Make VP8 DSP functions take two strides

This isn't useful for the C functions, but will allow re-using H and V functions
for HV functions without adding separate H and V wrappers.

Originally committed as revision 23782 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavcodec/dsputil.c

@@ -2657,18 +2657,6 @@ static void avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
 }
 #endif /* CONFIG_RV40_DECODER */
 
-#if CONFIG_VP8_DECODER
-void ff_put_vp8_pixels16_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y) {
-    put_pixels16_c(dst, src, stride, h);
-}
-void ff_put_vp8_pixels8_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y) {
-    put_pixels8_c(dst, src, stride, h);
-}
-void ff_put_vp8_pixels4_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y) {
-    put_pixels4_c(dst, src, stride, h);
-}
-#endif
-
 static void wmv2_mspel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int w){
     uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
     int i;

libavcodec/vp8.c

@@ -912,7 +912,7 @@ static inline void vp8_mc(VP8Context *s, int luma,
                           uint8_t *dst, uint8_t *src, const VP56mv *mv,
                           int x_off, int y_off, int block_w, int block_h,
                           int width, int height, int linesize,
-                          h264_chroma_mc_func mc_func[3][3])
+                          vp8_mc_func mc_func[3][3])
 {
     static const uint8_t idx[8] = { 0, 1, 2, 1, 2, 1, 2, 1 };
     int mx = (mv->x << luma)&7, mx_idx = idx[mx];
@@ -931,7 +931,7 @@ static inline void vp8_mc(VP8Context *s, int luma,
         src = s->edge_emu_buffer + 2 + linesize * 2;
     }
 
-    mc_func[my_idx][mx_idx](dst, src, linesize, block_h, mx, my);
+    mc_func[my_idx][mx_idx](dst, linesize, src, linesize, block_h, mx, my);
 }
 
 /**

libavcodec/vp8dsp.c

@@ -250,6 +250,16 @@ static const uint8_t subpel_filters[7][6] = {
     { 0,   1,  12, 123,   6,   0 },
 };
 
+#define PUT_PIXELS(WIDTH) \
+static void put_vp8_pixels ## WIDTH ##_c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int x, int y) { \
+    for (int y = 0; y < h; y++, dst+= dststride, src+= srcstride) { \
+        memcpy(dst, src, WIDTH); \
+    } \
+}
+
+PUT_PIXELS(16)
+PUT_PIXELS(8)
+PUT_PIXELS(4)
 
 #define FILTER_6TAP(src, F, stride) \
     av_clip_uint8((F[2]*src[x+0*stride] - F[1]*src[x-1*stride] + F[0]*src[x-2*stride] + \
@@ -260,7 +270,7 @@ static const uint8_t subpel_filters[7][6] = {
                    F[3]*src[x+1*stride] - F[4]*src[x+2*stride] + 64) >> 7)
 
 #define VP8_EPEL_H(SIZE, FILTER, FILTERNAME) \
-static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, uint8_t *src, int stride, int h, int mx, int my) \
+static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int mx, int my) \
 { \
     const uint8_t *filter = subpel_filters[mx-1]; \
     int x, y; \
@@ -268,37 +278,37 @@ static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, uint8_t
     for (y = 0; y < h; y++) { \
         for (x = 0; x < SIZE; x++) \
             dst[x] = FILTER(src, filter, 1); \
-        dst += stride; \
-        src += stride; \
+        dst += dststride; \
+        src += srcstride; \
     } \
 }
 #define VP8_EPEL_V(SIZE, FILTER, FILTERNAME) \
-static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, uint8_t *src, int stride, int h, int mx, int my) \
+static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int mx, int my) \
 { \
     const uint8_t *filter = subpel_filters[my-1]; \
     int x, y; \
 \
     for (y = 0; y < h; y++) { \
         for (x = 0; x < SIZE; x++) \
-            dst[x] = FILTER(src, filter, stride); \
-        dst += stride; \
-        src += stride; \
+            dst[x] = FILTER(src, filter, srcstride); \
+        dst += dststride; \
+        src += srcstride; \
     } \
 }
 #define VP8_EPEL_HV(SIZE, FILTERX, FILTERY, FILTERNAME) \
-static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, uint8_t *src, int stride, int h, int mx, int my) \
+static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, int dststride, uint8_t *src, int srcstride, int h, int mx, int my) \
 { \
     const uint8_t *filter = subpel_filters[mx-1]; \
     int x, y; \
     uint8_t tmp_array[(2*SIZE+5)*SIZE]; \
     uint8_t *tmp = tmp_array; \
-    src -= 2*stride; \
+    src -= 2*srcstride; \
 \
     for (y = 0; y < h+5; y++) { \
         for (x = 0; x < SIZE; x++) \
             tmp[x] = FILTERX(src, filter, 1); \
         tmp += SIZE; \
-        src += stride; \
+        src += srcstride; \
     } \
 \
     tmp = tmp_array + 2*SIZE; \
@@ -307,7 +317,7 @@ static void put_vp8_epel ## SIZE ## _ ## FILTERNAME ## _c(uint8_t *dst, uint8_t
     for (y = 0; y < h; y++) { \
         for (x = 0; x < SIZE; x++) \
             dst[x] = FILTERY(tmp, filter, SIZE); \
-        dst += stride; \
+        dst += dststride; \
         tmp += SIZE; \
     } \
 }
@@ -338,7 +348,7 @@ VP8_EPEL_HV(8,  FILTER_6TAP, FILTER_6TAP, h6v6)
 VP8_EPEL_HV(4,  FILTER_6TAP, FILTER_6TAP, h6v6)
 
 #define VP8_MC_FUNC(IDX, SIZE) \
-    dsp->put_vp8_epel_pixels_tab[IDX][0][0] = ff_put_vp8_pixels ## SIZE ## _c; \
+    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; \
     dsp->put_vp8_epel_pixels_tab[IDX][1][0] = put_vp8_epel ## SIZE ## _v4_c; \

libavcodec/vp8dsp.h

@@ -27,6 +27,8 @@
 
 #include "dsputil.h"
 
+typedef void (*vp8_mc_func)(uint8_t *dst/*align 8*/, int dstStride, uint8_t *src/*align 1*/, int srcStride, int h, int x, int y);
+
 typedef struct VP8DSPContext {
     void (*vp8_luma_dc_wht)(DCTELEM block[4][4][16], DCTELEM dc[16]);
     void (*vp8_idct_add)(uint8_t *dst, DCTELEM block[16], int stride);
@@ -55,7 +57,7 @@ typedef struct VP8DSPContext {
      * third dimension: same as second dimention, for horizontal interpolation
      * so something like put_vp8_epel_pixels_tab[width>>3][2*!!my-(my&1)][2*!!mx-(mx&1)](..., mx, my)
      */
-    h264_chroma_mc_func put_vp8_epel_pixels_tab[3][3][3];
+    vp8_mc_func put_vp8_epel_pixels_tab[3][3][3];
 } VP8DSPContext;
 
 void ff_put_vp8_pixels16_c(uint8_t *dst, uint8_t *src, int stride, int h, int x, int y);