Re: [RFC v2 1/8] video: tegra: Add nvhost driver

[Thierry Reding <thierry.reding-RM9K5IK7kjKj5M59NBduVrNAH6kLmebB@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 3133 bytes --]

On Sun, Dec 02, 2012 at 01:24:13PM +0200, Terje Bergström wrote:
> On 01.12.2012 23:42, Dave Airlie wrote:
> > Guys I think you guys might be overthniking things here.
> > 
> > I know you have some sort of upstream/downstream split, but really in
> > the upstream kernel, we don't care about that, so don't make it our
> > problem.
> 
> I am not trying to make anything your problem. Most of the issues we
> have already worked out with a good solution that all active
> participants have agreed with. We have only a couple of disagreements
> with Thierry.
> 
> My goal is to get a good open source co-operation and trying to prevent
> a code fork while still maintaining good design. That way everybody
> wins. The way to do that is to base our BSP on upstream kernel.

Yes, that's exactly what you should be doing.

> I'm not trying to here throw code over the fence and flee. This is a
> genuine attempt to work together. I want to prevent the "we" (kernel
> community excluding NVIDIA) and "you" (NVIDIA) that a split code base
> would cause in the long run. I'd like to just talk about "we" including
> NVIDIA.

FWIW I'm convinced that you're genuinely trying to make this work and
nobody welcomes this more than me. However it is only natural if you
dump such a large body of code on the community that people will
disagree with some of the design decisions.

So when I comment on the design or patches in general, it is not my
intention to exclude you or NVIDIA in any way. All I'm trying to do is
spot problematic or unclear parts that will make working with the code
any more difficult than it has to be.

> > There is no need for any sort of stable API between host1x and the sub
> > drivers, we change APIs in the kernel the whole time it isn't a
> > problem.
> > 
> > If you need to change the API, submit a single patch changing it
> > across all the drivers in the tree, collecting Acks or not as needed.
> > We do this the whole time, I've never had or seen a problem with it.
> > 
> > We don't do separate subsystems APIs set in stone bullshit, and all
> > subsystem maintainers are used to dealing with these sort of issues.
> > You get an ack from one maintainer and the other one sticks it in his
> > tree with a note to Linus.
> > 
> > You can put the code where you want, maybe just under drivers/gpu
> > instead of drivers/video or drivers/gpu/drm, just make sure you have a
> > path for it into the kernel.
> 
> Follows exactly my thinking, as the location of host1x driver has no
> practical consequence to me.
> 
> Thierry proposed drivers/gpu/host1x. I'd like to see a couple of
> comments on that proposal, and if it sticks, follow that.
> 
> Thierry, did you mean that host1x driver would be in drivers/gpu/host1x,
> and tegradrm in drivers/gpu/drm/tegra, or would we put both in same
> directory?

Since tegra-drm is a DRM driver it should stay in drivers/gpu/drm. I can
also live with the host1x driver staying in drivers/video, but I don't
think it's the proper location and drivers/gpu/host1x seems like a much
better fit.

Thierry

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]