[Intel-gfx] [4.4-rc1][Regression] drm/i915: Check live status before reading edid

[Joseph Salisbury]

On 03/02/2016 04:58 PM, Joseph Salisbury wrote:
> On 02/29/2016 04:33 AM, Jani Nikula wrote:
>> On Wed, 24 Feb 2016, Joseph Salisbury <joseph.salisbury at canonical.com> wrote:
>>> Hi Sonika,
>>>
>>> A kernel bug report was opened against Ubuntu [0].  After a kernel
>>> bisect, it was found that reverting the following commit resolved this bug:
>>>
>>> commit 237ed86c693d8a8e4db476976aeb30df4deac74b
>>> Author: Sonika Jindal <sonika.jindal at intel.com>
>>> Date:   Tue Sep 15 09:44:20 2015 +0530
>>>
>>>     drm/i915: Check live status before reading edid
>>>
>>>
>>>
>>> The regression was introduced as of v4.4-rc1.
>>>
>>> I was hoping to get your feedback, since you are the patch author.  Do
>>> think increasing the number of tries in intel_hdmi_detect() is worth
>>> trying?  Do you think gathering any additional data will help diagnose
>>> this issue, or would it be best to submit a revert request?
> Thanks for the info.  I will have all of these commits tested.
>
>
>> There are at least the following commits claiming to fix issues in the
>> above commit. Please make sure you have them.
>>
>> BR,
>> Jani.
>>
>>
>> commit 8d409cb3e8a24196be7271defafd4638f3e0b514
>> Author: Ville Syrjälä <ville.syrjala at linux.intel.com>
>> Date:   Wed Feb 10 19:59:05 2016 +0200
>>
>>     drm/i915: Fix hpd live status bits for g4x
>>
>> commit 3d8acd1f667b45c531401c8f0c2033072e32a05d
>> Author: Gary Wang <gary.c.wang at intel.com>
>> Date:   Wed Dec 23 16:11:35 2015 +0800
>>
>>     drm/i915: increase the tries for HDMI hotplug live status checking
>>
>> commit 97f9010af05c15e0b7e6b4ef6ff8cb0ebb7e7715
>> Author: Daniel Vetter <daniel.vetter at ffwll.ch>
>> Date:   Fri Dec 11 19:44:15 2015 +0100
>>
>>     drm/i915: mdelay(10) considered harmful
>>
>> commit 0f5a9be15797f78c9a34e432f26c796165b6e49a
>> Author: Imre Deak <imre.deak at intel.com>
>> Date:   Fri Nov 27 18:55:29 2015 +0200
>>
>>     drm/i915: take a power domain reference while checking the HDMI live status
>>
>>
>>
>>
Hi Jani,

Applying the following commit did indeed fix the original bug[0]:

commit 8d409cb3e8a24196be7271defafd4638f3e0b514
Author: Ville Syrjälä <ville.syrjala at linux.intel.com>
Date:   Wed Feb 10 19:59:05 2016 +0200

    drm/i915: Fix hpd live status bits for g4x



However, it also introduced a new bug, which is covered in that bug
report.  The new bug is that after locking/unlocking the screen, all the
windows get shuffled around in a manner that is consistent with X
deciding that it's single-headed again, and then back to double headed. 
This only happens when I build a kernel with both commit 237ed86c AND
commit 8d409cb3e.  If I revert 237ed86c and keep only 8d409cb3e both the
original bug and the new bug go away.  Do you think commit 237ed86c is
still even needed now that 8d409cb3e has landed?  Maybe the new bug is
due to the interaction between 237ed86c and 8d409cb3e.

Thanks,

Joe

[0] http://pad.lv/1543683