On Tue, 9 Jun 2009, Mauro Carvalho Chehab wrote:
> Em Tue, 9 Jun 2009 18:08:38 +0200
> Hans Verkuil <hverk...@xs4all.nl> escreveu:
> > > > Should I submit a patch that implement VIDIOC_S_JPEGCOMP support in the
> > > > UVC driver and implement a JPEG compression quality control later, or
> > > > would you prefer the driver not to implement VIDIOC_S_JPEGCOMP at all ?
> > > > As there are existing applications using that ioctl a few users are
> > > > pushing for VIDIOC_S_JPEGCOMP support.
> > >
> > > I prefer the later. Adding a new ioctl support just to deprecate it on
> > > the next kernel doesn't seem nice. Let's add directly the newer controls
> > > and add a patch marking this as deprecated.
> >
> > I'm not sure whether we can deprecate JPEGCOMP. It is used in too many
> > places. Perhaps we should create a set of JPEG controls that match what is
> > in v4l2_jpegcompression and convert all the drivers that use JPEGCOMP to
> > these new controls. Then the v4l core can map S/G_JPEGCOMP ioctls to a set
> > of control read/writes. I'm working on string control support, so that will
> > allow us to handle the APP_data and COM_data fields.
> This seems to be the correct approach. Implement it as controls, and let
> video_ioctl2 to handle the calls to the legacy ioctls, while marking it as
> deprecate.

One problem is that COM and APP segment actually have a string of bytes
associated with them.  Right now we only have boolean and interger
controls.  There is no way to set a control to the 32-bytes you want put
into a APP segment.

Now, if we added NUL terminated string and fixed length byte arrays as
control types this could be done.  I know that's something I've mentioned
before.  It would also be a step to allowing drivers to export metadata
about captured images (exposured info, focus tracking, etc.) via a
control-like interface.
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

Reply via email to