Bug 1375619

Summary: Constant plasma crashes since upgrade F23->F24
Product: [Fedora] Fedora Reporter: Martin Vuille <jpmv27>
Component: plasma-workspaceAssignee: KDE SIG <kde-sig>
Status: CLOSED WORKSFORME QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: urgent Docs Contact:
Priority: unspecified    
Version: 24CC: jgrulich, jpmv27, kde-sig, me, rdieter, than
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-09-15 10:11:10 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Martin Vuille 2016-09-13 14:19:40 UTC
Description of problem:
Upgraded from F23 to F24 a few days ago. Using KDE5/Plasma.
Since the upgrade, graphical shell is unusable, constantly "crashes."
Not sure what is really happening, most times there is no coredump
in /var/lib/systemd/coredump. On occasion I have also seen message
about DrKonqi crashing. After the crash, graphical shell window shows
a grey background and an active mouse cursor, but reacts to no input
other than mouse movements.

Please advise what information to collect and provide to enable
further analysis.

I will update this ticket if I am able to collect coredump or provide
additional information.


Version-Release number of selected component (if applicable):
Not really sure at this point which component is responsible.

How reproducible:
Random but frequent. May take an hour or a few seconds after logging in
to shell. No discernible pattern yet.


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:
Motherboard: Gigabyte GA-X48-DS4
Graphics: Advanced Micro Devices, Inc. [AMD/ATI] RV630 XT [Radeon HD 2600 XT]

Comment 1 Rex Dieter 2016-09-13 18:13:55 UTC
Can you provide any backtraces for these crashes?  Else, we'd can only speculate about what the problem could be.

Comment 2 Martin Vuille 2016-09-13 22:53:17 UTC
Yes, I realize this is not enough information. I was looking for some hints about where I could look for relevant information to provide since, as reported, there are no coredumps in /var/lib/systemd/coredump

I enabled /proc/sys/kernel/print-fatal-signals but do not see anything in messages or in journalctl

The only relevant information I could find was in .xsession-errors:

klauncher: Exiting on signal 15
crashreporter: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.
XIO:  fatal IO error 4 (Interrupted system call) on X server ":0"
      after 4916 requests (4916 known processed) with 0 events remaining.
XIO:  fatal IO error 11 (Resource temporarily unavailable) on X server ":0"
      after 17124 requests (17124 known processed) with 0 events remaining.
Application::crashHandler() called with signal 11; recent crashes: 1
KCrash: Attempting to start /usr/bin/kdeinit5 from kdeinit
Warning: connect() failed: : Connection refused
KCrash: Attempting to start /usr/bin/kdeinit5 directly
KCrash: Application 'kdeinit5' crashing...
KCrash: Attempting to start /usr/libexec/drkonqi from kdeinit
Warning: connect() failed: : Connection refused
KCrash: Attempting to start /usr/libexec/drkonqi directly
kdeinit5: Communication error with launcher. Exiting!
KCrash: Application 'kwin_x11' crashing...
/usr/bin/kwin_x11: FATAL ERROR while trying to open display :0

I will be happy to follow-up any suggestions you may have to track down this issue. If I am able to capture more information, I will update the ticket.

Comment 3 Rex Dieter 2016-09-14 00:06:40 UTC
Ah, based on your .xsession-errors content, X is crashing (or at least refusing new connections), and is taking down everything with it (including plasma)

Comment 4 Martin Vuille 2016-09-14 02:04:12 UTC
I can see that kwin_x11 isn't crashing. The PID is unchanged after the problem happens.

I also attached to kwin_x11 from gdb immediately after starting the session and waited for the problem to happen. Nothing seen in gdb. I can provide backtrace of the various threads if that might be helpful.

I tried "DISPLAY=:0 kwin_x11 --replace", which is suggested in KWin/Hacking Wiki page, and that resulted in the same FATAL ERROR seen above.

Comment 5 Martin Vuille 2016-09-14 19:51:40 UTC
Updates yesterday and this morning appear to have solved the issue.
Would like to close but not sure what status to put.

Comment 6 Rex Dieter 2016-09-15 10:11:10 UTC
Great, let's go with 'WORKSFORME'