indicator-applet crashed with SIGSEGV in g_simple_async_result_complete()

[David Planella]

Binary package hint: indicator-applet

This is the second time I've seen this happening in the course of two days.

I cannot provide much more information, unfortunately. I was simply editing a text file, not interfacing directly with the indicators at all, when all of them disappeared from the indicator area and then I got informed about the crash and whether I wanted to reload them. Reloading was successful.

ProblemType: Crash
DistroRelease: Ubuntu 11.04
Package: indicator-applet-complete 0.4.7-0ubuntu1
ProcVersionSignature: Ubuntu 2.6.37-12.26-generic 2.6.37
Uname: Linux 2.6.37-12-generic i686
NonfreeKernelModules: nvidia
Architecture: i386
Date: Sat Jan 29 13:40:23 2011
ExecutablePath: /usr/lib/indicator-applet/indicator-applet
ProcCmdline: /usr/lib/indicator-applet/indicator-applet
ProcEnviron:
 SHELL=/bin/bash
 LC_MESSAGES=ca_ES.utf8@valencia
 LANGUAGE=ca_ES@valencia:ca:en_GB:en
 LANG=ca_ES.UTF-8
SegvAnalysis:
 Segfault happened at: 0x70a053e: mov 0x1c(%esi),%eax
 PC (0x070a053e) ok
 source "0x1c(%esi)" (0xaaaaaac6) not located in a known VMA region (needed readable region)!
 destination "%eax" ok
SegvReason: reading unknown VMA
Signal: 11
SourcePackage: indicator-applet
StacktraceTop:
 ?? () from /usr/lib/libdbusmenu-glib.so.3
 g_simple_async_result_complete (simple=0x8fc6930) at /build/buildd/glib2.0-2.27.92/gio/gsimpleasyncresult.c:747
 reply_cb (connection=0x8fc3820, res=0x9147a38, user_data=0x8fc6930) at /build/buildd/glib2.0-2.27.92/gio/gdbusproxy.c:2158
 g_simple_async_result_complete (simple=0x9147a38) at /build/buildd/glib2.0-2.27.92/gio/gsimpleasyncresult.c:747
 g_dbus_connection_call_done (source=0x8fc3820, result=0x8fc6888, user_data=0x92836a0) at /build/buildd/glib2.0-2.27.92/gio/gdbusconnection.c:4995
Title: indicator-applet crashed with SIGSEGV in g_simple_async_result_complete()
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

[Loïc Minier (lool)]

I got this crasher in close sequence today; I couldn't stop it, so I had to debug it a bit. :-)

I tried running indicator-applet within valgrind, but apart of a conditional jump on startup in the gtk / zlib init, I would just get the backtrace at the time where it segfaults, even after installing indicator-applet, libdbusmenu-glib3 and libglib2.0-0 ddebs.

I tried running indicator-applet within gdb, and saw the same backtrace is actually in update_layout_cb(), where data is optimized out (it's a static function) but if I print *client, I see that it's corrupt.

I added a breakpoint where the debug is registered, but all the private callback data was correct at this point, so it seems corruption is happening later on.

My system has the gnome-power-manager icon updating erratically, so I tried killing gnome-power-manager; now that I've done that, it doesn't crash anymore.

[Loïc Minier (lool)]

bug #707804 has a retraced backtrace

[Sebastien Bacher (seb128)]

the issue should be fixed in the current versions