Created attachment 47656 [details] standard Xorg conf This bug seems to affect multiple distributions, across many different X and kernel versions, across all graphics hardware (Nvidia / ATI / Intel), across all driver implementations (proprietary / open). This bug has been reported many times over the last couple of years, but is normally (incorrectly) pinned to a specific graphics chipset, when it is evidently a more generic issue relating to Xorg / kernel interaction when a DRI/3D/Composite configuration is active. N.B. Without 3D/Desktop Effects activated, this bug DOES NOT occur. However, I would like to use 3D effects, as this year is 2011, and I have some high-end hardware (Lenovo W510), and I don't want to feel like I'm living in the digital stone age when all the Windows and Mac OS X people are having fun with their whizzbang effects. Please, Xorg devs - can we nail this? My particular setup (though not very relevant): - Ubuntu 11.04 64-bit - uptodate - KDE 4.6.2 - Gnome 2.32.0 - nvidia-current 270.41.06-0ubuntu1 - nouveau 0.0.16+git20110107+b795ca6e-0ubuntu7 Symptom: - hard X freeze with mouse cursor still working - box normally still SSH-able Replication procedure: - use default single-screen X config - use desktop in normal way, web browsing etc. No particular app or activity provokes the issue - the crash can come any time after logging into the X session Using following desktop setups: - KDE + Kwin (desktop effects enabled) - Gnome + Compiz - using either nvidia-current (proprietary) or nouveau graphics driver. Workaround: - use Gnome+Metacity or KWin with Desktop Effects disabled. Typical trace in Xorg log: 24110.967] [mi] EQ overflowing. The server is probably stuck in an infinite loop. [ 24110.967] Backtrace: [ 24110.999] 0: /usr/bin/X (xorg_backtrace+0x26) [0x4a2626] [ 24110.999] 1: /usr/bin/X (mieqEnqueue+0x1f4) [0x4a18e4] [ 24110.999] 2: /usr/bin/X (xf86PostButtonEventM+0xaf) [0x47dccf] [ 24110.999] 3: /usr/bin/X (xf86PostButtonEvent+0xdd) [0x47de9d] [ 24110.999] 4: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f3152885000+0x4c81) [0x7f3152889c81] [ 24110.999] 5: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f3152885000+0xb268) [0x7f3152890268] [ 24110.999] 6: /usr/lib/libutouch-grail.so.1 (grail_pull+0x382) [0x7f31530a5342] [ 24110.999] 7: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f3152885000+0xb034) [0x7f3152890034] [ 24110.999] 8: /usr/lib/xorg/modules/input/synaptics_drv.so (0x7f3152885000+0x6d7e) [0x7f315288bd7e] [ 24110.999] 9: /usr/bin/X (0x400000+0x6c657) [0x46c657] [ 24110.999] 10: /usr/bin/X (0x400000+0x124bbe) [0x524bbe] [ 24110.999] 11: /lib/x86_64-linux-gnu/libpthread.so.0 (0x7f3159631000+0xfc60) [0x7f3159640c60] [ 24110.999] 12: /lib/x86_64-linux-gnu/libc.so.6 (__select+0x13) [0x7f315863a123] [ 24110.999] 13: /usr/bin/X (WaitForSomething+0x19b) [0x45c76b] [ 24111.000] 14: /usr/bin/X (0x400000+0x2e032) [0x42e032] [ 24111.000] 15: /usr/bin/X (0x400000+0x21a7e) [0x421a7e] [ 24111.000] 16: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xff [0x7f315857aeff] [ 24111.000] 17: /usr/bin/X (0x400000+0x21629) [0x421629]
(In reply to comment #1) > This bug has been reported many times over the last couple of years, but is > normally (incorrectly) pinned to a specific graphics chipset, when it is > evidently a more generic issue relating to Xorg / kernel interaction when a > DRI/3D/Composite configuration is active. You have to be careful to differentiate between symptom and cause. The line > [mi] EQ overflowing. The server is probably stuck in an infinite loop. is a generic symptom which merely means the X server is taking a long time to process client requests, but it doesn't say anything about the cause. The most common causes are indeed specific to the GPU hardware/driver, e.g. a GPU hang resulting in the CPU spinning forever waiting for it to make progress.
Thanks for the extra info here Michel. In fact I had seen other notes about the "[mi] EQ overflowing" message being a generic one. However the rest of the stack trace is still noteworthy, I think. And there is a reliable replication procedure as well, at least on my current system: KDE+KWin, Desktop Effects enabled: Open Konsole, try re-sizing from any corner. Instant X freeze! I don't have a similar procedure for Gnome yet .....
Is this still an issue? Please provide your full Xorg.log which will include information on what video drivers you are using.
No, this is no longer an issue. Or rather, I no longer have this trace occuring in X logs, but I do have regular freezes with X (on up-to-date 64bit Ubuntu 11.10) with no error logs in Xorg log, syslog, or even any ability to dump register info etc with Magic SysRq facility. (When X freezes, eveything stops: mouse cursor stops moving, CapsLock stops working, everything on screen if frozen - only way out is a Hard Reset) Only seems to happen when I'm on battery, but I realise that's not much of a bug report. ("My computer freezes sometimes when on battery power.") Unfortunately, without logs or specific errors this is one of those critical bugs that nobody can or does give a monkey's about.
Created attachment 52561 [details] nvidia bug report on X server freeze during opengl app
I have seen X server freezes repeatedly when using opengl applications with the Xorg.0.log showing a backtrace after the "EQ overflowing message". Details are posted in the Debian bug #642497: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=642497 [with info about the specific packages used and how to reproduce] Additional info also posted on the NVIDIA forums: http://www.nvnews.net/vbulletin/showthread.php?t=167084 I found the crash below today running /usr/lib/xscreensaver/circuit: [ 93.931] (II) No input driver/identifier specified (ignoring) [ 7923.634] nvLock: client timed out, taking the lock [ 8068.591] nvLock: client timed out, taking the lock [ 8097.178] nvLock: client timed out, taking the lock [ 8224.735] nvLock: client timed out, taking the lock [ 8352.518] [mi] EQ overflowing. The server is probably stuck in an infinite loop. [ 8352.518] Backtrace: [ 8352.533] 0: /usr/bin/X (xorg_backtrace+0x37) [0x80a7f47] [ 8352.533] 1: /usr/bin/X (mieqEnqueue+0x1d1) [0x80a2231] [ 8352.533] 2: /usr/bin/X (xf86PostMotionEventM+0xb0) [0x80c9130] [ 8352.533] 3: /usr/bin/X (xf86PostMotionEventP+0x6f) [0x80c927f] [ 8352.533] 4: /usr/lib/xorg/modules/input/evdev_drv.so (0xb4ad5000+0x2cfe) [0xb4ad7cfe] [ 8352.533] 5: /usr/lib/xorg/modules/input/evdev_drv.so (0xb4ad5000+0x3e1d) [0xb4ad8e1d] [ 8352.533] 6: /usr/bin/X (0x8048000+0x6e081) [0x80b6081] [ 8352.533] 7: /usr/bin/X (0x8048000+0x1282ef) [0x81702ef] [ 8352.533] 8: (vdso) (__kernel_sigreturn+0x0) [0xb786c400] [ 8352.533] 9: (vdso) (__kernel_vsyscall+0x10) [0xb786c424] [ 8352.533] 10: /lib/i386-linux-gnu/i686/cmov/libc.so.6 (nanosleep+0x20) [0xb75b23e0] [ 8352.533] 11: /lib/i386-linux-gnu/i686/cmov/libc.so.6 (usleep+0x3c) [0xb75e166c] [ 8352.533] 12: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0xb501d000+0x4540e1) [0xb54710e1] [ 8352.533] 13: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0xb501d000+0x455871) [0xb5472871] [ 8352.533] 14: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0xb501d000+0x45bdc9) [0xb5478dc9] [ 8353.519] nvLock: client timed out, taking the lock [227640.543] (WW) NVIDIA(0): WAIT (2, 6, 0x8000, 0x00004fe4, 0x000051c4) The full nvidia-bug report for this problem is attached.
Marking as fixed per 3 comments above. The followup comments describe a bug in the nvidia drivers, so you should work with them or try nouveau. In any event, that is a separate issue with the same symptom.
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.