RET

TEXT	·Load8(SB),NOSPLIT|NOFRAME,$0-5
	MOVW	addr+0(FP), R0
	MOVB	(R0), R1

	MOVB	runtime·goarm(SB), R11
	CMP	$7, R11
	BLT	2(PC)
	DMB	MB_ISH

	MOVB	R1, ret+4(FP)
	RET

TEXT	·Store8(SB),NOSPLIT,$0-5
	MOVW	addr+0(FP), R1
	MOVB	v+4(FP), R2

	MOVB	runtime·goarm(SB), R8
	CMP	$7, R8
	BLT	2(PC)
	DMB	MB_ISH

	MOVB	R2, (R1)

	CMP	$7, R8
	BLT	2(PC)
	DMB	MB_ISH

	RET

TEXT	·Load8(SB),NOSPLIT,$0-5
	MOVW	addr+0(FP), R0
	MOVB	(R0), R1

	MOVB	runtime·goarm(SB), R11
	CMP	$7, R11
	BGE	native_barrier
	BL	memory_barrier<>(SB)
	B	end
native_barrier:
	DMB	MB_ISH
end:
	MOVB	R1, ret+4(FP)
	RET

TEXT	·Store8(SB),NOSPLIT,$0-5
	MOVW	addr+0(FP), R1
	MOVB	v+4(FP), R2

	MOVB	runtime·goarm(SB), R8
	CMP	$7, R8
	BGE	native_barrier

TEXT ·Store(SB), NOSPLIT, $0
	MOVD	ptr+0(FP), R2
	MOVWZ	val+8(FP), R3
	MOVW	R3, 0(R2)
	SYNC
	RET

// func Store8(ptr *uint8, val uint8)
TEXT ·Store8(SB), NOSPLIT, $0
	MOVD	ptr+0(FP), R2
	MOVB	val+8(FP), R3
	MOVB	R3, 0(R2)
	SYNC
	RET

// func Store64(ptr *uint64, val uint64)
TEXT ·Store64(SB), NOSPLIT, $0
	MOVD	ptr+0(FP), R2
	MOVD	val+8(FP), R3
	MOVD	R3, 0(R2)
	SYNC
	RET

	B	equal
lt_4:
	TBZ	$1, R2, lt_2
	MOVHU.P	2(R0), R4
	MOVHU.P	2(R1), R5
	CMP	R4, R5
	BNE	not_equal
lt_2:
	TBZ	$0, R2, equal
one:
	MOVBU	(R0), R4
	MOVBU	(R1), R5
	CMP	R4, R5
	BNE	not_equal
equal:
	MOVD	$1, R0
	RET
not_equal:
	MOVB	ZR, R0

	// R6 = b_cap (unused)
	// R7 = byte to find
	AND	$0xff, R7
	MOVV	R4, R6		// store base for later
	ADDV	R4, R5		// end
	ADDV	$-1, R4

	PCALIGN	$16
loop:
	ADDV	$1, R4
	BEQ	R4, R5, notfound
	MOVBU	(R4), R8
	BNE	R7, R8, loop

	SUBV	R6, R4		// remove base
	RET

notfound:
	MOVV	$-1, R4
	RET

TEXT ·IndexByteString<ABIInternal>(SB),NOSPLIT,$0-32
	// R4 = s_base
	// R5 = s_len

	"BGEZ",
	"BGT",
	"BGTU",
	"BGTZ",
	"BLE",
	"BLEU",
	"BLEZ",
	"BLTZ",
	"BNEZ",
	"FABSD",
	"FABSS",
	"FNEGD",
	"FNEGS",
	"FNED",
	"FNES",
	"MOV",
	"MOVB",
	"MOVBU",
	"MOVF",
	"MOVD",
	"MOVH",
	"MOVHU",
	"MOVW",
	"MOVWU",
	"NEG",
	"NEGW",
	"NOT",
	"SEQZ",
	"SNEZ",
	"LAST",

	// do one byte at a time until 8-aligned
	AND	$7, R4, R8
	BEQ	R8, words
	MOVB	R0, (R4)
	ADDV	$1, R4
	JMP	-4(PC)

words:
	// do 8 bytes at a time if there is room
	ADDV	$-7, R6, R5

	PCALIGN	$16
	SGTU	R5, R4, R8
	BEQ	R8, out
	MOVV	R0, (R4)
	ADDV	$8, R4
	JMP	-4(PC)

out:
	BEQ	R4, R6, done
	MOVB	R0, (R4)
	ADDV	$1, R4
	JMP	-3(PC)
done:

Daniel Grimm <******@****.***> 1708485021 +0100

	MOVHZ	(R5)(R8), R15
	XOR	R14, R15, R16
	MOVH	R16, (R3)(R8)
	ADD	$2,R8
	ADD	$-2,R6
xor1:
	CMP	R6, $0
	BC	12,2,LR		// BEQLR
	MOVBZ	(R4)(R8), R14	// R14 = a[i]
	MOVBZ	(R5)(R8), R15	// R15 = b[i]
	XOR	R14, R15, R16	// R16 = a[i] ^ b[i]
	MOVB	R16, (R3)(R8)	// Store to dst
#endif
done:

		{name: "MOVBreg", argLength: 1, reg: gp11, asm: "MOVB", typ: "Int64"},   // sign extend int8 to int64
		{name: "MOVBZreg", argLength: 1, reg: gp11, asm: "MOVBZ", typ: "Int64"}, // zero extend uint8 to uint64
		{name: "MOVHreg", argLength: 1, reg: gp11, asm: "MOVH", typ: "Int64"},   // sign extend int16 to int64