[glibc/release/2.32/master] AArch64: Improve strchrnul

[Wilco Dijkstra]

https://sourceware.org/git/gitweb.cgi?p=glibc.git;h=bb36cb21ef589ec1199b9382a94c1960af0aa90a

commit bb36cb21ef589ec1199b9382a94c1960af0aa90a
Author: Wilco Dijkstra <wilco.dijkstra@arm.com>
Date:   Wed Jan 11 13:52:23 2023 +0000

    AArch64: Improve strchrnul
    
    Unroll the main loop, which improves performance slightly.
    
    Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
    (cherry picked from commit 09ebd8549b2ce5a3a6c0c7c5f3e62227faf50a99)

Diff:
---
 sysdeps/aarch64/strchrnul.S | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/sysdeps/aarch64/strchrnul.S b/sysdeps/aarch64/strchrnul.S
index 3a29eb7692..c3817f45db 100644
--- a/sysdeps/aarch64/strchrnul.S
+++ b/sysdeps/aarch64/strchrnul.S
@@ -70,14 +70,22 @@ ENTRY (__strchrnul)
 
 	.p2align 4
 L(loop):
-	ldr	qdata, [src, 16]!
+	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