Re: [PATCH v5 00/39] i.MX Media Driver

[Niklas Söderlund]

On 2017-03-20 16:57:54 +0100, Hans Verkuil wrote:
> On 03/20/2017 03:11 PM, Russell King - ARM Linux wrote:
> > On Mon, Mar 20, 2017 at 02:57:03PM +0100, Hans Verkuil wrote:
> >> On 03/20/2017 02:29 PM, Russell King - ARM Linux wrote:
> >>> It's what I have - remember, not everyone is happy to constantly replace
> >>> their distro packages with random new stuff.
> >>
> >> This is a compliance test, which is continuously developed in tandem with
> >> new kernel versions. If you are working with an upstream kernel, then you
> >> should also use the corresponding v4l2-compliance test. What's the point
> >> of using an old one?
> >>
> >> I will not support driver developers that use an old version of the
> >> compliance test, that's a waste of my time.
> > 
> > The reason that people may _not_ wish to constantly update v4l-utils
> > is that it changes the libraries installed on their systems.
> > 
> > So, the solution to that is not to complain about developers not using
> > the latest version, but instead to de-couple it from the rest of the
> > package, and provide it as a separate, stand-alone package that doesn't
> > come with all the extra baggage.
> > 
> > Now, there's two possible answers to that:
> > 
> > 1. it depends on the libv4l2 version.  If that's so, then you are
> >    insisting that people constantly move to the latest libv4l2 because
> >    of API changes, and those API changes may upset applications they're
> >    using.  So this isn't really on.
> > 
> > 2. it doesn't depend on libv4l2 version, in which case there's no reason
> >    for it to be packaged with v4l-utils.
> 
> Run configure with --disable-v4l2-compliance-libv4l.
> 
> This avoids linking with libv4l and allows you to build it stand-alone.
> 
> Perhaps I should invert this option since in most cases you don't want to
> run v4l2-compliance with libv4l (it's off by default unless you pass the
> -w option to v4l2-compliance).
> 
> > 
> > The reality is that v4l2-compliance links with libv4l2, so I'm not sure
> > which it is.  What I am sure of is that I don't want to upgrade libv4l2
> > on an ad-hoc basis, potentially causing issues with applications.
> > 
> >>>> To test actual streaming you need to provide the -s option.
> >>>>
> >>>> Note: v4l2-compliance has been developed for 'regular' video devices,
> >>>> not MC devices. It may or may not work with the -s option.
> >>>
> >>> Right, and it exists to verify that the establised v4l2 API is correctly
> >>> implemented.  If the v4l2 API is being offered to user applications,
> >>> then it must be conformant, otherwise it's not offering the v4l2 API.
> >>> (That's very much a definition statement in itself.)
> >>>
> >>> So, are we really going to say MC devices do not offer the v4l2 API to
> >>> userspace, but something that might work?  We've already seen today
> >>> one user say that they're not going to use mainline because of the
> >>> crud surrounding MC.
> >>>
> >>
> >> Actually, my understanding was that he was stuck on the old kernel code.
> > 
> > Err, sorry, I really don't follow.  Who is "he"?
> 
> "one user say that they're not going to use mainline because of the
> crud surrounding MC."
> 
> > 
> > _I_ was the one who reported the EXPBUF problem.  Your comment makes no
> > sense.
> > 
> >> In the case of v4l2-compliance, I never had the time to make it work with
> >> MC devices. Same for that matter of certain memory to memory devices.
> >>
> >> Just like MC devices these too behave differently. They are partially
> >> supported in v4l2-compliance, but not fully.
> > 
> > It seems you saying that the API provided by /dev/video* for a MC device
> > breaks the v4l2-compliance tests?
> 
> There may be tests in the compliance suite that do not apply for MC devices
> and for which I never check. The compliance suite was never written with MC
> devices in mind, and it certainly hasn't been tested much with such devices.
> 
> It's only very recent that I even got hardware that has MC support...
> 
> From what I can tell from the feedback I got it seems to be OKish, but I
> just can't guarantee that the compliance utility is correct for such devices.
> 
> In particular I doubt the streaming tests (-s, -f, etc.) will work. The -s
> test *might* work if the pipeline is configured correctly before running
> v4l2-compliance. I can't imagine that the -f option would work at all since
> I would expect pipeline validation errors.

I successfully use v4l2-compliance with the -s option to test the 
Renesas R-Car Gen3 driver which uses MC, I first have to setup the 
pipeline using media-ctl. I have had much use of the tool and it have 
helped me catch many errors in the rcar-vin driver both on Gen2 (no MC 
involved) and Gen3. And yes the -f option is only usable on Gen2 where 
MC is not used.

For what it's worth, the first versions of the R-Car Gen3 patches did 
not use MC for anything else then setting up the pipeline, all format 
propagation and communication with subdevice where done using the 
/dev/videoX node and it was rather cumbersome. This was done in part due 
the me not understanding the MC framework. Later versions of the patches 
use MC and depends on the user configuring the pipeline and formats 
before starting the stream and it works better for me.

My ahh moment came when I realised that the pipeline can be different on 
different SoC but still be needed to be handled by the same drivers, I 
for example have two use-cases I look at.

Onboard components on Renesas R-Car Salvator-X. This was the only 
use-case I first considered and the early Gen3 patches handled this 
without MC support.

[ADV7482 (CSI-2 transmitter)] -> [R-Car CSI-2 (CSI-2 receiver)] -> [R-Car VIN (dma engine)]

Later I was presented with a more complex use-case where more devices 
would be involved.

[Sensor (Parallel output)] -> [Parallel to GSML] -> [GSML to CSI-2 (CSI-2 transmitter)] -> [R-Car CSI-2 (CSI-2 receiver)] -> [R-Car VIN (dma engine)]

And I could not see a reasonable way to support both use-cases (or 
others where a different pipeline would be connected to the CSI-2 
receiver present on the SoC itself) without the use of MC. There is no 
good way I can see where I can add the logic to the R-Car VIN driver 
which could cover all the knowledge needed to work with arbitrary 
pipelines. But then again I have lot to still learn about V4L2.

> 
> I've been gently pushing Helen Koike to finish her virtual MC driver
> (https://patchwork.kernel.org/patch/9312783/) since having a virtual driver
> makes writing compliance tests much easier.
> 
> > _No one_ has mentioned using v4l2-compliance on the subdevs.
> > 
> >> Complaining about this really won't help. We know it's a problem and unless
> >> someone (me perhaps?) manages to get paid to work on this it's unlikely to
> >> change for now.
> > 
> > Like the above comment, your comment makes no sense.  I'm not complaining,
> > I'm trying to find out the details.
> 
> Must be me then, it did feel like complaining...
> 
> > Yes, MC stuff sucks big time right now, the documentation is poor, there's
> > a lack of understanding on all sides of the issues (which can be seen by
> > the different opinions that people hold.)  The only way to resolve these
> > differences is via discussion, and if you're going to start thinking that
> > everyone is complaining, then there's not going to be any forward progress.
> > 
> 
> Regards,
> 
> 	Hans

-- 
Regards,
Niklas Söderlund