|Summary:||lockup with i915_wait_irq ERROR|
|Product:||xorg||Reporter:||Brice Goglin <brice.goglin>|
|Status:||RESOLVED FIXED||QA Contact:||Xorg Project Team <xorg-team>|
|i915 platform:||i915 features:|
Description Brice Goglin 2008-03-16 06:00:22 UTC
Francesco Poli reported this bug at the URL above on the Debian BTS. He has an Intel 965G integrated graphics chipset (on an ASUS P5B-VM motherboard). He can reproduce a lockup with a) Inside an X session, open an xterm b) run /usr/lib/xscreensaver/stoner view c) expand the screensaver window (so that it takes almost the whole screen estate) Also, opening a new window above the screensaver one (or otherwise sending an existing window above the screensaver one) seems to trigger the bug. The kernel log contain: Feb 24 17:23:23 homebrew kernel: [drm:i915_wait_irq] *ERROR* i915_wait_irq: EBUSY -- rec: 46388219 emitted: 46388226" He was initially using driver 2.1.0 but also reproduced recently with 2.2.1 on a 2.6.24 kernel on amd64. The corresponding log is available inside http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=25;bug=467319
Comment 1 Wang Zhenyu 2008-03-16 19:38:06 UTC
which debian package includes /usr/lib/xscreensaver/stoner in sid? I can't find it somehow...
Comment 2 Wang Zhenyu 2008-03-16 19:39:42 UTC
oh, found it in xscreensaver-gl.
Comment 3 Wang Zhenyu 2008-03-18 01:51:49 UTC
I can somehow reproduce this with xorg7.3 + mesa-7.0 branch, but can't produce with new drm kernel module 1.12 and xserver-1.5 + mesa master branch. This is clearly a 3d dri issue, reassign to haihao.
Comment 4 haihao 2008-05-14 00:01:16 UTC
fixed in mesa7.0 branch (709f24adbb21419881f0d857c8454814f85c2757)
Comment 5 Brice Goglin 2008-06-15 11:24:03 UTC
Unfortunately, Francesco reports that the bug isn't fixed in mesa_7_0_branch (Debian's Mesa 7.0.3-2 contains git commits upto 03447de3). See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=467319#71
Comment 6 haihao 2008-06-15 19:33:25 UTC
Could you try the commit 82a0e82232d61a75fee39d50333016d78938450f? So far I can't reproduce it.
Comment 7 Brice Goglin 2008-06-16 21:55:37 UTC
Right, Francesco tried 82a0e822 and couldn't reproduce: "And I was *not* able to reproduce the bug. No crash at all, even after about 30 min of stress tests (with huge amount of xterm windows opened in front of the stonerview window and then closed)." thanks haihao