swscale: x86: Drop pointless _mmx suffix from filenames
authorDiego Biurrun <diego@biurrun.de>
Sat, 21 Jul 2012 20:33:03 +0000 (22:33 +0200)
committerDiego Biurrun <diego@biurrun.de>
Sun, 22 Jul 2012 23:30:04 +0000 (01:30 +0200)
The files do not contain only MMX code.

libswscale/x86/Makefile
libswscale/x86/swscale.c [new file with mode: 0644]
libswscale/x86/swscale_mmx.c [deleted file]
libswscale/x86/yuv2rgb.c [new file with mode: 0644]
libswscale/x86/yuv2rgb_mmx.c [deleted file]

index 7f37799..5416d48 100644 (file)
@@ -1,8 +1,8 @@
 OBJS-$(CONFIG_XMM_CLOBBER_TEST) += x86/w64xmmtest.o
 
 MMX-OBJS                        += x86/rgb2rgb.o                        \
-                                   x86/swscale_mmx.o                    \
-                                   x86/yuv2rgb_mmx.o                    \
+                                   x86/swscale.o                        \
+                                   x86/yuv2rgb.o                        \
 
 YASM-OBJS                       += x86/input.o                          \
                                    x86/output.o                         \
diff --git a/libswscale/x86/swscale.c b/libswscale/x86/swscale.c
new file mode 100644 (file)
index 0000000..b57b7a4
--- /dev/null
@@ -0,0 +1,493 @@
+/*
+ * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
+ *
+ * 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 "libswscale/swscale.h"
+#include "libswscale/swscale_internal.h"
+#include "libavutil/intreadwrite.h"
+#include "libavutil/x86_cpu.h"
+#include "libavutil/cpu.h"
+#include "libavutil/pixdesc.h"
+
+#if HAVE_INLINE_ASM
+
+#define DITHER1XBPP
+
+DECLARE_ASM_CONST(8, uint64_t, bF8)=       0xF8F8F8F8F8F8F8F8LL;
+DECLARE_ASM_CONST(8, uint64_t, bFC)=       0xFCFCFCFCFCFCFCFCLL;
+DECLARE_ASM_CONST(8, uint64_t, w10)=       0x0010001000100010LL;
+DECLARE_ASM_CONST(8, uint64_t, w02)=       0x0002000200020002LL;
+
+const DECLARE_ALIGNED(8, uint64_t, ff_dither4)[2] = {
+    0x0103010301030103LL,
+    0x0200020002000200LL,};
+
+const DECLARE_ALIGNED(8, uint64_t, ff_dither8)[2] = {
+    0x0602060206020602LL,
+    0x0004000400040004LL,};
+
+DECLARE_ASM_CONST(8, uint64_t, b16Mask)=   0x001F001F001F001FLL;
+DECLARE_ASM_CONST(8, uint64_t, g16Mask)=   0x07E007E007E007E0LL;
+DECLARE_ASM_CONST(8, uint64_t, r16Mask)=   0xF800F800F800F800LL;
+DECLARE_ASM_CONST(8, uint64_t, b15Mask)=   0x001F001F001F001FLL;
+DECLARE_ASM_CONST(8, uint64_t, g15Mask)=   0x03E003E003E003E0LL;
+DECLARE_ASM_CONST(8, uint64_t, r15Mask)=   0x7C007C007C007C00LL;
+
+DECLARE_ALIGNED(8, const uint64_t, ff_M24A)         = 0x00FF0000FF0000FFLL;
+DECLARE_ALIGNED(8, const uint64_t, ff_M24B)         = 0xFF0000FF0000FF00LL;
+DECLARE_ALIGNED(8, const uint64_t, ff_M24C)         = 0x0000FF0000FF0000LL;
+
+#ifdef FAST_BGR2YV12
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YCoeff)   = 0x000000210041000DULL;
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UCoeff)   = 0x0000FFEEFFDC0038ULL;
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2VCoeff)   = 0x00000038FFD2FFF8ULL;
+#else
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YCoeff)   = 0x000020E540830C8BULL;
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UCoeff)   = 0x0000ED0FDAC23831ULL;
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2VCoeff)   = 0x00003831D0E6F6EAULL;
+#endif /* FAST_BGR2YV12 */
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YOffset)  = 0x1010101010101010ULL;
+DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UVOffset) = 0x8080808080808080ULL;
+DECLARE_ALIGNED(8, const uint64_t, ff_w1111)        = 0x0001000100010001ULL;
+
+//MMX versions
+#if HAVE_MMX
+#undef RENAME
+#define COMPILE_TEMPLATE_MMX2 0
+#define RENAME(a) a ## _MMX
+#include "swscale_template.c"
+#endif
+
+//MMX2 versions
+#if HAVE_MMX2
+#undef RENAME
+#undef COMPILE_TEMPLATE_MMX2
+#define COMPILE_TEMPLATE_MMX2 1
+#define RENAME(a) a ## _MMX2
+#include "swscale_template.c"
+#endif
+
+void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
+                           int lastInLumBuf, int lastInChrBuf)
+{
+    const int dstH= c->dstH;
+    const int flags= c->flags;
+    int16_t **lumPixBuf= c->lumPixBuf;
+    int16_t **chrUPixBuf= c->chrUPixBuf;
+    int16_t **alpPixBuf= c->alpPixBuf;
+    const int vLumBufSize= c->vLumBufSize;
+    const int vChrBufSize= c->vChrBufSize;
+    int32_t *vLumFilterPos= c->vLumFilterPos;
+    int32_t *vChrFilterPos= c->vChrFilterPos;
+    int16_t *vLumFilter= c->vLumFilter;
+    int16_t *vChrFilter= c->vChrFilter;
+    int32_t *lumMmxFilter= c->lumMmxFilter;
+    int32_t *chrMmxFilter= c->chrMmxFilter;
+    int32_t av_unused *alpMmxFilter= c->alpMmxFilter;
+    const int vLumFilterSize= c->vLumFilterSize;
+    const int vChrFilterSize= c->vChrFilterSize;
+    const int chrDstY= dstY>>c->chrDstVSubSample;
+    const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input
+    const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input
+
+    c->blueDither= ff_dither8[dstY&1];
+    if (c->dstFormat == PIX_FMT_RGB555 || c->dstFormat == PIX_FMT_BGR555)
+        c->greenDither= ff_dither8[dstY&1];
+    else
+        c->greenDither= ff_dither4[dstY&1];
+    c->redDither= ff_dither8[(dstY+1)&1];
+    if (dstY < dstH - 2) {
+        const int16_t **lumSrcPtr= (const int16_t **) lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize;
+        const int16_t **chrUSrcPtr= (const int16_t **) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
+        const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL;
+        int i;
+
+        if (firstLumSrcY < 0 || firstLumSrcY + vLumFilterSize > c->srcH) {
+            const int16_t **tmpY = (const int16_t **) lumPixBuf + 2 * vLumBufSize;
+            int neg = -firstLumSrcY, i, end = FFMIN(c->srcH - firstLumSrcY, vLumFilterSize);
+            for (i = 0; i < neg;            i++)
+                tmpY[i] = lumSrcPtr[neg];
+            for (     ; i < end;            i++)
+                tmpY[i] = lumSrcPtr[i];
+            for (     ; i < vLumFilterSize; i++)
+                tmpY[i] = tmpY[i-1];
+            lumSrcPtr = tmpY;
+
+            if (alpSrcPtr) {
+                const int16_t **tmpA = (const int16_t **) alpPixBuf + 2 * vLumBufSize;
+                for (i = 0; i < neg;            i++)
+                    tmpA[i] = alpSrcPtr[neg];
+                for (     ; i < end;            i++)
+                    tmpA[i] = alpSrcPtr[i];
+                for (     ; i < vLumFilterSize; i++)
+                    tmpA[i] = tmpA[i - 1];
+                alpSrcPtr = tmpA;
+            }
+        }
+        if (firstChrSrcY < 0 || firstChrSrcY + vChrFilterSize > c->chrSrcH) {
+            const int16_t **tmpU = (const int16_t **) chrUPixBuf + 2 * vChrBufSize;
+            int neg = -firstChrSrcY, i, end = FFMIN(c->chrSrcH - firstChrSrcY, vChrFilterSize);
+            for (i = 0; i < neg;            i++) {
+                tmpU[i] = chrUSrcPtr[neg];
+            }
+            for (     ; i < end;            i++) {
+                tmpU[i] = chrUSrcPtr[i];
+            }
+            for (     ; i < vChrFilterSize; i++) {
+                tmpU[i] = tmpU[i - 1];
+            }
+            chrUSrcPtr = tmpU;
+        }
+
+        if (flags & SWS_ACCURATE_RND) {
+            int s= APCK_SIZE / 8;
+            for (i=0; i<vLumFilterSize; i+=2) {
+                *(const void**)&lumMmxFilter[s*i              ]= lumSrcPtr[i  ];
+                *(const void**)&lumMmxFilter[s*i+APCK_PTR2/4  ]= lumSrcPtr[i+(vLumFilterSize>1)];
+                lumMmxFilter[s*i+APCK_COEF/4  ]=
+                lumMmxFilter[s*i+APCK_COEF/4+1]= vLumFilter[dstY*vLumFilterSize + i    ]
+                + (vLumFilterSize>1 ? vLumFilter[dstY*vLumFilterSize + i + 1]<<16 : 0);
+                if (CONFIG_SWSCALE_ALPHA && alpPixBuf) {
+                    *(const void**)&alpMmxFilter[s*i              ]= alpSrcPtr[i  ];
+                    *(const void**)&alpMmxFilter[s*i+APCK_PTR2/4  ]= alpSrcPtr[i+(vLumFilterSize>1)];
+                    alpMmxFilter[s*i+APCK_COEF/4  ]=
+                    alpMmxFilter[s*i+APCK_COEF/4+1]= lumMmxFilter[s*i+APCK_COEF/4  ];
+                }
+            }
+            for (i=0; i<vChrFilterSize; i+=2) {
+                *(const void**)&chrMmxFilter[s*i              ]= chrUSrcPtr[i  ];
+                *(const void**)&chrMmxFilter[s*i+APCK_PTR2/4  ]= chrUSrcPtr[i+(vChrFilterSize>1)];
+                chrMmxFilter[s*i+APCK_COEF/4  ]=
+                chrMmxFilter[s*i+APCK_COEF/4+1]= vChrFilter[chrDstY*vChrFilterSize + i    ]
+                + (vChrFilterSize>1 ? vChrFilter[chrDstY*vChrFilterSize + i + 1]<<16 : 0);
+            }
+        } else {
+            for (i=0; i<vLumFilterSize; i++) {
+                *(const void**)&lumMmxFilter[4*i+0]= lumSrcPtr[i];
+                lumMmxFilter[4*i+2]=
+                lumMmxFilter[4*i+3]=
+                ((uint16_t)vLumFilter[dstY*vLumFilterSize + i])*0x10001;
+                if (CONFIG_SWSCALE_ALPHA && alpPixBuf) {
+                    *(const void**)&alpMmxFilter[4*i+0]= alpSrcPtr[i];
+                    alpMmxFilter[4*i+2]=
+                    alpMmxFilter[4*i+3]= lumMmxFilter[4*i+2];
+                }
+            }
+            for (i=0; i<vChrFilterSize; i++) {
+                *(const void**)&chrMmxFilter[4*i+0]= chrUSrcPtr[i];
+                chrMmxFilter[4*i+2]=
+                chrMmxFilter[4*i+3]=
+                ((uint16_t)vChrFilter[chrDstY*vChrFilterSize + i])*0x10001;
+            }
+        }
+    }
+}
+
+#endif /* HAVE_INLINE_ASM */
+
+#define SCALE_FUNC(filter_n, from_bpc, to_bpc, opt) \
+extern void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt( \
+                                                SwsContext *c, int16_t *data, \
+                                                int dstW, const uint8_t *src, \
+                                                const int16_t *filter, \
+                                                const int32_t *filterPos, int filterSize)
+
+#define SCALE_FUNCS(filter_n, opt) \
+    SCALE_FUNC(filter_n,  8, 15, opt); \
+    SCALE_FUNC(filter_n,  9, 15, opt); \
+    SCALE_FUNC(filter_n, 10, 15, opt); \
+    SCALE_FUNC(filter_n, 16, 15, opt); \
+    SCALE_FUNC(filter_n,  8, 19, opt); \
+    SCALE_FUNC(filter_n,  9, 19, opt); \
+    SCALE_FUNC(filter_n, 10, 19, opt); \
+    SCALE_FUNC(filter_n, 16, 19, opt)
+
+#define SCALE_FUNCS_MMX(opt) \
+    SCALE_FUNCS(4, opt); \
+    SCALE_FUNCS(8, opt); \
+    SCALE_FUNCS(X, opt)
+
+#define SCALE_FUNCS_SSE(opt) \
+    SCALE_FUNCS(4, opt); \
+    SCALE_FUNCS(8, opt); \
+    SCALE_FUNCS(X4, opt); \
+    SCALE_FUNCS(X8, opt)
+
+#if ARCH_X86_32
+SCALE_FUNCS_MMX(mmx);
+#endif
+SCALE_FUNCS_SSE(sse2);
+SCALE_FUNCS_SSE(ssse3);
+SCALE_FUNCS_SSE(sse4);
+
+#define VSCALEX_FUNC(size, opt) \
+extern void ff_yuv2planeX_ ## size ## _ ## opt(const int16_t *filter, int filterSize, \
+                                               const int16_t **src, uint8_t *dest, int dstW, \
+                                               const uint8_t *dither, int offset)
+#define VSCALEX_FUNCS(opt) \
+    VSCALEX_FUNC(8,  opt); \
+    VSCALEX_FUNC(9,  opt); \
+    VSCALEX_FUNC(10, opt)
+
+#if ARCH_X86_32
+VSCALEX_FUNCS(mmx2);
+#endif
+VSCALEX_FUNCS(sse2);
+VSCALEX_FUNCS(sse4);
+VSCALEX_FUNC(16, sse4);
+VSCALEX_FUNCS(avx);
+
+#define VSCALE_FUNC(size, opt) \
+extern void ff_yuv2plane1_ ## size ## _ ## opt(const int16_t *src, uint8_t *dst, int dstW, \
+                                               const uint8_t *dither, int offset)
+#define VSCALE_FUNCS(opt1, opt2) \
+    VSCALE_FUNC(8,  opt1); \
+    VSCALE_FUNC(9,  opt2); \
+    VSCALE_FUNC(10, opt2); \
+    VSCALE_FUNC(16, opt1)
+
+#if ARCH_X86_32
+VSCALE_FUNCS(mmx, mmx2);
+#endif
+VSCALE_FUNCS(sse2, sse2);
+VSCALE_FUNC(16, sse4);
+VSCALE_FUNCS(avx, avx);
+
+#define INPUT_Y_FUNC(fmt, opt) \
+extern void ff_ ## fmt ## ToY_  ## opt(uint8_t *dst, const uint8_t *src, \
+                                       int w, uint32_t *unused)
+#define INPUT_UV_FUNC(fmt, opt) \
+extern void ff_ ## fmt ## ToUV_ ## opt(uint8_t *dstU, uint8_t *dstV, \
+                                       const uint8_t *src, const uint8_t *unused1, \
+                                       int w, uint32_t *unused2)
+#define INPUT_FUNC(fmt, opt) \
+    INPUT_Y_FUNC(fmt, opt); \
+    INPUT_UV_FUNC(fmt, opt)
+#define INPUT_FUNCS(opt) \
+    INPUT_FUNC(uyvy, opt); \
+    INPUT_FUNC(yuyv, opt); \
+    INPUT_UV_FUNC(nv12, opt); \
+    INPUT_UV_FUNC(nv21, opt); \
+    INPUT_FUNC(rgba, opt); \
+    INPUT_FUNC(bgra, opt); \
+    INPUT_FUNC(argb, opt); \
+    INPUT_FUNC(abgr, opt); \
+    INPUT_FUNC(rgb24, opt); \
+    INPUT_FUNC(bgr24, opt)
+
+#if ARCH_X86_32
+INPUT_FUNCS(mmx);
+#endif
+INPUT_FUNCS(sse2);
+INPUT_FUNCS(ssse3);
+INPUT_FUNCS(avx);
+
+void ff_sws_init_swScale_mmx(SwsContext *c)
+{
+    int cpu_flags = av_get_cpu_flags();
+
+#if HAVE_INLINE_ASM
+    if (cpu_flags & AV_CPU_FLAG_MMX)
+        sws_init_swScale_MMX(c);
+#if HAVE_MMX2
+    if (cpu_flags & AV_CPU_FLAG_MMX2)
+        sws_init_swScale_MMX2(c);
+#endif
+#endif /* HAVE_INLINE_ASM */
+
+#if HAVE_YASM
+#define ASSIGN_SCALE_FUNC2(hscalefn, filtersize, opt1, opt2) do { \
+    if (c->srcBpc == 8) { \
+        hscalefn = c->dstBpc <= 10 ? ff_hscale8to15_ ## filtersize ## _ ## opt2 : \
+                                     ff_hscale8to19_ ## filtersize ## _ ## opt1; \
+    } else if (c->srcBpc == 9) { \
+        hscalefn = c->dstBpc <= 10 ? ff_hscale9to15_ ## filtersize ## _ ## opt2 : \
+                                     ff_hscale9to19_ ## filtersize ## _ ## opt1; \
+    } else if (c->srcBpc == 10) { \
+        hscalefn = c->dstBpc <= 10 ? ff_hscale10to15_ ## filtersize ## _ ## opt2 : \
+                                     ff_hscale10to19_ ## filtersize ## _ ## opt1; \
+    } else /* c->srcBpc == 16 */ { \
+        hscalefn = c->dstBpc <= 10 ? ff_hscale16to15_ ## filtersize ## _ ## opt2 : \
+                                     ff_hscale16to19_ ## filtersize ## _ ## opt1; \
+    } \
+} while (0)
+#define ASSIGN_MMX_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
+    switch (filtersize) { \
+    case 4:  ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
+    case 8:  ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
+    default: ASSIGN_SCALE_FUNC2(hscalefn, X, opt1, opt2); break; \
+    }
+#define ASSIGN_VSCALEX_FUNC(vscalefn, opt, do_16_case, condition_8bit) \
+switch(c->dstBpc){ \
+    case 16:                          do_16_case;                          break; \
+    case 10: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_10_ ## opt; break; \
+    case 9:  if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_9_  ## opt; break; \
+    default: if (condition_8bit)      vscalefn = ff_yuv2planeX_8_  ## opt; break; \
+    }
+#define ASSIGN_VSCALE_FUNC(vscalefn, opt1, opt2, opt2chk) \
+    switch(c->dstBpc){ \
+    case 16: if (!isBE(c->dstFormat))            vscalefn = ff_yuv2plane1_16_ ## opt1; break; \
+    case 10: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_10_ ## opt2; break; \
+    case 9:  if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_9_  ## opt2;  break; \
+    default:                                     vscalefn = ff_yuv2plane1_8_  ## opt1;  break; \
+    }
+#define case_rgb(x, X, opt) \
+        case PIX_FMT_ ## X: \
+            c->lumToYV12 = ff_ ## x ## ToY_ ## opt; \
+            if (!c->chrSrcHSubSample) \
+                c->chrToYV12 = ff_ ## x ## ToUV_ ## opt; \
+            break
+#if ARCH_X86_32
+    if (cpu_flags & AV_CPU_FLAG_MMX) {
+        ASSIGN_MMX_SCALE_FUNC(c->hyScale, c->hLumFilterSize, mmx, mmx);
+        ASSIGN_MMX_SCALE_FUNC(c->hcScale, c->hChrFilterSize, mmx, mmx);
+        ASSIGN_VSCALE_FUNC(c->yuv2plane1, mmx, mmx2, cpu_flags & AV_CPU_FLAG_MMX2);
+
+        switch (c->srcFormat) {
+        case PIX_FMT_Y400A:
+            c->lumToYV12 = ff_yuyvToY_mmx;
+            if (c->alpPixBuf)
+                c->alpToYV12 = ff_uyvyToY_mmx;
+            break;
+        case PIX_FMT_YUYV422:
+            c->lumToYV12 = ff_yuyvToY_mmx;
+            c->chrToYV12 = ff_yuyvToUV_mmx;
+            break;
+        case PIX_FMT_UYVY422:
+            c->lumToYV12 = ff_uyvyToY_mmx;
+            c->chrToYV12 = ff_uyvyToUV_mmx;
+            break;
+        case PIX_FMT_NV12:
+            c->chrToYV12 = ff_nv12ToUV_mmx;
+            break;
+        case PIX_FMT_NV21:
+            c->chrToYV12 = ff_nv21ToUV_mmx;
+            break;
+        case_rgb(rgb24, RGB24, mmx);
+        case_rgb(bgr24, BGR24, mmx);
+        case_rgb(bgra,  BGRA,  mmx);
+        case_rgb(rgba,  RGBA,  mmx);
+        case_rgb(abgr,  ABGR,  mmx);
+        case_rgb(argb,  ARGB,  mmx);
+        default:
+            break;
+        }
+    }
+    if (cpu_flags & AV_CPU_FLAG_MMX2) {
+        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, mmx2, , 1);
+    }
+#endif
+#define ASSIGN_SSE_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
+    switch (filtersize) { \
+    case 4:  ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
+    case 8:  ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
+    default: if (filtersize & 4) ASSIGN_SCALE_FUNC2(hscalefn, X4, opt1, opt2); \
+             else                ASSIGN_SCALE_FUNC2(hscalefn, X8, opt1, opt2); \
+             break; \
+    }
+    if (cpu_flags & AV_CPU_FLAG_SSE2) {
+        ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse2, sse2);
+        ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse2, sse2);
+        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse2, ,
+                            HAVE_ALIGNED_STACK || ARCH_X86_64);
+        ASSIGN_VSCALE_FUNC(c->yuv2plane1, sse2, sse2, 1);
+
+        switch (c->srcFormat) {
+        case PIX_FMT_Y400A:
+            c->lumToYV12 = ff_yuyvToY_sse2;
+            if (c->alpPixBuf)
+                c->alpToYV12 = ff_uyvyToY_sse2;
+            break;
+        case PIX_FMT_YUYV422:
+            c->lumToYV12 = ff_yuyvToY_sse2;
+            c->chrToYV12 = ff_yuyvToUV_sse2;
+            break;
+        case PIX_FMT_UYVY422:
+            c->lumToYV12 = ff_uyvyToY_sse2;
+            c->chrToYV12 = ff_uyvyToUV_sse2;
+            break;
+        case PIX_FMT_NV12:
+            c->chrToYV12 = ff_nv12ToUV_sse2;
+            break;
+        case PIX_FMT_NV21:
+            c->chrToYV12 = ff_nv21ToUV_sse2;
+            break;
+        case_rgb(rgb24, RGB24, sse2);
+        case_rgb(bgr24, BGR24, sse2);
+        case_rgb(bgra,  BGRA,  sse2);
+        case_rgb(rgba,  RGBA,  sse2);
+        case_rgb(abgr,  ABGR,  sse2);
+        case_rgb(argb,  ARGB,  sse2);
+        default:
+            break;
+        }
+    }
+    if (cpu_flags & AV_CPU_FLAG_SSSE3) {
+        ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, ssse3, ssse3);
+        ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, ssse3, ssse3);
+        switch (c->srcFormat) {
+        case_rgb(rgb24, RGB24, ssse3);
+        case_rgb(bgr24, BGR24, ssse3);
+        default:
+            break;
+        }
+    }
+    if (cpu_flags & AV_CPU_FLAG_SSE4) {
+        /* Xto15 don't need special sse4 functions */
+        ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse4, ssse3);
+        ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse4, ssse3);
+        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse4,
+                            if (!isBE(c->dstFormat)) c->yuv2planeX = ff_yuv2planeX_16_sse4,
+                            HAVE_ALIGNED_STACK || ARCH_X86_64);
+        if (c->dstBpc == 16 && !isBE(c->dstFormat))
+            c->yuv2plane1 = ff_yuv2plane1_16_sse4;
+    }
+
+    if (cpu_flags & AV_CPU_FLAG_AVX) {
+        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, avx, ,
+                            HAVE_ALIGNED_STACK || ARCH_X86_64);
+        ASSIGN_VSCALE_FUNC(c->yuv2plane1, avx, avx, 1);
+
+        switch (c->srcFormat) {
+        case PIX_FMT_YUYV422:
+            c->chrToYV12 = ff_yuyvToUV_avx;
+            break;
+        case PIX_FMT_UYVY422:
+            c->chrToYV12 = ff_uyvyToUV_avx;
+            break;
+        case PIX_FMT_NV12:
+            c->chrToYV12 = ff_nv12ToUV_avx;
+            break;
+        case PIX_FMT_NV21:
+            c->chrToYV12 = ff_nv21ToUV_avx;
+            break;
+        case_rgb(rgb24, RGB24, avx);
+        case_rgb(bgr24, BGR24, avx);
+        case_rgb(bgra,  BGRA,  avx);
+        case_rgb(rgba,  RGBA,  avx);
+        case_rgb(abgr,  ABGR,  avx);
+        case_rgb(argb,  ARGB,  avx);
+        default:
+            break;
+        }
+    }
+#endif
+}
diff --git a/libswscale/x86/swscale_mmx.c b/libswscale/x86/swscale_mmx.c
deleted file mode 100644 (file)
index b57b7a4..0000000
+++ /dev/null
@@ -1,493 +0,0 @@
-/*
- * Copyright (C) 2001-2003 Michael Niedermayer <michaelni@gmx.at>
- *
- * 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 "libswscale/swscale.h"
-#include "libswscale/swscale_internal.h"
-#include "libavutil/intreadwrite.h"
-#include "libavutil/x86_cpu.h"
-#include "libavutil/cpu.h"
-#include "libavutil/pixdesc.h"
-
-#if HAVE_INLINE_ASM
-
-#define DITHER1XBPP
-
-DECLARE_ASM_CONST(8, uint64_t, bF8)=       0xF8F8F8F8F8F8F8F8LL;
-DECLARE_ASM_CONST(8, uint64_t, bFC)=       0xFCFCFCFCFCFCFCFCLL;
-DECLARE_ASM_CONST(8, uint64_t, w10)=       0x0010001000100010LL;
-DECLARE_ASM_CONST(8, uint64_t, w02)=       0x0002000200020002LL;
-
-const DECLARE_ALIGNED(8, uint64_t, ff_dither4)[2] = {
-    0x0103010301030103LL,
-    0x0200020002000200LL,};
-
-const DECLARE_ALIGNED(8, uint64_t, ff_dither8)[2] = {
-    0x0602060206020602LL,
-    0x0004000400040004LL,};
-
-DECLARE_ASM_CONST(8, uint64_t, b16Mask)=   0x001F001F001F001FLL;
-DECLARE_ASM_CONST(8, uint64_t, g16Mask)=   0x07E007E007E007E0LL;
-DECLARE_ASM_CONST(8, uint64_t, r16Mask)=   0xF800F800F800F800LL;
-DECLARE_ASM_CONST(8, uint64_t, b15Mask)=   0x001F001F001F001FLL;
-DECLARE_ASM_CONST(8, uint64_t, g15Mask)=   0x03E003E003E003E0LL;
-DECLARE_ASM_CONST(8, uint64_t, r15Mask)=   0x7C007C007C007C00LL;
-
-DECLARE_ALIGNED(8, const uint64_t, ff_M24A)         = 0x00FF0000FF0000FFLL;
-DECLARE_ALIGNED(8, const uint64_t, ff_M24B)         = 0xFF0000FF0000FF00LL;
-DECLARE_ALIGNED(8, const uint64_t, ff_M24C)         = 0x0000FF0000FF0000LL;
-
-#ifdef FAST_BGR2YV12
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YCoeff)   = 0x000000210041000DULL;
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UCoeff)   = 0x0000FFEEFFDC0038ULL;
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2VCoeff)   = 0x00000038FFD2FFF8ULL;
-#else
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YCoeff)   = 0x000020E540830C8BULL;
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UCoeff)   = 0x0000ED0FDAC23831ULL;
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2VCoeff)   = 0x00003831D0E6F6EAULL;
-#endif /* FAST_BGR2YV12 */
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YOffset)  = 0x1010101010101010ULL;
-DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UVOffset) = 0x8080808080808080ULL;
-DECLARE_ALIGNED(8, const uint64_t, ff_w1111)        = 0x0001000100010001ULL;
-
-//MMX versions
-#if HAVE_MMX
-#undef RENAME
-#define COMPILE_TEMPLATE_MMX2 0
-#define RENAME(a) a ## _MMX
-#include "swscale_template.c"
-#endif
-
-//MMX2 versions
-#if HAVE_MMX2
-#undef RENAME
-#undef COMPILE_TEMPLATE_MMX2
-#define COMPILE_TEMPLATE_MMX2 1
-#define RENAME(a) a ## _MMX2
-#include "swscale_template.c"
-#endif
-
-void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
-                           int lastInLumBuf, int lastInChrBuf)
-{
-    const int dstH= c->dstH;
-    const int flags= c->flags;
-    int16_t **lumPixBuf= c->lumPixBuf;
-    int16_t **chrUPixBuf= c->chrUPixBuf;
-    int16_t **alpPixBuf= c->alpPixBuf;
-    const int vLumBufSize= c->vLumBufSize;
-    const int vChrBufSize= c->vChrBufSize;
-    int32_t *vLumFilterPos= c->vLumFilterPos;
-    int32_t *vChrFilterPos= c->vChrFilterPos;
-    int16_t *vLumFilter= c->vLumFilter;
-    int16_t *vChrFilter= c->vChrFilter;
-    int32_t *lumMmxFilter= c->lumMmxFilter;
-    int32_t *chrMmxFilter= c->chrMmxFilter;
-    int32_t av_unused *alpMmxFilter= c->alpMmxFilter;
-    const int vLumFilterSize= c->vLumFilterSize;
-    const int vChrFilterSize= c->vChrFilterSize;
-    const int chrDstY= dstY>>c->chrDstVSubSample;
-    const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input
-    const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input
-
-    c->blueDither= ff_dither8[dstY&1];
-    if (c->dstFormat == PIX_FMT_RGB555 || c->dstFormat == PIX_FMT_BGR555)
-        c->greenDither= ff_dither8[dstY&1];
-    else
-        c->greenDither= ff_dither4[dstY&1];
-    c->redDither= ff_dither8[(dstY+1)&1];
-    if (dstY < dstH - 2) {
-        const int16_t **lumSrcPtr= (const int16_t **) lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize;
-        const int16_t **chrUSrcPtr= (const int16_t **) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
-        const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL;
-        int i;
-
-        if (firstLumSrcY < 0 || firstLumSrcY + vLumFilterSize > c->srcH) {
-            const int16_t **tmpY = (const int16_t **) lumPixBuf + 2 * vLumBufSize;
-            int neg = -firstLumSrcY, i, end = FFMIN(c->srcH - firstLumSrcY, vLumFilterSize);
-            for (i = 0; i < neg;            i++)
-                tmpY[i] = lumSrcPtr[neg];
-            for (     ; i < end;            i++)
-                tmpY[i] = lumSrcPtr[i];
-            for (     ; i < vLumFilterSize; i++)
-                tmpY[i] = tmpY[i-1];
-            lumSrcPtr = tmpY;
-
-            if (alpSrcPtr) {
-                const int16_t **tmpA = (const int16_t **) alpPixBuf + 2 * vLumBufSize;
-                for (i = 0; i < neg;            i++)
-                    tmpA[i] = alpSrcPtr[neg];
-                for (     ; i < end;            i++)
-                    tmpA[i] = alpSrcPtr[i];
-                for (     ; i < vLumFilterSize; i++)
-                    tmpA[i] = tmpA[i - 1];
-                alpSrcPtr = tmpA;
-            }
-        }
-        if (firstChrSrcY < 0 || firstChrSrcY + vChrFilterSize > c->chrSrcH) {
-            const int16_t **tmpU = (const int16_t **) chrUPixBuf + 2 * vChrBufSize;
-            int neg = -firstChrSrcY, i, end = FFMIN(c->chrSrcH - firstChrSrcY, vChrFilterSize);
-            for (i = 0; i < neg;            i++) {
-                tmpU[i] = chrUSrcPtr[neg];
-            }
-            for (     ; i < end;            i++) {
-                tmpU[i] = chrUSrcPtr[i];
-            }
-            for (     ; i < vChrFilterSize; i++) {
-                tmpU[i] = tmpU[i - 1];
-            }
-            chrUSrcPtr = tmpU;
-        }
-
-        if (flags & SWS_ACCURATE_RND) {
-            int s= APCK_SIZE / 8;
-            for (i=0; i<vLumFilterSize; i+=2) {
-                *(const void**)&lumMmxFilter[s*i              ]= lumSrcPtr[i  ];
-                *(const void**)&lumMmxFilter[s*i+APCK_PTR2/4  ]= lumSrcPtr[i+(vLumFilterSize>1)];
-                lumMmxFilter[s*i+APCK_COEF/4  ]=
-                lumMmxFilter[s*i+APCK_COEF/4+1]= vLumFilter[dstY*vLumFilterSize + i    ]
-                + (vLumFilterSize>1 ? vLumFilter[dstY*vLumFilterSize + i + 1]<<16 : 0);
-                if (CONFIG_SWSCALE_ALPHA && alpPixBuf) {
-                    *(const void**)&alpMmxFilter[s*i              ]= alpSrcPtr[i  ];
-                    *(const void**)&alpMmxFilter[s*i+APCK_PTR2/4  ]= alpSrcPtr[i+(vLumFilterSize>1)];
-                    alpMmxFilter[s*i+APCK_COEF/4  ]=
-                    alpMmxFilter[s*i+APCK_COEF/4+1]= lumMmxFilter[s*i+APCK_COEF/4  ];
-                }
-            }
-            for (i=0; i<vChrFilterSize; i+=2) {
-                *(const void**)&chrMmxFilter[s*i              ]= chrUSrcPtr[i  ];
-                *(const void**)&chrMmxFilter[s*i+APCK_PTR2/4  ]= chrUSrcPtr[i+(vChrFilterSize>1)];
-                chrMmxFilter[s*i+APCK_COEF/4  ]=
-                chrMmxFilter[s*i+APCK_COEF/4+1]= vChrFilter[chrDstY*vChrFilterSize + i    ]
-                + (vChrFilterSize>1 ? vChrFilter[chrDstY*vChrFilterSize + i + 1]<<16 : 0);
-            }
-        } else {
-            for (i=0; i<vLumFilterSize; i++) {
-                *(const void**)&lumMmxFilter[4*i+0]= lumSrcPtr[i];
-                lumMmxFilter[4*i+2]=
-                lumMmxFilter[4*i+3]=
-                ((uint16_t)vLumFilter[dstY*vLumFilterSize + i])*0x10001;
-                if (CONFIG_SWSCALE_ALPHA && alpPixBuf) {
-                    *(const void**)&alpMmxFilter[4*i+0]= alpSrcPtr[i];
-                    alpMmxFilter[4*i+2]=
-                    alpMmxFilter[4*i+3]= lumMmxFilter[4*i+2];
-                }
-            }
-            for (i=0; i<vChrFilterSize; i++) {
-                *(const void**)&chrMmxFilter[4*i+0]= chrUSrcPtr[i];
-                chrMmxFilter[4*i+2]=
-                chrMmxFilter[4*i+3]=
-                ((uint16_t)vChrFilter[chrDstY*vChrFilterSize + i])*0x10001;
-            }
-        }
-    }
-}
-
-#endif /* HAVE_INLINE_ASM */
-
-#define SCALE_FUNC(filter_n, from_bpc, to_bpc, opt) \
-extern void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt( \
-                                                SwsContext *c, int16_t *data, \
-                                                int dstW, const uint8_t *src, \
-                                                const int16_t *filter, \
-                                                const int32_t *filterPos, int filterSize)
-
-#define SCALE_FUNCS(filter_n, opt) \
-    SCALE_FUNC(filter_n,  8, 15, opt); \
-    SCALE_FUNC(filter_n,  9, 15, opt); \
-    SCALE_FUNC(filter_n, 10, 15, opt); \
-    SCALE_FUNC(filter_n, 16, 15, opt); \
-    SCALE_FUNC(filter_n,  8, 19, opt); \
-    SCALE_FUNC(filter_n,  9, 19, opt); \
-    SCALE_FUNC(filter_n, 10, 19, opt); \
-    SCALE_FUNC(filter_n, 16, 19, opt)
-
-#define SCALE_FUNCS_MMX(opt) \
-    SCALE_FUNCS(4, opt); \
-    SCALE_FUNCS(8, opt); \
-    SCALE_FUNCS(X, opt)
-
-#define SCALE_FUNCS_SSE(opt) \
-    SCALE_FUNCS(4, opt); \
-    SCALE_FUNCS(8, opt); \
-    SCALE_FUNCS(X4, opt); \
-    SCALE_FUNCS(X8, opt)
-
-#if ARCH_X86_32
-SCALE_FUNCS_MMX(mmx);
-#endif
-SCALE_FUNCS_SSE(sse2);
-SCALE_FUNCS_SSE(ssse3);
-SCALE_FUNCS_SSE(sse4);
-
-#define VSCALEX_FUNC(size, opt) \
-extern void ff_yuv2planeX_ ## size ## _ ## opt(const int16_t *filter, int filterSize, \
-                                               const int16_t **src, uint8_t *dest, int dstW, \
-                                               const uint8_t *dither, int offset)
-#define VSCALEX_FUNCS(opt) \
-    VSCALEX_FUNC(8,  opt); \
-    VSCALEX_FUNC(9,  opt); \
-    VSCALEX_FUNC(10, opt)
-
-#if ARCH_X86_32
-VSCALEX_FUNCS(mmx2);
-#endif
-VSCALEX_FUNCS(sse2);
-VSCALEX_FUNCS(sse4);
-VSCALEX_FUNC(16, sse4);
-VSCALEX_FUNCS(avx);
-
-#define VSCALE_FUNC(size, opt) \
-extern void ff_yuv2plane1_ ## size ## _ ## opt(const int16_t *src, uint8_t *dst, int dstW, \
-                                               const uint8_t *dither, int offset)
-#define VSCALE_FUNCS(opt1, opt2) \
-    VSCALE_FUNC(8,  opt1); \
-    VSCALE_FUNC(9,  opt2); \
-    VSCALE_FUNC(10, opt2); \
-    VSCALE_FUNC(16, opt1)
-
-#if ARCH_X86_32
-VSCALE_FUNCS(mmx, mmx2);
-#endif
-VSCALE_FUNCS(sse2, sse2);
-VSCALE_FUNC(16, sse4);
-VSCALE_FUNCS(avx, avx);
-
-#define INPUT_Y_FUNC(fmt, opt) \
-extern void ff_ ## fmt ## ToY_  ## opt(uint8_t *dst, const uint8_t *src, \
-                                       int w, uint32_t *unused)
-#define INPUT_UV_FUNC(fmt, opt) \
-extern void ff_ ## fmt ## ToUV_ ## opt(uint8_t *dstU, uint8_t *dstV, \
-                                       const uint8_t *src, const uint8_t *unused1, \
-                                       int w, uint32_t *unused2)
-#define INPUT_FUNC(fmt, opt) \
-    INPUT_Y_FUNC(fmt, opt); \
-    INPUT_UV_FUNC(fmt, opt)
-#define INPUT_FUNCS(opt) \
-    INPUT_FUNC(uyvy, opt); \
-    INPUT_FUNC(yuyv, opt); \
-    INPUT_UV_FUNC(nv12, opt); \
-    INPUT_UV_FUNC(nv21, opt); \
-    INPUT_FUNC(rgba, opt); \
-    INPUT_FUNC(bgra, opt); \
-    INPUT_FUNC(argb, opt); \
-    INPUT_FUNC(abgr, opt); \
-    INPUT_FUNC(rgb24, opt); \
-    INPUT_FUNC(bgr24, opt)
-
-#if ARCH_X86_32
-INPUT_FUNCS(mmx);
-#endif
-INPUT_FUNCS(sse2);
-INPUT_FUNCS(ssse3);
-INPUT_FUNCS(avx);
-
-void ff_sws_init_swScale_mmx(SwsContext *c)
-{
-    int cpu_flags = av_get_cpu_flags();
-
-#if HAVE_INLINE_ASM
-    if (cpu_flags & AV_CPU_FLAG_MMX)
-        sws_init_swScale_MMX(c);
-#if HAVE_MMX2
-    if (cpu_flags & AV_CPU_FLAG_MMX2)
-        sws_init_swScale_MMX2(c);
-#endif
-#endif /* HAVE_INLINE_ASM */
-
-#if HAVE_YASM
-#define ASSIGN_SCALE_FUNC2(hscalefn, filtersize, opt1, opt2) do { \
-    if (c->srcBpc == 8) { \
-        hscalefn = c->dstBpc <= 10 ? ff_hscale8to15_ ## filtersize ## _ ## opt2 : \
-                                     ff_hscale8to19_ ## filtersize ## _ ## opt1; \
-    } else if (c->srcBpc == 9) { \
-        hscalefn = c->dstBpc <= 10 ? ff_hscale9to15_ ## filtersize ## _ ## opt2 : \
-                                     ff_hscale9to19_ ## filtersize ## _ ## opt1; \
-    } else if (c->srcBpc == 10) { \
-        hscalefn = c->dstBpc <= 10 ? ff_hscale10to15_ ## filtersize ## _ ## opt2 : \
-                                     ff_hscale10to19_ ## filtersize ## _ ## opt1; \
-    } else /* c->srcBpc == 16 */ { \
-        hscalefn = c->dstBpc <= 10 ? ff_hscale16to15_ ## filtersize ## _ ## opt2 : \
-                                     ff_hscale16to19_ ## filtersize ## _ ## opt1; \
-    } \
-} while (0)
-#define ASSIGN_MMX_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
-    switch (filtersize) { \
-    case 4:  ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
-    case 8:  ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
-    default: ASSIGN_SCALE_FUNC2(hscalefn, X, opt1, opt2); break; \
-    }
-#define ASSIGN_VSCALEX_FUNC(vscalefn, opt, do_16_case, condition_8bit) \
-switch(c->dstBpc){ \
-    case 16:                          do_16_case;                          break; \
-    case 10: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_10_ ## opt; break; \
-    case 9:  if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_9_  ## opt; break; \
-    default: if (condition_8bit)      vscalefn = ff_yuv2planeX_8_  ## opt; break; \
-    }
-#define ASSIGN_VSCALE_FUNC(vscalefn, opt1, opt2, opt2chk) \
-    switch(c->dstBpc){ \
-    case 16: if (!isBE(c->dstFormat))            vscalefn = ff_yuv2plane1_16_ ## opt1; break; \
-    case 10: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_10_ ## opt2; break; \
-    case 9:  if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_9_  ## opt2;  break; \
-    default:                                     vscalefn = ff_yuv2plane1_8_  ## opt1;  break; \
-    }
-#define case_rgb(x, X, opt) \
-        case PIX_FMT_ ## X: \
-            c->lumToYV12 = ff_ ## x ## ToY_ ## opt; \
-            if (!c->chrSrcHSubSample) \
-                c->chrToYV12 = ff_ ## x ## ToUV_ ## opt; \
-            break
-#if ARCH_X86_32
-    if (cpu_flags & AV_CPU_FLAG_MMX) {
-        ASSIGN_MMX_SCALE_FUNC(c->hyScale, c->hLumFilterSize, mmx, mmx);
-        ASSIGN_MMX_SCALE_FUNC(c->hcScale, c->hChrFilterSize, mmx, mmx);
-        ASSIGN_VSCALE_FUNC(c->yuv2plane1, mmx, mmx2, cpu_flags & AV_CPU_FLAG_MMX2);
-
-        switch (c->srcFormat) {
-        case PIX_FMT_Y400A:
-            c->lumToYV12 = ff_yuyvToY_mmx;
-            if (c->alpPixBuf)
-                c->alpToYV12 = ff_uyvyToY_mmx;
-            break;
-        case PIX_FMT_YUYV422:
-            c->lumToYV12 = ff_yuyvToY_mmx;
-            c->chrToYV12 = ff_yuyvToUV_mmx;
-            break;
-        case PIX_FMT_UYVY422:
-            c->lumToYV12 = ff_uyvyToY_mmx;
-            c->chrToYV12 = ff_uyvyToUV_mmx;
-            break;
-        case PIX_FMT_NV12:
-            c->chrToYV12 = ff_nv12ToUV_mmx;
-            break;
-        case PIX_FMT_NV21:
-            c->chrToYV12 = ff_nv21ToUV_mmx;
-            break;
-        case_rgb(rgb24, RGB24, mmx);
-        case_rgb(bgr24, BGR24, mmx);
-        case_rgb(bgra,  BGRA,  mmx);
-        case_rgb(rgba,  RGBA,  mmx);
-        case_rgb(abgr,  ABGR,  mmx);
-        case_rgb(argb,  ARGB,  mmx);
-        default:
-            break;
-        }
-    }
-    if (cpu_flags & AV_CPU_FLAG_MMX2) {
-        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, mmx2, , 1);
-    }
-#endif
-#define ASSIGN_SSE_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
-    switch (filtersize) { \
-    case 4:  ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
-    case 8:  ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
-    default: if (filtersize & 4) ASSIGN_SCALE_FUNC2(hscalefn, X4, opt1, opt2); \
-             else                ASSIGN_SCALE_FUNC2(hscalefn, X8, opt1, opt2); \
-             break; \
-    }
-    if (cpu_flags & AV_CPU_FLAG_SSE2) {
-        ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse2, sse2);
-        ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse2, sse2);
-        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse2, ,
-                            HAVE_ALIGNED_STACK || ARCH_X86_64);
-        ASSIGN_VSCALE_FUNC(c->yuv2plane1, sse2, sse2, 1);
-
-        switch (c->srcFormat) {
-        case PIX_FMT_Y400A:
-            c->lumToYV12 = ff_yuyvToY_sse2;
-            if (c->alpPixBuf)
-                c->alpToYV12 = ff_uyvyToY_sse2;
-            break;
-        case PIX_FMT_YUYV422:
-            c->lumToYV12 = ff_yuyvToY_sse2;
-            c->chrToYV12 = ff_yuyvToUV_sse2;
-            break;
-        case PIX_FMT_UYVY422:
-            c->lumToYV12 = ff_uyvyToY_sse2;
-            c->chrToYV12 = ff_uyvyToUV_sse2;
-            break;
-        case PIX_FMT_NV12:
-            c->chrToYV12 = ff_nv12ToUV_sse2;
-            break;
-        case PIX_FMT_NV21:
-            c->chrToYV12 = ff_nv21ToUV_sse2;
-            break;
-        case_rgb(rgb24, RGB24, sse2);
-        case_rgb(bgr24, BGR24, sse2);
-        case_rgb(bgra,  BGRA,  sse2);
-        case_rgb(rgba,  RGBA,  sse2);
-        case_rgb(abgr,  ABGR,  sse2);
-        case_rgb(argb,  ARGB,  sse2);
-        default:
-            break;
-        }
-    }
-    if (cpu_flags & AV_CPU_FLAG_SSSE3) {
-        ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, ssse3, ssse3);
-        ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, ssse3, ssse3);
-        switch (c->srcFormat) {
-        case_rgb(rgb24, RGB24, ssse3);
-        case_rgb(bgr24, BGR24, ssse3);
-        default:
-            break;
-        }
-    }
-    if (cpu_flags & AV_CPU_FLAG_SSE4) {
-        /* Xto15 don't need special sse4 functions */
-        ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse4, ssse3);
-        ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse4, ssse3);
-        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse4,
-                            if (!isBE(c->dstFormat)) c->yuv2planeX = ff_yuv2planeX_16_sse4,
-                            HAVE_ALIGNED_STACK || ARCH_X86_64);
-        if (c->dstBpc == 16 && !isBE(c->dstFormat))
-            c->yuv2plane1 = ff_yuv2plane1_16_sse4;
-    }
-
-    if (cpu_flags & AV_CPU_FLAG_AVX) {
-        ASSIGN_VSCALEX_FUNC(c->yuv2planeX, avx, ,
-                            HAVE_ALIGNED_STACK || ARCH_X86_64);
-        ASSIGN_VSCALE_FUNC(c->yuv2plane1, avx, avx, 1);
-
-        switch (c->srcFormat) {
-        case PIX_FMT_YUYV422:
-            c->chrToYV12 = ff_yuyvToUV_avx;
-            break;
-        case PIX_FMT_UYVY422:
-            c->chrToYV12 = ff_uyvyToUV_avx;
-            break;
-        case PIX_FMT_NV12:
-            c->chrToYV12 = ff_nv12ToUV_avx;
-            break;
-        case PIX_FMT_NV21:
-            c->chrToYV12 = ff_nv21ToUV_avx;
-            break;
-        case_rgb(rgb24, RGB24, avx);
-        case_rgb(bgr24, BGR24, avx);
-        case_rgb(bgra,  BGRA,  avx);
-        case_rgb(rgba,  RGBA,  avx);
-        case_rgb(abgr,  ABGR,  avx);
-        case_rgb(argb,  ARGB,  avx);
-        default:
-            break;
-        }
-    }
-#endif
-}
diff --git a/libswscale/x86/yuv2rgb.c b/libswscale/x86/yuv2rgb.c
new file mode 100644 (file)
index 0000000..f534e0e
--- /dev/null
@@ -0,0 +1,114 @@
+/*
+ * software YUV to RGB converter
+ *
+ * Copyright (C) 2009 Konstantin Shishkov
+ *
+ * MMX/MMX2 template stuff (needed for fast movntq support),
+ * 1,4,8bpp support and context / deglobalize stuff
+ * by Michael Niedermayer (michaelni@gmx.at)
+ *
+ * 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 <stdio.h>
+#include <stdlib.h>
+#include <inttypes.h>
+#include <assert.h>
+
+#include "config.h"
+#include "libswscale/rgb2rgb.h"
+#include "libswscale/swscale.h"
+#include "libswscale/swscale_internal.h"
+#include "libavutil/x86_cpu.h"
+#include "libavutil/cpu.h"
+
+#if HAVE_INLINE_ASM
+
+#define DITHER1XBPP // only for MMX
+
+/* hope these constant values are cache line aligned */
+DECLARE_ASM_CONST(8, uint64_t, mmx_00ffw)   = 0x00ff00ff00ff00ffULL;
+DECLARE_ASM_CONST(8, uint64_t, mmx_redmask) = 0xf8f8f8f8f8f8f8f8ULL;
+DECLARE_ASM_CONST(8, uint64_t, mmx_grnmask) = 0xfcfcfcfcfcfcfcfcULL;
+DECLARE_ASM_CONST(8, uint64_t, pb_e0) = 0xe0e0e0e0e0e0e0e0ULL;
+DECLARE_ASM_CONST(8, uint64_t, pb_03) = 0x0303030303030303ULL;
+DECLARE_ASM_CONST(8, uint64_t, pb_07) = 0x0707070707070707ULL;
+
+//MMX versions
+#if HAVE_MMX
+#undef RENAME
+#undef COMPILE_TEMPLATE_MMX2
+#define COMPILE_TEMPLATE_MMX2 0
+#define RENAME(a) a ## _MMX
+#include "yuv2rgb_template.c"
+#endif /* HAVE_MMX */
+
+//MMX2 versions
+#if HAVE_MMX2
+#undef RENAME
+#undef COMPILE_TEMPLATE_MMX2
+#define COMPILE_TEMPLATE_MMX2 1
+#define RENAME(a) a ## _MMX2
+#include "yuv2rgb_template.c"
+#endif /* HAVE_MMX2 */
+
+#endif /* HAVE_INLINE_ASM */
+
+SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c)
+{
+#if HAVE_INLINE_ASM
+    int cpu_flags = av_get_cpu_flags();
+
+    if (c->srcFormat != PIX_FMT_YUV420P &&
+        c->srcFormat != PIX_FMT_YUVA420P)
+        return NULL;
+
+#if HAVE_MMX2
+    if (cpu_flags & AV_CPU_FLAG_MMX2) {
+        switch (c->dstFormat) {
+        case PIX_FMT_RGB24:  return yuv420_rgb24_MMX2;
+        case PIX_FMT_BGR24:  return yuv420_bgr24_MMX2;
+        }
+    }
+#endif
+
+    if (cpu_flags & AV_CPU_FLAG_MMX) {
+        switch (c->dstFormat) {
+            case PIX_FMT_RGB32:
+                if (c->srcFormat == PIX_FMT_YUVA420P) {
+#if HAVE_7REGS && CONFIG_SWSCALE_ALPHA
+                    return yuva420_rgb32_MMX;
+#endif
+                    break;
+                } else return yuv420_rgb32_MMX;
+            case PIX_FMT_BGR32:
+                if (c->srcFormat == PIX_FMT_YUVA420P) {
+#if HAVE_7REGS && CONFIG_SWSCALE_ALPHA
+                    return yuva420_bgr32_MMX;
+#endif
+                    break;
+                } else return yuv420_bgr32_MMX;
+            case PIX_FMT_RGB24:  return yuv420_rgb24_MMX;
+            case PIX_FMT_BGR24:  return yuv420_bgr24_MMX;
+            case PIX_FMT_RGB565: return yuv420_rgb16_MMX;
+            case PIX_FMT_RGB555: return yuv420_rgb15_MMX;
+        }
+    }
+#endif /* HAVE_INLINE_ASM */
+
+    return NULL;
+}
diff --git a/libswscale/x86/yuv2rgb_mmx.c b/libswscale/x86/yuv2rgb_mmx.c
deleted file mode 100644 (file)
index f534e0e..0000000
+++ /dev/null
@@ -1,114 +0,0 @@
-/*
- * software YUV to RGB converter
- *
- * Copyright (C) 2009 Konstantin Shishkov
- *
- * MMX/MMX2 template stuff (needed for fast movntq support),
- * 1,4,8bpp support and context / deglobalize stuff
- * by Michael Niedermayer (michaelni@gmx.at)
- *
- * 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 <stdio.h>
-#include <stdlib.h>
-#include <inttypes.h>
-#include <assert.h>
-
-#include "config.h"
-#include "libswscale/rgb2rgb.h"
-#include "libswscale/swscale.h"
-#include "libswscale/swscale_internal.h"
-#include "libavutil/x86_cpu.h"
-#include "libavutil/cpu.h"
-
-#if HAVE_INLINE_ASM
-
-#define DITHER1XBPP // only for MMX
-
-/* hope these constant values are cache line aligned */
-DECLARE_ASM_CONST(8, uint64_t, mmx_00ffw)   = 0x00ff00ff00ff00ffULL;
-DECLARE_ASM_CONST(8, uint64_t, mmx_redmask) = 0xf8f8f8f8f8f8f8f8ULL;
-DECLARE_ASM_CONST(8, uint64_t, mmx_grnmask) = 0xfcfcfcfcfcfcfcfcULL;
-DECLARE_ASM_CONST(8, uint64_t, pb_e0) = 0xe0e0e0e0e0e0e0e0ULL;
-DECLARE_ASM_CONST(8, uint64_t, pb_03) = 0x0303030303030303ULL;
-DECLARE_ASM_CONST(8, uint64_t, pb_07) = 0x0707070707070707ULL;
-
-//MMX versions
-#if HAVE_MMX
-#undef RENAME
-#undef COMPILE_TEMPLATE_MMX2
-#define COMPILE_TEMPLATE_MMX2 0
-#define RENAME(a) a ## _MMX
-#include "yuv2rgb_template.c"
-#endif /* HAVE_MMX */
-
-//MMX2 versions
-#if HAVE_MMX2
-#undef RENAME
-#undef COMPILE_TEMPLATE_MMX2
-#define COMPILE_TEMPLATE_MMX2 1
-#define RENAME(a) a ## _MMX2
-#include "yuv2rgb_template.c"
-#endif /* HAVE_MMX2 */
-
-#endif /* HAVE_INLINE_ASM */
-
-SwsFunc ff_yuv2rgb_init_mmx(SwsContext *c)
-{
-#if HAVE_INLINE_ASM
-    int cpu_flags = av_get_cpu_flags();
-
-    if (c->srcFormat != PIX_FMT_YUV420P &&
-        c->srcFormat != PIX_FMT_YUVA420P)
-        return NULL;
-
-#if HAVE_MMX2
-    if (cpu_flags & AV_CPU_FLAG_MMX2) {
-        switch (c->dstFormat) {
-        case PIX_FMT_RGB24:  return yuv420_rgb24_MMX2;
-        case PIX_FMT_BGR24:  return yuv420_bgr24_MMX2;
-        }
-    }
-#endif
-
-    if (cpu_flags & AV_CPU_FLAG_MMX) {
-        switch (c->dstFormat) {
-            case PIX_FMT_RGB32:
-                if (c->srcFormat == PIX_FMT_YUVA420P) {
-#if HAVE_7REGS && CONFIG_SWSCALE_ALPHA
-                    return yuva420_rgb32_MMX;
-#endif
-                    break;
-                } else return yuv420_rgb32_MMX;
-            case PIX_FMT_BGR32:
-                if (c->srcFormat == PIX_FMT_YUVA420P) {
-#if HAVE_7REGS && CONFIG_SWSCALE_ALPHA
-                    return yuva420_bgr32_MMX;
-#endif
-                    break;
-                } else return yuv420_bgr32_MMX;
-            case PIX_FMT_RGB24:  return yuv420_rgb24_MMX;
-            case PIX_FMT_BGR24:  return yuv420_bgr24_MMX;
-            case PIX_FMT_RGB565: return yuv420_rgb16_MMX;
-            case PIX_FMT_RGB555: return yuv420_rgb15_MMX;
-        }
-    }
-#endif /* HAVE_INLINE_ASM */
-
-    return NULL;
-}