On Thu, Dec 04, 2014 at 03:28:32PM +0100, Benoit Fouet wrote:
> Hi,
>
> Le 03/12/2014 14:31, Benoit Fouet a écrit :
> >Rely on the way memcpy is optimized for one's system instead of looping
> >on a byte buffer for buffer copies to handle P frames.
> >---
> > libavcodec/pngdec.c | 25 +++++++++++++------------
> > 1 file changed, 13 insertions(+), 12 deletions(-)
> >
> >diff --git a/libavcodec/pngdec.c b/libavcodec/pngdec.c
> >index 7e7b285..e6b7593 100644
> >--- a/libavcodec/pngdec.c
> >+++ b/libavcodec/pngdec.c
> >@@ -884,8 +884,7 @@ static int handle_p_frame_apng(AVCodecContext *avctx,
> >PNGDecContext *s,
> > ff_thread_await_progress(&s->previous_picture, INT_MAX, 0);
> > for (j = 0; j < s->y_offset; j++) {
> >- for (i = 0; i < ls; i++)
> >- pd[i] = pd_last[i];
> >+ memcpy(pd, pd_last, ls);
> > pd += s->image_linesize;
> > pd_last += s->image_linesize;
> > }
> >@@ -907,8 +906,9 @@ static int handle_p_frame_apng(AVCodecContext *avctx,
> >PNGDecContext *s,
> > }
> > for (j = s->y_offset; j < s->y_offset + s->cur_h; j++) {
> >- for (i = 0; i < s->x_offset * s->bpp; i++)
> >- pd[i] = pd_last[i];
> >+ i = s->x_offset * s->bpp;
> >+ if (i)
> >+ memcpy(pd, pd_last, i);
> > for (; i < (s->x_offset + s->cur_w) * s->bpp; i += s->bpp) {
> > uint8_t alpha = pd[i+ai];
> >@@ -930,26 +930,27 @@ static int handle_p_frame_apng(AVCodecContext *avctx,
> >PNGDecContext *s,
> > break;
> > }
> > }
> >- for (; i < ls; i++)
> >- pd[i] = pd_last[i];
> >+ if (ls - i)
> >+ memcpy(pd+i, pd_last+i, ls - i);
> > pd += s->image_linesize;
> > pd_last += s->image_linesize;
> > pd_last_region += s->image_linesize;
> > }
> > } else {
> > for (j = s->y_offset; j < s->y_offset + s->cur_h; j++) {
> >- for (i = 0; i < s->x_offset * s->bpp; i++)
> >- pd[i] = pd_last[i];
> >- for (i = (s->x_offset + s->cur_w) * s->bpp; i < ls; i++)
> >- pd[i] = pd_last[i];
> >+ int end_offset = (s->x_offset + s->cur_w) * s->bpp;
> >+ int end_len = ls - end_offset;
> >+ if (s->x_offset)
> >+ memcpy(pd, pd_last, s->x_offset * s->bpp);
> >+ if (end_len)
> >+ memcpy(pd+end_offset, pd_last+end_offset, end_len);
> > pd += s->image_linesize;
> > pd_last += s->image_linesize;
> > }
> > }
> > for (j = s->y_offset + s->cur_h; j < s->height; j++) {
> >- for (i = 0; i < ls; i++)
> >- pd[i] = pd_last[i];
> >+ memcpy(pd, pd_last, ls);
> > pd += s->image_linesize;
> > pd_last += s->image_linesize;
> > }
>
> Unless there is some more remarks on this one, I think it can be
> applied too.applied thx [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB Many things microsoft did are stupid, but not doing something just because microsoft did it is even more stupid. If everything ms did were stupid they would be bankrupt already.
signature.asc
Description: Digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
