AMANDV		R14, (R13), R12 // acb96238
	AMORW		R14, (R13), R12 // ac396338
	AMORV		R14, (R13), R12 // acb96338
	AMXORW		R14, (R13), R12 // ac396438
	AMXORV		R14, (R13), R12 // acb96438
	AMMAXW		R14, (R13), R12 // ac396538
	AMMAXV		R14, (R13), R12 // acb96538
	AMMINW		R14, (R13), R12 // ac396638
	AMMINV		R14, (R13), R12 // acb96638
	AMMAXWU		R14, (R13), R12 // ac396738
	AMMAXVU		R14, (R13), R12 // acb96738

				{1, 49151}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R15
			},
		},
	},
	{
		name:   "CMPW",
		argLen: 2,
		asm:    x86.ACMPW,
		reg: regInfo{
			inputs: []inputInfo{
				{0, 49151}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R15
				{1, 49151}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R15
			},
		},
	},
	{

	MOVD	0(R1), R3 // argc
	ADD	$8, R1, R4 // argv

	// Did the TLS pointer get set? If so, don't change it (e.g musl).
	CMP	R13, $0
	BNE	tls_and_argcv_in_reg

	MOVD	$runtime·m0+m_tls(SB), R13 // TLS
	ADD	$0x7000, R13

tls_and_argcv_in_reg:
	BR	main(SB)

TEXT main(SB),NOSPLIT,$-8
	MOVD	$runtime·rt0_go(SB), R12
	MOVD	R12, CTR

	MOVV	(g_sched+gobuf_sp)(g), R3

	// Now on a scheduling stack (a pthread-created stack).
g0:
	// Save room for two of our pointers.
	ADDV	$-16, R3
	MOVV	R13, 0(R3)	// save old g on stack
	MOVV	(g_stack+stack_hi)(R13), R13
	SUBVU	R12, R13
	MOVV	R13, 8(R3)	// save depth in old g stack (can't just save SP, as stack might be copied during a callback)
	JAL	(R25)

	// Restore g, stack pointer. R4 is return value.

	MOVQ	g(R12), R14
	MOVQ	g_m(R14), R13
	MOVQ	m_g0(R13), R15
	CMPQ	R13, R15
	JEQ	noswitch	// branch if already on g0
	MOVQ	R15, g(R12)	// g = m->g0
	MOVQ	R15, R14	// set g register
	PUSHQ	RARG1	// func arg
	PUSHQ	RARG0	// func arg
	CALL	runtime·racecallback(SB)
	POPQ	R12
	POPQ	R12
	// All registers are smashed after Go code, reload.
	get_tls(R12)
	MOVQ	g(R12), R13
	MOVQ	g_m(R13), R13

	// load_g will clobber R0, Save R0
	MOVD	R0, R13
	load_g
	// restore R0
	MOVD	R13, R0
	MOVD	g_m(g), R13
	MOVD	m_g0(R13), R14
	CMP	R14, g
	BEQ	noswitch	// branch if already on g0
	MOVD	R14, g

	MOVD	R0, 8(RSP)	// func arg
	MOVD	R1, 16(RSP)	// func arg
	BL	runtime·racecallback(SB)

	// All registers are smashed after Go code, reload.
	MOVD	g_m(g), R13
	MOVD	m_curg(R13), g	// g = m->curg
ret:

	// R4 = a_base
	// R5 = a_len
	// R6 = b_base
	// R7 = b_len
	JMP	cmpbody<>(SB)

// On entry:
// R5 length of a
// R7 length of b
// R4 points to the start of a
// R6 points to the start of b
// R13 points to the return value (-1/0/1)
TEXT cmpbody<>(SB),NOSPLIT|NOFRAME,$0
	BEQ	R4, R6, samebytes // same start of a and b

	SGTU	R5, R7, R9
	BNE	R0, R9, r2_lt_r1
	MOVV	R5, R14
	JMP	entry
r2_lt_r1:

	MOVD	R9, R0

	// Now on a scheduling stack (a pthread-created stack).
	// Save room for two of our pointers /*, plus 32 bytes of callee
	// save area that lives on the caller stack. */
	MOVD	RSP, R13
	SUB	$16, R13
	MOVD	R13, RSP
	MOVD	R4, 0(RSP)	// save old g on stack
	MOVD	(g_stack+stack_hi)(R4), R4
	SUB	R2, R4
	MOVD	R4, 8(RSP)	// save depth in old g stack (can't just save SP, as stack might be copied during a callback)

	// faster than having the caller spill these.
	MOVQ	R12, 96(SP)
	MOVQ	R13, 104(SP)
retry:
	// TODO: Consider passing g.m.p in as an argument so they can be shared
	// across a sequence of write barriers.
	MOVQ	g_m(R14), R13
	MOVQ	m_p(R13), R13
	// Get current buffer write position.
	MOVQ	(p_wbBuf+wbBuf_next)(R13), R12	// original next position
	ADDQ	R11, R12			// new next position
	// Is the buffer full?

	rbx    uint64
	rcx    uint64
	rdx    uint64
	rdi    uint64
	rsi    uint64
	rbp    uint64
	rsp    uint64
	r8     uint64
	r9     uint64
	r10    uint64
	r11    uint64
	r12    uint64
	r13    uint64
	r14    uint64
	r15    uint64
	rip    uint64
	rflags uint64
	cs     uint64
	fs     uint64
	gs     uint64
}

type floatstate64 struct {
	fpu_reserved  [2]int32