falkTX
/
FFmpeg
mirror of https://github.com/falkTX/FFmpeg.git
1
0
Fork 0
Code Issues 0 Releases 338 Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
80254 Commits
32 Branches
748MB
Tree: 218bb8b3f3
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '218bb8b3f3'
${ noResults }
FFmpeg/libswscale/x86/swscale.c

586 lines
23KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>

  3.  *

  4.  * This file is part of FFmpeg.

  5.  *

  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.

  10.  *

  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.

  15.  *

  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

  19.  */

  20. 

  21. #include <inttypes.h>

  22. #include "config.h"

  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/asm.h"

  29. #include "libavutil/x86/cpu.h"

  30. #include "libavutil/cpu.h"

  31. #include "libavutil/pixdesc.h"

  32. 

  33. #if HAVE_INLINE_ASM

  34. 

  35. #define DITHER1XBPP

  36. 

  37. DECLARE_ASM_CONST(8, uint64_t, bF8)=       0xF8F8F8F8F8F8F8F8LL;

  38. DECLARE_ASM_CONST(8, uint64_t, bFC)=       0xFCFCFCFCFCFCFCFCLL;

  39. DECLARE_ASM_CONST(8, uint64_t, w10)=       0x0010001000100010LL;

  40. DECLARE_ASM_CONST(8, uint64_t, w02)=       0x0002000200020002LL;

  41. 

  42. const DECLARE_ALIGNED(8, uint64_t, ff_dither4)[2] = {

  43.     0x0103010301030103LL,

  44.     0x0200020002000200LL,};

  45. 

  46. const DECLARE_ALIGNED(8, uint64_t, ff_dither8)[2] = {

  47.     0x0602060206020602LL,

  48.     0x0004000400040004LL,};

  49. 

  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;

  56. 

  57. DECLARE_ALIGNED(8, const uint64_t, ff_M24A)         = 0x00FF0000FF0000FFLL;

  58. DECLARE_ALIGNED(8, const uint64_t, ff_M24B)         = 0xFF0000FF0000FF00LL;

  59. DECLARE_ALIGNED(8, const uint64_t, ff_M24C)         = 0x0000FF0000FF0000LL;

  60. 

  61. DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YOffset)  = 0x1010101010101010ULL;

  62. DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UVOffset) = 0x8080808080808080ULL;

  63. DECLARE_ALIGNED(8, const uint64_t, ff_w1111)        = 0x0001000100010001ULL;

  64. 

  65. 

  66. //MMX versions

  67. #if HAVE_MMX_INLINE

  68. #undef RENAME

  69. #define COMPILE_TEMPLATE_MMXEXT 0

  70. #define RENAME(a) a ## _mmx

  71. #include "swscale_template.c"

  72. #endif

  73. 

  74. // MMXEXT versions

  75. #if HAVE_MMXEXT_INLINE

  76. #undef RENAME

  77. #undef COMPILE_TEMPLATE_MMXEXT

  78. #define COMPILE_TEMPLATE_MMXEXT 1

  79. #define RENAME(a) a ## _mmxext

  80. #include "swscale_template.c"

  81. #endif

  82. 

  83. void ff_updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,

  84.                            int lastInLumBuf, int lastInChrBuf)

  85. {

  86.     const int dstH= c->dstH;

  87.     const int flags= c->flags;

  88. 

  89.     SwsPlane *lumPlane = &c->slice[c->numSlice-2].plane[0];

  90.     SwsPlane *chrUPlane = &c->slice[c->numSlice-2].plane[1];

  91.     SwsPlane *alpPlane = &c->slice[c->numSlice-2].plane[3];

  92. 

  93.     int hasAlpha = c->needAlpha;

  94.     int32_t *vLumFilterPos= c->vLumFilterPos;

  95.     int32_t *vChrFilterPos= c->vChrFilterPos;

  96.     int16_t *vLumFilter= c->vLumFilter;

  97.     int16_t *vChrFilter= c->vChrFilter;

  98.     int32_t *lumMmxFilter= c->lumMmxFilter;

  99.     int32_t *chrMmxFilter= c->chrMmxFilter;

  100.     int32_t av_unused *alpMmxFilter= c->alpMmxFilter;

  101.     const int vLumFilterSize= c->vLumFilterSize;

  102.     const int vChrFilterSize= c->vChrFilterSize;

  103.     const int chrDstY= dstY>>c->chrDstVSubSample;

  104.     const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input

  105.     const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input

  106. 

  107.     c->blueDither= ff_dither8[dstY&1];

  108.     if (c->dstFormat == AV_PIX_FMT_RGB555 || c->dstFormat == AV_PIX_FMT_BGR555)

  109.         c->greenDither= ff_dither8[dstY&1];

  110.     else

  111.         c->greenDither= ff_dither4[dstY&1];

  112.     c->redDither= ff_dither8[(dstY+1)&1];

  113.     if (dstY < dstH - 2) {

  114.         const int16_t **lumSrcPtr  = (const int16_t **)(void*) lumPlane->line + firstLumSrcY - lumPlane->sliceY;

  115.         const int16_t **chrUSrcPtr = (const int16_t **)(void*) chrUPlane->line + firstChrSrcY - chrUPlane->sliceY;

  116.         const int16_t **alpSrcPtr  = (CONFIG_SWSCALE_ALPHA && hasAlpha) ? (const int16_t **)(void*) alpPlane->line + firstLumSrcY - alpPlane->sliceY : NULL;

  117. 

  118.         int i;

  119.         if (firstLumSrcY < 0 || firstLumSrcY + vLumFilterSize > c->srcH) {

  120.             const int16_t **tmpY = (const int16_t **) lumPlane->tmp;

  121. 

  122.             int neg = -firstLumSrcY, i, end = FFMIN(c->srcH - firstLumSrcY, vLumFilterSize);

  123.             for (i = 0; i < neg;            i++)

  124.                 tmpY[i] = lumSrcPtr[neg];

  125.             for (     ; i < end;            i++)

  126.                 tmpY[i] = lumSrcPtr[i];

  127.             for (     ; i < vLumFilterSize; i++)

  128.                 tmpY[i] = tmpY[i-1];

  129.             lumSrcPtr = tmpY;

  130. 

  131.             if (alpSrcPtr) {

  132.                 const int16_t **tmpA = (const int16_t **) alpPlane->tmp;

  133.                 for (i = 0; i < neg;            i++)

  134.                     tmpA[i] = alpSrcPtr[neg];

  135.                 for (     ; i < end;            i++)

  136.                     tmpA[i] = alpSrcPtr[i];

  137.                 for (     ; i < vLumFilterSize; i++)

  138.                     tmpA[i] = tmpA[i - 1];

  139.                 alpSrcPtr = tmpA;

  140.             }

  141.         }

  142.         if (firstChrSrcY < 0 || firstChrSrcY + vChrFilterSize > c->chrSrcH) {

  143.             const int16_t **tmpU = (const int16_t **) chrUPlane->tmp;

  144.             int neg = -firstChrSrcY, i, end = FFMIN(c->chrSrcH - firstChrSrcY, vChrFilterSize);

  145.             for (i = 0; i < neg;            i++) {

  146.                 tmpU[i] = chrUSrcPtr[neg];

  147.             }

  148.             for (     ; i < end;            i++) {

  149.                 tmpU[i] = chrUSrcPtr[i];

  150.             }

  151.             for (     ; i < vChrFilterSize; i++) {

  152.                 tmpU[i] = tmpU[i - 1];

  153.             }

  154.             chrUSrcPtr = tmpU;

  155.         }

  156. 

  157.         if (flags & SWS_ACCURATE_RND) {

  158.             int s= APCK_SIZE / 8;

  159.             for (i=0; i<vLumFilterSize; i+=2) {

  160.                 *(const void**)&lumMmxFilter[s*i              ]= lumSrcPtr[i  ];

  161.                 *(const void**)&lumMmxFilter[s*i+APCK_PTR2/4  ]= lumSrcPtr[i+(vLumFilterSize>1)];

  162.                 lumMmxFilter[s*i+APCK_COEF/4  ]=

  163.                 lumMmxFilter[s*i+APCK_COEF/4+1]= vLumFilter[dstY*vLumFilterSize + i    ]

  164.                 + (vLumFilterSize>1 ? vLumFilter[dstY*vLumFilterSize + i + 1]<<16 : 0);

  165.                 if (CONFIG_SWSCALE_ALPHA && hasAlpha) {

  166.                     *(const void**)&alpMmxFilter[s*i              ]= alpSrcPtr[i  ];

  167.                     *(const void**)&alpMmxFilter[s*i+APCK_PTR2/4  ]= alpSrcPtr[i+(vLumFilterSize>1)];

  168.                     alpMmxFilter[s*i+APCK_COEF/4  ]=

  169.                     alpMmxFilter[s*i+APCK_COEF/4+1]= lumMmxFilter[s*i+APCK_COEF/4  ];

  170.                 }

  171.             }

  172.             for (i=0; i<vChrFilterSize; i+=2) {

  173.                 *(const void**)&chrMmxFilter[s*i              ]= chrUSrcPtr[i  ];

  174.                 *(const void**)&chrMmxFilter[s*i+APCK_PTR2/4  ]= chrUSrcPtr[i+(vChrFilterSize>1)];

  175.                 chrMmxFilter[s*i+APCK_COEF/4  ]=

  176.                 chrMmxFilter[s*i+APCK_COEF/4+1]= vChrFilter[chrDstY*vChrFilterSize + i    ]

  177.                 + (vChrFilterSize>1 ? vChrFilter[chrDstY*vChrFilterSize + i + 1]<<16 : 0);

  178.             }

  179.         } else {

  180.             for (i=0; i<vLumFilterSize; i++) {

  181.                 *(const void**)&lumMmxFilter[4*i+0]= lumSrcPtr[i];

  182.                 lumMmxFilter[4*i+2]=

  183.                 lumMmxFilter[4*i+3]=

  184.                 ((uint16_t)vLumFilter[dstY*vLumFilterSize + i])*0x10001U;

  185.                 if (CONFIG_SWSCALE_ALPHA && hasAlpha) {

  186.                     *(const void**)&alpMmxFilter[4*i+0]= alpSrcPtr[i];

  187.                     alpMmxFilter[4*i+2]=

  188.                     alpMmxFilter[4*i+3]= lumMmxFilter[4*i+2];

  189.                 }

  190.             }

  191.             for (i=0; i<vChrFilterSize; i++) {

  192.                 *(const void**)&chrMmxFilter[4*i+0]= chrUSrcPtr[i];

  193.                 chrMmxFilter[4*i+2]=

  194.                 chrMmxFilter[4*i+3]=

  195.                 ((uint16_t)vChrFilter[chrDstY*vChrFilterSize + i])*0x10001U;

  196.             }

  197.         }

  198.     }

  199. }

  200. 

  201. #if HAVE_MMXEXT

  202. static void yuv2yuvX_sse3(const int16_t *filter, int filterSize,

  203.                            const int16_t **src, uint8_t *dest, int dstW,

  204.                            const uint8_t *dither, int offset)

  205. {

  206.     if(((uintptr_t)dest) & 15){

  207.         yuv2yuvX_mmxext(filter, filterSize, src, dest, dstW, dither, offset);

  208.         return;

  209.     }

  210.     filterSize--;

  211. #define MAIN_FUNCTION \

  212.         "pxor       %%xmm0, %%xmm0 \n\t" \

  213.         "punpcklbw  %%xmm0, %%xmm3 \n\t" \

  214.         "movd           %4, %%xmm1 \n\t" \

  215.         "punpcklwd  %%xmm1, %%xmm1 \n\t" \

  216.         "punpckldq  %%xmm1, %%xmm1 \n\t" \

  217.         "punpcklqdq %%xmm1, %%xmm1 \n\t" \

  218.         "psllw          $3, %%xmm1 \n\t" \

  219.         "paddw      %%xmm1, %%xmm3 \n\t" \

  220.         "psraw          $4, %%xmm3 \n\t" \

  221.         "movdqa     %%xmm3, %%xmm4 \n\t" \

  222.         "movdqa     %%xmm3, %%xmm7 \n\t" \

  223.         "movl           %3, %%ecx  \n\t" \

  224.         "mov                                 %0, %%"REG_d"  \n\t"\

  225.         "mov                        (%%"REG_d"), %%"REG_S"  \n\t"\

  226.         ".p2align                             4             \n\t" /* FIXME Unroll? */\

  227.         "1:                                                 \n\t"\

  228.         "movddup                  8(%%"REG_d"), %%xmm0      \n\t" /* filterCoeff */\

  229.         "movdqa              (%%"REG_S", %%"REG_c", 2), %%xmm2      \n\t" /* srcData */\

  230.         "movdqa            16(%%"REG_S", %%"REG_c", 2), %%xmm5      \n\t" /* srcData */\

  231.         "add                                $16, %%"REG_d"  \n\t"\

  232.         "mov                        (%%"REG_d"), %%"REG_S"  \n\t"\

  233.         "test                         %%"REG_S", %%"REG_S"  \n\t"\

  234.         "pmulhw                           %%xmm0, %%xmm2      \n\t"\

  235.         "pmulhw                           %%xmm0, %%xmm5      \n\t"\

  236.         "paddw                            %%xmm2, %%xmm3      \n\t"\

  237.         "paddw                            %%xmm5, %%xmm4      \n\t"\

  238.         " jnz                                1b             \n\t"\

  239.         "psraw                               $3, %%xmm3      \n\t"\

  240.         "psraw                               $3, %%xmm4      \n\t"\

  241.         "packuswb                         %%xmm4, %%xmm3      \n\t"\

  242.         "movntdq                          %%xmm3, (%1, %%"REG_c")\n\t"\

  243.         "add                         $16, %%"REG_c"         \n\t"\

  244.         "cmp                          %2, %%"REG_c"         \n\t"\

  245.         "movdqa                   %%xmm7, %%xmm3            \n\t" \

  246.         "movdqa                   %%xmm7, %%xmm4            \n\t" \

  247.         "mov                                 %0, %%"REG_d"  \n\t"\

  248.         "mov                        (%%"REG_d"), %%"REG_S"  \n\t"\

  249.         "jb                                  1b             \n\t"

  250. 

  251.     if (offset) {

  252.         __asm__ volatile(

  253.             "movq          %5, %%xmm3  \n\t"

  254.             "movdqa    %%xmm3, %%xmm4  \n\t"

  255.             "psrlq        $24, %%xmm3  \n\t"

  256.             "psllq        $40, %%xmm4  \n\t"

  257.             "por       %%xmm4, %%xmm3  \n\t"

  258.             MAIN_FUNCTION

  259.               :: "g" (filter),

  260.               "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset),

  261.               "m"(filterSize), "m"(((uint64_t *) dither)[0])

  262.               : XMM_CLOBBERS("%xmm0" , "%xmm1" , "%xmm2" , "%xmm3" , "%xmm4" , "%xmm5" , "%xmm7" ,)

  263.                 "%"REG_d, "%"REG_S, "%"REG_c

  264.               );

  265.     } else {

  266.         __asm__ volatile(

  267.             "movq          %5, %%xmm3   \n\t"

  268.             MAIN_FUNCTION

  269.               :: "g" (filter),

  270.               "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset),

  271.               "m"(filterSize), "m"(((uint64_t *) dither)[0])

  272.               : XMM_CLOBBERS("%xmm0" , "%xmm1" , "%xmm2" , "%xmm3" , "%xmm4" , "%xmm5" , "%xmm7" ,)

  273.                 "%"REG_d, "%"REG_S, "%"REG_c

  274.               );

  275.     }

  276. }

  277. #endif

  278. 

  279. #endif /* HAVE_INLINE_ASM */

  280. 

  281. #define SCALE_FUNC(filter_n, from_bpc, to_bpc, opt) \

  282. void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt( \

  283.                                                 SwsContext *c, int16_t *data, \

  284.                                                 int dstW, const uint8_t *src, \

  285.                                                 const int16_t *filter, \

  286.                                                 const int32_t *filterPos, int filterSize)

  287. 

  288. #define SCALE_FUNCS(filter_n, opt) \

  289.     SCALE_FUNC(filter_n,  8, 15, opt); \

  290.     SCALE_FUNC(filter_n,  9, 15, opt); \

  291.     SCALE_FUNC(filter_n, 10, 15, opt); \

  292.     SCALE_FUNC(filter_n, 12, 15, opt); \

  293.     SCALE_FUNC(filter_n, 14, 15, opt); \

  294.     SCALE_FUNC(filter_n, 16, 15, opt); \

  295.     SCALE_FUNC(filter_n,  8, 19, opt); \

  296.     SCALE_FUNC(filter_n,  9, 19, opt); \

  297.     SCALE_FUNC(filter_n, 10, 19, opt); \

  298.     SCALE_FUNC(filter_n, 12, 19, opt); \

  299.     SCALE_FUNC(filter_n, 14, 19, opt); \

  300.     SCALE_FUNC(filter_n, 16, 19, opt)

  301. 

  302. #define SCALE_FUNCS_MMX(opt) \

  303.     SCALE_FUNCS(4, opt); \

  304.     SCALE_FUNCS(8, opt); \

  305.     SCALE_FUNCS(X, opt)

  306. 

  307. #define SCALE_FUNCS_SSE(opt) \

  308.     SCALE_FUNCS(4, opt); \

  309.     SCALE_FUNCS(8, opt); \

  310.     SCALE_FUNCS(X4, opt); \

  311.     SCALE_FUNCS(X8, opt)

  312. 

  313. #if ARCH_X86_32

  314. SCALE_FUNCS_MMX(mmx);

  315. #endif

  316. SCALE_FUNCS_SSE(sse2);

  317. SCALE_FUNCS_SSE(ssse3);

  318. SCALE_FUNCS_SSE(sse4);

  319. 

  320. #define VSCALEX_FUNC(size, opt) \

  321. void ff_yuv2planeX_ ## size ## _ ## opt(const int16_t *filter, int filterSize, \

  322.                                         const int16_t **src, uint8_t *dest, int dstW, \

  323.                                         const uint8_t *dither, int offset)

  324. #define VSCALEX_FUNCS(opt) \

  325.     VSCALEX_FUNC(8,  opt); \

  326.     VSCALEX_FUNC(9,  opt); \

  327.     VSCALEX_FUNC(10, opt)

  328. 

  329. #if ARCH_X86_32

  330. VSCALEX_FUNCS(mmxext);

  331. #endif

  332. VSCALEX_FUNCS(sse2);

  333. VSCALEX_FUNCS(sse4);

  334. VSCALEX_FUNC(16, sse4);

  335. VSCALEX_FUNCS(avx);

  336. 

  337. #define VSCALE_FUNC(size, opt) \

  338. void ff_yuv2plane1_ ## size ## _ ## opt(const int16_t *src, uint8_t *dst, int dstW, \

  339.                                         const uint8_t *dither, int offset)

  340. #define VSCALE_FUNCS(opt1, opt2) \

  341.     VSCALE_FUNC(8,  opt1); \

  342.     VSCALE_FUNC(9,  opt2); \

  343.     VSCALE_FUNC(10, opt2); \

  344.     VSCALE_FUNC(16, opt1)

  345. 

  346. #if ARCH_X86_32

  347. VSCALE_FUNCS(mmx, mmxext);

  348. #endif

  349. VSCALE_FUNCS(sse2, sse2);

  350. VSCALE_FUNC(16, sse4);

  351. VSCALE_FUNCS(avx, avx);

  352. 

  353. #define INPUT_Y_FUNC(fmt, opt) \

  354. void ff_ ## fmt ## ToY_  ## opt(uint8_t *dst, const uint8_t *src, \

  355.                                 const uint8_t *unused1, const uint8_t *unused2, \

  356.                                 int w, uint32_t *unused)

  357. #define INPUT_UV_FUNC(fmt, opt) \

  358. void ff_ ## fmt ## ToUV_ ## opt(uint8_t *dstU, uint8_t *dstV, \

  359.                                 const uint8_t *unused0, \

  360.                                 const uint8_t *src1, \

  361.                                 const uint8_t *src2, \

  362.                                 int w, uint32_t *unused)

  363. #define INPUT_FUNC(fmt, opt) \

  364.     INPUT_Y_FUNC(fmt, opt); \

  365.     INPUT_UV_FUNC(fmt, opt)

  366. #define INPUT_FUNCS(opt) \

  367.     INPUT_FUNC(uyvy, opt); \

  368.     INPUT_FUNC(yuyv, opt); \

  369.     INPUT_UV_FUNC(nv12, opt); \

  370.     INPUT_UV_FUNC(nv21, opt); \

  371.     INPUT_FUNC(rgba, opt); \

  372.     INPUT_FUNC(bgra, opt); \

  373.     INPUT_FUNC(argb, opt); \

  374.     INPUT_FUNC(abgr, opt); \

  375.     INPUT_FUNC(rgb24, opt); \

  376.     INPUT_FUNC(bgr24, opt)

  377. 

  378. #if ARCH_X86_32

  379. INPUT_FUNCS(mmx);

  380. #endif

  381. INPUT_FUNCS(sse2);

  382. INPUT_FUNCS(ssse3);

  383. INPUT_FUNCS(avx);

  384. 

  385. av_cold void ff_sws_init_swscale_x86(SwsContext *c)

  386. {

  387.     int cpu_flags = av_get_cpu_flags();

  388. 

  389. #if HAVE_MMX_INLINE

  390.     if (INLINE_MMX(cpu_flags))

  391.         sws_init_swscale_mmx(c);

  392. #endif

  393. #if HAVE_MMXEXT_INLINE

  394.     if (INLINE_MMXEXT(cpu_flags))

  395.         sws_init_swscale_mmxext(c);

  396.     if (cpu_flags & AV_CPU_FLAG_SSE3){

  397.         if(c->use_mmx_vfilter && !(c->flags & SWS_ACCURATE_RND))

  398.             c->yuv2planeX = yuv2yuvX_sse3;

  399.     }

  400. #endif

  401. 

  402. #define ASSIGN_SCALE_FUNC2(hscalefn, filtersize, opt1, opt2) do { \

  403.     if (c->srcBpc == 8) { \

  404.         hscalefn = c->dstBpc <= 14 ? ff_hscale8to15_ ## filtersize ## _ ## opt2 : \

  405.                                      ff_hscale8to19_ ## filtersize ## _ ## opt1; \

  406.     } else if (c->srcBpc == 9) { \

  407.         hscalefn = c->dstBpc <= 14 ? ff_hscale9to15_ ## filtersize ## _ ## opt2 : \

  408.                                      ff_hscale9to19_ ## filtersize ## _ ## opt1; \

  409.     } else if (c->srcBpc == 10) { \

  410.         hscalefn = c->dstBpc <= 14 ? ff_hscale10to15_ ## filtersize ## _ ## opt2 : \

  411.                                      ff_hscale10to19_ ## filtersize ## _ ## opt1; \

  412.     } else if (c->srcBpc == 12) { \

  413.         hscalefn = c->dstBpc <= 14 ? ff_hscale12to15_ ## filtersize ## _ ## opt2 : \

  414.                                      ff_hscale12to19_ ## filtersize ## _ ## opt1; \

  415.     } 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)) { \

  416.         hscalefn = c->dstBpc <= 14 ? ff_hscale14to15_ ## filtersize ## _ ## opt2 : \

  417.                                      ff_hscale14to19_ ## filtersize ## _ ## opt1; \

  418.     } else { /* c->srcBpc == 16 */ \

  419.         av_assert0(c->srcBpc == 16);\

  420.         hscalefn = c->dstBpc <= 14 ? ff_hscale16to15_ ## filtersize ## _ ## opt2 : \

  421.                                      ff_hscale16to19_ ## filtersize ## _ ## opt1; \

  422.     } \

  423. } while (0)

  424. #define ASSIGN_MMX_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \

  425.     switch (filtersize) { \

  426.     case 4:  ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \

  427.     case 8:  ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \

  428.     default: ASSIGN_SCALE_FUNC2(hscalefn, X, opt1, opt2); break; \

  429.     }

  430. #define ASSIGN_VSCALEX_FUNC(vscalefn, opt, do_16_case, condition_8bit) \

  431. switch(c->dstBpc){ \

  432.     case 16:                          do_16_case;                          break; \

  433.     case 10: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_10_ ## opt; break; \

  434.     case 9:  if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_9_  ## opt; break; \

  435.     case 8: if ((condition_8bit) && !c->use_mmx_vfilter) vscalefn = ff_yuv2planeX_8_  ## opt; break; \

  436.     }

  437. #define ASSIGN_VSCALE_FUNC(vscalefn, opt1, opt2, opt2chk) \

  438.     switch(c->dstBpc){ \

  439.     case 16: if (!isBE(c->dstFormat))            vscalefn = ff_yuv2plane1_16_ ## opt1; break; \

  440.     case 10: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_10_ ## opt2; break; \

  441.     case 9:  if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_9_  ## opt2;  break; \

  442.     case 8:                                      vscalefn = ff_yuv2plane1_8_  ## opt1;  break; \

  443.     default: av_assert0(c->dstBpc>8); \

  444.     }

  445. #define case_rgb(x, X, opt) \

  446.         case AV_PIX_FMT_ ## X: \

  447.             c->lumToYV12 = ff_ ## x ## ToY_ ## opt; \

  448.             if (!c->chrSrcHSubSample) \

  449.                 c->chrToYV12 = ff_ ## x ## ToUV_ ## opt; \

  450.             break

  451. #if ARCH_X86_32

  452.     if (EXTERNAL_MMX(cpu_flags)) {

  453.         ASSIGN_MMX_SCALE_FUNC(c->hyScale, c->hLumFilterSize, mmx, mmx);

  454.         ASSIGN_MMX_SCALE_FUNC(c->hcScale, c->hChrFilterSize, mmx, mmx);

  455.         ASSIGN_VSCALE_FUNC(c->yuv2plane1, mmx, mmxext, cpu_flags & AV_CPU_FLAG_MMXEXT);

  456. 

  457.         switch (c->srcFormat) {

  458.         case AV_PIX_FMT_YA8:

  459.             c->lumToYV12 = ff_yuyvToY_mmx;

  460.             if (c->needAlpha)

  461.                 c->alpToYV12 = ff_uyvyToY_mmx;

  462.             break;

  463.         case AV_PIX_FMT_YUYV422:

  464.             c->lumToYV12 = ff_yuyvToY_mmx;

  465.             c->chrToYV12 = ff_yuyvToUV_mmx;

  466.             break;

  467.         case AV_PIX_FMT_UYVY422:

  468.             c->lumToYV12 = ff_uyvyToY_mmx;

  469.             c->chrToYV12 = ff_uyvyToUV_mmx;

  470.             break;

  471.         case AV_PIX_FMT_NV12:

  472.             c->chrToYV12 = ff_nv12ToUV_mmx;

  473.             break;

  474.         case AV_PIX_FMT_NV21:

  475.             c->chrToYV12 = ff_nv21ToUV_mmx;

  476.             break;

  477.         case_rgb(rgb24, RGB24, mmx);

  478.         case_rgb(bgr24, BGR24, mmx);

  479.         case_rgb(bgra,  BGRA,  mmx);

  480.         case_rgb(rgba,  RGBA,  mmx);

  481.         case_rgb(abgr,  ABGR,  mmx);

  482.         case_rgb(argb,  ARGB,  mmx);

  483.         default:

  484.             break;

  485.         }

  486.     }

  487.     if (EXTERNAL_MMXEXT(cpu_flags)) {

  488.         ASSIGN_VSCALEX_FUNC(c->yuv2planeX, mmxext, , 1);

  489.     }

  490. #endif /* ARCH_X86_32 */

  491. #define ASSIGN_SSE_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \

  492.     switch (filtersize) { \

  493.     case 4:  ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \

  494.     case 8:  ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \

  495.     default: if (filtersize & 4) ASSIGN_SCALE_FUNC2(hscalefn, X4, opt1, opt2); \

  496.              else                ASSIGN_SCALE_FUNC2(hscalefn, X8, opt1, opt2); \

  497.              break; \

  498.     }

  499.     if (EXTERNAL_SSE2(cpu_flags)) {

  500.         ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse2, sse2);

  501.         ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse2, sse2);

  502.         ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse2, ,

  503.                             HAVE_ALIGNED_STACK || ARCH_X86_64);

  504.         ASSIGN_VSCALE_FUNC(c->yuv2plane1, sse2, sse2, 1);

  505. 

  506.         switch (c->srcFormat) {

  507.         case AV_PIX_FMT_YA8:

  508.             c->lumToYV12 = ff_yuyvToY_sse2;

  509.             if (c->needAlpha)

  510.                 c->alpToYV12 = ff_uyvyToY_sse2;

  511.             break;

  512.         case AV_PIX_FMT_YUYV422:

  513.             c->lumToYV12 = ff_yuyvToY_sse2;

  514.             c->chrToYV12 = ff_yuyvToUV_sse2;

  515.             break;

  516.         case AV_PIX_FMT_UYVY422:

  517.             c->lumToYV12 = ff_uyvyToY_sse2;

  518.             c->chrToYV12 = ff_uyvyToUV_sse2;

  519.             break;

  520.         case AV_PIX_FMT_NV12:

  521.             c->chrToYV12 = ff_nv12ToUV_sse2;

  522.             break;

  523.         case AV_PIX_FMT_NV21:

  524.             c->chrToYV12 = ff_nv21ToUV_sse2;

  525.             break;

  526.         case_rgb(rgb24, RGB24, sse2);

  527.         case_rgb(bgr24, BGR24, sse2);

  528.         case_rgb(bgra,  BGRA,  sse2);

  529.         case_rgb(rgba,  RGBA,  sse2);

  530.         case_rgb(abgr,  ABGR,  sse2);

  531.         case_rgb(argb,  ARGB,  sse2);

  532.         default:

  533.             break;

  534.         }

  535.     }

  536.     if (EXTERNAL_SSSE3(cpu_flags)) {

  537.         ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, ssse3, ssse3);

  538.         ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, ssse3, ssse3);

  539.         switch (c->srcFormat) {

  540.         case_rgb(rgb24, RGB24, ssse3);

  541.         case_rgb(bgr24, BGR24, ssse3);

  542.         default:

  543.             break;

  544.         }

  545.     }

  546.     if (EXTERNAL_SSE4(cpu_flags)) {

  547.         /* Xto15 don't need special sse4 functions */

  548.         ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse4, ssse3);

  549.         ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse4, ssse3);

  550.         ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse4,

  551.                             if (!isBE(c->dstFormat)) c->yuv2planeX = ff_yuv2planeX_16_sse4,

  552.                             HAVE_ALIGNED_STACK || ARCH_X86_64);

  553.         if (c->dstBpc == 16 && !isBE(c->dstFormat))

  554.             c->yuv2plane1 = ff_yuv2plane1_16_sse4;

  555.     }

  556. 

  557.     if (EXTERNAL_AVX(cpu_flags)) {

  558.         ASSIGN_VSCALEX_FUNC(c->yuv2planeX, avx, ,

  559.                             HAVE_ALIGNED_STACK || ARCH_X86_64);

  560.         ASSIGN_VSCALE_FUNC(c->yuv2plane1, avx, avx, 1);

  561. 

  562.         switch (c->srcFormat) {

  563.         case AV_PIX_FMT_YUYV422:

  564.             c->chrToYV12 = ff_yuyvToUV_avx;

  565.             break;

  566.         case AV_PIX_FMT_UYVY422:

  567.             c->chrToYV12 = ff_uyvyToUV_avx;

  568.             break;

  569.         case AV_PIX_FMT_NV12:

  570.             c->chrToYV12 = ff_nv12ToUV_avx;

  571.             break;

  572.         case AV_PIX_FMT_NV21:

  573.             c->chrToYV12 = ff_nv21ToUV_avx;

  574.             break;

  575.         case_rgb(rgb24, RGB24, avx);

  576.         case_rgb(bgr24, BGR24, avx);

  577.         case_rgb(bgra,  BGRA,  avx);

  578.         case_rgb(rgba,  RGBA,  avx);

  579.         case_rgb(abgr,  ABGR,  avx);

  580.         case_rgb(argb,  ARGB,  avx);

  581.         default:

  582.             break;

  583.         }

  584.     }

  585. }