Bugzilla – Bug 113256
KDE crashes on reboot or system shutdown
Last modified: 2006-03-14 11:29:53 UTC
Easy to reproduce, shutdown of system or reboot leads to a KDE crash message: Signal 11 (SIGSEGV)
I wasn't able to reproduce this bug on my system. To be able to fix this bug we would need more information. Please check: http://www.opensuse.org/index.php/Submit_a_bug http://www.opensuse.org/index.php/Bugs:KDE
Created attachment 47920 [details] KDE crash log I attached the KDE crash backtrace. Anything else needed?
I was not able to reproduce this bug. I tried to shut down/reboot (init 0, 6) the computer via the KDE-menu both as root and as normal user. Worked well so far. Please give us more information, like: + Does this appear always? If no: when does it appear (maby after you run a special software or executed a specific function). Also check if it works when you login and shut down without running any program! + Does it appear both as user and as root? + Attach the last 500 lines of your syslog as root, execute: tail -500 /var/log/messages | gzip - > syslog.gz and attach syslog.gz (maby there is something else wrong with your sys) + Describe your hardware configuration
Created attachment 48031 [details] syslog
(In reply to comment #3) > I was not able to reproduce this bug. I tried to shut down/reboot (init 0, 6) > the computer via the KDE-menu both as root and as normal user. Worked well so far. > > Please give us more information, like: > + Does this appear always? Yes, everytime I start up KDE and end session, reboot or shutdown via K menu. It also works when I simply login and shut down without running any program! > + Does it appear both as user and as root? Yes. > + Attach the last 500 lines of your syslog > as root, execute: tail -500 /var/log/messages | gzip - > syslog.gz > and attach syslog.gz (maby there is something else wrong with your sys) s. attachement > + Describe your hardware configuration AMD 64 with running 32 Bit SUSE 10 beta 3 installation, details see attachment hardware.
Created attachment 48032 [details] hardware.txt
This is not the whole KDE but more specifically, the kicker as far as I can tell. I have the same crash handler at exit but if you look closely, it complains about the kicker. I diff'd Sven's crashlog from mine. All entries are the same except for the address location of the line. I'll attach mine also. This also happened to me in 9.3 but only happens to me at log out. 9.2 was fine. Currently using SuSE 10 BETA 3 32 Bit on P4.
Created attachment 48247 [details] backtrace of kicker crash at exit
Bug is still valid with beta 4 (AMD64 with 32 Bit system installed)
crash in XIM - are you guys using skim or any kind of different input method?
Created attachment 49099 [details] hardware info
Created attachment 49100 [details] strace by pid on kicker There aren't any XIM running, I am also on amd64 with 32-Bit SuSE Beta4
Created attachment 49347 [details] Bug is still valid for RC1
Does it due to the patch of qtimmodule?
How do you like this one? http://cvs.mandriva.com/cgi-bin/cvsweb.cgi/SPECS/qt3/qt-3.3.4-update-qimmodule. patch
Thanks to Wang, it works for me now.
Add kde-maintainers@suse.de to CC:.
I cannot reproduce this either (AMD x86_64 SMP machine running RC4).
Sven Schmidt> Thanks to Wang, it works for me now. I don't understand this. Please elaborate! What did you do to make it work? The patch mentioned lin comment #15 by Funda Wang is already applied to our qt3 package, see the patches fix-key-release-event-with-imm.diff fix-x11-immodule.diff which contain exactly the same as the patch mentioned in comment #15 *plus* the hunk --- src/kernel/qapplication_x11.cpp +++ src/kernel/qapplication_x11.cpp 2005/01/27 09:41:16 @@ -3289,11 +3289,7 @@ // filtering opportunity first to ensure all input methods work // properly regardless of application design. -#ifndef QT_NO_IM_EXTENSIONS if( keywidget && keywidget->isEnabled() && keywidget->isInputMethodEnabled() ) { -#else - if( keywidget && keywidget->isEnabled() ) { -#endif if( ( event->type==XKeyPress || event->type==XKeyRelease ) && sm_blockUserInput ) // block user interaction during session management return TRUE; So what did you do to make it work for you Sven? Did you remove the qt-immodule patch: qt-x11-immodule-unified-qt3.3.3-20041203.diff ?
I just applied the patch from Funda Wang and build a new qt-rpm. Afterwards I installed it via YaST and the crash disappeared.
Did you start with the suse-spec file for qt3? I don't understand how you can apply Funda Wang's patch because as I wrote in comment #15, it is already applied in the SuSE qt3 package. If you try to apply it again, the patch program must fail.
Yes I did, I took the SUSE qt3.spec file from the SUSE qt.src.rpm and I put in this patch as an additional entry. The outcome were 4 new RPM packages: qt3-3.3.4-11.3.i568.rpm qt3-debuginfo-3.3.4-11.3.i568.rpm qt3-devel-3.3.4-11.3.i568.rpm qt3-devel-doc-3.3.4-11.3.i568.rpm I`ve installed the first 2 on my RC1 (also build on that machine). I had to download the following additional packages from opensuse.org to do so: qt3-3.3.4-11.3.src.rpm qt-3.3.4-update-qimmodule.patch update-desktop-files-0.0-192.noarch.rpm cups-devel-1.1.23-7.i586.rpm
*** Bug 144643 has been marked as a duplicate of this bug. ***
Created attachment 69358 [details] Tracelog I have still this trouble in full updated SUSE 10.
*** This bug has been marked as a duplicate of 151831 ***