setOperationAction(ISD::FP_TO_SINT, VT.getSimpleVT(), Custom);
setOperationAction(ISD::FP_TO_UINT, VT.getSimpleVT(), Custom);
- // [SU][MIN|MAX] and [SU]ABSDIFF are available for all NEON types apart from
- // i64.
+ // [SU][MIN|MAX] are available for all NEON types apart from i64.
if (!VT.isFloatingPoint() &&
VT.getSimpleVT() != MVT::v2i64 && VT.getSimpleVT() != MVT::v1i64)
- for (unsigned Opcode : {ISD::SMIN, ISD::SMAX, ISD::UMIN, ISD::UMAX,
- ISD::SABSDIFF, ISD::UABSDIFF})
+ for (unsigned Opcode : {ISD::SMIN, ISD::SMAX, ISD::UMIN, ISD::UMAX})
setOperationAction(Opcode, VT.getSimpleVT(), Legal);
// F[MIN|MAX][NUM|NAN] are available for all FP NEON types (not f16 though!).
// in the cost tables.
EVT InVT = Op.getOperand(0).getValueType();
EVT VT = Op.getValueType();
+ unsigned NumElts = InVT.getVectorNumElements();
+
+ // f16 vectors are promoted to f32 before a conversion.
+ if (InVT.getVectorElementType() == MVT::f16) {
+ MVT NewVT = MVT::getVectorVT(MVT::f32, NumElts);
+ SDLoc dl(Op);
+ return DAG.getNode(
+ Op.getOpcode(), dl, Op.getValueType(),
+ DAG.getNode(ISD::FP_EXTEND, dl, NewVT, Op.getOperand(0)));
+ }
if (VT.getSizeInBits() < InVT.getSizeInBits()) {
SDLoc dl(Op);
}
}
-/// getFunctionAlignment - Return the Log2 alignment of this function.
-unsigned AArch64TargetLowering::getFunctionAlignment(const Function *F) const {
- return 2;
-}
-
//===----------------------------------------------------------------------===//
// Calling Convention Implementation
//===----------------------------------------------------------------------===//
Flag = Chain.getValue(1);
RetOps.push_back(DAG.getRegister(VA.getLocReg(), VA.getLocVT()));
}
+ const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
+ const MCPhysReg *I =
+ TRI->getCalleeSavedRegsViaCopy(&DAG.getMachineFunction());
+ 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.
case Intrinsic::aarch64_neon_ld4r: {
Info.opc = ISD::INTRINSIC_W_CHAIN;
// Conservatively set memVT to the entire set of vectors loaded.
- uint64_t NumElts = DL.getTypeAllocSize(I.getType()) / 8;
+ uint64_t NumElts = DL.getTypeSizeInBits(I.getType()) / 64;
Info.memVT = EVT::getVectorVT(I.getType()->getContext(), MVT::i64, NumElts);
Info.ptrVal = I.getArgOperand(I.getNumArgOperands() - 1);
Info.offset = 0;
Type *ArgTy = I.getArgOperand(ArgI)->getType();
if (!ArgTy->isVectorTy())
break;
- NumElts += DL.getTypeAllocSize(ArgTy) / 8;
+ NumElts += DL.getTypeSizeInBits(ArgTy) / 64;
}
Info.memVT = EVT::getVectorVT(I.getType()->getContext(), MVT::i64, NumElts);
Info.ptrVal = I.getArgOperand(I.getNumArgOperands() - 1);
const DataLayout &DL = LI->getModule()->getDataLayout();
VectorType *VecTy = Shuffles[0]->getType();
- unsigned VecSize = DL.getTypeAllocSizeInBits(VecTy);
+ unsigned VecSize = DL.getTypeSizeInBits(VecTy);
// Skip if we do not have NEON and skip illegal vector types.
if (!Subtarget->hasNEON() || (VecSize != 64 && VecSize != 128))
VectorType *SubVecTy = VectorType::get(EltTy, NumSubElts);
const DataLayout &DL = SI->getModule()->getDataLayout();
- unsigned SubVecSize = DL.getTypeAllocSizeInBits(SubVecTy);
+ unsigned SubVecSize = DL.getTypeSizeInBits(SubVecTy);
// Skip if we do not have NEON and skip illegal vector types.
if (!Subtarget->hasNEON() || (SubVecSize != 64 && SubVecSize != 128))
// (aarch64_neon_umull (extract_high (v2i64 vec)))
// (extract_high (v2i64 (dup128 scalar)))))
//
-static SDValue tryCombineLongOpWithDup(SDNode *N,
+static SDValue tryCombineLongOpWithDup(unsigned IID, SDNode *N,
TargetLowering::DAGCombinerInfo &DCI,
SelectionDAG &DAG) {
if (DCI.isBeforeLegalizeOps())
return SDValue();
- bool IsIntrinsic = N->getOpcode() == ISD::INTRINSIC_WO_CHAIN;
- SDValue LHS = N->getOperand(IsIntrinsic ? 1 : 0);
- SDValue RHS = N->getOperand(IsIntrinsic ? 2 : 1);
+ SDValue LHS = N->getOperand(1);
+ SDValue RHS = N->getOperand(2);
assert(LHS.getValueType().is64BitVector() &&
RHS.getValueType().is64BitVector() &&
"unexpected shape for long operation");
return SDValue();
}
- // N could either be an intrinsic or a sabsdiff/uabsdiff node.
- if (IsIntrinsic)
- return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, SDLoc(N), N->getValueType(0),
- N->getOperand(0), LHS, RHS);
- else
- 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) {
case Intrinsic::aarch64_neon_fmin:
return DAG.getNode(ISD::FMINNAN, SDLoc(N), N->getValueType(0),
N->getOperand(1), N->getOperand(2));
- case Intrinsic::aarch64_neon_sabd:
- return DAG.getNode(ISD::SABSDIFF, SDLoc(N), N->getValueType(0),
- N->getOperand(1), N->getOperand(2));
- case Intrinsic::aarch64_neon_uabd:
- return DAG.getNode(ISD::UABSDIFF, 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_umull:
case Intrinsic::aarch64_neon_pmull:
case Intrinsic::aarch64_neon_sqdmull:
- return tryCombineLongOpWithDup(N, DCI, DAG);
+ return tryCombineLongOpWithDup(IID, N, DCI, DAG);
case Intrinsic::aarch64_neon_sqshl:
case Intrinsic::aarch64_neon_uqshl:
case Intrinsic::aarch64_neon_sqshlu:
// 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).getOpcode() == ISD::SABSDIFF ||
- N->getOperand(0).getOpcode() == ISD::UABSDIFF)) {
+ N->getOperand(0).getOpcode() == ISD::INTRINSIC_WO_CHAIN) {
SDNode *ABDNode = N->getOperand(0).getNode();
- SDValue NewABD = tryCombineLongOpWithDup(ABDNode, DCI, DAG);
- if (!NewABD.getNode())
- return SDValue();
+ unsigned IID = getIntrinsicID(ABDNode);
+ if (IID == Intrinsic::aarch64_neon_sabd ||
+ IID == Intrinsic::aarch64_neon_uabd) {
+ SDValue NewABD = tryCombineLongOpWithDup(IID, ABDNode, DCI, DAG);
+ if (!NewABD.getNode())
+ return SDValue();
- return DAG.getNode(ISD::ZERO_EXTEND, SDLoc(N), N->getValueType(0),
- NewABD);
+ return DAG.getNode(ISD::ZERO_EXTEND, SDLoc(N), N->getValueType(0),
+ NewABD);
+ }
}
// This is effectively a custom type legalization for AArch64.
IRB.CreateConstGEP1_32(IRB.CreateCall(ThreadPointerFunc), TlsOffset),
Type::getInt8PtrTy(IRB.getContext())->getPointerTo(0));
}
+
+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();
+ 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!");
+
+ unsigned 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, Entry->begin(), DebugLoc(), TII->get(TargetOpcode::COPY),
+ NewVR)
+ .addReg(*I);
+
+ for (auto *Exit : Exits)
+ BuildMI(*Exit, Exit->begin(), DebugLoc(), TII->get(TargetOpcode::COPY),
+ *I)
+ .addReg(NewVR);
+ }
+}
projects / oota-llvm.git / blobdiff