ADDS	R4, R9
	MUL	R6, R3, R14
	ADCS	R14, R10
	MUL	R7, R3, R15
	LDP	16(R1), (R11, R12)
	ADCS	R15, R11
	MUL	R8, R3, R16
	ADCS	R16, R12
	UMULH	R8, R3, R20
	ADC	$0, R20

	MUL	R5, R3, R13
	ADDS	R13, R9
	UMULH	R5, R3, R17
	ADCS	R17, R10
	UMULH	R6, R3, R21
	STP.P	(R9, R10), 16(R1)
	ADCS	R21, R11
	UMULH	R7, R3, R19
	ADCS	R19, R12
	STP.P	(R11, R12), 16(R1)
	ADC	$0, R20, R4

	SUB	$4, R0

	#ifdef GOOS_ios
	MOVD R30, (RSP)
	#endif
	STP (R0, R1), 8(RSP)
	STP (R2, R3), 24(RSP)
	STP (R4, R5), 40(RSP)
	STP (R6, R7), 56(RSP)
	STP (R8, R9), 72(RSP)
	STP (R10, R11), 88(RSP)
	STP (R12, R13), 104(RSP)
	STP (R14, R15), 120(RSP)
	STP (R16, R17), 136(RSP)
	STP (R19, R20), 152(RSP)
	STP (R21, R22), 168(RSP)
	STP (R23, R24), 184(RSP)
	STP (R25, R26), 200(RSP)
	MOVD NZCV, R0
	MOVD R0, 216(RSP)

	OR	R3, R4, R11
	AND	$7, R11
	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

TEXT	asancall<>(SB), NOSPLIT, $0-0
	get_tls(R12)
	MOVQ	g(R12), R14
	MOVQ	SP, R12		// callee-saved, preserved across the CALL
	CMPQ	R14, $0
	JE	call	// no g; still on a system stack

	MOVQ	g_m(R14), R13
	// Switch to g0 stack.
	MOVQ	m_g0(R13), R10
	CMPQ	R10, R14
	JE	call	// already on g0

	MOVQ	(g_sched+gobuf_sp)(R10), SP
call:
	ANDQ	$~15, SP	// alignment for gcc ABI
	CALL	AX
	MOVQ	R12, SP

func (c *sigctxt) r10() uint64 { return c.regs().r10 }
func (c *sigctxt) r11() uint64 { return c.regs().r11 }
func (c *sigctxt) r12() uint64 { return c.regs().r12 }
func (c *sigctxt) r13() uint64 { return c.regs().r13 }
func (c *sigctxt) r14() uint64 { return c.regs().r14 }
func (c *sigctxt) r15() uint64 { return c.regs().r15 }

//go:nosplit
//go:nowritebarrierrec
func (c *sigctxt) rip() uint64 { return c.regs().rip }

	CSINV	HS, R1, R2, R3	// 232082da
	CSINVW	MI, R2, ZR, R2	// 42409f5a
	CINC	EQ, R4, R9	// 8914849a
	CINCW	PL, R2, ZR	// 5f44821a
	CINV	PL, R11, R22	// 76418bda
	CINVW	LS, R7, R13	// ed80875a
	CNEG	LS, R13, R7	// a7858dda
	CNEGW	EQ, R8, R13	// 0d15885a

// atomic ops
	LDARB	(R25), R2                            // 22ffdf08
	LDARH	(R5), R7                             // a7fcdf48

	print("r8     ", hex(c.r8()), "\n")
	print("r9     ", hex(c.r9()), "\n")
	print("r10    ", hex(c.r10()), "\n")
	print("r11    ", hex(c.r11()), "\n")
	print("r12    ", hex(c.r12()), "\n")
	print("r13    ", hex(c.r13()), "\n")
	print("r14    ", hex(c.r14()), "\n")
	print("r15    ", hex(c.r15()), "\n")
	print("rip    ", hex(c.rip()), "\n")
	print("rflags ", hex(c.rflags()), "\n")
	print("cs     ", hex(c.cs()), "\n")

	MOVW	a_base+0(FP), R2
	MOVW	a_len+4(FP), R0
	MOVW	b_base+12(FP), R3
	MOVW	b_len+16(FP), R1
	ADD	$28, R13, R7
	B	cmpbody<>(SB)

TEXT runtime·cmpstring(SB),NOSPLIT|NOFRAME,$0-20
	MOVW	a_base+0(FP), R2
	MOVW	a_len+4(FP), R0
	MOVW	b_base+8(FP), R3
	MOVW	b_len+12(FP), R1
	ADD	$20, R13, R7
	B	cmpbody<>(SB)

// On entry:
// R0 is the length of a
// R1 is the length of b
// R2 points to the start of a

	print("r8   ", hex(c.r8()), "\t")
	print("r9   ", hex(c.r9()), "\n")
	print("r10  ", hex(c.r10()), "\t")
	print("r11  ", hex(c.r11()), "\n")
	print("r12  ", hex(c.r12()), "\t")
	print("r13  ", hex(c.r13()), "\n")
	print("r14  ", hex(c.r14()), "\t")
	print("r15  ", hex(c.r15()), "\n")
	print("pc   ", hex(c.pc()), "\t")
	print("link ", hex(c.link()), "\n")
}

//go:nosplit
//go:nowritebarrierrec

chunk_loop:
	// chunk size is 16
	SGTU	$16, R3, R8
	BNE	R0, R8, chunk_tail_8
chunk_loop_1:
	MOVV	(R1), R6
	MOVV	(R2), R7
	BNE	R6, R7, not_eq
	MOVV	8(R1), R12
	MOVV	8(R2), R13
	ADDV	$16, R1
	ADDV	$16, R2
	SUBV	$16, R3
	BEQ	R12, R13, chunk_loop
	JMP	not_eq

chunk_tail_8:
	AND	$8, R3, R14
	BEQ	R0, R14, chunk_tail_4
	MOVV	(R1), R6
	MOVV	(R2), R7
	BNE	R6, R7, not_eq
	ADDV	$8, R1
	ADDV	$8, R2