Created attachment 33692 [details] Syslog after using wine After starting graphical windows applications using wine (or even wine-based applications like winecfg) my system becomes unresponsive for about 30s-5min. During this time mouse-movement is jerky and applications aren't painted and don't react at anymore. This happens with intel-2.10.901, but did also happen with intel-2.9.1 on my system. happens with - kernel-2.6.32.9-67.fc12.i686 - intel-2.10.901 (but also with 2.9.1) - FC12 + updates-testing
Created attachment 33693 [details] xorg log after using wine
Created attachment 33694 [details] lspci output
Forgot to mention, this happens on my Toshiba-Tecra-A8 with i945GM. After rebooting, the boot process takes unusual long with messages like: "Invalid EDID" beeing displayed at startup several times. To make this behaviour to go away, I have to completly power down (not reboot) the laptop.
Created attachment 33695 [details] [review] Cache EDID probes I'm not sure what is causing the sudden onslaught of EDID probing, presumably it is hotplug gone mad, but why? And to make matters worse it seems like we trigger a h/w bug by probing too ofter? This patch may help workaround the issue by caching the earlier EDID results.
Also of relevance is the patch by Yakui to ignore invalid EDID results, c38 bug 23099: http://bugs.freedesktop.org/attachment.cgi?id=33486
still happens with 2.6.34 + intel-2.12, but the blocking duration is now way shorter than it was and also syslog isn't flooded with "Invalid EDID" messages. I made a vide moving konsole, while starting a wine application: http://www.youtube.com/watch?v=DKAFxS5DRx4
jap, still there (2.17-git + linux-3.2.1)
Does it still happens with 3.4-rc2?
Hi Eugeni, It still happens with linux 3.3.1 - I'll give 3.4 a try when its released to fedora updates testing repo.
just tested with 3.4_rc2, unfourtunatly the problem still persists
Hmm, watching the movie... It looks like these problems could also be explained by UXA hitting software fallbacks (readbacks through the GTT), are the symptoms as severe with SNA?
Assuming the bug is still present and not alleviated by SNA, it should be fixed by Daniel's reworking of hotplug detection. *** This bug has been marked as a duplicate of bug 49906 ***
Fixed, I can't reproduce the issue on kernel-3.6.1 + intel-git =)
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.