Hi! Marcin Szewczyk <firstname.lastname@example.org> (03/03/2010): > It's probably the case (it sounds like I knew what I'm writing > about), because after a crash (on a text vt) I can see the same > backtrace as previously (at least the function from the frame #9 > doesn't appear). I'm a bit lost here. Trying to get things properly, can we talk about only gdb traces, so as to make sure we don't mix everything together? So my question: Can you now obtain two different gdb backtraces now? The patch I've just cherry-picked should help (that's for the extra function that gets called in the “new” trace) getting rid of the crash entirely. It could be that: - My first patch does something, but that's not sufficient. - My first patch serves no purposes at all. In order to test that, I've included both patches for now, and hopefully you might get a working system. If that's the case, I'll probably provide you with an extra set of packages with only the cherry-picked patch. Depending on the outcome, we should learn whether my first patch was useful at all. (So that I can forward it to either xorg-devel@ or /dev/null.) I've prepared binary packages for i386 including both my first patch, and the cherry-picked one. You can fetch them from: http://people.debian.org/~kibi/packages/xorg-572268/ (Those packages will probably go away once I'm done with this bug.) Installing both xserver-xorg-core and xserver-xorg-core-dbg should be sufficient. Mraw, KiBi.
Description: Digital signature