"MODUD",
	"MODUW",
	"MODSD",
	"MODSW",
	"EQV",
	"EQVCC",
	"EXTSB",
	"EXTSBCC",
	"EXTSH",
	"EXTSHCC",
	"FABS",
	"FABSCC",
	"FADD",
	"FADDCC",
	"FADDS",
	"FADDSCC",
	"FCMPO",
	"FCMPU",
	"FCTIW",
	"FCTIWCC",
	"FCTIWZ",
	"FCTIWZCC",
	"FDIV",
	"FDIVCC",
	"FDIVS",
	"FDIVSCC",
	"FMADD",
	"FMADDCC",
	"FMADDS",
	"FMADDSCC",
	"FMOVD",
	"FMOVDCC",
	"FMOVDU",
	"FMOVS",

	CMP R3, R4                      // 7c232000
	CMP R3, R0                      // 7c230000
	CMP R3, R0, CR1                 // CMP R3,CR1,R0   // 7ca30000
	CMPU R3, R4                     // 7c232040
	CMPU R3, R0                     // 7c230040
	CMPU R3, R0, CR2                // CMPU R3,CR2,R0  // 7d230040
	CMPW R3, R4                     // 7c032000
	CMPW R3, R0                     // 7c030000

			// guard against underflow.
			//
			//	CMPU	SP, $(framesize-StackSmall)
			//	BLT	label-of-call-to-morestack
			if offset <= 0xffff {
				p = obj.Appendp(p, c.newprog)
				p.As = ACMPU
				p.From.Type = obj.TYPE_REG
				p.From.Reg = REGSP
				p.To.Type = obj.TYPE_CONST
				p.To.Offset = offset
			} else {
				// Constant is too big for CMPU.
				p = obj.Appendp(p, c.newprog)
				p.As = AMOVD

TEXT ·xorBytes(SB), NOSPLIT, $0
	MOVD	dst+0(FP), R3	// R3 = dst
	MOVD	a+8(FP), R4	// R4 = a
	MOVD	b+16(FP), R5	// R5 = b
	MOVD	n+24(FP), R6	// R6 = n

	CMPU	R6, $64, CR7	// Check if n ≥ 64 bytes
	MOVD	R0, R8		// R8 = index
	CMPU	R6, $8, CR6	// Check if 8 ≤ n < 64 bytes
	BLE	CR6, small	// <= 8
	BLT	CR7, xor32	// Case for 32 ≤ n < 64 bytes

	// Case for n ≥ 64 bytes
preloop64:

(Neq(32|64)F x y) => (NotEqual (FCMPU x y))

(Less(8|16) x y)  => (LessThan (CMPW (SignExt(8|16)to32 x) (SignExt(8|16)to32 y)))
(Less(32|64) x y) => (LessThan ((CMPW|CMP) x y))
(Less(32|64)F x y) => (FLessThan (FCMPU x y))

(Less(8|16)U x y)  => (LessThan (CMPWU (ZeroExt(8|16)to32 x) (ZeroExt(8|16)to32 y)))
(Less(32|64)U x y) => (LessThan ((CMPWU|CMPU) x y))

	XXPERMDI VS44, VS46, $3, VS47
	XXPERMDI VS61, VS62, $0, VS44
	XXPERMDI VS61, VS62, $3, VS46

	VADDUWM V0, V16, V0
	VADDUWM V4, V17, V4
	VADDUWM V8, V18, V8
	VADDUWM V12, V19, V12

	CMPU LEN, $64
	BLT tail_vsx

	// Bottom of loop
	LXVW4X (INP)(R0), VS59
	LXVW4X (INP)(R8), VS60
	LXVW4X (INP)(R9), VS61
	LXVW4X (INP)(R10), VS62

	VXOR V27, V0, V27
	VXOR V28, V4, V28

#ifdef NEEDS_ESPERM
	MOVD	$·rcon(SB), R7
	LVX	(R7), ESPERM   // Permute value for P8_ macros.
#endif

	// Set CR{1,2,3}EQ to hold the key size information.
	CMPU	R6, $10, CR1
	CMPU	R6, $12, CR2
	CMPU	R6, $14, CR3

	MOVD	$16, R6
	MOVD	$32, R7
	MOVD	$48, R8
	MOVD	$64, R9
	MOVD	$80, R10
	MOVD	$96, R11
	MOVD	$112, R12

	// Load text in BE order
	P8_LXVB16X(R4, R0, V0)

}

// IsPPC64CMP reports whether the op (as defined by an ppc64.A* constant) is
// one of the CMP instructions that require special handling.
func IsPPC64CMP(op obj.As) bool {
	switch op {
	case ppc64.ACMP, ppc64.ACMPU, ppc64.ACMPW, ppc64.ACMPWU, ppc64.AFCMPO, ppc64.AFCMPU:
		return true
	}
	return false
}

// IsPPC64NEG reports whether the op (as defined by an ppc64.A* constant) is

	// amd64:"(MOVB\t[$]0)|(XORL\t[A-Z][A-Z0-9]+, [A-Z][A-Z0-9]+)",-"FCMP",-"MOVB\t[$]1"
	// s390x:"MOV(B|BZ|D)\t[$]0,",-"FCMPU",-"MOV(B|BZ|D)\t[$]1,"
	return 0.5 == float64(uint32(1)) || 1.5 > float64(uint64(1<<63))
}

func constantCheck32() bool {
	// amd64:"MOV(B|L)\t[$]1",-"FCMP",-"MOV(B|L)\t[$]0"
	// s390x:"MOV(B|BZ|D)\t[$]1,",-"FCMPU",-"MOV(B|BZ|D)\t[$]0,"
	return float32(0.5) <= float32(int64(1)) && float32(1.5) >= float32(int32(-1<<31))
}

		{name: "FMOVSconst", argLength: 0, reg: fp01, aux: "Float32", asm: "FMOVS", rematerializeable: true},           //
		{name: "FCMPU", argLength: 2, reg: fp2cr, asm: "FCMPU", typ: "Flags"},

		{name: "CMP", argLength: 2, reg: gp2cr, asm: "CMP", typ: "Flags"},     // arg0 compare to arg1
		{name: "CMPU", argLength: 2, reg: gp2cr, asm: "CMPU", typ: "Flags"},   // arg0 compare to arg1