Collection of patches for om-gta02-2.6.34

classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|

Collection of patches for om-gta02-2.6.34

Riccardo Magliocchetti
Hello,

i'd like to move from 2.6.29 debian kernel to something more recent so
this night i've started collecting patches found both in debian kernel
and in the last two months in this list. Haven't looked at shr repository.

Please be aware that still i haven't booted my fr with this kernel and i
think i'll try to add some debian packaging to it before doing so.

I've a couple of things to note:

- Very few people use git in the (proper?) kernel way: this mean adding
a signed off by line, a proper title and a proper description. Probably
because people who do the patches does not have a public git tree so
they don't care about that. That's bad imho.

- There's a lot of random patches available in trac and ml that have not
been applied. That's bad too. The cause for this _may_ be that only
larsc has write access. If is not possible to extend write access, what
if the other developers open a tree somewhere else so at least the
patches are stored in the proper way and the chances of having them lost
is reduced?

Anyway you can find the patches here:

http://people.freedesktop.org/~rm/2.6.34-patches/

Btw, if you see patches you have done with my name and signoff this is
not because i want to steal your work, i've cut & pasted stuff from
mailing list archives so haven't used proper tools like git am and my
git is configured to automatically add my signoff. Please understand that.

If something is wrong / missing, please shout :)

thanks,
riccardo

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Martin Jansa
On Thu, Aug 26, 2010 at 01:12:23AM +0200, Riccardo Magliocchetti wrote:
> Hello,

Hi,

> Anyway you can find the patches here:
>
> http://people.freedesktop.org/~rm/2.6.34-patches/

FYI: patches for SHR kernel are here:
http://gitorious.org/~jama/htc-msm-2-6-32/openmoko-kernel/commits/om-gta02-2.6.34

Regards,

--
Martin 'JaMa' Jansa     jabber: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Timo Jyrinki
In reply to this post by Riccardo Magliocchetti
2010/8/26 Riccardo Magliocchetti <[hidden email]>:
> - There's a lot of random patches available in trac and ml that have not
> been applied. That's bad too. The cause for this _may_ be that only larsc
> has write access. If is not possible to extend write access,

It is still possible to get new write access to git.openmoko.org [1].
I actually got one since I had an idea of trying to make a branch of
andy-tracking to something like 2.6.29-endusers. In the end I ended up
just committing to Debian's legacy pkg-fso 2.6.29 kernel branch. There
would be the job of applying 2.6.29rc3 -> 2.6.29.6 patch over our
2.6.29 [2], and I actually tried it but couldn't get rid of some
kernel panic at a quick try. The patch has really few conflicts.

[1] http://wiki.openmoko.org/wiki/Openmoko:Community_Portal#Git_Servers
[2] http://users.tkk.fi/~tajyrink/moko/2.6.29rc3-2.6.29.6.xz

> If something is wrong / missing, please shout :)

Well, in the case of Debian it was just recently noted [3] that it
would be best to concentrate efforts so that FreeRunner support could
become official part of Debian kernel. Therefore it's AFAIK
recommended to work on the s3c24xx branch [4], which doesn't boot yet
either. It's part of Thibaut Girka's Google SoC, debian-installer
support [5].

[3] http://lists.alioth.debian.org/pipermail/pkg-fso-maint/2010-August/003370.html
[4] http://git.debian.org/?p=pkg-fso/kernel.git;a=shortlog;h=refs/heads/s3c24xx
[5] http://wiki.debian.org/SummerOfCode2010/CrossbuildDebianInstaller/ThibautGirka

-Timo

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Vasily Khoruzhick
In reply to this post by Riccardo Magliocchetti
В сообщении от 26 августа 2010 02:12:23 автор Riccardo Magliocchetti написал:

> Hello,
>
> i'd like to move from 2.6.29 debian kernel to something more recent so
> this night i've started collecting patches found both in debian kernel
> and in the last two months in this list. Haven't looked at shr repository.
>
> Please be aware that still i haven't booted my fr with this kernel and i
> think i'll try to add some debian packaging to it before doing so.
>
> I've a couple of things to note:
>
> - Very few people use git in the (proper?) kernel way: this mean adding
> a signed off by line, a proper title and a proper description. Probably
> because people who do the patches does not have a public git tree so
> they don't care about that. That's bad imho.
>
> - There's a lot of random patches available in trac and ml that have not
> been applied. That's bad too. The cause for this _may_ be that only
> larsc has write access. If is not possible to extend write access, what
> if the other developers open a tree somewhere else so at least the
> patches are stored in the proper way and the chances of having them lost
> is reduced?
>
> Anyway you can find the patches here:
>
> http://people.freedesktop.org/~rm/2.6.34-patches/
>
> Btw, if you see patches you have done with my name and signoff this is
> not because i want to steal your work, i've cut & pasted stuff from
> mailing list archives so haven't used proper tools like git am and my
> git is configured to automatically add my signoff. Please understand that.
>
> If something is wrong / missing, please shout :)
>
> thanks,
> riccardo
It's pretty possible that patch 0004-Add-patch-from-debian-kernel-that-should-
help-with-W.patch is not neccessary when 0009 and 0010 patches are applied
(and DMA transfers option for s3cmci is enabled), however it needs testing.
Btw, update 0011 patch to enable CONFIG_MMC_S3C_DMA in kernel
(CONFIG_MMC_S3C_PIO should be disabled) if you want to get benefit from
0009-0010 patches.

Regards
Vasily

signature.asc (205 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Thibaut Girka
In reply to this post by Timo Jyrinki
> It is still possible to get new write access to git.openmoko.org [1].
Oh, maybe I should do that.

> > If something is wrong / missing, please shout :)
>
> Well, in the case of Debian it was just recently noted [3] that it
> would be best to concentrate efforts so that FreeRunner support could
> become official part of Debian kernel.
And to do that, it should be part of upstream kernel too.
I'm doing (I must admit, those days, really slowly) the work, but any
help is welcome!
We need to have what's in om-gta02-2.6.34 in mainline, especially the
glamo and jbt drivers. The jbt driver needs cleaning and reviewing, we
should also merge the changes made by Weiss for his KMS branch.
Another thing that is important is power management, gta02 files under
mach-s3c2440/. I'm not working on it yet, though...

> Therefore it's AFAIK recommended to work on the s3c24xx branch [4], which doesn't boot yet
> either.
I beg your pardon?

> It's part of Thibaut Girka's Google SoC, debian-installer support [5].

> [3] http://lists.alioth.debian.org/pipermail/pkg-fso-maint/2010-August/003370.html
> [4] http://git.debian.org/?p=pkg-fso/kernel.git;a=shortlog;h=refs/heads/s3c24xx
> [5] http://wiki.debian.org/SummerOfCode2010/CrossbuildDebianInstaller/ThibautGirka
Oh, my page is under the "CrossbuildDebianInstaller"? Interesting... need to change that :p

> -Timo

[Sorry for the double post, I've hit the wrong button...]


signature.asc (853 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Riccardo Magliocchetti
In reply to this post by Martin Jansa
Il 26/08/2010 06:56, Martin Jansa ha scritto:
> On Thu, Aug 26, 2010 at 01:12:23AM +0200, Riccardo Magliocchetti wrote:
>> Hello,
>
> Hi,

Hi Martin,

>> Anyway you can find the patches here:
>>
>> http://people.freedesktop.org/~rm/2.6.34-patches/
>
> FYI: patches for SHR kernel are here:
> http://gitorious.org/~jama/htc-msm-2-6-32/openmoko-kernel/commits/om-gta02-2.6.34
>

I have few questions:

- 113012b12e5188ae7d73857f6b1a214bb609b612 adds accelerometers but i see
there is a driver already in hwmon/lis3lv02d.c for what seems the same hw.

- 5cdee38c18074180f1958aaaa41f81140a8b2915 is something that should be
pushed upstream or is just a hack for ar6000?

thanks,
riccardo

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Martin Jansa
On Thu, Aug 26, 2010 at 10:48:04AM +0200, Riccardo Magliocchetti wrote:

> Il 26/08/2010 06:56, Martin Jansa ha scritto:
> > On Thu, Aug 26, 2010 at 01:12:23AM +0200, Riccardo Magliocchetti wrote:
> >> Hello,
> >
> > Hi,
>
> Hi Martin,
>
> >> Anyway you can find the patches here:
> >>
> >> http://people.freedesktop.org/~rm/2.6.34-patches/
> >
> > FYI: patches for SHR kernel are here:
> > http://gitorious.org/~jama/htc-msm-2-6-32/openmoko-kernel/commits/om-gta02-2.6.34
> >
>
> I have few questions:

Both patches were sent to this list
http://lists.openmoko.org/pipermail/openmoko-kernel/2010-April/010915.html
so better to discuss with original author

> - 113012b12e5188ae7d73857f6b1a214bb609b612 adds accelerometers but i see
> there is a driver already in hwmon/lis3lv02d.c for what seems the same hw.

with upstream driver I never got event interface available for userspace

> - 5cdee38c18074180f1958aaaa41f81140a8b2915 is something that should be
> pushed upstream or is just a hack for ar6000?

better to ask Radek

BTW: here is Radek's tree for qtmoko kernel
http://github.com/radekp/linux-2.6/commits/qtmoko-v25

Regards,
--
Martin 'JaMa' Jansa     jabber: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Riccardo Magliocchetti
In reply to this post by Timo Jyrinki
Il 26/08/2010 07:31, Timo Jyrinki ha scritto:

> Well, in the case of Debian it was just recently noted [3] that it
> would be best to concentrate efforts so that FreeRunner support could
> become official part of Debian kernel. Therefore it's AFAIK
> recommended to work on the s3c24xx branch [4], which doesn't boot yet
> either. It's part of Thibaut Girka's Google SoC, debian-installer
> support [5].
>
> [3] http://lists.alioth.debian.org/pipermail/pkg-fso-maint/2010-August/003370.html
> [4] http://git.debian.org/?p=pkg-fso/kernel.git;a=shortlog;h=refs/heads/s3c24xx
> [5] http://wiki.debian.org/SummerOfCode2010/CrossbuildDebianInstaller/ThibautGirka
>
> -Timo
>

I've read the discussion on list, the problem i have with [4] is that
Thibaut is committing patches the way debian kernel maintainers do and
not committing directly the code. I think it makes thing complicated,
even for him.

thanks,
riccardo

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Riccardo Magliocchetti
In reply to this post by Vasily Khoruzhick
Il 26/08/2010 07:47, Vasily Khoruzhick ha scritto:
> В сообщении от 26 августа 2010 02:12:23 автор Riccardo Magliocchetti написал:
>
> It's pretty possible that patch 0004-Add-patch-from-debian-kernel-that-should-
> help-with-W.patch is not neccessary when 0009 and 0010 patches are applied
> (and DMA transfers option for s3cmci is enabled), however it needs testing.
> Btw, update 0011 patch to enable CONFIG_MMC_S3C_DMA in kernel
> (CONFIG_MMC_S3C_PIO should be disabled) if you want to get benefit from
> 0009-0010 patches.

Thanks, will do that. I realized that the configure missed some pieces
while i was already in bed :)

riccardo

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Gaudenz Steinlin
In reply to this post by Riccardo Magliocchetti
Excerpts from Riccardo Magliocchetti's message of Don Aug 26 11:05:31 +0200 2010:

> Il 26/08/2010 07:31, Timo Jyrinki ha scritto:
> > Well, in the case of Debian it was just recently noted [3] that it
> > would be best to concentrate efforts so that FreeRunner support could
> > become official part of Debian kernel. Therefore it's AFAIK
> > recommended to work on the s3c24xx branch [4], which doesn't boot yet
> > either. It's part of Thibaut Girka's Google SoC, debian-installer
> > support [5].
> >
> > [3] http://lists.alioth.debian.org/pipermail/pkg-fso-maint/2010-August/003370.html
> > [4] http://git.debian.org/?p=pkg-fso/kernel.git;a=shortlog;h=refs/heads/s3c24xx
> > [5] http://wiki.debian.org/SummerOfCode2010/CrossbuildDebianInstaller/ThibautGirka
> >
> > -Timo
> >
>
> I've read the discussion on list, the problem i have with [4] is that
> Thibaut is committing patches the way debian kernel maintainers do and
> not committing directly the code. I think it makes thing complicated,
> even for him.

If you are refering to the pkg-fso/kernel.git, then this is because
this is not intended as a general kernel development tree. This tree
is only for the Debian packageing of the kernel and therefore follows
the SVN repository layout of the official Debian kernel package. The
goal is to ultimately integrate all changes there into the official
Debian kernel package. Patches commited there should be generated from
another repository (possibly openmokos kernel git repo) to minimize work.

I agree that this workflow is not optimal. But as long as the official
Debian kernel package is developed this way. It does not make things
easier for pkg-fso to deviate from this scheme. The complete
disconnection of the current pkg-fso kernel package
(linux-image-2.6.29-openmoko) from the Debian kernel package is one
reason why it has not been updated for a long time now.

IMHO all kernel development should be done in the openmoko kernel git
repository.

Gaudenz
--
Ever tried. Ever failed. No matter.
Try again. Fail again. Fail better.
~ Samuel Beckett ~

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Timo Jyrinki
In reply to this post by Thibaut Girka
2010/8/26 Thibaut Girka <[hidden email]>:
>> It is still possible to get new write access to git.openmoko.org [1].
> Oh, maybe I should do that.

I got my access just a month ago. Like noted by Sean at some point,
they still have one person paid to keep openmoko.org services running.

>> Therefore it's AFAIK recommended to work on the s3c24xx branch [4], which doesn't boot yet
>> either.
> I beg your pardon?

Sorry, I didn't mean to imply it's not working or anything, I just had
a vague memory that the branch didn't have yet all the drivers
required for booting. That, and writing in haste. Sorry for not
checking it properly.

> Oh, my page is under the "CrossbuildDebianInstaller"? Interesting... need to change that :p

:)

-Timo

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Mark Brown-3
In reply to this post by Gaudenz Steinlin
On Thu, Aug 26, 2010 at 02:35:27PM +0200, Gaudenz Steinlin wrote:

> I agree that this workflow is not optimal. But as long as the official
> Debian kernel package is developed this way. It does not make things
> easier for pkg-fso to deviate from this scheme. The complete
> disconnection of the current pkg-fso kernel package
> (linux-image-2.6.29-openmoko) from the Debian kernel package is one
> reason why it has not been updated for a long time now.

> IMHO all kernel development should be done in the openmoko kernel git
> repository.

As has already been pointed out in this thread the most direct way of
simplifying things here is probably to work to push the code directly
upstream.

Reply | Threaded
Open this post in threaded view
|

Re: Collection of patches for om-gta02-2.6.34

Gaudenz Steinlin
Excerpts from Mark Brown's message of Don Aug 26 14:59:04 +0200 2010:

> On Thu, Aug 26, 2010 at 02:35:27PM +0200, Gaudenz Steinlin wrote:
>
> > I agree that this workflow is not optimal. But as long as the official
> > Debian kernel package is developed this way. It does not make things
> > easier for pkg-fso to deviate from this scheme. The complete
> > disconnection of the current pkg-fso kernel package
> > (linux-image-2.6.29-openmoko) from the Debian kernel package is one
> > reason why it has not been updated for a long time now.
>
> > IMHO all kernel development should be done in the openmoko kernel git
> > repository.
>
> As has already been pointed out in this thread the most direct way of
> simplifying things here is probably to work to push the code directly
> upstream.

I totally agree. I just wanted to propose the openmoko repository as a
short term staging ground instead of the completely unsuitable  debian
packageing tree someone mentioned.

Gaudenz
--
Ever tried. Ever failed. No matter.
Try again. Fail again. Fail better.
~ Samuel Beckett ~