On Tuesday, September 27, 2011 19:08:38 Laurent Pinchart wrote:
> Hi Hans,
>
> On Tuesday 27 September 2011 12:34:20 Hans Verkuil wrote:
> > On Thursday, September 08, 2011 09:45:15 Guennadi Liakhovetski wrote:
> > > A possibility to preallocate and initialise buffers of different sizes
> > > in V4L2 is required for an efficient implementation of a snapshot
> > > mode. This patch adds two new ioctl()s: VIDIOC_CREATE_BUFS and
> > > VIDIOC_PREPARE_BUF and defines respective data structures.
> > >
> > > Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@xxxxxx>
> > > ---
>
> [snip]
>
> > > @@ -2096,6 +2098,33 @@ static long __video_do_ioctl(struct file *file,
> > >
> > > dbgarg(cmd, "type=0x%8.8x", sub->type);
> > > break;
> > >
> > > }
> > >
> > > + case VIDIOC_CREATE_BUFS:
> > > + {
> > > + struct v4l2_create_buffers *create = arg;
> > > +
> > > + if (!ops->vidioc_create_bufs)
> > > + break;
> >
> > Just as with REQBUFS you need to add code here to handle priority checking:
> >
> > if (ret_prio) {
> > ret = ret_prio;
> > break;
> > }
>
> Speaking of prio support, how is locking handled here ?
Define 'here'. Are you talking about video_ioctl2() as a function? Or prio
support in particular?
> Does video_ioctl2()
> require drivers to synchronize all ioctl calls or can it work with fine-grain
> locking ?
It should always work OK with fine-grained locking.
Regards,
Hans
>
> > > + ret = check_fmt(ops, create->format.type);
> > > + if (ret)
> > > + break;
> > > +
> > > + ret = ops->vidioc_create_bufs(file, fh, create);
> > > +
> > > + dbgarg(cmd, "count=%d @ %d\n", create->count, create->index);
> > > + break;
> > > + }
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
