[GIT PULL] More power management updates for 3.4

[GIT PULL] More power management updates for 3.4

[Rafael J. Wysocki]

Hi Linus,

I prepared this on Friday with the hope to send the pull request before the end
of the merge window, but then I had to go to San Francisco (for the LFCS) and
missed the opportunity.  That said, this includes fixes mostly (and -stable
ones for that matter), so I think it's appropriate for the -rc1 time frame too. 

Please pull additional power management updates for 3.4 since commit
735e941caa9a35f933297af0ab1e0ad6447411c4:

    Merge tag 'common-clk-api' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc

with top-most commit c4772d192c70b61d52262b0db76f7abd8aeb51c6:

    PM / QoS: add pm_qos_update_request_timeout() API

from the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git pm-for-3.4-part-2

This mostly consists of fixes and changes requisite for those fixes, with the
exception of the last commit, which is a new feature.  However, that new
feature is reasonably isolated and very unlikely to cause any regressions to
happen, although it is necessary for some future changes in the works now.
Included are:

* Patch series that hopefully fixes races between the freezer and request_firmware()
  and request_firmware_nowait() for good, with two cleanups from Stephen Boyd on top.

* Runtime PM fix from Alan Stern preventing tasks from getting stuck indefinitely
  in the runtime PM wait queue.

* Device PM QoS update from MyungJoo Ham introducing a new variant of
  pm_qos_update_request() allowing the callers to specify a timeout.

Thanks!


 drivers/base/firmware_class.c |  208 +++++++++++++++++++++--------------------
 drivers/base/power/runtime.c  |    3 +-
 include/linux/kmod.h          |   27 +++++-
 include/linux/pm_qos.h        |    4 +
 kernel/kmod.c                 |  117 +++++++++++++++++------
 kernel/power/hibernate.c      |   18 +---
 kernel/power/process.c        |    8 ++
 kernel/power/qos.c            |   50 ++++++++++
 kernel/power/suspend.c        |    7 --
 kernel/power/user.c           |   10 +--
 10 files changed, 284 insertions(+), 168 deletions(-)

---------------

Alan Stern (1):
      PM / Runtime: don't forget to wake up waitqueue on failure

MyungJoo Ham (1):
      PM / QoS: add pm_qos_update_request_timeout() API

Rafael J. Wysocki (6):
      firmware_class: Rework usermodehelper check
      firmware_class: Split _request_firmware() into three functions, v2
      firmware_class: Do not warn that system is not ready from async loads
      PM / Hibernate: Disable usermode helpers right before freezing tasks
      PM / Sleep: Move disabling of usermode helpers to the freezer
      PM / Sleep: Mitigate race between the freezer and request_firmware()

Stephen Boyd (2):
      firmware_class: Reorganize fw_create_instance()
      firmware_class: Move request_firmware_nowait() to workqueues

Re: [GIT PULL] More power management updates for 3.4

[Rafael J. Wysocki]

Hi,

On Sunday, April 01, 2012, Rafael J. Wysocki wrote:
> Hi Linus,
> 
> I prepared this on Friday with the hope to send the pull request before the end
> of the merge window, but then I had to go to San Francisco (for the LFCS) and
> missed the opportunity.  That said, this includes fixes mostly (and -stable
> ones for that matter), so I think it's appropriate for the -rc1 time frame too. 
> 
> Please pull additional power management updates for 3.4 since commit
> 735e941caa9a35f933297af0ab1e0ad6447411c4:
> 
>     Merge tag 'common-clk-api' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
> 
> with top-most commit c4772d192c70b61d52262b0db76f7abd8aeb51c6:
> 
>     PM / QoS: add pm_qos_update_request_timeout() API
> 
> from the git repository at:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git pm-for-3.4-part-2

Given the lack of feedback and the fact that you don't seem to have pulled this,
I wonder if there's anything wrong with the changes or am I simply supposed to
queue them up for 3.5?

Or would you like me to prepare another pull request without the top-most
commit that doesn't count as a fix (I rather won't be able to prepare it
before -rc2 is out, though, due to travel-related limitations)?

Rafael


> This mostly consists of fixes and changes requisite for those fixes, with the
> exception of the last commit, which is a new feature.  However, that new
> feature is reasonably isolated and very unlikely to cause any regressions to
> happen, although it is necessary for some future changes in the works now.
> Included are:
> 
> * Patch series that hopefully fixes races between the freezer and request_firmware()
>   and request_firmware_nowait() for good, with two cleanups from Stephen Boyd on top.
> 
> * Runtime PM fix from Alan Stern preventing tasks from getting stuck indefinitely
>   in the runtime PM wait queue.
> 
> * Device PM QoS update from MyungJoo Ham introducing a new variant of
>   pm_qos_update_request() allowing the callers to specify a timeout.
> 
> Thanks!
> 
> 
>  drivers/base/firmware_class.c |  208 +++++++++++++++++++++--------------------
>  drivers/base/power/runtime.c  |    3 +-
>  include/linux/kmod.h          |   27 +++++-
>  include/linux/pm_qos.h        |    4 +
>  kernel/kmod.c                 |  117 +++++++++++++++++------
>  kernel/power/hibernate.c      |   18 +---
>  kernel/power/process.c        |    8 ++
>  kernel/power/qos.c            |   50 ++++++++++
>  kernel/power/suspend.c        |    7 --
>  kernel/power/user.c           |   10 +--
>  10 files changed, 284 insertions(+), 168 deletions(-)
> 
> ---------------
> 
> Alan Stern (1):
>       PM / Runtime: don't forget to wake up waitqueue on failure
> 
> MyungJoo Ham (1):
>       PM / QoS: add pm_qos_update_request_timeout() API
> 
> Rafael J. Wysocki (6):
>       firmware_class: Rework usermodehelper check
>       firmware_class: Split _request_firmware() into three functions, v2
>       firmware_class: Do not warn that system is not ready from async loads
>       PM / Hibernate: Disable usermode helpers right before freezing tasks
>       PM / Sleep: Move disabling of usermode helpers to the freezer
>       PM / Sleep: Mitigate race between the freezer and request_firmware()
> 
> Stephen Boyd (2):
>       firmware_class: Reorganize fw_create_instance()
>       firmware_class: Move request_firmware_nowait() to workqueues
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 

Re: [GIT PULL] More power management updates for 3.4

[Linus Torvalds]

On Wed, Apr 4, 2012 at 2:26 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
>
> Given the lack of feedback and the fact that you don't seem to have pulled this,
> I wonder if there's anything wrong with the changes or am I simply supposed to
> queue them up for 3.5?

Actually, I was just a bit unhappy with it, and decided I wanted to
give it a second look, and then it just fell in between the cracks.

I still don't like the timing of that thing, but I saw no major reason
not to pull. So in it went,

                  Linus

Re: [GIT PULL] More power management updates for 3.4

[Rafael J. Wysocki]

On Wednesday, April 04, 2012, Linus Torvalds wrote:
> On Wed, Apr 4, 2012 at 2:26 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> >
> > Given the lack of feedback and the fact that you don't seem to have pulled this,
> > I wonder if there's anything wrong with the changes or am I simply supposed to
> > queue them up for 3.5?
> 
> Actually, I was just a bit unhappy with it, and decided I wanted to
> give it a second look, and then it just fell in between the cracks.
> 
> I still don't like the timing of that thing, but I saw no major reason
> not to pull. So in it went,

Thanks a lot!

Sorry about the timing, I'll do my best not to repeat that in the future.

Rafael