github.com/epfl-dcsl/gotee@v0.0.0-20200909122901-014b35f5e5e9/src/cmd/compile/internal/ssa/gen/AMD64Ops.go (about)

     1  // Copyright 2015 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  // +build ignore
     6  
     7  package main
     8  
     9  import "strings"
    10  
    11  // Notes:
    12  //  - Integer types live in the low portion of registers. Upper portions are junk.
    13  //  - Boolean types use the low-order byte of a register. 0=false, 1=true.
    14  //    Upper bytes are junk.
    15  //  - Floating-point types live in the low natural slot of an sse2 register.
    16  //    Unused portions are junk.
    17  //  - We do not use AH,BH,CH,DH registers.
    18  //  - When doing sub-register operations, we try to write the whole
    19  //    destination register to avoid a partial-register write.
    20  //  - Unused portions of AuxInt (or the Val portion of ValAndOff) are
    21  //    filled by sign-extending the used portion.  Users of AuxInt which interpret
    22  //    AuxInt as unsigned (e.g. shifts) must be careful.
    23  //  - All SymOff opcodes require their offset to fit in an int32.
    24  
    25  // Suffixes encode the bit width of various instructions.
    26  // Q (quad word) = 64 bit
    27  // L (long word) = 32 bit
    28  // W (word)      = 16 bit
    29  // B (byte)      = 8 bit
    30  
    31  // copied from ../../amd64/reg.go
    32  var regNamesAMD64 = []string{
    33  	"AX",
    34  	"CX",
    35  	"DX",
    36  	"BX",
    37  	"SP",
    38  	"BP",
    39  	"SI",
    40  	"DI",
    41  	"R8",
    42  	"R9",
    43  	"R10",
    44  	"R11",
    45  	"R12",
    46  	"R13",
    47  	"R14",
    48  	"R15",
    49  	"X0",
    50  	"X1",
    51  	"X2",
    52  	"X3",
    53  	"X4",
    54  	"X5",
    55  	"X6",
    56  	"X7",
    57  	"X8",
    58  	"X9",
    59  	"X10",
    60  	"X11",
    61  	"X12",
    62  	"X13",
    63  	"X14",
    64  	"X15",
    65  
    66  	// pseudo-registers
    67  	"SB",
    68  }
    69  
    70  func init() {
    71  	// Make map from reg names to reg integers.
    72  	if len(regNamesAMD64) > 64 {
    73  		panic("too many registers")
    74  	}
    75  	num := map[string]int{}
    76  	for i, name := range regNamesAMD64 {
    77  		num[name] = i
    78  	}
    79  	buildReg := func(s string) regMask {
    80  		m := regMask(0)
    81  		for _, r := range strings.Split(s, " ") {
    82  			if n, ok := num[r]; ok {
    83  				m |= regMask(1) << uint(n)
    84  				continue
    85  			}
    86  			panic("register " + r + " not found")
    87  		}
    88  		return m
    89  	}
    90  
    91  	// Common individual register masks
    92  	var (
    93  		ax         = buildReg("AX")
    94  		cx         = buildReg("CX")
    95  		dx         = buildReg("DX")
    96  		gp         = buildReg("AX CX DX BX BP SI DI R8 R9 R10 R11 R12 R13 R14 R15")
    97  		fp         = buildReg("X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15")
    98  		gpsp       = gp | buildReg("SP")
    99  		gpspsb     = gpsp | buildReg("SB")
   100  		callerSave = gp | fp
   101  	)
   102  	// Common slices of register masks
   103  	var (
   104  		gponly = []regMask{gp}
   105  		fponly = []regMask{fp}
   106  	)
   107  
   108  	// Common regInfo
   109  	var (
   110  		gp01      = regInfo{inputs: nil, outputs: gponly}
   111  		gp11      = regInfo{inputs: []regMask{gp}, outputs: gponly}
   112  		gp11sp    = regInfo{inputs: []regMask{gpsp}, outputs: gponly}
   113  		gp11sb    = regInfo{inputs: []regMask{gpspsb}, outputs: gponly}
   114  		gp21      = regInfo{inputs: []regMask{gp, gp}, outputs: gponly}
   115  		gp21sp    = regInfo{inputs: []regMask{gpsp, gp}, outputs: gponly}
   116  		gp21sb    = regInfo{inputs: []regMask{gpspsb, gpsp}, outputs: gponly}
   117  		gp21shift = regInfo{inputs: []regMask{gp, cx}, outputs: []regMask{gp}}
   118  		gp11div   = regInfo{inputs: []regMask{ax, gpsp &^ dx}, outputs: []regMask{ax, dx}}
   119  		gp21hmul  = regInfo{inputs: []regMask{ax, gpsp}, outputs: []regMask{dx}, clobbers: ax}
   120  
   121  		gp2flags = regInfo{inputs: []regMask{gpsp, gpsp}}
   122  		gp1flags = regInfo{inputs: []regMask{gpsp}}
   123  		flagsgp  = regInfo{inputs: nil, outputs: gponly}
   124  
   125  		gp11flags = regInfo{inputs: []regMask{gp}, outputs: []regMask{gp, 0}}
   126  
   127  		readflags = regInfo{inputs: nil, outputs: gponly}
   128  		flagsgpax = regInfo{inputs: nil, clobbers: ax, outputs: []regMask{gp &^ ax}}
   129  
   130  		gpload    = regInfo{inputs: []regMask{gpspsb, 0}, outputs: gponly}
   131  		gp21load  = regInfo{inputs: []regMask{gp, gpspsb, 0}, outputs: gponly}
   132  		gploadidx = regInfo{inputs: []regMask{gpspsb, gpsp, 0}, outputs: gponly}
   133  
   134  		gpstore         = regInfo{inputs: []regMask{gpspsb, gpsp, 0}}
   135  		gpstoreconst    = regInfo{inputs: []regMask{gpspsb, 0}}
   136  		gpstoreidx      = regInfo{inputs: []regMask{gpspsb, gpsp, gpsp, 0}}
   137  		gpstoreconstidx = regInfo{inputs: []regMask{gpspsb, gpsp, 0}}
   138  		gpstorexchg     = regInfo{inputs: []regMask{gp, gpspsb, 0}, outputs: []regMask{gp}}
   139  		cmpxchg         = regInfo{inputs: []regMask{gp, ax, gp, 0}, outputs: []regMask{gp, 0}, clobbers: ax}
   140  
   141  		fp01     = regInfo{inputs: nil, outputs: fponly}
   142  		fp21     = regInfo{inputs: []regMask{fp, fp}, outputs: fponly}
   143  		fp21load = regInfo{inputs: []regMask{fp, gpspsb, 0}, outputs: fponly}
   144  		fpgp     = regInfo{inputs: fponly, outputs: gponly}
   145  		gpfp     = regInfo{inputs: gponly, outputs: fponly}
   146  		fp11     = regInfo{inputs: fponly, outputs: fponly}
   147  		fp2flags = regInfo{inputs: []regMask{fp, fp}}
   148  
   149  		fpload    = regInfo{inputs: []regMask{gpspsb, 0}, outputs: fponly}
   150  		fploadidx = regInfo{inputs: []regMask{gpspsb, gpsp, 0}, outputs: fponly}
   151  
   152  		fpstore    = regInfo{inputs: []regMask{gpspsb, fp, 0}}
   153  		fpstoreidx = regInfo{inputs: []regMask{gpspsb, gpsp, fp, 0}}
   154  	)
   155  
   156  	var AMD64ops = []opData{
   157  		// fp ops
   158  		{name: "ADDSS", argLength: 2, reg: fp21, asm: "ADDSS", commutative: true, resultInArg0: true}, // fp32 add
   159  		{name: "ADDSD", argLength: 2, reg: fp21, asm: "ADDSD", commutative: true, resultInArg0: true}, // fp64 add
   160  		{name: "SUBSS", argLength: 2, reg: fp21, asm: "SUBSS", resultInArg0: true},                    // fp32 sub
   161  		{name: "SUBSD", argLength: 2, reg: fp21, asm: "SUBSD", resultInArg0: true},                    // fp64 sub
   162  		{name: "MULSS", argLength: 2, reg: fp21, asm: "MULSS", commutative: true, resultInArg0: true}, // fp32 mul
   163  		{name: "MULSD", argLength: 2, reg: fp21, asm: "MULSD", commutative: true, resultInArg0: true}, // fp64 mul
   164  		{name: "DIVSS", argLength: 2, reg: fp21, asm: "DIVSS", resultInArg0: true},                    // fp32 div
   165  		{name: "DIVSD", argLength: 2, reg: fp21, asm: "DIVSD", resultInArg0: true},                    // fp64 div
   166  
   167  		{name: "MOVSSload", argLength: 2, reg: fpload, asm: "MOVSS", aux: "SymOff", faultOnNilArg0: true, symEffect: "Read"}, // fp32 load
   168  		{name: "MOVSDload", argLength: 2, reg: fpload, asm: "MOVSD", aux: "SymOff", faultOnNilArg0: true, symEffect: "Read"}, // fp64 load
   169  		{name: "MOVSSconst", reg: fp01, asm: "MOVSS", aux: "Float32", rematerializeable: true},                               // fp32 constant
   170  		{name: "MOVSDconst", reg: fp01, asm: "MOVSD", aux: "Float64", rematerializeable: true},                               // fp64 constant
   171  		{name: "MOVSSloadidx1", argLength: 3, reg: fploadidx, asm: "MOVSS", aux: "SymOff", symEffect: "Read"},                // fp32 load indexed by i
   172  		{name: "MOVSSloadidx4", argLength: 3, reg: fploadidx, asm: "MOVSS", aux: "SymOff", symEffect: "Read"},                // fp32 load indexed by 4*i
   173  		{name: "MOVSDloadidx1", argLength: 3, reg: fploadidx, asm: "MOVSD", aux: "SymOff", symEffect: "Read"},                // fp64 load indexed by i
   174  		{name: "MOVSDloadidx8", argLength: 3, reg: fploadidx, asm: "MOVSD", aux: "SymOff", symEffect: "Read"},                // fp64 load indexed by 8*i
   175  
   176  		{name: "MOVSSstore", argLength: 3, reg: fpstore, asm: "MOVSS", aux: "SymOff", faultOnNilArg0: true, symEffect: "Write"}, // fp32 store
   177  		{name: "MOVSDstore", argLength: 3, reg: fpstore, asm: "MOVSD", aux: "SymOff", faultOnNilArg0: true, symEffect: "Write"}, // fp64 store
   178  		{name: "MOVSSstoreidx1", argLength: 4, reg: fpstoreidx, asm: "MOVSS", aux: "SymOff", symEffect: "Write"},                // fp32 indexed by i store
   179  		{name: "MOVSSstoreidx4", argLength: 4, reg: fpstoreidx, asm: "MOVSS", aux: "SymOff", symEffect: "Write"},                // fp32 indexed by 4i store
   180  		{name: "MOVSDstoreidx1", argLength: 4, reg: fpstoreidx, asm: "MOVSD", aux: "SymOff", symEffect: "Write"},                // fp64 indexed by i store
   181  		{name: "MOVSDstoreidx8", argLength: 4, reg: fpstoreidx, asm: "MOVSD", aux: "SymOff", symEffect: "Write"},                // fp64 indexed by 8i store
   182  
   183  		{name: "ADDSSmem", argLength: 3, reg: fp21load, asm: "ADDSS", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, symEffect: "Read"}, // fp32 arg0 + tmp, tmp loaded from arg1+auxint+aux, arg2 = mem
   184  		{name: "ADDSDmem", argLength: 3, reg: fp21load, asm: "ADDSD", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, symEffect: "Read"}, // fp64 arg0 + tmp, tmp loaded from arg1+auxint+aux, arg2 = mem
   185  		{name: "SUBSSmem", argLength: 3, reg: fp21load, asm: "SUBSS", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, symEffect: "Read"}, // fp32 arg0 - tmp, tmp loaded from arg1+auxint+aux, arg2 = mem
   186  		{name: "SUBSDmem", argLength: 3, reg: fp21load, asm: "SUBSD", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, symEffect: "Read"}, // fp64 arg0 - tmp, tmp loaded from arg1+auxint+aux, arg2 = mem
   187  		{name: "MULSSmem", argLength: 3, reg: fp21load, asm: "MULSS", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, symEffect: "Read"}, // fp32 arg0 * tmp, tmp loaded from arg1+auxint+aux, arg2 = mem
   188  		{name: "MULSDmem", argLength: 3, reg: fp21load, asm: "MULSD", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, symEffect: "Read"}, // fp64 arg0 * tmp, tmp loaded from arg1+auxint+aux, arg2 = mem
   189  
   190  		// binary ops
   191  		{name: "ADDQ", argLength: 2, reg: gp21sp, asm: "ADDQ", commutative: true, clobberFlags: true},                                                           // arg0 + arg1
   192  		{name: "ADDL", argLength: 2, reg: gp21sp, asm: "ADDL", commutative: true, clobberFlags: true},                                                           // arg0 + arg1
   193  		{name: "ADDQconst", argLength: 1, reg: gp11sp, asm: "ADDQ", aux: "Int32", typ: "UInt64", clobberFlags: true},                                            // arg0 + auxint
   194  		{name: "ADDLconst", argLength: 1, reg: gp11sp, asm: "ADDL", aux: "Int32", clobberFlags: true},                                                           // arg0 + auxint
   195  		{name: "ADDQconstmem", argLength: 2, reg: gpstoreconst, asm: "ADDQ", aux: "SymValAndOff", clobberFlags: true, faultOnNilArg0: true, symEffect: "Write"}, // add ValAndOff(AuxInt).Val() to arg0+ValAndOff(AuxInt).Off()+aux, arg1=mem
   196  		{name: "ADDLconstmem", argLength: 2, reg: gpstoreconst, asm: "ADDL", aux: "SymValAndOff", clobberFlags: true, faultOnNilArg0: true, symEffect: "Write"}, // add ValAndOff(AuxInt).Val() to arg0+ValAndOff(AuxInt).Off()+aux, arg1=mem
   197  
   198  		{name: "SUBQ", argLength: 2, reg: gp21, asm: "SUBQ", resultInArg0: true, clobberFlags: true},                    // arg0 - arg1
   199  		{name: "SUBL", argLength: 2, reg: gp21, asm: "SUBL", resultInArg0: true, clobberFlags: true},                    // arg0 - arg1
   200  		{name: "SUBQconst", argLength: 1, reg: gp11, asm: "SUBQ", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 - auxint
   201  		{name: "SUBLconst", argLength: 1, reg: gp11, asm: "SUBL", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 - auxint
   202  
   203  		{name: "MULQ", argLength: 2, reg: gp21, asm: "IMULQ", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 * arg1
   204  		{name: "MULL", argLength: 2, reg: gp21, asm: "IMULL", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 * arg1
   205  		{name: "MULQconst", argLength: 1, reg: gp11, asm: "IMULQ", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 * auxint
   206  		{name: "MULLconst", argLength: 1, reg: gp11, asm: "IMULL", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 * auxint
   207  
   208  		{name: "HMULQ", argLength: 2, reg: gp21hmul, commutative: true, asm: "IMULQ", clobberFlags: true}, // (arg0 * arg1) >> width
   209  		{name: "HMULL", argLength: 2, reg: gp21hmul, commutative: true, asm: "IMULL", clobberFlags: true}, // (arg0 * arg1) >> width
   210  		{name: "HMULQU", argLength: 2, reg: gp21hmul, commutative: true, asm: "MULQ", clobberFlags: true}, // (arg0 * arg1) >> width
   211  		{name: "HMULLU", argLength: 2, reg: gp21hmul, commutative: true, asm: "MULL", clobberFlags: true}, // (arg0 * arg1) >> width
   212  
   213  		{name: "AVGQU", argLength: 2, reg: gp21, commutative: true, resultInArg0: true, clobberFlags: true}, // (arg0 + arg1) / 2 as unsigned, all 64 result bits
   214  
   215  		{name: "DIVQ", argLength: 2, reg: gp11div, typ: "(Int64,Int64)", asm: "IDIVQ", clobberFlags: true},   // [arg0 / arg1, arg0 % arg1]
   216  		{name: "DIVL", argLength: 2, reg: gp11div, typ: "(Int32,Int32)", asm: "IDIVL", clobberFlags: true},   // [arg0 / arg1, arg0 % arg1]
   217  		{name: "DIVW", argLength: 2, reg: gp11div, typ: "(Int16,Int16)", asm: "IDIVW", clobberFlags: true},   // [arg0 / arg1, arg0 % arg1]
   218  		{name: "DIVQU", argLength: 2, reg: gp11div, typ: "(UInt64,UInt64)", asm: "DIVQ", clobberFlags: true}, // [arg0 / arg1, arg0 % arg1]
   219  		{name: "DIVLU", argLength: 2, reg: gp11div, typ: "(UInt32,UInt32)", asm: "DIVL", clobberFlags: true}, // [arg0 / arg1, arg0 % arg1]
   220  		{name: "DIVWU", argLength: 2, reg: gp11div, typ: "(UInt16,UInt16)", asm: "DIVW", clobberFlags: true}, // [arg0 / arg1, arg0 % arg1]
   221  
   222  		{name: "MULQU2", argLength: 2, reg: regInfo{inputs: []regMask{ax, gpsp}, outputs: []regMask{dx, ax}}, commutative: true, asm: "MULQ", clobberFlags: true}, // arg0 * arg1, returns (hi, lo)
   223  		{name: "DIVQU2", argLength: 3, reg: regInfo{inputs: []regMask{dx, ax, gpsp}, outputs: []regMask{ax, dx}}, asm: "DIVQ", clobberFlags: true},                // arg0:arg1 / arg2 (128-bit divided by 64-bit), returns (q, r)
   224  
   225  		{name: "ANDQ", argLength: 2, reg: gp21, asm: "ANDQ", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 & arg1
   226  		{name: "ANDL", argLength: 2, reg: gp21, asm: "ANDL", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 & arg1
   227  		{name: "ANDQconst", argLength: 1, reg: gp11, asm: "ANDQ", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 & auxint
   228  		{name: "ANDLconst", argLength: 1, reg: gp11, asm: "ANDL", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 & auxint
   229  
   230  		{name: "ORQ", argLength: 2, reg: gp21, asm: "ORQ", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 | arg1
   231  		{name: "ORL", argLength: 2, reg: gp21, asm: "ORL", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 | arg1
   232  		{name: "ORQconst", argLength: 1, reg: gp11, asm: "ORQ", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 | auxint
   233  		{name: "ORLconst", argLength: 1, reg: gp11, asm: "ORL", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 | auxint
   234  
   235  		{name: "XORQ", argLength: 2, reg: gp21, asm: "XORQ", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 ^ arg1
   236  		{name: "XORL", argLength: 2, reg: gp21, asm: "XORL", commutative: true, resultInArg0: true, clobberFlags: true}, // arg0 ^ arg1
   237  		{name: "XORQconst", argLength: 1, reg: gp11, asm: "XORQ", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 ^ auxint
   238  		{name: "XORLconst", argLength: 1, reg: gp11, asm: "XORL", aux: "Int32", resultInArg0: true, clobberFlags: true}, // arg0 ^ auxint
   239  
   240  		{name: "CMPQ", argLength: 2, reg: gp2flags, asm: "CMPQ", typ: "Flags"},                    // arg0 compare to arg1
   241  		{name: "CMPL", argLength: 2, reg: gp2flags, asm: "CMPL", typ: "Flags"},                    // arg0 compare to arg1
   242  		{name: "CMPW", argLength: 2, reg: gp2flags, asm: "CMPW", typ: "Flags"},                    // arg0 compare to arg1
   243  		{name: "CMPB", argLength: 2, reg: gp2flags, asm: "CMPB", typ: "Flags"},                    // arg0 compare to arg1
   244  		{name: "CMPQconst", argLength: 1, reg: gp1flags, asm: "CMPQ", typ: "Flags", aux: "Int32"}, // arg0 compare to auxint
   245  		{name: "CMPLconst", argLength: 1, reg: gp1flags, asm: "CMPL", typ: "Flags", aux: "Int32"}, // arg0 compare to auxint
   246  		{name: "CMPWconst", argLength: 1, reg: gp1flags, asm: "CMPW", typ: "Flags", aux: "Int16"}, // arg0 compare to auxint
   247  		{name: "CMPBconst", argLength: 1, reg: gp1flags, asm: "CMPB", typ: "Flags", aux: "Int8"},  // arg0 compare to auxint
   248  
   249  		{name: "UCOMISS", argLength: 2, reg: fp2flags, asm: "UCOMISS", typ: "Flags"}, // arg0 compare to arg1, f32
   250  		{name: "UCOMISD", argLength: 2, reg: fp2flags, asm: "UCOMISD", typ: "Flags"}, // arg0 compare to arg1, f64
   251  
   252  		{name: "BTL", argLength: 2, reg: gp2flags, asm: "BTL", typ: "Flags"},                   // test whether bit arg0 % 32 in arg1 is set
   253  		{name: "BTQ", argLength: 2, reg: gp2flags, asm: "BTQ", typ: "Flags"},                   // test whether bit arg0 % 64 in arg1 is set
   254  		{name: "BTLconst", argLength: 1, reg: gp1flags, asm: "BTL", typ: "Flags", aux: "Int8"}, // test whether bit auxint in arg0 is set, 0 <= auxint < 32
   255  		{name: "BTQconst", argLength: 1, reg: gp1flags, asm: "BTQ", typ: "Flags", aux: "Int8"}, // test whether bit auxint in arg0 is set, 0 <= auxint < 64
   256  
   257  		{name: "TESTQ", argLength: 2, reg: gp2flags, commutative: true, asm: "TESTQ", typ: "Flags"}, // (arg0 & arg1) compare to 0
   258  		{name: "TESTL", argLength: 2, reg: gp2flags, commutative: true, asm: "TESTL", typ: "Flags"}, // (arg0 & arg1) compare to 0
   259  		{name: "TESTW", argLength: 2, reg: gp2flags, commutative: true, asm: "TESTW", typ: "Flags"}, // (arg0 & arg1) compare to 0
   260  		{name: "TESTB", argLength: 2, reg: gp2flags, commutative: true, asm: "TESTB", typ: "Flags"}, // (arg0 & arg1) compare to 0
   261  		{name: "TESTQconst", argLength: 1, reg: gp1flags, asm: "TESTQ", typ: "Flags", aux: "Int32"}, // (arg0 & auxint) compare to 0
   262  		{name: "TESTLconst", argLength: 1, reg: gp1flags, asm: "TESTL", typ: "Flags", aux: "Int32"}, // (arg0 & auxint) compare to 0
   263  		{name: "TESTWconst", argLength: 1, reg: gp1flags, asm: "TESTW", typ: "Flags", aux: "Int16"}, // (arg0 & auxint) compare to 0
   264  		{name: "TESTBconst", argLength: 1, reg: gp1flags, asm: "TESTB", typ: "Flags", aux: "Int8"},  // (arg0 & auxint) compare to 0
   265  
   266  		{name: "SHLQ", argLength: 2, reg: gp21shift, asm: "SHLQ", resultInArg0: true, clobberFlags: true},              // arg0 << arg1, shift amount is mod 64
   267  		{name: "SHLL", argLength: 2, reg: gp21shift, asm: "SHLL", resultInArg0: true, clobberFlags: true},              // arg0 << arg1, shift amount is mod 32
   268  		{name: "SHLQconst", argLength: 1, reg: gp11, asm: "SHLQ", aux: "Int8", resultInArg0: true, clobberFlags: true}, // arg0 << auxint, shift amount 0-63
   269  		{name: "SHLLconst", argLength: 1, reg: gp11, asm: "SHLL", aux: "Int8", resultInArg0: true, clobberFlags: true}, // arg0 << auxint, shift amount 0-31
   270  		// Note: x86 is weird, the 16 and 8 byte shifts still use all 5 bits of shift amount!
   271  
   272  		{name: "SHRQ", argLength: 2, reg: gp21shift, asm: "SHRQ", resultInArg0: true, clobberFlags: true},              // unsigned arg0 >> arg1, shift amount is mod 64
   273  		{name: "SHRL", argLength: 2, reg: gp21shift, asm: "SHRL", resultInArg0: true, clobberFlags: true},              // unsigned uint32(arg0) >> arg1, shift amount is mod 32
   274  		{name: "SHRW", argLength: 2, reg: gp21shift, asm: "SHRW", resultInArg0: true, clobberFlags: true},              // unsigned uint16(arg0) >> arg1, shift amount is mod 32
   275  		{name: "SHRB", argLength: 2, reg: gp21shift, asm: "SHRB", resultInArg0: true, clobberFlags: true},              // unsigned uint8(arg0) >> arg1, shift amount is mod 32
   276  		{name: "SHRQconst", argLength: 1, reg: gp11, asm: "SHRQ", aux: "Int8", resultInArg0: true, clobberFlags: true}, // unsigned arg0 >> auxint, shift amount 0-63
   277  		{name: "SHRLconst", argLength: 1, reg: gp11, asm: "SHRL", aux: "Int8", resultInArg0: true, clobberFlags: true}, // unsigned uint32(arg0) >> auxint, shift amount 0-31
   278  		{name: "SHRWconst", argLength: 1, reg: gp11, asm: "SHRW", aux: "Int8", resultInArg0: true, clobberFlags: true}, // unsigned uint16(arg0) >> auxint, shift amount 0-15
   279  		{name: "SHRBconst", argLength: 1, reg: gp11, asm: "SHRB", aux: "Int8", resultInArg0: true, clobberFlags: true}, // unsigned uint8(arg0) >> auxint, shift amount 0-7
   280  
   281  		{name: "SARQ", argLength: 2, reg: gp21shift, asm: "SARQ", resultInArg0: true, clobberFlags: true},              // signed arg0 >> arg1, shift amount is mod 64
   282  		{name: "SARL", argLength: 2, reg: gp21shift, asm: "SARL", resultInArg0: true, clobberFlags: true},              // signed int32(arg0) >> arg1, shift amount is mod 32
   283  		{name: "SARW", argLength: 2, reg: gp21shift, asm: "SARW", resultInArg0: true, clobberFlags: true},              // signed int16(arg0) >> arg1, shift amount is mod 32
   284  		{name: "SARB", argLength: 2, reg: gp21shift, asm: "SARB", resultInArg0: true, clobberFlags: true},              // signed int8(arg0) >> arg1, shift amount is mod 32
   285  		{name: "SARQconst", argLength: 1, reg: gp11, asm: "SARQ", aux: "Int8", resultInArg0: true, clobberFlags: true}, // signed arg0 >> auxint, shift amount 0-63
   286  		{name: "SARLconst", argLength: 1, reg: gp11, asm: "SARL", aux: "Int8", resultInArg0: true, clobberFlags: true}, // signed int32(arg0) >> auxint, shift amount 0-31
   287  		{name: "SARWconst", argLength: 1, reg: gp11, asm: "SARW", aux: "Int8", resultInArg0: true, clobberFlags: true}, // signed int16(arg0) >> auxint, shift amount 0-15
   288  		{name: "SARBconst", argLength: 1, reg: gp11, asm: "SARB", aux: "Int8", resultInArg0: true, clobberFlags: true}, // signed int8(arg0) >> auxint, shift amount 0-7
   289  
   290  		{name: "ROLQ", argLength: 2, reg: gp21shift, asm: "ROLQ", resultInArg0: true, clobberFlags: true},              // arg0 rotate left arg1 bits.
   291  		{name: "ROLL", argLength: 2, reg: gp21shift, asm: "ROLL", resultInArg0: true, clobberFlags: true},              // arg0 rotate left arg1 bits.
   292  		{name: "ROLW", argLength: 2, reg: gp21shift, asm: "ROLW", resultInArg0: true, clobberFlags: true},              // arg0 rotate left arg1 bits.
   293  		{name: "ROLB", argLength: 2, reg: gp21shift, asm: "ROLB", resultInArg0: true, clobberFlags: true},              // arg0 rotate left arg1 bits.
   294  		{name: "RORQ", argLength: 2, reg: gp21shift, asm: "RORQ", resultInArg0: true, clobberFlags: true},              // arg0 rotate right arg1 bits.
   295  		{name: "RORL", argLength: 2, reg: gp21shift, asm: "RORL", resultInArg0: true, clobberFlags: true},              // arg0 rotate right arg1 bits.
   296  		{name: "RORW", argLength: 2, reg: gp21shift, asm: "RORW", resultInArg0: true, clobberFlags: true},              // arg0 rotate right arg1 bits.
   297  		{name: "RORB", argLength: 2, reg: gp21shift, asm: "RORB", resultInArg0: true, clobberFlags: true},              // arg0 rotate right arg1 bits.
   298  		{name: "ROLQconst", argLength: 1, reg: gp11, asm: "ROLQ", aux: "Int8", resultInArg0: true, clobberFlags: true}, // arg0 rotate left auxint, rotate amount 0-63
   299  		{name: "ROLLconst", argLength: 1, reg: gp11, asm: "ROLL", aux: "Int8", resultInArg0: true, clobberFlags: true}, // arg0 rotate left auxint, rotate amount 0-31
   300  		{name: "ROLWconst", argLength: 1, reg: gp11, asm: "ROLW", aux: "Int8", resultInArg0: true, clobberFlags: true}, // arg0 rotate left auxint, rotate amount 0-15
   301  		{name: "ROLBconst", argLength: 1, reg: gp11, asm: "ROLB", aux: "Int8", resultInArg0: true, clobberFlags: true}, // arg0 rotate left auxint, rotate amount 0-7
   302  
   303  		{name: "ADDLmem", argLength: 3, reg: gp21load, asm: "ADDL", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 + tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   304  		{name: "ADDQmem", argLength: 3, reg: gp21load, asm: "ADDQ", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 + tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   305  		{name: "SUBQmem", argLength: 3, reg: gp21load, asm: "SUBQ", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 - tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   306  		{name: "SUBLmem", argLength: 3, reg: gp21load, asm: "SUBL", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 - tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   307  		{name: "ANDLmem", argLength: 3, reg: gp21load, asm: "ANDL", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 & tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   308  		{name: "ANDQmem", argLength: 3, reg: gp21load, asm: "ANDQ", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 & tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   309  		{name: "ORQmem", argLength: 3, reg: gp21load, asm: "ORQ", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"},   // arg0 | tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   310  		{name: "ORLmem", argLength: 3, reg: gp21load, asm: "ORL", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"},   // arg0 | tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   311  		{name: "XORQmem", argLength: 3, reg: gp21load, asm: "XORQ", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 ^ tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   312  		{name: "XORLmem", argLength: 3, reg: gp21load, asm: "XORL", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, symEffect: "Read"}, // arg0 ^ tmp, tmp loaded from  arg1+auxint+aux, arg2 = mem
   313  
   314  		// unary ops
   315  		{name: "NEGQ", argLength: 1, reg: gp11, asm: "NEGQ", resultInArg0: true, clobberFlags: true}, // -arg0
   316  		{name: "NEGL", argLength: 1, reg: gp11, asm: "NEGL", resultInArg0: true, clobberFlags: true}, // -arg0
   317  
   318  		{name: "NOTQ", argLength: 1, reg: gp11, asm: "NOTQ", resultInArg0: true, clobberFlags: true}, // ^arg0
   319  		{name: "NOTL", argLength: 1, reg: gp11, asm: "NOTL", resultInArg0: true, clobberFlags: true}, // ^arg0
   320  
   321  		// BSF{L,Q} returns a tuple [result, flags]
   322  		// result is undefined if the input is zero.
   323  		// flags are set to "equal" if the input is zero, "not equal" otherwise.
   324  		{name: "BSFQ", argLength: 1, reg: gp11flags, asm: "BSFQ", typ: "(UInt64,Flags)"}, // # of low-order zeroes in 64-bit arg
   325  		{name: "BSFL", argLength: 1, reg: gp11flags, asm: "BSFL", typ: "(UInt32,Flags)"}, // # of low-order zeroes in 32-bit arg
   326  		{name: "BSRQ", argLength: 1, reg: gp11flags, asm: "BSRQ", typ: "(UInt64,Flags)"}, // # of high-order zeroes in 64-bit arg
   327  		{name: "BSRL", argLength: 1, reg: gp11flags, asm: "BSRL", typ: "(UInt32,Flags)"}, // # of high-order zeroes in 32-bit arg
   328  
   329  		// Note ASM for ops moves whole register
   330  		//
   331  		{name: "CMOVQEQ", argLength: 3, reg: gp21, asm: "CMOVQEQ", resultInArg0: true}, // if arg2 encodes "equal" return arg1 else arg0
   332  		{name: "CMOVLEQ", argLength: 3, reg: gp21, asm: "CMOVLEQ", resultInArg0: true}, // if arg2 encodes "equal" return arg1 else arg0
   333  
   334  		{name: "BSWAPQ", argLength: 1, reg: gp11, asm: "BSWAPQ", resultInArg0: true, clobberFlags: true}, // arg0 swap bytes
   335  		{name: "BSWAPL", argLength: 1, reg: gp11, asm: "BSWAPL", resultInArg0: true, clobberFlags: true}, // arg0 swap bytes
   336  
   337  		// POPCNT instructions aren't guaranteed to be on the target platform (they are SSE4).
   338  		// Any use must be preceded by a successful check of runtime.support_popcnt.
   339  		{name: "POPCNTQ", argLength: 1, reg: gp11, asm: "POPCNTQ", clobberFlags: true}, // count number of set bits in arg0
   340  		{name: "POPCNTL", argLength: 1, reg: gp11, asm: "POPCNTL", clobberFlags: true}, // count number of set bits in arg0
   341  
   342  		{name: "SQRTSD", argLength: 1, reg: fp11, asm: "SQRTSD"}, // sqrt(arg0)
   343  
   344  		// ROUNDSD instruction isn't guaranteed to be on the target platform (it is SSE4.1)
   345  		// Any use must be preceded by a successful check of runtime.support_sse41.
   346  		{name: "ROUNDSD", argLength: 1, reg: fp11, aux: "Int8", asm: "ROUNDSD"}, // rounds arg0 depending on auxint, 1 means math.Floor, 2 Ceil, 3 Trunc
   347  
   348  		{name: "SBBQcarrymask", argLength: 1, reg: flagsgp, asm: "SBBQ"}, // (int64)(-1) if carry is set, 0 if carry is clear.
   349  		{name: "SBBLcarrymask", argLength: 1, reg: flagsgp, asm: "SBBL"}, // (int32)(-1) if carry is set, 0 if carry is clear.
   350  		// Note: SBBW and SBBB are subsumed by SBBL
   351  
   352  		{name: "SETEQ", argLength: 1, reg: readflags, asm: "SETEQ"}, // extract == condition from arg0
   353  		{name: "SETNE", argLength: 1, reg: readflags, asm: "SETNE"}, // extract != condition from arg0
   354  		{name: "SETL", argLength: 1, reg: readflags, asm: "SETLT"},  // extract signed < condition from arg0
   355  		{name: "SETLE", argLength: 1, reg: readflags, asm: "SETLE"}, // extract signed <= condition from arg0
   356  		{name: "SETG", argLength: 1, reg: readflags, asm: "SETGT"},  // extract signed > condition from arg0
   357  		{name: "SETGE", argLength: 1, reg: readflags, asm: "SETGE"}, // extract signed >= condition from arg0
   358  		{name: "SETB", argLength: 1, reg: readflags, asm: "SETCS"},  // extract unsigned < condition from arg0
   359  		{name: "SETBE", argLength: 1, reg: readflags, asm: "SETLS"}, // extract unsigned <= condition from arg0
   360  		{name: "SETA", argLength: 1, reg: readflags, asm: "SETHI"},  // extract unsigned > condition from arg0
   361  		{name: "SETAE", argLength: 1, reg: readflags, asm: "SETCC"}, // extract unsigned >= condition from arg0
   362  		// Variants that store result to memory
   363  		{name: "SETEQmem", argLength: 3, reg: gpstoreconst, asm: "SETEQ", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // extract == condition from arg1 to arg0+auxint+aux, arg2=mem
   364  		{name: "SETNEmem", argLength: 3, reg: gpstoreconst, asm: "SETNE", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // extract != condition from arg1 to arg0+auxint+aux, arg2=mem
   365  		{name: "SETLmem", argLength: 3, reg: gpstoreconst, asm: "SETLT", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},  // extract signed < condition from arg1 to arg0+auxint+aux, arg2=mem
   366  		{name: "SETLEmem", argLength: 3, reg: gpstoreconst, asm: "SETLE", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // extract signed <= condition from arg1 to arg0+auxint+aux, arg2=mem
   367  		{name: "SETGmem", argLength: 3, reg: gpstoreconst, asm: "SETGT", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},  // extract signed > condition from arg1 to arg0+auxint+aux, arg2=mem
   368  		{name: "SETGEmem", argLength: 3, reg: gpstoreconst, asm: "SETGE", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // extract signed >= condition from arg1 to arg0+auxint+aux, arg2=mem
   369  		{name: "SETBmem", argLength: 3, reg: gpstoreconst, asm: "SETCS", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},  // extract unsigned < condition from arg1 to arg0+auxint+aux, arg2=mem
   370  		{name: "SETBEmem", argLength: 3, reg: gpstoreconst, asm: "SETLS", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // extract unsigned <= condition from arg1 to arg0+auxint+aux, arg2=mem
   371  		{name: "SETAmem", argLength: 3, reg: gpstoreconst, asm: "SETHI", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},  // extract unsigned > condition from arg1 to arg0+auxint+aux, arg2=mem
   372  		{name: "SETAEmem", argLength: 3, reg: gpstoreconst, asm: "SETCC", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // extract unsigned >= condition from arg1 to arg0+auxint+aux, arg2=mem
   373  		// Need different opcodes for floating point conditions because
   374  		// any comparison involving a NaN is always FALSE and thus
   375  		// the patterns for inverting conditions cannot be used.
   376  		{name: "SETEQF", argLength: 1, reg: flagsgpax, asm: "SETEQ", clobberFlags: true}, // extract == condition from arg0
   377  		{name: "SETNEF", argLength: 1, reg: flagsgpax, asm: "SETNE", clobberFlags: true}, // extract != condition from arg0
   378  		{name: "SETORD", argLength: 1, reg: flagsgp, asm: "SETPC"},                       // extract "ordered" (No Nan present) condition from arg0
   379  		{name: "SETNAN", argLength: 1, reg: flagsgp, asm: "SETPS"},                       // extract "unordered" (Nan present) condition from arg0
   380  
   381  		{name: "SETGF", argLength: 1, reg: flagsgp, asm: "SETHI"},  // extract floating > condition from arg0
   382  		{name: "SETGEF", argLength: 1, reg: flagsgp, asm: "SETCC"}, // extract floating >= condition from arg0
   383  
   384  		{name: "MOVBQSX", argLength: 1, reg: gp11, asm: "MOVBQSX"}, // sign extend arg0 from int8 to int64
   385  		{name: "MOVBQZX", argLength: 1, reg: gp11, asm: "MOVBLZX"}, // zero extend arg0 from int8 to int64
   386  		{name: "MOVWQSX", argLength: 1, reg: gp11, asm: "MOVWQSX"}, // sign extend arg0 from int16 to int64
   387  		{name: "MOVWQZX", argLength: 1, reg: gp11, asm: "MOVWLZX"}, // zero extend arg0 from int16 to int64
   388  		{name: "MOVLQSX", argLength: 1, reg: gp11, asm: "MOVLQSX"}, // sign extend arg0 from int32 to int64
   389  		{name: "MOVLQZX", argLength: 1, reg: gp11, asm: "MOVL"},    // zero extend arg0 from int32 to int64
   390  
   391  		{name: "MOVLconst", reg: gp01, asm: "MOVL", typ: "UInt32", aux: "Int32", rematerializeable: true}, // 32 low bits of auxint
   392  		{name: "MOVQconst", reg: gp01, asm: "MOVQ", typ: "UInt64", aux: "Int64", rematerializeable: true}, // auxint
   393  
   394  		{name: "CVTTSD2SL", argLength: 1, reg: fpgp, asm: "CVTTSD2SL"}, // convert float64 to int32
   395  		{name: "CVTTSD2SQ", argLength: 1, reg: fpgp, asm: "CVTTSD2SQ"}, // convert float64 to int64
   396  		{name: "CVTTSS2SL", argLength: 1, reg: fpgp, asm: "CVTTSS2SL"}, // convert float32 to int32
   397  		{name: "CVTTSS2SQ", argLength: 1, reg: fpgp, asm: "CVTTSS2SQ"}, // convert float32 to int64
   398  		{name: "CVTSL2SS", argLength: 1, reg: gpfp, asm: "CVTSL2SS"},   // convert int32 to float32
   399  		{name: "CVTSL2SD", argLength: 1, reg: gpfp, asm: "CVTSL2SD"},   // convert int32 to float64
   400  		{name: "CVTSQ2SS", argLength: 1, reg: gpfp, asm: "CVTSQ2SS"},   // convert int64 to float32
   401  		{name: "CVTSQ2SD", argLength: 1, reg: gpfp, asm: "CVTSQ2SD"},   // convert int64 to float64
   402  		{name: "CVTSD2SS", argLength: 1, reg: fp11, asm: "CVTSD2SS"},   // convert float64 to float32
   403  		{name: "CVTSS2SD", argLength: 1, reg: fp11, asm: "CVTSS2SD"},   // convert float32 to float64
   404  
   405  		// Move values between int and float registers, with no conversion.
   406  		// TODO: should we have generic versions of these?
   407  		{name: "MOVQi2f", argLength: 1, reg: gpfp, typ: "Float64"}, // move 64 bits from int to float reg
   408  		{name: "MOVQf2i", argLength: 1, reg: fpgp, typ: "UInt64"},  // move 64 bits from float to int reg
   409  		{name: "MOVLi2f", argLength: 1, reg: gpfp, typ: "Float32"}, // move 32 bits from int to float reg
   410  		{name: "MOVLf2i", argLength: 1, reg: fpgp, typ: "UInt32"},  // move 32 bits from float to int reg, zero extend
   411  
   412  		{name: "PXOR", argLength: 2, reg: fp21, asm: "PXOR", commutative: true, resultInArg0: true}, // exclusive or, applied to X regs for float negation.
   413  
   414  		{name: "LEAQ", argLength: 1, reg: gp11sb, asm: "LEAQ", aux: "SymOff", rematerializeable: true, symEffect: "Addr"}, // arg0 + auxint + offset encoded in aux
   415  		{name: "LEAQ1", argLength: 2, reg: gp21sb, commutative: true, aux: "SymOff", symEffect: "Addr"},                   // arg0 + arg1 + auxint + aux
   416  		{name: "LEAQ2", argLength: 2, reg: gp21sb, aux: "SymOff", symEffect: "Addr"},                                      // arg0 + 2*arg1 + auxint + aux
   417  		{name: "LEAQ4", argLength: 2, reg: gp21sb, aux: "SymOff", symEffect: "Addr"},                                      // arg0 + 4*arg1 + auxint + aux
   418  		{name: "LEAQ8", argLength: 2, reg: gp21sb, aux: "SymOff", symEffect: "Addr"},                                      // arg0 + 8*arg1 + auxint + aux
   419  		// Note: LEAQ{1,2,4,8} must not have OpSB as either argument.
   420  
   421  		{name: "LEAL", argLength: 1, reg: gp11sb, asm: "LEAL", aux: "SymOff", rematerializeable: true, symEffect: "Addr"}, // arg0 + auxint + offset encoded in aux
   422  
   423  		// auxint+aux == add auxint and the offset of the symbol in aux (if any) to the effective address
   424  		{name: "MOVBload", argLength: 2, reg: gpload, asm: "MOVBLZX", aux: "SymOff", typ: "UInt8", faultOnNilArg0: true, symEffect: "Read"},  // load byte from arg0+auxint+aux. arg1=mem.  Zero extend.
   425  		{name: "MOVBQSXload", argLength: 2, reg: gpload, asm: "MOVBQSX", aux: "SymOff", faultOnNilArg0: true, symEffect: "Read"},             // ditto, sign extend to int64
   426  		{name: "MOVWload", argLength: 2, reg: gpload, asm: "MOVWLZX", aux: "SymOff", typ: "UInt16", faultOnNilArg0: true, symEffect: "Read"}, // load 2 bytes from arg0+auxint+aux. arg1=mem.  Zero extend.
   427  		{name: "MOVWQSXload", argLength: 2, reg: gpload, asm: "MOVWQSX", aux: "SymOff", faultOnNilArg0: true, symEffect: "Read"},             // ditto, sign extend to int64
   428  		{name: "MOVLload", argLength: 2, reg: gpload, asm: "MOVL", aux: "SymOff", typ: "UInt32", faultOnNilArg0: true, symEffect: "Read"},    // load 4 bytes from arg0+auxint+aux. arg1=mem.  Zero extend.
   429  		{name: "MOVLQSXload", argLength: 2, reg: gpload, asm: "MOVLQSX", aux: "SymOff", faultOnNilArg0: true, symEffect: "Read"},             // ditto, sign extend to int64
   430  		{name: "MOVQload", argLength: 2, reg: gpload, asm: "MOVQ", aux: "SymOff", typ: "UInt64", faultOnNilArg0: true, symEffect: "Read"},    // load 8 bytes from arg0+auxint+aux. arg1=mem
   431  		{name: "MOVBstore", argLength: 3, reg: gpstore, asm: "MOVB", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},    // store byte in arg1 to arg0+auxint+aux. arg2=mem
   432  		{name: "MOVWstore", argLength: 3, reg: gpstore, asm: "MOVW", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},    // store 2 bytes in arg1 to arg0+auxint+aux. arg2=mem
   433  		{name: "MOVLstore", argLength: 3, reg: gpstore, asm: "MOVL", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},    // store 4 bytes in arg1 to arg0+auxint+aux. arg2=mem
   434  		{name: "MOVQstore", argLength: 3, reg: gpstore, asm: "MOVQ", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},    // store 8 bytes in arg1 to arg0+auxint+aux. arg2=mem
   435  		{name: "MOVOload", argLength: 2, reg: fpload, asm: "MOVUPS", aux: "SymOff", typ: "Int128", faultOnNilArg0: true, symEffect: "Read"},  // load 16 bytes from arg0+auxint+aux. arg1=mem
   436  		{name: "MOVOstore", argLength: 3, reg: fpstore, asm: "MOVUPS", aux: "SymOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"},  // store 16 bytes in arg1 to arg0+auxint+aux. arg2=mem
   437  
   438  		// indexed loads/stores
   439  		{name: "MOVBloadidx1", argLength: 3, reg: gploadidx, commutative: true, asm: "MOVBLZX", aux: "SymOff", typ: "UInt8", symEffect: "Read"},  // load a byte from arg0+arg1+auxint+aux. arg2=mem
   440  		{name: "MOVWloadidx1", argLength: 3, reg: gploadidx, commutative: true, asm: "MOVWLZX", aux: "SymOff", typ: "UInt16", symEffect: "Read"}, // load 2 bytes from arg0+arg1+auxint+aux. arg2=mem
   441  		{name: "MOVWloadidx2", argLength: 3, reg: gploadidx, asm: "MOVWLZX", aux: "SymOff", typ: "UInt16", symEffect: "Read"},                    // load 2 bytes from arg0+2*arg1+auxint+aux. arg2=mem
   442  		{name: "MOVLloadidx1", argLength: 3, reg: gploadidx, commutative: true, asm: "MOVL", aux: "SymOff", typ: "UInt32", symEffect: "Read"},    // load 4 bytes from arg0+arg1+auxint+aux. arg2=mem
   443  		{name: "MOVLloadidx4", argLength: 3, reg: gploadidx, asm: "MOVL", aux: "SymOff", typ: "UInt32", symEffect: "Read"},                       // load 4 bytes from arg0+4*arg1+auxint+aux. arg2=mem
   444  		{name: "MOVLloadidx8", argLength: 3, reg: gploadidx, asm: "MOVL", aux: "SymOff", typ: "UInt32", symEffect: "Read"},                       // load 4 bytes from arg0+8*arg1+auxint+aux. arg2=mem
   445  		{name: "MOVQloadidx1", argLength: 3, reg: gploadidx, commutative: true, asm: "MOVQ", aux: "SymOff", typ: "UInt64", symEffect: "Read"},    // load 8 bytes from arg0+arg1+auxint+aux. arg2=mem
   446  		{name: "MOVQloadidx8", argLength: 3, reg: gploadidx, asm: "MOVQ", aux: "SymOff", typ: "UInt64", symEffect: "Read"},                       // load 8 bytes from arg0+8*arg1+auxint+aux. arg2=mem
   447  		// TODO: sign-extending indexed loads
   448  		// TODO: mark the MOVXstoreidx1 ops as commutative.  Generates too many rewrite rules at the moment.
   449  		{name: "MOVBstoreidx1", argLength: 4, reg: gpstoreidx, asm: "MOVB", aux: "SymOff", symEffect: "Write"}, // store byte in arg2 to arg0+arg1+auxint+aux. arg3=mem
   450  		{name: "MOVWstoreidx1", argLength: 4, reg: gpstoreidx, asm: "MOVW", aux: "SymOff", symEffect: "Write"}, // store 2 bytes in arg2 to arg0+arg1+auxint+aux. arg3=mem
   451  		{name: "MOVWstoreidx2", argLength: 4, reg: gpstoreidx, asm: "MOVW", aux: "SymOff", symEffect: "Write"}, // store 2 bytes in arg2 to arg0+2*arg1+auxint+aux. arg3=mem
   452  		{name: "MOVLstoreidx1", argLength: 4, reg: gpstoreidx, asm: "MOVL", aux: "SymOff", symEffect: "Write"}, // store 4 bytes in arg2 to arg0+arg1+auxint+aux. arg3=mem
   453  		{name: "MOVLstoreidx4", argLength: 4, reg: gpstoreidx, asm: "MOVL", aux: "SymOff", symEffect: "Write"}, // store 4 bytes in arg2 to arg0+4*arg1+auxint+aux. arg3=mem
   454  		{name: "MOVLstoreidx8", argLength: 4, reg: gpstoreidx, asm: "MOVL", aux: "SymOff", symEffect: "Write"}, // store 4 bytes in arg2 to arg0+8*arg1+auxint+aux. arg3=mem
   455  		{name: "MOVQstoreidx1", argLength: 4, reg: gpstoreidx, asm: "MOVQ", aux: "SymOff", symEffect: "Write"}, // store 8 bytes in arg2 to arg0+arg1+auxint+aux. arg3=mem
   456  		{name: "MOVQstoreidx8", argLength: 4, reg: gpstoreidx, asm: "MOVQ", aux: "SymOff", symEffect: "Write"}, // store 8 bytes in arg2 to arg0+8*arg1+auxint+aux. arg3=mem
   457  		// TODO: add size-mismatched indexed loads, like MOVBstoreidx4.
   458  
   459  		// For storeconst ops, the AuxInt field encodes both
   460  		// the value to store and an address offset of the store.
   461  		// Cast AuxInt to a ValAndOff to extract Val and Off fields.
   462  		{name: "MOVBstoreconst", argLength: 2, reg: gpstoreconst, asm: "MOVB", aux: "SymValAndOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // store low byte of ValAndOff(AuxInt).Val() to arg0+ValAndOff(AuxInt).Off()+aux.  arg1=mem
   463  		{name: "MOVWstoreconst", argLength: 2, reg: gpstoreconst, asm: "MOVW", aux: "SymValAndOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // store low 2 bytes of ...
   464  		{name: "MOVLstoreconst", argLength: 2, reg: gpstoreconst, asm: "MOVL", aux: "SymValAndOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // store low 4 bytes of ...
   465  		{name: "MOVQstoreconst", argLength: 2, reg: gpstoreconst, asm: "MOVQ", aux: "SymValAndOff", typ: "Mem", faultOnNilArg0: true, symEffect: "Write"}, // store 8 bytes of ...
   466  
   467  		{name: "MOVBstoreconstidx1", argLength: 3, reg: gpstoreconstidx, asm: "MOVB", aux: "SymValAndOff", typ: "Mem", symEffect: "Write"}, // store low byte of ValAndOff(AuxInt).Val() to arg0+1*arg1+ValAndOff(AuxInt).Off()+aux.  arg2=mem
   468  		{name: "MOVWstoreconstidx1", argLength: 3, reg: gpstoreconstidx, asm: "MOVW", aux: "SymValAndOff", typ: "Mem", symEffect: "Write"}, // store low 2 bytes of ... arg1 ...
   469  		{name: "MOVWstoreconstidx2", argLength: 3, reg: gpstoreconstidx, asm: "MOVW", aux: "SymValAndOff", typ: "Mem", symEffect: "Write"}, // store low 2 bytes of ... 2*arg1 ...
   470  		{name: "MOVLstoreconstidx1", argLength: 3, reg: gpstoreconstidx, asm: "MOVL", aux: "SymValAndOff", typ: "Mem", symEffect: "Write"}, // store low 4 bytes of ... arg1 ...
   471  		{name: "MOVLstoreconstidx4", argLength: 3, reg: gpstoreconstidx, asm: "MOVL", aux: "SymValAndOff", typ: "Mem", symEffect: "Write"}, // store low 4 bytes of ... 4*arg1 ...
   472  		{name: "MOVQstoreconstidx1", argLength: 3, reg: gpstoreconstidx, asm: "MOVQ", aux: "SymValAndOff", typ: "Mem", symEffect: "Write"}, // store 8 bytes of ... arg1 ...
   473  		{name: "MOVQstoreconstidx8", argLength: 3, reg: gpstoreconstidx, asm: "MOVQ", aux: "SymValAndOff", typ: "Mem", symEffect: "Write"}, // store 8 bytes of ... 8*arg1 ...
   474  
   475  		// arg0 = pointer to start of memory to zero
   476  		// arg1 = value to store (will always be zero)
   477  		// arg2 = mem
   478  		// auxint = # of bytes to zero
   479  		// returns mem
   480  		{
   481  			name:      "DUFFZERO",
   482  			aux:       "Int64",
   483  			argLength: 3,
   484  			reg: regInfo{
   485  				inputs:   []regMask{buildReg("DI"), buildReg("X0")},
   486  				clobbers: buildReg("DI"),
   487  			},
   488  			faultOnNilArg0: true,
   489  		},
   490  		{name: "MOVOconst", reg: regInfo{nil, 0, []regMask{fp}}, typ: "Int128", aux: "Int128", rematerializeable: true},
   491  
   492  		// arg0 = address of memory to zero
   493  		// arg1 = # of 8-byte words to zero
   494  		// arg2 = value to store (will always be zero)
   495  		// arg3 = mem
   496  		// returns mem
   497  		{
   498  			name:      "REPSTOSQ",
   499  			argLength: 4,
   500  			reg: regInfo{
   501  				inputs:   []regMask{buildReg("DI"), buildReg("CX"), buildReg("AX")},
   502  				clobbers: buildReg("DI CX"),
   503  			},
   504  			faultOnNilArg0: true,
   505  		},
   506  
   507  		{name: "CALLstatic", argLength: 1, reg: regInfo{clobbers: callerSave}, aux: "SymOff", clobberFlags: true, call: true, symEffect: "None"},                          // call static function aux.(*obj.LSym).  arg0=mem, auxint=argsize, returns mem
   508  		{name: "CALLclosure", argLength: 3, reg: regInfo{inputs: []regMask{gpsp, buildReg("DX"), 0}, clobbers: callerSave}, aux: "Int64", clobberFlags: true, call: true}, // call function via closure.  arg0=codeptr, arg1=closure, arg2=mem, auxint=argsize, returns mem
   509  		{name: "CALLinter", argLength: 2, reg: regInfo{inputs: []regMask{gp}, clobbers: callerSave}, aux: "Int64", clobberFlags: true, call: true},                        // call fn by pointer.  arg0=codeptr, arg1=mem, auxint=argsize, returns mem
   510  
   511  		// arg0 = destination pointer
   512  		// arg1 = source pointer
   513  		// arg2 = mem
   514  		// auxint = offset from duffcopy symbol to call
   515  		// returns memory
   516  		{
   517  			name:      "DUFFCOPY",
   518  			aux:       "Int64",
   519  			argLength: 3,
   520  			reg: regInfo{
   521  				inputs:   []regMask{buildReg("DI"), buildReg("SI")},
   522  				clobbers: buildReg("DI SI X0"), // uses X0 as a temporary
   523  			},
   524  			clobberFlags:   true,
   525  			faultOnNilArg0: true,
   526  			faultOnNilArg1: true,
   527  		},
   528  
   529  		// arg0 = destination pointer
   530  		// arg1 = source pointer
   531  		// arg2 = # of 8-byte words to copy
   532  		// arg3 = mem
   533  		// returns memory
   534  		{
   535  			name:      "REPMOVSQ",
   536  			argLength: 4,
   537  			reg: regInfo{
   538  				inputs:   []regMask{buildReg("DI"), buildReg("SI"), buildReg("CX")},
   539  				clobbers: buildReg("DI SI CX"),
   540  			},
   541  			faultOnNilArg0: true,
   542  			faultOnNilArg1: true,
   543  		},
   544  
   545  		// (InvertFlags (CMPQ a b)) == (CMPQ b a)
   546  		// So if we want (SETL (CMPQ a b)) but we can't do that because a is a constant,
   547  		// then we do (SETL (InvertFlags (CMPQ b a))) instead.
   548  		// Rewrites will convert this to (SETG (CMPQ b a)).
   549  		// InvertFlags is a pseudo-op which can't appear in assembly output.
   550  		{name: "InvertFlags", argLength: 1}, // reverse direction of arg0
   551  
   552  		// Pseudo-ops
   553  		{name: "LoweredGetG", argLength: 1, reg: gp01}, // arg0=mem
   554  		// Scheduler ensures LoweredGetClosurePtr occurs only in entry block,
   555  		// and sorts it to the very beginning of the block to prevent other
   556  		// use of DX (the closure pointer)
   557  		{name: "LoweredGetClosurePtr", reg: regInfo{outputs: []regMask{buildReg("DX")}}},
   558  		// LoweredGetCallerPC evaluates to the PC to which its "caller" will return.
   559  		// I.e., if f calls g "calls" getcallerpc,
   560  		// the result should be the PC within f that g will return to.
   561  		// See runtime/stubs.go for a more detailed discussion.
   562  		{name: "LoweredGetCallerPC", reg: gp01, rematerializeable: true},
   563  		// LoweredGetCallerSP returns the SP of the caller of the current function.
   564  		{name: "LoweredGetCallerSP", reg: gp01, rematerializeable: true},
   565  		//arg0=ptr,arg1=mem, returns void.  Faults if ptr is nil.
   566  		{name: "LoweredNilCheck", argLength: 2, reg: regInfo{inputs: []regMask{gpsp}}, clobberFlags: true, nilCheck: true, faultOnNilArg0: true},
   567  
   568  		// LoweredWB invokes runtime.gcWriteBarrier. arg0=destptr, arg1=srcptr, arg2=mem, aux=runtime.gcWriteBarrier
   569  		// It saves all GP registers if necessary, but may clobber others.
   570  		{name: "LoweredWB", argLength: 3, reg: regInfo{inputs: []regMask{buildReg("DI"), ax}, clobbers: callerSave ^ gp}, clobberFlags: true, aux: "Sym", symEffect: "None"},
   571  
   572  		// MOVQconvert converts between pointers and integers.
   573  		// We have a special op for this so as to not confuse GC
   574  		// (particularly stack maps).  It takes a memory arg so it
   575  		// gets correctly ordered with respect to GC safepoints.
   576  		// arg0=ptr/int arg1=mem, output=int/ptr
   577  		{name: "MOVQconvert", argLength: 2, reg: gp11, asm: "MOVQ", resultInArg0: true},
   578  		{name: "MOVLconvert", argLength: 2, reg: gp11, asm: "MOVL", resultInArg0: true}, // amd64p32 equivalent
   579  
   580  		// Constant flag values. For any comparison, there are 5 possible
   581  		// outcomes: the three from the signed total order (<,==,>) and the
   582  		// three from the unsigned total order. The == cases overlap.
   583  		// Note: there's a sixth "unordered" outcome for floating-point
   584  		// comparisons, but we don't use such a beast yet.
   585  		// These ops are for temporary use by rewrite rules. They
   586  		// cannot appear in the generated assembly.
   587  		{name: "FlagEQ"},     // equal
   588  		{name: "FlagLT_ULT"}, // signed < and unsigned <
   589  		{name: "FlagLT_UGT"}, // signed < and unsigned >
   590  		{name: "FlagGT_UGT"}, // signed > and unsigned >
   591  		{name: "FlagGT_ULT"}, // signed > and unsigned <
   592  
   593  		// Atomic loads.  These are just normal loads but return <value,memory> tuples
   594  		// so they can be properly ordered with other loads.
   595  		// load from arg0+auxint+aux.  arg1=mem.
   596  		{name: "MOVLatomicload", argLength: 2, reg: gpload, asm: "MOVL", aux: "SymOff", faultOnNilArg0: true, symEffect: "Read"},
   597  		{name: "MOVQatomicload", argLength: 2, reg: gpload, asm: "MOVQ", aux: "SymOff", faultOnNilArg0: true, symEffect: "Read"},
   598  
   599  		// Atomic stores and exchanges.  Stores use XCHG to get the right memory ordering semantics.
   600  		// store arg0 to arg1+auxint+aux, arg2=mem.
   601  		// These ops return a tuple of <old contents of *(arg1+auxint+aux), memory>.
   602  		// Note: arg0 and arg1 are backwards compared to MOVLstore (to facilitate resultInArg0)!
   603  		{name: "XCHGL", argLength: 3, reg: gpstorexchg, asm: "XCHGL", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, hasSideEffects: true, symEffect: "RdWr"},
   604  		{name: "XCHGQ", argLength: 3, reg: gpstorexchg, asm: "XCHGQ", aux: "SymOff", resultInArg0: true, faultOnNilArg1: true, hasSideEffects: true, symEffect: "RdWr"},
   605  
   606  		// Atomic adds.
   607  		// *(arg1+auxint+aux) += arg0.  arg2=mem.
   608  		// Returns a tuple of <old contents of *(arg1+auxint+aux), memory>.
   609  		// Note: arg0 and arg1 are backwards compared to MOVLstore (to facilitate resultInArg0)!
   610  		{name: "XADDLlock", argLength: 3, reg: gpstorexchg, asm: "XADDL", typ: "(UInt32,Mem)", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, hasSideEffects: true, symEffect: "RdWr"},
   611  		{name: "XADDQlock", argLength: 3, reg: gpstorexchg, asm: "XADDQ", typ: "(UInt64,Mem)", aux: "SymOff", resultInArg0: true, clobberFlags: true, faultOnNilArg1: true, hasSideEffects: true, symEffect: "RdWr"},
   612  		{name: "AddTupleFirst32", argLength: 2}, // arg1=tuple <x,y>.  Returns <x+arg0,y>.
   613  		{name: "AddTupleFirst64", argLength: 2}, // arg1=tuple <x,y>.  Returns <x+arg0,y>.
   614  
   615  		// Compare and swap.
   616  		// arg0 = pointer, arg1 = old value, arg2 = new value, arg3 = memory.
   617  		// if *(arg0+auxint+aux) == arg1 {
   618  		//   *(arg0+auxint+aux) = arg2
   619  		//   return (true, memory)
   620  		// } else {
   621  		//   return (false, memory)
   622  		// }
   623  		// Note that these instructions also return the old value in AX, but we ignore it.
   624  		// TODO: have these return flags instead of bool.  The current system generates:
   625  		//    CMPXCHGQ ...
   626  		//    SETEQ AX
   627  		//    CMPB  AX, $0
   628  		//    JNE ...
   629  		// instead of just
   630  		//    CMPXCHGQ ...
   631  		//    JEQ ...
   632  		// but we can't do that because memory-using ops can't generate flags yet
   633  		// (flagalloc wants to move flag-generating instructions around).
   634  		{name: "CMPXCHGLlock", argLength: 4, reg: cmpxchg, asm: "CMPXCHGL", aux: "SymOff", clobberFlags: true, faultOnNilArg0: true, hasSideEffects: true, symEffect: "RdWr"},
   635  		{name: "CMPXCHGQlock", argLength: 4, reg: cmpxchg, asm: "CMPXCHGQ", aux: "SymOff", clobberFlags: true, faultOnNilArg0: true, hasSideEffects: true, symEffect: "RdWr"},
   636  
   637  		// Atomic memory updates.
   638  		{name: "ANDBlock", argLength: 3, reg: gpstore, asm: "ANDB", aux: "SymOff", clobberFlags: true, faultOnNilArg0: true, hasSideEffects: true, symEffect: "RdWr"}, // *(arg0+auxint+aux) &= arg1
   639  		{name: "ORBlock", argLength: 3, reg: gpstore, asm: "ORB", aux: "SymOff", clobberFlags: true, faultOnNilArg0: true, hasSideEffects: true, symEffect: "RdWr"},   // *(arg0+auxint+aux) |= arg1
   640  	}
   641  
   642  	var AMD64blocks = []blockData{
   643  		{name: "EQ"},
   644  		{name: "NE"},
   645  		{name: "LT"},
   646  		{name: "LE"},
   647  		{name: "GT"},
   648  		{name: "GE"},
   649  		{name: "ULT"},
   650  		{name: "ULE"},
   651  		{name: "UGT"},
   652  		{name: "UGE"},
   653  		{name: "EQF"},
   654  		{name: "NEF"},
   655  		{name: "ORD"}, // FP, ordered comparison (parity zero)
   656  		{name: "NAN"}, // FP, unordered comparison (parity one)
   657  	}
   658  
   659  	archs = append(archs, arch{
   660  		name:            "AMD64",
   661  		pkg:             "cmd/internal/obj/x86",
   662  		genfile:         "../../amd64/ssa.go",
   663  		ops:             AMD64ops,
   664  		blocks:          AMD64blocks,
   665  		regnames:        regNamesAMD64,
   666  		gpregmask:       gp,
   667  		fpregmask:       fp,
   668  		framepointerreg: int8(num["BP"]),
   669  		linkreg:         -1, // not used
   670  	})
   671  }