tracker-miner-fs crashed with SIGSEGV in g_type_check_instance_is_a()

[Laurent Bonnaud]

This crash (or bug #1324808) happens each time I log in.

ProblemType: Crash
DistroRelease: Ubuntu 14.04
Package: tracker-miner-fs 1.0.1-2ubuntu1~trusty1 [origin: LP-PPA-gnome3-team-gnome3-staging]
ProcVersionSignature: Ubuntu 3.13.0-29.53-generic 3.13.11.2
Uname: Linux 3.13.0-29-generic x86_64
ApportVersion: 2.14.1-0ubuntu3.2
Architecture: amd64
CurrentDesktop: KDE
Date: Fri Jun 6 11:31:15 2014
EcryptfsInUse: Yes
ExecutablePath: /usr/lib/tracker/tracker-miner-fs
ProcCmdline: /usr/lib/tracker/tracker-miner-fs
SegvAnalysis:
 Segfault happened at: 0x7fa5f92d7d6c <g_type_check_instance_is_a+60>: testb $0x4,0x16(%r8)
 PC (0x7fa5f92d7d6c) ok
 source "$0x4" ok
 destination "0x16(%r8)" (0x5320545245534e5e) not located in a known VMA region (needed writable region)!
SegvReason: writing unknown VMA
Signal: 11
SourcePackage: tracker
StacktraceTop:
 g_type_check_instance_is_a (type_instance=type_instance@entry=0x2888580, iface_type=iface_type@entry=33188960) at /build/buildd/glib2.0-2.40.0/./gobject/gtype.c:3969
 g_file_get_uri (file=0x2888580) at /build/buildd/glib2.0-2.40.0/./gio/gfile.c:546
 sparql_files_compose_query (files=<optimized out>, n_files=679) at tracker-file-notifier.c:740
 sparql_files_query_start (n_files=<optimized out>, files=<optimized out>, notifier=0x1fc3100) at tracker-file-notifier.c:759
 crawler_finished_cb (crawler=<optimized out>, was_interrupted=<optimized out>, user_data=0x1fc3100) at tracker-file-notifier.c:843
Title: tracker-miner-fs crashed with SIGSEGV in g_type_check_instance_is_a()
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm autopilot cdrom dip fuse libvirtd lpadmin plugdev sambashare staff sudo

[Ubuntu GNOME (ug-bot)]

StacktraceTop:
 type_check_is_value_type_U (type=<optimized out>) at /build/buildd/glib2.0-2.41.0~git20140523.570b27b9/./gobject/gtype.c:4107
 g_type_check_value_holds (value=0x28d3440, type=40443936) at /build/buildd/glib2.0-2.41.0~git20140523.570b27b9/./gobject/gtype.c:4153
 ?? ()
 ?? ()
 ?? ()

[Laurent Bonnaud (laurent-bonnaud)]

The following environment variables might be needed to reproduce this bug:

MALLOC_CHECK_=3
MALLOC_PERTURB_=117

[Tim Lunn (darkxst)]

Hey Laurent, the retracer got stumped on the retrace in this bag, can you try and get one locally using the following instructions, thanks!

Thank you for taking the time to report this bug and helping to make Ubuntu better. Please try to obtain a backtrace following the instructions at http://wiki.ubuntu.com/DebuggingProgramCrash and upload the backtrace (as an attachment) to the bug report. This will greatly help us in tracking down your problem.

[Laurent Bonnaud (laurent-bonnaud)]

The backtrace was in my initial report:

https://launchpadlibrarian.net/177071868/Stacktrace.txt

[Tim Lunn (darkxst)]

Laurent, can you file a bug upstream with that stacktrace, then link the upstream bug to this report.

[Laurent Bonnaud (laurent-bonnaud)]

I could not find tracker in launchpad (despite spending quite a lot of time trying to find it) therefore I linked it to some other GNOME project and provided the GNOME bugzilla URL (which is the useful part).

[Martyn Russell (martyn-lanedo)]

Hi, I am taking a look at this bug upstream, is there a consistent way to reproduce this?
Has anyone tried:

  /usr/libexec/tracker-miner-fs -v 3 -s 0

The output of that would help here so I can see if it is a file causing this or some unique scenario.
Thanks!

[Laurent Bonnaud (laurent-bonnaud)]

In Ubuntu the path to tracker-miner-fs is slightly different, therefore I used:

$ /usr/lib/tracker/tracker-miner-fs -v 3 -s 0
[...]
(tracker-miner-fs:20960): Tracker-DEBUG: Added monitor for path:'file:///home/bonnaudl/Logiciels/hevc.hhi.fraunhofer.de/svn/svn_HEVCSoftware/tags/HM-3.0/build/linux/test/TAppEncoder', total monitors:5401

(tracker-miner-fs:20960): GLib-GIO-CRITICAL **: g_file_get_uri: assertion 'G_IS_FILE (file)' failed
Segmentation fault (core dumped)

Here are details on concerned directories:

$ l /home/bonnaudl/Logiciels/hevc.hhi.fraunhofer.de/svn/svn_HEVCSoftware/tags/HM-3.0/build/linux/test/TAppEncoder/
total 8
-rw------- 1 bonnaudl bonnaudl 415 avril 6 2011 index.html
-rw------- 1 bonnaudl bonnaudl 1905 avril 6 2011 makefile

$ l /home/bonnaudl/Logiciels/hevc.hhi.fraunhofer.de/svn/svn_HEVCSoftware/tags/HM-3.0/build/linux/test
total 12
-rw------- 1 bonnaudl bonnaudl 450 avril 6 2011 index.html
drwx------ 2 bonnaudl bonnaudl 4096 mai 4 2011 TAppDecoder/
drwx------ 2 bonnaudl bonnaudl 4096 mai 4 2011 TAppEncoder/

[Laurent Bonnaud (laurent-bonnaud)]

I also ran tracker-miner-fs with valgrind: it allows tracker-miner-fs to finish without crashing.

And on a subsequent run without valgrind tracker-miner-fs also completed without crashing.

[Laurent Bonnaud (laurent-bonnaud)]

I then tried to reproduce the bug from scratch:
 - I killed the tracker processes
 - I deleted ~/.local/share/tracker/ and ~/.cache/tracker/
 - I ran /usr/lib/tracker/tracker-miner-fs -v 3 -s 0

and could not reproduce the crash.

The crash seems to be caused by some state of the databases created by tracker.

[Laurent Bonnaud (laurent-bonnaud)]

Today I had to log in again and the crash occurred again:

 - running /usr/lib/tracker/tracker-miner-fs -v 3 -s 0 led to a reproducible crash

 - unsetting the MALLOC_PERTURB_=117 variable allowed tracker-miner-fs to complete without crashing

 - running /usr/lib/tracker/tracker-miner-fs -v 3 -s 0 again led to no more crash

[Christian Dysthe (christian-dysthe)]

I'm seeing this crash during one out of four logins.