linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Regression: in v4l2 converter does not set the buffer.length anymore
@ 2014-09-29 21:23 Nicolas Dufresne
  2014-09-30 14:04 ` Hans de Goede
  0 siblings, 1 reply; 2+ messages in thread
From: Nicolas Dufresne @ 2014-09-29 21:23 UTC (permalink / raw)
  To: linux-media, Hans de Goede

This was initially reported to GStreamer project:
https://bugzilla.gnome.org/show_bug.cgi?id=737521

We track this down to be a regression introduced in v4l2-utils from 
version 1.4.0. In recent GStreamer we make sure the buffer.length field 
(retreived with QUERYBUF) is bigger or equal to the expected sizeimage 
(as obtained in S_FMT). This is to fail cleanly and avoid buffer 
overflow if a driver (or libv4l2) endup doing a short allocation. Since 
1.4.0, this field is always 0 if an emulated format is selected.

Reverting patch 10213c brings back normal behaviour:
http://git.linuxtv.org/cgit.cgi/v4l-utils.git/commit/?id=10213c975afdfcc90aa7de39e66c40cd7e8a57f7

This currently makes use of any emulated format impossible in GStreamer. 
v4l2-utils 1.4.0 is being shipped at least in debian/unstable at the moment.

cheers,
Nicolas

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Regression: in v4l2 converter does not set the buffer.length anymore
  2014-09-29 21:23 Regression: in v4l2 converter does not set the buffer.length anymore Nicolas Dufresne
@ 2014-09-30 14:04 ` Hans de Goede
  0 siblings, 0 replies; 2+ messages in thread
From: Hans de Goede @ 2014-09-30 14:04 UTC (permalink / raw)
  To: Nicolas Dufresne, linux-media

Hi,

On 09/29/2014 11:23 PM, Nicolas Dufresne wrote:
> This was initially reported to GStreamer project:
> https://bugzilla.gnome.org/show_bug.cgi?id=737521
> 
> We track this down to be a regression introduced in v4l2-utils from version 1.4.0. In recent GStreamer we make sure the buffer.length field (retreived with QUERYBUF) is bigger or equal to the expected sizeimage (as obtained in S_FMT). This is to fail cleanly and avoid buffer overflow if a driver (or libv4l2) endup doing a short allocation. Since 1.4.0, this field is always 0 if an emulated format is selected.
> 
> Reverting patch 10213c brings back normal behaviour:
> http://git.linuxtv.org/cgit.cgi/v4l-utils.git/commit/?id=10213c975afdfcc90aa7de39e66c40cd7e8a57f7
> 
> This currently makes use of any emulated format impossible in GStreamer. v4l2-utils 1.4.0 is being shipped at least in debian/unstable at the moment.

Oops, thanks for the bug report.

I've created a 3 patch patch-set fixing this, which I'll send right after
this mail. The actual fix is in the 2nd patch, the first patch and third
patches fix 2 unrelated bugs which I noticed while working on this.

Regards,

Hans

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2014-09-30 14:04 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-09-29 21:23 Regression: in v4l2 converter does not set the buffer.length anymore Nicolas Dufresne
2014-09-30 14:04 ` Hans de Goede

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).