path: root/llvm/lib/Target/SystemZ/SystemZOperators.td

//===-- SystemZOperators.td - SystemZ-specific operators ------*- tblgen-*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//

//===----------------------------------------------------------------------===//
// Type profiles
//===----------------------------------------------------------------------===//
def SDT_CallSeqStart        : SDCallSeqStart<[SDTCisVT<0, i64>,
                                              SDTCisVT<1, i64>]>;
def SDT_CallSeqEnd          : SDCallSeqEnd<[SDTCisVT<0, i64>,
                                            SDTCisVT<1, i64>]>;
def SDT_ZCall               : SDTypeProfile<0, -1, [SDTCisPtrTy<0>]>;
def SDT_ZCmp                : SDTypeProfile<1, 2,
                                            [SDTCisVT<0, i32>,
                                             SDTCisSameAs<1, 2>]>;
def SDT_ZICmp               : SDTypeProfile<1, 3,
                                            [SDTCisVT<0, i32>,
                                             SDTCisSameAs<1, 2>,
                                             SDTCisVT<3, i32>]>;
def SDT_ZBRCCMask           : SDTypeProfile<0, 4,
                                            [SDTCisVT<0, i32>,
                                             SDTCisVT<1, i32>,
                                             SDTCisVT<2, OtherVT>,
                                             SDTCisVT<3, i32>]>;
def SDT_ZSelectCCMask       : SDTypeProfile<1, 5,
                                            [SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<1, 2>,
                                             SDTCisVT<3, i32>,
                                             SDTCisVT<4, i32>,
                                             SDTCisVT<5, i32>]>;
def SDT_ZWrapPtr            : SDTypeProfile<1, 1,
                                            [SDTCisSameAs<0, 1>,
                                             SDTCisPtrTy<0>]>;
def SDT_ZWrapOffset         : SDTypeProfile<1, 2,
                                            [SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisPtrTy<0>]>;
def SDT_ZAdjDynAlloc        : SDTypeProfile<1, 0, [SDTCisVT<0, i64>]>;
def SDT_ZProbedAlloca       : SDTypeProfile<1, 2,
                                            [SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisPtrTy<0>]>;
def SDT_ZGR128Binary        : SDTypeProfile<1, 2,
                                            [SDTCisVT<0, untyped>,
                                             SDTCisInt<1>,
                                             SDTCisInt<2>]>;
def SDT_ZBinaryWithFlags    : SDTypeProfile<2, 2,
                                            [SDTCisInt<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>]>;
def SDT_ZBinaryWithCarry    : SDTypeProfile<2, 3,
                                            [SDTCisInt<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>,
                                             SDTCisVT<1, i32>]>;
def SDT_ZBinaryConv         : SDTypeProfile<1, 2,
                                            [SDTCisInt<0>,
                                             SDTCisInt<1>,
                                             SDTCisSameAs<1, 2>]>;
def SDT_ZTernary            : SDTypeProfile<1, 3,
                                            [SDTCisInt<0>,
                                             SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>]>;
def SDT_ZAtomicLoadBinaryW  : SDTypeProfile<1, 5,
                                            [SDTCisVT<0, i32>,
                                             SDTCisPtrTy<1>,
                                             SDTCisVT<2, i32>,
                                             SDTCisVT<3, i32>,
                                             SDTCisVT<4, i32>,
                                             SDTCisVT<5, i32>]>;
def SDT_ZAtomicCmpSwapW     : SDTypeProfile<2, 6,
                                            [SDTCisVT<0, i32>,
                                             SDTCisVT<1, i32>,
                                             SDTCisPtrTy<2>,
                                             SDTCisVT<3, i32>,
                                             SDTCisVT<4, i32>,
                                             SDTCisVT<5, i32>,
                                             SDTCisVT<6, i32>,
                                             SDTCisVT<7, i32>]>;
def SDT_ZAtomicCmpSwap      : SDTypeProfile<2, 3,
                                            [SDTCisInt<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisPtrTy<2>,
                                             SDTCisSameAs<0, 3>,
                                             SDTCisSameAs<0, 4>]>;
def SDT_ZAtomicLoad128      : SDTypeProfile<1, 1,
                                            [SDTCisVT<0, untyped>,
                                             SDTCisPtrTy<1>]>;
def SDT_ZAtomicStore128     : SDTypeProfile<0, 2,
                                            [SDTCisVT<0, untyped>,
                                             SDTCisPtrTy<1>]>;
def SDT_ZAtomicCmpSwap128   : SDTypeProfile<2, 3,
                                            [SDTCisVT<0, untyped>,
                                             SDTCisVT<1, i32>,
                                             SDTCisPtrTy<2>,
                                             SDTCisVT<3, untyped>,
                                             SDTCisVT<4, untyped>]>;
def SDT_ZMemMemLength       : SDTypeProfile<0, 3,
                                            [SDTCisPtrTy<0>,
                                             SDTCisPtrTy<1>,
                                             SDTCisVT<2, i64>]>;
def SDT_ZMemMemLengthCC     : SDTypeProfile<1, 3,
                                            [SDTCisVT<0, i32>,
                                             SDTCisPtrTy<1>,
                                             SDTCisPtrTy<2>,
                                             SDTCisVT<3, i64>]>;
def SDT_ZMemsetMVC          : SDTypeProfile<0, 3,
                                            [SDTCisPtrTy<0>,
                                             SDTCisVT<1, i64>,
                                             SDTCisVT<2, i32>]>;
def SDT_ZString             : SDTypeProfile<1, 3,
                                            [SDTCisPtrTy<0>,
                                             SDTCisPtrTy<1>,
                                             SDTCisPtrTy<2>,
                                             SDTCisVT<3, i32>]>;
def SDT_ZStringCC           : SDTypeProfile<2, 3,
                                            [SDTCisPtrTy<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisPtrTy<2>,
                                             SDTCisPtrTy<3>,
                                             SDTCisVT<4, i32>]>;
def SDT_ZIPM                : SDTypeProfile<1, 1,
                                            [SDTCisVT<0, i32>,
                                             SDTCisVT<1, i32>]>;
def SDT_ZPrefetch           : SDTypeProfile<0, 2,
                                            [SDTCisVT<0, i32>,
                                             SDTCisPtrTy<1>]>;
def SDT_ZStoreInherent      : SDTypeProfile<0, 1,
                                            [SDTCisPtrTy<0>]>;
def SDT_ZTBegin             : SDTypeProfile<1, 2,
                                            [SDTCisVT<0, i32>,
                                             SDTCisPtrTy<1>,
                                             SDTCisVT<2, i32>]>;
def SDT_ZADAENTRY           : SDTypeProfile<1, 3,
                                            [SDTCisPtrTy<0>,
                                             SDTCisPtrTy<1>,
                                             SDTCisPtrTy<2>,
                                             SDTCisVT<3, i64>]>;
def SDT_ZTEnd               : SDTypeProfile<1, 0,
                                            [SDTCisVT<0, i32>]>;
def SDT_ZInsertVectorElt    : SDTypeProfile<1, 3,
                                            [SDTCisVec<0>,
                                             SDTCisSameAs<0, 1>,
                                             SDTCisVT<3, i32>]>;
def SDT_ZExtractVectorElt   : SDTypeProfile<1, 2,
                                            [SDTCisVec<1>,
                                             SDTCisVT<2, i32>]>;
def SDT_ZReplicate          : SDTypeProfile<1, 1,
                                            [SDTCisVec<0>]>;
def SDT_ZVecUnpack          : SDTypeProfile<1, 1,
                                            [SDTCisVec<1>]>;
def SDT_ZVecUnaryConv       : SDTypeProfile<1, 1,
                                            [SDTCisVec<0>,
                                             SDTCisVec<1>]>;
def SDT_ZVecUnary           : SDTypeProfile<1, 1,
                                            [SDTCisVec<0>,
                                             SDTCisSameAs<0, 1>]>;
def SDT_ZVecUnaryCC         : SDTypeProfile<2, 1,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisSameAs<0, 2>]>;
def SDT_ZVecCompare         : SDTypeProfile<1, 2,
                                            [SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>]>;
def SDT_ZVecCompareCC       : SDTypeProfile<2, 2,
                                            [SDTCisVT<1, i32>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 2>]>;
def SDT_ZVecBinary          : SDTypeProfile<1, 2,
                                            [SDTCisVec<0>,
                                             SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>]>;
def SDT_ZVecBinaryCC        : SDTypeProfile<2, 2,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 2>]>;
def SDT_ZVecBinaryInt       : SDTypeProfile<1, 2,
                                            [SDTCisVec<0>,
                                             SDTCisSameAs<0, 1>,
                                             SDTCisVT<2, i32>]>;
def SDT_ZVecBinaryConv      : SDTypeProfile<1, 2,
                                            [SDTCisVec<0>,
                                             SDTCisVec<1>,
                                             SDTCisSameAs<1, 2>]>;
def SDT_ZVecBinaryConvCC    : SDTypeProfile<2, 2,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisVec<2>,
                                             SDTCisSameAs<2, 3>]>;
def SDT_ZVecBinaryConvIntCC : SDTypeProfile<2, 2,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisVec<2>,
                                             SDTCisVT<3, i32>]>;
def SDT_ZRotateMask         : SDTypeProfile<1, 2,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisVT<2, i32>]>;
def SDT_ZJoinDwords         : SDTypeProfile<1, 2,
                                            [SDTCisVT<0, v2i64>,
                                             SDTCisVT<1, i64>,
                                             SDTCisVT<2, i64>]>;
def SDT_ZVecTernary         : SDTypeProfile<1, 3,
                                            [SDTCisVec<0>,
                                             SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>]>;
def SDT_ZVecTernaryConvCC   : SDTypeProfile<2, 3,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisVec<2>,
                                             SDTCisSameAs<2, 3>,
                                             SDTCisSameAs<0, 4>]>;
def SDT_ZVecTernaryInt      : SDTypeProfile<1, 3,
                                            [SDTCisVec<0>,
                                             SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisVT<3, i32>]>;
def SDT_ZVecTernaryIntCC    : SDTypeProfile<2, 3,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>,
                                             SDTCisVT<4, i32>]>;
def SDT_ZVecQuaternaryInt   : SDTypeProfile<1, 4,
                                            [SDTCisVec<0>,
                                             SDTCisSameAs<0, 1>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>,
                                             SDTCisVT<4, i32>]>;
def SDT_ZVecQuaternaryIntCC : SDTypeProfile<2, 4,
                                            [SDTCisVec<0>,
                                             SDTCisVT<1, i32>,
                                             SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>,
                                             SDTCisSameAs<0, 4>,
                                             SDTCisVT<5, i32>]>;
def SDT_ZTest               : SDTypeProfile<1, 2,
                                            [SDTCisVT<0, i32>,
                                             SDTCisVT<2, i64>]>;

def SDT_ZSetJmp             : SDTypeProfile<1, 1,
                                            [SDTCisInt<0>,
                                             SDTCisPtrTy<1>]>;
def SDT_ZLongJmp            : SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>;


//===----------------------------------------------------------------------===//
// Node definitions
//===----------------------------------------------------------------------===//

// These are target-independent nodes, but have target-specific formats.
def callseq_start       : SDNode<"ISD::CALLSEQ_START", SDT_CallSeqStart,
                                 [SDNPHasChain, SDNPSideEffect, SDNPOutGlue]>;
def callseq_end         : SDNode<"ISD::CALLSEQ_END",   SDT_CallSeqEnd,
                                 [SDNPHasChain, SDNPSideEffect, SDNPOptInGlue,
                                  SDNPOutGlue]>;
def global_offset_table : SDNode<"ISD::GLOBAL_OFFSET_TABLE", SDTPtrLeaf>;

// Return with a glue operand.  Operand 0 is the chain operand.
def z_retglue           : SDNode<"SystemZISD::RET_GLUE", SDTNone,
                                 [SDNPHasChain, SDNPOptInGlue, SDNPVariadic]>;

// Calls a function.  Operand 0 is the chain operand and operand 1
// is the target address.  The arguments start at operand 2.
// There is an optional glue operand at the end.
def z_call              : SDNode<"SystemZISD::CALL", SDT_ZCall,
                                 [SDNPHasChain, SDNPOutGlue, SDNPOptInGlue,
                                  SDNPVariadic]>;
def z_sibcall           : SDNode<"SystemZISD::SIBCALL", SDT_ZCall,
                                 [SDNPHasChain, SDNPOutGlue, SDNPOptInGlue,
                                  SDNPVariadic]>;
// TLS calls.  Like regular calls, except operand 1 is the TLS symbol.
// (The call target is implicitly __tls_get_offset.)
def z_tls_gdcall        : SDNode<"SystemZISD::TLS_GDCALL", SDT_ZCall,
                                 [SDNPHasChain, SDNPInGlue, SDNPOutGlue,
                                  SDNPVariadic]>;
def z_tls_ldcall        : SDNode<"SystemZISD::TLS_LDCALL", SDT_ZCall,
                                 [SDNPHasChain, SDNPInGlue, SDNPOutGlue,
                                  SDNPVariadic]>;

// Wraps a TargetGlobalAddress that should be loaded using PC-relative
// accesses (LARL).  Operand 0 is the address.
def z_pcrel_wrapper     : SDNode<"SystemZISD::PCREL_WRAPPER", SDT_ZWrapPtr, []>;

// Used in cases where an offset is applied to a TargetGlobalAddress.
// Operand 0 is the full TargetGlobalAddress and operand 1 is a
// PCREL_WRAPPER for an anchor point.  This is used so that we can
// cheaply refer to either the full address or the anchor point
// as a register base.
def z_pcrel_offset      : SDNode<"SystemZISD::PCREL_OFFSET",
                                 SDT_ZWrapOffset, []>;

// Integer comparisons.  There are three operands: the two values
// to compare, and an integer of type SystemZICMP.
def z_icmp              : SDNode<"SystemZISD::ICMP", SDT_ZICmp>;

// Floating-point comparisons.  The two operands are the values to compare.
def z_fcmp              : SDNode<"SystemZISD::FCMP", SDT_ZCmp>;

let IsStrictFP = true in {
  // Strict variants of scalar floating-point comparisons.
  // Quiet and signaling versions.
  def z_strict_fcmp : SDNode<"SystemZISD::STRICT_FCMP", SDT_ZCmp,
                             [SDNPHasChain]>;
  def z_strict_fcmps : SDNode<"SystemZISD::STRICT_FCMPS", SDT_ZCmp,
                              [SDNPHasChain]>;
}

// Test under mask.  The first operand is ANDed with the second operand
// and the condition codes are set on the result.  The third operand is
// a boolean that is true if the condition codes need to distinguish
// between CCMASK_TM_MIXED_MSB_0 and CCMASK_TM_MIXED_MSB_1 (which the
// register forms do but the memory forms don't).
def z_tm                : SDNode<"SystemZISD::TM", SDT_ZICmp>;

// Branches if a condition is true.  Operand 0 is the chain operand;
// operand 1 is the 4-bit condition-code mask, with bit N in
// big-endian order meaning "branch if CC=N"; operand 2 is the
// target block and operand 3 is the flag operand.
def z_br_ccmask_1       : SDNode<"SystemZISD::BR_CCMASK", SDT_ZBRCCMask,
                                 [SDNPHasChain]>;

// Selects between operand 0 and operand 1.  Operand 2 is the
// mask of condition-code values for which operand 0 should be
// chosen over operand 1; it has the same form as BR_CCMASK.
// Operand 3 is the flag operand.
def z_select_ccmask_1   : SDNode<"SystemZISD::SELECT_CCMASK",
                                 SDT_ZSelectCCMask>;

// Store the CC value in bits 29 and 28 of an integer.
def z_ipm_1             : SDNode<"SystemZISD::IPM", SDT_ZIPM>;

// Evaluates to the gap between the stack pointer and the
// base of the dynamically-allocatable area.
def z_adjdynalloc       : SDNode<"SystemZISD::ADJDYNALLOC", SDT_ZAdjDynAlloc>;

// For allocating stack space when using stack clash protector.
// Allocation is performed by block, and each block is probed.
def z_probed_alloca     : SDNode<"SystemZISD::PROBED_ALLOCA", SDT_ZProbedAlloca,
                                 [SDNPHasChain]>;

// Count number of bits set in operand 0 per byte.
def z_popcnt            : SDNode<"SystemZISD::POPCNT", SDTIntUnaryOp>;

// Wrappers around the ISD opcodes of the same name.  The output is GR128.
// Input operands may be GR64 or GR32, depending on the instruction.
def z_smul_lohi         : SDNode<"SystemZISD::SMUL_LOHI", SDT_ZGR128Binary>;
def z_umul_lohi         : SDNode<"SystemZISD::UMUL_LOHI", SDT_ZGR128Binary>;
def z_sdivrem           : SDNode<"SystemZISD::SDIVREM", SDT_ZGR128Binary>;
def z_udivrem           : SDNode<"SystemZISD::UDIVREM", SDT_ZGR128Binary>;

// Add/subtract with overflow/carry.  These have the same operands as
// the corresponding standard operations, except with the carry flag
// replaced by a condition code value.
def z_saddo             : SDNode<"SystemZISD::SADDO", SDT_ZBinaryWithFlags>;
def z_ssubo             : SDNode<"SystemZISD::SSUBO", SDT_ZBinaryWithFlags>;
def z_uaddo             : SDNode<"SystemZISD::UADDO", SDT_ZBinaryWithFlags>;
def z_usubo             : SDNode<"SystemZISD::USUBO", SDT_ZBinaryWithFlags>;
def z_addcarry_1        : SDNode<"SystemZISD::ADDCARRY", SDT_ZBinaryWithCarry>;
def z_subcarry_1        : SDNode<"SystemZISD::SUBCARRY", SDT_ZBinaryWithCarry>;

// Compute carry/borrow indication for add/subtract.
def z_vacc              : SDNode<"SystemZISD::VACC", SDTIntBinOp>;
def z_vscbi             : SDNode<"SystemZISD::VSCBI", SDTIntBinOp>;

// Add/subtract with carry/borrow.
def z_vac               : SDNode<"SystemZISD::VAC", SDT_ZTernary>;
def z_vsbi              : SDNode<"SystemZISD::VSBI", SDT_ZTernary>;

// Compute carry/borrow indication for add/subtract with carry/borrow.
def z_vaccc             : SDNode<"SystemZISD::VACCC", SDT_ZTernary>;
def z_vsbcbi            : SDNode<"SystemZISD::VSBCBI", SDT_ZTernary>;

// High-word multiply-and-add.
def z_vmah              : SDNode<"SystemZISD::VMAH", SDT_ZTernary>;
def z_vmalh             : SDNode<"SystemZISD::VMALH", SDT_ZTernary>;

// Widen and multiply even/odd vector elements.
def z_vme               : SDNode<"SystemZISD::VME", SDT_ZBinaryConv>;
def z_vmle              : SDNode<"SystemZISD::VMLE", SDT_ZBinaryConv>;
def z_vmo               : SDNode<"SystemZISD::VMO", SDT_ZBinaryConv>;
def z_vmlo              : SDNode<"SystemZISD::VMLO", SDT_ZBinaryConv>;

// Byte swapping load/store.  Same operands as regular load/store.
def z_loadbswap        : SDNode<"SystemZISD::LRV", SDTLoad,
                                 [SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>;
def z_storebswap       : SDNode<"SystemZISD::STRV", SDTStore,
                                 [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;

// Element swapping load/store.  Same operands as regular load/store.
def z_loadeswap        : SDNode<"SystemZISD::VLER", SDTLoad,
                                 [SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>;
def z_storeeswap       : SDNode<"SystemZISD::VSTER", SDTStore,
                                 [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;

// Use STORE CLOCK FAST to store current TOD clock value.
def z_stckf            : SDNode<"SystemZISD::STCKF", SDT_ZStoreInherent,
                                [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;

// Test Data Class.
//
// Operand 0: the value to test
// Operand 1: the bit mask
def z_tdc               : SDNode<"SystemZISD::TDC", SDT_ZTest>;

def z_eh_sjlj_setjmp    : SDNode<"ISD::EH_SJLJ_SETJMP", SDT_ZSetJmp,
                                 [SDNPHasChain, SDNPSideEffect]>;
def z_eh_sjlj_longjmp   : SDNode<"ISD::EH_SJLJ_LONGJMP", SDT_ZLongJmp,
                                 [SDNPHasChain, SDNPSideEffect]>;


// Defined because the index is an i32 rather than a pointer.
def z_vector_insert     : SDNode<"ISD::INSERT_VECTOR_ELT",
                                 SDT_ZInsertVectorElt>;
def z_vector_extract    : SDNode<"ISD::EXTRACT_VECTOR_ELT",
                                 SDT_ZExtractVectorElt>;

// Create a vector constant by filling byte N of the result with bit
// 15-N of the single operand.
def z_byte_mask         : SDNode<"SystemZISD::BYTE_MASK", SDT_ZReplicate>;

// Create a vector constant by replicating an element-sized RISBG-style mask.
// The first operand specifies the starting set bit and the second operand
// specifies the ending set bit.  Both operands count from the MSB of the
// element.
def z_rotate_mask       : SDNode<"SystemZISD::ROTATE_MASK", SDT_ZRotateMask>;

// Replicate a GPR scalar value into all elements of a vector.
def z_replicate         : SDNode<"SystemZISD::REPLICATE", SDT_ZReplicate>;

// Create a vector from two i64 GPRs.
def z_join_dwords       : SDNode<"SystemZISD::JOIN_DWORDS", SDT_ZJoinDwords>;

// Replicate one element of a vector into all elements.  The first operand
// is the vector and the second is the index of the element to replicate.
def z_splat             : SDNode<"SystemZISD::SPLAT", SDT_ZVecBinaryInt>;

// Interleave elements from the high half of operand 0 and the high half
// of operand 1.
def z_merge_high        : SDNode<"SystemZISD::MERGE_HIGH", SDT_ZVecBinary>;

// Likewise for the low halves.
def z_merge_low         : SDNode<"SystemZISD::MERGE_LOW", SDT_ZVecBinary>;

// Concatenate the vectors in the first two operands, shift them left
// by the third operand, and take the first half of the result.
def z_shl_double        : SDNode<"SystemZISD::SHL_DOUBLE", SDT_ZVecTernaryInt>;

// Concatenate the vectors in the first two operands, shift them left/right
// bitwise by the third operand, and take the first/last half of the result.
def z_shl_double_bit    : SDNode<"SystemZISD::SHL_DOUBLE_BIT", SDT_ZVecTernaryInt>;
def z_shr_double_bit    : SDNode<"SystemZISD::SHR_DOUBLE_BIT", SDT_ZVecTernaryInt>;

// Take one element of the first v2i64 operand and the one element of
// the second v2i64 operand and concatenate them to form a v2i64 result.
// The third operand is a 4-bit value of the form 0A0B, where A and B
// are the element selectors for the first operand and second operands
// respectively.
def z_permute_dwords    : SDNode<"SystemZISD::PERMUTE_DWORDS",
                                 SDT_ZVecTernaryInt>;

// Perform a general vector permute on vector operands 0 and 1.
// Each byte of operand 2 controls the corresponding byte of the result,
// in the same way as a byte-level VECTOR_SHUFFLE mask.
def z_permute           : SDNode<"SystemZISD::PERMUTE", SDT_ZVecTernary>;

// Pack vector operands 0 and 1 into a single vector with half-sized elements.
def z_pack              : SDNode<"SystemZISD::PACK", SDT_ZVecBinaryConv>;

// Likewise, but saturate the result and set CC.  PACKS_CC does signed
// saturation and PACKLS_CC does unsigned saturation.
def z_packs_cc          : SDNode<"SystemZISD::PACKS_CC", SDT_ZVecBinaryConvCC>;
def z_packls_cc         : SDNode<"SystemZISD::PACKLS_CC", SDT_ZVecBinaryConvCC>;

// Unpack the first half of vector operand 0 into double-sized elements.
// UNPACK_HIGH sign-extends and UNPACKL_HIGH zero-extends.
def z_unpack_high       : SDNode<"SystemZISD::UNPACK_HIGH", SDT_ZVecUnpack>;
def z_unpackl_high      : SDNode<"SystemZISD::UNPACKL_HIGH", SDT_ZVecUnpack>;

// Likewise for the second half.
def z_unpack_low        : SDNode<"SystemZISD::UNPACK_LOW", SDT_ZVecUnpack>;
def z_unpackl_low       : SDNode<"SystemZISD::UNPACKL_LOW", SDT_ZVecUnpack>;

// Shift/rotate each element of vector operand 0 by the number of bits
// specified by scalar operand 1.
def z_vshl_by_scalar    : SDNode<"SystemZISD::VSHL_BY_SCALAR",
                                 SDT_ZVecBinaryInt>;
def z_vsrl_by_scalar    : SDNode<"SystemZISD::VSRL_BY_SCALAR",
                                 SDT_ZVecBinaryInt>;
def z_vsra_by_scalar    : SDNode<"SystemZISD::VSRA_BY_SCALAR",
                                 SDT_ZVecBinaryInt>;
def z_vrotl_by_scalar   : SDNode<"SystemZISD::VROTL_BY_SCALAR",
                                 SDT_ZVecBinaryInt>;

// For each element of the output type, sum across all sub-elements of
// operand 0 belonging to the corresponding element, and add in the
// rightmost sub-element of the corresponding element of operand 1.
def z_vsum              : SDNode<"SystemZISD::VSUM", SDT_ZBinaryConv>;

// Compare integer vector operands 0 and 1 to produce the usual 0/-1
// vector result.  VICMPE is for equality, VICMPH for "signed greater than"
// and VICMPHL for "unsigned greater than".
def z_vicmpe            : SDNode<"SystemZISD::VICMPE", SDT_ZVecCompare>;
def z_vicmph            : SDNode<"SystemZISD::VICMPH", SDT_ZVecCompare>;
def z_vicmphl           : SDNode<"SystemZISD::VICMPHL", SDT_ZVecCompare>;

// Likewise, but also set the condition codes on the result.
def z_vicmpes           : SDNode<"SystemZISD::VICMPES", SDT_ZVecCompareCC>;
def z_vicmphs           : SDNode<"SystemZISD::VICMPHS", SDT_ZVecCompareCC>;
def z_vicmphls          : SDNode<"SystemZISD::VICMPHLS", SDT_ZVecCompareCC>;

// Compare floating-point vector operands 0 and 1 to produce the usual 0/-1
// vector result.  VFCMPE is for "ordered and equal", VFCMPH for "ordered and
// greater than" and VFCMPHE for "ordered and greater than or equal to".
def z_vfcmpe            : SDNode<"SystemZISD::VFCMPE", SDT_ZVecBinaryConv>;
def z_vfcmph            : SDNode<"SystemZISD::VFCMPH", SDT_ZVecBinaryConv>;
def z_vfcmphe           : SDNode<"SystemZISD::VFCMPHE", SDT_ZVecBinaryConv>;

// Likewise, but also set the condition codes on the result.
def z_vfcmpes           : SDNode<"SystemZISD::VFCMPES", SDT_ZVecBinaryConvCC>;
def z_vfcmphs           : SDNode<"SystemZISD::VFCMPHS", SDT_ZVecBinaryConvCC>;
def z_vfcmphes          : SDNode<"SystemZISD::VFCMPHES", SDT_ZVecBinaryConvCC>;

// Extend the even f32 elements of vector operand 0 to produce a vector
// of f64 elements.
def z_vextend           : SDNode<"SystemZISD::VEXTEND", SDT_ZVecUnaryConv>;

// Round the f64 elements of vector operand 0 to f32s and store them in the
// even elements of the result.
def z_vround            : SDNode<"SystemZISD::VROUND", SDT_ZVecUnaryConv>;

let IsStrictFP = true in {
  // Strict variants of vector floating-point comparisons.
  // Quiet and signaling versions.
  def z_strict_vfcmpe   : SDNode<"SystemZISD::STRICT_VFCMPE",
                                 SDT_ZVecBinaryConv, [SDNPHasChain]>;
  def z_strict_vfcmph   : SDNode<"SystemZISD::STRICT_VFCMPH",
                                 SDT_ZVecBinaryConv, [SDNPHasChain]>;
  def z_strict_vfcmphe  : SDNode<"SystemZISD::STRICT_VFCMPHE",
                                 SDT_ZVecBinaryConv, [SDNPHasChain]>;
  def z_strict_vfcmpes  : SDNode<"SystemZISD::STRICT_VFCMPES",
                                 SDT_ZVecBinaryConv, [SDNPHasChain]>;
  def z_strict_vfcmphs  : SDNode<"SystemZISD::STRICT_VFCMPHS",
                                 SDT_ZVecBinaryConv, [SDNPHasChain]>;
  def z_strict_vfcmphes : SDNode<"SystemZISD::STRICT_VFCMPHES",
                                 SDT_ZVecBinaryConv, [SDNPHasChain]>;

  // Strict variants of VEXTEND and VROUND.
  def z_strict_vextend  : SDNode<"SystemZISD::STRICT_VEXTEND",
                                 SDT_ZVecUnaryConv, [SDNPHasChain]>;
  def z_strict_vround   : SDNode<"SystemZISD::STRICT_VROUND",
                                 SDT_ZVecUnaryConv, [SDNPHasChain]>;
}

// AND the two vector operands together and set CC based on the result.
def z_vtm               : SDNode<"SystemZISD::VTM", SDT_ZCmp>;

// i128 high integer comparisons.
def z_scmp128hi         : SDNode<"SystemZISD::SCMP128HI", SDT_ZCmp>;
def z_ucmp128hi         : SDNode<"SystemZISD::UCMP128HI", SDT_ZCmp>;

// String operations that set CC as a side-effect.
def z_vfae_cc           : SDNode<"SystemZISD::VFAE_CC", SDT_ZVecTernaryIntCC>;
def z_vfaez_cc          : SDNode<"SystemZISD::VFAEZ_CC", SDT_ZVecTernaryIntCC>;
def z_vfee_cc           : SDNode<"SystemZISD::VFEE_CC", SDT_ZVecBinaryCC>;
def z_vfeez_cc          : SDNode<"SystemZISD::VFEEZ_CC", SDT_ZVecBinaryCC>;
def z_vfene_cc          : SDNode<"SystemZISD::VFENE_CC", SDT_ZVecBinaryCC>;
def z_vfenez_cc         : SDNode<"SystemZISD::VFENEZ_CC", SDT_ZVecBinaryCC>;
def z_vistr_cc          : SDNode<"SystemZISD::VISTR_CC", SDT_ZVecUnaryCC>;
def z_vstrc_cc          : SDNode<"SystemZISD::VSTRC_CC",
                                 SDT_ZVecQuaternaryIntCC>;
def z_vstrcz_cc         : SDNode<"SystemZISD::VSTRCZ_CC",
                                 SDT_ZVecQuaternaryIntCC>;
def z_vstrs_cc          : SDNode<"SystemZISD::VSTRS_CC",
                                 SDT_ZVecTernaryConvCC>;
def z_vstrsz_cc         : SDNode<"SystemZISD::VSTRSZ_CC",
                                 SDT_ZVecTernaryConvCC>;

// Test floating-point data class for vectors.
def z_vftci             : SDNode<"SystemZISD::VFTCI", SDT_ZVecBinaryConvIntCC>;

class AtomicWOp<string name, SDTypeProfile profile = SDT_ZAtomicLoadBinaryW>
  : SDNode<"SystemZISD::"#name, profile,
           [SDNPHasChain, SDNPMayStore, SDNPMayLoad, SDNPMemOperand]>;

// Wrappers around the inner loop of an 8- or 16-bit ATOMIC_SWAP or
// ATOMIC_LOAD_<op>.
//
// Operand 0: the address of the containing 32-bit-aligned field
// Operand 1: the second operand of <op>, in the high bits of an i32
//            for everything except ATOMIC_SWAPW
// Operand 2: how many bits to rotate the i32 left to bring the first
//            operand into the high bits
// Operand 3: the negative of operand 2, for rotating the other way
// Operand 4: the width of the field in bits (8 or 16)
def z_atomic_swapw      : AtomicWOp<"ATOMIC_SWAPW">;
def z_atomic_loadw_add  : AtomicWOp<"ATOMIC_LOADW_ADD">;
def z_atomic_loadw_sub  : AtomicWOp<"ATOMIC_LOADW_SUB">;
def z_atomic_loadw_and  : AtomicWOp<"ATOMIC_LOADW_AND">;
def z_atomic_loadw_or   : AtomicWOp<"ATOMIC_LOADW_OR">;
def z_atomic_loadw_xor  : AtomicWOp<"ATOMIC_LOADW_XOR">;
def z_atomic_loadw_nand : AtomicWOp<"ATOMIC_LOADW_NAND">;
def z_atomic_loadw_min  : AtomicWOp<"ATOMIC_LOADW_MIN">;
def z_atomic_loadw_max  : AtomicWOp<"ATOMIC_LOADW_MAX">;
def z_atomic_loadw_umin : AtomicWOp<"ATOMIC_LOADW_UMIN">;
def z_atomic_loadw_umax : AtomicWOp<"ATOMIC_LOADW_UMAX">;

// Atomic compare-and-swap returning CC value.
// Val, CC, OUTCHAIN = ATOMIC_CMP_SWAP(INCHAIN, ptr, cmp, swap)
def z_atomic_cmp_swap   : SDNode<"SystemZISD::ATOMIC_CMP_SWAP",
                                 SDT_ZAtomicCmpSwap,
                                 [SDNPHasChain, SDNPMayStore, SDNPMayLoad,
                                  SDNPMemOperand]>;

// A wrapper around the inner loop of an ATOMIC_CMP_SWAP.
//
// Operand 0: the address of the containing 32-bit-aligned field
// Operand 1: the compare value, in the low bits of an i32
// Operand 2: the swap value, in the low bits of an i32
// Operand 3: how many bits to rotate the i32 left to bring the first
//            operand into the high bits
// Operand 4: the negative of operand 2, for rotating the other way
// Operand 5: the width of the field in bits (8 or 16)
def z_atomic_cmp_swapw  : SDNode<"SystemZISD::ATOMIC_CMP_SWAPW",
                                 SDT_ZAtomicCmpSwapW,
                                 [SDNPHasChain, SDNPMayStore, SDNPMayLoad,
                                  SDNPMemOperand]>;

// 128-bit atomic load.
// Val, OUTCHAIN = ATOMIC_LOAD_128(INCHAIN, ptr)
def z_atomic_load_128   : SDNode<"SystemZISD::ATOMIC_LOAD_128",
                                 SDT_ZAtomicLoad128,
                                 [SDNPHasChain, SDNPMayLoad, SDNPMemOperand]>;

// 128-bit atomic store.
// OUTCHAIN = ATOMIC_STORE_128(INCHAIN, val, ptr)
def z_atomic_store_128  : SDNode<"SystemZISD::ATOMIC_STORE_128",
                                 SDT_ZAtomicStore128,
                                 [SDNPHasChain, SDNPMayStore, SDNPMemOperand]>;

// 128-bit atomic compare-and-swap.
// Val, CC, OUTCHAIN = ATOMIC_CMP_SWAP(INCHAIN, ptr, cmp, swap)
def z_atomic_cmp_swap_128 : SDNode<"SystemZISD::ATOMIC_CMP_SWAP_128",
                                   SDT_ZAtomicCmpSwap128,
                                   [SDNPHasChain, SDNPMayStore, SDNPMayLoad,
                                    SDNPMemOperand]>;

// Use a series of MVCs to copy bytes from one memory location to another.
// The operands are:
// - the target address
// - the source address
// - the constant length
//
// This isn't a memory opcode because we'd need to attach two
// MachineMemOperands rather than one.
def z_mvc               : SDNode<"SystemZISD::MVC", SDT_ZMemMemLength,
                                 [SDNPHasChain, SDNPMayStore, SDNPMayLoad]>;

// Similar to MVC, but for logic operations (AND, OR, XOR).
def z_nc                : SDNode<"SystemZISD::NC", SDT_ZMemMemLength,
                                  [SDNPHasChain, SDNPMayStore, SDNPMayLoad]>;
def z_oc                : SDNode<"SystemZISD::OC", SDT_ZMemMemLength,
                                  [SDNPHasChain, SDNPMayStore, SDNPMayLoad]>;
def z_xc                : SDNode<"SystemZISD::XC", SDT_ZMemMemLength,
                                  [SDNPHasChain, SDNPMayStore, SDNPMayLoad]>;

// Use CLC to compare two blocks of memory, with the same comments
// as for MVC.
def z_clc               : SDNode<"SystemZISD::CLC", SDT_ZMemMemLengthCC,
                                 [SDNPHasChain, SDNPMayLoad]>;

// Use MVC to set a block of memory after storing the first byte.
def z_memset_mvc        : SDNode<"SystemZISD::MEMSET_MVC", SDT_ZMemsetMVC,
                                 [SDNPHasChain, SDNPMayStore, SDNPMayLoad]>;

// Use a CLST-based sequence to implement strcmp().  The two input operands
// are the addresses of the strings to compare.
def z_strcmp            : SDNode<"SystemZISD::STRCMP", SDT_ZStringCC,
                                 [SDNPHasChain, SDNPMayLoad]>;

// Use an MVST-based sequence to implement stpcpy().
def z_stpcpy            : SDNode<"SystemZISD::STPCPY", SDT_ZString,
                                 [SDNPHasChain, SDNPMayStore, SDNPMayLoad]>;

// Use an SRST-based sequence to search a block of memory.  The first
// operand is the end address, the second is the start, and the third
// is the character to search for.  CC is set to 1 on success and 2
// on failure.
def z_search_string     : SDNode<"SystemZISD::SEARCH_STRING", SDT_ZStringCC,
                                 [SDNPHasChain, SDNPMayLoad]>;

// Prefetch from the second operand using the 4-bit control code in
// the first operand.  The code is 1 for a load prefetch and 2 for
// a store prefetch.
def z_prefetch          : SDNode<"SystemZISD::PREFETCH", SDT_ZPrefetch,
                                 [SDNPHasChain, SDNPMayLoad, SDNPMayStore,
                                  SDNPMemOperand]>;

// Transaction begin.  The first operand is the chain, the second
// the TDB pointer, and the third the immediate control field.
// Returns CC value and chain.
def z_tbegin            : SDNode<"SystemZISD::TBEGIN", SDT_ZTBegin,
                                 [SDNPHasChain, SDNPMayStore, SDNPSideEffect]>;
def z_tbegin_nofloat    : SDNode<"SystemZISD::TBEGIN_NOFLOAT", SDT_ZTBegin,
                                 [SDNPHasChain, SDNPMayStore, SDNPSideEffect]>;

// Transaction end.  Just the chain operand.  Returns CC value and chain.
def z_tend              : SDNode<"SystemZISD::TEND", SDT_ZTEnd,
                                 [SDNPHasChain, SDNPSideEffect]>;

// z/OS XPLINK ADA Entry
// Wraps a TargetGlobalAddress that should be loaded from a function's
// AssociatedData Area (ADA). Tha ADA is passed to the function by the
// caller in the XPLink ABI defined register R5.
// Operand 0: the GlobalValue/External Symbol
// Operand 1: the ADA register
// Operand 2: the offset (0 for the first and 8 for the second element in the
// function descriptor)
def z_ada_entry         : SDNode<"SystemZISD::ADA_ENTRY",
                                  SDT_ZADAENTRY>;

def z_vshl              : SDNode<"ISD::SHL", SDT_ZVecBinary>;
def z_vsra              : SDNode<"ISD::SRA", SDT_ZVecBinary>;
def z_vsrl              : SDNode<"ISD::SRL", SDT_ZVecBinary>;

//===----------------------------------------------------------------------===//
// Pattern fragments
//===----------------------------------------------------------------------===//

def z_loadbswap16 : PatFrag<(ops node:$addr), (z_loadbswap node:$addr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_loadbswap32 : PatFrag<(ops node:$addr), (z_loadbswap node:$addr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def z_loadbswap64 : PatFrag<(ops node:$addr), (z_loadbswap node:$addr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

def z_storebswap16 : PatFrag<(ops node:$src, node:$addr),
                             (z_storebswap node:$src, node:$addr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_storebswap32 : PatFrag<(ops node:$src, node:$addr),
                             (z_storebswap node:$src, node:$addr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def z_storebswap64 : PatFrag<(ops node:$src, node:$addr),
                             (z_storebswap node:$src, node:$addr), [{
  return cast<MemIntrinsicSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

// Fragments including CC as an implicit source.
def z_br_ccmask
  : PatFrag<(ops node:$valid, node:$mask, node:$bb),
            (z_br_ccmask_1 node:$valid, node:$mask, node:$bb, CC)>;
def z_select_ccmask
  : PatFrag<(ops node:$true, node:$false, node:$valid, node:$mask),
            (z_select_ccmask_1 node:$true, node:$false,
                               node:$valid, node:$mask, CC)>;
def z_ipm : PatFrag<(ops), (z_ipm_1 CC)>;
def z_addcarry : PatFrag<(ops node:$lhs, node:$rhs),
                              (z_addcarry_1 node:$lhs, node:$rhs, CC)>;
def z_subcarry : PatFrag<(ops node:$lhs, node:$rhs),
                              (z_subcarry_1 node:$lhs, node:$rhs, CC)>;

// Signed and unsigned comparisons.
def z_scmp : PatFrag<(ops node:$a, node:$b), (z_icmp node:$a, node:$b, timm), [{
  unsigned Type = N->getConstantOperandVal(2);
  return Type != SystemZICMP::UnsignedOnly;
}]>;
def z_ucmp : PatFrag<(ops node:$a, node:$b), (z_icmp node:$a, node:$b, timm), [{
  unsigned Type = N->getConstantOperandVal(2);
  return Type != SystemZICMP::SignedOnly;
}]>;

// Register- and memory-based TEST UNDER MASK.
def z_tm_reg : PatFrag<(ops node:$a, node:$b), (z_tm node:$a, node:$b, timm)>;
def z_tm_mem : PatFrag<(ops node:$a, node:$b), (z_tm node:$a, node:$b, 0)>;

// Shifts by small immediate amounts.
def shl1 : PatFrag<(ops node:$src), (shl node:$src, (i32 1))>;
def shl2 : PatFrag<(ops node:$src), (shl node:$src, (i32 2))>;
def shl3 : PatFrag<(ops node:$src), (shl node:$src, (i32 3))>;
def shl4 : PatFrag<(ops node:$src), (shl node:$src, (i32 4))>;

// Register sign-extend operations.  Sub-32-bit values are represented as i32s.
def sext8  : PatFrag<(ops node:$src), (sext_inreg node:$src, i8)>;
def sext16 : PatFrag<(ops node:$src), (sext_inreg node:$src, i16)>;
def sext32 : PatFrag<(ops node:$src), (sext (i32 node:$src))>;

// Match extensions of an i32 to an i64, followed by an in-register sign
// extension from a sub-i32 value.
def sext8dbl : PatFrag<(ops node:$src), (sext8 (anyext node:$src))>;
def sext16dbl : PatFrag<(ops node:$src), (sext16 (anyext node:$src))>;

// Register zero-extend operations.  Sub-32-bit values are represented as i32s.
def zext8  : PatFrag<(ops node:$src), (and node:$src, 0xff)>;
def zext16 : PatFrag<(ops node:$src), (and node:$src, 0xffff)>;
def zext32 : PatFrag<(ops node:$src), (zext (i32 node:$src))>;

// Match a 64-bit value that is guaranteed to have been sign-
// or zero-extended from a 32-bit value.
def assertsext32 : PatFrag<(ops node:$src), (assertsext node:$src), [{
  return cast<VTSDNode>(N->getOperand(1))->getVT() == MVT::i32;
}]>;
def assertzext32 : PatFrag<(ops node:$src), (assertzext node:$src), [{
  return cast<VTSDNode>(N->getOperand(1))->getVT() == MVT::i32;
}]>;

// Match a load or a non-extending atomic load.
def z_load : PatFrags<(ops node:$ptr),
                      [(load node:$ptr),
                       (atomic_load node:$ptr)], [{
  if (auto *AL = dyn_cast<AtomicSDNode>(N))
    if (AL->getExtensionType() != ISD::NON_EXTLOAD)
      return false;
  return true;
}]>;

// Sign extending (atomic) loads.
def z_sextload : PatFrags<(ops node:$ptr),
                          [(unindexedload node:$ptr),
                           (atomic_load node:$ptr)], [{
  return getLoadExtType(N) == ISD::SEXTLOAD;
}]>;
def z_sextloadi8 : PatFrag<(ops node:$ptr), (z_sextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;
def z_sextloadi16 : PatFrag<(ops node:$ptr), (z_sextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_sextloadi32 : PatFrag<(ops node:$ptr), (z_sextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def z_sextloadi64 : PatFrag<(ops node:$ptr), (z_sextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

// Zero extending (atomic) loads.
def z_zextload : PatFrags<(ops node:$ptr),
                          [(unindexedload node:$ptr),
                           (atomic_load node:$ptr)], [{
  return getLoadExtType(N) == ISD::ZEXTLOAD;
}]>;
def z_zextloadi8 : PatFrag<(ops node:$ptr), (z_zextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;
def z_zextloadi16 : PatFrag<(ops node:$ptr), (z_zextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_zextloadi32 : PatFrag<(ops node:$ptr), (z_zextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def z_zextloadi64 : PatFrag<(ops node:$ptr), (z_zextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

// Extending (atomic) loads in which the extension type can be signed.
def z_asextload : PatFrags<(ops node:$ptr),
                           [(unindexedload node:$ptr),
                            (atomic_load node:$ptr)], [{
  ISD::LoadExtType ETy = getLoadExtType(N);
  return ETy == ISD::EXTLOAD || ETy == ISD::SEXTLOAD;
}]>;
def z_asextloadi8 : PatFrag<(ops node:$ptr), (z_asextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;
def z_asextloadi16 : PatFrag<(ops node:$ptr), (z_asextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_asextloadi32 : PatFrag<(ops node:$ptr), (z_asextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;

// Extending (atomic) loads in which the extension type can be unsigned.
def z_azextload : PatFrags<(ops node:$ptr),
                           [(unindexedload node:$ptr),
                           (atomic_load node:$ptr)], [{
  ISD::LoadExtType ETy = getLoadExtType(N);
  return ETy == ISD::EXTLOAD || ETy == ISD::ZEXTLOAD;
}]>;
def z_azextloadi8 : PatFrag<(ops node:$ptr), (z_azextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;
def z_azextloadi16 : PatFrag<(ops node:$ptr), (z_azextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_azextloadi32 : PatFrag<(ops node:$ptr), (z_azextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;

// Extending (atomic) loads in which the extension type doesn't matter.
def z_anyextload : PatFrags<(ops node:$ptr),
                            [(unindexedload node:$ptr),
                             (atomic_load node:$ptr)], [{
  return getLoadExtType(N) != ISD::NON_EXTLOAD;
}]>;
def z_anyextloadi8 : PatFrag<(ops node:$ptr), (z_anyextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;
def z_anyextloadi16 : PatFrag<(ops node:$ptr), (z_anyextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_anyextloadi32 : PatFrag<(ops node:$ptr), (z_anyextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def z_anyextloadi64 : PatFrag<(ops node:$ptr), (z_anyextload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

// Extending non-atomic loads in which the extension type doesn't matter.
def anyextload : PatFrag<(ops node:$ptr), (unindexedload node:$ptr), [{
  return cast<LoadSDNode>(N)->getExtensionType() != ISD::NON_EXTLOAD;
}]>;
def anyextloadi8 : PatFrag<(ops node:$ptr), (anyextload node:$ptr), [{
  return cast<LoadSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;
def anyextloadi16 : PatFrag<(ops node:$ptr), (anyextload node:$ptr), [{
  return cast<LoadSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def anyextloadi32 : PatFrag<(ops node:$ptr), (anyextload node:$ptr), [{
  return cast<LoadSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;

// Extending (atomic) loads that are not sign/zero extending.
def z_extload : PatFrags<(ops node:$ptr),
                         [(extload node:$ptr),
                          (atomic_load node:$ptr)], [{
  return getLoadExtType(N) == ISD::EXTLOAD;
}]>;
def z_extloadi8 : PatFrag<(ops node:$ptr), (z_extload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i8;
}]>;
def z_extloadi16 : PatFrag<(ops node:$ptr), (z_extload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i16;
}]>;
def z_extloadi32 : PatFrag<(ops node:$ptr), (z_extload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i32;
}]>;
def z_extloadi64 : PatFrag<(ops node:$ptr), (z_extload node:$ptr), [{
  return cast<MemSDNode>(N)->getMemoryVT() == MVT::i64;
}]>;

// Extending atomic FP loads.
def z_any_extloadf32 : PatFrags<(ops node:$ptr),
                                [(any_extloadf32 node:$ptr),
                                 (any_fpextend (f32 (atomic_load node:$ptr)))]>;
def z_any_extloadf64 : PatFrags<(ops node:$ptr),
                                [(any_extloadf64 node:$ptr),
                                 (any_fpextend (f64 (atomic_load node:$ptr)))]>;

// Aligned loads.
class AlignedLoad<SDPatternOperator load>
  : PatFrag<(ops node:$addr), (load node:$addr),
  [{ return storeLoadIsAligned(N); }]>;
def aligned_z_load         : AlignedLoad<z_load>;
def aligned_z_asextloadi16 : AlignedLoad<z_asextloadi16>;
def aligned_z_asextloadi32 : AlignedLoad<z_asextloadi32>;
def aligned_z_azextloadi16 : AlignedLoad<z_azextloadi16>;
def aligned_z_azextloadi32 : AlignedLoad<z_azextloadi32>;

// Aligned stores.
class AlignedStore<SDPatternOperator store>
  : PatFrag<(ops node:$src, node:$addr), (store node:$src, node:$addr),
  [{ return storeLoadIsAligned(N); }]>;
def aligned_store         : AlignedStore<store>;
def aligned_truncstorei16 : AlignedStore<truncstorei16>;
def aligned_truncstorei32 : AlignedStore<truncstorei32>;

// Non-volatile loads.  Used for instructions that might access the storage
// location multiple times.
class NonvolatileLoad<SDPatternOperator load>
  : PatFrag<(ops node:$addr), (load node:$addr), [{
  auto *Load = cast<LoadSDNode>(N);
  return !Load->isVolatile();
}]>;
def nonvolatile_anyextloadi8  : NonvolatileLoad<anyextloadi8>;
def nonvolatile_anyextloadi16 : NonvolatileLoad<anyextloadi16>;
def nonvolatile_anyextloadi32 : NonvolatileLoad<anyextloadi32>;

// Non-volatile stores.
class NonvolatileStore<SDPatternOperator store>
  : PatFrag<(ops node:$src, node:$addr), (store node:$src, node:$addr), [{
  auto *Store = cast<StoreSDNode>(N);
  return !Store->isVolatile();
}]>;
def nonvolatile_truncstorei8  : NonvolatileStore<truncstorei8>;
def nonvolatile_truncstorei16 : NonvolatileStore<truncstorei16>;
def nonvolatile_truncstorei32 : NonvolatileStore<truncstorei32>;

// A store of a load that can be implemented using MVC.
def mvc_store : PatFrag<(ops node:$value, node:$addr),
                        (unindexedstore node:$value, node:$addr),
                        [{ return storeLoadCanUseMVC(N); }]>;

// Binary read-modify-write operations on memory in which the other
// operand is also memory and for which block operations like NC can
// be used.  There are two patterns for each operator, depending on
// which operand contains the "other" load.
multiclass block_op<SDPatternOperator operator> {
  def "1" : PatFrag<(ops node:$value, node:$addr),
                    (unindexedstore (operator node:$value,
                                              (unindexedload node:$addr)),
                                    node:$addr),
                    [{ return storeLoadCanUseBlockBinary(N, 0); }]>;
  def "2" : PatFrag<(ops node:$value, node:$addr),
                    (unindexedstore (operator (unindexedload node:$addr),
                                              node:$value),
                                    node:$addr),
                    [{ return storeLoadCanUseBlockBinary(N, 1); }]>;
}
defm block_and : block_op<and>;
defm block_or  : block_op<or>;
defm block_xor : block_op<xor>;

// Insertions.
class insert_imm<int mask> : PatFrag<(ops node:$src1, node:$src2),
                                     (or (and node:$src1, mask), node:$src2)>;

def inserti8   : insert_imm<-256>;
def insertll   : insert_imm<0xffff0000>;
def insertlh   : insert_imm<0x0000ffff>;
def insertll64 : insert_imm<0xffffffffffff0000>;
def insertlh64 : insert_imm<0xffffffff0000ffff>;
def inserthl64 : insert_imm<0xffff0000ffffffff>;
def inserthh64 : insert_imm<0x0000ffffffffffff>;
def insertlf   : insert_imm<0xffffffff00000000>;
def inserthf   : insert_imm<0x00000000ffffffff>;

// ORs that can be treated as insertions.
def or_as_inserti8 : PatFrag<(ops node:$src1, node:$src2),
                             (or node:$src1, node:$src2), [{
  unsigned BitWidth = N->getValueType(0).getScalarSizeInBits();
  return CurDAG->MaskedValueIsZero(N->getOperand(0),
                                   APInt::getLowBitsSet(BitWidth, 8));
}]>;

// ORs that can be treated as reversed insertions.
def or_as_revinserti8 : PatFrag<(ops node:$src1, node:$src2),
                                (or node:$src1, node:$src2), [{
  unsigned BitWidth = N->getValueType(0).getScalarSizeInBits();
  return CurDAG->MaskedValueIsZero(N->getOperand(1),
                                   APInt::getLowBitsSet(BitWidth, 8));
}]>;

// Negative integer absolute.
def z_inegabs : PatFrag<(ops node:$src), (ineg (abs node:$src))>;

// Integer multiply-and-add
class z_muladd<SDPatternOperator mulop>
  : PatFrag<(ops node:$src1, node:$src2, node:$src3),
            (add (mulop node:$src1, node:$src2), node:$src3)>;

// Alternatives to match operations with or without an overflow CC result.
def z_sadd : PatFrags<(ops node:$src1, node:$src2),
                      [(z_saddo node:$src1, node:$src2),
                       (add node:$src1, node:$src2)]>;
def z_uadd : PatFrags<(ops node:$src1, node:$src2),
                      [(z_uaddo node:$src1, node:$src2),
                       (add node:$src1, node:$src2)]>;
def z_ssub : PatFrags<(ops node:$src1, node:$src2),
                      [(z_ssubo node:$src1, node:$src2),
                       (sub node:$src1, node:$src2)]>;
def z_usub : PatFrags<(ops node:$src1, node:$src2),
                      [(z_usubo node:$src1, node:$src2),
                       (sub node:$src1, node:$src2)]>;

// Combined logical operations.
def andc : PatFrag<(ops node:$src1, node:$src2),
                   (and node:$src1, (not node:$src2))>;
def orc  : PatFrag<(ops node:$src1, node:$src2),
                   (or node:$src1, (not node:$src2))>;
def nand : PatFrag<(ops node:$src1, node:$src2),
                   (not (and node:$src1, node:$src2))>;
def nor  : PatFrag<(ops node:$src1, node:$src2),
                   (not (or node:$src1, node:$src2))>;
def nxor : PatFrag<(ops node:$src1, node:$src2),
                   (not (xor node:$src1, node:$src2))>;

// Fused multiply-subtract, using the natural operand order.
def any_fms : PatFrag<(ops node:$src1, node:$src2, node:$src3),
                      (any_fma node:$src1, node:$src2, (fneg node:$src3))>;

// Fused multiply-add and multiply-subtract, but with the order of the
// operands matching SystemZ's MA and MS instructions.
def z_any_fma : PatFrag<(ops node:$src1, node:$src2, node:$src3),
                        (any_fma node:$src2, node:$src3, node:$src1)>;
def z_any_fms : PatFrag<(ops node:$src1, node:$src2, node:$src3),
                        (any_fma node:$src2, node:$src3, (fneg node:$src1))>;

// Negative fused multiply-add and multiply-subtract.
def any_fnma : PatFrag<(ops node:$src1, node:$src2, node:$src3),
                       (fneg (any_fma node:$src1, node:$src2, node:$src3))>;
def any_fnms : PatFrag<(ops node:$src1, node:$src2, node:$src3),
                       (fneg (any_fms node:$src1, node:$src2, node:$src3))>;

// Floating-point negative absolute.
def fnabs : PatFrag<(ops node:$ptr), (fneg (fabs node:$ptr))>;

// Floating-point operations which will not participate in reassociation, and
// therefore are candidates for reg/mem folding during isel.
def z_any_fadd_noreassoc : PatFrag<(ops node:$src1, node:$src2),
                                   (any_fadd node:$src1, node:$src2),
                                   [{ return !shouldSelectForReassoc(N); }]>;
def z_any_fsub_noreassoc : PatFrag<(ops node:$src1, node:$src2),
                                   (any_fsub node:$src1, node:$src2),
                                   [{ return !shouldSelectForReassoc(N); }]>;
def z_any_fmul_noreassoc : PatFrag<(ops node:$src1, node:$src2),
                                   (any_fmul node:$src1, node:$src2),
                                   [{ return !shouldSelectForReassoc(N); }]>;

// Strict floating-point fragments.
def z_any_fcmp    : PatFrags<(ops node:$lhs, node:$rhs),
                             [(z_strict_fcmp node:$lhs, node:$rhs),
                              (z_fcmp node:$lhs, node:$rhs)]>;
def z_any_vfcmpe  : PatFrags<(ops node:$lhs, node:$rhs),
                             [(z_strict_vfcmpe node:$lhs, node:$rhs),
                              (z_vfcmpe node:$lhs, node:$rhs)]>;
def z_any_vfcmph  : PatFrags<(ops node:$lhs, node:$rhs),
                             [(z_strict_vfcmph node:$lhs, node:$rhs),
                              (z_vfcmph node:$lhs, node:$rhs)]>;
def z_any_vfcmphe : PatFrags<(ops node:$lhs, node:$rhs),
                             [(z_strict_vfcmphe node:$lhs, node:$rhs),
                              (z_vfcmphe node:$lhs, node:$rhs)]>;
def z_any_vextend : PatFrags<(ops node:$src),
                             [(z_strict_vextend node:$src),
                              (z_vextend node:$src)]>;
def z_any_vround  : PatFrags<(ops node:$src),
                             [(z_strict_vround node:$src),
                              (z_vround node:$src)]>;

// Create a unary operator that loads from memory and then performs
// the given operation on it.
class loadu<SDPatternOperator operator, SDPatternOperator load = z_load>
  : PatFrag<(ops node:$addr), (operator (load node:$addr))>;

// Create a store operator that performs the given unary operation
// on the value before storing it.
class storeu<SDPatternOperator operator, SDPatternOperator store = store>
  : PatFrag<(ops node:$value, node:$addr),
            (store (operator node:$value), node:$addr)>;

// Create a store operator that performs the given inherent operation
// and stores the resulting value.
class storei<SDPatternOperator operator, SDPatternOperator store = store>
  : PatFrag<(ops node:$addr),
            (store (operator), node:$addr)>;

// Create a shift operator that optionally ignores an AND of the
// shift count with an immediate if the bottom 6 bits are all set.
def imm32bottom6set : PatLeaf<(i32 imm), [{
  return (N->getZExtValue() & 0x3f) == 0x3f;
}]>;
class shiftop<SDPatternOperator operator>
  : PatFrags<(ops node:$val, node:$count),
             [(operator node:$val, node:$count),
              (operator node:$val, (and node:$count, imm32bottom6set))]>;

// Create a shift operator that optionally ignores an AND of the
// shift count with an immediate if the bottom 7 bits are all set.
def imm32bottom7set : PatLeaf<(i32 imm), [{
  return (N->getZExtValue() & 0x7f) == 0x7f;
}]>;
class vshiftop<SDPatternOperator operator>
  : PatFrags<(ops node:$val, node:$count),
             [(operator node:$val, node:$count),
              (operator node:$val, (and node:$count, imm32bottom7set))]>;

def imm32mod64  : PatLeaf<(i32 imm), [{
  return (N->getZExtValue() % 64 == 0);
}]>;

def imm32nobits : PatLeaf<(i32 imm), [{
  return (N->getZExtValue() & 0x07) == 0;
}]>;
def imm32nobytes : PatLeaf<(i32 imm), [{
  return (N->getZExtValue() & 0x78) == 0;
}]>;

// Load a scalar and replicate it in all elements of a vector.
class z_replicate_load<ValueType scalartype, SDPatternOperator load>
  : PatFrag<(ops node:$addr),
            (z_replicate (scalartype (load node:$addr)))>;
def z_replicate_loadi8  : z_replicate_load<i32, z_anyextloadi8>;
def z_replicate_loadi16 : z_replicate_load<i32, z_anyextloadi16>;
def z_replicate_loadi32 : z_replicate_load<i32, z_load>;
def z_replicate_loadi64 : z_replicate_load<i64, z_load>;
def z_replicate_loadf32 : z_replicate_load<f32, z_load>;
def z_replicate_loadf64 : z_replicate_load<f64, z_load>;
// Byte-swapped replicated vector element loads.
def z_replicate_loadbswapi16 : z_replicate_load<i32, z_loadbswap16>;
def z_replicate_loadbswapi32 : z_replicate_load<i32, z_loadbswap32>;
def z_replicate_loadbswapi64 : z_replicate_load<i64, z_loadbswap64>;

// Load a scalar and insert it into a single element of a vector.
class z_vle<ValueType scalartype, SDPatternOperator load>
  : PatFrag<(ops node:$vec, node:$addr, node:$index),
            (z_vector_insert node:$vec, (scalartype (load node:$addr)),
                             node:$index)>;
def z_vlei8  : z_vle<i32, z_anyextloadi8>;
def z_vlei16 : z_vle<i32, z_anyextloadi16>;
def z_vlei32 : z_vle<i32, z_load>;
def z_vlei64 : z_vle<i64, z_load>;
def z_vlef32 : z_vle<f32, z_load>;
def z_vlef64 : z_vle<f64, z_load>;
// Byte-swapped vector element loads.
def z_vlebri16 : z_vle<i32, z_loadbswap16>;
def z_vlebri32 : z_vle<i32, z_loadbswap32>;
def z_vlebri64 : z_vle<i64, z_loadbswap64>;

// Load a scalar and insert it into the low element of the high i64 of a
// zeroed vector.
class z_vllez<ValueType scalartype, SDPatternOperator load, int index>
  : PatFrag<(ops node:$addr),
            (z_vector_insert immAllZerosV,
                             (scalartype (load node:$addr)), (i32 index))>;
def z_vllezi8  : z_vllez<i32, z_anyextloadi8, 7>;
def z_vllezi16 : z_vllez<i32, z_anyextloadi16, 3>;
def z_vllezi32 : z_vllez<i32, z_load, 1>;
def z_vllezi64 : PatFrags<(ops node:$addr),
                          [(z_vector_insert immAllZerosV,
                                            (i64 (z_load node:$addr)), (i32 0)),
                           (z_join_dwords (i64 (z_load node:$addr)), (i64 0))]>;
// We use high merges to form a v4f32 from four f32s.  Propagating zero
// into all elements but index 1 gives this expression.
def z_vllezf32 : PatFrag<(ops node:$addr),
                         (z_merge_high
                          (v2i64
                           (z_unpackl_high
                            (v4i32
                             (bitconvert
                              (v4f32 (scalar_to_vector
                                      (f32 (z_load node:$addr)))))))),
                          (v2i64
                           (bitconvert (v4f32 immAllZerosV))))>;
def z_vllezf64 : PatFrag<(ops node:$addr),
                         (z_merge_high
                          (v2f64 (scalar_to_vector (f64 (z_load node:$addr)))),
                          immAllZerosV)>;

// Similarly for the high element of a zeroed vector.
def z_vllezli32 : z_vllez<i32, z_load, 0>;
def z_vllezlf32 : PatFrag<(ops node:$addr),
                          (z_merge_high
                           (v2i64
                            (bitconvert
                             (z_merge_high
                              (v4f32 (scalar_to_vector
                                      (f32 (z_load node:$addr)))),
                              (v4f32 immAllZerosV)))),
                           (v2i64
                            (bitconvert (v4f32 immAllZerosV))))>;

// Byte-swapped variants.
def z_vllebrzi16  : z_vllez<i32, z_loadbswap16, 3>;
def z_vllebrzi32  : z_vllez<i32, z_loadbswap32, 1>;
def z_vllebrzli32 : z_vllez<i32, z_loadbswap32, 0>;
def z_vllebrzi64  : PatFrags<(ops node:$addr),
                             [(z_vector_insert immAllZerosV,
                                               (i64 (z_loadbswap64 node:$addr)),
                                               (i32 0)),
                              (z_join_dwords (i64 (z_loadbswap64 node:$addr)),
                                             (i64 0))]>;


// Store one element of a vector.
class z_vste<ValueType scalartype, SDPatternOperator store>
  : PatFrag<(ops node:$vec, node:$addr, node:$index),
            (store (scalartype (z_vector_extract node:$vec, node:$index)),
                   node:$addr)>;
def z_vstei8  : z_vste<i32, truncstorei8>;
def z_vstei16 : z_vste<i32, truncstorei16>;
def z_vstei32 : z_vste<i32, store>;
def z_vstei64 : z_vste<i64, store>;
def z_vstef32 : z_vste<f32, store>;
def z_vstef64 : z_vste<f64, store>;
// Byte-swapped vector element stores.
def z_vstebri16 : z_vste<i32, z_storebswap16>;
def z_vstebri32 : z_vste<i32, z_storebswap32>;
def z_vstebri64 : z_vste<i64, z_storebswap64>;

// Arithmetic negation on vectors.
def z_vneg : PatFrag<(ops node:$x), (sub immAllZerosV, node:$x)>;

// Bitwise negation on vectors.
def z_vnot : PatFrag<(ops node:$x), (xor node:$x, immAllOnesV)>;

// In-register element-wise zero extension from i1 on vectors.
def vsplat_imm_eq_1 : PatFrag<(ops), (build_vector), [{
  APInt Imm;
  return ISD::isConstantSplatVector(N, Imm) && Imm == 1;
}]>;
def z_vzext1 : PatFrag<(ops node:$x), (and node:$x, vsplat_imm_eq_1)>;

// Vector constants for saturating truncation, containing the minimum and
// maximum value for the integer type that is half of the element width.
def ssat_trunc_min_vec: PatFrag<(ops), (build_vector), [{
  APInt Imm;
  EVT EltTy = N->getValueType(0).getVectorElementType();
  unsigned SizeInBits = EltTy.getSizeInBits();
  APInt min = APInt::getSignedMinValue(SizeInBits / 2).sext(SizeInBits);
  return ISD::isConstantSplatVector(N, Imm) && APInt::isSameValue(Imm, min);
}]>;
def ssat_trunc_max_vec: PatFrag<(ops), (build_vector), [{
  APInt Imm;
  EVT EltTy = N->getValueType(0).getVectorElementType();
  unsigned SizeInBits = EltTy.getSizeInBits();
  APInt max = APInt::getSignedMaxValue(SizeInBits / 2).sext(SizeInBits);
  return ISD::isConstantSplatVector(N, Imm) && APInt::isSameValue(Imm, max);
}]>;

def usat_trunc_max_vec: PatFrag<(ops), (build_vector), [{
  APInt Imm;
  EVT EltTy = N->getValueType(0).getVectorElementType();
  unsigned SizeInBits = EltTy.getSizeInBits();
  APInt max = APInt::getMaxValue(SizeInBits / 2).zext(SizeInBits);
  return ISD::isConstantSplatVector(N, Imm) && APInt::isSameValue(Imm, max);
}]>;

// Signed "integer greater than zero" on vectors.
def z_vicmph_zero : PatFrag<(ops node:$x), (z_vicmph node:$x, immAllZerosV)>;

// Signed "integer less than zero" on vectors.
def z_vicmpl_zero : PatFrag<(ops node:$x), (z_vicmph immAllZerosV, node:$x)>;

// Sign-extend the i64 elements of a vector.
class z_vse<int shift>
  : PatFrag<(ops node:$src),
            (z_vsra_by_scalar (z_vshl_by_scalar node:$src, shift), shift)>;
def z_vsei8  : z_vse<56>;
def z_vsei16 : z_vse<48>;
def z_vsei32 : z_vse<32>;

// ...and again with the extensions being done on individual i64 scalars.
class z_vse_by_parts<SDPatternOperator operator, int index1, int index2>
  : PatFrag<(ops node:$src),
            (z_join_dwords
             (operator (z_vector_extract node:$src, index1)),
             (operator (z_vector_extract node:$src, index2)))>;
def z_vsei8_by_parts  : z_vse_by_parts<sext8dbl, 7, 15>;
def z_vsei16_by_parts : z_vse_by_parts<sext16dbl, 3, 7>;
def z_vsei32_by_parts : z_vse_by_parts<sext32, 1, 3>;