Re: [PATCH v5 15/39] [media] v4l2: add a frame interval error event

[Steve Longerbeam]

On 03/11/2017 10:51 AM, Russell King - ARM Linux wrote:
> On Sat, Mar 11, 2017 at 10:14:49AM -0800, Steve Longerbeam wrote:
> > On 03/11/2017 03:39 AM, Hans Verkuil wrote:
> > > It's fine to use an internal event as long as the end-user doesn't
> > > see it. But if you lose vsyncs, then you never capture another frame,
> > > right?
> >
> > No, that's not correct. By loss of vertical sync, I mean the IPU
> > captures portions of two different frames, resulting in a permanent
> > "split image", with one frame containing portions of two consecutive
> > images. Or, the video rolls continuously, if you remember the old CRT
> > television sets of yore, it's the same rolling effect.
>
> I have seen that rolling effect, but the iMX6 regains correct sync
> within one complete "roll" just fine here with IMX219.  However, it
> has always recovered.

I've seen permanent split images, and rolling that continues for
minutes.


> So, I don't think there's a problem with the iMX6 part of the
> processing, and so I don't think we should cripple the iMX6 capture
> drivers for this problem.
>
> It seems to me that the problem is with the source.

The the problem is almost surely in the CSI. It's handling of the
bt.656 sync codes is broken in some way.

Steve