Browse Source

H264: fix signed overflow in constant multiplication

This fixes libavcodec/h264.h:1100: warning: integer overflow in expression

Originally committed as revision 22558 to svn://svn.ffmpeg.org/ffmpeg/trunk
tags/v0.6
Måns Rullgård 15 years ago
parent
commit
404793f4ac
1 changed files with 1 additions and 1 deletions
  1. +1
    -1
      libavcodec/h264.h

+ 1
- 1
libavcodec/h264.h View File

@@ -1097,7 +1097,7 @@ static void fill_decode_caches(H264Context *h, int mb_type){
fill_rectangle(&h->direct_cache[scan8[0]], 4, 4, 8, MB_TYPE_16x16>>1, 1);

if(IS_DIRECT(top_type)){
AV_WN32A(&h->direct_cache[scan8[0] - 1*8], 0x01010101*(MB_TYPE_DIRECT2>>1));
AV_WN32A(&h->direct_cache[scan8[0] - 1*8], 0x01010101u*(MB_TYPE_DIRECT2>>1));
}else if(IS_8X8(top_type)){
int b8_xy = 4*top_xy;
h->direct_cache[scan8[0] + 0 - 1*8]= h->direct_table[b8_xy + 2];


Loading…
Cancel
Save