@@ -4428,9 +4428,6 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)

         c->vp3_v_loop_filter= ff_vp3_v_loop_filter_c;

         c->vp3_idct_dc_add= ff_vp3_idct_dc_add_c;

     }

-    if (CONFIG_VP6_DECODER) {

-        c->vp6_filter_diag4= ff_vp6_filter_diag4_c;

-    }

     c->h261_loop_filter= h261_loop_filter_c;

@@ -91,10 +91,6 @@ void ff_vp3_idct_dc_add_c(uint8_t *dest/*align 8*/, int line_size, const DCTELEM

 void ff_vp3_v_loop_filter_c(uint8_t *src, int stride, int *bounding_values);

 void ff_vp3_h_loop_filter_c(uint8_t *src, int stride, int *bounding_values);

-/* VP6 DSP functions */

-void ff_vp6_filter_diag4_c(uint8_t *dst, uint8_t *src, int stride,

-                           const int16_t *h_weights, const int16_t *v_weights);

-

 /* Bink functions */

 void ff_bink_idct_c    (DCTELEM *block);

 void ff_bink_idct_add_c(uint8_t *dest, int linesize, DCTELEM *block);

@@ -369,9 +365,6 @@ typedef struct DSPContext {

     void (*vp3_v_loop_filter)(uint8_t *src, int stride, int *bounding_values);

     void (*vp3_h_loop_filter)(uint8_t *src, int stride, int *bounding_values);

-    void (*vp6_filter_diag4)(uint8_t *dst, uint8_t *src, int stride,

-                             const int16_t *h_weights,const int16_t *v_weights);

-

     /* assume len is a multiple of 4, and arrays are 16-byte aligned */

     void (*vorbis_inverse_coupling)(float *mag, float *ang, int blocksize);

     void (*ac3_downmix)(float (*samples)[256], float (*matrix)[2], int out_ch, int in_ch, int len);

@@ -82,7 +82,12 @@ void ff_vp56dsp_init(VP56DSPContext *s, enum CodecID codec)

     } else {

         s->edge_filter_hor = vp6_edge_filter_hor;

         s->edge_filter_ver = vp6_edge_filter_ver;

+

+        if (CONFIG_VP6_DECODER) {

+            s->vp6_filter_diag4= ff_vp6_filter_diag4_c;

+        }

     }

     if (ARCH_ARM) ff_vp56dsp_init_arm(s, codec);

+    if (HAVE_MMX) ff_vp56dsp_init_x86(s, codec);

 }

@@ -22,13 +22,21 @@

 #define AVCODEC_VP56DSP_H

 #include <stdint.h>

+#include "avcodec.h"

 typedef struct VP56DSPContext {

     void (*edge_filter_hor)(uint8_t *yuv, int stride, int t);

     void (*edge_filter_ver)(uint8_t *yuv, int stride, int t);

+

+    void (*vp6_filter_diag4)(uint8_t *dst, uint8_t *src, int stride,

+                             const int16_t *h_weights,const int16_t *v_weights);

 } VP56DSPContext;

+void ff_vp6_filter_diag4_c(uint8_t *dst, uint8_t *src, int stride,

+                           const int16_t *h_weights, const int16_t *v_weights);

+

 void ff_vp56dsp_init(VP56DSPContext *s, enum CodecID codec);

 void ff_vp56dsp_init_arm(VP56DSPContext *s, enum CodecID codec);

+void ff_vp56dsp_init_x86(VP56DSPContext* c, enum CodecID codec);

 #endif /* AVCODEC_VP56DSP_H */

@@ -559,7 +559,7 @@ static void vp6_filter(VP56Context *s, uint8_t *dst, uint8_t *src,

             vp6_filter_hv4(dst, src+offset1, stride, stride,

                            vp6_block_copy_filter[select][y8]);

         } else {

-            s->dsp.vp6_filter_diag4(dst, src+offset1+((mv.x^mv.y)>>31), stride,

+            s->vp56dsp.vp6_filter_diag4(dst, src+offset1+((mv.x^mv.y)>>31), stride,

                              vp6_block_copy_filter[select][x8],

                              vp6_block_copy_filter[select][y8]);

         }

@@ -22,7 +22,7 @@

  */

 #include "libavutil/common.h"

-#include "dsputil.h"

+#include "vp56dsp.h"

 void ff_vp6_filter_diag4_c(uint8_t *dst, uint8_t *src, int stride,

@@ -29,11 +29,13 @@ MMX-OBJS-$(CONFIG_VC1_DECODER)         += x86/vc1dsp_mmx.o

 MMX-OBJS-$(CONFIG_VP3_DECODER)         += x86/vp3dsp_mmx.o              \

                                           x86/vp3dsp_sse2.o

 MMX-OBJS-$(CONFIG_VP5_DECODER)         += x86/vp3dsp_mmx.o              \

-                                          x86/vp3dsp_sse2.o

+                                          x86/vp3dsp_sse2.o             \

+                                          x86/vp56dsp_init.o

 MMX-OBJS-$(CONFIG_VP6_DECODER)         += x86/vp3dsp_mmx.o              \

                                           x86/vp3dsp_sse2.o             \

                                           x86/vp6dsp_mmx.o              \

-                                          x86/vp6dsp_sse2.o

+                                          x86/vp6dsp_sse2.o             \

+                                          x86/vp56dsp_init.o

 YASM-OBJS-$(CONFIG_VP8_DECODER)        += x86/vp8dsp.o

 MMX-OBJS-$(CONFIG_VP8_DECODER)         += x86/vp8dsp-init.o

 MMX-OBJS-$(HAVE_YASM)                  += x86/dsputil_yasm.o            \

@@ -30,8 +30,6 @@

 #include "dsputil_mmx.h"

 #include "vp3dsp_mmx.h"

 #include "vp3dsp_sse2.h"

-#include "vp6dsp_mmx.h"

-#include "vp6dsp_sse2.h"

 #include "idct_xvid.h"

 //#undef NDEBUG

@@ -2626,10 +2624,6 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)

         c->put_rv40_chroma_pixels_tab[0]= put_rv40_chroma_mc8_mmx;

         c->put_rv40_chroma_pixels_tab[1]= put_rv40_chroma_mc4_mmx;

-        if (CONFIG_VP6_DECODER) {

-            c->vp6_filter_diag4 = ff_vp6_filter_diag4_mmx;

-        }

-

         if (mm_flags & FF_MM_MMX2) {

             c->prefetch = prefetch_mmx2;

@@ -2812,10 +2806,6 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)

             H264_QPEL_FUNCS(3, 1, sse2);

             H264_QPEL_FUNCS(3, 2, sse2);

             H264_QPEL_FUNCS(3, 3, sse2);

-

-            if (CONFIG_VP6_DECODER) {

-                c->vp6_filter_diag4 = ff_vp6_filter_diag4_sse2;

-            }

         }

 #if HAVE_SSSE3

         if(mm_flags & FF_MM_SSSE3){

new file mode 100644

@@ -0,0 +1,42 @@

+/*

+ * VP6 MMX/SSE2 optimizations

+ * Copyright (C) 2009  Sebastien Lucas <sebastien.lucas@gmail.com>

+ * Copyright (C) 2009  Zuxy Meng <zuxy.meng@gmail.com>

+ *

+ * This file is part of FFmpeg.

+ *

+ * FFmpeg 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.

+ *

+ * FFmpeg 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 FFmpeg; if not, write to the Free Software

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

+ */

+

+#include "libavutil/x86_cpu.h"

+#include "libavcodec/dsputil.h"

+#include "libavcodec/vp56dsp.h"

+#include "vp6dsp_mmx.h"

+#include "vp6dsp_sse2.h"

+

+av_cold void ff_vp56dsp_init_x86(VP56DSPContext* c, enum CodecID codec)

+{

+    int mm_flags = mm_support();

+

+    if (CONFIG_VP6_DECODER && codec == CODEC_ID_VP6) {

+        if (mm_flags & FF_MM_MMX) {

+            c->vp6_filter_diag4 = ff_vp6_filter_diag4_mmx;

+        }

+

+        if (mm_flags & FF_MM_SSE2) {

+            c->vp6_filter_diag4 = ff_vp6_filter_diag4_sse2;

+        }

+    }

+}