Re: Nokia N900 cameras -- pipeline setup in python (was Re: [RFC PATCH 00/24] Make Nokia N900 cameras working)

From: Pavel Machek <pavel@ucw.cz>
To: Sakari Ailus <sakari.ailus@iki.fi>
Cc: Sebastian Reichel <sre@kernel.org>,
	Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com>,
	pali.rohar@gmail.com, linux-media@vger.kernel.org
Subject: Re: Nokia N900 cameras -- pipeline setup in python (was Re: [RFC PATCH 00/24] Make Nokia N900 cameras working)
Date: Tue, 21 Jun 2016 20:05:49 +0200	[thread overview]
Message-ID: <20160621180549.GA9602@amd> (raw)
In-Reply-To: <20160620205904.GL24980@valkosipuli.retiisi.org.uk>

Hi!

> > > First, I re-did pipeline setup in python, it seems slightly less hacky
> > > then in shell.
> > > 
> > > I tried to modify fcam-dev to work with the new interface, but was not
> > > successful so far. I can post patches if someone is interested
> > > (mplayer works for me, but that's not too suitable for taking photos).
> > > 
> > > I tried to get gstreamer to work, with something like:
> > 
> > While trying to debug gstreamer, I ran v4l2-compliance, and it seems
> > to suggest that QUERYCAP is required... but it is not present on
> > /dev/video2 or video6.
> 
> It's not saying that it wouldn't be present, but the content appears wrong.
> It should have the real bus information there rather than just "media".
> 
> See e.g. drivers/media/platform/vsp1/vsp1_drv.c . I suppose that should be
> right.
> 
> Feel free to submit a patch. :-)

For now I'd not know what to change, sorry :-(. Perhaps we can debug
it after the support is merged into mainline.

Another weirdness:

yavta, on v4l-subdev12 :

control 0x00a40906 `Sensivity' min 0 max 0 step 1 default 0 current 65536.

Min and max being the same, I don't think I can control the
sensitivity. I guess I'll have to provid  more light for the tests for
now...

Best regards,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html