"WFI",
	"SFENCEVMA",
	"ADDUW",
	"SH1ADD",
	"SH1ADDUW",
	"SH2ADD",
	"SH2ADDUW",
	"SH3ADD",
	"SH3ADDUW",
	"SLLIUW",
	"ANDN",
	"ORN",
	"XNOR",
	"CLZ",
	"CLZW",
	"CTZ",
	"CTZW",
	"CPOP",
	"CPOPW",
	"MAX",
	"MAXU",
	"MIN",
	"MINU",
	"SEXTB",
	"SEXTH",
	"ZEXTH",
	"ROL",
	"ROLW",
	"ROR",
	"RORI",
	"RORIW",
	"RORW",
	"ORCB",

		// 	SLLI $63, X10, X10
		var insSLLI *instruction
		if err := immIFits(ins.imm, 32); err != nil {
			ctz := bits.TrailingZeros64(uint64(ins.imm))
			if err := immIFits(ins.imm>>ctz, 32); err == nil {
				ins.imm = ins.imm >> ctz
				insSLLI = &instruction{as: ASLLI, rd: ins.rd, rs1: ins.rd, imm: int64(ctz)}
			}
		}

		low, high, err := Split32BitImmediate(ins.imm)
		if err != nil {

		{name: "F64Copysign", asm: "F64Copysign", argLength: 2, reg: fp64_21, typ: "Float64"}, // copysign(arg0, arg1)

		{name: "I64Ctz", asm: "I64Ctz", argLength: 1, reg: gp11, typ: "Int64"},       // ctz(arg0)
		{name: "I64Clz", asm: "I64Clz", argLength: 1, reg: gp11, typ: "Int64"},       // clz(arg0)
		{name: "I32Rotl", asm: "I32Rotl", argLength: 2, reg: gp21, typ: "Int32"},     // rotl(arg0, arg1)

	ANDN	X19, X20, X21				// b37a3a41
	ANDN	X19, X20				// 337a3a41
	CLZ	X20, X21				// 931a0a60
	CLZW	X21, X22				// 1b9b0a60
	CPOP	X22, X23				// 931b2b60
	CPOPW	X23, X24				// 1b9c2b60
	CTZ	X24, X25				// 931c1c60
	CTZW	X25, X26				// 1b9d1c60
	MAX	X26, X28, X29				// b36eae0b
	MAX	X26, X28				// 336eae0b
	MAXU	X28, X29, X30				// 33ffce0b
	MAXU	X28, X29				// b3fece0b
	MIN	X29, X30, X5				// b342df0b

(Ctz64 ...) => (I64Ctz ...)
(Ctz32 x) => (I64Ctz (I64Or x (I64Const [0x100000000])))
(Ctz16 x) => (I64Ctz (I64Or x (I64Const [0x10000])))
(Ctz8  x) => (I64Ctz (I64Or x (I64Const [0x100])))

(Ctz(64|32|16|8)NonZero ...) => (I64Ctz ...)

(BitLen64 x) => (I64Sub (I64Const [64]) (I64Clz x))

(PopCount64 ...) => (I64Popcnt ...)
(PopCount32 x) => (I64Popcnt (ZeroExt32to64 x))

(RotateLeft16 <t> x y) => (RORW <t> (ORshiftLL <typ.UInt32> (ZeroExt16to32 x) (ZeroExt16to32 x) [16]) (NEG <typ.Int64> y))
(RotateLeft32 x y) => (RORW x (NEG <y.Type> y))
(RotateLeft64 x y) => (ROR x (NEG <y.Type> y))

(Ctz(64|32|16|8)NonZero ...) => (Ctz(64|32|32|32) ...)

(Ctz64 <t> x) => (CLZ  (RBIT  <t> x))
(Ctz32 <t> x) => (CLZW (RBITW <t> x))
(Ctz16 <t> x) => (CLZW <t> (RBITW <typ.UInt32> (ORconst <typ.UInt32> [0x10000] x)))

	bytep, mask := s.allocBits.bitp(allocBitIndex)
	return markBits{bytep, mask, allocBitIndex}
}

// refillAllocCache takes 8 bytes s.allocBits starting at whichByte
// and negates them so that ctz (count trailing zeros) instructions
// can be used. It then places these 8 bytes into the cached 64 bit
// s.allocCache.
func (s *mspan) refillAllocCache(whichByte uint16) {

(OffPtr [off] ptr) && is32Bit(off) => (ADDconst [int32(off)] ptr)
(OffPtr [off] ptr) => (ADD (MOVDconst [off]) ptr)

// TODO: optimize these cases?
(Ctz64NonZero ...) => (Ctz64 ...)
(Ctz32NonZero ...) => (Ctz32 ...)

// Ctz(x) = 64 - findLeftmostOne((x-1)&^x)
(Ctz64 <t> x) => (SUB (MOVDconst [64]) (FLOGR (AND <t> (SUBconst <t> [1] x) (NOT <t> x))))
(Ctz32 <t> x) => (SUB (MOVDconst [64]) (FLOGR (MOVWZreg (ANDW <t> (SUBWconst <t> [1] x) (NOTW <t> x)))))

	freeIndexForScan uint16

	// Cache of the allocBits at freeindex. allocCache is shifted
	// such that the lowest bit corresponds to the bit freeindex.
	// allocCache holds the complement of allocBits, thus allowing
	// ctz (count trailing zero) to use it directly.
	// allocCache may contain bits beyond s.nelems; the caller must ignore
	// these.
	allocCache uint64

	// allocBits and gcmarkBits hold pointers to a span's mark and