[GIT BISECT] 2c1756b12edc19fdd75c833699cb752e1bbb641e - Display corruption after blank screen state

[GIT BISECT] 2c1756b12edc19fdd75c833699cb752e1bbb641e - Display corruption after blank screen state

[Emmanuel Benisty]

Hi,

The following commit causes display corruption once the machine goes
out from default X server power saving blank screen state. This is
with an i3 330m.

commit 2c1756b12edc19fdd75c833699cb752e1bbb641e
Merge: cf96e46 d74362c
Author: Keith Packard <keithp@keithp.com>
Date:   Thu Jul 28 16:30:41 2011 -0700

    Merge branch 'drm-intel-fixes' into drm-intel-next

Thanks.

Re: [GIT BISECT] 2c1756b12edc19fdd75c833699cb752e1bbb641e - Display corruption after blank screen state

[Keith Packard]

[-- Attachment #1: Type: text/plain, Size: 991 bytes --]

On Wed, 7 Sep 2011 23:33:13 +0700, Emmanuel Benisty <benisty.e@gmail.com> wrote:

(I'm sorry I missed this mail; I was preoccupied with family issues
that week).

> The following commit causes display corruption once the machine goes
> out from default X server power saving blank screen state. This is
> with an i3 330m.

Can you provide more information about which display is corrupted and
what the corruption looks like?

> commit 2c1756b12edc19fdd75c833699cb752e1bbb641e
> Merge: cf96e46 d74362c
> Author: Keith Packard <keithp@keithp.com>
> Date:   Thu Jul 28 16:30:41 2011 -0700
> 
>     Merge branch 'drm-intel-fixes' into drm-intel-next
> 

Can you provide the output of intel_reg_dumper from both working and
non-working kernels? Reverting this merge would be a fairly substantial
challenge at this point in the release cycle, so I'd like to see if we
can't narrow things down a bit and find the actual code change at fault.

-- 
keith.packard@intel.com

[-- Attachment #2: Type: application/pgp-signature, Size: 827 bytes --]

Re: [GIT BISECT] 2c1756b12edc19fdd75c833699cb752e1bbb641e - Display corruption after blank screen state

[Emmanuel Benisty]

On Fri, Oct 7, 2011 at 10:54 AM, Keith Packard <keithp@keithp.com> wrote:
> On Wed, 7 Sep 2011 23:33:13 +0700, Emmanuel Benisty <benisty.e@gmail.com> wrote:
>
> (I'm sorry I missed this mail; I was preoccupied with family issues
> that week).

And I missed your reply too, very sorry about this.

>> The following commit causes display corruption once the machine goes
>> out from default X server power saving blank screen state. This is
>> with an i3 330m.
>
> Can you provide more information about which display is corrupted and
> what the corruption looks like?
>
>> commit 2c1756b12edc19fdd75c833699cb752e1bbb641e
>> Merge: cf96e46 d74362c
>> Author: Keith Packard <keithp@keithp.com>
>> Date:   Thu Jul 28 16:30:41 2011 -0700
>>
>>     Merge branch 'drm-intel-fixes' into drm-intel-next
>>
>
> Can you provide the output of intel_reg_dumper from both working and
> non-working kernels? Reverting this merge would be a fairly substantial
> challenge at this point in the release cycle, so I'd like to see if we
> can't narrow things down a bit and find the actual code change at fault.

This is now fixed in Linus' tree. Just FTR, adding
i915.i915_enable_fbc=0 solved the issue (prior to mainline fix).

Re: [GIT BISECT] 2c1756b12edc19fdd75c833699cb752e1bbb641e - Display corruption after blank screen state

[Keith Packard]

[-- Attachment #1: Type: text/plain, Size: 397 bytes --]

On Fri, 21 Oct 2011 10:01:57 +0700, Emmanuel Benisty <benisty.e@gmail.com> wrote:

> This is now fixed in Linus' tree. Just FTR, adding
> i915.i915_enable_fbc=0 solved the issue (prior to mainline fix).

Good. We'll try some more FBC changes after 3.1 comes out and see if
those work, but we've always had machines which do weird things when you
enable it.

-- 
keith.packard@intel.com

[-- Attachment #2: Type: application/pgp-signature, Size: 827 bytes --]