[Intel-gfx] inux-next: Tree for Aug 21 (call-trace when suspending: PM?)

[Daniel Vetter]

On Tue, Aug 21, 2012 at 3:24 PM, Sedat Dilek <sedat.dilek at gmail.com> wrote:
> On Tue, Aug 21, 2012 at 1:53 PM, Daniel Vetter <daniel at ffwll.ch> wrote:
>> On Tue, Aug 21, 2012 at 1:14 PM, Sedat Dilek <sedat.dilek at gmail.com> wrote:
>>> On Tue, Aug 21, 2012 at 1:03 PM, Sedat Dilek <sedat.dilek at gmail.com> wrote:
>>>> On Tue, Aug 21, 2012 at 1:02 PM, Sedat Dilek <sedat.dilek at gmail.com> wrote:
>>>>> On Tue, Aug 21, 2012 at 8:04 AM, Stephen Rothwell <sfr at canb.auug.org.au> wrote:
>>>>>> Hi all,
>>>>>>
>>>>>> Changes since 20120820:
>>>>>>
>>>>>> The rr tree gained a conflict against Linus' tree.
>>>>>>
>>>>>> The tip tree still has its build failure so I used the version from
>>>>>> next-20120814.
>>>>>>
>>>>>> The workqueues tree gained a conflict against the hid tree.
>>>>>>
>>>>>> The drivers-x86 tree still has its build failure so I used the version
>>>>>> from next-20120817.
>>>>>>
>>>>>> The signal tree gained a conflict against Linus' tree.  I have still
>>>>>> reverted 3 commits from the signal tree at the request of the arm
>>>>>> maintainer.
>>>>>>
>>>>>> ----------------------------------------------------------------------------
>>>>>>
>>>>>
>>>>> Hi,
>>>>>
>>>>> I have compiled linux-next (next-20120821) and see the attached
>>>>> call-trace when suspending.
>>>>> Suspending did NOT work (Xorg seems to cause it) - machine came back to desktop.
>>>>>
>>>>> With yesterday's next-20120820 I haven't seen this.
>>>>>
>>>>> I am not sure what is this causing... PM, x86/sched or even VFS?
>>>>> Any help for debugging appreciated.
>>>>>
>>>>> I am on Ubuntu/precise AMD64 and use systemd-v43 as init-system.
>>>>>
>>>>> Regards,
>>>>> - Sedat -
>>>>
>>>> Forgot attachment!
>>>> If you don't succeed - try try try...
>>>>
>>>> - Sedat -
>>>
>>> [ CC danvet ]
>>>
>>> I have pulled in drm-intel-fixes into my local GIT tree and rebuilt
>>> i915 - this seems to fix the problem.
>>> Daniel any suggestion which patch in d-i-f did it?
>>
>> Without the backtrace it's kinda hard to tell ... Also, if you can
>> dump a git log of the commits from -fixes that you don't yet have.
>> -Daniel
>>
>
> Hi Daniel,
>
> $ git log --oneline v3.6.0-rc2-next20120821-1-iniza-generic..drm-intel-fixes
> 1ee9ae3 drm/i915: use hsw rps tuning values everywhere on gen6+
> f1a2f5b drm/i915: fall back to bit-banging if GMBUS fails in CRT EDID reads
> 4eab813 drm/i915: extract connector update from intel_ddc_get_modes() for reuse
> a843af1 drm/i915: fix hsw uncached pte
> b6c7488 drm/i915/contexts: fix list corruption
> 38ab8a2 drm/i915: fix EDID memory leak in SDVO
>
> Looks like "1ee9ae3 drm/i915: use hsw rps tuning values everywhere on
> gen6+" has PM fixes for i915.
>
> I tried with only that patch on top of today's linux-next - and I can
> suspend/resume again!

Well, this is slightly shocking - this patch /should/ only optimize
power consumption, it should in now way fix suspend/resume (i.e. it
doesn't even apply any h/w workaround). Do you have any more details
what's going wrong here (logs, ...)?

Thanks, Daniel
-- 
Daniel Vetter
daniel.vetter at ffwll.ch - +41 (0) 79 364 57 48 - http://blog.ffwll.ch