[Gta04-owner] Second trial with QtMoko2

[H. Nikolaus Schaller]

Hi,

> Am 04.04.2018 um 07:07 schrieb Andreas Kemnade <andreas at kemnade.info>:
> 
> On Tue, 3 Apr 2018 21:50:29 +0200
> "H. Nikolaus Schaller" <hns at goldelico.com> wrote:
> 
>>> Am 03.04.2018 um 19:48 schrieb Andreas Kemnade <andreas at kemnade.info>:
>>> 
>>> On Tue, 3 Apr 2018 19:19:31 +0200
>>> Sven Dyroff <S.Dyroff at phytec.de> wrote:
>>> 
>>>> Hello Andreas,
>>>> 
>>>>> So especially any regressions (compared to your v55 with old kernel) are
>>>> interesting.
>>>> 
>>>> Ok, thank you! Thus will focus on that, now.
>>>> 
>>>> But then please note that for me QtMaze is really an issue, because it's
>>>> not only a game. In fact it was a wonderful demonstration for the
>>>> accelerators of the GTA04.
>>>> 
>>> I found the piece of suspicious code:
>>> 
>>> in src/libraries/accelerometer/accelerometers.cpp
>>> char gta04_accel_file[] = "/dev/input/event1";
>> 
>> Oops! Well, event1 may be touch or rumble or something else.
>> On my device it is the "incoming call" event.
>> 
>>> 
>>> looks like something fixable. Probably it is better to use the iio
>>> interface now which will initially be a bit more work.
>> 
>> We can simply use "/dev/input/accel" since we have a special
>> iio-input bridge in Letux-Kernel which translates any iio based
>> accelerometer to input events.
>> 
> No, I do not think that it is a good idea.

Considering the small number of active developers, it is IMHO the
best approach. Just 1 line of code change.

> What is the upstreaming sitiuation for that?

I have not even attempted...

The problem with upstreaming such things is that if not even we do
not agree that it is something good, why should kernel maintainers?

And: I think we should have a higher priority to make QtMoko2 useable
than upstreaming. Upstreaming is a bonus...

If someone wants to give it a try, the bridge code is here:

http://git.goldelico.com/?p=gta04-kernel.git;a=shortlog;h=refs/heads/work/letux-base/hns/iio-input-bridge

> We have a standard interface in the kernel, so why we should not use
> it and use our own?

Because it is more work to take that way and we already have a
solution in the Letux OS ecosystem... So why not use what we have?

> The brigde is good for using old user spaces.

So far we have an "old user space". Even if we start to learn how
to improve.

> But if we change anything, I think it is good to use the correct
> interface.

If we had plenty of developers this would certainly be the best
approach. But for the situation we have, changing the gta04_accel_file
seems to be a much quicker fix.

I have done this and the result is available here:

http://git.goldelico.com/?p=gta04-qtmoko.git;a=blobdiff;f=src/libraries/accelerometer/accelerometers.cpp;h=33cb8b64f653acbe290e83de18a9f7848e3211d5;hp=0e22694ec894cdf9477f9cf4baf1af2575765a78;hb=79610ca47e8c46db5a0f8ba7bd3e5163c5152b64;hpb=4630e2c97f5141799967ee043dda602f133b4d75

NOTE: building new packages for apt-get update has started but
takes some time (hours)...

BR,
Nikolaus

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.goldelico.com/pipermail/gta04-owner/attachments/20180404/0b3508b2/attachment.asc>