h264dsp_init.c 20.1 KB
Newer Older
1 2 3
/*
 * Copyright (c) 2004-2005 Michael Niedermayer, Loren Merritt
 *
4 5 6
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
7 8
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
9
 * version 2.1 of the License, or (at your option) any later version.
10
 *
11
 * FFmpeg is distributed in the hope that it will be useful,
12 13 14 15 16
 * 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
17
 * License along with FFmpeg; if not, write to the Free Software
18
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 20
 */

21
#include "libavutil/attributes.h"
22
#include "libavutil/cpu.h"
23
#include "libavutil/x86/asm.h"
24
#include "libavutil/x86/cpu.h"
25
#include "libavcodec/h264dsp.h"
26 27 28

/***********************************/
/* IDCT */
29 30 31 32
#define IDCT_ADD_FUNC(NUM, DEPTH, OPT)                                  \
void ff_h264_idct ## NUM ## _add_ ## DEPTH ## _ ## OPT(uint8_t *dst,    \
                                                       int16_t *block,  \
                                                       int stride);
33

34
IDCT_ADD_FUNC(, 8, mmx)
35
IDCT_ADD_FUNC(, 8, sse2)
36
IDCT_ADD_FUNC(, 8, avx)
37
IDCT_ADD_FUNC(, 10, sse2)
38
IDCT_ADD_FUNC(_dc, 8, mmxext)
39
IDCT_ADD_FUNC(_dc, 8, sse2)
40
IDCT_ADD_FUNC(_dc, 8, avx)
41 42
IDCT_ADD_FUNC(_dc, 10, mmxext)
IDCT_ADD_FUNC(8_dc, 8, mmxext)
43
IDCT_ADD_FUNC(8_dc, 10, sse2)
44 45
IDCT_ADD_FUNC(8, 8, mmx)
IDCT_ADD_FUNC(8, 8, sse2)
46 47 48 49 50 51
IDCT_ADD_FUNC(8, 10, sse2)
IDCT_ADD_FUNC(, 10, avx)
IDCT_ADD_FUNC(8_dc, 10, avx)
IDCT_ADD_FUNC(8, 10, avx)


52 53 54
#define IDCT_ADD_REP_FUNC(NUM, REP, DEPTH, OPT)                         \
void ff_h264_idct ## NUM ## _add ## REP ## _ ## DEPTH ## _ ## OPT       \
    (uint8_t *dst, const int *block_offset,                             \
Diego Biurrun's avatar
Diego Biurrun committed
55
     int16_t *block, int stride, const uint8_t nnzc[6 * 8]);
56

57
IDCT_ADD_REP_FUNC(8, 4, 8, mmx)
58
IDCT_ADD_REP_FUNC(8, 4, 8, mmxext)
59
IDCT_ADD_REP_FUNC(8, 4, 8, sse2)
60 61
IDCT_ADD_REP_FUNC(8, 4, 10, sse2)
IDCT_ADD_REP_FUNC(8, 4, 10, avx)
62
IDCT_ADD_REP_FUNC(, 16, 8, mmx)
63
IDCT_ADD_REP_FUNC(, 16, 8, mmxext)
64
IDCT_ADD_REP_FUNC(, 16, 8, sse2)
65
IDCT_ADD_REP_FUNC(, 16, 10, sse2)
66
IDCT_ADD_REP_FUNC(, 16intra, 8, mmx)
67
IDCT_ADD_REP_FUNC(, 16intra, 8, mmxext)
68
IDCT_ADD_REP_FUNC(, 16intra, 8, sse2)
69 70 71 72 73
IDCT_ADD_REP_FUNC(, 16intra, 10, sse2)
IDCT_ADD_REP_FUNC(, 16, 10, avx)
IDCT_ADD_REP_FUNC(, 16intra, 10, avx)


74 75 76
#define IDCT_ADD_REP_FUNC2(NUM, REP, DEPTH, OPT)                      \
void ff_h264_idct ## NUM ## _add ## REP ## _ ## DEPTH ## _ ## OPT     \
    (uint8_t **dst, const int *block_offset,                          \
Diego Biurrun's avatar
Diego Biurrun committed
77
     int16_t *block, int stride, const uint8_t nnzc[6 * 8]);
78

79
IDCT_ADD_REP_FUNC2(, 8, 8, mmx)
80
IDCT_ADD_REP_FUNC2(, 8, 8, mmxext)
81
IDCT_ADD_REP_FUNC2(, 8, 8, sse2)
82 83
IDCT_ADD_REP_FUNC2(, 8, 10, sse2)
IDCT_ADD_REP_FUNC2(, 8, 10, avx)
84

85 86
IDCT_ADD_REP_FUNC2(, 8_422, 8, mmx)

87 88 89
IDCT_ADD_REP_FUNC2(, 8_422, 10, sse2)
IDCT_ADD_REP_FUNC2(, 8_422, 10, avx)

Diego Biurrun's avatar
Diego Biurrun committed
90 91
void ff_h264_luma_dc_dequant_idct_mmx(int16_t *output, int16_t *input, int qmul);
void ff_h264_luma_dc_dequant_idct_sse2(int16_t *output, int16_t *input, int qmul);
92

93 94 95
/***********************************/
/* deblocking */

96 97 98 99 100
void ff_h264_loop_filter_strength_mmxext(int16_t bS[2][4][4], uint8_t nnz[40],
                                         int8_t ref[2][40],
                                         int16_t mv[2][40][2],
                                         int bidir, int edges, int step,
                                         int mask_mv0, int mask_mv1, int field);
101

102 103
#define LF_FUNC(DIR, TYPE, DEPTH, OPT)                                        \
void ff_deblock_ ## DIR ## _ ## TYPE ## _ ## DEPTH ## _ ## OPT(uint8_t *pix,  \
104
                                                               ptrdiff_t stride, \
105 106 107
                                                               int alpha,     \
                                                               int beta,      \
                                                               int8_t *tc0);
108
#define LF_IFUNC(DIR, TYPE, DEPTH, OPT) \
109
void ff_deblock_ ## DIR ## _ ## TYPE ## _ ## DEPTH ## _ ## OPT(uint8_t *pix,  \
110
                                                               ptrdiff_t stride, \
111 112 113 114
                                                               int alpha,     \
                                                               int beta);

#define LF_FUNCS(type, depth)                   \
115 116
LF_FUNC(h,  chroma,          depth, mmxext)     \
LF_IFUNC(h, chroma_intra,    depth, mmxext)     \
117 118
LF_FUNC(h,  chroma422,       depth, mmxext)     \
LF_IFUNC(h, chroma422_intra, depth, mmxext)     \
119 120 121 122 123 124 125 126 127 128
LF_FUNC(v,  chroma,          depth, mmxext)     \
LF_IFUNC(v, chroma_intra,    depth, mmxext)     \
LF_FUNC(h,  luma,            depth, mmxext)     \
LF_IFUNC(h, luma_intra,      depth, mmxext)     \
LF_FUNC(h,  luma,            depth, sse2)       \
LF_IFUNC(h, luma_intra,      depth, sse2)       \
LF_FUNC(v,  luma,            depth, sse2)       \
LF_IFUNC(v, luma_intra,      depth, sse2)       \
LF_FUNC(h,  chroma,          depth, sse2)       \
LF_IFUNC(h, chroma_intra,    depth, sse2)       \
129 130
LF_FUNC(h,  chroma422,       depth, sse2)       \
LF_IFUNC(h, chroma422_intra, depth, sse2)       \
131 132 133 134 135 136 137 138
LF_FUNC(v,  chroma,          depth, sse2)       \
LF_IFUNC(v, chroma_intra,    depth, sse2)       \
LF_FUNC(h,  luma,            depth, avx)        \
LF_IFUNC(h, luma_intra,      depth, avx)        \
LF_FUNC(v,  luma,            depth, avx)        \
LF_IFUNC(v, luma_intra,      depth, avx)        \
LF_FUNC(h,  chroma,          depth, avx)        \
LF_IFUNC(h, chroma_intra,    depth, avx)        \
139 140
LF_FUNC(h,  chroma422,       depth, avx)        \
LF_IFUNC(h, chroma422_intra, depth, avx)        \
141 142
LF_FUNC(v,  chroma,          depth, avx)        \
LF_IFUNC(v, chroma_intra,    depth, avx)
143

144 145 146
LF_FUNC(h, luma_mbaff, 8, sse2)
LF_FUNC(h, luma_mbaff, 8, avx)

147
LF_FUNCS(uint8_t,   8)
148 149
LF_FUNCS(uint16_t, 10)

150
#if ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL
151
LF_FUNC(v8, luma, 8, mmxext)
152 153
static void deblock_v_luma_8_mmxext(uint8_t *pix, int stride, int alpha,
                                    int beta, int8_t *tc0)
154
{
155
    if ((tc0[0] & tc0[1]) >= 0)
156
        ff_deblock_v8_luma_8_mmxext(pix + 0, stride, alpha, beta, tc0);
157
    if ((tc0[2] & tc0[3]) >= 0)
158
        ff_deblock_v8_luma_8_mmxext(pix + 8, stride, alpha, beta, tc0 + 2);
159
}
160
LF_IFUNC(v8, luma_intra, 8, mmxext)
161 162
static void deblock_v_luma_intra_8_mmxext(uint8_t *pix, int stride,
                                          int alpha, int beta)
163
{
164 165
    ff_deblock_v8_luma_intra_8_mmxext(pix + 0, stride, alpha, beta);
    ff_deblock_v8_luma_intra_8_mmxext(pix + 8, stride, alpha, beta);
166
}
167
#endif /* ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL */
168

169 170
LF_FUNC(v,  luma,       10, mmxext)
LF_IFUNC(v, luma_intra, 10, mmxext)
171

172 173 174
/***********************************/
/* weighted prediction */

175
#define H264_WEIGHT(W, OPT)                                             \
176
void ff_h264_weight_ ## W ## _ ## OPT(uint8_t *dst, ptrdiff_t stride,   \
177 178
                                      int height, int log2_denom,       \
                                      int weight, int offset);
179

180 181
#define H264_BIWEIGHT(W, OPT)                                           \
void ff_h264_biweight_ ## W ## _ ## OPT(uint8_t *dst, uint8_t *src,     \
182
                                        ptrdiff_t stride, int height,   \
183 184
                                        int log2_denom, int weightd,    \
                                        int weights, int offset);
185

186
#define H264_BIWEIGHT_MMX(W)                    \
187 188
    H264_WEIGHT(W, mmxext)                      \
    H264_BIWEIGHT(W, mmxext)
189

190 191 192 193 194
#define H264_BIWEIGHT_MMX_SSE(W)                \
    H264_BIWEIGHT_MMX(W)                        \
    H264_WEIGHT(W, sse2)                        \
    H264_BIWEIGHT(W, sse2)                      \
    H264_BIWEIGHT(W, ssse3)
195 196

H264_BIWEIGHT_MMX_SSE(16)
197 198 199 200 201
H264_BIWEIGHT_MMX_SSE(8)
H264_BIWEIGHT_MMX(4)

#define H264_WEIGHT_10(W, DEPTH, OPT)                                   \
void ff_h264_weight_ ## W ## _ ## DEPTH ## _ ## OPT(uint8_t *dst,       \
202
                                                    ptrdiff_t stride,   \
203 204 205 206 207 208 209 210
                                                    int height,         \
                                                    int log2_denom,     \
                                                    int weight,         \
                                                    int offset);

#define H264_BIWEIGHT_10(W, DEPTH, OPT)                                 \
void ff_h264_biweight_ ## W ## _ ## DEPTH ## _ ## OPT(uint8_t *dst,     \
                                                      uint8_t *src,     \
211
                                                      ptrdiff_t stride, \
212 213 214 215 216 217 218 219 220 221 222
                                                      int height,       \
                                                      int log2_denom,   \
                                                      int weightd,      \
                                                      int weights,      \
                                                      int offset);

#define H264_BIWEIGHT_10_SSE(W, DEPTH)          \
    H264_WEIGHT_10(W, DEPTH, sse2)              \
    H264_WEIGHT_10(W, DEPTH, sse4)              \
    H264_BIWEIGHT_10(W, DEPTH, sse2)            \
    H264_BIWEIGHT_10(W, DEPTH, sse4)
223 224

H264_BIWEIGHT_10_SSE(16, 10)
225 226
H264_BIWEIGHT_10_SSE(8,  10)
H264_BIWEIGHT_10_SSE(4,  10)
227

228 229
av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth,
                                 const int chroma_format_idc)
230
{
231
#if HAVE_X86ASM
232
    int cpu_flags = av_get_cpu_flags();
233

234
    if (EXTERNAL_MMXEXT(cpu_flags) && chroma_format_idc <= 1)
235
        c->h264_loop_filter_strength = ff_h264_loop_filter_strength_mmxext;
236 237

    if (bit_depth == 8) {
238
        if (EXTERNAL_MMX(cpu_flags)) {
239 240 241 242 243 244 245
            c->h264_idct_dc_add   =
            c->h264_idct_add      = ff_h264_idct_add_8_mmx;
            c->h264_idct8_dc_add  =
            c->h264_idct8_add     = ff_h264_idct8_add_8_mmx;

            c->h264_idct_add16 = ff_h264_idct_add16_8_mmx;
            c->h264_idct8_add4 = ff_h264_idct8_add4_8_mmx;
246
            if (chroma_format_idc <= 1) {
247
                c->h264_idct_add8 = ff_h264_idct_add8_8_mmx;
248 249 250
            } else {
                c->h264_idct_add8 = ff_h264_idct_add8_422_8_mmx;
            }
251
            c->h264_idct_add16intra = ff_h264_idct_add16intra_8_mmx;
252
            if (cpu_flags & AV_CPU_FLAG_CMOV)
253
                c->h264_luma_dc_dequant_idct = ff_h264_luma_dc_dequant_idct_mmx;
254 255 256 257 258 259
        }
        if (EXTERNAL_MMXEXT(cpu_flags)) {
            c->h264_idct_dc_add  = ff_h264_idct_dc_add_8_mmxext;
            c->h264_idct8_dc_add = ff_h264_idct8_dc_add_8_mmxext;
            c->h264_idct_add16   = ff_h264_idct_add16_8_mmxext;
            c->h264_idct8_add4   = ff_h264_idct8_add4_8_mmxext;
260
            if (chroma_format_idc <= 1)
261 262 263 264 265
                c->h264_idct_add8 = ff_h264_idct_add8_8_mmxext;
            c->h264_idct_add16intra = ff_h264_idct_add16intra_8_mmxext;

            c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_8_mmxext;
            c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_mmxext;
266
            if (chroma_format_idc <= 1) {
267 268
                c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma_8_mmxext;
                c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma_intra_8_mmxext;
269 270
            } else {
                c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_8_mmxext;
271
                c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_mmxext;
272
            }
273
#if ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL
274 275 276 277
            c->h264_v_loop_filter_luma       = deblock_v_luma_8_mmxext;
            c->h264_h_loop_filter_luma       = ff_deblock_h_luma_8_mmxext;
            c->h264_v_loop_filter_luma_intra = deblock_v_luma_intra_8_mmxext;
            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_mmxext;
278
#endif /* ARCH_X86_32 && HAVE_MMXEXT_EXTERNAL */
279 280 281 282 283 284 285 286 287 288 289 290 291
            c->weight_h264_pixels_tab[0] = ff_h264_weight_16_mmxext;
            c->weight_h264_pixels_tab[1] = ff_h264_weight_8_mmxext;
            c->weight_h264_pixels_tab[2] = ff_h264_weight_4_mmxext;

            c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_mmxext;
            c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_mmxext;
            c->biweight_h264_pixels_tab[2] = ff_h264_biweight_4_mmxext;
        }
        if (EXTERNAL_SSE2(cpu_flags)) {
            c->h264_idct8_add  = ff_h264_idct8_add_8_sse2;

            c->h264_idct_add16 = ff_h264_idct_add16_8_sse2;
            c->h264_idct8_add4 = ff_h264_idct8_add4_8_sse2;
292
            if (chroma_format_idc <= 1)
293 294 295 296 297 298 299 300 301 302 303 304 305 306
                c->h264_idct_add8 = ff_h264_idct_add8_8_sse2;
            c->h264_idct_add16intra      = ff_h264_idct_add16intra_8_sse2;
            c->h264_luma_dc_dequant_idct = ff_h264_luma_dc_dequant_idct_sse2;

            c->weight_h264_pixels_tab[0] = ff_h264_weight_16_sse2;
            c->weight_h264_pixels_tab[1] = ff_h264_weight_8_sse2;

            c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_sse2;
            c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_sse2;

            c->h264_v_loop_filter_luma       = ff_deblock_v_luma_8_sse2;
            c->h264_h_loop_filter_luma       = ff_deblock_h_luma_8_sse2;
            c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_8_sse2;
            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_sse2;
307 308 309 310

#if ARCH_X86_64
            c->h264_h_loop_filter_luma_mbaff = ff_deblock_h_luma_mbaff_8_sse2;
#endif
311 312 313 314 315 316 317 318 319 320

            c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_8_sse2;
            c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_sse2;
            if (chroma_format_idc <= 1) {
                c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma_8_sse2;
                c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma_intra_8_sse2;
            } else {
                c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma422_8_sse2;
                c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_sse2;
            }
321 322 323

            c->h264_idct_add        = ff_h264_idct_add_8_sse2;
            c->h264_idct_dc_add     = ff_h264_idct_dc_add_8_sse2;
324 325 326 327 328 329 330 331 332 333
        }
        if (EXTERNAL_SSSE3(cpu_flags)) {
            c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_ssse3;
            c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_ssse3;
        }
        if (EXTERNAL_AVX(cpu_flags)) {
            c->h264_v_loop_filter_luma       = ff_deblock_v_luma_8_avx;
            c->h264_h_loop_filter_luma       = ff_deblock_h_luma_8_avx;
            c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_8_avx;
            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_avx;
334 335 336
#if ARCH_X86_64
            c->h264_h_loop_filter_luma_mbaff = ff_deblock_h_luma_mbaff_8_avx;
#endif
337 338

            c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_8_avx;
339
            c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_avx;
340 341
            if (chroma_format_idc <= 1) {
                c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma_8_avx;
342
                c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma_intra_8_avx;
343 344
            } else {
                c->h264_h_loop_filter_chroma       = ff_deblock_h_chroma422_8_avx;
345
                c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_avx;
346
            }
347 348

            c->h264_idct_add        = ff_h264_idct_add_8_avx;
349
            c->h264_idct_dc_add     = ff_h264_idct_dc_add_8_avx;
350 351
        }
    } else if (bit_depth == 10) {
352
        if (EXTERNAL_MMXEXT(cpu_flags)) {
353
#if ARCH_X86_32
354 355
            c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_10_mmxext;
            c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_mmxext;
356 357
            if (chroma_format_idc <= 1) {
                c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_mmxext;
358 359
            } else {
                c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_mmxext;
360
            }
361 362 363 364
            c->h264_v_loop_filter_luma         = ff_deblock_v_luma_10_mmxext;
            c->h264_h_loop_filter_luma         = ff_deblock_h_luma_10_mmxext;
            c->h264_v_loop_filter_luma_intra   = ff_deblock_v_luma_intra_10_mmxext;
            c->h264_h_loop_filter_luma_intra   = ff_deblock_h_luma_intra_10_mmxext;
365
#endif /* ARCH_X86_32 */
366 367 368 369 370 371 372
            c->h264_idct_dc_add = ff_h264_idct_dc_add_10_mmxext;
        }
        if (EXTERNAL_SSE2(cpu_flags)) {
            c->h264_idct_add     = ff_h264_idct_add_10_sse2;
            c->h264_idct8_dc_add = ff_h264_idct8_dc_add_10_sse2;

            c->h264_idct_add16 = ff_h264_idct_add16_10_sse2;
373
            if (chroma_format_idc <= 1) {
374
                c->h264_idct_add8 = ff_h264_idct_add8_10_sse2;
375 376 377
            } else {
                c->h264_idct_add8 = ff_h264_idct_add8_422_10_sse2;
            }
378
            c->h264_idct_add16intra = ff_h264_idct_add16intra_10_sse2;
379
#if HAVE_ALIGNED_STACK
380 381
            c->h264_idct8_add  = ff_h264_idct8_add_10_sse2;
            c->h264_idct8_add4 = ff_h264_idct8_add4_10_sse2;
382
#endif /* HAVE_ALIGNED_STACK */
383

384 385 386
            c->weight_h264_pixels_tab[0] = ff_h264_weight_16_10_sse2;
            c->weight_h264_pixels_tab[1] = ff_h264_weight_8_10_sse2;
            c->weight_h264_pixels_tab[2] = ff_h264_weight_4_10_sse2;
387

388 389 390
            c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_10_sse2;
            c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_10_sse2;
            c->biweight_h264_pixels_tab[2] = ff_h264_biweight_4_10_sse2;
391

392 393
            c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_10_sse2;
            c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_sse2;
394 395
            if (chroma_format_idc <= 1) {
                c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_sse2;
396 397
            } else {
                c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_sse2;
398
            }
399
#if HAVE_ALIGNED_STACK
400 401 402 403
            c->h264_v_loop_filter_luma       = ff_deblock_v_luma_10_sse2;
            c->h264_h_loop_filter_luma       = ff_deblock_h_luma_10_sse2;
            c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_10_sse2;
            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_10_sse2;
404
#endif /* HAVE_ALIGNED_STACK */
405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420
        }
        if (EXTERNAL_SSE4(cpu_flags)) {
            c->weight_h264_pixels_tab[0] = ff_h264_weight_16_10_sse4;
            c->weight_h264_pixels_tab[1] = ff_h264_weight_8_10_sse4;
            c->weight_h264_pixels_tab[2] = ff_h264_weight_4_10_sse4;

            c->biweight_h264_pixels_tab[0] = ff_h264_biweight_16_10_sse4;
            c->biweight_h264_pixels_tab[1] = ff_h264_biweight_8_10_sse4;
            c->biweight_h264_pixels_tab[2] = ff_h264_biweight_4_10_sse4;
        }
        if (EXTERNAL_AVX(cpu_flags)) {
            c->h264_idct_dc_add  =
            c->h264_idct_add     = ff_h264_idct_add_10_avx;
            c->h264_idct8_dc_add = ff_h264_idct8_dc_add_10_avx;

            c->h264_idct_add16 = ff_h264_idct_add16_10_avx;
421
            if (chroma_format_idc <= 1) {
422
                c->h264_idct_add8 = ff_h264_idct_add8_10_avx;
423 424 425
            } else {
                c->h264_idct_add8 = ff_h264_idct_add8_422_10_avx;
            }
426
            c->h264_idct_add16intra = ff_h264_idct_add16intra_10_avx;
427
#if HAVE_ALIGNED_STACK
428 429
            c->h264_idct8_add  = ff_h264_idct8_add_10_avx;
            c->h264_idct8_add4 = ff_h264_idct8_add4_10_avx;
430
#endif /* HAVE_ALIGNED_STACK */
431

432 433
            c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_10_avx;
            c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_10_avx;
434 435
            if (chroma_format_idc <= 1) {
                c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_10_avx;
436 437
            } else {
                c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_10_avx;
438
            }
439
#if HAVE_ALIGNED_STACK
440 441 442 443
            c->h264_v_loop_filter_luma         = ff_deblock_v_luma_10_avx;
            c->h264_h_loop_filter_luma         = ff_deblock_h_luma_10_avx;
            c->h264_v_loop_filter_luma_intra   = ff_deblock_v_luma_intra_10_avx;
            c->h264_h_loop_filter_luma_intra   = ff_deblock_h_luma_intra_10_avx;
444
#endif /* HAVE_ALIGNED_STACK */
445 446
        }
    }
447
#endif
448
}