Merge commit 'ac4b32df'

* commit 'ac4b32df':
  On2 VP7 decoder

Conflicts:
	Changelog
	libavcodec/arm/h264pred_init_arm.c
	libavcodec/arm/vp8dsp.h
	libavcodec/arm/vp8dsp_init_arm.c
	libavcodec/arm/vp8dsp_init_armv6.c
	libavcodec/arm/vp8dsp_init_neon.c
	libavcodec/avcodec.h
	libavcodec/h264pred.c
	libavcodec/version.h
	libavcodec/vp8.c
	libavcodec/vp8.h
	libavcodec/vp8data.h
	libavcodec/vp8dsp.c
	libavcodec/vp8dsp.h
	libavcodec/x86/h264_intrapred_init.c
	libavcodec/x86/vp8dsp_init.c

See: 89f2f5db and others
Merged-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/arm/Makefile

@@ -32,6 +32,7 @@ OBJS-$(CONFIG_MLP_DECODER)             += arm/mlpdsp_init_arm.o
 OBJS-$(CONFIG_VC1_DECODER)             += arm/vc1dsp_init_arm.o
 OBJS-$(CONFIG_VORBIS_DECODER)          += arm/vorbisdsp_init_arm.o
 OBJS-$(CONFIG_VP6_DECODER)             += arm/vp6dsp_init_arm.o
+OBJS-$(CONFIG_VP7_DECODER)             += arm/vp8dsp_init_arm.o
 OBJS-$(CONFIG_VP8_DECODER)             += arm/vp8dsp_init_arm.o
 OBJS-$(CONFIG_RV30_DECODER)            += arm/rv34dsp_init_arm.o
 OBJS-$(CONFIG_RV40_DECODER)            += arm/rv34dsp_init_arm.o        \
@@ -56,6 +57,9 @@ ARMV6-OBJS-$(CONFIG_HPELDSP)           += arm/hpeldsp_init_armv6.o      \
 ARMV6-OBJS-$(CONFIG_MPEGAUDIODSP)      += arm/mpegaudiodsp_fixed_armv6.o
 
 ARMV6-OBJS-$(CONFIG_MLP_DECODER)       += arm/mlpdsp_armv6.o
+ARMV6-OBJS-$(CONFIG_VP7_DECODER)       += arm/vp8_armv6.o               \
+                                          arm/vp8dsp_init_armv6.o       \
+                                          arm/vp8dsp_armv6.o
 ARMV6-OBJS-$(CONFIG_VP8_DECODER)       += arm/vp8_armv6.o               \
                                           arm/vp8dsp_init_armv6.o       \
                                           arm/vp8dsp_armv6.o
@@ -103,5 +107,7 @@ NEON-OBJS-$(CONFIG_VC1_DECODER)        += arm/vc1dsp_init_neon.o        \
                                           arm/vc1dsp_neon.o
 NEON-OBJS-$(CONFIG_VORBIS_DECODER)     += arm/vorbisdsp_neon.o
 NEON-OBJS-$(CONFIG_VP6_DECODER)        += arm/vp6dsp_neon.o
+NEON-OBJS-$(CONFIG_VP7_DECODER)        += arm/vp8dsp_init_neon.o        \
+                                          arm/vp8dsp_neon.o
 NEON-OBJS-$(CONFIG_VP8_DECODER)        += arm/vp8dsp_init_neon.o        \
                                           arm/vp8dsp_neon.o

libavcodec/arm/h264pred_init_arm.c

@@ -60,7 +60,8 @@ static av_cold void h264_pred_init_neon(H264PredContext *h, int codec_id,
     if (codec_id != AV_CODEC_ID_VP7 && codec_id != AV_CODEC_ID_VP8)
         h->pred8x8[PLANE_PRED8x8] = ff_pred8x8_plane_neon;
     h->pred8x8[DC_128_PRED8x8   ] = ff_pred8x8_128_dc_neon;
-    if (codec_id != AV_CODEC_ID_RV40 && codec_id != AV_CODEC_ID_VP7 && codec_id != AV_CODEC_ID_VP8) {
+    if (codec_id != AV_CODEC_ID_RV40 && codec_id != AV_CODEC_ID_VP7 &&
+        codec_id != AV_CODEC_ID_VP8) {
         h->pred8x8[DC_PRED8x8     ] = ff_pred8x8_dc_neon;
         h->pred8x8[LEFT_DC_PRED8x8] = ff_pred8x8_left_dc_neon;
         h->pred8x8[TOP_DC_PRED8x8 ] = ff_pred8x8_top_dc_neon;
@@ -77,7 +78,8 @@ static av_cold void h264_pred_init_neon(H264PredContext *h, int codec_id,
     h->pred16x16[LEFT_DC_PRED8x8] = ff_pred16x16_left_dc_neon;
     h->pred16x16[TOP_DC_PRED8x8 ] = ff_pred16x16_top_dc_neon;
     h->pred16x16[DC_128_PRED8x8 ] = ff_pred16x16_128_dc_neon;
-    if (codec_id != AV_CODEC_ID_SVQ3 && codec_id != AV_CODEC_ID_RV40 && codec_id != AV_CODEC_ID_VP7 && codec_id != AV_CODEC_ID_VP8)
+    if (codec_id != AV_CODEC_ID_SVQ3 && codec_id != AV_CODEC_ID_RV40 &&
+        codec_id != AV_CODEC_ID_VP7 && codec_id != AV_CODEC_ID_VP8)
         h->pred16x16[PLANE_PRED8x8  ] = ff_pred16x16_plane_neon;
 #endif // HAVE_NEON
 }

libavcodec/arm/vp8dsp.h

@@ -21,8 +21,10 @@
 
 #include "libavcodec/vp8dsp.h"
 
-void ff_vp8dsp_init_armv6(VP8DSPContext *dsp, int vp7);
-void ff_vp8dsp_init_neon(VP8DSPContext *dsp, int vp7);
+void ff_vp78dsp_init_armv6(VP8DSPContext *dsp);
+void ff_vp78dsp_init_neon(VP8DSPContext *dsp);
+void ff_vp8dsp_init_armv6(VP8DSPContext *dsp);
+void ff_vp8dsp_init_neon(VP8DSPContext *dsp);
 
 #define VP8_LF_Y(hv, inner, opt)                                             \
     void ff_vp8_##hv##_loop_filter16##inner##_##opt(uint8_t *dst,            \

libavcodec/arm/vp8dsp_init_arm.c

@@ -23,12 +23,22 @@
 #include "libavcodec/vp8dsp.h"
 #include "vp8dsp.h"
 
-av_cold void ff_vp8dsp_init_arm(VP8DSPContext *dsp, int vp7)
+av_cold void ff_vp78dsp_init_arm(VP8DSPContext *dsp)
 {
     int cpu_flags = av_get_cpu_flags();
 
     if (have_armv6(cpu_flags))
-        ff_vp8dsp_init_armv6(dsp, vp7);
+        ff_vp78dsp_init_armv6(dsp);
     if (have_neon(cpu_flags))
-        ff_vp8dsp_init_neon(dsp, vp7);
+        ff_vp78dsp_init_neon(dsp);
+}
+
+av_cold void ff_vp8dsp_init_arm(VP8DSPContext *dsp)
+{
+    int cpu_flags = av_get_cpu_flags();
+
+    if (have_armv6(cpu_flags))
+        ff_vp8dsp_init_armv6(dsp);
+    if (have_neon(cpu_flags))
+        ff_vp8dsp_init_neon(dsp);
 }

libavcodec/arm/vp8dsp_init_armv6.c

@@ -40,31 +40,8 @@ VP8_BILIN(16, armv6);
 VP8_BILIN(8,  armv6);
 VP8_BILIN(4,  armv6);
 
-av_cold void ff_vp8dsp_init_armv6(VP8DSPContext *dsp, int vp7)
+av_cold void ff_vp78dsp_init_armv6(VP8DSPContext *dsp)
 {
-    if (!vp7) {
-        dsp->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_armv6;
-        dsp->vp8_luma_dc_wht_dc = ff_vp8_luma_dc_wht_dc_armv6;
-
-        dsp->vp8_idct_add       = ff_vp8_idct_add_armv6;
-        dsp->vp8_idct_dc_add    = ff_vp8_idct_dc_add_armv6;
-        dsp->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_armv6;
-        dsp->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_armv6;
-
-        dsp->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16_armv6;
-        dsp->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16_armv6;
-        dsp->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_armv6;
-        dsp->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_armv6;
-
-        dsp->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16_inner_armv6;
-        dsp->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16_inner_armv6;
-        dsp->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_armv6;
-        dsp->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_armv6;
-
-        dsp->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter16_simple_armv6;
-        dsp->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter16_simple_armv6;
-    }
-
     dsp->put_vp8_epel_pixels_tab[0][0][0] = ff_put_vp8_pixels16_armv6;
     dsp->put_vp8_epel_pixels_tab[0][0][2] = ff_put_vp8_epel16_h6_armv6;
     dsp->put_vp8_epel_pixels_tab[0][2][0] = ff_put_vp8_epel16_v6_armv6;
@@ -120,3 +97,27 @@ av_cold void ff_vp8dsp_init_armv6(VP8DSPContext *dsp, int vp7)
     dsp->put_vp8_bilinear_pixels_tab[2][2][1] = ff_put_vp8_bilin4_hv_armv6;
     dsp->put_vp8_bilinear_pixels_tab[2][2][2] = ff_put_vp8_bilin4_hv_armv6;
 }
+
+av_cold void ff_vp8dsp_init_armv6(VP8DSPContext *dsp)
+{
+    dsp->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_armv6;
+    dsp->vp8_luma_dc_wht_dc = ff_vp8_luma_dc_wht_dc_armv6;
+
+    dsp->vp8_idct_add       = ff_vp8_idct_add_armv6;
+    dsp->vp8_idct_dc_add    = ff_vp8_idct_dc_add_armv6;
+    dsp->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_armv6;
+    dsp->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_armv6;
+
+    dsp->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16_armv6;
+    dsp->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16_armv6;
+    dsp->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_armv6;
+    dsp->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_armv6;
+
+    dsp->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16_inner_armv6;
+    dsp->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16_inner_armv6;
+    dsp->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_armv6;
+    dsp->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_armv6;
+
+    dsp->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter16_simple_armv6;
+    dsp->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter16_simple_armv6;
+}

libavcodec/arm/vp8dsp_init_neon.c

@@ -39,30 +39,8 @@ VP8_BILIN(16, neon);
 VP8_BILIN(8,  neon);
 VP8_BILIN(4,  neon);
 
-av_cold void ff_vp8dsp_init_neon(VP8DSPContext *dsp, int vp7)
+av_cold void ff_vp78dsp_init_neon(VP8DSPContext *dsp)
 {
-    if (!vp7) {
-        dsp->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_neon;
-
-        dsp->vp8_idct_add       = ff_vp8_idct_add_neon;
-        dsp->vp8_idct_dc_add    = ff_vp8_idct_dc_add_neon;
-        dsp->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_neon;
-        dsp->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_neon;
-
-        dsp->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16_neon;
-        dsp->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16_neon;
-        dsp->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_neon;
-        dsp->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_neon;
-
-        dsp->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16_inner_neon;
-        dsp->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16_inner_neon;
-        dsp->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_neon;
-        dsp->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_neon;
-
-        dsp->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter16_simple_neon;
-        dsp->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter16_simple_neon;
-    }
-
     dsp->put_vp8_epel_pixels_tab[0][0][0] = ff_put_vp8_pixels16_neon;
     dsp->put_vp8_epel_pixels_tab[0][0][2] = ff_put_vp8_epel16_h6_neon;
     dsp->put_vp8_epel_pixels_tab[0][2][0] = ff_put_vp8_epel16_v6_neon;
@@ -116,3 +94,26 @@ av_cold void ff_vp8dsp_init_neon(VP8DSPContext *dsp, int vp7)
     dsp->put_vp8_bilinear_pixels_tab[2][2][1] = ff_put_vp8_bilin4_hv_neon;
     dsp->put_vp8_bilinear_pixels_tab[2][2][2] = ff_put_vp8_bilin4_hv_neon;
 }
+
+av_cold void ff_vp8dsp_init_neon(VP8DSPContext *dsp)
+{
+    dsp->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_neon;
+
+    dsp->vp8_idct_add       = ff_vp8_idct_add_neon;
+    dsp->vp8_idct_dc_add    = ff_vp8_idct_dc_add_neon;
+    dsp->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_neon;
+    dsp->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_neon;
+
+    dsp->vp8_v_loop_filter16y = ff_vp8_v_loop_filter16_neon;
+    dsp->vp8_h_loop_filter16y = ff_vp8_h_loop_filter16_neon;
+    dsp->vp8_v_loop_filter8uv = ff_vp8_v_loop_filter8uv_neon;
+    dsp->vp8_h_loop_filter8uv = ff_vp8_h_loop_filter8uv_neon;
+
+    dsp->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16_inner_neon;
+    dsp->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16_inner_neon;
+    dsp->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_neon;
+    dsp->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_neon;
+
+    dsp->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter16_simple_neon;
+    dsp->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter16_simple_neon;
+}

libavcodec/avcodec.h

@@ -290,6 +290,7 @@ enum AVCodecID {
     AV_CODEC_ID_BRENDER_PIX_DEPRECATED,
     AV_CODEC_ID_PAF_VIDEO_DEPRECATED,
     AV_CODEC_ID_EXR_DEPRECATED,
+    AV_CODEC_ID_VP7_DEPRECATED,
 
     AV_CODEC_ID_BRENDER_PIX= MKBETAG('B','P','I','X'),
     AV_CODEC_ID_Y41P       = MKBETAG('Y','4','1','P'),

libavcodec/h264pred.c

@@ -420,7 +420,7 @@ av_cold void ff_h264_pred_init(H264PredContext *h, int codec_id,
 
 #define H264_PRED(depth) \
     if(codec_id != AV_CODEC_ID_RV40){\
-        if(codec_id == AV_CODEC_ID_VP7 || codec_id == AV_CODEC_ID_VP8) {\
+        if (codec_id == AV_CODEC_ID_VP7 || codec_id == AV_CODEC_ID_VP8) {\
             h->pred4x4[VERT_PRED       ]= FUNCD(pred4x4_vertical_vp8);\
             h->pred4x4[HOR_PRED        ]= FUNCD(pred4x4_horizontal_vp8);\
         } else {\
@@ -440,7 +440,7 @@ av_cold void ff_h264_pred_init(H264PredContext *h, int codec_id,
         } else\
             h->pred4x4[VERT_LEFT_PRED  ]= FUNCC(pred4x4_vertical_left     , depth);\
         h->pred4x4[HOR_UP_PRED         ]= FUNCC(pred4x4_horizontal_up     , depth);\
-        if(codec_id != AV_CODEC_ID_VP7 && codec_id != AV_CODEC_ID_VP8) {\
+        if (codec_id != AV_CODEC_ID_VP7 && codec_id != AV_CODEC_ID_VP8) {\
             h->pred4x4[LEFT_DC_PRED    ]= FUNCC(pred4x4_left_dc           , depth);\
             h->pred4x4[TOP_DC_PRED     ]= FUNCC(pred4x4_top_dc            , depth);\
         } else {\
@@ -498,7 +498,8 @@ av_cold void ff_h264_pred_init(H264PredContext *h, int codec_id,
         }\
     } else\
         h->pred8x8[PLANE_PRED8x8]= FUNCD(pred8x8_tm_vp8);\
-    if(codec_id != AV_CODEC_ID_RV40 && codec_id != AV_CODEC_ID_VP7 && codec_id != AV_CODEC_ID_VP8){\
+    if (codec_id != AV_CODEC_ID_RV40 && codec_id != AV_CODEC_ID_VP7 && \
+        codec_id != AV_CODEC_ID_VP8) {\
         if (chroma_format_idc <= 1) {\
             h->pred8x8[DC_PRED8x8     ]= FUNCC(pred8x8_dc                     , depth);\
             h->pred8x8[LEFT_DC_PRED8x8]= FUNCC(pred8x8_left_dc                , depth);\

libavcodec/ppc/Makefile

@@ -13,6 +13,7 @@ OBJS-$(CONFIG_VP3DSP)                  += ppc/vp3dsp_altivec.o
 
 OBJS-$(CONFIG_VC1_DECODER)             += ppc/vc1dsp_altivec.o
 OBJS-$(CONFIG_VORBIS_DECODER)          += ppc/vorbisdsp_altivec.o
+OBJS-$(CONFIG_VP7_DECODER)             += ppc/vp8dsp_altivec.o
 OBJS-$(CONFIG_VP8_DECODER)             += ppc/vp8dsp_altivec.o
 
 ALTIVEC-OBJS-$(CONFIG_DSPUTIL)         += ppc/dsputil_altivec.o         \

libavcodec/ppc/vp8dsp_altivec.c

@@ -311,7 +311,8 @@ static void put_vp8_pixels16_altivec(uint8_t *dst, ptrdiff_t dstride, uint8_t *s
 
 #endif /* HAVE_ALTIVEC */
 
-av_cold void ff_vp8dsp_init_ppc(VP8DSPContext *c)
+
+av_cold void ff_vp78dsp_init_ppc(VP8DSPContext *c)
 {
 #if HAVE_ALTIVEC
     if (!(av_get_cpu_flags() & AV_CPU_FLAG_ALTIVEC))

libavcodec/utils.c

@@ -2687,6 +2687,7 @@ static enum AVCodecID remap_deprecated_codec_id(enum AVCodecID id)
         case AV_CODEC_ID_PAF_VIDEO_DEPRECATED : return AV_CODEC_ID_PAF_VIDEO;
         case AV_CODEC_ID_WEBP_DEPRECATED: return AV_CODEC_ID_WEBP;
         case AV_CODEC_ID_HEVC_DEPRECATED: return AV_CODEC_ID_HEVC;
+        case AV_CODEC_ID_VP7_DEPRECATED : return AV_CODEC_ID_VP7;
         default                         : return id;
     }
 }

libavcodec/version.h

@@ -30,7 +30,7 @@
 
 #define LIBAVCODEC_VERSION_MAJOR 55
 #define LIBAVCODEC_VERSION_MINOR  56
-#define LIBAVCODEC_VERSION_MICRO 107
+#define LIBAVCODEC_VERSION_MICRO 108
 
 #define LIBAVCODEC_VERSION_INT  AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
                                                LIBAVCODEC_VERSION_MINOR, \

libavcodec/vp8.c

@@ -101,7 +101,7 @@ static int vp8_ref_frame(VP8Context *s, VP8Frame *dst, VP8Frame *src)
 
     return 0;
 }
-#endif
+#endif /* CONFIG_VP8_DECODER */
 
 static void vp8_decode_flush_impl(AVCodecContext *avctx, int free_mem)
 {
@@ -121,16 +121,16 @@ static void vp8_decode_flush(AVCodecContext *avctx)
     vp8_decode_flush_impl(avctx, 0);
 }
 
-static VP8Frame * vp8_find_free_buffer(VP8Context *s)
+static VP8Frame *vp8_find_free_buffer(VP8Context *s)
 {
     VP8Frame *frame = NULL;
     int i;
 
     // find a free buffer
     for (i = 0; i < 5; i++)
-        if (&s->frames[i] != s->framep[VP56_FRAME_CURRENT] &&
+        if (&s->frames[i] != s->framep[VP56_FRAME_CURRENT]  &&
             &s->frames[i] != s->framep[VP56_FRAME_PREVIOUS] &&
-            &s->frames[i] != s->framep[VP56_FRAME_GOLDEN] &&
+            &s->frames[i] != s->framep[VP56_FRAME_GOLDEN]   &&
             &s->frames[i] != s->framep[VP56_FRAME_GOLDEN2]) {
             frame = &s->frames[i];
             break;
@@ -145,7 +145,8 @@ static VP8Frame * vp8_find_free_buffer(VP8Context *s)
     return frame;
 }
 
-static int update_dimensions(VP8Context *s, int width, int height)
+static av_always_inline
+int update_dimensions(VP8Context *s, int width, int height, int is_vp7)
 {
     AVCodecContext *avctx = s->avctx;
     int i, ret;
@@ -162,8 +163,8 @@ static int update_dimensions(VP8Context *s, int width, int height)
     s->mb_width  = (s->avctx->coded_width  + 15) / 16;
     s->mb_height = (s->avctx->coded_height + 15) / 16;
 
-    s->mb_layout = s->vp7 || (avctx->active_thread_type == FF_THREAD_SLICE) &&
-                   (FFMIN(s->num_coeff_partitions, avctx->thread_count) > 1);
+    s->mb_layout = is_vp7 || avctx->active_thread_type == FF_THREAD_SLICE &&
+                   FFMIN(s->num_coeff_partitions, avctx->thread_count) > 1;
     if (!s->mb_layout) { // Frame threading and one thread
         s->macroblocks_base       = av_mallocz((s->mb_width + s->mb_height * 2 + 1) *
                                                sizeof(*s->macroblocks));
@@ -193,6 +194,16 @@ static int update_dimensions(VP8Context *s, int width, int height)
     return 0;
 }
 
+static int vp7_update_dimensions(VP8Context *s, int width, int height)
+{
+    return update_dimensions(s, width, height, IS_VP7);
+}
+
+static int vp8_update_dimensions(VP8Context *s, int width, int height)
+{
+    return update_dimensions(s, width, height, IS_VP8);
+}
+
 
 #if CONFIG_VP8_DECODER
 static void parse_segment_info(VP8Context *s)
@@ -357,8 +368,7 @@ static void vp78_reset_probability_tables(VP8Context *s)
     int i, j;
     for (i = 0; i < 4; i++)
         for (j = 0; j < 16; j++)
-            memcpy(s->prob->token[i][j],
-                   vp8_token_default_probs[i][vp8_coeff_band[j]],
+            memcpy(s->prob->token[i][j], vp8_token_default_probs[i][vp8_coeff_band[j]],
                    sizeof(s->prob->token[i][j]));
 }
 
@@ -370,7 +380,7 @@ static void vp78_update_probability_tables(VP8Context *s)
     for (i = 0; i < 4; i++)
         for (j = 0; j < 8; j++)
             for (k = 0; k < 3; k++)
-                for (l = 0; l < NUM_DCT_TOKENS - 1; l++)
+                for (l = 0; l < NUM_DCT_TOKENS-1; l++)
                     if (vp56_rac_get_prob_branchy(c, vp8_token_update_probs[i][j][k][l])) {
                         int prob = vp8_rac_get_uint(c, 8);
                         for (m = 0; vp8_coeff_band_indexes[j][m] >= 0; m++)
@@ -378,7 +388,11 @@ static void vp78_update_probability_tables(VP8Context *s)
                     }
 }
 
-static void vp78_update_pred16x16_pred8x8_mvc_probabilities(VP8Context *s)
+#define VP7_MVC_SIZE 17
+#define VP8_MVC_SIZE 19
+
+static void vp78_update_pred16x16_pred8x8_mvc_probabilities(VP8Context *s,
+                                                            int mvc_size)
 {
     VP56RangeCoder *c = &s->c;
     int i, j;
@@ -392,7 +406,7 @@ static void vp78_update_pred16x16_pred8x8_mvc_probabilities(VP8Context *s)
 
     // 17.2 MV probability update
     for (i = 0; i < 2; i++)
-        for (j = 0; j < (s->vp7 ? 17 : 19); j++)
+        for (j = 0; j < mvc_size; j++)
             if (vp56_rac_get_prob_branchy(c, vp8_mv_update_prob[i][j]))
                 s->prob->mvc[i][j] = vp8_rac_get_nn(c);
 }
@@ -410,15 +424,68 @@ static void update_refs(VP8Context *s)
 }
 #endif
 
-#if CONFIG_VP7_DECODER
-static void fade(uint8_t *dst, int dst_linesize, const uint8_t *src, int src_linesize, int width, int height, int alpha, int beta)
+static void copy_luma(AVFrame *dst, AVFrame *src, int width, int height)
+{
+    int i, j;
+
+    for (j = 1; j < 3; j++) {
+        for (i = 0; i < height / 2; i++)
+            memcpy(dst->data[j] + i * dst->linesize[j],
+                   src->data[j] + i * src->linesize[j], width / 2);
+    }
+}
+
+static void fade(uint8_t *dst, int dst_linesize,
+                 const uint8_t *src, int src_linesize,
+                 int width, int height,
+                 int alpha, int beta)
 {
     int i, j;
-    for (j = 0; j < height; j++)
+    for (j = 0; j < height; j++) {
         for (i = 0; i < width; i++) {
-            uint8_t y = src[j*src_linesize + i];
-            dst[j*dst_linesize + i] = av_clip_uint8(y + ((y * beta) >> 8) + alpha);
+            uint8_t y = src[j * src_linesize + i];
+            dst[j * dst_linesize + i] = av_clip_uint8(y + ((y * beta) >> 8) + alpha);
         }
+    }
+}
+
+static int vp7_fade_frame(VP8Context *s, VP56RangeCoder *c)
+{
+    int alpha = (int8_t) vp8_rac_get_uint(c, 8);
+    int beta  = (int8_t) vp8_rac_get_uint(c, 8);
+    int ret;
+
+    if (!s->keyframe && (alpha || beta)) {
+        int width  = s->mb_width * 16;
+        int height = s->mb_height * 16;
+        AVFrame *src, *dst;
+
+        if (!s->framep[VP56_FRAME_PREVIOUS] ||
+            !s->framep[VP56_FRAME_GOLDEN]) {
+            av_log(s->avctx, AV_LOG_WARNING, "Discarding interframe without a prior keyframe!\n");
+            return AVERROR_INVALIDDATA;
+        }
+
+        dst =
+        src = s->framep[VP56_FRAME_PREVIOUS]->tf.f;
+
+        /* preserve the golden frame, write a new previous frame */
+        if (s->framep[VP56_FRAME_GOLDEN] == s->framep[VP56_FRAME_PREVIOUS]) {
+            s->framep[VP56_FRAME_PREVIOUS] = vp8_find_free_buffer(s);
+            if ((ret = vp8_alloc_frame(s, s->framep[VP56_FRAME_PREVIOUS], 1)) < 0)
+                return ret;
+
+            dst = s->framep[VP56_FRAME_PREVIOUS]->tf.f;
+
+            copy_luma(dst, src, width, height);
+        }
+
+        fade(dst->data[0], dst->linesize[0],
+             src->data[0], src->linesize[0],
+             width, height, alpha, beta);
+    }
+
+    return 0;
 }
 
 static int vp7_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_size)
@@ -428,7 +495,7 @@ static int vp7_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
     int width  = s->avctx->width;
     int height = s->avctx->height;
 
-    s->profile   =  (buf[0]>>1) & 7;
+    s->profile = (buf[0] >> 1) & 7;
     if (s->profile > 1) {
         avpriv_request_sample(s->avctx, "Unknown profile %d", s->profile);
         return AVERROR_INVALIDDATA;
@@ -463,10 +530,13 @@ static int vp7_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
 
         s->update_golden = s->update_altref = VP56_FRAME_CURRENT;
         vp78_reset_probability_tables(s);
-        memcpy(s->prob->pred16x16, vp8_pred16x16_prob_inter, sizeof(s->prob->pred16x16));
-        memcpy(s->prob->pred8x8c , vp8_pred8x8c_prob_inter , sizeof(s->prob->pred8x8c));
+        memcpy(s->prob->pred16x16, vp8_pred16x16_prob_inter,
+               sizeof(s->prob->pred16x16));
+        memcpy(s->prob->pred8x8c, vp8_pred8x8c_prob_inter,
+               sizeof(s->prob->pred8x8c));
         for (i = 0; i < 2; i++)
-            memcpy(s->prob->mvc[i], vp7_mv_default_prob[i], sizeof(vp7_mv_default_prob[i]));
+            memcpy(s->prob->mvc[i], vp7_mv_default_prob[i],
+                   sizeof(vp7_mv_default_prob[i]));
         memset(&s->segmentation, 0, sizeof(s->segmentation));
         memset(&s->lf_delta, 0, sizeof(s->lf_delta));
         memcpy(s->prob[0].scan, zigzag_scan, sizeof(s->prob[0].scan));
@@ -482,24 +552,27 @@ static int vp7_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
              s->feature_present_prob[i] = vp8_rac_get_uint(c, 8);
 
              for (j = 0; j < 3; j++)
-                 s->feature_index_prob[i][j] = vp8_rac_get(c) ? vp8_rac_get_uint(c, 8) : 255;
+                 s->feature_index_prob[i][j] =
+                     vp8_rac_get(c) ? vp8_rac_get_uint(c, 8) : 255;
 
              if (vp7_feature_value_size[i])
                  for (j = 0; j < 4; j++)
-                     s->feature_value[i][j] = vp8_rac_get(c) ? vp8_rac_get_uint(c, vp7_feature_value_size[s->profile][i]) : 0;
+                     s->feature_value[i][j] =
+                        vp8_rac_get(c) ? vp8_rac_get_uint(c, vp7_feature_value_size[s->profile][i]) : 0;
         }
     }
 
-    s->segmentation.enabled = 0;
+    s->segmentation.enabled    = 0;
     s->segmentation.update_map = 0;
-    s->lf_delta.enabled = 0;
+    s->lf_delta.enabled        = 0;
 
     s->num_coeff_partitions = 1;
     ff_vp56_init_range_decoder(&s->coeff_partition[0], buf, buf_size);
 
     if (!s->macroblocks_base || /* first frame */
-        width != s->avctx->width || height != s->avctx->height || (width+15)/16 != s->mb_width || (height+15)/16 != s->mb_height) {
-        if ((ret = update_dimensions(s, width, height)) < 0)
+        width != s->avctx->width || height != s->avctx->height ||
+        (width + 15) / 16 != s->mb_width || (height + 15) / 16 != s->mb_height) {
+        if ((ret = vp7_update_dimensions(s, width, height)) < 0)
             return ret;
     }
 
@@ -512,7 +585,7 @@ static int vp7_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
         s->sign_bias[VP56_FRAME_GOLDEN] = 0;
     }
 
-    s->update_last = 1;
+    s->update_last          = 1;
     s->update_probabilities = 1;
     s->fade_present         = 1;
 
@@ -527,37 +600,8 @@ static int vp7_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
 
     /* E. Fading information for previous frame */
     if (s->fade_present && vp8_rac_get(c)) {
-        int alpha = (int8_t)vp8_rac_get_uint(c, 8);
-        int beta  = (int8_t)vp8_rac_get_uint(c, 8);
-        if (!s->keyframe && (alpha || beta)) {
-
-            if (!s->framep[VP56_FRAME_PREVIOUS] ||
-                !s->framep[VP56_FRAME_GOLDEN]) {
-                av_log(s->avctx, AV_LOG_WARNING, "Discarding interframe without a prior keyframe!\n");
-                return AVERROR_INVALIDDATA;
-            }
-
-            /* preserve the golden frame */
-            if (s->framep[VP56_FRAME_GOLDEN] == s->framep[VP56_FRAME_PREVIOUS]) {
-                AVFrame *gold = s->framep[VP56_FRAME_GOLDEN]->tf.f;
-                AVFrame *prev;
-                int i, j;
-
-                s->framep[VP56_FRAME_PREVIOUS] = vp8_find_free_buffer(s);
-                if ((ret = vp8_alloc_frame(s, s->framep[VP56_FRAME_PREVIOUS], 1)) < 0)
-                   return ret;
-                prev = s->framep[VP56_FRAME_PREVIOUS]->tf.f;
-
-                fade(prev->data[0], prev->linesize[0], gold->data[0], gold->linesize[0], s->mb_width * 16, s->mb_height * 16, alpha, beta);
-                for (j = 1; j < 3; j++)
-                    for (i = 0; i < s->mb_height * 8; i++)
-                        memcpy(prev->data[j] + i * prev->linesize[j], gold->data[j] + i * gold->linesize[j], s->mb_width * 8);
-            } else {
-                AVFrame *prev = s->framep[VP56_FRAME_PREVIOUS]->tf.f;
-                fade(prev->data[0], prev->linesize[0], prev->data[0], prev->linesize[0], s->mb_width * 16, s->mb_height * 16, alpha, beta);
-            }
-
-        }
+        if ((ret = vp7_fade_frame(s ,c)) < 0)
+            return ret;
     }
 
     /* F. Loop filter type */
@@ -584,14 +628,12 @@ static int vp7_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
     if (!s->keyframe) {
         s->prob->intra  = vp8_rac_get_uint(c, 8);
         s->prob->last   = vp8_rac_get_uint(c, 8);
-        vp78_update_pred16x16_pred8x8_mvc_probabilities(s);
+        vp78_update_pred16x16_pred8x8_mvc_probabilities(s, VP7_MVC_SIZE);
     }
 
     return 0;
 }
-#endif
 
-#if CONFIG_VP8_DECODER
 static int vp8_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_size)
 {
     VP56RangeCoder *c = &s->c;
@@ -680,7 +722,7 @@ static int vp8_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
     if (!s->macroblocks_base || /* first frame */
         width != s->avctx->width || height != s->avctx->height ||
         (width+15)/16 != s->mb_width || (height+15)/16 != s->mb_height)
-        if ((ret = update_dimensions(s, width, height)) < 0)
+        if ((ret = vp8_update_dimensions(s, width, height)) < 0)
             return ret;
 
     vp8_get_quants(s);
@@ -707,12 +749,11 @@ static int vp8_decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_si
         s->prob->intra  = vp8_rac_get_uint(c, 8);
         s->prob->last   = vp8_rac_get_uint(c, 8);
         s->prob->golden = vp8_rac_get_uint(c, 8);
-        vp78_update_pred16x16_pred8x8_mvc_probabilities(s);
+        vp78_update_pred16x16_pred8x8_mvc_probabilities(s, VP8_MVC_SIZE);
     }
 
     return 0;
 }
-#endif
 
 static av_always_inline
 void clamp_mv(VP8Context *s, VP56mv *dst, const VP56mv *src)
@@ -763,8 +804,11 @@ static int vp8_read_mv_component(VP56RangeCoder *c, const uint8_t *p)
 }
 
 static av_always_inline
-const uint8_t *get_submv_prob(uint32_t left, uint32_t top)
+const uint8_t *get_submv_prob(uint32_t left, uint32_t top, int is_vp7)
 {
+    if (is_vp7)
+        return vp7_submv_prob;
+
     if (left == top)
         return vp8_submv_prob[4 - !!left];
     if (!top)
@@ -777,7 +821,8 @@ const uint8_t *get_submv_prob(uint32_t left, uint32_t top)
  * @returns the number of motion vectors parsed (2, 4 or 16)
  */
 static av_always_inline
-int decode_splitmvs(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb, int layout, int vp7)
+int decode_splitmvs(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb,
+                    int layout, int is_vp7)
 {
     int part_idx;
     int n, num;
@@ -824,13 +869,15 @@ int decode_splitmvs(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb, int lay
         else
             above = AV_RN32A(&cur_mv[mbsplits_cur[k - 4]]);
 
-        submv_prob = vp7 ? vp7_submv_prob : get_submv_prob(left, above);
+        submv_prob = get_submv_prob(left, above, is_vp7);
 
         if (vp56_rac_get_prob_branchy(c, submv_prob[0])) {
             if (vp56_rac_get_prob_branchy(c, submv_prob[1])) {
                 if (vp56_rac_get_prob_branchy(c, submv_prob[2])) {
-                    mb->bmv[n].y = mb->mv.y + VPX(vp7, read_mv_component)(c, s->prob->mvc[0]);
-                    mb->bmv[n].x = mb->mv.x + VPX(vp7, read_mv_component)(c, s->prob->mvc[1]);
+                    mb->bmv[n].y = mb->mv.y +
+                                   read_mv_component(c, s->prob->mvc[0], is_vp7);
+                    mb->bmv[n].x = mb->mv.x +
+                                   read_mv_component(c, s->prob->mvc[1], is_vp7);
                 } else {
                     AV_ZERO32(&mb->bmv[n]);
                 }
@@ -846,15 +893,19 @@ int decode_splitmvs(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb, int lay
 }
 
 /**
- * the vp7 reference decoder uses a padding macroblock column (added to right
- * edge of the frame) to guard against illegal macroblock offsets. The algorithm
- * has bugs that permit offsets to straddle the padding column. This function
- * replicates those bugs.
+ * The vp7 reference decoder uses a padding macroblock column (added to right
+ * edge of the frame) to guard against illegal macroblock offsets. The
+ * algorithm has bugs that permit offsets to straddle the padding column.
+ * This function replicates those bugs.
+ *
  * @param[out] edge_x macroblock x address
  * @param[out] edge_y macroblock y address
+ *
  * @return macroblock offset legal (boolean)
  */
-static int vp7_calculate_mb_offset(int mb_x, int mb_y, int mb_width, int xoffset, int yoffset, int boundary, int *edge_x, int *edge_y)
+static int vp7_calculate_mb_offset(int mb_x, int mb_y, int mb_width,
+                                   int xoffset, int yoffset, int boundary,
+                                   int *edge_x, int *edge_y)
 {
     int vwidth = mb_width + 1;
     int new = (mb_y + yoffset) * vwidth + mb_x + xoffset;
@@ -865,13 +916,14 @@ static int vp7_calculate_mb_offset(int mb_x, int mb_y, int mb_width, int xoffset
     return 1;
 }
 
-static const VP56mv * get_bmv_ptr(const VP8Macroblock *mb, int subblock)
+static const VP56mv *get_bmv_ptr(const VP8Macroblock *mb, int subblock)
 {
     return &mb->bmv[mb->mode == VP8_MVMODE_SPLIT ? vp8_mbsplits[mb->partitioning][subblock] : 0];
 }
 
 static av_always_inline
-void vp7_decode_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y, int layout)
+void vp7_decode_mvs(VP8Context *s, VP8Macroblock *mb,
+                    int mb_x, int mb_y, int layout)
 {
     VP8Macroblock *mb_edge[12];
     enum { CNT_ZERO, CNT_NEAREST, CNT_NEAR };
@@ -890,10 +942,13 @@ void vp7_decode_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y, int la
         const VP7MVPred * pred = &vp7_mv_pred[i];
         int edge_x, edge_y;
 
-        if (vp7_calculate_mb_offset(mb_x, mb_y, s->mb_width, pred->xoffset, pred->yoffset, !s->profile, &edge_x, &edge_y)) {
-            VP8Macroblock *edge = mb_edge[i] = (s->mb_layout == 1) ?
-                                                   s->macroblocks_base + (s->mb_width+1)*(edge_y + 1) + 1 + edge_x :
-                                                   s->macroblocks + (s->mb_height - edge_y - 1)*2 + edge_x;
+        if (vp7_calculate_mb_offset(mb_x, mb_y, s->mb_width, pred->xoffset,
+                                    pred->yoffset, !s->profile, &edge_x, &edge_y)) {
+            VP8Macroblock *edge = mb_edge[i] = (s->mb_layout == 1)
+                                             ? s->macroblocks_base + 1 + edge_x +
+                                               (s->mb_width + 1) * (edge_y + 1)
+                                             : s->macroblocks + edge_x +
+                                               (s->mb_height - edge_y - 1) * 2;
             uint32_t mv = AV_RN32A(get_bmv_ptr(edge, vp7_mv_pred[i].subblock));
             if (mv) {
                 if (AV_RN32A(&near_mv[CNT_NEAREST])) {
@@ -936,7 +991,7 @@ void vp7_decode_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y, int la
 
                 if (vp56_rac_get_prob_branchy(c, vp7_mode_contexts[cnt[CNT_NEAR]][3])) {
                     mb->mode = VP8_MVMODE_SPLIT;
-                    mb->mv = mb->bmv[decode_splitmvs(s, c, mb, layout, 1) - 1];
+                    mb->mv = mb->bmv[decode_splitmvs(s, c, mb, layout, IS_VP7) - 1];
                 } else {
                     mb->mv.y += vp7_read_mv_component(c, s->prob->mvc[0]);
                     mb->mv.x += vp7_read_mv_component(c, s->prob->mvc[1]);
@@ -1036,7 +1091,7 @@ void vp8_decode_mvs(VP8Context *s, VP8Macroblock *mb,
 
                 if (vp56_rac_get_prob_branchy(c, vp8_mode_contexts[cnt[CNT_SPLITMV]][3])) {
                     mb->mode = VP8_MVMODE_SPLIT;
-                    mb->mv   = mb->bmv[decode_splitmvs(s, c, mb, layout, 0) - 1];
+                    mb->mv = mb->bmv[decode_splitmvs(s, c, mb, layout, IS_VP8) - 1];
                 } else {
                     mb->mv.y  += vp8_read_mv_component(c, s->prob->mvc[0]);
                     mb->mv.x  += vp8_read_mv_component(c, s->prob->mvc[1]);
@@ -1092,32 +1147,34 @@ void decode_intra4x4_modes(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb,
     }
 }
 
-static const char * vp7_feature_name[] = { "q-index", "lf-delta", "partial-golden-update", "blit-pitch" };
-
 static av_always_inline
 void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
-                    uint8_t *segment, uint8_t *ref, int layout, int vp7)
+                    uint8_t *segment, uint8_t *ref, int layout, int is_vp7)
 {
     VP56RangeCoder *c = &s->c;
-
-    if (vp7) {
+    const char *vp7_feature_name[] = { "q-index",
+                                       "lf-delta",
+                                       "partial-golden-update",
+                                       "blit-pitch" };
+    if (is_vp7) {
         int i;
         *segment = 0;
         for (i = 0; i < 4; i++) {
             if (s->feature_enabled[i]) {
                 if (vp56_rac_get_prob(c, s->feature_present_prob[i])) {
-                      int index = vp8_rac_get_tree(c, vp7_feature_index_tree, s->feature_index_prob[i]);
-                      av_log(s->avctx, AV_LOG_WARNING, "Feature %s present in macroblock (value 0x%x)\n", vp7_feature_name[i], s->feature_value[i][index]);
+                      int index = vp8_rac_get_tree(c, vp7_feature_index_tree,
+                                                   s->feature_index_prob[i]);
+                      av_log(s->avctx, AV_LOG_WARNING,
+                             "Feature %s present in macroblock (value 0x%x)\n",
+                             vp7_feature_name[i], s->feature_value[i][index]);
                 }
-            }
+           }
         }
-    } else {
-        if (s->segmentation.update_map) {
-            int bit  = vp56_rac_get_prob(c, s->prob->segmentid[0]);
-            *segment = vp56_rac_get_prob(c, s->prob->segmentid[1+bit]) + 2*bit;
-        } else if (s->segmentation.enabled)
-            *segment = ref ? *ref : *segment;
-    }
+    } else if (s->segmentation.update_map) {
+        int bit  = vp56_rac_get_prob(c, s->prob->segmentid[0]);
+        *segment = vp56_rac_get_prob(c, s->prob->segmentid[1+bit]) + 2*bit;
+    } else if (s->segmentation.enabled)
+        *segment = ref ? *ref : *segment;
     mb->segment = *segment;
 
     mb->skip = s->mbskip_enabled ? vp56_rac_get_prob(c, s->prob->mbskip) : 0;
@@ -1129,7 +1186,8 @@ void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
         if (mb->mode == MODE_I4x4) {
             decode_intra4x4_modes(s, c, mb, mb_x, 1, layout);
         } else {
-            const uint32_t modes = VPX(vp7, pred4x4_mode)[mb->mode] * 0x01010101u;
+            const uint32_t modes = (is_vp7 ? vp7_pred4x4_mode
+                                           : vp8_pred4x4_mode)[mb->mode] * 0x01010101u;
             if (s->mb_layout)
                 AV_WN32A(mb->intra4x4_pred_mode_top, modes);
             else
@@ -1144,15 +1202,14 @@ void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
         // inter MB, 16.2
         if (vp56_rac_get_prob_branchy(c, s->prob->last))
             mb->ref_frame =
-                (!vp7 && vp56_rac_get_prob(c, s->prob->golden))
-                                                      ? VP56_FRAME_GOLDEN2 /* altref */
-                                                      : VP56_FRAME_GOLDEN;
+                (!is_vp7 && vp56_rac_get_prob(c, s->prob->golden)) ? VP56_FRAME_GOLDEN2 /* altref */
+                                                                   : VP56_FRAME_GOLDEN;
         else
             mb->ref_frame = VP56_FRAME_PREVIOUS;
         s->ref_count[mb->ref_frame - 1]++;
 
         // motion vectors, 16.3
-        if (vp7)
+        if (is_vp7)
             vp7_decode_mvs(s, mb, mb_x, mb_y, layout);
         else
             vp8_decode_mvs(s, mb, mb_x, mb_y, layout);
@@ -1171,28 +1228,6 @@ void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
     }
 }
 
-static av_always_inline int inter_predict_dc(int16_t block[16], int16_t pred[2])
-{
-    int16_t dc = block[0];
-    int ret = 0;
-
-    if (pred[1] > 3) {
-        dc += pred[0];
-        ret = 1;
-    }
-
-    if (!pred[0] || !dc || FFSIGN(pred[0]) != FFSIGN(dc)) {
-        block[0] = pred[0] = dc;
-        pred[1] = 0;
-    } else {
-        if (pred[0] == dc)
-            pred[1]++;
-        block[0] = pred[0] = dc;
-    }
-
-    return ret;
-}
-
 /**
  * @param r     arithmetic bitstream reader context
  * @param block destination for block coefficients
@@ -1205,10 +1240,9 @@ static av_always_inline int inter_predict_dc(int16_t block[16], int16_t pred[2])
  */
 static av_always_inline
 int decode_block_coeffs_internal(VP56RangeCoder *r, int16_t block[16],
-                                        uint8_t probs[16][3][NUM_DCT_TOKENS - 1],
-                                        int i, uint8_t *token_prob,
-                                        int16_t qmul[2],
-                                        const uint8_t scan[16], int vp7)
+                                 uint8_t probs[16][3][NUM_DCT_TOKENS - 1],
+                                 int i, uint8_t *token_prob, int16_t qmul[2],
+                                 const uint8_t scan[16], int vp7)
 {
     VP56RangeCoder c = *r;
     goto skip_eob;
@@ -1264,20 +1298,49 @@ skip_eob:
     return i;
 }
 
-static int vp7_decode_block_coeffs_internal(VP56RangeCoder *r, int16_t block[16],
-                                        uint8_t probs[16][3][NUM_DCT_TOKENS - 1],
-                                        int i, uint8_t *token_prob, int16_t qmul[2],
-                                        const uint8_t scan[16])
+static av_always_inline
+int inter_predict_dc(int16_t block[16], int16_t pred[2])
 {
-    return decode_block_coeffs_internal(r, block, probs, i, token_prob, qmul, scan, 1);
+    int16_t dc = block[0];
+    int ret = 0;
+
+    if (pred[1] > 3) {
+        dc += pred[0];
+        ret = 1;
+    }
+
+    if (!pred[0] | !dc | ((int32_t)pred[0] ^ (int32_t)dc) >> 31) {
+        block[0] = pred[0] = dc;
+        pred[1] = 0;
+    } else {
+        if (pred[0] == dc)
+            pred[1]++;
+        block[0] = pred[0] = dc;
+    }
+
+    return ret;
+}
+
+static int vp7_decode_block_coeffs_internal(VP56RangeCoder *r,
+                                            int16_t block[16],
+                                            uint8_t probs[16][3][NUM_DCT_TOKENS - 1],
+                                            int i, uint8_t *token_prob,
+                                            int16_t qmul[2],
+                                            const uint8_t scan[16])
+{
+    return decode_block_coeffs_internal(r, block, probs, i,
+                                        token_prob, qmul, scan, IS_VP7);
 }
 
 #ifndef vp8_decode_block_coeffs_internal
-static int vp8_decode_block_coeffs_internal(VP56RangeCoder *r, int16_t block[16],
-                                        uint8_t probs[16][3][NUM_DCT_TOKENS - 1],
-                                        int i, uint8_t *token_prob, int16_t qmul[2])
+static int vp8_decode_block_coeffs_internal(VP56RangeCoder *r,
+                                            int16_t block[16],
+                                            uint8_t probs[16][3][NUM_DCT_TOKENS - 1],
+                                            int i, uint8_t *token_prob,
+                                            int16_t qmul[2])
 {
-    return decode_block_coeffs_internal(r, block, probs, i, token_prob, qmul, zigzag_scan, 0);
+    return decode_block_coeffs_internal(r, block, probs, i,
+                                        token_prob, qmul, zigzag_scan, IS_VP8);
 }
 #endif
 
@@ -1303,31 +1366,35 @@ int decode_block_coeffs(VP56RangeCoder *c, int16_t block[16],
     uint8_t *token_prob = probs[i][zero_nhood];
     if (!vp56_rac_get_prob_branchy(c, token_prob[0]))   // DCT_EOB
         return 0;
-    return vp7 ? vp7_decode_block_coeffs_internal(c, block, probs, i, token_prob, qmul, scan)
-               : vp8_decode_block_coeffs_internal(c, block, probs, i, token_prob, qmul);
+    return vp7 ? vp7_decode_block_coeffs_internal(c, block, probs, i,
+                                                  token_prob, qmul, scan)
+               : vp8_decode_block_coeffs_internal(c, block, probs, i,
+                                                  token_prob, qmul);
 }
 
 static av_always_inline
 void decode_mb_coeffs(VP8Context *s, VP8ThreadData *td, VP56RangeCoder *c,
                       VP8Macroblock *mb, uint8_t t_nnz[9], uint8_t l_nnz[9],
-                      int vp7)
+                      int is_vp7)
 {
     int i, x, y, luma_start = 0, luma_ctx = 3;
     int nnz_pred, nnz, nnz_total = 0;
     int segment = mb->segment;
     int block_dc = 0;
 
-    if (mb->mode != MODE_I4x4 && (vp7 || mb->mode != VP8_MVMODE_SPLIT)) {
+    if (mb->mode != MODE_I4x4 && (is_vp7 || mb->mode != VP8_MVMODE_SPLIT)) {
         nnz_pred = t_nnz[8] + l_nnz[8];
 
         // decode DC values and do hadamard
         nnz = decode_block_coeffs(c, td->block_dc, s->prob->token[1], 0,
                                   nnz_pred, s->qmat[segment].luma_dc_qmul,
-                                  zigzag_scan, vp7);
+                                  zigzag_scan, is_vp7);
         l_nnz[8] = t_nnz[8] = !!nnz;
 
-        if (vp7 && mb->mode > MODE_I4x4)
-            nnz |= inter_predict_dc(td->block_dc, s->inter_dc_pred[mb->ref_frame - 1]);
+        if (is_vp7 && mb->mode > MODE_I4x4) {
+            nnz |=  inter_predict_dc(td->block_dc,
+                                     s->inter_dc_pred[mb->ref_frame - 1]);
+        }
 
         if (nnz) {
             nnz_total += nnz;
@@ -1349,7 +1416,7 @@ void decode_mb_coeffs(VP8Context *s, VP8ThreadData *td, VP56RangeCoder *c,
                                       s->prob->token[luma_ctx],
                                       luma_start, nnz_pred,
                                       s->qmat[segment].luma_qmul,
-                                      s->prob[0].scan, vp7);
+                                      s->prob[0].scan, is_vp7);
             /* nnz+block_dc may be one more than the actual last index,
              * but we don't care */
             td->non_zero_count_cache[y][x] = nnz + block_dc;
@@ -1364,13 +1431,13 @@ void decode_mb_coeffs(VP8Context *s, VP8ThreadData *td, VP56RangeCoder *c,
         for (y = 0; y < 2; y++)
             for (x = 0; x < 2; x++) {
                 nnz_pred = l_nnz[i + 2 * y] + t_nnz[i + 2 * x];
-                nnz      = decode_block_coeffs(c, td->block[i][(y << 1) + x],
-                                               s->prob->token[2],
-                                               0, nnz_pred,
-                                               s->qmat[segment].chroma_qmul, s->prob[0].scan, vp7);
+                nnz = decode_block_coeffs(c, td->block[i][(y << 1) + x],
+                                          s->prob->token[2], 0, nnz_pred,
+                                          s->qmat[segment].chroma_qmul,
+                                          s->prob[0].scan, is_vp7);
                 td->non_zero_count_cache[i][(y << 1) + x] = nnz;
                 t_nnz[i + 2 * x] = l_nnz[i + 2 * y] = !!nnz;
-                nnz_total       += nnz;
+                nnz_total += nnz;
             }
 
     // if there were no coded coeffs despite the macroblock not being marked skip,
@@ -1471,7 +1538,8 @@ int check_tm_pred4x4_mode(int mode, int mb_x, int mb_y, int vp7)
 }
 
 static av_always_inline
-int check_intra_pred4x4_mode_emuedge(int mode, int mb_x, int mb_y, int *copy_buf, int vp7)
+int check_intra_pred4x4_mode_emuedge(int mode, int mb_x, int mb_y,
+                                     int *copy_buf, int vp7)
 {
     switch (mode) {
     case VERT_PRED:
@@ -1507,7 +1575,7 @@ int check_intra_pred4x4_mode_emuedge(int mode, int mb_x, int mb_y, int *copy_buf
 
 static av_always_inline
 void intra_predict(VP8Context *s, VP8ThreadData *td, uint8_t *dst[3],
-                   VP8Macroblock *mb, int mb_x, int mb_y, int vp7)
+                   VP8Macroblock *mb, int mb_x, int mb_y, int is_vp7)
 {
     int x, y, mode, nnz;
     uint32_t tr;
@@ -1520,13 +1588,13 @@ void intra_predict(VP8Context *s, VP8ThreadData *td, uint8_t *dst[3],
                        s->filter.simple, 1);
 
     if (mb->mode < MODE_I4x4) {
-        mode = check_intra_pred8x8_mode_emuedge(mb->mode, mb_x, mb_y, vp7);
+        mode = check_intra_pred8x8_mode_emuedge(mb->mode, mb_x, mb_y, is_vp7);
         s->hpc.pred16x16[mode](dst[0], s->linesize);
     } else {
         uint8_t *ptr = dst[0];
         uint8_t *intra4x4 = mb->intra4x4_pred_mode_mb;
-        const uint8_t lo = vp7 ? 128 : 127;
-        const uint8_t hi = vp7 ? 128 : 129;
+        const uint8_t lo = is_vp7 ? 128 : 127;
+        const uint8_t hi = is_vp7 ? 128 : 129;
         uint8_t tr_top[4] = { lo, lo, lo, lo };
 
         // all blocks on the right edge of the macroblock use bottom edge
@@ -1555,9 +1623,8 @@ void intra_predict(VP8Context *s, VP8ThreadData *td, uint8_t *dst[3],
                 } else if (x == 3)
                     topright = tr_right;
 
-                mode = check_intra_pred4x4_mode_emuedge(intra4x4[x],
-                                                        mb_x + x, mb_y + y,
-                                                        &copy, vp7);
+                mode = check_intra_pred4x4_mode_emuedge(intra4x4[x], mb_x + x,
+                                                        mb_y + y, &copy, is_vp7);
                 if (copy) {
                     dst      = copy_dst + 12;
                     linesize = 8;
@@ -1609,7 +1676,8 @@ void intra_predict(VP8Context *s, VP8ThreadData *td, uint8_t *dst[3],
         }
     }
 
-    mode = check_intra_pred8x8_mode_emuedge(mb->chroma_pred_mode, mb_x, mb_y, vp7);
+    mode = check_intra_pred8x8_mode_emuedge(mb->chroma_pred_mode,
+                                            mb_x, mb_y, is_vp7);
     s->hpc.pred8x8[mode](dst[1], s->uvlinesize);
     s->hpc.pred8x8[mode](dst[2], s->uvlinesize);
 
@@ -1765,7 +1833,9 @@ void vp8_mc_part(VP8Context *s, VP8ThreadData *td, uint8_t *dst[3],
                 s->put_pixels_tab[block_w == 8]);
 
     /* U/V */
-    if (s->profile == 3) { /* this block only applies VP8; it is safe to check only the profile, as VP7 profile <= 1 */
+    if (s->profile == 3) {
+        /* this block only applies VP8; it is safe to check
+         * only the profile, as VP7 profile <= 1 */
         uvmv.x &= ~7;
         uvmv.y &= ~7;
     }
@@ -1956,7 +2026,8 @@ chroma_idct_end:
 }
 
 static av_always_inline
-void filter_level_for_mb(VP8Context *s, VP8Macroblock *mb, VP8FilterStrength *f, int vp7)
+void filter_level_for_mb(VP8Context *s, VP8Macroblock *mb,
+                         VP8FilterStrength *f, int is_vp7)
 {
     int interior_limit, filter_level;
 
@@ -1983,13 +2054,13 @@ void filter_level_for_mb(VP8Context *s, VP8Macroblock *mb, VP8FilterStrength *f,
 
     f->filter_level = filter_level;
     f->inner_limit = interior_limit;
-    f->inner_filter = vp7 || !mb->skip || mb->mode == MODE_I4x4 ||
+    f->inner_filter = is_vp7 || !mb->skip || mb->mode == MODE_I4x4 ||
                       mb->mode == VP8_MVMODE_SPLIT;
 }
 
 static av_always_inline
 void filter_mb(VP8Context *s, uint8_t *dst[3], VP8FilterStrength *f,
-               int mb_x, int mb_y, int vp7)
+               int mb_x, int mb_y, int is_vp7)
 {
     int mbedge_lim, bedge_lim_y, bedge_lim_uv, hev_thresh;
     int filter_level = f->filter_level;
@@ -2011,14 +2082,14 @@ void filter_mb(VP8Context *s, uint8_t *dst[3], VP8FilterStrength *f,
     if (!filter_level)
         return;
 
-    if (vp7) {
-        bedge_lim_y  =     filter_level;
-        bedge_lim_uv = 2 * filter_level;
-        mbedge_lim   =     filter_level + 2;
+    if (is_vp7) {
+        bedge_lim_y  = filter_level;
+        bedge_lim_uv = filter_level * 2;
+        mbedge_lim   = filter_level + 2;
     } else {
         bedge_lim_y  =
-        bedge_lim_uv = 2 * filter_level + inner_limit;
-        mbedge_lim   =     bedge_lim_y + 4;
+        bedge_lim_uv = filter_level * 2 + inner_limit;
+        mbedge_lim   = bedge_lim_y + 4;
     }
 
     hev_thresh = hev_thresh_lut[s->keyframe][filter_level];
@@ -2030,20 +2101,23 @@ void filter_mb(VP8Context *s, uint8_t *dst[3], VP8FilterStrength *f,
                                        mbedge_lim, inner_limit, hev_thresh);
     }
 
-#define H_LOOP_FILTER_16Y_INNER(cond) \
-    if (cond && inner_filter) {\
-        s->vp8dsp.vp8_h_loop_filter16y_inner(dst[0] +  4, linesize, bedge_lim_y,\
-                                             inner_limit, hev_thresh);\
-        s->vp8dsp.vp8_h_loop_filter16y_inner(dst[0] +  8, linesize, bedge_lim_y,\
-                                             inner_limit, hev_thresh);\
-        s->vp8dsp.vp8_h_loop_filter16y_inner(dst[0] + 12, linesize, bedge_lim_y,\
-                                             inner_limit, hev_thresh);\
-        s->vp8dsp.vp8_h_loop_filter8uv_inner(dst[1] +  4, dst[2] + 4,\
-                                             uvlinesize, bedge_lim_uv,\
-                                             inner_limit, hev_thresh);\
+#define H_LOOP_FILTER_16Y_INNER(cond)                                         \
+    if (cond && inner_filter) {                                               \
+        s->vp8dsp.vp8_h_loop_filter16y_inner(dst[0] +  4, linesize,           \
+                                             bedge_lim_y, inner_limit,        \
+                                             hev_thresh);                     \
+        s->vp8dsp.vp8_h_loop_filter16y_inner(dst[0] +  8, linesize,           \
+                                             bedge_lim_y, inner_limit,        \
+                                             hev_thresh);                     \
+        s->vp8dsp.vp8_h_loop_filter16y_inner(dst[0] + 12, linesize,           \
+                                             bedge_lim_y, inner_limit,        \
+                                             hev_thresh);                     \
+        s->vp8dsp.vp8_h_loop_filter8uv_inner(dst[1] +  4, dst[2] + 4,         \
+                                             uvlinesize,  bedge_lim_uv,       \
+                                             inner_limit, hev_thresh);        \
     }
 
-    H_LOOP_FILTER_16Y_INNER(!vp7)
+    H_LOOP_FILTER_16Y_INNER(!is_vp7)
 
     if (mb_y) {
         s->vp8dsp.vp8_v_loop_filter16y(dst[0], linesize,
@@ -2068,7 +2142,7 @@ void filter_mb(VP8Context *s, uint8_t *dst[3], VP8FilterStrength *f,
                                              inner_limit, hev_thresh);
     }
 
-    H_LOOP_FILTER_16Y_INNER(vp7)
+    H_LOOP_FILTER_16Y_INNER(is_vp7)
 }
 
 static av_always_inline
@@ -2106,8 +2180,8 @@ void filter_mb_simple(VP8Context *s, uint8_t *dst, VP8FilterStrength *f,
 
 #define MARGIN (16 << 2)
 static av_always_inline
-void decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *curframe,
-                                   VP8Frame *prev_frame, int vp7)
+void vp78_decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *curframe,
+                                    VP8Frame *prev_frame, int is_vp7)
 {
     VP8Context *s = avctx->priv_data;
     int mb_x, mb_y;
@@ -2129,7 +2203,7 @@ void decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *curframe,
                          DC_PRED * 0x01010101);
             decode_mb_mode(s, mb, mb_x, mb_y, curframe->seg_map->data + mb_xy,
                            prev_frame && prev_frame->seg_map ?
-                           prev_frame->seg_map->data + mb_xy : NULL, 1, vp7);
+                           prev_frame->seg_map->data + mb_xy : NULL, 1, is_vp7);
             s->mv_min.x -= 64;
             s->mv_max.x -= 64;
         }
@@ -2138,21 +2212,17 @@ void decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *curframe,
     }
 }
 
-#if CONFIG_VP7_DECODER
-static void vp7_decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *curframe,
+static void vp7_decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *cur_frame,
                                    VP8Frame *prev_frame)
 {
-    decode_mv_mb_modes(avctx, curframe, prev_frame, 1);
+    vp78_decode_mv_mb_modes(avctx, cur_frame, prev_frame, IS_VP7);
 }
-#endif
 
-#if CONFIG_VP8_DECODER
-static void vp8_decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *curframe,
+static void vp8_decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *cur_frame,
                                    VP8Frame *prev_frame)
 {
-    decode_mv_mb_modes(avctx, curframe, prev_frame, 0);
+    vp78_decode_mv_mb_modes(avctx, cur_frame, prev_frame, IS_VP8);
 }
-#endif
 
 #if HAVE_THREADS
 #define check_thread_pos(td, otd, mb_x_check, mb_y_check)                     \
@@ -2195,7 +2265,7 @@ static void vp8_decode_mv_mb_modes(AVCodecContext *avctx, VP8Frame *curframe,
 #endif
 
 static av_always_inline void decode_mb_row_no_filter(AVCodecContext *avctx, void *tdata,
-                                        int jobnr, int threadnr, int vp7)
+                                        int jobnr, int threadnr, int is_vp7)
 {
     VP8Context *s = avctx->priv_data;
     VP8ThreadData *prev_td, *next_td, *td = &s->thread_data[threadnr];
@@ -2231,7 +2301,7 @@ static av_always_inline void decode_mb_row_no_filter(AVCodecContext *avctx, void
         AV_WN32A(s->intra4x4_pred_mode_left, DC_PRED * 0x01010101);
     }
 
-    if (!vp7 || mb_y == 0)
+    if (!is_vp7 || mb_y == 0)
         memset(td->left_nnz, 0, sizeof(td->left_nnz));
 
     s->mv_min.x = -MARGIN;
@@ -2241,10 +2311,13 @@ static av_always_inline void decode_mb_row_no_filter(AVCodecContext *avctx, void
         // Wait for previous thread to read mb_x+2, and reach mb_y-1.
         if (prev_td != td) {
             if (threadnr != 0) {
-                check_thread_pos(td, prev_td, mb_x + (vp7?2:1), mb_y - (vp7?2:1));
+                check_thread_pos(td, prev_td,
+                                 mb_x + (is_vp7 ? 2 : 1),
+                                 mb_y - (is_vp7 ? 2 : 1));
             } else {
                 check_thread_pos(td, prev_td,
-                                 (s->mb_width + 3) + (mb_x + (vp7?2:1)), mb_y - (vp7?2:1));
+                                 mb_x + (is_vp7 ? 2 : 1) + s->mb_width + 3,
+                                 mb_y - (is_vp7 ? 2 : 1));
             }
         }
 
@@ -2256,15 +2329,15 @@ static av_always_inline void decode_mb_row_no_filter(AVCodecContext *avctx, void
         if (!s->mb_layout)
             decode_mb_mode(s, mb, mb_x, mb_y, curframe->seg_map->data + mb_xy,
                            prev_frame && prev_frame->seg_map ?
-                           prev_frame->seg_map->data + mb_xy : NULL, 0, vp7);
+                           prev_frame->seg_map->data + mb_xy : NULL, 0, is_vp7);
 
         prefetch_motion(s, mb, mb_x, mb_y, mb_xy, VP56_FRAME_PREVIOUS);
 
         if (!mb->skip)
-            decode_mb_coeffs(s, td, c, mb, s->top_nnz[mb_x], td->left_nnz, vp7);
+            decode_mb_coeffs(s, td, c, mb, s->top_nnz[mb_x], td->left_nnz, is_vp7);
 
         if (mb->mode <= MODE_I4x4)
-            intra_predict(s, td, dst, mb, mb_x, mb_y, vp7);
+            intra_predict(s, td, dst, mb, mb_x, mb_y, is_vp7);
         else
             inter_predict(s, td, dst, mb, mb_x, mb_y);
 
@@ -2285,7 +2358,7 @@ static av_always_inline void decode_mb_row_no_filter(AVCodecContext *avctx, void
         }
 
         if (s->deblock_filter)
-            filter_level_for_mb(s, mb, &td->filter_strength[mb_x], vp7);
+            filter_level_for_mb(s, mb, &td->filter_strength[mb_x], is_vp7);
 
         if (s->deblock_filter && num_jobs != 1 && threadnr == num_jobs - 1) {
             if (s->filter.simple)
@@ -2329,7 +2402,7 @@ static void vp8_decode_mb_row_no_filter(AVCodecContext *avctx, void *tdata,
 #endif
 
 static av_always_inline void filter_mb_row(AVCodecContext *avctx, void *tdata,
-                              int jobnr, int threadnr, int vp7)
+                              int jobnr, int threadnr, int is_vp7)
 {
     VP8Context *s = avctx->priv_data;
     VP8ThreadData *td = &s->thread_data[threadnr];
@@ -2378,7 +2451,7 @@ static av_always_inline void filter_mb_row(AVCodecContext *avctx, void *tdata,
         if (s->filter.simple)
             filter_mb_simple(s, dst[0], f, mb_x, mb_y);
         else
-            filter_mb(s, dst, f, mb_x, mb_y, vp7);
+            filter_mb(s, dst, f, mb_x, mb_y, is_vp7);
         dst[0] += 16;
         dst[1] += 8;
         dst[2] += 8;
@@ -2403,8 +2476,9 @@ static void vp8_filter_mb_row(AVCodecContext *avctx, void *tdata,
 }
 #endif
 
-static int vp8_decode_mb_row_sliced(AVCodecContext *avctx, void *tdata,
-                                    int jobnr, int threadnr)
+static av_always_inline
+int vp78_decode_mb_row_sliced(AVCodecContext *avctx, void *tdata, int jobnr,
+                              int threadnr, int is_vp7)
 {
     VP8Context *s = avctx->priv_data;
     VP8ThreadData *td = &s->thread_data[jobnr];
@@ -2432,15 +2506,34 @@ static int vp8_decode_mb_row_sliced(AVCodecContext *avctx, void *tdata,
     return 0;
 }
 
-int ff_vp8_decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
-                        AVPacket *avpkt)
+static int vp7_decode_mb_row_sliced(AVCodecContext *avctx, void *tdata,
+                                    int jobnr, int threadnr)
+{
+    return vp78_decode_mb_row_sliced(avctx, tdata, jobnr, threadnr, IS_VP7);
+}
+
+static int vp8_decode_mb_row_sliced(AVCodecContext *avctx, void *tdata,
+                                    int jobnr, int threadnr)
+{
+    return vp78_decode_mb_row_sliced(avctx, tdata, jobnr, threadnr, IS_VP8);
+}
+
+
+static av_always_inline
+int vp78_decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
+                      AVPacket *avpkt, int is_vp7)
 {
     VP8Context *s = avctx->priv_data;
     int ret, i, referenced, num_jobs;
     enum AVDiscard skip_thresh;
     VP8Frame *av_uninit(curframe), *prev_frame;
 
-    if ((ret = VPX(s->vp7, decode_frame_header)(s, avpkt->data, avpkt->size)) < 0)
+    if (is_vp7)
+        ret = vp7_decode_frame_header(s, avpkt->data, avpkt->size);
+    else
+        ret = vp8_decode_frame_header(s, avpkt->data, avpkt->size);
+
+    if (ret < 0)
         goto err;
 
     prev_frame = s->framep[VP56_FRAME_CURRENT];
@@ -2529,7 +2622,10 @@ int ff_vp8_decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
         if (prev_frame && s->segmentation.enabled &&
             !s->segmentation.update_map)
             ff_thread_await_progress(&prev_frame->tf, 1, 0);
-        VPX(s->vp7, decode_mv_mb_modes)(avctx, curframe, prev_frame);
+        if (is_vp7)
+            vp7_decode_mv_mb_modes(avctx, curframe, prev_frame);
+        else
+            vp8_decode_mv_mb_modes(avctx, curframe, prev_frame);
     }
 
     if (avctx->active_thread_type == FF_THREAD_FRAME)
@@ -2545,8 +2641,12 @@ int ff_vp8_decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
         s->thread_data[i].thread_mb_pos = 0;
         s->thread_data[i].wait_mb_pos   = INT_MAX;
     }
-    avctx->execute2(avctx, vp8_decode_mb_row_sliced,
-                    s->thread_data, NULL, num_jobs);
+    if (is_vp7)
+        avctx->execute2(avctx, vp7_decode_mb_row_sliced, s->thread_data, NULL,
+                        num_jobs);
+    else
+        avctx->execute2(avctx, vp8_decode_mb_row_sliced, s->thread_data, NULL,
+                        num_jobs);
 
     ff_thread_report_progress(&curframe->tf, INT_MAX, 0);
     memcpy(&s->framep[0], &s->next_framep[0], sizeof(s->framep[0]) * 4);
@@ -2569,6 +2669,20 @@ err:
     return ret;
 }
 
+int ff_vp8_decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
+                        AVPacket *avpkt)
+{
+    return vp78_decode_frame(avctx, data, got_frame, avpkt, IS_VP8);
+}
+
+#if CONFIG_VP7_DECODER
+static int vp7_decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
+                            AVPacket *avpkt)
+{
+    return vp78_decode_frame(avctx, data, got_frame, avpkt, IS_VP7);
+}
+#endif /* CONFIG_VP7_DECODER */
+
 av_cold int ff_vp8_decode_free(AVCodecContext *avctx)
 {
     VP8Context *s = avctx->priv_data;
@@ -2592,7 +2706,8 @@ static av_cold int vp8_init_frames(VP8Context *s)
     return 0;
 }
 
-av_cold int ff_vp8_decode_init(AVCodecContext *avctx)
+static av_always_inline
+int vp78_decode_init(AVCodecContext *avctx, int is_vp7)
 {
     VP8Context *s = avctx->priv_data;
     int ret;
@@ -2603,11 +2718,22 @@ av_cold int ff_vp8_decode_init(AVCodecContext *avctx)
     avctx->internal->allocate_progress = 1;
 
     ff_videodsp_init(&s->vdsp, 8);
-    ff_h264_pred_init(&s->hpc, s->vp7 ? AV_CODEC_ID_VP7 : AV_CODEC_ID_VP8, 8, 1);
-    ff_vp8dsp_init(&s->vp8dsp, s->vp7);
 
-    s->decode_mb_row_no_filter = VPX(s->vp7, decode_mb_row_no_filter);
-    s->filter_mb_row           = VPX(s->vp7, filter_mb_row);
+    ff_vp78dsp_init(&s->vp8dsp);
+    if (CONFIG_VP7_DECODER && is_vp7) {
+        ff_h264_pred_init(&s->hpc, AV_CODEC_ID_VP7, 8, 1);
+        ff_vp7dsp_init(&s->vp8dsp);
+        s->decode_mb_row_no_filter = vp7_decode_mb_row_no_filter;
+        s->filter_mb_row           = vp7_filter_mb_row;
+    } else if (CONFIG_VP8_DECODER && !is_vp7) {
+        ff_h264_pred_init(&s->hpc, AV_CODEC_ID_VP8, 8, 1);
+        ff_vp8dsp_init(&s->vp8dsp);
+        s->decode_mb_row_no_filter = vp8_decode_mb_row_no_filter;
+        s->filter_mb_row           = vp8_filter_mb_row;
+    }
+
+    /* does not change for VP8 */
+    memcpy(s->prob[0].scan, zigzag_scan, sizeof(s->prob[0].scan));
 
     if ((ret = vp8_init_frames(s)) < 0) {
         ff_vp8_decode_free(avctx);
@@ -2617,6 +2743,18 @@ av_cold int ff_vp8_decode_init(AVCodecContext *avctx)
     return 0;
 }
 
+#if CONFIG_VP7_DECODER
+static int vp7_decode_init(AVCodecContext *avctx)
+{
+    return vp78_decode_init(avctx, IS_VP7);
+}
+#endif /* CONFIG_VP7_DECODER */
+
+av_cold int ff_vp8_decode_init(AVCodecContext *avctx)
+{
+    return vp78_decode_init(avctx, IS_VP8);
+}
+
 #if CONFIG_VP8_DECODER
 static av_cold int vp8_decode_init_thread_copy(AVCodecContext *avctx)
 {
@@ -2668,7 +2806,7 @@ static int vp8_decode_update_thread_context(AVCodecContext *dst,
 
     return 0;
 }
-#endif
+#endif /* CONFIG_VP8_DECODER */
 
 #if CONFIG_VP7_DECODER
 AVCodec ff_vp7_decoder = {
@@ -2677,13 +2815,13 @@ AVCodec ff_vp7_decoder = {
     .type                  = AVMEDIA_TYPE_VIDEO,
     .id                    = AV_CODEC_ID_VP7,
     .priv_data_size        = sizeof(VP8Context),
-    .init                  = ff_vp8_decode_init,
+    .init                  = vp7_decode_init,
     .close                 = ff_vp8_decode_free,
-    .decode                = ff_vp8_decode_frame,
+    .decode                = vp7_decode_frame,
     .capabilities          = CODEC_CAP_DR1,
     .flush                 = vp8_decode_flush,
 };
-#endif
+#endif /* CONFIG_VP7_DECODER */
 
 #if CONFIG_VP8_DECODER
 AVCodec ff_vp8_decoder = {
@@ -2700,5 +2838,4 @@ AVCodec ff_vp8_decoder = {
     .init_thread_copy      = ONLY_IF_THREADS_ENABLED(vp8_decode_init_thread_copy),
     .update_thread_context = ONLY_IF_THREADS_ENABLED(vp8_decode_update_thread_context),
 };
-#endif
-
+#endif /* CONFIG_VP7_DECODER */

libavcodec/vp8data.h

@@ -30,8 +30,7 @@
 #include "vp8.h"
 #include "h264pred.h"
 
-static const uint8_t vp7_pred4x4_mode[] =
-{
+static const uint8_t vp7_pred4x4_mode[] = {
     [DC_PRED8x8]    = DC_PRED,
     [VERT_PRED8x8]  = TM_VP8_PRED,
     [HOR_PRED8x8]   = TM_VP8_PRED,
@@ -59,9 +58,9 @@ static const int8_t vp8_pred16x16_tree_inter[4][2] = {
     { -PLANE_PRED8x8,   -MODE_I4x4 }, // '110', '111'
 };
 
-typedef struct {
-    int8_t  yoffset;
-    int8_t  xoffset;
+typedef struct VP7MVPred {
+    int8_t yoffset;
+    int8_t xoffset;
     uint8_t subblock;
     uint8_t score;
 } VP7MVPred;
@@ -147,7 +146,9 @@ static const uint8_t vp8_mbsplit_prob[3] = {
     110, 111, 150
 };
 
-static const uint8_t vp7_submv_prob[3] = { 180, 162, 25 };
+static const uint8_t vp7_submv_prob[3] = {
+    180, 162, 25
+};
 
 static const uint8_t vp8_submv_prob[5][3] = {
     { 147, 136,  18 },
@@ -161,7 +162,7 @@ static const uint8_t vp8_pred16x16_prob_intra[4] = {
     145, 156, 163, 128
 };
 static const uint8_t vp8_pred16x16_prob_inter[4] = {
-    112, 86, 140, 37
+    112,  86, 140,  37
 };
 
 static const int8_t vp8_pred4x4_tree[9][2] = {
@@ -708,7 +709,7 @@ static const uint8_t vp8_token_update_probs[4][8][3][NUM_DCT_TOKENS - 1] = {
 };
 
 // fixme: copied from h264data.h
-static const uint8_t zigzag_scan[16] = {
+static const uint8_t zigzag_scan[16]={
     0 + 0 * 4, 1 + 0 * 4, 0 + 1 * 4, 0 + 2 * 4,
     1 + 1 * 4, 2 + 0 * 4, 3 + 0 * 4, 2 + 1 * 4,
     1 + 2 * 4, 0 + 3 * 4, 1 + 3 * 4, 2 + 2 * 4,
@@ -771,59 +772,67 @@ static const uint8_t vp8_mv_default_prob[2][19] = {
 };
 
 static const uint8_t vp7_feature_value_size[2][4] = {
-    {7, 6, 0, 8},
-    {7, 6, 0, 5},
+    { 7, 6, 0, 8 },
+    { 7, 6, 0, 5 },
 };
 
 static const int8_t vp7_feature_index_tree[4][2] =
 {
-    { 1, 2 },
-     { -0, -1 }, // '00', '01'
-     { -2, -3 }, // '10', '11'
+    {  1,  2 },
+    { -0, -1 }, // '00', '01'
+    { -2, -3 }, // '10', '11'
 };
 
 static const uint16_t vp7_ydc_qlookup[] = {
-    4, 4, 5, 6, 6, 7, 8, 8, 9, 10, 11, 12, 13, 14, 15, 16,
-    17, 18, 19, 20, 21, 22, 23, 23, 24, 25, 26, 27, 28, 29, 30, 31,
-    32, 33, 33, 34, 35, 36, 36, 37, 38, 39, 39, 40, 41, 41, 42, 43,
-    43, 44, 45, 45, 46, 47, 48, 48, 49, 50, 51, 52, 53, 53, 54, 56,
-    57, 58, 59, 60, 62, 63, 65, 66, 68, 70, 72, 74, 76, 79, 81, 84,
-    87, 90, 93, 96, 100, 104, 108, 112, 116, 121, 126, 131, 136, 142, 148, 154,
-    160, 167, 174, 182, 189, 198, 206, 215, 224, 234, 244, 254, 265, 277, 288, 301,
-    313, 327, 340, 355, 370, 385, 401, 417, 434, 452, 470, 489, 509, 529, 550, 572,
+      4,   4,   5,   6,   6,   7,   8,   8,   9,  10,  11,  12,  13,  14,  15,
+     16,  17,  18,  19,  20,  21,  22,  23,  23,  24,  25,  26,  27,  28,  29,
+     30,  31,  32,  33,  33,  34,  35,  36,  36,  37,  38,  39,  39,  40,  41,
+     41,  42,  43,  43,  44,  45,  45,  46,  47,  48,  48,  49,  50,  51,  52,
+     53,  53,  54,  56,  57,  58,  59,  60,  62,  63,  65,  66,  68,  70,  72,
+     74,  76,  79,  81,  84,  87,  90,  93,  96, 100, 104, 108, 112, 116, 121,
+    126, 131, 136, 142, 148, 154, 160, 167, 174, 182, 189, 198, 206, 215, 224,
+    234, 244, 254, 265, 277, 288, 301, 313, 327, 340, 355, 370, 385, 401, 417,
+    434, 452, 470, 489, 509, 529, 550, 572,
 };
 
 static const uint16_t vp7_yac_qlookup[] = {
-    4, 4, 5, 5, 6, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17,
-    19, 20, 22, 23, 25, 26, 28, 29, 31, 32, 34, 35, 37, 38, 40, 41,
-    42, 44, 45, 46, 48, 49, 50, 51, 53, 54, 55, 56, 57, 58, 59, 61,
-    62, 63, 64, 65, 67, 68, 69, 70, 72, 73, 75, 76, 78, 80, 82, 84,
-    86, 88, 91, 93, 96, 99, 102, 105, 109, 112, 116, 121, 125, 130, 135, 140,
-    146, 152, 158, 165, 172, 180, 188, 196, 205, 214, 224, 234, 245, 256, 268, 281,
-    294, 308, 322, 337, 353, 369, 386, 404, 423, 443, 463, 484, 506, 529, 553, 578,
-    604, 631, 659, 688, 718, 749, 781, 814, 849, 885, 922, 960, 1000, 1041, 1083, 1127,
+       4,    4,   5,   5,   6,   6,   7,   8,   9,  10,  11,  12,   13,   15,
+      16,   17,  19,  20,  22,  23,  25,  26,  28,  29,  31,  32,   34,   35,
+      37,   38,  40,  41,  42,  44,  45,  46,  48,  49,  50,  51,   53,   54,
+      55,   56,  57,  58,  59,  61,  62,  63,  64,  65,  67,  68,   69,   70,
+      72,   73,  75,  76,  78,  80,  82,  84,  86,  88,  91,  93,   96,   99,
+     102,  105, 109, 112, 116, 121, 125, 130, 135, 140, 146, 152,  158,  165,
+     172,  180, 188, 196, 205, 214, 224, 234, 245, 256, 268, 281,  294,  308,
+     322,  337, 353, 369, 386, 404, 423, 443, 463, 484, 506, 529,  553,  578,
+     604,  631, 659, 688, 718, 749, 781, 814, 849, 885, 922, 960, 1000, 1041,
+    1083, 1127,
 };
 
 static const uint16_t vp7_y2dc_qlookup[] = {
-    7, 9, 11, 13, 15, 17, 19, 21, 23, 26, 28, 30, 33, 35, 37, 39,
-    42, 44, 46, 48, 51, 53, 55, 57, 59, 61, 63, 65, 67, 69, 70, 72,
-    74, 75, 77, 78, 80, 81, 83, 84, 85, 87, 88, 89, 90, 92, 93, 94,
-    95, 96, 97, 99, 100, 101, 102, 104, 105, 106, 108, 109, 111, 113, 114, 116,
-    118, 120, 123, 125, 128, 131, 134, 137, 140, 144, 148, 152, 156, 161, 166, 171,
-    176, 182, 188, 195, 202, 209, 217, 225, 234, 243, 253, 263, 274, 285, 297, 309,
-    322, 336, 350, 365, 381, 397, 414, 432, 450, 470, 490, 511, 533, 556, 579, 604,
-    630, 656, 684, 713, 742, 773, 805, 838, 873, 908, 945, 983, 1022, 1063, 1105, 1148,
+       7,    9,  11,  13,  15,  17,  19,  21,  23,  26,  28,  30,   33,   35,
+      37,   39,  42,  44,  46,  48,  51,  53,  55,  57,  59,  61,   63,   65,
+      67,   69,  70,  72,  74,  75,  77,  78,  80,  81,  83,  84,   85,   87,
+      88,   89,  90,  92,  93,  94,  95,  96,  97,  99, 100, 101,  102,  104,
+     105,  106, 108, 109, 111, 113, 114, 116, 118, 120, 123, 125,  128,  131,
+     134,  137, 140, 144, 148, 152, 156, 161, 166, 171, 176, 182,  188,  195,
+     202,  209, 217, 225, 234, 243, 253, 263, 274, 285, 297, 309,  322,  336,
+     350,  365, 381, 397, 414, 432, 450, 470, 490, 511, 533, 556,  579,  604,
+     630,  656, 684, 713, 742, 773, 805, 838, 873, 908, 945, 983, 1022, 1063,
+    1105, 1148,
 };
 
 static const uint16_t vp7_y2ac_qlookup[] = {
-    7, 9, 11, 13, 16, 18, 21, 24, 26, 29, 32, 35, 38, 41, 43, 46,
-    49, 52, 55, 58, 61, 64, 66, 69, 72, 74, 77, 79, 82, 84, 86, 88,
-    91, 93, 95, 97, 98, 100, 102, 104, 105, 107, 109, 110, 112, 113, 115, 116,
-    117, 119, 120, 122, 123, 125, 127, 128, 130, 132, 134, 136, 138, 141, 143, 146,
-    149, 152, 155, 158, 162, 166, 171, 175, 180, 185, 191, 197, 204, 210, 218, 226,
-    234, 243, 252, 262, 273, 284, 295, 308, 321, 335, 350, 365, 381, 398, 416, 435,
-    455, 476, 497, 520, 544, 569, 595, 622, 650, 680, 711, 743, 776, 811, 848, 885,
-    925, 965, 1008, 1052, 1097, 1144, 1193, 1244, 1297, 1351, 1407, 1466, 1526, 1588, 1652, 1719,
+       7,    9,   11,   13,   16,   18,   21,   24,   26,   29,   32,   35,
+      38,   41,   43,   46,   49,   52,   55,   58,   61,   64,   66,   69,
+      72,   74,   77,   79,   82,   84,   86,   88,   91,   93,   95,   97,
+      98,  100,  102,  104,  105,  107,  109,  110,  112,  113,  115,  116,
+     117,  119,  120,  122,  123,  125,  127,  128,  130,  132,  134,  136,
+     138,  141,  143,  146,  149,  152,  155,  158,  162,  166,  171,  175,
+     180,  185,  191,  197,  204,  210,  218,  226,  234,  243,  252,  262,
+     273,  284,  295,  308,  321,  335,  350,  365,  381,  398,  416,  435,
+     455,  476,  497,  520,  544,  569,  595,  622,  650,  680,  711,  743,
+     776,  811,  848,  885,  925,  965, 1008, 1052, 1097, 1144, 1193, 1244,
+    1297, 1351, 1407, 1466, 1526, 1588, 1652, 1719,
 };
 
 #endif /* AVCODEC_VP8DATA_H */

libavcodec/vp8dsp.c

@@ -31,21 +31,23 @@
 #include "mathops.h"
 #include "vp8dsp.h"
 
-#define MK_IDCT_DC_ADD4_C(name) \
-static void name ## _idct_dc_add4uv_c(uint8_t *dst, int16_t block[4][16], ptrdiff_t stride)\
-{\
-    name ## _idct_dc_add_c(dst + stride * 0 + 0, block[0], stride);\
-    name ## _idct_dc_add_c(dst + stride * 0 + 4, block[1], stride);\
-    name ## _idct_dc_add_c(dst + stride * 4 + 0, block[2], stride);\
-    name ## _idct_dc_add_c(dst + stride * 4 + 4, block[3], stride);\
-}\
-\
-static void name ## _idct_dc_add4y_c(uint8_t *dst, int16_t block[4][16], ptrdiff_t stride)\
-{\
-    name ## _idct_dc_add_c(dst + 0, block[0], stride);\
-    name ## _idct_dc_add_c(dst + 4, block[1], stride);\
-    name ## _idct_dc_add_c(dst + 8, block[2], stride);\
-    name ## _idct_dc_add_c(dst + 12, block[3], stride);\
+#define MK_IDCT_DC_ADD4_C(name)                                               \
+static void name ## _idct_dc_add4uv_c(uint8_t *dst, int16_t block[4][16],     \
+                                      ptrdiff_t stride)                       \
+{                                                                             \
+    name ## _idct_dc_add_c(dst + stride * 0 + 0, block[0], stride);           \
+    name ## _idct_dc_add_c(dst + stride * 0 + 4, block[1], stride);           \
+    name ## _idct_dc_add_c(dst + stride * 4 + 0, block[2], stride);           \
+    name ## _idct_dc_add_c(dst + stride * 4 + 4, block[3], stride);           \
+}                                                                             \
+                                                                              \
+static void name ## _idct_dc_add4y_c(uint8_t *dst, int16_t block[4][16],      \
+                                     ptrdiff_t stride)                        \
+{                                                                             \
+    name ## _idct_dc_add_c(dst +  0, block[0], stride);                       \
+    name ## _idct_dc_add_c(dst +  4, block[1], stride);                       \
+    name ## _idct_dc_add_c(dst +  8, block[2], stride);                       \
+    name ## _idct_dc_add_c(dst + 12, block[3], stride);                       \
 }
 
 #if CONFIG_VP7_DECODER
@@ -113,10 +115,14 @@ static void vp7_idct_add_c(uint8_t *dst, int16_t block[16], ptrdiff_t stride)
         b1 = (tmp[i + 0] - tmp[i + 8]) * 23170;
         c1 = tmp[i + 4] * 12540 - tmp[i + 12] * 30274;
         d1 = tmp[i + 4] * 30274 + tmp[i + 12] * 12540;
-        dst[0 * stride + i] = av_clip_uint8(dst[0 * stride + i] + ((a1 + d1 + 0x20000) >> 18));
-        dst[3 * stride + i] = av_clip_uint8(dst[3 * stride + i] + ((a1 - d1 + 0x20000) >> 18));
-        dst[1 * stride + i] = av_clip_uint8(dst[1 * stride + i] + ((b1 + c1 + 0x20000) >> 18));
-        dst[2 * stride + i] = av_clip_uint8(dst[2 * stride + i] + ((b1 - c1 + 0x20000) >> 18));
+        dst[0 * stride + i] = av_clip_uint8(dst[0 * stride + i] +
+                                            ((a1 + d1 + 0x20000) >> 18));
+        dst[3 * stride + i] = av_clip_uint8(dst[3 * stride + i] +
+                                            ((a1 - d1 + 0x20000) >> 18));
+        dst[1 * stride + i] = av_clip_uint8(dst[1 * stride + i] +
+                                            ((b1 + c1 + 0x20000) >> 18));
+        dst[2 * stride + i] = av_clip_uint8(dst[2 * stride + i] +
+                                            ((b1 - c1 + 0x20000) >> 18));
     }
 }
 
@@ -130,12 +136,12 @@ static void vp7_idct_dc_add_c(uint8_t *dst, int16_t block[16], ptrdiff_t stride)
         dst[1] = av_clip_uint8(dst[1] + dc);
         dst[2] = av_clip_uint8(dst[2] + dc);
         dst[3] = av_clip_uint8(dst[3] + dc);
-        dst += stride;
+        dst   += stride;
     }
 }
 
 MK_IDCT_DC_ADD4_C(vp7)
-#endif
+#endif /* CONFIG_VP7_DECODER */
 
 // TODO: Maybe add dequant
 #if CONFIG_VP8_DECODER
@@ -235,7 +241,7 @@ static void vp8_idct_dc_add_c(uint8_t *dst, int16_t block[16], ptrdiff_t stride)
 }
 
 MK_IDCT_DC_ADD4_C(vp8)
-#endif
+#endif /* CONFIG_VP8_DECODER */
 
 // because I like only having two parameters to pass functions...
 #define LOAD_PIXELS                                                           \
@@ -251,7 +257,7 @@ MK_IDCT_DC_ADD4_C(vp8)
 #define clip_int8(n) (cm[n + 0x80] - 0x80)
 
 static av_always_inline void filter_common(uint8_t *p, ptrdiff_t stride,
-                                           int is4tap, int vpn)
+                                           int is4tap, int is_vp7)
 {
     LOAD_PIXELS
     int a, f1, f2;
@@ -268,7 +274,7 @@ static av_always_inline void filter_common(uint8_t *p, ptrdiff_t stride,
     // since that's what libvpx does.
     f1 = FFMIN(a + 4, 127) >> 3;
 
-    if (vpn == 7)
+    if (is_vp7)
         f2 = f1 - ((a & 7) == 4);
     else
         f2 = FFMIN(a + 3, 127) >> 3;
@@ -280,19 +286,33 @@ static av_always_inline void filter_common(uint8_t *p, ptrdiff_t stride,
 
     // only used for _inner on blocks without high edge variance
     if (!is4tap) {
-        a = (f1 + 1) >> 1;
+        a              = (f1 + 1) >> 1;
         p[-2 * stride] = cm[p1 + a];
         p[ 1 * stride] = cm[q1 - a];
     }
 }
 
-static av_always_inline int vp7_simple_limit(uint8_t *p, ptrdiff_t stride, int flim)
+static av_always_inline void vp7_filter_common(uint8_t *p, ptrdiff_t stride,
+                                               int is4tap)
+{
+    filter_common(p, stride, is4tap, IS_VP7);
+}
+
+static av_always_inline void vp8_filter_common(uint8_t *p, ptrdiff_t stride,
+                                               int is4tap)
+{
+    filter_common(p, stride, is4tap, IS_VP8);
+}
+
+static av_always_inline int vp7_simple_limit(uint8_t *p, ptrdiff_t stride,
+                                             int flim)
 {
     LOAD_PIXELS
-    return FFABS(p0-q0) <= flim;
+    return FFABS(p0 - q0) <= flim;
 }
 
-static av_always_inline int vp8_simple_limit(uint8_t *p, ptrdiff_t stride, int flim)
+static av_always_inline int vp8_simple_limit(uint8_t *p, ptrdiff_t stride,
+                                             int flim)
 {
     LOAD_PIXELS
     return 2 * FFABS(p0 - q0) + (FFABS(p1 - q1) >> 1) <= flim;
@@ -302,18 +322,16 @@ static av_always_inline int vp8_simple_limit(uint8_t *p, ptrdiff_t stride, int f
  * E - limit at the macroblock edge
  * I - limit for interior difference
  */
-#define NORMAL_LIMIT(vpn) \
-static av_always_inline int vp ## vpn ## _normal_limit(uint8_t *p, ptrdiff_t stride, \
-                                                       int E, int I)\
-{                                                      \
-    LOAD_PIXELS                                        \
-    return vp ## vpn ## _simple_limit(p, stride, E) && \
-           FFABS(p3 - p2) <= I &&                      \
-           FFABS(p2 - p1) <= I &&                      \
-           FFABS(p1 - p0) <= I &&                      \
-           FFABS(q3 - q2) <= I &&                      \
-           FFABS(q2 - q1) <= I &&                      \
-           FFABS(q1 - q0) <= I;                        \
+#define NORMAL_LIMIT(vpn)                                                     \
+static av_always_inline int vp ## vpn ## _normal_limit(uint8_t *p,            \
+                                                       ptrdiff_t stride,      \
+                                                       int E, int I)          \
+{                                                                             \
+    LOAD_PIXELS                                                               \
+    return vp ## vpn ## _simple_limit(p, stride, E) &&                        \
+           FFABS(p3 - p2) <= I && FFABS(p2 - p1) <= I &&                      \
+           FFABS(p1 - p0) <= I && FFABS(q3 - q2) <= I &&                      \
+           FFABS(q2 - q1) <= I && FFABS(q1 - q0) <= I;                        \
 }
 
 NORMAL_LIMIT(7)
@@ -350,90 +368,89 @@ static av_always_inline void filter_mbedge(uint8_t *p, ptrdiff_t stride)
 
 #define LOOP_FILTER(vpn, dir, size, stridea, strideb, maybe_inline)           \
 static maybe_inline                                                           \
-void vp ## vpn ## _ ## dir ## _loop_filter ## size ## _c(uint8_t *dst,        \
-                                               ptrdiff_t stride,              \
-                                               int flim_E, int flim_I,        \
-                                               int hev_thresh)                \
+void vpn ## _ ## dir ## _loop_filter ## size ## _c(uint8_t *dst,              \
+                                                   ptrdiff_t stride,          \
+                                                   int flim_E, int flim_I,    \
+                                                   int hev_thresh)            \
 {                                                                             \
     int i;                                                                    \
     for (i = 0; i < size; i++)                                                \
-        if (vp ## vpn ## _normal_limit(dst + i * stridea, strideb, flim_E, flim_I)) {       \
+        if (vpn ## _normal_limit(dst + i * stridea, strideb,                  \
+                                 flim_E, flim_I)) {                           \
             if (hev(dst + i * stridea, strideb, hev_thresh))                  \
-                filter_common(dst + i * stridea, strideb, 1, vpn);            \
+                vpn ## _filter_common(dst + i * stridea, strideb, 1);         \
             else                                                              \
                 filter_mbedge(dst + i * stridea, strideb);                    \
         }                                                                     \
 }                                                                             \
                                                                               \
 static maybe_inline                                                           \
-void vp ## vpn ## _ ## dir ## _loop_filter ## size ## _inner_c(uint8_t *dst,  \
-                                                     ptrdiff_t stride,        \
-                                                     int flim_E, int flim_I,  \
-                                                     int hev_thresh)          \
+void vpn ## _ ## dir ## _loop_filter ## size ## _inner_c(uint8_t *dst,        \
+                                                         ptrdiff_t stride,    \
+                                                         int flim_E,          \
+                                                         int flim_I,          \
+                                                         int hev_thresh)      \
 {                                                                             \
     int i;                                                                    \
     for (i = 0; i < size; i++)                                                \
-        if (vp ## vpn ## _normal_limit(dst + i * stridea, strideb, flim_E, flim_I)) {       \
+        if (vpn ## _normal_limit(dst + i * stridea, strideb,                  \
+                                 flim_E, flim_I)) {                           \
             int hv = hev(dst + i * stridea, strideb, hev_thresh);             \
             if (hv)                                                           \
-                filter_common(dst + i * stridea, strideb, 1, vpn);            \
+                vpn ## _filter_common(dst + i * stridea, strideb, 1);         \
             else                                                              \
-                filter_common(dst + i * stridea, strideb, 0, vpn);            \
+                vpn ## _filter_common(dst + i * stridea, strideb, 0);         \
         }                                                                     \
 }
 
-#define UV_LOOP_FILTER(vpn, dir, stridea, strideb) \
-LOOP_FILTER(vpn, dir, 8, stridea, strideb, av_always_inline) \
-static void vp ## vpn ## _ ## dir ## _loop_filter8uv_c(uint8_t *dstU, uint8_t *dstV, \
-                                                       ptrdiff_t stride, int fE, \
-                                                       int fI, int hev_thresh)\
-{\
-  vp ## vpn ## _ ## dir ## _loop_filter8_c(dstU, stride, fE, fI, hev_thresh);\
-  vp ## vpn ## _ ## dir ## _loop_filter8_c(dstV, stride, fE, fI, hev_thresh);\
-}\
-static void vp ## vpn ## _ ## dir ## _loop_filter8uv_inner_c(uint8_t *dstU, \
-                                                             uint8_t *dstV, \
-                                                             ptrdiff_t stride, int fE, \
-                                                             int fI, int hev_thresh) \
-{\
-  vp ## vpn ## _ ## dir ## _loop_filter8_inner_c(dstU, stride, fE, fI, hev_thresh);\
-  vp ## vpn ## _ ## dir ## _loop_filter8_inner_c(dstV, stride, fE, fI, hev_thresh);\
+#define UV_LOOP_FILTER(vpn, dir, stridea, strideb)                            \
+LOOP_FILTER(vpn, dir, 8, stridea, strideb, av_always_inline)                  \
+static void vpn ## _ ## dir ## _loop_filter8uv_c(uint8_t *dstU,               \
+                                                 uint8_t *dstV,               \
+                                                 ptrdiff_t stride, int fE,    \
+                                                 int fI, int hev_thresh)      \
+{                                                                             \
+    vpn ## _ ## dir ## _loop_filter8_c(dstU, stride, fE, fI, hev_thresh);     \
+    vpn ## _ ## dir ## _loop_filter8_c(dstV, stride, fE, fI, hev_thresh);     \
+}                                                                             \
+                                                                              \
+static void vpn ## _ ## dir ## _loop_filter8uv_inner_c(uint8_t *dstU,         \
+                                                       uint8_t *dstV,         \
+                                                       ptrdiff_t stride,      \
+                                                       int fE, int fI,        \
+                                                       int hev_thresh)        \
+{                                                                             \
+    vpn ## _ ## dir ## _loop_filter8_inner_c(dstU, stride, fE, fI,            \
+                                             hev_thresh);                     \
+    vpn ## _ ## dir ## _loop_filter8_inner_c(dstV, stride, fE, fI,            \
+                                             hev_thresh);                     \
 }
 
-#define LOOP_FILTER_SIMPLE(vpn) \
-static void vp ## vpn ## _v_loop_filter_simple_c(uint8_t *dst, ptrdiff_t stride, int flim)\
-{\
-    int i;\
-\
-    for (i = 0; i < 16; i++)\
-        if (vp ## vpn ## _simple_limit(dst + i, stride, flim))\
-            filter_common(dst + i, stride, 1, vpn);\
-}\
-\
-static void vp ## vpn ## _h_loop_filter_simple_c(uint8_t *dst, ptrdiff_t stride, int flim)\
-{\
-    int i;\
-\
-    for (i = 0; i < 16; i++)\
-        if (vp ## vpn ## _simple_limit(dst + i * stride, 1, flim))\
-            filter_common(dst + i * stride, 1, 1, vpn);\
+#define LOOP_FILTER_SIMPLE(vpn)                                               \
+static void vpn ## _v_loop_filter_simple_c(uint8_t *dst, ptrdiff_t stride,    \
+                                           int flim)                          \
+{                                                                             \
+    int i;                                                                    \
+    for (i = 0; i < 16; i++)                                                  \
+        if (vpn ## _simple_limit(dst + i, stride, flim))                      \
+            vpn ## _filter_common(dst + i, stride, 1);                        \
+}                                                                             \
+                                                                              \
+static void vpn ## _h_loop_filter_simple_c(uint8_t *dst, ptrdiff_t stride,    \
+                                           int flim)                          \
+{                                                                             \
+    int i;                                                                    \
+    for (i = 0; i < 16; i++)                                                  \
+        if (vpn ## _simple_limit(dst + i * stride, 1, flim))                  \
+            vpn ## _filter_common(dst + i * stride, 1, 1);                    \
 }
 
-#if CONFIG_VP7_DECODER
-LOOP_FILTER(7, v, 16, 1, stride,)
-LOOP_FILTER(7, h, 16, stride, 1,)
-UV_LOOP_FILTER(7, v, 1, stride)
-UV_LOOP_FILTER(7, h, stride, 1)
-LOOP_FILTER_SIMPLE(7)
-#endif
-
-#if CONFIG_VP8_DECODER
-LOOP_FILTER(8, v, 16, 1, stride,)
-LOOP_FILTER(8, h, 16, stride, 1,)
-UV_LOOP_FILTER(8, v, 1, stride)
-UV_LOOP_FILTER(8, h, stride, 1)
-LOOP_FILTER_SIMPLE(8)
-#endif
+#define LOOP_FILTERS(vpn)                \
+    LOOP_FILTER(vpn, v, 16, 1, stride, ) \
+    LOOP_FILTER(vpn, h, 16, stride, 1, ) \
+    UV_LOOP_FILTER(vpn, v, 1, stride)    \
+    UV_LOOP_FILTER(vpn, h, stride, 1)    \
+    LOOP_FILTER_SIMPLE(vpn)              \
 
 static const uint8_t subpel_filters[7][6] = {
     { 0,  6, 123,  12,  1, 0 },
@@ -622,7 +639,7 @@ VP8_BILINEAR(16)
 VP8_BILINEAR(8)
 VP8_BILINEAR(4)
 
-#define VP8_MC_FUNC(IDX, SIZE)                                                \
+#define VP78_MC_FUNC(IDX, SIZE)                                               \
     dsp->put_vp8_epel_pixels_tab[IDX][0][0] = put_vp8_pixels ## SIZE ## _c;   \
     dsp->put_vp8_epel_pixels_tab[IDX][0][1] = put_vp8_epel ## SIZE ## _h4_c;  \
     dsp->put_vp8_epel_pixels_tab[IDX][0][2] = put_vp8_epel ## SIZE ## _h6_c;  \
@@ -633,8 +650,8 @@ VP8_BILINEAR(4)
     dsp->put_vp8_epel_pixels_tab[IDX][2][1] = put_vp8_epel ## SIZE ## _h4v6_c; \
     dsp->put_vp8_epel_pixels_tab[IDX][2][2] = put_vp8_epel ## SIZE ## _h6v6_c
 
-#define VP8_BILINEAR_MC_FUNC(IDX, SIZE)                                       \
-    dsp->put_vp8_bilinear_pixels_tab[IDX][0][0] = put_vp8_pixels ## SIZE ## _c; \
+#define VP78_BILINEAR_MC_FUNC(IDX, SIZE)                                      \
+    dsp->put_vp8_bilinear_pixels_tab[IDX][0][0] = put_vp8_pixels   ## SIZE ## _c; \
     dsp->put_vp8_bilinear_pixels_tab[IDX][0][1] = put_vp8_bilinear ## SIZE ## _h_c; \
     dsp->put_vp8_bilinear_pixels_tab[IDX][0][2] = put_vp8_bilinear ## SIZE ## _h_c; \
     dsp->put_vp8_bilinear_pixels_tab[IDX][1][0] = put_vp8_bilinear ## SIZE ## _v_c; \
@@ -644,48 +661,79 @@ VP8_BILINEAR(4)
     dsp->put_vp8_bilinear_pixels_tab[IDX][2][1] = put_vp8_bilinear ## SIZE ## _hv_c; \
     dsp->put_vp8_bilinear_pixels_tab[IDX][2][2] = put_vp8_bilinear ## SIZE ## _hv_c
 
-av_cold void ff_vp8dsp_init(VP8DSPContext *dsp, int vp7)
+av_cold void ff_vp78dsp_init(VP8DSPContext *dsp)
 {
-#if CONFIG_VP7_DECODER && CONFIG_VP8_DECODER
-#define VPX(f) vp7 ? vp7_ ## f : vp8_ ## f
-#elif CONFIG_VP7_DECODER
-#define VPX(f) vp7_ ## f
-#else // CONFIG_VP8_DECODER
-#define VPX(f) vp8_ ## f
-#endif
-
-    dsp->vp8_luma_dc_wht    = VPX(luma_dc_wht_c);
-    dsp->vp8_luma_dc_wht_dc = VPX(luma_dc_wht_dc_c);
-    dsp->vp8_idct_add       = VPX(idct_add_c);
-    dsp->vp8_idct_dc_add    = VPX(idct_dc_add_c);
-    dsp->vp8_idct_dc_add4y  = VPX(idct_dc_add4y_c);
-    dsp->vp8_idct_dc_add4uv = VPX(idct_dc_add4uv_c);
-
-    dsp->vp8_v_loop_filter16y = VPX(v_loop_filter16_c);
-    dsp->vp8_h_loop_filter16y = VPX(h_loop_filter16_c);
-    dsp->vp8_v_loop_filter8uv = VPX(v_loop_filter8uv_c);
-    dsp->vp8_h_loop_filter8uv = VPX(h_loop_filter8uv_c);
-
-    dsp->vp8_v_loop_filter16y_inner = VPX(v_loop_filter16_inner_c);
-    dsp->vp8_h_loop_filter16y_inner = VPX(h_loop_filter16_inner_c);
-    dsp->vp8_v_loop_filter8uv_inner = VPX(v_loop_filter8uv_inner_c);
-    dsp->vp8_h_loop_filter8uv_inner = VPX(h_loop_filter8uv_inner_c);
-
-    dsp->vp8_v_loop_filter_simple = VPX(v_loop_filter_simple_c);
-    dsp->vp8_h_loop_filter_simple = VPX(h_loop_filter_simple_c);
-
-    VP8_MC_FUNC(0, 16);
-    VP8_MC_FUNC(1, 8);
-    VP8_MC_FUNC(2, 4);
-
-    VP8_BILINEAR_MC_FUNC(0, 16);
-    VP8_BILINEAR_MC_FUNC(1, 8);
-    VP8_BILINEAR_MC_FUNC(2, 4);
+    VP78_MC_FUNC(0, 16);
+    VP78_MC_FUNC(1, 8);
+    VP78_MC_FUNC(2, 4);
+
+    VP78_BILINEAR_MC_FUNC(0, 16);
+    VP78_BILINEAR_MC_FUNC(1, 8);
+    VP78_BILINEAR_MC_FUNC(2, 4);
 
     if (ARCH_ARM)
-        ff_vp8dsp_init_arm(dsp, vp7);
+        ff_vp78dsp_init_arm(dsp);
     if (ARCH_PPC)
-        ff_vp8dsp_init_ppc(dsp);
+        ff_vp78dsp_init_ppc(dsp);
+    if (ARCH_X86)
+        ff_vp78dsp_init_x86(dsp);
+}
+
+#if CONFIG_VP7_DECODER
+LOOP_FILTERS(vp7)
+
+av_cold void ff_vp7dsp_init(VP8DSPContext *dsp)
+{
+    dsp->vp8_luma_dc_wht    = vp7_luma_dc_wht_c;
+    dsp->vp8_luma_dc_wht_dc = vp7_luma_dc_wht_dc_c;
+    dsp->vp8_idct_add       = vp7_idct_add_c;
+    dsp->vp8_idct_dc_add    = vp7_idct_dc_add_c;
+    dsp->vp8_idct_dc_add4y  = vp7_idct_dc_add4y_c;
+    dsp->vp8_idct_dc_add4uv = vp7_idct_dc_add4uv_c;
+
+    dsp->vp8_v_loop_filter16y = vp7_v_loop_filter16_c;
+    dsp->vp8_h_loop_filter16y = vp7_h_loop_filter16_c;
+    dsp->vp8_v_loop_filter8uv = vp7_v_loop_filter8uv_c;
+    dsp->vp8_h_loop_filter8uv = vp7_h_loop_filter8uv_c;
+
+    dsp->vp8_v_loop_filter16y_inner = vp7_v_loop_filter16_inner_c;
+    dsp->vp8_h_loop_filter16y_inner = vp7_h_loop_filter16_inner_c;
+    dsp->vp8_v_loop_filter8uv_inner = vp7_v_loop_filter8uv_inner_c;
+    dsp->vp8_h_loop_filter8uv_inner = vp7_h_loop_filter8uv_inner_c;
+
+    dsp->vp8_v_loop_filter_simple = vp7_v_loop_filter_simple_c;
+    dsp->vp8_h_loop_filter_simple = vp7_h_loop_filter_simple_c;
+}
+#endif /* CONFIG_VP7_DECODER */
+
+#if CONFIG_VP8_DECODER
+LOOP_FILTERS(vp8)
+
+av_cold void ff_vp8dsp_init(VP8DSPContext *dsp)
+{
+    dsp->vp8_luma_dc_wht    = vp8_luma_dc_wht_c;
+    dsp->vp8_luma_dc_wht_dc = vp8_luma_dc_wht_dc_c;
+    dsp->vp8_idct_add       = vp8_idct_add_c;
+    dsp->vp8_idct_dc_add    = vp8_idct_dc_add_c;
+    dsp->vp8_idct_dc_add4y  = vp8_idct_dc_add4y_c;
+    dsp->vp8_idct_dc_add4uv = vp8_idct_dc_add4uv_c;
+
+    dsp->vp8_v_loop_filter16y = vp8_v_loop_filter16_c;
+    dsp->vp8_h_loop_filter16y = vp8_h_loop_filter16_c;
+    dsp->vp8_v_loop_filter8uv = vp8_v_loop_filter8uv_c;
+    dsp->vp8_h_loop_filter8uv = vp8_h_loop_filter8uv_c;
+
+    dsp->vp8_v_loop_filter16y_inner = vp8_v_loop_filter16_inner_c;
+    dsp->vp8_h_loop_filter16y_inner = vp8_h_loop_filter16_inner_c;
+    dsp->vp8_v_loop_filter8uv_inner = vp8_v_loop_filter8uv_inner_c;
+    dsp->vp8_h_loop_filter8uv_inner = vp8_h_loop_filter8uv_inner_c;
+
+    dsp->vp8_v_loop_filter_simple = vp8_v_loop_filter_simple_c;
+    dsp->vp8_h_loop_filter_simple = vp8_h_loop_filter_simple_c;
+
+    if (ARCH_ARM)
+        ff_vp8dsp_init_arm(dsp);
     if (ARCH_X86)
-        ff_vp8dsp_init_x86(dsp, vp7);
+        ff_vp8dsp_init_x86(dsp);
 }
+#endif /* CONFIG_VP8_DECODER */

libavcodec/vp8dsp.h

@@ -88,9 +88,18 @@ void ff_put_vp8_pixels8_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
 void ff_put_vp8_pixels4_c(uint8_t *dst, uint8_t *src, ptrdiff_t stride,
                           int h, int x, int y);
 
-void ff_vp8dsp_init(VP8DSPContext *c, int vp7);
-void ff_vp8dsp_init_x86(VP8DSPContext *c, int vp7);
-void ff_vp8dsp_init_arm(VP8DSPContext *c, int vp7);
-void ff_vp8dsp_init_ppc(VP8DSPContext *c);
+void ff_vp7dsp_init(VP8DSPContext *c);
+
+void ff_vp78dsp_init(VP8DSPContext *c);
+void ff_vp78dsp_init_arm(VP8DSPContext *c);
+void ff_vp78dsp_init_ppc(VP8DSPContext *c);
+void ff_vp78dsp_init_x86(VP8DSPContext *c);
+
+void ff_vp8dsp_init(VP8DSPContext *c);
+void ff_vp8dsp_init_arm(VP8DSPContext *c);
+void ff_vp8dsp_init_x86(VP8DSPContext *c);
+
+#define IS_VP7 1
+#define IS_VP8 0
 
 #endif /* AVCODEC_VP8DSP_H */

libavcodec/x86/h264_intrapred_init.c

@@ -231,7 +231,8 @@ av_cold void ff_h264_pred_init_x86(H264PredContext *h, int codec_id,
             h->pred4x4  [VERT_RIGHT_PRED        ] = ff_pred4x4_vertical_right_8_mmxext;
             h->pred4x4  [HOR_DOWN_PRED          ] = ff_pred4x4_horizontal_down_8_mmxext;
             h->pred4x4  [DC_PRED                ] = ff_pred4x4_dc_8_mmxext;
-            if (codec_id == AV_CODEC_ID_VP7 || codec_id == AV_CODEC_ID_VP8 || codec_id == AV_CODEC_ID_H264) {
+            if (codec_id == AV_CODEC_ID_VP7 || codec_id == AV_CODEC_ID_VP8 ||
+                codec_id == AV_CODEC_ID_H264) {
                 h->pred4x4  [DIAG_DOWN_LEFT_PRED] = ff_pred4x4_down_left_8_mmxext;
             }
             if (codec_id == AV_CODEC_ID_SVQ3 || codec_id == AV_CODEC_ID_H264) {

libavcodec/x86/vp8dsp_init.c

@@ -315,44 +315,18 @@ DECLARE_LOOP_FILTER(sse4)
     c->put_vp8_bilinear_pixels_tab[IDX][2][2] = ff_put_vp8_bilinear ## SIZE ## _hv_ ## OPT
 
 
-av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c, int vp7)
+av_cold void ff_vp78dsp_init_x86(VP8DSPContext *c)
 {
 #if HAVE_YASM
     int cpu_flags = av_get_cpu_flags();
 
     if (EXTERNAL_MMX(cpu_flags)) {
-        if (!vp7) {
-            c->vp8_idct_dc_add    = ff_vp8_idct_dc_add_mmx;
-            c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx;
-        }
 #if ARCH_X86_32
-        if (!vp7) {
-            c->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_mmx;
-            c->vp8_idct_add       = ff_vp8_idct_add_mmx;
-            c->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_mmx;
-        }
         c->put_vp8_epel_pixels_tab[0][0][0]     =
         c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_mmx;
 #endif
         c->put_vp8_epel_pixels_tab[1][0][0]     =
         c->put_vp8_bilinear_pixels_tab[1][0][0] = ff_put_vp8_pixels8_mmx;
-
-#if ARCH_X86_32
-        if (!vp7) {
-            c->vp8_v_loop_filter_simple   = ff_vp8_v_loop_filter_simple_mmx;
-            c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_mmx;
-
-            c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmx;
-            c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmx;
-            c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmx;
-            c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmx;
-
-            c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmx;
-            c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmx;
-            c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmx;
-            c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmx;
-        }
-#endif
     }
 
     /* note that 4-tap width=16 functions are missing because w=16
@@ -365,29 +339,10 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c, int vp7)
         VP8_MC_FUNC(1, 8, mmxext);
         VP8_BILINEAR_MC_FUNC(0, 16, mmxext);
         VP8_BILINEAR_MC_FUNC(1,  8, mmxext);
-
-        if (!vp7) {
-            c->vp8_v_loop_filter_simple   = ff_vp8_v_loop_filter_simple_mmxext;
-            c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_mmxext;
-
-            c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmxext;
-            c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmxext;
-            c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmxext;
-            c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmxext;
-
-            c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmxext;
-            c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmxext;
-            c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmxext;
-            c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmxext;
-        }
 #endif
     }
 
     if (EXTERNAL_SSE(cpu_flags)) {
-        if (!vp7) {
-            c->vp8_idct_add                         = ff_vp8_idct_add_sse;
-            c->vp8_luma_dc_wht                      = ff_vp8_luma_dc_wht_sse;
-        }
         c->put_vp8_epel_pixels_tab[0][0][0]     =
         c->put_vp8_bilinear_pixels_tab[0][0][0] = ff_put_vp8_pixels16_sse;
     }
@@ -397,64 +352,114 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c, int vp7)
         VP8_MC_FUNC(1, 8, sse2);
         VP8_BILINEAR_MC_FUNC(0, 16, sse2);
         VP8_BILINEAR_MC_FUNC(1, 8, sse2);
+    }
 
-        if (!vp7) {
-            c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2;
+    if (EXTERNAL_SSSE3(cpu_flags)) {
+        VP8_LUMA_MC_FUNC(0, 16, ssse3);
+        VP8_MC_FUNC(1, 8, ssse3);
+        VP8_MC_FUNC(2, 4, ssse3);
+        VP8_BILINEAR_MC_FUNC(0, 16, ssse3);
+        VP8_BILINEAR_MC_FUNC(1, 8, ssse3);
+        VP8_BILINEAR_MC_FUNC(2, 4, ssse3);
+    }
+#endif /* HAVE_YASM */
+}
 
-            c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2;
-            c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_sse2;
+av_cold void ff_vp8dsp_init_x86(VP8DSPContext *c)
+{
+#if HAVE_YASM
+    int cpu_flags = av_get_cpu_flags();
 
-            c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_sse2;
-            c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_sse2;
-        }
+    if (EXTERNAL_MMX(cpu_flags)) {
+        c->vp8_idct_dc_add    = ff_vp8_idct_dc_add_mmx;
+        c->vp8_idct_dc_add4uv = ff_vp8_idct_dc_add4uv_mmx;
+#if ARCH_X86_32
+        c->vp8_idct_dc_add4y  = ff_vp8_idct_dc_add4y_mmx;
+        c->vp8_idct_add       = ff_vp8_idct_add_mmx;
+        c->vp8_luma_dc_wht    = ff_vp8_luma_dc_wht_mmx;
+
+        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_mmx;
+        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_mmx;
+
+        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmx;
+        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmx;
+        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmx;
+        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmx;
+
+        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmx;
+        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmx;
+        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmx;
+        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmx;
+#endif
     }
 
-    if (EXTERNAL_SSE2(cpu_flags)) {
-        if (!vp7) {
-            c->vp8_idct_dc_add4y          = ff_vp8_idct_dc_add4y_sse2;
+    /* note that 4-tap width=16 functions are missing because w=16
+     * is only used for luma, and luma is always a copy or sixtap. */
+    if (EXTERNAL_MMXEXT(cpu_flags)) {
+#if ARCH_X86_32
+        c->vp8_v_loop_filter_simple   = ff_vp8_v_loop_filter_simple_mmxext;
+        c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_mmxext;
+
+        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_mmxext;
+        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_mmxext;
+        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_mmxext;
+        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_mmxext;
+
+        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_mmxext;
+        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_mmxext;
+        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_mmxext;
+        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_mmxext;
+#endif
+    }
 
-            c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse2;
+    if (EXTERNAL_SSE(cpu_flags)) {
+        c->vp8_idct_add                         = ff_vp8_idct_add_sse;
+        c->vp8_luma_dc_wht                      = ff_vp8_luma_dc_wht_sse;
+    }
+
+    if (HAVE_SSE2_EXTERNAL && cpu_flags & (AV_CPU_FLAG_SSE2 | AV_CPU_FLAG_SSE2SLOW)) {
+        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2;
 
-            c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_sse2;
-            c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_sse2;
+        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2;
+        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_sse2;
 
-            c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse2;
-            c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse2;
-        }
+        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_sse2;
+        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_sse2;
     }
 
-    if (EXTERNAL_SSSE3(cpu_flags)) {
-        VP8_LUMA_MC_FUNC(0, 16, ssse3);
-        VP8_MC_FUNC(1, 8, ssse3);
-        VP8_MC_FUNC(2, 4, ssse3);
-        VP8_BILINEAR_MC_FUNC(0, 16, ssse3);
-        VP8_BILINEAR_MC_FUNC(1, 8, ssse3);
-        VP8_BILINEAR_MC_FUNC(2, 4, ssse3);
+    if (EXTERNAL_SSE2(cpu_flags)) {
+        c->vp8_idct_dc_add4y          = ff_vp8_idct_dc_add4y_sse2;
 
-        if (!vp7) {
-            c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_ssse3;
-            c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_ssse3;
+        c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse2;
 
-            c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_ssse3;
-            c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_ssse3;
-            c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_ssse3;
-            c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_ssse3;
+        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_sse2;
+        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_sse2;
 
-            c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_ssse3;
-            c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_ssse3;
-            c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_ssse3;
-            c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_ssse3;
-        }
+        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse2;
+        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse2;
+    }
+
+    if (EXTERNAL_SSSE3(cpu_flags)) {
+        c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_ssse3;
+        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_ssse3;
+
+        c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_ssse3;
+        c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_ssse3;
+        c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_ssse3;
+        c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_ssse3;
+
+        c->vp8_v_loop_filter16y       = ff_vp8_v_loop_filter16y_mbedge_ssse3;
+        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_ssse3;
+        c->vp8_v_loop_filter8uv       = ff_vp8_v_loop_filter8uv_mbedge_ssse3;
+        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_ssse3;
     }
 
     if (EXTERNAL_SSE4(cpu_flags)) {
-        if (!vp7) {
-            c->vp8_idct_dc_add            = ff_vp8_idct_dc_add_sse4;
+        c->vp8_idct_dc_add            = ff_vp8_idct_dc_add_sse4;
 
-            c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse4;
-            c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse4;
-            c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse4;
-        }
+        c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse4;
+        c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse4;
+        c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse4;
     }
 #endif /* HAVE_YASM */
 }