Hide Forgot
Created attachment 479855 [details] output of gdb --pid `pidof spicec` Description of problem: after a clean install of Win XP guest, I switched to full screen and back and then: - spicec window disappeared - X resolution was kept at guest's 800x600 instead of returning back to 1920x1080 Version-Release number of selected component (if applicable): spice-client-0.7.3-1.el6.x86_64 spice-server-0.7.2-4.el6.x86_64 qemu-kvm-0.12.1.2-2.147.el6.x86_64 xorg-x11-server-Xorg-1.7.7-28.el6.x86_64 How reproducible: didn't try to reproduce
Created attachment 479856 [details] spicec.log
Created attachment 479857 [details] X log
Created attachment 479859 [details] dmesg output
Created attachment 479860 [details] qemu-kvm command line output
Created attachment 480129 [details] bt of another crash in main.cpp:34 I got another crash in main.cpp:34, this time when switching rhel6/32 guest from console to X. Command line for qemu was: qemu-kvm -boot order=c -net nic -net tap,ifname=$TAP,script=no -smp 2 -m 1024 -name rhel6_32 -enable-kvm -spice port=3000,disable-ticketing -soundhw ac97 -vga qxl -device virtio-serial -chardev spicevmc,id=vdagent,debug=0,name=vdagent -device virtserialport,chardev=vdagent,name=com.redhat.spice.0 /dev/root_vg/q-rhel6_32
Created attachment 480165 [details] full backtrace, otherwise the same as in previous attachment
Splitting this bug because it involves two issues: * one is the bad parameter which we (Hans, myself) couldn't reproduce. * the other is the cause of the hang, which looks like we are taking the XLockDisplay in a non main thread, and then in the main thread it gets stuck on it. The proposed solution is to keep a counter on outstanding XLockDisplay, and if it's non zero in the X error handler don't do any further X operations. If you manage to reproduce this bug (this is left as the bug for the bad parameter) please reopen it.