1 /* Copyright (C) 2006-2013 Free Software Foundation, Inc.
2 This file is part of the GNU C Library.
4 Contributed by MontaVista Software, Inc. (written by Nicolas Pitre)
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.
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.
16 You should have received a copy of the GNU Lesser General Public
17 License along with the GNU C Library. If not, see
18 <http://www.gnu.org/licenses/>. */
20 /* Thumb requires excessive IT insns here. */
23 #include <arm-features.h>
26 * Data preload for architectures that support it (ARM V5TE and above)
28 #if (!defined (__ARM_ARCH_2__) && !defined (__ARM_ARCH_3__) \
29 && !defined (__ARM_ARCH_3M__) && !defined (__ARM_ARCH_4__) \
30 && !defined (__ARM_ARCH_4T__) && !defined (__ARM_ARCH_5__) \
31 && !defined (__ARM_ARCH_5T__))
32 #define PLD(code...) code
38 * This can be used to enable code to cacheline align the source pointer.
39 * Experiments on tested architectures (StrongARM and XScale) didn't show
40 * this a worthwhile thing to do. That might be different in the future.
42 //#define CALGN(code...) code
43 #define CALGN(code...)
46 * Endian independent macros for shifting bytes within registers.
60 * Prototype: void *memmove(void *dest, const void *src, size_t n);
64 * If the memory regions don't overlap, we simply branch to memcpy which is
65 * normally a bit faster. Otherwise the copy is done going downwards.
75 bls HIDDEN_JUMPTARGET(memcpy)
79 cfi_adjust_cfa_offset (12)
80 cfi_rel_offset (r4, 4)
81 cfi_rel_offset (lr, 8)
97 cfi_adjust_cfa_offset (16)
98 cfi_rel_offset (r5, 0)
99 cfi_rel_offset (r6, 4)
100 cfi_rel_offset (r7, 8)
101 cfi_rel_offset (r8, 12)
104 CALGN( ands ip, r1, #31 )
105 CALGN( sbcsne r4, ip, r2 ) @ C is always set here
108 CALGN( subs r2, r2, ip ) @ C is set here
109 #ifndef ARM_ALWAYS_BX
110 CALGN( add pc, r4, ip, lsl #(ARM_BX_ALIGN_LOG2 - 2))
112 CALGN( add r4, r4, ip, lsl #(ARM_BX_ALIGN_LOG2 - 2))
117 2: PLD( subs r2, r2, #96 )
118 PLD( pld [r1, #-32] )
120 PLD( pld [r1, #-64] )
121 PLD( pld [r1, #-96] )
123 3: PLD( pld [r1, #-128] )
124 4: ldmdb r1!, {r3, r4, r5, r6, r7, r8, ip, lr}
126 stmdb r0!, {r3, r4, r5, r6, r7, r8, ip, lr}
133 #ifndef ARM_ALWAYS_BX
134 /* C is always clear here. */
135 addne pc, pc, ip, lsl #(ARM_BX_ALIGN_LOG2 - 2)
140 cfi_adjust_cfa_offset (4)
141 cfi_rel_offset (r10, 0)
142 add r10, pc, ip, lsl #(ARM_BX_ALIGN_LOG2 - 2)
145 .p2align ARM_BX_ALIGN_LOG2
147 .p2align ARM_BX_ALIGN_LOG2
149 .p2align ARM_BX_ALIGN_LOG2
151 .p2align ARM_BX_ALIGN_LOG2
153 .p2align ARM_BX_ALIGN_LOG2
155 .p2align ARM_BX_ALIGN_LOG2
157 .p2align ARM_BX_ALIGN_LOG2
159 .p2align ARM_BX_ALIGN_LOG2
162 #ifndef ARM_ALWAYS_BX
163 add pc, pc, ip, lsl #(ARM_BX_ALIGN_LOG2 - 2)
166 add r10, pc, ip, lsl #(ARM_BX_ALIGN_LOG2 - 2)
169 .p2align ARM_BX_ALIGN_LOG2
171 .p2align ARM_BX_ALIGN_LOG2
173 .p2align ARM_BX_ALIGN_LOG2
175 .p2align ARM_BX_ALIGN_LOG2
177 .p2align ARM_BX_ALIGN_LOG2
179 .p2align ARM_BX_ALIGN_LOG2
181 .p2align ARM_BX_ALIGN_LOG2
183 .p2align ARM_BX_ALIGN_LOG2
188 cfi_adjust_cfa_offset (-4)
195 cfi_adjust_cfa_offset (-16)
201 8: movs r2, r2, lsl #31
202 ldrbne r3, [r1, #-1]!
203 ldrbcs r4, [r1, #-1]!
205 strbne r3, [r0, #-1]!
206 strbcs r4, [r0, #-1]!
209 #if ((defined (__ARM_ARCH_4T__) && defined (__THUMB_INTERWORK__)) \
210 || defined (ARM_ALWAYS_BX))
212 cfi_adjust_cfa_offset (-12)
223 ldrbgt r3, [r1, #-1]!
224 ldrbge r4, [r1, #-1]!
226 strbgt r3, [r0, #-1]!
227 strbge r4, [r0, #-1]!
241 .macro backward_copy_shift push pull
246 CALGN( ands ip, r1, #31 )
247 CALGN( rsb ip, ip, #32 )
248 CALGN( sbcsne r4, ip, r2 ) @ C is always set here
249 CALGN( subcc r2, r2, ip )
252 11: push {r5 - r8, r10}
253 cfi_adjust_cfa_offset (20)
254 cfi_rel_offset (r5, 0)
255 cfi_rel_offset (r6, 4)
256 cfi_rel_offset (r7, 8)
257 cfi_rel_offset (r8, 12)
258 cfi_rel_offset (r10, 16)
261 PLD( subs r2, r2, #96 )
262 PLD( pld [r1, #-32] )
264 PLD( pld [r1, #-64] )
265 PLD( pld [r1, #-96] )
267 12: PLD( pld [r1, #-128] )
268 13: ldmdb r1!, {r7, r8, r10, ip}
269 mov lr, r3, PUSH #\push
271 ldmdb r1!, {r3, r4, r5, r6}
272 orr lr, lr, ip, PULL #\pull
273 mov ip, ip, PUSH #\push
274 orr ip, ip, r10, PULL #\pull
275 mov r10, r10, PUSH #\push
276 orr r10, r10, r8, PULL #\pull
277 mov r8, r8, PUSH #\push
278 orr r8, r8, r7, PULL #\pull
279 mov r7, r7, PUSH #\push
280 orr r7, r7, r6, PULL #\pull
281 mov r6, r6, PUSH #\push
282 orr r6, r6, r5, PULL #\pull
283 mov r5, r5, PUSH #\push
284 orr r5, r5, r4, PULL #\pull
285 mov r4, r4, PUSH #\push
286 orr r4, r4, r3, PULL #\pull
287 stmdb r0!, {r4 - r8, r10, ip, lr}
293 cfi_adjust_cfa_offset (-20)
303 15: mov lr, r3, PUSH #\push
306 orr lr, lr, r3, PULL #\pull
312 16: add r1, r1, #(\pull / 8)
318 backward_copy_shift push=8 pull=24
320 17: backward_copy_shift push=16 pull=16
322 18: backward_copy_shift push=24 pull=8
326 libc_hidden_builtin_def (memmove)