glibc/sysdeps/x86_64/multiarch/wcschr-sse2.S

163 lines
3.0 KiB
ArmAsm

/* wcschr optimized with SSE2.
Copyright (C) 2017-2024 Free Software Foundation, Inc.
This file is part of the GNU C Library.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<https://www.gnu.org/licenses/>. */
#include <isa-level.h>
/* ISA level >= 2 because there is no wcschr-sse4 implementations. */
#if ISA_SHOULD_BUILD (2)
# ifndef WCSCHR
# define WCSCHR __wcschr_sse2
# endif
# include <sysdep.h>
.text
ENTRY (WCSCHR)
movd %rsi, %xmm1
pxor %xmm2, %xmm2
mov %rdi, %rcx
punpckldq %xmm1, %xmm1
punpckldq %xmm1, %xmm1
and $63, %rcx
cmp $48, %rcx
ja L(cross_cache)
movdqu (%rdi), %xmm0
pcmpeqd %xmm0, %xmm2
add $16, %rdi
pcmpeqd %xmm1, %xmm0
pmovmskb %xmm2, %rdx
pmovmskb %xmm0, %rax
or %rax, %rdx
jnz L(matches)
and $-16, %rdi
movdqa (%rdi), %xmm0
pcmpeqd %xmm0, %xmm2
add $16, %rdi
pcmpeqd %xmm1, %xmm0
pmovmskb %xmm2, %rdx
pmovmskb %xmm0, %rax
or %rax, %rdx
jnz L(matches)
jmp L(loop)
L(cross_cache):
and $15, %rcx
and $-16, %rdi
movdqa (%rdi), %xmm0
pcmpeqd %xmm0, %xmm2
pcmpeqd %xmm1, %xmm0
pmovmskb %xmm2, %rdx
pmovmskb %xmm0, %rax
sar %cl, %rdx
sar %cl, %rax
test %rax, %rax
je L(unaligned_no_match)
bsf %rax, %rax
test %rdx, %rdx
je L(unaligned_match)
bsf %rdx, %rdx
cmp %rdx, %rax
ja L(return_null)
L(unaligned_match):
add %rdi, %rax
add %rcx, %rax
ret
.p2align 4
L(unaligned_no_match):
test %rdx, %rdx
jne L(return_null)
pxor %xmm2, %xmm2
add $16, %rdi
.p2align 4
/* Loop start on aligned string. */
L(loop):
movdqa (%rdi), %xmm0
pcmpeqd %xmm0, %xmm2
add $16, %rdi
pcmpeqd %xmm1, %xmm0
pmovmskb %xmm2, %rdx
pmovmskb %xmm0, %rax
or %rax, %rdx
jnz L(matches)
movdqa (%rdi), %xmm0
pcmpeqd %xmm0, %xmm2
add $16, %rdi
pcmpeqd %xmm1, %xmm0
pmovmskb %xmm2, %rdx
pmovmskb %xmm0, %rax
or %rax, %rdx
jnz L(matches)
movdqa (%rdi), %xmm0
pcmpeqd %xmm0, %xmm2
add $16, %rdi
pcmpeqd %xmm1, %xmm0
pmovmskb %xmm2, %rdx
pmovmskb %xmm0, %rax
or %rax, %rdx
jnz L(matches)
movdqa (%rdi), %xmm0
pcmpeqd %xmm0, %xmm2
add $16, %rdi
pcmpeqd %xmm1, %xmm0
pmovmskb %xmm2, %rdx
pmovmskb %xmm0, %rax
or %rax, %rdx
jnz L(matches)
jmp L(loop)
.p2align 4
L(matches):
pmovmskb %xmm2, %rdx
test %rax, %rax
jz L(return_null)
bsf %rax, %rax
test %rdx, %rdx
je L(match)
bsf %rdx, %rcx
cmp %rcx, %rax
ja L(return_null)
L(match):
sub $16, %rdi
add %rdi, %rax
ret
.p2align 4
L(return_null):
xor %rax, %rax
ret
END (WCSCHR)
#endif