Move tegra in linux-next merge order

Move tegra in linux-next merge order

[Colin Cross]

I am going to push changes to tegra's for-next branch that will only
build after the asoc branch is merged, and will include a merge of a
small part of Ben's next-i2c branch.  Can you move tegra after these
branches?

Should I merge the asoc branch in below the patches that depend on it
in tegra's for-next tree, or should I push a branch that will not
build without merging into linux-next?
--
To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Move tegra in linux-next merge order

[Olof Johansson]

On Mon, Mar 14, 2011 at 1:37 PM, Colin Cross <ccross-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org> wrote:
> I am going to push changes to tegra's for-next branch that will only
> build after the asoc branch is merged, and will include a merge of a
> small part of Ben's next-i2c branch.  Can you move tegra after these
> branches?
>
> Should I merge the asoc branch in below the patches that depend on it
> in tegra's for-next tree, or should I push a branch that will not
> build without merging into linux-next?

My $.02:

If you want to push a branch that's not for-next that contains the
dependencies to make it easier for people to test, that might be a
good option.

For linux-next, if the branch is pulled by sfr after the asoc and i2c
trees, the option with least hassle is to just leave it out of your
branch (my board branch left them out, I just merged that branch into
linux-next to test it).


-Olof
--
To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Move tegra in linux-next merge order

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1479 bytes --]

Hi Colin,

On Mon, 14 Mar 2011 13:37:01 -0700 Colin Cross <ccross-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org> wrote:
>
> I am going to push changes to tegra's for-next branch that will only
> build after the asoc branch is merged, and will include a merge of a
> small part of Ben's next-i2c branch.  Can you move tegra after these
> branches?
> 
> Should I merge the asoc branch in below the patches that depend on it
> in tegra's for-next tree, or should I push a branch that will not
> build without merging into linux-next?

Sorry I didn't get back to you sooner.  I note that some of the tegra
code has gone into Linus' tree now.

Is the asoc tree you are talking about the sound-asoc tree
(git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound-2.6.git)?
If so, I could move the tegra tree merge to after that.   However, I
dilike having such dependencies in linux-next and much prefer for the
maintainers of the two trees to work it out themselves.

The usual way to do that is for (in this case) the asoc tree to have a
branch that contains the commits that the tegra tree depends on and for
both trees to merge the branch.   This is much better than (say)
cherry-picking the commits into the tegra tree.

The question in my mind is: How are you testing the tegra tree without
merging in the asoc tree anyway?
-- 
Cheers,
Stephen Rothwell                    sfr-3FnU+UHB4dNDw9hX6IcOSA@public.gmane.org
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

Re: Move tegra in linux-next merge order

[Mark Brown]

On Thu, Mar 17, 2011 at 09:10:48AM +1100, Stephen Rothwell wrote:

> The usual way to do that is for (in this case) the asoc tree to have a
> branch that contains the commits that the tegra tree depends on and for
> both trees to merge the branch.   This is much better than (say)
> cherry-picking the commits into the tegra tree.

We'd need to cherry pick the commits into a branch to create the new
ASoC branch anyway, I can't rebase the tree as it's merged elsewhere.
The need to use the code in the Tegra tree wasn't flagged until too
late.
--
To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Move tegra in linux-next merge order

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1182 bytes --]

Hi Mark,

On Wed, 16 Mar 2011 22:15:22 +0000 Mark Brown <broonie-yzvPICuk2AATkU/dhu1WVpmvS+d8x4Xz@public.gmane.orgcro.com> wrote:
>
> On Thu, Mar 17, 2011 at 09:10:48AM +1100, Stephen Rothwell wrote:
> 
> > The usual way to do that is for (in this case) the asoc tree to have a
> > branch that contains the commits that the tegra tree depends on and for
> > both trees to merge the branch.   This is much better than (say)
> > cherry-picking the commits into the tegra tree.
> 
> We'd need to cherry pick the commits into a branch to create the new
> ASoC branch anyway, I can't rebase the tree as it's merged elsewhere.
> The need to use the code in the Tegra tree wasn't flagged until too
> late.

All you would need to do is create a branch (ref) at the last of the
necessary commits in the asoc tree and Colin could pull that.  This means
that Colin will merge a (hopefully not to large) subset of the asoc tree.
Or he could just merge the asoc tree as it is today.  Either way, the
merge commit should contain an explanation.
-- 
Cheers,
Stephen Rothwell                    sfr-3FnU+UHB4dNDw9hX6IcOSA@public.gmane.org
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

Re: Move tegra in linux-next merge order

[Mark Brown]

On Thu, Mar 17, 2011 at 09:50:16AM +1100, Stephen Rothwell wrote:

> All you would need to do is create a branch (ref) at the last of the
> necessary commits in the asoc tree and Colin could pull that.  This means
> that Colin will merge a (hopefully not to large) subset of the asoc tree.
> Or he could just merge the asoc tree as it is today.  Either way, the
> merge commit should contain an explanation.

I wouldn't even need to create a branch, that doesn't make any practical
difference over working directly with the commit ID.  Colin wasn't
enthusiastic about merging the fairly active ASoC tree into Tegra, and
the Tegra side changes depend on other Tegra changes so the reverse
merge was an issue also.

Given how trivial the change is it may as well just get dropped in in
-rc1 - I really don't understand the handwringing here.
--
To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Move tegra in linux-next merge order

[Colin Cross]

On Wed, Mar 16, 2011 at 3:54 PM, Mark Brown
<broonie-yzvPICuk2AATkU/dhu1WVueM+bqZidxxQQ4Iyu8u01E@public.gmane.org> wrote:
> On Thu, Mar 17, 2011 at 09:50:16AM +1100, Stephen Rothwell wrote:
>
>> All you would need to do is create a branch (ref) at the last of the
>> necessary commits in the asoc tree and Colin could pull that.  This means
>> that Colin will merge a (hopefully not to large) subset of the asoc tree.
>> Or he could just merge the asoc tree as it is today.  Either way, the
>> merge commit should contain an explanation.
>
> I wouldn't even need to create a branch, that doesn't make any practical
> difference over working directly with the commit ID.  Colin wasn't
> enthusiastic about merging the fairly active ASoC tree into Tegra, and
> the Tegra side changes depend on other Tegra changes so the reverse
> merge was an issue also.
>
> Given how trivial the change is it may as well just get dropped in in
> -rc1 - I really don't understand the handwringing here.
>

The only issue is that the branch that is being merged by linux-next
is untestable on it's own.  But we can ignore the problem until the
asoc tree is merged into master, at which point the problem goes away.
 I plan to send a second pull request to Linus as soon as asoc is in,
we shouldn't need to wait for rc1.

We made sure the patches that depend on the asoc tree are at the top
of the stack for 2.6.39.  I do my testing on the last patch that
compiles without asoc (which is what got pulled into master today),
and rely on the people working with the board that depends on asoc to
test the last few patches.
--
To unsubscribe from this list: send the line "unsubscribe linux-tegra" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Move tegra in linux-next merge order

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1899 bytes --]

On Wed, 16 Mar 2011 15:58:26 -0700 Colin Cross <ccross-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org> wrote:
>
> On Wed, Mar 16, 2011 at 3:54 PM, Mark Brown
> <broonie-yzvPICuk2AATkU/dhu1WVueM+bqZidxxQQ4Iyu8u01E@public.gmane.org> wrote:
> > On Thu, Mar 17, 2011 at 09:50:16AM +1100, Stephen Rothwell wrote:
> >
> > I wouldn't even need to create a branch, that doesn't make any practical
> > difference over working directly with the commit ID.  Colin wasn't
> > enthusiastic about merging the fairly active ASoC tree into Tegra, and
> > the Tegra side changes depend on other Tegra changes so the reverse
> > merge was an issue also.

The advantage of a branch would be that if this was ongoing and fixes
needed to be applied to the asoc code, the they could be applied to that
branch and remerged into both trees.   This actually happened this time
round with the tip:irq tree and the net tree.

> > Given how trivial the change is it may as well just get dropped in in
> > -rc1 - I really don't understand the handwringing here.

Well, I wasn't aware of how much was involved ...

> The only issue is that the branch that is being merged by linux-next
> is untestable on it's own.  But we can ignore the problem until the
> asoc tree is merged into master, at which point the problem goes away.
>  I plan to send a second pull request to Linus as soon as asoc is in,
> we shouldn't need to wait for rc1.
> 
> We made sure the patches that depend on the asoc tree are at the top
> of the stack for 2.6.39.  I do my testing on the last patch that
> compiles without asoc (which is what got pulled into master today),
> and rely on the people working with the board that depends on asoc to
> test the last few patches.

OK, thanks guys.

-- 
Cheers,
Stephen Rothwell                    sfr-3FnU+UHB4dNDw9hX6IcOSA@public.gmane.org
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]