Hi Laurent
On 10/08/2012 10:42 PM, Laurent Pinchart wrote:
> Hi,
>
> When did the {get,set}_frame_desc subdev operations reach mainline ? Last I
> knew is that they were an RFC, and they're now suddenly in, without even one
> ack from an embedded v4l developer :-S I'm not totally happy with that.
Sorry to hear now you have issues with those patches. I've sent a pull request
on last Wednesday [1], after 3 RFC versions of the whole change set. I've Cced
many people on these patches, including you and Sakari. I have addressed comments
to some patches you raised and I felt you're generally OK or don't have
objections (since there was no comments on some patches and I explicitly
mentioned there I'd like to get them in v3.7). I'm sorry if it was otherwise.
The problem I'm trying to address with those patches have been there for us for
over _one year_ [1], making JPEG capture with s5p-fimc effectively unusable in
the mainline kernel.
We've decided extending struct v4l2_mbus_framefmt was not an option, since it
is exposed to user-space now. Also using controls turned out to not be helpful,
since multiple values need to be passed, where there are multiple logical
streams transmitted by e.g. CSI-2 transmitter in single video frame [2].
It seemed like there is a general agreement on LMML to use the frame
descriptor callbacks instead. These callbacks are all in-kernel API and can
be changed any time. I'll be happy adapt the drivers to whatever sane changes
are proposed. I've stated in the commit description it's just an preliminary
form. And it at least let's us to move forward and carry on with more serious
problems.
I'll try to do better on letting people know, when sending things upstream
in future.
--
Regards,
Sylwester
[1] http://patchwork.linuxtv.org/patch/14875
[2] https://patchwork.kernel.org/patch/1365451
--
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
