Created attachment 39554 [details] i915_error_state file (kernel debugfs) System environment: -- chipset: i945GM (45GM) -- system architecture: i686 -- xf86-video-intel: 2.13.0 -- xserver: 1.8.2 (xorg-server) -- mesa: 7.8.2 -- libdrm: 2.4.22 -- kernel: 2.6.35.6 -- Linux distribution: Chakra Linux -- Machine or mobo model: Dell Latitude E5500 -- Display connector: VGA (+ laptop screen, dual head) Reproducing steps: This happens only if I enable the KDE desktop effects. I can not really reproduce it, but it happens from alone when triggering desktop effects as the dashboard, sometimes it even freezes switching windows sometimes. Additional info: I'm using the "default" KDE 4.5.2 desktop effects for the intel chipset. Even though, we all know that the intel drivers are buggy in this concern, I still wanted to report it.
Created attachment 39555 [details] intel_gpu_dump (sorry, zip file, was too large ...)
Created attachment 39556 [details] vbios.dump
Created attachment 39557 [details] dmesg.txt Seems as if the driver still reacts, allthough, xorg does not. Moving the mouse works properly, just the rest ... does not.
I'm sorry to say this, you've attached an awful lot of information, none of which is relevant. The i915_error_state is interesting, but is a separate bug during ringbuffer initialisation that I have not seen reported before. What we don't have are any details pertaining to the hang during KDE.
What kind of information would you need? I now updated to 2.6.36, but this problem still occurs in some situations, and is .. more than less .. reproducable, so I can get backtraces of anything if needed. I also have some graphic glitches in example using "show all desktop windows", which is an effect on its own.
I can now 100 % reproduce it, but it makes me think on "what" is going wrong could also be connected to kwin. I filed a upstream bug at kde, maybe I find more answers there. https://bugs.kde.org/show_bug.cgi?id=254954 Thomas Lübking writes this may be a bug in mesa, I'll check that as well - and report back.
Updated to xorg-server 1.9.1, mesa 7.9 (final) and it works like charm.
Would seem to be have been a GPU hang fixed by the update to xorg and mesa. My guess would be it was one of the SRC_COPY_BLT bugs that tried to read from outside the source extents.
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.