Re: [PATCH] soc-camera: fix compile breakage on SH

From: Magnus Damm
Date: Wed Oct 15 2008 - 04:56:15 EST


On Wed, Oct 15, 2008 at 5:26 PM, Guennadi Liakhovetski
<g.liakhovetski@xxxxxx> wrote:
> On Wed, 15 Oct 2008, Magnus Damm wrote:
>> On Wed, Oct 15, 2008 at 3:41 PM, Guennadi Liakhovetski
>> <g.liakhovetski@xxxxxx> wrote:
>> > On Wed, 15 Oct 2008, Magnus Damm wrote:
>> >
>> >> Thanks for working on fixing the breakage. I'd prefer to wait a bit
>> >> since there are quite a few pinmux patches queued up that may break if
>> >> we merge a fix right now. I can fix it up later on.
>> >
>> > no, I would not leave the kernel in a non-compilable state even if just
>> > for one board. Please, test a new version of the patch below. And yes, You
>> > will have to rebase your patches, sorry. Another thing, could you also,
>> > please, add a license / copyright header to
>> > include/media/soc_camera_platform.h?
>>
>> I'm not asking you to keep the board broken forever. It's just a
>> question of in which order the trees are getting merged. Again, I'd
>> rather see that this fix is put _on_top_ of the patches that are
>> already queued up in the SuperH tree. Merging it before doesn't help
>> anything in my opinion - especially since the change should go though
>> the SuperH tree anyway.
>
> I think, compilation-breakage fixes should have higher priority than
> further enhancements. Think about bisection. If you now first commit
> several more patches, you make the interval where the tree is not
> compilable longer, and thus the probabiliy that someone hits it in their
> git.bisect higher. That's why I think any compilation breakage should be
> fixed ASAP. And which changes do you mean specifically? This one:
>
> http://marc.info/?l=linux-sh&m=122346619318532&w=2
>
> Yes, indeed they conflict, but it is trivial to fix. So, I would prefer to
> close the compile-breakage window ASAP, and then trivially update that one
> your patch. Let's see what others say. And as for through which tree it
> should go, if you insist the sh-part going through the sh-tree, then it
> has to be split into two parts - video and sh. Thus extending the
> breakage-window by one commit...

Yeah, that one plus a patch for the smc91x platform data and another
one for mmc (which needs updating anyway). So maybe it's not such a
big deal. And I see your point with closing the window ASAP to do
damage control. Otoh I wonder how big difference it will be extending
the breakage window with one commit - there must be zillions of
commits in after the breakage already.

Paul, any strong feelings regarding merging things though the SuperH tree?

>> Feel free to add any header you like. =)
>
> Thanks, but no thanks:-) I cannot add your copyright, at least not without
> your explicit agreement (I think). So, I'd prefer you submit a patch for
> that.

I wonder if it's a large enough bit sequence to actually copyright. =)
But sure, I'll do that.

Is this .29 material, or will there be a second v4l round with trivial
driver changes for .28?

I've already posted some vivi patches and two simple patches for the
sh_mobile_ceu driver - sorry about the timing - and i have one more
sh_mobile_ceu patch outstanding. Also, I think one of my coworkers may
post a soc_camera driver for ov772x chips soon too.

Is there any chance that can get included in .28?

Thank you!

/ magnus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/