Hi, On Thu, Jul 10, 2014 at 2:47 AM, Christophe Gisquet < christophe.gisq...@gmail.com> wrote:
> - mergecandlist[nb_merge_cand].mv[0].x = l0_cand.mv > [0].x; > - mergecandlist[nb_merge_cand].mv[0].y = l0_cand.mv > [0].y; > - mergecandlist[nb_merge_cand].mv[1].x = l1_cand.mv > [1].x; > - mergecandlist[nb_merge_cand].mv[1].y = l1_cand.mv > [1].y; > + AV_WN32(&mergecandlist[nb_merge_cand].mv[0], AV_RN32(& > l0_cand.mv[0])); > + AV_WN32(&mergecandlist[nb_merge_cand].mv[1], AV_RN32(& > l1_cand.mv[1])); AV_COPY32? Also use aligned writes/reads (AV_R/WN32A) in the other cases, since that prevents unaligned "per-byte" reads on some archs. Ronald _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
