On Sunday 01 June 2008 04:34, Andy Walls wrote:
> On Mon, 2008-05-26 at 18:46 +0200, Hans Verkuil wrote:
> > On Monday 26 May 2008 18:34:57 Mauro Carvalho Chehab wrote:
> > > In the specific case of ivtv and cx18, I think that the better
> > > would be to convert it first to video_ioctl2. Then, remove the
> > > BKL, with a video_ioctl2_unlocked version.
> > >
> > > Douglas already did an experimental patch converting ivtv to
> > > video_ioctl2 and sent to Hans. It needs testing, since he doesn't
> > > have any ivtv board. It should be trivial to port this to cx18,
> > > since both drivers have similar structures.
> > >
> > > Douglas,
> > >
> > > Could you send this patch to the ML for people to review and for
> > > Andy to port it to cx18?
> >
> > Unless there is an objection, I would prefer to take Douglas' patch
> > and merge it into the v4l-dvb ivtv driver myself. There were
> > several things in the patch I didn't like, but I need to 'work'
> > with it a bit to see how/if it can be done better.
> >
> > I can work on it tonight and tomorrow. Hopefully it is finished by
> > then. I can move the BKL down at the same time for ivtv. It is
> > unlikely that I will have time to do cx18 as well as I'm abroad
> > from Wednesday until Monday, but I think Andy can do that easily
> > based on the ivtv changes.
>
> I have attached a patch, made against Hans' v4l-dvb-ioctl2
> repository, to convert the cx18 driver to use video_ioctl2(). In the
> process I pushed down the priority checks and the debug messages into
> the individual functions. I did not remove the serialization lock as
> I have not had the time to assess if that would be safe. I "#if
> 0"'ed out some sliced VBI code that was being skipped in the original
> code.
>
> Comments are welcome.
>
> Many thanks to Hans, for without his changeset to ivtv for reference,
> this would have taken me much, much longer.
>
>
> Short term initial tests of this patch with MythTV, v4l-ctl, and
> v4l-dbg indicate things are OK. I did notice once strange artifact
> with MythTV. When switching from one analog channel to another, for
> about a second, I see the two television video fields non-interlaced,
> stacked one atop of the other in the frame. Weird, but not a big
> deal.
Thanks Andy!
I'll take a closer look on Tuesday or Wednesday, but I noticed one
thing: you set unused callbacks to NULL in cx18_set_funcs(), however
these can just be removed as they are NULL by default.
Regards,
Hans
_______________________________________________
ivtv-devel mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-devel
