86 lines
1.9 KiB
ArmAsm
86 lines
1.9 KiB
ArmAsm
|
/*
|
||
|
|
* strchrnul - find a character or nul in a string
|
||
|
|
*
|
||
|
|
* Copyright (c) 2020-2022, Arm Limited.
|
||
|
|
* SPDX-License-Identifier: MIT OR Apache-2.0 WITH LLVM-exception
|
||
|
|
*/
|
||
|
|
|
||
|
|
/* Assumptions:
|
||
|
|
*
|
||
|
|
* ARMv8-a, AArch64, Advanced SIMD.
|
||
|
|
* MTE compatible.
|
||
|
|
*/
|
||
|
|
|
||
|
|
#include "asmdefs.h"
|
||
|
|
|
||
|
|
#define srcin x0
|
||
|
|
#define chrin w1
|
||
|
|
#define result x0
|
||
|
|
|
||
|
|
#define src x2
|
||
|
|
#define tmp1 x1
|
||
|
|
#define tmp2 x3
|
||
|
|
|
||
|
|
#define vrepchr v0
|
||
|
|
#define vdata v1
|
||
|
|
#define qdata q1
|
||
|
|
#define vhas_nul v2
|
||
|
|
#define vhas_chr v3
|
||
|
|
#define vend v4
|
||
|
|
#define dend d4
|
||
|
|
|
||
|
|
/*
|
||
|
|
Core algorithm:
|
||
|
|
For each 16-byte chunk we calculate a 64-bit nibble mask value with four bits
|
||
|
|
per byte. We take 4 bits of every comparison byte with shift right and narrow
|
||
|
|
by 4 instruction. Since the bits in the nibble mask reflect the order in
|
||
|
|
which things occur in the original string, counting leading zeros identifies
|
||
|
|
exactly which byte matched. */
|
||
|
|
|
||
|
|
ENTRY (__strchrnul_aarch64_mte)
|
||
|
|
PTR_ARG (0)
|
||
|
|
bic src, srcin, 15
|
||
|
|
dup vrepchr.16b, chrin
|
||
|
|
ld1 {vdata.16b}, [src]
|
||
|
|
cmeq vhas_chr.16b, vdata.16b, vrepchr.16b
|
||
|
|
cmhs vhas_chr.16b, vhas_chr.16b, vdata.16b
|
||
|
|
lsl tmp2, srcin, 2
|
||
|
|
shrn vend.8b, vhas_chr.8h, 4 /* 128->64 */
|
||
|
|
fmov tmp1, dend
|
||
|
|
lsr tmp1, tmp1, tmp2 /* Mask padding bits. */
|
||
|
|
cbz tmp1, L(loop)
|
||
|
|
|
||
|
|
rbit tmp1, tmp1
|
||
|
|
clz tmp1, tmp1
|
||
|
|
add result, srcin, tmp1, lsr 2
|
||
|
|
ret
|
||
|
|
|
||
|
|
.p2align 4
|
||
|
|
L(loop):
|
||
|
|
ldr qdata, [src, 16]
|
||
|
|
cmeq vhas_chr.16b, vdata.16b, vrepchr.16b
|
||
|
|
cmhs vhas_chr.16b, vhas_chr.16b, vdata.16b
|
||
|
|
umaxp vend.16b, vhas_chr.16b, vhas_chr.16b
|
||
|
|
fmov tmp1, dend
|
||
|
|
cbnz tmp1, L(end)
|
||
|
|
ldr qdata, [src, 32]!
|
||
|
|
cmeq vhas_chr.16b, vdata.16b, vrepchr.16b
|
||
|
|
cmhs vhas_chr.16b, vhas_chr.16b, vdata.16b
|
||
|
|
umaxp vend.16b, vhas_chr.16b, vhas_chr.16b
|
||
|
|
fmov tmp1, dend
|
||
|
|
cbz tmp1, L(loop)
|
||
|
|
sub src, src, 16
|
||
|
|
L(end):
|
||
|
|
shrn vend.8b, vhas_chr.8h, 4 /* 128->64 */
|
||
|
|
add src, src, 16
|
||
|
|
fmov tmp1, dend
|
||
|
|
#ifndef __AARCH64EB__
|
||
|
|
rbit tmp1, tmp1
|
||
|
|
#endif
|
||
|
|
clz tmp1, tmp1
|
||
|
|
add result, src, tmp1, lsr 2
|
||
|
|
ret
|
||
|
|
|
||
|
|
END (__strchrnul_aarch64_mte)
|
||
|
|
|