indicator-messages-service crashed with SIGSEGV in dirname()

[LaMont Jones]

libnss-db does not seem to like how it's being called.

ProblemType: Crash
DistroRelease: Ubuntu 12.04
Package: indicator-messages 0.6.0-0ubuntu1
ProcVersionSignature: Ubuntu 3.2.0-23.36-generic 3.2.14
Uname: Linux 3.2.0-23-generic x86_64
ApportVersion: 2.0.1-0ubuntu2
Architecture: amd64
Date: Fri Apr 13 09:09:35 2012
ExecutablePath: /usr/lib/indicator-messages/indicator-messages-service
InstallationMedia: Ubuntu 11.10 "Oneiric Ocelot" - Release amd64 (20111011)
ProcCmdline: /usr/lib/indicator-messages/indicator-messages-service
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=en_US.UTF-8
SegvAnalysis:
 Segfault happened at: 0x6117c0 <dirname>: add %al,(%rax)
 PC (0x006117c0) in non-executable VMA region: 0x00611000-0x00612000 rw-p /usr/lib/indicator-messages/indicator-messages-service
 source "%al" ok
 destination "(%rax)" (0x015dd770) ok
SegvReason: executing writable VMA /usr/lib/indicator-messages/indicator-messages-service
Signal: 11
SourcePackage: indicator-messages
StacktraceTop:
 dirname ()
 ?? () from /usr/lib/libnss_db.so.2
 ?? () from /usr/lib/libnss_db.so.2
 _nss_db_getpwnam_r () from /usr/lib/libnss_db.so.2
 getpwnam_r () from /lib/x86_64-linux-gnu/libc.so.6
Title: indicator-messages-service crashed with SIGSEGV in dirname()
UpgradeStatus: Upgraded to precise on 2012-02-12 (61 days ago)
UserGroups: adm cdrom
XsessionErrors:
 gnome-session[5818]: WARNING: Could not parse desktop file /home/lamont/.config/autostart/xfce4-settings-helper-autostart.desktop: Key file does not have key 'Name'
 gnome-session[5818]: WARNING: could not read /home/lamont/.config/autostart/xfce4-settings-helper-autostart.desktop

[LaMont Jones (lamont)]

This is probably related to bug 761463.

[Apport retracing service (apport)]

StacktraceTop:
 dirname ()
 internal_setent (file=0x7f1a2ad436e7 "/var/lib/misc/passwd.db", dbp=0x7f1a2af45288, dbenvp=0x7f1a2af45290) at db-open.c:66
 lookup (key=0x7fff3e667a50, result=0x7fff3e667b30, buffer=0x15e0860, buflen=1024, errnop=0x7f1a33a7d8e0) at db-XXX.c:116
 _nss_db_getpwnam_r (name=0x7fff3e668fba "lamont", result=0x7fff3e667b30, buffer=0x15e0860 "snmp", buflen=1024, errnop=0x7f1a33a7d8e0) at files-pwd.c:33
 __getpwnam_r (name=0x7fff3e668fba "lamont", resbuf=0x7fff3e667b30, buffer=0x15e0860 "snmp", buflen=1024, result=0x7fff3e667b68) at ../nss/getXXbyYY_r.c:256

[Bilal Akhtar (bilalakhtar)]

The stacktrace looks weird. What were you doing at the time of the crash?

[Charles Kerr (charlesk)]

I don't know that this is an indicator-messages problem -- in indicator-message's seen-db.c, the crash comes by calling g_get_user_cache_dir(), which doesn't even take any arguments so I don't think we're using it wrong ;)

In glib, the crash is coming from this block:

> logname = (gchar *) g_getenv ("LOGNAME");
> ...
> if (logname) {
> error = getpwnam_r (logname, &pwd, buffer, bufsize, &pw);

Where logname appears to be "lamont"

[Charles Kerr (charlesk)]

Ahh, I knew this looked familiar -- I triaged an identical crash in bug #963870 :)

LaMont, I'm closing this ticket as a duplicate of Bug #761463. Please look at the other bug report to see if there is any missing information that you can provide, or to see if there is a workaround for the bug. Additionally, any further discussion regarding the bug should occur in the other report. Feel free to continue to report any other bugs you may find.