mbox

[GIT,PULL] ARM: BCM5301X: SoC changes for v3.19

Message ID 546298D9.7080800@hauke-m.de
State New
Headers show

Pull-request

https://github.com/hauke/linux.git tags/bcm5301x-soc-2014-11-11

Message

Hauke Mehrtens Nov. 11, 2014, 11:16 p.m. UTC
Hi Olof, Hi Arnd,

Please consider these BCM5301X ARM based SoC soc updates for v3.19.

The following changes since commit 206c5f60a3d902bc4b56dab2de3e88de5eb06108:

  Linux 3.18-rc4 (2014-11-09 14:55:29 -0800)

are available in the git repository at:

  https://github.com/hauke/linux.git tags/bcm5301x-soc-2014-11-11

for you to fetch changes up to 6f3a612d5ec2b2fd861c21d3c8ad708ffd533925:

  ARM: BCM5301X: select GPIOLIB as optional (2014-11-11 23:49:01 +0100)

----------------------------------------------------------------
ARM: BCM5301X: SoC: two patches for the SoC part

These patches were all send to the arm list and I haven't got any
negative responses.

----------------------------------------------------------------
Hauke Mehrtens (1):
      ARM: BCM5301X: fix early serial console

Rafał Miłecki (1):
      ARM: BCM5301X: select GPIOLIB as optional

 arch/arm/Kconfig.debug    | 4 ++--
 arch/arm/mach-bcm/Kconfig | 1 +
 2 files changed, 3 insertions(+), 2 deletions(-)

Comments

Arnd Bergmann Nov. 20, 2014, 9:27 a.m. UTC | #1
On Wednesday 12 November 2014, Hauke Mehrtens wrote:
> ----------------------------------------------------------------
> ARM: BCM5301X: SoC: two patches for the SoC part
> 
> These patches were all send to the arm list and I haven't got any
> negative responses.
> 
> ----------------------------------------------------------------
> Hauke Mehrtens (1):
>       ARM: BCM5301X: fix early serial console
> 
> Rafał Miłecki (1):
>       ARM: BCM5301X: select GPIOLIB as optional

Hi Hauke and Florian,

sorry for the late reply. The patches are both good and I want to merge them,
but I have two problems with your pull request:

- I'm unsure about who is sending us patches for mach-bcm. I was under the
  assumption that Florian would now handle all of them and pick up your
  patches to forward them to arm@kernel.org together with the other mach-bcm
  platforms. I'd much rather get only patches from one person per platform
  directory, though we can have multiple people if that's what you all want.

- The pull request is based on -rc4, but I see no strong reason why you did
  that. It helps us a lot of all pull requests in a branch are based off
  the same -rc, so please have them based on -rc1 by default or explain why
  you picked something else in the tag description.

If you want, I can just cherry-pick the two patches from your branch into
the next/soc branch so you don't have to rebase it this time.

	Arnd
Hauke Mehrtens Nov. 20, 2014, 8:45 p.m. UTC | #2
On 11/20/2014 10:27 AM, Arnd Bergmann wrote:
> On Wednesday 12 November 2014, Hauke Mehrtens wrote:
>> ----------------------------------------------------------------
>> ARM: BCM5301X: SoC: two patches for the SoC part
>>
>> These patches were all send to the arm list and I haven't got any
>> negative responses.
>>
>> ----------------------------------------------------------------
>> Hauke Mehrtens (1):
>>       ARM: BCM5301X: fix early serial console
>>
>> Rafał Miłecki (1):
>>       ARM: BCM5301X: select GPIOLIB as optional
> 
> Hi Hauke and Florian,
> 
> sorry for the late reply. The patches are both good and I want to merge them,
> but I have two problems with your pull request:
> 
> - I'm unsure about who is sending us patches for mach-bcm. I was under the
>   assumption that Florian would now handle all of them and pick up your
>   patches to forward them to arm@kernel.org together with the other mach-bcm
>   platforms. I'd much rather get only patches from one person per platform
>   directory, though we can have multiple people if that's what you all want.

I talked to Florian and he suggested me that I should send patches
directly to you. There could be some stuff shared with the other network
SoCs, there is nothing shared with the mobile SoCs.
We thought that it would be easier for us (Florian and me) if you pull
directly from me.

> - The pull request is based on -rc4, but I see no strong reason why you did
>   that. It helps us a lot of all pull requests in a branch are based off
>   the same -rc, so please have them based on -rc1 by default or explain why
>   you picked something else in the tag description.

I just thought that I should base it on a recent -rc, so I took -rc4, I
can rebase it on rc1 if you like, I do not depend on any new stuff.

> If you want, I can just cherry-pick the two patches from your branch into
> the next/soc branch so you don't have to rebase it this time.

That would be nice.

Hauke
Florian Fainelli Nov. 20, 2014, 8:48 p.m. UTC | #3
On 11/20/2014 12:45 PM, Hauke Mehrtens wrote:
> On 11/20/2014 10:27 AM, Arnd Bergmann wrote:
>> On Wednesday 12 November 2014, Hauke Mehrtens wrote:
>>> ----------------------------------------------------------------
>>> ARM: BCM5301X: SoC: two patches for the SoC part
>>>
>>> These patches were all send to the arm list and I haven't got any
>>> negative responses.
>>>
>>> ----------------------------------------------------------------
>>> Hauke Mehrtens (1):
>>>       ARM: BCM5301X: fix early serial console
>>>
>>> Rafał Miłecki (1):
>>>       ARM: BCM5301X: select GPIOLIB as optional
>>
>> Hi Hauke and Florian,
>>
>> sorry for the late reply. The patches are both good and I want to merge them,
>> but I have two problems with your pull request:
>>
>> - I'm unsure about who is sending us patches for mach-bcm. I was under the
>>   assumption that Florian would now handle all of them and pick up your
>>   patches to forward them to arm@kernel.org together with the other mach-bcm
>>   platforms. I'd much rather get only patches from one person per platform
>>   directory, though we can have multiple people if that's what you all want.
> 
> I talked to Florian and he suggested me that I should send patches
> directly to you. There could be some stuff shared with the other network
> SoCs, there is nothing shared with the mobile SoCs.
> We thought that it would be easier for us (Florian and me) if you pull
> directly from me.

Absolutely, I did it differently for the Cygnus SoC to help Scott with
his initial submission, but in the future, I would rather have Hauke,
Scott and myself send separate pull requests, leaving you with just
potential conflicts for arch/arm/mach-bcm/{Makefile,Kconfig}.

If this model is not desirable, I have no problems aggregating all
mach-bcm changes to appropriate pull requests, whether that means
pulling from Hauke and Scott directly, or compiling pull requests myself
as long as you guys CC me on your patches.

Thanks!

> 
>> - The pull request is based on -rc4, but I see no strong reason why you did
>>   that. It helps us a lot of all pull requests in a branch are based off
>>   the same -rc, so please have them based on -rc1 by default or explain why
>>   you picked something else in the tag description.
> 
> I just thought that I should base it on a recent -rc, so I took -rc4, I
> can rebase it on rc1 if you like, I do not depend on any new stuff.
> 
>> If you want, I can just cherry-pick the two patches from your branch into
>> the next/soc branch so you don't have to rebase it this time.
> 
> That would be nice.
> 
> Hauke
>
Arnd Bergmann Nov. 21, 2014, 11:42 a.m. UTC | #4
On Thursday 20 November 2014 21:45:29 Hauke Mehrtens wrote:
> On 11/20/2014 10:27 AM, Arnd Bergmann wrote:
> > - The pull request is based on -rc4, but I see no strong reason why you did
> >   that. It helps us a lot of all pull requests in a branch are based off
> >   the same -rc, so please have them based on -rc1 by default or explain why
> >   you picked something else in the tag description.
> 
> I just thought that I should base it on a recent -rc, so I took -rc4, I
> can rebase it on rc1 if you like, I do not depend on any new stuff.

No, you should base on the oldest -rc that works for you to avoid backmerges.
When I start the next/soc branch from -rc1 and I pull in a branch that
is based on -rc4, 'git request-pull' can no longer figure out what the
base changeset was and I have to manually edit the pull request.

> > If you want, I can just cherry-pick the two patches from your branch into
> > the next/soc branch so you don't have to rebase it this time.
> 
> That would be nice.
> 

I've cherry-picked the first one, but the "BCM5301X: select GPIOLIB as
optional" change no longer applies: as part of Scott's "ARM: mach-bcm:
Consolidate currently supported IPROC SoCs", bcm5301x now selects
ARCH_IPROC, which already uses ARCH_REQUIRE_GPIOLIB.

Let me know if that is a problem.

	Arnd
Arnd Bergmann Nov. 21, 2014, 11:50 a.m. UTC | #5
On Thursday 20 November 2014 12:48:35 Florian Fainelli wrote:
> On 11/20/2014 12:45 PM, Hauke Mehrtens wrote:
> > On 11/20/2014 10:27 AM, Arnd Bergmann wrote:
> >> On Wednesday 12 November 2014, Hauke Mehrtens wrote:
> >>> ----------------------------------------------------------------
> >>> ARM: BCM5301X: SoC: two patches for the SoC part
> >>>
> >>> These patches were all send to the arm list and I haven't got any
> >>> negative responses.
> >>>
> >>> ----------------------------------------------------------------
> >>> Hauke Mehrtens (1):
> >>>       ARM: BCM5301X: fix early serial console
> >>>
> >>> Rafał Miłecki (1):
> >>>       ARM: BCM5301X: select GPIOLIB as optional
> >>
> >> Hi Hauke and Florian,
> >>
> >> sorry for the late reply. The patches are both good and I want to merge them,
> >> but I have two problems with your pull request:
> >>
> >> - I'm unsure about who is sending us patches for mach-bcm. I was under the
> >>   assumption that Florian would now handle all of them and pick up your
> >>   patches to forward them to arm@kernel.org together with the other mach-bcm
> >>   platforms. I'd much rather get only patches from one person per platform
> >>   directory, though we can have multiple people if that's what you all want.
> > 
> > I talked to Florian and he suggested me that I should send patches
> > directly to you. There could be some stuff shared with the other network
> > SoCs, there is nothing shared with the mobile SoCs.
> > We thought that it would be easier for us (Florian and me) if you pull
> > directly from me.
> 
> Absolutely, I did it differently for the Cygnus SoC to help Scott with
> his initial submission, but in the future, I would rather have Hauke,
> Scott and myself send separate pull requests, leaving you with just
> potential conflicts for arch/arm/mach-bcm/{Makefile,Kconfig}.
> 
> If this model is not desirable, I have no problems aggregating all
> mach-bcm changes to appropriate pull requests, whether that means
> pulling from Hauke and Scott directly, or compiling pull requests myself
> as long as you guys CC me on your patches.

My preference is definitely to get fewer pull requests, and to get them
from fewer people. There are so many details that each maintainer needs
to know about to make the arm-soc workflow efficient, I mainly want to
avoid teaching more people about it. I've also discussed this with Lee
Jones yesterday, since he just took over maintenance for bcm2835.

We now have between five and seven platforms (depending how you count)
under mach-bcm that are mostly unrelated, so I can understand the desire
to keep them maintained separately on your end. From my perspective,
these are all 'small' platforms in the sense that I get very few patches
for each one, but they occasionally conflict on Makefile or Kconfig
changes, and I'd appreciate these all to be consolidated in common
pull requests.

	Arnd