glibc/sysdeps/x86_64/multiarch/strcat-strlen-evex.h.S
Paul Pluzhnikov 1e9d5987fd Fix misspellings in sysdeps/x86_64 -- BZ 25337.
Applying this commit results in bit-identical rebuild of libc.so.6
math/libm.so.6 elf/ld-linux-x86-64.so.2 mathvec/libmvec.so.1

Reviewed-by: Florian Weimer <fweimer@redhat.com>
2023-05-23 10:25:11 +00:00

111 lines
2.7 KiB
ArmAsm

/* strlen used for beginning of str{n}cat using EVEX 256/512.
Copyright (C) 2011-2023 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/>. */
/* NOTE: This file is meant to be included by strcat-evex or
strncat-evex and does not standalone. Before including %rdi
must be saved in %rax. */
/* Simple strlen implementation that ends at
L(strcat_strlen_done). */
vpxorq %VZERO_128, %VZERO_128, %VZERO_128
movq %rdi, %r8
andq $(VEC_SIZE * -1), %r8
VPCMPEQ (%r8), %VZERO, %k0
KMOV %k0, %VRCX
#ifdef USE_AS_WCSCPY
subl %r8d, %edi
shrl $2, %edi
#endif
shrx %VRDI, %VRCX, %VRCX
#ifdef USE_AS_WCSCPY
movq %rax, %rdi
#endif
test %VRCX, %VRCX
jnz L(bsf_and_done_v0)
VPCMPEQ VEC_SIZE(%r8), %VZERO, %k0
KMOV %k0, %VRCX
leaq (VEC_SIZE)(%r8), %rdi
test %VRCX, %VRCX
jnz L(bsf_and_done_v0)
VPCMPEQ (VEC_SIZE * 2)(%r8), %VZERO, %k0
KMOV %k0, %VRCX
test %VRCX, %VRCX
jnz L(bsf_and_done_v1)
VPCMPEQ (VEC_SIZE * 3)(%r8), %VZERO, %k0
KMOV %k0, %VRCX
test %VRCX, %VRCX
jnz L(bsf_and_done_v2)
VPCMPEQ (VEC_SIZE * 4)(%r8), %VZERO, %k0
KMOV %k0, %VRCX
test %VRCX, %VRCX
jnz L(bsf_and_done_v3)
andq $-(VEC_SIZE * 4), %rdi
.p2align 4,, 8
L(loop_2x_vec):
VMOVA (VEC_SIZE * 4)(%rdi), %VMM(0)
VPMIN (VEC_SIZE * 5)(%rdi), %VMM(0), %VMM(1)
VMOVA (VEC_SIZE * 6)(%rdi), %VMM(2)
VPMIN (VEC_SIZE * 7)(%rdi), %VMM(2), %VMM(3)
VPTESTN %VMM(1), %VMM(1), %k1
VPTESTN %VMM(3), %VMM(3), %k3
subq $(VEC_SIZE * -4), %rdi
KORTEST %k1, %k3
jz L(loop_2x_vec)
VPTESTN %VMM(0), %VMM(0), %k0
KMOV %k0, %VRCX
test %VRCX, %VRCX
jnz L(bsf_and_done_v0)
KMOV %k1, %VRCX
test %VRCX, %VRCX
jnz L(bsf_and_done_v1)
VPTESTN %VMM(2), %VMM(2), %k0
KMOV %k0, %VRCX
test %VRCX, %VRCX
jnz L(bsf_and_done_v2)
KMOV %k3, %VRCX
L(bsf_and_done_v3):
addq $VEC_SIZE, %rdi
L(bsf_and_done_v2):
bsf %VRCX, %VRCX
leaq (VEC_SIZE * 2)(%rdi, %rcx, CHAR_SIZE), %rdi
jmp L(strcat_strlen_done)
.p2align 4,, 4
L(bsf_and_done_v1):
addq $VEC_SIZE, %rdi
L(bsf_and_done_v0):
bsf %VRCX, %VRCX
#ifdef USE_AS_WCSCPY
leaq (%rdi, %rcx, CHAR_SIZE), %rdi
#else
addq %rcx, %rdi
#endif
L(strcat_strlen_done):