Target/AArch64/AArch64SelectionDAGInfo.cpp

//===-- AArch64SelectionDAGInfo.cpp - AArch64 SelectionDAG Info -----------===//
//
// 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 AArch64SelectionDAGInfo class.
//
//===----------------------------------------------------------------------===//

#include "AArch64TargetMachine.h"
using namespace llvm;

#define DEBUG_TYPE "aarch64-selectiondag-info"

SDValue AArch64SelectionDAGInfo::EmitTargetCodeForMemset(
    SelectionDAG &DAG, const SDLoc &dl, SDValue Chain, SDValue Dst, SDValue Src,
    SDValue Size, unsigned Align, bool isVolatile,
    MachinePointerInfo DstPtrInfo) const {
  // Check to see if there is a specialized entry-point for memory zeroing.
  ConstantSDNode *V = dyn_cast<ConstantSDNode>(Src);
  ConstantSDNode *SizeValue = dyn_cast<ConstantSDNode>(Size);
  const AArch64Subtarget &STI =
      DAG.getMachineFunction().getSubtarget<AArch64Subtarget>();
  const char *bzeroName = (V && V->isNullValue())
      ? DAG.getTargetLoweringInfo().getLibcallName(RTLIB::BZERO) : nullptr;
  // For small size (< 256), it is not beneficial to use bzero
  // instead of memset.
  if (bzeroName && (!SizeValue || SizeValue->getZExtValue() > 256)) {
    const AArch64TargetLowering &TLI = *STI.getTargetLowering();

    EVT IntPtr = TLI.getPointerTy(DAG.getDataLayout());
    Type *IntPtrTy = Type::getInt8PtrTy(*DAG.getContext());
    TargetLowering::ArgListTy Args;
    TargetLowering::ArgListEntry Entry;
    Entry.Node = Dst;
    Entry.Ty = IntPtrTy;
    Args.push_back(Entry);
    Entry.Node = Size;
    Args.push_back(Entry);
    TargetLowering::CallLoweringInfo CLI(DAG);
    CLI.setDebugLoc(dl)
        .setChain(Chain)
        .setLibCallee(CallingConv::C, Type::getVoidTy(*DAG.getContext()),
                      DAG.getExternalSymbol(bzeroName, IntPtr),
                      std::move(Args))
        .setDiscardResult();
    std::pair<SDValue, SDValue> CallResult = TLI.LowerCallTo(CLI);
    return CallResult.second;
  }
  return SDValue();
}
bool AArch64SelectionDAGInfo::generateFMAsInMachineCombiner(
    CodeGenOpt::Level OptLevel) const {
  return OptLevel >= CodeGenOpt::Aggressive;
}

static const int kSetTagLoopThreshold = 176;

static SDValue EmitUnrolledSetTag(SelectionDAG &DAG, const SDLoc &dl,
                                  SDValue Chain, SDValue Ptr, uint64_t ObjSize,
                                  const MachineMemOperand *BaseMemOperand,
                                  bool ZeroData) {
  MachineFunction &MF = DAG.getMachineFunction();
  unsigned ObjSizeScaled = ObjSize / 16;

  SDValue TagSrc = Ptr;
  if (Ptr.getOpcode() == ISD::FrameIndex) {
    int FI = cast<FrameIndexSDNode>(Ptr)->getIndex();
    Ptr = DAG.getTargetFrameIndex(FI, MVT::i64);
    // A frame index operand may end up as [SP + offset] => it is fine to use SP
    // register as the tag source.
    TagSrc = DAG.getRegister(AArch64::SP, MVT::i64);
  }

  const unsigned OpCode1 = ZeroData ? AArch64ISD::STZG : AArch64ISD::STG;
  const unsigned OpCode2 = ZeroData ? AArch64ISD::STZ2G : AArch64ISD::ST2G;

  SmallVector<SDValue, 8> OutChains;
  unsigned OffsetScaled = 0;
  while (OffsetScaled < ObjSizeScaled) {
    if (ObjSizeScaled - OffsetScaled >= 2) {
      SDValue AddrNode = DAG.getMemBasePlusOffset(Ptr, OffsetScaled * 16, dl);
      SDValue St = DAG.getMemIntrinsicNode(
          OpCode2, dl, DAG.getVTList(MVT::Other),
          {Chain, TagSrc, AddrNode},
          MVT::v4i64,
          MF.getMachineMemOperand(BaseMemOperand, OffsetScaled * 16, 16 * 2));
      OffsetScaled += 2;
      OutChains.push_back(St);
      continue;
    }

    if (ObjSizeScaled - OffsetScaled > 0) {
      SDValue AddrNode = DAG.getMemBasePlusOffset(Ptr, OffsetScaled * 16, dl);
      SDValue St = DAG.getMemIntrinsicNode(
          OpCode1, dl, DAG.getVTList(MVT::Other),
          {Chain, TagSrc, AddrNode},
          MVT::v2i64,
          MF.getMachineMemOperand(BaseMemOperand, OffsetScaled * 16, 16));
      OffsetScaled += 1;
      OutChains.push_back(St);
    }
  }

  SDValue Res = DAG.getNode(ISD::TokenFactor, dl, MVT::Other, OutChains);
  return Res;
}

SDValue AArch64SelectionDAGInfo::EmitTargetCodeForSetTag(
    SelectionDAG &DAG, const SDLoc &dl, SDValue Chain, SDValue Addr,
    SDValue Size, MachinePointerInfo DstPtrInfo, bool ZeroData) const {
  uint64_t ObjSize = cast<ConstantSDNode>(Size)->getZExtValue();
  assert(ObjSize % 16 == 0);

  MachineFunction &MF = DAG.getMachineFunction();
  MachineMemOperand *BaseMemOperand = MF.getMachineMemOperand(
      DstPtrInfo, MachineMemOperand::MOStore, ObjSize, 16);

  bool UseSetTagRangeLoop =
      kSetTagLoopThreshold >= 0 && (int)ObjSize >= kSetTagLoopThreshold;
  if (!UseSetTagRangeLoop)
    return EmitUnrolledSetTag(DAG, dl, Chain, Addr, ObjSize, BaseMemOperand,
                              ZeroData);

  if (ObjSize % 32 != 0) {
    SDNode *St1 = DAG.getMachineNode(
        ZeroData ? AArch64::STZGPostIndex : AArch64::STGPostIndex, dl,
        {MVT::i64, MVT::Other},
        {Addr, Addr, DAG.getTargetConstant(1, dl, MVT::i64), Chain});
    DAG.setNodeMemRefs(cast<MachineSDNode>(St1), {BaseMemOperand});
    ObjSize -= 16;
    Addr = SDValue(St1, 0);
    Chain = SDValue(St1, 1);
  }

  const EVT ResTys[] = {MVT::i64, MVT::i64, MVT::Other};
  SDValue Ops[] = {DAG.getConstant(ObjSize, dl, MVT::i64), Addr, Chain};
  SDNode *St = DAG.getMachineNode(
      ZeroData ? AArch64::STZGloop : AArch64::STGloop, dl, ResTys, Ops);

  DAG.setNodeMemRefs(cast<MachineSDNode>(St), {BaseMemOperand});
  return SDValue(St, 2);
}