Mir

Mir crashes from time to time when using Telegram while network changes [AndroidWindow: Caught exception at Mir/EGL driver boundary (in dequeueBuffer): Dynamic exception type: St11logic_error std::exception::what: state was not valid]

Bug #1630230 reported by zubozrout
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Mir
Expired
Undecided
Unassigned
mir (Ubuntu)
Expired
Undecided
Unassigned

Bug Description

Hi,
for more than 2 months I've been occasionally experiencing Mir crashes when I use Telegram and my network connection changes from none to cellular and wifi - which happens a lot to me while travelling by subway and/or passing by shopping centers/offices/school/public transport with wifi networks saved.

I've experienced this several times and it happened just after I sent a message via Telegram (the message actually got through, I could even see it being sent moments before my phone froze). It is however not unusual for me to experience this while I am composing a message either.

If this is not a pure Mir issue but something that Telegram causes so that Mir crashes I am pointing out that I only recall seeing this issue when Telegram is in the working state, that is not the broken one that ignores everything although network clearly works and incoming notifications appear and only way to fix this is to reopen the app.

Here is a Telegram log I have: http://paste.ubuntu.com/23274022/ (my system crashed at around 10:25)

Thank you very much for looking into this. I am running rc-proposed on MX4.

Gerry Boland (gerboland)
affects: mir → unity8
Revision history for this message
Gerry Boland (gerboland) wrote :

More likely it is something with unity8, but not impossible Mir is at fault. But without some kind of crash dump, and it being extremely hard to reproduce, this is hard to solve.

A crash dump was generated, and hopefully uploaded to our servers. You can see them by visiting System Settings, Privacy & Security, Diagnostics, Previous Error reports (opens webbrowser). Can you share any unity8 reports here?

Revision history for this message
zubozrout (zubozrout) wrote :

Hello, here are the Unity8 and Unity System Compositor logs. Apparently nothing was sent to your servers today except for Xmir log about 10 minutes prior to this crash - but I don't usually run Xmir apps and I am certain I've experienced most of the crashes like this when I didn't run Xmir.

Revision history for this message
zubozrout (zubozrout) wrote :
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Actually the telegram log suggests this is a Mir (libmirclient) crash:

[2016-10-04 10:25:44.729406] <ERROR> AndroidWindow: Caught exception at Mir/EGL driver boundary (in dequeueBuffer): Dynamic exception type: St11logic_error
std::exception::what: state was not valid

[2016-10-04 10:25:44.730737] <ERROR> AndroidWindow: Caught exception at Mir/EGL driver boundary (in dequeueBuffer): Dynamic exception type: St11logic_error
std::exception::what: state was not valid

affects: unity8 → mir
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Although the logs also suggest that Unity8/USC had a hiccup/restart at 10:25 too, but without actually restarting.

The only definite error is the one in the telegram log, which is a libmirclient exception.

summary: Mir crashes from time to time when using Telegram while network changes
+ [AndroidWindow: Caught exception at Mir/EGL driver boundary (in
+ dequeueBuffer): Dynamic exception type: St11logic_error
+ std::exception::what: state was not valid]
tags: added: android
Revision history for this message
zubozrout (zubozrout) wrote :

Another one just now, a bit different log (time index 8:34)
http://paste.ubuntu.com/23278517/

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

That one and the first one could plausibly be due to a server bug that is breaking the client.

Revision history for this message
zubozrout (zubozrout) wrote :

This one is probably the same

Revision history for this message
zubozrout (zubozrout) wrote :

And so is this ... hmm, sorry if it is a duplicate log but maybe there is something I don't see

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Actually I just hit a similar exception in a nested server (which throws continuously and infinitely):

[2016-10-05 15:49:42.260127] <ERROR> Mesa/NativeSurface: Caught exception at Mir/EGL driver boundary (in advance_buffer): Dynamic exception type: std::logic_error
std::exception::what: state was not valid

[2016-10-05 15:49:42.262044] <ERROR> Mesa/NativeSurface: Caught exception at Mir/EGL driver boundary (in advance_buffer): Dynamic exception type: std::logic_error
std::exception::what: state was not valid

[2016-10-05 15:49:42.264824] <ERROR> Mesa/NativeSurface: Caught exception at Mir/EGL driver boundary (in advance_buffer): Dynamic exception type: std::logic_error
std::exception::what: state was not valid

[2016-10-05 15:49:42.266636] <ERROR> Mesa/NativeSurface: Caught exception at Mir/EGL driver boundary (in advance_buffer): Dynamic exception type: std::logic_error
std::exception::what: state was not valid

[2016-10-05 15:49:42.269473] <ERROR> Mesa/NativeSurface: Caught exception at Mir/EGL driver boundary (in advance_buffer): Dynamic exception type: std::logic_error
std::exception::what: state was not valid

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

I'm not sure if this was an Android-specific bug...

Changed in mir:
status: New → Incomplete
Changed in mir (Ubuntu):
status: New → Incomplete
tags: removed: android
Revision history for this message
Launchpad Janitor (janitor) wrote :

[Expired for Mir because there has been no activity for 60 days.]

Changed in mir:
status: Incomplete → Expired
Revision history for this message
Launchpad Janitor (janitor) wrote :

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

Changed in mir (Ubuntu):
status: Incomplete → Expired
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.