unity-panel-service crashed with SIGSEGV in panel_indicator_entry_accessible_get_n_children()

[Edward Donovan]

unity-panel-service has crashed a few times on me, but I can't reproduce it consistently. This occured shortly after logging in; I don't think any action prompted it.

I'm running indicator-weather in the panel, which I think is an unofficial package at this point, and has occasionally crashed, itself. And indicator-multiload, not sure if that's unofficial as well. And glipper, which doesn't seem to integrate into the Unity panel just right, and has also occasionally crashed, itself.

None of these reported a crash at the moment, and the panel seems to have restarted successfully, this time. Apport prompted me to report, so I did. :-/

It may be a dupe some of the other u-p-s crashes, but those sounded like they had different particular triggers, so I wasn't sure. Thanks.

ProblemType: Crash
DistroRelease: Ubuntu 12.04
Package: unity-services 5.6.0-0ubuntu4
ProcVersionSignature: Ubuntu 3.2.0-19.30-generic 3.2.11
Uname: Linux 3.2.0-19-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 1.94.1-0ubuntu2
Architecture: amd64
CompizPlugins: [core,bailer,detection,composite,opengl,decor,mousepoll,resize,snap,imgpng,place,vpswitch,compiztoolbox,wall,regex,move,scale,gnomecompat,workarounds,expo,session,unityshell]
CrashCounter: 1
Date: Sun Mar 18 10:03:49 2012
ExecutablePath: /usr/lib/unity/unity-panel-service
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Release amd64 (20101007)
ProcCmdline: /usr/lib/unity/unity-panel-service
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
SegvAnalysis:
 Segfault happened at: 0x405d02: cmp %rsi,(%rdx)
 PC (0x00405d02) ok
 source "%rsi" ok
 destination "(%rdx)" (0x6c8948d8245c8948) not located in a known VMA region (needed writable region)!
SegvReason: writing unknown VMA
Signal: 11
SourcePackage: unity
StacktraceTop:
 ?? ()
 ?? () from /usr/lib/x86_64-linux-gnu/gtk-3.0/modules/libatk-bridge.so
 g_main_context_dispatch () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
 ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
 g_main_loop_run () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
Title: unity-panel-service crashed with SIGSEGV in g_main_context_dispatch()
UpgradeStatus: Upgraded to precise on 2012-03-13 (5 days ago)
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

[Apport retracing service (apport)]

StacktraceTop:
 panel_indicator_entry_accessible_get_n_children (accessible=<optimized out>) at /build/buildd/unity-5.6.0/services/panel-indicator-entry-accessible.c:254
 ?? ()
 ?? ()
 ?? ()
 ?? ()

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Edward Donovan (edward.donovan)]

It happened again, upon login, in new session. So far, I have no idea why it happens, on a minority of login events.

[Edward Donovan (edward.donovan)]

I see that I didn't note the installed version of unity, when the second crash happened. Oops. Looking at my logs, it probably happened shortly after I upgraded to unity-services 5.8.0-0ubuntu2. Sorry I can't be more sure about that. The crash log is gone from /var/crash, now.

It has not happened again, to date -- about two weeks since the second occurence.

[Edward Donovan (edward.donovan)]

Happened again, under Unity 5.10.0-0ubuntu6. Again shortly after login. I'm attaching the crashfile; don't know if it adds anything. There were plenty of logins in-between with no crash, and I see nothing distinctive about the times it occurred. Unfortunately.

I am still running some non-default indicators. None crashed themselves. indicator-multiload printed this to .xsession-errors.

 (indicator-multiload:28768): GLib-GIO-CRITICAL **: GApplication subclass 'Main' failed to chain up on ::startup (from start of override function)
 (compiz:28726): GConf-CRITICAL **: gconf_client_add_dir: assertion `gconf_valid_key (dirname, NULL)' failed
 (indicator-multiload:28768): GdkPixbuf-CRITICAL **: gdk_pixbuf_scale_simple: assertion `dest_width > 0' failed
 (indicator-multiload:28768): GLib-GObject-CRITICAL **: g_object_ref: assertion `G_IS_OBJECT (object)' failed

I will have to test later to see if it prints such warnings, at other times when the crash hasn't occured.

If I could reproduce this on demand, I would test it with and without various indicators. (Not that I'm sure they're involved.) But it happens very infrequently, and randomly from my POV.

This is causing me no real trouble, I'm just providing info in case it's of use. Thanks.

[Edward Donovan (edward.donovan)]

indicator-multiload does print those error outputs in other sessions, where no panel-service crash occurred. So that may not be too relevant.