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.
44689 Commits
32 Branches
1.0GB
Tree: c5714b51aa
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'c5714b51aa'
${ noResults }
FFmpeg/libswscale/x86/swscale_template.c

1637 lines
72KB
Raw Blame History 

  1. /*

  2.  * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>

  3.  *

  4.  * This file is part of Libav.

  5.  *

  6.  * Libav 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.  * Libav 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 Libav; if not, write to the Free Software

  18.  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

  19.  */

  20. 

  21. #include <stdint.h>

  22. 

  23. #include "libavutil/x86/asm.h"

  24. #include "libswscale/swscale_internal.h"

  25. 

  26. #undef REAL_MOVNTQ

  27. #undef MOVNTQ

  28. #undef PREFETCH

  29. 

  30. #if COMPILE_TEMPLATE_MMXEXT

  31. #define PREFETCH "prefetchnta"

  32. #else

  33. #define PREFETCH  " # nop"

  34. #endif

  35. 

  36. #if COMPILE_TEMPLATE_MMXEXT

  37. #define REAL_MOVNTQ(a,b) "movntq " #a ", " #b " \n\t"

  38. #else

  39. #define REAL_MOVNTQ(a,b) "movq " #a ", " #b " \n\t"

  40. #endif

  41. #define MOVNTQ(a,b)  REAL_MOVNTQ(a,b)

  42. 

  43. #define YSCALEYUV2PACKEDX_UV \

  44.     __asm__ volatile(\

  45.         "xor                %%"FF_REG_a", %%"FF_REG_a"  \n\t"\

  46.         ".p2align                      4                \n\t"\

  47.         "nop                                            \n\t"\

  48.         "1:                                             \n\t"\

  49.         "lea "CHR_MMX_FILTER_OFFSET"(%0), %%"FF_REG_d"  \n\t"\

  50.         "mov              (%%"FF_REG_d"), %%"FF_REG_S"  \n\t"\

  51.         "movq      "VROUNDER_OFFSET"(%0), %%mm3         \n\t"\

  52.         "movq                      %%mm3, %%mm4         \n\t"\

  53.         ".p2align                      4                \n\t"\

  54.         "2:                                             \n\t"\

  55.         "movq            8(%%"FF_REG_d"), %%mm0         \n\t" /* filterCoeff */\

  56.         "movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm2       \n\t" /* UsrcData */\

  57.         "add                       %6, %%"FF_REG_S"     \n\t" \

  58.         "movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm5       \n\t" /* VsrcData */\

  59.         "add                      $16, %%"FF_REG_d"     \n\t"\

  60.         "mov           (%%"FF_REG_d"), %%"FF_REG_S"     \n\t"\

  61.         "pmulhw                    %%mm0, %%mm2         \n\t"\

  62.         "pmulhw                    %%mm0, %%mm5         \n\t"\

  63.         "paddw                     %%mm2, %%mm3         \n\t"\

  64.         "paddw                     %%mm5, %%mm4         \n\t"\

  65.         "test               %%"FF_REG_S", %%"FF_REG_S"  \n\t"\

  66.         " jnz                         2b                \n\t"\

  67. 

  68. #define YSCALEYUV2PACKEDX_YA(offset,coeff,src1,src2,dst1,dst2) \

  69.     "lea                "offset"(%0), %%"FF_REG_d"  \n\t"\

  70.     "mov              (%%"FF_REG_d"), %%"FF_REG_S"  \n\t"\

  71.     "movq      "VROUNDER_OFFSET"(%0), "#dst1"       \n\t"\

  72.     "movq                    "#dst1", "#dst2"       \n\t"\

  73.     ".p2align                      4                \n\t"\

  74.     "2:                                             \n\t"\

  75.     "movq            8(%%"FF_REG_d"), "#coeff"      \n\t" /* filterCoeff */\

  76.     "movq  (%%"FF_REG_S", %%"FF_REG_a", 2), "#src1" \n\t" /* Y1srcData */\

  77.     "movq 8(%%"FF_REG_S", %%"FF_REG_a", 2), "#src2" \n\t" /* Y2srcData */\

  78.     "add                         $16, %%"FF_REG_d"  \n\t"\

  79.     "mov              (%%"FF_REG_d"), %%"FF_REG_S"  \n\t"\

  80.     "pmulhw                 "#coeff", "#src1"       \n\t"\

  81.     "pmulhw                 "#coeff", "#src2"       \n\t"\

  82.     "paddw                   "#src1", "#dst1"       \n\t"\

  83.     "paddw                   "#src2", "#dst2"       \n\t"\

  84.     "test               %%"FF_REG_S", %%"FF_REG_S"  \n\t"\

  85.     " jnz                         2b                \n\t"\

  86. 

  87. #define YSCALEYUV2PACKEDX \

  88.     YSCALEYUV2PACKEDX_UV \

  89.     YSCALEYUV2PACKEDX_YA(LUM_MMX_FILTER_OFFSET,%%mm0,%%mm2,%%mm5,%%mm1,%%mm7) \

  90. 

  91. #define YSCALEYUV2PACKEDX_END                     \

  92.         :: "r" (&c->redDither),                   \

  93.             "m" (dummy), "m" (dummy), "m" (dummy),\

  94.             "r" (dest), "m" (dstW_reg), "m"(uv_off) \

  95.         : "%"FF_REG_a, "%"FF_REG_d, "%"FF_REG_S   \

  96.     );

  97. 

  98. #define YSCALEYUV2PACKEDX_ACCURATE_UV \

  99.     __asm__ volatile(\

  100.         "xor %%"FF_REG_a", %%"FF_REG_a"                 \n\t"\

  101.         ".p2align                      4                \n\t"\

  102.         "nop                                            \n\t"\

  103.         "1:                                             \n\t"\

  104.         "lea "CHR_MMX_FILTER_OFFSET"(%0), %%"FF_REG_d"  \n\t"\

  105.         "mov              (%%"FF_REG_d"), %%"FF_REG_S"  \n\t"\

  106.         "pxor                      %%mm4, %%mm4         \n\t"\

  107.         "pxor                      %%mm5, %%mm5         \n\t"\

  108.         "pxor                      %%mm6, %%mm6         \n\t"\

  109.         "pxor                      %%mm7, %%mm7         \n\t"\

  110.         ".p2align                      4                \n\t"\

  111.         "2:                                             \n\t"\

  112.         "movq  (%%"FF_REG_S", %%"FF_REG_a"), %%mm0      \n\t" /* UsrcData */\

  113.         "add                             %6, %%"FF_REG_S" \n\t" \

  114.         "movq  (%%"FF_REG_S", %%"FF_REG_a"), %%mm2      \n\t" /* VsrcData */\

  115.         "mov "STR(APCK_PTR2)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\

  116.         "movq  (%%"FF_REG_S", %%"FF_REG_a"), %%mm1      \n\t" /* UsrcData */\

  117.         "movq                      %%mm0, %%mm3         \n\t"\

  118.         "punpcklwd                 %%mm1, %%mm0         \n\t"\

  119.         "punpckhwd                 %%mm1, %%mm3         \n\t"\

  120.         "movq "STR(APCK_COEF)"(%%"FF_REG_d"),%%mm1      \n\t" /* filterCoeff */\

  121.         "pmaddwd                   %%mm1, %%mm0         \n\t"\

  122.         "pmaddwd                   %%mm1, %%mm3         \n\t"\

  123.         "paddd                     %%mm0, %%mm4         \n\t"\

  124.         "paddd                     %%mm3, %%mm5         \n\t"\

  125.         "add                          %6, %%"FF_REG_S"  \n\t"\

  126.         "movq  (%%"FF_REG_S", %%"FF_REG_a"), %%mm3      \n\t" /* VsrcData */\

  127.         "mov "STR(APCK_SIZE)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\

  128.         "add           $"STR(APCK_SIZE)", %%"FF_REG_d"  \n\t"\

  129.         "test               %%"FF_REG_S", %%"FF_REG_S"  \n\t"\

  130.         "movq                      %%mm2, %%mm0         \n\t"\

  131.         "punpcklwd                 %%mm3, %%mm2         \n\t"\

  132.         "punpckhwd                 %%mm3, %%mm0         \n\t"\

  133.         "pmaddwd                   %%mm1, %%mm2         \n\t"\

  134.         "pmaddwd                   %%mm1, %%mm0         \n\t"\

  135.         "paddd                     %%mm2, %%mm6         \n\t"\

  136.         "paddd                     %%mm0, %%mm7         \n\t"\

  137.         " jnz                         2b                \n\t"\

  138.         "psrad                       $16, %%mm4         \n\t"\

  139.         "psrad                       $16, %%mm5         \n\t"\

  140.         "psrad                       $16, %%mm6         \n\t"\

  141.         "psrad                       $16, %%mm7         \n\t"\

  142.         "movq      "VROUNDER_OFFSET"(%0), %%mm0         \n\t"\

  143.         "packssdw                  %%mm5, %%mm4         \n\t"\

  144.         "packssdw                  %%mm7, %%mm6         \n\t"\

  145.         "paddw                     %%mm0, %%mm4         \n\t"\

  146.         "paddw                     %%mm0, %%mm6         \n\t"\

  147.         "movq                      %%mm4, "U_TEMP"(%0)  \n\t"\

  148.         "movq                      %%mm6, "V_TEMP"(%0)  \n\t"\

  149. 

  150. #define YSCALEYUV2PACKEDX_ACCURATE_YA(offset) \

  151.     "lea                "offset"(%0), %%"FF_REG_d"  \n\t"\

  152.     "mov              (%%"FF_REG_d"), %%"FF_REG_S"  \n\t"\

  153.     "pxor                      %%mm1, %%mm1         \n\t"\

  154.     "pxor                      %%mm5, %%mm5         \n\t"\

  155.     "pxor                      %%mm7, %%mm7         \n\t"\

  156.     "pxor                      %%mm6, %%mm6         \n\t"\

  157.     ".p2align                      4                \n\t"\

  158.     "2:                                             \n\t"\

  159.     "movq  (%%"FF_REG_S", %%"FF_REG_a", 2), %%mm0   \n\t" /* Y1srcData */\

  160.     "movq 8(%%"FF_REG_S", %%"FF_REG_a", 2), %%mm2   \n\t" /* Y2srcData */\

  161.     "mov "STR(APCK_PTR2)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\

  162.     "movq  (%%"FF_REG_S", %%"FF_REG_a", 2), %%mm4   \n\t" /* Y1srcData */\

  163.     "movq                      %%mm0, %%mm3         \n\t"\

  164.     "punpcklwd                 %%mm4, %%mm0         \n\t"\

  165.     "punpckhwd                 %%mm4, %%mm3         \n\t"\

  166.     "movq "STR(APCK_COEF)"(%%"FF_REG_d"), %%mm4     \n\t" /* filterCoeff */\

  167.     "pmaddwd                   %%mm4, %%mm0         \n\t"\

  168.     "pmaddwd                   %%mm4, %%mm3         \n\t"\

  169.     "paddd                     %%mm0, %%mm1         \n\t"\

  170.     "paddd                     %%mm3, %%mm5         \n\t"\

  171.     "movq 8(%%"FF_REG_S", %%"FF_REG_a", 2), %%mm3   \n\t" /* Y2srcData */\

  172.     "mov "STR(APCK_SIZE)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\

  173.     "add           $"STR(APCK_SIZE)", %%"FF_REG_d"  \n\t"\

  174.     "test               %%"FF_REG_S", %%"FF_REG_S"  \n\t"\

  175.     "movq                      %%mm2, %%mm0         \n\t"\

  176.     "punpcklwd                 %%mm3, %%mm2         \n\t"\

  177.     "punpckhwd                 %%mm3, %%mm0         \n\t"\

  178.     "pmaddwd                   %%mm4, %%mm2         \n\t"\

  179.     "pmaddwd                   %%mm4, %%mm0         \n\t"\

  180.     "paddd                     %%mm2, %%mm7         \n\t"\

  181.     "paddd                     %%mm0, %%mm6         \n\t"\

  182.     " jnz                         2b                \n\t"\

  183.     "psrad                       $16, %%mm1         \n\t"\

  184.     "psrad                       $16, %%mm5         \n\t"\

  185.     "psrad                       $16, %%mm7         \n\t"\

  186.     "psrad                       $16, %%mm6         \n\t"\

  187.     "movq      "VROUNDER_OFFSET"(%0), %%mm0         \n\t"\

  188.     "packssdw                  %%mm5, %%mm1         \n\t"\

  189.     "packssdw                  %%mm6, %%mm7         \n\t"\

  190.     "paddw                     %%mm0, %%mm1         \n\t"\

  191.     "paddw                     %%mm0, %%mm7         \n\t"\

  192.     "movq               "U_TEMP"(%0), %%mm3         \n\t"\

  193.     "movq               "V_TEMP"(%0), %%mm4         \n\t"\

  194. 

  195. #define YSCALEYUV2PACKEDX_ACCURATE \

  196.     YSCALEYUV2PACKEDX_ACCURATE_UV \

  197.     YSCALEYUV2PACKEDX_ACCURATE_YA(LUM_MMX_FILTER_OFFSET)

  198. 

  199. #define YSCALEYUV2RGBX \

  200.     "psubw  "U_OFFSET"(%0), %%mm3       \n\t" /* (U-128)8*/\

  201.     "psubw  "V_OFFSET"(%0), %%mm4       \n\t" /* (V-128)8*/\

  202.     "movq            %%mm3, %%mm2       \n\t" /* (U-128)8*/\

  203.     "movq            %%mm4, %%mm5       \n\t" /* (V-128)8*/\

  204.     "pmulhw "UG_COEFF"(%0), %%mm3       \n\t"\

  205.     "pmulhw "VG_COEFF"(%0), %%mm4       \n\t"\

  206.     /* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\

  207.     "pmulhw "UB_COEFF"(%0), %%mm2       \n\t"\

  208.     "pmulhw "VR_COEFF"(%0), %%mm5       \n\t"\

  209.     "psubw  "Y_OFFSET"(%0), %%mm1       \n\t" /* 8(Y-16)*/\

  210.     "psubw  "Y_OFFSET"(%0), %%mm7       \n\t" /* 8(Y-16)*/\

  211.     "pmulhw  "Y_COEFF"(%0), %%mm1       \n\t"\

  212.     "pmulhw  "Y_COEFF"(%0), %%mm7       \n\t"\

  213.     /* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\

  214.     "paddw           %%mm3, %%mm4       \n\t"\

  215.     "movq            %%mm2, %%mm0       \n\t"\

  216.     "movq            %%mm5, %%mm6       \n\t"\

  217.     "movq            %%mm4, %%mm3       \n\t"\

  218.     "punpcklwd       %%mm2, %%mm2       \n\t"\

  219.     "punpcklwd       %%mm5, %%mm5       \n\t"\

  220.     "punpcklwd       %%mm4, %%mm4       \n\t"\

  221.     "paddw           %%mm1, %%mm2       \n\t"\

  222.     "paddw           %%mm1, %%mm5       \n\t"\

  223.     "paddw           %%mm1, %%mm4       \n\t"\

  224.     "punpckhwd       %%mm0, %%mm0       \n\t"\

  225.     "punpckhwd       %%mm6, %%mm6       \n\t"\

  226.     "punpckhwd       %%mm3, %%mm3       \n\t"\

  227.     "paddw           %%mm7, %%mm0       \n\t"\

  228.     "paddw           %%mm7, %%mm6       \n\t"\

  229.     "paddw           %%mm7, %%mm3       \n\t"\

  230.     /* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\

  231.     "packuswb        %%mm0, %%mm2       \n\t"\

  232.     "packuswb        %%mm6, %%mm5       \n\t"\

  233.     "packuswb        %%mm3, %%mm4       \n\t"\

  234. 

  235. #define REAL_WRITEBGR32(dst, dstw, index, b, g, r, a, q0, q2, q3, t) \

  236.     "movq       "#b", "#q2"     \n\t" /* B */\

  237.     "movq       "#r", "#t"      \n\t" /* R */\

  238.     "punpcklbw  "#g", "#b"      \n\t" /* GBGBGBGB 0 */\

  239.     "punpcklbw  "#a", "#r"      \n\t" /* ARARARAR 0 */\

  240.     "punpckhbw  "#g", "#q2"     \n\t" /* GBGBGBGB 2 */\

  241.     "punpckhbw  "#a", "#t"      \n\t" /* ARARARAR 2 */\

  242.     "movq       "#b", "#q0"     \n\t" /* GBGBGBGB 0 */\

  243.     "movq      "#q2", "#q3"     \n\t" /* GBGBGBGB 2 */\

  244.     "punpcklwd  "#r", "#q0"     \n\t" /* ARGBARGB 0 */\

  245.     "punpckhwd  "#r", "#b"      \n\t" /* ARGBARGB 1 */\

  246.     "punpcklwd  "#t", "#q2"     \n\t" /* ARGBARGB 2 */\

  247.     "punpckhwd  "#t", "#q3"     \n\t" /* ARGBARGB 3 */\

  248. \

  249.     MOVNTQ(   q0,   (dst, index, 4))\

  250.     MOVNTQ(    b,  8(dst, index, 4))\

  251.     MOVNTQ(   q2, 16(dst, index, 4))\

  252.     MOVNTQ(   q3, 24(dst, index, 4))\

  253. \

  254.     "add      $8, "#index"      \n\t"\

  255.     "cmp "#dstw", "#index"      \n\t"\

  256.     " jb      1b                \n\t"

  257. #define WRITEBGR32(dst, dstw, index, b, g, r, a, q0, q2, q3, t)  REAL_WRITEBGR32(dst, dstw, index, b, g, r, a, q0, q2, q3, t)

  258. 

  259. static void RENAME(yuv2rgb32_X_ar)(SwsContext *c, const int16_t *lumFilter,

  260.                                    const int16_t **lumSrc, int lumFilterSize,

  261.                                    const int16_t *chrFilter, const int16_t **chrUSrc,

  262.                                    const int16_t **chrVSrc,

  263.                                    int chrFilterSize, const int16_t **alpSrc,

  264.                                    uint8_t *dest, int dstW, int dstY)

  265. {

  266.     x86_reg dummy=0;

  267.     x86_reg dstW_reg = dstW;

  268.     x86_reg uv_off = c->uv_off_byte;

  269. 

  270.     if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {

  271.         YSCALEYUV2PACKEDX_ACCURATE

  272.         YSCALEYUV2RGBX

  273.         "movq                      %%mm2, "U_TEMP"(%0)  \n\t"

  274.         "movq                      %%mm4, "V_TEMP"(%0)  \n\t"

  275.         "movq                      %%mm5, "Y_TEMP"(%0)  \n\t"

  276.         YSCALEYUV2PACKEDX_ACCURATE_YA(ALP_MMX_FILTER_OFFSET)

  277.         "movq               "Y_TEMP"(%0), %%mm5         \n\t"

  278.         "psraw                        $3, %%mm1         \n\t"

  279.         "psraw                        $3, %%mm7         \n\t"

  280.         "packuswb                  %%mm7, %%mm1         \n\t"

  281.         WRITEBGR32(%4, %5, %%FF_REGa, %%mm3, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm2, %%mm6)

  282.         YSCALEYUV2PACKEDX_END

  283.     } else {

  284.         YSCALEYUV2PACKEDX_ACCURATE

  285.         YSCALEYUV2RGBX

  286.         "pcmpeqd %%mm7, %%mm7 \n\t"

  287.         WRITEBGR32(%4, %5, %%FF_REGa, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)

  288.         YSCALEYUV2PACKEDX_END

  289.     }

  290. }

  291. 

  292. static void RENAME(yuv2rgb32_X)(SwsContext *c, const int16_t *lumFilter,

  293.                                 const int16_t **lumSrc, int lumFilterSize,

  294.                                 const int16_t *chrFilter, const int16_t **chrUSrc,

  295.                                 const int16_t **chrVSrc,

  296.                                 int chrFilterSize, const int16_t **alpSrc,

  297.                                 uint8_t *dest, int dstW, int dstY)

  298. {

  299.     x86_reg dummy=0;

  300.     x86_reg dstW_reg = dstW;

  301.     x86_reg uv_off = c->uv_off_byte;

  302. 

  303.     if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {

  304.         YSCALEYUV2PACKEDX

  305.         YSCALEYUV2RGBX

  306.         YSCALEYUV2PACKEDX_YA(ALP_MMX_FILTER_OFFSET, %%mm0, %%mm3, %%mm6, %%mm1, %%mm7)

  307.         "psraw                        $3, %%mm1         \n\t"

  308.         "psraw                        $3, %%mm7         \n\t"

  309.         "packuswb                  %%mm7, %%mm1         \n\t"

  310.         WRITEBGR32(%4, %5, %%FF_REGa, %%mm2, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm3, %%mm6)

  311.         YSCALEYUV2PACKEDX_END

  312.     } else {

  313.         YSCALEYUV2PACKEDX

  314.         YSCALEYUV2RGBX

  315.         "pcmpeqd %%mm7, %%mm7 \n\t"

  316.         WRITEBGR32(%4, %5, %%FF_REGa, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)

  317.         YSCALEYUV2PACKEDX_END

  318.     }

  319. }

  320. 

  321. #define REAL_WRITERGB16(dst, dstw, index) \

  322.     "pand "MANGLE(bF8)", %%mm2  \n\t" /* B */\

  323.     "pand "MANGLE(bFC)", %%mm4  \n\t" /* G */\

  324.     "pand "MANGLE(bF8)", %%mm5  \n\t" /* R */\

  325.     "psrlq           $3, %%mm2  \n\t"\

  326. \

  327.     "movq         %%mm2, %%mm1  \n\t"\

  328.     "movq         %%mm4, %%mm3  \n\t"\

  329. \

  330.     "punpcklbw    %%mm7, %%mm3  \n\t"\

  331.     "punpcklbw    %%mm5, %%mm2  \n\t"\

  332.     "punpckhbw    %%mm7, %%mm4  \n\t"\

  333.     "punpckhbw    %%mm5, %%mm1  \n\t"\

  334. \

  335.     "psllq           $3, %%mm3  \n\t"\

  336.     "psllq           $3, %%mm4  \n\t"\

  337. \

  338.     "por          %%mm3, %%mm2  \n\t"\

  339.     "por          %%mm4, %%mm1  \n\t"\

  340. \

  341.     MOVNTQ(%%mm2,  (dst, index, 2))\

  342.     MOVNTQ(%%mm1, 8(dst, index, 2))\

  343. \

  344.     "add             $8, "#index"   \n\t"\

  345.     "cmp        "#dstw", "#index"   \n\t"\

  346.     " jb             1b             \n\t"

  347. #define WRITERGB16(dst, dstw, index)  REAL_WRITERGB16(dst, dstw, index)

  348. 

  349. static void RENAME(yuv2rgb565_X_ar)(SwsContext *c, const int16_t *lumFilter,

  350.                                     const int16_t **lumSrc, int lumFilterSize,

  351.                                     const int16_t *chrFilter, const int16_t **chrUSrc,

  352.                                     const int16_t **chrVSrc,

  353.                                     int chrFilterSize, const int16_t **alpSrc,

  354.                                     uint8_t *dest, int dstW, int dstY)

  355. {

  356.     x86_reg dummy=0;

  357.     x86_reg dstW_reg = dstW;

  358.     x86_reg uv_off = c->uv_off_byte;

  359. 

  360.     YSCALEYUV2PACKEDX_ACCURATE

  361.     YSCALEYUV2RGBX

  362.     "pxor %%mm7, %%mm7 \n\t"

  363.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  364. #ifdef DITHER1XBPP

  365.     "paddusb "BLUE_DITHER"(%0), %%mm2\n\t"

  366.     "paddusb "GREEN_DITHER"(%0), %%mm4\n\t"

  367.     "paddusb "RED_DITHER"(%0), %%mm5\n\t"

  368. #endif

  369.     WRITERGB16(%4, %5, %%FF_REGa)

  370.     YSCALEYUV2PACKEDX_END

  371. }

  372. 

  373. static void RENAME(yuv2rgb565_X)(SwsContext *c, const int16_t *lumFilter,

  374.                                  const int16_t **lumSrc, int lumFilterSize,

  375.                                  const int16_t *chrFilter, const int16_t **chrUSrc,

  376.                                  const int16_t **chrVSrc,

  377.                                  int chrFilterSize, const int16_t **alpSrc,

  378.                                  uint8_t *dest, int dstW, int dstY)

  379. {

  380.     x86_reg dummy=0;

  381.     x86_reg dstW_reg = dstW;

  382.     x86_reg uv_off = c->uv_off_byte;

  383. 

  384.     YSCALEYUV2PACKEDX

  385.     YSCALEYUV2RGBX

  386.     "pxor %%mm7, %%mm7 \n\t"

  387.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  388. #ifdef DITHER1XBPP

  389.     "paddusb "BLUE_DITHER"(%0), %%mm2  \n\t"

  390.     "paddusb "GREEN_DITHER"(%0), %%mm4  \n\t"

  391.     "paddusb "RED_DITHER"(%0), %%mm5  \n\t"

  392. #endif

  393.     WRITERGB16(%4, %5, %%FF_REGa)

  394.     YSCALEYUV2PACKEDX_END

  395. }

  396. 

  397. #define REAL_WRITERGB15(dst, dstw, index) \

  398.     "pand "MANGLE(bF8)", %%mm2  \n\t" /* B */\

  399.     "pand "MANGLE(bF8)", %%mm4  \n\t" /* G */\

  400.     "pand "MANGLE(bF8)", %%mm5  \n\t" /* R */\

  401.     "psrlq           $3, %%mm2  \n\t"\

  402.     "psrlq           $1, %%mm5  \n\t"\

  403. \

  404.     "movq         %%mm2, %%mm1  \n\t"\

  405.     "movq         %%mm4, %%mm3  \n\t"\

  406. \

  407.     "punpcklbw    %%mm7, %%mm3  \n\t"\

  408.     "punpcklbw    %%mm5, %%mm2  \n\t"\

  409.     "punpckhbw    %%mm7, %%mm4  \n\t"\

  410.     "punpckhbw    %%mm5, %%mm1  \n\t"\

  411. \

  412.     "psllq           $2, %%mm3  \n\t"\

  413.     "psllq           $2, %%mm4  \n\t"\

  414. \

  415.     "por          %%mm3, %%mm2  \n\t"\

  416.     "por          %%mm4, %%mm1  \n\t"\

  417. \

  418.     MOVNTQ(%%mm2,  (dst, index, 2))\

  419.     MOVNTQ(%%mm1, 8(dst, index, 2))\

  420. \

  421.     "add             $8, "#index"   \n\t"\

  422.     "cmp        "#dstw", "#index"   \n\t"\

  423.     " jb             1b             \n\t"

  424. #define WRITERGB15(dst, dstw, index)  REAL_WRITERGB15(dst, dstw, index)

  425. 

  426. static void RENAME(yuv2rgb555_X_ar)(SwsContext *c, const int16_t *lumFilter,

  427.                                     const int16_t **lumSrc, int lumFilterSize,

  428.                                     const int16_t *chrFilter, const int16_t **chrUSrc,

  429.                                     const int16_t **chrVSrc,

  430.                                     int chrFilterSize, const int16_t **alpSrc,

  431.                                     uint8_t *dest, int dstW, int dstY)

  432. {

  433.     x86_reg dummy=0;

  434.     x86_reg dstW_reg = dstW;

  435.     x86_reg uv_off = c->uv_off_byte;

  436. 

  437.     YSCALEYUV2PACKEDX_ACCURATE

  438.     YSCALEYUV2RGBX

  439.     "pxor %%mm7, %%mm7 \n\t"

  440.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  441. #ifdef DITHER1XBPP

  442.     "paddusb "BLUE_DITHER"(%0), %%mm2\n\t"

  443.     "paddusb "GREEN_DITHER"(%0), %%mm4\n\t"

  444.     "paddusb "RED_DITHER"(%0), %%mm5\n\t"

  445. #endif

  446.     WRITERGB15(%4, %5, %%FF_REGa)

  447.     YSCALEYUV2PACKEDX_END

  448. }

  449. 

  450. static void RENAME(yuv2rgb555_X)(SwsContext *c, const int16_t *lumFilter,

  451.                                  const int16_t **lumSrc, int lumFilterSize,

  452.                                  const int16_t *chrFilter, const int16_t **chrUSrc,

  453.                                  const int16_t **chrVSrc,

  454.                                  int chrFilterSize, const int16_t **alpSrc,

  455.                                  uint8_t *dest, int dstW, int dstY)

  456. {

  457.     x86_reg dummy=0;

  458.     x86_reg dstW_reg = dstW;

  459.     x86_reg uv_off = c->uv_off_byte;

  460. 

  461.     YSCALEYUV2PACKEDX

  462.     YSCALEYUV2RGBX

  463.     "pxor %%mm7, %%mm7 \n\t"

  464.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  465. #ifdef DITHER1XBPP

  466.     "paddusb "BLUE_DITHER"(%0), %%mm2  \n\t"

  467.     "paddusb "GREEN_DITHER"(%0), %%mm4  \n\t"

  468.     "paddusb "RED_DITHER"(%0), %%mm5  \n\t"

  469. #endif

  470.     WRITERGB15(%4, %5, %%FF_REGa)

  471.     YSCALEYUV2PACKEDX_END

  472. }

  473. 

  474. #define WRITEBGR24MMX(dst, dstw, index) \

  475.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */\

  476.     "movq      %%mm2, %%mm1     \n\t" /* B */\

  477.     "movq      %%mm5, %%mm6     \n\t" /* R */\

  478.     "punpcklbw %%mm4, %%mm2     \n\t" /* GBGBGBGB 0 */\

  479.     "punpcklbw %%mm7, %%mm5     \n\t" /* 0R0R0R0R 0 */\

  480.     "punpckhbw %%mm4, %%mm1     \n\t" /* GBGBGBGB 2 */\

  481.     "punpckhbw %%mm7, %%mm6     \n\t" /* 0R0R0R0R 2 */\

  482.     "movq      %%mm2, %%mm0     \n\t" /* GBGBGBGB 0 */\

  483.     "movq      %%mm1, %%mm3     \n\t" /* GBGBGBGB 2 */\

  484.     "punpcklwd %%mm5, %%mm0     \n\t" /* 0RGB0RGB 0 */\

  485.     "punpckhwd %%mm5, %%mm2     \n\t" /* 0RGB0RGB 1 */\

  486.     "punpcklwd %%mm6, %%mm1     \n\t" /* 0RGB0RGB 2 */\

  487.     "punpckhwd %%mm6, %%mm3     \n\t" /* 0RGB0RGB 3 */\

  488. \

  489.     "movq      %%mm0, %%mm4     \n\t" /* 0RGB0RGB 0 */\

  490.     "movq      %%mm2, %%mm6     \n\t" /* 0RGB0RGB 1 */\

  491.     "movq      %%mm1, %%mm5     \n\t" /* 0RGB0RGB 2 */\

  492.     "movq      %%mm3, %%mm7     \n\t" /* 0RGB0RGB 3 */\

  493. \

  494.     "psllq       $40, %%mm0     \n\t" /* RGB00000 0 */\

  495.     "psllq       $40, %%mm2     \n\t" /* RGB00000 1 */\

  496.     "psllq       $40, %%mm1     \n\t" /* RGB00000 2 */\

  497.     "psllq       $40, %%mm3     \n\t" /* RGB00000 3 */\

  498. \

  499.     "punpckhdq %%mm4, %%mm0     \n\t" /* 0RGBRGB0 0 */\

  500.     "punpckhdq %%mm6, %%mm2     \n\t" /* 0RGBRGB0 1 */\

  501.     "punpckhdq %%mm5, %%mm1     \n\t" /* 0RGBRGB0 2 */\

  502.     "punpckhdq %%mm7, %%mm3     \n\t" /* 0RGBRGB0 3 */\

  503. \

  504.     "psrlq        $8, %%mm0     \n\t" /* 00RGBRGB 0 */\

  505.     "movq      %%mm2, %%mm6     \n\t" /* 0RGBRGB0 1 */\

  506.     "psllq       $40, %%mm2     \n\t" /* GB000000 1 */\

  507.     "por       %%mm2, %%mm0     \n\t" /* GBRGBRGB 0 */\

  508.     MOVNTQ(%%mm0, (dst))\

  509. \

  510.     "psrlq       $24, %%mm6     \n\t" /* 0000RGBR 1 */\

  511.     "movq      %%mm1, %%mm5     \n\t" /* 0RGBRGB0 2 */\

  512.     "psllq       $24, %%mm1     \n\t" /* BRGB0000 2 */\

  513.     "por       %%mm1, %%mm6     \n\t" /* BRGBRGBR 1 */\

  514.     MOVNTQ(%%mm6, 8(dst))\

  515. \

  516.     "psrlq       $40, %%mm5     \n\t" /* 000000RG 2 */\

  517.     "psllq        $8, %%mm3     \n\t" /* RGBRGB00 3 */\

  518.     "por       %%mm3, %%mm5     \n\t" /* RGBRGBRG 2 */\

  519.     MOVNTQ(%%mm5, 16(dst))\

  520. \

  521.     "add         $24, "#dst"    \n\t"\

  522. \

  523.     "add          $8, "#index"  \n\t"\

  524.     "cmp     "#dstw", "#index"  \n\t"\

  525.     " jb          1b            \n\t"

  526. 

  527. #define WRITEBGR24MMXEXT(dst, dstw, index) \

  528.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */\

  529.     "movq "MANGLE(ff_M24A)", %%mm0 \n\t"\

  530.     "movq "MANGLE(ff_M24C)", %%mm7 \n\t"\

  531.     "pshufw $0x50, %%mm2, %%mm1 \n\t" /* B3 B2 B3 B2  B1 B0 B1 B0 */\

  532.     "pshufw $0x50, %%mm4, %%mm3 \n\t" /* G3 G2 G3 G2  G1 G0 G1 G0 */\

  533.     "pshufw $0x00, %%mm5, %%mm6 \n\t" /* R1 R0 R1 R0  R1 R0 R1 R0 */\

  534. \

  535.     "pand   %%mm0, %%mm1        \n\t" /*    B2        B1       B0 */\

  536.     "pand   %%mm0, %%mm3        \n\t" /*    G2        G1       G0 */\

  537.     "pand   %%mm7, %%mm6        \n\t" /*       R1        R0       */\

  538. \

  539.     "psllq     $8, %%mm3        \n\t" /* G2        G1       G0    */\

  540.     "por    %%mm1, %%mm6        \n\t"\

  541.     "por    %%mm3, %%mm6        \n\t"\

  542.     MOVNTQ(%%mm6, (dst))\

  543. \

  544.     "psrlq     $8, %%mm4        \n\t" /* 00 G7 G6 G5  G4 G3 G2 G1 */\

  545.     "pshufw $0xA5, %%mm2, %%mm1 \n\t" /* B5 B4 B5 B4  B3 B2 B3 B2 */\

  546.     "pshufw $0x55, %%mm4, %%mm3 \n\t" /* G4 G3 G4 G3  G4 G3 G4 G3 */\

  547.     "pshufw $0xA5, %%mm5, %%mm6 \n\t" /* R5 R4 R5 R4  R3 R2 R3 R2 */\

  548. \

  549.     "pand "MANGLE(ff_M24B)", %%mm1 \n\t" /* B5       B4        B3    */\

  550.     "pand   %%mm7, %%mm3        \n\t" /*       G4        G3       */\

  551.     "pand   %%mm0, %%mm6        \n\t" /*    R4        R3       R2 */\

  552. \

  553.     "por    %%mm1, %%mm3        \n\t" /* B5    G4 B4     G3 B3    */\

  554.     "por    %%mm3, %%mm6        \n\t"\

  555.     MOVNTQ(%%mm6, 8(dst))\

  556. \

  557.     "pshufw $0xFF, %%mm2, %%mm1 \n\t" /* B7 B6 B7 B6  B7 B6 B6 B7 */\

  558.     "pshufw $0xFA, %%mm4, %%mm3 \n\t" /* 00 G7 00 G7  G6 G5 G6 G5 */\

  559.     "pshufw $0xFA, %%mm5, %%mm6 \n\t" /* R7 R6 R7 R6  R5 R4 R5 R4 */\

  560. \

  561.     "pand   %%mm7, %%mm1        \n\t" /*       B7        B6       */\

  562.     "pand   %%mm0, %%mm3        \n\t" /*    G7        G6       G5 */\

  563.     "pand "MANGLE(ff_M24B)", %%mm6 \n\t" /* R7       R6        R5    */\

  564. \

  565.     "por    %%mm1, %%mm3        \n\t"\

  566.     "por    %%mm3, %%mm6        \n\t"\

  567.     MOVNTQ(%%mm6, 16(dst))\

  568. \

  569.     "add      $24, "#dst"       \n\t"\

  570. \

  571.     "add       $8, "#index"     \n\t"\

  572.     "cmp  "#dstw", "#index"     \n\t"\

  573.     " jb       1b               \n\t"

  574. 

  575. #if COMPILE_TEMPLATE_MMXEXT

  576. #undef WRITEBGR24

  577. #define WRITEBGR24(dst, dstw, index)  WRITEBGR24MMXEXT(dst, dstw, index)

  578. #else

  579. #undef WRITEBGR24

  580. #define WRITEBGR24(dst, dstw, index)  WRITEBGR24MMX(dst, dstw, index)

  581. #endif

  582. 

  583. static void RENAME(yuv2bgr24_X_ar)(SwsContext *c, const int16_t *lumFilter,

  584.                                    const int16_t **lumSrc, int lumFilterSize,

  585.                                    const int16_t *chrFilter, const int16_t **chrUSrc,

  586.                                    const int16_t **chrVSrc,

  587.                                    int chrFilterSize, const int16_t **alpSrc,

  588.                                    uint8_t *dest, int dstW, int dstY)

  589. {

  590.     x86_reg dummy=0;

  591.     x86_reg dstW_reg = dstW;

  592.     x86_reg uv_off = c->uv_off_byte;

  593. 

  594.     YSCALEYUV2PACKEDX_ACCURATE

  595.     YSCALEYUV2RGBX

  596.     "pxor                          %%mm7, %%mm7         \n\t"

  597.     "lea (%%"FF_REG_a", %%"FF_REG_a", 2), %%"FF_REG_c"  \n\t" // FIXME optimize

  598.     "add                              %4, %%"FF_REG_c"  \n\t"

  599.     WRITEBGR24(%%FF_REGc, %5, %%FF_REGa)

  600.     :: "r" (&c->redDither),

  601.        "m" (dummy), "m" (dummy), "m" (dummy),

  602.        "r" (dest), "m" (dstW_reg), "m"(uv_off)

  603.     : "%"FF_REG_a, "%"FF_REG_c, "%"FF_REG_d, "%"FF_REG_S

  604.     );

  605. }

  606. 

  607. static void RENAME(yuv2bgr24_X)(SwsContext *c, const int16_t *lumFilter,

  608.                                 const int16_t **lumSrc, int lumFilterSize,

  609.                                 const int16_t *chrFilter, const int16_t **chrUSrc,

  610.                                 const int16_t **chrVSrc,

  611.                                 int chrFilterSize, const int16_t **alpSrc,

  612.                                 uint8_t *dest, int dstW, int dstY)

  613. {

  614.     x86_reg dummy=0;

  615.     x86_reg dstW_reg = dstW;

  616.     x86_reg uv_off = c->uv_off_byte;

  617. 

  618.     YSCALEYUV2PACKEDX

  619.     YSCALEYUV2RGBX

  620.     "pxor                          %%mm7, %%mm7         \n\t"

  621.     "lea (%%"FF_REG_a", %%"FF_REG_a", 2), %%"FF_REG_c"  \n\t" // FIXME optimize

  622.     "add                              %4, %%"FF_REG_c"  \n\t"

  623.     WRITEBGR24(%%FF_REGc, %5, %%FF_REGa)

  624.     :: "r" (&c->redDither),

  625.        "m" (dummy), "m" (dummy), "m" (dummy),

  626.        "r" (dest),  "m" (dstW_reg), "m"(uv_off)

  627.     : "%"FF_REG_a, "%"FF_REG_c, "%"FF_REG_d, "%"FF_REG_S

  628.     );

  629. }

  630. 

  631. #define REAL_WRITEYUY2(dst, dstw, index) \

  632.     "packuswb  %%mm3, %%mm3     \n\t"\

  633.     "packuswb  %%mm4, %%mm4     \n\t"\

  634.     "packuswb  %%mm7, %%mm1     \n\t"\

  635.     "punpcklbw %%mm4, %%mm3     \n\t"\

  636.     "movq      %%mm1, %%mm7     \n\t"\

  637.     "punpcklbw %%mm3, %%mm1     \n\t"\

  638.     "punpckhbw %%mm3, %%mm7     \n\t"\

  639. \

  640.     MOVNTQ(%%mm1, (dst, index, 2))\

  641.     MOVNTQ(%%mm7, 8(dst, index, 2))\

  642. \

  643.     "add          $8, "#index"  \n\t"\

  644.     "cmp     "#dstw", "#index"  \n\t"\

  645.     " jb          1b            \n\t"

  646. #define WRITEYUY2(dst, dstw, index)  REAL_WRITEYUY2(dst, dstw, index)

  647. 

  648. static void RENAME(yuv2yuyv422_X_ar)(SwsContext *c, const int16_t *lumFilter,

  649.                                      const int16_t **lumSrc, int lumFilterSize,

  650.                                      const int16_t *chrFilter, const int16_t **chrUSrc,

  651.                                      const int16_t **chrVSrc,

  652.                                      int chrFilterSize, const int16_t **alpSrc,

  653.                                      uint8_t *dest, int dstW, int dstY)

  654. {

  655.     x86_reg dummy=0;

  656.     x86_reg dstW_reg = dstW;

  657.     x86_reg uv_off = c->uv_off_byte;

  658. 

  659.     YSCALEYUV2PACKEDX_ACCURATE

  660.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  661.     "psraw $3, %%mm3    \n\t"

  662.     "psraw $3, %%mm4    \n\t"

  663.     "psraw $3, %%mm1    \n\t"

  664.     "psraw $3, %%mm7    \n\t"

  665.     WRITEYUY2(%4, %5, %%FF_REGa)

  666.     YSCALEYUV2PACKEDX_END

  667. }

  668. 

  669. static void RENAME(yuv2yuyv422_X)(SwsContext *c, const int16_t *lumFilter,

  670.                                   const int16_t **lumSrc, int lumFilterSize,

  671.                                   const int16_t *chrFilter, const int16_t **chrUSrc,

  672.                                   const int16_t **chrVSrc,

  673.                                   int chrFilterSize, const int16_t **alpSrc,

  674.                                   uint8_t *dest, int dstW, int dstY)

  675. {

  676.     x86_reg dummy=0;

  677.     x86_reg dstW_reg = dstW;

  678.     x86_reg uv_off = c->uv_off_byte;

  679. 

  680.     YSCALEYUV2PACKEDX

  681.     /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  682.     "psraw $3, %%mm3    \n\t"

  683.     "psraw $3, %%mm4    \n\t"

  684.     "psraw $3, %%mm1    \n\t"

  685.     "psraw $3, %%mm7    \n\t"

  686.     WRITEYUY2(%4, %5, %%FF_REGa)

  687.     YSCALEYUV2PACKEDX_END

  688. }

  689. 

  690. #define REAL_YSCALEYUV2RGB_UV(index, c) \

  691.     "xor            "#index", "#index"  \n\t"\

  692.     ".p2align              4            \n\t"\

  693.     "1:                                 \n\t"\

  694.     "movq     (%2, "#index"), %%mm2     \n\t" /* uvbuf0[eax]*/\

  695.     "movq     (%3, "#index"), %%mm3     \n\t" /* uvbuf1[eax]*/\

  696.     "add "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  697.     "movq     (%2, "#index"), %%mm5     \n\t" /* uvbuf0[eax+2048]*/\

  698.     "movq     (%3, "#index"), %%mm4     \n\t" /* uvbuf1[eax+2048]*/\

  699.     "sub "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  700.     "psubw             %%mm3, %%mm2     \n\t" /* uvbuf0[eax] - uvbuf1[eax]*/\

  701.     "psubw             %%mm4, %%mm5     \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048]*/\

  702.     "movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0    \n\t"\

  703.     "pmulhw            %%mm0, %%mm2     \n\t" /* (uvbuf0[eax] - uvbuf1[eax])uvalpha1>>16*/\

  704.     "pmulhw            %%mm0, %%mm5     \n\t" /* (uvbuf0[eax+2048] - uvbuf1[eax+2048])uvalpha1>>16*/\

  705.     "psraw                $4, %%mm3     \n\t" /* uvbuf0[eax] - uvbuf1[eax] >>4*/\

  706.     "psraw                $4, %%mm4     \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048] >>4*/\

  707.     "paddw             %%mm2, %%mm3     \n\t" /* uvbuf0[eax]uvalpha1 - uvbuf1[eax](1-uvalpha1)*/\

  708.     "paddw             %%mm5, %%mm4     \n\t" /* uvbuf0[eax+2048]uvalpha1 - uvbuf1[eax+2048](1-uvalpha1)*/\

  709.     "psubw  "U_OFFSET"("#c"), %%mm3     \n\t" /* (U-128)8*/\

  710.     "psubw  "V_OFFSET"("#c"), %%mm4     \n\t" /* (V-128)8*/\

  711.     "movq              %%mm3, %%mm2     \n\t" /* (U-128)8*/\

  712.     "movq              %%mm4, %%mm5     \n\t" /* (V-128)8*/\

  713.     "pmulhw "UG_COEFF"("#c"), %%mm3     \n\t"\

  714.     "pmulhw "VG_COEFF"("#c"), %%mm4     \n\t"\

  715.     /* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\

  716. 

  717. #define REAL_YSCALEYUV2RGB_YA(index, c, b1, b2) \

  718.     "movq  ("#b1", "#index", 2), %%mm0     \n\t" /*buf0[eax]*/\

  719.     "movq  ("#b2", "#index", 2), %%mm1     \n\t" /*buf1[eax]*/\

  720.     "movq 8("#b1", "#index", 2), %%mm6     \n\t" /*buf0[eax]*/\

  721.     "movq 8("#b2", "#index", 2), %%mm7     \n\t" /*buf1[eax]*/\

  722.     "psubw             %%mm1, %%mm0     \n\t" /* buf0[eax] - buf1[eax]*/\

  723.     "psubw             %%mm7, %%mm6     \n\t" /* buf0[eax] - buf1[eax]*/\

  724.     "pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm0  \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\

  725.     "pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm6  \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\

  726.     "psraw                $4, %%mm1     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  727.     "psraw                $4, %%mm7     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  728.     "paddw             %%mm0, %%mm1     \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\

  729.     "paddw             %%mm6, %%mm7     \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\

  730. 

  731. #define REAL_YSCALEYUV2RGB_COEFF(c) \

  732.     "pmulhw "UB_COEFF"("#c"), %%mm2     \n\t"\

  733.     "pmulhw "VR_COEFF"("#c"), %%mm5     \n\t"\

  734.     "psubw  "Y_OFFSET"("#c"), %%mm1     \n\t" /* 8(Y-16)*/\

  735.     "psubw  "Y_OFFSET"("#c"), %%mm7     \n\t" /* 8(Y-16)*/\

  736.     "pmulhw  "Y_COEFF"("#c"), %%mm1     \n\t"\

  737.     "pmulhw  "Y_COEFF"("#c"), %%mm7     \n\t"\

  738.     /* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\

  739.     "paddw             %%mm3, %%mm4     \n\t"\

  740.     "movq              %%mm2, %%mm0     \n\t"\

  741.     "movq              %%mm5, %%mm6     \n\t"\

  742.     "movq              %%mm4, %%mm3     \n\t"\

  743.     "punpcklwd         %%mm2, %%mm2     \n\t"\

  744.     "punpcklwd         %%mm5, %%mm5     \n\t"\

  745.     "punpcklwd         %%mm4, %%mm4     \n\t"\

  746.     "paddw             %%mm1, %%mm2     \n\t"\

  747.     "paddw             %%mm1, %%mm5     \n\t"\

  748.     "paddw             %%mm1, %%mm4     \n\t"\

  749.     "punpckhwd         %%mm0, %%mm0     \n\t"\

  750.     "punpckhwd         %%mm6, %%mm6     \n\t"\

  751.     "punpckhwd         %%mm3, %%mm3     \n\t"\

  752.     "paddw             %%mm7, %%mm0     \n\t"\

  753.     "paddw             %%mm7, %%mm6     \n\t"\

  754.     "paddw             %%mm7, %%mm3     \n\t"\

  755.     /* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\

  756.     "packuswb          %%mm0, %%mm2     \n\t"\

  757.     "packuswb          %%mm6, %%mm5     \n\t"\

  758.     "packuswb          %%mm3, %%mm4     \n\t"\

  759. 

  760. #define YSCALEYUV2RGB_YA(index, c, b1, b2) REAL_YSCALEYUV2RGB_YA(index, c, b1, b2)

  761. 

  762. #define YSCALEYUV2RGB(index, c) \

  763.     REAL_YSCALEYUV2RGB_UV(index, c) \

  764.     REAL_YSCALEYUV2RGB_YA(index, c, %0, %1) \

  765.     REAL_YSCALEYUV2RGB_COEFF(c)

  766. 

  767. /**

  768.  * vertical bilinear scale YV12 to RGB

  769.  */

  770. static void RENAME(yuv2rgb32_2)(SwsContext *c, const int16_t *buf[2],

  771.                                 const int16_t *ubuf[2], const int16_t *vbuf[2],

  772.                                 const int16_t *abuf[2], uint8_t *dest,

  773.                                 int dstW, int yalpha, int uvalpha, int y)

  774. {

  775.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  776.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];

  777. 

  778.     if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {

  779.         const int16_t *abuf0 = abuf[0], *abuf1 = abuf[1];

  780. #if ARCH_X86_64

  781.         __asm__ volatile(

  782.             YSCALEYUV2RGB(%%r8, %5)

  783.             YSCALEYUV2RGB_YA(%%r8, %5, %6, %7)

  784.             "psraw                  $3, %%mm1       \n\t" /* abuf0[eax] - abuf1[eax] >>7*/

  785.             "psraw                  $3, %%mm7       \n\t" /* abuf0[eax] - abuf1[eax] >>7*/

  786.             "packuswb            %%mm7, %%mm1       \n\t"

  787.             WRITEBGR32(%4, 8280(%5), %%r8, %%mm2, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm3, %%mm6)

  788.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "r" (dest),

  789.                "a" (&c->redDither),

  790.                "r" (abuf0), "r" (abuf1)

  791.             : "%r8"

  792.         );

  793. #else

  794.         *(const uint16_t **)(&c->u_temp)=abuf0;

  795.         *(const uint16_t **)(&c->v_temp)=abuf1;

  796.         __asm__ volatile(

  797.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  798.             "mov           %4, %%"FF_REG_b"         \n\t"

  799.             "push %%"FF_REG_BP"                     \n\t"

  800.             YSCALEYUV2RGB(%%FF_REGBP, %5)

  801.             "push                   %0              \n\t"

  802.             "push                   %1              \n\t"

  803.             "mov          "U_TEMP"(%5), %0          \n\t"

  804.             "mov          "V_TEMP"(%5), %1          \n\t"

  805.             YSCALEYUV2RGB_YA(%%FF_REGBP, %5, %0, %1)

  806.             "psraw                  $3, %%mm1       \n\t" /* abuf0[eax] - abuf1[eax] >>7*/

  807.             "psraw                  $3, %%mm7       \n\t" /* abuf0[eax] - abuf1[eax] >>7*/

  808.             "packuswb            %%mm7, %%mm1       \n\t"

  809.             "pop                    %1              \n\t"

  810.             "pop                    %0              \n\t"

  811.             WRITEBGR32(%%FF_REGb, 8280(%5), %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm3, %%mm6)

  812.             "pop %%"FF_REG_BP"                      \n\t"

  813.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  814.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  815.                "a" (&c->redDither)

  816.         );

  817. #endif

  818.     } else {

  819.         __asm__ volatile(

  820.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  821.             "mov           %4, %%"FF_REG_b"         \n\t"

  822.             "push %%"FF_REG_BP"                     \n\t"

  823.             YSCALEYUV2RGB(%%FF_REGBP, %5)

  824.             "pcmpeqd %%mm7, %%mm7                   \n\t"

  825.             WRITEBGR32(%%FF_REGb, 8280(%5), %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)

  826.             "pop %%"FF_REG_BP"                      \n\t"

  827.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  828.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  829.                "a" (&c->redDither)

  830.         );

  831.     }

  832. }

  833. 

  834. static void RENAME(yuv2bgr24_2)(SwsContext *c, const int16_t *buf[2],

  835.                                 const int16_t *ubuf[2], const int16_t *vbuf[2],

  836.                                 const int16_t *abuf[2], uint8_t *dest,

  837.                                 int dstW, int yalpha, int uvalpha, int y)

  838. {

  839.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  840.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];

  841. 

  842.     //Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(

  843.     __asm__ volatile(

  844.         "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  845.         "mov           %4, %%"FF_REG_b"         \n\t"

  846.         "push %%"FF_REG_BP"                     \n\t"

  847.         YSCALEYUV2RGB(%%FF_REGBP, %5)

  848.         "pxor    %%mm7, %%mm7                   \n\t"

  849.         WRITEBGR24(%%FF_REGb, 8280(%5), %%FF_REGBP)

  850.         "pop %%"FF_REG_BP"                      \n\t"

  851.         "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  852.         :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  853.            "a" (&c->redDither)

  854.     );

  855. }

  856. 

  857. static void RENAME(yuv2rgb555_2)(SwsContext *c, const int16_t *buf[2],

  858.                                  const int16_t *ubuf[2], const int16_t *vbuf[2],

  859.                                  const int16_t *abuf[2], uint8_t *dest,

  860.                                  int dstW, int yalpha, int uvalpha, int y)

  861. {

  862.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  863.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];

  864. 

  865.     //Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(

  866.     __asm__ volatile(

  867.         "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  868.         "mov           %4, %%"FF_REG_b"         \n\t"

  869.         "push %%"FF_REG_BP"                     \n\t"

  870.         YSCALEYUV2RGB(%%FF_REGBP, %5)

  871.         "pxor    %%mm7, %%mm7                   \n\t"

  872.         /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  873. #ifdef DITHER1XBPP

  874.         "paddusb "BLUE_DITHER"(%5), %%mm2      \n\t"

  875.         "paddusb "GREEN_DITHER"(%5), %%mm4      \n\t"

  876.         "paddusb "RED_DITHER"(%5), %%mm5      \n\t"

  877. #endif

  878.         WRITERGB15(%%FF_REGb, 8280(%5), %%FF_REGBP)

  879.         "pop %%"FF_REG_BP"                      \n\t"

  880.         "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  881.         :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  882.            "a" (&c->redDither)

  883.     );

  884. }

  885. 

  886. static void RENAME(yuv2rgb565_2)(SwsContext *c, const int16_t *buf[2],

  887.                                  const int16_t *ubuf[2], const int16_t *vbuf[2],

  888.                                  const int16_t *abuf[2], uint8_t *dest,

  889.                                  int dstW, int yalpha, int uvalpha, int y)

  890. {

  891.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  892.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];

  893. 

  894.     //Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(

  895.     __asm__ volatile(

  896.         "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  897.         "mov           %4, %%"FF_REG_b"         \n\t"

  898.         "push %%"FF_REG_BP"                     \n\t"

  899.         YSCALEYUV2RGB(%%FF_REGBP, %5)

  900.         "pxor    %%mm7, %%mm7                   \n\t"

  901.         /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  902. #ifdef DITHER1XBPP

  903.         "paddusb "BLUE_DITHER"(%5), %%mm2      \n\t"

  904.         "paddusb "GREEN_DITHER"(%5), %%mm4      \n\t"

  905.         "paddusb "RED_DITHER"(%5), %%mm5      \n\t"

  906. #endif

  907.         WRITERGB16(%%FF_REGb, 8280(%5), %%FF_REGBP)

  908.         "pop %%"FF_REG_BP"                      \n\t"

  909.         "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  910.         :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  911.            "a" (&c->redDither)

  912.     );

  913. }

  914. 

  915. #define REAL_YSCALEYUV2PACKED(index, c) \

  916.     "movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0              \n\t"\

  917.     "movq "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm1              \n\t"\

  918.     "psraw                $3, %%mm0                           \n\t"\

  919.     "psraw                $3, %%mm1                           \n\t"\

  920.     "movq              %%mm0, "CHR_MMX_FILTER_OFFSET"+8("#c") \n\t"\

  921.     "movq              %%mm1, "LUM_MMX_FILTER_OFFSET"+8("#c") \n\t"\

  922.     "xor            "#index", "#index"                        \n\t"\

  923.     ".p2align              4            \n\t"\

  924.     "1:                                 \n\t"\

  925.     "movq     (%2, "#index"), %%mm2     \n\t" /* uvbuf0[eax]*/\

  926.     "movq     (%3, "#index"), %%mm3     \n\t" /* uvbuf1[eax]*/\

  927.     "add "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  928.     "movq     (%2, "#index"), %%mm5     \n\t" /* uvbuf0[eax+2048]*/\

  929.     "movq     (%3, "#index"), %%mm4     \n\t" /* uvbuf1[eax+2048]*/\

  930.     "sub "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  931.     "psubw             %%mm3, %%mm2     \n\t" /* uvbuf0[eax] - uvbuf1[eax]*/\

  932.     "psubw             %%mm4, %%mm5     \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048]*/\

  933.     "movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0    \n\t"\

  934.     "pmulhw            %%mm0, %%mm2     \n\t" /* (uvbuf0[eax] - uvbuf1[eax])uvalpha1>>16*/\

  935.     "pmulhw            %%mm0, %%mm5     \n\t" /* (uvbuf0[eax+2048] - uvbuf1[eax+2048])uvalpha1>>16*/\

  936.     "psraw                $7, %%mm3     \n\t" /* uvbuf0[eax] - uvbuf1[eax] >>4*/\

  937.     "psraw                $7, %%mm4     \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048] >>4*/\

  938.     "paddw             %%mm2, %%mm3     \n\t" /* uvbuf0[eax]uvalpha1 - uvbuf1[eax](1-uvalpha1)*/\

  939.     "paddw             %%mm5, %%mm4     \n\t" /* uvbuf0[eax+2048]uvalpha1 - uvbuf1[eax+2048](1-uvalpha1)*/\

  940.     "movq  (%0, "#index", 2), %%mm0     \n\t" /*buf0[eax]*/\

  941.     "movq  (%1, "#index", 2), %%mm1     \n\t" /*buf1[eax]*/\

  942.     "movq 8(%0, "#index", 2), %%mm6     \n\t" /*buf0[eax]*/\

  943.     "movq 8(%1, "#index", 2), %%mm7     \n\t" /*buf1[eax]*/\

  944.     "psubw             %%mm1, %%mm0     \n\t" /* buf0[eax] - buf1[eax]*/\

  945.     "psubw             %%mm7, %%mm6     \n\t" /* buf0[eax] - buf1[eax]*/\

  946.     "pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm0  \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\

  947.     "pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm6  \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\

  948.     "psraw                $7, %%mm1     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  949.     "psraw                $7, %%mm7     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  950.     "paddw             %%mm0, %%mm1     \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\

  951.     "paddw             %%mm6, %%mm7     \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\

  952. 

  953. #define YSCALEYUV2PACKED(index, c)  REAL_YSCALEYUV2PACKED(index, c)

  954. 

  955. static void RENAME(yuv2yuyv422_2)(SwsContext *c, const int16_t *buf[2],

  956.                                   const int16_t *ubuf[2], const int16_t *vbuf[2],

  957.                                   const int16_t *abuf[2], uint8_t *dest,

  958.                                   int dstW, int yalpha, int uvalpha, int y)

  959. {

  960.     const int16_t *buf0  = buf[0],  *buf1  = buf[1],

  961.                   *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];

  962. 

  963.     //Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(

  964.     __asm__ volatile(

  965.         "mov %%"FF_REG_b", "ESP_OFFSET"(%5)        \n\t"

  966.         "mov           %4, %%"FF_REG_b"            \n\t"

  967.         "push %%"FF_REG_BP"                        \n\t"

  968.         YSCALEYUV2PACKED(%%FF_REGBP, %5)

  969.         WRITEYUY2(%%FF_REGb, 8280(%5), %%FF_REGBP)

  970.         "pop %%"FF_REG_BP"                         \n\t"

  971.         "mov "ESP_OFFSET"(%5), %%"FF_REG_b"        \n\t"

  972.         :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  973.            "a" (&c->redDither)

  974.     );

  975. }

  976. 

  977. #define REAL_YSCALEYUV2RGB1(index, c) \

  978.     "xor            "#index", "#index"  \n\t"\

  979.     ".p2align              4            \n\t"\

  980.     "1:                                 \n\t"\

  981.     "movq     (%2, "#index"), %%mm3     \n\t" /* uvbuf0[eax]*/\

  982.     "add "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  983.     "movq     (%2, "#index"), %%mm4     \n\t" /* uvbuf0[eax+2048]*/\

  984.     "sub "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  985.     "psraw                $4, %%mm3     \n\t" /* uvbuf0[eax] - uvbuf1[eax] >>4*/\

  986.     "psraw                $4, %%mm4     \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048] >>4*/\

  987.     "psubw  "U_OFFSET"("#c"), %%mm3     \n\t" /* (U-128)8*/\

  988.     "psubw  "V_OFFSET"("#c"), %%mm4     \n\t" /* (V-128)8*/\

  989.     "movq              %%mm3, %%mm2     \n\t" /* (U-128)8*/\

  990.     "movq              %%mm4, %%mm5     \n\t" /* (V-128)8*/\

  991.     "pmulhw "UG_COEFF"("#c"), %%mm3     \n\t"\

  992.     "pmulhw "VG_COEFF"("#c"), %%mm4     \n\t"\

  993.     /* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\

  994.     "movq  (%0, "#index", 2), %%mm1     \n\t" /*buf0[eax]*/\

  995.     "movq 8(%0, "#index", 2), %%mm7     \n\t" /*buf0[eax]*/\

  996.     "psraw                $4, %%mm1     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  997.     "psraw                $4, %%mm7     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  998.     "pmulhw "UB_COEFF"("#c"), %%mm2     \n\t"\

  999.     "pmulhw "VR_COEFF"("#c"), %%mm5     \n\t"\

  1000.     "psubw  "Y_OFFSET"("#c"), %%mm1     \n\t" /* 8(Y-16)*/\

  1001.     "psubw  "Y_OFFSET"("#c"), %%mm7     \n\t" /* 8(Y-16)*/\

  1002.     "pmulhw  "Y_COEFF"("#c"), %%mm1     \n\t"\

  1003.     "pmulhw  "Y_COEFF"("#c"), %%mm7     \n\t"\

  1004.     /* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\

  1005.     "paddw             %%mm3, %%mm4     \n\t"\

  1006.     "movq              %%mm2, %%mm0     \n\t"\

  1007.     "movq              %%mm5, %%mm6     \n\t"\

  1008.     "movq              %%mm4, %%mm3     \n\t"\

  1009.     "punpcklwd         %%mm2, %%mm2     \n\t"\

  1010.     "punpcklwd         %%mm5, %%mm5     \n\t"\

  1011.     "punpcklwd         %%mm4, %%mm4     \n\t"\

  1012.     "paddw             %%mm1, %%mm2     \n\t"\

  1013.     "paddw             %%mm1, %%mm5     \n\t"\

  1014.     "paddw             %%mm1, %%mm4     \n\t"\

  1015.     "punpckhwd         %%mm0, %%mm0     \n\t"\

  1016.     "punpckhwd         %%mm6, %%mm6     \n\t"\

  1017.     "punpckhwd         %%mm3, %%mm3     \n\t"\

  1018.     "paddw             %%mm7, %%mm0     \n\t"\

  1019.     "paddw             %%mm7, %%mm6     \n\t"\

  1020.     "paddw             %%mm7, %%mm3     \n\t"\

  1021.     /* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\

  1022.     "packuswb          %%mm0, %%mm2     \n\t"\

  1023.     "packuswb          %%mm6, %%mm5     \n\t"\

  1024.     "packuswb          %%mm3, %%mm4     \n\t"\

  1025. 

  1026. #define YSCALEYUV2RGB1(index, c)  REAL_YSCALEYUV2RGB1(index, c)

  1027. 

  1028. // do vertical chrominance interpolation

  1029. #define REAL_YSCALEYUV2RGB1b(index, c) \

  1030.     "xor            "#index", "#index"  \n\t"\

  1031.     ".p2align              4            \n\t"\

  1032.     "1:                                 \n\t"\

  1033.     "movq     (%2, "#index"), %%mm2     \n\t" /* uvbuf0[eax]*/\

  1034.     "movq     (%3, "#index"), %%mm3     \n\t" /* uvbuf1[eax]*/\

  1035.     "add "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  1036.     "movq     (%2, "#index"), %%mm5     \n\t" /* uvbuf0[eax+2048]*/\

  1037.     "movq     (%3, "#index"), %%mm4     \n\t" /* uvbuf1[eax+2048]*/\

  1038.     "sub "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  1039.     "paddw             %%mm2, %%mm3     \n\t" /* uvbuf0[eax] + uvbuf1[eax]*/\

  1040.     "paddw             %%mm5, %%mm4     \n\t" /* uvbuf0[eax+2048] + uvbuf1[eax+2048]*/\

  1041.     "psrlw                $5, %%mm3     \n\t" /*FIXME might overflow*/\

  1042.     "psrlw                $5, %%mm4     \n\t" /*FIXME might overflow*/\

  1043.     "psubw  "U_OFFSET"("#c"), %%mm3     \n\t" /* (U-128)8*/\

  1044.     "psubw  "V_OFFSET"("#c"), %%mm4     \n\t" /* (V-128)8*/\

  1045.     "movq              %%mm3, %%mm2     \n\t" /* (U-128)8*/\

  1046.     "movq              %%mm4, %%mm5     \n\t" /* (V-128)8*/\

  1047.     "pmulhw "UG_COEFF"("#c"), %%mm3     \n\t"\

  1048.     "pmulhw "VG_COEFF"("#c"), %%mm4     \n\t"\

  1049.     /* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\

  1050.     "movq  (%0, "#index", 2), %%mm1     \n\t" /*buf0[eax]*/\

  1051.     "movq 8(%0, "#index", 2), %%mm7     \n\t" /*buf0[eax]*/\

  1052.     "psraw                $4, %%mm1     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  1053.     "psraw                $4, %%mm7     \n\t" /* buf0[eax] - buf1[eax] >>4*/\

  1054.     "pmulhw "UB_COEFF"("#c"), %%mm2     \n\t"\

  1055.     "pmulhw "VR_COEFF"("#c"), %%mm5     \n\t"\

  1056.     "psubw  "Y_OFFSET"("#c"), %%mm1     \n\t" /* 8(Y-16)*/\

  1057.     "psubw  "Y_OFFSET"("#c"), %%mm7     \n\t" /* 8(Y-16)*/\

  1058.     "pmulhw  "Y_COEFF"("#c"), %%mm1     \n\t"\

  1059.     "pmulhw  "Y_COEFF"("#c"), %%mm7     \n\t"\

  1060.     /* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\

  1061.     "paddw             %%mm3, %%mm4     \n\t"\

  1062.     "movq              %%mm2, %%mm0     \n\t"\

  1063.     "movq              %%mm5, %%mm6     \n\t"\

  1064.     "movq              %%mm4, %%mm3     \n\t"\

  1065.     "punpcklwd         %%mm2, %%mm2     \n\t"\

  1066.     "punpcklwd         %%mm5, %%mm5     \n\t"\

  1067.     "punpcklwd         %%mm4, %%mm4     \n\t"\

  1068.     "paddw             %%mm1, %%mm2     \n\t"\

  1069.     "paddw             %%mm1, %%mm5     \n\t"\

  1070.     "paddw             %%mm1, %%mm4     \n\t"\

  1071.     "punpckhwd         %%mm0, %%mm0     \n\t"\

  1072.     "punpckhwd         %%mm6, %%mm6     \n\t"\

  1073.     "punpckhwd         %%mm3, %%mm3     \n\t"\

  1074.     "paddw             %%mm7, %%mm0     \n\t"\

  1075.     "paddw             %%mm7, %%mm6     \n\t"\

  1076.     "paddw             %%mm7, %%mm3     \n\t"\

  1077.     /* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\

  1078.     "packuswb          %%mm0, %%mm2     \n\t"\

  1079.     "packuswb          %%mm6, %%mm5     \n\t"\

  1080.     "packuswb          %%mm3, %%mm4     \n\t"\

  1081. 

  1082. #define YSCALEYUV2RGB1b(index, c)  REAL_YSCALEYUV2RGB1b(index, c)

  1083. 

  1084. #define REAL_YSCALEYUV2RGB1_ALPHA(index) \

  1085.     "movq  (%1, "#index", 2), %%mm7     \n\t" /* abuf0[index  ]     */\

  1086.     "movq 8(%1, "#index", 2), %%mm1     \n\t" /* abuf0[index+4]     */\

  1087.     "psraw                $7, %%mm7     \n\t" /* abuf0[index  ] >>7 */\

  1088.     "psraw                $7, %%mm1     \n\t" /* abuf0[index+4] >>7 */\

  1089.     "packuswb          %%mm1, %%mm7     \n\t"

  1090. #define YSCALEYUV2RGB1_ALPHA(index) REAL_YSCALEYUV2RGB1_ALPHA(index)

  1091. 

  1092. /**

  1093.  * YV12 to RGB without scaling or interpolating

  1094.  */

  1095. static void RENAME(yuv2rgb32_1)(SwsContext *c, const int16_t *buf0,

  1096.                                 const int16_t *ubuf[2], const int16_t *vbuf[2],

  1097.                                 const int16_t *abuf0, uint8_t *dest,

  1098.                                 int dstW, int uvalpha, int y)

  1099. {

  1100.     const int16_t *ubuf0 = ubuf[0];

  1101.     const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1

  1102. 

  1103.     if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster

  1104.         const int16_t *ubuf1 = ubuf[0];

  1105.         if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {

  1106.             __asm__ volatile(

  1107.                 "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1108.                 "mov           %4, %%"FF_REG_b"         \n\t"

  1109.                 "push %%"FF_REG_BP"                     \n\t"

  1110.                 YSCALEYUV2RGB1(%%FF_REGBP, %5)

  1111.                 YSCALEYUV2RGB1_ALPHA(%%FF_REGBP)

  1112.                 WRITEBGR32(%%FF_REGb, 8280(%5), %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)

  1113.                 "pop %%"FF_REG_BP"                      \n\t"

  1114.                 "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1115.                 :: "c" (buf0), "d" (abuf0), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1116.                    "a" (&c->redDither)

  1117.             );

  1118.         } else {

  1119.             __asm__ volatile(

  1120.                 "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1121.                 "mov           %4, %%"FF_REG_b"         \n\t"

  1122.                 "push %%"FF_REG_BP"                     \n\t"

  1123.                 YSCALEYUV2RGB1(%%FF_REGBP, %5)

  1124.                 "pcmpeqd %%mm7, %%mm7                   \n\t"

  1125.                 WRITEBGR32(%%FF_REGb, 8280(%5), %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)

  1126.                 "pop %%"FF_REG_BP"                      \n\t"

  1127.                 "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1128.                 :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1129.                    "a" (&c->redDither)

  1130.             );

  1131.         }

  1132.     } else {

  1133.         const int16_t *ubuf1 = ubuf[1];

  1134.         if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {

  1135.             __asm__ volatile(

  1136.                 "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1137.                 "mov           %4, %%"FF_REG_b"         \n\t"

  1138.                 "push %%"FF_REG_BP"                     \n\t"

  1139.                 YSCALEYUV2RGB1b(%%FF_REGBP, %5)

  1140.                 YSCALEYUV2RGB1_ALPHA(%%FF_REGBP)

  1141.                 WRITEBGR32(%%FF_REGb, 8280(%5), %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)

  1142.                 "pop %%"FF_REG_BP"                      \n\t"

  1143.                 "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1144.                 :: "c" (buf0), "d" (abuf0), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1145.                    "a" (&c->redDither)

  1146.             );

  1147.         } else {

  1148.             __asm__ volatile(

  1149.                 "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1150.                 "mov           %4, %%"FF_REG_b"         \n\t"

  1151.                 "push %%"FF_REG_BP"                     \n\t"

  1152.                 YSCALEYUV2RGB1b(%%FF_REGBP, %5)

  1153.                 "pcmpeqd %%mm7, %%mm7                   \n\t"

  1154.                 WRITEBGR32(%%FF_REGb, 8280(%5), %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)

  1155.                 "pop %%"FF_REG_BP"                      \n\t"

  1156.                 "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1157.                 :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1158.                    "a" (&c->redDither)

  1159.             );

  1160.         }

  1161.     }

  1162. }

  1163. 

  1164. static void RENAME(yuv2bgr24_1)(SwsContext *c, const int16_t *buf0,

  1165.                                 const int16_t *ubuf[2], const int16_t *vbuf[2],

  1166.                                 const int16_t *abuf0, uint8_t *dest,

  1167.                                 int dstW, int uvalpha, int y)

  1168. {

  1169.     const int16_t *ubuf0 = ubuf[0];

  1170.     const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1

  1171. 

  1172.     if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster

  1173.         const int16_t *ubuf1 = ubuf[0];

  1174.         __asm__ volatile(

  1175.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1176.             "mov           %4, %%"FF_REG_b"         \n\t"

  1177.             "push %%"FF_REG_BP"                     \n\t"

  1178.             YSCALEYUV2RGB1(%%FF_REGBP, %5)

  1179.             "pxor    %%mm7, %%mm7                   \n\t"

  1180.             WRITEBGR24(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1181.             "pop %%"FF_REG_BP"                      \n\t"

  1182.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1183.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1184.                "a" (&c->redDither)

  1185.         );

  1186.     } else {

  1187.         const int16_t *ubuf1 = ubuf[1];

  1188.         __asm__ volatile(

  1189.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1190.             "mov           %4, %%"FF_REG_b"         \n\t"

  1191.             "push %%"FF_REG_BP"                     \n\t"

  1192.             YSCALEYUV2RGB1b(%%FF_REGBP, %5)

  1193.             "pxor    %%mm7, %%mm7                   \n\t"

  1194.             WRITEBGR24(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1195.             "pop %%"FF_REG_BP"                      \n\t"

  1196.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1197.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1198.                "a" (&c->redDither)

  1199.         );

  1200.     }

  1201. }

  1202. 

  1203. static void RENAME(yuv2rgb555_1)(SwsContext *c, const int16_t *buf0,

  1204.                                  const int16_t *ubuf[2], const int16_t *vbuf[2],

  1205.                                  const int16_t *abuf0, uint8_t *dest,

  1206.                                  int dstW, int uvalpha, int y)

  1207. {

  1208.     const int16_t *ubuf0 = ubuf[0];

  1209.     const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1

  1210. 

  1211.     if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster

  1212.         const int16_t *ubuf1 = ubuf[0];

  1213.         __asm__ volatile(

  1214.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1215.             "mov           %4, %%"FF_REG_b"         \n\t"

  1216.             "push %%"FF_REG_BP"                     \n\t"

  1217.             YSCALEYUV2RGB1(%%FF_REGBP, %5)

  1218.             "pxor    %%mm7, %%mm7                   \n\t"

  1219.             /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  1220. #ifdef DITHER1XBPP

  1221.             "paddusb "BLUE_DITHER"(%5), %%mm2      \n\t"

  1222.             "paddusb "GREEN_DITHER"(%5), %%mm4      \n\t"

  1223.             "paddusb "RED_DITHER"(%5), %%mm5      \n\t"

  1224. #endif

  1225.             WRITERGB15(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1226.             "pop %%"FF_REG_BP"                      \n\t"

  1227.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1228.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1229.                "a" (&c->redDither)

  1230.         );

  1231.     } else {

  1232.         const int16_t *ubuf1 = ubuf[1];

  1233.         __asm__ volatile(

  1234.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1235.             "mov           %4, %%"FF_REG_b"         \n\t"

  1236.             "push %%"FF_REG_BP"                     \n\t"

  1237.             YSCALEYUV2RGB1b(%%FF_REGBP, %5)

  1238.             "pxor    %%mm7, %%mm7                   \n\t"

  1239.             /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  1240. #ifdef DITHER1XBPP

  1241.             "paddusb "BLUE_DITHER"(%5), %%mm2      \n\t"

  1242.             "paddusb "GREEN_DITHER"(%5), %%mm4      \n\t"

  1243.             "paddusb "RED_DITHER"(%5), %%mm5      \n\t"

  1244. #endif

  1245.             WRITERGB15(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1246.             "pop %%"FF_REG_BP"                      \n\t"

  1247.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1248.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1249.                "a" (&c->redDither)

  1250.         );

  1251.     }

  1252. }

  1253. 

  1254. static void RENAME(yuv2rgb565_1)(SwsContext *c, const int16_t *buf0,

  1255.                                  const int16_t *ubuf[2], const int16_t *vbuf[2],

  1256.                                  const int16_t *abuf0, uint8_t *dest,

  1257.                                  int dstW, int uvalpha, int y)

  1258. {

  1259.     const int16_t *ubuf0 = ubuf[0];

  1260.     const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1

  1261. 

  1262.     if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster

  1263.         const int16_t *ubuf1 = ubuf[0];

  1264.         __asm__ volatile(

  1265.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1266.             "mov           %4, %%"FF_REG_b"         \n\t"

  1267.             "push %%"FF_REG_BP"                     \n\t"

  1268.             YSCALEYUV2RGB1(%%FF_REGBP, %5)

  1269.             "pxor    %%mm7, %%mm7                   \n\t"

  1270.             /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  1271. #ifdef DITHER1XBPP

  1272.             "paddusb "BLUE_DITHER"(%5), %%mm2      \n\t"

  1273.             "paddusb "GREEN_DITHER"(%5), %%mm4      \n\t"

  1274.             "paddusb "RED_DITHER"(%5), %%mm5      \n\t"

  1275. #endif

  1276.             WRITERGB16(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1277.             "pop %%"FF_REG_BP"                      \n\t"

  1278.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1279.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1280.                "a" (&c->redDither)

  1281.         );

  1282.     } else {

  1283.         const int16_t *ubuf1 = ubuf[1];

  1284.         __asm__ volatile(

  1285.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1286.             "mov           %4, %%"FF_REG_b"         \n\t"

  1287.             "push %%"FF_REG_BP"                     \n\t"

  1288.             YSCALEYUV2RGB1b(%%FF_REGBP, %5)

  1289.             "pxor    %%mm7, %%mm7                   \n\t"

  1290.             /* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */

  1291. #ifdef DITHER1XBPP

  1292.             "paddusb "BLUE_DITHER"(%5), %%mm2      \n\t"

  1293.             "paddusb "GREEN_DITHER"(%5), %%mm4      \n\t"

  1294.             "paddusb "RED_DITHER"(%5), %%mm5      \n\t"

  1295. #endif

  1296.             WRITERGB16(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1297.             "pop %%"FF_REG_BP"                      \n\t"

  1298.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1299.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1300.                "a" (&c->redDither)

  1301.         );

  1302.     }

  1303. }

  1304. 

  1305. #define REAL_YSCALEYUV2PACKED1(index, c) \

  1306.     "xor            "#index", "#index"  \n\t"\

  1307.     ".p2align              4            \n\t"\

  1308.     "1:                                 \n\t"\

  1309.     "movq     (%2, "#index"), %%mm3     \n\t" /* uvbuf0[eax]*/\

  1310.     "add "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  1311.     "movq     (%2, "#index"), %%mm4     \n\t" /* uvbuf0[eax+2048]*/\

  1312.     "sub "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  1313.     "psraw                $7, %%mm3     \n\t" \

  1314.     "psraw                $7, %%mm4     \n\t" \

  1315.     "movq  (%0, "#index", 2), %%mm1     \n\t" /*buf0[eax]*/\

  1316.     "movq 8(%0, "#index", 2), %%mm7     \n\t" /*buf0[eax]*/\

  1317.     "psraw                $7, %%mm1     \n\t" \

  1318.     "psraw                $7, %%mm7     \n\t" \

  1319. 

  1320. #define YSCALEYUV2PACKED1(index, c)  REAL_YSCALEYUV2PACKED1(index, c)

  1321. 

  1322. #define REAL_YSCALEYUV2PACKED1b(index, c) \

  1323.     "xor "#index", "#index"             \n\t"\

  1324.     ".p2align              4            \n\t"\

  1325.     "1:                                 \n\t"\

  1326.     "movq     (%2, "#index"), %%mm2     \n\t" /* uvbuf0[eax]*/\

  1327.     "movq     (%3, "#index"), %%mm3     \n\t" /* uvbuf1[eax]*/\

  1328.     "add "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  1329.     "movq     (%2, "#index"), %%mm5     \n\t" /* uvbuf0[eax+2048]*/\

  1330.     "movq     (%3, "#index"), %%mm4     \n\t" /* uvbuf1[eax+2048]*/\

  1331.     "sub "UV_OFF_BYTE"("#c"), "#index"  \n\t" \

  1332.     "paddw             %%mm2, %%mm3     \n\t" /* uvbuf0[eax] + uvbuf1[eax]*/\

  1333.     "paddw             %%mm5, %%mm4     \n\t" /* uvbuf0[eax+2048] + uvbuf1[eax+2048]*/\

  1334.     "psrlw                $8, %%mm3     \n\t" \

  1335.     "psrlw                $8, %%mm4     \n\t" \

  1336.     "movq  (%0, "#index", 2), %%mm1     \n\t" /*buf0[eax]*/\

  1337.     "movq 8(%0, "#index", 2), %%mm7     \n\t" /*buf0[eax]*/\

  1338.     "psraw                $7, %%mm1     \n\t" \

  1339.     "psraw                $7, %%mm7     \n\t"

  1340. #define YSCALEYUV2PACKED1b(index, c)  REAL_YSCALEYUV2PACKED1b(index, c)

  1341. 

  1342. static void RENAME(yuv2yuyv422_1)(SwsContext *c, const int16_t *buf0,

  1343.                                   const int16_t *ubuf[2], const int16_t *vbuf[2],

  1344.                                   const int16_t *abuf0, uint8_t *dest,

  1345.                                   int dstW, int uvalpha, int y)

  1346. {

  1347.     const int16_t *ubuf0 = ubuf[0];

  1348.     const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1

  1349. 

  1350.     if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster

  1351.         const int16_t *ubuf1 = ubuf[0];

  1352.         __asm__ volatile(

  1353.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1354.             "mov           %4, %%"FF_REG_b"         \n\t"

  1355.             "push %%"FF_REG_BP"                     \n\t"

  1356.             YSCALEYUV2PACKED1(%%FF_REGBP, %5)

  1357.             WRITEYUY2(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1358.             "pop %%"FF_REG_BP"                      \n\t"

  1359.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1360.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1361.                "a" (&c->redDither)

  1362.         );

  1363.     } else {

  1364.         const int16_t *ubuf1 = ubuf[1];

  1365.         __asm__ volatile(

  1366.             "mov %%"FF_REG_b", "ESP_OFFSET"(%5)     \n\t"

  1367.             "mov           %4, %%"FF_REG_b"         \n\t"

  1368.             "push %%"FF_REG_BP"                     \n\t"

  1369.             YSCALEYUV2PACKED1b(%%FF_REGBP, %5)

  1370.             WRITEYUY2(%%FF_REGb, 8280(%5), %%FF_REGBP)

  1371.             "pop %%"FF_REG_BP"                      \n\t"

  1372.             "mov "ESP_OFFSET"(%5), %%"FF_REG_b"     \n\t"

  1373.             :: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),

  1374.                "a" (&c->redDither)

  1375.         );

  1376.     }

  1377. }

  1378. 

  1379. #if COMPILE_TEMPLATE_MMXEXT

  1380. static void RENAME(hyscale_fast)(SwsContext *c, int16_t *dst,

  1381.                                  int dstWidth, const uint8_t *src,

  1382.                                  int srcW, int xInc)

  1383. {

  1384.     int32_t *filterPos = c->hLumFilterPos;

  1385.     int16_t *filter    = c->hLumFilter;

  1386.     void    *mmxextFilterCode = c->lumMmxextFilterCode;

  1387.     int i;

  1388. #if defined(PIC)

  1389.     uint64_t ebxsave;

  1390. #endif

  1391. #if ARCH_X86_64

  1392.     uint64_t retsave;

  1393. #endif

  1394. 

  1395.     __asm__ volatile(

  1396. #if defined(PIC)

  1397.         "mov            %%"FF_REG_b", %5            \n\t"

  1398. #if ARCH_X86_64

  1399.         "mov               -8(%%rsp), %%"FF_REG_a"  \n\t"

  1400.         "mov            %%"FF_REG_a", %6            \n\t"

  1401. #endif

  1402. #else

  1403. #if ARCH_X86_64

  1404.         "mov               -8(%%rsp), %%"FF_REG_a" \n\t"

  1405.         "mov            %%"FF_REG_a", %5           \n\t"

  1406. #endif

  1407. #endif

  1408.         "pxor                  %%mm7, %%mm7        \n\t"

  1409.         "mov                      %0, %%"FF_REG_c" \n\t"

  1410.         "mov                      %1, %%"FF_REG_D" \n\t"

  1411.         "mov                      %2, %%"FF_REG_d" \n\t"

  1412.         "mov                      %3, %%"FF_REG_b" \n\t"

  1413.         "xor            %%"FF_REG_a", %%"FF_REG_a" \n\t" // i

  1414.         PREFETCH"     (%%"FF_REG_c")               \n\t"

  1415.         PREFETCH"   32(%%"FF_REG_c")               \n\t"

  1416.         PREFETCH"   64(%%"FF_REG_c")               \n\t"

  1417. 

  1418. #if ARCH_X86_64

  1419. #define CALL_MMXEXT_FILTER_CODE \

  1420.         "movl               (%%"FF_REG_b"), %%esi        \n\t"\

  1421.         "call                          *%4               \n\t"\

  1422.         "movl (%%"FF_REG_b", %%"FF_REG_a"), %%esi        \n\t"\

  1423.         "add                  %%"FF_REG_S", %%"FF_REG_c" \n\t"\

  1424.         "add                  %%"FF_REG_a", %%"FF_REG_D" \n\t"\

  1425.         "xor                  %%"FF_REG_a", %%"FF_REG_a" \n\t"\

  1426. 

  1427. #else

  1428. #define CALL_MMXEXT_FILTER_CODE \

  1429.         "movl               (%%"FF_REG_b"), %%esi        \n\t"\

  1430.         "call                          *%4               \n\t"\

  1431.         "addl (%%"FF_REG_b", %%"FF_REG_a"), %%"FF_REG_c" \n\t"\

  1432.         "add                  %%"FF_REG_a", %%"FF_REG_D" \n\t"\

  1433.         "xor                  %%"FF_REG_a", %%"FF_REG_a" \n\t"\

  1434. 

  1435. #endif /* ARCH_X86_64 */

  1436. 

  1437.         CALL_MMXEXT_FILTER_CODE

  1438.         CALL_MMXEXT_FILTER_CODE

  1439.         CALL_MMXEXT_FILTER_CODE

  1440.         CALL_MMXEXT_FILTER_CODE

  1441.         CALL_MMXEXT_FILTER_CODE

  1442.         CALL_MMXEXT_FILTER_CODE

  1443.         CALL_MMXEXT_FILTER_CODE

  1444.         CALL_MMXEXT_FILTER_CODE

  1445. 

  1446. #if defined(PIC)

  1447.         "mov                      %5, %%"FF_REG_b"  \n\t"

  1448. #if ARCH_X86_64

  1449.         "mov                      %6, %%"FF_REG_a"  \n\t"

  1450.         "mov            %%"FF_REG_a", -8(%%rsp)     \n\t"

  1451. #endif

  1452. #else

  1453. #if ARCH_X86_64

  1454.         "mov                      %5, %%"FF_REG_a"  \n\t"

  1455.         "mov            %%"FF_REG_a", -8(%%rsp)     \n\t"

  1456. #endif

  1457. #endif

  1458.         :: "m" (src), "m" (dst), "m" (filter), "m" (filterPos),

  1459.            "m" (mmxextFilterCode)

  1460. #if defined(PIC)

  1461.           ,"m" (ebxsave)

  1462. #endif

  1463. #if ARCH_X86_64

  1464.           ,"m"(retsave)

  1465. #endif

  1466.         : "%"FF_REG_a, "%"FF_REG_c, "%"FF_REG_d, "%"FF_REG_S, "%"FF_REG_D

  1467. #if !defined(PIC)

  1468.          ,"%"FF_REG_b

  1469. #endif

  1470.     );

  1471. 

  1472.     for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--)

  1473.         dst[i] = src[srcW-1]*128;

  1474. }

  1475. 

  1476. static void RENAME(hcscale_fast)(SwsContext *c, int16_t *dst1, int16_t *dst2,

  1477.                                  int dstWidth, const uint8_t *src1,

  1478.                                  const uint8_t *src2, int srcW, int xInc)

  1479. {

  1480.     int32_t *filterPos = c->hChrFilterPos;

  1481.     int16_t *filter    = c->hChrFilter;

  1482.     void    *mmxextFilterCode = c->chrMmxextFilterCode;

  1483.     int i;

  1484. #if defined(PIC)

  1485.     DECLARE_ALIGNED(8, uint64_t, ebxsave);

  1486. #endif

  1487. #if ARCH_X86_64

  1488.     DECLARE_ALIGNED(8, uint64_t, retsave);

  1489. #endif

  1490. 

  1491.     __asm__ volatile(

  1492. #if defined(PIC)

  1493.         "mov       %%"FF_REG_b", %7             \n\t"

  1494. #if ARCH_X86_64

  1495.         "mov          -8(%%rsp), %%"FF_REG_a"   \n\t"

  1496.         "mov       %%"FF_REG_a", %8             \n\t"

  1497. #endif

  1498. #else

  1499. #if ARCH_X86_64

  1500.         "mov          -8(%%rsp), %%"FF_REG_a"   \n\t"

  1501.         "mov       %%"FF_REG_a", %7             \n\t"

  1502. #endif

  1503. #endif

  1504.         "pxor             %%mm7, %%mm7          \n\t"

  1505.         "mov                 %0, %%"FF_REG_c"   \n\t"

  1506.         "mov                 %1, %%"FF_REG_D"   \n\t"

  1507.         "mov                 %2, %%"FF_REG_d"   \n\t"

  1508.         "mov                 %3, %%"FF_REG_b"   \n\t"

  1509.         "xor       %%"FF_REG_a", %%"FF_REG_a"   \n\t" // i

  1510.         PREFETCH"   (%%"FF_REG_c")              \n\t"

  1511.         PREFETCH" 32(%%"FF_REG_c")              \n\t"

  1512.         PREFETCH" 64(%%"FF_REG_c")              \n\t"

  1513. 

  1514.         CALL_MMXEXT_FILTER_CODE

  1515.         CALL_MMXEXT_FILTER_CODE

  1516.         CALL_MMXEXT_FILTER_CODE

  1517.         CALL_MMXEXT_FILTER_CODE

  1518.         "xor       %%"FF_REG_a", %%"FF_REG_a"   \n\t" // i

  1519.         "mov                 %5, %%"FF_REG_c"   \n\t" // src

  1520.         "mov                 %6, %%"FF_REG_D"   \n\t" // buf2

  1521.         PREFETCH"   (%%"FF_REG_c")              \n\t"

  1522.         PREFETCH" 32(%%"FF_REG_c")              \n\t"

  1523.         PREFETCH" 64(%%"FF_REG_c")              \n\t"

  1524. 

  1525.         CALL_MMXEXT_FILTER_CODE

  1526.         CALL_MMXEXT_FILTER_CODE

  1527.         CALL_MMXEXT_FILTER_CODE

  1528.         CALL_MMXEXT_FILTER_CODE

  1529. 

  1530. #if defined(PIC)

  1531.         "mov                 %7, %%"FF_REG_b"   \n\t"

  1532. #if ARCH_X86_64

  1533.         "mov                 %8, %%"FF_REG_a"   \n\t"

  1534.         "mov       %%"FF_REG_a", -8(%%rsp)      \n\t"

  1535. #endif

  1536. #else

  1537. #if ARCH_X86_64

  1538.         "mov                 %7, %%"FF_REG_a"   \n\t"

  1539.         "mov       %%"FF_REG_a", -8(%%rsp)      \n\t"

  1540. #endif

  1541. #endif

  1542.         :: "m" (src1), "m" (dst1), "m" (filter), "m" (filterPos),

  1543.            "m" (mmxextFilterCode), "m" (src2), "m"(dst2)

  1544. #if defined(PIC)

  1545.           ,"m" (ebxsave)

  1546. #endif

  1547. #if ARCH_X86_64

  1548.           ,"m"(retsave)

  1549. #endif

  1550.         : "%"FF_REG_a, "%"FF_REG_c, "%"FF_REG_d, "%"FF_REG_S, "%"FF_REG_D

  1551. #if !defined(PIC)

  1552.          ,"%"FF_REG_b

  1553. #endif

  1554.     );

  1555. 

  1556.     for (i=dstWidth-1; (i*xInc)>>16 >=srcW-1; i--) {

  1557.         dst1[i] = src1[srcW-1]*128;

  1558.         dst2[i] = src2[srcW-1]*128;

  1559.     }

  1560. }

  1561. #endif /* COMPILE_TEMPLATE_MMXEXT */

  1562. 

  1563. static av_cold void RENAME(sws_init_swscale)(SwsContext *c)

  1564. {

  1565.     enum AVPixelFormat dstFormat = c->dstFormat;

  1566. 

  1567.     if (!is16BPS(dstFormat) && !is9_15BPS(dstFormat) &&

  1568.         dstFormat != AV_PIX_FMT_NV12 && dstFormat != AV_PIX_FMT_NV21) {

  1569.         if (!(c->flags & SWS_BITEXACT)) {

  1570.             if (c->flags & SWS_ACCURATE_RND) {

  1571.                 if (!(c->flags & SWS_FULL_CHR_H_INT)) {

  1572.                     switch (c->dstFormat) {

  1573.                     case AV_PIX_FMT_RGB32:   c->yuv2packedX = RENAME(yuv2rgb32_X_ar);   break;

  1574.                     case AV_PIX_FMT_BGR24:   c->yuv2packedX = RENAME(yuv2bgr24_X_ar);   break;

  1575.                     case AV_PIX_FMT_RGB555:  c->yuv2packedX = RENAME(yuv2rgb555_X_ar);  break;

  1576.                     case AV_PIX_FMT_RGB565:  c->yuv2packedX = RENAME(yuv2rgb565_X_ar);  break;

  1577.                     case AV_PIX_FMT_YUYV422: c->yuv2packedX = RENAME(yuv2yuyv422_X_ar); break;

  1578.                     default: break;

  1579.                     }

  1580.                 }

  1581.             } else {

  1582.                 if (!(c->flags & SWS_FULL_CHR_H_INT)) {

  1583.                     switch (c->dstFormat) {

  1584.                     case AV_PIX_FMT_RGB32:   c->yuv2packedX = RENAME(yuv2rgb32_X);   break;

  1585.                     case AV_PIX_FMT_BGR24:   c->yuv2packedX = RENAME(yuv2bgr24_X);   break;

  1586.                     case AV_PIX_FMT_RGB555:  c->yuv2packedX = RENAME(yuv2rgb555_X);  break;

  1587.                     case AV_PIX_FMT_RGB565:  c->yuv2packedX = RENAME(yuv2rgb565_X);  break;

  1588.                     case AV_PIX_FMT_YUYV422: c->yuv2packedX = RENAME(yuv2yuyv422_X); break;

  1589.                     default: break;

  1590.                     }

  1591.                 }

  1592.             }

  1593.         }

  1594.         if (!(c->flags & SWS_FULL_CHR_H_INT)) {

  1595.             switch (c->dstFormat) {

  1596.             case AV_PIX_FMT_RGB32:

  1597.                 c->yuv2packed1 = RENAME(yuv2rgb32_1);

  1598.                 c->yuv2packed2 = RENAME(yuv2rgb32_2);

  1599.                 break;

  1600.             case AV_PIX_FMT_BGR24:

  1601.                 c->yuv2packed1 = RENAME(yuv2bgr24_1);

  1602.                 c->yuv2packed2 = RENAME(yuv2bgr24_2);

  1603.                 break;

  1604.             case AV_PIX_FMT_RGB555:

  1605.                 c->yuv2packed1 = RENAME(yuv2rgb555_1);

  1606.                 c->yuv2packed2 = RENAME(yuv2rgb555_2);

  1607.                 break;

  1608.             case AV_PIX_FMT_RGB565:

  1609.                 c->yuv2packed1 = RENAME(yuv2rgb565_1);

  1610.                 c->yuv2packed2 = RENAME(yuv2rgb565_2);

  1611.                 break;

  1612.             case AV_PIX_FMT_YUYV422:

  1613.                 c->yuv2packed1 = RENAME(yuv2yuyv422_1);

  1614.                 c->yuv2packed2 = RENAME(yuv2yuyv422_2);

  1615.                 break;

  1616.             default:

  1617.                 break;

  1618.             }

  1619.         }

  1620.     }

  1621. 

  1622.     if (c->srcBpc == 8 && c->dstBpc <= 12) {

  1623.     // Use the new MMX scaler if the MMXEXT one can't be used (it is faster than the x86 ASM one).

  1624. #if COMPILE_TEMPLATE_MMXEXT

  1625.     if (c->flags & SWS_FAST_BILINEAR && c->canMMXEXTBeUsed) {

  1626.         c->hyscale_fast = RENAME(hyscale_fast);

  1627.         c->hcscale_fast = RENAME(hcscale_fast);

  1628.     } else {

  1629. #endif /* COMPILE_TEMPLATE_MMXEXT */

  1630.         c->hyscale_fast = NULL;

  1631.         c->hcscale_fast = NULL;

  1632. #if COMPILE_TEMPLATE_MMXEXT

  1633.     }

  1634. #endif /* COMPILE_TEMPLATE_MMXEXT */

  1635.     }

  1636. }