mirror of
https://github.com/apache/nuttx.git
synced 2026-08-03 21:29:02 +00:00
603c87977f
kasan does not instrument the assembly function, so it is checked in advance before calling. If Kasan is not turned on, the speed and space will be almost unchanged under compiler optimization. Signed-off-by: anjiahao <anjiahao@xiaomi.com>
271 lines
6.8 KiB
ArmAsm
271 lines
6.8 KiB
ArmAsm
/****************************************************************************
|
|
* libs/libc/machine/xtensa/arch_strncpy.S
|
|
*
|
|
* SPDX-License-Identifier: Apache-2.0
|
|
*
|
|
* Licensed to the Apache Software Foundation (ASF) under one or more
|
|
* contributor license agreements. See the NOTICE file distributed with
|
|
* this work for additional information regarding copyright ownership. The
|
|
* ASF licenses this file to you under the Apache License, Version 2.0 (the
|
|
* "License"); you may not use this file except in compliance with the
|
|
* License. You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
|
|
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
|
|
* License for the specific language governing permissions and limitations
|
|
* under the License.
|
|
*
|
|
****************************************************************************/
|
|
|
|
/****************************************************************************
|
|
* Included Files
|
|
****************************************************************************/
|
|
|
|
#include "xtensa_asm.h"
|
|
|
|
#include <arch/chip/core-isa.h>
|
|
#include <arch/xtensa/xtensa_abi.h>
|
|
|
|
#include "libc.h"
|
|
|
|
#ifdef LIBC_BUILD_STRNCPY
|
|
|
|
/****************************************************************************
|
|
* Public Functions
|
|
****************************************************************************/
|
|
|
|
.section .text
|
|
.begin schedule
|
|
.align 4
|
|
.literal_position
|
|
__strncpy_aux:
|
|
|
|
.Lsrc1mod2: # src address is odd
|
|
l8ui a8, a3, 0 # get byte 0
|
|
addi a3, a3, 1 # advance src pointer
|
|
s8i a8, a10, 0 # store byte 0
|
|
addi a4, a4, -1 # decrement n
|
|
beqz a4, .Lret # if n is zero
|
|
addi a10, a10, 1 # advance dst pointer
|
|
beqz a8, .Lfill # if byte 0 is zero
|
|
bbci.l a3, 1, .Lsrcaligned # if src is now word-aligned
|
|
|
|
.Lsrc2mod4: # src address is 2 mod 4
|
|
l8ui a8, a3, 0 # get byte 0
|
|
addi a4, a4, -1 # decrement n
|
|
s8i a8, a10, 0 # store byte 0
|
|
beqz a4, .Lret # if n is zero
|
|
addi a10, a10, 1 # advance dst pointer
|
|
beqz a8, .Lfill # if byte 0 is zero
|
|
l8ui a8, a3, 1 # get byte 0
|
|
addi a3, a3, 2 # advance src pointer
|
|
s8i a8, a10, 0 # store byte 0
|
|
addi a4, a4, -1 # decrement n
|
|
beqz a4, .Lret # if n is zero
|
|
addi a10, a10, 1 # advance dst pointer
|
|
bnez a8, .Lsrcaligned
|
|
j .Lfill
|
|
|
|
.Lret:
|
|
RET(16)
|
|
|
|
.align 4
|
|
.global ARCH_LIBCFUN(strncpy)
|
|
.type ARCH_LIBCFUN(strncpy), @function
|
|
ARCH_LIBCFUN(strncpy):
|
|
ENTRY(16)
|
|
/* a2 = dst, a3 = src */
|
|
|
|
mov a10, a2 # leave dst in return value register
|
|
beqz a4, .Lret # if n is zero
|
|
|
|
movi a11, MASK0
|
|
movi a5, MASK1
|
|
movi a6, MASK2
|
|
movi a7, MASK3
|
|
bbsi.l a3, 0, .Lsrc1mod2
|
|
bbsi.l a3, 1, .Lsrc2mod4
|
|
.Lsrcaligned:
|
|
|
|
/* Check if the destination is aligned. */
|
|
movi a8, 3
|
|
bnone a10, a8, .Laligned
|
|
|
|
j .Ldstunaligned
|
|
|
|
/* Fill the dst with zeros -- n is at least 1. */
|
|
|
|
.Lfill:
|
|
movi a9, 0
|
|
bbsi.l a10, 0, .Lfill1mod2
|
|
bbsi.l a10, 1, .Lfill2mod4
|
|
.Lfillaligned:
|
|
blti a4, 4, .Lfillcleanup
|
|
|
|
/* Loop filling complete words with zero. */
|
|
#if XCHAL_HAVE_LOOPS
|
|
|
|
srai a8, a4, 2
|
|
loop a8, 1f
|
|
s32i a9, a10, 0
|
|
addi a10, a10, 4
|
|
|
|
1: slli a8, a8, 2
|
|
sub a4, a4, a8
|
|
|
|
#else /* !XCHAL_HAVE_LOOPS */
|
|
|
|
1: s32i a9, a10, 0
|
|
addi a10, a10, 4
|
|
addi a4, a4, -4
|
|
bgei a4, 4, 1b
|
|
|
|
#endif /* !XCHAL_HAVE_LOOPS */
|
|
|
|
beqz a4, 2f
|
|
|
|
.Lfillcleanup:
|
|
/* Fill leftover (1 to 3) bytes with zero. */
|
|
s8i a9, a10, 0 # store byte 0
|
|
addi a4, a4, -1 # decrement n
|
|
addi a10, a10, 1
|
|
bnez a4, .Lfillcleanup
|
|
|
|
2: RET(16)
|
|
|
|
.Lfill1mod2: # dst address is odd
|
|
s8i a9, a10, 0 # store byte 0
|
|
addi a4, a4, -1 # decrement n
|
|
beqz a4, 2b # if n is zero
|
|
addi a10, a10, 1 # advance dst pointer
|
|
bbci.l a10, 1, .Lfillaligned # if dst is now word-aligned
|
|
|
|
.Lfill2mod4: # dst address is 2 mod 4
|
|
s8i a9, a10, 0 # store byte 0
|
|
addi a4, a4, -1 # decrement n
|
|
beqz a4, 2b # if n is zero
|
|
s8i a9, a10, 1 # store byte 1
|
|
addi a4, a4, -1 # decrement n
|
|
beqz a4, 2b # if n is zero
|
|
addi a10, a10, 2 # advance dst pointer
|
|
j .Lfillaligned
|
|
|
|
/* dst is word-aligned; src is word-aligned; n is at least 1. */
|
|
|
|
.align 4
|
|
#if XCHAL_HAVE_LOOPS
|
|
#if XCHAL_HAVE_DENSITY
|
|
/* (2 mod 4) alignment for loop instruction */
|
|
#else
|
|
/* (1 mod 4) alignment for loop instruction */
|
|
.byte 0
|
|
.byte 0
|
|
#endif
|
|
#endif
|
|
.Laligned:
|
|
#if XCHAL_HAVE_LOOPS
|
|
#if XCHAL_HAVE_DENSITY
|
|
_movi.n a8, 0 # set up for the maximum loop count
|
|
#else
|
|
_movi a8, 0 # set up for the maximum loop count
|
|
#endif
|
|
loop a8, 1f # loop forever (almost anyway)
|
|
blti a4, 5, .Ldstunaligned # n is near limit; do one at a time
|
|
l32i a8, a3, 0 # get word from src
|
|
addi a3, a3, 4 # advance src pointer
|
|
bnone a8, a11, .Lz0 # if byte 0 is zero
|
|
bnone a8, a5, .Lz1 # if byte 1 is zero
|
|
bnone a8, a6, .Lz2 # if byte 2 is zero
|
|
s32i a8, a10, 0 # store word to dst
|
|
addi a4, a4, -4 # decrement n
|
|
addi a10, a10, 4 # advance dst pointer
|
|
bnone a8, a7, .Lfill # if byte 3 is zero
|
|
1:
|
|
|
|
#else /* !XCHAL_HAVE_LOOPS */
|
|
|
|
1: blti a4, 5, .Ldstunaligned # n is near limit; do one at a time
|
|
l32i a8, a3, 0 # get word from src
|
|
addi a3, a3, 4 # advance src pointer
|
|
bnone a8, a11, .Lz0 # if byte 0 is zero
|
|
bnone a8, a5, .Lz1 # if byte 1 is zero
|
|
bnone a8, a6, .Lz2 # if byte 2 is zero
|
|
s32i a8, a10, 0 # store word to dst
|
|
addi a4, a4, -4 # decrement n
|
|
addi a10, a10, 4 # advance dst pointer
|
|
bany a8, a7, 1b # no zeroes
|
|
#endif /* !XCHAL_HAVE_LOOPS */
|
|
|
|
j .Lfill
|
|
|
|
.Lz0: /* Byte 0 is zero. */
|
|
#if XCHAL_HAVE_BE
|
|
movi a8, 0
|
|
#endif
|
|
s8i a8, a10, 0
|
|
addi a4, a4, -1 # decrement n
|
|
addi a10, a10, 1 # advance dst pointer
|
|
j .Lfill
|
|
|
|
.Lz1: /* Byte 1 is zero. */
|
|
#if XCHAL_HAVE_BE
|
|
extui a8, a8, 16, 16
|
|
#endif
|
|
s16i a8, a10, 0
|
|
addi a4, a4, -2 # decrement n
|
|
addi a10, a10, 2 # advance dst pointer
|
|
j .Lfill
|
|
|
|
.Lz2: /* Byte 2 is zero. */
|
|
#if XCHAL_HAVE_BE
|
|
extui a8, a8, 16, 16
|
|
#endif
|
|
s16i a8, a10, 0
|
|
movi a8, 0
|
|
s8i a8, a10, 2
|
|
addi a4, a4, -3 # decrement n
|
|
addi a10, a10, 3 # advance dst pointer
|
|
j .Lfill
|
|
|
|
.align 4
|
|
#if XCHAL_HAVE_LOOPS
|
|
#if XCHAL_HAVE_DENSITY
|
|
/* (2 mod 4) alignment for loop instruction */
|
|
#else
|
|
/* (1 mod 4) alignment for loop instruction */
|
|
.byte 0
|
|
.byte 0
|
|
#endif
|
|
#endif
|
|
.Ldstunaligned:
|
|
|
|
#if XCHAL_HAVE_LOOPS
|
|
#if XCHAL_HAVE_DENSITY
|
|
_movi.n a8, 0 # set up for the maximum loop count
|
|
#else
|
|
_movi a8, 0 # set up for the maximum loop count
|
|
#endif
|
|
loop a8, 2f # loop forever (almost anyway)
|
|
#endif
|
|
1: l8ui a8, a3, 0
|
|
addi a3, a3, 1
|
|
s8i a8, a10, 0
|
|
addi a4, a4, -1
|
|
beqz a4, 3f
|
|
addi a10, a10, 1
|
|
#if XCHAL_HAVE_LOOPS
|
|
beqz a8, 2f
|
|
#else
|
|
bnez a8, 1b
|
|
#endif
|
|
2: j .Lfill
|
|
|
|
3: RET(16)
|
|
.end schedule
|
|
|
|
.size ARCH_LIBCFUN(strncpy), . - ARCH_LIBCFUN(strncpy)
|
|
|
|
#endif
|