Change linesize alignment method to ensure that linesize[0] == 2*linesize[1]

for YUV420P and YUV422P always.
Fixes MPEG-2 4:2:2 encoding on PPC and ARM, and should fix mxf_d10 regression.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18047 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

libavcodec/utils.c

@@ -242,6 +242,7 @@ int avcodec_default_get_buffer(AVCodecContext *s, AVFrame *pic){
         int h_chroma_shift, v_chroma_shift;
         int size[4] = {0};
         int tmpsize;
+        int unaligned;
         AVPicture picture;
         int stride_align[4];
 
@@ -254,8 +255,14 @@ int avcodec_default_get_buffer(AVCodecContext *s, AVFrame *pic){
             h+= EDGE_WIDTH*2;
         }
 
+        do {
+            // NOTE: do not align linesizes individually, this breaks e.g. assumptions
+            // that linesize[0] == 2*linesize[1] in the MPEG-encoder for 4:2:2
         ff_fill_linesize(&picture, s->pix_fmt, w);
+            // increase alignment of w for next try (rhs gives the lowest bit set in w)
+            w += w & ~(w-1);
 
+            unaligned = 0;
         for (i=0; i<4; i++){
 //STRIDE_ALIGN is 8 for SSE* but this does not work for SVQ1 chroma planes
 //we could change STRIDE_ALIGN to 16 for x86/sse but it would increase the
@@ -267,8 +274,9 @@ int avcodec_default_get_buffer(AVCodecContext *s, AVFrame *pic){
             else
 #endif
             stride_align[i] = STRIDE_ALIGN;
-            picture.linesize[i] = ALIGN(picture.linesize[i], stride_align[i]);
+                unaligned |= picture.linesize[i] % stride_align[i];
         }
+        } while (unaligned);
 
         tmpsize = ff_fill_pointer(&picture, NULL, s->pix_fmt, h);
         if (tmpsize < 0)