2014-06-18 16:14:59

by Theodore Ts'o

[permalink] [raw]
Subject: Display glitching (and possibly related lockup) in 3.16-rc1?

Is this a known problem?

On my T540p, occasionally the display will go black and then be
completely locked. One time when this happened, the display started
glitching and tearing (like what you might see if an analog display had
a loose connector and it was being wiggled) when the system locked up,
but the at least 75% of the time when the bug manifests, the screen goes
black instead.

This appears to only happen when the system is running on battery --- it
doesn't seem to manifest if I'm plugged into the AC mains. And I've
been typing an e-mail message when the system locks up, so I don't think
it's related to DPMS or the screen saver kicking in.

I can try to bisect this regression (3.15 didn't have this problem), but
I thought I would ask first if this was a known problem. I don't know
for certain that it's an i915 bug, but the symptoms do somewhat point in
that direction, especially the display glitches.

Thanks!!

- Ted


2014-06-19 16:09:55

by Daniel Vetter

[permalink] [raw]
Subject: Re: [Intel-gfx] Display glitching (and possibly related lockup) in 3.16-rc1?

On Wed, Jun 18, 2014 at 12:14:55PM -0400, Theodore Ts'o wrote:
> Is this a known problem?
>
> On my T540p, occasionally the display will go black and then be
> completely locked. One time when this happened, the display started
> glitching and tearing (like what you might see if an analog display had
> a loose connector and it was being wiggled) when the system locked up,
> but the at least 75% of the time when the bug manifests, the screen goes
> black instead.
>
> This appears to only happen when the system is running on battery --- it
> doesn't seem to manifest if I'm plugged into the AC mains. And I've
> been typing an e-mail message when the system locks up, so I don't think
> it's related to DPMS or the screen saver kicking in.
>
> I can try to bisect this regression (3.15 didn't have this problem), but
> I thought I would ask first if this was a known problem. I don't know
> for certain that it's an i915 bug, but the symptoms do somewhat point in
> that direction, especially the display glitches.

Doesn't ring a bell, but can you please boot with drm.debug=0xe and attach
dmesg here so we know what kind of hardware you have? Also any dying
breadths over netconsole or similar? Can't hurt to check that before doing
the bisect.

Thanks, Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch

2014-06-19 21:59:09

by Dave Airlie

[permalink] [raw]
Subject: Re: [Intel-gfx] Display glitching (and possibly related lockup) in 3.16-rc1?

On 20 June 2014 02:09, Daniel Vetter <[email protected]> wrote:
> On Wed, Jun 18, 2014 at 12:14:55PM -0400, Theodore Ts'o wrote:
>> Is this a known problem?
>>
>> On my T540p, occasionally the display will go black and then be
>> completely locked. One time when this happened, the display started
>> glitching and tearing (like what you might see if an analog display had
>> a loose connector and it was being wiggled) when the system locked up,
>> but the at least 75% of the time when the bug manifests, the screen goes
>> black instead.
>>
>> This appears to only happen when the system is running on battery --- it
>> doesn't seem to manifest if I'm plugged into the AC mains. And I've
>> been typing an e-mail message when the system locks up, so I don't think
>> it's related to DPMS or the screen saver kicking in.
>>
>> I can try to bisect this regression (3.15 didn't have this problem), but
>> I thought I would ask first if this was a known problem. I don't know
>> for certain that it's an i915 bug, but the symptoms do somewhat point in
>> that direction, especially the display glitches.
>
> Doesn't ring a bell, but can you please boot with drm.debug=0xe and attach
> dmesg here so we know what kind of hardware you have? Also any dying
> breadths over netconsole or similar? Can't hurt to check that before doing
> the bisect.
>

t540p are haswell.

Dave.