Since xf86-video-intel-2.6.0 I am experiencing weird X behaviour - starting almost any program after logging in to the X session will cause X to crash and restart or randomly freeze (switching to a vt via ctrl+alt+f1 and pressing ctrl+alt+del to reboot works in most cases, pushing the power button to shut down the computer works too - I don't mean instantly power off by holding the button a few seconds but pressing it shortly to make the computer shut down correctly). The problem still exists with driver version 2.6.2. The 2.5.* branch works. I don't have the logs of the machine right now, please tell me what you need and I will upload them. (dmesg is 100% clean, Xorg.log shows video ram detected as "-1" and a segfault of drv_intel.so everytime the server crashes) Hardware: Intel GMA X4500HD (Asus P5Q-EM board) Software: Ubuntu 8.10 Kernel 2.6.29-rc6 Standard Ubuntu 8.10 Xorg xf86-video-intel-2.6.* libdrm-2.4.5 mesa-7.3 / current git-master-repos tried as well
xorg.conf and X log.
dmesg Xorg.0.log backtrace from gdb
How do I use gdb? I read the man, but can't think of a way to start the X session with dgb (I'm using the default Debian/Ubuntu /etc/init.d/gdm script to start/stop the X session). Please let me know how I have to use gdm under these conditions to produce usable output/information. PS: I won't have access to the machine until Fri March 6th 19:00 GMT.
http://www.x.org/wiki/Development/Documentation/ServerDebugging
Created attachment 23587 [details] dmesg
Created attachment 23588 [details] Xorg.0.log
Created attachment 23589 [details] gdb backtrace
Created attachment 23590 [details] Xorg.0.log (after freeze)
Here's the information you requested. I even tried with the latest xf86-video-intel (2.6.3) and kernel 2.6.29-rc7, the problem still exists.
Peter, I'm afraid you forget to issue bt and/or bt full in gdb to actually get the backtrace. Would you be able to provide that info? Thanks.
Could you be a bit more specific on this? I am not a developer or skilled programmer, I know a few basic things about writing software with C and can read C code, but thats it. And I never used gdb before, so what exactly did I do wrong or maybe miss to do? I suppose I have to do 'bt' when I hit the bug, but what exactly do you mean by "full" bt? There's nothing about it in the gdb howto mentioned in post #4. Sorry for my lack of understanding here!
Created attachment 24889 [details] gdb backtrace (bt + bt f) I suppose this is what you are looking for, after Xorg crashes I did a "bt" and "bt f" on the gdb command line. See attachment. Software: Kernel 2.6.29.1 (tried 2.6.30-rc2 as well, same output there) xf86-video-intel-2.7.0 libdrm-2.4.9 Mesa-7.4 Please tell me if you need anything else, I'd really like to see this bug fixed! Btw: The graphics look all messed up with black lines, parts of the screen drawn somewhere else (like the gdm login textbox has the background in it and I can't see what I'm typing). After logging in it looks even worse and crashes within a second or so.
Is there any news/progress on this bug? Or maybe a patch I might want to try? This bug is quite annoying since I can't run kernels newer than 2.6.27 and I'd really like to see it fixed! Do you need any additional information?
Interesting backtrace. That code path is now completely different and so marking this as closed for lack of obsolete. Peter, if I am wrong and this is still a problem, please upload another bt with the current drivers.
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.