1 /* Copyright (C) 2006, 2008, 2009 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, write to the Free
18 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
24 * Data preload for architectures that support it (ARM V5TE and above)
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
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.
40 //#define CALGN(code...) code
41 #define CALGN(code...)
44 * Endian independent macros for shifting bytes within registers.
57 * Prototype: void *memmove(void *dest, const void *src, size_t n);
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.
72 bls HIDDEN_JUMPTARGET(memcpy)
75 stmfd sp!, {r0, r4, lr}
76 cfi_adjust_cfa_offset (12)
77 cfi_rel_offset (r4, 4)
78 cfi_rel_offset (lr, 8)
94 cfi_adjust_cfa_offset (16)
95 cfi_rel_offset (r5, 0)
96 cfi_rel_offset (r6, 4)
97 cfi_rel_offset (r7, 8)
98 cfi_rel_offset (r8, 12)
101 CALGN( ands ip, r1, #31 )
102 CALGN( sbcnes r4, ip, r2 ) @ C is always set here
105 CALGN( subs r2, r2, ip ) @ C is set here
106 CALGN( add pc, r4, ip )
109 2: PLD( subs r2, r2, #96 )
110 PLD( pld [r1, #-32] )
112 PLD( pld [r1, #-64] )
113 PLD( pld [r1, #-96] )
115 3: PLD( pld [r1, #-128] )
116 4: ldmdb r1!, {r3, r4, r5, r6, r7, r8, ip, lr}
118 stmdb r0!, {r3, r4, r5, r6, r7, r8, ip, lr}
125 addne pc, pc, ip @ C is always clear here
149 7: ldmfd sp!, {r5 - r8}
150 cfi_adjust_cfa_offset (-16)
156 8: movs r2, r2, lsl #31
157 ldrneb r3, [r1, #-1]!
158 ldrcsb r4, [r1, #-1]!
160 strneb r3, [r0, #-1]!
161 strcsb r4, [r0, #-1]!
164 #if defined (__ARM_ARCH_4T__) && defined (__THUMB_INTERWORK__)
165 ldmfd sp!, {r0, r4, lr}
166 cfi_adjust_cfa_offset (-12)
171 ldmfd sp!, {r0, r4, pc}
177 ldrgtb r3, [r1, #-1]!
178 ldrgeb r4, [r1, #-1]!
180 strgtb r3, [r0, #-1]!
181 strgeb r4, [r0, #-1]!
195 .macro backward_copy_shift push pull
200 CALGN( ands ip, r1, #31 )
201 CALGN( rsb ip, ip, #32 )
202 CALGN( sbcnes r4, ip, r2 ) @ C is always set here
203 CALGN( subcc r2, r2, ip )
206 11: stmfd sp!, {r5 - r9}
207 cfi_adjust_cfa_offset (20)
208 cfi_rel_offset (r5, 0)
209 cfi_rel_offset (r6, 4)
210 cfi_rel_offset (r7, 8)
211 cfi_rel_offset (r8, 12)
212 cfi_rel_offset (r9, 16)
215 PLD( subs r2, r2, #96 )
216 PLD( pld [r1, #-32] )
218 PLD( pld [r1, #-64] )
219 PLD( pld [r1, #-96] )
221 12: PLD( pld [r1, #-128] )
222 13: ldmdb r1!, {r7, r8, r9, ip}
223 mov lr, r3, push #\push
225 ldmdb r1!, {r3, r4, r5, r6}
226 orr lr, lr, ip, pull #\pull
227 mov ip, ip, push #\push
228 orr ip, ip, r9, pull #\pull
229 mov r9, r9, push #\push
230 orr r9, r9, r8, pull #\pull
231 mov r8, r8, push #\push
232 orr r8, r8, r7, pull #\pull
233 mov r7, r7, push #\push
234 orr r7, r7, r6, pull #\pull
235 mov r6, r6, push #\push
236 orr r6, r6, r5, pull #\pull
237 mov r5, r5, push #\push
238 orr r5, r5, r4, pull #\pull
239 mov r4, r4, push #\push
240 orr r4, r4, r3, pull #\pull
241 stmdb r0!, {r4 - r9, ip, lr}
247 cfi_adjust_cfa_offset (-20)
257 15: mov lr, r3, push #\push
260 orr lr, lr, r3, pull #\pull
266 16: add r1, r1, #(\pull / 8)
272 backward_copy_shift push=8 pull=24
274 17: backward_copy_shift push=16 pull=16
276 18: backward_copy_shift push=24 pull=8
280 libc_hidden_builtin_def (memmove)