llvm-project/libc/AOR_v20.02/string/aarch64/memcmp-sve.S
Kristof Beyls 0928368f62 [libc] Provide Arm Optimized Routines for the LLVM libc project.
This adds the Arm Optimized Routines (see
https://github.com/ARM-software/optimized-routines) source code under the
the LLVM license. The version of the code provided in this patch is v20.02
of the Arm Optimized Routines project.

This entire contribution is being committed as is even though it does
not currently fit the LLVM libc model and does not follow the LLVM
coding style. In the near future, implementations from this patch will be
moved over to their right place in the LLVM-libc tree. This will be done
over many small patches, all of which will go through the normal LLVM code
review process. See this libc-dev post for the plan:
http://lists.llvm.org/pipermail/libc-dev/2020-March/000044.html

Differential revision of the original upload: https://reviews.llvm.org/D75355
2020-03-16 12:19:31 -07:00

52 lines
1.2 KiB
ArmAsm

/*
* memcmp - compare memory
*
* Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
* See https://llvm.org/LICENSE.txt for license information.
* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
*/
#if __ARM_FEATURE_SVE
/* Assumptions:
*
* ARMv8-a, AArch64
* SVE Available.
*/
.arch armv8-a+sve
.text
.globl __memcmp_aarch64_sve
.type __memcmp_aarch64_sve, %function
.p2align 4
__memcmp_aarch64_sve:
mov x3, 0 /* initialize off */
0: whilelo p0.b, x3, x2 /* while off < max */
b.none 9f
ld1b z0.b, p0/z, [x0, x3] /* read vectors bounded by max. */
ld1b z1.b, p0/z, [x1, x3]
/* Increment for a whole vector, even if we've only read a partial.
This is significantly cheaper than INCP, and since OFF is not
used after the loop it is ok to increment OFF past MAX. */
incb x3
cmpne p1.b, p0/z, z0.b, z1.b /* while no inequalities */
b.none 0b
/* Found inequality. */
1: brkb p1.b, p0/z, p1.b /* find first such */
lasta w0, p1, z0.b /* extract each byte */
lasta w1, p1, z1.b
sub x0, x0, x1 /* return comparison */
ret
/* Found end-of-count. */
9: mov x0, 0 /* return equality */
ret
.size __memcmp_aarch64_sve, . - __memcmp_aarch64_sve
#endif