When running compiz Xorg will crash after some period of time. I have thusfar been unable to establish exactly what user action triggers the crash (it seems many actions will). While frequently Xorg.0.log doesn't actually give a backtrace (strangely), I have attached a log where the backtrace did get written.
Created attachment 12719 [details] Xorg log with backtrace from crashed session
Created attachment 12728 [details] Another log datapoint for a crashed server This was running Mesa a8fee3a498c8c4966d57a5273408477f3aa3ce73
I've been compiling Mesa with -ggdb -g3 -O1. Can anyone see any reason why my stack traces don't have line numbers?
(In reply to comment #3) > Can anyone see any reason why my stack traces don't have line numbers? The backtrace in the log file never provides more information. Use something like gdb.
It turns out that attachment 12728 [details] was actually from another bug pertaining to compiz. In this case, compiz crashed, which then triggered another bug in the intel driver, causing Xorg to crash. This is a very reproducible behavior (for me) so I created bug 13409 to track it. Otherwise, I do not believe I've seen a crash like the one in attachment 12719 [details] since the initial incident.
Well, it seems that the apparent cessation of crashes due to this bug was a mere coincidence. Just in the last day, compiz has crashed 7 times. Unfortunately, the server rarely provides any indication of an actual error in the log and almost never provides a backtrace. I'll try and get a backtrace from gdb later today
Hi Ben, We have fixed quite some bugs in mesa those days, Could you check if the crash on the latest mesa master branch?
Hey, sorry about the delayed response. I'm on a i965 now so I may or may not be able to reproduce the problem. Regardless, I'm just struggling to even get compiz to start at the moment with the recent DRI2 churn (with DRI2 disabled I get a bunch of "failed to create drawable" errors from mesa on start). Anyways, I'll let you know when I finally get compiz working again. (In reply to comment #7) > Hi Ben, > We have fixed quite some bugs in mesa those days, Could you check if the > crash on the latest mesa master branch? >
Ben, any news?
(In reply to comment #9) > Ben, any news? > Well, I now have compiz running on the 965. I have been getting crashes with git, although it doesn't seem to be the same bug. They are pretty bad though and produce multiple sporadic xserver crashes per day; reported in Bug 16190. Seems to be a mesa bug. All in all, though, I haven't seen this bug in particular on the 965. Should we close it for an inability to reproduce or just leave it open for someone else's reference?
So I'm closing this. btw, Compiz runs fine on my 915 and 965 with current master branch.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.