No apport report generated for kernel oops

Bug #344813 reported by Matt Zimmerman
4
Affects Status Importance Assigned to Milestone
kerneloops (Ubuntu)
Fix Released
Medium
Ken VanDine

Bug Description

Binary package hint: kerneloops

I've had two kernel oopses since I started using kerneloops. The first one (bug 344127) didn't trigger a response from kerneloops at all. The second (bug 344812) did trigger a response from kerneloops, and I saw its dialog asking whether I should submit it. I clicked "Always", and then was presented with a *second* dialog, which said "...sent" but still offered the same buttons (always, once, etc.). Presumably that is covered by bug 344377.

The problem I'm reporting in this bug is that no problem report was created in /var/crash for the oops. It may have been submitted to kerneloops.org, but it should have triggered a report to be submitted to Launchpad as well, but it didn't.

My configuration is the default:

perseus:[/tmp/kerneloops-0.12] grep apport /etc/kerneloops.conf
submit-pipe = /usr/share/apport/kernel_oops

ProblemType: Bug
Architecture: amd64
DistroRelease: Ubuntu 9.04
Package: kerneloops 0.12-0ubuntu3
ProcEnviron:
 LC_COLLATE=C
 PATH=(custom, user)
 LANG=en_US.UTF-8
 SHELL=/bin/zsh
SourcePackage: kerneloops
Uname: Linux 2.6.28-9-generic x86_64

Related branches

Revision history for this message
Matt Zimmerman (mdz) wrote :
Revision history for this message
Matt Zimmerman (mdz) wrote :
Revision history for this message
Matt Zimmerman (mdz) wrote :

Confirming based on comments in bug 340773

Changed in kerneloops (Ubuntu):
importance: Undecided → Medium
status: New → Triaged
Revision history for this message
Ken VanDine (ken-vandine) wrote :

There is an apport traceback in the /var/log/apport.log:

apport (pid 17813) Thu Mar 19 09:57:11 2009: Unhandled exception:
Traceback (most recent call last):
  File "/usr/share/apport/apport", line 220, in <module>
    drop_privileges(pid, True)
  File "/usr/share/apport/apport", line 54, in drop_privileges
    raise ValueError, 'Invalid process ID: ' + str(e)
ValueError: Invalid process ID: [Errno 2] No such file or directory: '/proc/3714'
pid: 17813, uid: 0, gid: 0, euid: 0, egid: 0
environment: {}
apport (pid 17855) Thu Mar 19 09:57:11 2009: another apport instance is already running, aborting
apport (pid 17853) Thu Mar 19 09:57:12 2009: another apport instance is already running, aborting

Revision history for this message
Matt Zimmerman (mdz) wrote :

The problem is that the config file parser is broken:

 (gdb) print submit_pipe
 $1 = 0x1c73360 "/usr/share/apport/kernel_oops\n"

This is true for the other string options as well. The attached patch fixes it.

Revision history for this message
Matt Zimmerman (mdz) wrote : Re: [Bug 344813] Re: No apport report generated for kernel oops

On Thu, Mar 19, 2009 at 04:53:45PM -0000, Ken VanDine wrote:
>
> There is an apport traceback in the /var/log/apport.log:
>
> apport (pid 17813) Thu Mar 19 09:57:11 2009: Unhandled exception:
> Traceback (most recent call last):
> File "/usr/share/apport/apport", line 220, in <module>
> drop_privileges(pid, True)
> File "/usr/share/apport/apport", line 54, in drop_privileges
> raise ValueError, 'Invalid process ID: ' + str(e)
> ValueError: Invalid process ID: [Errno 2] No such file or directory: '/proc/3714'
> pid: 17813, uid: 0, gid: 0, euid: 0, egid: 0
> environment: {}
> apport (pid 17855) Thu Mar 19 09:57:11 2009: another apport instance is already running, aborting
> apport (pid 17853) Thu Mar 19 09:57:12 2009: another apport instance is already running, aborting

Are you sure that is from kerneloops? It looks like it's from a segfault.
According to my testing, kerneloops has never successfully invoked apport.

--
 - mdz

Revision history for this message
Ken VanDine (ken-vandine) wrote :

Glad to see you figured it out. I am not certain, it does happen every time i resume which is when i get the oops. So I guess something else is crashing on suspend/resume as well.

Revision history for this message
Matt Zimmerman (mdz) wrote :

On Thu, Mar 19, 2009 at 05:16:13PM -0000, Ken VanDine wrote:
> Glad to see you figured it out. I am not certain, it does happen every
> time i resume which is when i get the oops. So I guess something else
> is crashing on suspend/resume as well.

I get it on every resume due to:

[47611.466014] WARNING: synaptics was reset on resume, see synaptics_resume_reset if you have trouble on resume

--
 - mdz

Revision history for this message
Ken VanDine (ken-vandine) wrote :

attached debdiff

Changed in kerneloops:
assignee: nobody → ken-vandine
status: Triaged → In Progress
Martin Pitt (pitti)
Changed in kerneloops (Ubuntu):
status: In Progress → Fix Committed
Revision history for this message
Manoj Iyer (manjo) wrote :

Attaching a patch to eliminate compiler error due to missing ctype.h in configfile.c

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package kerneloops - 0.12-0ubuntu4

---------------
kerneloops (0.12-0ubuntu4) jaunty; urgency=low

  [ Ken VanDine ]
  * debian/applied-patches/kerneloops-applet.c.patch: Convert notification to
    a dialog. Applied inline, since package has no patch system. (LP: #344377)
  * debian/rules: Remove test/*dbg, these should be cleaned by the Makefile
  * debian/applied-patches/submit.c.patch: Fix the config parser so that it
    will actually work with apport. Thanks to Matt Zimmerman for the patch!
    (LP: #344813)

  [ Martin Pitt ]
  * Moved *.patch to debian/applied-patches/, to be a little less confusing.

 -- Ken VanDine <email address hidden> Thu, 19 Mar 2009 08:48:36 -0400

Changed in kerneloops:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.