summaryrefslogtreecommitdiff
path: root/libavcodec/sh4/dsputil_align.c
blob: 333e5632d58b304fc9908fb9c911d23a0110c3dc (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
/*
 * aligned/packed access motion
 *
 * Copyright (c) 2001-2003 BERO <bero@geocities.co.jp>
 *
 * 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 "libavcodec/avcodec.h"
#include "libavcodec/dsputil.h"
#include "dsputil_sh4.h"


#define         LP(p)           *(uint32_t*)(p)
#define         LPC(p)          *(const uint32_t*)(p)


#define         UNPACK(ph,pl,tt0,tt1) do { \
        uint32_t t0,t1; t0=tt0;t1=tt1; \
        ph = ( (t0 & ~BYTE_VEC32(0x03))>>2) + ( (t1 & ~BYTE_VEC32(0x03))>>2); \
        pl = (t0 & BYTE_VEC32(0x03)) + (t1 & BYTE_VEC32(0x03)); } while(0)

#define         rnd_PACK(ph,pl,nph,npl) ph + nph + (((pl + npl + BYTE_VEC32(0x02))>>2) & BYTE_VEC32(0x03))
#define         no_rnd_PACK(ph,pl,nph,npl)      ph + nph + (((pl + npl + BYTE_VEC32(0x01))>>2) & BYTE_VEC32(0x03))

/* little-endian */
#define         MERGE1(a,b,ofs) (ofs==0)?a:( ((a)>>(8*ofs))|((b)<<(32-8*ofs)) )
#define         MERGE2(a,b,ofs) (ofs==3)?b:( ((a)>>(8*(ofs+1)))|((b)<<(32-8*(ofs+1))) )
/* big
#define         MERGE1(a,b,ofs) (ofs==0)?a:( ((a)<<(8*ofs))|((b)>>(32-8*ofs)) )
#define         MERGE2(a,b,ofs) (ofs==3)?b:( ((a)<<(8+8*ofs))|((b)>>(32-8-8*ofs)) )
*/


#define         put(d,s)        d = s
#define         avg(d,s)        d = rnd_avg32(s,d)

#define         OP_C4(ofs) \
        ref-=ofs; \
        do { \
                OP(LP(dest),MERGE1(LPC(ref),LPC(ref+4),ofs)); \
                ref+=stride; \
                dest+=stride; \
        } while(--height)

#define        OP_C40() \
        do { \
                OP(LP(dest),LPC(ref)); \
                ref+=stride; \
                dest+=stride; \
        } while(--height)


#define         OP      put

static void put_pixels4_c(uint8_t *dest,const uint8_t *ref, const int stride,int height)
{
        switch((int)ref&3){
        case 0: OP_C40(); return;
        case 1: OP_C4(1); return;
        case 2: OP_C4(2); return;
        case 3: OP_C4(3); return;
        }
}

#undef          OP
#define         OP      avg

static void avg_pixels4_c(uint8_t *dest,const uint8_t *ref, const int stride,int height)
{
        switch((int)ref&3){
        case 0: OP_C40(); return;
        case 1: OP_C4(1); return;
        case 2: OP_C4(2); return;
        case 3: OP_C4(3); return;
        }
}

#undef          OP

#define         OP_C(ofs,sz,avg2) \
{ \
        ref-=ofs; \
        do { \
                uint32_t        t0,t1; \
                t0 = LPC(ref+0); \
                t1 = LPC(ref+4); \
                OP(LP(dest+0), MERGE1(t0,t1,ofs)); \
                t0 = LPC(ref+8); \
                OP(LP(dest+4), MERGE1(t1,t0,ofs)); \
if (sz==16) { \
                t1 = LPC(ref+12); \
                OP(LP(dest+8), MERGE1(t0,t1,ofs)); \
                t0 = LPC(ref+16); \
                OP(LP(dest+12), MERGE1(t1,t0,ofs)); \
} \
                ref+=stride; \
                dest+= stride; \
        } while(--height); \
}

/* aligned */
#define         OP_C0(sz,avg2) \
{ \
        do { \
                OP(LP(dest+0), LPC(ref+0)); \
                OP(LP(dest+4), LPC(ref+4)); \
if (sz==16) { \
                OP(LP(dest+8), LPC(ref+8)); \
                OP(LP(dest+12), LPC(ref+12)); \
} \
                ref+=stride; \
                dest+= stride; \
        } while(--height); \
}

#define         OP_X(ofs,sz,avg2) \
{ \
        ref-=ofs; \
        do { \
                uint32_t        t0,t1; \
                t0 = LPC(ref+0); \
                t1 = LPC(ref+4); \
                OP(LP(dest+0), avg2(MERGE1(t0,t1,ofs),MERGE2(t0,t1,ofs))); \
                t0 = LPC(ref+8); \
                OP(LP(dest+4), avg2(MERGE1(t1,t0,ofs),MERGE2(t1,t0,ofs))); \
if (sz==16) { \
                t1 = LPC(ref+12); \
                OP(LP(dest+8), avg2(MERGE1(t0,t1,ofs),MERGE2(t0,t1,ofs))); \
                t0 = LPC(ref+16); \
                OP(LP(dest+12), avg2(MERGE1(t1,t0,ofs),MERGE2(t1,t0,ofs))); \
} \
                ref+=stride; \
                dest+= stride; \
        } while(--height); \
}

/* aligned */
#define         OP_Y0(sz,avg2) \
{ \
        uint32_t t0,t1,t2,t3,t; \
\
        t0 = LPC(ref+0); \
        t1 = LPC(ref+4); \
if (sz==16) { \
        t2 = LPC(ref+8); \
        t3 = LPC(ref+12); \
} \
        do { \
                ref += stride; \
\
                t = LPC(ref+0); \
                OP(LP(dest+0), avg2(t0,t)); t0 = t; \
                t = LPC(ref+4); \
                OP(LP(dest+4), avg2(t1,t)); t1 = t; \
if (sz==16) { \
                t = LPC(ref+8); \
                OP(LP(dest+8), avg2(t2,t)); t2 = t; \
                t = LPC(ref+12); \
                OP(LP(dest+12), avg2(t3,t)); t3 = t; \
} \
                dest+= stride; \
        } while(--height); \
}

#define         OP_Y(ofs,sz,avg2) \
{ \
        uint32_t t0,t1,t2,t3,t,w0,w1; \
\
        ref-=ofs; \
        w0 = LPC(ref+0); \
        w1 = LPC(ref+4); \
        t0 = MERGE1(w0,w1,ofs); \
        w0 = LPC(ref+8); \
        t1 = MERGE1(w1,w0,ofs); \
if (sz==16) { \
        w1 = LPC(ref+12); \
        t2 = MERGE1(w0,w1,ofs); \
        w0 = LPC(ref+16); \
        t3 = MERGE1(w1,w0,ofs); \
} \
        do { \
                ref += stride; \
\
                w0 = LPC(ref+0); \
                w1 = LPC(ref+4); \
                t = MERGE1(w0,w1,ofs); \
                OP(LP(dest+0), avg2(t0,t)); t0 = t; \
                w0 = LPC(ref+8); \
                t = MERGE1(w1,w0,ofs); \
                OP(LP(dest+4), avg2(t1,t)); t1 = t; \
if (sz==16) { \
                w1 = LPC(ref+12); \
                t = MERGE1(w0,w1,ofs); \
                OP(LP(dest+8), avg2(t2,t)); t2 = t; \
                w0 = LPC(ref+16); \
                t = MERGE1(w1,w0,ofs); \
                OP(LP(dest+12), avg2(t3,t)); t3 = t; \
} \
                dest+=stride; \
        } while(--height); \
}

#define OP_X0(sz,avg2) OP_X(0,sz,avg2)
#define OP_XY0(sz,PACK) OP_XY(0,sz,PACK)
#define         OP_XY(ofs,sz,PACK) \
{ \
        uint32_t        t2,t3,w0,w1; \
        uint32_t        a0,a1,a2,a3,a4,a5,a6,a7; \
\
        ref -= ofs; \
        w0 = LPC(ref+0); \
        w1 = LPC(ref+4); \
        UNPACK(a0,a1,MERGE1(w0,w1,ofs),MERGE2(w0,w1,ofs)); \
        w0 = LPC(ref+8); \
        UNPACK(a2,a3,MERGE1(w1,w0,ofs),MERGE2(w1,w0,ofs)); \
if (sz==16) { \
        w1 = LPC(ref+12); \
        UNPACK(a4,a5,MERGE1(w0,w1,ofs),MERGE2(w0,w1,ofs)); \
        w0 = LPC(ref+16); \
        UNPACK(a6,a7,MERGE1(w1,w0,ofs),MERGE2(w1,w0,ofs)); \
} \
        do { \
                ref+=stride; \
                w0 = LPC(ref+0); \
                w1 = LPC(ref+4); \
                UNPACK(t2,t3,MERGE1(w0,w1,ofs),MERGE2(w0,w1,ofs)); \
                OP(LP(dest+0),PACK(a0,a1,t2,t3)); \
                a0 = t2; a1 = t3; \
                w0 = LPC(ref+8); \
                UNPACK(t2,t3,MERGE1(w1,w0,ofs),MERGE2(w1,w0,ofs)); \
                OP(LP(dest+4),PACK(a2,a3,t2,t3)); \
                a2 = t2; a3 = t3; \
if (sz==16) { \
                w1 = LPC(ref+12); \
                UNPACK(t2,t3,MERGE1(w0,w1,ofs),MERGE2(w0,w1,ofs)); \
                OP(LP(dest+8),PACK(a4,a5,t2,t3)); \
                a4 = t2; a5 = t3; \
                w0 = LPC(ref+16); \
                UNPACK(t2,t3,MERGE1(w1,w0,ofs),MERGE2(w1,w0,ofs)); \
                OP(LP(dest+12),PACK(a6,a7,t2,t3)); \
                a6 = t2; a7 = t3; \
} \
                dest+=stride; \
        } while(--height); \
}

#define         DEFFUNC(op,rnd,xy,sz,OP_N,avgfunc) \
static void op##_##rnd##_pixels##sz##_##xy (uint8_t * dest, const uint8_t * ref, \
                                const int stride, int height) \
{ \
        switch((int)ref&3) { \
        case 0:OP_N##0(sz,rnd##_##avgfunc); return; \
        case 1:OP_N(1,sz,rnd##_##avgfunc); return; \
        case 2:OP_N(2,sz,rnd##_##avgfunc); return; \
        case 3:OP_N(3,sz,rnd##_##avgfunc); return; \
        } \
}

#define OP put

DEFFUNC(put,   rnd,o,8,OP_C,avg32)
DEFFUNC(put,   rnd,x,8,OP_X,avg32)
DEFFUNC(put,no_rnd,x,8,OP_X,avg32)
DEFFUNC(put,   rnd,y,8,OP_Y,avg32)
DEFFUNC(put,no_rnd,y,8,OP_Y,avg32)
DEFFUNC(put,   rnd,xy,8,OP_XY,PACK)
DEFFUNC(put,no_rnd,xy,8,OP_XY,PACK)
DEFFUNC(put,   rnd,o,16,OP_C,avg32)
DEFFUNC(put,   rnd,x,16,OP_X,avg32)
DEFFUNC(put,no_rnd,x,16,OP_X,avg32)
DEFFUNC(put,   rnd,y,16,OP_Y,avg32)
DEFFUNC(put,no_rnd,y,16,OP_Y,avg32)
DEFFUNC(put,   rnd,xy,16,OP_XY,PACK)
DEFFUNC(put,no_rnd,xy,16,OP_XY,PACK)

#undef OP
#define OP avg

DEFFUNC(avg,   rnd,o,8,OP_C,avg32)
DEFFUNC(avg,   rnd,x,8,OP_X,avg32)
DEFFUNC(avg,no_rnd,x,8,OP_X,avg32)
DEFFUNC(avg,   rnd,y,8,OP_Y,avg32)
DEFFUNC(avg,no_rnd,y,8,OP_Y,avg32)
DEFFUNC(avg,   rnd,xy,8,OP_XY,PACK)
DEFFUNC(avg,no_rnd,xy,8,OP_XY,PACK)
DEFFUNC(avg,   rnd,o,16,OP_C,avg32)
DEFFUNC(avg,   rnd,x,16,OP_X,avg32)
DEFFUNC(avg,no_rnd,x,16,OP_X,avg32)
DEFFUNC(avg,   rnd,y,16,OP_Y,avg32)
DEFFUNC(avg,no_rnd,y,16,OP_Y,avg32)
DEFFUNC(avg,   rnd,xy,16,OP_XY,PACK)
DEFFUNC(avg,no_rnd,xy,16,OP_XY,PACK)

#undef OP

#define         put_no_rnd_pixels8_o     put_rnd_pixels8_o
#define         put_no_rnd_pixels16_o    put_rnd_pixels16_o
#define         avg_no_rnd_pixels8_o     avg_rnd_pixels8_o
#define         avg_no_rnd_pixels16_o    avg_rnd_pixels16_o

#define         put_pixels8_c            put_rnd_pixels8_o
#define         put_pixels16_c           put_rnd_pixels16_o
#define         avg_pixels8_c            avg_rnd_pixels8_o
#define         avg_pixels16_c           avg_rnd_pixels16_o
#define         put_no_rnd_pixels8_c     put_rnd_pixels8_o
#define         put_no_rnd_pixels16_c    put_rnd_pixels16_o
#define         avg_no_rnd_pixels8_c     avg_rnd_pixels8_o
#define         avg_no_rnd_pixels16_c    avg_rnd_pixels16_o

#define         QPEL

#ifdef QPEL

#include "qpel.c"

#endif

void ff_dsputil_init_align(DSPContext* c, AVCodecContext *avctx)
{
        const int high_bit_depth = avctx->bits_per_raw_sample > 8;

        if (!high_bit_depth) {
        c->put_pixels_tab[0][0] = put_rnd_pixels16_o;
        c->put_pixels_tab[0][1] = put_rnd_pixels16_x;
        c->put_pixels_tab[0][2] = put_rnd_pixels16_y;
        c->put_pixels_tab[0][3] = put_rnd_pixels16_xy;
        c->put_pixels_tab[1][0] = put_rnd_pixels8_o;
        c->put_pixels_tab[1][1] = put_rnd_pixels8_x;
        c->put_pixels_tab[1][2] = put_rnd_pixels8_y;
        c->put_pixels_tab[1][3] = put_rnd_pixels8_xy;

        c->put_no_rnd_pixels_tab[0][0] = put_no_rnd_pixels16_o;
        c->put_no_rnd_pixels_tab[0][1] = put_no_rnd_pixels16_x;
        c->put_no_rnd_pixels_tab[0][2] = put_no_rnd_pixels16_y;
        c->put_no_rnd_pixels_tab[0][3] = put_no_rnd_pixels16_xy;
        c->put_no_rnd_pixels_tab[1][0] = put_no_rnd_pixels8_o;
        c->put_no_rnd_pixels_tab[1][1] = put_no_rnd_pixels8_x;
        c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y;
        c->put_no_rnd_pixels_tab[1][3] = put_no_rnd_pixels8_xy;

        c->avg_pixels_tab[0][0] = avg_rnd_pixels16_o;
        c->avg_pixels_tab[0][1] = avg_rnd_pixels16_x;
        c->avg_pixels_tab[0][2] = avg_rnd_pixels16_y;
        c->avg_pixels_tab[0][3] = avg_rnd_pixels16_xy;
        c->avg_pixels_tab[1][0] = avg_rnd_pixels8_o;
        c->avg_pixels_tab[1][1] = avg_rnd_pixels8_x;
        c->avg_pixels_tab[1][2] = avg_rnd_pixels8_y;
        c->avg_pixels_tab[1][3] = avg_rnd_pixels8_xy;

        c->avg_no_rnd_pixels_tab[0][0] = avg_no_rnd_pixels16_o;
        c->avg_no_rnd_pixels_tab[0][1] = avg_no_rnd_pixels16_x;
        c->avg_no_rnd_pixels_tab[0][2] = avg_no_rnd_pixels16_y;
        c->avg_no_rnd_pixels_tab[0][3] = avg_no_rnd_pixels16_xy;
        c->avg_no_rnd_pixels_tab[1][0] = avg_no_rnd_pixels8_o;
        c->avg_no_rnd_pixels_tab[1][1] = avg_no_rnd_pixels8_x;
        c->avg_no_rnd_pixels_tab[1][2] = avg_no_rnd_pixels8_y;
        c->avg_no_rnd_pixels_tab[1][3] = avg_no_rnd_pixels8_xy;
        }

#ifdef QPEL

#define dspfunc(PFX, IDX, NUM) \
    c->PFX ## _pixels_tab[IDX][ 0] = PFX ## NUM ## _mc00_sh4; \
    c->PFX ## _pixels_tab[IDX][ 1] = PFX ## NUM ## _mc10_sh4; \
    c->PFX ## _pixels_tab[IDX][ 2] = PFX ## NUM ## _mc20_sh4; \
    c->PFX ## _pixels_tab[IDX][ 3] = PFX ## NUM ## _mc30_sh4; \
    c->PFX ## _pixels_tab[IDX][ 4] = PFX ## NUM ## _mc01_sh4; \
    c->PFX ## _pixels_tab[IDX][ 5] = PFX ## NUM ## _mc11_sh4; \
    c->PFX ## _pixels_tab[IDX][ 6] = PFX ## NUM ## _mc21_sh4; \
    c->PFX ## _pixels_tab[IDX][ 7] = PFX ## NUM ## _mc31_sh4; \
    c->PFX ## _pixels_tab[IDX][ 8] = PFX ## NUM ## _mc02_sh4; \
    c->PFX ## _pixels_tab[IDX][ 9] = PFX ## NUM ## _mc12_sh4; \
    c->PFX ## _pixels_tab[IDX][10] = PFX ## NUM ## _mc22_sh4; \
    c->PFX ## _pixels_tab[IDX][11] = PFX ## NUM ## _mc32_sh4; \
    c->PFX ## _pixels_tab[IDX][12] = PFX ## NUM ## _mc03_sh4; \
    c->PFX ## _pixels_tab[IDX][13] = PFX ## NUM ## _mc13_sh4; \
    c->PFX ## _pixels_tab[IDX][14] = PFX ## NUM ## _mc23_sh4; \
    c->PFX ## _pixels_tab[IDX][15] = PFX ## NUM ## _mc33_sh4

    dspfunc(put_qpel, 0, 16);
    dspfunc(put_no_rnd_qpel, 0, 16);

    dspfunc(avg_qpel, 0, 16);
    /* dspfunc(avg_no_rnd_qpel, 0, 16); */

    dspfunc(put_qpel, 1, 8);
    dspfunc(put_no_rnd_qpel, 1, 8);

    dspfunc(avg_qpel, 1, 8);
    /* dspfunc(avg_no_rnd_qpel, 1, 8); */

    if (!high_bit_depth) {
    dspfunc(put_h264_qpel, 0, 16);
    dspfunc(put_h264_qpel, 1, 8);
    dspfunc(put_h264_qpel, 2, 4);
    dspfunc(avg_h264_qpel, 0, 16);
    dspfunc(avg_h264_qpel, 1, 8);
    dspfunc(avg_h264_qpel, 2, 4);
    }

#undef dspfunc
    if (!high_bit_depth) {
    c->put_h264_chroma_pixels_tab[0]= put_h264_chroma_mc8_sh4;
    c->put_h264_chroma_pixels_tab[1]= put_h264_chroma_mc4_sh4;
    c->put_h264_chroma_pixels_tab[2]= put_h264_chroma_mc2_sh4;
    c->avg_h264_chroma_pixels_tab[0]= avg_h264_chroma_mc8_sh4;
    c->avg_h264_chroma_pixels_tab[1]= avg_h264_chroma_mc4_sh4;
    c->avg_h264_chroma_pixels_tab[2]= avg_h264_chroma_mc2_sh4;
    }

    c->put_mspel_pixels_tab[0]= put_mspel8_mc00_sh4;
    c->put_mspel_pixels_tab[1]= put_mspel8_mc10_sh4;
    c->put_mspel_pixels_tab[2]= put_mspel8_mc20_sh4;
    c->put_mspel_pixels_tab[3]= put_mspel8_mc30_sh4;
    c->put_mspel_pixels_tab[4]= put_mspel8_mc02_sh4;
    c->put_mspel_pixels_tab[5]= put_mspel8_mc12_sh4;
    c->put_mspel_pixels_tab[6]= put_mspel8_mc22_sh4;
    c->put_mspel_pixels_tab[7]= put_mspel8_mc32_sh4;

    c->gmc1 = gmc1_c;

#endif
}