[Intel-gfx] 3.11.0+ Laptop screen goes blank during kernelboot

[Hans de Bruin]

On 09/13/2013 09:36 PM, Daniel Vetter wrote:
> On Fri, Sep 13, 2013 at 7:43 PM, Hans de Bruin <jmdebruin at xmsnet.nl> wrote:
>> On 09/12/2013 10:49 PM, Daniel Vetter wrote:
>>>
>>> On Thu, Sep 12, 2013 at 10:45:10PM +0200, Hans de Bruin wrote:
>>>>
>>>> On 09/10/2013 01:00 PM, Jani Nikula wrote:
>>>>>
>>>>>
>>>>> Hi Hans -
>>>>>
>>>>> On Sat, 07 Sep 2013, Hans de Bruin <jmdebruin at xmsnet.nl> wrote:
>>>>>>
>>>>>> Since a day or two the screen of my laptop goes blank when the kernel
>>>>>> switches resolution in the boot process. Happily the x-server turns it
>>>>>> on again.
>>>>>
>>>>>
>>>>> Please try the drm-intel-nightly branch of
>>>>> git://people.freedesktop.org/~danvet/drm-intel
>>>>
>>>>
>>>> Well the drm-intel-testing branch works fine but that was not the
>>>> question. lets see if I am able to use gits -b option and clone
>>>> drm-intel-nightly.
>>>
>>>
>>> -testing should be recent enough for the commit Jani mentioned. It's also
>>> already in Linus' git, so hopefully that works again for you.
>>
>>
>> drm-intel-nightly Works fine to (the compile was already running when  I
>> read  your mail). Today's Linus tree still has the defect. Time for a
>> bisect?
>
> Hm, so either it's a different feature patch in -nightly which -fixes
> this, or much more likely a regression somewhere in Linus' git tree. I
> think a bisect is asked for here indeed.
> -Daniel
>

This is the one:


137b944e100278d696826cf25c83014ac17473fe is the first bad commit
commit 137b944e100278d696826cf25c83014ac17473fe
Author: Daniel Lezcano <daniel.lezcano at linaro.org>
Date:   Wed Jun 12 15:08:48 2013 +0200

   cpuidle: Make it clear that governors cannot be modules

   cpufreq governors are defined as modules in the code, but the Kconfig
   options do not allow them to be built as modules.  This is not really
   a problem, but the cpuidle init ordering is: the cpuidle init
   functions (framework and driver) and then the governors.  That leads
   to some weirdness in the cpuidle framework.

   Namely,  cpuidle_register_device() calls cpuidle_enable_device() which
   fails at the first attempt, because governors have not been registered
   yet.  When a governor is registered, the framework calls
   cpuidle_enable_device() again which runs __cpuidle_register_device()
   only then.  Of course, for that to work, the cpuidle_enable_device()
   return value has to be ignored by cpuidle_register_device().

     Instead of having this cyclic call graph and relying on a positive
     side effects of the hackish back and forth cpuidle_enable_device()
     calls it is better to fix the cpuidle init ordering.

     To that end, replace the module init code with postcore_initcall()
     so we have:

      * cpuidle framework : core_initcall
      * cpuidle governors : postcore_initcall
      * cpuidle drivers   : device_initcall

     and remove the corresponding module exit code as it is dead anyway
     (governors can't be built as modules).

     [rjw: Changelog]
     Signed-off-by: Daniel Lezcano <daniel.lezcano at linaro.org>
     Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki at intel.com>

:040000 040000 45e7cb28411864c00b62296a7ff239f9f139d5d9 
86a36b01dee52a77cfbab41a973b7a2bfa44ce34 M      drivers



Reverting it makes the problem go away.

The pull request from Rafael had cc's to linux-acpi at vger.kernel.org 
linux-pci at vger.kernel.org.

-- 
Hans