Bug 16262 - [945GM] intermittent crashes
Summary: [945GM] intermittent crashes
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: 7.3 (2007.09)
Hardware: x86 (IA32) Linux (All)
: medium critical
Assignee: Jesse Barnes
QA Contact: Xorg Project Team
Keywords: NEEDINFO
Depends on:
Reported: 2008-06-06 16:56 UTC by celejar
Modified: 2009-05-13 13:32 UTC (History)
2 users (show)

See Also:
i915 platform:
i915 features:

X server log from crashed session (37.49 KB, text/plain)
2008-06-06 16:56 UTC, celejar
no flags Details
X config (2.59 KB, text/plain)
2008-06-06 16:57 UTC, celejar
no flags Details
dmesg (7.90 KB, text/plain)
2008-06-11 17:06 UTC, celejar
no flags Details
X server log from crashed session (70.96 KB, text/plain)
2009-03-19 12:19 UTC, celejar
no flags Details

Description celejar 2008-06-06 16:56:01 UTC
Created attachment 16976 [details]
X server log from crashed session

X crashes, not constantly, I'd say every few days.  I have not noticed any triggering condition or pattern.  I have seen bug #14709; I have no idea if this is related.  Enclosed is a copy of a log from a crashed session.
Comment 1 celejar 2008-06-06 16:57:24 UTC
Created attachment 16977 [details]
X config
Comment 2 Gordon Jin 2008-06-10 18:47:40 UTC
The logs says:
(EE) AIGLX error: dlopen of /usr/lib/dri/i915_dri.so failed (/usr/lib/dri/i915_dri.so: cannot open shared object file: No such file or directory)
(EE) AIGLX: reverting to software rendering

Can you check /usr/lib/dri/i915_dri.so ?
Comment 3 Gordon Jin 2008-06-10 18:51:25 UTC
And please provide more info according to http://www.intellinuxgraphics.org/how_to_report_bug.html, like the detailed driver component version.

Are you running any 3d apps? apps using render?
Does exanocomposite get rid of the crashes?
Comment 4 celejar 2008-06-11 17:06:10 UTC
Created attachment 17063 [details]

A current dmesg; I snipped a bunch of wireless networking lines from the end.
Comment 5 celejar 2008-06-11 17:25:49 UTC
This is an Acer Aspire 3690-2672 laptop.  I don't know exactly how to determine the MB, but lspci shows a lot of '82801G (ICH7 Family)' lines.

I'm not exactly sure how to determine the chipset; from lspci:

00:02.0 VGA compatible controller: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller (rev 03)

I'm running an i686 environment; processor is an Intel Celeron M 420.
Debian Sid (unstable).  xserver-xorg-core is version 2:1.4.1~git20080517-1; xserver-xorg-video-intel is at 2:2.3.1-1.  Kernel is currently 2.6.25 (self-built from Debian source packages).

There is no /usr/lib/dri/ directory on this system.

I am not running any 3D apps.  I don't know what render is, so I assume I am not using it :)

I'll look into exanocomposite, but as I said in the original report, the crashes are infrequent to begin with, so I won't have any easy way of determining whether the problem persists.

Thanks for the help; I'm sorry about the lack of information in the original report.  If there's anything further I should provide, please let me know.
Comment 6 Gordon Jin 2008-06-17 19:25:05 UTC
I'm not sure what we can for this. Re-assign to Jesse to see if he has better idea.
Comment 7 Jesse Barnes 2008-07-17 12:55:27 UTC
Have you been able to reproduce this after setting the ExaNoComposite option in your xorg.conf?  Or maybe it's gone away with a more recent version of the driver (I can hope can't I? :).
Comment 8 celejar 2008-07-20 17:33:40 UTC
I don't think I've seen any crashes since I reported the bug, although I have not set ExaNoComposite.  Perhaps it is indeed gone.

Note, though, that a version of this bug was just reported on the Debian User mailing list, and the reporter said that he reported it to the Debian BTS.  I suppose they'll be forwarding it to you, but here's the discussion on d-u:


Comment 9 Jesse Barnes 2008-07-21 11:46:44 UTC
Ok, thanks.  We'll track the other crash separately.
Comment 10 celejar 2009-03-19 12:19:35 UTC
Created attachment 24050 [details]
X server log from crashed session
Comment 11 celejar 2009-03-19 12:23:24 UTC
Hi, Jesse, Gordon, and anyone else following this,

I've been seeing the crash again.  I've added an X server log from a recent crash, in case there's anything interesting there that wasn't in the original log.

Note that X won't start up properly after a crash; the screen goes blank, but X doesn't come up.  I can return to VT1, or kill X.  I've tried searching the process list to see if any X stuff was left running, but I didn't recognize anything.  Is there anything specific I should look for?

I'll of course be glad to provide any additional information that will be helpful.
Comment 12 Gordon Jin 2009-04-15 01:20:52 UTC
(In reply to comment #11)
> Hi, Jesse, Gordon, and anyone else following this,
> I've been seeing the crash again.  I've added an X server log from a recent
> crash, in case there's anything interesting there that wasn't in the original
> log.

Have you set ExaNoComposite option?

If this only happens without ExaNoComposite, I'd suggest to mark this bug dup with bug#17638.
Comment 13 Jesse Barnes 2009-05-11 11:21:27 UTC
Adjusting severity: crashes & hangs should be marked critical.
Comment 14 Eric Anholt 2009-05-12 16:55:42 UTC
Does this occur with the current (linus master of kernel, 2D driver 2.7.0 or newer) code?  If so, please attach the output of intel_gpu_dump along with Xorg.0.log and dmesg, so we can take a closer look at what we're doing to the hardware that's making it so angry.

Note, that 2.6.30-rc kernel is required to run intel_gpu_dump and get the info we need.  Otherwise, should be fine.
Comment 15 celejar 2009-05-13 13:21:01 UTC
I have not seen the crash since my last message of Mar. 19, even though I had never gotten around to setting ExaNoComposite.  I'm currently running 2.6.30-rc2, built from the wireless-testing git repository.
Comment 16 Jesse Barnes 2009-05-13 13:32:37 UTC
That's quite awhile... I'll tentatively mark this as resolved.  Please re-open with the info Eric requested if you run into trouble again (or file new bugs for other issues).  Thanks.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.