1 /*
2 * Copyright (C) 2008-2009 Michal Simek <monstr
3 * Copyright (C) 2008-2009 PetaLogix
4 * Copyright (C) 2008 Jim Law - Iris LP All r
5 *
6 * This file is subject to the terms and condi
7 * Public License. See the file COPYING in th
8 * archive for more details.
9 *
10 * Written by Jim Law <jlaw@irispower.com>
11 *
12 * intended to replace:
13 * memcpy in memcpy.c and
14 * memmove in memmove.c
15 * ... in arch/microblaze/lib
16 *
17 *
18 * assly_fastcopy.S
19 *
20 * Attempt at quicker memcpy and memmove for M
21 * Input : Operand1 in Reg r5 - destinati
22 * Operand2 in Reg r6 - source ad
23 * Operand3 in Reg r7 - number of
24 * Output: Result in Reg r3 - starting de
25 *
26 *
27 * Explanation:
28 * Perform (possibly unaligned) copy of a
29 * between mem locations with size of xfe
30 */
31
32 #include <linux/linkage.h>
33 .text
34 .globl memcpy
35 .type memcpy, @function
36 .ent memcpy
37
38 memcpy:
39 fast_memcpy_ascending:
40 /* move d to return register as value
41 addi r3, r5, 0
42
43 addi r4, r0, 4 /* n = 4 */
44 cmpu r4, r4, r7 /* n = c - n
45 blti r4, a_xfer_end /* if n < 0, l
46
47 /* transfer first 0~3 bytes to get ali
48 andi r4, r5, 3 /* n =
49 /* if zero, destination already aligne
50 beqi r4, a_dalign_done
51 /* n = 4 - n (yields 3, 2, 1 transfers
52 rsubi r4, r4, 4
53 rsub r7, r4, r7 /* c =
54
55 a_xfer_first_loop:
56 /* if no bytes left to transfer, trans
57 beqi r4, a_dalign_done
58 lbui r11, r6, 0 /* h =
59 sbi r11, r5, 0 /* *d
60 addi r6, r6, 1 /* s++
61 addi r5, r5, 1 /* d++
62 brid a_xfer_first_loop /* loo
63 addi r4, r4, -1 /* n--
64
65 a_dalign_done:
66 addi r4, r0, 32 /* n =
67 cmpu r4, r4, r7 /* n =
68 /* if n < 0, less than one block to tr
69 blti r4, a_block_done
70
71 a_block_xfer:
72 andi r4, r7, 0xffffffe0 /* n =
73 rsub r7, r4, r7 /* c =
74
75 andi r9, r6, 3 /* t1
76 /* if temp != 0, unaligned transfers n
77 bnei r9, a_block_unaligned
78
79 a_block_aligned:
80 lwi r9, r6, 0 /* t1
81 lwi r10, r6, 4 /* t2
82 lwi r11, r6, 8 /* t3
83 lwi r12, r6, 12 /* t4
84 swi r9, r5, 0 /* *(d
85 swi r10, r5, 4 /* *(d
86 swi r11, r5, 8 /* *(d
87 swi r12, r5, 12 /* *(d
88 lwi r9, r6, 16 /* t1
89 lwi r10, r6, 20 /* t2
90 lwi r11, r6, 24 /* t3
91 lwi r12, r6, 28 /* t4
92 swi r9, r5, 16 /* *(d
93 swi r10, r5, 20 /* *(d
94 swi r11, r5, 24 /* *(d
95 swi r12, r5, 28 /* *(d
96 addi r6, r6, 32 /* s =
97 addi r4, r4, -32 /* n =
98 bneid r4, a_block_aligned /* whi
99 addi r5, r5, 32 /* d =
100 bri a_block_done
101
102 a_block_unaligned:
103 andi r8, r6, 0xfffffffc /* as
104 add r6, r6, r4 /* s =
105 lwi r11, r8, 0 /* h =
106
107 addi r9, r9, -1
108 beqi r9, a_block_u1 /* t1
109 addi r9, r9, -1
110 beqi r9, a_block_u2 /* t1
111
112 a_block_u3:
113 bslli r11, r11, 24 /* h = h << 24
114 a_bu3_loop:
115 lwi r12, r8, 4 /* v = *(as +
116 bsrli r9, r12, 8 /* t1 = v >> 8
117 or r9, r11, r9 /* t1 = h | t1
118 swi r9, r5, 0 /* *(d + 0) =
119 bslli r11, r12, 24 /* h = v << 24
120 lwi r12, r8, 8 /* v = *(as +
121 bsrli r9, r12, 8 /* t1 = v >> 8
122 or r9, r11, r9 /* t1 = h | t1
123 swi r9, r5, 4 /* *(d + 4) =
124 bslli r11, r12, 24 /* h = v << 24
125 lwi r12, r8, 12 /* v = *(as +
126 bsrli r9, r12, 8 /* t1 = v >> 8
127 or r9, r11, r9 /* t1 = h | t1
128 swi r9, r5, 8 /* *(d + 8) =
129 bslli r11, r12, 24 /* h = v << 24
130 lwi r12, r8, 16 /* v = *(as +
131 bsrli r9, r12, 8 /* t1 = v >> 8
132 or r9, r11, r9 /* t1 = h | t1
133 swi r9, r5, 12 /* *(d + 12) =
134 bslli r11, r12, 24 /* h = v << 24
135 lwi r12, r8, 20 /* v = *(as +
136 bsrli r9, r12, 8 /* t1 = v >> 8
137 or r9, r11, r9 /* t1 = h | t1
138 swi r9, r5, 16 /* *(d + 16) =
139 bslli r11, r12, 24 /* h = v << 24
140 lwi r12, r8, 24 /* v = *(as +
141 bsrli r9, r12, 8 /* t1 = v >> 8
142 or r9, r11, r9 /* t1 = h | t1
143 swi r9, r5, 20 /* *(d + 20) =
144 bslli r11, r12, 24 /* h = v << 24
145 lwi r12, r8, 28 /* v = *(as +
146 bsrli r9, r12, 8 /* t1 = v >> 8
147 or r9, r11, r9 /* t1 = h | t1
148 swi r9, r5, 24 /* *(d + 24) =
149 bslli r11, r12, 24 /* h = v << 24
150 lwi r12, r8, 32 /* v = *(as +
151 bsrli r9, r12, 8 /* t1 = v >> 8
152 or r9, r11, r9 /* t1 = h | t1
153 swi r9, r5, 28 /* *(d + 28) =
154 bslli r11, r12, 24 /* h = v << 24
155 addi r8, r8, 32 /* as = as + 3
156 addi r4, r4, -32 /* n = n - 32
157 bneid r4, a_bu3_loop /* while (n) l
158 addi r5, r5, 32 /* d = d + 32
159 bri a_block_done
160
161 a_block_u1:
162 bslli r11, r11, 8 /* h = h << 8
163 a_bu1_loop:
164 lwi r12, r8, 4 /* v = *(as +
165 bsrli r9, r12, 24 /* t1 = v >> 2
166 or r9, r11, r9 /* t1 = h | t1
167 swi r9, r5, 0 /* *(d + 0) =
168 bslli r11, r12, 8 /* h = v << 8
169 lwi r12, r8, 8 /* v = *(as +
170 bsrli r9, r12, 24 /* t1 = v >> 2
171 or r9, r11, r9 /* t1 = h | t1
172 swi r9, r5, 4 /* *(d + 4) =
173 bslli r11, r12, 8 /* h = v << 8
174 lwi r12, r8, 12 /* v = *(as +
175 bsrli r9, r12, 24 /* t1 = v >> 2
176 or r9, r11, r9 /* t1 = h | t1
177 swi r9, r5, 8 /* *(d + 8) =
178 bslli r11, r12, 8 /* h = v << 8
179 lwi r12, r8, 16 /* v = *(as +
180 bsrli r9, r12, 24 /* t1 = v >> 2
181 or r9, r11, r9 /* t1 = h | t1
182 swi r9, r5, 12 /* *(d + 12) =
183 bslli r11, r12, 8 /* h = v << 8
184 lwi r12, r8, 20 /* v = *(as +
185 bsrli r9, r12, 24 /* t1 = v >> 2
186 or r9, r11, r9 /* t1 = h | t1
187 swi r9, r5, 16 /* *(d + 16) =
188 bslli r11, r12, 8 /* h = v << 8
189 lwi r12, r8, 24 /* v = *(as +
190 bsrli r9, r12, 24 /* t1 = v >> 2
191 or r9, r11, r9 /* t1 = h | t1
192 swi r9, r5, 20 /* *(d + 20) =
193 bslli r11, r12, 8 /* h = v << 8
194 lwi r12, r8, 28 /* v = *(as +
195 bsrli r9, r12, 24 /* t1 = v >> 2
196 or r9, r11, r9 /* t1 = h | t1
197 swi r9, r5, 24 /* *(d + 24) =
198 bslli r11, r12, 8 /* h = v << 8
199 lwi r12, r8, 32 /* v = *(as +
200 bsrli r9, r12, 24 /* t1 = v >> 2
201 or r9, r11, r9 /* t1 = h | t1
202 swi r9, r5, 28 /* *(d + 28) =
203 bslli r11, r12, 8 /* h = v << 8
204 addi r8, r8, 32 /* as = as + 3
205 addi r4, r4, -32 /* n = n - 32
206 bneid r4, a_bu1_loop /* while (n) l
207 addi r5, r5, 32 /* d = d + 32
208 bri a_block_done
209
210 a_block_u2:
211 bslli r11, r11, 16 /* h = h << 16
212 a_bu2_loop:
213 lwi r12, r8, 4 /* v = *(as +
214 bsrli r9, r12, 16 /* t1 = v >> 1
215 or r9, r11, r9 /* t1 = h | t1
216 swi r9, r5, 0 /* *(d + 0) =
217 bslli r11, r12, 16 /* h = v << 16
218 lwi r12, r8, 8 /* v = *(as +
219 bsrli r9, r12, 16 /* t1 = v >> 1
220 or r9, r11, r9 /* t1 = h | t1
221 swi r9, r5, 4 /* *(d + 4) =
222 bslli r11, r12, 16 /* h = v << 16
223 lwi r12, r8, 12 /* v = *(as +
224 bsrli r9, r12, 16 /* t1 = v >> 1
225 or r9, r11, r9 /* t1 = h | t1
226 swi r9, r5, 8 /* *(d + 8) =
227 bslli r11, r12, 16 /* h = v << 16
228 lwi r12, r8, 16 /* v = *(as +
229 bsrli r9, r12, 16 /* t1 = v >> 1
230 or r9, r11, r9 /* t1 = h | t1
231 swi r9, r5, 12 /* *(d + 12) =
232 bslli r11, r12, 16 /* h = v << 16
233 lwi r12, r8, 20 /* v = *(as +
234 bsrli r9, r12, 16 /* t1 = v >> 1
235 or r9, r11, r9 /* t1 = h | t1
236 swi r9, r5, 16 /* *(d + 16) =
237 bslli r11, r12, 16 /* h = v << 16
238 lwi r12, r8, 24 /* v = *(as +
239 bsrli r9, r12, 16 /* t1 = v >> 1
240 or r9, r11, r9 /* t1 = h | t1
241 swi r9, r5, 20 /* *(d + 20) =
242 bslli r11, r12, 16 /* h = v << 16
243 lwi r12, r8, 28 /* v = *(as +
244 bsrli r9, r12, 16 /* t1 = v >> 1
245 or r9, r11, r9 /* t1 = h | t1
246 swi r9, r5, 24 /* *(d + 24) =
247 bslli r11, r12, 16 /* h = v << 16
248 lwi r12, r8, 32 /* v = *(as +
249 bsrli r9, r12, 16 /* t1 = v >> 1
250 or r9, r11, r9 /* t1 = h | t1
251 swi r9, r5, 28 /* *(d + 28) =
252 bslli r11, r12, 16 /* h = v << 16
253 addi r8, r8, 32 /* as = as + 3
254 addi r4, r4, -32 /* n = n - 32
255 bneid r4, a_bu2_loop /* while (n) l
256 addi r5, r5, 32 /* d = d + 32
257
258 a_block_done:
259 addi r4, r0, 4 /* n = 4 */
260 cmpu r4, r4, r7 /* n = c - n
261 blti r4, a_xfer_end /* if n < 0, l
262
263 a_word_xfer:
264 andi r4, r7, 0xfffffffc /* n =
265 addi r10, r0, 0 /* off
266
267 andi r9, r6, 3 /* t1
268 /* if temp != 0, unaligned transfers n
269 bnei r9, a_word_unaligned
270
271 a_word_aligned:
272 lw r9, r6, r10 /* t1
273 sw r9, r5, r10 /* *(d
274 addi r4, r4,-4 /* n--
275 bneid r4, a_word_aligned /* loo
276 addi r10, r10, 4 /* off
277
278 bri a_word_done
279
280 a_word_unaligned:
281 andi r8, r6, 0xfffffffc /* as
282 lwi r11, r8, 0 /* h =
283 addi r8, r8, 4 /* as
284
285 addi r9, r9, -1
286 beqi r9, a_word_u1 /* t1
287 addi r9, r9, -1
288 beqi r9, a_word_u2 /* t1
289
290 a_word_u3:
291 bslli r11, r11, 24 /* h = h << 24
292 a_wu3_loop:
293 lw r12, r8, r10 /* v = *(as +
294 bsrli r9, r12, 8 /* t1 = v >> 8
295 or r9, r11, r9 /* t1 = h | t1
296 sw r9, r5, r10 /* *(d + offse
297 bslli r11, r12, 24 /* h = v << 24
298 addi r4, r4,-4 /* n = n - 4 *
299 bneid r4, a_wu3_loop /* while (n) l
300 addi r10, r10, 4 /* offset = of
301
302 bri a_word_done
303
304 a_word_u1:
305 bslli r11, r11, 8 /* h = h << 8
306 a_wu1_loop:
307 lw r12, r8, r10 /* v = *(as +
308 bsrli r9, r12, 24 /* t1 = v >> 2
309 or r9, r11, r9 /* t1 = h | t1
310 sw r9, r5, r10 /* *(d + offse
311 bslli r11, r12, 8 /* h = v << 8
312 addi r4, r4,-4 /* n = n - 4 *
313 bneid r4, a_wu1_loop /* while (n) l
314 addi r10, r10, 4 /* offset = of
315
316 bri a_word_done
317
318 a_word_u2:
319 bslli r11, r11, 16 /* h = h << 16
320 a_wu2_loop:
321 lw r12, r8, r10 /* v = *(as +
322 bsrli r9, r12, 16 /* t1 = v >> 1
323 or r9, r11, r9 /* t1 = h | t1
324 sw r9, r5, r10 /* *(d + offse
325 bslli r11, r12, 16 /* h = v << 16
326 addi r4, r4,-4 /* n = n - 4 *
327 bneid r4, a_wu2_loop /* while (n) l
328 addi r10, r10, 4 /* offset = of
329
330 a_word_done:
331 add r5, r5, r10 /* d = d + off
332 add r6, r6, r10 /* s = s + off
333 rsub r7, r10, r7 /* c = c - off
334
335 a_xfer_end:
336 a_xfer_end_loop:
337 beqi r7, a_done /* whi
338 lbui r9, r6, 0 /* t1
339 addi r6, r6, 1 /* s++
340 sbi r9, r5, 0 /* *d
341 addi r7, r7, -1 /* c--
342 brid a_xfer_end_loop /* loo
343 addi r5, r5, 1 /* d++
344
345 a_done:
346 rtsd r15, 8
347 nop
348
349 .size memcpy, . - memcpy
350 .end memcpy
351 /*--------------------------------------------
352 .globl memmove
353 .type memmove, @function
354 .ent memmove
355
356 memmove:
357 cmpu r4, r5, r6 /* n = s - d *
358 bgei r4,fast_memcpy_ascending
359
360 fast_memcpy_descending:
361 /* move d to return register as value
362 addi r3, r5, 0
363
364 add r5, r5, r7 /* d = d + c *
365 add r6, r6, r7 /* s = s + c *
366
367 addi r4, r0, 4 /* n = 4 */
368 cmpu r4, r4, r7 /* n = c - n
369 blti r4,d_xfer_end /* if n < 0, l
370
371 /* transfer first 0~3 bytes to get ali
372 andi r4, r5, 3 /* n =
373 /* if zero, destination already aligne
374 beqi r4,d_dalign_done
375 rsub r7, r4, r7 /* c =
376
377 d_xfer_first_loop:
378 /* if no bytes left to transfer, trans
379 beqi r4,d_dalign_done
380 addi r6, r6, -1 /* s--
381 addi r5, r5, -1 /* d--
382 lbui r11, r6, 0 /* h =
383 sbi r11, r5, 0 /* *d
384 brid d_xfer_first_loop /* loo
385 addi r4, r4, -1 /* n--
386
387 d_dalign_done:
388 addi r4, r0, 32 /* n = 32 */
389 cmpu r4, r4, r7 /* n = c - n
390 /* if n < 0, less than one block to tr
391 blti r4, d_block_done
392
393 d_block_xfer:
394 andi r4, r7, 0xffffffe0 /* n =
395 rsub r7, r4, r7 /* c =
396
397 andi r9, r6, 3 /* t1
398 /* if temp != 0, unaligned transfers n
399 bnei r9, d_block_unaligned
400
401 d_block_aligned:
402 addi r6, r6, -32 /* s =
403 addi r5, r5, -32 /* d =
404 lwi r9, r6, 28 /* t1
405 lwi r10, r6, 24 /* t2
406 lwi r11, r6, 20 /* t3
407 lwi r12, r6, 16 /* t4
408 swi r9, r5, 28 /* *(d
409 swi r10, r5, 24 /* *(d
410 swi r11, r5, 20 /* *(d
411 swi r12, r5, 16 /* *(d
412 lwi r9, r6, 12 /* t1
413 lwi r10, r6, 8 /* t2
414 lwi r11, r6, 4 /* t3
415 lwi r12, r6, 0 /* t4
416 swi r9, r5, 12 /* *(d
417 swi r10, r5, 8 /* *(d
418 swi r11, r5, 4 /* *(d
419 addi r4, r4, -32 /* n =
420 bneid r4, d_block_aligned /* whi
421 swi r12, r5, 0 /* *(d
422 bri d_block_done
423
424 d_block_unaligned:
425 andi r8, r6, 0xfffffffc /* as
426 rsub r6, r4, r6 /* s =
427 lwi r11, r8, 0 /* h =
428
429 addi r9, r9, -1
430 beqi r9,d_block_u1 /* t1
431 addi r9, r9, -1
432 beqi r9,d_block_u2 /* t1
433
434 d_block_u3:
435 bsrli r11, r11, 8 /* h = h >> 8
436 d_bu3_loop:
437 addi r8, r8, -32 /* as = as - 3
438 addi r5, r5, -32 /* d = d - 32
439 lwi r12, r8, 28 /* v = *(as +
440 bslli r9, r12, 24 /* t1 = v << 2
441 or r9, r11, r9 /* t1 = h | t1
442 swi r9, r5, 28 /* *(d + 28) =
443 bsrli r11, r12, 8 /* h = v >> 8
444 lwi r12, r8, 24 /* v = *(as +
445 bslli r9, r12, 24 /* t1 = v << 2
446 or r9, r11, r9 /* t1 = h | t1
447 swi r9, r5, 24 /* *(d + 24) =
448 bsrli r11, r12, 8 /* h = v >> 8
449 lwi r12, r8, 20 /* v = *(as +
450 bslli r9, r12, 24 /* t1 = v << 2
451 or r9, r11, r9 /* t1 = h | t1
452 swi r9, r5, 20 /* *(d + 20) =
453 bsrli r11, r12, 8 /* h = v >> 8
454 lwi r12, r8, 16 /* v = *(as +
455 bslli r9, r12, 24 /* t1 = v << 2
456 or r9, r11, r9 /* t1 = h | t1
457 swi r9, r5, 16 /* *(d + 16) =
458 bsrli r11, r12, 8 /* h = v >> 8
459 lwi r12, r8, 12 /* v = *(as +
460 bslli r9, r12, 24 /* t1 = v << 2
461 or r9, r11, r9 /* t1 = h | t1
462 swi r9, r5, 12 /* *(d + 112)
463 bsrli r11, r12, 8 /* h = v >> 8
464 lwi r12, r8, 8 /* v = *(as +
465 bslli r9, r12, 24 /* t1 = v << 2
466 or r9, r11, r9 /* t1 = h | t1
467 swi r9, r5, 8 /* *(d + 8) =
468 bsrli r11, r12, 8 /* h = v >> 8
469 lwi r12, r8, 4 /* v = *(as +
470 bslli r9, r12, 24 /* t1 = v << 2
471 or r9, r11, r9 /* t1 = h | t1
472 swi r9, r5, 4 /* *(d + 4) =
473 bsrli r11, r12, 8 /* h = v >> 8
474 lwi r12, r8, 0 /* v = *(as +
475 bslli r9, r12, 24 /* t1 = v << 2
476 or r9, r11, r9 /* t1 = h | t1
477 swi r9, r5, 0 /* *(d + 0) =
478 addi r4, r4, -32 /* n = n - 32
479 bneid r4, d_bu3_loop /* while (n) l
480 bsrli r11, r12, 8 /* h = v >> 8
481 bri d_block_done
482
483 d_block_u1:
484 bsrli r11, r11, 24 /* h = h >> 24
485 d_bu1_loop:
486 addi r8, r8, -32 /* as = as - 3
487 addi r5, r5, -32 /* d = d - 32
488 lwi r12, r8, 28 /* v = *(as +
489 bslli r9, r12, 8 /* t1 = v << 8
490 or r9, r11, r9 /* t1 = h | t1
491 swi r9, r5, 28 /* *(d + 28) =
492 bsrli r11, r12, 24 /* h = v >> 24
493 lwi r12, r8, 24 /* v = *(as +
494 bslli r9, r12, 8 /* t1 = v << 8
495 or r9, r11, r9 /* t1 = h | t1
496 swi r9, r5, 24 /* *(d + 24) =
497 bsrli r11, r12, 24 /* h = v >> 24
498 lwi r12, r8, 20 /* v = *(as +
499 bslli r9, r12, 8 /* t1 = v << 8
500 or r9, r11, r9 /* t1 = h | t1
501 swi r9, r5, 20 /* *(d + 20) =
502 bsrli r11, r12, 24 /* h = v >> 24
503 lwi r12, r8, 16 /* v = *(as +
504 bslli r9, r12, 8 /* t1 = v << 8
505 or r9, r11, r9 /* t1 = h | t1
506 swi r9, r5, 16 /* *(d + 16) =
507 bsrli r11, r12, 24 /* h = v >> 24
508 lwi r12, r8, 12 /* v = *(as +
509 bslli r9, r12, 8 /* t1 = v << 8
510 or r9, r11, r9 /* t1 = h | t1
511 swi r9, r5, 12 /* *(d + 112)
512 bsrli r11, r12, 24 /* h = v >> 24
513 lwi r12, r8, 8 /* v = *(as +
514 bslli r9, r12, 8 /* t1 = v << 8
515 or r9, r11, r9 /* t1 = h | t1
516 swi r9, r5, 8 /* *(d + 8) =
517 bsrli r11, r12, 24 /* h = v >> 24
518 lwi r12, r8, 4 /* v = *(as +
519 bslli r9, r12, 8 /* t1 = v << 8
520 or r9, r11, r9 /* t1 = h | t1
521 swi r9, r5, 4 /* *(d + 4) =
522 bsrli r11, r12, 24 /* h = v >> 24
523 lwi r12, r8, 0 /* v = *(as +
524 bslli r9, r12, 8 /* t1 = v << 8
525 or r9, r11, r9 /* t1 = h | t1
526 swi r9, r5, 0 /* *(d + 0) =
527 addi r4, r4, -32 /* n = n - 32
528 bneid r4, d_bu1_loop /* while (n) l
529 bsrli r11, r12, 24 /* h = v >> 24
530 bri d_block_done
531
532 d_block_u2:
533 bsrli r11, r11, 16 /* h = h >> 16
534 d_bu2_loop:
535 addi r8, r8, -32 /* as = as - 3
536 addi r5, r5, -32 /* d = d - 32
537 lwi r12, r8, 28 /* v = *(as +
538 bslli r9, r12, 16 /* t1 = v << 1
539 or r9, r11, r9 /* t1 = h | t1
540 swi r9, r5, 28 /* *(d + 28) =
541 bsrli r11, r12, 16 /* h = v >> 16
542 lwi r12, r8, 24 /* v = *(as +
543 bslli r9, r12, 16 /* t1 = v << 1
544 or r9, r11, r9 /* t1 = h | t1
545 swi r9, r5, 24 /* *(d + 24) =
546 bsrli r11, r12, 16 /* h = v >> 16
547 lwi r12, r8, 20 /* v = *(as +
548 bslli r9, r12, 16 /* t1 = v << 1
549 or r9, r11, r9 /* t1 = h | t1
550 swi r9, r5, 20 /* *(d + 20) =
551 bsrli r11, r12, 16 /* h = v >> 16
552 lwi r12, r8, 16 /* v = *(as +
553 bslli r9, r12, 16 /* t1 = v << 1
554 or r9, r11, r9 /* t1 = h | t1
555 swi r9, r5, 16 /* *(d + 16) =
556 bsrli r11, r12, 16 /* h = v >> 16
557 lwi r12, r8, 12 /* v = *(as +
558 bslli r9, r12, 16 /* t1 = v << 1
559 or r9, r11, r9 /* t1 = h | t1
560 swi r9, r5, 12 /* *(d + 112)
561 bsrli r11, r12, 16 /* h = v >> 16
562 lwi r12, r8, 8 /* v = *(as +
563 bslli r9, r12, 16 /* t1 = v << 1
564 or r9, r11, r9 /* t1 = h | t1
565 swi r9, r5, 8 /* *(d + 8) =
566 bsrli r11, r12, 16 /* h = v >> 16
567 lwi r12, r8, 4 /* v = *(as +
568 bslli r9, r12, 16 /* t1 = v << 1
569 or r9, r11, r9 /* t1 = h | t1
570 swi r9, r5, 4 /* *(d + 4) =
571 bsrli r11, r12, 16 /* h = v >> 16
572 lwi r12, r8, 0 /* v = *(as +
573 bslli r9, r12, 16 /* t1 = v << 1
574 or r9, r11, r9 /* t1 = h | t1
575 swi r9, r5, 0 /* *(d + 0) =
576 addi r4, r4, -32 /* n = n - 32
577 bneid r4, d_bu2_loop /* while (n) l
578 bsrli r11, r12, 16 /* h = v >> 16
579
580 d_block_done:
581 addi r4, r0, 4 /* n = 4 */
582 cmpu r4, r4, r7 /* n = c - n
583 blti r4,d_xfer_end /* if n < 0, l
584
585 d_word_xfer:
586 andi r4, r7, 0xfffffffc /* n =
587 rsub r5, r4, r5 /* d =
588 rsub r6, r4, r6 /* s =
589 rsub r7, r4, r7 /* c =
590
591 andi r9, r6, 3 /* t1
592 /* if temp != 0, unaligned transfers n
593 bnei r9, d_word_unaligned
594
595 d_word_aligned:
596 addi r4, r4,-4 /* n--
597 lw r9, r6, r4 /* t1
598 bneid r4, d_word_aligned /* loo
599 sw r9, r5, r4 /* *(d
600
601 bri d_word_done
602
603 d_word_unaligned:
604 andi r8, r6, 0xfffffffc /* as
605 lw r11, r8, r4 /* h =
606
607 addi r9, r9, -1
608 beqi r9,d_word_u1 /* t1
609 addi r9, r9, -1
610 beqi r9,d_word_u2 /* t1
611
612 d_word_u3:
613 bsrli r11, r11, 8 /* h = h >> 8
614 d_wu3_loop:
615 addi r4, r4,-4 /* n = n - 4 *
616 lw r12, r8, r4 /* v = *(as +
617 bslli r9, r12, 24 /* t1 = v << 2
618 or r9, r11, r9 /* t1 = h | t1
619 sw r9, r5, r4 /* *(d + n) =
620 bneid r4, d_wu3_loop /* while (n) l
621 bsrli r11, r12, 8 /* h = v >> 8
622
623 bri d_word_done
624
625 d_word_u1:
626 bsrli r11, r11, 24 /* h = h >> 24
627 d_wu1_loop:
628 addi r4, r4,-4 /* n = n - 4 *
629 lw r12, r8, r4 /* v = *(as +
630 bslli r9, r12, 8 /* t1 = v << 8
631 or r9, r11, r9 /* t1 = h | t1
632 sw r9, r5, r4 /* *(d + n) =
633 bneid r4, d_wu1_loop /* while (n) l
634 bsrli r11, r12, 24 /* h = v >> 24
635
636 bri d_word_done
637
638 d_word_u2:
639 bsrli r11, r11, 16 /* h = h >> 16
640 d_wu2_loop:
641 addi r4, r4,-4 /* n = n - 4 *
642 lw r12, r8, r4 /* v = *(as +
643 bslli r9, r12, 16 /* t1 = v << 1
644 or r9, r11, r9 /* t1 = h | t1
645 sw r9, r5, r4 /* *(d + n) =
646 bneid r4, d_wu2_loop /* while (n) l
647 bsrli r11, r12, 16 /* h = v >> 16
648
649 d_word_done:
650
651 d_xfer_end:
652 d_xfer_end_loop:
653 beqi r7, a_done /* whi
654 addi r6, r6, -1 /* s--
655 lbui r9, r6, 0 /* t1
656 addi r5, r5, -1 /* d--
657 sbi r9, r5, 0 /* *d
658 brid d_xfer_end_loop /* loo
659 addi r7, r7, -1 /* c--
660
661 d_done:
662 rtsd r15, 8
663 nop
664
665 .size memmove, . - memmove
666 .end memmove
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.