init crashed with SIGSEGV in event_poll()

StacktraceTop:job_change_goal (job=0x7f70bacdb1e0, goal=JOB_STOP,
job_handle_event (emission=0x7f70bacd8ff0)
event_poll () at event.c:478
nih_main_loop () at main.c:615
main (argc=1, argv=<value optimized out>)

Could you attach the file from /var/crash for me? This should not yet have been deleted from your system.

The stack trace indicates that init is segfaulting on function entry, which generally implies badness

Sorry, I thought apport had already done that for me :)

It does, but then it removes the core dump before retracing it and making it public.

Unfortunately the core file isn't any more revealing, the crash happened on function entry and all of the passed variables look just fine to me.

Were there any other console messages before the crash?

Does this happen every time for you?

If you have other crashes from Upstart, could you submit them with apport as well?

Unfortunately currently I only have one file in /var/crash/ directory, the one attached above.
The crashes happened at random times, so I couldn't identify a usage pattern that triggers it.
I've seen the same crash at least 3 times, but ignored apport suggestion to post the bug and manually deleted the core dumps in order to avoid apport to continue annoying :)

I didn't pay so much attention until I disabled the bootsplash and seen the actual segfault message. Fortunately I configured sysklogd to dump *.* log messages to tty12 previously, so I could easily take a pic right after it happened switching to tty12. I haven't seen any other messages than those in the attached pics.

Right after the init crash (and before rebooting, of course) I logged in as root and tried invoking init directly to see what happens.

kinda foolish method.. to take pics from the screen :) I just didn't wanted to go to look in /var/log/messages

Unfortunately this bug has insufficient information to determine a cause; given the location of the crash, I can't even rule out hardware failure or a Kernel issue (perhaps caused by a binary module, or virtualisation software)

Please delete the file from /var/crash; if this happens again, please file a new report and hopefully that one might have more information in the trace.

>> I can't even rule out hardware failure or a Kernel issue (perhaps caused by a binary module, or virtualisation software)

I was thinking maybe it could have something to do with VirtualBox driver (vboxdrv), which is loaded at boot time.
The module iwlagn (intel wireless) is another not-so-stable driver which is loaded at boot, wich caused a lot of ugly backtraces to printed in tty01 when shutting down my previous Intrepid installation. But I feel inclined to point a finger on vboxdrv.

I'll keep an eye on those two miscreants.