| Summary: | Crash in KF5 with kwin_x11 --replace ( xcb ) | ||
|---|---|---|---|
| Product: | [Plasma] kwin | Reporter: | Nicolas L. <kde> |
| Component: | core | Assignee: | KWin default assignee <kwin-bugs-null> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | major | CC: | hrvoje.senjan |
| Priority: | NOR | ||
| Version: | 5.0.0 | ||
| Target Milestone: | --- | ||
| Platform: | Mageia RPMs | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: |
here is the log.
new conf file |
||
|
Description
Nicolas L.
2014-07-24 21:35:05 UTC
Is this reproducible and do you always get the same backtrace in case? (xcb_get_setup returning 0x0 somehow smells like the connection() is invalid at this point, but it's required before some times, eg. when querying the extension list. Are you running this in some nested X11 server? as i can't start KF5, it kills X, i tried to start kwin_x11 under IceWM but it killed X. So my last try ( the one that allowed me to have the trace ) was to open IceWM, go in a tty and do DISPLAY=:0 gdb --args kwin_x11 --replace For your question, yes i can reproduce each time and the trace is always the same. > as i can't start KF5, it kills X, i tried to start kwin_x11 under IceWM but it killed X. -> there's hopefully a backtrace in /var/log/Xorg.0.log This however explains the backtrace in comment #0 Since the X11 server is gone, accessing it won't work (the backtrace is therefore also irrelevant and *not* a particular bug in KWin) The X11 server crash is (by definition, *clients* are not suppsed to crash the server ever) an Xorg bug, probably in the driver. -> Attach mentioned log and i'll tell you more ;-) Created attachment 87945 [details]
here is the log.
The one you attached is from a running X11 server (sorry, I could have anticipated this ;-) When you start another X11 server, the log is lost, resp. moved to "/var/log/Xorg.1.log" Please save it right after the crash (and w/o restarting X11) or use the implicit .1. backup. Created attachment 87948 [details]
new conf file
i don't see any Xorg.1.log, but i can see a Xorg.0.log.old, but i don't see anything related to a crash inside.
That server exited cleanly. Was this obtained right after the server crashed (on starting kwin_x11) and restarted no more than once afterwards? i used kwin_x11 --replace and as soon as X restarted i copied the file. is there a X binary i could plug gdb with to have more infos ? Yes. The binary is (usually) "X" ps ax | grep X ensure to "continue" X11 before switching back to its VT btw it only crash for me in a DE "not" KF5, in KF5 all works as expected. how do you determine that X11 crashes? notice that for many sessions the WM could be the client that keeps the session alive, ie. if you exit (attempting to replace) it, the session exits and the DM will likely restart the server in return. *** This bug has been marked as a duplicate of bug 343844 *** |