Pandora audio using machine_is_omap3_pandora() check

Pandora audio using machine_is_omap3_pandora() check

[Tony Lindgren]

Hi Grazvydas,

Arnd noticed that pandora audio is using machine_is_omap3_pandora() check
that never succeeds for devicetree booting machines. Looks like this has
been broken at some point many years ago with the devicetree conversion.

Does anybody have an interest in fixing this driver?

Regards,

Tony

Re: Pandora audio using machine_is_omap3_pandora() check

[H. Nikolaus Schaller]

Hi Tony,

> Am 27.06.2022 um 11:52 schrieb Tony Lindgren <tony@atomide.com>:
> 
> Hi Grazvydas,
> 
> Arnd noticed that pandora audio is using machine_is_omap3_pandora() check
> that never succeeds for devicetree booting machines. Looks like this has
> been broken at some point many years ago with the devicetree conversion.
> 
> Does anybody have an interest in fixing this driver?

we already have fixes by Stefan Leichter and Grond here (incl. removing any call to machine_is_omap3_pandora):

	https://git.goldelico.com/?p=letux-kernel.git;a=shortlog;h=refs/heads/letux/sound-soc

But I don't know who can volunteer to run this series through the upstreaming discussions
and do regression tests (AFAIR the Pandora of Grond is broken and he has no replacement).

BR,
NIkolaus

Re: Pandora audio using machine_is_omap3_pandora() check

[Tony Lindgren]

* H. Nikolaus Schaller <hns@goldelico.com> [220627 11:36]:
> Hi Tony,
> 
> > Am 27.06.2022 um 11:52 schrieb Tony Lindgren <tony@atomide.com>:
> > 
> > Hi Grazvydas,
> > 
> > Arnd noticed that pandora audio is using machine_is_omap3_pandora() check
> > that never succeeds for devicetree booting machines. Looks like this has
> > been broken at some point many years ago with the devicetree conversion.
> > 
> > Does anybody have an interest in fixing this driver?
> 
> we already have fixes by Stefan Leichter and Grond here (incl. removing any call to machine_is_omap3_pandora):
> 
> 	https://git.goldelico.com/?p=letux-kernel.git;a=shortlog;h=refs/heads/letux/sound-soc

OK :)

> But I don't know who can volunteer to run this series through the upstreaming discussions
> and do regression tests (AFAIR the Pandora of Grond is broken and he has no replacement).

Probably best that Stefan and Grond do it :) Not sure what the minimal fix
for the mainline kernel might be to get things at least try to probe.

Regards,

Tony

Re: Pandora audio using machine_is_omap3_pandora() check

[Stefan Leichter]

Hi,

Am 27.06.22 um 15:53 schrieb Tony Lindgren:
> * H. Nikolaus Schaller <hns@goldelico.com> [220627 11:36]:
>> Hi Tony,
>>
>>> Am 27.06.2022 um 11:52 schrieb Tony Lindgren <tony@atomide.com>:
>>>
>>> Hi Grazvydas,
>>>
>>> Arnd noticed that pandora audio is using machine_is_omap3_pandora() check
>>> that never succeeds for devicetree booting machines. Looks like this has
>>> been broken at some point many years ago with the devicetree conversion.
>>>
>>> Does anybody have an interest in fixing this driver?
>>
>> we already have fixes by Stefan Leichter and Grond here (incl. removing any call to machine_is_omap3_pandora):
>>
>> 	https://git.goldelico.com/?p=letux-kernel.git;a=shortlog;h=refs/heads/letux/sound-soc
>
> OK :)
>
>> But I don't know who can volunteer to run this series through the upstreaming discussions
>> and do regression tests (AFAIR the Pandora of Grond is broken and he has no replacement).

I hope that Grond's Pandora didn't break because of experiments to get the sound working.

>
> Probably best that Stefan and Grond do it :) Not sure what the minimal fix
> for the mainline kernel might be to get things at least try to probe.
>

Well, I think I'm not the right person to try to mainline the patches, but if nobody else like to volunteer I will give it a try. Be warned, I have no experience with mainlining patches, even not with kernel programming at all. Therefore I need someone holding my hand. Does anybody volunteer?

I had a look to the branch Nikolaus pointed out. Here are some questions about the patches in the branch:

- It seams to me the branch doesn't support any longer to compile the driver directly into the kernel. Am I correct?
- There are three patches from Andreas and Nikolaus in the beginning of the branch. Are this patched needed to be mainlined too?
- Two of my patches are super seeded by the patches from Grond. My guess it that the patches needs to be squashed together?
- Where should I send the patches? The current audience might be a little bit large?

Regards Stefan

Re: Pandora audio using machine_is_omap3_pandora() check

[H. Nikolaus Schaller]

Hi,


> Am 05.07.2022 um 23:39 schrieb Stefan Leichter <sle85276@gmx.de>:
> 
> Hi,
> 
> Am 27.06.22 um 15:53 schrieb Tony Lindgren:
>> * H. Nikolaus Schaller <hns@goldelico.com> [220627 11:36]:
>>> Hi Tony,
>>> 
>>>> Am 27.06.2022 um 11:52 schrieb Tony Lindgren <tony@atomide.com>:
>>>> 
>>>> Hi Grazvydas,
>>>> 
>>>> Arnd noticed that pandora audio is using machine_is_omap3_pandora() check
>>>> that never succeeds for devicetree booting machines. Looks like this has
>>>> been broken at some point many years ago with the devicetree conversion.
>>>> 
>>>> Does anybody have an interest in fixing this driver?
>>> 
>>> we already have fixes by Stefan Leichter and Grond here (incl. removing any call to machine_is_omap3_pandora):
>>> 
>>> 	https://git.goldelico.com/?p=letux-kernel.git;a=shortlog;h=refs/heads/letux/sound-soc
>> 
>> OK :)
>> 
>>> But I don't know who can volunteer to run this series through the upstreaming discussions
>>> and do regression tests (AFAIR the Pandora of Grond is broken and he has no replacement).
> 
> I hope that Grond's Pandora didn't break because of experiments to get the sound working.

We tried to analyse but it is unilekely. It may just be wear somewhere else.
I also see XUDF error reports from I2C from time to time and it seems to depend a little on battery charge.
Difficult to pinpoint.

> 
>> 
>> Probably best that Stefan and Grond do it :) Not sure what the minimal fix
>> for the mainline kernel might be to get things at least try to probe.
>> 
> 
> Well, I think I'm not the right person to try to mainline the patches, but if nobody else like to volunteer I will give it a try. Be warned, I have no experience with mainlining patches, even not with kernel programming at all. Therefore I need someone holding my hand. Does anybody volunteer?

Yes, we all :)

Basically the task is to
a) shape (i.e. rebase, edit) the patches according the rules: https://www.kernel.org/doc/html/latest/process/submitting-patches.html
b) add Signed-Offs etc.
c) compile test
d) git format-patch
e) ./scripts/checkpatch.pl and fix any reported weaknesses
f) send by mail to this audience (perhaps check with ./MAINTAINERS)

> 
> I had a look to the branch Nikolaus pointed out. Here are some questions about the patches in the branch:
> 
> - It seams to me the branch doesn't support any longer to compile the driver directly into the kernel. Am I correct?

Hm. It still should? What makes you think so?

> - There are three patches from Andreas and Nikolaus in the beginning of the branch. Are this patched needed to be mainlined too?

No, they are not related to the Pandora but its cousin the GTA04.

> - Two of my patches are super seeded by the patches from Grond. My guess it that the patches needs to be squashed together?

Yes.

> - Where should I send the patches? The current audience might be a little bit large?

The audience is usually defined by ./MAINTAINERS
And whoever could contribute in review.

BR and thanks,
Nikolaus