Bug 47512 - Xorg crash while shutting down lightdm
Summary: Xorg crash while shutting down lightdm
Status: RESOLVED DUPLICATE of bug 47519
Alias: None
Product: xorg
Classification: Unclassified
Component: Server/General (show other bugs)
Version: 7.6 (2010.12)
Hardware: x86-64 (AMD64) Linux (All)
: high critical
Assignee: Xorg Project Team
QA Contact: Xorg Project Team
URL:
Whiteboard: 2012BRB_Reviewed
Keywords:
Depends on:
Blocks:
 
Reported: 2012-03-19 08:01 UTC by Torsten Krah
Modified: 2012-06-14 01:16 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Backtrace. (15.01 KB, text/plain)
2012-03-19 08:01 UTC, Torsten Krah
no flags Details
Xorg.log (71.18 KB, text/plain)
2012-03-19 08:02 UTC, Torsten Krah
no flags Details

Description Torsten Krah 2012-03-19 08:01:10 UTC
Created attachment 58686 [details]
Backtrace.

Hi,

environment is Ubuntu Oneiric and latest Xorg stuff.
Attached is Xorg.0.log and a backtrace - X does hang very often and mouse etc. is moving, but thats it - have to kill and restart X (vt switch does work) - to use it again (it seems to happen with gnome3 only, e.g. using lxde i have not seen this one yet - don't know if this is relevant or not).
Comment 1 Torsten Krah 2012-03-19 08:02:42 UTC
Created attachment 58687 [details]
Xorg.log
Comment 2 Alan Coopersmith 2012-03-19 08:08:59 UTC
The log indicates a segmentation fault/crash not a hang.
Did you attach the wrong long?

BTW, the X server sitting around doing nothing in a select() call from
WaitForSomething() is normal - that's the main X server event loop, 
waiting for select() to indicate there's input from a device or a client
to process, or for a timer to expire.
Comment 3 Torsten Krah 2012-03-19 08:50:02 UTC
The crash did happen on calling stop on the lightdm service which wanted to shutdown X. It was not there at "hang" time - so its the correct log.

Hm ok, if its the main loop - why i can "click" e.g. at my desktop (gnome shell) and nothing does happen -> don't know yet where to look for why its hanging otherwise.
Comment 4 Jeremy Huddleston Sequoia 2012-06-12 04:07:56 UTC
The provided backtrace is not useful.  Please see 
http://www.x.org/wiki/Development/Documentation/ServerDebugging

Is this a regression?
Comment 5 Torsten Krah 2012-06-12 04:50:28 UTC
This one is bypassed by me using the workaround from 

https://bugs.freedesktop.org/show_bug.cgi?id=47519

radeon.msi=0 does fix those hangs.

Enabling MSI again and the hangs are there again.

Why is the backtrace not useful, what do you miss exactly?
Did read the debugging guide, installed the symbols and took registers and thread dump - it does not crash, it does only "hang" at the main event loop forever - what else should be captured?

Don't know if this is a regression, used LXDE/gnome2 before which does not use the gnome 3 graphics features which does exhibit the msi problem, so it maybe there before.
Comment 6 Jeremy Huddleston Sequoia 2012-06-12 09:15:31 UTC
(In reply to comment #5)
> This one is bypassed by me using the workaround from 
> 
> https://bugs.freedesktop.org/show_bug.cgi?id=47519
> 
> radeon.msi=0 does fix those hangs.
> 
> Enabling MSI again and the hangs are there again.

Ok, then that aspect is a dupe of #47519

> Why is the backtrace not useful, what do you miss exactly?
> Did read the debugging guide, installed the symbols and took registers and
> thread dump - it does not crash, it does only "hang" at the main event loop
> forever - what else should be captured?

As Alan mentioned, the log you attached was of a crash, not a hang.  Perhaps you grabbed the wrong log file.  It has only addresses, no symbols, so it's not useful.

> Don't know if this is a regression, used LXDE/gnome2 before which does not use
> the gnome 3 graphics features which does exhibit the msi problem, so it maybe
> there before.

Ok, thanks.

*** This bug has been marked as a duplicate of bug 47519 ***
Comment 7 Torsten Krah 2012-06-13 00:47:49 UTC
Ah you meant Xorg.0.log - not like i said - the backtrace log attached is the LOG about the hang.
The Xorg.0.log is only attached for info purpose and the crash there did "happen" after i did want to stop the X server.
The trace was taken "before" those messages appeared, so thats all i can provide.
Comment 8 Torsten Krah 2012-06-13 00:49:13 UTC
About symbols - this is the trace from "backtrace" attachment:

Thread 1 (Thread 0x7fa65012e8a0 (LWP 1221)):
#0  0x00007fa64e4835f3 in __select_nocancel () at ../sysdeps/unix/syscall-template.S:82
#1  0x000000000045dffb in WaitForSomething (pClientsReady=0x212e620) at ../../os/WaitFor.c:232
#2  0x000000000042f912 in Dispatch () at ../../dix/dispatch.c:367
#3  0x00000000004232fe in main (argc=10, argv=<optimized out>, envp=<optimized out>) at ../../dix/main.c:287


Symbols are there, what do you miss here?
Comment 9 Jeremy Huddleston Sequoia 2012-06-13 17:32:17 UTC
(In reply to comment #8)
> 
> Symbols are there, what do you miss here?

I was requesting the symbolicated crash.  The backtrace you provided was from the hang (#47519) and not the server crash.
Comment 10 Torsten Krah 2012-06-14 00:27:00 UTC
Ah now i got it, sorry :-)
Unfortunately i did see the crash in the logs after i did a reboot and so i did not take/hava a trace for this case, sorry.
I'll will provide one if if can make this happen again.
Comment 11 Jeremy Huddleston Sequoia 2012-06-14 01:16:47 UTC
Ok, no worries.  I just cringe at server crashes, so if you can reproduce it, please report it.  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.