GNOME Bugzilla – Bug 663948
Applications abort with XI_BadDevice
Last modified: 2011-12-30 04:57:45 UTC
I have received this bug on the Debian BTS; I found several occurrences of it on other distros: http://bugs.debian.org/648562 https://bugzilla.redhat.com/show_bug.cgi?id=728044 https://bugs.launchpad.net/ubuntu/+source/gnome-session/+bug/882956 https://bugs.archlinux.org/task/24096 I’m forwarding this here to track all those bugs, because so far gtk+ has been blamed. The symptoms are that applications using gtk3 (including gnome-session) can exit suddenly in some cases: - suspend/resume cycle for some people, - dysfunctional keyboard for one occurrence. Errors look like this: (evolution:14840): Gdk-WARNING **: The program 'evolution' received an X Window System error. This probably reflects a bug in the program. The error was 'XI_BadDevice (invalid Device parameter)'. (Details: serial 63439 error_code 160 request_code 143 minor_code 48)
Still waiting for a usable stacktrace from where that X error happens.
Created attachment 201771 [details] Stack track of x-session-manager which was stopped in gdk_x_error.
Created attachment 201772 [details] kernel message at the time of the error In case it helps, I also attached the kernel messages which appeared at the time of the GDK error. Though I don't know the exact time when the gdk_x_error was catched.
Hello, unfortunately I don't know how to trigger the cause of the problem. This time it took a few hours. I first attached the debugger to evolution, then gnome-shell and set a break point in gdk_x_error. But the session got always killed before the error was trapped. Finally I attached the debugger to the x-session-manager. If You have any suggestions what I should do instead, let me know. Cheers, Götz Gaycken
... and in gdb I tried "finish" but that terminated the x-session-manager.
Thanks for that stacktrace. I've now pushed a patch that might help, would be great if you could test that.
Goetz, this is the patch: http://git.gnome.org/browse/gtk+/commit/?id=9b967ab8aff46b79494d75a3129fa8780545ecd1 I’ll push it in gtk+3.0/unstable, but it would be nice if you could test it beforehand.
Thank You very much for the patch ! I applied the patch and it seems to have solved the problem. I had a few disconnects of the touchpad/keyboard since then, but no aborts of any of the gtk3 applications. Cheers, Götz
Considering this fixed, then.