2 * Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>
4 * This file is part of FFmpeg.
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * FFmpeg is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * 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
23 #include "libswscale/swscale.h"
24 #include "libswscale/swscale_internal.h"
25 #include "libavutil/attributes.h"
26 #include "libavutil/avassert.h"
27 #include "libavutil/intreadwrite.h"
28 #include "libavutil/x86/cpu.h"
29 #include "libavutil/cpu.h"
30 #include "libavutil/mem_internal.h"
31 #include "libavutil/pixdesc.h"
33 const DECLARE_ALIGNED(8, uint64_t, ff_dither4)[2] = {
35 0x0200020002000200LL,};
37 const DECLARE_ALIGNED(8, uint64_t, ff_dither8)[2] = {
39 0x0004000400040004LL,};
45 DECLARE_ASM_CONST(8, uint64_t, bF8)= 0xF8F8F8F8F8F8F8F8LL;
46 DECLARE_ASM_CONST(8, uint64_t, bFC)= 0xFCFCFCFCFCFCFCFCLL;
47 DECLARE_ASM_CONST(8, uint64_t, w10)= 0x0010001000100010LL;
48 DECLARE_ASM_CONST(8, uint64_t, w02)= 0x0002000200020002LL;
50 DECLARE_ASM_CONST(8, uint64_t, b16Mask)= 0x001F001F001F001FLL;
51 DECLARE_ASM_CONST(8, uint64_t, g16Mask)= 0x07E007E007E007E0LL;
52 DECLARE_ASM_CONST(8, uint64_t, r16Mask)= 0xF800F800F800F800LL;
53 DECLARE_ASM_CONST(8, uint64_t, b15Mask)= 0x001F001F001F001FLL;
54 DECLARE_ASM_CONST(8, uint64_t, g15Mask)= 0x03E003E003E003E0LL;
55 DECLARE_ASM_CONST(8, uint64_t, r15Mask)= 0x7C007C007C007C00LL;
57 DECLARE_ASM_ALIGNED(8, const uint64_t, ff_M24A) = 0x00FF0000FF0000FFLL;
58 DECLARE_ASM_ALIGNED(8, const uint64_t, ff_M24B) = 0xFF0000FF0000FF00LL;
59 DECLARE_ASM_ALIGNED(8, const uint64_t, ff_M24C) = 0x0000FF0000FF0000LL;
61 DECLARE_ASM_ALIGNED(8, const uint64_t, ff_bgr2YOffset) = 0x1010101010101010ULL;
62 DECLARE_ASM_ALIGNED(8, const uint64_t, ff_bgr2UVOffset) = 0x8080808080808080ULL;
63 DECLARE_ASM_ALIGNED(8, const uint64_t, ff_w1111) = 0x0001000100010001ULL;
69 #define COMPILE_TEMPLATE_MMXEXT 0
70 #define RENAME(a) a ## _mmx
71 #include "swscale_template.c"
75 #if HAVE_MMXEXT_INLINE
77 #undef COMPILE_TEMPLATE_MMXEXT
78 #define COMPILE_TEMPLATE_MMXEXT 1
79 #define RENAME(a) a ## _mmxext
80 #include "swscale_template.c"
83 void ff_updateMMXDitherTables(SwsContext *c, int dstY)
85 const int dstH= c->dstH;
86 const int flags= c->flags;
88 SwsPlane *lumPlane = &c->slice[c->numSlice-2].plane[0];
89 SwsPlane *chrUPlane = &c->slice[c->numSlice-2].plane[1];
90 SwsPlane *alpPlane = &c->slice[c->numSlice-2].plane[3];
92 int hasAlpha = c->needAlpha;
93 int32_t *vLumFilterPos= c->vLumFilterPos;
94 int32_t *vChrFilterPos= c->vChrFilterPos;
95 int16_t *vLumFilter= c->vLumFilter;
96 int16_t *vChrFilter= c->vChrFilter;
97 int32_t *lumMmxFilter= c->lumMmxFilter;
98 int32_t *chrMmxFilter= c->chrMmxFilter;
99 int32_t av_unused *alpMmxFilter= c->alpMmxFilter;
100 const int vLumFilterSize= c->vLumFilterSize;
101 const int vChrFilterSize= c->vChrFilterSize;
102 const int chrDstY= dstY>>c->chrDstVSubSample;
103 const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input
104 const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input
106 c->blueDither= ff_dither8[dstY&1];
107 if (c->dstFormat == AV_PIX_FMT_RGB555 || c->dstFormat == AV_PIX_FMT_BGR555)
108 c->greenDither= ff_dither8[dstY&1];
110 c->greenDither= ff_dither4[dstY&1];
111 c->redDither= ff_dither8[(dstY+1)&1];
112 if (dstY < dstH - 2) {
113 const int16_t **lumSrcPtr = (const int16_t **)(void*) lumPlane->line + firstLumSrcY - lumPlane->sliceY;
114 const int16_t **chrUSrcPtr = (const int16_t **)(void*) chrUPlane->line + firstChrSrcY - chrUPlane->sliceY;
115 const int16_t **alpSrcPtr = (CONFIG_SWSCALE_ALPHA && hasAlpha) ? (const int16_t **)(void*) alpPlane->line + firstLumSrcY - alpPlane->sliceY : NULL;
118 if (firstLumSrcY < 0 || firstLumSrcY + vLumFilterSize > c->srcH) {
119 const int16_t **tmpY = (const int16_t **) lumPlane->tmp;
121 int neg = -firstLumSrcY, i, end = FFMIN(c->srcH - firstLumSrcY, vLumFilterSize);
122 for (i = 0; i < neg; i++)
123 tmpY[i] = lumSrcPtr[neg];
124 for ( ; i < end; i++)
125 tmpY[i] = lumSrcPtr[i];
126 for ( ; i < vLumFilterSize; i++)
131 const int16_t **tmpA = (const int16_t **) alpPlane->tmp;
132 for (i = 0; i < neg; i++)
133 tmpA[i] = alpSrcPtr[neg];
134 for ( ; i < end; i++)
135 tmpA[i] = alpSrcPtr[i];
136 for ( ; i < vLumFilterSize; i++)
137 tmpA[i] = tmpA[i - 1];
141 if (firstChrSrcY < 0 || firstChrSrcY + vChrFilterSize > c->chrSrcH) {
142 const int16_t **tmpU = (const int16_t **) chrUPlane->tmp;
143 int neg = -firstChrSrcY, i, end = FFMIN(c->chrSrcH - firstChrSrcY, vChrFilterSize);
144 for (i = 0; i < neg; i++) {
145 tmpU[i] = chrUSrcPtr[neg];
147 for ( ; i < end; i++) {
148 tmpU[i] = chrUSrcPtr[i];
150 for ( ; i < vChrFilterSize; i++) {
151 tmpU[i] = tmpU[i - 1];
156 if (flags & SWS_ACCURATE_RND) {
157 int s= APCK_SIZE / 8;
158 for (i=0; i<vLumFilterSize; i+=2) {
159 *(const void**)&lumMmxFilter[s*i ]= lumSrcPtr[i ];
160 *(const void**)&lumMmxFilter[s*i+APCK_PTR2/4 ]= lumSrcPtr[i+(vLumFilterSize>1)];
161 lumMmxFilter[s*i+APCK_COEF/4 ]=
162 lumMmxFilter[s*i+APCK_COEF/4+1]= vLumFilter[dstY*vLumFilterSize + i ]
163 + (vLumFilterSize>1 ? vLumFilter[dstY*vLumFilterSize + i + 1] * (1 << 16) : 0);
164 if (CONFIG_SWSCALE_ALPHA && hasAlpha) {
165 *(const void**)&alpMmxFilter[s*i ]= alpSrcPtr[i ];
166 *(const void**)&alpMmxFilter[s*i+APCK_PTR2/4 ]= alpSrcPtr[i+(vLumFilterSize>1)];
167 alpMmxFilter[s*i+APCK_COEF/4 ]=
168 alpMmxFilter[s*i+APCK_COEF/4+1]= lumMmxFilter[s*i+APCK_COEF/4 ];
171 for (i=0; i<vChrFilterSize; i+=2) {
172 *(const void**)&chrMmxFilter[s*i ]= chrUSrcPtr[i ];
173 *(const void**)&chrMmxFilter[s*i+APCK_PTR2/4 ]= chrUSrcPtr[i+(vChrFilterSize>1)];
174 chrMmxFilter[s*i+APCK_COEF/4 ]=
175 chrMmxFilter[s*i+APCK_COEF/4+1]= vChrFilter[chrDstY*vChrFilterSize + i ]
176 + (vChrFilterSize>1 ? vChrFilter[chrDstY*vChrFilterSize + i + 1] * (1 << 16) : 0);
179 for (i=0; i<vLumFilterSize; i++) {
180 *(const void**)&lumMmxFilter[4*i+0]= lumSrcPtr[i];
183 ((uint16_t)vLumFilter[dstY*vLumFilterSize + i])*0x10001U;
184 if (CONFIG_SWSCALE_ALPHA && hasAlpha) {
185 *(const void**)&alpMmxFilter[4*i+0]= alpSrcPtr[i];
187 alpMmxFilter[4*i+3]= lumMmxFilter[4*i+2];
190 for (i=0; i<vChrFilterSize; i++) {
191 *(const void**)&chrMmxFilter[4*i+0]= chrUSrcPtr[i];
194 ((uint16_t)vChrFilter[chrDstY*vChrFilterSize + i])*0x10001U;
201 static void yuv2yuvX_sse3(const int16_t *filter, int filterSize,
202 const int16_t **src, uint8_t *dest, int dstW,
203 const uint8_t *dither, int offset)
205 if(((uintptr_t)dest) & 15){
206 yuv2yuvX_mmxext(filter, filterSize, src, dest, dstW, dither, offset);
210 #define MAIN_FUNCTION \
211 "pxor %%xmm0, %%xmm0 \n\t" \
212 "punpcklbw %%xmm0, %%xmm3 \n\t" \
213 "movd %4, %%xmm1 \n\t" \
214 "punpcklwd %%xmm1, %%xmm1 \n\t" \
215 "punpckldq %%xmm1, %%xmm1 \n\t" \
216 "punpcklqdq %%xmm1, %%xmm1 \n\t" \
217 "psllw $3, %%xmm1 \n\t" \
218 "paddw %%xmm1, %%xmm3 \n\t" \
219 "psraw $4, %%xmm3 \n\t" \
220 "movdqa %%xmm3, %%xmm4 \n\t" \
221 "movdqa %%xmm3, %%xmm7 \n\t" \
222 "movl %3, %%ecx \n\t" \
223 "mov %0, %%"FF_REG_d" \n\t"\
224 "mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
225 ".p2align 4 \n\t" /* FIXME Unroll? */\
227 "movddup 8(%%"FF_REG_d"), %%xmm0 \n\t" /* filterCoeff */\
228 "movdqa (%%"FF_REG_S", %%"FF_REG_c", 2), %%xmm2 \n\t" /* srcData */\
229 "movdqa 16(%%"FF_REG_S", %%"FF_REG_c", 2), %%xmm5 \n\t" /* srcData */\
230 "add $16, %%"FF_REG_d" \n\t"\
231 "mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
232 "test %%"FF_REG_S", %%"FF_REG_S" \n\t"\
233 "pmulhw %%xmm0, %%xmm2 \n\t"\
234 "pmulhw %%xmm0, %%xmm5 \n\t"\
235 "paddw %%xmm2, %%xmm3 \n\t"\
236 "paddw %%xmm5, %%xmm4 \n\t"\
238 "psraw $3, %%xmm3 \n\t"\
239 "psraw $3, %%xmm4 \n\t"\
240 "packuswb %%xmm4, %%xmm3 \n\t"\
241 "movntdq %%xmm3, (%1, %%"FF_REG_c") \n\t"\
242 "add $16, %%"FF_REG_c" \n\t"\
243 "cmp %2, %%"FF_REG_c" \n\t"\
244 "movdqa %%xmm7, %%xmm3 \n\t" \
245 "movdqa %%xmm7, %%xmm4 \n\t" \
246 "mov %0, %%"FF_REG_d" \n\t"\
247 "mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
252 "movq %5, %%xmm3 \n\t"
253 "movdqa %%xmm3, %%xmm4 \n\t"
254 "psrlq $24, %%xmm3 \n\t"
255 "psllq $40, %%xmm4 \n\t"
256 "por %%xmm4, %%xmm3 \n\t"
259 "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset),
260 "m"(filterSize), "m"(((uint64_t *) dither)[0])
261 : XMM_CLOBBERS("%xmm0" , "%xmm1" , "%xmm2" , "%xmm3" , "%xmm4" , "%xmm5" , "%xmm7" ,)
262 "%"FF_REG_d, "%"FF_REG_S, "%"FF_REG_c
266 "movq %5, %%xmm3 \n\t"
269 "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset),
270 "m"(filterSize), "m"(((uint64_t *) dither)[0])
271 : XMM_CLOBBERS("%xmm0" , "%xmm1" , "%xmm2" , "%xmm3" , "%xmm4" , "%xmm5" , "%xmm7" ,)
272 "%"FF_REG_d, "%"FF_REG_S, "%"FF_REG_c
278 #endif /* HAVE_INLINE_ASM */
280 #define SCALE_FUNC(filter_n, from_bpc, to_bpc, opt) \
281 void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt( \
282 SwsContext *c, int16_t *data, \
283 int dstW, const uint8_t *src, \
284 const int16_t *filter, \
285 const int32_t *filterPos, int filterSize)
287 #define SCALE_FUNCS(filter_n, opt) \
288 SCALE_FUNC(filter_n, 8, 15, opt); \
289 SCALE_FUNC(filter_n, 9, 15, opt); \
290 SCALE_FUNC(filter_n, 10, 15, opt); \
291 SCALE_FUNC(filter_n, 12, 15, opt); \
292 SCALE_FUNC(filter_n, 14, 15, opt); \
293 SCALE_FUNC(filter_n, 16, 15, opt); \
294 SCALE_FUNC(filter_n, 8, 19, opt); \
295 SCALE_FUNC(filter_n, 9, 19, opt); \
296 SCALE_FUNC(filter_n, 10, 19, opt); \
297 SCALE_FUNC(filter_n, 12, 19, opt); \
298 SCALE_FUNC(filter_n, 14, 19, opt); \
299 SCALE_FUNC(filter_n, 16, 19, opt)
301 #define SCALE_FUNCS_MMX(opt) \
302 SCALE_FUNCS(4, opt); \
303 SCALE_FUNCS(8, opt); \
306 #define SCALE_FUNCS_SSE(opt) \
307 SCALE_FUNCS(4, opt); \
308 SCALE_FUNCS(8, opt); \
309 SCALE_FUNCS(X4, opt); \
313 SCALE_FUNCS_MMX(mmx);
315 SCALE_FUNCS_SSE(sse2);
316 SCALE_FUNCS_SSE(ssse3);
317 SCALE_FUNCS_SSE(sse4);
319 #define VSCALEX_FUNC(size, opt) \
320 void ff_yuv2planeX_ ## size ## _ ## opt(const int16_t *filter, int filterSize, \
321 const int16_t **src, uint8_t *dest, int dstW, \
322 const uint8_t *dither, int offset)
323 #define VSCALEX_FUNCS(opt) \
324 VSCALEX_FUNC(8, opt); \
325 VSCALEX_FUNC(9, opt); \
326 VSCALEX_FUNC(10, opt)
329 VSCALEX_FUNCS(mmxext);
333 VSCALEX_FUNC(16, sse4);
336 #define VSCALE_FUNC(size, opt) \
337 void ff_yuv2plane1_ ## size ## _ ## opt(const int16_t *src, uint8_t *dst, int dstW, \
338 const uint8_t *dither, int offset)
339 #define VSCALE_FUNCS(opt1, opt2) \
340 VSCALE_FUNC(8, opt1); \
341 VSCALE_FUNC(9, opt2); \
342 VSCALE_FUNC(10, opt2); \
343 VSCALE_FUNC(16, opt1)
346 VSCALE_FUNCS(mmx, mmxext);
348 VSCALE_FUNCS(sse2, sse2);
349 VSCALE_FUNC(16, sse4);
350 VSCALE_FUNCS(avx, avx);
352 #define INPUT_Y_FUNC(fmt, opt) \
353 void ff_ ## fmt ## ToY_ ## opt(uint8_t *dst, const uint8_t *src, \
354 const uint8_t *unused1, const uint8_t *unused2, \
355 int w, uint32_t *unused)
356 #define INPUT_UV_FUNC(fmt, opt) \
357 void ff_ ## fmt ## ToUV_ ## opt(uint8_t *dstU, uint8_t *dstV, \
358 const uint8_t *unused0, \
359 const uint8_t *src1, \
360 const uint8_t *src2, \
361 int w, uint32_t *unused)
362 #define INPUT_FUNC(fmt, opt) \
363 INPUT_Y_FUNC(fmt, opt); \
364 INPUT_UV_FUNC(fmt, opt)
365 #define INPUT_FUNCS(opt) \
366 INPUT_FUNC(uyvy, opt); \
367 INPUT_FUNC(yuyv, opt); \
368 INPUT_UV_FUNC(nv12, opt); \
369 INPUT_UV_FUNC(nv21, opt); \
370 INPUT_FUNC(rgba, opt); \
371 INPUT_FUNC(bgra, opt); \
372 INPUT_FUNC(argb, opt); \
373 INPUT_FUNC(abgr, opt); \
374 INPUT_FUNC(rgb24, opt); \
375 INPUT_FUNC(bgr24, opt)
385 #define YUV2NV_DECL(fmt, opt) \
386 void ff_yuv2 ## fmt ## cX_ ## opt(enum AVPixelFormat format, const uint8_t *dither, \
387 const int16_t *filter, int filterSize, \
388 const int16_t **u, const int16_t **v, \
389 uint8_t *dst, int dstWidth)
391 YUV2NV_DECL(nv12, avx2);
392 YUV2NV_DECL(nv21, avx2);
395 av_cold void ff_sws_init_swscale_x86(SwsContext *c)
397 int cpu_flags = av_get_cpu_flags();
400 if (INLINE_MMX(cpu_flags))
401 sws_init_swscale_mmx(c);
403 #if HAVE_MMXEXT_INLINE
404 if (INLINE_MMXEXT(cpu_flags))
405 sws_init_swscale_mmxext(c);
406 if (cpu_flags & AV_CPU_FLAG_SSE3){
407 if(c->use_mmx_vfilter && !(c->flags & SWS_ACCURATE_RND))
408 c->yuv2planeX = yuv2yuvX_sse3;
412 #define ASSIGN_SCALE_FUNC2(hscalefn, filtersize, opt1, opt2) do { \
413 if (c->srcBpc == 8) { \
414 hscalefn = c->dstBpc <= 14 ? ff_hscale8to15_ ## filtersize ## _ ## opt2 : \
415 ff_hscale8to19_ ## filtersize ## _ ## opt1; \
416 } else if (c->srcBpc == 9) { \
417 hscalefn = c->dstBpc <= 14 ? ff_hscale9to15_ ## filtersize ## _ ## opt2 : \
418 ff_hscale9to19_ ## filtersize ## _ ## opt1; \
419 } else if (c->srcBpc == 10) { \
420 hscalefn = c->dstBpc <= 14 ? ff_hscale10to15_ ## filtersize ## _ ## opt2 : \
421 ff_hscale10to19_ ## filtersize ## _ ## opt1; \
422 } else if (c->srcBpc == 12) { \
423 hscalefn = c->dstBpc <= 14 ? ff_hscale12to15_ ## filtersize ## _ ## opt2 : \
424 ff_hscale12to19_ ## filtersize ## _ ## opt1; \
425 } else if (c->srcBpc == 14 || ((c->srcFormat==AV_PIX_FMT_PAL8||isAnyRGB(c->srcFormat)) && av_pix_fmt_desc_get(c->srcFormat)->comp[0].depth<16)) { \
426 hscalefn = c->dstBpc <= 14 ? ff_hscale14to15_ ## filtersize ## _ ## opt2 : \
427 ff_hscale14to19_ ## filtersize ## _ ## opt1; \
428 } else { /* c->srcBpc == 16 */ \
429 av_assert0(c->srcBpc == 16);\
430 hscalefn = c->dstBpc <= 14 ? ff_hscale16to15_ ## filtersize ## _ ## opt2 : \
431 ff_hscale16to19_ ## filtersize ## _ ## opt1; \
434 #define ASSIGN_MMX_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
435 switch (filtersize) { \
436 case 4: ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
437 case 8: ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
438 default: ASSIGN_SCALE_FUNC2(hscalefn, X, opt1, opt2); break; \
440 #define ASSIGN_VSCALEX_FUNC(vscalefn, opt, do_16_case, condition_8bit) \
442 case 16: do_16_case; break; \
443 case 10: if (!isBE(c->dstFormat) && c->dstFormat != AV_PIX_FMT_P010LE) vscalefn = ff_yuv2planeX_10_ ## opt; break; \
444 case 9: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_9_ ## opt; break; \
445 case 8: if ((condition_8bit) && !c->use_mmx_vfilter) vscalefn = ff_yuv2planeX_8_ ## opt; break; \
447 #define ASSIGN_VSCALE_FUNC(vscalefn, opt1, opt2, opt2chk) \
449 case 16: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2plane1_16_ ## opt1; break; \
450 case 10: if (!isBE(c->dstFormat) && c->dstFormat != AV_PIX_FMT_P010LE && opt2chk) vscalefn = ff_yuv2plane1_10_ ## opt2; break; \
451 case 9: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_9_ ## opt2; break; \
452 case 8: vscalefn = ff_yuv2plane1_8_ ## opt1; break; \
453 default: av_assert0(c->dstBpc>8); \
455 #define case_rgb(x, X, opt) \
456 case AV_PIX_FMT_ ## X: \
457 c->lumToYV12 = ff_ ## x ## ToY_ ## opt; \
458 if (!c->chrSrcHSubSample) \
459 c->chrToYV12 = ff_ ## x ## ToUV_ ## opt; \
462 if (EXTERNAL_MMX(cpu_flags)) {
463 ASSIGN_MMX_SCALE_FUNC(c->hyScale, c->hLumFilterSize, mmx, mmx);
464 ASSIGN_MMX_SCALE_FUNC(c->hcScale, c->hChrFilterSize, mmx, mmx);
465 ASSIGN_VSCALE_FUNC(c->yuv2plane1, mmx, mmxext, cpu_flags & AV_CPU_FLAG_MMXEXT);
467 switch (c->srcFormat) {
469 c->lumToYV12 = ff_yuyvToY_mmx;
471 c->alpToYV12 = ff_uyvyToY_mmx;
473 case AV_PIX_FMT_YUYV422:
474 c->lumToYV12 = ff_yuyvToY_mmx;
475 c->chrToYV12 = ff_yuyvToUV_mmx;
477 case AV_PIX_FMT_UYVY422:
478 c->lumToYV12 = ff_uyvyToY_mmx;
479 c->chrToYV12 = ff_uyvyToUV_mmx;
481 case AV_PIX_FMT_NV12:
482 c->chrToYV12 = ff_nv12ToUV_mmx;
484 case AV_PIX_FMT_NV21:
485 c->chrToYV12 = ff_nv21ToUV_mmx;
487 case_rgb(rgb24, RGB24, mmx);
488 case_rgb(bgr24, BGR24, mmx);
489 case_rgb(bgra, BGRA, mmx);
490 case_rgb(rgba, RGBA, mmx);
491 case_rgb(abgr, ABGR, mmx);
492 case_rgb(argb, ARGB, mmx);
497 if (EXTERNAL_MMXEXT(cpu_flags)) {
498 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, mmxext, , 1);
500 #endif /* ARCH_X86_32 */
501 #define ASSIGN_SSE_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
502 switch (filtersize) { \
503 case 4: ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
504 case 8: ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
505 default: if (filtersize & 4) ASSIGN_SCALE_FUNC2(hscalefn, X4, opt1, opt2); \
506 else ASSIGN_SCALE_FUNC2(hscalefn, X8, opt1, opt2); \
509 if (EXTERNAL_SSE2(cpu_flags)) {
510 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse2, sse2);
511 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse2, sse2);
512 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse2, ,
513 HAVE_ALIGNED_STACK || ARCH_X86_64);
514 ASSIGN_VSCALE_FUNC(c->yuv2plane1, sse2, sse2, 1);
516 switch (c->srcFormat) {
518 c->lumToYV12 = ff_yuyvToY_sse2;
520 c->alpToYV12 = ff_uyvyToY_sse2;
522 case AV_PIX_FMT_YUYV422:
523 c->lumToYV12 = ff_yuyvToY_sse2;
524 c->chrToYV12 = ff_yuyvToUV_sse2;
526 case AV_PIX_FMT_UYVY422:
527 c->lumToYV12 = ff_uyvyToY_sse2;
528 c->chrToYV12 = ff_uyvyToUV_sse2;
530 case AV_PIX_FMT_NV12:
531 c->chrToYV12 = ff_nv12ToUV_sse2;
533 case AV_PIX_FMT_NV21:
534 c->chrToYV12 = ff_nv21ToUV_sse2;
536 case_rgb(rgb24, RGB24, sse2);
537 case_rgb(bgr24, BGR24, sse2);
538 case_rgb(bgra, BGRA, sse2);
539 case_rgb(rgba, RGBA, sse2);
540 case_rgb(abgr, ABGR, sse2);
541 case_rgb(argb, ARGB, sse2);
546 if (EXTERNAL_SSSE3(cpu_flags)) {
547 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, ssse3, ssse3);
548 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, ssse3, ssse3);
549 switch (c->srcFormat) {
550 case_rgb(rgb24, RGB24, ssse3);
551 case_rgb(bgr24, BGR24, ssse3);
556 if (EXTERNAL_SSE4(cpu_flags)) {
557 /* Xto15 don't need special sse4 functions */
558 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse4, ssse3);
559 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse4, ssse3);
560 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse4,
561 if (!isBE(c->dstFormat)) c->yuv2planeX = ff_yuv2planeX_16_sse4,
562 HAVE_ALIGNED_STACK || ARCH_X86_64);
563 if (c->dstBpc == 16 && !isBE(c->dstFormat))
564 c->yuv2plane1 = ff_yuv2plane1_16_sse4;
567 if (EXTERNAL_AVX(cpu_flags)) {
568 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, avx, ,
569 HAVE_ALIGNED_STACK || ARCH_X86_64);
570 ASSIGN_VSCALE_FUNC(c->yuv2plane1, avx, avx, 1);
572 switch (c->srcFormat) {
573 case AV_PIX_FMT_YUYV422:
574 c->chrToYV12 = ff_yuyvToUV_avx;
576 case AV_PIX_FMT_UYVY422:
577 c->chrToYV12 = ff_uyvyToUV_avx;
579 case AV_PIX_FMT_NV12:
580 c->chrToYV12 = ff_nv12ToUV_avx;
582 case AV_PIX_FMT_NV21:
583 c->chrToYV12 = ff_nv21ToUV_avx;
585 case_rgb(rgb24, RGB24, avx);
586 case_rgb(bgr24, BGR24, avx);
587 case_rgb(bgra, BGRA, avx);
588 case_rgb(rgba, RGBA, avx);
589 case_rgb(abgr, ABGR, avx);
590 case_rgb(argb, ARGB, avx);
597 if (EXTERNAL_AVX2_FAST(cpu_flags)) {
598 switch (c->dstFormat) {
599 case AV_PIX_FMT_NV12:
600 case AV_PIX_FMT_NV24:
601 c->yuv2nv12cX = ff_yuv2nv12cX_avx2;
603 case AV_PIX_FMT_NV21:
604 case AV_PIX_FMT_NV42:
605 c->yuv2nv12cX = ff_yuv2nv21cX_avx2;