[FFmpeg-devel] [PATCH] lavfi/gradfun: fix direct writing in buffer.
Clément Bœsch
ubitux at gmail.com
Sun Dec 2 15:33:46 CET 2012
On Sun, Dec 02, 2012 at 11:48:03AM +0100, Nicolas George wrote:
> Le duodi 12 frimaire, an CCXXI, Clément Bœsch a écrit :
> > - first, the use of AV_PERM_PRESERVE looks inappropriate, so we can remove it.
>
> I confirm.
>
> > - then it seems assumed that gradfun can do filtering in-place (see the direct
> > flag). Though, that direct flag isn't passed to the filtering functions, and
>
> The direct flag is there to check whether the input buffer must be freed.
>
> > the filter doesn't seem to assume/support dst=src, so we need to get a new
> > temporary buffer (allocated with ff_get_video_buffer) all the time.
>
> If the filter did not work when direct is set, I believe we would probably
> have noticed it. Right now, a quick glance seems to show that it does work.
>
In that case, we should be able to do something like this:
diff --git a/libavfilter/vf_gradfun.c b/libavfilter/vf_gradfun.c
index 260a44b..0483b6c 100644
--- a/libavfilter/vf_gradfun.c
+++ b/libavfilter/vf_gradfun.c
@@ -186,23 +186,8 @@ static int filter_frame(AVFilterLink *inlink, AVFilterBufferRef *in)
{
GradFunContext *gf = inlink->dst->priv;
AVFilterLink *outlink = inlink->dst->outputs[0];
- AVFilterBufferRef *out;
- int p, direct = 0;
-
- if ((in->perms & AV_PERM_WRITE) && !(in->perms & AV_PERM_PRESERVE)) {
- direct = 1;
- out = in;
- } else {
- out = ff_get_video_buffer(outlink, AV_PERM_WRITE, outlink->w, outlink->h);
- if (!out) {
- avfilter_unref_bufferp(&in);
- return AVERROR(ENOMEM);
- }
-
- avfilter_copy_buffer_ref_props(out, in);
- out->video->w = outlink->w;
- out->video->h = outlink->h;
- }
+ AVFilterBufferRef *out = in;
+ int p;
for (p = 0; p < 4 && in->data[p]; p++) {
int w = inlink->w;
@@ -216,13 +201,8 @@ static int filter_frame(AVFilterLink *inlink, AVFilterBufferRef *in)
if (FFMIN(w, h) > 2 * r)
filter(gf, out->data[p], in->data[p], w, h, out->linesize[p], in->linesize[p], r);
- else if (out->data[p] != in->data[p])
- av_image_copy_plane(out->data[p], out->linesize[p], in->data[p], in->linesize[p], w, h);
}
- if (!direct)
- avfilter_unref_bufferp(&in);
-
return ff_filter_frame(outlink, out);
}
@@ -232,7 +212,7 @@ static const AVFilterPad avfilter_vf_gradfun_inputs[] = {
.type = AVMEDIA_TYPE_VIDEO,
.config_props = config_input,
.filter_frame = filter_frame,
- .min_perms = AV_PERM_READ,
+ .min_perms = AV_PERM_READ | AV_PERM_WRITE,
},
{ NULL }
};
diff --git a/libavfilter/vf_hqdn3d.c b/libavfilter/vf_hqdn3d.c
index 939a31d..0b7e33c 100644
--- a/libavfilter/vf_hqdn3d.c
+++ b/libavfilter/vf_hqdn3d.c
@@ -325,24 +325,8 @@ static int filter_frame(AVFilterLink *inlink, AVFilterBufferRef *in)
{
HQDN3DContext *hqdn3d = inlink->dst->priv;
AVFilterLink *outlink = inlink->dst->outputs[0];
-
- AVFilterBufferRef *out;
- int direct = 0, c;
-
- if (in->perms & AV_PERM_WRITE) {
- direct = 1;
- out = in;
- } else {
- out = ff_get_video_buffer(outlink, AV_PERM_WRITE, outlink->w, outlink->h);
- if (!out) {
- avfilter_unref_bufferp(&in);
- return AVERROR(ENOMEM);
- }
-
- avfilter_copy_buffer_ref_props(out, in);
- out->video->w = outlink->w;
- out->video->h = outlink->h;
- }
+ int c;
+ AVFilterBufferRef *out = in;
for (c = 0; c < 3; c++) {
denoise(hqdn3d, in->data[c], out->data[c],
@@ -353,9 +337,6 @@ static int filter_frame(AVFilterLink *inlink, AVFilterBufferRef *in)
hqdn3d->coefs[c?2:0], hqdn3d->coefs[c?3:1]);
}
- if (!direct)
- avfilter_unref_bufferp(&in);
-
return ff_filter_frame(outlink, out);
}
@@ -365,6 +346,7 @@ static const AVFilterPad avfilter_vf_hqdn3d_inputs[] = {
.type = AVMEDIA_TYPE_VIDEO,
.config_props = config_input,
.filter_frame = filter_frame,
+ .min_perms = AV_PERM_READ | AV_PERM_WRITE,
},
{ NULL }
};
The (untested) code will now assume src = dst all the time, and we can
maybe do some more simplification. The min_perms is here to make sure that
a new buffer is allocated when the input is read-only (that was done
manually previously).
currently:
if inbuf is RO:
- filter requests a new write buffer
- processing with src ≠ dst
else if inbuf is RW:
- processing with src = dst
after:
if inbuf is RO:
- lavfi requests a new write buffer so a new RW inbuf is passed
- processing with src = dst
Does that make any sense?
--
Clément B.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20121202/146d2cea/attachment.asc>
More information about the ffmpeg-devel
mailing list