ADD	$65536, R4, R5		// 1e02001485781000
	ADD	$4096, R4, R5		// 3e00001485781000
	ADD	$65536, R4		// 1e02001484781000
	ADD	$4096, R4		// 3e00001484781000
	ADDV	$65536, R4, R5		// 1e02001485f81000
	ADDV	$4096, R4, R5		// 3e00001485f81000
	ADDV	$65536, R4		// 1e02001484f81000
	ADDV	$4096, R4		// 3e00001484f81000
	AND	$65536, R4, R5		// 1e02001485f81400
	AND	$4096, R4, R5		// 3e00001485f81400
	AND	$65536, R4		// 1e02001484f81400

	// do one byte at a time until 8-aligned
	AND	$7, R1, R3
	BEQ	R3, words
	MOVB	R0, (R1)
	ADDV	$1, R1
	JMP	-4(PC)

words:
	// do 8 bytes at a time if there is room
	ADDV	$-7, R4, R2

	SGTU	R2, R1, R3
	BEQ	R3, out
	MOVV	R0, (R1)
	ADDV	$8, R1
	JMP	-4(PC)

out:
	BEQ	R1, R4, done
	MOVB	R0, (R1)
	ADDV	$1, R1
	JMP	-3(PC)
done:

	// will clobber the test pattern created by the caller
	ADDV	$(1024*8), R3

	// Ask signaller to setgid
	MOVW	$1, R12
	DBAR
	MOVW	R12, ·Baton(SB)
	DBAR

	// Wait for setgid completion
loop:
	DBAR
	MOVW	·Baton(SB), R12
	OR	R13, R13, R13	// hint that we're in a spin loop
	BNE	R12, loop
	DBAR

	// Restore stack
	ADDV	$(-1024*8), R3

	// will clobber the test pattern created by the caller
	ADDV	$(1024*8), R29

	// Ask signaller to setgid
	MOVW	$1, R1
	SYNC
	MOVW	R1, ·Baton(SB)
	SYNC

	// Wait for setgid completion
loop:
	SYNC
	MOVW	·Baton(SB), R1
	OR	R2, R2, R2	// hint that we're in a spin loop
	BNE	R1, loop
	SYNC

	// Restore stack
	ADDV	$(-1024*8), R29

	BNE	R0, R15, byte_loop

	PCALIGN	$16
chunk16_loop:
	BEQ	R0, R14, byte_loop
	MOVV	(R4), R8
	MOVV	(R6), R9
	BNE	R8, R9, byte_loop
	MOVV	8(R4), R16
	MOVV	8(R6), R17
	ADDV	$16, R4
	ADDV	$16, R6
	SUBVU	$1, R14
	BEQ	R16, R17, chunk16_loop
	SUBV	$8, R4
	SUBV	$8, R6

byte_loop:
	BEQ	R4, R12, samebytes
	MOVBU	(R4), R8
	ADDVU	$1, R4
	MOVBU	(R6), R9
	ADDVU	$1, R6
	BEQ	R8, R9, byte_loop

byte_cmp:

	BNE	R0, R11, byte_loop

chunk16_loop:
	BEQ	R0, R10, byte_loop
	MOVV	(R3), R6
	MOVV	(R4), R7
	BNE	R6, R7, byte_loop
	MOVV	8(R3), R13
	MOVV	8(R4), R14
	ADDV	$16, R3
	ADDV	$16, R4
	SUBVU	$1, R10
	BEQ	R13, R14, chunk16_loop
	SUBV	$8, R3
	SUBV	$8, R4

byte_loop:
	BEQ	R3, R8, samebytes
	MOVBU	(R3), R6
	ADDVU	$1, R3
	MOVBU	(R4), R7
	ADDVU	$1, R4
	BEQ	R6, R7, byte_loop

byte_cmp:

	fmt.Fprintln(buf, "BNE R5, label23")

	for i := 0; i <= genSize1; i++ {
		fmt.Fprintln(buf, "ADDV $0, R0, R0")
	}

	fmt.Fprintln(buf, "label18:")
	for i := 0; i <= (genSize2 - genSize1); i++ {
		fmt.Fprintln(buf, "ADDV $0, R0, R0")
	}

	fmt.Fprintln(buf, "label23:")
	fmt.Fprintln(buf, "ADDV $0, R0, R0")
	fmt.Fprintln(buf, "RET")

	 *  push 3 args for fn (R4, R5, R7), skipping R6.
	 * Also note that at procedure entry in gc world, 8(R29) will be the
	 *  first arg.
	 */
#ifndef GOMIPS64_softfloat
	ADDV	$(-8*23), R29
#else
	ADDV	$(-8*15), R29
#endif
	MOVV	R4, (8*1)(R29) // fn unsafe.Pointer
	MOVV	R5, (8*2)(R29) // a unsafe.Pointer
	MOVV	R7, (8*3)(R29) // ctxt uintptr
	MOVV	R16, (8*4)(R29)
	MOVV	R17, (8*5)(R29)

	/* copy arguments to stack */		\
	MOVV	stackArgs+16(FP), R1;			\
	MOVWU	stackArgsSize+24(FP), R2;			\
	MOVV	R29, R3;				\
	ADDV	$8, R3;			\
	ADDV	R3, R2;				\
	BEQ	R3, R2, 6(PC);				\
	MOVBU	(R1), R4;			\
	ADDV	$1, R1;			\
	MOVBU	R4, (R3);			\
	ADDV	$1, R3;			\
	JMP	-5(PC);				\
	/* call function */			\
	MOVV	f+8(FP), REGCTXT;			\
	MOVV	(REGCTXT), R4;			\
	PCDATA  $PCDATA_StackMapIndex, $0;	\

	// amd64:"NEGL","ADCQ","SBBQ","NEGQ"
	// loong64: "ADDV", "SGTU"
	// ppc64x: "ADDC", "ADDE", "ADDZE"
	// s390x:"ADDE","ADDC\t[$]-1,"
	// mips64:"ADDV","SGTU"
	// riscv64: "ADD","SLTU"
	return bits.Add(x, 7, ci)
}

func AddZ(x, y uint) (r, co uint) {
	// arm64:"ADDS","ADC",-"ADCS",-"ADD\t",-"CMP"
	// amd64:"ADDQ","SBBQ","NEGQ",-"NEGL",-"ADCQ"
	// loong64: "ADDV", "SGTU"
	// ppc64x: "ADDC", -"ADDE", "ADDZE"