() == ISD::XOR) {
WorkList.push_back(std::make_pair(N->getOperand(0), N->getOperand(1)));
Num++;
return true;
}
// All the non-leaf nodes must be OR.
if (N->getOpcode() != ISD::OR || !N->hasOneUse())
return false;
if (isOrXorChain(N->getOperand(0), Num, WorkList) &&
isOrXorChain(N->getOperand(1), Num, WorkList))
return true;
return false;
}
// Transform chains of ORs and XORs, which usually outlined by memcmp/bmp.
static SDValue performOrXorChainCombine(SDNode *N, SelectionDAG &DAG) {
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
SDLoc DL(N);
EVT VT = N->getValueType(0);
SmallVector<std::pair<SDValue, SDValue>, 16> WorkList;
// Only handle integer compares.
if (N->getOpcode() != ISD::SETCC)
return SDValue();
ISD::CondCode Cond = cast<CondCodeSDNode>(N->getOperand(2))->get();
// Try to express conjunction "cmp 0 (or (xor A0 A1) (xor B0 B1))" as:
// sub A0, A1; ccmp B0, B1, 0, eq; cmp inv(Cond) flag
unsigned NumXors = 0;
if ((Cond == ISD::SETEQ || Cond == ISD::SETNE) && isNullConstant(RHS) &&
LHS->getOpcode() == ISD::OR && LHS->hasOneUse() &&
isOrXorChain(LHS, NumXors, WorkList)) {
SDValue XOR0, XOR1;
std::tie(XOR0, XOR1) = WorkList[0];
unsigned LogicOp = (Cond == ISD::SETEQ) ? ISD::AND : ISD::OR;
SDValue Cmp = DAG.getSetCC(DL, VT, XOR0, XOR1, Cond);
for (unsigned I = 1; I < WorkList.size(); I++) {
std::tie(XOR0, XOR1) = WorkList[I];
SDValue CmpChain = DAG.getSetCC(DL, VT, XOR0, XOR1, Cond);
Cmp = DAG.getNode(LogicOp, DL, VT, Cmp, CmpChain);
}
// Exit early by inverting the condition, which help reduce indentations.
return Cmp;
}
return SDValue();
}
SDValue AArch64TargetLowering::LowerSETCC(SDValue Op, SelectionDAG &DAG) const {
if (Op.getValueType().isVector())
return LowerVSETCC(Op, DAG);
bool IsStrict = Op->isStrictFPOpcode();
bool IsSignaling = Op.getOpcode() == ISD::STRICT_FSETCCS;
unsigned OpNo = IsStrict ? 1 : 0;
SDValue Chain;
if (IsStrict)
Chain = Op.getOperand(0);
SDValue LHS = Op.getOperand(OpNo + 0);
SDValue RHS = Op.getOperand(OpNo + 1);
ISD::CondCode CC = cast<CondCodeSDNode>(Op.getOperand(OpNo + 2))->get();
SDLoc DL(Op);
// We chose ZeroOrOneBooleanContents, so use zero and one.
EVT VT = Op.getValueType();
SDValue TVal = DAG.getConstant(1, DL, VT);
SDValue FVal = DAG.getConstant(0, DL, VT);
// Handle f128 first, since one possible outcome is a normal integer
// comparison which gets picked up by the next if statement.
if (LHS.getValueType() == MVT::f128) {
softenSetCCOperands(DAG, MVT::f128, LHS, RHS, CC, DL, LHS, RHS, Chain,
IsSignaling);
// If softenSetCCOperands returned a scalar, use it.
if (!RHS.getNode()) {
assert(LHS.getValueType() == Op.getValueType() &&
"Unexpected setcc expansion!");
return IsStrict ? DAG.getMergeValues({LHS, Chain}, DL) : LHS;
}
}
if (LHS.getValueType().isInteger()) {
simplifySetCCIntoEq(CC, LHS, RHS, DAG, DL);
SDValue CCVal;
SDValue Cmp = getAArch64Cmp(
LHS, RHS, ISD::getSetCCInverse(CC, LHS.getValueType()), CCVal, DAG, DL);
// Note that we inverted the condition above, so we reverse the order of
// the true and false operands here. This will allow the setcc to be
// matched to a single CSINC instruction.
SDValue Res = DAG.getNode(AArch64ISD::CSEL, DL, VT, FVal, TVal, CCVal, Cmp);
return IsStrict ? DAG.getMergeValues({Res, Chain}, DL) : Res;
}
// Now we know we're dealing with FP values.
assert(LHS.getValueType() == MVT::bf16 || LHS.getValueType() == MVT::f16 ||
LHS.getValueType() == MVT::f32 || LHS.getValueType() == MVT::f64);
// If that fails, we'll need to perform an FCMP + CSEL sequence. Go ahead
// and do the comparison.
SDValue Cmp;
if (IsStrict)
Cmp = emitStrictFPComparison(LHS, RHS, DL, DAG, Chain, IsSignaling);
else
Cmp = emitComparison(LHS, RHS, CC, DL, DAG);
AArch64CC::CondCode CC1, CC2;
changeFPCCToAArch64CC(CC, CC1, CC2);
SDValue Res;
if (CC2 == AArch64CC::AL) {
changeFPCCToAArch64CC(ISD::getSetCCInverse(CC, LHS.getValueType()), CC1,
CC2);
SDValue CC1Val = DAG.getConstant(CC1, DL, MVT::i32);
// Note that we inverted the condition above, so we reverse the order of
// the true and false operands here. This will allow the setcc to be
// matched to a single CSINC instruction.
Res = DAG.getNode(AArch64ISD::CSEL, DL, VT, FVal, TVal, CC1Val, Cmp);
} else {
// Unfortunately, the mapping of LLVM FP CC's onto AArch64 CC's isn't
// totally clean. Some of them require two CSELs to implement. As is in
// this case, we emit the first CSEL and then emit a second using the output
// of the first as the RHS. We're effectively OR'ing the two CC's together.
// FIXME: It would be nice if we could match the two CSELs to two CSINCs.
SDValue CC1Val = DAG.getConstant(CC1, DL, MVT::i32);
SDValue CS1 =
DAG.getNode(AArch64ISD::CSEL, DL, VT, TVal, FVal, CC1Val, Cmp);
SDValue CC2Val = DAG.getConstant(CC2, DL, MVT::i32);
Res = DAG.getNode(AArch64ISD::CSEL, DL, VT, TVal, CS1, CC2Val, Cmp);
}
return IsStrict ? DAG.getMergeValues({Res, Cmp.getValue(1)}, DL) : Res;
}
SDValue AArch64TargetLowering::LowerSETCCCARRY(SDValue Op,
SelectionDAG &DAG) const {
SDValue LHS = Op.getOperand(0);
SDValue RHS = Op.getOperand(1);
EVT VT = LHS.getValueType();
if (VT != MVT::i32 && VT != MVT::i64)
return SDValue();
SDLoc DL(Op);
SDValue Carry = Op.getOperand(2);
// SBCS uses a carry not a borrow so the carry flag should be inverted first.
SDValue InvCarry = valueToCarryFlag(Carry, DAG, true);
SDValue Cmp = DAG.getNode(AArch64ISD::SBCS, DL, DAG.getVTList(VT, MVT::Glue),
LHS, RHS, InvCarry);
EVT OpVT = Op.getValueType();
SDValue TVal = DAG.getConstant(1, DL, OpVT);
SDValue FVal = DAG.getConstant(0, DL, OpVT);
ISD::CondCode Cond = cast<CondCodeSDNode>(Op.getOperand(3))->get();
ISD::CondCode CondInv = ISD::getSetCCInverse(Cond, VT);
SDValue CCVal =
DAG.getConstant(changeIntCCToAArch64CC(CondInv), DL, MVT::i32);
// Inputs are swapped because the condition is inverted. This will allow
// matching with a single CSINC instruction.
return DAG.getNode(AArch64ISD::CSEL, DL, OpVT, FVal, TVal, CCVal,
Cmp.getValue(1));
}
/// Emit vector comparison for floating-point values, producing a mask.
static SDValue emitVectorComparison(SDValue LHS, SDValue RHS,
AArch64CC::CondCode CC, bool NoNans, EVT VT,
const SDLoc &DL, SelectionDAG &DAG) {
assert(VT.getSizeInBits() == LHS.getValueType().getSizeInBits() &&
"function only supposed to emit natural comparisons");
switch (CC) {
default:
return SDValue();
case AArch64CC::NE: {
SDValue Fcmeq = DAG.getNode(AArch64ISD::FCMEQ, DL, VT, LHS, RHS);
// Use vector semantics for the inversion to potentially save a copy between
// SIMD and regular registers.
if (!LHS.getValueType().isVector()) {
EVT VecVT =
EVT::getVectorVT(*DAG.getContext(), VT, 128 / VT.getSizeInBits());
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
SDValue MaskVec = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VecVT,
DAG.getUNDEF(VecVT), Fcmeq, Zero);
SDValue InvertedMask = DAG.getNOT(DL, MaskVec, VecVT);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, InvertedMask, Zero);
}
return DAG.getNOT(DL, Fcmeq, VT);
}
case AArch64CC::EQ:
return DAG.getNode(AArch64ISD::FCMEQ, DL, VT, LHS, RHS);
case AArch64CC::GE:
return DAG.getNode(AArch64ISD::FCMGE, DL, VT, LHS, RHS);
case AArch64CC::GT:
return DAG.getNode(AArch64ISD::FCMGT, DL, VT, LHS, RHS);
case AArch64CC::LE:
if (!NoNans)
return SDValue();
// If we ignore NaNs then we can use to the LS implementation.
[[fallthrough]];
case AArch64CC::LS:
return DAG.getNode(AArch64ISD::FCMGE, DL, VT, RHS, LHS);
case AArch64CC::LT:
if (!NoNans)
return SDValue();
// If we ignore NaNs then we can use to the MI implementation.
[[fallthrough]];
case AArch64CC::MI:
return DAG.getNode(AArch64ISD::FCMGT, DL, VT, RHS, LHS);
}
}
/// For SELECT_CC, when the true/false values are (-1, 0) and the compared
/// values are scalars, try to emit a mask generating vector instruction.
static SDValue emitFloatCompareMask(SDValue LHS, SDValue RHS, SDValue TVal,
SDValue FVal, ISD::CondCode CC, bool NoNaNs,
const SDLoc &DL, SelectionDAG &DAG) {
assert(!LHS.getValueType().isVector());
assert(!RHS.getValueType().isVector());
auto *CTVal = dyn_cast<ConstantSDNode>(TVal);
auto *CFVal = dyn_cast<ConstantSDNode>(FVal);
if (!CTVal || !CFVal)
return {};
if (!(CTVal->isAllOnes() && CFVal->isZero()) &&
!(CTVal->isZero() && CFVal->isAllOnes()))
return {};
if (CTVal->isZero())
CC = ISD::getSetCCInverse(CC, LHS.getValueType());
EVT VT = TVal.getValueType();
if (VT.getSizeInBits() != LHS.getValueType().getSizeInBits())
return {};
if (!NoNaNs && (CC == ISD::SETUO || CC == ISD::SETO)) {
bool OneNaN = false;
if (LHS == RHS) {
OneNaN = true;
} else if (DAG.isKnownNeverNaN(RHS)) {
OneNaN = true;
RHS = LHS;
} else if (DAG.isKnownNeverNaN(LHS)) {
OneNaN = true;
LHS = RHS;
}
if (OneNaN)
CC = (CC == ISD::SETUO) ? ISD::SETUNE : ISD::SETOEQ;
}
AArch64CC::CondCode CC1;
AArch64CC::CondCode CC2;
bool ShouldInvert = false;
changeVectorFPCCToAArch64CC(CC, CC1, CC2, ShouldInvert);
SDValue Cmp = emitVectorComparison(LHS, RHS, CC1, NoNaNs, VT, DL, DAG);
SDValue Cmp2;
if (CC2 != AArch64CC::AL) {
Cmp2 = emitVectorComparison(LHS, RHS, CC2, NoNaNs, VT, DL, DAG);
if (!Cmp2)
return {};
}
if (!Cmp2 && !ShouldInvert)
return Cmp;
EVT VecVT = EVT::getVectorVT(*DAG.getContext(), VT, 128 / VT.getSizeInBits());
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
Cmp = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VecVT, DAG.getUNDEF(VecVT), Cmp,
Zero);
if (Cmp2) {
Cmp2 = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VecVT, DAG.getUNDEF(VecVT),
Cmp2, Zero);
Cmp = DAG.getNode(ISD::OR, DL, VecVT, Cmp, Cmp2);
}
if (ShouldInvert)
Cmp = DAG.getNOT(DL, Cmp, VecVT);
Cmp = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, Cmp, Zero);
return Cmp;
}
SDValue AArch64TargetLowering::LowerSELECT_CC(
ISD::CondCode CC, SDValue LHS, SDValue RHS, SDValue TVal, SDValue FVal,
iterator_range<SDNode::user_iterator> Users, bool HasNoNaNs,
const SDLoc &DL, SelectionDAG &DAG) const {
// Handle f128 first, because it will result in a comparison of some RTLIB
// call result against zero.
if (LHS.getValueType() == MVT::f128) {
softenSetCCOperands(DAG, MVT::f128, LHS, RHS, CC, DL, LHS, RHS);
// If softenSetCCOperands returned a scalar, we need to compare the result
// against zero to select between true and false values.
if (!RHS.getNode()) {
RHS = DAG.getConstant(0, DL, LHS.getValueType());
CC = ISD::SETNE;
}
}
// Also handle f16, for which we need to do a f32 comparison.
if ((LHS.getValueType() == MVT::f16 && !Subtarget->hasFullFP16()) ||
LHS.getValueType() == MVT::bf16) {
LHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, LHS);
RHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, RHS);
}
// Next, handle integers.
if (LHS.getValueType().isInteger()) {
assert((LHS.getValueType() == RHS.getValueType()) &&
(LHS.getValueType() == MVT::i32 || LHS.getValueType() == MVT::i64));
ConstantSDNode *CFVal = dyn_cast<ConstantSDNode>(FVal);
ConstantSDNode *CTVal = dyn_cast<ConstantSDNode>(TVal);
ConstantSDNode *RHSC = dyn_cast<ConstantSDNode>(RHS);
// Check for sign pattern (SELECT_CC setgt, iN lhs, -1, 1, -1) and transform
// into (OR (ASR lhs, N-1), 1), which requires less instructions for the
// supported types.
if (CC == ISD::SETGT && RHSC && RHSC->isAllOnes() && CTVal && CFVal &&
CTVal->isOne() && CFVal->isAllOnes() &&
LHS.getValueType() == TVal.getValueType()) {
EVT VT = LHS.getValueType();
SDValue Shift =
DAG.getNode(ISD::SRA, DL, VT, LHS,
DAG.getConstant(VT.getSizeInBits() - 1, DL, VT));
return DAG.getNode(ISD::OR, DL, VT, Shift, DAG.getConstant(1, DL, VT));
}
// Check for SMAX(lhs, 0) and SMIN(lhs, 0) patterns.
// (SELECT_CC setgt, lhs, 0, lhs, 0) -> (BIC lhs, (SRA lhs, typesize-1))
// (SELECT_CC setlt, lhs, 0, lhs, 0) -> (AND lhs, (SRA lhs, typesize-1))
// Both require less instructions than compare and conditional select.
if ((CC == ISD::SETGT || CC == ISD::SETLT) && LHS == TVal &&
RHSC && RHSC->isZero() && CFVal && CFVal->isZero() &&
LHS.getValueType() == RHS.getValueType()) {
EVT VT = LHS.getValueType();
SDValue Shift =
DAG.getNode(ISD::SRA, DL, VT, LHS,
DAG.getConstant(VT.getSizeInBits() - 1, DL, VT));
if (CC == ISD::SETGT)
Shift = DAG.getNOT(DL, Shift, VT);
return DAG.getNode(ISD::AND, DL, VT, LHS, Shift);
}
unsigned Opcode = AArch64ISD::CSEL;
// If both the TVal and the FVal are constants, see if we can swap them in
// order to for a CSINV or CSINC out of them.
if (CTVal && CFVal && CTVal->isAllOnes() && CFVal->isZero()) {
std::swap(TVal, FVal);
std::swap(CTVal, CFVal);
CC = ISD::getSetCCInverse(CC, LHS.getValueType());
} else if (CTVal && CFVal && CTVal->isOne() && CFVal->isZero()) {
std::swap(TVal, FVal);
std::swap(CTVal, CFVal);
CC = ISD::getSetCCInverse(CC, LHS.getValueType());
} else if (TVal.getOpcode() == ISD::XOR) {
// If TVal is a NOT we want to swap TVal and FVal so that we can match
// with a CSINV rather than a CSEL.
if (isAllOnesConstant(TVal.getOperand(1))) {
std::swap(TVal, FVal);
std::swap(CTVal, CFVal);
CC = ISD::getSetCCInverse(CC, LHS.getValueType());
}
} else if (TVal.getOpcode() == ISD::SUB) {
// If TVal is a negation (SUB from 0) we want to swap TVal and FVal so
// that we can match with a CSNEG rather than a CSEL.
if (isNullConstant(TVal.getOperand(0))) {
std::swap(TVal, FVal);
std::swap(CTVal, CFVal);
CC = ISD::getSetCCInverse(CC, LHS.getValueType());
}
} else if (CTVal && CFVal) {
const int64_t TrueVal = CTVal->getSExtValue();
const int64_t FalseVal = CFVal->getSExtValue();
bool Swap = false;
// If both TVal and FVal are constants, see if FVal is the
// inverse/negation/increment of TVal and generate a CSINV/CSNEG/CSINC
// instead of a CSEL in that case.
if (TrueVal == ~FalseVal) {
Opcode = AArch64ISD::CSINV;
} else if (FalseVal > std::numeric_limits<int64_t>::min() &&
TrueVal == -FalseVal) {
Opcode = AArch64ISD::CSNEG;
} else if (TVal.getValueType() == MVT::i32) {
// If our operands are only 32-bit wide, make sure we use 32-bit
// arithmetic for the check whether we can use CSINC. This ensures that
// the addition in the check will wrap around properly in case there is
// an overflow (which would not be the case if we do the check with
// 64-bit arithmetic).
const uint32_t TrueVal32 = CTVal->getZExtValue();
const uint32_t FalseVal32 = CFVal->getZExtValue();
if ((TrueVal32 == FalseVal32 + 1) || (TrueVal32 + 1 == FalseVal32)) {
Opcode = AArch64ISD::CSINC;
if (TrueVal32 > FalseVal32) {
Swap = true;
}
}
} else {
// 64-bit check whether we can use CSINC.
const uint64_t TrueVal64 = TrueVal;
const uint64_t FalseVal64 = FalseVal;
if ((TrueVal64 == FalseVal64 + 1) || (TrueVal64 + 1 == FalseVal64)) {
Opcode = AArch64ISD::CSINC;
if (TrueVal > FalseVal) {
Swap = true;
}
}
}
// Swap TVal and FVal if necessary.
if (Swap) {
std::swap(TVal, FVal);
std::swap(CTVal, CFVal);
CC = ISD::getSetCCInverse(CC, LHS.getValueType());
}
if (Opcode != AArch64ISD::CSEL) {
// Drop FVal since we can get its value by simply inverting/negating
// TVal.
FVal = TVal;
}
}
// Avoid materializing a constant when possible by reusing a known value in
// a register. However, don't perform this optimization if the known value
// is one, zero or negative one in the case of a CSEL. We can always
// materialize these values using CSINC, CSEL and CSINV with wzr/xzr as the
// FVal, respectively.
ConstantSDNode *RHSVal = dyn_cast<ConstantSDNode>(RHS);
if (Opcode == AArch64ISD::CSEL && RHSVal && !RHSVal->isOne() &&
!RHSVal->isZero() && !RHSVal->isAllOnes()) {
AArch64CC::CondCode AArch64CC = changeIntCCToAArch64CC(CC);
// Transform "a == C ? C : x" to "a == C ? a : x" and "a != C ? x : C" to
// "a != C ? x : a" to avoid materializing C.
if (CTVal && CTVal == RHSVal && AArch64CC == AArch64CC::EQ)
TVal = LHS;
else if (CFVal && CFVal == RHSVal && AArch64CC == AArch64CC::NE)
FVal = LHS;
} else if (Opcode == AArch64ISD::CSNEG && RHSVal && RHSVal->isOne()) {
assert (CTVal && CFVal && "Expected constant operands for CSNEG.");
// Use a CSINV to transform "a == C ? 1 : -1" to "a == C ? a : -1" to
// avoid materializing C.
AArch64CC::CondCode AArch64CC = changeIntCCToAArch64CC(CC);
if (CTVal == RHSVal && AArch64CC == AArch64CC::EQ) {
Opcode = AArch64ISD::CSINV;
TVal = LHS;
FVal = DAG.getConstant(0, DL, FVal.getValueType());
}
}
SDValue CCVal;
SDValue Cmp = getAArch64Cmp(LHS, RHS, CC, CCVal, DAG, DL);
EVT VT = TVal.getValueType();
return DAG.getNode(Opcode, DL, VT, TVal, FVal, CCVal, Cmp);
}
// Now we know we're dealing with FP values.
assert(LHS.getValueType() == MVT::f16 || LHS.getValueType() == MVT::f32 ||
LHS.getValueType() == MVT::f64);
assert(LHS.getValueType() == RHS.getValueType());
EVT VT = TVal.getValueType();
// If the purpose of the comparison is to select between all ones
// or all zeros, try to use a vector comparison because the operands are
// already stored in SIMD registers.
if (Subtarget->isNeonAvailable() && all_of(Users, [](const SDNode *U) {
switch (U->getOpcode()) {
default:
return false;
case ISD::INSERT_VECTOR_ELT:
case ISD::SCALAR_TO_VECTOR:
case AArch64ISD::DUP:
return true;
}
})) {
bool NoNaNs = getTargetMachine().Options.NoNaNsFPMath || HasNoNaNs;
SDValue VectorCmp =
emitFloatCompareMask(LHS, RHS, TVal, FVal, CC, NoNaNs, DL, DAG);
if (VectorCmp)
return VectorCmp;
}
SDValue Cmp = emitComparison(LHS, RHS, CC, DL, DAG);
// Unfortunately, the mapping of LLVM FP CC's onto AArch64 CC's isn't totally
// clean. Some of them require two CSELs to implement.
AArch64CC::CondCode CC1, CC2;
changeFPCCToAArch64CC(CC, CC1, CC2);
if (DAG.getTarget().Options.UnsafeFPMath) {
// Transform "a == 0.0 ? 0.0 : x" to "a == 0.0 ? a : x" and
// "a != 0.0 ? x : 0.0" to "a != 0.0 ? x : a" to avoid materializing 0.0.
ConstantFPSDNode *RHSVal = dyn_cast<ConstantFPSDNode>(RHS);
if (RHSVal && RHSVal->isZero()) {
ConstantFPSDNode *CFVal = dyn_cast<ConstantFPSDNode>(FVal);
ConstantFPSDNode *CTVal = dyn_cast<ConstantFPSDNode>(TVal);
if ((CC == ISD::SETEQ || CC == ISD::SETOEQ || CC == ISD::SETUEQ) &&
CTVal && CTVal->isZero() && TVal.getValueType() == LHS.getValueType())
TVal = LHS;
else if ((CC == ISD::SETNE || CC == ISD::SETONE || CC == ISD::SETUNE) &&
CFVal && CFVal->isZero() &&
FVal.getValueType() == LHS.getValueType())
FVal = LHS;
}
}
// Emit first, and possibly only, CSEL.
SDValue CC1Val = DAG.getConstant(CC1, DL, MVT::i32);
SDValue CS1 = DAG.getNode(AArch64ISD::CSEL, DL, VT, TVal, FVal, CC1Val, Cmp);
// If we need a second CSEL, emit it, using the output of the first as the
// RHS. We're effectively OR'ing the two CC's together.
if (CC2 != AArch64CC::AL) {
SDValue CC2Val = DAG.getConstant(CC2, DL, MVT::i32);
return DAG.getNode(AArch64ISD::CSEL, DL, VT, TVal, CS1, CC2Val, Cmp);
}
// Otherwise, return the output of the first CSEL.
return CS1;
}
SDValue AArch64TargetLowering::LowerVECTOR_SPLICE(SDValue Op,
SelectionDAG &DAG) const {
EVT Ty = Op.getValueType();
auto Idx = Op.getConstantOperandAPInt(2);
int64_t IdxVal = Idx.getSExtValue();
assert(Ty.isScalableVector() &&
"Only expect scalable vectors for custom lowering of VECTOR_SPLICE");
// We can use the splice instruction for certain index values where we are
// able to efficiently generate the correct predicate. The index will be
// inverted and used directly as the input to the ptrue instruction, i.e.
// -1 -> vl1, -2 -> vl2, etc. The predicate will then be reversed to get the
// splice predicate. However, we can only do this if we can guarantee that
// there are enough elements in the vector, hence we check the index <= min
// number of elements.
std::optional<unsigned> PredPattern;
if (Ty.isScalableVector() && IdxVal < 0 &&
(PredPattern = getSVEPredPatternFromNumElements(std::abs(IdxVal))) !=
std::nullopt) {
SDLoc DL(Op);
// Create a predicate where all but the last -IdxVal elements are false.
EVT PredVT = Ty.changeVectorElementType(MVT::i1);
SDValue Pred = getPTrue(DAG, DL, PredVT, *PredPattern);
Pred = DAG.getNode(ISD::VECTOR_REVERSE, DL, PredVT, Pred);
// Now splice the two inputs together using the predicate.
return DAG.getNode(AArch64ISD::SPLICE, DL, Ty, Pred, Op.getOperand(0),
Op.getOperand(1));
}
// We can select to an EXT instruction when indexing the first 256 bytes.
unsigned BlockSize = AArch64::SVEBitsPerBlock / Ty.getVectorMinNumElements();
if (IdxVal >= 0 && (IdxVal * BlockSize / 8) < 256)
return Op;
return SDValue();
}
SDValue AArch64TargetLowering::LowerSELECT_CC(SDValue Op,
SelectionDAG &DAG) const {
ISD::CondCode CC = cast<CondCodeSDNode>(Op.getOperand(4))->get();
SDValue LHS = Op.getOperand(0);
SDValue RHS = Op.getOperand(1);
SDValue TVal = Op.getOperand(2);
SDValue FVal = Op.getOperand(3);
bool HasNoNans = Op->getFlags().hasNoNaNs();
SDLoc DL(Op);
return LowerSELECT_CC(CC, LHS, RHS, TVal, FVal, Op->users(), HasNoNans, DL,
DAG);
}
SDValue AArch64TargetLowering::LowerSELECT(SDValue Op,
SelectionDAG &DAG) const {
SDValue CCVal = Op->getOperand(0);
SDValue TVal = Op->getOperand(1);
SDValue FVal = Op->getOperand(2);
bool HasNoNans = Op->getFlags().hasNoNaNs();
SDLoc DL(Op);
EVT Ty = Op.getValueType();
if (Ty == MVT::aarch64svcount) {
TVal = DAG.getNode(ISD::BITCAST, DL, MVT::nxv16i1, TVal);
FVal = DAG.getNode(ISD::BITCAST, DL, MVT::nxv16i1, FVal);
SDValue Sel =
DAG.getNode(ISD::SELECT, DL, MVT::nxv16i1, CCVal, TVal, FVal);
return DAG.getNode(ISD::BITCAST, DL, Ty, Sel);
}
if (Ty.isScalableVector()) {
MVT PredVT = MVT::getVectorVT(MVT::i1, Ty.getVectorElementCount());
SDValue SplatPred = DAG.getNode(ISD::SPLAT_VECTOR, DL, PredVT, CCVal);
return DAG.getNode(ISD::VSELECT, DL, Ty, SplatPred, TVal, FVal);
}
if (useSVEForFixedLengthVectorVT(Ty, !Subtarget->isNeonAvailable())) {
// FIXME: Ideally this would be the same as above using i1 types, however
// for the moment we can't deal with fixed i1 vector types properly, so
// instead extend the predicate to a result type sized integer vector.
MVT SplatValVT = MVT::getIntegerVT(Ty.getScalarSizeInBits());
MVT PredVT = MVT::getVectorVT(SplatValVT, Ty.getVectorElementCount());
SDValue SplatVal = DAG.getSExtOrTrunc(CCVal, DL, SplatValVT);
SDValue SplatPred = DAG.getNode(ISD::SPLAT_VECTOR, DL, PredVT, SplatVal);
return DAG.getNode(ISD::VSELECT, DL, Ty, SplatPred, TVal, FVal);
}
// Optimize {s|u}{add|sub|mul}.with.overflow feeding into a select
// instruction.
if (ISD::isOverflowIntrOpRes(CCVal)) {
// Only lower legal XALUO ops.
if (!DAG.getTargetLoweringInfo().isTypeLegal(CCVal->getValueType(0)))
return SDValue();
AArch64CC::CondCode OFCC;
SDValue Value, Overflow;
std::tie(Value, Overflow) = getAArch64XALUOOp(OFCC, CCVal.getValue(0), DAG);
SDValue CCVal = DAG.getConstant(OFCC, DL, MVT::i32);
return DAG.getNode(AArch64ISD::CSEL, DL, Op.getValueType(), TVal, FVal,
CCVal, Overflow);
}
// Lower it the same way as we would lower a SELECT_CC node.
ISD::CondCode CC;
SDValue LHS, RHS;
if (CCVal.getOpcode() == ISD::SETCC) {
LHS = CCVal.getOperand(0);
RHS = CCVal.getOperand(1);
CC = cast<CondCodeSDNode>(CCVal.getOperand(2))->get();
} else {
LHS = CCVal;
RHS = DAG.getConstant(0, DL, CCVal.getValueType());
CC = ISD::SETNE;
}
// If we are lowering a f16 and we do not have fullf16, convert to a f32 in
// order to use FCSELSrrr
if ((Ty == MVT::f16 || Ty == MVT::bf16) && !Subtarget->hasFullFP16()) {
TVal = DAG.getTargetInsertSubreg(AArch64::hsub, DL, MVT::f32,
DAG.getUNDEF(MVT::f32), TVal);
FVal = DAG.getTargetInsertSubreg(AArch64::hsub, DL, MVT::f32,
DAG.getUNDEF(MVT::f32), FVal);
}
SDValue Res =
LowerSELECT_CC(CC, LHS, RHS, TVal, FVal, Op->users(), HasNoNans, DL, DAG);
if ((Ty == MVT::f16 || Ty == MVT::bf16) && !Subtarget->hasFullFP16()) {
return DAG.getTargetExtractSubreg(AArch64::hsub, DL, Ty, Res);
}
return Res;
}
SDValue AArch64TargetLowering::LowerJumpTable(SDValue Op,
SelectionDAG &DAG) const {
// Jump table entries as PC relative offsets. No additional tweaking
// is necessary here. Just get the address of the jump table.
JumpTableSDNode *JT = cast<JumpTableSDNode>(Op);
CodeModel::Model CM = getTargetMachine().getCodeModel();
if (CM == CodeModel::Large && !getTargetMachine().isPositionIndependent() &&
!Subtarget->isTargetMachO())
return getAddrLarge(JT, DAG);
if (CM == CodeModel::Tiny)
return getAddrTiny(JT, DAG);
return getAddr(JT, DAG);
}
SDValue AArch64TargetLowering::LowerBR_JT(SDValue Op,
SelectionDAG &DAG) const {
// Jump table entries as PC relative offsets. No additional tweaking
// is necessary here. Just get the address of the jump table.
SDLoc DL(Op);
SDValue JT = Op.getOperand(1);
SDValue Entry = Op.getOperand(2);
int JTI = cast<JumpTableSDNode>(JT.getNode())->getIndex();
auto *AFI = DAG.getMachineFunction().getInfo<AArch64FunctionInfo>();
AFI->setJumpTableEntryInfo(JTI, 4, nullptr);
// With aarch64-jump-table-hardening, we only expand the jump table dispatch
// sequence later, to guarantee the integrity of the intermediate values.
if (DAG.getMachineFunction().getFunction().hasFnAttribute(
"aarch64-jump-table-hardening")) {
CodeModel::Model CM = getTargetMachine().getCodeModel();
if (Subtarget->isTargetMachO()) {
if (CM != CodeModel::Small && CM != CodeModel::Large)
report_fatal_error("Unsupported code-model for hardened jump-table");
} else {
// Note that COFF support would likely also need JUMP_TABLE_DEBUG_INFO.
assert(Subtarget->isTargetELF() &&
"jump table hardening only supported on MachO/ELF");
if (CM != CodeModel::Small)
report_fatal_error("Unsupported code-model for hardened jump-table");
}
SDValue X16Copy = DAG.getCopyToReg(DAG.getEntryNode(), DL, AArch64::X16,
Entry, SDValue());
SDNode *B = DAG.getMachineNode(AArch64::BR_JumpTable, DL, MVT::Other,
DAG.getTargetJumpTable(JTI, MVT::i32),
X16Copy.getValue(0), X16Copy.getValue(1));
return SDValue(B, 0);
}
SDNode *Dest =
DAG.getMachineNode(AArch64::JumpTableDest32, DL, MVT::i64, MVT::i64, JT,
Entry, DAG.getTargetJumpTable(JTI, MVT::i32));
SDValue JTInfo = DAG.getJumpTableDebugInfo(JTI, Op.getOperand(0), DL);
return DAG.getNode(ISD::BRIND, DL, MVT::Other, JTInfo, SDValue(Dest, 0));
}
SDValue AArch64TargetLowering::LowerBRIND(SDValue Op, SelectionDAG &DAG) const {
SDValue Chain = Op.getOperand(0);
SDValue Dest = Op.getOperand(1);
// BR_JT is lowered to BRIND, but the later lowering is specific to indirectbr
// Skip over the jump-table BRINDs, where the destination is JumpTableDest32.
if (Dest->isMachineOpcode() &&
Dest->getMachineOpcode() == AArch64::JumpTableDest32)
return SDValue();
const MachineFunction &MF = DAG.getMachineFunction();
std::optional<uint16_t> BADisc =
Subtarget->getPtrAuthBlockAddressDiscriminatorIfEnabled(MF.getFunction());
if (!BADisc)
return SDValue();
SDLoc DL(Op);
SDValue Disc = DAG.getTargetConstant(*BADisc, DL, MVT::i64);
SDValue Key = DAG.getTargetConstant(AArch64PACKey::IA, DL, MVT::i32);
SDValue AddrDisc = DAG.getRegister(AArch64::XZR, MVT::i64);
SDNode *BrA = DAG.getMachineNode(AArch64::BRA, DL, MVT::Other,
{Dest, Key, Disc, AddrDisc, Chain});
return SDValue(BrA, 0);
}
SDValue AArch64TargetLowering::LowerConstantPool(SDValue Op,
SelectionDAG &DAG) const {
ConstantPoolSDNode *CP = cast<ConstantPoolSDNode>(Op);
CodeModel::Model CM = getTargetMachine().getCodeModel();
if (CM == CodeModel::Large) {
// Use the GOT for the large code model on iOS.
if (Subtarget->isTargetMachO()) {
return getGOT(CP, DAG);
}
if (!getTargetMachine().isPositionIndependent())
return getAddrLarge(CP, DAG);
} else if (CM == CodeModel::Tiny) {
return getAddrTiny(CP, DAG);
}
return getAddr(CP, DAG);
}
SDValue AArch64TargetLowering::LowerBlockAddress(SDValue Op,
SelectionDAG &DAG) const {
BlockAddressSDNode *BAN = cast<BlockAddressSDNode>(Op);
const BlockAddress *BA = BAN->getBlockAddress();
if (std::optional<uint16_t> BADisc =
Subtarget->getPtrAuthBlockAddressDiscriminatorIfEnabled(
*BA->getFunction())) {
SDLoc DL(Op);
// This isn't cheap, but BRIND is rare.
SDValue TargetBA = DAG.getTargetBlockAddress(BA, BAN->getValueType(0));
SDValue Disc = DAG.getTargetConstant(*BADisc, DL, MVT::i64);
SDValue Key = DAG.getTargetConstant(AArch64PACKey::IA, DL, MVT::i32);
SDValue AddrDisc = DAG.getRegister(AArch64::XZR, MVT::i64);
SDNode *MOV =
DAG.getMachineNode(AArch64::MOVaddrPAC, DL, {MVT::Other, MVT::Glue},
{TargetBA, Key, AddrDisc, Disc});
return DAG.getCopyFromReg(SDValue(MOV, 0), DL, AArch64::X16, MVT::i64,
SDValue(MOV, 1));
}
CodeModel::Model CM = getTargetMachine().getCodeModel();
if (CM == CodeModel::Large && !Subtarget->isTargetMachO()) {
if (!getTargetMachine().isPositionIndependent())
return getAddrLarge(BAN, DAG);
} else if (CM == CodeModel::Tiny) {
return getAddrTiny(BAN, DAG);
}
return getAddr(BAN, DAG);
}
SDValue AArch64TargetLowering::LowerDarwin_VASTART(SDValue Op,
SelectionDAG &DAG) const {
AArch64FunctionInfo *FuncInfo =
DAG.getMachineFunction().getInfo<AArch64FunctionInfo>();
SDLoc DL(Op);
SDValue FR = DAG.getFrameIndex(FuncInfo->getVarArgsStackIndex(),
getPointerTy(DAG.getDataLayout()));
FR = DAG.getZExtOrTrunc(FR, DL, getPointerMemTy(DAG.getDataLayout()));
const Value *SV = cast<SrcValueSDNode>(Op.getOperand(2))->getValue();
return DAG.getStore(Op.getOperand(0), DL, FR, Op.getOperand(1),
MachinePointerInfo(SV));
}
SDValue AArch64TargetLowering::LowerWin64_VASTART(SDValue Op,
SelectionDAG &DAG) const {
MachineFunction &MF = DAG.getMachineFunction();
AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
SDLoc DL(Op);
SDValue FR;
if (Subtarget->isWindowsArm64EC()) {
// With the Arm64EC ABI, we compute the address of the varargs save area
// relative to x4. For a normal AArch64->AArch64 call, x4 == sp on entry,
// but calls from an entry thunk can pass in a different address.
Register VReg = MF.addLiveIn(AArch64::X4, &AArch64::GPR64RegClass);
SDValue Val = DAG.getCopyFromReg(DAG.getEntryNode(), DL, VReg, MVT::i64);
uint64_t StackOffset;
if (FuncInfo->getVarArgsGPRSize() > 0)
StackOffset = -(uint64_t)FuncInfo->getVarArgsGPRSize();
else
StackOffset = FuncInfo->getVarArgsStackOffset();
FR = DAG.getNode(ISD::ADD, DL, MVT::i64, Val,
DAG.getConstant(StackOffset, DL, MVT::i64));
} else {
FR = DAG.getFrameIndex(FuncInfo->getVarArgsGPRSize() > 0
? FuncInfo->getVarArgsGPRIndex()
: FuncInfo->getVarArgsStackIndex(),
getPointerTy(DAG.getDataLayout()));
}
const Value *SV = cast<SrcValueSDNode>(Op.getOperand(2))->getValue();
return DAG.getStore(Op.getOperand(0), DL, FR, Op.getOperand(1),
MachinePointerInfo(SV));
}
SDValue AArch64TargetLowering::LowerAAPCS_VASTART(SDValue Op,
SelectionDAG &DAG) const {
// The layout of the va_list struct is specified in the AArch64 Procedure Call
// Standard, section B.3.
MachineFunction &MF = DAG.getMachineFunction();
AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
unsigned PtrSize = Subtarget->isTargetILP32() ? 4 : 8;
auto PtrMemVT = getPointerMemTy(DAG.getDataLayout());
auto PtrVT = getPointerTy(DAG.getDataLayout());
SDLoc DL(Op);
SDValue Chain = Op.getOperand(0);
SDValue VAList = Op.getOperand(1);
const Value *SV = cast<SrcValueSDNode>(Op.getOperand(2))->getValue();
SmallVector<SDValue, 4> MemOps;
// void *__stack at offset 0
unsigned Offset = 0;
SDValue Stack = DAG.getFrameIndex(FuncInfo->getVarArgsStackIndex(), PtrVT);
Stack = DAG.getZExtOrTrunc(Stack, DL, PtrMemVT);
MemOps.push_back(DAG.getStore(Chain, DL, Stack, VAList,
MachinePointerInfo(SV), Align(PtrSize)));
// void *__gr_top at offset 8 (4 on ILP32)
Offset += PtrSize;
int GPRSize = FuncInfo->getVarArgsGPRSize();
if (GPRSize > 0) {
SDValue GRTop, GRTopAddr;
GRTopAddr = DAG.getNode(ISD::ADD, DL, PtrVT, VAList,
DAG.getConstant(Offset, DL, PtrVT));
GRTop = DAG.getFrameIndex(FuncInfo->getVarArgsGPRIndex(), PtrVT);
GRTop = DAG.getNode(ISD::ADD, DL, PtrVT, GRTop,
DAG.getSignedConstant(GPRSize, DL, PtrVT));
GRTop = DAG.getZExtOrTrunc(GRTop, DL, PtrMemVT);
MemOps.push_back(DAG.getStore(Chain, DL, GRTop, GRTopAddr,
MachinePointerInfo(SV, Offset),
Align(PtrSize)));
}
// void *__vr_top at offset 16 (8 on ILP32)
Offset += PtrSize;
int FPRSize = FuncInfo->getVarArgsFPRSize();
if (FPRSize > 0) {
SDValue VRTop, VRTopAddr;
VRTopAddr = DAG.getNode(ISD::ADD, DL, PtrVT, VAList,
DAG.getConstant(Offset, DL, PtrVT));
VRTop = DAG.getFrameIndex(FuncInfo->getVarArgsFPRIndex(), PtrVT);
VRTop = DAG.getNode(ISD::ADD, DL, PtrVT, VRTop,
DAG.getSignedConstant(FPRSize, DL, PtrVT));
VRTop = DAG.getZExtOrTrunc(VRTop, DL, PtrMemVT);
MemOps.push_back(DAG.getStore(Chain, DL, VRTop, VRTopAddr,
MachinePointerInfo(SV, Offset),
Align(PtrSize)));
}
// int __gr_offs at offset 24 (12 on ILP32)
Offset += PtrSize;
SDValue GROffsAddr = DAG.getNode(ISD::ADD, DL, PtrVT, VAList,
DAG.getConstant(Offset, DL, PtrVT));
MemOps.push_back(
DAG.getStore(Chain, DL, DAG.getSignedConstant(-GPRSize, DL, MVT::i32),
GROffsAddr, MachinePointerInfo(SV, Offset), Align(4)));
// int __vr_offs at offset 28 (16 on ILP32)
Offset += 4;
SDValue VROffsAddr = DAG.getNode(ISD::ADD, DL, PtrVT, VAList,
DAG.getConstant(Offset, DL, PtrVT));
MemOps.push_back(
DAG.getStore(Chain, DL, DAG.getSignedConstant(-FPRSize, DL, MVT::i32),
VROffsAddr, MachinePointerInfo(SV, Offset), Align(4)));
return DAG.getNode(ISD::TokenFactor, DL, MVT::Other, MemOps);
}
SDValue AArch64TargetLowering::LowerVASTART(SDValue Op,
SelectionDAG &DAG) const {
MachineFunction &MF = DAG.getMachineFunction();
Function &F = MF.getFunction();
if (Subtarget->isCallingConvWin64(F.getCallingConv(), F.isVarArg()))
return LowerWin64_VASTART(Op, DAG);
else if (Subtarget->isTargetDarwin())
return LowerDarwin_VASTART(Op, DAG);
else
return LowerAAPCS_VASTART(Op, DAG);
}
SDValue AArch64TargetLowering::LowerVACOPY(SDValue Op,
SelectionDAG &DAG) const {
// AAPCS has three pointers and two ints (= 32 bytes), Darwin has single
// pointer.
SDLoc DL(Op);
unsigned PtrSize = Subtarget->isTargetILP32() ? 4 : 8;
unsigned VaListSize =
(Subtarget->isTargetDarwin() || Subtarget->isTargetWindows())
? PtrSize
: Subtarget->isTargetILP32() ? 20 : 32;
const Value *DestSV = cast<SrcValueSDNode>(Op.getOperand(3))->getValue();
const Value *SrcSV = cast<SrcValueSDNode>(Op.getOperand(4))->getValue();
return DAG.getMemcpy(Op.getOperand(0), DL, Op.getOperand(1), Op.getOperand(2),
DAG.getConstant(VaListSize, DL, MVT::i32),
Align(PtrSize), false, false, /*CI=*/nullptr,
std::nullopt, MachinePointerInfo(DestSV),
MachinePointerInfo(SrcSV));
}
SDValue AArch64TargetLowering::LowerVAARG(SDValue Op, SelectionDAG &DAG) const {
assert(Subtarget->isTargetDarwin() &&
"automatic va_arg instruction only works on Darwin");
const Value *V = cast<SrcValueSDNode>(Op.getOperand(2))->getValue();
EVT VT = Op.getValueType();
SDLoc DL(Op);
SDValue Chain = Op.getOperand(0);
SDValue Addr = Op.getOperand(1);
MaybeAlign Align(Op.getConstantOperandVal(3));
unsigned MinSlotSize = Subtarget->isTargetILP32() ? 4 : 8;
auto PtrVT = getPointerTy(DAG.getDataLayout());
auto PtrMemVT = getPointerMemTy(DAG.getDataLayout());
SDValue VAList =
DAG.getLoad(PtrMemVT, DL, Chain, Addr, MachinePointerInfo(V));
Chain = VAList.getValue(1);
VAList = DAG.getZExtOrTrunc(VAList, DL, PtrVT);
if (VT.isScalableVector())
report_fatal_error("Passing SVE types to variadic functions is "
"currently not supported");
if (Align && *Align > MinSlotSize) {
VAList = DAG.getNode(ISD::ADD, DL, PtrVT, VAList,
DAG.getConstant(Align->value() - 1, DL, PtrVT));
VAList =
DAG.getNode(ISD::AND, DL, PtrVT, VAList,
DAG.getSignedConstant(-(int64_t)Align->value(), DL, PtrVT));
}
Type *ArgTy = VT.getTypeForEVT(*DAG.getContext());
unsigned ArgSize = DAG.getDataLayout().getTypeAllocSize(ArgTy);
// Scalar integer and FP values smaller than 64 bits are implicitly extended
// up to 64 bits. At the very least, we have to increase the striding of the
// vaargs list to match this, and for FP values we need to introduce
// FP_ROUND nodes as well.
if (VT.isInteger() && !VT.isVector())
ArgSize = std::max(ArgSize, MinSlotSize);
bool NeedFPTrunc = false;
if (VT.isFloatingPoint() && !VT.isVector() && VT != MVT::f64) {
ArgSize = 8;
NeedFPTrunc = true;
}
// Increment the pointer, VAList, to the next vaarg
SDValue VANext = DAG.getNode(ISD::ADD, DL, PtrVT, VAList,
DAG.getConstant(ArgSize, DL, PtrVT));
VANext = DAG.getZExtOrTrunc(VANext, DL, PtrMemVT);
// Store the incremented VAList to the legalized pointer
SDValue APStore =
DAG.getStore(Chain, DL, VANext, Addr, MachinePointerInfo(V));
// Load the actual argument out of the pointer VAList
if (NeedFPTrunc) {
// Load the value as an f64.
SDValue WideFP =
DAG.getLoad(MVT::f64, DL, APStore, VAList, MachinePointerInfo());
// Round the value down to an f32.
SDValue NarrowFP =
DAG.getNode(ISD::FP_ROUND, DL, VT, WideFP.getValue(0),
DAG.getIntPtrConstant(1, DL, /*isTarget=*/true));
SDValue Ops[] = { NarrowFP, WideFP.getValue(1) };
// Merge the rounded value with the chain output of the load.
return DAG.getMergeValues(Ops, DL);
}
return DAG.getLoad(VT, DL, APStore, VAList, MachinePointerInfo());
}
SDValue AArch64TargetLowering::LowerFRAMEADDR(SDValue Op,
SelectionDAG &DAG) const {
MachineFrameInfo &MFI = DAG.getMachineFunction().getFrameInfo();
MFI.setFrameAddressIsTaken(true);
EVT VT = Op.getValueType();
SDLoc DL(Op);
unsigned Depth = Op.getConstantOperandVal(0);
SDValue FrameAddr =
DAG.getCopyFromReg(DAG.getEntryNode(), DL, AArch64::FP, MVT::i64);
while (Depth--)
FrameAddr = DAG.getLoad(VT, DL, DAG.getEntryNode(), FrameAddr,
MachinePointerInfo());
if (Subtarget->isTargetILP32())
FrameAddr = DAG.getNode(ISD::AssertZext, DL, MVT::i64, FrameAddr,
DAG.getValueType(VT));
return FrameAddr;
}
SDValue AArch64TargetLowering::LowerSPONENTRY(SDValue Op,
SelectionDAG &DAG) const {
MachineFrameInfo &MFI = DAG.getMachineFunction().getFrameInfo();
EVT VT = getPointerTy(DAG.getDataLayout());
int FI = MFI.CreateFixedObject(4, 0, false);
return DAG.getFrameIndex(FI, VT);
}
#define GET_REGISTER_MATCHER
#include "AArch64GenAsmMatcher.inc"
// FIXME? Maybe this could be a TableGen attribute on some registers and
// this table could be generated automatically from RegInfo.
Register AArch64TargetLowering::
getRegisterByName(const char* RegName, LLT VT, const MachineFunction &MF) const {
Register Reg = MatchRegisterName(RegName);
if (AArch64::X1 <= Reg && Reg <= AArch64::X28) {
const AArch64RegisterInfo *MRI = Subtarget->getRegisterInfo();
unsigned DwarfRegNum = MRI->getDwarfRegNum(Reg, false);
if (!Subtarget->isXRegisterReserved(DwarfRegNum) &&
!MRI->isReservedReg(MF, Reg))
Reg = Register();
}
return Reg;
}
SDValue AArch64TargetLowering::LowerADDROFRETURNADDR(SDValue Op,
SelectionDAG &DAG) const {
DAG.getMachineFunction().getFrameInfo().setFrameAddressIsTaken(true);
EVT VT = Op.getValueType();
SDLoc DL(Op);
SDValue FrameAddr =
DAG.getCopyFromReg(DAG.getEntryNode(), DL, AArch64::FP, VT);
SDValue Offset = DAG.getConstant(8, DL, getPointerTy(DAG.getDataLayout()));
return DAG.getNode(ISD::ADD, DL, VT, FrameAddr, Offset);
}
SDValue AArch64TargetLowering::LowerRETURNADDR(SDValue Op,
SelectionDAG &DAG) const {
MachineFunction &MF = DAG.getMachineFunction();
MachineFrameInfo &MFI = MF.getFrameInfo();
MFI.setReturnAddressIsTaken(true);
EVT VT = Op.getValueType();
SDLoc DL(Op);
unsigned Depth = Op.getConstantOperandVal(0);
SDValue ReturnAddress;
if (Depth) {
SDValue FrameAddr = LowerFRAMEADDR(Op, DAG);
SDValue Offset = DAG.getConstant(8, DL, getPointerTy(DAG.getDataLayout()));
ReturnAddress = DAG.getLoad(
VT, DL, DAG.getEntryNode(),
DAG.getNode(ISD::ADD, DL, VT, FrameAddr, Offset), MachinePointerInfo());
} else {
// Return LR, which contains the return address. Mark it an implicit
// live-in.
Register Reg = MF.addLiveIn(AArch64::LR, &AArch64::GPR64RegClass);
ReturnAddress = DAG.getCopyFromReg(DAG.getEntryNode(), DL, Reg, VT);
}
// The XPACLRI instruction assembles to a hint-space instruction before
// Armv8.3-A therefore this instruction can be safely used for any pre
// Armv8.3-A architectures. On Armv8.3-A and onwards XPACI is available so use
// that instead.
SDNode *St;
if (Subtarget->hasPAuth()) {
St = DAG.getMachineNode(AArch64::XPACI, DL, VT, ReturnAddress);
} else {
// XPACLRI operates on LR therefore we must move the operand accordingly.
SDValue Chain =
DAG.getCopyToReg(DAG.getEntryNode(), DL, AArch64::LR, ReturnAddress);
St = DAG.getMachineNode(AArch64::XPACLRI, DL, VT, Chain);
}
return SDValue(St, 0);
}
/// LowerShiftParts - Lower SHL_PARTS/SRA_PARTS/SRL_PARTS, which returns two
/// i32 values and take a 2 x i32 value to shift plus a shift amount.
SDValue AArch64TargetLowering::LowerShiftParts(SDValue Op,
SelectionDAG &DAG) const {
SDValue Lo, Hi;
expandShiftParts(Op.getNode(), Lo, Hi, DAG);
return DAG.getMergeValues({Lo, Hi}, SDLoc(Op));
}
bool AArch64TargetLowering::isOffsetFoldingLegal(
const GlobalAddressSDNode *GA) const {
// Offsets are folded in the DAG combine rather than here so that we can
// intelligently choose an offset based on the uses.
return false;
}
bool AArch64TargetLowering::isFPImmLegal(const APFloat &Imm, EVT VT,
bool OptForSize) const {
bool IsLegal = false;
// We can materialize #0.0 as fmov $Rd, XZR for 64-bit, 32-bit cases, and
// 16-bit case when target has full fp16 support.
// We encode bf16 bit patterns as if they were fp16. This results in very
// strange looking assembly but should populate the register with appropriate
// values. Let's say we wanted to encode 0xR3FC0 which is 1.5 in BF16. We will
// end up encoding this as the imm8 0x7f. This imm8 will be expanded to the
// FP16 1.9375 which shares the same bit pattern as BF16 1.5.
// FIXME: We should be able to handle f128 as well with a clever lowering.
const APInt ImmInt = Imm.bitcastToAPInt();
if (VT == MVT::f64)
IsLegal = AArch64_AM::getFP64Imm(ImmInt) != -1 || Imm.isPosZero();
else if (VT == MVT::f32)
IsLegal = AArch64_AM::getFP32Imm(ImmInt) != -1 || Imm.isPosZero();
else if (VT == MVT::f16 || VT == MVT::bf16)
IsLegal =
(Subtarget->hasFullFP16() && AArch64_AM::getFP16Imm(ImmInt) != -1) ||
Imm.isPosZero();
// If we can not materialize in immediate field for fmov, check if the
// value can be encoded as the immediate operand of a logical instruction.
// The immediate value will be created with either MOVZ, MOVN, or ORR.
// TODO: fmov h0, w0 is also legal, however we don't have an isel pattern to
// generate that fmov.
if (!IsLegal && (VT == MVT::f64 || VT == MVT::f32)) {
// The cost is actually exactly the same for mov+fmov vs. adrp+ldr;
// however the mov+fmov sequence is always better because of the reduced
// cache pressure. The timings are still the same if you consider
// movw+movk+fmov vs. adrp+ldr (it's one instruction longer, but the
// movw+movk is fused). So we limit up to 2 instrdduction at most.
SmallVector<AArch64_IMM::ImmInsnModel, 4> Insn;
AArch64_IMM::expandMOVImm(ImmInt.getZExtValue(), VT.getSizeInBits(), Insn);
assert(Insn.size() <= 4 &&
"Should be able to build any value with at most 4 moves");
unsigned Limit = (OptForSize ? 1 : (Subtarget->hasFuseLiterals() ? 4 : 2));
IsLegal = Insn.size() <= Limit;
}
LLVM_DEBUG(dbgs() << (IsLegal ? "Legal " : "Illegal ") << VT
<< " imm value: "; Imm.dump(););
return IsLegal;
}
//===----------------------------------------------------------------------===//
// AArch64 Optimization Hooks
//===----------------------------------------------------------------------===//
static SDValue getEstimate(const AArch64Subtarget *ST, unsigned Opcode,
SDValue Operand, SelectionDAG &DAG,
int &ExtraSteps) {
EVT VT = Operand.getValueType();
if ((ST->hasNEON() &&
(VT == MVT::f64 || VT == MVT::v1f64 || VT == MVT::v2f64 ||
VT == MVT::f32 || VT == MVT::v1f32 || VT == MVT::v2f32 ||
VT == MVT::v4f32)) ||
(ST->hasSVE() &&
(VT == MVT::nxv8f16 || VT == MVT::nxv4f32 || VT == MVT::nxv2f64))) {
if (ExtraSteps == TargetLoweringBase::ReciprocalEstimate::Unspecified) {
// For the reciprocal estimates, convergence is quadratic, so the number
// of digits is doubled after each iteration. In ARMv8, the accuracy of
// the initial estimate is 2^-8. Thus the number of extra steps to refine
// the result for float (23 mantissa bits) is 2 and for double (52
// mantissa bits) is 3.
constexpr unsigned AccurateBits = 8;
unsigned DesiredBits = APFloat::semanticsPrecision(VT.getFltSemantics());
ExtraSteps = DesiredBits <= AccurateBits
? 0
: Log2_64_Ceil(DesiredBits) - Log2_64_Ceil(AccurateBits);
}
return DAG.getNode(Opcode, SDLoc(Operand), VT, Operand);
}
return SDValue();
}
SDValue
AArch64TargetLowering::getSqrtInputTest(SDValue Op, SelectionDAG &DAG,
const DenormalMode &Mode) const {
SDLoc DL(Op);
EVT VT = Op.getValueType();
EVT CCVT = getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(), VT);
SDValue FPZero = DAG.getConstantFP(0.0, DL, VT);
return DAG.getSetCC(DL, CCVT, Op, FPZero, ISD::SETEQ);
}
SDValue
AArch64TargetLowering::getSqrtResultForDenormInput(SDValue Op,
SelectionDAG &DAG) const {
return Op;
}
SDValue AArch64TargetLowering::getSqrtEstimate(SDValue Operand,
SelectionDAG &DAG, int Enabled,
int &ExtraSteps,
bool &UseOneConst,
bool Reciprocal) const {
if (Enabled == ReciprocalEstimate::Enabled ||
(Enabled == ReciprocalEstimate::Unspecified && Subtarget->useRSqrt()))
if (SDValue Estimate = getEstimate(Subtarget, AArch64ISD::FRSQRTE, Operand,
DAG, ExtraSteps)) {
SDLoc DL(Operand);
EVT VT = Operand.getValueType();
SDNodeFlags Flags = SDNodeFlags::AllowReassociation;
// Newton reciprocal square root iteration: E * 0.5 * (3 - X * E^2)
// AArch64 reciprocal square root iteration instruction: 0.5 * (3 - M * N)
for (int i = ExtraSteps; i > 0; --i) {
SDValue Step = DAG.getNode(ISD::FMUL, DL, VT, Estimate, Estimate,
Flags);
Step = DAG.getNode(AArch64ISD::FRSQRTS, DL, VT, Operand, Step, Flags);
Estimate = DAG.getNode(ISD::FMUL, DL, VT, Estimate, Step, Flags);
}
if (!Reciprocal)
Estimate = DAG.getNode(ISD::FMUL, DL, VT, Operand, Estimate, Flags);
ExtraSteps = 0;
return Estimate;
}
return SDValue();
}
SDValue AArch64TargetLowering::getRecipEstimate(SDValue Operand,
SelectionDAG &DAG, int Enabled,
int &ExtraSteps) const {
if (Enabled == ReciprocalEstimate::Enabled)
if (SDValue Estimate = getEstimate(Subtarget, AArch64ISD::FRECPE, Operand,
DAG, ExtraSteps)) {
SDLoc DL(Operand);
EVT VT = Operand.getValueType();
SDNodeFlags Flags = SDNodeFlags::AllowReassociation;
// Newton reciprocal iteration: E * (2 - X * E)
// AArch64 reciprocal iteration instruction: (2 - M * N)
for (int i = ExtraSteps; i > 0; --i) {
SDValue Step = DAG.getNode(AArch64ISD::FRECPS, DL, VT, Operand,
Estimate, Flags);
Estimate = DAG.getNode(ISD::FMUL, DL, VT, Estimate, Step, Flags);
}
ExtraSteps = 0;
return Estimate;
}
return SDValue();
}
//===----------------------------------------------------------------------===//
// AArch64 Inline Assembly Support
//===----------------------------------------------------------------------===//
// Table of Constraints
// TODO: This is the current set of constraints supported by ARM for the
// compiler, not all of them may make sense.
//
// r - A general register
// w - An FP/SIMD register of some size in the range v0-v31
// x - An FP/SIMD register of some size in the range v0-v15
// I - Constant that can be used with an ADD instruction
// J - Constant that can be used with a SUB instruction
// K - Constant that can be used with a 32-bit logical instruction
// L - Constant that can be used with a 64-bit logical instruction
// M - Constant that can be used as a 32-bit MOV immediate
// N - Constant that can be used as a 64-bit MOV immediate
// Q - A memory reference with base register and no offset
// S - A symbolic address
// Y - Floating point constant zero
// Z - Integer constant zero
//
// Note that general register operands will be output using their 64-bit x
// register name, whatever the size of the variable, unless the asm operand
// is prefixed by the %w modifier. Floating-point and SIMD register operands
// will be output with the v prefix unless prefixed by the %b, %h, %s, %d or
// %q modifier.
const char *AArch64TargetLowering::LowerXConstraint(EVT ConstraintVT) const {
// At this point, we have to lower this constraint to something else, so we
// lower it to an "r" or "w". However, by doing this we will force the result
// to be in register, while the X constraint is much more permissive.
//
// Although we are correct (we are free to emit anything, without
// constraints), we might break use cases that would expect us to be more
// efficient and emit something else.
if (!Subtarget->hasFPARMv8())
return "r";
if (ConstraintVT.isFloatingPoint())
return "w";
if (ConstraintVT.isVector() &&
(ConstraintVT.getSizeInBits() == 64 ||
ConstraintVT.getSizeInBits() == 128))
return "w";
return "r";
}
enum class PredicateConstraint { Uph, Upl, Upa };
// Returns a {Reg, RegisterClass} tuple if the constraint is
// a specific predicate register.
//
// For some constraint like "{pn3}" the default path in
// TargetLowering::getRegForInlineAsmConstraint() leads it to determine that a
// suitable register class for this register is "PPRorPNR", after which it
// determines that nxv16i1 is an appropriate type for the constraint, which is
// not what we want. The code here pre-empts this by matching the register
// explicitly.
static std::optional<std::pair<unsigned, const TargetRegisterClass *>>
parseSVERegAsConstraint(StringRef Constraint) {
if (!Constraint.starts_with('{') || !Constraint.ends_with('}') ||
(Constraint[1] != 'p' && Constraint[1] != 'z'))
return std::nullopt;
bool IsPredicate = Constraint[1] == 'p';
Constraint = Constraint.substr(2, Constraint.size() - 3);
bool IsPredicateAsCount = IsPredicate && Constraint.starts_with("n");
if (IsPredicateAsCount)
Constraint = Constraint.drop_front(1);
unsigned V;
if (Constraint.getAsInteger(10, V) || V > 31)
return std::nullopt;
if (IsPredicateAsCount)
return std::make_pair(AArch64::PN0 + V, &AArch64::PNRRegClass);
if (IsPredicate)
return std::make_pair(AArch64::P0 + V, &AArch64::PPRRegClass);
return std::make_pair(AArch64::Z0 + V, &AArch64::ZPRRegClass);
}
static std::optional<PredicateConstraint>
parsePredicateConstraint(StringRef Constraint) {
return StringSwitch<std::optional<PredicateConstraint>>(Constraint)
.Case("Uph", PredicateConstraint::Uph)
.Case("Upl", PredicateConstraint::Upl)
.Case("Upa", PredicateConstraint::Upa)
.Default(std::nullopt);
}
static const TargetRegisterClass *
getPredicateRegisterClass(PredicateConstraint Constraint, EVT VT) {
if (VT != MVT::aarch64svcount &&
(!VT.isScalableVector() || VT.getVectorElementType() != MVT::i1))
return nullptr;
switch (Constraint) {
case PredicateConstraint::Uph:
return VT == MVT::aarch64svcount ? &AArch64::PNR_p8to15RegClass
: &AArch64::PPR_p8to15RegClass;
case PredicateConstraint::Upl:
return VT == MVT::aarch64svcount ? &AArch64::PNR_3bRegClass
: &AArch64::PPR_3bRegClass;
case PredicateConstraint::Upa:
return VT == MVT::aarch64svcount ? &AArch64::PNRRegClass
: &AArch64::PPRRegClass;
}
llvm_unreachable("Missing PredicateConstraint!");
}
enum class ReducedGprConstraint { Uci, Ucj };
static std::optional<ReducedGprConstraint>
parseReducedGprConstraint(StringRef Constraint) {
return StringSwitch<std::optional<ReducedGprConstraint>>(Constraint)
.Case("Uci", ReducedGprConstraint::Uci)
.Case("Ucj", ReducedGprConstraint::Ucj)
.Default(std::nullopt);
}
static const TargetRegisterClass *
getReducedGprRegisterClass(ReducedGprConstraint Constraint, EVT VT) {
if (!VT.isScalarInteger() || VT.getFixedSizeInBits() > 64)
return nullptr;
switch (Constraint) {
case ReducedGprConstraint::Uci:
return &AArch64::MatrixIndexGPR32_8_11RegClass;
case ReducedGprConstraint::Ucj:
return &AArch64::MatrixIndexGPR32_12_15RegClass;
}
llvm_unreachable("Missing ReducedGprConstraint!");
}
// The set of cc code supported is from
// https://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html#Flag-Output-Operands
static AArch64CC::CondCode parseConstraintCode(llvm::StringRef Constraint) {
AArch64CC::CondCode Cond = StringSwitch<AArch64CC::CondCode>(Constraint)
.Case("{@cchi}", AArch64CC::HI)
.Case("{@cccs}", AArch64CC::HS)
.Case("{@cclo}", AArch64CC::LO)
.Case("{@ccls}", AArch64CC::LS)
.Case("{@cccc}", AArch64CC::LO)
.Case("{@cceq}", AArch64CC::EQ)
.Case("{@ccgt}", AArch64CC::GT)
.Case("{@ccge}", AArch64CC::GE)
.Case("{@cclt}", AArch64CC::LT)
.Case("{@ccle}", AArch64CC::LE)
.Case("{@cchs}", AArch64CC::HS)
.Case("{@ccne}", AArch64CC::NE)
.Case("{@ccvc}", AArch64CC::VC)
.Case("{@ccpl}", AArch64CC::PL)
.Case("{@ccvs}", AArch64CC::VS)
.Case("{@ccmi}", AArch64CC::MI)
.Default(AArch64CC::Invalid);
return Cond;
}
/// Helper function to create 'CSET', which is equivalent to 'CSINC <Wd>, WZR,
/// WZR, invert(<cond>)'.
static SDValue getSETCC(AArch64CC::CondCode CC, SDValue NZCV, const SDLoc &DL,
SelectionDAG &DAG) {
return DAG.getNode(
AArch64ISD::CSINC, DL, MVT::i32, DAG.getConstant(0, DL, MVT::i32),
DAG.getConstant(0, DL, MVT::i32),
DAG.getConstant(getInvertedCondCode(CC), DL, MVT::i32), NZCV);
}
// Lower @cc flag output via getSETCC.
SDValue AArch64TargetLowering::LowerAsmOutputForConstraint(
SDValue &Chain, SDValue &Glue, const SDLoc &DL,
const AsmOperandInfo &OpInfo, SelectionDAG &DAG) const {
AArch64CC::CondCode Cond = parseConstraintCode(OpInfo.ConstraintCode);
if (Cond == AArch64CC::Invalid)
return SDValue();
// The output variable should be a scalar integer.
if (OpInfo.ConstraintVT.isVector() || !OpInfo.ConstraintVT.isInteger() ||
OpInfo.ConstraintVT.getSizeInBits() < 8)
report_fatal_error("Flag output operand is of invalid type");
// Get NZCV register. Only update chain when copyfrom is glued.
if (Glue.getNode()) {
Glue = DAG.getCopyFromReg(Chain, DL, AArch64::NZCV, MVT::i32, Glue);
Chain = Glue.getValue(1);
} else
Glue = DAG.getCopyFromReg(Chain, DL, AArch64::NZCV, MVT::i32);
// Extract CC code.
SDValue CC = getSETCC(Cond, Glue, DL, DAG);
SDValue Result;
// Truncate or ZERO_EXTEND based on value types.
if (OpInfo.ConstraintVT.getSizeInBits() <= 32)
Result = DAG.getNode(ISD::TRUNCATE, DL, OpInfo.ConstraintVT, CC);
else
Result = DAG.getNode(ISD::ZERO_EXTEND, DL, OpInfo.ConstraintVT, CC);
return Result;
}
/// getConstraintType - Given a constraint letter, return the type of
/// constraint it is for this target.
AArch64TargetLowering::ConstraintType
AArch64TargetLowering::getConstraintType(StringRef Constraint) const {
if (Constraint.size() == 1) {
switch (Constraint[0]) {
default:
break;
case 'x':
case 'w':
case 'y':
return C_RegisterClass;
// An address with a single base register. Due to the way we
// currently handle addresses it is the same as 'r'.
case 'Q':
return C_Memory;
case 'I':
case 'J':
case 'K':
case 'L':
case 'M':
case 'N':
case 'Y':
case 'Z':
return C_Immediate;
case 'z':
case 'S': // A symbol or label reference with a constant offset
return C_Other;
}
} else if (parsePredicateConstraint(Constraint))
return C_RegisterClass;
else if (parseReducedGprConstraint(Constraint))
return C_RegisterClass;
else if (parseConstraintCode(Constraint) != AArch64CC::Invalid)
return C_Other;
return TargetLowering::getConstraintType(Constraint);
}
/// Examine constraint type and operand type and determine a weight value.
/// This object must already have been set up with the operand type
/// and the current alternative constraint selected.
TargetLowering::ConstraintWeight
AArch64TargetLowering::getSingleConstraintMatchWeight(
AsmOperandInfo &info, const char *constraint) const {
ConstraintWeight weight = CW_Invalid;
Value *CallOperandVal = info.CallOperandVal;
// If we don't have a value, we can't do a match,
// but allow it at the lowest weight.
if (!CallOperandVal)
return CW_Default;
Type *type = CallOperandVal->getType();
// Look at the constraint type.
switch (*constraint) {
default:
weight = TargetLowering::getSingleConstraintMatchWeight(info, constraint);
break;
case 'x':
case 'w':
case 'y':
if (type->isFloatingPointTy() || type->isVectorTy())
weight = CW_Register;
break;
case 'z':
weight = CW_Constant;
break;
case 'U':
if (parsePredicateConstraint(constraint) ||
parseReducedGprConstraint(constraint))
weight = CW_Register;
break;
}
return weight;
}
std::pair<unsigned, const TargetRegisterClass *>
AArch64TargetLowering::getRegForInlineAsmConstraint(
const TargetRegisterInfo *TRI, StringRef Constraint, MVT VT) const {
if (Constraint.size() == 1) {
switch (Constraint[0]) {
case 'r':
if (VT.isScalableVector())
return std::make_pair(0U, nullptr);
if (Subtarget->hasLS64() && VT.getSizeInBits() == 512)
return std::make_pair(0U, &AArch64::GPR64x8ClassRegClass);
if (VT.getFixedSizeInBits() == 64)
return std::make_pair(0U, &AArch64::GPR64commonRegClass);
return std::make_pair(0U, &AArch64::GPR32commonRegClass);
case 'w': {
if (!Subtarget->hasFPARMv8())
break;
if (VT.isScalableVector()) {
if (VT.getVectorElementType() != MVT::i1)
return std::make_pair(0U, &AArch64::ZPRRegClass);
return std::make_pair(0U, nullptr);
}
if (VT == MVT::Other)
break;
uint64_t VTSize = VT.getFixedSizeInBits();
if (VTSize == 16)
return std::make_pair(0U, &AArch64::FPR16RegClass);
if (VTSize == 32)
return std::make_pair(0U, &AArch64::FPR32RegClass);
if (VTSize == 64)
return std::make_pair(0U, &AArch64::FPR64RegClass);
if (VTSize == 128)
return std::make_pair(0U, &AArch64::FPR128RegClass);
break;
}
// The instructions that this constraint is designed for can
// only take 128-bit registers so just use that regclass.
case 'x':
if (!Subtarget->hasFPARMv8())
break;
if (VT.isScalableVector())
return std::make_pair(0U, &AArch64::ZPR_4bRegClass);
if (VT.getSizeInBits() == 128)
return std::make_pair(0U, &AArch64::FPR128_loRegClass);
break;
case 'y':
if (!Subtarget->hasFPARMv8())
break;
if (VT.isScalableVector())
return std::make_pair(0U, &AArch64::ZPR_3bRegClass);
break;
}
} else {
if (const auto P = parseSVERegAsConstraint(Constraint)) {
// SME functions that are not in streaming mode, should
// still observe clobbers of Z-registers by clobbering
// the lower 128bits of those registers.
if (AArch64::ZPRRegClass.hasSubClassEq(P->second) &&
!Subtarget->isSVEorStreamingSVEAvailable())
return std::make_pair(TRI->getSubReg(P->first, AArch64::zsub),
&AArch64::FPR128RegClass);
return *P;
}
if (const auto PC = parsePredicateConstraint(Constraint))
if (const auto *RegClass = getPredicateRegisterClass(*PC, VT))
return std::make_pair(0U, RegClass);
if (const auto RGC = parseReducedGprConstraint(Constraint))
if (const auto *RegClass = getReducedGprRegisterClass(*RGC, VT))
return std::make_pair(0U, RegClass);
}
if (StringRef("{cc}").equals_insensitive(Constraint) ||
parseConstraintCode(Constraint) != AArch64CC::Invalid)
return std::make_pair(unsigned(AArch64::NZCV), &AArch64::CCRRegClass);
if (Constraint == "{za}") {
return std::make_pair(unsigned(AArch64::ZA), &AArch64::MPRRegClass);
}
if (Constraint == "{zt0}") {
return std::make_pair(unsigned(AArch64::ZT0), &AArch64::ZTRRegClass);
}
// Use the default implementation in TargetLowering to convert the register
// constraint into a member of a register class.
std::pair<unsigned, const TargetRegisterClass *> Res;
Res = TargetLowering::getRegForInlineAsmConstraint(TRI, Constraint, VT);
// Not found as a standard register?
if (!Res.second) {
unsigned Size = Constraint.size();
if ((Size == 4 || Size == 5) && Constraint[0] == '{' &&
tolower(Constraint[1]) == 'v' && Constraint[Size - 1] == '}') {
int RegNo;
bool Failed = Constraint.slice(2, Size - 1).getAsInteger(10, RegNo);
if (!Failed && RegNo >= 0 && RegNo <= 31) {
// v0 - v31 are aliases of q0 - q31 or d0 - d31 depending on size.
// By default we'll emit v0-v31 for this unless there's a modifier where
// we'll emit the correct register as well.
if (VT != MVT::Other && VT.getSizeInBits() == 64) {
Res.first = AArch64::FPR64RegClass.getRegister(RegNo);
Res.second = &AArch64::FPR64RegClass;
} else {
Res.first = AArch64::FPR128RegClass.getRegister(RegNo);
Res.second = &AArch64::FPR128RegClass;
}
}
}
}
if (Res.second && !Subtarget->hasFPARMv8() &&
!AArch64::GPR32allRegClass.hasSubClassEq(Res.second) &&
!AArch64::GPR64allRegClass.hasSubClassEq(Res.second))
return std::make_pair(0U, nullptr);
return Res;
}
EVT AArch64TargetLowering::getAsmOperandValueType(const DataLayout &DL,
llvm::Type *Ty,
bool AllowUnknown) const {
if (Subtarget->hasLS64() && Ty->isIntegerTy(512))
return EVT(MVT::i64x8);
return TargetLowering::getAsmOperandValueType(DL, Ty, AllowUnknown);
}
/// LowerAsmOperandForConstraint - Lower the specified operand into the Ops
/// vector. If it is invalid, don't add anything to Ops.
void AArch64TargetLowering::LowerAsmOperandForConstraint(
SDValue Op, StringRef Constraint, std::vector<SDValue> &Ops,
SelectionDAG &DAG) const {
SDValue Result;
// Currently only support length 1 constraints.
if (Constraint.size() != 1)
return;
char ConstraintLetter = Constraint[0];
switch (ConstraintLetter) {
default:
break;
// This set of constraints deal with valid constants for various instructions.
// Validate and return a target constant for them if we can.
case 'z': {
// 'z' maps to xzr or wzr so it needs an input of 0.
if (!isNullConstant(Op))
return;
if (Op.getValueType() == MVT::i64)
Result = DAG.getRegister(AArch64::XZR, MVT::i64);
else
Result = DAG.getRegister(AArch64::WZR, MVT::i32);
break;
}
case 'S':
// Use the generic code path for "s". In GCC's aarch64 port, "S" is
// supported for PIC while "s" isn't, making "s" less useful. We implement
// "S" but not "s".
TargetLowering::LowerAsmOperandForConstraint(Op, "s", Ops, DAG);
break;
case 'I':
case 'J':
case 'K':
case 'L':
case 'M':
case 'N':
ConstantSDNode *C = dyn_cast<ConstantSDNode>(Op);
if (!C)
return;
// Grab the value and do some validation.
uint64_t CVal = C->getZExtValue();
switch (ConstraintLetter) {
// The I constraint applies only to simple ADD or SUB immediate operands:
// i.e. 0 to 4095 with optional shift by 12
// The J constraint applies only to ADD or SUB immediates that would be
// valid when negated, i.e. if [an add pattern] were to be output as a SUB
// instruction [or vice versa], in other words -1 to -4095 with optional
// left shift by 12.
case 'I':
if (isUInt<12>(CVal) || isShiftedUInt<12, 12>(CVal))
break;
return;
case 'J': {
uint64_t NVal = -C->getSExtValue();
if (isUInt<12>(NVal) || isShiftedUInt<12, 12>(NVal)) {
CVal = C->getSExtValue();
break;
}
return;
}
// The K and L constraints apply *only* to logical immediates, including
// what used to be the MOVI alias for ORR (though the MOVI alias has now
// been removed and MOV should be used). So these constraints have to
// distinguish between bit patterns that are valid 32-bit or 64-bit
// "bitmask immediates": for example 0xaaaaaaaa is a valid bimm32 (K), but
// not a valid bimm64 (L) where 0xaaaaaaaaaaaaaaaa would be valid, and vice
// versa.
case 'K':
if (AArch64_AM::isLogicalImmediate(CVal, 32))
break;
return;
case 'L':
if (AArch64_AM::isLogicalImmediate(CVal, 64))
break;
return;
// The M and N constraints are a superset of K and L respectively, for use
// with the MOV (immediate) alias. As well as the logical immediates they
// also match 32 or 64-bit immediates that can be loaded either using a
// *single* MOVZ or MOVN , such as 32-bit 0x12340000, 0x00001234, 0xffffedca
// (M) or 64-bit 0x1234000000000000 (N) etc.
// As a note some of this code is liberally stolen from the asm parser.
case 'M': {
if (!isUInt<32>(CVal))
return;
if (AArch64_AM::isLogicalImmediate(CVal, 32))
break;
if ((CVal & 0xFFFF) == CVal)
break;
if ((CVal & 0xFFFF0000ULL) == CVal)
break;
uint64_t NCVal = ~(uint32_t)CVal;
if ((NCVal & 0xFFFFULL) == NCVal)
break;
if ((NCVal & 0xFFFF0000ULL) == NCVal)
break;
return;
}
case 'N': {
if (AArch64_AM::isLogicalImmediate(CVal, 64))
break;
if ((CVal & 0xFFFFULL) == CVal)
break;
if ((CVal & 0xFFFF0000ULL) == CVal)
break;
if ((CVal & 0xFFFF00000000ULL) == CVal)
break;
if ((CVal & 0xFFFF000000000000ULL) == CVal)
break;
uint64_t NCVal = ~CVal;
if ((NCVal & 0xFFFFULL) == NCVal)
break;
if ((NCVal & 0xFFFF0000ULL) == NCVal)
break;
if ((NCVal & 0xFFFF00000000ULL) == NCVal)
break;
if ((NCVal & 0xFFFF000000000000ULL) == NCVal)
break;
return;
}
default:
return;
}
// All assembler immediates are 64-bit integers.
Result = DAG.getTargetConstant(CVal, SDLoc(Op), MVT::i64);
break;
}
if (Result.getNode()) {
Ops.push_back(Result);
return;
}
return TargetLowering::LowerAsmOperandForConstraint(Op, Constraint, Ops, DAG);
}
//===----------------------------------------------------------------------===//
// AArch64 Advanced SIMD Support
//===----------------------------------------------------------------------===//
/// WidenVector - Given a value in the V64 register class, produce the
/// equivalent value in the V128 register class.
static SDValue WidenVector(SDValue V64Reg, SelectionDAG &DAG) {
EVT VT = V64Reg.getValueType();
unsigned NarrowSize = VT.getVectorNumElements();
MVT EltTy = VT.getVectorElementType().getSimpleVT();
MVT WideTy = MVT::getVectorVT(EltTy, 2 * NarrowSize);
SDLoc DL(V64Reg);
return DAG.getNode(ISD::INSERT_SUBVECTOR, DL, WideTy, DAG.getUNDEF(WideTy),
V64Reg, DAG.getConstant(0, DL, MVT::i64));
}
/// getExtFactor - Determine the adjustment factor for the position when
/// generating an "extract from vector registers" instruction.
static unsigned getExtFactor(SDValue &V) {
EVT EltType = V.getValueType().getVectorElementType();
return EltType.getSizeInBits() / 8;
}
// Check if a vector is built from one vector via extracted elements of
// another together with an AND mask, ensuring that all elements fit
// within range. This can be reconstructed using AND and NEON's TBL1.
SDValue ReconstructShuffleWithRuntimeMask(SDValue Op, SelectionDAG &DAG) {
assert(Op.getOpcode() == ISD::BUILD_VECTOR && "Unknown opcode!");
SDLoc DL(Op);
EVT VT = Op.getValueType();
assert(!VT.isScalableVector() &&
"Scalable vectors cannot be used with ISD::BUILD_VECTOR");
// Can only recreate a shuffle with 16xi8 or 8xi8 elements, as they map
// directly to TBL1.
if (VT != MVT::v16i8 && VT != MVT::v8i8)
return SDValue();
unsigned NumElts = VT.getVectorNumElements();
assert((NumElts == 8 || NumElts == 16) &&
"Need to have exactly 8 or 16 elements in vector.");
SDValue SourceVec;
SDValue MaskSourceVec;
SmallVector<SDValue, 16> AndMaskConstants;
for (unsigned i = 0; i < NumElts; ++i) {
SDValue V = Op.getOperand(i);
if (V.getOpcode() != ISD::EXTRACT_VECTOR_ELT)
return SDValue();
SDValue OperandSourceVec = V.getOperand(0);
if (!SourceVec)
SourceVec = OperandSourceVec;
else if (SourceVec != OperandSourceVec)
return SDValue();
// This only looks at shuffles with elements that are
// a) truncated by a constant AND mask extracted from a mask vector, or
// b) extracted directly from a mask vector.
SDValue MaskSource = V.getOperand(1);
if (MaskSource.getOpcode() == ISD::AND) {
if (!isa<ConstantSDNode>(MaskSource.getOperand(1)))
return SDValue();
AndMaskConstants.push_back(MaskSource.getOperand(1));
MaskSource = MaskSource->getOperand(0);
} else if (!AndMaskConstants.empty()) {
// Either all or no operands should have an AND mask.
return SDValue();
}
// An ANY_EXTEND may be inserted between the AND and the source vector
// extraction. We don't care about that, so we can just skip it.
if (MaskSource.getOpcode() == ISD::ANY_EXTEND)
MaskSource = MaskSource.getOperand(0);
if (MaskSource.getOpcode() != ISD::EXTRACT_VECTOR_ELT)
return SDValue();
SDValue MaskIdx = MaskSource.getOperand(1);
if (!isa<ConstantSDNode>(MaskIdx) ||
!cast<ConstantSDNode>(MaskIdx)->getConstantIntValue()->equalsInt(i))
return SDValue();
// We only apply this if all elements come from the same vector with the
// same vector type.
if (!MaskSourceVec) {
MaskSourceVec = MaskSource->getOperand(0);
if (MaskSourceVec.getValueType() != VT)
return SDValue();
} else if (MaskSourceVec != MaskSource->getOperand(0)) {
return SDValue();
}
}
// We need a v16i8 for TBL, so we extend the source with a placeholder vector
// for v8i8 to get a v16i8. As the pattern we are replacing is extract +
// insert, we know that the index in the mask must be smaller than the number
// of elements in the source, or we would have an out-of-bounds access.
if (NumElts == 8)
SourceVec = DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v16i8, SourceVec,
DAG.getUNDEF(VT));
// Preconditions met, so we can use a vector (AND +) TBL to build this vector.
if (!AndMaskConstants.empty())
MaskSourceVec = DAG.getNode(ISD::AND, DL, VT, MaskSourceVec,
DAG.getBuildVector(VT, DL, AndMaskConstants));
return DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, VT,
DAG.getConstant(Intrinsic::aarch64_neon_tbl1, DL, MVT::i32), SourceVec,
MaskSourceVec);
}
// Gather data to see if the operation can be modelled as a
// shuffle in combination with VEXTs.
SDValue AArch64TargetLowering::ReconstructShuffle(SDValue Op,
SelectionDAG &DAG) const {
assert(Op.getOpcode() == ISD::BUILD_VECTOR && "Unknown opcode!");
LLVM_DEBUG(dbgs() << "AArch64TargetLowering::ReconstructShuffle\n");
SDLoc DL(Op);
EVT VT = Op.getValueType();
assert(!VT.isScalableVector() &&
"Scalable vectors cannot be used with ISD::BUILD_VECTOR");
unsigned NumElts = VT.getVectorNumElements();
struct ShuffleSourceInfo {
SDValue Vec;
unsigned MinElt;
unsigned MaxElt;
// We may insert some combination of BITCASTs and VEXT nodes to force Vec to
// be compatible with the shuffle we intend to construct. As a result
// ShuffleVec will be some sliding window into the original Vec.
SDValue ShuffleVec;
// Code should guarantee that element i in Vec starts at element "WindowBase
// + i * WindowScale in ShuffleVec".
int WindowBase;
int WindowScale;
ShuffleSourceInfo(SDValue Vec)
: Vec(Vec), MinElt(std::numeric_limits<unsigned>::max()), MaxElt(0),
ShuffleVec(Vec), WindowBase(0), WindowScale(1) {}
bool operator ==(SDValue OtherVec) { return Vec == OtherVec; }
};
// First gather all vectors used as an immediate source for this BUILD_VECTOR
// node.
SmallVector<ShuffleSourceInfo, 2> Sources;
for (unsigned i = 0; i < NumElts; ++i) {
SDValue V = Op.getOperand(i);
if (V.isUndef())
continue;
else if (V.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
!isa<ConstantSDNode>(V.getOperand(1)) ||
V.getOperand(0).getValueType().isScalableVector()) {
LLVM_DEBUG(
dbgs() << "Reshuffle failed: "
"a shuffle can only come from building a vector from "
"various elements of other fixed-width vectors, provided "
"their indices are constant\n");
return SDValue();
}
// Add this element source to the list if it's not already there.
SDValue SourceVec = V.getOperand(0);
auto Source = find(Sources, SourceVec);
if (Source == Sources.end())
Source = Sources.insert(Sources.end(), ShuffleSourceInfo(SourceVec));
// Update the minimum and maximum lane number seen.
unsigned EltNo = V.getConstantOperandVal(1);
Source->MinElt = std::min(Source->MinElt, EltNo);
Source->MaxElt = std::max(Source->MaxElt, EltNo);
}
// If we have 3 or 4 sources, try to generate a TBL, which will at least be
// better than moving to/from gpr registers for larger vectors.
if ((Sources.size() == 3 || Sources.size() == 4) && NumElts > 4) {
// Construct a mask for the tbl. We may need to adjust the index for types
// larger than i8.
SmallVector<unsigned, 16> Mask;
unsigned OutputFactor = VT.getScalarSizeInBits() / 8;
for (unsigned I = 0; I < NumElts; ++I) {
SDValue V = Op.getOperand(I);
if (V.isUndef()) {
for (unsigned OF = 0; OF < OutputFactor; OF++)
Mask.push_back(-1);
continue;
}
// Set the Mask lanes adjusted for the size of the input and output
// lanes. The Mask is always i8, so it will set OutputFactor lanes per
// output element, adjusted in their positions per input and output types.
unsigned Lane = V.getConstantOperandVal(1);
for (unsigned S = 0; S < Sources.size(); S++) {
if (V.getOperand(0) == Sources[S].Vec) {
unsigned InputSize = Sources[S].Vec.getScalarValueSizeInBits();
unsigned InputBase = 16 * S + Lane * InputSize / 8;
for (unsigned OF = 0; OF < OutputFactor; OF++)
Mask.push_back(InputBase + OF);
break;
}
}
}
// Construct the tbl3/tbl4 out of an intrinsic, the sources converted to
// v16i8, and the TBLMask
SmallVector<SDValue, 16> TBLOperands;
TBLOperands.push_back(DAG.getConstant(Sources.size() == 3
? Intrinsic::aarch64_neon_tbl3
: Intrinsic::aarch64_neon_tbl4,
DL, MVT::i32));
for (unsigned i = 0; i < Sources.size(); i++) {
SDValue Src = Sources[i].Vec;
EVT SrcVT = Src.getValueType();
Src = DAG.getBitcast(SrcVT.is64BitVector() ? MVT::v8i8 : MVT::v16i8, Src);
assert((SrcVT.is64BitVector() || SrcVT.is128BitVector()) &&
"Expected a legally typed vector");
if (SrcVT.is64BitVector())
Src = DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v16i8, Src,
DAG.getUNDEF(MVT::v8i8));
TBLOperands.push_back(Src);
}
SmallVector<SDValue, 16> TBLMask;
for (unsigned i = 0; i < Mask.size(); i++)
TBLMask.push_back(DAG.getConstant(Mask[i], DL, MVT::i32));
assert((Mask.size() == 8 || Mask.size() == 16) &&
"Expected a v8i8 or v16i8 Mask");
TBLOperands.push_back(DAG.getBuildVector(
Mask.size() == 8 ? MVT::v8i8 : MVT::v16i8, DL, TBLMask));
SDValue Shuffle =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL,
Mask.size() == 8 ? MVT::v8i8 : MVT::v16i8, TBLOperands);
return DAG.getBitcast(VT, Shuffle);
}
if (Sources.size() > 2) {
LLVM_DEBUG(dbgs() << "Reshuffle failed: currently only do something "
<< "sensible when at most two source vectors are "
<< "involved\n");
return SDValue();
}
// Find out the smallest element size among result and two sources, and use
// it as element size to build the shuffle_vector.
EVT SmallestEltTy = VT.getVectorElementType();
for (auto &Source : Sources) {
EVT SrcEltTy = Source.Vec.getValueType().getVectorElementType();
if (SrcEltTy.bitsLT(SmallestEltTy)) {
SmallestEltTy = SrcEltTy;
}
}
unsigned ResMultiplier =
VT.getScalarSizeInBits() / SmallestEltTy.getFixedSizeInBits();
uint64_t VTSize = VT.getFixedSizeInBits();
NumElts = VTSize / SmallestEltTy.getFixedSizeInBits();
EVT ShuffleVT = EVT::getVectorVT(*DAG.getContext(), SmallestEltTy, NumElts);
// If the source vector is too wide or too narrow, we may nevertheless be able
// to construct a compatible shuffle either by concatenating it with UNDEF or
// extracting a suitable range of elements.
for (auto &Src : Sources) {
EVT SrcVT = Src.ShuffleVec.getValueType();
TypeSize SrcVTSize = SrcVT.getSizeInBits();
if (SrcVTSize == TypeSize::getFixed(VTSize))
continue;
// This stage of the search produces a source with the same element type as
// the original, but with a total width matching the BUILD_VECTOR output.
EVT EltVT = SrcVT.getVectorElementType();
unsigned NumSrcElts = VTSize / EltVT.getFixedSizeInBits();
EVT DestVT = EVT::getVectorVT(*DAG.getContext(), EltVT, NumSrcElts);
if (SrcVTSize.getFixedValue() < VTSize) {
assert(2 * SrcVTSize == VTSize);
// We can pad out the smaller vector for free, so if it's part of a
// shuffle...
Src.ShuffleVec =
DAG.getNode(ISD::CONCAT_VECTORS, DL, DestVT, Src.ShuffleVec,
DAG.getUNDEF(Src.ShuffleVec.getValueType()));
continue;
}
if (SrcVTSize.getFixedValue() != 2 * VTSize) {
LLVM_DEBUG(
dbgs() << "Reshuffle failed: result vector too small to extract\n");
return SDValue();
}
if (Src.MaxElt - Src.MinElt >= NumSrcElts) {
LLVM_DEBUG(
dbgs() << "Reshuffle failed: span too large for a VEXT to cope\n");
return SDValue();
}
if (Src.MinElt >= NumSrcElts) {
// The extraction can just take the second half
Src.ShuffleVec =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, DestVT, Src.ShuffleVec,
DAG.getConstant(NumSrcElts, DL, MVT::i64));
Src.WindowBase = -NumSrcElts;
} else if (Src.MaxElt < NumSrcElts) {
// The extraction can just take the first half
Src.ShuffleVec =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, DestVT, Src.ShuffleVec,
DAG.getConstant(0, DL, MVT::i64));
} else {
// An actual VEXT is needed
SDValue VEXTSrc1 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, DestVT, Src.ShuffleVec,
DAG.getConstant(0, DL, MVT::i64));
SDValue VEXTSrc2 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, DestVT, Src.ShuffleVec,
DAG.getConstant(NumSrcElts, DL, MVT::i64));
unsigned Imm = Src.MinElt * getExtFactor(VEXTSrc1);
if (!SrcVT.is64BitVector()) {
LLVM_DEBUG(
dbgs() << "Reshuffle failed: don't know how to lower AArch64ISD::EXT "
"for SVE vectors.");
return SDValue();
}
Src.ShuffleVec =
DAG.getNode(AArch64ISD::EXT, DL, DestVT, VEXTSrc1, VEXTSrc2,
DAG.getConstant(Imm, DL, MVT::i32));
Src.WindowBase = -Src.MinElt;
}
}
// Another possible incompatibility occurs from the vector element types. We
// can fix this by bitcasting the source vectors to the same type we intend
// for the shuffle.
for (auto &Src : Sources) {
EVT SrcEltTy = Src.ShuffleVec.getValueType().getVectorElementType();
if (SrcEltTy == SmallestEltTy)
continue;
assert(ShuffleVT.getVectorElementType() == SmallestEltTy);
if (DAG.getDataLayout().isBigEndian()) {
Src.ShuffleVec =
DAG.getNode(AArch64ISD::NVCAST, DL, ShuffleVT, Src.ShuffleVec);
} else {
Src.ShuffleVec = DAG.getNode(ISD::BITCAST, DL, ShuffleVT, Src.ShuffleVec);
}
Src.WindowScale =
SrcEltTy.getFixedSizeInBits() / SmallestEltTy.getFixedSizeInBits();
Src.WindowBase *= Src.WindowScale;
}
// Final check before we try to actually produce a shuffle.
LLVM_DEBUG({
for (auto Src : Sources)
assert(Src.ShuffleVec.getValueType() == ShuffleVT);
});
// The stars all align, our next step is to produce the mask for the shuffle.
SmallVector<int, 8> Mask(ShuffleVT.getVectorNumElements(), -1);
int BitsPerShuffleLane = ShuffleVT.getScalarSizeInBits();
for (unsigned i = 0; i < VT.getVectorNumElements(); ++i) {
SDValue Entry = Op.getOperand(i);
if (Entry.isUndef())
continue;
auto Src = find(Sources, Entry.getOperand(0));
int EltNo = cast<ConstantSDNode>(Entry.getOperand(1))->getSExtValue();
// EXTRACT_VECTOR_ELT performs an implicit any_ext; BUILD_VECTOR an implicit
// trunc. So only std::min(SrcBits, DestBits) actually get defined in this
// segment.
EVT OrigEltTy = Entry.getOperand(0).getValueType().getVectorElementType();
int BitsDefined = std::min(OrigEltTy.getScalarSizeInBits(),
VT.getScalarSizeInBits());
int LanesDefined = BitsDefined / BitsPerShuffleLane;
// This source is expected to fill ResMultiplier lanes of the final shuffle,
// starting at the appropriate offset.
int *LaneMask = &Mask[i * ResMultiplier];
int ExtractBase = EltNo * Src->WindowScale + Src->WindowBase;
ExtractBase += NumElts * (Src - Sources.begin());
for (int j = 0; j < LanesDefined; ++j)
LaneMask[j] = ExtractBase + j;
}
// Final check before we try to produce nonsense...
if (!isShuffleMaskLegal(Mask, ShuffleVT)) {
LLVM_DEBUG(dbgs() << "Reshuffle failed: illegal shuffle mask\n");
return SDValue();
}
SDValue ShuffleOps[] = { DAG.getUNDEF(ShuffleVT), DAG.getUNDEF(ShuffleVT) };
for (unsigned i = 0; i < Sources.size(); ++i)
ShuffleOps[i] = Sources[i].ShuffleVec;
SDValue Shuffle =
DAG.getVectorShuffle(ShuffleVT, DL, ShuffleOps[0], ShuffleOps[1], Mask);
SDValue V;
if (DAG.getDataLayout().isBigEndian()) {
V = DAG.getNode(AArch64ISD::NVCAST, DL, VT, Shuffle);
} else {
V = DAG.getNode(ISD::BITCAST, DL, VT, Shuffle);
}
LLVM_DEBUG(dbgs() << "Reshuffle, creating node: "; Shuffle.dump();
dbgs() << "Reshuffle, creating node: "; V.dump(););
return V;
}
// check if an EXT instruction can handle the shuffle mask when the
// vector sources of the shuffle are the same.
static bool isSingletonEXTMask(ArrayRef<int> M, EVT VT, unsigned &Imm) {
unsigned NumElts = VT.getVectorNumElements();
// Assume that the first shuffle index is not UNDEF. Fail if it is.
if (M[0] < 0)
return false;
Imm = M[0];
// If this is a VEXT shuffle, the immediate value is the index of the first
// element. The other shuffle indices must be the successive elements after
// the first one.
unsigned ExpectedElt = Imm;
for (unsigned i = 1; i < NumElts; ++i) {
// Increment the expected index. If it wraps around, just follow it
// back to index zero and keep going.
++ExpectedElt;
if (ExpectedElt == NumElts)
ExpectedElt = 0;
if (M[i] < 0)
continue; // ignore UNDEF indices
if (ExpectedElt != static_cast<unsigned>(M[i]))
return false;
}
return true;
}
// Detect patterns of a0,a1,a2,a3,b0,b1,b2,b3,c0,c1,c2,c3,d0,d1,d2,d3 from
// v4i32s. This is really a truncate, which we can construct out of (legal)
// concats and truncate nodes.
static SDValue ReconstructTruncateFromBuildVector(SDValue V, SelectionDAG &DAG) {
if (V.getValueType() != MVT::v16i8)
return SDValue();
assert(V.getNumOperands() == 16 && "Expected 16 operands on the BUILDVECTOR");
for (unsigned X = 0; X < 4; X++) {
// Check the first item in each group is an extract from lane 0 of a v4i32
// or v4i16.
SDValue BaseExt = V.getOperand(X * 4);
if (BaseExt.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
(BaseExt.getOperand(0).getValueType() != MVT::v4i16 &&
BaseExt.getOperand(0).getValueType() != MVT::v4i32) ||
!isa<ConstantSDNode>(BaseExt.getOperand(1)) ||
BaseExt.getConstantOperandVal(1) != 0)
return SDValue();
SDValue Base = BaseExt.getOperand(0);
// And check the other items are extracts from the same vector.
for (unsigned Y = 1; Y < 4; Y++) {
SDValue Ext = V.getOperand(X * 4 + Y);
if (Ext.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
Ext.getOperand(0) != Base ||
!isa<ConstantSDNode>(Ext.getOperand(1)) ||
Ext.getConstantOperandVal(1) != Y)
return SDValue();
}
}
// Turn the buildvector into a series of truncates and concates, which will
// become uzip1's. Any v4i32s we found get truncated to v4i16, which are
// concat together to produce 2 v8i16. These are both truncated and concat
// together.
SDLoc DL(V);
SDValue Trunc[4] = {
V.getOperand(0).getOperand(0), V.getOperand(4).getOperand(0),
V.getOperand(8).getOperand(0), V.getOperand(12).getOperand(0)};
for (SDValue &V : Trunc)
if (V.getValueType() == MVT::v4i32)
V = DAG.getNode(ISD::TRUNCATE, DL, MVT::v4i16, V);
SDValue Concat0 =
DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v8i16, Trunc[0], Trunc[1]);
SDValue Concat1 =
DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v8i16, Trunc[2], Trunc[3]);
SDValue Trunc0 = DAG.getNode(ISD::TRUNCATE, DL, MVT::v8i8, Concat0);
SDValue Trunc1 = DAG.getNode(ISD::TRUNCATE, DL, MVT::v8i8, Concat1);
return DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v16i8, Trunc0, Trunc1);
}
/// Check if a vector shuffle corresponds to a DUP instructions with a larger
/// element width than the vector lane type. If that is the case the function
/// returns true and writes the value of the DUP instruction lane operand into
/// DupLaneOp
static bool isWideDUPMask(ArrayRef<int> M, EVT VT, unsigned BlockSize,
unsigned &DupLaneOp) {
assert((BlockSize == 16 || BlockSize == 32 || BlockSize == 64) &&
"Only possible block sizes for wide DUP are: 16, 32, 64");
if (BlockSize <= VT.getScalarSizeInBits())
return false;
if (BlockSize % VT.getScalarSizeInBits() != 0)
return false;
if (VT.getSizeInBits() % BlockSize != 0)
return false;
size_t SingleVecNumElements = VT.getVectorNumElements();
size_t NumEltsPerBlock = BlockSize / VT.getScalarSizeInBits();
size_t NumBlocks = VT.getSizeInBits() / BlockSize;
// We are looking for masks like
// [0, 1, 0, 1] or [2, 3, 2, 3] or [4, 5, 6, 7, 4, 5, 6, 7] where any element
// might be replaced by 'undefined'. BlockIndices will eventually contain
// lane indices of the duplicated block (i.e. [0, 1], [2, 3] and [4, 5, 6, 7]
// for the above examples)
SmallVector<int, 8> BlockElts(NumEltsPerBlock, -1);
for (size_t BlockIndex = 0; BlockIndex < NumBlocks; BlockIndex++)
for (size_t I = 0; I < NumEltsPerBlock; I++) {
int Elt = M[BlockIndex * NumEltsPerBlock + I];
if (Elt < 0)
continue;
// For now we don't support shuffles that use the second operand
if ((unsigned)Elt >= SingleVecNumElements)
return false;
if (BlockElts[I] < 0)
BlockElts[I] = Elt;
else if (BlockElts[I] != Elt)
return false;
}
// We found a candidate block (possibly with some undefs). It must be a
// sequence of consecutive integers starting with a value divisible by
// NumEltsPerBlock with some values possibly replaced by undef-s.
// Find first non-undef element
auto FirstRealEltIter = find_if(BlockElts, [](int Elt) { return Elt >= 0; });
assert(FirstRealEltIter != BlockElts.end() &&
"Shuffle with all-undefs must have been caught by previous cases, "
"e.g. isSplat()");
if (FirstRealEltIter == BlockElts.end()) {
DupLaneOp = 0;
return true;
}
// Index of FirstRealElt in BlockElts
size_t FirstRealIndex = FirstRealEltIter - BlockElts.begin();
if ((unsigned)*FirstRealEltIter < FirstRealIndex)
return false;
// BlockElts[0] must have the following value if it isn't undef:
size_t Elt0 = *FirstRealEltIter - FirstRealIndex;
// Check the first element
if (Elt0 % NumEltsPerBlock != 0)
return false;
// Check that the sequence indeed consists of consecutive integers (modulo
// undefs)
for (size_t I = 0; I < NumEltsPerBlock; I++)
if (BlockElts[I] >= 0 && (unsigned)BlockElts[I] != Elt0 + I)
return false;
DupLaneOp = Elt0 / NumEltsPerBlock;
return true;
}
// check if an EXT instruction can handle the shuffle mask when the
// vector sources of the shuffle are different.
static bool isEXTMask(ArrayRef<int> M, EVT VT, bool &ReverseEXT,
unsigned &Imm) {
// Look for the first non-undef element.
const int *FirstRealElt = find_if(M, [](int Elt) { return Elt >= 0; });
// Benefit form APInt to handle overflow when calculating expected element.
unsigned NumElts = VT.getVectorNumElements();
unsigned MaskBits = APInt(32, NumElts * 2).logBase2();
APInt ExpectedElt = APInt(MaskBits, *FirstRealElt + 1, /*isSigned=*/false,
/*implicitTrunc=*/true);
// The following shuffle indices must be the successive elements after the
// first real element.
bool FoundWrongElt = std::any_of(FirstRealElt + 1, M.end(), [&](int Elt) {
return Elt != ExpectedElt++ && Elt != -1;
});
if (FoundWrongElt)
return false;
// The index of an EXT is the first element if it is not UNDEF.
// Watch out for the beginning UNDEFs. The EXT index should be the expected
// value of the first element. E.g.
// <-1, -1, 3, ...> is treated as <1, 2, 3, ...>.
// <-1, -1, 0, 1, ...> is treated as <2*NumElts-2, 2*NumElts-1, 0, 1, ...>.
// ExpectedElt is the last mask index plus 1.
Imm = ExpectedElt.getZExtValue();
// There are two difference cases requiring to reverse input vectors.
// For example, for vector <4 x i32> we have the following cases,
// Case 1: shufflevector(<4 x i32>,<4 x i32>,<-1, -1, -1, 0>)
// Case 2: shufflevector(<4 x i32>,<4 x i32>,<-1, -1, 7, 0>)
// For both cases, we finally use mask <5, 6, 7, 0>, which requires
// to reverse two input vectors.
if (Imm < NumElts)
ReverseEXT = true;
else
Imm -= NumElts;
return true;
}
/// isZIP_v_undef_Mask - Special case of isZIPMask for canonical form of
/// "vector_shuffle v, v", i.e., "vector_shuffle v, undef".
/// Mask is e.g., <0, 0, 1, 1> instead of <0, 4, 1, 5>.
static bool isZIP_v_undef_Mask(ArrayRef<int> M, EVT VT, unsigned &WhichResult) {
unsigned NumElts = VT.getVectorNumElements();
if (NumElts % 2 != 0)
return false;
WhichResult = (M[0] == 0 ? 0 : 1);
unsigned Idx = WhichResult * NumElts / 2;
for (unsigned i = 0; i != NumElts; i += 2) {
if ((M[i] >= 0 && (unsigned)M[i] != Idx) ||
(M[i + 1] >= 0 && (unsigned)M[i + 1] != Idx))
return false;
Idx += 1;
}
return true;
}
/// isUZP_v_undef_Mask - Special case of isUZPMask for canonical form of
/// "vector_shuffle v, v", i.e., "vector_shuffle v, undef".
/// Mask is e.g., <0, 2, 0, 2> instead of <0, 2, 4, 6>,
static bool isUZP_v_undef_Mask(ArrayRef<int> M, EVT VT, unsigned &WhichResult) {
unsigned Half = VT.getVectorNumElements() / 2;
WhichResult = (M[0] == 0 ? 0 : 1);
for (unsigned j = 0; j != 2; ++j) {
unsigned Idx = WhichResult;
for (unsigned i = 0; i != Half; ++i) {
int MIdx = M[i + j * Half];
if (MIdx >= 0 && (unsigned)MIdx != Idx)
return false;
Idx += 2;
}
}
return true;
}
/// isTRN_v_undef_Mask - Special case of isTRNMask for canonical form of
/// "vector_shuffle v, v", i.e., "vector_shuffle v, undef".
/// Mask is e.g., <0, 0, 2, 2> instead of <0, 4, 2, 6>.
static bool isTRN_v_undef_Mask(ArrayRef<int> M, EVT VT, unsigned &WhichResult) {
unsigned NumElts = VT.getVectorNumElements();
if (NumElts % 2 != 0)
return false;
WhichResult = (M[0] == 0 ? 0 : 1);
for (unsigned i = 0; i < NumElts; i += 2) {
if ((M[i] >= 0 && (unsigned)M[i] != i + WhichResult) ||
(M[i + 1] >= 0 && (unsigned)M[i + 1] != i + WhichResult))
return false;
}
return true;
}
static bool isINSMask(ArrayRef<int> M, int NumInputElements,
bool &DstIsLeft, int &Anomaly) {
if (M.size() != static_cast<size_t>(NumInputElements))
return false;
int NumLHSMatch = 0, NumRHSMatch = 0;
int LastLHSMismatch = -1, LastRHSMismatch = -1;
for (int i = 0; i < NumInputElements; ++i) {
if (M[i] == -1) {
++NumLHSMatch;
++NumRHSMatch;
continue;
}
if (M[i] == i)
++NumLHSMatch;
else
LastLHSMismatch = i;
if (M[i] == i + NumInputElements)
++NumRHSMatch;
else
LastRHSMismatch = i;
}
if (NumLHSMatch == NumInputElements - 1) {
DstIsLeft = true;
Anomaly = LastLHSMismatch;
return true;
} else if (NumRHSMatch == NumInputElements - 1) {
DstIsLeft = false;
Anomaly = LastRHSMismatch;
return true;
}
return false;
}
static bool isConcatMask(ArrayRef<int> Mask, EVT VT, bool SplitLHS) {
if (VT.getSizeInBits() != 128)
return false;
unsigned NumElts = VT.getVectorNumElements();
for (int I = 0, E = NumElts / 2; I != E; I++) {
if (Mask[I] != I)
return false;
}
int Offset = NumElts / 2;
for (int I = NumElts / 2, E = NumElts; I != E; I++) {
if (Mask[I] != I + SplitLHS * Offset)
return false;
}
return true;
}
static SDValue tryFormConcatFromShuffle(SDValue Op, SelectionDAG &DAG) {
SDLoc DL(Op);
EVT VT = Op.getValueType();
SDValue V0 = Op.getOperand(0);
SDValue V1 = Op.getOperand(1);
ArrayRef<int> Mask = cast<ShuffleVectorSDNode>(Op)->getMask();
if (VT.getVectorElementType() != V0.getValueType().getVectorElementType() ||
VT.getVectorElementType() != V1.getValueType().getVectorElementType())
return SDValue();
bool SplitV0 = V0.getValueSizeInBits() == 128;
if (!isConcatMask(Mask, VT, SplitV0))
return SDValue();
EVT CastVT = VT.getHalfNumVectorElementsVT(*DAG.getContext());
if (SplitV0) {
V0 = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, CastVT, V0,
DAG.getConstant(0, DL, MVT::i64));
}
if (V1.getValueSizeInBits() == 128) {
V1 = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, CastVT, V1,
DAG.getConstant(0, DL, MVT::i64));
}
return DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, V0, V1);
}
/// GeneratePerfectShuffle - Given an entry in the perfect-shuffle table, emit
/// the specified operations to build the shuffle. ID is the perfect-shuffle
//ID, V1 and V2 are the original shuffle inputs. PFEntry is the Perfect shuffle
//table entry and LHS/RHS are the immediate inputs for this stage of the
//shuffle.
static SDValue GeneratePerfectShuffle(unsigned ID, SDValue V1, SDValue V2,
unsigned PFEntry, SDValue LHS,
SDValue RHS, SelectionDAG &DAG,
const SDLoc &DL) {
unsigned OpNum = (PFEntry >> 26) & 0x0F;
unsigned LHSID = (PFEntry >> 13) & ((1 << 13) - 1);
unsigned RHSID = (PFEntry >> 0) & ((1 << 13) - 1);
enum {
OP_COPY = 0, // Copy, used for things like <u,u,u,3> to say it is <0,1,2,3>
OP_VREV,
OP_VDUP0,
OP_VDUP1,
OP_VDUP2,
OP_VDUP3,
OP_VEXT1,
OP_VEXT2,
OP_VEXT3,
OP_VUZPL, // VUZP, left result
OP_VUZPR, // VUZP, right result
OP_VZIPL, // VZIP, left result
OP_VZIPR, // VZIP, right result
OP_VTRNL, // VTRN, left result
OP_VTRNR, // VTRN, right result
OP_MOVLANE // Move lane. RHSID is the lane to move into
};
if (OpNum == OP_COPY) {
if (LHSID == (1 * 9 + 2) * 9 + 3)
return LHS;
assert(LHSID == ((4 * 9 + 5) * 9 + 6) * 9 + 7 && "Illegal OP_COPY!");
return RHS;
}
if (OpNum == OP_MOVLANE) {
// Decompose a PerfectShuffle ID to get the Mask for lane Elt
auto getPFIDLane = [](unsigned ID, int Elt) -> int {
assert(Elt < 4 && "Expected Perfect Lanes to be less than 4");
Elt = 3 - Elt;
while (Elt > 0) {
ID /= 9;
Elt--;
}
return (ID % 9 == 8) ? -1 : ID % 9;
};
// For OP_MOVLANE shuffles, the RHSID represents the lane to move into. We
// get the lane to move from the PFID, which is always from the
// original vectors (V1 or V2).
SDValue OpLHS = GeneratePerfectShuffle(
LHSID, V1, V2, PerfectShuffleTable[LHSID], LHS, RHS, DAG, DL);
EVT VT = OpLHS.getValueType();
assert(RHSID < 8 && "Expected a lane index for RHSID!");
unsigned ExtLane = 0;
SDValue Input;
// OP_MOVLANE are either D movs (if bit 0x4 is set) or S movs. D movs
// convert into a higher type.
if (RHSID & 0x4) {
int MaskElt = getPFIDLane(ID, (RHSID & 0x01) << 1) >> 1;
if (MaskElt == -1)
MaskElt = (getPFIDLane(ID, ((RHSID & 0x01) << 1) + 1) - 1) >> 1;
assert(MaskElt >= 0 && "Didn't expect an undef movlane index!");
ExtLane = MaskElt < 2 ? MaskElt : (MaskElt - 2);
Input = MaskElt < 2 ? V1 : V2;
if (VT.getScalarSizeInBits() == 16) {
Input = DAG.getBitcast(MVT::v2f32, Input);
OpLHS = DAG.getBitcast(MVT::v2f32, OpLHS);
} else {
assert(VT.getScalarSizeInBits() == 32 &&
"Expected 16 or 32 bit shuffle elements");
Input = DAG.getBitcast(MVT::v2f64, Input);
OpLHS = DAG.getBitcast(MVT::v2f64, OpLHS);
}
} else {
int MaskElt = getPFIDLane(ID, RHSID);
assert(MaskElt >= 0 && "Didn't expect an undef movlane index!");
ExtLane = MaskElt < 4 ? MaskElt : (MaskElt - 4);
Input = MaskElt < 4 ? V1 : V2;
// Be careful about creating illegal types. Use f16 instead of i16.
if (VT == MVT::v4i16) {
Input = DAG.getBitcast(MVT::v4f16, Input);
OpLHS = DAG.getBitcast(MVT::v4f16, OpLHS);
}
}
SDValue Ext = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL,
Input.getValueType().getVectorElementType(),
Input, DAG.getVectorIdxConstant(ExtLane, DL));
SDValue Ins =
DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, Input.getValueType(), OpLHS,
Ext, DAG.getVectorIdxConstant(RHSID & 0x3, DL));
return DAG.getBitcast(VT, Ins);
}
SDValue OpLHS, OpRHS;
OpLHS = GeneratePerfectShuffle(LHSID, V1, V2, PerfectShuffleTable[LHSID], LHS,
RHS, DAG, DL);
OpRHS = GeneratePerfectShuffle(RHSID, V1, V2, PerfectShuffleTable[RHSID], LHS,
RHS, DAG, DL);
EVT VT = OpLHS.getValueType();
switch (OpNum) {
default:
llvm_unreachable("Unknown shuffle opcode!");
case OP_VREV:
// VREV divides the vector in half and swaps within the half.
if (VT.getVectorElementType() == MVT::i32 ||
VT.getVectorElementType() == MVT::f32)
return DAG.getNode(AArch64ISD::REV64, DL, VT, OpLHS);
// vrev <4 x i16> -> REV32
if (VT.getVectorElementType() == MVT::i16 ||
VT.getVectorElementType() == MVT::f16 ||
VT.getVectorElementType() == MVT::bf16)
return DAG.getNode(AArch64ISD::REV32, DL, VT, OpLHS);
// vrev <4 x i8> -> REV16
assert(VT.getVectorElementType() == MVT::i8);
return DAG.getNode(AArch64ISD::REV16, DL, VT, OpLHS);
case OP_VDUP0:
case OP_VDUP1:
case OP_VDUP2:
case OP_VDUP3: {
EVT EltTy = VT.getVectorElementType();
unsigned Opcode;
if (EltTy == MVT::i8)
Opcode = AArch64ISD::DUPLANE8;
else if (EltTy == MVT::i16 || EltTy == MVT::f16 || EltTy == MVT::bf16)
Opcode = AArch64ISD::DUPLANE16;
else if (EltTy == MVT::i32 || EltTy == MVT::f32)
Opcode = AArch64ISD::DUPLANE32;
else if (EltTy == MVT::i64 || EltTy == MVT::f64)
Opcode = AArch64ISD::DUPLANE64;
else
llvm_unreachable("Invalid vector element type?");
if (VT.getSizeInBits() == 64)
OpLHS = WidenVector(OpLHS, DAG);
SDValue Lane = DAG.getConstant(OpNum - OP_VDUP0, DL, MVT::i64);
return DAG.getNode(Opcode, DL, VT, OpLHS, Lane);
}
case OP_VEXT1:
case OP_VEXT2:
case OP_VEXT3: {
unsigned Imm = (OpNum - OP_VEXT1 + 1) * getExtFactor(OpLHS);
return DAG.getNode(AArch64ISD::EXT, DL, VT, OpLHS, OpRHS,
DAG.getConstant(Imm, DL, MVT::i32));
}
case OP_VUZPL:
return DAG.getNode(AArch64ISD::UZP1, DL, VT, OpLHS, OpRHS);
case OP_VUZPR:
return DAG.getNode(AArch64ISD::UZP2, DL, VT, OpLHS, OpRHS);
case OP_VZIPL:
return DAG.getNode(AArch64ISD::ZIP1, DL, VT, OpLHS, OpRHS);
case OP_VZIPR:
return DAG.getNode(AArch64ISD::ZIP2, DL, VT, OpLHS, OpRHS);
case OP_VTRNL:
return DAG.getNode(AArch64ISD::TRN1, DL, VT, OpLHS, OpRHS);
case OP_VTRNR:
return DAG.getNode(AArch64ISD::TRN2, DL, VT, OpLHS, OpRHS);
}
}
static SDValue GenerateTBL(SDValue Op, ArrayRef<int> ShuffleMask,
SelectionDAG &DAG) {
// Check to see if we can use the TBL instruction.
SDValue V1 = Op.getOperand(0);
SDValue V2 = Op.getOperand(1);
SDLoc DL(Op);
EVT EltVT = Op.getValueType().getVectorElementType();
unsigned BytesPerElt = EltVT.getSizeInBits() / 8;
bool Swap = false;
if (V1.isUndef() || isZerosVector(V1.getNode())) {
std::swap(V1, V2);
Swap = true;
}
// If the V2 source is undef or zero then we can use a tbl1, as tbl1 will fill
// out of range values with 0s. We do need to make sure that any out-of-range
// values are really out-of-range for a v16i8 vector.
bool IsUndefOrZero = V2.isUndef() || isZerosVector(V2.getNode());
MVT IndexVT = MVT::v8i8;
unsigned IndexLen = 8;
if (Op.getValueSizeInBits() == 128) {
IndexVT = MVT::v16i8;
IndexLen = 16;
}
SmallVector<SDValue, 8> TBLMask;
for (int Val : ShuffleMask) {
for (unsigned Byte = 0; Byte < BytesPerElt; ++Byte) {
unsigned Offset = Byte + Val * BytesPerElt;
if (Swap)
Offset = Offset < IndexLen ? Offset + IndexLen : Offset - IndexLen;
if (IsUndefOrZero && Offset >= IndexLen)
Offset = 255;
TBLMask.push_back(DAG.getConstant(Offset, DL, MVT::i32));
}
}
SDValue V1Cst = DAG.getNode(ISD::BITCAST, DL, IndexVT, V1);
SDValue V2Cst = DAG.getNode(ISD::BITCAST, DL, IndexVT, V2);
SDValue Shuffle;
if (IsUndefOrZero) {
if (IndexLen == 8)
V1Cst = DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v16i8, V1Cst, V1Cst);
Shuffle = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, IndexVT,
DAG.getConstant(Intrinsic::aarch64_neon_tbl1, DL, MVT::i32), V1Cst,
DAG.getBuildVector(IndexVT, DL, ArrayRef(TBLMask.data(), IndexLen)));
} else {
if (IndexLen == 8) {
V1Cst = DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v16i8, V1Cst, V2Cst);
Shuffle = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, IndexVT,
DAG.getConstant(Intrinsic::aarch64_neon_tbl1, DL, MVT::i32), V1Cst,
DAG.getBuildVector(IndexVT, DL, ArrayRef(TBLMask.data(), IndexLen)));
} else {
// FIXME: We cannot, for the moment, emit a TBL2 instruction because we
// cannot currently represent the register constraints on the input
// table registers.
// Shuffle = DAG.getNode(AArch64ISD::TBL2, DL, IndexVT, V1Cst, V2Cst,
// DAG.getBuildVector(IndexVT, DL, &TBLMask[0],
// IndexLen));
Shuffle = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, IndexVT,
DAG.getConstant(Intrinsic::aarch64_neon_tbl2, DL, MVT::i32), V1Cst,
V2Cst,
DAG.getBuildVector(IndexVT, DL, ArrayRef(TBLMask.data(), IndexLen)));
}
}
return DAG.getNode(ISD::BITCAST, DL, Op.getValueType(), Shuffle);
}
static unsigned getDUPLANEOp(EVT EltType) {
if (EltType == MVT::i8)
return AArch64ISD::DUPLANE8;
if (EltType == MVT::i16 || EltType == MVT::f16 || EltType == MVT::bf16)
return AArch64ISD::DUPLANE16;
if (EltType == MVT::i32 || EltType == MVT::f32)
return AArch64ISD::DUPLANE32;
if (EltType == MVT::i64 || EltType == MVT::f64)
return AArch64ISD::DUPLANE64;
llvm_unreachable("Invalid vector element type?");
}
static SDValue constructDup(SDValue V, int Lane, SDLoc DL, EVT VT,
unsigned Opcode, SelectionDAG &DAG) {
// Try to eliminate a bitcasted extract subvector before a DUPLANE.
auto getScaledOffsetDup = [](SDValue BitCast, int &LaneC, MVT &CastVT) {
// Match: dup (bitcast (extract_subv X, C)), LaneC
if (BitCast.getOpcode() != ISD::BITCAST ||
BitCast.getOperand(0).getOpcode() != ISD::EXTRACT_SUBVECTOR)
return false;
// The extract index must align in the destination type. That may not
// happen if the bitcast is from narrow to wide type.
SDValue Extract = BitCast.getOperand(0);
unsigned ExtIdx = Extract.getConstantOperandVal(1);
unsigned SrcEltBitWidth = Extract.getScalarValueSizeInBits();
unsigned ExtIdxInBits = ExtIdx * SrcEltBitWidth;
unsigned CastedEltBitWidth = BitCast.getScalarValueSizeInBits();
if (ExtIdxInBits % CastedEltBitWidth != 0)
return false;
// Can't handle cases where vector size is not 128-bit
if (!Extract.getOperand(0).getValueType().is128BitVector())
return false;
// Update the lane value by offsetting with the scaled extract index.
LaneC += ExtIdxInBits / CastedEltBitWidth;
// Determine the casted vector type of the wide vector input.
// dup (bitcast (extract_subv X, C)), LaneC --> dup (bitcast X), LaneC'
// Examples:
// dup (bitcast (extract_subv v2f64 X, 1) to v2f32), 1 --> dup v4f32 X, 3
// dup (bitcast (extract_subv v16i8 X, 8) to v4i16), 1 --> dup v8i16 X, 5
unsigned SrcVecNumElts =
Extract.getOperand(0).getValueSizeInBits() / CastedEltBitWidth;
CastVT = MVT::getVectorVT(BitCast.getSimpleValueType().getScalarType(),
SrcVecNumElts);
return true;
};
MVT CastVT;
if (getScaledOffsetDup(V, Lane, CastVT)) {
V = DAG.getBitcast(CastVT, V.getOperand(0).getOperand(0));
} else if (V.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
V.getOperand(0).getValueType().is128BitVector()) {
// The lane is incremented by the index of the extract.
// Example: dup v2f32 (extract v4f32 X, 2), 1 --> dup v4f32 X, 3
Lane += V.getConstantOperandVal(1);
V = V.getOperand(0);
} else if (V.getOpcode() == ISD::CONCAT_VECTORS) {
// The lane is decremented if we are splatting from the 2nd operand.
// Example: dup v4i32 (concat v2i32 X, v2i32 Y), 3 --> dup v4i32 Y, 1
unsigned Idx = Lane >= (int)VT.getVectorNumElements() / 2;
Lane -= Idx * VT.getVectorNumElements() / 2;
V = WidenVector(V.getOperand(Idx), DAG);
} else if (VT.getSizeInBits() == 64) {
// Widen the operand to 128-bit register with undef.
V = WidenVector(V, DAG);
}
return DAG.getNode(Opcode, DL, VT, V, DAG.getConstant(Lane, DL, MVT::i64));
}
// Try to widen element type to get a new mask value for a better permutation
// sequence, so that we can use NEON shuffle instructions, such as zip1/2,
// UZP1/2, TRN1/2, REV, INS, etc.
// For example:
// shufflevector <4 x i32> %a, <4 x i32> %b,
// <4 x i32> <i32 6, i32 7, i32 2, i32 3>
// is equivalent to:
// shufflevector <2 x i64> %a, <2 x i64> %b, <2 x i32> <i32 3, i32 1>
// Finally, we can get:
// mov v0.d[0], v1.d[1]
static SDValue tryWidenMaskForShuffle(SDValue Op, SelectionDAG &DAG) {
SDLoc DL(Op);
EVT VT = Op.getValueType();
EVT ScalarVT = VT.getVectorElementType();
unsigned ElementSize = ScalarVT.getFixedSizeInBits();
SDValue V0 = Op.getOperand(0);
SDValue V1 = Op.getOperand(1);
ArrayRef<int> Mask = cast<ShuffleVectorSDNode>(Op)->getMask();
// If combining adjacent elements, like two i16's -> i32, two i32's -> i64 ...
// We need to make sure the wider element type is legal. Thus, ElementSize
// should be not larger than 32 bits, and i1 type should also be excluded.
if (ElementSize > 32 || ElementSize == 1)
return SDValue();
SmallVector<int, 8> NewMask;
if (widenShuffleMaskElts(Mask, NewMask)) {
MVT NewEltVT = VT.isFloatingPoint()
? MVT::getFloatingPointVT(ElementSize * 2)
: MVT::getIntegerVT(ElementSize * 2);
MVT NewVT = MVT::getVectorVT(NewEltVT, VT.getVectorNumElements() / 2);
if (DAG.getTargetLoweringInfo().isTypeLegal(NewVT)) {
V0 = DAG.getBitcast(NewVT, V0);
V1 = DAG.getBitcast(NewVT, V1);
return DAG.getBitcast(VT,
DAG.getVectorShuffle(NewVT, DL, V0, V1, NewMask));
}
}
return SDValue();
}
// Try to fold shuffle (tbl2, tbl2) into a single tbl4.
static SDValue tryToConvertShuffleOfTbl2ToTbl4(SDValue Op,
ArrayRef<int> ShuffleMask,
SelectionDAG &DAG) {
SDValue Tbl1 = Op->getOperand(0);
SDValue Tbl2 = Op->getOperand(1);
SDLoc DL(Op);
SDValue Tbl2ID =
DAG.getTargetConstant(Intrinsic::aarch64_neon_tbl2, DL, MVT::i64);
EVT VT = Op.getValueType();
if (Tbl1.getOpcode() != ISD::INTRINSIC_WO_CHAIN ||
Tbl1.getOperand(0) != Tbl2ID ||
Tbl2.getOpcode() != ISD::INTRINSIC_WO_CHAIN ||
Tbl2.getOperand(0) != Tbl2ID)
return SDValue();
if (Tbl1.getValueType() != MVT::v16i8 || Tbl2.getValueType() != MVT::v16i8)
return SDValue();
SDValue Mask1 = Tbl1.getOperand(3);
SDValue Mask2 = Tbl2.getOperand(3);
if (Mask1.getOpcode() != ISD::BUILD_VECTOR ||
Mask2.getOpcode() != ISD::BUILD_VECTOR)
return SDValue();
SmallVector<SDValue, 16> TBLMaskParts(16, SDValue());
for (unsigned I = 0; I < 16; I++) {
if (ShuffleMask[I] < 16)
TBLMaskParts[I] = Mask1.getOperand(ShuffleMask[I]);
else {
auto *C = dyn_cast<ConstantSDNode>(Mask2.getOperand(ShuffleMask[I] - 16));
if (!C)
return SDValue();
TBLMaskParts[I] = DAG.getConstant(C->getSExtValue() + 32, DL, MVT::i32);
}
}
SDValue TBLMask = DAG.getBuildVector(VT, DL, TBLMaskParts);
SDValue ID =
DAG.getTargetConstant(Intrinsic::aarch64_neon_tbl4, DL, MVT::i64);
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, MVT::v16i8,
{ID, Tbl1->getOperand(1), Tbl1->getOperand(2),
Tbl2->getOperand(1), Tbl2->getOperand(2), TBLMask});
}
// Baseline legalization for ZERO_EXTEND_VECTOR_INREG will blend-in zeros,
// but we don't have an appropriate instruction,
// so custom-lower it as ZIP1-with-zeros.
SDValue
AArch64TargetLowering::LowerZERO_EXTEND_VECTOR_INREG(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT VT = Op.getValueType();
SDValue SrcOp = Op.getOperand(0);
EVT SrcVT = SrcOp.getValueType();
assert(VT.getScalarSizeInBits() % SrcVT.getScalarSizeInBits() == 0 &&
"Unexpected extension factor.");
unsigned Scale = VT.getScalarSizeInBits() / SrcVT.getScalarSizeInBits();
// FIXME: support multi-step zipping?
if (Scale != 2)
return SDValue();
SDValue Zeros = DAG.getConstant(0, DL, SrcVT);
return DAG.getBitcast(VT,
DAG.getNode(AArch64ISD::ZIP1, DL, SrcVT, SrcOp, Zeros));
}
SDValue AArch64TargetLowering::LowerVECTOR_SHUFFLE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT VT = Op.getValueType();
ShuffleVectorSDNode *SVN = cast<ShuffleVectorSDNode>(Op.getNode());
if (useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()))
return LowerFixedLengthVECTOR_SHUFFLEToSVE(Op, DAG);
// Convert shuffles that are directly supported on NEON to target-specific
// DAG nodes, instead of keeping them as shuffles and matching them again
// during code selection. This is more efficient and avoids the possibility
// of inconsistencies between legalization and selection.
ArrayRef<int> ShuffleMask = SVN->getMask();
SDValue V1 = Op.getOperand(0);
SDValue V2 = Op.getOperand(1);
assert(V1.getValueType() == VT && "Unexpected VECTOR_SHUFFLE type!");
assert(ShuffleMask.size() == VT.getVectorNumElements() &&
"Unexpected VECTOR_SHUFFLE mask size!");
if (SDValue Res = tryToConvertShuffleOfTbl2ToTbl4(Op, ShuffleMask, DAG))
return Res;
if (SVN->isSplat()) {
int Lane = SVN->getSplatIndex();
// If this is undef splat, generate it via "just" vdup, if possible.
if (Lane == -1)
Lane = 0;
if (Lane == 0 && V1.getOpcode() == ISD::SCALAR_TO_VECTOR)
return DAG.getNode(AArch64ISD::DUP, DL, V1.getValueType(),
V1.getOperand(0));
// Test if V1 is a BUILD_VECTOR and the lane being referenced is a non-
// constant. If so, we can just reference the lane's definition directly.
if (V1.getOpcode() == ISD::BUILD_VECTOR &&
!isa<ConstantSDNode>(V1.getOperand(Lane)))
return DAG.getNode(AArch64ISD::DUP, DL, VT, V1.getOperand(Lane));
// Otherwise, duplicate from the lane of the input vector.
unsigned Opcode = getDUPLANEOp(V1.getValueType().getVectorElementType());
return constructDup(V1, Lane, DL, VT, Opcode, DAG);
}
// Check if the mask matches a DUP for a wider element
for (unsigned LaneSize : {64U, 32U, 16U}) {
unsigned Lane = 0;
if (isWideDUPMask(ShuffleMask, VT, LaneSize, Lane)) {
unsigned Opcode = LaneSize == 64 ? AArch64ISD::DUPLANE64
: LaneSize == 32 ? AArch64ISD::DUPLANE32
: AArch64ISD::DUPLANE16;
// Cast V1 to an integer vector with required lane size
MVT NewEltTy = MVT::getIntegerVT(LaneSize);
unsigned NewEltCount = VT.getSizeInBits() / LaneSize;
MVT NewVecTy = MVT::getVectorVT(NewEltTy, NewEltCount);
V1 = DAG.getBitcast(NewVecTy, V1);
// Construct the DUP instruction
V1 = constructDup(V1, Lane, DL, NewVecTy, Opcode, DAG);
// Cast back to the original type
return DAG.getBitcast(VT, V1);
}
}
unsigned NumElts = VT.getVectorNumElements();
unsigned EltSize = VT.getScalarSizeInBits();
if (isREVMask(ShuffleMask, EltSize, NumElts, 64))
return DAG.getNode(AArch64ISD::REV64, DL, V1.getValueType(), V1);
if (isREVMask(ShuffleMask, EltSize, NumElts, 32))
return DAG.getNode(AArch64ISD::REV32, DL, V1.getValueType(), V1);
if (isREVMask(ShuffleMask, EltSize, NumElts, 16))
return DAG.getNode(AArch64ISD::REV16, DL, V1.getValueType(), V1);
if (((NumElts == 8 && EltSize == 16) || (NumElts == 16 && EltSize == 8)) &&
ShuffleVectorInst::isReverseMask(ShuffleMask, ShuffleMask.size())) {
SDValue Rev = DAG.getNode(AArch64ISD::REV64, DL, VT, V1);
return DAG.getNode(AArch64ISD::EXT, DL, VT, Rev, Rev,
DAG.getConstant(8, DL, MVT::i32));
}
bool ReverseEXT = false;
unsigned Imm;
if (isEXTMask(ShuffleMask, VT, ReverseEXT, Imm)) {
if (ReverseEXT)
std::swap(V1, V2);
Imm *= getExtFactor(V1);
return DAG.getNode(AArch64ISD::EXT, DL, V1.getValueType(), V1, V2,
DAG.getConstant(Imm, DL, MVT::i32));
} else if (V2->isUndef() && isSingletonEXTMask(ShuffleMask, VT, Imm)) {
Imm *= getExtFactor(V1);
return DAG.getNode(AArch64ISD::EXT, DL, V1.getValueType(), V1, V1,
DAG.getConstant(Imm, DL, MVT::i32));
}
unsigned WhichResult;
if (isZIPMask(ShuffleMask, NumElts, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::ZIP1 : AArch64ISD::ZIP2;
return DAG.getNode(Opc, DL, V1.getValueType(), V1, V2);
}
if (isUZPMask(ShuffleMask, NumElts, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::UZP1 : AArch64ISD::UZP2;
return DAG.getNode(Opc, DL, V1.getValueType(), V1, V2);
}
if (isTRNMask(ShuffleMask, NumElts, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::TRN1 : AArch64ISD::TRN2;
return DAG.getNode(Opc, DL, V1.getValueType(), V1, V2);
}
if (isZIP_v_undef_Mask(ShuffleMask, VT, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::ZIP1 : AArch64ISD::ZIP2;
return DAG.getNode(Opc, DL, V1.getValueType(), V1, V1);
}
if (isUZP_v_undef_Mask(ShuffleMask, VT, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::UZP1 : AArch64ISD::UZP2;
return DAG.getNode(Opc, DL, V1.getValueType(), V1, V1);
}
if (isTRN_v_undef_Mask(ShuffleMask, VT, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::TRN1 : AArch64ISD::TRN2;
return DAG.getNode(Opc, DL, V1.getValueType(), V1, V1);
}
if (SDValue Concat = tryFormConcatFromShuffle(Op, DAG))
return Concat;
bool DstIsLeft;
int Anomaly;
int NumInputElements = V1.getValueType().getVectorNumElements();
if (isINSMask(ShuffleMask, NumInputElements, DstIsLeft, Anomaly)) {
SDValue DstVec = DstIsLeft ? V1 : V2;
SDValue DstLaneV = DAG.getConstant(Anomaly, DL, MVT::i64);
SDValue SrcVec = V1;
int SrcLane = ShuffleMask[Anomaly];
if (SrcLane >= NumInputElements) {
SrcVec = V2;
SrcLane -= NumElts;
}
SDValue SrcLaneV = DAG.getConstant(SrcLane, DL, MVT::i64);
EVT ScalarVT = VT.getVectorElementType();
if (ScalarVT.getFixedSizeInBits() < 32 && ScalarVT.isInteger())
ScalarVT = MVT::i32;
return DAG.getNode(
ISD::INSERT_VECTOR_ELT, DL, VT, DstVec,
DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ScalarVT, SrcVec, SrcLaneV),
DstLaneV);
}
if (SDValue NewSD = tryWidenMaskForShuffle(Op, DAG))
return NewSD;
// If the shuffle is not directly supported and it has 4 elements, use
// the PerfectShuffle-generated table to synthesize it from other shuffles.
if (NumElts == 4) {
unsigned PFIndexes[4];
for (unsigned i = 0; i != 4; ++i) {
if (ShuffleMask[i] < 0)
PFIndexes[i] = 8;
else
PFIndexes[i] = ShuffleMask[i];
}
// Compute the index in the perfect shuffle table.
unsigned PFTableIndex = PFIndexes[0] * 9 * 9 * 9 + PFIndexes[1] * 9 * 9 +
PFIndexes[2] * 9 + PFIndexes[3];
unsigned PFEntry = PerfectShuffleTable[PFTableIndex];
return GeneratePerfectShuffle(PFTableIndex, V1, V2, PFEntry, V1, V2, DAG,
DL);
}
// Check for a "select shuffle", generating a BSL to pick between lanes in
// V1/V2.
if (ShuffleVectorInst::isSelectMask(ShuffleMask, NumElts)) {
assert(VT.getScalarSizeInBits() <= 32 &&
"Expected larger vector element sizes to be handled already");
SmallVector<SDValue> MaskElts;
for (int M : ShuffleMask)
MaskElts.push_back(DAG.getConstant(
M >= static_cast<int>(NumElts) ? 0 : 0xffffffff, DL, MVT::i32));
EVT IVT = VT.changeVectorElementTypeToInteger();
SDValue MaskConst = DAG.getBuildVector(IVT, DL, MaskElts);
return DAG.getBitcast(VT, DAG.getNode(AArch64ISD::BSP, DL, IVT, MaskConst,
DAG.getBitcast(IVT, V1),
DAG.getBitcast(IVT, V2)));
}
// Fall back to generating a TBL
return GenerateTBL(Op, ShuffleMask, DAG);
}
SDValue AArch64TargetLowering::LowerSPLAT_VECTOR(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
if (useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()))
return LowerToScalableOp(Op, DAG);
assert(VT.isScalableVector() && VT.getVectorElementType() == MVT::i1 &&
"Unexpected vector type!");
// We can handle the constant cases during isel.
if (isa<ConstantSDNode>(Op.getOperand(0)))
return Op;
// There isn't a natural way to handle the general i1 case, so we use some
// trickery with whilelo.
SDLoc DL(Op);
SDValue SplatVal = DAG.getAnyExtOrTrunc(Op.getOperand(0), DL, MVT::i64);
SplatVal = DAG.getNode(ISD::SIGN_EXTEND_INREG, DL, MVT::i64, SplatVal,
DAG.getValueType(MVT::i1));
SDValue ID =
DAG.getTargetConstant(Intrinsic::aarch64_sve_whilelo, DL, MVT::i64);
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
if (VT == MVT::nxv1i1)
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::nxv1i1,
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, MVT::nxv2i1, ID,
Zero, SplatVal),
Zero);
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, VT, ID, Zero, SplatVal);
}
SDValue AArch64TargetLowering::LowerDUPQLane(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT VT = Op.getValueType();
if (!isTypeLegal(VT) || !VT.isScalableVector())
return SDValue();
// Current lowering only supports the SVE-ACLE types.
if (VT.getSizeInBits().getKnownMinValue() != AArch64::SVEBitsPerBlock)
return SDValue();
// The DUPQ operation is independent of element type so normalise to i64s.
SDValue Idx128 = Op.getOperand(2);
// DUPQ can be used when idx is in range.
auto *CIdx = dyn_cast<ConstantSDNode>(Idx128);
if (CIdx && (CIdx->getZExtValue() <= 3)) {
SDValue CI = DAG.getTargetConstant(CIdx->getZExtValue(), DL, MVT::i64);
return DAG.getNode(AArch64ISD::DUPLANE128, DL, VT, Op.getOperand(1), CI);
}
SDValue V = DAG.getNode(ISD::BITCAST, DL, MVT::nxv2i64, Op.getOperand(1));
// The ACLE says this must produce the same result as:
// svtbl(data, svadd_x(svptrue_b64(),
// svand_x(svptrue_b64(), svindex_u64(0, 1), 1),
// index * 2))
SDValue One = DAG.getConstant(1, DL, MVT::i64);
SDValue SplatOne = DAG.getNode(ISD::SPLAT_VECTOR, DL, MVT::nxv2i64, One);
// create the vector 0,1,0,1,...
SDValue SV = DAG.getStepVector(DL, MVT::nxv2i64);
SV = DAG.getNode(ISD::AND, DL, MVT::nxv2i64, SV, SplatOne);
// create the vector idx64,idx64+1,idx64,idx64+1,...
SDValue Idx64 = DAG.getNode(ISD::ADD, DL, MVT::i64, Idx128, Idx128);
SDValue SplatIdx64 = DAG.getNode(ISD::SPLAT_VECTOR, DL, MVT::nxv2i64, Idx64);
SDValue ShuffleMask = DAG.getNode(ISD::ADD, DL, MVT::nxv2i64, SV, SplatIdx64);
// create the vector Val[idx64],Val[idx64+1],Val[idx64],Val[idx64+1],...
SDValue TBL = DAG.getNode(AArch64ISD::TBL, DL, MVT::nxv2i64, V, ShuffleMask);
return DAG.getNode(ISD::BITCAST, DL, VT, TBL);
}
static bool resolveBuildVector(BuildVectorSDNode *BVN, APInt &CnstBits,
APInt &UndefBits) {
EVT VT = BVN->getValueType(0);
APInt SplatBits, SplatUndef;
unsigned SplatBitSize;
bool HasAnyUndefs;
if (BVN->isConstantSplat(SplatBits, SplatUndef, SplatBitSize, HasAnyUndefs)) {
unsigned NumSplats = VT.getSizeInBits() / SplatBitSize;
for (unsigned i = 0; i < NumSplats; ++i) {
CnstBits <<= SplatBitSize;
UndefBits <<= SplatBitSize;
CnstBits |= SplatBits.zextOrTrunc(VT.getSizeInBits());
UndefBits |= (SplatBits ^ SplatUndef).zextOrTrunc(VT.getSizeInBits());
}
return true;
}
return false;
}
// Try 64-bit splatted SIMD immediate.
static SDValue tryAdvSIMDModImm64(unsigned NewOp, SDValue Op, SelectionDAG &DAG,
const APInt &Bits) {
if (Bits.getHiBits(64) == Bits.getLoBits(64)) {
uint64_t Value = Bits.zextOrTrunc(64).getZExtValue();
EVT VT = Op.getValueType();
MVT MovTy = (VT.getSizeInBits() == 128) ? MVT::v2i64 : MVT::f64;
if (AArch64_AM::isAdvSIMDModImmType10(Value)) {
Value = AArch64_AM::encodeAdvSIMDModImmType10(Value);
SDLoc DL(Op);
SDValue Mov =
DAG.getNode(NewOp, DL, MovTy, DAG.getConstant(Value, DL, MVT::i32));
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, Mov);
}
}
return SDValue();
}
// Try 32-bit splatted SIMD immediate.
static SDValue tryAdvSIMDModImm32(unsigned NewOp, SDValue Op, SelectionDAG &DAG,
const APInt &Bits,
const SDValue *LHS = nullptr) {
EVT VT = Op.getValueType();
if (VT.isFixedLengthVector() &&
!DAG.getSubtarget<AArch64Subtarget>().isNeonAvailable())
return SDValue();
if (Bits.getHiBits(64) == Bits.getLoBits(64)) {
uint64_t Value = Bits.zextOrTrunc(64).getZExtValue();
MVT MovTy = (VT.getSizeInBits() == 128) ? MVT::v4i32 : MVT::v2i32;
bool isAdvSIMDModImm = false;
uint64_t Shift;
if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType1(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType1(Value);
Shift = 0;
}
else if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType2(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType2(Value);
Shift = 8;
}
else if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType3(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType3(Value);
Shift = 16;
}
else if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType4(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType4(Value);
Shift = 24;
}
if (isAdvSIMDModImm) {
SDLoc DL(Op);
SDValue Mov;
if (LHS)
Mov = DAG.getNode(NewOp, DL, MovTy,
DAG.getNode(AArch64ISD::NVCAST, DL, MovTy, *LHS),
DAG.getConstant(Value, DL, MVT::i32),
DAG.getConstant(Shift, DL, MVT::i32));
else
Mov =
DAG.getNode(NewOp, DL, MovTy, DAG.getConstant(Value, DL, MVT::i32),
DAG.getConstant(Shift, DL, MVT::i32));
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, Mov);
}
}
return SDValue();
}
// Try 16-bit splatted SIMD immediate.
static SDValue tryAdvSIMDModImm16(unsigned NewOp, SDValue Op, SelectionDAG &DAG,
const APInt &Bits,
const SDValue *LHS = nullptr) {
EVT VT = Op.getValueType();
if (VT.isFixedLengthVector() &&
!DAG.getSubtarget<AArch64Subtarget>().isNeonAvailable())
return SDValue();
if (Bits.getHiBits(64) == Bits.getLoBits(64)) {
uint64_t Value = Bits.zextOrTrunc(64).getZExtValue();
MVT MovTy = (VT.getSizeInBits() == 128) ? MVT::v8i16 : MVT::v4i16;
bool isAdvSIMDModImm = false;
uint64_t Shift;
if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType5(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType5(Value);
Shift = 0;
}
else if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType6(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType6(Value);
Shift = 8;
}
if (isAdvSIMDModImm) {
SDLoc DL(Op);
SDValue Mov;
if (LHS)
Mov = DAG.getNode(NewOp, DL, MovTy,
DAG.getNode(AArch64ISD::NVCAST, DL, MovTy, *LHS),
DAG.getConstant(Value, DL, MVT::i32),
DAG.getConstant(Shift, DL, MVT::i32));
else
Mov =
DAG.getNode(NewOp, DL, MovTy, DAG.getConstant(Value, DL, MVT::i32),
DAG.getConstant(Shift, DL, MVT::i32));
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, Mov);
}
}
return SDValue();
}
// Try 32-bit splatted SIMD immediate with shifted ones.
static SDValue tryAdvSIMDModImm321s(unsigned NewOp, SDValue Op,
SelectionDAG &DAG, const APInt &Bits) {
if (Bits.getHiBits(64) == Bits.getLoBits(64)) {
uint64_t Value = Bits.zextOrTrunc(64).getZExtValue();
EVT VT = Op.getValueType();
MVT MovTy = (VT.getSizeInBits() == 128) ? MVT::v4i32 : MVT::v2i32;
bool isAdvSIMDModImm = false;
uint64_t Shift;
if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType7(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType7(Value);
Shift = 264;
}
else if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType8(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType8(Value);
Shift = 272;
}
if (isAdvSIMDModImm) {
SDLoc DL(Op);
SDValue Mov =
DAG.getNode(NewOp, DL, MovTy, DAG.getConstant(Value, DL, MVT::i32),
DAG.getConstant(Shift, DL, MVT::i32));
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, Mov);
}
}
return SDValue();
}
// Try 8-bit splatted SIMD immediate.
static SDValue tryAdvSIMDModImm8(unsigned NewOp, SDValue Op, SelectionDAG &DAG,
const APInt &Bits) {
if (Bits.getHiBits(64) == Bits.getLoBits(64)) {
uint64_t Value = Bits.zextOrTrunc(64).getZExtValue();
EVT VT = Op.getValueType();
MVT MovTy = (VT.getSizeInBits() == 128) ? MVT::v16i8 : MVT::v8i8;
if (AArch64_AM::isAdvSIMDModImmType9(Value)) {
Value = AArch64_AM::encodeAdvSIMDModImmType9(Value);
SDLoc DL(Op);
SDValue Mov =
DAG.getNode(NewOp, DL, MovTy, DAG.getConstant(Value, DL, MVT::i32));
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, Mov);
}
}
return SDValue();
}
// Try FP splatted SIMD immediate.
static SDValue tryAdvSIMDModImmFP(unsigned NewOp, SDValue Op, SelectionDAG &DAG,
const APInt &Bits) {
if (Bits.getHiBits(64) == Bits.getLoBits(64)) {
uint64_t Value = Bits.zextOrTrunc(64).getZExtValue();
EVT VT = Op.getValueType();
bool isWide = (VT.getSizeInBits() == 128);
MVT MovTy;
bool isAdvSIMDModImm = false;
if ((isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType11(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType11(Value);
MovTy = isWide ? MVT::v4f32 : MVT::v2f32;
}
else if (isWide &&
(isAdvSIMDModImm = AArch64_AM::isAdvSIMDModImmType12(Value))) {
Value = AArch64_AM::encodeAdvSIMDModImmType12(Value);
MovTy = MVT::v2f64;
}
if (isAdvSIMDModImm) {
SDLoc DL(Op);
SDValue Mov =
DAG.getNode(NewOp, DL, MovTy, DAG.getConstant(Value, DL, MVT::i32));
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, Mov);
}
}
return SDValue();
}
// Specialized code to quickly find if PotentialBVec is a BuildVector that
// consists of only the same constant int value, returned in reference arg
// ConstVal
static bool isAllConstantBuildVector(const SDValue &PotentialBVec,
uint64_t &ConstVal) {
BuildVectorSDNode *Bvec = dyn_cast<BuildVectorSDNode>(PotentialBVec);
if (!Bvec)
return false;
ConstantSDNode *FirstElt = dyn_cast<ConstantSDNode>(Bvec->getOperand(0));
if (!FirstElt)
return false;
EVT VT = Bvec->getValueType(0);
unsigned NumElts = VT.getVectorNumElements();
for (unsigned i = 1; i < NumElts; ++i)
if (dyn_cast<ConstantSDNode>(Bvec->getOperand(i)) != FirstElt)
return false;
ConstVal = FirstElt->getZExtValue();
return true;
}
static bool isAllInactivePredicate(SDValue N) {
// Look through cast.
while (N.getOpcode() == AArch64ISD::REINTERPRET_CAST)
N = N.getOperand(0);
return ISD::isConstantSplatVectorAllZeros(N.getNode());
}
static bool isAllActivePredicate(SelectionDAG &DAG, SDValue N) {
unsigned NumElts = N.getValueType().getVectorMinNumElements();
// Look through cast.
while (N.getOpcode() == AArch64ISD::REINTERPRET_CAST) {
N = N.getOperand(0);
// When reinterpreting from a type with fewer elements the "new" elements
// are not active, so bail if they're likely to be used.
if (N.getValueType().getVectorMinNumElements() < NumElts)
return false;
}
if (ISD::isConstantSplatVectorAllOnes(N.getNode()))
return true;
// "ptrue p.<ty>, all" can be considered all active when <ty> is the same size
// or smaller than the implicit element type represented by N.
// NOTE: A larger element count implies a smaller element type.
if (N.getOpcode() == AArch64ISD::PTRUE &&
N.getConstantOperandVal(0) == AArch64SVEPredPattern::all)
return N.getValueType().getVectorMinNumElements() >= NumElts;
// If we're compiling for a specific vector-length, we can check if the
// pattern's VL equals that of the scalable vector at runtime.
if (N.getOpcode() == AArch64ISD::PTRUE) {
const auto &Subtarget = DAG.getSubtarget<AArch64Subtarget>();
unsigned MinSVESize = Subtarget.getMinSVEVectorSizeInBits();
unsigned MaxSVESize = Subtarget.getMaxSVEVectorSizeInBits();
if (MaxSVESize && MinSVESize == MaxSVESize) {
unsigned VScale = MaxSVESize / AArch64::SVEBitsPerBlock;
unsigned PatNumElts =
getNumElementsFromSVEPredPattern(N.getConstantOperandVal(0));
return PatNumElts == (NumElts * VScale);
}
}
return false;
}
// Attempt to form a vector S[LR]I from (or (and X, BvecC1), (lsl Y, C2)),
// to (SLI X, Y, C2), where X and Y have matching vector types, BvecC1 is a
// BUILD_VECTORs with constant element C1, C2 is a constant, and:
// - for the SLI case: C1 == ~(Ones(ElemSizeInBits) << C2)
// - for the SRI case: C1 == ~(Ones(ElemSizeInBits) >> C2)
// The (or (lsl Y, C2), (and X, BvecC1)) case is also handled.
static SDValue tryLowerToSLI(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (!VT.isVector())
return SDValue();
SDLoc DL(N);
SDValue And;
SDValue Shift;
SDValue FirstOp = N->getOperand(0);
unsigned FirstOpc = FirstOp.getOpcode();
SDValue SecondOp = N->getOperand(1);
unsigned SecondOpc = SecondOp.getOpcode();
// Is one of the operands an AND or a BICi? The AND may have been optimised to
// a BICi in order to use an immediate instead of a register.
// Is the other operand an shl or lshr? This will have been turned into:
// AArch64ISD::VSHL vector, #shift or AArch64ISD::VLSHR vector, #shift
// or (AArch64ISD::SHL_PRED || AArch64ISD::SRL_PRED) mask, vector, #shiftVec.
if ((FirstOpc == ISD::AND || FirstOpc == AArch64ISD::BICi) &&
(SecondOpc == AArch64ISD::VSHL || SecondOpc == AArch64ISD::VLSHR ||
SecondOpc == AArch64ISD::SHL_PRED ||
SecondOpc == AArch64ISD::SRL_PRED)) {
And = FirstOp;
Shift = SecondOp;
} else if ((SecondOpc == ISD::AND || SecondOpc == AArch64ISD::BICi) &&
(FirstOpc == AArch64ISD::VSHL || FirstOpc == AArch64ISD::VLSHR ||
FirstOpc == AArch64ISD::SHL_PRED ||
FirstOpc == AArch64ISD::SRL_PRED)) {
And = SecondOp;
Shift = FirstOp;
} else
return SDValue();
bool IsAnd = And.getOpcode() == ISD::AND;
bool IsShiftRight = Shift.getOpcode() == AArch64ISD::VLSHR ||
Shift.getOpcode() == AArch64ISD::SRL_PRED;
bool ShiftHasPredOp = Shift.getOpcode() == AArch64ISD::SHL_PRED ||
Shift.getOpcode() == AArch64ISD::SRL_PRED;
// Is the shift amount constant and are all lanes active?
uint64_t C2;
if (ShiftHasPredOp) {
if (!isAllActivePredicate(DAG, Shift.getOperand(0)))
return SDValue();
APInt C;
if (!ISD::isConstantSplatVector(Shift.getOperand(2).getNode(), C))
return SDValue();
C2 = C.getZExtValue();
} else if (ConstantSDNode *C2node =
dyn_cast<ConstantSDNode>(Shift.getOperand(1)))
C2 = C2node->getZExtValue();
else
return SDValue();
APInt C1AsAPInt;
unsigned ElemSizeInBits = VT.getScalarSizeInBits();
if (IsAnd) {
// Is the and mask vector all constant?
if (!ISD::isConstantSplatVector(And.getOperand(1).getNode(), C1AsAPInt))
return SDValue();
} else {
// Reconstruct the corresponding AND immediate from the two BICi immediates.
ConstantSDNode *C1nodeImm = dyn_cast<ConstantSDNode>(And.getOperand(1));
ConstantSDNode *C1nodeShift = dyn_cast<ConstantSDNode>(And.getOperand(2));
assert(C1nodeImm && C1nodeShift);
C1AsAPInt = ~(C1nodeImm->getAPIntValue() << C1nodeShift->getAPIntValue());
C1AsAPInt = C1AsAPInt.zextOrTrunc(ElemSizeInBits);
}
// Is C1 == ~(Ones(ElemSizeInBits) << C2) or
// C1 == ~(Ones(ElemSizeInBits) >> C2), taking into account
// how much one can shift elements of a particular size?
if (C2 > ElemSizeInBits)
return SDValue();
APInt RequiredC1 = IsShiftRight ? APInt::getHighBitsSet(ElemSizeInBits, C2)
: APInt::getLowBitsSet(ElemSizeInBits, C2);
if (C1AsAPInt != RequiredC1)
return SDValue();
SDValue X = And.getOperand(0);
SDValue Y = ShiftHasPredOp ? Shift.getOperand(1) : Shift.getOperand(0);
SDValue Imm = ShiftHasPredOp ? DAG.getTargetConstant(C2, DL, MVT::i32)
: Shift.getOperand(1);
unsigned Inst = IsShiftRight ? AArch64ISD::VSRI : AArch64ISD::VSLI;
SDValue ResultSLI = DAG.getNode(Inst, DL, VT, X, Y, Imm);
return ResultSLI;
}
SDValue AArch64TargetLowering::LowerVectorOR(SDValue Op,
SelectionDAG &DAG) const {
if (useSVEForFixedLengthVectorVT(Op.getValueType(),
!Subtarget->isNeonAvailable()))
return LowerToScalableOp(Op, DAG);
// Attempt to form a vector S[LR]I from (or (and X, C1), (lsl Y, C2))
if (SDValue Res = tryLowerToSLI(Op.getNode(), DAG))
return Res;
EVT VT = Op.getValueType();
if (VT.isScalableVector())
return Op;
SDValue LHS = Op.getOperand(0);
BuildVectorSDNode *BVN =
dyn_cast<BuildVectorSDNode>(Op.getOperand(1).getNode());
if (!BVN) {
// OR commutes, so try swapping the operands.
LHS = Op.getOperand(1);
BVN = dyn_cast<BuildVectorSDNode>(Op.getOperand(0).getNode());
}
if (!BVN)
return Op;
APInt DefBits(VT.getSizeInBits(), 0);
APInt UndefBits(VT.getSizeInBits(), 0);
if (resolveBuildVector(BVN, DefBits, UndefBits)) {
SDValue NewOp;
if ((NewOp = tryAdvSIMDModImm32(AArch64ISD::ORRi, Op, DAG,
DefBits, &LHS)) ||
(NewOp = tryAdvSIMDModImm16(AArch64ISD::ORRi, Op, DAG,
DefBits, &LHS)))
return NewOp;
if ((NewOp = tryAdvSIMDModImm32(AArch64ISD::ORRi, Op, DAG,
UndefBits, &LHS)) ||
(NewOp = tryAdvSIMDModImm16(AArch64ISD::ORRi, Op, DAG,
UndefBits, &LHS)))
return NewOp;
}
// We can always fall back to a non-immediate OR.
return Op;
}
// Normalize the operands of BUILD_VECTOR. The value of constant operands will
// be truncated to fit element width.
static SDValue NormalizeBuildVector(SDValue Op,
SelectionDAG &DAG) {
assert(Op.getOpcode() == ISD::BUILD_VECTOR && "Unknown opcode!");
SDLoc DL(Op);
EVT VT = Op.getValueType();
EVT EltTy= VT.getVectorElementType();
if (EltTy.isFloatingPoint() || EltTy.getSizeInBits() > 16)
return Op;
SmallVector<SDValue, 16> Ops;
for (SDValue Lane : Op->ops()) {
// For integer vectors, type legalization would have promoted the
// operands already. Otherwise, if Op is a floating-point splat
// (with operands cast to integers), then the only possibilities
// are constants and UNDEFs.
if (auto *CstLane = dyn_cast<ConstantSDNode>(Lane)) {
Lane = DAG.getConstant(
CstLane->getAPIntValue().trunc(EltTy.getSizeInBits()).getZExtValue(),
DL, MVT::i32);
} else if (Lane.getNode()->isUndef()) {
Lane = DAG.getUNDEF(MVT::i32);
} else {
assert(Lane.getValueType() == MVT::i32 &&
"Unexpected BUILD_VECTOR operand type");
}
Ops.push_back(Lane);
}
return DAG.getBuildVector(VT, DL, Ops);
}
static SDValue ConstantBuildVector(SDValue Op, SelectionDAG &DAG,
const AArch64Subtarget *ST) {
EVT VT = Op.getValueType();
assert((VT.getSizeInBits() == 64 || VT.getSizeInBits() == 128) &&
"Expected a legal NEON vector");
APInt DefBits(VT.getSizeInBits(), 0);
APInt UndefBits(VT.getSizeInBits(), 0);
BuildVectorSDNode *BVN = cast<BuildVectorSDNode>(Op.getNode());
if (resolveBuildVector(BVN, DefBits, UndefBits)) {
auto TryMOVIWithBits = [&](APInt DefBits) {
SDValue NewOp;
if ((NewOp =
tryAdvSIMDModImm64(AArch64ISD::MOVIedit, Op, DAG, DefBits)) ||
(NewOp =
tryAdvSIMDModImm32(AArch64ISD::MOVIshift, Op, DAG, DefBits)) ||
(NewOp =
tryAdvSIMDModImm321s(AArch64ISD::MOVImsl, Op, DAG, DefBits)) ||
(NewOp =
tryAdvSIMDModImm16(AArch64ISD::MOVIshift, Op, DAG, DefBits)) ||
(NewOp = tryAdvSIMDModImm8(AArch64ISD::MOVI, Op, DAG, DefBits)) ||
(NewOp = tryAdvSIMDModImmFP(AArch64ISD::FMOV, Op, DAG, DefBits)))
return NewOp;
APInt NotDefBits = ~DefBits;
if ((NewOp = tryAdvSIMDModImm32(AArch64ISD::MVNIshift, Op, DAG,
NotDefBits)) ||
(NewOp = tryAdvSIMDModImm321s(AArch64ISD::MVNImsl, Op, DAG,
NotDefBits)) ||
(NewOp =
tryAdvSIMDModImm16(AArch64ISD::MVNIshift, Op, DAG, NotDefBits)))
return NewOp;
return SDValue();
};
if (SDValue R = TryMOVIWithBits(DefBits))
return R;
if (SDValue R = TryMOVIWithBits(UndefBits))
return R;
// See if a fneg of the constant can be materialized with a MOVI, etc
auto TryWithFNeg = [&](APInt DefBits, MVT FVT) {
// FNegate each sub-element of the constant
assert(VT.getSizeInBits() % FVT.getScalarSizeInBits() == 0);
APInt Neg = APInt::getHighBitsSet(FVT.getSizeInBits(), 1)
.zext(VT.getSizeInBits());
APInt NegBits(VT.getSizeInBits(), 0);
unsigned NumElts = VT.getSizeInBits() / FVT.getScalarSizeInBits();
for (unsigned i = 0; i < NumElts; i++)
NegBits |= Neg << (FVT.getScalarSizeInBits() * i);
NegBits = DefBits ^ NegBits;
// Try to create the new constants with MOVI, and if so generate a fneg
// for it.
if (SDValue NewOp = TryMOVIWithBits(NegBits)) {
SDLoc DL(Op);
MVT VFVT = NumElts == 1 ? FVT : MVT::getVectorVT(FVT, NumElts);
return DAG.getNode(
AArch64ISD::NVCAST, DL, VT,
DAG.getNode(ISD::FNEG, DL, VFVT,
DAG.getNode(AArch64ISD::NVCAST, DL, VFVT, NewOp)));
}
return SDValue();
};
SDValue R;
if ((R = TryWithFNeg(DefBits, MVT::f32)) ||
(R = TryWithFNeg(DefBits, MVT::f64)) ||
(ST->hasFullFP16() && (R = TryWithFNeg(DefBits, MVT::f16))))
return R;
}
return SDValue();
}
SDValue AArch64TargetLowering::LowerFixedLengthBuildVectorToSVE(
SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
SDLoc DL(Op);
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
auto *BVN = cast<BuildVectorSDNode>(Op);
if (auto SeqInfo = BVN->isConstantSequence()) {
SDValue Start = DAG.getConstant(SeqInfo->first, DL, ContainerVT);
SDValue Steps = DAG.getStepVector(DL, ContainerVT, SeqInfo->second);
SDValue Seq = DAG.getNode(ISD::ADD, DL, ContainerVT, Start, Steps);
return convertFromScalableVector(DAG, VT, Seq);
}
unsigned NumElems = VT.getVectorNumElements();
if (!VT.isPow2VectorType() || VT.getFixedSizeInBits() > 128 ||
NumElems <= 1 || BVN->isConstant())
return SDValue();
auto IsExtractElt = [](SDValue Op) {
return Op.getOpcode() == ISD::EXTRACT_VECTOR_ELT;
};
// For integer types that are not already in vectors limit to at most four
// elements. This is an arbitrary restriction to avoid many fmovs from GPRs.
if (VT.getScalarType().isInteger() &&
NumElems - count_if(Op->op_values(), IsExtractElt) > 4)
return SDValue();
// Lower (pow2) BUILD_VECTORS that are <= 128-bit to a sequence of ZIP1s.
SDValue ZeroI64 = DAG.getConstant(0, DL, MVT::i64);
SmallVector<SDValue, 16> Intermediates = map_to_vector<16>(
Op->op_values(), [&, Undef = DAG.getUNDEF(ContainerVT)](SDValue Op) {
return Op.isUndef() ? Undef
: DAG.getNode(ISD::INSERT_VECTOR_ELT, DL,
ContainerVT, Undef, Op, ZeroI64);
});
ElementCount ZipEC = ContainerVT.getVectorElementCount();
while (Intermediates.size() > 1) {
EVT ZipVT = getPackedSVEVectorVT(ZipEC);
for (unsigned I = 0; I < Intermediates.size(); I += 2) {
SDValue Op0 = DAG.getBitcast(ZipVT, Intermediates[I + 0]);
SDValue Op1 = DAG.getBitcast(ZipVT, Intermediates[I + 1]);
Intermediates[I / 2] =
Op1.isUndef() ? Op0
: DAG.getNode(AArch64ISD::ZIP1, DL, ZipVT, Op0, Op1);
}
Intermediates.resize(Intermediates.size() / 2);
ZipEC = ZipEC.divideCoefficientBy(2);
}
assert(Intermediates.size() == 1);
SDValue Vec = DAG.getBitcast(ContainerVT, Intermediates[0]);
return convertFromScalableVector(DAG, VT, Vec);
}
SDValue AArch64TargetLowering::LowerBUILD_VECTOR(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
bool OverrideNEON = !Subtarget->isNeonAvailable() ||
cast<BuildVectorSDNode>(Op)->isConstantSequence();
if (useSVEForFixedLengthVectorVT(VT, OverrideNEON))
return LowerFixedLengthBuildVectorToSVE(Op, DAG);
// Try to build a simple constant vector.
Op = NormalizeBuildVector(Op, DAG);
// Thought this might return a non-BUILD_VECTOR (e.g. CONCAT_VECTORS), if so,
// abort.
if (Op.getOpcode() != ISD::BUILD_VECTOR)
return SDValue();
// Certain vector constants, used to express things like logical NOT and
// arithmetic NEG, are passed through unmodified. This allows special
// patterns for these operations to match, which will lower these constants
// to whatever is proven necessary.
BuildVectorSDNode *BVN = cast<BuildVectorSDNode>(Op.getNode());
if (BVN->isConstant()) {
if (ConstantSDNode *Const = BVN->getConstantSplatNode()) {
unsigned BitSize = VT.getVectorElementType().getSizeInBits();
APInt Val(BitSize,
Const->getAPIntValue().zextOrTrunc(BitSize).getZExtValue());
if (Val.isZero() || (VT.isInteger() && Val.isAllOnes()))
return Op;
}
if (ConstantFPSDNode *Const = BVN->getConstantFPSplatNode())
if (Const->isZero() && !Const->isNegative())
return Op;
}
if (SDValue V = ConstantBuildVector(Op, DAG, Subtarget))
return V;
// Scan through the operands to find some interesting properties we can
// exploit:
// 1) If only one value is used, we can use a DUP, or
// 2) if only the low element is not undef, we can just insert that, or
// 3) if only one constant value is used (w/ some non-constant lanes),
// we can splat the constant value into the whole vector then fill
// in the non-constant lanes.
// 4) FIXME: If different constant values are used, but we can intelligently
// select the values we'll be overwriting for the non-constant
// lanes such that we can directly materialize the vector
// some other way (MOVI, e.g.), we can be sneaky.
// 5) if all operands are EXTRACT_VECTOR_ELT, check for VUZP.
SDLoc DL(Op);
unsigned NumElts = VT.getVectorNumElements();
bool isOnlyLowElement = true;
bool usesOnlyOneValue = true;
bool usesOnlyOneConstantValue = true;
bool isConstant = true;
bool AllLanesExtractElt = true;
unsigned NumConstantLanes = 0;
unsigned NumDifferentLanes = 0;
unsigned NumUndefLanes = 0;
SDValue Value;
SDValue ConstantValue;
SmallMapVector<SDValue, unsigned, 16> DifferentValueMap;
unsigned ConsecutiveValCount = 0;
SDValue PrevVal;
for (unsigned i = 0; i < NumElts; ++i) {
SDValue V = Op.getOperand(i);
if (V.getOpcode() != ISD::EXTRACT_VECTOR_ELT)
AllLanesExtractElt = false;
if (V.isUndef()) {
++NumUndefLanes;
continue;
}
if (i > 0)
isOnlyLowElement = false;
if (!isIntOrFPConstant(V))
isConstant = false;
if (isIntOrFPConstant(V)) {
++NumConstantLanes;
if (!ConstantValue.getNode())
ConstantValue = V;
else if (ConstantValue != V)
usesOnlyOneConstantValue = false;
}
if (!Value.getNode())
Value = V;
else if (V != Value) {
usesOnlyOneValue = false;
++NumDifferentLanes;
}
if (PrevVal != V) {
ConsecutiveValCount = 0;
PrevVal = V;
}
// Keep different values and its last consecutive count. For example,
//
// t22: v16i8 = build_vector t23, t23, t23, t23, t23, t23, t23, t23,
// t24, t24, t24, t24, t24, t24, t24, t24
// t23 = consecutive count 8
// t24 = consecutive count 8
// ------------------------------------------------------------------
// t22: v16i8 = build_vector t24, t24, t23, t23, t23, t23, t23, t24,
// t24, t24, t24, t24, t24, t24, t24, t24
// t23 = consecutive count 5
// t24 = consecutive count 9
DifferentValueMap[V] = ++ConsecutiveValCount;
}
if (!Value.getNode()) {
LLVM_DEBUG(
dbgs() << "LowerBUILD_VECTOR: value undefined, creating undef node\n");
return DAG.getUNDEF(VT);
}
// Convert BUILD_VECTOR where all elements but the lowest are undef into
// SCALAR_TO_VECTOR, except for when we have a single-element constant vector
// as SimplifyDemandedBits will just turn that back into BUILD_VECTOR.
if (isOnlyLowElement && !(NumElts == 1 && isIntOrFPConstant(Value))) {
LLVM_DEBUG(dbgs() << "LowerBUILD_VECTOR: only low element used, creating 1 "
"SCALAR_TO_VECTOR node\n");
return DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, VT, Value);
}
if (AllLanesExtractElt) {
SDNode *Vector = nullptr;
bool Even = false;
bool Odd = false;
// Check whether the extract elements match the Even pattern <0,2,4,...> or
// the Odd pattern <1,3,5,...>.
for (unsigned i = 0; i < NumElts; ++i) {
SDValue V = Op.getOperand(i);
const SDNode *N = V.getNode();
if (!isa<ConstantSDNode>(N->getOperand(1))) {
Even = false;
Odd = false;
break;
}
SDValue N0 = N->getOperand(0);
// All elements are extracted from the same vector.
if (!Vector) {
Vector = N0.getNode();
// Check that the type of EXTRACT_VECTOR_ELT matches the type of
// BUILD_VECTOR.
if (VT.getVectorElementType() !=
N0.getValueType().getVectorElementType())
break;
} else if (Vector != N0.getNode()) {
Odd = false;
Even = false;
break;
}
// Extracted values are either at Even indices <0,2,4,...> or at Odd
// indices <1,3,5,...>.
uint64_t Val = N->getConstantOperandVal(1);
if (Val == 2 * i) {
Even = true;
continue;
}
if (Val - 1 == 2 * i) {
Odd = true;
continue;
}
// Something does not match: abort.
Odd = false;
Even = false;
break;
}
if (Even || Odd) {
SDValue LHS =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, SDValue(Vector, 0),
DAG.getConstant(0, DL, MVT::i64));
SDValue RHS =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, SDValue(Vector, 0),
DAG.getConstant(NumElts, DL, MVT::i64));
if (Even && !Odd)
return DAG.getNode(AArch64ISD::UZP1, DL, VT, LHS, RHS);
if (Odd && !Even)
return DAG.getNode(AArch64ISD::UZP2, DL, VT, LHS, RHS);
}
}
// Use DUP for non-constant splats. For f32 constant splats, reduce to
// i32 and try again.
if (usesOnlyOneValue) {
if (!isConstant) {
if (Value.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
Value.getValueType() != VT) {
LLVM_DEBUG(
dbgs() << "LowerBUILD_VECTOR: use DUP for non-constant splats\n");
return DAG.getNode(AArch64ISD::DUP, DL, VT, Value);
}
// This is actually a DUPLANExx operation, which keeps everything vectory.
SDValue Lane = Value.getOperand(1);
Value = Value.getOperand(0);
if (Value.getValueSizeInBits() == 64) {
LLVM_DEBUG(
dbgs() << "LowerBUILD_VECTOR: DUPLANE works on 128-bit vectors, "
"widening it\n");
Value = WidenVector(Value, DAG);
}
unsigned Opcode = getDUPLANEOp(VT.getVectorElementType());
return DAG.getNode(Opcode, DL, VT, Value, Lane);
}
if (VT.getVectorElementType().isFloatingPoint()) {
SmallVector<SDValue, 8> Ops;
EVT EltTy = VT.getVectorElementType();
assert ((EltTy == MVT::f16 || EltTy == MVT::bf16 || EltTy == MVT::f32 ||
EltTy == MVT::f64) && "Unsupported floating-point vector type");
LLVM_DEBUG(
dbgs() << "LowerBUILD_VECTOR: float constant splats, creating int "
"BITCASTS, and try again\n");
MVT NewType = MVT::getIntegerVT(EltTy.getSizeInBits());
for (unsigned i = 0; i < NumElts; ++i)
Ops.push_back(DAG.getNode(ISD::BITCAST, DL, NewType, Op.getOperand(i)));
EVT VecVT = EVT::getVectorVT(*DAG.getContext(), NewType, NumElts);
SDValue Val = DAG.getBuildVector(VecVT, DL, Ops);
LLVM_DEBUG(dbgs() << "LowerBUILD_VECTOR: trying to lower new vector: ";
Val.dump(););
Val = LowerBUILD_VECTOR(Val, DAG);
if (Val.getNode())
return DAG.getNode(ISD::BITCAST, DL, VT, Val);
}
}
// If we need to insert a small number of different non-constant elements and
// the vector width is sufficiently large, prefer using DUP with the common
// value and INSERT_VECTOR_ELT for the different lanes. If DUP is preferred,
// skip the constant lane handling below.
bool PreferDUPAndInsert =
!isConstant && NumDifferentLanes >= 1 &&
NumDifferentLanes < ((NumElts - NumUndefLanes) / 2) &&
NumDifferentLanes >= NumConstantLanes;
// If there was only one constant value used and for more than one lane,
// start by splatting that value, then replace the non-constant lanes. This
// is better than the default, which will perform a separate initialization
// for each lane.
if (!PreferDUPAndInsert && NumConstantLanes > 0 && usesOnlyOneConstantValue) {
// Firstly, try to materialize the splat constant.
SDValue Val = DAG.getSplatBuildVector(VT, DL, ConstantValue);
unsigned BitSize = VT.getScalarSizeInBits();
APInt ConstantValueAPInt(1, 0);
if (auto *C = dyn_cast<ConstantSDNode>(ConstantValue))
ConstantValueAPInt = C->getAPIntValue().zextOrTrunc(BitSize);
if (!isNullConstant(ConstantValue) && !isNullFPConstant(ConstantValue) &&
!ConstantValueAPInt.isAllOnes()) {
Val = ConstantBuildVector(Val, DAG, Subtarget);
if (!Val)
// Otherwise, materialize the constant and splat it.
Val = DAG.getNode(AArch64ISD::DUP, DL, VT, ConstantValue);
}
// Now insert the non-constant lanes.
for (unsigned i = 0; i < NumElts; ++i) {
SDValue V = Op.getOperand(i);
SDValue LaneIdx = DAG.getConstant(i, DL, MVT::i64);
if (!isIntOrFPConstant(V))
// Note that type legalization likely mucked about with the VT of the
// source operand, so we may have to convert it here before inserting.
Val = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VT, Val, V, LaneIdx);
}
return Val;
}
// This will generate a load from the constant pool.
if (isConstant) {
LLVM_DEBUG(
dbgs() << "LowerBUILD_VECTOR: all elements are constant, use default "
"expansion\n");
return SDValue();
}
// Detect patterns of a0,a1,a2,a3,b0,b1,b2,b3,c0,c1,c2,c3,d0,d1,d2,d3 from
// v4i32s. This is really a truncate, which we can construct out of (legal)
// concats and truncate nodes.
if (SDValue M = ReconstructTruncateFromBuildVector(Op, DAG))
return M;
// Empirical tests suggest this is rarely worth it for vectors of length <= 2.
if (NumElts >= 4) {
if (SDValue Shuffle = ReconstructShuffle(Op, DAG))
return Shuffle;
if (SDValue Shuffle = ReconstructShuffleWithRuntimeMask(Op, DAG))
return Shuffle;
}
if (PreferDUPAndInsert) {
// First, build a constant vector with the common element.
SmallVector<SDValue, 8> Ops(NumElts, Value);
SDValue NewVector = LowerBUILD_VECTOR(DAG.getBuildVector(VT, DL, Ops), DAG);
// Next, insert the elements that do not match the common value.
for (unsigned I = 0; I < NumElts; ++I)
if (Op.getOperand(I) != Value)
NewVector =
DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VT, NewVector,
Op.getOperand(I), DAG.getConstant(I, DL, MVT::i64));
return NewVector;
}
// If vector consists of two different values, try to generate two DUPs and
// (CONCAT_VECTORS or VECTOR_SHUFFLE).
if (DifferentValueMap.size() == 2 && NumUndefLanes == 0) {
SmallVector<SDValue, 2> Vals;
// Check the consecutive count of the value is the half number of vector
// elements. In this case, we can use CONCAT_VECTORS. For example,
//
// canUseVECTOR_CONCAT = true;
// t22: v16i8 = build_vector t23, t23, t23, t23, t23, t23, t23, t23,
// t24, t24, t24, t24, t24, t24, t24, t24
//
// canUseVECTOR_CONCAT = false;
// t22: v16i8 = build_vector t23, t23, t23, t23, t23, t24, t24, t24,
// t24, t24, t24, t24, t24, t24, t24, t24
bool canUseVECTOR_CONCAT = true;
for (auto Pair : DifferentValueMap) {
// Check different values have same length which is NumElts / 2.
if (Pair.second != NumElts / 2)
canUseVECTOR_CONCAT = false;
Vals.push_back(Pair.first);
}
// If canUseVECTOR_CONCAT is true, we can generate two DUPs and
// CONCAT_VECTORs. For example,
//
// t22: v16i8 = BUILD_VECTOR t23, t23, t23, t23, t23, t23, t23, t23,
// t24, t24, t24, t24, t24, t24, t24, t24
// ==>
// t26: v8i8 = AArch64ISD::DUP t23
// t28: v8i8 = AArch64ISD::DUP t24
// t29: v16i8 = concat_vectors t26, t28
if (canUseVECTOR_CONCAT) {
EVT SubVT = VT.getHalfNumVectorElementsVT(*DAG.getContext());
if (isTypeLegal(SubVT) && SubVT.isVector() &&
SubVT.getVectorNumElements() >= 2) {
SmallVector<SDValue, 8> Ops1(NumElts / 2, Vals[0]);
SmallVector<SDValue, 8> Ops2(NumElts / 2, Vals[1]);
SDValue DUP1 =
LowerBUILD_VECTOR(DAG.getBuildVector(SubVT, DL, Ops1), DAG);
SDValue DUP2 =
LowerBUILD_VECTOR(DAG.getBuildVector(SubVT, DL, Ops2), DAG);
SDValue CONCAT_VECTORS =
DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, DUP1, DUP2);
return CONCAT_VECTORS;
}
}
// Let's try to generate VECTOR_SHUFFLE. For example,
//
// t24: v8i8 = BUILD_VECTOR t25, t25, t25, t25, t26, t26, t26, t26
// ==>
// t27: v8i8 = BUILD_VECTOR t26, t26, t26, t26, t26, t26, t26, t26
// t28: v8i8 = BUILD_VECTOR t25, t25, t25, t25, t25, t25, t25, t25
// t29: v8i8 = vector_shuffle<0,1,2,3,12,13,14,15> t27, t28
if (NumElts >= 8) {
SmallVector<int, 16> MaskVec;
// Build mask for VECTOR_SHUFLLE.
SDValue FirstLaneVal = Op.getOperand(0);
for (unsigned i = 0; i < NumElts; ++i) {
SDValue Val = Op.getOperand(i);
if (FirstLaneVal == Val)
MaskVec.push_back(i);
else
MaskVec.push_back(i + NumElts);
}
SmallVector<SDValue, 8> Ops1(NumElts, Vals[0]);
SmallVector<SDValue, 8> Ops2(NumElts, Vals[1]);
SDValue VEC1 = DAG.getBuildVector(VT, DL, Ops1);
SDValue VEC2 = DAG.getBuildVector(VT, DL, Ops2);
SDValue VECTOR_SHUFFLE =
DAG.getVectorShuffle(VT, DL, VEC1, VEC2, MaskVec);
return VECTOR_SHUFFLE;
}
}
// If all else fails, just use a sequence of INSERT_VECTOR_ELT when we
// know the default expansion would otherwise fall back on something even
// worse. For a vector with one or two non-undef values, that's
// scalar_to_vector for the elements followed by a shuffle (provided the
// shuffle is valid for the target) and materialization element by element
// on the stack followed by a load for everything else.
if (!isConstant && !usesOnlyOneValue) {
LLVM_DEBUG(
dbgs() << "LowerBUILD_VECTOR: alternatives failed, creating sequence "
"of INSERT_VECTOR_ELT\n");
SDValue Vec = DAG.getUNDEF(VT);
SDValue Op0 = Op.getOperand(0);
unsigned i = 0;
// Use SCALAR_TO_VECTOR for lane zero to
// a) Avoid a RMW dependency on the full vector register, and
// b) Allow the register coalescer to fold away the copy if the
// value is already in an S or D register, and we're forced to emit an
// INSERT_SUBREG that we can't fold anywhere.
//
// We also allow types like i8 and i16 which are illegal scalar but legal
// vector element types. After type-legalization the inserted value is
// extended (i32) and it is safe to cast them to the vector type by ignoring
// the upper bits of the lowest lane (e.g. v8i8, v4i16).
if (!Op0.isUndef()) {
LLVM_DEBUG(dbgs() << "Creating node for op0, it is not undefined:\n");
Vec = DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, VT, Op0);
++i;
}
LLVM_DEBUG({
if (i < NumElts)
dbgs() << "Creating nodes for the other vector elements:\n";
});
for (; i < NumElts; ++i) {
SDValue V = Op.getOperand(i);
if (V.isUndef())
continue;
SDValue LaneIdx = DAG.getConstant(i, DL, MVT::i64);
Vec = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VT, Vec, V, LaneIdx);
}
return Vec;
}
LLVM_DEBUG(
dbgs() << "LowerBUILD_VECTOR: use default expansion, failed to find "
"better alternative\n");
return SDValue();
}
SDValue AArch64TargetLowering::LowerCONCAT_VECTORS(SDValue Op,
SelectionDAG &DAG) const {
if (useSVEForFixedLengthVectorVT(Op.getValueType(),
!Subtarget->isNeonAvailable()))
return LowerFixedLengthConcatVectorsToSVE(Op, DAG);
assert(Op.getValueType().isScalableVector() &&
isTypeLegal(Op.getValueType()) &&
"Expected legal scalable vector type!");
if (isTypeLegal(Op.getOperand(0).getValueType())) {
unsigned NumOperands = Op->getNumOperands();
assert(NumOperands > 1 && isPowerOf2_32(NumOperands) &&
"Unexpected number of operands in CONCAT_VECTORS");
if (NumOperands == 2)
return Op;
// Concat each pair of subvectors and pack into the lower half of the array.
SmallVector<SDValue> ConcatOps(Op->ops());
while (ConcatOps.size() > 1) {
for (unsigned I = 0, E = ConcatOps.size(); I != E; I += 2) {
SDValue V1 = ConcatOps[I];
SDValue V2 = ConcatOps[I + 1];
EVT SubVT = V1.getValueType();
EVT PairVT = SubVT.getDoubleNumVectorElementsVT(*DAG.getContext());
ConcatOps[I / 2] =
DAG.getNode(ISD::CONCAT_VECTORS, SDLoc(Op), PairVT, V1, V2);
}
ConcatOps.resize(ConcatOps.size() / 2);
}
return ConcatOps[0];
}
return SDValue();
}
SDValue AArch64TargetLowering::LowerINSERT_VECTOR_ELT(SDValue Op,
SelectionDAG &DAG) const {
assert(Op.getOpcode() == ISD::INSERT_VECTOR_ELT && "Unknown opcode!");
if (useSVEForFixedLengthVectorVT(Op.getValueType(),
!Subtarget->isNeonAvailable()))
return LowerFixedLengthInsertVectorElt(Op, DAG);
EVT VT = Op.getOperand(0).getValueType();
if (VT.getScalarType() == MVT::i1) {
EVT VectorVT = getPromotedVTForPredicate(VT);
SDLoc DL(Op);
SDValue ExtendedVector =
DAG.getAnyExtOrTrunc(Op.getOperand(0), DL, VectorVT);
SDValue ExtendedValue =
DAG.getAnyExtOrTrunc(Op.getOperand(1), DL,
VectorVT.getScalarType().getSizeInBits() < 32
? MVT::i32
: VectorVT.getScalarType());
ExtendedVector =
DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VectorVT, ExtendedVector,
ExtendedValue, Op.getOperand(2));
return DAG.getAnyExtOrTrunc(ExtendedVector, DL, VT);
}
// Check for non-constant or out of range lane.
ConstantSDNode *CI = dyn_cast<ConstantSDNode>(Op.getOperand(2));
if (!CI || CI->getZExtValue() >= VT.getVectorNumElements())
return SDValue();
return Op;
}
SDValue
AArch64TargetLowering::LowerEXTRACT_VECTOR_ELT(SDValue Op,
SelectionDAG &DAG) const {
assert(Op.getOpcode() == ISD::EXTRACT_VECTOR_ELT && "Unknown opcode!");
EVT VT = Op.getOperand(0).getValueType();
if (VT.getScalarType() == MVT::i1) {
// We can't directly extract from an SVE predicate; extend it first.
// (This isn't the only possible lowering, but it's straightforward.)
EVT VectorVT = getPromotedVTForPredicate(VT);
SDLoc DL(Op);
SDValue Extend =
DAG.getNode(ISD::ANY_EXTEND, DL, VectorVT, Op.getOperand(0));
MVT ExtractTy = VectorVT == MVT::nxv2i64 ? MVT::i64 : MVT::i32;
SDValue Extract = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ExtractTy,
Extend, Op.getOperand(1));
return DAG.getAnyExtOrTrunc(Extract, DL, Op.getValueType());
}
if (useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()))
return LowerFixedLengthExtractVectorElt(Op, DAG);
// Check for non-constant or out of range lane.
ConstantSDNode *CI = dyn_cast<ConstantSDNode>(Op.getOperand(1));
if (!CI || CI->getZExtValue() >= VT.getVectorNumElements())
return SDValue();
// Insertion/extraction are legal for V128 types.
if (VT == MVT::v16i8 || VT == MVT::v8i16 || VT == MVT::v4i32 ||
VT == MVT::v2i64 || VT == MVT::v4f32 || VT == MVT::v2f64 ||
VT == MVT::v8f16 || VT == MVT::v8bf16)
return Op;
if (VT != MVT::v8i8 && VT != MVT::v4i16 && VT != MVT::v2i32 &&
VT != MVT::v1i64 && VT != MVT::v2f32 && VT != MVT::v4f16 &&
VT != MVT::v4bf16)
return SDValue();
// For V64 types, we perform extraction by expanding the value
// to a V128 type and perform the extraction on that.
SDLoc DL(Op);
SDValue WideVec = WidenVector(Op.getOperand(0), DAG);
EVT WideTy = WideVec.getValueType();
EVT ExtrTy = WideTy.getVectorElementType();
if (ExtrTy == MVT::i16 || ExtrTy == MVT::i8)
ExtrTy = MVT::i32;
// For extractions, we just return the result directly.
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ExtrTy, WideVec,
Op.getOperand(1));
}
SDValue AArch64TargetLowering::LowerEXTRACT_SUBVECTOR(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() &&
"Only cases that extract a fixed length vector are supported!");
EVT InVT = Op.getOperand(0).getValueType();
// If we don't have legal types yet, do nothing
if (!isTypeLegal(InVT))
return SDValue();
if (InVT.is128BitVector()) {
assert(VT.is64BitVector() && "Extracting unexpected vector type!");
unsigned Idx = Op.getConstantOperandVal(1);
// This will get lowered to an appropriate EXTRACT_SUBREG in ISel.
if (Idx == 0)
return Op;
// If this is extracting the upper 64-bits of a 128-bit vector, we match
// that directly.
if (Idx * InVT.getScalarSizeInBits() == 64 && Subtarget->isNeonAvailable())
return Op;
}
if (InVT.isScalableVector() ||
useSVEForFixedLengthVectorVT(InVT, !Subtarget->isNeonAvailable())) {
SDLoc DL(Op);
SDValue Vec = Op.getOperand(0);
SDValue Idx = Op.getOperand(1);
EVT PackedVT = getPackedSVEVectorVT(InVT.getVectorElementType());
if (PackedVT != InVT) {
// Pack input into the bottom part of an SVE register and try again.
SDValue Container = DAG.getNode(ISD::INSERT_SUBVECTOR, DL, PackedVT,
DAG.getUNDEF(PackedVT), Vec,
DAG.getVectorIdxConstant(0, DL));
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, Container, Idx);
}
// This will get matched by custom code during ISelDAGToDAG.
if (isNullConstant(Idx))
return Op;
assert(InVT.isScalableVector() && "Unexpected vector type!");
// Move requested subvector to the start of the vector and try again.
SDValue Splice = DAG.getNode(ISD::VECTOR_SPLICE, DL, InVT, Vec, Vec, Idx);
return convertFromScalableVector(DAG, VT, Splice);
}
return SDValue();
}
SDValue AArch64TargetLowering::LowerINSERT_SUBVECTOR(SDValue Op,
SelectionDAG &DAG) const {
assert(Op.getValueType().isScalableVector() &&
"Only expect to lower inserts into scalable vectors!");
EVT InVT = Op.getOperand(1).getValueType();
unsigned Idx = Op.getConstantOperandVal(2);
SDValue Vec0 = Op.getOperand(0);
SDValue Vec1 = Op.getOperand(1);
SDLoc DL(Op);
EVT VT = Op.getValueType();
if (InVT.isScalableVector()) {
if (!isTypeLegal(VT))
return SDValue();
// Break down insert_subvector into simpler parts.
if (VT.getVectorElementType() == MVT::i1) {
unsigned NumElts = VT.getVectorMinNumElements();
EVT HalfVT = VT.getHalfNumVectorElementsVT(*DAG.getContext());
SDValue Lo, Hi;
Lo = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, HalfVT, Vec0,
DAG.getVectorIdxConstant(0, DL));
Hi = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, HalfVT, Vec0,
DAG.getVectorIdxConstant(NumElts / 2, DL));
if (Idx < (NumElts / 2))
Lo = DAG.getNode(ISD::INSERT_SUBVECTOR, DL, HalfVT, Lo, Vec1,
DAG.getVectorIdxConstant(Idx, DL));
else
Hi = DAG.getNode(ISD::INSERT_SUBVECTOR, DL, HalfVT, Hi, Vec1,
DAG.getVectorIdxConstant(Idx - (NumElts / 2), DL));
return DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, Lo, Hi);
}
// We can select these directly.
if (isTypeLegal(InVT) && Vec0.isUndef())
return Op;
// Ensure the subvector is half the size of the main vector.
if (VT.getVectorElementCount() != (InVT.getVectorElementCount() * 2))
return SDValue();
// Here narrow and wide refers to the vector element types. After "casting"
// both vectors must have the same bit length and so because the subvector
// has fewer elements, those elements need to be bigger.
EVT NarrowVT = getPackedSVEVectorVT(VT.getVectorElementCount());
EVT WideVT = getPackedSVEVectorVT(InVT.getVectorElementCount());
// NOP cast operands to the largest legal vector of the same element count.
if (VT.isFloatingPoint()) {
Vec0 = getSVESafeBitCast(NarrowVT, Vec0, DAG);
Vec1 = getSVESafeBitCast(NarrowVT, Vec1, DAG);
} else {
// Legal integer vectors are already their largest so Vec0 is fine as is.
Vec1 = DAG.getNode(ISD::ANY_EXTEND, DL, WideVT, Vec1);
Vec1 = DAG.getNode(AArch64ISD::NVCAST, DL, NarrowVT, Vec1);
}
// To replace the top/bottom half of vector V with vector SubV we widen the
// preserved half of V, concatenate this to SubV (the order depending on the
// half being replaced) and then narrow the result.
SDValue Narrow;
if (Idx == 0) {
SDValue HiVec0 = DAG.getNode(AArch64ISD::UUNPKHI, DL, WideVT, Vec0);
HiVec0 = DAG.getNode(AArch64ISD::NVCAST, DL, NarrowVT, HiVec0);
Narrow = DAG.getNode(AArch64ISD::UZP1, DL, NarrowVT, Vec1, HiVec0);
} else {
assert(Idx == InVT.getVectorMinNumElements() &&
"Invalid subvector index!");
SDValue LoVec0 = DAG.getNode(AArch64ISD::UUNPKLO, DL, WideVT, Vec0);
LoVec0 = DAG.getNode(AArch64ISD::NVCAST, DL, NarrowVT, LoVec0);
Narrow = DAG.getNode(AArch64ISD::UZP1, DL, NarrowVT, LoVec0, Vec1);
}
return getSVESafeBitCast(VT, Narrow, DAG);
}
if (Idx == 0 && isPackedVectorType(VT, DAG)) {
// This will be matched by custom code during ISelDAGToDAG.
if (Vec0.isUndef())
return Op;
std::optional<unsigned> PredPattern =
getSVEPredPatternFromNumElements(InVT.getVectorNumElements());
auto PredTy = VT.changeVectorElementType(MVT::i1);
SDValue PTrue = getPTrue(DAG, DL, PredTy, *PredPattern);
SDValue ScalableVec1 = convertToScalableVector(DAG, VT, Vec1);
return DAG.getNode(ISD::VSELECT, DL, VT, PTrue, ScalableVec1, Vec0);
}
return SDValue();
}
static bool isPow2Splat(SDValue Op, uint64_t &SplatVal, bool &Negated) {
if (Op.getOpcode() != AArch64ISD::DUP &&
Op.getOpcode() != ISD::SPLAT_VECTOR &&
Op.getOpcode() != ISD::BUILD_VECTOR)
return false;
if (Op.getOpcode() == ISD::BUILD_VECTOR &&
!isAllConstantBuildVector(Op, SplatVal))
return false;
if (Op.getOpcode() != ISD::BUILD_VECTOR &&
!isa<ConstantSDNode>(Op->getOperand(0)))
return false;
SplatVal = Op->getConstantOperandVal(0);
if (Op.getValueType().getVectorElementType() != MVT::i64)
SplatVal = (int32_t)SplatVal;
Negated = false;
if (isPowerOf2_64(SplatVal))
return true;
Negated = true;
if (isPowerOf2_64(-SplatVal)) {
SplatVal = -SplatVal;
return true;
}
return false;
}
SDValue AArch64TargetLowering::LowerDIV(SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
SDLoc DL(Op);
if (useSVEForFixedLengthVectorVT(VT, /*OverrideNEON=*/true))
return LowerFixedLengthVectorIntDivideToSVE(Op, DAG);
assert(VT.isScalableVector() && "Expected a scalable vector.");
bool Signed = Op.getOpcode() == ISD::SDIV;
unsigned PredOpcode = Signed ? AArch64ISD::SDIV_PRED : AArch64ISD::UDIV_PRED;
bool Negated;
uint64_t SplatVal;
if (Signed && isPow2Splat(Op.getOperand(1), SplatVal, Negated)) {
SDValue Pg = getPredicateForScalableVector(DAG, DL, VT);
SDValue Res =
DAG.getNode(AArch64ISD::SRAD_MERGE_OP1, DL, VT, Pg, Op->getOperand(0),
DAG.getTargetConstant(Log2_64(SplatVal), DL, MVT::i32));
if (Negated)
Res = DAG.getNode(ISD::SUB, DL, VT, DAG.getConstant(0, DL, VT), Res);
return Res;
}
if (VT == MVT::nxv4i32 || VT == MVT::nxv2i64)
return LowerToPredicatedOp(Op, DAG, PredOpcode);
// SVE doesn't have i8 and i16 DIV operations; widen them to 32-bit
// operations, and truncate the result.
EVT WidenedVT;
if (VT == MVT::nxv16i8)
WidenedVT = MVT::nxv8i16;
else if (VT == MVT::nxv8i16)
WidenedVT = MVT::nxv4i32;
else
llvm_unreachable("Unexpected Custom DIV operation");
unsigned UnpkLo = Signed ? AArch64ISD::SUNPKLO : AArch64ISD::UUNPKLO;
unsigned UnpkHi = Signed ? AArch64ISD::SUNPKHI : AArch64ISD::UUNPKHI;
SDValue Op0Lo = DAG.getNode(UnpkLo, DL, WidenedVT, Op.getOperand(0));
SDValue Op1Lo = DAG.getNode(UnpkLo, DL, WidenedVT, Op.getOperand(1));
SDValue Op0Hi = DAG.getNode(UnpkHi, DL, WidenedVT, Op.getOperand(0));
SDValue Op1Hi = DAG.getNode(UnpkHi, DL, WidenedVT, Op.getOperand(1));
SDValue ResultLo = DAG.getNode(Op.getOpcode(), DL, WidenedVT, Op0Lo, Op1Lo);
SDValue ResultHi = DAG.getNode(Op.getOpcode(), DL, WidenedVT, Op0Hi, Op1Hi);
SDValue ResultLoCast = DAG.getNode(AArch64ISD::NVCAST, DL, VT, ResultLo);
SDValue ResultHiCast = DAG.getNode(AArch64ISD::NVCAST, DL, VT, ResultHi);
return DAG.getNode(AArch64ISD::UZP1, DL, VT, ResultLoCast, ResultHiCast);
}
bool AArch64TargetLowering::shouldExpandBuildVectorWithShuffles(
EVT VT, unsigned DefinedValues) const {
if (!Subtarget->isNeonAvailable())
return false;
return TargetLowering::shouldExpandBuildVectorWithShuffles(VT, DefinedValues);
}
bool AArch64TargetLowering::isShuffleMaskLegal(ArrayRef<int> M, EVT VT) const {
// Currently no fixed length shuffles that require SVE are legal.
if (useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()))
return false;
if (VT.getVectorNumElements() == 4 &&
(VT.is128BitVector() || VT.is64BitVector())) {
unsigned Cost = getPerfectShuffleCost(M);
if (Cost <= 1)
return true;
}
bool DummyBool;
int DummyInt;
unsigned DummyUnsigned;
unsigned EltSize = VT.getScalarSizeInBits();
unsigned NumElts = VT.getVectorNumElements();
return (ShuffleVectorSDNode::isSplatMask(M) ||
isREVMask(M, EltSize, NumElts, 64) ||
isREVMask(M, EltSize, NumElts, 32) ||
isREVMask(M, EltSize, NumElts, 16) ||
isEXTMask(M, VT, DummyBool, DummyUnsigned) ||
isTRNMask(M, NumElts, DummyUnsigned) ||
isUZPMask(M, NumElts, DummyUnsigned) ||
isZIPMask(M, NumElts, DummyUnsigned) ||
isTRN_v_undef_Mask(M, VT, DummyUnsigned) ||
isUZP_v_undef_Mask(M, VT, DummyUnsigned) ||
isZIP_v_undef_Mask(M, VT, DummyUnsigned) ||
isINSMask(M, NumElts, DummyBool, DummyInt) ||
isConcatMask(M, VT, VT.getSizeInBits() == 128));
}
bool AArch64TargetLowering::isVectorClearMaskLegal(ArrayRef<int> M,
EVT VT) const {
// Just delegate to the generic legality, clear masks aren't special.
return isShuffleMaskLegal(M, VT);
}
/// getVShiftImm - Check if this is a valid build_vector for the immediate
/// operand of a vector shift operation, where all the elements of the
/// build_vector must have the same constant integer value.
static bool getVShiftImm(SDValue Op, unsigned ElementBits, int64_t &Cnt) {
// Ignore bit_converts.
while (Op.getOpcode() == ISD::BITCAST)
Op = Op.getOperand(0);
BuildVectorSDNode *BVN = dyn_cast<BuildVectorSDNode>(Op.getNode());
APInt SplatBits, SplatUndef;
unsigned SplatBitSize;
bool HasAnyUndefs;
if (!BVN || !BVN->isConstantSplat(SplatBits, SplatUndef, SplatBitSize,
HasAnyUndefs, ElementBits) ||
SplatBitSize > ElementBits)
return false;
Cnt = SplatBits.getSExtValue();
return true;
}
/// isVShiftLImm - Check if this is a valid build_vector for the immediate
/// operand of a vector shift left operation. That value must be in the range:
/// 0 <= Value < ElementBits for a left shift; or
/// 0 <= Value <= ElementBits for a long left shift.
static bool isVShiftLImm(SDValue Op, EVT VT, bool isLong, int64_t &Cnt) {
assert(VT.isVector() && "vector shift count is not a vector type");
int64_t ElementBits = VT.getScalarSizeInBits();
if (!getVShiftImm(Op, ElementBits, Cnt))
return false;
return (Cnt >= 0 && (isLong ? Cnt - 1 : Cnt) < ElementBits);
}
/// isVShiftRImm - Check if this is a valid build_vector for the immediate
/// operand of a vector shift right operation. The value must be in the range:
/// 1 <= Value <= ElementBits for a right shift; or
static bool isVShiftRImm(SDValue Op, EVT VT, bool isNarrow, int64_t &Cnt) {
assert(VT.isVector() && "vector shift count is not a vector type");
int64_t ElementBits = VT.getScalarSizeInBits();
if (!getVShiftImm(Op, ElementBits, Cnt))
return false;
return (Cnt >= 1 && Cnt <= (isNarrow ? ElementBits / 2 : ElementBits));
}
SDValue AArch64TargetLowering::LowerTRUNCATE(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
if (VT.getScalarType() == MVT::i1) {
// Lower i1 truncate to `(x & 1) != 0`.
SDLoc DL(Op);
EVT OpVT = Op.getOperand(0).getValueType();
SDValue Zero = DAG.getConstant(0, DL, OpVT);
SDValue One = DAG.getConstant(1, DL, OpVT);
SDValue And = DAG.getNode(ISD::AND, DL, OpVT, Op.getOperand(0), One);
return DAG.getSetCC(DL, VT, And, Zero, ISD::SETNE);
}
if (!VT.isVector() || VT.isScalableVector())
return SDValue();
if (useSVEForFixedLengthVectorVT(Op.getOperand(0).getValueType(),
!Subtarget->isNeonAvailable()))
return LowerFixedLengthVectorTruncateToSVE(Op, DAG);
return SDValue();
}
// Check if we can we lower this SRL to a rounding shift instruction. ResVT is
// possibly a truncated type, it tells how many bits of the value are to be
// used.
static bool canLowerSRLToRoundingShiftForVT(SDValue Shift, EVT ResVT,
SelectionDAG &DAG,
unsigned &ShiftValue,
SDValue &RShOperand) {
if (Shift->getOpcode() != ISD::SRL)
return false;
EVT VT = Shift.getValueType();
assert(VT.isScalableVT());
auto ShiftOp1 =
dyn_cast_or_null<ConstantSDNode>(DAG.getSplatValue(Shift->getOperand(1)));
if (!ShiftOp1)
return false;
ShiftValue = ShiftOp1->getZExtValue();
if (ShiftValue < 1 || ShiftValue > ResVT.getScalarSizeInBits())
return false;
SDValue Add = Shift->getOperand(0);
if (Add->getOpcode() != ISD::ADD || !Add->hasOneUse())
return false;
assert(ResVT.getScalarSizeInBits() <= VT.getScalarSizeInBits() &&
"ResVT must be truncated or same type as the shift.");
// Check if an overflow can lead to incorrect results.
uint64_t ExtraBits = VT.getScalarSizeInBits() - ResVT.getScalarSizeInBits();
if (ShiftValue > ExtraBits && !Add->getFlags().hasNoUnsignedWrap())
return false;
auto AddOp1 =
dyn_cast_or_null<ConstantSDNode>(DAG.getSplatValue(Add->getOperand(1)));
if (!AddOp1)
return false;
uint64_t AddValue = AddOp1->getZExtValue();
if (AddValue != 1ULL << (ShiftValue - 1))
return false;
RShOperand = Add->getOperand(0);
return true;
}
SDValue AArch64TargetLowering::LowerVectorSRA_SRL_SHL(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
SDLoc DL(Op);
int64_t Cnt;
if (!Op.getOperand(1).getValueType().isVector())
return Op;
unsigned EltSize = VT.getScalarSizeInBits();
switch (Op.getOpcode()) {
case ISD::SHL:
if (VT.isScalableVector() ||
useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable()))
return LowerToPredicatedOp(Op, DAG, AArch64ISD::SHL_PRED);
if (isVShiftLImm(Op.getOperand(1), VT, false, Cnt) && Cnt < EltSize)
return DAG.getNode(AArch64ISD::VSHL, DL, VT, Op.getOperand(0),
DAG.getConstant(Cnt, DL, MVT::i32));
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, VT,
DAG.getConstant(Intrinsic::aarch64_neon_ushl, DL,
MVT::i32),
Op.getOperand(0), Op.getOperand(1));
case ISD::SRA:
case ISD::SRL:
if (VT.isScalableVector() &&
(Subtarget->hasSVE2() ||
(Subtarget->hasSME() && Subtarget->isStreaming()))) {
SDValue RShOperand;
unsigned ShiftValue;
if (canLowerSRLToRoundingShiftForVT(Op, VT, DAG, ShiftValue, RShOperand))
return DAG.getNode(AArch64ISD::URSHR_I_PRED, DL, VT,
getPredicateForVector(DAG, DL, VT), RShOperand,
DAG.getTargetConstant(ShiftValue, DL, MVT::i32));
}
if (VT.isScalableVector() ||
useSVEForFixedLengthVectorVT(VT, !Subtarget->isNeonAvailable())) {
unsigned Opc = Op.getOpcode() == ISD::SRA ? AArch64ISD::SRA_PRED
: AArch64ISD::SRL_PRED;
return LowerToPredicatedOp(Op, DAG, Opc);
}
// Right shift immediate
if (isVShiftRImm(Op.getOperand(1), VT, false, Cnt) && Cnt < EltSize) {
unsigned Opc =
(Op.getOpcode() == ISD::SRA) ? AArch64ISD::VASHR : AArch64ISD::VLSHR;
return DAG.getNode(Opc, DL, VT, Op.getOperand(0),
DAG.getConstant(Cnt, DL, MVT::i32), Op->getFlags());
}
// Right shift register. Note, there is not a shift right register
// instruction, but the shift left register instruction takes a signed
// value, where negative numbers specify a right shift.
unsigned Opc = (Op.getOpcode() == ISD::SRA) ? Intrinsic::aarch64_neon_sshl
: Intrinsic::aarch64_neon_ushl;
// negate the shift amount
SDValue NegShift = DAG.getNode(ISD::SUB, DL, VT, DAG.getConstant(0, DL, VT),
Op.getOperand(1));
SDValue NegShiftLeft =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, VT,
DAG.getConstant(Opc, DL, MVT::i32), Op.getOperand(0),
NegShift);
return NegShiftLeft;
}
llvm_unreachable("unexpected shift opcode");
}
SDValue AArch64TargetLowering::LowerVSETCC(SDValue Op,
SelectionDAG &DAG) const {
if (Op.getValueType().isScalableVector())
return LowerToPredicatedOp(Op, DAG, AArch64ISD::SETCC_MERGE_ZERO);
if (useSVEForFixedLengthVectorVT(Op.getOperand(0).getValueType(),
!Subtarget->isNeonAvailable()))
return LowerFixedLengthVectorSetccToSVE(Op, DAG);
ISD::CondCode CC = cast<CondCodeSDNode>(Op.getOperand(2))->get();
SDValue LHS = Op.getOperand(0);
SDValue RHS = Op.getOperand(1);
EVT CmpVT = LHS.getValueType().changeVectorElementTypeToInteger();
SDLoc DL(Op);
if (LHS.getValueType().getVectorElementType().isInteger())
return Op;
assert(((!Subtarget->hasFullFP16() &&
LHS.getValueType().getVectorElementType() != MVT::f16) ||
LHS.getValueType().getVectorElementType() != MVT::bf16 ||
LHS.getValueType().getVectorElementType() != MVT::f128) &&
"Unexpected type!");
// Lower isnan(x) | isnan(never-nan) to x != x.
// Lower !isnan(x) & !isnan(never-nan) to x == x.
if (CC == ISD::SETUO || CC == ISD::SETO) {
bool OneNaN = false;
if (LHS == RHS) {
OneNaN = true;
} else if (DAG.isKnownNeverNaN(RHS)) {
OneNaN = true;
RHS = LHS;
} else if (DAG.isKnownNeverNaN(LHS)) {
OneNaN = true;
LHS = RHS;
}
if (OneNaN) {
CC = CC == ISD::SETUO ? ISD::SETUNE : ISD::SETOEQ;
}
}
// Unfortunately, the mapping of LLVM FP CC's onto AArch64 CC's isn't totally
// clean. Some of them require two branches to implement.
AArch64CC::CondCode CC1, CC2;
bool ShouldInvert;
changeVectorFPCCToAArch64CC(CC, CC1, CC2, ShouldInvert);
bool NoNaNs =
getTargetMachine().Options.NoNaNsFPMath || Op->getFlags().hasNoNaNs();
SDValue Cmp = emitVectorComparison(LHS, RHS, CC1, NoNaNs, CmpVT, DL, DAG);
if (!Cmp.getNode())
return SDValue();
if (CC2 != AArch64CC::AL) {
SDValue Cmp2 = emitVectorComparison(LHS, RHS, CC2, NoNaNs, CmpVT, DL, DAG);
if (!Cmp2.getNode())
return SDValue();
Cmp = DAG.getNode(ISD::OR, DL, CmpVT, Cmp, Cmp2);
}
Cmp = DAG.getSExtOrTrunc(Cmp, DL, Op.getValueType());
if (ShouldInvert)
Cmp = DAG.getNOT(DL, Cmp, Cmp.getValueType());
return Cmp;
}
static SDValue getReductionSDNode(unsigned Op, SDLoc DL, SDValue ScalarOp,
SelectionDAG &DAG) {
SDValue VecOp = ScalarOp.getOperand(0);
auto Rdx = DAG.getNode(Op, DL, VecOp.getSimpleValueType(), VecOp);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ScalarOp.getValueType(), Rdx,
DAG.getConstant(0, DL, MVT::i64));
}
static SDValue getVectorBitwiseReduce(unsigned Opcode, SDValue Vec, EVT VT,
SDLoc DL, SelectionDAG &DAG) {
unsigned ScalarOpcode;
switch (Opcode) {
case ISD::VECREDUCE_AND:
ScalarOpcode = ISD::AND;
break;
case ISD::VECREDUCE_OR:
ScalarOpcode = ISD::OR;
break;
case ISD::VECREDUCE_XOR:
ScalarOpcode = ISD::XOR;
break;
default:
llvm_unreachable("Expected bitwise vector reduction");
return SDValue();
}
EVT VecVT = Vec.getValueType();
assert(VecVT.isFixedLengthVector() && VecVT.isPow2VectorType() &&
"Expected power-of-2 length vector");
EVT ElemVT = VecVT.getVectorElementType();
SDValue Result;
unsigned NumElems = VecVT.getVectorNumElements();
// Special case for boolean reductions
if (ElemVT == MVT::i1) {
// Split large vectors into smaller ones
if (NumElems > 16) {
SDValue Lo, Hi;
std::tie(Lo, Hi) = DAG.SplitVector(Vec, DL);
EVT HalfVT = Lo.getValueType();
SDValue HalfVec = DAG.getNode(ScalarOpcode, DL, HalfVT, Lo, Hi);
return getVectorBitwiseReduce(Opcode, HalfVec, VT, DL, DAG);
}
// Results of setcc operations get widened to 128 bits if their input
// operands are 128 bits wide, otherwise vectors that are less than 64 bits
// get widened to neatly fit a 64 bit register, so e.g. <4 x i1> gets
// lowered to either <4 x i16> or <4 x i32>. Sign extending to this element
// size leads to the best codegen, since e.g. setcc results might need to be
// truncated otherwise.
unsigned ExtendedWidth = 64;
if (Vec.getOpcode() == ISD::SETCC &&
Vec.getOperand(0).getValueSizeInBits() >= 128) {
ExtendedWidth = 128;
}
EVT ExtendedVT = MVT::getIntegerVT(std::max(ExtendedWidth / NumElems, 8u));
// any_ext doesn't work with umin/umax, so only use it for uadd.
unsigned ExtendOp =
ScalarOpcode == ISD::XOR ? ISD::ANY_EXTEND : ISD::SIGN_EXTEND;
SDValue Extended = DAG.getNode(
ExtendOp, DL, VecVT.changeVectorElementType(ExtendedVT), Vec);
// The uminp/uminv and umaxp/umaxv instructions don't have .2d variants, so
// in that case we bitcast the sign extended values from v2i64 to v4i32
// before reduction for optimal code generation.
if ((ScalarOpcode == ISD::AND || ScalarOpcode == ISD::OR) &&
NumElems == 2 && ExtendedWidth == 128) {
Extended = DAG.getBitcast(MVT::v4i32, Extended);
ExtendedVT = MVT::i32;
}
switch (ScalarOpcode) {
case ISD::AND:
Result = DAG.getNode(ISD::VECREDUCE_UMIN, DL, ExtendedVT, Extended);
break;
case ISD::OR:
Result = DAG.getNode(ISD::VECREDUCE_UMAX, DL, ExtendedVT, Extended);
break;
case ISD::XOR:
Result = DAG.getNode(ISD::VECREDUCE_ADD, DL, ExtendedVT, Extended);
break;
default:
llvm_unreachable("Unexpected Opcode");
}
Result = DAG.getAnyExtOrTrunc(Result, DL, MVT::i1);
} else {
// Iteratively split the vector in half and combine using the bitwise
// operation until it fits in a 64 bit register.
while (VecVT.getSizeInBits() > 64) {
SDValue Lo, Hi;
std::tie(Lo, Hi) = DAG.SplitVector(Vec, DL);
VecVT = Lo.getValueType();
NumElems = VecVT.getVectorNumElements();
Vec = DAG.getNode(ScalarOpcode, DL, VecVT, Lo, Hi);
}
EVT ScalarVT = EVT::getIntegerVT(*DAG.getContext(), VecVT.getSizeInBits());
// Do the remaining work on a scalar since it allows the code generator to
// combine the shift and bitwise operation into one instruction and since
// integer instructions can have higher throughput than vector instructions.
SDValue Scalar = DAG.getBitcast(ScalarVT, Vec);
// Iteratively combine the lower and upper halves of the scalar using the
// bitwise operation, halving the relevant region of the scalar in each
// iteration, until the relevant region is just one element of the original
// vector.
for (unsigned Shift = NumElems / 2; Shift > 0; Shift /= 2) {
SDValue ShiftAmount =
DAG.getConstant(Shift * ElemVT.getSizeInBits(), DL, MVT::i64);
SDValue Shifted =
DAG.getNode(ISD::SRL, DL, ScalarVT, Scalar, ShiftAmount);
Scalar = DAG.getNode(ScalarOpcode, DL, ScalarVT, Scalar, Shifted);
}
Result = DAG.getAnyExtOrTrunc(Scalar, DL, ElemVT);
}
return DAG.getAnyExtOrTrunc(Result, DL, VT);
}
SDValue AArch64TargetLowering::LowerVECREDUCE(SDValue Op,
SelectionDAG &DAG) const {
SDValue Src = Op.getOperand(0);
// Try to lower fixed length reductions to SVE.
EVT SrcVT = Src.getValueType();
bool OverrideNEON = !Subtarget->isNeonAvailable() ||
Op.getOpcode() == ISD::VECREDUCE_AND ||
Op.getOpcode() == ISD::VECREDUCE_OR ||
Op.getOpcode() == ISD::VECREDUCE_XOR ||
Op.getOpcode() == ISD::VECREDUCE_FADD ||
(Op.getOpcode() != ISD::VECREDUCE_ADD &&
SrcVT.getVectorElementType() == MVT::i64);
if (SrcVT.isScalableVector() ||
useSVEForFixedLengthVectorVT(
SrcVT, OverrideNEON && Subtarget->useSVEForFixedLengthVectors())) {
if (SrcVT.getVectorElementType() == MVT::i1)
return LowerPredReductionToSVE(Op, DAG);
switch (Op.getOpcode()) {
case ISD::VECREDUCE_ADD:
return LowerReductionToSVE(AArch64ISD::UADDV_PRED, Op, DAG);
case ISD::VECREDUCE_AND:
return LowerReductionToSVE(AArch64ISD::ANDV_PRED, Op, DAG);
case ISD::VECREDUCE_OR:
return LowerReductionToSVE(AArch64ISD::ORV_PRED, Op, DAG);
case ISD::VECREDUCE_SMAX:
return LowerReductionToSVE(AArch64ISD::SMAXV_PRED, Op, DAG);
case ISD::VECREDUCE_SMIN:
return LowerReductionToSVE(AArch64ISD::SMINV_PRED, Op, DAG);
case ISD::VECREDUCE_UMAX:
return LowerReductionToSVE(AArch64ISD::UMAXV_PRED, Op, DAG);
case ISD::VECREDUCE_UMIN:
return LowerReductionToSVE(AArch64ISD::UMINV_PRED, Op, DAG);
case ISD::VECREDUCE_XOR:
return LowerReductionToSVE(AArch64ISD::EORV_PRED, Op, DAG);
case ISD::VECREDUCE_FADD:
return LowerReductionToSVE(AArch64ISD::FADDV_PRED, Op, DAG);
case ISD::VECREDUCE_FMAX:
return LowerReductionToSVE(AArch64ISD::FMAXNMV_PRED, Op, DAG);
case ISD::VECREDUCE_FMIN:
return LowerReductionToSVE(AArch64ISD::FMINNMV_PRED, Op, DAG);
case ISD::VECREDUCE_FMAXIMUM:
return LowerReductionToSVE(AArch64ISD::FMAXV_PRED, Op, DAG);
case ISD::VECREDUCE_FMINIMUM:
return LowerReductionToSVE(AArch64ISD::FMINV_PRED, Op, DAG);
default:
llvm_unreachable("Unhandled fixed length reduction");
}
}
// Lower NEON reductions.
SDLoc DL(Op);
switch (Op.getOpcode()) {
case ISD::VECREDUCE_AND:
case ISD::VECREDUCE_OR:
case ISD::VECREDUCE_XOR:
return getVectorBitwiseReduce(Op.getOpcode(), Op.getOperand(0),
Op.getValueType(), DL, DAG);
case ISD::VECREDUCE_ADD:
return getReductionSDNode(AArch64ISD::UADDV, DL, Op, DAG);
case ISD::VECREDUCE_SMAX:
return getReductionSDNode(AArch64ISD::SMAXV, DL, Op, DAG);
case ISD::VECREDUCE_SMIN:
return getReductionSDNode(AArch64ISD::SMINV, DL, Op, DAG);
case ISD::VECREDUCE_UMAX:
return getReductionSDNode(AArch64ISD::UMAXV, DL, Op, DAG);
case ISD::VECREDUCE_UMIN:
return getReductionSDNode(AArch64ISD::UMINV, DL, Op, DAG);
default:
llvm_unreachable("Unhandled reduction");
}
}
SDValue AArch64TargetLowering::LowerATOMIC_LOAD_AND(SDValue Op,
SelectionDAG &DAG) const {
auto &Subtarget = DAG.getSubtarget<AArch64Subtarget>();
// No point replacing if we don't have the relevant instruction/libcall anyway
if (!Subtarget.hasLSE() && !Subtarget.outlineAtomics())
return SDValue();
// LSE has an atomic load-clear instruction, but not a load-and.
SDLoc DL(Op);
MVT VT = Op.getSimpleValueType();
assert(VT != MVT::i128 && "Handled elsewhere, code replicated.");
SDValue RHS = Op.getOperand(2);
AtomicSDNode *AN = cast<AtomicSDNode>(Op.getNode());
RHS = DAG.getNode(ISD::XOR, DL, VT, DAG.getAllOnesConstant(DL, VT), RHS);
return DAG.getAtomic(ISD::ATOMIC_LOAD_CLR, DL, AN->getMemoryVT(),
Op.getOperand(0), Op.getOperand(1), RHS,
AN->getMemOperand());
}
SDValue
AArch64TargetLowering::LowerWindowsDYNAMIC_STACKALLOC(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
// Get the inputs.
SDNode *Node = Op.getNode();
SDValue Chain = Op.getOperand(0);
SDValue Size = Op.getOperand(1);
MaybeAlign Align =
cast<ConstantSDNode>(Op.getOperand(2))->getMaybeAlignValue();
EVT VT = Node->getValueType(0);
if (DAG.getMachineFunction().getFunction().hasFnAttribute(
"no-stack-arg-probe")) {
SDValue SP = DAG.getCopyFromReg(Chain, DL, AArch64::SP, MVT::i64);
Chain = SP.getValue(1);
SP = DAG.getNode(ISD::SUB, DL, MVT::i64, SP, Size);
if (Align)
SP =
DAG.getNode(ISD::AND, DL, VT, SP.getValue(0),
DAG.getSignedConstant(-(uint64_t)Align->value(), DL, VT));
Chain = DAG.getCopyToReg(Chain, DL, AArch64::SP, SP);
SDValue Ops[2] = {SP, Chain};
return DAG.getMergeValues(Ops, DL);
}
Chain = DAG.getCALLSEQ_START(Chain, 0, 0, DL);
EVT PtrVT = getPointerTy(DAG.getDataLayout());
SDValue Callee = DAG.getTargetExternalSymbol(Subtarget->getChkStkName(),
PtrVT, 0);
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
const uint32_t *Mask = TRI->getWindowsStackProbePreservedMask();
if (Subtarget->hasCustomCallingConv())
TRI->UpdateCustomCallPreservedMask(DAG.getMachineFunction(), &Mask);
Size = DAG.getNode(ISD::SRL, DL, MVT::i64, Size,
DAG.getConstant(4, DL, MVT::i64));
Chain = DAG.getCopyToReg(Chain, DL, AArch64::X15, Size, SDValue());
Chain =
DAG.getNode(AArch64ISD::CALL, DL, DAG.getVTList(MVT::Other, MVT::Glue),
Chain, Callee, DAG.getRegister(AArch64::X15, MVT::i64),
DAG.getRegisterMask(Mask), Chain.getValue(1));
// To match the actual intent better, we should read the output from X15 here
// again (instead of potentially spilling it to the stack), but rereading Size
// from X15 here doesn't work at -O0, since it thinks that X15 is undefined
// here.
Size = DAG.getNode(ISD::SHL, DL, MVT::i64, Size,
DAG.getConstant(4, DL, MVT::i64));
SDValue SP = DAG.getCopyFromReg(Chain, DL, AArch64::SP, MVT::i64);
Chain = SP.getValue(1);
SP = DAG.getNode(ISD::SUB, DL, MVT::i64, SP, Size);
if (Align)
SP = DAG.getNode(ISD::AND, DL, VT, SP.getValue(0),
DAG.getSignedConstant(-(uint64_t)Align->value(), DL, VT));
Chain = DAG.getCopyToReg(Chain, DL, AArch64::SP, SP);
Chain = DAG.getCALLSEQ_END(Chain, 0, 0, SDValue(), DL);
SDValue Ops[2] = {SP, Chain};
return DAG.getMergeValues(Ops, DL);
}
SDValue
AArch64TargetLowering::LowerInlineDYNAMIC_STACKALLOC(SDValue Op,
SelectionDAG &DAG) const {
// Get the inputs.
SDNode *Node = Op.getNode();
SDValue Chain = Op.getOperand(0);
SDValue Size = Op.getOperand(1);
MaybeAlign Align =
cast<ConstantSDNode>(Op.getOperand(2))->getMaybeAlignValue();
SDLoc DL(Op);
EVT VT = Node->getValueType(0);
// Construct the new SP value in a GPR.
SDValue SP = DAG.getCopyFromReg(Chain, DL, AArch64::SP, MVT::i64);
Chain = SP.getValue(1);
SP = DAG.getNode(ISD::SUB, DL, MVT::i64, SP, Size);
if (Align)
SP = DAG.getNode(ISD::AND, DL, VT, SP.getValue(0),
DAG.getSignedConstant(-(uint64_t)Align->value(), DL, VT));
// Set the real SP to the new value with a probing loop.
Chain = DAG.getNode(AArch64ISD::PROBED_ALLOCA, DL, MVT::Other, Chain, SP);
SDValue Ops[2] = {SP, Chain};
return DAG.getMergeValues(Ops, DL);
}
SDValue
AArch64TargetLowering::LowerDYNAMIC_STACKALLOC(SDValue Op,
SelectionDAG &DAG) const {
MachineFunction &MF = DAG.getMachineFunction();
if (Subtarget->isTargetWindows())
return LowerWindowsDYNAMIC_STACKALLOC(Op, DAG);
else if (hasInlineStackProbe(MF))
return LowerInlineDYNAMIC_STACKALLOC(Op, DAG);
else
return SDValue();
}
SDValue AArch64TargetLowering::LowerAVG(SDValue Op, SelectionDAG &DAG,
unsigned NewOp) const {
if (Subtarget->hasSVE2())
return LowerToPredicatedOp(Op, DAG, NewOp);
// Default to expand.
return SDValue();
}
SDValue AArch64TargetLowering::LowerVSCALE(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT != MVT::i64 && "Expected illegal VSCALE node");
SDLoc DL(Op);
APInt MulImm = Op.getConstantOperandAPInt(0);
return DAG.getZExtOrTrunc(DAG.getVScale(DL, MVT::i64, MulImm.sext(64)), DL,
VT);
}
/// Set the IntrinsicInfo for the `aarch64_sve_st<N>` intrinsics.
template <unsigned NumVecs>
static bool
setInfoSVEStN(const AArch64TargetLowering &TLI, const DataLayout &DL,
AArch64TargetLowering::IntrinsicInfo &Info, const CallInst &CI) {
Info.opc = ISD::INTRINSIC_VOID;
// Retrieve EC from first vector argument.
const EVT VT = TLI.getMemValueType(DL, CI.getArgOperand(0)->getType());
ElementCount EC = VT.getVectorElementCount();
#ifndef NDEBUG
// Check the assumption that all input vectors are the same type.
for (unsigned I = 0; I < NumVecs; ++I)
assert(VT == TLI.getMemValueType(DL, CI.getArgOperand(I)->getType()) &&
"Invalid type.");
#endif
// memVT is `NumVecs * VT`.
Info.memVT = EVT::getVectorVT(CI.getType()->getContext(), VT.getScalarType(),
EC * NumVecs);
Info.ptrVal = CI.getArgOperand(CI.arg_size() - 1);
Info.offset = 0;
Info.align.reset();
Info.flags = MachineMemOperand::MOStore;
return true;
}
/// getTgtMemIntrinsic - Represent NEON load and store intrinsics as
/// MemIntrinsicNodes. The associated MachineMemOperands record the alignment
/// specified in the intrinsic calls.
bool AArch64TargetLowering::getTgtMemIntrinsic(IntrinsicInfo &Info,
const CallInst &I,
MachineFunction &MF,
unsigned Intrinsic) const {
auto &DL = I.getDataLayout();
switch (Intrinsic) {
case Intrinsic::aarch64_sve_st2:
return setInfoSVEStN<2>(*this, DL, Info, I);
case Intrinsic::aarch64_sve_st3:
return setInfoSVEStN<3>(*this, DL, Info, I);
case Intrinsic::aarch64_sve_st4:
return setInfoSVEStN<4>(*this, DL, Info, I);
case Intrinsic::aarch64_neon_ld2:
case Intrinsic::aarch64_neon_ld3:
case Intrinsic::aarch64_neon_ld4:
case Intrinsic::aarch64_neon_ld1x2:
case Intrinsic::aarch64_neon_ld1x3:
case Intrinsic::aarch64_neon_ld1x4: {
Info.opc = ISD::INTRINSIC_W_CHAIN;
uint64_t NumElts = DL.getTypeSizeInBits(I.getType()) / 64;
Info.memVT = EVT::getVectorVT(I.getType()->getContext(), MVT::i64, NumElts);
Info.ptrVal = I.getArgOperand(I.arg_size() - 1);
Info.offset = 0;
Info.align.reset();
// volatile loads with NEON intrinsics not supported
Info.flags = MachineMemOperand::MOLoad;
return true;
}
case Intrinsic::aarch64_neon_ld2lane:
case Intrinsic::aarch64_neon_ld3lane:
case Intrinsic::aarch64_neon_ld4lane:
case Intrinsic::aarch64_neon_ld2r:
case Intrinsic::aarch64_neon_ld3r:
case Intrinsic::aarch64_neon_ld4r: {
Info.opc = ISD::INTRINSIC_W_CHAIN;
// ldx return struct with the same vec type
Type *RetTy = I.getType();
auto *StructTy = cast<StructType>(RetTy);
unsigned NumElts = StructTy->getNumElements();
Type *VecTy = StructTy->getElementType(0);
MVT EleVT = MVT::getVT(VecTy).getVectorElementType();
Info.memVT = EVT::getVectorVT(I.getType()->getContext(), EleVT, NumElts);
Info.ptrVal = I.getArgOperand(I.arg_size() - 1);
Info.offset = 0;
Info.align.reset();
// volatile loads with NEON intrinsics not supported
Info.flags = MachineMemOperand::MOLoad;
return true;
}
case Intrinsic::aarch64_neon_st2:
case Intrinsic::aarch64_neon_st3:
case Intrinsic::aarch64_neon_st4:
case Intrinsic::aarch64_neon_st1x2:
case Intrinsic::aarch64_neon_st1x3:
case Intrinsic::aarch64_neon_st1x4: {
Info.opc = ISD::INTRINSIC_VOID;
unsigned NumElts = 0;
for (const Value *Arg : I.args()) {
Type *ArgTy = Arg->getType();
if (!ArgTy->isVectorTy())
break;
NumElts += DL.getTypeSizeInBits(ArgTy) / 64;
}
Info.memVT = EVT::getVectorVT(I.getType()->getContext(), MVT::i64, NumElts);
Info.ptrVal = I.getArgOperand(I.arg_size() - 1);
Info.offset = 0;
Info.align.reset();
// volatile stores with NEON intrinsics not supported
Info.flags = MachineMemOperand::MOStore;
return true;
}
case Intrinsic::aarch64_neon_st2lane:
case Intrinsic::aarch64_neon_st3lane:
case Intrinsic::aarch64_neon_st4lane: {
Info.opc = ISD::INTRINSIC_VOID;
unsigned NumElts = 0;
// all the vector type is same
Type *VecTy = I.getArgOperand(0)->getType();
MVT EleVT = MVT::getVT(VecTy).getVectorElementType();
for (const Value *Arg : I.args()) {
Type *ArgTy = Arg->getType();
if (!ArgTy->isVectorTy())
break;
NumElts += 1;
}
Info.memVT = EVT::getVectorVT(I.getType()->getContext(), EleVT, NumElts);
Info.ptrVal = I.getArgOperand(I.arg_size() - 1);
Info.offset = 0;
Info.align.reset();
// volatile stores with NEON intrinsics not supported
Info.flags = MachineMemOperand::MOStore;
return true;
}
case Intrinsic::aarch64_ldaxr:
case Intrinsic::aarch64_ldxr: {
Type *ValTy = I.getParamElementType(0);
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::getVT(ValTy);
Info.ptrVal = I.getArgOperand(0);
Info.offset = 0;
Info.align = DL.getABITypeAlign(ValTy);
Info.flags = MachineMemOperand::MOLoad | MachineMemOperand::MOVolatile;
return true;
}
case Intrinsic::aarch64_stlxr:
case Intrinsic::aarch64_stxr: {
Type *ValTy = I.getParamElementType(1);
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::getVT(ValTy);
Info.ptrVal = I.getArgOperand(1);
Info.offset = 0;
Info.align = DL.getABITypeAlign(ValTy);
Info.flags = MachineMemOperand::MOStore | MachineMemOperand::MOVolatile;
return true;
}
case Intrinsic::aarch64_ldaxp:
case Intrinsic::aarch64_ldxp:
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::i128;
Info.ptrVal = I.getArgOperand(0);
Info.offset = 0;
Info.align = Align(16);
Info.flags = MachineMemOperand::MOLoad | MachineMemOperand::MOVolatile;
return true;
case Intrinsic::aarch64_stlxp:
case Intrinsic::aarch64_stxp:
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::i128;
Info.ptrVal = I.getArgOperand(2);
Info.offset = 0;
Info.align = Align(16);
Info.flags = MachineMemOperand::MOStore | MachineMemOperand::MOVolatile;
return true;
case Intrinsic::aarch64_sve_ldnt1: {
Type *ElTy = cast<VectorType>(I.getType())->getElementType();
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::getVT(I.getType());
Info.ptrVal = I.getArgOperand(1);
Info.offset = 0;
Info.align = DL.getABITypeAlign(ElTy);
Info.flags = MachineMemOperand::MOLoad | MachineMemOperand::MONonTemporal;
return true;
}
case Intrinsic::aarch64_sve_stnt1: {
Type *ElTy =
cast<VectorType>(I.getArgOperand(0)->getType())->getElementType();
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::getVT(I.getOperand(0)->getType());
Info.ptrVal = I.getArgOperand(2);
Info.offset = 0;
Info.align = DL.getABITypeAlign(ElTy);
Info.flags = MachineMemOperand::MOStore | MachineMemOperand::MONonTemporal;
return true;
}
case Intrinsic::aarch64_mops_memset_tag: {
Value *Dst = I.getArgOperand(0);
Value *Val = I.getArgOperand(1);
Info.opc = ISD::INTRINSIC_W_CHAIN;
Info.memVT = MVT::getVT(Val->getType());
Info.ptrVal = Dst;
Info.offset = 0;
Info.align = I.getParamAlign(0).valueOrOne();
Info.flags = MachineMemOperand::MOStore;
// The size of the memory being operated on is unknown at this point
Info.size = MemoryLocation::UnknownSize;
return true;
}
default:
break;
}
return false;
}
bool AArch64TargetLowering::shouldReduceLoadWidth(
SDNode *Load, ISD::LoadExtType ExtTy, EVT NewVT,
std::optional<unsigned> ByteOffset) const {
// TODO: This may be worth removing. Check regression tests for diffs.
if (!TargetLoweringBase::shouldReduceLoadWidth(Load, ExtTy, NewVT,
ByteOffset))
return false;
// If we're reducing the load width in order to avoid having to use an extra
// instruction to do extension then it's probably a good idea.
if (ExtTy != ISD::NON_EXTLOAD)
return true;
// Don't reduce load width if it would prevent us from combining a shift into
// the offset.
MemSDNode *Mem = dyn_cast<MemSDNode>(Load);
assert(Mem);
const SDValue &Base = Mem->getBasePtr();
if (Base.getOpcode() == ISD::ADD &&
Base.getOperand(1).getOpcode() == ISD::SHL &&
Base.getOperand(1).hasOneUse() &&
Base.getOperand(1).getOperand(1).getOpcode() == ISD::Constant) {
// It's unknown whether a scalable vector has a power-of-2 bitwidth.
if (Mem->getMemoryVT().isScalableVector())
return false;
// The shift can be combined if it matches the size of the value being
// loaded (and so reducing the width would make it not match).
uint64_t ShiftAmount = Base.getOperand(1).getConstantOperandVal(1);
uint64_t LoadBytes = Mem->getMemoryVT().getSizeInBits()/8;
if (ShiftAmount == Log2_32(LoadBytes))
return false;
}
// We have no reason to disallow reducing the load width, so allow it.
return true;
}
// Treat a sext_inreg(extract(..)) as free if it has multiple uses.
bool AArch64TargetLowering::shouldRemoveRedundantExtend(SDValue Extend) const {
EVT VT = Extend.getValueType();
if ((VT == MVT::i64 || VT == MVT::i32) && Extend->use_size()) {
SDValue Extract = Extend.getOperand(0);
if (Extract.getOpcode() == ISD::ANY_EXTEND && Extract.hasOneUse())
Extract = Extract.getOperand(0);
if (Extract.getOpcode() == ISD::EXTRACT_VECTOR_ELT && Extract.hasOneUse()) {
EVT VecVT = Extract.getOperand(0).getValueType();
if (VecVT.getScalarType() == MVT::i8 || VecVT.getScalarType() == MVT::i16)
return false;
}
}
return true;
}
// Truncations from 64-bit GPR to 32-bit GPR is free.
bool AArch64TargetLowering::isTruncateFree(Type *Ty1, Type *Ty2) const {
if (!Ty1->isIntegerTy() || !Ty2->isIntegerTy())
return false;
uint64_t NumBits1 = Ty1->getPrimitiveSizeInBits().getFixedValue();
uint64_t NumBits2 = Ty2->getPrimitiveSizeInBits().getFixedValue();
return NumBits1 > NumBits2;
}
bool AArch64TargetLowering::isTruncateFree(EVT VT1, EVT VT2) const {
if (VT1.isVector() || VT2.isVector() || !VT1.isInteger() || !VT2.isInteger())
return false;
uint64_t NumBits1 = VT1.getFixedSizeInBits();
uint64_t NumBits2 = VT2.getFixedSizeInBits();
return NumBits1 > NumBits2;
}
/// Check if it is profitable to hoist instruction in then/else to if.
/// Not profitable if I and it's user can form a FMA instruction
/// because we prefer FMSUB/FMADD.
bool AArch64TargetLowering::isProfitableToHoist(Instruction *I) const {
if (I->getOpcode() != Instruction::FMul)
return true;
if (!I->hasOneUse())
return true;
Instruction *User = I->user_back();
if (!(User->getOpcode() == Instruction::FSub ||
User->getOpcode() == Instruction::FAdd))
return true;
const TargetOptions &Options = getTargetMachine().Options;
const Function *F = I->getFunction();
const DataLayout &DL = F->getDataLayout();
Type *Ty = User->getOperand(0)->getType();
return !(isFMAFasterThanFMulAndFAdd(*F, Ty) &&
isOperationLegalOrCustom(ISD::FMA, getValueType(DL, Ty)) &&
(Options.AllowFPOpFusion == FPOpFusion::Fast ||
Options.UnsafeFPMath));
}
// All 32-bit GPR operations implicitly zero the high-half of the corresponding
// 64-bit GPR.
bool AArch64TargetLowering::isZExtFree(Type *Ty1, Type *Ty2) const {
if (!Ty1->isIntegerTy() || !Ty2->isIntegerTy())
return false;
unsigned NumBits1 = Ty1->getPrimitiveSizeInBits();
unsigned NumBits2 = Ty2->getPrimitiveSizeInBits();
return NumBits1 == 32 && NumBits2 == 64;
}
bool AArch64TargetLowering::isZExtFree(EVT VT1, EVT VT2) const {
if (VT1.isVector() || VT2.isVector() || !VT1.isInteger() || !VT2.isInteger())
return false;
unsigned NumBits1 = VT1.getSizeInBits();
unsigned NumBits2 = VT2.getSizeInBits();
return NumBits1 == 32 && NumBits2 == 64;
}
bool AArch64TargetLowering::isZExtFree(SDValue Val, EVT VT2) const {
EVT VT1 = Val.getValueType();
if (isZExtFree(VT1, VT2)) {
return true;
}
if (Val.getOpcode() != ISD::LOAD)
return false;
// 8-, 16-, and 32-bit integer loads all implicitly zero-extend.
return (VT1.isSimple() && !VT1.isVector() && VT1.isInteger() &&
VT2.isSimple() && !VT2.isVector() && VT2.isInteger() &&
VT1.getSizeInBits() <= 32);
}
bool AArch64TargetLowering::isExtFreeImpl(const Instruction *Ext) const {
if (isa<FPExtInst>(Ext))
return false;
// Vector types are not free.
if (Ext->getType()->isVectorTy())
return false;
for (const Use &U : Ext->uses()) {
// The extension is free if we can fold it with a left shift in an
// addressing mode or an arithmetic operation: add, sub, and cmp.
// Is there a shift?
const Instruction *Instr = cast<Instruction>(U.getUser());
// Is this a constant shift?
switch (Instr->getOpcode()) {
case Instruction::Shl:
if (!isa<ConstantInt>(Instr->getOperand(1)))
return false;
break;
case Instruction::GetElementPtr: {
gep_type_iterator GTI = gep_type_begin(Instr);
auto &DL = Ext->getDataLayout();
std::advance(GTI, U.getOperandNo()-1);
Type *IdxTy = GTI.getIndexedType();
// This extension will end up with a shift because of the scaling factor.
// 8-bit sized types have a scaling factor of 1, thus a shift amount of 0.
// Get the shift amount based on the scaling factor:
// log2(sizeof(IdxTy)) - log2(8).
if (IdxTy->isScalableTy())
return false;
uint64_t ShiftAmt =
llvm::countr_zero(DL.getTypeStoreSizeInBits(IdxTy).getFixedValue()) -
3;
// Is the constant foldable in the shift of the addressing mode?
// I.e., shift amount is between 1 and 4 inclusive.
if (ShiftAmt == 0 || ShiftAmt > 4)
return false;
break;
}
case Instruction::Trunc:
// Check if this is a noop.
// trunc(sext ty1 to ty2) to ty1.
if (Instr->getType() == Ext->getOperand(0)->getType())
continue;
[[fallthrough]];
default:
return false;
}
// At this point we can use the bfm family, so this extension is free
// for that use.
}
return true;
}
static bool createTblShuffleMask(unsigned SrcWidth, unsigned DstWidth,
unsigned NumElts, bool IsLittleEndian,
SmallVectorImpl<int> &Mask) {
if (DstWidth % 8 != 0 || DstWidth <= 16 || DstWidth > 64)
return false;
assert(DstWidth % SrcWidth == 0 &&
"TBL lowering is not supported for a conversion instruction with this "
"source and destination element type.");
unsigned Factor = DstWidth / SrcWidth;
unsigned MaskLen = NumElts * Factor;
Mask.clear();
Mask.resize(MaskLen, NumElts);
unsigned SrcIndex = 0;
for (unsigned I = IsLittleEndian ? 0 : Factor - 1; I < MaskLen; I += Factor)
Mask[I] = SrcIndex++;
return true;
}
static Value *createTblShuffleForZExt(IRBuilderBase &Builder, Value *Op,
FixedVectorType *ZExtTy,
FixedVectorType *DstTy,
bool IsLittleEndian) {
auto *SrcTy = cast<FixedVectorType>(Op->getType());
unsigned NumElts = SrcTy->getNumElements();
auto SrcWidth = cast<IntegerType>(SrcTy->getElementType())->getBitWidth();
auto DstWidth = cast<IntegerType>(DstTy->getElementType())->getBitWidth();
SmallVector<int> Mask;
if (!createTblShuffleMask(SrcWidth, DstWidth, NumElts, IsLittleEndian, Mask))
return nullptr;
auto *FirstEltZero = Builder.CreateInsertElement(
PoisonValue::get(SrcTy), Builder.getIntN(SrcWidth, 0), uint64_t(0));
Value *Result = Builder.CreateShuffleVector(Op, FirstEltZero, Mask);
Result = Builder.CreateBitCast(Result, DstTy);
if (DstTy != ZExtTy)
Result = Builder.CreateZExt(Result, ZExtTy);
return Result;
}
static Value *createTblShuffleForSExt(IRBuilderBase &Builder, Value *Op,
FixedVectorType *DstTy,
bool IsLittleEndian) {
auto *SrcTy = cast<FixedVectorType>(Op->getType());
auto SrcWidth = cast<IntegerType>(SrcTy->getElementType())->getBitWidth();
auto DstWidth = cast<IntegerType>(DstTy->getElementType())->getBitWidth();
SmallVector<int> Mask;
if (!createTblShuffleMask(SrcWidth, DstWidth, SrcTy->getNumElements(),
!IsLittleEndian, Mask))
return nullptr;
auto *FirstEltZero = Builder.CreateInsertElement(
PoisonValue::get(SrcTy), Builder.getIntN(SrcWidth, 0), uint64_t(0));
return Builder.CreateShuffleVector(Op, FirstEltZero, Mask);
}
static void createTblForTrunc(TruncInst *TI, bool IsLittleEndian) {
IRBuilder<> Builder(TI);
SmallVector<Value *> Parts;
int NumElements = cast<FixedVectorType>(TI->getType())->getNumElements();
auto *SrcTy = cast<FixedVectorType>(TI->getOperand(0)->getType());
auto *DstTy = cast<FixedVectorType>(TI->getType());
assert(SrcTy->getElementType()->isIntegerTy() &&
"Non-integer type source vector element is not supported");
assert(DstTy->getElementType()->isIntegerTy(8) &&
"Unsupported destination vector element type");
unsigned SrcElemTySz =
cast<IntegerType>(SrcTy->getElementType())->getBitWidth();
unsigned DstElemTySz =
cast<IntegerType>(DstTy->getElementType())->getBitWidth();
assert((SrcElemTySz % DstElemTySz == 0) &&
"Cannot lower truncate to tbl instructions for a source element size "
"that is not divisible by the destination element size");
unsigned TruncFactor = SrcElemTySz / DstElemTySz;
assert((SrcElemTySz == 16 || SrcElemTySz == 32 || SrcElemTySz == 64) &&
"Unsupported source vector element type size");
Type *VecTy = FixedVectorType::get(Builder.getInt8Ty(), 16);
// Create a mask to choose every nth byte from the source vector table of
// bytes to create the truncated destination vector, where 'n' is the truncate
// ratio. For example, for a truncate from Yxi64 to Yxi8, choose
// 0,8,16,..Y*8th bytes for the little-endian format
SmallVector<Constant *, 16> MaskConst;
for (int Itr = 0; Itr < 16; Itr++) {
if (Itr < NumElements)
MaskConst.push_back(Builder.getInt8(
IsLittleEndian ? Itr * TruncFactor
: Itr * TruncFactor + (TruncFactor - 1)));
else
MaskConst.push_back(Builder.getInt8(255));
}
int MaxTblSz = 128 * 4;
int MaxSrcSz = SrcElemTySz * NumElements;
int ElemsPerTbl =
(MaxTblSz > MaxSrcSz) ? NumElements : (MaxTblSz / SrcElemTySz);
assert(ElemsPerTbl <= 16 &&
"Maximum elements selected using TBL instruction cannot exceed 16!");
int ShuffleCount = 128 / SrcElemTySz;
SmallVector<int> ShuffleLanes;
for (int i = 0; i < ShuffleCount; ++i)
ShuffleLanes.push_back(i);
// Create TBL's table of bytes in 1,2,3 or 4 FP/SIMD registers using shuffles
// over the source vector. If TBL's maximum 4 FP/SIMD registers are saturated,
// call TBL & save the result in a vector of TBL results for combining later.
SmallVector<Value *> Results;
while (ShuffleLanes.back() < NumElements) {
Parts.push_back(Builder.CreateBitCast(
Builder.CreateShuffleVector(TI->getOperand(0), ShuffleLanes), VecTy));
if (Parts.size() == 4) {
Parts.push_back(ConstantVector::get(MaskConst));
Results.push_back(
Builder.CreateIntrinsic(Intrinsic::aarch64_neon_tbl4, VecTy, Parts));
Parts.clear();
}
for (int i = 0; i < ShuffleCount; ++i)
ShuffleLanes[i] += ShuffleCount;
}
assert((Parts.empty() || Results.empty()) &&
"Lowering trunc for vectors requiring different TBL instructions is "
"not supported!");
// Call TBL for the residual table bytes present in 1,2, or 3 FP/SIMD
// registers
if (!Parts.empty()) {
Intrinsic::ID TblID;
switch (Parts.size()) {
case 1:
TblID = Intrinsic::aarch64_neon_tbl1;
break;
case 2:
TblID = Intrinsic::aarch64_neon_tbl2;
break;
case 3:
TblID = Intrinsic::aarch64_neon_tbl3;
break;
}
Parts.push_back(ConstantVector::get(MaskConst));
Results.push_back(Builder.CreateIntrinsic(TblID, VecTy, Parts));
}
// Extract the destination vector from TBL result(s) after combining them
// where applicable. Currently, at most two TBLs are supported.
assert(Results.size() <= 2 && "Trunc lowering does not support generation of "
"more than 2 tbl instructions!");
Value *FinalResult = Results[0];
if (Results.size() == 1) {
if (ElemsPerTbl < 16) {
SmallVector<int> FinalMask(ElemsPerTbl);
std::iota(FinalMask.begin(), FinalMask.end(), 0);
FinalResult = Builder.CreateShuffleVector(Results[0], FinalMask);
}
} else {
SmallVector<int> FinalMask(ElemsPerTbl * Results.size());
if (ElemsPerTbl < 16) {
std::iota(FinalMask.begin(), FinalMask.begin() + ElemsPerTbl, 0);
std::iota(FinalMask.begin() + ElemsPerTbl, FinalMask.end(), 16);
} else {
std::iota(FinalMask.begin(), FinalMask.end(), 0);
}
FinalResult =
Builder.CreateShuffleVector(Results[0], Results[1], FinalMask);
}
TI->replaceAllUsesWith(FinalResult);
TI->eraseFromParent();
}
bool AArch64TargetLowering::optimizeExtendOrTruncateConversion(
Instruction *I, Loop *L, const TargetTransformInfo &TTI) const {
// shuffle_vector instructions are serialized when targeting SVE,
// see LowerSPLAT_VECTOR. This peephole is not beneficial.
if (!EnableExtToTBL || Subtarget->useSVEForFixedLengthVectors())
return false;
// Try to optimize conversions using tbl. This requires materializing constant
// index vectors, which can increase code size and add loads. Skip the
// transform unless the conversion is in a loop block guaranteed to execute
// and we are not optimizing for size.
Function *F = I->getParent()->getParent();
if (!L || L->getHeader() != I->getParent() || F->hasOptSize())
return false;
auto *SrcTy = dyn_cast<FixedVectorType>(I->getOperand(0)->getType());
auto *DstTy = dyn_cast<FixedVectorType>(I->getType());
if (!SrcTy || !DstTy)
return false;
// Convert 'zext <Y x i8> %x to <Y x i8X>' to a shuffle that can be
// lowered to tbl instructions to insert the original i8 elements
// into i8x lanes. This is enabled for cases where it is beneficial.
auto *ZExt = dyn_cast<ZExtInst>(I);
if (ZExt && SrcTy->getElementType()->isIntegerTy(8)) {
auto DstWidth = DstTy->getElementType()->getScalarSizeInBits();
if (DstWidth % 8 != 0)
return false;
auto *TruncDstType =
cast<FixedVectorType>(VectorType::getTruncatedElementVectorType(DstTy));
// If the ZExt can be lowered to a single ZExt to the next power-of-2 and
// the remaining ZExt folded into the user, don't use tbl lowering.
auto SrcWidth = SrcTy->getElementType()->getScalarSizeInBits();
if (TTI.getCastInstrCost(I->getOpcode(), DstTy, TruncDstType,
TargetTransformInfo::getCastContextHint(I),
TTI::TCK_SizeAndLatency, I) == TTI::TCC_Free) {
if (SrcWidth * 2 >= TruncDstType->getElementType()->getScalarSizeInBits())
return false;
DstTy = TruncDstType;
}
// mul(zext(i8), sext) can be transformed into smull(zext, sext) which
// performs one extend implicitly. If DstWidth is at most 4 * SrcWidth, at
// most one extra extend step is needed and using tbl is not profitable.
// Similarly, bail out if partial_reduce(acc, zext(i8)) can be lowered to a
// udot instruction.
if (SrcWidth * 4 <= DstWidth) {
if (all_of(I->users(), [&](auto *U) {
auto *SingleUser = cast<Instruction>(&*U);
if (match(SingleUser, m_c_Mul(m_Specific(I), m_SExt(m_Value()))))
return true;
if (match(SingleUser,
m_Intrinsic<
Intrinsic::experimental_vector_partial_reduce_add>(
m_Value(), m_Specific(I))))
return true;
return false;
}))
return false;
}
if (DstTy->getScalarSizeInBits() >= 64)
return false;
IRBuilder<> Builder(ZExt);
Value *Result = createTblShuffleForZExt(
Builder, ZExt->getOperand(0), cast<FixedVectorType>(ZExt->getType()),
DstTy, Subtarget->isLittleEndian());
if (!Result)
return false;
ZExt->replaceAllUsesWith(Result);
ZExt->eraseFromParent();
return true;
}
auto *UIToFP = dyn_cast<UIToFPInst>(I);
if (UIToFP && ((SrcTy->getElementType()->isIntegerTy(8) &&
DstTy->getElementType()->isFloatTy()) ||
(SrcTy->getElementType()->isIntegerTy(16) &&
DstTy->getElementType()->isDoubleTy()))) {
IRBuilder<> Builder(I);
Value *ZExt = createTblShuffleForZExt(
Builder, I->getOperand(0), FixedVectorType::getInteger(DstTy),
FixedVectorType::getInteger(DstTy), Subtarget->isLittleEndian());
assert(ZExt && "Cannot fail for the i8 to float conversion");
auto *UI = Builder.CreateUIToFP(ZExt, DstTy);
I->replaceAllUsesWith(UI);
I->eraseFromParent();
return true;
}
auto *SIToFP = dyn_cast<SIToFPInst>(I);
if (SIToFP && SrcTy->getElementType()->isIntegerTy(8) &&
DstTy->getElementType()->isFloatTy()) {
IRBuilder<> Builder(I);
auto *Shuffle = createTblShuffleForSExt(Builder, I->getOperand(0),
FixedVectorType::getInteger(DstTy),
Subtarget->isLittleEndian());
assert(Shuffle && "Cannot fail for the i8 to float conversion");
auto *Cast = Builder.CreateBitCast(Shuffle, VectorType::getInteger(DstTy));
auto *AShr = Builder.CreateAShr(Cast, 24, "", true);
auto *SI = Builder.CreateSIToFP(AShr, DstTy);
I->replaceAllUsesWith(SI);
I->eraseFromParent();
return true;
}
// Convert 'fptoui <(8|16) x float> to <(8|16) x i8>' to a wide fptoui
// followed by a truncate lowered to using tbl.4.
auto *FPToUI = dyn_cast<FPToUIInst>(I);
if (FPToUI &&
(SrcTy->getNumElements() == 8 || SrcTy->getNumElements() == 16) &&
SrcTy->getElementType()->isFloatTy() &&
DstTy->getElementType()->isIntegerTy(8)) {
IRBuilder<> Builder(I);
auto *WideConv = Builder.CreateFPToUI(FPToUI->getOperand(0),
VectorType::getInteger(SrcTy));
auto *TruncI = Builder.CreateTrunc(WideConv, DstTy);
I->replaceAllUsesWith(TruncI);
I->eraseFromParent();
createTblForTrunc(cast<TruncInst>(TruncI), Subtarget->isLittleEndian());
return true;
}
// Convert 'trunc <(8|16) x (i32|i64)> %x to <(8|16) x i8>' to an appropriate
// tbl instruction selecting the lowest/highest (little/big endian) 8 bits
// per lane of the input that is represented using 1,2,3 or 4 128-bit table
// registers
auto *TI = dyn_cast<TruncInst>(I);
if (TI && DstTy->getElementType()->isIntegerTy(8) &&
((SrcTy->getElementType()->isIntegerTy(32) ||
SrcTy->getElementType()->isIntegerTy(64)) &&
(SrcTy->getNumElements() == 16 || SrcTy->getNumElements() == 8))) {
createTblForTrunc(TI, Subtarget->isLittleEndian());
return true;
}
return false;
}
bool AArch64TargetLowering::hasPairedLoad(EVT LoadedType,
Align &RequiredAlignment) const {
if (!LoadedType.isSimple() ||
(!LoadedType.isInteger() && !LoadedType.isFloatingPoint()))
return false;
// Cyclone supports unaligned accesses.
RequiredAlignment = Align(1);
unsigned NumBits = LoadedType.getSizeInBits();
return NumBits == 32 || NumBits == 64;
}
/// A helper function for determining the number of interleaved accesses we
/// will generate when lowering accesses of the given type.
unsigned AArch64TargetLowering::getNumInterleavedAccesses(
VectorType *VecTy, const DataLayout &DL, bool UseScalable) const {
unsigned VecSize = 128;
unsigned ElSize = DL.getTypeSizeInBits(VecTy->getElementType());
unsigned MinElts = VecTy->getElementCount().getKnownMinValue();
if (UseScalable && isa<FixedVectorType>(VecTy))
VecSize = std::max(Subtarget->getMinSVEVectorSizeInBits(), 128u);
return std::max<unsigned>(1, (MinElts * ElSize + 127) / VecSize);
}
MachineMemOperand::Flags
AArch64TargetLowering::getTargetMMOFlags(const Instruction &I) const {
if (Subtarget->getProcFamily() == AArch64Subtarget::Falkor &&
I.hasMetadata(FALKOR_STRIDED_ACCESS_MD))
return MOStridedAccess;
return MachineMemOperand::MONone;
}
bool AArch64TargetLowering::isLegalInterleavedAccessType(
VectorType *VecTy, const DataLayout &DL, bool &UseScalable) const {
unsigned ElSize = DL.getTypeSizeInBits(VecTy->getElementType());
auto EC = VecTy->getElementCount();
unsigned MinElts = EC.getKnownMinValue();
UseScalable = false;
if (isa<FixedVectorType>(VecTy) && !Subtarget->isNeonAvailable() &&
(!Subtarget->useSVEForFixedLengthVectors() ||
!getSVEPredPatternFromNumElements(MinElts)))
return false;
if (isa<ScalableVectorType>(VecTy) &&
!Subtarget->isSVEorStreamingSVEAvailable())
return false;
// Ensure the number of vector elements is greater than 1.
if (MinElts < 2)
return false;
// Ensure the element type is legal.
if (ElSize != 8 && ElSize != 16 && ElSize != 32 && ElSize != 64)
return false;
if (EC.isScalable()) {
UseScalable = true;
return isPowerOf2_32(MinElts) && (MinElts * ElSize) % 128 == 0;
}
unsigned VecSize = DL.getTypeSizeInBits(VecTy);
if (Subtarget->useSVEForFixedLengthVectors()) {
unsigned MinSVEVectorSize =
std::max(Subtarget->getMinSVEVectorSizeInBits(), 128u);
if (VecSize % MinSVEVectorSize == 0 ||
(VecSize < MinSVEVectorSize && isPowerOf2_32(MinElts) &&
(!Subtarget->isNeonAvailable() || VecSize > 128))) {
UseScalable = true;
return true;
}
}
// Ensure the total vector size is 64 or a multiple of 128. Types larger than
// 128 will be split into multiple interleaved accesses.
return Subtarget->isNeonAvailable() && (VecSize == 64 || VecSize % 128 == 0);
}
static ScalableVectorType *getSVEContainerIRType(FixedVectorType *VTy) {
if (VTy->getElementType() == Type::getDoubleTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 2);
if (VTy->getElementType() == Type::getFloatTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 4);
if (VTy->getElementType() == Type::getBFloatTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 8);
if (VTy->getElementType() == Type::getHalfTy(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 8);
if (VTy->getElementType() == Type::getInt64Ty(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 2);
if (VTy->getElementType() == Type::getInt32Ty(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 4);
if (VTy->getElementType() == Type::getInt16Ty(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 8);
if (VTy->getElementType() == Type::getInt8Ty(VTy->getContext()))
return ScalableVectorType::get(VTy->getElementType(), 16);
llvm_unreachable("Cannot handle input vector type");
}
static Function *getStructuredLoadFunction(Module *M, unsigned Factor,
bool Scalable, Type *LDVTy,
Type *PtrTy) {
assert(Factor >= 2 && Factor <= 4 && "Invalid interleave factor");
static const Intrinsic::ID SVELoads[3] = {Intrinsic::aarch64_sve_ld2_sret,
Intrinsic::aarch64_sve_ld3_sret,
Intrinsic::aarch64_sve_ld4_sret};
static const Intrinsic::ID NEONLoads[3] = {Intrinsic::aarch64_neon_ld2,
Intrinsic::aarch64_neon_ld3,
Intrinsic::aarch64_neon_ld4};
if (Scalable)
return Intrinsic::getOrInsertDeclaration(M, SVELoads[Factor - 2], {LDVTy});
return Intrinsic::getOrInsertDeclaration(M, NEONLoads[Factor - 2],
{LDVTy, PtrTy});
}
static Function *getStructuredStoreFunction(Module *M, unsigned Factor,
bool Scalable, Type *STVTy,
Type *PtrTy) {
assert(Factor >= 2 && Factor <= 4 && "Invalid interleave factor");
static const Intrinsic::ID SVEStores[3] = {Intrinsic::aarch64_sve_st2,
Intrinsic::aarch64_sve_st3,
Intrinsic::aarch64_sve_st4};
static const Intrinsic::ID NEONStores[3] = {Intrinsic::aarch64_neon_st2,
Intrinsic::aarch64_neon_st3,
Intrinsic::aarch64_neon_st4};
if (Scalable)
return Intrinsic::getOrInsertDeclaration(M, SVEStores[Factor - 2], {STVTy});
return Intrinsic::getOrInsertDeclaration(M, NEONStores[Factor - 2],
{STVTy, PtrTy});
}
/// Lower an interleaved load into a ldN intrinsic.
///
/// E.g. Lower an interleaved load (Factor = 2):
/// %wide.vec = load <8 x i32>, <8 x i32>* %ptr
/// %v0 = shuffle %wide.vec, undef, <0, 2, 4, 6> ; Extract even elements
/// %v1 = shuffle %wide.vec, undef, <1, 3, 5, 7> ; Extract odd elements
///
/// Into:
/// %ld2 = { <4 x i32>, <4 x i32> } call llvm.aarch64.neon.ld2(%ptr)
/// %vec0 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 0
/// %vec1 = extractelement { <4 x i32>, <4 x i32> } %ld2, i32 1
bool AArch64TargetLowering::lowerInterleavedLoad(
LoadInst *LI, ArrayRef<ShuffleVectorInst *> Shuffles,
ArrayRef<unsigned> Indices, unsigned Factor) const {
assert(Factor >= 2 && Factor <= getMaxSupportedInterleaveFactor() &&
"Invalid interleave factor");
assert(!Shuffles.empty() && "Empty shufflevector input");
assert(Shuffles.size() == Indices.size() &&
"Unmatched number of shufflevectors and indices");
const DataLayout &DL = LI->getDataLayout();
VectorType *VTy = Shuffles[0]->getType();
// Skip if we do not have NEON and skip illegal vector types. We can
// "legalize" wide vector types into multiple interleaved accesses as long as
// the vector types are divisible by 128.
bool UseScalable;
if (!isLegalInterleavedAccessType(VTy, DL, UseScalable))
return false;
// Check if the interleave is a zext(shuffle), that can be better optimized
// into shift / and masks. For the moment we do this just for uitofp (not
// zext) to avoid issues with widening instructions.
if (Shuffles.size() == 4 && all_of(Shuffles, [](ShuffleVectorInst *SI) {
return SI->hasOneUse() && match(SI->user_back(), m_UIToFP(m_Value())) &&
SI->getType()->getScalarSizeInBits() * 4 ==
SI->user_back()->getType()->getScalarSizeInBits();
}))
return false;
unsigned NumLoads = getNumInterleavedAccesses(VTy, DL, UseScalable);
auto *FVTy = cast<FixedVectorType>(VTy);
// A pointer vector can not be the return type of the ldN intrinsics. Need to
// load integer vectors first and then convert to pointer vectors.
Type *EltTy = FVTy->getElementType();
if (EltTy->isPointerTy())
FVTy =
FixedVectorType::get(DL.getIntPtrType(EltTy), FVTy->getNumElements());
// If we're going to generate more than one load, reset the sub-vector type
// to something legal.
FVTy = FixedVectorType::get(FVTy->getElementType(),
FVTy->getNumElements() / NumLoads);
auto *LDVTy =
UseScalable ? cast<VectorType>(getSVEContainerIRType(FVTy)) : FVTy;
IRBuilder<> Builder(LI);
// The base address of the load.
Value *BaseAddr = LI->getPointerOperand();
Type *PtrTy = LI->getPointerOperandType();
Type *PredTy = VectorType::get(Type::getInt1Ty(LDVTy->getContext()),
LDVTy->getElementCount());
Function *LdNFunc = getStructuredLoadFunction(LI->getModule(), Factor,
UseScalable, LDVTy, PtrTy);
// Holds sub-vectors extracted from the load intrinsic return values. The
// sub-vectors are associated with the shufflevector instructions they will
// replace.
DenseMap<ShuffleVectorInst *, SmallVector<Value *, 4>> SubVecs;
Value *PTrue = nullptr;
if (UseScalable) {
std::optional<unsigned> PgPattern =
getSVEPredPatternFromNumElements(FVTy->getNumElements());
if (Subtarget->getMinSVEVectorSizeInBits() ==
Subtarget->getMaxSVEVectorSizeInBits() &&
Subtarget->getMinSVEVectorSizeInBits() == DL.getTypeSizeInBits(FVTy))
PgPattern = AArch64SVEPredPattern::all;
auto *PTruePat =
ConstantInt::get(Type::getInt32Ty(LDVTy->getContext()), *PgPattern);
PTrue = Builder.CreateIntrinsic(Intrinsic::aarch64_sve_ptrue, {PredTy},
{PTruePat});
}
for (unsigned LoadCount = 0; LoadCount < NumLoads; ++LoadCount) {
// If we're generating more than one load, compute the base address of
// subsequent loads as an offset from the previous.
if (LoadCount > 0)
BaseAddr = Builder.CreateConstGEP1_32(LDVTy->getElementType(), BaseAddr,
FVTy->getNumElements() * Factor);
CallInst *LdN;
if (UseScalable)
LdN = Builder.CreateCall(LdNFunc, {PTrue, BaseAddr}, "ldN");
else
LdN = Builder.CreateCall(LdNFunc, BaseAddr, "ldN");
// Extract and store the sub-vectors returned by the load intrinsic.
for (unsigned i = 0; i < Shuffles.size(); i++) {
ShuffleVectorInst *SVI = Shuffles[i];
unsigned Index = Indices[i];
Value *SubVec = Builder.CreateExtractValue(LdN, Index);
if (UseScalable)
SubVec = Builder.CreateExtractVector(FVTy, SubVec, uint64_t(0));
// Convert the integer vector to pointer vector if the element is pointer.
if (EltTy->isPointerTy())
SubVec = Builder.CreateIntToPtr(
SubVec, FixedVectorType::get(SVI->getType()->getElementType(),
FVTy->getNumElements()));
SubVecs[SVI].push_back(SubVec);
}
}
// Replace uses of the shufflevector instructions with the sub-vectors
// returned by the load intrinsic. If a shufflevector instruction is
// associated with more than one sub-vector, those sub-vectors will be
// concatenated into a single wide vector.
for (ShuffleVectorInst *SVI : Shuffles) {
auto &SubVec = SubVecs[SVI];
auto *WideVec =
SubVec.size() > 1 ? concatenateVectors(Builder, SubVec) : SubVec[0];
SVI->replaceAllUsesWith(WideVec);
}
return true;
}
template <typename Iter>
bool hasNearbyPairedStore(Iter It, Iter End, Value *Ptr, const DataLayout &DL) {
int MaxLookupDist = 20;
unsigned IdxWidth = DL.getIndexSizeInBits(0);
APInt OffsetA(IdxWidth, 0), OffsetB(IdxWidth, 0);
const Value *PtrA1 =
Ptr->stripAndAccumulateInBoundsConstantOffsets(DL, OffsetA);
while (++It != End) {
if (It->isDebugOrPseudoInst())
continue;
if (MaxLookupDist-- == 0)
break;
if (const auto *SI = dyn_cast<StoreInst>(&*It)) {
const Value *PtrB1 =
SI->getPointerOperand()->stripAndAccumulateInBoundsConstantOffsets(
DL, OffsetB);
if (PtrA1 == PtrB1 &&
(OffsetA.sextOrTrunc(IdxWidth) - OffsetB.sextOrTrunc(IdxWidth))
.abs() == 16)
return true;
}
}
return false;
}
/// Lower an interleaved store into a stN intrinsic.
///
/// E.g. Lower an interleaved store (Factor = 3):
/// %i.vec = shuffle <8 x i32> %v0, <8 x i32> %v1,
/// <0, 4, 8, 1, 5, 9, 2, 6, 10, 3, 7, 11>
/// store <12 x i32> %i.vec, <12 x i32>* %ptr
///
/// Into:
/// %sub.v0 = shuffle <8 x i32> %v0, <8 x i32> v1, <0, 1, 2, 3>
/// %sub.v1 = shuffle <8 x i32> %v0, <8 x i32> v1, <4, 5, 6, 7>
/// %sub.v2 = shuffle <8 x i32> %v0, <8 x i32> v1, <8, 9, 10, 11>
/// call void llvm.aarch64.neon.st3(%sub.v0, %sub.v1, %sub.v2, %ptr)
///
/// Note that the new shufflevectors will be removed and we'll only generate one
/// st3 instruction in CodeGen.
///
/// Example for a more general valid mask (Factor 3). Lower:
/// %i.vec = shuffle <32 x i32> %v0, <32 x i32> %v1,
/// <4, 32, 16, 5, 33, 17, 6, 34, 18, 7, 35, 19>
/// store <12 x i32> %i.vec, <12 x i32>* %ptr
///
/// Into:
/// %sub.v0 = shuffle <32 x i32> %v0, <32 x i32> v1, <4, 5, 6, 7>
/// %sub.v1 = shuffle <32 x i32> %v0, <32 x i32> v1, <32, 33, 34, 35>
/// %sub.v2 = shuffle <32 x i32> %v0, <32 x i32> v1, <16, 17, 18, 19>
/// call void llvm.aarch64.neon.st3(%sub.v0, %sub.v1, %sub.v2, %ptr)
bool AArch64TargetLowering::lowerInterleavedStore(StoreInst *SI,
ShuffleVectorInst *SVI,
unsigned Factor) const {
assert(Factor >= 2 && Factor <= getMaxSupportedInterleaveFactor() &&
"Invalid interleave factor");
auto *VecTy = cast<FixedVectorType>(SVI->getType());
assert(VecTy->getNumElements() % Factor == 0 && "Invalid interleaved store");
unsigned LaneLen = VecTy->getNumElements() / Factor;
Type *EltTy = VecTy->getElementType();
auto *SubVecTy = FixedVectorType::get(EltTy, LaneLen);
const DataLayout &DL = SI->getDataLayout();
bool UseScalable;
// Skip if we do not have NEON and skip illegal vector types. We can
// "legalize" wide vector types into multiple interleaved accesses as long as
// the vector types are divisible by 128.
if (!isLegalInterleavedAccessType(SubVecTy, DL, UseScalable))
return false;
unsigned NumStores = getNumInterleavedAccesses(SubVecTy, DL, UseScalable);
Value *Op0 = SVI->getOperand(0);
Value *Op1 = SVI->getOperand(1);
IRBuilder<> Builder(SI);
// StN intrinsics don't support pointer vectors as arguments. Convert pointer
// vectors to integer vectors.
if (EltTy->isPointerTy()) {
Type *IntTy = DL.getIntPtrType(EltTy);
unsigned NumOpElts =
cast<FixedVectorType>(Op0->getType())->getNumElements();
// Convert to the corresponding integer vector.
auto *IntVecTy = FixedVectorType::get(IntTy, NumOpElts);
Op0 = Builder.CreatePtrToInt(Op0, IntVecTy);
Op1 = Builder.CreatePtrToInt(Op1, IntVecTy);
SubVecTy = FixedVectorType::get(IntTy, LaneLen);
}
// If we're going to generate more than one store, reset the lane length
// and sub-vector type to something legal.
LaneLen /= NumStores;
SubVecTy = FixedVectorType::get(SubVecTy->getElementType(), LaneLen);
auto *STVTy = UseScalable ? cast<VectorType>(getSVEContainerIRType(SubVecTy))
: SubVecTy;
// The base address of the store.
Value *BaseAddr = SI->getPointerOperand();
auto Mask = SVI->getShuffleMask();
// Sanity check if all the indices are NOT in range.
// If mask is `poison`, `Mask` may be a vector of -1s.
// If all of them are `poison`, OOB read will happen later.
if (llvm::all_of(Mask, [](int Idx) { return Idx == PoisonMaskElem; })) {
return false;
}
// A 64bit st2 which does not start at element 0 will involved adding extra
// ext elements making the st2 unprofitable, and if there is a nearby store
// that points to BaseAddr+16 or BaseAddr-16 then it can be better left as a
// zip;ldp pair which has higher throughput.
if (Factor == 2 && SubVecTy->getPrimitiveSizeInBits() == 64 &&
(Mask[0] != 0 ||
hasNearbyPairedStore(SI->getIterator(), SI->getParent()->end(), BaseAddr,
DL) ||
hasNearbyPairedStore(SI->getReverseIterator(), SI->getParent()->rend(),
BaseAddr, DL)))
return false;
Type *PtrTy = SI->getPointerOperandType();
Type *PredTy = VectorType::get(Type::getInt1Ty(STVTy->getContext()),
STVTy->getElementCount());
Function *StNFunc = getStructuredStoreFunction(SI->getModule(), Factor,
UseScalable, STVTy, PtrTy);
Value *PTrue = nullptr;
if (UseScalable) {
std::optional<unsigned> PgPattern =
getSVEPredPatternFromNumElements(SubVecTy->getNumElements());
if (Subtarget->getMinSVEVectorSizeInBits() ==
Subtarget->getMaxSVEVectorSizeInBits() &&
Subtarget->getMinSVEVectorSizeInBits() ==
DL.getTypeSizeInBits(SubVecTy))
PgPattern = AArch64SVEPredPattern::all;
auto *PTruePat =
ConstantInt::get(Type::getInt32Ty(STVTy->getContext()), *PgPattern);
PTrue = Builder.CreateIntrinsic(Intrinsic::aarch64_sve_ptrue, {PredTy},
{PTruePat});
}
for (unsigned StoreCount = 0; StoreCount < NumStores; ++StoreCount) {
SmallVector<Value *, 5> Ops;
// Split the shufflevector operands into sub vectors for the new stN call.
for (unsigned i = 0; i < Factor; i++) {
Value *Shuffle;
unsigned IdxI = StoreCount * LaneLen * Factor + i;
if (Mask[IdxI] >= 0) {
Shuffle = Builder.CreateShuffleVector(
Op0, Op1, createSequentialMask(Mask[IdxI], LaneLen, 0));
} else {
unsigned StartMask = 0;
for (unsigned j = 1; j < LaneLen; j++) {
unsigned IdxJ = StoreCount * LaneLen * Factor + j * Factor + i;
if (Mask[IdxJ] >= 0) {
StartMask = Mask[IdxJ] - j;
break;
}
}
// Note: Filling undef gaps with random elements is ok, since
// those elements were being written anyway (with undefs).
// In the case of all undefs we're defaulting to using elems from 0
// Note: StartMask cannot be negative, it's checked in
// isReInterleaveMask
Shuffle = Builder.CreateShuffleVector(
Op0, Op1, createSequentialMask(StartMask, LaneLen, 0));
}
if (UseScalable)
Shuffle = Builder.CreateInsertVector(STVTy, PoisonValue::get(STVTy),
Shuffle, uint64_t(0));
Ops.push_back(Shuffle);
}
if (UseScalable)
Ops.push_back(PTrue);
// If we generating more than one store, we compute the base address of
// subsequent stores as an offset from the previous.
if (StoreCount > 0)
BaseAddr = Builder.CreateConstGEP1_32(SubVecTy->getElementType(),
BaseAddr, LaneLen * Factor);
Ops.push_back(BaseAddr);
Builder.CreateCall(StNFunc, Ops);
}
return true;
}
bool AArch64TargetLowering::lowerDeinterleaveIntrinsicToLoad(
LoadInst *LI, ArrayRef<Value *> DeinterleavedValues) const {
unsigned Factor = DeinterleavedValues.size();
if (Factor != 2 && Factor != 4) {
LLVM_DEBUG(dbgs() << "Matching ld2 and ld4 patterns failed\n");
return false;
}
VectorType *VTy = cast<VectorType>(DeinterleavedValues[0]->getType());
const DataLayout &DL = LI->getModule()->getDataLayout();
bool UseScalable;
if (!isLegalInterleavedAccessType(VTy, DL, UseScalable))
return false;
// TODO: Add support for using SVE instructions with fixed types later, using
// the code from lowerInterleavedLoad to obtain the correct container type.
if (UseScalable && !VTy->isScalableTy())
return false;
unsigned NumLoads = getNumInterleavedAccesses(VTy, DL, UseScalable);
VectorType *LdTy =
VectorType::get(VTy->getElementType(),
VTy->getElementCount().divideCoefficientBy(NumLoads));
Type *PtrTy = LI->getPointerOperandType();
Function *LdNFunc = getStructuredLoadFunction(LI->getModule(), Factor,
UseScalable, LdTy, PtrTy);
IRBuilder<> Builder(LI);
Value *Pred = nullptr;
if (UseScalable)
Pred =
Builder.CreateVectorSplat(LdTy->getElementCount(), Builder.getTrue());
Value *BaseAddr = LI->getPointerOperand();
if (NumLoads > 1) {
// Create multiple legal small ldN.
SmallVector<Value *, 4> ExtractedLdValues(Factor, PoisonValue::get(VTy));
for (unsigned I = 0; I < NumLoads; ++I) {
Value *Offset = Builder.getInt64(I * Factor);
Value *Address = Builder.CreateGEP(LdTy, BaseAddr, {Offset});
Value *LdN = nullptr;
if (UseScalable)
LdN = Builder.CreateCall(LdNFunc, {Pred, Address}, "ldN");
else
LdN = Builder.CreateCall(LdNFunc, Address, "ldN");
Value *Idx =
Builder.getInt64(I * LdTy->getElementCount().getKnownMinValue());
for (unsigned J = 0; J < Factor; ++J) {
ExtractedLdValues[J] = Builder.CreateInsertVector(
VTy, ExtractedLdValues[J], Builder.CreateExtractValue(LdN, J), Idx);
}
LLVM_DEBUG(dbgs() << "LdN4 res: "; LdN->dump());
}
// Replace output of deinterleave2 intrinsic by output of ldN2/ldN4
for (unsigned J = 0; J < Factor; ++J)
DeinterleavedValues[J]->replaceAllUsesWith(ExtractedLdValues[J]);
} else {
Value *Result;
if (UseScalable)
Result = Builder.CreateCall(LdNFunc, {Pred, BaseAddr}, "ldN");
else
Result = Builder.CreateCall(LdNFunc, BaseAddr, "ldN");
// Replace output of deinterleave2 intrinsic by output of ldN2/ldN4
for (unsigned I = 0; I < Factor; I++) {
Value *NewExtract = Builder.CreateExtractValue(Result, I);
DeinterleavedValues[I]->replaceAllUsesWith(NewExtract);
}
}
return true;
}
bool AArch64TargetLowering::lowerInterleaveIntrinsicToStore(
StoreInst *SI, ArrayRef<Value *> InterleavedValues) const {
unsigned Factor = InterleavedValues.size();
if (Factor != 2 && Factor != 4) {
LLVM_DEBUG(dbgs() << "Matching st2 and st4 patterns failed\n");
return false;
}
VectorType *VTy = cast<VectorType>(InterleavedValues[0]->getType());
const DataLayout &DL = SI->getModule()->getDataLayout();
bool UseScalable;
if (!isLegalInterleavedAccessType(VTy, DL, UseScalable))
return false;
// TODO: Add support for using SVE instructions with fixed types later, using
// the code from lowerInterleavedStore to obtain the correct container type.
if (UseScalable && !VTy->isScalableTy())
return false;
unsigned NumStores = getNumInterleavedAccesses(VTy, DL, UseScalable);
VectorType *StTy =
VectorType::get(VTy->getElementType(),
VTy->getElementCount().divideCoefficientBy(NumStores));
Type *PtrTy = SI->getPointerOperandType();
Function *StNFunc = getStructuredStoreFunction(SI->getModule(), Factor,
UseScalable, StTy, PtrTy);
IRBuilder<> Builder(SI);
Value *BaseAddr = SI->getPointerOperand();
Value *Pred = nullptr;
if (UseScalable)
Pred =
Builder.CreateVectorSplat(StTy->getElementCount(), Builder.getTrue());
auto ExtractedValues = InterleavedValues;
SmallVector<Value *, 4> StoreOperands(InterleavedValues);
if (UseScalable)
StoreOperands.push_back(Pred);
StoreOperands.push_back(BaseAddr);
for (unsigned I = 0; I < NumStores; ++I) {
Value *Address = BaseAddr;
if (NumStores > 1) {
Value *Offset = Builder.getInt64(I * Factor);
Address = Builder.CreateGEP(StTy, BaseAddr, {Offset});
Value *Idx =
Builder.getInt64(I * StTy->getElementCount().getKnownMinValue());
for (unsigned J = 0; J < Factor; J++) {
StoreOperands[J] =
Builder.CreateExtractVector(StTy, ExtractedValues[J], Idx);
}
// update the address
StoreOperands[StoreOperands.size() - 1] = Address;
}
Builder.CreateCall(StNFunc, StoreOperands);
}
return true;
}
EVT AArch64TargetLowering::getOptimalMemOpType(
const MemOp &Op, const AttributeList &FuncAttributes) const {
bool CanImplicitFloat = !FuncAttributes.hasFnAttr(Attribute::NoImplicitFloat);
bool CanUseNEON = Subtarget->hasNEON() && CanImplicitFloat;
bool CanUseFP = Subtarget->hasFPARMv8() && CanImplicitFloat;
// Only use AdvSIMD to implement memset of 32-byte and above. It would have
// taken one instruction to materialize the v2i64 zero and one store (with
// restrictive addressing mode). Just do i64 stores.
bool IsSmallMemset = Op.isMemset() && Op.size() < 32;
auto AlignmentIsAcceptable = [&](EVT VT, Align AlignCheck) {
if (Op.isAligned(AlignCheck))
return true;
unsigned Fast;
return allowsMisalignedMemoryAccesses(VT, 0, Align(1),
MachineMemOperand::MONone, &Fast) &&
Fast;
};
if (CanUseNEON && Op.isMemset() && !IsSmallMemset &&
AlignmentIsAcceptable(MVT::v16i8, Align(16)))
return MVT::v16i8;
if (CanUseFP && !IsSmallMemset && AlignmentIsAcceptable(MVT::f128, Align(16)))
return MVT::f128;
if (Op.size() >= 8 && AlignmentIsAcceptable(MVT::i64, Align(8)))
return MVT::i64;
if (Op.size() >= 4 && AlignmentIsAcceptable(MVT::i32, Align(4)))
return MVT::i32;
return MVT::Other;
}
LLT AArch64TargetLowering::getOptimalMemOpLLT(
const MemOp &Op, const AttributeList &FuncAttributes) const {
bool CanImplicitFloat = !FuncAttributes.hasFnAttr(Attribute::NoImplicitFloat);
bool CanUseNEON = Subtarget->hasNEON() && CanImplicitFloat;
bool CanUseFP = Subtarget->hasFPARMv8() && CanImplicitFloat;
// Only use AdvSIMD to implement memset of 32-byte and above. It would have
// taken one instruction to materialize the v2i64 zero and one store (with
// restrictive addressing mode). Just do i64 stores.
bool IsSmallMemset = Op.isMemset() && Op.size() < 32;
auto AlignmentIsAcceptable = [&](EVT VT, Align AlignCheck) {
if (Op.isAligned(AlignCheck))
return true;
unsigned Fast;
return allowsMisalignedMemoryAccesses(VT, 0, Align(1),
MachineMemOperand::MONone, &Fast) &&
Fast;
};
if (CanUseNEON && Op.isMemset() && !IsSmallMemset &&
AlignmentIsAcceptable(MVT::v2i64, Align(16)))
return LLT::fixed_vector(2, 64);
if (CanUseFP && !IsSmallMemset && AlignmentIsAcceptable(MVT::f128, Align(16)))
return LLT::scalar(128);
if (Op.size() >= 8 && AlignmentIsAcceptable(MVT::i64, Align(8)))
return LLT::scalar(64);
if (Op.size() >= 4 && AlignmentIsAcceptable(MVT::i32, Align(4)))
return LLT::scalar(32);
return LLT();
}
// 12-bit optionally shifted immediates are legal for adds.
bool AArch64TargetLowering::isLegalAddImmediate(int64_t Immed) const {
if (Immed == std::numeric_limits<int64_t>::min()) {
return false;
}
// Same encoding for add/sub, just flip the sign.
return isLegalArithImmed((uint64_t)std::abs(Immed));
}
bool AArch64TargetLowering::isLegalAddScalableImmediate(int64_t Imm) const {
// We will only emit addvl/inc* instructions for SVE2
if (!Subtarget->hasSVE2())
return false;
// addvl's immediates are in terms of the number of bytes in a register.
// Since there are 16 in the base supported size (128bits), we need to
// divide the immediate by that much to give us a useful immediate to
// multiply by vscale. We can't have a remainder as a result of this.
if (Imm % 16 == 0)
return isInt<6>(Imm / 16);
// Inc[b|h|w|d] instructions take a pattern and a positive immediate
// multiplier. For now, assume a pattern of 'all'. Incb would be a subset
// of addvl as a result, so only take h|w|d into account.
// Dec[h|w|d] will cover subtractions.
// Immediates are in the range [1,16], so we can't do a 2's complement check.
// FIXME: Can we make use of other patterns to cover other immediates?
// inch|dech
if (Imm % 8 == 0)
return std::abs(Imm / 8) <= 16;
// incw|decw
if (Imm % 4 == 0)
return std::abs(Imm / 4) <= 16;
// incd|decd
if (Imm % 2 == 0)
return std::abs(Imm / 2) <= 16;
return false;
}
// Return false to prevent folding
// (mul (add x, c1), c2) -> (add (mul x, c2), c2*c1) in DAGCombine,
// if the folding leads to worse code.
bool AArch64TargetLowering::isMulAddWithConstProfitable(
SDValue AddNode, SDValue ConstNode) const {
// Let the DAGCombiner decide for vector types and large types.
const EVT VT = AddNode.getValueType();
if (VT.isVector() || VT.getScalarSizeInBits() > 64)
return true;
// It is worse if c1 is legal add immediate, while c1*c2 is not
// and has to be composed by at least two instructions.
const ConstantSDNode *C1Node = cast<ConstantSDNode>(AddNode.getOperand(1));
const ConstantSDNode *C2Node = cast<ConstantSDNode>(ConstNode);
const int64_t C1 = C1Node->getSExtValue();
const APInt C1C2 = C1Node->getAPIntValue() * C2Node->getAPIntValue();
if (!isLegalAddImmediate(C1) || isLegalAddImmediate(C1C2.getSExtValue()))
return true;
SmallVector<AArch64_IMM::ImmInsnModel, 4> Insn;
// Adapt to the width of a register.
unsigned BitSize = VT.getSizeInBits() <= 32 ? 32 : 64;
AArch64_IMM::expandMOVImm(C1C2.getZExtValue(), BitSize, Insn);
if (Insn.size() > 1)
return false;
// Default to true and let the DAGCombiner decide.
return true;
}
// Integer comparisons are implemented with ADDS/SUBS, so the range of valid
// immediates is the same as for an add or a sub.
bool AArch64TargetLowering::isLegalICmpImmediate(int64_t Immed) const {
return isLegalAddImmediate(Immed);
}
/// isLegalAddressingMode - Return true if the addressing mode represented
/// by AM is legal for this target, for a load/store of the specified type.
bool AArch64TargetLowering::isLegalAddressingMode(const DataLayout &DL,
const AddrMode &AMode, Type *Ty,
unsigned AS, Instruction *I) const {
// AArch64 has five basic addressing modes:
// reg
// reg + 9-bit signed offset
// reg + SIZE_IN_BYTES * 12-bit unsigned offset
// reg1 + reg2
// reg + SIZE_IN_BYTES * reg
// No global is ever allowed as a base.
if (AMode.BaseGV)
return false;
// No reg+reg+imm addressing.
if (AMode.HasBaseReg && AMode.BaseOffs && AMode.Scale)
return false;
// Canonicalise `1*ScaledReg + imm` into `BaseReg + imm` and
// `2*ScaledReg` into `BaseReg + ScaledReg`
AddrMode AM = AMode;
if (AM.Scale && !AM.HasBaseReg) {
if (AM.Scale == 1) {
AM.HasBaseReg = true;
AM.Scale = 0;
} else if (AM.Scale == 2) {
AM.HasBaseReg = true;
AM.Scale = 1;
} else {
return false;
}
}
// A base register is required in all addressing modes.
if (!AM.HasBaseReg)
return false;
if (Ty->isScalableTy()) {
if (isa<ScalableVectorType>(Ty)) {
// See if we have a foldable vscale-based offset, for vector types which
// are either legal or smaller than the minimum; more work will be
// required if we need to consider addressing for types which need
// legalization by splitting.
uint64_t VecNumBytes = DL.getTypeSizeInBits(Ty).getKnownMinValue() / 8;
if (AM.HasBaseReg && !AM.BaseOffs && AM.ScalableOffset && !AM.Scale &&
(AM.ScalableOffset % VecNumBytes == 0) && VecNumBytes <= 16 &&
isPowerOf2_64(VecNumBytes))
return isInt<4>(AM.ScalableOffset / (int64_t)VecNumBytes);
uint64_t VecElemNumBytes =
DL.getTypeSizeInBits(cast<VectorType>(Ty)->getElementType()) / 8;
return AM.HasBaseReg && !AM.BaseOffs && !AM.ScalableOffset &&
(AM.Scale == 0 || (uint64_t)AM.Scale == VecElemNumBytes);
}
return AM.HasBaseReg && !AM.BaseOffs && !AM.ScalableOffset && !AM.Scale;
}
// No scalable offsets allowed for non-scalable types.
if (AM.ScalableOffset)
return false;
// check reg + imm case:
// i.e., reg + 0, reg + imm9, reg + SIZE_IN_BYTES * uimm12
uint64_t NumBytes = 0;
if (Ty->isSized()) {
uint64_t NumBits = DL.getTypeSizeInBits(Ty);
NumBytes = NumBits / 8;
if (!isPowerOf2_64(NumBits))
NumBytes = 0;
}
return Subtarget->getInstrInfo()->isLegalAddressingMode(NumBytes, AM.BaseOffs,
AM.Scale);
}
// Check whether the 2 offsets belong to the same imm24 range, and their high
// 12bits are same, then their high part can be decoded with the offset of add.
int64_t
AArch64TargetLowering::getPreferredLargeGEPBaseOffset(int64_t MinOffset,
int64_t MaxOffset) const {
int64_t HighPart = MinOffset & ~0xfffULL;
if (MinOffset >> 12 == MaxOffset >> 12 && isLegalAddImmediate(HighPart)) {
// Rebase the value to an integer multiple of imm12.
return HighPart;
}
return 0;
}
bool AArch64TargetLowering::shouldConsiderGEPOffsetSplit() const {
// Consider splitting large offset of struct or array.
return true;
}
bool AArch64TargetLowering::isFMAFasterThanFMulAndFAdd(
const MachineFunction &MF, EVT VT) const {
VT = VT.getScalarType();
if (!VT.isSimple())
return false;
switch (VT.getSimpleVT().SimpleTy) {
case MVT::f16:
return Subtarget->hasFullFP16();
case MVT::f32:
case MVT::f64:
return true;
default:
break;
}
return false;
}
bool AArch64TargetLowering::isFMAFasterThanFMulAndFAdd(const Function &F,
Type *Ty) const {
switch (Ty->getScalarType()->getTypeID()) {
case Type::FloatTyID:
case Type::DoubleTyID:
return true;
default:
return false;
}
}
bool AArch64TargetLowering::generateFMAsInMachineCombiner(
EVT VT, CodeGenOptLevel OptLevel) const {
return (OptLevel >= CodeGenOptLevel::Aggressive) && !VT.isScalableVector() &&
!useSVEForFixedLengthVectorVT(VT);
}
const MCPhysReg *
AArch64TargetLowering::getScratchRegisters(CallingConv::ID) const {
// LR is a callee-save register, but we must treat it as clobbered by any call
// site. Hence we include LR in the scratch registers, which are in turn added
// as implicit-defs for stackmaps and patchpoints.
static const MCPhysReg ScratchRegs[] = {
AArch64::X16, AArch64::X17, AArch64::LR, 0
};
return ScratchRegs;
}
ArrayRef<MCPhysReg> AArch64TargetLowering::getRoundingControlRegisters() const {
static const MCPhysReg RCRegs[] = {AArch64::FPCR};
return RCRegs;
}
bool
AArch64TargetLowering::isDesirableToCommuteWithShift(const SDNode *N,
CombineLevel Level) const {
assert((N->getOpcode() == ISD::SHL || N->getOpcode() == ISD::SRA ||
N->getOpcode() == ISD::SRL) &&
"Expected shift op");
SDValue ShiftLHS = N->getOperand(0);
EVT VT = N->getValueType(0);
if (!ShiftLHS->hasOneUse())
return false;
if (ShiftLHS.getOpcode() == ISD::SIGN_EXTEND &&
!ShiftLHS.getOperand(0)->hasOneUse())
return false;
// If ShiftLHS is unsigned bit extraction: ((x >> C) & mask), then do not
// combine it with shift 'N' to let it be lowered to UBFX except:
// ((x >> C) & mask) << C.
if (ShiftLHS.getOpcode() == ISD::AND && (VT == MVT::i32 || VT == MVT::i64) &&
isa<ConstantSDNode>(ShiftLHS.getOperand(1))) {
uint64_t TruncMask = ShiftLHS.getConstantOperandVal(1);
if (isMask_64(TruncMask)) {
SDValue AndLHS = ShiftLHS.getOperand(0);
if (AndLHS.getOpcode() == ISD::SRL) {
if (auto *SRLC = dyn_cast<ConstantSDNode>(AndLHS.getOperand(1))) {
if (N->getOpcode() == ISD::SHL)
if (auto *SHLC = dyn_cast<ConstantSDNode>(N->getOperand(1)))
return SRLC->getZExtValue() == SHLC->getZExtValue();
return false;
}
}
}
}
return true;
}
bool AArch64TargetLowering::isDesirableToCommuteXorWithShift(
const SDNode *N) const {
assert(N->getOpcode() == ISD::XOR &&
(N->getOperand(0).getOpcode() == ISD::SHL ||
N->getOperand(0).getOpcode() == ISD::SRL) &&
"Expected XOR(SHIFT) pattern");
// Only commute if the entire NOT mask is a hidden shifted mask.
auto *XorC = dyn_cast<ConstantSDNode>(N->getOperand(1));
auto *ShiftC = dyn_cast<ConstantSDNode>(N->getOperand(0).getOperand(1));
if (XorC && ShiftC) {
unsigned MaskIdx, MaskLen;
if (XorC->getAPIntValue().isShiftedMask(MaskIdx, MaskLen)) {
unsigned ShiftAmt = ShiftC->getZExtValue();
unsigned BitWidth = N->getValueType(0).getScalarSizeInBits();
if (N->getOperand(0).getOpcode() == ISD::SHL)
return MaskIdx == ShiftAmt && MaskLen == (BitWidth - ShiftAmt);
return MaskIdx == 0 && MaskLen == (BitWidth - ShiftAmt);
}
}
return false;
}
bool AArch64TargetLowering::shouldFoldConstantShiftPairToMask(
const SDNode *N, CombineLevel Level) const {
assert(((N->getOpcode() == ISD::SHL &&
N->getOperand(0).getOpcode() == ISD::SRL) ||
(N->getOpcode() == ISD::SRL &&
N->getOperand(0).getOpcode() == ISD::SHL)) &&
"Expected shift-shift mask");
// Don't allow multiuse shift folding with the same shift amount.
if (!N->getOperand(0)->hasOneUse())
return false;
// Only fold srl(shl(x,c1),c2) iff C1 >= C2 to prevent loss of UBFX patterns.
EVT VT = N->getValueType(0);
if (N->getOpcode() == ISD::SRL && (VT == MVT::i32 || VT == MVT::i64)) {
auto *C1 = dyn_cast<ConstantSDNode>(N->getOperand(0).getOperand(1));
auto *C2 = dyn_cast<ConstantSDNode>(N->getOperand(1));
return (!C1 || !C2 || C1->getZExtValue() >= C2->getZExtValue());
}
// We do not need to fold when this shifting used in specific load case:
// (ldr x, (add x, (shl (srl x, c1) 2)))
if (N->getOpcode() == ISD::SHL && N->hasOneUse()) {
if (auto C2 = dyn_cast<ConstantSDNode>(N->getOperand(1))) {
unsigned ShlAmt = C2->getZExtValue();
if (auto ShouldADD = *N->user_begin();
ShouldADD->getOpcode() == ISD::ADD && ShouldADD->hasOneUse()) {
if (auto ShouldLOAD = dyn_cast<LoadSDNode>(*ShouldADD->user_begin())) {
unsigned ByteVT = ShouldLOAD->getMemoryVT().getSizeInBits() / 8;
if ((1ULL << ShlAmt) == ByteVT &&
isIndexedLoadLegal(ISD::PRE_INC, ShouldLOAD->getMemoryVT()))
return false;
}
}
}
}
return true;
}
bool AArch64TargetLowering::shouldFoldSelectWithIdentityConstant(
unsigned BinOpcode, EVT VT, unsigned SelectOpcode, SDValue X,
SDValue Y) const {
return VT.isScalableVector() && isTypeLegal(VT) &&
SelectOpcode == ISD::VSELECT;
}
bool AArch64TargetLowering::shouldConvertConstantLoadToIntImm(const APInt &Imm,
Type *Ty) const {
assert(Ty->isIntegerTy());
unsigned BitSize = Ty->getPrimitiveSizeInBits();
if (BitSize == 0)
return false;
int64_t Val = Imm.getSExtValue();
if (Val == 0 || AArch64_AM::isLogicalImmediate(Val, BitSize))
return true;
if ((int64_t)Val < 0)
Val = ~Val;
if (BitSize == 32)
Val &= (1LL << 32) - 1;
unsigned Shift = llvm::Log2_64((uint64_t)Val) / 16;
// MOVZ is free so return true for one or fewer MOVK.
return Shift < 3;
}
bool AArch64TargetLowering::isExtractSubvectorCheap(EVT ResVT, EVT SrcVT,
unsigned Index) const {
if (!isOperationLegalOrCustom(ISD::EXTRACT_SUBVECTOR, ResVT))
return false;
return (Index == 0 || Index == ResVT.getVectorMinNumElements());
}
/// Turn vector tests of the signbit in the form of:
/// xor (sra X, elt_size(X)-1), -1
/// into:
/// cmge X, X, #0
static SDValue foldVectorXorShiftIntoCmp(SDNode *N, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
EVT VT = N->getValueType(0);
if (!Subtarget->hasNEON() || !VT.isVector())
return SDValue();
// There must be a shift right algebraic before the xor, and the xor must be a
// 'not' operation.
SDValue Shift = N->getOperand(0);
SDValue Ones = N->getOperand(1);
if (Shift.getOpcode() != AArch64ISD::VASHR || !Shift.hasOneUse() ||
!ISD::isBuildVectorAllOnes(Ones.getNode()))
return SDValue();
// The shift should be smearing the sign bit across each vector element.
auto *ShiftAmt = dyn_cast<ConstantSDNode>(Shift.getOperand(1));
EVT ShiftEltTy = Shift.getValueType().getVectorElementType();
if (!ShiftAmt || ShiftAmt->getZExtValue() != ShiftEltTy.getSizeInBits() - 1)
return SDValue();
SDLoc DL(N);
SDValue Zero = DAG.getConstant(0, DL, Shift.getValueType());
return DAG.getSetCC(DL, VT, Shift.getOperand(0), Zero, ISD::SETGE);
}
// Given a vecreduce_add node, detect the below pattern and convert it to the
// node sequence with UABDL, [S|U]ADB and UADDLP.
//
// i32 vecreduce_add(
// v16i32 abs(
// v16i32 sub(
// v16i32 [sign|zero]_extend(v16i8 a), v16i32 [sign|zero]_extend(v16i8 b))))
// =================>
// i32 vecreduce_add(
// v4i32 UADDLP(
// v8i16 add(
// v8i16 zext(
// v8i8 [S|U]ABD low8:v16i8 a, low8:v16i8 b
// v8i16 zext(
// v8i8 [S|U]ABD high8:v16i8 a, high8:v16i8 b
static SDValue performVecReduceAddCombineWithUADDLP(SDNode *N,
SelectionDAG &DAG) {
// Assumed i32 vecreduce_add
if (N->getValueType(0) != MVT::i32)
return SDValue();
SDValue VecReduceOp0 = N->getOperand(0);
unsigned Opcode = VecReduceOp0.getOpcode();
// Assumed v16i32 abs
if (Opcode != ISD::ABS || VecReduceOp0->getValueType(0) != MVT::v16i32)
return SDValue();
SDValue ABS = VecReduceOp0;
// Assumed v16i32 sub
if (ABS->getOperand(0)->getOpcode() != ISD::SUB ||
ABS->getOperand(0)->getValueType(0) != MVT::v16i32)
return SDValue();
SDValue SUB = ABS->getOperand(0);
unsigned Opcode0 = SUB->getOperand(0).getOpcode();
unsigned Opcode1 = SUB->getOperand(1).getOpcode();
// Assumed v16i32 type
if (SUB->getOperand(0)->getValueType(0) != MVT::v16i32 ||
SUB->getOperand(1)->getValueType(0) != MVT::v16i32)
return SDValue();
// Assumed zext or sext
bool IsZExt = false;
if (Opcode0 == ISD::ZERO_EXTEND && Opcode1 == ISD::ZERO_EXTEND) {
IsZExt = true;
} else if (Opcode0 == ISD::SIGN_EXTEND && Opcode1 == ISD::SIGN_EXTEND) {
IsZExt = false;
} else
return SDValue();
SDValue EXT0 = SUB->getOperand(0);
SDValue EXT1 = SUB->getOperand(1);
// Assumed zext's operand has v16i8 type
if (EXT0->getOperand(0)->getValueType(0) != MVT::v16i8 ||
EXT1->getOperand(0)->getValueType(0) != MVT::v16i8)
return SDValue();
// Pattern is detected. Let's convert it to sequence of nodes.
SDLoc DL(N);
// First, create the node pattern of UABD/SABD.
SDValue UABDHigh8Op0 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v8i8, EXT0->getOperand(0),
DAG.getConstant(8, DL, MVT::i64));
SDValue UABDHigh8Op1 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v8i8, EXT1->getOperand(0),
DAG.getConstant(8, DL, MVT::i64));
SDValue UABDHigh8 = DAG.getNode(IsZExt ? ISD::ABDU : ISD::ABDS, DL, MVT::v8i8,
UABDHigh8Op0, UABDHigh8Op1);
SDValue UABDL = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::v8i16, UABDHigh8);
// Second, create the node pattern of UABAL.
SDValue UABDLo8Op0 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v8i8, EXT0->getOperand(0),
DAG.getConstant(0, DL, MVT::i64));
SDValue UABDLo8Op1 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v8i8, EXT1->getOperand(0),
DAG.getConstant(0, DL, MVT::i64));
SDValue UABDLo8 = DAG.getNode(IsZExt ? ISD::ABDU : ISD::ABDS, DL, MVT::v8i8,
UABDLo8Op0, UABDLo8Op1);
SDValue ZExtUABD = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::v8i16, UABDLo8);
SDValue UABAL = DAG.getNode(ISD::ADD, DL, MVT::v8i16, UABDL, ZExtUABD);
// Third, create the node of UADDLP.
SDValue UADDLP = DAG.getNode(AArch64ISD::UADDLP, DL, MVT::v4i32, UABAL);
// Fourth, create the node of VECREDUCE_ADD.
return DAG.getNode(ISD::VECREDUCE_ADD, DL, MVT::i32, UADDLP);
}
static SDValue
performActiveLaneMaskCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *ST) {
if (DCI.isBeforeLegalize())
return SDValue();
if (SDValue While = optimizeIncrementingWhile(N, DCI.DAG, /*IsSigned=*/false,
/*IsEqual=*/false))
return While;
if (!N->getValueType(0).isScalableVector() ||
(!ST->hasSVE2p1() && !(ST->hasSME2() && ST->isStreaming())))
return SDValue();
if (!N->hasNUsesOfValue(2, 0))
return SDValue();
const uint64_t HalfSize = N->getValueType(0).getVectorMinNumElements() / 2;
if (HalfSize < 2)
return SDValue();
auto It = N->user_begin();
SDNode *Lo = *It++;
SDNode *Hi = *It;
if (Lo->getOpcode() != ISD::EXTRACT_SUBVECTOR ||
Hi->getOpcode() != ISD::EXTRACT_SUBVECTOR)
return SDValue();
uint64_t OffLo = Lo->getConstantOperandVal(1);
uint64_t OffHi = Hi->getConstantOperandVal(1);
if (OffLo > OffHi) {
std::swap(Lo, Hi);
std::swap(OffLo, OffHi);
}
if (OffLo != 0 || OffHi != HalfSize)
return SDValue();
EVT HalfVec = Lo->getValueType(0);
if (HalfVec != Hi->getValueType(0) ||
HalfVec.getVectorElementCount() != ElementCount::getScalable(HalfSize))
return SDValue();
SelectionDAG &DAG = DCI.DAG;
SDLoc DL(N);
SDValue ID =
DAG.getTargetConstant(Intrinsic::aarch64_sve_whilelo_x2, DL, MVT::i64);
SDValue Idx = N->getOperand(0);
SDValue TC = N->getOperand(1);
if (Idx.getValueType() != MVT::i64) {
Idx = DAG.getZExtOrTrunc(Idx, DL, MVT::i64);
TC = DAG.getZExtOrTrunc(TC, DL, MVT::i64);
}
auto R =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL,
{Lo->getValueType(0), Hi->getValueType(0)}, {ID, Idx, TC});
DCI.CombineTo(Lo, R.getValue(0));
DCI.CombineTo(Hi, R.getValue(1));
return SDValue(N, 0);
}
// Turn a v8i8/v16i8 extended vecreduce into a udot/sdot and vecreduce
// vecreduce.add(ext(A)) to vecreduce.add(DOT(zero, A, one))
// vecreduce.add(mul(ext(A), ext(B))) to vecreduce.add(DOT(zero, A, B))
// If we have vectors larger than v16i8 we extract v16i8 vectors,
// Follow the same steps above to get DOT instructions concatenate them
// and generate vecreduce.add(concat_vector(DOT, DOT2, ..)).
static SDValue performVecReduceAddCombine(SDNode *N, SelectionDAG &DAG,
const AArch64Subtarget *ST) {
if (!ST->isNeonAvailable())
return SDValue();
if (!ST->hasDotProd())
return performVecReduceAddCombineWithUADDLP(N, DAG);
SDValue Op0 = N->getOperand(0);
if (N->getValueType(0) != MVT::i32 || Op0.getValueType().isScalableVT() ||
Op0.getValueType().getVectorElementType() != MVT::i32)
return SDValue();
unsigned ExtOpcode = Op0.getOpcode();
SDValue A = Op0;
SDValue B;
unsigned DotOpcode;
if (ExtOpcode == ISD::MUL) {
A = Op0.getOperand(0);
B = Op0.getOperand(1);
if (A.getOperand(0).getValueType() != B.getOperand(0).getValueType())
return SDValue();
auto OpCodeA = A.getOpcode();
if (OpCodeA != ISD::ZERO_EXTEND && OpCodeA != ISD::SIGN_EXTEND)
return SDValue();
auto OpCodeB = B.getOpcode();
if (OpCodeB != ISD::ZERO_EXTEND && OpCodeB != ISD::SIGN_EXTEND)
return SDValue();
if (OpCodeA == OpCodeB) {
DotOpcode =
OpCodeA == ISD::ZERO_EXTEND ? AArch64ISD::UDOT : AArch64ISD::SDOT;
} else {
// Check USDOT support support
if (!ST->hasMatMulInt8())
return SDValue();
DotOpcode = AArch64ISD::USDOT;
if (OpCodeA == ISD::SIGN_EXTEND)
std::swap(A, B);
}
} else if (ExtOpcode == ISD::ZERO_EXTEND) {
DotOpcode = AArch64ISD::UDOT;
} else if (ExtOpcode == ISD::SIGN_EXTEND) {
DotOpcode = AArch64ISD::SDOT;
} else {
return SDValue();
}
EVT Op0VT = A.getOperand(0).getValueType();
bool IsValidElementCount = Op0VT.getVectorNumElements() % 8 == 0;
bool IsValidSize = Op0VT.getScalarSizeInBits() == 8;
if (!IsValidElementCount || !IsValidSize)
return SDValue();
SDLoc DL(Op0);
// For non-mla reductions B can be set to 1. For MLA we take the operand of
// the extend B.
if (!B)
B = DAG.getConstant(1, DL, Op0VT);
else
B = B.getOperand(0);
unsigned IsMultipleOf16 = Op0VT.getVectorNumElements() % 16 == 0;
unsigned NumOfVecReduce;
EVT TargetType;
if (IsMultipleOf16) {
NumOfVecReduce = Op0VT.getVectorNumElements() / 16;
TargetType = MVT::v4i32;
} else {
NumOfVecReduce = Op0VT.getVectorNumElements() / 8;
TargetType = MVT::v2i32;
}
// Handle the case where we need to generate only one Dot operation.
if (NumOfVecReduce == 1) {
SDValue Zeros = DAG.getConstant(0, DL, TargetType);
SDValue Dot = DAG.getNode(DotOpcode, DL, Zeros.getValueType(), Zeros,
A.getOperand(0), B);
return DAG.getNode(ISD::VECREDUCE_ADD, DL, N->getValueType(0), Dot);
}
// Generate Dot instructions that are multiple of 16.
unsigned VecReduce16Num = Op0VT.getVectorNumElements() / 16;
SmallVector<SDValue, 4> SDotVec16;
unsigned I = 0;
for (; I < VecReduce16Num; I += 1) {
SDValue Zeros = DAG.getConstant(0, DL, MVT::v4i32);
SDValue Op0 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v16i8, A.getOperand(0),
DAG.getConstant(I * 16, DL, MVT::i64));
SDValue Op1 = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v16i8, B,
DAG.getConstant(I * 16, DL, MVT::i64));
SDValue Dot =
DAG.getNode(DotOpcode, DL, Zeros.getValueType(), Zeros, Op0, Op1);
SDotVec16.push_back(Dot);
}
// Concatenate dot operations.
EVT SDot16EVT =
EVT::getVectorVT(*DAG.getContext(), MVT::i32, 4 * VecReduce16Num);
SDValue ConcatSDot16 =
DAG.getNode(ISD::CONCAT_VECTORS, DL, SDot16EVT, SDotVec16);
SDValue VecReduceAdd16 =
DAG.getNode(ISD::VECREDUCE_ADD, DL, N->getValueType(0), ConcatSDot16);
unsigned VecReduce8Num = (Op0VT.getVectorNumElements() % 16) / 8;
if (VecReduce8Num == 0)
return VecReduceAdd16;
// Generate the remainder Dot operation that is multiple of 8.
SDValue Zeros = DAG.getConstant(0, DL, MVT::v2i32);
SDValue Vec8Op0 =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v8i8, A.getOperand(0),
DAG.getConstant(I * 16, DL, MVT::i64));
SDValue Vec8Op1 = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v8i8, B,
DAG.getConstant(I * 16, DL, MVT::i64));
SDValue Dot =
DAG.getNode(DotOpcode, DL, Zeros.getValueType(), Zeros, Vec8Op0, Vec8Op1);
SDValue VecReduceAdd8 =
DAG.getNode(ISD::VECREDUCE_ADD, DL, N->getValueType(0), Dot);
return DAG.getNode(ISD::ADD, DL, N->getValueType(0), VecReduceAdd16,
VecReduceAdd8);
}
// Given an (integer) vecreduce, we know the order of the inputs does not
// matter. We can convert UADDV(add(zext(extract_lo(x)), zext(extract_hi(x))))
// into UADDV(UADDLP(x)). This can also happen through an extra add, where we
// transform UADDV(add(y, add(zext(extract_lo(x)), zext(extract_hi(x))))).
static SDValue performUADDVAddCombine(SDValue A, SelectionDAG &DAG) {
auto DetectAddExtract = [&](SDValue A) {
// Look for add(zext(extract_lo(x)), zext(extract_hi(x))), returning
// UADDLP(x) if found.
assert(A.getOpcode() == ISD::ADD);
EVT VT = A.getValueType();
SDValue Op0 = A.getOperand(0);
SDValue Op1 = A.getOperand(1);
if (Op0.getOpcode() != Op1.getOpcode() ||
(Op0.getOpcode() != ISD::ZERO_EXTEND &&
Op0.getOpcode() != ISD::SIGN_EXTEND))
return SDValue();
SDValue Ext0 = Op0.getOperand(0);
SDValue Ext1 = Op1.getOperand(0);
if (Ext0.getOpcode() != ISD::EXTRACT_SUBVECTOR ||
Ext1.getOpcode() != ISD::EXTRACT_SUBVECTOR ||
Ext0.getOperand(0) != Ext1.getOperand(0))
return SDValue();
// Check that the type is twice the add types, and the extract are from
// upper/lower parts of the same source.
if (Ext0.getOperand(0).getValueType().getVectorNumElements() !=
VT.getVectorNumElements() * 2)
return SDValue();
if ((Ext0.getConstantOperandVal(1) != 0 ||
Ext1.getConstantOperandVal(1) != VT.getVectorNumElements()) &&
(Ext1.getConstantOperandVal(1) != 0 ||
Ext0.getConstantOperandVal(1) != VT.getVectorNumElements()))
return SDValue();
unsigned Opcode = Op0.getOpcode() == ISD::ZERO_EXTEND ? AArch64ISD::UADDLP
: AArch64ISD::SADDLP;
return DAG.getNode(Opcode, SDLoc(A), VT, Ext0.getOperand(0));
};
if (SDValue R = DetectAddExtract(A))
return R;
if (A.getOperand(0).getOpcode() == ISD::ADD && A.getOperand(0).hasOneUse())
if (SDValue R = performUADDVAddCombine(A.getOperand(0), DAG))
return DAG.getNode(ISD::ADD, SDLoc(A), A.getValueType(), R,
A.getOperand(1));
if (A.getOperand(1).getOpcode() == ISD::ADD && A.getOperand(1).hasOneUse())
if (SDValue R = performUADDVAddCombine(A.getOperand(1), DAG))
return DAG.getNode(ISD::ADD, SDLoc(A), A.getValueType(), R,
A.getOperand(0));
return SDValue();
}
// We can convert a UADDV(add(zext(64-bit source), zext(64-bit source))) into
// UADDLV(concat), where the concat represents the 64-bit zext sources.
static SDValue performUADDVZextCombine(SDValue A, SelectionDAG &DAG) {
// Look for add(zext(64-bit source), zext(64-bit source)), returning
// UADDLV(concat(zext, zext)) if found.
assert(A.getOpcode() == ISD::ADD);
EVT VT = A.getValueType();
if (VT != MVT::v8i16 && VT != MVT::v4i32 && VT != MVT::v2i64)
return SDValue();
SDValue Op0 = A.getOperand(0);
SDValue Op1 = A.getOperand(1);
if (Op0.getOpcode() != ISD::ZERO_EXTEND || Op0.getOpcode() != Op1.getOpcode())
return SDValue();
SDValue Ext0 = Op0.getOperand(0);
SDValue Ext1 = Op1.getOperand(0);
EVT ExtVT0 = Ext0.getValueType();
EVT ExtVT1 = Ext1.getValueType();
// Check zext VTs are the same and 64-bit length.
if (ExtVT0 != ExtVT1 ||
VT.getScalarSizeInBits() != (2 * ExtVT0.getScalarSizeInBits()))
return SDValue();
// Get VT for concat of zext sources.
EVT PairVT = ExtVT0.getDoubleNumVectorElementsVT(*DAG.getContext());
SDValue Concat =
DAG.getNode(ISD::CONCAT_VECTORS, SDLoc(A), PairVT, Ext0, Ext1);
switch (VT.getSimpleVT().SimpleTy) {
case MVT::v2i64:
case MVT::v4i32:
return DAG.getNode(AArch64ISD::UADDLV, SDLoc(A), VT, Concat);
case MVT::v8i16: {
SDValue Uaddlv =
DAG.getNode(AArch64ISD::UADDLV, SDLoc(A), MVT::v4i32, Concat);
return DAG.getNode(AArch64ISD::NVCAST, SDLoc(A), MVT::v8i16, Uaddlv);
}
default:
llvm_unreachable("Unhandled vector type");
}
}
static SDValue performUADDVCombine(SDNode *N, SelectionDAG &DAG) {
SDValue A = N->getOperand(0);
if (A.getOpcode() == ISD::ADD) {
if (SDValue R = performUADDVAddCombine(A, DAG))
return DAG.getNode(N->getOpcode(), SDLoc(N), N->getValueType(0), R);
else if (SDValue R = performUADDVZextCombine(A, DAG))
return R;
}
return SDValue();
}
static SDValue performXorCombine(SDNode *N, SelectionDAG &DAG,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
if (DCI.isBeforeLegalizeOps())
return SDValue();
return foldVectorXorShiftIntoCmp(N, DAG, Subtarget);
}
SDValue
AArch64TargetLowering::BuildSDIVPow2(SDNode *N, const APInt &Divisor,
SelectionDAG &DAG,
SmallVectorImpl<SDNode *> &Created) const {
AttributeList Attr = DAG.getMachineFunction().getFunction().getAttributes();
if (isIntDivCheap(N->getValueType(0), Attr))
return SDValue(N, 0); // Lower SDIV as SDIV
EVT VT = N->getValueType(0);
// If SVE is available, we can generate
// sdiv(x,y) -> ptrue + asrd , where 'y' is positive pow-2 divisor.
// sdiv(x,y) -> ptrue + asrd + subr , where 'y' is negative pow-2 divisor.
if (VT.isVector() && Subtarget->isSVEorStreamingSVEAvailable())
return SDValue(N, 0);
// fold (sdiv X, pow2)
if ((VT != MVT::i32 && VT != MVT::i64) ||
!(Divisor.isPowerOf2() || Divisor.isNegatedPowerOf2()))
return SDValue();
// If the divisor is 2 or -2, the default expansion is better. It will add
// (N->getValueType(0) >> (BitWidth - 1)) to it before shifting right.
if (Divisor == 2 ||
Divisor == APInt(Divisor.getBitWidth(), -2, /*isSigned*/ true))
return SDValue();
return TargetLowering::buildSDIVPow2WithCMov(N, Divisor, DAG, Created);
}
SDValue
AArch64TargetLowering::BuildSREMPow2(SDNode *N, const APInt &Divisor,
SelectionDAG &DAG,
SmallVectorImpl<SDNode *> &Created) const {
AttributeList Attr = DAG.getMachineFunction().getFunction().getAttributes();
if (isIntDivCheap(N->getValueType(0), Attr))
return SDValue(N, 0); // Lower SREM as SREM
EVT VT = N->getValueType(0);
// For scalable and fixed types, mark them as cheap so we can handle it much
// later. This allows us to handle larger than legal types.
if (VT.isScalableVector() || Subtarget->useSVEForFixedLengthVectors())
return SDValue(N, 0);
// fold (srem X, pow2)
if ((VT != MVT::i32 && VT != MVT::i64) ||
!(Divisor.isPowerOf2() || Divisor.isNegatedPowerOf2()))
return SDValue();
unsigned Lg2 = Divisor.countr_zero();
if (Lg2 == 0)
return SDValue();
SDLoc DL(N);
SDValue N0 = N->getOperand(0);
SDValue Pow2MinusOne = DAG.getConstant((1ULL << Lg2) - 1, DL, VT);
SDValue Zero = DAG.getConstant(0, DL, VT);
SDValue CCVal, CSNeg;
if (Lg2 == 1) {
SDValue Cmp = getAArch64Cmp(N0, Zero, ISD::SETGE, CCVal, DAG, DL);
SDValue And = DAG.getNode(ISD::AND, DL, VT, N0, Pow2MinusOne);
CSNeg = DAG.getNode(AArch64ISD::CSNEG, DL, VT, And, And, CCVal, Cmp);
Created.push_back(Cmp.getNode());
Created.push_back(And.getNode());
} else {
SDValue CCVal = DAG.getConstant(AArch64CC::MI, DL, MVT_CC);
SDVTList VTs = DAG.getVTList(VT, MVT::i32);
SDValue Negs = DAG.getNode(AArch64ISD::SUBS, DL, VTs, Zero, N0);
SDValue AndPos = DAG.getNode(ISD::AND, DL, VT, N0, Pow2MinusOne);
SDValue AndNeg = DAG.getNode(ISD::AND, DL, VT, Negs, Pow2MinusOne);
CSNeg = DAG.getNode(AArch64ISD::CSNEG, DL, VT, AndPos, AndNeg, CCVal,
Negs.getValue(1));
Created.push_back(Negs.getNode());
Created.push_back(AndPos.getNode());
Created.push_back(AndNeg.getNode());
}
return CSNeg;
}
static std::optional<unsigned> IsSVECntIntrinsic(SDValue S) {
switch(getIntrinsicID(S.getNode())) {
default:
break;
case Intrinsic::aarch64_sve_cntb:
return 8;
case Intrinsic::aarch64_sve_cnth:
return 16;
case Intrinsic::aarch64_sve_cntw:
return 32;
case Intrinsic::aarch64_sve_cntd:
return 64;
}
return {};
}
/// Calculates what the pre-extend type is, based on the extension
/// operation node provided by \p Extend.
///
/// In the case that \p Extend is a SIGN_EXTEND or a ZERO_EXTEND, the
/// pre-extend type is pulled directly from the operand, while other extend
/// operations need a bit more inspection to get this information.
///
/// \param Extend The SDNode from the DAG that represents the extend operation
///
/// \returns The type representing the \p Extend source type, or \p MVT::Other
/// if no valid type can be determined
static EVT calculatePreExtendType(SDValue Extend) {
switch (Extend.getOpcode()) {
case ISD::SIGN_EXTEND:
case ISD::ZERO_EXTEND:
case ISD::ANY_EXTEND:
return Extend.getOperand(0).getValueType();
case ISD::AssertSext:
case ISD::AssertZext:
case ISD::SIGN_EXTEND_INREG: {
VTSDNode *TypeNode = dyn_cast<VTSDNode>(Extend.getOperand(1));
if (!TypeNode)
return MVT::Other;
return TypeNode->getVT();
}
case ISD::AND: {
ConstantSDNode *Constant =
dyn_cast<ConstantSDNode>(Extend.getOperand(1).getNode());
if (!Constant)
return MVT::Other;
uint32_t Mask = Constant->getZExtValue();
if (Mask == UCHAR_MAX)
return MVT::i8;
else if (Mask == USHRT_MAX)
return MVT::i16;
else if (Mask == UINT_MAX)
return MVT::i32;
return MVT::Other;
}
default:
return MVT::Other;
}
}
/// Combines a buildvector(sext/zext) or shuffle(sext/zext, undef) node pattern
/// into sext/zext(buildvector) or sext/zext(shuffle) making use of the vector
/// SExt/ZExt rather than the scalar SExt/ZExt
static SDValue performBuildShuffleExtendCombine(SDValue BV, SelectionDAG &DAG) {
EVT VT = BV.getValueType();
if (BV.getOpcode() != ISD::BUILD_VECTOR &&
BV.getOpcode() != ISD::VECTOR_SHUFFLE)
return SDValue();
// Use the first item in the buildvector/shuffle to get the size of the
// extend, and make sure it looks valid.
SDValue Extend = BV->getOperand(0);
unsigned ExtendOpcode = Extend.getOpcode();
bool IsAnyExt = ExtendOpcode == ISD::ANY_EXTEND;
bool IsSExt = ExtendOpcode == ISD::SIGN_EXTEND ||
ExtendOpcode == ISD::SIGN_EXTEND_INREG ||
ExtendOpcode == ISD::AssertSext;
if (!IsAnyExt && !IsSExt && ExtendOpcode != ISD::ZERO_EXTEND &&
ExtendOpcode != ISD::AssertZext && ExtendOpcode != ISD::AND)
return SDValue();
// Shuffle inputs are vector, limit to SIGN_EXTEND/ZERO_EXTEND/ANY_EXTEND to
// ensure calculatePreExtendType will work without issue.
if (BV.getOpcode() == ISD::VECTOR_SHUFFLE &&
ExtendOpcode != ISD::SIGN_EXTEND && ExtendOpcode != ISD::ZERO_EXTEND)
return SDValue();
// Restrict valid pre-extend data type
EVT PreExtendType = calculatePreExtendType(Extend);
if (PreExtendType == MVT::Other ||
PreExtendType.getScalarSizeInBits() != VT.getScalarSizeInBits() / 2)
return SDValue();
// Make sure all other operands are equally extended.
bool SeenZExtOrSExt = !IsAnyExt;
for (SDValue Op : drop_begin(BV->ops())) {
if (Op.isUndef())
continue;
if (calculatePreExtendType(Op) != PreExtendType)
return SDValue();
unsigned Opc = Op.getOpcode();
if (Opc == ISD::ANY_EXTEND)
continue;
bool OpcIsSExt = Opc == ISD::SIGN_EXTEND || Opc == ISD::SIGN_EXTEND_INREG ||
Opc == ISD::AssertSext;
if (SeenZExtOrSExt && OpcIsSExt != IsSExt)
return SDValue();
IsSExt = OpcIsSExt;
SeenZExtOrSExt = true;
}
SDValue NBV;
SDLoc DL(BV);
if (BV.getOpcode() == ISD::BUILD_VECTOR) {
EVT PreExtendVT = VT.changeVectorElementType(PreExtendType);
EVT PreExtendLegalType =
PreExtendType.getScalarSizeInBits() < 32 ? MVT::i32 : PreExtendType;
SmallVector<SDValue, 8> NewOps;
for (SDValue Op : BV->ops())
NewOps.push_back(Op.isUndef() ? DAG.getUNDEF(PreExtendLegalType)
: DAG.getAnyExtOrTrunc(Op.getOperand(0), DL,
PreExtendLegalType));
NBV = DAG.getNode(ISD::BUILD_VECTOR, DL, PreExtendVT, NewOps);
} else { // BV.getOpcode() == ISD::VECTOR_SHUFFLE
EVT PreExtendVT = VT.changeVectorElementType(PreExtendType.getScalarType());
NBV = DAG.getVectorShuffle(PreExtendVT, DL, BV.getOperand(0).getOperand(0),
BV.getOperand(1).isUndef()
? DAG.getUNDEF(PreExtendVT)
: BV.getOperand(1).getOperand(0),
cast<ShuffleVectorSDNode>(BV)->getMask());
}
unsigned ExtOpc = !SeenZExtOrSExt
? ISD::ANY_EXTEND
: (IsSExt ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND);
return DAG.getNode(ExtOpc, DL, VT, NBV);
}
/// Combines a mul(dup(sext/zext)) node pattern into mul(sext/zext(dup))
/// making use of the vector SExt/ZExt rather than the scalar SExt/ZExt
static SDValue performMulVectorExtendCombine(SDNode *Mul, SelectionDAG &DAG) {
// If the value type isn't a vector, none of the operands are going to be dups
EVT VT = Mul->getValueType(0);
if (VT != MVT::v8i16 && VT != MVT::v4i32 && VT != MVT::v2i64)
return SDValue();
SDValue Op0 = performBuildShuffleExtendCombine(Mul->getOperand(0), DAG);
SDValue Op1 = performBuildShuffleExtendCombine(Mul->getOperand(1), DAG);
// Neither operands have been changed, don't make any further changes
if (!Op0 && !Op1)
return SDValue();
SDLoc DL(Mul);
return DAG.getNode(Mul->getOpcode(), DL, VT, Op0 ? Op0 : Mul->getOperand(0),
Op1 ? Op1 : Mul->getOperand(1));
}
// Combine v4i32 Mul(And(Srl(X, 15), 0x10001), 0xffff) -> v8i16 CMLTz
// Same for other types with equivalent constants.
static SDValue performMulVectorCmpZeroCombine(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (VT != MVT::v2i64 && VT != MVT::v1i64 && VT != MVT::v2i32 &&
VT != MVT::v4i32 && VT != MVT::v4i16 && VT != MVT::v8i16)
return SDValue();
if (N->getOperand(0).getOpcode() != ISD::AND ||
N->getOperand(0).getOperand(0).getOpcode() != ISD::SRL)
return SDValue();
SDValue And = N->getOperand(0);
SDValue Srl = And.getOperand(0);
APInt V1, V2, V3;
if (!ISD::isConstantSplatVector(N->getOperand(1).getNode(), V1) ||
!ISD::isConstantSplatVector(And.getOperand(1).getNode(), V2) ||
!ISD::isConstantSplatVector(Srl.getOperand(1).getNode(), V3))
return SDValue();
unsigned HalfSize = VT.getScalarSizeInBits() / 2;
if (!V1.isMask(HalfSize) || V2 != (1ULL | 1ULL << HalfSize) ||
V3 != (HalfSize - 1))
return SDValue();
EVT HalfVT = EVT::getVectorVT(*DAG.getContext(),
EVT::getIntegerVT(*DAG.getContext(), HalfSize),
VT.getVectorElementCount() * 2);
SDLoc DL(N);
SDValue In = DAG.getNode(AArch64ISD::NVCAST, DL, HalfVT, Srl.getOperand(0));
SDValue Zero = DAG.getConstant(0, DL, In.getValueType());
SDValue CM = DAG.getSetCC(DL, HalfVT, Zero, In, ISD::SETGT);
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, CM);
}
// Transform vector add(zext i8 to i32, zext i8 to i32)
// into sext(add(zext(i8 to i16), zext(i8 to i16)) to i32)
// This allows extra uses of saddl/uaddl at the lower vector widths, and less
// extends.
static SDValue performVectorExtCombine(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (!VT.isFixedLengthVector() || VT.getSizeInBits() <= 128 ||
(N->getOperand(0).getOpcode() != ISD::ZERO_EXTEND &&
N->getOperand(0).getOpcode() != ISD::SIGN_EXTEND) ||
(N->getOperand(1).getOpcode() != ISD::ZERO_EXTEND &&
N->getOperand(1).getOpcode() != ISD::SIGN_EXTEND) ||
N->getOperand(0).getOperand(0).getValueType() !=
N->getOperand(1).getOperand(0).getValueType())
return SDValue();
if (N->getOpcode() == ISD::MUL &&
N->getOperand(0).getOpcode() != N->getOperand(1).getOpcode())
return SDValue();
SDValue N0 = N->getOperand(0).getOperand(0);
SDValue N1 = N->getOperand(1).getOperand(0);
EVT InVT = N0.getValueType();
EVT S1 = InVT.getScalarType();
EVT S2 = VT.getScalarType();
if ((S2 == MVT::i32 && S1 == MVT::i8) ||
(S2 == MVT::i64 && (S1 == MVT::i8 || S1 == MVT::i16))) {
SDLoc DL(N);
EVT HalfVT = EVT::getVectorVT(*DAG.getContext(),
S2.getHalfSizedIntegerVT(*DAG.getContext()),
VT.getVectorElementCount());
SDValue NewN0 = DAG.getNode(N->getOperand(0).getOpcode(), DL, HalfVT, N0);
SDValue NewN1 = DAG.getNode(N->getOperand(1).getOpcode(), DL, HalfVT, N1);
SDValue NewOp = DAG.getNode(N->getOpcode(), DL, HalfVT, NewN0, NewN1);
return DAG.getNode(N->getOpcode() == ISD::MUL ? N->getOperand(0).getOpcode()
: (unsigned)ISD::SIGN_EXTEND,
DL, VT, NewOp);
}
return SDValue();
}
static SDValue performMulCombine(SDNode *N, SelectionDAG &DAG,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
if (SDValue Ext = performMulVectorExtendCombine(N, DAG))
return Ext;
if (SDValue Ext = performMulVectorCmpZeroCombine(N, DAG))
return Ext;
if (SDValue Ext = performVectorExtCombine(N, DAG))
return Ext;
if (DCI.isBeforeLegalizeOps())
return SDValue();
// Canonicalize X*(Y+1) -> X*Y+X and (X+1)*Y -> X*Y+Y,
// and in MachineCombiner pass, add+mul will be combined into madd.
// Similarly, X*(1-Y) -> X - X*Y and (1-Y)*X -> X - Y*X.
SDLoc DL(N);
EVT VT = N->getValueType(0);
SDValue N0 = N->getOperand(0);
SDValue N1 = N->getOperand(1);
SDValue MulOper;
unsigned AddSubOpc;
auto IsAddSubWith1 = [&](SDValue V) -> bool {
AddSubOpc = V->getOpcode();
if ((AddSubOpc == ISD::ADD || AddSubOpc == ISD::SUB) && V->hasOneUse()) {
SDValue Opnd = V->getOperand(1);
MulOper = V->getOperand(0);
if (AddSubOpc == ISD::SUB)
std::swap(Opnd, MulOper);
if (auto C = dyn_cast<ConstantSDNode>(Opnd))
return C->isOne();
}
return false;
};
if (IsAddSubWith1(N0)) {
SDValue MulVal = DAG.getNode(ISD::MUL, DL, VT, N1, MulOper);
return DAG.getNode(AddSubOpc, DL, VT, N1, MulVal);
}
if (IsAddSubWith1(N1)) {
SDValue MulVal = DAG.getNode(ISD::MUL, DL, VT, N0, MulOper);
return DAG.getNode(AddSubOpc, DL, VT, N0, MulVal);
}
// The below optimizations require a constant RHS.
if (!isa<ConstantSDNode>(N1))
return SDValue();
ConstantSDNode *C = cast<ConstantSDNode>(N1);
const APInt &ConstValue = C->getAPIntValue();
// Allow the scaling to be folded into the `cnt` instruction by preventing
// the scaling to be obscured here. This makes it easier to pattern match.
if (IsSVECntIntrinsic(N0) ||
(N0->getOpcode() == ISD::TRUNCATE &&
(IsSVECntIntrinsic(N0->getOperand(0)))))
if (ConstValue.sge(1) && ConstValue.sle(16))
return SDValue();
// Multiplication of a power of two plus/minus one can be done more
// cheaply as shift+add/sub. For now, this is true unilaterally. If
// future CPUs have a cheaper MADD instruction, this may need to be
// gated on a subtarget feature. For Cyclone, 32-bit MADD is 4 cycles and
// 64-bit is 5 cycles, so this is always a win.
// More aggressively, some multiplications N0 * C can be lowered to
// shift+add+shift if the constant C = A * B where A = 2^N + 1 and B = 2^M,
// e.g. 6=3*2=(2+1)*2, 45=(1+4)*(1+8)
// TODO: lower more cases.
// TrailingZeroes is used to test if the mul can be lowered to
// shift+add+shift.
unsigned TrailingZeroes = ConstValue.countr_zero();
if (TrailingZeroes) {
// Conservatively do not lower to shift+add+shift if the mul might be
// folded into smul or umul.
if (N0->hasOneUse() && (isSignExtended(N0, DAG) ||
isZeroExtended(N0, DAG)))
return SDValue();
// Conservatively do not lower to shift+add+shift if the mul might be
// folded into madd or msub.
if (N->hasOneUse() && (N->user_begin()->getOpcode() == ISD::ADD ||
N->user_begin()->getOpcode() == ISD::SUB))
return SDValue();
}
// Use ShiftedConstValue instead of ConstValue to support both shift+add/sub
// and shift+add+shift.
APInt ShiftedConstValue = ConstValue.ashr(TrailingZeroes);
unsigned ShiftAmt;
auto Shl = [&](SDValue N0, unsigned N1) {
if (!N0.getNode())
return SDValue();
// If shift causes overflow, ignore this combine.
if (N1 >= N0.getValueSizeInBits())
return SDValue();
SDValue RHS = DAG.getConstant(N1, DL, MVT::i64);
return DAG.getNode(ISD::SHL, DL, VT, N0, RHS);
};
auto Add = [&](SDValue N0, SDValue N1) {
if (!N0.getNode() || !N1.getNode())
return SDValue();
return DAG.getNode(ISD::ADD, DL, VT, N0, N1);
};
auto Sub = [&](SDValue N0, SDValue N1) {
if (!N0.getNode() || !N1.getNode())
return SDValue();
return DAG.getNode(ISD::SUB, DL, VT, N0, N1);
};
auto Negate = [&](SDValue N) {
if (!N0.getNode())
return SDValue();
SDValue Zero = DAG.getConstant(0, DL, VT);
return DAG.getNode(ISD::SUB, DL, VT, Zero, N);
};
// Can the const C be decomposed into (1+2^M1)*(1+2^N1), eg:
// C = 45 is equal to (1+4)*(1+8), we don't decompose it into (1+2)*(16-1) as
// the (2^N - 1) can't be execused via a single instruction.
auto isPowPlusPlusConst = [](APInt C, APInt &M, APInt &N) {
unsigned BitWidth = C.getBitWidth();
for (unsigned i = 1; i < BitWidth / 2; i++) {
APInt Rem;
APInt X(BitWidth, (1 << i) + 1);
APInt::sdivrem(C, X, N, Rem);
APInt NVMinus1 = N - 1;
if (Rem == 0 && NVMinus1.isPowerOf2()) {
M = X;
return true;
}
}
return false;
};
// Can the const C be decomposed into (2^M + 1) * 2^N + 1), eg:
// C = 11 is equal to (1+4)*2+1, we don't decompose it into (1+2)*4-1 as
// the (2^N - 1) can't be execused via a single instruction.
auto isPowPlusPlusOneConst = [](APInt C, APInt &M, APInt &N) {
APInt CVMinus1 = C - 1;
if (CVMinus1.isNegative())
return false;
unsigned TrailingZeroes = CVMinus1.countr_zero();
APInt SCVMinus1 = CVMinus1.ashr(TrailingZeroes) - 1;
if (SCVMinus1.isPowerOf2()) {
unsigned BitWidth = SCVMinus1.getBitWidth();
M = APInt(BitWidth, SCVMinus1.logBase2());
N = APInt(BitWidth, TrailingZeroes);
return true;
}
return false;
};
// Can the const C be decomposed into (1 - (1 - 2^M) * 2^N), eg:
// C = 29 is equal to 1 - (1 - 2^3) * 2^2.
auto isPowMinusMinusOneConst = [](APInt C, APInt &M, APInt &N) {
APInt CVMinus1 = C - 1;
if (CVMinus1.isNegative())
return false;
unsigned TrailingZeroes = CVMinus1.countr_zero();
APInt CVPlus1 = CVMinus1.ashr(TrailingZeroes) + 1;
if (CVPlus1.isPowerOf2()) {
unsigned BitWidth = CVPlus1.getBitWidth();
M = APInt(BitWidth, CVPlus1.logBase2());
N = APInt(BitWidth, TrailingZeroes);
return true;
}
return false;
};
if (ConstValue.isNonNegative()) {
// (mul x, (2^N + 1) * 2^M) => (shl (add (shl x, N), x), M)
// (mul x, 2^N - 1) => (sub (shl x, N), x)
// (mul x, (2^(N-M) - 1) * 2^M) => (sub (shl x, N), (shl x, M))
// (mul x, (2^M + 1) * (2^N + 1))
// => MV = (add (shl x, M), x); (add (shl MV, N), MV)
// (mul x, (2^M + 1) * 2^N + 1))
// => MV = add (shl x, M), x); add (shl MV, N), x)
// (mul x, 1 - (1 - 2^M) * 2^N))
// => MV = sub (x - (shl x, M)); sub (x - (shl MV, N))
APInt SCVMinus1 = ShiftedConstValue - 1;
APInt SCVPlus1 = ShiftedConstValue + 1;
APInt CVPlus1 = ConstValue + 1;
APInt CVM, CVN;
if (SCVMinus1.isPowerOf2()) {
ShiftAmt = SCVMinus1.logBase2();
return Shl(Add(Shl(N0, ShiftAmt), N0), TrailingZeroes);
} else if (CVPlus1.isPowerOf2()) {
ShiftAmt = CVPlus1.logBase2();
return Sub(Shl(N0, ShiftAmt), N0);
} else if (SCVPlus1.isPowerOf2()) {
ShiftAmt = SCVPlus1.logBase2() + TrailingZeroes;
return Sub(Shl(N0, ShiftAmt), Shl(N0, TrailingZeroes));
}
if (Subtarget->hasALULSLFast() &&
isPowPlusPlusConst(ConstValue, CVM, CVN)) {
APInt CVMMinus1 = CVM - 1;
APInt CVNMinus1 = CVN - 1;
unsigned ShiftM1 = CVMMinus1.logBase2();
unsigned ShiftN1 = CVNMinus1.logBase2();
// ALULSLFast implicate that Shifts <= 4 places are fast
if (ShiftM1 <= 4 && ShiftN1 <= 4) {
SDValue MVal = Add(Shl(N0, ShiftM1), N0);
return Add(Shl(MVal, ShiftN1), MVal);
}
}
if (Subtarget->hasALULSLFast() &&
isPowPlusPlusOneConst(ConstValue, CVM, CVN)) {
unsigned ShiftM = CVM.getZExtValue();
unsigned ShiftN = CVN.getZExtValue();
// ALULSLFast implicate that Shifts <= 4 places are fast
if (ShiftM <= 4 && ShiftN <= 4) {
SDValue MVal = Add(Shl(N0, CVM.getZExtValue()), N0);
return Add(Shl(MVal, CVN.getZExtValue()), N0);
}
}
if (Subtarget->hasALULSLFast() &&
isPowMinusMinusOneConst(ConstValue, CVM, CVN)) {
unsigned ShiftM = CVM.getZExtValue();
unsigned ShiftN = CVN.getZExtValue();
// ALULSLFast implicate that Shifts <= 4 places are fast
if (ShiftM <= 4 && ShiftN <= 4) {
SDValue MVal = Sub(N0, Shl(N0, CVM.getZExtValue()));
return Sub(N0, Shl(MVal, CVN.getZExtValue()));
}
}
} else {
// (mul x, -(2^N - 1)) => (sub x, (shl x, N))
// (mul x, -(2^N + 1)) => - (add (shl x, N), x)
// (mul x, -(2^(N-M) - 1) * 2^M) => (sub (shl x, M), (shl x, N))
APInt SCVPlus1 = -ShiftedConstValue + 1;
APInt CVNegPlus1 = -ConstValue + 1;
APInt CVNegMinus1 = -ConstValue - 1;
if (CVNegPlus1.isPowerOf2()) {
ShiftAmt = CVNegPlus1.logBase2();
return Sub(N0, Shl(N0, ShiftAmt));
} else if (CVNegMinus1.isPowerOf2()) {
ShiftAmt = CVNegMinus1.logBase2();
return Negate(Add(Shl(N0, ShiftAmt), N0));
} else if (SCVPlus1.isPowerOf2()) {
ShiftAmt = SCVPlus1.logBase2() + TrailingZeroes;
return Sub(Shl(N0, TrailingZeroes), Shl(N0, ShiftAmt));
}
}
return SDValue();
}
static SDValue performVectorCompareAndMaskUnaryOpCombine(SDNode *N,
SelectionDAG &DAG) {
// Take advantage of vector comparisons producing 0 or -1 in each lane to
// optimize away operation when it's from a constant.
//
// The general transformation is:
// UNARYOP(AND(VECTOR_CMP(x,y), constant)) -->
// AND(VECTOR_CMP(x,y), constant2)
// constant2 = UNARYOP(constant)
// Early exit if this isn't a vector operation, the operand of the
// unary operation isn't a bitwise AND, or if the sizes of the operations
// aren't the same.
EVT VT = N->getValueType(0);
if (!VT.isVector() || N->getOperand(0)->getOpcode() != ISD::AND ||
N->getOperand(0)->getOperand(0)->getOpcode() != ISD::SETCC ||
VT.getSizeInBits() != N->getOperand(0)->getValueType(0).getSizeInBits())
return SDValue();
// Now check that the other operand of the AND is a constant. We could
// make the transformation for non-constant splats as well, but it's unclear
// that would be a benefit as it would not eliminate any operations, just
// perform one more step in scalar code before moving to the vector unit.
if (BuildVectorSDNode *BV =
dyn_cast<BuildVectorSDNode>(N->getOperand(0)->getOperand(1))) {
// Bail out if the vector isn't a constant.
if (!BV->isConstant())
return SDValue();
// Everything checks out. Build up the new and improved node.
SDLoc DL(N);
EVT IntVT = BV->getValueType(0);
// Create a new constant of the appropriate type for the transformed
// DAG.
SDValue SourceConst = DAG.getNode(N->getOpcode(), DL, VT, SDValue(BV, 0));
// The AND node needs bitcasts to/from an integer vector type around it.
SDValue MaskConst = DAG.getNode(ISD::BITCAST, DL, IntVT, SourceConst);
SDValue NewAnd = DAG.getNode(ISD::AND, DL, IntVT,
N->getOperand(0)->getOperand(0), MaskConst);
SDValue Res = DAG.getNode(ISD::BITCAST, DL, VT, NewAnd);
return Res;
}
return SDValue();
}
/// Tries to replace scalar FP <-> INT conversions with SVE in streaming
/// functions, this can help to reduce the number of fmovs to/from GPRs.
static SDValue
tryToReplaceScalarFPConversionWithSVE(SDNode *N, SelectionDAG &DAG,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
if (N->isStrictFPOpcode())
return SDValue();
if (DCI.isBeforeLegalizeOps())
return SDValue();
if (!Subtarget->isSVEorStreamingSVEAvailable() ||
(!Subtarget->isStreaming() && !Subtarget->isStreamingCompatible()))
return SDValue();
auto isSupportedType = [](EVT VT) {
return !VT.isVector() && VT != MVT::bf16 && VT != MVT::f128;
};
SDValue SrcVal = N->getOperand(0);
EVT SrcTy = SrcVal.getValueType();
EVT DestTy = N->getValueType(0);
if (!isSupportedType(SrcTy) || !isSupportedType(DestTy))
return SDValue();
EVT SrcVecTy;
EVT DestVecTy;
if (DestTy.bitsGT(SrcTy)) {
DestVecTy = getPackedSVEVectorVT(DestTy);
SrcVecTy = DestVecTy.changeVectorElementType(SrcTy);
} else {
SrcVecTy = getPackedSVEVectorVT(SrcTy);
DestVecTy = SrcVecTy.changeVectorElementType(DestTy);
}
// Ensure the resulting src/dest vector type is legal.
if (SrcVecTy == MVT::nxv2i32 || DestVecTy == MVT::nxv2i32)
return SDValue();
SDLoc DL(N);
SDValue ZeroIdx = DAG.getVectorIdxConstant(0, DL);
SDValue Vec = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, SrcVecTy,
DAG.getUNDEF(SrcVecTy), SrcVal, ZeroIdx);
SDValue Convert = DAG.getNode(N->getOpcode(), DL, DestVecTy, Vec);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, DestTy, Convert, ZeroIdx);
}
static SDValue performIntToFpCombine(SDNode *N, SelectionDAG &DAG,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
// First try to optimize away the conversion when it's conditionally from
// a constant. Vectors only.
if (SDValue Res = performVectorCompareAndMaskUnaryOpCombine(N, DAG))
return Res;
if (SDValue Res =
tryToReplaceScalarFPConversionWithSVE(N, DAG, DCI, Subtarget))
return Res;
EVT VT = N->getValueType(0);
if (VT != MVT::f32 && VT != MVT::f64)
return SDValue();
// Only optimize when the source and destination types have the same width.
if (VT.getSizeInBits() != N->getOperand(0).getValueSizeInBits())
return SDValue();
// If the result of an integer load is only used by an integer-to-float
// conversion, use a fp load instead and a AdvSIMD scalar {S|U}CVTF instead.
// This eliminates an "integer-to-vector-move" UOP and improves throughput.
SDValue N0 = N->getOperand(0);
if (Subtarget->isNeonAvailable() && ISD::isNormalLoad(N0.getNode()) &&
N0.hasOneUse() &&
// Do not change the width of a volatile load.
!cast<LoadSDNode>(N0)->isVolatile()) {
LoadSDNode *LN0 = cast<LoadSDNode>(N0);
SDValue Load = DAG.getLoad(VT, SDLoc(N), LN0->getChain(), LN0->getBasePtr(),
LN0->getPointerInfo(), LN0->getAlign(),
LN0->getMemOperand()->getFlags());
// Make sure successors of the original load stay after it by updating them
// to use the new Chain.
DAG.ReplaceAllUsesOfValueWith(SDValue(LN0, 1), Load.getValue(1));
unsigned Opcode =
(N->getOpcode() == ISD::SINT_TO_FP) ? AArch64ISD::SITOF : AArch64ISD::UITOF;
return DAG.getNode(Opcode, SDLoc(N), VT, Load);
}
return SDValue();
}
/// Fold a floating-point multiply by power of two into floating-point to
/// fixed-point conversion.
static SDValue performFpToIntCombine(SDNode *N, SelectionDAG &DAG,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
if (SDValue Res =
tryToReplaceScalarFPConversionWithSVE(N, DAG, DCI, Subtarget))
return Res;
if (!Subtarget->isNeonAvailable())
return SDValue();
if (!N->getValueType(0).isSimple())
return SDValue();
SDValue Op = N->getOperand(0);
if (!Op.getValueType().isSimple() || Op.getOpcode() != ISD::FMUL)
return SDValue();
if (!Op.getValueType().is64BitVector() && !Op.getValueType().is128BitVector())
return SDValue();
SDValue ConstVec = Op->getOperand(1);
if (!isa<BuildVectorSDNode>(ConstVec))
return SDValue();
MVT FloatTy = Op.getSimpleValueType().getVectorElementType();
uint32_t FloatBits = FloatTy.getSizeInBits();
if (FloatBits != 32 && FloatBits != 64 &&
(FloatBits != 16 || !Subtarget->hasFullFP16()))
return SDValue();
MVT IntTy = N->getSimpleValueType(0).getVectorElementType();
uint32_t IntBits = IntTy.getSizeInBits();
if (IntBits != 16 && IntBits != 32 && IntBits != 64)
return SDValue();
// Avoid conversions where iN is larger than the float (e.g., float -> i64).
if (IntBits > FloatBits)
return SDValue();
BitVector UndefElements;
BuildVectorSDNode *BV = cast<BuildVectorSDNode>(ConstVec);
int32_t Bits = IntBits == 64 ? 64 : 32;
int32_t C = BV->getConstantFPSplatPow2ToLog2Int(&UndefElements, Bits + 1);
if (C == -1 || C == 0 || C > Bits)
return SDValue();
EVT ResTy = Op.getValueType().changeVectorElementTypeToInteger();
if (!DAG.getTargetLoweringInfo().isTypeLegal(ResTy))
return SDValue();
if (N->getOpcode() == ISD::FP_TO_SINT_SAT ||
N->getOpcode() == ISD::FP_TO_UINT_SAT) {
EVT SatVT = cast<VTSDNode>(N->getOperand(1))->getVT();
if (SatVT.getScalarSizeInBits() != IntBits || IntBits != FloatBits)
return SDValue();
}
SDLoc DL(N);
bool IsSigned = (N->getOpcode() == ISD::FP_TO_SINT ||
N->getOpcode() == ISD::FP_TO_SINT_SAT);
unsigned IntrinsicOpcode = IsSigned ? Intrinsic::aarch64_neon_vcvtfp2fxs
: Intrinsic::aarch64_neon_vcvtfp2fxu;
SDValue FixConv =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, ResTy,
DAG.getConstant(IntrinsicOpcode, DL, MVT::i32),
Op->getOperand(0), DAG.getConstant(C, DL, MVT::i32));
// We can handle smaller integers by generating an extra trunc.
if (IntBits < FloatBits)
FixConv = DAG.getNode(ISD::TRUNCATE, DL, N->getValueType(0), FixConv);
return FixConv;
}
static SDValue tryCombineToBSL(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
const AArch64TargetLowering &TLI) {
EVT VT = N->getValueType(0);
SelectionDAG &DAG = DCI.DAG;
SDLoc DL(N);
const auto &Subtarget = DAG.getSubtarget<AArch64Subtarget>();
if (!VT.isVector())
return SDValue();
if (VT.isScalableVector() && !Subtarget.hasSVE2())
return SDValue();
if (VT.isFixedLengthVector() &&
(!Subtarget.isNeonAvailable() || TLI.useSVEForFixedLengthVectorVT(VT)))
return SDValue();
SDValue N0 = N->getOperand(0);
if (N0.getOpcode() != ISD::AND)
return SDValue();
SDValue N1 = N->getOperand(1);
if (N1.getOpcode() != ISD::AND)
return SDValue();
// InstCombine does (not (neg a)) => (add a -1).
// Try: (or (and (neg a) b) (and (add a -1) c)) => (bsl (neg a) b c)
// Loop over all combinations of AND operands.
for (int i = 1; i >= 0; --i) {
for (int j = 1; j >= 0; --j) {
SDValue O0 = N0->getOperand(i);
SDValue O1 = N1->getOperand(j);
SDValue Sub, Add, SubSibling, AddSibling;
// Find a SUB and an ADD operand, one from each AND.
if (O0.getOpcode() == ISD::SUB && O1.getOpcode() == ISD::ADD) {
Sub = O0;
Add = O1;
SubSibling = N0->getOperand(1 - i);
AddSibling = N1->getOperand(1 - j);
} else if (O0.getOpcode() == ISD::ADD && O1.getOpcode() == ISD::SUB) {
Add = O0;
Sub = O1;
AddSibling = N0->getOperand(1 - i);
SubSibling = N1->getOperand(1 - j);
} else
continue;
if (!ISD::isConstantSplatVectorAllZeros(Sub.getOperand(0).getNode()))
continue;
// Constant ones is always righthand operand of the Add.
if (!ISD::isConstantSplatVectorAllOnes(Add.getOperand(1).getNode()))
continue;
if (Sub.getOperand(1) != Add.getOperand(0))
continue;
return DAG.getNode(AArch64ISD::BSP, DL, VT, Sub, SubSibling, AddSibling);
}
}
// (or (and a b) (and (not a) c)) => (bsl a b c)
// We only have to look for constant vectors here since the general, variable
// case can be handled in TableGen.
unsigned Bits = VT.getScalarSizeInBits();
uint64_t BitMask = Bits == 64 ? -1ULL : ((1ULL << Bits) - 1);
for (int i = 1; i >= 0; --i)
for (int j = 1; j >= 0; --j) {
APInt Val1, Val2;
if (ISD::isConstantSplatVector(N0->getOperand(i).getNode(), Val1) &&
ISD::isConstantSplatVector(N1->getOperand(j).getNode(), Val2) &&
(BitMask & ~Val1.getZExtValue()) == Val2.getZExtValue()) {
return DAG.getNode(AArch64ISD::BSP, DL, VT, N0->getOperand(i),
N0->getOperand(1 - i), N1->getOperand(1 - j));
}
BuildVectorSDNode *BVN0 = dyn_cast<BuildVectorSDNode>(N0->getOperand(i));
BuildVectorSDNode *BVN1 = dyn_cast<BuildVectorSDNode>(N1->getOperand(j));
if (!BVN0 || !BVN1)
continue;
bool FoundMatch = true;
for (unsigned k = 0; k < VT.getVectorNumElements(); ++k) {
ConstantSDNode *CN0 = dyn_cast<ConstantSDNode>(BVN0->getOperand(k));
ConstantSDNode *CN1 = dyn_cast<ConstantSDNode>(BVN1->getOperand(k));
if (!CN0 || !CN1 ||
CN0->getZExtValue() != (BitMask & ~CN1->getZExtValue())) {
FoundMatch = false;
break;
}
}
if (FoundMatch)
return DAG.getNode(AArch64ISD::BSP, DL, VT, N0->getOperand(i),
N0->getOperand(1 - i), N1->getOperand(1 - j));
}
return SDValue();
}
// Given a tree of and/or(csel(0, 1, cc0), csel(0, 1, cc1)), we may be able to
// convert to csel(ccmp(.., cc0)), depending on cc1:
// (AND (CSET cc0 cmp0) (CSET cc1 (CMP x1 y1)))
// =>
// (CSET cc1 (CCMP x1 y1 !cc1 cc0 cmp0))
//
// (OR (CSET cc0 cmp0) (CSET cc1 (CMP x1 y1)))
// =>
// (CSET cc1 (CCMP x1 y1 cc1 !cc0 cmp0))
static SDValue performANDORCSELCombine(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
SDValue CSel0 = N->getOperand(0);
SDValue CSel1 = N->getOperand(1);
if (CSel0.getOpcode() != AArch64ISD::CSEL ||
CSel1.getOpcode() != AArch64ISD::CSEL)
return SDValue();
if (!CSel0->hasOneUse() || !CSel1->hasOneUse())
return SDValue();
if (!isNullConstant(CSel0.getOperand(0)) ||
!isOneConstant(CSel0.getOperand(1)) ||
!isNullConstant(CSel1.getOperand(0)) ||
!isOneConstant(CSel1.getOperand(1)))
return SDValue();
SDValue Cmp0 = CSel0.getOperand(3);
SDValue Cmp1 = CSel1.getOperand(3);
AArch64CC::CondCode CC0 = (AArch64CC::CondCode)CSel0.getConstantOperandVal(2);
AArch64CC::CondCode CC1 = (AArch64CC::CondCode)CSel1.getConstantOperandVal(2);
if (!Cmp0->hasOneUse() || !Cmp1->hasOneUse())
return SDValue();
if (Cmp1.getOpcode() != AArch64ISD::SUBS &&
Cmp0.getOpcode() == AArch64ISD::SUBS) {
std::swap(Cmp0, Cmp1);
std::swap(CC0, CC1);
}
if (Cmp1.getOpcode() != AArch64ISD::SUBS)
return SDValue();
SDLoc DL(N);
SDValue CCmp, Condition;
unsigned NZCV;
if (N->getOpcode() == ISD::AND) {
AArch64CC::CondCode InvCC0 = AArch64CC::getInvertedCondCode(CC0);
Condition = DAG.getConstant(InvCC0, DL, MVT_CC);
NZCV = AArch64CC::getNZCVToSatisfyCondCode(CC1);
} else {
AArch64CC::CondCode InvCC1 = AArch64CC::getInvertedCondCode(CC1);
Condition = DAG.getConstant(CC0, DL, MVT_CC);
NZCV = AArch64CC::getNZCVToSatisfyCondCode(InvCC1);
}
SDValue NZCVOp = DAG.getConstant(NZCV, DL, MVT::i32);
auto *Op1 = dyn_cast<ConstantSDNode>(Cmp1.getOperand(1));
if (Op1 && Op1->getAPIntValue().isNegative() &&
Op1->getAPIntValue().sgt(-32)) {
// CCMP accept the constant int the range [0, 31]
// if the Op1 is a constant in the range [-31, -1], we
// can select to CCMN to avoid the extra mov
SDValue AbsOp1 =
DAG.getConstant(Op1->getAPIntValue().abs(), DL, Op1->getValueType(0));
CCmp = DAG.getNode(AArch64ISD::CCMN, DL, MVT_CC, Cmp1.getOperand(0), AbsOp1,
NZCVOp, Condition, Cmp0);
} else {
CCmp = DAG.getNode(AArch64ISD::CCMP, DL, MVT_CC, Cmp1.getOperand(0),
Cmp1.getOperand(1), NZCVOp, Condition, Cmp0);
}
return DAG.getNode(AArch64ISD::CSEL, DL, VT, CSel0.getOperand(0),
CSel0.getOperand(1), DAG.getConstant(CC1, DL, MVT::i32),
CCmp);
}
static SDValue performORCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget,
const AArch64TargetLowering &TLI) {
SelectionDAG &DAG = DCI.DAG;
EVT VT = N->getValueType(0);
if (SDValue R = performANDORCSELCombine(N, DAG))
return R;
if (!DAG.getTargetLoweringInfo().isTypeLegal(VT))
return SDValue();
if (SDValue Res = tryCombineToBSL(N, DCI, TLI))
return Res;
return SDValue();
}
static bool isConstantSplatVectorMaskForType(SDNode *N, EVT MemVT) {
if (!MemVT.getVectorElementType().isSimple())
return false;
uint64_t MaskForTy = 0ull;
switch (MemVT.getVectorElementType().getSimpleVT().SimpleTy) {
case MVT::i8:
MaskForTy = 0xffull;
break;
case MVT::i16:
MaskForTy = 0xffffull;
break;
case MVT::i32:
MaskForTy = 0xffffffffull;
break;
default:
return false;
break;
}
if (N->getOpcode() == AArch64ISD::DUP || N->getOpcode() == ISD::SPLAT_VECTOR)
if (auto *Op0 = dyn_cast<ConstantSDNode>(N->getOperand(0)))
return Op0->getAPIntValue().getLimitedValue() == MaskForTy;
return false;
}
static SDValue performReinterpretCastCombine(SDNode *N) {
SDValue LeafOp = SDValue(N, 0);
SDValue Op = N->getOperand(0);
while (Op.getOpcode() == AArch64ISD::REINTERPRET_CAST &&
LeafOp.getValueType() != Op.getValueType())
Op = Op->getOperand(0);
if (LeafOp.getValueType() == Op.getValueType())
return Op;
return SDValue();
}
static SDValue performSVEAndCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
SelectionDAG &DAG = DCI.DAG;
SDValue Src = N->getOperand(0);
unsigned Opc = Src->getOpcode();
// Zero/any extend of an unsigned unpack
if (Opc == AArch64ISD::UUNPKHI || Opc == AArch64ISD::UUNPKLO) {
SDValue UnpkOp = Src->getOperand(0);
SDValue Dup = N->getOperand(1);
if (Dup.getOpcode() != ISD::SPLAT_VECTOR)
return SDValue();
SDLoc DL(N);
ConstantSDNode *C = dyn_cast<ConstantSDNode>(Dup->getOperand(0));
if (!C)
return SDValue();
uint64_t ExtVal = C->getZExtValue();
auto MaskAndTypeMatch = [ExtVal](EVT VT) -> bool {
return ((ExtVal == 0xFF && VT == MVT::i8) ||
(ExtVal == 0xFFFF && VT == MVT::i16) ||
(ExtVal == 0xFFFFFFFF && VT == MVT::i32));
};
// If the mask is fully covered by the unpack, we don't need to push
// a new AND onto the operand
EVT EltTy = UnpkOp->getValueType(0).getVectorElementType();
if (MaskAndTypeMatch(EltTy))
return Src;
// If this is 'and (uunpklo/hi (extload MemTy -> ExtTy)), mask', then check
// to see if the mask is all-ones of size MemTy.
auto MaskedLoadOp = dyn_cast<MaskedLoadSDNode>(UnpkOp);
if (MaskedLoadOp && (MaskedLoadOp->getExtensionType() == ISD::ZEXTLOAD ||
MaskedLoadOp->getExtensionType() == ISD::EXTLOAD)) {
EVT EltTy = MaskedLoadOp->getMemoryVT().getVectorElementType();
if (MaskAndTypeMatch(EltTy))
return Src;
}
// Truncate to prevent a DUP with an over wide constant
APInt Mask = C->getAPIntValue().trunc(EltTy.getSizeInBits());
// Otherwise, make sure we propagate the AND to the operand
// of the unpack
Dup = DAG.getNode(ISD::SPLAT_VECTOR, DL, UnpkOp->getValueType(0),
DAG.getConstant(Mask.zextOrTrunc(32), DL, MVT::i32));
SDValue And = DAG.getNode(ISD::AND, DL,
UnpkOp->getValueType(0), UnpkOp, Dup);
return DAG.getNode(Opc, DL, N->getValueType(0), And);
}
if (DCI.isBeforeLegalizeOps())
return SDValue();
// If both sides of AND operations are i1 splat_vectors then
// we can produce just i1 splat_vector as the result.
if (isAllActivePredicate(DAG, N->getOperand(0)))
return N->getOperand(1);
if (isAllActivePredicate(DAG, N->getOperand(1)))
return N->getOperand(0);
if (!EnableCombineMGatherIntrinsics)
return SDValue();
SDValue Mask = N->getOperand(1);
if (!Src.hasOneUse())
return SDValue();
EVT MemVT;
// SVE load instructions perform an implicit zero-extend, which makes them
// perfect candidates for combining.
switch (Opc) {
case AArch64ISD::LD1_MERGE_ZERO:
case AArch64ISD::LDNF1_MERGE_ZERO:
case AArch64ISD::LDFF1_MERGE_ZERO:
MemVT = cast<VTSDNode>(Src->getOperand(3))->getVT();
break;
case AArch64ISD::GLD1_MERGE_ZERO:
case AArch64ISD::GLD1_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1_SXTW_MERGE_ZERO:
case AArch64ISD::GLD1_SXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1_UXTW_MERGE_ZERO:
case AArch64ISD::GLD1_UXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1_IMM_MERGE_ZERO:
case AArch64ISD::GLDFF1_MERGE_ZERO:
case AArch64ISD::GLDFF1_SCALED_MERGE_ZERO:
case AArch64ISD::GLDFF1_SXTW_MERGE_ZERO:
case AArch64ISD::GLDFF1_SXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLDFF1_UXTW_MERGE_ZERO:
case AArch64ISD::GLDFF1_UXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLDFF1_IMM_MERGE_ZERO:
case AArch64ISD::GLDNT1_MERGE_ZERO:
MemVT = cast<VTSDNode>(Src->getOperand(4))->getVT();
break;
default:
return SDValue();
}
if (isConstantSplatVectorMaskForType(Mask.getNode(), MemVT))
return Src;
return SDValue();
}
// Transform and(fcmp(a, b), fcmp(c, d)) into fccmp(fcmp(a, b), c, d)
static SDValue performANDSETCCCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
// This function performs an optimization on a specific pattern involving
// an AND operation and SETCC (Set Condition Code) node.
SDValue SetCC = N->getOperand(0);
EVT VT = N->getValueType(0);
SelectionDAG &DAG = DCI.DAG;
// Checks if the current node (N) is used by any SELECT instruction and
// returns an empty SDValue to avoid applying the optimization to prevent
// incorrect results
for (auto U : N->users())
if (U->getOpcode() == ISD::SELECT)
return SDValue();
// Check if the operand is a SETCC node with floating-point comparison
if (SetCC.getOpcode() == ISD::SETCC &&
SetCC.getOperand(0).getValueType() == MVT::f32) {
SDValue Cmp;
AArch64CC::CondCode CC;
// Check if the DAG is after legalization and if we can emit the conjunction
if (!DCI.isBeforeLegalize() &&
(Cmp = emitConjunction(DAG, SDValue(N, 0), CC))) {
AArch64CC::CondCode InvertedCC = AArch64CC::getInvertedCondCode(CC);
SDLoc DL(N);
return DAG.getNode(AArch64ISD::CSINC, DL, VT, DAG.getConstant(0, DL, VT),
DAG.getConstant(0, DL, VT),
DAG.getConstant(InvertedCC, DL, MVT::i32), Cmp);
}
}
return SDValue();
}
static SDValue performANDCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
SelectionDAG &DAG = DCI.DAG;
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
EVT VT = N->getValueType(0);
if (SDValue R = performANDORCSELCombine(N, DAG))
return R;
if (SDValue R = performANDSETCCCombine(N,DCI))
return R;
if (!DAG.getTargetLoweringInfo().isTypeLegal(VT))
return SDValue();
if (VT.isScalableVector())
return performSVEAndCombine(N, DCI);
// The combining code below works only for NEON vectors. In particular, it
// does not work for SVE when dealing with vectors wider than 128 bits.
if (!VT.is64BitVector() && !VT.is128BitVector())
return SDValue();
BuildVectorSDNode *BVN = dyn_cast<BuildVectorSDNode>(RHS.getNode());
if (!BVN)
return SDValue();
// AND does not accept an immediate, so check if we can use a BIC immediate
// instruction instead. We do this here instead of using a (and x, (mvni imm))
// pattern in isel, because some immediates may be lowered to the preferred
// (and x, (movi imm)) form, even though an mvni representation also exists.
APInt DefBits(VT.getSizeInBits(), 0);
APInt UndefBits(VT.getSizeInBits(), 0);
if (resolveBuildVector(BVN, DefBits, UndefBits)) {
SDValue NewOp;
// Any bits known to already be 0 need not be cleared again, which can help
// reduce the size of the immediate to one supported by the instruction.
KnownBits Known = DAG.computeKnownBits(LHS);
APInt ZeroSplat(VT.getSizeInBits(), 0);
for (unsigned I = 0; I < VT.getSizeInBits() / Known.Zero.getBitWidth(); I++)
ZeroSplat |= Known.Zero.zext(VT.getSizeInBits())
<< (Known.Zero.getBitWidth() * I);
DefBits = ~(DefBits | ZeroSplat);
if ((NewOp = tryAdvSIMDModImm32(AArch64ISD::BICi, SDValue(N, 0), DAG,
DefBits, &LHS)) ||
(NewOp = tryAdvSIMDModImm16(AArch64ISD::BICi, SDValue(N, 0), DAG,
DefBits, &LHS)))
return NewOp;
UndefBits = ~(UndefBits | ZeroSplat);
if ((NewOp = tryAdvSIMDModImm32(AArch64ISD::BICi, SDValue(N, 0), DAG,
UndefBits, &LHS)) ||
(NewOp = tryAdvSIMDModImm16(AArch64ISD::BICi, SDValue(N, 0), DAG,
UndefBits, &LHS)))
return NewOp;
}
return SDValue();
}
static SDValue performFADDCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
SelectionDAG &DAG = DCI.DAG;
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
EVT VT = N->getValueType(0);
SDLoc DL(N);
if (!N->getFlags().hasAllowReassociation())
return SDValue();
// Combine fadd(a, vcmla(b, c, d)) -> vcmla(fadd(a, b), b, c)
auto ReassocComplex = [&](SDValue A, SDValue B) {
if (A.getOpcode() != ISD::INTRINSIC_WO_CHAIN)
return SDValue();
unsigned Opc = A.getConstantOperandVal(0);
if (Opc != Intrinsic::aarch64_neon_vcmla_rot0 &&
Opc != Intrinsic::aarch64_neon_vcmla_rot90 &&
Opc != Intrinsic::aarch64_neon_vcmla_rot180 &&
Opc != Intrinsic::aarch64_neon_vcmla_rot270)
return SDValue();
SDValue VCMLA = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, VT, A.getOperand(0),
DAG.getNode(ISD::FADD, DL, VT, A.getOperand(1), B, N->getFlags()),
A.getOperand(2), A.getOperand(3));
VCMLA->setFlags(A->getFlags());
return VCMLA;
};
if (SDValue R = ReassocComplex(LHS, RHS))
return R;
if (SDValue R = ReassocComplex(RHS, LHS))
return R;
return SDValue();
}
static bool hasPairwiseAdd(unsigned Opcode, EVT VT, bool FullFP16) {
switch (Opcode) {
case ISD::STRICT_FADD:
case ISD::FADD:
return (FullFP16 && VT == MVT::f16) || VT == MVT::f32 || VT == MVT::f64;
case ISD::ADD:
return VT == MVT::i64;
default:
return false;
}
}
static SDValue getPTest(SelectionDAG &DAG, EVT VT, SDValue Pg, SDValue Op,
AArch64CC::CondCode Cond);
static bool isPredicateCCSettingOp(SDValue N) {
if ((N.getOpcode() == ISD::SETCC) ||
// get_active_lane_mask is lowered to a whilelo instruction.
(N.getOpcode() == ISD::GET_ACTIVE_LANE_MASK) ||
(N.getOpcode() == ISD::INTRINSIC_WO_CHAIN &&
(N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilege ||
N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilegt ||
N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilehi ||
N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilehs ||
N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilele ||
N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilelo ||
N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilels ||
N.getConstantOperandVal(0) == Intrinsic::aarch64_sve_whilelt)))
return true;
return false;
}
// Materialize : i1 = extract_vector_elt t37, Constant:i64<0>
// ... into: "ptrue p, all" + PTEST
static SDValue
performFirstTrueTestVectorCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
assert(N->getOpcode() == ISD::EXTRACT_VECTOR_ELT);
// Make sure PTEST can be legalised with illegal types.
if (!Subtarget->hasSVE() || DCI.isBeforeLegalize())
return SDValue();
SDValue N0 = N->getOperand(0);
EVT VT = N0.getValueType();
if (!VT.isScalableVector() || VT.getVectorElementType() != MVT::i1 ||
!isNullConstant(N->getOperand(1)))
return SDValue();
// Restricted the DAG combine to only cases where we're extracting from a
// flag-setting operation.
if (!isPredicateCCSettingOp(N0))
return SDValue();
// Extracts of lane 0 for SVE can be expressed as PTEST(Op, FIRST) ? 1 : 0
SelectionDAG &DAG = DCI.DAG;
SDValue Pg = getPTrue(DAG, SDLoc(N), VT, AArch64SVEPredPattern::all);
return getPTest(DAG, N->getValueType(0), Pg, N0, AArch64CC::FIRST_ACTIVE);
}
// Materialize : Idx = (add (mul vscale, NumEls), -1)
// i1 = extract_vector_elt t37, Constant:i64<Idx>
// ... into: "ptrue p, all" + PTEST
static SDValue
performLastTrueTestVectorCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
assert(N->getOpcode() == ISD::EXTRACT_VECTOR_ELT);
// Make sure PTEST is legal types.
if (!Subtarget->hasSVE() || DCI.isBeforeLegalize())
return SDValue();
SDValue N0 = N->getOperand(0);
EVT OpVT = N0.getValueType();
if (!OpVT.isScalableVector() || OpVT.getVectorElementType() != MVT::i1)
return SDValue();
// Idx == (add (mul vscale, NumEls), -1)
SDValue Idx = N->getOperand(1);
if (Idx.getOpcode() != ISD::ADD || !isAllOnesConstant(Idx.getOperand(1)))
return SDValue();
SDValue VS = Idx.getOperand(0);
if (VS.getOpcode() != ISD::VSCALE)
return SDValue();
unsigned NumEls = OpVT.getVectorElementCount().getKnownMinValue();
if (VS.getConstantOperandVal(0) != NumEls)
return SDValue();
// Extracts of lane EC-1 for SVE can be expressed as PTEST(Op, LAST) ? 1 : 0
SelectionDAG &DAG = DCI.DAG;
SDValue Pg = getPTrue(DAG, SDLoc(N), OpVT, AArch64SVEPredPattern::all);
return getPTest(DAG, N->getValueType(0), Pg, N0, AArch64CC::LAST_ACTIVE);
}
static SDValue
performExtractLastActiveCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
assert(N->getOpcode() == ISD::EXTRACT_VECTOR_ELT);
SelectionDAG &DAG = DCI.DAG;
SDValue Vec = N->getOperand(0);
SDValue Idx = N->getOperand(1);
if (DCI.isBeforeLegalize() || Idx.getOpcode() != ISD::VECTOR_FIND_LAST_ACTIVE)
return SDValue();
// Only legal for 8, 16, 32, and 64 bit element types.
EVT EltVT = Vec.getValueType().getVectorElementType();
if (!is_contained(ArrayRef({MVT::i8, MVT::i16, MVT::i32, MVT::i64, MVT::f16,
MVT::bf16, MVT::f32, MVT::f64}),
EltVT.getSimpleVT().SimpleTy))
return SDValue();
SDValue Mask = Idx.getOperand(0);
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
if (!TLI.isOperationLegal(ISD::VECTOR_FIND_LAST_ACTIVE, Mask.getValueType()))
return SDValue();
return DAG.getNode(AArch64ISD::LASTB, SDLoc(N), N->getValueType(0), Mask,
Vec);
}
static SDValue
performExtractVectorEltCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
assert(N->getOpcode() == ISD::EXTRACT_VECTOR_ELT);
if (SDValue Res = performFirstTrueTestVectorCombine(N, DCI, Subtarget))
return Res;
if (SDValue Res = performLastTrueTestVectorCombine(N, DCI, Subtarget))
return Res;
if (SDValue Res = performExtractLastActiveCombine(N, DCI, Subtarget))
return Res;
SelectionDAG &DAG = DCI.DAG;
SDValue N0 = N->getOperand(0), N1 = N->getOperand(1);
EVT VT = N->getValueType(0);
const bool FullFP16 = Subtarget->hasFullFP16();
bool IsStrict = N0->isStrictFPOpcode();
// extract(dup x) -> x
if (N0.getOpcode() == AArch64ISD::DUP)
return VT.isInteger() ? DAG.getZExtOrTrunc(N0.getOperand(0), SDLoc(N), VT)
: N0.getOperand(0);
// Rewrite for pairwise fadd pattern
// (f32 (extract_vector_elt
// (fadd (vXf32 Other)
// (vector_shuffle (vXf32 Other) undef <1,X,...> )) 0))
// ->
// (f32 (fadd (extract_vector_elt (vXf32 Other) 0)
// (extract_vector_elt (vXf32 Other) 1))
// For strict_fadd we need to make sure the old strict_fadd can be deleted, so
// we can only do this when it's used only by the extract_vector_elt.
if (isNullConstant(N1) && hasPairwiseAdd(N0->getOpcode(), VT, FullFP16) &&
(!IsStrict || N0.hasOneUse())) {
SDLoc DL(N0);
SDValue N00 = N0->getOperand(IsStrict ? 1 : 0);
SDValue N01 = N0->getOperand(IsStrict ? 2 : 1);
ShuffleVectorSDNode *Shuffle = dyn_cast<ShuffleVectorSDNode>(N01);
SDValue Other = N00;
// And handle the commutative case.
if (!Shuffle) {
Shuffle = dyn_cast<ShuffleVectorSDNode>(N00);
Other = N01;
}
if (Shuffle && Shuffle->getMaskElt(0) == 1 &&
Other == Shuffle->getOperand(0)) {
SDValue Extract1 = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, Other,
DAG.getConstant(0, DL, MVT::i64));
SDValue Extract2 = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, Other,
DAG.getConstant(1, DL, MVT::i64));
if (!IsStrict)
return DAG.getNode(N0->getOpcode(), DL, VT, Extract1, Extract2);
// For strict_fadd we need uses of the final extract_vector to be replaced
// with the strict_fadd, but we also need uses of the chain output of the
// original strict_fadd to use the chain output of the new strict_fadd as
// otherwise it may not be deleted.
SDValue Ret = DAG.getNode(N0->getOpcode(), DL,
{VT, MVT::Other},
{N0->getOperand(0), Extract1, Extract2});
DAG.ReplaceAllUsesOfValueWith(SDValue(N, 0), Ret);
DAG.ReplaceAllUsesOfValueWith(N0.getValue(1), Ret.getValue(1));
return SDValue(N, 0);
}
}
return SDValue();
}
static SDValue performConcatVectorsCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
SDLoc DL(N);
EVT VT = N->getValueType(0);
SDValue N0 = N->getOperand(0), N1 = N->getOperand(1);
unsigned N0Opc = N0->getOpcode(), N1Opc = N1->getOpcode();
if (VT.isScalableVector())
return SDValue();
if (N->getNumOperands() == 2 && N0Opc == ISD::TRUNCATE &&
N1Opc == ISD::TRUNCATE) {
SDValue N00 = N0->getOperand(0);
SDValue N10 = N1->getOperand(0);
EVT N00VT = N00.getValueType();
unsigned N00Opc = N00.getOpcode(), N10Opc = N10.getOpcode();
// Optimize concat_vectors of truncated vectors, where the intermediate
// type is illegal, to avoid said illegality, e.g.,
// (v4i16 (concat_vectors (v2i16 (truncate (v2i64))),
// (v2i16 (truncate (v2i64)))))
// ->
// (v4i16 (truncate (vector_shuffle (v4i32 (bitcast (v2i64))),
// (v4i32 (bitcast (v2i64))),
// <0, 2, 4, 6>)))
// This isn't really target-specific, but ISD::TRUNCATE legality isn't keyed
// on both input and result type, so we might generate worse code.
// On AArch64 we know it's fine for v2i64->v4i16 and v4i32->v8i8.
if (N00VT == N10.getValueType() &&
(N00VT == MVT::v2i64 || N00VT == MVT::v4i32) &&
N00VT.getScalarSizeInBits() == 4 * VT.getScalarSizeInBits()) {
MVT MidVT = (N00VT == MVT::v2i64 ? MVT::v4i32 : MVT::v8i16);
SmallVector<int, 8> Mask(MidVT.getVectorNumElements());
for (size_t i = 0; i < Mask.size(); ++i)
Mask[i] = i * 2;
return DAG.getNode(ISD::TRUNCATE, DL, VT,
DAG.getVectorShuffle(
MidVT, DL,
DAG.getNode(ISD::BITCAST, DL, MidVT, N00),
DAG.getNode(ISD::BITCAST, DL, MidVT, N10), Mask));
}
// Optimize two large shifts and a combine into a single combine and shift
// For AArch64 architectures, sequences like the following:
//
// ushr v0.4s, v0.4s, #20
// ushr v1.4s, v1.4s, #20
// uzp1 v0.8h, v0.8h, v1.8h
//
// Can be optimized to:
//
// uzp2 v0.8h, v0.8h, v1.8h
// ushr v0.8h, v0.8h, #4
//
// This optimization reduces instruction count.
if (N00Opc == AArch64ISD::VLSHR && N10Opc == AArch64ISD::VLSHR &&
N00->getOperand(1) == N10->getOperand(1)) {
SDValue N000 = N00->getOperand(0);
SDValue N100 = N10->getOperand(0);
uint64_t N001ConstVal = N00->getConstantOperandVal(1),
N101ConstVal = N10->getConstantOperandVal(1),
NScalarSize = N->getValueType(0).getScalarSizeInBits();
if (N001ConstVal == N101ConstVal && N001ConstVal > NScalarSize) {
N000 = DAG.getNode(AArch64ISD::NVCAST, DL, VT, N000);
N100 = DAG.getNode(AArch64ISD::NVCAST, DL, VT, N100);
SDValue Uzp = DAG.getNode(AArch64ISD::UZP2, DL, VT, N000, N100);
SDValue NewShiftConstant =
DAG.getConstant(N001ConstVal - NScalarSize, DL, MVT::i32);
return DAG.getNode(AArch64ISD::VLSHR, DL, VT, Uzp, NewShiftConstant);
}
}
}
if (N->getOperand(0).getValueType() == MVT::v4i8 ||
N->getOperand(0).getValueType() == MVT::v2i16 ||
N->getOperand(0).getValueType() == MVT::v2i8) {
EVT SrcVT = N->getOperand(0).getValueType();
// If we have a concat of v4i8 loads, convert them to a buildvector of f32
// loads to prevent having to go through the v4i8 load legalization that
// needs to extend each element into a larger type.
if (N->getNumOperands() % 2 == 0 &&
all_of(N->op_values(), [SrcVT](SDValue V) {
if (V.getValueType() != SrcVT)
return false;
if (V.isUndef())
return true;
LoadSDNode *LD = dyn_cast<LoadSDNode>(V);
return LD && V.hasOneUse() && LD->isSimple() && !LD->isIndexed() &&
LD->getExtensionType() == ISD::NON_EXTLOAD;
})) {
EVT FVT = SrcVT == MVT::v2i8 ? MVT::f16 : MVT::f32;
EVT NVT = EVT::getVectorVT(*DAG.getContext(), FVT, N->getNumOperands());
SmallVector<SDValue> Ops;
for (unsigned i = 0; i < N->getNumOperands(); i++) {
SDValue V = N->getOperand(i);
if (V.isUndef())
Ops.push_back(DAG.getUNDEF(FVT));
else {
LoadSDNode *LD = cast<LoadSDNode>(V);
SDValue NewLoad = DAG.getLoad(FVT, DL, LD->getChain(),
LD->getBasePtr(), LD->getMemOperand());
DAG.ReplaceAllUsesOfValueWith(SDValue(LD, 1), NewLoad.getValue(1));
Ops.push_back(NewLoad);
}
}
return DAG.getBitcast(N->getValueType(0),
DAG.getBuildVector(NVT, DL, Ops));
}
}
// Canonicalise concat_vectors to replace concatenations of truncated nots
// with nots of concatenated truncates. This in some cases allows for multiple
// redundant negations to be eliminated.
// (concat_vectors (v4i16 (truncate (not (v4i32)))),
// (v4i16 (truncate (not (v4i32)))))
// ->
// (not (concat_vectors (v4i16 (truncate (v4i32))),
// (v4i16 (truncate (v4i32)))))
if (N->getNumOperands() == 2 && N0Opc == ISD::TRUNCATE &&
N1Opc == ISD::TRUNCATE && N->isOnlyUserOf(N0.getNode()) &&
N->isOnlyUserOf(N1.getNode())) {
auto isBitwiseVectorNegate = [](SDValue V) {
return V->getOpcode() == ISD::XOR &&
ISD::isConstantSplatVectorAllOnes(V.getOperand(1).getNode());
};
SDValue N00 = N0->getOperand(0);
SDValue N10 = N1->getOperand(0);
if (isBitwiseVectorNegate(N00) && N0->isOnlyUserOf(N00.getNode()) &&
isBitwiseVectorNegate(N10) && N1->isOnlyUserOf(N10.getNode())) {
return DAG.getNOT(
DL,
DAG.getNode(ISD::CONCAT_VECTORS, DL, VT,
DAG.getNode(ISD::TRUNCATE, DL, N0.getValueType(),
N00->getOperand(0)),
DAG.getNode(ISD::TRUNCATE, DL, N1.getValueType(),
N10->getOperand(0))),
VT);
}
}
// Wait till after everything is legalized to try this. That way we have
// legal vector types and such.
if (DCI.isBeforeLegalizeOps())
return SDValue();
// Optimise concat_vectors of two identical binops with a 128-bit destination
// size, combine into an binop of two contacts of the source vectors. eg:
// concat(uhadd(a,b), uhadd(c, d)) -> uhadd(concat(a, c), concat(b, d))
if (N->getNumOperands() == 2 && N0Opc == N1Opc && VT.is128BitVector() &&
DAG.getTargetLoweringInfo().isBinOp(N0Opc) && N0->hasOneUse() &&
N1->hasOneUse()) {
SDValue N00 = N0->getOperand(0);
SDValue N01 = N0->getOperand(1);
SDValue N10 = N1->getOperand(0);
SDValue N11 = N1->getOperand(1);
if (!N00.isUndef() && !N01.isUndef() && !N10.isUndef() && !N11.isUndef()) {
SDValue Concat0 = DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, N00, N10);
SDValue Concat1 = DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, N01, N11);
return DAG.getNode(N0Opc, DL, VT, Concat0, Concat1);
}
}
auto IsRSHRN = [](SDValue Shr) {
if (Shr.getOpcode() != AArch64ISD::VLSHR)
return false;
SDValue Op = Shr.getOperand(0);
EVT VT = Op.getValueType();
unsigned ShtAmt = Shr.getConstantOperandVal(1);
if (ShtAmt > VT.getScalarSizeInBits() / 2 || Op.getOpcode() != ISD::ADD)
return false;
APInt Imm;
if (Op.getOperand(1).getOpcode() == AArch64ISD::MOVIshift)
Imm = APInt(VT.getScalarSizeInBits(),
Op.getOperand(1).getConstantOperandVal(0)
<< Op.getOperand(1).getConstantOperandVal(1));
else if (Op.getOperand(1).getOpcode() == AArch64ISD::DUP &&
isa<ConstantSDNode>(Op.getOperand(1).getOperand(0)))
Imm = APInt(VT.getScalarSizeInBits(),
Op.getOperand(1).getConstantOperandVal(0));
else
return false;
if (Imm != 1ULL << (ShtAmt - 1))
return false;
return true;
};
// concat(rshrn(x), rshrn(y)) -> rshrn(concat(x, y))
if (N->getNumOperands() == 2 && IsRSHRN(N0) &&
((IsRSHRN(N1) &&
N0.getConstantOperandVal(1) == N1.getConstantOperandVal(1)) ||
N1.isUndef())) {
SDValue X = N0.getOperand(0).getOperand(0);
SDValue Y = N1.isUndef() ? DAG.getUNDEF(X.getValueType())
: N1.getOperand(0).getOperand(0);
EVT BVT =
X.getValueType().getDoubleNumVectorElementsVT(*DCI.DAG.getContext());
SDValue CC = DAG.getNode(ISD::CONCAT_VECTORS, DL, BVT, X, Y);
SDValue Add = DAG.getNode(
ISD::ADD, DL, BVT, CC,
DAG.getConstant(1ULL << (N0.getConstantOperandVal(1) - 1), DL, BVT));
SDValue Shr =
DAG.getNode(AArch64ISD::VLSHR, DL, BVT, Add, N0.getOperand(1));
return Shr;
}
// concat(zip1(a, b), zip2(a, b)) is zip1(a, b)
if (N->getNumOperands() == 2 && N0Opc == AArch64ISD::ZIP1 &&
N1Opc == AArch64ISD::ZIP2 && N0.getOperand(0) == N1.getOperand(0) &&
N0.getOperand(1) == N1.getOperand(1)) {
SDValue E0 = DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, N0.getOperand(0),
DAG.getUNDEF(N0.getValueType()));
SDValue E1 = DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, N0.getOperand(1),
DAG.getUNDEF(N0.getValueType()));
return DAG.getNode(AArch64ISD::ZIP1, DL, VT, E0, E1);
}
// If we see a (concat_vectors (v1x64 A), (v1x64 A)) it's really a vector
// splat. The indexed instructions are going to be expecting a DUPLANE64, so
// canonicalise to that.
if (N->getNumOperands() == 2 && N0 == N1 && VT.getVectorNumElements() == 2) {
assert(VT.getScalarSizeInBits() == 64);
return DAG.getNode(AArch64ISD::DUPLANE64, DL, VT, WidenVector(N0, DAG),
DAG.getConstant(0, DL, MVT::i64));
}
// Canonicalise concat_vectors so that the right-hand vector has as few
// bit-casts as possible before its real operation. The primary matching
// destination for these operations will be the narrowing "2" instructions,
// which depend on the operation being performed on this right-hand vector.
// For example,
// (concat_vectors LHS, (v1i64 (bitconvert (v4i16 RHS))))
// becomes
// (bitconvert (concat_vectors (v4i16 (bitconvert LHS)), RHS))
if (N->getNumOperands() != 2 || N1Opc != ISD::BITCAST)
return SDValue();
SDValue RHS = N1->getOperand(0);
MVT RHSTy = RHS.getValueType().getSimpleVT();
// If the RHS is not a vector, this is not the pattern we're looking for.
if (!RHSTy.isVector())
return SDValue();
LLVM_DEBUG(
dbgs() << "aarch64-lower: concat_vectors bitcast simplification\n");
MVT ConcatTy = MVT::getVectorVT(RHSTy.getVectorElementType(),
RHSTy.getVectorNumElements() * 2);
return DAG.getNode(ISD::BITCAST, DL, VT,
DAG.getNode(ISD::CONCAT_VECTORS, DL, ConcatTy,
DAG.getNode(ISD::BITCAST, DL, RHSTy, N0),
RHS));
}
static SDValue
performExtractSubvectorCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (DCI.isBeforeLegalizeOps())
return SDValue();
EVT VT = N->getValueType(0);
if (!VT.isScalableVector() || VT.getVectorElementType() != MVT::i1)
return SDValue();
SDValue V = N->getOperand(0);
// NOTE: This combine exists in DAGCombiner, but that version's legality check
// blocks this combine because the non-const case requires custom lowering.
//
// ty1 extract_vector(ty2 splat(const))) -> ty1 splat(const)
if (V.getOpcode() == ISD::SPLAT_VECTOR)
if (isa<ConstantSDNode>(V.getOperand(0)))
return DAG.getNode(ISD::SPLAT_VECTOR, SDLoc(N), VT, V.getOperand(0));
return SDValue();
}
static SDValue
performInsertSubvectorCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Vec = N->getOperand(0);
SDValue SubVec = N->getOperand(1);
uint64_t IdxVal = N->getConstantOperandVal(2);
EVT VecVT = Vec.getValueType();
EVT SubVT = SubVec.getValueType();
// Promote fixed length vector zeros.
if (VecVT.isScalableVector() && SubVT.isFixedLengthVector() &&
Vec.isUndef() && isZerosVector(SubVec.getNode()))
return VecVT.isInteger() ? DAG.getConstant(0, DL, VecVT)
: DAG.getConstantFP(0, DL, VecVT);
// Only do this for legal fixed vector types.
if (!VecVT.isFixedLengthVector() ||
!DAG.getTargetLoweringInfo().isTypeLegal(VecVT) ||
!DAG.getTargetLoweringInfo().isTypeLegal(SubVT))
return SDValue();
// Ignore widening patterns.
if (IdxVal == 0 && Vec.isUndef())
return SDValue();
// Subvector must be half the width and an "aligned" insertion.
unsigned NumSubElts = SubVT.getVectorNumElements();
if ((SubVT.getSizeInBits() * 2) != VecVT.getSizeInBits() ||
(IdxVal != 0 && IdxVal != NumSubElts))
return SDValue();
// Fold insert_subvector -> concat_vectors
// insert_subvector(Vec,Sub,lo) -> concat_vectors(Sub,extract(Vec,hi))
// insert_subvector(Vec,Sub,hi) -> concat_vectors(extract(Vec,lo),Sub)
SDValue Lo, Hi;
if (IdxVal == 0) {
Lo = SubVec;
Hi = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, SubVT, Vec,
DAG.getVectorIdxConstant(NumSubElts, DL));
} else {
Lo = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, SubVT, Vec,
DAG.getVectorIdxConstant(0, DL));
Hi = SubVec;
}
return DAG.getNode(ISD::CONCAT_VECTORS, DL, VecVT, Lo, Hi);
}
static SDValue tryCombineFixedPointConvert(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
// Wait until after everything is legalized to try this. That way we have
// legal vector types and such.
if (DCI.isBeforeLegalizeOps())
return SDValue();
// Transform a scalar conversion of a value from a lane extract into a
// lane extract of a vector conversion. E.g., from foo1 to foo2:
// double foo1(int64x2_t a) { return vcvtd_n_f64_s64(a[1], 9); }
// double foo2(int64x2_t a) { return vcvtq_n_f64_s64(a, 9)[1]; }
//
// The second form interacts better with instruction selection and the
// register allocator to avoid cross-class register copies that aren't
// coalescable due to a lane reference.
// Check the operand and see if it originates from a lane extract.
SDValue Op1 = N->getOperand(1);
if (Op1.getOpcode() != ISD::EXTRACT_VECTOR_ELT)
return SDValue();
// Yep, no additional predication needed. Perform the transform.
SDValue IID = N->getOperand(0);
SDValue Shift = N->getOperand(2);
SDValue Vec = Op1.getOperand(0);
SDValue Lane = Op1.getOperand(1);
EVT ResTy = N->getValueType(0);
EVT VecResTy;
SDLoc DL(N);
// The vector width should be 128 bits by the time we get here, even
// if it started as 64 bits (the extract_vector handling will have
// done so). Bail if it is not.
if (Vec.getValueSizeInBits() != 128)
return SDValue();
if (Vec.getValueType() == MVT::v4i32)
VecResTy = MVT::v4f32;
else if (Vec.getValueType() == MVT::v2i64)
VecResTy = MVT::v2f64;
else
return SDValue();
SDValue Convert =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, VecResTy, IID, Vec, Shift);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ResTy, Convert, Lane);
}
// AArch64 high-vector "long" operations are formed by performing the non-high
// version on an extract_subvector of each operand which gets the high half:
//
// (longop2 LHS, RHS) == (longop (extract_high LHS), (extract_high RHS))
//
// However, there are cases which don't have an extract_high explicitly, but
// have another operation that can be made compatible with one for free. For
// example:
//
// (dupv64 scalar) --> (extract_high (dup128 scalar))
//
// This routine does the actual conversion of such DUPs, once outer routines
// have determined that everything else is in order.
// It also supports immediate DUP-like nodes (MOVI/MVNi), which we can fold
// similarly here.
static SDValue tryExtendDUPToExtractHigh(SDValue N, SelectionDAG &DAG) {
MVT VT = N.getSimpleValueType();
if (N.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
N.getConstantOperandVal(1) == 0)
N = N.getOperand(0);
switch (N.getOpcode()) {
case AArch64ISD::DUP:
case AArch64ISD::DUPLANE8:
case AArch64ISD::DUPLANE16:
case AArch64ISD::DUPLANE32:
case AArch64ISD::DUPLANE64:
case AArch64ISD::MOVI:
case AArch64ISD::MOVIshift:
case AArch64ISD::MOVIedit:
case AArch64ISD::MOVImsl:
case AArch64ISD::MVNIshift:
case AArch64ISD::MVNImsl:
break;
default:
// FMOV could be supported, but isn't very useful, as it would only occur
// if you passed a bitcast' floating point immediate to an eligible long
// integer op (addl, smull, ...).
return SDValue();
}
if (!VT.is64BitVector())
return SDValue();
SDLoc DL(N);
unsigned NumElems = VT.getVectorNumElements();
if (N.getValueType().is64BitVector()) {
MVT ElementTy = VT.getVectorElementType();
MVT NewVT = MVT::getVectorVT(ElementTy, NumElems * 2);
N = DAG.getNode(N->getOpcode(), DL, NewVT, N->ops());
}
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, N,
DAG.getConstant(NumElems, DL, MVT::i64));
}
static bool isEssentiallyExtractHighSubvector(SDValue N) {
if (N.getOpcode() == ISD::BITCAST)
N = N.getOperand(0);
if (N.getOpcode() != ISD::EXTRACT_SUBVECTOR)
return false;
if (N.getOperand(0).getValueType().isScalableVector())
return false;
return N.getConstantOperandAPInt(1) ==
N.getOperand(0).getValueType().getVectorNumElements() / 2;
}
/// Helper structure to keep track of ISD::SET_CC operands.
struct GenericSetCCInfo {
const SDValue *Opnd0;
const SDValue *Opnd1;
ISD::CondCode CC;
};
/// Helper structure to keep track of a SET_CC lowered into AArch64 code.
struct AArch64SetCCInfo {
const SDValue *Cmp;
AArch64CC::CondCode CC;
};
/// Helper structure to keep track of SetCC information.
union SetCCInfo {
GenericSetCCInfo Generic;
AArch64SetCCInfo AArch64;
};
/// Helper structure to be able to read SetCC information. If set to
/// true, IsAArch64 field, Info is a AArch64SetCCInfo, otherwise Info is a
/// GenericSetCCInfo.
struct SetCCInfoAndKind {
SetCCInfo Info;
bool IsAArch64;
};
/// Check whether or not \p Op is a SET_CC operation, either a generic or
/// an
/// AArch64 lowered one.
/// \p SetCCInfo is filled accordingly.
/// \post SetCCInfo is meanginfull only when this function returns true.
/// \return True when Op is a kind of SET_CC operation.
static bool isSetCC(SDValue Op, SetCCInfoAndKind &SetCCInfo) {
// If this is a setcc, this is straight forward.
if (Op.getOpcode() == ISD::SETCC) {
SetCCInfo.Info.Generic.Opnd0 = &Op.getOperand(0);
SetCCInfo.Info.Generic.Opnd1 = &Op.getOperand(1);
SetCCInfo.Info.Generic.CC = cast<CondCodeSDNode>(Op.getOperand(2))->get();
SetCCInfo.IsAArch64 = false;
return true;
}
// Otherwise, check if this is a matching csel instruction.
// In other words:
// - csel 1, 0, cc
// - csel 0, 1, !cc
if (Op.getOpcode() != AArch64ISD::CSEL)
return false;
// Set the information about the operands.
// TODO: we want the operands of the Cmp not the csel
SetCCInfo.Info.AArch64.Cmp = &Op.getOperand(3);
SetCCInfo.IsAArch64 = true;
SetCCInfo.Info.AArch64.CC =
static_cast<AArch64CC::CondCode>(Op.getConstantOperandVal(2));
// Check that the operands matches the constraints:
// (1) Both operands must be constants.
// (2) One must be 1 and the other must be 0.
ConstantSDNode *TValue = dyn_cast<ConstantSDNode>(Op.getOperand(0));
ConstantSDNode *FValue = dyn_cast<ConstantSDNode>(Op.getOperand(1));
// Check (1).
if (!TValue || !FValue)
return false;
// Check (2).
if (!TValue->isOne()) {
// Update the comparison when we are interested in !cc.
std::swap(TValue, FValue);
SetCCInfo.Info.AArch64.CC =
AArch64CC::getInvertedCondCode(SetCCInfo.Info.AArch64.CC);
}
return TValue->isOne() && FValue->isZero();
}
// Returns true if Op is setcc or zext of setcc.
static bool isSetCCOrZExtSetCC(const SDValue& Op, SetCCInfoAndKind &Info) {
if (isSetCC(Op, Info))
return true;
return ((Op.getOpcode() == ISD::ZERO_EXTEND) &&
isSetCC(Op->getOperand(0), Info));
}
// The folding we want to perform is:
// (add x, [zext] (setcc cc ...) )
// -->
// (csel x, (add x, 1), !cc ...)
//
// The latter will get matched to a CSINC instruction.
static SDValue performSetccAddFolding(SDNode *Op, SelectionDAG &DAG) {
assert(Op && Op->getOpcode() == ISD::ADD && "Unexpected operation!");
SDValue LHS = Op->getOperand(0);
SDValue RHS = Op->getOperand(1);
SetCCInfoAndKind InfoAndKind;
// If both operands are a SET_CC, then we don't want to perform this
// folding and create another csel as this results in more instructions
// (and higher register usage).
if (isSetCCOrZExtSetCC(LHS, InfoAndKind) &&
isSetCCOrZExtSetCC(RHS, InfoAndKind))
return SDValue();
// If neither operand is a SET_CC, give up.
if (!isSetCCOrZExtSetCC(LHS, InfoAndKind)) {
std::swap(LHS, RHS);
if (!isSetCCOrZExtSetCC(LHS, InfoAndKind))
return SDValue();
}
// FIXME: This could be generatized to work for FP comparisons.
EVT CmpVT = InfoAndKind.IsAArch64
? InfoAndKind.Info.AArch64.Cmp->getOperand(0).getValueType()
: InfoAndKind.Info.Generic.Opnd0->getValueType();
if (CmpVT != MVT::i32 && CmpVT != MVT::i64)
return SDValue();
SDValue CCVal;
SDValue Cmp;
SDLoc DL(Op);
if (InfoAndKind.IsAArch64) {
CCVal = DAG.getConstant(
AArch64CC::getInvertedCondCode(InfoAndKind.Info.AArch64.CC), DL,
MVT::i32);
Cmp = *InfoAndKind.Info.AArch64.Cmp;
} else
Cmp = getAArch64Cmp(
*InfoAndKind.Info.Generic.Opnd0, *InfoAndKind.Info.Generic.Opnd1,
ISD::getSetCCInverse(InfoAndKind.Info.Generic.CC, CmpVT), CCVal, DAG,
DL);
EVT VT = Op->getValueType(0);
LHS = DAG.getNode(ISD::ADD, DL, VT, RHS, DAG.getConstant(1, DL, VT));
return DAG.getNode(AArch64ISD::CSEL, DL, VT, RHS, LHS, CCVal, Cmp);
}
// ADD(UADDV a, UADDV b) --> UADDV(ADD a, b)
static SDValue performAddUADDVCombine(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
// Only scalar integer and vector types.
if (N->getOpcode() != ISD::ADD || !VT.isScalarInteger())
return SDValue();
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
if (LHS.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
RHS.getOpcode() != ISD::EXTRACT_VECTOR_ELT || LHS.getValueType() != VT)
return SDValue();
auto *LHSN1 = dyn_cast<ConstantSDNode>(LHS->getOperand(1));
auto *RHSN1 = dyn_cast<ConstantSDNode>(RHS->getOperand(1));
if (!LHSN1 || LHSN1 != RHSN1 || !RHSN1->isZero())
return SDValue();
SDValue Op1 = LHS->getOperand(0);
SDValue Op2 = RHS->getOperand(0);
EVT OpVT1 = Op1.getValueType();
EVT OpVT2 = Op2.getValueType();
if (Op1.getOpcode() != AArch64ISD::UADDV || OpVT1 != OpVT2 ||
Op2.getOpcode() != AArch64ISD::UADDV ||
OpVT1.getVectorElementType() != VT)
return SDValue();
SDValue Val1 = Op1.getOperand(0);
SDValue Val2 = Op2.getOperand(0);
EVT ValVT = Val1->getValueType(0);
SDLoc DL(N);
SDValue AddVal = DAG.getNode(ISD::ADD, DL, ValVT, Val1, Val2);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT,
DAG.getNode(AArch64ISD::UADDV, DL, ValVT, AddVal),
DAG.getConstant(0, DL, MVT::i64));
}
/// Perform the scalar expression combine in the form of:
/// CSEL(c, 1, cc) + b => CSINC(b+c, b, cc)
/// CSNEG(c, -1, cc) + b => CSINC(b+c, b, cc)
static SDValue performAddCSelIntoCSinc(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (!VT.isScalarInteger() || N->getOpcode() != ISD::ADD)
return SDValue();
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
// Handle commutivity.
if (LHS.getOpcode() != AArch64ISD::CSEL &&
LHS.getOpcode() != AArch64ISD::CSNEG) {
std::swap(LHS, RHS);
if (LHS.getOpcode() != AArch64ISD::CSEL &&
LHS.getOpcode() != AArch64ISD::CSNEG) {
return SDValue();
}
}
if (!LHS.hasOneUse())
return SDValue();
AArch64CC::CondCode AArch64CC =
static_cast<AArch64CC::CondCode>(LHS.getConstantOperandVal(2));
// The CSEL should include a const one operand, and the CSNEG should include
// One or NegOne operand.
ConstantSDNode *CTVal = dyn_cast<ConstantSDNode>(LHS.getOperand(0));
ConstantSDNode *CFVal = dyn_cast<ConstantSDNode>(LHS.getOperand(1));
if (!CTVal || !CFVal)
return SDValue();
if (!(LHS.getOpcode() == AArch64ISD::CSEL &&
(CTVal->isOne() || CFVal->isOne())) &&
!(LHS.getOpcode() == AArch64ISD::CSNEG &&
(CTVal->isOne() || CFVal->isAllOnes())))
return SDValue();
// Switch CSEL(1, c, cc) to CSEL(c, 1, !cc)
if (LHS.getOpcode() == AArch64ISD::CSEL && CTVal->isOne() &&
!CFVal->isOne()) {
std::swap(CTVal, CFVal);
AArch64CC = AArch64CC::getInvertedCondCode(AArch64CC);
}
SDLoc DL(N);
// Switch CSNEG(1, c, cc) to CSNEG(-c, -1, !cc)
if (LHS.getOpcode() == AArch64ISD::CSNEG && CTVal->isOne() &&
!CFVal->isAllOnes()) {
APInt C = -1 * CFVal->getAPIntValue();
CTVal = cast<ConstantSDNode>(DAG.getConstant(C, DL, VT));
CFVal = cast<ConstantSDNode>(DAG.getAllOnesConstant(DL, VT));
AArch64CC = AArch64CC::getInvertedCondCode(AArch64CC);
}
// It might be neutral for larger constants, as the immediate need to be
// materialized in a register.
APInt ADDC = CTVal->getAPIntValue();
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
if (!TLI.isLegalAddImmediate(ADDC.getSExtValue()))
return SDValue();
assert(((LHS.getOpcode() == AArch64ISD::CSEL && CFVal->isOne()) ||
(LHS.getOpcode() == AArch64ISD::CSNEG && CFVal->isAllOnes())) &&
"Unexpected constant value");
SDValue NewNode = DAG.getNode(ISD::ADD, DL, VT, RHS, SDValue(CTVal, 0));
SDValue CCVal = DAG.getConstant(AArch64CC, DL, MVT::i32);
SDValue Cmp = LHS.getOperand(3);
return DAG.getNode(AArch64ISD::CSINC, DL, VT, NewNode, RHS, CCVal, Cmp);
}
// ADD(UDOT(zero, x, y), A) --> UDOT(A, x, y)
static SDValue performAddDotCombine(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (N->getOpcode() != ISD::ADD)
return SDValue();
SDValue Dot = N->getOperand(0);
SDValue A = N->getOperand(1);
// Handle commutivity
auto isZeroDot = [](SDValue Dot) {
return (Dot.getOpcode() == AArch64ISD::UDOT ||
Dot.getOpcode() == AArch64ISD::SDOT) &&
isZerosVector(Dot.getOperand(0).getNode());
};
if (!isZeroDot(Dot))
std::swap(Dot, A);
if (!isZeroDot(Dot))
return SDValue();
return DAG.getNode(Dot.getOpcode(), SDLoc(N), VT, A, Dot.getOperand(1),
Dot.getOperand(2));
}
static bool isNegatedInteger(SDValue Op) {
return Op.getOpcode() == ISD::SUB && isNullConstant(Op.getOperand(0));
}
static SDValue getNegatedInteger(SDValue Op, SelectionDAG &DAG) {
SDLoc DL(Op);
EVT VT = Op.getValueType();
SDValue Zero = DAG.getConstant(0, DL, VT);
return DAG.getNode(ISD::SUB, DL, VT, Zero, Op);
}
// Try to fold
//
// (neg (csel X, Y)) -> (csel (neg X), (neg Y))
//
// The folding helps csel to be matched with csneg without generating
// redundant neg instruction, which includes negation of the csel expansion
// of abs node lowered by lowerABS.
static SDValue performNegCSelCombine(SDNode *N, SelectionDAG &DAG) {
if (!isNegatedInteger(SDValue(N, 0)))
return SDValue();
SDValue CSel = N->getOperand(1);
if (CSel.getOpcode() != AArch64ISD::CSEL || !CSel->hasOneUse())
return SDValue();
SDValue N0 = CSel.getOperand(0);
SDValue N1 = CSel.getOperand(1);
// If both of them is not negations, it's not worth the folding as it
// introduces two additional negations while reducing one negation.
if (!isNegatedInteger(N0) && !isNegatedInteger(N1))
return SDValue();
SDValue N0N = getNegatedInteger(N0, DAG);
SDValue N1N = getNegatedInteger(N1, DAG);
SDLoc DL(N);
EVT VT = CSel.getValueType();
return DAG.getNode(AArch64ISD::CSEL, DL, VT, N0N, N1N, CSel.getOperand(2),
CSel.getOperand(3));
}
// The basic add/sub long vector instructions have variants with "2" on the end
// which act on the high-half of their inputs. They are normally matched by
// patterns like:
//
// (add (zeroext (extract_high LHS)),
// (zeroext (extract_high RHS)))
// -> uaddl2 vD, vN, vM
//
// However, if one of the extracts is something like a duplicate, this
// instruction can still be used profitably. This function puts the DAG into a
// more appropriate form for those patterns to trigger.
static SDValue performAddSubLongCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
SelectionDAG &DAG = DCI.DAG;
if (DCI.isBeforeLegalizeOps())
return SDValue();
MVT VT = N->getSimpleValueType(0);
if (!VT.is128BitVector()) {
if (N->getOpcode() == ISD::ADD)
return performSetccAddFolding(N, DAG);
return SDValue();
}
// Make sure both branches are extended in the same way.
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
if ((LHS.getOpcode() != ISD::ZERO_EXTEND &&
LHS.getOpcode() != ISD::SIGN_EXTEND) ||
LHS.getOpcode() != RHS.getOpcode())
return SDValue();
unsigned ExtType = LHS.getOpcode();
// It's not worth doing if at least one of the inputs isn't already an
// extract, but we don't know which it'll be so we have to try both.
if (isEssentiallyExtractHighSubvector(LHS.getOperand(0))) {
RHS = tryExtendDUPToExtractHigh(RHS.getOperand(0), DAG);
if (!RHS.getNode())
return SDValue();
RHS = DAG.getNode(ExtType, SDLoc(N), VT, RHS);
} else if (isEssentiallyExtractHighSubvector(RHS.getOperand(0))) {
LHS = tryExtendDUPToExtractHigh(LHS.getOperand(0), DAG);
if (!LHS.getNode())
return SDValue();
LHS = DAG.getNode(ExtType, SDLoc(N), VT, LHS);
}
return DAG.getNode(N->getOpcode(), SDLoc(N), VT, LHS, RHS);
}
static bool isCMP(SDValue Op) {
return Op.getOpcode() == AArch64ISD::SUBS &&
!Op.getNode()->hasAnyUseOfValue(0);
}
// (CSEL 1 0 CC Cond) => CC
// (CSEL 0 1 CC Cond) => !CC
static std::optional<AArch64CC::CondCode> getCSETCondCode(SDValue Op) {
if (Op.getOpcode() != AArch64ISD::CSEL)
return std::nullopt;
auto CC = static_cast<AArch64CC::CondCode>(Op.getConstantOperandVal(2));
if (CC == AArch64CC::AL || CC == AArch64CC::NV)
return std::nullopt;
SDValue OpLHS = Op.getOperand(0);
SDValue OpRHS = Op.getOperand(1);
if (isOneConstant(OpLHS) && isNullConstant(OpRHS))
return CC;
if (isNullConstant(OpLHS) && isOneConstant(OpRHS))
return getInvertedCondCode(CC);
return std::nullopt;
}
// (ADC{S} l r (CMP (CSET HS carry) 1)) => (ADC{S} l r carry)
// (SBC{S} l r (CMP 0 (CSET LO carry))) => (SBC{S} l r carry)
static SDValue foldOverflowCheck(SDNode *Op, SelectionDAG &DAG, bool IsAdd) {
SDValue CmpOp = Op->getOperand(2);
if (!isCMP(CmpOp))
return SDValue();
if (IsAdd) {
if (!isOneConstant(CmpOp.getOperand(1)))
return SDValue();
} else {
if (!isNullConstant(CmpOp.getOperand(0)))
return SDValue();
}
SDValue CsetOp = CmpOp->getOperand(IsAdd ? 0 : 1);
auto CC = getCSETCondCode(CsetOp);
if (CC != (IsAdd ? AArch64CC::HS : AArch64CC::LO))
return SDValue();
return DAG.getNode(Op->getOpcode(), SDLoc(Op), Op->getVTList(),
Op->getOperand(0), Op->getOperand(1),
CsetOp.getOperand(3));
}
// (ADC x 0 cond) => (CINC x HS cond)
static SDValue foldADCToCINC(SDNode *N, SelectionDAG &DAG) {
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
SDValue Cond = N->getOperand(2);
if (!isNullConstant(RHS))
return SDValue();
EVT VT = N->getValueType(0);
SDLoc DL(N);
// (CINC x cc cond) <=> (CSINC x x !cc cond)
SDValue CC = DAG.getConstant(AArch64CC::LO, DL, MVT::i32);
return DAG.getNode(AArch64ISD::CSINC, DL, VT, LHS, LHS, CC, Cond);
}
static SDValue performBuildVectorCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
SDLoc DL(N);
EVT VT = N->getValueType(0);
if (DAG.getSubtarget<AArch64Subtarget>().isNeonAvailable() &&
(VT == MVT::v4f16 || VT == MVT::v4bf16)) {
SDValue Elt0 = N->getOperand(0), Elt1 = N->getOperand(1),
Elt2 = N->getOperand(2), Elt3 = N->getOperand(3);
if (Elt0->getOpcode() == ISD::FP_ROUND &&
Elt1->getOpcode() == ISD::FP_ROUND &&
isa<ConstantSDNode>(Elt0->getOperand(1)) &&
isa<ConstantSDNode>(Elt1->getOperand(1)) &&
Elt0->getConstantOperandVal(1) == Elt1->getConstantOperandVal(1) &&
Elt0->getOperand(0)->getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
Elt1->getOperand(0)->getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
// Constant index.
isa<ConstantSDNode>(Elt0->getOperand(0)->getOperand(1)) &&
isa<ConstantSDNode>(Elt1->getOperand(0)->getOperand(1)) &&
Elt0->getOperand(0)->getOperand(0) ==
Elt1->getOperand(0)->getOperand(0) &&
Elt0->getOperand(0)->getConstantOperandVal(1) == 0 &&
Elt1->getOperand(0)->getConstantOperandVal(1) == 1) {
SDValue LowLanesSrcVec = Elt0->getOperand(0)->getOperand(0);
if (LowLanesSrcVec.getValueType() == MVT::v2f64) {
SDValue HighLanes;
if (Elt2->getOpcode() == ISD::UNDEF &&
Elt3->getOpcode() == ISD::UNDEF) {
HighLanes = DAG.getUNDEF(MVT::v2f32);
} else if (Elt2->getOpcode() == ISD::FP_ROUND &&
Elt3->getOpcode() == ISD::FP_ROUND &&
isa<ConstantSDNode>(Elt2->getOperand(1)) &&
isa<ConstantSDNode>(Elt3->getOperand(1)) &&
Elt2->getConstantOperandVal(1) ==
Elt3->getConstantOperandVal(1) &&
Elt2->getOperand(0)->getOpcode() ==
ISD::EXTRACT_VECTOR_ELT &&
Elt3->getOperand(0)->getOpcode() ==
ISD::EXTRACT_VECTOR_ELT &&
// Constant index.
isa<ConstantSDNode>(Elt2->getOperand(0)->getOperand(1)) &&
isa<ConstantSDNode>(Elt3->getOperand(0)->getOperand(1)) &&
Elt2->getOperand(0)->getOperand(0) ==
Elt3->getOperand(0)->getOperand(0) &&
Elt2->getOperand(0)->getConstantOperandVal(1) == 0 &&
Elt3->getOperand(0)->getConstantOperandVal(1) == 1) {
SDValue HighLanesSrcVec = Elt2->getOperand(0)->getOperand(0);
HighLanes =
DAG.getNode(AArch64ISD::FCVTXN, DL, MVT::v2f32, HighLanesSrcVec);
}
if (HighLanes) {
SDValue DoubleToSingleSticky =
DAG.getNode(AArch64ISD::FCVTXN, DL, MVT::v2f32, LowLanesSrcVec);
SDValue Concat = DAG.getNode(ISD::CONCAT_VECTORS, DL, MVT::v4f32,
DoubleToSingleSticky, HighLanes);
return DAG.getNode(ISD::FP_ROUND, DL, VT, Concat,
Elt0->getOperand(1));
}
}
}
}
if (VT == MVT::v2f64) {
SDValue Elt0 = N->getOperand(0), Elt1 = N->getOperand(1);
if (Elt0->getOpcode() == ISD::FP_EXTEND &&
Elt1->getOpcode() == ISD::FP_EXTEND &&
Elt0->getOperand(0)->getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
Elt1->getOperand(0)->getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
Elt0->getOperand(0)->getOperand(0) ==
Elt1->getOperand(0)->getOperand(0) &&
// Constant index.
isa<ConstantSDNode>(Elt0->getOperand(0)->getOperand(1)) &&
isa<ConstantSDNode>(Elt1->getOperand(0)->getOperand(1)) &&
Elt0->getOperand(0)->getConstantOperandVal(1) + 1 ==
Elt1->getOperand(0)->getConstantOperandVal(1) &&
// EXTRACT_SUBVECTOR requires that Idx be a constant multiple of
// ResultType's known minimum vector length.
Elt0->getOperand(0)->getConstantOperandVal(1) %
VT.getVectorMinNumElements() ==
0) {
SDValue SrcVec = Elt0->getOperand(0)->getOperand(0);
if (SrcVec.getValueType() == MVT::v4f16 ||
SrcVec.getValueType() == MVT::v4bf16) {
SDValue HalfToSingle =
DAG.getNode(ISD::FP_EXTEND, DL, MVT::v4f32, SrcVec);
SDValue SubvectorIdx = Elt0->getOperand(0)->getOperand(1);
SDValue Extract = DAG.getNode(
ISD::EXTRACT_SUBVECTOR, DL, VT.changeVectorElementType(MVT::f32),
HalfToSingle, SubvectorIdx);
return DAG.getNode(ISD::FP_EXTEND, DL, VT, Extract);
}
}
}
// A build vector of two extracted elements is equivalent to an
// extract subvector where the inner vector is any-extended to the
// extract_vector_elt VT.
// (build_vector (extract_elt_iXX_to_i32 vec Idx+0)
// (extract_elt_iXX_to_i32 vec Idx+1))
// => (extract_subvector (anyext_iXX_to_i32 vec) Idx)
// For now, only consider the v2i32 case, which arises as a result of
// legalization.
if (VT != MVT::v2i32)
return SDValue();
SDValue Elt0 = N->getOperand(0), Elt1 = N->getOperand(1);
// Reminder, EXTRACT_VECTOR_ELT has the effect of any-extending to its VT.
if (Elt0->getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
Elt1->getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
// Constant index.
isa<ConstantSDNode>(Elt0->getOperand(1)) &&
isa<ConstantSDNode>(Elt1->getOperand(1)) &&
// Both EXTRACT_VECTOR_ELT from same vector...
Elt0->getOperand(0) == Elt1->getOperand(0) &&
// ... and contiguous. First element's index +1 == second element's index.
Elt0->getConstantOperandVal(1) + 1 == Elt1->getConstantOperandVal(1) &&
// EXTRACT_SUBVECTOR requires that Idx be a constant multiple of
// ResultType's known minimum vector length.
Elt0->getConstantOperandVal(1) % VT.getVectorMinNumElements() == 0) {
SDValue VecToExtend = Elt0->getOperand(0);
EVT ExtVT = VecToExtend.getValueType().changeVectorElementType(MVT::i32);
if (!DAG.getTargetLoweringInfo().isTypeLegal(ExtVT))
return SDValue();
SDValue SubvectorIdx = DAG.getVectorIdxConstant(Elt0->getConstantOperandVal(1), DL);
SDValue Ext = DAG.getNode(ISD::ANY_EXTEND, DL, ExtVT, VecToExtend);
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v2i32, Ext,
SubvectorIdx);
}
return SDValue();
}
static SDValue performTruncateCombine(SDNode *N, SelectionDAG &DAG,
TargetLowering::DAGCombinerInfo &DCI) {
SDLoc DL(N);
EVT VT = N->getValueType(0);
SDValue N0 = N->getOperand(0);
if (VT.isFixedLengthVector() && VT.is64BitVector() && N0.hasOneUse() &&
N0.getOpcode() == AArch64ISD::DUP) {
SDValue Op = N0.getOperand(0);
if (VT.getScalarType() == MVT::i32 &&
N0.getOperand(0).getValueType().getScalarType() == MVT::i64)
Op = DAG.getNode(ISD::TRUNCATE, DL, MVT::i32, Op);
return DAG.getNode(N0.getOpcode(), DL, VT, Op);
}
// Performing the following combine produces a preferable form for ISEL.
// i32 (trunc (extract Vi64, idx)) -> i32 (extract (nvcast Vi32), idx*2))
if (DCI.isAfterLegalizeDAG() && N0.getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
N0.hasOneUse()) {
SDValue Op = N0.getOperand(0);
SDValue ExtractIndexNode = N0.getOperand(1);
if (!isa<ConstantSDNode>(ExtractIndexNode))
return SDValue();
// For a legal DAG, EXTRACT_VECTOR_ELT can only have produced an i32 or i64.
// So we can only expect: i32 (trunc (i64 (extract Vi64, idx))).
assert((VT == MVT::i32 && N0.getValueType() == MVT::i64) &&
"Unexpected legalisation result!");
EVT SrcVectorType = Op.getValueType();
// We also assume that SrcVectorType cannot be a V64 (see
// LowerEXTRACT_VECTOR_ELT).
assert((SrcVectorType == MVT::v2i64 || SrcVectorType == MVT::nxv2i64) &&
"Unexpected legalisation result!");
unsigned ExtractIndex =
cast<ConstantSDNode>(ExtractIndexNode)->getZExtValue();
MVT CastVT = SrcVectorType.isScalableVector() ? MVT::nxv4i32 : MVT::v4i32;
Op = DAG.getNode(AArch64ISD::NVCAST, DL, CastVT, Op);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, Op,
DAG.getVectorIdxConstant(ExtractIndex * 2, DL));
}
return SDValue();
}
// Check an node is an extend or shift operand
static bool isExtendOrShiftOperand(SDValue N) {
unsigned Opcode = N.getOpcode();
if (ISD::isExtOpcode(Opcode) || Opcode == ISD::SIGN_EXTEND_INREG) {
EVT SrcVT;
if (Opcode == ISD::SIGN_EXTEND_INREG)
SrcVT = cast<VTSDNode>(N.getOperand(1))->getVT();
else
SrcVT = N.getOperand(0).getValueType();
return SrcVT == MVT::i32 || SrcVT == MVT::i16 || SrcVT == MVT::i8;
} else if (Opcode == ISD::AND) {
ConstantSDNode *CSD = dyn_cast<ConstantSDNode>(N.getOperand(1));
if (!CSD)
return false;
uint64_t AndMask = CSD->getZExtValue();
return AndMask == 0xff || AndMask == 0xffff || AndMask == 0xffffffff;
} else if (Opcode == ISD::SHL || Opcode == ISD::SRL || Opcode == ISD::SRA) {
return isa<ConstantSDNode>(N.getOperand(1));
}
return false;
}
// (N - Y) + Z --> (Z - Y) + N
// when N is an extend or shift operand
static SDValue performAddCombineSubShift(SDNode *N, SDValue SUB, SDValue Z,
SelectionDAG &DAG) {
auto IsOneUseExtend = [](SDValue N) {
return N.hasOneUse() && isExtendOrShiftOperand(N);
};
// DAGCombiner will revert the combination when Z is constant cause
// dead loop. So don't enable the combination when Z is constant.
// If Z is one use shift C, we also can't do the optimization.
// It will falling to self infinite loop.
if (isa<ConstantSDNode>(Z) || IsOneUseExtend(Z))
return SDValue();
if (SUB.getOpcode() != ISD::SUB || !SUB.hasOneUse())
return SDValue();
SDValue Shift = SUB.getOperand(0);
if (!IsOneUseExtend(Shift))
return SDValue();
SDLoc DL(N);
EVT VT = N->getValueType(0);
SDValue Y = SUB.getOperand(1);
SDValue NewSub = DAG.getNode(ISD::SUB, DL, VT, Z, Y);
return DAG.getNode(ISD::ADD, DL, VT, NewSub, Shift);
}
static SDValue performAddCombineForShiftedOperands(SDNode *N,
SelectionDAG &DAG) {
// NOTE: Swapping LHS and RHS is not done for SUB, since SUB is not
// commutative.
if (N->getOpcode() != ISD::ADD)
return SDValue();
// Bail out when value type is not one of {i32, i64}, since AArch64 ADD with
// shifted register is only available for i32 and i64.
EVT VT = N->getValueType(0);
if (VT != MVT::i32 && VT != MVT::i64)
return SDValue();
SDLoc DL(N);
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
if (SDValue Val = performAddCombineSubShift(N, LHS, RHS, DAG))
return Val;
if (SDValue Val = performAddCombineSubShift(N, RHS, LHS, DAG))
return Val;
uint64_t LHSImm = 0, RHSImm = 0;
// If both operand are shifted by imm and shift amount is not greater than 4
// for one operand, swap LHS and RHS to put operand with smaller shift amount
// on RHS.
//
// On many AArch64 processors (Cortex A78, Neoverse N1/N2/V1, etc), ADD with
// LSL shift (shift <= 4) has smaller latency and larger throughput than ADD
// with LSL (shift > 4). For the rest of processors, this is no-op for
// performance or correctness.
if (isOpcWithIntImmediate(LHS.getNode(), ISD::SHL, LHSImm) &&
isOpcWithIntImmediate(RHS.getNode(), ISD::SHL, RHSImm) && LHSImm <= 4 &&
RHSImm > 4 && LHS.hasOneUse())
return DAG.getNode(ISD::ADD, DL, VT, RHS, LHS);
return SDValue();
}
// The mid end will reassociate sub(sub(x, m1), m2) to sub(x, add(m1, m2))
// This reassociates it back to allow the creation of more mls instructions.
static SDValue performSubAddMULCombine(SDNode *N, SelectionDAG &DAG) {
if (N->getOpcode() != ISD::SUB)
return SDValue();
SDValue Add = N->getOperand(1);
SDValue X = N->getOperand(0);
if (Add.getOpcode() != ISD::ADD)
return SDValue();
if (!Add.hasOneUse())
return SDValue();
if (DAG.isConstantIntBuildVectorOrConstantInt(X))
return SDValue();
SDValue M1 = Add.getOperand(0);
SDValue M2 = Add.getOperand(1);
if (M1.getOpcode() != ISD::MUL && M1.getOpcode() != AArch64ISD::SMULL &&
M1.getOpcode() != AArch64ISD::UMULL)
return SDValue();
if (M2.getOpcode() != ISD::MUL && M2.getOpcode() != AArch64ISD::SMULL &&
M2.getOpcode() != AArch64ISD::UMULL)
return SDValue();
EVT VT = N->getValueType(0);
SDValue Sub = DAG.getNode(ISD::SUB, SDLoc(N), VT, X, M1);
return DAG.getNode(ISD::SUB, SDLoc(N), VT, Sub, M2);
}
// Combine into mla/mls.
// This works on the patterns of:
// add v1, (mul v2, v3)
// sub v1, (mul v2, v3)
// for vectors of type <1 x i64> and <2 x i64> when SVE is available.
// It will transform the add/sub to a scalable version, so that we can
// make use of SVE's MLA/MLS that will be generated for that pattern
static SDValue
performSVEMulAddSubCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI) {
SelectionDAG &DAG = DCI.DAG;
// Make sure that the types are legal
if (!DCI.isAfterLegalizeDAG())
return SDValue();
// Before using SVE's features, check first if it's available.
if (!DAG.getSubtarget<AArch64Subtarget>().hasSVE())
return SDValue();
if (N->getOpcode() != ISD::ADD && N->getOpcode() != ISD::SUB)
return SDValue();
if (!N->getValueType(0).isFixedLengthVector())
return SDValue();
auto performOpt = [&DAG, &N](SDValue Op0, SDValue Op1) -> SDValue {
if (Op1.getOpcode() != ISD::EXTRACT_SUBVECTOR)
return SDValue();
if (!cast<ConstantSDNode>(Op1->getOperand(1))->isZero())
return SDValue();
SDValue MulValue = Op1->getOperand(0);
if (MulValue.getOpcode() != AArch64ISD::MUL_PRED)
return SDValue();
if (!Op1.hasOneUse() || !MulValue.hasOneUse())
return SDValue();
EVT ScalableVT = MulValue.getValueType();
if (!ScalableVT.isScalableVector())
return SDValue();
SDValue ScaledOp = convertToScalableVector(DAG, ScalableVT, Op0);
SDValue NewValue =
DAG.getNode(N->getOpcode(), SDLoc(N), ScalableVT, {ScaledOp, MulValue});
return convertFromScalableVector(DAG, N->getValueType(0), NewValue);
};
if (SDValue res = performOpt(N->getOperand(0), N->getOperand(1)))
return res;
else if (N->getOpcode() == ISD::ADD)
return performOpt(N->getOperand(1), N->getOperand(0));
return SDValue();
}
// Given a i64 add from a v1i64 extract, convert to a neon v1i64 add. This can
// help, for example, to produce ssra from sshr+add.
static SDValue performAddSubIntoVectorOp(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (VT != MVT::i64 ||
DAG.getTargetLoweringInfo().isOperationExpand(N->getOpcode(), MVT::v1i64))
return SDValue();
SDValue Op0 = N->getOperand(0);
SDValue Op1 = N->getOperand(1);
// At least one of the operands should be an extract, and the other should be
// something that is easy to convert to v1i64 type (in this case a load).
if (Op0.getOpcode() != ISD::EXTRACT_VECTOR_ELT &&
Op0.getOpcode() != ISD::LOAD)
return SDValue();
if (Op1.getOpcode() != ISD::EXTRACT_VECTOR_ELT &&
Op1.getOpcode() != ISD::LOAD)
return SDValue();
SDLoc DL(N);
if (Op0.getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
Op0.getOperand(0).getValueType() == MVT::v1i64) {
Op0 = Op0.getOperand(0);
Op1 = DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, MVT::v1i64, Op1);
} else if (Op1.getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
Op1.getOperand(0).getValueType() == MVT::v1i64) {
Op0 = DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, MVT::v1i64, Op0);
Op1 = Op1.getOperand(0);
} else
return SDValue();
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i64,
DAG.getNode(N->getOpcode(), DL, MVT::v1i64, Op0, Op1),
DAG.getConstant(0, DL, MVT::i64));
}
static bool isLoadOrMultipleLoads(SDValue B, SmallVector<LoadSDNode *> &Loads) {
SDValue BV = peekThroughOneUseBitcasts(B);
if (!BV->hasOneUse())
return false;
if (auto *Ld = dyn_cast<LoadSDNode>(BV)) {
if (!Ld || !Ld->isSimple())
return false;
Loads.push_back(Ld);
return true;
} else if (BV.getOpcode() == ISD::BUILD_VECTOR ||
BV.getOpcode() == ISD::CONCAT_VECTORS) {
for (unsigned Op = 0; Op < BV.getNumOperands(); Op++) {
auto *Ld = dyn_cast<LoadSDNode>(BV.getOperand(Op));
if (!Ld || !Ld->isSimple() || !BV.getOperand(Op).hasOneUse())
return false;
Loads.push_back(Ld);
}
return true;
} else if (B.getOpcode() == ISD::VECTOR_SHUFFLE) {
// Try to find a tree of shuffles and concats from how IR shuffles of loads
// are lowered. Note that this only comes up because we do not always visit
// operands before uses. After that is fixed this can be removed and in the
// meantime this is fairly specific to the lowering we expect from IR.
// t46: v16i8 = vector_shuffle<0,1,2,3,4,5,6,7,8,9,10,11,16,17,18,19> t44, t45
// t44: v16i8 = vector_shuffle<0,1,2,3,4,5,6,7,16,17,18,19,u,u,u,u> t42, t43
// t42: v16i8 = concat_vectors t40, t36, undef:v4i8, undef:v4i8
// t40: v4i8,ch = load<(load (s32) from %ir.17)> t0, t22, undef:i64
// t36: v4i8,ch = load<(load (s32) from %ir.13)> t0, t18, undef:i64
// t43: v16i8 = concat_vectors t32, undef:v4i8, undef:v4i8, undef:v4i8
// t32: v4i8,ch = load<(load (s32) from %ir.9)> t0, t14, undef:i64
// t45: v16i8 = concat_vectors t28, undef:v4i8, undef:v4i8, undef:v4i8
// t28: v4i8,ch = load<(load (s32) from %ir.0)> t0, t2, undef:i64
if (B.getOperand(0).getOpcode() != ISD::VECTOR_SHUFFLE ||
B.getOperand(0).getOperand(0).getOpcode() != ISD::CONCAT_VECTORS ||
B.getOperand(0).getOperand(1).getOpcode() != ISD::CONCAT_VECTORS ||
B.getOperand(1).getOpcode() != ISD::CONCAT_VECTORS ||
B.getOperand(1).getNumOperands() != 4)
return false;
auto SV1 = cast<ShuffleVectorSDNode>(B);
auto SV2 = cast<ShuffleVectorSDNode>(B.getOperand(0));
int NumElts = B.getValueType().getVectorNumElements();
int NumSubElts = NumElts / 4;
for (int I = 0; I < NumSubElts; I++) {
// <0,1,2,3,4,5,6,7,8,9,10,11,16,17,18,19>
if (SV1->getMaskElt(I) != I ||
SV1->getMaskElt(I + NumSubElts) != I + NumSubElts ||
SV1->getMaskElt(I + NumSubElts * 2) != I + NumSubElts * 2 ||
SV1->getMaskElt(I + NumSubElts * 3) != I + NumElts)
return false;
// <0,1,2,3,4,5,6,7,16,17,18,19,u,u,u,u>
if (SV2->getMaskElt(I) != I ||
SV2->getMaskElt(I + NumSubElts) != I + NumSubElts ||
SV2->getMaskElt(I + NumSubElts * 2) != I + NumElts)
return false;
}
auto *Ld0 = dyn_cast<LoadSDNode>(SV2->getOperand(0).getOperand(0));
auto *Ld1 = dyn_cast<LoadSDNode>(SV2->getOperand(0).getOperand(1));
auto *Ld2 = dyn_cast<LoadSDNode>(SV2->getOperand(1).getOperand(0));
auto *Ld3 = dyn_cast<LoadSDNode>(B.getOperand(1).getOperand(0));
if (!Ld0 || !Ld1 || !Ld2 || !Ld3 || !Ld0->isSimple() || !Ld1->isSimple() ||
!Ld2->isSimple() || !Ld3->isSimple())
return false;
Loads.push_back(Ld0);
Loads.push_back(Ld1);
Loads.push_back(Ld2);
Loads.push_back(Ld3);
return true;
}
return false;
}
static bool areLoadedOffsetButOtherwiseSame(SDValue Op0, SDValue Op1,
SelectionDAG &DAG,
unsigned &NumSubLoads) {
if (!Op0.hasOneUse() || !Op1.hasOneUse())
return false;
SmallVector<LoadSDNode *> Loads0, Loads1;
if (isLoadOrMultipleLoads(Op0, Loads0) &&
isLoadOrMultipleLoads(Op1, Loads1)) {
if (NumSubLoads && Loads0.size() != NumSubLoads)
return false;
NumSubLoads = Loads0.size();
return Loads0.size() == Loads1.size() &&
all_of(zip(Loads0, Loads1), [&DAG](auto L) {
unsigned Size = get<0>(L)->getValueType(0).getSizeInBits();
return Size == get<1>(L)->getValueType(0).getSizeInBits() &&
DAG.areNonVolatileConsecutiveLoads(get<1>(L), get<0>(L),
Size / 8, 1);
});
}
if (Op0.getOpcode() != Op1.getOpcode())
return false;
switch (Op0.getOpcode()) {
case ISD::ADD:
case ISD::SUB:
return areLoadedOffsetButOtherwiseSame(Op0.getOperand(0), Op1.getOperand(0),
DAG, NumSubLoads) &&
areLoadedOffsetButOtherwiseSame(Op0.getOperand(1), Op1.getOperand(1),
DAG, NumSubLoads);
case ISD::SIGN_EXTEND:
case ISD::ANY_EXTEND:
case ISD::ZERO_EXTEND:
EVT XVT = Op0.getOperand(0).getValueType();
if (XVT.getScalarSizeInBits() != 8 && XVT.getScalarSizeInBits() != 16 &&
XVT.getScalarSizeInBits() != 32)
return false;
return areLoadedOffsetButOtherwiseSame(Op0.getOperand(0), Op1.getOperand(0),
DAG, NumSubLoads);
}
return false;
}
// This method attempts to fold trees of add(ext(load p), shl(ext(load p+4))
// into a single load of twice the size, that we extract the bottom part and top
// part so that the shl can use a shll2 instruction. The two loads in that
// example can also be larger trees of instructions, which are identical except
// for the leaves which are all loads offset from the LHS, including
// buildvectors of multiple loads. For example the RHS tree could be
// sub(zext(buildvec(load p+4, load q+4)), zext(buildvec(load r+4, load s+4)))
// Whilst it can be common for the larger loads to replace LDP instructions
// (which doesn't gain anything on it's own), the larger loads can help create
// more efficient code, and in buildvectors prevent the need for ld1 lane
// inserts which can be slower than normal loads.
static SDValue performExtBinopLoadFold(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (!VT.isFixedLengthVector() ||
(VT.getScalarSizeInBits() != 16 && VT.getScalarSizeInBits() != 32 &&
VT.getScalarSizeInBits() != 64))
return SDValue();
SDValue Other = N->getOperand(0);
SDValue Shift = N->getOperand(1);
if (Shift.getOpcode() != ISD::SHL && N->getOpcode() != ISD::SUB)
std::swap(Shift, Other);
APInt ShiftAmt;
if (Shift.getOpcode() != ISD::SHL || !Shift.hasOneUse() ||
!ISD::isConstantSplatVector(Shift.getOperand(1).getNode(), ShiftAmt))
return SDValue();
if (!ISD::isExtOpcode(Shift.getOperand(0).getOpcode()) ||
!ISD::isExtOpcode(Other.getOpcode()) ||
Shift.getOperand(0).getOperand(0).getValueType() !=
Other.getOperand(0).getValueType() ||
!Other.hasOneUse() || !Shift.getOperand(0).hasOneUse())
return SDValue();
SDValue Op0 = Other.getOperand(0);
SDValue Op1 = Shift.getOperand(0).getOperand(0);
unsigned NumSubLoads = 0;
if (!areLoadedOffsetButOtherwiseSame(Op0, Op1, DAG, NumSubLoads))
return SDValue();
// Attempt to rule out some unprofitable cases using heuristics (some working
// around suboptimal code generation), notably if the extend not be able to
// use ushll2 instructions as the types are not large enough. Otherwise zip's
// will need to be created which can increase the instruction count.
unsigned NumElts = Op0.getValueType().getVectorNumElements();
unsigned NumSubElts = NumElts / NumSubLoads;
if (NumSubElts * VT.getScalarSizeInBits() < 128 ||
(Other.getOpcode() != Shift.getOperand(0).getOpcode() &&
Op0.getValueType().getSizeInBits() < 128 &&
!DAG.getTargetLoweringInfo().isTypeLegal(Op0.getValueType())))
return SDValue();
// Recreate the tree with the new combined loads.
std::function<SDValue(SDValue, SDValue, SelectionDAG &)> GenCombinedTree =
[&GenCombinedTree](SDValue Op0, SDValue Op1, SelectionDAG &DAG) {
EVT DVT =
Op0.getValueType().getDoubleNumVectorElementsVT(*DAG.getContext());
SmallVector<LoadSDNode *> Loads0, Loads1;
if (isLoadOrMultipleLoads(Op0, Loads0) &&
isLoadOrMultipleLoads(Op1, Loads1)) {
EVT LoadVT = EVT::getVectorVT(
*DAG.getContext(), Op0.getValueType().getScalarType(),
Op0.getValueType().getVectorNumElements() / Loads0.size());
EVT DLoadVT = LoadVT.getDoubleNumVectorElementsVT(*DAG.getContext());
SmallVector<SDValue> NewLoads;
for (const auto &[L0, L1] : zip(Loads0, Loads1)) {
SDValue Load = DAG.getLoad(DLoadVT, SDLoc(L0), L0->getChain(),
L0->getBasePtr(), L0->getPointerInfo(),
L0->getBaseAlign());
DAG.makeEquivalentMemoryOrdering(L0, Load.getValue(1));
DAG.makeEquivalentMemoryOrdering(L1, Load.getValue(1));
NewLoads.push_back(Load);
}
return DAG.getNode(ISD::CONCAT_VECTORS, SDLoc(Op0), DVT, NewLoads);
}
SmallVector<SDValue> Ops;
for (const auto &[O0, O1] : zip(Op0->op_values(), Op1->op_values()))
Ops.push_back(GenCombinedTree(O0, O1, DAG));
return DAG.getNode(Op0.getOpcode(), SDLoc(Op0), DVT, Ops);
};
SDValue NewOp = GenCombinedTree(Op0, Op1, DAG);
SmallVector<int> LowMask(NumElts, 0), HighMask(NumElts, 0);
int Hi = NumSubElts, Lo = 0;
for (unsigned i = 0; i < NumSubLoads; i++) {
for (unsigned j = 0; j < NumSubElts; j++) {
LowMask[i * NumSubElts + j] = Lo++;
HighMask[i * NumSubElts + j] = Hi++;
}
Lo += NumSubElts;
Hi += NumSubElts;
}
SDLoc DL(N);
SDValue Ext0, Ext1;
// Extract the top and bottom lanes, then extend the result. Possibly extend
// the result then extract the lanes if the two operands match as it produces
// slightly smaller code.
if (Other.getOpcode() != Shift.getOperand(0).getOpcode()) {
SDValue SubL = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, Op0.getValueType(),
NewOp, DAG.getConstant(0, DL, MVT::i64));
SDValue SubH =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, Op0.getValueType(), NewOp,
DAG.getConstant(NumSubElts * NumSubLoads, DL, MVT::i64));
SDValue Extr0 =
DAG.getVectorShuffle(Op0.getValueType(), DL, SubL, SubH, LowMask);
SDValue Extr1 =
DAG.getVectorShuffle(Op0.getValueType(), DL, SubL, SubH, HighMask);
Ext0 = DAG.getNode(Other.getOpcode(), DL, VT, Extr0);
Ext1 = DAG.getNode(Shift.getOperand(0).getOpcode(), DL, VT, Extr1);
} else {
EVT DVT = VT.getDoubleNumVectorElementsVT(*DAG.getContext());
SDValue Ext = DAG.getNode(Other.getOpcode(), DL, DVT, NewOp);
SDValue SubL = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, Ext,
DAG.getConstant(0, DL, MVT::i64));
SDValue SubH =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, Ext,
DAG.getConstant(NumSubElts * NumSubLoads, DL, MVT::i64));
Ext0 = DAG.getVectorShuffle(VT, DL, SubL, SubH, LowMask);
Ext1 = DAG.getVectorShuffle(VT, DL, SubL, SubH, HighMask);
}
SDValue NShift =
DAG.getNode(Shift.getOpcode(), DL, VT, Ext1, Shift.getOperand(1));
return DAG.getNode(N->getOpcode(), DL, VT, Ext0, NShift);
}
static SDValue performAddSubCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
// Try to change sum of two reductions.
if (SDValue Val = performAddUADDVCombine(N, DCI.DAG))
return Val;
if (SDValue Val = performAddDotCombine(N, DCI.DAG))
return Val;
if (SDValue Val = performAddCSelIntoCSinc(N, DCI.DAG))
return Val;
if (SDValue Val = performNegCSelCombine(N, DCI.DAG))
return Val;
if (SDValue Val = performVectorExtCombine(N, DCI.DAG))
return Val;
if (SDValue Val = performAddCombineForShiftedOperands(N, DCI.DAG))
return Val;
if (SDValue Val = performSubAddMULCombine(N, DCI.DAG))
return Val;
if (SDValue Val = performSVEMulAddSubCombine(N, DCI))
return Val;
if (SDValue Val = performAddSubIntoVectorOp(N, DCI.DAG))
return Val;
if (SDValue Val = performExtBinopLoadFold(N, DCI.DAG))
return Val;
return performAddSubLongCombine(N, DCI);
}
// Massage DAGs which we can use the high-half "long" operations on into
// something isel will recognize better. E.g.
//
// (aarch64_neon_umull (extract_high vec) (dupv64 scalar)) -->
// (aarch64_neon_umull (extract_high (v2i64 vec)))
// (extract_high (v2i64 (dup128 scalar)))))
//
static SDValue tryCombineLongOpWithDup(unsigned IID, SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (DCI.isBeforeLegalizeOps())
return SDValue();
SDValue LHS = N->getOperand((IID == Intrinsic::not_intrinsic) ? 0 : 1);
SDValue RHS = N->getOperand((IID == Intrinsic::not_intrinsic) ? 1 : 2);
assert(LHS.getValueType().is64BitVector() &&
RHS.getValueType().is64BitVector() &&
"unexpected shape for long operation");
// Either node could be a DUP, but it's not worth doing both of them (you'd
// just as well use the non-high version) so look for a corresponding extract
// operation on the other "wing".
if (isEssentiallyExtractHighSubvector(LHS)) {
RHS = tryExtendDUPToExtractHigh(RHS, DAG);
if (!RHS.getNode())
return SDValue();
} else if (isEssentiallyExtractHighSubvector(RHS)) {
LHS = tryExtendDUPToExtractHigh(LHS, DAG);
if (!LHS.getNode())
return SDValue();
} else
return SDValue();
if (IID == Intrinsic::not_intrinsic)
return DAG.getNode(N->getOpcode(), SDLoc(N), N->getValueType(0), LHS, RHS);
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, SDLoc(N), N->getValueType(0),
N->getOperand(0), LHS, RHS);
}
static SDValue tryCombineShiftImm(unsigned IID, SDNode *N, SelectionDAG &DAG) {
MVT ElemTy = N->getSimpleValueType(0).getScalarType();
unsigned ElemBits = ElemTy.getSizeInBits();
int64_t ShiftAmount;
if (BuildVectorSDNode *BVN = dyn_cast<BuildVectorSDNode>(N->getOperand(2))) {
APInt SplatValue, SplatUndef;
unsigned SplatBitSize;
bool HasAnyUndefs;
if (!BVN->isConstantSplat(SplatValue, SplatUndef, SplatBitSize,
HasAnyUndefs, ElemBits) ||
SplatBitSize != ElemBits)
return SDValue();
ShiftAmount = SplatValue.getSExtValue();
} else if (ConstantSDNode *CVN = dyn_cast<ConstantSDNode>(N->getOperand(2))) {
ShiftAmount = CVN->getSExtValue();
} else
return SDValue();
// If the shift amount is zero, remove the shift intrinsic.
if (ShiftAmount == 0 && IID != Intrinsic::aarch64_neon_sqshlu)
return N->getOperand(1);
unsigned Opcode;
bool IsRightShift;
switch (IID) {
default:
llvm_unreachable("Unknown shift intrinsic");
case Intrinsic::aarch64_neon_sqshl:
Opcode = AArch64ISD::SQSHL_I;
IsRightShift = false;
break;
case Intrinsic::aarch64_neon_uqshl:
Opcode = AArch64ISD::UQSHL_I;
IsRightShift = false;
break;
case Intrinsic::aarch64_neon_srshl:
Opcode = AArch64ISD::SRSHR_I;
IsRightShift = true;
break;
case Intrinsic::aarch64_neon_urshl:
Opcode = AArch64ISD::URSHR_I;
IsRightShift = true;
break;
case Intrinsic::aarch64_neon_sqshlu:
Opcode = AArch64ISD::SQSHLU_I;
IsRightShift = false;
break;
case Intrinsic::aarch64_neon_sshl:
case Intrinsic::aarch64_neon_ushl:
// For positive shift amounts we can use SHL, as ushl/sshl perform a regular
// left shift for positive shift amounts. For negative shifts we can use a
// VASHR/VLSHR as appropriate.
if (ShiftAmount < 0) {
Opcode = IID == Intrinsic::aarch64_neon_sshl ? AArch64ISD::VASHR
: AArch64ISD::VLSHR;
ShiftAmount = -ShiftAmount;
} else
Opcode = AArch64ISD::VSHL;
IsRightShift = false;
break;
}
EVT VT = N->getValueType(0);
SDValue Op = N->getOperand(1);
SDLoc DL(N);
if (VT == MVT::i64) {
Op = DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, MVT::v1i64, Op);
VT = MVT::v1i64;
}
if (IsRightShift && ShiftAmount <= -1 && ShiftAmount >= -(int)ElemBits) {
Op = DAG.getNode(Opcode, DL, VT, Op,
DAG.getSignedConstant(-ShiftAmount, DL, MVT::i32));
if (N->getValueType(0) == MVT::i64)
Op = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i64, Op,
DAG.getConstant(0, DL, MVT::i64));
return Op;
} else if (!IsRightShift && ShiftAmount >= 0 && ShiftAmount < ElemBits) {
Op = DAG.getNode(Opcode, DL, VT, Op,
DAG.getConstant(ShiftAmount, DL, MVT::i32));
if (N->getValueType(0) == MVT::i64)
Op = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i64, Op,
DAG.getConstant(0, DL, MVT::i64));
return Op;
}
return SDValue();
}
// The CRC32[BH] instructions ignore the high bits of their data operand. Since
// the intrinsics must be legal and take an i32, this means there's almost
// certainly going to be a zext in the DAG which we can eliminate.
static SDValue tryCombineCRC32(unsigned Mask, SDNode *N, SelectionDAG &DAG) {
SDValue AndN = N->getOperand(2);
if (AndN.getOpcode() != ISD::AND)
return SDValue();
ConstantSDNode *CMask = dyn_cast<ConstantSDNode>(AndN.getOperand(1));
if (!CMask || CMask->getZExtValue() != Mask)
return SDValue();
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, SDLoc(N), MVT::i32,
N->getOperand(0), N->getOperand(1), AndN.getOperand(0));
}
static SDValue combineAcrossLanesIntrinsic(unsigned Opc, SDNode *N,
SelectionDAG &DAG) {
SDLoc DL(N);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, N->getValueType(0),
DAG.getNode(Opc, DL, N->getOperand(1).getSimpleValueType(),
N->getOperand(1)),
DAG.getConstant(0, DL, MVT::i64));
}
static SDValue LowerSVEIntrinsicIndex(SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Op1 = N->getOperand(1);
SDValue Op2 = N->getOperand(2);
EVT ScalarTy = Op2.getValueType();
if ((ScalarTy == MVT::i8) || (ScalarTy == MVT::i16))
ScalarTy = MVT::i32;
// Lower index_vector(base, step) to mul(step step_vector(1)) + splat(base).
SDValue StepVector = DAG.getStepVector(DL, N->getValueType(0));
SDValue Step = DAG.getNode(ISD::SPLAT_VECTOR, DL, N->getValueType(0), Op2);
SDValue Mul = DAG.getNode(ISD::MUL, DL, N->getValueType(0), StepVector, Step);
SDValue Base = DAG.getNode(ISD::SPLAT_VECTOR, DL, N->getValueType(0), Op1);
return DAG.getNode(ISD::ADD, DL, N->getValueType(0), Mul, Base);
}
static SDValue LowerSVEIntrinsicDUP(SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Scalar = N->getOperand(3);
EVT ScalarTy = Scalar.getValueType();
if ((ScalarTy == MVT::i8) || (ScalarTy == MVT::i16))
Scalar = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, Scalar);
SDValue Passthru = N->getOperand(1);
SDValue Pred = N->getOperand(2);
return DAG.getNode(AArch64ISD::DUP_MERGE_PASSTHRU, DL, N->getValueType(0),
Pred, Scalar, Passthru);
}
static SDValue LowerSVEIntrinsicEXT(SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
LLVMContext &Ctx = *DAG.getContext();
EVT VT = N->getValueType(0);
assert(VT.isScalableVector() && "Expected a scalable vector.");
// Current lowering only supports the SVE-ACLE types.
if (VT.getSizeInBits().getKnownMinValue() != AArch64::SVEBitsPerBlock)
return SDValue();
unsigned ElemSize = VT.getVectorElementType().getSizeInBits() / 8;
unsigned ByteSize = VT.getSizeInBits().getKnownMinValue() / 8;
EVT ByteVT =
EVT::getVectorVT(Ctx, MVT::i8, ElementCount::getScalable(ByteSize));
// Convert everything to the domain of EXT (i.e bytes).
SDValue Op0 = DAG.getNode(ISD::BITCAST, DL, ByteVT, N->getOperand(1));
SDValue Op1 = DAG.getNode(ISD::BITCAST, DL, ByteVT, N->getOperand(2));
SDValue Op2 = DAG.getNode(ISD::MUL, DL, MVT::i32, N->getOperand(3),
DAG.getConstant(ElemSize, DL, MVT::i32));
SDValue EXT = DAG.getNode(AArch64ISD::EXT, DL, ByteVT, Op0, Op1, Op2);
return DAG.getNode(ISD::BITCAST, DL, VT, EXT);
}
static SDValue tryConvertSVEWideCompare(SDNode *N, ISD::CondCode CC,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (DCI.isBeforeLegalize())
return SDValue();
SDValue Comparator = N->getOperand(3);
if (Comparator.getOpcode() == AArch64ISD::DUP ||
Comparator.getOpcode() == ISD::SPLAT_VECTOR) {
unsigned IID = getIntrinsicID(N);
EVT VT = N->getValueType(0);
EVT CmpVT = N->getOperand(2).getValueType();
SDValue Pred = N->getOperand(1);
SDValue Imm;
SDLoc DL(N);
switch (IID) {
default:
llvm_unreachable("Called with wrong intrinsic!");
break;
// Signed comparisons
case Intrinsic::aarch64_sve_cmpeq_wide:
case Intrinsic::aarch64_sve_cmpne_wide:
case Intrinsic::aarch64_sve_cmpge_wide:
case Intrinsic::aarch64_sve_cmpgt_wide:
case Intrinsic::aarch64_sve_cmplt_wide:
case Intrinsic::aarch64_sve_cmple_wide: {
if (auto *CN = dyn_cast<ConstantSDNode>(Comparator.getOperand(0))) {
int64_t ImmVal = CN->getSExtValue();
if (ImmVal >= -16 && ImmVal <= 15)
Imm = DAG.getSignedConstant(ImmVal, DL, MVT::i32);
else
return SDValue();
}
break;
}
// Unsigned comparisons
case Intrinsic::aarch64_sve_cmphs_wide:
case Intrinsic::aarch64_sve_cmphi_wide:
case Intrinsic::aarch64_sve_cmplo_wide:
case Intrinsic::aarch64_sve_cmpls_wide: {
if (auto *CN = dyn_cast<ConstantSDNode>(Comparator.getOperand(0))) {
uint64_t ImmVal = CN->getZExtValue();
if (ImmVal <= 127)
Imm = DAG.getConstant(ImmVal, DL, MVT::i32);
else
return SDValue();
}
break;
}
}
if (!Imm)
return SDValue();
SDValue Splat = DAG.getNode(ISD::SPLAT_VECTOR, DL, CmpVT, Imm);
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, DL, VT, Pred,
N->getOperand(2), Splat, DAG.getCondCode(CC));
}
return SDValue();
}
static SDValue getPTest(SelectionDAG &DAG, EVT VT, SDValue Pg, SDValue Op,
AArch64CC::CondCode Cond) {
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
SDLoc DL(Op);
assert(Op.getValueType().isScalableVector() &&
TLI.isTypeLegal(Op.getValueType()) &&
"Expected legal scalable vector type!");
assert(Op.getValueType() == Pg.getValueType() &&
"Expected same type for PTEST operands");
// Ensure target specific opcodes are using legal type.
EVT OutVT = TLI.getTypeToTransformTo(*DAG.getContext(), VT);
SDValue TVal = DAG.getConstant(1, DL, OutVT);
SDValue FVal = DAG.getConstant(0, DL, OutVT);
// Ensure operands have type nxv16i1.
if (Op.getValueType() != MVT::nxv16i1) {
if ((Cond == AArch64CC::ANY_ACTIVE || Cond == AArch64CC::NONE_ACTIVE) &&
isZeroingInactiveLanes(Op))
Pg = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, MVT::nxv16i1, Pg);
else
Pg = getSVEPredicateBitCast(MVT::nxv16i1, Pg, DAG);
Op = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, MVT::nxv16i1, Op);
}
// Set condition code (CC) flags.
SDValue Test = DAG.getNode(
Cond == AArch64CC::ANY_ACTIVE ? AArch64ISD::PTEST_ANY : AArch64ISD::PTEST,
DL, MVT::i32, Pg, Op);
// Convert CC to integer based on requested condition.
// NOTE: Cond is inverted to promote CSEL's removal when it feeds a compare.
SDValue CC = DAG.getConstant(getInvertedCondCode(Cond), DL, MVT::i32);
SDValue Res = DAG.getNode(AArch64ISD::CSEL, DL, OutVT, FVal, TVal, CC, Test);
return DAG.getZExtOrTrunc(Res, DL, VT);
}
static SDValue combineSVEReductionInt(SDNode *N, unsigned Opc,
SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Pred = N->getOperand(1);
SDValue VecToReduce = N->getOperand(2);
// NOTE: The integer reduction's result type is not always linked to the
// operand's element type so we construct it from the intrinsic's result type.
EVT ReduceVT = getPackedSVEVectorVT(N->getValueType(0));
SDValue Reduce = DAG.getNode(Opc, DL, ReduceVT, Pred, VecToReduce);
// SVE reductions set the whole vector register with the first element
// containing the reduction result, which we'll now extract.
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, N->getValueType(0), Reduce,
Zero);
}
static SDValue combineSVEReductionFP(SDNode *N, unsigned Opc,
SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Pred = N->getOperand(1);
SDValue VecToReduce = N->getOperand(2);
EVT ReduceVT = VecToReduce.getValueType();
SDValue Reduce = DAG.getNode(Opc, DL, ReduceVT, Pred, VecToReduce);
// SVE reductions set the whole vector register with the first element
// containing the reduction result, which we'll now extract.
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, N->getValueType(0), Reduce,
Zero);
}
static SDValue combineSVEReductionOrderedFP(SDNode *N, unsigned Opc,
SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Pred = N->getOperand(1);
SDValue InitVal = N->getOperand(2);
SDValue VecToReduce = N->getOperand(3);
EVT ReduceVT = VecToReduce.getValueType();
// Ordered reductions use the first lane of the result vector as the
// reduction's initial value.
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
InitVal = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, ReduceVT,
DAG.getUNDEF(ReduceVT), InitVal, Zero);
SDValue Reduce = DAG.getNode(Opc, DL, ReduceVT, Pred, InitVal, VecToReduce);
// SVE reductions set the whole vector register with the first element
// containing the reduction result, which we'll now extract.
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, N->getValueType(0), Reduce,
Zero);
}
// If a merged operation has no inactive lanes we can relax it to a predicated
// or unpredicated operation, which potentially allows better isel (perhaps
// using immediate forms) or relaxing register reuse requirements.
static SDValue convertMergedOpToPredOp(SDNode *N, unsigned Opc,
SelectionDAG &DAG, bool UnpredOp = false,
bool SwapOperands = false) {
assert(N->getOpcode() == ISD::INTRINSIC_WO_CHAIN && "Expected intrinsic!");
assert(N->getNumOperands() == 4 && "Expected 3 operand intrinsic!");
SDValue Pg = N->getOperand(1);
SDValue Op1 = N->getOperand(SwapOperands ? 3 : 2);
SDValue Op2 = N->getOperand(SwapOperands ? 2 : 3);
// ISD way to specify an all active predicate.
if (isAllActivePredicate(DAG, Pg)) {
if (UnpredOp)
return DAG.getNode(Opc, SDLoc(N), N->getValueType(0), Op1, Op2);
return DAG.getNode(Opc, SDLoc(N), N->getValueType(0), Pg, Op1, Op2);
}
// FUTURE: SplatVector(true)
return SDValue();
}
SDValue tryLowerPartialReductionToDot(SDNode *N,
const AArch64Subtarget *Subtarget,
SelectionDAG &DAG) {
assert(N->getOpcode() == ISD::INTRINSIC_WO_CHAIN &&
getIntrinsicID(N) ==
Intrinsic::experimental_vector_partial_reduce_add &&
"Expected a partial reduction node");
bool Scalable = N->getValueType(0).isScalableVector();
if (Scalable && !Subtarget->isSVEorStreamingSVEAvailable())
return SDValue();
if (!Scalable && (!Subtarget->isNeonAvailable() || !Subtarget->hasDotProd()))
return SDValue();
SDLoc DL(N);
SDValue Op2 = N->getOperand(2);
unsigned Op2Opcode = Op2->getOpcode();
SDValue MulOpLHS, MulOpRHS;
bool MulOpLHSIsSigned, MulOpRHSIsSigned;
if (ISD::isExtOpcode(Op2Opcode)) {
MulOpLHSIsSigned = MulOpRHSIsSigned = (Op2Opcode == ISD::SIGN_EXTEND);
MulOpLHS = Op2->getOperand(0);
MulOpRHS = DAG.getConstant(1, DL, MulOpLHS.getValueType());
} else if (Op2Opcode == ISD::MUL) {
SDValue ExtMulOpLHS = Op2->getOperand(0);
SDValue ExtMulOpRHS = Op2->getOperand(1);
unsigned ExtMulOpLHSOpcode = ExtMulOpLHS->getOpcode();
unsigned ExtMulOpRHSOpcode = ExtMulOpRHS->getOpcode();
if (!ISD::isExtOpcode(ExtMulOpLHSOpcode) ||
!ISD::isExtOpcode(ExtMulOpRHSOpcode))
return SDValue();
MulOpLHSIsSigned = ExtMulOpLHSOpcode == ISD::SIGN_EXTEND;
MulOpRHSIsSigned = ExtMulOpRHSOpcode == ISD::SIGN_EXTEND;
MulOpLHS = ExtMulOpLHS->getOperand(0);
MulOpRHS = ExtMulOpRHS->getOperand(0);
if (MulOpLHS.getValueType() != MulOpRHS.getValueType())
return SDValue();
} else
return SDValue();
SDValue Acc = N->getOperand(1);
EVT ReducedVT = N->getValueType(0);
EVT MulSrcVT = MulOpLHS.getValueType();
// Dot products operate on chunks of four elements so there must be four times
// as many elements in the wide type
if (!(ReducedVT == MVT::nxv4i64 && MulSrcVT == MVT::nxv16i8) &&
!(ReducedVT == MVT::nxv4i32 && MulSrcVT == MVT::nxv16i8) &&
!(ReducedVT == MVT::nxv2i64 && MulSrcVT == MVT::nxv8i16) &&
!(ReducedVT == MVT::v4i64 && MulSrcVT == MVT::v16i8) &&
!(ReducedVT == MVT::v4i32 && MulSrcVT == MVT::v16i8) &&
!(ReducedVT == MVT::v2i32 && MulSrcVT == MVT::v8i8))
return SDValue();
// If the extensions are mixed, we should lower it to a usdot instead
unsigned Opcode = 0;
if (MulOpLHSIsSigned != MulOpRHSIsSigned) {
if (!Subtarget->hasMatMulInt8())
return SDValue();
bool Scalable = N->getValueType(0).isScalableVT();
// There's no nxv2i64 version of usdot
if (Scalable && ReducedVT != MVT::nxv4i32 && ReducedVT != MVT::nxv4i64)
return SDValue();
Opcode = AArch64ISD::USDOT;
// USDOT expects the signed operand to be last
if (!MulOpRHSIsSigned)
std::swap(MulOpLHS, MulOpRHS);
} else
Opcode = MulOpLHSIsSigned ? AArch64ISD::SDOT : AArch64ISD::UDOT;
// Partial reduction lowering for (nx)v16i8 to (nx)v4i64 requires an i32 dot
// product followed by a zero / sign extension
if ((ReducedVT == MVT::nxv4i64 && MulSrcVT == MVT::nxv16i8) ||
(ReducedVT == MVT::v4i64 && MulSrcVT == MVT::v16i8)) {
EVT ReducedVTI32 =
(ReducedVT.isScalableVector()) ? MVT::nxv4i32 : MVT::v4i32;
SDValue DotI32 =
DAG.getNode(Opcode, DL, ReducedVTI32,
DAG.getConstant(0, DL, ReducedVTI32), MulOpLHS, MulOpRHS);
SDValue Extended = DAG.getSExtOrTrunc(DotI32, DL, ReducedVT);
return DAG.getNode(ISD::ADD, DL, ReducedVT, Acc, Extended);
}
return DAG.getNode(Opcode, DL, ReducedVT, Acc, MulOpLHS, MulOpRHS);
}
SDValue tryLowerPartialReductionToWideAdd(SDNode *N,
const AArch64Subtarget *Subtarget,
SelectionDAG &DAG) {
assert(N->getOpcode() == ISD::INTRINSIC_WO_CHAIN &&
getIntrinsicID(N) ==
Intrinsic::experimental_vector_partial_reduce_add &&
"Expected a partial reduction node");
if (!Subtarget->hasSVE2() && !Subtarget->isStreamingSVEAvailable())
return SDValue();
SDLoc DL(N);
if (!ISD::isExtOpcode(N->getOperand(2).getOpcode()))
return SDValue();
SDValue Acc = N->getOperand(1);
SDValue Ext = N->getOperand(2);
EVT AccVT = Acc.getValueType();
EVT ExtVT = Ext.getValueType();
if (ExtVT.getVectorElementType() != AccVT.getVectorElementType())
return SDValue();
SDValue ExtOp = Ext->getOperand(0);
EVT ExtOpVT = ExtOp.getValueType();
if (!(ExtOpVT == MVT::nxv4i32 && AccVT == MVT::nxv2i64) &&
!(ExtOpVT == MVT::nxv8i16 && AccVT == MVT::nxv4i32) &&
!(ExtOpVT == MVT::nxv16i8 && AccVT == MVT::nxv8i16))
return SDValue();
bool ExtOpIsSigned = Ext.getOpcode() == ISD::SIGN_EXTEND;
unsigned BottomOpcode =
ExtOpIsSigned ? AArch64ISD::SADDWB : AArch64ISD::UADDWB;
unsigned TopOpcode = ExtOpIsSigned ? AArch64ISD::SADDWT : AArch64ISD::UADDWT;
SDValue BottomNode = DAG.getNode(BottomOpcode, DL, AccVT, Acc, ExtOp);
return DAG.getNode(TopOpcode, DL, AccVT, BottomNode, ExtOp);
}
static SDValue combineSVEBitSel(unsigned IID, SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
EVT VT = N->getValueType(0);
SDValue Op1 = N->getOperand(1);
SDValue Op2 = N->getOperand(2);
SDValue Op3 = N->getOperand(3);
switch (IID) {
default:
llvm_unreachable("Called with wrong intrinsic!");
case Intrinsic::aarch64_sve_bsl:
return DAG.getNode(AArch64ISD::BSP, DL, VT, Op3, Op1, Op2);
case Intrinsic::aarch64_sve_bsl1n:
return DAG.getNode(AArch64ISD::BSP, DL, VT, Op3, DAG.getNOT(DL, Op1, VT),
Op2);
case Intrinsic::aarch64_sve_bsl2n:
return DAG.getNode(AArch64ISD::BSP, DL, VT, Op3, Op1,
DAG.getNOT(DL, Op2, VT));
case Intrinsic::aarch64_sve_nbsl:
return DAG.getNOT(DL, DAG.getNode(AArch64ISD::BSP, DL, VT, Op3, Op1, Op2),
VT);
}
}
static SDValue performIntrinsicCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
SelectionDAG &DAG = DCI.DAG;
unsigned IID = getIntrinsicID(N);
switch (IID) {
default:
break;
case Intrinsic::experimental_vector_partial_reduce_add: {
if (SDValue Dot = tryLowerPartialReductionToDot(N, Subtarget, DAG))
return Dot;
if (SDValue WideAdd = tryLowerPartialReductionToWideAdd(N, Subtarget, DAG))
return WideAdd;
SDLoc DL(N);
SDValue Input = N->getOperand(2);
return DAG.getNode(ISD::PARTIAL_REDUCE_UMLA, DL, N->getValueType(0),
N->getOperand(1), Input,
DAG.getConstant(1, DL, Input.getValueType()));
}
case Intrinsic::aarch64_neon_vcvtfxs2fp:
case Intrinsic::aarch64_neon_vcvtfxu2fp:
return tryCombineFixedPointConvert(N, DCI, DAG);
case Intrinsic::aarch64_neon_saddv:
return combineAcrossLanesIntrinsic(AArch64ISD::SADDV, N, DAG);
case Intrinsic::aarch64_neon_uaddv:
return combineAcrossLanesIntrinsic(AArch64ISD::UADDV, N, DAG);
case Intrinsic::aarch64_neon_sminv:
return combineAcrossLanesIntrinsic(AArch64ISD::SMINV, N, DAG);
case Intrinsic::aarch64_neon_uminv:
return combineAcrossLanesIntrinsic(AArch64ISD::UMINV, N, DAG);
case Intrinsic::aarch64_neon_smaxv:
return combineAcrossLanesIntrinsic(AArch64ISD::SMAXV, N, DAG);
case Intrinsic::aarch64_neon_umaxv:
return combineAcrossLanesIntrinsic(AArch64ISD::UMAXV, N, DAG);
case Intrinsic::aarch64_neon_fmax:
return DAG.getNode(ISD::FMAXIMUM, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_fmin:
return DAG.getNode(ISD::FMINIMUM, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_fmaxnm:
return DAG.getNode(ISD::FMAXNUM, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_fminnm:
return DAG.getNode(ISD::FMINNUM, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_smull:
return DAG.getNode(AArch64ISD::SMULL, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_umull:
return DAG.getNode(AArch64ISD::UMULL, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_pmull:
return DAG.getNode(AArch64ISD::PMULL, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_sqdmull:
return tryCombineLongOpWithDup(IID, N, DCI, DAG);
case Intrinsic::aarch64_neon_sqshl:
case Intrinsic::aarch64_neon_uqshl:
case Intrinsic::aarch64_neon_sqshlu:
case Intrinsic::aarch64_neon_srshl:
case Intrinsic::aarch64_neon_urshl:
case Intrinsic::aarch64_neon_sshl:
case Intrinsic::aarch64_neon_ushl:
return tryCombineShiftImm(IID, N, DAG);
case Intrinsic::aarch64_neon_sabd:
return DAG.getNode(ISD::ABDS, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_neon_uabd:
return DAG.getNode(ISD::ABDU, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_crc32b:
case Intrinsic::aarch64_crc32cb:
return tryCombineCRC32(0xff, N, DAG);
case Intrinsic::aarch64_crc32h:
case Intrinsic::aarch64_crc32ch:
return tryCombineCRC32(0xffff, N, DAG);
case Intrinsic::aarch64_sve_saddv:
// There is no i64 version of SADDV because the sign is irrelevant.
if (N->getOperand(2)->getValueType(0).getVectorElementType() == MVT::i64)
return combineSVEReductionInt(N, AArch64ISD::UADDV_PRED, DAG);
else
return combineSVEReductionInt(N, AArch64ISD::SADDV_PRED, DAG);
case Intrinsic::aarch64_sve_uaddv:
return combineSVEReductionInt(N, AArch64ISD::UADDV_PRED, DAG);
case Intrinsic::aarch64_sve_smaxv:
return combineSVEReductionInt(N, AArch64ISD::SMAXV_PRED, DAG);
case Intrinsic::aarch64_sve_umaxv:
return combineSVEReductionInt(N, AArch64ISD::UMAXV_PRED, DAG);
case Intrinsic::aarch64_sve_sminv:
return combineSVEReductionInt(N, AArch64ISD::SMINV_PRED, DAG);
case Intrinsic::aarch64_sve_uminv:
return combineSVEReductionInt(N, AArch64ISD::UMINV_PRED, DAG);
case Intrinsic::aarch64_sve_orv:
return combineSVEReductionInt(N, AArch64ISD::ORV_PRED, DAG);
case Intrinsic::aarch64_sve_eorv:
return combineSVEReductionInt(N, AArch64ISD::EORV_PRED, DAG);
case Intrinsic::aarch64_sve_andv:
return combineSVEReductionInt(N, AArch64ISD::ANDV_PRED, DAG);
case Intrinsic::aarch64_sve_index:
return LowerSVEIntrinsicIndex(N, DAG);
case Intrinsic::aarch64_sve_dup:
return LowerSVEIntrinsicDUP(N, DAG);
case Intrinsic::aarch64_sve_dup_x:
return DAG.getNode(ISD::SPLAT_VECTOR, SDLoc(N), N->getValueType(0),
N->getOperand(1));
case Intrinsic::aarch64_sve_ext:
return LowerSVEIntrinsicEXT(N, DAG);
case Intrinsic::aarch64_sve_mul_u:
return DAG.getNode(AArch64ISD::MUL_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_smulh_u:
return DAG.getNode(AArch64ISD::MULHS_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_umulh_u:
return DAG.getNode(AArch64ISD::MULHU_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_smin_u:
return DAG.getNode(AArch64ISD::SMIN_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_umin_u:
return DAG.getNode(AArch64ISD::UMIN_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_smax_u:
return DAG.getNode(AArch64ISD::SMAX_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_umax_u:
return DAG.getNode(AArch64ISD::UMAX_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_lsl_u:
return DAG.getNode(AArch64ISD::SHL_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_lsr_u:
return DAG.getNode(AArch64ISD::SRL_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_asr_u:
return DAG.getNode(AArch64ISD::SRA_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fadd_u:
return DAG.getNode(AArch64ISD::FADD_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fdiv_u:
return DAG.getNode(AArch64ISD::FDIV_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fmax_u:
return DAG.getNode(AArch64ISD::FMAX_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fmaxnm_u:
return DAG.getNode(AArch64ISD::FMAXNM_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fmla_u:
return DAG.getNode(AArch64ISD::FMA_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(3), N->getOperand(4),
N->getOperand(2));
case Intrinsic::aarch64_sve_fmin_u:
return DAG.getNode(AArch64ISD::FMIN_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fminnm_u:
return DAG.getNode(AArch64ISD::FMINNM_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fmul_u:
return DAG.getNode(AArch64ISD::FMUL_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_fsub_u:
return DAG.getNode(AArch64ISD::FSUB_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_add_u:
return DAG.getNode(ISD::ADD, SDLoc(N), N->getValueType(0), N->getOperand(2),
N->getOperand(3));
case Intrinsic::aarch64_sve_sub_u:
return DAG.getNode(ISD::SUB, SDLoc(N), N->getValueType(0), N->getOperand(2),
N->getOperand(3));
case Intrinsic::aarch64_sve_subr:
return convertMergedOpToPredOp(N, ISD::SUB, DAG, true, true);
case Intrinsic::aarch64_sve_and_u:
return DAG.getNode(ISD::AND, SDLoc(N), N->getValueType(0), N->getOperand(2),
N->getOperand(3));
case Intrinsic::aarch64_sve_bic_u:
return DAG.getNode(AArch64ISD::BIC, SDLoc(N), N->getValueType(0),
N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_saddwb:
return DAG.getNode(AArch64ISD::SADDWB, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_saddwt:
return DAG.getNode(AArch64ISD::SADDWT, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_uaddwb:
return DAG.getNode(AArch64ISD::UADDWB, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_uaddwt:
return DAG.getNode(AArch64ISD::UADDWT, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_eor_u:
return DAG.getNode(ISD::XOR, SDLoc(N), N->getValueType(0), N->getOperand(2),
N->getOperand(3));
case Intrinsic::aarch64_sve_orr_u:
return DAG.getNode(ISD::OR, SDLoc(N), N->getValueType(0), N->getOperand(2),
N->getOperand(3));
case Intrinsic::aarch64_sve_sabd_u:
return DAG.getNode(ISD::ABDS, SDLoc(N), N->getValueType(0),
N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_uabd_u:
return DAG.getNode(ISD::ABDU, SDLoc(N), N->getValueType(0),
N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_sdiv_u:
return DAG.getNode(AArch64ISD::SDIV_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_udiv_u:
return DAG.getNode(AArch64ISD::UDIV_PRED, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_sqadd:
return convertMergedOpToPredOp(N, ISD::SADDSAT, DAG, true);
case Intrinsic::aarch64_sve_sqsub_u:
return DAG.getNode(ISD::SSUBSAT, SDLoc(N), N->getValueType(0),
N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_uqadd:
return convertMergedOpToPredOp(N, ISD::UADDSAT, DAG, true);
case Intrinsic::aarch64_sve_uqsub_u:
return DAG.getNode(ISD::USUBSAT, SDLoc(N), N->getValueType(0),
N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_sqadd_x:
return DAG.getNode(ISD::SADDSAT, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_sqsub_x:
return DAG.getNode(ISD::SSUBSAT, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_uqadd_x:
return DAG.getNode(ISD::UADDSAT, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_uqsub_x:
return DAG.getNode(ISD::USUBSAT, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_asrd:
return DAG.getNode(AArch64ISD::SRAD_MERGE_OP1, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_cmphs:
if (!N->getOperand(2).getValueType().isFloatingPoint())
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, SDLoc(N),
N->getValueType(0), N->getOperand(1), N->getOperand(2),
N->getOperand(3), DAG.getCondCode(ISD::SETUGE));
break;
case Intrinsic::aarch64_sve_cmphi:
if (!N->getOperand(2).getValueType().isFloatingPoint())
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, SDLoc(N),
N->getValueType(0), N->getOperand(1), N->getOperand(2),
N->getOperand(3), DAG.getCondCode(ISD::SETUGT));
break;
case Intrinsic::aarch64_sve_fcmpge:
case Intrinsic::aarch64_sve_cmpge:
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, SDLoc(N),
N->getValueType(0), N->getOperand(1), N->getOperand(2),
N->getOperand(3), DAG.getCondCode(ISD::SETGE));
break;
case Intrinsic::aarch64_sve_fcmpgt:
case Intrinsic::aarch64_sve_cmpgt:
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, SDLoc(N),
N->getValueType(0), N->getOperand(1), N->getOperand(2),
N->getOperand(3), DAG.getCondCode(ISD::SETGT));
break;
case Intrinsic::aarch64_sve_fcmpeq:
case Intrinsic::aarch64_sve_cmpeq:
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, SDLoc(N),
N->getValueType(0), N->getOperand(1), N->getOperand(2),
N->getOperand(3), DAG.getCondCode(ISD::SETEQ));
break;
case Intrinsic::aarch64_sve_fcmpne:
case Intrinsic::aarch64_sve_cmpne:
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, SDLoc(N),
N->getValueType(0), N->getOperand(1), N->getOperand(2),
N->getOperand(3), DAG.getCondCode(ISD::SETNE));
break;
case Intrinsic::aarch64_sve_fcmpuo:
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, SDLoc(N),
N->getValueType(0), N->getOperand(1), N->getOperand(2),
N->getOperand(3), DAG.getCondCode(ISD::SETUO));
break;
case Intrinsic::aarch64_sve_fadda:
return combineSVEReductionOrderedFP(N, AArch64ISD::FADDA_PRED, DAG);
case Intrinsic::aarch64_sve_faddv:
return combineSVEReductionFP(N, AArch64ISD::FADDV_PRED, DAG);
case Intrinsic::aarch64_sve_fmaxnmv:
return combineSVEReductionFP(N, AArch64ISD::FMAXNMV_PRED, DAG);
case Intrinsic::aarch64_sve_fmaxv:
return combineSVEReductionFP(N, AArch64ISD::FMAXV_PRED, DAG);
case Intrinsic::aarch64_sve_fminnmv:
return combineSVEReductionFP(N, AArch64ISD::FMINNMV_PRED, DAG);
case Intrinsic::aarch64_sve_fminv:
return combineSVEReductionFP(N, AArch64ISD::FMINV_PRED, DAG);
case Intrinsic::aarch64_sve_sel:
return DAG.getNode(ISD::VSELECT, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sve_cmpeq_wide:
return tryConvertSVEWideCompare(N, ISD::SETEQ, DCI, DAG);
case Intrinsic::aarch64_sve_cmpne_wide:
return tryConvertSVEWideCompare(N, ISD::SETNE, DCI, DAG);
case Intrinsic::aarch64_sve_cmpge_wide:
return tryConvertSVEWideCompare(N, ISD::SETGE, DCI, DAG);
case Intrinsic::aarch64_sve_cmpgt_wide:
return tryConvertSVEWideCompare(N, ISD::SETGT, DCI, DAG);
case Intrinsic::aarch64_sve_cmplt_wide:
return tryConvertSVEWideCompare(N, ISD::SETLT, DCI, DAG);
case Intrinsic::aarch64_sve_cmple_wide:
return tryConvertSVEWideCompare(N, ISD::SETLE, DCI, DAG);
case Intrinsic::aarch64_sve_cmphs_wide:
return tryConvertSVEWideCompare(N, ISD::SETUGE, DCI, DAG);
case Intrinsic::aarch64_sve_cmphi_wide:
return tryConvertSVEWideCompare(N, ISD::SETUGT, DCI, DAG);
case Intrinsic::aarch64_sve_cmplo_wide:
return tryConvertSVEWideCompare(N, ISD::SETULT, DCI, DAG);
case Intrinsic::aarch64_sve_cmpls_wide:
return tryConvertSVEWideCompare(N, ISD::SETULE, DCI, DAG);
case Intrinsic::aarch64_sve_ptest_any:
return getPTest(DAG, N->getValueType(0), N->getOperand(1), N->getOperand(2),
AArch64CC::ANY_ACTIVE);
case Intrinsic::aarch64_sve_ptest_first:
return getPTest(DAG, N->getValueType(0), N->getOperand(1), N->getOperand(2),
AArch64CC::FIRST_ACTIVE);
case Intrinsic::aarch64_sve_ptest_last:
return getPTest(DAG, N->getValueType(0), N->getOperand(1), N->getOperand(2),
AArch64CC::LAST_ACTIVE);
case Intrinsic::aarch64_sve_whilelo:
return DAG.getNode(ISD::GET_ACTIVE_LANE_MASK, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
case Intrinsic::aarch64_sve_bsl:
case Intrinsic::aarch64_sve_bsl1n:
case Intrinsic::aarch64_sve_bsl2n:
case Intrinsic::aarch64_sve_nbsl:
return combineSVEBitSel(IID, N, DAG);
}
return SDValue();
}
static bool isCheapToExtend(const SDValue &N) {
unsigned OC = N->getOpcode();
return OC == ISD::LOAD || OC == ISD::MLOAD ||
ISD::isConstantSplatVectorAllZeros(N.getNode());
}
static SDValue
performSignExtendSetCCCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
// If we have (sext (setcc A B)) and A and B are cheap to extend,
// we can move the sext into the arguments and have the same result. For
// example, if A and B are both loads, we can make those extending loads and
// avoid an extra instruction. This pattern appears often in VLS code
// generation where the inputs to the setcc have a different size to the
// instruction that wants to use the result of the setcc.
assert(N->getOpcode() == ISD::SIGN_EXTEND &&
N->getOperand(0)->getOpcode() == ISD::SETCC);
const SDValue SetCC = N->getOperand(0);
const SDValue CCOp0 = SetCC.getOperand(0);
const SDValue CCOp1 = SetCC.getOperand(1);
if (!CCOp0->getValueType(0).isInteger() ||
!CCOp1->getValueType(0).isInteger())
return SDValue();
ISD::CondCode Code =
cast<CondCodeSDNode>(SetCC->getOperand(2).getNode())->get();
ISD::NodeType ExtType =
isSignedIntSetCC(Code) ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND;
if (isCheapToExtend(SetCC.getOperand(0)) &&
isCheapToExtend(SetCC.getOperand(1))) {
const SDValue Ext1 =
DAG.getNode(ExtType, SDLoc(N), N->getValueType(0), CCOp0);
const SDValue Ext2 =
DAG.getNode(ExtType, SDLoc(N), N->getValueType(0), CCOp1);
return DAG.getSetCC(
SDLoc(SetCC), N->getValueType(0), Ext1, Ext2,
cast<CondCodeSDNode>(SetCC->getOperand(2).getNode())->get());
}
return SDValue();
}
// Convert zext(extract(shuffle a, b, [0,4,8,12])) -> and(uzp1(a, b), 255)
// This comes from interleaved vectorization. It is performed late to capture
// uitofp converts too.
static SDValue performZExtDeinterleaveShuffleCombine(SDNode *N,
SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if ((VT != MVT::v4i32 && VT != MVT::v8i16) ||
N->getOpcode() != ISD::ZERO_EXTEND ||
N->getOperand(0).getOpcode() != ISD::EXTRACT_SUBVECTOR)
return SDValue();
unsigned ExtOffset = N->getOperand(0).getConstantOperandVal(1);
if (ExtOffset != 0 && ExtOffset != VT.getVectorNumElements())
return SDValue();
EVT InVT = N->getOperand(0).getOperand(0).getValueType();
auto *Shuffle = dyn_cast<ShuffleVectorSDNode>(N->getOperand(0).getOperand(0));
if (!Shuffle ||
InVT.getVectorNumElements() != VT.getVectorNumElements() * 2 ||
InVT.getScalarSizeInBits() * 2 != VT.getScalarSizeInBits())
return SDValue();
unsigned Idx;
bool IsDeInterleave = ShuffleVectorInst::isDeInterleaveMaskOfFactor(
Shuffle->getMask().slice(ExtOffset, VT.getVectorNumElements()), 4, Idx);
// An undef interleave shuffle can come up after other canonicalizations,
// where the shuffle has been converted to
// zext(extract(shuffle b, undef, [u,u,0,4]))
bool IsUndefDeInterleave = false;
if (!IsDeInterleave)
IsUndefDeInterleave =
Shuffle->getOperand(1).isUndef() &&
all_of(
Shuffle->getMask().slice(ExtOffset, VT.getVectorNumElements() / 2),
[](int M) { return M < 0; }) &&
ShuffleVectorInst::isDeInterleaveMaskOfFactor(
Shuffle->getMask().slice(ExtOffset + VT.getVectorNumElements() / 2,
VT.getVectorNumElements() / 2),
4, Idx);
if ((!IsDeInterleave && !IsUndefDeInterleave) || Idx >= 4)
return SDValue();
SDLoc DL(N);
SDValue BC1 = DAG.getNode(AArch64ISD::NVCAST, DL, VT,
Shuffle->getOperand(IsUndefDeInterleave ? 1 : 0));
SDValue BC2 = DAG.getNode(AArch64ISD::NVCAST, DL, VT,
Shuffle->getOperand(IsUndefDeInterleave ? 0 : 1));
SDValue UZP = DAG.getNode(Idx < 2 ? AArch64ISD::UZP1 : AArch64ISD::UZP2, DL,
VT, BC1, BC2);
if ((Idx & 1) == 1)
UZP = DAG.getNode(ISD::SRL, DL, VT, UZP,
DAG.getConstant(InVT.getScalarSizeInBits(), DL, VT));
return DAG.getNode(
ISD::AND, DL, VT, UZP,
DAG.getConstant((1 << InVT.getScalarSizeInBits()) - 1, DL, VT));
}
// This comes up similar to the above when lowering deinterleaving shuffles from
// zexts. We have legalized the operations in the generally case to
// zext(extract_subvector(uzp(a, b))), which can be converted to and(a, mask) if
// the extract is to the low half and the uzp is uzp1. There would be an extra
// shift if the uzp was uzp2 to grab the upper half. Due to the combine above
// there could also be an existing and / shift that can be combined in, either
// before of after the extract.
static SDValue performZExtUZPCombine(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
if (N->getOpcode() != ISD::ZERO_EXTEND ||
(VT != MVT::v2i64 && VT != MVT::v4i32 && VT != MVT::v8i16))
return SDValue();
SDValue Op = N->getOperand(0);
unsigned ExtOffset = (unsigned)-1;
if (Op.getOpcode() == ISD::EXTRACT_SUBVECTOR) {
ExtOffset = Op.getConstantOperandVal(1);
Op = Op.getOperand(0);
}
unsigned Shift = 0;
APInt Mask = APInt::getLowBitsSet(VT.getScalarSizeInBits(),
Op.getValueType().getScalarSizeInBits());
if (Op.getOpcode() == AArch64ISD::VLSHR) {
Shift = Op.getConstantOperandVal(1);
Op = Op.getOperand(0);
Mask = Mask.lshr(Shift);
}
if (Op.getOpcode() == ISD::AND &&
ISD::isConstantSplatVector(Op.getOperand(1).getNode(), Mask)) {
Op = Op.getOperand(0);
Mask = Mask.zext(VT.getScalarSizeInBits());
} else if (Op.getOpcode() == AArch64ISD::BICi) {
Mask = ~APInt(Op.getValueType().getScalarSizeInBits(),
Op.getConstantOperandVal(1) << Op.getConstantOperandVal(2));
Mask = Mask.zext(VT.getScalarSizeInBits());
Op = Op.getOperand(0);
}
if (ExtOffset == (unsigned)-1) {
if (Op.getOpcode() == ISD::EXTRACT_SUBVECTOR) {
ExtOffset = Op.getConstantOperandVal(1);
Op = Op.getOperand(0);
} else
return SDValue();
}
if (ExtOffset != 0 && ExtOffset != VT.getVectorNumElements())
return SDValue();
if (Op.getOpcode() != AArch64ISD::UZP1 && Op.getOpcode() != AArch64ISD::UZP2)
return SDValue();
if (Op.getOpcode() == AArch64ISD::UZP2)
Shift += VT.getScalarSizeInBits() / 2;
SDLoc DL(N);
SDValue BC = DAG.getNode(AArch64ISD::NVCAST, DL, VT,
Op.getOperand(ExtOffset == 0 ? 0 : 1));
if (Shift != 0)
BC = DAG.getNode(AArch64ISD::VLSHR, DL, VT, BC,
DAG.getConstant(Shift, DL, MVT::i32));
return DAG.getNode(ISD::AND, DL, VT, BC, DAG.getConstant(Mask, DL, VT));
}
static SDValue performExtendCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
// If we see something like (zext (sabd (extract_high ...), (DUP ...))) then
// we can convert that DUP into another extract_high (of a bigger DUP), which
// helps the backend to decide that an sabdl2 would be useful, saving a real
// extract_high operation.
if (!DCI.isBeforeLegalizeOps() && N->getOpcode() == ISD::ZERO_EXTEND &&
N->getOperand(0).getValueType().is64BitVector() &&
(N->getOperand(0).getOpcode() == ISD::ABDU ||
N->getOperand(0).getOpcode() == ISD::ABDS)) {
SDNode *ABDNode = N->getOperand(0).getNode();
SDValue NewABD =
tryCombineLongOpWithDup(Intrinsic::not_intrinsic, ABDNode, DCI, DAG);
if (!NewABD.getNode())
return SDValue();
return DAG.getNode(ISD::ZERO_EXTEND, SDLoc(N), N->getValueType(0), NewABD);
}
if (SDValue R = performZExtDeinterleaveShuffleCombine(N, DAG))
return R;
if (SDValue R = performZExtUZPCombine(N, DAG))
return R;
if (N->getValueType(0).isFixedLengthVector() &&
N->getOpcode() == ISD::SIGN_EXTEND &&
N->getOperand(0)->getOpcode() == ISD::SETCC)
return performSignExtendSetCCCombine(N, DCI, DAG);
// If we see (any_extend (bswap ...)) with bswap returning an i16, we know
// that the top half of the result register must be unused, due to the
// any_extend. This means that we can replace this pattern with (rev16
// (any_extend ...)). This saves a machine instruction compared to (lsr (rev
// ...)), which is what this pattern would otherwise be lowered to.
// Only apply this optimisation if any_extend in original pattern to i32 or
// i64, because this type will become the input type to REV16 in the new
// pattern, so must be a legitimate REV16 input type.
SDValue Bswap = N->getOperand(0);
if (N->getOpcode() == ISD::ANY_EXTEND && Bswap.getOpcode() == ISD::BSWAP &&
Bswap.getValueType() == MVT::i16 &&
(N->getValueType(0) == MVT::i32 || N->getValueType(0) == MVT::i64)) {
SDLoc DL(N);
SDValue NewAnyExtend = DAG.getNode(ISD::ANY_EXTEND, DL, N->getValueType(0),
Bswap->getOperand(0));
return DAG.getNode(AArch64ISD::REV16, SDLoc(N), N->getValueType(0),
NewAnyExtend);
}
return SDValue();
}
static SDValue splitStoreSplat(SelectionDAG &DAG, StoreSDNode &St,
SDValue SplatVal, unsigned NumVecElts) {
assert(!St.isTruncatingStore() && "cannot split truncating vector store");
Align OrigAlignment = St.getAlign();
unsigned EltOffset = SplatVal.getValueType().getSizeInBits() / 8;
// Create scalar stores. This is at least as good as the code sequence for a
// split unaligned store which is a dup.s, ext.b, and two stores.
// Most of the time the three stores should be replaced by store pair
// instructions (stp).
SDLoc DL(&St);
SDValue BasePtr = St.getBasePtr();
uint64_t BaseOffset = 0;
const MachinePointerInfo &PtrInfo = St.getPointerInfo();
SDValue NewST1 =
DAG.getStore(St.getChain(), DL, SplatVal, BasePtr, PtrInfo,
OrigAlignment, St.getMemOperand()->getFlags());
// As this in ISel, we will not merge this add which may degrade results.
if (BasePtr->getOpcode() == ISD::ADD &&
isa<ConstantSDNode>(BasePtr->getOperand(1))) {
BaseOffset = cast<ConstantSDNode>(BasePtr->getOperand(1))->getSExtValue();
BasePtr = BasePtr->getOperand(0);
}
unsigned Offset = EltOffset;
while (--NumVecElts) {
Align Alignment = commonAlignment(OrigAlignment, Offset);
SDValue OffsetPtr =
DAG.getNode(ISD::ADD, DL, MVT::i64, BasePtr,
DAG.getConstant(BaseOffset + Offset, DL, MVT::i64));
NewST1 = DAG.getStore(NewST1.getValue(0), DL, SplatVal, OffsetPtr,
PtrInfo.getWithOffset(Offset), Alignment,
St.getMemOperand()->getFlags());
Offset += EltOffset;
}
return NewST1;
}
// Returns an SVE type that ContentTy can be trivially sign or zero extended
// into.
static MVT getSVEContainerType(EVT ContentTy) {
assert(ContentTy.isSimple() && "No SVE containers for extended types");
switch (ContentTy.getSimpleVT().SimpleTy) {
default:
llvm_unreachable("No known SVE container for this MVT type");
case MVT::nxv2i8:
case MVT::nxv2i16:
case MVT::nxv2i32:
case MVT::nxv2i64:
case MVT::nxv2f32:
case MVT::nxv2f64:
return MVT::nxv2i64;
case MVT::nxv4i8:
case MVT::nxv4i16:
case MVT::nxv4i32:
case MVT::nxv4f32:
return MVT::nxv4i32;
case MVT::nxv8i8:
case MVT::nxv8i16:
case MVT::nxv8f16:
case MVT::nxv8bf16:
return MVT::nxv8i16;
case MVT::nxv16i8:
return MVT::nxv16i8;
}
}
static SDValue performLD1Combine(SDNode *N, SelectionDAG &DAG, unsigned Opc) {
SDLoc DL(N);
EVT VT = N->getValueType(0);
if (VT.getSizeInBits().getKnownMinValue() > AArch64::SVEBitsPerBlock)
return SDValue();
EVT ContainerVT = VT;
if (ContainerVT.isInteger())
ContainerVT = getSVEContainerType(ContainerVT);
SDVTList VTs = DAG.getVTList(ContainerVT, MVT::Other);
SDValue Ops[] = { N->getOperand(0), // Chain
N->getOperand(2), // Pg
N->getOperand(3), // Base
DAG.getValueType(VT) };
SDValue Load = DAG.getNode(Opc, DL, VTs, Ops);
SDValue LoadChain = SDValue(Load.getNode(), 1);
if (ContainerVT.isInteger() && (VT != ContainerVT))
Load = DAG.getNode(ISD::TRUNCATE, DL, VT, Load.getValue(0));
return DAG.getMergeValues({ Load, LoadChain }, DL);
}
static SDValue performLDNT1Combine(SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
EVT VT = N->getValueType(0);
EVT PtrTy = N->getOperand(3).getValueType();
EVT LoadVT = VT;
if (VT.isFloatingPoint())
LoadVT = VT.changeTypeToInteger();
auto *MINode = cast<MemIntrinsicSDNode>(N);
SDValue PassThru = DAG.getConstant(0, DL, LoadVT);
SDValue L = DAG.getMaskedLoad(LoadVT, DL, MINode->getChain(),
MINode->getOperand(3), DAG.getUNDEF(PtrTy),
MINode->getOperand(2), PassThru,
MINode->getMemoryVT(), MINode->getMemOperand(),
ISD::UNINDEXED, ISD::NON_EXTLOAD, false);
if (VT.isFloatingPoint()) {
SDValue Ops[] = { DAG.getNode(ISD::BITCAST, DL, VT, L), L.getValue(1) };
return DAG.getMergeValues(Ops, DL);
}
return L;
}
template <unsigned Opcode>
static SDValue performLD1ReplicateCombine(SDNode *N, SelectionDAG &DAG) {
static_assert(Opcode == AArch64ISD::LD1RQ_MERGE_ZERO ||
Opcode == AArch64ISD::LD1RO_MERGE_ZERO,
"Unsupported opcode.");
SDLoc DL(N);
EVT VT = N->getValueType(0);
EVT LoadVT = VT;
if (VT.isFloatingPoint())
LoadVT = VT.changeTypeToInteger();
SDValue Ops[] = {N->getOperand(0), N->getOperand(2), N->getOperand(3)};
SDValue Load = DAG.getNode(Opcode, DL, {LoadVT, MVT::Other}, Ops);
SDValue LoadChain = SDValue(Load.getNode(), 1);
if (VT.isFloatingPoint())
Load = DAG.getNode(ISD::BITCAST, DL, VT, Load.getValue(0));
return DAG.getMergeValues({Load, LoadChain}, DL);
}
static SDValue performST1Combine(SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Data = N->getOperand(2);
EVT DataVT = Data.getValueType();
EVT HwSrcVt = getSVEContainerType(DataVT);
SDValue InputVT = DAG.getValueType(DataVT);
if (DataVT.isFloatingPoint())
InputVT = DAG.getValueType(HwSrcVt);
SDValue SrcNew;
if (Data.getValueType().isFloatingPoint())
SrcNew = DAG.getNode(ISD::BITCAST, DL, HwSrcVt, Data);
else
SrcNew = DAG.getNode(ISD::ANY_EXTEND, DL, HwSrcVt, Data);
SDValue Ops[] = { N->getOperand(0), // Chain
SrcNew,
N->getOperand(4), // Base
N->getOperand(3), // Pg
InputVT
};
return DAG.getNode(AArch64ISD::ST1_PRED, DL, N->getValueType(0), Ops);
}
static SDValue performSTNT1Combine(SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Data = N->getOperand(2);
EVT DataVT = Data.getValueType();
EVT PtrTy = N->getOperand(4).getValueType();
if (DataVT.isFloatingPoint())
Data = DAG.getNode(ISD::BITCAST, DL, DataVT.changeTypeToInteger(), Data);
auto *MINode = cast<MemIntrinsicSDNode>(N);
return DAG.getMaskedStore(MINode->getChain(), DL, Data, MINode->getOperand(4),
DAG.getUNDEF(PtrTy), MINode->getOperand(3),
MINode->getMemoryVT(), MINode->getMemOperand(),
ISD::UNINDEXED, false, false);
}
/// Replace a splat of zeros to a vector store by scalar stores of WZR/XZR. The
/// load store optimizer pass will merge them to store pair stores. This should
/// be better than a movi to create the vector zero followed by a vector store
/// if the zero constant is not re-used, since one instructions and one register
/// live range will be removed.
///
/// For example, the final generated code should be:
///
/// stp xzr, xzr, [x0]
///
/// instead of:
///
/// movi v0.2d, #0
/// str q0, [x0]
///
static SDValue replaceZeroVectorStore(SelectionDAG &DAG, StoreSDNode &St) {
SDValue StVal = St.getValue();
EVT VT = StVal.getValueType();
// Avoid scalarizing zero splat stores for scalable vectors.
if (VT.isScalableVector())
return SDValue();
// It is beneficial to scalarize a zero splat store for 2 or 3 i64 elements or
// 2, 3 or 4 i32 elements.
int NumVecElts = VT.getVectorNumElements();
if (!(((NumVecElts == 2 || NumVecElts == 3) &&
VT.getVectorElementType().getSizeInBits() == 64) ||
((NumVecElts == 2 || NumVecElts == 3 || NumVecElts == 4) &&
VT.getVectorElementType().getSizeInBits() == 32)))
return SDValue();
if (StVal.getOpcode() != ISD::BUILD_VECTOR)
return SDValue();
// If the zero constant has more than one use then the vector store could be
// better since the constant mov will be amortized and stp q instructions
// should be able to be formed.
if (!StVal.hasOneUse())
return SDValue();
// If the store is truncating then it's going down to i16 or smaller, which
// means it can be implemented in a single store anyway.
if (St.isTruncatingStore())
return SDValue();
// If the immediate offset of the address operand is too large for the stp
// instruction, then bail out.
if (DAG.isBaseWithConstantOffset(St.getBasePtr())) {
int64_t Offset = St.getBasePtr()->getConstantOperandVal(1);
if (Offset < -512 || Offset > 504)
return SDValue();
}
for (int I = 0; I < NumVecElts; ++I) {
SDValue EltVal = StVal.getOperand(I);
if (!isNullConstant(EltVal) && !isNullFPConstant(EltVal))
return SDValue();
}
// Use a CopyFromReg WZR/XZR here to prevent
// DAGCombiner::MergeConsecutiveStores from undoing this transformation.
SDLoc DL(&St);
unsigned ZeroReg;
EVT ZeroVT;
if (VT.getVectorElementType().getSizeInBits() == 32) {
ZeroReg = AArch64::WZR;
ZeroVT = MVT::i32;
} else {
ZeroReg = AArch64::XZR;
ZeroVT = MVT::i64;
}
SDValue SplatVal =
DAG.getCopyFromReg(DAG.getEntryNode(), DL, ZeroReg, ZeroVT);
return splitStoreSplat(DAG, St, SplatVal, NumVecElts);
}
/// Replace a splat of a scalar to a vector store by scalar stores of the scalar
/// value. The load store optimizer pass will merge them to store pair stores.
/// This has better performance than a splat of the scalar followed by a split
/// vector store. Even if the stores are not merged it is four stores vs a dup,
/// followed by an ext.b and two stores.
static SDValue replaceSplatVectorStore(SelectionDAG &DAG, StoreSDNode &St) {
SDValue StVal = St.getValue();
EVT VT = StVal.getValueType();
// Don't replace floating point stores, they possibly won't be transformed to
// stp because of the store pair suppress pass.
if (VT.isFloatingPoint())
return SDValue();
// We can express a splat as store pair(s) for 2 or 4 elements.
unsigned NumVecElts = VT.getVectorNumElements();
if (NumVecElts != 4 && NumVecElts != 2)
return SDValue();
// If the store is truncating then it's going down to i16 or smaller, which
// means it can be implemented in a single store anyway.
if (St.isTruncatingStore())
return SDValue();
// Check that this is a splat.
// Make sure that each of the relevant vector element locations are inserted
// to, i.e. 0 and 1 for v2i64 and 0, 1, 2, 3 for v4i32.
std::bitset<4> IndexNotInserted((1 << NumVecElts) - 1);
SDValue SplatVal;
for (unsigned I = 0; I < NumVecElts; ++I) {
// Check for insert vector elements.
if (StVal.getOpcode() != ISD::INSERT_VECTOR_ELT)
return SDValue();
// Check that same value is inserted at each vector element.
if (I == 0)
SplatVal = StVal.getOperand(1);
else if (StVal.getOperand(1) != SplatVal)
return SDValue();
// Check insert element index.
ConstantSDNode *CIndex = dyn_cast<ConstantSDNode>(StVal.getOperand(2));
if (!CIndex)
return SDValue();
uint64_t IndexVal = CIndex->getZExtValue();
if (IndexVal >= NumVecElts)
return SDValue();
IndexNotInserted.reset(IndexVal);
StVal = StVal.getOperand(0);
}
// Check that all vector element locations were inserted to.
if (IndexNotInserted.any())
return SDValue();
return splitStoreSplat(DAG, St, SplatVal, NumVecElts);
}
static SDValue splitStores(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
StoreSDNode *S = cast<StoreSDNode>(N);
if (S->isVolatile() || S->isIndexed())
return SDValue();
SDValue StVal = S->getValue();
EVT VT = StVal.getValueType();
if (!VT.isFixedLengthVector())
return SDValue();
// If we get a splat of zeros, convert this vector store to a store of
// scalars. They will be merged into store pairs of xzr thereby removing one
// instruction and one register.
if (SDValue ReplacedZeroSplat = replaceZeroVectorStore(DAG, *S))
return ReplacedZeroSplat;
// FIXME: The logic for deciding if an unaligned store should be split should
// be included in TLI.allowsMisalignedMemoryAccesses(), and there should be
// a call to that function here.
if (!Subtarget->isMisaligned128StoreSlow())
return SDValue();
// Don't split at -Oz.
if (DAG.getMachineFunction().getFunction().hasMinSize())
return SDValue();
// Don't split v2i64 vectors. Memcpy lowering produces those and splitting
// those up regresses performance on micro-benchmarks and olden/bh.
if (VT.getVectorNumElements() < 2 || VT == MVT::v2i64)
return SDValue();
// Split unaligned 16B stores. They are terrible for performance.
// Don't split stores with alignment of 1 or 2. Code that uses clang vector
// extensions can use this to mark that it does not want splitting to happen
// (by underspecifying alignment to be 1 or 2). Furthermore, the chance of
// eliminating alignment hazards is only 1 in 8 for alignment of 2.
if (VT.getSizeInBits() != 128 || S->getAlign() >= Align(16) ||
S->getAlign() <= Align(2))
return SDValue();
// If we get a splat of a scalar convert this vector store to a store of
// scalars. They will be merged into store pairs thereby removing two
// instructions.
if (SDValue ReplacedSplat = replaceSplatVectorStore(DAG, *S))
return ReplacedSplat;
SDLoc DL(S);
// Split VT into two.
EVT HalfVT = VT.getHalfNumVectorElementsVT(*DAG.getContext());
unsigned NumElts = HalfVT.getVectorNumElements();
SDValue SubVector0 = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, HalfVT, StVal,
DAG.getConstant(0, DL, MVT::i64));
SDValue SubVector1 = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, HalfVT, StVal,
DAG.getConstant(NumElts, DL, MVT::i64));
SDValue BasePtr = S->getBasePtr();
SDValue NewST1 =
DAG.getStore(S->getChain(), DL, SubVector0, BasePtr, S->getPointerInfo(),
S->getAlign(), S->getMemOperand()->getFlags());
SDValue OffsetPtr = DAG.getNode(ISD::ADD, DL, MVT::i64, BasePtr,
DAG.getConstant(8, DL, MVT::i64));
return DAG.getStore(NewST1.getValue(0), DL, SubVector1, OffsetPtr,
S->getPointerInfo(), S->getAlign(),
S->getMemOperand()->getFlags());
}
static SDValue performSpliceCombine(SDNode *N, SelectionDAG &DAG) {
assert(N->getOpcode() == AArch64ISD::SPLICE && "Unexpected Opcode!");
// splice(pg, op1, undef) -> op1
if (N->getOperand(2).isUndef())
return N->getOperand(1);
return SDValue();
}
static SDValue performUnpackCombine(SDNode *N, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
assert((N->getOpcode() == AArch64ISD::UUNPKHI ||
N->getOpcode() == AArch64ISD::UUNPKLO) &&
"Unexpected Opcode!");
// uunpklo/hi undef -> undef
if (N->getOperand(0).isUndef())
return DAG.getUNDEF(N->getValueType(0));
// If this is a masked load followed by an UUNPKLO, fold this into a masked
// extending load. We can do this even if this is already a masked
// {z,}extload.
if (N->getOperand(0).getOpcode() == ISD::MLOAD &&
N->getOpcode() == AArch64ISD::UUNPKLO) {
MaskedLoadSDNode *MLD = cast<MaskedLoadSDNode>(N->getOperand(0));
SDValue Mask = MLD->getMask();
SDLoc DL(N);
if (MLD->isUnindexed() && MLD->getExtensionType() != ISD::SEXTLOAD &&
SDValue(MLD, 0).hasOneUse() && Mask->getOpcode() == AArch64ISD::PTRUE &&
(MLD->getPassThru()->isUndef() ||
isZerosVector(MLD->getPassThru().getNode()))) {
unsigned MinSVESize = Subtarget->getMinSVEVectorSizeInBits();
unsigned PgPattern = Mask->getConstantOperandVal(0);
EVT VT = N->getValueType(0);
// Ensure we can double the size of the predicate pattern
unsigned NumElts = getNumElementsFromSVEPredPattern(PgPattern);
if (NumElts &&
NumElts * VT.getVectorElementType().getSizeInBits() <= MinSVESize) {
Mask =
getPTrue(DAG, DL, VT.changeVectorElementType(MVT::i1), PgPattern);
SDValue PassThru = DAG.getConstant(0, DL, VT);
SDValue NewLoad = DAG.getMaskedLoad(
VT, DL, MLD->getChain(), MLD->getBasePtr(), MLD->getOffset(), Mask,
PassThru, MLD->getMemoryVT(), MLD->getMemOperand(),
MLD->getAddressingMode(), ISD::ZEXTLOAD);
DAG.ReplaceAllUsesOfValueWith(SDValue(MLD, 1), NewLoad.getValue(1));
return NewLoad;
}
}
}
return SDValue();
}
static bool isHalvingTruncateAndConcatOfLegalIntScalableType(SDNode *N) {
if (N->getOpcode() != AArch64ISD::UZP1)
return false;
SDValue Op0 = N->getOperand(0);
EVT SrcVT = Op0->getValueType(0);
EVT DstVT = N->getValueType(0);
return (SrcVT == MVT::nxv8i16 && DstVT == MVT::nxv16i8) ||
(SrcVT == MVT::nxv4i32 && DstVT == MVT::nxv8i16) ||
(SrcVT == MVT::nxv2i64 && DstVT == MVT::nxv4i32);
}
// Try to combine rounding shifts where the operands come from an extend, and
// the result is truncated and combined into one vector.
// uzp1(rshrnb(uunpklo(X),C), rshrnb(uunpkhi(X), C)) -> urshr(X, C)
static SDValue tryCombineExtendRShTrunc(SDNode *N, SelectionDAG &DAG) {
assert(N->getOpcode() == AArch64ISD::UZP1 && "Only UZP1 expected.");
SDValue Op0 = N->getOperand(0);
SDValue Op1 = N->getOperand(1);
EVT ResVT = N->getValueType(0);
unsigned RshOpc = Op0.getOpcode();
if (RshOpc != AArch64ISD::RSHRNB_I)
return SDValue();
// Same op code and imm value?
SDValue ShiftValue = Op0.getOperand(1);
if (RshOpc != Op1.getOpcode() || ShiftValue != Op1.getOperand(1))
return SDValue();
// Same unextended operand value?
SDValue Lo = Op0.getOperand(0);
SDValue Hi = Op1.getOperand(0);
if (Lo.getOpcode() != AArch64ISD::UUNPKLO &&
Hi.getOpcode() != AArch64ISD::UUNPKHI)
return SDValue();
SDValue OrigArg = Lo.getOperand(0);
if (OrigArg != Hi.getOperand(0))
return SDValue();
SDLoc DL(N);
return DAG.getNode(AArch64ISD::URSHR_I_PRED, DL, ResVT,
getPredicateForVector(DAG, DL, ResVT), OrigArg,
ShiftValue);
}
// Try to simplify:
// t1 = nxv8i16 add(X, 1 << (ShiftValue - 1))
// t2 = nxv8i16 srl(t1, ShiftValue)
// to
// t1 = nxv8i16 rshrnb(X, shiftvalue).
// rshrnb will zero the top half bits of each element. Therefore, this combine
// should only be performed when a following instruction with the rshrnb
// as an operand does not care about the top half of each element. For example,
// a uzp1 or a truncating store.
static SDValue trySimplifySrlAddToRshrnb(SDValue Srl, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
EVT VT = Srl->getValueType(0);
if (!VT.isScalableVector() || !Subtarget->hasSVE2())
return SDValue();
EVT ResVT;
if (VT == MVT::nxv8i16)
ResVT = MVT::nxv16i8;
else if (VT == MVT::nxv4i32)
ResVT = MVT::nxv8i16;
else if (VT == MVT::nxv2i64)
ResVT = MVT::nxv4i32;
else
return SDValue();
SDLoc DL(Srl);
unsigned ShiftValue;
SDValue RShOperand;
if (!canLowerSRLToRoundingShiftForVT(Srl, ResVT, DAG, ShiftValue, RShOperand))
return SDValue();
SDValue Rshrnb = DAG.getNode(
AArch64ISD::RSHRNB_I, DL, ResVT,
{RShOperand, DAG.getTargetConstant(ShiftValue, DL, MVT::i32)});
return DAG.getNode(AArch64ISD::NVCAST, DL, VT, Rshrnb);
}
static SDValue isNVCastToHalfWidthElements(SDValue V) {
if (V.getOpcode() != AArch64ISD::NVCAST)
return SDValue();
SDValue Op = V.getOperand(0);
if (!Op.getValueType().isVector() ||
V.getValueType().getVectorElementCount() !=
Op.getValueType().getVectorElementCount() * 2)
return SDValue();
return Op;
}
static SDValue performUzpCombine(SDNode *N, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
SDLoc DL(N);
SDValue Op0 = N->getOperand(0);
SDValue Op1 = N->getOperand(1);
EVT ResVT = N->getValueType(0);
// uzp(extract_lo(x), extract_hi(x)) -> extract_lo(uzp x, x)
if (Op0.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
Op1.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
Op0.getOperand(0) == Op1.getOperand(0)) {
SDValue SourceVec = Op0.getOperand(0);
uint64_t ExtIdx0 = Op0.getConstantOperandVal(1);
uint64_t ExtIdx1 = Op1.getConstantOperandVal(1);
uint64_t NumElements = SourceVec.getValueType().getVectorMinNumElements();
if (ExtIdx0 == 0 && ExtIdx1 == NumElements / 2) {
EVT OpVT = Op0.getOperand(1).getValueType();
EVT WidenedResVT = ResVT.getDoubleNumVectorElementsVT(*DAG.getContext());
SDValue Uzp = DAG.getNode(N->getOpcode(), DL, WidenedResVT, SourceVec,
DAG.getUNDEF(WidenedResVT));
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, ResVT, Uzp,
DAG.getConstant(0, DL, OpVT));
}
}
// Following optimizations only work with uzp1.
if (N->getOpcode() == AArch64ISD::UZP2)
return SDValue();
// uzp1(x, undef) -> concat(truncate(x), undef)
if (Op1.getOpcode() == ISD::UNDEF) {
EVT BCVT = MVT::Other, HalfVT = MVT::Other;
switch (ResVT.getSimpleVT().SimpleTy) {
default:
break;
case MVT::v16i8:
BCVT = MVT::v8i16;
HalfVT = MVT::v8i8;
break;
case MVT::v8i16:
BCVT = MVT::v4i32;
HalfVT = MVT::v4i16;
break;
case MVT::v4i32:
BCVT = MVT::v2i64;
HalfVT = MVT::v2i32;
break;
}
if (BCVT != MVT::Other) {
SDValue BC = DAG.getBitcast(BCVT, Op0);
SDValue Trunc = DAG.getNode(ISD::TRUNCATE, DL, HalfVT, BC);
return DAG.getNode(ISD::CONCAT_VECTORS, DL, ResVT, Trunc,
DAG.getUNDEF(HalfVT));
}
}
if (SDValue Urshr = tryCombineExtendRShTrunc(N, DAG))
return Urshr;
if (SDValue PreCast = isNVCastToHalfWidthElements(Op0)) {
if (SDValue Rshrnb = trySimplifySrlAddToRshrnb(PreCast, DAG, Subtarget)) {
Rshrnb = DAG.getNode(AArch64ISD::NVCAST, DL, ResVT, Rshrnb);
return DAG.getNode(AArch64ISD::UZP1, DL, ResVT, Rshrnb, Op1);
}
}
if (SDValue PreCast = isNVCastToHalfWidthElements(Op1)) {
if (SDValue Rshrnb = trySimplifySrlAddToRshrnb(PreCast, DAG, Subtarget)) {
Rshrnb = DAG.getNode(AArch64ISD::NVCAST, DL, ResVT, Rshrnb);
return DAG.getNode(AArch64ISD::UZP1, DL, ResVT, Op0, Rshrnb);
}
}
// uzp1<ty>(nvcast(unpklo(uzp1<ty>(x, y))), z) => uzp1<ty>(x, z)
if (SDValue PreCast = isNVCastToHalfWidthElements(Op0)) {
if (PreCast.getOpcode() == AArch64ISD::UUNPKLO) {
if (PreCast.getOperand(0).getOpcode() == AArch64ISD::UZP1) {
SDValue X = PreCast.getOperand(0).getOperand(0);
return DAG.getNode(AArch64ISD::UZP1, DL, ResVT, X, Op1);
}
}
}
// uzp1<ty>(x, nvcast(unpkhi(uzp1<ty>(y, z)))) => uzp1<ty>(x, z)
if (SDValue PreCast = isNVCastToHalfWidthElements(Op1)) {
if (PreCast.getOpcode() == AArch64ISD::UUNPKHI) {
if (PreCast.getOperand(0).getOpcode() == AArch64ISD::UZP1) {
SDValue Z = PreCast.getOperand(0).getOperand(1);
return DAG.getNode(AArch64ISD::UZP1, DL, ResVT, Op0, Z);
}
}
}
// These optimizations only work on little endian.
if (!DAG.getDataLayout().isLittleEndian())
return SDValue();
// uzp1(bitcast(x), bitcast(y)) -> uzp1(x, y)
// Example:
// nxv4i32 = uzp1 bitcast(nxv4i32 x to nxv2i64), bitcast(nxv4i32 y to nxv2i64)
// to
// nxv4i32 = uzp1 nxv4i32 x, nxv4i32 y
if (isHalvingTruncateAndConcatOfLegalIntScalableType(N) &&
Op0.getOpcode() == ISD::BITCAST && Op1.getOpcode() == ISD::BITCAST) {
if (Op0.getOperand(0).getValueType() == Op1.getOperand(0).getValueType()) {
return DAG.getNode(AArch64ISD::UZP1, DL, ResVT, Op0.getOperand(0),
Op1.getOperand(0));
}
}
if (ResVT != MVT::v2i32 && ResVT != MVT::v4i16 && ResVT != MVT::v8i8)
return SDValue();
SDValue SourceOp0 = peekThroughBitcasts(Op0);
SDValue SourceOp1 = peekThroughBitcasts(Op1);
// truncating uzp1(x, y) -> xtn(concat (x, y))
if (SourceOp0.getValueType() == SourceOp1.getValueType()) {
EVT Op0Ty = SourceOp0.getValueType();
if ((ResVT == MVT::v4i16 && Op0Ty == MVT::v2i32) ||
(ResVT == MVT::v8i8 && Op0Ty == MVT::v4i16)) {
SDValue Concat =
DAG.getNode(ISD::CONCAT_VECTORS, DL,
Op0Ty.getDoubleNumVectorElementsVT(*DAG.getContext()),
SourceOp0, SourceOp1);
return DAG.getNode(ISD::TRUNCATE, DL, ResVT, Concat);
}
}
// uzp1(xtn x, xtn y) -> xtn(uzp1 (x, y))
if (SourceOp0.getOpcode() != ISD::TRUNCATE ||
SourceOp1.getOpcode() != ISD::TRUNCATE)
return SDValue();
SourceOp0 = SourceOp0.getOperand(0);
SourceOp1 = SourceOp1.getOperand(0);
if (SourceOp0.getValueType() != SourceOp1.getValueType() ||
!SourceOp0.getValueType().isSimple())
return SDValue();
EVT ResultTy;
switch (SourceOp0.getSimpleValueType().SimpleTy) {
case MVT::v2i64:
ResultTy = MVT::v4i32;
break;
case MVT::v4i32:
ResultTy = MVT::v8i16;
break;
case MVT::v8i16:
ResultTy = MVT::v16i8;
break;
default:
return SDValue();
}
SDValue UzpOp0 = DAG.getNode(ISD::BITCAST, DL, ResultTy, SourceOp0);
SDValue UzpOp1 = DAG.getNode(ISD::BITCAST, DL, ResultTy, SourceOp1);
SDValue UzpResult =
DAG.getNode(AArch64ISD::UZP1, DL, UzpOp0.getValueType(), UzpOp0, UzpOp1);
EVT BitcastResultTy;
switch (ResVT.getSimpleVT().SimpleTy) {
case MVT::v2i32:
BitcastResultTy = MVT::v2i64;
break;
case MVT::v4i16:
BitcastResultTy = MVT::v4i32;
break;
case MVT::v8i8:
BitcastResultTy = MVT::v8i16;
break;
default:
llvm_unreachable("Should be one of {v2i32, v4i16, v8i8}");
}
return DAG.getNode(ISD::TRUNCATE, DL, ResVT,
DAG.getNode(ISD::BITCAST, DL, BitcastResultTy, UzpResult));
}
static SDValue performGLD1Combine(SDNode *N, SelectionDAG &DAG) {
unsigned Opc = N->getOpcode();
const bool Scaled = Opc == AArch64ISD::GLD1_SCALED_MERGE_ZERO ||
Opc == AArch64ISD::GLD1S_SCALED_MERGE_ZERO;
const bool Signed = Opc == AArch64ISD::GLD1S_MERGE_ZERO ||
Opc == AArch64ISD::GLD1S_SCALED_MERGE_ZERO;
const bool Extended = Opc == AArch64ISD::GLD1_SXTW_MERGE_ZERO ||
Opc == AArch64ISD::GLD1_SXTW_SCALED_MERGE_ZERO ||
Opc == AArch64ISD::GLD1_UXTW_MERGE_ZERO ||
Opc == AArch64ISD::GLD1_UXTW_SCALED_MERGE_ZERO;
SDLoc DL(N);
SDValue Chain = N->getOperand(0);
SDValue Pg = N->getOperand(1);
SDValue Base = N->getOperand(2);
SDValue Offset = N->getOperand(3);
SDValue Ty = N->getOperand(4);
EVT ResVT = N->getValueType(0);
const auto OffsetOpc = Offset.getOpcode();
const bool OffsetIsZExt =
OffsetOpc == AArch64ISD::ZERO_EXTEND_INREG_MERGE_PASSTHRU;
const bool OffsetIsSExt =
OffsetOpc == AArch64ISD::SIGN_EXTEND_INREG_MERGE_PASSTHRU;
// Fold sign/zero extensions of vector offsets into GLD1 nodes where possible.
if (!Extended && (OffsetIsSExt || OffsetIsZExt)) {
SDValue ExtPg = Offset.getOperand(0);
VTSDNode *ExtFrom = cast<VTSDNode>(Offset.getOperand(2).getNode());
EVT ExtFromEVT = ExtFrom->getVT().getVectorElementType();
// If the predicate for the sign- or zero-extended offset is the
// same as the predicate used for this load and the sign-/zero-extension
// was from a 32-bits...
if (ExtPg == Pg && ExtFromEVT == MVT::i32) {
SDValue UnextendedOffset = Offset.getOperand(1);
unsigned NewOpc = getGatherVecOpcode(Scaled, OffsetIsSExt, true);
if (Signed)
NewOpc = getSignExtendedGatherOpcode(NewOpc);
return DAG.getNode(NewOpc, DL, {ResVT, MVT::Other},
{Chain, Pg, Base, UnextendedOffset, Ty});
}
}
return SDValue();
}
/// Optimize a vector shift instruction and its operand if shifted out
/// bits are not used.
static SDValue performVectorShiftCombine(SDNode *N,
const AArch64TargetLowering &TLI,
TargetLowering::DAGCombinerInfo &DCI) {
assert(N->getOpcode() == AArch64ISD::VASHR ||
N->getOpcode() == AArch64ISD::VLSHR);
SDValue Op = N->getOperand(0);
unsigned OpScalarSize = Op.getScalarValueSizeInBits();
unsigned ShiftImm = N->getConstantOperandVal(1);
assert(OpScalarSize > ShiftImm && "Invalid shift imm");
// Remove sign_extend_inreg (ashr(shl(x)) based on the number of sign bits.
if (N->getOpcode() == AArch64ISD::VASHR &&
Op.getOpcode() == AArch64ISD::VSHL &&
N->getOperand(1) == Op.getOperand(1))
if (DCI.DAG.ComputeNumSignBits(Op.getOperand(0)) > ShiftImm)
return Op.getOperand(0);
// If the shift is exact, the shifted out bits matter.
if (N->getFlags().hasExact())
return SDValue();
APInt ShiftedOutBits = APInt::getLowBitsSet(OpScalarSize, ShiftImm);
APInt DemandedMask = ~ShiftedOutBits;
if (TLI.SimplifyDemandedBits(Op, DemandedMask, DCI))
return SDValue(N, 0);
return SDValue();
}
static SDValue performSunpkloCombine(SDNode *N, SelectionDAG &DAG) {
// sunpklo(sext(pred)) -> sext(extract_low_half(pred))
// This transform works in partnership with performSetCCPunpkCombine to
// remove unnecessary transfer of predicates into standard registers and back
if (N->getOperand(0).getOpcode() == ISD::SIGN_EXTEND &&
N->getOperand(0)->getOperand(0)->getValueType(0).getScalarType() ==
MVT::i1) {
SDValue CC = N->getOperand(0)->getOperand(0);
auto VT = CC->getValueType(0).getHalfNumVectorElementsVT(*DAG.getContext());
SDValue Unpk = DAG.getNode(ISD::EXTRACT_SUBVECTOR, SDLoc(N), VT, CC,
DAG.getVectorIdxConstant(0, SDLoc(N)));
return DAG.getNode(ISD::SIGN_EXTEND, SDLoc(N), N->getValueType(0), Unpk);
}
return SDValue();
}
/// Target-specific DAG combine function for post-increment LD1 (lane) and
/// post-increment LD1R.
static SDValue performPostLD1Combine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
bool IsLaneOp) {
if (DCI.isBeforeLegalizeOps())
return SDValue();
SelectionDAG &DAG = DCI.DAG;
EVT VT = N->getValueType(0);
if (!VT.is128BitVector() && !VT.is64BitVector())
return SDValue();
// If it is not LOAD, can not do such combine.
unsigned LoadIdx = IsLaneOp ? 1 : 0;
LoadSDNode *LD = dyn_cast<LoadSDNode>(N->getOperand(LoadIdx).getNode());
if (!LD)
return SDValue();
// If the Generic combiner already helped form a pre- or post-indexed load,
// skip forming one here.
if (LD->isIndexed())
return SDValue();
// The vector lane must be a constant in the LD1LANE opcode.
SDValue Lane;
if (IsLaneOp) {
Lane = N->getOperand(2);
auto *LaneC = dyn_cast<ConstantSDNode>(Lane);
if (!LaneC || LaneC->getZExtValue() >= VT.getVectorNumElements())
return SDValue();
if (LaneC->getZExtValue() == 0 && isNullOrNullSplat(N->getOperand(0)))
return SDValue();
}
LoadSDNode *LoadSDN = cast<LoadSDNode>(LD);
EVT MemVT = LoadSDN->getMemoryVT();
// Check if memory operand is the same type as the vector element.
if (MemVT != VT.getVectorElementType())
return SDValue();
// Check if there are other uses. If so, do not combine as it will introduce
// an extra load.
for (SDUse &U : LD->uses()) {
if (U.getResNo() == 1) // Ignore uses of the chain result.
continue;
if (U.getUser() != N)
return SDValue();
}
// If there is one use and it can splat the value, prefer that operation.
// TODO: This could be expanded to more operations if they reliably use the
// index variants.
if (N->hasOneUse()) {
unsigned UseOpc = N->user_begin()->getOpcode();
if (UseOpc == ISD::FMUL || UseOpc == ISD::FMA)
return SDValue();
}
SDValue Addr = LD->getOperand(1);
SDValue Vector = N->getOperand(0);
// Search for a use of the address operand that is an increment.
for (SDUse &Use : Addr->uses()) {
SDNode *User = Use.getUser();
if (User->getOpcode() != ISD::ADD || Use.getResNo() != Addr.getResNo())
continue;
// If the increment is a constant, it must match the memory ref size.
SDValue Inc = User->getOperand(User->getOperand(0) == Addr ? 1 : 0);
if (ConstantSDNode *CInc = dyn_cast<ConstantSDNode>(Inc.getNode())) {
uint32_t IncVal = CInc->getZExtValue();
unsigned NumBytes = VT.getScalarSizeInBits() / 8;
if (IncVal != NumBytes)
continue;
Inc = DAG.getRegister(AArch64::XZR, MVT::i64);
}
// To avoid cycle construction make sure that neither the load nor the add
// are predecessors to each other or the Vector.
SmallPtrSet<const SDNode *, 32> Visited;
SmallVector<const SDNode *, 16> Worklist;
Visited.insert(Addr.getNode());
Worklist.push_back(User);
Worklist.push_back(LD);
Worklist.push_back(Vector.getNode());
if (SDNode::hasPredecessorHelper(LD, Visited, Worklist) ||
SDNode::hasPredecessorHelper(User, Visited, Worklist))
continue;
SmallVector<SDValue, 8> Ops;
Ops.push_back(LD->getOperand(0)); // Chain
if (IsLaneOp) {
Ops.push_back(Vector); // The vector to be inserted
Ops.push_back(Lane); // The lane to be inserted in the vector
}
Ops.push_back(Addr);
Ops.push_back(Inc);
EVT Tys[3] = { VT, MVT::i64, MVT::Other };
SDVTList SDTys = DAG.getVTList(Tys);
unsigned NewOp = IsLaneOp ? AArch64ISD::LD1LANEpost : AArch64ISD::LD1DUPpost;
SDValue UpdN = DAG.getMemIntrinsicNode(NewOp, SDLoc(N), SDTys, Ops,
MemVT,
LoadSDN->getMemOperand());
// Update the uses.
SDValue NewResults[] = {
SDValue(LD, 0), // The result of load
SDValue(UpdN.getNode(), 2) // Chain
};
DCI.CombineTo(LD, NewResults);
DCI.CombineTo(N, SDValue(UpdN.getNode(), 0)); // Dup/Inserted Result
DCI.CombineTo(User, SDValue(UpdN.getNode(), 1)); // Write back register
break;
}
return SDValue();
}
/// Simplify ``Addr`` given that the top byte of it is ignored by HW during
/// address translation.
static bool performTBISimplification(SDValue Addr,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
APInt DemandedMask = APInt::getLowBitsSet(64, 56);
KnownBits Known;
TargetLowering::TargetLoweringOpt TLO(DAG, !DCI.isBeforeLegalize(),
!DCI.isBeforeLegalizeOps());
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
if (TLI.SimplifyDemandedBits(Addr, DemandedMask, Known, TLO)) {
DCI.CommitTargetLoweringOpt(TLO);
return true;
}
return false;
}
static SDValue foldTruncStoreOfExt(SelectionDAG &DAG, SDNode *N) {
assert((N->getOpcode() == ISD::STORE || N->getOpcode() == ISD::MSTORE) &&
"Expected STORE dag node in input!");
if (auto Store = dyn_cast<StoreSDNode>(N)) {
if (!Store->isTruncatingStore() || Store->isIndexed())
return SDValue();
SDValue Ext = Store->getValue();
auto ExtOpCode = Ext.getOpcode();
if (ExtOpCode != ISD::ZERO_EXTEND && ExtOpCode != ISD::SIGN_EXTEND &&
ExtOpCode != ISD::ANY_EXTEND)
return SDValue();
SDValue Orig = Ext->getOperand(0);
if (Store->getMemoryVT() != Orig.getValueType())
return SDValue();
return DAG.getStore(Store->getChain(), SDLoc(Store), Orig,
Store->getBasePtr(), Store->getMemOperand());
}
return SDValue();
}
// A custom combine to lower load <3 x i8> as the more efficient sequence
// below:
// ldrb wX, [x0, #2]
// ldrh wY, [x0]
// orr wX, wY, wX, lsl #16
// fmov s0, wX
//
// Note that an alternative sequence with even fewer (although usually more
// complex/expensive) instructions would be:
// ld1r.4h { v0 }, [x0], #2
// ld1.b { v0 }[2], [x0]
//
// Generating this sequence unfortunately results in noticeably worse codegen
// for code that extends the loaded v3i8, due to legalization breaking vector
// shuffle detection in a way that is very difficult to work around.
// TODO: Revisit once v3i8 legalization has been improved in general.
static SDValue combineV3I8LoadExt(LoadSDNode *LD, SelectionDAG &DAG) {
EVT MemVT = LD->getMemoryVT();
if (MemVT != EVT::getVectorVT(*DAG.getContext(), MVT::i8, 3) ||
LD->getBaseAlign() >= 4)
return SDValue();
SDLoc DL(LD);
MachineFunction &MF = DAG.getMachineFunction();
SDValue Chain = LD->getChain();
SDValue BasePtr = LD->getBasePtr();
MachineMemOperand *MMO = LD->getMemOperand();
assert(LD->getOffset().isUndef() && "undef offset expected");
// Load 2 x i8, then 1 x i8.
SDValue L16 = DAG.getLoad(MVT::i16, DL, Chain, BasePtr, MMO);
TypeSize Offset2 = TypeSize::getFixed(2);
SDValue L8 = DAG.getLoad(MVT::i8, DL, Chain,
DAG.getMemBasePlusOffset(BasePtr, Offset2, DL),
MF.getMachineMemOperand(MMO, 2, 1));
// Extend to i32.
SDValue Ext16 = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i32, L16);
SDValue Ext8 = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i32, L8);
// Pack 2 x i8 and 1 x i8 in an i32 and convert to v4i8.
SDValue Shl = DAG.getNode(ISD::SHL, DL, MVT::i32, Ext8,
DAG.getConstant(16, DL, MVT::i32));
SDValue Or = DAG.getNode(ISD::OR, DL, MVT::i32, Ext16, Shl);
SDValue Cast = DAG.getNode(ISD::BITCAST, DL, MVT::v4i8, Or);
// Extract v3i8 again.
SDValue Extract = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MemVT, Cast,
DAG.getConstant(0, DL, MVT::i64));
SDValue TokenFactor = DAG.getNode(
ISD::TokenFactor, DL, MVT::Other,
{SDValue(cast<SDNode>(L16), 1), SDValue(cast<SDNode>(L8), 1)});
return DAG.getMergeValues({Extract, TokenFactor}, DL);
}
// Perform TBI simplification if supported by the target and try to break up
// nontemporal loads larger than 256-bits loads for odd types so LDNPQ 256-bit
// load instructions can be selected.
static SDValue performLOADCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
if (Subtarget->supportsAddressTopByteIgnored())
performTBISimplification(N->getOperand(1), DCI, DAG);
LoadSDNode *LD = cast<LoadSDNode>(N);
EVT RegVT = LD->getValueType(0);
EVT MemVT = LD->getMemoryVT();
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
SDLoc DL(LD);
// Cast ptr32 and ptr64 pointers to the default address space before a load.
unsigned AddrSpace = LD->getAddressSpace();
if (AddrSpace == ARM64AS::PTR64 || AddrSpace == ARM64AS::PTR32_SPTR ||
AddrSpace == ARM64AS::PTR32_UPTR) {
MVT PtrVT = TLI.getPointerTy(DAG.getDataLayout());
if (PtrVT != LD->getBasePtr().getSimpleValueType()) {
SDValue Cast =
DAG.getAddrSpaceCast(DL, PtrVT, LD->getBasePtr(), AddrSpace, 0);
return DAG.getExtLoad(LD->getExtensionType(), DL, RegVT, LD->getChain(),
Cast, LD->getPointerInfo(), MemVT,
LD->getBaseAlign(),
LD->getMemOperand()->getFlags());
}
}
if (LD->isVolatile() || !Subtarget->isLittleEndian())
return SDValue(N, 0);
if (SDValue Res = combineV3I8LoadExt(LD, DAG))
return Res;
if (!LD->isNonTemporal())
return SDValue(N, 0);
if (MemVT.isScalableVector() || MemVT.getSizeInBits() <= 256 ||
MemVT.getSizeInBits() % 256 == 0 ||
256 % MemVT.getScalarSizeInBits() != 0)
return SDValue(N, 0);
SDValue Chain = LD->getChain();
SDValue BasePtr = LD->getBasePtr();
SDNodeFlags Flags = LD->getFlags();
SmallVector<SDValue, 4> LoadOps;
SmallVector<SDValue, 4> LoadOpsChain;
// Replace any non temporal load over 256-bit with a series of 256 bit loads
// and a scalar/vector load less than 256. This way we can utilize 256-bit
// loads and reduce the amount of load instructions generated.
MVT NewVT =
MVT::getVectorVT(MemVT.getVectorElementType().getSimpleVT(),
256 / MemVT.getVectorElementType().getSizeInBits());
unsigned Num256Loads = MemVT.getSizeInBits() / 256;
// Create all 256-bit loads starting from offset 0 and up to Num256Loads-1*32.
for (unsigned I = 0; I < Num256Loads; I++) {
unsigned PtrOffset = I * 32;
SDValue NewPtr = DAG.getMemBasePlusOffset(
BasePtr, TypeSize::getFixed(PtrOffset), DL, Flags);
Align NewAlign = commonAlignment(LD->getAlign(), PtrOffset);
SDValue NewLoad = DAG.getLoad(
NewVT, DL, Chain, NewPtr, LD->getPointerInfo().getWithOffset(PtrOffset),
NewAlign, LD->getMemOperand()->getFlags(), LD->getAAInfo());
LoadOps.push_back(NewLoad);
LoadOpsChain.push_back(SDValue(cast<SDNode>(NewLoad), 1));
}
// Process remaining bits of the load operation.
// This is done by creating an UNDEF vector to match the size of the
// 256-bit loads and inserting the remaining load to it. We extract the
// original load type at the end using EXTRACT_SUBVECTOR instruction.
unsigned BitsRemaining = MemVT.getSizeInBits() % 256;
unsigned PtrOffset = (MemVT.getSizeInBits() - BitsRemaining) / 8;
MVT RemainingVT = MVT::getVectorVT(
MemVT.getVectorElementType().getSimpleVT(),
BitsRemaining / MemVT.getVectorElementType().getSizeInBits());
SDValue NewPtr = DAG.getMemBasePlusOffset(
BasePtr, TypeSize::getFixed(PtrOffset), DL, Flags);
Align NewAlign = commonAlignment(LD->getAlign(), PtrOffset);
SDValue RemainingLoad =
DAG.getLoad(RemainingVT, DL, Chain, NewPtr,
LD->getPointerInfo().getWithOffset(PtrOffset), NewAlign,
LD->getMemOperand()->getFlags(), LD->getAAInfo());
SDValue UndefVector = DAG.getUNDEF(NewVT);
SDValue InsertIdx = DAG.getVectorIdxConstant(0, DL);
SDValue ExtendedRemainingLoad =
DAG.getNode(ISD::INSERT_SUBVECTOR, DL, NewVT,
{UndefVector, RemainingLoad, InsertIdx});
LoadOps.push_back(ExtendedRemainingLoad);
LoadOpsChain.push_back(SDValue(cast<SDNode>(RemainingLoad), 1));
EVT ConcatVT =
EVT::getVectorVT(*DAG.getContext(), MemVT.getScalarType(),
LoadOps.size() * NewVT.getVectorNumElements());
SDValue ConcatVectors =
DAG.getNode(ISD::CONCAT_VECTORS, DL, ConcatVT, LoadOps);
// Extract the original vector type size.
SDValue ExtractSubVector =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MemVT,
{ConcatVectors, DAG.getVectorIdxConstant(0, DL)});
SDValue TokenFactor =
DAG.getNode(ISD::TokenFactor, DL, MVT::Other, LoadOpsChain);
return DAG.getMergeValues({ExtractSubVector, TokenFactor}, DL);
}
static EVT tryGetOriginalBoolVectorType(SDValue Op, int Depth = 0) {
EVT VecVT = Op.getValueType();
assert(VecVT.isVector() && VecVT.getVectorElementType() == MVT::i1 &&
"Need boolean vector type.");
if (Depth > 3)
return MVT::INVALID_SIMPLE_VALUE_TYPE;
// We can get the base type from a vector compare or truncate.
if (Op.getOpcode() == ISD::SETCC || Op.getOpcode() == ISD::TRUNCATE)
return Op.getOperand(0).getValueType();
// If an operand is a bool vector, continue looking.
EVT BaseVT = MVT::INVALID_SIMPLE_VALUE_TYPE;
for (SDValue Operand : Op->op_values()) {
if (Operand.getValueType() != VecVT)
continue;
EVT OperandVT = tryGetOriginalBoolVectorType(Operand, Depth + 1);
if (!BaseVT.isSimple())
BaseVT = OperandVT;
else if (OperandVT != BaseVT)
return MVT::INVALID_SIMPLE_VALUE_TYPE;
}
return BaseVT;
}
// When converting a <N x iX> vector to <N x i1> to store or use as a scalar
// iN, we can use a trick that extracts the i^th bit from the i^th element and
// then performs a vector add to get a scalar bitmask. This requires that each
// element's bits are either all 1 or all 0.
static SDValue vectorToScalarBitmask(SDNode *N, SelectionDAG &DAG) {
SDLoc DL(N);
SDValue ComparisonResult(N, 0);
EVT VecVT = ComparisonResult.getValueType();
assert(VecVT.isVector() && "Must be a vector type");
unsigned NumElts = VecVT.getVectorNumElements();
if (NumElts != 2 && NumElts != 4 && NumElts != 8 && NumElts != 16)
return SDValue();
if (VecVT.getVectorElementType() != MVT::i1 &&
!DAG.getTargetLoweringInfo().isTypeLegal(VecVT))
return SDValue();
// If we can find the original types to work on instead of a vector of i1,
// we can avoid extend/extract conversion instructions.
if (VecVT.getVectorElementType() == MVT::i1) {
VecVT = tryGetOriginalBoolVectorType(ComparisonResult);
if (!VecVT.isSimple()) {
unsigned BitsPerElement = std::max(64 / NumElts, 8u); // >= 64-bit vector
VecVT = MVT::getVectorVT(MVT::getIntegerVT(BitsPerElement), NumElts);
}
}
VecVT = VecVT.changeVectorElementTypeToInteger();
// Large vectors don't map directly to this conversion, so to avoid too many
// edge cases, we don't apply it here. The conversion will likely still be
// applied later via multiple smaller vectors, whose results are concatenated.
if (VecVT.getSizeInBits() > 128)
return SDValue();
// Ensure that all elements' bits are either 0s or 1s.
ComparisonResult = DAG.getSExtOrTrunc(ComparisonResult, DL, VecVT);
SmallVector<SDValue, 16> MaskConstants;
if (DAG.getSubtarget<AArch64Subtarget>().isNeonAvailable() &&
VecVT == MVT::v16i8) {
// v16i8 is a special case, as we have 16 entries but only 8 positional bits
// per entry. We split it into two halves, apply the mask, zip the halves to
// create 8x 16-bit values, and the perform the vector reduce.
for (unsigned Half = 0; Half < 2; ++Half) {
for (unsigned MaskBit = 1; MaskBit <= 128; MaskBit *= 2) {
MaskConstants.push_back(DAG.getConstant(MaskBit, DL, MVT::i32));
}
}
SDValue Mask = DAG.getNode(ISD::BUILD_VECTOR, DL, VecVT, MaskConstants);
SDValue RepresentativeBits =
DAG.getNode(ISD::AND, DL, VecVT, ComparisonResult, Mask);
SDValue UpperRepresentativeBits =
DAG.getNode(AArch64ISD::EXT, DL, VecVT, RepresentativeBits,
RepresentativeBits, DAG.getConstant(8, DL, MVT::i32));
SDValue Zipped = DAG.getNode(AArch64ISD::ZIP1, DL, VecVT,
RepresentativeBits, UpperRepresentativeBits);
Zipped = DAG.getNode(ISD::BITCAST, DL, MVT::v8i16, Zipped);
return DAG.getNode(ISD::VECREDUCE_ADD, DL, MVT::i16, Zipped);
}
// All other vector sizes.
unsigned MaxBitMask = 1u << (VecVT.getVectorNumElements() - 1);
for (unsigned MaskBit = 1; MaskBit <= MaxBitMask; MaskBit *= 2) {
MaskConstants.push_back(DAG.getConstant(MaskBit, DL, MVT::i64));
}
SDValue Mask = DAG.getNode(ISD::BUILD_VECTOR, DL, VecVT, MaskConstants);
SDValue RepresentativeBits =
DAG.getNode(ISD::AND, DL, VecVT, ComparisonResult, Mask);
EVT ResultVT = MVT::getIntegerVT(std::max<unsigned>(
NumElts, VecVT.getVectorElementType().getSizeInBits()));
return DAG.getNode(ISD::VECREDUCE_ADD, DL, ResultVT, RepresentativeBits);
}
static SDValue combineBoolVectorAndTruncateStore(SelectionDAG &DAG,
StoreSDNode *Store) {
if (!Store->isTruncatingStore())
return SDValue();
SDLoc DL(Store);
SDValue VecOp = Store->getValue();
EVT VT = VecOp.getValueType();
EVT MemVT = Store->getMemoryVT();
if (!MemVT.isVector() || !VT.isVector() ||
MemVT.getVectorElementType() != MVT::i1)
return SDValue();
// If we are storing a vector that we are currently building, let
// `scalarizeVectorStore()` handle this more efficiently.
if (VecOp.getOpcode() == ISD::BUILD_VECTOR)
return SDValue();
VecOp = DAG.getNode(ISD::TRUNCATE, DL, MemVT, VecOp);
SDValue VectorBits = vectorToScalarBitmask(VecOp.getNode(), DAG);
if (!VectorBits)
return SDValue();
EVT StoreVT =
EVT::getIntegerVT(*DAG.getContext(), MemVT.getStoreSizeInBits());
SDValue ExtendedBits = DAG.getZExtOrTrunc(VectorBits, DL, StoreVT);
return DAG.getStore(Store->getChain(), DL, ExtendedBits, Store->getBasePtr(),
Store->getMemOperand());
}
bool isHalvingTruncateOfLegalScalableType(EVT SrcVT, EVT DstVT) {
return (SrcVT == MVT::nxv8i16 && DstVT == MVT::nxv8i8) ||
(SrcVT == MVT::nxv4i32 && DstVT == MVT::nxv4i16) ||
(SrcVT == MVT::nxv2i64 && DstVT == MVT::nxv2i32);
}
// Combine store (trunc X to <3 x i8>) to sequence of ST1.b.
static SDValue combineI8TruncStore(StoreSDNode *ST, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
SDValue Value = ST->getValue();
EVT ValueVT = Value.getValueType();
if (ST->isVolatile() || !Subtarget->isLittleEndian() ||
Value.getOpcode() != ISD::TRUNCATE ||
ValueVT != EVT::getVectorVT(*DAG.getContext(), MVT::i8, 3))
return SDValue();
assert(ST->getOffset().isUndef() && "undef offset expected");
SDLoc DL(ST);
auto WideVT = EVT::getVectorVT(
*DAG.getContext(),
Value->getOperand(0).getValueType().getVectorElementType(), 4);
SDValue UndefVector = DAG.getUNDEF(WideVT);
SDValue WideTrunc = DAG.getNode(
ISD::INSERT_SUBVECTOR, DL, WideVT,
{UndefVector, Value->getOperand(0), DAG.getVectorIdxConstant(0, DL)});
SDValue Cast = DAG.getNode(
ISD::BITCAST, DL, WideVT.getSizeInBits() == 64 ? MVT::v8i8 : MVT::v16i8,
WideTrunc);
MachineFunction &MF = DAG.getMachineFunction();
SDValue Chain = ST->getChain();
MachineMemOperand *MMO = ST->getMemOperand();
unsigned IdxScale = WideVT.getScalarSizeInBits() / 8;
SDValue E2 = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i8, Cast,
DAG.getConstant(2 * IdxScale, DL, MVT::i64));
TypeSize Offset2 = TypeSize::getFixed(2);
SDValue Ptr2 = DAG.getMemBasePlusOffset(ST->getBasePtr(), Offset2, DL);
Chain = DAG.getStore(Chain, DL, E2, Ptr2, MF.getMachineMemOperand(MMO, 2, 1));
SDValue E1 = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i8, Cast,
DAG.getConstant(1 * IdxScale, DL, MVT::i64));
TypeSize Offset1 = TypeSize::getFixed(1);
SDValue Ptr1 = DAG.getMemBasePlusOffset(ST->getBasePtr(), Offset1, DL);
Chain = DAG.getStore(Chain, DL, E1, Ptr1, MF.getMachineMemOperand(MMO, 1, 1));
SDValue E0 = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, MVT::i8, Cast,
DAG.getConstant(0, DL, MVT::i64));
Chain = DAG.getStore(Chain, DL, E0, ST->getBasePtr(),
MF.getMachineMemOperand(MMO, 0, 1));
return Chain;
}
static unsigned getFPSubregForVT(EVT VT) {
assert(VT.isSimple() && "Expected simple VT");
switch (VT.getSimpleVT().SimpleTy) {
case MVT::aarch64mfp8:
return AArch64::bsub;
case MVT::f16:
return AArch64::hsub;
case MVT::f32:
return AArch64::ssub;
case MVT::f64:
return AArch64::dsub;
default:
llvm_unreachable("Unexpected VT!");
}
}
static SDValue performSTORECombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
StoreSDNode *ST = cast<StoreSDNode>(N);
SDValue Chain = ST->getChain();
SDValue Value = ST->getValue();
SDValue Ptr = ST->getBasePtr();
EVT ValueVT = Value.getValueType();
EVT MemVT = ST->getMemoryVT();
const TargetLowering &TLI = DAG.getTargetLoweringInfo();
SDLoc DL(ST);
auto hasValidElementTypeForFPTruncStore = [](EVT VT) {
EVT EltVT = VT.getVectorElementType();
return EltVT == MVT::f32 || EltVT == MVT::f64;
};
// Cast ptr32 and ptr64 pointers to the default address space before a store.
unsigned AddrSpace = ST->getAddressSpace();
if (AddrSpace == ARM64AS::PTR64 || AddrSpace == ARM64AS::PTR32_SPTR ||
AddrSpace == ARM64AS::PTR32_UPTR) {
MVT PtrVT = TLI.getPointerTy(DAG.getDataLayout());
if (PtrVT != Ptr.getSimpleValueType()) {
SDValue Cast = DAG.getAddrSpaceCast(DL, PtrVT, Ptr, AddrSpace, 0);
return DAG.getStore(Chain, DL, Value, Cast, ST->getPointerInfo(),
ST->getBaseAlign(), ST->getMemOperand()->getFlags(),
ST->getAAInfo());
}
}
if (SDValue Res = combineI8TruncStore(ST, DAG, Subtarget))
return Res;
// If this is an FP_ROUND followed by a store, fold this into a truncating
// store. We can do this even if this is already a truncstore.
// We purposefully don't care about legality of the nodes here as we know
// they can be split down into something legal.
if (DCI.isBeforeLegalizeOps() && Value.getOpcode() == ISD::FP_ROUND &&
Value.getNode()->hasOneUse() && ST->isUnindexed() &&
Subtarget->useSVEForFixedLengthVectors() &&
ValueVT.isFixedLengthVector() &&
ValueVT.getFixedSizeInBits() >= Subtarget->getMinSVEVectorSizeInBits() &&
hasValidElementTypeForFPTruncStore(Value.getOperand(0).getValueType()))
return DAG.getTruncStore(Chain, DL, Value.getOperand(0), Ptr, MemVT,
ST->getMemOperand());
if (SDValue Split = splitStores(N, DCI, DAG, Subtarget))
return Split;
if (Subtarget->supportsAddressTopByteIgnored() &&
performTBISimplification(N->getOperand(2), DCI, DAG))
return SDValue(N, 0);
if (SDValue Store = foldTruncStoreOfExt(DAG, N))
return Store;
if (SDValue Store = combineBoolVectorAndTruncateStore(DAG, ST))
return Store;
if (ST->isTruncatingStore() &&
isHalvingTruncateOfLegalScalableType(ValueVT, MemVT)) {
if (SDValue Rshrnb =
trySimplifySrlAddToRshrnb(ST->getOperand(1), DAG, Subtarget)) {
return DAG.getTruncStore(ST->getChain(), ST, Rshrnb, ST->getBasePtr(),
MemVT, ST->getMemOperand());
}
}
// This is an integer vector_extract_elt followed by a (possibly truncating)
// store. We may be able to replace this with a store of an FP subregister.
if (DCI.isAfterLegalizeDAG() && ST->isUnindexed() &&
Value.getOpcode() == ISD::EXTRACT_VECTOR_ELT) {
SDValue Vector = Value.getOperand(0);
SDValue ExtIdx = Value.getOperand(1);
EVT VectorVT = Vector.getValueType();
EVT ElemVT = VectorVT.getVectorElementType();
if (!ValueVT.isInteger())
return SDValue();
// Propagate zero constants (applying this fold may miss optimizations).
if (ISD::isConstantSplatVectorAllZeros(Vector.getNode())) {
SDValue ZeroElt = DAG.getConstant(0, DL, ValueVT);
DAG.ReplaceAllUsesWith(Value, ZeroElt);
return SDValue();
}
if (ValueVT != MemVT && !ST->isTruncatingStore())
return SDValue();
// This could generate an additional extract if the index is non-zero and
// the extracted value has multiple uses.
auto *ExtCst = dyn_cast<ConstantSDNode>(ExtIdx);
if ((!ExtCst || !ExtCst->isZero()) && !Value.hasOneUse())
return SDValue();
// These can lower to st1, which is preferable if we're unlikely to fold the
// addressing into the store.
if (Subtarget->isNeonAvailable() && ElemVT == MemVT &&
(VectorVT.is64BitVector() || VectorVT.is128BitVector()) && ExtCst &&
!ExtCst->isZero() && ST->getBasePtr().getOpcode() != ISD::ADD)
return SDValue();
if (MemVT == MVT::i64 || MemVT == MVT::i32) {
// Heuristic: If there are other users of w/x integer scalars extracted
// from this vector that won't fold into the store -- abandon folding.
// Applying this fold may disrupt paired stores.
for (const auto &Use : Vector->uses()) {
if (Use.getResNo() != Vector.getResNo())
continue;
const SDNode *User = Use.getUser();
if (User->getOpcode() == ISD::EXTRACT_VECTOR_ELT &&
(!User->hasOneUse() ||
(*User->user_begin())->getOpcode() != ISD::STORE))
return SDValue();
}
}
SDValue ExtVector = Vector;
if (!ExtCst || !ExtCst->isZero()) {
// Handle extracting from lanes != 0.
SDValue Ext = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL,
Value.getValueType(), Vector, ExtIdx);
SDValue Zero = DAG.getVectorIdxConstant(0, DL);
ExtVector = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, VectorVT,
DAG.getUNDEF(VectorVT), Ext, Zero);
}
EVT FPMemVT = MemVT == MVT::i8
? MVT::aarch64mfp8
: EVT::getFloatingPointVT(MemVT.getSizeInBits());
SDValue FPSubreg = DAG.getTargetExtractSubreg(getFPSubregForVT(FPMemVT), DL,
FPMemVT, ExtVector);
return DAG.getStore(ST->getChain(), DL, FPSubreg, ST->getBasePtr(),
ST->getMemOperand());
}
return SDValue();
}
static SDValue performMSTORECombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
MaskedStoreSDNode *MST = cast<MaskedStoreSDNode>(N);
SDValue Value = MST->getValue();
SDValue Mask = MST->getMask();
SDLoc DL(N);
// If this is a UZP1 followed by a masked store, fold this into a masked
// truncating store. We can do this even if this is already a masked
// truncstore.
if (Value.getOpcode() == AArch64ISD::UZP1 && Value->hasOneUse() &&
MST->isUnindexed() && Mask->getOpcode() == AArch64ISD::PTRUE &&
Value.getValueType().isInteger()) {
Value = Value.getOperand(0);
if (Value.getOpcode() == ISD::BITCAST) {
EVT HalfVT =
Value.getValueType().getHalfNumVectorElementsVT(*DAG.getContext());
EVT InVT = Value.getOperand(0).getValueType();
if (HalfVT.widenIntegerVectorElementType(*DAG.getContext()) == InVT) {
unsigned MinSVESize = Subtarget->getMinSVEVectorSizeInBits();
unsigned PgPattern = Mask->getConstantOperandVal(0);
// Ensure we can double the size of the predicate pattern
unsigned NumElts = getNumElementsFromSVEPredPattern(PgPattern);
if (NumElts && NumElts * InVT.getVectorElementType().getSizeInBits() <=
MinSVESize) {
Mask = getPTrue(DAG, DL, InVT.changeVectorElementType(MVT::i1),
PgPattern);
return DAG.getMaskedStore(MST->getChain(), DL, Value.getOperand(0),
MST->getBasePtr(), MST->getOffset(), Mask,
MST->getMemoryVT(), MST->getMemOperand(),
MST->getAddressingMode(),
/*IsTruncating=*/true);
}
}
}
}
if (MST->isTruncatingStore()) {
EVT ValueVT = Value->getValueType(0);
EVT MemVT = MST->getMemoryVT();
if (!isHalvingTruncateOfLegalScalableType(ValueVT, MemVT))
return SDValue();
if (SDValue Rshrnb = trySimplifySrlAddToRshrnb(Value, DAG, Subtarget)) {
return DAG.getMaskedStore(MST->getChain(), DL, Rshrnb, MST->getBasePtr(),
MST->getOffset(), MST->getMask(),
MST->getMemoryVT(), MST->getMemOperand(),
MST->getAddressingMode(), true);
}
}
return SDValue();
}
/// \return true if part of the index was folded into the Base.
static bool foldIndexIntoBase(SDValue &BasePtr, SDValue &Index, SDValue Scale,
SDLoc DL, SelectionDAG &DAG) {
// This function assumes a vector of i64 indices.
EVT IndexVT = Index.getValueType();
if (!IndexVT.isVector() || IndexVT.getVectorElementType() != MVT::i64)
return false;
// Simplify:
// BasePtr = Ptr
// Index = X + splat(Offset)
// ->
// BasePtr = Ptr + Offset * scale.
// Index = X
if (Index.getOpcode() == ISD::ADD) {
if (auto Offset = DAG.getSplatValue(Index.getOperand(1))) {
Offset = DAG.getNode(ISD::MUL, DL, MVT::i64, Offset, Scale);
BasePtr = DAG.getNode(ISD::ADD, DL, MVT::i64, BasePtr, Offset);
Index = Index.getOperand(0);
return true;
}
}
// Simplify:
// BasePtr = Ptr
// Index = (X + splat(Offset)) << splat(Shift)
// ->
// BasePtr = Ptr + (Offset << Shift) * scale)
// Index = X << splat(shift)
if (Index.getOpcode() == ISD::SHL &&
Index.getOperand(0).getOpcode() == ISD::ADD) {
SDValue Add = Index.getOperand(0);
SDValue ShiftOp = Index.getOperand(1);
SDValue OffsetOp = Add.getOperand(1);
if (auto Shift = DAG.getSplatValue(ShiftOp))
if (auto Offset = DAG.getSplatValue(OffsetOp)) {
Offset = DAG.getNode(ISD::SHL, DL, MVT::i64, Offset, Shift);
Offset = DAG.getNode(ISD::MUL, DL, MVT::i64, Offset, Scale);
BasePtr = DAG.getNode(ISD::ADD, DL, MVT::i64, BasePtr, Offset);
Index = DAG.getNode(ISD::SHL, DL, Index.getValueType(),
Add.getOperand(0), ShiftOp);
return true;
}
}
return false;
}
// Analyse the specified address returning true if a more optimal addressing
// mode is available. When returning true all parameters are updated to reflect
// their recommended values.
static bool findMoreOptimalIndexType(const MaskedGatherScatterSDNode *N,
SDValue &BasePtr, SDValue &Index,
SelectionDAG &DAG) {
// Try to iteratively fold parts of the index into the base pointer to
// simplify the index as much as possible.
bool Changed = false;
while (foldIndexIntoBase(BasePtr, Index, N->getScale(), SDLoc(N), DAG))
Changed = true;
// Only consider element types that are pointer sized as smaller types can
// be easily promoted.
EVT IndexVT = Index.getValueType();
if (IndexVT.getVectorElementType() != MVT::i64 || IndexVT == MVT::nxv2i64)
return Changed;
// Can indices be trivially shrunk?
EVT DataVT = N->getOperand(1).getValueType();
// Don't attempt to shrink the index for fixed vectors of 64 bit data since it
// will later be re-extended to 64 bits in legalization
if (DataVT.isFixedLengthVector() && DataVT.getScalarSizeInBits() == 64)
return Changed;
if (ISD::isVectorShrinkable(Index.getNode(), 32, N->isIndexSigned())) {
EVT NewIndexVT = IndexVT.changeVectorElementType(MVT::i32);
Index = DAG.getNode(ISD::TRUNCATE, SDLoc(N), NewIndexVT, Index);
return true;
}
// Match:
// Index = step(const)
int64_t Stride = 0;
if (Index.getOpcode() == ISD::STEP_VECTOR) {
Stride = cast<ConstantSDNode>(Index.getOperand(0))->getSExtValue();
}
// Match:
// Index = step(const) << shift(const)
else if (Index.getOpcode() == ISD::SHL &&
Index.getOperand(0).getOpcode() == ISD::STEP_VECTOR) {
SDValue RHS = Index.getOperand(1);
if (auto *Shift =
dyn_cast_or_null<ConstantSDNode>(DAG.getSplatValue(RHS))) {
int64_t Step = (int64_t)Index.getOperand(0).getConstantOperandVal(1);
Stride = Step << Shift->getZExtValue();
}
}
// Return early because no supported pattern is found.
if (Stride == 0)
return Changed;
if (Stride < std::numeric_limits<int32_t>::min() ||
Stride > std::numeric_limits<int32_t>::max())
return Changed;
const auto &Subtarget = DAG.getSubtarget<AArch64Subtarget>();
unsigned MaxVScale =
Subtarget.getMaxSVEVectorSizeInBits() / AArch64::SVEBitsPerBlock;
int64_t LastElementOffset =
IndexVT.getVectorMinNumElements() * Stride * MaxVScale;
if (LastElementOffset < std::numeric_limits<int32_t>::min() ||
LastElementOffset > std::numeric_limits<int32_t>::max())
return Changed;
EVT NewIndexVT = IndexVT.changeVectorElementType(MVT::i32);
// Stride does not scale explicitly by 'Scale', because it happens in
// the gather/scatter addressing mode.
Index = DAG.getStepVector(SDLoc(N), NewIndexVT, APInt(32, Stride, true));
return true;
}
static SDValue performMaskedGatherScatterCombine(
SDNode *N, TargetLowering::DAGCombinerInfo &DCI, SelectionDAG &DAG) {
if (!DCI.isBeforeLegalize())
return SDValue();
MaskedGatherScatterSDNode *MGS = cast<MaskedGatherScatterSDNode>(N);
SDLoc DL(MGS);
SDValue Chain = MGS->getChain();
SDValue Scale = MGS->getScale();
SDValue Index = MGS->getIndex();
SDValue Mask = MGS->getMask();
SDValue BasePtr = MGS->getBasePtr();
ISD::MemIndexType IndexType = MGS->getIndexType();
if (!findMoreOptimalIndexType(MGS, BasePtr, Index, DAG))
return SDValue();
// Here we catch such cases early and change MGATHER's IndexType to allow
// the use of an Index that's more legalisation friendly.
if (auto *MGT = dyn_cast<MaskedGatherSDNode>(MGS)) {
SDValue PassThru = MGT->getPassThru();
SDValue Ops[] = {Chain, PassThru, Mask, BasePtr, Index, Scale};
return DAG.getMaskedGather(
DAG.getVTList(N->getValueType(0), MVT::Other), MGT->getMemoryVT(), DL,
Ops, MGT->getMemOperand(), IndexType, MGT->getExtensionType());
}
if (auto *MSC = dyn_cast<MaskedScatterSDNode>(MGS)) {
SDValue Data = MSC->getValue();
SDValue Ops[] = {Chain, Data, Mask, BasePtr, Index, Scale};
return DAG.getMaskedScatter(DAG.getVTList(MVT::Other), MSC->getMemoryVT(),
DL, Ops, MSC->getMemOperand(), IndexType,
MSC->isTruncatingStore());
}
auto *HG = cast<MaskedHistogramSDNode>(MGS);
SDValue Ops[] = {Chain, HG->getInc(), Mask, BasePtr,
Index, Scale, HG->getIntID()};
return DAG.getMaskedHistogram(DAG.getVTList(MVT::Other), HG->getMemoryVT(),
DL, Ops, HG->getMemOperand(), IndexType);
}
/// Target-specific DAG combine function for NEON load/store intrinsics
/// to merge base address updates.
static SDValue performNEONPostLDSTCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (DCI.isBeforeLegalize() || DCI.isCalledByLegalizer())
return SDValue();
unsigned AddrOpIdx = N->getNumOperands() - 1;
SDValue Addr = N->getOperand(AddrOpIdx);
// Search for a use of the address operand that is an increment.
for (SDUse &Use : Addr->uses()) {
SDNode *User = Use.getUser();
if (User->getOpcode() != ISD::ADD || Use.getResNo() != Addr.getResNo())
continue;
// Check that the add is independent of the load/store. Otherwise, folding
// it would create a cycle.
SmallPtrSet<const SDNode *, 32> Visited;
SmallVector<const SDNode *, 16> Worklist;
Visited.insert(Addr.getNode());
Worklist.push_back(N);
Worklist.push_back(User);
if (SDNode::hasPredecessorHelper(N, Visited, Worklist) ||
SDNode::hasPredecessorHelper(User, Visited, Worklist))
continue;
// Find the new opcode for the updating load/store.
bool IsStore = false;
bool IsLaneOp = false;
bool IsDupOp = false;
unsigned NewOpc = 0;
unsigned NumVecs = 0;
unsigned IntNo = N->getConstantOperandVal(1);
switch (IntNo) {
default: llvm_unreachable("unexpected intrinsic for Neon base update");
case Intrinsic::aarch64_neon_ld2: NewOpc = AArch64ISD::LD2post;
NumVecs = 2; break;
case Intrinsic::aarch64_neon_ld3: NewOpc = AArch64ISD::LD3post;
NumVecs = 3; break;
case Intrinsic::aarch64_neon_ld4: NewOpc = AArch64ISD::LD4post;
NumVecs = 4; break;
case Intrinsic::aarch64_neon_st2: NewOpc = AArch64ISD::ST2post;
NumVecs = 2; IsStore = true; break;
case Intrinsic::aarch64_neon_st3: NewOpc = AArch64ISD::ST3post;
NumVecs = 3; IsStore = true; break;
case Intrinsic::aarch64_neon_st4: NewOpc = AArch64ISD::ST4post;
NumVecs = 4; IsStore = true; break;
case Intrinsic::aarch64_neon_ld1x2: NewOpc = AArch64ISD::LD1x2post;
NumVecs = 2; break;
case Intrinsic::aarch64_neon_ld1x3: NewOpc = AArch64ISD::LD1x3post;
NumVecs = 3; break;
case Intrinsic::aarch64_neon_ld1x4: NewOpc = AArch64ISD::LD1x4post;
NumVecs = 4; break;
case Intrinsic::aarch64_neon_st1x2: NewOpc = AArch64ISD::ST1x2post;
NumVecs = 2; IsStore = true; break;
case Intrinsic::aarch64_neon_st1x3: NewOpc = AArch64ISD::ST1x3post;
NumVecs = 3; IsStore = true; break;
case Intrinsic::aarch64_neon_st1x4: NewOpc = AArch64ISD::ST1x4post;
NumVecs = 4; IsStore = true; break;
case Intrinsic::aarch64_neon_ld2r: NewOpc = AArch64ISD::LD2DUPpost;
NumVecs = 2; IsDupOp = true; break;
case Intrinsic::aarch64_neon_ld3r: NewOpc = AArch64ISD::LD3DUPpost;
NumVecs = 3; IsDupOp = true; break;
case Intrinsic::aarch64_neon_ld4r: NewOpc = AArch64ISD::LD4DUPpost;
NumVecs = 4; IsDupOp = true; break;
case Intrinsic::aarch64_neon_ld2lane: NewOpc = AArch64ISD::LD2LANEpost;
NumVecs = 2; IsLaneOp = true; break;
case Intrinsic::aarch64_neon_ld3lane: NewOpc = AArch64ISD::LD3LANEpost;
NumVecs = 3; IsLaneOp = true; break;
case Intrinsic::aarch64_neon_ld4lane: NewOpc = AArch64ISD::LD4LANEpost;
NumVecs = 4; IsLaneOp = true; break;
case Intrinsic::aarch64_neon_st2lane: NewOpc = AArch64ISD::ST2LANEpost;
NumVecs = 2; IsStore = true; IsLaneOp = true; break;
case Intrinsic::aarch64_neon_st3lane: NewOpc = AArch64ISD::ST3LANEpost;
NumVecs = 3; IsStore = true; IsLaneOp = true; break;
case Intrinsic::aarch64_neon_st4lane: NewOpc = AArch64ISD::ST4LANEpost;
NumVecs = 4; IsStore = true; IsLaneOp = true; break;
}
EVT VecTy;
if (IsStore)
VecTy = N->getOperand(2).getValueType();
else
VecTy = N->getValueType(0);
// If the increment is a constant, it must match the memory ref size.
SDValue Inc = User->getOperand(User->getOperand(0) == Addr ? 1 : 0);
if (ConstantSDNode *CInc = dyn_cast<ConstantSDNode>(Inc.getNode())) {
uint32_t IncVal = CInc->getZExtValue();
unsigned NumBytes = NumVecs * VecTy.getSizeInBits() / 8;
if (IsLaneOp || IsDupOp)
NumBytes /= VecTy.getVectorNumElements();
if (IncVal != NumBytes)
continue;
Inc = DAG.getRegister(AArch64::XZR, MVT::i64);
}
SmallVector<SDValue, 8> Ops;
Ops.push_back(N->getOperand(0)); // Incoming chain
// Load lane and store have vector list as input.
if (IsLaneOp || IsStore)
for (unsigned i = 2; i < AddrOpIdx; ++i)
Ops.push_back(N->getOperand(i));
Ops.push_back(Addr); // Base register
Ops.push_back(Inc);
// Return Types.
EVT Tys[6];
unsigned NumResultVecs = (IsStore ? 0 : NumVecs);
unsigned n;
for (n = 0; n < NumResultVecs; ++n)
Tys[n] = VecTy;
Tys[n++] = MVT::i64; // Type of write back register
Tys[n] = MVT::Other; // Type of the chain
SDVTList SDTys = DAG.getVTList(ArrayRef(Tys, NumResultVecs + 2));
MemIntrinsicSDNode *MemInt = cast<MemIntrinsicSDNode>(N);
SDValue UpdN = DAG.getMemIntrinsicNode(NewOpc, SDLoc(N), SDTys, Ops,
MemInt->getMemoryVT(),
MemInt->getMemOperand());
// Update the uses.
std::vector<SDValue> NewResults;
for (unsigned i = 0; i < NumResultVecs; ++i) {
NewResults.push_back(SDValue(UpdN.getNode(), i));
}
NewResults.push_back(SDValue(UpdN.getNode(), NumResultVecs + 1));
DCI.CombineTo(N, NewResults);
DCI.CombineTo(User, SDValue(UpdN.getNode(), NumResultVecs));
break;
}
return SDValue();
}
// Checks to see if the value is the prescribed width and returns information
// about its extension mode.
static
bool checkValueWidth(SDValue V, unsigned width, ISD::LoadExtType &ExtType) {
ExtType = ISD::NON_EXTLOAD;
switch(V.getNode()->getOpcode()) {
default:
return false;
case ISD::LOAD: {
LoadSDNode *LoadNode = cast<LoadSDNode>(V.getNode());
if ((LoadNode->getMemoryVT() == MVT::i8 && width == 8)
|| (LoadNode->getMemoryVT() == MVT::i16 && width == 16)) {
ExtType = LoadNode->getExtensionType();
return true;
}
return false;
}
case ISD::AssertSext: {
VTSDNode *TypeNode = cast<VTSDNode>(V.getNode()->getOperand(1));
if ((TypeNode->getVT() == MVT::i8 && width == 8)
|| (TypeNode->getVT() == MVT::i16 && width == 16)) {
ExtType = ISD::SEXTLOAD;
return true;
}
return false;
}
case ISD::AssertZext: {
VTSDNode *TypeNode = cast<VTSDNode>(V.getNode()->getOperand(1));
if ((TypeNode->getVT() == MVT::i8 && width == 8)
|| (TypeNode->getVT() == MVT::i16 && width == 16)) {
ExtType = ISD::ZEXTLOAD;
return true;
}
return false;
}
case ISD::Constant:
case ISD::TargetConstant: {
return std::abs(cast<ConstantSDNode>(V.getNode())->getSExtValue()) <
1LL << (width - 1);
}
}
return true;
}
// This function does a whole lot of voodoo to determine if the tests are
// equivalent without and with a mask. Essentially what happens is that given a
// DAG resembling:
//
// +-------------+ +-------------+ +-------------+ +-------------+
// | Input | | AddConstant | | CompConstant| | CC |
// +-------------+ +-------------+ +-------------+ +-------------+
// | | | |
// V V | +----------+
// +-------------+ +----+ | |
// | ADD | |0xff| | |
// +-------------+ +----+ | |
// | | | |
// V V | |
// +-------------+ | |
// | AND | | |
// +-------------+ | |
// | | |
// +-----+ | |
// | | |
// V V V
// +-------------+
// | CMP |
// +-------------+
//
// The AND node may be safely removed for some combinations of inputs. In
// particular we need to take into account the extension type of the Input,
// the exact values of AddConstant, CompConstant, and CC, along with the nominal
// width of the input (this can work for any width inputs, the above graph is
// specific to 8 bits.
//
// The specific equations were worked out by generating output tables for each
// AArch64CC value in terms of and AddConstant (w1), CompConstant(w2). The
// problem was simplified by working with 4 bit inputs, which means we only
// needed to reason about 24 distinct bit patterns: 8 patterns unique to zero
// extension (8,15), 8 patterns unique to sign extensions (-8,-1), and 8
// patterns present in both extensions (0,7). For every distinct set of
// AddConstant and CompConstants bit patterns we can consider the masked and
// unmasked versions to be equivalent if the result of this function is true for
// all 16 distinct bit patterns of for the current extension type of Input (w0).
//
// sub w8, w0, w1
// and w10, w8, #0x0f
// cmp w8, w2
// cset w9, AArch64CC
// cmp w10, w2
// cset w11, AArch64CC
// cmp w9, w11
// cset w0, eq
// ret
//
// Since the above function shows when the outputs are equivalent it defines
// when it is safe to remove the AND. Unfortunately it only runs on AArch64 and
// would be expensive to run during compiles. The equations below were written
// in a test harness that confirmed they gave equivalent outputs to the above
// for all inputs function, so they can be used determine if the removal is
// legal instead.
//
// isEquivalentMaskless() is the code for testing if the AND can be removed
// factored out of the DAG recognition as the DAG can take several forms.
static bool isEquivalentMaskless(unsigned CC, unsigned width,
ISD::LoadExtType ExtType, int AddConstant,
int CompConstant) {
// By being careful about our equations and only writing the in term
// symbolic values and well known constants (0, 1, -1, MaxUInt) we can
// make them generally applicable to all bit widths.
int MaxUInt = (1 << width);
// For the purposes of these comparisons sign extending the type is
// equivalent to zero extending the add and displacing it by half the integer
// width. Provided we are careful and make sure our equations are valid over
// the whole range we can just adjust the input and avoid writing equations
// for sign extended inputs.
if (ExtType == ISD::SEXTLOAD)
AddConstant -= (1 << (width-1));
switch(CC) {
case AArch64CC::LE:
case AArch64CC::GT:
if ((AddConstant == 0) ||
(CompConstant == MaxUInt - 1 && AddConstant < 0) ||
(AddConstant >= 0 && CompConstant < 0) ||
(AddConstant <= 0 && CompConstant <= 0 && CompConstant < AddConstant))
return true;
break;
case AArch64CC::LT:
case AArch64CC::GE:
if ((AddConstant == 0) ||
(AddConstant >= 0 && CompConstant <= 0) ||
(AddConstant <= 0 && CompConstant <= 0 && CompConstant <= AddConstant))
return true;
break;
case AArch64CC::HI:
case AArch64CC::LS:
if ((AddConstant >= 0 && CompConstant < 0) ||
(AddConstant <= 0 && CompConstant >= -1 &&
CompConstant < AddConstant + MaxUInt))
return true;
break;
case AArch64CC::PL:
case AArch64CC::MI:
if ((AddConstant == 0) ||
(AddConstant > 0 && CompConstant <= 0) ||
(AddConstant < 0 && CompConstant <= AddConstant))
return true;
break;
case AArch64CC::LO:
case AArch64CC::HS:
if ((AddConstant >= 0 && CompConstant <= 0) ||
(AddConstant <= 0 && CompConstant >= 0 &&
CompConstant <= AddConstant + MaxUInt))
return true;
break;
case AArch64CC::EQ:
case AArch64CC::NE:
if ((AddConstant > 0 && CompConstant < 0) ||
(AddConstant < 0 && CompConstant >= 0 &&
CompConstant < AddConstant + MaxUInt) ||
(AddConstant >= 0 && CompConstant >= 0 &&
CompConstant >= AddConstant) ||
(AddConstant <= 0 && CompConstant < 0 && CompConstant < AddConstant))
return true;
break;
case AArch64CC::VS:
case AArch64CC::VC:
case AArch64CC::AL:
case AArch64CC::NV:
return true;
case AArch64CC::Invalid:
break;
}
return false;
}
// (X & C) >u Mask --> (X & (C & (~Mask)) != 0
// (X & C) <u Pow2 --> (X & (C & ~(Pow2-1)) == 0
static SDValue performSubsToAndsCombine(SDNode *N, SDNode *SubsNode,
SDNode *AndNode, SelectionDAG &DAG,
unsigned CCIndex, unsigned CmpIndex,
unsigned CC) {
ConstantSDNode *SubsC = dyn_cast<ConstantSDNode>(SubsNode->getOperand(1));
if (!SubsC)
return SDValue();
APInt SubsAP = SubsC->getAPIntValue();
if (CC == AArch64CC::HI) {
if (!SubsAP.isMask())
return SDValue();
} else if (CC == AArch64CC::LO) {
if (!SubsAP.isPowerOf2())
return SDValue();
} else
return SDValue();
ConstantSDNode *AndC = dyn_cast<ConstantSDNode>(AndNode->getOperand(1));
if (!AndC)
return SDValue();
APInt MaskAP = CC == AArch64CC::HI ? SubsAP : (SubsAP - 1);
SDLoc DL(N);
APInt AndSMask = (~MaskAP) & AndC->getAPIntValue();
SDValue ANDS = DAG.getNode(
AArch64ISD::ANDS, DL, SubsNode->getVTList(), AndNode->getOperand(0),
DAG.getConstant(AndSMask, DL, SubsC->getValueType(0)));
SDValue AArch64_CC =
DAG.getConstant(CC == AArch64CC::HI ? AArch64CC::NE : AArch64CC::EQ, DL,
N->getOperand(CCIndex)->getValueType(0));
// For now, only performCSELCombine and performBRCONDCombine call this
// function. And both of them pass 2 for CCIndex, 3 for CmpIndex with 4
// operands. So just init the ops direct to simplify the code. If we have some
// other case with different CCIndex, CmpIndex, we need to use for loop to
// rewrite the code here.
// TODO: Do we need to assert number of operand is 4 here?
assert((CCIndex == 2 && CmpIndex == 3) &&
"Expected CCIndex to be 2 and CmpIndex to be 3.");
SDValue Ops[] = {N->getOperand(0), N->getOperand(1), AArch64_CC,
ANDS.getValue(1)};
return DAG.getNode(N->getOpcode(), N, N->getVTList(), Ops);
}
static
SDValue performCONDCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG, unsigned CCIndex,
unsigned CmpIndex) {
unsigned CC = cast<ConstantSDNode>(N->getOperand(CCIndex))->getSExtValue();
SDNode *SubsNode = N->getOperand(CmpIndex).getNode();
unsigned CondOpcode = SubsNode->getOpcode();
if (CondOpcode != AArch64ISD::SUBS || SubsNode->hasAnyUseOfValue(0) ||
!SubsNode->hasOneUse())
return SDValue();
// There is a SUBS feeding this condition. Is it fed by a mask we can
// use?
SDNode *AndNode = SubsNode->getOperand(0).getNode();
unsigned MaskBits = 0;
if (AndNode->getOpcode() != ISD::AND)
return SDValue();
if (SDValue Val = performSubsToAndsCombine(N, SubsNode, AndNode, DAG, CCIndex,
CmpIndex, CC))
return Val;
if (ConstantSDNode *CN = dyn_cast<ConstantSDNode>(AndNode->getOperand(1))) {
uint32_t CNV = CN->getZExtValue();
if (CNV == 255)
MaskBits = 8;
else if (CNV == 65535)
MaskBits = 16;
}
if (!MaskBits)
return SDValue();
SDValue AddValue = AndNode->getOperand(0);
if (AddValue.getOpcode() != ISD::ADD)
return SDValue();
// The basic dag structure is correct, grab the inputs and validate them.
SDValue AddInputValue1 = AddValue.getNode()->getOperand(0);
SDValue AddInputValue2 = AddValue.getNode()->getOperand(1);
SDValue SubsInputValue = SubsNode->getOperand(1);
// The mask is present and the provenance of all the values is a smaller type,
// lets see if the mask is superfluous.
if (!isa<ConstantSDNode>(AddInputValue2.getNode()) ||
!isa<ConstantSDNode>(SubsInputValue.getNode()))
return SDValue();
ISD::LoadExtType ExtType;
if (!checkValueWidth(SubsInputValue, MaskBits, ExtType) ||
!checkValueWidth(AddInputValue2, MaskBits, ExtType) ||
!checkValueWidth(AddInputValue1, MaskBits, ExtType) )
return SDValue();
if(!isEquivalentMaskless(CC, MaskBits, ExtType,
cast<ConstantSDNode>(AddInputValue2.getNode())->getSExtValue(),
cast<ConstantSDNode>(SubsInputValue.getNode())->getSExtValue()))
return SDValue();
// The AND is not necessary, remove it.
SDVTList VTs = DAG.getVTList(SubsNode->getValueType(0),
SubsNode->getValueType(1));
SDValue Ops[] = { AddValue, SubsNode->getOperand(1) };
SDValue NewValue = DAG.getNode(CondOpcode, SDLoc(SubsNode), VTs, Ops);
DAG.ReplaceAllUsesWith(SubsNode, NewValue.getNode());
return SDValue(N, 0);
}
// Optimize compare with zero and branch.
static SDValue performBRCONDCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
MachineFunction &MF = DAG.getMachineFunction();
// Speculation tracking/SLH assumes that optimized TB(N)Z/CB(N)Z instructions
// will not be produced, as they are conditional branch instructions that do
// not set flags.
if (MF.getFunction().hasFnAttribute(Attribute::SpeculativeLoadHardening))
return SDValue();
if (SDValue NV = performCONDCombine(N, DCI, DAG, 2, 3))
N = NV.getNode();
SDValue Chain = N->getOperand(0);
SDValue Dest = N->getOperand(1);
SDValue CCVal = N->getOperand(2);
SDValue Cmp = N->getOperand(3);
assert(isa<ConstantSDNode>(CCVal) && "Expected a ConstantSDNode here!");
unsigned CC = CCVal->getAsZExtVal();
if (CC != AArch64CC::EQ && CC != AArch64CC::NE)
return SDValue();
// Fold away brcond(NE, cmp(csel(1, 0, CC, Cmp), 1)) -> brcond(~CC, Cmp)
if (isCMP(Cmp) && CC == AArch64CC::NE && isOneConstant(Cmp.getOperand(1))) {
SDValue CSel = Cmp.getOperand(0);
auto CSelCC = getCSETCondCode(CSel);
if (CSelCC) {
SDLoc DL(N);
return DAG.getNode(
N->getOpcode(), DL, N->getVTList(), Chain, Dest,
DAG.getConstant(getInvertedCondCode(*CSelCC), DL, MVT::i32),
CSel.getOperand(3));
}
}
unsigned CmpOpc = Cmp.getOpcode();
if (CmpOpc != AArch64ISD::ADDS && CmpOpc != AArch64ISD::SUBS)
return SDValue();
// Only attempt folding if there is only one use of the flag and no use of the
// value.
if (!Cmp->hasNUsesOfValue(0, 0) || !Cmp->hasNUsesOfValue(1, 1))
return SDValue();
SDValue LHS = Cmp.getOperand(0);
SDValue RHS = Cmp.getOperand(1);
assert(LHS.getValueType() == RHS.getValueType() &&
"Expected the value type to be the same for both operands!");
if (LHS.getValueType() != MVT::i32 && LHS.getValueType() != MVT::i64)
return SDValue();
if (isNullConstant(LHS))
std::swap(LHS, RHS);
if (!isNullConstant(RHS))
return SDValue();
if (LHS.getOpcode() == ISD::SHL || LHS.getOpcode() == ISD::SRA ||
LHS.getOpcode() == ISD::SRL)
return SDValue();
// Fold the compare into the branch instruction.
SDValue BR;
if (CC == AArch64CC::EQ)
BR = DAG.getNode(AArch64ISD::CBZ, SDLoc(N), MVT::Other, Chain, LHS, Dest);
else
BR = DAG.getNode(AArch64ISD::CBNZ, SDLoc(N), MVT::Other, Chain, LHS, Dest);
// Do not add new nodes to DAG combiner worklist.
DCI.CombineTo(N, BR, false);
return SDValue();
}
static SDValue foldCSELofCTTZ(SDNode *N, SelectionDAG &DAG) {
unsigned CC = N->getConstantOperandVal(2);
SDValue SUBS = N->getOperand(3);
SDValue Zero, CTTZ;
if (CC == AArch64CC::EQ && SUBS.getOpcode() == AArch64ISD::SUBS) {
Zero = N->getOperand(0);
CTTZ = N->getOperand(1);
} else if (CC == AArch64CC::NE && SUBS.getOpcode() == AArch64ISD::SUBS) {
Zero = N->getOperand(1);
CTTZ = N->getOperand(0);
} else
return SDValue();
if ((CTTZ.getOpcode() != ISD::CTTZ && CTTZ.getOpcode() != ISD::TRUNCATE) ||
(CTTZ.getOpcode() == ISD::TRUNCATE &&
CTTZ.getOperand(0).getOpcode() != ISD::CTTZ))
return SDValue();
assert((CTTZ.getValueType() == MVT::i32 || CTTZ.getValueType() == MVT::i64) &&
"Illegal type in CTTZ folding");
if (!isNullConstant(Zero) || !isNullConstant(SUBS.getOperand(1)))
return SDValue();
SDValue X = CTTZ.getOpcode() == ISD::TRUNCATE
? CTTZ.getOperand(0).getOperand(0)
: CTTZ.getOperand(0);
if (X != SUBS.getOperand(0))
return SDValue();
unsigned BitWidth = CTTZ.getOpcode() == ISD::TRUNCATE
? CTTZ.getOperand(0).getValueSizeInBits()
: CTTZ.getValueSizeInBits();
SDValue BitWidthMinusOne =
DAG.getConstant(BitWidth - 1, SDLoc(N), CTTZ.getValueType());
return DAG.getNode(ISD::AND, SDLoc(N), CTTZ.getValueType(), CTTZ,
BitWidthMinusOne);
}
// (CSEL l r EQ (CMP (CSEL x y cc2 cond) x)) => (CSEL l r cc2 cond)
// (CSEL l r EQ (CMP (CSEL x y cc2 cond) y)) => (CSEL l r !cc2 cond)
// Where x and y are constants and x != y
// (CSEL l r NE (CMP (CSEL x y cc2 cond) x)) => (CSEL l r !cc2 cond)
// (CSEL l r NE (CMP (CSEL x y cc2 cond) y)) => (CSEL l r cc2 cond)
// Where x and y are constants and x != y
static SDValue foldCSELOfCSEL(SDNode *Op, SelectionDAG &DAG) {
SDValue L = Op->getOperand(0);
SDValue R = Op->getOperand(1);
AArch64CC::CondCode OpCC =
static_cast<AArch64CC::CondCode>(Op->getConstantOperandVal(2));
SDValue OpCmp = Op->getOperand(3);
if (!isCMP(OpCmp))
return SDValue();
SDValue CmpLHS = OpCmp.getOperand(0);
SDValue CmpRHS = OpCmp.getOperand(1);
if (CmpRHS.getOpcode() == AArch64ISD::CSEL)
std::swap(CmpLHS, CmpRHS);
else if (CmpLHS.getOpcode() != AArch64ISD::CSEL)
return SDValue();
SDValue X = CmpLHS->getOperand(0);
SDValue Y = CmpLHS->getOperand(1);
if (!isa<ConstantSDNode>(X) || !isa<ConstantSDNode>(Y) || X == Y) {
return SDValue();
}
// If one of the constant is opaque constant, x,y sdnode is still different
// but the real value maybe the same. So check APInt here to make sure the
// code is correct.
ConstantSDNode *CX = cast<ConstantSDNode>(X);
ConstantSDNode *CY = cast<ConstantSDNode>(Y);
if (CX->getAPIntValue() == CY->getAPIntValue())
return SDValue();
AArch64CC::CondCode CC =
static_cast<AArch64CC::CondCode>(CmpLHS->getConstantOperandVal(2));
SDValue Cond = CmpLHS->getOperand(3);
if (CmpRHS == Y)
CC = AArch64CC::getInvertedCondCode(CC);
else if (CmpRHS != X)
return SDValue();
if (OpCC == AArch64CC::NE)
CC = AArch64CC::getInvertedCondCode(CC);
else if (OpCC != AArch64CC::EQ)
return SDValue();
SDLoc DL(Op);
EVT VT = Op->getValueType(0);
SDValue CCValue = DAG.getConstant(CC, DL, MVT::i32);
return DAG.getNode(AArch64ISD::CSEL, DL, VT, L, R, CCValue, Cond);
}
// Reassociate the true/false expressions of a CSEL instruction to obtain a
// common subexpression with the comparison instruction. For example, change
// (CSEL (ADD (ADD x y) -c) f LO (SUBS x c)) to
// (CSEL (ADD (SUBS x c) y) f LO (SUBS x c)) such that (SUBS x c) is a common
// subexpression.
static SDValue reassociateCSELOperandsForCSE(SDNode *N, SelectionDAG &DAG) {
SDValue SubsNode = N->getOperand(3);
if (SubsNode.getOpcode() != AArch64ISD::SUBS || !SubsNode.hasOneUse())
return SDValue();
SDValue CmpOpToMatch = SubsNode.getOperand(1);
SDValue CmpOpOther = SubsNode.getOperand(0);
EVT VT = N->getValueType(0);
unsigned ExpectedOpcode;
SDValue ExpectedOp;
SDValue SubsOp;
auto *CmpOpConst = dyn_cast<ConstantSDNode>(CmpOpToMatch);
if (CmpOpConst) {
ExpectedOpcode = ISD::ADD;
ExpectedOp =
DAG.getConstant(-CmpOpConst->getAPIntValue(), SDLoc(CmpOpConst),
CmpOpConst->getValueType(0));
SubsOp = DAG.getConstant(CmpOpConst->getAPIntValue(), SDLoc(CmpOpConst),
CmpOpConst->getValueType(0));
} else {
ExpectedOpcode = ISD::SUB;
ExpectedOp = CmpOpToMatch;
SubsOp = CmpOpToMatch;
}
// Get the operand that can be reassociated with the SUBS instruction.
auto GetReassociationOp = [&](SDValue Op, SDValue ExpectedOp) {
if (Op.getOpcode() != ExpectedOpcode)
return SDValue();
if (Op.getOperand(0).getOpcode() != ISD::ADD ||
!Op.getOperand(0).hasOneUse())
return SDValue();
SDValue X = Op.getOperand(0).getOperand(0);
SDValue Y = Op.getOperand(0).getOperand(1);
if (X != CmpOpOther)
std::swap(X, Y);
if (X != CmpOpOther)
return SDValue();
if (ExpectedOp != Op.getOperand(1))
return SDValue();
return Y;
};
// Try the reassociation using the given constant and condition code.
auto Fold = [&](AArch64CC::CondCode NewCC, SDValue ExpectedOp,
SDValue SubsOp) {
SDValue TReassocOp = GetReassociationOp(N->getOperand(0), ExpectedOp);
SDValue FReassocOp = GetReassociationOp(N->getOperand(1), ExpectedOp);
if (!TReassocOp && !FReassocOp)
return SDValue();
SDValue NewCmp = DAG.getNode(AArch64ISD::SUBS, SDLoc(SubsNode),
DAG.getVTList(VT, MVT_CC), CmpOpOther, SubsOp);
auto Reassociate = [&](SDValue ReassocOp, unsigned OpNum) {
if (!ReassocOp)
return N->getOperand(OpNum);
SDValue Res = DAG.getNode(ISD::ADD, SDLoc(N->getOperand(OpNum)), VT,
NewCmp.getValue(0), ReassocOp);
DAG.ReplaceAllUsesWith(N->getOperand(OpNum), Res);
return Res;
};
SDValue TValReassoc = Reassociate(TReassocOp, 0);
SDValue FValReassoc = Reassociate(FReassocOp, 1);
return DAG.getNode(AArch64ISD::CSEL, SDLoc(N), VT, TValReassoc, FValReassoc,
DAG.getConstant(NewCC, SDLoc(N->getOperand(2)), MVT_CC),
NewCmp.getValue(1));
};
auto CC = static_cast<AArch64CC::CondCode>(N->getConstantOperandVal(2));
// First, try to eliminate the compare instruction by searching for a
// subtraction with the same constant.
if (SDValue R = Fold(CC, ExpectedOp, SubsOp))
return R;
if (!CmpOpConst) {
// Try again with the operands of the SUBS instruction and the condition
// swapped. Due to canonicalization, this only helps for non-constant
// operands of the SUBS instruction.
std::swap(CmpOpToMatch, CmpOpOther);
if (SDValue R = Fold(getSwappedCondition(CC), CmpOpToMatch, CmpOpToMatch))
return R;
return SDValue();
}
if ((CC == AArch64CC::EQ || CC == AArch64CC::NE) && !CmpOpConst->isZero())
return SDValue();
// Next, search for a subtraction with a slightly different constant. By
// adjusting the condition code, we can still eliminate the compare
// instruction. Adjusting the constant is only valid if it does not result
// in signed/unsigned wrap for signed/unsigned comparisons, respectively.
// Since such comparisons are trivially true/false, we should not encounter
// them here but check for them nevertheless to be on the safe side.
auto CheckedFold = [&](bool Check, APInt NewCmpConst,
AArch64CC::CondCode NewCC) {
auto ExpectedOp = DAG.getConstant(-NewCmpConst, SDLoc(CmpOpConst),
CmpOpConst->getValueType(0));
auto SubsOp = DAG.getConstant(NewCmpConst, SDLoc(CmpOpConst),
CmpOpConst->getValueType(0));
return Check ? Fold(NewCC, ExpectedOp, SubsOp) : SDValue();
};
switch (CC) {
case AArch64CC::EQ:
case AArch64CC::LS:
return CheckedFold(!CmpOpConst->getAPIntValue().isMaxValue(),
CmpOpConst->getAPIntValue() + 1, AArch64CC::LO);
case AArch64CC::NE:
case AArch64CC::HI:
return CheckedFold(!CmpOpConst->getAPIntValue().isMaxValue(),
CmpOpConst->getAPIntValue() + 1, AArch64CC::HS);
case AArch64CC::LO:
return CheckedFold(!CmpOpConst->getAPIntValue().isZero(),
CmpOpConst->getAPIntValue() - 1, AArch64CC::LS);
case AArch64CC::HS:
return CheckedFold(!CmpOpConst->getAPIntValue().isZero(),
CmpOpConst->getAPIntValue() - 1, AArch64CC::HI);
case AArch64CC::LT:
return CheckedFold(!CmpOpConst->getAPIntValue().isMinSignedValue(),
CmpOpConst->getAPIntValue() - 1, AArch64CC::LE);
case AArch64CC::LE:
return CheckedFold(!CmpOpConst->getAPIntValue().isMaxSignedValue(),
CmpOpConst->getAPIntValue() + 1, AArch64CC::LT);
case AArch64CC::GT:
return CheckedFold(!CmpOpConst->getAPIntValue().isMaxSignedValue(),
CmpOpConst->getAPIntValue() + 1, AArch64CC::GE);
case AArch64CC::GE:
return CheckedFold(!CmpOpConst->getAPIntValue().isMinSignedValue(),
CmpOpConst->getAPIntValue() - 1, AArch64CC::GT);
default:
return SDValue();
}
}
static SDValue foldCSELofLASTB(SDNode *Op, SelectionDAG &DAG) {
AArch64CC::CondCode OpCC =
static_cast<AArch64CC::CondCode>(Op->getConstantOperandVal(2));
if (OpCC != AArch64CC::NE)
return SDValue();
SDValue PTest = Op->getOperand(3);
if (PTest.getOpcode() != AArch64ISD::PTEST_ANY)
return SDValue();
SDValue TruePred = PTest.getOperand(0);
SDValue AnyPred = PTest.getOperand(1);
if (TruePred.getOpcode() == AArch64ISD::REINTERPRET_CAST)
TruePred = TruePred.getOperand(0);
if (AnyPred.getOpcode() == AArch64ISD::REINTERPRET_CAST)
AnyPred = AnyPred.getOperand(0);
if (TruePred != AnyPred && !isAllActivePredicate(DAG, TruePred))
return SDValue();
SDValue LastB = Op->getOperand(0);
SDValue Default = Op->getOperand(1);
if (LastB.getOpcode() != AArch64ISD::LASTB || LastB.getOperand(0) != AnyPred)
return SDValue();
return DAG.getNode(AArch64ISD::CLASTB_N, SDLoc(Op), Op->getValueType(0),
AnyPred, Default, LastB.getOperand(1));
}
// Optimize CSEL instructions
static SDValue performCSELCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
// CSEL x, x, cc -> x
if (N->getOperand(0) == N->getOperand(1))
return N->getOperand(0);
if (SDValue R = foldCSELOfCSEL(N, DAG))
return R;
// Try to reassociate the true/false expressions so that we can do CSE with
// a SUBS instruction used to perform the comparison.
if (SDValue R = reassociateCSELOperandsForCSE(N, DAG))
return R;
// CSEL 0, cttz(X), eq(X, 0) -> AND cttz bitwidth-1
// CSEL cttz(X), 0, ne(X, 0) -> AND cttz bitwidth-1
if (SDValue Folded = foldCSELofCTTZ(N, DAG))
return Folded;
// CSEL a, b, cc, SUBS(x, y) -> CSEL a, b, swapped(cc), SUBS(y, x)
// if SUB(y, x) already exists and we can produce a swapped predicate for cc.
SDValue Cond = N->getOperand(3);
if (DCI.isAfterLegalizeDAG() && Cond.getOpcode() == AArch64ISD::SUBS &&
Cond.hasOneUse() && Cond->hasNUsesOfValue(0, 0) &&
DAG.doesNodeExist(ISD::SUB, N->getVTList(),
{Cond.getOperand(1), Cond.getOperand(0)}) &&
!DAG.doesNodeExist(ISD::SUB, N->getVTList(),
{Cond.getOperand(0), Cond.getOperand(1)}) &&
!isNullConstant(Cond.getOperand(1))) {
AArch64CC::CondCode OldCond =
static_cast<AArch64CC::CondCode>(N->getConstantOperandVal(2));
AArch64CC::CondCode NewCond = getSwappedCondition(OldCond);
if (NewCond != AArch64CC::AL) {
SDLoc DL(N);
SDValue Sub = DAG.getNode(AArch64ISD::SUBS, DL, Cond->getVTList(),
Cond.getOperand(1), Cond.getOperand(0));
return DAG.getNode(AArch64ISD::CSEL, DL, N->getVTList(), N->getOperand(0),
N->getOperand(1),
DAG.getConstant(NewCond, DL, MVT::i32),
Sub.getValue(1));
}
}
// CSEL (LASTB P, Z), X, NE(ANY P) -> CLASTB P, X, Z
if (SDValue CondLast = foldCSELofLASTB(N, DAG))
return CondLast;
return performCONDCombine(N, DCI, DAG, 2, 3);
}
// Try to re-use an already extended operand of a vector SetCC feeding a
// extended select. Doing so avoids requiring another full extension of the
// SET_CC result when lowering the select.
static SDValue tryToWidenSetCCOperands(SDNode *Op, SelectionDAG &DAG) {
EVT Op0MVT = Op->getOperand(0).getValueType();
if (!Op0MVT.isVector() || Op->use_empty())
return SDValue();
// Make sure that all uses of Op are VSELECTs with result matching types where
// the result type has a larger element type than the SetCC operand.
SDNode *FirstUse = *Op->user_begin();
if (FirstUse->getOpcode() != ISD::VSELECT)
return SDValue();
EVT UseMVT = FirstUse->getValueType(0);
if (UseMVT.getScalarSizeInBits() <= Op0MVT.getScalarSizeInBits())
return SDValue();
if (any_of(Op->users(), [&UseMVT](const SDNode *N) {
return N->getOpcode() != ISD::VSELECT || N->getValueType(0) != UseMVT;
}))
return SDValue();
APInt V;
if (!ISD::isConstantSplatVector(Op->getOperand(1).getNode(), V))
return SDValue();
SDLoc DL(Op);
SDValue Op0ExtV;
SDValue Op1ExtV;
ISD::CondCode CC = cast<CondCodeSDNode>(Op->getOperand(2))->get();
// Check if the first operand of the SET_CC is already extended. If it is,
// split the SET_CC and re-use the extended version of the operand.
SDNode *Op0SExt = DAG.getNodeIfExists(ISD::SIGN_EXTEND, DAG.getVTList(UseMVT),
Op->getOperand(0));
SDNode *Op0ZExt = DAG.getNodeIfExists(ISD::ZERO_EXTEND, DAG.getVTList(UseMVT),
Op->getOperand(0));
if (Op0SExt && (isSignedIntSetCC(CC) || isIntEqualitySetCC(CC))) {
Op0ExtV = SDValue(Op0SExt, 0);
Op1ExtV = DAG.getNode(ISD::SIGN_EXTEND, DL, UseMVT, Op->getOperand(1));
} else if (Op0ZExt && (isUnsignedIntSetCC(CC) || isIntEqualitySetCC(CC))) {
Op0ExtV = SDValue(Op0ZExt, 0);
Op1ExtV = DAG.getNode(ISD::ZERO_EXTEND, DL, UseMVT, Op->getOperand(1));
} else
return SDValue();
return DAG.getNode(ISD::SETCC, DL, UseMVT.changeVectorElementType(MVT::i1),
Op0ExtV, Op1ExtV, Op->getOperand(2));
}
static SDValue
performVecReduceBitwiseCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
SDValue Vec = N->getOperand(0);
if (DCI.isBeforeLegalize() &&
Vec.getValueType().getVectorElementType() == MVT::i1 &&
Vec.getValueType().isFixedLengthVector() &&
Vec.getValueType().isPow2VectorType()) {
SDLoc DL(N);
return getVectorBitwiseReduce(N->getOpcode(), Vec, N->getValueType(0), DL,
DAG);
}
return SDValue();
}
static SDValue performSETCCCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
assert(N->getOpcode() == ISD::SETCC && "Unexpected opcode!");
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
ISD::CondCode Cond = cast<CondCodeSDNode>(N->getOperand(2))->get();
SDLoc DL(N);
EVT VT = N->getValueType(0);
if (SDValue V = tryToWidenSetCCOperands(N, DAG))
return V;
// setcc (csel 0, 1, cond, X), 1, ne ==> csel 0, 1, !cond, X
if (Cond == ISD::SETNE && isOneConstant(RHS) &&
LHS->getOpcode() == AArch64ISD::CSEL &&
isNullConstant(LHS->getOperand(0)) && isOneConstant(LHS->getOperand(1)) &&
LHS->hasOneUse()) {
// Invert CSEL's condition.
auto OldCond =
static_cast<AArch64CC::CondCode>(LHS.getConstantOperandVal(2));
auto NewCond = getInvertedCondCode(OldCond);
// csel 0, 1, !cond, X
SDValue CSEL =
DAG.getNode(AArch64ISD::CSEL, DL, LHS.getValueType(), LHS.getOperand(0),
LHS.getOperand(1), DAG.getConstant(NewCond, DL, MVT::i32),
LHS.getOperand(3));
return DAG.getZExtOrTrunc(CSEL, DL, VT);
}
// setcc (srl x, imm), 0, ne ==> setcc (and x, (-1 << imm)), 0, ne
if (Cond == ISD::SETNE && isNullConstant(RHS) &&
LHS->getOpcode() == ISD::SRL && isa<ConstantSDNode>(LHS->getOperand(1)) &&
LHS->hasOneUse()) {
EVT TstVT = LHS->getValueType(0);
if (TstVT.isScalarInteger() && TstVT.getFixedSizeInBits() <= 64 &&
LHS->getConstantOperandVal(1) < TstVT.getFixedSizeInBits()) {
// this pattern will get better opt in emitComparison
uint64_t TstImm = -1ULL << LHS->getConstantOperandVal(1);
SDValue TST = DAG.getNode(ISD::AND, DL, TstVT, LHS->getOperand(0),
DAG.getSignedConstant(TstImm, DL, TstVT));
return DAG.getNode(ISD::SETCC, DL, VT, TST, RHS, N->getOperand(2));
}
}
// setcc (iN (bitcast (vNi1 X))), 0, (eq|ne)
// ==> setcc (iN (zext (i1 (vecreduce_or (vNi1 X))))), 0, (eq|ne)
// setcc (iN (bitcast (vNi1 X))), -1, (eq|ne)
// ==> setcc (iN (sext (i1 (vecreduce_and (vNi1 X))))), -1, (eq|ne)
if (DCI.isBeforeLegalize() && VT.isScalarInteger() &&
(Cond == ISD::SETEQ || Cond == ISD::SETNE) &&
(isNullConstant(RHS) || isAllOnesConstant(RHS)) &&
LHS->getOpcode() == ISD::BITCAST) {
EVT ToVT = LHS->getValueType(0);
EVT FromVT = LHS->getOperand(0).getValueType();
if (FromVT.isFixedLengthVector() &&
FromVT.getVectorElementType() == MVT::i1) {
bool IsNull = isNullConstant(RHS);
LHS = DAG.getNode(IsNull ? ISD::VECREDUCE_OR : ISD::VECREDUCE_AND,
DL, MVT::i1, LHS->getOperand(0));
LHS = DAG.getNode(IsNull ? ISD::ZERO_EXTEND : ISD::SIGN_EXTEND, DL, ToVT,
LHS);
return DAG.getSetCC(DL, VT, LHS, RHS, Cond);
}
}
// Try to perform the memcmp when the result is tested for [in]equality with 0
if (SDValue V = performOrXorChainCombine(N, DAG))
return V;
EVT CmpVT = LHS.getValueType();
// NOTE: This exists as a combine only because it proved too awkward to match
// splat(1) across all the NEON types during isel.
APInt SplatLHSVal;
if (CmpVT.isInteger() && Cond == ISD::SETGT &&
ISD::isConstantSplatVector(LHS.getNode(), SplatLHSVal) &&
SplatLHSVal.isOne())
return DAG.getSetCC(DL, VT, DAG.getConstant(0, DL, CmpVT), RHS, ISD::SETGE);
return SDValue();
}
// Replace a flag-setting operator (eg ANDS) with the generic version
// (eg AND) if the flag is unused.
static SDValue performFlagSettingCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
unsigned GenericOpcode) {
SDLoc DL(N);
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
EVT VT = N->getValueType(0);
// If the flag result isn't used, convert back to a generic opcode.
if (!N->hasAnyUseOfValue(1)) {
SDValue Res = DCI.DAG.getNode(GenericOpcode, DL, VT, N->ops());
return DCI.DAG.getMergeValues({Res, DCI.DAG.getConstant(0, DL, MVT::i32)},
DL);
}
// Combine identical generic nodes into this node, re-using the result.
if (SDNode *Generic = DCI.DAG.getNodeIfExists(
GenericOpcode, DCI.DAG.getVTList(VT), {LHS, RHS}))
DCI.CombineTo(Generic, SDValue(N, 0));
return SDValue();
}
static SDValue performSetCCPunpkCombine(SDNode *N, SelectionDAG &DAG) {
// setcc_merge_zero pred
// (sign_extend (extract_subvector (setcc_merge_zero ... pred ...))), 0, ne
// => extract_subvector (inner setcc_merge_zero)
SDValue Pred = N->getOperand(0);
SDValue LHS = N->getOperand(1);
SDValue RHS = N->getOperand(2);
ISD::CondCode Cond = cast<CondCodeSDNode>(N->getOperand(3))->get();
if (Cond != ISD::SETNE || !isZerosVector(RHS.getNode()) ||
LHS->getOpcode() != ISD::SIGN_EXTEND)
return SDValue();
SDValue Extract = LHS->getOperand(0);
if (Extract->getOpcode() != ISD::EXTRACT_SUBVECTOR ||
Extract->getValueType(0) != N->getValueType(0) ||
Extract->getConstantOperandVal(1) != 0)
return SDValue();
SDValue InnerSetCC = Extract->getOperand(0);
if (InnerSetCC->getOpcode() != AArch64ISD::SETCC_MERGE_ZERO)
return SDValue();
// By this point we've effectively got
// zero_inactive_lanes_and_trunc_i1(sext_i1(A)). If we can prove A's inactive
// lanes are already zero then the trunc(sext()) sequence is redundant and we
// can operate on A directly.
SDValue InnerPred = InnerSetCC.getOperand(0);
if (Pred.getOpcode() == AArch64ISD::PTRUE &&
InnerPred.getOpcode() == AArch64ISD::PTRUE &&
Pred.getConstantOperandVal(0) == InnerPred.getConstantOperandVal(0) &&
Pred->getConstantOperandVal(0) >= AArch64SVEPredPattern::vl1 &&
Pred->getConstantOperandVal(0) <= AArch64SVEPredPattern::vl256)
return Extract;
return SDValue();
}
static SDValue
performSetccMergeZeroCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI) {
assert(N->getOpcode() == AArch64ISD::SETCC_MERGE_ZERO &&
"Unexpected opcode!");
SelectionDAG &DAG = DCI.DAG;
SDValue Pred = N->getOperand(0);
SDValue LHS = N->getOperand(1);
SDValue RHS = N->getOperand(2);
ISD::CondCode Cond = cast<CondCodeSDNode>(N->getOperand(3))->get();
if (SDValue V = performSetCCPunpkCombine(N, DAG))
return V;
if (Cond == ISD::SETNE && isZerosVector(RHS.getNode()) &&
LHS->getOpcode() == ISD::SIGN_EXTEND &&
LHS->getOperand(0)->getValueType(0) == N->getValueType(0)) {
// setcc_merge_zero(
// pred, extend(setcc_merge_zero(pred, ...)), != splat(0))
// => setcc_merge_zero(pred, ...)
if (LHS->getOperand(0)->getOpcode() == AArch64ISD::SETCC_MERGE_ZERO &&
LHS->getOperand(0)->getOperand(0) == Pred)
return LHS->getOperand(0);
// setcc_merge_zero(
// all_active, extend(nxvNi1 ...), != splat(0))
// -> nxvNi1 ...
if (isAllActivePredicate(DAG, Pred))
return LHS->getOperand(0);
// setcc_merge_zero(
// pred, extend(nxvNi1 ...), != splat(0))
// -> nxvNi1 and(pred, ...)
if (DCI.isAfterLegalizeDAG())
// Do this after legalization to allow more folds on setcc_merge_zero
// to be recognized.
return DAG.getNode(ISD::AND, SDLoc(N), N->getValueType(0),
LHS->getOperand(0), Pred);
}
return SDValue();
}
// Optimize some simple tbz/tbnz cases. Returns the new operand and bit to test
// as well as whether the test should be inverted. This code is required to
// catch these cases (as opposed to standard dag combines) because
// AArch64ISD::TBZ is matched during legalization.
static SDValue getTestBitOperand(SDValue Op, unsigned &Bit, bool &Invert,
SelectionDAG &DAG) {
if (!Op->hasOneUse())
return Op;
// We don't handle undef/constant-fold cases below, as they should have
// already been taken care of (e.g. and of 0, test of undefined shifted bits,
// etc.)
// (tbz (trunc x), b) -> (tbz x, b)
// This case is just here to enable more of the below cases to be caught.
if (Op->getOpcode() == ISD::TRUNCATE &&
Bit < Op->getValueType(0).getSizeInBits()) {
return getTestBitOperand(Op->getOperand(0), Bit, Invert, DAG);
}
// (tbz (any_ext x), b) -> (tbz x, b) if we don't use the extended bits.
if (Op->getOpcode() == ISD::ANY_EXTEND &&
Bit < Op->getOperand(0).getValueSizeInBits()) {
return getTestBitOperand(Op->getOperand(0), Bit, Invert, DAG);
}
if (Op->getNumOperands() != 2)
return Op;
auto *C = dyn_cast<ConstantSDNode>(Op->getOperand(1));
if (!C)
return Op;
switch (Op->getOpcode()) {
default:
return Op;
// (tbz (and x, m), b) -> (tbz x, b)
case ISD::AND:
if ((C->getZExtValue() >> Bit) & 1)
return getTestBitOperand(Op->getOperand(0), Bit, Invert, DAG);
return Op;
// (tbz (shl x, c), b) -> (tbz x, b-c)
case ISD::SHL:
if (C->getZExtValue() <= Bit &&
(Bit - C->getZExtValue()) < Op->getValueType(0).getSizeInBits()) {
Bit = Bit - C->getZExtValue();
return getTestBitOperand(Op->getOperand(0), Bit, Invert, DAG);
}
return Op;
// (tbz (sra x, c), b) -> (tbz x, b+c) or (tbz x, msb) if b+c is > # bits in x
case ISD::SRA:
Bit = Bit + C->getZExtValue();
if (Bit >= Op->getValueType(0).getSizeInBits())
Bit = Op->getValueType(0).getSizeInBits() - 1;
return getTestBitOperand(Op->getOperand(0), Bit, Invert, DAG);
// (tbz (srl x, c), b) -> (tbz x, b+c)
case ISD::SRL:
if ((Bit + C->getZExtValue()) < Op->getValueType(0).getSizeInBits()) {
Bit = Bit + C->getZExtValue();
return getTestBitOperand(Op->getOperand(0), Bit, Invert, DAG);
}
return Op;
// (tbz (xor x, -1), b) -> (tbnz x, b)
case ISD::XOR:
if ((C->getZExtValue() >> Bit) & 1)
Invert = !Invert;
return getTestBitOperand(Op->getOperand(0), Bit, Invert, DAG);
}
}
// Optimize test single bit zero/non-zero and branch.
static SDValue performTBZCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
unsigned Bit = N->getConstantOperandVal(2);
bool Invert = false;
SDValue TestSrc = N->getOperand(1);
SDValue NewTestSrc = getTestBitOperand(TestSrc, Bit, Invert, DAG);
if (TestSrc == NewTestSrc)
return SDValue();
unsigned NewOpc = N->getOpcode();
if (Invert) {
if (NewOpc == AArch64ISD::TBZ)
NewOpc = AArch64ISD::TBNZ;
else {
assert(NewOpc == AArch64ISD::TBNZ);
NewOpc = AArch64ISD::TBZ;
}
}
SDLoc DL(N);
return DAG.getNode(NewOpc, DL, MVT::Other, N->getOperand(0), NewTestSrc,
DAG.getConstant(Bit, DL, MVT::i64), N->getOperand(3));
}
// Swap vselect operands where it may allow a predicated operation to achieve
// the `sel`.
//
// (vselect (setcc ( condcode) (_) (_)) (a) (op (a) (b)))
// => (vselect (setcc (!condcode) (_) (_)) (op (a) (b)) (a))
static SDValue trySwapVSelectOperands(SDNode *N, SelectionDAG &DAG) {
auto SelectA = N->getOperand(1);
auto SelectB = N->getOperand(2);
auto NTy = N->getValueType(0);
if (!NTy.isScalableVector())
return SDValue();
SDValue SetCC = N->getOperand(0);
if (SetCC.getOpcode() != ISD::SETCC || !SetCC.hasOneUse())
return SDValue();
switch (SelectB.getOpcode()) {
default:
return SDValue();
case ISD::FMUL:
case ISD::FSUB:
case ISD::FADD:
break;
}
if (SelectA != SelectB.getOperand(0))
return SDValue();
ISD::CondCode CC = cast<CondCodeSDNode>(SetCC.getOperand(2))->get();
ISD::CondCode InverseCC =
ISD::getSetCCInverse(CC, SetCC.getOperand(0).getValueType());
auto InverseSetCC =
DAG.getSetCC(SDLoc(SetCC), SetCC.getValueType(), SetCC.getOperand(0),
SetCC.getOperand(1), InverseCC);
return DAG.getNode(ISD::VSELECT, SDLoc(N), NTy,
{InverseSetCC, SelectB, SelectA});
}
// vselect (v1i1 setcc) ->
// vselect (v1iXX setcc) (XX is the size of the compared operand type)
// FIXME: Currently the type legalizer can't handle VSELECT having v1i1 as
// condition. If it can legalize "VSELECT v1i1" correctly, no need to combine
// such VSELECT.
static SDValue performVSelectCombine(SDNode *N, SelectionDAG &DAG) {
if (auto SwapResult = trySwapVSelectOperands(N, DAG))
return SwapResult;
SDValue N0 = N->getOperand(0);
EVT CCVT = N0.getValueType();
if (isAllActivePredicate(DAG, N0))
return N->getOperand(1);
if (isAllInactivePredicate(N0))
return N->getOperand(2);
// Check for sign pattern (VSELECT setgt, iN lhs, -1, 1, -1) and transform
// into (OR (ASR lhs, N-1), 1), which requires less instructions for the
// supported types.
SDValue SetCC = N->getOperand(0);
if (SetCC.getOpcode() == ISD::SETCC &&
SetCC.getOperand(2) == DAG.getCondCode(ISD::SETGT)) {
SDValue CmpLHS = SetCC.getOperand(0);
EVT VT = CmpLHS.getValueType();
SDNode *CmpRHS = SetCC.getOperand(1).getNode();
SDNode *SplatLHS = N->getOperand(1).getNode();
SDNode *SplatRHS = N->getOperand(2).getNode();
APInt SplatLHSVal;
if (CmpLHS.getValueType() == N->getOperand(1).getValueType() &&
VT.isSimple() &&
is_contained(ArrayRef({MVT::v8i8, MVT::v16i8, MVT::v4i16, MVT::v8i16,
MVT::v2i32, MVT::v4i32, MVT::v2i64}),
VT.getSimpleVT().SimpleTy) &&
ISD::isConstantSplatVector(SplatLHS, SplatLHSVal) &&
SplatLHSVal.isOne() && ISD::isConstantSplatVectorAllOnes(CmpRHS) &&
ISD::isConstantSplatVectorAllOnes(SplatRHS)) {
unsigned NumElts = VT.getVectorNumElements();
SmallVector<SDValue, 8> Ops(
NumElts, DAG.getConstant(VT.getScalarSizeInBits() - 1, SDLoc(N),
VT.getScalarType()));
SDValue Val = DAG.getBuildVector(VT, SDLoc(N), Ops);
auto Shift = DAG.getNode(ISD::SRA, SDLoc(N), VT, CmpLHS, Val);
auto Or = DAG.getNode(ISD::OR, SDLoc(N), VT, Shift, N->getOperand(1));
return Or;
}
}
EVT CmpVT = N0.getOperand(0).getValueType();
if (N0.getOpcode() != ISD::SETCC ||
CCVT.getVectorElementCount() != ElementCount::getFixed(1) ||
CCVT.getVectorElementType() != MVT::i1 ||
CmpVT.getVectorElementType().isFloatingPoint())
return SDValue();
EVT ResVT = N->getValueType(0);
// Only combine when the result type is of the same size as the compared
// operands.
if (ResVT.getSizeInBits() != CmpVT.getSizeInBits())
return SDValue();
SDValue IfTrue = N->getOperand(1);
SDValue IfFalse = N->getOperand(2);
SetCC = DAG.getSetCC(SDLoc(N), CmpVT.changeVectorElementTypeToInteger(),
N0.getOperand(0), N0.getOperand(1),
cast<CondCodeSDNode>(N0.getOperand(2))->get());
return DAG.getNode(ISD::VSELECT, SDLoc(N), ResVT, SetCC,
IfTrue, IfFalse);
}
/// A vector select: "(select vL, vR, (setcc LHS, RHS))" is best performed with
/// the compare-mask instructions rather than going via NZCV, even if LHS and
/// RHS are really scalar. This replaces any scalar setcc in the above pattern
/// with a vector one followed by a DUP shuffle on the result.
static SDValue performSelectCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
SelectionDAG &DAG = DCI.DAG;
SDValue N0 = N->getOperand(0);
EVT ResVT = N->getValueType(0);
if (N0.getOpcode() != ISD::SETCC)
return SDValue();
if (ResVT.isScalableVT())
return SDValue();
// Make sure the SETCC result is either i1 (initial DAG), or i32, the lowered
// scalar SetCCResultType. We also don't expect vectors, because we assume
// that selects fed by vector SETCCs are canonicalized to VSELECT.
assert((N0.getValueType() == MVT::i1 || N0.getValueType() == MVT::i32) &&
"Scalar-SETCC feeding SELECT has unexpected result type!");
// If NumMaskElts == 0, the comparison is larger than select result. The
// largest real NEON comparison is 64-bits per lane, which means the result is
// at most 32-bits and an illegal vector. Just bail out for now.
EVT SrcVT = N0.getOperand(0).getValueType();
// Don't try to do this optimization when the setcc itself has i1 operands.
// There are no legal vectors of i1, so this would be pointless. v1f16 is
// ruled out to prevent the creation of setcc that need to be scalarized.
if (SrcVT == MVT::i1 ||
(SrcVT.isFloatingPoint() && SrcVT.getSizeInBits() <= 16))
return SDValue();
int NumMaskElts = ResVT.getSizeInBits() / SrcVT.getSizeInBits();
if (!ResVT.isVector() || NumMaskElts == 0)
return SDValue();
SrcVT = EVT::getVectorVT(*DAG.getContext(), SrcVT, NumMaskElts);
EVT CCVT = SrcVT.changeVectorElementTypeToInteger();
// Also bail out if the vector CCVT isn't the same size as ResVT.
// This can happen if the SETCC operand size doesn't divide the ResVT size
// (e.g., f64 vs v3f32).
if (CCVT.getSizeInBits() != ResVT.getSizeInBits())
return SDValue();
// Make sure we didn't create illegal types, if we're not supposed to.
assert(DCI.isBeforeLegalize() ||
DAG.getTargetLoweringInfo().isTypeLegal(SrcVT));
// First perform a vector comparison, where lane 0 is the one we're interested
// in.
SDLoc DL(N0);
SDValue LHS =
DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, SrcVT, N0.getOperand(0));
SDValue RHS =
DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, SrcVT, N0.getOperand(1));
SDValue SetCC = DAG.getNode(ISD::SETCC, DL, CCVT, LHS, RHS, N0.getOperand(2));
// Now duplicate the comparison mask we want across all other lanes.
SmallVector<int, 8> DUPMask(CCVT.getVectorNumElements(), 0);
SDValue Mask = DAG.getVectorShuffle(CCVT, DL, SetCC, SetCC, DUPMask);
Mask = DAG.getNode(ISD::BITCAST, DL,
ResVT.changeVectorElementTypeToInteger(), Mask);
return DAG.getSelect(DL, ResVT, Mask, N->getOperand(1), N->getOperand(2));
}
static SDValue performDUPCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI) {
EVT VT = N->getValueType(0);
SDLoc DL(N);
// If "v2i32 DUP(x)" and "v4i32 DUP(x)" both exist, use an extract from the
// 128bit vector version.
if (VT.is64BitVector() && DCI.isAfterLegalizeDAG()) {
EVT LVT = VT.getDoubleNumVectorElementsVT(*DCI.DAG.getContext());
SmallVector<SDValue> Ops(N->ops());
if (SDNode *LN = DCI.DAG.getNodeIfExists(N->getOpcode(),
DCI.DAG.getVTList(LVT), Ops)) {
return DCI.DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, SDValue(LN, 0),
DCI.DAG.getConstant(0, DL, MVT::i64));
}
}
if (N->getOpcode() == AArch64ISD::DUP) {
// If the instruction is known to produce a scalar in SIMD registers, we can
// duplicate it across the vector lanes using DUPLANE instead of moving it
// to a GPR first. For example, this allows us to handle:
// v4i32 = DUP (i32 (FCMGT (f32, f32)))
SDValue Op = N->getOperand(0);
// FIXME: Ideally, we should be able to handle all instructions that
// produce a scalar value in FPRs.
if (Op.getOpcode() == AArch64ISD::FCMEQ ||
Op.getOpcode() == AArch64ISD::FCMGE ||
Op.getOpcode() == AArch64ISD::FCMGT) {
EVT ElemVT = VT.getVectorElementType();
EVT ExpandedVT = VT;
// Insert into a 128-bit vector to match DUPLANE's pattern.
if (VT.getSizeInBits() != 128)
ExpandedVT = EVT::getVectorVT(*DCI.DAG.getContext(), ElemVT,
128 / ElemVT.getSizeInBits());
SDValue Zero = DCI.DAG.getConstant(0, DL, MVT::i64);
SDValue Vec = DCI.DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, ExpandedVT,
DCI.DAG.getUNDEF(ExpandedVT), Op, Zero);
return DCI.DAG.getNode(getDUPLANEOp(ElemVT), DL, VT, Vec, Zero);
}
if (DCI.isAfterLegalizeDAG()) {
// If scalar dup's operand is extract_vector_elt, try to combine them into
// duplane. For example,
//
// t21: i32 = extract_vector_elt t19, Constant:i64<0>
// t18: v4i32 = AArch64ISD::DUP t21
// ==>
// t22: v4i32 = AArch64ISD::DUPLANE32 t19, Constant:i64<0>
SDValue EXTRACT_VEC_ELT = N->getOperand(0);
if (EXTRACT_VEC_ELT.getOpcode() == ISD::EXTRACT_VECTOR_ELT) {
if (VT == EXTRACT_VEC_ELT.getOperand(0).getValueType()) {
unsigned Opcode = getDUPLANEOp(VT.getVectorElementType());
return DCI.DAG.getNode(Opcode, DL, VT, EXTRACT_VEC_ELT.getOperand(0),
EXTRACT_VEC_ELT.getOperand(1));
}
}
}
return performPostLD1Combine(N, DCI, false);
}
return SDValue();
}
/// Get rid of unnecessary NVCASTs (that don't change the type).
static SDValue performNVCASTCombine(SDNode *N, SelectionDAG &DAG) {
if (N->getValueType(0) == N->getOperand(0).getValueType())
return N->getOperand(0);
if (N->getOperand(0).getOpcode() == AArch64ISD::NVCAST)
return DAG.getNode(AArch64ISD::NVCAST, SDLoc(N), N->getValueType(0),
N->getOperand(0).getOperand(0));
return SDValue();
}
// If all users of the globaladdr are of the form (globaladdr + constant), find
// the smallest constant, fold it into the globaladdr's offset and rewrite the
// globaladdr as (globaladdr + constant) - constant.
static SDValue performGlobalAddressCombine(SDNode *N, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget,
const TargetMachine &TM) {
auto *GN = cast<GlobalAddressSDNode>(N);
if (Subtarget->ClassifyGlobalReference(GN->getGlobal(), TM) !=
AArch64II::MO_NO_FLAG)
return SDValue();
uint64_t MinOffset = -1ull;
for (SDNode *N : GN->users()) {
if (N->getOpcode() != ISD::ADD)
return SDValue();
auto *C = dyn_cast<ConstantSDNode>(N->getOperand(0));
if (!C)
C = dyn_cast<ConstantSDNode>(N->getOperand(1));
if (!C)
return SDValue();
MinOffset = std::min(MinOffset, C->getZExtValue());
}
uint64_t Offset = MinOffset + GN->getOffset();
// Require that the new offset is larger than the existing one. Otherwise, we
// can end up oscillating between two possible DAGs, for example,
// (add (add globaladdr + 10, -1), 1) and (add globaladdr + 9, 1).
if (Offset <= uint64_t(GN->getOffset()))
return SDValue();
// Check whether folding this offset is legal. It must not go out of bounds of
// the referenced object to avoid violating the code model, and must be
// smaller than 2^20 because this is the largest offset expressible in all
// object formats. (The IMAGE_REL_ARM64_PAGEBASE_REL21 relocation in COFF
// stores an immediate signed 21 bit offset.)
//
// This check also prevents us from folding negative offsets, which will end
// up being treated in the same way as large positive ones. They could also
// cause code model violations, and aren't really common enough to matter.
if (Offset >= (1 << 20))
return SDValue();
const GlobalValue *GV = GN->getGlobal();
Type *T = GV->getValueType();
if (!T->isSized() ||
Offset > GV->getDataLayout().getTypeAllocSize(T))
return SDValue();
SDLoc DL(GN);
SDValue Result = DAG.getGlobalAddress(GV, DL, MVT::i64, Offset);
return DAG.getNode(ISD::SUB, DL, MVT::i64, Result,
DAG.getConstant(MinOffset, DL, MVT::i64));
}
static SDValue performCTLZCombine(SDNode *N, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
SDValue BR = N->getOperand(0);
if (!Subtarget->hasCSSC() || BR.getOpcode() != ISD::BITREVERSE ||
!BR.getValueType().isScalarInteger())
return SDValue();
SDLoc DL(N);
return DAG.getNode(ISD::CTTZ, DL, BR.getValueType(), BR.getOperand(0));
}
// Turns the vector of indices into a vector of byte offstes by scaling Offset
// by (BitWidth / 8).
static SDValue getScaledOffsetForBitWidth(SelectionDAG &DAG, SDValue Offset,
SDLoc DL, unsigned BitWidth) {
assert(Offset.getValueType().isScalableVector() &&
"This method is only for scalable vectors of offsets");
SDValue Shift = DAG.getConstant(Log2_32(BitWidth / 8), DL, MVT::i64);
SDValue SplatShift = DAG.getNode(ISD::SPLAT_VECTOR, DL, MVT::nxv2i64, Shift);
return DAG.getNode(ISD::SHL, DL, MVT::nxv2i64, Offset, SplatShift);
}
/// Check if the value of \p OffsetInBytes can be used as an immediate for
/// the gather load/prefetch and scatter store instructions with vector base and
/// immediate offset addressing mode:
///
/// [<Zn>.[S|D]{, #<imm>}]
///
/// where <imm> = sizeof(<T>) * k, for k = 0, 1, ..., 31.
inline static bool isValidImmForSVEVecImmAddrMode(unsigned OffsetInBytes,
unsigned ScalarSizeInBytes) {
// The immediate is not a multiple of the scalar size.
if (OffsetInBytes % ScalarSizeInBytes)
return false;
// The immediate is out of range.
if (OffsetInBytes / ScalarSizeInBytes > 31)
return false;
return true;
}
/// Check if the value of \p Offset represents a valid immediate for the SVE
/// gather load/prefetch and scatter store instructiona with vector base and
/// immediate offset addressing mode:
///
/// [<Zn>.[S|D]{, #<imm>}]
///
/// where <imm> = sizeof(<T>) * k, for k = 0, 1, ..., 31.
static bool isValidImmForSVEVecImmAddrMode(SDValue Offset,
unsigned ScalarSizeInBytes) {
ConstantSDNode *OffsetConst = dyn_cast<ConstantSDNode>(Offset.getNode());
return OffsetConst && isValidImmForSVEVecImmAddrMode(
OffsetConst->getZExtValue(), ScalarSizeInBytes);
}
static SDValue performScatterStoreCombine(SDNode *N, SelectionDAG &DAG,
unsigned Opcode,
bool OnlyPackedOffsets = true) {
const SDValue Src = N->getOperand(2);
const EVT SrcVT = Src->getValueType(0);
assert(SrcVT.isScalableVector() &&
"Scatter stores are only possible for SVE vectors");
SDLoc DL(N);
MVT SrcElVT = SrcVT.getVectorElementType().getSimpleVT();
// Make sure that source data will fit into an SVE register
if (SrcVT.getSizeInBits().getKnownMinValue() > AArch64::SVEBitsPerBlock)
return SDValue();
// For FPs, ACLE only supports _packed_ single and double precision types.
// SST1Q_[INDEX_]PRED is the ST1Q for sve2p1 and should allow all sizes.
if (SrcElVT.isFloatingPoint())
if ((SrcVT != MVT::nxv4f32) && (SrcVT != MVT::nxv2f64) &&
((Opcode != AArch64ISD::SST1Q_PRED &&
Opcode != AArch64ISD::SST1Q_INDEX_PRED) ||
((SrcVT != MVT::nxv8f16) && (SrcVT != MVT::nxv8bf16))))
return SDValue();
// Depending on the addressing mode, this is either a pointer or a vector of
// pointers (that fits into one register)
SDValue Base = N->getOperand(4);
// Depending on the addressing mode, this is either a single offset or a
// vector of offsets (that fits into one register)
SDValue Offset = N->getOperand(5);
// For "scalar + vector of indices", just scale the indices. This only
// applies to non-temporal scatters because there's no instruction that takes
// indices.
if (Opcode == AArch64ISD::SSTNT1_INDEX_PRED) {
Offset =
getScaledOffsetForBitWidth(DAG, Offset, DL, SrcElVT.getSizeInBits());
Opcode = AArch64ISD::SSTNT1_PRED;
} else if (Opcode == AArch64ISD::SST1Q_INDEX_PRED) {
Offset =
getScaledOffsetForBitWidth(DAG, Offset, DL, SrcElVT.getSizeInBits());
Opcode = AArch64ISD::SST1Q_PRED;
}
// In the case of non-temporal gather loads there's only one SVE instruction
// per data-size: "scalar + vector", i.e.
// * stnt1{b|h|w|d} { z0.s }, p0/z, [z0.s, x0]
// Since we do have intrinsics that allow the arguments to be in a different
// order, we may need to swap them to match the spec.
if ((Opcode == AArch64ISD::SSTNT1_PRED || Opcode == AArch64ISD::SST1Q_PRED) &&
Offset.getValueType().isVector())
std::swap(Base, Offset);
// SST1_IMM requires that the offset is an immediate that is:
// * a multiple of #SizeInBytes,
// * in the range [0, 31 x #SizeInBytes],
// where #SizeInBytes is the size in bytes of the stored items. For
// immediates outside that range and non-immediate scalar offsets use SST1 or
// SST1_UXTW instead.
if (Opcode == AArch64ISD::SST1_IMM_PRED) {
if (!isValidImmForSVEVecImmAddrMode(Offset,
SrcVT.getScalarSizeInBits() / 8)) {
if (MVT::nxv4i32 == Base.getValueType().getSimpleVT().SimpleTy)
Opcode = AArch64ISD::SST1_UXTW_PRED;
else
Opcode = AArch64ISD::SST1_PRED;
std::swap(Base, Offset);
}
}
auto &TLI = DAG.getTargetLoweringInfo();
if (!TLI.isTypeLegal(Base.getValueType()))
return SDValue();
// Some scatter store variants allow unpacked offsets, but only as nxv2i32
// vectors. These are implicitly sign (sxtw) or zero (zxtw) extend to
// nxv2i64. Legalize accordingly.
if (!OnlyPackedOffsets &&
Offset.getValueType().getSimpleVT().SimpleTy == MVT::nxv2i32)
Offset = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::nxv2i64, Offset).getValue(0);
if (!TLI.isTypeLegal(Offset.getValueType()))
return SDValue();
// Source value type that is representable in hardware
EVT HwSrcVt = getSVEContainerType(SrcVT);
// Keep the original type of the input data to store - this is needed to be
// able to select the correct instruction, e.g. ST1B, ST1H, ST1W and ST1D. For
// FP values we want the integer equivalent, so just use HwSrcVt.
SDValue InputVT = DAG.getValueType(SrcVT);
if (SrcVT.isFloatingPoint())
InputVT = DAG.getValueType(HwSrcVt);
SDVTList VTs = DAG.getVTList(MVT::Other);
SDValue SrcNew;
if (Src.getValueType().isFloatingPoint())
SrcNew = DAG.getNode(ISD::BITCAST, DL, HwSrcVt, Src);
else
SrcNew = DAG.getNode(ISD::ANY_EXTEND, DL, HwSrcVt, Src);
SDValue Ops[] = {N->getOperand(0), // Chain
SrcNew,
N->getOperand(3), // Pg
Base,
Offset,
InputVT};
return DAG.getNode(Opcode, DL, VTs, Ops);
}
static SDValue performGatherLoadCombine(SDNode *N, SelectionDAG &DAG,
unsigned Opcode,
bool OnlyPackedOffsets = true) {
const EVT RetVT = N->getValueType(0);
assert(RetVT.isScalableVector() &&
"Gather loads are only possible for SVE vectors");
SDLoc DL(N);
// Make sure that the loaded data will fit into an SVE register
if (RetVT.getSizeInBits().getKnownMinValue() > AArch64::SVEBitsPerBlock)
return SDValue();
// Depending on the addressing mode, this is either a pointer or a vector of
// pointers (that fits into one register)
SDValue Base = N->getOperand(3);
// Depending on the addressing mode, this is either a single offset or a
// vector of offsets (that fits into one register)
SDValue Offset = N->getOperand(4);
// For "scalar + vector of indices", scale the indices to obtain unscaled
// offsets. This applies to non-temporal and quadword gathers, which do not
// have an addressing mode with scaled offset.
if (Opcode == AArch64ISD::GLDNT1_INDEX_MERGE_ZERO) {
Offset = getScaledOffsetForBitWidth(DAG, Offset, DL,
RetVT.getScalarSizeInBits());
Opcode = AArch64ISD::GLDNT1_MERGE_ZERO;
} else if (Opcode == AArch64ISD::GLD1Q_INDEX_MERGE_ZERO) {
Offset = getScaledOffsetForBitWidth(DAG, Offset, DL,
RetVT.getScalarSizeInBits());
Opcode = AArch64ISD::GLD1Q_MERGE_ZERO;
}
// In the case of non-temporal gather loads and quadword gather loads there's
// only one addressing mode : "vector + scalar", e.g.
// ldnt1{b|h|w|d} { z0.s }, p0/z, [z0.s, x0]
// Since we do have intrinsics that allow the arguments to be in a different
// order, we may need to swap them to match the spec.
if ((Opcode == AArch64ISD::GLDNT1_MERGE_ZERO ||
Opcode == AArch64ISD::GLD1Q_MERGE_ZERO) &&
Offset.getValueType().isVector())
std::swap(Base, Offset);
// GLD{FF}1_IMM requires that the offset is an immediate that is:
// * a multiple of #SizeInBytes,
// * in the range [0, 31 x #SizeInBytes],
// where #SizeInBytes is the size in bytes of the loaded items. For
// immediates outside that range and non-immediate scalar offsets use
// GLD1_MERGE_ZERO or GLD1_UXTW_MERGE_ZERO instead.
if (Opcode == AArch64ISD::GLD1_IMM_MERGE_ZERO ||
Opcode == AArch64ISD::GLDFF1_IMM_MERGE_ZERO) {
if (!isValidImmForSVEVecImmAddrMode(Offset,
RetVT.getScalarSizeInBits() / 8)) {
if (MVT::nxv4i32 == Base.getValueType().getSimpleVT().SimpleTy)
Opcode = (Opcode == AArch64ISD::GLD1_IMM_MERGE_ZERO)
? AArch64ISD::GLD1_UXTW_MERGE_ZERO
: AArch64ISD::GLDFF1_UXTW_MERGE_ZERO;
else
Opcode = (Opcode == AArch64ISD::GLD1_IMM_MERGE_ZERO)
? AArch64ISD::GLD1_MERGE_ZERO
: AArch64ISD::GLDFF1_MERGE_ZERO;
std::swap(Base, Offset);
}
}
auto &TLI = DAG.getTargetLoweringInfo();
if (!TLI.isTypeLegal(Base.getValueType()))
return SDValue();
// Some gather load variants allow unpacked offsets, but only as nxv2i32
// vectors. These are implicitly sign (sxtw) or zero (zxtw) extend to
// nxv2i64. Legalize accordingly.
if (!OnlyPackedOffsets &&
Offset.getValueType().getSimpleVT().SimpleTy == MVT::nxv2i32)
Offset = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::nxv2i64, Offset).getValue(0);
// Return value type that is representable in hardware
EVT HwRetVt = getSVEContainerType(RetVT);
// Keep the original output value type around - this is needed to be able to
// select the correct instruction, e.g. LD1B, LD1H, LD1W and LD1D. For FP
// values we want the integer equivalent, so just use HwRetVT.
SDValue OutVT = DAG.getValueType(RetVT);
if (RetVT.isFloatingPoint())
OutVT = DAG.getValueType(HwRetVt);
SDVTList VTs = DAG.getVTList(HwRetVt, MVT::Other);
SDValue Ops[] = {N->getOperand(0), // Chain
N->getOperand(2), // Pg
Base, Offset, OutVT};
SDValue Load = DAG.getNode(Opcode, DL, VTs, Ops);
SDValue LoadChain = SDValue(Load.getNode(), 1);
if (RetVT.isInteger() && (RetVT != HwRetVt))
Load = DAG.getNode(ISD::TRUNCATE, DL, RetVT, Load.getValue(0));
// If the original return value was FP, bitcast accordingly. Doing it here
// means that we can avoid adding TableGen patterns for FPs.
if (RetVT.isFloatingPoint())
Load = DAG.getNode(ISD::BITCAST, DL, RetVT, Load.getValue(0));
return DAG.getMergeValues({Load, LoadChain}, DL);
}
static SDValue
performSignExtendInRegCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Src = N->getOperand(0);
unsigned Opc = Src->getOpcode();
// Sign extend of an unsigned unpack -> signed unpack
if (Opc == AArch64ISD::UUNPKHI || Opc == AArch64ISD::UUNPKLO) {
unsigned SOpc = Opc == AArch64ISD::UUNPKHI ? AArch64ISD::SUNPKHI
: AArch64ISD::SUNPKLO;
// Push the sign extend to the operand of the unpack
// This is necessary where, for example, the operand of the unpack
// is another unpack:
// 4i32 sign_extend_inreg (4i32 uunpklo(8i16 uunpklo (16i8 opnd)), from 4i8)
// ->
// 4i32 sunpklo (8i16 sign_extend_inreg(8i16 uunpklo (16i8 opnd), from 8i8)
// ->
// 4i32 sunpklo(8i16 sunpklo(16i8 opnd))
SDValue ExtOp = Src->getOperand(0);
auto VT = cast<VTSDNode>(N->getOperand(1))->getVT();
EVT EltTy = VT.getVectorElementType();
(void)EltTy;
assert((EltTy == MVT::i8 || EltTy == MVT::i16 || EltTy == MVT::i32) &&
"Sign extending from an invalid type");
EVT ExtVT = VT.getDoubleNumVectorElementsVT(*DAG.getContext());
SDValue Ext = DAG.getNode(ISD::SIGN_EXTEND_INREG, DL, ExtOp.getValueType(),
ExtOp, DAG.getValueType(ExtVT));
return DAG.getNode(SOpc, DL, N->getValueType(0), Ext);
}
if (DCI.isBeforeLegalizeOps())
return SDValue();
if (!EnableCombineMGatherIntrinsics)
return SDValue();
// SVE load nodes (e.g. AArch64ISD::GLD1) are straightforward candidates
// for DAG Combine with SIGN_EXTEND_INREG. Bail out for all other nodes.
unsigned NewOpc;
unsigned MemVTOpNum = 4;
switch (Opc) {
case AArch64ISD::LD1_MERGE_ZERO:
NewOpc = AArch64ISD::LD1S_MERGE_ZERO;
MemVTOpNum = 3;
break;
case AArch64ISD::LDNF1_MERGE_ZERO:
NewOpc = AArch64ISD::LDNF1S_MERGE_ZERO;
MemVTOpNum = 3;
break;
case AArch64ISD::LDFF1_MERGE_ZERO:
NewOpc = AArch64ISD::LDFF1S_MERGE_ZERO;
MemVTOpNum = 3;
break;
case AArch64ISD::GLD1_MERGE_ZERO:
NewOpc = AArch64ISD::GLD1S_MERGE_ZERO;
break;
case AArch64ISD::GLD1_SCALED_MERGE_ZERO:
NewOpc = AArch64ISD::GLD1S_SCALED_MERGE_ZERO;
break;
case AArch64ISD::GLD1_SXTW_MERGE_ZERO:
NewOpc = AArch64ISD::GLD1S_SXTW_MERGE_ZERO;
break;
case AArch64ISD::GLD1_SXTW_SCALED_MERGE_ZERO:
NewOpc = AArch64ISD::GLD1S_SXTW_SCALED_MERGE_ZERO;
break;
case AArch64ISD::GLD1_UXTW_MERGE_ZERO:
NewOpc = AArch64ISD::GLD1S_UXTW_MERGE_ZERO;
break;
case AArch64ISD::GLD1_UXTW_SCALED_MERGE_ZERO:
NewOpc = AArch64ISD::GLD1S_UXTW_SCALED_MERGE_ZERO;
break;
case AArch64ISD::GLD1_IMM_MERGE_ZERO:
NewOpc = AArch64ISD::GLD1S_IMM_MERGE_ZERO;
break;
case AArch64ISD::GLDFF1_MERGE_ZERO:
NewOpc = AArch64ISD::GLDFF1S_MERGE_ZERO;
break;
case AArch64ISD::GLDFF1_SCALED_MERGE_ZERO:
NewOpc = AArch64ISD::GLDFF1S_SCALED_MERGE_ZERO;
break;
case AArch64ISD::GLDFF1_SXTW_MERGE_ZERO:
NewOpc = AArch64ISD::GLDFF1S_SXTW_MERGE_ZERO;
break;
case AArch64ISD::GLDFF1_SXTW_SCALED_MERGE_ZERO:
NewOpc = AArch64ISD::GLDFF1S_SXTW_SCALED_MERGE_ZERO;
break;
case AArch64ISD::GLDFF1_UXTW_MERGE_ZERO:
NewOpc = AArch64ISD::GLDFF1S_UXTW_MERGE_ZERO;
break;
case AArch64ISD::GLDFF1_UXTW_SCALED_MERGE_ZERO:
NewOpc = AArch64ISD::GLDFF1S_UXTW_SCALED_MERGE_ZERO;
break;
case AArch64ISD::GLDFF1_IMM_MERGE_ZERO:
NewOpc = AArch64ISD::GLDFF1S_IMM_MERGE_ZERO;
break;
case AArch64ISD::GLDNT1_MERGE_ZERO:
NewOpc = AArch64ISD::GLDNT1S_MERGE_ZERO;
break;
default:
return SDValue();
}
EVT SignExtSrcVT = cast<VTSDNode>(N->getOperand(1))->getVT();
EVT SrcMemVT = cast<VTSDNode>(Src->getOperand(MemVTOpNum))->getVT();
if ((SignExtSrcVT != SrcMemVT) || !Src.hasOneUse())
return SDValue();
EVT DstVT = N->getValueType(0);
SDVTList VTs = DAG.getVTList(DstVT, MVT::Other);
SmallVector<SDValue, 5> Ops;
for (unsigned I = 0; I < Src->getNumOperands(); ++I)
Ops.push_back(Src->getOperand(I));
SDValue ExtLoad = DAG.getNode(NewOpc, SDLoc(N), VTs, Ops);
DCI.CombineTo(N, ExtLoad);
DCI.CombineTo(Src.getNode(), ExtLoad, ExtLoad.getValue(1));
// Return N so it doesn't get rechecked
return SDValue(N, 0);
}
/// Legalize the gather prefetch (scalar + vector addressing mode) when the
/// offset vector is an unpacked 32-bit scalable vector. The other cases (Offset
/// != nxv2i32) do not need legalization.
static SDValue legalizeSVEGatherPrefetchOffsVec(SDNode *N, SelectionDAG &DAG) {
const unsigned OffsetPos = 4;
SDValue Offset = N->getOperand(OffsetPos);
// Not an unpacked vector, bail out.
if (Offset.getValueType().getSimpleVT().SimpleTy != MVT::nxv2i32)
return SDValue();
// Extend the unpacked offset vector to 64-bit lanes.
SDLoc DL(N);
Offset = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::nxv2i64, Offset);
SmallVector<SDValue, 5> Ops(N->ops());
// Replace the offset operand with the 64-bit one.
Ops[OffsetPos] = Offset;
return DAG.getNode(N->getOpcode(), DL, DAG.getVTList(MVT::Other), Ops);
}
/// Combines a node carrying the intrinsic
/// `aarch64_sve_prf<T>_gather_scalar_offset` into a node that uses
/// `aarch64_sve_prfb_gather_uxtw_index` when the scalar offset passed to
/// `aarch64_sve_prf<T>_gather_scalar_offset` is not a valid immediate for the
/// sve gather prefetch instruction with vector plus immediate addressing mode.
static SDValue combineSVEPrefetchVecBaseImmOff(SDNode *N, SelectionDAG &DAG,
unsigned ScalarSizeInBytes) {
const unsigned ImmPos = 4, OffsetPos = 3;
// No need to combine the node if the immediate is valid...
if (isValidImmForSVEVecImmAddrMode(N->getOperand(ImmPos), ScalarSizeInBytes))
return SDValue();
// ...otherwise swap the offset base with the offset...
SmallVector<SDValue, 5> Ops(N->ops());
std::swap(Ops[ImmPos], Ops[OffsetPos]);
// ...and remap the intrinsic `aarch64_sve_prf<T>_gather_scalar_offset` to
// `aarch64_sve_prfb_gather_uxtw_index`.
SDLoc DL(N);
Ops[1] = DAG.getConstant(Intrinsic::aarch64_sve_prfb_gather_uxtw_index, DL,
MVT::i64);
return DAG.getNode(N->getOpcode(), DL, DAG.getVTList(MVT::Other), Ops);
}
// Return true if the vector operation can guarantee only the first lane of its
// result contains data, with all bits in other lanes set to zero.
static bool isLanes1toNKnownZero(SDValue Op) {
switch (Op.getOpcode()) {
default:
return false;
case AArch64ISD::ANDV_PRED:
case AArch64ISD::EORV_PRED:
case AArch64ISD::FADDA_PRED:
case AArch64ISD::FADDV_PRED:
case AArch64ISD::FMAXNMV_PRED:
case AArch64ISD::FMAXV_PRED:
case AArch64ISD::FMINNMV_PRED:
case AArch64ISD::FMINV_PRED:
case AArch64ISD::ORV_PRED:
case AArch64ISD::SADDV_PRED:
case AArch64ISD::SMAXV_PRED:
case AArch64ISD::SMINV_PRED:
case AArch64ISD::UADDV_PRED:
case AArch64ISD::UMAXV_PRED:
case AArch64ISD::UMINV_PRED:
return true;
}
}
static SDValue removeRedundantInsertVectorElt(SDNode *N) {
assert(N->getOpcode() == ISD::INSERT_VECTOR_ELT && "Unexpected node!");
SDValue InsertVec = N->getOperand(0);
SDValue InsertElt = N->getOperand(1);
SDValue InsertIdx = N->getOperand(2);
// We only care about inserts into the first element...
if (!isNullConstant(InsertIdx))
return SDValue();
// ...of a zero'd vector...
if (!ISD::isConstantSplatVectorAllZeros(InsertVec.getNode()))
return SDValue();
// ...where the inserted data was previously extracted...
if (InsertElt.getOpcode() != ISD::EXTRACT_VECTOR_ELT)
return SDValue();
SDValue ExtractVec = InsertElt.getOperand(0);
SDValue ExtractIdx = InsertElt.getOperand(1);
// ...from the first element of a vector.
if (!isNullConstant(ExtractIdx))
return SDValue();
// If we get here we are effectively trying to zero lanes 1-N of a vector.
// Ensure there's no type conversion going on.
if (N->getValueType(0) != ExtractVec.getValueType())
return SDValue();
if (!isLanes1toNKnownZero(ExtractVec))
return SDValue();
// The explicit zeroing is redundant.
return ExtractVec;
}
static SDValue
performInsertVectorEltCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI) {
if (SDValue Res = removeRedundantInsertVectorElt(N))
return Res;
return performPostLD1Combine(N, DCI, true);
}
static SDValue performFPExtendCombine(SDNode *N, SelectionDAG &DAG,
TargetLowering::DAGCombinerInfo &DCI,
const AArch64Subtarget *Subtarget) {
SDValue N0 = N->getOperand(0);
EVT VT = N->getValueType(0);
// If this is fp_round(fpextend), don't fold it, allow ourselves to be folded.
if (N->hasOneUse() && N->user_begin()->getOpcode() == ISD::FP_ROUND)
return SDValue();
auto hasValidElementTypeForFPExtLoad = [](EVT VT) {
EVT EltVT = VT.getVectorElementType();
return EltVT == MVT::f32 || EltVT == MVT::f64;
};
// fold (fpext (load x)) -> (fpext (fptrunc (extload x)))
// We purposefully don't care about legality of the nodes here as we know
// they can be split down into something legal.
if (DCI.isBeforeLegalizeOps() && ISD::isNormalLoad(N0.getNode()) &&
N0.hasOneUse() && Subtarget->useSVEForFixedLengthVectors() &&
VT.isFixedLengthVector() && hasValidElementTypeForFPExtLoad(VT) &&
VT.getFixedSizeInBits() >= Subtarget->getMinSVEVectorSizeInBits()) {
LoadSDNode *LN0 = cast<LoadSDNode>(N0);
SDValue ExtLoad = DAG.getExtLoad(ISD::EXTLOAD, SDLoc(N), VT,
LN0->getChain(), LN0->getBasePtr(),
N0.getValueType(), LN0->getMemOperand());
DCI.CombineTo(N, ExtLoad);
DCI.CombineTo(
N0.getNode(),
DAG.getNode(ISD::FP_ROUND, SDLoc(N0), N0.getValueType(), ExtLoad,
DAG.getIntPtrConstant(1, SDLoc(N0), /*isTarget=*/true)),
ExtLoad.getValue(1));
return SDValue(N, 0); // Return N so it doesn't get rechecked!
}
return SDValue();
}
static SDValue performBSPExpandForSVE(SDNode *N, SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
EVT VT = N->getValueType(0);
// Don't expand for NEON, SVE2 or SME
if (!VT.isScalableVector() || Subtarget->hasSVE2() || Subtarget->hasSME())
return SDValue();
SDLoc DL(N);
SDValue Mask = N->getOperand(0);
SDValue In1 = N->getOperand(1);
SDValue In2 = N->getOperand(2);
SDValue InvMask = DAG.getNOT(DL, Mask, VT);
SDValue Sel = DAG.getNode(ISD::AND, DL, VT, Mask, In1);
SDValue SelInv = DAG.getNode(ISD::AND, DL, VT, InvMask, In2);
return DAG.getNode(ISD::OR, DL, VT, Sel, SelInv);
}
static SDValue performDupLane128Combine(SDNode *N, SelectionDAG &DAG) {
EVT VT = N->getValueType(0);
SDValue Insert = N->getOperand(0);
if (Insert.getOpcode() != ISD::INSERT_SUBVECTOR)
return SDValue();
if (!Insert.getOperand(0).isUndef())
return SDValue();
uint64_t IdxInsert = Insert.getConstantOperandVal(2);
uint64_t IdxDupLane = N->getConstantOperandVal(1);
if (IdxInsert != 0 || IdxDupLane != 0)
return SDValue();
SDValue Bitcast = Insert.getOperand(1);
if (Bitcast.getOpcode() != ISD::BITCAST)
return SDValue();
SDValue Subvec = Bitcast.getOperand(0);
EVT SubvecVT = Subvec.getValueType();
if (!SubvecVT.is128BitVector())
return SDValue();
EVT NewSubvecVT =
getPackedSVEVectorVT(Subvec.getValueType().getVectorElementType());
SDLoc DL(N);
SDValue NewInsert =
DAG.getNode(ISD::INSERT_SUBVECTOR, DL, NewSubvecVT,
DAG.getUNDEF(NewSubvecVT), Subvec, Insert->getOperand(2));
SDValue NewDuplane128 = DAG.getNode(AArch64ISD::DUPLANE128, DL, NewSubvecVT,
NewInsert, N->getOperand(1));
return DAG.getNode(ISD::BITCAST, DL, VT, NewDuplane128);
}
// Try to combine mull with uzp1.
static SDValue tryCombineMULLWithUZP1(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (DCI.isBeforeLegalizeOps())
return SDValue();
SDValue LHS = N->getOperand(0);
SDValue RHS = N->getOperand(1);
SDValue ExtractHigh;
SDValue ExtractLow;
SDValue TruncHigh;
SDValue TruncLow;
SDLoc DL(N);
// Check the operands are trunc and extract_high.
if (isEssentiallyExtractHighSubvector(LHS) &&
RHS.getOpcode() == ISD::TRUNCATE) {
TruncHigh = RHS;
if (LHS.getOpcode() == ISD::BITCAST)
ExtractHigh = LHS.getOperand(0);
else
ExtractHigh = LHS;
} else if (isEssentiallyExtractHighSubvector(RHS) &&
LHS.getOpcode() == ISD::TRUNCATE) {
TruncHigh = LHS;
if (RHS.getOpcode() == ISD::BITCAST)
ExtractHigh = RHS.getOperand(0);
else
ExtractHigh = RHS;
} else
return SDValue();
// If the truncate's operand is BUILD_VECTOR with DUP, do not combine the op
// with uzp1.
// You can see the regressions on test/CodeGen/AArch64/aarch64-smull.ll
SDValue TruncHighOp = TruncHigh.getOperand(0);
EVT TruncHighOpVT = TruncHighOp.getValueType();
if (TruncHighOp.getOpcode() == AArch64ISD::DUP ||
DAG.isSplatValue(TruncHighOp, false))
return SDValue();
// Check there is other extract_high with same source vector.
// For example,
//
// t18: v4i16 = extract_subvector t2, Constant:i64<0>
// t12: v4i16 = truncate t11
// t31: v4i32 = AArch64ISD::SMULL t18, t12
// t23: v4i16 = extract_subvector t2, Constant:i64<4>
// t16: v4i16 = truncate t15
// t30: v4i32 = AArch64ISD::SMULL t23, t1
//
// This dagcombine assumes the two extract_high uses same source vector in
// order to detect the pair of the mull. If they have different source vector,
// this code will not work.
// TODO: Should also try to look through a bitcast.
bool HasFoundMULLow = true;
SDValue ExtractHighSrcVec = ExtractHigh.getOperand(0);
if (ExtractHighSrcVec->use_size() != 2)
HasFoundMULLow = false;
// Find ExtractLow.
for (SDNode *User : ExtractHighSrcVec.getNode()->users()) {
if (User == ExtractHigh.getNode())
continue;
if (User->getOpcode() != ISD::EXTRACT_SUBVECTOR ||
!isNullConstant(User->getOperand(1))) {
HasFoundMULLow = false;
break;
}
ExtractLow.setNode(User);
}
if (!ExtractLow || !ExtractLow->hasOneUse())
HasFoundMULLow = false;
// Check ExtractLow's user.
if (HasFoundMULLow) {
SDNode *ExtractLowUser = *ExtractLow.getNode()->user_begin();
if (ExtractLowUser->getOpcode() != N->getOpcode()) {
HasFoundMULLow = false;
} else {
if (ExtractLowUser->getOperand(0) == ExtractLow) {
if (ExtractLowUser->getOperand(1).getOpcode() == ISD::TRUNCATE)
TruncLow = ExtractLowUser->getOperand(1);
else
HasFoundMULLow = false;
} else {
if (ExtractLowUser->getOperand(0).getOpcode() == ISD::TRUNCATE)
TruncLow = ExtractLowUser->getOperand(0);
else
HasFoundMULLow = false;
}
}
}
// If the truncate's operand is BUILD_VECTOR with DUP, do not combine the op
// with uzp1.
// You can see the regressions on test/CodeGen/AArch64/aarch64-smull.ll
EVT TruncHighVT = TruncHigh.getValueType();
EVT UZP1VT = TruncHighVT.getDoubleNumVectorElementsVT(*DAG.getContext());
SDValue TruncLowOp =
HasFoundMULLow ? TruncLow.getOperand(0) : DAG.getUNDEF(UZP1VT);
EVT TruncLowOpVT = TruncLowOp.getValueType();
if (HasFoundMULLow && (TruncLowOp.getOpcode() == AArch64ISD::DUP ||
DAG.isSplatValue(TruncLowOp, false)))
return SDValue();
// Create uzp1, extract_high and extract_low.
if (TruncHighOpVT != UZP1VT)
TruncHighOp = DAG.getNode(ISD::BITCAST, DL, UZP1VT, TruncHighOp);
if (TruncLowOpVT != UZP1VT)
TruncLowOp = DAG.getNode(ISD::BITCAST, DL, UZP1VT, TruncLowOp);
SDValue UZP1 =
DAG.getNode(AArch64ISD::UZP1, DL, UZP1VT, TruncLowOp, TruncHighOp);
SDValue HighIdxCst =
DAG.getConstant(TruncHighVT.getVectorNumElements(), DL, MVT::i64);
SDValue NewTruncHigh =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, TruncHighVT, UZP1, HighIdxCst);
DAG.ReplaceAllUsesWith(TruncHigh, NewTruncHigh);
if (HasFoundMULLow) {
EVT TruncLowVT = TruncLow.getValueType();
SDValue NewTruncLow = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, TruncLowVT,
UZP1, ExtractLow.getOperand(1));
DAG.ReplaceAllUsesWith(TruncLow, NewTruncLow);
}
return SDValue(N, 0);
}
static SDValue performMULLCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (SDValue Val =
tryCombineLongOpWithDup(Intrinsic::not_intrinsic, N, DCI, DAG))
return Val;
if (SDValue Val = tryCombineMULLWithUZP1(N, DCI, DAG))
return Val;
return SDValue();
}
static SDValue
performScalarToVectorCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
// Let's do below transform.
//
// t34: v4i32 = AArch64ISD::UADDLV t2
// t35: i32 = extract_vector_elt t34, Constant:i64<0>
// t7: i64 = zero_extend t35
// t20: v1i64 = scalar_to_vector t7
// ==>
// t34: v4i32 = AArch64ISD::UADDLV t2
// t39: v2i32 = extract_subvector t34, Constant:i64<0>
// t40: v1i64 = AArch64ISD::NVCAST t39
if (DCI.isBeforeLegalizeOps())
return SDValue();
EVT VT = N->getValueType(0);
if (VT != MVT::v1i64)
return SDValue();
SDValue ZEXT = N->getOperand(0);
if (ZEXT.getOpcode() != ISD::ZERO_EXTEND || ZEXT.getValueType() != MVT::i64)
return SDValue();
SDValue EXTRACT_VEC_ELT = ZEXT.getOperand(0);
if (EXTRACT_VEC_ELT.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
EXTRACT_VEC_ELT.getValueType() != MVT::i32)
return SDValue();
if (!isNullConstant(EXTRACT_VEC_ELT.getOperand(1)))
return SDValue();
SDValue UADDLV = EXTRACT_VEC_ELT.getOperand(0);
if (UADDLV.getOpcode() != AArch64ISD::UADDLV ||
UADDLV.getValueType() != MVT::v4i32 ||
UADDLV.getOperand(0).getValueType() != MVT::v8i8)
return SDValue();
// Let's generate new sequence with AArch64ISD::NVCAST.
SDLoc DL(N);
SDValue EXTRACT_SUBVEC =
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, MVT::v2i32, UADDLV,
DAG.getConstant(0, DL, MVT::i64));
SDValue NVCAST =
DAG.getNode(AArch64ISD::NVCAST, DL, MVT::v1i64, EXTRACT_SUBVEC);
return NVCAST;
}
/// If the operand is a bitwise AND with a constant RHS, and the shift has a
/// constant RHS and is the only use, we can pull it out of the shift, i.e.
///
/// (shl (and X, C1), C2) -> (and (shl X, C2), (shl C1, C2))
///
/// We prefer this canonical form to match existing isel patterns.
static SDValue performSHLCombine(SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (DCI.isBeforeLegalizeOps())
return SDValue();
SDValue Op0 = N->getOperand(0);
if (Op0.getOpcode() != ISD::AND || !Op0.hasOneUse())
return SDValue();
SDValue C1 = Op0->getOperand(1);
SDValue C2 = N->getOperand(1);
if (!isa<ConstantSDNode>(C1) || !isa<ConstantSDNode>(C2))
return SDValue();
// Might be folded into shifted op, do not lower.
if (N->hasOneUse()) {
unsigned UseOpc = N->user_begin()->getOpcode();
if (UseOpc == ISD::ADD || UseOpc == ISD::SUB || UseOpc == ISD::SETCC ||
UseOpc == AArch64ISD::ADDS || UseOpc == AArch64ISD::SUBS)
return SDValue();
}
SDLoc DL(N);
EVT VT = N->getValueType(0);
// Don't combine unless (shl C1, C2) can be constant folded. Otherwise,
// DAGCombiner will simplify (and (op x...), (op y...)) -> (op (and x, y))
// causing infinite loop. Result may also be worse.
SDValue NewRHS = DAG.getNode(ISD::SHL, DL, VT, C1, C2);
if (!isa<ConstantSDNode>(NewRHS))
return SDValue();
SDValue X = Op0->getOperand(0);
SDValue NewShift = DAG.getNode(ISD::SHL, DL, VT, X, C2);
return DAG.getNode(ISD::AND, DL, VT, NewShift, NewRHS);
}
SDValue AArch64TargetLowering::PerformDAGCombine(SDNode *N,
DAGCombinerInfo &DCI) const {
SelectionDAG &DAG = DCI.DAG;
switch (N->getOpcode()) {
default:
LLVM_DEBUG(dbgs() << "Custom combining: skipping\n");
break;
case ISD::VECREDUCE_AND:
case ISD::VECREDUCE_OR:
case ISD::VECREDUCE_XOR:
return performVecReduceBitwiseCombine(N, DCI, DAG);
case ISD::ADD:
case ISD::SUB:
return performAddSubCombine(N, DCI);
case ISD::BUILD_VECTOR:
return performBuildVectorCombine(N, DCI, DAG);
case ISD::TRUNCATE:
return performTruncateCombine(N, DAG, DCI);
case AArch64ISD::ANDS:
return performFlagSettingCombine(N, DCI, ISD::AND);
case AArch64ISD::ADC:
if (auto R = foldOverflowCheck(N, DAG, /* IsAdd */ true))
return R;
return foldADCToCINC(N, DAG);
case AArch64ISD::SBC:
return foldOverflowCheck(N, DAG, /* IsAdd */ false);
case AArch64ISD::ADCS:
if (auto R = foldOverflowCheck(N, DAG, /* IsAdd */ true))
return R;
return performFlagSettingCombine(N, DCI, AArch64ISD::ADC);
case AArch64ISD::SBCS:
if (auto R = foldOverflowCheck(N, DAG, /* IsAdd */ false))
return R;
return performFlagSettingCombine(N, DCI, AArch64ISD::SBC);
case AArch64ISD::BICi: {
APInt DemandedBits =
APInt::getAllOnes(N->getValueType(0).getScalarSizeInBits());
APInt DemandedElts =
APInt::getAllOnes(N->getValueType(0).getVectorNumElements());
if (DAG.getTargetLoweringInfo().SimplifyDemandedBits(
SDValue(N, 0), DemandedBits, DemandedElts, DCI))
return SDValue();
break;
}
case ISD::XOR:
return performXorCombine(N, DAG, DCI, Subtarget);
case ISD::MUL:
return performMulCombine(N, DAG, DCI, Subtarget);
case ISD::SINT_TO_FP:
case ISD::UINT_TO_FP:
return performIntToFpCombine(N, DAG, DCI, Subtarget);
case ISD::FP_TO_SINT:
case ISD::FP_TO_UINT:
case ISD::FP_TO_SINT_SAT:
case ISD::FP_TO_UINT_SAT:
return performFpToIntCombine(N, DAG, DCI, Subtarget);
case ISD::OR:
return performORCombine(N, DCI, Subtarget, *this);
case ISD::AND:
return performANDCombine(N, DCI);
case ISD::FADD:
return performFADDCombine(N, DCI);
case ISD::INTRINSIC_WO_CHAIN:
return performIntrinsicCombine(N, DCI, Subtarget);
case ISD::ANY_EXTEND:
case ISD::ZERO_EXTEND:
case ISD::SIGN_EXTEND:
return performExtendCombine(N, DCI, DAG);
case ISD::SIGN_EXTEND_INREG:
return performSignExtendInRegCombine(N, DCI, DAG);
case ISD::CONCAT_VECTORS:
return performConcatVectorsCombine(N, DCI, DAG);
case ISD::EXTRACT_SUBVECTOR:
return performExtractSubvectorCombine(N, DCI, DAG);
case ISD::INSERT_SUBVECTOR:
return performInsertSubvectorCombine(N, DCI, DAG);
case ISD::SELECT:
return performSelectCombine(N, DCI);
case ISD::VSELECT:
return performVSelectCombine(N, DCI.DAG);
case ISD::SETCC:
return performSETCCCombine(N, DCI, DAG);
case ISD::LOAD:
return performLOADCombine(N, DCI, DAG, Subtarget);
case ISD::STORE:
return performSTORECombine(N, DCI, DAG, Subtarget);
case ISD::MSTORE:
return performMSTORECombine(N, DCI, DAG, Subtarget);
case ISD::MGATHER:
case ISD::MSCATTER:
case ISD::EXPERIMENTAL_VECTOR_HISTOGRAM:
return performMaskedGatherScatterCombine(N, DCI, DAG);
case ISD::FP_EXTEND:
return performFPExtendCombine(N, DAG, DCI, Subtarget);
case AArch64ISD::BRCOND:
return performBRCONDCombine(N, DCI, DAG);
case AArch64ISD::TBNZ:
case AArch64ISD::TBZ:
return performTBZCombine(N, DCI, DAG);
case AArch64ISD::CSEL:
return performCSELCombine(N, DCI, DAG);
case AArch64ISD::DUP:
case AArch64ISD::DUPLANE8:
case AArch64ISD::DUPLANE16:
case AArch64ISD::DUPLANE32:
case AArch64ISD::DUPLANE64:
return performDUPCombine(N, DCI);
case AArch64ISD::DUPLANE128:
return performDupLane128Combine(N, DAG);
case AArch64ISD::NVCAST:
return performNVCASTCombine(N, DAG);
case AArch64ISD::SPLICE:
return performSpliceCombine(N, DAG);
case AArch64ISD::UUNPKLO:
case AArch64ISD::UUNPKHI:
return performUnpackCombine(N, DAG, Subtarget);
case AArch64ISD::UZP1:
case AArch64ISD::UZP2:
return performUzpCombine(N, DAG, Subtarget);
case AArch64ISD::SETCC_MERGE_ZERO:
return performSetccMergeZeroCombine(N, DCI);
case AArch64ISD::REINTERPRET_CAST:
return performReinterpretCastCombine(N);
case AArch64ISD::GLD1_MERGE_ZERO:
case AArch64ISD::GLD1_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1_UXTW_MERGE_ZERO:
case AArch64ISD::GLD1_SXTW_MERGE_ZERO:
case AArch64ISD::GLD1_UXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1_SXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1_IMM_MERGE_ZERO:
case AArch64ISD::GLD1S_MERGE_ZERO:
case AArch64ISD::GLD1S_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1S_UXTW_MERGE_ZERO:
case AArch64ISD::GLD1S_SXTW_MERGE_ZERO:
case AArch64ISD::GLD1S_UXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1S_SXTW_SCALED_MERGE_ZERO:
case AArch64ISD::GLD1S_IMM_MERGE_ZERO:
return performGLD1Combine(N, DAG);
case AArch64ISD::VASHR:
case AArch64ISD::VLSHR:
return performVectorShiftCombine(N, *this, DCI);
case AArch64ISD::SUNPKLO:
return performSunpkloCombine(N, DAG);
case AArch64ISD::BSP:
return performBSPExpandForSVE(N, DAG, Subtarget);
case ISD::INSERT_VECTOR_ELT:
return performInsertVectorEltCombine(N, DCI);
case ISD::EXTRACT_VECTOR_ELT:
return performExtractVectorEltCombine(N, DCI, Subtarget);
case ISD::VECREDUCE_ADD:
return performVecReduceAddCombine(N, DCI.DAG, Subtarget);
case ISD::GET_ACTIVE_LANE_MASK:
return performActiveLaneMaskCombine(N, DCI, Subtarget);
case AArch64ISD::UADDV:
return performUADDVCombine(N, DAG);
case AArch64ISD::SMULL:
case AArch64ISD::UMULL:
case AArch64ISD::PMULL:
return performMULLCombine(N, DCI, DAG);
case ISD::INTRINSIC_VOID:
case ISD::INTRINSIC_W_CHAIN:
switch (N->getConstantOperandVal(1)) {
case Intrinsic::aarch64_sve_prfb_gather_scalar_offset:
return combineSVEPrefetchVecBaseImmOff(N, DAG, 1 /*=ScalarSizeInBytes*/);
case Intrinsic::aarch64_sve_prfh_gather_scalar_offset:
return combineSVEPrefetchVecBaseImmOff(N, DAG, 2 /*=ScalarSizeInBytes*/);
case Intrinsic::aarch64_sve_prfw_gather_scalar_offset:
return combineSVEPrefetchVecBaseImmOff(N, DAG, 4 /*=ScalarSizeInBytes*/);
case Intrinsic::aarch64_sve_prfd_gather_scalar_offset:
return combineSVEPrefetchVecBaseImmOff(N, DAG, 8 /*=ScalarSizeInBytes*/);
case Intrinsic::aarch64_sve_prfb_gather_uxtw_index:
case Intrinsic::aarch64_sve_prfb_gather_sxtw_index:
case Intrinsic::aarch64_sve_prfh_gather_uxtw_index:
case Intrinsic::aarch64_sve_prfh_gather_sxtw_index:
case Intrinsic::aarch64_sve_prfw_gather_uxtw_index:
case Intrinsic::aarch64_sve_prfw_gather_sxtw_index:
case Intrinsic::aarch64_sve_prfd_gather_uxtw_index:
case Intrinsic::aarch64_sve_prfd_gather_sxtw_index:
return legalizeSVEGatherPrefetchOffsVec(N, DAG);
case Intrinsic::aarch64_neon_ld2:
case Intrinsic::aarch64_neon_ld3:
case Intrinsic::aarch64_neon_ld4:
case Intrinsic::aarch64_neon_ld1x2:
case Intrinsic::aarch64_neon_ld1x3:
case Intrinsic::aarch64_neon_ld1x4:
case Intrinsic::aarch64_neon_ld2lane:
case Intrinsic::aarch64_neon_ld3lane:
case Intrinsic::aarch64_neon_ld4lane:
case Intrinsic::aarch64_neon_ld2r:
case Intrinsic::aarch64_neon_ld3r:
case Intrinsic::aarch64_neon_ld4r:
case Intrinsic::aarch64_neon_st2:
case Intrinsic::aarch64_neon_st3:
case Intrinsic::aarch64_neon_st4:
case Intrinsic::aarch64_neon_st1x2:
case Intrinsic::aarch64_neon_st1x3:
case Intrinsic::aarch64_neon_st1x4:
case Intrinsic::aarch64_neon_st2lane:
case Intrinsic::aarch64_neon_st3lane:
case Intrinsic::aarch64_neon_st4lane:
return performNEONPostLDSTCombine(N, DCI, DAG);
case Intrinsic::aarch64_sve_ldnt1:
return performLDNT1Combine(N, DAG);
case Intrinsic::aarch64_sve_ld1rq:
return performLD1ReplicateCombine<AArch64ISD::LD1RQ_MERGE_ZERO>(N, DAG);
case Intrinsic::aarch64_sve_ld1ro:
return performLD1ReplicateCombine<AArch64ISD::LD1RO_MERGE_ZERO>(N, DAG);
case Intrinsic::aarch64_sve_ldnt1_gather_scalar_offset:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLDNT1_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldnt1_gather:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLDNT1_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldnt1_gather_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLDNT1_INDEX_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldnt1_gather_uxtw:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLDNT1_MERGE_ZERO);
case Intrinsic::aarch64_sve_ld1:
return performLD1Combine(N, DAG, AArch64ISD::LD1_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldnf1:
return performLD1Combine(N, DAG, AArch64ISD::LDNF1_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldff1:
return performLD1Combine(N, DAG, AArch64ISD::LDFF1_MERGE_ZERO);
case Intrinsic::aarch64_sve_st1:
return performST1Combine(N, DAG);
case Intrinsic::aarch64_sve_stnt1:
return performSTNT1Combine(N, DAG);
case Intrinsic::aarch64_sve_stnt1_scatter_scalar_offset:
return performScatterStoreCombine(N, DAG, AArch64ISD::SSTNT1_PRED);
case Intrinsic::aarch64_sve_stnt1_scatter_uxtw:
return performScatterStoreCombine(N, DAG, AArch64ISD::SSTNT1_PRED);
case Intrinsic::aarch64_sve_stnt1_scatter:
return performScatterStoreCombine(N, DAG, AArch64ISD::SSTNT1_PRED);
case Intrinsic::aarch64_sve_stnt1_scatter_index:
return performScatterStoreCombine(N, DAG, AArch64ISD::SSTNT1_INDEX_PRED);
case Intrinsic::aarch64_sve_ld1_gather:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLD1_MERGE_ZERO);
case Intrinsic::aarch64_sve_ld1q_gather_scalar_offset:
case Intrinsic::aarch64_sve_ld1q_gather_vector_offset:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLD1Q_MERGE_ZERO);
case Intrinsic::aarch64_sve_ld1q_gather_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLD1Q_INDEX_MERGE_ZERO);
case Intrinsic::aarch64_sve_ld1_gather_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLD1_SCALED_MERGE_ZERO);
case Intrinsic::aarch64_sve_ld1_gather_sxtw:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLD1_SXTW_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ld1_gather_uxtw:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLD1_UXTW_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ld1_gather_sxtw_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLD1_SXTW_SCALED_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ld1_gather_uxtw_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLD1_UXTW_SCALED_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ld1_gather_scalar_offset:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLD1_IMM_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldff1_gather:
return performGatherLoadCombine(N, DAG, AArch64ISD::GLDFF1_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldff1_gather_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLDFF1_SCALED_MERGE_ZERO);
case Intrinsic::aarch64_sve_ldff1_gather_sxtw:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLDFF1_SXTW_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ldff1_gather_uxtw:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLDFF1_UXTW_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ldff1_gather_sxtw_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLDFF1_SXTW_SCALED_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ldff1_gather_uxtw_index:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLDFF1_UXTW_SCALED_MERGE_ZERO,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_ldff1_gather_scalar_offset:
return performGatherLoadCombine(N, DAG,
AArch64ISD::GLDFF1_IMM_MERGE_ZERO);
case Intrinsic::aarch64_sve_st1q_scatter_scalar_offset:
case Intrinsic::aarch64_sve_st1q_scatter_vector_offset:
return performScatterStoreCombine(N, DAG, AArch64ISD::SST1Q_PRED);
case Intrinsic::aarch64_sve_st1q_scatter_index:
return performScatterStoreCombine(N, DAG, AArch64ISD::SST1Q_INDEX_PRED);
case Intrinsic::aarch64_sve_st1_scatter:
return performScatterStoreCombine(N, DAG, AArch64ISD::SST1_PRED);
case Intrinsic::aarch64_sve_st1_scatter_index:
return performScatterStoreCombine(N, DAG, AArch64ISD::SST1_SCALED_PRED);
case Intrinsic::aarch64_sve_st1_scatter_sxtw:
return performScatterStoreCombine(N, DAG, AArch64ISD::SST1_SXTW_PRED,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_st1_scatter_uxtw:
return performScatterStoreCombine(N, DAG, AArch64ISD::SST1_UXTW_PRED,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_st1_scatter_sxtw_index:
return performScatterStoreCombine(N, DAG,
AArch64ISD::SST1_SXTW_SCALED_PRED,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_st1_scatter_uxtw_index:
return performScatterStoreCombine(N, DAG,
AArch64ISD::SST1_UXTW_SCALED_PRED,
/*OnlyPackedOffsets=*/false);
case Intrinsic::aarch64_sve_st1_scatter_scalar_offset:
return performScatterStoreCombine(N, DAG, AArch64ISD::SST1_IMM_PRED);
case Intrinsic::aarch64_rndr:
case Intrinsic::aarch64_rndrrs: {
unsigned IntrinsicID = N->getConstantOperandVal(1);
auto Register =
(IntrinsicID == Intrinsic::aarch64_rndr ? AArch64SysReg::RNDR
: AArch64SysReg::RNDRRS);
SDLoc DL(N);
SDValue A = DAG.getNode(
AArch64ISD::MRS, DL, DAG.getVTList(MVT::i64, MVT::i32, MVT::Other),
N->getOperand(0), DAG.getConstant(Register, DL, MVT::i32));
SDValue B = DAG.getNode(
AArch64ISD::CSINC, DL, MVT::i32, DAG.getConstant(0, DL, MVT::i32),
DAG.getConstant(0, DL, MVT::i32),
DAG.getConstant(AArch64CC::NE, DL, MVT::i32), A.getValue(1));
return DAG.getMergeValues(
{A, DAG.getZExtOrTrunc(B, DL, MVT::i1), A.getValue(2)}, DL);
}
case Intrinsic::aarch64_sme_ldr_zt:
return DAG.getNode(AArch64ISD::RESTORE_ZT, SDLoc(N),
DAG.getVTList(MVT::Other), N->getOperand(0),
N->getOperand(2), N->getOperand(3));
case Intrinsic::aarch64_sme_str_zt:
return DAG.getNode(AArch64ISD::SAVE_ZT, SDLoc(N),
DAG.getVTList(MVT::Other), N->getOperand(0),
N->getOperand(2), N->getOperand(3));
default:
break;
}
break;
case ISD::GlobalAddress:
return performGlobalAddressCombine(N, DAG, Subtarget, getTargetMachine());
case ISD::CTLZ:
return performCTLZCombine(N, DAG, Subtarget);
case ISD::SCALAR_TO_VECTOR:
return performScalarToVectorCombine(N, DCI, DAG);
case ISD::SHL:
return performSHLCombine(N, DCI, DAG);
}
return SDValue();
}
// Check if the return value is used as only a return value, as otherwise
// we can't perform a tail-call. In particular, we need to check for
// target ISD nodes that are returns and any other "odd" constructs
// that the generic analysis code won't necessarily catch.
bool AArch64TargetLowering::isUsedByReturnOnly(SDNode *N,
SDValue &Chain) const {
if (N->getNumValues() != 1)
return false;
if (!N->hasNUsesOfValue(1, 0))
return false;
SDValue TCChain = Chain;
SDNode *Copy = *N->user_begin();
if (Copy->getOpcode() == ISD::CopyToReg) {
// If the copy has a glue operand, we conservatively assume it isn't safe to
// perform a tail call.
if (Copy->getOperand(Copy->getNumOperands() - 1).getValueType() ==
MVT::Glue)
return false;
TCChain = Copy->getOperand(0);
} else if (Copy->getOpcode() != ISD::FP_EXTEND)
return false;
bool HasRet = false;
for (SDNode *Node : Copy->users()) {
if (Node->getOpcode() != AArch64ISD::RET_GLUE)
return false;
HasRet = true;
}
if (!HasRet)
return false;
Chain = TCChain;
return true;
}
// Return whether the an instruction can potentially be optimized to a tail
// call. This will cause the optimizers to attempt to move, or duplicate,
// return instructions to help enable tail call optimizations for this
// instruction.
bool AArch64TargetLowering::mayBeEmittedAsTailCall(const CallInst *CI) const {
return CI->isTailCall();
}
bool AArch64TargetLowering::isIndexingLegal(MachineInstr &MI, Register Base,
Register Offset, bool IsPre,
MachineRegisterInfo &MRI) const {
auto CstOffset = getIConstantVRegVal(Offset, MRI);
if (!CstOffset || CstOffset->isZero())
return false;
// All of the indexed addressing mode instructions take a signed 9 bit
// immediate offset. Our CstOffset is a G_PTR_ADD offset so it already
// encodes the sign/indexing direction.
return isInt<9>(CstOffset->getSExtValue());
}
bool AArch64TargetLowering::getIndexedAddressParts(SDNode *N, SDNode *Op,
SDValue &Base,
SDValue &Offset,
SelectionDAG &DAG) const {
if (Op->getOpcode() != ISD::ADD && Op->getOpcode() != ISD::SUB)
return false;
// Non-null if there is exactly one user of the loaded value (ignoring chain).
SDNode *ValOnlyUser = nullptr;
for (SDUse &U : N->uses()) {
if (U.getResNo() == 1)
continue; // Ignore chain.
if (ValOnlyUser == nullptr)
ValOnlyUser = U.getUser();
else {
ValOnlyUser = nullptr; // Multiple non-chain uses, bail out.
break;
}
}
auto IsUndefOrZero = [](SDValue V) {
return V.isUndef() || isNullOrNullSplat(V, /*AllowUndefs*/ true);
};
// If the only user of the value is a scalable vector splat, it is
// preferable to do a replicating load (ld1r*).
if (ValOnlyUser && ValOnlyUser->getValueType(0).isScalableVector() &&
(ValOnlyUser->getOpcode() == ISD::SPLAT_VECTOR ||
(ValOnlyUser->getOpcode() == AArch64ISD::DUP_MERGE_PASSTHRU &&
IsUndefOrZero(ValOnlyUser->getOperand(2)))))
return false;
Base = Op->getOperand(0);
// All of the indexed addressing mode instructions take a signed
// 9 bit immediate offset.
if (ConstantSDNode *RHS = dyn_cast<ConstantSDNode>(Op->getOperand(1))) {
int64_t RHSC = RHS->getSExtValue();
if (Op->getOpcode() == ISD::SUB)
RHSC = -(uint64_t)RHSC;
if (!isInt<9>(RHSC))
return false;
// When big-endian VLD1/VST1 are used for vector load and store, and these
// only allow an offset that's equal to the store size.
EVT MemType = cast<MemSDNode>(N)->getMemoryVT();
if (!Subtarget->isLittleEndian() && MemType.isVector() &&
(uint64_t)RHSC != MemType.getStoreSize())
return false;
// Always emit pre-inc/post-inc addressing mode. Use negated constant offset
// when dealing with subtraction.
Offset = DAG.getConstant(RHSC, SDLoc(N), RHS->getValueType(0));
return true;
}
return false;
}
bool AArch64TargetLowering::getPreIndexedAddressParts(SDNode *N, SDValue &Base,
SDValue &Offset,
ISD::MemIndexedMode &AM,
SelectionDAG &DAG) const {
EVT VT;
SDValue Ptr;
if (LoadSDNode *LD = dyn_cast<LoadSDNode>(N)) {
VT = LD->getMemoryVT();
Ptr = LD->getBasePtr();
} else if (StoreSDNode *ST = dyn_cast<StoreSDNode>(N)) {
VT = ST->getMemoryVT();
Ptr = ST->getBasePtr();
} else
return false;
if (!getIndexedAddressParts(N, Ptr.getNode(), Base, Offset, DAG))
return false;
AM = ISD::PRE_INC;
return true;
}
bool AArch64TargetLowering::getPostIndexedAddressParts(
SDNode *N, SDNode *Op, SDValue &Base, SDValue &Offset,
ISD::MemIndexedMode &AM, SelectionDAG &DAG) const {
EVT VT;
SDValue Ptr;
if (LoadSDNode *LD = dyn_cast<LoadSDNode>(N)) {
VT = LD->getMemoryVT();
Ptr = LD->getBasePtr();
} else if (StoreSDNode *ST = dyn_cast<StoreSDNode>(N)) {
VT = ST->getMemoryVT();
Ptr = ST->getBasePtr();
} else
return false;
if (!getIndexedAddressParts(N, Op, Base, Offset, DAG))
return false;
// Post-indexing updates the base, so it's not a valid transform
// if that's not the same as the load's pointer.
if (Ptr != Base)
return false;
AM = ISD::POST_INC;
return true;
}
static void replaceBoolVectorBitcast(SDNode *N,
SmallVectorImpl<SDValue> &Results,
SelectionDAG &DAG) {
SDLoc DL(N);
SDValue Op = N->getOperand(0);
EVT VT = N->getValueType(0);
[[maybe_unused]] EVT SrcVT = Op.getValueType();
assert(SrcVT.isVector() && SrcVT.getVectorElementType() == MVT::i1 &&
"Must be bool vector.");
// Special handling for Clang's __builtin_convertvector. For vectors with <8
// elements, it adds a vector concatenation with undef(s). If we encounter
// this here, we can skip the concat.
if (Op.getOpcode() == ISD::CONCAT_VECTORS && !Op.getOperand(0).isUndef()) {
bool AllUndef = true;
for (unsigned I = 1; I < Op.getNumOperands(); ++I)
AllUndef &= Op.getOperand(I).isUndef();
if (AllUndef)
Op = Op.getOperand(0);
}
SDValue VectorBits = vectorToScalarBitmask(Op.getNode(), DAG);
if (VectorBits)
Results.push_back(DAG.getZExtOrTrunc(VectorBits, DL, VT));
}
static void CustomNonLegalBITCASTResults(SDNode *N,
SmallVectorImpl<SDValue> &Results,
SelectionDAG &DAG, EVT ExtendVT,
EVT CastVT) {
SDLoc DL(N);
SDValue Op = N->getOperand(0);
EVT VT = N->getValueType(0);
// Use SCALAR_TO_VECTOR for lane zero
SDValue Vec = DAG.getNode(ISD::SCALAR_TO_VECTOR, DL, ExtendVT, Op);
SDValue CastVal = DAG.getNode(ISD::BITCAST, DL, CastVT, Vec);
SDValue IdxZero = DAG.getVectorIdxConstant(0, DL);
Results.push_back(
DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, CastVal, IdxZero));
}
void AArch64TargetLowering::ReplaceBITCASTResults(
SDNode *N, SmallVectorImpl<SDValue> &Results, SelectionDAG &DAG) const {
SDLoc DL(N);
SDValue Op = N->getOperand(0);
EVT VT = N->getValueType(0);
EVT SrcVT = Op.getValueType();
if (VT == MVT::v2i16 && SrcVT == MVT::i32) {
CustomNonLegalBITCASTResults(N, Results, DAG, MVT::v2i32, MVT::v4i16);
return;
}
if (VT == MVT::v4i8 && SrcVT == MVT::i32) {
CustomNonLegalBITCASTResults(N, Results, DAG, MVT::v2i32, MVT::v8i8);
return;
}
if (VT == MVT::v2i8 && SrcVT == MVT::i16) {
CustomNonLegalBITCASTResults(N, Results, DAG, MVT::v4i16, MVT::v8i8);
return;
}
if (VT.isScalableVector() && !isTypeLegal(VT) && isTypeLegal(SrcVT)) {
assert(!VT.isFloatingPoint() && SrcVT.isFloatingPoint() &&
"Expected fp->int bitcast!");
// Bitcasting between unpacked vector types of different element counts is
// not a NOP because the live elements are laid out differently.
// 01234567
// e.g. nxv2i32 = XX??XX??
// nxv4f16 = X?X?X?X?
if (VT.getVectorElementCount() != SrcVT.getVectorElementCount())
return;
SDValue CastResult = getSVESafeBitCast(getSVEContainerType(VT), Op, DAG);
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, CastResult));
return;
}
if (SrcVT.isVector() && SrcVT.getVectorElementType() == MVT::i1 &&
!VT.isVector())
return replaceBoolVectorBitcast(N, Results, DAG);
if (VT != MVT::i16 || (SrcVT != MVT::f16 && SrcVT != MVT::bf16))
return;
Op = DAG.getTargetInsertSubreg(AArch64::hsub, DL, MVT::f32,
DAG.getUNDEF(MVT::i32), Op);
Op = DAG.getNode(ISD::BITCAST, DL, MVT::i32, Op);
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, MVT::i16, Op));
}
static void ReplaceAddWithADDP(SDNode *N, SmallVectorImpl<SDValue> &Results,
SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
EVT VT = N->getValueType(0);
if (!VT.is256BitVector() ||
(VT.getScalarType().isFloatingPoint() &&
!N->getFlags().hasAllowReassociation()) ||
(VT.getScalarType() == MVT::f16 && !Subtarget->hasFullFP16()) ||
VT.getScalarType() == MVT::bf16)
return;
SDValue X = N->getOperand(0);
auto *Shuf = dyn_cast<ShuffleVectorSDNode>(N->getOperand(1));
if (!Shuf) {
Shuf = dyn_cast<ShuffleVectorSDNode>(N->getOperand(0));
X = N->getOperand(1);
if (!Shuf)
return;
}
if (Shuf->getOperand(0) != X || !Shuf->getOperand(1)->isUndef())
return;
// Check the mask is 1,0,3,2,5,4,...
ArrayRef<int> Mask = Shuf->getMask();
for (int I = 0, E = Mask.size(); I < E; I++)
if (Mask[I] != (I % 2 == 0 ? I + 1 : I - 1))
return;
SDLoc DL(N);
auto LoHi = DAG.SplitVector(X, DL);
assert(LoHi.first.getValueType() == LoHi.second.getValueType());
SDValue Addp = DAG.getNode(AArch64ISD::ADDP, N, LoHi.first.getValueType(),
LoHi.first, LoHi.second);
// Shuffle the elements back into order.
SmallVector<int> NMask;
for (unsigned I = 0, E = VT.getVectorNumElements() / 2; I < E; I++) {
NMask.push_back(I);
NMask.push_back(I);
}
Results.push_back(
DAG.getVectorShuffle(VT, DL,
DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, Addp,
DAG.getUNDEF(LoHi.first.getValueType())),
DAG.getUNDEF(VT), NMask));
}
static void ReplaceReductionResults(SDNode *N,
SmallVectorImpl<SDValue> &Results,
SelectionDAG &DAG, unsigned InterOp,
unsigned AcrossOp) {
EVT LoVT, HiVT;
SDValue Lo, Hi;
SDLoc DL(N);
std::tie(LoVT, HiVT) = DAG.GetSplitDestVTs(N->getValueType(0));
std::tie(Lo, Hi) = DAG.SplitVectorOperand(N, 0);
SDValue InterVal = DAG.getNode(InterOp, DL, LoVT, Lo, Hi);
SDValue SplitVal = DAG.getNode(AcrossOp, DL, LoVT, InterVal);
Results.push_back(SplitVal);
}
void AArch64TargetLowering::ReplaceExtractSubVectorResults(
SDNode *N, SmallVectorImpl<SDValue> &Results, SelectionDAG &DAG) const {
SDValue In = N->getOperand(0);
EVT InVT = In.getValueType();
// Common code will handle these just fine.
if (!InVT.isScalableVector() || !InVT.isInteger())
return;
SDLoc DL(N);
EVT VT = N->getValueType(0);
// The following checks bail if this is not a halving operation.
ElementCount ResEC = VT.getVectorElementCount();
if (InVT.getVectorElementCount() != (ResEC * 2))
return;
auto *CIndex = dyn_cast<ConstantSDNode>(N->getOperand(1));
if (!CIndex)
return;
unsigned Index = CIndex->getZExtValue();
if ((Index != 0) && (Index != ResEC.getKnownMinValue()))
return;
unsigned Opcode = (Index == 0) ? AArch64ISD::UUNPKLO : AArch64ISD::UUNPKHI;
EVT ExtendedHalfVT = VT.widenIntegerVectorElementType(*DAG.getContext());
SDValue Half = DAG.getNode(Opcode, DL, ExtendedHalfVT, N->getOperand(0));
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, Half));
}
void AArch64TargetLowering::ReplaceGetActiveLaneMaskResults(
SDNode *N, SmallVectorImpl<SDValue> &Results, SelectionDAG &DAG) const {
assert((Subtarget->hasSVE2p1() ||
(Subtarget->hasSME2() && Subtarget->isStreaming())) &&
"Custom lower of get.active.lane.mask missing required feature.");
assert(N->getValueType(0) == MVT::nxv32i1 &&
"Unexpected result type for get.active.lane.mask");
SDLoc DL(N);
SDValue Idx = N->getOperand(0);
SDValue TC = N->getOperand(1);
assert(Idx.getValueType().getFixedSizeInBits() <= 64 &&
"Unexpected operand type for get.active.lane.mask");
if (Idx.getValueType() != MVT::i64) {
Idx = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i64, Idx);
TC = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i64, TC);
}
SDValue ID =
DAG.getTargetConstant(Intrinsic::aarch64_sve_whilelo_x2, DL, MVT::i64);
EVT HalfVT = N->getValueType(0).getHalfNumVectorElementsVT(*DAG.getContext());
auto WideMask =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, {HalfVT, HalfVT}, {ID, Idx, TC});
Results.push_back(DAG.getNode(ISD::CONCAT_VECTORS, DL, N->getValueType(0),
{WideMask.getValue(0), WideMask.getValue(1)}));
}
// Create an even/odd pair of X registers holding integer value V.
static SDValue createGPRPairNode(SelectionDAG &DAG, SDValue V) {
SDLoc DL(V.getNode());
auto [VLo, VHi] = DAG.SplitScalar(V, DL, MVT::i64, MVT::i64);
if (DAG.getDataLayout().isBigEndian())
std::swap (VLo, VHi);
SDValue RegClass =
DAG.getTargetConstant(AArch64::XSeqPairsClassRegClassID, DL, MVT::i32);
SDValue SubReg0 = DAG.getTargetConstant(AArch64::sube64, DL, MVT::i32);
SDValue SubReg1 = DAG.getTargetConstant(AArch64::subo64, DL, MVT::i32);
const SDValue Ops[] = { RegClass, VLo, SubReg0, VHi, SubReg1 };
return SDValue(
DAG.getMachineNode(TargetOpcode::REG_SEQUENCE, DL, MVT::Untyped, Ops), 0);
}
static void ReplaceCMP_SWAP_128Results(SDNode *N,
SmallVectorImpl<SDValue> &Results,
SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
assert(N->getValueType(0) == MVT::i128 &&
"AtomicCmpSwap on types less than 128 should be legal");
MachineMemOperand *MemOp = cast<MemSDNode>(N)->getMemOperand();
if (Subtarget->hasLSE() || Subtarget->outlineAtomics()) {
// LSE has a 128-bit compare and swap (CASP), but i128 is not a legal type,
// so lower it here, wrapped in REG_SEQUENCE and EXTRACT_SUBREG.
SDValue Ops[] = {
createGPRPairNode(DAG, N->getOperand(2)), // Compare value
createGPRPairNode(DAG, N->getOperand(3)), // Store value
N->getOperand(1), // Ptr
N->getOperand(0), // Chain in
};
unsigned Opcode;
switch (MemOp->getMergedOrdering()) {
case AtomicOrdering::Monotonic:
Opcode = AArch64::CASPX;
break;
case AtomicOrdering::Acquire:
Opcode = AArch64::CASPAX;
break;
case AtomicOrdering::Release:
Opcode = AArch64::CASPLX;
break;
case AtomicOrdering::AcquireRelease:
case AtomicOrdering::SequentiallyConsistent:
Opcode = AArch64::CASPALX;
break;
default:
llvm_unreachable("Unexpected ordering!");
}
MachineSDNode *CmpSwap = DAG.getMachineNode(
Opcode, SDLoc(N), DAG.getVTList(MVT::Untyped, MVT::Other), Ops);
DAG.setNodeMemRefs(CmpSwap, {MemOp});
unsigned SubReg1 = AArch64::sube64, SubReg2 = AArch64::subo64;
if (DAG.getDataLayout().isBigEndian())
std::swap(SubReg1, SubReg2);
SDValue Lo = DAG.getTargetExtractSubreg(SubReg1, SDLoc(N), MVT::i64,
SDValue(CmpSwap, 0));
SDValue Hi = DAG.getTargetExtractSubreg(SubReg2, SDLoc(N), MVT::i64,
SDValue(CmpSwap, 0));
Results.push_back(
DAG.getNode(ISD::BUILD_PAIR, SDLoc(N), MVT::i128, Lo, Hi));
Results.push_back(SDValue(CmpSwap, 1)); // Chain out
return;
}
unsigned Opcode;
switch (MemOp->getMergedOrdering()) {
case AtomicOrdering::Monotonic:
Opcode = AArch64::CMP_SWAP_128_MONOTONIC;
break;
case AtomicOrdering::Acquire:
Opcode = AArch64::CMP_SWAP_128_ACQUIRE;
break;
case AtomicOrdering::Release:
Opcode = AArch64::CMP_SWAP_128_RELEASE;
break;
case AtomicOrdering::AcquireRelease:
case AtomicOrdering::SequentiallyConsistent:
Opcode = AArch64::CMP_SWAP_128;
break;
default:
llvm_unreachable("Unexpected ordering!");
}
SDLoc DL(N);
auto Desired = DAG.SplitScalar(N->getOperand(2), DL, MVT::i64, MVT::i64);
auto New = DAG.SplitScalar(N->getOperand(3), DL, MVT::i64, MVT::i64);
SDValue Ops[] = {N->getOperand(1), Desired.first, Desired.second,
New.first, New.second, N->getOperand(0)};
SDNode *CmpSwap = DAG.getMachineNode(
Opcode, SDLoc(N), DAG.getVTList(MVT::i64, MVT::i64, MVT::i32, MVT::Other),
Ops);
DAG.setNodeMemRefs(cast<MachineSDNode>(CmpSwap), {MemOp});
Results.push_back(DAG.getNode(ISD::BUILD_PAIR, DL, MVT::i128,
SDValue(CmpSwap, 0), SDValue(CmpSwap, 1)));
Results.push_back(SDValue(CmpSwap, 3));
}
static unsigned getAtomicLoad128Opcode(unsigned ISDOpcode,
AtomicOrdering Ordering) {
// ATOMIC_LOAD_CLR only appears when lowering ATOMIC_LOAD_AND (see
// LowerATOMIC_LOAD_AND). We can't take that approach with 128-bit, because
// the type is not legal. Therefore we shouldn't expect to see a 128-bit
// ATOMIC_LOAD_CLR at any point.
assert(ISDOpcode != ISD::ATOMIC_LOAD_CLR &&
"ATOMIC_LOAD_AND should be lowered to LDCLRP directly");
assert(ISDOpcode != ISD::ATOMIC_LOAD_ADD && "There is no 128 bit LDADD");
assert(ISDOpcode != ISD::ATOMIC_LOAD_SUB && "There is no 128 bit LDSUB");
if (ISDOpcode == ISD::ATOMIC_LOAD_AND) {
// The operand will need to be XORed in a separate step.
switch (Ordering) {
case AtomicOrdering::Monotonic:
return AArch64::LDCLRP;
break;
case AtomicOrdering::Acquire:
return AArch64::LDCLRPA;
break;
case AtomicOrdering::Release:
return AArch64::LDCLRPL;
break;
case AtomicOrdering::AcquireRelease:
case AtomicOrdering::SequentiallyConsistent:
return AArch64::LDCLRPAL;
break;
default:
llvm_unreachable("Unexpected ordering!");
}
}
if (ISDOpcode == ISD::ATOMIC_LOAD_OR) {
switch (Ordering) {
case AtomicOrdering::Monotonic:
return AArch64::LDSETP;
break;
case AtomicOrdering::Acquire:
return AArch64::LDSETPA;
break;
case AtomicOrdering::Release:
return AArch64::LDSETPL;
break;
case AtomicOrdering::AcquireRelease:
case AtomicOrdering::SequentiallyConsistent:
return AArch64::LDSETPAL;
break;
default:
llvm_unreachable("Unexpected ordering!");
}
}
if (ISDOpcode == ISD::ATOMIC_SWAP) {
switch (Ordering) {
case AtomicOrdering::Monotonic:
return AArch64::SWPP;
break;
case AtomicOrdering::Acquire:
return AArch64::SWPPA;
break;
case AtomicOrdering::Release:
return AArch64::SWPPL;
break;
case AtomicOrdering::AcquireRelease:
case AtomicOrdering::SequentiallyConsistent:
return AArch64::SWPPAL;
break;
default:
llvm_unreachable("Unexpected ordering!");
}
}
llvm_unreachable("Unexpected ISDOpcode!");
}
static void ReplaceATOMIC_LOAD_128Results(SDNode *N,
SmallVectorImpl<SDValue> &Results,
SelectionDAG &DAG,
const AArch64Subtarget *Subtarget) {
// LSE128 has a 128-bit RMW ops, but i128 is not a legal type, so lower it
// here. This follows the approach of the CMP_SWAP_XXX pseudo instructions
// rather than the CASP instructions, because CASP has register classes for
// the pairs of registers and therefore uses REG_SEQUENCE and EXTRACT_SUBREG
// to present them as single operands. LSE128 instructions use the GPR64
// register class (because the pair does not have to be sequential), like
// CMP_SWAP_XXX, and therefore we use TRUNCATE and BUILD_PAIR.
assert(N->getValueType(0) == MVT::i128 &&
"AtomicLoadXXX on types less than 128 should be legal");
if (!Subtarget->hasLSE128())
return;
MachineMemOperand *MemOp = cast<MemSDNode>(N)->getMemOperand();
const SDValue &Chain = N->getOperand(0);
const SDValue &Ptr = N->getOperand(1);
const SDValue &Val128 = N->getOperand(2);
std::pair<SDValue, SDValue> Val2x64 =
DAG.SplitScalar(Val128, SDLoc(Val128), MVT::i64, MVT::i64);
const unsigned ISDOpcode = N->getOpcode();
const unsigned MachineOpcode =
getAtomicLoad128Opcode(ISDOpcode, MemOp->getMergedOrdering());
if (ISDOpcode == ISD::ATOMIC_LOAD_AND) {
SDLoc DL(Val128);
Val2x64.first =
DAG.getNode(ISD::XOR, DL, MVT::i64,
DAG.getAllOnesConstant(DL, MVT::i64), Val2x64.first);
Val2x64.second =
DAG.getNode(ISD::XOR, DL, MVT::i64,
DAG.getAllOnesConstant(DL, MVT::i64), Val2x64.second);
}
SDValue Ops[] = {Val2x64.first, Val2x64.second, Ptr, Chain};
if (DAG.getDataLayout().isBigEndian())
std::swap(Ops[0], Ops[1]);
MachineSDNode *AtomicInst =
DAG.getMachineNode(MachineOpcode, SDLoc(N),
DAG.getVTList(MVT::i64, MVT::i64, MVT::Other), Ops);
DAG.setNodeMemRefs(AtomicInst, {MemOp});
SDValue Lo = SDValue(AtomicInst, 0), Hi = SDValue(AtomicInst, 1);
if (DAG.getDataLayout().isBigEndian())
std::swap(Lo, Hi);
Results.push_back(DAG.getNode(ISD::BUILD_PAIR, SDLoc(N), MVT::i128, Lo, Hi));
Results.push_back(SDValue(AtomicInst, 2)); // Chain out
}
void AArch64TargetLowering::ReplaceNodeResults(
SDNode *N, SmallVectorImpl<SDValue> &Results, SelectionDAG &DAG) const {
switch (N->getOpcode()) {
default:
llvm_unreachable("Don't know how to custom expand this");
case ISD::BITCAST:
ReplaceBITCASTResults(N, Results, DAG);
return;
case ISD::VECREDUCE_ADD:
case ISD::VECREDUCE_SMAX:
case ISD::VECREDUCE_SMIN:
case ISD::VECREDUCE_UMAX:
case ISD::VECREDUCE_UMIN:
Results.push_back(LowerVECREDUCE(SDValue(N, 0), DAG));
return;
case ISD::VECTOR_COMPRESS:
if (SDValue Res = LowerVECTOR_COMPRESS(SDValue(N, 0), DAG))
Results.push_back(Res);
return;
case ISD::ADD:
case ISD::FADD:
ReplaceAddWithADDP(N, Results, DAG, Subtarget);
return;
case ISD::CTPOP:
case ISD::PARITY:
if (SDValue Result = LowerCTPOP_PARITY(SDValue(N, 0), DAG))
Results.push_back(Result);
return;
case AArch64ISD::SADDV:
ReplaceReductionResults(N, Results, DAG, ISD::ADD, AArch64ISD::SADDV);
return;
case AArch64ISD::UADDV:
ReplaceReductionResults(N, Results, DAG, ISD::ADD, AArch64ISD::UADDV);
return;
case AArch64ISD::SMINV:
ReplaceReductionResults(N, Results, DAG, ISD::SMIN, AArch64ISD::SMINV);
return;
case AArch64ISD::UMINV:
ReplaceReductionResults(N, Results, DAG, ISD::UMIN, AArch64ISD::UMINV);
return;
case AArch64ISD::SMAXV:
ReplaceReductionResults(N, Results, DAG, ISD::SMAX, AArch64ISD::SMAXV);
return;
case AArch64ISD::UMAXV:
ReplaceReductionResults(N, Results, DAG, ISD::UMAX, AArch64ISD::UMAXV);
return;
case ISD::MULHS:
if (useSVEForFixedLengthVectorVT(SDValue(N, 0).getValueType()))
Results.push_back(
LowerToPredicatedOp(SDValue(N, 0), DAG, AArch64ISD::MULHS_PRED));
return;
case ISD::MULHU:
if (useSVEForFixedLengthVectorVT(SDValue(N, 0).getValueType()))
Results.push_back(
LowerToPredicatedOp(SDValue(N, 0), DAG, AArch64ISD::MULHU_PRED));
return;
case ISD::FP_TO_UINT:
case ISD::FP_TO_SINT:
case ISD::STRICT_FP_TO_SINT:
case ISD::STRICT_FP_TO_UINT:
assert(N->getValueType(0) == MVT::i128 && "unexpected illegal conversion");
// Let normal code take care of it by not adding anything to Results.
return;
case ISD::ATOMIC_CMP_SWAP:
ReplaceCMP_SWAP_128Results(N, Results, DAG, Subtarget);
return;
case ISD::ATOMIC_LOAD_CLR:
assert(N->getValueType(0) != MVT::i128 &&
"128-bit ATOMIC_LOAD_AND should be lowered directly to LDCLRP");
break;
case ISD::ATOMIC_LOAD_AND:
case ISD::ATOMIC_LOAD_OR:
case ISD::ATOMIC_SWAP: {
assert(cast<AtomicSDNode>(N)->getVal().getValueType() == MVT::i128 &&
"Expected 128-bit atomicrmw.");
// These need custom type legalisation so we go directly to instruction.
ReplaceATOMIC_LOAD_128Results(N, Results, DAG, Subtarget);
return;
}
case ISD::ADDRSPACECAST: {
SDValue V = LowerADDRSPACECAST(SDValue(N, 0), DAG);
Results.push_back(V);
return;
}
case ISD::ATOMIC_LOAD:
case ISD::LOAD: {
MemSDNode *LoadNode = cast<MemSDNode>(N);
EVT MemVT = LoadNode->getMemoryVT();
// Handle lowering 256 bit non temporal loads into LDNP for little-endian
// targets.
if (LoadNode->isNonTemporal() && Subtarget->isLittleEndian() &&
MemVT.getSizeInBits() == 256u &&
(MemVT.getScalarSizeInBits() == 8u ||
MemVT.getScalarSizeInBits() == 16u ||
MemVT.getScalarSizeInBits() == 32u ||
MemVT.getScalarSizeInBits() == 64u)) {
SDValue Result = DAG.getMemIntrinsicNode(
AArch64ISD::LDNP, SDLoc(N),
DAG.getVTList({MemVT.getHalfNumVectorElementsVT(*DAG.getContext()),
MemVT.getHalfNumVectorElementsVT(*DAG.getContext()),
MVT::Other}),
{LoadNode->getChain(), LoadNode->getBasePtr()},
LoadNode->getMemoryVT(), LoadNode->getMemOperand());
SDValue Pair = DAG.getNode(ISD::CONCAT_VECTORS, SDLoc(N), MemVT,
Result.getValue(0), Result.getValue(1));
Results.append({Pair, Result.getValue(2) /* Chain */});
return;
}
if ((!LoadNode->isVolatile() && !LoadNode->isAtomic()) ||
LoadNode->getMemoryVT() != MVT::i128) {
// Non-volatile or atomic loads are optimized later in AArch64's load/store
// optimizer.
return;
}
if (SDValue(N, 0).getValueType() == MVT::i128) {
auto *AN = dyn_cast<AtomicSDNode>(LoadNode);
bool isLoadAcquire =
AN && AN->getSuccessOrdering() == AtomicOrdering::Acquire;
unsigned Opcode = isLoadAcquire ? AArch64ISD::LDIAPP : AArch64ISD::LDP;
if (isLoadAcquire)
assert(Subtarget->hasFeature(AArch64::FeatureRCPC3));
SDValue Result = DAG.getMemIntrinsicNode(
Opcode, SDLoc(N), DAG.getVTList({MVT::i64, MVT::i64, MVT::Other}),
{LoadNode->getChain(), LoadNode->getBasePtr()},
LoadNode->getMemoryVT(), LoadNode->getMemOperand());
unsigned FirstRes = DAG.getDataLayout().isBigEndian() ? 1 : 0;
SDValue Pair =
DAG.getNode(ISD::BUILD_PAIR, SDLoc(N), MVT::i128,
Result.getValue(FirstRes), Result.getValue(1 - FirstRes));
Results.append({Pair, Result.getValue(2) /* Chain */});
}
return;
}
case ISD::EXTRACT_SUBVECTOR:
ReplaceExtractSubVectorResults(N, Results, DAG);
return;
case ISD::INSERT_SUBVECTOR:
case ISD::CONCAT_VECTORS:
// Custom lowering has been requested for INSERT_SUBVECTOR and
// CONCAT_VECTORS -- but delegate to common code for result type
// legalisation
return;
case ISD::GET_ACTIVE_LANE_MASK:
ReplaceGetActiveLaneMaskResults(N, Results, DAG);
return;
case ISD::INTRINSIC_WO_CHAIN: {
EVT VT = N->getValueType(0);
Intrinsic::ID IntID =
static_cast<Intrinsic::ID>(N->getConstantOperandVal(0));
switch (IntID) {
default:
return;
case Intrinsic::aarch64_sve_clasta_n: {
assert((VT == MVT::i8 || VT == MVT::i16) &&
"custom lowering for unexpected type");
SDLoc DL(N);
auto Op2 = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, N->getOperand(2));
auto V = DAG.getNode(AArch64ISD::CLASTA_N, DL, MVT::i32,
N->getOperand(1), Op2, N->getOperand(3));
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, V));
return;
}
case Intrinsic::aarch64_sve_clastb_n: {
assert((VT == MVT::i8 || VT == MVT::i16) &&
"custom lowering for unexpected type");
SDLoc DL(N);
auto Op2 = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, N->getOperand(2));
auto V = DAG.getNode(AArch64ISD::CLASTB_N, DL, MVT::i32,
N->getOperand(1), Op2, N->getOperand(3));
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, V));
return;
}
case Intrinsic::aarch64_sve_lasta: {
assert((VT == MVT::i8 || VT == MVT::i16) &&
"custom lowering for unexpected type");
SDLoc DL(N);
auto V = DAG.getNode(AArch64ISD::LASTA, DL, MVT::i32,
N->getOperand(1), N->getOperand(2));
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, V));
return;
}
case Intrinsic::aarch64_sve_lastb: {
assert((VT == MVT::i8 || VT == MVT::i16) &&
"custom lowering for unexpected type");
SDLoc DL(N);
auto V = DAG.getNode(AArch64ISD::LASTB, DL, MVT::i32,
N->getOperand(1), N->getOperand(2));
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, V));
return;
}
case Intrinsic::aarch64_sme_in_streaming_mode: {
SDLoc DL(N);
SDValue Chain = DAG.getEntryNode();
SDValue RuntimePStateSM =
getRuntimePStateSM(DAG, Chain, DL, N->getValueType(0));
Results.push_back(
DAG.getNode(ISD::TRUNCATE, DL, MVT::i1, RuntimePStateSM));
return;
}
case Intrinsic::experimental_vector_match: {
if (!VT.isFixedLengthVector() || VT.getVectorElementType() != MVT::i1)
return;
// NOTE: Only trivial type promotion is supported.
EVT NewVT = getTypeToTransformTo(*DAG.getContext(), VT);
if (NewVT.getVectorNumElements() != VT.getVectorNumElements())
return;
SDLoc DL(N);
auto V = DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, NewVT, N->ops());
Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, V));
return;
}
}
}
case ISD::READ_REGISTER: {
SDLoc DL(N);
assert(N->getValueType(0) == MVT::i128 &&
"READ_REGISTER custom lowering is only for 128-bit sysregs");
SDValue Chain = N->getOperand(0);
SDValue SysRegName = N->getOperand(1);
SDValue Result = DAG.getNode(
AArch64ISD::MRRS, DL, DAG.getVTList({MVT::i64, MVT::i64, MVT::Other}),
Chain, SysRegName);
// Sysregs are not endian. Result.getValue(0) always contains the lower half
// of the 128-bit System Register value.
SDValue Pair = DAG.getNode(ISD::BUILD_PAIR, DL, MVT::i128,
Result.getValue(0), Result.getValue(1));
Results.push_back(Pair);
Results.push_back(Result.getValue(2)); // Chain
return;
}
}
}
bool AArch64TargetLowering::useLoadStackGuardNode(const Module &M) const {
if (Subtarget->isTargetAndroid() || Subtarget->isTargetFuchsia())
return TargetLowering::useLoadStackGuardNode(M);
return true;
}
unsigned AArch64TargetLowering::combineRepeatedFPDivisors() const {
// Combine multiple FDIVs with the same divisor into multiple FMULs by the
// reciprocal if there are three or more FDIVs.
return 3;
}
TargetLoweringBase::LegalizeTypeAction
AArch64TargetLowering::getPreferredVectorAction(MVT VT) const {
// During type legalization, we prefer to widen v1i8, v1i16, v1i32 to v8i8,
// v4i16, v2i32 instead of to promote.
if (VT == MVT::v1i8 || VT == MVT::v1i16 || VT == MVT::v1i32 ||
VT == MVT::v1f32)
return TypeWidenVector;
return TargetLoweringBase::getPreferredVectorAction(VT);
}
// In v8.4a, ldp and stp instructions are guaranteed to be single-copy atomic
// provided the address is 16-byte aligned.
bool AArch64TargetLowering::isOpSuitableForLDPSTP(const Instruction *I) const {
if (!Subtarget->hasLSE2())
return false;
if (auto LI = dyn_cast<LoadInst>(I))
return LI->getType()->getPrimitiveSizeInBits() == 128 &&
LI->getAlign() >= Align(16);
if (auto SI = dyn_cast<StoreInst>(I))
return SI->getValueOperand()->getType()->getPrimitiveSizeInBits() == 128 &&
SI->getAlign() >= Align(16);
return false;
}
bool AArch64TargetLowering::isOpSuitableForLSE128(const Instruction *I) const {
if (!Subtarget->hasLSE128())
return false;
// Only use SWPP for stores where LSE2 would require a fence. Unlike STP, SWPP
// will clobber the two registers.
if (const auto *SI = dyn_cast<StoreInst>(I))
return SI->getValueOperand()->getType()->getPrimitiveSizeInBits() == 128 &&
SI->getAlign() >= Align(16) &&
(SI->getOrdering() == AtomicOrdering::Release ||
SI->getOrdering() == AtomicOrdering::SequentiallyConsistent);
if (const auto *RMW = dyn_cast<AtomicRMWInst>(I))
return RMW->getValOperand()->getType()->getPrimitiveSizeInBits() == 128 &&
RMW->getAlign() >= Align(16) &&
(RMW->getOperation() == AtomicRMWInst::Xchg ||
RMW->getOperation() == AtomicRMWInst::And ||
RMW->getOperation() == AtomicRMWInst::Or);
return false;
}
bool AArch64TargetLowering::isOpSuitableForRCPC3(const Instruction *I) const {
if (!Subtarget->hasLSE2() || !Subtarget->hasRCPC3())
return false;
if (auto LI = dyn_cast<LoadInst>(I))
return LI->getType()->getPrimitiveSizeInBits() == 128 &&
LI->getAlign() >= Align(16) &&
LI->getOrdering() == AtomicOrdering::Acquire;
if (auto SI = dyn_cast<StoreInst>(I))
return SI->getValueOperand()->getType()->getPrimitiveSizeInBits() == 128 &&
SI->getAlign() >= Align(16) &&
SI->getOrdering() == AtomicOrdering::Release;
return false;
}
bool AArch64TargetLowering::shouldInsertFencesForAtomic(
const Instruction *I) const {
if (isOpSuitableForRCPC3(I))
return false;
if (isOpSuitableForLSE128(I))
return false;
if (isOpSuitableForLDPSTP(I))
return true;
return false;
}
bool AArch64TargetLowering::shouldInsertTrailingFenceForAtomicStore(
const Instruction *I) const {
// Store-Release instructions only provide seq_cst guarantees when paired with
// Load-Acquire instructions. MSVC CRT does not use these instructions to
// implement seq_cst loads and stores, so we need additional explicit fences
// after memory writes.
if (!Subtarget->getTargetTriple().isWindowsMSVCEnvironment())
return false;
switch (I->getOpcode()) {
default:
return false;
case Instruction::AtomicCmpXchg:
return cast<AtomicCmpXchgInst>(I)->getSuccessOrdering() ==
AtomicOrdering::SequentiallyConsistent;
case Instruction::AtomicRMW:
return cast<AtomicRMWInst>(I)->getOrdering() ==
AtomicOrdering::SequentiallyConsistent;
case Instruction::Store:
return cast<StoreInst>(I)->getOrdering() ==
AtomicOrdering::SequentiallyConsistent;
}
}
// Loads and stores less than 128-bits are already atomic; ones above that
// are doomed anyway, so defer to the default libcall and blame the OS when
// things go wrong.
TargetLoweringBase::AtomicExpansionKind
AArch64TargetLowering::shouldExpandAtomicStoreInIR(StoreInst *SI) const {
unsigned Size = SI->getValueOperand()->getType()->getPrimitiveSizeInBits();
if (Size != 128)
return AtomicExpansionKind::None;
if (isOpSuitableForRCPC3(SI))
return AtomicExpansionKind::None;
if (isOpSuitableForLSE128(SI))
return AtomicExpansionKind::Expand;
if (isOpSuitableForLDPSTP(SI))
return AtomicExpansionKind::None;
return AtomicExpansionKind::Expand;
}
// Loads and stores less than 128-bits are already atomic; ones above that
// are doomed anyway, so defer to the default libcall and blame the OS when
// things go wrong.
TargetLowering::AtomicExpansionKind
AArch64TargetLowering::shouldExpandAtomicLoadInIR(LoadInst *LI) const {
unsigned Size = LI->getType()->getPrimitiveSizeInBits();
if (Size != 128)
return AtomicExpansionKind::None;
if (isOpSuitableForRCPC3(LI))
return AtomicExpansionKind::None;
// No LSE128 loads
if (isOpSuitableForLDPSTP(LI))
return AtomicExpansionKind::None;
// At -O0, fast-regalloc cannot cope with the live vregs necessary to
// implement atomicrmw without spilling. If the target address is also on the
// stack and close enough to the spill slot, this can lead to a situation
// where the monitor always gets cleared and the atomic operation can never
// succeed. So at -O0 lower this operation to a CAS loop.
if (getTargetMachine().getOptLevel() == CodeGenOptLevel::None)
return AtomicExpansionKind::CmpXChg;
// Using CAS for an atomic load has a better chance of succeeding under high
// contention situations. So use it if available.
return Subtarget->hasLSE() ? AtomicExpansionKind::CmpXChg
: AtomicExpansionKind::LLSC;
}
// Return true if the atomic operation expansion will lower to use a library
// call, and is thus ineligible to use an LLSC expansion.
static bool rmwOpMayLowerToLibcall(const AArch64Subtarget &Subtarget,
const AtomicRMWInst *RMW) {
if (!RMW->isFloatingPointOperation())
return false;
switch (RMW->getType()->getScalarType()->getTypeID()) {
case Type::FloatTyID:
case Type::DoubleTyID:
case Type::HalfTyID:
case Type::BFloatTyID:
// Will use soft float
return !Subtarget.hasFPARMv8();
default:
// fp128 will emit library calls.
return true;
}
llvm_unreachable("covered type switch");
}
// The "default" for integer RMW operations is to expand to an LL/SC loop.
// However, with the LSE instructions (or outline-atomics mode, which provides
// library routines in place of the LSE-instructions), we can directly emit many
// operations instead.
TargetLowering::AtomicExpansionKind
AArch64TargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
Type *Ty = AI->getType();
unsigned Size = Ty->getPrimitiveSizeInBits();
assert(Size <= 128 && "AtomicExpandPass should've handled larger sizes.");
bool CanUseLSE128 = Subtarget->hasLSE128() && Size == 128 &&
(AI->getOperation() == AtomicRMWInst::Xchg ||
AI->getOperation() == AtomicRMWInst::Or ||
AI->getOperation() == AtomicRMWInst::And);
if (CanUseLSE128)
return AtomicExpansionKind::None;
// If LSFE available, use atomic FP instructions in preference to expansion
if (Subtarget->hasLSFE() && (AI->getOperation() == AtomicRMWInst::FAdd ||
AI->getOperation() == AtomicRMWInst::FMax ||
AI->getOperation() == AtomicRMWInst::FMin ||
AI->getOperation() == AtomicRMWInst::FMaximum ||
AI->getOperation() == AtomicRMWInst::FMinimum))
return AtomicExpansionKind::None;
// Nand is not supported in LSE.
// Leave 128 bits to LLSC or CmpXChg.
if (AI->getOperation() != AtomicRMWInst::Nand && Size < 128 &&
!AI->isFloatingPointOperation()) {
if (Subtarget->hasLSE())
return AtomicExpansionKind::None;
if (Subtarget->outlineAtomics()) {
// [U]Min/[U]Max RWM atomics are used in __sync_fetch_ libcalls so far.
// Don't outline them unless
// (1) high level <atomic> support approved:
// http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p0493r1.pdf
// (2) low level libgcc and compiler-rt support implemented by:
// min/max outline atomics helpers
if (AI->getOperation() != AtomicRMWInst::Min &&
AI->getOperation() != AtomicRMWInst::Max &&
AI->getOperation() != AtomicRMWInst::UMin &&
AI->getOperation() != AtomicRMWInst::UMax) {
return AtomicExpansionKind::None;
}
}
}
// At -O0, fast-regalloc cannot cope with the live vregs necessary to
// implement atomicrmw without spilling. If the target address is also on the
// stack and close enough to the spill slot, this can lead to a situation
// where the monitor always gets cleared and the atomic operation can never
// succeed. So at -O0 lower this operation to a CAS loop. Also worthwhile if
// we have a single CAS instruction that can replace the loop.
if (getTargetMachine().getOptLevel() == CodeGenOptLevel::None ||
Subtarget->hasLSE() || rmwOpMayLowerToLibcall(*Subtarget, AI))
return AtomicExpansionKind::CmpXChg;
return AtomicExpansionKind::LLSC;
}
TargetLowering::AtomicExpansionKind
AArch64TargetLowering::shouldExpandAtomicCmpXchgInIR(
AtomicCmpXchgInst *AI) const {
// If subtarget has LSE, leave cmpxchg intact for codegen.
if (Subtarget->hasLSE() || Subtarget->outlineAtomics())
return AtomicExpansionKind::None;
// At -O0, fast-regalloc cannot cope with the live vregs necessary to
// implement cmpxchg without spilling. If the address being exchanged is also
// on the stack and close enough to the spill slot, this can lead to a
// situation where the monitor always gets cleared and the atomic operation
// can never succeed. So at -O0 we need a late-expanded pseudo-inst instead.
if (getTargetMachine().getOptLevel() == CodeGenOptLevel::None)
return AtomicExpansionKind::None;
// 128-bit atomic cmpxchg is weird; AtomicExpand doesn't know how to expand
// it.
unsigned Size = AI->getCompareOperand()->getType()->getPrimitiveSizeInBits();
if (Size > 64)
return AtomicExpansionKind::None;
return AtomicExpansionKind::LLSC;
}
Value *AArch64TargetLowering::emitLoadLinked(IRBuilderBase &Builder,
Type *ValueTy, Value *Addr,
AtomicOrdering Ord) const {
Module *M = Builder.GetInsertBlock()->getParent()->getParent();
bool IsAcquire = isAcquireOrStronger(Ord);
// Since i128 isn't legal and intrinsics don't get type-lowered, the ldrexd
// intrinsic must return {i64, i64} and we have to recombine them into a
// single i128 here.
if (ValueTy->getPrimitiveSizeInBits() == 128) {
Intrinsic::ID Int =
IsAcquire ? Intrinsic::aarch64_ldaxp : Intrinsic::aarch64_ldxp;
Value *LoHi =
Builder.CreateIntrinsic(Int, Addr, /*FMFSource=*/nullptr, "lohi");
Value *Lo = Builder.CreateExtractValue(LoHi, 0, "lo");
Value *Hi = Builder.CreateExtractValue(LoHi, 1, "hi");
auto *Int128Ty = Type::getInt128Ty(Builder.getContext());
Lo = Builder.CreateZExt(Lo, Int128Ty, "lo64");
Hi = Builder.CreateZExt(Hi, Int128Ty, "hi64");
Value *Or = Builder.CreateOr(
Lo, Builder.CreateShl(Hi, ConstantInt::get(Int128Ty, 64)), "val64");
return Builder.CreateBitCast(Or, ValueTy);
}
Type *Tys[] = { Addr->getType() };
Intrinsic::ID Int =
IsAcquire ? Intrinsic::aarch64_ldaxr : Intrinsic::aarch64_ldxr;
const DataLayout &DL = M->getDataLayout();
IntegerType *IntEltTy = Builder.getIntNTy(DL.getTypeSizeInBits(ValueTy));
CallInst *CI = Builder.CreateIntrinsic(Int, Tys, Addr);
CI->addParamAttr(0, Attribute::get(Builder.getContext(),
Attribute::ElementType, IntEltTy));
Value *Trunc = Builder.CreateTrunc(CI, IntEltTy);
return Builder.CreateBitCast(Trunc, ValueTy);
}
void AArch64TargetLowering::emitAtomicCmpXchgNoStoreLLBalance(
IRBuilderBase &Builder) const {
Builder.CreateIntrinsic(Intrinsic::aarch64_clrex, {});
}
Value *AArch64TargetLowering::emitStoreConditional(IRBuilderBase &Builder,
Value *Val, Value *Addr,
AtomicOrdering Ord) const {
Module *M = Builder.GetInsertBlock()->getParent()->getParent();
bool IsRelease = isReleaseOrStronger(Ord);
// Since the intrinsics must have legal type, the i128 intrinsics take two
// parameters: "i64, i64". We must marshal Val into the appropriate form
// before the call.
if (Val->getType()->getPrimitiveSizeInBits() == 128) {
Intrinsic::ID Int =
IsRelease ? Intrinsic::aarch64_stlxp : Intrinsic::aarch64_stxp;
Function *Stxr = Intrinsic::getOrInsertDeclaration(M, Int);
Type *Int64Ty = Type::getInt64Ty(M->getContext());
Type *Int128Ty = Type::getInt128Ty(M->getContext());
Value *CastVal = Builder.CreateBitCast(Val, Int128Ty);
Value *Lo = Builder.CreateTrunc(CastVal, Int64Ty, "lo");
Value *Hi =
Builder.CreateTrunc(Builder.CreateLShr(CastVal, 64), Int64Ty, "hi");
return Builder.CreateCall(Stxr, {Lo, Hi, Addr});
}
Intrinsic::ID Int =
IsRelease ? Intrinsic::aarch64_stlxr : Intrinsic::aarch64_stxr;
Type *Tys[] = { Addr->getType() };
Function *Stxr = Intrinsic::getOrInsertDeclaration(M, Int, Tys);
const DataLayout &DL = M->getDataLayout();
IntegerType *IntValTy = Builder.getIntNTy(DL.getTypeSizeInBits(Val->getType()));
Val = Builder.CreateBitCast(Val, IntValTy);
CallInst *CI = Builder.CreateCall(
Stxr, {Builder.CreateZExtOrBitCast(
Val, Stxr->getFunctionType()->getParamType(0)),
Addr});
CI->addParamAttr(1, Attribute::get(Builder.getContext(),
Attribute::ElementType, Val->getType()));
return CI;
}
bool AArch64TargetLowering::functionArgumentNeedsConsecutiveRegisters(
Type *Ty, CallingConv::ID CallConv, bool isVarArg,
const DataLayout &DL) const {
if (!Ty->isArrayTy()) {
const TypeSize &TySize = Ty->getPrimitiveSizeInBits();
return TySize.isScalable() && TySize.getKnownMinValue() > 128;
}
// All non aggregate members of the type must have the same type
SmallVector<EVT> ValueVTs;
ComputeValueVTs(*this, DL, Ty, ValueVTs);
return all_equal(ValueVTs);
}
bool AArch64TargetLowering::shouldNormalizeToSelectSequence(LLVMContext &,
EVT) const {
return false;
}
static Value *UseTlsOffset(IRBuilderBase &IRB, unsigned Offset) {
Module *M = IRB.GetInsertBlock()->getParent()->getParent();
Function *ThreadPointerFunc = Intrinsic::getOrInsertDeclaration(
M, Intrinsic::thread_pointer, IRB.getPtrTy());
return IRB.CreatePointerCast(
IRB.CreateConstGEP1_32(IRB.getInt8Ty(), IRB.CreateCall(ThreadPointerFunc),
Offset),
IRB.getPtrTy(0));
}
Value *AArch64TargetLowering::getIRStackGuard(IRBuilderBase &IRB) const {
// Android provides a fixed TLS slot for the stack cookie. See the definition
// of TLS_SLOT_STACK_GUARD in
// https://android.googlesource.com/platform/bionic/+/main/libc/platform/bionic/tls_defines.h
if (Subtarget->isTargetAndroid())
return UseTlsOffset(IRB, 0x28);
// Fuchsia is similar.
// <zircon/tls.h> defines ZX_TLS_STACK_GUARD_OFFSET with this value.
if (Subtarget->isTargetFuchsia())
return UseTlsOffset(IRB, -0x10);
return TargetLowering::getIRStackGuard(IRB);
}
void AArch64TargetLowering::insertSSPDeclarations(Module &M) const {
// MSVC CRT provides functionalities for stack protection.
if (Subtarget->getTargetTriple().isWindowsMSVCEnvironment()) {
// MSVC CRT has a global variable holding security cookie.
M.getOrInsertGlobal("__security_cookie",
PointerType::getUnqual(M.getContext()));
// MSVC CRT has a function to validate security cookie.
FunctionCallee SecurityCheckCookie =
M.getOrInsertFunction(Subtarget->getSecurityCheckCookieName(),
Type::getVoidTy(M.getContext()),
PointerType::getUnqual(M.getContext()));
if (Function *F = dyn_cast<Function>(SecurityCheckCookie.getCallee())) {
F->setCallingConv(CallingConv::Win64);
F->addParamAttr(0, Attribute::AttrKind::InReg);
}
return;
}
TargetLowering::insertSSPDeclarations(M);
}
Value *AArch64TargetLowering::getSDagStackGuard(const Module &M) const {
// MSVC CRT has a global variable holding security cookie.
if (Subtarget->getTargetTriple().isWindowsMSVCEnvironment())
return M.getGlobalVariable("__security_cookie");
return TargetLowering::getSDagStackGuard(M);
}
Function *AArch64TargetLowering::getSSPStackGuardCheck(const Module &M) const {
// MSVC CRT has a function to validate security cookie.
if (Subtarget->getTargetTriple().isWindowsMSVCEnvironment())
return M.getFunction(Subtarget->getSecurityCheckCookieName());
return TargetLowering::getSSPStackGuardCheck(M);
}
Value *
AArch64TargetLowering::getSafeStackPointerLocation(IRBuilderBase &IRB) const {
// Android provides a fixed TLS slot for the SafeStack pointer. See the
// definition of TLS_SLOT_SAFESTACK in
// https://android.googlesource.com/platform/bionic/+/master/libc/private/bionic_tls.h
if (Subtarget->isTargetAndroid())
return UseTlsOffset(IRB, 0x48);
// Fuchsia is similar.
// <zircon/tls.h> defines ZX_TLS_UNSAFE_SP_OFFSET with this value.
if (Subtarget->isTargetFuchsia())
return UseTlsOffset(IRB, -0x8);
return TargetLowering::getSafeStackPointerLocation(IRB);
}
/// If a physical register, this returns the register that receives the
/// exception address on entry to an EH pad.
Register AArch64TargetLowering::getExceptionPointerRegister(
const Constant *PersonalityFn) const {
// FIXME: This is a guess. Has this been defined yet?
return AArch64::X0;
}
/// If a physical register, this returns the register that receives the
/// exception typeid on entry to a landing pad.
Register AArch64TargetLowering::getExceptionSelectorRegister(
const Constant *PersonalityFn) const {
// FIXME: This is a guess. Has this been defined yet?
return AArch64::X1;
}
bool AArch64TargetLowering::isMaskAndCmp0FoldingBeneficial(
const Instruction &AndI) const {
// Only sink 'and' mask to cmp use block if it is masking a single bit, since
// this is likely to be fold the and/cmp/br into a single tbz instruction. It
// may be beneficial to sink in other cases, but we would have to check that
// the cmp would not get folded into the br to form a cbz for these to be
// beneficial.
ConstantInt* Mask = dyn_cast<ConstantInt>(AndI.getOperand(1));
if (!Mask)
return false;
return Mask->getValue().isPowerOf2();
}
bool AArch64TargetLowering::
shouldProduceAndByConstByHoistingConstFromShiftsLHSOfAnd(
SDValue X, ConstantSDNode *XC, ConstantSDNode *CC, SDValue Y,
unsigned OldShiftOpcode, unsigned NewShiftOpcode,
SelectionDAG &DAG) const {
// Does baseline recommend not to perform the fold by default?
if (!TargetLowering::shouldProduceAndByConstByHoistingConstFromShiftsLHSOfAnd(
X, XC, CC, Y, OldShiftOpcode, NewShiftOpcode, DAG))
return false;
// Else, if this is a vector shift, prefer 'shl'.
return X.getValueType().isScalarInteger() || NewShiftOpcode == ISD::SHL;
}
TargetLowering::ShiftLegalizationStrategy
AArch64TargetLowering::preferredShiftLegalizationStrategy(
SelectionDAG &DAG, SDNode *N, unsigned int ExpansionFactor) const {
if (DAG.getMachineFunction().getFunction().hasMinSize() &&
!Subtarget->isTargetWindows() && !Subtarget->isTargetDarwin())
return ShiftLegalizationStrategy::LowerToLibcall;
return TargetLowering::preferredShiftLegalizationStrategy(DAG, N,
ExpansionFactor);
}
void AArch64TargetLowering::initializeSplitCSR(MachineBasicBlock *Entry) const {
// Update IsSplitCSR in AArch64unctionInfo.
AArch64FunctionInfo *AFI = Entry->getParent()->getInfo<AArch64FunctionInfo>();
AFI->setIsSplitCSR(true);
}
void AArch64TargetLowering::insertCopiesSplitCSR(
MachineBasicBlock *Entry,
const SmallVectorImpl<MachineBasicBlock *> &Exits) const {
const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
const MCPhysReg *IStart = TRI->getCalleeSavedRegsViaCopy(Entry->getParent());
if (!IStart)
return;
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
MachineRegisterInfo *MRI = &Entry->getParent()->getRegInfo();
MachineBasicBlock::iterator MBBI = Entry->begin();
for (const MCPhysReg *I = IStart; *I; ++I) {
const TargetRegisterClass *RC = nullptr;
if (AArch64::GPR64RegClass.contains(*I))
RC = &AArch64::GPR64RegClass;
else if (AArch64::FPR64RegClass.contains(*I))
RC = &AArch64::FPR64RegClass;
else
llvm_unreachable("Unexpected register class in CSRsViaCopy!");
Register NewVR = MRI->createVirtualRegister(RC);
// Create copy from CSR to a virtual register.
// FIXME: this currently does not emit CFI pseudo-instructions, it works
// fine for CXX_FAST_TLS since the C++-style TLS access functions should be
// nounwind. If we want to generalize this later, we may need to emit
// CFI pseudo-instructions.
assert(Entry->getParent()->getFunction().hasFnAttribute(
Attribute::NoUnwind) &&
"Function should be nounwind in insertCopiesSplitCSR!");
Entry->addLiveIn(*I);
BuildMI(*Entry, MBBI, DebugLoc(), TII->get(TargetOpcode::COPY), NewVR)
.addReg(*I);
// Insert the copy-back instructions right before the terminator.
for (auto *Exit : Exits)
BuildMI(*Exit, Exit->getFirstTerminator(), DebugLoc(),
TII->get(TargetOpcode::COPY), *I)
.addReg(NewVR);
}
}
bool AArch64TargetLowering::isIntDivCheap(EVT VT, AttributeList Attr) const {
// Integer division on AArch64 is expensive. However, when aggressively
// optimizing for code size, we prefer to use a div instruction, as it is
// usually smaller than the alternative sequence.
// The exception to this is vector division. Since AArch64 doesn't have vector
// integer division, leaving the division as-is is a loss even in terms of
// size, because it will have to be scalarized, while the alternative code
// sequence can be performed in vector form.
bool OptSize = Attr.hasFnAttr(Attribute::MinSize);
return OptSize && !VT.isVector();
}
bool AArch64TargetLowering::canMergeStoresTo(unsigned AddressSpace, EVT MemVT,
const MachineFunction &MF) const {
// Avoid merging stores into fixed-length vectors when Neon is unavailable.
// In future, we could allow this when SVE is available, but currently,
// the SVE lowerings for BUILD_VECTOR are limited to a few specific cases (and
// the general lowering may introduce stack spills/reloads).
if (MemVT.isFixedLengthVector() && !Subtarget->isNeonAvailable())
return false;
// Do not merge to float value size (128 bytes) if no implicit float attribute
// is set.
bool NoFloat = MF.getFunction().hasFnAttribute(Attribute::NoImplicitFloat);
return !NoFloat || MemVT.getSizeInBits() <= 64;
}
bool AArch64TargetLowering::preferIncOfAddToSubOfNot(EVT VT) const {
// We want inc-of-add for scalars and sub-of-not for vectors.
return VT.isScalarInteger();
}
bool AArch64TargetLowering::shouldConvertFpToSat(unsigned Op, EVT FPVT,
EVT VT) const {
// v8f16 without fp16 need to be extended to v8f32, which is more difficult to
// legalize.
if (FPVT == MVT::v8f16 && !Subtarget->hasFullFP16())
return false;
if (FPVT == MVT::v8bf16)
return false;
return TargetLowering::shouldConvertFpToSat(Op, FPVT, VT);
}
bool AArch64TargetLowering::shouldExpandCmpUsingSelects(EVT VT) const {
// Expand scalar and SVE operations using selects. Neon vectors prefer sub to
// avoid vselect becoming bsl / unrolling.
return !VT.isFixedLengthVector();
}
MachineInstr *
AArch64TargetLowering::EmitKCFICheck(MachineBasicBlock &MBB,
MachineBasicBlock::instr_iterator &MBBI,
const TargetInstrInfo *TII) const {
assert(MBBI->isCall() && MBBI->getCFIType() &&
"Invalid call instruction for a KCFI check");
switch (MBBI->getOpcode()) {
case AArch64::BLR:
case AArch64::BLRNoIP:
case AArch64::TCRETURNri:
case AArch64::TCRETURNrix16x17:
case AArch64::TCRETURNrix17:
case AArch64::TCRETURNrinotx16:
break;
default:
llvm_unreachable("Unexpected CFI call opcode");
}
MachineOperand &Target = MBBI->getOperand(0);
assert(Target.isReg() && "Invalid target operand for an indirect call");
Target.setIsRenamable(false);
return BuildMI(MBB, MBBI, MBBI->getDebugLoc(), TII->get(AArch64::KCFI_CHECK))
.addReg(Target.getReg())
.addImm(MBBI->getCFIType())
.getInstr();
}
bool AArch64TargetLowering::enableAggressiveFMAFusion(EVT VT) const {
return Subtarget->hasAggressiveFMA() && VT.isFloatingPoint();
}
unsigned
AArch64TargetLowering::getVaListSizeInBits(const DataLayout &DL) const {
if (Subtarget->isTargetDarwin() || Subtarget->isTargetWindows())
return getPointerTy(DL).getSizeInBits();
return 3 * getPointerTy(DL).getSizeInBits() + 2 * 32;
}
void AArch64TargetLowering::finalizeLowering(MachineFunction &MF) const {
MachineFrameInfo &MFI = MF.getFrameInfo();
// If we have any vulnerable SVE stack objects then the stack protector
// needs to be placed at the top of the SVE stack area, as the SVE locals
// are placed above the other locals, so we allocate it as if it were a
// scalable vector.
// FIXME: It may be worthwhile having a specific interface for this rather
// than doing it here in finalizeLowering.
if (MFI.hasStackProtectorIndex()) {
for (unsigned int i = 0, e = MFI.getObjectIndexEnd(); i != e; ++i) {
if (MFI.getStackID(i) == TargetStackID::ScalableVector &&
MFI.getObjectSSPLayout(i) != MachineFrameInfo::SSPLK_None) {
MFI.setStackID(MFI.getStackProtectorIndex(),
TargetStackID::ScalableVector);
MFI.setObjectAlignment(MFI.getStackProtectorIndex(), Align(16));
break;
}
}
}
MFI.computeMaxCallFrameSize(MF);
TargetLoweringBase::finalizeLowering(MF);
}
// Unlike X86, we let frame lowering assign offsets to all catch objects.
bool AArch64TargetLowering::needsFixedCatchObjects() const {
return false;
}
bool AArch64TargetLowering::shouldLocalize(
const MachineInstr &MI, const TargetTransformInfo *TTI) const {
auto &MF = *MI.getMF();
auto &MRI = MF.getRegInfo();
auto maxUses = [](unsigned RematCost) {
// A cost of 1 means remats are basically free.
if (RematCost == 1)
return std::numeric_limits<unsigned>::max();
if (RematCost == 2)
return 2U;
// Remat is too expensive, only sink if there's one user.
if (RematCost > 2)
return 1U;
llvm_unreachable("Unexpected remat cost");
};
unsigned Opc = MI.getOpcode();
switch (Opc) {
case TargetOpcode::G_GLOBAL_VALUE: {
// On Darwin, TLS global vars get selected into function calls, which
// we don't want localized, as they can get moved into the middle of a
// another call sequence.
const GlobalValue &GV = *MI.getOperand(1).getGlobal();
if (GV.isThreadLocal() && Subtarget->isTargetMachO())
return false;
return true; // Always localize G_GLOBAL_VALUE to avoid high reg pressure.
}
case TargetOpcode::G_FCONSTANT:
case TargetOpcode::G_CONSTANT: {
const ConstantInt *CI;
unsigned AdditionalCost = 0;
if (Opc == TargetOpcode::G_CONSTANT)
CI = MI.getOperand(1).getCImm();
else {
LLT Ty = MRI.getType(MI.getOperand(0).getReg());
// We try to estimate cost of 32/64b fpimms, as they'll likely be
// materialized as integers.
if (Ty.getScalarSizeInBits() != 32 && Ty.getScalarSizeInBits() != 64)
break;
auto APF = MI.getOperand(1).getFPImm()->getValueAPF();
bool OptForSize = MF.getFunction().hasOptSize();
if (isFPImmLegal(APF, EVT::getFloatingPointVT(Ty.getScalarSizeInBits()),
OptForSize))
return true; // Constant should be cheap.
CI =
ConstantInt::get(MF.getFunction().getContext(), APF.bitcastToAPInt());
// FP materialization also costs an extra move, from gpr to fpr.
AdditionalCost = 1;
}
APInt Imm = CI->getValue();
InstructionCost Cost = TTI->getIntImmCost(
Imm, CI->getType(), TargetTransformInfo::TCK_CodeSize);
assert(Cost.isValid() && "Expected a valid imm cost");
unsigned RematCost = Cost.getValue();
RematCost += AdditionalCost;
Register Reg = MI.getOperand(0).getReg();
unsigned MaxUses = maxUses(RematCost);
// Don't pass UINT_MAX sentinel value to hasAtMostUserInstrs().
if (MaxUses == std::numeric_limits<unsigned>::max())
--MaxUses;
return MRI.hasAtMostUserInstrs(Reg, MaxUses);
}
// If we legalized G_GLOBAL_VALUE into ADRP + G_ADD_LOW, mark both as being
// localizable.
case AArch64::ADRP:
case AArch64::G_ADD_LOW:
// Need to localize G_PTR_ADD so that G_GLOBAL_VALUE can be localized too.
case TargetOpcode::G_PTR_ADD:
return true;
default:
break;
}
return TargetLoweringBase::shouldLocalize(MI, TTI);
}
bool AArch64TargetLowering::fallBackToDAGISel(const Instruction &Inst) const {
// Fallback for scalable vectors.
// Note that if EnableSVEGISel is true, we allow scalable vector types for
// all instructions, regardless of whether they are actually supported.
if (!EnableSVEGISel) {
if (Inst.getType()->isScalableTy()) {
return true;
}
for (unsigned i = 0; i < Inst.getNumOperands(); ++i)
if (Inst.getOperand(i)->getType()->isScalableTy())
return true;
if (const AllocaInst *AI = dyn_cast<AllocaInst>(&Inst)) {
if (AI->getAllocatedType()->isScalableTy())
return true;
}
}
// Checks to allow the use of SME instructions
if (auto *Base = dyn_cast<CallBase>(&Inst)) {
auto CallAttrs = SMECallAttrs(*Base);
if (CallAttrs.requiresSMChange() || CallAttrs.requiresLazySave() ||
CallAttrs.requiresPreservingZT0() ||
CallAttrs.requiresPreservingAllZAState())
return true;
}
return false;
}
// Return the largest legal scalable vector type that matches VT's element type.
static EVT getContainerForFixedLengthVector(SelectionDAG &DAG, EVT VT) {
assert(VT.isFixedLengthVector() &&
DAG.getTargetLoweringInfo().isTypeLegal(VT) &&
"Expected legal fixed length vector!");
switch (VT.getVectorElementType().getSimpleVT().SimpleTy) {
default:
llvm_unreachable("unexpected element type for SVE container");
case MVT::i8:
return EVT(MVT::nxv16i8);
case MVT::i16:
return EVT(MVT::nxv8i16);
case MVT::i32:
return EVT(MVT::nxv4i32);
case MVT::i64:
return EVT(MVT::nxv2i64);
case MVT::bf16:
return EVT(MVT::nxv8bf16);
case MVT::f16:
return EVT(MVT::nxv8f16);
case MVT::f32:
return EVT(MVT::nxv4f32);
case MVT::f64:
return EVT(MVT::nxv2f64);
}
}
// Return a predicate with active lanes corresponding to the extent of VT.
static SDValue getPredicateForFixedLengthVector(SelectionDAG &DAG, SDLoc &DL,
EVT VT) {
assert(VT.isFixedLengthVector() &&
DAG.getTargetLoweringInfo().isTypeLegal(VT) &&
"Expected legal fixed length vector!");
std::optional<unsigned> PgPattern =
getSVEPredPatternFromNumElements(VT.getVectorNumElements());
assert(PgPattern && "Unexpected element count for SVE predicate");
// For vectors that are exactly getMaxSVEVectorSizeInBits big, we can use
// AArch64SVEPredPattern::all, which can enable the use of unpredicated
// variants of instructions when available.
const auto &Subtarget = DAG.getSubtarget<AArch64Subtarget>();
unsigned MinSVESize = Subtarget.getMinSVEVectorSizeInBits();
unsigned MaxSVESize = Subtarget.getMaxSVEVectorSizeInBits();
if (MaxSVESize && MinSVESize == MaxSVESize &&
MaxSVESize == VT.getSizeInBits())
PgPattern = AArch64SVEPredPattern::all;
MVT MaskVT;
switch (VT.getVectorElementType().getSimpleVT().SimpleTy) {
default:
llvm_unreachable("unexpected element type for SVE predicate");
case MVT::i8:
MaskVT = MVT::nxv16i1;
break;
case MVT::i16:
case MVT::f16:
case MVT::bf16:
MaskVT = MVT::nxv8i1;
break;
case MVT::i32:
case MVT::f32:
MaskVT = MVT::nxv4i1;
break;
case MVT::i64:
case MVT::f64:
MaskVT = MVT::nxv2i1;
break;
}
return getPTrue(DAG, DL, MaskVT, *PgPattern);
}
static SDValue getPredicateForScalableVector(SelectionDAG &DAG, SDLoc &DL,
EVT VT) {
assert(VT.isScalableVector() && DAG.getTargetLoweringInfo().isTypeLegal(VT) &&
"Expected legal scalable vector!");
auto PredTy = VT.changeVectorElementType(MVT::i1);
return getPTrue(DAG, DL, PredTy, AArch64SVEPredPattern::all);
}
static SDValue getPredicateForVector(SelectionDAG &DAG, SDLoc &DL, EVT VT) {
if (VT.isFixedLengthVector())
return getPredicateForFixedLengthVector(DAG, DL, VT);
return getPredicateForScalableVector(DAG, DL, VT);
}
// Grow V to consume an entire SVE register.
static SDValue convertToScalableVector(SelectionDAG &DAG, EVT VT, SDValue V) {
assert(VT.isScalableVector() &&
"Expected to convert into a scalable vector!");
assert(V.getValueType().isFixedLengthVector() &&
"Expected a fixed length vector operand!");
SDLoc DL(V);
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
return DAG.getNode(ISD::INSERT_SUBVECTOR, DL, VT, DAG.getUNDEF(VT), V, Zero);
}
// Shrink V so it's just big enough to maintain a VT's worth of data.
static SDValue convertFromScalableVector(SelectionDAG &DAG, EVT VT, SDValue V) {
assert(VT.isFixedLengthVector() &&
"Expected to convert into a fixed length vector!");
assert(V.getValueType().isScalableVector() &&
"Expected a scalable vector operand!");
SDLoc DL(V);
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, V, Zero);
}
// Convert all fixed length vector loads larger than NEON to masked_loads.
SDValue AArch64TargetLowering::LowerFixedLengthVectorLoadToSVE(
SDValue Op, SelectionDAG &DAG) const {
auto Load = cast<LoadSDNode>(Op);
SDLoc DL(Op);
EVT VT = Op.getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
EVT LoadVT = ContainerVT;
EVT MemVT = Load->getMemoryVT();
auto Pg = getPredicateForFixedLengthVector(DAG, DL, VT);
if (VT.isFloatingPoint()) {
LoadVT = ContainerVT.changeTypeToInteger();
MemVT = MemVT.changeTypeToInteger();
}
SDValue NewLoad = DAG.getMaskedLoad(
LoadVT, DL, Load->getChain(), Load->getBasePtr(), Load->getOffset(), Pg,
DAG.getUNDEF(LoadVT), MemVT, Load->getMemOperand(),
Load->getAddressingMode(), Load->getExtensionType());
SDValue Result = NewLoad;
if (VT.isFloatingPoint() && Load->getExtensionType() == ISD::EXTLOAD) {
EVT ExtendVT = ContainerVT.changeVectorElementType(
Load->getMemoryVT().getVectorElementType());
Result = getSVESafeBitCast(ExtendVT, Result, DAG);
Result = DAG.getNode(AArch64ISD::FP_EXTEND_MERGE_PASSTHRU, DL, ContainerVT,
Pg, Result, DAG.getUNDEF(ContainerVT));
} else if (VT.isFloatingPoint()) {
Result = DAG.getNode(ISD::BITCAST, DL, ContainerVT, Result);
}
Result = convertFromScalableVector(DAG, VT, Result);
SDValue MergedValues[2] = {Result, NewLoad.getValue(1)};
return DAG.getMergeValues(MergedValues, DL);
}
static SDValue convertFixedMaskToScalableVector(SDValue Mask,
SelectionDAG &DAG) {
SDLoc DL(Mask);
EVT InVT = Mask.getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, InVT);
SDValue Pg = getPredicateForFixedLengthVector(DAG, DL, InVT);
if (ISD::isBuildVectorAllOnes(Mask.getNode()))
return Pg;
bool InvertCond = false;
if (isBitwiseNot(Mask)) {
InvertCond = true;
Mask = Mask.getOperand(0);
}
SDValue Op1, Op2;
ISD::CondCode CC;
// When Mask is the result of a SETCC, it's better to regenerate the compare.
if (Mask.getOpcode() == ISD::SETCC) {
Op1 = convertToScalableVector(DAG, ContainerVT, Mask.getOperand(0));
Op2 = convertToScalableVector(DAG, ContainerVT, Mask.getOperand(1));
CC = cast<CondCodeSDNode>(Mask.getOperand(2))->get();
} else {
Op1 = convertToScalableVector(DAG, ContainerVT, Mask);
Op2 = DAG.getConstant(0, DL, ContainerVT);
CC = ISD::SETNE;
}
if (InvertCond)
CC = getSetCCInverse(CC, Op1.getValueType());
return DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, DL, Pg.getValueType(),
{Pg, Op1, Op2, DAG.getCondCode(CC)});
}
// Convert all fixed length vector loads larger than NEON to masked_loads.
SDValue AArch64TargetLowering::LowerFixedLengthVectorMLoadToSVE(
SDValue Op, SelectionDAG &DAG) const {
auto Load = cast<MaskedLoadSDNode>(Op);
SDLoc DL(Op);
EVT VT = Op.getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
SDValue Mask = Load->getMask();
// If this is an extending load and the mask type is not the same as
// load's type then we have to extend the mask type.
if (VT.getScalarSizeInBits() > Mask.getValueType().getScalarSizeInBits()) {
assert(Load->getExtensionType() != ISD::NON_EXTLOAD &&
"Incorrect mask type");
Mask = DAG.getNode(ISD::SIGN_EXTEND, DL, VT, Mask);
}
Mask = convertFixedMaskToScalableVector(Mask, DAG);
SDValue PassThru;
bool IsPassThruZeroOrUndef = false;
if (Load->getPassThru()->isUndef()) {
PassThru = DAG.getUNDEF(ContainerVT);
IsPassThruZeroOrUndef = true;
} else {
if (ContainerVT.isInteger())
PassThru = DAG.getConstant(0, DL, ContainerVT);
else
PassThru = DAG.getConstantFP(0, DL, ContainerVT);
if (isZerosVector(Load->getPassThru().getNode()))
IsPassThruZeroOrUndef = true;
}
SDValue NewLoad = DAG.getMaskedLoad(
ContainerVT, DL, Load->getChain(), Load->getBasePtr(), Load->getOffset(),
Mask, PassThru, Load->getMemoryVT(), Load->getMemOperand(),
Load->getAddressingMode(), Load->getExtensionType());
SDValue Result = NewLoad;
if (!IsPassThruZeroOrUndef) {
SDValue OldPassThru =
convertToScalableVector(DAG, ContainerVT, Load->getPassThru());
Result = DAG.getSelect(DL, ContainerVT, Mask, Result, OldPassThru);
}
Result = convertFromScalableVector(DAG, VT, Result);
SDValue MergedValues[2] = {Result, NewLoad.getValue(1)};
return DAG.getMergeValues(MergedValues, DL);
}
// Convert all fixed length vector stores larger than NEON to masked_stores.
SDValue AArch64TargetLowering::LowerFixedLengthVectorStoreToSVE(
SDValue Op, SelectionDAG &DAG) const {
auto Store = cast<StoreSDNode>(Op);
SDLoc DL(Op);
EVT VT = Store->getValue().getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
EVT MemVT = Store->getMemoryVT();
auto Pg = getPredicateForFixedLengthVector(DAG, DL, VT);
auto NewValue = convertToScalableVector(DAG, ContainerVT, Store->getValue());
if (VT.isFloatingPoint() && Store->isTruncatingStore()) {
EVT TruncVT = ContainerVT.changeVectorElementType(
Store->getMemoryVT().getVectorElementType());
MemVT = MemVT.changeTypeToInteger();
NewValue = DAG.getNode(AArch64ISD::FP_ROUND_MERGE_PASSTHRU, DL, TruncVT, Pg,
NewValue, DAG.getTargetConstant(0, DL, MVT::i64),
DAG.getUNDEF(TruncVT));
NewValue =
getSVESafeBitCast(ContainerVT.changeTypeToInteger(), NewValue, DAG);
} else if (VT.isFloatingPoint()) {
MemVT = MemVT.changeTypeToInteger();
NewValue =
getSVESafeBitCast(ContainerVT.changeTypeToInteger(), NewValue, DAG);
}
return DAG.getMaskedStore(Store->getChain(), DL, NewValue,
Store->getBasePtr(), Store->getOffset(), Pg, MemVT,
Store->getMemOperand(), Store->getAddressingMode(),
Store->isTruncatingStore());
}
SDValue AArch64TargetLowering::LowerFixedLengthVectorMStoreToSVE(
SDValue Op, SelectionDAG &DAG) const {
auto *Store = cast<MaskedStoreSDNode>(Op);
SDLoc DL(Op);
EVT VT = Store->getValue().getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
auto NewValue = convertToScalableVector(DAG, ContainerVT, Store->getValue());
SDValue Mask = convertFixedMaskToScalableVector(Store->getMask(), DAG);
return DAG.getMaskedStore(
Store->getChain(), DL, NewValue, Store->getBasePtr(), Store->getOffset(),
Mask, Store->getMemoryVT(), Store->getMemOperand(),
Store->getAddressingMode(), Store->isTruncatingStore());
}
SDValue AArch64TargetLowering::LowerFixedLengthVectorIntDivideToSVE(
SDValue Op, SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT VT = Op.getValueType();
EVT EltVT = VT.getVectorElementType();
bool Signed = Op.getOpcode() == ISD::SDIV;
unsigned PredOpcode = Signed ? AArch64ISD::SDIV_PRED : AArch64ISD::UDIV_PRED;
bool Negated;
uint64_t SplatVal;
if (Signed && isPow2Splat(Op.getOperand(1), SplatVal, Negated)) {
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
SDValue Op1 = convertToScalableVector(DAG, ContainerVT, Op.getOperand(0));
SDValue Op2 = DAG.getTargetConstant(Log2_64(SplatVal), DL, MVT::i32);
SDValue Pg = getPredicateForFixedLengthVector(DAG, DL, VT);
SDValue Res =
DAG.getNode(AArch64ISD::SRAD_MERGE_OP1, DL, ContainerVT, Pg, Op1, Op2);
if (Negated)
Res = DAG.getNode(ISD::SUB, DL, ContainerVT,
DAG.getConstant(0, DL, ContainerVT), Res);
return convertFromScalableVector(DAG, VT, Res);
}
// Scalable vector i32/i64 DIV is supported.
if (EltVT == MVT::i32 || EltVT == MVT::i64)
return LowerToPredicatedOp(Op, DAG, PredOpcode);
// Scalable vector i8/i16 DIV is not supported. Promote it to i32.
EVT HalfVT = VT.getHalfNumVectorElementsVT(*DAG.getContext());
EVT PromVT = HalfVT.widenIntegerVectorElementType(*DAG.getContext());
unsigned ExtendOpcode = Signed ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND;
// If the wider type is legal: extend, op, and truncate.
EVT WideVT = VT.widenIntegerVectorElementType(*DAG.getContext());
if (DAG.getTargetLoweringInfo().isTypeLegal(WideVT)) {
SDValue Op0 = DAG.getNode(ExtendOpcode, DL, WideVT, Op.getOperand(0));
SDValue Op1 = DAG.getNode(ExtendOpcode, DL, WideVT, Op.getOperand(1));
SDValue Div = DAG.getNode(Op.getOpcode(), DL, WideVT, Op0, Op1);
return DAG.getNode(ISD::TRUNCATE, DL, VT, Div);
}
auto HalveAndExtendVector = [&DAG, &DL, &HalfVT, &PromVT,
&ExtendOpcode](SDValue Op) {
SDValue IdxZero = DAG.getConstant(0, DL, MVT::i64);
SDValue IdxHalf =
DAG.getConstant(HalfVT.getVectorNumElements(), DL, MVT::i64);
SDValue Lo = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, HalfVT, Op, IdxZero);
SDValue Hi = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, HalfVT, Op, IdxHalf);
return std::pair<SDValue, SDValue>(
{DAG.getNode(ExtendOpcode, DL, PromVT, Lo),
DAG.getNode(ExtendOpcode, DL, PromVT, Hi)});
};
// If wider type is not legal: split, extend, op, trunc and concat.
auto [Op0LoExt, Op0HiExt] = HalveAndExtendVector(Op.getOperand(0));
auto [Op1LoExt, Op1HiExt] = HalveAndExtendVector(Op.getOperand(1));
SDValue Lo = DAG.getNode(Op.getOpcode(), DL, PromVT, Op0LoExt, Op1LoExt);
SDValue Hi = DAG.getNode(Op.getOpcode(), DL, PromVT, Op0HiExt, Op1HiExt);
SDValue LoTrunc = DAG.getNode(ISD::TRUNCATE, DL, HalfVT, Lo);
SDValue HiTrunc = DAG.getNode(ISD::TRUNCATE, DL, HalfVT, Hi);
return DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, {LoTrunc, HiTrunc});
}
SDValue AArch64TargetLowering::LowerFixedLengthVectorIntExtendToSVE(
SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
SDLoc DL(Op);
SDValue Val = Op.getOperand(0);
EVT ContainerVT = getContainerForFixedLengthVector(DAG, Val.getValueType());
Val = convertToScalableVector(DAG, ContainerVT, Val);
bool Signed = Op.getOpcode() == ISD::SIGN_EXTEND;
unsigned ExtendOpc = Signed ? AArch64ISD::SUNPKLO : AArch64ISD::UUNPKLO;
// Repeatedly unpack Val until the result is of the desired element type.
switch (ContainerVT.getSimpleVT().SimpleTy) {
default:
llvm_unreachable("unimplemented container type");
case MVT::nxv16i8:
Val = DAG.getNode(ExtendOpc, DL, MVT::nxv8i16, Val);
if (VT.getVectorElementType() == MVT::i16)
break;
[[fallthrough]];
case MVT::nxv8i16:
Val = DAG.getNode(ExtendOpc, DL, MVT::nxv4i32, Val);
if (VT.getVectorElementType() == MVT::i32)
break;
[[fallthrough]];
case MVT::nxv4i32:
Val = DAG.getNode(ExtendOpc, DL, MVT::nxv2i64, Val);
assert(VT.getVectorElementType() == MVT::i64 && "Unexpected element type!");
break;
}
return convertFromScalableVector(DAG, VT, Val);
}
SDValue AArch64TargetLowering::LowerFixedLengthVectorTruncateToSVE(
SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
SDLoc DL(Op);
SDValue Val = Op.getOperand(0);
EVT ContainerVT = getContainerForFixedLengthVector(DAG, Val.getValueType());
Val = convertToScalableVector(DAG, ContainerVT, Val);
// Repeatedly truncate Val until the result is of the desired element type.
switch (ContainerVT.getSimpleVT().SimpleTy) {
default:
llvm_unreachable("unimplemented container type");
case MVT::nxv2i64:
Val = DAG.getNode(ISD::BITCAST, DL, MVT::nxv4i32, Val);
Val = DAG.getNode(AArch64ISD::UZP1, DL, MVT::nxv4i32, Val, Val);
if (VT.getVectorElementType() == MVT::i32)
break;
[[fallthrough]];
case MVT::nxv4i32:
Val = DAG.getNode(ISD::BITCAST, DL, MVT::nxv8i16, Val);
Val = DAG.getNode(AArch64ISD::UZP1, DL, MVT::nxv8i16, Val, Val);
if (VT.getVectorElementType() == MVT::i16)
break;
[[fallthrough]];
case MVT::nxv8i16:
Val = DAG.getNode(ISD::BITCAST, DL, MVT::nxv16i8, Val);
Val = DAG.getNode(AArch64ISD::UZP1, DL, MVT::nxv16i8, Val, Val);
assert(VT.getVectorElementType() == MVT::i8 && "Unexpected element type!");
break;
}
return convertFromScalableVector(DAG, VT, Val);
}
SDValue AArch64TargetLowering::LowerFixedLengthExtractVectorElt(
SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
EVT InVT = Op.getOperand(0).getValueType();
assert(InVT.isFixedLengthVector() && "Expected fixed length vector type!");
SDLoc DL(Op);
EVT ContainerVT = getContainerForFixedLengthVector(DAG, InVT);
SDValue Op0 = convertToScalableVector(DAG, ContainerVT, Op->getOperand(0));
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, VT, Op0, Op.getOperand(1));
}
SDValue AArch64TargetLowering::LowerFixedLengthInsertVectorElt(
SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
SDLoc DL(Op);
EVT InVT = Op.getOperand(0).getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, InVT);
SDValue Op0 = convertToScalableVector(DAG, ContainerVT, Op->getOperand(0));
auto ScalableRes = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, ContainerVT, Op0,
Op.getOperand(1), Op.getOperand(2));
return convertFromScalableVector(DAG, VT, ScalableRes);
}
// Convert vector operation 'Op' to an equivalent predicated operation whereby
// the original operation's type is used to construct a suitable predicate.
// NOTE: The results for inactive lanes are undefined.
SDValue AArch64TargetLowering::LowerToPredicatedOp(SDValue Op,
SelectionDAG &DAG,
unsigned NewOp) const {
EVT VT = Op.getValueType();
SDLoc DL(Op);
auto Pg = getPredicateForVector(DAG, DL, VT);
if (VT.isFixedLengthVector()) {
assert(isTypeLegal(VT) && "Expected only legal fixed-width types");
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
// Create list of operands by converting existing ones to scalable types.
SmallVector<SDValue, 4> Operands = {Pg};
for (const SDValue &V : Op->op_values()) {
if (isa<CondCodeSDNode>(V)) {
Operands.push_back(V);
continue;
}
if (const VTSDNode *VTNode = dyn_cast<VTSDNode>(V)) {
EVT VTArg = VTNode->getVT().getVectorElementType();
EVT NewVTArg = ContainerVT.changeVectorElementType(VTArg);
Operands.push_back(DAG.getValueType(NewVTArg));
continue;
}
assert(isTypeLegal(V.getValueType()) &&
"Expected only legal fixed-width types");
Operands.push_back(convertToScalableVector(DAG, ContainerVT, V));
}
if (isMergePassthruOpcode(NewOp))
Operands.push_back(DAG.getUNDEF(ContainerVT));
auto ScalableRes = DAG.getNode(NewOp, DL, ContainerVT, Operands);
return convertFromScalableVector(DAG, VT, ScalableRes);
}
assert(VT.isScalableVector() && "Only expect to lower scalable vector op!");
SmallVector<SDValue, 4> Operands = {Pg};
for (const SDValue &V : Op->op_values()) {
assert((!V.getValueType().isVector() ||
V.getValueType().isScalableVector()) &&
"Only scalable vectors are supported!");
Operands.push_back(V);
}
if (isMergePassthruOpcode(NewOp))
Operands.push_back(DAG.getUNDEF(VT));
return DAG.getNode(NewOp, DL, VT, Operands, Op->getFlags());
}
// If a fixed length vector operation has no side effects when applied to
// undefined elements, we can safely use scalable vectors to perform the same
// operation without needing to worry about predication.
SDValue AArch64TargetLowering::LowerToScalableOp(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && isTypeLegal(VT) &&
"Only expected to lower fixed length vector operation!");
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
// Create list of operands by converting existing ones to scalable types.
SmallVector<SDValue, 4> Ops;
for (const SDValue &V : Op->op_values()) {
assert(!isa<VTSDNode>(V) && "Unexpected VTSDNode node!");
// Pass through non-vector operands.
if (!V.getValueType().isVector()) {
Ops.push_back(V);
continue;
}
// "cast" fixed length vector to a scalable vector.
assert(V.getValueType().isFixedLengthVector() &&
isTypeLegal(V.getValueType()) &&
"Only fixed length vectors are supported!");
Ops.push_back(convertToScalableVector(DAG, ContainerVT, V));
}
auto ScalableRes = DAG.getNode(Op.getOpcode(), SDLoc(Op), ContainerVT, Ops);
return convertFromScalableVector(DAG, VT, ScalableRes);
}
SDValue AArch64TargetLowering::LowerVECREDUCE_SEQ_FADD(SDValue ScalarOp,
SelectionDAG &DAG) const {
SDLoc DL(ScalarOp);
SDValue AccOp = ScalarOp.getOperand(0);
SDValue VecOp = ScalarOp.getOperand(1);
EVT SrcVT = VecOp.getValueType();
EVT ResVT = SrcVT.getVectorElementType();
EVT ContainerVT = SrcVT;
if (SrcVT.isFixedLengthVector()) {
ContainerVT = getContainerForFixedLengthVector(DAG, SrcVT);
VecOp = convertToScalableVector(DAG, ContainerVT, VecOp);
}
SDValue Pg = getPredicateForVector(DAG, DL, SrcVT);
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
// Convert operands to Scalable.
AccOp = DAG.getNode(ISD::INSERT_VECTOR_ELT, DL, ContainerVT,
DAG.getUNDEF(ContainerVT), AccOp, Zero);
// Perform reduction.
SDValue Rdx = DAG.getNode(AArch64ISD::FADDA_PRED, DL, ContainerVT,
Pg, AccOp, VecOp);
return DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ResVT, Rdx, Zero);
}
SDValue AArch64TargetLowering::LowerPredReductionToSVE(SDValue ReduceOp,
SelectionDAG &DAG) const {
SDLoc DL(ReduceOp);
SDValue Op = ReduceOp.getOperand(0);
EVT OpVT = Op.getValueType();
EVT VT = ReduceOp.getValueType();
if (!OpVT.isScalableVector() || OpVT.getVectorElementType() != MVT::i1)
return SDValue();
SDValue Pg = getPredicateForVector(DAG, DL, OpVT);
switch (ReduceOp.getOpcode()) {
default:
return SDValue();
case ISD::VECREDUCE_OR:
if (isAllActivePredicate(DAG, Pg) && OpVT == MVT::nxv16i1)
// The predicate can be 'Op' because
// vecreduce_or(Op & <all true>) <=> vecreduce_or(Op).
return getPTest(DAG, VT, Op, Op, AArch64CC::ANY_ACTIVE);
else
return getPTest(DAG, VT, Pg, Op, AArch64CC::ANY_ACTIVE);
case ISD::VECREDUCE_AND: {
Op = DAG.getNode(ISD::XOR, DL, OpVT, Op, Pg);
return getPTest(DAG, VT, Pg, Op, AArch64CC::NONE_ACTIVE);
}
case ISD::VECREDUCE_XOR: {
SDValue ID =
DAG.getTargetConstant(Intrinsic::aarch64_sve_cntp, DL, MVT::i64);
if (OpVT == MVT::nxv1i1) {
// Emulate a CNTP on .Q using .D and a different governing predicate.
Pg = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, MVT::nxv2i1, Pg);
Op = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, MVT::nxv2i1, Op);
}
SDValue Cntp =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, MVT::i64, ID, Pg, Op);
return DAG.getAnyExtOrTrunc(Cntp, DL, VT);
}
}
return SDValue();
}
SDValue AArch64TargetLowering::LowerReductionToSVE(unsigned Opcode,
SDValue ScalarOp,
SelectionDAG &DAG) const {
SDLoc DL(ScalarOp);
SDValue VecOp = ScalarOp.getOperand(0);
EVT SrcVT = VecOp.getValueType();
if (useSVEForFixedLengthVectorVT(
SrcVT,
/*OverrideNEON=*/Subtarget->useSVEForFixedLengthVectors())) {
EVT ContainerVT = getContainerForFixedLengthVector(DAG, SrcVT);
VecOp = convertToScalableVector(DAG, ContainerVT, VecOp);
}
// Lower VECREDUCE_ADD of nxv2i1-nxv16i1 to CNTP rather than UADDV.
if (ScalarOp.getOpcode() == ISD::VECREDUCE_ADD &&
VecOp.getOpcode() == ISD::ZERO_EXTEND) {
SDValue BoolVec = VecOp.getOperand(0);
if (BoolVec.getValueType().getVectorElementType() == MVT::i1) {
// CNTP(BoolVec & BoolVec) <=> CNTP(BoolVec & PTRUE)
SDValue CntpOp = DAG.getNode(
ISD::INTRINSIC_WO_CHAIN, DL, MVT::i64,
DAG.getTargetConstant(Intrinsic::aarch64_sve_cntp, DL, MVT::i64),
BoolVec, BoolVec);
return DAG.getAnyExtOrTrunc(CntpOp, DL, ScalarOp.getValueType());
}
}
// UADDV always returns an i64 result.
EVT ResVT = (Opcode == AArch64ISD::UADDV_PRED) ? MVT::i64 :
SrcVT.getVectorElementType();
EVT RdxVT = SrcVT;
if (SrcVT.isFixedLengthVector() || Opcode == AArch64ISD::UADDV_PRED)
RdxVT = getPackedSVEVectorVT(ResVT);
SDValue Pg = getPredicateForVector(DAG, DL, SrcVT);
SDValue Rdx = DAG.getNode(Opcode, DL, RdxVT, Pg, VecOp);
SDValue Res = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ResVT,
Rdx, DAG.getConstant(0, DL, MVT::i64));
// The VEC_REDUCE nodes expect an element size result.
if (ResVT != ScalarOp.getValueType())
Res = DAG.getAnyExtOrTrunc(Res, DL, ScalarOp.getValueType());
return Res;
}
SDValue
AArch64TargetLowering::LowerFixedLengthVectorSelectToSVE(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
SDLoc DL(Op);
EVT InVT = Op.getOperand(1).getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, InVT);
SDValue Op1 = convertToScalableVector(DAG, ContainerVT, Op->getOperand(1));
SDValue Op2 = convertToScalableVector(DAG, ContainerVT, Op->getOperand(2));
// Convert the mask to a predicated (NOTE: We don't need to worry about
// inactive lanes since VSELECT is safe when given undefined elements).
EVT MaskVT = Op.getOperand(0).getValueType();
EVT MaskContainerVT = getContainerForFixedLengthVector(DAG, MaskVT);
auto Mask = convertToScalableVector(DAG, MaskContainerVT, Op.getOperand(0));
Mask = DAG.getNode(ISD::TRUNCATE, DL,
MaskContainerVT.changeVectorElementType(MVT::i1), Mask);
auto ScalableRes = DAG.getNode(ISD::VSELECT, DL, ContainerVT,
Mask, Op1, Op2);
return convertFromScalableVector(DAG, VT, ScalableRes);
}
SDValue AArch64TargetLowering::LowerFixedLengthVectorSetccToSVE(
SDValue Op, SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT InVT = Op.getOperand(0).getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, InVT);
assert(InVT.isFixedLengthVector() && isTypeLegal(InVT) &&
"Only expected to lower fixed length vector operation!");
assert(Op.getValueType() == InVT.changeTypeToInteger() &&
"Expected integer result of the same bit length as the inputs!");
auto Op1 = convertToScalableVector(DAG, ContainerVT, Op.getOperand(0));
auto Op2 = convertToScalableVector(DAG, ContainerVT, Op.getOperand(1));
auto Pg = getPredicateForFixedLengthVector(DAG, DL, InVT);
EVT CmpVT = Pg.getValueType();
auto Cmp = DAG.getNode(AArch64ISD::SETCC_MERGE_ZERO, DL, CmpVT,
{Pg, Op1, Op2, Op.getOperand(2)});
EVT PromoteVT = ContainerVT.changeTypeToInteger();
auto Promote = DAG.getBoolExtOrTrunc(Cmp, DL, PromoteVT, InVT);
return convertFromScalableVector(DAG, Op.getValueType(), Promote);
}
SDValue
AArch64TargetLowering::LowerFixedLengthBitcastToSVE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
auto SrcOp = Op.getOperand(0);
EVT VT = Op.getValueType();
EVT ContainerDstVT = getContainerForFixedLengthVector(DAG, VT);
EVT ContainerSrcVT =
getContainerForFixedLengthVector(DAG, SrcOp.getValueType());
SrcOp = convertToScalableVector(DAG, ContainerSrcVT, SrcOp);
Op = DAG.getNode(ISD::BITCAST, DL, ContainerDstVT, SrcOp);
return convertFromScalableVector(DAG, VT, Op);
}
SDValue AArch64TargetLowering::LowerFixedLengthConcatVectorsToSVE(
SDValue Op, SelectionDAG &DAG) const {
SDLoc DL(Op);
unsigned NumOperands = Op->getNumOperands();
assert(NumOperands > 1 && isPowerOf2_32(NumOperands) &&
"Unexpected number of operands in CONCAT_VECTORS");
auto SrcOp1 = Op.getOperand(0);
auto SrcOp2 = Op.getOperand(1);
EVT VT = Op.getValueType();
EVT SrcVT = SrcOp1.getValueType();
// Match a splat of 128b segments that fit in a single register.
if (SrcVT.is128BitVector() && all_equal(Op.getNode()->op_values())) {
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
SDValue Splat =
DAG.getNode(AArch64ISD::DUPLANE128, DL, ContainerVT,
convertToScalableVector(DAG, ContainerVT, SrcOp1),
DAG.getConstant(0, DL, MVT::i64, /*isTarget=*/true));
return convertFromScalableVector(DAG, VT, Splat);
}
if (NumOperands > 2) {
SmallVector<SDValue, 4> Ops;
EVT PairVT = SrcVT.getDoubleNumVectorElementsVT(*DAG.getContext());
for (unsigned I = 0; I < NumOperands; I += 2)
Ops.push_back(DAG.getNode(ISD::CONCAT_VECTORS, DL, PairVT,
Op->getOperand(I), Op->getOperand(I + 1)));
return DAG.getNode(ISD::CONCAT_VECTORS, DL, VT, Ops);
}
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
SDValue Pg = getPredicateForFixedLengthVector(DAG, DL, SrcVT);
SrcOp1 = convertToScalableVector(DAG, ContainerVT, SrcOp1);
SrcOp2 = convertToScalableVector(DAG, ContainerVT, SrcOp2);
Op = DAG.getNode(AArch64ISD::SPLICE, DL, ContainerVT, Pg, SrcOp1, SrcOp2);
return convertFromScalableVector(DAG, VT, Op);
}
SDValue
AArch64TargetLowering::LowerFixedLengthFPExtendToSVE(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
SDLoc DL(Op);
SDValue Val = Op.getOperand(0);
SDValue Pg = getPredicateForVector(DAG, DL, VT);
EVT SrcVT = Val.getValueType();
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
EVT ExtendVT = ContainerVT.changeVectorElementType(
SrcVT.getVectorElementType());
Val = DAG.getNode(ISD::BITCAST, DL, SrcVT.changeTypeToInteger(), Val);
Val = DAG.getNode(ISD::ANY_EXTEND, DL, VT.changeTypeToInteger(), Val);
Val = convertToScalableVector(DAG, ContainerVT.changeTypeToInteger(), Val);
Val = getSVESafeBitCast(ExtendVT, Val, DAG);
Val = DAG.getNode(AArch64ISD::FP_EXTEND_MERGE_PASSTHRU, DL, ContainerVT,
Pg, Val, DAG.getUNDEF(ContainerVT));
return convertFromScalableVector(DAG, VT, Val);
}
SDValue
AArch64TargetLowering::LowerFixedLengthFPRoundToSVE(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
SDLoc DL(Op);
SDValue Val = Op.getOperand(0);
EVT SrcVT = Val.getValueType();
EVT ContainerSrcVT = getContainerForFixedLengthVector(DAG, SrcVT);
EVT RoundVT = ContainerSrcVT.changeVectorElementType(
VT.getVectorElementType());
SDValue Pg = getPredicateForVector(DAG, DL, RoundVT);
Val = convertToScalableVector(DAG, ContainerSrcVT, Val);
Val = DAG.getNode(AArch64ISD::FP_ROUND_MERGE_PASSTHRU, DL, RoundVT, Pg, Val,
Op.getOperand(1), DAG.getUNDEF(RoundVT));
Val = getSVESafeBitCast(ContainerSrcVT.changeTypeToInteger(), Val, DAG);
Val = convertFromScalableVector(DAG, SrcVT.changeTypeToInteger(), Val);
Val = DAG.getNode(ISD::TRUNCATE, DL, VT.changeTypeToInteger(), Val);
return DAG.getNode(ISD::BITCAST, DL, VT, Val);
}
SDValue
AArch64TargetLowering::LowerFixedLengthIntToFPToSVE(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
bool IsSigned = Op.getOpcode() == ISD::SINT_TO_FP;
unsigned Opcode = IsSigned ? AArch64ISD::SINT_TO_FP_MERGE_PASSTHRU
: AArch64ISD::UINT_TO_FP_MERGE_PASSTHRU;
SDLoc DL(Op);
SDValue Val = Op.getOperand(0);
EVT SrcVT = Val.getValueType();
EVT ContainerDstVT = getContainerForFixedLengthVector(DAG, VT);
EVT ContainerSrcVT = getContainerForFixedLengthVector(DAG, SrcVT);
if (VT.bitsGE(SrcVT)) {
SDValue Pg = getPredicateForFixedLengthVector(DAG, DL, VT);
Val = DAG.getNode(IsSigned ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND, DL,
VT.changeTypeToInteger(), Val);
// Safe to use a larger than specified operand because by promoting the
// value nothing has changed from an arithmetic point of view.
Val =
convertToScalableVector(DAG, ContainerDstVT.changeTypeToInteger(), Val);
Val = DAG.getNode(Opcode, DL, ContainerDstVT, Pg, Val,
DAG.getUNDEF(ContainerDstVT));
return convertFromScalableVector(DAG, VT, Val);
} else {
EVT CvtVT = ContainerSrcVT.changeVectorElementType(
ContainerDstVT.getVectorElementType());
SDValue Pg = getPredicateForFixedLengthVector(DAG, DL, SrcVT);
Val = convertToScalableVector(DAG, ContainerSrcVT, Val);
Val = DAG.getNode(Opcode, DL, CvtVT, Pg, Val, DAG.getUNDEF(CvtVT));
Val = getSVESafeBitCast(ContainerSrcVT, Val, DAG);
Val = convertFromScalableVector(DAG, SrcVT, Val);
Val = DAG.getNode(ISD::TRUNCATE, DL, VT.changeTypeToInteger(), Val);
return DAG.getNode(ISD::BITCAST, DL, VT, Val);
}
}
SDValue
AArch64TargetLowering::LowerVECTOR_DEINTERLEAVE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT OpVT = Op.getValueType();
assert(OpVT.isScalableVector() &&
"Expected scalable vector in LowerVECTOR_DEINTERLEAVE.");
// Are multi-register uzp instructions available?
if (Subtarget->hasSME2() && Subtarget->isStreaming() &&
OpVT.getVectorElementType() != MVT::i1) {
Intrinsic::ID IntID;
switch (Op->getNumOperands()) {
default:
return SDValue();
case 2:
IntID = Intrinsic::aarch64_sve_uzp_x2;
break;
case 4:
if (Subtarget->getMinSVEVectorSizeInBits() < 256 &&
OpVT.getScalarSizeInBits() == 64)
return SDValue();
IntID = Intrinsic::aarch64_sve_uzp_x4;
break;
}
SmallVector<SDValue, 5> Ops;
Ops.push_back(DAG.getTargetConstant(IntID, DL, MVT::i64));
Ops.append(Op->op_values().begin(), Op->op_values().end());
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, Op->getVTList(), Ops);
}
if (Op->getNumOperands() != 2)
return SDValue();
SDValue Even = DAG.getNode(AArch64ISD::UZP1, DL, OpVT, Op.getOperand(0),
Op.getOperand(1));
SDValue Odd = DAG.getNode(AArch64ISD::UZP2, DL, OpVT, Op.getOperand(0),
Op.getOperand(1));
return DAG.getMergeValues({Even, Odd}, DL);
}
SDValue AArch64TargetLowering::LowerVECTOR_INTERLEAVE(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT OpVT = Op.getValueType();
assert(OpVT.isScalableVector() &&
"Expected scalable vector in LowerVECTOR_INTERLEAVE.");
// Are multi-register zip instructions available?
if (Subtarget->hasSME2() && Subtarget->isStreaming() &&
OpVT.getVectorElementType() != MVT::i1) {
Intrinsic::ID IntID;
switch (Op->getNumOperands()) {
default:
return SDValue();
case 2:
IntID = Intrinsic::aarch64_sve_zip_x2;
break;
case 4:
if (Subtarget->getMinSVEVectorSizeInBits() < 256 &&
OpVT.getScalarSizeInBits() == 64)
return SDValue();
IntID = Intrinsic::aarch64_sve_zip_x4;
break;
}
SmallVector<SDValue, 5> Ops;
Ops.push_back(DAG.getTargetConstant(IntID, DL, MVT::i64));
Ops.append(Op->op_values().begin(), Op->op_values().end());
return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, Op->getVTList(), Ops);
}
if (Op->getNumOperands() != 2)
return SDValue();
SDValue Lo = DAG.getNode(AArch64ISD::ZIP1, DL, OpVT, Op.getOperand(0),
Op.getOperand(1));
SDValue Hi = DAG.getNode(AArch64ISD::ZIP2, DL, OpVT, Op.getOperand(0),
Op.getOperand(1));
return DAG.getMergeValues({Lo, Hi}, DL);
}
SDValue AArch64TargetLowering::LowerVECTOR_HISTOGRAM(SDValue Op,
SelectionDAG &DAG) const {
// FIXME: Maybe share some code with LowerMGather/Scatter?
MaskedHistogramSDNode *HG = cast<MaskedHistogramSDNode>(Op);
SDLoc DL(HG);
SDValue Chain = HG->getChain();
SDValue Inc = HG->getInc();
SDValue Mask = HG->getMask();
SDValue Ptr = HG->getBasePtr();
SDValue Index = HG->getIndex();
SDValue Scale = HG->getScale();
SDValue IntID = HG->getIntID();
// The Intrinsic ID determines the type of update operation.
[[maybe_unused]] ConstantSDNode *CID = cast<ConstantSDNode>(IntID.getNode());
// Right now, we only support 'add' as an update.
assert(CID->getZExtValue() == Intrinsic::experimental_vector_histogram_add &&
"Unexpected histogram update operation");
EVT IndexVT = Index.getValueType();
LLVMContext &Ctx = *DAG.getContext();
ElementCount EC = IndexVT.getVectorElementCount();
EVT MemVT = EVT::getVectorVT(Ctx, HG->getMemoryVT(), EC);
EVT IncExtVT =
EVT::getIntegerVT(Ctx, AArch64::SVEBitsPerBlock / EC.getKnownMinValue());
EVT IncSplatVT = EVT::getVectorVT(Ctx, IncExtVT, EC);
bool ExtTrunc = IncSplatVT != MemVT;
SDValue Zero = DAG.getConstant(0, DL, MVT::i64);
SDValue PassThru = DAG.getSplatVector(IncSplatVT, DL, Zero);
SDValue IncSplat = DAG.getSplatVector(
IncSplatVT, DL, DAG.getAnyExtOrTrunc(Inc, DL, IncExtVT));
SDValue Ops[] = {Chain, PassThru, Mask, Ptr, Index, Scale};
MachineMemOperand *MMO = HG->getMemOperand();
// Create an MMO for the gather, without load|store flags.
MachineMemOperand *GMMO = DAG.getMachineFunction().getMachineMemOperand(
MMO->getPointerInfo(), MachineMemOperand::MOLoad, MMO->getSize(),
MMO->getAlign(), MMO->getAAInfo());
ISD::MemIndexType IndexType = HG->getIndexType();
SDValue Gather = DAG.getMaskedGather(
DAG.getVTList(IncSplatVT, MVT::Other), MemVT, DL, Ops, GMMO, IndexType,
ExtTrunc ? ISD::EXTLOAD : ISD::NON_EXTLOAD);
SDValue GChain = Gather.getValue(1);
// Perform the histcnt, multiply by inc, add to bucket data.
SDValue ID =
DAG.getTargetConstant(Intrinsic::aarch64_sve_histcnt, DL, IncExtVT);
SDValue HistCnt =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, IndexVT, ID, Mask, Index, Index);
SDValue Mul = DAG.getNode(ISD::MUL, DL, IncSplatVT, HistCnt, IncSplat);
SDValue Add = DAG.getNode(ISD::ADD, DL, IncSplatVT, Gather, Mul);
// Create an MMO for the scatter, without load|store flags.
MachineMemOperand *SMMO = DAG.getMachineFunction().getMachineMemOperand(
MMO->getPointerInfo(), MachineMemOperand::MOStore, MMO->getSize(),
MMO->getAlign(), MMO->getAAInfo());
SDValue ScatterOps[] = {GChain, Add, Mask, Ptr, Index, Scale};
SDValue Scatter = DAG.getMaskedScatter(DAG.getVTList(MVT::Other), MemVT, DL,
ScatterOps, SMMO, IndexType, ExtTrunc);
return Scatter;
}
/// If a PARTIAL_REDUCE_MLA node comes in with an accumulator-input type pairing
/// of (nx)v2i64/(nx)v16i8, we cannot directly lower it to a (u|s)dot. We can
/// however still make use of the dot product instruction by instead
/// accumulating over two steps: (nx)v16i8 -> (nx)v4i32 -> (nx)v2i64.
/// If available, make use of the (U|S)ADDW(B|T) instructions, otherwise
/// the following pattern is emitted:
/// add(add(Acc, ext(EXTRACT_SUBVECTOR(N, 0)), ext(EXTRACT_SUBVECTOR(N,
/// NTy/2))))
SDValue
AArch64TargetLowering::LowerPARTIAL_REDUCE_MLA(SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
SDValue Acc = Op.getOperand(0);
SDValue LHS = Op.getOperand(1);
SDValue RHS = Op.getOperand(2);
EVT ResultVT = Op.getValueType();
EVT OrigResultVT = ResultVT;
EVT OpVT = LHS.getValueType();
bool ConvertToScalable =
ResultVT.isFixedLengthVector() &&
useSVEForFixedLengthVectorVT(ResultVT, /*OverrideNEON=*/true);
if (ConvertToScalable) {
ResultVT = getContainerForFixedLengthVector(DAG, ResultVT);
OpVT = getContainerForFixedLengthVector(DAG, LHS.getValueType());
Acc = convertToScalableVector(DAG, ResultVT, Acc);
LHS = convertToScalableVector(DAG, OpVT, LHS);
RHS = convertToScalableVector(DAG, OpVT, RHS);
Op = DAG.getNode(Op.getOpcode(), DL, ResultVT, {Acc, LHS, RHS});
}
// Two-way and four-way partial reductions are supported by patterns.
// We only need to handle the 8-way partial reduction.
if (ResultVT.getScalarType() != MVT::i64 || OpVT.getScalarType() != MVT::i8)
return ConvertToScalable ? convertFromScalableVector(DAG, OrigResultVT, Op)
: Op;
EVT DotVT = ResultVT.isScalableVector() ? MVT::nxv4i32 : MVT::v4i32;
SDValue DotNode = DAG.getNode(Op.getOpcode(), DL, DotVT,
DAG.getConstant(0, DL, DotVT), LHS, RHS);
SDValue Res;
bool IsUnsigned = Op.getOpcode() == ISD::PARTIAL_REDUCE_UMLA;
if (Subtarget->hasSVE2() || Subtarget->isStreamingSVEAvailable()) {
unsigned LoOpcode = IsUnsigned ? AArch64ISD::UADDWB : AArch64ISD::SADDWB;
unsigned HiOpcode = IsUnsigned ? AArch64ISD::UADDWT : AArch64ISD::SADDWT;
SDValue Lo = DAG.getNode(LoOpcode, DL, ResultVT, Acc, DotNode);
Res = DAG.getNode(HiOpcode, DL, ResultVT, Lo, DotNode);
} else {
// Fold (nx)v4i32 into (nx)v2i64
auto [DotNodeLo, DotNodeHi] = DAG.SplitVector(DotNode, DL);
if (IsUnsigned) {
DotNodeLo = DAG.getZExtOrTrunc(DotNodeLo, DL, ResultVT);
DotNodeHi = DAG.getZExtOrTrunc(DotNodeHi, DL, ResultVT);
} else {
DotNodeLo = DAG.getSExtOrTrunc(DotNodeLo, DL, ResultVT);
DotNodeHi = DAG.getSExtOrTrunc(DotNodeHi, DL, ResultVT);
}
auto Lo = DAG.getNode(ISD::ADD, DL, ResultVT, Acc, DotNodeLo);
Res = DAG.getNode(ISD::ADD, DL, ResultVT, Lo, DotNodeHi);
}
return ConvertToScalable ? convertFromScalableVector(DAG, OrigResultVT, Res)
: Res;
}
SDValue
AArch64TargetLowering::LowerGET_ACTIVE_LANE_MASK(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
assert(Subtarget->isSVEorStreamingSVEAvailable() &&
"Lowering fixed length get_active_lane_mask requires SVE!");
// There are no dedicated fixed-length instructions for GET_ACTIVE_LANE_MASK,
// but we can use SVE when available.
SDLoc DL(Op);
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
EVT WhileVT = ContainerVT.changeElementType(MVT::i1);
SDValue Mask = DAG.getNode(ISD::GET_ACTIVE_LANE_MASK, DL, WhileVT,
Op.getOperand(0), Op.getOperand(1));
SDValue MaskAsInt = DAG.getNode(ISD::SIGN_EXTEND, DL, ContainerVT, Mask);
return DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, VT, MaskAsInt,
DAG.getVectorIdxConstant(0, DL));
}
SDValue
AArch64TargetLowering::LowerFixedLengthFPToIntToSVE(SDValue Op,
SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
bool IsSigned = Op.getOpcode() == ISD::FP_TO_SINT;
unsigned Opcode = IsSigned ? AArch64ISD::FCVTZS_MERGE_PASSTHRU
: AArch64ISD::FCVTZU_MERGE_PASSTHRU;
SDLoc DL(Op);
SDValue Val = Op.getOperand(0);
EVT SrcVT = Val.getValueType();
EVT ContainerDstVT = getContainerForFixedLengthVector(DAG, VT);
EVT ContainerSrcVT = getContainerForFixedLengthVector(DAG, SrcVT);
if (VT.bitsGT(SrcVT)) {
EVT CvtVT = ContainerDstVT.changeVectorElementType(
ContainerSrcVT.getVectorElementType());
SDValue Pg = getPredicateForFixedLengthVector(DAG, DL, VT);
Val = DAG.getNode(ISD::BITCAST, DL, SrcVT.changeTypeToInteger(), Val);
Val = DAG.getNode(ISD::ANY_EXTEND, DL, VT, Val);
Val = convertToScalableVector(DAG, ContainerDstVT, Val);
Val = getSVESafeBitCast(CvtVT, Val, DAG);
Val = DAG.getNode(Opcode, DL, ContainerDstVT, Pg, Val,
DAG.getUNDEF(ContainerDstVT));
return convertFromScalableVector(DAG, VT, Val);
} else {
EVT CvtVT = ContainerSrcVT.changeTypeToInteger();
SDValue Pg = getPredicateForFixedLengthVector(DAG, DL, SrcVT);
// Safe to use a larger than specified result since an fp_to_int where the
// result doesn't fit into the destination is undefined.
Val = convertToScalableVector(DAG, ContainerSrcVT, Val);
Val = DAG.getNode(Opcode, DL, CvtVT, Pg, Val, DAG.getUNDEF(CvtVT));
Val = convertFromScalableVector(DAG, SrcVT.changeTypeToInteger(), Val);
return DAG.getNode(ISD::TRUNCATE, DL, VT, Val);
}
}
static SDValue GenerateFixedLengthSVETBL(SDValue Op, SDValue Op1, SDValue Op2,
ArrayRef<int> ShuffleMask, EVT VT,
EVT ContainerVT, SelectionDAG &DAG) {
auto &Subtarget = DAG.getSubtarget<AArch64Subtarget>();
SDLoc DL(Op);
unsigned MinSVESize = Subtarget.getMinSVEVectorSizeInBits();
unsigned MaxSVESize = Subtarget.getMaxSVEVectorSizeInBits();
bool IsSingleOp =
ShuffleVectorInst::isSingleSourceMask(ShuffleMask, ShuffleMask.size());
if (!Subtarget.isNeonAvailable() && !MinSVESize)
MinSVESize = 128;
// Ignore two operands if no SVE2 or all index numbers couldn't
// be represented.
if (!IsSingleOp && !Subtarget.hasSVE2())
return SDValue();
EVT VTOp1 = Op.getOperand(0).getValueType();
unsigned BitsPerElt = VTOp1.getVectorElementType().getSizeInBits();
unsigned IndexLen = MinSVESize / BitsPerElt;
unsigned ElementsPerVectorReg = VTOp1.getVectorNumElements();
uint64_t MaxOffset = maxUIntN(BitsPerElt);
EVT MaskEltType = VTOp1.getVectorElementType().changeTypeToInteger();
EVT MaskType = EVT::getVectorVT(*DAG.getContext(), MaskEltType, IndexLen);
bool MinMaxEqual = (MinSVESize == MaxSVESize);
assert(ElementsPerVectorReg <= IndexLen && ShuffleMask.size() <= IndexLen &&
"Incorrectly legalised shuffle operation");
SmallVector<SDValue, 8> TBLMask;
// If MinSVESize is not equal to MaxSVESize then we need to know which
// TBL mask element needs adjustment.
SmallVector<SDValue, 8> AddRuntimeVLMask;
// Bail out for 8-bits element types, because with 2048-bit SVE register
// size 8 bits is only sufficient to index into the first source vector.
if (!IsSingleOp && !MinMaxEqual && BitsPerElt == 8)
return SDValue();
for (int Index : ShuffleMask) {
// Handling poison index value.
if (Index < 0)
Index = 0;
// If the mask refers to elements in the second operand, then we have to
// offset the index by the number of elements in a vector. If this is number
// is not known at compile-time, we need to maintain a mask with 'VL' values
// to add at runtime.
if ((unsigned)Index >= ElementsPerVectorReg) {
if (MinMaxEqual) {
Index += IndexLen - ElementsPerVectorReg;
} else {
Index = Index - ElementsPerVectorReg;
AddRuntimeVLMask.push_back(DAG.getConstant(1, DL, MVT::i64));
}
} else if (!MinMaxEqual)
AddRuntimeVLMask.push_back(DAG.getConstant(0, DL, MVT::i64));
// For 8-bit elements and 1024-bit SVE registers and MaxOffset equals
// to 255, this might point to the last element of in the second operand
// of the shufflevector, thus we are rejecting this transform.
if ((unsigned)Index >= MaxOffset)
return SDValue();
TBLMask.push_back(DAG.getConstant(Index, DL, MVT::i64));
}
// Choosing an out-of-range index leads to the lane being zeroed vs zero
// value where it would perform first lane duplication for out of
// index elements. For i8 elements an out-of-range index could be a valid
// for 2048-bit vector register size.
for (unsigned i = 0; i < IndexLen - ElementsPerVectorReg; ++i) {
TBLMask.push_back(DAG.getConstant((int)MaxOffset, DL, MVT::i64));
if (!MinMaxEqual)
AddRuntimeVLMask.push_back(DAG.getConstant(0, DL, MVT::i64));
}
EVT MaskContainerVT = getContainerForFixedLengthVector(DAG, MaskType);
SDValue VecMask =
DAG.getBuildVector(MaskType, DL, ArrayRef(TBLMask.data(), IndexLen));
SDValue SVEMask = convertToScalableVector(DAG, MaskContainerVT, VecMask);
SDValue Shuffle;
if (IsSingleOp)
Shuffle =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, ContainerVT,
DAG.getConstant(Intrinsic::aarch64_sve_tbl, DL, MVT::i32),
Op1, SVEMask);
else if (Subtarget.hasSVE2()) {
if (!MinMaxEqual) {
unsigned MinNumElts = AArch64::SVEBitsPerBlock / BitsPerElt;
SDValue VScale = (BitsPerElt == 64)
? DAG.getVScale(DL, MVT::i64, APInt(64, MinNumElts))
: DAG.getVScale(DL, MVT::i32, APInt(32, MinNumElts));
SDValue VecMask =
DAG.getBuildVector(MaskType, DL, ArrayRef(TBLMask.data(), IndexLen));
SDValue MulByMask = DAG.getNode(
ISD::MUL, DL, MaskType,
DAG.getNode(ISD::SPLAT_VECTOR, DL, MaskType, VScale),
DAG.getBuildVector(MaskType, DL,
ArrayRef(AddRuntimeVLMask.data(), IndexLen)));
SDValue UpdatedVecMask =
DAG.getNode(ISD::ADD, DL, MaskType, VecMask, MulByMask);
SVEMask = convertToScalableVector(
DAG, getContainerForFixedLengthVector(DAG, MaskType), UpdatedVecMask);
}
Shuffle =
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, ContainerVT,
DAG.getConstant(Intrinsic::aarch64_sve_tbl2, DL, MVT::i32),
Op1, Op2, SVEMask);
}
Shuffle = convertFromScalableVector(DAG, VT, Shuffle);
return DAG.getNode(ISD::BITCAST, DL, Op.getValueType(), Shuffle);
}
SDValue AArch64TargetLowering::LowerFixedLengthVECTOR_SHUFFLEToSVE(
SDValue Op, SelectionDAG &DAG) const {
EVT VT = Op.getValueType();
assert(VT.isFixedLengthVector() && "Expected fixed length vector type!");
auto *SVN = cast<ShuffleVectorSDNode>(Op.getNode());
auto ShuffleMask = SVN->getMask();
SDLoc DL(Op);
SDValue Op1 = Op.getOperand(0);
SDValue Op2 = Op.getOperand(1);
EVT ContainerVT = getContainerForFixedLengthVector(DAG, VT);
Op1 = convertToScalableVector(DAG, ContainerVT, Op1);
Op2 = convertToScalableVector(DAG, ContainerVT, Op2);
auto MinLegalExtractEltScalarTy = [](EVT ScalarTy) -> EVT {
if (ScalarTy == MVT::i8 || ScalarTy == MVT::i16)
return MVT::i32;
return ScalarTy;
};
if (SVN->isSplat()) {
unsigned Lane = std::max(0, SVN->getSplatIndex());
EVT ScalarTy = MinLegalExtractEltScalarTy(VT.getVectorElementType());
SDValue SplatEl = DAG.getNode(ISD::EXTRACT_VECTOR_ELT, DL, ScalarTy, Op1,
DAG.getConstant(Lane, DL, MVT::i64));
Op = DAG.getNode(ISD::SPLAT_VECTOR, DL, ContainerVT, SplatEl);
return convertFromScalableVector(DAG, VT, Op);
}
bool ReverseEXT = false;
unsigned Imm;
if (isEXTMask(ShuffleMask, VT, ReverseEXT, Imm) &&
Imm == VT.getVectorNumElements() - 1) {
if (ReverseEXT)
std::swap(Op1, Op2);
EVT ScalarTy = MinLegalExtractEltScalarTy(VT.getVectorElementType());
SDValue Scalar = DAG.getNode(
ISD::EXTRACT_VECTOR_ELT, DL, ScalarTy, Op1,
DAG.getConstant(VT.getVectorNumElements() - 1, DL, MVT::i64));
Op = DAG.getNode(AArch64ISD::INSR, DL, ContainerVT, Op2, Scalar);
return convertFromScalableVector(DAG, VT, Op);
}
unsigned EltSize = VT.getScalarSizeInBits();
for (unsigned BlockSize : {64U, 32U, 16U}) {
if (isREVMask(ShuffleMask, EltSize, VT.getVectorNumElements(), BlockSize)) {
unsigned RevOp;
if (EltSize == 8)
RevOp = AArch64ISD::BSWAP_MERGE_PASSTHRU;
else if (EltSize == 16)
RevOp = AArch64ISD::REVH_MERGE_PASSTHRU;
else
RevOp = AArch64ISD::REVW_MERGE_PASSTHRU;
EVT BlockedVT =
getPackedSVEVectorVT(EVT::getIntegerVT(*DAG.getContext(), BlockSize));
SDValue Pg = getPredicateForVector(DAG, DL, BlockedVT);
SDValue BlockedOp1 = DAG.getNode(ISD::BITCAST, DL, BlockedVT, Op1);
SDValue BlockedRev = DAG.getNode(RevOp, DL, BlockedVT, Pg, BlockedOp1,
DAG.getUNDEF(BlockedVT));
SDValue Container =
DAG.getNode(ISD::BITCAST, DL, ContainerVT, BlockedRev);
return convertFromScalableVector(DAG, VT, Container);
}
}
if (Subtarget->hasSVE2p1() && EltSize == 64 &&
isREVMask(ShuffleMask, EltSize, VT.getVectorNumElements(), 128)) {
SDValue Pg = getPredicateForVector(DAG, DL, VT);
SDValue Revd = DAG.getNode(AArch64ISD::REVD_MERGE_PASSTHRU, DL, ContainerVT,
Pg, Op1, DAG.getUNDEF(ContainerVT));
return convertFromScalableVector(DAG, VT, Revd);
}
unsigned WhichResult;
if (isZIPMask(ShuffleMask, VT.getVectorNumElements(), WhichResult) &&
WhichResult == 0)
return convertFromScalableVector(
DAG, VT, DAG.getNode(AArch64ISD::ZIP1, DL, ContainerVT, Op1, Op2));
if (isTRNMask(ShuffleMask, VT.getVectorNumElements(), WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::TRN1 : AArch64ISD::TRN2;
return convertFromScalableVector(
DAG, VT, DAG.getNode(Opc, DL, ContainerVT, Op1, Op2));
}
if (isZIP_v_undef_Mask(ShuffleMask, VT, WhichResult) && WhichResult == 0)
return convertFromScalableVector(
DAG, VT, DAG.getNode(AArch64ISD::ZIP1, DL, ContainerVT, Op1, Op1));
if (isTRN_v_undef_Mask(ShuffleMask, VT, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::TRN1 : AArch64ISD::TRN2;
return convertFromScalableVector(
DAG, VT, DAG.getNode(Opc, DL, ContainerVT, Op1, Op1));
}
// Functions like isZIPMask return true when a ISD::VECTOR_SHUFFLE's mask
// represents the same logical operation as performed by a ZIP instruction. In
// isolation these functions do not mean the ISD::VECTOR_SHUFFLE is exactly
// equivalent to an AArch64 instruction. There's the extra component of
// ISD::VECTOR_SHUFFLE's value type to consider. Prior to SVE these functions
// only operated on 64/128bit vector types that have a direct mapping to a
// target register and so an exact mapping is implied.
// However, when using SVE for fixed length vectors, most legal vector types
// are actually sub-vectors of a larger SVE register. When mapping
// ISD::VECTOR_SHUFFLE to an SVE instruction care must be taken to consider
// how the mask's indices translate. Specifically, when the mapping requires
// an exact meaning for a specific vector index (e.g. Index X is the last
// vector element in the register) then such mappings are often only safe when
// the exact SVE register size is know. The main exception to this is when
// indices are logically relative to the first element of either
// ISD::VECTOR_SHUFFLE operand because these relative indices don't change
// when converting from fixed-length to scalable vector types (i.e. the start
// of a fixed length vector is always the start of a scalable vector).
unsigned MinSVESize = Subtarget->getMinSVEVectorSizeInBits();
unsigned MaxSVESize = Subtarget->getMaxSVEVectorSizeInBits();
if (MinSVESize == MaxSVESize && MaxSVESize == VT.getSizeInBits()) {
if (ShuffleVectorInst::isReverseMask(ShuffleMask, ShuffleMask.size()) &&
Op2.isUndef()) {
Op = DAG.getNode(ISD::VECTOR_REVERSE, DL, ContainerVT, Op1);
return convertFromScalableVector(DAG, VT, Op);
}
if (isZIPMask(ShuffleMask, VT.getVectorNumElements(), WhichResult) &&
WhichResult != 0)
return convertFromScalableVector(
DAG, VT, DAG.getNode(AArch64ISD::ZIP2, DL, ContainerVT, Op1, Op2));
if (isUZPMask(ShuffleMask, VT.getVectorNumElements(), WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::UZP1 : AArch64ISD::UZP2;
return convertFromScalableVector(
DAG, VT, DAG.getNode(Opc, DL, ContainerVT, Op1, Op2));
}
if (isZIP_v_undef_Mask(ShuffleMask, VT, WhichResult) && WhichResult != 0)
return convertFromScalableVector(
DAG, VT, DAG.getNode(AArch64ISD::ZIP2, DL, ContainerVT, Op1, Op1));
if (isUZP_v_undef_Mask(ShuffleMask, VT, WhichResult)) {
unsigned Opc = (WhichResult == 0) ? AArch64ISD::UZP1 : AArch64ISD::UZP2;
return convertFromScalableVector(
DAG, VT, DAG.getNode(Opc, DL, ContainerVT, Op1, Op1));
}
if ((Subtarget->hasSVE2p1() || Subtarget->hasSME2p1()) &&
Subtarget->isSVEorStreamingSVEAvailable()) {
assert(VT.getFixedSizeInBits() % AArch64::SVEBitsPerBlock == 0 &&
"Unsupported SVE vector size");
unsigned Segments = VT.getFixedSizeInBits() / AArch64::SVEBitsPerBlock;
unsigned SegmentElts = VT.getVectorNumElements() / Segments;
if (std::optional<unsigned> Lane =
isDUPQMask(ShuffleMask, Segments, SegmentElts)) {
SDValue IID =
DAG.getConstant(Intrinsic::aarch64_sve_dup_laneq, DL, MVT::i64);
return convertFromScalableVector(
DAG, VT,
DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, ContainerVT,
{IID, Op1,
DAG.getConstant(*Lane, DL, MVT::i64,
/*isTarget=*/true)}));
}
}
}
// Try to widen the shuffle before generating a possibly expensive SVE TBL.
// This may allow the shuffle to be matched as something cheaper like ZIP1.
if (SDValue WideOp = tryWidenMaskForShuffle(Op, DAG))
return WideOp;
// Avoid producing TBL instruction if we don't know SVE register minimal size,
// unless NEON is not available and we can assume minimal SVE register size is
// 128-bits.
if (MinSVESize || !Subtarget->isNeonAvailable())
return GenerateFixedLengthSVETBL(Op, Op1, Op2, ShuffleMask, VT, ContainerVT,
DAG);
return SDValue();
}
SDValue AArch64TargetLowering::getSVESafeBitCast(EVT VT, SDValue Op,
SelectionDAG &DAG) const {
SDLoc DL(Op);
EVT InVT = Op.getValueType();
assert(VT.isScalableVector() && isTypeLegal(VT) &&
InVT.isScalableVector() && isTypeLegal(InVT) &&
"Only expect to cast between legal scalable vector types!");
assert(VT.getVectorElementType() != MVT::i1 &&
InVT.getVectorElementType() != MVT::i1 &&
"For predicate bitcasts, use getSVEPredicateBitCast");
if (InVT == VT)
return Op;
EVT PackedVT = getPackedSVEVectorVT(VT.getVectorElementType());
EVT PackedInVT = getPackedSVEVectorVT(InVT.getVectorElementType());
// Safe bitcasting between unpacked vector types of different element counts
// is currently unsupported because the following is missing the necessary
// work to ensure the result's elements live where they're supposed to within
// an SVE register.
// 01234567
// e.g. nxv2i32 = XX??XX??
// nxv4f16 = X?X?X?X?
assert((VT.getVectorElementCount() == InVT.getVectorElementCount() ||
VT == PackedVT || InVT == PackedInVT) &&
"Unexpected bitcast!");
// Pack input if required.
if (InVT != PackedInVT)
Op = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, PackedInVT, Op);
if (Subtarget->isLittleEndian() ||
PackedVT.getScalarSizeInBits() == PackedInVT.getScalarSizeInBits())
Op = DAG.getNode(ISD::BITCAST, DL, PackedVT, Op);
else {
EVT PackedVTAsInt = PackedVT.changeTypeToInteger();
EVT PackedInVTAsInt = PackedInVT.changeTypeToInteger();
// Simulate the effect of casting through memory.
Op = DAG.getNode(ISD::BITCAST, DL, PackedInVTAsInt, Op);
if (PackedInVTAsInt.getScalarSizeInBits() != 8)
Op = DAG.getNode(ISD::BSWAP, DL, PackedInVTAsInt, Op);
Op = DAG.getNode(AArch64ISD::NVCAST, DL, PackedVTAsInt, Op);
if (PackedVTAsInt.getScalarSizeInBits() != 8)
Op = DAG.getNode(ISD::BSWAP, DL, PackedVTAsInt, Op);
Op = DAG.getNode(ISD::BITCAST, DL, PackedVT, Op);
}
// Unpack result if required.
if (VT != PackedVT)
Op = DAG.getNode(AArch64ISD::REINTERPRET_CAST, DL, VT, Op);
return Op;
}
bool AArch64TargetLowering::isAllActivePredicate(SelectionDAG &DAG,
SDValue N) const {
return ::isAllActivePredicate(DAG, N);
}
EVT AArch64TargetLowering::getPromotedVTForPredicate(EVT VT) const {
return ::getPromotedVTForPredicate(VT);
}
bool AArch64TargetLowering::SimplifyDemandedBitsForTargetNode(
SDValue Op, const APInt &OriginalDemandedBits,
const APInt &OriginalDemandedElts, KnownBits &Known, TargetLoweringOpt &TLO,
unsigned Depth) const {
unsigned Opc = Op.getOpcode();
switch (Opc) {
case AArch64ISD::VSHL: {
// Match (VSHL (VLSHR Val X) X)
SDValue ShiftL = Op;
SDValue ShiftR = Op->getOperand(0);
if (ShiftR->getOpcode() != AArch64ISD::VLSHR)
return false;
if (!ShiftL.hasOneUse() || !ShiftR.hasOneUse())
return false;
unsigned ShiftLBits = ShiftL->getConstantOperandVal(1);
unsigned ShiftRBits = ShiftR->getConstantOperandVal(1);
// Other cases can be handled as well, but this is not
// implemented.
if (ShiftRBits != ShiftLBits)
return false;
unsigned ScalarSize = Op.getScalarValueSizeInBits();
assert(ScalarSize > ShiftLBits && "Invalid shift imm");
APInt ZeroBits = APInt::getLowBitsSet(ScalarSize, ShiftLBits);
APInt UnusedBits = ~OriginalDemandedBits;
if ((ZeroBits & UnusedBits) != ZeroBits)
return false;
// All bits that are zeroed by (VSHL (VLSHR Val X) X) are not
// used - simplify to just Val.
return TLO.CombineTo(Op, ShiftR->getOperand(0));
}
case AArch64ISD::BICi: {
// Fold BICi if all destination bits already known to be zeroed
SDValue Op0 = Op.getOperand(0);
KnownBits KnownOp0 =
TLO.DAG.computeKnownBits(Op0, OriginalDemandedElts, Depth + 1);
// Op0 &= ~(ConstantOperandVal(1) << ConstantOperandVal(2))
APInt BitsToClear =
(Op->getConstantOperandAPInt(1) << Op->getConstantOperandAPInt(2))
.trunc(KnownOp0.getBitWidth());
APInt AlreadyZeroedBitsToClear = BitsToClear & KnownOp0.Zero;
if (BitsToClear.isSubsetOf(AlreadyZeroedBitsToClear))
return TLO.CombineTo(Op, Op0);
Known = KnownOp0 & KnownBits::makeConstant(~BitsToClear);
return false;
}
case ISD::INTRINSIC_WO_CHAIN: {
if (auto ElementSize = IsSVECntIntrinsic(Op)) {
unsigned MaxSVEVectorSizeInBits = Subtarget->getMaxSVEVectorSizeInBits();
if (!MaxSVEVectorSizeInBits)
MaxSVEVectorSizeInBits = AArch64::SVEMaxBitsPerVector;
unsigned MaxElements = MaxSVEVectorSizeInBits / *ElementSize;
// The SVE count intrinsics don't support the multiplier immediate so we
// don't have to account for that here. The value returned may be slightly
// over the true required bits, as this is based on the "ALL" pattern. The
// other patterns are also exposed by these intrinsics, but they all
// return a value that's strictly less than "ALL".
unsigned RequiredBits = llvm::bit_width(MaxElements);
unsigned BitWidth = Known.Zero.getBitWidth();
if (RequiredBits < BitWidth)
Known.Zero.setHighBits(BitWidth - RequiredBits);
return false;
}
}
}
return TargetLowering::SimplifyDemandedBitsForTargetNode(
Op, OriginalDemandedBits, OriginalDemandedElts, Known, TLO, Depth);
}
bool AArch64TargetLowering::isTargetCanonicalConstantNode(SDValue Op) const {
return Op.getOpcode() == AArch64ISD::DUP ||
Op.getOpcode() == AArch64ISD::MOVI ||
(Op.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
Op.getOperand(0).getOpcode() == AArch64ISD::DUP) ||
TargetLowering::isTargetCanonicalConstantNode(Op);
}
bool AArch64TargetLowering::isComplexDeinterleavingSupported() const {
return Subtarget->hasSVE() || Subtarget->hasSVE2() ||
Subtarget->hasComplxNum();
}
bool AArch64TargetLowering::isComplexDeinterleavingOperationSupported(
ComplexDeinterleavingOperation Operation, Type *Ty) const {
auto *VTy = dyn_cast<VectorType>(Ty);
if (!VTy)
return false;
// If the vector is scalable, SVE is enabled, implying support for complex
// numbers. Otherwise, we need to ensure complex number support is available
if (!VTy->isScalableTy() && !Subtarget->hasComplxNum())
return false;
auto *ScalarTy = VTy->getScalarType();
unsigned NumElements = VTy->getElementCount().getKnownMinValue();
// We can only process vectors that have a bit size of 128 or higher (with an
// additional 64 bits for Neon). Additionally, these vectors must have a
// power-of-2 size, as we later split them into the smallest supported size
// and merging them back together after applying complex operation.
unsigned VTyWidth = VTy->getScalarSizeInBits() * NumElements;
if ((VTyWidth < 128 && (VTy->isScalableTy() || VTyWidth != 64)) ||
!llvm::isPowerOf2_32(VTyWidth))
return false;
if (ScalarTy->isIntegerTy() && Subtarget->hasSVE2() && VTy->isScalableTy()) {
unsigned ScalarWidth = ScalarTy->getScalarSizeInBits();
if (Operation == ComplexDeinterleavingOperation::CDot)
return ScalarWidth == 32 || ScalarWidth == 64;
return 8 <= ScalarWidth && ScalarWidth <= 64;
}
// CDot is not supported outside of scalable/sve scopes
if (Operation == ComplexDeinterleavingOperation::CDot)
return false;
return (ScalarTy->isHalfTy() && Subtarget->hasFullFP16()) ||
ScalarTy->isFloatTy() || ScalarTy->isDoubleTy();
}
Value *AArch64TargetLowering::createComplexDeinterleavingIR(
IRBuilderBase &B, ComplexDeinterleavingOperation OperationType,
ComplexDeinterleavingRotation Rotation, Value *InputA, Value *InputB,
Value *Accumulator) const {
VectorType *Ty = cast<VectorType>(InputA->getType());
if (Accumulator == nullptr)
Accumulator = Constant::getNullValue(Ty);
bool IsScalable = Ty->isScalableTy();
bool IsInt = Ty->getElementType()->isIntegerTy();
unsigned TyWidth =
Ty->getScalarSizeInBits() * Ty->getElementCount().getKnownMinValue();
assert(((TyWidth >= 128 && llvm::isPowerOf2_32(TyWidth)) || TyWidth == 64) &&
"Vector type must be either 64 or a power of 2 that is at least 128");
if (TyWidth > 128) {
int Stride = Ty->getElementCount().getKnownMinValue() / 2;
int AccStride = cast<VectorType>(Accumulator->getType())
->getElementCount()
.getKnownMinValue() /
2;
auto *HalfTy = VectorType::getHalfElementsVectorType(Ty);
auto *LowerSplitA = B.CreateExtractVector(HalfTy, InputA, uint64_t(0));
auto *LowerSplitB = B.CreateExtractVector(HalfTy, InputB, uint64_t(0));
auto *UpperSplitA = B.CreateExtractVector(HalfTy, InputA, Stride);
auto *UpperSplitB = B.CreateExtractVector(HalfTy, InputB, Stride);
Value *LowerSplitAcc = nullptr;
Value *UpperSplitAcc = nullptr;
Type *FullTy = Ty;
FullTy = Accumulator->getType();
auto *HalfAccTy = VectorType::getHalfElementsVectorType(
cast<VectorType>(Accumulator->getType()));
LowerSplitAcc = B.CreateExtractVector(HalfAccTy, Accumulator, uint64_t(0));
UpperSplitAcc = B.CreateExtractVector(HalfAccTy, Accumulator, AccStride);
auto *LowerSplitInt = createComplexDeinterleavingIR(
B, OperationType, Rotation, LowerSplitA, LowerSplitB, LowerSplitAcc);
auto *UpperSplitInt = createComplexDeinterleavingIR(
B, OperationType, Rotation, UpperSplitA, UpperSplitB, UpperSplitAcc);
auto *Result = B.CreateInsertVector(FullTy, PoisonValue::get(FullTy),
LowerSplitInt, uint64_t(0));
return B.CreateInsertVector(FullTy, Result, UpperSplitInt, AccStride);
}
if (OperationType == ComplexDeinterleavingOperation::CMulPartial) {
if (IsScalable) {
if (IsInt)
return B.CreateIntrinsic(
Intrinsic::aarch64_sve_cmla_x, Ty,
{Accumulator, InputA, InputB, B.getInt32((int)Rotation * 90)});
auto *Mask = B.getAllOnesMask(Ty->getElementCount());
return B.CreateIntrinsic(
Intrinsic::aarch64_sve_fcmla, Ty,
{Mask, Accumulator, InputA, InputB, B.getInt32((int)Rotation * 90)});
}
Intrinsic::ID IdMap[4] = {Intrinsic::aarch64_neon_vcmla_rot0,
Intrinsic::aarch64_neon_vcmla_rot90,
Intrinsic::aarch64_neon_vcmla_rot180,
Intrinsic::aarch64_neon_vcmla_rot270};
return B.CreateIntrinsic(IdMap[(int)Rotation], Ty,
{Accumulator, InputA, InputB});
}
if (OperationType == ComplexDeinterleavingOperation::CAdd) {
if (IsScalable) {
if (Rotation == ComplexDeinterleavingRotation::Rotation_90 ||
Rotation == ComplexDeinterleavingRotation::Rotation_270) {
if (IsInt)
return B.CreateIntrinsic(
Intrinsic::aarch64_sve_cadd_x, Ty,
{InputA, InputB, B.getInt32((int)Rotation * 90)});
auto *Mask = B.getAllOnesMask(Ty->getElementCount());
return B.CreateIntrinsic(
Intrinsic::aarch64_sve_fcadd, Ty,
{Mask, InputA, InputB, B.getInt32((int)Rotation * 90)});
}
return nullptr;
}
Intrinsic::ID IntId = Intrinsic::not_intrinsic;
if (Rotation == ComplexDeinterleavingRotation::Rotation_90)
IntId = Intrinsic::aarch64_neon_vcadd_rot90;
else if (Rotation == ComplexDeinterleavingRotation::Rotation_270)
IntId = Intrinsic::aarch64_neon_vcadd_rot270;
if (IntId == Intrinsic::not_intrinsic)
return nullptr;
return B.CreateIntrinsic(IntId, Ty, {InputA, InputB});
}
if (OperationType == ComplexDeinterleavingOperation::CDot && IsInt &&
IsScalable) {
return B.CreateIntrinsic(
Intrinsic::aarch64_sve_cdot, Accumulator->getType(),
{Accumulator, InputA, InputB, B.getInt32((int)Rotation * 90)});
}
return nullptr;
}
bool AArch64TargetLowering::preferScalarizeSplat(SDNode *N) const {
unsigned Opc = N->getOpcode();
if (ISD::isExtOpcode(Opc)) {
if (any_of(N->users(),
[&](SDNode *Use) { return Use->getOpcode() == ISD::MUL; }))
return false;
}
return true;
}
unsigned AArch64TargetLowering::getMinimumJumpTableEntries() const {
return Subtarget->getMinimumJumpTableEntries();
}
MVT AArch64TargetLowering::getRegisterTypeForCallingConv(LLVMContext &Context,
CallingConv::ID CC,
EVT VT) const {
bool NonUnitFixedLengthVector =
VT.isFixedLengthVector() && !VT.getVectorElementCount().isScalar();
if (!NonUnitFixedLengthVector || !Subtarget->useSVEForFixedLengthVectors())
return TargetLowering::getRegisterTypeForCallingConv(Context, CC, VT);
EVT VT1;
MVT RegisterVT;
unsigned NumIntermediates;
getVectorTypeBreakdownForCallingConv(Context, CC, VT, VT1, NumIntermediates,
RegisterVT);
return RegisterVT;
}
unsigned AArch64TargetLowering::getNumRegistersForCallingConv(
LLVMContext &Context, CallingConv::ID CC, EVT VT) const {
bool NonUnitFixedLengthVector =
VT.isFixedLengthVector() && !VT.getVectorElementCount().isScalar();
if (!NonUnitFixedLengthVector || !Subtarget->useSVEForFixedLengthVectors())
return TargetLowering::getNumRegistersForCallingConv(Context, CC, VT);
EVT VT1;
MVT VT2;
unsigned NumIntermediates;
return getVectorTypeBreakdownForCallingConv(Context, CC, VT, VT1,
NumIntermediates, VT2);
}
unsigned AArch64TargetLowering::getVectorTypeBreakdownForCallingConv(
LLVMContext &Context, CallingConv::ID CC, EVT VT, EVT &IntermediateVT,
unsigned &NumIntermediates, MVT &RegisterVT) const {
int NumRegs = TargetLowering::getVectorTypeBreakdownForCallingConv(
Context, CC, VT, IntermediateVT, NumIntermediates, RegisterVT);
if (!RegisterVT.isFixedLengthVector() ||
RegisterVT.getFixedSizeInBits() <= 128)
return NumRegs;
assert(Subtarget->useSVEForFixedLengthVectors() && "Unexpected mode!");
assert(IntermediateVT == RegisterVT && "Unexpected VT mismatch!");
assert(RegisterVT.getFixedSizeInBits() % 128 == 0 && "Unexpected size!");
// A size mismatch here implies either type promotion or widening and would
// have resulted in scalarisation if larger vectors had not be available.
if (RegisterVT.getSizeInBits() * NumRegs != VT.getSizeInBits()) {
EVT EltTy = VT.getVectorElementType();
EVT NewVT = EVT::getVectorVT(Context, EltTy, ElementCount::getFixed(1));
if (!isTypeLegal(NewVT))
NewVT = EltTy;
IntermediateVT = NewVT;
NumIntermediates = VT.getVectorNumElements();
RegisterVT = getRegisterType(Context, NewVT);
return NumIntermediates;
}
// SVE VLS support does not introduce a new ABI so we should use NEON sized
// types for vector arguments and returns.
unsigned NumSubRegs = RegisterVT.getFixedSizeInBits() / 128;
NumIntermediates *= NumSubRegs;
NumRegs *= NumSubRegs;
switch (RegisterVT.getVectorElementType().SimpleTy) {
default:
llvm_unreachable("unexpected element type for vector");
case MVT::i8:
IntermediateVT = RegisterVT = MVT::v16i8;
break;
case MVT::i16:
IntermediateVT = RegisterVT = MVT::v8i16;
break;
case MVT::i32:
IntermediateVT = RegisterVT = MVT::v4i32;
break;
case MVT::i64:
IntermediateVT = RegisterVT = MVT::v2i64;
break;
case MVT::f16:
IntermediateVT = RegisterVT = MVT::v8f16;
break;
case MVT::f32:
IntermediateVT = RegisterVT = MVT::v4f32;
break;
case MVT::f64:
IntermediateVT = RegisterVT = MVT::v2f64;
break;
case MVT::bf16:
IntermediateVT = RegisterVT = MVT::v8bf16;
break;
}
return NumRegs;
}
bool AArch64TargetLowering::hasInlineStackProbe(
const MachineFunction &MF) const {
return !Subtarget->isTargetWindows() &&
MF.getInfo<AArch64FunctionInfo>()->hasStackProbing();
}
bool AArch64TargetLowering::isTypeDesirableForOp(unsigned Opc, EVT VT) const {
switch (Opc) {
case ISD::TRUNCATE_SSAT_S:
case ISD::TRUNCATE_SSAT_U:
case ISD::TRUNCATE_USAT_U:
if (VT == MVT::v8i8 || VT == MVT::v4i16 || VT == MVT::v2i32)
return true;
}
return TargetLowering::isTypeDesirableForOp(Opc, VT);
}
bool AArch64TargetLowering::shouldPreservePtrArith(const Function &F,
EVT VT) const {
return Subtarget->hasCPA() && UseFEATCPACodegen;
}