libc/machine: introduce memchr/memmove/memset/strlen from newlib for armv7-m

N/A

Signed-off-by: Huang Qi <huangqi3@xiaomi.com>
Change-Id: If901402e3dc4e6776521ac18fcd2c61e06e6493b
This commit is contained in:
Huang Qi
2021-03-05 16:03:30 +08:00
parent 817bf3b28b
commit ab81410372
8 changed files with 988 additions and 2 deletions
+36
View File
@@ -5,6 +5,15 @@
if ARCH_ARMV7M
config ARMV7M_MEMCHR
bool "Enable optimized memchr() for ARMv7-M"
default n
select MACHINE_OPTS_ARMV7M
select LIBC_ARCH_MEMCHR
depends on ARCH_TOOLCHAIN_GNU
---help---
Enable optimized ARMv7-M specific memchr() library function
config ARMV7M_MEMCPY
bool "Enable optimized memcpy() for ARMv7-M"
default n
@@ -14,6 +23,24 @@ config ARMV7M_MEMCPY
---help---
Enable optimized ARMv7-M specific memcpy() library function
config ARMV7M_MEMSET
bool "Enable optimized memset() for ARMv7-M"
default n
select MACHINE_OPTS_ARMV7M
select LIBC_ARCH_MEMSET
depends on ARCH_TOOLCHAIN_GNU
---help---
Enable optimized ARMv7-M specific memset() library function
config ARMV7M_MEMMOVE
bool "Enable optimized memmove() for ARMv7-M"
default n
select MACHINE_OPTS_ARMV7M
select LIBC_ARCH_MEMMOVE
depends on ARCH_TOOLCHAIN_GNU
---help---
Enable optimized ARMv7-M specific memmove() library function
config ARMV7M_STRCMP
bool "Enable optimized strcmp() for ARMv7-M"
default n
@@ -23,6 +50,15 @@ config ARMV7M_STRCMP
---help---
Enable optimized ARMv7-M specific strcmp() library function
config ARMV7M_STRLEN
bool "Enable optimized strlen() for ARMv7-M"
default n
select MACHINE_OPTS_ARMV7M
select LIBC_ARCH_STRLEN
depends on ARCH_TOOLCHAIN_GNU
---help---
Enable optimized ARMv7-M specific strlen() library function
config ARMV7M_LIBM
bool "Architecture specific FPU optimizations"
default n
+16
View File
@@ -33,14 +33,30 @@
#
############################################################################
ifeq ($(CONFIG_ARMV7M_MEMCHR),y)
ASRCS += arch_memchr.S
endif
ifeq ($(CONFIG_ARMV7M_MEMCPY),y)
ASRCS += arch_memcpy.S
endif
ifeq ($(CONFIG_ARMV7M_MEMSET),y)
ASRCS += arch_memset.S
endif
ifeq ($(CONFIG_ARMV7M_MEMMOVE),y)
ASRCS += arch_memmove.S
endif
ifeq ($(CONFIG_ARMV7M_STRCMP),y)
ASRCS += arch_strcmp.S
endif
ifeq ($(CONFIG_ARMV7M_STRLEN),y)
ASRCS += arch_strlen.S
endif
ifeq ($(CONFIG_MACHINE_OPTS_ARMV7M),y)
DEPPATH += --dep-path machine/arm/armv7-m/gnu
VPATH += :machine/arm/armv7-m/gnu
@@ -0,0 +1,181 @@
/*
* Copyright (c) 2014 ARM Ltd
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the company may not be used to endorse or promote
* products derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY ARM LTD ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL ARM LTD BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
* TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __ARM_ARCH
/* ACLE standardises a set of pre-defines that describe the ARM architecture.
These were mostly implemented in GCC around GCC-4.8; older versions
have no, or only partial support. To provide a level of backwards
compatibility we try to work out what the definitions should be, given
the older pre-defines that GCC did produce. This isn't complete, but
it should be enough for use by routines that depend on this header. */
/* No need to handle ARMv8, GCC had ACLE support before that. */
# ifdef __ARM_ARCH_7__
/* The common subset of ARMv7 in all profiles. */
# define __ARM_ARCH 7
# define __ARM_ARCH_ISA_THUMB 2
# define __ARM_FEATURE_CLZ
# define __ARM_FEATURE_LDREX 7
# define __ARM_FEATURE_UNALIGNED
# endif
# if defined (__ARM_ARCH_7A__) || defined (__ARM_ARCH_7R__)
# define __ARM_ARCH 7
# define __ARM_ARCH_ISA_THUMB 2
# define __ARM_ARCH_ISA_ARM
# define __ARM_FEATURE_CLZ
# define __ARM_FEATURE_SIMD32
# define __ARM_FEATURE_DSP
# define __ARM_FEATURE_QBIT
# define __ARM_FEATURE_SAT
# define __ARM_FEATURE_LDREX 15
# define __ARM_FEATURE_UNALIGNED
# ifdef __ARM_ARCH_7A__
# define __ARM_ARCH_PROFILE 'A'
# else
# define __ARM_ARCH_PROFILE 'R'
# endif
# endif
# ifdef __ARM_ARCH_7EM__
# define __ARM_ARCH 7
# define __ARM_ARCH_ISA_THUMB 2
# define __ARM_FEATURE_CLZ
# define __ARM_FEATURE_SIMD32
# define __ARM_FEATURE_DSP
# define __ARM_FEATURE_QBIT
# define __ARM_FEATURE_SAT
# define __ARM_FEATURE_LDREX 7
# define __ARM_FEATURE_UNALIGNED
# define __ARM_ARCH_PROFILE 'M'
# endif
# ifdef __ARM_ARCH_7M__
# define __ARM_ARCH 7
# define __ARM_ARCH_ISA_THUMB 2
# define __ARM_FEATURE_CLZ
# define __ARM_FEATURE_QBIT
# define __ARM_FEATURE_SAT
# define __ARM_FEATURE_LDREX 7
# define __ARM_FEATURE_UNALIGNED
# define __ARM_ARCH_PROFILE 'M'
# endif
# ifdef __ARM_ARCH_6T2__
# define __ARM_ARCH 6
# define __ARM_ARCH_ISA_THUMB 2
# define __ARM_ARCH_ISA_ARM
# define __ARM_FEATURE_CLZ
# define __ARM_FEATURE_SIMD32
# define __ARM_FEATURE_DSP
# define __ARM_FEATURE_QBIT
# define __ARM_FEATURE_SAT
# define __ARM_FEATURE_LDREX 4
# define __ARM_FEATURE_UNALIGNED
# endif
# ifdef __ARM_ARCH_6M__
# define __ARM_ARCH 6
# define __ARM_ARCH_ISA_THUMB 1
# define __ARM_ARCH_PROFILE 'M'
# endif
# if defined (__ARM_ARCH_6__) || defined (__ARM_ARCH_6J__) \
|| defined (__ARM_ARCH_6K__) || defined (__ARM_ARCH_6Z__) \
|| defined (__ARM_ARCH_6ZK__)
# define __ARM_ARCH 6
# define __ARM_ARCH_ISA_THUMB 1
# define __ARM_ARCH_ISA_ARM
# define __ARM_FEATURE_CLZ
# define __ARM_FEATURE_SIMD32
# define __ARM_FEATURE_DSP
# define __ARM_FEATURE_QBIT
# define __ARM_FEATURE_SAT
# define __ARM_FEATURE_UNALIGNED
# ifndef __thumb__
# if defined (__ARM_ARCH_6K__) || defined (__ARM_ARCH_6ZK__)
# define __ARM_FEATURE_LDREX 15
# else
# define __ARM_FEATURE_LDREX 4
# endif
# endif
# endif
# if defined (__ARM_ARCH_5TE__) || defined (__ARM_ARCH_5E__)
# define __ARM_ARCH 5
# define __ARM_ARCH_ISA_ARM
# ifdef __ARM_ARCH_5TE__
# define __ARM_ARCH_ISA_THUMB 1
# endif
# define __ARM_FEATURE_CLZ
# define __ARM_FEATURE_DSP
# endif
# if defined (__ARM_ARCH_5T__) || defined (__ARM_ARCH_5__)
# define __ARM_ARCH 5
# define __ARM_ARCH_ISA_ARM
# ifdef __ARM_ARCH_5TE__
# define __ARM_ARCH_ISA_THUMB 1
# endif
# define __ARM_FEATURE_CLZ
# endif
# ifdef __ARM_ARCH_4T__
# define __ARM_ARCH 4
# define __ARM_ARCH_ISA_ARM
# define __ARM_ARCH_ISA_THUMB 1
# endif
# ifdef __ARM_ARCH_4__
# define __ARM_ARCH 4
# define __ARM_ARCH_ISA_ARM
# endif
# if defined (__ARM_ARCH_3__) || defined (__ARM_ARCH_3M__)
# define __ARM_ARCH 3
# define __ARM_ARCH_ISA_ARM
# endif
# ifdef __ARM_ARCH_2__
# define __ARM_ARCH 2
# define __ARM_ARCH_ISA_ARM
# endif
# ifdef __ARMEB__
# define __ARM_BIG_ENDIAN
# endif
/* If we still don't know what the target architecture is, then we're
probably not using GCC. */
# ifndef __ARM_ARCH
# error Unable to determine architecture version.
# endif
#endif /* __ARM_ARCH */
@@ -0,0 +1,386 @@
/* Copyright (c) 2010-2011, Linaro Limited
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of Linaro Limited nor the names of its
contributors may be used to endorse or promote products derived
from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Written by Dave Gilbert <david.gilbert@linaro.org>
This memchr routine is optimised on a Cortex-A9 and should work on
all ARMv7 processors. It has a fast path for short sizes, and has
an optimised path for large data sets; the worst case is finding the
match early in a large data set. */
/* Copyright (c) 2015 ARM Ltd.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the Linaro nor the
names of its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
@ 2011-02-07 david.gilbert@linaro.org
@ Extracted from local git a5b438d861
@ 2011-07-14 david.gilbert@linaro.org
@ Import endianness fix from local git ea786f1b
@ 2011-10-11 david.gilbert@linaro.org
@ Import from cortex-strings bzr rev 63
@ Flip to ldrd (as suggested by Greta Yorsh)
@ Make conditional on CPU type
@ tidy
@ This code requires armv6t2 or later. Uses Thumb2.
.syntax unified
#include "acle-compat.h"
@ NOTE: This ifdef MUST match the one in memchr-stub.c
#if defined (__ARM_NEON__) || defined (__ARM_NEON)
#if __ARM_ARCH >= 8 && __ARM_ARCH_PROFILE == 'R'
.arch armv8-r
#else
.arch armv7-a
#endif
.fpu neon
/* Arguments */
#define srcin r0
#define chrin r1
#define cntin r2
/* Retval */
#define result r0 /* Live range does not overlap with srcin */
/* Working registers */
#define src r1 /* Live range does not overlap with chrin */
#define tmp r3
#define synd r0 /* No overlap with srcin or result */
#define soff r12
/* Working NEON registers */
#define vrepchr q0
#define vdata0 q1
#define vdata0_0 d2 /* Lower half of vdata0 */
#define vdata0_1 d3 /* Upper half of vdata0 */
#define vdata1 q2
#define vdata1_0 d4 /* Lower half of vhas_chr0 */
#define vdata1_1 d5 /* Upper half of vhas_chr0 */
#define vrepmask q3
#define vrepmask0 d6
#define vrepmask1 d7
#define vend q4
#define vend0 d8
#define vend1 d9
/*
* Core algorithm:
*
* For each 32-byte chunk we calculate a 32-bit syndrome value, with one bit per
* byte. Each bit is set if the relevant byte matched the requested character
* and cleared otherwise. Since the bits in the syndrome reflect exactly the
* order in which things occur in the original string, counting trailing zeros
* allows to identify exactly which byte has matched.
*/
.text
.thumb_func
.align 4
.p2align 4,,15
.global memchr
.type memchr,%function
memchr:
.cfi_sections .debug_frame
.cfi_startproc
/* Use a simple loop if there are less than 8 bytes to search. */
cmp cntin, #7
bhi .Llargestr
and chrin, chrin, #0xff
.Lsmallstr:
subs cntin, cntin, #1
blo .Lnotfound /* Return not found if reached end. */
ldrb tmp, [srcin], #1
cmp tmp, chrin
bne .Lsmallstr /* Loop again if not found. */
/* Otherwise fixup address and return. */
sub result, result, #1
bx lr
.Llargestr:
vdup.8 vrepchr, chrin /* Duplicate char across all lanes. */
/*
* Magic constant 0x8040201008040201 allows us to identify which lane
* matches the requested byte.
*/
movw tmp, #0x0201
movt tmp, #0x0804
lsl soff, tmp, #4
vmov vrepmask0, tmp, soff
vmov vrepmask1, tmp, soff
/* Work with aligned 32-byte chunks */
bic src, srcin, #31
ands soff, srcin, #31
beq .Lloopintro /* Go straight to main loop if it's aligned. */
/*
* Input string is not 32-byte aligned. We calculate the syndrome
* value for the aligned 32 bytes block containing the first bytes
* and mask the irrelevant part.
*/
vld1.8 {vdata0, vdata1}, [src:256]!
sub tmp, soff, #32
adds cntin, cntin, tmp
vceq.i8 vdata0, vdata0, vrepchr
vceq.i8 vdata1, vdata1, vrepchr
vand vdata0, vdata0, vrepmask
vand vdata1, vdata1, vrepmask
vpadd.i8 vdata0_0, vdata0_0, vdata0_1
vpadd.i8 vdata1_0, vdata1_0, vdata1_1
vpadd.i8 vdata0_0, vdata0_0, vdata1_0
vpadd.i8 vdata0_0, vdata0_0, vdata0_0
vmov synd, vdata0_0[0]
/* Clear the soff lower bits */
lsr synd, synd, soff
lsl synd, synd, soff
/* The first block can also be the last */
bls .Lmasklast
/* Have we found something already? */
cbnz synd, .Ltail
.Lloopintro:
vpush {vend}
/* 264/265 correspond to d8/d9 for q4 */
.cfi_adjust_cfa_offset 16
.cfi_rel_offset 264, 0
.cfi_rel_offset 265, 8
.p2align 3,,7
.Lloop:
vld1.8 {vdata0, vdata1}, [src:256]!
subs cntin, cntin, #32
vceq.i8 vdata0, vdata0, vrepchr
vceq.i8 vdata1, vdata1, vrepchr
/* If we're out of data we finish regardless of the result. */
bls .Lend
/* Use a fast check for the termination condition. */
vorr vend, vdata0, vdata1
vorr vend0, vend0, vend1
vmov synd, tmp, vend0
orrs synd, synd, tmp
/* We're not out of data, loop if we haven't found the character. */
beq .Lloop
.Lend:
vpop {vend}
.cfi_adjust_cfa_offset -16
.cfi_restore 264
.cfi_restore 265
/* Termination condition found, let's calculate the syndrome value. */
vand vdata0, vdata0, vrepmask
vand vdata1, vdata1, vrepmask
vpadd.i8 vdata0_0, vdata0_0, vdata0_1
vpadd.i8 vdata1_0, vdata1_0, vdata1_1
vpadd.i8 vdata0_0, vdata0_0, vdata1_0
vpadd.i8 vdata0_0, vdata0_0, vdata0_0
vmov synd, vdata0_0[0]
cbz synd, .Lnotfound
bhi .Ltail
.Lmasklast:
/* Clear the (-cntin) upper bits to avoid out-of-bounds matches. */
neg cntin, cntin
lsl synd, synd, cntin
lsrs synd, synd, cntin
it eq
moveq src, #0 /* If no match, set src to 0 so the retval is 0. */
.Ltail:
/* Count the trailing zeros using bit reversing */
rbit synd, synd
/* Compensate the last post-increment */
sub src, src, #32
/* Count the leading zeros */
clz synd, synd
/* Compute the potential result and return */
add result, src, synd
bx lr
.Lnotfound:
/* Set result to NULL if not found and return */
mov result, #0
bx lr
.cfi_endproc
.size memchr, . - memchr
#elif __ARM_ARCH_ISA_THUMB >= 2 && defined (__ARM_FEATURE_DSP)
#if __ARM_ARCH_PROFILE == 'M'
.arch armv7e-m
#else
.arch armv6t2
#endif
@ this lets us check a flag in a 00/ff byte easily in either endianness
#ifdef __ARMEB__
#define CHARTSTMASK(c) 1<<(31-(c*8))
#else
#define CHARTSTMASK(c) 1<<(c*8)
#endif
.text
.thumb
@ ---------------------------------------------------------------------------
.thumb_func
.align 2
.p2align 4,,15
.global memchr
.type memchr,%function
memchr:
@ r0 = start of memory to scan
@ r1 = character to look for
@ r2 = length
@ returns r0 = pointer to character or NULL if not found
and r1,r1,#0xff @ Don't trust the caller to pass a char
cmp r2,#16 @ If short don't bother with anything clever
blt 20f
tst r0, #7 @ If it's already aligned skip the next bit
beq 10f
@ Work up to an aligned point
5:
ldrb r3, [r0],#1
subs r2, r2, #1
cmp r3, r1
beq 50f @ If it matches exit found
tst r0, #7
cbz r2, 40f @ If we run off the end, exit not found
bne 5b @ If not aligned yet then do next byte
10:
@ We are aligned, we know we have at least 8 bytes to work with
push {r4,r5,r6,r7}
orr r1, r1, r1, lsl #8 @ expand the match word across all bytes
orr r1, r1, r1, lsl #16
bic r4, r2, #7 @ Number of double words to work with * 8
mvns r7, #0 @ all F's
movs r3, #0
15:
ldrd r5,r6,[r0],#8
subs r4, r4, #8
eor r5,r5, r1 @ r5,r6 have 00's where bytes match the target
eor r6,r6, r1
uadd8 r5, r5, r7 @ Par add 0xff - sets GE bits for bytes!=0
sel r5, r3, r7 @ bytes are 00 for none-00 bytes,
@ or ff for 00 bytes - NOTE INVERSION
uadd8 r6, r6, r7 @ Par add 0xff - sets GE bits for bytes!=0
sel r6, r5, r7 @ chained....bytes are 00 for none-00 bytes
@ or ff for 00 bytes - NOTE INVERSION
cbnz r6, 60f
bne 15b @ (Flags from the subs above)
pop {r4,r5,r6,r7}
and r1,r1,#0xff @ r1 back to a single character
and r2,r2,#7 @ Leave the count remaining as the number
@ after the double words have been done
20:
cbz r2, 40f @ 0 length or hit the end already then not found
21: @ Post aligned section, or just a short call
ldrb r3,[r0],#1
subs r2,r2,#1
eor r3,r3,r1 @ r3 = 0 if match - doesn't break flags from sub
cbz r3, 50f
bne 21b @ on r2 flags
40:
movs r0,#0 @ not found
bx lr
50:
subs r0,r0,#1 @ found
bx lr
60: @ We're here because the fast path found a hit
@ now we have to track down exactly which word it was
@ r0 points to the start of the double word after the one tested
@ r5 has the 00/ff pattern for the first word, r6 has the chained value
cmp r5, #0
itte eq
moveq r5, r6 @ the end is in the 2nd word
subeq r0,r0,#3 @ Points to 2nd byte of 2nd word
subne r0,r0,#7 @ or 2nd byte of 1st word
@ r0 currently points to the 2nd byte of the word containing the hit
tst r5, # CHARTSTMASK(0) @ 1st character
bne 61f
adds r0,r0,#1
tst r5, # CHARTSTMASK(1) @ 2nd character
ittt eq
addeq r0,r0,#1
tsteq r5, # (3<<15) @ 2nd & 3rd character
@ If not the 3rd must be the last one
addeq r0,r0,#1
61:
pop {r4,r5,r6,r7}
subs r0,r0,#1
bx lr
#else
/* Defined in memchr-stub.c. */
#endif
@@ -0,0 +1,63 @@
/*
* Copyright (c) 2015 ARM Ltd
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the company may not be used to endorse or promote
* products derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY ARM LTD ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL ARM LTD BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
* TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
.thumb
.syntax unified
.global memmove
.type memmove, %function
memmove:
cmp r0, r1
push {r4}
bls 3f
adds r3, r1, r2
cmp r0, r3
bcs 3f
adds r1, r0, r2
cbz r2, 2f
subs r2, r3, r2
1:
ldrb r4, [r3, #-1]!
cmp r2, r3
strb r4, [r1, #-1]!
bne 1b
2:
pop {r4}
bx lr
3:
cmp r2, #0
beq 2b
add r2, r2, r1
subs r3, r0, #1
4:
ldrb r4, [r1], #1
cmp r2, r1
strb r4, [r3, #1]!
bne 4b
pop {r4}
bx lr
.size memmove, . - memmove
@@ -0,0 +1,105 @@
/*
* Copyright (c) 2015 ARM Ltd
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the company may not be used to endorse or promote
* products derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY ARM LTD ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL ARM LTD BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
* TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
.thumb
.syntax unified
.global memset
.type memset, %function
memset:
push {r4, r5, r6}
lsls r4, r0, #30
beq 10f
subs r4, r1, #1
cmp r1, #0
beq 9f
uxtb r5, r2
mov r3, r0
b 2f
1:
subs r1, r4, #1
cbz r4, 9f
mov r4, r1
2:
strb r5, [r3], #1
lsls r1, r3, #30
bne 1b
3:
cmp r4, #3
bls 7f
uxtb r5, r2
orr r5, r5, r5, lsl #8
cmp r4, #15
orr r5, r5, r5, lsl #16
bls 5f
mov r6, r4
add r1, r3, #16
4:
subs r6, r6, #16
cmp r6, #15
str r5, [r1, #-16]
str r5, [r1, #-12]
str r5, [r1, #-8]
str r5, [r1, #-4]
add r1, r1, #16
bhi 4b
sub r1, r4, #16
bic r1, r1, #15
and r4, r4, #15
adds r1, r1, #16
cmp r4, #3
add r3, r3, r1
bls 7f
5:
mov r6, r3
mov r1, r4
6:
subs r1, r1, #4
cmp r1, #3
str r5, [r6], #4
bhi 6b
subs r1, r4, #4
bic r1, r1, #3
adds r1, r1, #4
add r3, r3, r1
and r4, r4, #3
7:
cbz r4, 9f
uxtb r2, r2
add r4, r4, r3
8:
strb r2, [r3], #1
cmp r3, r4
bne 8b
9:
pop {r4, r5, r6}
bx lr
10:
mov r4, r1
mov r3, r0
b 3b
.size memset, . - memset
@@ -0,0 +1,179 @@
/* Copyright (c) 2010-2011,2013 Linaro Limited
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of Linaro Limited nor the names of its
contributors may be used to endorse or promote products derived
from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Assumes:
ARMv6T2 or ARMv7E-M, AArch32
*/
/* Copyright (c) 2015 ARM Ltd.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the Linaro nor the
names of its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */
#include "acle-compat.h"
.macro def_fn f p2align=0
.text
.p2align \p2align
.global \f
.type \f, %function
\f:
.endm
#ifdef __ARMEB__
#define S2LO lsl
#define S2HI lsr
#else
#define S2LO lsr
#define S2HI lsl
#endif
/* This code requires Thumb. */
#if __ARM_ARCH_PROFILE == 'M'
.arch armv7e-m
#else
.arch armv6t2
#endif
.eabi_attribute Tag_ARM_ISA_use, 0
.thumb
.syntax unified
/* Parameters and result. */
#define srcin r0
#define result r0
/* Internal variables. */
#define src r1
#define data1a r2
#define data1b r3
#define const_m1 r12
#define const_0 r4
#define tmp1 r4 /* Overlaps const_0 */
#define tmp2 r5
def_fn strlen p2align=6
pld [srcin, #0]
strd r4, r5, [sp, #-8]!
bic src, srcin, #7
mvn const_m1, #0
ands tmp1, srcin, #7 /* (8 - bytes) to alignment. */
pld [src, #32]
bne.w .Lmisaligned8
mov const_0, #0
mov result, #-8
.Lloop_aligned:
/* Bytes 0-7. */
ldrd data1a, data1b, [src]
pld [src, #64]
add result, result, #8
.Lstart_realigned:
uadd8 data1a, data1a, const_m1 /* Saturating GE<0:3> set. */
sel data1a, const_0, const_m1 /* Select based on GE<0:3>. */
uadd8 data1b, data1b, const_m1
sel data1b, data1a, const_m1 /* Only used if d1a == 0. */
cbnz data1b, .Lnull_found
/* Bytes 8-15. */
ldrd data1a, data1b, [src, #8]
uadd8 data1a, data1a, const_m1 /* Saturating GE<0:3> set. */
add result, result, #8
sel data1a, const_0, const_m1 /* Select based on GE<0:3>. */
uadd8 data1b, data1b, const_m1
sel data1b, data1a, const_m1 /* Only used if d1a == 0. */
cbnz data1b, .Lnull_found
/* Bytes 16-23. */
ldrd data1a, data1b, [src, #16]
uadd8 data1a, data1a, const_m1 /* Saturating GE<0:3> set. */
add result, result, #8
sel data1a, const_0, const_m1 /* Select based on GE<0:3>. */
uadd8 data1b, data1b, const_m1
sel data1b, data1a, const_m1 /* Only used if d1a == 0. */
cbnz data1b, .Lnull_found
/* Bytes 24-31. */
ldrd data1a, data1b, [src, #24]
add src, src, #32
uadd8 data1a, data1a, const_m1 /* Saturating GE<0:3> set. */
add result, result, #8
sel data1a, const_0, const_m1 /* Select based on GE<0:3>. */
uadd8 data1b, data1b, const_m1
sel data1b, data1a, const_m1 /* Only used if d1a == 0. */
cmp data1b, #0
beq .Lloop_aligned
.Lnull_found:
cmp data1a, #0
itt eq
addeq result, result, #4
moveq data1a, data1b
#ifndef __ARMEB__
rev data1a, data1a
#endif
clz data1a, data1a
ldrd r4, r5, [sp], #8
add result, result, data1a, lsr #3 /* Bits -> Bytes. */
bx lr
.Lmisaligned8:
ldrd data1a, data1b, [src]
and tmp2, tmp1, #3
rsb result, tmp1, #0
lsl tmp2, tmp2, #3 /* Bytes -> bits. */
tst tmp1, #4
pld [src, #64]
S2HI tmp2, const_m1, tmp2
orn data1a, data1a, tmp2
itt ne
ornne data1b, data1b, tmp2
movne data1a, const_m1
mov const_0, #0
b .Lstart_realigned
.size strlen, . - strlen
+22 -2
View File
@@ -37,8 +37,8 @@
# Add the string C files to the build
CSRCS += lib_ffs.c lib_ffsl.c lib_ffsll.c lib_fls.c lib_flsl.c
CSRCS += lib_flsll.c lib_isbasedigit.c lib_memset.c lib_memchr.c
CSRCS += lib_memccpy.c lib_memcmp.c lib_memmove.c lib_memrchr.c
CSRCS += lib_flsll.c lib_isbasedigit.c
CSRCS += lib_memccpy.c lib_memcmp.c lib_memrchr.c
CSRCS += lib_popcount.c lib_popcountl.c lib_popcountll.c
CSRCS += lib_skipspace.c lib_stpcpy.c lib_stpncpy.c lib_strcasecmp.c
CSRCS += lib_strcat.c lib_strchr.c lib_strcpy.c lib_strcspn.c
@@ -51,6 +51,10 @@ CSRCS += lib_anbstr2cstr.c lib_ancstr2bstr.c lib_bmem2cmem.c
CSRCS += lib_bstrnlen.c lib_cmem2bmem.c lib_nbstr2cstr.c lib_ncstr2bstr.c
CSRCS += lib_index.c lib_rindex.c lib_strlcpy.c
ifneq ($(CONFIG_LIBC_ARCH_MEMCHR),y)
CSRCS += lib_memchr.c
endif
ifneq ($(CONFIG_LIBC_ARCH_MEMCPY),y)
ifeq ($(CONFIG_MEMCPY_VIK),y)
CSRCS += lib_vikmemcpy.c
@@ -59,10 +63,26 @@ CSRCS += lib_memcpy.c
endif
endif
ifneq ($(CONFIG_LIBC_ARCH_MEMSET),y)
CSRCS += lib_memset.c
endif
ifneq ($(CONFIG_LIBC_ARCH_MEMMOVE),y)
CSRCS += lib_memmove.c
endif
ifneq ($(CONFIG_LIBC_ARCH_STRCPY),y)
CSRCS += lib_strcpy.c
endif
ifneq ($(CONFIG_LIBC_ARCH_STRCMP),y)
CSRCS += lib_strcmp.c
endif
ifneq ($(CONFIG_LIBC_ARCH_STRLEN),y)
CSRCS += lib_strlen.c
endif
ifeq ($(CONFIG_LIBC_LOCALE),y)
CSRCS += lib_strcoll.c lib_strxfrm.c
endif