[AArch64] Identify SAD pattern

[mivnay]

Summary

(ABS (SUB (ZERO_EXTEND a), (ZERO_EXTEND b))) to ZERO_EXTEND((UABD a, b))

This partially solves the bug: https://bugs.llvm.org/show_bug.cgi?id=46888
Meta ticket: https://bugs.llvm.org/show_bug.cgi?id=46929

[dmgreen]

So this converts v16i32 abs(sub(zext(v16i8), zext(v16i8)))) to v16i32 zext(abd(v16i8, v16i8))? And doesn't start with a trunk? That sounds like it should work OK.

Can you add other types and sign extends? And more tests. And a AArch64ISD node (they are not very complex to add :) )

[mivnay]

In D88742#2312075, @dmgreen wrote:

So this converts v16i32 abs(sub(zext(v16i8), zext(v16i8)))) to v16i32 zext(abd(v16i8, v16i8))? And doesn't start with a trunk? That sounds like it should work OK.

Yes.

Can you add other types and sign extends? And more tests. And a AArch64ISD node (they are not very complex to add :) )

Few patterns already exists in .td files for uabdl, etc. I tried to add the ISD node for UABD (posted it in https://reviews.llvm.org/D88897) but there are unintended consequences (ISel is generating different patterns). Can I take a look at it later (when other pending patches are done)? About the sign extend variants, I am not sure whether it is legal to do this transformation.

[dmgreen]

In D88742#2314350, @mivnay wrote:

In D88742#2312075, @dmgreen wrote:

So this converts v16i32 abs(sub(zext(v16i8), zext(v16i8)))) to v16i32 zext(abd(v16i8, v16i8))? And doesn't start with a trunk? That sounds like it should work OK.

Yes.

Can you add other types and sign extends? And more tests. And a AArch64ISD node (they are not very complex to add :) )

Few patterns already exists in .td files for uabdl, etc. I tried to add the ISD node for UABD (posted it in https://reviews.llvm.org/D88897) but there are unintended consequences (ISel is generating different patterns). Can I take a look at it later (when other pending patches are done)? About the sign extend variants, I am not sure whether it is legal to do this transformation.

What part of sadb are you worried about? I thought they could be treated the same, given you are extended from enough extra bits. But I may be mistaken, they can be somewhat difficult.

Can D88897 be folder back into here now?

[mivnay]

What part of sadb are you worried about? I thought they could be treated the same, given you are extended from enough extra bits. But I may be mistaken, they can be somewhat difficult.

I am worried about the semantics of sabd instruction especially about v16i32 (sub(sext(v16i8), sext(v16i8))) to v16i32 sext(sub(v16i8, v16i8)) part. Can this be looked at later once my other patches are through?

Can D88897 be folder back into here now?

There seems to be more issues for adding SABD ISD node. Since it is a separate patch now, can this one move forward?

[dmgreen]

In D88742#2314735, @mivnay wrote:

What part of sadb are you worried about? I thought they could be treated the same, given you are extended from enough extra bits. But I may be mistaken, they can be somewhat difficult.

I am worried about the semantics of sabd instruction especially about v16i32 (sub(sext(v16i8), sext(v16i8))) to v16i32 sext(sub(v16i8, v16i8)) part. Can this be looked at later once my other patches are through?

The return type of the abd is positive unsigned, so I think if turning it into v16i32 zext(sabd(v16i8, v16i8)) should be valid? But it's worth testing that, I may be mistaken..

[mivnay]

In D88742#2316222, @dmgreen wrote:

In D88742#2314735, @mivnay wrote:

What part of sadb are you worried about? I thought they could be treated the same, given you are extended from enough extra bits. But I may be mistaken, they can be somewhat difficult.

I am worried about the semantics of sabd instruction especially about v16i32 (sub(sext(v16i8), sext(v16i8))) to v16i32 sext(sub(v16i8, v16i8)) part. Can this be looked at later once my other patches are through?

The return type of the abd is positive unsigned, so I think if turning it into v16i32 zext(sabd(v16i8, v16i8)) should be valid? But it's worth testing that, I may be mistaken..

I am not sure about the validity either and need to test it. Can that be added as enhancement later and go ahead with the current pattern?

[dmgreen]

Does this still only handle v16i8?

[RKSimon]

llvm/lib/Target/AArch64/AArch64ISelLowering.cpp

11067 ↗

(On Diff #296632)

Any chance we could move some of this pattern matching into SelectionDAG? I'd love to get rid of some of the almost identical code from X86ISelLowering.

[mivnay]

In D88742#2318989, @dmgreen wrote:

Does this still only handle v16i8?

Currently the patterns written are for v16i8 to v16i16 and v16i8 to v16i32 types. I am not sure what are the other types needed to support along with this as there are at least 24 more patterns if we consider signed and floating points excluding the vabdl ones which are already in td files (Reference : https://developer.arm.com/architectures/instruction-sets/simd-isas/neon/intrinsics?search=vabd).

There are 12 type pair possible for unsigned types alone:

1. v8i8 to i16,i32,i64 vector types with same number of elements
2. v16i8 to i16,i32,i64 vector types
3. v4i16 to i32,i64 vector types
4. v8i16 to i32, i64 vector types
5. v2i32 to v2i64 vector type
6. v4i32 to v4i64 vector type

I have handled the current patterns to solve the bug mentioned in the ticket. Which other types do you have in mind?

[mivnay]

llvm/lib/Target/AArch64/AArch64ISelLowering.cpp

11067 ↗

(On Diff #296632)

I have only handled one of the ABD patterns here. There are more in td files and in other places, which would need a bigger refactoring. I think this is out of purview for this patch and can be filed as an enhancement.

[dmgreen]

Currently the patterns written are for v16i8 to v16i16 and v16i8 to v16i32 types. I am not sure what are the other types needed to support along with this as there are at least 24 more patterns if we consider signed and floating points excluding the vabdl ones which are already in td files (Reference : https://developer.arm.com/architectures/instruction-sets/simd-isas/neon/intrinsics?search=vabd).

There are 12 type pair possible for unsigned types alone:

1. v8i8 to i16,i32,i64 vector types with same number of elements
2. v16i8 to i16,i32,i64 vector types
3. v4i16 to i32,i64 vector types
4. v8i16 to i32, i64 vector types
5. v2i32 to v2i64 vector type
6. v4i32 to v4i64 vector type

I have handled the current patterns to solve the bug mentioned in the ticket. Which other types do you have in mind?

If we are adding UABD/SABD pattern matching I would expect us to generalize it to support for all the types that instruction supports. You can ignore UABDL though, I agree that is a separate issue (and seems to be handled elsewhere).

The "to" type isn't very important - it just needs to be big enough to make the results equivalent. I think in this case that's only 1 or 2 bits, and it could even not be a power 2 (although that usually isn't very important to support). The "from" type would be i8, i16 or i32, handling either 64 or 128bit vectors. I think if we effectively ignore the "to" type (because it gets handled by the ZERO_EXTEND), what you have here is already pretty close.

llvm/lib/Target/AArch64/AArch64ISelLowering.cpp

11051 ↗

(On Diff #296632)

You might want to simplify this and keep it in a single performABSCombine function.

llvm/lib/Target/AArch64/AArch64ISelLowering.h

225 ↗

(On Diff #296632)

Can you move this past SRHADD? SHADD and SRHADD deserve to live next to each other.

llvm/test/CodeGen/AArch64/arm64-vabs.ll

165 ↗

(On Diff #296632)

The tests needn't do this load if you pass the value as a function argument directly.

[dmgreen]

Thanks. https://alive2.llvm.org/ce/z/R7TNGK suggests this should be fine for sign extends too, which would be great to add.

@RKSimon As I mentioned in https://reviews.llvm.org/D88897#2316214 I'd like to make this target independent, If you don't mind waiting for me to port this over I am happy to do that later.

lib/Target/AArch64/AArch64ISelLowering.cpp

11085–11086

Can you change this to checking for sign extend or zero extend, then produce AArch64ISD::UABD or AArch64ISD::SABD as needed.

I believe that the final ISD::ZERO_EXTEND would remain as a zext for UABD or SABD.

[mivnay]

Adding support for SABD patterns..

[dmgreen]

Thank you very much. With one little nitpick, LGTM.

lib/Target/AArch64/AArch64ISelLowering.cpp

11092

size -> Size

lib/Target/AArch64/AArch64ISelLowering.h

enum NodeType : unsigned {
   // Vector halving addition
   SHADD,
   UHADD,
 
   // Vector rounding halving addition
   SRHADD,
   URHADD,
 
+  // Absolute difference
+  UABD,
+  SABD,
+
   // Vector across-lanes min/max
   // Only the lower result lane is defined.
   SMINV,
   UMINV,
   SMAXV,
   UMAXV,
 
   SADDV_PRED,
EVT getSetCCResultType(const DataLayout &DL, LLVMContext &Context,
                          EVT VT) const override;
 
   SDValue ReconstructShuffle(SDValue Op, SelectionDAG &DAG) const;
 
   MachineBasicBlock *EmitF128CSEL(MachineInstr &MI,
                                   MachineBasicBlock *BB) const;
 
   MachineBasicBlock *EmitLoweredCatchRet(MachineInstr &MI,
                                            MachineBasicBlock *BB) const;
 
   MachineBasicBlock *
   EmitInstrWithCustomInserter(MachineInstr &MI,
                               MachineBasicBlock *MBB) const override;
 
   bool getTgtMemIntrinsic(IntrinsicInfo &Info, const CallInst &I,
                           MachineFunction &MF,
                           unsigned Intrinsic) const override;
bool shouldFormOverflowOp(unsigned Opcode, EVT VT,
                             bool MathUsed) const override {
     // Using overflow ops for overflow checks only should beneficial on
     // AArch64.
     return TargetLowering::shouldFormOverflowOp(Opcode, VT, true);
   }
 
   Value *emitLoadLinked(IRBuilder<> &Builder, Value *Addr,
                         AtomicOrdering Ord) const override;
   Value *emitStoreConditional(IRBuilder<> &Builder, Value *Val,
                               Value *Addr, AtomicOrdering Ord) const override;
 
   void emitAtomicCmpXchgNoStoreLLBalance(IRBuilder<> &Builder) const override;
 
   TargetLoweringBase::AtomicExpansionKind
   shouldExpandAtomicLoadInIR(LoadInst *LI) const override;
   bool shouldExpandAtomicStoreInIR(StoreInst *SI) const override;
   TargetLoweringBase::AtomicExpansionKind
bool canMergeStoresTo(unsigned AddressSpace, EVT MemVT,
     bool NoFloat = DAG.getMachineFunction().getFunction().hasFnAttribute(
         Attribute::NoImplicitFloat);
 
     if (NoFloat)
       return (MemVT.getSizeInBits() <= 64);
     return true;
   }
 
   bool isCheapToSpeculateCttz() const override {
     return true;
   }
 
   bool isCheapToSpeculateCtlz() const override {
     return true;
   }
 
   bool isMaskAndCmp0FoldingBeneficial(const Instruction &AndI) const override;
 
   bool hasAndNotCompare(SDValue V) const override {
     // We can use bics for any scalar.
     return V.getValueType().isScalarInteger();
bool supportSplitCSR(MachineFunction *MF) const override {
     return MF->getFunction().getCallingConv() == CallingConv::CXX_FAST_TLS &&
            MF->getFunction().hasFnAttribute(Attribute::NoUnwind);
   }
   void initializeSplitCSR(MachineBasicBlock *Entry) const override;
   void insertCopiesSplitCSR(
       MachineBasicBlock *Entry,
       const SmallVectorImpl<MachineBasicBlock *> &Exits) const override;
 
   bool supportSwiftError() const override {
     return true;
   }
 
   /// Enable aggressive FMA fusion on targets that want it.
   bool enableAggressiveFMAFusion(EVT VT) const override;
 
   /// Returns the size of the platform's va_list object.
   unsigned getVaListSizeInBits(const DataLayout &DL) const override;
 
   /// Returns true if \p VecTy is a legal interleaved access type. This
   /// function checks the vector element type and the overall width of the
   /// vector.
   bool isLegalInterleavedAccessType(VectorType *VecTy,
                                     const DataLayout &DL) const;
 
   /// Returns the number of interleaved accesses that will be generated when
   /// lowering accesses of the given type.
   unsigned getNumInterleavedAccesses(VectorType *VecTy,
                                      const DataLayout &DL) const;
 
   MachineMemOperand::Flags getTargetMMOFlags(
     const Instruction &I) const override;
 
   bool functionArgumentNeedsConsecutiveRegisters(Type *Ty,
                                                  CallingConv::ID CallConv,
                                                  bool isVarArg) const override;
   /// Used for exception handling on Win64.
   bool needsFixedCatchObjects() const override;
 
private:
 
   SDValue LowerFixedLengthVectorIntDivideToSVE(SDValue Op,
                                                SelectionDAG &DAG) const;
   SDValue LowerFixedLengthVectorIntExtendToSVE(SDValue Op,
                                                SelectionDAG &DAG) const;
   SDValue LowerFixedLengthVectorLoadToSVE(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFixedLengthReductionToSVE(unsigned Opcode, SDValue ScalarOp,
                                          SelectionDAG &DAG) const;
   SDValue LowerFixedLengthVectorSelectToSVE(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFixedLengthVectorSetccToSVE(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFixedLengthVectorStoreToSVE(SDValue Op, SelectionDAG &DAG) const;
   SDValue LowerFixedLengthVectorTruncateToSVE(SDValue Op,
                                               SelectionDAG &DAG) const;
 
   SDValue BuildSDIVPow2(SDNode *N, const APInt &Divisor, SelectionDAG &DAG,
                         SmallVectorImpl<SDNode *> &Created) const override;
   SDValue getSqrtEstimate(SDValue Operand, SelectionDAG &DAG, int Enabled,
                           int &ExtraSteps, bool &UseOneConst,
                           bool Reciprocal) const override;
   SDValue getRecipEstimate(SDValue Operand, SelectionDAG &DAG, int Enabled,
                            int &ExtraSteps) const override;
   unsigned combineRepeatedFPDivisors() const override;
 
   ConstraintType getConstraintType(StringRef Constraint) const override;
   Register getRegisterByName(const char* RegName, LLT VT,
                              const MachineFunction &MF) const override;
 
   /// Examine constraint string and operand type and determine a weight value.
   /// The operand object must already have been set up with the operand type.
   ConstraintWeight
   getSingleConstraintMatchWeight(AsmOperandInfo &info,
                                  const char *constraint) const override;
 

lib/Target/AArch64/AArch64ISelLowering.cpp

 // 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));
 
 /// Value type used for condition codes.
 static const MVT MVT_CC = MVT::i32;
 
 static inline EVT getPackedSVEVectorVT(EVT VT) {
if (useSVEForFixedLengthVectors()) {
         if (useSVEForFixedLengthVectorVT(VT))
           addRegisterClass(VT, &AArch64::ZPRRegClass);
 
       for (MVT VT : MVT::fp_fixedlen_vector_valuetypes())
         if (useSVEForFixedLengthVectorVT(VT))
           addRegisterClass(VT, &AArch64::ZPRRegClass);
     }
 
     for (auto VT : { MVT::nxv16i8, MVT::nxv8i16, MVT::nxv4i32, MVT::nxv2i64 }) {
       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);
     }
 
     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::nxv2f16, MVT::nxv4f16, MVT::nxv8f16, MVT::nxv2f32, MVT::nxv4f32,
            MVT::nxv2f64 }) {
       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);
AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
   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);
   else
     setOperationAction(ISD::FCOPYSIGN, MVT::f16, Promote);
 
   setOperationAction(ISD::FREM,    MVT::f16,   Promote);
   setOperationAction(ISD::FREM,    MVT::v4f16, Expand);
   setOperationAction(ISD::FREM,    MVT::v8f16, Expand);
   setOperationAction(ISD::FPOW,    MVT::f16,   Promote);
   setOperationAction(ISD::FPOW,    MVT::v4f16, Expand);
   setOperationAction(ISD::FPOW,    MVT::v8f16, Expand);
   setOperationAction(ISD::FPOWI,   MVT::f16,   Promote);
   setOperationAction(ISD::FPOWI,   MVT::v4f16, Expand);
   setOperationAction(ISD::FPOWI,   MVT::v8f16, Expand);
   setOperationAction(ISD::FCOS,    MVT::f16,   Promote);
   setOperationAction(ISD::FCOS,    MVT::v4f16, Expand);
   setOperationAction(ISD::FCOS,    MVT::v8f16, Expand);
   setOperationAction(ISD::FSIN,    MVT::f16,   Promote);
   setOperationAction(ISD::FSIN,    MVT::v4f16, Expand);
   setOperationAction(ISD::FSIN,    MVT::v8f16, Expand);
   setOperationAction(ISD::FSINCOS, MVT::f16,   Promote);
   setOperationAction(ISD::FSINCOS, MVT::v4f16, Expand);
   setOperationAction(ISD::FSINCOS, MVT::v8f16, Expand);
   setOperationAction(ISD::FEXP,    MVT::f16,   Promote);
   setOperationAction(ISD::FEXP,    MVT::v4f16, Expand);
   setOperationAction(ISD::FEXP,    MVT::v8f16, Expand);
   setOperationAction(ISD::FEXP2,   MVT::f16,   Promote);
   setOperationAction(ISD::FEXP2,   MVT::v4f16, Expand);
   setOperationAction(ISD::FEXP2,   MVT::v8f16, Expand);
   setOperationAction(ISD::FLOG,    MVT::f16,   Promote);
   setOperationAction(ISD::FLOG,    MVT::v4f16, Expand);
   setOperationAction(ISD::FLOG,    MVT::v8f16, Expand);
   setOperationAction(ISD::FLOG2,   MVT::f16,   Promote);
   setOperationAction(ISD::FLOG2,   MVT::v4f16, Expand);
   setOperationAction(ISD::FLOG2,   MVT::v8f16, Expand);
   setOperationAction(ISD::FLOG10,  MVT::f16,   Promote);
   setOperationAction(ISD::FLOG10,  MVT::v4f16, Expand);
   setOperationAction(ISD::FLOG10,  MVT::v8f16, Expand);
 
   if (!Subtarget->hasFullFP16()) {
     setOperationAction(ISD::SELECT,      MVT::f16,  Promote);
     setOperationAction(ISD::SELECT_CC,   MVT::f16,  Promote);
     setOperationAction(ISD::SETCC,       MVT::f16,  Promote);
     setOperationAction(ISD::BR_CC,       MVT::f16,  Promote);
     setOperationAction(ISD::FADD,        MVT::f16,  Promote);
     setOperationAction(ISD::FSUB,        MVT::f16,  Promote);
     setOperationAction(ISD::FMUL,        MVT::f16,  Promote);
     setOperationAction(ISD::FDIV,        MVT::f16,  Promote);
     setOperationAction(ISD::FMA,         MVT::f16,  Promote);
     setOperationAction(ISD::FNEG,        MVT::f16,  Promote);
     setOperationAction(ISD::FABS,        MVT::f16,  Promote);
     setOperationAction(ISD::FCEIL,       MVT::f16,  Promote);
     setOperationAction(ISD::FSQRT,       MVT::f16,  Promote);
     setOperationAction(ISD::FFLOOR,      MVT::f16,  Promote);
     setOperationAction(ISD::FNEARBYINT,  MVT::f16,  Promote);
     setOperationAction(ISD::FRINT,       MVT::f16,  Promote);
     setOperationAction(ISD::FROUND,      MVT::f16,  Promote);
     setOperationAction(ISD::FTRUNC,      MVT::f16,  Promote);
     setOperationAction(ISD::FMINNUM,     MVT::f16,  Promote);
     setOperationAction(ISD::FMAXNUM,     MVT::f16,  Promote);
     setOperationAction(ISD::FMINIMUM,    MVT::f16,  Promote);
     setOperationAction(ISD::FMAXIMUM,    MVT::f16,  Promote);
 
     // promote v4f16 to v4f32 when that is known to be safe.
     setOperationAction(ISD::FADD,        MVT::v4f16, Promote);
     setOperationAction(ISD::FSUB,        MVT::v4f16, Promote);
     setOperationAction(ISD::FMUL,        MVT::v4f16, Promote);
     setOperationAction(ISD::FDIV,        MVT::v4f16, Promote);
     AddPromotedToType(ISD::FADD,         MVT::v4f16, MVT::v4f32);
     AddPromotedToType(ISD::FSUB,         MVT::v4f16, MVT::v4f32);
     AddPromotedToType(ISD::FMUL,         MVT::v4f16, MVT::v4f32);
     AddPromotedToType(ISD::FDIV,         MVT::v4f16, MVT::v4f32);
 
for (MVT Ty : {MVT::f32, MVT::f64}) {
     setOperationAction(ISD::LROUND, Ty, Legal);
     setOperationAction(ISD::LLROUND, Ty, Legal);
     setOperationAction(ISD::LRINT, Ty, Legal);
     setOperationAction(ISD::LLRINT, Ty, Legal);
   }
 
   if (Subtarget->hasFullFP16()) {
     setOperationAction(ISD::FNEARBYINT, MVT::f16, Legal);
     setOperationAction(ISD::FFLOOR,  MVT::f16, Legal);
     setOperationAction(ISD::FCEIL,   MVT::f16, Legal);
     setOperationAction(ISD::FRINT,   MVT::f16, Legal);
     setOperationAction(ISD::FTRUNC,  MVT::f16, Legal);
     setOperationAction(ISD::FROUND,  MVT::f16, Legal);
     setOperationAction(ISD::FMINNUM, MVT::f16, Legal);
     setOperationAction(ISD::FMAXNUM, MVT::f16, Legal);
     setOperationAction(ISD::FMINIMUM, MVT::f16, Legal);
     setOperationAction(ISD::FMAXIMUM, MVT::f16, Legal);
AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
   // We combine OR nodes for bitfield operations.
   setTargetDAGCombine(ISD::OR);
   // Try to create BICs for vector ANDs.
   setTargetDAGCombine(ISD::AND);
 
   // Vector add and sub nodes may conceal a high-half opportunity.
   // Also, try to fold ADD into CSINC/CSINV..
   setTargetDAGCombine(ISD::ADD);
+  setTargetDAGCombine(ISD::ABS);
   setTargetDAGCombine(ISD::SUB);
   setTargetDAGCombine(ISD::SRL);
   setTargetDAGCombine(ISD::XOR);
   setTargetDAGCombine(ISD::SINT_TO_FP);
   setTargetDAGCombine(ISD::UINT_TO_FP);
 
   setTargetDAGCombine(ISD::FP_TO_SINT);
   setTargetDAGCombine(ISD::FP_TO_UINT);
AArch64TargetLowering::AArch64TargetLowering(const TargetMachine &TM,
   setTargetDAGCombine(ISD::INSERT_VECTOR_ELT);
   setTargetDAGCombine(ISD::EXTRACT_VECTOR_ELT);
   setTargetDAGCombine(ISD::VECREDUCE_ADD);
 
   setTargetDAGCombine(ISD::GlobalAddress);
 
   // 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 = MaxStoresPerMemmove = 4;
 
   MaxLoadsPerMemcmpOptSize = 4;
   MaxLoadsPerMemcmp = Subtarget->requiresStrictAlign()
                       ? MaxLoadsPerMemcmpOptSize : 8;
 
   setStackPointerRegisterToSaveRestore(AArch64::SP);
 
   setSchedulingPreference(Sched::Hybrid);
 
   EnableExtLdPromotion = true;
 
if (Subtarget->hasFullFP16()) {
       // when AArch64 doesn't have fullfp16 support, promote the input
       // to i32 first.
       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);
 
     // AArch64 doesn't have MUL.2d:
     setOperationAction(ISD::MUL, MVT::v2i64, Expand);
     // 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);
 
     // Saturates
     for (MVT VT : { MVT::v8i8, MVT::v4i16, MVT::v2i32,
                     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);
     }
 
     // Vector reductions
     for (MVT VT : { MVT::v4f16, MVT::v2f32,
                     MVT::v8f16, MVT::v4f32, MVT::v2f64 }) {
       setOperationAction(ISD::VECREDUCE_FMAX, VT, Custom);
       setOperationAction(ISD::VECREDUCE_FMIN, VT, Custom);
     }
     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_ADD, MVT::v2i64, 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] 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})
       setOperationAction(Opcode, VT, Legal);
 
   if (Subtarget->isLittleEndian()) {
     for (unsigned im = (unsigned)ISD::PRE_INC;
static bool optimizeLogicalImm(SDValue Op, unsigned Size, uint64_t Imm,
   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);
   }
 
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::CSEL: {
     KnownBits Known2;
     Known = DAG.computeKnownBits(Op->getOperand(0), Depth + 1);
     Known2 = DAG.computeKnownBits(Op->getOperand(1), Depth + 1);
     Known.Zero &= Known2.Zero;
     Known.One &= Known2.One;
     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 ISD::INTRINSIC_W_CHAIN: {
     ConstantSDNode *CN = cast<ConstantSDNode>(Op->getOperand(1));
     Intrinsic::ID IntID = static_cast<Intrinsic::ID>(CN->getZExtValue());
     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;
     }
case AArch64ISD::FIRST_NUMBER:
     MAKE_CASE(AArch64ISD::SST1_IMM_PRED)
     MAKE_CASE(AArch64ISD::SSTNT1_PRED)
     MAKE_CASE(AArch64ISD::SSTNT1_INDEX_PRED)
     MAKE_CASE(AArch64ISD::LDP)
     MAKE_CASE(AArch64ISD::STP)
     MAKE_CASE(AArch64ISD::STNP)
     MAKE_CASE(AArch64ISD::DUP_MERGE_PASSTHRU)
     MAKE_CASE(AArch64ISD::INDEX_VECTOR)
+    MAKE_CASE(AArch64ISD::UABD)
+    MAKE_CASE(AArch64ISD::SABD)
   }
 #undef MAKE_CASE
   return nullptr;
 }
 
 MachineBasicBlock *
 AArch64TargetLowering::EmitF128CSEL(MachineInstr &MI,
                                     MachineBasicBlock *MBB) const {
BuildMI(*EndBB, EndBB->begin(), DL, TII->get(AArch64::PHI), DestReg)
       .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::EmitInstrWithCustomInserter(
unsigned Opcode =
       IsSignaling ? AArch64ISD::STRICT_FCMPE : AArch64ISD::STRICT_FCMP;
   return DAG.getNode(Opcode, dl, {VT, 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 =
     static_cast<const AArch64Subtarget &>(DAG.getSubtarget()).hasFullFP16();
 
   if (VT.isFloatingPoint()) {
     assert(VT != MVT::f128);
     if (VT == MVT::f16 && !FullFP16) {
       LHS = DAG.getNode(ISD::FP_EXTEND, dl, MVT::f32, LHS);
       RHS = DAG.getNode(ISD::FP_EXTEND, dl, MVT::f32, RHS);
       VT = MVT::f32;
     }
if (isCMN(RHS, CC)) {
     // 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

 /// 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 =
     static_cast<const AArch64Subtarget &>(DAG.getSubtarget()).hasFullFP16();
 
   if (LHS.getValueType().isFloatingPoint()) {
     assert(LHS.getValueType() != MVT::f128);
     if (LHS.getValueType() == MVT::f16 && !FullFP16) {
       LHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, LHS);
       RHS = DAG.getNode(ISD::FP_EXTEND, DL, MVT::f32, RHS);
     }
     Opcode = AArch64ISD::FCCMP;
static bool canEmitConjunction(const SDValue Val, bool &CanNegate,
   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.

 /// 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 (!isLegalArithImmed(C)) {
           C = (VT == MVT::i32) ? (uint32_t)(C - 1) : C - 1;
           RHS = DAG.getConstant(C, dl, VT);
         }
         break;
       case ISD::SETLE:
       case ISD::SETGT:
         if ((VT == MVT::i32 && C != INT32_MAX &&
              isLegalArithImmed((uint32_t)(C + 1))) ||
             (VT == MVT::i64 && C != INT64_MAX &&
              isLegalArithImmed(C + 1ULL))) {
           CC = (CC == ISD::SETLE) ? ISD::SETLT : ISD::SETGE;
           C = (VT == MVT::i32) ? (uint32_t)(C + 1) : C + 1;
           RHS = DAG.getConstant(C, dl, VT);
         }
         break;
       case ISD::SETULE:
       case ISD::SETUGT:
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 = cast<ConstantSDNode>(RHS)->getZExtValue();
       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.getConstant(ValueofRHS, dl,
                                                    RHS.getValueType()),
                              CC, dl, DAG);
         AArch64CC = changeIntCCToAArch64CC(CC);
       }
     }
 
     if (!Cmp && (RHSC->isNullValue() || RHSC->isOne())) {
       if ((Cmp = emitConjunction(DAG, LHS, AArch64CC))) {
if (Op.getValueType() == MVT::i32) {
         // The overflow check for unsigned multiply is easy. We only need to
         // check if any of the upper 32 bits are set. This can be done with a
         // CMP (shifted register). For that we need to generate the following
         // pattern:
         // (i64 AArch64ISD::SUBS i64 0, (i64 srl i64 %Mul, i64 32)
         SDValue UpperBits = DAG.getNode(ISD::SRL, DL, MVT::i64, Mul,
                                         DAG.getConstant(32, DL, MVT::i64));
         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;
     }
     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 (...)
 
SDValue AArch64TargetLowering::LowerF128Call(SDValue Op, SelectionDAG &DAG,
                                              RTLIB::Libcall Call) const {
   bool IsStrict = Op->isStrictFPOpcode();
   unsigned Offset = IsStrict ? 1 : 0;
   SDValue Chain = IsStrict ? Op.getOperand(0) : SDValue();
   SmallVector<SDValue, 2> Ops(Op->op_begin() + Offset, Op->op_end());
   MakeLibCallOptions CallOptions;
   SDValue Result;
   SDLoc dl(Op);
   std::tie(Result, Chain) = makeLibCall(DAG, Call, Op.getValueType(), Ops,
                                         CallOptions, dl, Chain);
   return IsStrict ? DAG.getMergeValues({Result, Chain}, dl) : Result;
 }
 
 SDValue AArch64TargetLowering::LowerXOR(SDValue Op, SelectionDAG &DAG) const {
   if (useSVEForFixedLengthVectorVT(Op.getValueType()))
     return LowerToScalableOp(Op, DAG);
 
static SDValue LowerXALUO(SDValue Op, SelectionDAG &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
if (Locality) {
     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.getConstant(PrfOp, DL, MVT::i32), Op.getOperand(1));
 }
 
 SDValue AArch64TargetLowering::LowerFP_EXTEND(SDValue Op,
SDValue AArch64TargetLowering::LowerFP_ROUND(SDValue Op,
 
   // FP_ROUND node has a second operand indicating whether it is known to be
   // precise. That doesn't take part in the LibCall so we can't directly use
   // LowerF128Call.
   MakeLibCallOptions CallOptions;
   SDValue Chain = IsStrict ? Op.getOperand(0) : SDValue();
   SDValue Result;
   SDLoc dl(Op);
   std::tie(Result, Chain) = makeLibCall(DAG, LC, Op.getValueType(), SrcVal,
                                         CallOptions, dl, Chain);
   return IsStrict ? DAG.getMergeValues({Result, Chain}, dl) : Result;
 }
 
 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.
   EVT InVT = Op.getOperand(0).getValueType();
   EVT VT = Op.getValueType();
 
   if (VT.isScalableVector()) {
     unsigned Opcode = Op.getOpcode() == ISD::FP_TO_UINT
                           ? AArch64ISD::FCVTZU_MERGE_PASSTHRU
                           : AArch64ISD::FCVTZS_MERGE_PASSTHRU;
     return LowerToPredicatedOp(Op, DAG, Opcode);
   }
 
   unsigned NumElts = InVT.getVectorNumElements();
 
   // f16 conversions are promoted to f32 when full fp16 is not supported.
   if (InVT.getVectorElementType() == MVT::f16 &&
       !Subtarget->hasFullFP16()) {
     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)));
   }
 
SDValue AArch64TargetLowering::LowerFP_TO_INT(SDValue Op,
 
   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()) {
     assert(!IsStrict && "Lowering of strict fp16 not yet implemented");
     SDLoc dl(Op);
     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;
   }
if (VT.getSizeInBits() > InVT.getSizeInBits()) {
     In = DAG.getNode(CastOpc, dl, CastVT, In);
     return DAG.getNode(Op.getOpcode(), dl, VT, In);
   }
 
   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);
 
   // f16 conversions are promoted to f32 when full fp16 is not supported.
   if (Op.getValueType() == MVT::f16 &&
       !Subtarget->hasFullFP16()) {
     assert(!IsStrict && "Lowering of strict fp16 not yet implemented");
     SDLoc dl(Op);
     return DAG.getNode(
         ISD::FP_ROUND, dl, MVT::f16,
         DAG.getNode(Op.getOpcode(), dl, MVT::f32, SrcVal),
         DAG.getIntPtrConstant(0, dl));
   }
 
   // i128 conversions are libcalls.
   if (SrcVal.getValueType() == MVT::i128)
     return SDValue();
SDValue AArch64TargetLowering::LowerFSINCOS(SDValue Op,
   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);
   CLI.setDebugLoc(dl)

 static EVT getExtensionTo64Bits(const EVT &OrigVT) {
   if (OrigVT.getSizeInBits() >= 64)
     return OrigVT;
 
   assert(OrigVT.isSimple() && "Expecting a simple value type");
 
   MVT::SimpleValueType OrigSimpleTy = OrigVT.getSimpleVT().SimpleTy;
   switch (OrigSimpleTy) {
   default: llvm_unreachable("Unexpected Vector Type");
   case MVT::v2i8:
   case MVT::v2i16:
      return MVT::v2i32;
   case MVT::v4i8:
     return  MVT::v4i16;
   }
 }
 
 static SDValue addRequiredExtensionForVectorMULL(SDValue N, SelectionDAG &DAG,
                                                  const EVT &OrigTy,
                                                  const EVT &ExtTy,
                                                  unsigned ExtOpcode) {
   // The vector originally had a size of OrigTy. It was then extended to ExtTy.
for (const SDValue &Elt : N->op_values()) {
     return false;
   }
 
   return true;
 }
 
 static SDValue skipExtensionForVectorMULL(SDNode *N, SelectionDAG &DAG) {
   if (N->getOpcode() == ISD::SIGN_EXTEND || N->getOpcode() == ISD::ZERO_EXTEND)
     return addRequiredExtensionForVectorMULL(N->getOperand(0), DAG,
                                              N->getOperand(0)->getValueType(0),
                                              N->getValueType(0),
                                              N->getOpcode());
 
   assert(N->getOpcode() == ISD::BUILD_VECTOR && "expected BUILD_VECTOR");
   EVT VT = N->getValueType(0);
   SDLoc dl(N);
   unsigned EltSize = VT.getScalarSizeInBits() / 2;
return N->getOpcode() == ISD::ZERO_EXTEND ||
          isExtendedBUILD_VECTOR(N, DAG, false);
 }
 
 static bool isAddSubSExt(SDNode *N, SelectionDAG &DAG) {
   unsigned Opcode = N->getOpcode();
   if (Opcode == ISD::ADD || Opcode == ISD::SUB) {
     SDNode *N0 = N->getOperand(0).getNode();
     SDNode *N1 = N->getOperand(1).getNode();
     return N0->hasOneUse() && N1->hasOneUse() &&
       isSignExtended(N0, DAG) && isSignExtended(N1, DAG);
   }
   return false;
 }
 
 static bool isAddSubZExt(SDNode *N, SelectionDAG &DAG) {
   unsigned Opcode = N->getOpcode();
   if (Opcode == ISD::ADD || Opcode == ISD::SUB) {
     SDNode *N0 = N->getOperand(0).getNode();
     SDNode *N1 = N->getOperand(1).getNode();
     return N0->hasOneUse() && N1->hasOneUse() &&
       isZeroExtended(N0, DAG) && isZeroExtended(N1, DAG);
   }
   return false;
 }
 
 SDValue AArch64TargetLowering::LowerFLT_ROUNDS_(SDValue Op,
                                                 SelectionDAG &DAG) const {
   // The rounding mode is in bits 23:22 of the FPSCR.
if (isN0ZExt && isN1ZExt)
       NewOpc = AArch64ISD::UMULL;
     else if (isN1SExt || isN1ZExt) {
       // 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)) {
         NewOpc = AArch64ISD::SMULL;
         isMLA = true;
       } else if (isN1ZExt && isAddSubZExt(N0, DAG)) {
         NewOpc =  AArch64ISD::UMULL;
         isMLA = true;
       } else if (isN0ZExt && isAddSubZExt(N1, DAG)) {
         std::swap(N0, N1);
         NewOpc =  AArch64ISD::UMULL;
         isMLA = true;
       }
     }
 
     if (!NewOpc) {
       if (VT == MVT::v2i64)
         // Fall through to expand this.  It is not legal.
         return SDValue();
if (!isMLA) {
     return DAG.getNode(NewOpc, DL, VT, Op0, Op1);
   }
   // 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).getNode(), DAG);
   SDValue N01 = skipExtensionForVectorMULL(N0->getOperand(1).getNode(), DAG);
   EVT Op1VT = Op1.getValueType();
   return 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));
 }
 
 static inline SDValue getPTrue(SelectionDAG &DAG, SDLoc DL, EVT VT,
                                int Pattern) {
   return DAG.getNode(AArch64ISD::PTRUE, DL, VT,
                      DAG.getTargetConstant(Pattern, DL, MVT::i32));
 }
 
 SDValue AArch64TargetLowering::LowerINTRINSIC_WO_CHAIN(SDValue Op,
                                                      SelectionDAG &DAG) const {
   unsigned IntNo = cast<ConstantSDNode>(Op.getOperand(0))->getZExtValue();
   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 intrinic");
     }
   }
   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_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));
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(AArch64ISD::REV, 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::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_neon_uhadd: {
     bool IsRoundingAdd = (IntNo == Intrinsic::aarch64_neon_srhadd ||
                           IntNo == Intrinsic::aarch64_neon_urhadd);
     unsigned Opcode =
         IsSignedAdd ? (IsRoundingAdd ? AArch64ISD::SRHADD : AArch64ISD::SHADD)
                     : (IsRoundingAdd ? AArch64ISD::URHADD : AArch64ISD::UHADD);
     return DAG.getNode(Opcode, dl, Op.getValueType(), Op.getOperand(1),
                        Op.getOperand(2));
   }
+
+  case Intrinsic::aarch64_neon_uabd: {
+    return DAG.getNode(AArch64ISD::UABD, dl, Op.getValueType(),
+                       Op.getOperand(1), Op.getOperand(2));
+  }
+  case Intrinsic::aarch64_neon_sabd: {
+    return DAG.getNode(AArch64ISD::SABD, dl, Op.getValueType(),
+                       Op.getOperand(1), Op.getOperand(2));
+  }
   }
 }
 
 bool AArch64TargetLowering::isVectorLoadExtDesirable(SDValue ExtVal) const {
   return ExtVal.getValueType().isScalableVector();
 }
 
 // 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());
 }
 
 // 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))
case CallingConv::Swift:
     if (Subtarget->isTargetWindows() && IsVarArg)
       return CC_AArch64_Win64_VarArg;
     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:
     return IsVarArg ? CC_AArch64_Win64_VarArg : CC_AArch64_AAPCS;
    case CallingConv::CFGuard_Check:
      return CC_AArch64_Win64_CFGuard_Check;
    case CallingConv::AArch64_VectorCall:
    case CallingConv::AArch64_SVE_VectorCall:
      return CC_AArch64_AAPCS;
   }
 }
 
 CCAssignFn *
 AArch64TargetLowering::CCAssignFnForReturn(CallingConv::ID CC) const {
   return CC == CallingConv::WebKit_JS ? RetCC_AArch64_WebKit_JS
                                       : RetCC_AArch64_AAPCS;
 }
 
 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();
   MachineFrameInfo &MFI = MF.getFrameInfo();
   bool IsWin64 = Subtarget->isCallingConvWin64(MF.getFunction().getCallingConv());
 
   // Assign locations to all of the incoming arguments.
   SmallVector<CCValAssign, 16> ArgLocs;
   DenseMap<unsigned, SDValue> CopiedRegs;
   CCState CCInfo(CallConv, isVarArg, DAG.getMachineFunction(), ArgLocs,
                  *DAG.getContext());
 
   // At this point, Ins[].VT may already be promoted to i32. To correctly
for (unsigned i = 0, e = ArgLocs.size(); i != e; ++i) {
 
     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;
     }
 
     SDValue ArgValue;
     if (VA.isRegLoc()) {
if (VA.isRegLoc()) {
         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;
 
       int FI = MFI.CreateFixedObject(ArgSize, ArgOffset + BEAlign, true);
 
if (VA.isRegLoc()) {
         ExtType = ISD::EXTLOAD;
         break;
       }
 
       ArgValue = DAG.getExtLoad(
           ExtType, DL, VA.getLocVT(), Chain, FIN,
           MachinePointerInfo::getFixedStack(DAG.getMachineFunction(), FI),
           MemVT);
 
     }
 
     if (VA.getLocInfo() == CCValAssign::Indirect) {
       assert(VA.getValVT().isScalableVector() &&
            "Only scalable vectors can be passed indirectly");
       // If value is passed via pointer - do a load.
       ArgValue =
           DAG.getLoad(VA.getValVT(), DL, Chain, ArgValue, MachinePointerInfo());
     }
 
     if (Subtarget->isTargetILP32() && Ins[i].Flags.isPointer())
       ArgValue = DAG.getNode(ISD::AssertZext, DL, ArgValue.getValueType(),
                              ArgValue, DAG.getValueType(MVT::i32));
     InVals.push_back(ArgValue);
   }
 
   // varargs
   AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
   if (isVarArg) {
     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 StackOffset = CCInfo.getNextStackOffset();
     // We currently pass all varargs at 8-byte alignment, or 4 for ILP32
     StackOffset = alignTo(StackOffset, Subtarget->isTargetILP32() ? 4 : 8);
     FuncInfo->setVarArgsStackIndex(MFI.CreateFixedObject(4, StackOffset, 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)) {
         unsigned X8VReg = MF.addLiveIn(AArch64::X8, &AArch64::GPR64RegClass);
         Forwards.push_back(ForwardedRegister(X8VReg, AArch64::X8, MVT::i64));
       }

 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());
   bool IsWin64 = Subtarget->isCallingConvWin64(MF.getFunction().getCallingConv());
 
   SmallVector<SDValue, 8> MemOps;
 
   static const MCPhysReg GPRArgRegs[] = { AArch64::X0, AArch64::X1, AArch64::X2,
                                           AArch64::X3, AArch64::X4, AArch64::X5,
                                           AArch64::X6, AArch64::X7 };
   static const unsigned NumGPRArgRegs = array_lengthof(GPRArgRegs);
   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 = DAG.getFrameIndex(GPRIdx, PtrVT);
 
     for (unsigned i = FirstVariadicGPR; i < NumGPRArgRegs; ++i) {
       unsigned VReg = MF.addLiveIn(GPRArgRegs[i], &AArch64::GPR64RegClass);
       SDValue Val = DAG.getCopyFromReg(Chain, DL, VReg, MVT::i64);
AArch64TargetLowering::LowerCall(CallLoweringInfo &CLI,
 
   AArch64FunctionInfo *FuncInfo = MF.getInfo<AArch64FunctionInfo>();
   bool TailCallOpt = MF.getTarget().Options.GuaranteedTailCallOpt;
   bool IsSibCall = false;
 
   // Check callee args/returns for SVE registers and set calling convention
   // accordingly.
   if (CallConv == CallingConv::C) {
     bool CalleeOutSVE = any_of(Outs, [](ISD::OutputArg &Out){
       return Out.VT.isScalableVector();
     });
     bool CalleeInSVE = any_of(Ins, [](ISD::InputArg &In){
       return In.VT.isScalableVector();
     });
 
     if (CalleeInSVE || CalleeOutSVE)
       CallConv = CallingConv::AArch64_SVE_VectorCall;
   }
 
   if (IsTailCall) {
     // Check if it's really possible to do a tail call.
     IsTailCall = isEligibleForTailCallOptimization(
         Callee, CallConv, IsVarArg, Outs, OutVals, Ins, DAG);
     if (!IsTailCall && CLI.CB && CLI.CB->isMustTailCall())
       report_fatal_error("failed to perform tail call elimination on a call "
                          "site marked musttail");
 
     // 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)
if (IsVarArg) {
     // Handle fixed and variable vector arguments differently.
     // Variable vector arguments always go into memory.
     unsigned NumArgs = Outs.size();
 
     for (unsigned i = 0; i != NumArgs; ++i) {
       MVT ArgVT = Outs[i].VT;
       ISD::ArgFlagsTy ArgFlags = Outs[i].Flags;
       CCAssignFn *AssignFn = CCAssignFnForCall(CallConv,
                                                /*IsVarArg=*/ !Outs[i].IsFixed);
       bool Res = AssignFn(i, ArgVT, ArgVT, CCValAssign::Full, ArgFlags, CCInfo);
       assert(!Res && "Call operand has unhandled type");
       (void)Res;
     }
   } else {
     // At this point, Outs[].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.
AArch64TargetLowering::LowerCall(CallLoweringInfo &CLI,
   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.
   for (unsigned i = 0, e = ArgLocs.size(); i != e; ++i) {
     CCValAssign &VA = ArgLocs[i];
     SDValue Arg = OutVals[i];
     ISD::ArgFlagsTy Flags = Outs[i].Flags;
AArch64TargetLowering::LowerCall(CallLoweringInfo &CLI,
 
   if (!MemOpChains.empty())
     Chain = DAG.getNode(ISD::TokenFactor, DL, MVT::Other, MemOpChains);
 
   // 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.
   SDValue InFlag;
   for (auto &RegToPass : RegsToPass) {
     Chain = DAG.getCopyToReg(Chain, DL, RegToPass.first,
                              RegToPass.second, InFlag);
     InFlag = 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.
   if (auto *G = dyn_cast<GlobalAddressSDNode>(Callee)) {
if (IsTailCall) {
     // this information must travel along with the operation for eventual
     // consumption by emitEpilogue.
     Ops.push_back(DAG.getTargetConstant(FPDiff, DL, MVT::i32));
   }
 
   // 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 (unsigned SRetReg = FuncInfo->getSRetReturnReg()) {
     SDValue Val = DAG.getCopyFromReg(RetOps[0], DL, SRetReg,
                                      getPointerTy(MF.getDataLayout()));
 
     unsigned RetValReg = AArch64::X0;
     Chain = DAG.getCopyToReg(Chain, DL, RetValReg, Val, Flag);
     Flag = Chain.getValue(1);
 
     RetOps.push_back(
       DAG.getRegister(RetValReg, getPointerTy(DAG.getDataLayout())));
   }
 
   const AArch64RegisterInfo *TRI = Subtarget->getRegisterInfo();
   const MCPhysReg *I =
       TRI->getCalleeSavedRegsViaCopy(&DAG.getMachineFunction());
   if (I) {
     for (; *I; ++I) {
       if (AArch64::GPR64RegClass.contains(*I))

 
 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);
 }
 
 // (loadGOT sym)
 template <class NodeTy>
 SDValue AArch64TargetLowering::getGOT(NodeTy *N, SelectionDAG &DAG,
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
AArch64TargetLowering::LowerWindowsGlobalTLSAddress(SDValue Op,
   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();
if (SrcVT.bitsLT(VT))
     In2 = DAG.getNode(ISD::FP_EXTEND, DL, VT, In2);
   else if (SrcVT.bitsGT(VT))
     In2 = DAG.getNode(ISD::FP_ROUND, DL, VT, In2, DAG.getIntPtrConstant(0, DL));
 
   EVT VecVT;
   uint64_t EltMask;
   SDValue VecVal1, VecVal2;
 
   auto setVecVal = [&] (int Idx) {
     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 = DAG.getNode(ISD::BITCAST, DL, VecVT, In1);
       VecVal2 = DAG.getNode(ISD::BITCAST, DL, VecVT, In2);
     }
   };
if (Opcode == AArch64ISD::CSEL && RHSVal && !RHSVal->isOne() &&
       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());
       }
if (getTargetMachine().getCodeModel() == CodeModel::Large &&
       !Subtarget->isTargetMachO()) {
     return getAddrLarge(JT, DAG);
   } else if (getTargetMachine().getCodeModel() == 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();
 
SDValue AArch64TargetLowering::LowerConstantPool(SDValue Op,
   } else if (getTargetMachine().getCodeModel() == CodeModel::Tiny) {
     return getAddrTiny(CP, DAG);
   } else {
     return getAddr(CP, DAG);
   }
 }
 
 SDValue AArch64TargetLowering::LowerBlockAddress(SDValue Op,
                                                SelectionDAG &DAG) const {
   BlockAddressSDNode *BA = cast<BlockAddressSDNode>(Op);
   if (getTargetMachine().getCodeModel() == CodeModel::Large &&
       !Subtarget->isTargetMachO()) {
     return getAddrLarge(BA, DAG);
   } else if (getTargetMachine().getCodeModel() == CodeModel::Tiny) {
     return getAddrTiny(BA, DAG);
   }
   return getAddr(BA, 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();
SDValue FR = DAG.getFrameIndex(FuncInfo->getVarArgsGPRSize() > 0
                                      : 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>();
   auto PtrVT = getPointerTy(DAG.getDataLayout());
   SDLoc DL(Op);
 
   SDValue Chain = Op.getOperand(0);

 }
 
 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 : 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, false,
                        MachinePointerInfo(DestSV), MachinePointerInfo(SrcSV));
SDValue AArch64TargetLowering::LowerVAARG(SDValue Op, SelectionDAG &DAG) const {
   // 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));
     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,
SDValue AArch64TargetLowering::LowerSPONENTRY(SDValue Op,
   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 MCRegisterInfo *MRI = Subtarget->getRegisterInfo();
     unsigned DwarfRegNum = MRI->getDwarfRegNum(Reg, false);
     if (!Subtarget->isXRegisterReserved(DwarfRegNum))
       Reg = 0;
   }
   if (Reg)
     return Reg;
   report_fatal_error(Twine("Invalid register name \""
                               + StringRef(RegName)  + "\"."));
 }
 
 SDValue AArch64TargetLowering::LowerADDROFRETURNADDR(SDValue Op,
                                                      SelectionDAG &DAG) const {
   DAG.getMachineFunction().getFrameInfo().setFrameAddressIsTaken(true);
 
   EVT VT = Op.getValueType();
SDValue AArch64TargetLowering::LowerShiftRightParts(SDValue Op,
   SDValue LoForBigShift = DAG.getNode(Opc, dl, VT, ShOpHi, ExtraShAmt);
   SDValue Lo = DAG.getNode(AArch64ISD::CSEL, dl, VT, LoForBigShift,
                            LoForNormalShift, CCVal, Cmp);
 
   // AArch64 shifts larger than the register width are wrapped rather than
   // clamped, so we can't just emit "hi >> x".
   SDValue HiForNormalShift = DAG.getNode(Opc, dl, VT, ShOpHi, ShAmt);
   SDValue HiForBigShift =
       Opc == ISD::SRA
           ? DAG.getNode(Opc, dl, VT, ShOpHi,
                         DAG.getConstant(VTBits - 1, dl, MVT::i64))
           : DAG.getConstant(0, dl, VT);
   SDValue Hi = DAG.getNode(AArch64ISD::CSEL, dl, VT, HiForBigShift,
                            HiForNormalShift, CCVal, Cmp);
 
   SDValue Ops[2] = { Lo, Hi };
   return DAG.getMergeValues(Ops, dl);
 }
 
 /// LowerShiftLeftParts - Lower SHL_PARTS, which returns two
 /// i64 values and take a 2 x i64 value to shift plus a shift amount.
 SDValue AArch64TargetLowering::LowerShiftLeftParts(SDValue Op,
                                                    SelectionDAG &DAG) const {
   assert(Op.getNumOperands() == 3 && "Not a double-shift!");
SDValue AArch64TargetLowering::LowerShiftLeftParts(SDValue Op,
 
   // AArch64 shifts of larger than register sizes are wrapped rather than
   // clamped, so we can't just emit "lo << a" if a is too big.
   SDValue LoForBigShift = DAG.getConstant(0, dl, VT);
   SDValue LoForNormalShift = DAG.getNode(ISD::SHL, dl, VT, ShOpLo, ShAmt);
   SDValue Lo = DAG.getNode(AArch64ISD::CSEL, dl, VT, LoForBigShift,
                            LoForNormalShift, CCVal, Cmp);
 
   SDValue Ops[2] = { Lo, Hi };
   return DAG.getMergeValues(Ops, dl);
 }
 
 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,
   // The immediate value will be created with either MOVZ, MOVN, or ORR.
   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);
     unsigned Limit = (OptForSize ? 1 : (Subtarget->hasFuseLiterals() ? 5 : 2));
     IsLegal = Insn.size() <= Limit;
   }
 
   LLVM_DEBUG(dbgs() << (IsLegal ? "Legal " : "Illegal ") << VT.getEVTString()
                     << " 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)) {
     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.
       ExtraSteps = VT.getScalarType() == MVT::f64 ? 3 : 2;
if (SDValue Estimate = getEstimate(Subtarget, AArch64ISD::FRSQRTE, Operand,
       EVT VT = Operand.getValueType();
 
       SDNodeFlags Flags;
       Flags.setAllowReassociation(true);
 
       // 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) {
         EVT CCVT = getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(),
                                       VT);
         SDValue FPZero = DAG.getConstantFP(0.0, DL, VT);
         SDValue Eq = DAG.getSetCC(DL, CCVT, Operand, FPZero, ISD::SETEQ);
 
         Estimate = DAG.getNode(ISD::FMUL, DL, VT, Operand, Estimate, Flags);
         // Correct the result if the operand is 0.0.
         Estimate = DAG.getNode(VT.isVector() ? ISD::VSELECT : ISD::SELECT, DL,
                                VT, Eq, Operand, Estimate);
if (SDValue Estimate = getEstimate(Subtarget, AArch64ISD::FRECPE, Operand,
       EVT VT = Operand.getValueType();
 
       SDNodeFlags Flags;
       Flags.setAllowReassociation(true);
 
       // 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;
     }
 
const char *AArch64TargetLowering::LowerXConstraint(EVT ConstraintVT) const {
   // 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 PredicateConstraint {
   Upl,
   Upa,
   Invalid
 };
 
 static PredicateConstraint parsePredicateConstraint(StringRef Constraint) {
   PredicateConstraint P = PredicateConstraint::Invalid;
   if (Constraint == "Upa")
if (Constraint.size() == 1) {
     case 'Z':
       return C_Immediate;
     case 'z':
     case 'S': // A symbolic address
       return C_Other;
     }
   } else if (parsePredicateConstraint(Constraint) !=
              PredicateConstraint::Invalid)
       return C_RegisterClass;
   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(
case 'y':
       break;
     }
   } else {
     PredicateConstraint PC = parsePredicateConstraint(Constraint);
     if (PC != PredicateConstraint::Invalid) {
       assert(VT.isScalableVector());
       bool restricted = (PC == PredicateConstraint::Upl);
       return restricted ? std::make_pair(0U, &AArch64::PPR_3bRegClass)
                           : std::make_pair(0U, &AArch64::PPRRegClass);
     }
   }
   if (StringRef("{cc}").equals_lower(Constraint))
     return std::make_pair(unsigned(AArch64::NZCV), &AArch64::CCRRegClass);
 
   // Use the default implementation in TargetLowering to convert the register
   // constraint into a member of a register class.
   std::pair<unsigned, const TargetRegisterClass *> Res;
struct ShuffleSourceInfo {
     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())
if (Src.MinElt >= NumSrcElts) {
                       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 (unsigned i = 0; i < VT.getVectorNumElements(); ++i) {
 
     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 = DAG.getNode(ISD::BITCAST, dl, VT, Shuffle);
 
   LLVM_DEBUG(dbgs() << "Reshuffle, creating node: "; Shuffle.dump();
              dbgs() << "Reshuffle, creating node: "; V.dump(););
 
   return V;
 }
static bool isEXTMask(ArrayRef<int> M, EVT VT, bool &ReverseEXT,
   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);
   // The following shuffle indices must be the successive elements after the
   // first real element.
   const int *FirstWrongElt = std::find_if(FirstRealElt + 1, M.end(),
       [&](int Elt) {return Elt != ExpectedElt++ && Elt != -1;});
   if (FirstWrongElt != M.end())
     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, ...>.
static bool isTRN_v_undef_Mask(ArrayRef<int> M, EVT VT, unsigned &WhichResult) {
   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 (IndexLen == 8) {
       // 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,
                                     makeArrayRef(TBLMask.data(), IndexLen)));
     }
   }
   return DAG.getNode(ISD::BITCAST, DL, Op.getValueType(), Shuffle);
 }
 
 static unsigned getDUPLANEOp(EVT EltType) {
   if (EltType == MVT::i8)
if (auto *ConstVal = dyn_cast<ConstantSDNode>(SplatVal)) {
         return getPTrue(DAG, dl, VT, AArch64SVEPredPattern::all);
       // TODO: Add special case for constant false
     }
     // The general case of i1.  There isn't any natural way to do this,
     // so we use some trickery with whilelo.
     SplatVal = DAG.getAnyExtOrTrunc(SplatVal, 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);
     return DAG.getNode(ISD::INTRINSIC_WO_CHAIN, dl, VT, ID,
                        DAG.getConstant(0, dl, MVT::i64), SplatVal);
   }
   case MVT::i8:
   case MVT::i16:
   case MVT::i32:
     SplatVal = DAG.getAnyExtOrTrunc(SplatVal, dl, MVT::i32);
SDValue AArch64TargetLowering::LowerDUPQLane(SDValue Op,
   //   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 Zero = DAG.getConstant(0, DL, MVT::i64);
   SDValue SV = DAG.getNode(AArch64ISD::INDEX_VECTOR,
                            DL, MVT::nxv2i64, Zero, One);
   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) {
   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::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, *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) {
   if (Bits.getHiBits(64) == Bits.getLoBits(64)) {
     uint64_t Value = Bits.zextOrTrunc(64).getZExtValue();
     EVT VT = Op.getValueType();
     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, *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();
 }
 
SDValue AArch64TargetLowering::LowerVectorOR(SDValue Op,
   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)) {
       APInt LowBits(EltTy.getSizeInBits(),
                     CstLane->getZExtValue());
       Lane = DAG.getConstant(LowBits.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");
     }
if (!isConstant) {
 
       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 AArch64TargetLowering::LowerINSERT_VECTOR_ELT(SDValue Op,
   assert(Op.getOpcode() == ISD::INSERT_VECTOR_ELT && "Unknown opcode!");
 
   // Check for non-constant or out of range lane.
   EVT VT = Op.getOperand(0).getValueType();
   ConstantSDNode *CI = dyn_cast<ConstantSDNode>(Op.getOperand(2));
   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 &&
AArch64TargetLowering::LowerEXTRACT_VECTOR_ELT(SDValue Op,
   assert(Op.getOpcode() == ISD::EXTRACT_VECTOR_ELT && "Unknown opcode!");
 
   // Check for non-constant or out of range lane.
   EVT VT = Op.getOperand(0).getValueType();
   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 &&

 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
SDValue AArch64TargetLowering::LowerVectorSRA_SRL_SHL(SDValue Op,
 
   case ISD::SHL:
     if (VT.isScalableVector() || useSVEForFixedLengthVectorVT(VT))
       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() || useSVEForFixedLengthVectorVT(VT)) {
       unsigned Opc = Op.getOpcode() == ISD::SRA ? AArch64ISD::SRA_PRED
                                                 : AArch64ISD::SRL_PRED;
case ISD::SRL:
 
     // 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(AArch64ISD::NEG, 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;
   }
 
   return SDValue();
 }
if (LHS.getValueType().getVectorElementType().isInteger()) {
     assert(LHS.getValueType() == RHS.getValueType());
     AArch64CC::CondCode AArch64CC = changeIntCCToAArch64CC(CC);
     SDValue Cmp =
         EmitVectorComparison(LHS, RHS, AArch64CC, false, CmpVT, dl, DAG);
     return DAG.getSExtOrTrunc(Cmp, dl, Op.getValueType());
   }
 
   const bool FullFP16 =
     static_cast<const AArch64Subtarget &>(DAG.getSubtarget()).hasFullFP16();
 
   // Make v4f16 (only) fcmp operations utilise vector instructions
   // v8f16 support will be a litle more complicated
   if (!FullFP16 && LHS.getValueType().getVectorElementType() == MVT::f16) {
     if (LHS.getValueType().getVectorNumElements() == 4) {
       LHS = DAG.getNode(ISD::FP_EXTEND, dl, MVT::v4f32, LHS);
       RHS = DAG.getNode(ISD::FP_EXTEND, dl, MVT::v4f32, RHS);
       SDValue NewSetcc = DAG.getSetCC(dl, MVT::v4i16, LHS, RHS, CC);
       DAG.ReplaceAllUsesWith(Op, NewSetcc);
       CmpVT = MVT::v4i32;
     } else
       return SDValue();
   }
 
   assert((!FullFP16 && LHS.getValueType().getVectorElementType() != MVT::f16) ||
           LHS.getValueType().getVectorElementType() != MVT::f128);
 
   // 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;
   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());
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);
   case ISD::VECREDUCE_FMAX: {
     return DAG.getNode(
         ISD::INTRINSIC_WO_CHAIN, dl, Op.getValueType(),
         DAG.getConstant(Intrinsic::aarch64_neon_fmaxnmv, dl, MVT::i32),
         Src);
   }
   case ISD::VECREDUCE_FMIN: {
     return DAG.getNode(
         ISD::INTRINSIC_WO_CHAIN, dl, Op.getValueType(),
         DAG.getConstant(Intrinsic::aarch64_neon_fminnmv, dl, MVT::i32),
         Src);
   }
   default:
     llvm_unreachable("Unhandled reduction");
   }
 }
 
 SDValue AArch64TargetLowering::LowerATOMIC_LOAD_SUB(SDValue Op,
bool AArch64TargetLowering::shouldReduceLoadWidth(SDNode *Load,
   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) {
     // 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;
 }
 
 // Truncations from 64-bit GPR to 32-bit GPR is free.
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 &&
       !(User->getOpcode() == Instruction::FSub ||
         User->getOpcode() == Instruction::FAdd))
     return true;
 
   const TargetOptions &Options = getTargetMachine().Options;
   const Function *F = I->getFunction();
   const DataLayout &DL = F->getParent()->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 {
for (const Use &U : Ext->uses()) {
     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->getModule()->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).
       uint64_t ShiftAmt =
         countTrailingZeros(DL.getTypeStoreSizeInBits(IdxTy).getFixedSize()) - 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.

 bool AArch64TargetLowering::isLegalAddImmediate(int64_t Immed) const {
   if (Immed == std::numeric_limits<int64_t>::min()) {
     LLVM_DEBUG(dbgs() << "Illegal add imm " << Immed
                       << ": avoid UB for INT64_MIN\n");
     return false;
   }
   // Same encoding for add/sub, just flip the sign.
   Immed = std::abs(Immed);
   bool IsLegal = ((Immed >> 12) == 0 ||
                   ((Immed & 0xfff) == 0 && Immed >> 24 == 0));
   LLVM_DEBUG(dbgs() << "Is " << Immed
                     << " legal add imm: " << (IsLegal ? "yes" : "no") << "\n");
   return IsLegal;
 }
 
 // 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 &AM, 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.
bool AArch64TargetLowering::isFMAFasterThanFMulAndFAdd(const Function &F,
   }
 }
 
 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;
 }
 
 bool
 AArch64TargetLowering::isDesirableToCommuteWithShift(const SDNode *N,
                                                      CombineLevel Level) const {
   N = N->getOperand(0).getNode();
   EVT VT = N->getValueType(0);
     // If N is unsigned bit extraction: ((x >> C) & mask), then do not combine
     // it with shift to let it be lowered to UBFX.
   if (N->getOpcode() == ISD::AND && (VT == MVT::i32 || VT == MVT::i64) &&
       isa<ConstantSDNode>(N->getOperand(1))) {
     uint64_t TruncMask = N->getConstantOperandVal(1);
     if (isMask_64(TruncMask) &&
       N->getOperand(0).getOpcode() == ISD::SRL &&
       isa<ConstantSDNode>(N->getOperand(0)->getOperand(1)))
       return false;
   }
   return true;
 }
 
 bool AArch64TargetLowering::shouldConvertConstantLoadToIntImm(const APInt &Imm,
auto DotIntrisic = (ExtOpcode == ISD::ZERO_EXTEND)
                          ? Intrinsic::aarch64_neon_udot
                          : Intrinsic::aarch64_neon_sdot;
   SDValue Dot = DAG.getNode(ISD::INTRINSIC_WO_CHAIN, DL, Zeros.getValueType(),
                             DAG.getConstant(DotIntrisic, DL, MVT::i32), Zeros,
                             Ones, Op0.getOperand(0));
   return DAG.getNode(ISD::VECREDUCE_ADD, DL, N->getValueType(0), Dot);
 }
 
+// Given a ABS node, detect the following pattern:
+// (ABS (SUB (EXTEND a), (EXTEND b))).
+// Generates UABD/SABD instruction.
+static SDValue performABSCombine(SDNode *N, SelectionDAG &DAG,
+                                 TargetLowering::DAGCombinerInfo &DCI,
+                                 const AArch64Subtarget *Subtarget) {
+  SDValue AbsOp1 = N->getOperand(0);
+  SDValue Op0, Op1;
+
+  if (AbsOp1.getOpcode() != ISD::SUB)
+    return SDValue();
+
+  Op0 = AbsOp1.getOperand(0);
+  Op1 = AbsOp1.getOperand(1);
+
+  unsigned Opc0 = Op0.getOpcode();
+  // Check if the operands of the sub are (zero|sign)-extended.
+  if (Opc0 != Op1.getOpcode() ||
+      (Opc0 != ISD::ZERO_EXTEND && Opc0 != ISD::SIGN_EXTEND))
+    return SDValue();
+
+  EVT VectorT1 = Op0.getOperand(0).getValueType();
+  EVT VectorT2 = Op1.getOperand(0).getValueType();
+  // Check if vectors are of same type and valid size.
+  uint64_t size = VectorT1.getFixedSizeInBits();
+  if (VectorT1 != VectorT2 || (size != 64 && size != 128))
+    return SDValue();
+
+  // Check if vector element types are valid.
+  EVT VT1 = VectorT1.getVectorElementType();
+  if (VT1 != MVT::i8 && VT1 != MVT::i16 && VT1 != MVT::i32)
+    return SDValue();
+
+  Op0 = Op0.getOperand(0);
+  Op1 = Op1.getOperand(0);
+  unsigned ABDOpcode =
+      (Opc0 == ISD::SIGN_EXTEND) ? AArch64ISD::SABD : AArch64ISD::UABD;
+  SDValue ABD =
+      DAG.getNode(ABDOpcode, SDLoc(N), Op0->getValueType(0), Op0, Op1);
+  return DAG.getNode(ISD::ZERO_EXTEND, SDLoc(N), N->getValueType(0), ABD);
+}
+
 static SDValue performXorCombine(SDNode *N, SelectionDAG &DAG,
                                  TargetLowering::DAGCombinerInfo &DCI,
                                  const AArch64Subtarget *Subtarget) {
   if (DCI.isBeforeLegalizeOps())
     return SDValue();
 
   if (SDValue Cmp = foldVectorXorShiftIntoCmp(N, DAG, Subtarget))
     return Cmp;
 
   return performIntegerAbsCombine(N, DAG);
 }
 
 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
 
   // fold (sdiv X, pow2)
   EVT VT = N->getValueType(0);
   if ((VT != MVT::i32 && VT != MVT::i64) ||
       !(Divisor.isPowerOf2() || (-Divisor).isPowerOf2()))
     return SDValue();
 
   SDLoc DL(N);
AArch64TargetLowering::BuildSDIVPow2(SDNode *N, const APInt &Divisor,
   if (Divisor.isNonNegative())
     return SRA;
 
   Created.push_back(SRA.getNode());
   return DAG.getNode(ISD::SUB, DL, VT, DAG.getConstant(0, DL, VT), SRA);
 }
 
 static bool IsSVECntIntrinsic(SDValue S) {
   switch(getIntrinsicID(S.getNode())) {
   default:
     break;
   case Intrinsic::aarch64_sve_cntb:
   case Intrinsic::aarch64_sve_cnth:
   case Intrinsic::aarch64_sve_cntw:
   case Intrinsic::aarch64_sve_cntd:
     return true;
   }
if (!isa<ConstantSDNode>(N->getOperand(1)))
     return SDValue();
 
   SDValue N0 = N->getOperand(0);
   ConstantSDNode *C = cast<ConstantSDNode>(N->getOperand(1));
   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 as shift+add/sub. For now, this is true unilaterally. If
   // future CPUs have a cheaper MADD instruction, this may need to be
static SDValue performIntToFpCombine(SDNode *N, SelectionDAG &DAG,
   // 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->hasNEON() && 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->getAlignment(),
                                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
if (FloatBits != 32 && FloatBits != 64)
     return SDValue();
 
   // Avoid conversions where iN is larger than the float (e.g., i64 -> float).
   if (IntBits > FloatBits)
     return SDValue();
 
   BitVector UndefElements;
   BuildVectorSDNode *BV = cast<BuildVectorSDNode>(ConstVec);
   int32_t C = BV->getConstantFPSplatPow2ToLog2Int(&UndefElements, FloatBits + 1);
   if (C == -1 || C == 0 || C > FloatBits)
     return SDValue();
 
   MVT ResTy;
   unsigned NumLanes = Op.getValueType().getVectorNumElements();
   switch (NumLanes) {
   default:
     return SDValue();
if (LHSFromHi) {
     std::swap(ShiftLHS, ShiftRHS);
   }
 
   return DAG.getNode(AArch64ISD::EXTR, DL, VT, LHS, RHS,
                      DAG.getConstant(ShiftRHS, DL, MVT::i64));
 }
 
 static SDValue tryCombineToBSL(SDNode *N,
                                 TargetLowering::DAGCombinerInfo &DCI) {
   EVT VT = N->getValueType(0);
   SelectionDAG &DAG = DCI.DAG;
   SDLoc DL(N);
 
   if (!VT.isVector())
     return SDValue();
 
   SDValue N0 = N->getOperand(0);
if ((ExtVal == 0xFF && EltTy == MVT::i8) ||
         (ExtVal == 0xFFFFFFFF && EltTy == MVT::i32))
       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(AArch64ISD::DUP, 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);
   }
 
   SDValue Mask = N->getOperand(1);
 
   if (!Src.hasOneUse())
if (!TValue->isOne()) {
     std::swap(TValue, FValue);
     SetCCInfo.Info.AArch64.CC =
         AArch64CC::getInvertedCondCode(SetCCInfo.Info.AArch64.CC);
   }
   return TValue->isOne() && FValue->isNullValue();
 }
 
 // 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.

 //                     (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(1);
-  SDValue RHS = N->getOperand(2);
+  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();
   }
 
+  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();
 
static SDValue tryCombineCRC32(unsigned Mask, SDNode *N, SelectionDAG &DAG) {
   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 = Op1.getValueType();
 
   if ((ScalarTy == MVT::i8) || (ScalarTy == MVT::i16)) {
     Op1 = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, Op1);
     Op2 = DAG.getNode(ISD::ANY_EXTEND, DL, MVT::i32, Op2);
   }
 
   return DAG.getNode(AArch64ISD::INDEX_VECTOR, DL, N->getValueType(0),
                      Op1, Op2);
 }
 
 static SDValue LowerSVEIntrinsicDUP(SDNode *N, SelectionDAG &DAG) {
   SDLoc dl(N);
   SDValue Scalar = N->getOperand(3);
   EVT ScalarTy = Scalar.getValueType();
 
case Intrinsic::aarch64_sve_cmple_wide: {
           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;

 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).getOpcode() == ISD::INTRINSIC_WO_CHAIN) {
+      (N->getOperand(0).getOpcode() == AArch64ISD::UABD ||
+       N->getOperand(0).getOpcode() == AArch64ISD::SABD)) {
     SDNode *ABDNode = N->getOperand(0).getNode();
-    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();
+    SDValue NewABD =
+        tryCombineLongOpWithDup(Intrinsic::not_intrinsic, ABDNode, DCI, DAG);
 
-      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.
   //
   // Type legalization will split an extend of a small, legal, type to a larger
   // illegal type by first splitting the destination type, often creating
   // illegal source types, which then get legalized in isel-confusing ways,
   // leading to really terrible codegen. E.g.,
static SDValue performExtendCombine(SDNode *N,
   EVT LoVT, HiVT;
   SDValue Lo, Hi;
   LoVT = HiVT = ResVT.getHalfNumVectorElementsVT(*DAG.getContext());
 
   EVT InNVT = EVT::getVectorVT(*DAG.getContext(), SrcVT.getVectorElementType(),
                                LoVT.getVectorElementCount());
   Lo = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, InNVT, Src,
                    DAG.getConstant(0, DL, MVT::i64));
   Hi = DAG.getNode(ISD::EXTRACT_SUBVECTOR, DL, InNVT, Src,
                    DAG.getConstant(InNVT.getVectorMinNumElements(), DL, MVT::i64));
   Lo = DAG.getNode(N->getOpcode(), DL, LoVT, Lo);
   Hi = DAG.getNode(N->getOpcode(), DL, HiVT, Hi);
 
   // Now combine the parts back together so we still have a single result
   // like the combiner expects.
   return DAG.getNode(ISD::CONCAT_VECTORS, DL, ResVT, Lo, Hi);
 }
 
 static SDValue splitStoreSplat(SelectionDAG &DAG, StoreSDNode &St,
                                SDValue SplatVal, unsigned NumVecElts) {
   assert(!St.isTruncatingStore() && "cannot split truncating vector store");
   unsigned OrigAlignment = St.getAlignment();
   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);
static SDValue performLD1Combine(SDNode *N, SelectionDAG &DAG, unsigned Opc) {
   if (VT.getSizeInBits().getKnownMinSize() > 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();
 
   if (VT == MVT::nxv8bf16 &&
       !static_cast<const AArch64Subtarget &>(DAG.getSubtarget()).hasBF16())
     return SDValue();
 
   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>
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);
 }
for (unsigned I = 0; I < NumVecElts; ++I) {
     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) {
 
if (UI.getUse().getResNo() == 1) // Ignore uses of the chain result.
       continue;
     if (*UI != N)
       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 (SDNode::use_iterator UI = Addr.getNode()->use_begin(), UE =
        Addr.getNode()->use_end(); UI != UE; ++UI) {
     SDNode *User = *UI;
     if (User->getOpcode() != ISD::ADD
         || UI.getUse().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;
for (SDNode::use_iterator UI = Addr.getNode()->use_begin(), UE =
     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
static SDValue performSTORECombine(SDNode *N,
 
   if (Subtarget->supportsAddressTopByteIgnored() &&
       performTBISimplification(N->getOperand(2), DCI, DAG))
     return SDValue(N, 0);
 
   return SDValue();
 }
 
 
 /// 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 (SDNode::use_iterator UI = Addr.getNode()->use_begin(),
        UE = Addr.getNode()->use_end(); UI != UE; ++UI) {
     SDNode *User = *UI;
     if (User->getOpcode() != ISD::ADD ||
         UI.getUse().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 = cast<ConstantSDNode>(N->getOperand(1))->getZExtValue();
     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;
for (SDNode::use_iterator UI = Addr.getNode()->use_begin(),
     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(makeArrayRef(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: {
static bool isEquivalentMaskless(unsigned CC, unsigned width,
   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 &&
case AArch64CC::NV:
     return true;
   case AArch64CC::Invalid:
     break;
   }
 
   return false;
 }
 
 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();
 
SDValue performCONDCombine(SDNode *N,
   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);
 }
static SDValue performVSelectCombine(SDNode *N, SelectionDAG &DAG) {
   EVT CmpVT = N0.getOperand(0).getValueType();
   // 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);
   SDValue 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
static SDValue performSelectCombine(SDNode *N,
 
   // 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));
 }
 
 /// Get rid of unnecessary NVCASTs (that don't change the type).
 static SDValue performNVCASTCombine(SDNode *N) {
   if (N->getValueType(0) == N->getOperand(0).getValueType())
performSignExtendInRegCombine(SDNode *N, TargetLowering::DAGCombinerInfo &DCI,
 
   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)

 
 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::ABS:
+    return performABSCombine(N, DAG, DCI, Subtarget);
   case ISD::ADD:
   case ISD::SUB:
     return performAddSubLongCombine(N, DCI, DAG);
   case ISD::XOR:
     return performXorCombine(N, DAG, DCI, Subtarget);
   case ISD::MUL:
     return performMulCombine(N, DAG, DCI, Subtarget);
   case ISD::SINT_TO_FP:
void AArch64TargetLowering::ReplaceExtractSubVectorResults(
   SDValue Half = DAG.getNode(Opcode, DL, ExtendedHalfVT, N->getOperand(0));
   Results.push_back(DAG.getNode(ISD::TRUNCATE, DL, VT, Half));
 }
 
 // Create an even/odd pair of X registers holding integer value V.
 static SDValue createGPRPairNode(SelectionDAG &DAG, SDValue V) {
   SDLoc dl(V.getNode());
   SDValue VLo = DAG.getAnyExtOrTrunc(V, dl, MVT::i64);
   SDValue VHi = DAG.getAnyExtOrTrunc(
       DAG.getNode(ISD::SRL, dl, MVT::i128, V, DAG.getConstant(64, dl, MVT::i64)),
       dl, 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");
 
   if (Subtarget->hasLSE()) {
     // 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
     };
 
     MachineMemOperand *MemOp = cast<MemSDNode>(N)->getMemOperand();
 
     unsigned Opcode;
     switch (MemOp->getOrdering()) {
     case AtomicOrdering::Monotonic:
case ISD::INTRINSIC_WO_CHAIN: {
     ConstantSDNode *CN = cast<ConstantSDNode>(N->getOperand(0));
     Intrinsic::ID IntID = static_cast<Intrinsic::ID>(CN->getZExtValue());
     switch (IntID) {
     default:
       return;
     case Intrinsic::aarch64_sve_clasta_n: {
       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: {
       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: {
       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: {
       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;
     }
     }
   }
   }
 }

 
 // For the real atomic operations, we have ldxr/stxr up to 128 bits,
 TargetLowering::AtomicExpansionKind
 AArch64TargetLowering::shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const {
   if (AI->isFloatingPointOperation())
     return AtomicExpansionKind::CmpXChg;
 
   unsigned Size = AI->getType()->getPrimitiveSizeInBits();
   if (Size > 128) return AtomicExpansionKind::None;
   // Nand not supported in LSE.
   if (AI->getOperation() == AtomicRMWInst::Nand) return AtomicExpansionKind::LLSC;
   // Leave 128 bits to LLSC.
   return (Subtarget->hasLSE() && Size < 128) ? AtomicExpansionKind::None : AtomicExpansionKind::LLSC;
 }
 
 TargetLowering::AtomicExpansionKind
 AArch64TargetLowering::shouldExpandAtomicCmpXchgInIR(
     AtomicCmpXchgInst *AI) const {
   // If subtarget has LSE, leave cmpxchg intact for codegen.
   if (Subtarget->hasLSE())
     return AtomicExpansionKind::None;
if (ValTy->getPrimitiveSizeInBits() == 128) {
     Value *Lo = Builder.CreateExtractValue(LoHi, 0, "lo");
     Value *Hi = Builder.CreateExtractValue(LoHi, 1, "hi");
     Lo = Builder.CreateZExt(Lo, ValTy, "lo64");
     Hi = Builder.CreateZExt(Hi, ValTy, "hi64");
     return Builder.CreateOr(
         Lo, Builder.CreateShl(Hi, ConstantInt::get(ValTy, 64)), "val64");
   }
 
   Type *Tys[] = { Addr->getType() };
   Intrinsic::ID Int =
       IsAcquire ? Intrinsic::aarch64_ldaxr : Intrinsic::aarch64_ldxr;
   Function *Ldxr = Intrinsic::getDeclaration(M, Int, Tys);
 
   Type *EltTy = cast<PointerType>(Addr->getType())->getElementType();
 
   const DataLayout &DL = M->getDataLayout();
   IntegerType *IntEltTy = Builder.getIntNTy(DL.getTypeSizeInBits(EltTy));
if (Val->getType()->getPrimitiveSizeInBits() == 128) {
     Value *Lo = Builder.CreateTrunc(Val, Int64Ty, "lo");
     Value *Hi = Builder.CreateTrunc(Builder.CreateLShr(Val, 64), Int64Ty, "hi");
     Addr = Builder.CreateBitCast(Addr, Type::getInt8PtrTy(M->getContext()));
     return Builder.CreateCall(Stxr, {Lo, Hi, Addr});
   }
 
   Intrinsic::ID Int =
       IsRelease ? Intrinsic::aarch64_stlxr : Intrinsic::aarch64_stxr;
   Type *Tys[] = { Addr->getType() };
   Function *Stxr = Intrinsic::getDeclaration(M, Int, Tys);
 
   const DataLayout &DL = M->getDataLayout();
   IntegerType *IntValTy = Builder.getIntNTy(DL.getTypeSizeInBits(Val->getType()));
   Val = Builder.CreateBitCast(Val, IntValTy);
 
   return Builder.CreateCall(Stxr,
                             {Builder.CreateZExtOrBitCast(
                                  Val, Stxr->getFunctionType()->getParamType(0)),
                              Addr});
 }
 

 
 Function *AArch64TargetLowering::getSSPStackGuardCheck(const Module &M) const {
   // MSVC CRT has a function to validate security cookie.
   if (Subtarget->getTargetTriple().isWindowsMSVCEnvironment())
     return M.getFunction("__security_check_cookie");
   return TargetLowering::getSSPStackGuardCheck(M);
 }
 
 Value *AArch64TargetLowering::getSafeStackPointerLocation(IRBuilder<> &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);
 }
 
 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,
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)

 }
 
 void AArch64TargetLowering::finalizeLowering(MachineFunction &MF) const {
   MF.getFrameInfo().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 {
   switch (MI.getOpcode()) {
   case TargetOpcode::G_GLOBAL_VALUE: {
     // On Darwin, TLS global vars get selected into function calls, which
SDValue AArch64TargetLowering::LowerFixedLengthVectorIntDivideToSVE(
   SDValue Op0Hi = DAG.getNode(UnpkHi, dl, ScalableWidenedVT, Op0);
   SDValue Op1Hi = DAG.getNode(UnpkHi, dl, ScalableWidenedVT, Op1);
 
   // Convert back to fixed vectors so the DIV can be further lowered.
   Op0Lo = convertFromScalableVector(DAG, FixedWidenedVT, Op0Lo);
   Op1Lo = convertFromScalableVector(DAG, FixedWidenedVT, Op1Lo);
   Op0Hi = convertFromScalableVector(DAG, FixedWidenedVT, Op0Hi);
   Op1Hi = convertFromScalableVector(DAG, FixedWidenedVT, Op1Hi);
   SDValue ResultLo = DAG.getNode(Op.getOpcode(), dl, FixedWidenedVT,
                                  Op0Lo, Op1Lo);
   SDValue ResultHi = DAG.getNode(Op.getOpcode(), dl, FixedWidenedVT,
                                  Op0Hi, Op1Hi);
 
   // Convert again to scalable vectors to truncate.
   ResultLo = convertToScalableVector(DAG, ScalableWidenedVT, ResultLo);
   ResultHi = convertToScalableVector(DAG, ScalableWidenedVT, ResultHi);
   SDValue ScalableResult = DAG.getNode(AArch64ISD::UZP1, dl, ContainerVT,
                                        ResultLo, ResultHi);
 
   return convertFromScalableVector(DAG, VT, ScalableResult);
 }
 
 SDValue AArch64TargetLowering::LowerFixedLengthVectorIntExtendToSVE(
     SDValue Op, SelectionDAG &DAG) const {
   EVT VT = Op.getValueType();
if (useSVEForFixedLengthVectorVT(VT, OverrideNEON)) {
     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));
 
assert(useSVEForFixedLengthVectorVT(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::LowerFixedLengthReductionToSVE(unsigned Opcode,
     SDValue ScalarOp, SelectionDAG &DAG) const {
   SDLoc DL(ScalarOp);
   SDValue VecOp = ScalarOp.getOperand(0);
   EVT SrcVT = VecOp.getValueType();
 
   SDValue Pg = getPredicateForVector(DAG, DL, SrcVT);
   EVT ContainerVT = getContainerForFixedLengthVector(DAG, SrcVT);
   VecOp = convertToScalableVector(DAG, ContainerVT, VecOp);
 
   // UADDV always returns an i64 result.
   EVT ResVT = (Opcode == AArch64ISD::UADDV_PRED) ? MVT::i64 :
                                                    SrcVT.getVectorElementType();
 
   SDValue Rdx = DAG.getNode(Opcode, DL, getPackedSVEVectorVT(ResVT), 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);

lib/Target/AArch64/AArch64InstrInfo.td

 def AArch64smaxv    : SDNode<"AArch64ISD::SMAXV", SDT_AArch64UnaryVec>;
 def AArch64umaxv    : SDNode<"AArch64ISD::UMAXV", SDT_AArch64UnaryVec>;
 
 def AArch64srhadd   : SDNode<"AArch64ISD::SRHADD", SDT_AArch64binvec>;
 def AArch64urhadd   : SDNode<"AArch64ISD::URHADD", SDT_AArch64binvec>;
 def AArch64shadd   : SDNode<"AArch64ISD::SHADD", SDT_AArch64binvec>;
 def AArch64uhadd   : SDNode<"AArch64ISD::UHADD", SDT_AArch64binvec>;
 
+def AArch64uabd_n   : SDNode<"AArch64ISD::UABD", SDT_AArch64binvec>;
+def AArch64sabd_n   : SDNode<"AArch64ISD::SABD", SDT_AArch64binvec>;
+
+def AArch64uabd     : PatFrags<(ops node:$lhs, node:$rhs),
+                               [(AArch64uabd_n node:$lhs, node:$rhs),
+                                (int_aarch64_neon_uabd node:$lhs, node:$rhs)]>;
+def AArch64sabd     : PatFrags<(ops node:$lhs, node:$rhs),
+                               [(AArch64sabd_n node:$lhs, node:$rhs),
+                                (int_aarch64_neon_sabd node:$lhs, node:$rhs)]>;
+
 def SDT_AArch64SETTAG : SDTypeProfile<0, 2, [SDTCisPtrTy<0>, SDTCisPtrTy<1>]>;
 def AArch64stg : SDNode<"AArch64ISD::STG", SDT_AArch64SETTAG, [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;
 def AArch64stzg : SDNode<"AArch64ISD::STZG", SDT_AArch64SETTAG, [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;
 def AArch64st2g : SDNode<"AArch64ISD::ST2G", SDT_AArch64SETTAG, [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;
 def AArch64stz2g : SDNode<"AArch64ISD::STZ2G", SDT_AArch64SETTAG, [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;
 
 def SDT_AArch64unpk : SDTypeProfile<1, 1, [
     SDTCisInt<0>, SDTCisInt<1>, SDTCisOpSmallerThanOp<1, 0>

 defm FMOV : FPMoveImmediate<"fmov">;
 }
 
 //===----------------------------------------------------------------------===//
 // Advanced SIMD two vector instructions.
 //===----------------------------------------------------------------------===//
 
 defm UABDL   : SIMDLongThreeVectorBHSabdl<1, 0b0111, "uabdl",
-                                          int_aarch64_neon_uabd>;
+                                          AArch64uabd>;
 // Match UABDL in log2-shuffle patterns.
 def : Pat<(abs (v8i16 (sub (zext (v8i8 V64:$opA)),
                            (zext (v8i8 V64:$opB))))),
           (UABDLv8i8_v8i16 V64:$opA, V64:$opB)>;
 def : Pat<(xor (v8i16 (AArch64vashr v8i16:$src, (i32 15))),
                (v8i16 (add (sub (zext (v8i8 V64:$opA)),
                                 (zext (v8i8 V64:$opB))),
                            (AArch64vashr v8i16:$src, (i32 15))))),

 
 // MLA and MLS are generated in MachineCombine
 defm MLA      : SIMDThreeSameVectorBHSTied<0, 0b10010, "mla", null_frag>;
 defm MLS      : SIMDThreeSameVectorBHSTied<1, 0b10010, "mls", null_frag>;
 
 defm MUL      : SIMDThreeSameVectorBHS<0, 0b10011, "mul", mul>;
 defm PMUL     : SIMDThreeSameVectorB<1, 0b10011, "pmul", int_aarch64_neon_pmul>;
 defm SABA     : SIMDThreeSameVectorBHSTied<0, 0b01111, "saba",
-      TriOpFrag<(add node:$LHS, (int_aarch64_neon_sabd node:$MHS, node:$RHS))> >;
-defm SABD     : SIMDThreeSameVectorBHS<0,0b01110,"sabd", int_aarch64_neon_sabd>;
+      TriOpFrag<(add node:$LHS, (AArch64sabd node:$MHS, node:$RHS))> >;
+defm SABD     : SIMDThreeSameVectorBHS<0,0b01110,"sabd", AArch64sabd>;
 defm SHADD    : SIMDThreeSameVectorBHS<0,0b00000,"shadd", AArch64shadd>;
 defm SHSUB    : SIMDThreeSameVectorBHS<0,0b00100,"shsub", int_aarch64_neon_shsub>;
 defm SMAXP    : SIMDThreeSameVectorBHS<0,0b10100,"smaxp", int_aarch64_neon_smaxp>;
 defm SMAX     : SIMDThreeSameVectorBHS<0,0b01100,"smax", smax>;
 defm SMINP    : SIMDThreeSameVectorBHS<0,0b10101,"sminp", int_aarch64_neon_sminp>;
 defm SMIN     : SIMDThreeSameVectorBHS<0,0b01101,"smin", smin>;
 defm SQADD    : SIMDThreeSameVector<0,0b00001,"sqadd", int_aarch64_neon_sqadd>;
 defm SQDMULH  : SIMDThreeSameVectorHS<0,0b10110,"sqdmulh",int_aarch64_neon_sqdmulh>;
 defm SQRDMULH : SIMDThreeSameVectorHS<1,0b10110,"sqrdmulh",int_aarch64_neon_sqrdmulh>;
 defm SQRSHL   : SIMDThreeSameVector<0,0b01011,"sqrshl", int_aarch64_neon_sqrshl>;
 defm SQSHL    : SIMDThreeSameVector<0,0b01001,"sqshl", int_aarch64_neon_sqshl>;
 defm SQSUB    : SIMDThreeSameVector<0,0b00101,"sqsub", int_aarch64_neon_sqsub>;
 defm SRHADD   : SIMDThreeSameVectorBHS<0,0b00010,"srhadd", AArch64srhadd>;
 defm SRSHL    : SIMDThreeSameVector<0,0b01010,"srshl", int_aarch64_neon_srshl>;
 defm SSHL     : SIMDThreeSameVector<0,0b01000,"sshl", int_aarch64_neon_sshl>;
 defm SUB      : SIMDThreeSameVector<1,0b10000,"sub", sub>;
 defm UABA     : SIMDThreeSameVectorBHSTied<1, 0b01111, "uaba",
-      TriOpFrag<(add node:$LHS, (int_aarch64_neon_uabd node:$MHS, node:$RHS))> >;
-defm UABD     : SIMDThreeSameVectorBHS<1,0b01110,"uabd", int_aarch64_neon_uabd>;
+      TriOpFrag<(add node:$LHS, (AArch64uabd node:$MHS, node:$RHS))> >;
+defm UABD     : SIMDThreeSameVectorBHS<1,0b01110,"uabd", AArch64uabd>;
 defm UHADD    : SIMDThreeSameVectorBHS<1,0b00000,"uhadd", AArch64uhadd>;
 defm UHSUB    : SIMDThreeSameVectorBHS<1,0b00100,"uhsub", int_aarch64_neon_uhsub>;
 defm UMAXP    : SIMDThreeSameVectorBHS<1,0b10100,"umaxp", int_aarch64_neon_umaxp>;
 defm UMAX     : SIMDThreeSameVectorBHS<1,0b01100,"umax", umax>;
 defm UMINP    : SIMDThreeSameVectorBHS<1,0b10101,"uminp", int_aarch64_neon_uminp>;
 defm UMIN     : SIMDThreeSameVectorBHS<1,0b01101,"umin", umin>;
 defm UQADD    : SIMDThreeSameVector<1,0b00001,"uqadd", int_aarch64_neon_uqadd>;
 defm UQRSHL   : SIMDThreeSameVector<1,0b01011,"uqrshl", int_aarch64_neon_uqrshl>;

 //===----------------------------------------------------------------------===//
 
 defm ADDHN  : SIMDNarrowThreeVectorBHS<0,0b0100,"addhn", int_aarch64_neon_addhn>;
 defm SUBHN  : SIMDNarrowThreeVectorBHS<0,0b0110,"subhn", int_aarch64_neon_subhn>;
 defm RADDHN : SIMDNarrowThreeVectorBHS<1,0b0100,"raddhn",int_aarch64_neon_raddhn>;
 defm RSUBHN : SIMDNarrowThreeVectorBHS<1,0b0110,"rsubhn",int_aarch64_neon_rsubhn>;
 defm PMULL  : SIMDDifferentThreeVectorBD<0,0b1110,"pmull",int_aarch64_neon_pmull>;
 defm SABAL  : SIMDLongThreeVectorTiedBHSabal<0,0b0101,"sabal",
-                                             int_aarch64_neon_sabd>;
+                                             AArch64sabd>;
 defm SABDL   : SIMDLongThreeVectorBHSabdl<0, 0b0111, "sabdl",
-                                          int_aarch64_neon_sabd>;
+                                          AArch64sabd>;
 defm SADDL   : SIMDLongThreeVectorBHS<   0, 0b0000, "saddl",
             BinOpFrag<(add (sext node:$LHS), (sext node:$RHS))>>;
 defm SADDW   : SIMDWideThreeVectorBHS<   0, 0b0001, "saddw",
                  BinOpFrag<(add node:$LHS, (sext node:$RHS))>>;
 defm SMLAL   : SIMDLongThreeVectorTiedBHS<0, 0b1000, "smlal",
     TriOpFrag<(add node:$LHS, (int_aarch64_neon_smull node:$MHS, node:$RHS))>>;
 defm SMLSL   : SIMDLongThreeVectorTiedBHS<0, 0b1010, "smlsl",
     TriOpFrag<(sub node:$LHS, (int_aarch64_neon_smull node:$MHS, node:$RHS))>>;
 defm SMULL   : SIMDLongThreeVectorBHS<0, 0b1100, "smull", int_aarch64_neon_smull>;
 defm SQDMLAL : SIMDLongThreeVectorSQDMLXTiedHS<0, 0b1001, "sqdmlal",
                                                int_aarch64_neon_sqadd>;
 defm SQDMLSL : SIMDLongThreeVectorSQDMLXTiedHS<0, 0b1011, "sqdmlsl",
                                                int_aarch64_neon_sqsub>;
 defm SQDMULL : SIMDLongThreeVectorHS<0, 0b1101, "sqdmull",
                                      int_aarch64_neon_sqdmull>;
 defm SSUBL   : SIMDLongThreeVectorBHS<0, 0b0010, "ssubl",
                  BinOpFrag<(sub (sext node:$LHS), (sext node:$RHS))>>;
 defm SSUBW   : SIMDWideThreeVectorBHS<0, 0b0011, "ssubw",
                  BinOpFrag<(sub node:$LHS, (sext node:$RHS))>>;
 defm UABAL   : SIMDLongThreeVectorTiedBHSabal<1, 0b0101, "uabal",
-                                              int_aarch64_neon_uabd>;
+                                              AArch64uabd>;
 defm UADDL   : SIMDLongThreeVectorBHS<1, 0b0000, "uaddl",
                  BinOpFrag<(add (zext node:$LHS), (zext node:$RHS))>>;
 defm UADDW   : SIMDWideThreeVectorBHS<1, 0b0001, "uaddw",
                  BinOpFrag<(add node:$LHS, (zext node:$RHS))>>;
 defm UMLAL   : SIMDLongThreeVectorTiedBHS<1, 0b1000, "umlal",
     TriOpFrag<(add node:$LHS, (int_aarch64_neon_umull node:$MHS, node:$RHS))>>;
 defm UMLSL   : SIMDLongThreeVectorTiedBHS<1, 0b1010, "umlsl",
     TriOpFrag<(sub node:$LHS, (int_aarch64_neon_umull node:$MHS, node:$RHS))>>;

test/CodeGen/AArch64/arm64-vabs.ll

;CHECK: uabdl.2d
   %tmp1 = shufflevector <4 x i32> %load1, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
   %tmp2 = shufflevector <4 x i32> %load2, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
   %tmp3 = call <2 x i32> @llvm.aarch64.neon.uabd.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
   %tmp4 = zext <2 x i32> %tmp3 to <2 x i64>
   ret <2 x i64> %tmp4
 }
 
 declare i16 @llvm.vector.reduce.add.v16i16(<16 x i16>)
+declare i32 @llvm.experimental.vector.reduce.add.v16i32(<16 x i32>)
 
-define i16 @uabdl8h_rdx(<16 x i8>* %a, <16 x i8>* %b) {
-; CHECK-LABEL: uabdl8h_rdx
-; CHECK: uabdl2.8h
-; CHECK: uabdl.8h
+define i16 @uabd16b_rdx(<16 x i8>* %a, <16 x i8>* %b) {
+; CHECK-LABEL: uabd16b_rdx
+; CHECK: uabd.16b
   %aload = load <16 x i8>, <16 x i8>* %a, align 1
   %bload = load <16 x i8>, <16 x i8>* %b, align 1
   %aext = zext <16 x i8> %aload to <16 x i16>
   %bext = zext <16 x i8> %bload to <16 x i16>
   %abdiff = sub nsw <16 x i16> %aext, %bext
   %abcmp = icmp slt <16 x i16> %abdiff, zeroinitializer
   %ababs = sub nsw <16 x i16> zeroinitializer, %abdiff
   %absel = select <16 x i1> %abcmp, <16 x i16> %ababs, <16 x i16> %abdiff
   %reduced_v = call i16 @llvm.vector.reduce.add.v16i16(<16 x i16> %absel)
   ret i16 %reduced_v
 }
 
+define i32 @uabd16b_rdx_i32(<16 x i8> %a, <16 x i8> %b) {
+; CHECK-LABEL: uabd16b_rdx_i32
+; CHECK: uabd.16b
+  %aext = zext <16 x i8> %a to <16 x i32>
+  %bext = zext <16 x i8> %b to <16 x i32>
+  %abdiff = sub nsw <16 x i32> %aext, %bext
+  %abcmp = icmp slt <16 x i32> %abdiff, zeroinitializer
+  %ababs = sub nsw <16 x i32> zeroinitializer, %abdiff
+  %absel = select <16 x i1> %abcmp, <16 x i32> %ababs, <16 x i32> %abdiff
+  %reduced_v = call i32 @llvm.experimental.vector.reduce.add.v16i32(<16 x i32> %absel)
+  ret i32 %reduced_v
+}
+
+define i32 @sabd16b_rdx_i32(<16 x i8> %a, <16 x i8> %b) {
+; CHECK-LABEL: sabd16b_rdx_i32
+; CHECK: sabd.16b
+  %aext = sext <16 x i8> %a to <16 x i32>
+  %bext = sext <16 x i8> %b to <16 x i32>
+  %abdiff = sub nsw <16 x i32> %aext, %bext
+  %abcmp = icmp slt <16 x i32> %abdiff, zeroinitializer
+  %ababs = sub nsw <16 x i32> zeroinitializer, %abdiff
+  %absel = select <16 x i1> %abcmp, <16 x i32> %ababs, <16 x i32> %abdiff
+  %reduced_v = call i32 @llvm.experimental.vector.reduce.add.v16i32(<16 x i32> %absel)
+  ret i32 %reduced_v
+}
+
+
 declare i32 @llvm.vector.reduce.add.v8i32(<8 x i32>)
+declare i32 @llvm.experimental.vector.reduce.add.v4i32(<4 x i32>)
 
-define i32 @uabdl4s_rdx(<8 x i16>* %a, <8 x i16>* %b) {
-; CHECK-LABEL: uabdl4s_rdx
-; CHECK: uabdl2.4s
-; CHECK: uabdl.4s
+define i32 @uabd8h_rdx(<8 x i16>* %a, <8 x i16>* %b) {
+; CHECK-LABEL: uabd8h_rdx
+; CHECK: uabd.8h
   %aload = load <8 x i16>, <8 x i16>* %a, align 1
   %bload = load <8 x i16>, <8 x i16>* %b, align 1
   %aext = zext <8 x i16> %aload to <8 x i32>
   %bext = zext <8 x i16> %bload to <8 x i32>
   %abdiff = sub nsw <8 x i32> %aext, %bext
   %abcmp = icmp slt <8 x i32> %abdiff, zeroinitializer
   %ababs = sub nsw <8 x i32> zeroinitializer, %abdiff
   %absel = select <8 x i1> %abcmp, <8 x i32> %ababs, <8 x i32> %abdiff
   %reduced_v = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %absel)
   ret i32 %reduced_v
 }
 
+define i32 @sabd8h_rdx(<8 x i16> %a, <8 x i16> %b) {
+; CHECK-LABEL: sabd8h_rdx
+; CHECK: sabd.8h
+  %aext = sext <8 x i16> %a to <8 x i32>
+  %bext = sext <8 x i16> %b to <8 x i32>
+  %abdiff = sub nsw <8 x i32> %aext, %bext
+  %abcmp = icmp slt <8 x i32> %abdiff, zeroinitializer
+  %ababs = sub nsw <8 x i32> zeroinitializer, %abdiff
+  %absel = select <8 x i1> %abcmp, <8 x i32> %ababs, <8 x i32> %abdiff
+  %reduced_v = call i32 @llvm.vector.reduce.add.v8i32(<8 x i32> %absel)
+  ret i32 %reduced_v
+}
+
+define i32 @uabdl4s_rdx_i32(<4 x i16> %a, <4 x i16> %b) {
+; CHECK-LABEL: uabdl4s_rdx_i32
+; CHECK: uabdl.4s
+  %aext = zext <4 x i16> %a to <4 x i32>
+  %bext = zext <4 x i16> %b to <4 x i32>
+ %abdiff = sub nsw <4 x i32> %aext, %bext
+  %abcmp = icmp slt <4 x i32> %abdiff, zeroinitializer
+  %ababs = sub nsw <4 x i32> zeroinitializer, %abdiff
+  %absel = select <4 x i1> %abcmp, <4 x i32> %ababs, <4 x i32> %abdiff
+  %reduced_v = call i32 @llvm.experimental.vector.reduce.add.v4i32(<4 x i32> %absel)
+  ret i32 %reduced_v
+}
+
 declare i64 @llvm.vector.reduce.add.v4i64(<4 x i64>)
+declare i64 @llvm.experimental.vector.reduce.add.v2i64(<2 x i64>)
 
-define i64 @uabdl2d_rdx(<4 x i32>* %a, <4 x i32>* %b, i32 %h) {
-; CHECK: uabdl2d_rdx
-; CHECK: uabdl2.2d
-; CHECK: uabdl.2d
+define i64 @uabd4s_rdx(<4 x i32>* %a, <4 x i32>* %b, i32 %h) {
+; CHECK: uabd4s_rdx
+; CHECK: uabd.4s
   %aload = load <4 x i32>, <4 x i32>* %a, align 1
   %bload = load <4 x i32>, <4 x i32>* %b, align 1
   %aext = zext <4 x i32> %aload to <4 x i64>
   %bext = zext <4 x i32> %bload to <4 x i64>
   %abdiff = sub nsw <4 x i64> %aext, %bext
   %abcmp = icmp slt <4 x i64> %abdiff, zeroinitializer
   %ababs = sub nsw <4 x i64> zeroinitializer, %abdiff
   %absel = select <4 x i1> %abcmp, <4 x i64> %ababs, <4 x i64> %abdiff
   %reduced_v = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %absel)
   ret i64 %reduced_v
 }
 
+define i64 @sabd4s_rdx(<4 x i32> %a, <4 x i32> %b) {
+; CHECK: sabd4s_rdx
+; CHECK: sabd.4s
+  %aext = sext <4 x i32> %a to <4 x i64>
+  %bext = sext <4 x i32> %b to <4 x i64>
+  %abdiff = sub nsw <4 x i64> %aext, %bext
+  %abcmp = icmp slt <4 x i64> %abdiff, zeroinitializer
+  %ababs = sub nsw <4 x i64> zeroinitializer, %abdiff
+  %absel = select <4 x i1> %abcmp, <4 x i64> %ababs, <4 x i64> %abdiff
+  %reduced_v = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %absel)
+  ret i64 %reduced_v
+}
+
+define i64 @uabdl2d_rdx_i64(<2 x i32> %a, <2 x i32> %b) {
+; CHECK-LABEL: uabdl2d_rdx_i64
+; CHECK: uabdl.2d
+  %aext = zext <2 x i32> %a to <2 x i64>
+  %bext = zext <2 x i32> %b to <2 x i64>
+  %abdiff = sub nsw <2 x i64> %aext, %bext
+  %abcmp = icmp slt <2 x i64> %abdiff, zeroinitializer
+  %ababs = sub nsw <2 x i64> zeroinitializer, %abdiff
+  %absel = select <2 x i1> %abcmp, <2 x i64> %ababs, <2 x i64> %abdiff
+  %reduced_v = call i64 @llvm.experimental.vector.reduce.add.v2i64(<2 x i64> %absel)
+  ret i64 %reduced_v
+}
+
 define <2 x float> @fabd_2s(<2 x float>* %A, <2 x float>* %B) nounwind {
 ;CHECK-LABEL: fabd_2s:
 ;CHECK: fabd.2s
         %tmp1 = load <2 x float>, <2 x float>* %A
         %tmp2 = load <2 x float>, <2 x float>* %B
         %tmp3 = call <2 x float> @llvm.aarch64.neon.fabd.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
         ret <2 x float> %tmp3
 }