[Gta04-owner] We need an Android kernel

[Dr. H. Nikolaus Schaller]

Hi,

Am 17.07.2013 um 15:31 schrieb Paul Kocialkowski:

> Le mercredi 17 juillet 2013 à 14:52 +0200, Dr. H. Nikolaus Schaller a
> écrit :
>> You mean touch screen orientation? I think this should be corrected/calibrated
>> somehow in user space so that we can keep the tsc2007 driver untouched.
> 
> There are several things to adapt on input for Android:
> * Monotonic time for event time stamps
> * Invert Y axis

Well, this may fail on the Letux 7004 anyways since there you have to swap
the axes in addition.

> * Ajust min/max values for each axis
> 
> This cannot be done in userspace: the Android framework expect every
> value that comes out of the input event node to be directly usable.

Ah, I see.

Well, we could do that also in the original neil-plus driver since Debian can
handle an already calibrated touch as well :) It could even simplify the X11
config as well.

So please share your patch.

> 
> So I quickly added support for all this in the tsc2007-gta04 driver,
> which works just fine.
> 
>>> The 3.7 kernel lacks
>>> Android USB Gadget support (the code is missing).
>> 
>> Ok, that may be one of the patches that are not yet in mainline Linux.
>> I have expected that we would not get 100% for free...
> 
> We need that badly, along with the other Android USB Composite
> functions. Our best hope is probably using the code from 3.8, but it
> won't be anything clean.
> 
>> Generally I think we now have a great basis to continue from!
> 
> It is far from perfect.

I didn't expect that it would be perfect...

> The amount of errors shown in userspace is
> incredible, many components complain about missing kernel interfaces,

Maybe /sys names are different.

> it
> seems that wakelocks are unusable with that kernel and the 4.0
> userspace. The UI is slower than on other versions,

That is unexpected... But maybe some process is running on 100%
because of one of the other bugs?

How does it manifest that wakelocks are unuseable?

> applications hang
> more than with any other kernel. On top of that, it seems that audio
> doesn't work most of the time (it did work once or twice). Some of these
> issues can be solved, some can't.
> 
> That's what I was mentioning when I said the 3.8 Android features are
> not complete, not stable, and that nobody uses them anyway.

I agree with the first 2 points, but the last one is the reason for the first
two.

If nobody decides to go this way, uses and tests the kernel and others
are working to fixing things, it will never become stable.

> 
> So the situation is the same as it was when we tested Neil's 3.7 kernel.
> However, it is true that suspend/resume works better than on the 3.0
> kernel.

That is IMHO an important point since I think nobody will be willing to
invest much into 3.0 or 3.4 suspend/resume issues if it works on 3.7 and
later.

> 
> I think I'll give 3.4 a try, depending on how suspend/resume works
> there, it might be worth investigating that way instead.

Yes, for a comparison this will be very important to know. 

Nevertheless I think it is more valuable investing time to fix the issues
that appear in the 3.7 kernel than working on (different) bugs of 3.4...

Maybe you (and we as testers) can file the issues here:

http://projects.goldelico.com/p/gta04-kernel/issues/

> 
>> As soon as you are happy with it, please push to the repo so that I (we)
>> can sync and try ourselves...

Is it possible to "push" several different kernel trees to the Replicant repo?
I.e. that we can choose to repo-sync / compile the 3.0, 3.4 and 3.7 kernels?

BR,
Nikolaus