totem crashed with SIGSEGV in tcache_get() [assertion "nqueue != NULL" failed in g_object_new_internal]

[Rachel Greenham]

https://errors.ubuntu.com/problem/e6f21a21cd523f0c489de07464069cea59d69427

I made no attempt to run Videos. This just came up out of the blue. Maybe a video embedded in Tweetdeck tried to trigger it? (Firefox 66.0b4 from firefox-next ppa, installed for the working CSD when buttons on left.)

ProblemType: Crash
DistroRelease: Ubuntu 19.04
Package: totem 3.30.0-4ubuntu1
ProcVersionSignature: Ubuntu 4.18.0-13.14-generic 4.18.17
Uname: Linux 4.18.0-13-generic x86_64
ApportVersion: 2.20.10-0ubuntu19
Architecture: amd64
CurrentDesktop: ubuntu:GNOME
Date: Wed Jan 23 13:41:15 2019
ExecutablePath: /usr/bin/totem
InstallationDate: Installed on 2018-09-11 (141 days ago)
InstallationMedia: Ubuntu 18.04 LTS "Bionic Beaver" - Alpha amd64 (20180214)
ProcCmdline: /usr/bin/totem --gapplication-service
ProcEnviron:
 XDG_RUNTIME_DIR=<set>
 SHELL=/bin/bash
 LANGUAGE=en_GB:en
 PATH=(custom, user)
 LANG=en_GB.UTF-8
SegvAnalysis:
 Segfault happened at: 0x7f609e404d36 <__GI___libc_malloc+422>: mov (%rdx),%rsi
 PC (0x7f609e404d36) ok
 source "(%rdx)" (0x561c1b1ae92000) not located in a known VMA region (needed readable region)!
 destination "%rsi" ok
SegvReason: reading unknown VMA
Signal: 11
SourcePackage: totem
StacktraceTop:
 tcache_get (tc_idx=1) at malloc.c:2927
 __GI___libc_malloc (bytes=35) at malloc.c:3034
 g_malloc () at /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
 g_strconcat () at /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
 g_assertion_message_expr () at /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0
Title: totem crashed with SIGSEGV in tcache_get()
UpgradeStatus: Upgraded to disco on 2019-01-13 (17 days ago)
UserGroups: adm cdrom dip lpadmin plugdev sambashare sudo
XorgLog: Error: [Errno 2] No such file or directory: '/var/log/Xorg.0.log'

[Apport retracing service (apport)]

StacktraceTop:
 tcache_get (tc_idx=1) at malloc.c:2927
 __GI___libc_malloc (bytes=bytes@entry=35) at malloc.c:3034
 g_malloc (n_bytes=n_bytes@entry=35) at ../../../../glib/gmem.c:99
 g_strconcat (string1=string1@entry=0x7f609f01d969 "assertion failed: (") at ../../../../glib/gstrfuncs.c:585
 g_assertion_message_expr (domain=domain@entry=0x7f609f0cf07b "GLib-GObject", file=file@entry=0x7f609f0d12ce "../../../../gobject/gobject.c", line=line@entry=1813, func=func@entry=0x7f609f0d28c0 <__func__.14400> "g_object_new_internal", expr=expr@entry=0x7f609f0d148e "nqueue != NULL") at ../../../../glib/gtestutils.c:2618

[Sebastien Bacher (seb128)]

Thank you for your bug report. Without steps to reproduce the issue it's going to be a bit difficult to debug that though, could see if you can trigger it again by doing the same things you were doing?

[Rachel Greenham (rachel-strangenoises)]

I agree it's a bit of a nightmare. It was the start of the day and I
literally just had Firefox-Next open, and Slack (using the app, not via
Firefox). No video had been posted to slack. I can only think some
random embedded video went by on the timeline or activities column in
tweetdeck (the only tab) on firefox. Nothing else was going on, and said
embedded videos usually either work fine or not at all. In the past I've
had stuff like this happen when opening a Nautilus window onto a
directory with videos in it, so it being the preview of said videos that
probably crashed... but not this time.

So I agree, there's probably not much to be done about this unless it
starts happening more. So far just that one time. (I didn't set it to
ignore similar in future.)

(NB: replied via email because on-site form is timing out for me, all
else working)

[Sebastien Bacher (seb128)]

Thanks for the reply. It's a bit weird if it comes from firefox because the totem webbrowser plugin got deprecated and there is no reason firefox would call to totem to play video...

[Rachel Greenham (rachel-strangenoises)]

True. And I can confirm it's not listed in my plugins at all. (There's
just Cisco OpenH264 and Widevine). Only potentially-relevant extension
is Disable HTML5 Autoplay, which if anything should *stop* any video
being played before it gets to the software that does it.

Only other thing that seems odd, is the mimetime mapping Firefox has for
MP3 audio (in prefs) *defaults* to Videos, but in my case it's set to
"Use env"... but I wouldn't have set that (is first time I've seen it).
It wouldn't be surprising if "Use env" gets it to be sent to Videos too,
by a system default? So maybe it was an MP3 file. But not one I
knowingly clicked on.

... but finding some MP3 files and playing them in the browser
deliberately is working fine.

So no, I'm sorry, I'm mystified.

[Launchpad Janitor (janitor)]

[Expired for totem (Ubuntu) because there has been no activity for 60 days.]

[Daniel van Vugt (vanvugt)]

Also tracking in https://errors.ubuntu.com/problem/e6f21a21cd523f0c489de07464069cea59d69427

[Daniel van Vugt (vanvugt)]

Confirmed, but the number of crash reports is incredibly low right now:

artful = 1
bionic = 1
cosmic = 1
disco = 1

Also it sounds like a memory corruption issue. So the location it crashed in may not be the problem at all. If you are able to ever reproduce the issue then please try adding this to your /etc/environment first:

  MALLOC_CHECK_=3

Then produce the same crash and that should give a better idea of the root cause.