« back to all changes in this revision
Viewing changes to src/reference/glibc/memmove.S
- browse files at revision 1
- compare with another revision
- download diff
- download tarball
- view history from revision 1
- Committer: Michael Hope
- Date: 2010-08-26 22:19:29 UTC
- Revision ID: michael.hope@linaro.org-20100826221929-ppeg01mnpx34aqrp
Pulled in the initial versions
- files added:
- src
- src/reference
- src/reference/bionic
- src/reference/glibc
- src/reference/helpers
- src/reference/newlib
- src/reference/plain
added
removed
src/reference/glibc/memmove.S
1
/* Copyright (C) 2006, 2008, 2009 Free Software Foundation, Inc.
2
This file is part of the GNU C Library.
3
4
Contributed by MontaVista Software, Inc. (written by Nicolas Pitre)
5
6
The GNU C Library is free software; you can redistribute it and/or
7
modify it under the terms of the GNU Lesser General Public
8
License as published by the Free Software Foundation; either
9
version 2.1 of the License, or (at your option) any later version.
10
11
The GNU C Library is distributed in the hope that it will be useful,
12
but WITHOUT ANY WARRANTY; without even the implied warranty of
13
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14
Lesser General Public License for more details.
15
16
You should have received a copy of the GNU Lesser General Public
17
License along with the GNU C Library; if not, write to the Free
18
Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
19
02111-1307 USA. */
20
21
#include <sysdep.h>
22
23
/*
24
* Data preload for architectures that support it (ARM V5TE and above)
25
*/
26
#if (!defined (__ARM_ARCH_2__) && !defined (__ARM_ARCH_3__) \
27
&& !defined (__ARM_ARCH_3M__) && !defined (__ARM_ARCH_4__) \
28
&& !defined (__ARM_ARCH_4T__) && !defined (__ARM_ARCH_5__) \
29
&& !defined (__ARM_ARCH_5T__))
30
#define PLD(code...) code
31
#else
32
#define PLD(code...)
33
#endif
34
35
/*
36
* This can be used to enable code to cacheline align the source pointer.
37
* Experiments on tested architectures (StrongARM and XScale) didn't show
38
* this a worthwhile thing to do. That might be different in the future.
39
*/
40
//#define CALGN(code...) code
41
#define CALGN(code...)
42
43
/*
44
* Endian independent macros for shifting bytes within registers.
45
*/
46
#ifndef __ARMEB__
47
#define pull lsr
48
#define push lsl
49
#else
50
#define pull lsl
51
#define push lsr
52
#endif
53
54
.text
55
56
/*
57
* Prototype: void *memmove(void *dest, const void *src, size_t n);
58
*
59
* Note:
60
*
61
* If the memory regions don't overlap, we simply branch to memcpy which is
62
* normally a bit faster. Otherwise the copy is done going downwards.
63
*/
64
65
ENTRY(memmove)
66
67
subs ip, r0, r1
68
cmphi r2, ip
69
#ifdef NOT_IN_libc
70
bls memcpy
71
#else
72
bls HIDDEN_JUMPTARGET(memcpy)
73
#endif
74
75
stmfd sp!, {r0, r4, lr}
76
add r1, r1, r2
77
add r0, r0, r2
78
subs r2, r2, #4
79
blt 8f
80
ands ip, r0, #3
81
PLD( pld [r1, #-4] )
82
bne 9f
83
ands ip, r1, #3
84
bne 10f
85
86
1: subs r2, r2, #(28)
87
stmfd sp!, {r5 - r8}
88
blt 5f
89
90
CALGN( ands ip, r1, #31 )
91
CALGN( sbcnes r4, ip, r2 ) @ C is always set here
92
CALGN( bcs 2f )
93
CALGN( adr r4, 6f )
94
CALGN( subs r2, r2, ip ) @ C is set here
95
CALGN( add pc, r4, ip )
96
97
PLD( pld [r1, #-4] )
98
2: PLD( subs r2, r2, #96 )
99
PLD( pld [r1, #-32] )
100
PLD( blt 4f )
101
PLD( pld [r1, #-64] )
102
PLD( pld [r1, #-96] )
103
104
3: PLD( pld [r1, #-128] )
105
4: ldmdb r1!, {r3, r4, r5, r6, r7, r8, ip, lr}
106
subs r2, r2, #32
107
stmdb r0!, {r3, r4, r5, r6, r7, r8, ip, lr}
108
bge 3b
109
PLD( cmn r2, #96 )
110
PLD( bge 4b )
111
112
5: ands ip, r2, #28
113
rsb ip, ip, #32
114
addne pc, pc, ip @ C is always clear here
115
b 7f
116
6: nop
117
ldr r3, [r1, #-4]!
118
ldr r4, [r1, #-4]!
119
ldr r5, [r1, #-4]!
120
ldr r6, [r1, #-4]!
121
ldr r7, [r1, #-4]!
122
ldr r8, [r1, #-4]!
123
ldr lr, [r1, #-4]!
124
125
add pc, pc, ip
126
nop
127
nop
128
str r3, [r0, #-4]!
129
str r4, [r0, #-4]!
130
str r5, [r0, #-4]!
131
str r6, [r0, #-4]!
132
str r7, [r0, #-4]!
133
str r8, [r0, #-4]!
134
str lr, [r0, #-4]!
135
136
CALGN( bcs 2b )
137
138
7: ldmfd sp!, {r5 - r8}
139
140
8: movs r2, r2, lsl #31
141
ldrneb r3, [r1, #-1]!
142
ldrcsb r4, [r1, #-1]!
143
ldrcsb ip, [r1, #-1]
144
strneb r3, [r0, #-1]!
145
strcsb r4, [r0, #-1]!
146
strcsb ip, [r0, #-1]
147
#if defined (__ARM_ARCH_4T__) && defined (__THUMB_INTERWORK__)
148
ldmfd sp!, {r0, r4, lr}
149
bx lr
150
#else
151
ldmfd sp!, {r0, r4, pc}
152
#endif
153
154
9: cmp ip, #2
155
ldrgtb r3, [r1, #-1]!
156
ldrgeb r4, [r1, #-1]!
157
ldrb lr, [r1, #-1]!
158
strgtb r3, [r0, #-1]!
159
strgeb r4, [r0, #-1]!
160
subs r2, r2, ip
161
strb lr, [r0, #-1]!
162
blt 8b
163
ands ip, r1, #3
164
beq 1b
165
166
10: bic r1, r1, #3
167
cmp ip, #2
168
ldr r3, [r1, #0]
169
beq 17f
170
blt 18f
171
172
173
.macro backward_copy_shift push pull
174
175
subs r2, r2, #28
176
blt 14f
177
178
CALGN( ands ip, r1, #31 )
179
CALGN( rsb ip, ip, #32 )
180
CALGN( sbcnes r4, ip, r2 ) @ C is always set here
181
CALGN( subcc r2, r2, ip )
182
CALGN( bcc 15f )
183
184
11: stmfd sp!, {r5 - r9}
185
186
PLD( pld [r1, #-4] )
187
PLD( subs r2, r2, #96 )
188
PLD( pld [r1, #-32] )
189
PLD( blt 13f )
190
PLD( pld [r1, #-64] )
191
PLD( pld [r1, #-96] )
192
193
12: PLD( pld [r1, #-128] )
194
13: ldmdb r1!, {r7, r8, r9, ip}
195
mov lr, r3, push #\push
196
subs r2, r2, #32
197
ldmdb r1!, {r3, r4, r5, r6}
198
orr lr, lr, ip, pull #\pull
199
mov ip, ip, push #\push
200
orr ip, ip, r9, pull #\pull
201
mov r9, r9, push #\push
202
orr r9, r9, r8, pull #\pull
203
mov r8, r8, push #\push
204
orr r8, r8, r7, pull #\pull
205
mov r7, r7, push #\push
206
orr r7, r7, r6, pull #\pull
207
mov r6, r6, push #\push
208
orr r6, r6, r5, pull #\pull
209
mov r5, r5, push #\push
210
orr r5, r5, r4, pull #\pull
211
mov r4, r4, push #\push
212
orr r4, r4, r3, pull #\pull
213
stmdb r0!, {r4 - r9, ip, lr}
214
bge 12b
215
PLD( cmn r2, #96 )
216
PLD( bge 13b )
217
218
ldmfd sp!, {r5 - r9}
219
220
14: ands ip, r2, #28
221
beq 16f
222
223
15: mov lr, r3, push #\push
224
ldr r3, [r1, #-4]!
225
subs ip, ip, #4
226
orr lr, lr, r3, pull #\pull
227
str lr, [r0, #-4]!
228
bgt 15b
229
CALGN( cmp r2, #0 )
230
CALGN( bge 11b )
231
232
16: add r1, r1, #(\pull / 8)
233
b 8b
234
235
.endm
236
237
238
backward_copy_shift push=8 pull=24
239
240
17: backward_copy_shift push=16 pull=16
241
242
18: backward_copy_shift push=24 pull=8
243
244
245
END(memmove)
246
libc_hidden_builtin_def (memmove)
Loggerhead is a web-based interface for Breezy
Version: 2.0.1