1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
|
/*
* Copyright (c) 2008 Mans Rullgard <mans@mansr.com>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "asm.S"
.fpu neon
/* chroma_mc8(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y) */
.macro h264_chroma_mc8 avg=0
push {r4-r7, lr}
ldrd r4, [sp, #20]
.if \avg
mov lr, r0
.endif
pld [r1]
pld [r1, r2]
muls r7, r4, r5
rsb r6, r7, r5, lsl #3
rsb ip, r7, r4, lsl #3
sub r4, r7, r4, lsl #3
sub r4, r4, r5, lsl #3
add r4, r4, #64
beq 2f
add r5, r1, r2
vdup.8 d0, r4
lsl r4, r2, #1
vdup.8 d1, ip
vld1.64 {d4, d5}, [r1], r4
vdup.8 d2, r6
vld1.64 {d6, d7}, [r5], r4
vdup.8 d3, r7
vext.8 d5, d4, d5, #1
vext.8 d7, d6, d7, #1
1: pld [r5]
vmull.u8 q8, d4, d0
vmlal.u8 q8, d5, d1
vld1.64 {d4, d5}, [r1], r4
vmlal.u8 q8, d6, d2
vext.8 d5, d4, d5, #1
vmlal.u8 q8, d7, d3
vmull.u8 q9, d6, d0
subs r3, r3, #2
vmlal.u8 q9, d7, d1
vmlal.u8 q9, d4, d2
vmlal.u8 q9, d5, d3
vrshrn.u16 d16, q8, #6
vld1.64 {d6, d7}, [r5], r4
pld [r1]
vrshrn.u16 d17, q9, #6
.if \avg
vld1.64 {d20}, [lr,:64], r2
vld1.64 {d21}, [lr,:64], r2
vrhadd.u8 q8, q8, q10
.endif
vext.8 d7, d6, d7, #1
vst1.64 {d16}, [r0,:64], r2
vst1.64 {d17}, [r0,:64], r2
bgt 1b
pop {r4-r7, pc}
2: tst r6, r6
add ip, ip, r6
vdup.8 d0, r4
vdup.8 d1, ip
beq 4f
add r5, r1, r2
lsl r4, r2, #1
vld1.64 {d4}, [r1], r4
vld1.64 {d6}, [r5], r4
3: pld [r5]
vmull.u8 q8, d4, d0
vmlal.u8 q8, d6, d1
vld1.64 {d4}, [r1], r4
vmull.u8 q9, d6, d0
vmlal.u8 q9, d4, d1
vld1.64 {d6}, [r5], r4
vrshrn.u16 d16, q8, #6
vrshrn.u16 d17, q9, #6
.if \avg
vld1.64 {d20}, [lr,:64], r2
vld1.64 {d21}, [lr,:64], r2
vrhadd.u8 q8, q8, q10
.endif
subs r3, r3, #2
pld [r1]
vst1.64 {d16}, [r0,:64], r2
vst1.64 {d17}, [r0,:64], r2
bgt 3b
pop {r4-r7, pc}
4: vld1.64 {d4, d5}, [r1], r2
vld1.64 {d6, d7}, [r1], r2
vext.8 d5, d4, d5, #1
vext.8 d7, d6, d7, #1
5: pld [r1]
subs r3, r3, #2
vmull.u8 q8, d4, d0
vmlal.u8 q8, d5, d1
vld1.64 {d4, d5}, [r1], r2
vmull.u8 q9, d6, d0
vmlal.u8 q9, d7, d1
pld [r1]
vext.8 d5, d4, d5, #1
vrshrn.u16 d16, q8, #6
vrshrn.u16 d17, q9, #6
.if \avg
vld1.64 {d20}, [lr,:64], r2
vld1.64 {d21}, [lr,:64], r2
vrhadd.u8 q8, q8, q10
.endif
vld1.64 {d6, d7}, [r1], r2
vext.8 d7, d6, d7, #1
vst1.64 {d16}, [r0,:64], r2
vst1.64 {d17}, [r0,:64], r2
bgt 5b
pop {r4-r7, pc}
.endm
/* chroma_mc4(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y) */
.macro h264_chroma_mc4 avg=0
push {r4-r7, lr}
ldrd r4, [sp, #20]
.if \avg
mov lr, r0
.endif
pld [r1]
pld [r1, r2]
muls r7, r4, r5
rsb r6, r7, r5, lsl #3
rsb ip, r7, r4, lsl #3
sub r4, r7, r4, lsl #3
sub r4, r4, r5, lsl #3
add r4, r4, #64
beq 2f
add r5, r1, r2
vdup.8 d0, r4
lsl r4, r2, #1
vdup.8 d1, ip
vld1.64 {d4}, [r1], r4
vdup.8 d2, r6
vld1.64 {d6}, [r5], r4
vdup.8 d3, r7
vext.8 d5, d4, d5, #1
vext.8 d7, d6, d7, #1
vtrn.32 d4, d5
vtrn.32 d6, d7
vtrn.32 d0, d1
vtrn.32 d2, d3
1: pld [r5]
vmull.u8 q8, d4, d0
vmlal.u8 q8, d6, d2
vld1.64 {d4}, [r1], r4
vext.8 d5, d4, d5, #1
vtrn.32 d4, d5
vmull.u8 q9, d6, d0
vmlal.u8 q9, d4, d2
vld1.64 {d6}, [r5], r4
vadd.i16 d16, d16, d17
vadd.i16 d17, d18, d19
vrshrn.u16 d16, q8, #6
subs r3, r3, #2
pld [r1]
.if \avg
vld1.32 {d20[0]}, [lr,:32], r2
vld1.32 {d20[1]}, [lr,:32], r2
vrhadd.u8 d16, d16, d20
.endif
vext.8 d7, d6, d7, #1
vtrn.32 d6, d7
vst1.32 {d16[0]}, [r0,:32], r2
vst1.32 {d16[1]}, [r0,:32], r2
bgt 1b
pop {r4-r7, pc}
2: tst r6, r6
add ip, ip, r6
vdup.8 d0, r4
vdup.8 d1, ip
vtrn.32 d0, d1
beq 4f
vext.32 d1, d0, d1, #1
add r5, r1, r2
lsl r4, r2, #1
vld1.32 {d4[0]}, [r1], r4
vld1.32 {d4[1]}, [r5], r4
3: pld [r5]
vmull.u8 q8, d4, d0
vld1.32 {d4[0]}, [r1], r4
vmull.u8 q9, d4, d1
vld1.32 {d4[1]}, [r5], r4
vadd.i16 d16, d16, d17
vadd.i16 d17, d18, d19
vrshrn.u16 d16, q8, #6
.if \avg
vld1.32 {d20[0]}, [lr,:32], r2
vld1.32 {d20[1]}, [lr,:32], r2
vrhadd.u8 d16, d16, d20
.endif
subs r3, r3, #2
pld [r1]
vst1.32 {d16[0]}, [r0,:32], r2
vst1.32 {d16[1]}, [r0,:32], r2
bgt 3b
pop {r4-r7, pc}
4: vld1.64 {d4}, [r1], r2
vld1.64 {d6}, [r1], r2
vext.8 d5, d4, d5, #1
vext.8 d7, d6, d7, #1
vtrn.32 d4, d5
vtrn.32 d6, d7
5: vmull.u8 q8, d4, d0
vmull.u8 q9, d6, d0
subs r3, r3, #2
vld1.64 {d4}, [r1], r2
vext.8 d5, d4, d5, #1
vtrn.32 d4, d5
vadd.i16 d16, d16, d17
vadd.i16 d17, d18, d19
pld [r1]
vrshrn.u16 d16, q8, #6
.if \avg
vld1.32 {d20[0]}, [lr,:32], r2
vld1.32 {d20[1]}, [lr,:32], r2
vrhadd.u8 d16, d16, d20
.endif
vld1.64 {d6}, [r1], r2
vext.8 d7, d6, d7, #1
vtrn.32 d6, d7
pld [r1]
vst1.32 {d16[0]}, [r0,:32], r2
vst1.32 {d16[1]}, [r0,:32], r2
bgt 5b
pop {r4-r7, pc}
.endm
.text
.align
function ff_put_h264_chroma_mc8_neon, export=1
h264_chroma_mc8
.endfunc
function ff_avg_h264_chroma_mc8_neon, export=1
h264_chroma_mc8 avg=1
.endfunc
function ff_put_h264_chroma_mc4_neon, export=1
h264_chroma_mc4
.endfunc
function ff_avg_h264_chroma_mc4_neon, export=1
h264_chroma_mc4 avg=1
.endfunc
/* H.264 loop filter */
.macro h264_loop_filter_start
ldr ip, [sp]
tst r2, r2
ldr ip, [ip]
tstne r3, r3
vmov.32 d24[0], ip
and ip, ip, ip, lsl #16
bxeq lr
ands ip, ip, ip, lsl #8
bxlt lr
.endm
.macro align_push_regs
and ip, sp, #15
add ip, ip, #32
sub sp, sp, ip
vst1.64 {d12-d15}, [sp,:128]
sub sp, sp, #32
vst1.64 {d8-d11}, [sp,:128]
.endm
.macro align_pop_regs
vld1.64 {d8-d11}, [sp,:128]!
vld1.64 {d12-d15}, [sp,:128], ip
.endm
.macro h264_loop_filter_luma
vdup.8 q11, r2 @ alpha
vmovl.u8 q12, d24
vabd.u8 q6, q8, q0 @ abs(p0 - q0)
vmovl.u16 q12, d24
vabd.u8 q14, q9, q8 @ abs(p1 - p0)
vsli.16 q12, q12, #8
vabd.u8 q15, q1, q0 @ abs(q1 - q0)
vsli.32 q12, q12, #16
vclt.u8 q6, q6, q11 @ < alpha
vdup.8 q11, r3 @ beta
vclt.s8 q7, q12, #0
vclt.u8 q14, q14, q11 @ < beta
vclt.u8 q15, q15, q11 @ < beta
vbic q6, q6, q7
vabd.u8 q4, q10, q8 @ abs(p2 - p0)
vand q6, q6, q14
vabd.u8 q5, q2, q0 @ abs(q2 - q0)
vclt.u8 q4, q4, q11 @ < beta
vand q6, q6, q15
vclt.u8 q5, q5, q11 @ < beta
vand q4, q4, q6
vand q5, q5, q6
vand q12, q12, q6
vrhadd.u8 q14, q8, q0
vsub.i8 q6, q12, q4
vqadd.u8 q7, q9, q12
vhadd.u8 q10, q10, q14
vsub.i8 q6, q6, q5
vhadd.u8 q14, q2, q14
vmin.u8 q7, q7, q10
vqsub.u8 q11, q9, q12
vqadd.u8 q2, q1, q12
vmax.u8 q7, q7, q11
vqsub.u8 q11, q1, q12
vmin.u8 q14, q2, q14
vmovl.u8 q2, d0
vmax.u8 q14, q14, q11
vmovl.u8 q10, d1
vsubw.u8 q2, q2, d16
vsubw.u8 q10, q10, d17
vshl.i16 q2, q2, #2
vshl.i16 q10, q10, #2
vaddw.u8 q2, q2, d18
vaddw.u8 q10, q10, d19
vsubw.u8 q2, q2, d2
vsubw.u8 q10, q10, d3
vrshrn.i16 d4, q2, #3
vrshrn.i16 d5, q10, #3
vbsl q4, q7, q9
vbsl q5, q14, q1
vneg.s8 q7, q6
vmovl.u8 q14, d16
vmin.s8 q2, q2, q6
vmovl.u8 q6, d17
vmax.s8 q2, q2, q7
vmovl.u8 q11, d0
vmovl.u8 q12, d1
vaddw.s8 q14, q14, d4
vaddw.s8 q6, q6, d5
vsubw.s8 q11, q11, d4
vsubw.s8 q12, q12, d5
vqmovun.s16 d16, q14
vqmovun.s16 d17, q6
vqmovun.s16 d0, q11
vqmovun.s16 d1, q12
.endm
function ff_h264_v_loop_filter_luma_neon, export=1
h264_loop_filter_start
vld1.64 {d0, d1}, [r0,:128], r1
vld1.64 {d2, d3}, [r0,:128], r1
vld1.64 {d4, d5}, [r0,:128], r1
sub r0, r0, r1, lsl #2
sub r0, r0, r1, lsl #1
vld1.64 {d20,d21}, [r0,:128], r1
vld1.64 {d18,d19}, [r0,:128], r1
vld1.64 {d16,d17}, [r0,:128], r1
align_push_regs
h264_loop_filter_luma
sub r0, r0, r1, lsl #1
vst1.64 {d8, d9}, [r0,:128], r1
vst1.64 {d16,d17}, [r0,:128], r1
vst1.64 {d0, d1}, [r0,:128], r1
vst1.64 {d10,d11}, [r0,:128]
align_pop_regs
bx lr
.endfunc
function ff_h264_h_loop_filter_luma_neon, export=1
h264_loop_filter_start
sub r0, r0, #4
vld1.64 {d6}, [r0], r1
vld1.64 {d20}, [r0], r1
vld1.64 {d18}, [r0], r1
vld1.64 {d16}, [r0], r1
vld1.64 {d0}, [r0], r1
vld1.64 {d2}, [r0], r1
vld1.64 {d4}, [r0], r1
vld1.64 {d26}, [r0], r1
vld1.64 {d7}, [r0], r1
vld1.64 {d21}, [r0], r1
vld1.64 {d19}, [r0], r1
vld1.64 {d17}, [r0], r1
vld1.64 {d1}, [r0], r1
vld1.64 {d3}, [r0], r1
vld1.64 {d5}, [r0], r1
vld1.64 {d27}, [r0], r1
vtrn.32 q3, q0
vtrn.32 q10, q1
vtrn.32 q9, q2
vtrn.32 q8, q13
vtrn.16 q3, q9
vtrn.16 q10, q8
vtrn.16 q0, q2
vtrn.16 q1, q13
vtrn.8 q3, q10
vtrn.8 q9, q8
vtrn.8 q0, q1
vtrn.8 q2, q13
align_push_regs
sub sp, sp, #16
vst1.64 {d4, d5}, [sp,:128]
sub sp, sp, #16
vst1.64 {d20,d21}, [sp,:128]
h264_loop_filter_luma
vld1.64 {d20,d21}, [sp,:128]!
vld1.64 {d4, d5}, [sp,:128]!
vtrn.32 q3, q0
vtrn.32 q10, q5
vtrn.32 q4, q2
vtrn.32 q8, q13
vtrn.16 q3, q4
vtrn.16 q10, q8
vtrn.16 q0, q2
vtrn.16 q5, q13
vtrn.8 q3, q10
vtrn.8 q4, q8
vtrn.8 q0, q5
vtrn.8 q2, q13
sub r0, r0, r1, lsl #4
vst1.64 {d6}, [r0], r1
vst1.64 {d20}, [r0], r1
vst1.64 {d8}, [r0], r1
vst1.64 {d16}, [r0], r1
vst1.64 {d0}, [r0], r1
vst1.64 {d10}, [r0], r1
vst1.64 {d4}, [r0], r1
vst1.64 {d26}, [r0], r1
vst1.64 {d7}, [r0], r1
vst1.64 {d21}, [r0], r1
vst1.64 {d9}, [r0], r1
vst1.64 {d17}, [r0], r1
vst1.64 {d1}, [r0], r1
vst1.64 {d11}, [r0], r1
vst1.64 {d5}, [r0], r1
vst1.64 {d27}, [r0], r1
align_pop_regs
bx lr
.endfunc
.macro h264_loop_filter_chroma
vdup.8 d22, r2 @ alpha
vmovl.u8 q12, d24
vabd.u8 d26, d16, d0 @ abs(p0 - q0)
vmovl.u8 q2, d0
vabd.u8 d28, d18, d16 @ abs(p1 - p0)
vsubw.u8 q2, q2, d16
vsli.16 d24, d24, #8
vshl.i16 q2, q2, #2
vabd.u8 d30, d2, d0 @ abs(q1 - q0)
vaddw.u8 q2, q2, d18
vclt.u8 d26, d26, d22 @ < alpha
vsubw.u8 q2, q2, d2
vdup.8 d22, r3 @ beta
vclt.s8 d25, d24, #0
vrshrn.i16 d4, q2, #3
vclt.u8 d28, d28, d22 @ < beta
vbic d26, d26, d25
vclt.u8 d30, d30, d22 @ < beta
vand d26, d26, d28
vneg.s8 d25, d24
vand d26, d26, d30
vmin.s8 d4, d4, d24
vmovl.u8 q14, d16
vand d4, d4, d26
vmax.s8 d4, d4, d25
vmovl.u8 q11, d0
vaddw.s8 q14, q14, d4
vsubw.s8 q11, q11, d4
vqmovun.s16 d16, q14
vqmovun.s16 d0, q11
.endm
function ff_h264_v_loop_filter_chroma_neon, export=1
h264_loop_filter_start
sub r0, r0, r1, lsl #1
vld1.64 {d18}, [r0,:64], r1
vld1.64 {d16}, [r0,:64], r1
vld1.64 {d0}, [r0,:64], r1
vld1.64 {d2}, [r0,:64]
h264_loop_filter_chroma
sub r0, r0, r1, lsl #1
vst1.64 {d16}, [r0,:64], r1
vst1.64 {d0}, [r0,:64], r1
bx lr
.endfunc
function ff_h264_h_loop_filter_chroma_neon, export=1
h264_loop_filter_start
sub r0, r0, #2
vld1.32 {d18[0]}, [r0], r1
vld1.32 {d16[0]}, [r0], r1
vld1.32 {d0[0]}, [r0], r1
vld1.32 {d2[0]}, [r0], r1
vld1.32 {d18[1]}, [r0], r1
vld1.32 {d16[1]}, [r0], r1
vld1.32 {d0[1]}, [r0], r1
vld1.32 {d2[1]}, [r0], r1
vtrn.16 d18, d0
vtrn.16 d16, d2
vtrn.8 d18, d16
vtrn.8 d0, d2
h264_loop_filter_chroma
vtrn.16 d18, d0
vtrn.16 d16, d2
vtrn.8 d18, d16
vtrn.8 d0, d2
sub r0, r0, r1, lsl #3
vst1.32 {d18[0]}, [r0], r1
vst1.32 {d16[0]}, [r0], r1
vst1.32 {d0[0]}, [r0], r1
vst1.32 {d2[0]}, [r0], r1
vst1.32 {d18[1]}, [r0], r1
vst1.32 {d16[1]}, [r0], r1
vst1.32 {d0[1]}, [r0], r1
vst1.32 {d2[1]}, [r0], r1
bx lr
.endfunc
|
