x86: xvid: K&R formatting cosmetics
authorDiego Biurrun <diego@biurrun.de>
Wed, 27 Aug 2014 00:58:08 +0000 (02:58 +0200)
committerDiego Biurrun <diego@biurrun.de>
Wed, 27 Aug 2014 12:58:04 +0000 (05:58 -0700)
libavcodec/x86/xvididct_mmx.c
libavcodec/x86/xvididct_sse2.c

dissimilarity index 82%
index 7fa7f71..e371142 100644 (file)
-/*
- * XVID MPEG-4 VIDEO CODEC
- * - MMX and XMM forward discrete cosine transform -
- *
- * Copyright(C) 2001 Peter Ross <pross@xvid.org>
- *
- * Originally provided by Intel at AP-922
- * http://developer.intel.com/vtune/cbts/strmsimd/922down.htm
- * (See more app notes at http://developer.intel.com/vtune/cbts/strmsimd/appnotes.htm)
- * but in a limited edition.
- * New macro implements a column part for precise iDCT
- * The routine precision now satisfies IEEE standard 1180-1990.
- *
- * Copyright(C) 2000-2001 Peter Gubanov <peter@elecard.net.ru>
- * Rounding trick Copyright(C) 2000 Michel Lespinasse <walken@zoy.org>
- *
- * http://www.elecard.com/peter/idct.html
- * http://www.linuxvideo.org/mpeg2dec/
- *
- * These examples contain code fragments for first stage iDCT 8x8
- * (for rows) and first stage DCT 8x8 (for columns)
- *
- * conversion to gcc syntax by Michael Niedermayer
- *
- * This file is part of Libav.
- *
- * Libav is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * Libav is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public License
- * along with Libav; if not, write to the Free Software Foundation,
- * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include <inttypes.h>
-
-#include "config.h"
-#include "libavcodec/avcodec.h"
-#include "libavutil/mem.h"
-
-#include "idctdsp.h"
-#include "xvididct.h"
-
-#if HAVE_MMX_INLINE
-
-//-----------------------------------------------------------------------------
-// Various memory constants (trigonometric values or rounding values)
-//-----------------------------------------------------------------------------
-
-
-DECLARE_ALIGNED(8, static const int16_t, tg_1_16)[4*4] = {
-  13036,13036,13036,13036,        // tg * (2<<16) + 0.5
-  27146,27146,27146,27146,        // tg * (2<<16) + 0.5
-  -21746,-21746,-21746,-21746,    // tg * (2<<16) + 0.5
-  23170,23170,23170,23170};       // cos * (2<<15) + 0.5
-
-DECLARE_ALIGNED(8, static const int32_t, rounder_0)[2*8] = {
-  65536,65536,
-  3597,3597,
-  2260,2260,
-  1203,1203,
-  0,0,
-  120,120,
-  512,512,
-  512,512};
-
-//-----------------------------------------------------------------------------
-//
-// The first stage iDCT 8x8 - inverse DCTs of rows
-//
-//-----------------------------------------------------------------------------
-// The 8-point inverse DCT direct algorithm
-//-----------------------------------------------------------------------------
-//
-// static const short w[32] = {
-//       FIX(cos_4_16),  FIX(cos_2_16),  FIX(cos_4_16),  FIX(cos_6_16),
-//       FIX(cos_4_16),  FIX(cos_6_16), -FIX(cos_4_16), -FIX(cos_2_16),
-//       FIX(cos_4_16), -FIX(cos_6_16), -FIX(cos_4_16),  FIX(cos_2_16),
-//       FIX(cos_4_16), -FIX(cos_2_16),  FIX(cos_4_16), -FIX(cos_6_16),
-//       FIX(cos_1_16),  FIX(cos_3_16),  FIX(cos_5_16),  FIX(cos_7_16),
-//       FIX(cos_3_16), -FIX(cos_7_16), -FIX(cos_1_16), -FIX(cos_5_16),
-//       FIX(cos_5_16), -FIX(cos_1_16),  FIX(cos_7_16),  FIX(cos_3_16),
-//       FIX(cos_7_16), -FIX(cos_5_16),  FIX(cos_3_16), -FIX(cos_1_16) };
-//
-// #define DCT_8_INV_ROW(x, y)
-// {
-//       int a0, a1, a2, a3, b0, b1, b2, b3;
-//
-//       a0 =x[0]*w[0]+x[2]*w[1]+x[4]*w[2]+x[6]*w[3];
-//       a1 =x[0]*w[4]+x[2]*w[5]+x[4]*w[6]+x[6]*w[7];
-//       a2 = x[0] * w[ 8] + x[2] * w[ 9] + x[4] * w[10] + x[6] * w[11];
-//       a3 = x[0] * w[12] + x[2] * w[13] + x[4] * w[14] + x[6] * w[15];
-//       b0 = x[1] * w[16] + x[3] * w[17] + x[5] * w[18] + x[7] * w[19];
-//       b1 = x[1] * w[20] + x[3] * w[21] + x[5] * w[22] + x[7] * w[23];
-//       b2 = x[1] * w[24] + x[3] * w[25] + x[5] * w[26] + x[7] * w[27];
-//       b3 = x[1] * w[28] + x[3] * w[29] + x[5] * w[30] + x[7] * w[31];
-//
-//       y[0] = SHIFT_ROUND ( a0 + b0 );
-//       y[1] = SHIFT_ROUND ( a1 + b1 );
-//       y[2] = SHIFT_ROUND ( a2 + b2 );
-//       y[3] = SHIFT_ROUND ( a3 + b3 );
-//       y[4] = SHIFT_ROUND ( a3 - b3 );
-//       y[5] = SHIFT_ROUND ( a2 - b2 );
-//       y[6] = SHIFT_ROUND ( a1 - b1 );
-//       y[7] = SHIFT_ROUND ( a0 - b0 );
-// }
-//
-//-----------------------------------------------------------------------------
-//
-// In this implementation the outputs of the iDCT-1D are multiplied
-//       for rows 0,4 - by cos_4_16,
-//       for rows 1,7 - by cos_1_16,
-//       for rows 2,6 - by cos_2_16,
-//       for rows 3,5 - by cos_3_16
-// and are shifted to the left for better accuracy
-//
-// For the constants used,
-//       FIX(float_const) = (short) (float_const * (1<<15) + 0.5)
-//
-//-----------------------------------------------------------------------------
-
-//-----------------------------------------------------------------------------
-// Tables for mmx processors
-//-----------------------------------------------------------------------------
-
-// Table for rows 0,4 - constants are multiplied by cos_4_16
-DECLARE_ALIGNED(8, static const int16_t, tab_i_04_mmx)[32*4] = {
-  16384,16384,16384,-16384,       // movq-> w06 w04 w02 w00
-  21407,8867,8867,-21407,         // w07 w05 w03 w01
-  16384,-16384,16384,16384,       // w14 w12 w10 w08
-  -8867,21407,-21407,-8867,       // w15 w13 w11 w09
-  22725,12873,19266,-22725,       // w22 w20 w18 w16
-  19266,4520,-4520,-12873,        // w23 w21 w19 w17
-  12873,4520,4520,19266,          // w30 w28 w26 w24
-  -22725,19266,-12873,-22725,     // w31 w29 w27 w25
-// Table for rows 1,7 - constants are multiplied by cos_1_16
-  22725,22725,22725,-22725,       // movq-> w06 w04 w02 w00
-  29692,12299,12299,-29692,       // w07 w05 w03 w01
-  22725,-22725,22725,22725,       // w14 w12 w10 w08
-  -12299,29692,-29692,-12299,     // w15 w13 w11 w09
-  31521,17855,26722,-31521,       // w22 w20 w18 w16
-  26722,6270,-6270,-17855,        // w23 w21 w19 w17
-  17855,6270,6270,26722,          // w30 w28 w26 w24
-  -31521,26722,-17855,-31521,     // w31 w29 w27 w25
-// Table for rows 2,6 - constants are multiplied by cos_2_16
-  21407,21407,21407,-21407,       // movq-> w06 w04 w02 w00
-  27969,11585,11585,-27969,       // w07 w05 w03 w01
-  21407,-21407,21407,21407,       // w14 w12 w10 w08
-  -11585,27969,-27969,-11585,     // w15 w13 w11 w09
-  29692,16819,25172,-29692,       // w22 w20 w18 w16
-  25172,5906,-5906,-16819,        // w23 w21 w19 w17
-  16819,5906,5906,25172,          // w30 w28 w26 w24
-  -29692,25172,-16819,-29692,     // w31 w29 w27 w25
-// Table for rows 3,5 - constants are multiplied by cos_3_16
-  19266,19266,19266,-19266,       // movq-> w06 w04 w02 w00
-  25172,10426,10426,-25172,       // w07 w05 w03 w01
-  19266,-19266,19266,19266,       // w14 w12 w10 w08
-  -10426,25172,-25172,-10426,     // w15 w13 w11 w09
-  26722,15137,22654,-26722,       // w22 w20 w18 w16
-  22654,5315,-5315,-15137,        // w23 w21 w19 w17
-  15137,5315,5315,22654,          // w30 w28 w26 w24
-  -26722,22654,-15137,-26722,     // w31 w29 w27 w25
-};
-//-----------------------------------------------------------------------------
-// Tables for xmm processors
-//-----------------------------------------------------------------------------
-
-// %3 for rows 0,4 - constants are multiplied by cos_4_16
-DECLARE_ALIGNED(8, static const int16_t, tab_i_04_xmm)[32*4] = {
-  16384,21407,16384,8867,      // movq-> w05 w04 w01 w00
-  16384,8867,-16384,-21407,    // w07 w06 w03 w02
-  16384,-8867,16384,-21407,    // w13 w12 w09 w08
-  -16384,21407,16384,-8867,    // w15 w14 w11 w10
-  22725,19266,19266,-4520,     // w21 w20 w17 w16
-  12873,4520,-22725,-12873,    // w23 w22 w19 w18
-  12873,-22725,4520,-12873,    // w29 w28 w25 w24
-  4520,19266,19266,-22725,     // w31 w30 w27 w26
-// %3 for rows 1,7 - constants are multiplied by cos_1_16
-  22725,29692,22725,12299,     // movq-> w05 w04 w01 w00
-  22725,12299,-22725,-29692,   // w07 w06 w03 w02
-  22725,-12299,22725,-29692,   // w13 w12 w09 w08
-  -22725,29692,22725,-12299,   // w15 w14 w11 w10
-  31521,26722,26722,-6270,     // w21 w20 w17 w16
-  17855,6270,-31521,-17855,    // w23 w22 w19 w18
-  17855,-31521,6270,-17855,    // w29 w28 w25 w24
-  6270,26722,26722,-31521,     // w31 w30 w27 w26
-// %3 for rows 2,6 - constants are multiplied by cos_2_16
-  21407,27969,21407,11585,     // movq-> w05 w04 w01 w00
-  21407,11585,-21407,-27969,   // w07 w06 w03 w02
-  21407,-11585,21407,-27969,   // w13 w12 w09 w08
-  -21407,27969,21407,-11585,   // w15 w14 w11 w10
-  29692,25172,25172,-5906,     // w21 w20 w17 w16
-  16819,5906,-29692,-16819,    // w23 w22 w19 w18
-  16819,-29692,5906,-16819,    // w29 w28 w25 w24
-  5906,25172,25172,-29692,     // w31 w30 w27 w26
-// %3 for rows 3,5 - constants are multiplied by cos_3_16
-  19266,25172,19266,10426,     // movq-> w05 w04 w01 w00
-  19266,10426,-19266,-25172,   // w07 w06 w03 w02
-  19266,-10426,19266,-25172,   // w13 w12 w09 w08
-  -19266,25172,19266,-10426,   // w15 w14 w11 w10
-  26722,22654,22654,-5315,     // w21 w20 w17 w16
-  15137,5315,-26722,-15137,    // w23 w22 w19 w18
-  15137,-26722,5315,-15137,    // w29 w28 w25 w24
-  5315,22654,22654,-26722,     // w31 w30 w27 w26
-};
-//=============================================================================
-// Helper macros for the code
-//=============================================================================
-
-//-----------------------------------------------------------------------------
-// DCT_8_INV_ROW_MMX( INP, OUT, TABLE, ROUNDER
-//-----------------------------------------------------------------------------
-
-#define DCT_8_INV_ROW_MMX(A1,A2,A3,A4)\
-  "movq " #A1 ",%%mm0              \n\t"/* 0 ; x3 x2 x1 x0*/\
-  "movq 8+" #A1 ",%%mm1            \n\t"/* 1 ; x7 x6 x5 x4*/\
-  "movq %%mm0,%%mm2            \n\t"/* 2  ; x3 x2 x1 x0*/\
-  "movq " #A3 ",%%mm3              \n\t"/* 3 ; w06 w04 w02 w00*/\
-  "punpcklwd %%mm1,%%mm0       \n\t"/* x5 x1 x4 x0*/\
-  "movq %%mm0,%%mm5            \n\t"/* 5 ; x5 x1 x4 x0*/\
-  "punpckldq %%mm0,%%mm0       \n\t"/* x4 x0 x4 x0*/\
-  "movq 8+" #A3 ",%%mm4            \n\t"/* 4 ; w07 w05 w03 w01*/\
-  "punpckhwd %%mm1,%%mm2       \n\t"/* 1 ; x7 x3 x6 x2*/\
-  "pmaddwd %%mm0,%%mm3         \n\t"/* x4*w06+x0*w04 x4*w02+x0*w00*/\
-  "movq %%mm2,%%mm6            \n\t"/* 6 ; x7 x3 x6 x2*/\
-  "movq 32+" #A3 ",%%mm1           \n\t"/* 1 ; w22 w20 w18 w16*/\
-  "punpckldq %%mm2,%%mm2       \n\t"/* x6 x2 x6 x2*/\
-  "pmaddwd %%mm2,%%mm4         \n\t"/* x6*w07+x2*w05 x6*w03+x2*w01*/\
-  "punpckhdq %%mm5,%%mm5       \n\t"/* x5 x1 x5 x1*/\
-  "pmaddwd 16+" #A3 ",%%mm0        \n\t"/* x4*w14+x0*w12 x4*w10+x0*w08*/\
-  "punpckhdq %%mm6,%%mm6       \n\t"/* x7 x3 x7 x3*/\
-  "movq 40+" #A3 ",%%mm7           \n\t"/* 7 ; w23 w21 w19 w17*/\
-  "pmaddwd %%mm5,%%mm1         \n\t"/* x5*w22+x1*w20 x5*w18+x1*w16*/\
-  "paddd " #A4 ",%%mm3             \n\t"/* +%4*/\
-  "pmaddwd %%mm6,%%mm7         \n\t"/* x7*w23+x3*w21 x7*w19+x3*w17*/\
-  "pmaddwd 24+" #A3 ",%%mm2        \n\t"/* x6*w15+x2*w13 x6*w11+x2*w09*/\
-  "paddd %%mm4,%%mm3           \n\t"/* 4 ; a1=sum(even1) a0=sum(even0)*/\
-  "pmaddwd 48+" #A3 ",%%mm5        \n\t"/* x5*w30+x1*w28 x5*w26+x1*w24*/\
-  "movq %%mm3,%%mm4            \n\t"/* 4 ; a1 a0*/\
-  "pmaddwd 56+" #A3 ",%%mm6        \n\t"/* x7*w31+x3*w29 x7*w27+x3*w25*/\
-  "paddd %%mm7,%%mm1           \n\t"/* 7 ; b1=sum(odd1) b0=sum(odd0)*/\
-  "paddd " #A4 ",%%mm0             \n\t"/* +%4*/\
-  "psubd %%mm1,%%mm3           \n\t"/* a1-b1 a0-b0*/\
-  "psrad $11,%%mm3 \n\t"/* y6=a1-b1 y7=a0-b0*/\
-  "paddd %%mm4,%%mm1           \n\t"/* 4 ; a1+b1 a0+b0*/\
-  "paddd %%mm2,%%mm0           \n\t"/* 2 ; a3=sum(even3) a2=sum(even2)*/\
-  "psrad $11,%%mm1 \n\t"/* y1=a1+b1 y0=a0+b0*/\
-  "paddd %%mm6,%%mm5           \n\t"/* 6 ; b3=sum(odd3) b2=sum(odd2)*/\
-  "movq %%mm0,%%mm4            \n\t"/* 4 ; a3 a2*/\
-  "paddd %%mm5,%%mm0           \n\t"/* a3+b3 a2+b2*/\
-  "psubd %%mm5,%%mm4           \n\t"/* 5 ; a3-b3 a2-b2*/\
-  "psrad $11,%%mm0 \n\t"/* y3=a3+b3 y2=a2+b2*/\
-  "psrad $11,%%mm4 \n\t"/* y4=a3-b3 y5=a2-b2*/\
-  "packssdw %%mm0,%%mm1        \n\t"/* 0 ; y3 y2 y1 y0*/\
-  "packssdw %%mm3,%%mm4        \n\t"/* 3 ; y6 y7 y4 y5*/\
-  "movq %%mm4,%%mm7            \n\t"/* 7 ; y6 y7 y4 y5*/\
-  "psrld $16,%%mm4            \n\t"/* 0 y6 0 y4*/\
-  "pslld $16,%%mm7            \n\t"/* y7 0 y5 0*/\
-  "movq %%mm1," #A2 "              \n\t"/* 1 ; save y3 y2 y1 y0*/\
-  "por %%mm4,%%mm7             \n\t"/* 4 ; y7 y6 y5 y4*/\
-  "movq %%mm7,8            +" #A2 "\n\t"/* 7 ; save y7 y6 y5 y4*/\
-
-
-//-----------------------------------------------------------------------------
-// DCT_8_INV_ROW_XMM( INP, OUT, TABLE, ROUNDER
-//-----------------------------------------------------------------------------
-
-#define DCT_8_INV_ROW_XMM(A1,A2,A3,A4)\
-  "movq " #A1 ",%%mm0                  \n\t"/* 0     ; x3 x2 x1 x0*/\
-  "movq 8+" #A1 ",%%mm1                \n\t"/* 1     ; x7 x6 x5 x4*/\
-  "movq %%mm0,%%mm2                \n\t"/* 2     ; x3 x2 x1 x0*/\
-  "movq " #A3 ",%%mm3                  \n\t"/* 3     ; w05 w04 w01 w00*/\
-  "pshufw $0x88,%%mm0,%%mm0        \n\t"/* x2 x0 x2 x0*/\
-  "movq 8+" #A3 ",%%mm4                \n\t"/* 4     ; w07 w06 w03 w02*/\
-  "movq %%mm1,%%mm5                \n\t"/* 5     ; x7 x6 x5 x4*/\
-  "pmaddwd %%mm0,%%mm3             \n\t"/* x2*w05+x0*w04 x2*w01+x0*w00*/\
-  "movq 32+" #A3 ",%%mm6               \n\t"/* 6     ; w21 w20 w17 w16*/\
-  "pshufw $0x88,%%mm1,%%mm1        \n\t"/* x6 x4 x6 x4*/\
-  "pmaddwd %%mm1,%%mm4             \n\t"/* x6*w07+x4*w06 x6*w03+x4*w02*/\
-  "movq 40+" #A3 ",%%mm7               \n\t"/* 7    ; w23 w22 w19 w18*/\
-  "pshufw $0xdd,%%mm2,%%mm2        \n\t"/* x3 x1 x3 x1*/\
-  "pmaddwd %%mm2,%%mm6             \n\t"/* x3*w21+x1*w20 x3*w17+x1*w16*/\
-  "pshufw $0xdd,%%mm5,%%mm5        \n\t"/* x7 x5 x7 x5*/\
-  "pmaddwd %%mm5,%%mm7             \n\t"/* x7*w23+x5*w22 x7*w19+x5*w18*/\
-  "paddd " #A4 ",%%mm3                 \n\t"/* +%4*/\
-  "pmaddwd 16+" #A3 ",%%mm0            \n\t"/* x2*w13+x0*w12 x2*w09+x0*w08*/\
-  "paddd %%mm4,%%mm3               \n\t"/* 4     ; a1=sum(even1) a0=sum(even0)*/\
-  "pmaddwd 24+" #A3 ",%%mm1            \n\t"/* x6*w15+x4*w14 x6*w11+x4*w10*/\
-  "movq %%mm3,%%mm4                \n\t"/* 4     ; a1 a0*/\
-  "pmaddwd 48+" #A3 ",%%mm2            \n\t"/* x3*w29+x1*w28 x3*w25+x1*w24*/\
-  "paddd %%mm7,%%mm6               \n\t"/* 7     ; b1=sum(odd1) b0=sum(odd0)*/\
-  "pmaddwd 56+" #A3 ",%%mm5            \n\t"/* x7*w31+x5*w30 x7*w27+x5*w26*/\
-  "paddd %%mm6,%%mm3               \n\t"/* a1+b1 a0+b0*/\
-  "paddd " #A4 ",%%mm0                 \n\t"/* +%4*/\
-  "psrad $11,%%mm3     \n\t"/* y1=a1+b1 y0=a0+b0*/\
-  "paddd %%mm1,%%mm0               \n\t"/* 1     ; a3=sum(even3) a2=sum(even2)*/\
-  "psubd %%mm6,%%mm4               \n\t"/* 6     ; a1-b1 a0-b0*/\
-  "movq %%mm0,%%mm7                \n\t"/* 7     ; a3 a2*/\
-  "paddd %%mm5,%%mm2               \n\t"/* 5     ; b3=sum(odd3) b2=sum(odd2)*/\
-  "paddd %%mm2,%%mm0               \n\t"/* a3+b3 a2+b2*/\
-  "psrad $11,%%mm4     \n\t"/* y6=a1-b1 y7=a0-b0*/\
-  "psubd %%mm2,%%mm7               \n\t"/* 2     ; a3-b3 a2-b2*/\
-  "psrad $11,%%mm0     \n\t"/* y3=a3+b3 y2=a2+b2*/\
-  "psrad $11,%%mm7     \n\t"/* y4=a3-b3 y5=a2-b2*/\
-  "packssdw %%mm0,%%mm3            \n\t"/* 0     ; y3 y2 y1 y0*/\
-  "packssdw %%mm4,%%mm7            \n\t"/* 4     ; y6 y7 y4 y5*/\
-  "movq %%mm3, " #A2 "                  \n\t"/* 3     ; save y3 y2 y1 y0*/\
-  "pshufw $0xb1,%%mm7,%%mm7        \n\t"/* y7 y6 y5 y4*/\
-  "movq %%mm7,8                +" #A2 "\n\t"/* 7     ; save y7 y6 y5 y4*/\
-
-
-//-----------------------------------------------------------------------------
-//
-// The first stage DCT 8x8 - forward DCTs of columns
-//
-// The %2puts are multiplied
-// for rows 0,4 - on cos_4_16,
-// for rows 1,7 - on cos_1_16,
-// for rows 2,6 - on cos_2_16,
-// for rows 3,5 - on cos_3_16
-// and are shifted to the left for rise of accuracy
-//
-//-----------------------------------------------------------------------------
-//
-// The 8-point scaled forward DCT algorithm (26a8m)
-//
-//-----------------------------------------------------------------------------
-//
-// #define DCT_8_FRW_COL(x, y)
-//{
-// short t0, t1, t2, t3, t4, t5, t6, t7;
-// short tp03, tm03, tp12, tm12, tp65, tm65;
-// short tp465, tm465, tp765, tm765;
-//
-// t0 = LEFT_SHIFT ( x[0] + x[7] );
-// t1 = LEFT_SHIFT ( x[1] + x[6] );
-// t2 = LEFT_SHIFT ( x[2] + x[5] );
-// t3 = LEFT_SHIFT ( x[3] + x[4] );
-// t4 = LEFT_SHIFT ( x[3] - x[4] );
-// t5 = LEFT_SHIFT ( x[2] - x[5] );
-// t6 = LEFT_SHIFT ( x[1] - x[6] );
-// t7 = LEFT_SHIFT ( x[0] - x[7] );
-//
-// tp03 = t0 + t3;
-// tm03 = t0 - t3;
-// tp12 = t1 + t2;
-// tm12 = t1 - t2;
-//
-// y[0] = tp03 + tp12;
-// y[4] = tp03 - tp12;
-//
-// y[2] = tm03 + tm12 * tg_2_16;
-// y[6] = tm03 * tg_2_16 - tm12;
-//
-// tp65 =(t6 +t5 )*cos_4_16;
-// tm65 =(t6 -t5 )*cos_4_16;
-//
-// tp765 = t7 + tp65;
-// tm765 = t7 - tp65;
-// tp465 = t4 + tm65;
-// tm465 = t4 - tm65;
-//
-// y[1] = tp765 + tp465 * tg_1_16;
-// y[7] = tp765 * tg_1_16 - tp465;
-// y[5] = tm765 * tg_3_16 + tm465;
-// y[3] = tm765 - tm465 * tg_3_16;
-//}
-//
-//-----------------------------------------------------------------------------
-
-//-----------------------------------------------------------------------------
-// DCT_8_INV_COL_4  INP,OUT
-//-----------------------------------------------------------------------------
-
-#define DCT_8_INV_COL(A1,A2)\
-  "movq 2*8(%3),%%mm0\n\t"\
-  "movq 16*3+" #A1 ",%%mm3\n\t"\
-  "movq %%mm0,%%mm1            \n\t"/* tg_3_16*/\
-  "movq 16*5+" #A1 ",%%mm5\n\t"\
-  "pmulhw %%mm3,%%mm0          \n\t"/* x3*(tg_3_16-1)*/\
-  "movq (%3),%%mm4\n\t"\
-  "pmulhw %%mm5,%%mm1          \n\t"/* x5*(tg_3_16-1)*/\
-  "movq 16*7+" #A1 ",%%mm7\n\t"\
-  "movq %%mm4,%%mm2            \n\t"/* tg_1_16*/\
-  "movq 16*1+" #A1 ",%%mm6\n\t"\
-  "pmulhw %%mm7,%%mm4          \n\t"/* x7*tg_1_16*/\
-  "paddsw %%mm3,%%mm0          \n\t"/* x3*tg_3_16*/\
-  "pmulhw %%mm6,%%mm2          \n\t"/* x1*tg_1_16*/\
-  "paddsw %%mm3,%%mm1          \n\t"/* x3+x5*(tg_3_16-1)*/\
-  "psubsw %%mm5,%%mm0          \n\t"/* x3*tg_3_16-x5 = tm35*/\
-  "movq 3*8(%3),%%mm3\n\t"\
-  "paddsw %%mm5,%%mm1          \n\t"/* x3+x5*tg_3_16 = tp35*/\
-  "paddsw %%mm6,%%mm4          \n\t"/* x1+tg_1_16*x7 = tp17*/\
-  "psubsw %%mm7,%%mm2          \n\t"/* x1*tg_1_16-x7 = tm17*/\
-  "movq %%mm4,%%mm5            \n\t"/* tp17*/\
-  "movq %%mm2,%%mm6            \n\t"/* tm17*/\
-  "paddsw %%mm1,%%mm5          \n\t"/* tp17+tp35 = b0*/\
-  "psubsw %%mm0,%%mm6          \n\t"/* tm17-tm35 = b3*/\
-  "psubsw %%mm1,%%mm4          \n\t"/* tp17-tp35 = t1*/\
-  "paddsw %%mm0,%%mm2          \n\t"/* tm17+tm35 = t2*/\
-  "movq 1*8(%3),%%mm7\n\t"\
-  "movq %%mm4,%%mm1            \n\t"/* t1*/\
-  "movq %%mm5,3*16         +" #A2 "\n\t"/* save b0*/\
-  "paddsw %%mm2,%%mm1          \n\t"/* t1+t2*/\
-  "movq %%mm6,5*16         +" #A2 "\n\t"/* save b3*/\
-  "psubsw %%mm2,%%mm4          \n\t"/* t1-t2*/\
-  "movq 2*16+" #A1 ",%%mm5\n\t"\
-  "movq %%mm7,%%mm0            \n\t"/* tg_2_16*/\
-  "movq 6*16+" #A1 ",%%mm6\n\t"\
-  "pmulhw %%mm5,%%mm0          \n\t"/* x2*tg_2_16*/\
-  "pmulhw %%mm6,%%mm7          \n\t"/* x6*tg_2_16*/\
-  "pmulhw %%mm3,%%mm1          \n\t"/* ocos_4_16*(t1+t2) = b1/2*/\
-  "movq 0*16+" #A1 ",%%mm2\n\t"\
-  "pmulhw %%mm3,%%mm4          \n\t"/* ocos_4_16*(t1-t2) = b2/2*/\
-  "psubsw %%mm6,%%mm0          \n\t"/* t2*tg_2_16-x6 = tm26*/\
-  "movq %%mm2,%%mm3            \n\t"/* x0*/\
-  "movq 4*16+" #A1 ",%%mm6\n\t"\
-  "paddsw %%mm5,%%mm7          \n\t"/* x2+x6*tg_2_16 = tp26*/\
-  "paddsw %%mm6,%%mm2          \n\t"/* x0+x4 = tp04*/\
-  "psubsw %%mm6,%%mm3          \n\t"/* x0-x4 = tm04*/\
-  "movq %%mm2,%%mm5            \n\t"/* tp04*/\
-  "movq %%mm3,%%mm6            \n\t"/* tm04*/\
-  "psubsw %%mm7,%%mm2          \n\t"/* tp04-tp26 = a3*/\
-  "paddsw %%mm0,%%mm3          \n\t"/* tm04+tm26 = a1*/\
-  "paddsw %%mm1,%%mm1          \n\t"/* b1*/\
-  "paddsw %%mm4,%%mm4          \n\t"/* b2*/\
-  "paddsw %%mm7,%%mm5          \n\t"/* tp04+tp26 = a0*/\
-  "psubsw %%mm0,%%mm6          \n\t"/* tm04-tm26 = a2*/\
-  "movq %%mm3,%%mm7            \n\t"/* a1*/\
-  "movq %%mm6,%%mm0            \n\t"/* a2*/\
-  "paddsw %%mm1,%%mm3          \n\t"/* a1+b1*/\
-  "paddsw %%mm4,%%mm6          \n\t"/* a2+b2*/\
-  "psraw $6,%%mm3 \n\t"/* dst1*/\
-  "psubsw %%mm1,%%mm7          \n\t"/* a1-b1*/\
-  "psraw $6,%%mm6 \n\t"/* dst2*/\
-  "psubsw %%mm4,%%mm0          \n\t"/* a2-b2*/\
-  "movq 3*16+" #A2 ",%%mm1         \n\t"/* load b0*/\
-  "psraw $6,%%mm7 \n\t"/* dst6*/\
-  "movq %%mm5,%%mm4            \n\t"/* a0*/\
-  "psraw $6,%%mm0 \n\t"/* dst5*/\
-  "movq %%mm3,1*16+" #A2 "\n\t"\
-  "paddsw %%mm1,%%mm5          \n\t"/* a0+b0*/\
-  "movq %%mm6,2*16+" #A2 "\n\t"\
-  "psubsw %%mm1,%%mm4          \n\t"/* a0-b0*/\
-  "movq 5*16+" #A2 ",%%mm3         \n\t"/* load b3*/\
-  "psraw $6,%%mm5 \n\t"/* dst0*/\
-  "movq %%mm2,%%mm6            \n\t"/* a3*/\
-  "psraw $6,%%mm4 \n\t"/* dst7*/\
-  "movq %%mm0,5*16+" #A2 "\n\t"\
-  "paddsw %%mm3,%%mm2          \n\t"/* a3+b3*/\
-  "movq %%mm7,6*16+" #A2 "\n\t"\
-  "psubsw %%mm3,%%mm6          \n\t"/* a3-b3*/\
-  "movq %%mm5,0*16+" #A2 "\n\t"\
-  "psraw $6,%%mm2 \n\t"/* dst3*/\
-  "movq %%mm4,7*16+" #A2 "\n\t"\
-  "psraw $6,%%mm6 \n\t"/* dst4*/\
-  "movq %%mm2,3*16+" #A2 "\n\t"\
-  "movq %%mm6,4*16+" #A2 "\n\t"
-
-//=============================================================================
-// Code
-//=============================================================================
-
-//-----------------------------------------------------------------------------
-// void idct_mmx(uint16_t block[64]);
-//-----------------------------------------------------------------------------
-
-
-void ff_xvid_idct_mmx(short *block)
-{
-__asm__ volatile(
-            //# Process each row
-    DCT_8_INV_ROW_MMX(0*16(%0), 0*16(%0), 64*0(%2), 8*0(%1))
-    DCT_8_INV_ROW_MMX(1*16(%0), 1*16(%0), 64*1(%2), 8*1(%1))
-    DCT_8_INV_ROW_MMX(2*16(%0), 2*16(%0), 64*2(%2), 8*2(%1))
-    DCT_8_INV_ROW_MMX(3*16(%0), 3*16(%0), 64*3(%2), 8*3(%1))
-    DCT_8_INV_ROW_MMX(4*16(%0), 4*16(%0), 64*0(%2), 8*4(%1))
-    DCT_8_INV_ROW_MMX(5*16(%0), 5*16(%0), 64*3(%2), 8*5(%1))
-    DCT_8_INV_ROW_MMX(6*16(%0), 6*16(%0), 64*2(%2), 8*6(%1))
-    DCT_8_INV_ROW_MMX(7*16(%0), 7*16(%0), 64*1(%2), 8*7(%1))
-
-            //# Process the columns (4 at a time)
-    DCT_8_INV_COL(0(%0), 0(%0))
-    DCT_8_INV_COL(8(%0), 8(%0))
-    :: "r"(block), "r"(rounder_0), "r"(tab_i_04_mmx), "r"(tg_1_16));
-}
-
-void ff_xvid_idct_mmx_put(uint8_t *dest, int line_size, int16_t *block)
-{
-    ff_xvid_idct_mmx(block);
-    ff_put_pixels_clamped_mmx(block, dest, line_size);
-}
-
-void ff_xvid_idct_mmx_add(uint8_t *dest, int line_size, int16_t *block)
-{
-    ff_xvid_idct_mmx(block);
-    ff_add_pixels_clamped_mmx(block, dest, line_size);
-}
-
-#endif /* HAVE_MMX_INLINE */
-
-#if HAVE_MMXEXT_INLINE
-
-//-----------------------------------------------------------------------------
-// void idct_xmm(uint16_t block[64]);
-//-----------------------------------------------------------------------------
-
-
-void ff_xvid_idct_mmxext(short *block)
-{
-__asm__ volatile(
-            //# Process each row
-    DCT_8_INV_ROW_XMM(0*16(%0), 0*16(%0), 64*0(%2), 8*0(%1))
-    DCT_8_INV_ROW_XMM(1*16(%0), 1*16(%0), 64*1(%2), 8*1(%1))
-    DCT_8_INV_ROW_XMM(2*16(%0), 2*16(%0), 64*2(%2), 8*2(%1))
-    DCT_8_INV_ROW_XMM(3*16(%0), 3*16(%0), 64*3(%2), 8*3(%1))
-    DCT_8_INV_ROW_XMM(4*16(%0), 4*16(%0), 64*0(%2), 8*4(%1))
-    DCT_8_INV_ROW_XMM(5*16(%0), 5*16(%0), 64*3(%2), 8*5(%1))
-    DCT_8_INV_ROW_XMM(6*16(%0), 6*16(%0), 64*2(%2), 8*6(%1))
-    DCT_8_INV_ROW_XMM(7*16(%0), 7*16(%0), 64*1(%2), 8*7(%1))
-
-            //# Process the columns (4 at a time)
-    DCT_8_INV_COL(0(%0), 0(%0))
-    DCT_8_INV_COL(8(%0), 8(%0))
-    :: "r"(block), "r"(rounder_0), "r"(tab_i_04_xmm), "r"(tg_1_16));
-}
-
-void ff_xvid_idct_mmxext_put(uint8_t *dest, int line_size, int16_t *block)
-{
-    ff_xvid_idct_mmxext(block);
-    ff_put_pixels_clamped_mmx(block, dest, line_size);
-}
-
-void ff_xvid_idct_mmxext_add(uint8_t *dest, int line_size, int16_t *block)
-{
-    ff_xvid_idct_mmxext(block);
-    ff_add_pixels_clamped_mmx(block, dest, line_size);
-}
-
-#endif /* HAVE_MMXEXT_INLINE */
+/*
+ * XVID MPEG-4 VIDEO CODEC
+ * - MMX and XMM forward discrete cosine transform -
+ *
+ * Copyright(C) 2001 Peter Ross <pross@xvid.org>
+ *
+ * Originally provided by Intel at AP-922
+ * http://developer.intel.com/vtune/cbts/strmsimd/922down.htm
+ * (See more app notes at http://developer.intel.com/vtune/cbts/strmsimd/appnotes.htm)
+ * but in a limited edition.
+ * New macro implements a column part for precise iDCT
+ * The routine precision now satisfies IEEE standard 1180-1990.
+ *
+ * Copyright(C) 2000-2001 Peter Gubanov <peter@elecard.net.ru>
+ * Rounding trick Copyright(C) 2000 Michel Lespinasse <walken@zoy.org>
+ *
+ * http://www.elecard.com/peter/idct.html
+ * http://www.linuxvideo.org/mpeg2dec/
+ *
+ * These examples contain code fragments for first stage iDCT 8x8
+ * (for rows) and first stage DCT 8x8 (for columns)
+ *
+ * conversion to gcc syntax by Michael Niedermayer
+ *
+ * This file is part of Libav.
+ *
+ * Libav is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * Libav is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with Libav; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <inttypes.h>
+
+#include "config.h"
+
+#include "libavutil/mem.h"
+
+#include "libavcodec/avcodec.h"
+
+#include "idctdsp.h"
+#include "xvididct.h"
+
+#if HAVE_MMX_INLINE
+
+// -----------------------------------------------------------------------------
+// Various memory constants (trigonometric values or rounding values)
+// -----------------------------------------------------------------------------
+
+DECLARE_ALIGNED(8, static const int16_t, tg_1_16)[4 * 4] = {
+     13036,  13036,  13036,  13036, // tg * (2 << 16) + 0.5
+     27146,  27146,  27146,  27146, // tg * (2 << 16) + 0.5
+    -21746, -21746, -21746, -21746, // tg * (2 << 16) + 0.5
+     23170,  23170,  23170,  23170
+};                                  // cos * (2 << 15) + 0.5
+
+DECLARE_ALIGNED(8, static const int32_t, rounder_0)[2 * 8] = {
+    65536, 65536,
+    3597,   3597,
+    2260,   2260,
+    1203,   1203,
+    0,         0,
+    120,     120,
+    512,     512,
+    512, 512
+};
+
+// -----------------------------------------------------------------------------
+//
+// The first stage iDCT 8x8 - inverse DCTs of rows
+//
+// -----------------------------------------------------------------------------
+// The 8-point inverse DCT direct algorithm
+// -----------------------------------------------------------------------------
+//
+// static const short w[32] = {
+//     FIX(cos_4_16),  FIX(cos_2_16),  FIX(cos_4_16),  FIX(cos_6_16),
+//     FIX(cos_4_16),  FIX(cos_6_16), -FIX(cos_4_16), -FIX(cos_2_16),
+//     FIX(cos_4_16), -FIX(cos_6_16), -FIX(cos_4_16),  FIX(cos_2_16),
+//     FIX(cos_4_16), -FIX(cos_2_16),  FIX(cos_4_16), -FIX(cos_6_16),
+//     FIX(cos_1_16),  FIX(cos_3_16),  FIX(cos_5_16),  FIX(cos_7_16),
+//     FIX(cos_3_16), -FIX(cos_7_16), -FIX(cos_1_16), -FIX(cos_5_16),
+//     FIX(cos_5_16), -FIX(cos_1_16),  FIX(cos_7_16),  FIX(cos_3_16),
+//     FIX(cos_7_16), -FIX(cos_5_16),  FIX(cos_3_16), -FIX(cos_1_16) };
+//
+// #define DCT_8_INV_ROW(x, y)
+// {
+//     int a0, a1, a2, a3, b0, b1, b2, b3;
+//
+//     a0 = x[0] * w[0]  + x[2] * w[1]  + x[4] * w[2]  + x[6] * w[3];
+//     a1 = x[0] * w[4]  + x[2] * w[5]  + x[4] * w[6]  + x[6] * w[7];
+//     a2 = x[0] * w[8]  + x[2] * w[9]  + x[4] * w[10] + x[6] * w[11];
+//     a3 = x[0] * w[12] + x[2] * w[13] + x[4] * w[14] + x[6] * w[15];
+//     b0 = x[1] * w[16] + x[3] * w[17] + x[5] * w[18] + x[7] * w[19];
+//     b1 = x[1] * w[20] + x[3] * w[21] + x[5] * w[22] + x[7] * w[23];
+//     b2 = x[1] * w[24] + x[3] * w[25] + x[5] * w[26] + x[7] * w[27];
+//     b3 = x[1] * w[28] + x[3] * w[29] + x[5] * w[30] + x[7] * w[31];
+//
+//     y[0] = SHIFT_ROUND(a0 + b0);
+//     y[1] = SHIFT_ROUND(a1 + b1);
+//     y[2] = SHIFT_ROUND(a2 + b2);
+//     y[3] = SHIFT_ROUND(a3 + b3);
+//     y[4] = SHIFT_ROUND(a3 - b3);
+//     y[5] = SHIFT_ROUND(a2 - b2);
+//     y[6] = SHIFT_ROUND(a1 - b1);
+//     y[7] = SHIFT_ROUND(a0 - b0);
+// }
+//
+// -----------------------------------------------------------------------------
+//
+// In this implementation the outputs of the iDCT-1D are multiplied
+//     for rows 0,4 - by cos_4_16,
+//     for rows 1,7 - by cos_1_16,
+//     for rows 2,6 - by cos_2_16,
+//     for rows 3,5 - by cos_3_16
+// and are shifted to the left for better accuracy.
+//
+// For the constants used,
+//     FIX(float_const) = (short) (float_const * (1 << 15) + 0.5)
+//
+// -----------------------------------------------------------------------------
+
+// -----------------------------------------------------------------------------
+// Tables for mmx processors
+// -----------------------------------------------------------------------------
+
+// Table for rows 0,4 - constants are multiplied by cos_4_16
+DECLARE_ALIGNED(8, static const int16_t, tab_i_04_mmx)[32 * 4] = {
+     16384,  16384,  16384, -16384, // movq-> w06 w04 w02 w00
+     21407,   8867,   8867, -21407, // w07 w05 w03 w01
+     16384, -16384,  16384,  16384, // w14 w12 w10 w08
+     -8867,  21407, -21407,  -8867, // w15 w13 w11 w09
+     22725,  12873,  19266, -22725, // w22 w20 w18 w16
+     19266,   4520,  -4520, -12873, // w23 w21 w19 w17
+     12873,   4520,   4520,  19266, // w30 w28 w26 w24
+    -22725,  19266, -12873, -22725, // w31 w29 w27 w25
+// Table for rows 1,7 - constants are multiplied by cos_1_16
+     22725,  22725,  22725, -22725, // movq-> w06 w04 w02 w00
+     29692,  12299,  12299, -29692, // w07 w05 w03 w01
+     22725, -22725,  22725,  22725, // w14 w12 w10 w08
+    -12299,  29692, -29692, -12299, // w15 w13 w11 w09
+     31521,  17855,  26722, -31521, // w22 w20 w18 w16
+     26722,   6270,  -6270, -17855, // w23 w21 w19 w17
+     17855,   6270,   6270,  26722, // w30 w28 w26 w24
+    -31521,  26722, -17855, -31521, // w31 w29 w27 w25
+// Table for rows 2,6 - constants are multiplied by cos_2_16
+     21407,  21407,  21407, -21407, // movq-> w06 w04 w02 w00
+     27969,  11585,  11585, -27969, // w07 w05 w03 w01
+     21407, -21407,  21407,  21407, // w14 w12 w10 w08
+    -11585,  27969, -27969, -11585, // w15 w13 w11 w09
+     29692,  16819,  25172, -29692, // w22 w20 w18 w16
+     25172,   5906,  -5906, -16819, // w23 w21 w19 w17
+     16819,   5906,   5906,  25172, // w30 w28 w26 w24
+    -29692,  25172, -16819, -29692, // w31 w29 w27 w25
+// Table for rows 3,5 - constants are multiplied by cos_3_16
+     19266,  19266,  19266, -19266, // movq-> w06 w04 w02 w00
+     25172,  10426,  10426, -25172, // w07 w05 w03 w01
+     19266, -19266,  19266,  19266, // w14 w12 w10 w08
+    -10426,  25172, -25172, -10426, // w15 w13 w11 w09
+     26722,  15137,  22654, -26722, // w22 w20 w18 w16
+     22654,   5315,  -5315, -15137, // w23 w21 w19 w17
+     15137,   5315,   5315,  22654, // w30 w28 w26 w24
+    -26722,  22654, -15137, -26722, // w31 w29 w27 w25
+};
+// -----------------------------------------------------------------------------
+// Tables for xmm processors
+// -----------------------------------------------------------------------------
+
+// %3 for rows 0,4 - constants are multiplied by cos_4_16
+DECLARE_ALIGNED(8, static const int16_t, tab_i_04_xmm)[32 * 4] = {
+     16384,  21407,  16384,   8867, // movq-> w05 w04 w01 w00
+     16384,   8867, -16384, -21407, // w07 w06 w03 w02
+     16384,  -8867,  16384, -21407, // w13 w12 w09 w08
+    -16384,  21407,  16384,  -8867, // w15 w14 w11 w10
+     22725,  19266,  19266,  -4520, // w21 w20 w17 w16
+     12873,   4520, -22725, -12873, // w23 w22 w19 w18
+     12873, -22725,   4520, -12873, // w29 w28 w25 w24
+      4520,  19266,  19266, -22725, // w31 w30 w27 w26
+// %3 for rows 1,7 - constants are multiplied by cos_1_16
+     22725,  29692,  22725,  12299, // movq-> w05 w04 w01 w00
+     22725,  12299, -22725, -29692, // w07 w06 w03 w02
+     22725, -12299,  22725, -29692, // w13 w12 w09 w08
+    -22725,  29692,  22725, -12299, // w15 w14 w11 w10
+     31521,  26722,  26722,  -6270, // w21 w20 w17 w16
+     17855,   6270, -31521, -17855, // w23 w22 w19 w18
+     17855, -31521,   6270, -17855, // w29 w28 w25 w24
+      6270,  26722,  26722, -31521, // w31 w30 w27 w26
+// %3 for rows 2,6 - constants are multiplied by cos_2_16
+     21407,  27969,  21407,  11585, // movq-> w05 w04 w01 w00
+     21407,  11585, -21407, -27969, // w07 w06 w03 w02
+     21407, -11585,  21407, -27969, // w13 w12 w09 w08
+    -21407,  27969,  21407, -11585, // w15 w14 w11 w10
+     29692,  25172,  25172,  -5906, // w21 w20 w17 w16
+     16819,   5906, -29692, -16819, // w23 w22 w19 w18
+     16819, -29692,   5906, -16819, // w29 w28 w25 w24
+      5906,  25172,  25172, -29692, // w31 w30 w27 w26
+// %3 for rows 3,5 - constants are multiplied by cos_3_16
+     19266,  25172,  19266,  10426, // movq-> w05 w04 w01 w00
+     19266,  10426, -19266, -25172, // w07 w06 w03 w02
+     19266, -10426,  19266, -25172, // w13 w12 w09 w08
+    -19266,  25172,  19266, -10426, // w15 w14 w11 w10
+     26722,  22654,  22654,  -5315, // w21 w20 w17 w16
+     15137,   5315, -26722, -15137, // w23 w22 w19 w18
+     15137, -26722,   5315, -15137, // w29 w28 w25 w24
+      5315,  22654,  22654, -26722, // w31 w30 w27 w26
+};
+// =============================================================================
+// Helper macros for the code
+// =============================================================================
+
+// -----------------------------------------------------------------------------
+// DCT_8_INV_ROW_MMX( INP, OUT, TABLE, ROUNDER
+// -----------------------------------------------------------------------------
+
+#define DCT_8_INV_ROW_MMX(A1, A2, A3, A4)                                       \
+    "movq       "#A1", %%mm0    \n\t" /* 0 ; x3 x2 x1 x0 */                     \
+    "movq     8+"#A1", %%mm1    \n\t" /* 1 ; x7 x6 x5 x4 */                     \
+    "movq       %%mm0, %%mm2    \n\t" /* 2 ; x3 x2 x1 x0 */                     \
+    "movq       "#A3", %%mm3    \n\t" /* 3 ; w06 w04 w02 w00 */                 \
+    "punpcklwd  %%mm1, %%mm0    \n\t" /* x5 x1 x4 x0 */                         \
+    "movq       %%mm0, %%mm5    \n\t" /* 5 ; x5 x1 x4 x0 */                     \
+    "punpckldq  %%mm0, %%mm0    \n\t" /* x4 x0 x4 x0 */                         \
+    "movq     8+"#A3", %%mm4    \n\t" /* 4 ; w07 w05 w03 w01 */                 \
+    "punpckhwd  %%mm1, %%mm2    \n\t" /* 1 ; x7 x3 x6 x2 */                     \
+    "pmaddwd    %%mm0, %%mm3    \n\t" /* x4*w06+x0*w04 x4*w02+x0*w00 */         \
+    "movq       %%mm2, %%mm6    \n\t" /* 6 ; x7 x3 x6 x2 */                     \
+    "movq    32+"#A3", %%mm1    \n\t" /* 1 ; w22 w20 w18 w16 */                 \
+    "punpckldq  %%mm2, %%mm2    \n\t" /* x6 x2 x6 x2 */                         \
+    "pmaddwd    %%mm2, %%mm4    \n\t" /* x6*w07+x2*w05 x6*w03+x2*w01 */         \
+    "punpckhdq  %%mm5, %%mm5    \n\t" /* x5 x1 x5 x1 */                         \
+    "pmaddwd 16+"#A3", %%mm0    \n\t" /* x4*w14+x0*w12 x4*w10+x0*w08 */         \
+    "punpckhdq  %%mm6, %%mm6    \n\t" /* x7 x3 x7 x3 */                         \
+    "movq 40+   "#A3", %%mm7    \n\t" /* 7 ; w23 w21 w19 w17 */                 \
+    "pmaddwd    %%mm5, %%mm1    \n\t" /* x5*w22+x1*w20 x5*w18+x1*w16 */         \
+    "paddd      "#A4", %%mm3    \n\t" /* +%4 */                                 \
+    "pmaddwd    %%mm6, %%mm7    \n\t" /* x7*w23+x3*w21 x7*w19+x3*w17 */         \
+    "pmaddwd 24+"#A3", %%mm2    \n\t" /* x6*w15+x2*w13 x6*w11+x2*w09 */         \
+    "paddd      %%mm4, %%mm3    \n\t" /* 4 ; a1=sum(even1) a0=sum(even0) */     \
+    "pmaddwd 48+"#A3", %%mm5    \n\t" /* x5*w30+x1*w28 x5*w26+x1*w24 */         \
+    "movq       %%mm3, %%mm4    \n\t" /* 4 ; a1 a0 */                           \
+    "pmaddwd 56+"#A3", %%mm6    \n\t" /* x7*w31+x3*w29 x7*w27+x3*w25 */         \
+    "paddd      %%mm7, %%mm1    \n\t" /* 7 ; b1=sum(odd1) b0=sum(odd0) */       \
+    "paddd      "#A4", %%mm0    \n\t" /* +%4 */                                 \
+    "psubd      %%mm1, %%mm3    \n\t" /* a1-b1 a0-b0 */                         \
+    "psrad        $11, %%mm3    \n\t" /* y6=a1-b1 y7=a0-b0 */                   \
+    "paddd      %%mm4, %%mm1    \n\t" /* 4 ; a1+b1 a0+b0 */                     \
+    "paddd      %%mm2, %%mm0    \n\t" /* 2 ; a3=sum(even3) a2=sum(even2) */     \
+    "psrad        $11, %%mm1    \n\t" /* y1=a1+b1 y0=a0+b0 */                   \
+    "paddd      %%mm6, %%mm5    \n\t" /* 6 ; b3=sum(odd3) b2=sum(odd2) */       \
+    "movq       %%mm0, %%mm4    \n\t" /* 4 ; a3 a2 */                           \
+    "paddd      %%mm5, %%mm0    \n\t" /* a3+b3 a2+b2 */                         \
+    "psubd      %%mm5, %%mm4    \n\t" /* 5 ; a3-b3 a2-b2 */                     \
+    "psrad        $11, %%mm0    \n\t" /* y3=a3+b3 y2=a2+b2 */                   \
+    "psrad        $11, %%mm4    \n\t" /* y4=a3-b3 y5=a2-b2 */                   \
+    "packssdw   %%mm0, %%mm1    \n\t" /* 0 ; y3 y2 y1 y0 */                     \
+    "packssdw   %%mm3, %%mm4    \n\t" /* 3 ; y6 y7 y4 y5 */                     \
+    "movq       %%mm4, %%mm7    \n\t" /* 7 ; y6 y7 y4 y5 */                     \
+    "psrld        $16, %%mm4    \n\t" /* 0 y6 0 y4 */                           \
+    "pslld        $16, %%mm7    \n\t" /* y7 0 y5 0 */                           \
+    "movq       %%mm1, "#A2"    \n\t" /* 1 ; save y3 y2 y1 y0 */                \
+    "por        %%mm4, %%mm7    \n\t" /* 4 ; y7 y6 y5 y4 */                     \
+    "movq       %%mm7, 8+"#A2"  \n\t" /* 7 ; save y7 y6 y5 y4 */                \
+
+
+// -----------------------------------------------------------------------------
+// DCT_8_INV_ROW_XMM( INP, OUT, TABLE, ROUNDER
+// -----------------------------------------------------------------------------
+
+#define DCT_8_INV_ROW_XMM(A1, A2, A3, A4)                                       \
+    "movq       "#A1", %%mm0        \n\t" /* 0 ; x3 x2 x1 x0 */                 \
+    "movq     8+"#A1", %%mm1        \n\t" /* 1 ; x7 x6 x5 x4 */                 \
+    "movq       %%mm0, %%mm2        \n\t" /* 2 ; x3 x2 x1 x0 */                 \
+    "movq       "#A3", %%mm3        \n\t" /* 3 ; w05 w04 w01 w00 */             \
+    "pshufw     $0x88, %%mm0, %%mm0 \n\t" /* x2 x0 x2 x0 */                     \
+    "movq     8+"#A3", %%mm4        \n\t" /* 4 ; w07 w06 w03 w02 */             \
+    "movq       %%mm1, %%mm5        \n\t" /* 5 ; x7 x6 x5 x4 */                 \
+    "pmaddwd    %%mm0, %%mm3        \n\t" /* x2*w05+x0*w04 x2*w01+x0*w00 */     \
+    "movq    32+"#A3", %%mm6        \n\t" /* 6 ; w21 w20 w17 w16 */             \
+    "pshufw     $0x88, %%mm1, %%mm1 \n\t" /* x6 x4 x6 x4 */                     \
+    "pmaddwd    %%mm1, %%mm4        \n\t" /* x6*w07+x4*w06 x6*w03+x4*w02 */     \
+    "movq    40+"#A3", %%mm7        \n\t" /* 7; w23 w22 w19 w18 */              \
+    "pshufw     $0xdd, %%mm2, %%mm2 \n\t" /* x3 x1 x3 x1 */                     \
+    "pmaddwd    %%mm2, %%mm6        \n\t" /* x3*w21+x1*w20 x3*w17+x1*w16 */     \
+    "pshufw     $0xdd, %%mm5, %%mm5 \n\t" /* x7 x5 x7 x5 */                     \
+    "pmaddwd    %%mm5, %%mm7        \n\t" /* x7*w23+x5*w22 x7*w19+x5*w18 */     \
+    "paddd      "#A4", %%mm3        \n\t" /* +%4 */                             \
+    "pmaddwd 16+"#A3", %%mm0        \n\t" /* x2*w13+x0*w12 x2*w09+x0*w08 */     \
+    "paddd      %%mm4, %%mm3        \n\t" /* 4 ; a1=sum(even1) a0=sum(even0) */ \
+    "pmaddwd 24+"#A3", %%mm1        \n\t" /* x6*w15+x4*w14 x6*w11+x4*w10 */     \
+    "movq       %%mm3, %%mm4        \n\t" /* 4 ; a1 a0 */                       \
+    "pmaddwd 48+"#A3", %%mm2        \n\t" /* x3*w29+x1*w28 x3*w25+x1*w24 */     \
+    "paddd      %%mm7, %%mm6        \n\t" /* 7 ; b1=sum(odd1) b0=sum(odd0) */   \
+    "pmaddwd 56+"#A3", %%mm5        \n\t" /* x7*w31+x5*w30 x7*w27+x5*w26 */     \
+    "paddd      %%mm6, %%mm3        \n\t" /* a1+b1 a0+b0 */                     \
+    "paddd      "#A4", %%mm0        \n\t" /* +%4 */                             \
+    "psrad        $11, %%mm3        \n\t" /* y1=a1+b1 y0=a0+b0 */               \
+    "paddd      %%mm1, %%mm0        \n\t" /* 1 ; a3=sum(even3) a2=sum(even2) */ \
+    "psubd      %%mm6, %%mm4        \n\t" /* 6 ; a1-b1 a0-b0 */                 \
+    "movq       %%mm0, %%mm7        \n\t" /* 7 ; a3 a2 */                       \
+    "paddd      %%mm5, %%mm2        \n\t" /* 5 ; b3=sum(odd3) b2=sum(odd2) */   \
+    "paddd      %%mm2, %%mm0        \n\t" /* a3+b3 a2+b2 */                     \
+    "psrad        $11, %%mm4        \n\t" /* y6=a1-b1 y7=a0-b0 */               \
+    "psubd      %%mm2, %%mm7        \n\t" /* 2 ; a3-b3 a2-b2 */                 \
+    "psrad        $11, %%mm0        \n\t" /* y3=a3+b3 y2=a2+b2 */               \
+    "psrad        $11, %%mm7        \n\t" /* y4=a3-b3 y5=a2-b2 */               \
+    "packssdw   %%mm0, %%mm3        \n\t" /* 0 ; y3 y2 y1 y0 */                 \
+    "packssdw   %%mm4, %%mm7        \n\t" /* 4 ; y6 y7 y4 y5 */                 \
+    "movq       %%mm3, "#A2"        \n\t" /* 3 ; save y3 y2 y1 y0 */            \
+    "pshufw     $0xb1, %%mm7, %%mm7 \n\t" /* y7 y6 y5 y4 */                     \
+    "movq       %%mm7, 8+"#A2"      \n\t" /* 7 ; save y7 y6 y5 y4 */            \
+
+
+// -----------------------------------------------------------------------------
+//
+// The first stage DCT 8x8 - forward DCTs of columns
+//
+// The %2puts are multiplied
+// for rows 0,4 - on cos_4_16,
+// for rows 1,7 - on cos_1_16,
+// for rows 2,6 - on cos_2_16,
+// for rows 3,5 - on cos_3_16
+// and are shifted to the left for rise of accuracy
+//
+// -----------------------------------------------------------------------------
+//
+// The 8-point scaled forward DCT algorithm (26a8m)
+//
+// -----------------------------------------------------------------------------
+//
+//#define DCT_8_FRW_COL(x, y)
+// {
+//     short t0, t1, t2, t3, t4, t5, t6, t7;
+//     short tp03, tm03, tp12, tm12, tp65, tm65;
+//     short tp465, tm465, tp765, tm765;
+//
+//     t0 = LEFT_SHIFT(x[0] + x[7]);
+//     t1 = LEFT_SHIFT(x[1] + x[6]);
+//     t2 = LEFT_SHIFT(x[2] + x[5]);
+//     t3 = LEFT_SHIFT(x[3] + x[4]);
+//     t4 = LEFT_SHIFT(x[3] - x[4]);
+//     t5 = LEFT_SHIFT(x[2] - x[5]);
+//     t6 = LEFT_SHIFT(x[1] - x[6]);
+//     t7 = LEFT_SHIFT(x[0] - x[7]);
+//
+//     tp03 = t0 + t3;
+//     tm03 = t0 - t3;
+//     tp12 = t1 + t2;
+//     tm12 = t1 - t2;
+//
+//     y[0] = tp03 + tp12;
+//     y[4] = tp03 - tp12;
+//
+//     y[2] = tm03 + tm12 * tg_2_16;
+//     y[6] = tm03 * tg_2_16 - tm12;
+//
+//     tp65 = (t6 + t5) * cos_4_16;
+//     tm65 = (t6 - t5) * cos_4_16;
+//
+//     tp765 = t7 + tp65;
+//     tm765 = t7 - tp65;
+//     tp465 = t4 + tm65;
+//     tm465 = t4 - tm65;
+//
+//     y[1] = tp765 + tp465 * tg_1_16;
+//     y[7] = tp765 * tg_1_16 - tp465;
+//     y[5] = tm765 * tg_3_16 + tm465;
+//     y[3] = tm765 - tm465 * tg_3_16;
+// }
+//
+// -----------------------------------------------------------------------------
+
+// -----------------------------------------------------------------------------
+// DCT_8_INV_COL_4  INP,OUT
+// -----------------------------------------------------------------------------
+
+#define DCT_8_INV_COL(A1, A2)                                                   \
+    "movq    2*8(%3), %%mm0         \n\t"                                       \
+    "movq 16*3+"#A1", %%mm3         \n\t"                                       \
+    "movq      %%mm0, %%mm1         \n\t" /* tg_3_16 */                         \
+    "movq 16*5+"#A1", %%mm5         \n\t"                                       \
+    "pmulhw    %%mm3, %%mm0         \n\t" /* x3*(tg_3_16-1) */                  \
+    "movq       (%3), %%mm4         \n\t"                                       \
+    "pmulhw    %%mm5, %%mm1         \n\t" /* x5*(tg_3_16-1) */                  \
+    "movq 16*7+"#A1", %%mm7         \n\t"                                       \
+    "movq      %%mm4, %%mm2         \n\t" /* tg_1_16 */                         \
+    "movq 16*1+"#A1", %%mm6         \n\t"                                       \
+    "pmulhw    %%mm7, %%mm4         \n\t" /* x7*tg_1_16 */                      \
+    "paddsw    %%mm3, %%mm0         \n\t" /* x3*tg_3_16 */                      \
+    "pmulhw    %%mm6, %%mm2         \n\t" /* x1*tg_1_16 */                      \
+    "paddsw    %%mm3, %%mm1         \n\t" /* x3+x5*(tg_3_16-1) */               \
+    "psubsw    %%mm5, %%mm0         \n\t" /* x3*tg_3_16-x5 = tm35 */            \
+    "movq    3*8(%3), %%mm3         \n\t"                                       \
+    "paddsw    %%mm5, %%mm1         \n\t" /* x3+x5*tg_3_16 = tp35 */            \
+    "paddsw    %%mm6, %%mm4         \n\t" /* x1+tg_1_16*x7 = tp17 */            \
+    "psubsw    %%mm7, %%mm2         \n\t" /* x1*tg_1_16-x7 = tm17 */            \
+    "movq      %%mm4, %%mm5         \n\t" /* tp17 */                            \
+    "movq      %%mm2, %%mm6         \n\t" /* tm17 */                            \
+    "paddsw    %%mm1, %%mm5         \n\t" /* tp17+tp35 = b0 */                  \
+    "psubsw    %%mm0, %%mm6         \n\t" /* tm17-tm35 = b3 */                  \
+    "psubsw    %%mm1, %%mm4         \n\t" /* tp17-tp35 = t1 */                  \
+    "paddsw    %%mm0, %%mm2         \n\t" /* tm17+tm35 = t2 */                  \
+    "movq    1*8(%3), %%mm7         \n\t"                                       \
+    "movq      %%mm4, %%mm1         \n\t" /* t1 */                              \
+    "movq      %%mm5, 3*16+"#A2"    \n\t" /* save b0 */                         \
+    "paddsw    %%mm2, %%mm1         \n\t" /* t1+t2 */                           \
+    "movq      %%mm6, 5*16+"#A2"    \n\t" /* save b3 */                         \
+    "psubsw    %%mm2, %%mm4         \n\t" /* t1-t2 */                           \
+    "movq 2*16+"#A1", %%mm5         \n\t"                                       \
+    "movq      %%mm7, %%mm0         \n\t" /* tg_2_16 */                         \
+    "movq 6*16+"#A1", %%mm6         \n\t"                                       \
+    "pmulhw    %%mm5, %%mm0         \n\t" /* x2*tg_2_16 */                      \
+    "pmulhw    %%mm6, %%mm7         \n\t" /* x6*tg_2_16 */                      \
+    "pmulhw    %%mm3, %%mm1         \n\t" /* ocos_4_16*(t1+t2) = b1/2 */        \
+    "movq 0*16+"#A1", %%mm2         \n\t"                                       \
+    "pmulhw    %%mm3, %%mm4         \n\t" /* ocos_4_16*(t1-t2) = b2/2 */        \
+    "psubsw    %%mm6, %%mm0         \n\t" /* t2*tg_2_16-x6 = tm26 */            \
+    "movq      %%mm2, %%mm3         \n\t" /* x0 */                              \
+    "movq 4*16+"#A1", %%mm6         \n\t"                                       \
+    "paddsw    %%mm5, %%mm7         \n\t" /* x2+x6*tg_2_16 = tp26 */            \
+    "paddsw    %%mm6, %%mm2         \n\t" /* x0+x4 = tp04 */                    \
+    "psubsw    %%mm6, %%mm3         \n\t" /* x0-x4 = tm04 */                    \
+    "movq      %%mm2, %%mm5         \n\t" /* tp04 */                            \
+    "movq      %%mm3, %%mm6         \n\t" /* tm04 */                            \
+    "psubsw    %%mm7, %%mm2         \n\t" /* tp04-tp26 = a3 */                  \
+    "paddsw    %%mm0, %%mm3         \n\t" /* tm04+tm26 = a1 */                  \
+    "paddsw    %%mm1, %%mm1         \n\t" /* b1 */                              \
+    "paddsw    %%mm4, %%mm4         \n\t" /* b2 */                              \
+    "paddsw    %%mm7, %%mm5         \n\t" /* tp04+tp26 = a0 */                  \
+    "psubsw    %%mm0, %%mm6         \n\t" /* tm04-tm26 = a2 */                  \
+    "movq      %%mm3, %%mm7         \n\t" /* a1 */                              \
+    "movq      %%mm6, %%mm0         \n\t" /* a2 */                              \
+    "paddsw    %%mm1, %%mm3         \n\t" /* a1+b1 */                           \
+    "paddsw    %%mm4, %%mm6         \n\t" /* a2+b2 */                           \
+    "psraw        $6, %%mm3         \n\t" /* dst1 */                            \
+    "psubsw    %%mm1, %%mm7         \n\t" /* a1-b1 */                           \
+    "psraw        $6, %%mm6         \n\t" /* dst2 */                            \
+    "psubsw    %%mm4, %%mm0         \n\t" /* a2-b2 */                           \
+    "movq 3*16+"#A2", %%mm1         \n\t" /* load b0 */                         \
+    "psraw        $6, %%mm7         \n\t" /* dst6 */                            \
+    "movq      %%mm5, %%mm4         \n\t" /* a0 */                              \
+    "psraw        $6, %%mm0         \n\t" /* dst5 */                            \
+    "movq      %%mm3, 1*16+"#A2"    \n\t"                                       \
+    "paddsw    %%mm1, %%mm5         \n\t" /* a0+b0 */                           \
+    "movq      %%mm6, 2*16+"#A2"    \n\t"                                       \
+    "psubsw    %%mm1, %%mm4         \n\t" /* a0-b0 */                           \
+    "movq 5*16+"#A2", %%mm3         \n\t" /* load b3 */                         \
+    "psraw        $6, %%mm5         \n\t" /* dst0 */                            \
+    "movq      %%mm2, %%mm6         \n\t" /* a3 */                              \
+    "psraw        $6, %%mm4         \n\t" /* dst7 */                            \
+    "movq      %%mm0, 5*16+"#A2"    \n\t"                                       \
+    "paddsw    %%mm3, %%mm2         \n\t" /* a3+b3 */                           \
+    "movq      %%mm7, 6*16+"#A2"    \n\t"                                       \
+    "psubsw    %%mm3, %%mm6         \n\t" /* a3-b3 */                           \
+    "movq      %%mm5, 0*16+"#A2"    \n\t"                                       \
+    "psraw        $6, %%mm2         \n\t" /* dst3 */                            \
+    "movq      %%mm4, 7*16+"#A2"    \n\t"                                       \
+    "psraw        $6, %%mm6         \n\t" /* dst4 */                            \
+    "movq      %%mm2, 3*16+"#A2"    \n\t"                                       \
+    "movq      %%mm6, 4*16+"#A2"    \n\t"                                       \
+
+// =============================================================================
+// Code
+// =============================================================================
+
+// -----------------------------------------------------------------------------
+// void idct_mmx(uint16_t block[64]);
+// -----------------------------------------------------------------------------
+
+void ff_xvid_idct_mmx(short *block)
+{
+    __asm__ volatile (
+        // # Process each row
+        DCT_8_INV_ROW_MMX(0 * 16(%0), 0 * 16(%0), 64 * 0(%2), 8 * 0(%1))
+        DCT_8_INV_ROW_MMX(1 * 16(%0), 1 * 16(%0), 64 * 1(%2), 8 * 1(%1))
+        DCT_8_INV_ROW_MMX(2 * 16(%0), 2 * 16(%0), 64 * 2(%2), 8 * 2(%1))
+        DCT_8_INV_ROW_MMX(3 * 16(%0), 3 * 16(%0), 64 * 3(%2), 8 * 3(%1))
+        DCT_8_INV_ROW_MMX(4 * 16(%0), 4 * 16(%0), 64 * 0(%2), 8 * 4(%1))
+        DCT_8_INV_ROW_MMX(5 * 16(%0), 5 * 16(%0), 64 * 3(%2), 8 * 5(%1))
+        DCT_8_INV_ROW_MMX(6 * 16(%0), 6 * 16(%0), 64 * 2(%2), 8 * 6(%1))
+        DCT_8_INV_ROW_MMX(7 * 16(%0), 7 * 16(%0), 64 * 1(%2), 8 * 7(%1))
+
+        // # Process the columns (4 at a time)
+        DCT_8_INV_COL(0(%0), 0(%0))
+        DCT_8_INV_COL(8(%0), 8(%0))
+        :: "r" (block), "r" (rounder_0), "r" (tab_i_04_mmx), "r" (tg_1_16));
+}
+
+void ff_xvid_idct_mmx_put(uint8_t *dest, int line_size, int16_t *block)
+{
+    ff_xvid_idct_mmx(block);
+    ff_put_pixels_clamped_mmx(block, dest, line_size);
+}
+
+void ff_xvid_idct_mmx_add(uint8_t *dest, int line_size, int16_t *block)
+{
+    ff_xvid_idct_mmx(block);
+    ff_add_pixels_clamped_mmx(block, dest, line_size);
+}
+
+#endif /* HAVE_MMX_INLINE */
+
+#if HAVE_MMXEXT_INLINE
+
+// -----------------------------------------------------------------------------
+// void idct_xmm(uint16_t block[64]);
+// -----------------------------------------------------------------------------
+
+void ff_xvid_idct_mmxext(short *block)
+{
+    __asm__ volatile (
+        // # Process each row
+        DCT_8_INV_ROW_XMM(0 * 16(%0), 0 * 16(%0), 64 * 0(%2), 8 * 0(%1))
+        DCT_8_INV_ROW_XMM(1 * 16(%0), 1 * 16(%0), 64 * 1(%2), 8 * 1(%1))
+        DCT_8_INV_ROW_XMM(2 * 16(%0), 2 * 16(%0), 64 * 2(%2), 8 * 2(%1))
+        DCT_8_INV_ROW_XMM(3 * 16(%0), 3 * 16(%0), 64 * 3(%2), 8 * 3(%1))
+        DCT_8_INV_ROW_XMM(4 * 16(%0), 4 * 16(%0), 64 * 0(%2), 8 * 4(%1))
+        DCT_8_INV_ROW_XMM(5 * 16(%0), 5 * 16(%0), 64 * 3(%2), 8 * 5(%1))
+        DCT_8_INV_ROW_XMM(6 * 16(%0), 6 * 16(%0), 64 * 2(%2), 8 * 6(%1))
+        DCT_8_INV_ROW_XMM(7 * 16(%0), 7 * 16(%0), 64 * 1(%2), 8 * 7(%1))
+
+        // # Process the columns (4 at a time)
+        DCT_8_INV_COL(0(%0), 0(%0))
+        DCT_8_INV_COL(8(%0), 8(%0))
+        :: "r" (block), "r" (rounder_0), "r" (tab_i_04_xmm), "r" (tg_1_16));
+}
+
+void ff_xvid_idct_mmxext_put(uint8_t *dest, int line_size, int16_t *block)
+{
+    ff_xvid_idct_mmxext(block);
+    ff_put_pixels_clamped_mmx(block, dest, line_size);
+}
+
+void ff_xvid_idct_mmxext_add(uint8_t *dest, int line_size, int16_t *block)
+{
+    ff_xvid_idct_mmxext(block);
+    ff_add_pixels_clamped_mmx(block, dest, line_size);
+}
+
+#endif /* HAVE_MMXEXT_INLINE */
index 9a87df0..d4f0169 100644 (file)
  * @brief SSE2 IDCT compatible with the Xvid IDCT
  */
 
-#define X8(x)     x,x,x,x,x,x,x,x
+#define X8(x) x, x, x, x, x, x, x, x
 
-DECLARE_ASM_CONST(16, int16_t, tan1)[] = {X8(13036)}; // tan( pi/16)
-DECLARE_ASM_CONST(16, int16_t, tan2)[] = {X8(27146)}; // tan(2pi/16) = sqrt(2)-1
-DECLARE_ASM_CONST(16, int16_t, tan3)[] = {X8(43790)}; // tan(3pi/16)-1
-DECLARE_ASM_CONST(16, int16_t, sqrt2)[]= {X8(23170)}; // 0.5/sqrt(2)
-DECLARE_ASM_CONST(8,  uint8_t, m127)[] = {X8(127)};
+DECLARE_ASM_CONST(16, int16_t, tan1)[]  = { X8(13036) }; // tan( pi/16)
+DECLARE_ASM_CONST(16, int16_t, tan2)[]  = { X8(27146) }; // tan(2pi/16) = sqrt(2)-1
+DECLARE_ASM_CONST(16, int16_t, tan3)[]  = { X8(43790) }; // tan(3pi/16)-1
+DECLARE_ASM_CONST(16, int16_t, sqrt2)[] = { X8(23170) }; // 0.5/sqrt(2)
+DECLARE_ASM_CONST(8,  uint8_t, m127)[]  = { X8(127) };
 
 DECLARE_ASM_CONST(16, int16_t, iTab1)[] = {
- 0x4000, 0x539f, 0xc000, 0xac61, 0x4000, 0xdd5d, 0x4000, 0xdd5d,
- 0x4000, 0x22a3, 0x4000, 0x22a3, 0xc000, 0x539f, 0x4000, 0xac61,
- 0x3249, 0x11a8, 0x4b42, 0xee58, 0x11a8, 0x4b42, 0x11a8, 0xcdb7,
- 0x58c5, 0x4b42, 0xa73b, 0xcdb7, 0x3249, 0xa73b, 0x4b42, 0xa73b
   0x4000, 0x539f, 0xc000, 0xac61, 0x4000, 0xdd5d, 0x4000, 0xdd5d,
   0x4000, 0x22a3, 0x4000, 0x22a3, 0xc000, 0x539f, 0x4000, 0xac61,
   0x3249, 0x11a8, 0x4b42, 0xee58, 0x11a8, 0x4b42, 0x11a8, 0xcdb7,
   0x58c5, 0x4b42, 0xa73b, 0xcdb7, 0x3249, 0xa73b, 0x4b42, 0xa73b
 };
 
 DECLARE_ASM_CONST(16, int16_t, iTab2)[] = {
- 0x58c5, 0x73fc, 0xa73b, 0x8c04, 0x58c5, 0xcff5, 0x58c5, 0xcff5,
- 0x58c5, 0x300b, 0x58c5, 0x300b, 0xa73b, 0x73fc, 0x58c5, 0x8c04,
- 0x45bf, 0x187e, 0x6862, 0xe782, 0x187e, 0x6862, 0x187e, 0xba41,
- 0x7b21, 0x6862, 0x84df, 0xba41, 0x45bf, 0x84df, 0x6862, 0x84df
   0x58c5, 0x73fc, 0xa73b, 0x8c04, 0x58c5, 0xcff5, 0x58c5, 0xcff5,
   0x58c5, 0x300b, 0x58c5, 0x300b, 0xa73b, 0x73fc, 0x58c5, 0x8c04,
   0x45bf, 0x187e, 0x6862, 0xe782, 0x187e, 0x6862, 0x187e, 0xba41,
   0x7b21, 0x6862, 0x84df, 0xba41, 0x45bf, 0x84df, 0x6862, 0x84df
 };
 
 DECLARE_ASM_CONST(16, int16_t, iTab3)[] = {
- 0x539f, 0x6d41, 0xac61, 0x92bf, 0x539f, 0xd2bf, 0x539f, 0xd2bf,
- 0x539f, 0x2d41, 0x539f, 0x2d41, 0xac61, 0x6d41, 0x539f, 0x92bf,
- 0x41b3, 0x1712, 0x6254, 0xe8ee, 0x1712, 0x6254, 0x1712, 0xbe4d,
- 0x73fc, 0x6254, 0x8c04, 0xbe4d, 0x41b3, 0x8c04, 0x6254, 0x8c04
   0x539f, 0x6d41, 0xac61, 0x92bf, 0x539f, 0xd2bf, 0x539f, 0xd2bf,
   0x539f, 0x2d41, 0x539f, 0x2d41, 0xac61, 0x6d41, 0x539f, 0x92bf,
   0x41b3, 0x1712, 0x6254, 0xe8ee, 0x1712, 0x6254, 0x1712, 0xbe4d,
   0x73fc, 0x6254, 0x8c04, 0xbe4d, 0x41b3, 0x8c04, 0x6254, 0x8c04
 };
 
 DECLARE_ASM_CONST(16, int16_t, iTab4)[] = {
- 0x4b42, 0x6254, 0xb4be, 0x9dac, 0x4b42, 0xd746, 0x4b42, 0xd746,
- 0x4b42, 0x28ba, 0x4b42, 0x28ba, 0xb4be, 0x6254, 0x4b42, 0x9dac,
- 0x3b21, 0x14c3, 0x587e, 0xeb3d, 0x14c3, 0x587e, 0x14c3, 0xc4df,
- 0x6862, 0x587e, 0x979e, 0xc4df, 0x3b21, 0x979e, 0x587e, 0x979e
   0x4b42, 0x6254, 0xb4be, 0x9dac, 0x4b42, 0xd746, 0x4b42, 0xd746,
   0x4b42, 0x28ba, 0x4b42, 0x28ba, 0xb4be, 0x6254, 0x4b42, 0x9dac,
   0x3b21, 0x14c3, 0x587e, 0xeb3d, 0x14c3, 0x587e, 0x14c3, 0xc4df,
   0x6862, 0x587e, 0x979e, 0xc4df, 0x3b21, 0x979e, 0x587e, 0x979e
 };
 
 DECLARE_ASM_CONST(16, int32_t, walkenIdctRounders)[] = {
- 65536, 65536, 65536, 65536,
-  3597,  3597,  3597,  3597,
-  2260,  2260,  2260,  2260,
-  1203,  1203,  1203,  1203,
-   120,   120,   120,   120,
-   512,   512,   512,   512
   65536, 65536, 65536, 65536,
+     3597,  3597,  3597,  3597,
+     2260,  2260,  2260,  2260,
+     1203,  1203,  1203,  1203,
+      120,   120,   120,   120,
+      512,   512,   512,   512
 };
 
 // Temporary storage before the column pass
@@ -164,46 +164,46 @@ DECLARE_ASM_CONST(16, int32_t, walkenIdctRounders)[] = {
     "pmovmskb %%mm1, "reg"            \n\t"
 
 #define TEST_TWO_ROWS(row1, row2, reg1, reg2, clear1, clear2) \
-    clear1                                  \
-    clear2                                  \
-    "movq     "row1", %%mm1           \n\t" \
-    "por    8+"row1", %%mm1           \n\t" \
-    "movq     "row2", %%mm2           \n\t" \
-    "por    8+"row2", %%mm2           \n\t" \
-    "paddusb   %%mm0, %%mm1           \n\t" \
-    "paddusb   %%mm0, %%mm2           \n\t" \
-    "pmovmskb  %%mm1, "reg1"          \n\t" \
+    clear1                                                    \
+    clear2                                                    \
+    "movq     "row1", %%mm1           \n\t"                   \
+    "por    8+"row1", %%mm1           \n\t"                   \
+    "movq     "row2", %%mm2           \n\t"                   \
+    "por    8+"row2", %%mm2           \n\t"                   \
+    "paddusb   %%mm0, %%mm1           \n\t"                   \
+    "paddusb   %%mm0, %%mm2           \n\t"                   \
+    "pmovmskb  %%mm1, "reg1"          \n\t"                   \
     "pmovmskb  %%mm2, "reg2"          \n\t"
 
-///IDCT pass on rows.
-#define iMTX_MULT(src, table, rounder, put) \
-    "movdqa        "src", %%xmm3      \n\t" \
-    "movdqa       %%xmm3, %%xmm0      \n\t" \
+/// IDCT pass on rows.
+#define iMTX_MULT(src, table, rounder, put)            \
+    "movdqa        "src", %%xmm3      \n\t"            \
+    "movdqa       %%xmm3, %%xmm0      \n\t"            \
     "pshufd   $0x11, %%xmm3, %%xmm1   \n\t" /* 4602 */ \
     "punpcklqdq   %%xmm0, %%xmm0      \n\t" /* 0246 */ \
-    "pmaddwd     "table", %%xmm0      \n\t" \
-    "pmaddwd  16+"table", %%xmm1      \n\t" \
+    "pmaddwd     "table", %%xmm0      \n\t"            \
+    "pmaddwd  16+"table", %%xmm1      \n\t"            \
     "pshufd   $0xBB, %%xmm3, %%xmm2   \n\t" /* 5713 */ \
     "punpckhqdq   %%xmm3, %%xmm3      \n\t" /* 1357 */ \
-    "pmaddwd  32+"table", %%xmm2      \n\t" \
-    "pmaddwd  48+"table", %%xmm3      \n\t" \
-    "paddd        %%xmm1, %%xmm0      \n\t" \
-    "paddd        %%xmm3, %%xmm2      \n\t" \
-    rounder",     %%xmm0              \n\t" \
-    "movdqa       %%xmm2, %%xmm3      \n\t" \
-    "paddd        %%xmm0, %%xmm2      \n\t" \
-    "psubd        %%xmm3, %%xmm0      \n\t" \
-    "psrad           $11, %%xmm2      \n\t" \
-    "psrad           $11, %%xmm0      \n\t" \
-    "packssdw     %%xmm0, %%xmm2      \n\t" \
-    put                                     \
+    "pmaddwd  32+"table", %%xmm2      \n\t"            \
+    "pmaddwd  48+"table", %%xmm3      \n\t"            \
+    "paddd        %%xmm1, %%xmm0      \n\t"            \
+    "paddd        %%xmm3, %%xmm2      \n\t"            \
+    rounder",     %%xmm0              \n\t"            \
+    "movdqa       %%xmm2, %%xmm3      \n\t"            \
+    "paddd        %%xmm0, %%xmm2      \n\t"            \
+    "psubd        %%xmm3, %%xmm0      \n\t"            \
+    "psrad           $11, %%xmm2      \n\t"            \
+    "psrad           $11, %%xmm0      \n\t"            \
+    "packssdw     %%xmm0, %%xmm2      \n\t"            \
+    put                                                \
     "1:                               \n\t"
 
 #define iLLM_HEAD                           \
     "movdqa   "MANGLE(tan3)", "TAN3"  \n\t" \
     "movdqa   "MANGLE(tan1)", "TAN1"  \n\t" \
 
-///IDCT pass on columns.
+/// IDCT pass on columns.
 #define iLLM_PASS(dct)                      \
     "movdqa   "TAN3", %%xmm1          \n\t" \
     "movdqa   "TAN1", %%xmm3          \n\t" \
@@ -281,7 +281,7 @@ DECLARE_ASM_CONST(16, int32_t, walkenIdctRounders)[] = {
     "movdqa   "REG4", 4*16("dct")     \n\t" \
     "movdqa   %%xmm7, 7*16("dct")     \n\t"
 
-///IDCT pass on columns, assuming rows 4-7 are zero.
+/// IDCT pass on columns, assuming rows 4-7 are zero.
 #define iLLM_PASS_SPARSE(dct)               \
     "pmulhw   %%xmm4, "TAN3"          \n\t" \
     "paddsw   %%xmm4, "TAN3"          \n\t" \
@@ -344,51 +344,50 @@ DECLARE_ASM_CONST(16, int32_t, walkenIdctRounders)[] = {
 
 inline void ff_xvid_idct_sse2(short *block)
 {
-    __asm__ volatile(
-    "movq     "MANGLE(m127)", %%mm0                              \n\t"
-    iMTX_MULT("(%0)",     MANGLE(iTab1), ROUND(walkenIdctRounders),      PUT_EVEN(ROW0))
-    iMTX_MULT("1*16(%0)", MANGLE(iTab2), ROUND(walkenIdctRounders+1*16), PUT_ODD(ROW1))
-    iMTX_MULT("2*16(%0)", MANGLE(iTab3), ROUND(walkenIdctRounders+2*16), PUT_EVEN(ROW2))
-
-    TEST_TWO_ROWS("3*16(%0)", "4*16(%0)", "%%eax", "%%ecx", CLEAR_ODD(ROW3), CLEAR_EVEN(ROW4))
-    JZ("%%eax", "1f")
-    iMTX_MULT("3*16(%0)", MANGLE(iTab4), ROUND(walkenIdctRounders+3*16), PUT_ODD(ROW3))
-
-    TEST_TWO_ROWS("5*16(%0)", "6*16(%0)", "%%eax", "%%edx", CLEAR_ODD(ROW5), CLEAR_EVEN(ROW6))
-    TEST_ONE_ROW("7*16(%0)", "%%esi", CLEAR_ODD(ROW7))
-    iLLM_HEAD
-    ".p2align 4 \n\t"
-    JNZ("%%ecx", "2f")
-    JNZ("%%eax", "3f")
-    JNZ("%%edx", "4f")
-    JNZ("%%esi", "5f")
-    iLLM_PASS_SPARSE("%0")
-    "jmp 6f                                                      \n\t"
-    "2:                                                          \n\t"
-    iMTX_MULT("4*16(%0)", MANGLE(iTab1), "#", PUT_EVEN(ROW4))
-    "3:                                                          \n\t"
-    iMTX_MULT("5*16(%0)", MANGLE(iTab4), ROUND(walkenIdctRounders+4*16), PUT_ODD(ROW5))
-    JZ("%%edx", "1f")
-    "4:                                                          \n\t"
-    iMTX_MULT("6*16(%0)", MANGLE(iTab3), ROUND(walkenIdctRounders+5*16), PUT_EVEN(ROW6))
-    JZ("%%esi", "1f")
-    "5:                                                          \n\t"
-    iMTX_MULT("7*16(%0)", MANGLE(iTab2), ROUND(walkenIdctRounders+5*16), PUT_ODD(ROW7))
+    __asm__ volatile (
+        "movq     "MANGLE (m127) ", %%mm0                              \n\t"
+        iMTX_MULT("(%0)",     MANGLE(iTab1), ROUND(walkenIdctRounders),          PUT_EVEN(ROW0))
+        iMTX_MULT("1*16(%0)", MANGLE(iTab2), ROUND(walkenIdctRounders + 1 * 16), PUT_ODD(ROW1))
+        iMTX_MULT("2*16(%0)", MANGLE(iTab3), ROUND(walkenIdctRounders + 2 * 16), PUT_EVEN(ROW2))
+
+        TEST_TWO_ROWS("3*16(%0)", "4*16(%0)", "%%eax", "%%ecx", CLEAR_ODD(ROW3), CLEAR_EVEN(ROW4))
+        JZ("%%eax", "1f")
+        iMTX_MULT("3*16(%0)", MANGLE(iTab4), ROUND(walkenIdctRounders + 3 * 16), PUT_ODD(ROW3))
+
+        TEST_TWO_ROWS("5*16(%0)", "6*16(%0)", "%%eax", "%%edx", CLEAR_ODD(ROW5), CLEAR_EVEN(ROW6))
+        TEST_ONE_ROW("7*16(%0)", "%%esi", CLEAR_ODD(ROW7))
+        iLLM_HEAD
+        ".p2align 4 \n\t"
+        JNZ("%%ecx", "2f")
+        JNZ("%%eax", "3f")
+        JNZ("%%edx", "4f")
+        JNZ("%%esi", "5f")
+        iLLM_PASS_SPARSE("%0")
+        "jmp 6f                                                      \n\t"
+        "2:                                                          \n\t"
+        iMTX_MULT("4*16(%0)", MANGLE(iTab1), "#", PUT_EVEN(ROW4))
+        "3:                                                          \n\t"
+        iMTX_MULT("5*16(%0)", MANGLE(iTab4), ROUND(walkenIdctRounders + 4 * 16), PUT_ODD(ROW5))
+        JZ("%%edx", "1f")
+        "4:                                                          \n\t"
+        iMTX_MULT("6*16(%0)", MANGLE(iTab3), ROUND(walkenIdctRounders + 5 * 16), PUT_EVEN(ROW6))
+        JZ("%%esi", "1f")
+        "5:                                                          \n\t"
+        iMTX_MULT("7*16(%0)", MANGLE(iTab2), ROUND(walkenIdctRounders + 5 * 16), PUT_ODD(ROW7))
 #if ARCH_X86_32
-    iLLM_HEAD
+        iLLM_HEAD
 #endif
-    iLLM_PASS("%0")
-    "6:                                                          \n\t"
-    : "+r"(block)
-    :
-    : XMM_CLOBBERS("%xmm0" , "%xmm1" , "%xmm2" , "%xmm3" ,
-                   "%xmm4" , "%xmm5" , "%xmm6" , "%xmm7" ,)
+        iLLM_PASS("%0")
+        "6:                                                          \n\t"
+        : "+r" (block)
+        :
+        : XMM_CLOBBERS("%xmm0", "%xmm1", "%xmm2", "%xmm3",
+                       "%xmm4", "%xmm5", "%xmm6", "%xmm7", )
 #if ARCH_X86_64
-      XMM_CLOBBERS("%xmm8" , "%xmm9" , "%xmm10", "%xmm11",
-                   "%xmm12", "%xmm13", "%xmm14",)
+          XMM_CLOBBERS("%xmm8", "%xmm9", "%xmm10", "%xmm11",
+                       "%xmm12", "%xmm13", "%xmm14", )
 #endif
-      "%eax", "%ecx", "%edx", "%esi", "memory"
-    );
+          "%eax", "%ecx", "%edx", "%esi", "memory");
 }
 
 void ff_xvid_idct_sse2_put(uint8_t *dest, int line_size, short *block)