Internal error - gpu hung

[habana]

I received some updates yesterday and among them was some type of fix for the gpu false hang. But when I read through the reports and documentation, they seem to consider it an annoying false alert. When your system is frozen or stops for 5 to 10 seconds at a time multiple times an hour, it does not feel like a false alert. I am relatively new to Linux and might be misinterpreting it. I am going to wait a little bit. I would really appreciate it if you gave us an update in a week to see if you were still past this problem.

I understand there are two separate bugs. One is the the false crash report which has been fixed by a workaround. The other, which you, I and others are/were experiencing, has not yet been fixed.

If I remember (!), I will report back next week to confirm all OK in 13.04

[habana]

If I remember (!), I will report back next week to confirm all OK in 13.04

Not quite a week but I am off on holiday for a few days. I can confirm this bug is not present on my machine in 13.04

[VisualThoy]

Not quite a week but I am off on holiday for a few days. I can confirm this bug is not present on my machine in 13.04

I'm using 13.04 and had not had this crash until the upgrade. Now it happens nearly every day.

[hongquan1987]

I'm using 13.04 and had not had this crash until the upgrade. Now it happens nearly every day.

Me too

[ryanvade]

13.04, latest intel graphics from x-swats, kernel 3.10.1, bumblebee 3.2.1, and nvidia 319. Still having this issue.

[Bashing-om]

et all;

I have not checked lately with launchpad on an updated status, will do so on my morrow.

we be look'n

[Bashing-om]

Guys;

Near as I can tell the old bugs referring to;
"drm:i915_hangcheck_hung" have been closed out, as fixed in upstream. This from the lucid era in respect to the i915 driver for Intel hardware.

Maybe a new bug report should be opened, as it looks as if a number of people are affected (again) ?

just a thought

[Pawka]

13.04, Linux yoda 3.8.0-29-generic. Still the same issue here.

[Hamish.MacEwan]

Hi,

Have the same problem, fortunately it only impairs the graphic X interface. I can ctrl-alt-Fx and login to a text screen where dmesg reports:

[11527.486513] [drm:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung
[11527.486517] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state
[11527.886179] openshot[23160]: segfault at 7f5549ab005c ip 00007f5549ab005c sp 00007f5529d22dc0 error 14 in libmltvorbis.so[7f5549cb6000+3000]

There's nothing there I'm competent to recognise, but the machine isn't hung, but X is.

Attached is the output of sudo cat /sys/kernel/debug/dri/0/i915_error_state. Well it would be, the 307KB zip seems to break some rule, so I can't sorry.

Update: Went back to X, not surprisingly OpenShot was gone. Alt-K-SysRq got me a screenshot and a restart of X. Logged in and all good.

[zazuge]

I have the same problem I won't repost the same dmesg and syslog error lines
I have tried to corelate the bug with what upgraded package
for me it only started in Oct 21 2013 (according to syslog) never seen this bug before but i think it happened less frequently even before but i have no trace of that (oldest syslog is Oct 19)
there 2 upgrades that are related to this
on in 19Oct upgraded:
procps (3.3.3-2ubuntu5.1,3.3.3-2ubuntu5.2)
xserver-common (1.13.3-0ubuntu6, 1.13.3-0ubuntu6.2)
xserver-xorg-core (1.13.3-0ubuntu6, 1.13.3-0ubuntu6.2)
libprocps0 (3.3.3-2ubuntu5.1, 3.3.3-ubuntu5.2)
xserver-xorg-intel (2.21.6-0ubuntu4.1, 2.21.6-0ubuntu4.3)
the other on 23 Oct:
Linux-image-3.8.0-32-generic (3.8.0-32.47, automatic)
.... etc
these two are th most relevant to the bug
I suspect the xserver-xorg-intel package is the culprit
but the new kernel made things worse (GPU hangs happen more frequently) and when i returned to the older kernel to this point it didn't happen again