summaryrefslogtreecommitdiffstats
path: root/media/ffvpx/libavcodec/x86/vp8dsp_init.c
blob: 897d5a0e77c88fd71dc95141ae68b551e37eec57 (plain)
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
/*
 * VP8 DSP functions x86-optimized
 * Copyright (c) 2010 Ronald S. Bultje <rsbultje@gmail.com>
 * Copyright (c) 2010 Fiona Glaser <fiona@x264.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 "libavutil/attributes.h"
#include "libavutil/cpu.h"
#include "libavutil/mem.h"
#include "libavutil/x86/cpu.h"
#include "libavcodec/vp8dsp.h"

#if HAVE_YASM

/*
 * MC functions
 */
void ff_put_vp8_epel4_h4_mmxext(uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel4_h6_mmxext(uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel4_v4_mmxext(uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel4_v6_mmxext(uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);

void ff_put_vp8_epel8_h4_sse2  (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel8_h6_sse2  (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel8_v4_sse2  (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel8_v6_sse2  (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);

void ff_put_vp8_epel4_h4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel4_h6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel4_v4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel4_v6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel8_h4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel8_h6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel8_v4_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);
void ff_put_vp8_epel8_v6_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                uint8_t *src, ptrdiff_t srcstride,
                                int height, int mx, int my);

void ff_put_vp8_bilinear4_h_mmxext(uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);
void ff_put_vp8_bilinear8_h_sse2  (uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);
void ff_put_vp8_bilinear4_h_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);
void ff_put_vp8_bilinear8_h_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);

void ff_put_vp8_bilinear4_v_mmxext(uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);
void ff_put_vp8_bilinear8_v_sse2  (uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);
void ff_put_vp8_bilinear4_v_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);
void ff_put_vp8_bilinear8_v_ssse3 (uint8_t *dst, ptrdiff_t dststride,
                                   uint8_t *src, ptrdiff_t srcstride,
                                   int height, int mx, int my);


void ff_put_vp8_pixels8_mmx (uint8_t *dst, ptrdiff_t dststride,
                             uint8_t *src, ptrdiff_t srcstride,
                             int height, int mx, int my);
void ff_put_vp8_pixels16_mmx(uint8_t *dst, ptrdiff_t dststride,
                             uint8_t *src, ptrdiff_t srcstride,
                             int height, int mx, int my);
void ff_put_vp8_pixels16_sse(uint8_t *dst, ptrdiff_t dststride,
                             uint8_t *src, ptrdiff_t srcstride,
                             int height, int mx, int my);

#define TAP_W16(OPT, FILTERTYPE, TAPTYPE) \
static void ff_put_vp8_ ## FILTERTYPE ## 16_ ## TAPTYPE ## _ ## OPT( \
    uint8_t *dst,  ptrdiff_t dststride, uint8_t *src, \
    ptrdiff_t srcstride, int height, int mx, int my) \
{ \
    ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
        dst,     dststride, src,     srcstride, height, mx, my); \
    ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
        dst + 8, dststride, src + 8, srcstride, height, mx, my); \
}
#define TAP_W8(OPT, FILTERTYPE, TAPTYPE) \
static void ff_put_vp8_ ## FILTERTYPE ## 8_ ## TAPTYPE ## _ ## OPT( \
    uint8_t *dst,  ptrdiff_t dststride, uint8_t *src, \
    ptrdiff_t srcstride, int height, int mx, int my) \
{ \
    ff_put_vp8_ ## FILTERTYPE ## 4_ ## TAPTYPE ## _ ## OPT( \
        dst,     dststride, src,     srcstride, height, mx, my); \
    ff_put_vp8_ ## FILTERTYPE ## 4_ ## TAPTYPE ## _ ## OPT( \
        dst + 4, dststride, src + 4, srcstride, height, mx, my); \
}

#if ARCH_X86_32
TAP_W8 (mmxext, epel, h4)
TAP_W8 (mmxext, epel, h6)
TAP_W16(mmxext, epel, h6)
TAP_W8 (mmxext, epel, v4)
TAP_W8 (mmxext, epel, v6)
TAP_W16(mmxext, epel, v6)
TAP_W8 (mmxext, bilinear, h)
TAP_W16(mmxext, bilinear, h)
TAP_W8 (mmxext, bilinear, v)
TAP_W16(mmxext, bilinear, v)
#endif

TAP_W16(sse2,  epel, h6)
TAP_W16(sse2,  epel, v6)
TAP_W16(sse2,  bilinear, h)
TAP_W16(sse2,  bilinear, v)

TAP_W16(ssse3, epel, h6)
TAP_W16(ssse3, epel, v6)
TAP_W16(ssse3, bilinear, h)
TAP_W16(ssse3, bilinear, v)

#define HVTAP(OPT, ALIGN, TAPNUMX, TAPNUMY, SIZE, MAXHEIGHT) \
static void ff_put_vp8_epel ## SIZE ## _h ## TAPNUMX ## v ## TAPNUMY ## _ ## OPT( \
    uint8_t *dst, ptrdiff_t dststride, uint8_t *src, \
    ptrdiff_t srcstride, int height, int mx, int my) \
{ \
    LOCAL_ALIGNED(ALIGN, uint8_t, tmp, [SIZE * (MAXHEIGHT + TAPNUMY - 1)]); \
    uint8_t *tmpptr = tmp + SIZE * (TAPNUMY / 2 - 1); \
    src -= srcstride * (TAPNUMY / 2 - 1); \
    ff_put_vp8_epel ## SIZE ## _h ## TAPNUMX ## _ ## OPT( \
        tmp, SIZE,      src,    srcstride, height + TAPNUMY - 1, mx, my); \
    ff_put_vp8_epel ## SIZE ## _v ## TAPNUMY ## _ ## OPT( \
        dst, dststride, tmpptr, SIZE,      height,               mx, my); \
}

#if ARCH_X86_32
#define HVTAPMMX(x, y) \
HVTAP(mmxext, 8, x, y,  4,  8) \
HVTAP(mmxext, 8, x, y,  8, 16)

HVTAP(mmxext, 8, 6, 6, 16, 16)
#else
#define HVTAPMMX(x, y) \
HVTAP(mmxext, 8, x, y,  4,  8)
#endif

HVTAPMMX(4, 4)
HVTAPMMX(4, 6)
HVTAPMMX(6, 4)
HVTAPMMX(6, 6)

#define HVTAPSSE2(x, y, w) \
HVTAP(sse2,  16, x, y, w, 16) \
HVTAP(ssse3, 16, x, y, w, 16)

HVTAPSSE2(4, 4, 8)
HVTAPSSE2(4, 6, 8)
HVTAPSSE2(6, 4, 8)
HVTAPSSE2(6, 6, 8)
HVTAPSSE2(6, 6, 16)

HVTAP(ssse3, 16, 4, 4, 4, 8)
HVTAP(ssse3, 16, 4, 6, 4, 8)
HVTAP(ssse3, 16, 6, 4, 4, 8)
HVTAP(ssse3, 16, 6, 6, 4, 8)

#define HVBILIN(OPT, ALIGN, SIZE, MAXHEIGHT) \
static void ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT( \
    uint8_t *dst, ptrdiff_t dststride, uint8_t *src, \
    ptrdiff_t srcstride, int height, int mx, int my) \
{ \
    LOCAL_ALIGNED(ALIGN, uint8_t, tmp, [SIZE * (MAXHEIGHT + 2)]); \
    ff_put_vp8_bilinear ## SIZE ## _h_ ## OPT( \
        tmp, SIZE,      src, srcstride, height + 1, mx, my); \
    ff_put_vp8_bilinear ## SIZE ## _v_ ## OPT( \
        dst, dststride, tmp, SIZE,      height,     mx, my); \
}

HVBILIN(mmxext,  8,  4,  8)
#if ARCH_X86_32
HVBILIN(mmxext,  8,  8, 16)
HVBILIN(mmxext,  8, 16, 16)
#endif
HVBILIN(sse2,  8,  8, 16)
HVBILIN(sse2,  8, 16, 16)
HVBILIN(ssse3, 8,  4,  8)
HVBILIN(ssse3, 8,  8, 16)
HVBILIN(ssse3, 8, 16, 16)

void ff_vp8_idct_dc_add_mmx(uint8_t *dst, int16_t block[16],
                            ptrdiff_t stride);
void ff_vp8_idct_dc_add_sse4(uint8_t *dst, int16_t block[16],
                             ptrdiff_t stride);
void ff_vp8_idct_dc_add4y_mmx(uint8_t *dst, int16_t block[4][16],
                               ptrdiff_t stride);
void ff_vp8_idct_dc_add4y_sse2(uint8_t *dst, int16_t block[4][16],
                               ptrdiff_t stride);
void ff_vp8_idct_dc_add4uv_mmx(uint8_t *dst, int16_t block[2][16],
                               ptrdiff_t stride);
void ff_vp8_luma_dc_wht_mmx(int16_t block[4][4][16], int16_t dc[16]);
void ff_vp8_luma_dc_wht_sse(int16_t block[4][4][16], int16_t dc[16]);
void ff_vp8_idct_add_mmx(uint8_t *dst, int16_t block[16], ptrdiff_t stride);
void ff_vp8_idct_add_sse(uint8_t *dst, int16_t block[16], ptrdiff_t stride);

#define DECLARE_LOOP_FILTER(NAME)                                       \
void ff_vp8_v_loop_filter_simple_ ## NAME(uint8_t *dst,                 \
                                          ptrdiff_t stride,             \
                                          int flim);                    \
void ff_vp8_h_loop_filter_simple_ ## NAME(uint8_t *dst,                 \
                                          ptrdiff_t stride,             \
                                          int flim);                    \
void ff_vp8_v_loop_filter16y_inner_ ## NAME (uint8_t *dst,              \
                                             ptrdiff_t stride,          \
                                             int e, int i, int hvt);    \
void ff_vp8_h_loop_filter16y_inner_ ## NAME (uint8_t *dst,              \
                                             ptrdiff_t stride,          \
                                             int e, int i, int hvt);    \
void ff_vp8_v_loop_filter8uv_inner_ ## NAME (uint8_t *dstU,             \
                                             uint8_t *dstV,             \
                                             ptrdiff_t s,               \
                                             int e, int i, int hvt);    \
void ff_vp8_h_loop_filter8uv_inner_ ## NAME (uint8_t *dstU,             \
                                             uint8_t *dstV,             \
                                             ptrdiff_t s,               \
                                             int e, int i, int hvt);    \
void ff_vp8_v_loop_filter16y_mbedge_ ## NAME(uint8_t *dst,              \
                                             ptrdiff_t stride,          \
                                             int e, int i, int hvt);    \
void ff_vp8_h_loop_filter16y_mbedge_ ## NAME(uint8_t *dst,              \
                                             ptrdiff_t stride,          \
                                             int e, int i, int hvt);    \
void ff_vp8_v_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU,             \
                                             uint8_t *dstV,             \
                                             ptrdiff_t s,               \
                                             int e, int i, int hvt);    \
void ff_vp8_h_loop_filter8uv_mbedge_ ## NAME(uint8_t *dstU,             \
                                             uint8_t *dstV,             \
                                             ptrdiff_t s,               \
                                             int e, int i, int hvt);

DECLARE_LOOP_FILTER(mmx)
DECLARE_LOOP_FILTER(mmxext)
DECLARE_LOOP_FILTER(sse2)
DECLARE_LOOP_FILTER(ssse3)
DECLARE_LOOP_FILTER(sse4)

#endif /* HAVE_YASM */

#define VP8_LUMA_MC_FUNC(IDX, SIZE, OPT) \
    c->put_vp8_epel_pixels_tab[IDX][0][2] = ff_put_vp8_epel ## SIZE ## _h6_ ## OPT; \
    c->put_vp8_epel_pixels_tab[IDX][2][0] = ff_put_vp8_epel ## SIZE ## _v6_ ## OPT; \
    c->put_vp8_epel_pixels_tab[IDX][2][2] = ff_put_vp8_epel ## SIZE ## _h6v6_ ## OPT

#define VP8_MC_FUNC(IDX, SIZE, OPT) \
    c->put_vp8_epel_pixels_tab[IDX][0][1] = ff_put_vp8_epel ## SIZE ## _h4_ ## OPT; \
    c->put_vp8_epel_pixels_tab[IDX][1][0] = ff_put_vp8_epel ## SIZE ## _v4_ ## OPT; \
    c->put_vp8_epel_pixels_tab[IDX][1][1] = ff_put_vp8_epel ## SIZE ## _h4v4_ ## OPT; \
    c->put_vp8_epel_pixels_tab[IDX][1][2] = ff_put_vp8_epel ## SIZE ## _h6v4_ ## OPT; \
    c->put_vp8_epel_pixels_tab[IDX][2][1] = ff_put_vp8_epel ## SIZE ## _h4v6_ ## OPT; \
    VP8_LUMA_MC_FUNC(IDX, SIZE, OPT)

#define VP8_BILINEAR_MC_FUNC(IDX, SIZE, OPT) \
    c->put_vp8_bilinear_pixels_tab[IDX][0][1] = ff_put_vp8_bilinear ## SIZE ## _h_ ## OPT; \
    c->put_vp8_bilinear_pixels_tab[IDX][0][2] = ff_put_vp8_bilinear ## SIZE ## _h_ ## OPT; \
    c->put_vp8_bilinear_pixels_tab[IDX][1][0] = ff_put_vp8_bilinear ## SIZE ## _v_ ## OPT; \
    c->put_vp8_bilinear_pixels_tab[IDX][1][1] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT; \
    c->put_vp8_bilinear_pixels_tab[IDX][1][2] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT; \
    c->put_vp8_bilinear_pixels_tab[IDX][2][0] = ff_put_vp8_bilinear ## SIZE ## _v_ ## OPT; \
    c->put_vp8_bilinear_pixels_tab[IDX][2][1] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT; \
    c->put_vp8_bilinear_pixels_tab[IDX][2][2] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT


av_cold void ff_vp78dsp_init_x86(VP8DSPContext *c)
{
#if HAVE_YASM
    int cpu_flags = av_get_cpu_flags();

    if (EXTERNAL_MMX(cpu_flags)) {
#if ARCH_X86_32
        c->put_vp8_epel_pixels_tab[0][0][0]     =
        c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_mmx;
#endif
        c->put_vp8_epel_pixels_tab[1][0][0]     =
        c->put_vp8_bilinear_pixels_tab[1][0][0] = ff_put_vp8_pixels8_mmx;
    }

    /* note that 4-tap width=16 functions are missing because w=16
     * is only used for luma, and luma is always a copy or sixtap. */
    if (EXTERNAL_MMXEXT(cpu_flags)) {
        VP8_MC_FUNC(2, 4, mmxext);
        VP8_BILINEAR_MC_FUNC(2, 4, mmxext);
#if ARCH_X86_32
        VP8_LUMA_MC_FUNC(0, 16, mmxext);
        VP8_MC_FUNC(1, 8, mmxext);
        VP8_BILINEAR_MC_FUNC(0, 16, mmxext);
        VP8_BILINEAR_MC_FUNC(1,  8, mmxext);
#endif
    }

    if (EXTERNAL_SSE(cpu_flags)) {
        c->put_vp8_epel_pixels_tab[0][0][0]     =
        c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_sse;
    }

    if (HAVE_SSE2_EXTERNAL && cpu_flags & (AV_CPU_FLAG_SSE2 | AV_CPU_FLAG_SSE2SLOW)) {
        VP8_LUMA_MC_FUNC(0, 16, sse2);
        VP8_MC_FUNC(1, 8, sse2);
        VP8_BILINEAR_MC_FUNC(0, 16, sse2);
        VP8_BILINEAR_MC_FUNC(1, 8, sse2);
    }

    if (EXTERNAL_SSSE3(cpu_flags)) {
        VP8_LUMA_MC_FUNC(0, 16, ssse3);
        VP8_MC_FUNC(1, 8, ssse3);
        VP8_MC_FUNC(2, 4, ssse3);
        VP8_BILINEAR_MC_FUNC(0, 16, ssse3);
        VP8_BILINEAR_MC_FUNC(1, 8, ssse3);
        VP8_BILINEAR_MC_FUNC(2, 4, ssse3);
    }
#endif /* HAVE_YASM */
}

av_cold void ff_vp8dsp_init_x86(VP8DSPContext *c)
{
#if HAVE_YASM
    int cpu_flags = av_get_cpu_flags();

    if (EXTERNAL_MMX(cpu_flags)) {
        c->vp8_idct_dc_add    = ff_vp8_idct_dc_add_mmx;
        c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx;
#if ARCH_X86_32
        c->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_mmx;
        c->vp8_idct_add       = ff_vp8_idct_add_mmx;
        c->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_mmx;

        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmx;
        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmx;

        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmx;
        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmx;
        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmx;
        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmx;

        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmx;
        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmx;
        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmx;
        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmx;
#endif
    }

    /* note that 4-tap width=16 functions are missing because w=16
     * is only used for luma, and luma is always a copy or sixtap. */
    if (EXTERNAL_MMXEXT(cpu_flags)) {
#if ARCH_X86_32
        c->vp8_v_loop_filter_simple   = ff_vp8_v_loop_filter_simple_mmxext;
        c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_mmxext;

        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmxext;
        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmxext;
        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmxext;
        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmxext;

        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmxext;
        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmxext;
        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmxext;
        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmxext;
#endif
    }

    if (EXTERNAL_SSE(cpu_flags)) {
        c->vp8_idct_add                         = ff_vp8_idct_add_sse;
        c->vp8_luma_dc_wht                      = ff_vp8_luma_dc_wht_sse;
    }

    if (HAVE_SSE2_EXTERNAL && cpu_flags & (AV_CPU_FLAG_SSE2 | AV_CPU_FLAG_SSE2SLOW)) {
        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2;

        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2;
        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_sse2;

        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_sse2;
        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_sse2;
    }

    if (EXTERNAL_SSE2(cpu_flags)) {
        c->vp8_idct_dc_add4y          = ff_vp8_idct_dc_add4y_sse2;

        c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse2;

        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_sse2;
        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_sse2;

        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse2;
        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse2;
    }

    if (EXTERNAL_SSSE3(cpu_flags)) {
        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_ssse3;
        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_ssse3;

        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_ssse3;
        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_ssse3;
        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_ssse3;
        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_ssse3;

        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_ssse3;
        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_ssse3;
        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_ssse3;
        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_ssse3;
    }

    if (EXTERNAL_SSE4(cpu_flags)) {
        c->vp8_idct_dc_add            = ff_vp8_idct_dc_add_sse4;

        c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse4;
        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse4;
        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse4;
    }
#endif /* HAVE_YASM */
}