[GIT PULL 0/5] Second batch of arm-soc changes

[GIT PULL 0/5] Second batch of arm-soc changes

[Olof Johansson]

Hi Linus,

Here is the second batch of arm-soc changes for this merge window. As Arnd
mentioned with the first cover letter, most of these either had external
dependencies that have now been pulled in, or just needed a bit more
linux-next bake time due to getting picked up close to the 3.3 release.


There are 128 non-merge patches, with a total diffstat of:

 346 files changed, 10790 insertions(+), 2792 deletions(-)

So, while there's been some good cleanups especially in the first branch,
there's been a net addition of code:

* New SoC family support for Samsung EXYNOS5250: clock tables, driver
  updates, etc means net gain in code.
* Added device tree conversions. Once we start to sunset non-device tree
  boards on the enabled platforms, we should see payback in code removal.
* Common clock framework. As SoCs get converted over to the framework,
  we should see more code removal from this as well.


There are a handful of merge conflicts on the branches, nothing that
should be too messy to sort out. I've described each conflict in the
corresponding pull request, including how we've resolved them in our
for-next branch.


Thanks,

-Olof

[GIT PULL 1/5] ARM: cleanups of io includes

[Russell King - ARM Linux]

On Wed, Mar 28, 2012 at 12:13:58AM -0700, Olof Johansson wrote:
> Rob Herring has done a sweeping change cleaning up all of the mach/io.h includes,
> moving some of the oft-repeated macros to a common location and removing a bunch of
> boiler plate. This is another step closer to a common zImage for multiple platforms.

NAK.  Linus, do not pull this.  This breaks PXA compilation, which has
been known about for more than a week.

[GIT PULL 1/5] ARM: cleanups of io includes

[Olof Johansson]

On Wed, Mar 28, 2012 at 1:25 AM, Russell King - ARM Linux
<linux@arm.linux.org.uk> wrote:
> On Wed, Mar 28, 2012 at 12:13:58AM -0700, Olof Johansson wrote:
>> Rob Herring has done a sweeping change cleaning up all of the mach/io.h includes,
>> moving some of the oft-repeated macros to a common location and removing a bunch of
>> boiler plate. This is another step closer to a common zImage for multiple platforms.
>
> NAK. ?Linus, do not pull this. ?This breaks PXA compilation, which has
> been known about for more than a week.

The branch contains fixes for the compilation failures, they were
pulled in right before I staged the tag for the pull request:

http://marc.info/?l=linux-arm-kernel&m=133272592723334

If you prefer to let it sit in linux-next for a day or two before
going in, that's fine. But I verified that the original build problem
introduced was fixed before I sent it out.

Unfortunately, PXA is also broken due to the regulator changes
mentioned in the above referenced mail, which is already upstream and
unrelated to the changes in this branch.

Linus; the rest of the pull requests should be OK to go in
independently of this, so please go ahead with those if you don't
mind. The only changes to those branches in the last few days is the
merge resolution I did to the drivers2 branch/tag.


-Olof

[GIT PULL 1/5] ARM: cleanups of io includes

[Arnd Bergmann]

On Wednesday 28 March 2012, Olof Johansson wrote:
> If you prefer to let it sit in linux-next for a day or two before
> going in, that's fine. But I verified that the original build problem
> introduced was fixed before I sent it out.
> 
> Unfortunately, PXA is also broken due to the regulator changes
> mentioned in the above referenced mail, which is already upstream and
> unrelated to the changes in this branch.
> 
> Linus; the rest of the pull requests should be OK to go in
> independently of this, so please go ahead with those if you don't
> mind. The only changes to those branches in the last few days is the
> merge resolution I did to the drivers2 branch/tag.

Yes, sounds good. I've also verified that none of the other branches
depends on this one.

	Arnd

[GIT PULL 1/5] ARM: cleanups of io includes

[Russell King - ARM Linux]

On Wed, Mar 28, 2012 at 02:05:50AM -0700, Olof Johansson wrote:
> On Wed, Mar 28, 2012 at 1:25 AM, Russell King - ARM Linux
> <linux@arm.linux.org.uk> wrote:
> > On Wed, Mar 28, 2012 at 12:13:58AM -0700, Olof Johansson wrote:
> >> Rob Herring has done a sweeping change cleaning up all of the mach/io.h includes,
> >> moving some of the oft-repeated macros to a common location and removing a bunch of
> >> boiler plate. This is another step closer to a common zImage for multiple platforms.
> >
> > NAK. ?Linus, do not pull this. ?This breaks PXA compilation, which has
> > been known about for more than a week.
> 
> The branch contains fixes for the compilation failures, they were
> pulled in right before I staged the tag for the pull request:
> 
> http://marc.info/?l=linux-arm-kernel&m=133272592723334
> 
> If you prefer to let it sit in linux-next for a day or two before
> going in, that's fine. But I verified that the original build problem
> introduced was fixed before I sent it out.

Right, having merged the latest arm-soc tree this evening into the build
tree and forced a rebuild, the errors induced by Rob's patches are now
gone, so I'm fine with this going in now.

> Unfortunately, PXA is also broken due to the regulator changes
> mentioned in the above referenced mail, which is already upstream and
> unrelated to the changes in this branch.

Yes, I noticed that appearing when I updated the build tree on the 27th.
Not impressed by that.

[GIT PULL 1/5] ARM: cleanups of io includes

[Linus Torvalds]

On Wed, Mar 28, 2012 at 12:28 PM, Russell King - ARM Linux
<linux@arm.linux.org.uk> wrote:
>
> Right, having merged the latest arm-soc tree this evening into the build
> tree and forced a rebuild, the errors induced by Rob's patches are now
> gone, so I'm fine with this going in now.

Ok, I'm getting back to this pull request, and apparently everything
is ok with it.

Holler if not, because it's not back on my queue of trees to merge
(but I haven't actually gotten to it yet - this is just a heads-up
that I will get to it later today).

                        Linus

[GIT PULL 1/5] ARM: cleanups of io includes

[Olof Johansson]

On Thu, Mar 29, 2012 at 5:09 PM, Linus Torvalds
<torvalds@linux-foundation.org> wrote:
> On Wed, Mar 28, 2012 at 12:28 PM, Russell King - ARM Linux
> <linux@arm.linux.org.uk> wrote:
>>
>> Right, having merged the latest arm-soc tree this evening into the build
>> tree and forced a rebuild, the errors induced by Rob's patches are now
>> gone, so I'm fine with this going in now.
>
> Ok, I'm getting back to this pull request, and apparently everything
> is ok with it.
>
> Holler if not, because it's not back on my queue of trees to merge
> (but I haven't actually gotten to it yet - this is just a heads-up
> that I will get to it later today).

I was going to send a fresh pull request for it and include some of
the other PXA breakage fixes with it, but I haven't received those
from the PXA maintainer yet. I'll send a separate pull request for a
fixes branch when I get it instead (this pull request does not add any
new breakage).

So please, go ahead and pull. It hasn't grown any new merge conflicts
since before as of now, so see the original pull request for those.


Thanks!

-Olof

[GIT PULL 1/5] ARM: cleanups of io includes

[Olof Johansson]

On Thu, Mar 29, 2012 at 5:27 PM, Olof Johansson <olof@lixom.net> wrote:

> It hasn't grown any new merge conflicts
> since before as of now, so see the original pull request for those.

D'oh, I did a git remote update but didn't checkout the new head for
your tree when I verified the merge conflict situation. It does have a
couple of new trivial ones:

* arch/arm/mach-tegra/Makefile: keep reset.o, remove localtimer.o
* arch/arm/mach-vexpress/include/mach/io.h: line remove vs file remove
conflict: file is gone, but a patch in rmk's tree had removed a line
from it. It's safe to still remove the file.

+ the previous ones from original pull request:

arch/arm/Kconfig: New selects added for ARCH_S3C24X. Include both
sides of the merge.

arch/arm/mach-shmobile/smp-sh73a0.c: A readl vs readw fix collided
with the io accessor cleanup. Ditch the __io() use but keep readl.

arch/arm/mach-tegra/include/mach/debug-macro.S: include of mach/io.h
is going away, iomap.h is kept.

-Olof

[GIT PULL 0/5] Second batch of arm-soc changes

[Rob Herring]

On 03/28/2012 02:13 AM, Olof Johansson wrote:
> Hi Linus,
> 
> Here is the second batch of arm-soc changes for this merge window. As Arnd
> mentioned with the first cover letter, most of these either had external
> dependencies that have now been pulled in, or just needed a bit more
> linux-next bake time due to getting picked up close to the 3.3 release.
> 
> 
> There are 128 non-merge patches, with a total diffstat of:
> 
>  346 files changed, 10790 insertions(+), 2792 deletions(-)
> 
> So, while there's been some good cleanups especially in the first branch,
> there's been a net addition of code:
> 
> * New SoC family support for Samsung EXYNOS5250: clock tables, driver
>   updates, etc means net gain in code.
> * Added device tree conversions. Once we start to sunset non-device tree
>   boards on the enabled platforms, we should see payback in code removal.
> * Common clock framework. As SoCs get converted over to the framework,
>   we should see more code removal from this as well.
> 
> 
> There are a handful of merge conflicts on the branches, nothing that
> should be too messy to sort out. I've described each conflict in the
> corresponding pull request, including how we've resolved them in our
> for-next branch.
> 

Why isn't lakml getting copied on emails 1-5? The same thing happened
with Arnd's pull request. Are they getting hung up for moderator approval?

Rob

[GIT PULL 0/5] Second batch of arm-soc changes

[Arnd Bergmann]

On Wednesday 28 March 2012, Rob Herring wrote:
> On 03/28/2012 02:13 AM, Olof Johansson wrote:
> > There are a handful of merge conflicts on the branches, nothing that
> > should be too messy to sort out. I've described each conflict in the
> > corresponding pull request, including how we've resolved them in our
> > for-next branch.
> > 
> 
> Why isn't lakml getting copied on emails 1-5? The same thing happened
> with Arnd's pull request. Are they getting hung up for moderator approval?

For some reason, the mailing list rejected the emails with 'suspicious header'
warnings, but did not list which headers were suspicious. I think I've
had this problem before and solved it, but don't remember the solution
any more now. linux-kernel took the emails though.

	Arnd

[GIT PULL 4/5] ARM: More device tree support updates

[Linus Torvalds]

On Wed, Mar 28, 2012 at 12:14 AM, Olof Johansson <olof@lixom.net> wrote:
>
> Merge conflicts and proposed solutions:
>
> arch/arm/mach-ux500/Kconfig: Two new config options. Keep both.

Guys, I'm seeing merge problems that go way beyond this.

Look at that MACH_U8500 config option. It got renamed to MACH_MOP500,
but it continues to live in the tree.

This rename conflict happened earlier and I didn't notice (grep for
it: it's used by a staging driver), so it's already in my pushed out
tree. But it seems to be getting worse, and your proposed merge
solution doesn't cover it.

I'll try to fix things up, but you need to verify my merge. And you
guys should have noticed and noted it, rather than leave it to my
random bumbling about.

                     Linus

[GIT PULL 4/5] ARM: More device tree support updates

[Arnd Bergmann]

On Wednesday 28 March 2012, Linus Torvalds wrote:
> On Wed, Mar 28, 2012 at 12:14 AM, Olof Johansson <olof@lixom.net> wrote:
> >
> > Merge conflicts and proposed solutions:
> >
> > arch/arm/mach-ux500/Kconfig: Two new config options. Keep both.
> 
> Guys, I'm seeing merge problems that go way beyond this.
> 
> Look at that MACH_U8500 config option. It got renamed to MACH_MOP500,
> but it continues to live in the tree.
> 
> This rename conflict happened earlier and I didn't notice (grep for
> it: it's used by a staging driver), so it's already in my pushed out
> tree. But it seems to be getting worse, and your proposed merge
> solution doesn't cover it.

Your solution looks good. I had the MACH_UX500_DT option above U5500
instead of below it in our for-next branch, but either way makes sense.

> I'll try to fix things up, but you need to verify my merge. And you
> guys should have noticed and noted it, rather than leave it to my
> random bumbling about.

Yes, noted for next time. We had a lot of merge conflicts because of
being relatively late in the merge window (this one was purely
self made though), and it's not always easy to decide which ones
to eliminate in advance, leave in or describe the resolution we found.

Since we usually see the same conflicts and resolve them in the
for-next branch, I guess it's easy enough to just always provide
the branch with the resolutions as a help for cases like this one.

	Arnd