summaryrefslogtreecommitdiffstats
path: root/libavcodec/x86/cavsdsp_mmx.c
blob: 3bc62ea156916b030256854d6052239fe2f76aa8 (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
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
/*
 * Chinese AVS video (AVS1-P2, JiZhun profile) decoder.
 * Copyright (c) 2006  Stefan Gehrer <stefan.gehrer@gmx.de>
 *
 * MMX-optimized DSP functions, based on H.264 optimizations by
 * Michael Niedermayer and Loren Merritt
 *
 * This file is part of Libav.
 *
 * Libav 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.
 *
 * Libav 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 Libav; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "libavutil/common.h"
#include "libavutil/cpu.h"
#include "libavutil/x86_cpu.h"
#include "libavcodec/dsputil.h"
#include "libavcodec/cavsdsp.h"
#include "dsputil_mmx.h"

/*****************************************************************************
 *
 * inverse transform
 *
 ****************************************************************************/

static inline void cavs_idct8_1d(int16_t *block, uint64_t bias)
{
    __asm__ volatile(
        "movq 112(%0), %%mm4  \n\t" /* mm4 = src7 */
        "movq  16(%0), %%mm5  \n\t" /* mm5 = src1 */
        "movq  80(%0), %%mm2  \n\t" /* mm2 = src5 */
        "movq  48(%0), %%mm7  \n\t" /* mm7 = src3 */
        "movq   %%mm4, %%mm0  \n\t"
        "movq   %%mm5, %%mm3  \n\t"
        "movq   %%mm2, %%mm6  \n\t"
        "movq   %%mm7, %%mm1  \n\t"

        "paddw  %%mm4, %%mm4  \n\t" /* mm4 = 2*src7 */
        "paddw  %%mm3, %%mm3  \n\t" /* mm3 = 2*src1 */
        "paddw  %%mm6, %%mm6  \n\t" /* mm6 = 2*src5 */
        "paddw  %%mm1, %%mm1  \n\t" /* mm1 = 2*src3 */
        "paddw  %%mm4, %%mm0  \n\t" /* mm0 = 3*src7 */
        "paddw  %%mm3, %%mm5  \n\t" /* mm5 = 3*src1 */
        "paddw  %%mm6, %%mm2  \n\t" /* mm2 = 3*src5 */
        "paddw  %%mm1, %%mm7  \n\t" /* mm7 = 3*src3 */
        "psubw  %%mm4, %%mm5  \n\t" /* mm5 = 3*src1 - 2*src7 = a0 */
        "paddw  %%mm6, %%mm7  \n\t" /* mm7 = 3*src3 + 2*src5 = a1 */
        "psubw  %%mm2, %%mm1  \n\t" /* mm1 = 2*src3 - 3*src5 = a2 */
        "paddw  %%mm0, %%mm3  \n\t" /* mm3 = 2*src1 + 3*src7 = a3 */

        "movq   %%mm5, %%mm4  \n\t"
        "movq   %%mm7, %%mm6  \n\t"
        "movq   %%mm3, %%mm0  \n\t"
        "movq   %%mm1, %%mm2  \n\t"
        SUMSUB_BA( %%mm7, %%mm5 )   /* mm7 = a0 + a1  mm5 = a0 - a1 */
        "paddw  %%mm3, %%mm7  \n\t" /* mm7 = a0 + a1 + a3 */
        "paddw  %%mm1, %%mm5  \n\t" /* mm5 = a0 - a1 + a2 */
        "paddw  %%mm7, %%mm7  \n\t"
        "paddw  %%mm5, %%mm5  \n\t"
        "paddw  %%mm6, %%mm7  \n\t" /* mm7 = b4 */
        "paddw  %%mm4, %%mm5  \n\t" /* mm5 = b5 */

        SUMSUB_BA( %%mm1, %%mm3 )   /* mm1 = a3 + a2  mm3 = a3 - a2 */
        "psubw  %%mm1, %%mm4  \n\t" /* mm4 = a0 - a2 - a3 */
        "movq   %%mm4, %%mm1  \n\t" /* mm1 = a0 - a2 - a3 */
        "psubw  %%mm6, %%mm3  \n\t" /* mm3 = a3 - a2 - a1 */
        "paddw  %%mm1, %%mm1  \n\t"
        "paddw  %%mm3, %%mm3  \n\t"
        "psubw  %%mm2, %%mm1  \n\t" /* mm1 = b7 */
        "paddw  %%mm0, %%mm3  \n\t" /* mm3 = b6 */

        "movq  32(%0), %%mm2  \n\t" /* mm2 = src2 */
        "movq  96(%0), %%mm6  \n\t" /* mm6 = src6 */
        "movq   %%mm2, %%mm4  \n\t"
        "movq   %%mm6, %%mm0  \n\t"
        "psllw  $2,    %%mm4  \n\t" /* mm4 = 4*src2 */
        "psllw  $2,    %%mm6  \n\t" /* mm6 = 4*src6 */
        "paddw  %%mm4, %%mm2  \n\t" /* mm2 = 5*src2 */
        "paddw  %%mm6, %%mm0  \n\t" /* mm0 = 5*src6 */
        "paddw  %%mm2, %%mm2  \n\t"
        "paddw  %%mm0, %%mm0  \n\t"
        "psubw  %%mm0, %%mm4  \n\t" /* mm4 = 4*src2 - 10*src6 = a7 */
        "paddw  %%mm2, %%mm6  \n\t" /* mm6 = 4*src6 + 10*src2 = a6 */

        "movq    (%0), %%mm2  \n\t" /* mm2 = src0 */
        "movq  64(%0), %%mm0  \n\t" /* mm0 = src4 */
        SUMSUB_BA( %%mm0, %%mm2 )   /* mm0 = src0+src4  mm2 = src0-src4 */
        "psllw  $3,    %%mm0  \n\t"
        "psllw  $3,    %%mm2  \n\t"
        "paddw  %1,    %%mm0  \n\t" /* add rounding bias */
        "paddw  %1,    %%mm2  \n\t" /* add rounding bias */

        SUMSUB_BA( %%mm6, %%mm0 )   /* mm6 = a4 + a6  mm0 = a4 - a6 */
        SUMSUB_BA( %%mm4, %%mm2 )   /* mm4 = a5 + a7  mm2 = a5 - a7 */
        SUMSUB_BA( %%mm7, %%mm6 )   /* mm7 = dst0  mm6 = dst7 */
        SUMSUB_BA( %%mm5, %%mm4 )   /* mm5 = dst1  mm4 = dst6 */
        SUMSUB_BA( %%mm3, %%mm2 )   /* mm3 = dst2  mm2 = dst5 */
        SUMSUB_BA( %%mm1, %%mm0 )   /* mm1 = dst3  mm0 = dst4 */
        :: "r"(block), "m"(bias)
    );
}

static void cavs_idct8_add_mmx(uint8_t *dst, int16_t *block, int stride)
{
    int i;
    DECLARE_ALIGNED(8, int16_t, b2)[64];

    for(i=0; i<2; i++){
        DECLARE_ALIGNED(8, uint64_t, tmp);

        cavs_idct8_1d(block+4*i, ff_pw_4.a);

        __asm__ volatile(
            "psraw     $3, %%mm7  \n\t"
            "psraw     $3, %%mm6  \n\t"
            "psraw     $3, %%mm5  \n\t"
            "psraw     $3, %%mm4  \n\t"
            "psraw     $3, %%mm3  \n\t"
            "psraw     $3, %%mm2  \n\t"
            "psraw     $3, %%mm1  \n\t"
            "psraw     $3, %%mm0  \n\t"
            "movq   %%mm7,    %0   \n\t"
            TRANSPOSE4( %%mm0, %%mm2, %%mm4, %%mm6, %%mm7 )
            "movq   %%mm0,  8(%1)  \n\t"
            "movq   %%mm6, 24(%1)  \n\t"
            "movq   %%mm7, 40(%1)  \n\t"
            "movq   %%mm4, 56(%1)  \n\t"
            "movq    %0,    %%mm7  \n\t"
            TRANSPOSE4( %%mm7, %%mm5, %%mm3, %%mm1, %%mm0 )
            "movq   %%mm7,   (%1)  \n\t"
            "movq   %%mm1, 16(%1)  \n\t"
            "movq   %%mm0, 32(%1)  \n\t"
            "movq   %%mm3, 48(%1)  \n\t"
            : "=m"(tmp)
            : "r"(b2+32*i)
            : "memory"
        );
    }

    for(i=0; i<2; i++){
        cavs_idct8_1d(b2+4*i, ff_pw_64.a);

        __asm__ volatile(
            "psraw     $7, %%mm7  \n\t"
            "psraw     $7, %%mm6  \n\t"
            "psraw     $7, %%mm5  \n\t"
            "psraw     $7, %%mm4  \n\t"
            "psraw     $7, %%mm3  \n\t"
            "psraw     $7, %%mm2  \n\t"
            "psraw     $7, %%mm1  \n\t"
            "psraw     $7, %%mm0  \n\t"
            "movq   %%mm7,    (%0)  \n\t"
            "movq   %%mm5,  16(%0)  \n\t"
            "movq   %%mm3,  32(%0)  \n\t"
            "movq   %%mm1,  48(%0)  \n\t"
            "movq   %%mm0,  64(%0)  \n\t"
            "movq   %%mm2,  80(%0)  \n\t"
            "movq   %%mm4,  96(%0)  \n\t"
            "movq   %%mm6, 112(%0)  \n\t"
            :: "r"(b2+4*i)
            : "memory"
        );
    }

    ff_add_pixels_clamped_mmx(b2, dst, stride);
}

/*****************************************************************************
 *
 * motion compensation
 *
 ****************************************************************************/

/* vertical filter [-1 -2 96 42 -7  0]  */
#define QPEL_CAVSV1(A,B,C,D,E,F,OP,MUL2) \
        "movd (%0), "#F"            \n\t"\
        "movq "#C", %%mm6           \n\t"\
        "pmullw %5, %%mm6           \n\t"\
        "movq "#D", %%mm7           \n\t"\
        "pmullw "MANGLE(MUL2)", %%mm7\n\t"\
        "psllw $3, "#E"             \n\t"\
        "psubw "#E", %%mm6          \n\t"\
        "psraw $3, "#E"             \n\t"\
        "paddw %%mm7, %%mm6         \n\t"\
        "paddw "#E", %%mm6          \n\t"\
        "paddw "#B", "#B"           \n\t"\
        "pxor %%mm7, %%mm7          \n\t"\
        "add %2, %0                 \n\t"\
        "punpcklbw %%mm7, "#F"      \n\t"\
        "psubw "#B", %%mm6          \n\t"\
        "psraw $1, "#B"             \n\t"\
        "psubw "#A", %%mm6          \n\t"\
        "paddw %4, %%mm6            \n\t"\
        "psraw $7, %%mm6            \n\t"\
        "packuswb %%mm6, %%mm6      \n\t"\
        OP(%%mm6, (%1), A, d)            \
        "add %3, %1                 \n\t"

/* vertical filter [ 0 -1  5  5 -1  0]  */
#define QPEL_CAVSV2(A,B,C,D,E,F,OP,MUL2) \
        "movd (%0), "#F"            \n\t"\
        "movq "#C", %%mm6           \n\t"\
        "paddw "#D", %%mm6          \n\t"\
        "pmullw %5, %%mm6           \n\t"\
        "add %2, %0                 \n\t"\
        "punpcklbw %%mm7, "#F"      \n\t"\
        "psubw "#B", %%mm6          \n\t"\
        "psubw "#E", %%mm6          \n\t"\
        "paddw %4, %%mm6            \n\t"\
        "psraw $3, %%mm6            \n\t"\
        "packuswb %%mm6, %%mm6      \n\t"\
        OP(%%mm6, (%1), A, d)            \
        "add %3, %1                 \n\t"

/* vertical filter [ 0 -7 42 96 -2 -1]  */
#define QPEL_CAVSV3(A,B,C,D,E,F,OP,MUL2) \
        "movd (%0), "#F"            \n\t"\
        "movq "#C", %%mm6           \n\t"\
        "pmullw "MANGLE(MUL2)", %%mm6\n\t"\
        "movq "#D", %%mm7           \n\t"\
        "pmullw %5, %%mm7           \n\t"\
        "psllw $3, "#B"             \n\t"\
        "psubw "#B", %%mm6          \n\t"\
        "psraw $3, "#B"             \n\t"\
        "paddw %%mm7, %%mm6         \n\t"\
        "paddw "#B", %%mm6          \n\t"\
        "paddw "#E", "#E"           \n\t"\
        "pxor %%mm7, %%mm7          \n\t"\
        "add %2, %0                 \n\t"\
        "punpcklbw %%mm7, "#F"      \n\t"\
        "psubw "#E", %%mm6          \n\t"\
        "psraw $1, "#E"             \n\t"\
        "psubw "#F", %%mm6          \n\t"\
        "paddw %4, %%mm6            \n\t"\
        "psraw $7, %%mm6            \n\t"\
        "packuswb %%mm6, %%mm6      \n\t"\
        OP(%%mm6, (%1), A, d)            \
        "add %3, %1                 \n\t"


#define QPEL_CAVSVNUM(VOP,OP,ADD,MUL1,MUL2)\
    int w= 2;\
    src -= 2*srcStride;\
    \
    while(w--){\
      __asm__ volatile(\
        "pxor %%mm7, %%mm7          \n\t"\
        "movd (%0), %%mm0           \n\t"\
        "add %2, %0                 \n\t"\
        "movd (%0), %%mm1           \n\t"\
        "add %2, %0                 \n\t"\
        "movd (%0), %%mm2           \n\t"\
        "add %2, %0                 \n\t"\
        "movd (%0), %%mm3           \n\t"\
        "add %2, %0                 \n\t"\
        "movd (%0), %%mm4           \n\t"\
        "add %2, %0                 \n\t"\
        "punpcklbw %%mm7, %%mm0     \n\t"\
        "punpcklbw %%mm7, %%mm1     \n\t"\
        "punpcklbw %%mm7, %%mm2     \n\t"\
        "punpcklbw %%mm7, %%mm3     \n\t"\
        "punpcklbw %%mm7, %%mm4     \n\t"\
        VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP, MUL2)\
        VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP, MUL2)\
        VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP, MUL2)\
        VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP, MUL2)\
        VOP(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP, MUL2)\
        VOP(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP, MUL2)\
        VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP, MUL2)\
        VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP, MUL2)\
        \
        : "+a"(src), "+c"(dst)\
        : "S"((x86_reg)srcStride), "r"((x86_reg)dstStride), "m"(ADD), "m"(MUL1)\
        : "memory"\
     );\
     if(h==16){\
        __asm__ volatile(\
            VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP, MUL2)\
            VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP, MUL2)\
            VOP(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP, MUL2)\
            VOP(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP, MUL2)\
            VOP(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP, MUL2)\
            VOP(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP, MUL2)\
            VOP(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP, MUL2)\
            VOP(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP, MUL2)\
            \
           : "+a"(src), "+c"(dst)\
           : "S"((x86_reg)srcStride), "r"((x86_reg)dstStride), "m"(ADD),  "m"(MUL1)\
           : "memory"\
        );\
     }\
     src += 4-(h+5)*srcStride;\
     dst += 4-h*dstStride;\
   }

#define QPEL_CAVS(OPNAME, OP, MMX)\
static void OPNAME ## cavs_qpel8_h_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    int h=8;\
    __asm__ volatile(\
        "pxor %%mm7, %%mm7          \n\t"\
        "movq %5, %%mm6             \n\t"\
        "1:                         \n\t"\
        "movq    (%0), %%mm0        \n\t"\
        "movq   1(%0), %%mm2        \n\t"\
        "movq %%mm0, %%mm1          \n\t"\
        "movq %%mm2, %%mm3          \n\t"\
        "punpcklbw %%mm7, %%mm0     \n\t"\
        "punpckhbw %%mm7, %%mm1     \n\t"\
        "punpcklbw %%mm7, %%mm2     \n\t"\
        "punpckhbw %%mm7, %%mm3     \n\t"\
        "paddw %%mm2, %%mm0         \n\t"\
        "paddw %%mm3, %%mm1         \n\t"\
        "pmullw %%mm6, %%mm0        \n\t"\
        "pmullw %%mm6, %%mm1        \n\t"\
        "movq   -1(%0), %%mm2       \n\t"\
        "movq    2(%0), %%mm4       \n\t"\
        "movq %%mm2, %%mm3          \n\t"\
        "movq %%mm4, %%mm5          \n\t"\
        "punpcklbw %%mm7, %%mm2     \n\t"\
        "punpckhbw %%mm7, %%mm3     \n\t"\
        "punpcklbw %%mm7, %%mm4     \n\t"\
        "punpckhbw %%mm7, %%mm5     \n\t"\
        "paddw %%mm4, %%mm2         \n\t"\
        "paddw %%mm3, %%mm5         \n\t"\
        "psubw %%mm2, %%mm0         \n\t"\
        "psubw %%mm5, %%mm1         \n\t"\
        "movq %6, %%mm5             \n\t"\
        "paddw %%mm5, %%mm0         \n\t"\
        "paddw %%mm5, %%mm1         \n\t"\
        "psraw $3, %%mm0            \n\t"\
        "psraw $3, %%mm1            \n\t"\
        "packuswb %%mm1, %%mm0      \n\t"\
        OP(%%mm0, (%1),%%mm5, q)         \
        "add %3, %0                 \n\t"\
        "add %4, %1                 \n\t"\
        "decl %2                    \n\t"\
        " jnz 1b                    \n\t"\
        : "+a"(src), "+c"(dst), "+m"(h)\
        : "d"((x86_reg)srcStride), "S"((x86_reg)dstStride), "m"(ff_pw_5), "m"(ff_pw_4)\
        : "memory"\
    );\
}\
\
static inline void OPNAME ## cavs_qpel8or16_v1_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
  QPEL_CAVSVNUM(QPEL_CAVSV1,OP,ff_pw_64,ff_pw_96,ff_pw_42)      \
}\
\
static inline void OPNAME ## cavs_qpel8or16_v2_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
  QPEL_CAVSVNUM(QPEL_CAVSV2,OP,ff_pw_4,ff_pw_5,ff_pw_5)         \
}\
\
static inline void OPNAME ## cavs_qpel8or16_v3_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
  QPEL_CAVSVNUM(QPEL_CAVSV3,OP,ff_pw_64,ff_pw_96,ff_pw_42)      \
}\
\
static void OPNAME ## cavs_qpel8_v1_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    OPNAME ## cavs_qpel8or16_v1_ ## MMX(dst  , src  , dstStride, srcStride, 8);\
}\
static void OPNAME ## cavs_qpel16_v1_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    OPNAME ## cavs_qpel8or16_v1_ ## MMX(dst  , src  , dstStride, srcStride, 16);\
    OPNAME ## cavs_qpel8or16_v1_ ## MMX(dst+8, src+8, dstStride, srcStride, 16);\
}\
\
static void OPNAME ## cavs_qpel8_v2_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    OPNAME ## cavs_qpel8or16_v2_ ## MMX(dst  , src  , dstStride, srcStride, 8);\
}\
static void OPNAME ## cavs_qpel16_v2_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    OPNAME ## cavs_qpel8or16_v2_ ## MMX(dst  , src  , dstStride, srcStride, 16);\
    OPNAME ## cavs_qpel8or16_v2_ ## MMX(dst+8, src+8, dstStride, srcStride, 16);\
}\
\
static void OPNAME ## cavs_qpel8_v3_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    OPNAME ## cavs_qpel8or16_v3_ ## MMX(dst  , src  , dstStride, srcStride, 8);\
}\
static void OPNAME ## cavs_qpel16_v3_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    OPNAME ## cavs_qpel8or16_v3_ ## MMX(dst  , src  , dstStride, srcStride, 16);\
    OPNAME ## cavs_qpel8or16_v3_ ## MMX(dst+8, src+8, dstStride, srcStride, 16);\
}\
\
static void OPNAME ## cavs_qpel16_h_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
    OPNAME ## cavs_qpel8_h_ ## MMX(dst  , src  , dstStride, srcStride);\
    OPNAME ## cavs_qpel8_h_ ## MMX(dst+8, src+8, dstStride, srcStride);\
    src += 8*srcStride;\
    dst += 8*dstStride;\
    OPNAME ## cavs_qpel8_h_ ## MMX(dst  , src  , dstStride, srcStride);\
    OPNAME ## cavs_qpel8_h_ ## MMX(dst+8, src+8, dstStride, srcStride);\
}\

#define CAVS_MC(OPNAME, SIZE, MMX) \
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc20_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
    OPNAME ## cavs_qpel ## SIZE ## _h_ ## MMX(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc01_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
    OPNAME ## cavs_qpel ## SIZE ## _v1_ ## MMX(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc02_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
    OPNAME ## cavs_qpel ## SIZE ## _v2_ ## MMX(dst, src, stride, stride);\
}\
\
static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc03_ ## MMX(uint8_t *dst, uint8_t *src, int stride){\
    OPNAME ## cavs_qpel ## SIZE ## _v3_ ## MMX(dst, src, stride, stride);\
}\

#define PUT_OP(a,b,temp, size) "mov" #size " " #a ", " #b "    \n\t"
#define AVG_3DNOW_OP(a,b,temp, size) \
"mov" #size " " #b ", " #temp "   \n\t"\
"pavgusb " #temp ", " #a "        \n\t"\
"mov" #size " " #a ", " #b "      \n\t"
#define AVG_MMX2_OP(a,b,temp, size) \
"mov" #size " " #b ", " #temp "   \n\t"\
"pavgb " #temp ", " #a "          \n\t"\
"mov" #size " " #a ", " #b "      \n\t"

QPEL_CAVS(put_,       PUT_OP, 3dnow)
QPEL_CAVS(avg_, AVG_3DNOW_OP, 3dnow)
QPEL_CAVS(put_,       PUT_OP, mmx2)
QPEL_CAVS(avg_,  AVG_MMX2_OP, mmx2)

CAVS_MC(put_, 8, 3dnow)
CAVS_MC(put_, 16,3dnow)
CAVS_MC(avg_, 8, 3dnow)
CAVS_MC(avg_, 16,3dnow)
CAVS_MC(put_, 8, mmx2)
CAVS_MC(put_, 16,mmx2)
CAVS_MC(avg_, 8, mmx2)
CAVS_MC(avg_, 16,mmx2)

static void ff_cavsdsp_init_mmx2(CAVSDSPContext* c, AVCodecContext *avctx) {
#define dspfunc(PFX, IDX, NUM) \
    c->PFX ## _pixels_tab[IDX][ 0] = ff_ ## PFX ## NUM ## _mc00_mmx2; \
    c->PFX ## _pixels_tab[IDX][ 2] = ff_ ## PFX ## NUM ## _mc20_mmx2; \
    c->PFX ## _pixels_tab[IDX][ 4] = ff_ ## PFX ## NUM ## _mc01_mmx2; \
    c->PFX ## _pixels_tab[IDX][ 8] = ff_ ## PFX ## NUM ## _mc02_mmx2; \
    c->PFX ## _pixels_tab[IDX][12] = ff_ ## PFX ## NUM ## _mc03_mmx2; \

    dspfunc(put_cavs_qpel, 0, 16);
    dspfunc(put_cavs_qpel, 1, 8);
    dspfunc(avg_cavs_qpel, 0, 16);
    dspfunc(avg_cavs_qpel, 1, 8);
#undef dspfunc
    c->cavs_idct8_add = cavs_idct8_add_mmx;
}

static void ff_cavsdsp_init_3dnow(CAVSDSPContext* c, AVCodecContext *avctx) {
#define dspfunc(PFX, IDX, NUM) \
    c->PFX ## _pixels_tab[IDX][ 0] = ff_ ## PFX ## NUM ## _mc00_mmx2; \
    c->PFX ## _pixels_tab[IDX][ 2] = ff_ ## PFX ## NUM ## _mc20_3dnow; \
    c->PFX ## _pixels_tab[IDX][ 4] = ff_ ## PFX ## NUM ## _mc01_3dnow; \
    c->PFX ## _pixels_tab[IDX][ 8] = ff_ ## PFX ## NUM ## _mc02_3dnow; \
    c->PFX ## _pixels_tab[IDX][12] = ff_ ## PFX ## NUM ## _mc03_3dnow; \

    dspfunc(put_cavs_qpel, 0, 16);
    dspfunc(put_cavs_qpel, 1, 8);
    dspfunc(avg_cavs_qpel, 0, 16);
    dspfunc(avg_cavs_qpel, 1, 8);
#undef dspfunc
    c->cavs_idct8_add = cavs_idct8_add_mmx;
}

void ff_cavsdsp_init_mmx(CAVSDSPContext *c, AVCodecContext *avctx)
{
    int mm_flags = av_get_cpu_flags();

    if (mm_flags & AV_CPU_FLAG_MMX2)  ff_cavsdsp_init_mmx2 (c, avctx);
    if (mm_flags & AV_CPU_FLAG_3DNOW) ff_cavsdsp_init_3dnow(c, avctx);
}
OpenPOWER on IntegriCloud