[Gta04-owner] Updates for gta04 DTS file.

Dr. H. Nikolaus Schaller hns at goldelico.com
Wed Jan 7 21:22:42 CET 2015


Am 07.01.2015 um 20:27 schrieb NeilBrown <neilb at suse.de>:

> On Tue, 6 Jan 2015 21:28:28 +0100 "Dr. H. Nikolaus Schaller"
> <hns at goldelico.com> wrote:
>> Hi,
>> Am 06.01.2015 um 20:54 schrieb NeilBrown <neilb at suse.de>:
>>> Where do you suppose this “discussion during development" might happen?  
>> here on this list.
>>> As
>>> I've said before I think the best place for discussion is on the lists
>>> relevant to the specific subsystem. e.g. linux-mmc for mmc related things.
>> But do they know the GTA04 and what the GTA04 users want?
> Nothing in the GTA04 is really specific to the GTA04.  Every problem we face
> has been faced by someone else, or will be, or probably is being at this
> moment.  Every component in the GTA04 also exist in some other device, though
> possibly in a slightly different configuration.
> All that is specific to the GTA04 is the particular selection of components
> and the exact interconnections.  That is almost completely described by the
> DTS file.  Everything else is generic and so people with other hardware could
> be interested.
> *All* of my interactions with the broader community show that people want to
> help make my device work, but they want it to be done in a way that is a
> general and idiomatic as possible so that it won't just work for me but will
> improve Linux for anyone with a vaguely similar device.  This is exactly as
> it should be.
>>> We *need* to regularly interact with relevant
>>> parts of the broader community - to see ourselves as simply a part of that
>>> community and not something separate - to have any hope of producing anything
>>> “mature".
>> That is not excluded and I have had many discussions with Tomi Valkainen to
>> fix issues with the display subsystem.
>> So this all is not excluded by the development model or the gta04-kernel project.
> You obviously have had these very productive discussions outside the GTA04
> community, yet you seem to be telling me that I should be discussing things
> more inside the community.
> I am confused by this.

1. I discussed with Tomi some time ago (I think it was still for the 2.6.32 kernel) long before we started the current discussion
2. my memory is not optimal but I think there was nobody to discuss DSS inside the gta04 community (or we even did discuss and can find in the archive)
3. my recent discussion with Tomi were about OMAP5 panel drivers (so not GTA04 related) and were copied on the Pyra-kernel developer list
4. the GTA04 discussions about TVout (VENC, OPA362) were on LKML and I think the last patches were copied here as we did agree

I hope this solves the confusion.

> I think it is exactly right that you should  talk to Tomi about dss and I
> should talk to Uli about mmc and the rest of the gta04-community is welcome
> to listen in but is unlikely to contribute very much.
>> The problem is that you appear to think in “patches” while I think in “results” (i.e.
>> flat files in a tree).
> Definitely agree here.  To me, the patches are as much a part of the result
> as the working code is.   The "patch" is the basic unit of communication
> between developers.  

Which is quite a high barrier for new developers… To me it appears that long-term
kernel developers are not even aware of this effect.

It is like looking at your bank account and seeing only the many + and - .
Until there is a line “balance” you don’t know if the account is good or bancrupt.

I mean: from reading 10 patches I don’t get an understanding how let’s say USB works
and what the patches do optimize or have side-effects.

> The comment at the top of the patch is just as important
> as the code  in the patch (and sometimes more important).

Yes, this is my observation as well. Most documentation is done in commit messages.
Which makes them quite difficult to digest and pinpoint a problem :)

Before git was introduced (2.4) there were nice text books and blogs which did describe subsystems
on a higher level and did describe their interactions and APIs. But since git was used more and
more, there is nothing comparable which is up-to date and only a tendency to refer people with
questions to code and commit messages. Obviously it has become either uninteresting or too
difficult for text-book writers to follow.

But that is something we have to live with. And almost off-topic.

>  The patch must
> make sense in itself as well as it must improve the code that is being
> patched.

Of course.

Maybe there is also some more quantitative differentiation. There are patches
that introduce new functions. Maybe for the first time. And maybe not in an optimal
way. And others optimize existing things.

> Until I have a patch that is ready to post, I don't have anything of value to
> anyone but me.

Well, it could be a starting point for me to dig deeper into a problem, especially if
you don’t have time to continuously work on something until it is 100% done (I usually
can’t). And I might find the solution you have not yet seen. And I might not have to
restart from scratch and repeat your mistakes myself…

But I think we already have discussed these different views well enough to have a
better understanding.


More information about the Gta04-owner mailing list