Bug 206882 - gdm - segfault at 0000000000000010
Summary: gdm - segfault at 0000000000000010
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: gdm
Version: rawhide
Hardware: x86_64
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Ray Strode [halfline]
QA Contact: Mike McLean
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2006-09-17 18:16 UTC by Michal Jaegermann
Modified: 2007-11-30 22:11 UTC (History)
1 user (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2006-09-18 17:37:52 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
debug output from gdm (7.99 KB, text/plain)
2006-09-18 16:37 UTC, Michal Jaegermann
no flags Details

Description Michal Jaegermann 2006-09-17 18:16:08 UTC
Description of problem:

After the latest update to gdm-2.16.0-4.fc6 I started to collect
in /var/log/messages entries like that:

Sep 17 11:40:38 dyna0 kernel: gdm-binary[2612]: segfault at 0000000000000010 rip
000000322f2082b9 rsp 00007fffcfd28940 error 4
Sep 17 11:48:28 dyna0 kernel: gdm-binary[3215]: segfault at 0000000000000010 rip
000000322f2082b9 rsp 00007fff5a48f680 error 4
Sep 17 11:49:00 dyna0 kernel: gdm-binary[3314]: segfault at 0000000000000010 rip
000000322f2082b9 rsp 00007ffffd890a80 error 4

This happens every time gdm starts and it is enough to 'pkill -f gdm'
to get the next segfault entry.  After that gdm seems to start, shows
a login screen and everything looks fine save those segfaults.
'pgrep -l -f gdm' after the last restart shows:

3273 /usr/sbin/gdm-binary -nodaemon
3308 /usr/sbin/gdm-binary -nodaemon
3317 /usr/bin/Xorg :0 -audit 0 -auth /var/gdm/:0.Xauth -nolisten tcp vt7

Version-Release number of selected component (if applicable):
gdm-2.16.0-4.fc6

How reproducible:
on every gdm start

Additional info:
I have gdm currently configured to use gdmlogin plus BackgroundProgram,
which happens to be a shell script, and not gdmgreeter.

I do not have other architecure test system to check if the issue
is x86_64 specific.

Comment 1 Ray Strode [halfline] 2006-09-18 14:22:56 UTC
if you set Enable=true in the [debug] section of /etc/gdm/custom.conf what does
/var/log/messages say before the crash?

Do you have  the coolkey  package installed?

Comment 2 Michal Jaegermann 2006-09-18 16:37:26 UTC
Created attachment 136554 [details]
debug output from gdm

> ... what does /var/log/messages say before the crash?

It says:

gdm_server_spawn: Forked server on pid 2624
gdm-binary[2623]: segfault at 0000000000000010 ....

The whole relevant excerpt from /var/log/messages is attached.
This is the only place where 2623 process is mentioned in logs.

Can you tell if this happens in such location that a NULL pointer
would be valid and would prevent a memory access?  But if this is
a single bit flip that I would expect to see more troubles and
not only, consistently, a new version of gdm.

> Do you have  the coolkey  package installed?

No, I do not.

Comment 3 Ray Strode [halfline] 2006-09-18 17:37:52 UTC
Okay, i think I found the problem.  Can you tell me if you still see the problem
in tomorrow's rawhide?  If so, reopen.

Comment 4 Michal Jaegermann 2006-09-20 19:23:20 UTC
Indeed, after an update to gdm-2.16.0-8.fc6 segfaults are gone.  Thanks!

Comment 5 Xiaohong Wang 2006-10-15 09:22:31 UTC
Confirmed that it has been fixed in gdm-2.16.0-8.fc6 and gdm-2.16.0-10.el5


Note You need to log in before you can comment on or make changes to this bug.