llvm/lib/Target/X86/X86ISelLowering.h

//===-- X86ISelLowering.h - X86 DAG Lowering Interface ----------*- C++ -*-===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file defines the interfaces that X86 uses to lower LLVM code into a
// selection DAG.
//
//===----------------------------------------------------------------------===//

#ifndef X86ISELLOWERING_H
#define X86ISELLOWERING_H

#include "X86Subtarget.h"
#include "llvm/CodeGen/CallingConvLower.h"
#include "llvm/CodeGen/SelectionDAG.h"
#include "llvm/Target/TargetLowering.h"
#include "llvm/Target/TargetOptions.h"

namespace llvm {
  class X86TargetMachine;

  namespace X86ISD {
    // X86 Specific DAG Nodes
    enum NodeType {
      // Start the numbering where the builtin ops leave off.
      FIRST_NUMBER = ISD::BUILTIN_OP_END,

      /// BSF - Bit scan forward.
      /// BSR - Bit scan reverse.
      BSF,
      BSR,

      /// SHLD, SHRD - Double shift instructions. These correspond to
      /// X86::SHLDxx and X86::SHRDxx instructions.
      SHLD,
      SHRD,

      /// FAND - Bitwise logical AND of floating point values. This corresponds
      /// to X86::ANDPS or X86::ANDPD.
      FAND,

      /// FOR - Bitwise logical OR of floating point values. This corresponds
      /// to X86::ORPS or X86::ORPD.
      FOR,

      /// FXOR - Bitwise logical XOR of floating point values. This corresponds
      /// to X86::XORPS or X86::XORPD.
      FXOR,

      /// FANDN - Bitwise logical ANDNOT of floating point values. This
      /// corresponds to X86::ANDNPS or X86::ANDNPD.
      FANDN,

      /// FSRL - Bitwise logical right shift of floating point values. These
      /// corresponds to X86::PSRLDQ.
      FSRL,

      /// CALL - These operations represent an abstract X86 call
      /// instruction, which includes a bunch of information.  In particular the
      /// operands of these node are:
      ///
      ///     #0 - The incoming token chain
      ///     #1 - The callee
      ///     #2 - The number of arg bytes the caller pushes on the stack.
      ///     #3 - The number of arg bytes the callee pops off the stack.
      ///     #4 - The value to pass in AL/AX/EAX (optional)
      ///     #5 - The value to pass in DL/DX/EDX (optional)
      ///
      /// The result values of these nodes are:
      ///
      ///     #0 - The outgoing token chain
      ///     #1 - The first register result value (optional)
      ///     #2 - The second register result value (optional)
      ///
      CALL,

      /// RDTSC_DAG - This operation implements the lowering for
      /// readcyclecounter
      RDTSC_DAG,

      /// X86 compare and logical compare instructions.
      CMP, COMI, UCOMI,

      /// X86 bit-test instructions.
      BT,

      /// X86 SetCC. Operand 0 is condition code, and operand 1 is the EFLAGS
      /// operand, usually produced by a CMP instruction.
      SETCC,

      /// X86 Select
      SELECT,

      // Same as SETCC except it's materialized with a sbb and the value is all
      // one's or all zero's.
      SETCC_CARRY,  // R = carry_bit ? ~0 : 0

      /// X86 FP SETCC, implemented with CMP{cc}SS/CMP{cc}SD.
      /// Operands are two FP values to compare; result is a mask of
      /// 0s or 1s.  Generally DTRT for C/C++ with NaNs.
      FSETCC,

      /// X86 MOVMSK{pd|ps}, extracts sign bits of two or four FP values,
      /// result in an integer GPR.  Needs masking for scalar result.
      FGETSIGNx86,

      /// X86 conditional moves. Operand 0 and operand 1 are the two values
      /// to select from. Operand 2 is the condition code, and operand 3 is the
      /// flag operand produced by a CMP or TEST instruction. It also writes a
      /// flag result.
      CMOV,

      /// X86 conditional branches. Operand 0 is the chain operand, operand 1
      /// is the block to branch if condition is true, operand 2 is the
      /// condition code, and operand 3 is the flag operand produced by a CMP
      /// or TEST instruction.
      BRCOND,

      /// Return with a flag operand. Operand 0 is the chain operand, operand
      /// 1 is the number of bytes of stack to pop.
      RET_FLAG,

      /// REP_STOS - Repeat fill, corresponds to X86::REP_STOSx.
      REP_STOS,

      /// REP_MOVS - Repeat move, corresponds to X86::REP_MOVSx.
      REP_MOVS,

      /// GlobalBaseReg - On Darwin, this node represents the result of the popl
      /// at function entry, used for PIC code.
      GlobalBaseReg,

      /// Wrapper - A wrapper node for TargetConstantPool,
      /// TargetExternalSymbol, and TargetGlobalAddress.
      Wrapper,

      /// WrapperRIP - Special wrapper used under X86-64 PIC mode for RIP
      /// relative displacements.
      WrapperRIP,

      /// MOVDQ2Q - Copies a 64-bit value from the low word of an XMM vector
      /// to an MMX vector.  If you think this is too close to the previous
      /// mnemonic, so do I; blame Intel.
      MOVDQ2Q,

      /// MMX_MOVD2W - Copies a 32-bit value from the low word of a MMX
      /// vector to a GPR.
      MMX_MOVD2W,

      /// PEXTRB - Extract an 8-bit value from a vector and zero extend it to
      /// i32, corresponds to X86::PEXTRB.
      PEXTRB,

      /// PEXTRW - Extract a 16-bit value from a vector and zero extend it to
      /// i32, corresponds to X86::PEXTRW.
      PEXTRW,

      /// INSERTPS - Insert any element of a 4 x float vector into any element
      /// of a destination 4 x floatvector.
      INSERTPS,

      /// PINSRB - Insert the lower 8-bits of a 32-bit value to a vector,
      /// corresponds to X86::PINSRB.
      PINSRB,

      /// PINSRW - Insert the lower 16-bits of a 32-bit value to a vector,
      /// corresponds to X86::PINSRW.
      PINSRW, MMX_PINSRW,

      /// PSHUFB - Shuffle 16 8-bit values within a vector.
      PSHUFB,

      /// ANDNP - Bitwise Logical AND NOT of Packed FP values.
      ANDNP,

      /// PSIGN - Copy integer sign.
      PSIGN,

      /// BLENDV - Blend where the selector is a register.
      BLENDV,

      /// BLENDI - Blend where the selector is an immediate.
      BLENDI,

      // SUBUS - Integer sub with unsigned saturation.
      SUBUS,

      /// HADD - Integer horizontal add.
      HADD,

      /// HSUB - Integer horizontal sub.
      HSUB,

      /// FHADD - Floating point horizontal add.
      FHADD,

      /// FHSUB - Floating point horizontal sub.
      FHSUB,

      /// UMAX, UMIN - Unsigned integer max and min.
      UMAX, UMIN,

      /// SMAX, SMIN - Signed integer max and min.
      SMAX, SMIN,

      /// FMAX, FMIN - Floating point max and min.
      ///
      FMAX, FMIN,

      /// FMAXC, FMINC - Commutative FMIN and FMAX.
      FMAXC, FMINC,

      /// FRSQRT, FRCP - Floating point reciprocal-sqrt and reciprocal
      /// approximation.  Note that these typically require refinement
      /// in order to obtain suitable precision.
      FRSQRT, FRCP,

      // TLSADDR - Thread Local Storage.
      TLSADDR,

      // TLSBASEADDR - Thread Local Storage. A call to get the start address
      // of the TLS block for the current module.
      TLSBASEADDR,

      // TLSCALL - Thread Local Storage.  When calling to an OS provided
      // thunk at the address from an earlier relocation.
      TLSCALL,

      // EH_RETURN - Exception Handling helpers.
      EH_RETURN,

      // EH_SJLJ_SETJMP - SjLj exception handling setjmp.
      EH_SJLJ_SETJMP,

      // EH_SJLJ_LONGJMP - SjLj exception handling longjmp.
      EH_SJLJ_LONGJMP,

      /// TC_RETURN - Tail call return. See X86TargetLowering::LowerCall for
      /// the list of operands.
      TC_RETURN,

      // VZEXT_MOVL - Vector move to low scalar and zero higher vector elements.
      VZEXT_MOVL,

      // VZEXT - Vector integer zero-extend.
      VZEXT,

      // VSEXT - Vector integer signed-extend.
      VSEXT,

      // VTRUNC - Vector integer truncate.
      VTRUNC,

      // VTRUNC - Vector integer truncate with mask.
      VTRUNCM,

      // VFPEXT - Vector FP extend.
      VFPEXT,

      // VFPROUND - Vector FP round.
      VFPROUND,

      // VSHL, VSRL - 128-bit vector logical left / right shift
      VSHLDQ, VSRLDQ,

      // VSHL, VSRL, VSRA - Vector shift elements
      VSHL, VSRL, VSRA,

      // VSHLI, VSRLI, VSRAI - Vector shift elements by immediate
      VSHLI, VSRLI, VSRAI,

      // CMPP - Vector packed double/float comparison.
      CMPP,

      // PCMP* - Vector integer comparisons.
      PCMPEQ, PCMPGT,
      // PCMP*M - Vector integer comparisons, the result is in a mask vector.
      PCMPEQM, PCMPGTM,

      /// CMPM, CMPMU - Vector comparison generating mask bits for fp and
      /// integer signed and unsigned data types.
      CMPM,
      CMPMU,

      // ADD, SUB, SMUL, etc. - Arithmetic operations with FLAGS results.
      ADD, SUB, ADC, SBB, SMUL,
      INC, DEC, OR, XOR, AND,

      BZHI,   // BZHI - Zero high bits
      BEXTR,  // BEXTR - Bit field extract

      UMUL, // LOW, HI, FLAGS = umul LHS, RHS

      // MUL_IMM - X86 specific multiply by immediate.
      MUL_IMM,

      // PTEST - Vector bitwise comparisons.
      PTEST,

      // TESTP - Vector packed fp sign bitwise comparisons.
      TESTP,

      // TESTM, TESTNM - Vector "test" in AVX-512, the result is in a mask vector.
      TESTM,
      TESTNM,

      // OR/AND test for masks
      KORTEST,

      // Several flavors of instructions with vector shuffle behaviors.
      PALIGNR,
      PSHUFD,
      PSHUFHW,
      PSHUFLW,
      SHUFP,
      MOVDDUP,
      MOVSHDUP,
      MOVSLDUP,
      MOVLHPS,
      MOVLHPD,
      MOVHLPS,
      MOVLPS,
      MOVLPD,
      MOVSD,
      MOVSS,
      UNPCKL,
      UNPCKH,
      VPERMILP,
      VPERMV,
      VPERMV3,
      VPERMIV3,
      VPERMI,
      VPERM2X128,
      VBROADCAST,
      // masked broadcast
      VBROADCASTM,
      // Insert/Extract vector element
      VINSERT,
      VEXTRACT,

      // PMULUDQ - Vector multiply packed unsigned doubleword integers
      PMULUDQ,

      // FMA nodes
      FMADD,
      FNMADD,
      FMSUB,
      FNMSUB,
      FMADDSUB,
      FMSUBADD,

      // VASTART_SAVE_XMM_REGS - Save xmm argument registers to the stack,
      // according to %al. An operator is needed so that this can be expanded
      // with control flow.
      VASTART_SAVE_XMM_REGS,

      // WIN_ALLOCA - Windows's _chkstk call to do stack probing.
      WIN_ALLOCA,

      // SEG_ALLOCA - For allocating variable amounts of stack space when using
      // segmented stacks. Check if the current stacklet has enough space, and
      // falls back to heap allocation if not.
      SEG_ALLOCA,

      // WIN_FTOL - Windows's _ftol2 runtime routine to do fptoui.
      WIN_FTOL,

      // Memory barrier
      MEMBARRIER,
      MFENCE,
      SFENCE,
      LFENCE,

      // FNSTSW16r - Store FP status word into i16 register.
      FNSTSW16r,

      // SAHF - Store contents of %ah into %eflags.
      SAHF,

      // RDRAND - Get a random integer and indicate whether it is valid in CF.
      RDRAND,

      // RDSEED - Get a NIST SP800-90B & C compliant random integer and
      // indicate whether it is valid in CF.
      RDSEED,

      // PCMP*STRI
      PCMPISTRI,
      PCMPESTRI,

      // XTEST - Test if in transactional execution.
      XTEST,

      // ATOMADD64_DAG, ATOMSUB64_DAG, ATOMOR64_DAG, ATOMAND64_DAG,
      // ATOMXOR64_DAG, ATOMNAND64_DAG, ATOMSWAP64_DAG -
      // Atomic 64-bit binary operations.
      ATOMADD64_DAG = ISD::FIRST_TARGET_MEMORY_OPCODE,
      ATOMSUB64_DAG,
      ATOMOR64_DAG,
      ATOMXOR64_DAG,
      ATOMAND64_DAG,
      ATOMNAND64_DAG,
      ATOMMAX64_DAG,
      ATOMMIN64_DAG,
      ATOMUMAX64_DAG,
      ATOMUMIN64_DAG,
      ATOMSWAP64_DAG,

      // LCMPXCHG_DAG, LCMPXCHG8_DAG, LCMPXCHG16_DAG - Compare and swap.
      LCMPXCHG_DAG,
      LCMPXCHG8_DAG,
      LCMPXCHG16_DAG,

      // VZEXT_LOAD - Load, scalar_to_vector, and zero extend.
      VZEXT_LOAD,

      // FNSTCW16m - Store FP control world into i16 memory.
      FNSTCW16m,

      /// FP_TO_INT*_IN_MEM - This instruction implements FP_TO_SINT with the
      /// integer destination in memory and a FP reg source.  This corresponds
      /// to the X86::FIST*m instructions and the rounding mode change stuff. It
      /// has two inputs (token chain and address) and two outputs (int value
      /// and token chain).
      FP_TO_INT16_IN_MEM,
      FP_TO_INT32_IN_MEM,
      FP_TO_INT64_IN_MEM,

      /// FILD, FILD_FLAG - This instruction implements SINT_TO_FP with the
      /// integer source in memory and FP reg result.  This corresponds to the
      /// X86::FILD*m instructions. It has three inputs (token chain, address,
      /// and source type) and two outputs (FP value and token chain). FILD_FLAG
      /// also produces a flag).
      FILD,
      FILD_FLAG,

      /// FLD - This instruction implements an extending load to FP stack slots.
      /// This corresponds to the X86::FLD32m / X86::FLD64m. It takes a chain
      /// operand, ptr to load from, and a ValueType node indicating the type
      /// to load to.
      FLD,

      /// FST - This instruction implements a truncating store to FP stack
      /// slots. This corresponds to the X86::FST32m / X86::FST64m. It takes a
      /// chain operand, value to store, address, and a ValueType to store it
      /// as.
      FST,

      /// VAARG_64 - This instruction grabs the address of the next argument
      /// from a va_list. (reads and modifies the va_list in memory)
      VAARG_64

      // WARNING: Do not add anything in the end unless you want the node to
      // have memop! In fact, starting from ATOMADD64_DAG all opcodes will be
      // thought as target memory ops!
    };
  }

  /// Define some predicates that are used for node matching.
  namespace X86 {
    /// isVEXTRACT128Index - Return true if the specified
    /// EXTRACT_SUBVECTOR operand specifies a vector extract that is
    /// suitable for input to VEXTRACTF128, VEXTRACTI128 instructions.
    bool isVEXTRACT128Index(SDNode *N);

    /// isVINSERT128Index - Return true if the specified
    /// INSERT_SUBVECTOR operand specifies a subvector insert that is
    /// suitable for input to VINSERTF128, VINSERTI128 instructions.
    bool isVINSERT128Index(SDNode *N);

    /// isVEXTRACT256Index - Return true if the specified
    /// EXTRACT_SUBVECTOR operand specifies a vector extract that is
    /// suitable for input to VEXTRACTF64X4, VEXTRACTI64X4 instructions.
    bool isVEXTRACT256Index(SDNode *N);

    /// isVINSERT256Index - Return true if the specified
    /// INSERT_SUBVECTOR operand specifies a subvector insert that is
    /// suitable for input to VINSERTF64X4, VINSERTI64X4 instructions.
    bool isVINSERT256Index(SDNode *N);

    /// getExtractVEXTRACT128Immediate - Return the appropriate
    /// immediate to extract the specified EXTRACT_SUBVECTOR index
    /// with VEXTRACTF128, VEXTRACTI128 instructions.
    unsigned getExtractVEXTRACT128Immediate(SDNode *N);

    /// getInsertVINSERT128Immediate - Return the appropriate
    /// immediate to insert at the specified INSERT_SUBVECTOR index
    /// with VINSERTF128, VINSERT128 instructions.
    unsigned getInsertVINSERT128Immediate(SDNode *N);

    /// getExtractVEXTRACT256Immediate - Return the appropriate
    /// immediate to extract the specified EXTRACT_SUBVECTOR index
    /// with VEXTRACTF64X4, VEXTRACTI64x4 instructions.
    unsigned getExtractVEXTRACT256Immediate(SDNode *N);

    /// getInsertVINSERT256Immediate - Return the appropriate
    /// immediate to insert at the specified INSERT_SUBVECTOR index
    /// with VINSERTF64x4, VINSERTI64x4 instructions.
    unsigned getInsertVINSERT256Immediate(SDNode *N);

    /// isZeroNode - Returns true if Elt is a constant zero or a floating point
    /// constant +0.0.
    bool isZeroNode(SDValue Elt);

    /// isOffsetSuitableForCodeModel - Returns true of the given offset can be
    /// fit into displacement field of the instruction.
    bool isOffsetSuitableForCodeModel(int64_t Offset, CodeModel::Model M,
                                      bool hasSymbolicDisplacement = true);


    /// isCalleePop - Determines whether the callee is required to pop its
    /// own arguments. Callee pop is necessary to support tail calls.
    bool isCalleePop(CallingConv::ID CallingConv,
                     bool is64Bit, bool IsVarArg, bool TailCallOpt);
  }

  //===--------------------------------------------------------------------===//
  //  X86TargetLowering - X86 Implementation of the TargetLowering interface
  class X86TargetLowering final : public TargetLowering {
  public:
    explicit X86TargetLowering(X86TargetMachine &TM);

    unsigned getJumpTableEncoding() const override;

    MVT getScalarShiftAmountTy(EVT LHSTy) const override { return MVT::i8; }

    const MCExpr *
    LowerCustomJumpTableEntry(const MachineJumpTableInfo *MJTI,
                              const MachineBasicBlock *MBB, unsigned uid,
                              MCContext &Ctx) const override;

    /// getPICJumpTableRelocaBase - Returns relocation base for the given PIC
    /// jumptable.
    SDValue getPICJumpTableRelocBase(SDValue Table,
                                     SelectionDAG &DAG) const override;
    const MCExpr *
    getPICJumpTableRelocBaseExpr(const MachineFunction *MF,
                                 unsigned JTI, MCContext &Ctx) const override;

    /// getByValTypeAlignment - Return the desired alignment for ByVal aggregate
    /// function arguments in the caller parameter area. For X86, aggregates
    /// that contains are placed at 16-byte boundaries while the rest are at
    /// 4-byte boundaries.
    unsigned getByValTypeAlignment(Type *Ty) const override;

    /// getOptimalMemOpType - Returns the target specific optimal type for load
    /// and store operations as a result of memset, memcpy, and memmove
    /// lowering. If DstAlign is zero that means it's safe to destination
    /// alignment can satisfy any constraint. Similarly if SrcAlign is zero it
    /// means there isn't a need to check it against alignment requirement,
    /// probably because the source does not need to be loaded. If 'IsMemset' is
    /// true, that means it's expanding a memset. If 'ZeroMemset' is true, that
    /// means it's a memset of zero. 'MemcpyStrSrc' indicates whether the memcpy
    /// source is constant so it does not need to be loaded.
    /// It returns EVT::Other if the type should be determined using generic
    /// target-independent logic.
    EVT getOptimalMemOpType(uint64_t Size, unsigned DstAlign, unsigned SrcAlign,
                            bool IsMemset, bool ZeroMemset, bool MemcpyStrSrc,
                            MachineFunction &MF) const override;

    /// isSafeMemOpType - Returns true if it's safe to use load / store of the
    /// specified type to expand memcpy / memset inline. This is mostly true
    /// for all types except for some special cases. For example, on X86
    /// targets without SSE2 f64 load / store are done with fldl / fstpl which
    /// also does type conversion. Note the specified type doesn't have to be
    /// legal as the hook is used before type legalization.
    bool isSafeMemOpType(MVT VT) const override;

    /// allowsUnalignedMemoryAccesses - Returns true if the target allows
    /// unaligned memory accesses. of the specified type. Returns whether it
    /// is "fast" by reference in the second argument.
    bool allowsUnalignedMemoryAccesses(EVT VT, unsigned AS,
                                       bool *Fast) const override;

    /// LowerOperation - Provide custom lowering hooks for some operations.
    ///
    SDValue LowerOperation(SDValue Op, SelectionDAG &DAG) const override;

    /// ReplaceNodeResults - Replace the results of node with an illegal result
    /// type with new values built out of custom code.
    ///
    void ReplaceNodeResults(SDNode *N, SmallVectorImpl<SDValue>&Results,
                            SelectionDAG &DAG) const override;


    SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const override;

    /// isTypeDesirableForOp - Return true if the target has native support for
    /// the specified value type and it is 'desirable' to use the type for the
    /// given node type. e.g. On x86 i16 is legal, but undesirable since i16
    /// instruction encodings are longer and some i16 instructions are slow.
    bool isTypeDesirableForOp(unsigned Opc, EVT VT) const override;

    /// isTypeDesirable - Return true if the target has native support for the
    /// specified value type and it is 'desirable' to use the type. e.g. On x86
    /// i16 is legal, but undesirable since i16 instruction encodings are longer
    /// and some i16 instructions are slow.
    bool IsDesirableToPromoteOp(SDValue Op, EVT &PVT) const override;

    MachineBasicBlock *
      EmitInstrWithCustomInserter(MachineInstr *MI,
                                  MachineBasicBlock *MBB) const override;


    /// getTargetNodeName - This method returns the name of a target specific
    /// DAG node.
    const char *getTargetNodeName(unsigned Opcode) const override;

    /// getSetCCResultType - Return the value type to use for ISD::SETCC.
    EVT getSetCCResultType(LLVMContext &Context, EVT VT) const override;

    /// computeMaskedBitsForTargetNode - Determine which of the bits specified
    /// in Mask are known to be either zero or one and return them in the
    /// KnownZero/KnownOne bitsets.
    void computeMaskedBitsForTargetNode(const SDValue Op,
                                        APInt &KnownZero,
                                        APInt &KnownOne,
                                        const SelectionDAG &DAG,
                                        unsigned Depth = 0) const override;

    // ComputeNumSignBitsForTargetNode - Determine the number of bits in the
    // operation that are sign bits.
    unsigned ComputeNumSignBitsForTargetNode(SDValue Op,
                                             const SelectionDAG &DAG,
                                             unsigned Depth) const override;

    bool isGAPlusOffset(SDNode *N, const GlobalValue* &GA,
                        int64_t &Offset) const override;

    SDValue getReturnAddressFrameIndex(SelectionDAG &DAG) const;

    bool ExpandInlineAsm(CallInst *CI) const override;

    ConstraintType
      getConstraintType(const std::string &Constraint) 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;

    const char *LowerXConstraint(EVT ConstraintVT) const override;

    /// LowerAsmOperandForConstraint - Lower the specified operand into the Ops
    /// vector.  If it is invalid, don't add anything to Ops. If hasMemory is
    /// true it means one of the asm constraint of the inline asm instruction
    /// being processed is 'm'.
    void LowerAsmOperandForConstraint(SDValue Op,
                                      std::string &Constraint,
                                      std::vector<SDValue> &Ops,
                                      SelectionDAG &DAG) const override;

    /// getRegForInlineAsmConstraint - Given a physical register constraint
    /// (e.g. {edx}), return the register number and the register class for the
    /// register.  This should only be used for C_Register constraints.  On
    /// error, this returns a register number of 0.
    std::pair<unsigned, const TargetRegisterClass*>
      getRegForInlineAsmConstraint(const std::string &Constraint,
                                   MVT VT) const override;

    /// isLegalAddressingMode - Return true if the addressing mode represented
    /// by AM is legal for this target, for a load/store of the specified type.
    bool isLegalAddressingMode(const AddrMode &AM, Type *Ty) const override;

    /// isLegalICmpImmediate - Return true if the specified immediate is legal
    /// icmp immediate, that is the target has icmp instructions which can
    /// compare a register against the immediate without having to materialize
    /// the immediate into a register.
    bool isLegalICmpImmediate(int64_t Imm) const override;

    /// isLegalAddImmediate - Return true if the specified immediate is legal
    /// add immediate, that is the target has add instructions which can
    /// add a register and the immediate without having to materialize
    /// the immediate into a register.
    bool isLegalAddImmediate(int64_t Imm) const override;


    bool isVectorShiftByScalarCheap(Type *Ty) const override;

    /// isTruncateFree - Return true if it's free to truncate a value of
    /// type Ty1 to type Ty2. e.g. On x86 it's free to truncate a i32 value in
    /// register EAX to i16 by referencing its sub-register AX.
    bool isTruncateFree(Type *Ty1, Type *Ty2) const override;
    bool isTruncateFree(EVT VT1, EVT VT2) const override;

    bool allowTruncateForTailCall(Type *Ty1, Type *Ty2) const override;

    /// isZExtFree - Return true if any actual instruction that defines a
    /// value of type Ty1 implicit zero-extends the value to Ty2 in the result
    /// register. This does not necessarily include registers defined in
    /// unknown ways, such as incoming arguments, or copies from unknown
    /// virtual registers. Also, if isTruncateFree(Ty2, Ty1) is true, this
    /// does not necessarily apply to truncate instructions. e.g. on x86-64,
    /// all instructions that define 32-bit values implicit zero-extend the
    /// result out to 64 bits.
    bool isZExtFree(Type *Ty1, Type *Ty2) const override;
    bool isZExtFree(EVT VT1, EVT VT2) const override;
    bool isZExtFree(SDValue Val, EVT VT2) const override;

    /// isFMAFasterThanFMulAndFAdd - Return true if an FMA operation is faster
    /// than a pair of fmul and fadd instructions. fmuladd intrinsics will be
    /// expanded to FMAs when this method returns true, otherwise fmuladd is
    /// expanded to fmul + fadd.
    bool isFMAFasterThanFMulAndFAdd(EVT VT) const override;

    /// isNarrowingProfitable - Return true if it's profitable to narrow
    /// operations of type VT1 to VT2. e.g. on x86, it's profitable to narrow
    /// from i32 to i8 but not from i32 to i16.
    bool isNarrowingProfitable(EVT VT1, EVT VT2) const override;

    /// isFPImmLegal - Returns true if the target can instruction select the
    /// specified FP immediate natively. If false, the legalizer will
    /// materialize the FP immediate as a load from a constant pool.
    bool isFPImmLegal(const APFloat &Imm, EVT VT) const override;

    /// isShuffleMaskLegal - Targets can use this to indicate that they only
    /// support *some* VECTOR_SHUFFLE operations, those with specific masks.
    /// By default, if a target supports the VECTOR_SHUFFLE node, all mask
    /// values are assumed to be legal.
    bool isShuffleMaskLegal(const SmallVectorImpl<int> &Mask,
                            EVT VT) const override;

    /// isVectorClearMaskLegal - Similar to isShuffleMaskLegal. This is
    /// used by Targets can use this to indicate if there is a suitable
    /// VECTOR_SHUFFLE that can be used to replace a VAND with a constant
    /// pool entry.
    bool isVectorClearMaskLegal(const SmallVectorImpl<int> &Mask,
                                EVT VT) const override;

    /// ShouldShrinkFPConstant - If true, then instruction selection should
    /// seek to shrink the FP constant of the specified type to a smaller type
    /// in order to save space and / or reduce runtime.
    bool ShouldShrinkFPConstant(EVT VT) const override {
      // Don't shrink FP constpool if SSE2 is available since cvtss2sd is more
      // expensive than a straight movsd. On the other hand, it's important to
      // shrink long double fp constant since fldt is very slow.
      return !X86ScalarSSEf64 || VT == MVT::f80;
    }

    const X86Subtarget* getSubtarget() const {
      return Subtarget;
    }

    /// isScalarFPTypeInSSEReg - Return true if the specified scalar FP type is
    /// computed in an SSE register, not on the X87 floating point stack.
    bool isScalarFPTypeInSSEReg(EVT VT) const {
      return (VT == MVT::f64 && X86ScalarSSEf64) || // f64 is when SSE2
      (VT == MVT::f32 && X86ScalarSSEf32);   // f32 is when SSE1
    }

    /// isTargetFTOL - Return true if the target uses the MSVC _ftol2 routine
    /// for fptoui.
    bool isTargetFTOL() const {
      return Subtarget->isTargetKnownWindowsMSVC() && !Subtarget->is64Bit();
    }

    /// isIntegerTypeFTOL - Return true if the MSVC _ftol2 routine should be
    /// used for fptoui to the given type.
    bool isIntegerTypeFTOL(EVT VT) const {
      return isTargetFTOL() && VT == MVT::i64;
    }

    /// \brief Returns true if it is beneficial to convert a load of a constant
    /// to just the constant itself.
    bool shouldConvertConstantLoadToIntImm(const APInt &Imm,
                                           Type *Ty) const override;

    /// Intel processors have a unified instruction and data cache
    const char * getClearCacheBuiltinName() const {
      return 0; // nothing to do, move along.
    }

    /// createFastISel - This method returns a target specific FastISel object,
    /// or null if the target does not support "fast" ISel.
    FastISel *createFastISel(FunctionLoweringInfo &funcInfo,
                             const TargetLibraryInfo *libInfo) const override;

    /// getStackCookieLocation - Return true if the target stores stack
    /// protector cookies at a fixed offset in some non-standard address
    /// space, and populates the address space and offset as
    /// appropriate.
    bool getStackCookieLocation(unsigned &AddressSpace,
                                unsigned &Offset) const override;

    SDValue BuildFILD(SDValue Op, EVT SrcVT, SDValue Chain, SDValue StackSlot,
                      SelectionDAG &DAG) const;

    bool isNoopAddrSpaceCast(unsigned SrcAS, unsigned DestAS) const override;

    /// \brief Reset the operation actions based on target options.
    void resetOperationActions() override;

  protected:
    std::pair<const TargetRegisterClass*, uint8_t>
    findRepresentativeClass(MVT VT) const override;

  private:
    /// Subtarget - Keep a pointer to the X86Subtarget around so that we can
    /// make the right decision when generating code for different targets.
    const X86Subtarget *Subtarget;
    const DataLayout *TD;

    /// Used to store the TargetOptions so that we don't waste time resetting
    /// the operation actions unless we have to.
    TargetOptions TO;

    /// X86ScalarSSEf32, X86ScalarSSEf64 - Select between SSE or x87
    /// floating point ops.
    /// When SSE is available, use it for f32 operations.
    /// When SSE2 is available, use it for f64 operations.
    bool X86ScalarSSEf32;
    bool X86ScalarSSEf64;

    /// LegalFPImmediates - A list of legal fp immediates.
    std::vector<APFloat> LegalFPImmediates;

    /// addLegalFPImmediate - Indicate that this x86 target can instruction
    /// select the specified FP immediate natively.
    void addLegalFPImmediate(const APFloat& Imm) {
      LegalFPImmediates.push_back(Imm);
    }

    SDValue LowerCallResult(SDValue Chain, SDValue InFlag,
                            CallingConv::ID CallConv, bool isVarArg,
                            const SmallVectorImpl<ISD::InputArg> &Ins,
                            SDLoc dl, SelectionDAG &DAG,
                            SmallVectorImpl<SDValue> &InVals) const;
    SDValue LowerMemArgument(SDValue Chain,
                             CallingConv::ID CallConv,
                             const SmallVectorImpl<ISD::InputArg> &ArgInfo,
                             SDLoc dl, SelectionDAG &DAG,
                             const CCValAssign &VA,  MachineFrameInfo *MFI,
                              unsigned i) const;
    SDValue LowerMemOpCallTo(SDValue Chain, SDValue StackPtr, SDValue Arg,
                             SDLoc dl, SelectionDAG &DAG,
                             const CCValAssign &VA,
                             ISD::ArgFlagsTy Flags) const;

    // Call lowering helpers.

    /// IsEligibleForTailCallOptimization - Check whether the call is eligible
    /// for tail call optimization. Targets which want to do tail call
    /// optimization should implement this function.
    bool IsEligibleForTailCallOptimization(SDValue Callee,
                                           CallingConv::ID CalleeCC,
                                           bool isVarArg,
                                           bool isCalleeStructRet,
                                           bool isCallerStructRet,
                                           Type *RetTy,
                                    const SmallVectorImpl<ISD::OutputArg> &Outs,
                                    const SmallVectorImpl<SDValue> &OutVals,
                                    const SmallVectorImpl<ISD::InputArg> &Ins,
                                           SelectionDAG& DAG) const;
    bool IsCalleePop(bool isVarArg, CallingConv::ID CallConv) const;
    SDValue EmitTailCallLoadRetAddr(SelectionDAG &DAG, SDValue &OutRetAddr,
                                SDValue Chain, bool IsTailCall, bool Is64Bit,
                                int FPDiff, SDLoc dl) const;

    unsigned GetAlignedArgumentStackSize(unsigned StackSize,
                                         SelectionDAG &DAG) const;

    std::pair<SDValue,SDValue> FP_TO_INTHelper(SDValue Op, SelectionDAG &DAG,
                                               bool isSigned,
                                               bool isReplace) const;

    SDValue LowerBUILD_VECTOR(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerBUILD_VECTORvXi1(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerVECTOR_SHUFFLE(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerEXTRACT_VECTOR_ELT(SDValue Op, SelectionDAG &DAG) const;
    SDValue ExtractBitFromMaskVector(SDValue Op, SelectionDAG &DAG) const;
    SDValue InsertBitToMaskVector(SDValue Op, SelectionDAG &DAG) const;

    SDValue LowerINSERT_VECTOR_ELT(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerConstantPool(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerBlockAddress(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerGlobalAddress(const GlobalValue *GV, SDLoc dl,
                               int64_t Offset, SelectionDAG &DAG) const;
    SDValue LowerGlobalAddress(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerGlobalTLSAddress(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerExternalSymbol(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerSINT_TO_FP(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerUINT_TO_FP(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerUINT_TO_FP_i64(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerUINT_TO_FP_i32(SDValue Op, SelectionDAG &DAG) const;
    SDValue lowerUINT_TO_FP_vec(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerTRUNCATE(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerFP_TO_SINT(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerFP_TO_UINT(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerToBT(SDValue And, ISD::CondCode CC,
                      SDLoc dl, SelectionDAG &DAG) const;
    SDValue LowerSETCC(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerSELECT(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerBRCOND(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerMEMSET(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerJumpTable(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerDYNAMIC_STACKALLOC(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerVASTART(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerVAARG(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerRETURNADDR(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerFRAMEADDR(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerFRAME_TO_ARGS_OFFSET(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerEH_RETURN(SDValue Op, SelectionDAG &DAG) const;
    SDValue lowerEH_SJLJ_SETJMP(SDValue Op, SelectionDAG &DAG) const;
    SDValue lowerEH_SJLJ_LONGJMP(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerINIT_TRAMPOLINE(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerFLT_ROUNDS_(SDValue Op, SelectionDAG &DAG) const;
    SDValue LowerSIGN_EXTEND_INREG(SDValue Op, SelectionDAG &DAG) const;

    SDValue
      LowerFormalArguments(SDValue Chain,
                           CallingConv::ID CallConv, bool isVarArg,
                           const SmallVectorImpl<ISD::InputArg> &Ins,
                           SDLoc dl, SelectionDAG &DAG,
                           SmallVectorImpl<SDValue> &InVals) const override;
    SDValue LowerCall(CallLoweringInfo &CLI,
                      SmallVectorImpl<SDValue> &InVals) const override;

    SDValue LowerReturn(SDValue Chain,
                        CallingConv::ID CallConv, bool isVarArg,
                        const SmallVectorImpl<ISD::OutputArg> &Outs,
                        const SmallVectorImpl<SDValue> &OutVals,
                        SDLoc dl, SelectionDAG &DAG) const override;

    bool isUsedByReturnOnly(SDNode *N, SDValue &Chain) const override;

    bool mayBeEmittedAsTailCall(CallInst *CI) const override;

    MVT getTypeForExtArgOrReturn(MVT VT,
                                 ISD::NodeType ExtendKind) const override;

    bool CanLowerReturn(CallingConv::ID CallConv, MachineFunction &MF,
                        bool isVarArg,
                        const SmallVectorImpl<ISD::OutputArg> &Outs,
                        LLVMContext &Context) const override;

    const MCPhysReg *getScratchRegisters(CallingConv::ID CC) const override;

    /// Utility function to emit atomic-load-arith operations (and, or, xor,
    /// nand, max, min, umax, umin). It takes the corresponding instruction to
    /// expand, the associated machine basic block, and the associated X86
    /// opcodes for reg/reg.
    MachineBasicBlock *EmitAtomicLoadArith(MachineInstr *MI,
                                           MachineBasicBlock *MBB) const;

    /// Utility function to emit atomic-load-arith operations (and, or, xor,
    /// nand, add, sub, swap) for 64-bit operands on 32-bit target.
    MachineBasicBlock *EmitAtomicLoadArith6432(MachineInstr *MI,
                                               MachineBasicBlock *MBB) const;

    // Utility function to emit the low-level va_arg code for X86-64.
    MachineBasicBlock *EmitVAARG64WithCustomInserter(
                       MachineInstr *MI,
                       MachineBasicBlock *MBB) const;

    /// Utility function to emit the xmm reg save portion of va_start.
    MachineBasicBlock *EmitVAStartSaveXMMRegsWithCustomInserter(
                                                   MachineInstr *BInstr,
                                                   MachineBasicBlock *BB) const;

    MachineBasicBlock *EmitLoweredSelect(MachineInstr *I,
                                         MachineBasicBlock *BB) const;

    MachineBasicBlock *EmitLoweredWinAlloca(MachineInstr *MI,
                                              MachineBasicBlock *BB) const;

    MachineBasicBlock *EmitLoweredSegAlloca(MachineInstr *MI,
                                            MachineBasicBlock *BB,
                                            bool Is64Bit) const;

    MachineBasicBlock *EmitLoweredTLSCall(MachineInstr *MI,
                                          MachineBasicBlock *BB) const;

    MachineBasicBlock *emitLoweredTLSAddr(MachineInstr *MI,
                                          MachineBasicBlock *BB) const;

    MachineBasicBlock *emitEHSjLjSetJmp(MachineInstr *MI,
                                        MachineBasicBlock *MBB) const;

    MachineBasicBlock *emitEHSjLjLongJmp(MachineInstr *MI,
                                         MachineBasicBlock *MBB) const;

    MachineBasicBlock *emitFMA3Instr(MachineInstr *MI,
                                     MachineBasicBlock *MBB) const;

    /// Emit nodes that will be selected as "test Op0,Op0", or something
    /// equivalent, for use with the given x86 condition code.
    SDValue EmitTest(SDLoc dl, SDValue Op0, unsigned X86CC,
                     SelectionDAG &DAG) const;

    /// Emit nodes that will be selected as "cmp Op0,Op1", or something
    /// equivalent, for use with the given x86 condition code.
    SDValue EmitCmp(SDLoc dl, SDValue Op0, SDValue Op1, unsigned X86CC,
                    SelectionDAG &DAG) const;

    /// Convert a comparison if required by the subtarget.
    SDValue ConvertCmpIfNecessary(SDValue Cmp, SelectionDAG &DAG) const;
  };

  namespace X86 {
    FastISel *createFastISel(FunctionLoweringInfo &funcInfo,
                             const TargetLibraryInfo *libInfo);
  }
}

#endif    // X86ISELLOWERING_H