Em Mon, 15 Sep 2014 08:56:40 -0400 Nicolas Dufresne <nicolas.dufre...@collabora.com> escreveu:
> > Le 2014-09-15 08:49, Laurent Pinchart a écrit : > > Reverting the patch will also be a regression, as that would break > > applications that now rely on the new behaviour (I've developed this patch > > to > > fix a problem I've noticed with gstreamer). One way or another, we're > > screwed > > and we'll break userspace. Well, VB1 is working with the old behavior, as the breakage on saa7134 only happened after its migration to VB2. So, the Gstreamer version you tested is still very likely broken with VB1. So, I still think that the less damage is to revert the POLLERR patch, with, according to Hans, is also a violation at the documented API. > We have worked around this issue in GStreamer 1.4+, Good to know. > for older version, > the problem may be faced again by users, specially if using a newer > libv4l2 where the locking has been fixed (or no libv4l2). The DQBUF locking fixup was merged on libv4l2 for version 1.2. So, the potential breakage happens when libv4l2 is 1.2 and Gstreamer versions before 1.4. Do you have any procedure on gstreamer to fix a bug on stable releases? Those VBI applications don't have any, as they're not actively maintained anymore. Even if we patch them today, I guess it could take a long time for those changes to be propagated on distros. So, I guess that the best is to try to fix Gstreamer on the distros that are using libv4l version 1.2 and a pre-1.4 Gstreamer version. > > cheers, > Nicolas > -- > To unsubscribe from this list: send the line "unsubscribe linux-media" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
